BRIDGE MANAGEMENT

Inspection, Maintenance, Assessment and Repair

This volume consists of papers presented at the First International
Conference on Bridge Management, held at the University of Surrey,
Guildford, UK, from 28th to 30th March 1990.

ORGANISING COMMITTEE

PROFESSOR J. E. HARDING
DR G. A. R. PARKE
MR M. J. RYALL

INTERNATIONAL TECHNICAL COMMITTEE

MR P. ANDREWS, UK
PROFESSOR J. C. BADOUX, SWITZERLAND
MR J. D. COOPER, USA
PROFESSOR P. J. DOWLING, UK
DR H. lNGVARSSON, SWEDEN
MR D. E. LEBEK, GERMANY
DR P. LINDSELL, UK
MR K. MOIJANEN, FINLAND
MR B. PRITCHARD, UK
MR R. REEL, CANADA
MR S. SAEKI, JAPAN

SPONSORS

INSTITUTION OF CIVIL ENGINEERS

1, Great George Street, London, SWJP 3AA

INSTITUTION OF STRUCTURAL ENGINEERS

11, Upper Belgraue Street, London, SWJX 8BH

DEPARTMENT OF TRANSPORT
2, Marsharn Street, London, SWJP 3EB
BRIDGE MANAGEMENT
Inspection, Maintenance,
Assessment and Repair

Edited by

J. E. HARDING, G. A. R. PARKE
and
M. J. RYALL

Civil Engineering Department, University of Surrey,

Guildford, Surrey, UK

Springer-Science+Business Media, B.V.

First edition 1990
Reprinted 1994
© 1990 Springer Science+Business Media Cordrecht
Ursprünglich erschienen bei Chapman & Hall1990
Softcoverreprint of the hardcover1st edition 1990

ISBN 978-0-419-16050-2 ISBN 978-1-4899-7232-3 (eBook)

DOI 10.1007/978-1-4899-7232-3

Apart from any fair dealing for the purposes of research or private
study, or criticism or review, as permitted under the UK Copyright
Designs and Patents Act, 1988, this publication may not be
reproduced, stored, or transmitted, in any form or by any means,
without the prior permission in writing of the publishers, or in the case
of reprographic reproduction only in accordance with the terms of the
licences issued by the Copyright Licensing Agency in the UK, or in
accordance with the terms of licences issued by the appropriate
Reproduction Rights Organization outside the UK. Enquiries concerning
reproduction outside the terms stated here should be sent to the
publishers at the London address printed on this page.
The publisher makes no representation, express or implied, with
regard to the accuracy of the information contained in this book and
cannot accept any legal responsibility or liability for any errors or
omissions that may be made.

A catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data available

 Preface

The last 10 to 15 years ha ve wi tnessed a growing a wareness amongst bridge

engineers around the world ofthe need for a properly formulated strategy of
bridge management to ensure that the existing stock of road and railway
bridges remains in service for as long as possible. Bridges represent a !arge
capital outlay in any road or rail network, and are important not only
because of their location, but by virtue of the cost implications if their
capacity is impaired or if they fail outright.
Propermanagement is essential if this capital investment is tobe protected
and begins from the moment the bridge is first conceived until the moment it
is replaced. In the past, such management has not been properly perceived
and many bridges were constructed with little or no thought given to their
inevitable deterioration due to ageing, accidental damage, and overstressing
from increased weights and volume of traffic.
Bridge inspection of the last decade has revealed an alarming rate of
deterioration to the world's stock of bridges, and the fields of maintenance,
repair and strength assessment are growing in importance as the volume of
traffic on our roads increases and heavier vehicles are sanctioned by
government authorities. This is particularly so on secondary road systems
where many older bridges are located and on primary roads which have
attracted a density of commercial vehicles greatly in excess ofthat envisaged
in the original design.
The management task is a daunting one and very often engineers have
worked in isolation to try and provide an adequate programme of
Inspection and Repair within strict financiallimitations and with little idea
of how their Counterparts are tackling the problern elsewhere.
The aim ot' the First International Conjcrence on Bric(«e Management,
held at the University of Surrey during March 1990, was to improve
VI Preface

communication by providing a focus for the presentation and discussion of

the various management tasks facing bridge engineers throughout the world
and to help them in the formulation of a sound Bridge Management
Strategy at both national and local levels. We believe that this aim was
realized and the papers presented provide a wide pool of knowledge from
which engineers can now draw in order to develop a management system
which will work best within the framework of their own particular
constraints.
The volume contains 67 papers emanating from more than 25 countries,
covering the fields of Inspection, Protection, Structural Assessment and
Evaluation, Maintenance, Repair, Strengthening, Rehabilitation and
Performance Monitaring as well as important papers on Bridge Manage-
ment Systems currently operating in several countries.
The book provides a very useful reference manual for engineers working
with Transportation Departments, Consultants, Contractors, Regional
Councils and Local Authorities, who are interested in developing anything
from the latest 'state of the art' expert system for Bridge Management on a
cost effective basis, to the sharp end tasks of ensuring that bridges are
adequately protected against the vagaries of aggressive environments.
The editors and conference o. ganisers all come from the University of
Surrey which has had a varied interest in bridge design, construction and
performance spread over many years. A Master's Degree course in Bridge
Engineering has operated for the last 22 years. The course covers the main
areas of analysis, design and construction by means oflectures, tutorials and
Iabaratory sessions. The course structure has recently been remodelled to
include design study modules where bridge designs are covered in-depth to
latest codes of practice; and elements covering the major Bridge
Management tasks have also been introduced.
Research in Bridge Engineering at the University has concentrated on the
areas of material behaviour and structural assessment. In particular, a novel
non-destructive method of determining the residual stresses in pre-stressed
concrete bridges was developed here. The method is currently being
developed to increase its application to other types of structure and
materials such as reinforced concrete and masonry. Research is also in
progress into the viability of composite bridge design using concrete and
timber.
Also in preparation by the editors is a Manual ofBridge Engineering which
will be of general use to bridge engineers worldwide, and should provide in
one volume a mass ofinformation which at present is very widely spread and
often difficult to find. lt will include comprehensive chapters on
Construction, Analysis, Design, Management and Planning.
Finally, we would wish to offer sincere thanks to our sponsors who
 Preface vii

provided great encouragement and help with the conference; to our

Technical Committee for their patience and diligence in poring over the
many abstracts which were submitted and the comments and advice which
they tendered on each; and last, but by no means least, to our Conference
Secretary, Mrs E. Ryan who managed to cope admirably with the typing and
administration of the numerous conference documents.
 Contents

Pref'ace V

Management Systems

I. The Centenary of the Forth Rail Bridge (1890--1990) 3

0. BECKETT
2. Bridge Management--An Overview 17
K. SRISKANDAN
3. lmplemcntation of Bridge Management and Maintenance Systems
(BMMS) in Europe and thc Far East 29
A. ß. SORENSEN and F. BERTHELSEN
4. Management of the Bridge Stock of a UK County for the 1990s 39
J. PALMERand G. COGSWELL
5. Bridge Management within the Swedish National Road
Administration 51
L. LINDBLADH
6. Local Agency Experience with the Utilization of Bridge Management
Systems in Finland and the United States 63
A. R. MARSHALL and M.-K. SöDERQVIST

7. The Pennsylvania Bridge Management System 75

R. M. McCLURE and G. L. HoFFMAN
8. Data Information System for Structures: DISK 89
M. EL-MARASY
1\
X Contents

9. Optimization of Bridge Maintenance Appropriations with the Help of

a Management System-Development of a Bridge Management
System in Finland 101
A. KÄHKÖNEN and A. R. MARSHALL

10. Highway Bridge Management . 113

J. W. S. MAXWELL

11. Bridge Management in Cyprus 121

P. H. MA Y and S. VRAHIMIS

Maintenance Strategies

12. Bridge Rehabilitation: Department of Transport's Fifteen-Year

Strategy 135
D. A. HoLLAND and P. H. DA WE

13. Comparative Maintenance Costs of Different Bridge Types 145

D. LEE

14. Programmed Maintenance of Motorway Bridges: Italian Experience

in the use of 'Expert Systems' . 155
G. CAMOMILLA, A. DRAGOTTI, G. NEBBIA and M. ROMAGNOLO
15. Engineering Management of the Tamar Bridge 173
W. I. HALSE and R. L. C. STEPHENS

16. ModeHing and Predicting Bridge Repair and Maintenance Costs 187
M. BouABAZ and R. M. W. HoRNER

17. Bridge Operation and Maintenance Costs 199

H. INGVARSSON
18. Clifton Suspension Bridge: An Historie Monument that Earns its
Keep 205
D. MITCHELL-BAKER and S. CULLIMORE

19. A Systematic Approach to Future Maintenance 215

A. VAN DER TOORN and A. W. f. REIJ

20. Management of Bridgeworks Maintenance in the UK 223

N. J. SMITH

Protection

21. Crack Bridging by Surface Treatments to Concrete 233

J. G. KEER and B. H. LE PAGE
 Contents XI

22. Keeping Water Out of Concrete-The Key to Durability 243

M. B. LEEMING

23. Reinforced Concrete Bridge Protection in Northern Ireland 259

F. R. MoNTGOMERY and A. McC. MuRRAY
24. Rebar Corrosion-FBECR: The Fight to Cure the Problem 267
J. A. READ

Inspection and Monitoring

25. Experiences with the First Generation ofPrestressed Concrete Bridges

in Germany 287
B. GöHLER

26. Movable Bridge Machinery Inspection and Rehabilitation 295

C. ßiRNSTIEL

27. Application of Radar and Thermography to Bridge Deck Condition

Surveys 305
D. G. MANNING and T. MASLIWEC
28. Inspection Based Reliability Updating for Fatigue of Steel Bridges 319
A. G. TALLIN and M. CESARE
29. Diagnostic Dynamic Testing of Bridges on Brenner Motorway 327
R. FLESCH and K. KERNBICHLER
30. Experience with the Management of Cable Stayed Bridges in Korea 339
H. WENZEL

31. Performance Monitoring of Glued Segmental Box Girder Bridges 349

P. WALDRON, M. RAMEZANKHANI and B. BARR

32. Remote Computer-i\ided Bridge Performance Monitaring 361

T. D. SU)AN, J. KIRKPATRICK and A. THOMPSON

33. lnspcction and Repair of some Highway Bridges in ltaly 373

M. P. PETRANGELI

34. lnspection and Strength Evaluation of Concrete Highway Bridges in

Czechoslovakia 383
K. DAHINTER

35. Prestressing with Fibre Composite Materials and Monitaring of

Bridges with Sensors 395
R. WoLFFand H.-J. MIESSELER
Xll Contents

Assessment and Evaluation

36. Bridge Capacity Assessment and Control ofPosting, Permit and Legal
Vehicle Loads 405
F. MosES

37. The Use ofReliability Analysis in the Assessment ofExisting Bridges 417
C. MIDDLETON and A. Low

38. Strength Assessment Methods for Concrete Bridges 429

P. A. JACKSON and R. J. CoPE

39. Assessment of Stresses in Post-Tensioned Concrete Bridges 439

C. L. BROOKES, S. H. BucHNER and S. MEHRKAR-ASL

40. Assessment of Prestressed Bridge Beams 447

D. CULLINGTON

41. Fatigue Assessment of Orthotropic Steel Bridge Decks 459

C. ßEALES and J. R. CUNINGHAME

42. Assessment and Rehabilitation of Suspension Bridges 475

P. G. BUCKLAND

43. Structural Assessment of a Bridge with Transversal Cracks 489

C. ABDUNUR and J.-L. DucHENE

44. Rehability Analysis Applied to Deteriorating Bridge Structures 501

J. G. M. Wooo, R. A. JoHNSON and C. ELLINAS

45. Computer-Aided Sketching of Load Paths: An Approach to the

Analysis of Multi-Span Arch Bridges 515
W. J. HARYEY and F. W. SMITH

46. The Assessment of Masonry Arch Bridges-The Effects of Defects 523

C. MELBOURNE

47. Theoretical and Experimental Investigations on Railway Bridges

Dating from 1856 to 1895 533
F. MANG and Ö. BUCAK

48. Structural and Material Darnage to Concrete Highway Bridge Decks

in Saudi Arabia 549
M. Y. AL-MANDIL, A. K. AZAD, M. H. BALUCH, A. M. SHARIF
and 0. PEARSON-KIRK
49. Trafik Load Simulation Programme 563
D. LEBEK
 Contents xiii

50. Canada's Advanced National Standard on Bridge Evaluation 575

P. G. BUCKLAND

51. Serviceability Assessment of Masonry Arch Bridges Using Vibration

Tests 585
A. J. PRETLOVE and J. C. A. ELLICK

52. Assessing the Dynamic Properties of Existing Bridge Structures by

Hammer Testing 595
J. R. MAGUIRE
53. Serviceability Performance of a Steel Highway Bridge 607
I. RosENTHAL and M. lTZKOVITCH
54. Monitaring of Tratnc Induced Strain in the Steel Reinforcement of a
Concrete Bridge Deck 619
J. CA!RNS

Repair and Rehabilitation

55. Cracks in Steel Orthotropic Decks 633

P. MEHUE

56. An Analysis of the Behaviour of Reinforced Concrete Beams

Following Deterioration and Repair 643
J. C AIRNS
57. An Investigation into the Effectiveness of Silane for Reducing
Corrosion Activity in a Chloride-Contaminated Reinforced Concrete
Bridge Structure 655
G. P. HAMMERSLEY, M. J. DILL and J. J. DARBY
58. Bridge Strengthening Using Load Relieving Techniques 667
B. PRITCHARD

59. The Integrated Construction and Conversion of Single and Multiple

Span Bridges 677
M. P. BURKE JR

60. Inspection and Rehabilitation of Steel Trusses for Highway Bridges 695
A. G. LICHTE]';STEIN

61. The Renovation of a Yictorian Swing Bridge 705

B. SIMPSON and M. F. BLYTH
62. Cost-Effective Strategies in Bridge Management 715
R. S. REEL and C. MuRUGANANDAN
XIV Contents

63. Tension Drop in Cable-Band Bolts on Suspension Bridges 725

Y. KAGA W A and A. FUKUSHI
64. The Design of a Flexible Surface Mix for Use at Bridge Expansion
Joints 737
A. R. WooDSIDE and W. D. H. WooDWARD

65. The Repair of a Composite Concrete-Steel Bridge 747

P. H. BESEM, M. WouTERS and C. WARNON

66. Hydrodemolition-A Modern Technique of Concrete Removal in

Bridge Repair . 765
R. MEDEOT
67. Aluminium Extrusion Bridge Rehabilitation System 777
L. SVENSSON and L. PETTERSSON

Index of Contributors 785

Subject Index . 7'(1,7

MANAGEMENT SYSTEMS
 1

The Centenary of the Forth Rail Bridge (1890-1990)

DERRICK BECKETT

Schoo! of' Civil Engineering, Faculty of the Built Environment,

Thames Polytechnic, Oakfield Lane, Dartford, Kent DAJ 2SZ, UK

ABSTRACT

The Forth Rail Bridge, opened on 4 March 1890, ce!ebrates its centenary at a
time 1rhen bridge engineers are facing the need to jormulate a strategy of
bridge management to ensure that the C.'(isting stock of road and rail bridges
remains in servicej(Jr as long aspossible. The bridge has over 54 000 t ofsteel in
its superstructure with a swface areaj(Jr painting of over 6 300 000ft 2 . This
paper descrihes aspects of' the design and construction ol the hridge and the
current maintenance programme.

HISTORICAL BACKGROUND

In the period that Thomas Telford was working on major projects to

improve transportation in Scotland~the Caledonian Canal and a network
of over 1000 miles of roads~there were two proposals to provide an
alternative to ferry boats crossing the Forth between South and North
Queensferry. The first, in 1805, wasforadouble tunnel~one for comers and
one for goers. Outline details were produced but nothing further came ofthe
scheme.
In 1818 1ames Anderson, a civil engineer and surveyor, produced a design
for a three-span suspension bridge with a main span of 2000 ft, at a location
close to that of the present rail bridge. It appears that the design required
about 2500 t of iron and Westhafen 1 suggests that this quantity of iron, if
distributed over the totallength, would have given the structure a very light
and slender appearance, so light indeed that on a dull day it would hardly
have been visible, and after a heavy gale probably no Iongertobe seenon a
4 Derrick Beckett

clear day either. The project was fortunately abandoned and the same
applied to a scheme drawn up by Thomas Bauch some 40 years later to cross
the F orth 6 miles to the west ofSouth Queensferry via a series of 500-ft spans
with a total length of 2·5 miles.
In 1873 the Forth Bridge Company was formed for the purpose of
constructing a suspension bridge with two spans of 1600 ft, again to the
design of Thomas Bauch. Work on the structure was started, but on the
night of 28 December 1879 the bridge over the River Tay, completed by
Bauch some 18 months previously, collapsed during a severe gale. This
shook public confidence in Sir Thomas Bouch's design and the suspension
bridge project was abandoned.
The Bridge Company instructed their engineers, Messrs Barlow, Harrison
and F owler, to examine various forms of construction for both a bridge and
tunnel crossing. The tunnel solution was not practical and cantilever
construction was wisely preferred to the suspension form. The contract was
Iet on 21 December 1882.

DESIGN

At the time of construction of the F orth Bridge, the world's largest spanwas
John Roebling's Brooklyn suspension bridge (1595 ft), completed in 1883.
Nineteenth-century engineers were aware of the problems of maintaining
the stability of suspension bridges under heavy moving Ioads (the test Ioad
for the Forth Bridge was two 900-t trains) and thus Fowler & Baker chose
the cantileverform of construction with a main span of 1710 ft. It remains
the second largest of its type in the world and, apart from James B. Eads'
St Louis bridge (the arch ribs were built out in cantilever), completed in
1873, it was the first major structure to use steel as a replacement for
wrought and cast iron. Over 54 000 t were used in the superstructure.
The principle of the cantilever form of construction was clearly
demonstrated by the 'human cantilever' devised by Benjamin Baker (see Fig.
1) .... Two men sitting on chairs extend their arms and support the same by
grasping sticks which are butted against the chairs. There are thus two
complete piers, as represented by the outline drawing above their heads. The
centre girder is represented by a stick suspended or slung from the two inner
hands of the men, while the anchorage provided by the counterpoise in the
cantileverend piers is represented here by a pile of bricks at each end. When
a Ioad is put on the central girder by a person sitting on it, the men's arms and
the anchorage ropes come into tension, and the men's bodies from the
shoulders downwards and the sticks come into compression. The chairs are
representative of the circular granite piers. Imagine the chairs are a third of a
 The Centenary of the Forth Rail Bridge 5

FrG. I. Benjamin Bakcr dcviscd this 'human cantilever', which elegantly illustrates thc
structural principlc of thc bridgc.

mile a part and the men's heads as high as the cross of St Paul's, their arms
represented by huge lattice steel girders and the sticks by tubes 12ft in
diameter at the base, and a very good notion of the structure is obtained.
The man supported by the two human cantilevers was Kaichi Watanbe, a
Japanese student of Fowler & Baker who was invited to participate in the
model to remind audiences of the Far Eastern origins of cantilever
construction.
The principal dimensions of the bridge in elevation are shown in Fig. 2
and the view looking along the bridge from the South Queensferry shore
(Fig. 3) clearly indicates the I in 7·5 batter of the central towers which gives
the structure its lateral stability. The tubular columns to the central towers
are spaced 120ft apart at the bottom and 33ft at the top. Extensive
experimental work led Baker to adopt, with some exceptions, a tubular form
for the compression members and open lattice girders for the tension
members (see Fig. 4).
In order to achieve balance for the dead Ioad ofthe outer Queensferry and
Fife cantilevers, it was necessary to Ioad the ends at the junction with the
approach viaducts to compensate for half the weight of the central girder,
plus the effects of train loading. The central or Inchgarvie Cantilever is
balanced for dead Ioad and the out of balance loading due to trains passing
over a centred girder is allowed for by making the base of the central tower
 689' 9" 145' 661' 9" 346' ' 6" 681' ' 9" 260' 1710' 145' 689' 9"
- t- ... - - .... .... - "'
.. T Elevation

South North
Oueensferry Oueensferry
(Dalmeny) (File)

Clay Inch Garvie

825' 1030' 940' 680' 1030' 825' e fixed point

prov1sion for
r: ~
= movement
•
b
·- -· ~
..,..,
;::;·
.. 1000' ;>;-
Test Load ___ - · - __ _ __ - - - - -
~_
2 X 900 TONS - - ··- - •. - - ---~- --:=-~ tc

POSITION 1
~
~
... 1000'
Testload

2 X900TONS
POSITION 2
67" 67" Plan
~ .. .JL..
r-·---~~~ r ::::::j N
s t --~-=

deflectlon measurement positions 2 3 4

F1G. 2. General arrangement of the bridge superstructure showing the fixed points and test Ioad positions.
 The Centenary ol the Forth Rail Bridge 7

FIG. 3. A vicw from the South Quccnsfcrry shorc shows the batter (I in 7·5) of thc central
towcrs .

.Fig. m. Clw$ &ction or Jliy. TJ6. Cro&s St.ction

$tn;u in C<Jntilr-u~ of Top M1!171bv.

Pi,g.118. C,O&S S.Ct.ion of mclm~d

tie.t in Conülrvua.

m:TAIU 01' ("\:STILE\ F.Rl.

FIG. 4. Details of the cantilcvcrs, tubular construction for thc struts and latticc girders for
the tics.
8 Derrick Beckett

260ft long in contrast to 145ft for the Queensferry and Fife central towers.
Thus uplift is avoided under the worst out of balance loading conditions.
The Ioads from the central girders and the internal viaduct carrying the
permanent way are transmitted via the cantilevers into the central towers. At
the base of each central tower there are four circular masonry piers. A
skewback is provided over each pier to transmit Ioads from five tubular and
five lattice members into the foundations.

CONSTRUCTION

Space permits only abrief outline of the construction of the bridge, which is
described in great detail by Westhofen. 1 The logistics of the construction
work listed at the end of the paper are staggering, even by contemporary
standards. About 60 acres were used for workshops, storage areas, etc., on
the south side of the river and a 2000-ft jetty was built out to the site of the
Queensferry piers. Additional temporary works were required on the Fife
shore. Construction of the foundations for the approach viaduct piers and
the central towers supporting the cantilever structure required extensive
work under water. Ordinary cofferdam work was used whenever possible,
but for two of the lnchgarvie and South Queensferry circular piers it was
necessary to construct pneumatic caissons. These were built on the shore and
towed into position. Each circular pier is 70ft in diameter at the base,
reducing to 49ft at the top. From low water point upwards the piers have a
granite facing to the central mass of Arbroath stone set in cement. The piers
were built to a height of about 18ft above high water Ievel and at this stage
erection of the superstructure could begin.
The superstructure was built to a height of about 100ft above high water
Ievel and then a Iifting platform was constructed to allow the towers to be
built to their full height. Astart was then made on the erection ofthe bottom
members of the cantilevers. As soon as the 'vertical' columns had reached
their full height, work on the top members of the cantileverswas started. In
the meantime, the approach viaduct piers were being raised. The 168-ft span
double-lattice girders were erected to a convenient Ievel and then lifted at the
same time as the building of the masonry piers.
The finalstagein the construction was the connection of the 350-ft span
central girders to the cantilevers. From Fig. 2 it can be seen that the central
girders were fixed at the Queensferry and Fife ends with provision for
expansion at the other ends. Since the connection of each half of the central
girders wastobe made at midspan, both ends were temporarily fixed to the
cantilevers until they were joined at the centre of the span. The expansion
ends could then be released. A half bay of the central girder was constructed
 The Centenary of the Forth Rail Bridge 9

on temporary girders attached to the bottom members of the cantilevers and

then moved back and connected to the cantilever. Work on the girders was
then continued using cranes which had moved downhill from the top
members of the cantilever arms. The connection of the two halves of the
central girders depended on the surrounding air temperature. The lengths of
the bottom booms were fixed so as to lea ve a gap of 4 in between the ends at a
temperature of 60° F.
On 10 October 1889, when the temperature was about 5SOF, alignment
was achieved with the west boom, but the gap on the east boom was still
about 0·25 in. A quantity ofwaste was soaked in naphtha and placed in the
bottarn booms for about 60ft each side ofthe gap and set on fire. Alignment
was achieved and the bolts inserted. The top booms were subsequently
connected and the final stage in the construction was to rivet up all the
connections which were temporarily bolted. Over 6 500 000 rivets were
required in the superstructure, over half of which were fixed by means of
hydraulic rivetting machines.

PROVISION FüR MOVEMENT

Observations at the time of construction of the bridge 1 indicated that

expansion or contraction amounted to about 1 ~ 0 th of an inch for each
degree temperature change for every 100ft of girder length. This is
equivalent to a coefficient of expansion of 0·0000052tF. The various
'inftuences' to which the structure is exposed are: (1) expansion and
contraction by changes of temperature acting in the direction of the
longitudinal axis ofthe bridge, and to some extent also transversely upon the
circular masonry piers; (2) inftuence of the sun's rays to one side or the other
ofthe structure; and (3) wind pressure acting at right angles or nearly so to
the centre line of the bridge. The fixed points on the bridge were established
as (a) the southeast circular pier of the Fife cantilever, (b) the northeast
circular pier of the Inchgarvie cantilever, and (c) the northeast circular pier
of the Queensferry cantilever. The connection of the superstructure to the
rrmaining piers was detailed to allow sliding movement in a longitudinal
and transverse direction. The fixed points are shown in Fig. 2, and the
expansion/contraction lengths are:

(1) 825ft from the fixed point on the southeast circular pier of the Fife
cantilever to the south end of the north approach viaduct (3·6 in for
70')F temperature change).
(2) 1030 ft from the fixed point on the southeast circular pier of the Fife
cantilever to the south end ofthe north central girder and 680ft from
10 Derrick Beckett

the northeast circular pier of the lnchgarvie cantilever to the south

end ofthe north central girder. The two movements overlap, giving a
totallength of 1710 ft (7·5 in for 70oF temperature change).
(3) 940ft from the northeast circular pier of the Inchgarvie cantilever to
the north end of the south central girder and 1030 ft from the
northeast pier of the Queensferry cantilever to the north end of the
south central girder. This gives a total length of 1970 ft (8·6 in for
70aF temperature change).
(4) The movement from the northeast circular pier of the Queensferry
cantilever to the north end of the south approach viaduct is 825ft
(3·6 in for 70°F temperature change).
Thus expansionjoints were provided at the south end ofthe north central
girder, the north end of the south central girder, the north end of the
Queensferry approach viaduct and the south end of the Fife approach
viaduct. Expansionjointsare also provided in the approach viaduct girders
every second span.

TEST LOADING

The specification called for the contractor to ... test the strength of the
girders as directed by the engineer, and ifthe girders fail to bear satisfactorily
such tests as shall have been required, the work so failing shall be rejected
and replaced at the cost of the contractor ... This onerous task was
commenced on 21 January 1890. Two trains, in parallel, entered the bridge
from the south end. Each train comprised two 72-t locomotives at the head
followed by 50 wagons, each weighing 13-5 t, and a single rear locomotive of
72 t. Each train weighed 900 t in total with a total length of about 1000 ft.
The trains were moved forward until the front locomotives were three-
quarters through the central girder connecting the Queensferry north
cantilever and the Inchgarvie south cantilever (see Fig. 2, test load position

TABLE 1
Observed vertical deflections for test Ioad positions 1 and 2
(see Fig. 2)

Test Ioad Vertical deflection (in) downward at location

position
2 3 4

1 5
2
-~---~------- -- - - -
 Thc Ccntcnary olthc Forth Rail Bridge II

1). This was considered to be the most unfavourable Ioad position for the
Queensferry north cantilever. The deftections recorded are shown in Table 1.
The train was moved to Ioad position 2 (see Fig. 2) and the deftections
recorded are shown in Table I. The tests (as was the whole of the
construction period) were monitared by two Board of Trade engineers,
Major Marindan and Major-General Hutchinson. The observed deftections
were well within the ca!culated values.

THE WORKFORCE

The construction work at its peak involved 4600 men, and working on a 360-
ft high superstructure in adverse weather conditions, below sea Ievel andjor
in compressed air is a hazardous business and the construction ofthe bridge
cost 57 Jives. Surprisingly, not one death can be attributed to working in
compressed air. 2 Benjamin Baker stated ... it was impossible to carry out a
gigantic work without paying for it not merely in money, but in men's
Jives ... He also refers to the Ha wes Inn ... the Ha wes lnn flourishes too well
for being in the middle of our works, its attractions prove irresistible for a
]arge proportion of our 3000 workmen. The accident ward adjoins the pretty
garden with hawthorns and many dead and injured men have been carried
there, who would have escaped had it not been for the whisky of the Ha wes
Inn ...
The workforce was in part drawn from the shipyards and engineering
works in the surrounding area and !arge numbers ftocked in from England
and Ireland. A French engineer, L. Coiseau, who had extensive experience of
sinking foundations hy means of compressed air, assisted in the suhstructure
works. This attracted Italian, French, Belgian, Austrian and German
workers who had experience of deep foundation work.
Housing and lodgings were provided at North and South Queensferry,
and special trains ran from Edinburgh, Dunfermline and Leith accom-
modating hoth the day and night shift. Electric light was used for night work,
but there were frequent failures and productivity was not high. A paddle
steamer ran between South Queensferry and Fife, and there was a summer
service between South Queensferry and Leith calling both ways at
Inchgarvie, where accommodation was provided for 90 men during the
sinking of the caissons.
Westhafen 1 refers to ... black sheep who were mere birds ofpassage, who
arrived on the tramp, worked for a week or two, and passed on again to other
parts, making little use of their hands except for Iifting the Saturday pay
packet and wiping their mouths clean at the pothouse ...
A sick and accident club was sct up in the summer of 1883 and
12 Derrick Beckett

membership was compulsory with a contribution of 1 hour's pay per week,

the maximum contribution being 8 pence per week. Free clothing was
provided for men working on the foundations and those on the
superstructure with warm clothing at a nominal charge. Shelters and dining
areas were erected on the superstructure, and wages were paid by the
contractor until any injured man returned to work.
There were several strikes instigated by the organising committees of
various trades unions, and an outbreak of smallpox. The Hougomont,
originally used to transport and store cement, was converted to a hospital
ship and towed to an isolated position.

MAINTENANCE

It is a tribute to the designers, contractors and subsequent maintenance

teams that a century after completion of the bridge no major structural
repairs have been necessary. Some bracings and angle-plates, corroded by
steam and smoke, have been replaced. It was estimated that the total internal
(tubular members) and external surface area to be painted is about
6300000ft 2 (145 acres) and the original procedure was as follows.
All the steel components on passing through the fabrication shops or
yards were scraped with steel scrapers and wire brushes and then coated with
boiled linseed oil applied hot. Prior to, or immediately after erection, they
received two coats of red lead paint. This was followed by two further coats
(iron oxide)-the first a primary coat of dark chocolate brown and the
second a finishing coat of bright Indian or Persian red. The paint suppliers
were as follows: at Fife, Craig & Rose's; at Inchgarvie, Calley's Torbay; at
Queensferry, Carson's; and for central girders, Wollaston's Torbay.
Internally, the tubular members received one coat of red and two coats of
white Iead paint.
One problern with the batter given to the sides of the structure isthat the
slope ofthe lattice girder ftanges will allow rain water to collect (see Fig. 4). In
locations where rain water is unable to drain, an asphalt-concrete
compound was applied to allow water to run off by gravity. Alternatively,
holes were drilled and the asphaltwas laid to falls to allow water to drain to
the holes.
Access hatches were provided to all the tubular members. Over 35 000
gallons of paint oils and 250 t of paint were required during the erection of
the bridge. Access to many parts of the structure has always presented a
problem, and the Health and Safety at Work Act has stopped the practice of
using bosuns' chairs and rope ladders.
Currently maintenance work is controlled from an office close to
 The Centenary of the Forth Rail Bridge 13

Dalmeny Station and the Forth Bridge supervisor, James Sinclair, is

responsible for a maintenance team of about 40 men 2-16 painters, 4
riggers, 2 platers, 3 welders, 2 woodworkers, 2 drillers, 2 boatmen, 1 bridge
examiner's labourer, 2 watchmen and 6 permanent way men. Considering
the risks of working at heights up to 360ft above water Ievel, the safety
record has been excellent. The last fall was in 1970 and the last fatality Mr
Sinclair's predecessor, who was unaccountably dragged along by a train
when conducting visitors along the walkway.
The author visited the bridge at the end of May 1989 and soon
appreciated that the climate can change from a gentle breeze on Dalmeny
Station platform to strong winds on the piers at South and North
Queensferry. Conditions would be worse on the bridge itself and, for safety,
hand-held anemometers are used to check wind speeds. Above a speed of
about 40 mph painting is stopped or not started. It is unusual to get more
than 100 days a year for painting exposed surfaces.
Repainting is currently taking place on the south approach viaduct and a
temporary materials hoist has been erected at the junction of the four-span
masonry viaduct and the first steellattice girder. Shotblasting down to base
meta! is being carried out behind plastic screens. A five-coat paint treatment
is currently specified and spray application has now superseded the brush. A
primer coat is applied within 4 h of reaching base meta!, and for health
reasons the use of red Iead has been banned. This is followed by two
undercoats, a coat of micaceous iron oxide and a final coat of the famous
Forth Bridge Red. lt is hoped that this specification will Iead to a 15-year
cycle by 1996. Scotrail is weil advanced with improved means of access to the
main spans, including suspended platforms to replace ropes, blocks and
tackle, and a trolley system over the top of the bridge.
Another ongoing maintenance item is the replacement ofthe longitudinal
timbers in trough girders which form an integral part ofthe internal viaduct.
The original design with teak longitudinal sleepers has subsequently been
modified. The timbers are now 28-ft long oak 'Iogs' and there are about 590
on the bridge. Replacement is carried out on Sunday when trains are worked
over a single line.
The annual cost of maintaining the bridge, about f600000 (wages and
materials), compares very favourably with other major bridges, including
the F orth Road Bridge, which employs 90 men to run and maintain it. 2
Speed Iimits on the bridge are 50 mph for high-speed trains and DMUs,
40 mph for standard diesel-hauled passenger trains and 30 mph for freight
trains. There is no doubt that the bridge will meet the BS 5400: Part 1
requirement of a u:-.efullife of 120 years, with reasonable maintenance, and
serve the Scotrail network for decades to come. However, it is of interest to
reflect on the fact that the 1983 Review of Railway Finances 3 (Serpell
14 Derrick Beckett

Report) included an option containing only 1630 route miles with no line
north of Edinburgh. Thomas Telford, John Fowler, Benjamin Baker,
William Arrol and all other pioneers of transportation in Scotland would
have greeted this proposal with incredulity.

LOGISTICS

Total Length
South approach viaduct (Dalmeny)-4 masonry arches 300ft, 10 steel
spans 1680 ft.
Three main cantilevers-South Queensferry-Inchgarvie-Fife-5340 ft.
North approach viaduct-3 masonry arches 126ft, 5 steel spans 840ft.
Total length of bridge = 8295 ft.

Heights
Rail Ievel from high water mark 158ft
Clear headway for shipping 150ft
Top of main cantilever towers from high water mark 361ft
Top of main cantilever towers to lowest foundation 450ft

Materials
Steel 54160t from the Welsh Landore Works, the Scottish
Steel Company, and Dalzells Iron and Steel Works,
Motherwell.
Rivets 6 500 000 (4270 t) made by the Clyde Rivet Company.
Granite 27 400 yd 3 from Aberdeen and Cornwall.
Ordinary stone 48 400 yd 3 quarried locally.
Concrete 64 300 yd 3 local aggregate. The Portland cement was
manufactured on the Medway, transported by sea and
stored in an old hulk, the Hougomont, with a capacity of
1200 t.
Paint Interna! and external surface area 6 316 200 ft 2 (145
acres) paint supplied by Craig & Rose's, Calley's and
Wollaston's Torbay, and Carson's. Craig & Rose's still
supply paint for the superstructure.

ENGINEERS AND CONTRACTORS

The engineers to the Forth Bridge Railway Company were Sir John Fowler,
Bart, KCMG, CE, and Sir Benjamin Baker, KCB, KCMG, CE, and the contractors
 The Centenary ol the Forth Rail Bridge 15

were Sir Thomas S. Tancred, Bart, Mr W. Arrol, Mr T. H. Falkiner and Mr J.

Phillips. On the contractors' staff was Mr W. Westhofen, who wrote the
authoritative account of the bridge's construction 1 which appeared in
Engineering on 28 February 1890.

ACK NOWLEDG EM ENTS

The assistance of the British Steel Corporation, the Director of Thames

Polytechnic, the staff of Scotrail and Craig & Rose's in the preparation of
this paper is gratefully acknowledged.

REFERENCES

I. WESTHOFEN, W., The Forth Bridge. Engineering, 28 February 1890.

2. GRANT, W. D. F. and DARGUE, L. B., The Forth Bridge: its history, construction
and maintcnance. Proc. Jnstn Civ. Engrs (November 1985).
3. SERPELL, SIR DAVID, The Review of Railway Finances and Supplementary
Volume. HMSO, London, 1983.
 2
Bridge Management-An Overview

K. SRISKANDAN

Bridge Division, Mott MacDonald Ltd,

St Anne House, 20/26 Welles/ey Road, Ct·oydon CR9 2UL, UK

ABSTRACT

Bridge management is a comparative/y neH· concept. This paper attempts to

define the term and revie~t·s the actions that constitute bridge management
during the mrious stages in the /if'e ol a hridge. lt comments on some ol the
issues and makes recommendations fin· the fillure.

INTRODUCTION

Bridgemanagement is a term covering all the actions that need to be carried

out to ensure that the bridge remains fit for its purpose throughout its design
life without the need for excessive maintenance. Generallymanagement has
been considered as beginning only when a bridge has been built and is
brought into service. However, much can go wrang with a bridge as a
result of actions and or decisions taken at the concept, design and
construction stages. lt is essential therefore that bridge management is
brought into play right at the very beginning.
In its wider sense, fitness for purpose includes safety, Serviceability and
durability. This paperwill reviewactionstaken to ensure fitness for purpose
and comment on improvements that can be made.

CONCEPTUAL STAGE

Fortunately thousands of bridges have been built successfully in the past

and therefore errors in concept arerare if one follows the well-trodden paths.
17
18 K. Sriskandan

Errors in concept are more likely when producing a new or innovative

design. Such cases need to be properly reviewed and tests undertaken where
necessary to verify the adeq uacy of the concept.

DESIGN STAGE

lt is standard practice in this country for the principles of design, including

the concept, to be approved, at least from the safety and Serviceability
viewpoint, and for the designtobe then certified and independently checked
in appropriate cases. A similar form of review of the design is carried out in
many countries. This tends to minimise errors in concept and the design of
the structure. These checks are generally concerned with the safety and
Serviceability of the structure and may not fully cover all aspects of
durability.
Until a few years ago it was considered that, provided steel in bridgeswas
painted regularly and concrete was of good quality providing adequate
cover to the reinforcement, there was no need for any further maintenance to
modern bridges. However, recent experience all over the world has shown
this to be a fallacy, and that durability and future inspection and
maintenance should also be considered at the design stage. This includes the
following:

(a) Design details that give rise to structural or durability problems.

(b) Design details that cause difficulties for inspection.
(c) Provision of access for inspection.
(d) Aspects of durability that arenot covered by 'deemed to satisfy' rules
in design codes.
(e) Use of appropriate materials-whole life cost versus initial cost.
(f) Design for maintenance.

Typicai examples of each of the above are:

(a) Welding details that give rise to fatigue cracks, parapet base plates
that are fixed in depressions in the concrete plinth, etc.
(b) Lack of verticai and lateral space between superstructure and
substructure:
(i) to inspect bearings, and
(ii) to inspect (and paint) ends of steei beams and also to inspect the
underside of expansion joints.
(c) Inspection and maintenance gantries for !arge bridges and provision
ofwalkways, etc., or runway beams to provide access; manholes from
boxes to top of very tall piers to inspect bearings.
 Bridge Management-An Overview 19

(d) Specifying strength, water/cement ratio and cover to reinforeerneut

is not sufficient to ensure durable concrete. Improvements in
permeability and resistance to freeze-thaw cycles can be achieved by
suitable design of the mix. Designers should specify the mix and
associated curing required.
(e) Apart from cement replacement materials, it is also possible to use
polymers, etc., to improve durability of concrete, and protective
coatings on reinforcing steel and concrete to further enhance
durability. The value for money ofthisextra expenditure should be
tested by considering whole life cost of the structure, taking account
of future maintenance and traffic delay costs, etc., in carrying out the
maintenance.
(f) It is now accepted that many parts of bridges from bearings,
expansion joints, waterproofing, surfacing, parapets, etc., to hangers,
cable stays and external prestressing tendons all have tobe replaced,
some more than once during the lifetime of the structure. Account
should be taken ofthis at the designstage so that replacements can be
carried out with ease and with minimum delay to traffic.

CONSTRUCTION STAGE

lt is necessary to ensure that materials and components manufactured on

and off the site, and in fact the whole of the construction process, is carried
out in accordance with the specification.
The concept of quality assurance (QA), which originated in the
manufacturing industries, is being introduced into the construction
industry. Quality assurance is not worth doing unless it also includes quality
control, which has traditionally, in the UK, been exercised by an
independent party, such as the engineer. lt is therefore essential that
responsibilities are clearly defined if quality assurance is called for within a
contract.
The extent of the control that can be exercised will depend on the type of
contract. In a normal three-part arrangement, aspects of durability which
are not specified in codes, etc., can be introduced into the design and
specifications following discussions between the client and the engineer.
However, in cases where the design is also competitive, such as in a design
and build contract, performance specifications may not be adequate for
durability requirements. All aspects affecting durabilitywill have tobe very
closely specified in the design brief.
Going onestepfurther to the finance, design, build and operate cases, the
client will need to be satisfied that the bridge is maintained adequately
20 K. Sriskandan

during the concession period and will be in good condition when it is handed
over at the end of it. Sufficient and suitable clauses need tobe written into the
contract to ensure this.
Whatever the type of contract, every structure will need tobe maintained
from the time it is brought into service. Invariably the parties concerned with
maintenance are different to those who carried out the design and/or
construction, even if in some cases they are from the same organisation. lt is
necessary, therefore, for the maintaining parties to have a full set of the 'as-
built' drawings and maintenance schedules which will indicate frequency of
inspection and maintenance for each part of the structure. lt should
highlight areas where something has gonewrang during construction and
which therefore require special monitoring. In large structures individual
parts should be referenced, as should accesses for inspection.

IN-SERVICE STAGE

Once a bridge is completed and brought into service, it not only starts to
carry traffic but is also exposed to the environment. lt is subject to wind, rain
and temperature changes, and also to chemical and (in some cases) biological
attack. In time deterioration may occur and/or the bridge may have to carry
heavier traffic loading than it was originally designed for.
Bridge management, in order to cope with these problems, not on one but
thousands or even tens or hundreds of thousands of bridges, cannot be a
haphazard reaction to something that happens. It must be a systematic
consideration of all the problems to ensure best value for the money spent.
The first requirement is information about all the bridges on the network
in the form of a bridge inventory. Most ofthe developed countries and states/
counties have an inventory of their bridges, some of which are more
complete than others. However, the information stored varies from the very
basic to many items of information which are peripheral to the bridge
structure itself. The aim should be to plan what information is required for
use and store only that, and not each and every item which might be thought
could become useful in the future.
The next requirement is to know the state ofthe bridges at any given time
and this can only be obtained by bridge inspection.

Inspection and Testing

Until about 20 years ago there was no systematic inspection ofbridges. So
much so that in the UK inspection was treated as part of bridge
maintenance. Following the OECD Report on Bridge Inspection, 1 most
 Bridge Management-An Overview 21

countries have adopted a strategy based on a hierarchy of bridge inspection

as follows:
(a) General inspection which is a visual examination at not more than 2-
yearly intervals but without the use of special access equipment.
(b) Principal inspections which are a detailed examination of all exposed
parts ofthe structure from within touching distance, generally at not
more than 6-yearly intervals.
(c) Special inspections are again a close examination but of specific parts
of the structure following the findings of another inspection or some
special loading event such as passage of abnormal Ioads, ftoods,
heavy wind or mining.
In the UK, the Department of Transport has published its requirements
for the inspection and records of its bridges 2 and also an Inspection Guide. 3
Up to now it has been considered that special investigations including
non-destructive testing need to be carried out only as part of special
inspections. However, recent experience has shown that concrete in
structures and the reinforeerneut and/or prestressing are subject to
considerable deterioration due to carbonation, sulphation, chloride ingress
from de-icing salts and/or alkali aggregate reaction. It is therefore necessary
to undertake some testing as part of the principal inspection. This should
include testing dust samples for chloride and sulphate content, haif-eeil
potential measurements. cover metre survey and carbonation tests. Where
necessary concrete cores should be obtained for strength tests and also to
confirm from thin slices the presence of AAR if any map cracking is
observed on the surface.
When inspecting steel bridges it is necessary to check whether corrosion is
taking place. In some of the older meta! (cast iron, wrought iron or steel)
bridges the road surfacing is laid on fill material support by troughing or
hogging plates or jack arches. In all cases the top ftange is hidden andin the
last the web is also enclosed. Therefore the road pavement and part ofthe fill
will need to be excavated for a proper inspection.
On older bridges and even in many modern bridges, particularly the
smaller ones, there is no permanent provision of walkways and the like for
inspection. The increased recognition of the need for inspection from close
quarters has created a demand for different types of access equipment. There
are many types oflorry-mounted aerial platforms capable of carrying two to
three people up to heights of 60 m. Theseare generally used to inspect bridge
soffits and facias where standing space is available underneath the bridge.
Where this is difficult, such as over water or over railways, over and under
platforms with a reach of up to 14m and more are available.
Inspection reports are generally written down on standard forms and then
22 K. Sriskandan

transferred to a computer data bank. lt should be possible to produce hand-

held data capture devices which would facilitate this process.

Maintenance Management Systems

The bridge inspection will identify the areas where maintenance is
required for all the bridges in the network, but what it cannot do is to
identify the most cost-effective remedy in each case and then rank them in
order of priority.
There are many bridge management systems in operation, but few claim
to carry out explicit economic analysis to decide the best maintenance
operation. In most management systems the type of maintenance is selected
by experience and judgement, and priority between bridges and different
maintenance actions is based on rankings for various factors which are
assigned subjectively and then added up to give an overall ranking.
However, there are certain routine maintenance operations, such as
cleaning of drainage channels, washing down bearing seatings, etc., which
are obviously good value for money, especially in bridges where chlorides
have not permeated into the concrete. This work should be carried out on all
bridges. Frequency and extent of each operation should be obtained from
trials and a code of practice drawn up for routine maintenance.
For other maintenance work such as repair of deteriorating concrete,
cost/benefit analysis of different maintenance strategies needs to be carried
out to determine the optimum treatment. This is best done on a computer
and when associated with the full bridge data base it becomes a proper
bridge management system.
In order to compare the merits of different maintenance options for a
given problem, it is necessary to know the maintenance profile and hence the
whole life cost for each option. The whole life cost will be actual costs of
maintenance plus the cost of delays and additional operating costs plus the
increased costs of accidents, all discounted to present values. The benefits
have tobe taken as the disbenefits that would accrue ifthe particular item of
maintenance was not done, also discounted to present values.
In order to evaluate the various options it is necessary to know the
following:
(a) The maintenance profile for each option, i.e. the selected mainten-
ance treatment, its cost, its effectiveness, subsequent actions and
costs, and so on.
(b) Trafik delays and operating costs, and accident costs.
Assessments need to be carried out to determine suitable maintenance
treatments. Realistic values for costs, etc., can only be obtained from historic
records. It is essential that maintenance histories are stored in the computer,
 Bridge Management-An Overview 23

along with the bridge inventory and inspection records, to form the total
bridge data base.
The costs related to traffic and accidents should be obtained after carrying
out a traffic redistribution analysis given the extent and duration of closures
for a given maintenance action. This again is a very complicated
computation. The Department of Transport have written a program
(QUADR0) 4 which is the ideal tool to use in cases where the traffic-related
costs will be high. Some simplification should be possible for smaller
schemes.

Structural Deterioration
The degradation of paint and other protective treatment and the
subsequent corrosion of steel is the most common form of deterioration in
structural steelwork. Inspection combined with regular local repair and total
repainting before all the paint films break down will reduce the cost of
maintenance and preserve the service life of the structure.
The other common defect is fatigue cracking. Again regular inspection
combined with NOT methods are necessary to detect these cracks. It may be
possible to arrest the progress of the cracks temporarily by various means,
the simplest of which is to drill a hole at the end of the crack. However, in
highly stressed areas proper repairs will be necessary. lf the detail is repeated
many times in a structure, it is weil worth while testing the fatigue life of a
proposed repair detail in the Iabaratory before implementing on site.
Cracks due to brittle fracture occur in steels which do not have the
necessary notch ductility. Since the much publicised failure of Kings Bridge
in Melbourne, 5 designersareweil aware oftbis problem. However, there are
many old steel bridges in existence which could weil fail by this method in
very low temperatures. They should therefore be inspected following a
severe cold spell.
The process of corrosion of steel embedded in concrete is weil known, and
many papers have been written on the subject. 6 - 8
Corrosion due to carbonation is generally due to Iack ofcover and/or high
water/cement ratios. Testing for depth of carbonation is quite a simple
process; however, if early action is not taken, it may result in general
corrosion and subsequent spalling requiring repairs over !arge areas.
General corrosion due to chloride attack will also Iead to expansion of
steel followed by spalling, all of which can be seen before the integrity of the
structure is afTected. However, local pitting corrosion could result in a
significant loss of steel area without any outward sign. F or this reason it is
necessary to carry out some investigations such as testing dust samples for
chlorides at various depths and taking half cut measurements to determine
whether there is risk of corrosion in the steel. U nfortunately there are no
24 K. Sriskandan

positive non-destructive means of determining whether corrosion is taking

place and if so at what rate. Therefore selective examination will be
necessary to determine the extent of the corrosion.
The need torepairsmall areas of concrete has produced a large number of
so-called 'repair materials'. Care should be exercised in the selection of
repair materials. Wherever possible it is best to use concrete similar tothat in
the parent structure. Coarse aggregates could be of smaller size than that in
the original concrete.
Deterioration due to alkali-silica reaction (ASR) has also been a
comparatively recent phenomenon in bridges in the UK. This has created
much research, and guidance has been published on the diagnosis, 9
avoidance in new structures 10 and management of structures affected by it.U
All of this cannot be properly summarised in this overview paper, except to
state that ASR is probably not as serious as first thought.
Corrosion of tendons in post-tensioned concrete occurs when salt water
gets into ungrouted or improperly grouted ducts. There is a paucity of non-
destructive methods to detect such corrosion. Radiography can be used to
detect whether ducts are fully grouted. However, this method is expensive
and has health hazards and cannot therefore be used for routine
investigations. If visual surveys and potential measurements indicate that
corrosion may be taking place, the prestressing ducts can be exposed by
careful drilling at selected places. This again is not recommended as part of
routine inspections.

Assessments
Structural assessments are required if a structure has deteriorated, been
subject to aceidentat damage, and/or is called on to carry a greater loading
than it was designed for. The greater loading may be due to a general increase
in vehicle weights and/or traffic densities or merely for a single event.
In the UK, the recent need for assessments arose from the increase in
vehicle weights from 32 to 38 t in the early 1980s and the subsequent increase
in European vehicle weights to 40 t. The Department ofTransport produces
its Standard for Assessment (BD21/84), 12 which was later amended to take
in the 40-t lorry.
1t is generally agreed that assessments should be done to some Iimit state
method. In its Standards mentioned above, the Department has advocated
the use of the relevant parts of BS 5400 for the assessment of concrete and
steel structures. The publication of the Department's Standards and the
deterioration of concrete bridges have generated a whole series of con-
ferences in the UK and their proceedings 13 - 15 contain interesting papers
on the subject. International organisations such as CEB and the OECD
have also produced publications 16 · 17 on the subject of concrete bridges.
 Bridge Management-An Overview 25

lt is now considered that the use of design codes may be too conservative
for the assessment of existing structures. Some consider that the
Serviceability Iimit state need not be considered at all. If the assessment is
required due to increased loading, the Serviceability Iimits which are
required for new designs should be maintained for existing structures also. If
the increased loading included a significant increase in axle weights, the
fatigue life of certain details might be affected. Some assessment should be
made ofthe fatigue life that has already been used up so that an estimate can
be made of the time when close inspection of the structure for fatigue
darnage should commence.
lt is also believed that the uncertainty about material strengths and
weights in existing structures is reduced. If records exist of the plates and
other members of a steel structure, it may be possible to make a better
estimate of its characteristic strength. A !arge number of cores need to be
taken from a concrete structure if its characteristic strength is tobe updated.
Weighing a whole structure by jacking it off its bearings is one means of
determining its true weight and hence reducing the partial factor on self
weight for assessment.
If upper bound methods and non-linear methods of analysis are used to
analyse the Ioad effects, these should be multiplied by a factor greater than 1
(y.f~ in BS 5400). Complex equations for the strength of a member are
sometimes simplified in design codes. Where necessary the basic equations
could be used in the assessment.
In deteriorated structures, allowance has to be made for the effects of
deterioration. In corroded steel structures, the member size should be
measured and the assessment carried out on the basis of the reduced section
properties. Judgements need to be made about corrosion of steel in
reinforced and prestressed concrete structures.
It is also advocated that probability methods be used in the assessment of
major structures. In Ref. 16 it is recommended that the safety Ievel for the
actual current use of existing structures should be equal to the one generally
required for new structures. The economic considerations that apply to the
safety Ievel of existing structures are different to the economic consider-
ations that apply for new structures. lt is therefore the author's contention
that there could be a difference in safety Ievels for new and existing
structures, especially when existing structures are being called on to carry
greater traffic loading.
If the assessment shows that the structure is not capable of carrying the
fullloading, either weight restrictions or lane closures should be introduced
to ensure that the totalloading that could come on to the bridge does not
exceed the assessed carrying capacity. In extreme cases it may be necessary to
close the bridge altogether.
26 K. Sriskandan

Increases in Loading
From time to time there is pressure, on the grounds of efficiency, from the
haulage industry and vehicle manufacturers to increase the gross vehicle
weight and also axle weights. The effect of this on highways and bridges are
considered before increased weights are permitted. Wherever possible axle
Ioads and spacings are adjusted so that the total effect on bridges is not
increased. Clearly general increases in vehicle weights need to be managed in
this way if bridges are to remain fit for their purpose.
There is, of course, illegal exceedence of vehicle and/or axle loadings
Enforcement, using weighbridges at selected spots, can be a deterrent; the
resource effort required for total eradication can be quite high. Therefore
data should be collected about these overloaded vehicles and allowance
made either in the loading or the partial factor on loading.

CONCLUSIONS

1. Bridge management covers the whole period from concept to ultimate

demolition or replacement.
2. At concept and design stage consideration should be given not only to
safety and serviceability but also to durability, inspectability and ease of
maintenance, including the replacement of various parts of the structure.
3. At every stage the whole life cost of the various options should be
considered. This should include costs of traffic delays and other similar
indirect costs.
4. Some investigations should be carried out during the principal inspection.
5. Bridge maintenance management systems should include an explicit
economic analysis of the various maintenance options.
6. Routine maintenance is obviously good value for money. There should be
a code of practice for routine maintenance which should give frequency
and standards for such maintenance.
7. There should be a code of practice for the assessment of existing
structures.
8. Axle weights and spacings should be controlled when gross vehicle
weights are increased.

REFERENCES

1. Road Research: Bridge Inspection. OECD, Paris, 1976.

2. Trunk Road Management and Maintenance Notice: Trunk Road and
Motorway Structures-Records and Inspection (TRMM 2/88). Department of
Transport, London, 1988.
 Bridge Management-An Overview 27

3. Bridge Jnspection Guide. Department of Transport, HMSO, London, 1983.

4. Queues and Delays at Roadworks (QUADR02). Department of Transport,
London, April 1982 (and subsequent revisions).
5. Report ofthe Royal Commission into the Failure ofKings Bridge. Government
Printer, Melbourne, 1963.
6. The Durability of Steel in Concrete. Part 1: Mechanism of Protection and
Corrosion (Digest 263); Part 2: Diagnosis and Assessment of Corrosion
Cracked Concrete (Digest 264). Building Research Station, Garston, UK, 1982.
7. VASSIE, P. R., Reinforcement corrosion and durability of concrete bridges. Proc.
Jnstn Civ. Engrs, Part 1 (August 1984) p. 76.
8. PULLAR-STRECKER, P., Corrosion Damaged Concrete: Assessment and Repair.
CIRIA, Butterworths, London, 1987.
9. PALMER, D., The Diagnosis oj' Alkali-Silica Reaction-Report oj' a Warking
Party. British Cement Association, Slough, 1988.
10. Alkali-silica reaction-Minimising the risk of darnage to concrete: guidance
notes and model specification clauses. Technical Report 30, Concrete Society,
London, October 1987.
11. Structural effects of alkali-silica reaction-Interim technical guidance on
appraisal of existing structures. Institution of Structural Engineers, London,
December 1988.
12. The assessment of highway bridges and structures. Departmental Standard
BD21/84, Department of Transport, London, March 1984.
13. Assessment of reinforced and prestressed concrete bridges. Papers presented at
a seminar organised by the Institution ofStructural Engineers, September 1988.
14. Concrete bridges-Management, maintenance and renovation. Proceedings of
one-day conference, Concrete Society, London, February 1989.
15. Bridge assessment symposium, Leamington Spa. Construction Marketing, June
1989.
16. Diagnosis and assessment of concrete structures, Bulletin d'Information
No. 192. Committee Euro International du Beton, Lausanne, January 1989.
17. Road Transport Rescarch-Durahility o(Concretc Road Bridges. OECD, Paris,
1989.
 3

Implementation of Bridge Management and

Maintenance Systems (BMMS) in
Europe and the Far East

ANDERS B. SoRENSEN and FINN ßERTHELSEN

CO Wlconsult Consulting Engineers & Planners AS,
45, Teknikerbyen, DK-2830 Virum, Denmark

ABSTRACT

This paper describes the DANBRO bridge management and maintenance

system (BMMS) and the aspects of implementing the BMMS at bridge
authorities.
The system prouides a rational and systematic approach to organising and
carrying out activities related to bridge management. lt consists o{
-A database system.
- Three modules:
• The inventory module.
• The inspection and heuring capacity module.
• The ranking and hudgeting module.
-Manuals for a/1 activities.
The implementation of'BM M S is illustrated by two case studies: the Danish
State Railways (2500 bridges) and the Department ofHighways in Thailand
( 10 000 hridf{es).

INTRODUCTION

lt is the objectives ofthe BMMS to give the bridge authority a tool that helps
to:
-Ensure the safety and network capacity.
-Ensure objective information on all bridges.
--Optimise utilisation of allocated funds.
-Ensure technical-economical feedback.
The BMMS is run on personal computers and minicomputers.
29
30 Anders B. Sorensen and Finn Berthelsen

The DANBRO BMMS has been developed in cooperation between the

Danish Road Directorate, the Danish State Railways and COWiconsult.

CONCEPT OF THE SYSTEM

Figure 1 shows the main activities in the system.

General
Manualsare available for all the activities shown in Fig. 1. Suchmanuals
are especially needed during the inspections in order to ensure objectivity in
the inspector's evaluations. Results from the inspections must obviously be
comparable from bridge to bridge if the ranking of structures in need of
repair are to be reliable.
Information on a bridge is categorised (administrative, geometry,
materials, condition, etc.) and associated to a specific element. The elements
are arranged in a hierarchic order as shown in an extract from the BMMS in
Fig. 2.
Elements are selected at the highest Ievel for which the information is
representative, e.g. conditions can be given for Ievel 1 (the bridge in general)
andjor for an element on a lower Ievel as required by varying conditions (e.g.
Ievel 4, element, foundation, if deteriorated to an extent different from the

lnventory Recordlng r-- Bridge Data Base

Bridge Bridge lnventory
Overvlew
Superflclal lnspec.
I / Malntenanee
r-- I I I I I
f-- I I I I I
Date

Prlnclpal lnspectlon 1- I I I I I
I I I
Special lnspectlon 1- -- Management t - Ranklng f- Budgetlng

I I I
Bridge Rating 1- lmplementatlon of Malntenance Works

FIG. I. Main activitics covcrcd by thc systcm.

 BMMS in Europc and thc Far East 31

Level I: Bridge

L
Level 2: Substructure

Level 3: Structures

II I~ II
Level 4: Pile Foundation Pier Slab Girder Truss Insulat ion

FI<;. 2. Hierarchie elcmcn\ s\rue\ure, ox\rac\ only.

bridge in general). This ensures the minimum amount of informationtobe

entered and stored in the database.

The System Modules

The invcnton modu!e
By means ofthe bridge information in the database, the inventory module
facilitates predefined output forms corresponding to routine enquiries. A
typical predefined form is shown in Fig. 3. Further, the user may create
individual reports comprising selected information.
A graphical interface is being implemented to the system. The userwill on
the screen be able to zoom "down' from a map of all the routes to
information on a specific bridge. This facility provides an easy and logical
access to the bridge information. The graphical interface is illustrated in
Fig.4.

Area Code: 610 Nakhon Ratchaaima Division Rev. Date: 31-06-01

Route No.: ALL , Control Section No.: ALL , km:
km: BK

Route ICtrliStructure 1 km I Name IFICon/1 Superstructure

ISectl PI km BK I IelChan I - design
No. 1 No. 1 No. SI I lnl I - material

Sida-Phon 1 2506 -slab

IHuaisokkabuong I I 1-reinforced concrete

19 PI352.6301Ban Sida-Phon 11125061-slab

I IHuaiphailon I I 1-reinforced concrete

1-reinforced concrete

1-reinforced concrete

FICi. 3. Bridge ovcrvic\\ (c.xamplc from the BMMS in Thailand). (Thc Thai year 2531
corrcsponds lo lhe Grcgorian year 19R8.)
32 Anders B. Sorensen and Finn Berthelsen

w ~ ~ 9
lF::I--====--D t
Administrati c data

• L;ncti on/ gcomctric data

FIG. 4. Zoom from a map of all the routes to information on a specific bridge.

The inspection and bearing capacity module

There are three types of inspections: superficial inspection, principal
inspection and special inspection.
- The superficial inspection is carried out by the local road or railway
personnel with short intervals. They clean the structural members and
repair minor darnage in accordance with manuals. The maintenance
Ievel for each bridge may depend on the importance of the bridge.
 BMMS in Europe and the Far East 33

- The principal inspection is carried out by local well-trained engineers,

who typically deal with about 500-1000 bridges. Inspection intervals
are normally 3 years with a range of 1-6 years depending on the
condition of the bridge.
The inspector evaluates the darnage and gives a condition mark to
selected elements in order to describe the condition of the bridge.
Significant darnage may be registered if required for documentation.
Further, he estimates the remaining life and corresponding repair costs
of each element, and recommends the time for the next principal
inspection. Finally, it is registered if a specia1 inspection is required.
- The special inspection is a detailed investigation of a bridge structure or
parts of it, when the condition or load-bearing capacity is about to
reach an unacceptable level. This inspection is either initiated on the
basis ofthe recommendations by the principal inspector or on the basis
of the ranking list. Labaratory tests will normally be required. The
findings in connection with an inspection will normally lead to
alternative maintenance strategies for rehabilitation with corre-
sponding budgets (lifetime and repair costs for elements) (Fig. 5).
The system comprises a bearing capacity programme for rating of the
bridges based on the inventory data, including the inspectors' reports on
current conditions, and for rating of actual trucks (Fig. 6).
The bridge rating programme calculates the bearing capacity of a bridge
and compares the result with the loading from a standard truck. A bridge-
bearing capacitydass is given as a percentage oftbis standard truck loading.
Material deterioration, as reftected by the condition mark, will often cause
lower material strength and thereby lower bearing capacity. The system
includes a facility for adjusting the bearing capacity in relation to condition
marks.

DEPARTHENT OF HIGHWAYS - SPECIAL INSPECTION REPORT - THIS DATE: 31/08/02

Route No.: 2 , Control Section No.: 0702 , Structure No.: 19

Stationing: km 352.630 and km BK

Repair Proposal No.: 1

Superticial Inspection/Maintenance:
Haintenance level (I minimum, II normal, III high) II
Yearly budget (1,000 Baht) 5

Repair Works in 1,000 Baht:

Year Element Hethod Estimated repair costs

Code Name (1,000 Baht)

2531 BRIDGE Bridge Re pair 850

2535 ABUWAL Abutment Walls Strenghtening 250
2541 EXPJOT Expansion Joints Exchange 60
2560 BRIDGE Bridge Exchange 3000

FIG. 5. Special inspcction rcport (cxample from thc BMMS in Thailand).

34 Anders B. Sorensen and Finn Berthelsen

DBPARTMENT OF HIGHWAYS - BRIDGE RATING - THIS DATE:31/08/07

Route No.: 2 , Control Section No.: 0702 , Structure No.: 16
Stationing: km 350.077 and km BK

Rating Oocument

Element: 3.10 Carrying Superstructure

Last Principal/Special Inspection: 31-08-02
Element'& Condition Mark: 1
Remarks:
Load Case 1:
- Load factor on Dead Load • 1.10
- Load factor on Standard Truck • 1.25
Inventory Rating Class: 133 '
Operating Rating Class: 104 '
Rating Mark: 0
FIG. 6. Results from the bridge rating on a single bridge (example from the BMMS in
Thailand).

The vehicle rating programme calculates for an actual vehicle the effects
on a set of bridg~ spans and compares the results with the standard truck's
effects on the same bridge spans. The maximum value expresses the actual
vehicle's dass in terms of the standard truck.
The system thereby facilitates the administration of heavy transports. By
a simple comparison of an actual vehide's dass with the bridge dass, it is
checked whether a given truck or a specific truck type can pass over the bridge.

The ranking and budgeting module

An initial and rough long-term budget calculation is based upon the
replacement costs for elements estimated by the principal inspection.
The condition marks and the rating mark of each bridge or its elements
form the basis ofthe ranking ofbridges in need ofrepair. The importance of
each element for the function and safety ofthe bridge and the importance of
the route is included in a model that calculates the ranking point for the
bridge.
The rank of each bridge (Fig. 7) thus reflects:

- The current condition.

- The actual bearing capacity in relation to the required capacity.
- The importance of each element for the function of the bridge.
- The importance of the route that passes over the bridge.
Alternative maintenance strategies are requested to be prepared by a
special inspection of the bridges in the top of the ranking Iist and of all
 BMMS in Europe and the Far East 35

DEPARTnENT OF HIGHWAYS - RAHKING OF STRUCTURES - REPORT DATE : 31/09/12

Ranking Point• have been calculated Year/nonth/Day: 31-09-08

Structures are reported for the folloving unit<sl only:

- Code: &10 Nakhon Ratchaaima Division

Route No. Control Structure No. Km Km No Renking Points Cond.

Section nark
No. BK Total Cond. Bear
Cap. BRIDGE
Pr Pc Pb

219 0500 14 14.9&1 3 49.2 24.& 24.& 3

2057 0100 27 1&.117 10 11.0 12.3 o.o 2

2149 0101 28 11.933 43 5.5 &. 1 o.o 2

FIG. 7. Ranking of the bridgcs in need of repair (example from the BMMS in Thailand).

bridges with a need for immediate maintenance works in order to prepare an

accurate short-term budget. Initial selection among alternative maintenance
strategies are normally based on net present value, including possible
considerations of traffic costs, imposed on users by the work programme
(detours, etc.).
The investment schedule for the selected maintenance strategies or
inspection estimates from the principal inspection are summarised for all
bridges and compared with available budgets (Fig. 8).
When discrepancies between available funds and estimated maintenance
costs are found, the user indicates from which period and to which period
maintenance costs shall be moved. The system then directs the identification

DEPARTMENT OF HIGHWAYS - BUDGET - R!PORT DATE : 31/09/12

The budget haa been accepted Year/Month/Day: 31-10-01

Budget Report tor the following unit(a) only:

- Code: 610 Nakhon Ratchaaima Division

Year 2531 2532 2533 2534 2535 2536- 2547- 2568-

Cast in mlo. Baht 2546 2567 2593

Route No. Contol Sec-

tion No.

2 0702 0.9 0.2 0.3 0.0 0.3 2 4 2

2 0800 o.o 0.0 0.0 0.0 0.0 1 3 6
2 0901 0.0 0.0 0.0 1.0 0.0 0 0 2
23 0101 0.0 0.4 0.0 0.0 0.2 3 0 0
23 0102 0.0 0.0 0.0 0.5 0.0 0 1 4
23 0103 0.0 0.0 2.4 0.3 0.0 1 0 0
202 0500 1.0 0.5 0.0 0.0 0.0 0 0 0
207 0202 0.0 0.7 0.0 0.0 0.0 0 0 1
207 0300 0.0 0.0 0.0 0.0 0.0 0 0 0
219 0500 2.8 0.0 0.5 0.0 1.0 3 2 3

FIG. 8. Final budgct for a flcld division (example from the BMMS in Thailand). (The Thai
year 2531 corrcsponds to the Gregorian year 1988.)
36 Anders B. Sorensen and Finn Berthelsen

of bridges where alternative maintenance strategies comprises maintenance

works in the 'move to' period instead of the previously selected ones in the
'move from' period.

Establishment of the BMMS

The outlines for the system were settled by a working group comprising:
~Representatives of the future users:
• Inspector.
• Coordinator of maintenance works.
• Technical assistant.
• Budget planner.
~System planner, who has long-term experience of bridge design and
maintenance projects.
~Database specialist

The definition of the system was made top-down. The user group defined
and described the main items first and detailed later. Preliminary
programmes and simple registers were prepared in order to present the
ideas:
~Firstly, the users decided the modular activities.
~Secondly, the following selection of information types was settled by
the working group as the typical data set:
• Administrative data.
• Function/geometric data.
• Material data.
• Load-hearing data.
• Condition data.
• Maintenance strategy.
~Thirdly, the users decided that all components of the BMMS should be
operational immediately after installation and thus be able to perform
ranking and budgeting even with limited information. The amount of
information that can be stored for old and damaged bridges should, on
the other hand, be unrestricted.
The above-mentioned data set was therefore used to describe a
bridge element as illustrated earlier in the paper.
~Fourthly, all activities connected to bridge management were described
in detail by the system planner and the database specialist Foreach of
the activities the following were described:
• The contents of the activity.
• Which information is needed prior to the activity.
• Which information is updated or created during the activity.
 BM M S in Europe and the Far East 37

Implementation of BMMS in the Danish State Railways

The Danish State Railways (DSB) is responsible for the maintenance of
approximately 2500 bridges on the overall railway network ofDenmark. In
the past inspections were carried out on all bridges and the reports were held
in !arge manual archives. The DSB decided to set up a BMMS in 1985 in
order to secure an updated overview oftheir bridges, the proper function of
the bridges, and to establish a ranking Iist of maintenance works and a tool
for budgeting of maintenance works. The implementation was completed in
1987.
The following activities were carried out:
-Description of the existing organisation.
-Establishment of new routines and teaching in these.
-lmplementation of EDP system and manuals.
- The existing inspection reports were entered into the database.
-A preliminary ranking Iist was calculated in order to select the bridges
where special inspections were required.
-Special inspections, including a set-up of maintenance strategies, were
carried out and an accurate short-term budget was prepared.
Implementation in the Department of Highways-Thailand
The Department of Highways (DoH) in Thailand is responsible for the
maintenance of approximately 10 000 bridges on the routes that form the
superior road network of Thailand. In the past only the seriously damaged
bridges were reported-ifidentified-and subsequent budgets for the repair
works were given. This situation caused the DoH to hold a budget reserve
each year, and the unforeseen amount varied from year to year. To manage
the maintenance budget in the most effective way and to keep the bridges in
good condition, the DoH decided to set up a BMMS.
The implementation covers all the modules described in this paper and
was initiated in November 1987. From April 1988 to February 1989 a full
inventory recording of all the DoH bridges in Thailand took place.
Furthermore, a country-wide bridge management and maintenance
Organisation has been established and the future members have been
involved in the design of the system from the very beginning in order to
secure its acceptance in the Central Administration as weil as in the districts.
The majority of the 10000 bridges have now been inspected. The total
implementation is planned to be completed early in 1992.
The following activities are carried out:
-Description of the existing organisation and routines.
-Introduction of a new organisation and new routines.
-Implementation of EDP system and manuals.
- Teaching in new routines in Bangkak and Denmark.
38 Anders B. Sorensen and Finn Berthelsen

CONCLUSION

lt is our experience that an early involvement of the users and well-planned

educational sessions are of considerable importance to a successful
implementation of a bridge management and maintenance system (BMMS).
This will be carried out in order to secure the daily use of the system and
avoid a return to 'the old way'.
The implementation both from Denmark and Thailand proves that the
BMMS willlead to:
-Establishment of an efficient up-to-date overview of the bridges.
-Consistently updated and objective information on each bridge and
easy access to it.
-Systematic evaluation of the condition carried out on the bridges but
only when needed.
-Improved basis for budgeting and maintenance planning by objective
ranking of the bridges.
-Savings from more reliable and flexible budgeting due to the
minimisation of unexpected repair works.
-Savings from a more rational administration of the bridge network
with a minimised need for inspections and collection of information.
 4
Management of the Bridge Stock of a UK County
for the 1990s

JoHN PALMERand GRAHAM CoGSWELL

County Engineer Department ol Surrey County Council,
HighH·ay House, 21 Chessington Road, West Ewe!!, Epsom, Surrey, UK

ABSTRACT

The aim ol this paper is to shmr how Surrey County Council has developed a
hridge management system in order to improve control ol the stock ol
structures so that optimum use is made of'resources and the overall condition is
enhanced.

INTRODUCTION

The bridge office in Surrey has some 2500 bridges or highway structures to
manage. These vary from traditional structures such as masonry arches and
walls through the whole spectrum to modern post-tensioned structures (see
Fig. 1).
Maintenance in the past was the Cinderella of bridge works, often
neglected or left until resources were available (or may be just left! If you
don't know there is a problern you can't deal with it!). However, ignorance
has proved not to be bliss but more like a nightmare.
There is now a growing requirement for local highway authorities to
operate on a business footing and target the limited available cash.
The majority of the problems with bridges built over the last few decades
or so are material, workmanship or detail related. It is important to record
suppliers, types and origins of materials, and details changed on site or
during fabrication.
The regular inspection process which gathers information means that the
facts have tobe assessed and action decided on. It is ofvital importance that
the right information is collected, stored and used in the correct way.
39
40 lohn Palmer and Graham Cogswe/1

NO DESIGN CODES DESIGN COO(S

173
NISCELLANEOUS FODTBR!OG(S
ARMCD, Lorge doo pope~
GANlR ! [ 5 ETC.

46 PRESTRESSEO C D N C R E T E - - - - - - - + - - - - -

347 HINFDRCED CONCRETE - - - - - IL _I________;

46 TROUGHI NG TRUSSES RSJ'~---- ~,I : r: ., . itJi j
· r. 't ' tu_~ J

2~ C O I ' t P D S l H - - - - - - -- - 1
L __ __ _ __ __ _ ;
I

95 LATTICE PLATE ANO ~

STEEL GRJDERS ~

.,, •••,.,..,••••, ••,.•, I

1780 1870 190El 1922 1970

1880 l9Hl 1983
FIG. I. Bridge deck construction materials and date for county road bridge stock (total
number 1172).

A management system is a tool that will allow the best use of resources to
ensure that the aim of effective management is achieved cost effectively. It is
important also to develop a culture within the organisation that fully utilises
all available resources with a business-like approach. The paper shows how
this has evolved in Surrey and the future strategy.

FUNDING

An historical Iook at bridge construction since 1780 in Surrey is shown in

Fig. 2.
The prosperity of the Victorian age is reflected in the gradual increase in
Lridge construction, the growth in canal, railway building and eventually
roads showing a peak at the turn of the century. The depression between the
wars resulted in a reduction in new bridge works although there were some
public works using unemployed miners in the Guildford area during this
period.
 Management of the Bridge Stock of a UK County 41

0::
<(
w
,.. a
0::
w
11.

1-
:::! 6
::J
!Dg
V1 X
- 5
w
~ I.
0::
(IJ

l!5
0
z

18()) 182~ 161.0 186('. 1800 1920 19LO 1960 1980 20:XJ
FIG. 2. Distribution of new bridge construction dates.

The bridge construction programme reached a low point in Surrey during

the 1940s and 1950s. The national boom in bridge reconstruction of the
1960s and 1970s was not reflected in Surrey until the 1980s.
Assuming a 100-120-year life, then the peak at the turn ofthe century will
be reflected in the next decade's workload.
Since the turn of the century traffic volumes have soared, and !arger and
heavier lorries are using the roads. This has increased the rate of
deterioration of bridges in Surrey.
The capital value of the bridgestock of road bridges only, i.e. the cost to
replace them, is estimated in 1989 at:
Number f(m)
Department of Transport structures 400 300
County structures 1173 250
550

Assuming an average 120-year life for each structure, then the cost per
annum to replace the stock alone is f4·6m, i.e. 0·84% of the capital value.
This does not include any maintenance costs to ensure that the average life
is reached, i.e. painting of steel, concrete repairs and other maintenance
works. The figures given in the OECD report on bridge maintenance 1 in
1989 gave the average amount spent on routine maintenance in the UK at
0·5% of the replacement value, i.e. f2·75m/annum.
The total estimated cost of replacement and maintenance based on the
above is therefore f4·6m + f2·75m = f7·35m/year, i.e. 1-4% of capital value.
lt must be stressed that this does not allow for deterioration due to increases
in traffic Ioads and weights.
42 lohn Palmer and Graham Cogsll'ell

Many ofthe older structures are exceeding the estimated design life of 120
years, but many of the newer, larger span and more sophisticated bridges,
which are more costly to repair, are not. The average age of those requiring
restrictions or reconstruction is 77 years, and if the older and inherently
stronger arches are ignored this reduces to a life of 63 years. Basedon a 77-
year life, then the sum req'uired for replacement and maintenance would rise
to nearly flOm per year.

CURRENT ALLOCATION IN SURREY

At present schemes in the capital programme (Fig. 3) that replace existing

structures, i.e. not new development on 'green field' sites and not including
the cost of associated road works, averages f0·5m/annum. The maintenance
budgets for 1989/90 is:
f(m)
DTp 0·94
sec 1·oo
1·94
+ Capital schemes 0·50
2-44

This represents 0-45% of the capital value of the stock. It compares

reasonably weil with the OECD figure for the UK in 1981 but weil under half
of the estimated amount given above to maintain the stock in a reasonable
condition.
This broad brush approach deals with average figures and average
bridges. lt assumes that similar sums have been ailocated in the past when in

§
* "0

_2,~1000
ro"O>
Ii; 1200
.0 1100

...5 0~ ~
u
~

l1i
900
800
28~
t:~"O
700
Due to DTp
~ <,1 ~ 600
moratorium
Cl~ 2 500 County
~ ~
~ ~·.;:;
5 400
ma1ntenance

~~~ ~Projected
Q) c:.~
~ expend iture
~·ro '-
wE-2
84 85 86 87 88 89 90 91 92
Fmancial year

FIG. 3. Expenditure on structures in Surrey (does not include new 'green field' construction).
 Management o{ the Bridge Stock of a UK County 43

reality the Ievel is well below current Ievels Iet alone required Ievels of
funding. There is therefore a catching-up process to deal with and fund.
This is a very subjective measure of the Ievel of funding required to
maintain the stock in good order and is often difficult to use as a persuasive
argument when fighting for more funds.
What is required is a management system that records actual cost of
maintenance in a form that can be easily interrogated and gives a
quantitative measure of the condition of the stock so that the right Ievel of
funding can be given to increase or maintain their condition. This is one of
the main elements of BRIDGIT.

MANAGEMENT SYSTEM DEVELOPMENT

Prior to 1980 the bridge management system in Surrey was not electronically
based. There was a card indexing system and individual bridge files.
Maintenance lists and programmes were all manually produced, which was
adequate at that time to deal with the relatively small budgets and work
involved. As-built drawings were not always properly made and valuable
information had been lost, particularly during the various local government
reorganisations The inspection cycle was often ad hoc or reactive to need or
public demand. This changed in 1977 with BE 4/77, which introduced the
notion ofthe principal inspections. Financial control was separate and again
largely paper based.
The increase in the number of structures to manage consequent upon the
opening of the M25 (Fig. 4) and the necessity for some form of discipline in
inspection work led Surrey to develop STREG in 1980.
STREG is a computer management system which stores information on
individual structures. lt is a batch system on the Unisys mainframe
computer which generates standard or ad hoc reports as required. Inspection
forms are prompted to a predefined programme, completed on site,
returned, maintenance work decided, prioritised and programmed.
Maintenance lists are then produced. There is some degree of very coarse
financial recording. The main financial control is at present carried out using
separate computer-based systems on a commitment accounting basis.
STREG is very much a system for the 1980s, it is electronically based and has
the security that goes with that type of system.
The next development involved converting STREG to BRIDGIT. 1t is a
more advanced management system, giving the engineer a whole
kaleidoscope of information to help him in his decisions on maintenance
work in the 1990s. The use of data capture devices that will prompt and
record inspections electronically on site will cut out the inefficiencies of the
44 John Palmer and Graham Cogswe/1

current paper-based system although it will work hand in hand with

photography, sketches and other methods of recording defects in the
inspection process. The system will be interactive and easily accessible to
view, change or update the range of information held. This will be extended
to include such things as material type, suppliers and stats information,
tagether with location maps linked to the road network. Financial
information will be stored in more detail for better future estimating and
forecasting. Site instructions for minor works will be printed by the system
direct to the maintenance contractor.
A quantitative measurement ofthe conditions ofthe individual and total
bridge stock will be obtained. The condition of each element of the bridge
will be given at the time of inspection on a scale of 1-5, with each
classification being well defined. A multiplier is applied to this to obtain the
condition factor for each element of the bridge. Each element is given a
location factor depending on its structural importance. The bridge itself is
given a road factor depending upon dass, i.e. motorway, A-class road, etc.
These three factors are multiplied tagether to give a priority rating to that
element of the bridge. These are totalled to give a condition factor for the
individual bridge and then processed to give the overallstock condition. The
key to knowing how the stock is performing is to use a rating system, as from
this flows the possibility of making statistical comparisons.
By comparing the overall condition factor on a regular basis, probably
annually, then a measure ofthe effect ofmaintenance work can be obtained.
Cash can be more easily targeted to have the greatest effect. 'BRIDGIT' will
give a good objective measure to present to government committees, etc., of
the need for extra funds.
Development of 'BRIDGIT' for the 21st century involves a link up with
CAD for as-built drawings, drawings of defects, etc., and distance data
access either in the inspector's van or home via telephone links. Direct
financial management will be within the system and direct feed of traffic,
accident and other relevant information to assist easier maintenance
decisions from the highway database. lt will also be linked directly to
digitised OS maps.

INSPECTION STRATEGY

The number of bridges on the register over the past 10 years has increased
dramatically with the opening of new motorways and from a concerted
effort to find all our bridges (Fig. 4). Unlike other counties, inspections in
Surrey are undertaken by technicians throughout the year with ~upport
from chartered engineers as required.
 Management of the Bridge Stock of a UK County 45

3000

2500

Vl 2000
~
::::l
I-
u
::::l
a:
1--
1500 j
Vl
i
u_
0 1ooo I
0
z
500

1977 78 79 80 81 82 83 81. 85 86 87 88 89 90

YEAR
FIG. 4. Structures inspectcd by Surrcy County Council, both county and DTp.

The county is split into four main areas of operation. Motorway and
trunk road bridges are the responsibility of the senior bridge inspector. He
arranges the programme and makes sure targets and Standards are met. The
rest of the county is split geographically between three inspectors/clerks of
works who report to the senior inspector. The inspectors/clerks ofworks are
responsible for all bridges in their area and making sure that they are
adequately inspected and maintained, and records kept.
Principal inspections are undertaken on a 6-yearly cycle with general and
superficial inspections in intervening years. The inspection forms are prompted
quarterly in an area so that adjacent parishes are done together. This
minimises travel costs and time. The individual inspectors plan their work
within that quarter. They have to ensure that the work is done to a high
standard and target dates met. Within a short time they gain an intimacy and
attachment (often sentimental) with individual structures.
The backhone of all the inspections is the forms sketches, and most
importantly photographs, which can tell more than any amount of written
words.
The inspectors, who tend to have a practical background, issue
instructions to the maintenance contractor for minor works, i.e. pointing,
brickwork repairs, timher deck, repairs to parapets, etc., and supervise the
works. This continues the philosophy of keeping them involved and
committed.
We are constantly finding substantial structures on rights ofway. Students
46 lohn Palmer and Graham Cogswell

have been employed for the last two summers to walk the rights ofway and
pick up qualifying structures they find. This programmewill be completed in
1989.
The cost effectiveness ofthe inspection process is constantly appraised. An
underbridge unit that gives access to three bridges in a possession (although
twice the price) may be more cost effective than a scaffold tower which
requires manpower for repeated movement. Annual tenders are let for
various categories of inspection plant.

ASSESSMENT STRATEGY

Surrey made funds available in 1986 for assessments to the new Code of
Assessment which takes account of the increased weight and numbers of
heavy goods vehicles. Eighty-four out of a possible 650 county structures
have been done to date with another 40 programmed for 1989. Acting as
agent to the DTp, 44 trunk roads and motorway structures aretobe assessed
in 1989.
This programme presented a great opportunity to thoroughly examine
the condition of the bridges. lt was decided from the start to combine a
thorough inspection and testing programme with the assessment.
This gave an intimate knowledge of the structure. This philosophy was
proposed to the DTp in 1986. They were unable to pursue the full testing
programme recommended because of shortage of funds. lt is interesting to
note that the condition report of 200 bridges by the DTp on the condition of
bridges recommends just such an approach. 2
The new Code is conservative in many areas. Structures with no apparent
structural distress, only durability problems, are being recommended for
heavy restrictions. By using the intimate knowledge gained from this
investigation areturn to 'basics' has been essential. How is the bridge acting?
What load is it actually taking? What happens if it collapses? Is further
testing required?
The bridges assessed to date are the older bridges in the county but are of a
range of materials (Fig. 5).
While by no means a representative sample because of the way they have
been selected (i.e. oldest or known problems first), it is interesting to note that
restrictions or recommended works to date are mainly concentrated on
metal bridges. The early reinforced concrete bridges fair reasonably well.
These are arches or beam and slab constructions with built-in ends, often
with no joints and their associated problems.
Ifthere are any lingering doubts then load tests can be carried out. Often a
cheap practical method of showing that the bridge is capable of carrying
 Management of the Bridge Stock of a UK County 47

D WEIGHT RESTRieTION IMPOSED OR

STRENGTHENING WORKS PLRNNED
OR OONE
30

25

<J) 20 w
w >--
'-" w
"'
0
~ w u
>-- z::
"'
Cl) w 0
u
'"-
0
15 "'uz:: 0
0 w
u
"'w u
"''"-0
"":c::> Z::I
z: 1 e

FIG. 5. Bridges asscsscd to date (undcr DTp memo BD 21/84).

imposed Ioads. Two have been done to date in Surrey at a cost of f5000. This
is not a thorough scientific examination ofthe structure but a good practical
comparison of actual response to expected response.
The average cost of assessment to date, which includes a principal
inspection, is fllOO for county bridges and f3000 for DTp bridges. Extensive
testing has also been undertaken, at an average cost of f1200 for county
bridges and f21 00 for DTp bridges.
Where restrictions have been applied the average costs have been (f/m 2 ):
(a) Weight restriction 19
(b) Strengthening works 1006
(c) Replacement (bridge works only) 3900
The latter high cost of replacement compared with average new works
cost ofbetween flOOO and f2000/m 2 is due to the substantial costs of dealing
with statutory undertakers' equipment and maintaining traffic flow.

MAINTENANCE STRATEGY

The thinking in Surrey is for good basic cost-effective schemes which

maximise available expertise and resources. The programme is flexible, as is
48 lohn Pa/mer and Graham Cogswe/1

the budget for individual maintenance schemes. Why use a complicated

design which can cause future maintenance problems?
A maintenance contractor is appointed annually following competitive
tendering. Basically a dayworks contract, the rates are usually very
competitive. On average three two-man gangs are used, one for each
inspector/clerk of works. These can be increased or decreased to suit
requirements and the work given to the contractor varies from minor works
to fairly substantial emergency works.
Larger works are Iet by quotation, using simplified conditions of contract
or normal tender by ICE, 5th edition. A 3-year programme is compiled
which allows for replacement, strengthening or maintenance works.
Wherever possible existing structures are strengthened, particularly the
more traditional types of structures, i.e. Town Bridge and Onslow Bridge,
Guildford, and Leatherhead Town Bridge. They are part of our heritage, are
often more pleasing and have Iasted Ionger than some moremodern forms
of construction.
If a new bridge is inevitable then the opportunity to widen or improve is
taken. Attention to detailing in new structures is encouraged to minimise the
durability problems of the past. A minimum cover of 40 mm is used
throughout with better quality concrete. For the future the use of epoxy-
coated rebars could make for more durable structures.

OTHER INFLUENCES

The gathering of data via the inspection and assessment process gives the
basis for technical decisions for maintenance or strengthening works. Yet
more and more of these decisions rely on other influences.
The amount of traffic dictates the type of work. The effects on the local
economy of a particularly low weight restriction, for instance, can result in
rapid reconstruction or strengthening from below. We are about to Iet our
first contract for minor works (f50 000), where the contractor will be given a
bonus for early completion and penalised heavily for being late. A very
simple lane rental approach yet vital in the future if traffic is to be kept
movmg.
The environment is an emotive issue at present and growing in
importance. Bridges are invariable Iandmarks and should be treated
accordingly.
Political pressures can be hard to accommodate, yet with weil thought-out
designs, good communication and justification the right scheme will win
through. This is tied up with the selling approach-we want people to buy
 Management of the Bridge Stock of a UK County 49

our schemes. All products need marketing, promoting and selling to often
highly articulate, well-informed and well-organised groups.

PRIORITIES

In Surrey we establish priorities based on the following factors:

(a) Need, i.e. condition of bridge.
(b) sec primary routes.
(c) Volume of traffic.
(d) Cost/benefit (including traffic delay costs and inspection/main-
tenance costs).
(e) Availability of alternative routes.
(f) Legal requirements.
(g) Environmental impact.
There are a range of tools that can be used at any particular site, as
follows:
(1) Monitoring.
(2) Further testing and investigation.
(3) Load tests.
(4) Weight or width restrictions (money has recently been given to
trading standards to appoint staff to 'police' these restrictions).
(5) Strengthening, local or major.
(6) Reconstruction to existing standard.
(7) New improved bridge scheme.
(8) Closure.
We try to assess the end of the life of a structure and then include it on a
programme for replacement. Clearly it is no use spending vast sums on a
structure with limited life. Whole life costing techniques are used to
determine the most cost-effective solution for a particular structure for a
particular time in its life cycle. The Department ofTransport has recognised
this necessity in its publication BD 36/88, which gives average maintenance
costs. The use of QUADRO for traffic delay cost is also required. Ten years
extra life in a structure may give time to construct that bypass.

APPRAISAL

lt is important at the end of the exercise to take time to reftect. Has the
money been weil spent, and the aim achieved? Could it have been done better
50 lohn Palmer and Graham Cogswell

or more cost effectively? What mistakes were made and can these be shared
so that others don't make the same mistakes? Has the targeting exercise been
successful?

SUMMARY

Bridge maintenance has been the Cinderella ofbridge work in the past, being
overridden by the demands of design work for new structures. This boom in
new works, sometimes with poor detailing and over-complex structures and
materials, has left many problems. Manifestation of these problems,
combined with improved inspection techniques and a need to rectify, has
meant that maintenance has thrown off its old rags and is now the belle of
the balltobe courted. However, unless funds are made available midnight
could strike and Cinderella will be in rags once more.
Effective maintenance relies on the gathering of information, an
awareness of all the influences and the effects of works so that the right
decisions can be made and cash effectively targeted. The management
strategy evolving in Surrey is modern and business-like, and will be
enhanced considerably once the attractive 'BRIDGIT' is up and running.
Book early to secure your copy!

ACKNOWLEDGEMENT

The authors wish to thank Dr J. Bergg, County Engineer of Surrey County

Council, for his support over the contents of this report.

REFERENCES

I. OECD Bridge Maintenance. A report prepared by an OECD Road Research

Group, 1981.
2. WALLBANK, E. J., The Performance of Concrete in Bridges. A survey of 200
highway bridges for the Department ofTransport, HMSO, London, Apri11989.
 5
Bridge Management within the Swedish National
Road Administration

LENNART LINDBLADH
Swedish National Road Administration, Head Office, Borlänge, Sweden

ABSTRACT

The Swedish National Road Administration is responsiblefor a bridgestock

comprising some 11000 hridges, representing a replacement value of some
SEK 25 hillion at the 1987 price Ievel. During the coming 10-year period the
Administration expects to invest ahout SEK I hillion per annum in
maintenance, repair, reconstruction and replacement. It is of the utmost
importance for these allocations to he utilised e.fficiently in order to be of
maximum henefit from a socio-economic viewpoint.
Within the Administration a new aid to systematic administration of the
actions required to manage the hridge stock in an optimum manner is currently
heing developed. Developments hitherto haue led to the introduction of an
ADP-hased in{ormation system, Bridge Data,facilitating a number ofmanual
administrative routines during the planning and operafing stage. Further
developments invo!ve, among other things, an overview and possible
computerisation of these manual routines, leading to a complete bridge
management system ( BMS).

BRIDGE STOCK

The Swedish National Road Administration is responsible for the

management of around 11 000 bridges included in the national road
network. In this context, a 'bridge' is a structure with a free opening of at
least 3·0 m in the Iongest span. Of these bridges, some 8500 are made of
concrete, 1800 of steel and 700 of stone. About 400 of the concrete bridges
are made of prestressed concrete.
51
52 Lennart Lindbladh

GOOO SALT- FROST

RESISTANCE
Num ber ot bridges 0/o
100
90 ·-
' /
V I

80 - -
I -+ V '
70 ---+
~
- -- - -

60 ---

~ j
so - --

-- - -- - --
--
l;T -
40 -

.,........ V I
30 - - - - r-- - /--
-· -

-t t Tl
20
10
0
1900 -os -10 -15
-r 4-
-20 -25 -30
I
-35
- -f--

-40 -45
----

-so -55 -60

I
-65 -70 -75
I
-so -8
Yea
HIGHER BEARING
CAPACITY

FIG. I. Bridge age distribution. Statc and state-municipal bridges.

The age composition reveals that about 3200 bridges were constructed
before nationalisation of the road network in 1944 (see Fig. 1). As a rule,
these bridges have a relatively low bearing capacity. It can also be stated that
some 6600 bridges were constructed before the year 1964, when the
requirement of air entrainment in the concrete was introduced, resulting in a
higher salt-frost resistance.
The load-carrying capacity of all bridges has been classified in order to
assess their trafficability by actual vehicles and vehicle trains. The
classification, which specifies the factual bearing capacity with due regard to
damage, etc., is expressedas the permissible axle and bogie loading, A/B(t).
The road network is normally capable of carrying A/ B = 10/16 t, a Ioad
which can be carried by 97% of the bridge stock.
During 1988 actions have been initiated with the aim of raising the load-
carrying capacity standard on the main road network and on the secondary
and tertiary road networks in the forest counties. The purpose is primarily to
make possible, by reinforcements, reconstruction and new construction,
adaptation of the Swedish weight regulations to those ratified by the EC.
Roughly 1300 older bridges are affected.

FINANCING

In the current maintenance plan for 1988-92, SEK 170 million per year has
been allocated at planning Ievel in the 1987 price Ievel (0-7% of the
replacement value) for bridge maintenance. In addition to this sum is SEK
60 million for reinvestments insmall bridges with an object cost of less than
SEK 3·0 million. Reinvestment is understood in this context to mean
 Bridge Management in Sweden 53

TABLE 1
Annual extent of bridge activities according to current plans
- ------------- -

Action Number of Cost Costa

(financial) bridges (SEK m/year) (%)
per year

Maintenance 170 0·7

Reinvestments
Durability 60 160 0·6
Bearing capacity 100 320 1·3
Investments 80 250 1·0
Total 240 900 3·6
-------

a In relation to replacement value.

reconstruction or new construction solely on account of shortcomings in

durability.
Major reinvestments and new investments areevident from the 10-year
multi-year plan. According to the plan valid for 1988-97, the annual volume
of reinvestment on account of shortcomings in durability is about 30 bridges
or SEK 100 million (0-4%). New investmentstotal SEK 250 million (1·0%)
or about 80 bridges.
Over and above the regular plans, a special bearing capacityplan has been
drawn up for implementation ofthe aforesaid commitment to load-carrying
capacity. From this it is evident that reinvestments will be made in about 100
bridges per year on account of insufficient load-carrying capacity for the
future EC-adapted weight regulations.
All in all, the above implies an annual bridge volume during the coming
10-year period as in Table 1.

BRIDGE MANAGEMENT

The term bridge management is used to describe all the activities, both
administrative and productive, required to construct, assign to traffic and
demolish a bridge (see Fig. 2).
The administrative activities include every stage of economic planning,
namely orientation planning, action planning and budgeting, as weil as
physical planning in the form of preliminary and detailed planning. Also
included among the administrative activities is the continuous description of
the bridge objects in the form of inspections, condition assessments,
classification of load-carrying capacity, etc. Following-up involves an
evaluation of the various actions taken from the standpoint of both
54 Lennart Lindbladh

POLITICAL

BUDGET IHG

FIG. 2. Bridge management elements.

economy and technique, and forms the basis for formulation and
reconsiderati9n respectively of road-keeping goals, strategies, maintenance
standards, performance standards, etc.
The allocations placed at the disposal of the Swedish National Road
Administration for implementation of its road-keeping assignment are and
will in all probability continue tobe more or less meagre. In view ofthis, it is
essential for the allocations to be used in such a way that they are of the
greatest possible benefit to society at the lowest possible cost. The actions to
be carried out must be technically correct, and take place at the correct
points in time and on the right objects.
In order for the optimisation tobe carried through all the way, an ADP-
based bridge management system (BMS), which will make possible
systematic administration of requisite actions on the bridges, is currently
being developed within the Swedish National Road Administration. The
system is being built up araund a database, Bridge Data, containing a large
quantity of data necessary for the BMS. Campare the description in the
section entitled 'Bridge Data'.
In addition to Bridge Data, a fully developed bridge management system,
BMS, should include routines for:
-inspections, condition assessment, load-carrying capacity classifica-
tion;
-selection of planned action, optimisation per shortcoming and bridge;
-prioritisation, optimisation per road network/bridge stock;
-specification of commonly performed maintenance and minor repair
tasks;
 Bridge ManagementinSweden 55

-economic and technical follow-up;

-reporting; and
-route finding for heavy transports.

Hitherto, developments have been concentrated on designing Bridge

Data, which is now largely complete. The ADP system also generates a
number of fixed and selected reports.
The inspection routines may be regarded as ready for use with the
introduction of new inspection regulations and a system for assessment and
documentation of faults and shortcomings.
An aid to the initiation of optimum actions per shortcoming and bridge is
also available through the preparation of a standard for assessment of
degrees of urgency for requisite actions.

BRIDGE DATA

Bridge Data is included as a subsystem in the road data bank (RDB). The
system is stored on a mainframe computer (Sperry) and is built up largely in
databases. This will facilitate the transition to base computer storage (VAX),
which is expected to take place in 1991.
Bridge Data currently consist of five components, and two new ones are
being planned. These components are:

-drawing section;
-administrative section;
-technical section;
-load-carrying capacity section;
-damage section;
-planning section (under planning); and
-projecting section (under planning).

The bridge register, which describes bridges, road tunnels, ferries and
jetties along state and state-municipal roads, contains output from Bridge
Data.
A brief description of the different sections now follows.

Drawing Section
The drawing section contains information on about 120 000 bridge
drawings. Among the stored information, mention may be made of drawing
content, format, date of drawing approval, ifit has been micro-filmed, and if
so, when.
56 Lennart Lindbladh

Administrative Section
The administrative section contains information on about 12 000 bridges
and about 1500 road tunnels on the state and state-municipal road network
(except in the municipalities of Stockholm and Gothenburg). In addition,
information is stored on roughly 6000 other bearing structures, induding
bridges, etc., which have been projected but not yet constructed or bridges
which have been reassigned to the private road network or demolished. The
administrative data indude such things as the bridge name, responsibility
for bridge maintenance, year of construction, free distance from underlying
road or watercourse, etc.

Technical Section
The technical information relates to, for instance, type of static system,
material in superstructure, spans, waterproofing and paving types, support
data, mode of foundation, type of bearings and expansion joints, bridge
area, etc.

Load-Carrying Capacity Section

The load-carrying capacity figures concern, for instance, permissible axle
and bogie loadings, trafficability data for certain heavy types of vehide,
carriageway width, bridge being repaired, etc.

Darnage Section
This section contains condition and darnage information. The inform-
ation relates to, for instance, type of inspection, condition dass for 15
different structural parts, summary condition dass for bridge, type of
darnage and where it is situated, extent of darnage and estimated repair cost,
degree of urgency for repair action, etc.

Planning Section
This section will contain in the first instance information that is necessary
for economic planning, such as planned and factual action costs, road-user
costs, index tables, discount rates, planned action times, etc.

Projecting Section
An expansion of the database with information to facilitate the physical
design of actions on the bridges has also been discussed. This section could
be used, for instance, for storage ofwork drawings, specifications, etc., which
could be used as source material for implementation of maintenance and
minor repair actions.
 Bridge Management in Sweden 57

INSPECTION SYSTEM

Information on the condition of the bridges and any necessary actions

comprises the base for the entire management system. The quality of this
information is completely decisive for the quality ofthe result generated by
the system, for instance source documentation for plans, transport permits,
etc. lt is therefore of the utmost importance for regular inspections to be
carried out and appropriately documented. To ensure this a new inspection
systemwas introduced within the Swedish National Road Administration in
1987. The rules governing implementation of the inspections are temporary
but will become definitive as from 1991.
The new rules will cover the following four types of inspection:
-superficial inspection,
-general inspection,
-principal inspection, and
-special inspection.
With regard to the intervals between different types of inspection, the
following will apply in all probability: maximum 6 years between principal
inspections and maximum 3 years between a principal inspection and a
general inspection and between general inspections respectively. In the
normal case, with time intervals of 6 and 3 years respectively, this means
alternately principal inspections and general inspections at intervals of 3
years. Superficial inspections will be carried out at least once a year.
The principal inspection comprises a new type of inspection with
considerably high er req uirements as regards extent and quality in relation to
the former full inspection. The purpose of the principal inspection is to
detect faults and shortcomings which can affect the function or trafficability
of the bridge within a 10-year period. Even such faults and shortcomings
which, ifthey have notbeendealt with within this period oftime, can Iead to
higher management costs shall be detectable. All free structural elements
both above andin the water that are accessible by means of visual methods
shall be inspected. Parts adjoining the bridge, such as road embankments,
slopes, cones, fillings and erosion protection, shall be inspected.
The principal inspection shall be carried out visually at 'hand-close
distance' and under conditions similar to daylight. Provided that anticipated
types of darnage are definitely detected, a slightly greater inspection distance,
but not more than 3·0 m, may be approved in exceptional cases under
extremely good lighting conditions. As a rule, the inspection requirements
imply that mobile inspection aidssuch as bridge Iifts, boats and ladders will
be required in addition to any fixed devices. Normally divers will be needed
for inspection of structural elements in water.
58 Lennart Lindbladh

In conjunction with the principal inspections, measurements are also

carried out in order to determine

-bottom profiles (foundation in water),

-qtrbonation depth,
---chloride content,
-plus heights of edge beams, and
-inclinations of supports.

These inspections are performed by personnel with thorough knowledge

ofthe design and mode offunction ofthe bridges and oftheir durability. The
personnel should also be thoroughly familiar with the bridge management
system used by the Administration and should be particularly qualified to
deal with issues relating to choice of actions from both the technical and the
financial angle.
The general inspection is also a new concept. In terms of extent and
quality, this type of inspection corresponds roughly to the former full
inspection. Special inspections of structural elements are carried out when so
considered necessary in order to investigate more closely any faults,
shortcomings or other observations found in conjunction with the aforesaid
regular inspections. Special inspections arealso carried out on the following
structural parts or elements, regardless of their condition:

-mechanical and electrical equipment on movable bridges,

-welds in certain steel structures,
-older bridge deck slabs of concrete, and
-foundations in water.

Any faults and shortcomings detected in the course of the inspections

shall be assessed and documented in accordance with special standards and
lists of codes. The results are registered in the Bridge Data information
system.
Allfaultsand shortcomings are described in terms oflocation, type, cause
and action. A fault or shortcoming shall also be assessed on the basis of the
functional requirements valid for the structural part in question. On the
occasion of the inspection, each structural component is allocated a
condition dass, which may vary from 0 to 3.

0 Defective function after 10 years (undamaged on the occasion of the

inspection).
1 Defective function within 3-10 years.
2 Defective function within 3 years.
3 Defective function on the occasion of the inspection.
 Bridge Management in Sweden 59

Type of damage: corrosion, scaling, spalling

Condition dass (CC): the highest CC as below is chosen.

Reinforcement area
The reduction in area (r%} is determined in one section. CC as per table.

Bonding
The reduction in bonding (r%) is determined on a length = the height of the
element. CC as per table.

Tahle r>% CC

20 3
15 2
0 I

FIG. 3. Condition standard.

Condition classifications are carried out in accordance with a special

standard. An example of such a standard is presented in Fig. 3.
In conjunction with the inspection, a record shall also be made for each
fault or shortcoming of information concerning the planned action,
including the degree of urgency and the cost. Compare below under the
section headed 'Planning-Goals'.

LOAD-CARRYING CAPACITY CLASSIFICATION

All bridges in operation have a load-carrying capacity classification as

evident from the Bridge Data information system. This classification reflects
the actual bearing capacity of a bridge with due regard to all known faults
and shortcomings.
The bearing capacity is expressed as the permissible axle and bogie
loading, A/ B (t), for a number of real classification vehicles. The
classification also includes 14 types of heavy vehicles for which the
permissible axle loadings and gross weights are specified.
If a bridge is classified as belanging to condition class 3 in conjunction
with an inspection, an investigation is carried out immediately in order to
determine the extent to which it may be necessary to reduce the permissible
traffic Ioad, i.e. the classification, ofthe bridge concerned in view ofthe faults
or shortcomings detected. This investigation may vary from a simple static
assessment to a regular control calculation. If the investigation results in an
60 Lennart Lindbladh

indication that the classification needs tobe altered this is done immediately
in Bridge Data.

PLANNING-GOALS

Prescribed in the economic plan is both the orientation road management

should ha ve d uring a period of 10-15 years and the factual actions necessary
in order to attain the road management goals. These actions are evident
from 10-year investment plans, multi-year plans and bearing capacity plans,
for new investments and reinvestments, as weil as a 5-year maintenance plan.
According to the valid orientation decision from 1986, the following
standards and actions shall be aspired to for maintenance in the planning of
road management:
- The present technical standard ofthe road network shall be maintained
(status quo).
-Heavily worn parts of the road network shall be restored to an
appropriate technical standard. The lag should largely be recovered by
the end of the century.
-Severely worn bridges should be replaced by new ones when this is
appropriate in view of the costs for replacement investment and future
maintenance.
When the faults and shortcomings established during the inspections are
documented in Bridge Data, the degree of urgency and costs, among other
things, shall be specified for the planned remedial actions. The degree of
urgency indicates the priority with which a fault or shortcoming shall be
remedied in order for the established goals for management of the bridge
stock to be attained.
The following degrees of urgency are applied:
3 Action required as soon as possible (on the occasion ofthe inspection).
2 Action required within 3 years.
1 Action req uired within 3.-10 years.
0 Action required after 10 years.
The aforesaid goals have been concretised in the form of a standard, in
which the prerequisites for selection of degree of urgency are specified for
individual actions in a number of fault situations. The costs for the most
commonly performed actions are evident from an established price Iist.
During a second planning stage the actions per bridge are optimised.
Various action alternatives, all ofwhich satisfy the requirements imposed on
the bridge, are studied both technically and economically, making due
 Bridge Management in Sweden 61

Condition index
0,70.-------------------------,
0 All counties

(2j County F
0,60

0,50

0,40

0,30

0,20

0.10

Europenn highways Primary Secondary Tertinry All

National highways county roads county roads county roads roads

FIG. 4. Functional condition ofthc bridgcs. Condition index (mean valuc ofmain condition
class) pcr road catcgory. (Thc main condition class is specified per bridge and refers to
weighted mean value of thc condition classes of the various structural parts.)

allowance also for the road-user costs. The various alternatives are
compared economically with the aid of their current values and a discount
rate of 5°/tl. The optimum alternative from a socio-economic viewpoint is
selected as a line of action for the bridge and registered in Bridge Data.
In the case of major actions, the finally developed BMS will indicate at
least one further alternative in describing the consequences in cases when it
is impossible to choose the optimum solution in view of limited funds.
The information registered in Bridge Data as above can, together with
other information in the system, be used in different ways as the bridge
management process goes on. Suchinformation can, for example, form the
basis ofprioritisation of objects, drawing-up ofnew maintenance and multi-
year plans, etc. The information can be processed and compiled in
accordance with the example in Fig. 4.
 6

Local Agency Experience with the Utilization of

Bridge Management Systems in Finland and the
United States

R. MARSHALL
ALLEN
Cambridge Systematics Incorporated, American Twine Building,
222 Third Street, Cambridge, Massachusetts 02142, USA
and
MARJA-KAARINA SÖDERQVIST
Finland Roads and Waterways Administration,
Opastinsilta 12, PO Box 33, SF-00521 Helsinki, Finland

ABSTRACT

In recent years significant efforts haue been made to collect, reuiew, manage
and analyse information on bridge structures in many countries. In many cases
these data are organized into computerized bridge management systems which
usuafly include additional analytic and programmatic capabilities beyond mere
record-keeping.for structures. These systems often are intended to define and
establish national hridge in{rastructure priorities. In many cases local
gouernment agencies haue been the primary collection agents for the uarious
data maintained in hridge management systems which include inuentory,
structural conditions and appraisal in{ormation.
The nationalusesj(Jr hridge management systems tend tofocus on network
Ievel programming andj(Jrecasting ofresource utilization and haue been widely
discussed in a uariety of different ./(Jrums and professional publications. This
paper discusses the actual uses that hridge management professionals working
at the regional or locct! Ievel hauefor BM S data hases. The paper is hased on
interviews conducted with hridge management staffin a local highway agency
and an urban transit authority in the US and 1rith distrief bridge engineers in
Finland. The case study discussion is intended to provide BMS system users
and designers ll'ith a hetter understanding oj' the positive and negative
63
64 Allen R. Marshall and Marja-Kaarina Söderqvist

consequences of the elaboration of national bridge management systems at the

grassroots Ievel.

INTRODUCTION

This paper discusses the essential ways that local agencies utilize their bridge
management systems (now andin the future) in two government agencies in
the United States andin Finnish RWA districts. Both the US government
entities and the Finnish districts have direct responsibility for the collection
and organization of BMS data as well as actual construction, rehabilitation
and bridge maintenance roles. The first discussion will highlight the uses for
BMS systems in the US and the latter section will provide an overview ofthe
Finnish situation.
Public agencies in many countries are utilizing computers in different
ways to manage information about their infrastructure, including bridges. A
variety of different information is gathered and maintained according to
agency information demands, but the basic set of information is usually
consistent to a great degree. The computerized information is used in a
variety of ways by different divisions of public infrastructure management
agencies. In spite of the growth of computerized management information
systems, in many cases the data items defined for inclusion in the
computerized systems are not totally suited to the information needs of
different bureaucratic hierarchy Ievels and responsibilities. More often than
not this mismatch is a result of the system design created by the original
bureaucratic division which provided primary impetus and funding for the
system development.

US Experience
In the United States most computerized bridge management systems have
focused at the most basic Ievel on managing the core set of information first
defined by the Federal Highway Administration (FHWA). This set of
information has been thoroughly standardized for several years. A recent
TRB research report on bridge management systems included a bridge
database design that incorporates essentially this fundamental set of
information. The initial purpose ofthe FHWA bridge database organization
was to create a national inventory with adequate administrative and
structure description information as well as basic condition information
which would enable all the states to report the nature and condition of their
bridges in a consistent fashion and to ensure that regular bridge inspections
are performed in all the states.
In most state governments there is a department of public works that has
 Brid~e Management Systems in Finland and US 65

direct state Ievel responsibility for performing bridge inspections and

transferring updated information to the FHWA. In many cases the state
agency also develops needs estimates and sets priorities for the repair,
rehabilitation and reconstruction ofbridges eligible for federal funds within
the state. The inspection reports provided to the federal government usually
provide one of the basic means of deciding on bridge infrastructure
improvement priorities.
lt is apparent that more information is required by states to perform
actual bridge system management as compared to basic structural inventory
and appraisal. The complexity of the most advanced systems in the US
(Pennsylvania, North Carolina, etc.) indicates that the information needs
defined by the FHWA fall short ofthe requirements ofthe states themselves.
These systems maintain a great deal of information and perform actual
management tasks (for example, overweight permit tracking). The
Commonwealth of Massachusetts, for instance, has extended the basic
information required by the FHWA to include specific information
necessary to perform basic bridge maintenance such as the public works
district responsible for the bridge and paint type and area.
This reftects a common federal/state bureaucratic hierarchy problem.
States have much different information requirements from those of federal
agencies. The result in the area of bridge management is that individual
states have been forced to take the Iead in developing bridge management
systems to address their particular bridge management concerns.

Finland
A somewhat different situation exists in Finland. All management of
public bridges is performed by national and regional administrative
divisions ofthe same central government agency, the Roads and Waterways
Administration (RWA). The districts and road master subdistricts actually
perform construction and rehabilitation while the national administration is
responsible for the allocation of public funds for these projects.
In many ways the Finnish bridge system administrative hierarchy is
comparable in scale at least to a single !arge US state, such as New York. The
districts and central office have always worked together closely in
establishing and carrying out bridge management programs and policy. As a
consequence, the types of information and reporting needed are relatively
consistent at the different hierarchy Ievels. Both districts and the national
office have collaborated in the development of existing bridge information
systems and the new system currently being tested in a prototype version.
This has led to the happy consequence that the types ofinformation needed
at the district Ievel are a core element of the overall information system
specification for the national database.
66 Allen R. Marshall and Marja-Kaarina Söderqvist

CASE STUDIES

Metropolitan District Commission-Boston, Massachusettst

The Metropolitan District Commission is a division of Massachusetts
state government with ownership and management of 153 miles of parkways
and 157 bridges. From 100000 to 120 000 vehicles utilize the MDC system of
highways on a daily basis since the great majority of the bridges under the
aegis of the MDC are located on urban parkways. The MDC bridge stock
includes a variety of design types and ages, including some bridges that are
exclusively utilized by pedestrian traffic to cross MDC parkways. Many of
the bridges are eligible for federal funding for rehabilitation or replacement
so there is a strong need within the MDC bridge group to maintain strict
compliance with federal inspection standards and to generate regular timely
inventory, condition and appraisal reports for federal review. Furthermore,
the state mandates a regular inspection of the MDC's bridges so that
Massachusetts as a whole will be in compliance with federal inspection
guidelines.
As part of the development of a major MDC parkways information
system,t a simple bridge database systemwas created to assist the MDC in
recording information on their bridges as well as to schedule inspections and
perform automatic reporting of federally required information. The system
was developed with substantial input by MDC bridge engineers, including
specification of information content and reports. The basic organization
starts with the federal reporting requirement and extends that data item
specification to include several administrative items required by the MDC.
The system also includes two additional files of information-a repair
project history and an itemization of bridge utilities. Thesefilesare intended
to help the MDC to track the nature and extent of repairs being performed
and to coordinate structural repairs or rehabilitation with utility companies
and other joint users of MDC structures such as the local transit authority.
MDC engineers are currently utilizing the bridge inventory for two major
purposes: recording and reporting bridge condition based on required
periodic inspections, and automatic inspection scheduling and program

t The authors wish to recognize the assistance of Mr David Lenhardt, chief bridge
engineer at the MDC, and Mr Dominic Anidi, project manager for the MBTA
Bridge Inspection Program, in preparing this and the following sections
respectively.
t The MDC parkways system, like the others described in this paper, was developed
by Cambridge Systematics Incorporated of Cambridge, Massachusetts, USA (617)
354-0167. The system consists of a comprehensive pavement management system
and bridge, traffic signal and street light inventories as weil as a capital project
development system.
 Bridge Management Systems in Finland and US 67

tracking. The MDC employs individuals with specific responsibility for

performing bridge inspections throughout the jurisdiction. These indi-
viduals have received bridge inspector training from the Massachusetts
Department of Public Works (MDPW). They utilize reports from the
computer database in the field and mark up the data items, mostly condition
items, that may have changed in the interim since the previous inspection.
These sheets are processed by an assistant engineer at MDC headquarters
and the database is updated to reflect current condition of the bridges. No
historical record of inspections is maintained by the MDC other than paper
records, although a need has been recognized to extend the current
computerized condition inventory to track inspection findings over time. No
condition items are collected other than those that are federally required but
the system does allow an inspector to enter a paragraph of comments in
order to extend and amplify the basic inspection condition report. These
comments can be viewed or printed at any time.
The inspection scheduling system is quite simple and does not optimize
the period of inspection to reflect the condition of different bridges. lt allows
either a regular 2-year inspection or a morefrequent interval to be defined.
The type of inspection (regular, in-depth, special) can also be assigned for a
bridge and this appears on the report. Two years is the current maximum
period between inspections required by the FHWA. At any time the
inspection interval can be altered to reftect deterioration or rehabilitation/
replacement of bridges. Besides scheduling the inspections, the system can
display a completion indicator as weil as an estimate of the days or weeks
that an inspection is overdue. The MDC is largely satisfied with this simple
system but they have planned to add an abilitytoset a time window for the
inspection scheduling. This would enable them to prepare, for example,
a 2-week inspection work schedule for the field inspectors at any time.
The subtiles for recording repair projects and utility tracking have been
used somewhat less than anticipated. This is due to several factors. First,
recording repair projects requires a new procedure to be implemented for
providing field reports. Currently the tracking is performed intermittently
for outside contractors and almost no in-house/own force repair projects are
recorded for subsequent review. Through time the MDC management staff
hope to develop a solid record of all repairs performed to their bridges,
including nature, extent and cost of the repair actions. They want to utilize
this information to help them decide whether a bridge should be replaced or
rehabilitated, instead of some other bridge in comparable condition because
of its historical repair log.
The M DC has completed an inventory of all utility fixtures on their
bridges as a gradual process over the last 3 years. The information on
utilities was collected as part of the regular inspections and where the
68 Allen R. Marshall and Marja-Kaarina Söderqvist

utilities were not readily visible to the eye the information was taken from
bridge drawings. The utility type, owner, approximate location and a
comment about each utility is maintained in a computer file. This allows the
MDC to easily prepare reports on any bridge that shows all co-users, the
number of utilities and any special comments related to the various utilities.
The reports can be sorted by bridge, utility type or utility owner.
The need for information on utilities is most acute for those bridges
undergoing minor repairs or rehabilitation, e.g. shoring where a complete
utility survey prior to construction is not likely to be performed. In a total
reconstruction or rehabilitation situation a utility survey will always be part
ofthe design/pre-construction work. For minor rehabilitation projects, the
MDC can anticipate any utilities that may interfere with the work and
proceed accordingly. The comprehensive utility inventory also helps reduce
or eliminate potentially problematic situations where the MDC and the
utility do not coordinate repair projects. Without an inventory it is
conceivable, for example, that a bridge that has undergone deck replacement
could have a utility company opening the newly-replaced deck to fix a line or
p!pe.

Massachusetts Bay Transportation Authority-Boston, Massachusetts

The Massachusetts Bay Transportalion Authority (MBTA) has been a
Ieader in developing computerized bridge management systems for transit,
railroad and highway bridges. The MBTA has varying Ievels of
responsibility for over 350 bridges, ofwhich approximately 100 are highway
bridges over their tracks or property. The MBTA is one of the nation's
largest transit properties with a subway and surface trolley system of 62·5
track-miles on five lines and a !arge commuter rail system of 479 track-miles
extending over 50 miles out from Boston to the north, south and west as weil
as a fleet of over 1153 buses. Daily over 420 000 people ride on MBTA
subway lines and 64 000 utilize commuter rail. An additional 435 000 daily
riders utilize MBTA bus lines.
Due to expansion of the state's infrastructure management responsi-
bilities under the Executive Office of Transportation and Construction, the
MBTA has acquired responsibility for many of the bridges and culverts on
the commuter rail lines that were formerly owned and maintained by the
Boston & Maine railroad (Guilford Industries) as weil as a number of
bridges formerly maintained and in some cases owned by other state
agencies such as the MDPW and the MDC. In many cases, particularly on
the commuter rail system, ownership ofthe structures is questionable at best
and the Authority has found out only recently that it actually owns a
number of bridges and should be maintaining them. This clarification of
responsibility has also affected the MDC. In fact the MDC has recently
 Bridge Management Systems in Finland and US 69

determined that it is the agency responsible for three bridges of great

importance to the MBTA.
The acquisition of these bridges and the clarification of ownership
responsibilities has fomented a management crisis for the MBTA bridge
engineers. A number of deficient bridges have come under MBTA control in
the last few years with extensive and expensive implications for the
Authority. As a result the MBTA Construction Directorate has launched an
extensive bridge inspection and reconstruction program which started in
earnest only in the last 2 or 3 years. This program is intended to install a
regular program of bridge inspections and initiate construction projects to
replace or rehabilitate deficient structures.
As part of the bridge management program, the MBTA has developed a
computerized database running on a microcomputer to track various types
ofdata related to its bridges. Theinformation is stored in a loosely organized
set of computer files tied tagether by primary bridge identification codes.
Each of the files is intended to address specific problems such as tracking
bridge inspection contracts or recording construction projects. The differing
characteristics of transit and railroad bridges, particularly with respect to
operating loadings, as compared to highway bridges has led the Authority to
develop new item specifications that are somewhat different from the
standard federal data items used for highway bridges. The data files are
organized into a menu system that provides an extremely simple means to
access the data and print reports but is not comprehensive and does not
perform any error checking or house-keeping tasks. No formal database
design exists, and the integration of the system is rudimentary. Most use of
the database req uires a fair degree of sophisticated knowledge of the
underlying database software.t As a result the system is utilized almost
exclusively in an ad hoc fashion.t
There are a wide variety of reports that can be generated from the system
at present. A partial Iist of available reports is shown in Table 1. lt is
apparent that several ofthe reports are in fact quite similar and differ only in
the order in which bridge information is presented, e.g. sorted by line, by type
of bridge, or some other ordering criteria. These reports are pre-set and can
only be displayed in the order shown. The final BMS will allow different sort
criteria and querying. As the system is further developed it is anticipated that
summary reports will be developed to show yearly construction figures,

t The software used by the MBTA is dBase IV, a product of Ashton-Tate

Incorporated. The system operatcs on a 80386-based microcomputer.
t Cambridge Systematics is currently documenting the existing system and will be
intcgrating and extending its capabilities so that users can easily access and update
the information and prepare reports. The MBTA project manager is Mr Dominic
Anidi, who can bc reachcd at (617) 722-5806 for further information.
70 Allen R. Marshall and Marja-Kaarina Söderqvist

TABLE 1
Typical M BTA bridge management system reports~Prototype system

Transit bridges listed by line

Highway bridges listed by line
Highway bridges that are owned by the MBTA and maintained by the
MassDPW
Transit and highway bridges owned entirely or in part by the MBTA
Transit and highway bridges that have questionable ownership or maintenance
responsibility
Inspection schedule for transit and highway bridges by month
Bridges that have not been inspected in the last 2 years
Bridges in the MBTA database that have been Ioad rated
Transit bridges ordered by condition indicators
Highway bridges ordered by condition indicators
Rehabilitation program for transit and highway bridges by Iine by year
Rehabilitation program for railroad bridges by line by year

expenditures and other summary indicators needed by MBTA management

to document progress in bridge management within the Authority.
The type of information stored will also be extended. At present no
computerized historical records of any type on previous bridge inspections,
maintenance or repair are maintained, nor does the MBTA have a
comprehensive utility inventory like that ofthe MDC. The enhancements to
the system will address these and other gaps in the information.
In addition to storing more comprehensive information about their
bridges, the MBTA is extremely interested in simplifying use ofthe system at
the same time. The current menuing system will be replaced in the near
future to allow the project engineers to easily access, update and print out
information they need. Capabilities will be provided to sort and browse
through the variety of different data items and the system will minimize or
eliminate the need for specialized computer knowledge in order to operate
the system.
It is also expected that a comprehensive inspection and construction
project monitaring system will be created as part of the system. All bridges
will be programmed automatically for inspection based on their condition
and the legally required inspection frequency as well as an inspection
priority indicator that will be developed to determine inspection job order.
Another subsystem will track the completion of the programmed
inspections. A significant portion of bridge inspections are performed by
outside contractors for the MBTA and there is a great need to track
contractor invoices and performance as well as the inspection work
performed by MBTA's own force.
 Bridge Management Systems in Finland and US 71

The MBTA also expects to utilize microcomputers to track bridge

construction projects from the design phase through final completion of
repairs or structure replacement. The computer will record standard
tracking information for each bridge project such as contractor, MBTA
project manager/engineer, project scope and budget, milestones and various
contractual information, and will particularly emphasize invoice and change
order tracking, which are of critical importance to MBTA project managers.
The prototype system currently in use at the MBTA will be replaced with a
final version by Fall 1989. A clerical staff personwill have primary day-to-
day responsibility for the system but several MBTA bridge engineers and
outside contractors will be drawing information from the databank.

Finnish Highway Districts-Finland Road and Waterways Administration

The Finnish Roads and Waterways Administration (RWA) is responsible
for almost 12 000 bridges on public roads in Finland. The maintenance and
inspection responsibilities for the bridges lie with 13 local highway districts.
These districts are shown in Fig. 1. Because of varying population and road
mileage densities in Finland, the types of bridges and bridge materials in
districts is variable and the allocations are shown in Table 2. This table also
distinguishes between bridges by their functional road classification. For

RWA Highway Distriels

CodeName
U Uusimaa
T Turku
H Häme
Ky Kymi
M Mikkeli
PK Pohjois-Karjala
Ku Kuopio
KS Keski-Suomi
V Vaasa
KP Keski-Pohjanmaa
0 Oulu
Kn Kainuu
L Lappi

FIG. I. Finnish Road and Watcrways Administration Highway Districts.

72 Allen R. Marshall and Marja-Kaarina Söderqvist

TABLE 2
Total bridges and posted bridges, tabulated by road classification and primary
construction material"

District Highway bridges Local roads All public All

roads bridges
II III IV II III IV

u 770 28 11 143 30 12 22 1 017 1 131

posted---> 2 1 9 10 14 36 36
T 694 51 31 6 224 57 39 122 1224 1 291
2 1 1 1 10 5 12 32 33
H 624 20 24 19 110 16 12 96 921 958
4 1 1 1 4 31 42 44
Ky 322 31 16 93 18 9 19 508 561
2 3 5 I 6 17 17
M 266 19 2 49 18 2 35 392 400
1 3 1 5 10 10
PK 260 17 3 2 110 17 2 49 460 480
1 5 3 3 12 13
Ku 385 18 6 12 80 19 2 80 602 619
2 1 2 2 3 10 10
KS 382 24 10 5 141 24 3 53 642 664
1 1 2 4 4
V 431 32 5 9 198 92 2 26 795 801
5 2 2 3 27 I 14 54 54
KP 293 7 3 120 68 30 521 526
4 2 6 6
0 478 15 3 I 112 60 58 727 736
I I 14 4 20 20
Kn 284 13 2 3 83 22 48 455 461
2 I 3 6 6
L 632 55 3 25 142 55 62 I 037 1047
6 4 19 27 86 86
Total 5 821 330 119 84 I 605 496 83 700
posted---> 18 II 3 10 42 95 2 124
Road type 6 354 2 884 9 301 9 675
posted---. 42 263 335 339
- - -

Material types: I, concrete; H, steel; III, stone; IV, wood.

"TVH Tuotanto-Osasto, S!LLAT 1.1.89 (annual bridgc system summary report);
TVH Tuotanto-Osasto (Road and Waterways Administration Bridge Management
Section), Helsinki, 1989, pp. I, 19.
 Bridge Management Systems in Finland and US 73

each district the number of posted or weight limited bridges is also

shown in Table 2. It is apparent that the bridge stock composition is
somewhat different in the various districts and that the proportion of
'deficient' bridges also varies. Lappi (Lapland) district, located as it is in a
harsh arctic environment that hinders maintenance and accelerates
deterioration, shows the greatest absolute number of posted bridges, while
U usimaa, the Helsinki area district, has a very low number of posted bridges.
The relatively small number of posted or weight limited bridges reftects the
RWA's existing policy to minimize or eliminate postings throughout the
system.
Finland has had a database of bridge information for several years which
consists largely of administrative and simple descriptive information. The
system is currently being revised both to support a national bridge
management system as weil as to support the information needs of local
bridge engineers. (This database and the Finnish bridge management system
as used at the national Ievel is discussed in the paper by Kahkönen and
Marshall elsewhere in this volume.) The local engineers will be supporting an
elaborate new inspection program which will be collecting a great deal of
specific condition information on the nation's bridges during the next few
years. This informationwill be used at both the national and locallevels. In
particular, the local engineers will be able to track specific darnage on
their bridges as weil as the completion of repairs to the damage. Their
recommendations for structure repairs, rehabilitation or replacement will be
incorporated within the national planning optimization system as weil as the
local Ievel project analysis system.
Besides tracking damage, the bridge management system used in the local
office will enable bridge engineers to completely inventory their bridges,
including all structural elements, bridge fixtures such as railing or signage,
and any associated bridgc utilities and special uses such as bikepaths.
Collecting the latter inventory items enables them to anticipate confticts
with other users during any repair or reconstruction projects and to assist in
the design of new structures. Once the inventory of associated uses is
complete, particularly public utilities, railroads and forest products users,
the RWA expects to formalize the arrangements for bridge use by other
parties and possibly share operating costs.
Bridge engineers in districts also are required to track their inspection
programs. The systemwill print out inspection reports for all the bridges in
the district at any time and a basic report will show the planned schedule for
bridge inspections. The inspections arealso recorded historically so that an
inspector can see the inspection results from a prior bridge visit and also to
view the most current inspection. This will help the experienced engineers to
highlight any potentially dangeraus situations.
74 Allen R. Marshall and Mwia-Kaarina Söderqvist

A major difference in the new bridge management system is the improved

access to information for local district managers. Previously all the
information was processed and prepared in the national Road and
Waterways Administration Office. Now local engineers will have direct
access and control over all information related to the bridges in their district.
They will be able to analyse data and prepare their own reports and graphs
to display patterns in the condition oftheir bridges. Up to this pointallsuch
reports or special tabulations required the local bridge officials to file a
request with the national office. In turn this will free the national RWA
officials from the work of preparing these reports for the districts. lt is
expected that every district in Finland will be equipped with a relatively
sophisticated personal computer to operate the bridge management system.
In most cases the data will actually be stored on a minicomputer within the
district office.

SUMMARY

We have reviewed three different types of bridge management systems

ranging from a relatively simple database established to meet federal
reporting requirements and manage inspections to a comprehensive
nationwide database that has both an inventory and optimization system as
major components. A single major finding of this overview is that bridge
management systems that are intended tobe used at the locallevel must be
designed in close cooperation with bridge management officials. The failure
of the federal bridge register to address local needs in the US agencies is
apparent and these systems are growing away from the national standard as
new local applications for computerized bridge management systems
develop. In Finland the close cooperation between the national and district
bridge management staffhas led to the development of a system that largely
satisfies both parties. There is of coursestill some material being collected by
the district engineers that they have little use for in the short run. However,
as the network optimization and project development systems that are yet to
be developed come on line, these systems that rely heavily on accurate data
collection at the locallevel will have direct utility in assisting local staff to
prioritize their bridge projects and effectively utilize the resources available
to them.
 7
The Pennsylvania Bridge Management System

RICHARD M. McCLURE
The Pennsylvania State University, 212 Sackett Building,
University Park, Pennsylvania 16802, USA
and
ÜARY L. HOFFMAN
Pennsylvania Department o{ Transportation, 1009 Transportalion and
Safety Building, Harrisburg, Pennsylvania 17120, USA

ABSTRACT

A Bridge Management Work Group, as researchers for the Pennsylvania

Department of Transportation, developed the engineering concepts and
requirements and assisted the system contractor in the development ofa total
bridge management system ( BMS). A contractor working for PennDOT
provided system design, development, testing, implementation and training on
the use of BM S so{tware.
The generat objectivefor the BMS was to develop a managernerzt too! which
will enable a SJ'Stematic determination of present needs for maintenance,
rehabilitation and replacement of bridges in Pennsylvania, and to predict
future needs using various scenarios, along with a prioritization for
maintenance, rehabi!itation and replacement, ~t-!zich ll'ill provide guidance in
the effective use of designatedfunds.
The sofhrare implementation date was 24 December 1986, with full
implementation completed 28 February 1987. The department assumed
responsibilityfor the BMS software on 1 March 1987.

INTRODUCTION

Background
A seven-member Bridge Management Task Group was convened in
1983-84 to consider the development of a bridge management system for
75
76 Richard M. McClure and Gary L. Hoffman

Pennsylvania. In their report, entitled 'Pennsylvania Bridge Management

Systems', the group unanimously agreed that the development of such a
bridge management system is feasible and that it is a very important and
urgently needed tool for better management and engineering of the state's
!arge and antiquated system of bridges. 1
HPR fundingwas secured for a work group of ten persons to develop the
concepts and technical requirements, to pilot test, and to guide statewide
implementation of a total bridge management system (BMS). The group
worked in Harrisburg approximately 3 days per week. The work group
consisted of six department employees and four outside consultants.
The work group prepared a report, entitled 'Engineering Concepts and
Requirements for a Bridge Management System', which formed the basis for
a request for proposal to develop software for the BMS. 2 McDonnell
Douglas Professional Services was selected to provide the development,
testing, implementation and training on the use of EDP software. Software
work by McDonnell Douglas was performed using state funding.

Needs
The Commonwealth ofPennsylvania currently has approximately 54 500
bridges on the state and local highway systems. Thesebridges include 22 500
with span lengths of 20ft or greater and 32 000 having a length between 8
and 20ft. Approximately 30% of the bridges in the state are classified as
structurally deficient or functionally obsolete. 3 The estimated cost to put
these deficient or obsolete highway bridges in a minimum acceptable
condition is approximately $4·0 billion. The costs to put the system in a first-
class condition could approach twice this value. 2
This information provides a single snapshot of the magnitude of the
problern of aging and the accompanying decay of highway bridges in
Pennsylvania. To attack the problem, annual funding is generated from
federal, state and local sources. The basic management challenge is to make
the best use of the available funds in an overall program of maintenance,
rehabilitation and replacement.

Objective
The general objective for the BMS was to develop a management tool
which will enable a systematic determination ofthe present and future needs
for maintenance, rehabilitation and replacement of bridges in Pennsylvania
using various scenarios, along with a prioritization which will provide
guidance in the effective use of designated funds.
The specific objectives for a BMS are to develop a system that on demand
but at least annually:
 The Pennsylvania Bridge Management System 77

- Yields recommendations, with associated cost estimates, for activities

required to enable all bridges on public highways and roads to perform
their functions in the most cost-effective manner. These activities
include various Ievels of maintenance, various modes of rehabilitation
and replacement.
-Predicts present and future needs and associated costs to perform the
above activities for all bridges in at least two scenarios, including
'minimum acceptable' and 'desirable'.
-Sets statewide and regional priorities for each of the above activities,
based upon functional and physical needs for each highway
classification system, and provides a listing of candidate bridges.
-Provides a basis for recommending regional distribution of budgeted
funds.

Overall System
The basic parts of the BMS are shown in Fig. 1. The department's
computerized structure inventory records system (SIRS) will form the base

Integration Structure
Enhanced
with other cost data
SIRS
dcpartment inventory
data base
systems file

/ ~
Bridge
Bridge
rehabilitation/
maintenance
replacement
subsystem
subsystem
(BMTS)
(BRRS)

FICi. 1. Flow diagram showing the basic parts of a BMS.

for the development of an overall BMS. 4 Enhancement of the SIRS data

base, integration with other data bases and the development of a structure
cost data inventory filewill form the complete data base. The two main parts
of the BMS will consist of the subsystems for maintenance (BMTS) and
rehabilitation/replacement (BRRS), as shown in Fig. 1.

DATA BASE

Structural Iuventory Records System (SIRS)

The entire BMS project has been structured under the premise that the
Pennsylvania system should basically be a derivative of available
78 Richard M. McClure and Gary L. Hoffman

information. Therefore the department's computerized SIRS will form the

base or starting point for the development ofthe overall BMS. 4 This system
includes other data in addition to the 88 data items required by the Federal
Surface Transportation Assistance Act of 1978. 3
Significant modifications in the form of enhancements and additions are
also required to satisfy the present and future needs of all users. Since the
implementation of a SIRS in 1982, the users have recommended numerous
changes or enhancements 'to improve the utility and quality of the system. 5
These enhancements can be effectively and efficiently added to the system
during extensive software modifications needed for the development of the
BMS.

Integration with Other Systems

The redundant storage of data elements between SIRS and other
department systems, along with the fact that the SIRS does not currently
interface with other application systems, indicates that a certain amount of
duplicate effort is required to maintain the redundant data. This duplicate
effort can Iead to data inaccuracies between application systems. Therefore
the BMS will be integrated with other PennDOT application systems during
the development of the BMS.

Cost Iuventory File

Bridge maintenance unit costs for the bridge maintenance activities will be
calculated from daily foreman input of labor, materials and equipment. If
storage of such data is not provided in the maintenance system, the BMS will
be required to provide such storage in the 'structure cost data inventory file'
for needs projections and other uses.
New or replacement costs as weB as rehabilitation costs, including
widening costs, are also needed to predict needs and priorities for
rehabilitation/replacement. The 'structure cost data inventory file' will
provide storage for these costs which are on the basis of dollars per square
foot of bridge deck area. Estimated costs can then be calculated by the
system from the square foot costs which are multiplied by the length and
width parameters from the SIRS. A bridge manager needs reliable cost data
and must be able to track all bridge costs for future use.

MAINTENANCE SUBSYSTEM

Maintenance Inspection
PennDOT has taken steps to expand the inspection effort to more
descriptively cover bridge maintenance needs. Seventy-six potential
 The Pennsylvania Bridge Management System 79

maintenance activities have been identified. SIRS individual bridge data files
are being expanded so that the user can include a tabulation of those
maintenance needs with each item quantified and prioritized by urgency.
This will ensure that routine maintenance needs and their urgencies are
adequately documented. A form will be completed by the field inspector
giving estimated quantities and priorities for the applicable maintenance
activities. A unit price table for the 76 activities is also established. It will be
periodically updated based on actual cost experience. The total bridge
maintenance needs can be determined from the estimated quantities and
unit price values. Maintenance activities completed will also be included in
the enhanced SIRS. 6

The Subsystem
The extent and urgency of overall bridge maintenance needs can be
determined and sorting of this work for efficient implementation by either
contractor or department forces can be handled automatically by the
system. The system will automatically notify the county maintenance
manager and monitor the implementation of critical repairs. The work
backlogthat exists far exceeds that which the department can physically and
financially handle. Hence it is important to rationally prioritize this work to
apportion limited available funding and manpower, or to relate to
non-bridge needs. 6

Prioritization Method
It is important that guidance be provided to the district and county offices
to assist them in selecting the best candidate bridges for maintenance work
as well as which activities to perform first. This will help ensure that those
deficiencies deemed tobe the most critical to the safety features ofthe bridge,
and hence to its users, are brought to the attention of management.
A simple prioritization procedure has been developed. 6 It considers the
effect of the most structurally critical maintenance activity needed on the
bridge as weil as the individual bridge's impact on the road system. The
components of the procedure are as follows:
-activity ranking,
-activity urgency,
-bridge criticality, and
-bridge adeq uacy.

Activity Rauking
The bridge maintenance activities themselves vary in their importance to
and effect on the structural integrity of the bridge. Activities such as
80 Richard M. McClure and Gary L. Hoffman

repairing stringers or repairing abutment underscour would generally be

performed on a priority basis while activities like applying protective
coatings and constructing abutment slopewalls would tend to be deferred.
As a general rule, activities that most directly, immediately and positively
impact the continued safety and structural adequacy ofthe bridge would be
performed first while those that have minimal immediate impacts would
tend to be performed last.
The activities have been divided into groups from A to E based on their
generalized relative importance to the current structural stability of the
bridge. Group A has the highest priority and group E the lowest. The
activities repair/replace: steel stringers, ftoorbeams, girders or truss
members could be related to existing or potential fatigue damage. If the
needs are indeed fatigue related, they are more important and hence would
be given a higher deficiency point assignment. This determination can be
made by comparing these maintenance activity needs with the type of
fatigue prone member that controls the inventory rating. SIRS data
provides space for recording the controlling member type as well as the
fatigue and Ioad data related to it. If the activity is fatigue related, it will be
assigned as group AF and hence given additional deficiency points.

Activity U rgency
The severity of a deficiency can be a reason to increase its priority for
repair. The urgency factor for each activity need is coded by the district
bridge inspection unit. lt yields an informed assessment of how soon the
work needs tobe completed. As such it is also a measure of the severity of the
deficiency. It will judgementally define the promptness of action that is
needed for each specific maintenance activity need. The priority code used
will range from 0 to 5. Priority code 0 is for a critical safety deficiency where
prompt action is required and priority code 5 is for routine nonstructural
bridge maintenance which can be delayed.

Bridge Criticality
The importance of a bridge to the road network, as well as the impact of
the lass ofbridge service to traffic, are other factors that must be considered
in deciding the order in which they are to be repaired. lt is readily apparent
that the road system hierarchy realistically defines importance. That is, if a
bridge on the interstate and a bridge on the local access system have similar
deficiencies, it is obvious that the interstate highway bridge would be
repaired first. However, the impact of a bridge's closure also needs to be
weighed. lf the detour length is excessive, and hence in tolerable, the priority
for repair should be raised.
The assessment of the importance of the bridge will be based on the
 The Pennsylvania Bridge Management System 81

classification of the highway, its ADT and the detour length that would be
imposed on traffic ifthe bridge were tobe closed. Multiplying the ADT times
the detour length results in a relative measure of this importance.
Bridge Adequacy
The capability of the bridge to safely carry the Ioads that traverse the
route, and to continue to do so, weigh in a manager's decision ofwhether or
not repairs should be implemented. The Ioad capacity rating indicates the
current strength ofthe bridge. It does not give any indication ofwhat can be
expected in the future. The condition rating of the most critical component
of the bridge can be used to generally assess degradation. By considering
both the current Ioad capacity and the lowest condition rating of the
structure's components a measure of the inadequacy of the bridge can be
obtained.
Deficiency Point Assignment
Most ofthe data that will be needed to define the above components ofthe
prioritization procedure arealready in the SIRS. The only new items are the
maintenance activities themselves and their individually assigned urgency
ranking. They are important components of the proposed BMS.
Having . 1 ~fined the major parameters that are to be considered, the
relative weights to be assigned to them and their elements must be
established. To be consistent with the general philosophy of the
rehabilitationjreplacement prioritization system, a deficiency point concept
will be used for the maintenance activity prioritization system. However, it is
readily apparent that the factors and methodology being used in each system
are quite different. Although it is possible, numerically, for a single bridge to
be assigned in excess of 100 deficiency points, the deficiency point
assignment will be limited to a maximum of 100. The higher the assignment
on a bridge, the higher the priority.
Table 1 summarizes the four major components of the prioritization
system, defines the elements in their makeup and indicates the initial or trail
weights that have been assigned to each. As the procedure is tested,
evaluated and refined the weight assignments could and probably will
change.
The maintenance deficiency point assignment forabridgewill be based on
the bridge maintenance activity that has the largest sum of deficiency points
for activity rank and urgency. The deficiency point assignment and its
ranking within the county will be recorded on the bridge maintenance
activity needs screen. Hence, when a manager views the subject screen for
individual bridges, he has an immediate indication ofthe relative priority of
the most critical repair needs of one bridge compared to another and
compared to the worst possible case ( 100 deficiency points).
82 Richard M. McClure and Gary L. Hoffman

TABLE 1
Maintenance deficiency points assignment

Deficiency Component Element Deficiency

points point
assignment

25 Bridge maintenance activity rank Group AF 40

(Note: AF = group A activity that is A 25
fatigue prone and controls the B 20
inventory rating) c 15
D 10
E 5
25 Activity urgency factor Code 0 25
1 20
2 15
3 10
4 5
5 0
25 Bridge criticality
Part A: Interstate 5
US numbered highway 4
State highway 3
County highway 2
City, Borough St & Twp Rd 1
Part B: PCN 5
PCN/coal haul 5
Agricultural access 3
Industrial access 3
Part C: ADT x detour length 2 30 000 15
;::::15000 but <30000 10
> 3 000 but < 15 000 5
<3000 0
25 Bridge adequacy
Part A: Lowest condition rating < 3 15
>3 but :-::;4 10
>4 but :-::;5 5
>5 0
Part B: Load capacity (inventory rating)
(H configuration) s; 12 t 10
(H configuration) > 12 s; 20 7
(M L 80 configura tion) > 20 to s; 30 4
(ML 80 configuration) > 30 0
 The Pennsylvania Bridge Management System 83

With a deficiency point assignment being stored in the BMS for every
bridge, prioritized listings can be easily generated using the particular
parameters desired. To facilitate this reporting, friendly programmed report
generators with user defined variables are being developed.

REHABILITATION/REPLACEMENT SUBSYSTEM

The Pennsylvania System

The prioritization of bridges for rehabilitation and replacement is based
upon the degree to which each bridge is deficient in meeting public needs.
Deficiencies are evaluated in three general categories:
--level-of-service ca pa bili ties,
-bridge condition, and
-other related characteristics.
These deficiencies are then combined to yield a total deficiency rating
(TDR) on a scale which ranges from 0 to 100. Basically, the framework for
the determination of the TDR was patterned after parts of the Federal
Sufficiency Rating System (FSRS) 3 and parts of the system developed by
North Carolina State University for use by the North Carolina DOT
(NCS). 7

Level-of-Service Deficiencies
There are four characteristics included for the level-of-service capabilities:
Ioad capacity, clear deck width, vertical clearance for traffic carried by the
bridge, and vertical clearance for traffic passing under the bridge. The level-
of-service deficiencies are based on comparisons ofthe actualload capacity,
clear deck width and vertical clearances of the bridge with level-of-service
criteria which have been developed for the Pennsylvania Bridge Manage-
ment System. 8 These criteria have been set at three Ievels: minimum
acceptable, minimum design and desirable design. They are primarily
dependent upon the functional classification of the highway carried by the
bridge, with some additional dependence on volume of traffic. Equations
have been developed to calculate Ioad capacity deficiency (LCD), clear deck
width deficiency (WD), over clearance deficiency (VCOD) and under
clearance deficiency (VCUD). 8

Bridge Condition Deficiencies

The second category of deficiencies in the Pennsylvania prioritization
methodology is based on bridge condition. Bridge condition is included in
the Federal Sufficiency Rating System (FSRS), 3 and can comprise a total
84 Richard M. McC/ure and Gary L. Hoffinan

reduction of up to 64 points in the sufficiency rating. On the other hand, in

the North Carolina system (NCS), 7 bridge condition is addressed only
indirectly, through the assignment of up to six deficiency points to the
estimated remaining life.
In the Pennsylvania system, the evaluation of the bridge condition
deficiency (BCD) includes the assessment of the condition of each of the
three primary elements of the bridge: superstructure, substructure and
bridge deck. The deficiency points for each element, which are directly
related to the individual condition ratings, are given by equations. 8
Finally, the bridge condition deficiency (BCD) is determined as the sum of
the conditions deficiencies for the superstructure (SPO), substructure (SBD)
and deck (BDD). 8

Other Deficiencies
Other deficiencies are related to the remaining life, approach roadway
alignment and waterway adequacy.
The estimated remaining life entered into the BMS data base is developed
by the system as a function of the condition ratings of the superstructure,
substructure and bridge deck. The remaining life deficiency (RLD) is then
calcula ted using an eq ua tion. 8
The approach roadway aligilment may be the source of additional
deficiency points. This deficiency (AAD), which is directly related to the
appraisal rating contained in the BMS data base, is also given by an
equation. 8
The adequacy of the waterway is the final characteristic included as a
source of deficiency points. This deficiency (WAD) is directly related to the
appraisal rating included in the BMS data base and is also given by an
equation. 8

Total Deficiency Rating

Once the deficiency points of the individual categories have been
determined, the next step in the Pennsylvania system is the determination of
the total deficiency rating (TOR) for each bridge. This is done by summing
the deficiency points of the individual categories, applying limiting
conditions to several combinations of deficiency points, and applying a
factor which reflects the functional classification of the highway carried by
the bridge.
The summation of deficiency points is represented by the simple equation
TOR= cj>[LCO +WO+ VCOD + VCUO + BCD + RLO + AAD +WAD]
where BCO = SPO + SBO + BDD.
The last step in the determination ofTDE is to apply the factor cj>, which is
 The Pennsylvania Bridge Management System 85

TABLE 2
Functional classification .factors

Functiönal classification cjJ

Interstate 1·00
Arterial 0·95
Collector 0·85
Local 0·75

dependent upon the functional classification of the highway carried by the

bridge. Values of cp are given in Table 2.
A tabular representation of the development of the total deficiency rating
for bridges is presented in Table 3. Table 3 also shows four combinations of
deficiency points which have governing conditions.

Prior~tizationListings
After the total deficiency has been established for all bridges, cost
informatioh is needed in order to develop the indexes which will be used in
the prioritized listings. Initially two costs will be requested: (1) the
replacement cost and (2) the cost of rehabilitation. Also requested will be the
number of deficiency points removed by the rehabilitation or replacement.
Total deficiency ratings, combined with cost information and other
86 Richard M. McClure and Gary L. Hoffman

factors, will yield listings of bridge rehabilitation and replacement projects

prioritized in order to enable effective management of the bridge system.

RECOMMENDATIONS

The Pennsylvania BMS has been operational since December 1986. The
following recommendations reflect knowledge gained through the develop-
ment and operation of the system which could be valuable to others
developing similar systems.

- The system must have administration and top management support.

-Get the users of the system involved early in the development of the
system. In Pennsylvania the users were included as members of the
bridge management work group during the development ofthe system.
A bridge management coordination has also been set up in the central
office and each of the 11 engineering districts to coordinate the
operation of the system. Regular meetings are required to discuss
problems with the system.
-A commitment must be made to the maintenance and enhancement of
the system. At least three full-time system people are needed for the
Pennsylvania system.

ACKNOWLEDGEMENTS

The work described in this paper is being funded by the Pennsylvania

Department of Transportation and the Federal Highway Administration.

REFERENCES

1. BRIDGE MANAGEMENT TASK GROUP, Pennsylvania Bridge Management

System-proposed strategies to inspect, design, maintain, rehabilitate, replace,
prioritize and fund bridge systems. Pennsylvania Department ofTransportation,
March 1984.
2. BRIDGE MANAGEMENT WoRK GROUP, Engineering concepts and requirements for
a bridge management system. Pennsylvania Department of Transportation,
January 1985.
3. US DEPARTMENT OF TRANSPORTATION/FEDERAL HIGHWAY ADMINISTRATION,
Recording and coding guide for the structure inventory and appraisal of the
nation's bridges. Washington, DC, January 1979.
4. PENNSYLVANIA DEPARTMENT OF TRANSPORTATION, Structure inventory record
system coding manual. Publication Number 100, Harrisburg, PA, June 1982.
 The Pennsylvania Bridge Management System 87

5. BRIDGE MANAGEMENT WoRK GROUP, The Pennsylvania Bridge Management

System. Pennsylvania Department of Transportation, July 1985.
6. ARNER, R. C., KRUEGLER, J. M., McCLURE, R. M. and PATEL, K. R.,
Pennsylvania's Bridge Maintenance Management System. Transportation
Research Record 1083, 1986.
7. JoHNSTON, D. W. and ZIA, P., A Ievel of service system for bridge evaluation.
Transportation Research Board, 63rd Annual Meeting, Washington, DC,
January 1984.
8. BRIDGE MANAGEMENT WoRK GROUP, The Pennsylvania Bridge Management
System. Final Report, Pennsylvania Department of Transportation, February
1987.
 8
Data Information System for Structures: DISK

M. EL- MARASY

Ministry of Transport and Public Works, Rijkswaterstaat,

Bridges Department, Voorburg, The Netherlands

ABSTRACT

The data information system for structures is an automated information

system developed by the Dutch Ministry of Transport and Public Works to
support the management o{ bridges and other types of structures. The
information stored in the system consists ol
• Basic administrative and technical data which are almost unchangeable
hut haue to be updated in case oj' change.
• Changeahle in{ormation 1rhich is related to daily management activities
of hridges such as inspection and maintenance.
• Financial infärmation ahout maintenance.
Most oj' the data haue hl'en standardised to meet the requirement.for uniform
description and C'Va!uation o( the deficiencies and the possibility to select and
analyse the data.

INTRODUCTION

Bridges are a very important part of the road infrastructure and lack of
maintenance can have far-reaching consequences. The developments in
science, in building materials and in electronic computation during the last
decades have resulted in highly advanced design and construction methods
andin the building of very sophisticated bridges. Environmental pollution,
the increase in both traffic intensity and heavy transport and the ageing of
existing bridgestock have a significant impact on the bearing capacity and
life duration of bridges.
89
90 M. EI-M arasy

The deterioration of the condition of bridges, and the increase in

inspection and maintenance costs in view of funding limitations, has in
recent years attracted the attention ofthe highways administration. There is
a need for an instrument which can support managers and decision makers
in carrying out their work more efficiently and in granting the functionality
and safety of bridges in the most reliable and economic way. Therefore it is
of great importance to find a model for the management of bridges through
which the technical and economic problems related to inspection,
maintenance, rehabilitation, strengthening and replacement can be solved.
Moreover, it is also required to have an instrument for the assessment ofthe
life duration and the bearing capacity of bridges or bridge components.
The Dutch Ministry of Transport and Public Works has recognised the
above-mentioned needs and therefore developed an automated manage-
ment system, DISK, for all structures under its management. The
development ofthe system has been completed and the implementation has
started.

GOALS OF THE DISK SYSTEM

The goals of the system are determined by carrying out an information

analysis for the daily activities related to the design, construction, inspection
and maintenance of structures under the management of the Ministry of
Transport and Public Works. The analyses are carried out by representatives
of the different disciplines engaged in the above-mentioned activities and
have resulted in the following goals:
• Easily accessible, complete and up-to-date databank.
• Proper planning for the inspection of structures.
• Proper planning for the maintenance of structures and estimated
budget requirement for a period of 5 years.
• Contribution to the effort to predict the functional or economic life of
the structures.
• Feedback to the designers and other users of the system with relevant
information acquired from inspection and maintenance experience.

Types of Structures
The infrastructures consist of various types of structures having different
functions:
• Structures for road traffic such as bridges, viaducts, tunnels, etc.
• Structures for navigation and water control such as locks, movable
bridges, storm surge barriers, etc.
 Data Information System for Structures: DISK 91

The above-mentioned structures are built of various materials such as

concrete, steel, wood, etc. Moreover, the movable structures include
mechanical and electrical installations. All these varieties in form, material
and function ofthe structures justify the development of several information
systems. However, because all these structures fall under the management of
the same administration it was preferred to integrate all the information
systems in the DISK system, which consists of:
• Iuventory and administrative information.
• Inspection.
• Maintenance.
• Historical information.
• Special transport.

INVENTORY AND ADMINISTRATIVE INFORMATION

Knowledge of the existing bridge stock is the basic factor for proper
management and it will be impossible without it to carry out inspection,
maintenance and to plan the required budget and manpower.
The inventory program ofthe DISK system has facilitated the storage of
the bridge stock with the relevant information which supports the users on
different managing Ievels. The stored data are almost unchangeable and
consist of:

Identification
• Topographie number related to the topographic map.
• The name of the structure.

Location
• Number of the traffic road or the waterway.
• X and Y coordinates.
• The municipality and province where the structure is located.

Administrative Information
• Names of the departments responsible for design, inspection,
maintenance and management.
• Number and location of structure file.
• Dates of construction, modification and demolition of the structure.

Technical Information
• Type, material and dimensions of the structure.
• Relevant members, main members, material, type and manufacturer.
92 M. El-Marasy

Heavy Transport
• A number of factors for the calculation of bridges for the dispensation
of heavy transport.
• Design loading dass and actual bearing capacity.
• Allowable passable width and height clearances under and above the
bridge.

INSPECTION

Inspection of bridges is a fundamental task for the managers. It is not

supposed to be a target but an unavoidable medium to grant the
functionality and safety of bridges and necessary activity to supply the
decision makers with reliable planning for manpower and budget. The
DISK inspection system helps the inspectors by supplying them with several
reports and information.

Inspection Planning
The DISK system produces a Iist of the structures which have to be
inspected in a certain period. This selection can be made from the
registration of the last inspection date and the required inspection period.
On the report there is valuable information given such as the
administrative, expected inspection duration (manpower), the required
materials for inspection (crane, boat, etc.) and the possible restrictions for
inspection if there are any.

Inspection Drawings
To support the inspectors in doing their work and to facilitate the
inspection activity, the data information system introduced the making of
inspection drawings. Each structure must have a number of drawings
projecting the elevation, side view, bottarn and top view and cross-sections
ofthe structure. All structural members which must be inspected are shown
on the drawings. The location of the members is defined by two axes. .
Each member has its local number. All members which must be inspected
are mentioned in a table on each drawing which isadoptedas the 'checklist'.
The checklist, the main members to which the members belang, the
number ofthe drawing and the local numbers ofthe members are stored in
the computer.
These inspection drawings are made once and they remain valid as long as
the structure is not modified.
The checklists on the drawings are used as guidelines for the inspector.
 Data Information System for Structures: DISK 93

Deficiencies Report
The inspection of the structures can start according to the plan, the
technical instruction for inspection and the available documents such as the
drawings, general information, technical reports, preceding inspection
report, etc. The results of the inspection and the interim recommendations
for required maintenance work are stored in the computer.
The inspection procedure includes the following steps:
• To locate the deficiency of the inspected member according to the
method mentioned previously.
• To describe the deficiency according to the standard types of
deficiencies.
• To determine the possible reason for the deficiency according to the
standard reasons.
• To recommend the necessary maintenance work.
• To assess the reliability of the members with respect to safety and
functioning. This is expressed in a numerical code 0-6 (see condition
rating).
• To obtain photos of the deficiencies if the inspector considers this
necessary.
All this information completed with other general information (identifi-
cation, location and description of the structure-weather conditions--the
inspection directorate-the name of the inspector) forms the interim
deficiency report.
Together with the inspection drawings and other documents, the
deficiency report forms the inspection report.

Condition Rating
Evidently it is important for the decision maker responsible for
maintenance to know the infl.uence of the deficiency on the condition of the
member and subsequently on the structure. The condition rating of the
member and the structure is expressed in a numerical code for:
• Safety.
• Functioning.
The numerical code of the condition rating is used to get a quick impression
of the structure's condition and to use it as an instrument for the
determination of the technical maintenance priorities.
There are other factors which may infl.uence the above-mentioned
technical priorities, such as the available budget for maintenance and the
importance of the structure.
The possible numerical codes for safety and functioning are related to a
94 M. El-Marasy

certain period. The period indicates how long the safety and the functioning
of the member of the structure is guaranteed. In other words, within which
period the maintenance has tobe carried out. The possible numerical codes
are as follows:
0 Safety or functioning is guaranteed.
1 There are deficiencies but safety or functioning is guaranteed.
2 Safety or functioning is guaranteed for the next 5 years.
3 Safety or functioning is guaranteed for the next 2 years.
4 Safety or functioning is guaranteed for the next 1 year.
5 Safety or functioning is guaranteed for the next 0·5 year.
6 Safety or functioning is in danger.
lt is clear that the choice of one ofthe above-mentioned codesisamatter of
experience and knowledge. In order to be able to make a reliable choice the
inspector has to consider the following factors:
• Deterioration of the quality.
• Deviation from the design Standards.
• Change in the original conditions of the structure.
To facilitate the work ofthe manager, the structural members are grouped
into a number of main members (about 20). The relation between members
and main members is established in the system via the checklist for each
structure. Because of this relation it is possible for the system to produce a
short condition rating report for the main members and the whole structure.
The system selects automatically for each main member the highest
condition rating code given to the members which belong to the main
member and subsequently for the structure the highest condition rating code
given to the main members.
The condition rating for the structure can be adjusted manually, if needed,
to avoid undue influence from the bad condition rating of a secondary
member reflecting on the main members or on the structure.

Evaluation
The quality of the inspection results, the recommendations for
maintenance and the condition rating depend on the experience and the
knowledge of the inspectors. To ensure good inspection results and good
quality ofinformation stored in the system and subsequently to supply the
users of the system and the managers of the structures with reliable
information, it was decided to form a group of experts on design, inspection
and maintenance of structures which have to judge the results of inspection.
This group is called the 'evaluation committee' and is formed for each
discipline (concrete bridges, steel structures including electrical and
 Data Information System for Structures: DISK 95

mechanical installations, water control structures). The task of the

evaluation committee is:
• To judge the results of the inspection such as the deficiencies, the reason
and the recommendations for maintenance.
• To judge the condition rating of the structures.
• To decide if special inspection and more research are needed if the
results of the inspection are not satisfactory.
• To determine the inspection period.
• To determine the technical priority of maintenance.
The results ofthe inspection stored in the system must be adjusted according
to the evaluation work. The interim deficiency report, condition rating
report and the inspection report will be replaced by the definitive reports. lt
is also the task of the evaluation committee to expand, delete or adjust the
standard information in the system such as:
• Structural members and main members.
• Types of deficiencies.
• Recommendations for maintenance.

MAINTENANCE

The inspection procedure is completed after the evaluation ofthe results and
the production of the definitive reports. Then the maintenance procedure
can start. Theinformationsystem can support the maintenance procedure in
a quick and efficient way by supplying the following information:
• Deficiencies report in priority order.
• Report with maintenance units.
• Planning of the maintenance units.
• Report with maintenance projects.
• Planning of maintenance projects and required budget.
• Report with postponed maintenance work.

Deficiencies Report in Priority Order

The condition rating ofthe structural members for safety and functioning
after the observed deficiencies is given in the definitive deficiencies reports.
The information system can produce a report with all the observed
deficiencies and rearrange them in priority order ofthe condition rating.lt is
also possible to produce a report with a certain condition rating or in
between two condition ratings. This report is used to select the (similar)
deficiencies which can be repaired in one maintenance action. This group of
96 M. EI-M arasy

deficiencies is called 'maintenance unit'. In this report we obtain the

administrative information, the identification and the description of the
structure, the number of the deficiency, the impaired structure member,
description and the reason for the deficiency, the required maintenance work
and the final date by which the maintenance has to be carried out.

Maintenance U nits
The deficiencies which form a maintenance unit are grouped manually on
the deficiencies priority report. The maintenance cost of each unit has to be
estimated. The information system can group the maintenance units per
structure or per managing department.
These reports contain, besides the administrative information, the
description of the maintenance units, the final date for maintenance, the
estimated cost, and its state (rough or accurate estimation) and type of
budget.

Planning of Maintenance U nits

To Support the managing of structures, the system can produce a report
with expected maintenance and its estimated cost. The maintenance units
per structure are given in the report. The necessary cost is planned for a
period of 5 years. This report can be adjusted by the manager for practical
reasons and because of the available budget for maintenance.
The following information is mentioned in the report: administrative
information, maintenance units per structure per managing directorate,
estimated cost per maintenance unit, type of budget required per year for a
5-year period and the total budget required per managing directorate.

Maintenance Project
For practical and efficiency reasons it is possible to combine several
maintenance units within one structure to form a maintenance project. In
certain circumstances, when the deficiencies are similar or can be repaired by
the same contractor, the maintenance units of different structures in a certain
areaarealso combined into one maintenance project. For economic reasons
the manager can decide to aceeierate the maintenance of certain deficiencies
which were planned for later years and combine them with the maintenance
project which must be carried out. The report of the maintenance projects
contains, besides the administrative information, the project number, the
number and the description of the maintenance unit, and the ultimate date
for repair.

Maintenance and Budget Planning

The DISK system can support the various managing directorates of
Rijkswaterstaat in the formation of their yearly work and budget planning.
 Data Information System .for Structures: DISK 97

The system can produce a maintenance planning report for each department.
The report gives the total required budget for a 5-year period divided into
the different budget types.

Postponed Maintenance
All maintenance projects tobe carried out must be entered in the system.
The real maintenance cost has also to be entered. These projects do not
appear in subsequent maintenance projects reports. Other maintenance
projects which are not carried out for one reason or another appear in the
postponed maintenance report. This report can be produced per directorate.

HISTORICAL INFORMATION

It is essential for the designers, maintenance staff and the managers to know
the history of the structures; therefore there are some activities which are
entered and kept in the system, such as:
• Each inspection date.
• Each maintenance date.
• Condition rating of the main structural members after each main-
tenance.
• lnspection period.
• Real maintenance cost.
All this information is produced on the historical information report. This
information accumulates and will never be deleted from the system.

REGISTRATION OF THE INFORMATION IN THE SYSTEM

The information which has to be entered in the system consists, as

mentioned before, of two types:
• Constant data.
• Variable data.
The functioning of the DISK system depends on the registration of the
above-mentioned data. To ensure the functioning and the reliability of the
system, the following conditions should be met:
• The data must be complete and up to date.
• The data must be unambiguous.
• The results of inspection and maintenance must be registered in the
system as quickly as possible.
98 M. El-Marasy

To meet the first condition, an inventory of all structures under the

management of Rijkswaterstaat has taken place. The inventory of other
information has started and it will be gradually registered in the system. It is
expected to be completed within 5 years.
The second and third conditions can be met by simplifying the
registration method of the data in the system, especially the results of
inspection and maintenance. Therefore it was decided to supply the users of
the system (spread over the whole country) with personal computers, which
can be connected to the central computer located at the bridge directorate,
suchthat they can enter the data directly in the system. The advantages of
this method are:
• Minimise the administrative work of the users.
• The user can directly control his input.
• The output is directly available.

HOW CAN THE SYSTEM SUPPORT THE MANAGEMENT?

The infrastructure of roads and waterways has great national economic

value. The first task of the management is to ensure the safety of these
structures for the public.
For an efficient and reliable management certain actions need tobe taken,
such as an inventory of all structures, inspection planning, periodic
inspection of the structures, allocation of required maintenance funds and
carrying out the maintenance on time.
With the development of the DISK information system many of these
actions have been realised and others have been started and are in the
development phase. The system provides the management with the
following information:
• All structures under the management of Rijkswaterstaat are inven-
toried and stored in the system with many administrative and technical
data. With this information, which is easily accessible, the manager can
do his job efficiently, such as inspection, maintenance and budget
planning.
• lnspection planning is facilitated by the list of structures which must be
inspected in a certain period for a certain region. The manager of the
discipline which is to be inspected makes his detailed inspection
planning, such as the personal planning, the necessary inspection
material and, if any, the required measurements.
• The system can supply the inspectors with technical data on the
structures which must be inspected, such as identification, location,
 Data Information System for Structures: DISK 99

results of preceding inspection, maintenance carried out and the

technical history if needed.
• The standardisation ofthe structural members, main members, types of
deficiencies, the reasons, etc., facilitates the work of the inspectors and
minimises the administrative work. It also creates uniformity in the
inspection and results in uniform reports of all inspection disciplines.
• The planning of maintenance depending on the technical priorities
given by the inspectors and the expert evaluation group, as well as the
planning of the necessary maintenance budget for a period of 5 years,
are fundamental to the work plan and the allocation of the necessary
funds.
• A complete and easily accessible databank is a good source of very
useful information for designers and for the maintenance staff. They
can observe the behaviour of certain members, certain types of
structures or the effectiveness ofmaintenance undertaken in the past. It
will also add basic information to studies on the life of structures and
the determination of a suitable inspection period.
 9
Optimization of Bridge Maintenance Appropriations
with the Help of a Management System-
Development of a Bridge Management System in
Finland

ARI KÄHKÖNEN
Vialek Ltd (VIASYS), Ahvenlie 4A, SF-02170 Espoo, Finland
and
ALLEN R. MARSHALL

Cambridge Systematics Incorporated, American Twine Building,

222 Third Street, Cambridge, Massachusetts 02142, USA

ABSTRACT

Inslead of huilding new infrastructure components such as hridges and

highways, public works agencies nowadays are putting more effort into
maintaining and repairing the existing structures under their charge. This also
pertains to bridges. A basic managementproblern is how to direct the hudget
andin what order ofpriority repair actions to structures should be done. The
priority can be determined hased on a variety ofd(!Jerent agency o~jectives. In
Finland at present, it is particularly important to set a Iang-term minimum
acceptahle condition levelfor public hridges. As a matter ofpolicy, Finland has
soughl to eliminate posfing on puhlic hridges hut this objective is obviously
more difficult to achieve as the bridge stock ages.
To solve these sorts ofmanagement prohlems many d(fj'erent countries have
developed bridge management systems ( BMS). Finland isfollowing in these
footsteps, drmt·ing on their mt·n experience with BMS and those of other
countries such as Sweden, Denmark and the United States. The Road and
Watenmys Administration ( R WA) is developing a two-level hridge
infrastructure improvement management system which will serve the RWA
and the 13 highway districts in Finland. The new management system
necessitates the development of a new database of bridge information ( the
hridge director_v in RWA terms).In order to support an effective deterioration
modefing and optimization capahility lhe BMS requires a more thorough
101
102 Ari Kähkönen and Allen R. Marshall

bridge inspection system, including the collection ofspecific darnage conditions

for individual structural parts, as weil as more comprehensive generat condition
data.

INTRODUCTION

In Finland the Road and Waterways Administration (RWA) builds and

upkeeps bridges on public roads. Presently there are about 12 000 bridges on
public roads. The bridge stock is quite varied and consists of bridges of
stone, wood and moremodern construction materials. A basic overview of
the bridge stock is shown in Table 1.
The present value of the bridges is over 10 billion Finnish marks (Fmk).
(There are approximately 4·2 Finnish marks to the US dollar, 6·9 to the f
sterling and 1·5 to the French franc.) Sixty million Fmk are used yearly for
the maintenance and repair of the bridges (0·6% of the present value).
Another 100 million Fmk/year are used for the rehabilitation (meaning
widening, strel).gthening and replacement) of the bridges, which is 1% of
their present value. The weight restrictions on the bridges in Finland will
change at the beginning of 1990 in order to conform to EEC requirements.
Presently there are 350 bridges with weight restrictions under current
posting standards, but due to increasing vehicle weights it has been
estimated that weight restrictions will need to be placed on another 560
bridges. Ifthe new weight restrictions aretobe removed by strengthening or
replacing bridges within the next 5 years, an additional 80 million Fmk/year
will be necessary.

TABLE 1
Statistics on Finland's bridges

Number Length and deck area

9 496 bridges 177·1km

3 216 culverts 2·2 million m 2

Material Length Average span Deck

(km) (m) (m2)

Concrete bridges 1 501 408 106·7 15·0

Reinforced concrete bridges 273055 22·3 46·0
Steel bridges 365 501 36·0 41·7
Timher bridges 88004 10·5 12·6
Stone bridges 26081 1·6 7·5
 Deve/opment of a Bridge Management System in Finland 103

The cost of repairing, upkeeping and rehabilitating bridges in Finland will

be in the neighborhood of 240 million Fmkjyear. Determining the
effectiveness of the current upkeep and repair system of the bridges requires
that the following questions should be considered:

-What is the present condition of the bridges and what will it be in the
future with:
(a) no repair actions undertaken?
(b) a mix of different action strategies?
-What is the optimum Ievel at which bridges should be kept?
-What budget Ievel is required so that:
(a) bridges remain in their current states?
(b) bridges are upgraded to the optimum Ievel?
-What are the right targets, actions and timing at a given investment
Ievel?
-Are the chosen actions correct and economical, and have all
alternatives been explored?

In order to provide answers to the questibns mentioned above, the RWA

began the development of a new Bridge Management System (BMS) in
1986-87. Previously the RWA kept detailed inventories oftheir bridges in a
bridge directory, but it was feit that the directory was inadequate as a means
to address the questions above. The original bridge directory was a fairly
static inventory and did not effectively address changing bridge condition,
loading impacts related to damages, repair and maintenance histories, and
other crucial factors. The new bridge register is far more extensive and is
designed to meet all types of information needs.
The function of the management system based on the enhanced directory
information is to help decision makers and maintenance personnet to
determine those goals and actions by which the safety, the Ievel of service
and the condition ofbridges can be kept at the desired Ievel. The system can
also optimize and allocate correctly the available funding at the network
Ievel. On the network Ievel, the system is also able to find the optimum
condition Ievel which the bridges should be kept at on a long-term basis.
In Finland it has been decided that the complete management system
should perform its work at two Ievels: the network Ievel, where bridges are
examined as a system, and the project Ievel, which deals with individual
bridges or groups of bridges. The system at the network Ievel will help the
RWA and the upper level management in districts. The system can find that
optimal condition Ievel where bridges should be kept in the long term. The
project Ievel system is a tool for bridge engineers to use in development work
programs (1-, 3- and 5-year programs).
104 Ari Kähkönen and Allen R. Marshall

SYSTEM DESCRIPTION

The Bridge Management System is a separate program that works in

conjunction with two other modules: a bridge inspection system and a
bridge directory. The general interrelationship of BMS components is
shown in Fig. 1.
When work on the bridge management system in Finland began, we
found that the directory did not satisfy the information needs of the
management system. The RWA already had a database-type register on
their DPS8 mainframe, but only the R WA could use it. The systemwas batch
oriented and not very flexible in the types of reports that could be generated
or the type of detailed analysis that could be performed (even ifthe data were
available). Every year a routine report was sent to the districts telling them
about the bridges in their own district. The register was composed of 37
different types of data (all of it was basically administrative information).
The main problern with the register was the fact that all the information
about bridges was centralized in the RWA, it was difficult to use and, in
particular, it did not have any information about either the condition or the
continuing repair/reconstruction costs of the bridges. These shortcomings
have been corrected in the new directory.
In Finland there was a regular and thorough bridge inspection system in
place long before the managementsystemwas even considered. The problern
with this system was that during the inspection only subjective condition
information was gathered. The system was changed and now objective
darnage reports are gathered, which the management system can use as is or
after they have been converted into condition information. For more
objectivity, darnage dass tables were produced to help the inspector. Bridge

M
A

llridR~ A plimal
l n~pcclion G 1-e\el or
E Strvicf
M
E
T

FIG. I. General description of the Finnish RWA Bridge Management System.

 Development ol a Bridge Management System in Finland 105

damage, necessary repair actions and their costs are estimated with these
tables.
The basic data for the management system are the darnage and condition
reports generated from bridge inspections. The inspection information is
recorded in the directory, and the management system programs can access
this information directly. The directory also contains other information the
management system uses, such as traffic information, bearing capacity,
repair history, utilities on bridges and other items.

BRIDGE INSPECTION PROGRAM (Fig. 2)

A systematic program ofbridge inspections was initiated in Finland in 1970.

The primary purpose of the inspections is to ensure acceptable Ievels of
traffic safety. There are other purposes, too, such as making sure that bridges
do not deteriorate unnecessarily and maintaining an acceptable structure
appearance. The inspection system includes different kinds of inspections,
such as:
-Final technical inspection.
- Yearly inspection.
-General inspection.
-Special inspection.
-Special control.

Final Technical Inspection

This is done as soon as the bridge is complete. The inspection rates the
construction results, and notes any possible mistakes and darnage due to
construction. The inspection is done by a bridge engineer.

Yearly Inspection
This is done for the maintenance and upkeep ofthe bridge. The inspection
is done by sight according to a set of maintenance instructions. This
inspection is by the highway resident. The yearly inspections amend the
general inspections.

General Inspection
A major inspection every 4-8 years which:
-~Checks the condition of the bridge and its parts. The findings/
recommendations of the bridge inspector are the first phase in project
Ievel programming for the bridge.
-Checks and completes the information in the bridge directory.
106 Ari Kähkönen and Allen R. Marshall

l Final TKhnical
Inspec:tion
I I Guaranleed Repalrs
I
I Yearly lnspec:tlon
I Malntenance

I Generallnspectlon
I Repair

I I
Spec:lallnspec:tion Rehabllltation

I Special Control
I Rebullding

Jnspection System Management System

FIG. 2. RWA bridge inspection system.

The inspection is basically done by sight, but the investigation can be

augmented with simple tests performed by specialized machines. This
inspection is by a bridge engineer.

Special Inspection
This is done to fill in general inspections and when a repair plan is being
made. Underwater bridge inspections are considered to be special
inspections. This type of inspection is done by a bridge engineer and
specialist from the RWA and the Technical Research Center, and a
consultant also is present in most cases.

Special Control
Any changes, repairs or darnage done to the bridge are checked on a
regular basis 2-6 tim es a year. The purpose of these inspections is to make it
possible to go over the calculated bearing capacity ofthe bridge or apart of
it (e.g. if we want the weight restrictions on a bridge to be removed). This
inspection is done by the highway resident or by a bridge engineer. This
inspection basically deals with any form changes, cracking and erosion.
During inspections information is recorded onto a sheet. Each type of
inspection has its own sheet, except for special inspections which are always
different each time. The forms are quite detailed and will eventually be
preprinted with the findings ofthe previous inspection so that the inspector
can simply mark the items (condition items) that have changed between
inspections. Information items obtained during a generat inspection are
shown in Table 2.
Theinformation recorded by the district engineers on the field forms is put
into the bridge directory by the engineers for later use. The centrat
 Development of a Bridge Management System in Finland 107

TABLE 2
General inspection-lnformation collected

Inspection number Inspection date

Inspection type Inspector
Next scheduled inspection Special equipment needed
Comments Special control recommendation
Condition assessment of:
Substructure Railings
Edge beam Expansion joints
Other substructure Other fixtures
Surfacing Other bridge site structures
Other surface structure General condition
Calculated condition (rule-based) of:
Substructure Railings
Edge beam Expansion joints
Other substructure Other fixtures
Surfacing Other bridge site structures
Other surface structure General condition
Comments on generat bridge deficiencies
Specific assessment (for certain 'typical' bridges) of:
Structural part type Urgency class
Structural part material (inspector's code for repair urgency)
Darnage identification Recommended treatment
Darnageversion number (action proposal)
Cause of darnage (coded) Urgency of repair need
Type of darnage (coded) Unit cost
Extent (area, length, width) Effect on bearing capacity of darnage
of darnage (yes or no)
Darnage location on bridge Special inspection put into effect
Darnage class (slight, moderate, Photograph of darnage
severe) Jnspector's comments about damages
and treatments

government can obtain copies of the data via network links. The forms and
photos are also stored in paper archives.

BRIDGE DIRECTORY

The bridge directory is where all information about bridges is kept. The
directory is composed of technical and administrative information, any
information (mainly condition and darnage information) gathered during
inspections, along with geometric and traffic data which affect the usability
108 Ari Kähkönen and Allen R. Marshall

•
.. .
Highway
Resldence
l
l---------l'
Highway
Residence Highway
11
II Residence

~
Distriel

\~-~-
RWA
Distriel

• Distriel

FIG. 3. The bridge directory information connections.

of the bridge. The directory is the database of the management system. A

good directory, along with its programs for reports, is a useful tool for bridge
engineers and any other people dealing with bridges.
The bridge directory in Finland is being completely redone, both the
database structure and the information in it, as weil as the user interface and
reporting capa bilities. * The new directory will be compiled so that it will suit
the needs of different types of environments and users. The directory will be
placed in all 13 districts and also in the RWA. Information about all the
bridges in Finland will be given to the RWA, while the districts will only have
information about the bridges in their own districts. The districts and the
RWA are connected through a high-speed data network. Highway
residences, which are smaller administrative units within districts, are
connected to district offices with modems. The Connections are pictured in
Fig. 3.
lt is not anticipated that the highway districts will have any access to the
system other than to prepare reports and review data. The districts are
responsible for updating and other maintenance of the bridge directory

* Cambridge Systematics Incorporated of Cambridge, Massachusetts, USA (617)

354-0167, is preparing the new register and transferring the information from the
old database. The directory utilizes Oracle database software and operates on a
80286- or 80386-based microcomputer. More information can be obtained by
writing to either of the authors of this paper.
 Development of a Bridge Management System in Finland 109

Longterm
Performance Type

I Road Use Info

I Type of Sign ~ ITrame Sign

Abutments and Piers,
Foundation Type

FIG. 4. The entity attribute relationship of the bridge directory.

(except for some bearing capacity and administrative information). This

insures that the information is correct and current.
The general entity attribute model for the database is shown in Fig. 4. The
prototype of the directory is being produced on a microcomputer, but the
production version will be on a UNIX mainframe. The prototype is being
tested now (J une 1989) and the production version will be in place by early
1990.

MANAGEMENTSYSTEM
General
The main purpose of the maintenance and repair of bridges is safety.
Bridges must be kept in good condition so that they can be used safely and so
that they are not a danger to crossing traffic. After the safety requirements
have been fulfilled, one must also think about darnage and in what order
they should be repaired. The appearance of a bridge is also important, and of
course its ability to move traffic. Other problems are brought about by the
budget, because there is never enough money for the necessary repair and
maintenance actions. There is only a certain amount of money which must
be used as effectively as possible. We must find a tool which can help decide
what the budget should be, and what effect various spending seenarios
would have on the bridge stock on both a long- and on a short-term basis.

Network Level System

One of the most important things the network Ievel system does is to
110 Ari Kähkönen and Allen R. Marshall

define the optimum condition Ievel of the bridge system. It is evident that
bridges in bad condition are safety risks and cause extra costs for the user
(for example detouring the bridge because ofweight restrictions) and also for
society (higher accident costs, etc.). Operating expenses are higher due to
more repairs. On the other hand, it is not cost-effective to maintain all
bridges in exemplary condition, because more money would have to be
poured into the maintenance of the bridge than would be economically
beneficial to the users. The optimum level can be figured out through a
benefitjcost analysis with some degree of confidence. In the simplest sense,
the collective benefits that bridge users and society obtain from bridges
should be equivalent to the bridge maintenance and repair expenses.
It is not simple to estimate the user and society costs and benefits, of
course, and this poses one of the biggest challenges for the development of
the management system. The system must seek to calculate costs and
benefits objectively yet maintain a sensitivity to bridge system standards
imposed by government policy makers that may make little or no use of
formal analysis, e.g. 'all bridges on major public roads should be free of
posting no matter what the cost implications may be'. The implications of
the optimization seenarios can be determined and rationalized with respect
to the vagaries of bridge 'politics' and policies.
The network Ievel system analysis will also be geographically sensitive.
This makes it possible to differentiate the optimization to recognize RWA
district level costs and benefits. The basic principle is that money is put into
areas which will generate the highest Ievel of benefits. The network Ievel
system mainly serves the RWA and the management of the districts. With
this system the RWA will be able to explain the higher budget needs to
politicians who decide about such matters.

Project Level Systems

The project Ievel system is a tool for bridge engineers when they are
planning a work and action schedule. It is based on action recommendations
received from the bridge system at the network level, condition and darnage
information gathered during bridge inspections, and on given budget Iimits.
Based on a combination of job experience and help from the system a
highway district bridge engineer can draw up a proposal for a repair
program for the next year or even prepare a long-term action program.
The management system offers good tools for generating work and action
programs. The engineer can immediately find out what effects adding,
removing or changing an action will have on costs and condition. The
system can, if necessary, show the condition and darnage information on a
particular bridge, information on previously performed actions, bridge
aging factors, etc.
 Development of a Bridge Management System in Finland 111

Other Management Capabilities

The programming of bridge inspections is also part of the management
system. Historically, particularly in the United States, bridges have been
inspected whether there is a need to inspect the bridge or not, for example,
every 2 years. This fixed schedule has had some negative consequences for
inspection programs, since visits are made to bridges in excellent condition
while questionable structures are not seen until their turn in the fixed
rotation. The Finnish RWA feels strongly that it is not an effective policy to
inspect bridges on a set schedule. Instead the computer system can figure out
what kind of inspection should be performed and schedule it based on the
time since the last general inspection, the findings of that inspection,
structure deterioration models and on aging factors.
The management system can also be used to help manage permits and
routings for special transport. By determining absolutely critical bridges
using high volume heavy transport routes and reviewing possible detour
options stored in the register, it is possible to identify bridges that require
special attention because of their site and Situation.
An important part of the management system is the establishment of
programs for strengthening and rebuilding bridges to satisfy new
transportation system usages. The management system programs can help
determine in what order weight restricted bridges, or bridges having some
other problern with bearing capacity, should be rebuilt. This capability is
very timely and important because of the new axle and total weight
restrictions which will be put into effect in Finland at the beginning of the
1990s to meet EEC standards.
 10

Highway Bridge Management

JOHN W. S. MAXWELL

Grampian Regional Council, Woodhill Hause,

Westburn Road, Aberdeen AB9 2LU, UK

ABSTRACT

Managerial, administrative, financial and technical information must be

effectively translated into 'marketing' language to persuade client organ-
isations to allocate adequatefinance for bridges on the required time schedule.
Clients must produce policies and objectivesfor maintenance, improvement
and replacement. Programmes of work and priorities will evolve from
inspection and assessment information. E"\:penditure estimates will then enable
programme revision as per financial constraints using relevant database and
information systems.
Validity o{traffic predictions must be checked and quality assurance systems
applied to design. specifications and construction. Behaviour monitoring
information has an important role to play, hence the urgent need to improve
instrumentation and techniques.
If hridge management is a coin, 'heads' is the markefing dealt with in this
paper and 'tails' is the production or inspection, assessment, maintenance and
repair li'ork ( Fig. I).

'Marketing' 'Production'

Management lnspection
Administration Assessment
Financial Mai ntenance
Technical Replacement

FIG. 1. Bridge management criteria.

113
114 John WS. Maxwell

INTRODUCTION

One hundred million pounds corrosion costs identified in 1960 increased to

f600 million by 1970. No costs are available for 1980 but the new assessment
code has placed severe weight restrictions on older steel, wrought iron and
cast iron bridges.
Fifteen billionpoundswas the estimated cost of replacing 120000 UK
bridges 10 years ago. In 1978 the Scottish Development Department
suggested that estimated replacement costs of their trunk road bridgeswas
f300 million with annual maintenance expenditure 0·28% of replacement
cost. Non-trunk road bridges had a projected f1700 million replacement
cost with an annual maintenance expenditure of 0·07%. Continuing
deterioration suggests that not enough finance has been made available.
The design life for British bridges is 120 years whilst the French claim a 50-
year life. Americans, because of salt attack, finance replacement on a 30-year
cycle. Hence British bridges over 10 years old with no waterproofing may
require an earlier and higher Ievel of investment.
Access, inspection and assessment costs are now being addressed but
expenditure on actual maintenance, repair or replacement work has not
increased. The travelling public, commerce and industry find increasingly
unacceptable any disruption to transportation and its associated costs. This
paper identifies what needs to be done to ensure that the essential financial
resources will be made available when required.
The main headings dealt with are as follows:
• Highway bridge management flow chart
• Palieies and objectives
• Computerised database and information systems
• Technical issues for managers to consider
• Summary
• Conclusion
• References

HIGHWAY BRIDGE MANAGEMENT

A flow chart for highway bridge management is given in Fig. 2.

POLICIES AND OBJECTIVES

Central and local government strategy is to facilitate the transportation of

goods and people within and through the highway networks which are their
 Highll'ay Bridge Management 115

Management of work, managers, workers and other resources, i.e.

I. Plan, identify, confirm objectives, priorities and programmes for Iang-
term plans, 5-year programmes and emergencies.
2. Organise resources to achieve objectives, i.e. administrative procedures,
information intelligen{:e, staffing, finance, etc.
3. Direct, control, monitor, amend action from resources using control,
coordination and communication techniques.

Administration systems and procedures are essential to rece1ve, record,

arrange, analyse and give relevant information. Ensure that they are kept up
to date and are always readily available, e.g.
1. Produce systems and proccdures for inspection ofbridges and recording
of results with follow-up action necessary.
2. Acquire or replace hasic information on contracts, drawings, schedules
and instructions for maintenance and improvements.

Finance
1. Annual cstimated vcrsus actual and projected expenditure budget
suh-parts include cstimates, carryovers, new starts, current actual,
projected balance, engineering administration and final cost.
2. Corrclatc information from the suh-parts ofthe budgets and update on
at least a 2-monthly cycle. Thereafter predict overspend/underspend and
takc corrective action rcquired to achieve expenditure targets.

Technical
I. Datahase and information systems should include inventory and
location, inspection. tcsting and behavioural monitoring, structural
assessment and loading capacities, programmes and costs for mainten-
ance, repairs, strengthening, reconstruction and replacement.
2. Programme priorities usually motorways, trunk roads, principal roads,
hus and emergency service routes and non-principal roads.
3. Interim measurcs vary from do nothing to temporary supports, single
lane working, Bailey bridging or closure with diversion routes.
4. Planning, dcsign, supervision of construction and maintenance works
then implemented.

FIG. 2. Highway bridgc managcment flow chart.

116 lohn W S. Maxwell

responsibility. Hence the policy that the first priority for road expenditure is
to maintain and improve the structural fabric of the existing road network,
including bridges, to a safe standard.
Objectives arising from this policy should include the following:
(a) achievement of requisite safety standards;
(b) reduction in rate of deterioration;
(c) cost-effective maintenance and improvement service;
(d) effective substantiation and marketing for increased finance;
(e) changing from unplanned emergencies to stable planned main-
tenance; and
(f) maximum in-service life from work done.
Different Organisations will have to determine the policies and objectives
for maintenance, improvement and replacement of bridges which are most
relevant to their problems and circumstances. These should apply to existing
bridges, those currently being designed or constructed, and those proposed
for the future.
The problems of deterioration now with us suggest a failure in bridge
management. A more realistic approach will be required to achieve and
sustain optimum operational use of the highway network over the next few
decades. Road Engineering Intelligence and Research (J uly /August 1989)
highlighted f800m torepair Department ofTransport concrete bridges over
the next 15 years after a 2-year study. Similar sample surveys or studies to
this and those previously carried out by the County Surveyors' Society on
different types and size ofbridges should be extended to statistically confirm
the previous estimates of the scale and cost of the work to be done. Positive
management should then impltment the policies and objectives by
translating the collated technical information from surveys, inspections and
assessments into work programmes and estimated expenditure.

COMPUTERISED DATABASE AND INFORMATION SYSTEMS

Will computerised database and information systems fulfil their intended

purpose for highway bridge management? The answer is yes and no.
For new bridges built since 1984 which have been subjected to the
statutory system of cyclic inspections the answer should be yes. This assumes
that reliable basic inventory and inspection data have been made available,
checked as being correct, and used to substantiate the case for finance to
initiate cyclic and structural maintenance where and when appropriate. In
theory this should help to achieve maximum in-service life for the new
bridges and safeguard the financial investment which enabled them to be
built.
 Highway Bridge Management 117

Prior to 1984 few systems had been developed and hence data were not
necessarily kept in the format now preferred. The cyclic system of
inspections introduced in 1978 had not been fully implemented, partly due to
access problems and associated costs. Inventory information was either
non-existent, incorrect or only partially available, requiring tobecheckedas
valid and updated. Considerable expenditure and time would therefore be
required for the collation of basic data before any system could effectively
start to contribute to preventing collapse and minimising traffic disruptions,
arresting deterioration and restoring bridges to a stable condition. This
situation in particular applies to at least 75% of the regional and county
bridges in Britain.
Most systems developed recently require large quantities of detailed
information for the inventory and inspection data. Assessments of carrying
capacity are also unlikely to be completed for some years yet. A more
streamlined spartan approach is therefore required in the interim to produce
draft programmes of work with cost estimates to substantiate the case for
increased allocations of finance for bridges now before the deterioration
identified for some time develops into an embarrassing crisis in the 1990s.
The author initiated a system with Dunsmore Data and Information Ltd
some years ago 1 which was accepted by the County Surveyors' Society
(Scotland) for their regions and subsequently adapted with the author by the
Scottish Development Department for motorway and trunk road bridges.

Location L
J l

lnventory

Structural
Testing ~ Assessments
lnspections ~---l measu rements E-------7 &.
Loading
Capacities

y Maintenance
Repairs
Strengthen
Reconstruct
Replace
Costs

FIG. 3. Bridges database and information system.

118 lohn WS. Maxwell

In conjunction with Dunsmore Data and Information Ltd, Edinburgh, the

author and a colleague, Mr Ron Lee, have now produced a streamlined
spartan system more suited to the immediate purposes of the Scottish
regions and English counties. lt consists of the basic subparts shown in
Fig. 3.

TECHNICAL ISSDES FüR MANAGERS TO CONSIDER

Too many repairs are required too soon or too often to relatively new road
and bridge projects. This suggests that something may be wrong with our
planning estimates, our design and specification criteria, or construction
quality. Behaviour monitoring information is essential in such circum-
stances.

Traflic
Volume, mix and weight estimates require checking.

Volume
The increase in the number of licensed vehicles (UK) this century is as
follows: 1900 (0·018m), 1920 (0·650m), 1940 (2·325m), 1960 (9-439m), 1980
(19·210m) and year 2000 (>30m?). New national road traffic forecasts in
May 1989 superseding the 1984 figures suggest that growth in the 1988-2025
periodwill increase previous predictions by somewhere between 83% and
142%.

Mix
Having determined that the damaging power of commercial traffic was
30% heavier in 1982 than in 1974, predictions for 1990 suggested double for
each category! 2 British Road Federation (1987) determined an increase in
lorry traffic 1977-85 of 10% in rural areas, 37% on motorways and then
suggested a further 12% increase by the year 2000. 3

Weight
Construction and use weight increased recently from 32 to 38 t with the
suggestion of 40-50 t by 1996. Moreinformation is required on the number
and frequency of vehicles exceeding these weights in addition to abnormal
Ioad notifications.

Design, Specification, Construction and Quality Assurance

Where current design and specification criteria may not accommodate the
traffic volume, mix or weight, some 'horses for courses' intuitive allowance
should be added to the computerised optimum solution. Materials and
workmanship specifications must not be relaxed with records of compliance
 Highway Bridge Management 119

kept on database to facilitate any problern identification later. Too many

new materials and methods are marketed on the basis oflaboratory tests or
applications abroad. Properly monitared on-site trials with recommended
applicators must form the basis of any specifications written.
Construction quality should improve if dient organisations are more
careful with their start dates and contract periods. Selection of contractors
requires a more professional approach also. Carefully control and Iimit
the amount of subcontracting and do not necessarily accept the lowest
tenderer.
Design or construction firms with accredited and properly implemented
quality assurance systems should present less risk of mistakes being made
than those without. Clients, however, should not be lulled into a false sense
of security. They must continue to strive to satisfy themselves about the
attitude ofthe people they are dealing with, and the pride they take in doing
a good job.

Bridge Behaviour Monitoring

In addition to the foregoing concern expressed about planning, design,
construction, etc., the improved braking efficiency of modern vehicles has
significantly increased the longitudinal force on highway bridge articulation
and substructures, i.e.
BS 153 HA maximum Ioad 253 kN, i.e. 100 kN plus 17 kN/m span >3m.
BS 5400 HA maximum Ioad 700 kN, i.e. 200 kN plus 8 kN/m loaded
length.
Thermal movements of curved or highly skewed decks can also introduce
complexities. If unexpected forces, movements or deterioration in condition
of materials arise, it seems sensible that instruments be installed to give a
case history before, during and after structural distress andremedial works.
This information should help confirm the extent to which the bridge is
behaving in relation to current design and code parameters. Decisions can
then be taken on the need to merely replace 'like with like', or strengthen.
Such matters will, of course, greatly influence the estimated cost of remedial
works.
The author has developed computerised systems 4 for monitaring (a) deck
deflection and temperature with joint movement and column tilt, (b) wind
speed and direction, and (c) tide rise and fall, current speed and water quality.
Experience has also been acquired on structural vibration of 'lively' decks,
monitaring retaining walls and Ioad testing arch bridges to destruction.
Such experiences confirm the need for the installation of instruments to be
dealt with at the design stage to facilitate provision of information, giving
advance warning of deterioration or imminent structural distress.
120 lohn W S. Maxwell

SUMMARY

The introduction highlights the serious deteriorating magnitude ofthe crisis

now emerging. A method of alleviating this crisis is identified in the highway
bridge management ftow chart (Fig. 2). Organisational policies to be
complied with and objectives tobe met are then clarified and confirmed. An
interim streamlined database and information system has now been
developed to facilitate dealing with the aforesaid to resolve the crisis. In
conclusion, the suggestion that something basic may be wrang is raised on
technical issues for managers to consider.

CONCLUSION

Highway bridge management to be effective must include the 'marketing'

and 'production' criteria identified in Fig. l. Implementation of the Fig. 2
management, administration, finance and technical ftow chart guidelines
will stimulate the organisations and managers involved to comply with the
policies and achieve the objectives required.

ACKNOWLEDGEMENTS

The author thanks George Kirkbride, Director of Roads, for his permission
to contribute this paper, and Mrs Lesley Higazi and Mrs Kath Butler for
their excellent assistance in the production of the paper.

REFERENCES

l. MAXWELL, J. W. S., A micro-computer database and information system for

engineering management. Civil-Comp 85, Proceedings Second International
Conference Civil and Structural Engineering Computing, Vol. l. Civil-Comp
Press, Edinburgh, pp. 115-22.
2. WRATHALL, D. and WALLIS, C. C., The practical impact ofreduced investment in
highway maintenance. The Highway Engineer, April 1982.
3. BRITISH RoAD FEDERATION LTD, Fact, London.
4. MAXWELL, J. W. S. and FYFFE, S., Essential monitaring for essential repairs-Tay
Road Bridge, Dundee. Civil-Comp 87, Proceedings Third International
Conf'erence Civil and Structural Engineering Computing, Vol. 2. Civil-Comp
Press, Edinburgh, pp. 253-9.
 11

Bridge Management in Cyprus

P. H. MAY
lohn Burroll' & Partners, Exeter, Devon, UK
and
S. VRAHIMIS

Department of Public Works, Nicosia, Cyprus

ABSTRACT

An attempt is being made in Cyprus to introduce an objective approach to

bridge management through the development and setfing up ofa computerised
bridge inuentory system. A schedule of' condition inspections has been
introduced ll'hich picks up overdue routine maintenance and provides datafor
an objective assessment of repairs required. Work is costed and an order of
priority estahlished to assist hudget preparation and planning offuture work in
various time scales. lt is anticipated that hy centralising re:-.ponsihilityfor the
condition inspection procedures and bringing the hridge inventory system
umler the control of a nell'ly .fcmned headquarters-based maintenance
management uni! the non-availahility o/skilled technical sta.ff at distrief Ievel
can he Ot'ercome. Impfementalion of the system developed is 1Wll' proceeding.

BACKGROUND

Cyprus is the third largest island in the Mediterranean with an area of

9250 km 2 . It has an intense climate with very hot, dry summers (when the
inland temperatures frequently exceed 40°C) and cooler, rainy, rather
changeable winters, separated by short spring and autumn seasons. Snow
falls frequently every winter in the Troodos mountains on ground over
1000 m above sea Ievel. These falls usually commence in December and
continue through until the end of April.
The island's population is currently estimated to be 673 000, of whom
77% are Greek Cypriots, 19°1!, Turkish Cypriots and 4% other minorities.
The political situation is complex and outside the scope ofthis paper, but it is
121
122 P. H. May and S. Vrahimis

sufficient to state that since the Turkish invasion of 1974 the Republic of
Cyprus has effective control over only the southern portion of the island,
representing some 63% of the land mass.
The Cyprus Department of Public Works (under the Ministry of
Communication and Works) is responsible for the preparation and
development of the government's road improvement programme as weil as
the design, construction and maintenance of roads, airports, fishing shelters,
coastal protection works and government buildings throughout the island.
In 1987 these responsibilities included the maintenance of 1898 km of
bituminous-surfaced roads and their associated bridges and culverts in the
southern portion of the island. Ofthis total, 209 km were within designated
municipal areas.
The diverse history of Cyprus over the last 500 years has resulted in a
fascinating variety of bridge and culvert forms and material types being
used. Inevitably, on an island where warkable Iimestones and calcareous
sandstones predominate, traditional masonry arches are common. Several
of thesedateback to the Lusignan and Venetian periods of Cypriot history
(AD 1192-1571), and though not now taking heavy Ioads some arestill in
service on minor roads. There are, of course, many examples of masonry
arches constructed more recently (particularly during the British colonial
era), when skilled craftsmen were more readily available. Most of these are
still giving excellent service even on roads carrying quite heavy traffic. More
recent bridges have favoured beam and slab-type concrete construction
(both reinforced and prestressed), and there are many composite structures
of in-situ concrete decks on steel girders and, in some cases, on steel trough
deck sections.
Since the mid-1970s considerable emphasis has been placed on the need to

Exprcssways:
Papho '· Alrcady Finished
Undcr Construction
Bcing Planned

FIG. I. Extcnt of currcnt expressway development.

 Bridxc Management in Cyprus 123

improve and extend the road network. Using financial assistance provided
under four World Bankhighway projects, as weil as help from the Kuwaiti
Fund and using internal resources, much ofthe primary trunk road network
either has been or is being strengthened and/or upgraded to international
standards. An indication of the present situation is shown in Fig. 1.
In 1984, under the Third Highway Project, attentionwas directed towards
the need for improved road maintenance and a project was funded to
develop and implement a computerised road maintenance management
system. The system implemented adequately covered the department's
requirement for inventory and condition information on road pavements
and drainage but did not cover bridges or culverts in any detail. Accordingly,
under the Fourth Highway Project, the same consultants were appointed to
develop a complementary bridge inventory system which could cover these
additional aspects.

ORGANISATION AND BUDGET ARRANGEMENTS

The Department of Public Works acts through a Nicosia-based head-

quarters and five operational district offices, each under the direction of an
experienced qualified engineer. Though not all districts are of similar size
(due to the present divided nature of the island), a typical district
organisation is shown, together with the headquarters arrangements, in
Fig.2.
The establishment of a road maintenance unit in 1987 und er the control of
a senior executive engineer confirms the department's commitment to
improve maintenance of roads and bridges. Though initially established to
operate the road maintenance management system, it is now being
developed to include responsibility for the bridge inventory system.
The construction of new roads and their associated bridges and culverts,
tagether with the periodic maintenance of road pavements (resealing,
overlaying, etc.), is carried out by private contractors from designs prepared
by the department at headquarters Ievel, sometimes with the assistance of
external consultants. Only routine maintenance is, therefore, undertaken by
the departmenfs own direct labour force. The district engineers have the
dual responsibility of supervising works being undertaken by private
contractors and operating a small, but by no means insignificant, direct
labour establishment. This is reflected in the organisation shown in Fig. 2.
No separate budget provisions are made for the maintenance of bridges
and culverts (as distinct from the maintenance of roads). In 1987 the
provisions for road maintenance were as follows:
Routine maintenance $1 700 000 (equivalent)
Periodic maintenance $2 800000 (equivalent)
124 P. H. May and S. Vrahimis

DIRECTOR
I
I
CHIEF EXECUTIVE
ENGINEER
I
I

SENIOR EXEC. ENGINEER : SENIOR EXEC. ENGINEER SENIOR EXEC. ENGINEER

I I
I I

·--------------------------------------------------------'
I
I I I I
I
I
I I I I I
: (District Engineer s Establishments)
1 :
I I
I I

NICOSIA LIMASSOL LARNACA PAPHOS FAMAGUSlA

HEADQUARTER'S ARRANGEMENTS

DISTRICT
ENGINEER

--------------·-----------------
1
I
I
SITE ENGINEER SITE ENGINEER SITE ENGINEER
ROADS MAINTENANCE BUILDINGS

TECHNICIAN TECHNICIAN
ROADS & BRIDGES BUILDINGS
TVPICAL DISTRICT ORGANISATION
FIG. 2. Typical district Organisation and headquarters arrangements.

Of the above figures, though no detailed breakdown of expenditure on

bridges is available, less than 5% of the provision for maintenance is
believed to have been spent on minor routine maintenance of bridges and
culverts. This low Ievel of expenditure refiects the priorities for attention
perceived by engineers at the district Ievel on the basis of only irregular
reports on bridge condition given by district technicians (from inspections
carried out during the course of their other duties). lt is not an indication of
the true condition of those structures. Until establishment of the bridge
inventory system no objective surveys to establish bridge condition had been
undertaken. Undoubtedly many of the district engineers had wished for
some time to carry out such surveys but had found a shortage of skilled
technical staff at the district Ievel a major barrier. lt is anticipated that the
establishment of the bridge inventory system under a central road
maintenance unit will circumvent this problern by centralising the staffing
requirement and hence making better use ofthe trained personnet available.

INSPECTION PROCEDURES

It was known that in the UK, where skilled technical resources can perhaps
be more readily assigned to the monitaring of bridge conditions,
 Bridge Management in Cyprus 125

management is based on a cycle of 'general' inspections (from ground and

deck Ievel) at 2-year intervals with 'principal', more detailed, inspections at
6-year intervals. (More frequent inspections are required for bridges with cast
iron members but these do not exist in Cyprus.) Similar inspection cycles
appear to be adopted by other industrialised countries, though in several
cases 'general' inspections are carried out on an annual rather than biannual
basis.
The less developed countries of the world, on the other hand, have the
combined problern of

(i) not having sufficient skilled technical resources available to carry out
routine inspections, and
(ii) having generally adverse climatic conditions (high temperature
variations and concentrated rainy seasons) which necessitate some
form of inspection being carried out at least annually and in many
countries at 6-monthly or quarterly intervals.

These problems may (arguably) be mitigated, to some extent, in the less

developed countries by the reduced economic (and possibly social)
consequences of bridge failure.
The Department of Public Works in Cyprus has a significant shortage of
skilled technical staff brought about by factors outside their control. lt was
considered, therefore, essential that the time of those skilled technical staff
that were available should be reserved for inspection of bridges which had
been screened by less skilled staff for likely defects of importance. In
addition, it was feit desirable, in view ofthe aggressive climatic conditions on
the island, to carry out some form of q uarterly checking procedure to pick
up darnage to bridges and culverts resulting from seasonal and, indeed, daily
expansion/contraction cycles, from the efTects of winter rains and from the
spring thaw, which results every year in considerable erosion problems on
rivers leading south from the Troodos mountains.
An inspection cycle has therefore been set up which provides for
-quarterly bridge and culvert checks (by district-based foremen from the
direct labour force in the course of their other maintenance duties);
-follow-up bridge and culvert checks (by district technicians in response
to a quarterly check showing the need for immediate further
inspection);
-annual routine condition inspections (by headquarters-based trained
technicians); and
-detailed condition inspections (by headquarters-based bridge engineers
in response to routine condition inspections showing the presence of
significant defects).
126 P. H. May and S. Vrahimis

Initially resources will be allocated to enable routine and detailed

condition inspections to be carried out on bridges but not culverts, the
separation being taken at a combined clear span of 4 m. Eventually it is
hoped that resources can be made available to extend this to include all
crossroad highway structures irrespective of span, since the difference
between bridges and culverts is often quite arbitrary.
Quarterly bridge checks are considered adequate to keep track ofroutine
maintenance works which may be overdue or have been overlooked, and to
pick up channel blockages or bank erosion before and after the winter rains.
In addition, thesefrequent checks ensure that no bridge condition requiring
urgent attention (to keep road users and general public safe or to preserve
the structural integrity ofthe bridges and culverts concerned) goes unnoticed
for any length of time. Responsibility for this element of the overall bridge
management process is retained entirely at district Ievel and is not
computerised.
The format for routine and detailed condition inspections is basically the
same, differing only in the extent of data collected. The format used is similar
to that developed by the Northern Ireland Roads Service and includes
identification of defects by codes according to the type of defect, the part,
area and position affected, the extent of deterioration and its severity.
Structures are inspected in a strict order to ensure no elements are forgotten
and no defects overlooked. Concrete, steel, masonry and timher structures
or a combination of these materials can all be accommodated.
The level of decision making required for the routine condition
inspections is not considered beyond the scope of trained technicians and,
indeed, such staff are preferred. The survey requirements for the detailed
condition inspections are extended to include decisions on the type of
treatment required to correct observed defects, the time scale within which
the work should be undertaken, the degree of relief the repairs will provide
and the estimated costs involved. This latter item can be either in the form of
broad bands of cost (less than $1000,$1000-$3000, etc.) or in terms ofa more
exact estimate. lt is intended that these decisions should be made by
experienced engineers, wherever possible, on site.

BRIDGE MANAGEMENT APPROACH

The management approach adopted in Cyprus is based on the following

sequence:

(i) bridge inventory collection,

(ii) quarterly checks (for urgent and overdue routine maintenance),
 Bridge Management in Cyprus 127

---------------''
INVENTORY :
COLLECTION :
---------------''
'
'
'----------------------------------'
'' ''
' '
'-------------'
I I ·---------------' I I
: QUARTERL Y : ANNUAL :
BRIDGE ROUTINE :
CHECKS CONDITION :
------------- : INSPECTIONS :
'' I
'
'---------------'

' '
I
'
'--------------------' I I
'
'--------------------' I
I I I I
I I I I

---------'' '--------'
' ' ------------''
I OVERDUE : : URGENT : DETAILED : MONITORING :
: ROUTINE : : WORKS CONDITION : PROGRAMME
WORKS INSPECTION

'-------------''
BRIDGE :' REPAIRS :
ASSESSMENT : PROGRAMMES :
'-------------'
' '
'------------' ·-----------'
---:' STRUCTURAL :' ---:' THIS YEAR :'

I---·-------- I '-----------'
---:' GEOMETRie :' ---:' NEXT YEAR :'
-----------:
'-----------' --------------'
---: NEXT 3 YEARS :'
' '
---: HYDRAULIC :

' ·---------------'
' AFTER 3 YEARS :'
:---:

Fru. 3. Logic diagram of bridgc managemcnt approach.

(iii) annual routine condition inspections,

(iv) detailed condition inspections (where required),
(v) prioritisation procedures,
(vi) repair works programming,
(vii) sufficiency assessments,
(viii) bridge rating procedures, and
(ix) bridge replacement programming.
A logic diagram showing the above approach is shown in Fig. 3.
From the outset there appeared a sound basis on which all the above
could be computerised with the exception of the quarterly checks (the
128 P. H. May and S. Vrahimis

analysis of which, as outlined in Fig. 3, is best held at district not

headquarters Ievel and which cannot therefore be sensibly computerised in
the Cyprus context at the present time) and the sufficiency assessments which
form the basis of the bridge rating procedure and replacement program-
ming. The sufficiency assessments, if done correctly, involve comprehensive
investigations into the geometric, hydraulic and structural adequacy of the
bridges concerned. At present these assessments appear better undertaken
manually.
In looking round at available computerised systems which could handle
the inventory and condition inspections, prioritisation procedures and
repair works programming, it became apparent that the work undertaken by
the Northern Ireland Roads Service was very much consistent with the
approach it was intended should be adopted in Cyprus. The availability of
this system on an IBM PC microcomputer proved tobe an added attraction.
In the event, the system made available from Northern Ireland was
considerably amended and extended for use in Cyprus to the extent that
complete rewriting of the software became necessary. None the less, the
system being implemented in Cyprus is based on almost identical inventory
collection and condition inspection procedures to those used in Northern
Ireland and a senior engineer from the Roads Service was made available in
Cyprus to assist with the training of local staff in the techniques of bridge
condition inspection.

SYSTEM DETAILS

The bridge inventory system as developed for use in Cyprus is a

computerised database of bridge information with associated interrogation
and analysis programmes. The database contains

(i) a static library of bridge records which does not change with time,
(ii) a dynamic library of condition inspection records updated annually,
and
(iii) an archive of repair records giving historical details of various works
carried out.

These three elements of the database are linked by a common referencing

system which allocates a unique reference number for each bridge and
culvert in the network.
Interrogation programs included in the bridge inventory system allow
the static and dynamic libraries of bridge and condition inspection records
 Bridge Management in Cyprus 129

tobe sifted and sorted according to a variety of procedures. These enable the
user to obtain

-district bridge registers,

-individual bridgedetails,
-lists of bridges with specific physical characteristics,
-lists of bridges constructed within specified years,
-lists of bridges with Ioad or height restrictions, and
-lists of bridges with defects that require monitoring.

The analysis programs allow prioritisation of defects noted in the

routine and detailed condition inspections on the basis of the part of bridge
affected, the type of defect observed, the importance ofthe defect (measured

PUBLIC w.:>RKS DE:PAA'IMENl' BRIDGE INIJENl'OIW SYS'mol

NICOS IA, CYPRUS
BRICGE: ROCORDS - DATA SKEEl' ·rvPE 2

DEI'AILS

BRICGß NO. [_][_1_1_1_1-U_I_ll(_l [_] m.1S REF. l_l_l

BRICGß NAME l_I)_I_I_I_I_I_I_U_I_I_I_I_I_!_I_I_I_I_I_I_I_I_l
DIS'rRICl' [_) FU!Cl'ION 2 - R01\D OIJER RIIJER

NUMBER OF ROAD OIJßl{ NAME Or' lUIJE:R UNDE:R

[_][_1_1_1_] l_l_l_l_l_l_l_j_l_l_l_l_l_l_l_l_l_l_l
ORIE:NT~l'ION [ )/[ ) YEAA BUILT U_l_l_l

DE:SIGNE:D BY: BUILT BY:

l_I_I_I_I_I_I_I_I_I_I_I_I_U_I_I_U r_I_I_I_!_I_I_I_I_I_I_I_I_I_U_I_I_l
G.A.DRG.NO. [_j_l_l __l_l_l_l_l LOAD RßSTRICTION [_1_1_1_1_]
SURIJEYED BY: l_l_l_l_l_l_l_l_l DATE OF RßCORD U_l/U_l/[_1_1

DIMENSIONS

NUMBE!<. OF S~ANS (MAX 5) [_) U_l_._l_l

SKEW ANGLE: (DEX:J.)

SQUARE: U 1_._1 1 U 1_._1 l U I_. I l U !_. I l U I . I l

m l_I_I_._T_l !_T_I_._T_l l_T_!_._T_l T_T_I_._T_l T_T_I_._T_T T_T_I_._I_l
SKEW u l_._l J u l_._l J u l_._l 1 u !_._I J u l_._l J
(2) r_u_._Ll u_L._Ll r_LL._Ll r_LL._u UJ_._Ll r_u_._u
DEI'AILS OF ROAD OIJER DE!'AILS OF RIIJER UNDE:R

PARAPEr WIDI'H [
-- --- ) AVERAGE WID'l'H OF RIIJE:R [
-- --- JJ
IJERGE WIDl'H [
-- --- JJ AIJG. DEPI'H (NORM. F!JJII) [
-- --- J
[ --
F(XJrPATH WIDTH [ AIJG. DEPI'H (MAX. FWJO) [
HAAOSHOOLDE:R WI!Jl'H
- - --- BED TO SOFFIT (MAX.) ---
[
-- --- )J - - ---JJ
[ --
CAARIAGEWAY WIDTH [ CONSTRUCTION DEPTH [
-- - - - --- J
[ -- --- J
MEDIAN WIDl'H [
-- --- JJ BED 'l'O SOF.FIT (M!N.)
CAARIAGEWAY WIDl'H
HARDSHOOLDE:R W!Dl'H
[
[ -- ---
- - --- )
CONSTRUCTION DEPTH
LENGl'H O.F BE:D PROl'Eel'ION c: -- :)
- - --- J
F<XJl'PATH W!Dl'H [

[ --
VERGE W!Dl'H [
---JJ
PARAPEr WIDl'H
-- ---
- - ---
OVERALL WIDTH [ J
LENGI'H OF PARAPErs N/W [
- - - - -)
LENGl'H OF PARAPErs S/E [
-- --- )
FIG. 4. Typical bridge record data sheet.
130 P. H. May and S. Vrahimis

PUBLIC WORKS DEPARTHENT BRIDGE INVENTORY SYSTEM

NICOSIA, CYPRUS
INSPECT ION RECORDS - DATA SHEET
== === === == ==== == == =====: ========== == ====== =:::: === =====::: ===: == == ===== ===== :: ===

BRIDGE NUHBER 1 H / / / H / / l/1 l

FUNCT ION NUHBER 1 / J

ORIENTAT!ON l_l/l_J

NUHBER OF SPANS 1_)

I NSPECTED BY: I / / / / / / / J DATE OF INSPECTION I / l/1 / l/1 / I

PART AREA POSITION DEFECT SEVERITY TIHESCALE COST

EllE II! TREATMENT RELIEF

l. FOUNDAT!ONS
:-- ... -----------------: :--------------------------
[: J [ l. [:: : J [: J [ J [ J [ J [ J [ J [ J

[: l : [ J [ : : : J : [: J [ J [ J [ l [ J [ J [ J
l
(I l I [ l [ l I I l I [: J [ J [ l [_l [_] [_] [_)
1----.. ---------------l I--------------------------
OPT!ONAL lNPUT ENGINEER 1S INSPECTION ONLY

FIG. 5. Inspcction rccord data shcet.

in terms of its extent and severity), the value for money the repair will
represent (obtained from the degree ofreliefthe work is likely to provide and
the expected cost involved) and the importance of the bridge in the network
(a function of road class).
The user can then obtain a costed Iist of urgent work, or of works to be
carried out in the next financial year, or of work for Ionger term action in
(say) the next 3 years each in order of priority. These listings can form the
basis of objectively assessed budget requests and can be used for Iongerterm
financial planning.
By defining budget cut-off Iimits "scalped' priority lists can be obtained of
all works within the budget Iimit rearranged so as to be listed in bridge
number order. Such listings a,re prepared to assist work planning.
Examples of typical system data sheets and some of the reports are given
in Figs 4~7.

IMPLEMENTATION

The system developed is now complete and the programs have been installed
on a computer in the Department of Public Works. Testing on a trial of
CYPI<IIS PUlil lC WlJf:l:s l!Ef'AI(I tll.ill Date : 22/ 6/89
ßkiDGE INVOITORY SYSTE11 DISTRICT f:EPORT Page No : l
===== ;::::: ===:: == === === ::.:::::: ==::::::::::: :-:: ===:. ====== ==== ===== ==-: ================== == === =:: =============== == ======== === == === ======== ==== ===
D!Sl~ICT : I
FUN Cl. LOidJ I'LAHD Si:EII ------- SKE~ S~AN DIMS. ---------- END
B~IDGF. NO : YEAR RSTf:. HCI :JHT A/11), Sf'AN I SPAN ·t SPAII 3 SPA/1 4 SPAII j H11JNO. SUPPORTS PIERS DECK SURFACE ANC I LLARY
-------------------------------------------------------------------------- .. -----------------------------
AO(IO J-00~/1 I IIUNF 11,1111 b' :J(I f:' 411 b,4(1 f,' 3(1 [: A A BEC oE r EEB 0 AB CCBEF
AüOOI-(113/1 :· NtlNf: ~ /. (J(J 1,,(1(1 f,' 7~ AAA CED [I Ec E GC EAB B AA F F
AOlJOJ-014/IA :1 11011[" j(l, (11) ')' /I) ')' 8(1 'j' öl) ~. 80 9' 80 8 AA BED DE F Er c 0 AB 0 0 BCE
AtJOOI-014/IH 2 11(1/lt :)1);(\(J ')'8ft 'j. f,(l Ii A A BE0 oE r EEC 0 A B D0 BCE
I.,. Jt/ I'), ]I) A t~ A
...,t:l:l
40(1(1J-(117/I ? I I. •/4 I 9. /II 19' 30 BE 0 BEG EEB CA B CGBAA
A0003-004/I l u. (ltj F,' 50 6. 50 ~ A A BE0 ß EC DE B 0 A& AAAAA ~~
BOOO 1-006/1 1 1101-if <II, I) II 5, (H) AAA ur o AAA 0 E0 EAB AA EF F
~000 1-008/ I 2 NONE (l,(lll ~' (II) A II A IJ F 6 AAA E6 B EAB & &CFF ~
:::::,
15, (II) AAA 8 E D ;::
BOOO 1-009/1 2 t!ONE b' 51) f.. 80 6. 60 0 E0 E GB EAB B AB F F
IJOOO 1-013/1 2 NONE 1),(1(1 9' •10 ~' 40 9. 45 9.45 9. 45 AAA BED 0 EE EE 8 DA B BD EF f ~
~

8000 1-0 13/1 B 2 IIGNE (1, 00 'l. 40 'l. 40 'l' 45 t, A A BED DE E EEB 0 AB BO EU Si
~
BOOO 1-01711 2 NONE 30. (t(l 3' 60 3. 60 AAA B E
r Br c DE B EAB BAAf F ~
&0001-018/1 2 Nt)/ I~ 45. 1)(1 [,' 31) 1, A A BE D AAA 0E B EAB A B A FF
Si'
80001-019/1 2 NOtiE 42. (IQ 8. ~0 AAA BED AAA EGB EAB AAAFf
80009-021/1 2 NOUE 18. Oll 5. 7(1 6.00 f.. 01) 6' 0() s. 90 AAA BE0 BEC E GB EAB B A BE F Q
"'::l
...,
B0009-02S/I 2 NOt/E o. ou 7' 7\1 8. [,0 8. f,Q 8. [,(I 8. 60 AAA BE0 BEC EEB 0 AB CA E FF
:::::
80009-025/18 2 NONE o. 0(1 7' 70 AAA BE 0 8EC EEB 0 AB c A Ef r
IJ0009-02f./ 1A 2 N!iNE 0.1/ll &. u5 AAA AAE AAA CCA 0 AB AAAf F
"'
B0009-\l2f.t I B 2 NOIIE tl,li) 8. 10 AAA BG0 AA A EGB 0 AB AAAF f
IJ0009-02E.I 1c '/ II OII[ (J.t'.lfl \J. •II) L, 3(1 r, A A 8 ED BEC EEB 0 AB AABEF
R000~-0'27/1 2 NONE (1, IJH 5. ~~1.1 AAA BE0 AAA EGB EAB A A AF F
ß(l(l09-(f29/l 2 NO Nt \1,(111 ~. 75 9. 45 9. 50 ~. 4~ 9. 35 t, A A BE D 8EC EEB 0 AB BABEf
ll000~-035/1 A l NONE 11,1)(1 4' f,l) t\ A A 8 E0 AAA EEB DAB AABF f
B0009-IJ35/I B 2 NUNE (),(1(1 4. i.(l 1\ A A Ah 1) A AA CCA 0 AB AAAF f
UO\lll'J-03:i/2A .. tiUIIE .,L, (>IJ :1.•11) AAA AAD Ah A CCA 0 AB A A AF F
D0009-03S/2D 2 NOiiF. ltl. (1(1 3. 50 [1 C A BE D AAA DE D DA B A D EF f
FIG. 6. District bridge rcgister. <.;.>
-
132 P. H. May and S. Vrahimis

CYPRUS PUBLIC WORKS DEPARlMENT Date : 81 UB~

BRIDGE INVENTORY SYSTEM Page : I

PkiORITY LISTING BY BRIDGE NUHBER

=== == ======:: ====: ==:: == =:: =======:: ======= ===== ====:: =:: ======= == === ====== == ======
from Timescale : A
To Ti~esca1e : D
Budqet Limit 4500(1
Oi stri ct

PART OEFECT SE VER 1TY TIHESC, CUHULATIVE BALANC!NG

BRIDSE NO I AREA : EXTENT I TREA T. I COST COST fACTOR
.... --- -- --- ........ --- --------- --------- -----------
1A0001-00&11 PO 8 (:fj A A 800 800 315.000
IAOOO 1-006/1 JV HB E 300 1100 243.000
1A0001-00G/1 ON u CA c A 1200 2300 220.500
1A0001·01311 JU Q DA A 0 c 400 2700 "/2'l.OOO
IA0001-013/1 JV II& A 8 450 3150 243,000
1A0001-014/1 Ok T G8 B B 500 3[,50 364.500
I AOOO 1-0 17/ 1 JV R HA c c B 1360 5010 1012.500
1A0001-01711 Hf 0 CA 8 8 A 100 5710 850.500
1AOOO 1-0 1'11 1 CR E GA B fj 200 5~10 G56, I 00
180001-00811 JR T GB c B 350 6260 2'2(1, 500
180001-008/1 JR u GB c B f 250 6510 220. 500

FIG. 7. Defect listing in bridge number order.

about 40 bridges and culverts near Nicosia proceeded smoothly and

preparations are now being made for wider implementation.

ACKNOWLEDGEMENTS

The bridge inventory system prepared for use in Cyprus is based on a similar
system developed by the Northern Ireland Roads Service under its Director,
Mr I. J oiner. The assistance of the Roads Service in allowing use of the
system and in providing the services of a senior engineer to assist with the
training of local staff is gratefully acknowledged. The authors also wish to
thank the Director of the Department of Public Works in Cyprus, Mr M.
Christodoulides, for his support during the course of the project and
permission to publish this paper.
MAINTENANCE STRATEGIES
 12

Bridge Rehabilitation: Department of Transport's

Fifteen-Year Strategy

D. A. HOLLAND and P. H. DAWE

Department of Transport, London, UK

ABSTRACT

The Department of Transport has embarked an a 15-year programme for the

rehabilitation of motorway and other trunk road bridges in England. This
paper gives the background to the programme and describes the different types
of remedial work to be carried out. It outlines the various studies which have
been undertaken to determine the extent ofthe problems and the research and
deve!opment work which is being undertaken to support the remedial work. The
organisation and management of the programme is described. Finally, the
paper Iooks at the problems whichface a major public bridge owner seeking to
keep an important stock ofbridges in a serviceable condition in the face ofeuer-
changing demands and with limited resources.

INTRODUCTION

In November 1987 the Minister for Roads and Trafik announced a 15-year
programme for the rehabilitation of bridges on motorways and all-purpose
trunk roads. This is an indication of the importance which is now being
given to maintaining our existing and ageing facilities and systems in the face
of ever-increasing demands. It is a problern which affects all parts of our
infrastructure and which is being faced by the developed countries
throughout the world.
This paper describes the background to the 15-year rehabilitation
programme for the highway bridges belonging to central government in
England and gives details ofthe various items in the programme. The paper
does not aim to provide a model of bridge management to be followed by
135
136 D. A. Holland and P. H. Dawe

other bridge owners but shows how one country with its own particular
problems and circumstances is setting about the task of managing its stock
of bridges. Moreover, the programme is still in its early stages so there is
much to be learnt about the best ways of tackling the various tasks. In
particular, the problems associated with the assessment and strengthening
of existing structures, some often quite old, are entirely different from those
connected with new design and construction. The paperalso gives details of
the studies which have been undertaken to determine the size and scope of
the programme, and the research which is being carried out to assist in the
efficient and effective execution of the programme.

CENTRAL GOVERNMENT BRIDGES

The transport departments ofEngland, Scotland and Walesare responsible

for 9500 miles of motorways and other trunk roads. Although these
represent only about 4% of the total road network in Great Britain, they
carry 30% of all traffic and 60% of all heavy goods traffic. They are therefore
heavily used routes and vital components of the country's road network.
The Department of Transport in England owns about 8500 highway
bridges and about 3000 other structures, including retaining walls, sign/
signal gantries and tunnels. About 75% of the bridges are concrete, 15%
steel (mainly steeljconcrete composite) and 10% masonry arches. They
range from the Severn Bridge with a main span of 988 m down to small
culverts with spans of 3m. A Iarge number of the bridges are over 100 years
old, the oldest being built in 1185, though the majority are modern bridges
built within the last 20-30 years. Details of all the department's structures
are held on a computerised data base, which also holds information from
inspection reports and details of maintenance expenditure.
There are about 100000 highway bridges altogether in the UK. The
majority belong to local authorities although British Rail owns over 10 000
highway bridges. About 50% of these are concrete, 15% steel (or steel/
concrete composite) and 35% masonry arches. Since most masonry arches
were built over 100 years ago, there is thus a high proportion of older
structures within the national bridge stock, most of these being on local
roads.

FIFTEEN-YEAR REHABILITATION PROGRAMME

The various items of work included in the programme have been grouped
under the following main headings:
 Bridge Rehabilitation: DTp's Fifteen- Year Strategy 137

(a) Steady-state maintenance.

(b) Assessment and strengthening.
(c) Upgrading substandard features.

Steady-State Maintenance
This forms the core ofthe ongoing programme ofrepair and replacement
ofthe various elements ofbridges which have deteriorated or been damaged
as a result of time or use. 1t includes such things as the repair of reinforced
concrete, painting of steel structures, replacement of bearings, expansion
joints, etc. Most of this work is carried out under agreement by agent
authorities and their annual bids for this work have risen considerably over
recent years, reflecting increasing rates of deterioration, particularly of
concrete structures.

Assessment and Strengthening

This part of the programme is concerned with structures which were
designed for loading criteria which are no Ionger adequate for current traffic.
A programme for assessing the older short span structures has started and is
mainly concerned with bridges which were built before the introduction of
national loading standards in 1922. Many of these bridges are masonry
arches built in the late 1800s. A subsequent programmewill deal with long
span bridges, with spans greater than 50 m, where traffic loading has
increased due to the increased numbers ofheavy goods vehicles on the roads.
lt may also be necessary to Iook at some of the moremodern short/medium
span bridges which could be deficient in shear resistance.

Upgrading Sub-standard Features

This is concerned with rectifying deficiencies in certain structures where
current design standards and specifications arenot being met, mainly those
involving safety and durability. lt is also intended to deal with particular
problems which have been identified for prestressed concrete bridges. The
following are some of the topics to be dealt with under this heading:
(i) Waterproofing unprotected bridge decks.
(ii) Rehabilitation of post-tensioned psc bridges.
(iii) Repairs to prestressed precast concrete beams with deflected
tendons.
(iv) Replacement of sub-standard vehicle parapets.
(v) Countermeasures (structural) to the 'bashing' of low headroom
bridges.
(vi) Strengthening of piers and columns to resist higher impact forces.
(vii) Health and safety aspects of access to structures.
lt should be noted that some of these items, although of a 'one-off' nature,
138 D. A. Holland and P. H. Dawe

can be carried out as part of the on-going steady-state maintenance

programme.

Programme
A comprehensive 15-year programme has been developed to cover all the
items of work identified above. The length and formulation of the
programme has been determined by the need to even out the demands on
resources and to avoid too much disruption to the road network at any one
time. Butthis has had tobe balanced by the need to complete certain items of
work where safety is at risk in a reasonable time. It is estimated that the total
cost ofthe programmewill be between flOOOm and f1500m (1987 estimate).
Other bridge owners have similar problems to those of the department and
will need to take similar steps to restore the state of their bridges. However,
the total costs of their work cannot be estimated at present.

BACKGROUND STUDIES

By the late l960s it had become clear that the great expansion of road
transport which had taken place as a result of post-war economic
development was putting the UK's bridge stock under considerable strain.
'Operation Bridgeguard' was an interim programme to identify and deal
with a limited number of the weakest bridges, pending implementation of a
more thorough programme of assessment and strengthening. 'Bridgeguard'
commenced around 1970 and by the early 1980s the bulk of the work had
been completed. Preparations were then put in hand for the basis of the
present 15-year programme. The first problern was the need to assess and
strengthen the older short span bridges as a result of the introduction of a
new Bridge Assessment Code in 1984. The codewas produced by a working
party representing all the major public bridge owners and was to replace an
earlier code which had allowed the use of reduced margins of safety. There
was also a need to organise, in a rational way, the work necessary to tackle a
number of diverse bridge problems which had by then come to light.

Bridge Assessment Code

This is a comprehensive document which covers all aspects of the
assessment of the load-carrying capacity of a highway bridge. It comprises a
mandatory standard which is supported by a complementary advice note.
The code adopts the Iimit state format and deals with inspection, Joading
and strength assessment. The code is intended to be used with a rationalised
system of weight Iimit signing for those structures which are found to be
incapable of carrying the full traffic loading. The code at present makes
reference to existing design codes for the strength assessment of the various
structural elements.
 Bridge Rehahilitation: DTp's Fifteen-Year Strategy 139

Because it is intended for older structures the code deals with such
materials as cast iron, wrought iron and early steel. lt also contains an
empirical method for assessing the capacity of masonry arch bridges. In
addition, it has simplified, but conservative, methods for the assessment of
beam and slab-type bridges. The loading was re-derived from scratch and
takes account of overloading and lateral bunching of vehicles and is
intended to be fully representative of the effects of current traffic.

Bridge Census and Sampie Survey

In order to assess the implications of applying the code to the UK stock of
bridges a study was undertaken with the help of the other major bridge
owners. This consisted of a census to determine the number ofbridges likely
tobe affected by the code together with the assessment of a sample of about
550 structures randomly selccted tobe representative ofthe bridgestock as a
whole. The results ofthe study showed that about 50000 bridges in the UK
were likely to be affected by the code, of which about 11 000 would need
strengthening to meet current standards. No DTp bridges were included in
the study but from the results and knowledge of the DTp stock it was
estimated that about 2000 trunk road bridges would need tobe assessed, of
which about I 000 might need strengthening or replacement.

Long Span Bridges

The loading developed for the assessment code has been adopted as the
basis for a revision to the short span end of the loading for design. At the
sametime a new loading for long span (greater than 50 m) bridges has been
developed since it was realised that the existing loading did not fully
represent thc effects of current traffic which now contains a much higher
proportion of heavy goods vehicles than originally assumcd. The loading
was derived using probabilistic methods operating on up-to-date vehicle
and traffic data and showed that the design loading needed tobe doubled for
the Ionger spans. Thcre is therefore a need to carry out fairly urgently a
programme of assessment of cxisting long span bridges to the new loading,
though it must be remernbered that for thcse structures the dead Ioad
provieles a fairly high proportion of the total Ioad. There are about 150 long
span bridges belonging to the department which need investigation. The first
of these is the Severn suspension bridge, which is nearing the end of an
extensive repair programme to bring it up to current standards.

Condition of Concrete Bridges

Over the last5--1 0 years most bridge owners will have become aware ofthe
increasing signs of deterioration being exhibited by fairly modern concrete
bridges. M uch ofthe detcrioration can be attributed to the widespread use of
deicing salt causing chloridc damage. Other causcs include carbonation,
alkali-silica reaction, sulphate attack and frost damage. Deterioration can
140 D. A. Holland and P H. Dawe

also be due to poor design and detailing, poor materials, poor workmanship
or inadequate maintenance.
The department owns about 6500 concrete bridges and these are subject
to a general inspection every 2 years and a more rigorous principal
inspection every 6-10 years. However, such inspections involve close visual
inspection and thus only record darnage or deterioration which has
manifested itself. Such inspections can only give a snapshot view ofthe state
ofthe bridgestock at any moment and do not give a clearpicture ofpotential
deterioration and future maintenance needs.
In order to obtain better information on the overall condition of its
concrete bridges the department commissioned a study by consultants of
200 randomly selected but representative concrete bridges. Besides a visual
examination of each structure, half-cell potentials, depths of cover and
depths of carbonation were measured. Samples were taken for analysis of
the cement, chloride and sulphate contents, and cores were examined
petrographically to check for the presence of or susceptibility to alkali-silica
reaction.
The study did not reveal any new problems but indicated that the general
state of the concrete bridge stock was worse than had been anticipated. The
report estimates the level of maintenance funding likely to be needed to
tackle these pro blems over the next 10-15 years. The report also
recommends a programme of maintenance works, including a crash
programme to replace or repair damaged expansion joints and the
impregnation with a hydrophobilising material of some existing structures.
Much ofthe work identified in the report had been allowed for in the 15-year
rehabilitation programme but some adjustments have had to be made to
ensure that the recommendations are covered on a priority basis.

RESEARCH AND DEVELOPMENT

The studies described above have been mainly concerned with identifying
the extent of various bridge problems and developing assessment criteria.
Other work is taking place which will help in finding solutions to the
problems or help to avoid them in the future. For instance, there is a
comprehensive programme of research being conducted by the Transport
and Road Research Labaratory concerned with concrete repair methods
and materials. This is complemented by work which is aimed at improving
the durability of reinforced concrete in new construction.
The assessment of the carrying capacity of existing structures is a much
more difficult and demanding task than the design of new structures, since
the financial conseq uences of any decision can be serious.lt isamatter which
demands the application of considerable judgement by the engineer,
 Bridge Rehabilitation: DTp's Fifteen- Year Strategy 141

especially when dealing with older structures whose forms of construction

are very different from modern practice. TRRL have nearly completed a
comprehensive investigation into the real strength of masonry arch bridges.
This has involved full-scale collapse tests as well as model tests and the
development of analytical methods. The outcome will be a more realistic
method for assessing the safe carrying capacity of this form of structure
which represents about a third of the UK bridge stock. Further full-scale
testing is being done on structures incorporating steel trough decking and on
beam bridges with brick jack arches between the beams. Work is also being
done on testing elements such as precast beams salvaged from existing
structures. All this work will enable the real strength of existing structures to
be determined with greater confidence and prevent them from being
unnecessarily replaced or strengthened.
In assessing the strength of the elements in existing structures it is realised
that design codes may not provide the most appropriate criteria, even
though modern codes are based on the latest research. Very often the criteria
in design codes have been subject to simplifying and conservative
adjustments which are oflittle consequence in the design but may be critical
for assessment. Design code models may also be associated with a partial
safety factor which is intended to give the required margin of safety over a
range of element sizes and dimensions, whereas for the structure in question
the margin may be greater than intended. The department has therefore
sponsored the production of an assessment version of the national concrete
bridge design code, BS 5400: Part 4. This has involved a clause by clause
examination of the code and the production of alternative assessment
clauses where appropriate and worthwhile. lt is intended to carry out a
similar procedure for the steel design code, BS 5400: Part 3. Here the main
problern is that detailing practices, provision of stiffeners, etc., in the older
structures may not be covered by current codes.
In 1987 a post-tensioned prestressed concrete segmental bridge in Wales
collapsed due to corrosion of the prestressing tendons. The corrosion
originated from the use of deicing salts which had percolated into the joints
between the segments. As a result TRRL have instigated a study of nine
typical segmental bridges to see whether there are similar signs of corrosion.
The study will also develop procedures for the inspection of this type of
bridge so that if it is feit necessary to examine all post-tensioned bridges the
programme can be instituted without delay.

MANAGEMENTOFPROGRAMME

The maintenance of the department's bridges and roads is coordinated,

programmed and funded by the Highway's Maintenance Division. The
142 D. A. Holland and P. H. Dawe

work is supervised by staff in the nine regional offices and is normally carried
out by agents acting for the department. These maintenance agents are
usually the local authorities (county or metropolitan) although British Rail
and private consultants act as agents in some cases. The maintenance agents
are also responsible for handling the annual round of bridge inspections.
Basedon the results of these inspections each agent prepares, for the region's
approval, bids for the work it proposes to undertake in the following year.
The bids are assigned priorities according to a fairly crude and subjective
five-point rating system and the available funds allocated accordingly.
Through the use of a system of standard forms all this information is stored
in the structures database and can be interrogated as desired.
Many of the items identified in the 15-year programme will be handled
under the procedures described above for steady-state maintenance. For
instance, a certain amount of replacement of sub-standard parapets will be
carried out each year as part of the general maintenance cycle. However,
other items require a separately identified programme and dedicated
funding, as for the assessment and strengthening of the older short span
bridges. Here the bridges concerned have been identified and programmes
prepared to complete their assessment in 3-5 years. Any strengthening work
found necessary will be carried out on a priority basis and the aim is to
complete all the assessments and strengthening work on trunk road bridges
within about 10 years. Completion of this work is expected to have been
achieved prior to the admission of the heavier 'European' lorries, both
domestic and from other community countries, in 1999.
With the possibility of a structure suffhing from more than one shortfall it
will be important to ensure that all the required work is carried out at one go.
This will be coordinated at a locallevel by the regional offices, who will also
determine priorities based on advice produced by HQ divisions. The recent
announcement by the government of an expanded road programme
includes the widening of many existing heavily used routes. lt is expected
that a good nurober ofthe structures earmarked for remedial work under the
15-year programmewill also be affected by the widening programme. The
regional offices are also likely to be required to liaise with the local
authorities over the assessment and strengthening of the older bridges on
local roads. The object would be to provide a comprehensive network ofthe
more important roads for use by the heavier vehicles from 1999.

MANAGING A STOCK OF BRIDGES

Although this paper has concentrated on the department's 15-year bridge

rehabilitation programme it has in a sense just been describing some of the
 Bridge Rehabilitation: DTp's Fifteen- Year Strategy 143

typical problems faced by the owner of a !arge stock of highway bridges.

Because bridges are long-life structures it is difficult to forecast accurately
the loading they will be required to carry throughout their entire life. Design
standards may change as the results of new research become available.
Public expectations in terms of safety and reliability may change. Procedures
for ensuring the free movement of traffic in all weathers may have
unforeseen effects on rates of deterioration. Thus, in addition to the normal
on-going maintenance required to keep structures exposed to the elements in
good condition, the owner also has to be prepared for discrete programmes
of work to bring his stock up to current standards.
The problems ofthe owner are exacerbated by the need to minimise traffic
disruption while he carries out this work. In the case of the department's
bridges the problern is even more severe due to the heavy volumes oftraffle
carried on the national roads. The owner also has limited resources at his
disposal and the nature of the work is often time consuming and labour
intensive. All this demands great professionalism and expertise from the
bridge owner in looking after his bridge stock. He needs to be sure that the
money spent on remedial work is spent in the most cost effective manner,
taking account of traffic delay costs, and that repairs are carried out at the
opportune time without compromising public safety.
In order to improve the way that it tackles its bridge management the
department is undertaking a study of the economics of bridge maintenance.
Part 1 of the study has been completed and has reviewed the present
structures management procedures. The report has recommended the
development of a computerised bridge management system which would
optimise maintenance strategies based on best value for money and rank
maintenance activities according to benefit/cost ratios. The BMS would be
based on the existing structures data base and would be operated by the
regional offices and agent authorities on a network basis. The report
contains a specification for a proposed system which will form the basis of
the development work tobe carried out in Part 2 of the study. At the time of
writing, it is expected that the work will be started in 1989 for completion in
1990. Meanwhile, management decisions are continuing to be made on the
basis of the existing evaluation system. Besides optimising the effectiveness
of bridge maintenance expenditure, the systemwill provide the information
necessary to justify the bids made for bridge maintenance funding m
competition with the other demands upon the government's purse.
 13

Comparative Maintenance Costs of Different Bridge

Types

DAVID LEE

Chairman, G. Maunsell & Partners,

Yeaman Hause, 63 Croydon Raad, Landan SE20 7TP, UK

ABSTRACT

Absolute maintenance casts. excluding thasefor routine cleaning, painting and

resurj'acing activities, are not generally available. In practice maintenance af
bridges is virtually non-existent and bridge owners are only driven to take
action when .. ;>rious deterioration hecomes apparent.
This paper discusses aspects of hridge design and construction \rhich from
experience have an impact 011 deterioration. It is important to take stock of
these aspects so that hoth maintenance strategies and design in future are
improved. (See The Assessment oj' Highlt'ay Bridges and Structures. Bridge
Census a11d Smnple Survey. Department of Transport. January 1987; Repart
011 the State of' Roads and Bridges in the United Kingdam- The Case fbr
Action. The Institution of Ciril Enginccrs, Ja11uary 1988.)

INTRODUCTION

This paper attempts to review various problems of maintenance in different

bridge types and the related costs involved. Inevitably the obtaining of
maintenance cost information is difficult. 1 Some of the comments made
must therefore bc subjective in nature with the expectation that discussion of
the subject may be enriched by the experience of maintenance engineers
willing to support or, in some cases, refute some of the assertions in the
paper.
In general, there are three main maintenance aspects arising from different
types of bridge structure.
(a) Those designed and built in the hope that maintenance costs will be
non-existent or very low.
145
146 DaPid Lee

(b) Those low first cost structures built without regard for ongoing
maintenance issues.
(c) Structures of either (a) and (b) above but which are Iet down by design
deficiencies and construction defects.
lt follows that for the three types (a), (b) and (c) noted above there is a
different maintenance cost involved in each dass. In some cases the long-
term application of experience is necessary to improve structures for the
future. At the present time many maintenance programmes require an
element offirefighting action to refurbish as weil as maintain structures in a
reasonable condition.
Whilst there may be exceptions it is a fair generalisation, in the author's
opinion, to say that bridges were not designed in the past to be maintained
and even if a maintenance schedule was prepared it was never followed. The
painting of steelwork generates a sense of realism, normally through the
appearance of ugly rust on bridge steelwork. Repainting is skimped by not
properly removing corrosion material and in some cases painting over dirt
and bird droppings.
The Iack of maintenance on expansion and rotation joints in bridges is
particularly noticeable. In these and many other details it is better to spend
more up front to achieve quality to reduce maintenance costs. To expect
inferior products to perform weil over a long life is usually expensive.

THE DE-ICING PROBLEM

For concrete bridge decks, columns and abutments the primary form of
attack in the UK is by de-icing salts which Iead to an irreversible chloride
degradation of the concrete. 2 The maintenance cost of a structure has to
balance the cost of de-icing darnage by chlorides. The alternative of using
urea is ten tim es the cost of salt. The additional cost can be justified when in
highly significant areas of the road system. Urea is being used on the
Midland Links Viaducts and the Severn and Avonmouth Bridges, for
example. lt is difficult to imagine that the increase in the cost of de-icing
material is acceptable where there are only limited numbers of bridges per
kilometre of road. Another alternative is the use of calcium magnesium
acetate, which is in fact 40 times the cost of salt.

THE WATERPROOFING PROBLEM

Much darnage follows from inadequacy ofthe waterproofing membrane on

the bridge deck. In some cases relatively efficient waterproofmembranes are
subject to faulty applicat:on or damage. Currently spray-applied plastic
 Maintenance Costs of D(fferent Bridge Types 147

waterproof membranes are coming into favour and certainly offer the
advantage of full bondingover the whole deck without the seam joints. The
drainage of bridge decks usually suffers from the difficulty of sealing
membranes around gullies.
This country has been very slow to adopt the use of concrete surface
treatments and with hindsight it can be seen that ifthese had been applied at
the time of construction much of the maintenance costs now arising might
have been substantially reduced.

PERFORMANCE OF CONCRETE BRIDGES

The Department of Transport is responsible for about 8900 bridges on the

motorway and trunk road network in England. Some 5900 of these are
nominally of concrete with a span cf at least 3m and excluding masonry
arches. Over 75% ofthem have been built since 1960. Two hundred bridges
were chosen at random and the maintenance costs of concrete in them has
been studied. 3 By establishing the performance of the concrete in service the
potential repair and preventative maintenance costs have been estimated.
Of the 200 sampled, 30% were classified as in good condition, 50% as fair
and 20% as in poor condition. In the next 10 years at present-day prices
some f21 million needs to be spent on the 200 bridges in the survey.
The good bridges indicate that investigation, surface treatmentandrepair
of joints req uires on average f 16 150 per bridge. F or bridges in fair condition
it is necessary to monitor condition as weil as the above items, leading to a
mean cost of approximately f69 150 per bridge. For the bridges in poor
condition maintenance strategies would also embrace cutting-out and
repair, cathodic protection where appropriate and replacement of whole or
part ofthe bridge, leading to a mean cost of f314 950 per bridge. The average
cost of a typical unit bridge would be approximately f600 000 at present-day
pnces.
Concrete bridge repairs expenditure has been running in England at the
rate of about f20 million per annum but the study recommends that an
averagerate off60 million per annum be implemented over the next 10-15
years, with a peak expenditure in any one year of nearly f140 million. The
rate of expenditure is thus very approximately flOOOO per bridge per
annum.

MAINTENANCE STRATEGIES

An enormous influence on dealing with actual or potential maintenance

costs is how one determines the past, present and future strategies. It has
148 David Lee

been previously stated that much maintenance was ignored which might be
termed the 'do nothing' strategy. Such a strategy is just one of the options
which may be setout in a form of array table where the maintenance option
can be selected on one axis and the time and date runs along the other. This
technique has been used in an unpublished joint working paper for the
Midlands Links Motorway Viaducts. In view of the assistance this array
method gives to optimising maintenance costs and the date they are carried
out, a brief description will be given belowI
for a particular reinforced
concrete portal frame. There are five potential classifications.
1. Uncorroded-no sign of deterioration.
2. Corroding but not delaminated.
3. Corroding with delamination of cover.
4. Severe corrosion with delamination.
5. Deiamination under the main steel causing loss of bond and structural
integrity.
It is possible to split each classification into sub-categories. For example,
item 2 might have a division into the percentage of area of the element
contaminated ·by chlorides such as 25%, 50% or 75%. For structural
integrity conditions 1, 2 and 3 might lead to no action. Condition 4 is not
satisfactory, however, and a condition 3 element requires maintenance or
refurbishment to ensure it does not deteriorate to 4 or 5. In classification 1
there will be some elements which have been contaminated by chloride and
are therefore suspect but which can be determined tobe unlikely or likely to
further deteriorate seriously in the future.
The maintenance cost must be applied to each option such as in the
following example for a concrete crossbeam.
(a) Do nothing, apart from inspection. The maintenance costs are
therefore related to inspection and access costs.
(b) The application of surface coating or impregnation. This option may
be adopted ifthe crossbeam is not already deteriorating and can be of
surface coating such as urethane or acrylic paint types or silane or
siloxane impregnations.
(c) Application of cathodic protection. 4 There is possibly still a
reluctance to take this method of protection seriously but the trials
and tests that have been carried out over a nurober of years suggest
that it has a use in arresting further deterioration. The cost of
cathodic protection is substantially less than options (d) and (e).
(d) Repair of concrete (including patching).
(e) Replacement of the structural member.
The costs of such maintenance options have been estimated for various
 Maintenance Costs of Different Bridge Types 149

TABLE 1
-------------~-~~---~------------

Maintenance option Number of Remarks

cost units

Inspection only and no action 1

Application of coating 3 Maintenance of coating-10 years
2 units
Apply impregnation 1 Maintenance not known
Install cathodic protection 30-40 Propping not required
I nstall cathodic protection 210360 Props required (one use only)
Install cathodic protection II 0-260 (ten re-uses)
Maintenance of CP 5 Over I 0 years
Replacement of concrete 25% 200
100% 340

elements of the Midland Links Viaducts and for work on the Tees Viaduct.
The quotation of costs is fraught with misinterpretation. To consider the
order of magnitude, assume that one unit represents approximately f1000
and Table 1 suggests the units that various options will generate. Similar
figures can be evaluated from strategies of strengthening and repair, and
cathodic protection in combination.
For larger viaducts (such as the Tees Viaduct) the crossbeam replacement
figures have been estimated to increase by 100-150 units per item. Such
estimates indicate that refurbishment may require the expenditure of sums
commensurate with or exceeding the first cost of construction.

DESIGN STRATEGfES

The following suggestions are offered to improve the design of bridge

schemes.

Durability, Simplicity and Cost

There is room for improvement in many bridge designs by adoption of
simple solutions. Examples include the greater use of simple, straight right
spans rather than unnecessarily curved and skewed. The saving in cost of
building simple structures which act and flex in a predictable fashion have to
be measured against any theoretical saving of a difficult design.

Spans and Articulation

Where appearance allows it is preferable to use larger spans to reduce the
number of intermediate columns. Given a free choice every additional
column detail is a recipe for additional maintenance costs.
150 David Lee

Continuity is preferable provided the potential shrinkage and creep

cracking in concrete decks is adequately combatted. Greater consideration
could be given to simply supported spans ifthey are linked over the column
supports by a continuity detail of the deck which would resist leaking.
The use of continuous bridge decks has always been a useful criterion to
assist waterproofing and by minimising the number of expansion joints in a
bridge deck. In the case ofthe Tees Viaduct the simply supported composite
steel spans are being changed in the refurbishment contract to allow
continuity over each pair of spans with an improved joint between the pairs.

Maintenance as a Design Criterion

We do not put on one shirt and wonder what to do when it needs
laundering. Yet this is precisely what we do for bridges. Much greater
provision for inspection and access for maintenance is required.
All bridges require adequate safe access. Expensive scaffolding is
frequently necessary and the cost of providing this often delays and deters
inspection programmes. The cliche that prevention is better than eure is
particularly forcefully demonstrated in the maintenance of bridge
construction.

Prestressed Bridges
lt seems that external tendons for prestressing will be preferred in the
future to ducts within the concrete section to facilitate simple inspection and
maintenance. Interna! ducting will presumably follow the design of concrete
nuclear pressure vessels where tendons are protected with special grease and
also provide the non-bonded action.
Prestressed concrete bridge beams produced on the longline system have
a relatively good record with high durability. Owing to the low capital
investment in our prestressed concrete factories, however, there have been
instances of poor quality through Iack of adequate quality control.
In reviewing composite steel construction this paper mentions the use of
cathodic protection but with prestressed concrete this is not recommended
in case the generation of hydrogen Ieads to hydrogen embrittlement of the
prestressing tendons.
The quality of long line precast beams should be matched by a higher
quality of erection and on-site completion.

Cable-supported Bridges
The primary support of cable-stayed bridges is naturally the cables
themselves and it is logical to locate these so that inspection, maintenance or
replacement can be performed clear of any traffic on the deck. A similar
consideration would be applied to the hangers of suspension bridges.
 Maintenance Costs of Different Bridge Types 151

The current state of the art for cable-supported bridges has not
established a particular standard method of tower design and the cables
connected to it. lt would seem logical to splice cables at the tower in
appropriate cases so making them easier to replace. Of course, in other
smaller bridges continuity of the cable stays may not present an access or
weight problem.

Composite Steel Bridges

Steel beams are supplied from the fabricator and he is not generally
involved in the quality of erection and completion ofthe composite deck and
other details. The danger of such division of responsibility is usually
compounded if alternative elements or designs are accepted whether in steel
or concrete.
Many steel details in existing bridges have not properly identified the
requirement ofwater-shedding and weathering. lt is fair to claim in principle
that with a proper painting schedule and an allowance for loss of parent
meta! owing to corrosion a steel bridge can be brought back to the as-new
condition. This is a lesson that should also be applied to concrete elements of
bridges. One of the dramatic features of the current maintenance and
refurbishment programme is the realisation that concrete elements very
often perform extremely weil and are virtually maintenance free but that
many suffer from chloride attack or cracking and in these cases it is
extremely difficult tobring the structure back to an as-new condition. At the
very least protective coatings should be considered far more seriously.
For composite construction the development of a glass-reinforced plastic
system to enclose the steelwork, and at the same time provide access, is
generating considerable interest currently on the refurbishment contract for
the Tees Viaduct.

Bridge Furniture
lt would seem logical to provide more space to allow robust detailing and
maintenance of parapets and bridge edges, lighting masts and sign gantries.
The detailing of safety barriers to allow tidal ftow, repair lanes and access of
that type has not had any attention. With major routes having at least four
lanes in each direction, investigation should proceed on intermediate safety
barriers between groups of lanes.

Regional Influences
The international nature of much bridge design is self evident. Each
region of the UK should have design criteria to enhance robustness and
durability. The British weather requires designs which reftect a response to
the service conditions. lt is important to clarify attitudes before the free for
all onslaught of post-1992.
152 David Lee

BRIDGE PROCUREMENT

On bridge construction in the UK it is somewhat unfortunate that

contractors have fallen into the habit by the competitive tendering and
claims procedures of only aiming at the lowest possible standard of
construction that could possibly be approved by the dient or his agents. This
is abetted by the nature of the standards being used which encourage the
notion of only just requiring them tobe exceeded to achieve compliance. It is
not economical to demand high quality when it is not required but equally
when engineers demand a certain standard of quality there is usually an
outhurst by the suppliers and contractors in that unnecessary restrictions
are being placed and unnecessarily high quality is being requested.
Whilst good relations between contractors and site supervision is the rule
rather than the exception this happy state of affairs seems tobe only due to
the essential good nature of personal working relationships. In reality these
relationships are strained because contractors do not expect to supervise
their own work; they leave it to the resident engineer and his staffto do all the
enforcing of the specification. This accounts for the negative image
consultant site staff present to many workmen on site. Another adverse
feature of bridge construction contracts is that frequently the time
programme is paramount and with the lowest possible cost puts quality well
into third place.
These remarks may be regarded as over-critical but the upshot is not
conducive to lowering maintenance costs during service life of bridges.

CONCLUSION

lt will be seen that in the author's view maintenance costs for bridges are
similar to estimating the length of a piece of string. Thorough inspection will
elucidate the quality of the materials of the string and a rough idea of its
length but costs would clearly not be reportable in an absolute manner. For
this reason this paper has attempted to indicate areas where maintenance
costs may be both increased or reduced according to type and detailed
design of the bridge, and with the ultimate aim of seeking bridge designs
which may have a forecastable maintenance cost in the future.

REFERENCES

1. Evaluation of maintenance costs in comparing alternative designs for highway

structures. Departmental Standard BD36/88 and Departmental Advice Note
BA28/88, Department of Transport, 1988.
 Maintenanc·e Costs of Different Bridge Types 153

2. BoAM, K. J., Concrete Highway Structures-Current Experience oflnvestigation

and Repair. Concrete Repair-Problems, Questions and Answers. Palladium
Publications, London, March 1988.
3. WALLBANK, E. J., The performance of concrete in bridges. A Survey of 200
Highway Bridges. HMSO, London, April 1989.
4. BoAM, K. J., Impact of cathodic protection on civil engineering. Second
International Conference on Cathodic Protection, Stratford upon Avon, June
1989.
 14
Programmed Maintenance of Motorway Bridges:
I talian Experience in the use of 'Expert Systems'

G. CAMOMILLA,a A. DRAGOTTI,a G. NEBB!Ab and M. RüMAGNOLOa

aAutostrade SpA, Via A. Bergamini, 50-00159 Rome, ltaly
b SPEA-lngegneria Europea, Via Cornaggia 10, Milan 20123, Italy

ABSTRACT

The maintenance managementprogram is designed to prouide the motorway

network manager with the tools needed to program maintenance of an entire
aggregate of structures of widely uarying structural, enuironmental and
'generational' characteristics. Using specially designed management sojiware
one can then deuelop ouerall assessments ofstructure reliability and thus obtain
a classification on the basis oj' their state of conseruation. The reliability
assessment automatically expressed by the computer ( the 'expert' system) is of
a global nature. i.e. aimed at creating an objectiue Iist of'interuention priorities
for the entire population oj' the structures under management.

INTRODUCTION

The Autostrade Company of the IRI-ITALSTAT Group, responsible for

managing some 3000 km of motorway containing around 3000 bridges and
viaducts exceeding 10m in length, utilises a series of control and
maintenance systems coordinated within the framework of the partially
automated SAMOA program (surveillance, auscultation, maintenance of
structures ).
Under this program, which requires classification and recording of
motorway structures according to their different structural components, one
can obtain either a detailed analysis of their construction characteristics or
an accurate survey of the defects present, selected according to ease of
identification or their importance in assessing the state of health of the
structure. To avoid subjective interpretations by different inspectors, data
155
156 G. Camomi/la, A. Dragotti, G. Nebbia and M. Romagnolo

forms are compiled according to codified methods spelled out in special

operator manuals, and then input into a system of compatible computers
located throughout the territory covered by the Autostrade network.
The maintenance works are always of a preventive nature, ranging from
ordinary 'conservative' maintenance to more elaborate repair interventions.
In some cases this may be preferable to the effect of complete restoration or a
functional improvement, to l!Pgrade the existing structure to cope with
increased load conditions or to implement antiseismic protection utilising
dissipative systems of more recent development. At this point one can also
assess the advisability of modifying the original static configuration, with
the advantage of extending the presumable working life of the structure.

AUTOMATED MANAGEMENT PROGRAM

The Autostrade Company has organised its maintenance of motorway

bridges and viaducts within the framework ofthe so-called SAMOA project,
which is aimed at developing the tools to achieve completely programmed
operations.
The high average age of the structures and their growing numbers have
reached such a point that surveillance based solely on visual inspection of
the state of the structures and diagnosis based on individual inspector
experience has become increasingly unreliable. lt thus became necessary to
devise specific criteria to ensure more uniform and so und assessment of the
state of 'health' of certain structures, but also to predict the probable
evolution of this state over time.
Starting from this global assessment or diagnosis of the entire
'population' of structures under surveillance, it is possible to extract three
distinct subgroups of structures (or structure components):
---certainly reliable;
---certainly requiring maintenance; and
-not perfectly defined in terms of reliability.
The two subgroups in precarious or suspect condition, presumably
comprising a very limited number of items with respect to the total
'population', are then subjected to more detailed inspection (and thence
diagnosis) using instruments and tests more sophisticated than simple visual
inspection.
The specific aims of the SAMOA program are as follows:
(a) creation of a data base and related software for the management of
the morphological data on the structures and the maintenance
interventions performed on them;
 Programmed Maintenance of Motorway Bridges 157

(b) research and development of rapid, non-destructive control systems

for automatic acquisition of such data; and
(c) development of structural verification programs to assess the Ievel of
safety and need for intervention.

Implementation of this project permits, starting from the comparative

analysis of the various data contained in the data base, the rational
definition of intervention needs and priorities. The flow chart of this
procedure (which constitutes the entire SAMOA project) is shown in Fig. 1.
The first part of management activity consisted in recording the
registration data on all the structures in the network. Drawing on design
and cost data from existing files, combined with information obtained on
possible site visits, the next step was to compile special 'morphological
records' for each ofthe structural elements comprising the structure. Various
type groupings were defined such as, for example, piers, foundations and all
other elements making up the structures, indicating the dimensions and
compositions occurring in the individual structure in question.
Each structure was thus sub-divided into its constitutive elements and
these latter reported, case by case, on the number of records required.
Generalsummary records refer to these latter and combine them with other
information either pre-existing or acquired subsequently such as, for
example, general data on the environment where the structure is situated, the
construction methods employed, static tests conducted, subsequent
maintenance and repair interventions.
These data, subsequently memorised on computer, constitute the
historical data base ofthe structures. Employing software able to select/sort
on the basis ofthe historical data, the structures can then be classified on the
basis of element types in such a way as to visualise the appearance in the
manner in which the human operator is more accustomed to viewing it. As
far as inspections are concerned, the surveillance is performed essentially by
means of close visual examination or binoculars of the individual structural
parts, to spot and take note of any defects which may be present, with
particular care to observe the development of those already noted
previously. Visual checks, despite their limitations with respect to the
difficult but not always impossible task of identifying hidden defects or
accessing parts concealed from view (e.g. parts below ground Ievel,
inaccessible heads of prestressed concrete beams, etc.), remain the
fundamental method of surveillance able to provide at least general
indications of the overall state of conservation of the structures.
In the case of those structures for which further detailed examination is
considered advisable, however, recourse is made to more sophisticated
methods based essentially on local and/or global non-destructive type tests.
 .......
V.
00

I
Data bank
r------1----l
I I
Instrument Required List of structures 0
I surveys I inspections (or parts thereof)
I Ir- and survey in good condition
I I of defects
~c
I I :::!
I I ~~
I I ::t:..
I I
Instrumental b
....,
Morphology Seriousness ..------
of structure of defects analysis ~
c
......
~I f-
......
1
I
I .-·
I good
I good 0
I I
List of structures Assessment
~
<::J-
: Environment H- Algorithm <::J-
1--------_.J for global s (or parts thereof) of condition i:;•
assessment suspected §
~ 1::>...

~
bad
~
Special Weli-defined Undefined c
f-- ~
intervention Intervention defects defects :::!
1---
techniques costs ~c
l ~ l
c-
I Updating I I Works I
1 program 1
Budget
1
I data base
I
FIG. 1. Flow chart.
 Programmed Maintenance of Motorway Bridges 159

Considerable advances are occurring and are under study in this sector, but
there already exist a number of different weil known and widely utilised
techniques, which will be described in a following section. In order to
guarantee the greatest possible uniformity in classifying the different types
of defects and the most important parameters for their description and
control, use is made of 'defect charts' on which such characteristics are
defined in univocal fashion.
In fact the inspections are conducted in accordance with a special manual
on darnage survey methods containing not just the record forms for
recording the defects but also detailed explanations of these defects
(including descriptions and photographs), as weil as instructions on how to
fill out the forms reporting the locations and extent of the deterioration
observed. By breaking down the structure into its individual structural
components it is possible to conduct an accurate survey of all the defects
present, using precise and codified methods such as to Iimit subjective
interpretations by the inspectors.
The record forms, as can be seen from the example shown in Fig. 2 of a
deck consisting of sliding cable prestressed reinforced concrete beams, are
organised in such a way as to divide the structural parts into further specific
elements to which the defects observed are referred. In the case in question,
the deck is first divided into its constitutive elements (beams, crossbeams and
slab) and each ofthese is then divided into fields: for each beam, for example,
a field consists of a section located between two consecutive crossbeams.
After the forms have been compiled in the field, the data are then
entered into the computer, where they are processed according to special
interpretative algorithms. The resulting output can take the form of an
overall but at the same time objective assessment of condition of the single
element, of the single structural component or of the structure as a whole,
depending on the information desired.
The processing software contains defect assessments of varying degrees
of seriousness in relation to the type of structure involved (supported
beam, continuous beam, framework, arch, etc.), the component materials
(reinforced concrete or prestressed reinforced concrete, steel, masonry, etc.),
as weil as the extent and location ofthe deterioration. This is combined with
instrument measurements and/or considerations regarding the evolution of
certain physical parameters measured geometrically (e.g. inertia moments)
or instrumentally (e.g. measuring vibration modes or using other non-
destructive methods) on the structures themselves.
The processing criteria are such as to permit a 'global assessment' of the
state of the structure or of its single component parts, which serves as the
basis for establishing criteria for intervention priority. In fact the global
assessment, which is quick to use and yet reasonably accurate, serves to
SCHEDA E 3: DIFETTI IMPALCATI 0\
Numero elemento tl Numero campata @Jil Foglio n. 8J 0
-
IMPALCATI A TRAVATA IN CA P. IFili scorrev I Carr D0 Sg E0 Numero di travi IT Solette· CAO IRl CAP. D
Codice opera ffi~l~zl Data rilevamento l2J4J iil~J l[[j Numero di trasversi I() 141 Trasversi CAO. gj CAP. D
TRAVI TRASVERSI SOLETTE
1 2 3 5 6 sx A B c D E DX SX 1 2 3 5 DX
• • 0
::-~ -- ./ v f.--
./ ]:-.. --~- Q
V ."'- !.,...- SF .-
_"'b V f.--
31
0
~ I -~-
"" ...... ....... 31
,_- · / V I--
D
--- ~-- k:::: - --- !"""
"" 1"-...
,..... D
--r- f-" ~~
--- - ::-,.., I-- L.--"' -=<
I f.-- r-- ::t...
L ....... _...-r: r--. f-" r-..
~ .J&I Ii IIL .4q t /.1 " l-1'1 ....<i ~ --·-- f- t;:,
....,
J ~-14-
+~ 3/t 19 jl, "'
J 19 -t- r-
1-1~ -- - Jt. II__ 4
II ~l_ ~
-" ~ t4- f4 ' ~
I--I--
~4
~
"- VOJ tt t1 -
f- I-- -- 1- - 1--- -r-- g
so
~ -!4 .q I A .Ii
.-·
~--
S_j~ :H IJ ~~ &~ Jlt "' _1_ ~ -~ z.u 'I
/"}_ A_ ~
I!'_ ~ 19 J _B_ I
1 rr rf= -· -- - - -· 0
8
.:1-~ r---
-t- - -·
~
c:J-
t.l.ft_ .q 3J, A~ ~q_
z. IJ-' _'j I A4 Al; ." 1-9 !II} c:J-
"' ·-t-- I-- -1-- f--1 ~
1'4 I"' ~ II ~~ .3z 19 ss·
A A
!::>
;:::
!::>..
CODICI DIFETII TRAVI ll :A I 3 19
"' !>'A
.5~,q J ""
11 Macehre dt umrdrta 561 Lesront rn cornspondpn SI<, II 3Ft, ,q )I<, II f.-- ~
21 Cis drlavato/ammalorato
1... ·-
31 v~spar
za der ferrr d'armatura
951 Rrduzrone sezrone ar
I
1_<9lt!::!
__J_ __ .,__
SI
f- t-- ·-
[::::-
,......
- ~
111 Atmatura ordrnana sco
OErta e;o ossrdata
matura precompressa
971 Les10nr longrtudrnalr
u .__
-- 0
31
51 lesronr rn cornsoondenza estradosso bulbo travr COOICI OIFEITI TRASVERSI CODICI DIFEITI SOLETTE
staffc 981 Orstacco sprgolr
61 Lesronr caprllan ancoraggr 11 Macchte dt umtdtta 11 Macchte dt umidttj
lestate dt ancoraggro
n 1031 Umrdrt.3 dall"rntcrno 21 Cis dliavato/d'llmalorato 21 Cis dllavato/ammalorato
~0
non srgtllate I solo per c a p cav1 )I Ve:>P,ll 31 vespat
81 Drstacco tamponr testate scorrev J 41 Armatura ordtnana scoperta elo osstdata 41 Armatura ordtnana scoperta elo osstdata Cl
91 Lesrom su anrma lungo 1061 Lestont attacco trave so 51 Lestonl tn cornsponden?a slaffe 71 Testate dt ancoraggto non stgtllate
cavr !etta 151 Lestont vert1calt 141 lestOnt modeste/diffuse
101 Lesron1 lungo suola bulbo 1001 Go; dtlavato/ammalorato 161 Lestont diagonal! 161 Les1ont dtagonall
11 I Guarne rn vrsta testate 171 Dtsti'!Cco trav1 - travers1 301 L€SIOnt longttud1nali
121 Guarne marote F1li ossJ· 1091 Armature scooertelos 341 Staffe scopertelosstdate 721 Les1ont trasversali
datr stdate testate 561 Les1ont 1n cornspondenza det fem d1 armatura
341 Staffe scopcrtelossrd<~te 1101 GU<ltne non mtas<lt(' 1071 Fuortusctta barre ancoraggto

FIG. 2
 Programmed Maintenance of Motorway Bridges 161

separate the certainly reliable structures from those which require more
detailed examination.
These latter must be examined with special attention (measurements,
determination of restoration measures to be undertaken) and hence will be
subject to more frequent surveillance or, if necessary, constant monitoring.
The reliable structures, on the other hand, will only be subjected to ordinary
routine surveillance, save for the repetition every 2 years of the reliability
inspection related to the 'global assessment'.
Before examining the data processing methods in detail, it would be
advisable to look first at the special techniques employed in structure
surveillance and the criteria employed in determining restoration
interventions, so as to have a better understanding of the procedures
followed in coping with the various problems of structure management.

SPECIAL TESTS

In presenting the surveillance methods it was mentioned earlier that,

contrary to visual inspection procedures, the specialised control techniques
are employed on an ad hoc basis when the need arises to have more detailed
information on a single structure or on those located on a particular section.
Such needs may be dictated by different circumstances:
(a) visual inspection and subsequent data processing may leave room for
various interpretations as to the actual state of the structure and the
seriousness of the situation;
(b) planning maintenance or restoration operations may require more
specific measurements; and
(c) special research projects.
Examples ofsituations (a) and (b) are quite similar insofar as the problern
is to define either the intervention priority or the manner of intervention; in
the case of point (c), however, the situation is much more varied, and weshall
limit ourselves here simply to a few references. In all cases, however, it is
advisable to deal with the subject starting from the single tests or, better,
from the description ofthe criteria for the utilisation ofthese in an integrated
manner, without lengthy treatment of the single methods. Proper
integration of the various methods used is fundamental for valid
interpretation of the data measured; indeed only in this way is it possible
to effect a type of iteration which, taking into account the various
approximation factors proper to each ofthe systems, permits one to obtain a
realistic 'quantification' of the state of conservation.
In particular, for type (a) and type (b) situations the following tests are
normally conducted either on specific parts or sample areas.
162 G. Camomilla, A. Dragotti, G. Nehhia and M. Romagnolo

-Ultrasonic: for the following purposes:

• to supply indications ofthe homogeneity ofthe concrete by means of
transparency and/or indirect tests;
• to provide indications ofthe depth of cracks with indirect tests across
the cracks or alongside them; and
• to determine, via correlation of the data obtained, the ultrasonic
transmission velocities along with the recoil index obtained at the
same points using sclerometric tests, concrete compression strength
values.
-Sclerometer: this traditional non-destructive test provides information
on the compression resistance correlated with the recoil index, but if
used by itselfit is limited to providing information tied very closely only
to the 'cortical' layer. For this reason it must always be used in
conjunction with ultrasonic tests. Such integration has proven to be
optimal in 'filtering' the results from the inftuence of a number of
variables (granulometry, moisture, hinder dosage, type of aggregate)
which in many instances register contrary values with the two
methods.
-Pull-out: this is utilised mainly in cases where it is impossible to effect
uhrasonie tests by transparency (on 'chunky' elements), or where it is
deemed useful to obtain compression strength data for comparison
with that obtained from other non-destructive tests; these data are
always combined with pacometer data regarding the position of the
superficial reinforcing, so as to ensure that the piece is inserted in a
suitable position such that extraction is not blocked by possible
presence of reinforcing rods or mesh.
-Windsor probe test: in using this test the same applies as was said in the
case ofthe pull-out test; the combination oftbistest with the preceding
one is always useful, especially if there is a Iack of information
regarding the aggregate grain size and characteristics.
-Measurement of carbonation depth: this test is always employed, as it
provides an important index of the aggressivity of a particular
environment on a given composition of concrete, and thus provides
extremely useful infonnation for detailed studies of mix design for
durable concretes; it is superftuous to stress the value of a test which
permits the designer to know whether or not one can still hypothesise
passivation around the reinforcing.
-Measurement of chloride penetration: this is used generally only on
exposed parts: pulvinos, heads of beams, heads of crossbeams. lt may
be direct (samples taken from the structure) or indirect, based on
electrochemical potential measurements conducted on accessible
structure surfaces and linked tagether by isopotential curves (mapping).
 Programmed Maintenance of Motorway Bridges 163

Obviously this Iist of tests is far from complete, but mention is made here
only of those which constitute part of the routine system which, as set out
above, automatically comes into play once certain 'alarm thresholds',
defined in assembling the visual inspection data, are exceeded.
Besides these tests, which can be defined as 'local' insofar as they provide
punctual results on the state of the structure, we also utilise global-type
systems which yield information on the characteristics of the structure as a
whole, obtained from static loading tests or measurements under dynamic
exci ta tion.
Dynamic analysis is the most commonly used test, aimed at determining
the vibration modes ofthe various parts ofthe bridge under a known source
of vibration. The method is not very sensitive to deterioration of the
structure itself and cannot, at the present time, be used to assess the actual
state of structures. lt can, however, be used to memorise the state of the
structure at a certain moment, so as then to draw a comparison with the
results of other dynamic tests conducted subsequently in time. Certain
recent developments in survey systems and finite element modelling would
seem tobe promising for the practical use ofthis method. Besides their use in
maintenance design, these measurements can provide new criteria for
assessing 'sample defects' which can subsequently be incorporated in the
global assessment and applied in expert systems to determine SAMOA
intervention priorities.
In this way the bridge management system is continually improved over
time: constantly increasing its data base and consequently improving the
work of the human operators who continue to manage weil the more
traditional tasks; thus we have surveillance, auscultation and intervention as
integrated moments of a single process, with the substantial but hardly
"tyrannical' assistance of the computer.
In concluding this section, mention should be made of the various
research objectives we will be aiming at in the use of the above tests and
others. On the one hand, we will be trying to understand better the Iimits of
the various systems utilised and obtain more detailed information on those
not routinely used and, on the other hand, we will also attempt to develop
hypotheses regarding the process of deterioration with respect to pre-
determined types of environment and structure.
In the case of the first of these two aspects we are already conducting
careful analyses ofthe data as they are gathered, so as to improve design and
programming of interventions, also by increasing the number and
distribution of these tests and checking the results by means of destructive
tests and controls employed during partial demolitions conducted in the
course of repair operations.
So as to gain a better understanding of other systems (e.g. potential
164 G. Camomilla, A. Dragotti, G. Nebbia and M. Romagnolo

mapping, release of tensions, etc.), trials are conducted in conjunction with

both destructive and non-destructive tests on a case by case basis in specific
situations.
Work continues toward the development of 'deterioration curves over
time' (a sort of measurement of the speed of deterioration under standard
conditions-of crucial import~nce for a company responsible for the
maintenance of such a large nurober of structures) as a function of the
environment and the structural and technological characteristics. We have
begun to conduct on-site tests to determine the depth of carbonation on
sample elements of selected structures of particular significance. These tests
will naturally be repeated at regular intervals (around every 2 years, as
already stated), so as to construct curves of the type indicated, also with a
view toward future projections.

INTERVENTIONS ON STRUCTURES

Starting with the list of priorities provided automatically by the computer

program, and following whatever specialised testing may be required, we
proceed to the design ofthe intervention in accordance with defined general
reference criteria.
Distinctions are made among different possible interventions, which can
be grouped as follows:

-Ordinary or extraordinary maintenance: these are distinguished

according to the type and importance of the intervention to be
performed, and are intended to maintain the structure at full efficiency
in accordance with its original design characteristics.
-Static restoration: when one wishes to restore the original bearing
capacity of the structure.
-Retrofitting (functional maintenance): when one decides to maintain
unchanged the original geometry and static scheme while at the same
time permitting the structure tobe subjected to actionseither greater or
differing from those for which it was initially designed.
-Restructuration: this entails the alteration of the original static scheme
of the structure and/or its geometric characteristics so as to restore or
increase its bearing capacity.

Naturally the last two items can be adopted also where the state of
conservation of the structure is judged to be good.
At this point we shall illustrate several special intervention techniques
employed on various parts of a bridge structure.
 Programmed Maintenance of Motorway Bridges 165

With regard to protective techniques (to restore the protective function of

certain parts ofthe structure, which may have deteriorated or not have been
supplied in the first place), the following may be employed:

For piers:
-plating with meta! mesh and rheoplastic mortars to restore protective
covering of exposed reinforcing;
-light hooping where it is necessary to provide a certain additional
strength to columns (the reinforcing is rendered continuous by
suitable overlapping beyond the edges); and
-localised repairs, followed by application of protective paint to prevent
carbonation of the concrete cover layer over the reinforcing.
For slabs:
-surface repairs with stuccos or small castings andjor waterproofing of
the deck with synthetic membranes or other systems;
-more extensive repairs and waterproofing of specific points (drains,
heads of prestressed beams, etc.); and
-diffuse anode cathodic protection of outer reinforcing to eliminate
chlorine ions present in the existing concrete.
For beams:
-application of protective paints with or without repairs of unfilled
sheaths.

In the case of repairs to restore bearing or bonding function which has been
lost or not suitably provided initially, on the other hand, operations might
include the following:
For piers:
-Substitute lining (new castings which by themselves support the total
Ioad on the pier) of columns with reinforcing and form casting using
non-shrink rheoplastic concretes;
-transverse strengthening of piers with additional septa; and
-upgrading according to more strict criteria to better resist transversal
seismic action.
For slabs:
-reconstruction of the upper layers of the slab, including elimination of
degraded material by high-power hydrodemolition, application of
new reinforcing and substitution with non-shrink rheoplastic
concrete.
For beams:
-strengthening with steel plate elements glued and/or bolted to the
existing structure (thin plate reinforcing); and
166 G. Camomilla, A. Dragotti, G. Nebbia and M. Romagnolo

-addition of external cables to existing prestressed or non-prestressed

structures torestorelost prestressing or provide it anew (positioning
of the cables with respect to the beam cross-section, and that of
anchor and drawing points, will vary depending on the space
available).

The preceding provides an overview of possible interventions normally

adopted. It must be stressed, however, that in the majority of cases the aim is
essentially to obtain a functional improvement ofthe structure as a whole, as
well as of its component parts.
Types of intervention which could be cited as particularly successful
examples are integral decks (jointing of decks), the introduction of special
unidirectional and multidirectional bearing devices as well as seismic
protection devices, and lastly the installation of breakthrough~proof
guardrails capable of dissipating impact energy.
The problems faced by the designer of maintenance interventions on
existing structures are considerably more complex than is the case with the
design of new ones. Some of the additional constraints encountered include
the need to operate in the presence of traffic; the need to select materials
which yield reliable results within a very short period of time and in the
presence of traffic-induced vibrations; the very strict Iimits on intervention
time; the need to assess material strengths in situations where degradation is
present; and the need to achieve higher performance standards than those
adopted in the original design of the structure.
It is easy to see from the above considerations that the concept of
maintenance thus assumes the more 'noble' connotations of the term
'project' (i.e. that of 'projecting the mind into the future so as to foresee in
advance' what the work will require), thus superseding the rather simplistic
notion of those who see no need for a specific design stagein the concept of
maintenance.

DATA PROCESSING METHOD

At this point we would like to provide an example of how one goes about
setting up an expert system of classification for purposes of controlling the
state of the art of the structures under management.
Given the fact that the assessment is conducted separately for each
component part of the structure, we shall concentrate our attention on a
specific application, in this case that of a deck.
As shown in the attached sample ftow chart for decks (see Figs 3 and 4),
starting from the analysis of the morphological characteristics, the
 set.l!cr BY
UGI$TIUoTIOW ~TA

~
~
SEl.fCT BY i:5
iEGJ5fAA.TIO N DATA
~
!tE.LfCT &Y r---I!FIN AL ~
iEGISTRATlON n-TA
a-.IWTM
l:l..
"'
t:=J ..-
SEU:CTrN
~
RE.Ciii~TitATION DATo\ CJ = s·
c=:~:rv ;;;
SELECT BY ;::
REGI$fRATION O.TA 1::>
;::
SEUCTI>V
REC:I&S.T1l-.TION DATA
"''"'
~
~
8"
....
~"'
~
~
~

,_
0\
-.l
FIG. 3
168 G. Camomilla, A. Dragotti, G. Nebbia and M. Romagnolo

CAlCUl.ATlON OF "SlGtJaiTS• BY BEAM VS DEFECT (l!EAM i,2, ... )

ß,~-------------------------J

ßt

CAI.CUIATIOW Of "SlGM:NTS" 8Y CJlOSS BE.AM VS DEI'ECT (CR8. i 12, ... )

VIEWPOINT: "SMETY"

~------------·---------~---------------------
ß~~-------------------------J

FIG. 4
 Programmed Maintenance of Motorway Bridges 169

~_::>=---.. --...

I SIAl! J
CALC. Of 'SEGIIENIS' BY ELEMENTS IIS tlEF'ECT(D .. S•,S,,S., lnt-1 s,..n.) I
". .. ... J ......... .
I TABLE Of DEFECTS BY ElEMENTS'
.' . .. .
,.----r.____, J l
I<-----.----'
VP "SAFETY"I I VP: 'CONSERVATION"I
,.---'--1-__,
:'WAltRPROOF'ING~ I
I VP lvP: ........................ I
~----~~1 Ir~----~
I OE fINIT ION QF DEFECT a.ASSES I
~~------------~~

I ELEt.4E'IT c:v.ss .I
1:.~~-~~::_··: ::: I
--~;;=

~~p-s---~J ~~~if"'~v
lt::, I
1...~,.,....,.......,..,...,.,......~
I =-~~i;~~·:..ll I
&r•S.: ................•........
I SPAN CLASS
CIAU :4"'.!'1······ .......
...... :11, .. , ............ .
II
l~··_··_ _ _ _ _ _ _ L_I_ _ _ _ _ __ ,_ _ _ _ _ _ _ _ _ _~-----

I CAL.WlATION r:JI' 'SEGMENTS' 8Y STRIP VS DEFECT (S..O., 1,2, ....)

.............. 1. ............ .
I
I TAIU: Of oEfECIS BY STRIP I
..... ..
( ( l
I VP: 'CONSERVATION • VP:'IriA~· VP: ....................... .

'l 'l ,----J("--------'

I DEFINITION OF DEFECTOASSES I
r~----~~----TI'~==~~====~----
STRIP CIASS t, :l,.. . . r:.S::-::T-=cllt:::-P-::C,-:V:-=155::-- 5• o, t t

:::: t::::. :=:~~---- I

·-· ............... ..
OVEIW.J. ~NIILY
.. ·······
MJm4L ASSEt.4BI.Y
.. ······ .............. .
SPAN CV-55

rm~~
C'l..üS :4~2"":-----··.
PJt,Mtol: rt,f\, ....

-~~~;;;l'-~--_··___·_·_·_rL·_·_··_·_··_·_·_··_··_··_,·,_·-··_.. _··_··_·_··_··_··-----···_·_T..___· ~~-~~

~6
FIG. 4~contd.
170 G. Camomilla, A. Dragotti, G. Nebbia and M. Romagnolo

processing algorithm performs different specific tasks depending on the type

of deck itself. Selecting as our example the case of a deck consisting of simply
supported beams of prestressed reinforced concrete (form E3) with sliding
cables (a structure quite common on the Autostrade network), we now
proceed to a detailed description of the procedure.
The steps followed are not always the same but will depend on the
objective one wishes to achieve. The choice among the different analysis
viewpoints takes the form of a selection from among the distinct 'viewpoints'
presented: structural safety, state of conservation, waterproofing, etc. lt is
necessary to select the viewpoint (VP) at the very beginning, as in this way the
processing operation can be carried out automatically without interruption.
At this selection Ievel it is also possible to choose whether one wishes to
have only the final output, or also the intermediate outputs corresponding to
the various partial processing stages. The intermediate outputs constitute a
partial processing of the codified defects reported on the forms compiled
during the site inspections. These partial outputs, e.g. the 'table of defects by
beam' (or by crossbeam or by slab element) provide, in the case of each
element (beam, crossbeam or slab field), a 'photograph' of the associated
defects and related intensity/extent parameters.
The assernblies of the various elements are more complex, as they will
differ for each of the above-mentioned viewpoints, and will provide data
strings indicating a Ievel of 'seriousness' and function parameters of the
intensity /extent of the defects for all of the beams, crossbeams or slab
elements taken together.
Another important partial output useful for further assessment of a non-
automatic nature is the schematic diagram, which summarises the defect
distribution of an entire span (beams plus crossbeams plus slab).
The final partial output consists of the aggregation of all the spans of a
structure. Theseare sub-divided into groups, each of which is distinguished
by appropriate separation thresholds. These latter serve to filter the
parameters defined in processing each ofthe spans, even where thesearenot
morphologically similar (simple reinforced concrete beams, prestressed
reinforced concrete beams, box girders, etc.).
The groups associated with each span are organised according to two
different output possibilities:
-aggregation of spans by single structure, with related indication of
group; and
-aggregation of spans by group, with related indication of the structure
to which they belong.
These outputs provide a sufficiently complete indication of the state of the
structure, always in accordance with the viewpoint selected. In this case, to
 Programmed Maintenance of Motorway Bridges 171

obtain a more synthetic unit view, we have opted here for assembly of all the
spans of each structure. In fact the final output is represented by an overall
assessment of the conditions of all the decks of a structure, and the ranking
of this latter in a condition classification in relation to the condition ratings
ofvarious structures of a selected population. The assembly operation takes
into account spans having different structural characteristics, and hence
different processing procedures. 'Homogenisation' of the spans is achieved
by attribution of thresholds which aredifferent foreachtype of record form;
the thresholds thus constitute the point of equivalence in comparing spans
of different types.
The operation proceeds as follows for the different viewpoints:
'Safety': The group of spans of a structure is classified in a determined
order equivalent to the preceding groups. If at least one span is classed as
group 1, the whole group ofspans is offirst order, and so on. Furthermore,
each order is attributed percentages relating to the spans contained in each
group, with respect to the total number of spans in the structure as a whole.
'Conservation': The group of spans of a structure is classified in a
determined order resulting from processing the data on the group.; to which
each span belongs. First, the frequencies of the groups are calculated. The
order of a given structure corresponds to the mean of the groups present
weighted according to their relative frequencies as in the following
expresswn:
- 'ffg
g=- (1)
n
The order value is linked to a second parameter consisting ofthe variance of
the frequency distribution, calculated as follows:
s2 = 'if(g- g)2
(2)
n
Depending on the viewpoint, each structure thus remains associated with an
order and certain percentages or a variance.
In classifying all the structures of a selected population, these are first
sorted in descending rank, from those of highest order to those of lowest
order. In the case of the safety VP the ranking is determined by relative
percentages in reference to the order, whereas for the conservation VP the
position is determined by the variance, in inverse proportion to the value.
 15
Engineering Management of the Tamar Bridge

W. I. HALSE
Bridges Division. Mott MacDonald Civil Ltd,
20/26 Wellesley Road, Croydon, Surrey CR9 2UL, UK
and
R. L. C. STEPHENS
County Surveyors Department, Cornwall County Council,
County Hall, Truro, Cornwall TRI 3BE, UK

ABSTRACT

The Tamar Bridge (Fig. 1) was constructed within the period 1959-61 as a
private toll crossing for the Tamar Bridge and Torpoint Ferry Joint
Committee, comprising representatives of the Cornwall County Council and
the Plymouth City Council. Since the end of the maintenance period, the
engineers responsible for the design and supervision of construction ( M ott,
Hay and Anderson) haue been retained by thejoint committee to undertake an
annual inspection of the bridge and its immediate approaches, and to report and
make recommendations on any necessary maintenance. Apart from small or ad
hoc items arranged through the Joint engineers, the county surveyor of
Cornwall and the City Engineer of Plymouth, subsequent works haue invariably
been arranged and supervised by those same consultants who, through close
Iiaison with the bridge owners and their staff, haue effectively provided a
professional service for the long-term management of maintenance activities
byforecasting.financial commitments and arranging the operationsaraund the
increasingly restrictive limitations imposed by traffic densities.

INTRODUCTION

When opened to traffic on 24 October 1961 the Tamar Bridge (Fig. 1), with its
335m main span, became the Iongest span road bridge in the UK and 26th
equal in the league table for 'suspension bridges of the world'. Since then it
173
174 W I. Halse and R. L. C. Stephens

FIG. I. Tamar Bridge.

has been successively relegated in the UK by the Forth, Severn and Humber
road bridges, and has disappeared without trace from world rankings.
However, the joint enterprise with the Torpoint Ferry still maintains a
unique record amongst UK tolled estuarial crossings,l with no capital debts
(discharged by 1981 ), a regular operating surplus without subsirlies and the
lowest tolls. 2 Since 1980 tolls on the bridge have been payable in one
direction only, and whereas the charge per car was 3 shillings (15p) each way
in 1961 the current cost is 40p cash to east-bound travellers, effectively
reducing to 20p for regular users taking advantage of the concessionary
vouchers available for pre-purchase.

PRE-CONSTRUCTION

Vehicular ferry crossings ofthe River Tamar existed at Torpoint and Saltash
from the early 19th century, but by the 1920s local agitation had already
begun for a fixed crossing to supplement or replace those increasingly
inadequate facilities and as early as 1931 a scheme for a high-level bridge
between Torpoint and Devonport was prepared but abandoned on
objection from the Admiralty. The post-war boom in traffic produced
further impetusandin 1950 the Cornwall County Council and the Plymouth
City Council formed a committee to pursue the matter with the Ministry of
Transport.
 Engineering Management of the Tamar Bridge 175

By 1955 it became obvious that, despite a favourable report the previous

year from the technical panel appointed by the minister, the matter was
unlikely to be pursued in the immediate future and the two councils
determined to investigate a private joint enterprise.
The consultants, appointed in September 1955, reported on bridge and
tunnel alternatives in April 1956, recommending a fixed crossing on the
present alignment, and a firm decision to proceed was taken by the councils
in July. Such was the considered degree of urgency and of the cooperation
afforded by all parties that the consultants and parliamentary agents
prepared and deposited a bill in Parliament by November 1956 and Royal
Assent was obtained at the end of July 1957.
Unfortunately the then current financial climate delayed Treasury
consent to even private borrowing, and it was not until the autumn of 1958
that the engineers were appointed to prepare the design, following which,
continuing the afore-mentioned cooperation, tenders were invited in March,
a contract was Iet in June and work commenced on site on 7 July 1959.
The construction details have been described elsewhere. 3 Suffice it to state
that the 152 mm thick deck in 43 N/mm 2 concrete is supported on steel
stringers carried on cross-girders at 9·17 m centres slung by suspenders from
the main cables spaced at 15·24 m centres, with traditional steel Warren truss
stiffening girders. Extensive use was made of box sections in high tensile
(BS 968) steel, shop-fabricated by welding, with site Connectionsmade with
friction-grip Torshear bolts. The 10·06 m wide three-lane carriageway and
twin 1·83 m footways were surfaced in hand-laid mastic asphalt at 38 and
20 mm thickness respectively.

TRAFFIC CONDITIONS

Hindsight is a wonderful gift and a golden opportunity to the local media,

who from time to time, prompted by some traffic delay through accident or
other, pontificate on the folly ofthe planners in building a three-lane bridge.
Figure 2 provides a histogram of recorded annual traffic fl.ows across the
Tamar Bridge, rising sharply and inexorably from 1·8 million vehicles in
1962 (the first full year of trafficking) to the anticipated total of 12 million
vehicles in 1989. The authors venture to suggest that had the engineers
included such a prediction in their original report then at very least their
professional judgement would have been called into question, if not their
report rejected outright.
The effect of the surprisingly regular increment of some 375 000 vehicles
per annum exhibited every year for 27 years, although marginally alleviated
overall by the introduction of tidal fl.ow in 1974 and one-way toll collection
176 W /. Halse and R. L. C. Stephens

12 UNIFORM RATE ~,----

- 11 OF 375,000 VEH./YR. .....--i""=
10
0 10 PER YEAR -------........ - __.._ : -

~ 9- "'"'"/r-

8-
------- r-
:::>
:::>
z 7
z<(
6
er:
~ 5
4
"'
w
-'
u 3
:;:
~ 2
1
0 il
I I I I I I I I I I I I I I I
1960 1965 1970 1975 1980 1985 1990
YEAR

FIG. 2. Histogram of annual traffic ftows.

in 1980, has increasingly contributed to problems in arranging maintenance

works requiring lane possessions. Whereas up to the late 1970s it was
possible to consider and implement closure of one lane for up to a week, for
example to construct and eure replacement epoxy nosings at damaged deck
joints, a decade later lane closures for the major refurbishment works
subsequently described had tobe stringently restricted to avoid traffic chaos
and maintain reasonable public relations.
These latter operations had to be executed outside the commuter peak
hours, within working sessions from 09.30 to 15.30 hours and again after
18.30 hours, on weekdays. The main opportunity for progress ostensibly lay
on Saturdays and Sundays when a dawn-to-dusk lane closure was
acceptable, unless Plymouth Argyle were playing at home (most of their
supporters appear to reside in Cornwall) and the returning fans merged with
their compatriots returning from Saturday shopping in Plymouth at around
16.30 hours, or unless Sunday morning dawned dry and sunny, at which a
substantial proportion of south Devon residents travelled to Cornwall,
spreading their departures through the morning but always reappearing at
the west end of the bridge around teatime.
Again, come the first weekend in July, the summer visitors arrive in droves
and it becomes impossible to even consider lane closures until school
holidays end in mid-September but latterly extended into early October by a
high proportion of senior citizens. Recuperating from the combination of
grandchildren and school holidays? Any work during January, always a
hostile month weather-wise, can be discounted. Leave out 10 days at Easter
and a week each at the MayDay and spring bank holidays to cater for local
holiday-takers and a proportion of early tourists, mix in a few wet weekdays
in east Cornwall, when shopping in Plymouth becomes an attraction to
those early visitors, and garnish, as in 1987, with the wettest 1une for 70 years
 Engineering Management of the Tamar Bridge 177

and the overall mixture becomes a recipe for ulcers to the programmers
charged with limiting traffic disruption to a bare minimum.

ROUTINE INSPECTION

A longitudinal walkway is provided each side of the bridge immediately

below deck Ievel extending to some 600m from back-to-back of the
anchorages and two more at bottarn chord Ievel across the suspended
structure. These, supplemented by hand ropes along the main cables, ladders
from deck Ievel up to the saddles and down to the caissons inside both main
towers and ready access at both river banks for low tide inspection and
binocular viewing, provide adequate permanent facilities for the routine
annual inspections carried out by the engineers.
A reduction in the main span camber became apparent early on, and in
1964 an extensive set of annual measurements was instituted to record
relative levels of 'fixed' points, e.g. anchorages, side and main towers;
cambers of the side and main spans; out-of-plumb of the main towers (by
nearly 70 m long plumb lines, permanently installed within the north legs);
and relative dimensions between the ends of the suspended structure
steelwork and the concrete faces of the towers.
After a decade and with the only significant recorded variations being a
continuing lass ofmain span camber andinward leaning ofthe main towers,
it was concluded that the effect was due, despite repeated prestressing by the

250

•f 200
0 ~TAMAR SU SPEN Sia-l
z 335m. 1961
..:
~ 150 I
Vl
I
~u
378m. 1 954
I'
100 '
/ ~DEUTZ ISEVERINSI
I, CAEJLE~STAYED

50
,(

10 20 15 25 30
LO'ENED TO ;RAFFIG TIME IYEARSI

FIG. 3. Creep-time relationship (locked-coil ropes).

178 W I. Halse and R. L. C. Stephens

rope makers prior to delivery, to permanent stretch (creep) under load in the
locked coil wire ropes forming the main cables and suspenders. Little
published data were available, but such as could be obtained appeared to
agree in reasonable measure with these early observations. The effect has
continued but at a reducing rate, as illustrated by the derived creep~time
relationship plotted in Fig. 3, tagether with comparable data. 4

ROUTINE MAINTENANCE

Apart from weekly sweeping, quarterly gulley cleaning and 6-monthly

charging of the grease nipples to the pins of the suspended structure,
arranged by the joint engineers, repainting of the structure has been
undertaken every 7 years (1968, 1975, 1982 and 1989) on the recommenda-
tion of the engineers, virtually as a routine maintenance operation.
The choice of the initial protection system, consisting of grit-blasting, zinc
metal spray, PVB etch primer followed by two undercoats and one gloss coat
(each of phenolic white Iead), had to be made at an early stage of the trials,
which eventually resulted in the development of the 'Forth Bridge' system
and which subsequently reigned unchallenged in bridgeworks protection for
over a decade. Unfortunately, despite a good early performance in these
trials, the white Iead coats proved less impermeable than the eventually
chosen micaceous iron oxidetop coats, and by 1968 some considerable areas
of the steelwork had to be blast cleaned to remove the underlying zinc
corrosion from the meta! spray coating. These and defective areas in
subsequent repaints were two-coat primed with zinc epoxy initially and
latterly zinc phosphate, followed by two coats of MIO, and completed in
1968 and 1975 with white Iead undercoat and finish, varied in 1982 to an
alkyd medium and in 1989 to a silicone alkyd.
As a result of this regular sequence, repainting has largely become
cosmetic, and the cost of remedial treatment of isolated defective areas,
principally on the north side below deck through Iack of exposure to drying
by the sun, does not normally exceed 10% of the total contract sum. The
close proximity of houses and parked cars at each end of the bridge and
moarings in the river beneath preclude spray application of paint, but the
extensive use of box sections in the construction allows under- and finish-
coating to some 90% ofthe area ofthe suspended structure tobe applied by
roller, reducing time and hence labour and scaffold hire cost.
Conversely, the successive repaints have highlighted the increasing
pressure for enhanced safety measures in the construction industry. The cost
of providing and maintaining the necessary scaffolding and safety nets
below deck, expressed as an on-cost percentage to other (labour, plant and
 Engineering Management of the Tamar Bridge 179

materials) charges, has risen from 15% in 1975, through 47% in 1982 to 64%
in 1989, and actually reached 91% in one unsuccessful tender for the latter.

1960s

The early years proved free of major problems other than that of zinc
corrosion, affecting the 1968 repaint.
By this date the annual traffic density had already more than doubled and
preliminary consideration was given to some upgrading of the facility. A
widening ofthe carriageway would be restricted to 11·6 m by the main tower
legs, providing 2·9 m wide substandard lanes, and there seemed little
advantage in any alteration on the bridge without attention to both
approaches, in the responsibility of Cornwall and Devon County Councils.

1970s

Various schemes were pursued, none ofwhich proved acceptable to all three
parties. Eventually a tidal flow systemwas introduced on the bridge in 1974,
controlled by signals on new overhead gantries, more acceptable than
sub-standard lanes, equally effective in commuter peak hours and less
expensive.
By 1978 densities had again more than doubled in a decade and thoughts
were directed to further alleviation. The toll booths formed a restriction but
tolls had to be maintained to pay off the (rapidly diminishing) construction
loan, provide for maintenance and subsidise the uneconomic unified toll
charges on the Torpoint Ferry. Accordingly an enabling act was prepared,
receiving Royal Assent in 1979, by which in 1980 toll collection was limited
to one way (eastbound), thus relieving afternoon congestion on the
Plymouth side which, on occasion of accident or other mishap, had been
known to back up some 5 miles to the city centre.
Over the firsthalf ofthis decade the 60 no. air-gap articulationjoints in the
concrete deck of the suspended span, closed by rubber bitumen filler in the
depth ofthe surfacing, began to cause increasing problems with local asphalt
break-up. In 1975 the worst affected areas were cutout and provided, in lane
widths, with nosings in a recently developed epoxy mortar (Febplate SLS),
which proved more durable than traditional rigid mortars. The remainder
were similarly treated within a 3-year programme, completing in 1979.
The severe winter of 1978/79 created some problems with kerb and
asphalt damage, and raised attention to the not-too-distant need for
resurfacing. Soon after, coinciding with the onset of ASR problems in the
180 W I. Halse and R. L. C. Stephens

South-West, a series of some 60 cores were cut from the deck. When tested
these demonstrated that, as a result of the use of sea-dredged coarse and fine
aggregates in the construction, the reactive content was too high to sustain
an alkali reaction under normal conditions butthat chloride contamination
from de-icing salt had already penetrated some millimetres into the concrete,
although not yet to top reinforcement Ievel. Both conclusions indicated the
need to provide an effective waterproof membrane below any future
res urfacing.

THE EARLY 1980s

These were marked by the continuing incidence of local darnage to the

surfacing and a resurgence of earlier problems at the joints, although
generally to localised areas within wheel-tracks, not directly through failure
of the epoxy mortar but as a result of darnage in the underlying concrete
under the ever-increasing speeds and shear gross weight of traffic.
The formerwas checked in 1983 by a surface dressing with chippings and
the latter contained by virtually biannual spring and autumn ad hoc repairs,
effectively 'buying time' before the now obvious major refurbishment could
be programmed.
In May 1981 one such repair was undertaken using a magnesium
phosphate concrete (Febset 45) on a trial basis and this appeared
encouraging. Without coarse aggregate its flow characteristics obviated
traditional compaction operations and its fast curing under exotherrnie
reaction subsequently proved it capable of direct trafficking within 2·5 h. By
extending its use in subsequent patch repairs, often under conditions and by
procedures not encouraged by the manufacturers, it was demonstrated tobe
virtually idiot-proof, subject to rigid control of water content and mixing
time, which was simplified by the purchase of domestic stainless steel
measuring jugs and a cooking timer from a local branch of Boots.
In late 1983 the engineers were formally instructed by the joint committee
to prepare a report which eventually provided detailed proposals and
estimates for a staged refurbishment of the bridge. These comprised, at
Phase 1, the permanent reconstruction by break-out and reconcreting over a
600 mm width of each of the 64 no. deck articulation joints including, as a
result ofthe reducing rate of cable creep, the 'closing up' oftwo in every three
over the suspended deck sections. Phase 2 would then comprise carriageway
resurfacing and provision of a deck waterproof membrane and include
attention to the footways and kerbing, and drainage improvements.
Following discussion with the joint treasurers on finance availability,
these phases were programmed around February to June working in 1987
 Engineering Management of the Tamar Bridge 181

and 1988, within the lane possession Iimits previously outlined, and the work
extended to include, at Phase 3, the repaint due in 1989. The package
dovetailed into the projected mid/late 1988 completion data for adjacent
DTp works on the Saltashand St Budeaux bypasses immediately either side
of the river.
Fortuitously, the preparation of this report was aided firstly by the
cooperation and practical experiences of the F orth Road BridgeJoint Board
and their consultants, then engaged on a resurfacing ofthe side spansofthat
bridge, including the provision of a sprayed MMA acrylic membrane
(Eliminator) originally developed by British Rail and applied to some 200 of
their bridges before its existence became more widely known. Secondly, by
an unexpected and sudden deterioration ofthe main articulationjoint at the
western side tower at Tamar, requiring the rapid arrangement of a
permanent full-width reconstruction of that joint in late 1984. This
interruptionwas unwelcome at the time but in the event provided a valuable
insight into procedures, materials, plant and labour content of such
reconstruction, leading to proven recommendations and more realistic
estimates for the Phase 1 operations.

A MAJOR REFURBISHMENT

The report was presented in May 1985 and formally accepted by the joint
committee in July, with a commission to proceed with arrangement of the
necessary contracts on an unusual basis, as subsequently described.
As recommended in the report, three 9·17 m bays of one-lane width and
the adjacent footway were stripped of surfacing for a trial in the autumn of
1985, overlain with the acrylic membrane, regulated with a nominal 20 mm
of dense bitumen macadam and surfaced with 20 mm of pervious bitumen
macadam. The membranewas continued up the back-of-kerb and across the
footway area, which was then treated with four alternative proprietary 'thin'
surfacings.
Although not previously used on a full-scale road surfacing in the UK,
pervious macadam wearing course had a long-standing track record on
airport runways and was then being laid extensively on highways in Hong
Kong. lts reduced spray characteristic and virtual freedom from risk of
aqua-planing were proven, but perhaps ofmore immediate advantage to the
maintenance operation was the Iack of need for a chipping spreader, the
charging of which always disrupts traffic in the adjacent lane and the
machine itself inevitably appears prone to breakdown or other malfunction,
disrupting the whole surfacing operation.
The joint committee then accepted a recommendation by the consultants
182 W I. Halse and R. L. C. Stephens

to undertake, und er a pilot works in April 1986, reconstruction of eight joints

in the Saltash side span. This was principally to relocate the starting end of
the 1987 (Phase 1) works some 80m further eastwards, isolating them from
works adjacent to the westend of the bridge, due to commence in March of
that year under the Saltash bypass scheme. After consultation, it was agreed
that this work should be executed under the direct control ofthe engineer, by
the same 'bridge gang' from Cornwall D LO as had undertaken most of the
previous ad hoc deck repairs. This large-scale 'trial' added further realism to
the assessment of optimum sequences, methods, plant and labour
requirements for the main works in the following year, and incidentally
provided sites for trials of three proprietary types of continuous joint
surfacing then being considered for use in Phase 2.
Movement joints are a perpetual headache to maintenance engineers.
There is no universal solution, as witnessed by the multiplicity of proprietary
types on offer at any one time which inevitably go off the market within a
decade or so. The authors believe that for small movements the asphaltic
plug joints as finally used in Phase 2 are currently the best value for money,
despite some defects which occurred during the maintenance period and
which may recur. Laid with reasonable care the riding quality is superb, they
are relatively inexpensive compared with any alternatives on offer and, more
importantly, defective areas can be cutout and replaced within a matter of a
few hours.
The joint reconstruction and the resurfacing works subsequently followed
as planned. The Phase 3 repaint is in progress at the time of drafting and it is
a matter of some satisfaction that the overall refurbishment cost is
anticipated to be just less than the 1985 estimate, after adjustment for
subsequent price fluctuations by published indices. 5

THE APPROACH TO MAINTENANCE

Maintenance works can be tedious and to some engineers may not be as

attractive as heading-up some multi-million pound scheme, but can provide
job satisfaction if approached in the right manner.
A differing degree ofinvolvement is required from that for new works in a
green field. The engineer must become more deeply concerned with minor
details, operational planning and in site supervision. He should be prepared
to consult with the contractors, but must specify the programme in some
detail and maintain a rigid control. Where the works are on a time-and-
materials basis, he must go further by providing and insisting on compliance
with a detailed operational programme to make the most efficient and
economic use of plant and operatives. By avoiding sophistication and
 Engineering Management oj' the Tamar Bridge 183

limiting site works to simple, clearly defined operations in a strict sequence,

small contractors relatively inexperienced in bridgeworks can be used to
advantage and with economy. Common sense and a near instinct to foresee
potential problems are a prerequisite, along with an ability to adapt and
improvise as difficulties are encountered. With a lesser degree of supervisory
assistance than is normal in new works, he must be prepared to spend time
when problems are encountered, observing and suggesting alternative
sequences, procedures or plant. In maintenance works it is neither
acceptable nor economic for the engineer to sit back and wait for a
contractor to put forward solutions.
With new works there is some scope for the engineer to drive the
contractor to provide a superior job at no extra over the contract sum, but in
maintenance this will rarely happen and the dient will only get what he is
prepared to pay for. What the engineer cannot fully see, he cannot describe
and quantify, and what is not fully specified in the tender document cannot
be fairly priced by a tenderer, leaving a potential for later, expensive, claims.
Hence much of the work must be on a time-and-materials or provisional
sum basis, and the experience and judgement of the engineer will be critical
in providing realistic assessments ofthe appropriate amounts tobe included
in the document. Competitive tendering should and can be financially
advantageaus but for maintenance it can also be disastrous. Wherever
possible the contractors should be locally based, and preference should
always be given to those who have previously worked on the site to the
satisfaction of the engineer.

CONTRACTPROCEDURES

Over the years and on the basis of results, the joint committee has gradually
accepted the somewhat unusual procedures recommended by the engineer.
Competitive tendering on a national basis has been restricted to repainting
and even then the same local firm has won the contract on the last three
occasions. Selective tendering by a few local contractors has been limited to
types or extent ofworks not previously or recently undertaken on the bridge,
and on a time-and-materials basis where appropriate. More often
negotiated tenders have been arranged with single contractors, either with
specialists after adequate trials or with those ofprevious proven ability and
experience of working conditions on the bridge.
For the recent refurbishment, the joint committee further extended this
enlightened approach, in acknowledgement of their public responsibility to
road users. With the certain knowledge ofthe Ievel oftraffle due on the first
weekend in July, extensions oftime were not permissible and contracts had a
184 W I. Halse and R. L. C. Stephens

completion daterather than a commencement and a contract period. Whilst

under anormal arrangement the main contractor would take contractual
responsibility for any overrun, the issue would as usual be clouded by alleged
defaults of his subcontractors, weather, deliveries and myriads of other
excuses, and meanwhile traffic chaos could still ensue. Accordingly, the joint
committee determined that the works should be arranged around individual
contracts, with the engineer effectively undertaking the contract manage-
ment and the resident engineer being additionally charged with responsi-
bility for ensuring the works were finished on time each year.
The contractual implications vis-a-vis one contractor's work and
another's, or due to default or poor workmanship, were daunting. The
logistics were no less so. The Phase 2 works finally involved 25 separate
contracts ranging in value from f600 to nearly f300 000. Of these, eight (after
selective or negotiated tender) were sealed and bonded under modified ICE
Conditions of Contract whilst the remainder were against written order
following prior quotations. Six were for the supply of specialist materials,
leaving the engineer directly responsible for ensuring that 80-odd operatives
from the other 19 contractors were present on site at the right time and
executing the work in order and to programme, between mid-February and
end of J une 1988.

MAINTENANCE COSTS

Expenditure on all maintenance has been researched, separately accounting

the costs of repainting, other contract works and a combined allowance for
minor works, fees and supervision costs, which latter has been averaged for
simplicity.
These have been expressed in histogram form in Fig. 4 as a percentage of
the then current bridge valuation, being the original construction cost of
f1·8m at 1959 prices annually updated by output price indices, 5 which has
also been plotted for information on aseparate scale. These do not include
'running costs', e.g. cleansing, lighting, insurance, etc., nor special items such
as maintenance of toll equipment or buildings.
From the histogram it can be deduced that the mean annual cost ofminor
maintenance initially increased but over the past decade has remained
constant at some 0·35% of the annual bridge valuation; that the
corresponding cost of repainting on a 7-year cycle has been 0·3% per annum;
and that other contract maintenance, including a major refurbishment after
more than a quarter of a century, has amounted to some 0·5% per annum. It
should perhaps be noted that these percentages have been related to the
updated construction cost rather than the annual insurance valuation,
 Engineering Management of the Tamar Bridge 185

LEGEND

20
~ REPAINTING D OlHER C<l-ITRACT
W<RKS
~ MIN<R MAIN TENANO:.
~ FE S ANO SUPERV1SIO'l

18

16

...... __ ... -...

CQ-ISTRUCTION
COST- UPOAT 0
- 12 BY 0 . 0 . .INOO:S
~ 10

FIG. 4. Maintenance costs and bridge valuation.

which includes allowance for demolition, fees and supervision, and which in
the instance of Tamar is assessed some 40% higher.

AN OVERVIEW

Above all eise, experience over the 28 years of life of the bridge has
demonstrated the value of long-standing relationships between the owners,
operators, contractors and the engineer in executing, managing and financ-
ing maintenance works. This sharply contrasts to difficulties encountered
elsewhere, arising through the alternative of regularly seeking competitive
tenders for all works, where the sole criterion for acceptance is on the
basis of minimum cost in the short term, and where successful tenderers
have little interest and no incentive to provide a good working relationship.
Previous mention has been made of such cooperation in respect of the
short gestation period between the first serious initiatives in 1955 and
commencement on site in mid-1959. Subsequently the joint committee has
almost invariably accepted the recommendations of the consultant, which
have always been formulated after close, often informal, consultation with
the joint engineers and the bridge and ferry manager. Again the joint
committee has always been prepared to authorise expenditure on trials of
new or unfamiliar materials and has actively encouraged use of more
expensive products or procedures where these would Iead to a lessening of
traffic disturbance.
Over the years this rapport has extended down through the engineers to
the various contractors and their operatives, and the permanent bridge staff.
The toll collectors are at the 'sharp end' when any traffic problems occur and
186 W I. Halse and R. L. C. Stephens

can then be subject to verbal, and sometimes physical, abuse. Invariably

their reaction has been to laugh it off rather than shift the blame to the
engineers and contractors.
Problems can and do arise from time to time during any works but always,
within the harmonious environment that has evolved, these have been
overcome. The nurober of persons who, through past involvement, now feel
they have a personal stake in the success of the Tamar Bridge is legion.

ACKNOWLEDGEMENTS

The authors wish to acknowledge the cooperation afforded them in

preparation of this paper by the joint engineers, Mr B. W. Mansell, County
Surveyor of Cornwall, and Mr R. Fairclough, Acting City Engineer of
Plymouth, and the bridge and ferry manager, Mr A. R. Warren, together
with members of their respective staffs.

REFERENCES

1. Tolled Crossings, Second Report from the Transport Committee, Session

1985-86. House of Commons 250-1. HMSO, London.
2. TRUSCOTT, R. P. and WARREN, A. R., The Tamar crossings. Highways and
Transportation, 35(5) (May 1988) 83-8.
3. ANDERSON, J. K., Tamar Bridge. Proc. Instn Civ. Engrs, 31 (August 1965) 337-60.
4. MosER, K., Time-dependent response of suspension and cable stayed bridges.
Proc. 8th Congress IABSE, New York, 1968, Final Report, pp. 119-29.
5. DEPARTMENT OF THE ENVIRONMENT, Housing and Construction Statistics Part 2,
Quarterly, HMSO, London.
 16

ModeHing and Predicting Bridge Repair and

Maintenance Costs

MüHAMED BüUABAZ a and R. MALCOLM w. HORNERb

a Department of Civil Engineering, b Department of Engineering Management,
University of Dundee, Dundee DD1 4HN, UK

ABSTRACT

Rational decisions about cost-effective bridge designs, optimum replacement

ages and recurrent cost budgets are hampered by the absence of reliable data
on which to base forecasts of repair and maintenance costs. Research into
simple models for predicting the new-build cost of bridges has led to the
development of equally simple models for predicting the costs of repair
contracts exceeding ilOk in value. These models, based on the principle of cost-
sign(ficance, contain only 17 or 18 elements yet are accurate to within 10%.
Analysis of historical data has allowed us to propose a tentative relationship
between the area of a bridge deck and the cost of repairs. The results !end
weight to the view that repair and maintenance costs for masonry bridges are
less than those for reinforced concrete bridges.

INTRODUCTION

Repair and maintenance now account for more than 50% of the
construction industry's turnover, yet very little data are available on which
to base predictions of future costs. Local authorities with large building
stocks have therefore considerable difficulty in justifying maintenance
budgets. Nowhere is this more true than in the roads and bridges
departments. In many cases historical records are incomplete or non-
existent, especially before local government reorganisation. Even when
records do exist they are not consistently structured. The inability to
quantify maintenance and repair costs, however, has further ramifications.
First, it makes the prediction of total life cycle costs impossible. Thus
187
188 Mohamed Bouabaz and R. Maleolm W Horner

investment decisions are based solely on the criterion ofinitial capital cost, a
particularly dangeraus practice when the authorities responsible for capital
and recurrent costs may not be one and the same party. Second, there is no
rational basis on which to choose between replacing and repairing a road or
bridge.
This paper describes some firstfaltering steps towards the identification of
those elements of a bridge which give rise to the majority of repair work and
towards the prediction of its cost. Eventually it is hoped that the work will
lead to a better understanding of the relationship between capital and
recurrent costs. Tentative relationships between deck size, repair costs and
age are also reported.

PREDICTING NEW-BUILD COSTS

Recent work at Dundee University has demonstrated the feasibility of

developing a simpler model of the new-build costs of bridges than those
which currently exist. 1 The models are based on the principle of cost-
significance. It has been known for many years that 80% ofthe value of a bill
of quantities is contained within only 20% of the items; those which are
'cost-significant'. The cost-significant items are easily identified as those
whose value is greater than the mean. The relationship between value and

TABLE 1
Cost model for feasihility design stage ( hridges)
(cost modelfactor=0·73)

CSWP CSWP description Unit

number

Supply and driving and Ioad testing of piles for main piling,
including establishment and moving of piling equipment Item
2 Vertical (85-90°) formwork > 300 mm wide for end supports
and intermediate support
3 Horizontal (0-5°) formwork > 300 mm wide for deck
4 Curved formwork at any inclination > 300 mm wide for
intermediate supports
5 I n-situ concrete
6 Precast concrete members for deck
7 Bar reinforeerneut
8 Paving in paved areas to surfaces > 10° to the horizontal
9 Waterproofing on surfaces > 300 mm sloping up to 45u to
the horizontal mz
10 Supply of parapets Lin.m
 Modelling and Predicting Bridge Repair and Maintenance Costs 189

number of the cost-significant items is closely approximated by a Pareto

curve.
Analysis of several hundred bills has shown that projects can be
categorised in such a way that the cost-significant items within any one
category are more or less the same. Moreover, the cost-significant items can
frequently be grouped into packages exhibiting two important features:
(1) they closely resemble a contractor's site operation; and
(2) a single unit rate can be applied to them.
Thus, within each category of project, cost-significant work packages
(CSWPs) can be identified whose value is a constant proportion (typically
80% or thereabouts) of the total project cost. This has spawned the
development of a new-build cost model containing only ten elements which

TABLE 2
Cost modelfor detailed designstage (hridges)
(cost modelfactor = 0·82)
--------- ~---~-----~ ----------------

CSWP CSWP description Unit

number

1 Establishment of piling equipment for piles in the main piling ltem

2 Moving piling plant for piles in the main piling No.
3 Supply and driving of piles for main piling m3
4 Load testing of piles in main piling No.
5 Vertical (85 90 ) formwork > 300 mm wide for end supports m2
6 Vertical (85-90 ) formwork > 300 mm wide for
intermediate supports
7 Horizontal (0- 5) formwork > 300 mm wide for deck
8 Curvcd formwork at any inclination > 300 mm wide for
intermediate supports
9 In-situ concrete for end supports
10 In-situ concrete for intermediate supports
11 ln-situ concrete for deck
12 Precast concrctc mcmbers for deck
13 Bar reinforcement for end supports
14 Bar reinforcement for intermediate supports
15 Bar reinforcement for deck
16 Paving in paved arcas to surfaces > 10' to the horizontal
17 Waterproofing on surfaccs > 300 mm sloping up to 45o to
the horizontal m2
18 Supply of parapets Lin.m
19 Drainage of end supports Item
20 lmported fill deposited adjacent to structures, including
around structural foundations
21 Void formcrs
190 Mohamed Bouabaz and R. Maleolm W Horner

can be used at the feasibility design stage to predict the cost of a new bridge
to 10%. Refinement ofthe model to include an additional11 elements allows
the cost of a bridge at the detailed design stage to be predicted with an
accuracy of 5%. The feasibility and design stage models for reinforced
concrete bridges are shown in Tables 1 and 2 respectively.
The ratio of the value of packages in the model to the total bill value is
known as the cost model factor (CMF).
On behalf ofthe Scottish Development Department, andin collaboration
with Babtie Shaw and Morton, the model has been built into a computer
package called BRIDGET which allows the price of a new bridge to be
calculated in less than 15 min.Z BRIDGET is now available to all Scottish
Regional Councils.

MODELLING REPAIR COSTS

Success in modelling new-build costs encouraged us to investigate the

possibility of modelling maintenance and repair costs using similar
techniques. Earlier work 3 had shown that the system was most likely to
work on bills of quantities containing at least 50 and preferably over 100
items. Accordingly our study was confined to repair contracts worth more
than about flOk. From our experience of new-build bills we decided to
divide repair work into three categories: masonry arches, masonry concrete
arches (i.e. masonry arches with an infilled concrete deck) and reinforced
concrete bridges. Bills of quantities were obtained from Tayside and Lothian
Regional Councils. The total number of bills available were for masonry
arches 14, masonry concrete arches 13 and for reinforced concrete bridges
24.

60 Predicted bill value (e><1000)

• Actual data
60 - Regression line

40

so

20

10 Coefficient of correlation: r • 0.97

10 20 30 40 60 60
Actual bill value (e><1000)
FIG. I. Predicted versus actual bill value for masonry bridges.
 Modelling and Predicting Bridge Repair and Maintenanc·e Costs 191

TABLE 3
Unit rates for rolled asphall 1vearing courses of'
diff'eren t thicknesses

Thickness Unit rate Unit rate

(mm) (f/m 2 ) (f/m 3 )

40 4·10 102·50
50 5·00 100·00
60 6·24 104·00
----

Methodology
The methodology was similar tothat used in the new-build analyses. Cost-
significant items were identified for each bill, and the results for bills within
each category were inspected for consistency. A variety of techniques were
then used to determine the minimum number of cost-significant work
packages which represented a constant proportion of the total bill value.

Changing units of measurement

In some cases it was found that the unit rate was linearly proportional to a
unit of measurement different from that recommended in the Method of
Measurement for Road and Bridge Works (MMRB). For instance, typical
items for rolled asphalt wearing courses are shown in Table 3. The unit of
measurement for rolled asphalt wearing courses was therefore changed to
cubic metres.

Minor dijferences in unit rates

Differences in rates for items of marginally different specification were
known from our earlier work to be statistically insignificant. Table 4 shows
typical rates for formwork providing different qualities of surface finish.

Trial and error

The cost-significant work packages contributing the smallest average

TABLE 4
Unit rat es for hori:ontal j(Jrmwork ol
diff'eren t c/asses

Class Unit rate

(f/m 3 }

F1 15·28
F2 16·19
F3 16·53
192 Mohamed Bouabaz and R. Maleolm W Horner

proportion ofthe total bill value were progressively deleted from the models
until a balance was struck between size and accuracy of the model. In
general, an accuracy of 10% was our goal.

Rates applied to the model

The accuracy of an estimating system depends on two factors: the
accuracy of the cost model itself and the accuracy of the rat es applied to the
model. Because in the first instance our objective was to test only the
accuracy of the model, the rates used to develop and test the models were
abstracted from the bills of quantities. In those cases where items with
slightly differing rates were combined into a single work package the
weighted mean unit rate was used.

Testing
The models were developed from an analysis for 37 bills of quantities and
tested on a further 14.

RESULTS

Models
The cost-significant work packages for masonry and masonry concrete
arches proved to be identical, so the two were combined into one single
category, 'masonry'. The resulting models for masonry and reinforced
concrete bridges are shown in Tables 5 and 6 respectively.

Accuracy
Foreachbill the ratio ofthe value ofthe cost-significant work packages to
the total bill value was calculated. The results for masonry bridges are shown
in Table 7.
The mean value of the ratio (value of CSWPs)/(total bill value) is the cost
model factor (CMF). The CMF for masonry bridges is thus 0·76. Similar
analysis of reinforced concrete bridge projects yields a CMF of 0·82 with a
standard deviation of 0·08.

Testing
Table 8 shows the results oftesting the cost model for masonry bridges on
the seven bills retained for that purpose.
Figure 1 shows the results of linearly regressing predicted on the actual
bill value.
For reinforced concrete bridges the mean error of prediction was 2·27%,
with a standard deviation of 8·66%. The coefficient of correlation for the
regression of predicted on actual bill value was 0·98.
 Modelling and Predicting Bridge Repair and Maintenance Costs 193

TABLE 5
Cost model for repair of masonry hridges

CSWP CSWP description Unit

number
-----·--·-

1 Preliminaries ltem
2 Excavate any material except rock or reinforced concrete
on bridge superstructure m3
3 Rolled asphalt wearing course in carriageway m3
4 Dense bitumen macadam base course in carriageway m3
5 Precast concrete Lin.m
6 Horizontal formwork more than 300 mm wide mz
7 Vertical formwork more than 300 mm wide mz
8 In-situ concrete m3
9 Gunite 40 N/mm 2 to soffit and vertical surfaces m3
10 Repair of concrete surfaces m3
II Bar reinforcement of any diameter t
12 Tie bars of any diameter No.
13 New masonry with hattered or vertical face m3
14 New general random rubble masonry previously set aside m3
15 Existing general random rubble masonry m3
16 Hand pointing on arch and soffit mz
17 Waterproofing more than 300 mm wide mz
18 Dayworks Sum
--- - - - - - - - ----·--· - - -

TABLE 6
Cost model for repair of reinforced concrete hridges
--···------------------- -------

CSWP CSWP description Unit

number

I Preliminaries Item
2 Excavate unsuitable material in flexible surfacing on
bridgedeck m3
3 Disposal of unsuitable material in tips off site m3
4 Rolled asphalt wearing course in carriageway m3
5 Dense bitumen macadam in carriageway m3
6 Horizontal formwork more than 300 mm wide mz
7 Cut out and scabble off unsound concrete in deck m3
8 Application of epoxy bonding aid to reinforced concrete mz
9 In-situ concrete in screeds m3
10 In-situ concrete m3
II Precast prestressed beams m3
12 Movement joints to bridge (expansion joints) No.
13 Waterproofing more than 300 mm wide mz
14 Dayworks Sum
194 Mohamed Bouaba:: and R. Maleolm W. Horner

TABLE 7
Percentage value of cost-significant work packages for masonry bridges

Project Actual value Value of CSWPs Value of CSWPs

number (f) (f) Actual value
(%)

1 21 709·70 14159·50 65·22

2 7631·55 6242·75 81·80
3 20240·00 15351·00 75·84
4 20087·12 14161·21 70·50
5 50725·90 32208·00 63·50
6 20437·92 14 827·66 72·55
7 54884·50 44550·00 81·17
8 20600·00 14480·31 70·29
9 51 070·60 38 302·80 75·00
10 15900·00 11426·00 71·86
11 18152·35 15 388·05 84·77
12 53 285·00 35661·95 66·93
13 49 592·00 43 770·67 88·26
14 16 632·87 13196·81 79·34
15 23 621·50 21 016·00 88·97
16 30401·15 20 386·86 67·06
17 13621·75 10 362·00 76·07
18 22 595·82 18 509-44 81·91
19 27 269·00 24205·00 88·76
20 14670·95 12210·25 83-23
Mean 76·00
Standard deviation 8·10

TABLE 8
Model tests on masonry bridges
(CMF0-76)

Project Value of Predicted Actual Difference Percentage

number CSWPs value value (f) difference
(f) (f) (f) (%)

1 29196·90 38415·78 45 796·90 -7 381·12 - 16·11

2 16841·80 22160·26 22 231·79 -71·54 -0·32
3 16267·09 21404·06 19625·96 1 778·10 9·05
4 13 032·15 17147·56 18 791·25 -1643·69 -8·74
5 13 600·00 17 894·73 15391·50 2 503·23 16·26
6 17 266·63 22719·25 21421-49 1 297·76 6·05
7 27011·30 35541·18 34 283·30 1 257-88 3·66
Mean 1-40
Standard deviation 10·94
 Modelling and Predicting Bridge Repair and Maintenance Costs 195

DISCUSSION

The average number of items in the bills of quantities analysed was 40. The
simple models developed for repair work therefore represent a reduction in
model complexity of some 60%. Their accuracy, however, is still ofthe order
of 10%. This is of the same order as the accuracy of quantity surveyors'
estimates ofthe cost of new construction, which is reported tobe about 13%
(Ref. 4).
If the models are to be used to predict the cost of future repairs, the
estimator will be obliged to insert a single unit rate against each item. The
extent to which this is possible, and the resulting change in accuracy, arestill
to be tested. Nevertheless, there is reason to believe that no great loss of
accuracy will ensue. Clearly the estimator will not be able to calculate the
weighted mean rate for, say, formwork of classes F1 and F3, since the
quantities of classes Fl and F3 formwork will not be differentiated.
However, preliminary tests using the arithmetic rather than the weighted
mean unit rates indicate no significant loss of accuracy. We anticipate that
the centrallimit theorem will work to our advantage, and that positive and
negative errors will cancel out.

ALTERNATIVE MODELS

lt would clearly be an advantage if the cost of bridge repairs could be

predicted before any detailed design was executed. Figure 2 shows the
relationship between costs per square metre of deck area and deck area for
the two categories of bridges studied. Best fit curves were derived from linear
regression analyses of the logarithms of the variables. The coefficient of
correlation for masonry bridgeswas -0·96 and for reinforced concrete -0·84.
The geometry of only 22 bridgeswas available to us, but although much
more data would be required before predictions could be made with
certainty, the curves do provide at least a starting point for the prediction of
the cost of repairs expected to exceed :Elük in value.
It is of interest to note that whilst the average age of the masonry bridges
at the time of repair was 63 years compared with 18 years for the reinforced
concrete bridges, the cost per square metre of repairing masonry bridges,
since we know that each bridge in our sample underwent major repairs only
once in its life, is no different from the cost for reinforced concrete bridges.
This is further confirmation that maintenance costs of masonry bridges are
less than for reinforced concrete bridges, a beliefwhich led Tayside Regional
Council to specify in 1987 the first masonry arch tobe built in Scotland for
more than 50 years.
196 Mohamed Bouabaz and R. Maleolm W Horner

Cost (E:/sq.m) _ _ _ _ _ _ _____,

1800 .,-----=._...:.......cc___

• Masonry
1600 + Reinforced concrete

1400

Masonry, r ·-0.96
1200

1000
Reinforced concrete, r =-0.84

800

600-
+
400

200

0
0 500 1000 1500 2000
Surface of deck (sq.m)

FIG. 2. Relationship between repair cost and area of deck for masonry and reinforced
concrete bridges.

CONCLUSIONS

1. It is imperative that the movement by local authorities to record repair

and maintenance costs on a consistent basis is sustained. Without good
quality historical data it is impossible to justify repair and maintenance
budgets or to make rational decisions about the cost effectiveness of new
bridge designs or about the optimumtime at which old bridges should be
replaced.
2. Bridges can be categorised in a way which reduces by 60% the number of
elements in the repair cost model and yet which still allows the cost of
repairs to be predicted to an accuracy of about 10%.
3. The tentative relationship between repair cost and surface area of bridge
deck proposed in the paper may provide the basis of an even simpler way
of predicting the costs of repairs expected to exceed flOk in value.
4. Our data lend further support ofthe view that the repair and maintenance
costs of masonry bridges are significantly less than those for reinforced
concrete bridges. There is therefore good reason to thoroughly reappraise
the relative merits of masonry and reinforced concrete for new bridges of
small spans.
5. Much work remains tobe done before the relationship between the age of
a bridge and the cost of maintenance and repair can be reliably
established.
 Modelling and Predicting Bridge Repair and Maintenance Costs 197

RE FE RENCES
1. HORNER, R. M. W. and AsiF, M., Economical designs using simple cost models.
Proc. 4th Jnt. Conf on Structural Faultsand Repair, London, 1989 (in press).
2. MuRRAY, M., HoRNER, R. M. W. and McLAUGHLIN, A., BRIDGET~a cost
estimating suite for highway structures. J. Jnst. Highwaysand Transportalion (in
press).
3. DuLAIMI, M. F., Towards simple contracting and estimating procedures. MSc
thesis, University of Dundee, 1986.
4. AsHWORTH, A. and SKITMORE, R. M., Accuracy in estimating. CIOB Occasional
Paper No. 27, 1983.
 17
Bridge Operation and Maintenance Costs

HANS INGVARSSON*

Swedish National Road Administration, S-78187 Borlänge, Sweden

ABSTRACT

Within theframework ofthe Swedish Commission an Maintenance and Costs

( DKU ), an autonomaus alliance between six city authorities, a comparison
study with regard to maintenance costs has been carried out by the bridge
committee of the commission.
The comparison study was concerned with operation and maintenance costs,
and their relation to the following parameters:
-amount of de-icing salts spread out.
-climatic and geographical conditions,
-age and size of bridge stock, and
-traffic intensity.
An explicit analyticalfunction wasfound which estimates the annual operation
and maintenance costs probably needed with respect to the parameters
mentioned above.

INTRODUCTION

Within the framework of the Swedish Commission on Maintenance and

Costs (DKU), an autonomaus alliance between six city authorities, a
comparison study 1 of bridge maintenance costs has been carried out by the
bridge committee of the commission. Members of this committee were

*Adjunct Professor at the Royal Institute ofTechnology, Department ofStructural

Engineering, S-1 00 44 Stockholm, Sweden.
199
200 Hans Ingvarsson

Messrs J. Gustavsson (Gothenburg), I. Karlsved (Västeräs), A. Malmberg

(Stockholm), C.-H. Silfwerbrand (Stockholm), W. Skottke (Solna) and G.
WegreH (Västeräs). In order to make the cost comparison study by the
committee more comprehensive, not only the street authorities of
Stockholm, Gothenburg, Västeräs and Solna participated in the commission
but also the Swedish National Road Administration was included. The
author of this paperwas therefore requested to join the committee, which
also acted as an advisory group to Ingvarsson and Westerberg 2 when this
state-of-the-art report on operation and maintenance of bridges was
prepared.
This paper comprises a brief summary of the committee report 1
mentioned above and its conclusions. This report, comprising cost statistics
from 1977 to 1984, was published in 1986. In this paper corresponding
statistics from 1985 to 1988 are also included on behalf of the committee.

RECORDED OPERATION AND MAINTENANCE COSTS

In the study of bridge operation and maintenance costs, 1 it was stated that
the following parameters were of interest:

-Total bridge deck area, D (m 2 ).

-Age of bridge stock on average, A (years).
-Average number of freeze-thaw cycles per year (F). In this case F
denotes the number of days during which the maximum and minimum
temperatures recorded are above and below the freezing point (Ooq
respectively.
-Amount of de-icing salt (sodium chloride) spread out per year, S (g/m 2 ).
- Traffic intensity ( T), described as total amount of vehicle-kilometres
per m 2 street area.
-Percentage (P) ofbridge deck area constructed before 1965. This year is
of special interest from the Swedish point of view as previously no air-
entraining agents were used and the typical bridge concrete had a
water-cement ratio of about 0·6. After 1965 this ratio was normally
about 0·5 and the concrete was air-entrained. Accordingly, it can be
noted that the durability of the concrete with regard to freezing and
thawing was significantly increased in 1965.

For each city, as weil as the Swedish National Road Administration

(SNRA), the parameters listed above are shown in Tables 1(A) and 1(B).
With these basic facts as a background the annual bridge operation and
maintenance costs recorded, from the year 1977 until 1988, are shown in
 Bridge Operation and Maintenance Costs 201

TABLE l(A)
Size of bridge stock and other basic facts

City Number of Total bridge Age on Number of

bridges deck area, average, freeze-thaw cycles
D (m 2 ) A (years) per year (F)
-~-~~--------

Solna 83 81000 18 37
Västeräs 177 57000 19 30
Gothenburg 580 260000 20 34
Stockholm 745 677000 27 37
SNRN 11600 2 780000 27 37b

aSNRA = Swedish National Road Administration.

bAverage value for the whole of Sweden.

TABLE 1(B)
Size of bridge stock and other basic facts

City Amount of sodium Trafik intensity, T Percentage of bridge

chloride spread out (vehicle-kilometres deck area constructed
per year, S (g/m 2 ) per m 2 street area) before 1965 (P%)

Solna 260 240 9

Västeräs 200 90 37
Gothenburg 765 170 22
Stockholm 600 230 40
SNRN 220 60 67

a SNRA = Swedish National Road Administration.

Fig. 1. In this figure all costs are adjusted in order to correspond to the 1988
price level.

DISCUSSION

From Fig. 1 it is rather difficult to draw any firm conclusions concerning the
maintenance costs as these vary considerably because of different
circumstances. In order to solve this problern it may be convenient to define
a bridge maintenance cost index as described below.
The recorded operation and maintenance costs are primarily due to
maintenance and repair of concrete bridge deck slabs suffering from
waterproofing systems no Ionger being watertight. Consequently, if the
202 Hans Ingvarsson

SEK/m 2

\
\
\
\
\

\
\

30 - _...... - ' \ ___.--.

20
10
-----~~::~~>-:: ______ /\, ----
., ______________ '
/_/''.. '-./--
/
0 +--~-~-..----+---r--r--~~----l---.---.---+-- Year
77 80 85 88
FIG. 1. Annual maintenance and operation cost per m2
bridgedeck area (SEK/m 2 ) for each
city between 1977 and 1988 (1988 price Ievel, I SEK ~ 0·15 US$).

concrete is not air-entrained, it will be damaged by freezing and thawing.

Bearing this typical deterioration mechanism in mind, the following bridge
deterioration index (B01 ) is proposed in (1):
B 01 = AFSTP/100 (1)

where A, F, S, Tand P are defined as described above. In Table 2 the resulting

bridge deterioration index for each city as well as that of the Swedish
National Road Administration (SNRA) are shown. In order to make it
possible to compare this index with the recorded operation and maintenance
costs the average costs (CA) valid for 1983-87 arealso shown (see Table 2 and
Fig. 1).
As found by DKU,l the expected annual maintenance cost (SEK/m 2
bridge deck area) can be calculated as follows:

BMcJ = 72·5Jlog B01 - 6-45 (2)

This cost refers to the 1984 price level. In order to correlate it to 1988 it
must be increased by 15%, obtaining the bridge maintenance cost index
(B~CI):

B~CJ = 83·4Jlog Bm - 6·45 (3)

As can be seen from Table 2, where this index is shown, the bridge
 Bridge Operation and Maintenance Costs 203

TABLE 2
Bridge deterioration and maintenance cost indices

City Bm Bt,.o CA CAIB~Cl

(X 10 6 ) (SEK/m 2 ) (SEK/m 2 )

Solna J7 28·7 12·5 0-44

Västeräs 3·8 30·0 42·7 1·42
Gothenburg 19-4 76·3 88·8 1·17
Stockholm 55·1 94·8 110·6 1·17
SNRA 8·8 58·6 65·7 1·12

Bm =Bridge deterioration index (eqn (1 )).

Bt,_c 1 =Bridge maintenance cost index (eqn (3)).
CA= Average recorded operation and maintenance cost 1983~87.

rnaintenance cost index (SEK/rn 2 bridgedeck area) roughly represents the

need for annual funding for operation and rnaintenance of bridges, if the
recorded costs are regarded as the proper Ievel of funding.
The results obtained by DKU 1 can probably be used by other local or
national authorities in the financial planning of operation and rnaintenance
ofbridges in other rnunicipalities. This can be done either by strict use ofthe
bridge rnaintenance cost index (B~CI) or by rnodification thereof ifthe typical
bridge deterioration is different frorn that in Sweden. If, for exarnple, the
rnajor problern being dealt with is reinforeerneut corrosion, the pararneter P
can be defined based on a critical age or service life with respect to this
phenornenon. Furtherrnore, the pararneter F is perhaps irrelevant in this
case.

CONCLUSION

Through the cost cornparison study 1 carried out by the bridge cornrnittee of
the Swedish Cornrnission on Maintenance and Costs (DKU) an explicit
analytical function was found which estirnates the annual operation and
rnaintenance costs probably needed with respect to the following
pararneters:
-total bridge deck area,
-age of bridge stock on average,
-average nurnber of freeze~thaw cycles per year,
-arnount of de-icing salt spread out per year, and
-traffic intensity.
The results of this study can thus probably be used by other local or
204 Hans Ingvarsson

national authorities in the financial planning of operation and maintenance

of bridges in other municipalities.

REFERENCES
1. DKU/7, Driftkostnadsutredningens konstbyggnadsgrupp, Specialrapport Maj
1986. Kommunförbundet, Stockholm (in Swedish).
2. INGVARSSON, H. and WESTERBERG, B., Operation and maintenance ofbridges and
other bearing structures. Publ. No. 42 from the Swedish Transport Research
Board, Stockholm, Sweden, 1985 (in English).
 18

Clifton Suspension Bridge: An Historie Monument

tha t Earns its Keep

0 AVID MITCHELL-BAKER
Howard Humphreys and Partners, Thorncroft Manor,
Darking Road, Leatherhead, Surrey KT22 8JB, UK
and
STUART CULLIMORE
Clifton Suspension Bridge Trust, 66 Queens Square,
Bristol BS14JB, UK

ABSTRACT

Completed in 1864, the Clifton Swpension Bridge now carries annually over
3·6 million vehicles at speeds not then envisaged. lt is as much an important
communications link aspart ofBristol's heritage. As the integrity ofthe design,
materials and construction haue become more fully understood and
appreciated, the Clifton Suspension Bridge Trust haue euolued a policy o.f; in
efj'ect, indefinite preseruation. The main historic events of darnage andrepair
are described, Iogether with uarious tests and analyses which haue encouraged
the trust to their present polic_v. The origins and organisation ofthe trust itself
are also of interest, offering a solution particularly suited to such a
responsibility as this bridge.

INTRODUCTION

The foundation stone to a bridge designed by Isambard Kingdom Brunel

was laid in 1837, almost a century after a bequest of flOOO by William Vick in
1753. lt remained unfinished at Brunel's death in 1859 and shortly after the
Institution of Civil Engineers led a commercial initiative to complete it to
remove 'what was considered a slur upon the engineering talent of the
205
206 David Mitchell-Baker and Stuart Cullimore

country'. They appointed Hawkshaw and Barlow to design and construct a

bridge on Brunel's abutments using ironwork, bought at a favourable price,
from Brunel's footbridge across the Thames at Hungerford. The Clifton
Bridge was finally opened in 1864 and was operated by a company financed
by tolls. After 80 years, accrued dividends were used to redeem the shares
and the bridgewas vestedunder the 1952 Clifton Suspension Bridge Act 1 in
the present Clifton Suspension Bridge Trust, comprising up to 12 trustees.
One trustee is appointed by each ofthe riparian local authorities-currently
Avon County, Bristol City and Woodspring District; the remainder have to
reside within 20 miles ofthe bridge. They bring to the direction ofits affairs a
wide range of responsible experience in management, commerce, engineer-
ing and public service.
Because of the bridge's historic importance, its basic integrity and a
growing interest in its conservation, the trust has developed a policy of
indefinite preservation. The objectives of this policy may be summarised as
follows.

Technical
Maintain the bridge to a high standard by
-regular and frequent inspection,
-careful protection against corrosion and rot,
-protection against overload, impact and fire,
-timely and careful repair and maintenance,
-analysis and testing using latest techniques, and
-application of experience from elsewhere.

CLIFTON SUSPENSION BRIDGE TRUST

comprising 12 Trustees
H
Clerk to the Trustees
Coopers & lybrand
I
I
operating through the
Technical, Investment
and Staff Committees. ~ Consulting Engineers to the Trust
Howard Humphreys & Partners
I '
I Bridge Master I
I
I

I
I
I

-----------------------, I

I I I
I

REVENUE COLLECTION MAINTENANCE &MINOR REPAIR MAJOR REPAIR

13 Tollkeepers Head of Maintenance Contractors
operating on 4 Regular Staff supplying
shi fts specific
l services &
I extra seasonal staff J goods

FIG. 1. Organisation of Trust.

 Clifton Suspension Bridge 207

Financial
Provide adequate resources to maintain and replace by
-efficient toll collection,
-effective financial and investment management, and
-insurance against external risks.
The organisation set up by the trust to achieve these objectives is shown in
Fig. 1.

EARLY HISTORY

The construction of the bridge, which skilfully incorporated the chains and
saddles from the original H ungerford bridge, is weil described by Barlow. 2 It
has been possible, without serious problems, to keep what is virtually the
original structure in a serviceable condition because of the inherent
soundness of the original design and construction procedures, the low
stresses in most members, and the high quality and nature of the wrought
1ron.

Attached to _ _ _ -
chain straps

4716

Rlght and
left hand
threads

Longitudinal
girder

FIG. 2. Suspcndcr rod connection to longitudinal girdcr.

208 David Mitchell-Baker and Stuart Cullimore

The only serious darnage ever reported was caused by extreme wind
conditions in 1877 and again 10 years later, when a total of five suspender
rods failed between the turnbuckle and eye connection to the longitudinal
girder (Fig. 2).
The transverse timher planks on the deck were replaced in 1884, in 1897
when the deck was first surfaced with mastic asphalt and again in 1948. Some
longitudinal baulks were replaced at various times and the whole deck was
replaced in 1958. Concern about the strength of the anchorages led to
reinforcement with additional chains in the Leighwoods anchorages in 1925
andin the Clifton anchorages in 1939. Corrosion ofthe chain links at ground
Ievel in the days of horse-drawn transport made necessary strengthening at
the land saddles in 1932.
These and subsequent major works are more fully described by Mitchell-
Baker and Cullimore. 3

THE POST-WAR PERIOD

Use during the Second World War by heavy vehicles, over the 6-t Iimit set
under the 1861 Act, exacerbated darnage to the already deteriorating deck
surfacing and timbers. Continuing restrictions on labour to maintain the
bridge and the shortage of replacement timher led Howard Humphreys &
Sons, who had been responsible since 1910 for engineering inspections of the
bridge, to place Iimits of 2! t per axle and 4 t maximum vehicle Ioad on traffic
but keeping the overall distributed Ioad of 28 t unchanged from the 1864
value. The formerwas based on the poor conditions ofthe deck timbers and
was vigorously resisted by the company. It was at this stage that the present
trust assumed responsibility for the bridge and ordered the major inspection
which was carried out in 1953 and supported by an extensive testing
programme, which has been described by Flint and Pugsley. 4 Concern over
the extent of corrosion of the lattice cross-girders resulted in the end lattice
girders, which were the worst affected because of inaccessibility of the back
face, being replaced by rolled steel joists. These lattice girders were then
tested to failure in the laboratory at Bristol University, with the result that
the earlier fears regarding loss of strength from corrosion were allayed. The
extent of corrosion on the remaining cross-girders, which with the timher
had been coated with pitch, was easily determined when they were grit
blasted in 1955 by Bristol Metal Spraying Company preparatory to zinc
metal spraying. The extent of corrosion was found to be less than expected
and not to be significant.
The trustees then considered alternatives to timher for the deck. These
included the concept of an alumini um deck which would have eliminated the
 Clifton Su.1pension Bridge 209

problern of providing a waterproof running surface on the timher deck and

so remove the problern of rot and generally reduce maintenance. The
resulting reduction in self-weight ofthe deck might have allowed an increase
in the permitted vehicle weight. This was, however, rejected in favour of an
improved timher deck, without increasing the permitted traffic loading. A
number of important benefits have resulted from this decision-
repairability by locally available building industry tradesmen has been
retained and the character and original design of the bridge has not been
compromised. lt is believed that the damping imparted by interface friction
in the heavy timher deck, by limiting vibration under dynamic loading
caused by wind and traffic, reduces fatigue darnage to the ironwork.
A small number of longitudinal baulks and cross-planks on the
Leighwoods end of the deck, thought to have been damaged by water
penetrating through the mastic asphalt, were replaced in 1988. In relaying
this section of the surfacing a polymer-modified mastic asphalt was used. It
was considered that this material would be more resistant to the continual
ftexing of the deck under traffic and assist in maintaining a waterproof seal.

s~.·

7Suspender
rod

wornhole
bulllupon
weldmelal

Maid steel Nyloc

nylon lockrng nut

Glacrer MB2525 DU
bush rn reamed hole

'

FIG. 3. Rcviscd bolt assembly on suspender rod.

210 David Mitchell-Baker and Stuart Cullimore

Longitudinal differential movement between the chains and the deck is

taken by bending ofthe Ionger suspender rods and by rotation at the upper
and lower bolts of the shorter ones. The resulting wear in the bolts, rod eyes
and suspender straps which led to replacement ofbolts at various times had
by 1970 become serious in the shorter rods. F ollowing extensive experiments
the revised arrangement, shown in Fig. 3, was adopted for fixing on the 33
shorter rods on each catenary spanning the centre of the bridge.
An important, and unexpected, environmental benefit has resulted from
the retention of the weight restriction. This Iimits traffic to private cars and
light commercial vehicles, compelling heavier through traffic to use the
alternative Cumberland Basin and other routes. The 'Clifton Village' has
therefore been largely spared the effect of such traffic, a significant factor in
maintaining its character.

OPERATION

Toll collection provides an effective means of traffic control, limiting the

total number of vehicles on the bridge and, with a weighbeam system
identifying overweight vehicles, is a crucial factor in preserving the bridge.
The automated toll collection system installed in 1975 was based on widely
used car park control barrier equipment, using electromechanical coin
rejector 'slug' mechanisms. These were very satisfactory as they could
operate in exposed conditions, so avoiding unacceptable canopies near the
listed structure. As electronic replaced electromechanical devices, severe
operating problems developed. New coinage with recessed faces, as opposed
to embossed detail, adhered to plastic chutes in the presence of even light
rain. Minor dimensional variations in coins, such as between earlier silver
and later cupronickel 10p pieces, led to high rejection rates and ingress of
small amounts of dirt and foreign matter caused unacceptably frequent
cleaning of chutes.
Festoon lighting has illuminated the bridge for many years. Originally
used only for festive occasions, it is now provided every evening as a
contribution to local civic amen~ties. The life ofthe tungstenrobust filament
lamps used in a horizontal position and subjected to vibration in these
exposed conditions is weil below the rated value. The labour costs for
replacing lamps, particularly in positions in the upper parts ofthe structure,
is high and forms a significant part of the overall maintenance labour costs.
In recent years there has been a considerable investment in new equipment
to improve safety and increase efficiency of maintenance operations. This
has included two demountable cradles to traverse the chains and these and
the cradle for underbridge inspection are now electrically driven. Steel cables
 Clifton Suspension Bridge 211

on each chain for use with a running harness now greatly improve safety for
those inspecting and painting the chains. Gantries and cradles for work on
faces of towers and abutments have been improved for safety of operatives
and the structure. Improvements in access arrangements to the anchorage
chambers are in progress.

TESTING AND ANALYSIS

The consulting engineers, in advising the trustees on major repairs and
maintenance, have been able to base their recommendations on extensive
programmes of testing and analysis commissioned by the Trust. The first of
these, reported by Flint and Pugsley, 4 comprising the load testing of the
bridgedeck tagether with tests to destruction of two cross-girders removed
from the bridge, followed the major survey of the bridge in 1953. It yielded
information about the structural action of the superstructure which
confirmed the interim weight limit and was useful in designing fixings for the
new timher deck.
A thorough investigation ofthe foundations ofthe masonry abutment on
the Leighwoods side was ordered in 1969 as a precautionary measure
because of reported 'slips' in other parts of the Avon Gorge and of the
geological conditions which were known to exist in the area. Such possible
risks as were found to exist were extremely small and considered to be
negligible.
An appraisal of the risk of fatigue failure, initiated in 1972, was later
extended to examine susceptibility to brittle fracture and, subsequently, to
the propagation of existing cracks by fatigue. The work, extending over
some 12 years, is described by Cullimore and Mason. 5 It included strain and
deflection measurements on the bridge under controlled loading, laboratory
testing of specimens and of full-size replicas of elements of the structure and
a fracture mechanics-based research programme supported by the Science
and Engineering Research Council.
The most fatigue-sensitive element ofthe structure was found tobe the eye
of the chain link joined to the tower saddle and regular inspection of this
area by fibre-optic endoscope was instituted. It was concluded that there was
a satisfactory margin of safety against fatigue crack initiation in the chain
eye and a negligible risk of the propagation of a pre-existing crack. In
reaching this conclusion account was taken of the compressive residual
stresses resulting from the initial proof loadings of each chain link to over
thrice its design load and, after construction, by 500 t of stone distributed
over the deck. Encouraged by these results, and bearing in mind the
increasing pressure of conservation interests, the Trust has adopted the
policy based, in effect, on achieving indefinite preservation described above.
212 David Mitchell-Baker and Stuart Cullimore

4.0
- ,....
Cii
z
0 .----
_,
::::i
-
! .----
(/)
.---- .----..---- ,--
"ii:iz
(/) '-- - ,.-
.----
3.0 -
0
II:
0 - f--
- -
...
111
0
i
111
> -
IL
0
ci
z
2.0
1976 77 78 79 80 81 82 83 84 85 86 87 88
YEAR
FIG. 4. Vehicle crossings 1976-88.

The traffic over the last 13 years, during which the toll equipment has
permitted accurate records to be kept, is shown in Fig. 4 and reflects the
growth in size and population ofthe local catchment areas and ofvehicle use
generally. The drop in 1981 coincides with the opening of the Avonmouth
Bridge on the M5 to the west of Clifton. Prior to this the trend was upward,
the peaks corresponding to periods when the A4 portway was closed for
remedial works necessitated by rockfalls in the Avon Gorge. There were 3·7
million crossings in 1988, up 7% on 1987 and up 12% on 1986.
These increases, and particularly the peak hour throughputs, now over
1200 vehicles in one lane, have causes the trustees to examine the adequacy
ofthe traffic control in these conditions. Consequently Howard Humphreys
have undertaken a large deftection analysis ofthe bridge using the LUSAS
finite element program. This method of modeHing the behaviour of the
structure gives a better assessment of the actions in its component elements
than was previously possible. This is confirmed by the comparisons (Table 1)
with Flint and Pugsley's 1953/54 measurements. lt is noted that the
linearised defiection theory calculations, which although not as close, still
give reasonable values. Having obtained a satisfactory calibration of the
model with isolated loads, the most severe lengths of distributed live load
have been identified and a quasi 3D analysis has been adapted to model the
more heavily loaded ofthe two chain/girder systems under eccentric loading.
All analyses are being carried out with total working loads and their effects
will be assessed against the strength, rotation and defiection capacities ofthe
relevant components of the structure.
 Clifton Suspension Bridge 213

TABLE 1
Comparison of measured and calculated deftections

Location of Deflection Deflection in main girders caused by

symmetrical measured at 1Gt-t vehicle (mm)
Ioad
Large Flint and Pugsley 1953/54
deflection
analysis Measured Calculated

±span ±span 81 84 78
±span :i span 55 55 62
mid-span mid-span 58 66 54

CONCLUSIONS

Because of its splendid setting and its historical associations with Brunel the
Clifton Suspension Bridge has become an object oflocal pride and national
interest. lt is a tourist attraction and a symbol for the promotion of Bristol
commerce. Its importance as a working element of Bristol's communication
system is confirmed by the !arge and steadily increasing usage. For these
reasons it is incumbent on the trustees to keep the structure fully functional
and, as far as possible, in its as-built condition. Therefore in repairing or
replacing a component its form is retained and the original material used
whenever possible. Structural quality wrought iron is no Ionger obtainable
and mild steel is used instead. Similarly, in repairing ajoint, bolts might have
to replace rivets but arc welding would not be used.
In the event of major structural damage, making the bridge unserviceable,
it is considered that conservation interests would strongly favour the
replacement ofthe bridge in its present form. Consequently the trustees have
insured the bridge against accidental darnage in a sum which would provide
for building a replica using modern materials.
The trustees are able to practise such policies because they have financial
autonomy and, unlike a public authority, they Iack competing demands on
their resources. Income is buoyant and capital reserves, although insufficient
to build a new bridge-a duty laid on the original trustees-are adequate to
provide a very high standard of maintenance and to carry out major repairs.
The prestige of this part of the local and national heritage attracts as
trustees persans who are able to bring to its affairs a high Ievel of relevant
expertise and, being locally resident, are inescapably accountable for the
engineering and environmental effects of their decisions. Over the years
therefore an active and responsive organisation has evolved to manage and
maintain this unique civil engineering structure, providing an interesting
214 David Mitchell-Baker and Stuart Cullimore

example of how an important, useful and historic structure, taken out of

direct public control, may be effectively maintained for public benefit andin
the public's interest.

ACKNOWLEDGEMENTS

The authors wish to thank the trustees of the Clifton Suspension Bridge for
allowing them access to the Trust's records and Howard Humphreys and
Partners for assistance in preparing the paper. The views expressed in the
paper are wholly those of the authors.

REFERENCES

1. Clifton Suspension Bridge Act 1952.

2. BARLOW, W. H., Description of Clifton Suspension Bridge. Mins Proc. Instn Civ.
Engrs, 26 (1867) 243-57.
3. MITCHELL-BAKER, D. and CuLLIMORE, M. S. G., Operation and maintenance of
the Clifton Suspension Bridge. Proc. Instn Civ. Engrs, Part I, 84 (Apri11988) 291-
308.
4. FLINT, A. R. and PUGSLEY, A. G., Some experiments on Clifton Suspension
Bridge. Correlation between calculated and observed stresses and displacements
in structures. Institution of Civil Engineers, London, 1955, Preliminary Vol.,
pp. 124-34.
5. CuLLIMORE, M. S. G. and MASON, P. J., Fatigue and fracture investigation carried
out on Clifton Suspension Bridge. Proc. Instn Civ. Engrs, Part I, 84 (April 1988)
309-29.
 19
A Systematic Approach to Future Maintenance

A. VAN DER TOORN and A. W. f. REIJ

Department for Structural Research, Ministerie van Verkeen en Waterstaat,
Rijkswaterstaat, PO Box 20.000, 3502LA Utrecht, The Netherlands

ABSTRACT

lf bridge management is the art of ensuring a good connection between two

opposite sides, the technical aspects are just a small (but essential) part.
One of these technical aspects is the prediction of the future behaviour of the
structure in the light of ageing mechanisms, and a second related aspect is how
to react to this in the form ofpreventive or corrective maintenance measures.
This paper deals with models which take into account the ageing of the
structure, the possible consequences offailure and the maintenance required to
return the bridge hack to an acceptahle condition.

INTRODUCTION

Up until 10 years ago the maintenance of civil engineering structures was a

low-profile requirement based on experience without any theoretical
background and executed by technician engineers working with small
financial budgets.
F orced by the fast-growing number of structures which have fundamental
and costly maintenance problems, nowadays there is a strong push to
upgrade our knowledge about maintenance or, even broader, to have a good
overall bridge management system in which all relevant factors are
presented and can be weighted to assist in taking the right decisions.
In Holland the Rijkswaterstaat (State Public Works) is responsible for the
maintenance of about 4000 structures that vary from simple viaducts to
huge infrastructural works like tunnels, bridges across the River Rhine and
storm surge barriers along the North Sea.
215
216 A. van der Toorn and A. W F Reij

All these structures tagether represent a replacement value of about 10

billion US$. Assuming a maintenance budget of about 1% of the original
cost, the future maintenance budget will be of the order of 100 million US$.
With the aid of a good bridge management system this amount of money
will be spent rationally, so that short-term economies do not give long-term
problems.

BRIDGE MANAGEMENT

In general, bridge management should contain the following basic topics:

1. Registration and description of the individual structures.
2. Initial determination of their actual condition.
3. Judgement of (some) structures in terms of safety, etc.
4. Prediction of their future behaviour resulting from ageing.
5. Definition of specific maintenance strategies.
6. Allocation of the (limited) budget to specific structures.
7. Execution of these strategies with adequate men and means.
8. Registration and evaluation of findings.
9. Back coupling with maintenance strategies and design.
10. Scenarios for the replacement of structures for economical, technical or
other reasons (increased traffic, new demands).
Although for the management of 4000 structures each ofthese ten topics is a
problern in itself, in this paper particular attentionwill be given to points 4
and 5.

MAINTENANCE MODELS

Maintenance models are intended to predict the time and scope of future
maintenance actions such as inspection, repair or replacement. In the first
instance these actions follow from an historically based maintenance
strategy, but later these models give the opportunity to come up with a more
or less optimised maintenance strategy.
Maintenance models need the following structure-related inputs:
1. A characterisation of the actual condition.
2. A prediction of future behaviour.
3. An estimation of maintenance-related costs such as inspection, repair,
replacement, loss of production and damage.
With the help of maintenance models the life cycle cost for a (historically)
given maintenance strategy can be determined.
 A Systematic Approach to Future Maintenance 217

Failure based Use based Condition based

FIG. 1. Qualitative decision tree for maintenance strategies.

The right choice of maintenance strategy (see Fig. 1) and, within that
strategy, the right adjustment of steering variables (inspection intervals,
action boundaries, duster size and time of replacement) make it possible to
optimise the strategy by minimising the cost.
Depending on the period of use, the cost can be expressed as cost in a fixed
period of time, per life cycle, per unit of time or as 'market value'.
Sometimes decisions about the maintenance strategy are not based on a
pure economical weighting but also on external demands such as a minimum
Serviceability Ievel, regulations, aesthetics, etc.
So far on this global Ievel all maintenance models are the same, but as
soon as one is predicted for specific categories of structure the models
diverge. A basic difference occurs between the civil technical and
electromechanical parts of the bridge structure.
Civil technical parts are long-term (50-100 years), usually complete
elements, and their lifetime can only be estimated if the underlying ageing
mechanism is known and described parametrically.
Electromechanical parts have a medium life span (10-50 years) and are
more or less standard components with a known rate offailure but without a
precise or measurable underlying ageing mechanism.
So the future behaviour of these two categories of construction are
organised differently in the maintenance models (see Fig. 2).
For both types of component (civil technical and electromechanical) there

t1---=..LLJ....W-~
-~
t ------------
: Tmean
1
tailure
-time
i~~-~.J..I..I..l.l..
t
-~. 1Tmin
:
1Tm.,an
:
-t1me

0
-t:iöuriCfry- <lJ
+'LrtJ
:
I
l /
1 /
~ I : /
·-o GJ I
c L I 7\
0 -~ Bathtub curve : /I
u &' 1
'C_ ...- I
I

-time
FIG. 2. Ageing of civil and electromechanical parts.
218 A. van der Toorn and A. W: F. Reij

are now computer models which give the life cycle cost for a single ageing
element which has only one renewal maintenance action (see Appendix).
Although sometimes the degeneration of a structure can be simplified to
that level (by considering only dominant mechanisms), most ofthe practical
situations are more complicated:

1. There are more different elements in one structure.

2. There is more than one ageing mechanism working.
3. The ageing mechanism is not directly affecting the function of the
element but has two stages.
4. Malfunctioning of an element does not always lead to malfunctioning
of the system as a whole because of redundancy.
5. There is more than one repair action possible, varying from a limited
local repair to total replacement of the element or structure.

When there are more ageing elements in a structure, the maintenance

strategy for the structure as a whole will normally differ from the sum ofthe
individual maintenance strategies. On the one hand, inspection andrepair
actions clus~ered together will give lower cost (less mobilisation cost, less
production loss, etc.), but alternatively individual elements are not always
optimally maintained.
In many civil technical structures the load-bearing construction is
protected against the aggressive environment by some form of protection
layer. For example, coatings on a steel structure, the concrete cover on the
reinforeerneut and the asphalt wearing layer on a bridge deck provide
protection. The necessity of maintenance in the first stage of ageing is not
because of the threat of production loss or the cost of darnage but for
prevention of extra maintenance cost in the next stage.
Although for a single element an optimal repair cycle can be found, the
way in which an individual element which is part of a system is maintained
depends on the redundancy, the condition of the accompanying elements
and the savings that can be achieved by joint action. To find the real
optimum is very difficult.

APPLICATION

To check the validity ofmaintenance models and assess their shortcomings,

an attempt was undertaken to apply the models on a 25-year-old steel bridge
crossing the River Rhine (see Fig. 3).
First of all, an analysis was undertaken considering the present
maintenance costs. After a considerable amount of work the cumulative
 A Systematic Approach to Future Maintenance 219

FIG. 3. Bridge used for maintenance analyses.

maintenance Ievel seemed tobe between 1% and 2% ofthe initial cost ofthe
structure (see Fig. 4).
A second analysiswas done of the different ageing mechanisms relative
to the amount of money spent to maintain the structure. In the top 12
maintenance categories, the repainting of the bridge was first, followed
by repair ofthe wooden deck adjacent to the movable part ofthe bridge. The
third highest category was replacement of the expansion joints.
If the use of maintenance models can reduce cost, it is clear that the most
benefit can be expected for the highest absolute contribution to the total
maintenance cost. It was therefore decided to first model the painting of the
bridge.

Ageing Mechanism
Considering a small area of the bridge structure, the ageing mechanism
consists of two stages. First, the degeneration of the paint layer occurs
without further consequences for the functioning of the bridge. Secondly,
the corrosion of the underlying steel structure results, which Ieads to a
decrease in structural safety.
Considering the total bridge structure, elements with different degra-
dation velocity can be distinguished. There are three main parts with their
own paint system:
1. The carriageway }
2. The arches Totalbridge area 80000m 2 •
3. The basculing bridge
Within the main parts degradation differences are caused by: sharp edges
and bolt nuts have minimallayer thickness; vertical web plates can easily

r--

,.......,.----...J'
I

---
- ;.;.·;, funding cost

1965 1970 1985 1990

FIG. 4. Development of the maintenance cost.

220 A. van der Toorn and A. W F. Reij

FIG. 5.
0 5 10
Time (years)
15

Elements with different degradation velocity.

20
-
dry; horizontal plates can accumulate (salt)water; welded joints have less
suture, etc. (see Fig. 5).

Actual Condition
Although there are some techniques for measuring the initial condition
(layer thickness, etc.), in practice the condition or darnage parameters in the
second stage are used, namely the percentage of corroded area and if needed
the loss of material in a cross-section.

Maintenance-related Cost
The cost related to the painting of the bridge consists of the cost of
scaffolding, inspection, (blast) cleaning if a certain degradation is detected,
cost of environmental mcasures, the paint itself, application, the processing
of the waste and the replacement of steel parts if considerable degradation is
detected.

Maintenance Model
Because of the number of different elements and the different but
correlated ageing mechanisms considered, the maintenance model used here
was based on Monte Carlo simulation.
Inspection of the condition (corrosion percentage and loss of steel) could
take place after a 'certain' period of time. Local painting was prescribed if
inspection of an element gives a 'certain' percentage of corrosion, and total
painting was prescribed if the group of elements which needed painting
exceeded a 'certain' Ievel.
Replacement of heavily corroded construction parts took place if a
'certain' safety Ievel was exceeded.
By varying the 'certain' values ofthe steering variables in the Monte Carlo
simulation a few optimal strategies were selected (see Fig. 6).
F or the first strategy the aim is to prevent the high cost of (blast) cleaning
and cost of other measures by means of frequent inspection and
considerable local intermittent painting.
 A Systematic Approach to Future Maintenance 221

iL
, .....
-------- ---
c::
I /
'
0

' ' ....... __ -

·-- I I
·-E \_,I
A A
+'
Ul 32 30
0
V

11
+'
0
1-
0 2 4 G B 10 12 14 16 18 20
At lnspection period

FIG. 6. Total maintcnance cost versus inspection period.

The second strategy is directed at prevention of the replacement of steel

parts by means oflow cycle inspections coupled with the total painting ofthe
bridge, including blast cleaning, etc.
Although the second strategy is somewhat better, there is only a small
difference between the schemes. The cumulative effect of steelloss during the
lifetime of 100 years is not critical.

CONCLUSIONS

1. ModeHing ofmaintenance procedures to optimise a maintenance strategy

for a given structure is possible but still needs to be extended when there
are more elements and ageing mechanisms.
2. Maintenance models force decisions to be made which were not
previously considered by inspectors.
3. Maintenance models require condition parameters as input data which
demand good measuring instruments and protocols.
4. Maintenance models demand the consideration of maintenance
alternatives and require consideration of their price and period of
effectiveness.
5. Maintenance models allow consideration of the consequences of bad
maintenance, or alternatively 'doing nothing'.

APPENDIX: MAINTENANCE MODELS

On the basis of a knowledge of the undisturbed future behaviour of an

element in a structure (input by means of ageing mechanisms or failure
222 A. van der Toorn and A. W F. Reij

GF tftl: _F~ilure ~oundari'

Re jection bound a r
GA - -- - - - -- - -

FIG. Al

rates), these models account for the probability of failure and rejection
within certain time intervals, depending on the maintenance strategy.

Failure-based maintenance

where

Use-based maintenance

where

and

Condition-based maintenance

where

in which
Cv = the cost of repair or replacement of the construction (part),
Ci = the cost of an inspection,
es= the cost of darnage (direct or indirect),
E(c) = the cost assessment parameter,
PrU) = the probability of failure in the ith interval,
Pa(i) = the probability of rejection in the ith interval,
tL = the expected value of the lifetime,
tmean = the lifetime in case of mean ageing,
t 0 = the previously determined moment of maintenance, and
n = the number of inspections during the lifetime.
 20

Management of Bridgeworks Maintenance in the

UK
N. 1. SMITH

Project Management Group, Department of Civil and Structural Engineering,

University of Manchester Institute of Science and Technology,
PO Box 88, Manchester M60 JQD, UK

ABSTRACT

This paper revielt·s current methods of managing the process of bridgeworks

maintenance in the UK at a time when there isanational movement in the
allocation of construction funds away from new works towards maintenance.
Bridgeworks maintenance has been the subject ofparticular attention in recent
years with the combined effects of increased deterioration in recent concrete
structures and the recent increases in the permissible loadings for highway
vehicles. These demands outstrip the available budget and hence many owners
of bridgeworks stock have adopted formal bridge management systems and
cooperated in research programmes.

INTRODUCTION

Bridges form a key part of the infrastructure of the United Kingdom (UK),
which facilitates the movement of people, goods and services vital to the
national economy. The total UK bridgeworks stock, probably one of the
mostdiverse bridge stocks in the world, is estimated to be about 150 000, of
which about 8900 are the responsibility of the Department of Transport
(DTp) 1 and 129000 the responsibility of the local authorities. In general
terms, the stock consists of about 70 000 masonry and brick arch and culvert
structures, mostly constructed prior to 1922, and about 60000 concrete and
25 000 metal bridges, largely constructed post-1922. Over 90% of the
bridgeworks stock consists of small bridges with a span of less than 10m,
although due to the motorway and trunk road networks the majority ofthe
major structures are within the DTp stock. The major bridges tend to have
223
224 N. J. Smith

more serious maintenance problems and in terms of percentage of initial

capital cost require higher maintenance expenditure. 2
Bridges have traditionally been designed to operate for long periods of
time without requiring major maintenance, indeed the current design life for
concrete and steel road bridges is 120 years. 3 However, bridges which were
constructed before 1922, and several which were constructed shortly
afterwards, were not designed to standards which would be acceptable
today. These older structures are now subjected to much greater loadings
than could have been envisaged at the time of their design. In addition, they
are subjected to the same problems as all other bridges, for example poor
workmanship, poor materials and accidental damage, which aceeierate the
need for maintenance work.
The design of bridges is a process of development and new structural
forms, new materials, new codes of practice and new methods of
maintenance all influence the final structure. Consequently the bridge stock
contains many structures which would not be constructed in a similar way
today, and which are disproportionately expensive to maintain using current
techniques. Many of these structures are unique and some are listed
buildings, hence the materials, working procedures and workforce skills
required to undertake maintenance are specialised and costly. Indeed the
cost of reconstruction has been shown to exceed the cost of initial
construction by a factor of between 1·5 and 8·0. 4
Further to these existing maintenance problems, the decision to permit by
1999 increased maximum vehicle loading in line with other European
countries 5 has caused new problems. In 1987 the DTp announced a
programme of strengthening works for the 38-t vehicle Iasting 15 years and
costing f2000m. lt is estimated that the increase from 38 t to 40 t would cost
an additional flOOm for DTp trunk roads and f600m for local authority
roads. Although the total vehicle weight has been the subject of political
attention, it is the axle weight that is important for bridge loading. Therefore
the existing 30·5-t four-axle vehicle provided a maximum axle loading of
10·5 t and more recently a partially loaded 38-t vehicle would provide an axle
Ioad of 11·5 t, which exceeds the axle loading for the 40-t vehicle and causes
the maximum Ioad case for short span structures.
Research work into the causes of deterioration and alternative methods of
physical repair has been undertaken by academia and the industry for many
years. In 1981 the OECD Report 6 divided bridgeworks maintenance into
the objectives of safety, Serviceability and economy. For public safety it is
estimated that in the UK the chances of a fatality due to a bridge failure are 1
in 10 000 000. Serviceability is defined as the optimal amenity for traffic with
minimal interference and the economy consideration is a measure of
financial decision of the trade-off between an existing structure with high
 Bridgeworks M aintenance in the UK 225

maintenance costs and the capital cost but predicted low maintenance costs
of a new structure.
The Maunsell Report published in 1989 7 was the most recent attempt to
try to determine the causes and the extent ofthe deterioration ofthe concrete
bridgeworks stock. This report investigated some 200 concrete bridges,
many of which were less than 25 years old, demonstrating considerable
deterioration due to chloride attack, alkali-silica reaction and carbonation.
However, the increasing pressures on maintenance budgets and the new
requirements have concentrated current research on the investigation ofthe
management of the maintenance of the bridge stock.

EXISTING SITUATION

A bridgeworks maintenance programme has to meet certain objectives, one

of the most important being the statutory requirements for public safety.
This has tobe achieved by making the most effective use oflimited resources,
by efficient planning and selection of priorities, and by using the most
appropriate contract strategy, and all within strict budgetary constraints.
Bridgeworks maintenance can be defined as ensuring the safe, unrestricted
passage of people, animals and vehicles as specified in the construction and
use regulations without limitations. 8 The work is a combination ofplanned
maintenance, refurbishment, replacement and the repair of accidental
damage. The first three items are regarded as traditional bridgeworks
maintenance and would usually be funded from the annual budget.
However, the last item is often regarded as a special case, sometimes
requiring immediate action, and therefore special funding arrangements
usually exist for these cases.
The expenditure on bridgeworks maintenance can be divided into costs
for structural elements, physical access elements and elements associated
with continued usage of the bridge during maintenance. The structural
elementsarenot always easy to quantify precisely as many abutments form
part oflarge retaining walls and substructure foundation works may not be
immediately adjacent to the structure. Access to the bridge is often a major
item of expenditure and sometimes inspections have been combined with
routine maintenance work. OECD indicate that the maintenance costs
should be of the order of 3% of the asset value whereas recent UK local
authority bridgeworks maintenance budgets have been about 0·3%,1
however these figures are too global to be of much value for assessing the
maintenance cost of a particular bridgeworks stock.
Maintenance work could be regarded as the action taken to prolong the
useful life of a bridge at a minimum cost with least interference to its
226 N. J. Smith

operational function. This approach contains contradictory aims as the

minimum cost of maintenance is often only achievable if the bridge is closed.
Frequently the cost of diversions of pedestrians and vehicles exceeds the cost
of the physical bridge repair. The expertise in the management of the
maintenance process is in deciding upon cost-effective minimum works
which permit the bridge to öe used without restrictions.
It appears that the rate of deterioration of bridges is not standard or
uniform. Although there are few published data to justify this statement,
there are many examples of arch bridges over 200 years old in good
condition whilst some bridges less than 25 years old require extensive
refurbishment or replacement. 9 lt is suspected that the relatively old and
relatively new bridges account for a higher proportion of the maintenance
budget than would be expected.
In the case of the older bridges, particularly pre-1922, this could be due to
poor materials, but the introduction of the DTp's new loading code means
that all70000 ofthe pre-1922 masonry and brick arch bridges and culverts,
and about 30% of the 15 000 pre-1922 metal bridges, require attention or
loading restrictions. 1 0 The more recent bridges, post-1960, were constructed
at a time when the specification had been revised and alternative bids were
allowed. This system, used in conjunction with tendering procedures which
concentrated on one parameter, the minimum capital cost, may have
affected durability adversely. Additionally, more sophisticated designs were
being produced with novel uses ofprestressed concrete, leading to minimum
weight and hence low priced bridge decks but with a tendency for high
maintenance costs if problems should arise.U The implication of these
factors isthat more attention should be given to the consideration ofthe life
cycle cost of the bridge. A bridge which is easily and cost-effectively
maintained is not likely tobe the bridge with the lowest capital cost at tender
stage. This raises the question as to whether or not the cost of maintenance
should be treated as capital or revenue expenditure.

MANAGEMENTSTRATEGY

Traditionally maintenance has been regarded as an 'after-the-fact' activity

and was something done in response to an unacceptable condition arising.
However, the life cycle costing approach includes the total management of
the maintenance process, consisting of maintenance planning, design
maintenance considerations, maintenance in construction and maintenance
in operation. 12 For new structures it is possible to consider the complete
cycle, but for the existing bridgeworks stock only maintenance in operation
can be implemented.
 Bridgeworks Maintenance in the UK 227

The fundamental requirement for the management ofthe maintenance in

the Operation stage is relevant base data. These data are largely obtained
from visual generaland principal inspections, and it is the frequency ofthese
inspections and hence the cost of collecting the data, and the reliability and
accessibility ofthe data, which inftuence the effectiveness ofthe management
process. There are a number of existing computer-based maintenance
databases operated by the DTp and several local authorities and
consultants, 13 however it should be noted that a database records the data
available at the time ofinspection or works and cannot be completely up to
date. There are also several computer-based maintenance management
systems for bridgeworks; 14 these mainly originated in the United States but
are being adopted in the EC countries in increasing numbers. Thesesystems
provide instant access to bridge data to monitor rates of deterioration, or
assess damage, or consider the maintenance spend profile on a structure over
time to assist in the decision-making process and also to facilitate feedback
to develop the database as work progresses.
On the basis of the bridge record data one of the key management
decisions is to determine the Ievel of intervention, which can be described as
the appropriate time for particular types of maintenance work: repair,
refurbishment or replacement. This aspect is the central problern for
management and illustrates the difficulties of decision making under
confticting constraints. The decisions are influenced by the quality of data
available, by the experience and expertise of the staff of the maintenance
section, and by the importance of the route using the bridge. The budgetwill
not allow all the works identified tobe treated at the sametime and hence the
expenditure has tobe made in such a way as to produce the 'outcome ofleast
regret'.
The importance of knowledgeable staff should be emphasised because in
order to save money on site and use the budget efficiently the expertise must
be available. Apparently similar problems may occur on two bridges. The
decision has tobe made as to which can be patched and which requires major
work. This decision should reflect not only which structure is a priority due
to structural safety considerations but also if a temporary repair is carried
out whether the structure is likely to deteriorate rapidly, incurring excessive
maintenance costs in future years. The cost of employing qualified engineers
is relatively small compared with the magnitude of the cost savings in
maintenance work which can be achieved through good pre-planning and
assessment of the works.
The interrelationship between the available budget and priority spending,
the age and type of the bridgeworks stock, statutory requirements and the
vehicular flows using the bridges is not well defined. Detailed planning is
difficult as, particularly on the older bridges, the full extent of the
228 N J. Smith

maintenance works required may not become known until the work has
started on site. Problems can also occur on some ofthe minor elements ofthe
structure, which whilst not serious for the structural safety ofthe bridge can
block or obstruct footpaths or highways, thus necessitating immediate
action. Thesetypes ofuncertainty mean that management has tobe flexible
in the scheduling of routine maintenance work and will need additional
funding from time to time.

IMPLICATIONS FOR NEW BRIDGES

Maintenance procedures have significant implications for the design of new

works, design being considered in its widest sense, incorporating feasibility,
technical assessment and detailing. There are two main areas of interest:
first, design which prolongs the life ofthe bridge and prevents deterioration,
and secondly, design features which make routine inspections and routine
maintenance, such as replacing bridge bearings, simple and low cost
operations. The main conceptual design is usually weil prepared and
considers the 120-year design life, however it is in poor detailing, usually
carried out by junior engineers, that the source of future maintenance
problems can be found. A particular problern in recent years has been
adequate drainage and waterproofing of decks, a problern which becomes
more serious in box beam bridges. 15 The cost of including features in a
bridge design to facilitate maintenance are small if these requirements are
identified early in the design process. Many of the larger bridges have
included access ways, lighting, water and other services in the original
designs. 16
Recent sophisticated structural designs, frequently using new materials,
have been successful in reducing the initial capital cost of the bridge but are
now seen to be absorbing a disproportionate amount of the repair and
maintenance budget, and causing problems in the operation ofthe facilities.
Two factors may be of significance: first, there has been a trend to separate
the bridge design from bridge maintenance, with different firms being
responsible for each phase. This has the disadvantages of a Iack of feedback
to the designers of the maintenance performance and of a Iack of
maintenance considerations during the construction on site phase.

CURRENT RESEARCH

The increasing problems of the management of bridgeworks maintenance

and strengthening programmes required for the 40-t vehicle are reflected in
 Bridgeworks Maintenance in the UK 229

the growing investment into research in this area. A major study is being
carried out by UMIST funded by the Repair, Operations and Maintenance
Programme of the Science and Engineering Research Council.
The research work at UMIST is concerned with the management of the
maintenance process for bridgeworks and gives particular attention to
information from a specific subdivision of this sector, the bridge
maintenance sections of Sheffield City Council and Manchester City
Council. The researchwill commence in 1989 and is due for completion in
1991.
The principal objectives of the work involve four main areas of interest,
which are described below:
(a) To investigate the existing management procedures, particularly the
organisation, planning and priority, contract strategy and legal
requirements associated with bridgeworks.
(b) To examine the criteria used in the decision-making process to
determine the Ievel of intervention, which can be described as the
appropriate time for particular types of maintenance work: repair,
refurbishment or replacement.
(c) To study maintenance budgets, the cost breakdown of maintenance
work into structural elements, physical access elements and elements
associated with continued usage of the bridge during maintenance.
The specific problern of mechanisms for the funding of emergency
maintenance work would also be included.
(d) To consider methods of assessing potential maintenance require-
ments in both the technical appraisal of a design of new works andin
the evaluation of contract tender bids for new works.

ACKNOWLEDGEMENTS

The author gratefully acknowledges the assistance ofthe bridges sections of

the City ofSheffield and the City of Manchester in the current research work.

REFERENCES

1. JoNES, C. J. F. P., Bridgeworks maintenance. Lecture to the Yorkshire

Association of the Concrete Society, 24 November 1988.
2. ANDREWS, 1. C. and GRINDALL, A. M., Maintenance and reliability congress
discussion: maintenance and renewal. Proc. lnstn Ch'. Engrs, 74 (1983) 562.
3. BRITISH STANDARDS INSTITUTION, Code of Practice for Design of Steel Bridges,
BS 5400, 1982.
4. MALLETT, G. P., Bridge Maintenance-Value for Money. National Workshop
230 N. J. Smith

on Bridge Maintenance Initiatives, Leamington Spa, Institution of Highways

and Transportation, 8 April 1986.
5. More bridge cash for 40-t lorries. New Civil Engineer (8 June 1989) 14.
6. OECD Report, Bridge Maintenance, 1981.
7. DEPARTMENT OF TRANSPORT, The Performance of Concrete in Bridges. Prepared
by G. Maunsell & Partners, HMSO, London, 1989.
8. LEADBEATER, A. D., The Practical Use of Bridge Assessments. National
Workshop on Bridge Maintenance Initiatives, Leamington Spa, Institution of
Highwaysand Transportation, 8 April 1986.
9. JoNES, H., New lease of life for Severn Bridge. New Civil Engineer (17 March
1988) 16.
10. New loading code to put squeeze on old bridges. New Civil Engineer (31 March
1983) 5.
11. JONES, C. J. F. P., Preventative Maintenance. National Workshop on Bridge
Maintenance Initiatives, Leamington Spa, Institution of Highways and
Transportation, 8 April 1986.
12. BLANCHARD, B. S., Total maintenance management. Tertoecchnia, 2(2) (1981)
139.
13. BRIDGET bridges data to cut out labour. New Civil Engineer (25 February
1988) 48.
14. Ho, C. K., Management of maintenance. MSc dissertation, University of
Sheffi.eld, 1988.
15. WELSH, N., Bridges affected by concrete problems. Construction News (3
October 1985) 12.
16. RUFFORD, N., Team assembles for Severn start. New Civil Engineer (16 April
1987) 30.
PROTECTION
 21

Crack Bridging by Surface Treatments to Concrete

1. G. KEER and B. H. LE PAGE

Department of Civil Engineering, University of Surrey,
Guildford, Surrey GU2 SXH, UK

ABSTRACT

Cracks offer an easy pathfor the ingress ofwater and chlorides into concrete.
In certain situations on bridge structures, coatings which are capable of
bridging over cracks and maintaining a continuous film are desirable. Although
the relationship between crack width and corrosion has been the subject of
debate, it is argued that a crack-bridging coating is preferable in critical
situations. Current tests in the UK which can be used to assess crack-bridging
ability are reviewed. West Germantests and specifications arealso summarised
and form a good basis for the development of better UK test methods and
spec(fications for crack bridging. The needfor further work on the effect of
cracks an hydrophohic surface treatments is highlighted.

INTRODUCTION

Surface treatments are applied to a concrete structure to improve its

appearance or to protect the structure from potentially aggressive agents. In
the latter case, past emphasis has perhaps been on the protection of concrete
against chemical attack when concrete is used in non-typical, rather special
situations. 1.2 In recent years, however, attention has focused on the
application of surface treatments to improve the performance of concrete
structures in typical environments which form common uses of concrete.
The principal objective has been to reduce the occurrence of reinforeerneut
corrosion and associated damage. In bridges, the depth of cover and
concrete quality is normally sufficient to eliminate problems due to
carbonation, although corrosion initiated by Carbonation has been
233
234 J. G. Keer and B. H. Le Page

observed. 3 The main corrosion problern in concrete bridge structures results

from the ingress of deicing salt solutions. The distinction in the source of the
problern is important, because some treatments may be effective only as anti-
carbonation coatings, otheFs only as anti-water/chloride treatments.
There are two principal types oftreatment. Coatingsjsealers rely upon the
formation of a pinhole-free film offinite thickness over the concrete surface,
which can act as a barrier to the diffusion into concrete of C0 2 andjor
solutions. Those materials described as sealers claim some penetration into
the pores ofthe concrete. Coatings would normally have a dry film thickness
of 150-500 J.lffi, although thicker 'high-build' coatings arealso used. Pore-
lining treatments are hydrophobic materials which line surface pores of the
concrete and repel moisture. The most widely used materials are commonly
called silanes, a term which embraces monomeric silane and oligomeric
siloxane formulations. Whilst among coating materials both anti-
carbonation and anti-waterjchloride treatments can be found, the pore-
lining materials are not effective as anti-carbonation coatings, only for
keeping waterjchloride out.
Preventing carbonation and waterjchloride ingress is dearly beneficial for
the durability of concrete bridge structures. Keeping concrete in a dryish
state is also beneficial in restricting ASR and freeze-thaw damage. Equally
clearly, cracks which cause breaks in coatings or which locally reduce water
repellency of the surface become paths for the ingress of moisture and

TABLE 1
Parameters affecting coating selection 4

Protection Durability

Diffusion of C0 2 , chloride, Proven use on concrete

0 2 , sulphate Case histories
Acid resistance Ability to span over passive cracks
Water vapour permeability Ability to seallive cracks
Water permeability Durability under strong UV, rain/wind
condensation, temperature and immersion
cycles, freeze/thaw, salt crystallisation

Application Cost

Tolerance to surface preparation, Special application needs

moisture on application and Relative costjbenefit at applied thickness
during eure Maintenance period
Resistant to alkaline conditions Ease of recoating
Ease of application, toxicity,
flammability
 Crack Bridging by Surface Treatments to Concrete 235

potentially destructive agents. A crack-bridging ability is therefore among

the many parameters influencing coating selection in Table 1 suggested by
Browne. 4 lt is, however, worth reviewing the role of cracks in corrosion of
reinforeerneut and concrete deterioration to establish the necessity of crack-
bridging ability. Crack bridging is a term generally applied to coatings when
the film spans across a crack, but the term is used here to include the ability
of hydrophobic treatments to repel water from penetrating down cracks.

THE NEED FOR CRACK BRIDGING

The role of cracks in the corrosion of reinforeerneut has been debated for
many years. Beeby has argued that the checks carried out on crack width
intended to control corrosion are based on 'no sound foundation of data
relating crack width to corrosion'. 5 While crack width may have an effect on
the time it takes for the depassivating agents (C0 2 or chlorides) to reach the
steel, once corrosion has started, the rate of corrosion is not very sensitive to
crack width. If the difference in time to initiation of corrosion between
cracked and uncracked concrete is small compared to the design life of the
concrete, crack width can have an insignificant role in corrosion damage.
The reason why the role of corrosion may not be sensitive to crack width is
that the corrosion process may be controlled by the cathodic reaction
requiring the diffusion of oxygen and water to steel in sound concrete and by
the electrical resistance of the path between cathode and anode. In this
respect treatments which are non-crack bridging may nonetheless be
beneficial in restricting oxygen and moisture penetration between cracks.
Another point in the argument about the influence of crack width isthat the
crack widths referred to are surface crack widths. Crack widths decrease
toward the bar surface and the width at the bar surface can be considerably
less than at the external concrete surface.
The argument above applies to cracks running perpendicular to the
reinforcing bar and is about the influence of crack width, not of the presence
of cracks. Thus crack-bridging coatings which negate the presence of cracks,
irrespective ofwidth, are likely tobe beneficial, particularly ifthe protection
is such that the initiation period is significantly extended beyond the design
life ofthe structure. Also, as Table 1 suggests, a continuous film will enhance
the durability ofthe film itself, preventing peeling at the edges, suchthat the
vital roJe ofthe coating between cracks is maintained. For cracksparallel to
reinforcing bars, both anodic and cathodic regions may be more easily
accessible and such cracks are potentially very dangerous. 6 lt may be true
again, however, that the danger is not very sensitive to crack width (e.g.
whether the crack is 0·1 or 0· 3 mm wide), merely to whether a crack is present
236 J. G. Keer and B. H. Le Page

or not. It is worth remembering that a loading crack perpendicular to one

reinforcing bar is likely to run along another. Also cracks resulting from
plastic shrinkage and Settlement may follow the lines of reinforeerneut If it is
then beneficial to use surface treatments which are capable of crack bridging,
the engineer must be able to relate his requirements for crack bridging in
practice to the results of laboratory tests which effectively assess crack-
bridging ability.

TEST METHODS FOR CRACK BRIDGING OF COATINGS

The crack-bridging ability of a coating is principally related to the

extensional characteristics ofthe coating material, the film thickness and the
adhesion to the substrate, all ofwhich may change with ageing. While these
parameters can be measured independently (for example the extensibility of
a free film of coating can be assessed using ASTM D2370 7 ), their complex
interaction means that tests modeHing films bridging cracks in mortar/
concrete substrates are essential. Tests can be broadly divided into three
types: static tests which measure the crack width at which coatings 'fail';
dynamic tests which subject the coating to cycles of measured crack
movement; compliance pass/fail tests where the crack widths at failure are
not explicitly measured.

UK Tests
In a test developed at BRE, 8 a coating is laid on a mortar block with a
weak section at its centre (Fig. 1). The block is slowly extended until a crack
forms, and is controlled so that the initial crack width does not exceed

FIG. 1. BRE crack-bridging test. 8

 Crack Bridging by Surface Treatments to Concrete 237

0·02 mm. The extension is then continued until the film cracks or small splits
develop. The crack width at failure is recorded. This approach has been used
by the British Board of Agrement to assess masonry paints for extensibility,
although MOAT No. 33 9 refers only to specimens of the coating on
aluminium strips which are extended in a tensile testing machine such as an
Instron. The coating is examined microscopically for signs of failure. By
either method extensibility is determined on unaged and artificially
weathered specimens. A styrene acrylate copolymer textured coating breaks
at a crack width of 1·0 mm when tested in the unweathered condition in the
BRE test. The crack width at coating failure is reduced to 0·5 mm after a
period of artificial weathering.
In MOAT No. 24/ 0 for plastic renderings drawn up by a body including
the British Board of Agrement, the test piece substrate comprises three fibre
cement segments carefully aligned and clamped so that the substrate has two
'cracks' less than 0·1 mm wide. After application and curing of the coating,
the clamp is removed and the cracks are widened by the introduction of a
wedge between the segments. The crack width at failure of the coating is
recorded.
A Iabaratory test was developed by TRRL to assess the resistance of
waterproofing membranes for concrete bridge decks to substrate cracking. 11
The membrane is laid on a small reinforced slab, which is then loaded to
produce cracking (Fig. 2). Loading is increased to widen the crack width,
with pauses of 30 min at crack widths of 0·25 and 0·60 mm to allow for
delayed fracture and at other stages where there were indications that
fracture of the material was commencing or developing. Three test
temperatures, 20, 3 and -10°C, were used. A satisfactory performance for a
membrane was withstanding a crack width of 0·60 mm. The basis of this
requirement was that the membrane should withstand fracture at the
permitted crack width of 0·25 mm factared by 1·2 to allow for widening
under repetitive Ioad and 2·0 to allow for variability in membrane
performance and for some crack widths in excess of 0·25 mm. (The
maximum observed width of crack has been reported tobe approximately
double the average crack width. 12 ) There are proposals to update this test to
include a dynamic effect.

-4-. J ~~ Sp~,;~,
coating crack reinforcement

15~l2-.
·~ _______
crack ~
~-- . 23Umm w1de
I
j.
inducer 70iJmm ~

FrG. 2. TRRL crack-bridging test. 11

238 J. G. Keer and B. H. Le Page

West German Tests

The protection of concrete by surface treatments is popular in West
Germany. In the very harsh climate of Berlin, bridges are given a protective
covering including a polymer deck membrane, silane or siloxane on soffi.ts
and an elastic crack-bridging coating on abutments and columns which can
be subject tosalt spray. Crack movements of0·4 mm over a 12-h period have
been recorded. 13 The Federal German Institute for Materials Testing
(BAM) have developed tests to assess the crack-bridging ability of coatings

coating
po 1ymer s1eeve
crack i nducer

40 x 40 x 16J mortar prism

+ (a)

(b)

FIG. 3. Crack-bridging test to West German specifications. 14 (a) Specimen type and (b)
specimen under static test crack width ~ 0·5 mm.
 Crack Bridging hy Surface Treatments to Concrete 239

over static and dynamic cracks in concrete. The test procedures have largely
been incorporated into regulations introduced by the Federal German
Transport Ministry for bridge surfacings on concrete and for the protection
and maintenance of concrete structures generally. 14
The specimen used is a small concrete prism reinforced centrally in which
a crack is induced (Fig. 3(a)). One face of the prism is coated. The prism is
cracked by a tensile Ioad applied through the ends of the steel bar. Figure
3(b) shows a specimen during a static test at the University ofSurrey. Cyclic
loading/unloading in tension results in crack opening/closing. The test is
carried out at - 20' C and coatings may be subjected to accelerated
weathering prior to test. Another specimen design used in West Germany
allows greater widths of coating tobe examined. 15 The coating is applied to
a layer of mortar, itself laidontop of two abutting steel plates. The plates
are moved apart gradually, inducing a crack in the mortar over which the
coating bridges. Static and dynamic tests can be carried out.
Draft German proposals classify coatings according to their performance
under a number of test regimes which relate to crack movement caused by
temperature cycles or temperature and Ioad cycles. The lowest dass of
coating, to cover a crack subject to thermal cycling only, would have to
withstand 1000 cycles at 0·03 Hz between crack widths ofO·l and 0·15mm,
i.e. 0·05 mm movement. The highest class is required to resist 10 5 cycles at
high frequency (5Hz) to simulate traffic effects superimposed on the low
frequency thermal movement, with a maximum crack movement of0-3 mm.
Also, under static conditions, a crack opening of 1·0 mm must be
accommodated without film failure.
It is suggested that these approaches should be the basis of improved test
procedures and specifications in the UK. The specimens in Refs 14 or 15
have advantages over !arger reinforced slab specimens in ease of
manufacture, handling and testing, particularly when tests are conducted at
low temperatures and when specimens are to be artificially aged prior to
test.

CRACKSAND HYDROPHOBIC TREATMENTS

There is little published Iiterature on the effects of cracks on the performance

of hydrophobic surface treatments in keeping maisture from penetrating
below surface layers. The Department of Transport Advice Note
BA 23/86 16 suggests that cracks up to 0·3 mm wide may be treated by
impregnation of the concrete surface with a silane treatment to prevent
chloride ingress. The basis of the 0·3 mm width is not clear, but it may be
some US research 17 in which a silane and a methyl methacrylate treatment
240 J. G. Keer and B. H. Le Page

provided added protection for embedded bars in cracked reinforced

concrete slabs. It was suggested that these two materials penetrate existing
cracks and could provide added protection for bridge surfaces containing
cracks of up to 0·25 mm. There is likely to be a significant difference in
performance between horizontal and vertical treated surfaces, since the
possibility of ponding in the former is likely to be infiuential.
A simple technique has been used at the University of Surrey based on
specimens offibre-reinforced mortar in which fine cracks ( <0·1 mm) can be
induced. 18 Untreated specimens laid on a moist surface draw water up
through fine cracks very quickly so that the cracks are highlighted on the top
'dry' surface. A silane treatment applied pre-cracking is not effective when
crack width exceeds about 0·05 mm. When applied post-cracking, the
treatment is much more effective and can prevent maisture penetration
through the crack up to about 0·1 mm and possibly beyond this, although
an upper bound has yet to be established. Further work is proceeding at
Surrey to develop tests to relate crack width and maisture penetration for
hydrophobic treatment.

CONCLUDING REMARKS

1. Surface treatments for concrete are likely to play an increasing role in

maintaining concrete bridge structures.
2. In certain situations it will be desirable to use coating films which are
capable ofbridging over cracks or, in the case ofhydrophobic treatments,
preventing by surface effects water or chloride ingress via cracks.
3. Tests for coatings should include cycles of crack opening/closing at low
and high temperatures representative of site conditions, in addition to the
assessment of the maximum crack width at coating failure, following
cyclic movement. Performance after accelerated ageing should be
assessed.
4. For hydrophobic surface treatments, further work is needed to establish
the effect of crack width on the performance of an otherwise water-
repellent surface.
5. Results of laboratory tests to assess crack-bridging ability should be
correlated with field performance of treatments.

ACKNOWLEDGEMENT

The authors would like to acknowledge the financial support of the Science
and Engineering Research Council.
 Crack Bridging hy Surface Treatments to Concrete 241

REFERENCES

1. ACI, Committee 515. J. Am. Concr. Inst., 63(12) (1966) 1305-91.

2. AMERICAN CoNCRETE INSTITUTE, Manual ofConcrete Practice, Part 5, ACI 515
IR-79, 515 IR-1, 515 IR-41, 1985.
3. VASSIE, P., Reinforcement corrosion and the durability of concrete bridges. Proc.
Jnstn Civ. Engrs, Part 1, 76 (1984) 713-23.
4. BROWNE, R., Building deteriology~the study and prediction ofbuilding life and
performance. Chem. Ind. (December 1986) 837-44.
5. BEEBY, A. W., Cracking and Corrosion. Concrete in the Oceans Tech. Report
No.1, Cement and Concrete Association, 1978, 77pp.
6. DARWIN, D., Debate: crack width, cover and corrosion. Concrete International
(May 1985) 20-35.
7. AMERICAN SociETY FüR TESTING AND MATERIALS, ASTM 02370-82, Standard
Test Method for Tensile Properties of Organic Coatings, 1982.
8. WHITELEY, P. and RoTHWELL, G. W., Appearancc and performance factors in
coatings for buildings. J. Oil Colour Chem. Assoc., 54 (1971) 855-78.
9. MOAT No. 33: 1986. The assessment of masonry coatings. British Board of
Agrement, 1986.
10. MOAT No. 24: 1983. Directives for the assessment of plastic renderings.
European Union of Agrement, 1983.
11. McDONALD, M. D., Waterproofing concrete bridge decks: maintenance and
methods. TRRL Labaratory Report 636, Transport and Road Research
Laboratory, Crowthorne, Berks, 1974.
12. ILLSTON, J. M. and STEVENS, R. F., Long-term cracking in reinforced concrete
beams. BRE Current Paper CP14/73, Building Research Establishment, May
1973.
13. MIDDELBOE, S., Berlin bridges take cover. New Civil Engineer, No. 639 (9 May
1985) 14 15.
14. HEROLD, C., Bridging of cracks in concrete using liquid plastic coatings.
Kunststoffe German Plastics, 78(7) (1988) 30-2 (English translation of paper in
Kunststofll>, 78(7) (I n7) 631 4).
15. ENGELFRIED, R., Requirements profile and suitability examinations for
materials for concrete repair. Proc. Int. Colloq. Materials Science and
Restoration, Esslingen, September 1986, pp. 175-82.
16. Departmental Advice Note BA 23/86. The investigation andrepair of concrete
highway structures. Department of Transport, 1986.
17. National Cooperative Highway Research Program Report 244. Concrete
Sealcrs for Protection of Bridge Structures. Transportation Research Board,
Washington, 1981.
18. KEER. J. G., Xu GuoDONG and FILIP, R., Cracking and maisture penetration in
fibre cement shceting. Proc. Int. Conf on Recent Developments in Fihre
Reinfärced Cements and Concretes, Cardiff, September 1989. Elsevier Applied
Sciencc Publishcrs, London (in press).
 22
Keeping Water Out of Concrete-The Key to
Durability

M. B. LEEMING
Arup Research and Devefopment, 13 Fitzroy Street, London WJP6BQ, UK

ABSTRACT

Alldeterioration mechanisms to reinforced and prestressed concrete bridges

and other structures are influenced or promoted by ~vater. Chloride ingress,
corrosion ofthe reinforcement,freeze/thaw, afkali-silica reaction and sufphate
attack can he controffed by restricting maisture movement. Carbonation being
dependent on the d(f/ilsion oj'a gas into the concrete is the oddity yet even this
onfy causes corrosion of the reinforcement at certain moi.Hure states of the
concrete. The rate oj'carho11ation is also moisture dependent. Moi.~ture has a
similar e_f/('Ct 011 other deterioration mechanisms and a proper understanding of
these sometimes confiicting influences and the way concrete takes up water can
Iead to a proper strategyf(Jr dealing with deterioration. The whole philosophy
o{ deali11g 1rith water i11gress will he covered from desig11 details, material
specificatiol1, constructio11 se11sitivity. non-destructive testi11gjor diagnosis and
rej'urhishment. Durability has a greater influence than strength 011 the design of
concrete mixes and 011 constructionfor most severe exposures. There are many
methods and materia/.1· which can he used to waterproof concrete but few of
these arej(Ji!saf'e a11d mostfäil atjoi11ts.junctions a11d edges. The economics o{
various approaches, hoth preventatil'e and remedial, will he discussed.

INTRODUCTION

Water is at the very heart of civil engineering; houses and factories provide a
dry environment, bridges and tunnels are needed to cross rivers and canals,
dams provide drinking water, pipes convey water and sewage, the sea
req uires harbours, docks and coastal defences, while roads provide all-
weather surfaces. When considering the deterioration of concrete, water is
243
244 M. B. Leeming

found to be a major, if not the main, cause of the problem. This statement
will be studied in detail in the remainder of the paper. Perhaps with some
understanding of the mechanisms involved there will be a better chance of
controlling deterioration.

THE DETERIORATION OF MATURE CONCRETE

The deterioration of concrete can be divided into two categories, the

chemical or physical deterioration of the concrete itself and corrosion of the
reinforcement. These two forms of deterioration have essentially different
mechanisms yet moisture still remains a common factor.

Chemicai/Physical Deterioration of the Concrete ltself

These forms of attack can be due either to internal or to external sources.

Sulphate attack
An external form of deterioration when sulphates 1 leached out of the
ground or waters high in sulphates react with the cement matrix to weaken it
and cause disintegration. Often dealt with by the use of sulphate-resisting
cement low in C 3 A or alternatively by dense high-strength concretes with
high cement contents and/or the use ofpfa or ggbs as a cement replacement
which additionally helps to lower the C 3 A content. The action ofwater is to
provide the supply of sulphate.

Freeze/thaw darnage
An external environmental form of attack by freezing of water within the
pores of the concrete. 2 •3 Normally air entrainment is used to combat the
problem. However, hydraulically pressed concrete such as paving slabs are
generallyimmune to freezejthaw darnage due to their inherent density and a
coarse granular structure. Dense well compacted concretes which have a
degree of Saturation below a critical Ievel usually survive freeze/thaw
conditions weiL

Alkali-silica reaction
An internal reaction as a result of mixing certain reactive aggregates with
cements ofhigh alkali metal content. 4 This phenomenon has been known for
a number of years but research is still going on as to the exact mechanisms
involved and the structural consequences ofthe resultant deterioration. The
concrete swells and cracks in a characteristic pattern when unreinforced.
The presence of moisture is essential to the disruptive swelling.
 Keeping Wafer Out of Concrete 245

High alumina cement conversion

An internal reaction within high alumina cement concretes. 5 This cement
has not been used in structural elements for many years and this form of
deterioration is now rarely experienced. However, the conversion of the
cement causing considerable loss of strength was most common in areas of
high humidity.

Saft crystallisation
The surface of the concrete just above ground or water Ievel can
deteriorate due to the capillary rise of salt-bearing water which evaparates
at a higher Ievel leaving the salts behind which crystallise and disrupt the
concrete.

Saft scaling
A similar phenomenon to that above but found in horizontal slabs and
caused by the application of deicing salts. The dissolved salts increase the
severity of freeze/thaw action. The salt finds its way into the concrete in
solution.
Popovics 6 lists six classifications of deteriorating mechanisms in concrete
andin all but one mechanism, mechanical deterioration, water is cited as an
influence.

The Corrosion of Reinforeerneut

Moisture can have a complex influence on corrosion, as shown in Fig. 1. 7
Concrete provides the ideal alkaline environment for reinforcement where a
Oxygen _
diffusi on -
CarbonatiOn----
-......

Corrosion

40 60 80 100
Relatrve humrdlty (•!.)

FIG. l. The influence of moisture on corrosion of reinforcement.

246 M. B. Leeming

stable film of corrosion products forms on the surface of the steel, strongly
adhering, which stiftes further corrosion (passivity). This situation can
remain indefinitely and requires a certain amount ofmoisture and oxygen to
sustain the situation. However, two forms of attack can upset the balance.

Chloride ingress
Chlorideions at the surface ofthe steel which breaksdown the passive film
causing corrosion in pits. The chlorides in concrete can result from one or
both of two causes. First, chlorides can be cast into the concrete due to salt
contamination ofthe aggregates or be added as an accelerator, a practice no
Ionger permitted. Secondly, chlorides can diffuse into the concrete in
solution either from seawater in marine situations or from traffic spray
containing deicing salts.

Carbonation
Carbon dioxide gas from the airdiffuses into the concrete and reacts with
the cement in the presence ofwater to make the concrete less alkaline. When
the pH ofthe concrete falls below about 9-10, the passivity breaksdown and
general corrosion can occur.
However, there are two further ways in which moisture influences
corroswn.

Resistivity
Corrosion is an electrochemical process and requires an electrolyte
between the anodes and cathodes on different parts of the reinforeerneut
This electrolyte is the pore water in the concrete which contains various
dissolved salts. A large amount of pore water provides a low resistance to
corrosion currents, hence corrosion can occur apace. Dry concrete, on the
other band, is highly resistive and corrosion is limited.

Gas diffusion
Saturated concrete severely restricts the diffusion of gases, in particular
carbon dioxide and oxygen. Wet concrete restricts the rate of carbonation
(loss of alkalinity) of the concrete, delaying the onset of corrosion. Oxygen is
required at the cathode to allow the electrochemical reaction to proceed. In
saturated concrete Iack of oxygen controls the rate of corrosion in spite of a
low resistance in the electrolyte.
Hence, as can be seen from Fig. 1, little corrosion occurs below about
50% humidity because the concrete is too dry to provide an adequate
electrolyte to the corrosion cell in spite of high Carbonation rates and an
ample supply of oxygen to the cathode. When the concrete is saturated as
stated above, the Iack of oxygen Iimits corrosion to very low rates. At
 Keeping Water Out of Concrete 247

relative humidities between these Iimits rapid corrosion can occur.

Alternative wetting and drying can aggravate the situation with rapid
carbonation and oxygen ingress during dry periods followed by corrosion at
high relative humidities.

THE INFLUENCE OF WATER ON FRESH CONCRETE

Water is a vital constituent of concrete. Furthermore, concreting is a 'wet

trade'; control of water content is required at early ages to influence both
strength development and durability. Most designers seem to understand
that water chemically reacts with the cement (hydration) yet many
practitioners seem to think that it is a drying process and as a result the
concrete does not get properly cured.
Strengthis usually the prime consideration when specifying concrete. The
grade number usually relates to the required 28-day strength and the main
compliance requirement is that cubes made of the concrete reach the
specified compressive strength. Modern codes, 8 however, now recognise that
durability is an equally important criterion and give guidance on the choice
of the correct cement content, waterjcement ratio and cover to the
reinforcement for the various exposures. There are no generally recognised
tests for fresh concrete that indicate whether the required durability criteria
have been met other than the mix has been correctly proportioned. As a
result the cube test remains the solepass or fail criterion yet a strong concrete
is not necessarily a durable concrete and vice versa. Hence it often happens
that when the cubes exceed the required strength the contractor seeks to
reduce the cement content or to increase the water/cement ratio, paying
scant attention to the durability requirements. This point will not be got
across to the contractor until the drawings state not only the strength
requirements but also the durability requirements.
The main factors at casting which inftuence the durability of concrete are
as follows.

The Water/Cement Ratio

Literatureis full of references to the fact that the lower the water/cement
ratio the moredurable the concrete. More water is required in the mix to
allow proper mixing and placing than is required for hydration of the
cement. The excess water eventually dries out of the concrete and leaves a
coarse structure behind.

Workability /Compaction
The amount of water m the m1x determines the ability to place the
248 M. B. Leeming

concrete (workability) around the reinforcement in a dense homogeneous

mass. lf the concrete is too dry voids and honeycombing of the concrete
results. lf it is too wet it tends to segregate and is more prone to plastic
cracking.

Curing/Maturity
The concrete needs tobe kept moist for sufficient time to allow the cement
to hydrate properly. Curing 9 - 12 has greatest effect on the durability of
concrete as the surface layers are most affected. The inner parts of concrete
are virtually self-curing. Maturity is part ofthe same mechanism although it
is normally affected by age and strength development, the striking time of
formwork being the main criterion for fast construction. Monitoring the
heat of hydration is used in the maturity meter 13 as a diagnostic tool. The
'Capo' test and other pull-out tests monitor the strength of the concrete at
early ages. lt is also possible to monitor the resistivity of the concrete where
significant changes can be seen as the hydration of the cement occurs.

MECHANISMS OF INGRESS OF WATER INTO OR THROUGH

CONCRETE

The movement of water into a porous medium such as concrete is highly

complex and is weil described in detail in Ref. 14 and will only be briefly
summarised here before concentrating on what the author believes tobe the
most important factor as far as durability of concrete is concerned.
Fermeability is a term which is often used loosely to cover a number of
mechanisms by which water passes through concrete, but has been defined 14
asthat property of a porous medium which characterises the ease with which
a fluid will pass through it under the action of a pressure differential.
Permeability is quantified in Darcy's law.
This mechanism assumes the concrete to be saturated for the section
under consideration with a head of waterat least on one faceund er a steady
state. The tests which measure this property are usually done with the
specimen submerged. The property is of importance when estimating the
amount ofwater passing through concrete in structures such as dams, water-
retaining structures and offshore oil platforms.
D(ffusion is the process by which a liquid, gas or ion can pass through
concrete due to a concentration gradient. Fick's first law describes the
process. Because a concentration gradient is required the process is
important with respect to concrete for solutions of different strengths,
vapour diffusion under differing relative humidities, differing ion concentra-
tions in saturated concrete and the diffusion of carbon dioxide and oxygen.
 Keeping Water Out of Concrete 249

Adsorption occurs when molecules of water adhere to the surface of the

concrete, held there by Van der Waals forces. lt is a transitory process and
not of much interest in the mass transfer of significant amounts of water in
the deteriorating process.
Capillaryflow or ahsorption occurs in narrow channels when a meniscus
forms and capillary forces are set up drawing the liquid through the channel.
The rate of flow is given by Washburn's equation.
Osmotic forces can be set up across thin semi-permeable membranes
subject to concentration gradients. These forces can be quite high but are of
importance only as far as water transport into concrete is concerned for
waterproof membranes and surface coatings.
Of the above mechanisms, the most important with regard to the
deteriorating mechanisms of concrete is believed to be capillary flow. The
pore structure of concrete is highly complex and consists of many voids
interconnected by micropores and microcracks up to cracks visible to the
naked eye. Many of these pores/cracks can be of capillary size and will
therefore suck free water on the surface of the concrete into the body of the
concrete. However, the variable size of these pores/cracks does not !end itself
to simple estimation offorces and flow rates involved. Capillary pressures of
up to 2 bar (20m depth of water) have been observed due to capillary
suction. 15
Diffusion and permeability are of importance in fully submerged concrete
and for concrete which retains or resists the ingress of water. However,
research has shown that under fully submerged conditions corrosion is
severely limited due to Iack of oxygen diffusion even with high chloride
Ievels. Diffusion and permeability are of more importance for sulphate
attack and for alkali aggregate reactions where the reaction can continue in
saturated conditions.
However, the deteriorating mechanisms of concrete under fully
submerged conditions are relatively simple as they are steady state and are
subject to analysis to obtain a reasonable estimate of the problem. Most
deterioration problems occur in the aerial zone of the structure where the
darnage is more readily seen. lt is weil known that offshore and marine
structures are most vulnerable in the splash or tidal zones. Onshore
structures have a similar zone which is much less well defined. The problern
is no Ionger steady state and the worst affected zone is subject to wetting and
drying on an entirely random basis. lt is believed that the important
mechanisms are absorption of surface water due to capillary suction and
subsequent drying by vapour diffusion. Water take-up is also possible by
vapour diffusion at times of high humidity. However, this effect is not
significant for most forms of deterioration except in swimming pools.
Capillary absorption and vapour diffusion have been studied by the
250 M. B. Leeming

author with respect to the use of surface coatings on concrete to Iimit

carbonation 16 and ingress of deicing salts into concrete. These same
mechanisms are also inftuenced by changes in mix design and other
measures to make concrete more impermeable, hence more durable.

CONTROLLING MOVEMENT OF WATER IN CONCRETE

Mix design. Correct specification and good workmanship are the most
important methods of achieving a durable concrete of low permeability. 1t
requires a low waterjcement ratio which is consistent with adequate
workability, allowing the concrete to be weil compacted, providing a dense
structure to the concrete. It needs to be properly cured and have the right
cover. These measures, combined with the choice of the right cement
content, are recognised by CP 8110 8 as the means of achieving the required
durability. When the exposure requires a stronger concrete than is needed
for structural purposes, and when workmanship is not of the best quality,
further methods to improve durability are required.
Chemical admixtures included in the mix can help to improve the
durability of concrete. Water-reducing admixtures help in maintaining a low
water/cement ratio with high workability. Air-eutraining admixtures are
mainly used in providing freeze/thaw durability. The precise action of air
entrainment in achieving this durability is still open to debate, however it
also affects workability to some extent and also the absorption of water.
Cement replacements such as pulverised fuel ash, ground blast furnace slag
cement and silica fume can produce moreimpermeable concretes. However,
the gain of strength of these cement replacement materials is slower than
normal OPCs and as a result they generally require Ionger curing to achieve
the improvement in durability.
Surface treatments have been extensively studied with a view to inhibiting
ingress of chlorides and carbon dioxide. They come in many forms from
penetrants through sealers to coatings. Treatments can have various
properties and it is necessary tobe clear as to what they are required to do
and to choose the right material accordingly. Most treatments can
significantly reduce the water absorption of an average concrete by a factor
of at least 20 and for the mostimpermeable coatings by a factor of 100. The
penetrant materials generally have little effect on carbon dioxide ingress
while sealers and coatings can give increasing resistance so that further
carbonation is limited to very low Ievels. Treatmentsgenerally have a lower
resistance to water vapour than to carbon dioxide and this property can
allow the concrete to breathe. Resistance to oxygen lies between that of
carbon dioxide and water vapour but the effect is insufficient to have any
significant influence on deterioration processes.
 Keeping Water Out of Concrete 251

Surface treatments are, however, applied in very thin layers up to 500 J..lffi
and are subject to deterioration themselves, such as embrittlement,
weathering dueto ultraviolet light and lack of adhesion. Thesesystems have
much shorter lives than the structure they are applied to and therefore need
recoating. These materials have not been in use long enough to generate
sufficient reliable data on their usefullives and on the change of protective
properties with time. There is evidence that ingress of chlorides, sulphates
and carbon dioxide is greater when the concrete is young so that this is the
period when treatments are most needed. However, to rely on a thin coating
for durability in lieu of normal covers of sound concrete seems poor
engineering judgement.
Mortar renderings have been used from very early times to provide a
durable skin to brickwork and other building materials. However, bond with
the substrate can be a problern and in the UK their use has strong regional
variations. Generally, in this country, durability requires concretes of a
similar strength tothat required for structural purposes and therefore to use
fair-faced concrete with the same mix throughout the member makes
economic sense. But where structural requirements are modest in severe
environments it seems sensible to use a durable skin to the minimum
requirement for the structured core. This philosophy, used in many
Mediterraneall countries, has a certain logic. The inherent durability ofthin
ferrocement members constructed by plastering techniques using high
cement contents and low waterjcement ratios surely points to possible
methods of achieving durability through renders. The corrosion protection
applied to the steel external columns of the Hong Kong Shanghai Bank 17 is
an extreme example of the use of renders.
Cladding. There are many other finishes, such as stone cladding, tiles,
mosaics, dry cladding, etc., which can be effective in keeping water out of
concrete. Where these systems provide an air gap at the face ofthe concrete,
consideration must be given to the higherrate of carbonation in the dryer
concrete. Corrosion could possibly occur unnoticed behind the covering in
periods of high humidity.
Waterproof membranes are widely used in roofs, for lining tanks and other
water-retaining structures. In bridges their use in the UK, in marked
contrast to the USA, has limited the amount of corrosion occurring in bridge
decks. However, problems have occurred when they begin to leak at
vulnerable points. These membranes have been found to have a finite life
and need renewing after 20 years or so.
Structural detailing can be a source of water ingress into concrete. A
Department of Transport report 18 found that most bridge joints leak,
causing darnage to piers and abutments due to chloride ingress. This darnage
can be eliminated by making bridge decks continuous. Ifstructural economy
252 M. B. Leeming

dictates simply supported spans, with a little ingenuity the deck slab can be
made continuous. Service troughs complicate waterproof membrane details
and are also a source of leakage of chloride-bearing water on to piers and
abutments. Again some ingenuity could be applied to eliminate these by
building in sufficient electricity and telephone ducts for all possible future
needs and building in gas and water pipes below the deck with extra wall
thickness so that they require no maintenance. Drips at cantilever edges are
correctly detailed but in many cases they channel the water along the edge
where it runs to the joint over a pier and down the beam face to be
concentrated on the pier top. lt is essential that the top of waterproofing
membranes have an escape path for the water that collects above them. This
is one of the causes of failure and corrosion darnage at bridge expansion
joints. The same problern can also be seen at drain outlets to buildings.
Drainage systems need to be simple, easily maintained and preferably
requiring no maintenance at all. Drainagesystems on bridges usually give
trouble due to the difficulty of accommodating standard gulley outlets and
lack of adequate falls. Consideration should be given to combining the
expansion joint and the drainage in one robust cross carriageway drainage
system.
Having prevented the ingress of water into one concrete face,
consideration must be given to all other possible water paths in the structure.
Many causes of the failure of surface coatings on concrete are due to
maisture on the rear side of the coating, causing lass of adhesion. Having
stopped the main source of water ingress, other minor sources become
apparent. Any method of contraHing water ingress on the surface of
concrete is more likely to fail if vapour pressure at the concrete/system
interface cannot be relieved. This is why 'breathability', i.e. high vapour
permeance, of a surface coating is an advantage.
One ofthe main objectives of contraHing the ingress ofwater into concrete
must be to try and achieve a steady and uniform state of maisture in the
concrete, particularly as far as corrosion is concerned. For instance,
corrosion in a tunnellining could be controlled with a vapour barrier on the
inside so that the concrete will remain in a permanently saturated state
where lack of oxygen restricts reinforcement corrosion. Maintaining the
concrete in a dry condition equivalent to an internal office environment
again limits corrosion.

RELEVANT TESTS
Absorption
As capillary flow or absorption has been identified as the primary
parameter with regard to the durability of concrete it is necessary to consider
 Keeping Water Out of Concrete 253

how this parameter can be measured. The simplest method of doing this is to
take a sample of concrete, seal the sides, place it face down on capillary
matting or in a shallow depth of water and to weigh it at predetermined
intervals. This is the method adopted by DIN 52617. 19 The results are
expressed in a graph as water gain plotted against the square root of time
which usually gives a straight line, the slope ofwhich is the water absorption
coefficient. This method was used by the author to compare the relative
performance of surface coatings on concrete in inhibiting the ingress of
water-containing deicing salts for the Transport and Road Research
Laboratory. In this instance a 15% salt solutionwas used and all faces ofthe
specimen except the test surface were sealed. The test ran for 21 days and
measurable gains ofweight were recorded provided that an accurate balance
was used. The uncoated plain concrete control specimens, however, rapidly
gained weight to near saturation within a few hours. This method is only
suitable for use in a Iabaratory but has the advantage that the specimens can
be preconditioned relative to a known humidity before the test. Other
similar tests 20 measure the depth of penetration of water against time.
There are other methods of test which monitor the volume of water
entering the test face. The simplest ofthese is the Karstentube which is fixed
to the test surface and measures the absorption ofwater to the nearest 0·1 ml
over an area of 450 mm 2 at a head of 115 mm. The coarse measurement over
a small area makes it only useful for very absorptive materials. A more
advanced test is the initial surface absorption test (ISAT) 21 which uses a
narrow capillary tube on a test area of 5000 mm 2 at a head of 200 mm. This
test, however, gives results in units of mllm 2 ls at fixed intervals of lümin,
30 min, 1 h and 2 h. These four figures are not very meaningful except in a
relative sense for comparison with other results on concrete. However, these
results can be expressedas a water absorption coefficient related to weight of
water absorbed per square metre ofsurface area related to the square root of
time by the following relationship:
X X 60l[(j(t + 1)160)- (jtl60)] (glm 2 ljh) (1)

where X is the ISAT reading in ml/m 2 ls and t is the time of reading in

minutes from start oftest. The units of glm 2 I Jh have been used as they give
relatively sensible figures on concretes which have been coated of up to
20glm 2 1Jh. An average concrete 14 would have a value between 750 and
1500 glm 2 I Jh.
The tests on surface coatings have been carried out by setting up the test as
specified in BS 1881: Part 5 until the point when the valve is closed after
lOmin. From that point on the test is modified so that the valve to the
reservoir remains closed and the total movement of the meniscus in the
capillary tube is recorded from the start point to the various recording times.
254 M. B. Leeming

The water absorption coefficient can be calculated in g/m 2 /Jh as follows:

X 1 x0·6/(Jt 2 -Jt1 ) (g/m 2 /jh) (2)
where X 1 is the movement in the capillary tube in numbers of units of
0·01 ml/m 2 /s in the period from t 1 h (normally 10 min) to t 2 h.
Tests have been carried out using the ISAT apparatus on the same
specimens that were used for the capillary test and reasonable correlation of
the results were obtained. However, the capillary test is basically a uniaxial
flow test while the ISAT method includes radial components. As a result the
simple root time relationship may be less valid. The ISAT method can be
used bothin the Iabaratory and on site, although it is not widely used as the
equipment is cumbersome; it can take up to 2 h to carry out a test and
achieving a good seal with the concrete surface can be difficult. However, the
principle of the test is sound and some of the above problems can be
overcome. More work is required in improving the test to make it more
convenient for site use.
There are other versions ofthe initial absorption test, some ofwhich apply
various pressure heads to the water which are reviewed in Ref. 14. As
capillary forces of up to 20m head of water have been measured in
concrete, 15 the application of small heads to the water on the surface arenot
significant. A simple test that has simple equipment and is easy to do has an
advantage.
The Figg water permeability test 22 works on much the same principle
except that the absorption is measured in a 10 mm diameter hole in the
concrete from 20 to 40 mm deep. This gives a relative value for absorption
within the cover zone ofthe concrete, but as the flow is radial in all directions
the results are not directly comparable with other methods of absorption
testing. This test has another difference-it tests the absorption into a freshly
cut surface of concrete, which has been found to give different values to the
absorption into cast surfaces. 23

Vapour Permeability (Diffusion)

Having measured the rate at which water is absorbed into the surface of a
concrete, the other parameter of interest is the rate at which it dries out. In
the tests described earlier on surface coatings, the specimens were taken off
the capillary matting and placed face up in achamberat 50% RH. Weight
loss over the next 21 days was recorded. This loss was found tobe up to five
times slower than weight gain for some specimens. The test is not rigorous as
the specimens did not all start from the same initial state of saturation.
This parameter is of importance in understanding the mechanisms of
water transport into and out of concrete in the environment but there are
few tests to measure it. lt is clear from the above that it can take Ionger for
 Keeping Wafer Out of Concrete 255

concrete to lose the water it gains. Vapour diffusion is a rnechanisrn obeying

Fick's first law. A water vapour diffusion test is described in Ref. 14, Section
4.6-the dry cup test. Another variant of the sarne test is the wet cup test,
which rnay be rnore relevant to actual conditions in that the relative
hurnidity difference is frorn 100% RH to, say, 50% RH as opposed to the dry
cup test where the difference is frorn, say, 50% RH to 0% RH. The test can
only be carried out in the laboratory on specially prepared specirnens which
rnay have cut or cast surfaces or a cornbination of both. The difference
between cut and cast faces rnentioned above rnay well apply to these tests, as
has been found for oxygen diffusion 24 which was found to be about ten
tirnes lower through a specirnen with two cast faces than through a specirnen
with two cut faces.
The Figg air perrneability test 22 gives a relative rneasure of the air
diffusion within the cover zone of concrete, but for the sarne reasons given
above for the Figg water perrneability test it cannot be related to an absolute
rneasure of vapour diffusion for cornparison with the rate of water
absorption. Schonbin and Hilsdorf 11 describe a sirnilartest which is applied
to the surface of the concrete.

Moisture Content of Concrete

All the above tests are dependent on the rnoisture content of concrete. In a
laboratory situation the specirnens can be conditioned to a known relative
hurnidity, allowing sorne cornparison between various results. Deterrnining
rnoisture content in the laboratory can only be done accurately by weighing
and oven drying, which is destructive with respect to further tests. On site the
rnoisture content of concrete is not easily deterrnined. Maisture rneters are
available which work on the principle of resistance or capacitance of the
concrete but these are far frorn accurate. Part ofthe problern ofthese rneters
isthat they do not rneasure water directly and can hence be affected by other
constituents. Parrot 25 uses a rnethod based on the relative hurnidity of the
air in a sealed hole in the concrete but this rnethod cannot be quickly used on
site. A rnethod based on the attenuation of rnicrowaves is under
developrnent and offers sorne hope of rnore accurate rneasurernents.

SUMMARY

In all forrns of deterioration of concrete water is a rnajor influence, either

encouraging the chernical action or conveying salts causing the darnage. lt is
a rnajor influence in the corrosion of reinforeerneut in concrete. As a rnajor
constituent of the concrete rnix it influences the durability in several ways.
The rnovernent of water into concrete is a cornplex rnechanisrn but, where
256 M. B. Leeming

concrete is not permanently submerged, capillary absorption is the

dominant factor in water take-up followed by vapour diffusion on drying.
The movement ofwater into or out of concrete can be controlled in many
ways by basic mix design, addition of cement replacements and other
chemical admixtures, surface treatments, renderings and other claddings.
Inadequate detailing can allow water to enter the fabric of the structure,
causing Jocal damage. There are tests to measure the capillary absorption of
concrete, both in the Iabaratory and on site, which arenot used extensively.
Cube strength is the main acceptance criteria for newly cast concrete but
durability rather than structural strength often determines the mix criteria.
Yet strength and durability do not necessarily go hand in hand. The drawings
should state the durability requirements as well as those for strength.
The initial surface absorption test (BS 1881: Part 5) is based on a sound
principle and has the possibility ofbeing a useful measure ofthe durability of
concrete but as presently configured is cumbersome, Jengthy and prone to
Jeakage. lmprovements in the method and equipment are possible, Jeading
to a greater versatility with meaningful camparisans ofresults from both the
Iabaratory and the field.
The vapour diffusion of concrete is much Jess easy to measure in the
Iabaratory and is not possible to measure on site. Research has shown that
water can be taken up by concrete much more quickly than it can lose it by
vapour diffusion. Alternate wetting and drying can be more damaging to
concrete than a steady state of maisture content, particularly with regard to
corrosion. However, the environmental regime at the surface of concrete is
little known.
The maisture content of concrete is an important factor in the
mechanisms of deteriorating influences yet its measurement in the
Iabaratory is destructive. Maisturemeters are available for site use but are
very inaccurate. Microwave attenuation may provide a more accurate
means of measurement.

REFERENCES
1. Concretes in sulphate-bearing soils and groundwaters. Building Research
Establishment Digest 250.
2. SAWAN, J., Cracking due to frost action in Portland cement concrete
pavements-a Iiterature survey. The Katharine and Bryant Mather Inter-
national Conference on Concrete Durability. American Concrete Institute,
SP100, Vol. 1, 1987, SP100-44, p. 781.
3. BJEGOVIC, D., MIKULIC, D. and KRAINCIK, V., Theoretical aspect and methods of
testing concrete resistance to freezing and deicing chemicals. The Katharine and
Bryant Mather International Conference on Concrete Durability. American
Concrete Institute, SP100, Vol.l, 1987, SP100-53, p. 947.
 Keeping Water Out of Concrete 257

4. Proceedings 8th International Conference on Alkali-Aggregate Reaction, Kyoto,

Japan, 17-20 July 1989. The Society of Material Science, Japan, and
proceedings of several earlier conferences in the series.
5. BATE, S. C. C., High alumina cement concrete-An assessment from Iabaratory
and field studies. The Structural Engineer, 58A(12) (December 1980).
6. PoPOVICS, S., A classification of the deterioration of concrete based on
mechanism. The Katharine and Bryant Mather International Conference on
Concrete Durability. American Concrete Institute, SP100, Vol. 1, 1987, SP100-
10, p. 131.
7. NuRNBERGER, V., Chloride corrosion of steel in concrete. Betonwerke und
Fetigteil-Technik, Part 1, No. 9 (1984); Part 2, No. 10 (1984).
8. Structural use of concrete, Part I. Code of practice for design and construction.
British Standards Institution, BS 8110: Part 1, 1985.
9. MEYER, A., The importance of the surface layer for the durability of concrete
structures. The Katharine and Bryant Mather International Conference on
Concrete Durability. American Concrete Institute, SP100, Vol. 1, 1987, SP100-
5, p. 49.
10. SENBETTA, E. and MALCHOW, G., Studies on control of durability of concrete
through proper curing. The Katharine and Bryant Mather International
Conference on Concrete Durability. American Concrete Institute, SP100,
Vol.1, 1987, SP100-7, p. 73.
11. ScHONBIN, K. and HILSDORF, H., Evaluation of the effectiveness of curing of
concrete structures. The Katharine and Bryant Mather International
Conference on Concrete Durability. American Concrete Institute, SP100,
Vol. 1, 1987, SP100-14, p. 207.
12. PoTTER, R. and Ho, D., Quality of cover concrete and its influence on durability.
The Katharine and Bryant Mather International Conference on Concrete
Durability. American Concrete Institute, SP100, Vol. 1, 1987, SP100-25, p. 423.
13. NAIK, T. R., Concrete strength prediction by the maturity method. American
Society of Civil Engineers, ASCE Convention, Boston, April 1979, preprint
3576.
14. Permeability testing of site concrete-A review of methods and experience.
Concrete Society, Technical Report No. 31, 1987.
15. GuNTER, M. and HILSDORF, H. K., Stresses due to physical and chemical actions
in polymer coatings on a concrete substrate. ISAP 86 .. Adhesion hetween
polymers and concrete, RILEM, Aix-en-Provence, September 1986, ed. H. R.
Sasse. Chapman and Hall.
16. Protection ofreinforced concrete by surface treatments. Cdnstruction Industry
ResearchandInformation Association (CIRIA), Technical Note 130, 1987.
17. ZuNz, G. J., GLOVER, M. J. and FITZPATRICK, A. 1., The structure of the new
headquarters for the Hong Kong and Shanghai Banking Corporation, Hong·
Kong. The Structural Engineer, 63A(9) (September 1985) 255-84.
18. The performance of concrete in bridges. A survey of 200 highway bridges.
Department of Transport, HMSO, April 1989.
19. Determination of the water absorption coefficient of building materials. DIN
52617, December 1984, Deutsches Institute für Normung e.v.
20. BAMFORTH, P. B., PococK, D. C. and RoBERY, P. C., The sorptivity of concrete.
Our World in Concrete and Structures Conference, Singapore, 27-28 August
1985.
258 M. B. Leeming

21. Methods oftesting hardened concrete for other than strength. British Standards
Institution, BS 1881: Part 5, 1970.
22. FIGG, J. W., Methods of measuring the air and water permeability of concrete.
Magazine ofConcrete Research, 25(85) (December 1973) 213-19; also 36(129)
(December 1984).
23. KREIJGER, P. C., The skin of concrete, composition and properties. Materiaux et
Constructions, 17(1 00) (1984) 275-83.
24. GJORV, 0. E., VENNESLAND, 0. and EL-BusAIDY, A. H. S., Diffusion of dissolved
oxygen through concrete. NACE Corrosion 76, Paper 17, Houston, March
1976.
25. PARROT, L. J., Maisture profiles in drying concretes. Advances in Cement
Research, 3 (July 1988).
 23
Reinforced Concrete Bridge Protection in
N orthern Ireland

F. R. MONTGOMERY
Department of Civil Engineering, the Queen's University of Be/fast,
Belfast BT7 1NN, UK
and
A. McC. MURRAY
Department of the Environment Roads Service, Commonwealth Hause,
35 Castle Street, Be(fast BT11GU, UK

ABSTRACT

Of the 7700 road bridges in Northern Ireland, approximately 1300 are of

reinforced concrete construction. Same of thesedateback to the early years of
this century but many !arge reinforced concrete bridges were constructed in the
1960s and early 1970s during the development of our motorway system. Since
1973 all of these bridges have been maintained by the Roads Service of the
Department of the Environment for Northern Ireland. In 1987 the DoE (NI)
commissioned the Civil Engineering Department of the Queen's University of
Belfast to investigate the current condition of the bridges and to make
recommendations an a programme for proteelive measures. This has
concentrated an surface treatments and an cathodic protection. lt is planned
that a coordinated maintenance management strategy will result from this
work.

INTRODUCTION

The Northern Ireland road network consists of some 23 700 km of adopted

road carriageway on which there are more than 7700 bridges with spans
greater than 2m. Prior to 1973 the maintenance of the bridge stock was
organised along similar lines to the system in Great Britain, with the various
259
260 F. R. Montgomery and A. McC. Murray

councils undertaking or arranging the maintenance and the Ministry of

Development fulfilling a role similar tothat ofthe Department ofTransport.
When local government was reorganised in 1973, responsibility for the
maintenance ofthe province's road network and the associated bridgeswas
transferred to the Roads Service of the Department of the Environment for
Northern Ireland. Roads Service has a headquarters office in Belfast and six
divisional headquarters throughout the province. lt employs almost 1400
professional, technical and administrative staff, and around 2000 industrial
staff.
This paper includes details of the existing bridge stock in Northern
Ireland and reports on the initial results of a research project which is
currently being undertaken by the Department of Civil Engineering of the
Queen's University of Belfast in conjunction with DoE (NI). The research
has included an appraisal of the conditions of the reinforced concrete
bridges in the province and an investigation of the various means of
protecting them against further damage, so that an effective maintenance
strategy can be developed.

THE NORTHERN IRELAND BRIDGE STOCK

The relatively high number ofbridges in Northern Ireland is accounted for

by the dense network of roads. However, of the 23 700 km of carriageway
less than 10% is either motorway or dass A while almost 60% is unclassified.
One consequence of this largely rural and static network is that a high
proportion ofthe bridges, for which the department is responsible, are both
old and minor. For example, approximately 75% of the bridges are of
masonry arch construction and the vast majority of these have short single
spans. Of the remaining 1970 bridges, 1300 are of reinforced concrete
construction and 165 are a combination of reinforced and prestressed
concrete.
The reinforced and prestressed concrete bridges, which are the subject of
this paper, are fairly evenly distributed around the province but the larger
and newer structures are heavily concentrated in the greater Belfast area or
on the motorway network. While there are some relatively old reinforced
concrete bridges in the province, more than half of the total number were
built during the 1960s and 1970s.
The condition of the reinforced concrete bridges is exceptionally good and
so far large-scale repairs have only been required to a very small number of
structures. In recent years, however, Roads Service bridge inspectors have
increasingly noted the early warning signs of problems, such as rust staining,
limited cracking or spalling of concrete. These observations, along with
 Reinforced Concrete Bridge Profeetion in Northern Ireland 261

some more detailed reports on the worst affected structures, led to the view
that more widespread problems could be on the horizon and that a formal
maintenance policy would be required to Iimit their impact.

INSPECTION PROGRAMME AND DATA HELD

When the department took responsibility for the bridge stock in 1973,
records and inspection data were held in inconsistent forms, were of poor
quality in many areas and often the local foreman was the best source of
information. As the importance of holding bridges data was recognised,
divisions endeavoured to improve their systems, but it was not until 1980
that a standard procedure was adopted for the whole province. This
procedure, which involved general inspections every 2 years and principal
inspections every 6 years, was updated in 1985 to take account of the
Department of Transport's record and inspection systems. The data, which
are now largely complete for the province, is held on DEC Rainbow
computers using the sensible solution database manager. By 1991, however,
the data will have been transferred to the integrated computer system which
Roads Service is currently installing at a cost of approximately f8 million.
This system will offer greatly enhanced interrogation facilities.

DEVELOPMENT OF A MAINTENANCE STRATEGY

As the routine inspection programme built up a databank of results, it

became clear that some of the older reinforced concrete bridges in
vulnerable positions were beginning to show signs of distress. Detailed tests
carried out on both damaged and so und parts of these bridges revealed the
need for some form of protective systems either to Iimit further chloride
ingress or to prevent rebar corrosion from taking place by some other
means. An initial assessment of the available Iiterature revealed that there
was little consensus of opinion on the preferred protective system. It was,
however, clear that the most cost effective form of action would depend
heavily on the amount of chloride ingress or darnage which the bridge had
already suffered.
While some policy decisions were made to commence modest pro-
grammes of silane treatment and deck joint repairs, it was decided that a
maintenance strategy could not be formalised until the condition of the
Northern Ireland bridge stock had been objectively assessed and until the
various methods of protection had been appraised. lt was this background
which led in 1987 to the Department of Civil Engineering of the Queen's
262 FR. Montgomery and A. McC. Murray

University of Belfast being commissioned to undertake a 3-year research

project to examine the Northern Ireland bridge stock and make
recommendations on its future maintenance, especially in relation to
protection against further damage.
Although the funding available for this project was relatively modest, the
aims were broadly similar to those described to G. Maunsell and Partners
for their survey, the results of which have recently been published 1 and are
being examined with interest. The Queen's University project will, however,
concentrate on the Northern Ireland context.

PROTECTIVE SYSTEMS AVAILABLE

M uch has been learned from hard experience in the last few years on how to
construct concrete bridges to make them durable, but only some of the
techniques available for new construction are of use to improve existing
structures. The idea of applying liquid sealants or penetrants to the surface
of reinforced concrete to provide protection against the ingress of marine or
deicing salts has been common for a number ofyears in West Germany and
North America. The results of research which examined the performance of
a range of concrete sealants for the protection of bridge structures were
published by the National Co-operative Highway Research Program ofthe
USA as Report 244 2 in 1981. Subsequently other work has been reported,
much ofit in the USA. A useful recent summary ofthis work is contained in
Ref. 3.
lt seems that three systems, a monomeric alkylalkoxy silane, a silane-
siloxane overcoated with a methacrylic polymer and an epoxy casting, are
being particularly recommended. The silane system is getting more use in
USA than the others, although the two-coat silane and acrylic is being used
for specialhigh quality work only, due to its greater expense. Bothofthese
systems are being seriously considered for Northern Ireland's bridges,
indeed the silane has already been applied ahead ofthe results ofthis present
study. However, it is envisaged that they will be used only for recent
construction or where there has not been much chloride penetration. In the
case of silane it is quite likely that the criteria detailed in recent DTp draft
publications 4 •5 will be adopted.
For those structures showing mild to modest amounts of chloride or
carbonation-induced rust staining or spalling one of a number of
proprietary concrete cover replacement systems is envisaged, together with
removal of the cause of the problem, if possible, often found to be badly
maintained deck expansion joints.
As a protective measure for those structures showing severe rebar
 Reinforced Concrete Bridge Profeetion in Northern Ireland 263

corrosion and spalling from chloride penetration a last resort is the

installation of an induced current cathodic protection system. Most of our
problems where this may be warranted are on bridge substructures. Even in
the USA the experiences of this technique on substructures is limited 3 and
very much in the developmental stage. One suchsystemwas installed during
1988 on a motorway bridge in Northern Ireland. All three piers have been
protected after first having had allloose concrete and rust removed and then
been gunnite repaired. The cathodic protection system used comprises a
conductive polymer strip bonded vertically, at 24-in centres, to the pier
faces. They are powered by a conventional controller rectifier. However,
problems with this initial installation have been discovered during routine
monitoring. It has emerged that high current concentrations in persistently
wet areas have led to acid formations at the anode/concrete interface. This
has caused partial debonding of some anodes. At present an assessment of
alternative anode systems is being finalised and it is hoped that the
installation of a new systemwill begin shortly. Experience gained from this
trial project will enable morevalid decisions tobe made on any further use of
the technique.

LABORATORY INVESTIGATIONS

The sealants and penetrants mentioned previously appear on the market in

many forms and under many names. It is possible to apply them in a number
of ways, such as single coat, two coat, full strength or dilute. They may be
applied to dry surfaces or wet surfaces and anything in between. To some
extent their performance is inftuenced by the quality ofthe concrete to begin
with. All of these factors are being investigated in an extensive test pro-
gramme designed to assess change of surface permeability due to treatment,
improvement in chloride penetration resistance and, very importantly, the
likely duration of any derived benefit.
The Iabaratory investigations are currently being extended to field work
where similar, though less all-embracing, measurements are being made on
trial structures already treated to detect any effects due to variability in field
applications.

CONDITION SURVEY

It is obviously impossibly expensive to perform a very detailed survey and

analysis of all our concrete bridges but they are all subject to the periodic
inspection mentioned previously.
264 FR. Montgomery and A. McC. Murray

A second level of inspection is presently being performed where a sample

selection of bridges is being submitted to a range of test techniques. The
sample has been chosen to represent the range of bridge types by size and
construction form, by type of environment and by age. Within the sample
there are those with no known problems and others with obvious problems.
lt is hoped to include about 20 bridges in this survey, which is about 50%
complete at the time of writing and is tending to confirm the beliefthat the
bridge stock is in reasonably good condition.
The test techniques being used include chloride measurements on exposed
faces, surface permeability and strength measurements, rebar cover
assessment and half-cell potential mapping. Visual observation of sources of
trouble, if any, and photographing and measuring local factors such as
distance from carriageway, etc., help to place all measurements in
perspective.
A third level of inspection has been performed on a small number of
bridges, all of which are showing varying degrees of distress. Basically the
same techniques as above have been used but here many more readings have
been taken. Theinformation gained in this series will enable threshold values
to be placed on measurable parameters above which darnage is likely to
occur.

CONCLUSIONS

Much information has yet to be properly analysed and correlated but

already some factors are beginning to emerge. lt is feit that, in broad terms,
likely performance can be quantified by combinations of measurable
parameters.
Knowing these parameters and having available on the data base many
factors which will enable the grouping tagether of structures of similar age,
similar construction, similar environment and similar maintenance history
will, it is hoped, enable a programme of planned protection tobe carried out.
The results available from the early Iabaratory tests of the various
protective systems indicate differences of performance on relatively new
concrete. Accelerated weathering is proving some data on the likely long-
term performance, but just how some of the applied protective systems will
actually fare under real conditions is not yet clear. There is a lack of
information in the Iiterature on the useful life of most applied protective
systems. This is probably the aspect which we find most difficult to
incorporate in our programme of planned maintenance.
However, with the best information we can find or can generate a strategy
is being devised and the hope isthat time will prove it to have been profitable.
 Reinforced Concrete Bridge Profeetion in Northern Ireland 265

RE FE RENCES
1. WALLBANK, E. J., The performance of concrete in bridges. A survey of 200
highway bridges. A report prepared for the Department of Transport by G.
Maunsell and Partners. HMSO, London, April 1989.
2. Concrete sealers for protection of bridge structures. National Co-operative
Highway Research Program, Transportation Research Board, National
Research Council, Washington, DC, December 1981, Report 244.
3. TASKER, J., HUMPHREY, M., McANOY, R. and MüNTGOMERY, F. R., Coatings for
Concrete and Cathodic Protection. Thomas Telford, London, 1989.
4. Criteria and material for the impregnation of concrete highway structures.
Departmental Standard, Department of Transport, London, 1989 (in draft).
5. Impregnation of concrete highway structures. Departmental Advice Note,
Department of Transport, London, 1989 (in draft).
 24
Rebar Corrosion-FBECR: The Fight to Cure the
Problem

JAMES A. READ

W S. Atkins Cansultants Limited, Waadcate Grave, Ashley Raad,

Epsom, Surrey KT18 5 B W, UK

ABSTRACT

The recent histary o{rebar corrosion an bridges in the U K and N orth America
due to ingress a{chloridesfrom road deicing is outlined. The UK investigatian
into the salution usingfusion-bonded epoxy-coated reinforcement ( FBECR)
bJ' both TRRL and the author's company are referred to. The main prablems
andfaults to the reinforcement and concrete an bridges is autlined tagether
with generat remedial measures.
An example ol a propased replacement cross-beamfor a bridge substructure
is putforward using FBEC Rasareplacement far uncaated bars. Finally, the
main methods o{FBE application are described 1vith sur{ace preparation and
homogeneity of the curedfilm heing discussed. A n?sume of U K application
using FBEC R is given at the end of the paper.

THE RECENT HISTORY OF REBAR CORROSION ON BRIDGES

The UK situation on concrete deterioration of bridges constructed with

uncoated reinforcement and subjected to road de-icing using chlorides
and in the marine environment is increasing. All parts of the UK are
experiencing these problems.
The recent report on the Midlands links motorway viaducts reported in
the New Civil Engineer (February 1989) 1 is a typical example, with the DTp
reported as saying 'it is a latter-day Forth Bridge requiring monitoring,
maintenance and repair for the rest of its useful life'.
Another example is the Tay Bridge, where salt from the marine waters has
penetrated the concrete piers of the bridge in the splash zone, leading to
severe attack on the uncoated reinforcement.
267
268 James A. Read

The North American situation is as bad if not worse. The recent study
mission organised by the Institution ofCivil Engineers and supported by the
Department of Trade and Industry 2 to study surface coating, cathodic
protection and epoxy coating of reinforeerneut to concrete reported that
'the overall picture of highway bridges, however, is one of deteriorating
condition ... Figures quoted for the USA of between f9000 and f14000
million for necessary repair and replacement of bridges....
One state alone-Pennsylvania-has 22 500 bridges with spans in
excess of 8 metres and 35 percent classified as structurally deficient or
functionally obsolete.'
The report goes on to state after field surveys:
'Epoxy coating was found tobe effective in protecting the reinforcing steel
in chloride-contaminated concrete during the nine years of exposure, and
even where the coating was damaged there was only superficial corrosion.'
In New York State:
'Confidence about the performance of epoxy coatings is so high that no
evaluation of the technique is being included in the current Strategie
Highways Research Programme.'
Babaei and Hawkins, in their 'Evaluation of bridge deck protective
strategies', 3 after examining ten different strategies used throughout the
USA and noting that a number of states bad discontinued using certain
types of protective strategy because of
'problems such as cracking in and debonding of overlays, wear and
stripping of asphalt overlays, or the ineffectiveness of some types of
sealers ... .'
goes on to state:
'Among the strategies used as standard practice, epoxy coating of bars is
the most popular. Forty-one states use this method, either alone or in
combination with other strategies.'
The TRRL Report 667 on 'FBECR in bridge decks' by J. Willis, 4
published in 1982, concluded that
'epoxy coating of the top steel in addition to current waterproofing
practice would provide-at relatively little extra cast-additional
assurance that the reinforcement would be adequately protected
throughout the life of the bridge.'
An investigation by W. S. Atkins into the specification performance of
 Curing Rehar Corrosion using FBECR 269

FBECR and the merits of ASTM A 775 was reported in the author's paper
presented at the CIRIA/BSE Conference in Bahrain in October 1987. 5
The work carried out by R. R. Bisbop of TRRL on behalf of the
Department of Transport is published in their Application Guide 6: The
specification of epoxy-coated reinforcement bars, 1987. 6 This concludes
that:
'lt is not practical to use the American Standard ASTM 775 (in the UK)
because the division oftests between those intended for quality assurance
of coated bars and those for certification of powder is not satisfactory.'
The work carried out by both Atkins and TRRL reach similar
conclusions on the shortcomings of the American specification in regard to
its suitability for use in a UK or European context.lncluding as reasons the
different types of steel used for rebar between North America and the UK/
Europe, bond/adhesion tests, measurement of coating thickness, coating
darnage and repair, etc., Bisbop makes the point that:
'This is not to be construed as a criticism of epoxy-coated bars. On the
contrary, there is much evidence of the value of epoxy-coated
reinforcement bars in concrete exposed to a corrosive environment.'

MAIN PROBLEMS WITH CONCRETE DETERIORATION ON

BRIDGES

The main cause of premature deterioration in the reinforced concrete road

and bridge stock in both the UK and North America is penetration of
chlorides, used for deicing, into the concrete and reaching the uncoated
reinforcement, depassivating the normally alkaline layer around the surface
ofthe bar after which, provided moisture and oxygen are present, corrosion
of the rebar takes place, leading to loss of bond between steel and concrete,
concrete cracking and eventually spalling of the concrete and loss of
structural integrity. Whilst different waterproofing strategies have been
employed between and within the UK and North America, to a ]arge extent,
the above effects have occurred, are still occurring and will certainly
continue for the foreseeable future on those existing structures having
uncoated reinforcement and where chlorides are used for road deicing.
Secondary causes may be attributed to a number of other factors, such as
reduced cover to the reinforcement, new design techniques, changes in raw
materials, the constant search for more slender structures coupled with
commercial pressures associated with new types of contract, performance-
type specifications, and an increasing tendency towards a Iack of on-site
impartial and independent inspection and checking during construction.
270 James A. Read

The methods and techniques of inspection, surveying and monitoring of

concrete deterioration in bridge structures have been developed over the past
20-odd years but there is still a Iot to learn. They include visual surveys of
surface condition, permeability and porosity measurements, testing for
chloride Ievels and carbonation of the concrete at various depths, haif-eeil
potential surveys to locate sites of active reinforeerneut corrosion,
petrographic examination of aggregates to check for alkali~silica reaction,
delamination surveys and measurement of ultrasonic pulse velocities. The
interpretation of electric potential methods and non-destructive permea-
bility and porosity measurements are particularly difficult. They are
sensitive to the methods used when carrying out the tests and results can
vary due to changes in environmental conditions.

MAIN FAULTS TO CONCRETE ON BRIDGES

Notwithstanding the differing designs and techniques being utilised, the

concrete deterioration on bridge structures, which nearly all use concrete
with uncoated reinforcement, is remarkably similar and the main faults
exposed over the last 10~15 years can be listedas follows:

• Cracking and spalling of concrete columns together with corrosion of

reinforeerneut due to chloride ingress either from spray from passing
vehicles, leakage down columns from above or within the splash zone
when columns are standing in marine waters. This is often associated
with Iack of specified cover and a loss of cementitious material in the
concrete matrix (see Fig. 1).
• Cracking and spalling of concrete in cross-beams and shear walls, often
with effiorescence and corrosion of reinforcement, again due to chloride
ingress from spray and condensation from below, leakage from above
via construction/expansion joints, etc., and often associated with less
than specified cover (see Figs 2~4).
• Cracking and spalling to deck slabs tending to be limited to ends of
slabs at construction and expansion joints; also along kerb lines,
particularly where there are drainage outlets. Deterioration darnage
has also occurred where waterproof membranes have been damaged
either during initial construction or during resurfacing operations.
• Cantamination and deterioration of concrete to plinths and bearings
supporting deck beams due to leakage from above and bad drainage
around bearings.
• Disintegration of central reserve upstands which often carry a safety
fence and lighting columns. Due to initially frost darnage and then
 Curing Rebar Corrosion using FBECR 271

chloride ingress followed by cracking and spalling. This also occurs to

precast concrete edge parapets (see Fig. 5).
Often associated with the above, most ofwhich can be detected because of
their surface visual effects, is widespread localised pitting corrosion. This can
be much more dangeraus as it can proceed undetected because the products
arenot expansive in the early stages and therefore the usual tell-tale signs of
cracking and spalling do not occur.

GENERAL REMEDIAL MEASURES

If detected early enough many remedial measures may be effective in halting

and preventing further deterioration progressing. These include providing
better drainage, sometimes by installing extra guttering and downpipes;

F1G. I. Spalling at the base of a column.

272 James A. Read

FIG. 2. Typical cross-beam with cracking and rust stain.

FIG. 3. Leakage of water onto a cross-beam.

 Curing Rehar Corrosion using FBECR 273

FIG. 4. Shcar wall with cracking and cffioresccnce.

FIG. 5. Disintegration of thc central rcserve upstand.

274 James A. Read

cutting out damaged concrete and replacing with or without additives; and
the use ofvarious coatings, some surface, some ofthe impregnation type and
sometimes both. Remaking joints using better sealants and mastics and
cathodic systems may also be effective for certain situations.

REPLACEMENT OF CONCRETE MEMBERS

However, if the deterioration has progressed beyond the stage of any repair
remedial measure being suitable, the replacement of the concrete members
becomes necessary. If this is the case then the author would recommend that
coated reinforeerneut could be used in the replacement members even if
other measures are incorporated. The extremely high disruptive costs
associated with repairs to road-running surfaces and replacement of
substructure members whilst traffic is allowed to continue overhead make
the small extra cost to coat the rebar insignificant when weighed against the
possibility of recurrence in the future.
The replacement of road-running members such as parapets, central
reserve upstands and parts of the deck are relatively Straightforward
although highly disruptive. The replacement of substructure elements,
however, is even more costly and therefore a solution is given here for a
typical composite concrete deck supported on steel longitudinal beams
carried by reinforced concrete cross-beams and columns, and where the
cross-beams have tobe replaced (see Fig. 6). Figure 6 shows a typical cross-
beam Iayout.
The major operation necessary before the existing cross-beam and/or
SHEAR WALLSTOBE REMOVED AND REPLACED R C DECK
WITH STEEL TRIMMER BEAM AND K BRACING
BETWEEN ALL DECK BEAMS \\STEEL DECK BE AMS

--- _l\ e-

\
CANTILEVERS TOBE\
1
\ IÄ CROSSBEAM TOBE
DEMOLISHED FIRST llio.:.. PROPPED DURING
oc~ 0

STEEL ROCKER BEARINGS.\ I DEMOLITION ~00

TOBE REPLACED BY
SPHERICAL BEARINGS
!
~~
STARTER BARS FR0~1 COLUMN.J
TOBE RETAINm !Cl CROSSBEAM RC CROSSBEAM TO 8 E---
CUT INTO SUITABLE
--- BEA RINGS

LENGTHS FOR SAFE

----~E?UND
EXISTING SPREAD-
FOOTING
--

lGROUIID
LEVEL
l__ - I
PART ELEVATION SECTION A-A

FIG. 6. Typical cross-beam Iayout.

 Curing Rebar Corrosion using FBECR 275

''
:
' I

E ISTitiG CROSSBEA~t-
TOBE REPLACED BY A
!. / •!
o ,
SIMILAR CRDSSBEAt·l l :
CAST IN SI TU i 1

PORTIO lOFCDLUMtl~ l
10 BE DEMOUSHED •

PART ELEVATION
SECTION A- A

FIG. 7. Proposed propping solution.

columns can be removedjrepaired is to relieve the Ioads from the running

deck and this has to be done by temporary supports. Figure 7 shows a
proposed solution for propping the carriageway. This is obviously very
expensive, but if a number of beams have tobe replaced then the propping
can be designed to be largely reusable.

2 LAY ERS OF 20- 1:Va"0 20 - 1:Vs·0

5;a•0EACH FACE
Y2 '0 EACH FACE

PART ELEVATION SECTION A-A

FIG. 8. Rcinforccment of a cross-beam.

276 James A. Read

A typical uncoated rebar in a cross-beam is shown diagrammatically in

Fig. 8.

CHANGES NECESSARY IF FBECR IS USED

Assuming a cross-beam of 33m overalllength and 1·5 m cross-section with

reinforeerneut as in Fig. 8 using mainly 32, 20 and 12 mm rebar, then:
• All rebars to be changed to high yield deformed type II to BS 4449,
coated to meet the draft British Standard but amended to meet DTp
and W. S. Atkins Consultants Limited specification, cut and bent to the
new BS 4466 complete with patch kits, tying wire and plastic end caps.
• Lap lengths to be increased by 20%, in order to allow for the present
doubts on bond criteria.

Bond
The following results (Table 1) have been produced from current unisteel
deformed bars coated by Allied Bar Coaters of Cardiff and tested in
accordance with BS 4449, Appendix B1, requiring six barstobe tested with
none exhibiting free end slip greater than 0·2 mm when loaded.

TABLE 1
----------------

Bar diameter Free end slip (mm)

(mm)
Uncoated E-bar

32 0·017 0·082
20 0·005 0·062
-------------

DISCUSSION ON COSTS OF USING FBECR

1t is estimated that the cost of replacing a cross-beam of the type described

above can be as much as f500 000 if the full cost of temporary works is
included. The rebar tonnage is about 25 t. Replacement of uncoated rebar
would cost about f9000 delivered to site. The cost of coating would add
about f7000, also cut, bent and delivered to site. There should be no
difference in fixing costs as the handling underneath a live roadway will have
to be done with care and by skilled steel fixers anyway.
1t is obvious from these figures that the small increase in the overall cost of
replacement is insignificant when weighed against the increased assurance
 Curing Rehar Corrosion using FBECR 277

that the use of eoated bar will give for durability. Whilst this is a small
pereentage of the total eost of replaeement for new works with no temporary
works the pereentage additional eost will be signifieantly higher but still
justifiable in terms of the eonfidenee in the future durability and life.

SOME POINTS ON APPLICATION/QUALITY OF FBECR

There are a number of points on the manufaeture of FBECR whieh if not

properly eontrolled and monitored ean Iead to the possibility of disastrous
breakdown in the proteelive properties of the eoating. A few of the main
ones (in the author's experienee/viewpoint) are given as follows.

Method of FBE Application

The three main methods of applieation of FBE on rebar are:
(i) Eleetrostatie spraying~for straight bars.
(ii) Tribostatie spraying~for straight bars.
(iii) Fluidised bed dipping-for eomplex shapes, fittings and welded
mesh reinforeerneut eages.
The first two involve spraying FBE granules eleetrieally eharged from
speeially designed nozzles on to the preheated bars. The bars have to be
temperature eontrolled between 220 and 260oC at the time the granules
toueh to allow proper melting, flowing and then euring. The eleetrie eharge
indueed in the granules is eonsiderably different by the two methods. In
eleetrostatie spraying it is indueed by injeeting epoxy powder granules into
an airstream around an eleetrieally eharged discrete eleetrode which imparts
a eharge of about 45 000-50 000 V to the granules. For tribostatie spraying
the eharge is indueed by cireulating the powder in a spiral mode between
PTFE surfaces, the granules beeoming eharged by frietion; in this case the
charge imparted is about 25 000-30 000 V.
The plant, equipment and nozzles for the two methods are eonsiderably
different, but what is even more important is that the granules themselves
have tobe speeially formulated for the two different methods. For fluidised
bed dipping the granules are held in suspension in a bath and the preheated
eage dipped into the bath and heldforatime suffieient to allow the fusion to
take plaee. F or this method the granules do not need to be eleetrieally
eharged and as there is usually no requirement for bending of the bars after
application the formulation ofthe granules is againdifferent to the other two
methods. (See Fig. 9 showing an example of a eage eoated by this method.)
If the granules formulated for one of these proeesses is used for another
then it is likely that the fusion bonding may not be effeeted properly.
278 James A. Read

FIG. 9. A prefabricated wclded reinforcement cage just having been fusion-bonded epoxy-
coated by the fluidised bed dipping technique at Webb Coatings Systems. Birmingham, on 29
November 1988 (courtesy of 3M UK).

Surface Preparation
The surface preparation of the bar prior to coating is perhaps one of the
most potentially dangerous. If this is not carefully monitared then defects
result which will often only become apparent in use. The three main points to
watch are as follows.
Surface defects such as shelling, rolling Japs, blowholes, fissures, slag
inclusions, slivers, scabs, etc., can be detrimental to the adhesion of the
coating and its subsequent protective properties. These defects are lia ble to
be more prevalent on rebar because much reinforcement is manufactured
from steel of a 'lower quality' than would be permitted for other structural
uses and the rolling processes used to produce the deformations will often
promote these defects. The abrasive blasting used for cleaning can expose
these defects. The profile and amplitude of the blasted surface will be
dependent on the type of grit and shot used, pressures, number of times the
material is recycled and the equipment used. It is suggested that an angular
profile produced by a mixture of grit and shot abrasive blast cleaning to give
an amplitude in the range 50- 75 flm should give a suitable surface for
subsequent coating.
Of course the surface must be clean of rust, oils, greases, chemical
contaminants and dust residues, and the cleanliness should be checked
immediately prior to the coating process.
Figures lO and 11 show two examples of the above-mentioned defects.
 Cu ring Rebar Corrosion using FBEC R 279

FIG. 10. Blast rrofilc showing corrosion deposits ( x 200).

Fici. II . Blast profilc showing finc cracking of substratc ( x 200).

280 James A. Read

CROSS SECTION INTERFACE WITH

BEST FILM
THROUGH FILM STEEL SUBSTRATE

FIG. 12. Foam Evaluation Guide (voids/porosity).

Obviously eure of the coating must be effected and the only method of
satisfactorily checking this is by differential scanning calorimetry to ensure
that the powder has been taken through its glass transmission temperature.
This has been amply dealt with elsewhere. 7
Homogeneity and density ofthe coating are often missed and are of equal
importance to the satisfactory performance of the finished coating.
Voids within the film can affect its adhesion, elasticity and protective
qualities. In electrostatic and tribostatic spraying the speed at which the

FJG. 13. Cross-section through coating showing voids.

 Curing Rehar Corrosion using FBECR 281

rebar is passed through the spraying area and time allowed for the powder to
gel are critical.
A rating system was developed in America by Bell & Stephens
Laboratories ofHouston, Texas, called a Foam Evaluation Guide, to form a
basis for evaluation on a five-point system. Both the through film and
coating substrate interface foam level are given (foam = porosity/voids).
Figure 12 shows the Bell & Stephens Foam Evaluation Guide, produced by
courtesy of3M UK, and Fig. 13 shows an example ofthe voids in a coating.
It is suggested that this isafurther significant point affecting the quality of
coating and is worthy of consideration as an addition to the testing of
coatings.

UK APPLICATION USING FBECR

The first UK applications werein the early 1980s using FBECR imported
from North America in order to evaluate the material. Some were used on a
bridge on the Colwich loop road as a trial organised by Nottinghamshire
County Council and 42 t was used on a test section of concrete reinforced
pavement on the M18 organised by the DTp. This material was American
rebar coated in the USA.
The first major application was in 1987 for the 'Cardiff peripheral
distribution link road' built by South Glamorgan County Council. Two
hundred tonnes of FBECR were used in the parapets of a 1·5 km long dual
carriageway viaduct. FBE starter bars were cast into the precast deck
segments and then further FBECR was fixed to the starter bars once in situ
to provide a cast in-situ bridge parapet. This material was mainly 16 mm
diameter produced by Allied Steel and Wire shipped to the USA and coated
on Mid-West Pipelines coating plant with 3M powder.
In 1988 the first two railway /road bridges extensively using FBECR were
built. One carrying rail traffic over the new Wareester southern link road
spanning 15·0 m used 26 t of FBECR. The other allowing a dual carriageway
to be built under the London to Cardiff main line at Reading spanning
27·2 m used 130 t of FBECR. Both bridges used the FBECR in the soffit of
the deck and parapets, i.e. the areas of the bridge most liable to chloride
penetration from road traffic spray. The lower mat of reinforeerneut in the
decks, all links between upper and lower mats of reinforcement, all chairs
and spacers to support prestressing ducts, and all parapet steel used coated
bar. Diameters used went up to 40 mm for some ofthe straight bar with links
out of 10 mm. British Rail designed and specified both bridges, with the
coated bar being supplied by Allied Bar Coaters (ABC) ofCardiff and HD34
epoxy powder coating by International Paints plc. Theseare the first UK
bridges using all UK-produced coated bar (see Ref. 8 and Fig. 14).
282 James A. Read

FIG. 14. First UK bridge using all UK-produccd coated bar.

In 1989 seven box culverts were constructed to carry a road across the
River Itchen at Winchester; one of the seven has been built using FBECR,
allowing a direct comparison with other box culverts using uncoated rebar
in the same project. These were by Hampshire County Council with material
supplied by ABC. Nottinghamshire County Council specified epoxy-coated
dowel bars for concrete carriageway repairs, and during the summer of 1989
over 16 000 dowel bars were used in the new construction of expansion and
contraction joints on sections of the Al north road.
Two projects specified to use FBECR are:
(i) Dornie Bridge for the Scottish Development Department, a
replacement bridge near the Kyle of Lochalsh in western Scotland
crossing Loch Long wi th ten spans each of 26m. Work started in the
summer of 1989 and will use over 150 t of coated bar in the piers and
edge beams. FBECR was chosen because of the severe marine
environment and the poor durability of the existing structure as a
result of chloride contamination.
(ii) The Ness Viaduct for Scottish Railways, being built in 1989/90 to
replace a bridge washed away during flooding in the spring of 1989, is
using 30 t of coated bar.
In addition to road and bridge works, a number of other applications
where coated bar has been used in the UK are worthy of note, namely:
• The Craibstone-Dyce link road phase 3B has used 28 t of coated bar in
a structure to carry the new road on an embankment over the 38-in
 Curing Rehar Corrosion using FBECR 283

diameter oil pipeline from the Forties field to the Grangemouth

refinery. Grampian Regional Council chose this solution to protect the
steel in the structure from stray currents in the ground, which might
result because the pipeline is cathodically protected. If the structure
collapsed and the pipeline was damaged, the resulting environmental
and oil production Iosses could be catastrophic.
• Coated bar is also being used on electrical sub-station foundations
where there is a danger of inducing stray current corrosion in the
reinforcement.
• Two 22-m high reinforced concrete salt silos were built in 1988 at the
Weston PointSalt Works, Runcorn, Cheshire, for ICI Chemieals and
Polymers Ltd. U sed 70 t.
• Treated effluent outfall for the NWWA built in 1989 at Sandon Docks
in Liverpool. Used 50 t.
• Water storage tanks built in 1989 for the Central Scottish Water Board
at Blairlinnans, Scotland. Used 50 t.
• Sea defences for the Anglian Water Authority.

DISCUSSION

It is the author's personal view that unless and until we as engineers are able
to produce impermeable concrete that does not crack, and whilst chlorides
are used for deicing, then protection of the reinforeerneut at the surface of
the reinforeerneut will be necessary. This is in addition to any other measures
employed, including surface coatings and cement additives.
The sciencejart oftetrology, which has progressed since the days of Brunel
and the great engineers of the past, is demonstrably still unable to allow the
engineering and accounting professions to predict with any reasonable
expectation of accuracy the true maintenance costs of major structures. Yet
qualitative and financial decisions affecting the basic durability and
engineering designs continue to be used based on false premise.
Civil engineering construction still relies on labour-intensive operations
carried out by alllevels of skills, in all types of weather, and often against
heavy time and cost pressures.
It has always been necessary to employ safety margins and extra
protective measures, often in parallel, to give added assurance to the safety
and integrity of our structures.
Engineers learned many, many years ago that the best method of
preventing corrosion was to put the protection where the corrosion occurs.
This is why fusion-bonded epoxy-coated reinforeerneut should be our
answer to the current problem.
284 James A. Read

ACKNOWLEDGEMENTS

The author would like to express thanks to his company, W. S. Atkins

Consultants Limited, plus many outside organisations that have been
involved in the research into the production of FBECR and its testing which
have led to the comments and recommendations within this paper. He would
also like to express his special thanks to Mr N. Woodman, Allied Bar
Coaters; Mr B. Cruse and Dr P. Langford, 3M UK; Dr K. McLeod,
International Paint Powder Coatings; and Mr J. Wallace, NEI Research and
Development.

REFERENCES

1. Record repair planned for Midlands Links, New Civil Engineer (28 February
1989). Repair and Maintenance of the Midlands Links Viaduct. Working party
report 1988 by Department of Transport, W. S. Atkins & Partners and G.
Maunsell and Partners.
2. Coatings for concrete and cathodic protection of reinforcement, report of a
mission sponsored by the Institution of Civil Engineers and supported by the
Department of Trade and Industry held at ICE, London, 25 January 1989.
3. BABAEI, K. and HAWKINS, NEIL M., Evaluation of bridge deck protective
strategie~. Concrete International Design and Construction. J. American
Concrete Institute (December 1988).
4. WrLLIS, J., Epoxy-coated reinforcement in bridge decks. Transport and Road
Research Labaratory Supplementary Report 667, 1982. Department of the
Environment, Department of Transport.
5. READ, J. A., Examination of FBECR and ASTM A 775 for use in the UK and
Middle East. CIRIA/BSE Conference, Bahrain, October 1987. FBECR, The
Need for Correct Specification and Quality Control, University ofSheffield, May
1989.
6. BrsHOP, R. R., Application Guide 6: The specification of epoxy-coated
reinforeerneut bars. Department of Transport. Transport and Road Research
Laboratory, 1987.
7. MANLEY, T. R. and ScuRR, G., Thermalanalysis of epoxy anti-corrosive coatings.
In Coatings and Surface Treatment of Corrosive and Wear Resistance. Ellis
Horwood, London, 1984.
8. HoRSELER, J., British Rail Western Region, epoxy-coated reinforcement-A
designer's viewpoint. University of Sheffield, May 1989.
INSPECTION AND MONITORING
 25
Experiences with the First Generation of Prestressed
Concrete Bridges in Germany

BERNHARD GÖHLER
Leonhardt, Andrä und Partner, Lenzhalde 16, 7000 Stuftgart 1, FRG

ABSTRACT

Four major bridges built in the years 1950-1953 had to be inspected and
refurbished by the author. The bridges and their main problems are described.
Thin webs and slabs and sma!l construction depths were typical ol the designs.
The main dej"ects encountered were: uncompacted carbonated concrete,
ungrouted tendons and delective waterproolmembranes. Theseare presented
and compared for similar bridges. Conclusions are given regarding
requirementsfor maintenance, guidesfor inspection and lessonsfor the design
(~lnew bridges. The refitrbishments described give a generat indication ol the
'state ol the art' ol bridges in Germany.

DESCRIPTION OF THE BRIDGES AND THEIR MAIN

PROBLEMS

Bridge over the Danube at Ulm-Gänstorbrücke-built in 1950 (Fig. 1)

The bridge system is a frame with a deck span of 82·4 m. Ties anchor the
cantilever sections to the footings.
With a depth of 1·2 m at mid-span the deck is very slender. Remarkable
oscillations were registered by the pedestrians as the first trucks passed over
the new bridge, and discussions started on how lang such a bridge would
last.
Similar discussions have been held about temperature effects in such a
wide spanned continuous structure. The information gathered within the
last 40 years has proved encouraging, both in relation to oscillations and
temperature.
The concrete proved tobe excellent and the ultimate cube strength is now
287
288 Bernhard Göhler

ELEVATION
ano - - - - - -

CROSS- SECTIONS

• 2 2~ • \ 4 • 1 69 1 69 • , 40 • 1 1S -

FIG. I. Bridge over the Danube at Ulm, Gänstorbrückc, elevation and sections.

about 70 MPa. The Carbonation was only a few millimetres. In the deck
more than 30% ofthe ducts are ungrouted or partially grouted. The ducts in
the ties running down to the footings were filled with water up to the Ievel of
the Danube. One of the 26-mm diameterbarswas taken for testing and we
were pleased that the original designers had called for an extremely good
quality steel with an ultimate strength of 900 MPa. The test showed no
significant Iosses in strength but it was not possible to determine the
percentage of deterioration which had taken place as only the outer bars
could be inspected within limited areas.
The amount of reinforeerneut in the bridge is extremely small, as in all
bridges ofthat time, so there was no risk of a serious ultimate Iimit state
occurring. After some discussion it was decided to strengthen the ties but not
the deck as it could be monitared more easily and possible cracks would
indicate Iosses of Ioad capacity.

Bridge over the Main-channel at Bamberg, built in 1953 (Fig. 2)

The prestressing system used was the same as in the first bridge. The span
is 68·0 m but the cross-section of the deck is a twin box girder.
At first it was thought that similar defects would be found, but the ducts in
the ties were carefully grouted and there were only a few instances of
insufficient grouting in the deck. It was obvious that the bridgewas heavily
deteriorated. Rusted rebars and ducts could be seenindifferent places from
 First Generation Prestressed Bridges in Germany 289

ELEVATION

CROSS -SECTION

IC
115

225 .ilL

F1G. 2. Bridge over the Main-channel at Bamberg, elevation and section.

underneath, and water was dripping through top and bottom slabs. The
concrete was fairly sound but large parts of the bottom slabs showed
evidence of incomplete compaction.
Carbonation was up to 10 mm deep but in the uncompacted areas
carbonated zones were found up to about 150 mm deep. Several honeycomb
areas were found at the couplings of the bars on the bottom of the webs, as
shown in Fig. 3.
The waterproofing was defective and water mixed with deicing salt
dissolved the cement at the top of the deck slab. Pieces of wood left in the
slab were rotten and along these most of the water dripped into the box and
then through the bottom slab. In some parts ofthe surface the concrete had
changed into sand to a depth of 30 mm. F ortunately there was not much
deterioration at the bars as they were mostly arranged in the webs outside
the wet zones.
lf the defective waterproofing and uncompacted concrete zones had
allowed water to run along ungrouted ducts the result of the inspection
would almost certainly have been to recommend demolition.

Bridge over the Danube at Untermarchtal (Fig. 4)

This bridge, designed by Professor Leonhardt and built in 1952- 53, was
one ofthe first and for some time the Iongest continuous concrete girder with
a totallength of 375m. The end spans are 62 m long and the mid-spans are
290 Bernhard Göhler

FIG. 3. Bridge over the Main-channel at Bamberg, honeycomb zones.

70m long. The cross-section isadouble T-beam with a depth of 4·1 m and
therefore quite slender.
There were long discussions about how a continuous concrete girder
could withstand settlements. After nearly 40 years the bridge showed no
problems due to settlements or temperature but there were other problems.
First, it was obvious that something was wrong at the top of the deck.
There were wide cracks in the asphalt. The waterproofing was destroyed in
places and in other areas the upper 20-30mm of the concrete had
disintegrated. lt had changed into sand and parts of the reinforcement were

(oncrn trated Tenclon Tf:ndon Anchor ~ge Rtg1on

f - ::=::=zs
Web Strengthenong Addol oon al E.ternal Bonded endans
ELE VA TION OF END SPAN

Local Web
Strengthenong

CROSS - SECTJON

FIG. 4. Bridge over the Danube at Untermarchtal; strengthened end span.

 First Generation Prestressed Bridges in Germany 291

shining as if sandblasted, as the sand moved by the traffic was rubbing

against the steel.
After repairing the concrete surface and installing a new waterproof
membrane and asphalt layer the bridge was inspected carefully from
underneath. In the first span a long crackwas found on the inner side ofthe
web along the line of the cable. The crack had been injected some years ago
and had opened again about 1 mm.
By partially removing the concrete cover to the cable we found that on a
length of about 4-5 m the cable was only half grouted. The ungrouted
strands were heavily corroded- some of them broken- and corrosive
chemieals such as chlorides could be found on the surface of the steel.
Drainage ofwater from the deck was not possible and the excess water in
the grout had dissolved the chemieals from the graut and started corrosion
on the surface of the cables. About 50% of the ungrouted strands had
deteriorated and the darnage was so severe that the cable had tobe inspected
on its total length. At least the same defect in the same position and with
similar deterioration was found at the other web.

Bridge over the Neckar at Stuttgart, Rosensteinbrücke (Fig. 5)

This bridge, designed by Leonhardt in 1952, is quite similar to that of
Finsterwalder but the cross-section with longitudinal boxes, transverse
beams and again longitudinal beams looked more vulnerable. The concrete
quality again was excellent and only minor defects werein evidence, such as
water draining through the deck and some corrosion of bars and tendons.

ELEVAT ION

48

CROSS- SECTION

FIG. 5. Bridge over the Neckar at Stuttgart, system and section.

292 Bernhard Göhler

The main reason for this was the difficulty m fixing the tramrails
waterproofing system on to the thin deck slab.

COMPARISON OF THE DIFFERENT BRIDGESAND THEIR

MAIN DEFECTS

In Table 1 the problems found at the four bridges are compared,

supplemented by experiences with similar old bridges. Table 1 shows that the
most common problern was that of defective waterproof membranes.
Honeycomb concrete areas which carbonate after some decades are in
combination with ungrouted tendons. In some cases the resulting corrosion
was so severe that some smaller bridges had to be demolished.

TABLE 1

Bridge over river Main problems

Uncompacted Ungrouted Defective

carbonated corroded waterproofing
concrete tendons

Danube at Ulm Rare Often None

Main-channel at Bamberg Often Some Often
Danube at Untermarchtal Rare Some but one heavy Often
Neckar at Stuttgart Rare Some Often
Other similar bridges Some Some Often

This deadly combination of leaking membranes, porous concrete and

ungrouted tendons put bridges at severe risk. The smaller the tendons the
more often there will be ungrouted areas but corrosion will be slow.
Bigger tendons will have less ungrouted areas, but if there is one near the
concrete surface deterioration will be relatively fast. Through small cracks
wet air will be sucked inside the tube and continue the corrosion process.
With modern rehabilitation techniques it is not difficult torepair such local
defects, as shown in the next section.
lt is difficult to find and rehabilitate small ungrouted ducts, especially if
there are severallayers. Mostly they are used for transverse post-tensioning
of deck slabs and therefore additional safety in such parts of bridges is
recommended.
Figure 6 shows ungrouted transverse tendons in a cantilever slab, which
was cut off with a diamond saw.
Concerning the fear of carbonation, the experience with these bridges, and
 First Generation Prestressed Bridges in Germany 293

Ftü. 6. Ungrouted tendons in a broken-off bridge.

in general, is that carbonation in bridges is not a problern if the W/C ratio is

below 0-45 and compaction is done carefully.

REHABILITATION OF THE BRIDGES

The state of the art nowadays in Germany is to use an elastic waterproof

membrane glued to the concrete, either a bitumen or a polyurethane
membrane. Such membranes were installed on all four bridges. In three cases
it was a bitumen membrane and for the bridge at Untermarchtal it was
polyurethane.
Replacement of deteriorated concrete zones is weil known, but materials
based on epoxy resin are avoided as it is difficult to get a mixture which is
simultaneously dense and with similar properties to the parent concrete. We
prefer materials which are cement based, probably modified by poly-
carbonates.
Ungrouted tendons are a special problern of prestressed concrete
demonstrated on all four bridges. The first aim is to keep moisture away
from the tendons. Waterproofing on the top is vital, and possibly coating of
vertical surfaces. Grouting of all hollow spaces is also vital to avoid the
sucking of moisture inside the ducts by temperature effects.
For the two bridges designed by Finsterwalder tensioned with the 26-mm
diameter bars it was not possible to grout all the small empty ducts. The
bridge over the Main therefore had to be coated outside.
For the ties ofthe Danube bridge at Ulm grouting ofthe empty ducts was
attempted and additionalload capacitywas installed by assembling ground
294 Bernhard Göhler

anchors beneath and parallel to the ties drilled from the deck and through
the spread footing into the rock.
Typical rehabilitation of corroded tendons is illustrated in the Danube
bridge at UntermarchtaL The remaining Ioad capacity of the concentrated
cable had to be calculated at 75% of the original. Inside each web three
additional cables were installed, each with 12 no. 12-mm diameter wires.
They were assembled inside an additional thickening of the web (see Fig. 5)
and anchored at additional corbels at both ends of the span.
Rehars grouted with epoxy resin into drilled holes transfer .shear forces
from the new concrete to the old. At the corbels additional post-tensioned
26-mm diameter bars transfer the anchor forces to the webs.
To avoid bursting due to the corroded and broken strands additional
heavy reinforced concrete thickenings were assembled at the defective zone.
Summing up, it can be said that all rehabilitations were successful and it
can be expected that all bridges will survive the next four decades.
 26
Movable Bridge Machinery Inspection and
Rehabilitation

CHARLES BIRNSTIEL
Consulting Engineer, Forest Hills, New York 11375, USA

ABSTRACT

Age and marine environment, exacerbated by inadequate maintenance, have

deteriorated movable bridge machinery and controls. The scope of six Ievels of
.field inspections pe~formed to quantify and document the deterioration is
described. For one of these, an intermediate Ievel inspection recently adopted
for the biennial bridge inspection program of New York State, the work items
are outlined in detail. A discussion of the current approach to machinery and
cantrot rehabilitation in New York State concludes the paper.

INTRODUCTION

Approximately 1700 movable bridges span navigable channels in the

continental United States. Ofthese 1100 are highway bridges and 600 carry
railroad traffic. Eighty percent are at least 40 years old. Within New York
State there are 60 movable highway bridges, ofwhich over 70% are 40 years
or older.
lt is seldom that the agency that caused a bridge to be built is still fully
responsible for that bridge. Many of the older movable highway bridges
were built as privately-owned toll facilities and after some years under
private operation were acquired by a state or other governmental agency.
Railroad bridges also changed ownership as a result of railroad company
mergers and through acquisition by states. In some cases the states took over
the railroad passenger service and the property. In other cases the states
acquired the bridges but other agencies operate the trains running over
them.
295
296 Charles Birnstiel

The responsibilities oflocal governments for the older movable road and
street bridges has also changed. A bridge built, rehabilitated and maintained
by a city for many years may have been transferred to a state highway
system, in which case the responsibility for operation and maintenance now
lies with the city but rehabilitation would be by the state, with federal
assistance.
The machinery of many of the older movable street bridges in the
northeast United States has deteriorated at an accelerated rate since 1950 for
two reasons. First, the replacement of solid timher decking by open steel
grating in the 1950s results in more debris accumulating on the machinery.
In winter the debris often contains deicing salts. Moisture falling through the
open grating is retained in the accumulated roadway debris and promotes
corrosion of the machinery. The second reason for the accelerated
deterioration of movable bridge machinery in the northeast was the
reduction in maintenance forces two decades ago. The social upheavals in
the 1960s created conditions in which budgets for maintenance of bridges
and other transportation facilities had to be drastically reduced in order to
make funds available to mitigate social unrest. lt is only recently that
maintenance forces are being rebuilt.
Before describing the scopes of machinery and electrical inspections, and
discussing the rehabilitation of bridge machinery, the different types of
movable bridges and the kinds of mechanical and electrical equipment
found on these bridges will be reviewed.

MOVABLE BRIDGE TYPES, MACHINERY AND CONTROLS

Types of Movable Bridges

Movable bridges may be broadly categorized as swing, vertical Iift,
bascule and retractile, although some bridges do not fit into these categories.
In the United States there are about 750 swing, 230 verticallift, 720 bascule
and five retractile bridges. Construction features of each type are described
subsequently.

Swing bridges
In swing bridges the movable span, often termed the draw, rotates about a
vertical axis. The type may be subdivided as to the manner of draw support
when swung open (permitting navigation). If the dead Ioad (self-weight) is
supported on a pivot bearing at the axis of rotation it is termed 'center
bearing'. The draw is balanced on this pivot. To keep the draw from tipping
under unbalanced Ioads, such as wind, balance wheels are provided that roll
on a circular track concentric with the pivot bearing.
Swingbridges when open, in which the dead Ioad is supported by a nest of
 Movable Bridge Machinery Inspection and Rehabilitation 297

tapered wheels, are said to be rim bearing. In these bridges the superstructure
is supported by a circular girder called a drum girder. A tapered plate (tread
plate) is fastened to the underside of the drum girder. It bears on a nest of
tapered wheels whose axes are oriented radially to the axis of span rotation.
The wheels, in turn, roll on a tapered plate, called a track plate, that is
fastened to a chair casting. The drum girder and the nest of tapered rollers
are all held concentric by radial members connected to bearings rotating
about a central pivot post.
In the closed position, the ends ofthe main longitudinal bending members
(usually trusses) are lifted at the rest piers so that there will be an upward
reaction at the truss ends for all combinations oflive load and temperature.

Vertical Iift bridge

The movable span of a vertical lift bridge moves vertically upward to
provide clearance for passage of vessels. The ends of the lift span are
attached to wire ropes passing over sheaves at the tops ofthe towers with the
far ends connected to counterweights. The vertical lift type is subdivided
into tower, span and tower-span drive.
In tower drive lift bridges there is drive machinery in each tower
connected to the counterweight sheaves. The forces necessary to raise the
span are transmitted to the counterweight ropes by friction. The action is
identical to that of a traction drive office building elevator.
Span drive lift bridges have the machinery located on the lift span itself,
usually above the roadway or tracks. There is one drive and the span is
moved by a wire rope drive which hauls the lift span up or down.
Alternatively, shafts extend longitudinally from the spandrive machinery to
the towers, where pinions engage a rack and the span crawls up or down the
towers.
In tower-span drive lift bridges there is also only one drive. It is fixed on a
structure spanning between the towers. Usually this drive is coupled to the
counterweight sheaves and the action is that of a friction drive.

Bascule bridges
There are two principal categories of bascules: rolling lift bascules and
trunnion bascules. Rolling lift bascules have the distinguishing feature that
the ends of the main spanning members (bascule girders) are cylindrically
curved and the movable span (leaf) rolls on these curved surfaces during
opening and closing. As the curved ends ofthe girders roll shareward the leaf
tilts open to clear the channel. The leaf simultaneously rotates and
translates. To close the bridge the leaf rolls toward the channel. This bridge
type was developed and promoted by the Scherzer brothers in Chicago at the
end of the 19th century.
298 Charles Bimstief

The other, older, type ofbascule is the trunnion bascule. Trunnion bascule
leaves rotate about large shafts (trunnions) that are usually inserted through
the webs ofthe bascule girders. The shafts may rotate with the girders, or the
bascule girders may rotate about fixed trunnions. The leaves of trunnion
bascules only rotate, they do not translate.

Retractile bridges
Thesebridges roll horizontally on tracks in order to clear a waterway. The
movable spans of retractile bridges only translate, they do not rotate. The
axes of the roadway and the navigation channel usually intersect at about
45° so that the movable leaf can be rolled clear ofthe channel without having
to be rolled backwards over or under the approach.

Span Drive Machinery

The movable span can be driven by either a mechanical or a hydraulic
drive. In mechanical drives reduction gearing is used to convert the high-
speed, low-torque output of electrical motors to the low-speed, high-torque
output required to drive the movable span. Two approaches are used for
hydraulic operation. In one, hydraulic motors produce the high torque
necessary to drive the movable span. An electric motor drives a hydraulic
pump that supplies the high-torque hydraulic motor. This eliminates the
need for all but the last or the last two stages of reduction gearing. In the
other approach, the movable span is driven by the movement of pistons in
the cylinders. Suitable articulation converts the linear action of the cylinder
to motion rotating the movable span. Wire ropes are sometimes used to
transmit drive forces in both mechanical and hydraulic systems.

Stabilizing Machinery
The machinery which stabilizes the movable span in the closed and open
positions varies with the bridge type. For center bearing swing spans it
comprises end Iifts at the rest piers and center wedges at the pivot pier. These
are used to Iift the draw so that the trusses act as continuous spanning
members when the draw is closed, from rest pier to rest pier with an
intermediate support. Rim bearing swingspansalso have end Iifts but do not
have center wedges. Both center and rim bearing swing bridges have
centering latches or centering mechanisms to aid in positioning the draw in
the closed position and prevent it from rotating while the end Iifts are
extended or retracted.
Vertical Iift bridges have span locks to lock the Iift span in the closed
(lowered) position, guide wheels to guide the Iift span as it rolls up the tower
and, sometimes, buffers to decelerate the moving span as it approaches the
upper and lower Iimits of travel.
 Movable Bridge Machinery Inspection and Rehabilitation 299

The curved treads fastened to the bottom of the bascule girders of

Scherzer bascules and the tracks on which they roll are usually considered
stabilizing machinery. Other stabilizing machinery of bascules, both rolling
Iift and trunnion, are midspan Jocks, tail locks and live Ioad reactions.

Electrical Power Equipment and Controls

The electrical equipment may be considered power equipment and
control equipment, although it is sometimes difficult to distinguish between
the power and control functions. Electrical power equipment comprises
transformers, circuit breakers, motors, overload relays, resistor banks,
thyristors, electrically-released brakes, panelboards, electromagnetic
switches, conduit, boxes, submarine cables and wiring. The control desk,
control switches, pushbuttons, indicating lights, meters, Iimit switches,
navigation lights and air horns are considered control devices. There is
obvious overlap of function, as for example with submarine cables which
may contain control wires in addition to power wires.
Most movable bridges are powered by electric motors that are either
directly coupled to gearing in order to drive a mechanism or coupled to
hydraulic pumps which perform the same basic function, that of multiplying
the torque output of the motor at the expense of speed. Few bridges are now
powered by internal combustion engines or are hand-operated. Alternating
current drive systems are most common, either as original equipment on
newer bridges or as part of rehabilitation of older bridges. AC drive motors
are usually of the wound-rotor type to permit the connection of external
secondary resistance so as to obtain the desired speed-torque motor
characteristics and, often, for speed control. In older AC systems speed is
usually varied stepwise by switching resistance into or out of the secondary
circuit using a drum switch and relays. In the state-of-the-art controls the
secondary resistance is fixed in order to obtain the desired motor
characteristic and speed is controlled by varying the input power to the
motor by means of thyristors responding to a feedback signal.
Span drives of some ]arger Iift bridges and a few other smaller movable
bridges are powered by direct current. The direct current is usually supplied
from a motor-generator set, the motor ofwhich is usually an AC motor. DC
drives were installed in the !arger Iift bridges in order to take advantage of
the superior speed control capability provided by the DC generator-DC
drive motor arrangement. This was the only practical means of stepless
motor speed control commercially available before the development of the
thyristor.
The motors of stabilizing machinery are usually AC squirrel cage motors
because speed control is not needed. The operation of these motors is
interlocked with the traffic control devices and the span drive by means of
Iimit switches.
300 Charles Birnstiel

Traffic Control Devices

The traffic control devices on highway bridges comprise traffic signals,
warning gates, warning gongs, resistance gates and warning signs. Warning
gates are usually of the semaphore type, similar to electrically-operated
railroad crossing gates. Resistance gates are intended to absorb impact of a
moving vehicle which has passed beyond the lowered warning gates. There
are three varieties of resistance gates: pintel gates, semaphore gates or
vertical Iift. Modes of operation also vary: hand, direct-geared electric or
hydraulic.

INSPECTIONS

Types of Inspections
The machinery and controls of movable bridges are inspected in
connection with bridge rehabilitation projects, to satisfy insurance
requirements, and as part of the biennial bridge inspection program
mandated by the US Department of Transportation. The thoroughness of
the inspection varies, depending on its purpose. Wehave classified them as
Types I, 1-R, II, 11-A and 111.

Type I
A cursory visual inspection of the drive and stabilizing machinery, the
traffic control devices and the electrical controls for this equipment. No
parts hidden by guards or housings that require tools for removal are
inspected. No mechanical or electrical measurements are made.

Type I-R
This is a visual reinspection along the lines of the Type I inspection. lt is
made some years after a higher-order inspection and has two main
objectives. First, to ascertain the condition of components or systems found
to be deficient in the prior inspection. Second, to discover other defects
which may have developed since the last inspection.

Type II
A visual inspection during which access covers are opened for the
inspection ofhidden components. An electrical performance test is made for
at least one opening/closing cycle, during which the power consumed by
each normal span drive motor is measured and recorded on a strip chart.

Type II-A
An intermediate inspection which is essentially a Type II but includes
 Movable Bridge Machinery Inspection and Rehabilitation 301

more measurements. lt is in accordance with the intent of the Bridge

lnspector's Manua!j(Jr Moveable Bridges published by the Federal Highway
Administration (FHWA), hereinafter termed the Manual. The scope of the
Type II -A inspection conforms essentially with the Manual but the extent of
machinery disassembly and measurements is less and the field data are
presented in a condensed form. On the other hand, electrical measurements
that arenot referred to in the Manual, but which have been found valuable in
assessing movable span behavior, are included.

Type///
An in-depth inspection in which virtually every mechanical and electrical
component is inspected visually and measurements are made to determine
wear of mechanical parts, and some electrical equipment is tested. lt is an
expansion of the Type 11-A inspection to meet the recommendations of the
Manual, especially as to reporting field data.
The intermediate inspection, Type Il-A, is being considered by New York
State as a standard for biennial inspections. The scope of this inspection is
described in more detail subsequently.

Scope of Type 11-A Inspection and Report

The Type II-A inspection is designed to conform to the intent of the
FHWA Manual but the amount of machinery disassembly is less than shown
in the Manual and the field data are presented in a more concise format. The
work items included in the scope are described subsequently.

Preparation
Prior to actual field inspection the team Ieader visits the bridge in order to
assess inspection difficulties. The available mechanical and electrical
drawings are studied in planning the inspection and to determine the
probable original tooth thicknesses of all open gearing.

Bridge operation
Bridgeoperation is observed in the normal mode so as to assess condition
of the normal mode drive, functioning and effectiveness of traffic control
devices, interferences between movable and stationary parts of the bridge,
controllability of the moving span and the effectiveness of stabilizing
machinery. During bridge operation observers monitor the machinery for
abnormal noises and vibration. The manner in which the bridge operator on
duty handles the controls is also observed.

Performance tests
The power consumed by the normalspandrive motors is measured and
302 Charles Hirnstiel

recorded Oll a strip chart durillg at least Olle opellillg/closillg cycle for each
movable spall. The object of the test is to detect excessive self-weight
imbalallce of bascule alld vertical Iift spalls or severe billdillg of the
machillery. Electrical colltrol deficiellcies are sometimes foulld from these
tests.

Span drive machinery

Every movillg alld statiollary component of the span drive is visually
illspected. The types of components inspected depend Oll the type of
movable bridge but, generally, the illspection includes brakes, speed
reducers, open gearing, drive shafts and couplillgs, bearings, mountillg bolts,
sheaves alld wire ropes. The following measuremellts are made: gear tooth
thickllesses, gearset backlash and clearance, and sleeve bearing clearallces.
Shafts are visually inspected for cracking at keyways alld shoulders, alld
for movement betweell coupling hubs and shafts.
The mechanicallinkages of brakes are inspected, as are the brake wheels.
Clearallces between shoes and wheels are measured with the brakes set and
halld released.
Sampies oflubricatillg oil are recovered from speed reducers and sent to a
laboratory for chemical allalysis to determine the amoullt and sources of
colltaminants.

Stabilizing machinery
The mechanical components that stabilize the movable span when it is in
motion alld at rest are inspected. Dependillg on the type ofmovable bridge,
these include treads and tracks, spall locks alld drives, cellterillg devices,
buffers, live Ioad supports, trunlliolls, wheels and axles, end Iifts, span guides,
alld wire ropes and adjusting devices.
The coullterweight sheave trunniolls of verticallift bridges are inspected
visually and ultrasonically. Ultrasonic testing is performed by a recognized
testing laboratory experienced in this work.
Aspart ofthe machinery inspectioll movements at midspan locks and tail
locks alld live Ioad shoe clearances are measured.

Electrical system
The major electrical components are visually inspected on the Ioad side of
the service discollnect. Span drive motors and motor-generators are
observed while running to check bearings alld for excessive noise and
vibration. Brushes, commutators and slip rings are inspected. The
installation and components of contactors, circuit breakers and drum
switches are examined. The brakes are inspected for thrustorjsolenoid
operation, Iimit switch operation and deterioration ofwiring. The extellsion
 Movahle Bridge Machinery Inspeerion and Rehabilitation 303

of springs of spring-set thrustor brakes is measured. Span lock motors, end

Iift motors, navigation lights and traffic control devices are inspected and
their operation observed.
Operability of accessible Iimit switches is determined and interlock Iimit
switches are subjected to a simulated test where appropriate.
The exposed portion of the submarine cable is inspected for corroded
armor wires and other damage. The electrical insulation resistance of
submarine cable conductors is measured.
The control desk is examined to assess operability of control functions
and safety with respect to misoperation and unauthorized use.

Tra!Jic control devices

Traffic lights, warning gates and resistance gates are inspected for
structural condition, mechanical operation and effectiveness in stopping
vehicular traffic. Layout of the traffic devices is recorded for comparison
with the relevant Manual of Unif'orm Tra.ffic Contro! Devices.
The field measurements and condition observations are recorded on field
inspection forms which were developed for this purpose and are included in
an appendix to the report. Where appropriate deficiencies are photographed.
The typical report contains the following:

• Descriptions of the existing mechanical, electrical and traffic control

systems illustrated with a schematic span drive machinery diagram, a
stabilizing machinery Iayout and a traffic control device Iayout.
• Photographs of the general arrangement of machinery, electrical
equipment and traffic control devices, and of defects.
• Reduced size photocopies of annotated pcrformance test strip charts.
• Discussion of the condition of the mechanical and electrical
components emphasizing defects and corrective measures.
• Recommendations for in-depth inspections ofinadequate components
and a Iisting of defects and items that require attention.
• Evaluation of the traffic control devices with respect to the relevant
Manual of' Unij(mn Traffic Contro! Devices.
• Ratings of the major components of the mechanical and electrical
systems and the tram.c control devices according to the state numerical
condition rating system. The ratings are based solely on visual
observations and measurements. No numerieil analyses are made.

The Type 11-A inspections were developed to satisfy the requirements of

the federal biennial bridge inspection program. They are primarily intended
to discover defects that may compromise the safety ofthe bridge. Dangeraus
conditions are "ftagged' in writing to the state. However, local agencies are
304 Charles Birnstiel

using these reports as the basis for improvements in maintenance and for
minor rehabilitation programs.

REHABILITATION

Many of the older movable bridges have been rehabilitated to some degree.
Usually the electrical systems have been revised. We found 90-year-old
bridges with the basicoriginal spandrive and stabilizing system in place but
with successive changes to the motive power-from reciprocating steam
engines to direct current, to two-phase alternating current and then to three-
phase alternating current. Wehave inspected some fifty different movable
bridges from 5 to 100 years old and not found two alike, even though some
were originally constructed from a single set of plans.
The usual criterion for movable bridge machinery and control
rehabilitation in the northeast United States is to design for a 30-year
extension ofbridge life. For the older bridges the span drives may have tobe
completely replaced. For bridges less than 50 years old the original span
drive is usually retained and specific components found to be worn or
defective are replaced. For bridges whose electrical systems have not been
rehabilitated in the last 40 years it is usual to replace the complete electrical
system with a thyristor-controlled alternating current system. The
unavailability of replacements for electrical components older than 40 years
and the fact that the older panelboards have live fronts, which are now
considered dangerous, play apart in such decisions. Ifthe traffic controls are
more than 30 years old they are usually replaced because they are often
damaged and deteriorated, and because the existing traffic control device
Iayouts ofthat age seldom conform to the current State DOT requirements.

CONCLUSION

The major types of movable bridges and their mechanical and electrical
components were described. Because of aging, repeated stressing from traffic
and the deleterious effect of marine environments, coupled with inadequate
maintenance, the machinery deteriorates. In order to identify dangerous
conditions and quantify the deterioration for repair and rehabilitation
programs, field inspection and reporting procedures at various Ievels of
thoroughness have been developed. Six of these were briefly described and
one, the Type II-A, was outlined in detail. The report contents forthistype of
inspection was also included. Finally, current approaches to machinery and
control rehabilitation in New York State were described.
 27
Application of Radar and Thermography to Bridge
Deck Condition Surveys

D. G. MANNING and T. MASLIWEC

Research and Development Branch, Ontario Ministry of Transportation,

1201 Wilson Avenue, Dmvnsview, Ontario, Canada, M3M JJ8

ABSTRACT

This paper summarizes the theoretical and practical aspects ofthe application
of radar and thermography to bridge deck condition surveys. A prolotype
vehicle was developed. The radar wavej(Jrms and the thermographic images
were calibrated in the{ield to variousforms ofphysical distress. lt is shmvn that
radar and thermography are usefid toolsj(Jr assessing the condition of bridge
decks and that the tH·o technologies are complementary; thermography is most
usefulfor locating suhsurface defccts on exposed concrete swfaces and radar is
most applicable to detecting defccts in concrete deck s/ahs which have a
hituminous surfacing.

INTRODUCTION

An effective bridge management system requires a comprehensive data base.

At the network Ievel, the data are needed in order to assess the overall
condition of the bridge network and to predict the effect of alternative
preservation and improvement actions. At the project Ievel, more detailed
data are required to select the most appropriate method of rehabilitating a
particular structure and to prepare the contract documents.
Collecting data on the condition of bridges by traditional methods is
expensive and the results are not always dependable. In the late 1970s,
research studies were initiated by the Ontario Ministry ofTransportation to
develop rapid and reliable methods for collecting data on the condition of
bridge decks. The studies were part of a !arger program to improve the
technology for rehabilitating bridges. 1
305
306 D. G. Manning and T. Masliwec

The first studies were conducted on exposed concrete decks and were
designed specifically to investigate rnethods of detecting delarnination.
Infrared therrnography was found tobe capable ofidentifying delarninated
areas over a wide range of arnbient ternperatures. 2 During the period
1980~82, the studies were extended to investigate rnethods capable of
detecting deterioration in concrete deck slabs which have a surfacing of
biturninous concrete. Over 90% of the bridges in Ontario now have a
waterproofing rnernbrane and two 40 rnrn thick Iifts of biturninous
surfacing. A test site was constructed by selecting a typical bridge which was
exhibiting corrosion-induced distress, surveying the bridge and then paving
it without first rnaking repairs. Areas of scaling and debonding were
sirnulated prior to paving. The capabilities of eight techniques to detect the
locations and types of deterioration in the test deck were evaluated. The
rnost prornising techniques were found to be radar and therrnography. 3
Further research concentrated on developing autornated processing
techniques to analyse the therrnograrns and radar waveforrns and present
the results in a form usable by bridge engineers. The product of this work
was a self-contained vehicle narned DART (deck assessrnent by radar and
therrnography) which, as the acronyrn irnplies, is equipped with both radar
and therrnography. 4

MAJOR TYPES OF DETERIORATION IN BRIDGE DECKS

The rnost serious form of deterioration in bridge decks in North Arnerica is
that caused by corrosion of ernbedded reinforeerneut As the reinforeerneut
steel corrodes, it expands and creates a crack or subsurface fracture plane in
the concrete at, or just above, the Ievel of the reinforeerneut The fracture
plane, or delarnination, rnay be localized or rnay extend over a substantial
area.
Scaling, which isabreakdown ofthe cernent~paste rnatrix, is also a serious
problern wherever it occurs. The disintegration of the concrete, which is
caused by the freezing of concrete critically saturated with water, begins at
the surface and gradually progresses so that the full depth of deck slab rnay
be affected. It is rnost cornrnon in older bridges which do not have an effective
air void systern.
On asphalt-covered decks, bond failure rnay occur between the concrete
deck and biturninous surfacing. Debonding can result in maisture being
trapped on the surface of the concrete and, where thin surfacings are used,
can Iead to failure of the biturninous surfacing. Although the consequences
of debonding are not as serious as either delarnination or scaling, it can be
confused with these two phenornena in surveys. Consequently, it is
irnportant to be able to identify and define debonded areas.
 Bridge Deck Condition Surveys 307

HOW THERMOGRAPHY AND RADAR DETECT

DETERIORATION

Thermography
Infrared thermography is a remote method of sensing the energy emitted
from the surface of an object. Consequently, the detection of deterioration
by infrared thermography is based on the difference in surface temperature
that exists between deteriorated and sound concrete under certain
atmospheric conditions. Concrete is a poor conductor of heat and the
differences in surface temperature develop on a bridge deck as a result of
different rates of heat transfer occurring inward from the surface.
Discontinuities which impede the conductive heat flow from the surface,
such as a delamination or debonding, will result in a higher surface
temperature during periods ofheating. The reverse is true during periods of
cooling. Because of the breakdown of the matrix, scaled concrete has a
different coefficient ofthermal conductivity than sound concrete, which also
results in a difference in surface temperature.
The differences in temperature have been confirmed by implanting
tliermocouples in deck slabs and by the direct measurement of surface
temperatures over known areas of so und and deteriorated concrete. 2 •3
Measured temperature differences were also compared with a theoretical
heat transfermodelas a means of establishing operational constraints on the
use of thermography for bridge deck surveys. An approximate solution of
the general equation for transient heat flow 5 indicated that the most critical
parameters affecting temperature differentials are solar irradiance and
emissivity. However, wind speed is also critical to the theoretical calculation
of emissivity.
Deiaminations appear as well defined white areas in an infrared
thermogram which are hotter (during daylight hours) than the surrounding
areas of solid deck which are cooler and appear dark.
Scaling of concrete beneath a bituminous surfacing produces a
characteristic thermal image which is a mottled grey-white tone. The more
severe and extensive the scaling, the more mottled the thermal image
appears.
Debonding is not easily identified by infrared thermography. Depending
upon the depth of any delaminations, debonding may appear hotter or
identical to the delaminations and, at other times, may go undetected.

Radar
The detection of deterioration by radar is based on reflections of a high
frequency electromagnetic wave caused by changes in the electromagnetic
properties of the material being probed. Whenever a transmitted wave
308 D. G. Manning and T. Masliwec

encounters an electromagnetic discontinuity (deterioration) or change of

material dielectric, it is partially reflected. The patterns created by the
reflected waves are received by the radar antenna. The radar receiver takes a
measurement of the time required for the transmitted pulse to travel to a
target discontinuity and for the reflected pulse to return. 4 The time delay of
the radar echo from a structural discontinuity is directly related to the depth
of the fault.
The im pulse radar used in the DART system has been described in detail
elsewhere. 6 The radar is monostatic with the antenna design being based on
a constant flare angle, variable-width open horn. The transmitted signal is a
pulse which has a duration of approximately 1 ns and a repetition rate of 5
million pulses per second.
The reflected signal is sampled using the sliding gate technique, and
averaged, producing an output voltage pulse waveform of approximately
1ms duration. The actual radar waveform reflected from a metal plate on
the deck surface is shown in Fig. 1. The signal has a duration of
approximately 0·9 ns between negative peaks P and Q with the main positive
peak at A. The peak at A represents the surface reflection and is used to
calibrate the system. The transients at the beginning and end of the main
pulse restriet the resolution in measuring the thickness of a material. The
response after point R is very small, indicating the variations due to
discontinuities can be detected in this region without too much interference.
Figure 2 represents, from a theoretical standpoint, the principal
reflections which occur in a bridge deck. The peak at A represents the
reflection from asphalt, the peak at C is produced by the asphalt-concrete
interface and the reflection at D is from the rebars. Although the reflections
are shown as being quite distinct, there may be considerable overlap of the
transient portians ofthe waveforms such as at point B. However, this is not a
serious problern provided that the main peaks are weil defined. The
thickness of the asphalt is determined by measuring the time between point
0.5
A

"""'0>
ai
"0
:::1
reflected signal
0.
·"'c. A
E
<

p
-.5
0 10
Time, nanoseconds characteristic W

FIG. 1. Waveform reftectcd from a mctal FIG. 2. Theoretical reftection from a sound
platc on the deck surface. bridge deck.
 Bridge Deck Condition Surveys 309

.20 , - - - - - - - - - - - - - - - - - - ,
peak from asphall surface

peak from asphalt-concrete interface

-.20 OL__:__ _____._ _ _ _--:':10,-----------'

Time, nanoseconds

FrG. 3. Actual reflection from a sound bridge deck.

A and point C, taking account of the dielectric constant of asphalt which is

calculated from field measurements. 7 The reflections from the asphalt-
concrete interface and the rebars produce a T·V-shaped waveform which is
called the characteristic W This characteristic W is useful in identifying the
difference between sound and delaminated concrete.
The radar reflection from a sound bridge deck is illustrated in Fig. 3. The
peak at A represents the surface reftection from the asphalt, the peak at C
indicates the reftection from the asphalt-concrete interface and the peak at
D is the reflection from the rebars.
For analysis, two ratios are defined, the first being Rl, the ratio of the
magnitude at point B to the magnitude at point A, where pointBis defined
tobe the firstnegative peak after peak A. The second ratio, R2, is the ratio of
the magnitude at point C to the magnitude at point A, where point C is
defined to be the first positive peak after point B. These two ratios have
different uses.

Debonding
Since debonding affects the interface between the asphalt and the
concrete, the reflection from the asphalt surface will remain essentially
unchanged. However, the reftection from the asphalt-concrete interfacewill
be altered, producing a reflected waveform with either increased or
decreased magnitude at point C depending on whether the gap is filled with
water or air. This gap will effectively change the R 1 and R2 ratios of the
radar waveform.

Scaling
Scaling breaksdown the concrete matrix which may contain either air or
water. In this case, the gap is much larger than occurs in debonding. For the
purposes ofmodelling, it is assumed that the reflection from the gap is due to
a single reflection from the asphalt-gap interface. With air in the gap, the
310 D. G. Manning and T. Masliwec

reflection not only changes polarity, compared with a good asphalt-concrete

interface, but the magnitude of the Rl ratio is increased substantially. With
water in the gap, the polarity remains the same, but the R2 ratio is much
larger. Since it is not known whether the scaled concrete contains air or
water, both the Rl and R2 ratios are calculated and a search made for values
larger than normal.

Deiamination
Deiamination occurs in the region between the surface of the concrete and
the top mat of steel. There may be one or more fracture planes in the concrete
containing air or water, resulting in at least one discontinuity between the
concrete surface and the rebars. The additional reflections from the
delamination mean that the characteristic W present in a good structure is
no Ionger present.
The range of values used for evaluation are as follows:
Rl R2
Good <0-4 <0·28
Debonded <0-4 >0·28
Scaled >0-4 >0·28

THE DART PROTOTYPE VEHICLE

The exterior and interior of the DART vehicle are shown in Figs 4 and 5
respectively. The essential features of the unit are as follows:
(i) A hydraulically-operated telescopic mast mounted on the front of
the vehicle. The mast is lowered for attaching the infrared scanner
and refilling the liquid nitrogen. It is raised to its operating height of
Sm.
(ii) Theinfrared scanner. When equipped with a 20° lens and operated at
an angle of 45o and height of 5 m, the scanner is capable of viewing
the width of one complete traffic lane at a single pass. The scanner is
connected through a control unit and filter to a highresolutionvideo
tape recorder.
(iii) The radar antenna is attached to a rail on the front of the vehicle
using a sliding bracket so that it can be positioned anywhere within
the width of the vehicle. The mounting height of 150 mm results in an
elliptical footprint of 300 x 175 mm on the deck surface. The
transmitter and receiver are also mounted on the sliding bracket.
Returning radar echoes are received by the antenna and transmitted
to the control unit inside the vehicle. The waveform is simultaneously
 Bridge Deck Condition Surveys 311

FIG. 4. Exterior of DART vehicle.

FIG. 5. Interior of DART vehicle.

displayed on an oscilloscope and recorded on a seven-channel FM

tape recorder. The operator can monitor the signal from the control
unit and from the tape recorder, thus permitting a check of the
q uality of the signal before and after recording.
(iv) A fifth wheel is attached to the rear of the vehicle for speed and
distance measurements. Custom interface devices enable distance
pulses to be stored on both the video and FM recorders.
(v) A microcomputer for processing the data.
(vi) The vehicle is equipped with heavy duty air conditioning, a generator
to supply electrical power and racks with vibration-isolation mounts
for the electronic equipment.

A schematic of the components of the thermography and the radar is

given in Figs 6 and 7 respectively.
312 D. G. Manning and T. Masliwec

To Audio Channel1

High
Personal Resolution
Computer 1---.-t Colour Monitor

FIG. 6. Infrared system schematic. FIG. 7. Radar system schematic.

Data Acquisition/Processing
Radar
A central problern in the system software development is in reducing the
volume of radar data while preserving all desired information. Data
reduction in the data acquisition programs is accomplished in two different
ways. First, for all waveforms only 10 ms of data before and after the surface
reflection is considered relevant. After internal noise subtraction, only the
second 10 ms of data are permitted to enter data processing subroutines. The
second 10-ms slot is also used to compute the average Ievel ofthe trace and
adjust the trace with respect to zero. Secondly, for data that are processed
using fifth wheel information, the 10-cm fifth wheel waveform is used to
ensure that waveforms are sampled only every lücm.
The program MT088, written in standard Fortran 77language, performs
all processing functions such as calculating the depth of asphalt, determining
the R1/R2 ratios for scaling and debonding, and providing the zero crossing
count for delamination detection. The software is designed to handle
analysis for a bridgedeck of any length with 12 grid lines. The processed
results are stored as formatted data files for future reference and are also
output, in tabular form, to the monitor screen and the line printer.

Thermography
For processing the data, the video record is fed into an on-board
computer through an image digitizer interface. Using custom software, the
 Bridge Deck Condition Surveys 313

infrared image is sampled in such a way that the oblique angle of view and
other distortions of the infrared image are eliminated. Deterioration is
identified through a combination of computer- and operator-assisted
machine interpretation. 5 The computer produces a scaled graphics image of
each lane of the deck, showing areas of delamination and scaling.

FIELD OPERATIONSAND EXPERIENCE

The DART vehicle is normally operated between May and November, with
the infrared scanner being rented only for the months of June, July and
August. However, the scanner has also been used in February to investigate
its application in cold weather.
The vehicle is operated at a speed of between 3 and 4·5 km/h, making
traffic protection necessary. At faster speeds the digitized version of the
infrared image is distorted and the radar surface coverage capability is
impeded.
Whereas one pass with the equipment permits an infrared scan of one
complete lane width, the radar collects information only along a grid line
approximately 300 mm wide. The number of grid lines required with the
radar is a function of the condition of the deck (the worse the condition of
the deck, the more grid lines are needed). In general, a minimum of two
passes are made in each traffic lane. If the data indicate significant
deterioration (as determined by the operator reviewing the waveforms on an
oscilloscope), additional passes are made and, on badly deteriorated decks, a
grid-line spacing of as little as I m may be used.
On most bridges the time taken to set up the traffic protection usually
exceeds that taken to complete the infrared and radar surveys. Infrared
surveys should always be performed before radar surveys. This is because
any obstructions to solar radiation on the bridge, such as people or
stationary vehicles, can cause false readings in the infrared image.

Thermography
In general, areas of delamination on exposed concrete decks can be
identified at almost any time of day except near sunrise and sunset. Near
sunrise the surface temperature of the delamination changes from being
cooler to hotter than the average deck temperature. The reverse situation
occurs shortly after sunset. The 'window' in which delaminations can be
detected through an 80-mm thickness of bituminous concrete is quite
narrow. The optimumtime of day (in southern Ontario) is between 12 noon
and 1 pm, when the delaminations arenot only most easily distinguished but
314 D. G. Manning and T. Masliwec

the area of the 'hot spots' approximates the actual size of the delaminated
areas. Although the maximum difference in temperature usually occurs later
in the day, the outline of the delaminations becomes less distinct because of
heat transfer within the surfacing. After about 3 pm, the deck begins to cool
and the delaminations fade and are no Ionger detectable by the scanner.
The optimum time of year for thermography will depend upon local
climate and latitude. In southern Ontario, maximum temperature
differentials were observed in early or late summer when clear skies
prevailed, rather than in midsummer when overcast conditions were more
prevalent. Ambient temperature in the range of -33°C (in February) to
32oC were investigated and were found not to have a significant effect on the
detectability of deterioration. Differentials were, however, greater in
summer than winter because of the greater intensity of solar radiation.
Although the equipment has a sensitivity of about O·l°C, a temperature
differential of 1·5-2oC is a reasonable practical minimum to counteract the
effects of differences in emissivity caused by such factors as polishing in the
wheel tracks, staining and patching.
A summary of the factors affecting the quality of the infrared image is
contained in Table 1. More specific information is given in Ref. 8.

Radar
The major advantage of radar is the ability to propagate through layered
media, thereby giving information on bituminous surfacings, concrete deck
slabs, the interface between the two and any discontinuities. In fact, a
bituminous surfacing more than 25 mm thick enhances waveform
interpretation (compared with an exposed concrete deck) because
interference between the radar echo reflected from the deck surface and the
radar echo from deterioration below the surface is reduced substantially. On
exposed concrete bridge decks, reflected signals from defects interact with
surface returns. The ministry's existing software is applicable only to
asphalt-covered bridge decks.
Investigations have been made to assess the radar's operating character-
istics under a range of climatic conditions on a variety of structures. A
comparative evaluation of the relative strengths and weaknesses of the
DART's radar andinfrared systems is provided in Table 1.
lt is apparent from Table 1 that the radar andinfrared systems used in the
DART prototype complement each other very weil. Taken separately, the
radar system provides only the dimension of depth, while the infrared
thermography provides a two-dimensional surface image. By using both
techniques, it is possible to obtain a three-dimensional perspective of the
condition of a bridge deck.
 Bridge Deck Condition Surveys 315

TABLE 1
Operational characteristics of radar and infrared thermography

Operational Effect
characteristic
Radar Thermography

Meteorological limitations Adversely affected by Needs sun, little cloud

standing water only and wind; inoperable
with standing water
When can system be used? Year round Year round but must
have sun and no rain
What hours of the day can All day Usually 9.00 to 15.00;
system be used? optimum 12.00 to 13.00
Influence of external
parameters
Paint markings None Easily compensated for
Isolated debris None Easily compensated for
Oil stains None Easily compensated for
Skid marks None None
Snow/icejwater Adverse Adverse
Trafik None Adverse if solar
irradiance affected
Signal noise Ievel Waveforms need careful Hot spots may not be
monitaring and detected due to
calibration meteorological factors
or secondary
reflections
Signal reproducibility Independent of time, Dependent on solar
excellent heating, variable
Selectivity
Asphalt-covered deck Excellent results with Good if surfacing
surfacing > 25 mm <80mm
Exposed concrete decks Signal processing soft- Excellent results
ware not applicable
Penetration Detects deep-seated Detects only surface
effects; difficulty with effects; no penetration
surface effects
Inspection speed Maximum 8 km/h Maximum 4·5 km/h
(limiting factors) (pulse repetition rate) (video digitization
rate/camera scan rate)
Number of scans Many, depending on One scan per lane
required bridge width and
coverage required
Expertise required High Ievel of technical Basic training plus
competence experience on bridges
316 D. G. Manning and T. Masliwec

COSTS

The DART system represents a significant investment. The approximate

value of the major items (in 1989 Canadian dollars) is summarized below:

Radar $75000
Thermal scanner and converter $54 000 (or $2400 per month rental)
Microcomputer and peripherals $27000
(including extra circuit boards)
FM recorder $13000
Industrial VCR $5000
Fifth wheel $7000
Oscilloscope $2400
Generator $1500

Adding the cost ofthe vehicle, generator, external and interior racks, and
numerous smaller items such as cables, filters and interface devices would
make the replacement cost of the fully equipped DART vehicle
approximately $180000. lt should be noted that the ministry did not
purchase the thermal scanner and converter but rents the equipment during
the summer months for $2400 per month. In addition, the cost of developing
the software was $96 000 for the radar and $30 000 for processing the
thermal images.

CONCLUDING REMARKS

The operating experience with the DART system has shown that, within
known limitations, the prototype constitutes a viable, rapid, non-contact
and non-destructive method for performing bridge deck condition surveys.
lt is envisaged that the systemwill play a major role in providing data for the
ministry's bridge management system, initially at the project Ievel with
possible future expansion to ~he network Ievel. lt is anticipated that more
reliable data will result in improvements in both the decision-making
process and in fewer cost overruns in rehabilitation contracts.
Future work will involve upgrading the microcomputer. This will not only
increase the processing speed, but the enhanced memory will eliminate the
need for the FM recorder. Any other changes are likely to focus on making
the system simpler to operate.
Other studies will be undertaken to determine the feasibility of applying
radar to other components of bridge structures and to pavements.
 Bridge Deck Condition Surveys 317

REFERENCES

I. MANNING, D. G. and RYELL, J., Decision criteria for the rehabilitation of concrete
bridge decks. Transportation Research Record No. 762, 1980, pp. 1-9.
2. MANNING, D. G. and HoLT. F. B., Detecting delamination in concrete bridge
decks. Concrete International, 2(11) ( 1980) 34-42.
3. MANNING, D. G. and HoLT, F. B., Detecting deterioration in asphalt-covered
bridge decks. Transportalion Research Record No. 899, 1983, pp. 10-20.
4. MANNING, D. G. and HoLT, F. B., Deck assessment by radar and thermography.
Transporta tion Research Record No. 1083, 1986, pp. 13- 20.
5. MASLIWEC, T., An experimental and theoretical evaluation of IR thermography
for surveying the condi tion of bridge decks. SPIE, 934, Thermosense X (1988)
19-27.
6. ALONGI, A. V., CANTOR, T. R., KNEETER, C. P. and ALONGI, A. JR, Concrete
cvaluation by radar theoretical analysis. Transportalion Research Record No.
853, 1981, pp. 31-7.
7. CARTER, C. R., CHUNG, T., HoLT, F. B. and MANNING, D. G., An automated signal
processing system for the signature analysis of radar waveforms from bridge
decks. Canadian Electrical Engineering Journal, 11(3) (1986) 128-37.
8. MAsuwr:c, R. and MA!\iNJNG, D. G., Bridge deck condition surveys using the
DART prototype vehiclc. Annual Conferencc Proceedings, Roads and
Transportalion Association of Canada, 1987, pp. C3-C36.
 28
Inspection Based Rehability Updating for Fatigue of
Steel Bridges

ANDREW G. TALLIN and MARK CESARE

Polylechnie University ol Brooklyn, 333 Jay Street, Ne~t· York 11201, USA

ABSTRACT

This paper reports on a study of the fatigue reliability ofstandard AASHTO

bridge girder details. In this study, the linear elastic fracture mechanics
(LEFM) model offatigue crack growth was used. Reliabilities were found
using techniques based onfirst-order reliability methods (FORM) modified to
calculate inspection updated estimates of' reliability. Three example analyses
of two different details are given. Analysis results of' these details indicate that
even at inspection qualities significantly heller than might be expected in the
.field, that effect of'inspection which detects no damage, has only a limited effect
on the estimated reliahility.

INTRODUCTION
About one-third of steel bridges in the US are 50 years old, or older, and
many more are nearing that age. 1 As the number of bridges nearing old age
increases the need for inspection and maintenance becomes increasingly
important. At the same time the resources which can be allocated to the
proper maintenance of bridges is shrinking. Older bridges are more
susceptible to problems of aging such as corrosion and fatigue. This paper
demonstrates the use of first-order reliability methods (FORM) and linear
elastic fracture mechanics (LEFM) to update the estimated probability of
fatigue failure of steel bridge details based on the results of inspections.

CRACK GROWTH

There are two common approaches to the fatigue of steel, S-N analysis and
fracture mechanics crack growth analysis. The S-N approach relates the
319
320 Andrew G. Tal/in and Mark Cesare

amplitude of the stress range at a fatigue sensitive point to the life time in
stress cycles. This method has been used extensively in bridge fatigue
studies. 2 •3 Because S-N analysis does not relate to a measurable indicator of
damage, it is difficult to incorporate inspection observations into the fatigue
analysis. The use of an LEFM model for fatigue crack growth allows
information on the presence or size of observed cracks to be incorporated
into descriptions ofboth failure and inspection events. The LEFM approach
to the fatigue of steel bridges has been described by a number of
researchers. 4 · 5 The LEFM approach to fatigue crack growth relates the
range in the stress intensity !1K at the crack tip to the rate of crack growth
dajdN by the Paris-Erdogan equation: 6
da/dN = C !1Km (la)
= C(SY(a, y)foyn (lb)
where C and marematerial constants. The factor Y(a, y)fo is the stress
intensity factor and depends on the crack size a and the detail geometry y.
Equation (1) can be rearranged and integrated to result in

cI N
sim = fa" ~-~--
da (2a)
i=l a0 (jna)"'Y(a,yt
(2b)

where S; is the stress range ofthe ith cycle, a 0 is the initial crack size and acr is
the crack size in the Nth Ioad cycle. The function \fl(acrl is the darnage
accumulated during the growth of a crack from the initial size a 0 to a crack
of size acr· The sum CL: S;m is the cumulative Ioad effect on the component
causing the damage. Because of the complicated forms of Y(a, y) which
occur for practical geometries, integration of eqn (2) must be performed
numerically.
In the case of bridge girder details a number of stress intensity functions
(SIF) have been compiled by Albrecht and Yazdani 7 for a number of
AASHTO fatigue sensitive details. These SIF were used in this study. These
SIF take into account a number of factors.
• Surface or edge crack effects. Because cracks in components usually
initiate at surface ftaws, cracks begin as surface cracks during the first
stage of growth. Cracks which initiate at points such as the end of a
ftange are edge cracks even after the crack has become a through crack.
• Elliptical crack effects. Surface cracks during the first stage of growth
are elliptical and, due to three-dimensional effects, have a varying stress
intensity along the tip of the crack.
• Width effects which adjust for the finite thickness of the webs and
ftanges.
 Inspection Based Reliability Updating for Fatigue 321

• Stress gradient effects which account for non-uniform stress distri-

bution near the crack tip.
• Evolution of the crack as it grows. A crack in a girder detail is assumed
to begin as a surface crack, develop into a through crack and continue
to grow as a through crack until the critical crack size is exceeded. As a
result different geometric factors Y(a, y) are used during each stage of
crack growth.

RE LI ABILITY ANALYSIS

The reliability analyses donein this study were performed using first-order
reliability methods (FORM). Reliability analysis using FORM begins by
defining the failure event of each element or component using a limit state
function graii(Z). Whenever gra; 1(Z) :s; 0 failure occurs. The limit state is a
function of a vector of physical random variables Z such as loads, material
and geometric properties. The surface gra; 1(Z) = 0, called the limit state
surface, divides the space of random variables into the failure set and the safe
set. The probability offailure is the probability content ofthe failure set Pr=
P(graii(Z) :s; 0).
In FORM the vector Z is transformed into a vector of uncorrelated zero
mean unit normal variables U, called u-space. The minimum distance from
the origin to the limit state surface in u-space is equal to the first-order
approximation of the safety index ß = - <D- 1(Pr ). The point on the limit state
surface which is closest to the origin is called the design point and is denoted
U*. FORM can be extended to the reliability analysis of parallel systems
(g 1 :s; 0 n g 2 :s; 0 n · · · n gn :s; 0), where gi is the limit state function for the ith
event. For parallel systems U* is the point closest to the origin which satisfies
all the constraints. The probability of the joint event is estimated by
linearizing each of the Iimit states which are zero at the design point U*,
called active constraints. The resulting estimate of the probability of failure
is Pr= <D[- ß; R], where ß and R are the vector of ß values and the
correlation matrix of the active limit states, respectively. More detailed
descriptions of FORM can be seen in a number of references, for example
Ref. 8. A computer program, PROINSP, 9 based on the general purpose
reliability program PROBAN was used to perform reliability and updated
reliability calculations.

FATIGUE RELIABILITY

The fatigue failure of a component is defined as the event that a crack

exceeds some critical crack length acr· However, since the darnage 'P(a),
322 Andrew G. Tal/in and Mark Cesare

defined in eqn (2b), is monotonically increasing with crack length, the

capacity of the component can be defined as the amount of darnage
absorbed by the component as the crack grows from a0 to aw <l>(acJ The
accumulated load CL Si exceeds the darnage evaluated at acr:
N
gfail(Z) = 'P(acr)- L S[" ::=;; 0
i= 1
(3)

where Z is a vector of random physical properties, including material

properties, loads and geometric parameters.
Inspections are similar to failure events and depend on the result of the
inspection. Two inspection results considered here are (1) a crack is not
detected and (2) a crack is detected and its size is measured. The first event is
described by the limit state
N
gnf(Z) = c i=L1 sim- 'P(ao)::::;; 0 (4)
where a 0 is the minimum detectable crack. The event gnr(Z)::::;; 0 occurs when
the accumulated load CL S[" is smaller than the amount of darnage which
must be accumulated to grow a crack from a 0 to the detectable crack size a 0 .
The second event is described by the limit state

c i=L1 sim = 0
N
gfd(Z) = 'P(al)- (5)

where a 1 is the crack size measured during the inspection. The event
grd(Z) = 0 occurs whenever the darnage accumulated in growing a crack
from a0 to a 1 equals the accumulated load C"Lß[".
The result of an inspection can be used to update the probability of failure.
If I and F denote the inspection event and the failure events, respectively,
then the probability offailure given that the inspection event occurs is from
the definition of conditional probability:
P[Fni]
Pfii = P[I] (6)

The probabilities ofthejoint events F n I and I can be found using FORM, 10

where the event F n I is modeled as the parallel system consisting of the
failure event (eqn (3)) and inspection events (eqns (5) and (6)).

EXAMPLES
Cover Plate
A 1·25-in welded cover plate terminus (AASHTO category E) 11 on a plate
girder was analysed. This cover plate is similar to the cover plates which were
 Inspection Based Reliability Updating for Fatigue 323

TABLE 1
Yellow Mill Pond Bridge

Quantity Distribution

Miner's stress N( 1-4, V= 0·2)

log C (m) N2(p = <-9·17, 3-2); CJ = <0·069, 0·055);
p = -0·97148)
Initial crack size LN(O·O 18 43, 0·002 08)
Average daily truck traffic N(5669·0, V=0·1)

Fixed values
Final crack size 4·3 in
Crack aspect ratio 0·25
Thickness of cover plate 1·25 in
Thickness of ftange 1·26in
Width of ftange 16·47in
Thickness of web 0·76in
Weid size 0·5in

observed to develop cracks after only 12 years of service on the Yellow Mill
Pond Bridge in Connecticut. 7 Table 1 shows the distributions for each ofthe
random variables used in both the failure and inspection Iimit states. The
failure criterion was the development of a through crack greater than 8·6 in
in length. The inspection times were selected at the time when the reliability
index ß fell below 2·0 or when the failure probability Pr rose above 0·023.
Figure 1 shows reliabilities of the uninspected detail along with the
updated reliabilities due to inspections resulting in no cracks found at 24, 33
and 40 years. The inspection intervals decrease from 24 to 7 years after three

x No Iimit state
o No find at 24 years
6 No find at 33 years
0 No find at 40 years

FIG. 1. Updatcd reliabilitics for thc Yellow Mill Pond Bridge.

324 Andrew G. Tal/in and Mark Cesare

0·2 in crack at 12 years

cn.
2·0-f---------'h----------

1·0

0
Time (year)

FIG. 2. Updated reliabilities for the Ycllow Mill Pond Bridge wherc 0·2-in crack is found at
12 years.

inspections. This does not show the characteristic increase in inspection

tim es for the fatigue crack growth Iimit states investigated by Madsen et a/. 1 0
These results show that the failure event Fand the no-crack detected event I
have a strong enough degree of stochastic dependence for only a short time
after the inspection.
The detectable crack criterion used here is of especially high quality for
bridge inspection, having a mean minimum detectable crack of only 0-35 in.
With such an inspection precision the probability of detecting a crack at 24
years is over 30%. Because the crack growth rate is much higher for !arger
cracks, a decrease in the inspection quality (i.e. an increase in the detectable

TABLE 2
Rolled hcam ( WJO x 360)

Quantity Distribution

Miner's stress N(lO·O, V= 0·2)

log C (m) N2(f.l = <- 9·17, 3·2); O" = <0·069, 0·055);
p = -0·97148)
Initial crack size LN(0·001 2, 0·000 58)
Average daily truck traffic N(500·0, V=0·1)

Fixed ualues
Final crack size 2·5in
Crack aspect ratio 0·67
Thickness of fiange 1·68 in
Width of fiange 16·665 in
Thickness of web 0·945 in
 Inspeerion Based Reliability Updatingfor Fatigue 325

o Lim'rt state
6 lnspection at 17 years

FIG. 3. Updatcd rcliabilitics for a W30 x 360 rolled section.

crack size) causes the time between inspections to decrease. For example, if
the detectable crack size is equal to the through thickness the probability
that there exists a crack with a short remaining life time is quite high.
Figure 2 shows the estimated reliability of the same detail after a crack of
0·2 in was found after 12 years of service. The reliability immediately
following the inspection is elevated, ß ~ 20 at 15 years. However, the updated
reliability falls to below ß = 0 by 24 years.

Rolled Beam
A W30 x 360 rolled section (Table 2) was analysed at several Ievels of
applied stress range and for a single inspection where no crackwas detected.
The failure criterion was the development of an edge crack of 2·5 in in the
flange. As in the case ofthe cover plate the inspection time was selected at the
point where the reliability falls below ß = 3·0.
Figure 3 shows reliability index for both the inspected and uninspected
detail due to a Miner's stress range equal to 10 ksi. The inspection at 17 years
Iifts the reliability immediately after the inspection; however, the reliability
soon approaches the initial uninspected reliability.

CONCLUSIONS

The method of updating reliability described here can be used to estimate

reliabilities of fatigue sensitive details conditioned on the results of
inspections which result in either no detection of a crack or a measured size
of a crack. Because thc LEFM based fatigue analysis relates physical
quantities such as crack size and stress range the method of updating
estimated reliabilities using LEFM is straightforward.
326 Andrew G. Tallin and Mark Cesare

The examples here show the limited effectiveness of even fairly precise
inspections of details which had been observed to fatigue easily. Because the
inspection qualities used for bridges can only detect cracks which are
advanced in age, the effect of an inspection which detects no darnage on the
estimated reliability is limited to a short time past the inspection.
The example of the inspection which resulted in a detected crack gave
updated reliabilities which quickly fell off following the discovery of the
crack. In such cases the rate of deterioration of the estimated reliability can
be used to determine the speed at which repairs must be made in order to
maintain an acceptable level of safety.

ACKNOWLEDGEMENTS

The work described in this paperwas supported by grants from the National
Science Foundation (grant number MSM-8657854) and Det Norske Veritas.

REFERENCES

1. GALAMBOS, C. F., Bridge design, maintenance and management. Public Roads,

50(4) (1987).
2. MosEs, F., Probabilistic Ioad modeling for bridge fatigue studies. IABSE
Colloquium on Fatigue of Steel and Concrete Structures, Lausanne, Switzerland,
1982.
3. NYMAN, W. and MosEs, F., Calibration of bridge fatigue design model. Journal
of Structural Engineering, 111(6) (1985).
4. FISHER, J. W., Fatigue and Fracture in Steel Bridges-Case Studie.\', John Wiley
& Sons lnc., New York (1984).
5. Y AZDANI, N. and ALBRECHT, P., Risk analysis offatigue failure ofhighway steel
bridges. Journal of Structural Engineering, 113(3) ( 1987).
6. PARIS, P. and ERDOGAN, F., A critical analysis of crack propagation laws. Journal
of Basic Engineering, Trans. ASM E, 85 (1963).
7. ALBRECHT, P. and YAZDANI, N., Risk analysis of extending the service life of steel
bridges. Maryland DoT Report No. FHWA/MD-84/01, 1986.
8. MADSEN, H. 0., KRENK, S. and LIND, N. C., Methods of Structural Safety.
Prentice-Hall, Englewood Cliffs, NJ, 1986.
9. TALLIN, A. G. and SKJONG, R., PROINSP-U~er's Manual, A.S. Veritas
Research Report 87-2018, Hovik, Norway, 1987.
10. MADSEN, H. 0., TALLIN, A. G., SKJONG, R. and KIRKEMO, F., Probabilistic fatigue
crack growth analysis of offshore structures with reliability updating.
Proceedings Marine Structural Reliability S)'mposium, Arlington, VA, 1987.
11. AASHTO, Standard Specifications for Highway Bridges (1977), 12th edn.
American Association of State Highway and Transportation Officials,
Washington, DC.
 29
Diagnostic Dynamic Testing of Bridges on Brenner
Motorway

RAINER FLESCH

Structural Dynamics/BVFA Arsenal. A-1030 Wien, Faradaygasse 3, Austria

and
KARL KERNBICHLER

Institute for Reinfärced Concrete, Technical UniL,ersity Gra::.

A-8010 Gra::, Technikerstrasse 4, Austria

ABSTRACT

A dynamic methodfär the danwge evaluationollarge hridges is det'eloped br

BVFA and Technical Universirr Gra::. The meth()(l is a comhination ol
dynamic in-situ tests and dynamic calculations. The hasic idea is to detect
damage to hridges via changes of' the {zrnamic parameters. The steps in the
process are dynamic in-situ testing. mathematical modelling, fitting ol thc
mathcmatical model to the fest results and sensitirity inrestigations. The
development of'the method ispresented hy discussing the history of'the process.
Further, the steps are discussed, gh·ing ref'erences to detailed inf(mnation and
some examples .fi'om past projects. The tatest incestigations H·ere started in
Octoher /988 onfive hridges on the Brenner motonray. The tests lt·ere(inished
in April 1989 and analysis of'measurements is stillumfer 1my. More detailed
results ll'il! soon he availah/e.

OBJECT

Due to the drastic increase of the number of vehicles passing over bridges.
especially heavy trucks, methods for bridge inspection become more and
more important. The main aims are:
• safety inspection of the bridge, and
• detection of darnage at an early stage for minimising the cost of repair.
327
328 Rainer Flesch and Kar/ Kernbichler

At the moment mainly visual inspection techniques with limited scope for
quantification of darnage are used. This situation is unsatisfactory in our
world of high technology. Hence in 1981 BVFA and Technical University
Graz started the development of a global method for the inspection of
structures using vibrations. The method has been greatly improved over the
years using the increasing capabilities of computers and measuring
equipment. Experience mainly with prestressed bridges has been obtained.
The basic concept of the method is to formulate a dynamic model for the
virgin state of the structure. In general, darnage decreases the stiffness and
increases damping, resulting in changes of dynamic properlies being
different for different modes ofvibration. The dynamic properties aremodal
frequencies (eigenfrequencies), mode shapes and modal damping ratios.
Hence darnage will Iead to a certain pattern of deviation of dynamic
parameters and can be used for localisation and quantification of darnage in
a global manner.
The following steps are carried out:

• establishment of a structural dynamic model for the vertical direction

(FE model),
• experimental determination of modal parameters by dynamic in-situ
tests, and
• systematic system identification (fitting of the structural model to the
modal test model).

The procedure described above is the baseline evaluation of the dynamic

properties. lt should be done when the construction of the bridge is finished
and before the bridge is opened for traffic.
After a certain time of use (e.g. 2-5 years) the investigations should be
repeated. If changes of modal frequencies, mode shapes and damping ratios
are found in the experiment, the structural model must be fitted again to the
test results. Comparing the old and new mathematical model, changes of
stiffness in certain parts of the structure can be detected and q uantified in a
global sense (over the FE element, including the actual damage).
If changes exceed a certain Iimit, which has to be fixed for any particular
case, the darnage area found by the global method should be investigated in
detail by a local method.
With the detailed information (e.g. exact location of cracks, crack width,
depth of cracks, etc.) the fitting of the structural model to the experimental
results can be further improved. With the resulting optimum model the
redistribution of internal moments and forces for design Ioads due to the
stiffness change can be calculated. Basedon the results, in any particular case
a decision must be made on whether the structure is still safe or not.
 Testing ol Bridges on Brenner Motorway 329

The concept described above can be strictly followed only for new bridges,
when the baseline was evaluated after finishing construction and before
opening ofthe bridge, so that no disturbance occurs as a result ofthe traffic.
Formost existing bridges there is no baseline available. For some bridges
used for several years first measurements were carried out during the last few
years, giving the baseline forthat certain point in time. Often the evaluation
was done after major rebuilding or visual inspection. lt is argued that severe
damage, currently invisible, will increase by the next routine measurement to
such an extent that it can be detected via the change of dynamic parameters.
For older bridges it would be helpful to have a variant of the method
where darnage can be detected from a single in-situ test. At the moment only
ideas exist (e.g. comparison of tests carried out with different Ievels of
excitation force or with different locations of the excitation, or by the
systematic placing of additional masses during the test to increase the
opening of cracks, resulting in a further stiffness decrease and hence
increasing the probability of detection).
The method consists of the following parts (see Fig. 1):

• structural (mathematical) modelling,

• in-situ testing,
• analysis of test results (first step of system identification: elaboration of
modal test model), and
• systematic system identification (fitting of mathematical model to the
modal test model).

Systematic system identification is still a developing field with many

unsolved problems. Hence additional approaches which can help to localise
and quantify darnage are most desirable. One possibility is the use of
structural dynamics modification software for sensitivity investigations.
Using only the modal test model, a decrease of the bending stiffness is
applied systematically to the structure, giving the change of the dy.namic
properties due to a local stiffness change. This fruitful approach will be
discussed later in more detail.
Depending on the mode shapes, the single modes have a different local
sensitivity to stiffness changes. The more modes with different shapes that
are available. the easier will be the localisation and quantification of any
damage. In the past investigations have been focused on the change ofmodal
frequencies although this parameter only varies in proportion to the square
root of the stiffness change. In the future the changes of the mode shapes,
which are much more pronounced, will be used for interpretation. Before
changes of damping ratios can be interpreted the damping mechanism of
R/C structures must be better understood.
 t.H
t.H
0
I
Strueture
modifieation
Dynamie in-situ tests FFT analysis Experimental model r--- software
Parametrie
Reeorded time Non-parametrie model identifieation Parametrie model ......
histories transfer funetions modal data
:::tl
::::,
;:;·
• Hydraulie exeiter 'll
• Digital zooming SM S software + ....
BVFA software ~
'll
'- "'...,
:::-
Indireet
,------------------------- ---------------- identifieation §
I Improved parametric model ::::,_
I
I ~
....
• Global fitting --~
• Trial and error
~ • Systematie identifieatior ~
<::l-
Mathematieal Calculation of Results: ;::;·
modelling :::-
eigen frequeneies Modal data ~
....
and Parametrie
Parametrie model mode shapes model
------·- - --

• Beam elements FLASH

• Modelling of SAPIV
transverse eoupling ADINA
• Soil springs DYNA

FIG. I. Steps of the dynamic method for safety inspection.

 Testing ol Bridges on Brenner Motonray 331

THE HISTORY OF APPLICATION

The work was started in 1981. For the Raach Bridge dynamic and static tests
were carried out before and after cutting of a number of tendons. The
artificial darnage was possible because the bridge had tobe removed later for
a new traffic scheme. Compared with today the test equipment was quite
primitive, but it showed that the method works in principle. Much more
information was obtained by the dynamic method than by the static
approaches. The results are reported in Ref. 1.
During the first investigations ofthe Obernberg Bridge 2 the test technique
was improved significantly. A maximum difference (1~f= 0·006 Hz) for
modes 1-7 was obtained between measured and calculated modal
frequencies. Two of the higher modes were also closely fitted to the test
results by a more detailed modelling of the open caisson foundation of pier 2.
For the Gänstorbrücke in Ulm (FRG) tests were carried out before and
after major repair work. Measurements were carried out for three states.
Changes of modal bending frequencies and modal torsional frequencies
were detected. The increase of stiffness resulting from the repair work was
established by comparing the structural models for the different states. The
results are given in Refs 3-5.
In November 1985 investigations of the Lavant Bridge were started.
Lavant Bridge is one of the greatest prestressed framed bridges in Europe.
The maximum span is 160m and the maximum shaft height is 130·35 m. The
tests could be carried out in the virgin state before it was opened for traffic.
For the analysis of the measurements a modal analysis software package
from SMS was used for the first time. Further, sensitivity investigations were
carried out using the structural dynamics modification software. The results
are given in Refs 6 and 7.
Starting in October 1988 a baseline evaluation was carried out for five
bridges on Brenner motorways. The bridge investigated first was again the
Obernberg, this time after widening ofthe cross-section (adding one lane). In
the latest test series a new reaction mass exciter driven by a hydraulic
actuator was used. For ali previous projects an eccentric mass exciter had to
be used.
To obtain experience with other types of bridge a R/C arch bridge
(Äussere Nösslachbrücke) and a composite bridge with steel girders and a
concrete slab (Gschnitztalbrücke) were included in the programme. Further,
for one mushroom slab bridge (Nösslachbrücke) tests were carried out to
study the behaviour of gaps and bearings between the six substructures of
the bridge. Finally, another R/C frame bridge was tested to provide a
baseline after reconstruction in areas of sliding foundations due to hillside
creep.
332 Rainer Flesch and Kar! Kernbichler

Hence the latest series could be carried out with excellent test equipment
and analysis capability, but with disturbance by traffic. As routine tests have
normally to be carried out under conditions of traffic disturbance, it is
important from the latest investigations to find procedures to eliminate these
disturbances. The problern will be discussed in more detail later. The
analysis for all five bridges is still under way, with more results available
soon.
For the next 2 years three new projects are planned. Projects one and two
are baseline evaluations for a three-span prestressed bridge and a composite
bridge in the virgin state in Budapest (Hungary). Project three is an
investigation of a multi-span prestressed bridge at the border between
Austria and Yugoslavia. In this project tests of substructures during the
construction are also planned which will be very fruitful for improving the
indirect system identification approaches.

DESCRIPTION OF THE METHOD

Focus and Limitations for the Application

All bridges tested before the latest series were !arge prestressed bridges.
Investigations focused on that type, because the mostfrequent darnage to
prestressed bridges is bending cracks having areas of distorted bond at both
sides of a crack, giving some kind of smeared stiffness decrease in keeping
with the assumption in the structural model. Hence the method seems to be
suitable for prestressed bridges. A more detailed discussion of different
forms of darnage and expected effects on dynamic properties are given in
Refs 6 and 8.
In principle the method can also be used for R/C bridges, steel bridges or
composite bridges, especially if changes in boundary areas (foundation of
piers, bearings, etc.) is detected. In the Iiterature some cases of loss of
bearings are reported, darnage which could be quickly identified by the
dynamic method.
As it is not possible to close a bridge completely for testing, the
measurements need to be carried out with traffic disturbance. Hence
procedures to eliminate disturbance are necessary. Early results are
discussed at the end oftbis paper. It is only possible to close one lane in each
direction for about 2-5 days, the sametime span used for visual inspection
using the bridge inspection truck.
Next, influences from environmental conditions on the measured results
must be considered. Experience has been gained relating to the influence of
temperature. There is an influence if a stiffness change (e.g. due to an
elongation in the longitudinal direction) occurs in the directioll' of a
 Testing of' Bridges on Brenner Motorway 333

pronounced modal movement. The experience from Gäustorbrücke is

reported in Refs 3-5 and 8. It can be concluded that temperature has an
inftuence, but the single modes are inftuenced to a different extent. lt is
assumed that after exclusion of modes which are very sensitive to
temperature changes that enough modes are still available to give the
expected information.

Test Technique
As frequency changes due to darnage are often small a precise test
technique is necessary. In the past an eccentric mass exciter was used. Using a
static frequency changer the frequency could be controlled with an accuracy
of 0·003 Hz. The disadvantage of the eccentric mass exciter is the quadratic
force and the very low force amplitude in the low frequency range. From
experience the minimum force should be 1 kN to obtain good results.
Recently a reaction mass exciter driven by a hydraulic actuator was
developed by BVFA. The excitation force can be kept constant during the
frequency sweep using a control program. The exciter, the hydraulic pump
and the control equipment are mounted on a ftat-bed lorry which is used for
transportation of the heavy equipment to the excitation point. The
excitation can be carried out in vertical, longitudinal or transverse
directions. For inspection the bridges are mainly excited in the vertical
direction.
The response is measured by veiocity transducers (Hottinger SMU 30A).
The signals are integrated to obtain the displacement response. As there is
often a long distance between transducer and magnetic tape recorder the
amplifiers are mounted in boxes tagether with the transducers. The
frequency response function for each transducer was obtained on the BVFA
shaking table. The function is approximated analytically and used for the
correction of the measured transfer functions via a complex division. To
have a mesh of measurement points adequate for sensitivity investigations
the distance between measurement profiles should be about 5-lOm.
Normally 4-6 transducers are distributed over the cross-section in each
profile. First, 2-3 positions for the exciter, adequate to excite as many modes
as possible, are selected. Then the transducers are placed in one profile after
the other. Foreach configuration a frequency sweep is carried out in the
frequency range 0-10 Hz Iasting about lOmin. The time histories of
excitation force and of the responses are recorded on magnetic tape.
From these records transfer functions are calculated using a FFT
analyser. In the latest series about l 000 transfer functions were obtained per
bridge within l week.
From 2-3 positions of the exciter 40-60 modes were obtained. More
detailed information is given in Refs 6 and 8.
334 Rainer Flesch and Kar! Kernbichler

Analysis of Measurements
From the records of the response and of the excitation force, transfer
functions are obtained by FFT analysis. Before the Lavant project, peak
picking and circle fit algorithms were used to elaborate the eigen frequencies,
mode shapes and damping ratios.
For the analysis of the Lavänt Bridge measurements the SMS modal
analysissoftwareMODAL 3·0 could be used for the first time. The software
is very modular and provides a bestfit algorithm and is very easy to use. The
total concept seems tobe aimed at mechanical engineering problems. Special
features of tests of very !arge civil engineering structures had to be handled
by additional software. Detailed information is given in Refs 6 and 8.

Dynamic Calculations
The dynamic calculations were carried out by Technical University Graz.
SAP IV and FLASH were used for mathematical modelling. During every
project attempts were made to find the method with an adequate Ievel of
accuracy. Only beam elementsandadditional spring elements were used. In
the opinion of the authors, the inclusion of more complicated elements
would not improve the quality of modelling since material parameters of
concrete structures are often quite uncertain (deviations from plan cross-
sections, variance ofmass density, etc.). These uncertainties must be allowed
for using correction factors.
About 1200 DOFs were used for each model. In general, the bending
stiffness can be modelled weil but problems can arise for torsional stiffness
and cross-section deformation for open cross-sections.
To get precise results important details must be modelled weiL In some
cases the transversal coupling of adjacent bridges via the carriageway slab
had to be modelled. The problern was solved by eccentrically connected
beam elements. 2 •4 The influence ofthe foundation was sometimes modelled
weil by additional springs. 2

Systematic System Identification

Systematic system identification means the fitting of structural models
(FE models) to test results. The state ofthe art is given in Refs 9 and 10, and
in the proceedings of the workshop held in Lambrecht (FRG) in 1987 (see
Ref. 6).
In past projects systematic system identification was carried out in two
relatively simple steps:
• In a global manner using the modulus of elasticity. Hence the modulus
was used as central fit parameter, representing also uncertainties and
deviations of cross-sections, mass densities and boundary conditions,
etc.
 Testing ol Bridges on Brenner Motorway 335

• By trial and error in the areas of repair work, especially in the case of
Gänstorbrücke. 3 - 5
For routine inspection systematic methods are necessary. Methods are
under development at Curt Risch Institute (University Hannover, FRG) but
they are limited to models with 30 DOFs at the moment.

Sensitivity lnvestigations
lt was shown before that systematic system identification can be quite
difficult for !arge systems. Hence the application of the structural dynamics
modification (SDM) software provides a powerful tool for interpretation of
changes to modal properties. For the investigations only the modal test
model is necessary, hence no FE model has tobe fitted to the test results. The
software is applied to the baseline test model. To investigate the influence of
local cracks on the modal parameters the decrease ofbending stiffness has to
be modelled. A 3-DOF beam element (negative rib stiffener) is 'roved' over
the bridge, being implemented in three adjacent DOFs at one time. In that
way the influence of a local stiffness decrease (smeared over two adjacent
distances between measurement profiles) on the modal parameters can be
elaborated. The basic equations used for the software are given in Refs 6 and
8, tagether with all references. Changes of modal frequencies due to a local
stiffness decrease are presented in Refs 6 and 8. Changes of mode shapes are
given in Refs 7 and 8.
In the future a systematic approach using the values mentioned above for
the interpretation of measured changes of dynamic parameters will be
elaborated.
As an example some changes of modal frequencies obtained for the
Lavant Bridge are shown in Fig. 2. For every position given by the
horizontal axis you can find the decrease ofmodal frequencies on the vertical
axis, the negative element being centred at the enumerated position.
In the future SDM will also be used for systematic system identification.

Disturbances by Traffic
To establish a practicable routine test technique the method must be
applicable during traffic ftow. The influence oftraffic was studied in detail in
the latest test series. The results are still being processed. The first lesson
learnt isthat at least one fixed transduccr should be provided, which remains
in the same position during all frequency sweeps (e.g. in the profile of
excitation). Comparing the resulting transfer functions, disturbances can be
detected. Probably a procedure for data correction can be established. The
imaginary part ofthe transfer function in the driving point must always have
positive values, hence negative values at certain frequencies are a criterion
for disturbances during that sweep.
 336 Rainer Flesch and Kar! Kernbichler

Modal Modal
Modal frequency damping
number (Hz) ratio (Ofo)
• 5- 1·256 1-906 /X\
> X 7 -- 1·487 2·094 .
'-'
cQ) 0 9 ----
0 12 - -
2·386
2-895
1·262
1·145
1
:::J
cr ö 15- 3·685 0·959 I
~
c N
Q) :r:
Ol
c

u"'
.!:

Position

Modal Modal
Modal frequency damping
number (Hz) ratio ("lo)
-0·20
• 6 -1·361 2·216
x8 --2·224 1·221
> 0 13 ---- 3·060 1·205
'-'
cQ)
-0·15 •16 -3·739 1·244
:::J
cr ·20 - - 4·601 1-232
~ 0 25 ----- 5·419 1-475
c ~ -0·10 _,"R\

c
Q)
Ol f '
' Simulated stiffness
u"'
.!: ''
b.. __ ~ decrease
10"/o of local st1ffness
/...._
' il
~~~~~~~~~Ä~~-~~ -c-~.· -~
Position

I I I I
2 3 4 5

FIG. 2. Decrcasc of modal frcquencics due to a local stiffness decreasc, calculated for
Lavant Bridge.

In general, the traffic has the following inftuences on the results:

• Additional excitation. As for the calculation of the transfer functions a
cross-spectrum is used. The results are only distorted if the disturbance
has a momentary frequency component equal to the actual sweep
frequency. As the sweep has to be repeated for every measurement
profile, there is a low probability that disturbances will occur at the
same modal frequencies in each profile. Due to the !arge number of
 Testing of Bridges on Brenner Motorway 337

profiles distorted values can be recognised and eliminated, with enough

information remaining to determine the mode shapes. In addition, ifthe
baseline was evaluated for the virgin state without traffic, the original
mode shapes are known, making the analysis of routine measurements
much easier.
• As the additional excitation is not synchronaus with the artificial
excitation, traffic can also produce an additional damping mechanism.
• Additional moving mass. The inftuence from heavy trucks could be of
special importance.

The inftuence of additional moving masses seems tobe most severe but it
can be assumed that modal frequencies are always decreased. Hence, using
the !arge number of results, the most probable values for each modal
frequency and damping ratio can be established. During the analysis of
measurements additional modes (computational modes) are used during
curve fitting to take into account the disturbances. These additional modes,
often having negative damping values, are removed finally.

CONCLUSION

The history of application illustrates the development of the dynamic

method. lt seems to be possible to produce a technologically advanced
approach with greater capability than visual inspection but with
comparable time required for the inspection. lt is necessary to carry out the
routine tests under traffic. The first results of the Jatest series promise a way
to establish the most probable values for modal frequencies and damping
ratios.
At the moment analysis is very time consuming, but a further automation
and reduction of the time necessary for analysis will be sought. At present
the method is mainly applied to !arge prestressed bridges because their main
darnage mechanism fits quite weil with the model assumptions.
The changes of modal freq uencies d ue to structural darnage are often very
small but the changes ofthe mode shapes are often much more pronounced
due to the shifting of modal nodes.
Before changes of damping ratios can be interpreted the damping
mechanism of R/C structures must be better understood. The structural
dynamics modification provides a powerful tool to study the inftuence of
Jocal stiffness changes on the modal properties. This information can be the
basis for a localisation and quantification of darnage at a later point in time.
The latest test results for bridges of the Brenner motorway will be
presented in forthcoming papers.
338 Rainer Flesch and Kar! Kernbichler

REFERENCES
1. KERNBICHLER, K. and FLESCH, R., Static and dynamic tests, their qualification
for bridges inspection and long-term observations ofbridge structures. RILEM
Symposium, Budapest, 1984.
2. FLESCH, R., KERNBICHLER, K. and DüRNSTEINER, CH., Dynamic testing and
modeHing ofObernberg Bridge. Proc. Int. Conf on Num. Meth.for Transient and
Coupted Problems, Venice, 1984.
3. FLESCH, R., KERNBICHLER, K. and GRÜBL, P., Brückeninspektion mittels
dynamischer Untersuchungen. 8. GESA-Symp., VDI-Berichte, Nr. 514, 1984.
4. KERNBICHLER, K., FLESCH, R. and RAUSCHER, G., Dynamische Untersuchungen
von Großbrücken (Massivbrücken), in-situ Versuche und Rechenmodelle.
Tagung Dynamische Probleme, Universität Hannover, 1984.
5. FLESCH, R., KERNBICHLER, K. and RAUSCHER, G., Dynamische in-situ Versuche
und Rechenmodelle-Praktische Anwendung auf GroßbrückenjMassiv-
brücken, ÖIAZ, 131. Jg., Heft 10, 1986.
6. FLESCH, R. G. and KERNBICHLER, K., Bridge inspection by dynamic tests and
calculations-dynamic investigations of Lavant Bridge. Proc. Workshop on
Struct. Safety Evaluation Based on System Jdentification Approaches. Friedr.
Vieweg & Sohn, Braunschweig, Wiesbaden, 1988.
7. FLESCH, R. G. and KERNBICHLER, K., A dynamic method for the safety
inspection of !arge prestressed bridges. Proc. Int. Workshop on Nondestructive
Evaluation for Performance in Civil Structures. Department of Civil
Engineering, University of Southern California, Los Angeles, 1988.
8. FLESCH, R., Die Methoden der Baudynamik mit spezieller Berücksichtigung
ihrer Anwendbarkeit zur Bauwerksinspektion. Habilitationsschrift, Teil2,
Technical University Graz, 1988.
9. NATKE, H. G., Die systematische Anpassung von Rechenmodellen an
Versuchswerte als Verfahren zum Nachweis des dynamischen System-
verhaltens. Bauingenieur, 57 (1982).
10. NATKE, H. G., Einführung in Theorie und Praxis der Zeitreihen- und
Modalanalyse. Vieweg & Sohn, Braunschweig, 1983.
 30
Experience with the Management of Cable Stayed
Bridges in Korea

HELMUT WENZEL

VCE, Vienna Consulting Engineers, Vienna, Austria

ABSTRACT

Cable stayed bridges are sophisticated structures and more attention has tobe
paid to their management, i.e. inspection, maintenance and repair. In Korea
several bridges of this type were designed by foreign consultants in co-operation
with local companies. This turned out to be a procedure that did not produce
satisfactory results. Two different types of structures, steel structures versus
concrete structures, and two methods of design and management are presented
here and discussed.

INTRODUCTION

In the period from 1985 till 1989 the author's company had the chance to
work on four cable stayed bridges in Korea. Two ofthem are steel structures,
designed by British consultants and built by local contractors. The other two
are concrete bridges, designed by VCE and also built by local contractors,
but under the supervision of the designer.
There are different owners and so the method of managing the structures
is different. In general, we found that the understanding ofthe requirements
of bridge management is too small and it is difficult to succeed with new
concepts. The following report shows that it is nearly impossible to
introduce a standard close to the European state of the art.
After a description ofthe structures the paperwill present how the bridges
were designed, built and supervised, which inspections were carried out,
which defects were found and which repair activities took place. The final
section deals with a proposal for more effective management.
339
340 Helmut Wen::el

DESCRIPTION OF THE STRUCTURES

Jindo Bridge (see Fig. 1)

The spans are 70 + 346 + 70 m. The pylon is also a steel structure. The
cables are of the locked coil type. The pavement consists of 5 cm asphalt
only.

Dolsan Bridge
This bridge is similar to the Jindo Bridge, with smaller spans only
(70 + 260 + 70 m). The structural system and appearance are the same.

Olympic Grand Bridge (see Figs 2 and 3)

The 30m wide deck is supported by cable pairs anchored in the median.
The concrete structure is very slender. The cables are made of single strands

484 0158 } Link

~-70_0_0_0------------------~3~43~9~7~0~-------------------;-7_0_0_9_8_~~
"'14 11·4 ~
! I

...._ Jin Do rsrand Elevation an bridge Mainland----...

Scale 1, 1000
Guld~d bear~s_~r~ sl1d1ng bearmg

Deck cross section

FIG. I. Jindo Bridge. gcncral arrangcmcnt.

 The Management of Cable Stayed Bridges in Korea 341

FICi. 2. Olympic Grand Bridge, front and side view.

FICi. 3. Olympic Grand Bridge, typical cross-section.

0·6 in, 37-61 of them in one PE sheath tube, and were grouted after erection
with cement mortar. The pylon is a concrete structure. The main span is
300m. Construction was by the cast in-situ free cantilever method.

Haeng Ju Bridge (see Fig. 4)

This is a two-lane bridge built by the incrementallaunching method using
auxiliary piers pushing 60 m spans. The main spans were converted into a
cable stayed bridge using prestressed concrete members. The width is 16·5 m,
the deck height 4·5 m.

r---
320000

.J
100000 120000 100000
soooo l 50000 20000 5oooo I 5oooo

:~
I I

FIG. 4. Hacng Ju Bridge, sidc view.

342 Helmut Wenzel

DESIGN

The arrangements for designing bridges in Korea is typical of the new

industrialised countries. There are some well trained engineers, but the
amount of work is huge, and whoever is successful in one job will be
promoted and all his skills are lost. In Korean society a great distance can be
observed between generations and know-how is not transferred to potential
rivals. This fact makes it impossible to build up a strong design capacity and
it is very common to meet new engineers in the most demanding jobs.
Foreign consultants are engaged to design structures in co-operation with
local firms. Training programmes are arranged and specialists educated.
But, for the above reasons, these activities do not bring the required success.
The specialists are promoted away from their jobs and the next one starts
from the beginning.
This happened for all the bridges described in the previous section.
The two steel bridges were designed by a well-known British consultant in
co-operation with a big local engineering firm. The intention was to create
substantial local knowledge to deal with all future requirements for these
structures. In fact the author found after only 4 years that even minor things
could not be handled by the local firm. It was impossible to find the relevant
drawings in their files. A good design got a bad reputation, due to lack of
understanding.
The design ofthe Olympic Grand Bridge has a long history and was finally
produced by VCE in co-operation with the contractor. Engineers were
trained but disappeared after completion of the job. The designwas handed
over to the same local engineering company and site supervisionwas carried
out together with VCE. Again the author's company was confronted with a
number of engineers who had to be trained from the beginning and who
disappeared after some time. The author's company therefore failed to
install a strong engineering force capable of handling the requirements of
inspection and maintenance of this bridge.
A similar story applied to the Haeng J u Bridge.
The design ofthe Haeng Ju Bridge bearings were modified to save money.
The new design did not consider local stresses and the bearings were
deformed and had to be exchanged. The result was a 4-month delay in
construction. Figure 5 shows the bearing design before and after
modification.
Although there were different clients for all the structures, the same Iack of
understanding was found. A better procedure was recommended to them,
but accusations were made that the local situationwas being misunderstood.
The argument of higher overall costs for the structures was ignored because
different activities were covered by different departments from different
 The Management of Cable Stayed Bridges in Korea 343

1000

Rubber bearing

De.sign~d pushing bearing

_L____

J4o-~=]ß:~~~~~~~~~~~~~~PO~~~
~ /' 's \
I '\Rubber bearing
Fabricated pushing bearing

FIG. 5. Bridge bearing as originally designed and finally fabricated.

budgets. This is also common in Europe, but the understanding at least

exists.
lt is essential to put more emphasis on client education, rather than
consultant training, to improve the situation.

CONSTRUCTION AND SUPERVISION

The construction industry in Korea is one of the strongest in the world.

Accordingly, their self-confidence is disproportionately developed. This
Ieads to problems when it comes to the building of structures not fully
understood by the engineers. Plenty of 'good' ideas are born to reduce costs
and essential parts ofthe structure arealtered or completely deleted. In most
cases there is no authority to set Iimits and important parts are wrongly
produced.

Jindo Bridge (see Fig. 6)

The biggest Korean steel manufacturer received the order to build this
bridge. Foreign supervisionwas not considered. Several items were executed
differently to the original design. The cables were ordered in Japan. A proper
stressing procedure was not available and no adjustment took place after the
whole structure had settled. This, combined with Iack of tolerance control,
led to the problems described in the section titled 'Inspection'.
The problern of vibrations, which is normally handled by the designers
after erection, was completely neglected. Rumours were spread that there
was something wrong with the design. Only after 4 years, when the problems
grew worse, were experts askcd to rcappraise the situation.
344 Helmut Wenzel

FlG. 6. Jindo Bridge.

Good design ideas disappeared before completion of the structures and

much money now has to be invested to upgrade the bridges to the desired
Ievel. The engineering is alsomoreexpensive now than proper supervision
would ha ve cost.

Olympic Grand Bridge (see Fig. 7)

It was possible to convince the client that proper supervision is well worth
the investment. The main reason was not related to the lifetime costs; it was
the fear of problems with a structure important for the Olympic Games.
In this case the designer who had the basic ideas was appointed to
supervise the construction works. A Iack of information did not occur and
many of the 'good' ideas could be avoided.
The contractor was one of the big bridge contractors in Korea, and due to
the fact that the alternativedesignwas backed by them good understanding
made co-operation easy. All the new methods introduced by the author's
company were understood and applied seriously. There was no major
incident during the whole construction period.
Again education ofthe local engineeringpartnerwas attempted for him to
take over the functions of bridge management, but again a satisfactory
status was not reached.

Haeng J u Bridge
Experience gained through proper Supervision of the Olympic Grand
 The Management of Cable Stayed Bridges in Korea 345

FIG. 7. Olympic Grand Bridge during construction.

Bridge was ignored at the Haeng Ju Bridge. Again the contractor pointed
out that the material suppliers would provide supervision free of charge,
which led to considerably lower quality at higher prices.

INSPECTION, MAINTENANCE AND REPAIR

The understanding of these items was completely lacking. Inspections are

done only if something strange happens or darnage is done by vehicles.
A proposal put forward by the author's company for a minor inspection
every year and a major inspection every 6 years was not accepted.
The idea ofinspecting before the warranty period expires was also new to
them. The main argument put forward was that very seldom was any
company held liable for darnage to their work. That this requires better
Supervision during construction was not accepted.

Jindo and Dolsan Bridges

The call for an inspection came after complaints about vibrations during
strong winds. A team of engineers was sent to Korea and a kind of routine
inspection was performed. Three major defects were found .
Firstly, most of the cable bearings were loose, because the complicated
fixing proposed by the designerwas changed to a simple method that did not
work. The final fabrication is shown in Fig. 8.
346 Helmut Wenzel

FIG. 8. Anchorage of the backstays.

The elastomeric blocks were inserted on top of the anchor structure and
the fixing device was not installed. lt is not shown in the contractor's shop
drawings. Due to the vibrations the pads went down along the cables. At the
first and last cable the client's workers installed a wooden box to support the
pads, which was better than nothing. The slipped pads can be seen on the
bottom of the anchor tube.
At the pylon the samesituationwas found by VCE. The pads had come
down all the way to the deck anchor.
Due to this fact no damping was available to the cables and there is doubt
about the number of Ioad cycles endured to date and also whether there is
any darnage to the cables at the anchors, which cannot be inspected with
simple methods.
The amplitude of the cable vibration was not excessive but too much for
comfort.
The second factwas that the cables are stressed very differently and not
according to the design. This can be seen particularly at the backstays, where
only three out of six cables bear 80% of the Ioads. The loose cables showed
unusual behaviour during strong winds. Their vibration was more
horizontal than vertical and an uncomfortable horizontal impact was
observed at the slender bridge deck. Due to the small width ofthe bridge and
the long span, the horizontal excitation can be considerable.
The rubber cable bearings that were provided to damp the vibrations
should have been fixed in position by a tiny steel structure to avoid
 The Management of Cable Stayed Bridges in Korea 347

'II

,,,
Designed cable bearing (fixing device)

--tR '"
'jj - -
'"
"'
:-
-~!~-:~1 =::::::=J-9

Fabricated cable bearing ( friction only)

FIG. 9. Hearing as originally designed and finally fabricated.

movement during vibrations. These angles could be installed only after the
erection and the work was difficult. On the other hand, the cable bearing has
tobe pushed in and the friction is high after installation. Tosave the difficult
fixing device erection the simpler detail was adopted, neglecting the long-
term behaviour of the bearing. Figure 9 shows the detail before and after
modification.
The third item found was a cracked asphalt surface and other damage to
lighting poles and hand rails. This can also be explained by the excessive
vibrations during the typhoon season. One of the poles was buckled
completely.
After the first visit areportwas presented that pointed out what measures
would be necessary to maintain and repair the bridge, tagether with a cost
estimate and schedule. Since then nothing further has happened because
there could be no budget allocated. The suggestion that the cables might
break without warning after a certain number of Ioad cycles did not help.
The proposed inspection procedure was not introduced. The bridge is still
inspected by the relevant road master, who has no technical education
related to bridge engineering.

Olympic Grand Bridge

Due to sufficient site supervision during construction and execution in line
with the original concept, including an inspection after erection, it was
possible to create a structure of excellent quality with low expected lifetime
costs.
The design and supervision concept was a success.
During construction attempts were made to save money using inferior
quality materials and alteration of difficult items, but subsequent
348 Helmut Wenzel

supervision hindered all attempts to do so. The success was satisfactory for
everyone. The dient got an excellent structure and the contractor had no
difficulties and finished weil ahead of the schedule, which is very unusual in
Korea.
This successful experience was not, however, adopted as a future
standard.

Haeng Ju Bridge
Even the good example a few kilometres upstream, the Olympic Grand
Bridge, did not result in the installation of a similar procedure for this bridge.
Attempts were again made to save money by deleting consultancy services.
Basic design, detail design and site supervision were carried out by three
different parties without any communication.
The result was a 4-month delay at the beginning of the works, because the
contractor did not understand the design and ordered wrong materials.
Then came construction 150% slower than the concept. At present the
construction has had tobe stopped due to difficulties with the accuracy. One
can imagine what kind of quality will be handed over to the dient.

CONCLUSION

lt is sad that bridge management does not exist in countries like Korea,
because there would be the chance of avoiding all the mistakes that
happened in Europe during the past 40 years. The hopeful beginnings
introduced by foreign consultants and supported by a few local engineers
have not been successful due to an unfortunate promotion system and split
competences and budgets for activities that should be seen in a combined
way.
It will be found that an average of 1·5% of the erection costs have to be
spent for maintenance every year after the growth ofthe countries' highway
network slows down. At that time costs for the work done now will be
considerably higher than the average. But it is too much to expect
understanding of these long-term problems.
The VCE proposal given to the clients in the Far East is:
-Provide an inspection by reasonably educated engineers (locals) every
year.
--Provide a detailed high-dass inspection every 6 years.
-Allocate an average budget of 1·5% of the investment costs of all
bridges for inspection and maintenance.
 31
Performance Monitoring of Glued
Segmental Box Girder Bridges

PETER WALDRON, MAHMOUD RAMEZANKHANI

Department of' Civil Engineering, University of Bristol,

Queen's Building, University Walk, Bristol BS8 ITR, UK
and
BEN BARR
University of' Wales, School of' Engineering,
PO Box 917, CardiffCF2/XM, UK

ABSTRACT

Over the past 5 years three major glued segmental bridges haue been
instrumentedfär the measurement ofstrain and temperature. A comprehensive
data hase of'results 1101\' exists containing strains and temperatures measured
atamanher of'important cross-sections in each of'the hridges. Since reference
readings 1\'ere recordedjust afell' days after casting, the resultsform a unique
recordj(Jr the assessment of'time-dependent eflects such as creep, shrinkage
and lass of' prestress at a/1 stages hef'ore, during and after erection. Recent
developments reported here include Ioad fesring of the structures, to provide
confirmation ol structura/ perfärmance, and the insta/lation of' a telemetry
system to enahle the continuous monitaring of' instrumentation to he
com'eniently managedfi·om the otfice.

INTRODUCTION

Segmental methods for constructing concrete box girder bridges are now
commonly used throughout the world. The term segmental construction
refers to any concrete bridge structure that is cast in a number of discrete
longitudinal segments. Several alternative techniques have evolved over the
years. 1 These are usually categorised according to the method of erection,
namely (i) balanced cantilever, (ii) progressive placing, (iii) span-by-span or
349
350 Peter Waldron, Mahmoud Ramezankhani and Ben Barr

(iv) incrementallaunching, and by the method of casting, either (a) cast in situ
or (b) precast. Of thesedifferent techniques, the balanced cantilever method
is by far the most common, accounting to date for approximately 85% of
segmental bridges built in the USA and two-thirds of those in the UK.
Whereas the other three methods of erection are best suited for continuous
viaducts with spans less than 60 m, the balanced cantilever approach has
been used extensively for spans up to approximately 250m, using in-situ
concrete, or 140m where precast segments ha ve been employed. 1

Structures Monitored
The research described here concentrates on the balanced cantilever
method as the predominant erection technique. More specifically, the study
is centred on glued segmental construction in which match-cast precast
segments are prestressed together through a thin layer of epoxy resin (Fig. 1).

FIG. I. Grangetown Viaduct under construction by the glued segmental method as a

balanccd cantilcvcr.

This technique is becoming increasingly competttlve for medium span

bridges and has been employed in the construction of seven of the 16
segmental bridges completed in the U K in the last decade. 1
The three bridges chosen for this study reflect the full range of structural
configurations for which glued segmental construction is a viable option.
The various features incorporated in the three bridges included straight or
highly curved alignment, constant or variable section depth, reetangular or
 Performance Monitorinr; of' Glued Segmental Box Girder Bridges 351

trapezoidal cross-section and differing methods of erection either by

launehing girder or crane.

( a) River Torridge Bridge

Located 1 km north of Bideford, North Devon, the structure formspart of
the 8-4-km Bideford bypass. 2 Completed in May 1987, it carries two lanes of
traffic 29m above mean high water Ievel over the Torridge estuary.
The bridge consists of eight continuous spans, each up to 90 m long, with a
totallength of 645 m. The superstructure, which is straight in plan, is formed
from 251 segments each weighing up to 105 t, varying in depth from 6·1 m at
the supports to J I m at midspan. Segments were match-cast on site and
erected using a purpose-built launehing gantry, itselfweighing 150 t. Aseach
segment was positioned a thin Jayer of epoxy resin was applied to the
matehing surfaces immediately prior to prestressing, thus ensunng
watertightness and a uniform transfer of stress across the joint.

( h) Grangctmrn Viaduct
The two remaining viaducts, which have been eompleted more recently,
are both located on the peripheral distributor road currently under
eonstruction around Cardiff.
The Grangetown Viaduet is over I km long and is the Iongest glued
segmental bridge in the UK. The twin trapezoidal box girder superstructure
is made up from 641 match-cast segments each weighing up to 74 t with a
near-constant depth of 2·8 m increased locally to 3·5 m at the supports. 8oth
the eastbound and westhound earriageways consist of 17 spans up to 72 m in
Jength.

( c) Cogan Viaduc I
The Cogan Viaduct, which is the second strueture to be instrumented at
the Cardiff site, provides a 15m navigahle clearance to the River Ely at high
water. With a 95 m main span and a radius of eurvature of only 285m for
some of thc 60 m approach spans, it is more complicated than the adjacent
structure although only one-third of the total length. U nlike the
Grangetown Viaduet, a reetangular hox section was selected to accom-
modate the variable section depth made necessary hy the !arge main span,
which is the Iongest in the UK for this type of structure. In all some 300
segments were used to construct the twin carriageways, varying in weight
from 43 to 117 t.
Segments for hoth of the Cardiff viaducts were match-cast on site by the
short line method. Although deck erection of Grangetown Viaduct
eommenced using alaunehing girder, the majority ofunits in both structures
werc ercctcd hy creme.
352 Peter Waldron, Mahmoud Ramezankhani and Ben Barr

PROGRAMME OBJECTIVES

The principal objectives of the research were threefold:

(a) the assessment oflong-term time-dependent effects such as creep and
shrinkage in the concrete and loss of prestress in the steel;
(b) verification of short-term structural performance due to the
application of vehicular loading; and
(c) the measurement of differential temperature in concrete box girder
bridges and an assessment of its structural significance.
All three objectives are best met by the development of generalised
computational models validated against high quality field data. Tothis end
an extensive programme offield monitaring was undertaken complemented
by a parallel study in the Iabaratory to provide the necessary input data on
the creep and shrinkage of the concretes used for construction.

INSTRUMENTATION

(a) Time-dependent Effects

Instrumentation for the long-term (20 years) measurement of time-
dependent concrete strain was similar in all three bridges. In the Torridge
Bridge four segments were selected for monitaring within a half-span of one
of the central 90 m spans. 2 Three further segments at approximately
midspan, quarter-span and support sections were instrumented within a
single half-span in both the remaining bridges. A typical 72 m span was
selected in the Grangetown Viaduct and a highly curved 60 m side span of
constant depth in the Cogan Viaduct. 3
Each ofthe ten segments selected was instrumented with embedment-type
vibrating wire strain gauges cast in the concrete during construction. These
were Jocated at a number of discrete points on the median line of the walls
forming the box section. In some segments gauges were deployed only in the
axial direction for the measurement of longitudinal strain; in others the
gauges were arranged as three-element rosettes to determine the component
of shear strain as weiL Nearly 300 gauges were used altogether. These proved
to be very robust and reliable with a failure rate of only 1%.
Due to the Iack of security on site and the remoteness of individual
segments prior to erection, data were acquired and entered into a computer
data base manually. Although inconvenient for the first few months after
casting and during erection, when readings were required every few days, the
procedure was the best that could be achieved and worked weiL
Once in a data base the raw results were reduced automatically to units of
 Performance Monitaring of Glued Segmental Box Girder Bridges 353

strain and adjusted for the effects of temperature. They could then be
considered in a variety of different formats on the computer screen before
being produced as hard copy for the validation of long-term performance.
The two most useful formats for presentation of data have proved to be the
distribution of strain (axial, shear or principal) around the cross-section at
any number of ages, and a strain/time plot for any number of gauges.
An example of the former is given in Fig. 2, which shows the distribution
of axial strain around the instrumented segment adjacent to pier 4 in the
Torridge Bridge immediately prior to erection (160 days), after erection of
the half-span containing the instrumented segment (197 days), after
continuity was established with the adjacent span (246 days) and at
completion ofthe bridge (450 days). Figure 3 is an example ofthe complete
strain/time plot for the four axial strain gauges cast in the corners of the
same segment.

~x. -..-1+------><
--- _.:.:
. ~:::~'-~-=-----+--+ ---- ~
>.::----- "'X"
_y- ,.q.. - .. - - -V- - 9 - - - - - . . 7 1
v---- -?---

Ir----~~-- -Js- ----Ir- -Ir---- -A

~ 't t 4 t II
+ I I II
II I I
I I I II
II I
t H
*t
~
f I I II
U I I I /II
l I I I I II

u tl~t
t t#
l w

PLOT SC ALE 1------11. OOm

STRAIN SC ALE 1------10.25 x 10_,

KEY

REFERENCE AGE • 5.00 DAYS

tr------t. 160.00DAYS

v- - - - - - -Ii 197. 00 DA YS
+ - - - - - + 246. 00 DA YS

><------ ·>< 450.00 DAYS

FIG. 2. Axial strain distribution in a Torridge Bridge pier segment.
354 Peter Waldron, Mahmoud Ramezankhani and Ben Barr

z
H

l>',EJr
;;i
H
2 6
"'00::
u
H
:;::

-"
800

~
~
15 II
600

400
~~~ y
V
.....
'~-" ·..........-
-* ~

-
200
.dJ.

0
l/
0 so 100 150 200 250 300 350 400 450
AGE SINGE CASTING (days)
FIG. 3. Strain/time plot for the axial strain gauges in the four corners ofthe Torridge Bridge
pier segment.

The basic time-rlependent properties were obtained from numerous prism

specimens, each with a single vibrating wire strain gauge located at its centre,
manufactured from the batches of concrete used for the construction ofthe
instrumented units. Some of the prisms were sealed fully soon after casting
to prevent the loss ofmoisture; the remainder have either been left unsealed
or partially sealed to represent more accurately the environmental
conditions at the various bridge sites.
A number of these prism specimens are being used for the assessment of
Young's modulus, Poisson's ratio and the coefficient ofthermal expansion
and their variations with age. The remainder are being used for the long-
term measurement of creep and shrinkage effects. Once again data are
collected manually and entered on a computer data base for later
interpretation. In addition to straight strain/time plots, it is then a simple
matter to obtain plots of other parameters such as creep coefficient, which
are necessary for the validation of the time-rlependent analysis.
By assessing all ofthe data, evidence is beginning to emerge regarding the
accuracy of the available methods of analysis and the validity of the various
creep and shrinkage algorithms used for design. A time-step analysis has
been developed for the assessment of any segmental bridge constructed
using concretes of different ages. Strain at any point and at any stage of
construction can be calculated from a knowledge of the physical properties
 Performance Monitaring of Glued Segmental Box Girder Bridges 355

of the actual materials used for construction and the sequence and age at
which certain erection events occurred. Results of this analysis can then be
compared with those obtained from field measurement and from those
derived by using algorithms recommended in different international codes of
practice. Early results from these comparative analyses indicate that all
major code recommendations have shortcomings but that the ACI-78
recommendations provide the best fit to experimental data. 4 • 5

(b) Short-term Performance

In the two Cardiffbridges short-term structural performancewas assessed
by load testing. This has provided sufficient data against which the original
design calculations and any alternative analytical approaches could be
validated.
For each test, the transporter used for the delivery of segments during
construction was loaded to 130 t and stopped at a number ofpositions along
the viaduct on the instrumented and adjacent spans. In both Ioad cases the
tests were first performed on the bridge centreline, to provide information on
flexural performance, and then at maximum eccentricity, for an assessment
of torsional behaviour. Results were recorded manually as before and
entered on to aseparate data base for later consideration. Figure 4 shows an
example of axial strain in the four corners at one of the instrumented
sections within the Grangetown Viaduct due to the application of the Ioad
test vehicle at eight different longitudinal positions.
In this way, by using existing instrumentation installed for the monitaring
oflong-term performance, and by making use of existing site equipment and
staff, it was possible to provide valuable confirmation on performance at
very modest cost.

LOAD GASE
40 A B c D E F G H

z
H
;;:i 20
H
"'0
"':;::u
H
0
top corners
-20

-40

-60

FIG. 4. Axial strains in the corners of a segment of the Grangetown Viaduct du ring Ioad
testing.
356 Peter Waldron, Mahmoud Ramezankhani and Ben Barr

(c) Differential Temperature Effects

One of the four segments instrumented for strain in the Torridge Bridge
was also used as a pilot study for a possible investigation ofthermal effects.
A number of thermocouples were installed within the concrete of one of the
segments for the measurement of temperature. Early results indicated that,
under certain climatic conditions, significant differentials occurred both
through the thickness ofthe individual walls and over the entire depth ofthe
bridge.
Since a more detailed investigation was warranted, a further segmentwas
selected for monitaring temperature effects in the Cogan Viaduct. This
segmentwas within the same half-spantobe instrumented for strain but was
in addition to the three segments containing vibrating wire gauges. The
instrumentation consisted principally of arrays of thermocouples placed
across the thickness of each wall element forming the box section,
supplemented by a number of individual thermocouples located at the
corners and near the outside surface of the concrete (Fig. 5). The segment
was also heavily instrumented with electrical resistance strain gauges placed
orthogonally in pairs adjacent to the thermocouples for the purpose of
monitaring short-term longitudinal and transverse strains due to diurnal
temperature variations.
Since the purpose of the thermal investigation was to monitordifferential
temperatures caused by rapid changes in the environmental conditions, a
manual system of data acquisition was no Ionger practicable. As there was
no particular merit in monitaring the bridge in the difficult site conditions
prior to completion, this aspect of the monitaring has only recently started.
A Solatron data acquisition systemdriven by a PC is now installed in the
relatively secure closed cell of the bridge, powered from the mains circuit
running through the box for lighting.
After running the system automatically for several weeks and collecting
data from each of the 150 sensors every hour, it became apparent that

r--------'-1 i

L6*~·r5· 1 7
10.
9

8
60i

1
1
48

46. r .
47f 4 9
• 45
$~

~: ~L17 1 _)38~::
19 + 20 ,. 36 + 35

22
21 +
+
+ 3

•
24
30
25
f
. 34.

•
31
+ 33
• 32

Fru. 5. The position of thermocouplcs in a scgment of the Cogan Viaduct.

 Performance Monitaring of Glued Segmental Box Girder Bridges 357

handling and interpreting the !arge amount of stored data would be a major
task. Moreover, data would ideally need tobe collected over several seasons
in order to monitor the most onerous loading conditions due to differential
temperature effects. The problern has recently been overcome by the
installation of a telemetry system using a standard telephone line. This
enables the remote computer on site to mirnie the hostcomputerback in the
office. lt is now possible to switch the data logging system on or off, to
change the frequency of readings, and to transfer data back to the office
automatically. This innovation, shown schematically in Fig. 6, has provided
greatly increased flexibility and efficiency in data acquisition at very little
additional cost.

-----~

I
I
I Solar
Hlcrocomputer I Hlcrocomputer Oata LoggJn•tl----< IntensJty
I RS232 System
Remote Hachlne I Host Hachlne f----CI E.R Straln
Gauges
I
I
I Thermocoup Je
Hot JunctJon

I I I
I I Amblent Air
l_ ~FICE_ _ 1 I_ B~DC::, ~T=-- _ _ _ _ _ _ _ _ _ ~mperature _I
FIG. 6. Data acquisition and tclcmctry system cmployed in the Cogan Viaduct for
monitoring dif!'crcntial tempcraturcs.

Although the system has only been installed recently, interesting data are
already being collected. Figure 7 shows the variation oftemperature over the
depth ofthe box during a relatively hotspell in June 1989. The surface ofthe
100 mm thick blacktop is considerably hotter than the ambient temperature;
the differential between the top and bottarn concrete surfaces is in line with
the maximum design values given in BS 5400. 6
Another point of interest is the effect of shading on the webs. Figure 8
shows the outside surface temperatures at the centre of both webs, one of
which is always in the shade whilst the other is shaded for the greater part of
the day by the overhanging side cantilever. The very rapid change in
temperature caused by this effect is just one ofthe many factors complicating
a detailed thermal analysis.
One such method of computer analysis currently under development is
being validated against this field data. This can calculate and allow
358 Peter Waldron, M ahmoud Rame::.ankhani and Ben Barr

so concretc

101
~

u
0 -~~~--
concret•
45 (\
~ /I ----~~.'~!~---··
~ I I /'I
::0
H 40 I
I I
I
I I
1\
I
blacktap
sur(ace.

~
~
P-<
~ 35
J
't
I
I I
!
.
I\ I
r J .:···J ! I
~
H i\, I / t.. ,'\ I
·. I f i ': j / \\. I \\.. j I
30
\ . . \ .....,;. ./ \1_-~ // .... i:' i\ ,\ _,•''l
. . . . . ._ ,_: \_ :/ i ·- / / \ \__ i / \''·
25 ·-~" ·' \ . . . .;-: . .\ ... / . !' .:·' f_,/ \
\·-L......-···/\ \~\·.:··_L.---···
V ·-••
I ,/ '\··-~·"t·;·'.'.....-······, ' . . · . _)../
' ..... ~.' I . I
.. l:
·-- .. /
\
20 .....r '-.,. __ _:... -·······' .,\,\ 1 •••••••''\••
\) '•,)'<:.:/~ ... --
15 LOw...~w.u~.~luu.lluuluuuuu~•••••••••••l••••"•~~=••••••l,,,,,,,,,,4=~tuu.
5 DAYS
FKi. 7. Tcmpcraturc di!Tcrcntials mcasurcd ovcr thc dcpth of thc instrumcntcd scgmcnt in
thc Cogan Yiaduct.

automatically for shading due to bridge configuration as weil as for the

effects of bridge location and orientation, time of day and season, degree of
cloud cover and turbidity, thermal properlies of the constituent materials,
and the nature and colour of each surface in terms of solar radiation
absorptivity, emissivity and convectivity. Based on a finite element
approach, the method requires design or measured values for these various
q uantities to be input. Then, in order to validate the analysis against
measured differential temperatures, values of those parameters which vary

3 4 DAYS
Fici. ~. Thc c!Tccts of full and partial shading on tcmpcraturcs mcasurcd on thc outside
surfaccs of both wcbs ovcr a 5-day pcriod (.Tune 1989).
 Performance MonitorinK of G/ued Segmental Box Gin/er Bridges 359

(;' 800 f--

8
......_
~
'-'
j\
'' ''
:>-< 600 f-- '' ''
':' ''
:
:>-< E-<
E-< H ''' '''
H rJ)
''' ''
u z ''
''
0 ~ 400 - '' '''
.....:1 ''' ''
I:LI z '''
''
''
> H
i !
Q : ~

z ~ 200 - i \
H .....:1 : ~
~ DAY 1 DAY 2 0
rJ) IDAY 1\
'' ''
/DAY 2\
0
} \ j \

(a) (h)

FIG. 9. Variations of(a) wind spccd and (h) solar intcnsity mcasurcd on thc Cogan Viaduct
ovcr a 2-day pcriod (.I unc 19l-:9).

continuously, such as wind speed and solar intensity, must also be input.
These are measured on site with an anemometer and solarimeter, and
recorded automatically with the other data as continuous functions against
time, as shown in Fig. 9.
Results from the analysis are cncouraging and permit temperatures to bc
estimated within a few degrces anywhere in a concrete section. A
conventional finite element strcss analysis may thcn be used to compute the
distribution of stress both longitudinally and transverscly. Early results
indicate that the transversc strcsscs duc to thc framc action around thc
closed cell of a box girdcr may be equally important as the longitudinal
stresses, which have already received some attention.

CONCLUSIONS

Meaningful instrumentation has bcen installed in three segmental bridgc

decks at modest cost which may be used for studying both the long-term
time-dependent behaviour and the short-term response due to applicd Ioad.
With the introduction ofinexpensivc portablc computers, it is timely to start
creating data bases of actual performancc during construction and service
life to assist with the technical management of real structures. Such
instrumentation systems should be installcd as a matter of routine, not only
in bridges of complex geometry or novel design but also in a number of
typical bridges of standard configuration.
360 Peter Waldron, Mahmoud Ramezankhani and Ben Barr

ACKNOWLEDGEMENTS

This research is being undertaken as part ofthe co-ordinated programme of

!arge- and full-scale testing of structures funded by the Science and
Engineering Research Council. Additional funding and support has been
provided by South Glamorgan County Council.

REFERENCES

1. ZIADAT, G. S. and WALDRON, P., Segmental construction-a state-of-the-art

report. Department of Civil Engineering, University of Bristol, Report No.
UBCE/C/86/1, December 1986, p. 159.
2. ZIADAT, G. S. and WALDRON, P., Measurement of time-dependent behaviour in
the River Torridge Bridge-instrumentation and early results. Department of
Civil Engineering, University of Bristol, Report No. UBCE/C/87/4, October
1987.
3. BARR, B. I. G., WALDRON, P. and EVANS, H. R., Instrumentation of glued
segmental box girder bridges. Int. Assoc. Bridge Struct. Engng Colloquium,
Bergamo, Italy, 1987, pp. 175-89.
4. WALDRON, P. and ZIADAT, G. S., Assessment of the long-term behaviour of
segmental bridges. Proc. /nst. Struct. Engrs, Building Res. Est., Seminar on the
Life of Structures, Brighton, April 1989.
5. ZIADAT, G. S., Time-dependent analysis of prestressed concrete segmental
bridges. PhD thesis, University of Bristol, September 1988.
6. BRITISH STANDARDS INSTITUTION, BS 5400: The design ofhighway bridges, Part 2,
1984.
 32
Remote Computer-Aided Bridge Performance
Monitoring

T. 0. SLOAN

Civil Engineering Department, The Queen's University ol Belfast,

Bellast BT7 JNN, UK
J. KIRKPATRICK

Department ol the Environment Roads Service, Commonwealth Hause,

35 Castle Street, Belfast BTJJGU, UK
and
A. THOMPSON

Civil Engineering Department, The Queen's University ol Be/fast,

Bellast BT7 JNN, UK

ABSTRACT

A project to monitor the in-service hehaviour ol the Foyle Bridge, near

Londonderry, is described. Special instrumentation to measure the dynamic
movements of'very /arge structures has heen developed, Iogether with sojill'are
to control it and to process the data obtained. The system can be run
unattended and can he operared hy remote control. Typical results are
presented.

INTRODUCTION

In 1984 the Environment Committee of SERC announced a programme of

structural research based on !arge- and full-scale testing of structures. As
part ofthis overall programme a project to monitor the behaviour, in service,
of the Foyle Bridge is being carried out. The work is funded jointly by the
Science and Engineering Research Council and the Department of the
Environment (NI) Roads Service, and is being run by the Department of
Civil Engineering, The Queen's University of Belfast
The new Foyle Bridge, near Londonderry, NI, opened to traffic in May
361
362 T. D. Sloan, J. Kirkpatrick and A. Thompson

1984. lt is a high level crossing of the River Foyle, some 3 km downstream

from the Craigavon Bridge in the city centre. Designed by Freeman Fox and
Partnersand built by RDL-Graham JointVenture, the bridge is 866 m long
and comprises three main spans totalling 522 m constructed in steel, tagether
with a 344m approach viaduct of 13 spans in prestressed concrete. Full
details ofthe design and construction have been published elsewhere, 1 - 4 but
the principal features of the structure are summarised here.
The whole bridge, both steel and concrete, consists of two independent
parallel structures, as can be seen in Fig. 1. The main steel structure consists
of twin box girders which vary in depth from 3m at midspan to 9 m at the
main piers, over each of which there is a diaphragm. Each girder was
fabricated, off site, in three sections, which were then floated to site and
winched into position. The site splices are located at the sixth points of the
centre span and are bolted, except for the deck which is site welded. To
achieve the desired bending moment conditions under zero applied load
conditions, the ends ofthe side spans were raised 6·5 m before the centre span
was fixed and lowered to the final position only after splicing was complete.
Both girders are fixed at the west abutment and there are expansionjoints
at the junction ofthe steel and concrete sections ofthe structure. At the main
piers the girders are carried on pinned bearings, the horizontal movements
being accommodated by flexure of the piers.
As there is no interconnection between the two girders they move
independently of each other and, at midspan, a vertical differential
movement of up to 500 mm can be expected under severe storm conditions.
At 30m above water Ievel the site is very exposed, being completely open
to both north and south and with almost no shelter from any other direction.
The basic design wind speed given in CP3, Chapter 5, for the area is 53 mjs,
compared with 50 m/s for the F orth Bridge and 43 m/s for the Severn Bridge.
Occupying as it does the most exposed site of any major bridge in the British
lsles, the Foyle Bridge offers an excellent opportunity to monitor the in-
service performance of a box girder structure.
The project is concerned only with the steel section of the bridge. The
objectives are:

• To investigate methods of continuously monitaring long-span bridges

with the aim of developing a cost-effective solution which could have
application in similar situations elsewhere.
• To establish a 'footprint' of the structural response to a range of test
Ioads when the structure is in the new condition.
• To monitor the behaviour of the structure under a variety of wind and
traffic loadings.
• To compare the measured and predicted behaviour of the structure.
 Remote Computer-Aided Bridge Performance Monitoring 363

22600

7300 2100 7300 501800 200

carriagewo y ltghting
colu rnn

0
0 '<
0 0
m E

main pier

FrG. I a. Cross-sections of stcel girders.

865·700m oo bnd
sl~~l box superstrucfure

233 ·626m 141.·267m 7n Sn

West poer 1 2 L7 5 5().3 503 55 55 35 41. Southbox
obutment
3 5 6 7s 8s 9 East
Elevation obut ment

FrG. I b. G eneral arrangement of thc Foyle Bridge.

364 T. D. Sloan, J. Kirkpatrick and A. Thompson

REQUIREMENTS OF THE MONITORING SYSTEM

Preliminary studies indicated that, as an initial requirement, data would be

needed on:
• deflection at the midspan point of the centre span;
• strains at midspan and at the intermediate supports;
• temperature changes in the structure;
• wind speed and direction; and
• movement of the expansion joint.
Several other requirements of a moregeneralnature soon became apparent.
The bridge site is remote both from the university (130 km) and the DoE
base in Coleraine (50 km). For security and safety reasons access to the
interior of the structure must be restricted. Since, in addition, studies of
response to wind, temperature and traffic have tobe made on a continuous
basis, the system had tobe capable of operating unattended for long periods
and of storing the resulting data for subsequent analysis.
Interruptions to the power supply, while infrequent, can occur at any time,
but are more likely during storm conditions when it is particularly important
to collect data. As funds were not available for an uninterruptible power
supply, the system had to be capable of restarting automatically after a
power failure. Furthermore, as transient voltages on a power supply are
liable to produce unpredictable effects in a computerised system, it badtobe
possible, as a last resort, to reset and restart the system under manual remote
control.
Automatie data collection generates very large volumes of data, up to 500
bytes/s in this case. Since, at this rate, even the largest storage devices would
soon be filled, it was essential to have significant processing capacity on site
so that a preliminary analysis ofthe raw data could be done immediately and
only the significant parts retained for transmission to the university for
further analysis.
In any system exceptional conditions of various types will occur from
time to time, caused either by malfunctioning of equipment or external
interference. In an unattended system such events must be logged so that
corrective action can be taken.
Finally, the system bad tobe capable of expansion and refinement in the
light of experience gained.

HARDWARE

A schematic diagram of the monitaring system is shown in Fig. 2. The core

of the system is a DEC PDP-11/53 minicomputer fitted with 512 kb of
 Remote Computer-Aided Bridge Performance Monitaring 365

telephone

H
remote user

FIG. 2. Foyle Bridge equipment schematic.

memory, a 20Mb Winchester disk drive and six RS232 serialline ports. The
computer runs under the Micro/RSX-11 M real-time, multi-tasking
operating system and can be accessed either locally from a terminal in the
bridge or remotely via a modern and dial-up telephone line from either the
university or the DoE offices.
The modems are fitted with error correction logic which provides an
error-free connection though at the cost of a slight reduction in effective
Operating speed. The hardware is configured so that a 'break' signal on the
modern line will cause the processor to halt, in which state a re-boot can be
initiated by the user.
Intermittent problems, caused by interference on the mains or the
telephone line, can cause the modern to 'lock up' and fail to answer incoming
calls. Purpose-built circuitry has been developed to reset the modern
periodically and so ensure that such a condition cannot persist for more than
4 h. A more complete description of the system can be found in previous
publications. 5 · 6
Measuring the deflections of such a !arge and inaccessible structure
presents considerable difficulty; the fundamental problern lies in establishing
a frame of reference from which to make measurements. Conventional
methods, such as levelling or the use of displacement transducers, were
clearly impractical, due both to the distance to the nearest fixed point and to
the need to record the response to dynamic loadings.
Initially consideration was given to the use of accelerometers but these
had to be rejected for several reasons; they will not measure the very small
accelerations experienced during the slow-moving deflections caused by
366 T. D. Sloan, J. Kirkpatrick and A. Thompson

slave slave
computer computer

pier 2 pier 1

FIG. 3. Layout of cquipment to measure deftection of main span.

heavy vehicles and even with fast-moving traffic or wind loading they would
be operating at the Iimit of their resolution. The solution adopted is shown
diagrammatically in Fig. 3. Helium neon Iasers fitted with beam expander
and focusing optics are mounted inside the box girders at midspan so as to
project a spot of light of approximately 3 mm diameter on to targets which
are fixed over the main piers. Any movement (either linear or rotational) of
the Iasers will thus cause the position of the light spot on the target to vary.
The movement of each spot is tracked using a solid-state camera in which
the film carrier is replaced by a light-sensitive computer chip. The chip
presents to the Jens an array of 128 x 256 light-sensitive cells which are
examined in turn by a specially developed program running in a BBC
microcomputer. These computers, one for each camera, scan the complete
arrays repeatedly and store the resulting coordinates of the light spot in
memory for later transmission to the central computer.
All the other data are collected using transducers which generate electrical
analogue signals. These are connected to a pair of free-standing 16-channel
analogue-to-digital converters, each of which is connected, in turn, to the
main computer via a serialline. This arrangement was chosen primarily to
allow the A/D converters tobe located as close as possible to the transducers
and so minimise the length of the cables carrying the analogue signals. The
alternative, which was to incorporate the A/D converter in the main
computer, would have involved cable runs of up to 150m with all the
attendant problems of signal attenuation and noise pickup.
An anemometerwas already available as part of an ice alert system which
was installed when the bridgewas built. It has been possible to tap into this
system to record the wind conditions.
The movement of the expansion joint is monitored using a linear
potentiometer which is linked to the A/D converter via a current loop line
driver as the cable is 150m long.
The relative movement between the two girders at midspan is also
monitored using a rotational potentiometer mounted on one girder with the
 Remote Computer-Aided Bridge Performance Monitaring 367

wire connected to the other. This has been included as a check on the
performance of the Iaser deftection system.

SOFTWARE

The whole system operates under the control of the Micro/RSX operating
system in the PDP-11 minicomputer. This is a real-time, multi-tasking,
multi-user operating system in which the following features are specially
useful:
• Tasks can be scheduled to run at specific times.
• One task can initiate and monitor the running of another.
• There are very good facilities for controlling the hardware without the
need for programming in assernhier language.
• The operating system can be configured so that, on start-up, a user-
supplied command file is executed to allow the work schedule to be re-
established after a power failure.
All data acquisition is under the control of a single program, SCNALL,
which accepts requests from either the user or from another program,
performs the scan and writes the results to a file in which full details of the
operation-time, date, scan rate, channel identifiers-are recorded.
Calibration factors are applied to all data as they are recorded so that
readings taken at different times are directly comparable even though the
equipment may have had to be reconfigured in the meantime. Data
compression techniques are used to reduce the size of the data files to a
mmtmum.
The maximum scanning rate of the cameras is 8·32 Hz and up to 8192
readings may be taken at one time. This allows the analysis offrequencies of
up to 4Hz, weil above the structure's fundamental frequency of 0-4 Hz.

Application Programs
When studying the dynamic response of the structure it is obviously
necessary that all the readings from a scan be available for analysis.
However, for static or 'pseudo-static' (such as temperature effect) studies
only a single value is normally needed. But because the strudure is
constantly in motion, even on the calmest day, such a single observation
needs tobe the average of a set of repeated observations. I t isasimple matter
to write an application program which will generate arequest to SCNALL
to perform a scan and then process the data file to extract the required
information. Multiple applications can be concurrently scheduled for
repeated runs; competing requests for the use of SCNALL are resolved via
the operating system.
 368 T. D. Sloan, J. Kirkpatrick and A. Thompson

RESULTS

At the time of writing only a limited amount of data have been collected.
Some preliminary results are shown in Figs 4-7.
Figure 4 shows the infiuence line for static midspan deflections due to a
75-t load moving across the bridge. This testwas done on a calm day when
the bridgewas closed to traffic. Each reading is the mean of 1024 camera
scans, to remove the effect of vibrations induced in the structure when the
load was moved from one position to the next. The position of the load is
measured from the east end of the steel structure.
Figure 5 shows the change in strain measured at the top and bottom ofthe
box section at midspan during the same test.
Figure 6 shows the dynamic deflections recorded as a test load of 100 t was
driven over the bridge from east to westat approximately 43 km/h. Initially
the structure was almost motionless, what small movement there was being
caused, probably, by a light breeze. As the load passed over the bridge, the
midspan point first rose and then, as the load reached the centre span,
deflected downwards by 220 mm before rising again as the load moved to the
second side span, At the right-hand side of the trace, the structure can be
seen vibrating at its natural frequency of 0·4 Hz, the amplitude reducing
after the source of the disturbance was removed.

60
A.
50
40 / '.,
/ ">\
30 f \ / \
20 I \ I
_A '\ \
10
0 \ I \
-10 \ l
\ I
i -20
\ I
~ -30
z -40 \ I
0
-50 \ I
F
u
w -60 '\ 1
.J
u. -70 \ I
w
0 -80 I
-90 1\ /
-100 \ I
-110 \ I
-120 ~
-130 \
-140 \ I
\ I
-150
-160
y
0 200 400 600
POSITION (METRES)

FIG. 4. Deftections at midspan.

 Remote Computer-Aided Bridge Performance Monitaring 369

120
110
100 ~~
90 I \
80 I \
70 \
60 I \
50 I \
40
; \
30 I \
I
.""' ~
20
10 ~ f --6--

0 "J ~
-10 .___ -....- -<!
/'\
"\.
/"" "'-- h-----4./
-20
-30 \ I
-40 ~ I
-50
\ I
-60 I
-70 \ f
-80 ~
0 200 400 600

POSITION (METRES)
0 Bottom 1::. Top

FtG. 5. Strain at midspan.

Figure 7 is a plot produced from data collected on the movement of the

expansion joint caused by temperature changes in the structure. A clear
trend can be seen, with the 'best fit' straight line having a gradient of 7·9 mm
per degree change in steel temperature. This figure is about 2·0 mm larger
than that tobe expected from the expansion of the steel alone; the difference
can be accounted for by the expansion of the concrete section of the bridge.

~~~--~--------------~
E
E

·-
-&
0.&

öl

<SI
<SI
?+---------.----------.---------.---------.----
0 50 100 150 200
Time (Seconda)

FKi. 6. Dcllcctions duc to I 00-t moving Ioad.

370 T. D. Sloan, J. Kirkpatrick and A. Thompson

lSI
lSI
10
10
~

~ISI
~s

g.~
D
+'ISI
c .
-ISI
010
...,",

c
OISI
•ISI
CISI
""'
Q.
X
WISI
lSI
~+---------~-----------r----------T-----
5.11.1 111.1.11.1 15.11.1 211.1.11.1
Temperature (degree• Cl

FIG. 7. Thermal movement.

A long-term study of this behaviour is currently under way and should allow
confident predictions ofthe expansionjoint movement with temperature to
be developed. In the future this knowledge will form a useful check on the
performance of the joint.

CONCLUSIONS

• A fully automatic monitaring system has been developed for a major

structure. lt can be run equally weil under direct operator control or
unattended.
• The system is flexible in that additional sensors may be added easily
without the need for major reprogramming. The principles used could be
adapted for use in other projects.
• The Iaser displacement system has been shown to be reliable in operation
and to give consistent results.

ACKNOWLEDGEMENTS

The work described in this paper has been supported jointly by the
Department of the Environment for Northern Ireland and the Science and
Engineering Research Council. The authors also wish to thank Professor
A. E. Long ofthe Department ofCivil Engineering, The Queen's University
of Belfast, for his support and encouragement throughout the project.
 Remale Computer-Aided Bridge Performance Monitaring 371

REFERENCES
1. PRESCOTT, T. A. N., STEVENSON, W. M. C. and NISSEN, J., Foyle Bridge: its history,
and the strategy ofthe design and build concept. Proc. Inst. Civ. Engrs, 76(l)(May
1984) 351-61.
2. WEx, B. P., GILLESPIE, N. M. and KINSELLA, J., Foyle Bridge: design and tender in
a design and build competition. Proc. Inst. Civ. Engrs, 76(1) (May 1984) 363-83.
3. QUINN, N. W., Foyle Bridge: construction of the foundations and viaduct. Proc.
Inst. Civ. Engrs, 76(1) (May 1984) 387 -409.
4. HUNTER, I. E. and McKEOWN, M. E., Foyle Bridge: fabrication and construction
of the main spans. Proc. Inst. Civ. Engrs, 76(1) (May 1984) 411-48.
5. LEITCH, J. G., THOMPSON, A. and SLOAN, T. D., A novel dynamic deftection
measurement system for )arge structures. Proc. Civil-Comp. 89, Civii-Comp.
Press, Edinburgh, September 1989 (to be published).
6. SLOAN, T. D. and THOMPSON, A., Development of an automatic data collection
system for a majorbox girder bridge. Proc. Civil-Comp. 89, Civii-Comp. Press,
Edinburgh, September 1989 (to be published).
 33
Inspection and Repair of some Highway Bridges in
ltaly

MARIO P. PETRANGELI
Department of Structural and Geotechnical Engineering,
'La Sapienza' University, Via Eudossiana 18, Rome, Italy

ABSTRACT

This paper reports about the inspection andrepair ofa number of 25-year-old
viaducts in service along the A3 highway linking central to southern Italy.
Particular reference is made to dynamic tests that allowed useful information
tobe collected about the e.ffective prestressingforces acting an the beams. The
results obtained are compared with those derivedfrom other in-situ tests. The
criteria followed for the repair ( or demolition) of the decks and for their
seismic retrofitting are finall_v hriefly presented.

INTRODUCTION

The 44-km long A3 Salerno-R. Calabria highway is the main road

connection between the centre and the south of ltaly, including Sicily. This
highwaywas built about 25 years ago. lt reaches the Ievel of 1015 m above
the sea, crossing the Apennines in a zone where deicing salt is often used.
The government agency (ANAS) responsible for the management of the
road has invested about 200 million US dollars to inspect and rehabilitate
the existing 264 main bridges, which have a totallength of about 56 km, most
of them with pc decks spanning over 35-40 m.
In this paper the results ofthe inspections carried out on seven viaducts as
well as their retrofitting are illustrated.
The 32 simply supported pc decks were all ofthe sametype (see Fig. 1) and
showed severe damage, such as deterioration of the concrete surface and
cracking of the beams.
First of all, an accurate visual inspection by means of a by-bridge was
373
374 Mario P. Petrangeli

~.-·
0 31·30
....'-
~ c:
::> §-I
Vl
I tl I Ii Vlj
I a [ u I
I
n II L
I
I I I I II II !~! u1i1
I II II
~---- - ·' •L 31·30 ~I
FIG. 1. Schemc of thc decks.

performed; as a result of that three groups of decks, according to their

deterioration, were defined:
(A) the pc beams had no visible cracks;
(B) the beams were cracked but the total width ofthe cracks measured on
the four beams was less than 4 mm, moreover the total width of the
cracks in a single beam was less than 2 mm; and
(C) the beams were cracked and one or both of the two previous Iimits
were exceeded.
This preliminary classification has been kept as a reference for the
instrumented tests.

METHODS OF INSPECTION

Materials
The following non- (or moderately) destructive tests on the decks were
performed:
(a) Boring of 100-mm cores from the webs of the beams with
measurements taken of the released stress; in the class C decks
compression stresses as low as 30% of the theoretical values were
found.
(b) Determination of the mechanical properties of the concrete of the
cores as well as the carbonation depth, which ranged between 8 and
19mm.
(c) Measurements of the ultrasonic pulse velocity in conjunction with
the surface hardness.
(d) Windsor penetration tests.
(e) Systematic inspections of the state of the prestressing cables by
means of an endoscope. The Jatter proved very useful since it showed
when the injection of some cables was badly made or, in some cases,
completely lacking.
 lnspection and Repair of Some Highway Bridges in Italy 375

Dynamic Tests
Fifteen spans of seven different viaducts underwent dynamic tests. The
decks analysed had been previously classified (i) three in group A, (ii) eight in
group B (moderate damage) and (iii) four in group C (severe cracks).
The investigation of the dynamic behaviour of each deck was carried out
by means of low-frequency geophones (range of resolution 1·5-3000 Hz).
Most of the tests were performed with two geophones, one on each side of
the deck at the midspan, in order to establish the first modes of vibration
(flexural and torsional); in a few cases eight geophones (four at the supports,
two at the quarter-point and two at midspan) were employed to collect
information about modes of higher order and about the behaviour of the
neoprene bearing devices.
The data collecting system comprised an analogue-to-digital converter
connected to a microcomputer capa ble of handling simultaneously up to 16
channels.
The maximum sampling frequency was 3-5kHz when all the channels
were operating, reaching 17kHz if only one channel was active. To excite the
decks vehicular traffic was utilised. The highway has never closed, at least
one lane being always open. This procedure did not permit external actions
to be precisely defined. Hence the analysis of the recorded signals has
focused mainly on the free vibrations occurring after the passage of some
hea vy vehicles.
Figure 2 shows some typical time histories ofthe amplitude ofthe signals;
the segments of these diagrams selected as 'meaningful' allowed the
derivation of the Fourier spectra of the type shown in Fig. 3.
Table 1 presents the most important results for each deck: the vibration

Grotta Della Pag11a I- 4a campat" lato RC a 4_2

C.1

c 2

(C1•C2)/2

(C1-C2)/2

~-----"-'S 2
_J
'-"'9n-=a-=ls--'a"-'n"-a1-"-'yz=-:ce-=-d_ _ _
3 4

FIG. 2
376 Mario P. Petrangeli

Fourier 3pectra
file : Pag 4_2.CD

Frequency

Fig. 3

period 1'r for the flexural mode and ~ for the torsional mode. The three
numbers in each column represent the extreme and average values found for
all the analyses (generally four or five) carried out for each deck. In the table
the ratio ~/I'r and the classification previously mentioned arealso reported.
One can see that the dass A decks (light or no damage) have 1'r values
practically equal to the theoretical value expected of 0·27 s, while those
showing severe cracks have 1'r higher, up to 30% more.
lt must be pointed out also that the ratio ~/I'r could be useful in assessing
the deck soundness since it seems to decrease as the darnage increases.

TABLE I

Deck Tr (s) ~ (s) ~/Tr Initial dass

I. S. Venere 3 0·31-0·31-0·31 0·27-0·27-0·27 0·87 c

2. S. Venere 4 0·30-0·31-0·31 0·27-0·28-0·27 0·87 B
3. S. Venere 5 0· 31--0· 32-0· 32 0·27-0·28-0·28 0·88 c
4. Grotta Io 1 0· 27--0· 30--0· 28 0· 24-0· 25-0· 24 0·86 A
5. Grotta Io 4 0· 30-0· 32-0· 31 0·25-0·26-0·25 0·81 c
6. Grotta Io 5 0· 34-0· 35-0· 35 0·26-0·27-0·26 0·74 c
7. Grotta Ilo 1 0·30-0·31-0·31 0· 24-0· 25-0· 24 0·77 c
8. Ranta 3 0·28-0·28-0·28 0· 25-0· 25-0· 25 0·89 B
9. Ranta 4 0· 33--0· 34-0· 34 0· 28-0· 29-0· 28 0·82 c
10. Vomice 4 0·34-0·35-0·35 0·27-0·28-0·27 0·77 c
11. Vomice 5 0·30-0·31-0·31 0·26-0·27-0·26 0·84 B
12. Vomice 6 0· 33-0· 34-0· 34 0·26-0·27-0·26 0·76 B
13. Bodetti 1 0·27-0·27-0·27 0· 24-0· 24-0· 24 0·89 A
14. Spatolette 2 0·35-0·35-0·35 0·27-0·27-0·27 0·77 c
15. Spatolette 5 0·35-0·35-0·35 0·27-0·28-0·27 0·77 c
 lnspection and Repair of Some Highway Bridges in ltaly 377

RESULTS OF THE INSPECTIONS

In order to decide the level of repair, knowledge of the effective prestressing

force acting in the bearns was of prirne irnportance. Sorne data were derived
frorn the stress releasing already rnentioned. Further inforrnation was
provided frorn the analysis of the dynarnic tests. Since the natural
frequencies of a grid frarne depend on the stiffness EJ of the bearns, it is
possible to relate 'Fr to the flexural rnornent of inertia J, provided that the
value of the E modulus is known with good accuracy by rneans of the non-
destructive tests and frorn the cores.
Linking J to the height of the cracks and this last pararneter to the actual
prestressing force NP, the rneasured 'Fr can be correlated with the unknown
NP. Figure 4 shows these diagrarns for the decks referred to.
The tensile strength of the concrete has been assurned equal to zero
because of the repeated live Ioads the decks have been subjected to in their
life. The curve relating NP to 'Fr shows how, for NP lower than Npc (value of
NP that cornbined with the dead Ioad only gives zero cornpression at the
bottorn of the bearn), 'Fr increases quickly with reduction of NP. No
inforrnation is available for values of NP located between Npc and NP,
(theoretical val ue ).
Better results are possible if a known static Ioad is located on the deck
during the dynarnic tests since a higher value of Npc is achieved. In this case
the problern is to have a rnass rigidly connected to the deck or a vehicle
whose dynamic characteristic is precisely known. The reliability of the
dynarnic tests has been confirrned from static Ioad tests that were permitted
on a few decks with the closure of one carriageway.
The results of these tests are represented in Fig. 5, giving the load-
deftection curves in the midspan of a lateral bearn. Line 't' gives the
Q
7<>r! ! N(ton)
I ~:rrTl JO "
I
I IJ -- <>i~l l
0.6Q I""' ''·
1
f-1.+--
,. --
"-
- ;-..
r-~
Q 50
-- ---~- I I !
r.,
_.".-
-·"" I
! 25
..-,-
I
Q ~\
·-.... I
0-~
\ _..-8' ' I I \<;l
fa""'
'· ..
~ l

:tt±±rfL-L
0 .20 ~

20
,..
0 I r-~ 1--+ 1\.
I. I1
0
020 040 OSJ 0.80 1.00 l20
15 ... ···r ..
hl(M) 3 0.4 0.5 6
To( ltC)
FIG. 4(a). Prestressing force N, funda-
mental period T0 and moment of incrtia J FIG. 4(b). Prcslressing force N versus
vcrsus hcight of thc cracks. pcriod T0 .
378 Mario P. Petrangeli

car I car2 car3 car4

FIG. 5. Defiection d versus Ioad in a lateral beam.

theoretical answer for the decks while line 'I' gives the same but with a value
of EJ reduced by 20%; this was assumed tobe a Iimit that, if passed, would
dictate that the test be terminated.
These curves show how decks 4 and 13, with a natural frequency of0·27 s
(equal to the theoretical one), performed well during static tests.
Decks 3 and 5, which had 'Fr greater than 0·3 s, had the loading interrupted
after two (of four) vehicles were located on the bridge.

METHOD OF REPAIR

On the basis of the experimental tests and of the subsequent computations

carried out, three Ievels of repair were decided according to the residual
prestressing force in the beams:
-Less than 40%; in this case the removal of the old deck and the
construction of a new one was found to be eheaper than the repair.
-Between 40% and 90%; new prestressing cables were added to the
beams by enlarging their bottom flange (see Fig. 5). Number and size of
these new cables were based upon the existing prestressing force; their
anchorage was produced by a steel frame or by a concrete block
depending on the distance between the beams (Fig. 6).
 lnspection and Repair of Some Highway Bridges in Italy 379

o) EXISTING CABLES
b) NEW CABLES
c) HOLES ~ILLED IN 'THE
EXISTING BEAMS

SECTJ<lll

L ror1 L
l .I l
==--
--, Ia-
~~=
ANCHORAGE BY STEEL f--' ~
: """-.-
FRAME
u
"' : t:
i
N

--. ,.__
___, L-
--- l ~~

PLAN

113

L
'(
oo , l 88
PLAN
ANCHORAGE ßY CONCRETE BLOCK

FrG. 6. Detail of thc additional prcstressing of the beams.

- More than 90%; no additional cables were provided but the repair of
the beams' surfaces as weil as reinjection of the existing cables was
carried out.
In addition, a substantial variation in structure was adopted for all the
viaducts in order to improve their response to seismic actions.
The decks, originally simply supported and separated on each pile by
380 Mario P. Petrangeli

a) INITIAL SCHEME

ENERiY DISSIPATING DEVICE

b) FINAL SCHEME
FIG. 7

Jomts, were connected tagether by the slab made continuous on the

supports. All the bearings were changed, the new ones being free in the
longitudinal direction. Fixed points only now occur on the abutments where
energy-dissipating devices have been located.
These devices are of the elastoplastic type and have been designed to
remain in the elastic range for earthquakes with areturn period of 50 years;
they will yield for stronger earthquakes andin this case they will be replaced.
Obviously the alterations to the abutments were necessary in order to
improve their strength against the horizontal forces. This was done by
adding new structures on the back of the existing ones, as shown in Fig. 7.

CONCLUSION

The 25-year-old bridges studied were found to be in a bad condition due to

the extensive use of de-icing salt and to the absence of any kind of
 Inspeerion andRepair ol Some High11·ay Bridges in ltaly 381

waterproofing between the deck slab and the surfacing. Besides that a
dramatic reduction in the expected value of the prestressing force in the
beams was found. To quantify this reduction the dynamic tests extensively
used were of great importance since they allowed meaningful comparisons
between decks of the same type.
The addition of new cables was the solution for the cases in which the loss
ofprestress was in the medium range. Economic reasons, i.e. the high cost of
manpower and the high industrialisation reached in ltaly in the construction
of pc prefabricated bridge beams, mean that for the worst cases the
replacement of the old deck with a new one was more economical.
Better behaviour under seismic action was obtained by making the slab
continuous on the supports and by adding energy-dissipating devices on the
abutments.
 34

Inspection and Strength Evaluation of Concrete

Highway Bridges in Czechoslovakia

KAREL DAHINTER
Pragoprojekt-Design, Engineering and Consulting Inc.,
K. RySimce 16, 147 54 Prague 4, Czechoslovakia

ABSTRACT

This paper describes some experiences and results of the investigation and
evaluation of concrete highway bridges carried out under the state research
programme. The complex diagnostic method is explained and illustrative
examples of reinforced and prestressed concrete bridges are appended.

INTRODUCTION

There are about 46 000 highway bridges in Czechoslovakia with a span of

more than 2·0m, of which approximately two-thirds are plain concrete,
reinforced concrete (rc) or prestressed concrete (pc) structures. The first of
them were built at the beginning of this century, the main part of rc bridges
between World Wars I and II, and bridges with pc superstructures after 1950.
The management of all highway bridges is controlled by two standards:
ON 736220, 'Register of Bridges on Motorways, Highways and Urban
Roads', and ON 736221, 'Maintenance ofBridges on Motorways, Highways
and Urban Roads', and methodological guidelines appended to them.
According to these standards a physical condition is classified in seven
rating Ievels:

-perfect, very good and good, with surface impairments only and non-
reduced load-carrying capacity, when preventive maintenance is
sufficient;
-satisfactory and bad, with structural deterioration and reduced load-
carrying capacity, when rehabilitation is needed; and
383
384 Kare! Dahinter

-very bad and poor, with heavy deterioration and distress signs, when
immediate traffic measures or closing-up must be done and extensive
rehabilitation or replacement of the structure follows.
In the standards there arealso basic specifications for three types of live
loads for strength evaluation (legal weights of a vehicle in tonnes for dass
A/B bridges):
-the 'normal'-six two-axle vehicles in two lanes or three in one lane for
narrow bridges (32/22);
-the 'exclusive'-one four-axle vehicle (80/40); and
-the 'exceptional'-special set for heavy loads with two three-axle
tractors and one 14-axle trailer (280).
According to the decision ofthe central state authorities the load-carrying
capacity of all existing bridges should have been checked by the end of 1993.
This means that at first the condition survey and at second the strength
evaluation should be done. For that purpose and for increasing the
management level in general, special departments in Czech and Slovak
republics were established as weil as inspection teams in all regions.
As most of these bridges were designed according to former codes, with
lower loading actionsandmaterial properties, their load-carrying capacity is
not sufficient.
Due to this fact one of the state research programmes for the period
1986-90, directed by the Federal Ministry ofTransport, has been aimed for
'Research of methods and technological measures for increasing the load-
carrying capacity ofhighway bridges, both one-off and permanent'. The first
of four partial programmes contains 'Methods of physical condition
assessment' and was finished in 1988. The results are presented as follows.

COMPLEX DIAGNOSTIC METHOD

In general, the following investigations forming the complex diagnostics are

used.
Visual inspections according to the national standards: routine checking,
and main and extraordinary inspections carried out twice a year, once a year,
every 4 years and for special purposes. Same devices are used, such as
precision measuring equipment, crack microscope, telescope, etc.
Surveying of the shape, position, translations and deformations, when
both classical measuring devices and a Iaser beam are used.
Geotechnical and geophysical surveys are always needed, at least as a
reconnaissance in place with a geological data bank study. If necessary,
 Concrete Highway Bridges in Czechoslovakia 385

explanatory boring is carried out to sample the core (core cutting), using pipe
inspection TV camera with a video recorder and a device measuring strain
characteristics ofthe earth body in situ, or the seismic methods of geophones
or four-electrode resistivity measurement and stray current measurement.
Material testing includes the Schmidt hammer and CAPO test concrete
surface strength techniques, and the ultrasonic pulse velocity test to check
concrete quality. For checking the location and condition of reinforeerneut
gamma radiography, cover-meter, half-cell potential and electrical
resistance probes are used. Chemical effects of C0 2 carbonation and
chloride ions for both concrete and reinforeerneut are measured by means of
German instrument sets, 'Rainbow indicator' and 'rapid chloride test',
respectively. Very rarely the techniques of concrete core cutting or
determination of cement content are used, but always small holes are made
to evaluate diameter and to check corrosion of rebars.
Strength evaluation. Tables presenting the load-carrying data on
superstructures of precast units are available as well as on some in-situ
concreted reinforced concrete slabs. All other bridges are to be reanalysed.
Dynamical analysis is provided to determine the characteristics of
oscillation behaviour, natural frequencies and shapes, dynamical impact
and logarithmic damping decrement.
Loading tests are used if there are any doubts as to the reliability of the
structure. From the viewpoint of costs dynamical testing or response
measurement are preferred to the statical if possible. However, the statical
loading tests are in controversial cases irreplaceable.

The Results of the Complex Diagnostics

-Rating of physical condition with a Iist of damages and deteriorations.
-Determination of load-carrying capacity according to three types of
highway live Ioad.
-Recommendations for maintenance, repairs or rehabilitations if need
be.
The final technical report for the bridge containing these conclusions is
complemented by general arrangement drawings and photodocumentation.

CONCLUSION

The methods used for investigation were checked in field conditions, and
their advantages and disadvantages were considered from the viewpoint of
their reliability, speed, destructivity and costs. From the aspects of
assessment of the structure behaviour and condition visual inspections,
surveying and structural reanalysis with statical testing are the most
386 Kare/ Dahinter

important investigations. However, all other methods are important as well

to create a whole picture of the structure, physical condition and Ioad-
hearing capacity.

ILLUSTRATIVE EXAMPLES

1 Reinforced Concrete Continuous Girder Bridge over the Sazava River

near Kacov (Fig. 1) (completed 1914)
Deterioration and damage:
-water leakage and leaching due to damaged waterproofing membrane
in the connection between the deck and main girders with cracking in all
parts of superstructure;
-heavy darnage in the bottom of main girders near the central piers,
corroded rebars, spalling, scaling and delamination of concrete due to
long-term carbonation, weathering, attacks by leaking water solution
and, last but not least, the effect of heavy transport; and
-cast steel bearings on the abutments heavily corroded, damaged and
one of them fully loose, not working as a support.

L
I 20,25
t 2~.50
tI 20.25
]
-1
M.
N

-.:I
t--.
N

r
ci

FIG. I
 Concrete Highway Bridges in Czechoslovakia 387

Rehabilitation process:

~All deteriorated parts of concrete were removed, reinforcement cleaned

and fully corroded stirrups complemented.
~Special epoxy solutionwas used for penetrating paint and injection of
cracks.
~The missing parts of the concrete section were complemented by
epoxy-mortar.
~The bridge deck was strengthened by an epoxy-mortar layer 10 cm
thick, reinforced by welded mesh.

Camplex diagnostics and rehabilitation results:

After finishing the rehabilitation work a reanalysis was done to

determine the new load-carrying capacity ofthe bridge. The permissible
vehicle weights of 20/40/80 t were checked by staticalloading test and
the bridge was opened to public traffic.

2 Three-hinged Arch Reinforced Concrete Bridge over the Ohfe River in

Karlovy Vary (Fig. 2) (completed 1932)
Deterioration and damage:

~defl.ection in the top hinge of about 10 cm;

~water leakage, leaching, cracks and spalling in the region of expansion
joints, especially in the top hinge (Fig. 3(a));
~leaching in the constructionjoints ofthe deckandin the cable channel;
~corroded rebars, spalling and delamination of concrete of the deck
supports (Fig. 3(b));
~local darnage of the concrete surfaces of arches.

Camplex diagnostic process and results:

~Repeated surveying of the shape determined the long-term and

temperature defl.ection.
~Reanalysis of the structure on a space calculation model explained the
long-term defl.ection as a consequence of creep and shrinkage of the
arch after completion of the deck.
~Load-carrying capacity was calculated and performance of the
structure was checked by nine positions of loading vehicles during the
statical loading test; results were used for correction of calculation
models by reanalysis.
~Dynamical analysis and measurements were carried out to determine
dynamical performance.
388 Kare! Dahinter

82 .0 uoo~ 13.5 I
*

FIG. 2

(u) (h)

FIG. 3. (a) Corroded top hinge; (b) damaged deck support.

 Concrete Highway Bridges in Czechoslovakia 389

-Principal geological and geophysical surveys were executed to check the

foundation and corrosion conditions, including occurrence of stray
currents.
-Schmidt hardness tests, ultrasonic pulse velocity tests, cover-meter and
half-potential methods, and phenolphthalein carbonation tests with
Jocal probes were applied for material testing of the structure.
-Total rehabilitation of the bridge is indispensable and a preliminary
design was prepared.
-Up to the rehabilitation a posted traffic (20/40/ 100 t) with reduced
velocity (40 km/h) for a maximum of 2 years is permitted.

3 Precast Prestressed Concrete Composite 1-Girder Bridge over the

Berounka River in Beroun (Fig. 4) (completed 1954)
Camplex diagnostic process and results:

- The old masonry piers and abutments from the year 1864 show only
surface deterioration due to weathering and leakage of water from the
bridge roadway.

r-rn
~-------46 ~~~0~·-5~1---------
n .3~ __5__,.~2_.3_6_____,.33

FIG. 4
390 Kare/ Dahinter

-Most of the steel bearings, made of old rails, are slightly corroded, and
need cleaning and protection paint.
-Superstructure is in a very good condition, concrete strength more than
50 MPa, statical and dynamical performance according to loading test
fully corresponding with theoretical analysis.
- General rehabilitation ofthe roadway layers, waterproofing membrane
and expansion joints are needed.
-Repair of cornice, railing and endpart surface of the girder under
expansion joints.
- Load-carrying capacity needs no posting.

4 Precast Prestressed Concrete Flyover Crossing the DC Electrified

Railway in Pardubice (Fig. 5) (completed 1964)
Camplex diagnostic process and results:
- Very strong stray currents were measured.
- Special measures for application of haif-eeil potential methods were
examined to determine the corrosion state of reinforcement.

H131-1WAY
~y
RM..WAY , SffiEET

-- I I I
19 18 17 '6 15 14 I 13
I : I: ~ I l ~
2 11 I 10
li1" l : i
3 2 , PERS
3 :4--- I
I

~ 20 " 15.0 1 9 ..1o

F1G. 5
 Concrete Highway Bridges in Czechoslovakia 391

-Cracks in connecting joints between the box beam due to long-term

repeated Jive loading, especially in the area of quarter and three-quarter
cross-sections, cause lowering of the distribution of Ioads.
-Statical and dynamical reanalysis and dynamical testing are to the date
of compilation of this paper still in preparation.
-Special measures must be taken to protect the structure against
corroswn.
- General rehabilitation of roadway and waterproofing is intended.
-Supplementary cast-in-place reinforced concrete slab on the box beams
to impose the Ioad distribution is recommended.

5 Cast-in-Piace Prestressed Concrete Segmental Box Girder Bridge over

the Ohre River in Drahovice (Fig. 6) (completed 1960)
Complex diagnostic process and results:
- Physical condition is very good.
-Long-term deformation due to creep and shrinkage led to excessive
deftection of about 19 cm in the midspan which caused water draining
difficulties on the roadway.

gl u :t::i §t
-r-
-l7-'-'-
.7t- 1 _ _....:....;;
66..:..;_.6 _ __ +-'-
17.7--t-
1 ,:11!.3.3. 1 m3.3,
1 1 ,, .5

FIG. 6
392 Kare! Dahinter

-Unpleasant oscillation for users and considerably high Iive-load impact

were observed. Dynamical response was measured in 1960, 1963 and
1987; the firstnatural frequency (2·40- 2·33) and logarithmic damping
decrement (0·055- 0·036) remained principally unchanged. The ob-
served dynamical impact for one vehicle is up to 75%.
-General rehabilitation of the roadway layers with waterproofing
membranes and drainage system is intended.

6 Three-span Prestressed Concrete Frame Bridge with Cast in Place on

Centring Box Girder (Fig. 7) (com pleted 1971)
With regard to the experience from bridge 5 one twin box girder with
increased haunch height was designed, steel bearings were substituted by
rigid connection and elastic walls respectively, and an increased coefficient of
creep and shrinkage for the constructional Superelevation in midspan was
applied (total 280 mm). During construction one superstructure was
damaged and several all-section cracks were caused in midspan.

ll ~~ !'
~
11q 16,0· 1· 62.0 •
1
16.o
l~p
27,0/2
·I
u-s
I .
I

;:.I :::
rr
...111,5 I 71,0 N I
t~
11,51 ..l2·C1.f 8,50 12·~~

FtG. 7
 Concrete Highway Bridges in Czechoslovakia 393

Preliminary results of complex diagnostics:

-cracks fully closed with leakage signs, otherwise physical condition very
good;
-deftection to theoretical elevation of so und superstructure 80 mm
(superelevation), the damaged one ±0; and
-observed dynamical impact for sound superstructure up to 40%,
=
damaged up to 20%; natural frequencies .1; 2·5 and f~ 4·2, =
=
logarithmic damping decrement 0·07 (the same).
 35
Prestressing with Fibre Composite Materials and
Monitoring of Bridges with Sensors

REINHARD WüLFF and HANS-JOACHIM MIESSELER

Strabag Bau-AG, Siegburger Strasse 241, D-5000 Cologne 21, FRG

ABSTRACT

On several structures heavJ·-duty composite materiaL~ have proved their

applicability as a corrosion-resistant alternative to conventional prestressing
steel.
lt is also necessary to monitor these concrete structures on a permanent
basis in order to guarantee their durability over a lang period of time. In
prestressed structures cracks in the concrete and changes of the stress state in
the tendons must be observed.
Bridges can he monitored in this way using sensor systems 1vhich are
incorporated in the high-performance/ihre composite prestressing elements or
directly into the concrete.

INTRODUCTION

Progress in bridge construction is nowadays demonstrated by structures

which are becoming increasingly audacious. This is particularly evident in
the greater span widths of bridges, although the aesthetics of the actual
structures and their integration in the natural setting of the landscape also
play a decisive role.
At the same time, by making the optimal choice ofthe materials available,
one should attempt to achieve a considerable increase in the useful life of
these structures and to recognise any darnage to the load-bearing structure
at an early stage so that specific counter-measures can be initiated.
To meet these demands good results have already been achieved with the
aid of optical fibre and capacitative sensors, both in the case of fibre
395
396 Reinhard Walff and Hans-Jaachim Miesseler

TABLE 1
Prestressing with fibre campasites, material characteristics and camparisans

Reinforcing Prestressing Polystal" Twaron" Carbon

steel steel (68% glass (aramid fibres"
BSt SOOS St 1470/1670 fibres) fibres)"

Tensile strength (N/mm 2 ) >500 > 1670 1670 3150 2800

Yield strength (N/mm 2 ) >500 > 1470
Ultimatestrain (%) 10 6 3-3 2·0 0·7
Modulus of elasticity (N/mm 2 ) 210000 210000 51000 125 000 400000
Specific weight (g/cm 3 ) 7-85 7·85 2·0 1-45 1·75
Fields of application Reinforced Prestressed structures Stay cables,
concrete bracings
structures

"Material characteristics corresponding to the fibres.

composite materials in prestressed structures and also in the monitoring of

concrete load-bearing structures and prestressing tendons. According to the
glass, carbon or aramide fibre composition, and with full exploitation of the
individual material characteristics (Table 1), fibre composite materials open
a possibility for the fulfilment of the most differing requirements.
Previous utilisation of fibre composite materials mainly concemed the
glass fibre composite material which had till then been applied in the form of
prestressing tendons for prestressed structures, tensioned mast bracing and
as tie-rods for the rehabilitation of arches.
Further monitaring of the performance of the structure is made possible
by integrating optical fibre and copper wire sensors in the tensioning bars.

THE PRESTRESSING TENDONS (HLV TENDONS)

Two companies-Strabag Bau-AG, Köln, and Bayer AG, Leverkusen-
have been working on the development of glass fibre composite bars since
1978. These bars (trade name 'Polystal') have a diameter of 7·5 mm and
comprise 60 000 E-glass fibres embedded in unsaturated polyester resin with
a thickness of just a few microns strictly oriented along the direction of the
bar. The fibre content is around 68%.
The tensile strength of the Polystal bars is comparable to that of high-
grade prestressing steels, roughly 1670 N/mm 2 , but in the other properties
Polystal shows significant advantages. The composite prestressing elements
are considerably lighter than their steel equivalents (Polystal has a density of
2 g/cm 3 versus 7·85 g/cm 3 forthat of steel), are electromagnetically neutral
and can be used in aggressive environments. Protection of the glass fibre
against chemical attack and mechanical darnage has been assured by a
specially developed coating technique.
 Prestressing with Fihre Composite Materials 397

-... ... ec""""'•'

2000

~
!':-st1470f1e70
I- -·
~ r--

~r-
..Dnl YJf'fN...
- • 11m1 Nmwn• -
E • 51000Ntmm'
fu• 3.3~ -
--
aoo

400
~ .. :c BSt~
;-
0
0 • 8 12 18
~8('1

FrG. I. Stress- strain diagram of a PolystaF~ bar in comparison with steel.

At 51000 N/mm 2 the modulus of elasticity is only a quarter that of steel;

tension force Iosses arising from creeping and contraction of the concrete
drop correspondingly to one-quarter. In contrast to conventional steel, the
composite glass fibre material has no plasticity Iimit; the stress/strain
behaviour of Polystal is linear practically up to the point of failure (Fig. 1).
In the case of structural elements which are prestressed with steel, high
degrees of plastic deformation occur when the plasticity Iimit is exceeded,
providing the desired warning prior to the incidence of failure. Trials on
structural elements have shown that, as a consequence ofthe low modulus of
elasticity, a structural element prestressed with Polystal bars possesses a high
degree of elastic expansion when permitted Ioads are exceeded. These
deformations also give a prewarning of failure (Fig. 2).
A total of 19 ofthese glass fibre bars are bundled to form one prestressing

AMtlc defoml8tion of
~steel

FIG. 2. Load- deformation rclation of a prestressed concretc structure.

398 Reinhard Woljf and Hans-Joachim Miesseler

tendon with a working Ioad of 600 kN. In the field of anchorage engineering
completely new solutions had to be found because the composite glass fibre
material will only bear transverse pressure to the extent of 10% of its
longitudinal tensile strength. Thus the Iack of cold workability rules out the
utilisation of upset heads, rolled-on threads or even the utilization of steel
wedges 'biting' directly into.the 'soft' composite glass fibre bar material. The
relatively low interlaminar shearing strength of the resin matrix requires a
comparatively greater anchorage length than would be the case with steel
prestressing tendons. Anchorage elements specific to the material had
therefore to be developed for the anchorage of this new material.

FIG. 3. Anchor head of a compositc glass fibre prestressing tendon.

The development of a tubular grouted anchorage (Fig. 3) by Strabag Bau-

AG heralded a breakthrough for the anchorage ofHLV composite bars. The
composite bar is set inside a sectional steel tube in an artificial resin-based
grouting mortar which has been specially developed for this purpose. The
use of these grouted prestressing tendons covers the entire spectrum of the
lightweight and mediumweight prestressing tendons (up to 1000 kN
working Ioad) and the complete field of soil and rock anchoring.
 Prestressing with Fibre Composite Materials 399

THE SENSORS

Nowadays increasingly high demands are made on monitaring methods for

the assessment and control of the stress/strain behaviour of concrete
structures. With the use of glass fibre composite bars it is possible, for the
first time, to integrate sensors capable of indicating the stress state for the
monitaring and control of the whole structure.
Today two different methods of monitaring (based on two different types
of sensors) are carried out successfully.

Optical Fibre Sensors (Fig. 4)

This type of sensor is a specially coated optical fibre covered with a spiral
ofthin wire. The pitch of the spiral is so dimensioned that, if there is axial
Stretching of the sensor, the wire spiral will press on to the optical fibre and
generate micro-bendings in it. These influence the permeability of light,
which is directly proportional to the sensor's tensile strain. These
circumstances also allow the localisation of a failure, which up to now was
not possible with the employment of Ioad cells. The sensor, which is
approximately 2 mm thick, is either embedded in the centre of a glass fibre
bar or directly in the concrete, or both.

Copper Wire Sensors (Fig. 5)

The individual bars of the prestressing tendons are permanently
monitared by copper wire sensors incorporated in the cross-section of the
composite bars, with four copper wires in each bar. As with a capacitor, an
electric field is thereby generated between the respective wires located in the
centre and one of the wires arranged externally. The glass fibre material
between the wires (a dielectric) is defined by the dielectric constant.
Assessment of the change in strain in the prestressing tendons is now based

rrcmer

0 0 .
I'IICII hhr IUIII

FIG. 4. Diagrammatic illustration of thc mcasuring and monitoring systcm using optical
fi brc scnsors.
400 Reinhard Woljf and Hans-Joachim Miesseler

._ur""' •f "'- , ...,"J ~ "- •oM~ fii i ,~IH( CII'O• I•"I I

...... l'tle - ' " - · .. .,. (Mfftl ..,.. ..,".,.. •ft<CI
11 lf'l•tc•t • '~~~"• IN•t • ,..,. rvptwe ::.:.:u ~!
.... . '·'
I •' r••--'••Mr•liion l

FIG. 5. Sketch showing the principle for the monitoring of prestressing tendons using
copper wire scnsors.

on the principle of the change in capacity produced by means of the centrat

copper wire relative to the others; the change in capacity and the change in
strain have a mutually proportional behaviour.

PRACTICAL EXAMPLES

The Ulenbergstrasse Bridge in Düsseldorf (Fig. 6), the first prestressed

concrete bridge designed worldwide for extremely heavy road traffic Ioads
(bridge dass 60/30), is a demonstration construction project for the new
material Polystal.
lts outline design provides for a 15m wide two-span solid slab bridge with
spans of 21· 30 and 25·60 m, and a slab thickness of approximately 1-44m.
In the longitudinal direction the bridge is prestressed by the use of 59
hea vy-duty composite prestressing tendons (working Ioad 660 k N each) with
19 heavy-duty composite bars per tendon. The bridge is prestressed by post-
bonding achieved by means of an artificial resin-based grouting mortar
specially converted for this purpose.
With regard to bridges, the next application was the construction of the
pedestrian bridge to Marienfelde Ieisure park in Berlin in 1988 (Fig. 7). This
is a two-span double-webbed slab/beam bridge with spans of 22·98 and
27·61 m; its superstructure has an overall height of 1·10 m and a slab width of
4·80 m. This bridgewas executed with partial prestressing without bond for
the first time in G ermany. A total of seven prestressing tendons with a
working Ioad of 600 kN each run underneath the spans, around each of the
 Prestressing with Fibre Composite Materials 401

FrG. 6. Ulenbcrgstrasse Bridge, overall view.

two transverse beams in the bay section, and then upwards along the central
column anc over the transverse beam.
The most recent application to date has been the stabilisation ofthe arch
ofthe 'Marie d'Ivry' metro-station in Paris (Fig. 8). Excavation in the vicinity
of the tunnel tube resulted in ground settlement, which in turn caused the
arch of the station's cross-section to crack. In order to improve the safety

FrG. 7. Berlin-Maricnfcldc Bridge, overall view.

402 Reinhard Wo(ff and Hans-Joachim Miesseler

FIG. 8. 'Marie d'lvry' metro-station in Paris, tie-rod view.

required for the arch, the dient decided on a tie-rod comprising glass fibre
composite prestressing tendons. A total of 36 tendons with a working Ioad of
650 kN each were installed in April 1989.

REFERENCES

I. WAASER, E. and WOLFF, R., Ein neuer Werkstoff für den Spannbeton, HLV-
Hochleistungs-Verbundstab aus Glasfasern. 'Beton' 36 (1986).
2. MIESSELER, H.-J. and PREIS, L., Hochleistungsverbundwerkstäbe als Bewehrung
im Beton- und Erd bau. Bauen mit Kunststoffen, 2 (1988) 4-14.
3. KöNIG, G. and WoLFF, R., Hochleistungsverbundwerkstoff für die Vorspannung
von Brückenbauwerken. IABSE Congress, Paris-Versailles, 1987.
4. LESSING, R. and MIESSELER, H.-J., Experiences in monitoring of Ioad-hearing
structures with optical fiber sensors. IABSE Congress, Lisbon, 1989.
ASSESSMENT AND EVALUATION
 36
Bridge Capacity Assessment and Control of Posting,
Permit and Legal Vehicle Loads

FRED MosEs
Department of Civil Engineering, Case Western Reserve University,
Cleveland, Ohio 44106, USA

ABSTRACT

Several recent projects conducted by the author haue concerned new methods
for bridge evaluation, including strength capacity and safe life assessment.
These projects haue led to new AASHTO proposals for capacity evaluation
andflexible methodsfor regulating saf'e Ioads on bridges. Economic pressure to
reuise truck weight regulations haue also heen considered. Proposals for
evaluating permit trucks for d(fferent highway classifications haue been
reported. A comprehensive study of' optimal truck ~veight regulations to
balance vehicle productivity with increased costs for hridge repairs and
replacement haue heen studied. These projects haue been supported by the US
Transportalion Research Board and various state agencies.

INTRODUCTION

The emphasis for the highway industry in the United States has shifted to
maintenance, rehabilitation and conserving the existing road network.
Bridges are a vitallink in the highway system and, in part, because of their
conservative design, bridges have been allowed to deteriorate over many
years because of deferred maintenance and repairs. There are some 600 000
bridges in the United States under a wide variety of ownerships and control.
The Federal Highway Administration now estimates that more than 200 000
bridges are inadequate and lists over 125 000 as structurally deficient on the
National Bridge Inventory System. There are some 5000-8000 replacements
per year, so that for the foreseeable future the inventory of bridges in the
United States will contain numerous structures incapable of carrying
today's standards of truck weight.
405
406 Fred Moses

In the present bridge inventory, about halfthe structures are more than 50
years old, which partly explains why so many are deficient. During recent
decades, truck weights and volumes have grown enormously, while funds for
inspection, maintenance, repair and rehabilitation were often not available.
Despite this situation, bridges have maintained relatively high safety records
because traditionally engineers used conservative methods of design which
produced high Ievels of reserve strength. With increasing truck loadings,
these reserves are eroded and many bridges must be replaced or eise Ioad
posted for restricted traffic usage.
At the sametimethat budgetary pressures are restricting the upgrading of
the bridge system, there are numerous proposals to allow increased demand
on the system. Although the US Interstate System allows a legal Ioad of
80 000 lb (356 kN), many states permit heavier Ioads. This is either done
through a liberal permit system or eise a 'grandfather' exemption clause
which allows higher loading for certain classes of vehicles. A review of state
regulations indicates over half utilize exemptions to allow vehicles on a
routine basis up to 140000lb gross weight (623kN). Requests for special
permits fortruck movements even above 300000lb (1330kN) have also
become relatively routine.
Two important aspects of bridge safety are impacted by these heavy
vehicles, including the overall capacity of the structure to withstand the
heaviest Ioad combinations and also the reduced remaining life that may
occur due to repetitive loadings which induce a cumulative fatigue damage.
Because of the wide variety of structure type, span, material, geometry, age
and design Ievel the response of bridges to new truck demands will vary.
In recognition of both the increasing safety risk in bridges and the !arger
demand being placed on the system, there has been considerable research
sponsored in recent years to promote better understanding of bridge Ioads,
response analysis and strength capacity, and fatigue life. This research had
led to improved methods of Ioad capacity evaluations and bridge
management policy to optimize utilization and conserve resources. This
paper reviews recent work by the author in several ofthese areas. The topics
to be covered include:
(a) guidelines for Ioad capacity evaluation,
(b) guidelines for estimating safe remaining Jives,
(c) evaluating heavy vehicle permits, and
(d) impact analysis of proposed truck weight regulations.

SAFETY METHODOLOGY

In the applications that follow, a consistent safety strategy is needed for

analysis. This is necessary because traditional calculation methods by bridge
 Bridge Capacity Assessment and Control of Loads 407

engineers have widely varying Ievels of conservativeness contained within

their design procedures. This variability in the safety Ievel may be due to
selection of Ioad criteria, analysis techniques or assessment of safety factors
for structure components and systems. Unlike pavement or roadway
management programs which optimize only the economic components of
the pavement, bridge management must relate to safety as well as economy.
It is well recognized that traditional 'safety factors', whether in bridges or
other structures, are not absolute criteria for safety. A major development
for structural codes in recent years has been the introduction of structural
reliability concepts to assist code writers to formulate safety checking
models and derive appropriate safety factors. These reliability models
recognize the uncertainties in modeling extreme Ioad events, inexact analysis
predictions, and the variability in material and system strength capacity. The
goal is to establish a reliability or safety index which incorporates the actual
safety margins and the uncertainties or randomness in the strength, Ioad and
analysis procedures. Safety margins are calibrated to ensure that target
reliability indices are satisfied. To the practising engineer, such reliability
based codes retain the traditional format of tabulated safety factors.
Although instituted worldwide for design codes, reliability is also being used
for evaluation procedures for existing structures. The uncertainties in
evaluation are naturally different from those encountered in design
situations.
Currently, guidelines for evaluating existing bridges are provided in the
AASHTO Manualfor Maintenance lnspection of Bridges. 1 Recommenda-
tions for introducing flexibility into the evaluation procedures are given in
general qualitative terms. Quoting directly from the manual, ' ... a higher
safety factor for a bridge carrying a !arge volume of traffic especially if it
includes many heavy Ioads ... '. Similarly, 'factors of safety used in rating
must provide for reasonably possible overloads ... and Iack ofknowledge as
to the distribution of stresses'. Those quotes demonstrate the awareness of
differences between design and rating, and that the latter should recognize
that an existing structure should be influenced by observations and analysis
that can directly be made. Nevertheless, the checking procedures in the
manual are deterministic and do not take advantage of the apparent data
base of the actual structure.
The intent of structural reliability theory is to characterize the
uncertainties in Ioad intensity, Ioad effect analysis and strength capacity, and
then allow for consistent and rational codified safety decisions. These
methods have been adopted in many codes in the United States where it is
known as Ioad and resistance factor design (LRFD), in Canada and the
United Kingdom where it is known as Iimit state design (LSD) andin other
countries where it is called a partial safety factor format. To make it feasible
for implementing reliability in rating a !arge number of structures, it is
408 Fred Moses

necessary (a) that the structures will be analysed by conventional methods

and (b) that Ioad and resistance factors be tabulated in a form that the
engineer need not be concerned with probability or reliability theory in
performing the checking equation.
The opportunity to introduce a comprehensive reliability approach to
evaluating existing bridge st~uctures is due to several circumstances. Recent
code changes, such as the Ontario Highway Bridge Design Code introduced
in 1979, 2 demonstrated that formal reliability methods can be used to
calibrate safety factors based on uncertainty Ievels for all components in the
design and evaluation process. Further, extensive field performance on
bridges is available ranging from in-service weigh-in-motion (WIM) 3 to
full-scale ultimate capacity determination. 4 These data show that current
design and evaluation parameters represent idealizations and approxi-
mations that, while broadly applicable to design, do not necessarily apply to
specific conditions of the bridge being evaluated. The flexibility of acquiring
and using site-specific inspections, and statistical loading and analysis
information, during the evaluation calculations become readily apparent.

LOAD CAPACITY EVALUATION

In the United States, all existing bridges must be Ioad evaluated on a

frequent periodic basis and recorded in the national inventory. This
evaluation or rating follows a physical inspection of the bridge's condition.
Following the evaluation, a bridge found deficient may be posted for
reduced legal traffic Ioads or else scheduled for rehabilitation or
replacement. The AASHTO manual allows evaluations at an inventory level
corresponding to the design stress condition (55% of yield stress) or at a
higher operating Ievel, typically 75% of yield stress. The latter is used by
many but not all jurisdictions in its posting decision. The selection of rating
stress level is usually fixed for all structures by an agency and does not
provide flexibility to account for the actual condition of the structure or to
quantify other important site or engineering efforts that might be considered
in the ra ting process.
As a consequence of these limitations, the author and his colleagues
introduced a reliability based LRFD oriented rating procedure. 5 •6 These
have recently been approved by AASHTO as a guide specification. 7 The
procedures utilized reliability theory and an existing data base to calibrate
load and resistance factors which are intended to achieve a uniform and
consistent evaluation procedure. Resistance factors ranging from 0·55 to
0·95 are affected by the bridge condition survey or level of deterioration
 Bridge Capacity Assessment and Contra! of Loads 409

noted, inspection and maintenance effort, and especially by the presence of

redundant load paths. Dead load factors are influenced by direct field
determination of overlay thickness. Live Ioad factor selection ranging from
1·3 to 1·8 is based on traffic volume, sources of overload or control ofheavy
truck traffic and structural analysis method used. Impact values depend on
deck roadway roughness and range from 0·1 to 0·3. Bridges with computed
rating factors below 1·0 must be load posted.
With the factors tabulated in the guide specification, bridges may reach or
even exceed previous operating ratings for those bridges which receive
frequent qualified inspection and have adequate maintenance programs and
Ioads corresponding to reasonable Ievels of traffic and enforcement.
Conversely, bridges which do not maintain these conditions or have
nonredundant components will find their ratings falling possibly to
inventory Ievels or even lower. Evaluators will find options in these
guidelines by which ratings can be improved by recommendations for more
frequent and detailed inspection and maintenance, improved structure
analysis and especially control ofheavy overweight vehicles. In all instances,
it is expected that the use of the tabulated factors will provide rating
evaluations which meet the target reliability Ievels. The target Ievels used in
establishing the factors were calibrated to reliability analysis with specific
field performance experience. 8

SAFE REMAINING BRIDGE LIVES

In most specifications, fatigue checking provisions are given in the form of

allowable stresses which must be satisfied by a designated loading case. Such
provisions do not indicate how to calculate the remaining life of an existing
bridge which is needed to make cost effective decisions regarding inspection,
repair, rehabilitation and replacement. Many agencies indicated in a survey
that remaining life calculations are needed especially to assess rehabilitation
options and in some cases to assist in making bridge posting decisions.
The author and his colleagues recently presented methods for assessing
remaining life of steel bridges which were accepted by AASHTO as a guide
specification. 9 - 11 The proced ures in these guidelines used pro ba bilistic
concepts which provide realistic description of the fatigue conditions in a
bridge. Appropriate target reliabilities were calibrated for both redundant
and nonredundant categories. Further recommendations were given to
quantify the effect of different Ievels of effort in reducing uncertainties
(analysis or site data acquisition) and thereby improve the predictions of
remaining life. To simplify implementation, definitions of detail categories
410 Fred Moses

were the same as present AASHTO rules. The options available to the
evaluator if the computed remaining life is inadequate include;
(1) calculating fatigue life more accurately with recommended
procedures;
(2) restricting traffic on the bridge;
(3) repairing the bridge; or
(4) instituting periodic detailed inspections.
The advantages of the procedures developed in the remaining life study
include:
(1) Methods which realistically reftect actual fatigue stress conditions in
bridges.
(2) Consistent procedures for both design and remaining life evaluation.
(3) Suitable flexible procedures for using site data.
(4) Allow record keeping of remaining life to be periodically updated.
(5) Extensive use of recent traffic and fatigue research.
(6) Reliability targets calibrated from current performance.
(7) Use of appropriate Ievels of analysis depending on application.
(8) Do not significantly depart from current methods of fatigue design.
Numerous examples and illustrations ofthe fatigue evaluation (as well as
corresponding design provisions) were provided.

PERMIT REVIEW

Permits for overloaded vehicles are frequently issued by highway agencies in

the United States and elsewhere. Most states face increasing pressures to
allow heavier and greater numbers of overweight truck permits. This raises
questions regarding bridge safety and management policy and especially the
methods for reviewing bridges for permit Ioads. Typically, permit checks
utilize the stress Ievels corresponding to operating Ievels (75% ofyield) in the
AASHTO specifications or, in some cases, agencies permit even higher stress
Ievels. It is clear that such simple allowable stress checking methods may not
produce uniform reliability due to considerations of bridge dead load,
bridge geometry and normal truck traffic. Reliability based load factors offer
advantages for more uniform permit checking rules. Presently, AASHTO is
reviewing reliability based load and resistance factor methods as a possible
alternative in its bridge design manual, and as stated above has reviewed
favorably its use in bridge evaluation methods.
A project conducted by the author and his colleagues is under way to
derive appropriate Ioad factors for permit decisions. 12 Among the situations
 Bridge Capacity Assessment and Control of Loads 411

considered are permits for (a) single passage (no control of movement), (b)
single passage with speed and lane position control, and (c) routine
overloads with no restrictions on frequency of movements.
The maximum Ioad effect may occur with apermit vehicle simultaneously
on the structure with one or more adjacent random heavy vehicles. The
occurrence probability of such multiple presence events depends on traffi.c
volume, bridge span and geometry, speed and control of permit vehicles.
Vehicle arrival simulation models have been developed to predict the
maximum loading response and the corresponding distribution of Ioad
effects (fatigue life impact is considered separately in the permit study).
Using target reliability Ievels calibrated from other bridge safety studies, it is
possible to determine appropriate live Ioad factors for permit checking.
These permit review methods are being currently put in a specification
format for appropriate review. The aim is to maintain within a bridge
management system options for assisting in moving specialloads without
introducing unacceptable bridge risk Ievels.

TRUCK WEIGHT REGULATIONS

Many industrialized countries face increasing political pressures to raise

current allowable legal vehicle Ioads. The economic incentives in terms of
increased productivity appear to be large since for a small percentage
increase in fuel consumption a much greater percentage increase in payload
and profit is possible. The considerations for increased truck Ioads are based
on safety, pavement darnage and risk to bridges. One proposal in the United
States which minimizes pavement darnage is the so-called Turner proposal.
This plan increases the number of vehicle axles which distributes the Ioad
and reduces pavement darnage (since the latter is approximately propor-
tional to axle weight raised to the fourth power). The effect on bridges of
spreading the Ioad is minor, and unless overall truck lengths are increased
bridges will be adversely impacted by any heavier truck loadings. Length
increases usually are restricted by problems of vehicle maneuverability and
safety. lt is possible, however, that Turner vehicles will provide enough
savings in pavement darnage to offset the larger bridge costs. In an effort to
define the bridge costs, the author and his colleagues have participated in a
recent TRB and FHWA study. 13 The main goalwas estimating bridge costs
that are likely tobe incurred by state agencies following the introduction of
new vehicle regulations. Sensitivity studies to compute costs if agencies
adopted new bridge evaluation methods such as described above were also
considered and are discussed below.
Bridge costs were divided into four categories. These include (a)
412 Fred Moses

overstress-new bridges, (b) overstress-ex1stmg bridges, (c) remammg

lives-new bridges and (d) remaining lives-existing bridges. Over ten
different proposed truck weight regulations (or seenarios as they are called)
have been studied, including Turner configurations (increased weight with
increased number of axles), as weil as other variations of proposed weight
rules such as the Canadian interprovincial proposal, the TTI model,
extensions of the existing federal highway bridge formula and proposals
suggested by various trucking associations. In each scenario, the costs were
evaluated for each of the cost categories.
For new bridges, an estimate was made that $3 billion per year is now
needed for new bridge replacements and this sum would be affected by any
new regulations. That is, design Ioads would have to be increased which
would raise costs. For example, we found that increases from present HS20
design Ioad to HS25 (25% increase in Ioads) would raise average new bridge
costs by about 4%. Applying such extrapolations to the Ioad impact of each
proposed regulation gave an estimate of increased bridge costs for new
structures.
The major cost item, however, proved tobe the impact on existing bridges.
Using the FHWA computerized NBIS, a breakdown was performed ofthe
almost 600 000 existing bridges to produce distributions by category of span,
material, type (single or continuous) and by current operating load capacity.
The present deficiencies under current weight laws were first identified. The
next step in each scenario was to compare for each category the live Ioad
effect for the new scenario with that of existing weight regulations. Hence,
the number of additional bridge deficiencies was estimated and the
corresponding replacement costs computed. (In a survey, states indicated
that replacement is often more economical than rehabilitation whenever
strength upgrading is required.) The costs for the replacement of deficient
existing bridges provide a major part ofthe overall cost impact.lt was found
that steel bridges incurred about two-thirds of the total cost impact
(although they are a lower percentage of the population) with reinforced
concrete and prestressed concrete the remainder. Continuous span bridges
were a significant component of the cost, especially for truck vehicle
proposals, which tend to cause large increased moments in the negative
moment regions.
In general, it was found that the important factor in any new weight
regulation was the proposed gross weight and the corresponding vehicle
length (exclusive of distance from steering axle to drive axle). Some of the
seenarios were found to more than double the existing number ofstructural
deficiencies (currently about 125 000 out of the total population) with
corresponding increases in bridge replacement costs. Breakdowns of impact
by interstate and primary routes and also nonprimary routes was also done.
 Bridge Capacity Assessment and Contra! of Loads 413

A sensitivity study also found that any deterministic changes in bridge

evaluation procedures are not likely to cause significant reduction or
mitigation of the impact of new weight regulations. For example, more
liberal stress Ievels in evaluation rules will reduce the number of currently
deficient bridges but many of these same bridges will then become deficient
under the proposed new regulations.
The above-cited study used a distribution of current Operating Ioad
capacity Ievels as a measure of evaluation. These are, in current practice,
deterministic stress criteria. A study now under way by Dr Michel Ghosn of
the City University ofNew York with the assistance ofthe author is looking
at modeling the bridge population according to a distribution of reliability
indices. 14 Hence, bridges would be replaced under either present truck
regulations or the new regulations only if their reliability Ievel was deficient.
As described above, the bridge reliability index depends on traffic volume,
methods of analysis, and bridge condition and maintenance. It appears
possible to create a new truck weight formula which trades off increased
reliability for shorter span bridges (where most deficiencies now occur) with
lower reliability Ievels for Ionger span bridges. The latter structures have
high dead/live Ioad ratios and have been found in reliability studies to often
exceed required target reliability levels. 8 Such a trade-off in formulating
truck weight guidelines could produce a new truck weight regulation which
has relatively small cost impact on bridges but does increase truck
productivity, i.e.longer vehicles may be allowed heavier Ioads than presently
allowed with some reduction in gross weight for shorter vehicles. These
studies are still under way but some preliminary results support these
conclusions.
The other two cost items mentioned above werein the fatigue life areas. A
model of the distribution of steel span bridges was created which reflects
their fatigue life. The model was calibrated to the present annual
expenditures on fatigue related problems and associated darnage costs in
steel bridges. New weight regulations (both weight and volume changes are
needed for fatigue) were used to find the expected corresponding future costs
in fatigue life.
The impact of these fatigue costs was studied with different discount
factors and other variables. The overall cost impacts, although significant,
were generally small compared to the cost effect on the strength capacity of
existing bridges. Similarly, the last cost item, namely cost impact of fatigue
rules on new steel bridges, was small. That is, required strength increases to
carry heavier vehicle Ioads also implicitly increases fatigue life so that this
additional cost item appeared small. Further study, however, is under way of
possible impacts on acceptable fatigue details under heavier loading
conditions.
414 Fred Moses

CONCLUSIONS

A variety of bridge studies are discussed above related to bridge safety,

evaluation, fatigue life estimation, review ofheavy overload permit requests
and consideration of new truck weight regulations. Details of these studies
are contained in the references and guideline specifications that have arisen
out ofthis work. A common thread in these studies is a need for a systematic
and defensible approach to bridge safety. Risk analysis has now become an
accepted tool for code writers and reliability based design procedures a
technique for implementing uniform safety criteria. Such work needs tobe
elaborated on so that confidence by users is increased. Risk criteria or
corresponding reliability indices become an important basis for a rational
bridge management policy. Otherwise, engineers are continually being asked
to stretch the Iimits of the bridge system by permitting heavier permit
loadings or even expanded truck weight regulations. Allowable stress
procedures fail as criteria for bridge safety management since proponents of
higher Ioads are aware ofthe conservative nature ofthese stress calculations.
Bridge tests have shown that predicted stresses are often much lower than
measured stresses. Also, proponents ofhigher Ioads claim that if a high stress
is acceptable in one instance it should be acceptable in all cases. A reliability
approach serves to incorporate all aspects ofthe load~capacity equation and
identify the uncertainties which merit the requisite safety margins to avoid
undue risk. Applications of reliability analysis described herein for Ioad
capacity evaluation, assessment of safe remaining life, permit review and
recommendations for truck weight regulations have reinforced past
engineering judgment but also allow recent research and test findings to be
incorporated.
Recent developments in the United States and Canada in reviewing and
adopting Ioad and resistance factor formats for both bridge design and
evaluation show the promise of such procedures being incorporated within a
bridge management system.

ACKNOWLEDGEMENTS

The author wishes to acknowledge colleagues and associates who

participated in these efforts. In particular, at Case Western Reserve
University, Dr D. Verma assisted with the Ioad capacity studies, K. S. Raju
with the fatigue life assessment and Dr G. Fu with the permit studies and
truck weight regulations. Dr Michel Ghosn also helped in the basic
reliability modeling and analysis for several projects. C. Schilling, consulting
engineer, helped with all aspects of fatigue modeling. The firm of A. G.
 Bridge Capacity Assessment and Contra/ of Loads 415

Lichtenstein made several important contributions in development and

implementations, and lmbsen and Associates was a valued collaborator in
the Ioad capacity studies. Finally, the important help oflan Friedland ofthe
NCHRP staff, Harry Cohen of the TRB, and L. I albert and V. Dalal of the
Ohio Department of Transportation are appreciated.

REFERENCES

I. Manual of Maintenanc·e lnspection of Bridges. American Association of State

Highway Transportation Officials, Washington, DC, 1982.
2. Ontario Highway Bridge Design Code. Ontario Ministry ofTransportation and
Communication, Downsview, Ontario, Canada, 1983.
3. MosEs, F., GHOSN, M. and GoBIESKI, J., Evaluation of steel bridges using in-
service testing. Presented at TRB Annual Meeting, Washington, DC, January
1986. Published in Transportation Research Record 1072, Washington, DC.
4. BAKHT, B. and CSAGOLY, P. F., Diagnostic testing of a bridge. ASCE Journal of
Structural Engineering, 106 (7) (J uly 1980).
5. MosEs, F. and VERMA, D., Loadcapacity eva1uation of existing bridges. NCHRP
Report 301, NCHRP 12-28(1) Final Report, Transportation Research Board,
Washington, DC, December 1987.
6. VERMA, D. and MosEs, F., Calibration of a bridge strength evaluation code.
ASCE Journal of Structural Engineering, 115(6) (June 1989).
7. Guide Specification for Strength Evaluation of Existing Steel and Concrete
Bridges. AASHTO, Washington, DC, 1989.
8. GHOSN, M. and MosEs, F., A reliability calibration of a bridge design code.
A SCE Journal of Structural Engineering, 112(3) (1986).
9. MosEs, F., ScHILLING, C. G. and RAJU, K. S., Fatigue evaluation procedures for
steel bridges. NCHRP Report 299, Transportalion Research Board, Washing-
ton, DC, November 1987.
I 0. MOSES, F., SCHILLING, C. G. and RAJU, K. S., Reliability-based bridge life
assessment. ICCOSAR '89, San Francisco, May 1989.
11. Guide Specification for Fatigue Evaluation ofExisting Steel Bridges, approved
by AASHTO, Washington, DC, 1989.
12. MosEs, F. and Fu, G., A reliability analysis of permit Ioads on bridges. In
progress, report to Ohio Department ofTransportation, Case Western Reserve
University.
13. MosEs, F., Effects on bridges of alternative truck configurations and weights. In
progress, National Academy of Science, Transportalion Research Board.
14. GHOSN, M., Bridge overstress criteria. In progress, sponsored by Federal
Highway Administration at City University of New York.
 37
The Use of Reliability Analysis in the Assessment of
Existing Bridges

CAMPBELL MIDDLETON
Department of Engineering, Cambridge University,
Trumpington Street, Cambridge CB21PZ, UK
and
ANaus Low
Ove Arup & Partners, 13 Fitzroy Street, London Wl P 6BQ, UK

ABSTRACT

Aspart ofa contract for the U K Transport and Road Research Labaratory the
authors haue developed procedures for assessing the reliability of some
common types of concrete bridges.
Initially upper-bound plastic assessment techniques were developed which
could be used as an alternative to lower-bound elastic or non-linear finite
element methods to estimate the collapse strength of a bridge. However, this
deterministic technique makes no provision for the many uncertainties facing
an engineer when assessing an existing structure. These could concern the
strengths of materials used, the deterioration of these materials, the Ioad
history and the form of hidden parts of the structure.
Structural reliability theory allows the effects of such uncertainties to be
assessed rationally. An advanced Ievel II reliability procedure following the
method outlined in CIRIA Report No. 63 (The Rationalisation of Safety and
Serviceability Factars in Structural Codes, 1977) was applied to the plastic
collapse analysis to determine the national probability offailure of a range of
bridge types and dimensional configurations.

1 INTRODUCTION

Reliability theory has not as yet been widely used for the assessment of
bridges. The complexity of the computations, limitations in the theory and
417
418 Campbell Middleton and Angus Low

general unfamiliarity have all been constraints on its use. With recent
extensions in the theory 5 and readily available computing power it may now
be possible to develop practical procedures to complement the more
traditional methods used in the assessment of existing bridge structures.
In this work a number of possible plastic collapse modes were examined
for each ofthe chosen bridge types. This provided a deterministic calculation
ofthe strength ofthe bridge in relation to the specified loading and predicted
the most likely collapse mode geometry. The probability of this failure
occurring was then derived using reliability theory which examined the
statistical variability of each of the basic parameters that contributed to the
loading on the bridge and its strength to resist the applied Ioads.
The notional probability offailure for a bridge (typically 10- 5-10- 25 ) has
little relevance to the reallikelihood of collapse as most actual failures relate
to gross errors or catastrophic events. However, the relative values can be
used to rank the safety of a number of bridges and thus provide a basis for
assessing the priorities for remedial work or bridge replacement in relation
to different bridge configurations of deck type, span length and width. In
addition, by using these methods an engineer assessing a complex bridge can
explore the sensitivity of its reliability by making allowances for his
knowledge or ignorance of the strengths and variability of individual
components.
The methods described here are applicable for use in assessing both newly
designed and existing structures, provided some details of the dimensions,
material types and reinforcement are known.

I• 9.3 n g. )
•I
os., "'~'
1·3 1.s11 o.s
C:? r.:- c:'l
,.. 25.1 I I
OX!kx<&kXU
I•
11.3
• I

FIG. 1. Dual two-lane portal slab bridge. Fru. 2. Single two-lanc M-beam bridge. All
All dimcnsions in metres. dimensions in metres.
 Reliability Analysis in the Assessment of Bridges 419

2 BRIDGE TYPESAND DIMENSIONS

Two ofthe most common forms ofhighway bridge structures were selected
for detailed examination. These were in-situ reinforced concrete portal slab
bridges and simply supported, precast, prestressed M -beam bridges with in-
situ reinforced concrete deck slabs (see Figs 1 and 2 for typical examples).
By reference to the Department of Transport's bridges database and
departmental standards for cross-section design, a number of different span
lengths, cross-section widths and spanjdepth ratios were selected as
representative of the current and likely future stock of these types of
structures in the UK. These cross-sections included single two-lane, dual
two-lane and dual three-lane structures with span lengths ranging from 6 to
28m and span/depth ratios from 11 to 30.

3 REINFORCEMENT DETAILS

The reinforcement details required for use in the collapse and reliability
analysis were obtained by designing a new bridge corresponding to each of
the selected dimensional configurations. Equally these details could have
been obtained from existing bridges by reference to design plans or by site
measurement.
Longitudinal and transverse reinforcement in the slabs, as weil as
prestress details and shear links in the M-beams, were designed using
standard design practice based on elastic design moments and code
provisions for shear reinforcement.

4 LOADING

In addition to dead and superimposed dead Ioads, two live loadcases from
BS 5400: Part 2 were considered in the collapse analysis. These were the
various lane loading combinations (HA with knife edge) and the heavy
vehicle Ioad (HB with 25 units). The most severe combination of HA lane
configuration and HB vehicle location was found by the automated
computer procedures in the collapse analysis and subsequently used in the
reliability analysis.

5 COLLAPSE ANALYSIS

A number of possible plastic collapse mode shapes were selected as being

appropriate for the bridge deck types examined. A set of geometric
420 Campbell Middleton and Angus Low

parameters describing these failure modes were optimised to give the lowest
load factor against failure based on minimum energy assumptions:
Load factor LF
ED Energy dissipated in the plastic zones and/or yieldlines
WD Work done by the applied Ioads

5.1 Portal Slab Failure Modes

Two geometric yieldline failure modes were considered for each of the
portal slab bridges:
Mode 1 Full-width (FW) transverse yieldlines across midspan and both
supports perpendicular to the traffic direction (Fig. 3).
Mode 2 Partial-width (PW) yieldline failure mode with variable
geometric failure mechanism parameters to define the apex
angle (AL) of the triangular plate and width of failure across
the deck (CC) (Fig. 4).

5.2 M-beam Failure Modes

An M-beam deck comprises a number of discrete beams connected by an
in-situ deck slab. At collapse each individual beam will be either intact or
failed.
The method of analysis adopted in the collapse program considered a
progressive failure of adjacent beams across the deck. lt allowed the edge
beam to collapse first and then continued calculating the Ioad factor for the
specific failure geometry after each successive beam failure until all beams
had failed to form a full-width collapse mode. Two failure modes were
considered for the M-beam deck bridges (see Figs 5 and 6).
Mode 3: Combined shear and bending failure (Fig. 5)
This failure resulted from propagation of an inclined crack, the location
of which was defined by points A and B, as shown in Fig. 5.

FIG. 3. Full-width portal slab failure modc. FIG. 4. Partial-width portal slab failure
mode.
 Reliahility Analysis in the Assessment of Bridges 421

LB

/(sE J ZsE /
~A C

FIG. 5. M-beam Iailure modc 3, combincd FIG. 6. M-bcam failure mode 4, shear.
shcar and bcnding.

For the work/energy evaluation a unit virtual displacement (6 = 1) from

point B to B' was assumed in the top of the failed beam.
Initially the program introduced a vertical crack on one edge beam,
resulting in a pure flexural failure. The crackwas then inclined by increasing
dimension LB and the same progressive failure across the deck assessed.
This procedure of optimising crack inclination and the number of failed M-
beams was repeated for the crack initiating at the support, f span, span and ±
midspan in order to find the lowest Ioad factor for this collapse mode. For
modes with an inclined crack (LB >LA, BE< 90'') the shear steel intersected
the failure plane and added to the energy dissipated in the section.
The deck slab between the failed and unfailed M-beams was assumed to
fail in a series of triangular plates. This modewas found to produce a lower
energy state than using a single triangular plate between beams.

Mode 4: Shearfailure (Fig. 6)

This failure mode had a zone failing in shear (ABCD) in which energy was
dissipated in the distortion ofthe concrete andin the straining of shear steel.
The energy dissipated in the plastic shear zone ofthe beams was obtained by
reference to N ielsen. 3
The failure geometry was defined by assuming a unit virtual displacement
downwards from point D to D'. This resulted in both points B' and D'
422 Camphell Middleton and Angus Low

rotating through the same angle, AL. A similar iterative procedure to that
adopted for calculating the lowest factor for mode 3 failures was adopted
here, except one extra geometric variable was introduced-the width of the
shear zone parallelogram (LC).
The position of initiation of the shear zone, defined by a line from point A
i t
to B on Fig. 6, was examined at the support, span and span. In addition,
the inclination of the parallelogram (defined by dimensions LA and LB),
width of shear zone (LC) and number of failed beams across the deck were
optimised until the lowest Ioad factor defining the most likely shear failure
mode was determined.

5.3 Governing Failure Mode

The governing or most likely overall failure mode was selected by
comparing the Ioad factors obtained for all the different failure modes under
each ofthe specified loadcases and choosing the lowest for each bridge. This
modewas then used in the reliability analysis to evaluate the probability of
such a failure occurring.

6 RELIABILITY ANALYSIS

6.1 Basic Variables

Reliability theory was used to determine a 'national' probability offailure
for each of the bridges examined.
In normal practice the major parameters used in structural design and
analysis are considered constant (e.g. dimensions, material strengths, applied
Ioads). However, in reality many ofthese parameters are subject to statistical
variation. For example, although a design may specify 40 MPa concrete
cube strength, the concrete that is actually placed in the structure may well
have a range of values between, say, 38 and 50 MPa. These values could be
represented by a normal or log-normal distribution curve with a particular
mean and standard deviation.
The Ievel II method of reliability analysis used here revolves araund
selecting a set ofthe most important variable parameters, which are referred
to as basic variables, and investigating the effect of this variability on the
strength of the structure studied.
Allowance was made in the reliability analysis for uncertainties in the
strength (or resistance) of the structure and applied ultimate collapse Ioads
arising from possible variations in the values of yield stress of the concrete
and steel, position of the main reinforcement, magnitude of the applied live
Ioads and permanent Ioads, and the accuracy of the structural model.
 Reliability Analysis in the Assessment of Bridges 423

Nine basic variables (XJ were selected for the bridges examined:
xl = FY = characteristic yield Stress of steel (kPa)
X 2 = DERR = error in placement of reinforeerneut bars from specified
position (m)
X3 = FCU = characteristic concrete compressive strength (kPa)
X4 = WDLL = work done by live Ioads (kNm)
X5 = WDSDL = work done by superimposed dead Ioads (kNm)
X6 = GC = density of concrete (kNjm 3 )
X7 = H = thickness of deck slab (m)
X8 = UNC = structural model uncertainty (non-dimensional)
x9 = FPY = characteristic yield Stress of prestressing strand (kPa)
6.2 Limit State Function, Z
Using the principles of plastic analysis, failure was defined to have
occurred when the work done by the Ioads (WD) equalled the energy
dissipated in the yieldlines or plastic zones (ED).
By defining the failure function Z such that
Z=ED-WD
then
Z ~ 0 => failure Z > Ü=> safe
As in the collapse analysis, the expressions for energy dissipated and work
done in the plastic zones were related to moment capacities, strain energies,
virtual displacements of the Ioads and geometry. These parameters were
themselves functions ofthe basicvariables such as steel yield stress, concrete
strength and reinforcement location which were described as probabilistic
functions to account for their variability. Hence
Z = g(Xi)
-
= 0 at failure where Xi
-
= set of all basic variables
A failure function corresponding to each of the four postulated failure
mechanisms was incorporated into the program.

6.3 Advanced Level II Method of ReHability Analysis

In the Ievel II method of reliability analysis the magnitude of each of the
basic variables is changed from their original mean value until a
combination of values is obtained for which the failure function Z equals
zero and collapse has occurred. The manner in which these parameters are
varied depends on their standard deviation and also their relative
importance to the loading on, or strength of, the structure. The difference
between the value of each variable at collapse and its original mean value
can be assessed to determine the likelihood ofthe event occurring and hence
424 Campbell Middleton and Angus Low

the probability of failure of the bridge. The reliability index (ß) gives a
quantitative measure of this difference and can be directly related to the
probability offailure ofthe bridge. Sensitivity factors (IX) were derived in the
reliability analysis which give an indication of the relative importance of
each of the basic variables to the reliability of the bridge.

7 RESULTS

7.1 Portal Slab Bridges

7.1.1 Col!apse analysis
The full-width (FW) failure mode was found to govern the most likely
collapse mode of all the portal slab bridges examined except for the two
shortest span structures (L = 7 and 9 m) for the wider dual two- and three-
lane bridges. In these cases the partial-width (PW) failure mode governed.
Basically this is because the lengths of yieldline in which energy can be
dissipated tend to be greater in the partial-width mode relative to the full-
width mode for all but the shortest, widest bridges in which the partial-width
failure occurs at the very edge of the bridge under a heavy HB vehicle.

7.1.2 Loadcase governing collapse

The HB loadcase was the dominant loadcase determining the most likely
collapse mode in all the portal slab bridges examined (Table 1). With the
wider, Iongerbridges it was less clear which would be the dominant loadcase.
As width and span increased the effect of multiple lane loadings in the HA
case became more significant and the relative importance of the
concentrated HB load compared to the HA uniform lane Ioad reduced. In
these bridges it was found that the Ioad factors for HA and HB loading were
much closer together but the HB case still predominated (i.e. the LF for the
HB loadcase was the lowest).

7.1.3 Reliability analysis

The failure probabilities of the portal slab bridges examined differed
markedly for different span lengths, widths and spanjdepth ratios. F or
portal slab bridges with a span/depth ratio of 30 values ranging between
10- 7 for a short (7 m), narrow (single two-lane) bridge and w- 24 for the
Iongest (28m), widest (dual three-Jane) bridge were obtained.
Steel yield stress, live Ioad and the structural model uncertainty were
found tobe the dominantbasicvariables governing the collapse ofthe portal
slab bridges.
The reliability index (ß) can be directly related to the probability offailure
of the bridge with increasing ß value corresponding to increasing reliability.
 Rehability Analysis in the Assessment of Bridges 425

TABLE 1
Example Ioadfactars at collapse for portal slah bridges with span/depth ratio
SDR=30

Bridge type Failure mode Loadcase Load factor at

collapse

1. Dual three-lane (wide) FW HA 3·04

Span L = 7 m (short) HB 2·93
PW HA 2·59
HB 2·54"
2. Dual three-lane (wide) FW HA 1·93
Span L =28m (long) HB 1·93a
PW HA 2-44
HB 2-44
3. Single two-lane (narrow) FW HA 2·30
Span L = 7 m (short) HB l·93a
PW HA 2·74
HB 2·18
4. Single two-lane (narrow) FW HA 1·73
Span L =28m (long) HB 1·67"
PW HA 5·00
HB 4·74
- -~~-

aIndicates governing failure mode and loadcase.

For example, the graph in Fig. 7 plots the reliability of some portal slab
bridges against span length. The results for both partial- and full-width
modes are shown. For structures failing by the full-width mode the
reliability is relatively constant with span, whereas for the partial-width
mode it greatly increases. lt can be seen that the structures of most
significance are the narrow, short span portal slabs with reliability indices
near 5 corresponding to a probability of failure of approximately 10- 7•

7.2 M-beam Bridges

The highest nominal probability offailure obtained in the M-beam study
(10- 19 ) was many orders of magnitude less than those obtained for the
majority of portal slab bridges. The results indicated that the narrow, single
two-lane M-beam bridges had a significantly higher probability of failure
than the wider dual two- and three-lane bridges. This is because the heavy
HB vehicle Ioad is far more concentrated over a narrow structure than over a
wider dual three-lane bridge, resulting in a higher likelihood of failure.
The most likely failure mode for the single two-lane bridgeswas found to
be a full-width combined shear and bending mode with an inclined crack
426 Campbell Middleton and Angus Low

-o- 03-FW

//
12 --<:r- 52 -FW
1033
-o- 03-PW
>-
10-24 ~
"""*"" 52 -PW
~ 10
"0
.s
.0
0
.0
0
>.
~ 8 -16 ct
.0 10 ~
0 ::J
Qi 0
er
10-10 lJ...
6

4 10-5
5 10 15 20 25 30
Span (m)
FIG. 7. Reliability index (ß) versus span, both failure modes SDR = 30.

initiating from the i span point. Some partial-width failures were observed
in the wider dual three-lane bridges. The pure shear failure mode did not
govern failure for any of the bridges examined.
The probability of failure of a given structure was not very sensitive to M-
beam size, indicating that the adjustments made in prestress and shear link
design to account for the different beam size adequately compensate for this
change. The dominant variable in the M-beam analysis was the live Ioad
followed by the model uncertainty factor. It was interesting to note that the
yield stress ofthe prestressing strand had a comparatively small effect on the
results.

8 DISCUSSION

A significant finding in this study was the high probability of failure of short
span, narrow portal slab bridges in comparison to the other bridges studied.

8.1 Sensitivity of Reliabilities to Basic Variable Parameters

A sensitivity analysis was performed for the portal slab bridges by
separately doubling the standard deviation for steel yield stress and then for
work done by the live Ioads. These changes resulted in significant increases in
the nominal probability of failure of individual bridges.

8.2 Limitations on Assessment Technique

Clearly this sensitivity to basic variable parameters is unsatisfactory
considering that their statistical properties are chosen in an imprecise
manner from quite limited data and a degree of engineering judgement.
 Reliability Analysis in the Assessment of Bridges 427

However, the methods outlined here are not intended for finding absolute
reliabilities. It is proposed that they be used for ranking reliabilities of
different bridges or identifying the most likely failure mode within one
structure. The strength ofthe methodisthat it takesquantitative account of
aspects such as brittleness, ductility and known variabilities which are not
covered directly by more conventional methods of strength assessment.
Another difficulty in this project was developing realistic collapse models.
There is very limited research available into the actual mechanisms offailure
of concrete bridges and much of this relies on quite simplistic loading and
failure models. In practice few engineers actually employ upper-bound
methods when assessing bridge structures, possibly because they fear they
cannot, from judgement, assess the lowest upper-bound failure mode. To a
degree this can be overcome, using powerful modern computers, by selecting
a !arge number of possible yielding/plastic collapse patterns and searching
for the lowest upper bound.
It is also difficult to accurately model the ultimate strength of a bridge
when factors such as ductility and membrane action can contribute
significantly to the type of collapse mode and actual strength of the bridge.

9 FURTHER RESEARCH AND PRACTICAL APPLICATIONS

Currently a TRRL funded research programme at Cambridge University is

working towards solving some ofthe problems outlined here and developing
a more generalised computer program for assessing concrete bridges. This
program will be calibrated against existing non-linear finite element
programs with the aim ofproducing a program that can realistically model
the lowest upper-bound failure Ioads for some common types of concrete
highway bridges. It is intended to examine a broad range of collapse modes
and refine the energy dissipation calculation to model the behaviour of the
concrete more realistically. By generalising the loading subroutine any
selected live Ioad model could be examined in detail. Reliability theory will
be applied to the derived collapse modes to allow for variations in the
loading and strength parameters ofthe structure. Within this it is planned to
make provision for the deterioration of reinforcement due to corrosion.

10 CONCLUSION

In this project procedures were developed for assessing the reliability of a

number of common types of concrete bridges. Basedon this work, a current
research contract at Cam bridge U niversity is developing a suite of computer
428 Campbell Middleton and Angus Low

programs aimed at assessing the strength and reliability of a moreextensive

range of bridges subject to differing levels of physical deteriora tion und er a
variety of possible loadcases.
It is anticipated that the derived probabilities of failure may provide a
basis for assessing the relative priorities of remedial work or bridge
replacement in relation to different bridge types and spans, and thus
contribute to the vital task facing our transport authorities in assessing the
strength, safety and reliability of the existing population of concrete bridges.
In this way bridges which fail crude assessment procedures may be allowed
to safely continue serving the community.

ACKNOWLEDGEMENTS

The authors would like to thank the Transport and Road Research
Labaratory and Ove Arup & Partners for permission to publish this paper.
The work reported herein was carried out under a contract placed on Ove
Arup & Partners by the Transport and Road Research Laboratory. Any
views expressed are not necessarily those of the Transport and Road
Research Labaratory nor the Department of Transport.

REFERENCES

1. JOHANSEN, K. W., Yield-line Formulae .for Stabs. Cement and Concrete

Association, Publication 12.044, 1972.
2. CONSTRUCTION INDUSTRY RESEARCH AND INFORMATION ÄSSOCIATION, The
Rationalisation ofSafety and Serviceability Factors in Structural Codes. CIRIA
Report 63, July 1977.
3. NIELSEN, M. P., Limit Analysis and Concrete Plasticity. Prentice-Hall, New York,
1984.
4. THOFT-CHRISTENSEN, P. and ßAKER, M. J., Structural Reliability Theory and its
Applications. Springer-Verlag, Berlin, 1982.
5. THOFT-CHRJSTENSEN, P. and MuROTSU, Y., Application of' Structural Systems
Reliability Theory. Springer-Verlag, Berlin, 1986.
 38
Strength Assessment Methods for Concrete Bridges

P. A. JACKSON
Gifford and Partners. Southampton, UK
and
R. J. COPE
Polylechnie South West. Plymouth, UK

ABSTRACT

Two half-scale models of bridges H'ith uery lightly reinforced deck slabs haue
heen tested tofailure under both the British 'HB' design Ioad and single wheel
Ioads. The strength of the models is assessed using a uariety of approaches:
conuentional elastic analysis using British and American practice, yield-line
analysis, a semi-empirical approach deueloped in Northern Ireland which
alfowsfor memhrane action in the slab, the Ontario Highway Bridge Design
Code (which also uses memhrane action) andfinally non-linear analysis. The
predicted strength isfound to differ greatly between the approaches, the ratio
o{ highest to lowest prediction approaching ten in one case.
Under single wheel Ioads the least conseruatiue prediction, that obtained
using non-linear analysis. proued to be most accurate although the predictions
obtained using the other methods which allowfor membrane action were also
acceptable. Under the H B Ioad the non-linear analysis again gaue the best
prediction and the other predictions which considered memhrane action proued
to he unsa{e, apparently due to failure to model the interaction ofglohal and
local effects. The conuentional methods hased on elastic theory were
conseruatiue in all cases.

INTRODUCTION

In Britain the usual approach to the assessment of concrete bridges is the

same as the design approach. The moments and forces are obtained from a
linear elastic analysis and the sections are then checked against the
requirements of a code of practice, normally BS 5400. 1 This approach, and
429
430 P. A. Jackson and R. J. Cape

the codes of practice used with it, has evolved over many years. Structures
which comply with its requirements are always safe and serviceable but the
reverse does not apply; structures which do not comply arenot always either
unsafe or unserviceable. This is relatively unimportant in design where
ensuring that structures comply with code requirements is Straightforward
and the cost of providing additional Ioad capacity is comparatively small.
However, in assessment the situation is quite different. The cost ofincreasing
the strength of an existing bridge by even 10% may be millians of pounds
and similar effort may be required to make a structure comply with code
requirements which are merely arbitrary rules defining what is now
considered 'good practice'.
In the extreme case of the deck slabs of beam and slab bridges, the actual
strength can be as much as ten times higher than is implied by conventional
assessment methods. In this paper this is demonstrated using the results of
tests on two half-scale concrete bridges and possible alternative assessment
methods are considered.

DETAIL OF MODELS

Two half-scale models of M-beam 2 type bridges were constructed, using

standard T2-beams 2 as approximate half-scale models of M4-beams. The
beams were precast in a normal casting yard, transported to the Iabaratory
and placed on elastomeric bearings. Formwork was then constructed for the
deck slab, which was cast in situ using a half-size concrete mix with a 10 mm
maximum aggregate size. The reinforeerneut and prestressing for the beams
were designed in the normal way using BS 5400. However, because it was
primarily the behaviour ofthe deck slab which was being investigated, 25%
more than normal prestress was provided whilst the reinforeerneut in the
deck slab was substantially lighter than would normally be provided.
The major reason why deck slabs are stronger than conventional design
methods isthat they are able to work by compressive membrane action, which
is otherwise known as arehing action or dome effect. In order to realise this
effect some restraint is required and previous research has suggested that
diaphragms are needed to develop this restraint Analysis using the
approach which will be considered in the section 'Non-linear Analysis'
suggested, however, that this was not the case; the understressed concrete
and steel surrounding the critical areas would provide adequate restraint In
order to investigate this, the first model, which is illustrated in Fig. 1, was
deliberately made a worst case for restraint and had no diaphragms. The
second model had five beams and it also had support diaphragms as well as
parapet upstands. Both models had the full-scale equivalent of a 160 mm
 Strength Assessment Methods for Concrete Bridges 431

SECTION
FrG. I. Detail of first deck.

deck slab with the beams placed at 2m centres, which is the widest spacing
ever likely to be used for this type of deck.

ASSESSMENT

The strength of the bridges was assessed both for HB Ioad and for single
wheelloads. Four different assessment methods were tried and thesewill be
considered in turn.

Conventional Approach
A conventional assessment of the first deck was performed. This was
based on British design practice and used Westergaard's approach for the
local analysis of the deck slab and a linear grillage for the global analysis of
the bridge. The only departure from normal practice was that, in order to
facilitate direct comparison with test results, measured rather than nominal
material properties were used and all partial safety factors were set to one.
This procedure implied a failure Ioad for the firstdeck of 14 kN per wheel
when all16 wheels ofthe HB rig were loaded and 21 kN when only one wheel
was loaded. The difference was due to the influence of global transverse
moments, the moments induced in the deck slab by its action in distributing
Ioad between the beams. These moments are not normally considered in
North American practice. The normal AASHT0 3 approach to assessing
deck slabs assumes a different shape patch Ioad to that used in the tests.
However, the analytical approach it is based on is very similar, apart from
not considering global transverse moments. lt would therefore predict a
failure Ioad of approximately 21 kN per wheel, both under single wheelloads
and under the HB bogies.
The reason for the very low predicted failure Ioad, only some 34% of the
BS 5400 requirement for a 45-unit HB Ioad, was that the reinforcement was
432 P. A. Jackson and R. J. Cape

lighter than normal. The second deck was provided with only one layer of
reinforcement each way instead of two. The reinforcement was slightly
heavier, however, with 8 mm high yield bars at 125 mm centres instead of
6 mm as in the first deck. Because this steel was located 10 mm below mid
depth, the sagging moment capacitywas greater than for the first deck. The
strength assessed using .westergaard's approach and considering only
sagging was therefore higher at approximately 16 kN per wheel with a1116
wheels loaded and 25 kN under a single wheel. However, this implied a
hagging moment in excess of the slab's capacity and the strength assessed
using Pucher's charts was similar to that quoted above for the first deck,
hagging moments being critical.

Yield-Line Analysis
A normal yield-line analysis, which ignores membrane effects, implied a
failure Ioad under a single wheel of approximately 100 kN for the firstdeck
and 85 kN for the second.
Plastic theory, upon which yield-line analysis is based, suggests that global
moments and forces could redistribute away from the critical areas of slab
and thus that the local strength ofthe slab would not be reduced when all16
wheels ofthe HB vehicle were loaded. Indeed, with such a lightly reinforced
deck slab, it might be argued that the global longitudinal compression
induced in the deck slab by the full HB Ioad would increase the local
longitudinal moment capacity of the slab and hence increase its strength.
However, this force makes the slab behave as though more heavily
reinforced and thus reduces its ductility. Because of this it is debatable
whether yield-line theory is a valid means of analysis for this case.

Punching Theory
Previous research into deck slab behaviour 4 •5 has shown that slabs
subjected to localloads fail by punching at Ioads which can be substantially
above those predicted even by yield-line theory. The reason for this high
strength isthat the slabs are able to support load by compressive membrane
action and the new draft assessment version of BS 5400 6 suggests that this
effect should be considered in assessment.
Two approaches have been developed for predicting the failure Ioads, one
by Kirkpatrick et al. 4 and the other by Hewitt and Batchelor. 5 Kirkpatrick's
approach gave a failure Ioad of approximately 150 kN for the second deck.
Since Kirkpatrick recommended that diaphragms should be provided to
develop the restraint, his approach is not strictly applicable to the first deck.
However, ignoring this, the failure load it predicted was 185 kN. The reason
for this being higher than for the second deck was that the concrete was
stronger.
 Strength Assessment Methods for Concrete Bridges 433

Although Kirkpatrick did perform some tests with two wheels loaded
(and he found that the failure Ioad per wheel could be reduced) his approach
does not enable the effect of loading more than one wheel to be quantified.
Like Kirkpatrick, Hewitt and Batchelor recommended the provision of
diaphragms and therefore their approach is not strictly applicable to the first
deck. Their approachalso requires an empirical restraint factor. Setting this
to 0·6, as suggested for concrete bridges, gave failure Ioads which were very
similar to those predicted by Kirkpatrick's approach.
Hewitt and Batchelor's approach has been used as the basis of an
assessment method given in the Ontario Highway Bridge Design Code. 7
This recommends that the local strength ofthe slab should be assessed using
charts which are based on Hewitt and Batchelor's approach using the
'conservative' restraint factor of0·5. The global strength ofthe bridge is then
assessed independently using a conventionallinear elastic grillage analysis.
This approach gave a failure Ioad for the second deck of approximately
150 kN per wheel, the implication being that failure would take the form of
one beam failing in flexure. For the firstdeck the prediction would have been
approximately 140 kN per wheel but again, because of the Iack of
diaphragms, the use of the approach would not have been recommended.
For the first deck, the best interpretation ofthe Ontario code appears tobe
that yield-line methods should be used to assess the local strength ofthe slab.

Non-linear Analysis
Both models were analysed using a non-linear program which has been
developed by the authors. This program uses comparatively simple line
elements but, because of novel features of the elements and because it
considers all six degrees of freedom at each node, it is still able to model in-
plane forces reasonably realistically. This enables it to model membrane
action. The program was developed from non-linear finite element programs
which have proved capable of analysing deck slabs allowing for membrane
action. 8
Because the research project was originally aimed at developing design
methods rather than assessment methods, and because Serviceability is
critical in design, the computer models were designed to give the best
predictions for the Serviceability behaviour. Two computer models were
used for the first deck, one with a fine element mesh and the other with a
coarse element mesh. The fine mesh analysis was used primarily for
serviceability analysis although it was taken up to failure under a single
wheelload. The coarse mesh was used for the failure analysis under all 16
wheels of the HB Ioad.
The analysis predicted that at service Ioad Ievels the slab would be most
heavily stressedunder thc whecl ncarest the centre of the deck. However, it
434 P. A. Jackson and R. J. Cape

predicted that failure would take the form of a local brittle bending
compression failure under the wheel farthest from the centre ofthe deck. The
predicted failure Ioad was approximately 100 kN per wheel. Analysis using
the finer mesh predicted that the failure Ioad under a single wheel applied
near to midspan would be approximately 220 kN.
Only one computermodelwas used for the analysis of the second deck
and this was intermediate in detail between the two models used for the first
deck. As for the first deck, it predicted that failure would take the form of a
brittle bending compression failure under one wheel. Unlike for the first
deck, the critical wheel was one of those nearest midspan of the deck.
Concrete crushing first occurred on the soffit over the web of a beam rather
than on top ofthe slab immediately under the wheel, as in the first deck. The
predicted failure Ioad was marginally higher than for the first deck.
Experience has shown that the program tends to be conservative when a
realistic element mesh is used, as in the analysis ofthe second deck. Indeed it
had been intended that the program should tend to err in the safe direction.
However, the use of an over-coarse mesh, as in the analysis of the first deck,
makes it less conservative.

TESTS

First Deck
The deck was first loaded with the design service HB Ioad applied in a
critical position; 120 cycles ofthis Ioad, 5000 cycles of a lower Ioad and three
cycles of a 20% greater Ioad were then applied in this and several other
positions. The loading rig was then returned to its first position and the Ioad
reapplied. Thesetests were intended primarily to investigate the service Ioad
behaviour. However, they also served to ensure that any cracking which
would be likely to occur in the real bridge would occur in the model.
On completion of these tests the bridge was still in good serviceable
condition, despite having been loaded to over three times the 'ultimate
strength' of the slab as given by conventionallinear analysis. The Ioad was
then increased until failure occurred. At a Ioad of approximately 103 kN per
wheel the wheel farthest from the centre of the bridge punched through the
deck. This failure looked like a dassie 'punching shear' failure. However, a
close study of the slab immediately before the finalload increment had been
applied had revealed a line of crushing concrete extending from the wheel
which later failed towards the adjacent wheel. It appeared that this crushing
was the cause of the failure.
As failure had approached, extensive flexural and shear cracks developed
in the two most heavily loaded beams and the maximum beam deflection
 Strength Assessment Methods for Concrete Bridges 435

was approximately 60 mm.lt did not appear, however, that beam failure was
imminent. In a subsequent test of a single beam with the appropriate width
of in-situ top flange, failure occurred from a deflection of 110 mm. The Ioad
was then over 20% greater than it had been at a deflection of 60 mm.
Since the failure had taken such a local form the bridge was still in good
enough condition to enable two single wheel tests to be performed. In both
cases failure took the form ofthe wheel punching through the deck with little
warning. The failure Ioads were 204 and 226 kN, approximately twice the
failure Ioads per wheel as when all 16 wheels had been loaded.

Second Deck
The second deck was subjected to a similar Ioad history to the first. On
completion of the service Ioad tests the slab was more extensively cracked
than the first had been. In particular, there were cracks on the top ofthe slab
which failed to close fully when the bridge was unloaded. No cracks had
appeared in the top of the slab of the first deck until the final loading to
failure was weil advanced.
As with the first deck, the eventual failure took the form of one wheel, this
time a wheel near to midspan, punching through the deck. Again as with the
first deck, concrete crushing had been visible a Ioad stage before the failure
occurred, this time on the soffit along the edge of a beam. The failure Ioad
was somewhat greater than for the first deck, at approximately 122 kN per
wheel.
The sudden failure under a wheel reduced the total Ioad on the bridge and
hence reduced the deflection ofthe beams. There was insufficient time for the
hydraulic pressure in the four jacks ofthe loading rig to equalise before three
further wheels, one under each jack, punched through the deck. Despite this
there was enough slab left in reasonably good condition to enable some local
tests to be performed. These included two single wheel tests which were
directly comparable with those which had been performed on the first deck.
Punching failures occurred at Ioads of 176 and 184 kN.

DISCUSSION

As expected, the tests showed that conventional assessment methods using

linear elastic theory are very conservative for this type of structure. The
margin of conservatism is so great that it would be unreasonable to
condemn a deck slab as unsafe on the basis of such an assessment. They
should only be used to dass a deck slab as safe. In the authors' view it is
reasonable, even then, to smooth out the peaks of the elastic moment
distribution over a finite width of slab. That is, to base the section check on
436 P. A. Jackson and R. J. Cope

the average moment intensity over a width equal to, say, the lesser of six
times the effective depth of the slab or half the slab span. 9
The yield-line analysis of the slab of the first deck gave an answer which
was very close to the actual failure load when all16 wheels ofthe HB vehicle
were loaded. However, the failure mode was so different from that implied
by yield-line analysis that this can be little more than coincidence; yield line
theory depends on ductility whilst the failure mode was very brittle. The
predictions of yield-line theory were conservative for all the other tests.
However, this may not always be the case; there are indications (including a
non-linear analysis ofthe firstdeck with 40% ofthe tendons removed) 9 that
ifthe beams had been weaker the bridges would have failed in the same mode
but at a lower Ioad.
Kirkpatrick's approach gave good predictions for the failure Ioads under
single wheelloads. The predictions were slightly conservative with a mean
ratio of actual failure Ioad to predicted failure load of approximately 1·2.
This is very similar to that which he observed for his own tests. The ratios
were similar for the two decks, indicating that, contrary to Kirkpatrick's
suggestion, his method is applicable in bridges without diaphragms.
However, it appears theoretically that this may not be the case if the wheel
load is applied very near to the end of the bridge, that is near to the
unsupported edge of the slab. The test performed with the wheel only 1-1m
from the end of the deck gave a 10% lower failure Ioad than the test
performed near the centre of the deck.
In both bridges, when all16 wheels ofthe HB vehicle were loaded, failure
of the deck slab occurred at a significantly lower Ioad than predicted by the
punching theories considered in the section titled 'Punching Theory'. The
implication ofthe Ontario code that the high Jocal strength ofthe slab would
cause a global failure, that is a failure of a beam, to precede slab failure
proved incorrect. What appeared to happen was that the global transverse
moments induced by the differential beam deftections reduced the local
strength of the slab.
The non-linear analysis gave good predictions for the failure Ioad under
full HB Ioad. Also, unlike the other methods, it predicted the failure mode
very well; it predicted precisely where concrete crushing would first occur
and which wheel would punch through the slab first. Although the failure
mode is relatively unimportant to someone assessing a bridge, this is
significant; prediction methods which predict the correct failure Ioad but the
wrong mode must be considered highly suspect.
The non-linear analysis also gave a good prediction for the result of the
one single wheel test for which it was investigated fully. However, analysis of
other tests 9 suggests that the methods considered in 'Punching Theory' tend
to be slightly better for such cases and, since they are also far simpler to
 Strength Assessment Methods for Concrete Bridges 437

obtain, they may be preferred. Punching theory could also be used for
assessment under full HB Ioad in cases where there are no significant global
transverse moments, for example in a bridge with intermediate diaphragms.
However, even then there are indications 4 · 9 that the interaction between
adjacent wheels could reduce the strength and some allowance should be
made for this.

CONCLUSIONS

Conventional analysis is very conservative for this type of structure. 1t

should be used only to class a bridge as safe, not unsafe. Even then it is
reasonable to make the assessment less conservative by evening out the
peaks in the predicted moments.
Both punching theory and non-linear analysis gave acceptable predic-
tions for the strength under single wheel Ioads whilst yield-line analysis
proved to be conservative by a factor of up to two.
Only non-linear analysis gave good predictions for the behaviour under
full HB Ioad. Predictions based on assessing the global and local strength of
the slab independently, as is recommended by the Ontario code, could be
unsafe.

ACKNOWLEDGEMENTS

The test work considered in this paperwas undertaken at the British Cement
Association whilst P.A.J. was an employee ofthat organisation. He would
like to thank the association for supporting the project and for allowing him
to publish the paper. He would also like to thank the many staff of the
association who assisted with the test work.

REFERENCES

I. BRITISH STANDARDS INSTITUTION. Code ol practice för the design ol concrete

hridges. BS 5400, Part 4, BSI, London, 1984.
2. PRESTRESSED CoNCRETE AssoCIATION. Prestressed concrete hridge heams.
Prestresscd Concrcte Association, Leicestcr, 2nd Edition, January 1984.
3. ÄMERICAN ÄSSOCIATION OF STATE HIGHWAY AND TRANSPORTATION ÜFFICIALS.
Standard specificationf(Jr highll'a\' hridges. Washington, DC, 13th Edition, 1983.
4. KIRKPATRICK, J., RANKIN, G. I. B. and LoNG, A. E., Strength evaluation of M-
beam bridge deck slabs. The Structural Engineer, 628(4) (1984) 86-8.
5. HEWITT, B. E. and BATCHELOR, B. DE V., Punching shear strength of restrained
slabs. Journal of tlze Structural Dil'ision, American Societr of' Civil Engineers,
101(ST9) (1975) lt\37 52.
438 P. A. Jackson and R. J. Cope

6. DEPARTMENT OF TRANSPORT, Draft BD/88. The assessment of concrete highway

bridges and structures. HMSO, London, 1988.
7. ÜNTARIO MINISTRY OF TRANSPORTATION AND COMMUNICATIONS. Ontario
Highway Bridge Design Code. Downsview, Ontario, Canada, 1983.
8. CoPE, R. J. and EDWARDS, K. R., Non-linear finite element analysis of
eccentrically stiffened bridge decks. Proceedings of international conference on
finite element analysis in computational mechanics. Pergarnon Press, Oxford,
1985, pp. 431-48.
9. JACKSON, P. A., Compressive membrane action in bridge deck slabs. PhD thesis,
Polytechnic South West/CNAA/BCA, April 1989.
 39
Assessment of Stresses in Post-Tensioned Concrete
Bridges

C. L. BROOKES, S. H. BUCHNER and S. MEHRKAR-ASL

Gifford and Partners, Southampton, UK

ABSTRACT

This paper presents the results oftrials carried out on a post-tensioned concrete
six-cefl box structure to determine the in-situ stress Ievels. The techniques used
were hased on the stress-relie{11/hich occurs during the removal of concrete
cores and the re-estahlishment of the stress fields using a novel jacking
technique. The estimated stresses were compared with those predicted using a
finite element analysis and a good correlation was ohtained. The measured
stresse.\' generally ranged between 6 and 9 N/mm 2 although local variations
were ohserved. The elastic constantsfor the concrete were assessed using the
results of the jacking tests and lahoratory compression tests.

INTRODUCTION

The long-term condition of a number of prestressed concrete structures has

been the subject of several research projects in the United Kingdom in recent
years. 1 - 3 These projects have been aimed at determining the in-situ stresses
in the structures and the residual Ievels of prestress. M uch of the work has
been carried out during the demolition of the structures, which provides an
ideal testing ground and a unique loading regime.
The research has involved the development of methods for determining
the in-situ state of stress in concrete structures based upon measurements of
released strains obtained during the removal of standard cores. In addition,
a uniq ue jacking system has been evolved which can be used to determine the
in-plane elastic properties of the concrete and provides a direct measure of
the existing stresses. Although these techniques have been used for residual
439
440 C. L. Brookes, S. H. Buchner and S. Mehrkar-Ast

stress determination in steel structures and rock mechanics, they have not
been previously applied with any great success to concrete structures.
In the current study these techniques have been primarily used for the
investigation of structures suffering distress and in conjunction with
computer assessments related to principal inspections. The site data has
been supported by extensive Iabaratory calibration trials. 4 · 5

IN-SITU STRESS ASSESSMENT

Measurements ofthe elastic response ofa stressed medium to the removal of

a core can be used to provide an estimate of the existing stresses. Once the
core is removed a specialised jacking system is inserted within the remaining
hole. The jacking tests use the application of Ioad to the periphery of the
remaining hole to assess the strain response ofthe material and hence the in-
plane elastic constants for the concrete. In addition, re-establishing the
strains released on the surface and the use of superposition for the different
loading directions provides a direct measure of the existing stresses.
The selection of core size and gauge length is generally controlled by the
size of the largest aggregate particle in the material under test, although
other physicallimitations may exist. For the Iabaratory and site tests, cores
of 150 and 75 mm diameter have been used in conjunction with 140 and

FIG. I. Typical gauging arrangement.

 Stresses in Post-Tensioned Concrete Bridges 441

64 mm long vibrating wire (VW) gauges, and 50, 100 and 200 mm long demec
gauges. A typical gauging and coring arrangement for a 150 mm core
position is shown in Fig. 1.
Analysis of the strains released after coring utilises the equations
developed by Muskhelishvili 6 for the displacements around a hole in a plate
under a known stress field. In order to analyse the results, the concrete is
assumed tobe an elastic homogeneaus isotropic material under plane stress
or plane strain loading conditions. This assumption cannot be considered as
wholly valid for concrete, and as a result anomalies arise due to the presence
ofmicrocracks and the effects ofheterogeneity and anisotropy caused by the
aggregate particles. However, the Iabaratory test results have indicated that
the methods can be used with a good degree of confidence, especially when
the member is subject to an axial loading regime.

IN-SITU TESTS

An assessment of a 144m long prestressed concrete bridge compared the

stresses predicted using the in-situ stress-relief technique with those
obtained from an in-depth computer study. The viaduct was constructed
with three continuous in-situ spans of 43, 58 and 43 m set on a sagging
vertical curve and supported on inclined abutment legs. The section of the
bridge used in the computer model is shown in Fig. 2.
The main prestressing cables were applied externally within the box
sections, passing beneath a number of diaphragms in the midspan section,
thereby inducing longitudinal compressive stresses in the bottom flange of
the deck. There are 12 cables, one on each face ofthe internal webs and one
on the inner face ofthe edge webs. Each cable consists of 17/28 mm 19-wire
strands and are Gifford-Burrow (CCL spiral strand) type. The strands were
simultaneously tensioned from each end in two stages before being covered
with concrete. In addition to the main cables, there are 168 PSC Freyssi

-at:l±im /Iif-1--1
m Hif\111
E I I I I __.... =--= : -

/ I
:; / i
1/--f----1
I; I /
LI-!
;;i·' .·I Ii
~
f i I
L.L_;
FI(i. 2. Computer modcl showing section through bridge.
442 C. L. Brookes, S. H. Buchner and S. M ehrkar-Asl

strand type cap cables, 84 over each leg, each cable consisting of 12/12 mm
seven-wire strands. Two profiles were used for these strands, each stressed
from either end alternately. A generallass of prestress of 16% was allowed
for in the analysis. This low figure took into account the advanced age of the
concrete at the time of stressing.
During the 25 years since the bridge was constructed water had seeped
through manholes in the top slab and caused some corrosion ofthe external
prestressing tendons. lt was anticipated that any significant loss of prestress
would be detected by use of the stress-relief technique.
Due to restrictions imposed by the depth of slab and the spacing of the
main reinforcement, 75 mm diameter cores were removed. Care was taken to
ensure that the stresses released were unaffected by the presence of the
central diaphragm and so the coring positions were sited on both the north
(N) and south (S) sides some 2-3m away. Arrays of demec gauges were
placed on and around the core positions, with additional rosette arrays of
64 mm long VW strain gauges placed around each position. As coring
progressed strain release readings were takenon the VW arrays. Once coring
was completed and a number of strain release readings had been taken, the
core holes were used for jacking tests to determine the in-situ elastic
properties and to obtain a second measure of the in-situ stresses.

COMPUTER ANALYSIS

The structure, excluding the abutments and foundations, was modelled

elastically using the finite element method. The structure and permanent
loading were symmetrically partitioned transversely to give a problern size
of approximately 5000 degrees of freedom. First-order thin shell elements
were used to represent all features of the deck and piers. Precast concrete
hinges attach the legs to the deck and were incorporated in the model by
coupling degrees of freedom to give similar articulation. The main and cap
cables were included in their actual positions using 3D truss/spar elements
attached to the shell mesh using rigid elements (multipoint constraint
equations). Using a combination of rigid and truss elements to model the
curved cable profile allowed the near uniform web mesh tobe uninterrupted,
thus maintaining good shell element aspect ratios.
Body forcc and pressure loading were used to apply dead and super-
imposed dead loads respectively, the pressure intensity being derived
from a survey of the thicknesses of the finishes. Prestress Ioads were
simulated directly by stressing the modelled cables. This was achieved by
calculating the cable force profile immediately after locking off at the
anchorages and estimating an equivalent axial temperature loading for each
 Stresses in Post-Tensioned Concrete Bridges 443

truss element to give the required force profile along the length ofthe cables.
Once the temperature Ioad vector had been applied to the model, the
resulting cable forces were modified to account for secondary flexural effects.
Second and subsequent sets of equivalent temperature Ioads are derived by
introducing adjustment factors, resolving and checking cable forces.
Generally cable forces converged within 0·5% of those required after four
iterations. This procedure was automated with the aid of a spreadsheet
computer program acting as a pre- and post-processor to the analysis.

RESULTS

During coring through the bottarn slab it became evident that a reasonably
high Ievel of longitudinal compressive stress existed within this structure.
The resulting stresses determined from the strains released during the
incremental drilling test at a depth of 125 mm and those obtained from all
the gauges after coring are shown in Table 1. These results indicated that the
values oflongitudinal stress predicted after only 125 mm ofpenetration were
in good agreement with those obtained from all the gauges after coring had
finished.
Once coring was completed, at a depth of approximately 175 mm, the
major principal stresses on the north side of the central diaphragm were
estimated to range between 6 and 9 Njmm 2 . The results from the south side
were generally higher than the north side. The estimated stresses ranged

TABLE 1
Estimated longitudinal stresses in viaduct

Core Estimated stresses (N/mm 2 )

number
vw Demec gauges vw Jacking
increment -- -----------~~.---
array test
Core Hole
-----~----~--------------------~---

IN 7·2 9-4 7·4 7-1 8·0

2N 7·1 7-5 5·1 6·3 5-4
2S 4·6 9·3 7·0 6·7 3·3
3N 4·5 6·6 7·0
3S 10·5 11·3 7·5 12·1 8·2
4N 5·3 8·8 7·5 8·7
4S 12·7 no 11-4 12·8
SN 5·6 9·0 4·9 5·5 J8
5S 6-4 9·4 11·6 6·9
6N 7·6 8·5 11·5 8·3
444 C. L. Brookes, S. H. Buchner and S. Mehrkar-Ast

TABLE 2
Comparison of jacking technique and standard compression
tests for determining elastic modulus

Box In-situ Labaratory Comparison

number jacking test compression test (Ec/Ei)
(E) (EJ

1N 31·0 36-4 1·17

2N 35·7 34-4 0·96
2S 35·5
3N
3S 33-0 37·3 1·13
4N
4S 38·0
SN 35·6 37·0 1·04
5S 41·4 38·0 0·92
6N 33-0 35·3 1·07
Average 34·95 36·4 1·04

between 8 and 11 N/mm 2 , and rose to a peak of approximately 13 Njmm 2 in

box 4S. Peak values were seen to occur in the outer boxes of each
carriageway with lower values in boxes 2 and 5. The principal stresses were
generally estimated to be parallel to the longitudinal direction.
These results were obtained using the estimates for in-situ Young's
modulus determined from the jacking tests. The results from these tests and
subsequent Iabaratory compression tests are shown in Table 2. The average
value ofYoung's modulus obtained from the jacking testswas within 4% of
that obtained from the compression tests with maximum deviations of
-8% and 17%. The results ofthe jacking test were also used to estimate the
existing Ievels of stress in a number of the boxes. Table 2 shows that the
estimates obtained using this technique were similar to those obtained
from the other gauging arrangements and provided a further method of
assessment.
The results of the computer analysis, based on a long-term reduction in
prestress of 16%, are illustrated in Fig. 3 and compared with the average
stresses determined from the in-situ tests. Using the reduced Ievel of
prestress, the predicted distribution of longitudinal stresses across the deck
ranged between 7·2 and 7·9 Njmm 2 . lt was noticeable that the estimated
stresses were in keeping with these predictions although they were slightly
higher in places. A sensitivity study was carried out assuming constant loss
of prestress across the section. The study indicated that in-situ stresses of 6
and 4 Njmm 2 would result from Iosses of prestress of 33% and 49%
respectively. Therefore it appeared that there had been a localised loss of
 Stresses in Post-Tensioned Concrete Bridges 445

16 F.E. rnodel
-+- Av. in situ stresses
14

12
n -+-
E 10
-€_ -+-
;;;:;;
-+-
8
]cn -+-
-+-
6
J'i
4

0
BOX 1 BOX 2 BOX 3 BOX 4 BOX 5 BOX 6
F!G. 3. Comparison of stresses obtained from the computer analysis and in situ.

stress of approximately 33% in the area ofbox 2, resulting in a redistribution

of stresses into the adjacent cells.
There were a number of possible reasons why the concrete stresses
determined from the analysis differed from those measured. Casting the deck
in situ meant that there was a variation in the age of the concrete, both
longitudinally and transversely across the deck, at the time of stressing. This
would cause a variation in the short-term elastic behaviour of the concrete
and in the long-term creep and shrinkage. In addition, the cable stressing
sequence was such that elastic Iosses could accumulate on one side of the
deck, causing lower stresses on that side. Other environmental factors could
have contributed as the wind and rain direction tended to be from the west.
Despite the apparent localised reduction in the Ievels ofloss ofprestress, the
structure was assessed as being capable of carrying both HA loading and
HA with 45 units ofHB in accordance with the requirements ofBS 5400 and
BD 23/84.

CONCLUSIONS

The coring stress-relief technique provides a relatively quick and simple

method of determining the in-situ stresses in prestressed concrete bridges. In
conjunction with the jacking technique, this method can be used to
determine the stresses in a deteriorating concrete structure or where the
complexity of the problern requires the computer model to be verified. In
addition, the jacking test provides a quick and reliable method of obtaining
a measure of the in-situ modulus for the concrete.
The bridge structure tested was subject to high stresses due to the nature
of the construction. The sagging vertical curvature of the deck and the high
446 C. L. Brookes, S. H. Buchner and S. Mehrkar-Ast

Ievel of prestress resulted in estimates for the existing midspan stresses of

6-9 N/mm 2 , with peak values of 11-13 Njmm 2 on the south side. These
values were generally in keeping with the results obtained from a finite
element analysis. However, localised anomalies were discovered which
suggested that differential loss of stress had occurred between the boxes.
Despite these localised effects, the structure was assessed tobe able to carry
full HA and HB type loadings as required by BS 5400. The use of the coring
stress-relief technique allowed the assessment of the structure to be carried
out on the basis of known facts rather than relying solely on a computer
analysis. Thus the future management strategy for the maintenance of the
bridge could be determined.

ACKNOWLEDGEMENT

The authors of this paper would like to thank the Science and Engineering
Research Council for their support over a 5-year period during the
development of the stress-relief technique.

REFERENCES

I. BucHNER, S. H. and LINDSELL, P., Demolition monitaring ofthe TafFawr Bridge.

Final report, Gifford and Partners, Southampton, October 1986.
2. BucHNER, S. H. and LINDSELL, P., Testing of prestressed concrete structures
during demolition. I.Struct.E./BRE Seminar on Structural Assessment-Based
on Full and Large Scale Testing, Watford, UK, April 1987.
3. LINDSELL, P. and BucHNER, S. H., Prestressed concrete beams: controlled
demolition and prestress loss assessment. CIRIA Technical Note 129, London,
1987.
4. MEHRKAR-ASL, S., Direct measurement of stresses in concrete structures. PhD
thesis, University of Surrey, September 1988.
5. BucHNER, S. H., Full-scale testing of prestressed concrete structures. PhD thesis,
University of Surrey, July 1989.
6. MusKHELISHVILI, N. 1., Some Basic Problems of the Mathematical Theory of
Elasticity. P. Noordhoff Ltd, Gröningen, Netherlands, 1963, pp. 202-10.
 40
Assessment of Prestressed Bridge Beams

DAVID CULLINGTON

Bridges Division. Transport and Road Research Laboratory.

Crmrthorne, Berkshire RG ll 6A U, U K

ABSTRACT

A number oj' pretensioned beams haue been recovered from an M63

underbridge during demo/ition. They have no link reinf(m·ement in the web. For
this reason, ij'assessed using BS 5400: Part 4, they would not complv in shear.
This paper contains a description of'shearfailure tests carried out on the 16m
long bemns. The results indicate that the beams possess reserves oj' strength
because of'conservative assumptions in the method of'calculation and the high
strength of' the concrete. Whereas it may be prudent to provide a minimum
amount oj'shear reinforcement in the form oj'links.for new design, there is a
strong casc for dispensing H'ith this rcquiremcnt in some assessments.

INTRODUCTION

Four motorway underbridges have been replaced at the Peel Green

roundabout, to the west ofManchester, as part ofthe M63 widening scheme.
Consideration was given to retaining the existing structures and widening
them, but the pretensioned beams forming the decks contained no shear
reinforcement, which is required by BS 5400: Part 4: 1984. 1 It was decided to
replace the bridge decks primarily for this reason.
The consultant responsible for the scheme gave some thought to retaining
the beams in spite of the absence of shear reinforcement. Calculations
showed that the shear forces under factared HB loading were, at maximum,

Any views expressed in this papcr are not necessarily those of the Department of
Transport.
447
448 David Cullington

about 70% of the factared shear capacity as found using BS 5400: Part 4.
Had the proportion been lower (perhaps 50%) the bridges might have been
retained simply by means of an agreed departure from standard.
It was decided to test the beams when the decks were demolished for the
following reasons:

- They were 30 years old and likely targets for assessment.

- They would contain the effects, if any, of time and heavy traffic.
- They might exhibit a shape or composite action effect which could cause
them to depart from code predictions.
- Their strength, the Ioad at onset of visible distress and the manner of
failure would be of interest for future assessments.

The Shear Strength of Prestressed Beams

The behaviour of prestressed concrete beams in shear is complex but
approximate methods are available for finding their strength for design and
assessment purposes. BS 5400 requires two types of shear tobe considered.
These are indicated by the symbols ~o and ~r·
The first ofthese, ~ 0 , generally applies to short shear spans.lt relates to a
type ofshear failure that initiates when the principal tensile stress in the web
of a beam exceeds the ultimate tensile strength of the concrete. lts
calculation is based on a number of simplifying assumptions. For instance,
an 1-beam is taken to be reetangular and the maximum principal tensile
stress is assumed to occur at the centroid.
The second, ~r' generally applies to !arger shear spans and involves the
combined action of bending and shear. lt has been reported that the shear
strength of a section of beam is reduced to below ~o when bending cracks
are present. An empirically-based method of calculating the reduced
strength appears in BS 5400, the basic data for which were obtained by tests
on 190 prestressed beams. 2
Failure in the ~o mode is expected to follow the appearance of inclined
cracks in the web and to occur when a crack path is formed between the Ioad
and a point at or near the support. Failure in the ~r mode is expected to
follow the extension of a flexural crack in the bottarn flange into an inclined
crack in the web, and to occur when a path is formed between the Ioad point
and the bottarn flange. 3

Assessment Using BS 5400: Part 4

BS 5400: Part 4: 1984 is a design code, but at present it has tobe used for
assessment calculations as weil. One important difference between design
and assessment concerns the provision, or avoidance, of specific details. A
detail being assessed may not be desirable according to current design
 Assessment of Prestressed Bridge Beams 449

praetiee, but the strength ofthe eomponent may nevertheless be suffieient to

resist the imposed Ioads.
The provision ofweb reinforeerneut in prestressed beams is a ease in point.
The eurrent version of BS 5400: Part 4 requires web reinforeerneut to be
provided in all eases. The version of BS 5400: Part 4 issued in 1978 was less
rigid in its requirements. It permitted the use ofbeams with no reinforeerneut
when the applied shear was less than 50% of the resistanee and where tests
had shown that reinforeerneut was not required. It would be helpful to
relax the eurrent design requirements for use in assessment provided that
reliability did not suffer.

SPECIMENS AND TESTING METHOD

The Bridges
The M63 crosses the roundabout earriageway twiee and there are thus
four bridges on the site. Eaeh bridge eonsists of three simply supported
spans of 8, 16 and 8 m. Twenty-two beams form eaeh span, plaeed side by
side and transversely post-stressed through in-situ diaphragms. The top
slab, east on permanent formwork, is largely unreinforeed. A 'triangular
saw-tooth' shaped reinforeing bar provides interfaeial shear eonneetion.

The Beams
The beams (see Fig. 1) have 11 hol es along their length for transverse
prestressing. Apart from a small amount of reinforeerneut in the end bloeks,
the webs are free from shear reinforeement. For the tests, the transverse
wires and diaphragms were removed, and the transverse holes refilled with

Cross-section of beams

Displacement transducer
FIG. I. Details of beams and lest arrangement.
450 David Cullington

fl.owing mortar. Loadpositions were selected for the shear tests so that the
principal bending cracks avoided the hole positions in areas of high shear.

Obtaining the Beams for Testing

The method of demolition was to breakout the beams from the bridge in
units of three joined tagether by the top slab. To recover the middle beam
intact for testing required the slab tobe cut through mid-way between the
beams and the outer beams to be prised off.
In practice considerable difficulty was experienced. Cracking took place in
the webs of the middle beams, in the vicinity of the transverse prestress, as
the outer beams were removed. This was thought to be a result of the
combined action of the release of prestress and the forces imposed by the
separation process. Separation was finally achieved successfully by coring
out the prestressing wires prior to prising the beams apart. Four beams
suitable for testing were recovered.

Test Configurations
The configurations for the first four shear tests are summarised below:
Beam No. 1 end 1: point load at 2·9 m from support; beam supported over
full span (15·7 m).
Beam No.1 end 2: point load at 4·1 m from support; beam supported over
span of 12·3 m, beam fulllength.
Beam No. 2 end 1: point load at 5·7 m from support; beam supported over
full span (15·7 m).
Beam No. 2 end 2: point load at 1·6 m from support; beam supported over
span of 8·1 m, beam 11 m long.
The beams were 0·86 m deep, and consequently the shear span to depth
ratios for the tests were 1·9, 3-3, 4·8 and 6·6.

Method of Testing
The test arrangement is shown in Fig. 1. Loading was provided by a
hydraulic jack of 1000 kN capacity operated through a servo-control panel.
Alltests were carried out under displacement control. There are a number of
advantages in this: measured strains and displacements form a consistent
set, failure is more controlled, and it is safer for observers who approach the
beam to plot crack positions.

Instrumentation and Methods of Observation

The instrumentation consisted of load cells under the jack and at each
support, transducers registering vertical displacement of the bottarn fl.ange
on the beam centreline, and demec studs to determine bending and shear
 Assessment of Prestressed Bridge Beams 451

strains. Loads and displacements were recorded on a data logger. In

addition, an X- Y plotter was connected to the jack load cell and a
displacement transducer beneath the load point to give an independent
indication of load-displacement behaviour.
Acoustic emission microphones were attached to the web of the beam in
two of the tests: to collect basic data and give an indication of prolonged
cracking after the application of each increment was completed-a sign of
impending failure. Cracks were marked on the beam at various increments.
Video cameras were used to record the failures.

RESULTS

Description of Failures
The crack pattems at failure are shown in Fig. 2. In all cases, in their final
state, the beams contain two systems of cracking. Bendingcracks are present

"4%i" ' '" 2 9m

t I J I I 3?::::1 I
t
FIG. 2. Beam cracks at failure.

beneath the applied load, emanating from the bottom flange, and a number
of pronounced cracks are present in the web extending between the load
point and the near support. The two systems of cracking are independent
and did not combine. lt was the web cracks that caused failure. This is
consistent with behaviour in the Vco mode. As expected the bending cracks
are more developed for the larger shear spans.

Shear spans 2·9 and 4·1 m

The progression to failure was uncomplicated. Bendingcracks appeared
first and extended as displacementwas increased. After further increments,
web cracks formed suddenly as maximum load was reached and developed
into their final shape under increasing displacement and falling load. lt can
be seen from Fig. 3 for the 2·9 m test that there was a considerable residual
452 David Cullington

z 400
.Y.

ai
u
0
:: 200
CO
Shear span Shear span
"'
.!:
(f) 2.9m 5.7m

0
0 50 100 150 0 50 100 150
Vertical displacement, (mm)

FIG. 3. Graphs of shear force against displacement.

load capacity after the point of maximum load had been passed. Figure 2
shows the cracking after the retraction of the jack. Most of the bending
cracks had closed up and the web cracks had closed up partially.

Shear span 5·7 m

Cracking due to bending was more pronounced and, although a minor
shear crack formed in the web, bending failure appeared to be a possibility.
Before this could happen, however, new severe cracks appeared in the web at
a shallow angle to the horizontal. A further small increase in load and a
moderate increase in displacement took place as testing proceeded and the
cracks widened. At this point the Ioad-displacement curve was almost flat,
as can be seen in Fig. 3. Finally, the topflange buckled upwards and split the
beam from the Ioad point to the support.
Figure 2 shows the position of the cracks in test 3 as they would have
appeared ifthe split had closed up. Final failure was not in bending because
there was no significant crushing of the top flange nor yielding in the
prestressing wires. The bending cracks closed partially when the jack was
retracted.
Failure was not considered tobe ofthe ~r type because the failure planes
extended directly to the support and did not follow a flexural crack to the
soffit of the beam. The shear crack nearest to the Ioad point initiated in the
web and its role in the final collapse was secondary tothat of the two main
failure planes indicated in Fig. 2.

Shear span 1-6m

Failure was initiated by web cracking, as was the case for the other tests.
The main difference was that the failure Ioad was much greater than the web
cracking Ioad. Maximumload coincided with the development of severe web
cracks. Subsequently the beam was able to sustain a modest increase in
displacement without significant loss of resistance. Minor flexural cracking
occurred before failure.
 Assessment of Prestressed Bridge Beams 453

800 Cracking Ioad o

Maximum Ioad •

600 Unfactored
z
.:<:

Cll
~
.E
~ 400
•
"'
~

J Vcob

--
(fJ

...__ ----·----
200 Factared

0
0 2 4 6 8 10
Shear span, (m)

Frc. 4. Graph of shear rcsistance against shear span.

Measured and Calculated Shear Strengths

Figure 4 contains a comparison between the measured shear forces at
failure and the shear resistance calculated by a number of different methods.
In calculating the shear resistance it has been assumed that the composite
section resists the total Ioad, i.e. the dead Ioad and the applied Ioad. Values
have been obtained in accordance with BS 5400 with partial factors included
and assumed concrete characteristic strengths of 45 Njmm 2 for the slab and
60 N/mm 2 for the beams. These curves are termed 'factored' in Fig. 4. The
strength values were deduced from the original design calculations with an
allowance for ageing. Assumptions have also been made about the Ievel of
prestress. Further calculations have produced the unfactored curves. These
use mean concrete strengths of80 Njmm 2 for the slab and 90 Njmm 2 for the
beams obtained from core tests. Partial factors for concrete strength and
prestressing force, and some other conservative assumptions implicit in
BS 5400, have been removed. The latter comprise allowances for shrinkage,
for the effective depth adjustment in ~r and for the use of a reetangular
section in V:Oo (Ref. 3).

DISCUSSION

Performance of the Beams

The measured values of shear resistance lie between 1·6 and 2·7 times the
factared calculated resistance (see Fig. 4). At critical points in the bridgedeck
the measured resistance is at least 2·7 times the required ultimate Iimit state
454 David Cullington

shear foree. On this basis the in-situ resistanee after the applieation ofpartial
faetors exeeeds twiee the applied forees. The beams eomplied with BS 5400:
Part 4: 1978 for use in the Peel Green bridges.
lt was eneouraging that the resistanee of the beams did not fall to zero on
the appearanee ofweb eraeks. Beyond the point of eraeking the beams were
able to sustain a eonsiderable Ioad as displaeement was inereased. This
abilitywas least apparent at shear span of 1·6 m, but was eompensated for by
the highstrengthin this test. In all eases the appearanee and behaviour ofthe
beams after the formation of shear eraeking suggested that there might have
been arehing between the Ioad and the support balaneed by tension in the
wires. This is deseribed by Neilsen 4 in his work on Iimit methods of analysis.
It eould have been the souree of resistanee in the later stages of the tests.
This behaviour eonfers a measure of duetility to the beams. Duetile failure
is generally preferred to sudden brittle failure in struetures. If there is an
unexpeeted struetural inadequaey or overload, displaeements beeome !arge,
eraeking oeeurs and warning of a problern is given. lt also enables
redistribution ofloads to take plaee. The presenee ofweb reinforeerneut in a
prestressed beam is known to enable higher Ioads to be earried beyond the
onset ofweb eraeking. The absenee ofweb reinforeerneut in the beams tested
may have redueed duetility but the behaviour gave no eause for eoneern in
this respeet. There is no evidenee in the results to suggest that the absenee of
reinforeerneut led to premature failure.

Comparisons with Theory

Aeeording to BS 5400 the resistanee of the beams to web failure, ~o'
should be independent of shear span. This is beeause the prestress and the
shear reinforeerneut (none) are eonstant along the length and the eonerete
seetion is praetieally eonstant. Consequently, if the beams had failed in a
simple ~o mode, the shear forees at failure should have been roughly the
samein all eases. Some allowanee must be made for seatter in the results and
the faet that only four tests are reported. However, Fig. 4 clearly indieates
that the shear foree at failure reduees with inerease in shear span.
The ~r ealculation does not eorreetly identify the failure mode for these
beams as this is supposed to depend on ftexural eraeks developing into shear
failure. In praetiee the ~r ealculation serves a purpose beeause it eorreetly
indieates a lower shear resistanee at higher shear spans.
With allowanee for experimental seatter the measured strengths follow
the ~r eurve tolerably throughout the range of shear spans, including the
region where ~o mode is expeeted to be eritieal. It is not known to what
extent web modes of failure of the kind reported here were present in
Hawkins data. 2 For magnitude alone the data fit moderately weil (see Fig. 4).
The arehing/limit analysis method also Ieads to the predietion of a
 Assessment ol Prestressed Bridge Beams 455

progressive reduction in shear resistance as shear span increases. In this case,

however, there is no assumption about the interaction of ftexural and shear
cracking systems as implied by the V::r calculation.
An alternative interpretation for the observed behaviour is offered in Fig.
4. The curve Iabelied Vcob indicates the shear resistance calculated on the
basis that failure occurs when the principal tensile stress exceeds the tensile
strength of concrete at the bottom of the web rather than at the centroid of
the section. This possibility is discounted in the calculation for V::o given in
BS 5400: Part 4. It is not particularly significant in practice because V:,, as
calculated according to BS 5400 provides a lower, more conservative, value
of resistance.
As the shear span increases, ftexural stresses increase. The highest ftexural
stresses occur below the Ioad and so it is at this point that the curve for V..-ob
has been calculated. This provides a simple basis for the calculations, but in
practice cracking is likely to initiate closer to the support. Figure 4 also
indicates the Ioad at which a major crack formed in the web for the test at
1·6 m shear span. Cracking is evidence of a tensile stress exceeding the tensile
strength of concrete and is the criterion for web shear failure of a beam. The
V..ob curve lies below this and the other experimental points but follows the
trend observed in the tests.

Performance of Bridges
To date the tests have shown that the factared shear resistance derived
from the test is more than twice the design shear force at relevant points in
the Peel Green bridges. There appeared to be no adverse effects from the
absence of shear reinforcement, including any influence there might have
been from traffic loading, time-dependent changes or deterioration. On this
basis the Peel Greenbridges could have been allowed to pass an assessment
for further service. This conclusion could be applied to similar structures
provided that a reliable estimate of concrete strength was available.
However, the particular case of the Peel Green bridgeswas complicated by
the fact that they were tobe widened using BS 5400 design rules rather than
simply being assessed for further service as they stood.
To allow the beams to continue in service after widening would have
produced hybrid structures. In addition, the need to widen the structures
necessitated the presence of civil engineering works on the site and the
disturbance of traffic. The cost and effect of this presence may have been
reduced by rctaining the existing decks, but it would not have been
eliminated. The balance of advantage was therefore shifted towards
replacing the complete decks.
Had the bridges been subjected to service overloads, the tests indicate that
cracking would have been visible on the bottom flanges before failure
456 David Cullington

occurred. However, the cracks would not have been readily visible from the
ground, only from a close inspection. Web cracks would not have been
visible because the beams were placed side by side and the webs could not be
seen. As the tests were observed to terminate in a different form of shear
failure to that expected, some care is needed in inspections. lt illustrates a
problern ofusing a design code for assessment. If destructive tests are carried
out on single elements to verify an assessment, design codes that do not
identify the correct method of failure can be confusing.
The failure of a bridgedeck is likely tobe more ductile than the failure of a
single beam. This is because a deck generally has a capacity to redistribute
Ioad effects transversely and some of these paths may have ductility.
Provided that there is sufficient strength in the alternative Ioad paths the
failure Ioad would also be higher. A non-linear analysis would be necessary
to demonstrate Ioad redistribution effects theoretically.

CONCLUSIONS

1. The Peel Green bridges would not have passed an assessment using
BS 5400: Part 4: 1984 because they have no shear links in the webs. Tests
on the beams indicated that they would have been satisfactory for further
serv1ce.
2. The beams showed no ill effects from the absence of links, nor from 30
years ofmotorway service. Measured shear resistances, after allowing for
partial factors of safety, were at least twice the required resistances. The
beams possessed a measure of ductility.
3. The measured resistance reduced with increase in shear span and failure
occurred in the webs. Bendingcracks were visible on the bottom flanges in
all cases but did not participate in the failures. The combination of
characteristics is not consistent with V::o or V.:r behaviour as normally
understood.
4. The decrease of resistance with increasing shear span, accompanied by
failure in the webs, can be explained by considering the principal tensile
stress at the bottom of the web rather than at the centroid of the section.
Arehing action also offers an explanation for the decrease of resistance
with increasing shear span and may also explain the high resistance
sustained after cracking.
5. Other bridges that contain prestressed beams without shear reinforee-
rneut could be assessedas satisfactory provided they were similar to those
tested and the factored calculated resistance was sufficient. Unless the
beams can be inspected thoroughly and show no signs of distress in spite
of previous heavy loading, and the structures contain alternative Ioad
 Assessment of Prestressed Bridge Beams 457

paths, it may be preferable to ensure that the factored resistance is twice

the required resistance. This may be the case if allowance is made for the
increase in concrete strength with age.

ACKNOWLEDGEMENTS

The work described in this paper forms part of the programme of the
Transport and Road Research Laboratory and is published by permission
of the director. Help is gratefully acknowledged from Martin Crowe and
Charlie Parkinsou of Parkman Consulting Engineers, Les Clark, members
of TRRL Bridges Division who have helped with the programme and staff
from the Department of Transport, North West Regional Office.

REFERENCES

I. BS 5400: Part 4: Steel, concrete and composite bridges, Part 4. Code of practice
for the design of concrete bridges. British Standards Institution, London, 1984.
2. HAWKINS, N. M., The shear provision of AS CA35-SAA Code for prestressed
concrete. Civil Engineering Transactions, Institute of Engineers, Australia, CE6
(Sept. 1964) 103-16.
3. CLARK, L. A., Concrete Bridge Design to BS 5400. Construction Press, London,
1983.
4. NEILSEN, M. P., Limit Analysis und Concrete Plasticity. Prentice-Hall, New Jersey,
1984.
 41
Fatigue Assessment of Orthotropic Steel Bridge
Decks

C. BEALES and J. R. CUNINGHAME

Transport and Road Research Laboratory,
Crowthorne, Berkshire RG116AU, UK

ABSTRACT

Fatiguefailures haue occurred in the steel orthotropic decks of seueral bridges

in Europe, some after little more than a decade in service. This has led not only
to expensive inspection and repair procedures but has also highlighted the
d(fficulty of assessing these decks, which lie outside the scope of most design
codes.
Test procedures haue been developed at TRRL which haue been used to
determine the fatigue Jives of welded connections on particular bridges. The
data co/lected and experience gained haue led to a greater understanding of the
behauiour of orthotropic decks which can be used for a more generat
assessment offatigue lives. In particular, the in.fluence offactors which arenot
easy to model mathematically, such as the effect of the surfacing, can be
evaluated.
This paper describes the experimental techniques and the application of
these procedures to the deck of the Severn Crossing.
Load tests, with a vehicle ofknown axle weights, are used to determine the
influence surface of'.stress at the lt'elded connections. Tests carried out with the
bridgedeck sw{acing removed are compared with similar tests on the surfaced
bridge with the asphall at summer and winter temperatures. Where such trials
on actual hridges are impractical, data may he ohtainedfrom laboratory tests
on afull-scale deck panel. The merits o{ the different methods are discussed.
Fatigue tests onfull-scale specimens representing sections of the hridge deck
are used to determine the weid classification of the connection.

Any views expressed in this paper are not necessarily those of the Department of
Transport.
459
460 C. Beales and J. R. Cuninghame

A computerprogram developed at TRRL applies the procedures of BS 5400

(Part 10) to the measured stresses and the S-N data to estimatefatigue lives.
Fatigue lives have also been estimated from stress spectra obtained under
traffic loading.
The research highlights the sensitivity of the calculated fatigue lives to the
position of the wheel tracks and to the asphalt surfacing. It is concluded that,
for 'warst case' conditions,few connections in the decks of existing long span
bridges meet the 120-year UK design life for fatigue.

INTRODUCTION

Steel orthotropic bridge decks are used where lightweight structures are
required, that is for long span river crossings and for Iifting bridges.
Although few in number the long span bridges usually form vital road links
and represent major capital investments. The Severn Crossing (the adjacent
structures ofSevern Bridge, Beachley Viaduct and Wye Bridge), for example,
carries around 2 000 000 heavy goods vehicles per year between England and
Wales, and cost f11m to build in 1966. The second Severn crossing is
expected to cost around f150m plus flOOm for approach roads.
The earlier long span bridges, such as the Sevem Crossing and F orth
Bridge (mid-1960s), were built to British Standard 153. 1 This standardwas a
specification for steel girder bridges and the fatigue clauses made no special
allowance for the complexities of orthotropic decks. The current standard,
BS 5400: Part 10 (code of practice for fatigue), 2 issued in 1980, specifically
excludes orthotropic decks because of the complex stress analysis and
classification of details. Against this background TRRL developed
experimental techniques to assess the decks of the Severn Crossing for
fatigue prior to decisions being made about the requirements for
strengthening this part of the structure. These techniques have since been
used to assess welded connections on other bridges.

THE ORTHOTROPIC BRIDGE DECK

A typical orthotropic bridge deck is illustrated in Fig. 1; the main welded

connections are identified. For most long span bridges the longitudinal
stiffeners (troughs) are either trapezoidal or 'V' shaped. Bridge decks with
open stiffeners (bulb flats or angled plates) are not considered here.
Transverse stiffeners (crossbeams) are welded to the deck plate and troughs
during the fabrication of the deck panel. The crossbeams are subsequently
welded or bolted to cross-girders or diaphragms.
 Fatigue Assessment of Orthotropic Steel Bridge Decks 461

Several different designs of trough to crossbeam connection have been

used. For the Severn Crossing and Forth Bridge the troughs are in short
lengths, around 3-4·5 m long, which are butted up to the crossbeams. In later
bridges the troughs are around 14m long and pass through cut-outs in the
crossbeams, the troughs being 'spliced' tagether some distance away from
the crossbeams. In the UK this later type also has cope holes around the
apex of the trough and around the trough to deck plate weid. In other
European countries the troughs are often welded all round. These differences
are illustrated in Fig. 1.

Longitudinal deck plate butt weid

Longitudinal
stitfener

deck plate weid

Alternatove connectoons Crossbeam Doaphragm

F1G. 1. Main wcldcd connections in a typical orthotropic bridgc deck.

Two forms oflongitudinal stiffener splice joints are commonly used, either
an infill plate, butt welded to the troughs on a backing bar, or an overlapping
plate, fillet welded to the troughs.
Single-pass manual metal arc (MMA) welds were typically used for fillet
welded connections though automatic or semi-automatic welding is used for
long weid runs in modern bridges. Multi-pass butt welds are used to join the
sections of deck plate.
Temporary attachments employed during construction can cause fatigue
problems. One such attachment on the Severn Bridge was between a
temporary diaphragm and the soffit of the trapezoidal troughs. This
temporary 'ftotation' diaphragm was used to seal the bridgebox sections so
that they could be ftoated down river from a temporary storage site to a
position beneath the bridge from where they were lifted into place.
462 C. Beales and J. R. Cuninghame

FATIGUE ASSESSMENT OF ORTHOTROPIC BRIDGE DECKS

The fatigue design of UK bridges is governed by Part 10 of BS 5400. 2 While

the highway loading included in Part 10 is applicable to orthotropic decks,
clause 1.5.1 of the code states that 'the stress analysis and classification of
details in such a deck is very complex and is beyond the scope ofthis Part of
this British Standard'.
Analytical methods can be used to calculate stresses at welded
connections due to traffic loading but such calculations often require
verification from 'measured' stresses (stresses calculated from measured
strains). The accuracy with which the stress is determined is of great
importance since fatigue life is inversely proportional to the third or fifth
power of the stress range. Consequently, experimental stress measurements
by load tests may be the simplest and most accurate method of assessing
existing structures.
Some of the experimental methods have the advantage that the stress
reduction (and fatigue life enhancement) of the surfacing may also be
considered. This effect is extremely difficult to model mathematically and as
such 'this effect should only be taken into account on the evidence of special
tests or specialist advice' (BS 5400: Part 10, clause 6.1.4.2). lt will be shown
later that ifthis effect is not taken into account most connections close to the
wheel tracks would fail to meet the required 120-year fatigue life.
Fatigue tests on full-scale specimens representing sections of the deck
have been carried out to determine the classification of the connections.
These tests are described later in this paper.
With the loading defined in the code and experimental methods used to
determine the stress spectra and weid classifications, fatigue lives may be
calculated for details on a bridge.
A choice of fatigue assessment procedures are given in the code. The first
two methods are relatively simple and conservative, and involve the
calculation of stress at the detail under the loading of a defined 'standard
fatigue vehicle'. In the first method calculated stresses are compared with
values of limiting stress foreachdass of detail. In the second method stress
spectra are calculated and a simplified darnage calculation performed. The
most rigorous and least conservative method involves the calculation of
stress influence lines for a specified set of commercial vehicles followed by a
darnage summation using the Palmgren~Miner rule.
The third procedure has been used as the basis for the assessment of
details on the Severn Crossing using a range of techniques to determine
stress spectra. These techniques and procedures can be applied to any bridge
of this type.
 Fatigue Assessment of Orthotropic Steel Bridge Decks 463

DETERMINING STRESS SPECTRA

Stress spectra were obtained in a number of ways from measured strains

using gauges installed close to the welded connections.

1 Static Tests in the Laboratory

A full-scale deck panel, 14·3m long by 3·7m wide, was tested at TRRL
under a single static wheelload. The panelwas supported in a reaction frame
carrying a single wheel and axle assembly which could be moved to any
position over the panel. The wheel and tyre were typical of those used on
heavy goods vehicles in the UK. The wheel was loaded to 20 kN though
Ioads up to 100 kN were possible.
The Ioad was applied at over 500 locations on the panel to determine the
full influence surface of stress for the gauge positions. Typicallongitudinal
and transverse influence lines obtained in the tests are shown in Fig. 2(a) for
the crossbeam to deck plate connection. They show that the effect of the
wheelload is very localised, the effective length of the influence lines being
around 1 m; this is the case for most connections close to the deck plate.
Influence lines for the vehicle types of BS 5400 (Part 10, Table 11) were
calculated by superposition of the single wheel stresses and stress spectra
calculated using the 'rainflow' cycle counting method.
The laboratory method enables precise control of the wheel Ioad and
position to be obtained. Environmental effects are eliminated. The method
does not allow the effect of surfacing tobe assessed. Asphaltic materials are
viscoelastic, that is the properties depend on the rate of loading.
Consequently, the composite effect of the surfacing on the steel deck is quite
different under a static Ioad from that under a moving vehicle. For this
reason laboratory tests were carried out with the panel unsurfaced.
Connections near the centre of the panel were tested to avoid problems
with edge effects. Aceurate modelling of support conditions is important,
especially for connections at the crossbeam. The web ofbox and diaphragm
plates are difficult to accommodate in a test panel ofthistype and were not
incorporated in this particular panel.

2 Static Tests on the Bridge

These tests were carried out with a two-axle 16-t test vehicle of known
wheel Ioads and axle spacings. An area of the bridge deck surfacing (20m
long by 3·7 m wide), over the instrumented details, was removed. Although
the influence ofthe surfacing issmall under a static Ioad, strain readings take
many minutes to stabilise because of creep of the asphalt. 3
Strains were recorded with the vehicle stationary at a number oflocations
464 C. Beales and J. R. Cuninghame

300mm
1-----1

Gaugeposition

'"
r.n
-20

Tranverse influence line

-40

40

20
Gaugeposition

1m
~-------.-1

'"
"'-20
longitudinal influence lines

-40

-60 --Stress for 20 kN single wheel Ioad

- - -calculated stress for 32kN Ioad

(a) Measurements on test panel

Stress for front wheel of test vehicle

whee1 Ioad = 32kN
20

'"
UJ-20

-40

_ _ Surfac1ng rem oved

-60
- -38mm of mastic asphalt surfacing
temperature 38°C
38mm of mast1c asphalt surfacing
- - - temperature 14°C
(b) Measurements on bridge

FIG. 2. Typical influence lincs--crossbeam to deck plate connection.

on the deck, the Ioad position being referenced to one of the front wheels.
Interaction between the front and rear or nearside and offside wheels is
negligible for most details because the influence of the wheel Ioad is very
localised. Consequently, single wheel influence line data can be obtained
from these tests and the data treated in the same way as for Iabaratory static
tests.
 Fatigue Assessment of Orthotropic Steel Bridge Decks 465

Exceptions are the connection between the trough and crossbeam and the
stiffener splice joint, where the longitudinal influence line for a single wheel is
very long. In these cases it is not possible toseparate the strains for a single
wheel from the vehicle loading.
Good agreement was generally found between the stresses measured in
the Iabaratory and on the bridge. In the longitudinal influence line of Fig.
2(a) the 20 kN 'measured' stresses from the Iabaratory deck panel tests have
been factared to the eq uivalent loading of the front wheel of the test vehicle
(32 kN). These data can be compared with the influence line obtained from
the test vehicle on the unsurfaced deck (Fig. 2(b)). The longitudinal scale of
Fig. 2(b) has been increased but it can be seen that the peak stresses are in
good agreement with the factared peak stresses from the Iabaratory tests.
The advantages oftesting on the bridge are that there are no doubts about
the accuracy ofthe model, and details which cannot easily be accommodated
in a test panel can be tested. Bridge tests are also considerably eheaper and
quicker to organise than Iabaratory tests, which may require the fabrication
of a deck panel. The disadvantage of bridge tests is that lane closures are
necessary to accommodate the test vehicle and the removal of the surfacing.
In addition, loading from vehicles in adjacent lanes or carriageways, and
changes in t~mperature, cause noise or drift of the strain gauge readings.
Aceurate positioning of the test vehicle is also diffi.cult; errors in excess of
10 mm are normally noticeable on influence line plots.

3 Dynamic Tests on the Bridge

After the tests on the unsurfaced decks were complete the area was
resurfaced. Hand-laid mastic asphalt was used for one half of the area and
an epoxy asphalt on the other. In both cases a nominal 4 mm layer of an
epoxy-based material was used as a waterproofing membrane.
To assess the effect of the bridgedeck surfacings, strains were monitared
using high-speed tape recording equipment as the test vehicle was driven
over the strain gauged Connections. A simple but effective method of
recording the transverse position of the vehicle relative to the connection
was by measuring from a reference line to the tyre imprint in a patch of sand.
The vehicle was driven at constant speed and longitudinal scaling was
calculated from electrical pulses on the magnetic tape generated as the
vehicle passed over detector strips on the road located before and after the
connection. Skilful driving was needed to run consistently close to the target
position. Around 20-30 'runs' of the test vehicle were required to achieve a
reasonable transverse distribution of points to describe accurately the full
influence surface of stress.
Typical longitudinal influence lines are given in Fig. 2(b) for the mastic
surfacing at 14 and 38 C. There is a considerable reduction in stress
466 C. Beales and J. R. Cuninghame

compared with the data from the unsurfaced deck, the greatest effect being
when the surfacing was cold and stiff. Because ofthis temperature sensitivity
load test data were required at a range of surfacing temperatures to assess
fully the surfacing effect.
Previous tests 3 had shown that the effect of the surfacing was also
dependent on vehicle speed, the greatest reduction in stress being obtained
for fast-moving vehicles. In practice the changewas found to be small for
speeds in excess of 32 km/h (20 mph) and this speedwas adopted for the test
vehicle runs. Since most vehicles travel at speeds in excess of 32 km/h a
calculation based on these data will tend to be conservative.
Single wheel influence line data were obtained following the analogue-to-
digital conversion of the raw data. Once again this was not possible for
details with long infl.uence lines.
Each test requires a lane closure for only a few hours. Compared to
laboratory methods, the uncertainties associated with modelling the deck
and surfacing are avoided. The temperature and traffic data are collected
separately from the strain data so that revised estimates offatigue life can be
produced to reflect changes, for example in traffic composition, without the
necessity of repeating Ioad tests. The disadvantage of this method is the
logistical one öf planning tests at the right times to obtain data at a range of
temperatures.

4 Direct Measurements under Traffic Loading

'Intelligent' electronic equipment was used to monitor the response of the
strain gauges to dynamic Ioads from the moving traffic. Stress spectra were
output directly.
The advantage of this methodisthat the effect of temperature and vehicle
speed on the composite action of the surfacing are automatically taken into
account. The only disadvantage is that data are recorded in real time and
many months ofrecordings were needed to assess the behaviour ofthe deck
over the full range of surfacing temperatures and to take account of
variations in traffic. An estimate oflong-term trends in traffic can be used to
predict fatigue lives from these data.

FATIGUE TESTSAND JOINT CLASSIFICATIONS

The fatigue classes in BS 5400: Part 10 2 are based on data from small
laboratory specimens tested under axial loading. Only tests under tensile
loading were included and the statistical analysis ofthe data took account of
the absence of residual stresses in the specimens. 4
The classification of welded joints in orthotropic decks may differ from
 Fatigue Assessment of Orthotrupie Steel Bridge Decks 467

the code because of the complex stress distribution and high stress gradients
around the joint. Also there is evidence that fatigue strengthisdifferent for a
joint subject to bending stress, 5 as is often the case in an orthotropic deck,
compared with axial stress.
In tests at TRRL fatigue specimens were made to full scale and large
enough to retain high residual stresses after welding. Care was taken to
obtain good fit-up and to follow the correct weld procedures so that the
specimens represented joints on a bridge without significant defects.

I
I
I
I
I
I
J

'\
Spl ice plate
Fillet weid

Single lap slice joint

I I
I I
I I
I I ~
r -:---------- f- ~ ""-Fillet weid
I
I
I
I I
~---------y

Outer splice plate

Double lap splice joint

300,_----------------------------------------------------------,

"'E
~
z

0 Double lap splice joint

/>.Single lap splice joint

30~--------------~------~--------------_J_______L~------~
10 5 5 106 5 10 7

Endurance (cycles)

FIG. 3. Fatigue test results for the fillet welded trough splice joints.
468 C. Beales and J. R. Cuninghame

Loading was arranged to replicate the stress distribution around the joint as
measured on the bridge or deck panel. Strain gauges were installed at the
same locations on fatigue test specimens as on the deck panel or bridge, so
that the stress used to define the fatigue strength was directly comparable to
that obtained from wheelloads.
Constant amplitude fatigue tests have been carried out on a number of
deck plate joints. For example, tests on joints between lengths of
longitudinal stiffener (trough splice joints) were reported by Cuninghame 6
for butt welded joints. Further tests have been carried out on fillet welded
splice joints as these may be more economical to fabricate and less sensitive
to misalignment ofthe troughs. Specimens consisting of2 m long by 600 mm
wide sections of deck containing a trough with a splice joint at mid-length
were loaded in four-point bending. Results are shown in Fig. 3 in the form of
S-N curves for longitudinal stress at the apex of the trough adjacent to the
splice weid.
lt will be seen that the fillet weldedjoints have lower fatigue strength than
those with butt welds and so are less likely to be suitable for heavily
trafficked bridges. In terms of BS 5400 dassifications the butt weid is dass D,
the fillet welded joint with inner and outer splice plates falls within dass E,
and the joint with outer splice only in dass G. Fatigue cracks in joints with
outer splice only have been reported by Darnell et al. 7

FATIGUE LIFE CALCULATION METHODS

Three calculation methods were developed to suit the methods used to

obtain the stress data. Method 1 uses the single wheel influence line data
from the Iabaratory or bridge tests to determine the fatigue life for the
unsurfaced deck. Method 2 uses a similar procedure to calculate Jives from
the data from the surfaced deck but goes on to assess the overalllife, taking
into account the flow of vehides across the bridge at different temperatures
throughout the year. Method 3 uses the data obtained under traffic loading
directly.

Method 1: Unsurfaced Deck-Static Test Data

A computer program has been developed at TRRL to calculate fatigue
Jives from experimentally determined stress influence surface data and the
loading of BS 5400: Part 10. Data preparation mainly involves the entry of
the singie wheel stress data and the coordinates of the loading positions.
Influence lines for the vehide types described in Table 11 of the code are
calculated by superposition of the single wheel data. The vehide influence
lines are calculated at a number of transverse positions across the deck and
 Fatigue Assessment of Orthotropic Steel Bridge Decks 469

due allowance is made for the transverse distribution of vehicles defined in

the 'multiple paths' clause (C.l.4) of the code. F or the assessment of existing
structures the centreline of the transverse distribution of traffic can be
determined by observation on the bridge.
A stress spectrum is calculated from the vehicle influence line data using
the 'rainflow' cycle counting technique. Fatigue Jives are calculated using
appropriate S-N data and applying the Palmgren-Miner method of darnage
summation.
Once the data are assembled the program runtime is short. lt is therefore
easy to repeat the calculations with traffic centred at a number of different
transverse positions to determine the effect on fatigue life. This facility can be
used to assess the effect of proposed changes to the position of traffic lanes
on a bridge.

Method 2: Surfaced Deck-Dynamic Test Data

Fatigue Jives may be calculated in a similar way from data obtained from
Ioad tests on the surfaced bridge. The data processing is more complex,
analogue-to-digital conversion of the strains recorded on to magnetic tape
first being required. It is also necessary to deal with the fact that the influence
lines from the test vehicle are at random transverse positions.
Fatigue Jives thus calculated relate to the surfacing temperature at the
time of the Ioad test. To determine an overall life relating to conditions
throughout the year, allowance must be made for the variation of
temperature throughout the day and year, and the variation in the flow of
traffic at different times of the day.
Howells 8 recorded temperatures for a 2-year period (1969-71) on the Wye
Bridge, presenting the data as the proportion oftime, in 5'C bands, for each
hour of the day. Data were also obtained giving the flow of HGVs over the
Severn Bridgeforeach hour ofthe day, averaged over a year. These two sets
of data were combined to obtain an estimate of the traffic flow at different
temperatures. It was then possible to apportion the fatigue Jives at specific
temperatures to determine an overall value relating to the whole year.

Method 3: Surfaced Deck-Traffic Stress Data

Stresses at 12 weldedjoints in the deck ofthe Wye Bridge were monitored
for 13 months. Stress spectra were stored by the recording equipment at 1 h
intervals. Traffic flow was monitored during the recording period and the
time and duration of interruptions to the flow over the instrumented area
due, for example, to lane closures or traffic management schemes was noted.
Fatigue darnage was calculated from the stress spectra for the total
recording period and factored to take account of interruptions to give an
estimate of fatigue life for each joint.
470 C. Reales and J. R. Cuninghame

CALCULATED FATIGUE LIVES

Table 1 gives fatigue lives for six connections on the Wye Bridge, calculated
using the three methods described above. All Jives relate to a 2·3%
probability of failure (using mean-2 s.d. S-N data) and a traffic flow of
800000 HGVs per annum. This is the observed traffic flow and is less than
the 1·5 million vehicles required by BS 5400. The centre of the transverse
distributionoftraffle is assumed (in Methods 1 and 2) tobe at the centreline
ofthe carriageway. Where appropriate, surfaced lives relate to the hand-laid
mastic experimental surfacing. Weld classes are given in brackets.

TABLE 1
Estimated fatigue lives

Detail Fatigue life (years)

(weid dass)
Unsurfaced deck Surfaced deck

Method 1 Method 2 Method 3

Static Dynamic Trafik
test data test data stress data
- - - - - - - - - -·----- ·----- --------------

Troughjdeck (F) 6·5 94 > 120

Longitudinal butt weid (F) 5·9 > 120° > 120
Web of box/deck (0) 41 > 120 40b
Trough/crossbeam (G) 4·3 13 18
Crossbeam/deck (D) 94 > 120 >120
Transverse butt weid (F) 35 > 120° > 120
--~~-···-

"Insufficient data for full assessment but expected value > 120 years.
b Surfacing cracked over web.

The lives given in Table 1 do not take account of darnage sustained prior
to resurfacing but would apply to repaired joints or to similar connections
on a new bridge. It can be seen that, for the unsurfaced deck, none of the
details meet the 120-year design life required by the code.
Some ofthe assumptions in Method 2 were chosen to give a conservative
(low) estimate of fatigue life. For two of the connections the data obtained
were at an insufficient range oftemperatures to enable overall fatigue Jives to
be calculated. Nevertheless, the data suggest that both these joints would
have fatigue Jives in excess of 120 years on the surfaced deck.
Data for the Method 2 calculation were obtained within 6 months of
resurfacing the strain gauged area. Collection of data for the Method 3
calculation began 12 months after resurfacing, by which time a longitudinal
 Fatigue Assessment (J[ Orthotropic Steel Bridge Decks 471

crack had developed over the web of the box. This explains the loss of
effectiveness of the surfacing for this connection and the estimated life
similar to that for the unsurfaced deck in this case.

Effect of Surfacing
The influence lines in Fig. 2(b) illustrate the large reduction in stress in the
steel deck due to the bridgedeck surfacing. They also show the effect of the
variability ofthe properties ofthe surfacing with temperature. Table 1 shows
that, for most connections, very short fatigue Jives calculated for the
unsurfaced deck are increased to above the 120-year design life by the
surfacing. One exception is the trough to crossbeam connection where the
stiffening effect of the surfacing has only a small influence on the stresses at
the apex ofthe trough. Failures oftbis connection have occurred in service.
The results suggest that failure of the trough to deck plate connection
should not occur within the lifetime of the bridge (strictly there is a
calculated 2·3% probability offailure within 94 years) surfaced with mastic
asphalt on an epoxy waterproofing layer. lt should be noted that this
surfacing system is stiffer than that originally used on the Severn Crossing
and this difference may account for the fact that cracks have occurred in this
connection after less than 20 years in service.
The results also show that the effectiveness of the surfacing can be
completely lost if cracks develop in the surfacing over the welded
connection. On some bridges it is quite common for cracks to develop over
the web of the box within weeks of resurfacing. These are normally
controlled by sawcutting the surfacing and sealing the joint with a flexible
bituminous material. Cracks can also develop over other hard spots, such as
the troughs and crossbeams, that is at the very connections the surfacing
should be helping. This partly explains the difficulty of incorporating a
surfacing factor in the codes. The search continues for a surfacing material
with high stiffness (preferably at high and low temperatures) and a long
fatigue life at the cold/brittle and hot/high strain extremes.

Transverse Position of Traffic

The transverse influence line shown in Fig. 2(a) illustrates the very
localised effect of the wheelload on the stresses at a typical connection near
the deck plate. Consequently, the fatigue life is very sensitive to the trans-
verse position of the traffic. Using the fatigue assessment program it can be
shown that the fatigue life ofthe web ofbox to deck plate connection on Wye
Bridge could be increased from 12 to 80 years (unsurfaced deck, weid class F)
by moving the traffic lanes 300 mm towards the centre of the bridge.
Consideration may be given at the designstage to the relative positions of
the traffic lanes and connections such as the web of box to deck and the
472 C. Beales and J. R. Cuninghame

longitudinal deck plate butt welds. F or more freq uently occurring

connections, such as the trough to deck plate connection (300 mm intervals
on Wye Bridge), there is little scope for improvement by this method.

CONCLUSIONS

The analysis of stresses and classification of details in orthotropic decks is

complex and beyond the scope of current codes of practice for fatigue. Tests
are TRRL and elsewhere have produced fatigue data for the classification of
most of the welded connections. Test procedures have been used to obtain
stress spectra for these connections on a number of bridges from which
fatigue lives have been calculated.
The data have been used to quantify the effect ofthe bridgedeck surfacing
and to show the sensitivity of the calculated fatigue life to the transverse
position of the traffic.
The results suggest that few connections close to the wheel tracks meet the
120-year design life if the effect of the surfacing is ignored. The difficulty of
incorporating a surfacing factor in the codes is recognised.

ACKNOWLEDGEMENTS

The work described in this paper forms part of the programme of the
Transport and Road Research Laboratory and the paper is published by
permission of the director. The work was carried out in the Fatigue Section
of the Bridges Division, TRRL. The leadership and guidance of Mr D. E.
Nunn is gratefully acknowledged.

REFERENCES

1. BS 153, Specification for steel girder bridges. Part 38: Stresses. British Standards
Institution, London, 1958.
2. BS 5400, Steel, concrete and composite bridges. Part 10: Code of practice for
fatigue. British Standards Institution, London, 1980.
3. MoRRIS, S. A. H., Stresses under dynamic wheel loading in a surfaced steel
orthotropic deck with V-stiffeners. TRRL Report SR237, Transport and Road
Research Laboratory, Crowthorne, 1976.
4. GuRNEY, T. R. and MADDOX, S. J., Are-analysis offatigue data for welded joints
in steel. Welding Research International, 3(4) (1973) 1-54.
5. MADDOX, S. J., Fatigue of welded joints loaded in bending. TRRL Report
SR84UC, Transport and Road Research Laboratory, Crowthorne, 1974.
 Fatigue Assessment oj' Orthotropic Steel Bridge Decks 473

6. CuNINGHAME, J. R., Steel bridge decks: fatigue performance of joints between

longitudinal stiffeners. TRRL Report LR1066, Transport and Road Research
Laboratory, Crowthorne, 1982.
7. DARNELL, W. J., EDMONDS, F. D., RUTLEDGE, J. C. and YONG, P. M. F., Fatigue
problems in the orthotropic deck of the Auckland Harbour Bridge. Pacific
structural steel conference, Auckland, New Zealand, 1986.
8. HowELLS, H., Temperature spectra recorded in surfacings on steel bridge decks.
TRRL Report LR587, Transport and Road Research Laboratory, Crowthorne,
1973.
 42
Assessment and Rehabilitation of Suspension
Bridges

PETER G. BOCKLAND
Ruckland and Taylor Ltd,
1591 Bowser Avenue, North Vancouver, BC, Canada V7P 2Y4

ABSTRACT

The upgrading of suspension bridges, if required, is an extremely costly

procedure. lt is therefore worth some considerable effort on the part of the
engineer to determine if rehabilitation is in fact really needed for the safety of
the bridge or whether it is only required in order to satisfy a code.
The importance is emphasized of taking a criticallook at criteria, including
traffic and wind Ioads, aerodynamic stability and seismic response. Recent
developments in wind tunnel testing are presented, with observations on Iang-
term cable stretch, reliability oftemperature measurements and the use of Iimit
states design. Three examples are taken from suspension bridges ranging in
span from under 200m to over 800 m.

INTRODUCTION

Suspensionbridgesare no moreimmune to the ravages oftime and the need

for upgrading than other bridges-but they are potentially far more
expensive to renovate. Worse than that, suspension bridges are usually on
major routes where any interruption of service can have a huge effect on the
community. lt is therefore necessary to be doubly sure about a suspension
bridge before committing torepair it. Is repair really necessary? How best to
do it?
Techniques have been evolved over the last few years to answer these two
questions with more certainty than before.
475
476 Peter G. Ruckland

LOADING

The first item might best be to take a hard look at the loading. Is it really
what the code says? What conventional wisdom says? Is it more? Less? Since
the only function of a bridge is to support safely and effectively the loads
imposed on it, particularly traffic loads, this is a useful starting place.
When the Lions' Gate Bridge in Vancouver, BC, was examined in the mid-
1970s it was noticed that although it was designed for two lanes it was in fact
carrying three lanes of densely packed traffic-about 60 000 vehicles/day. At
first sight this appeared to be a 50% overload.
Instead of closing the bridge it was decided to take a new look at the
loading. The trick is in formulating and then solving the probabilistic
equations which describe the maximum loading expected to occur during
the lifetime of the bridge-or some other defined period-with a certain
degree of probability. This work has been reported elsewhere.1.2 The results
for the Lions' Gate Bridge are shown in Fig. 1.
The interesting points here are that for the major components of the

40

JS

30

25

~
la
A-ASHTO HS 25 ( EXTRAPOLATED}

20
g
~

I
~
:
>
r--
IS DESIGN LENGTH FOR TOWERS
~

I
DESIGN UNGTH FOR
CA BUS' (AIU-IENT

10

100 200 300 400 soo 600 700 800 900 1000 1100

LOADED L<NGTH (m)

FIG. I. Traftic loading for Lions' Gate Bridge. 'Design loading' based on 'observed Ioad' was
used for evaluation. The bridge was originally designed for 'normal' and 'eongested' Ioads.
 Assessment and Rehabilitation of Suspension Bridges 477

bridge-towers and cables-which are governed by long loaded lengths, the

original designers' loading was in fact adequate. Such is not the case for
shorter loaded lengths such as affect the design of the stiffening truss.
Whether or not this situation is acceptable, it is at least the right way round;
failure ofthe stiffening truss would be far less catastrophic than failure ofthe
towers or cables (or anchorages).
Wind loading should not escape scrutiny either. The wind force is a
function of several parameters, including expected wind speed, wind
direction, turbulence of the wind, the shape of the bridge cross-section and
the dynamic properties of the bridge.
The cost of determining these factors is tiny compared to the cost of
modifying the bridge. Although the bridge engineer is trained to accept wind
loadings from the codes, this is an instance where the engineer's involvement
with criteria may pay handsome dividends to the dient.

ANALYSIS AND SURVEY

Having decided with at least some confidence on the Ioads to be applied to

the bridge, the next significant step is to check the condition of the bridge.
Some aspects of this are obvious and relatively easy: looking for corrosion
and fatigue cracking, measuring the tension in cable-band bolts, and even
inspecting some of the cable itself.
Some other areas are virtually impossible to inspect with today's
technology: embedded steel in the anchorages, for example, and the insides
of spiral strands.
However, the real key to success comes in determining the dead Ioad
stresses in the structure. The technique is as follows, and it depends on the
creation of a computer program that will accurately model the effects of
changes on a suspension bridge.
Suspension bridges are difficult to analyse. They are non-linear in their
behaviour and it is not possible, even theoretically, to have a condition in
which all members are stressless at the same time. This makes the use of
conventional non-linear analysis programs clumsy, at best.
For the examples herein the programSABER was developed. SABER not
only calculates the effects of loading, it also allows the structure to be
changed. Each loading or change is calculated directly from the preceding
condition in a step-by-step manner. The steps are:

(a) Model the structure in the computer as it was built, which is not
necessarily as it was designed.
(b) Impose on the model all known changes that have occurred to the
478 Peter G. Ruckland

bridge-extra dead Ioad, strengthened members, settlement of the

foundations, drag of the anchorages, and so on.
(c) At this stage the computerwill predict the expected geometry of the
bridge.
(d) The bridge is also surveyed, at least for the key points such as
elevations of midspans and tower-top deftections. The surveyed
geometry, corrected for temperature effects, should be identical to
that predicted by the computer. Experience with about a dozen
suspension bridges suggests that such is almost never the case.
(e) The reasons for discrepancies between the predicted and observed
geometries must now be sought. They will be caused by all those
changes to the bridge that have occurred but which were not known.
Apart from those suggested in (b), these could include long-term
stretch ofthe cables or hangers, slippage of cables in saddles, errors in
the construction ofthe bridge and incorrect information having been
given to the evaluating engineer.
In some cases the causes of discrepancies cannot be pinned down
exactly, but so far experience has shown that the effects of the causes
can be bracketted so that there are upper and lower Iimits.
(f) Finally, when all permanent effects on the bridge are known, the
resulting stresses can be calculated and added to the stresses from the
various live and other loadings to obtain a complete picture of the
real condition of the bridge.

The benefits of this powerful technique will be shown in the examples

presented.
In some cases it is useful to supplement the information gained by site
measurements. For example, ifthe polygon representing the cable geometry
is known, all cable and hanger forces in a span can be determined by
measuring the tension in one hanger.

TEMPERATURE EFFECTS

Surveying a suspension bridge only yields information if the temperature of

the bridge is uniform, known and constant during the time of the survey.
Conventional wisdom has been to survey the bridge during an overcast
windless night, typically around 02:00 hours when the ambient temperature
is fairly constant.
An interesting study on the Lions' Gate Bridge, however, revealed that the
bridge continued to experience radiant cooling even though its temperature
was soc below ambient. The measurements taken are shown in Fig. 2.
 Assessment and Rehabilitation ol Suspension Bridges 479

(OC)
N.E TOWER LEG, DECK LEVEL
2•

X AM81ENT AT GROUND UNDER VIADUtT

22
0 AMBIENT AT H MCS DISCOVERY, VANCOUVE R

20

S.E. TOWER LEG, DECK LEVEL

18

16
CLEAR,
SOME
CLOUD RAIN
I"

12
X

10 0
EAST CABLE
0 NEAR NORTH e e 0
CABLE BENT 0 .,

6 0

"+-~.- . .~-.-.-..-.-.-.-~~-.-.-.-.-..-.-.-.-~~--~---
0600 080D 1000 1200 1400 1600 1800 2000 2200 2400 0200 0400 0600 0800 1000 1200 h
I
NOON pm MIDNIGHT am
80 APRIL 03 180 APRIL 04

Fru. 2. Temperature variation with time, Lions' Gate Bridge.

SAFETY F ACTORS

Tothis point, no mention has been made ofsafety factors. Since this question
is addressed in another paper, 3 only two points will be made here.
Safety margins are provided to allow for all the unknowns in a situation,
and Iimit states design reflects this philosophy by providing !arger Ioad
factors for the Ioads that are least weil known. Most major suspension
bridges now in existence were not designed by Iimit states design, however,
and it is possible to derive some benefit from this situation when considering
the main members, which carry mostly dead Ioad, and to which the smallest
Ioad factors usually apply. As a caveat, note that the safety factors should
also retlect the importance of the members considered.
The second pointisthat if safety margins reflect ignorance then, when the
ignorance is reduced by the methods described herein, perhaps the safety
margins may also be reduced. Sensible judgement is required.

EXAMPLES

To illustrate the points made some examples of their use are given.

Lions' Gate Bridge, Vancouver, BC, Canada (Fig. 3)

In 1975 the concrete deck on the 670 m approach viaduct was replaced by
a lighter, wider steel orthotropic deck in a series of 61-hour closures at night.
480 Peter G. Ruckland

FIG. 3. Lions' Gate Bridge, Vancouver, BC, Canada.

By making the deck act compositely with the steel girders their ultimate
capacity was enhanced. This technique was later used on the George
Washington Bridge, the Golden Gate Bridge and others, and has even been
patented (not by the originators of the idea) in the USA!
But the most interesting aspects of this study were the research into
loading mentioned earlier, and the first use of the analysis-and-survey
technique.
In this case the following known changes were modelled:
(i) unstressed lengths of members as per shop drawings;
(ii) settlement of the north tower foundation and the north cable bent
(side tower) as surveyed;
(iii) some addition to the dead load; and
(iv) no movement of the anchorages as observed initially from a
measurement of expansionjoints and confirmed later by triangulated
survey.
The geometry surveyed could only be attained in the computer model if the
following changes were also added:
(v) long-term stretch of the main cables of about 0·038%;
(vi) long-term stretch of the hangers of about 0·075 % ; and
(vii) slippage ofthe cables through the north cable bent saddles of60 mm.
Stretch of the main cables (a hexagonal arrangement of helical wound
 Assessment and Rehabilitation of Suspension Bridges 481

structural strand) induced deftections to the tower tops, which added to the
P-b effects. By using Iimit states design, even with conservative factors, the
towers were deemed capable of carrying the induced moment.
Stretch of the hangers is most pronounced where the hangers are long,
next to the towers, and throws extra load on to the bearings. This was later
confirmed by measuring the bearing reactions, which were double what they
were intended tobe under dead load only. A solution is to lower the bearings
when the bridge is next renovated, to throw more load back on to the
hangers.
Slippage of the cables through the cable bent saddles was known to have
occurred during construction in 1938, but was thought to have been stopped.
Its continuance imposed some alarmingly high stresses on the cable bent
legs. Those stresses could not ha ve been revealed by any other method andin
fact had been missed in an evaluation of the bridge 4 years earlier by a
different firm of suspension bridge engineers.
The solution to the cable bent problern (Fig. 4) was to rotate the two
foundations. This was mostly done under traffic by cutting through the mass
concrete footings and inserting 3m long rocker bearings (Fig. 5).
At the south end ofthe bridge a short rocker bisects the angle ofthe cables
(Fig. 6). There was no slippage here, but the long-term lengthening of the
cables must have caused the tops of the rockers to move towards the centre
ofthe bridge. Close inspection, however, yielded no evidence that the pin had
ever rotated. The 20-year-old paint film was still intact. The stresses in the

SLIPPAGE TO 1938 22mm

SLIPPAGE 1938 TO 1974 38 mm
CABLE STRETCH 65mm
TOTAL 125 mm

\
\
\
\
..,E
0
"'

100
~.193

EAST LEG ONLY

FIG. 4. Distortion of north cable bent Lions" Gate Bridge.

482 Peter G. Ruckland

FIG. 5. The north cable bent footing, formerly of mass concrete, was cut and 'pinned' to
allow straightening of the bent.

FIG. 6. South rocker posts with cover plates added.

 Assessment and Rehabilitation of Suspension Bridges 483

rocker were thus indeterminate, depending on when the cables stretched and
when the rocker last moved.
Jacking up the cable in order to attend totherocker pin would have been
very expensive. The pin had a half-shell bronze bushing but its bearing
pressure was ten times the allowable pressure of the 1970s.
Instead it was decided to add plates to the flanges oftherocker so that its
ultimate capacity would be greater than that of the bronze bushing. This
way, if failure should occur, it would be a non-catastrophic failure of the
bushing rather than catastrophic failure of the rocker itself.
Also on Lions' Gate Bridge, an exhaustive study was made of the natural
frequencies, mode shapes and damping values of the bridge. 4 lt was
interesting that the initial measurements gave frequencies ofup to 1·38 times
the predicted values. Since this ratio is close to J2 and frequency is
proportional to j(K/ M), where K is the stiffness and M the mass, it was
thought that an error of 2 had been made.
Such was not the case, however. The main reason was that the breather
joints in the deck were not 'breathing' at the small amplitudes measured.
This significantly altered both the lateral and torsional frequencies, which
were found to be coupled, and showed up the danger of relying on
measurements alone, which in this case would predict, for example, a critical
velocity for aerodynamic instability of about 38% greater than the true
expected value.

Tacoma Narrows Bridge, USA

In professional collaboration with Arvid Grant and Associates, Buckland
and Taylor Ltd used the analysis and survey technique to evaluate the
Tacoma Narrows Bridge (Fig. 7) in 1984. The bridge has been well
maintained, but what the evaluation showed was that the hangers have, as
expected, elongated and thus increased shears in the stiffening trusses near
the towers.
There was another learning experience in the survey. The bridge profile
was surveyed from end to end and back during one night with no traffic on
the bridge while readings were taken. The survey did not close by 1 ft and a
simple error of one digit was suspected.
More careful analysis, however, revealed that the bridge temperature had
been changing slightly while the survey was conducted. Thus the elevation of
each point changed slightly between successive readings at that point. When
a correction was made for this the survey closed.

Belgo Bridge, Quebec, Canada

This Iittle suspension bridge, which carries Iogs on a conveyor over a span
of 183m, was built in 1917 and examined in 1982. By using the techniq ues
484 Peter G. Buckland

FIG. 7. Tacoma Narrows Bridge, WA, USA.

just described it was found that because of increases in dead Ioad, seizing of
rollers supporting cable saddles and differential Settlement of foundations
some tower members were carrying 2·5 times their calculated allowable
Ioads under dead Ioad only.
Because steel entering the anchorages, buried under a newsprint mill
building, was known to be corroded but was not fully inspectable, it was
decided to transform the suspension bridge into a cable-stayed bridge,
apparently the first time this has been done.
The project took place during two closures of the bridge in successive
summers and has been described in more detail elsewhere, but one of the
critical aspects was the necessity of having the bridge partly carried by the
old suspension system and partly by the new cable-stayed system with a
central tower for almost a year (Fig. 8).
 Assessment and Rehabilitation of Suspension Bridges 485

FIG. 8. Belgo Bridge in transition from a suspension bridge to a cable-stayed bridge.

During the hot summer months the suspension bridge sheds its Ioad to the
cable-staye d system, thus tending to overload the stiffening trusses in shear
at the temporary end ofthe cable-stayed system. In the harsh Quebec winter
the opposite occurs: the suspension bridge tries to take alJ the Ioad, including
that added for the cable-stayed system. This would overload the suspension
system. Tuning the bridgetobe always within these Iimits took some careful
engineering!
Another comment on safety factors is appropriate here. If it had been
decided that the stresses in the tower must be limited to about 0·6 of
theoretical capacity, as is normal, the bridge would have had to be closed.
As the bridge is the only feed of Iogs to the mill for 8 months of the year,
this would have had a disastraus effect on the economy of the mill and the
town.
In fact some stresses were at least 50% greater than the theoretical
capacities (i.e. 2·5 times normal allowable) under dead Ioad alone. A
temporary prop was installed at midspan which reduced the critical stresses
by 10%, and the bridge operated with the stresses at least 35% above
theoretical capacity (i.e. 2·25 times normal allowable stress) for 2 years. The
argument was that the bridge had, in essence, survived a Ioad test, and that
486 Peter G. Ruckland

reducing the Ioad by 10% would be satisfactory for the short term. The client
was made aware ofthe risks and the consequences ofthe alternatives.lt was
an informed client who made the decision as to which alternative to accept.
The benefits of continued operation were the client's. lt was not desirable for
the risks to belong to the engineer.

AERODYNAMICS

Several existing suspension bridges have been deemed tobe inadequate for
aerodynamic stability. Some bridges, such as Golden Gate and the Bronx-
Whitestone Bridge, have been modified to improve performance, and for
one (the original Tacoma Narrows) modifications came too late.
It has been found that by using the most sophisticated wind tunnel
techniques and by testing in turbulent wind flow with full aeroelastic models,
taut strip or taut tube models, the expense of solving a problernthat does not
really exist can be avoided.
Once again, Lions' Gate Bridge provides a good example. Following
movement during a hurricane in 1962, section modelwind tunnel tests were
performed the following year. The bridgewas given a clean bill ofhealth, but
interpretation ofthe results was open to doubt. Basedon the test results, this
author estimated a critical velocity of only 55 mph (25 m/s). Such a number
was consistent with incipient motion recorded in winds gusting to 72 mph
(32 m/s) as these gusts would coincide with a mean wind of 50-60 mph
(22-27 m/s).
What became apparent from the full model test in turbulent flow-
believed to be the largest model tested in turbulent flow-was that the
observed motion was in fact buffeting response and not the onset of
instability. 5

SUMMARY

Suspensionbridgesare expensive and disruptive to repair. Before embarking

on a repair programme it pays to be very sure about the condition of the
bridge, including those conditions that cannot be determined by routine
inspection. For this purpose the analysis-and-survey technique is
indispensable.
It also pays to take a carefullook at all design criteria, including loads and
safety Ievels. These conclusions are based on experience gained on a dozen
existing suspension bridges. The examples in this paper represent a variety of
sizes and degrees of rehabilitation required.
 Assessment and Rehabilitation of Suspension Bridges 487

ACKNOWLEDGEMENTS

The figures relating to Lions' Gate Bridge originally appeared in two papers
published in 1981 by the Canadian Journal of' Civil Engineering. Permission
to reproduce them here is gratefully acknowledged.

REFERENCES

1. BUCKLAND, P. G., McBRYDE, J. P., NAVIN, F. P. 0. and ZIDEK, J. V., Trafik loading
oflong span bridges. Proceedings ofConference on Bridge Engineering, St Louis.
TRB Transportalion Research Record 665, Vol. 2, Washington, DC, 1978.
2. BucKLAND, P. G., NAVIN, F. P. D., ZIDEK, J. V. and McBRYDE, J. P., Proposed
vehicle loading of long span bridges. J. Struct. Div. A SCE (April 1980).
3. BucKLAND, P. G., Canada's advanced national standard on bridge evaluation.
In Procccdings International Confcrcncc on Brich!,e Management, ed. J. E.
Harding ct al.. Elsevier A pplied Science Publishers, Barking, 1990, pp. 575-583.
4. BucKLAND, P. G., HooLEY, R., MoRGENSTERN, B. D., RAINER, J. H. and VAN SELST,
A. M., Suspension bridge vibrations: computer and measured. J. Struct. Div.
ASCE(May 1979).
5. IRWIN, H. P. A. H. and ScHUYLER, G. D., Experiments on a full aeroelastic model
of Lions' Gate Bridge in smooth and turbulent fiow. National Research Council
of Canada Labaratory Technical Report LTR-LA-206, Ottawa, 1977.
 43

Structural Assessment of a Bridge with Transversal

Cracks

CHARLES ABDUNUR
Laboratoire Centrat des Ponts et Chaussees, Paris, France
and
JEAN-LOUIS DUCHENE
Laboratoire Regional des Ponts et Chaussees, Le Bourget, France

ABSTRACT

To estimate the influence ol cracks on the mechanical hehaviour ol a bridge,

these discontinuities are assimilated to a series ol elastic or plastic hinges,
alternating with sound heam segments andjointly setting up a new system in
equilibrium. The main difjiculty resides in the calculation ol the hinge residual
fiexural stiffness, which is a .fimction subject to various parameters and
assumptions. To off'er a practical solution andfacilitate realistic modelling,
we developed an experimental method hased on the relationship hetween
bending moment and resulting curvature. Under a lest Ioad the measured
curvature redistrihution re{iects the new mechanical response ol the structure
and Ieads to the actual stiffnessfunctions ol the hinges. Theseare introduced
into a program ol structural analysis. The new statical SJ'Stem, thus defined,
enables the real stresses and residual strength to he estimated. In the
technologicalfield, prolotype inclinometry instruments of'high accuracy had to
be developed. This method was tested in the lahoratory then successfully
applied on site.

INTRODUCTION

Flexure cracks are among the main structural defects observed while
inspecting reinforced or prestressed concrete bridges. To assess the relative
residual strength and optimise the strengthening of such defective structures,
489
490 Charles Abdunur and Jean-Louis Duchene

it is first necessary to explore their new statical systems. The actual stresses
can thus be predicted under a given loading.

BASIC ASSUMPTION

The statical system may be obtained by the following basic assumption

illustrated in Fig. 1: the cracked sections and their disturbed vicinities are
assimilated to a series of elastic or plastic hinges H, alternating with sound
beam segments B and jointly setting up a new system in equilibrium.

CRACKED SECTION PROFILE

l.... l"--ool;;
l
HINGE H SEGMENT B

HINGEH
~hM~~~
STATICAL SYSTEM
FIG. I. ModeHing assumptions of cracked sections of a bridge.

The resulting stress redistribution depends on the reduction in the flexural

stiffnesses of the cracked sections and on the positions of these
discontinuities with respect to the initial moment diagram.

EVALUATION OF FLEXURAL STIFFNESS

The sound beam segments B can generally be assumed to retain their initial
flexural rigidities [E/] 0 , which are theoretically given. For concrete bridges
the calculation of [E/] 0 sometimes remains approximate, because of the
varying modulus and the random contribution of non-structural elements to
the deck moment of inertia. However, as the relative effect of cracking is
under discussion, a reasonable value for intact sections can be adopted for
both the initial and new mechanical systems.
The main difficulty obviously resides in the realistic calculation of the
residual stiffness [E/h for the cracked sections or hinges. In reinforced or
prestressed concrete bridges, this quantity greatly depends on the extent and
geometry of cracks, the constitutive laws of materials under cyclic Ioads, the
redistribution of bond stresses, and the sense and 'viscosity' of crack
movement.
 Structural Assessment of a Bridge with Transversal Cracks 491

The Experimental Option

Prior to theoretical modelling, it is essential to find an accurate
experimental method for evaluating the flexural stiffness of a cracked
section. Three main reasons can be given:
(a) Several variables of the complex function of residual stiffness are
subject to assumptions which can only be verified by experiment.
(b) Rapid applications are often required to assess, in pragmatic terms,
the actual mechanical behaviour of a structure.
(c) A specific metrology is at present being successfully developed for
this purpose.

The Assessment Principle

The proposed experimental evaluation of stiffness, at a point x along the
bridge, is based on the simple relationship between the bending moment
M(x), corresponding curvature O'(x) and flexural rigidity El(x).
If, under a convenient test loading, the measured curvature diagram ofthe
bridge is established, and ifthe stiffnesses [E/] 0 ofthe sound beam segments
are given, then the actual bending moments and the cracked section
stiffnesses [E/]H can be deduced through a chain application ofthe classical
beam equation. The procedure is explained below.

Adopted Rotation Measurements

Curvature variation is usually obtained from the strain profile, measured
with strain gauges in a sound section and with additional displacement
sensors in a cracked section. This extensometry technique is obviously
inconceivable for most bridges concerned, with multiple cracks and
inaccessible residual sections.
To cope with the nature of the problem, we turned to broader aspects of
measurement as provided by inclinometry, where substantial technological
progress has recently been achieved: 10- 6 rad accuracy, miniaturisation,
robustness, simplicity and movability. Theserotation measurements enable
the detection of discontinuities in the bridge profile, thereby obtaining main
angular deformations and sieving out local disturbances.
From a theoretical viewpoint, the following expressions of curvature O'(x)
are equivalent for asound beam section of height h and position x, with a
bending moment M(x) and a strain difference ~~> between the extreme fibres:

O'(x) = dO = ~s = M(x)
dx h El(x)
lndinomclry / ""' Extcnsomctry
For a cracked section the extensometry expression is very difficult to
492 Charles Ahdunur and Jean-Louis Duchene

apply, owing to stress concentration and complex strain redistribution. The

inclinometry expression remains reasonably valid.

Procedure and Interpretation

A test load cycle is applied and the resulting rotations are measured, at
sufficiently close paces, to detect discontinuities and deduce curvature
throughout the bridge spans.
Figure 2 outlines the interpretation of results at a given discontinuity or
hinge. Fora convenient Ioad configuration producing a moment M(x), and
at every rotation discontinuity 11.0, the curvature variation e'(x) is plotted
over a length covering the hingeH and the near parts ofboth adjacent so und
beam segments B 1 and B2 , conserving their initial stiffnesses [E/] 0 .

(a) (b) EI
[E 1) 0 [E1Ux) I

)M2
X
0
8, H 82 ·X 8, H 82 X

FIG. 2. (a) Moments M and stiffnesscs EI in cracked scctions H and sound scgmcnts B. (b)
Redistribution of curvature 0' and stiffness EI.

The curvature response, in its redistributed form, refl.ects the new

mechanical behaviour of the structure.
Applying the beam equation in B 1 , Hand B2 Ieads easily to the relative
residual stiffness of the hinge or cracked section:
[E/]H(x) 0~
[E/] 0 O~(x)

the right-hand expression being deduced from the plotted curve.

This quantity is not only a function of position within the hinge but may
also depend on the applied moment M(x) if the hinge is plasticised.
With a reliable value of [E/] 0 the residual stiffness can be expressed in
absolute form, [E/]H(x).

THE NEW STATICAL SYSTEM

The relative or absolute stiffness function of each cracked or plasticised

section is scrutinised from a durability viewpoint then inserted into a
 Structural Assessment ol a Bridge with Transversal Cracks 493

computer program of structural analysis. The new statical system, thus

defined, enables
-the prediction of the real stresses in the structure under any given
loading, and
-the assessment of the Ioad-hearing capacity and the optimal needs for
strengthening.
The procedure can be repeated at a later stage to verify the effectiveness of
eventual repairs or simply to follow up a time-dependent mechanical
change.

PRELIMINARY FIELD INVESTIGATION

Two identically constructed bridges, one with transversal cracks and one
apparently sound, were scrutinised under loading to test the feasibility and
sensitivity of the proposed method as regards instrumentation and adopted
criteria.
Under the same Ioad system of a convoy, the observed rotations and
deflections were respectively 25% and 30% greater in the damaged bridge
(Fig. 3). Extensometry measurements, simultaneously carried out for mere
comparison, confirmed the same percentage difference between the strain
profiles of corresponding sound sections of the two structures.
A third damaged bridge of a similar three-span configuration was tested
by inducing a controlled vertical fluctuation at one end support. Section
rotations 0 and crack gap () remained linear with Ievel change, except for
a short interval around zero (Fig. 4).
The instantaneous crack width and strain responses to flexure are more
pronounced on opening the crack than on closing it (Figs 4 and 5).
These general observations show the feasibility of consistent rotation
measurements on site and their aptitude to detect crack effects.

e ·6
(10 rad)
300
SOUND STRUCTURE
CRACKED STRUCTURE

FIG. 3. Effcct of cracks on thc rotation influcncc linc for a bridgc scction X undcr a passing
convoy.
494 Charles Abdunur and Jean-Louis Duchene

-4
8 (1 0 rad) 0 (11 m)
OPENif'.X>
4 20

---
8 --

15
t. Y(mm)

CLOSING

FIG. 4. Gap 6 and rotation 0 of a cracked section due to Ievel change ± ßy at an end
support.

neutral axis )
------------

FIG. 5. Inftuence of the sense of ftexure on strain redistribution in a cracked section.

LABORATORY TESTS

The influence of cracks on curvature and stiffness redistribution is

examined in the laboratory, under a gradually increasing number of
parameters, to explore further possibilities of rotation measurements and to
verify or modify certain assumptions in view of a more realistic modelling.
Strains and displacements were also measured but strictly for confirmation
purposes of test results.
Research began on the preliminary model of a slotted steel section and
continued on a prestressed concrete beam.

The Preliminary Model

A simply supported steel 1-beam was subjected to a four-point flexure
producing a constant moment over the span centre. Rotations 00 and
deflections y 0 were measured.
A vertical slot was then cut at midspan, starting from the tension flange
(Fig. 6). Under the same loading, the same measurements were repeated and
completed by rotations 8; and 0,, respectively along the intact and slotted
flanges, taken at close intervals throughout the span.
The results, summarised in Fig. 7(a), show the aptitude of rotation
 Structural Assessment of a Bridge with Transversal Cracks 495

600N 600N

:-._sTRAIN -i
MOBILE
INCLINOMETER
;
"'1-
T
GAUGES
1
20

~==~==~==Jtj10~&=0=T======~
1 1 FIG. 6. Slotted steel rolled section under
'.53.,.!.....--- 67------.! ( dimensions in cm) four-point bending.

-6
CURVATURE ( 10 rad/cm)
100

d 6; INTACT FLANGE
dx
50
8i (INTACT FLANGE)

8 (INITIAL)
/ 0 --------- X(CM) X(CM)
-_._I!-----
5 10 15 20 5 10 15 20
SLOT SLOT

(a) (b)
FIG. 7. (a) Measured rotation rcdistribution araund the slot in a loaded steel 1-beam. (b)
Dcduced curvaturc redistribution.

measurements to define the stiffness redistribution zone araund the slot.

Rotations Oi' at the intact ftange, can be differentiated (Fig. 7(b)) to obtain the
curvature or the inverse stiffness function (since the moment is constant).
Rotations a,,
at the slotted flange, evidently have a load-dependent
discontinuity.
The theoretical model ofthe slotted beam, using the new stiffness function,
yielded a mechanical behaviour in perfect agreement with direct
measurements such as deftections and strain profiles of sound sections.

The Concrete Beam

We proceeded in the same way to test a simply supported micro-concrete
T-beam prestressed with an unbonded tendon and lightly reinforced (Fig. 8).
Before cracking convenient four-point bending was applied. Rotations
and deftections were measured to determine the actual initial stiffness [E/] 0
of the beam.
Three flexure cracks were induced under a temporary, cautiously
increasing, point Ioad at midspan. The cracked sections were instrumented
with strain gauges and displacement sensors. 1 At one crack, after a
minimum removal of the concrete cover, strain gauges were also fixed on
both the unbonded tendon and the reinforcement bars to estimate the tensile
forces actually developed.
496 Charles Abdunur and Jean-Louis Duchene

F -so-

I I
I SENSORSAND I
I I
+9+
~4 0 __.., STRAIN GAUGES ~ DIMENSIONS IN CM
-------600------· NOTTOSCALE

FIG. 8. Instrumented cracks of a prestressed beam.

Direct determinatian af curvature

The initial four-point bending was resumed. Corresponding differential
rotations were recorded by two coupled inclinometers at close intervals
along the intact upper flange of the cracked sections and their vicinities.
Other measurements were maintained.
The curvature redistribution thus obtained, under a constant bending
moment, leads to theinverserelative stiffness [E/] 0 /[E/]H as a function of
position x over the influence length of the three cracks. The experimental
diagram is shown in Fig. 9.

Camparisan with ather data

As in the previous steel beam, the theoretical model based on the present
stiffness redistribution was used to calculate the increase in midspan
deflection, support rotations and normal fibre strain. These computed
increments, owing to cracking, fully agree with direct parallel measurements.
Moreover, the curvature-deduced residual stiffness (Fig. 9) coincides with
that estimated from the combined response of strain gauges and
displacement sensors at each crack.

LOAD, F = 7 KN

MEASUREMENTPACE:
0.1 h
- 0.5 h

(x)cm
CRACKS

FIG. 9. Redistribution curve of the inverse relative ftexural stiffness over a beam segment
with thrce cracks.
 Structural Assessment ol a Bridge with Transversal Cracks 497

INTACT ~ / CO'JCRETE
------~--~ ·~'--.:::::-.;;;:::s:::=~'====--
' ' FROM ROTATION
'' ........_ DEDUCED CURVATURE
BEFORE CRACKING /
CRACKED \~ FOR ALL MATERIALS ..........
, ""- REINFORCEMENT

\e-UN~DEDTENDON i....
.............................. f (10- 6)

0 +500 EXTENSION

FIG. I0. Strain profilcs and strain point valucs for concretc and steel at a cracked section.
Thc rotation-deduccd curvature coincides with strain data.

Figure 10 shows the logical yet complicated concrete strain profile

obtained at a cracked section, hence the impracticality of classical
extensometry for operational use in similar cases. It can be seen, however,
that strain measurements on the upper remaining concrete section,
combined with those on reinforcement bars, do confirm the curvature
deduced from rotations.

Other Parameters
To study the effects of statical indeterminancy and band redistribution,
research is being extended respectively to continuous spans with a similar
concrete section and to simply supported beams with various reinforcement
patterns.

FIRST APPLICATIONS ON SITE

The rotation method is now being used on reinforced and prestressed

concrete bridges with transversal cracks. At damaged sections the deduced
relative residual stiffnesses confirm the bell-shaped distribution obtained in
the Iabaratory and shown in Fig. 9. For the samerelative height of cracks,
field and Iabaratory stiffness values are very close at corresponding
positions.
Over damaged segments the optimal spacing of rotation readings varies
from 0-1 to 0-5 of the beam height.
In some cases a simplified version ofthe method may be sufficient, as in the
following example. At a flexure crack of an inspected bridge, the actual
rotation discontinuity MI, measured by coupled inclinometers on either side,
is represented in Fig. 11 as a function of bending moment, applied by a test
Ioad. This relationship is generally non-linear but can be linear for cracks
already open under dead Ioad. Hence the non-linear deviation, at the
498 Charles Abdunur and Jean-Louis Duchene

ANGULAR -•
OPENING Ae·eA- e 8 (1o rad)

100

CRACKED
50

THEORETICAL UNCRACKED

~-----------MOMENT
0 4 6

FIG. II. Crack angular opening (measured on a bridge by two inclinometers) versus applied
moment.

extremity of the curve, marks the advance of the crack tip under critical
loading.
We can therefore replace the residual fiexural stiffness [ E/]H(x) by another
experimental coefficient, k = M/ 118, which in this case is constant.
As already explained, the coefficients thus obtained for similar cracks
enable the bridge to be recalculated on a more realistic basis.

CONCLUSION

An experimental method was developed to determine, under a test Ioad, the

curvature redistribution throughout the spans of a damagcd bridge. This
redistribution Ieads to the evaluation of the actual residual stiffnesses of the
cracked sections.
The new statical system of the structure can thus be defined to enable

-the prediction ofthe real stresses in the bridge under any given loading,
and
-the assessment of the Ioad-hearing capacity and the optimal needs for
strengthening.

The whole procedure can be repeated at a later stage to verify the

effectiveness of repairs or simply to follow up a time-dependent mechanical
change.
Supported by numerical techniques, this experimental approach is
suitable not only for operational ends but also in research as a means to
tackle the various parameters involved and to establish a more realistic
theoretical model.
 Structural Assessment of a Bridge with Transversal Cracks 499

REFERENCES
1. CHATELAIN, J., BRUNEAU, J. and DucHE:NE, J.-L., Estimation par des essais de
chargement du defaut de resistance a Ia ftexion de certains tabliers en beton
precontraint. International Conference on Inspection, Maintenance and Repair
of Road and Railway Bridges, Brussels-Paris, 1981.
2. GoDART, B. and DucHE:NE, J-L., Intervention sur Je Pont de Champigny/Yonne.
Conference on the Inspection and Testing of Structures, EPNC, Paris, 1987.
3. GüDART, B., Approche par l'auscultation et Je calcul du fonctionnement de ponts
en beton precontraint fissures. Euro-American Conference on the Rehabilitation
of Structures, CEBTP, Saint- Remy-les-Chevreuse, France, 1987.
4. ABDUNUR, C. and DucHE:NE, J.-L., Mesures de rotations pour Je schema statique
d'un ouvrage fissure. International Conference on Measurements and Testing on
Civil Engineering, Lyon, 1988.
5. CHATELAIN, J. and GüDART, B., Evaluation de l'etat mecanique reel de ponts en
beton precontraint. IABSE Symposium, Helsinki, 1988.
 44
Reliability Analysis Applied to Deteriorating Bridge
Structures

JoNATHAN G. M. Wooo, R. AsHLEY JoHNSON

Mott MacDonald Special Services Division,
20-26 Wellesley Road, Croydon, Surrey CR9 2UL, UK
and
CHARLES ELLINAS
Advanced Mechanics and Engineering. 4 Frederick Sanger Road,
Surrey Research Park, Guildf'ord, Surrey GU2 5YT, UK

ABSTRACT

Reliahility analysis techniqucs dcvelopcd for steel hridgcs and offshorc

structures prol'idc a methodoloJ.;y 1rhich can also he used to el'a!uate the safety
ol suhstandard or detcriorating reinforccd concretc structurcs and arch
hridgcs. This is illustratcd hy rcf'ercncc to cascs ol structurcs with alkali
aggregate reaction and chloridc-induccd corrosion damage. The shortage of'
data on the statistics and defimnation characteristics of'concretefailure Iimits
the quantitath·c application and hiJ.;hliJ.;hts the nccessity for more testing ol
deteriorated reinfi>rccd concrcte. Considcration of' ovcrall reliahility must
inc!ude assessment ol risks inhercnt in road closures for inspection and
rcmedia/1\'ork, the unccrtaintics ofinspection procedures, and the reliahility ol
repair and corrosion control.

INTRODUCTION

In the UK the CIRIA Report 63 1 has provided the framework for

probabilistic methods for calibrating partial factors for Ioads and strength
to achieve more consistent reliability in design. This partial factor approach,
in a somewhat simplified form, has been incorporated in British Standards
like BS 8110 2 for reinforced concrete and BS 5400 3 for bridges with
501
502 J. G. M. Wood, R. A. Johnwn and C. Ellinas

widespread, but not universal, support in the engineering profession.

Neither code provides a margin for deterioration.
This approachwas used more rigorously in the aftermath of the steel box
girder failures in the UK, Germany and Australia for design of remedial
works, and to guide the major Merrison programme 4 · 5 of research and
testing on steel box girders which led to the simplified BS 5400: Part 3 steel
bridge code.
The methodology has since been further developed and refined in the UK
offshore industry, predominantly for steel structures in the North Sea. 6 The
logical framework provided by a probabilistic approach to determining the
reliability of structures is tailor-made for resolving the major problems
which arise when deterioration in bridges creates uncertainty about their
current and future safety. In this paper we outline how techniques used in the
design and appraisal of steel bridge structures and offshore structures are
being applied to the particular problems of deterioration in reinforced
concrete bridges with substandard details, alkali aggregate reaction andjor
chloride-induced corrosion, and the potential for their application to other
structures like arch bridges.
It is our view that reliability analysis must be based on a thorough
knowledge and understanding of those factors which actually Iead to

Original
safety
Ievei and
design life

FIG. I. The cyclc of appraisal and reliability assessment.

 Reliability Analysis of Deteriorating Bridge Structures 503

structural failure as distinct from the convenient oversimplifications of

structural behaviour in design codes. The analysis of well-recorded cases of
structural failure 7 - 9 provides the first source of data. However, it soon
becomes clear that actual failures are fortunately too infrequent to provide
statistical data on loading, strength or the reliability of inspection.
As there are not enough case studies we must use site measurements and
Iabaratory tests to provide us with the statistical data on the variability of
loading and strength that is needed for proper analysis. Data on the
reliability of inspection and monitaring techniques in detecting the severity
and rate of structural deterioration will also be required. While general
corrosion produces spalling and AARproduces cracking as warning signs,
pitting corrosion is not easily detectable. Many of the most structurally
sensitive parts of structures are not readily accessible for inspection.
Determination of actual construction details and ground conditions
provides another area of substantial uncertainty, especially for arch bridges.
Because the methodology of reliability analysis has been weil established
in other disciplines 10 · 11 it is not necessary to reinvent the framework of
analysis when applying it to concrete structures (Fig. 1).

RELIABILITY IN DETERIORATING STRUCTURES

The majority of deteriorating reinforced concrete structures are old enough

to have been analysed for design without benefit of a computer, often for
smaller Ioads and temperature ranges than currently required. Some oftheir
designers used robust over design as a substitute for precision. Others,
without benefit of current knowledge, provided inadequate designs,
particularly in shear, corbel design and in reinforeerneut detailing.
Inadequate provision for articulation and thermal effects is a frequent
source of massive overstress in appraisals but with a small serviceability risk
and no collapse hazard.
The application of reliability approach to determining the rising risks
associated with a deteriorating structure, which may also have been built to
a lower design standard, identifies these types of uncertainty:

(a) Inherent uncertainties of variability of loading and strength.

(b) Uncertainties in our knowledge ofloads, structural response, ground
conditions, initial strength and deterioration which can be reduced
by control analysis or testing.
(c) Uncertainties in the risks to the public arising indirectly from
highway restrictions during inspection or remedial works and
directly from falling spalled concrete or collapse.
504 J. G. M. Wood, R. A. Johnwn and C. Ellinas

Loading
Dead Ioad can be determined more accurately for a built structure.
Increases in the weight of services and surfacing can be controlled to enable
Ioads and partial factors to be reduced.
The most severe traffic loadings on short span structures often arise from
short special vehicles (e.g. cranes) in the 50-100 ton range, not from the
variability inherent in normal C& U commerciallorry traffic. Traffic control
of these 'over 50 ton' vehicles and lane restrictions, evaluated from Ioad
effects from the most adverse actual vehicle combinations, enable more
accurate and reduced loadings to be adopted for substandard structures.
However, reliability becomes very sensitive to the ability of police and
highway authorities to enforce controls. For reliability analysis HA and HB
loading are completely inappropriate. Garages of onerous real vehicles with
associated frequencies of occurrence must be used.
In normal design a simplified approach to environmental Ioads (wind,
overall temperature and differential temperature) is possible. For reliability
analysis these must be considered explicitly with the refined traffic Ioad
analysis, particularly where probability of coincidence is high (e.g. high
winds and/or .cold increase the risk of accidents, which cause high Ioad
conditions of close-packed slow-moving vehicles on long span bridges).
Impact loading from overheight or crashing vehicles, which is poorly
represented in standards, gives a substandard risk of darnage from
overloading and must be considered.
When considering members in which AAR is developing the effects of
restrained expansion must be considered as 'Ioads' in a similar way to
thermal expansions.

Analysis of Overall Structural Response

Because most deteriorating structures predate computer analysis a
detailed elastic appraisal inevitably shows areas of substantial over-
provision of strength and areas of underprovision. Typically in bridge decks
longitudinal ftexural steel is overprovided except at the edge, while there is
underprovision of transverse ftexural steel. However, concrete is not as
elastic as computer elements. Creep and cracking permit redistribution,
provided ductility is available from suitably detailed reinforeerneut which
has not been degraded by deterioration. The initial reserves of strength and
ductility enable many deteriorating structures to continue in service without
undue risks, despite severe local deterioration away from critical structural
elements. The risks arising from localised deterioration in a structure are
very sensitive to the availability of alternative Ioad paths due to structural
redundancy. The non-linear analysis of the structure from the local distress
'serviceability' stage through to the 'collapse' Iimit state enables this to be
 Reliability Analysis of Deteriorating Bridge Structures 505

evaluated. It also provides an indication of potential cracking conditions

which inspections may detect as a record of overloading and/or distress.
When comparing the result of detailed computer analysis with simplified
code analysis the implicit assumptions of concrete design codes must be
considered. The writers of codes carefully exclude all difficult analytical
problems by placing Iimits on the detailing and proportioning of structures
to ensure that

(a) strength calculations are based on simplified rules which do not

relate to the actual mechanism of failure;
(b) structures are sufficiently ductile so that even if the idealised
distributions of forces and stressesthat come from structural analysis
do not occur in the real structure the ductility will enable them tobe
redistributed before the ultimate Iimit state is reached; and
(c) many second-order effects like differential temperature and locked-in
stresses from early thermal effects can be discounted due to the
d uctili ty provided.

This characteristic of concrete structural codes mirrors the detailing

requirements for proportioning stiffeners and plates in steel codes to prevent
the initiation ofbuckling before yield has occurred. The lessons learned from

"\ 25 years
corrosion
"0 started
ro
0
..J
30 years corrosion
severe pitting brittle

Elongation
FIG. 2. Loss of ductility and strength in reinforced concrete with chloride ingress.

steel structures (in particular box girders) is that when buckling is initiated
collapses occur suddenly without warning. The embrittlement of concrete
structures (Fig. 2) by corrosion of steel or concrete deterioration produces
the same type of rapid, catastrophic failure mode as buckling. Inspection
and/or proof loading provide no warning of impending failures. lt is the
embrittlement from deterioration in concrete structures which provides the
greatest increase in risk, not just change in strength.
506 J. G. M. Wood, R. A. Johnson and C. Ellinas

EFFECTS OF ALKALI AGGREGATE REACTION ON MATERIAL

PROPERTIES AND CONSTITUTIVE MODELLING

Our knowledge of the structural effects of alkali aggregate reaction has

increased substantially since the early 1980s. AAR is probably the most
difficult deterioration phenomenon to be analysed in concrete. In 1983
initial results of the effects of simple uniform AAR expansion in the finite
element analysis of a pile cap were presentedY Following a review
of those results and the research Iiterature we stopped finite element
analysis of alkali aggregate reaction until sufficient data on the actual
stiffness, expansion and strength characteristics of concrete with AAR were
available for a valid constitutive model to be evolved for finite element
analysis. Over the last two years we have restarted finite element analysis
work on AAR using data now available 13 This detailed evaluation is only
appropriate to the most severely expansive and sensitively detailed parts of
structures as identified using the Institution of Structural Engineers'
approach. 14
Firstly, it has been necessary to evaluate the expansive characteristics of
the concrete and the extent to which this is restrained by the reinforcement.
It is now clear that the forces generated by restraint of a severe reaction can
be sufficient to yield the steel in most reinforced concrete members and it can
produce compressive failures at expansion joints. Therefore these forces
must be evaluated and considered as explicit forces in the analysis. In some
structures they serve to prestress the concrete and have a beneficial effect. In
other structures it is the magnitude of these forces combined with the
reduction in strength which darnage the structure. The magnitude of forces
developed are very sensitive to the reactive aggregate and gel characteristics
and to the conditions of exposure.
lt has become clear that there is substantial variability in expansion at all
Ievels, ranging from that araund individual reactive particles up to the
variation between adjacent pours. Even between nominally identical
concrete pours this variability is high. The expansion is randomly generated
by the variations in reactive aggregate concentration and alkali concen-
tration within the materials. The local rate of expansion is very sensitive to
maisture supply. It is this variability of expansion which gives rise to the
characteristic map cracking in AAR. At its smallest scale it gives rise to the
microcracking which changes the Young's modulus, tensile strength and
compressive strength of concrete.
The stiffness changes in concrete with AAR are probably the most
sensitive measure of microcracking. lt isasimpler and morequantitative test
and more representative of a structural volume of material than the
qualitative techniques of petrography which help us understand the
 Reliability Analysis of Deteriorating Bridge Structures 507

N'
~ 5 Normal Severe AAR
z
::1
~
-:;;
iii
c:
E
0
z OL--~~---------=-' OL---~0~~-------~500
0 500
Averagestrain (microstrain) Average strain (microstrain)

FIG. 3. Effect of AAR on stiffness of concrete.

deterioration. Figure 3 shows the change in Ioad deformation characteristics

of concretes with alkali aggregate reaction for a normal concrete and a
severely deteriorated concrete. It can be seen that there is a reduction in the
Young's modulus, a development of increased hysteresis and a progressive
packing down of the core. This approach and its correlation with strength
changes is described in Ref. 15.
In evaluating the tensile strength of concrete our understanding has been
hampered by the difficulties of sample preparation and testing. The Brazilian
splitting test is simple but unrepresentative of normal structural behaviour.
lt is sensitive to the inaccuracy of normal coring from existing structures. lts
only merit seems tobethat lots of people have done it before. Gaspressure
testing is an interesting measure of tensile strength but difficult to relate to
normal structural behaviour. Because the mode offailure ofmost concern in
structures with alkali aggregate reaction is shear the torsion testing of
cores 15 as illustrated in Fig. 4 gives us a much better measure ofthe changed

FIG. 4. Torsion testing of concrete cores.

508 J. G. M. Wood, R. A. Johnson and C. Ellinas

"'E
-
z
E
-""
u
w
1/)
1/)
41
....
c
:!::: x After Ono
Vi • Trend line

Cyl. compressive strength Fcyl ( N/mm 2 )

FIG. 5. Compressive strength test comparison, southwest concrete (5% reactive chart).

strength for evaluation of shear and bond behaviour. Results related to the
microcracking as indicated by reduced Young's modulus are shown in Fig. 5.
The growing train of strength below the design characteristic substantially
increases failure risk.
Having identified the changes in concrete properties from AAR we can
consider structural forms which will be sensitive to these changes.

PUNCHING SHEAR

An example is a fl.at slab bridge deck, with developing AAR, reinforced only
in its top and bottarn faces and supported on discrete columns. The rules for
the evaluation of punching shear strength of normal concretes are largely
empirically based on a limited series of tests. We were surprised to find in a
review of the literature, and in discussions with those who carried out the
work, that until recently none of these tests determined the Ioad
displacement characteristics to failure. Testing we have recently carried out
shows that
(a) this punching shear failure mode is extremely brittle, as shown in Fig.
6; there is no inspectable warning of failure;
(b) the softening and unloading part of the curve is infl.uenced by
detailing of the fl.exural reinforcement but not sufficiently to make it
ductile; and
(c) compressive membrane action can have a beneficial effect on
punching strength in tests but in structures the high fl.exural stresses
in adjacent concrete will Iimit this.
 Re!iahility Analysis oj" Deteriorating Bridge Structures 509

'0
l1l I
I
I
I
I
1
/
,11
I
I
I
I
I
I
I
r. ~.,J•.

Model A
l~1
~~~~w·
j,. . llf
t. m
Model B
0 I I
..J I I
I
I
I
I
I-------- A
\
-------
' ------------------
I
I
I

5
Disp. (mm)

FIG. 6. Effccts of rcinforccmcnt slabs on punching shcar rcsponsc.

Testing to determine changes in behaviour in slabs with AAR IS m

progress. 1 6
Overallanalysisshows that the deformation prior to fracture at a column
head is not sufficient to significantly redistribute Ioad to neighbouring
columns. Once the ultimate strength is exceeded the residual strength
capacity is weil below the dcad Ioad component for concrete structures. This
is the classic situation for progressive collapse. The normal ductility inherent
in concrete frame design is no Ionger present. This type of progressive
collapse from punching shear failures has occurred in the USA. One instance
was precipitated by corrosion damage to a slab in a car park.

CORRELATION OF DETERIORATION WITH STRESS STATE

Another aspect of reliability analysis that needs to be considered is the

correlation of deteriorationrate with areas ofhigh stress. In a bridge slab the
reliability ofwaterproofing is sensitive to any cracking in the concrete deck.
Cracking is most likely to occur at the most highly stressed locations and
waterproofing breakdown can then preferentially supply waterandsalt to
aceeierate the AAR or corrosion. The basic reaction of alkali aggregate
reaction continues in dry conditions and then can swell very rapidly with this
subsequent introduction of water. The crosslinking of highly stressed areas
with cracking Ieads to water ingress which then generates further cracking
and creates a substantially more serious reduction in reliability than a
random development of damage.
 Vl

0
-
Slabs C halving joints J

corrosion of main Corrosion of link from

longitudinal st-1 water and salt through
causing loss ot __ defective surfacing or

:-..
0
u ~
~
0
_:;::,..
(;:!
:.:...
u
...,
"0
~
., ;::...
;:
t<.. "'0;:
u
c:
0 1:)
u ;:
:;::,_
n
~
s·
E;

first I ink from

water in structural Typical dia!ijonal
crack structural crack

FIG. 7. Half-joint corrosion.

 Reliability Analysis of Deteriorating Bridge Structures 511

Similar phenomena occur with chlorides where cracking in highly stressed

components permits preferential channelling of chlorides to the most highly
stressed steel. This aspect needs particular consideration in assessment for
reliability. A dassie example ofthis is the half-joint shown in Fig. 7. Here the
structural effects of the corrosion are further aggravated by damp, low
oxygen conditions which favour the development of localised pitting
corrosion. Recent research into corrosion in cracks suggests general rules on
acceptable crack widths may underestimate their effect on the corrosion
rates where chlorides are present.

DETERIORATION RATE MODELLING

The modelling of the reliability of deteriorating structures must introduce

the third element, which is the predictive modelling of the deterioration
phenomena and their rate of development. 17 That brings us round the circle
to new design where the failure to consider durability and deterioration
phenomena explicitly, and to provide a margin on initial strength to offset
for deterioration, has been a major factor contributing to the problems we
face in many structures today. The corrosion rate is strongly influenced by
environment (oxygen supply, humidity, pH, chloride concentration and
temperature). Similar AAR darnage rates can be related to temperature and
maisture conditions for a specific aggregate type. Reliability analysis shows
that the control ofthe environment to Iimit deterioration rates, where this is
possible, is one of the most elfective methods of maintaining safety. This
attention to detailing and maintenance of expansion joints and water-
proofing is essential for safety.

CONCLUSION

The challenge for the coming years must be

(i) to build on our experience of reliability methods for steel structures;
(ii) to carry out testing to evaluate changes in the physical strength
properties and the way in which deterioration phenomena change
them;
(iii) to model the mechanisms of deterioration in terms of chloride ingress
rates, corrosion rates or rates of development of alkali aggregate
reaction; and
(iv) to ensure that a proper quantification of durability phenomena is
introduced into the ctesign process.
512 J. G. M. Wood, R. A. Johnson and C. Ellinas

If one compares the number of concrete structures which are being

demolished because of deterioration and lack of durability and those which
have given rise to problems because they are under strength, it is clear that
the main emphasis of education in the last 20 years has been too preoccupied
with matters of structural analysis and too little attention has been given to
durability.
To answer the client's question 'How long will it safely last?' we must
adopt reliability analysis methods, extend our body of test and case study
data, and hope we can learn faster than the corrosion and alkali aggregate
reaction progresses.

REFERENCES

1. CIRIA Report 63, Rationahmtion of Safety and Serviceability Factars in

Structural Codes. CIRIA, London, 1977.
2. BS 8110, Structural Use of Concrete. British Standards Institution, London,
1985.
3. BS 5400, Steel, Concrete and Composite Bridges. Part 1: General Statement 1988;
Part 2: Loads 1978; Part 3: Steel Bridges 1982; Part 4: Concrete Bridges 1984.
British Standards Institution, London.
4. INSTITUTION OF CrviL ENGINEERS, Steel box girder bridges. In Proc. Int. Conf
ICE, London, 1972.
5. RocKEY, K. C. and EVANS, H. R., The Design of Steel Bridges. Granada
Publishing, London, 1981.
6. SMITH, D., CsENKI, A. and ELLINAS, C. P., U1timate Iimit state of analysis of
unstiffened and stiffened structural components. In Integrity of Offshore
Structures, ed. D. Faulkner et al. Elsevier Applied Science, London, 1987,
pp. 145-65.
7. WALKER, A. C., Study and analysis of 120 cases of structural failures. In Proc. of
Symposium on Structural Failures in Buildings. Institution of Structural
Engineers, London, April 1980.
8. Report of the Royal Commission, Failure of West Gate Bridge, Melbourne,
1971.
9. WOODWARD, R. J. and WILLIAMS, F. W., Collapse ofYnys-Y-Gwas Bridge, West
Glamorgan. In Proc. 1nstn Civ. Engrs, Part 1, 1988, pp. 635-69.
10. THOFT-CHRISTENSEN, P. and BAKER, M. J., Structural Reliahility Theory and
Application. Springer-Verlag, Berlin, 1982.
11. SMITH, D. and ELLINAS, C. P., Rationalisation of tolerances through reliability
analysis. In Rational Fahrication Specificationsfor Steel Structures. Institution
of Mechanical Engineers, London, July 1988.
12. Woon, J. G. M. and WICKENS, P. J., Structural effects of AAR and remedial
actions. Alkalis in Concrete 6th International Conference, DBF, Copenhagen,
1983.
13. WooD, J. G. M., WILSON, J. R. and LEEK, D. S., Physical behaviour of AAR
damaged concrete in structure andin test conditions. In Proc. of8th Int. Conf on
Alkali-Aggregate Reaction, ed. K. Okada et al. Kyoto, Japan, July 1989.
 Reliability Analysis ol Deteriorating Bridge Structures 513

14. INSTITUTION OF STRUCTURAL ENGINEERS, Structural Effects of Alkali-Silica

Reaction-Interim Technical Guidance on Appraisal ofExisting Structures. ISE,
London, 1988.
15. CHRISP, T. M., Wooo, J. G. M. and NüRRIS, P., Towards quantification of
microstructural darnage in AAR deteriorated concrete. In Fracture olConcrete
and Rock, Recent Deve/opments, ed. S. Shah et al. Elsevier Applied Science,
1989.
16. CLARK, L. A. and Nu, K. E., Some factors influencing expansion and strength of
the SERC/BRE Standard ASR concrete mix. In Conf Proc. SERC-RMO, 29
June 1989.
17. Wooo, J. G. M., WILSON, J. R. and LEEK, 0. S., Improved testing for chloride
ingress resistance of concrete and relation of results to calculated behaviour. In
3rd Int. Conf on DeteriorationandRepair olReinlorced Concrete in the Arabian
Gut{; Bahrain Soc. of Engrs/CIRIA, October 1989.
 45
Computer-Aided Sketching of Load paths: An
Approach to the Analysis of Multi-span Arch
Bridges

W. J. HAR VEY and F. W. SMITH

Department of Civil Engineering, The University,
Dundee DD14HN, UK

ABSTRACT

Single-span masonry arches are notionally three times redundant. This

neglects the fact that the abutments arenot rigid, so the degree ofredundancy is
actually greater. In multi-span bridges the problern is compounded. Direct
analytical techniques based on a study of deformation are of limited value in
these circumstances.
Arches naturally release their redundancy by pushing the abutments
outwards. They become three hinge structures. This approach does not allow
closed solutions to the problern but rather indicates Iimits to the engineer's
knowledge and understanding of the structure.
Same details of the approach are described, using Robert Stephenson's
Royal Border Bridge at Berwick-upon-Tweed as an examp!e.

INTRODUCTION

Thrust Lines and Elastic Analyses

In 1676 Robert Hooke 1 found a solution to the problern of the most
important statically indeterminate structure then known. He kept his finding
secret for fear of academic rivalry from those whose stature should have
precluded such plagiarism. Of course we now know that his solution was of
limited value because it provided only understanding, not a scheme of
computation. This paper aims to show that understanding can be more
useful than essentially meaningless calculated results.
In 1879 Castigliano 2 provided engineers with a computational scheme for
indeterminate structures. In various guises his method has held engineers in
515
516 W J. Harvey and F. W Smith

thrall, and incidentally terrified students, ever since. Many of us recognise

that elastic studies may obscure the true behaviour of a structure, but they
do allow us to carry out closed-form calculations on structures which might
otherwise be intractable.
Castigliano's theorems are particularly inappropriate for masonry
structures. Masonry is hardly elastic, and arches in particular are very
susceptible to small movements of their foundations, movements which
must take place.
It is of course possible to use an elastic analysis in an exploratory way.
That is to investigate the effect of movements of the supports and cracking,
and different distributions of stiffness. Unfortunately this is rarely done
because it requires too much effort on the part of the engineer.

Load Path Analysis

Engineers use Ioad path analysis to assist in the conceptual design of
complex structures. The relative stiffness of different paths from point of
application to point of support is assessed on the basis of experience. The
results are checked by more 'rigorous' methods as the design proceeds. The
method proposed here allows the engineer to 'sketch' Ioad paths very
quickly on a computer screen. He receives quantitative feedback as to the
implications of the Ioad distribution he is assuming.

The Zone of Thrust

Hooke's line of thrust is of course a Ioad path. The line of thrust ignores
material strength, but it can be enhanced to show the material which would
be required to support the thrust concerned. 3 In a two-dimensional
structure, that is to say when studying the arch in elevation only, it is
assumed that stress is uniform into the depth of the paper and so only the
width ofthe thrust zone on the paper need be considered. This is obviously a
very rapid process.

Graphical Indeterminacy
Barlow 4 gave a fine graphical representa~ion ofthe indeterminacy ofarch
structures. He demonstrated with a physical model that there are many
potentiallines ofthrust within an arch, any one ofwhich may represent the
actual performance of the structure. The freedom to investigate all the
possible thrust lines in a structure may allow the engineer the flexibility he
needs to understand its behaviour.

THE INDETERMINATE ARCH

Arches are notionally three times statically indeterminate. In modern
structures engineers habitually release these indeterminacies by introducing
 Computer-A ided Sketchin[; ol Load Paths 517

hinges at the springings and the apex of the arch. Pippard and Chitty 5
showed that real arches will approach the three-hinged condition as they
deform after the removal of their centring. Having demonstrated that
masonry arches are not elastic rings they proceeded to use Castigliano to
produce the MEXE analysis beloved of British bridge engineers. More
recently considerable effort has been invested 6 · 7 in developing complex finite
element programs for arch analysis. These still assume that the arch is
supported on rigid foundations, which is manifestly untrue.
A force-based analytical technique is much more readily adapted to the
problern of arches. The flexibility method consists in releasing indeterminate
resultants until the structure becomes statically determinate, then analysing
the effect of unit resultants. The stiffness method is the converse of this. All
'joints' are made rigid in space and the fixity reactions computed. Unit
displacements are considered in order to find the 'true' deformed shape
of the structure. Stress resultants (forces) are computed from the
displacements.
A structure which tends to release its redundant reactions after
construction may be best analysed by the flexibility method. This is
especially true if a path can be opened to the exploration of the effect of
reactions which cannot sensibly be computed. We will endeavour to
demonstrate that such a technique is not just possible but actually easier
than somc others to implement.

THE RELEASED STRUCTURE

Natural Release: The Minimum Thrust View

A masonry arch deflects towards a three-hinge structure on release of the
centring. This must be the case because the arch ring will shorten under stress
and the abutments will deflect outwards. The shorter ring will no Ionger fit
perfectly into the springings. Asthisdeformation takes place, the horizontal
component of thrust in the arch will reduce towards a minimum value. The
minimum corresponds to the maximum possible rise of the zone of thrust
within the arch ring (Fig. 1). Analysis of the three-hinged arch, which is the
Iimit ofthis process, is very simple. It involves writing and solving two linear
simultaneaus moment cquations. This structure may sensibly be regarded as
the releascd form for the application of redundant actions.

Forced Release: The Maximum Thrust Case

It is of course conceivable, and was demonstrated in practice by Chettoe
and Henderson, 8 that the springings may be forced tagether by external
efTects. This might be caused by the approach of a heavy vehicle. The Iimit
518 W J. Harvey and F. W Smith

FIG. I. Arch with minimum thrust.

-\--1- 'f--r· ·...

... -,.,io-- ~...:h -,'
_;/j?,;>·- ·-·~~.
~~~.. y·' \'\.:
·,.\,
·.

8/ \~
>ij
~-
\\\
f~

f/ I
-L...J
iL.1l
""I\_

FIG. 2. Arch with maximum thrust.

case ofthis movement is when the arch begins to Iift in a three-hingedjacking

mechanism (Fig. 2). The result is still a stable structure which can be analysed
as described a bove. In this case the rise of the zone of thrust is the minimum
possible within the arch ring. The horizontal component ofthrust in the arch
is thus maximised.
The two cases described are exactly analogaus to the limiting active and
passive pressures exerted by soil on moving walls.

RESTORING REDUNDANCY

A Favoured Case: Minimum Stress

Having produced a released structure, we must now restore its redundant
actions and reactions to model the real behaviour. Heyman 9 used the
mechanism analysis to compute the line of thrust, which had a minimum
deviation from the curve of the arch. Both Heyman and Harvey 3 chose to
follow the curve of the intrados of the arch, but the arch centreline might
equally weil be used. This approach yields a line or zone of thrust which is
wholly contained within the arch ring, thus implying the fully redundant
structure (Fig. 3). A gross error is made in assuming that this is the correct
solution. Heyman clearly did not fall into this trap. The implication is of
minimum value of maximum compressive stress, which is not the form a
structure naturally seeks.
 Computer-Aided Sketching of Load Paths 519

The Infinity of Real Solutions

The line of thrust defined by the above method lies at a point between the
two extremes already discussed. Barlow 4 showed that these three solutions
are only three of the infinite possible set which exist in an arch which is not
on the point of failure.

The lmportance of Reactions

The reactions which result at the supports of an arch are important
because they control the behaviour of both arch and abutments. In some
Ioad cases a )arger thrust will be needed, in others the interaction willlead to
smaller values. If gross movements are excluded, the arch can deliver
whatever reactions are required within the bounds set above. The two Iimits
are reached with a range of abutment movements of less than 0·1% of the
arch span. 1 0

Exploratory Analysis
The illustrations presented thus far have all been produced using
microcomputer software developed at the Wolfson Bridge ResearchUnitin
Dundee. Wehave a program in commercial use based on the minimum stress
theme. We have recently developed the maximum and minimum thrust
models which were used for the diagrams. The final step has been to allow
the user to alter the location, direction and value of the reactions for the
arch. This encourages the engineer to explore the effect of redundancy by
observing the relationship between reactions and the thrustline. Because the
only computation involved is vector addition, results are presented as a
relatively smooth animation even on an IBM PC AT.

MULTIPLE REDUNDANCY

Most real structures have rather more redundancy than the simple arch
described above. lt is perfectly practicable to use the exploratory technique
on multiple spans.
520 W J. Harvey and F. W Smith

FIG. 4. Double span-both arches falling.

A two-span arch is the next step in complication. Figure 4 shows two

adjacent spans of Stephenson's Royal Border Bridge at Berwick.
Here there are three reactions and six redundancies. The statically
determinate release structure will have both arches descending (Fig. 4).1t will
be necessary to investigate the interaction oftwo resultants and so they must

:\
FIG. 5. Load to lcft-right span propping.

/~~--,~)
[/

FIG. 6. Two spans with reactions adjustcd.

 Computer-Aided Sketching of Load Paths 521

be changed in turn. Each arch is dealt with separately. Reactions are

adjusted to produce an acceptable zone of thrust. The two arches are then
combined and the path of the resultant forcedown the pier is traced. If the
two arches are analysed separately, the combined picture can be displayed
adequately on a Hercules screen, and very attractively on VGA. The figures
presented here are screen grabs from VGA using Word Perfeet 5. Figure 5
shows the effect of a heavy Ioad in the left-hand span of Fig. 4. In Fig. 6 the
reactions in the right-hand arch have been adjusted to produce stability.
Note that the maximum propping action has not been used.

Extension of the process

The graphical quality available on microcomputers is still somewhat
limited. When two spans are being worked on it is usually advantageaus to
treat them individually before combining the graphics on the screen. A
similar approach will of course be possible for a viaduct. Each arch can be
treated as a separate entity and then the structure combined to show the
effect at the piers. As screen resolution improves it will be possible to work
on multiple spans on screen.

THE VALUE OF THE OUTPUT

No doubt many of the readers who have borne with us thus far will wish to
ask the question 'so what?' What does a picture of the polygon of forces
offer the engineer? Our view isthat it allows him to explore the Iimits within
which his structure is actually working. It will not give an absolute solution
to the analytical problem, but the problern is not susceptible to analysis in
that way. So many of the parameters concerned can only be defined in terms
of Iimits. lt is foolhardy to rely on even a few analytical solutions which do
not clearly show what the implications are throughout the structure.

THE NEXT STAGES

This approach has been developed in a !wo-dimensional sense dealing with a

very specific type of structure. The computation is based on a vectorial
representation of the Ioad path within the structure. As such it is definitely
not limited to two dimensions. We hope to find the resources to develop the
approach, first for complex three-dimensional masonry structures and then
to structures where bending plays an important part in performance. The
complexity in this latter stage is in representing the Ioad path. The
traditional approach of bending moment and shear force diagrams in
522 W J. Harvey and F. W Smith

different orthogonal axes is quite inappropriate and a major part of the

development effort will be in finding more appropriate representations.

CONCLUSIONS

1. Slender gravity structures are fundamentally different from modern

continuous structures and demand unique analytical tools.
2. The line ofthrust view has long been popular and Ioad path analysis is its
modern descendent.
3. The flexibility method provides a basis for a semi-automatic form ofload
path analysis for masonry structures.
4. The method is capable of considerable development, especially as
computers become faster and as display techniques improve.

REFERENCES

1. HooKE, R., A Description o.f Helioscopes, and some other Instruments. London,
1676.
2. CASTIGLIANO, C. A. P., Theorie de l'Equilibre des Systemes Elastiques et ses
Applications. Augustos Frederico Negro, Turin, 1879. Translated by E. S.
Andrews, Elastic Stresses in Structures. Scott, Greenwood & Son, London,
1919. Also with an introduction by G. E. A. Oravas, The Theory of Equilibrium
of Elastic Systemsand its Applications. Dover, New York, 1966.
3. HARVEY, W. J., The application of the mechanism analysis to arch bridges. J.
Inst. Struct. Engng, 66(5) (March 1988).
4. BARLOW, W. H., On the existence (practically) of the line of equal horizontal
thrust in arches and modes of determining it by geometric construction. Min.
Proc. Inst. Civ. Engng, 5 (1846) 162.
5. PIPPARD, A. J. S. and CHITTY, L., A study ofthe voussoir arch. National Building
Studies Research Paper 11, HMSO, London, 1951.
6. CHRISFIELD, M. A., A finite element computer program for the analysis of
masonry arches. Transport and Road Res. Lab. Report LR 1115, Crowthorne,
Berks, UK, 1984.
7. CHRISFIELD, M. A., Computer methods for the analysis ofmasonry arches. Proc.
2nd Int. Conl on Civil and Structural Engineering Computing, Vol. 2. Civii-
Comp. Press, Edinburgh, 1985, pp. 213-20.
8. CHETTOE, C. S. and HENDERSON, W. H., Masonry arch bridges: a study. Proc.
Inst. Civ. Engng, 8 (1957) 723-55.
9. HEYMAN, J., The estimation of the strength of masonry arches. Proc. Inst. Civ.
Engng, 69 (Dec. 1980) 921-37.
10. HARVEY, W. J. and BARTHEL, R., The relationship between thrust and springing
movement in arches (to be published).
 46
The Assessment of Masonry Arch Bridges-The
Effects of Defects

CLIVE MELBOURNE
Department of Civil Engineering and Building, Bolton Institute of
Higher Education, Bolton, Lancashire, UK

ABSTRACT

A !arge number of the existing stock of masonry arch bridges in the UK suffer
from a variety of defects. The paper considers two of these defects-ring
separation and spandrei wall separation. Ringseparation in multi-ring arches
is the loss ofbond between successive rings. Aseries oflaboratory tests on I, 3
and 6 m span arch bridges is described. The significance of soil/structure
interaction is discussed and the importance of defects quantified. The relative
importance of the various parameters influencing the mode of behaviour and
strength of the arch is discussed, thus allowing the engineer to formulate a
mathematical model of the arch bridge to be assessed.

INTRODUCTION

Presently, in Britain, a programme of research is being undertaken to study

the behaviour ofmasonry arch bridges under Ioad. The programme is being
funded primarily by SERC, British Rail and the Department of Transport.
The aim of the research is to improve the methods of assessing the load-
carrying capacity of this type of bridge as present methods appear overly
conservative. 1•2
The Bolton Institute Arch Research Team has been studying the
behaviour of masonry arch bridges for some years as part of the national
research programme. The work has encompassed small- and large-scale
Iabaratory tests (up to 6 m span) as weil as field tests.
Much of the earlier work related to the importance of the soil structure
interaction and spandrei wall stiffening. However, an integral part of any
523
524 Clive Melbourne

assessment is an appraisal ofthe significance of defects which exist within the

structure. Recent work at the Bolton Institute has concentrated on the
significance ofthese defects on the load-carrying capacity. Initially two types
of defect have been considered, ring separation in multi-ring brick arches
and spandrei wall separation. Both types of defects have been identified by
the industry as being commonplace and significant although, to date,
engineers have only been able to take cognisance of them in the form of a
subjective condition factor.
Ring separation occurs in multi-ring brick arches and is associated with
the loss ofbond between successive rings caused by weathering and/or stress
cycling of the mortar.

MODEL TESTS

The models were of a parabolic profile with a span of 1000 mm, span/rise
ratio of 3:1 and a width of 500 mm. The arch ring comprised two rings or
courses ofbrickwork, either bonded or unbonded, around the full arc ofthe
arch. In total the ring thickness was approximately llOmm. The models
were constructed using half-scale fletton bricks with an average compressive
strength of 30 Njmm 2 . They were constructed without spandrei walls.
Because there were no spandrei walls the fill was contained by the perspex
and wooden sides ofthe test rig. Prior to laying the bed face, each brick was
coated with oil in order to minimise the effects of the bond strength on the
model behaviour. A 1:1:6 (cement:lime:sand) mortar was used, it was mixed
by volume and achieved an average 28-day compressive strength of
4 Njmm 2 • Sand was used as backfill and compacted by vibration.
A knife edge Ioad (KEL) was applied incrementally and monotonically up
to failure at either the 1/4 or crown point.
Table 1 presents the results of the tests. Arches 1-3 were built suchthat

TABLE 1
Model test results

Ar eh Inter ring Loading Experimental

number bed-joint position ultimate Ioad
material (kN)

1 Mortar 1/4 13·2

2 Mortar 1/4 18·6
3 Mortar Crown 31·0
4 Sand 1/4 6·9
5 Sand 1/4 8·1
6 Sand Crown 8·8
 The Assessrnent ol Masonry Arch Bridges 525

the two rings of brickwork were fully bonded together using a mortar. (By
bonding it is meant adhesion rather than brick bonding using 'headers'.)
Arches 4-6 were built with the mortar between the rings, being replaced
by damp sand to simulate loss of adhesion.
All of the models which were loaded at the 1/4 span failed due to the
formation of a four-hinge mechanism. For those models loaded at the
crown, failure was due to the development of a 'classical' five-hinge
mechanism. At failure in the 'bonded' arches no or little ring separation
occurred and formation ofthe hinges was as a monolithic 'single' ring. In the
'unbonded' arches, other than in the region of the applied Ioad, the
unbonded rings were measured by embedment strain gauges as separating
with increasing Ioad. Development of the hinges in the models with initial
ringseparationwas suchthat two thrust lines developed, one in each of the
inner and outer rings. Hinges in the two rings were coincident and developed
simultaneously. Collapse was sudden and caused by almost total physical
separation of the two rings and full development of the hinges.

THREE-METRE SPAN BRIDGE TESTS

Two 3-rn span bridges were built and loaded to failure in the institute's large-
scale testing laboratory. The segmental arch barre! (radius 1875 mm)
consisted of two rings of brickwork using class A engineering bricks. The
brickwork was built in a 'stretcher' bond with no bonding between the rings
other than through the mortar in the 'bonded' case and damp sand in the
'ring separation' case. The spandrel, wing and retaining walls were built in
English bond using concrete commons. The walls were not attached to the
arch ring. An average gap of 10 mm was provided between the spandrei walls
and the arch ring. This ensured that both the effects of ring separation and
spandrei wall separation could be studied.
Campaction of the 'graded 50 mm' Iimestone backfill was achieved using
lüümm layers and a vibrating compacting 'wacker' plate. The bridgewas
filled to 300 mm above the crown.
Both bridges were subjected to three loading conditions. Firstly, a 25·7 kN
KEL was applied at 250 mm centres across the span to simulate a rolling
Ioad. Secondly, a 50 kN KEL was incrementally applied at the north quarter
point, crown and south third point. Finally, a KEL was applied
incrementally at the quarter point through to collapse. The elastic tests
confirmed that the structure responded to the loading as a local effect, with
soil pressure and brickwork strain changes being confined to the vicinity of
the loading. This has been observed in field studies. 3
Both arches failed by the formation of four-hinge mechanisms. In each
526 Clive Melbourne

FJG. 1. Three-metre span arch barrel- ring separation.

case the spandrei walls cracked in the vicinity of the crown and rotated
about the abutment remote from the KEL (Fig. 1). The sequence of hinge
formation is given in Table 2.
In the bonded arch some ring separation occurred at the crown but the
hinges formed at intrados and extrados. On the other hand, the unbonded
arch produced extensive ring separation shortly after the formation of the
second hinge. lt is significant to note that the first hinge formed at
approximately the same Ioad in each test. As no ring separation cracking
had occurred at this stage, it confirmed that the two arches were comparable.
The unbonded arch deteriorated more rapidly after the formation ofthe first
hinge and carried an ultimate Ioad of 360 kN- a 33% reduction in carrying
capacity compared with the bonded arch.

TABLE 2
Sequence ol hinge formation
- - · ·- --- - - - - - - - - -- -- - - - - -- - -- - ----

Hinge Position Load (kN)

number
Bonded arch Unbonded arch

I Under Ioad point, north quarter point 240 220

2 South quarter point 300 240
3 South abutment 400 320
4 North abutment 480 320
(540 max.) (360 max.)
 The Assessment of Masonry Arch Bridges 527

As with the model tests, once ring separation occurred each ring formed
its own pattern of hinges which interacted with each other.
An assessment of the bridge using the presently accepted 'MEXE'
method 1 •2 gave a modified axle Ioad of200 kN. This represents a Ioad which
is 85% ofthat required to cause the formation of the first hinge.
The soiljstructure interactionwas monitored using pressure cells, not only
in the extrados but also in the spandrei walls and the backfill. The overall
picture which emerged was one of a compacted backfill exerting pressure
greater than earth pressures at rest and which restrained the arch and
dispersed the applied KEL. Additionally, there was a frictionaljcohesive
resistance between the backfill and the spandrei walls and the extrados,
which restrained the arch initially and which increased as the arch swayed
into the backfill as hinges formed. This movement into the backfill not only
increased the longitudinal horizontal soil pressure but also the deviatoric
stress, as on the back of the spandrei walls and hence the resistance to
movement.
A mechanism analysis using methods developed by Pippard and Baker 4
and Heyman, 5 and modified to incorporate the above factors, was used to
predict an ultimate load-carrying capacity of 470 kN.

SIX-METRE SPAN BRIDGE TEST

The bridgewas built and loaded to failure at the institute's large-scale testing
laboratory and represented one of a series of full-scale tests coordinated by
the TRRL. I t had a span of 6 m with a rise of 1 m and an overall width of 6 m.
The two-ring brickwork arch was constructed using solid engineering class
concrete bricks. The brickwork was built in a stretcher bond with no
bonding between the rings other than the mortar bed joint. The total
thickness of the arch barre! was 220 mm. This represented quite a slender
arch with a span:ring thickness ratio of27. The spandrei walls were built in
an English bond. A 50 mm graded Iimestone was used for the backfill
material compacted in 100 mm layers to 200 mm above the crown; this was
surfaced with a 100 mm thick bituminous layer. The overalllength of the
bridge was 14m.
Throughout the construction process and subsequent testing instrumen-
tation was provided to monitor all aspects of the bridge behaviour. Two
separate series ofload tests were conducted on the bridge. Full details ofthe
tests are given elsewhere. 6
The first comprised the application of a point Ioad at various positions
across the width and span of the bridge; these tests were carried out to
simulate a wheel loading.
528 Clive Melbourne

Initially it had been hoped to apply a point load of 100 kN; however,
during the initialloading cycle adjacent to the spandrei wall at the quarter
span, cracking was recorded at 70 kN. Loading caused separation of the
spandrei wall and arch ring, and the cracking manifested itself inside the
arch barrelas ring separation along the line ofthe quarter point, detected by
strain gauges embedded in the brickwork. Further point loading cycles at
different locations on the bridge caused the initial crack to open and close as
the load increased and decreased respectively.
Following the point loading tests, the bridgewas subjected to a KEL at
the quarter span; loading was applied monotonically through to collapse.
Separation of the spandreljbarrel interface commenced on the east side at
360 kN. Further loading caused the cracking to spread around the arch
barrel. At 400 kN the first hinge in the barrel was observed underneath the
load point. At 640 kN cracking of the spandrei walls was observed.
Failure ofthe bridgewas due to the formation ofa four-hinge mechanism
at a total applied load of 1173 kN. The spandrei walls were lifted and
rotated, as shown in Fig. 2. Extensive ring separation was coincidental with
failure and the arch barreljspandrel wall was almost completely separated
around the full arc of the barrel. It is worth noting that the mode of failure
was similar to that of model tests previously reported by the author and
others. 7
Typical graphs of deftection, soil pressure and brickwork strain are given
in Fig. 3.
Formation of the first hinge under the KEL at 400 kN can be clearly
deduced from the loadjdeftection graph, which shows a marked change in
stiffness from the initial elastic behaviour. At the quarter point, remote from
the loading, little movement took place until 800 kN, at which stage there
was a rapid increase in deftection corresponding to extensive cracking ofthe
arch and further hinge formation.
Surface-mounted vibrating wire strain gauges confirmed the stiffening

FIG. 2. Crack pattern at failure.

 The Assessment of Masonry Arch Bridges 529

effect ofthe spandrei wall, which confined the thrust line to within the middle
third at the three-quarter span.
Ring separation was monitared and propagation of the local ring
separation (caused during the point Ioad test) commenced at 400 kN.
Soil pressures were monitared not only on the extradas of the arch but
also on the spandrei walls. Upon completion of construction the soil
pressures were comparable with earth pressure at rest/unloading (i.e. K0 yh or
Kryh, Kr= 1/ K0 ). The oil pressures under the KEL increased as the Ioad was

Thrust
line
1200 position
1200 1 00
_1000
~ 800
"0 600
0
0 400
..J
200

____ _
/
/
/
/.:.. /

FIG. 3. Typical graphs.

applied and were compatible with a dispersal angle of 45'. There was a
significant increase in the lateral pressure against the spandrei wall in the
vicinity of the loading beam as the Ioad increased. This was a contributory
factor to the cracking in the wall. More significantly, there was little change
in the soil pressure elsewhere in the bridge until the Ioad exceeded 600 kN, at
which stage there was a general increase in pressure on the side remote from
the Ioad caused by the arch barre! trying to 'sway' into the fill. It is important
to note that even at failure the soil pressures did not reach passive pressures
but did provide significant restraint to the arch.
Using a modified mechanism analysis adopting Kr pressures, tagether
with the stiffening effects of the spandrei wall and cohesive/frictional
resistance, an ultimate Ioad capacity of 1010 kN was predicted and a Ioad of
730 kN for the onsct of mechanism bchaviour.
530 Clive Melbourne

CONCLUSIONS

(1) If free to do so, an arch bridgewill fail due to the formation of a four-
hinge mechanism.
(2) Using a modified mechanism analysis incorporating the lateral backfill
pressures, spandrei wall stiffening and backfill cohesion/friction
structural interaction, the onset of mechanism behaviour and the
collapse load can be predicted. The prediction ofthe onset ofmechanism
behaviour could be used to set a Serviceability Iimit state.
(3) At all stages ofloading consideration must be given to the possibility of
local failure (e.g. punching shear, ring separation, snap-through) and
adequate factors of safety applied.
(4) Ring separation caused a reduction of between 56% and 33% in the
ultimate load-carrying capacity of the model and full-scale two-ring
brick arch bridges respectively.
(5) Passive soil pressures were not observed in any ofthe tests, even at gross
deformation.
(6) Surface strain measurements indicated that the spandrei wall stiffening
delayed hinge formation and infl.uenced hinge positions.
(7) Where spandrei wall separation exists the cohesion/friction resistance
on the back ofthe wall makes a significant contribution to stiffening the
arch.

ACKNOWLEDGEMENTS

The author wishes to acknowledge the financial support of SERC, British

Rail, TRRL and NAB, also the encouragement and support from the staff of
the Department ofCivil Engineering and Building at the Bolton Institute of
Higher Education.

REFERENCES

1. The Assessment of Highway Bridges and Structures. Department of Transport

Roads and Local Transport Directorate, Departmental Standard BD 21/84,
Department of Transport, March 1984.
2. The Assessment of Highway Bridges and Structures. Department of Transport
Roads and Local Transport Directorate, Advice Note BA 16/84, Department of
Transport. HMSO, London, March 1984.
3. MELBOURNE, C., The construction of mass concrete arch bridges. Structural
FaultsandRepair Conference, London, June 1989, Engineering Technics, 1989.
4. PIPPARD, A. J. S. and BAKER, J. F., The Analysis of Engineering Structures, 2nd
edn. London, 1943.
 The Assessment of Masonry Arch Bridges 531

5. HEYMAN, J., The Masonry Arch. Ellis Horwood Ltd, London, 1982.
6. MELBOURNE, C. and WALKER, P. J., Loadtest to collapse of a full-scale brickwork
masonry arch. TRRL Contractor's Report (tobe published).
7. MELBOURNE, C., QAZZAZ, A. and WALKER, P. J., Load testing to collapse ofmodel
and full-scale brickwork masonry arches. SERC, Repair Maintenance and
Operation in Civil Engineering Conference, London, June 1989.
 47

Theoretical and Experimental Investigations on

Railway Bridges Dating from 1856 to 1895

F. MANG and Ö. BUCAK

Versuchsanstalt für Stahl, Holz und Steine
(Testing Centre for Steel, Wood and Stone),
Universität Karlsruhe, Kaiserstrasse 12, D-7500 Karlsruhe, FRG

ABSTRACT

When assessing the residual service life of old bridge structures, the question
arises whether they can be still used after having exceeded the theoretical
service life, or if they should be replaced by a new structure after having been
newly class(fied, or il they arestill usable afler reinforcement or an the basis of
shorter service intervals.
The determination of the residual service l(fe of a bridge results in the
difficulty ofrecording the exact history ofthe Ioads and the state ofthe bridge.
Insufficient knmt·!edge about the strength andfatigue behaviour of old steels
and constructions applied afu;ravate the above decision.
This paper dea/s with experimental investigations an a complete bridge
structure ofthe 'museum rai/way · ofthe community ofB!umberg as weil as with
two developed bridge structures of the Federal Railways. The results of these
component tests are compared with data knownfrom literat~;re as weil as with
other results of similar investigations performed at the institute.

INTRODUCTION

Today structures subjected to fatigue loads, for example railway bridges, are
usually designed for a certain service life. The reason for this is the
knowledge that after exceeding a critical number of load cycles with a
sufficiently high loading level fatigue fractures occur. This calculated failure
is covered with a corresponding 'safety factor'.
For older iron and steel structures, design 'for a set time' was not usual.
533
FIG. l. General view of KoblenzjWald shut railway bridge.

FIG. 2. Detail of the Koblenz/ Waldshut railway bridgc.

 lnvestigations on Railway Bridges Dating from 1856 to 1895 535

There are numerous structures which have been subjected to fatigue loading
for more than a century.
After reaching and exceeding the 'standard service life', fatigue-loaded
structures often cannot be reliably assessed with regard to possible residual
life. One reason for this is insufficient knowledge about the behaviour ofthe
static and fatigue strength of steels (wrought iron, puddled steel) and the
structures used in the 19th century. The o\dest railway bridge, the Koblenz/
Waidshut between the Federal Republic of Germany and Switzerland,
which is in the charge of the Versuchsansta lt für Stahl, Holz und Steine,
dates from 1858 and is still in operation today. This bridge is shown in Figs 1
and 2.

FIG. 1 Local rust formation.

536 F. Mang and Ö. Bucak

Extensive measurements under traffic Ioad and investigations on the

material allow the operationoftbis bridge to continue-at least for the next
few years. Presently the gap in single elements due to rust formation (Fig. 3)
seems to be a problem. After complete removal of the rust these areas are
filled with a flexible lute (Fig. 4). The original values of the forces caused by
the rust formation are being investigated and will be reported on later.
Since the design and strength of joints substantially influence the Ioad-
hearing behaviour of the total structure, such joints from old steel bridges
have tobe investigated more precisely. When assessing old existing bridges,
material specimens have been taken from the braces or from less stressed

F1G. 4. Flexible lute.

 Investigations on Railway Bridges Dating from 1856 to 1895 537

areas of the construction for examination. The data obtained were and still
are subjected to very high safety factors since a safe assessment is required.
The aim of recent investigations in Karlsruhe is to obtain real residual
service life by studies on original joints and structural members or full
structures, and comparison with data on simple specimens already tested in
the institute or which are available from literature. Some characteristics may
have to be newly determined.

GENERAL

Knowledge of the residual service life of old bridge structures is required by

their operators, or by people who need to know how much Ionger these
structures can be operated under present and projected loading conditions.
The protection of historical monuments and economic considerations are
the decisive factors.
One recent project was the so-called gun railway, also called 'pig-tail
railway' (museum railway), which runs in the area of the Wutach Valley in
the south of Germany. Figure 5 shows one of the bridge structures of the
museum railway.
Until recently the question of residual service life has been tackled by
gathering the following information:
(1) Survey of the bridge by visual inspection (state of corrosion). This
was the determining factor for some of the bridges.
(2) Ifstatic calculations are available, determination ofthe peak stressed
areas.
(3) Sampling of less critical but representative areas of the structure.
(4) Test investigations, for example chemical analysis, tensile tests and
fatigue tests on original material.
(5) Additional calculations based on actual material properties.
(6) Determination of the previous loading on the bridge and expected
future loading (Ioad spectrum).
(7) Application of the Miner rule (linear darnage accumulation
hypothesis) for the determination ofthe residual service life as weil as
the safety factor for this particular bridge.
(8) Fracture mechanics investigations, for example COD tests.
Sufficient data for material properties have been established by
preliminary surveying of various highly dynamically stressed structures
which can be used for initial assessment. Significant investigations need only
be done in the final phase for the purpose of confirming the assumptions
made. Presently there is still a gap in knowledge about the behaviour ofsuch
538 F Mang and Ö. Bucak

FIG. 5. A bridge structure of the museum railway.

bridges as a total system or on the bearing capacity of complete structural

members in the original state. Questions about the rivet slip, the rivet initial
Ioad, the displacement under shear stress and Ioad distribution with various
elements of a structural member are still unknown. These parameters
influence the fatigue behaviour. For this reason checks on full-scale
structural members typical of an actual bridge are needed.
For this purpose a complete bridge with a length of 4·50 m dating from
1877 has been selected from various bridges of the museum railway together
with the longitudinal girders and cross-girder joints of a further bridge
system dating from 1895 with a span of 4·8 m which have been investigated
in the Versuchanstalt für Stahl, Holz und Steine.

INVESTIGATIONS ON THE MATERIAL

Sampies have been taken from five specimens ofthree bridges ofthe museum
railway and three specimens have also been taken from the Stahringen
Bridge.
With the exception of one specimen, which was taken from a base plate
made of cast iron, investigations were carried out on semi-finished products
made from puddled steel. The purpose of the investigations on these
specimens was to find out whether the intensive investigations made on older
materials from other bridges of the same period is possible.
 Investigations on Railway Bridges Dating Jrom 1856 to 1895 539

The results ofthe tensile tests carried out on specimens from various parts
of the bridge confirm the material data obtained from other similar
structures.

PERFORMANCE OF FATIGUE TESTS

Tests on the Museum Railway Bridge

The bridge (Fig. 6) was installed in the 50 MN press of the institute. The
photograph of Fig. 7 shows the installation corresponding to the critical
Ioad configuration LF 1.
Stress results of static tests under both Ioad configuration 1 and Ioad
configuration 2 compared weil with calculated values (O" = 135 N/mm 2
calculated, 137 N/mm 2 measured).
Calculations carried out with the actual bridge dimensions before testing
indicated that LF 1 was the critical Ioad configuration.
Forthis reason a fatigue testwas performed with this Ioad configuration.

r-------------4885 -----------1
105

=
N
~
N

72 riv~ts

LF 1

LF 2

FIG. 6. Test spccimcn and loading arrangcments.

540 F. Mang and Ö. Bucak

FIG. 7. Bridge in the 50 MN testing machine.

The test Ioad was twice the maximum Ioad expected with a limiting stress
ratio R = + 0-35. Under this Ioad an edge stress of (J = 137 N/mm 2 existed on
the main girder. The stresses have been calculated with the Ioads applied and
the net iron section calculated according to the normal assumptions of static
behaviour.
Flexible bearings were used in order to allow for possible irregularities of
the bridge. To determine the stress state on the main girder important areas
were monitared by strain gauges. With a Ioad cycle of 108 070 the testwas
terminated since a correct Ioad application was not possible because of the
occurrence of a !arge crack. The cross-girder/ main girder jointwas where the
failure occurred. This is shown in Fig. 8.
The bridge was cut into small pieces in order to investigate individually
the main girders, longitudinal girders and the cross-girderjmain girder joints
(Fig. 9).

Tests on both Main Girders and Longitudinal Girders

The tests on the main girders with a span of 3300 mm were performedas a
four-point bending test. The distance between points ofload application was
 Investigations on Railway Bridges Dating from 1856 to 1895 541

Fru. 8. Cross-girder/ main girder joint---Hacked web plate.

800 mm. With this arrangement a constant bending stress occurred in the
central area of the main girder (Fig. 10).
The Ioad for the first main girder was selected to induce a tensile stress of
(J = 137 N/ mm 2 and later 165 N/ mm 2 at the extreme fibre. The limiting stress

ratiowas R=0·2. After 10000000 Ioad cycles with an edge stress of 135

~!:CD moin girder

:z;:c: :x.::x
I+++ +I
::::S.::L!

T -j- 1-t + -j- + -t +J

+ -t + + + ""t" -;rr; .... + + +++J

II n
I I (]) second girder
I
(!) ][
ti;' tro nsverse
'-V
t ro nsverse Q)
girder girder

I I
JI
ZL
TI
®
z .=z
second girder
li
001.x
I I
1-t-++-+- + + I 1++++++1
++++- + + I+- -j- +++-+I
t t! Q) main girder

Fru. 9. Blumbcrg Bridgc- classification of thc lest spccimcns aftcr dismantling.

542 F. Mang and Ö. Bucak

t
800 ~m

t t
3300 mm

FIG. 10. Thc ccntral arca of thc main girdcr.

N/mm 2 and without any discernible cracking, the Ioad was increased to rJ =
165N/mm 2 •
Under this load a crack in the web ofthe main girder occurred with a load
cycle of 1 534 000. With this the load could be sustained despite the crack in
the main girder. Forthis reason the testwas continued until both angles and
the chord plate of the main girder's tension flange were also broken. This
occurred at a total load cycle of 1572 600. The girder therefore took an
additional 38 600 load cycles from the clearly visible incipient crack of the
web up to the total failure of the main girder (Fig. 11).
The web was first to break followed by the angle section and last the chord
plate. The cracks were random and not on a line.

FIG. II. Rupture portions in main girder I aftcr the tcst.

 SR [N/mm 2)
300 _tl
c=r=========~~. ?
~
.",
7o<s<ao I I -._ ~~d I IJ
e { 21<d<31 I lso-J _ j ~·
10<t <20 -l50 ~­
200 bridge Blumberg §
faltgue tesls on lhe full bndge ® ""
~
fahgue Ies ls on the matn gtrder • :::;
150 faltgue lests on rematnmg ptece 0 :::>;:,
of lhe mam gtrders
faltgue tests on parts of lhe <> ~
langtludmal gtrder ·~
~
....,
Ty~e
of fatlure:
100 1--------t (onnec lton of cross gtrder -1----t----11--+-+-----"...._:---+ !!ttto.l----+-+-"""'.,+---1-- ~
to matn gtrder ""'
8 ""
b
--I I I I I I ~
- u ... ,c ..
~·
70~ +--
~
0
+----- · ~ cross gtrder
60 2i
Oo
OoL-.R V,
--
50 Ou~ ,_____ _.d!L ·•·H·•,' ~~·ot•• t--t---t---t-----i-l~-++--1 0\
R=•0.1 \.IIOaao • ., N 0
·~~ ~
40~~==~~=r~~====~~~~====~--_j~_j~_j_j_l~--~ Oo
\Q
--
10 4 2 4 6 8 10s 2 4 6 8 10 6 2x10 6 4 6 8 107 2x107 4 6 8 10 8 2x'to8 V,

Ftu. 12. Comparison of rcsults ofthc tcsts on structural mcmbcrs ofthe Blumbcrg Bridge with form er invcstigations on membcrs with drillcd holcs.

u.
.j:>.
w
544 F Mang and Ö. Bucak

With the second main girder, approximately 100000 cycles were needed
from the clearly visible incipient crack in the web up to total failure.
Three further test specimens were taken from the unbroken area of the
main girder and tested in the three-point bending test with a span of
1500 mm. The upper loads were 450 or 600 kN, and the corresponding
limiting stress ratio was R = 0·35 or + 0·1.
F our test specimens were taken from the longitudinal girders and tested in
a three-point bendingtest with a span of 1500mm. With two specimens the
upper stresswas 165 N/mm 2 and with the others 135 N/mm 2 . The tests were
performed with a limiting stress ratio of R = + 0·1.
As with the tests on the main girder specimens, the cracks start from the
bearing point and run diagonally through the web. Afterwards cracks
occurred at the angle sections of some specimens.

TABLE 1
Results of the tests on dead-end transverse girder Joints
p
2

--------~·-

Test State Po Pu Load cycles Remarks

number (kN) (kN)
-·------- --------- ~~-
-- ---~~

3 Dead-end 450·0 157·5 2 560400 Crack in the web

transverse between main
girder girder 2 and
longitudinal girder 7
4 Dead-end 300·0 105·0 II 479 500 No crack
transverse 450·0 157·5 3 366400 Diagonal crack
girder on the web plate
between I and 5
8 Non-broken 450·0 157·5 27 237 400 Crack in the web
joints screwed between main
tagether girder 2 and
longitudinal girder 7
7 Change of Ioad 450·0 157·5 25 375 400 No crack
arrangement so
that broken area
is not stressed
------- ~---------
 lnvestigations on Railway Bridges Datingfrom 1856 to 1895 545

The results ofthese investigations were compared with older test data on
puddle steel specimens with drilled holes existing in the institute. lt is evident
that the results from tests on the main girder and longitudinal girder are
within the scatter range of the tests on small specimens (see Fig. 12).
Since the failure of the cross-girder joints occurred under test on the full
bridge ofthe museuro railway and investigations on suchjoints did not exist
or rather were unknown from literature, emphasis has been placed on this
type of failure in these investigations.
At first both ends of the bridge were tested with a limiting stress ratio of
+ 0·35 (as with the test on the complete bridge) with various shearing forces,
both as the three-point bending test and as the four-point bending test. The
results with the corresponding test data are given in Table 1.
Afterwards the ends of the dead-end transverse girders were screwed
tagether to a new specimen to enable joints with !arger dimensions to be
tested. Figure 13 shows the test specimen after failure.
From the second bridge a girder was subjected to bending with four cross-
girder joints, and to a fatigue test. The limiting stress ratio was + 0·1. The
results are given in Table 2.

Tests on Wind Bracing Parts of the Bridge of the Museum Railway as

weil as on Small Specimens taken from the Bridge Structures
From both wind bracings ofthe bridge the central areas were kept in the

FIG. I 3. Crackcd cross-girder joint.

 Vl
~
0\

S.. [N/mm2 l
500 I r:
I I I I I I I_) &
e'
u
_,_ ·~
400 (1001 ""e .<:
::t
I! iiS a
"""- r0<8<80 I
• 20<d<31 ...• •
."
·;:
-1511
~~ 10<t<20 I -tl0-r9 .a
f.a
300 """ '- newly drilled holes ... ... :->'l
originil hol• with rivet 0 e
250 drilled •embers tilken II ~
........, tr~ th• lli!in gird•rs
• ;::
~ ~
"'"' . '
~~ drill•d •••bers tilhn )( $::)
~ tro111 the secondwy girders ;::
200 ...... i::l...
0
' ' ' '' I"' ..
~~ ~ ~:
156,1 N/mm2 l::l::l
150 f-
" r-:; ~~ ~-
.......,.
~
~I t
::::
$::)
"';>;-
124,C N/mm 2
~ ~ ·~ ~
Ou
~~ ~~
100 - R=+0,1 98.1 N/mm2
fl
"' J
s~." 2 4 6 8Ht 2 4 6 81o' 2 4 6 81o' 2•1o' 4 6 8107
FlG. 14. Results of investigations on webs with drilled hol es and on members with original rivets.
 lnuestigations on Railway Bridges Dating from 1856 to 1895 547

TABLE 2
---· - -- - - -- --- -- --- ------ --- - - - - - - - - --
Test State Po Pu Load cycles Remarks
number (kN) (kN)
-· ---- - - - - - - - -·- - - - - - - - - - - - - - - - - - - - - -
Connection 350·0 35·0 15 083 712 No crack
Cross-girder to 600·0 60·0 73440 Crack under the
main girder Ioad application

original state (with rivets) and subjected to a fatigue test with a pulsating
machine. Two tests with a nominal upper stress of 250 and 200 N j mm 2 were
performed. The results are plotted in Fig. 14 and compared with the results
on punched webs of previous bridge investigations existing at the institute. lt
can be seen that the results are on the favourable side ofthe scatter range. At
the sametime it should be pointed out that the scatter is relatively high, as is
usual with old structures.
Residual material from the bracings with newly drilled holes of 20 mm
diameter was removed and subjected to fatigue tests under various stress
Ievels. From the results plotted in Fig. 14 good correlation with old test data
is evident. Several small specimens from the L-sections and chord plates of
the main girders were taken and tested in fatigue. They also show good
correlation with previous test values. With all specimens the incipient fatigue
crack started from the rivet hole, as expected. Figure 15 shows the fracture
surface of test specimens after failure.

FIG. 15. Rupturc surface of a solid rod--polishcd section on a rivet.

548 F Mang and Ö. Bucak

REFERENCES
1. MANG, F., Stahl im Altbau und Wohnungsbau. Abschlußbericht zur Forschungs-
studie des Landes Nordrhein-Wes}falen, VBI-72.02-92/77.
2. MANG, F., STEIDL, G. and BUCAK, Ö., Altstahl im Bauwesen. Schweißen und
Schneiden, 1 (1985) 1-5.
3. BRÜHWILER, E. and HIRT, M. A., Das Ermüdungsverhalten genieteter
Brückenbauteile. Der Stahlbau, 1 (1987) 1-8.
4. HERZOG, M., Erwiderung zur Zuschrift von Tschumi, M. auf Herzog, M.,
Abschätzung der Restlebensdauer älterer genieteter Eisenbahnbrücken. Der
Stahlbau, 5 ( 1986) 159-60.
5. BAEHRE, R. and KosTEAS, D., Einfluß der Vorbelastung auf die Restnutzungs-
dauer schweißeiserner Brücken. Bericht Nr. 7496 der Versuchsanstalt für Stahl,
Holz und Steine der Universität Karlsruhe, January 1979 (unpublished).
6. STEINHARDT, 0., Festigkeitsverhalten von Schweißeisen aus Brückenbauwerken
des 19. Jahrhunderts. ETR 26, 6 (1977) 383-7.
7. N.N.: Unveröffentlichte Untersuchungen der Versuchsanstalt für Stahl, Holz
und Steine der Universität Karlsruhe.
8. STIER, W., KosTEAS, D. and GRAF, U., Ermüdungsverhalten von Brücken aus
Schweißeisen. Der Stahlbau, 5 (1983) 136-42.
9. WENZEL, F., Erhalten historisch bedeutsamer Bauwerke. Jahrbuch 1987, SFB
315, und der Universität Karlsruhe (TH), Ernst & Sohn, 1988.
 48
Structural and Material Darnage to CoQ,crete
Highway Bridge Decks in Saudi Arabia

M. Y. AL-MANDIL, A. K. AZAD, M. H. BALUCH,

A. M. SHARIF and D. PEARSON-KIRK
Department of Civil Engineering, King Fahd University of
Petroleum and Minerals, Dhahran, Saudi Arabia

ABSTRACT

A seemingly !arge number of concrete highway bridges in Saudi Arabia are

suffering signs of cracking, deterioration and, in some cases, localizedfailures
to their deck slabs. Several reasons are believed to haue contributed to these
phenomena, including lmr quality materials, poor construction practices, Iack
ol control over vehicular Ioads and the severe environmental factors.
A research programme has been initiated to categorize the types of darnage
to bridge decks and to identif.'v the prevailing causalfactors. Basedon detailed
in-situ and laboratory investigation of' several def'ective bridge decks araund
the kingdom, darnage may be broadly classified into (i) structural darnage
resulting from overloading and ( ii) material darnage resulting from
environmental impact on the durability of concrete. This paper presents a
typical bridge case studyfor each type of damage. Each case study involves a
detailed description of' darnage inflicted on the bridge deck, along with the
investigative approach adopted by the authors to determine the major causal
factors. The paper is concluded hy a 'damage likelihood chart'for the various
types of'bridge decks around the kingdom. The chart accountsfor variations in
the bridge geometry as ll'ell as its geographic location.

INTRODUCTION

Over the past two decades several hundred concrete highway bridges have
been built in Saudi Arabia as part of the development of a modern highway
network. Some of these bridges have suffered prematurely from excessive
cracking, deterioration, loss of Serviceability and failure. Several reasons are
549
550 M. Y. Al-Mandil et al.

believed to have contributed to these phenomena, among which are low

quality of concrete constituents, poor construction practices, lack of control
over vehicular loads and the harsh environmental factors.
The impact of the harsh environmental factors on the durability of
concrete structures in the Arabian Gulf region is well recognized in the
literature. 1 - 3 Although these.durability problems and the causal factors for
deterioration which are applicable to a concrete structure are equally
applicable to concrete bridges, the severity and unpredictability of bridge
loadings and the direct exposure ofthe bridges to hostile environments have
tagether paved the way for an aggravated assault on durability of bridge
decks.
Realizing the seriousness and the scope of deterioration of concrete bridge
decks, the Ministry of Communications (MOC) in association with King
Fahd University ofPetroleum and Minerals (KFUPM) have embarked on a
national project to study the bridge deck cracking phenomenon in the
kingdom of Saudi Arabia. The project was sponsored in 1984 by the King
Abdulaziz City for Science and Technology (KACST) for a duration of 4
years.
As part of this national project, in-depth studies were conducted on 21
defective bridge decks around the kingdom. These bridges were selected as
representative examples of the types of darnage identified to be most
prevalent in short-span bridge decks. Darnage to the deck system has been
broadly classified into (i) structural darnage and (ii) material damage.
Structural darnage to girder-slab bridges evinces itself as localized failure
in the form of potholes resulting from punching shear or combined shear
flexure failure. Darnage in slab-type decks, with slab thickness appreciably
greater than that in girder-slab decks, manifests itself in the form of a
reetangular grid pattern of cracks in the soffit, with crack widths up to
2-3mm.
Material darnage in the form of corrosion of deck slab reinforeerneut
occurred in the presence of a chloride and sulphate contaminated
environment, hypothesized to have transpired either due to lack of control
of mix ingredients and curing water or due to a subsequent ingress by the
aggressive salts in the coastal regions.
This paper presents a representative bridge deck case study for each type
of darnage identified to be prevailing in short-span bridges in the kingdom,
i.e. structural darnage and material damage. The D2 bridge deck is a vivid
example of localized failures inflicted on girder-slab bridge decks by the
grossly overweight trucks. The EP2 bridge deck, on the other hand, presents
an example of the concrete degradation in the environmentally hostile
regions of eastern Saudi Arabia.
The paper concludes with a darnage likelihood chart for the various types
 Darnage to Concrete Highway Bridge Decks in Saudi Arabia 551

of existing bridge decks in Saudi Arabia. The chart is derived from the
extensive surveys and inspections of damaged bridges around the kingdom.

STRUCTURALLY DAMAGED BRIDGE DECKS: AL-DARB

BRIDGE CASE STUDY

General Description
This bridge is located on the Al-Darb to Abba road, 12 km northwest of
Al-Darb Village. The MOC reference number is 081-0035-0211 and is code
named D2. The bridge spans over a deep wadi and has three simply
supported spans of 16m and a skew angle of 30°. The deck has a road width
of 8·0 m for two traffic lanes. The bridge consists of a reinforced concrete
slab cast monolithically over four main girders. The girders are connected
transversely by three cross diaphragms in each span. A plan view and a
cross-section of the bridge are given in Figs 1 and 2 respectively.

Span & panel loc;otions

Locat•ons of cor ing positions & potholes

FIG. I. Plan vicw of deck slab for bridge D 2.

FIG. 2. Cross-scction of deck slab for bridge 02.

552 M. Y. Al-Mandil et al.

FIG. 3. Plate showing localized failures (potholes) in the deck slab of 02.

At the time of our inspection there were two existing potholes in the deck
slab where concrete has completely fractured and disintegrated, leaving the
rebars exposed (Fig. 3). The position of the potholes is shown in Fig. 1.
Numerous grid pattern cracks are noted on the soffit of the deck. Crack
widths in the slabs ranged from 0-4 to 1·1 mm. This bridge had been
abandoned due to the localized failures in the deck slab, and was bypassed
by a road across the wadi.

Assessment of Concrete Quality and Strength

Twelve cores were selected from various locations on the deck slab (as
shown in Fig. 1) for determination of concrete compressive strength. These
core strengths ranged from 11·5 to 20·6 MPa with a mean of 17-4 MPa and a
standard deviation of the mean of 2·7 MPa. Four extra cores were taken for
crack-depth determination and all of these cores exhibited full-depth
cracking.
The design specification required a minimum cover to reinforcement of
40 mm for top reinforcement and 30 mm for bottarn reinforcement. The cover
to the top reinforcement averaged 46 mm and the cover to the bottom
reinforcement averaged 30 mm. Cover to the steel was clearly adequate.
The coarse aggregate gradation results showed a relatively oversanded
mix. The ratio of coarse to fine aggregate (CA/ FA) was 1·5:1.

Structural Analysis
A finite element analysis using ICES package program 'STRUDL' was
carried out for bridge 02 to determine the service load moments, and to
assess the ftexural stresses in the deck slab as caused by truck loadings. Three
 Darnage to Concrete Highway Bridge Decks in Saudi Arabia 553

AXLE N O . - - - 2 3
NO. OF TYRES - - 2 4 4
AXLE LOAD CTONNESJ 6.0 22.0 22.0
AXLE LOAD CKNJ .59 215.7 215.7

TOTAL VEHICLE LOAD = 50 TONNES

: 490.4 KN

(a)

3.20m
ii40~
AXLE NO. 1 2 3 4 5 6
NO. OF TYRES - - 2 4 4 4 4 4
AXLE LOAD CTONNESl7.6 15.0 15.0 21.0 21.0 21.0
AXLE LOAD (KN) 74.5 147 147 206 206 206

TOTAL VEHICLE LOAD 100.6 TONNES

9136.5 KN

(b)
FIG. 4. (a) Bridgedesign vehicle DPK No. 1; (b) bridge design vehicle DPK No. 2.

loadings were considered: (i) AASHTO HS20-44 + 10% truck loading,

which was the prescribed design loading; (ii) DPK vehicle No. 1; and (iii)
DPK vehicle No. 2. The last two truck loadings (Fig. 4) represent some ofthe
high vehicular loadings as observed from the axle Ioad measurement study. 4 · 5
Table 1 presents the results of the final analysis of the deck for the three
different loadings. The maximum moments are listed for the deck slab and
girders. Only the positive slab moment in both directions (x and y) are listed
in order to check the stresses in the bottom reinforcement. The
reinforcements in the deck slab are running orthogonal to each other and
parallel to the x and y axes, as shown in Fig. 5.
554 M. Y. Al-Mandil et al.

TABLE 1
Live Ioad analysis for bridge D2

Type of loading Slab maximumpositive moment Maximum girder

(kNm/m) moment
(kNm)

AASHTO + 10% H20 17·67 12·18 454

DPK No.1 31·15 16·65 823
DPK No.2 33-09 17·65 1120

The stresses in the bottom reinforeerneut ofthe deck slab were checked in
both directions due to the maximum positive moment. In the transverse
direction, the maximumpositive moment is due to DPK No. 2 vehicle, Mx=
MoL+ (MLL +I)= 48·74 kN m/m, where MoL= 5·72 kN m/m. Taking a slab
width of 1m and effective depth = 180-38 = 142 mm, the depth of the
neutral axis from the top of the slab = 46 mm. This gives a steel stress (fs) of
220-4 N/mm 2 (i.e. 0·64.!;, ifJ;, = 345 N/mm 2 ), which is not critically high. Thus
the transverse moment would produce a number offine longitudinal tension
cracks since the stress level is moderate.
In the longitudinal direction, the maximumpositive moment is also due to
DPK No.2 vehicle, My= MoL +(MLL +l)=31·06kNm/m, where M 0 L=
8·11 kN m/m. Taking a slab width of 1m and effective depth = 180- 53=
127 mm, the depth ofthe neutral axis from the top ofthe slab = 49 mm. This

-f~~=---~--'1> 14 mm (ll 0·25 m. Top

~ 14 m m Cl 0·14 m Bottom

f-----if----4> 16 mm (il. O·IOm·

Top & Bottom
X
3d'

FIG. 5. Steel arrangements in the deck slab of bridge 02.

 Darnage to Concrete Highway Bridge Decks in Saudi Arabia 555

gives a steel stress of 252·0 N/mm 2 (i.e. 0·73/y), which will cause severe
transverse tension cracks.

Discussion of D2 Case Study

The deck slab of bridge 02 is characterized by the presence of two
potholes, with the soffit exhibiting signs of extensive structural cracking,
often in grid form. The crack widths in the slabs ranged from 0-4 to 1·0 mm.
The assessment of the strength of the cores does reveal that the concrete
was indeed below specified strength. The core strengths ranged from 14 to
20 MPa with a mean of 17·2 MPa, whilst the design specification called for
24·6 MPa. lt may be argued that core strengths will be lower than the in-situ
concrete strengths as a result of the coring process; nevertheless, the concrete
strength definitely falls short of the design specified strength, not only on an
average strength basis but also because over 50% of the cores had strengths
which were less than 75% of the design specified strength.
The cement content of the concrete estimated from chemical analysis of
the cores was 255 kg/m 3 , which is some 32% below the current MOC
specification requirement of 375 kg/m 3 for 21 MPa strength concrete.
The ratio of coarse to fineaggregatewas 1·5: 1, which is acceptable but
does indicate slight oversanding. The !arger aggregates were extremely
elongated, some exceeding 60 mm in linear dimension, and thus were far
from being ideal materials for concrete.
The detailed structural analysis ofthe deck system for D2 did not provide
us with conclusive evidence as to the causes of the formation of the two
potholes. lt does, however, vividly illustrate that deck overloading is the
main cause for the shear and ftexural cracks in the deck slab. Such high steel
stresses under service Ioads (/~ = 0·73/;) would most certainly Iead to
extensive ftexural cracking.
However, at such high stress Ievels the linear elastic FE model as used
herein becomes merely an indicator ofthe actual situationrather than a true
index of stress distribution. A more appropriate criterion for structural
behaviour in this seriously cracked deck would be in-situ monitaring of
strain distribution under a vehicle of known Ioad. Results of such a test
would indicate the nature of redistribution of stresses in the deck system,
and a more realistic evaluation of the actual situation could be made.

Hypothesized Mechanism for Pothole Formation

Turning our attention towards the assessment of the nature and
mechanism of the spalling and the pothole formation, it is noted that most
potholes are formed in the severely cracked regions as a consequence of a
punching shear type of failure, where pieces of concrete are pushed through
the slab reinforcing bars. A quick check on the deck's punching shear
556 M. Y. Al-Mandil et al.

capacity reveals that this deck entertains a high margin of safety against
punching shear type of failure on the basis of the ACI punching shear
formula. 6 Therefore a punching shear design deficiency can be discounted.
However, it has been shown in a relevant study 7 •8 that the punching
capacity can be impaired by the existence of a fl.aw within the slab intricately
developed by the active process of crack growth and crack nucleation. A
deck slab will normally be subjected to two types of cracking: (i) non-
structural cracking related to environmental factors such as plastic
settlement, shrinkage and thermal effects, and (ii) structural cracking caused
by the tensile stresses from the vehicular load action. Superimposed crack
prints of these two represent the current state of cracking, which is
continually beingaltered by the progressive crack growth due to dynamic
overloading effects. This may lead to the formation of a concrete zone which
is separated from the surrounding body along an enclosed perimeter of the
nucleated crack surface. This separation along a closed perimeter
constitutes, in effect, a fl.aw. Existence of random cracking at the slab soffit is
conducive to the formation of such a fl.awed zone, which may be dangerous
from the punching viewpoint.

MATERIALLY DAMAGED BRIDGE DECKS: EP2 BRIDGE

CASE STUDY

General Description
This bridge is located in the Dammam district of the eastern province
along the Abu-Hadriyah-Dammam road. The MOC reference number is
023-0093-0234 and is code named EP2. The bridge has three spans of
continuous slab type with a central span of 13·45 m and two end spans of
8 m. The bridgedeck is a voided slab system with a slab thickness of 110 cm.
The voids are circular, 20 in number across the section, with each void being
70 cm in diameter. Shown in Fig. 6 is a typical cross-sectional view of the
bridge deck.

Dimensions in cm

FIG. 6. Cross-section of deck slab for bridge EP2.

 Darnage to Concrete Highway Bridge Decks in Saudi Arabia 557

FIG. 7. View of asphalt cracking in EP2.

Attention was drawn to distress in the deck slab by evidence of severe

cracking in places in the asphalt road surface (Fig. 7). The asphaltcoverwas
stripped in such cracked regions in order to expose the concrete deck slab.
The concrete slab itself showed few signs of cracking, but sounding a
hammer on the concrete surface indicated delamination ofthe slab. Figure 8
shows the extent of the delamination in certain zones.
Since all indications pointed to corrosion-induced darnage in the bridge
deck, field measurements centred around corrosion-related parameters.
Initially concrete was chiselled away in the delaminated zones and steel
exposed at these sites. The reinforcement showed definite signs of corrosion
and in some places the reinforcement had disintegrated entirely.

FIG. 8. Close-up view of delaminated concrete on deck slab of EP2.

558 M. Y. Al-Mandil et al.

er Ltvtl in lbs /yd3

10 12 14 16
1·00 -

1·00

E
u

19.()
20.0
FIG. 9. Chloride profile in top 20 cm of EP2 deck slab.

Calomel half-cell potential measurements were made and all readings

indicated a state of active corrosion. In addition, cores 5 cm in diameter and
20 cm in depth were taken from the bridge deck for chloride profile
determination. lt was not felt advisable to take full-depth cores due to the
presence ofthe voided pipes. Figure 9 shows the chloride profiles for the top
concrete depth of 20 cm. Bridge EP2 shows an extremely high gradient of
Cl- level, with maximum measured values of 9·12lb/yd 3 (5-42 kg/m 3 ).

Discussion of EP2 Case Study

The threshold level of Cl- to activate corrosion depends on the cement
content, but for the normal range of cement content (from 335 to 490 kg/m 3 )
the threshold level of Cl- varies from 1·13lb/yd 3 (0·67 kg/m 3 ) to 1·65lb/yd 3
(0·98 kg/m 3 ), with lower threshold values for lower cement content. Thus the
Cl- level far exceeds the threshold level throughout the top 20 cm region in
EP2, which indicates that the top reinforeerneut is susceptible to corrosion.
What is of great interest is the fact that there is a definite indication of
chloride ingress into the deck from the top surface. If the chloride
contamination were only to be from aggregate contamination and/or mix
water, the gradient would not have been as pronounced. One can only
hypothesize the source of this ingress, which includes the following:

(i) Chloride-contaminated curing water.

(ii) Airborne-contaminated pollutants deposited on the deck surface
periodically, which are then washed into the permeable deck with
condensation of moisture.
 Darnage to Concrete Highway Bridge Decks in Saudi Arabia 559

(iii) Periodic passage of transport vehicles loaded with saline water

dropping on to the bridge surface.

Presence of contaminated aggregates and mix water would only serve to

aggravate the problern further.
Inasmuch as up to 36 bridges in this sector are involved and several are
showing similar symptoms it appears that items (i) and/or (ii) mentioned
above may be plausible explanations. Item (ii) is considered a possibility
because the entire region is identified as a Sabkha region and the likelihood
of chloride-contaminated curing water (item (i)) should never be discounted
in a dry and arid region.
It may be that the mechanism of thermal incompatibility of concrete
components (TICC), as identified in the laboratory research component of
this project, 9 is playing a roJe in increasing the permeability of the concrete
decks over a period of time under repeated thermal cycling. This effect is
known to exist in concretes made with Iimestone aggregates (as in the eastern
province) due to the difference in the coefficient ofthermal expansion ofthe
aggregate and the cement paste. Diurnal and seasonal temperature
variations induce microcracking in the concrete structure over a period of
time, making it easy for chloride to ingress into the deck slabs.

Conclusions of EP2 Case Study

At this stage, based on analysis conducted thus far, it appears that there is
a very good probability that top reinforcement in the EP2 bridge is in an
active state of corrosion. Nothing can be said about the bottom
reinforcement in the absence of full-depth cores for the chloride profile. In
view of the fact that the bridge is continuous, corrosion of top steel in
negative moment regions can have very serious implications. In terms of
what needs to be done, one cannot say for certain that removal of
delaminated sections of the deck with subsequent corrosion-associated
repair would solve the problem. The reason for this isthat there has been a
substantial build-up of Cl- Ievel (at least in the top 20 cm of the slabs), to a
Ievel above the threshold deemed necessary for activating galvanic
corrosion. In addition, there isadefinite possibility that chemical corrosion
may also be present in view of the dangerously high Ievels of sulphates.
In retrospect, it may be stated that there is an urgent need for detailed
investigations of all affected bridges in this region. This should include full-
depth chloride and sulphate profiles, and half-cell potential contour plots.
Based on these results, the mode of repair can be decided, which could
include conventional corrosion repair in decks where the Cl- Ievel has not
built up excessively with chlorides and sulphates. Performance of cathodic
protection in SOl · -related chemical corrosion is not weil known.
560 M. Y. Al-Mandil et al.

CONCLUDING REMARKS

During the course ofthe national project, 5 21 bridgedeck systems have been
subjected to detailed investigations. Thesebridgescover a wide spectrum of
varying parameters, as they not only differ in their geographic locations
within the kingdom but their deck types, span lengths and deck widths also
vary. The common denominator amongst these bridges is that they were
undergoing a certain degree of distress to their decks at the time of
investigation. Each of these bridges was treated as a unique case study and
was diagnosed independently of other bridges in its region or of its own deck
type.
The o bjective of this section is to seek inferences from the case studies so
as to help identify certain global parameters common to the observed
darnage phenomena in an effort to minimize the probability of such
occurrences in the future.
lt may be concluded, in retrospect, that the shape and form of darnage
most likely to occur in a given situation is a function of bridge geometry,
proportioning and environment. Shown in Table 2 is a darnage likelihood
chart based on the evidence collected as a result of the detailed study of the
21 bridges in the kingdom.
In order to minimize the likelihood ofthe occurrence ofthe most common
forms of darnage as identified during the tenure of this project, certain
precautionary steps need to be exercised in order to avoid darnage

TABLE 2
Darnage likelihood chart for bridge decks in the kingdom of Saudi Arabia

Type of Environment Possible mode of Causa! factors

bridgedeck darnage

(i) Girderjslab Non-aggressive (i) Pothole in slab (i) Overloading

(ii) Box girder (ii) Rectangular-grid (ii) Low-strength concrete
cracking on slab (iii) U nder-design
soffit
(iii) Shear/ftexural
cracking in girder
(iii) Slab Non-aggressive Rectangular-grid (i) Overloading
cracking on slab (ii) Low-strength concrete
soffit (iii) Under-design
(iv) Thermal effects
All types of Aggressive Reinforcement (i) Chloride-contaminated
concrete bridges corrosion materials andjor mix,
curing water
(ii) Improper mix design
 Darnage to Concrete Highway Bridge Decks in Saudi Arabia 56!

occurrence in the future. These steps must cover all phases of the deck's life,
namely its design, construction and operational maintenance.

ACKNOWLEDGEMENTS

The authors wish to thank Y. N. Ziraba and G. S. Al-Sourti, research

associates of the national project, who have assisted in various activities of
the project, including their involvement in the case studies. Thanks also are
due to MOC, KFUPM and KACST for their technical and financial support
to the project.

REFERENCES

I. AL-TA YYIB, A. J., RASHEEDUZZAFAR and AL-MANA, A. r., Deterioration of

concrete structures in the Gulf States. First International Conference on the
Deterioration and Repair of R/C in the Arabian Gulf, Bahrain Society of
Engineers, Bahrain, October 1985, pp. 27-47.
2. RASHEEDUZZAFAR, DAKIL, F. H. and AL-GAHTANI, A. S., Deterioration of
concrete structures in the environment of the Middle East. Journal of the
American Concrete Institute, 81(1) (January 1984) 13-20.
3. The CIRIA Guide to Concrete Construction in the Gulf Region. CIRIA Special
Publ. No. 31, London.
4. PEARSON-KIRK, D., AL-MANDIL, M. Y., AZAD, A. K., BALUCH, M. H., SAHIR,
A. M. and ZIRABA, Y. N., Truck loadings for design of concrete bridges in Saudi
Arabia. Proceedings of a Symposium on Concrete and Concrete Structures in the
Middle East, King Saud University, March 1987. Riyadh, Saudi Arabia, 1987.
5. AL-MANDIL, M. Y., AZAD, A. K., BALUCH, M. H., SHARIF, A. M., PEARSON-KIRK,
D. and AL-DHALAAN, M. A., A study of cracking of concrete bridge decks in
Saudi Arabia. National Project Final Report, KFUPM, Dhahran, Saudi
Arabia, ~pril 1989.
6. ACI Standard, Building Code Requirements for Reinforced Concrete (1983).
American Concrete Institute, ACI 318-83, Detroit, Michigan, 1983.
7. KAREEM, K., Load-induced cracking and failure of concrete deck slabs in girder-
slab type bridges. MSc thesis, Department of Civil Engineering, KFUPM,
Dhahran, Saudi Arabia, March 1989.
8. AL-MANDIL, M. Y., AzAD, A. K., BALUCH, M. H., PEARSON-KIRK, D. and SHARIF,
A. M., Punching shear failure of concrete girder-slab type bridge decks in Saudi
Arabia. Proceedings of the Fourth International Conference on Structural Faults
and Repair, University of London, June 1989.
9. BALUCH, M. H., AL-NOUR, L. A., AZAD, A. K., AL-MANDIL, M. Y., SHARIF, A. M.
and PEARSON-KIRK, D., Concrete degradation due tothermal incompatibility of
its components. Journal o{Materials, ASCE, 115 (August 1989).
 49
Traffic Load Simulation Programme

DIETRICH LEBEK
Kaliwitzweg 11, D-5000 Köln 91, FRG

ABSTRACT

Most existing bridge codes are almost exclusively orientated towards bridge
design = to-be-built hridges. The reliabi/ity degree is not the samefor decisions
on to-be-built bridges and existing bridges. A 'service concept' must be pul at
equal rank side by side with the 'design concept'. Both concepts complement
each other. Neither can replace the other one. A traffic Ioad simulation
programme is an integral part of the 'service concept', allowing the evaluation
of the individual bridge for its actual everyday traffic Ioad. The summarised
generat considerations and a state-o{the-art review in brief underline the need
for a 'service concept'. The aim of a simulation programme is dejined and the
basic requirementsfor its development are described. A proposal is presented
for a complete trafjic simulation programme and guidance givenfor its usage.

INTRODUCTION

Maintenance of road bridges in a condition to provide safe and

uninterrupted traffic flow is the primary aim of every transportation agency.
A growing task facing bridge engineers in the near future is the rational
assessment of existing bridges and optimum allocation of resources among
decisions regarding unqualified acceptance, restricting traffic, rehabilitation
or closing existing bridges. Primary among the considerations is the safety
aspect. The need is for flexible evaluation options which recognise the site-
specific behaviour and account for the frequency and configuration of
everyday traffic. The everyday safe load-carrying capacity serves as a basis
for many keynote decisions about the bridge. A powerful tool for achieving
563
564 Dietrich Lebek

this goal is a flexible traffic Ioad simulation programme which is easily

adjustable to the specific site conditions of the individual bridge.

SERVICE CONCEPT

General Considerations
Bridges are designed as structures with a potential for long service life.
Bridgedesign is based upon codes and standardsvalid at the time of design.
However, codes and standards must be reviewed, changed or adapted to
changing requirements from time to time, i.e. the duration oftheir validity is
limitedas compared with the expected service life of a bridge. At present we
experience a rather rapid change of service conditions and a principal
change of our basic safety concept from a predominantly deterministic to a
predominantly probabilistic safety concept.
The Ievel of safety of a bridge and its components is determined by the
ratio of the sum of acting service Ioads and the resistance. The resistance is
not necessarily a time-independent value. Frequently there exists an
interrelationship between the resistance and the kind and duration of service
Ioads and service conditions (e.g. fatigue problems).
In addition, experience has shown that most technical products suffer
from decreasing 'reliability' (R 1) with time. There are strong indications that
bridges are no exception to this general rule. Such a decrease could be
conveniently expressedas an exponential function in the generalforme- ;.,r,
where t = number of years of the bridge in service and ;,; = function of
failure mode and maintenance intensity.

I0 15 20 23 30 40 50 50 70 80 90 100 t I years)

'i lunctian al Iailure mades and maintenance intensity

e lass al 17.3% and 26.7 '/, alter 15 and 23years al serv1ce respectively
due ta delective rivets

FIG. 1. Decrease in bridge 'rcliability' over the years with regard to load-carrying capacity.
- - , Decrease of the reliability due to wear and tear and other factors; ---, erratic decrease
duc to overloading; - - - , duc to complctc omission of maintenancc.
 Traffic Load Simulation Programme 565

The failure mode will be influenced by parameters such as live load

intensity and frequency, structural defects, fatigue, damage, deterioration,
change of assumed structural system, etc. Figure 1 illustrates possible trends
by a few examples ofbridges in service. The starting point of curves, R(t) = 1,
corresponds to the original reliability according to design assumptions and a
basically sound structure. The four cases, randomly chosen, may illustrate
the likelihood of the general validity of a function of failure mode and
maintenance intensity for bridges in service and typify what may be called
the 'creeping' or the 'erratic' change of the reliability of a bridge affecting its
load-carrying capacity. Figure 1 also provides an indication of the
'individuality' of bridges in service.

State of the Art

In the federal road system (i.e. motorways and federal roads), the Federal
Republic of Germany has a relatively new and modern bridge stock. The
first 'official' bridge census was made in 1965. Tables 1-3 illustrate the
development of the bridge stock over more than two decades.
Table 1 shows that about half of the present bridgestock has a service life
of25 years or more. Prestressed concrete bridges represent the youngest kind
ofbridge within the entire bridge stock. Only a marginal nurober ofnew steel
and composite bridges have been built in the last 15 years. About 90% of all
bridges presently in service are reinforced or prestressed concrete bridges.
Table 1 also indicates that the mean service age ofthe totalbridgestock will
increase more rapidly in the future since new construction has reduced.
Table 2 indicates the tendency within the last two decades to adapt bridge

TABLE I
DevC'lopmC'nt ol hrid[;e stock (lederal road systC'm onlr)
( percentage hased on total numher o{ hridges in service in 1987)
-------------- ----------

Year Kind of bridge Total

·-----·---~-----
bridge
Steel Composite Concrete Prestressed stock
concrete
--------------- · - -----------

1965 76 55 68 15 49
1970 95 79 80 32 63
1975 93 I 09" 87 54 76
1980 103" 103" 94 75 88
1987 100 100 100 100 100
------ ·- - -----·-----
"Between 1975 and 1987 some steel and composite bridges have been
replaced by other kinds of bridge structures, foremost by prestressed
concrcte bridgcs.
566 Dietrich Lebek

TABLE 2
Development of the statistical mean bridge dimensions in metres
(jederat road system only)

Year Kind of bridge Total bridge

stock
Steel Composite Concrete Prestressed
concrete

Span Width Span Width Span Width Span Width Span Width

1965 51·9 11· 3 83·2 18-4 14·5 13·1 64·1 15·1 25·1 13-7
1970 49·4 13-5 88·9 18·7 15·0 13·1 69·1 15-4 29·4 14·5
1975 51·4 14·8 82·2 18·9 14·7 14·7 74·9 18·0 34·6 16·7
1980 47·6 15·7 94·3 18·8 14·5 16·3 76·2 18·0 37·5 17-4
1987 46·6 17·5 101·1 18·7 14·2 17·2 75·6 17·8 39·6 17·7

dimensions to modern traffic requirements. One criterion is the development

of the bridge deck width. With the exception of composite bridges, the
statistical width of the bridge deck increased from about 11 to 14m in 1965
to about 17 and 18m in 1987.
This fact has implications as regards the total amount of designlive load
moment. The design live load consists of two components: an axle load
configuration and a uniformly distributed live Ioad. In the German bridge
design code, the axle load configuration is independent of the bridge deck
width. The amount of uniformly distributed live Ioad is directly
proportional to the bridge deck width. Taking the total design live Ioad
moment for the mean bridge in 1965 as unity, i.e. equal to 1·0, the total design
live Ioad moment for the mean bridge in 1987 would be 1·09, or about 10%
greater. However, both bridges have to carry the same daily traffic service
Ioads since the actual number oftruck lanes remained unchanged for both
bridges.
Another interesting aspect is: the present-day valid code would require a
factor of 1·35 for the same bridge, or a 35% greater total design live Ioad
moment. This underlines the statement made above about the limited
duration as compared with the expected service life.
Table 3 indicates that about three-quarters of the existing bridge stock
belongs to the range ofbridge lengths between 5 and 100m, i.e. the range of
small to medium span bridges. This is the range mainly affected by modern
truck traffic.
Intensive measurements have revealed interesting facts. As regards
directional traffic on motorway bridges, 75% have to cope with a yearly
transport volume of 10 million t or more, 50% with 20 million t or more,
15% with 30 million t or more and 1% with 40 million t or more. This again
 Traffic Load Simulation Programme 567

TABLE 3
Range of hridge lengths (federal road system only)
(percentage hasedontotal numher of hridges in service in 1988)

Length range Kind of bridge Total bridge

between -------- stock
Steel Composite Concrete Prestressed
concrete
-------- - - - - - - - - -

5 and 30m 30 26 61 26 45
30 and 100m 12 45 9 61 29
5 and lOOm 42 71 70 87 74

underlines the 'individuality' of bridges in service. Between 1978 and 1984

the mean gross vehicle weight increased about 10% for a representative
truck traffic configuration. The darnage equivalent value, used for fatigue
loading, increased by about the same amount. This corresponds to an
average increase rate of more than 1·5% pa at present. Predictions made
public recently indicate an increase of 30% for truck traffic intensity in the
European Community in the near future.

SIMULATION PROGRAMME

The Aim of a Traflic Simulation Programme

The primary aim of a traffic simulation programme is the evaluation of
the safety (or reliability) of road bridges and their components under the
action of 'everyday service Ioads', taking into consideration the actual
condition of the bridge and its components.
The expression 'everyday service Ioads' includes the following service
conditions:

-'Normal' traffic conditions, i.e. free-flowing truck traffic in the usual

configuration at the bridge site.
-'Specific' traffic conditions, i.e. truck traffic flow influenced by specific
measures or conditions such as specific road alignment conditions,
access to industrial or harbour arcas, etc.
-Conditions caused by specific infrastructure conditions or require-
ments, either limited to a certain duration or unlimited in time.
-Conditions caused by construction measures or structural require-
ments.

The evaluation ofthe safety (or reliability) should include static as weil as
568 Dietrich Lebek

dynamic aspects and, if required, fatigue aspects for the entire structure,
giving due consideration to redundancy aspects.
Another essential objective is the usage of the traffic simulation
programme as an everyday tool by the practising engineer (foremost in the
fields of public service and consultant engineering). Expertly used, this tool
can provide the much needed data serving as a basis for many keynote
decisions about the bridge.

Principles of a Simulation Programme and a Design Load Concept

A traffic simulation programme and a design Ioad concept are based upon
quite different principles. It is essential that this fact is clearly understood for
the development of a simulation programme as weil as for its proper usage.
This fact is completely independent of the underlying basic philosophy on
which the design Ioad concept is based, e.g. on working stress design, on Ioad
factor design and on a basically deterministic or basically probabilistic
concept.
- The principle of a simulation programme is the individuality.
- The principle of a design Ioad concept is the general validity.
In consequence of this, the following statements can be made:
- The design of a bridge orientated towards a long service life should be
based upon the principle of the general validity.
-An effective management for bridges should be based upon the
principle of individuality.
-Design codes or standards and a traffic simulation programme should
be viewed as sensible complementary principles rather than competitive
or, even worse, contradictory principles.
-Design codes or standards can be based upon the regulation principle
whereas a simulation programme must not be based upon the
regulation principle.

Basis and Requirements for the Development of a Traffic Simulation

Programme
The development of a traffic simulation programme is dependent upon
certain requirements. The following Iist of general requirements may not be
complete in every detail. However, the Iist contains the more important
items. The requirements concern such main items as Ioad characteristics,
traffic composition, traffic ftow characteristics, calibration of basic data,
accuracy of the models, and hardware and software requirements.
--Distinction between typesoftruck (see also Fig. 2). Twenty-one types of
truck form the basis for the simulation programme.
-Acquisition of accurate and sufficient data-preferably at sites which
are representative of normal traffic flow-for the following items:
• vehicle gross weight}
• axle load (separate for each truck type)
• axle spacing
• vehicle speed
• distance between vehicles (separate for each lane)
- The load functions can be described for each type of truck by
parameters that are independent of the location.
- The distribution functions of the load function parameters can be
transformed into mathematical functions with sufficient accuracy.
- Verification and recalibration ofthe above data are required at suitable
intervals (preferably measurements should be made at intervals not
exceeding 10 years ).
- Traffic composition and traffic frequency must be assessed separately at
each bridge site and be verified at suitable intervals.
-Specific traffic conditions must be assessed at each bridge site and
transformed into suitable mathematical models.
570 Dietrich Lebek

-Structural models are required that reflect the actual condition of the
bridge at the time ofthe investigation. It is essential that the accuracy of
the structure's model(s) and the accuracy of the load models are
compatible.
-Bach module of the simulation programme should be developed with
regard to the practical capacity of a suitable type of personal computer.
Consideration should be given to the possibilities offered by modern
hardware to guide the user by including a suitable form of diagram or
'pictures' in the programme for the various modules.
- The resulting strains and stresses should be given in the form of
distribution functions independent of the kind or the form desired, e.g.
stresses, moments, forces, etc. Only distribution functions can make it
possible to evaluate extreme values and their likely frequency of
occurrence, and thus assist in a reliable assessment ofthe bridge and its
components.

LAYOUT OF A SIMULATION PROGRAMME

The need for flexibility has been stressed repeatedly above. Flexibility is an
essential element for every possible form of a suitable layout. The number of
modulestobe included in the programmeisnot a decisive item as long as the
programme is capable of achieving its desired objective.
Two general ways appear to be possible for a layout:
-A clear definition of a desired objective resulting in a programme that
suits this objective only. As a consequence each objective would require
a specific programme.
-A complete layout for a simulation programme that is built up step by
step whereby the necessary number and sequence of modules can be
chosen to suit the desired objective(s).
In the following, the second method will be illustrated. However, the first
may offer advantages for certain conditions. The proposed layout considers
the step-by-step method. The modules 1-3 and 6 are devoted tomodeHing
traffic loads and traffic conditions. Module 1 is the 'foundation' for all traffic
loads and conditions. Module 6 is of similar importance if fatigue is
considered as an essential item. Module 2 covers the traffic stream for
normal conditions. Module 3 is intended to cover specific traffic conditions.
Module 4 should be interpreted as a summary of the model(s) needed to
describe the structural characteristics. This module will only suit certain
conditions, i.e. in general each bridge will require its own specific built-up
module. Nevertheless, specific elements could be programmed in this
 Traffic Load Simulation Programme 571

Trafik Ioads Structure

Module 1

r
Trafik Ioad
~
characteristics
I
I Module 2 Module 4
I
I
I
- Trafik 'fiow'
for normal
traffic conditions
- r--- Structural model

....__
I
I Module 3
I
L
- Traffic 'flow'
for specific
traffic conditions lt
I
I
t

-
Module 5

r Superposition with other decisive Ioad conditions -l

I I
I I
I Module 6 Module 4
I
L- Fatigue Ioads and
fatigue conditions
Structural model - .J

~
,
Module 7
Summation of results (distribution curves)
(stress, strain, moments, forces)

FIG. 3. Proposed Iayout for a complete traffic simulation programme.

572 Dietrich Lebek

modulein such a way as to be applicable for more general use. Module 5 is

reserved for the superposition of traffic loads with other defined loads or
load combinations. This module will be needed for more in-depth
investigations. Defined loads or load combinations could be, for example,
prestressing, creep, shrinkage, temperature, stress redistribution, uneven
settlement, etc. Finally, module 7 summarises all the modules needed to give
final results for a given objective in the form of distribution curves of
whatever is required, e.g. stress, strain, moments, forces, etc. The proposed

TABLE 4
Description of the modules forming the simulation programme

Module Descriptive remarks Dependent on,

independent of

Load characteristics of each type of Dependent on development

truck: distribution function of gross in course of time,
vehicle weight, axle Ioad and geometry independent of site
2 Trafik ftow (QJ in each lane for Traffic composition
normal traffic conditions, Q1 in truck dependent on local
lane and Q2 in passing lane(s); Q1 conditions, speed and
reflects the local traffic composition distance, independent of
and its linear distribution functions, local conditions
i.e. constant speed and distance
between vehicles according to
distribution functions
3 Traffic ftow for specific conditions not Dependent on location, on
included in module 2, e.g. specific specific situation; variable
infrastructure, gradients and repair or constant with time
4 Structural model, e.g. inftuence lines Dependent on structural
for moments, shear, torsion or their characteristic, material,
combined actions; FE models, grid condition, etc.
models, etc.
5 Superposition of Ioads; Ioad Dependent on structural
combinations, including temperature, characteristic, material
creep, shrinkage and stress property, time and
redistribution, could be of importance condition
6 Fatigue Ioads and/or fatigue Dependent on traffic
conditions for fatigue-prone composition and density,
components kind of material and detail
7 Summary of resulting stresses, strains, Same as for modules
moments or forces in the form of 5 and 6
distribution functions
 Traffic Load Simulation Programme 573

Iayout in Fig. 3 shows a possible arrangement of the seven modules briefly

described above. Table 4 sums up the modules in a descriptive manner.

USAGE OF A SIMULATION PROGRAMME

The scope of usage is largely dependent upon the imagination and flexibility
of the user. An extension of the number of modules is just as possible, as is
the enlargement of an individual module by introducing submodules. In any
case further development ofthe simulation programme is most desirable. An
interrelationship with management systems should be planned as a further
step of development.
The following Iist of possible usage is meant as guidance to some practical
applications, which may range from the individual bridge to a set ofbridges
up to the network Ievel.

Individual bridge
-Safety of the structure and its components
-Redundancy and critical sensitivity
-Suitability and/or sensitivity to heavy transport
-Urgency and scope of special investigations
-Scope of repair, rehabilitation or strengthening
-Necessary Ioad reductions, but also possible upgrading in allowable
Ioads
-Possible change in inspection intervals

Route Ievel or set of bridges

-Kind, degree and scope of critical 'weak points' in a route
-Degree of Joad-carrying capacity of a route
-Suitability for heavy transport (route criterion)
--Kind, scope and justification for measures to be taken (rehabilitation,
strengthening, upgrading or downgrading)
-Sensitivity to possible changes of the route infrastructure

Network Ievel
-Number and location of the 'safe', 'critical' and 'unsafe' bridges
-Kind and degree of changes with time in the Ievel of safety in certain
types of structure and in certain structures designed on the same
principle
-Strong and weak points of comparable types of structure
574 Dietrich Lebek

-Implications of changes in the legally allowed loads

-Implications of changes in the traffic infrastructure

ACKNOWLEDGEMENTS

The author gratefully acknowledges the highly competent advice given and
the helpful support received by Dr rer. nat. H. Steinhilberand Dring. Lehrke
from the Fraunhofer-Institut für Betriebsfestigkeit, Darmstadt
 50
Canada's Advanced National Standard on Bridge
Evaluation

PETER G. BOCKLAND
Ruckland and Taylor Ltd,
1591 Bowser Avenue, North Vancouver, British Columbia, Canada V7P 2Y4

ABSTRACT

With time bridges deteriorate but traffic Ioads increase. There is insufficient
money available to replace all deficient bridges, so in the last decade efforts
haue been under way in Canada to identify with more accuracy than before
which bridges are deficient, a·hich parts are critical and by hmt· much theyfall
short.
In a simple step-by-step procedure several firsts' are accomplished for a
bridge evaluation code. The safety index ß is selected as a function of the
expected behaviour of a member and the consequences of its failure. Load
factors reflect rhe confidence with which the Ioads can he predicted. A hridge on
an ore hau! route in the Yukon Territory serves as an example.

INTRODUCTION

There are two reasons for evaluating a bridge, either the Ioads are increasing
or the bridge is deteriorating-sometimes both.
The only Ioads over which mankind can exercise control are Ioads caused
by traffic. But unfortunately for the bridge engineer there are all sorts of
pressures to keep increasing the allowable traffic Ioads.
Because strengthening of bridges is a very costly business and there is
never enough money anyway, it was decided in Canada to develop a more
accurate method of deciding when a bridge really needs tobe upgraded and
when, even though it does not meet the provisions of the design code, the
bridge is in fact adequate.
575
576 Peter G. Buckland

The result is known as clause 12, 'Existing Bridge Evaluation', of

Canadian Standard CAN/CSA-S6-88: 'Design of Highway Bridges'.
In order to determine the reasonableness of answers given by the theory, it
was calibrated against a series of bridges which had previously been
evaluated by other means.

PROCEDURE

Clause 12 is in typical Iimit states format. That is to say,

aRR?: rxLL(1 + /) + rxvp 1 + !Xv 2 D2 (1)

where

R = resistance
L =live (traffic) Ioad effects
I= dynamic Ioad allowance
D 1 = effects of all dead Ioads except unmeasured asphalt
D 2 = effect of unmeasured asphalt
rx = the appropriate factor
rxR = Uc/J, where c/J is the resistance factor given in the code and U is an
adjustment factor to 'fine tune' the values of cfJ to increase their
accuracy.
lt is common to consider only dead and live Ioads when using clause 12,
but ifit is thought necessary other Ioads, such as wind or earthquake, may be
included.
The difference between clause 12 and other codes is that the Ioad and
resistance factors are selected within clause 12, and arenot the same as those
in the main body of the standard. The reason for this is that because of the
high cost of upgrading a bridge it pays tobe more exact in the calculation of
factors. The basic equation (1) is often written in the form

LLRF = Uc/JR- (rxvl Dt + rxv2Dz) (2)

rxLL(l + !)

where LLRF, the live Ioad rating factor, is equal to 1·0 if the element of
bridge being considered can exactly carry the Ioad required and more than
1·0 if there .is spare capacity. If it is less than one, the element being
considered is substandard.
lt can be seen that the LLRF is sensitive to the difference in two !arge
numbers, the factored resistance and the factored dead Ioads. This explains
the value of fine tuning cfJ with the modifierU.
 Canada's Advanced National Standard on Bridge Evaluation 577

LIVE LOAD FACTORS

Live Ioad factors depend on three things:

(1) How weil the live Ioad is known. Factors are !arge when Ioads are
uncertain and smaller when Ioads areweil known. This is in keeping
with Iimit states philosophy.
(2) The amount of warning likely as collapse is approached; a sudden
failure in a non-redundant structure commands a !arger factor than a
ductile failure in a multiple-path system, for example.
(3) The consequences of failure. The failure of a deck element is a
nuisance but not catastrophic; the failure of the main compression
chord of a truss is far more serious and a greater safety margin is
appropriate.

TRAFFIC LOADS

Clause 12 defines four kinds of traffic:

-NP (non-permit), normal 'legal' traffic; maximum Ioads arenot known
with confidence.
-PS (permit, single-trip), an overload for which a special permit is
required, normaily a !arge indivisible Ioad with axle Ioads exceeding
legal Iimits; Ioads often not weil known.
-PM (permit, multiple-trip), a series of overloaded vehicles on a special
permit, usuaily a bulk haul from a mine, for example; individuallegal
axle Ioads are not exceeded; Ioads are usuaily weil controlled.
-PC (permit, controiled), an unusually heavy Ioad, escorted; no other
vehicles on the bridge; Ioad often weighed and therefore accurately
known.
The significance of these traffic groupings can be understood conceptuaily
by reference to Fig. 1. Figure l(a) shows a nominal resistance and nominal
Ioad. They are separated by a 'safety factor'. In fact the actual resistances and
the actualloads can be described by curves representing probability density
functions: most, but not all, resistances are greater than the nominal; most,
but not all, Ioads are less than the nominal.
The shaded area where curves overlap is the area where the resistance can
be exceeded by the Ioad, which would Iead to failure. The probability of this
occurring must be kept sufficiently small as to be an acceptable risk (which
will never be zero so long as the curves have no Iimits). Seen graphicaily, the
'safety factor' in Fig. l(a) must be sufficiently !arge that the shaded area is
acceptably small.
578 Peter G. Ruckland

NOMINAL
RESISTANCE
i) ACTUAL
;3 LOADS
::J
0
w
e:

EFFECT EFFECT

(o) LOADS AND RESISTANCES (b) LOADS KNOWN WITH MORE

EQUALLY VARIABLE PRECISION THAN IN F.ig.1(o)
FIG. 1. The relationship between 'safety factor' and variability ofloads and resistances with
the real safety (i.e. probability of failure) kept constant.

The probability of failure, demonstrated by the shaded area and usually

represented by ß, the safety index, is chosen in clause 12 as a function of
structural behaviour (e.g. redundant or not) and the importance of the
consequences of failure. This is done by simple reference to tables and it
requires some intelligent engineering judgement.
Figure 1(b) represents the case where the traffic load is well known-for
example an escorted weighed vehicle at crawl speed with no other traffic on
the bridge. In this case the safety factor can be reduced without changing the
shaded area or the probability of failure. In other words, despite the
reduction ofthe 'safety factor', the risk is unchanged and the bridge's safety
is not altered from the condition in Fig. l(a).

EXAMPLE-THE YUKON TERRITORY

One of the eight bridges evaluated in the Yukon Territory will serve as an
example. The Yukon has an area more than double that of the United
Kingdom and a population of 28 000. The two main industries are tourism
and mining. The few roads are vital to the economy.
Lead zinc concentrate from the territory's largest mine is trucked 300
miles (500 km) to tidewater along the only north/south road through the
mountainous wilderness. In order to stay competitive the mine wished to
increase the amount it could carry on each truck. A truck leaves the mine
every 20min on average.
The trucks are defined by clause 12 as category PM (permit, multiple-trip)
and the loads are well regulated. Figure 2 shows a PM truck from Curragh
Mine at Faro passing over the Pelly River Bridge.
 Canada's Advanced National Standard on Bridge Evaluation 579

FIG. 2. A bulk haul truck with overload permit (PM) on the Pelly River Bridge at Faro,
Yukon Territory.

Figure 3 compares the global distribution of the heaviest 10% of non-

permit trucks with the distribution of the PM trucks from the mine. It can
clearly be seen that to cover the range of vehicles a !arger Ioad factor is
appropriate for the NP (typical) traffic than for the PM traffic.
Cla use 12 is in fact built around the distributions of 'annual maxima', as
shown in Fig. 4. Since one would expect the nominal Ioads tobe exceeded at
least once per year, it is not surprising that the entire distributions are greater
than the nominal. But here again it can be seen that NP traffic should be
given a greater Ioad factor than the better regulated PM traffic.

Selecting the Safety Index and Load Factors

Table 1, which is a reproduction of Table 47A of CAN/ CSA-S6,
demonstrates how the safety index is chosen. INSPl, INSP2 and INSP3

0.5

0.4 0.4 ...,

~ "! ~
zw ffi
~

::::> O.J )( ::::> O.J

0 0
w w
ff: 0.2 g
0

ff: 0.2 ~
-'
0.1
~ 0.1
~
"'-'
o~--~~~~~~~ o~-.--.-..---.--.-
~
..........~--~
0 40 80 1 20 160 200 240 0 40 80 120 160 200 240
GROSS VEHICLE WEIGHT (1000 lb.) GROSS VEHICLE WEIGHT (1000 lb.)

(a) NP TRAFFIC - t.4EASURED GVW (b) PM TRAFFIC - MEASURED GVW

FIG. 3. Mcasured distributions of gross vehicle weight for NP and PM traffic (1000 lb =
4·45 kN).
580 Peter G. Ruckland

NOMINAL SAFElY
0.5 0.5 IAARGIN
f\IOIAINAL SAFElY IAARGIN
~ 0.4
z
~ 0.3 ~
..Ii! ~ 0.4

5~ 0.3 ~
g
~

g " "
e: 0.2 g e: 0.2
0
~
L&J
~ ~
~
0.1

0+--o--+-.....-~--.-- ........---..--..,....,....
0.1
~
o+-----......-.....-+.....,.........,,.........~,....,....
120 140 160 180 200 220 240 120 140 160 180 200 220 240
GROSS VEHICLE WEIGHT (1000 lb.) GROSS VEHICLE WEIGHT (1000 lb.)

(a) NP TRAfFlC - MEASURED GVN (b) PM TRAfFIC - MEASURED G\fN

FIG. 4. Distribution of annual maxima of gross vehicle weights for NP and PM traffic
(1000 lb = 4-45 kN).

denote the levels of inspection to which the bridge has been subjected. lt is
assumed that all bridges have routine inspection. INSPl refers to elements
that are not inspectable, INSP2 is routine inspection and INSP3 is for
critical or substandard elements that have been inspected by the evaluator
(who may notice clues about structural performance).
Sl, S2 and S3 relate to system behaviour. Sl is where failure of one
element can lead to total collapse, S2 is where one element failure will
probably not lead to total collapse (e.g. multiple load paths) and S3 is where
element failure leads to local failure only.
El, E2 and E3 refer to behaviour of the element being considered as it
fails. An El element is subject to sudden failure with little or no warning. An
E2 element also fails suddenly but will retain post-failure capacity. An E3
element is subject to gradual failure with warning of failure probable.

TABLE 1
Target reliability index ßfor NP, PM and PS traffic

System Element Inspection level

behaviour behaviour
INSPl INSP2 INSP3

SI EI 3-75 3·50 3·50

E2 3·50 3-25 3·00
E3 3·25 3·00 2·75
S2 EI 3·50 3·25 3·25
E2 3-25 3·00 2·75
E3 3·00 2·75 2·50
S3 El 3-25 JOO JOO
E2 JOO 2·75 2·50
E3 2·75 2·50 2·25
 Canada's Advanced National Standard on Bridge Evaluation 581

TABLE 2
Live and dead loadfactors
-----~ -- --- -----

Load Load Target reliability index fJ

factors ------~--- ---·--

2·25 2·5 2·75 3·00 3·25 3·50 3·75

~------- ----------

NP (XL 1·31 1·37 1-43 1-49 1·56 1·64 1·71

PM (XL 1·01 1·05 1·09 1·14 1·19 1·25 I· 31
PS (XL 1·13 1·18 1·23 1·29 1·35 1-41 1·47
PC (XL 1·05 1·09 1·14 1·18 1-23
01 cxv, 1·05 1·07 1·08 1·09 1·10 1·12 1·13
02 Cl.v2 2·32 2·51 2·70 2·90 3-09 3-28 3-47
------ ---- -- -- - - - - - · -

Once f3 has been selected, the Ioad factors are selected from Table 2. There
is in fact an intermediate step which depends on the analysis method used,
omitted here for simplicity.

Evaluating the Pelly River Bridge at Faro, Yukon Territory

The Pelly River Bridge at Faro, Yukon Territory (Fig. 2), provides an
example of the use of clause 12. The bridge was evaluated for the heaviest
allowed non-permit (NP) 'legal' vehicle, a truck with a gross vehicle weight
(GVW) of 140000lb (63500kg) on eight axles, and for bulk haul trucks
operating on a multiple-use overload permit (PM) and a GVW of 170 000 lb
(77 000 kg) on eight axles.
Interestingly, it was found that in virtually all cases the lighter NP vehicle
governed because of the !arge Ioad factors it commanded. The figures
following are for a top chord member in compression and for an interior
floorbeam in bending:
Chord Floorbeam
Inspection Ievel INSP2 INSP2
Element behaviour El E3
(failure sudden) (failure ductile)
System behaviour Sl S3
(single Ioad path) (multiple Ioad paths)
Safety index {J (from Table 1) 3·5 2·5

Appropriate factors (from Table 2)

:xL-NP 1·61 1·37
:xL-PM 1·23 1·05
:Xv, 1·12 1·07
':XD2 3·28 2·51
u 1·00 1·06
582 Peter G. Ruckland

When these numbers were inserted into the formula (2) for live load rating
factor, they yielded
Chord LLRF = 0·56-1·29 for NP traffic
= 0·62-1·35 for PM traffic
Floorbeam LLRF = 1·08 for NP traffic
= 1·27 for PM traffic
From this it can be seen that:
(1) NP traffic has a lower live load rating factor than PM traffic even
though the nominal applied load is less.
(2) Some chords (those with LLRF less than 1·0) must be strengthened
or, in some cases, braced to reduce the slenderness ratio. Both are
fairly simple operations.
(3) The floorbeams do not need strengthening, which avoids an
expensive procedure. However-and here is the important point-if
conventionalload and resistance factors had been used the LLRFs
would have been 0·92 and 1·08 for NP and PM traffic respectively,
and strengthening would have been required.
(4) For chords with an LLRF of 1·0 or greater the annual probability of
failure is 0·000 23, corresponding to ß= 3·5. For the floorbeams the
annual probability of failure, without modification, is about 0·0014.
In other words, by accepting a lower value of ßfor the ftoorbeams the
cost of strengthening now is avoided, but a 1 in 750 chance per year is
accepted that the floorbeams may yield in bending, an event that will
cause only inconvenience. The philosophy of clause 12 is that the
preferable choice for the taxpayers' money is to not strengthen the
floorbeams.

SUMMARY

(1) Clause 12 of CAN/CSA-S6-88 provides a simple to use method of

evaluating bridges which takes account of the type of traffic and the
expected warning and consequences of a failure.
(2) Using the Pelly River Bridge in the Yukon Territory as an example, it is
shown that because multiple-trip bulk haul vehicles are better
controlled than random 'legal' traffic higher nominal vehicle weights
may be permitted without reducing the real safety Ievels.
(3) By accepting a higher probability of failure for members which give
warning of failure and have multiple Ioad paths which, if they should
fail, will not cause failure of the bridge, expensive strengthening can be
avoided.
 Canada's Advanced N<~tional Standard on Bridge Evaluation 583

Forthose having difficulty with this concept the questiontobe answered, at

least qualitatively, is: would you prefer to have a floorbeam yield in bending
or a main chord fail by compression buckling? If the former, by how much
more would you prefer it?

ACKNOWLEDGEMENTS

The author wishes to express his appreciation to Mr Eric Gibson, manager

of technical services, Yukon Transportation and Community Services, for
permission to reproduce information about the Pelly River Bridge and for
the opportunity to evaluate the eight bridges in the Yukon Territory, and to
the Canadian Standards Association and Mr Lorne J. Hamblin, chairman of
the Technical Committee of CAN/CSA-S6, for permission to reproduce
parts ofthe standard. Clause 12 is a supplement to CAN/CSA-S6-88 and is
available with its commentary from Canadian Standards Association, 178
Rexdale Boulevard, Rexdale, Ontario, Canada M9W 1R3.
The author is greatly indebted to F. Michael Bartlett, formerly of
Buckland and Taylor Ltd, for his assistance in developing clause 12 and in
the preparation of this paper.

APPENDIX: CHOICE OF TARGET RELIABILITY INDEX ß

The target reliability index ß is a function of the notional target annual

probability offailure P. Note that clause 12 uses an annual probability, not a
lifetime probability. P is based on life safety criteria and is defined as
P=AK/(Wjn)
where
A activity factor, a measure of the risk associated with the activity (i.e.
=
driving a car)
= 3·0 for NP, PM and PS traffic
= 10·0 for PC traffic
K = a calibration factor = 10- 4
W = a warning factor
= 1·0 for no warning of failure expected
n = the number of people at risk
= 10 for NP, PM and PS traffic on spans up to 100m
= 1 for PC traffic (other traffic kept off the bridge).
For elements supporting !arger or more important bridges, n can be
584 Peter G. Ruckland

increased but ß is insensitive to n and a value of 10 was chosen as a typical

value for most bridges.
These values lead to:
For NP, PM and PS traffic P;:;:,9·5xl0- 5 ß;:;::, 3·75
For PC traffic p 0::::, 1·0 X 10- 3 ß;:;::, 3·25

These values of ß are then reduced in a systematic way to account for

improved warning offailure (which comes from inspection and from ductile
behaviour) and for consequences of failure that are other than catastrophic.
 51
Serviceability Assessment of Masonry Arch Bridges
Using Vibration Tests

A. J. PRETLOVE and J. C. A. ELLICK

Department of Engineering, University of Reading, Reading, UK

ABSTRACT

Over the last 5 years an extensive programme ofresearch at the University of

Reading has explored the use of vibration methods in the assessment of the
Serviceability state o{ masonry arch bridges. Bridges in Derbyshire, West
Sussex, Hampshire and Berkshire have been tested in the studies. Not only has
a variety ofmethods 4excitation been used but also a range ofmeasurement
and analysis techniques. The resultsfrom this work have been compared with
the standard MEXE method ofassessment and with building vibrationdarnage
criteria such as those contained in the German Standard DIN 4150. An account
of the techniques which have been developed and used is given here. Same of
them have been shown to be ofno sign[ficant value whilst others are seen to be
more promising. The better methods can, at the least, give additional
information which can be considered alongside the present standard method.
Recommendations are made as to their use. These methods require some
development and verification before they can become suitable for routine
assessments.

INTRODUCTION

There is considerable concern that some masonry arch bridges in Britain

may be deteriorating rapidly as a result of increasingly heavy traffic and axle
Ioads. An accurate assessment of the structural condition of these bridges is
known to be extraordinarily difficult because of the wide range of
configurations and materials that have been used in their construction.
There is also a wide range of ages and, indeed, some of these bridges are
585
586 A. J. Pretlove and J. C. A. Ellick

classified as ancient monuments. Aceurate methods of structural condition

assessment are required if timely and cost-effective maintenance is to be
achieved. Ifthis cannot be done with sufficient precision then, in some cases,
these structures may deteriorate increasingly rapidly as a result of
developing structural faults, not only in the pavements but also in the arch
structure itself. However, this is not to say that they will immediately fall
down. The static strength of masonry arches is generally very much higher
than the applied traffic Ioads, as has been shown by some recent tests to
destruction. 1 The real problern is the progressive wear and tear of the
structure. This can take many forms and some frequently observed ones
are:

(a) mortar decay and erosion leading to arch barre! deformation or

cracking;
(b) separation of spandrei walls, together with the outer parts of the
barre!, from the rest of the arch barre!;
(c) fill deformation leading to an uneven pavement with consequent
increases in dynamic wheel Ioads and possible darnage to services,
such as gas pipes; and
(d) foundation settlement.

The problern is to determine accurately which structures have reached a

point at which they require maintenance or major repair. If the methods
used in this determination are not accurate then, on the one hand, money
may be needlessly spent on structures which are in adequate condition
whilst, on the other, structures on the point of collapse may not be detected.
A simple empirical method for assessing the capacity of masonry arches to
carry traffic was developed by the Military Engineering Experimental
Establishment (MEXE) in the 1940s. lt was based on Pippard's theory 2 and
various experimental studies. Details of the method, as currently used, are
setout in Ref. 3. In the method a provisional axle Ioad is calculated from the
span, ring thickness and depth of fill using a nomogram. This value is then
modified by a series of factors which takes into account the actual shape of
the arch, the type and condition of the materials, joints and workmanship,
and the presence and position of cracks. This Ieads to a permissible axle Ioad.
Generally the method is reckoned to be conservative and it only assesses the
load-carrying capacity of the arch barre!.
In this programme of work a variety of vibration techniques have been
studied with the aim of supplementing the information given by a MEXE
assessment and hence providing a more accurate indicator of state of
serviceability. The principle behind the use of vibration techniques is simply
 Serviceability Assessment Using Vibration Tests 587

that the vibration properties may change if the structural condition

deteriorates. Thus the following properties have been studied:
(a) natural frequencies,
(b) damping,
(c) comparisons of time traces,
(d) non-linear behaviour, and
(e) response values to given inputs.
The earliest work made use ofvibration excitation by normal traffic because
this had the obvious advantages ofzero traffic disruption. However, the wide
variability of this excitation, depending as it does on axle Ioads, speeds and
lines of action, precluded its further use. This was followed by work using
standard vehicles to provide excitation but this was also unsatisfactory
because response measurements made on the bridge included the dynamic
properties ofthe vehicle as weil as those ofthe bridge. A fresh approachwas
required: a testwas needed that would give a repeatable response that was
characteristic ofthat bridge and of no other. For this purpose an impact
device was developed which dropped a mass through a fixed height on to the
bridge surface. This has the disadvantage of requiring more equipment on
the bridge. However, not only could spectral techniques be applied to the
data but also the response traces could be subjected to various forms of
analysis. A series of tests on a number of arches showed that each one
produced a distinctly characteristic response.
In the latest work the impact test technique has been exclusively used with
a standard drop weight which is designed to apply a known and fixed
impulse to the bridge structure. This is described in more detail in the next
section.
A previous article on this work 4 has described the full range of work on
the various vibration response quantities listed above. The most promising
indicators of condition were found to be:
(a) Peak-to-peak velocity in response to a standard impulse. This ties in
with the German Standard DIN 4150, 5 which gives criteria for
building darnage due to vibration in terms of velocity.
(b) Fundamental natural frequency. The fundamental mode ofvibration
has in these tests always consisted ofvertical motion coherent across
the arch width and is thus termed 'bending'.
This paper concentrates on the analysis of these two features.
A wide range of tests has been made on arch bridges in England. Most of
these arches are in a reasonably sound structural state but one or two are
known to be reaching the point where maintenance or repair is needed. In
588 A. J. Pretlove and J. C. A. Ellick

this work the aim is to show that the two vibration methods listed above can
pick out these less sound structures.

DESIGN OF THE IMPACT DEVICE

In the latest version of the impa~t device a mass, which can be varied from
12·5 to 75 kg in six steps, is lifted by pulley and held by a bomb release on a
tripod system. lt is dropped through 1 m on to a thin bed of damp sand that has
been placed on the bridge pavement. The purpose of the sand is to prevent
bounce and hence provide a known and repeatable impulse. Some care is
needed with the dampness and thickness ofthe sand ifrepeatable results are
to be achieved. This impact device is not as complicated as the force
instrumented type developed recently by Bruel and Kjaer, and for this
purpose does not need tobe. The impulsetime has been made short (10ms)
compared with the fundamental period for this type of structure, usually
about 50 ms. The movement resulting from the impact is measured using a
seismometer. In most tests the impact is applied at the arch crown adjacent
to one parapet and the measurement is made adjacent to the opposite
parapet. This offers not only minimal disturbance ofthe traffic flow but also
simultaneaus excitation of bending and torsion modes of the arch ring. The
seismometer output is recorded on tape, and is taken back to the laboratory
for digitisation and computer analysis.
The 75 kg drop test has been shown in earlier tests to give a bridge
response roughly equivalent to that obtained from the passage of a 10-t
lorry.lt has therefore been used in these tests as it will give realistic values for
response, particularly necessary in the analysis of vibration velocities.

TEST MEASUREMENTS AND ANALYSIS

A typical responsetime trace is shown in Fig. 1 complete with its calibration

marker. The results of a Fourieranalysis of this trace can be seen in Fig. 2,
giving information on the natural frequencies of the structure. In this case
the fundamental (bending) frequency is at 12Hz. The torsion frequency can
also be detected though it is beyond the edge of the measurement range of
the instrumentation used, which is from 1 to 50 Hz.
The basic results from this procedure are shown for bridges numbered
1-11 in Table 1. For the purpose of comparison the table also shows the
modified axle load (MAL) calculated using the MEXE procedure. This gives
some indication of the state of Serviceability of the bridge because it includes
 Velocity

I 1 mm/sec
interval

0 0.1 0.2 0.3

Time Seconds
FIG. 1. Vclocity-time trace for bridgc 1.

dB

0
0 50 100
Frequency Hz
FIG. 2. Frcqucncy spectrum for bridge 1, derived from Fig. 1.

TABLE 1
Summary of the major dimensions and results obtainedfor the bridges included in
this study
------- -------- - - - - -- --- ------·---

Bridge d c SpanS CF MAL P-P6 Natural

(mm) (mm) (m) (t) velocity frequency
(mm/s) (Hz)
--·-----·

I 405 862 8·38 0·9 18 1" 12

2 355 I 064 5·26 0·9 38 3·33 28·9
3 457 737 9-40 0·9 14 4·92 11·8
4 470 800 8·28 1·0 12 3·60
5 405 1 230 6·70 0·65 12 2·27 14·6
6 335 830 6·03 0·7 12 1·96 17·4
7 565 785 8·62 0·7 7·5 5·51 15·3
8 340 760 6·57 0·8 12 3-41 15
9 353 648 6·55 0·8 11 4·69 15
10 340 740 6·55 0·6 8·5 1·34 15
11 432 686 6·71 5" 22

For a definition of the dimensions see Fig. 3.

"This figure has been estimated.
"Figures not available; bridge demolished 1987.
590 A. J. Pretlove and J. C. A. Eliick

a condition factor (CF) which is obtained by visual inspection of the

structure.

VIBRATION VELOCITY MEASUREMENTS

The peak-to-peak velocity values for a 75 kg drop test are shown in Table 1.
lt can be seen that there is not, in general, a clear relationship between peak-
to-peak velocities and the modified axle Ioad (MAL) obtained by the MEXE
assessment method. However, there are some useful remarks which can be
made. Bridge 7 has not only the highest measured velocity (5·5 mm/s) but
also the lowest MAL (and there are strong grounds for weight restriction
here). Bridges 3 and 9 have the next highest velocity values and both have
relatively low MAL, particularly so in the case of bridge 9.
The peak-to-peak values can be directly compared with the DIN 4150
guideline Iimit of3-8 mm/s. Several bridges are seen to have responses in this
band, but not beyond it. Thesebridgesare therefore close to the boundary of
vulnerability. Values quoted here have come from drop tests only. lt would
be useful to verify these results using real traffic excitation. This would
require the measurement ofpeak-to-peak vibration Ievels during busy times
ofthe day for a large nurober of arch bridges. The data so obtained could be
used not only as an indicator of the state of the structure but also as a
measure of the current rate of darnage (and hence the prospective
serviceability). Studies of this kind would confirm the relevance of drop-
weight tests to Serviceability assessment.

FUNDAMENTAL NATURAL FREQUENCIES

The values given in the last two columns of Table 1 may be referred to as
'raw' data. They take no account of the different dimensions of the various
arch bridges. In the case of natural frequencies this has to be taken into
account if valid comparisons are to be made. This can be done by
determining the mass and stiffness of an arch in relation to its dimensions.
Figure 3 shows the symbols used for the relevant dimensions of an arch
bridge.
In the fundamental (bending) vibration of the structure the major
movement is confined to a volume near the crown.lt is therefore the mass in
this area which is important in the vibration. lt is estimated that the
generalised mass M is proportional to hcS. The shallowness, or otherwise, of
the arch will only play a secondary part. The stiffness at the crown to a
vertical force can be derived from Pippard's theory for the elastic action in
 Serviceability Assessment Using Vibration Tests 591

Road Surface
I I I
I I

I~ s
FIG. 3. Arch bridge nomcnclature: h (not shown) = arch/roadway width; c = crown depth
including the arch ring; d = arch ring thickness, assumed constant; and S = span.

the arch ring as given in Ref. 1. In summary, the defiection at the central
point of load is given by Castigliano's theorem:
6= au1aw
With strain energy

U= 2 f S/2 M2

0
2E;0 dx
F or the arch ring
10 =hd 3 112
The arch ring bending moment

Mx
= Wx:
128
[too x - 36]
S

Evaluation of the elastic defiection 6 in this way shows that the stiffness K is
proportional to hd 3 IS 3 .
The fundamental natural frequency is

w 1 =filM
and this will therefore be approximately proportional to d 3 12 IS 2 c112 . We
have called this quantity the bending frequency factor (BFF).
Table 2 shows, for the bridges considered here, the calculated BFF, the test
natural frequency from Table 1 and R, the ratio ofthe two. One would expect
this ratio to be constant. If it is less than the general norm then it suggests
that the testnatural frequency is Jow and this may be an indicator of poor
structural condition.
Formost of the bridges the ratio R is remarkably consistent at a value of
about 3 (in the mixed units used), indicating that the theory above is
592 A. J. Pretlove and J. C. A. Ellick

TABLE 2
Test frequencies and calculated factors for bending

Bridge Test bending BFP R=TBF/BFF

frequency (TBF)
(Hz)

1 12 3·95 3·0
2 28·9 7-41 3-2
3 11·8 4·07 2·9
4 5·25
5 14·6 5·18 2·8
6 17-4 5·85 3·0
7 15·3 6·45 2-4
8 15 5·27 2·8
9 15 6·07 2·5
10 15 5·37 2·8
11 22 7·61 2·9

a BFF is defined in the text.

appropriate and reasonably accurate. Bridges 7 and 9 fall significantly below

the mark. Thesebridges also were shown to have high velocity values and, in
the case of bridge 7, the lowest modified axle load. There is, therefore, a
consistency about all ofthese results which supports the view that these tests
provide valuable indicators of serviceability state.

CONCLUSIONS

This work has concentrated on the use of impulsive tests on masonry arch
bridges and analysis of the conseq uent vibration to assess the Serviceability
states of these structures. Two features have proved to be consistent and
promising indicators of structural condition.
First, the measured peak-to-peak velocities (see Table 1) have shown that
the highest value (bridge 7) corresponds to the least permitted axle load from
the MEXE assessments. This value (5·51 mm/s) is not above the guideline
band recommended in DIN 4150 for structures in this dass (3-8 mm/s).
Second, the measured fundamental (bending) frequency of vibration has
been found in all cases to be close to a prediction based on

where the dimensions and their units are defined in Fig. 3 and Table 1. The
constant K is consistently close to a value of 3. The bridge which falls most
 Serviceability Assessment Using Vibration Tests 593

short of the prediction given by this formula is again bridge 7, with a

constant of 2-4. This indicates a possible Iack of stiffness in this structure.
Two tests are therefore recommended for common use which may
supplement the information given by a MEXE assessment.

(1) To measure the peak-to-peak vibration velocity at centre span in

response to a drop-weight test using 75 kg dropped from a height of
1 m on to a 50 mm sand bed on the bridge at its centre span. The
mass is tobe dropped near one parapet and the measurement tobe
made near the other parapet. Any result in excess of 8 mm/s will give
rise to concern about the Serviceability state of the structure.
(2) The fundamental natural freq uency of the structure iJ (Hz) is to be
measured by any suitable and available means. A drop-weight test on
a 50 mm sand bed on the bridge at its centre span, as described here, is
recommended. The constant Kin the above formula is then to be
calculated using measured values of c (mm), d (mm) and S (m). A
value of Kless than 2 will give rise to concern about the Serviceability
state of the structure.

It is an advantage to local authorities that these proposals are for

relatively siruple one-off tests. However, ~here is a need to develop standard
procedures and software for the analysis of results. These developments
should be pursued in further work.
In some respects the value of these recommended procedures requires
some supporting verification.ln the case ofprocedure 1 it would be useful to
compare drop-weight velocity values with those caused by normal traffic.
The case for procedure 2 would be strengthened if a comprehensively
accurate method can be developed for distinct definition of the bending
natural frequency. It is proposed that this could be achieved by means of
simultaneaus measurement using two seismometers, one at each parapet,
and dropping the weight on the bridge centreline. However, this will
unfortunately involve stopping the traffic in order to make the test.

ACKNOWLEDGEMENTS

The authors are grateful for help and advice from the County Surveyors
Departments ofBerkshire, Derbyshire, Hampshire and West Sussex County
Councils. Various colleagues at Reading and Dr D. W. Cullington and Mr R.
Eyre at TRRL have contributed to the work with discussions and physical
assistance.
The work was carried out under contract to TRRL and is published by
594 A. J. Pretlove and J. C. A. Ellick

permission of the Director, TRRL, Crown Copyright 1990. Any views

expressed are not necessarily those of the Department of Transport.

REFERENCES

1. PAGE, J., Load tests to collapse on two arch bridges at Strathmashie and Barlae.
TRRL Research Report RR201, 1989.
2. HEYMAN, J., The Masonry Arch. Ellis Horwood, London, 1982.
3. DEPARTMENT OF TRANSPORT, The Assessment. of Highway Bridges and
Structures: Advice Note BA 16/84 and Departmental Standard BD 21/84, 1984.
4. PRETLOVE, A. J. and ELLICK, J. C. A., Vibration techniques to assess the structural
condition of masonry arch bridges. Proc. Inst. Acoust., 10(2) (1988) 501-8.
5. German Standard DIN 4150, Part 3: Structural Vibrations in Buildings; Effects
on Structures, 1986.
 52
Assessing the Dynamic Properties of Existing Bridge
Structures by Hammer Testing

J. R. MAGUIRE
Lloyds Register ( Industrial Division), Lloyds Register Hause,
29 Wellesley Road, Croydon, Surrey, UK.

ABSTRACT

This paper describes the use ofhammer testing to assess the dynamic properlies
of existing bridge structures. Afterabrief reference to the background theory,
data acquisition and processing considerations are examined. One case history
is then presented relating to bridge beams at Basingstoke. The natural
frequencies, mode shapes and damping values of the beams are summarised. lt
is concluded that hammer testing provides a quick and accurate method of
assessing as-built structural dynamic properties, and it is envisaged that this
technique could be successfi.Llly used on many existing bridge structures.

NOTATION

c Damping coefficient (% critical)

c(n) nth damping coefficient (% critical)
E Young's modulus
f Frequency (Hz)
f(t) Applied force
F(w) Fourier transform of f(t)
h(t) Linear differential operator
H(w) Frequency response function
I Second moment of area
x(t) Displacement
X(w) Fourier transform of x(t)
595
596 J. R. M aguire

INTRODUCTION

Developments in the use of computers for structural design and analysis

have indicated 1 an urgent need for the knowledge of material properties and
the actual behaviour of as-built structures, which can only be provided by
tests carried out on those structures. Such tests which make use of applied
static Ioads are often difficult and expensive to carry out, whereas the
application of dynamic Ioads, whether steady-state or transient, is relatively
simple; although it has to be said that the collection and processing of the
response data require skill and experience if pitfalls are to be avoided.
Dynamic testing also usually operates at low input force Ievels, producing
low amplitude vibrations, usually in the elastic range of behaviour. The
dynamic force applied to the structure under test can be basically of two
types: first, steady-state forces such as are produced by rotating eccentric
mass exciters or hydraulic actuators; second, transient dynamic forces which
are produced by wind, explosions or direct impact. Steady-state studies have
been previously reported. 2 This paper describes the use of transient forces,
input to the structure by an instrumented hammer, to assess the dynamic
properties of existing bridge structures, and it presents one case history of
the use of this technique.
The vibration of a linear elastic structure may be described by the
combination of different modes of vibration, each of which has the modal
characteristics of frequency, mode shape and damping. Once these
parameters are known the dynamic behaviour ofthe structure subjected to a
known input may be predicted. The method of modal analysis 3 - 5 can be
used to derive these modal parameters from a hammer test, as has been
successfully demonstrated in the aerospace and automobile industries. 6 •7
This method is extended here to the bridge engineering field.

BACKGROUND THEORY

One of the main objectives of modal analysis is to break down a signal

measured from a vibrating structure into its components at various
frequencies. This is usually carried out using Fast Fourier Transform (FFT)

Excitation System Response

f(t) characterisbcs
h(t)
f-+-- x(t)

FJG. 1. Time domain representation of a linear SDOF system.

 Assessing Dynamic Properfies by Hammer Testing 597

Transtormed Frequency Transformed

excitation response response
F(w) ~ tunction -+- X(w)
H(w)
I
FtG. 2. Frequency domain representation of a linear SDOF system.

techniques, 8 based on the theory that any periodic signal may be looked on
as a combination of a number of pure sinusoidal curves with harmonically
related frequencies. 9 lt may be demonstrated that any system (and signal) in
the time domain (Fig. 1) has an equivalent representation in the frequency
domain (Fig. 2). The theory of dynamic systems is weil known and described
elsewhere. 1 o- 13

DATA ACQUISITION AND PROCESSING CONSIDERATIONS

A full discussion of modal analysis and digital signal processing is outside

the scope of this paper but is covered comprehensively by specialised
textbooks. 11 · 13 However, a few of the more important practical considera-
tions will now be briefly mentioned.

Digital Spectral Analysis

Aceurate practical evaluation of spectral functions involves considerable
computation and manipulation and is, therefore, ideally suited to digital
computer analysis. It should be noted, however, that the digitisation
procedure possesses a number of fundamental constraints (such as aliasing,
Jeakage and noise) which can severely Iimit the quality of results and Iead to
wrang interpretation.

Resolution
The single most important factor affecting the accuracy of calculated
modal parameters is the accuracy ofthe frequency response evaluation.lt is
not possible to extract the correct values ofthe modal parameters when there
is inadequate and/or insufficient information to process. Adequate selection
of frequency resolution, sampling interval and record length is therefore of
prime importance.

Aliasing
Adecisionon the sampling rate for digital data acquisition is dependent
on the analysis of the structure and is usually performed at equally spaced
time intervals. One task is to determine this interval; too short an interval
598 J. R. M aguire

willlead to more data than can be economically processed, whereas too long
an interval will lead to confusion between low and high frequency
components in the original data. This latter problern is known as aliasing,
and the preferred method of solving it is to filter the data before digitisation,
using low pass filters which cut out unwanted high frequency components.

Leakage and Windowing

To convert from the time domain to the frequency domain, the FFT is
used. One ofthe greatest sources of error in any digitally computed spectrum
results from the fact that the measured signal is probably not periodic in the
measurement period chosen and, therefore, violates a prime requirement of
the FFT. Spectral estimates may consequently be modified by power leaking
from other frequency components. A nurober of undesirable properties are
exhibited, namely a broadening ofpeaks and the appearance ofside Iobes. In
a spectrum containing a nurober of closely packed frequency components,
leakage may smear tagether peaks and mask important detail. Leakage can
be significantly reduced by taper-windowing the sampled time-domain
record-that is, shaping it to become periodic in the measurement period
chosen.

Noise and Averaging

One of the major characteristics of any modal testing system is that
extraneous noise from a variety of sources is always measured along with the
desired excitation and response signals. By taking a nurober of power
spectrum averages, it may be shown 4 that the measured frequency response
function estimates more accurately the true frequency response function,
assuming the noise has a zero mean value and is uncorrelated to ihe
measured input signal.

The Coherence Function

To determine the quality of the frequency response function, it is not
sufficient to know only the relationship between input and output; 3 a major
question is the degree to which the system output is caused by the system
input. Noise and/or non-linear effects can cause large outputs at various
frequencies, thus introducing errors in estimating the frequency response
functions. The influence of noise and/or non-linearities, and thus the degree
of noise contamination in the frequency response function, is measured by
calculating the coherence function denoted by yl, where
2 Response power caused by input
y =-------------------------
Measured response power
The coherence function is easily calculable on a digital Fourier analyser
 Assessing Dynamic Properlies by Hammer Testing 599

(spectrum analyser) when frequency response functions are being evaluated.

As the coherence function indicates the degree of noise in a frequency
response function, it has two very important uses: first, it can be used
qualitatively to determine how much averaging is required to reduce
measurement noise; second, it can serve as a monitor on the quality of the
frequency response function measurements.

Illustrative Example
A typical hammer testing setup is shown in Fig. 3. The hammer has a force
transducer attached and is used to impact the structure. Excitation occurs
with a nearly constant force over a limited frequency range. The effect of
different types of hammer head is to alter this frequency range, as shown in
Fig. 4. A very soft head will concentrate a high density of energy in a narrow
Hamme and

"'"~''~" ~
~ Aeeelerometers
Iore

;I

Fo reeoutput
TEST
s
STRUCTURE Aecel eration
outpu t

l CHARGE
AMPLIFIER
I I
CHARGE
AMPLIFIER
I
L0 00 Analog uedata
TAPE (baeku p eopy)
RECORDER

..
LOW PASS Anti-ali asmg
FILTERS

L
DIGITAL FOURIER ANALYSER

•
Analogue to d1g1tal
convers1on

•
Appllealion of
w1ndows and FFT

•
Computation of Data dig1tized,
poNer/eross proc essed and
speetra 1 averages anal ysed

•
Computation of
frequeney response
tunellans

FIG. 3. Typical hammer testing setup.

600 J. R. M aguire

~ 10t-----~
g_
,g
8' ~
t1
01 1 - - - - - - - - - ' .

A0·01
20 200 2000
--+ Log. frequency: Hz

FIG. 4. Hammer autopower spectrum.

frequency range, whereas a very hard head will spread energy evenly across a
wide frequency range. Typically, a !arge sledgehammer with a rubber tip will
excite uniformly over a range of0-100 Hz, whereas the same sledgehammer
with a hard plastic tip will excite uniformly over 0-400 Hz. Details of the
sledgehammer used during the prototype testing are given in the Appendix.
If there is no mains supply near to the test structure, power required for the
setup shown in Fig. 3 may be provided by a small portable generator.

TESTS ON BRIDGE BEAMS AT BASINGSTOKE

Hammer tests were conducted 15 during November 1983 on four simply

supported precast post-tensioned concrete bridge beams at Basingstoke
(shown in Fig. 5). Each beam weighed approximately 40 t and effectively
spanned 27·6 m over timber (sleeper) supports. Prestressing was by three
tendons, two straight and one draped, designed for a total force of 573 t.
Each tendon (consisting of twelve 15·2 mm strands) was soundly grouted
into its own corrugated duct. These beams were left over at the end of the
Basingstoke ringway construction and were to be demolished as they were
excess to requirements. The first three modes ofvibration for the undamaged
beams were determined, and their frequencies and damping values are given
in Table 1. A limited number of mode shape measurements (at mid- and
quarter-span points) showed that these three modes corresponded to the
expected classical mode shapes for a simply supported beam. 20 Also given in
Table 1 are theoretical frequencies based on the concrete cross-section,
including the tendons (modular ratio = 15), which show close agreement
with the measured values. The dynamic Young's modulus for the concrete 15
was taken as 43 kN/mm 2 .
Hammer tests were also carried out during tendon exposure and cutting,
and results for beam 3 are presented in Fig. 6. The effect of darnage may be
 Assessing Dynamic Properlies by Hammer Testing 601

28200 J

-ir~ rA r
riß

Timber support
(600 wide)
lA l.s
ELEVATION

•r
17r
J 1000 J 1000
" r

•
215
650 1254 1254
1079

J
r 770
SECTION A-A
Jr

I Tendon
l • •
770
SECTION B-B

FIG. 5. Basingstoke beams (all dimensions in mm).

TABLE 1
Basingstoke beam tests (undamaged state)
---~---···-----------------

(a) Measured and theoretical frequencies (Hz)

f(l) f(2) f(3)

Beam 1 (measured) 4·32 14·88 27-48

Beam 2 (measured) 4·40 15·20 28·10
Beam 3 (measured) 4·28 15·12 28·52
Beam 4 (measured) 4-40 15·56 29·64
Beams 1-4 (theoretical) 4·06 16·26 36·60

(b) Measured damping (% critical)

c(l) c(2) c(3)

Beam 1 1·85 1·21 1·97

Beam 2 1·99 1-48 1·95
Beam 3 1·87 1·06 1·68
Beam 4 1·59 1-41 1-69
602 J. R. M aguire

x Measured value - - Leastsquares linear fit

-undamaged

N
J:

~
Q)
:::l
g- 4·22
u:

4 " 18o'-------o~-----1:":2:------~24
Strands
Tendons cut
exposed (12 strands = 1 tendon)
(a)

- Undamaged
15·1
N
J:
>-
(.)
c
Q)
:J
g- 14·9
u:

14·7
0 0 12 24
Strands
Tendons cut
Jr exposed Jr (12 Strands= 1 tendon)
(b)

28 .5 - Undamaged

N
J:
~
c
Q)

§. 28·3
['!
u.

28·1o'-------o'---S-tra_n_d_s_ __,12'------~24

Tendons cut
exposed
( 12 strands = 1 tendon)
(C)

FIG. 6. Change in frequency during tendon exposure and cutting: (a) mode 1; (b) mode 2; (c)
mode 3.
 Assessing Dynamic Properlies by Hammer Testing 603

seen to affect the frequencies in two different ways: first, the effect of exposing
the tendons (removing concrete and thereby removing stiffness) is to
decrease the frequency; second, the effect of cutting the tendons (reducing
axialload and thereby increasing stiffness) is to increase the frequency. It
may be seen from the above that integrity monitoring is possible using
hammer testing.

CONCLUSIONS

This paper has described the use of hammer testing to assess the dynamic
properties of existing bridge structures and has presented one case history. I t
has been found that, given sensitive instrumentation, hammer testing is able
to determine the structural dynamic characteristics ofbridge beams. It is felt
that hammer testing provides a quick and accurate method if assessing as-
built structural dynamic properties, and it is envisaged that this technique
could be successfully used on many existing bridge structures.

REFERENCES

1. SciENCE AND ENGINEERING RESEARCH CouNCIL, Lang-term R and D in Civil

Engineering. London, June 1982, Memo P:BG:123.
2. ELLIS, B. R. et al., Forced vibration tests and theoretical studies on dams. Proc.
Jnstn Civ. Engrs, Part 2, 69 (Sept. 1980) 605-34; 71 (June 1981) 575-95.
3. RAMSAY, K. A., Effective measurements for structural dynamics testing. Sound
and Vibration (Nov. 1975) 24-35.
4. RICHARDSON, M. and PoTTER, R., Identification of the modal properlies of an
elastic structure from measured transfer function data. 20th Int. Symp. on
Instrumentation, Albuquerque, New Mexico, May 1974, pp. 239-46.
5. WALGRAVE, S. C. and EHLBECK, J. M., Understanding modal analysis. American
Society of Automobile Engineering, West Coast Meeting, August 1978,
Technical Paper Series 780695.
6. KNAUER, C. D. et al., Space vehicle experimental modal definition using transfer
function techniques. American Society of Automobile Engineering, National
Aerospace Engineering and Manufacturing Meeting, Culver City, LA,
November 1975.
7. RICHARDSON, M. and KNISKERN, J., Identifying modes of !arge structures from
multiple input and response measurements. American Society of Automobile
Engineering, National Aerospace Engineering and Manufacturing Meeting,
San Diego, November-December 1976, Paper 760875.
8. CooLEY, J. W. and TUKEY, J. W., An algorithm for the machine calculation of
complex Fourier series. Math. Comput., 19(90) (1965) 297-301.
9. FouRIER, J. B. J., The Analytical Theory of Heat. Didot, Paris, 1822 (in French).
10. BENDAT, J. S. and PIERSOL, A. G., Random Data Analysis and Measurement
Procedures. John Wiley, New York, 1971.
604 J. R. M aguire

11. CLOUGH, R. W. and PENZIEN, J., Dynamics of Structures. McGraw-Hill, New

York, 1975.
12. RANDALL, R. B., Application of Band K Equipment to Frequency Analysis. Bruel
and Kjaer, Hounslow, 1977.
13. MEIROVITCH, L., Elements of Vibration Analysis. McGraw-Hill, New York,
1975.
14. BEAUCHAMP, K. and YEN, C., Digital Methods for Signal Analysis. Allen and
Unwin, London, 1979.
15. MAGUIRE, J. R., The dynamic characteristics of elevated piled tanks and other
selected prototype structures. PhD thesis, University of Bristol, May 1984.
16. MAGUIRE, J. R. et al., Assessing the dynamic properties and integrity of
structures by the use of transient data. Proc. 8th World Conf on Earthquake
Engineering, San Francisco, 21-28 July 1984.
17. BATHE, K. et al., SAPIV-a structural analysisprogram for static and dynamic
response of linear systems. Earthquake Engineering Research Centre,
University ofCalifornia, June 1973 (revised April1974). Report to the National
Science Foundation, No. EERC-73-11.
18. RICHART, F. E. et al., Vibrations of Soils and Foundations. Prentice-Hall,
Englewood Cliffs, New Jersey, 1970.
19. HousNER, G. W., The dynamical behaviour of water tanks. Bull. Seism. Soc.
Amer., 53(2) (Feb. 1963) 381-7.
20. BROCH, J. T., Application of Band K Equipment to Mechanical Vibration and
Shock Measurement. Bruel and Kjaer, Hounslow, 1972.

APPENDIX: SLEDGEHAMMER USED DURING BRIDGE BEAM

TESTING

The sledgehammer used during the bridge beam testing was a PCB type
(GK291B50) and is shown in Fig. Al. The sledgehammer mass was 5-4 kg,
providing a maximum impact force of 22 kN (typically 17 kN during
testing). The shape of the force pulse generated by the hammer was
approximately that of a half sine wave, of 5 ms duration when hitting a
concrete structure through a soft plastic tip. The sledgehammer required no
power supply, although the signal from the quartz force transducer was
amplified by a battery-operated charge amplifier supplied with the
sledgehammer.
 Assessing Dynamic Properlies by Hammer Testing 605

260 mm
Removable Spare
plaslic 1mpact tlp
tlp

PCB modei206M06
quartzforce
transducer

E
E
~

Output s1gnal
from cable

FIG. AI. Outlinc drawing of instrumented sledgehammer.

 53
Serviceability Performance of a Steel Highway
Bridge

I. ROSENTHAL and M. lTZKOVITCH

Faculty of Civil Engineering, Technion-Israel Institute of Technology,
Haifa 32000, Israel

ABSTRACT

The response behaviour of a steel single span highway bridge ( 45·60 m lang)
ll'as examined hecause ofcomplaints ofheat'Y vihration. The study consisted of
field measurements under conditions of regular and controlled motor traffic as
weil as under forced vibration induced by a vibration generator. The data
yie!ded the dynamic characteristics of the hridge. The peak acceleration
(0·25g), and vibration velocity ( 150mm/s), obtained were compared with
mrious code seruiceahi/ity requirements.

INTRODUCTION

The steel bridge over the Kishon River (leading to Haifa airfield) was
observed to vibrate heavily under motor traffic. In view of the almosttotal
absence of pedestrians, the problern was not one of discomfort but rather
whether or not this vibration could endanger the structure itself.
Accordingly, the local authority sponsored an investigation on the
Serviceability performance of the bridge, including various field
meas uremen ts.

THE BRIDGE AND TESTING PROGRAMME

The 15-year-old bridge is a 7 m wide (two-lane), 45·60 m long single span

composite structure comprising of six 2·50 m deep welded steel girders and a
607
608 l Rosenthai and M. Itzkovitch

0·20m r.c. deck; because of the small volume of pedestrian traffic a 1·50m
wide single sidewalk was included in the design (Fig. 1).
The test programme included field measurements using the instrumen-
tation described in Table 1 so as to ensure coverage of the entire possible
(and still unknown) range of vibration.
Three separate series were run:
Series A. Vibration measurements under regular traffic over 4 days,
mainly during the morning and afternoon rush hours.
Series B. Vibration measurements under a controlled regime, for which
purpose the bridge was closed to regular traffic for 7 h at night. The
regime consisted of running convoys comprising two 32-t trucks and a
62-t semitrailer in various sequences (the 126-t total being equivalent to
the design load) across the bridge at four specified speeds, 10, 30, 50 and
70km/h.

FrG. 1. The bridge viewed from its northeastern corner.

FrG. 2. The vibration generator mounted on the bridge deck.

 Serviceability Performance of a Steel Highway Bridge 609

TABLE 1
Measuring instruments ( installed on third girder) and their locations
(I= 45·60 m, length of bridge)

Measuring Distance from Instruments Range Recorder

point southern support channel

6·90 m (0·15/) Accelerometer 2g 9

Vertical geophone ±1mm 1
LVDT (southern riverbank) ±50mm 7
2 22·80 m (0· 50/) Accelerometer 2g 10
Vertical geophone ±3mm 2
Horizontal geophone ±3mm 3
Strain gauge ±2ÜÜÜflill 6
3 30· 30m (0·66/) Vertical geophone ±3mm 4
4 34·80 m (0· 76/) Accelerometer 2g 11
Vertical geophone ±lmm 5
LVDT (northern riverbank) ±50mm 8

Series C. Forced vibration (sweep sine) applied to the bridge immediately

following Series B, by means of a vibration generator attached to the
deck at measuring point 4, so as to yield both the first and second
vibration modes (Fig. 2).

RESULTS

All readings were recorded on magnetic tape and by a multichannel

recorder, as illustrated for Series A in Fig. 3.
Under large vibration all vertical geophones went off the scale but
operated again after passage of the vehicle in the vibration-free stage. This
yielded the fundamental frequency ofthe bridge (2·6 Hz) as well as its critical
damping ratio ( 1·1% ). The vertical velocities calculated from the
accelerations were throughout about ten times higher than those in the
horizontal direction.
The tensile strain and displacement plots over the river banks (channels
6-8) represented the static component due to the load, on which the
vibration induced by the inertia forces was superimposed. Their sum yielded
the maximum stress while the ratio of the dynamic component to the static
one yielded the impact of the moving vehicle or the dynamic increment. It
can be seen that although high Ievels of acceleration were measured the
tensile stresses in the girders were low. Table 2 lists some of the peak results
of Series A.
 .,..._.",Vfl<~«t::~ -1/~\Vt" •t.'f.W#MI" ·

lt ~ . ". ;;,

...... l, ,l~

~·-••uJ. . ,........ ' ....

"*"'• r n n
-...
,

.•.. . . . .
...,..

FIG. 3. Analogue recording of traffic vibration in Series A: heavy truck (left peak), semi-
trailer tanker (middle peak) and bus (right peak).
 TABLE 2
Peak midspan vibration results obtained in Series A

Vehicle Speed" Girder tensile stresses Impact Aceeieration Velocities Static

(km/h) (MPa) ratio (g) (mm/s) deflection c
(mm)
Static Dynamic Total
~
...,
VerticaJb Horizontal ~
R.
Heavy truck 84 17-4 2·9 20·3 0·17 0·14 84 11 14 :;::,
Semi-trailer, tanker 72 16·5 3·9 20·4 0·24 0·10 60 8 13 ~
:::::
Bus 84 3·8 1·9 5·7 0·50 0·08 48 5 3 ~

"Calculated according to traverse time from points 1 to 4. ~

S-,
Cl
h Calculated from acc.j2nj; with f= 2·6 Hz.
'Average value calculated from static displacement components measured at both river banks.
~;:::
~
'"'
TABLE
3
~
Main midspan vibration results obtained in Series B :;::,
V:l
Sequence Moving Speed" Girdcr tcnsile stresses Impact Aceeieration Velocities Static ~
.._
of Ioad (km/h) (MPa) ratio (g) (mmjs) deflectionc
vehicles (t) (mm) ::X::
Static Dynamic Total VerticaJh Horizontal i;;::
T+SM+T 126 10 18·5 1·0 19·5 0·05
~
0·06 36 3 21
T+SM+T 126 30 19·4 1·9 21·3 0·10 0·09
~
54 4 20
T 32 50 9·7 1·9 11·6 0·20 0·15
~~
90 7 6
SM+T 94 50 13·6 3·9 17·5 0·29 0·17 102 7 13
T 32 70 9·7 1·9 11·6 0·20 0·20 120 11 7
SM+T 94 70 11·6 3·9 15·5 0·34 0·25 150 12 13

Notes: See Table 2 above. 0\

,_.
T = truck; SM = semi-trailer.
 :--
1
~ ~~,~ ~~~WNNrlW.l
v • li.Z~

l.

FIG. 4. Analogue recording of vibration of convoy in Series B running at 10 km/h (left) and
30 km/h (right).
 • . u.• :C

rn rt• W4f4.
r..,.. 1 :~

FIG. 5. Analogue recording ofvibration of convoy in Series B running ar 50 km/ h (left) and
70 km j h (right).
614 1 Rosenthai and M. Itzkovitch

The traffic volume during rushhourswas about 1500 vehicles per hour.
The percentage ofheavy-duty vehicles (buses, trucks, etc.) was 15-25% in the
mornings and 8% in the afternoons; the remainder were small vehicles (vans,
private cars, etc.).
The main results of Series B are given in Table 3 and Figs 4 and 5,
indicating that the traffic vibration-directly proportional to speed-
reached acceleration Ievels as high as 0·25g. On the other hand, girder
stresses due to the moving Ioads again did not exceed 21 MPa.
At the lower speeds (10 and 30 km/h, Fig. 4) each convoy ran close
together, creating the effect of a single Ioad, while at the higher speeds (50 and
70 km/h, Fig. 5) the vehicles were spaced, creating the effect of two distinct
Ioads. The results in Table 3 are therefore arranged accordingly.
The Fourier spectrum of the accelerograms for points 1, 2 and 4 in the
70 km/h test is shown in Fig. 6. lt clearly yields the first and second

2.19
1.18

2.18
1.4i

1.18 Ull

.111 .na

.211 .388

21.1 21.8 Hz 3. i !.H 15.8 21.8 27.1 Hz

(b)

1.811

1.41
...... ;.

1.811

.lli

.211

27 .1 Hz

FIG. 6. Fourier amplitude spectra of accelcrations at points (a) 1, (b) 2 and (c) 4 from
70 km/h test (Series B).
 ~
...,
<:::
r;·
::::,
""<::J-
:::.:
~-

~
s.,
0...,
~
::::,
;:
;:;;
~
::::,
V)

--::r::~
i;t
~
~
~
""
FrG. 7. Accelerogra ms recorded at points I (top), 2 (middlc) and 4 (bottom) during sweep sine forced vibration test (Series C).

0\
lh
616 I Rosenthai and M. Itzkovitch

frequencies ofthe bridge, 2·6 and 9·0 Hz respectively, and indicates the third
frequency at around 19Hz. The amplitudes obtained at 2·6 Hz, lower at the
extremities of the bridge (points 1 and 4) and highest at midspan (point 2),
describe the first mode of vibration shaped as a half sine wave. The second
mode, at 9·0 Hz, has the form of a complete sine wave as amplitudesexist at
points 1 and 4 but none at point 2.
In Series C the bridge went into first resonance at 2·6 Hz, which is its
fundamental frequency according to the accelerograms at points 1, 2 and 4
(Fig. 7). The second resonance state, at 9Hz, could not be reached because of
the mechanical limitations of the generator (maximum rotational speed
380 rpm or 6·3 Hz).
Significan tl y, the three am pli tudes a t points 1, 2 and 4 a t 2·6Hz were all of the
same phase, with the peak at point 2 (midspan), indicating the half sine wave
ofthe first mode, while those at points 1 and 4 at 6·3 Hz had opposite phases
with an almost zero amplitude at point 2, indicating the complete sine wave
of the second mode.
With the first frequency established beyond doubt the stiffness of the
bridge was found from K = (2nf) 2 W/g = 76·6 kN/mm, with f = 2·6 Hz and
W = 282 t, the total dead load of the bridge.

DISCUSSION

The vibration components, measured in the present study under regular and
controlled traffic, exceeded all known relevant standards. Insofar as
dynamic behaviour is referred to, if at all, in highway bridge codes, no
detailed treatment is given, but only to footbridges, where the approach is
physiologicaU- 3 In some codes emphasis is on prevention of darnage in
structures, by limiting the velocity component to 20 or 30 mm/s in the
frequency range up to 30 Hz, 4 - 6 while the US Bureau of Mines criterion for
structural safety against darnage from blasting limits velocity to 50 mm/s up
to 3Hz, and acceleration to 0·10g in the range between 3 and 100 Hz. 7 By
contrast, the Ontario Bridge Code 8 allows for the Serviceabilitylimit state by
limiting the maximum static deflection due to factared highway live load
(including the dynamic load allowance) as a function of the first flexural
frequency and the anticipated degree of pedestrian use. For the bridge in
question, withf = 2·6 Hz and hardly any pedestrian traffic, this code would
Iimit the static deflection to about 23 mm. However, as the latter refers to a
standard loading vehicle, there is no common basis for comparison with the
actual 21 mm deflection of Series B, obtained with a convoy of vehicles.
In spite of the heavy vibration observed in the bridge (acceleration 0·25g
or velocity 150mm/s), the problern is not yet one ofstructural safety, due to
 Serviceability Performance of a Steel Highway Bridge 617

the fact that the girder stresses are low. The traffic causes 21 MPa including
the dynamic component and, tagether with 65 MPa from the dead load,
yield a maximum stress of 86 MPa, namely 60% only of the allowable level
for steel. As a result this case may be considered a fatigue problem,
particularly regarding the connections, which are mainly welds. 9
If the present situation continues-with vehicles running at 70 km/hat 2-s
intervals during most of the day (and with the already large traffic volume
likely to increase further in the future)--cumulative darnage is bound to
endanger the structure in the long run.

CONCLUSIONS

At first an attempt was made to reduce vibration by imposing a speed limit

of 40 km/h, but this proved unenforceable and the attempt was abandoned
after a trial period of 6 months. As an alternative design measures were
recommended for moderating the response ofthe bridge, mainly with the aid
of vibration dampers, 10 so as to improve the presently very low damping
ratio (1-1%).

REFERENCES

1. THE STANDARDS INSTITUTION OF ISRAEL, Loads on Bridges: Highway Bridges: IS

1227, Part 1, 1985.
2. BRITISH STANDARDS INSTITUTION, Steel, Concrete and Composite Bridges: BS
5400, 1978.
3. National Standard of Canada: CAN 3-S6-M78: Design of Highway Bridges,
1978.
4. GERMAN INSTITUTE FOR STANDARDS, DIN 4150: Vibration in Civil Engin-
eering-Part 3: Effects on Structures, May 1986.
5. INSTITUTION OF Swrss HIGHWAY ENGINEERS (VSS), Swiss Standard SN 640312:
Vibration Effects on Structures, 1978.
6. GOR CHAMBER OF TECHNOLOGY, Directive KOT 046/72: Effects of Blasting
Operations on Buildings, East Berlin, 1972.
7. ESHELMAN, R. L., Vibration standards. In Shock and Vibration Handbook, 2nd
edn, ed. Harris and Crede, Chapter 19 (Fig. 19.8, p.19-12). McGraw-Hill
International, New York, 1976.
8. ÜNTARIO MrNISTRY OF TRANSPORTATION AND CoMMUNICATIONS, Ontario
Highway Bridge Design Code, OHBDC, 1983 (updated 1985).
9. TILLY, G. P. (ed.), Dynamic behaviour of concrete structures-report of the
RILEM 65 MOB Committee, Chapter 5.1: Bridges. Elsevier, Amsterdam, 1986,
p.225.
10. JoNES, R. T. and PRETLOVE, A. J., Vibrationabsorbersand bridges. The Highway
Engineer, 26(1) (January 1979) 2~9.
 54

Monitoring of Traffic Induced Strain in the Steel

Reinforeerneut of a Concrete Bridge Deck

JOHN CAIRNS
Department of Civil Engineering, Heriot- Watt University, Edinburgh, UK

ABSTRACT

This report describes measurements undertaken on a steeljconcrete composite

bridge to determine the range of stress to which reinforcement in the deck slab
of the bridge is subject under vehicle loading. The study was prompted by the
proposal that afatigue classijicationfor unwelded reinforcement be included in
the Code of Practice for Fatigue. Measurements were undertaken with a
vehicle of known weight prior to opening to general traffic and under service
loading after 10 months of use. Measured strains were weil below the Ievel at
which fatigue darnage might occur, and were also below values that would be
calculatedfor design purposes. The infiuence of surfacing and of oscillations of
the bridge deck are discussed.

INTRODUCTION

In 1980 the Code of Practice for Fatigue, BS 5400: Part 10, 1 apart of the
Code of Practice for Bridges, was published. The bulk of its requirements
affected steel construction, for which cumulative darnage calculations
deemed to satisfy equivalent rules were specified. Although a classification
for butt-welded reinforeerneut was included in the detailed rules, normal
reinforeerneut was covered by a limiting stress range only. Subsequently the
intention to establish a full classification for unwelded reinforeerneut was
declared. 2 This paper describes measurements taken on a highway bridge
structure to examine the need for such a rule. Measurements were
considered necessary because of imprecise assumptions made in analysis in
respect of Ioad dispersion, impact factors, material properties, etc.
619
620 lohn Cairns

An earlier study in which an M-beam typebridgewas instrumented found

measured strains in deck slab reinforeerneut tobe considerably lower than
anticipated. 3

DESCRIPTION OF BRIDGE DECK

Milton Bridge No. 1 is a composite steel and concrete bridge on the M74 at
Lesmahagow, approximately 15 miles to the south of Glasgow. Two
independent bridges carry northbound and southbound carriageways over
the River Netham in three continuous spans of 47, 54 and 31m. Each
carriageway comprised hard shoulder, two traffic lanes and a 1m safety
strip. A section through the deck is shown in Fig. 1.

200

1
2·5m

3·5 m --::l<-1'-- 3·3 m

FJG. 1. Milton Bridge~section through deck.

The reinforced concrete deck slab of the bridge is supported on and acts
compositely with twin longitudinal plate girders and a central stringer
member, itself supported off the plate girders by cross-bracing spaced at
intervals of between 6 and 7 m along the length of the bridge. Depth of the
slab varies from 320 mm over the plate girders to 200 mm over the stringer.
The deckslabwas cast in five sections during August/September 1986 and
opened to traffic during October that year.
Transverse reinforeerneut is 20mm bars at 150mm centres in the bottom
and 32 mm diameter bars at 125 mm centres in the top of the slab.
Longitudinal reinforeerneut is 16mm diameter bars at 150mm centrestop
and bottom within the spans, increasing to bottom reinforeerneut of 25 mm
diameter bars at 150mm centres in hogging regions near piers. Total depth
of surfacing to the deck including waterproofing membrane was 120 mm.
Grade 37·5 structural concrete with 40 mm minimum cover to reinforeerneut
was specified for deck slab concrete.
 Traffic Induced Strain in Steel Reinforcement 621

INSTRUMENTATION OF BRIDGE DECK

Instrumentation to monitor reinforcement strains was installed during

construction of the deck slab. Two sets of gauges were installed within the
47 m end span of the bridge. One set, denoted Rl, was positioned 16·5 m
from the southern abutment in a region where the deck is subject to sagging
moments and the concrete deck slab will be maintained in longitudinal
compression. Gauge set R2 was positioned 2·7 m from the centre of pier B,
where loading will tend to place the deck slab in longitudinal tension. In both
cases gauges were approximately midway between cross-bracings, as the
greater ftexibility of support to the slab from the stringer away from the
cross-bracing was expected to give greater strains in reinforcement. Gauges
were centred 2·0 m from the centreline ofthe plate girder and 1·5 m from the
stringer, under the nearside lane. Preliminary calculations showed the
maximum stress range to occur with the nearside wheelline around 2·0 m
from the plate girder. BS 5400 rules for fatigue assessment require the
nearside wheelline to be between 1-4 and 2·0 m from the line of the plate
girder.
There were three gauge units within each ofthe two sets ofreinforcement
gauges R1 and R2. At both locations two gauge units measured strains of
longitudinal and transverse reinforcement in the bottom mat. The third
gauge unit at Rl gave a repeat measurement of bottom transverse
reinforcement strains.
At location R2 the third gauge unit measured strains in the top layer of
transverse reinforcement directly over the transverse gauge unit on the lower
mat. The position ofthe neutral axis ofthe slab for transverse bending could
be found from the results ofthe top and bottom transversepair to determine
whether compression membraneactionwas occurring and if surfacing had a
significant stiffening effect on deck behaviour.
Gauge units were made from a section of welded wire fabric to which
electrical resistance gauges were fixed (Fig. 2). The technique therefore did
not require any treatment or modification to the main reinforcement, nor
interfere with the contractors' programme. Gauge units were fully calibrated
in the Iabaratory prior to installation. Two full strain gauge bridges were
fixed to each unit, one set providing a back-up in case of damage. Values
measured on first and back-up gauges were in good agreement. Crack
ind ucers 40 mm high x 300 mm long were also installed directly under and
perpendicular to each gauge unit to persuade the slab to crack at gauge
locations. Strains measured on a gauge unit cast in a slab constructed and
tested in the Iabaratory showed satisfactory agreement with calculated
values and with strains measured on instrumented reinforcement.
After suitable conditioning, signals from the various gauges were
622 lohn Cairns

Brom dia

6 No transverse wires Brom dia

at 100m centres

FIG. 2. Detail of reinforeerneut gauge unit.

recorded to tape on a seven-channel portable instrumentation recorder for

analysis in the laboratory. Six channels recorded data from the six
reinforeerneut gauges. The seventh track was used for an audio recording
under the expansion joint at the southern abutment, and could be used to
trigger a logging sequence. Noise level recorded was equivalent to a strain of
8 X 10- 6 •

THEORETICAL CALCULATION OF DECKSTRAINS

Analysis of Deck
For design purposes, the behaviour of the deck slab of beam and slab
bridge decks under concentrated loads is usually split into local and global
components. Local moments arise from behaviour of the slab under
individual wheel loads between beams, global components arise from
overalllongitudinal and transverse fiexing of the whole deck under vehicle
loading.
A grillage model consisting oftwo longitudinal members continuousover
three spans and neglecting the central stringerwas used to calculate global
moments. To this were added the results of a local analysis using a finite
element model consisting of plate elements for the slab and offset beam
elements for the stringer. Cross-bracing was represented by beam or spring
elements, the stiffness ofwhich was determined from a substructure analysis.
Maximummoment in a deck slab under a concentrated load depends on
the area over which the concentrated load is applied. BS 5400 specifies wheel
loads and contact pressures for a standard fatigue vehicle, and the manner in
which the load may be assumed to disperse through surfacing and concrete
to the neutral axis ofthe deck slab. The finite element model could not model
load dispersion without refining the mesh to a much higher degree than
otherwise necessary. lt was therefore necessary to correct values calculated
 Traffic Induced Strain in Steel Reinforcement 623

by the finite element model for the deck by deducting an allowance for Ioads
applied to an area of finite size. The deduction was calculated as the
difference between moments calculated using Westergaard's charts 4 for a
point load on a thin slab and that foraloaded area determined by the BS
5400 rules and the appropriate spanjdepth ratio.
Only vehicle live Ioads were considered in the analysis, as it is only stress
range that is of importance for fatigue.

Reinforeerneut Strains
Strains in reinforeerneut were calculated for slab bending moments
determined from the structural analysis above. Analysis was carried out on
the assumption that concrete had no tensile strength, representing concrete
cracked in flexure, and on the assumption that the concrete was uncracked.
Poisson's ratio was taken as 0·2 for uncracked concrete and 0·0 for cracked
concrete. Strains were calculated at the centroid of the appropriate
reinforcement. Where concrete was to be considered uncracked, reinforce-
ment and concrete strains were assumed tobe the same. A linear variation of
strain through the depth of the section was assumed in both cases.

Vibrations of Bridge Deck

A further analysis of bridge deck behaviour was carried out to estimate
the natural frequency of the deck. The deck was modelled as a three-span
beam continuousover two internal supports. Two analyses were carried out,
with section properties based on the full composite section throughout the
three spans in one case and using values for steel only over supports (to allow
for cracking of the deck slab in these areas) in the other. Lowest frequencies
were determined as 2·20 and 1·74 Hz, respectively, bounding the measured
value of 2·0 Hz.

TEST PROGRAMME

Measurements
Four sets of measurements were undertaken: sets 1-3 used a measured
vehicle Ioad and were conducted prior to opening of the road to general
traffic, while set 4 readings were taken with the bridge in service.
Measurements taken were:

1. Static measurements prior to surfacing of the deck (Rl only), taken 20

days after completion of deck concreting.
2. Quasi-static measurements of surfaced deck, 42 days after completion
of deck concreting.
624 lohn Cairns

3. Dynamic measurements with vehicle travelling at approximately

30km/h (20mph) and 60km/h (40mph).
The Ioad vehicle used in 1-3 was a Leyland Atlantean double-deck
bus. Details are:
Wheelbase: 4·95 m
Front axle loading: 3·28 t, single tyres, 2·05 m c/c
Rear axle loading: 6·37 t, double tyres, 1·85 m cjc of pair
Tyre pressures: 0·7 N/mm 2 (lOOpsi) all round;
tyre size 1000 x 20.
4. Continuous monitoring of strains with bridge in service.

Analysis
Analysis of reinforcement strains was carried out using the procedure
described earlier for both the Ioad vehicle used in sections 1-3 and the
standard fatigue vehicle described in BS 5400: Part 10. 1

RESULTS

Load Tests 1-3: Measured Vehicle Loads

The maximum value of strain recorded in tension reinforcement as the
Ioad vehicle moved at walking pace across the deck along successive
transverse lines is shown in Fig. 3. Maximum values occurred with the rear
axle in line with respective gauges. Strains calculated analytically are also
shown in Fig. 3 for the Ioad vehicle at corresponding locations. The values
are for the slab in the uncracked condition.
Measured strains in longitudinal reinforcement at R2 near pier B were
around double the longitudinal values measured at R1 near midspan. It
appears likely that the concrete section near the pier was at least partially
cracked due to tensile stresses arising from deck surfacing loading and
restrained shrinkage ofthe deck slab. Loads from deck surfacing at midspan
gauges would tend to place the slab in compression and offset shrinkage
stresses.
Transverse strains measured at the two sets of gauges are in closer
agreement although differences may be noted, particularly with the
compressive strains recorded at R2 with wheelloads on the far side of the
stringer from the gauges. The difference may be attributable to increased
stiffness of support from the stringer to the deck slab from a slightly heavier
cross-bracing over the pier and from resistance to lateral movement of
bearings under the plate girder.
A comparison of strains in top and bottom transverse reinforcement at
R2 showed that the neutral axis ofthe slab lay around 120mm from the top
 Traffic Induced Strain in Steel Reinforcement 625

jWheels Vehicle Wheels

!

20

-5L-----~------~------~----~
0 10 2·0 3-a 4·0
Transverse Position (m.J
Longitudinal reinforcement

20,_-----,-------,-------,------,
0Rl
..,R 2 measured
~15r-----~------~r
I
DRl,R2 calculated
~

0
.s; 5
~

~
0

-5
0 10 2·0 3·0 4·0
Transverse Position (m.)
Transverse reinforcement

FIG. 3. Variation of strain with transverse position of Ioad vehicle.

surface, an x/h ratio of0·45. An x/h value ofjust under0·5 would be expected
for an uncracked slab. The measured value is in close enough agreement for
it to be concluded that the slab behaved as if uncracked transversely.
lt is considered unwise to attempt detailed comparison of measured and
theoretically calculated strains in view of the very low Ievels of strain. A
strain of 12 x 10- 6 , the largest measured value, is equivalent to a
reinforeerneut stress of only 2·5 N/mm 2 .
It was anticipated that surfacing of the deck might reduce strains on two
626 lohn Cairns

counts. Greater dispersion of local wheelloads through the thickness of the

surfacing will increase the effective contact area of the wheel and reduce its
intensity, thus reducing local moments; and composite action of concrete
slab and asphalt surfacing may create a slab of increased section modulus,
thus reducing measured strains under a given loading.
Lower strains were recorded in the surfaced slab, but the position of the
neutral axis was not altered. It follows that the surfacing did not alter the
flexural stiffness of the slab. Strains on transverse reinforeerneut were
approximately 35% less on the surfaced deck. Longitudinal strains,
however, showed a reduction of only 10%. Using Westergaard's method for
calculation ofmoments under concentrated Ioads and dispersion allowances
in BS 5400, the difference between surfaced and unsurfaced conditions
would not be expected to exceed 10%. Bearing in mind the possible
development of transverse cracking and its influence on longitudinal strains,
and the increase in modulus of elasticity of concrete between the two sets of
readings, it is difficult to draw significant conclusions from these results.
Studies on fatigue of steel bridge decks have shown that the dispersion
characteristics of the surfacing are temperature and rate of loading
dependent. No significant difference was found between quasi-static values
measured in series 2 and corresponding values measured with the vehicle at
speed in series 3.

Service Loadings
Results of measurements during service are summarised in the histograms
offrequency of occurrence ofstress rangein Fig. 4. Number and intensity of
the stressrangein longitudinalandin transverse reinforeerneut was similar
at both locations. It is apparent from Fig. 4 that measured strains are low.

300 300
~ ~

<.: <.:
~ ~

t-200 't-200
<:: <::
ClJ ClJ
::,
5-
"'
ClJ
Li:
ClJ
Li:
100 100

0 '----'----:-::---'=~---:-~
40 60
Strain (xi0- 6I
Longitudinal Transverse

FIG. 4. Histograms of measured strain range.

 Trajjic Induced Strain in Steel Reinforcement 627

The largest strain range reeorded was less than 60 x 10- 6 , equivalent to a
stress of 12·5 N/rnrn 2 • BS 5400: Part 10 requires that the range of strain to
whieh high yield reinforeerneut is subjeet should not exeeed 325 N/rnrn 2 .
Measured strains therefore did not exeeed 4% of this value.
Longitudinal reinforeerneut showed greater nurnbers of greater stress
ranges at both rnidspan and near the pier than did the eorresponding
transverse gauges. A detailed seetion of the eornplete reeord is presented in
Fig. 5, showing variations in strains as two artieulated vehicles with three-
axle trailers erossed the bridge at around 1·5 s apart. Only the variation in
strain is of signifieanee, as it was not possible to deterrnine true da turn values.

200

u:J
.M A ~Mtom ong i tudinal
- 150
""
~ ~\ )VV fJV \) \. ._. .,
I ßO ~ /\ )-o.Ä./\Bott m transverse
=
~
C'J
Top ransverse
50

'

Time ( seconds l
FIG. 5. Typical strains rccorded by reinforcement gauges R2 near pier B for thc passage of
threc-axle trailer articulated vehicles.

Traees for the longitudinal gauges clearly show strains ansmg frorn
longitudinal oscillations of the deck at a frequency of approxirnately 2Hz.
Subjectively, it had been noticed that the deck was 'bouncy' enough to have
an unsteadying effect while walking. The vibrations accounted for the rnuch
larger nurnber of events recorded for longitudinal gauges (Fig. 4), and
eonstructive reinforeerneut of the pattern of strains arising directly frorn
wheel Ioads and frorn oscillation rnay also aceount for the higher stress
ranges recorded overall on longitudinal gauges.
Readings frorn the pair of transverse reinforeerneut gauges at R2 were
eornpared for several of the larger events recorded. The rneasurernents
indieated a neutral axis depth of 85 rnrn, against 120 rnrn deterrnined frorn
rneasurernents 9 rnonths earlier. U sing an elastic section analysis for a
eracked slab assurning zero concrete tensile strength, the neutral axis was
estirnated at 65 rnrn frorn the top surfaee ofthe concrete. The rneasured value
of 85 rnrn would therefore be reasonable if either the concrete were assurned
to have a srnall tensile strength or if sorne eornpression rnernbrane behaviour
were occurnng.
628 lohn Cairns

TABLE 1
Calculated strain ranges for standard fatigue vehicle

Calculated strains ( x 10- 6 )

R1 R2

Transverse Longitudinal Transverse Longitudinal

Cracked section 250 130 250 130

Uncracked section 53 11 53 12

Strain range calculated for the BS 5400 standard fatigue vehicle by the
methods outlined in section 4 are listed in Table 1. Vehicles using the bridge
could (legally) have been up to 25% heavier in both overall and individual
axle loadings.
A comparison of Table 1 values with maximum values in Fig. 4 shows
transverse reinforcement strains were smaller than calculated, even if the
concrete slab were assumed tobe uncracked. Measurements at R2 made it
appear that the section was cracked. Measured values oflongitudinal strain
lie between those calculated for the slab in the cracked and uncracked
condition.

SUMMARY OF FINDINGS

1. Measured stress ranges were less than those calculated analytically.

2. Oscillations of the bridge deck served to increase the number and
intensity of measured stress range on longitudinal reinforcement.
3. Surfacing does not influence the stiffness of the deck but does assist in
dispersion of the contact load over a wider area.
4. Fatigue of deck slab reinforcement is unlikely to be a problern unless
accompanied by other deterioration.

ACKNOWLEDGEMENTS

Financial support from the Transport and Road Research Laboratory and
the Science and Engineering Research Council and the assistance of the
Director of Roads, Strathclyde Regional Council, and his staff is gratefully
acknowledged.
 Traffic Induced Strain in Steel Reinforcement 629

REFERENCES
1. BS 5400, Steel, Concrete and Composite Bridges: Part 10. Code of Practice for
Fatigue, British Standards Institution, London, 1980.
2. ScoTTISH DEVELOPMENT DEPARTMENT, Technical Memorandum (Bridges), SB
12/83, Edinburgh, 1983.
3. CAIRNS, J., Fatigue of bridge decks: in-service measurements in a reinforced
concrete bridge deck. Proc. 2nd Int. Conf on Short and Medium Span Bridges,
Ottawa, 1986. Canadian Society for Civil Engineering, Montreal, 1986.
4. WESTERGAARD, H. M., Computation of stresses in bridge slabs due to wheelloads.
Puhlic Roads, VII( 1) ( 1930) 1-23.
REPAIR ANDREHABILITATION
 55

Cracks in Steel Orthotropic Decks

PIERRE MEHUE
Service d'Etudes Techniques des Routes et Autoroutes, Bagneux, France

ABSTRACT

M any cracks have been discovered in steel orthotropic decks during the past 12
years. They occurred eilher at the rib-to-deck plate junction or at the rib-to-
floor beam junction, depending on local facts. Frequent inspections on a /arge
number of decks allowed observation of their appearance and progress in order
toset up a plan of action and repair. In the light offinding this darnage certain
measures can be recommended to ensure good structural behaviour of
orthotropic decks, such as using thicker plates and making edge preparation
compulsory.

INTRODUCTION

Orthotropic decks were introduced in France towards the end of the 1960s
on long-span highway bridges and on movable bridges because ofthe dead-
weight savings they made possible in the designs. lt is for the same reason
they were used on span units for the temporary flyover viaducts which were
erected in the early 1970s as a provisional solution to traffic problems at
many urban crossroads all over the country, with a total area of
approximately 120000m 2 from 1970 to 1976.
Several of these latter units, which carried heavy traffic, were found to
have cracked a few years after they were put in service, showing a repetitive
character in damage. Lately similar cracks have also been discovered in the
deck of a motorway bridge built in 1966. In both cases cracks occurred (1) at
the rib-to-deck plate junction and (2) at the rib-to-floor beam junction, just
in the zones where the wheels of trucks pass, which made people wonder
about the reliability of some construction details of such structures.
633
634 Pierre M ehue

MAIN CHARACTERISTICS OF BRIDGES

Temporary Flyover Viaducts

The span units for demountable viaducts, which rangein length from 6 to
30m for a single width of3·50 m and adepth of1 m, consist oftwo main plate
girders with an orthotropic deck employing closed trapezoidallongitudinal
ribs and transverse floor beams. They are joined end to end according to the
crossing conditions, and can be used separately for single-lane viaducts or
assembled side by side in order to form two- or three-lane viaducts with one-
or two-way traffic (Fig. 1).

TYPE I[
Skid resistant
thin surfacing

1 '90 J
3' 50 3,5 0

FIG. 1. Cross-sections of span units and location of cracks.

The span units are generally one of two types, which only differ basically
in the distance between girders, the number and size of ribs, and the depth of
floor beams. Type I structures are used most frequently. Both types have a
floor beam spacing of 3m and use a 10 mm thick deck plate with a skid-
resistant thin surfacing, and are entirely shop manufactured so as to be as
light as possible and easy to lift and handle. The ribs, which are made of
630 x 6 mm folded plates for type I and 570 x 7 mm for type II, are generally
continuous through the floor beams, except for some curve on plan span
units in which they are cut at the floor beams to be welded to the webs in
order to accommodate the curvature of the road. The rib-to-deck plate
junction is made by one-sided fillet welds. Steel grade A 52 S was used for all
the components, only the deck plate of the span units und er 20m long being
made of steel grade A 42 S.

Richemont Bridge
This is a three-span continuous 52·65~58·00~52·65 m bridge which was
built over the Moselle River in 1966 to carry a one-way two-lane roadway of
 Cracks in Steel Orthotropic Decks 635

12 03
'
10,0 0
2,00 7,00 1,00
3,50 3, 50
I
I I I I I II ®
.11 ~
® IIII II ~-
~ V
vv V vvvvvv vVVV ~

J loe
9, 50
660

FIG. 2. Cross-section of Richemont Bridge and location of cracks.

the A31 motorway with an overall clear width of 10m. lt consists of twin
plate girders 9· 50 m apart and approximately 2·80 m deep, and comprises an
orthotropic deck employing longitudinal closed ribs (Fig. 2).
The 12 mm thick deck plate is supported by trapezoidal ribs spaced 0·66 m
and floor beams spaced 3-625 m in the centre span and 3·51 in the two end
spans, with a 50 mm thick wearing surface. The ribs, which are made of
550 x 6 mm folded plates, are discontinuous and welded to the floor beam
webs by fillet welds, the connection with the deck plate being also made by
one-sided fillet welds. The structure is fully welded, using steel grade A 52 S
except for the deck plate, which is fabricated from steel grade A42 S.

CRACKS AT THE RIB-TO-DECK PLATE JUNCTION

The cracks were found either in the deck plate or in the fillet welds at the top
of the trapezoidal ribs, depending on the bridges and the structure
concerned.

Cracks in the Deck Plate

These cracks have been discovered only on span units for temporary
viaducts which were subjected to heavy traffic. They are exactly located
above the weid lines of ribs and affect sometimes one side sometimes both
sides, with a relatively symmetrical disposition on straight single-lane unit
spans (mark A). On two-way viaducts they are slightly shifted towards the
axis of the roadway (mark B), and the shift is moreevident towards the left
on one-way two- or three-lane viaducts, and towards the inner edge of the
deck on curved on plan units (mark C).
636 Pierre M ehue

Over 500 cracks were checked within the past 12 years, both on short and
long spans, between floor beams or right above them, in steel grade A 42 S
and in steel grade A 52 S, with no regularity in occurrence. Investigations
carried out on a sample of deck taken from a damaged unit showed that
these cracks resulted from fatigue phenomena initiated by local poor
welding ~nd Iack of penetration, due to no edge preparation and excessive
gap between rib and deck plate, which led to a notch effect in the root of the
fillet weld. So the cracksstartat the underface ofthe deck plate inside the rib,
where they cannot be seen, coming out under the road surfacing which is
quickly damaged. This obviously means that water can infiltrate into the
cracks and soon fill the ribs (Fig. 3). Undersuch circumstances these cracks

Deck plate

Fi llet weid

FIG. 3. Crack in deck plate.

are relatively easy to find, positioned at the bottarn of a scar or a crevice in

the road surfacing whenever they are over 40 mm long. Most of the time the
water drained into the rib maintains a permanent humidity between the lips
of the cracks and, even in summer, oozes at its lower end, accompanied in
very hot weather by a constant bubbling. The first cracks were found in 1977
on two viaducts erected in 1971 which supported a very heavy traffic Ioad
(800-1500 trucks per lane per day), but as the cracks were already long (420-
560 mm) it is likely that they were then at least 1 year old. Many other cracks
have been discovered later in the deck plate of many viaducts 8, 10 or 12
years after their construction, depending on the intensity ofloading imposed
on the structures. The way in which cracks progress from the moment they
have been detected is difficult to estimate because any accurate investigation
requires the traffic tobe stopped, which is politically difficult, particularly on
single-lane viaducts. However, it was possible on a few occasions to
undertake some measurements (as shown in Table 1) which gave three
examples of crack growth with quite different rates of extension.
The width of the crack is generally about 0·5 mm but it may reach 1 or
2 mm at the central part of long or old cracks, and mostly with roughly
bevelled edges. At the sametime a slight difference oflevel (1-4 mm) between
the cracks lips can be observed, due to deck plate deformation under heavy
 Cracks in Steel Orthotropic Decks 637

TABLE 1
Evolution of cracks in deck plate

Viaduct Cracks Length (mm) and date of measurement

17/2/81 29/4/81 1/6/81

-----

1 310 350 440

2 350 350 400
3 480 520 600
4 480 750 870
------ -~-- -------------·-----·--

22/6/82 6/10/83 24/5/84

- -----·---------------·-----

2 1 130 240 240

2 190 270 270
3 200 320 410
4 240 330 370
------ ~--~---

16/11/84 20/6/86 10/10/86

3 1 120 200 240

2 120 410 460
3 270 430 550
4 270 500 675
------------ ---- ---- -------·-----·----------

wheelloads (Fig. 4). Finding the first cracks created some concern with the
viaduct owners, but due to frequent inspections it was evident there was no
danger to users as long as the cracks were few in number, rather short and
fine, and grew at a slow gradual rate. Consequently it was generally possible
to wait for fair conditions to undertake repairs, after careful examination of
risks and possibilities. As a matter of fact the situation only became
worrying when separated successive cracks were growing towards each
other to form a single crack, Ionger and wider, or when several parallel
cracks were rapidly progressing in line and slit the deck plate into strips,
about 300 mm wide and over 1 m long, which were likely to sag considerably
under heavy Ioads or bow up with thermal effects. In those cases, fortunately
infrequent, trucks were obviously no Ionger allowed on the viaducts.
Lastly, attention must be drawn to the fact that, in spite of severe
inspections when the wearing surface was removed in order to renew the
surfacing, no cracks have been found in the 12 mm thick deck plates of four
long-span bridges 14 and 17 years after construction.

Cracks at Welds
In this case the cracks are located at the lower toe of the fillet weld at the
638 Pierre Mehue

FIG. 4. Crack in a !Omm thick deck plate.

junction ofthe rib and the deck plate, or very near the toe in the weid (Fig. 5).
Thesecracksare generally difficult to find because ofthe dim light under the
bridges, the dark colour of paint coatings and the poor state of finish, and also
because of difficult access to the bridge soffit. However, about two dozen cracks
were found on temporary overpass bridges, 8 years after erection for the first
ones, ranging in length from 40 to 350 mm and with an opening of less than
1 mm. Although only a few growth measurements were made, it seems that
cracks are able to extend rapidly, reaching 1·50 m over a 2-year period. For
example, the crack in the photograph shown in Fig. 6 was 300 mm lang on 8
J uly 1988 and 500 mm on next 12 September. In most cases cracks progress
by running along the weid toe, but it may happen that they propagate into
the rib wall following a curved line. As before, these cracks are likely due to
the thinness of the deck plate, which causes high transverse flexural stresses,
and to poor weid penetration, which makes the junction rather unsym-
metrical.

Fillet
weid
Rib

FIG. 5. Crack in fillet weid.

 Cracks in Steel Orthotropic Decks 639

FIG. 6. Crack at the toc of a fillet weid.

In 1984 a dozen similar cracks were observed on Richemont Bridge

(shown as mark A in Fig. 2). In 1987 fifteen cracks of varying lengths from
570 to 2240 mm had occurred in the bridge. Because a previous inspection in
1981 had revealed no damage, it would seem tobe evident that crack growth
progressed very quickly. The early investigations carried out to try to
determine the reason for these failures showed that the fillet welds were thin
and poorly made at many locations, and also that the ribs were very
economically designed.

CRACKS AT THE RIB-TO-FLOOR BEAM JUNCTION

The cracks were found with both continuous and discontinuous ribs, being
more widespread in the latter disposition.

Continuous Ribs
About a dozen cracks have been discovered on span units in temporary
viaducts within the past 8 years. Thesecracks were located at the toe of the
junction weid to the rib wall close to the ftoor beam web. The cracks propagated
at the lower end of the weid, on both sides, and were 30- 60 mm long when
they were discovered. The cracks generally progressed slowly towards the
deck plate andin only two cases were found to affect the whole length ofthe
weld (Fig. 7).
640 Pierre Mehue

FIG. 7. Crack at the continuous rib-to-ftoor beam junction.

Discontinuous Ribs
The cracks developed in the fillet welds joining the ribs positioned
between the webs of the floor cross-beams. They started at the lower end of
the round outline, at or very near the toe ofthe weld, close to the floor beam
web (Figs 8 and 9), then they gradually propagated into the rib-to-deck plate
and floor beam-to-deck plate welds. Most of the time they occurred on one
side ofthe rib and the floor beam, but in some cases they were discovered on
each face of the web, or on both sides of the ribs. In these latter cases a crack
rapidly appeared at the bottom weid, which made the end of the rib quite
free. Finally, it may happen that the cracks in the fillet weid escape from the
round outline, into the web of the floor beam, or into the bottom flange of
the rib.
About a hundred of these cracks have been found on span units of
temporary bridges, the first ones in 1978 with a viaduct erected in 1971 (Fig.
10). In 1987 ninety cracks were repaired on Richemont Bridge (shown as
mark B in Fig. 2) and twenty new cracks have appeared subsequently,
growing at a rate of 5-10 mm per month (Fig. 11). The causes oftbis darnage
may be due to both weld shrinkage and fatigue effect, together with the
rib discontinuity generating high residual stresses due to the welding
arrangement and severe stresses due to live Ioads, being transmitted through
poorly made fillet welds.

Rib

Floor beam
web

FIG. 8. Crack in the weid of discontinuous FIG. 9. Propagation of cracks in the welds
ribs. of discontinuous ribs.
 Cracks in Steel Orthotropic Decks 641

FIG. 10. Crack in the weid of a discontinuous rib of a temporary viaduct.

FIG. II . Crack in the weid of a discontinuous rib of Richemont Bridge.

642 Pierre M ehue

REFLECTIONS AND COMMENTS

The damaged bridge structures previously described could give the

impression that orthotropic steel bridge decks are unwise for bridge
designers to consider. So it is important to draw attention to the point that
darnage affects only a very low proportion ofwelds-for example the cracks
in the deck plate of span units for overpass bridges account for less than
0·2% of the whole length of the rib fillet welds. Moreover, as interesting as
they may be from a pathological angle, the cracks described above generally
neither damaged the structural integrity of the decks nor endaugered the
user's safety, and most of the time they can be easily repaired. It is also
important to mention that when the Richemont Bridge and the span units
for the temporary viaducts were designed, in 1965 and 1969 respectively,
information about the behaviour of orthotropic decks was limited. With
regard to the span units, which were to be used successively on several
viaducts, they were, contrary to widespread opinion, designed and
fabricated according to the same rules as permanent bridges, except for the
thickness of the deck plate. Due to early darnage in the wearing surface it
soon became apparent that the deck plate was at fault. At this time the deck
plate of orthotropic decks was always considered by the steel designers and
fabricators to be redundant.
Severallessons were drawn from this experience in the early 1980s, and it
was strongly advised that designers ensured that (1) the deck plate is not
thinner than 12 mm, a minimum thickness of 14 mm is recommended to
minimise defiections; (2) the edges ofthe plates forming the ribs are bevelled
in order to improve penetration of fillet welds; (3) the ribs are continuous,
crossing the fioor beams through appropriately shaped cutouts; and that
welding procedures are specified and monitored to ensure high quality
performance.
Although several bridges have been built since 1980 in compliance with
these specifications, it must be stated that most of the time fabricators are
reluctant to use a deck plate thicker than 12 mm and to bevel the rib edges.
They argue that there is no proofthat this work is absolutely necessary and
that, in addition, it will make the deck more expensive.
In conclusion, as there has been no darnage ofthe kind observed, on many
orthotropic decks built since 1971, it can be considered that the cracks
described above, which affect the two oldest orthotropic steel decks built in
France, are due to inappropriate technical details. Consequently there
should be no reason to be concerned about the reliability of well designed
and properly detailed orthotropic bridge decks.
 56
An Analysis of the Behaviour of Reinforced
Concrete Beams Following Deterioration and Repair

JOHN CAIRNS
Department of Civil Engineering, Heriot- Watt University,
Edinburgh EH14 4AS, UK

ABSTRACT

Results of a study of the redistribution of stress in reinforced concrete beams

following repair are presented. A non-linear analytical model which permits
removaland replacement of parts of a reinforced concrete section has been set
up, and is used to compare the performance of repaired and equivalent 'as new'
beams. The effects of relieving dead Ioads during a repair are considered.

INTRODUCTION

Corrosion of reinforeerneut is the main eause of deterioration of struetural

eonerete. In some instanees loss of seetion of the reinforeerneut requires
provision of additional or replaeement bars. At present little guidanee is
available to the engineer on the redistribution of stresses within a member
following repair, on the effeetiveness of replaeement reinforeement, and on
the struetural integrity of repaired members.
The aim of this paper is to report findings of an analytieal study of this
problem, in whieh relative deformations and stresses at seetions of 'as new'
and 'repaired' reinforeed eonerete beams were examined.
Where deterioration of a member is of a severity to require that additional
reinforeerneut be provided, eurrent praetiee is generally to prop the member
during the repair. 1
These temporary props may perform either or both of two funetions:
(a) to support a member in orderthat it ean eontinue to earry the loads
imposed despite loss of seetion and henee of strength, and
(b) to unstress a member while a repair is earried out.
643
644 lohn Cairns

The support role (a) may not be required if the member has adequate
reserves of strength, or if the loads the member is required to support can be
reduced for the repair period. Provided props are relatively stiff, it will be
adequate if they are initially placed in contact only with the member, and
take up load as the member deflects.
The object of unstressing a member (b) is to ensure that replacement or
supplementary reinforcement provided as part of the repair will assist in
carrying dead load. Unless dead loads are relieved by propping the member
during the repair and while the replacement concrete gains strength, new
reinforcement and concrete cast in during the repair will be unstressed when
the member carries dead load. For props tobe effective in this role they must
actively impose a load on the member to counteract dead loads. Hydraulic
jacks with pressure gauges to monitor applied prop forces will generally be
required.
Even under carefully controlled propping it is possible that the load may
not be fully removed from a member. The nature of the structure may be
such that the prop force is shared between the member under repair and
other parts ofthe structure. It will then be difficult to ensure that a member is
unstressed without resort to stress relief strain measurement techniques.
Cracks in reinforced concrete members may become filled with fine
detritus or crystalline growth which wedges a crack open and prevents
closure on unloading. 2 Reinforeerneut crossing such wedged crackswill then
be unable to shed load when prop forces are applied.
Concrete is not a linear elastic material. Loading and unloading
relationships are different, as is apparent from the short-term stress-strain
plot in Fig. 1. Creep will accentuate the difference between the loading and
unloading portions. Shrinkage of concrete also occurs over a period of time.
It is evident that even ifthe member can be unstressed by propping it cannot
be unstrained.
It becomes clear that even in the best controlled work similarity between
'as new' and 'repaired' members will be imperfect, and that any assumption
of structural integrity on this basis is questionable.
Is propping to counteract dead Ioads necessary for a repair? An analogy

Stress

FIG. 1. Short-term stress-strain relation-

ships for concrete loading and unloading. Strain
 Analysis of Reinforced Concrete Beams after Repair 645

with cornposite construction would suggest that although stresses and

deforrnations at service loads would differ, ultirnate strength would not. The
analogy is irnperfect, however, and there are differences between cornposite
construction and structural repairs that rnust be borne in rnind. For
exarnple, a precast concrete bearn supporting an in-situ cornposite slab will
be relatively young when the slab is cast, and differential shrinkage will be
less than that between the original concrete of a rnernber under repair and
the replacernent concrete, unless shrinkage-cornpensated repair rnaterials
are to be used.
This paper describes the first stage of a study to exarnine behaviour of
structurally repaired rnernbers. The terrn structurally repaired is intended to
denote repairs to rnernbers that have been significantly weakened by
deterioration. The repair area will generally belarger than patch size, defined
as 0·25 rn 2 .
A rnathernatical rnodel has been constructed to analyse stresses at a
section of a reinforced concrete bearn in a constant rnornent region at
various stages of its history, through long-terrn deforrnations, loss of section,
cutting out of concrete and casting in of supplernentary reinforcernent, etc.
The analyses reported here assurne the bearn tobe free of external restraints.
Results presented in the report relate only to cases of deterioration where
cornpression reinforeerneut is affected. Results of an analysis relating to
deterioration of tension reinforeerneut have previously been reported. 3 The
rnodel is being used to design test specirnens and procedures for physical
tests in the laboratory, results of which will be reported in due course.

GENERAL DESCRIPTION OF MATHEMATICAL MODEL

The rnathernatical rnodel is largely based on rnethods of analysis of

reinforced concrete sections outlined in BS 8110. 4 Two variations of the
rnodel are used, one for Serviceability and one for ultirnate behaviour. Both
consider only bending behaviour and are based on the assurnption that
plane sections rernain plane. It is assurned that an adequate anchorage
length is provided to supplernentary reinforeerneut to realise the assurnption
that plane sections rernain plane. Non-linear stress-strain relationships are
used for concrete and reinforeerneut The rnodels are based on a finite
elernent type approach. The rnodel allows parts of the section of a bearn or
colurnn rnernber to be rernoved, to represent spalling of concrete, loss of
reinforeerneut section due to corrosion or cutting out of contarninated
concrete, and subsequently to be reinstated, to represent replacernent or
supplernentary reinforeerneut and replacernent concrete.
The ultirnate strength analysis uses the reetangular parabolic stress-strain
646 John Cairns

0·67fcu

Ym Pöirölbolic
curve '

111
111

"
L.
+'
111

2·4 x 1 o- 4/'f; Str01in 0·0035

FIG. 2. Stress-strain curve for concrete. 4

curve for concrete of BS 8110 (reproduced in Fig. 2) and the bilinear stress-
strain curve for reinforeerneut (reproduced in Fig. 3). The tensile strength of
concrete is ignored. This model is used in calculations for reinforeerneut
stress at all loads and for all calculations at reinforeerneut stresses at or
above the characteristic yield strength.
Creep and shrinkage were taken into account by the use of an effective
modulus method and an equivalent force method respectively. Short-term
unloadingjreloading follows the slope of the initial tangent modulus up to
the short-term curve. Coefficients ofthermal expansion are fairly similar for
steel and many concretes, and changes in temperature will produce
negligible changes in stress in members that are not externally restrained.
Thermal strains have therefore been neglected in the analysis. Cracks in
reinforced concrete may become filled with fine detritus or crystalline
growth. 2 On unloading of the member, the cracks attempt to close but

Tension

200 kN/ mm 2
Str01in

Compression
1---....:,__ _---~..-- 1y I ym
FIG. 3. Stress-strain curve for reinforcement. 4
 Analysis of Reinforced Concrete Beams after Repair 647

k 11 2
f- O·Sf ( 1l - )
- CU 1 +(k-2)1]
E: E:
1l : E;;" : 0·0022
1·4 Ec.1 E 0 3E 0
k: :-
fCU' fcu
0·8 fcu

FIG. 4. Alternative stress-strain curve for concrete. 4

remain wedged open. The analysis can allow for this by effectively reducing
the width of an open crack.
The model for Serviceability behaviour differed from that for ultimate
behaviour in that a different stress-strain relationship was used, and
allowance was made for tension stiffening of the concrete. The stress-strain
relationship was again taken from BS 8110 and is reproduced in Fig. 4.
Where stress in the concrete at the Ievel of tension reinforeerneut would
exceed a value of 1·0 N/mm 2 in the short term or 0·55 N/mm 2 under Iang-
term loading, stresses in the tension zone were calculated on the assumption
of a triangular stress distribution varying from zero at the neutral axis to 1·0
or 0·55 N/mm 2 at the centroid of tension reinforeerneut in short and long
term respectively.
Chloride-contaminated concrete must be removed from around the
reinforeerneut before replacement concrete can be cast. A compression bar
not confined by surrounding concrete will be a slender compression
member, and its contribution to member strength will be limited by elastic
instability considerations. For simplicity the model assumes that com-
pression reinforeerneut carries no stress unless it is confined by concrete.

ASSESSMENT PROCEDURE

The performance of a 'repaired' beam with supplementary reinforeerneut is

assessed by comparison with the performance of an identical beam in an 'as
new' condition subjected to similar long-term loading. In making the
648 lohn Cairns

comparison it has been assumed that both original and replacement

concrete and reinforeerneut in the repaired beam have the same properties as
those in the original. It is also assumed that the repaired member will be
rebuilt to the same section profile as the old. These assumptions will
frequently not hold in practical situations, as a repaired member may need to
be of greater dimensions than the original where coverwas inadequate, and
replacement concrete will differ from original, often with reduced shrinkage
properties and greater strengths. These differences would improve the
performance of a repaired beam.
Service loads were calculated by first determining the ultimate bending
strength ofthe section under consideration, with allpartial safety factors on
material strengths included in calculations. Partial safety factors of 1·5 on
concrete strength and 1·15 on steel strength were used. This ultimate bending
strength was divided by a load factor of 1·5 to obtain the service bending
moment, on the assumption that dead and imposed loads are of similar
magnitude. Permanent loads used in the calculation of long-term
deformations were taken to be full dead load plus a proportion of imposed
load, usually taken as 25% in results presented here.
Partial safety factors were taken as 1·0 in all further calculations.
The procedure followed in analyses of 'as new' sections was:
(a) Apply permanent load, long term.
(b) Apply increment to full service load, short term.
(c) Increase load to 'first yield' of reinforeerneut
(d) Increase load to ultimate, taken to occur when the maximum
compressive strain in concrete reaches a value of 0·0035.
The procedure for analysis of 'repaired' sections was:
(e) Apply permanent load, long term.
(f) Allow loss of section from corrosion of reinforeerneut A 25%
generallass of reinforeerneut section along a bar has been assumed
throughout
(g) Allow for any crack wedging.
(h) Apply increment to full service load, short term.
(j) Reduce load to level at which repairs are conducted, short term. This
loadwas taken as full dead load, equivalent to 50% of service load,
for unpropped repairs and 10% of service load for propped repairs.
(k) Cut out concrete around bars.
(1) Install supplementary reinforeerneut and replacement concrete.
(m) Apply load increment to permanent load, long term.
(n) Apply load increment to full service load, short term.
(o) Increase load to first yield of reinforeerneut
(p) Increase load to ultimate.
 Analysis of Reinforced Concrete Beams after Repair 649

The assessment of the performance of beams by this analysis is based on a

limit-state approach. Bending strength is compared at the ultimate limit
state, stages (p) and (d) above for 'repaired' and 'as new' beams respectively.
Crack width and deflection are compared at service load, determined as
described above. Section curvature calculated in the analysis is used as the
basis for deflection, and mean strain in the extreme tension fibre of the
concrete is used for crack width comparisons.
In addition, the stiffness of sections as the load is increased from
permanent load to full service load are compared, based on the secant to the
moment/curvature relationship between (m) and (n) for the 'repaired' section
and between (a) and (b) for the 'as new' beam. This comparison is introduced
as a better measure of the deformability of a member during service than
total deflection.
Figure 5 shows a schematic stress~strain history for concrete at a point in
the compression zone of a repaired beam. The notation of Fig. 5 follows the
stages outlined above for a 'repaired' beam. Initial elastic deformation is
followed by creep and shrinkage, with little change in stress under sustained
application of Ioad. As load is briefly increased to full service Ioad, the
stress~strain plot follows a line parallel to the initial tangent modulus until it
reaches the short-term stress~strain curve, at which point it turns to follow
that curve. Unloading as imposed Ioad is removed and props are inserted
follows down a route parallel to the initial tangent modulus. Stresses may,
under certain circumstances, go negative, although strains will remain
positive. On reloading/removal of props, the line rises parallel to the initial

111
111
41 ~o~------------------P
...
I
L.
VI

Strain
FIG. 5. Stress-strain history for clement of concrete in compression zone.
650 lohn Cairns

tangent modulus until the short-term curve is reached, when that curve is
followed until ultimate load is reached.
An analysiswas carried out forareetangular beam section under different
environments and loading conditions. Creep and shrinkage strains are
infiuenced by the relative humidity ofthe environment to which the beam is
subject. While dead Ioads act permanently, the proportion of imposed load
which should be considered as permanent will depend on the nature of
the use of the structure. Each analysis wa!3 run for a member in 'as new'
condition and for a member repaired under propped or unpropped
conditions. The section used had 2·4% tension reinforeerneut and 0·9%
compression reinforeerneut when 'new'. Ratio of effective to overall depth of
section was 0·92. Corrosion was taken to reduce the area of compression
reinforeerneut by 25%. This loss was balanced by supplementary
reinforeerneut installed at repair. All reinforeerneut was taken to have a
characteristic strength of 460 N/mm 2 • Concrete was taken to have a
characteristic cube strength of 30 N/mm 2 . When analysed with partial safety
coefficients taken as unity, the neutral axis ofthe section at ultimate loadwas
at 42% ofthe effective depth. Concrete was taken tobe removed to a depth
of0·10 times overall section depth, equivalent to removal of 75 mm depth of
concrete from the compression zone of a beam of 750mm overall depth.
Results are presented in Tables 1-3 as the ratio of values for 'repaired'
beams, either propped to relieve dead Ioad or unpropped during repair, to
corresponding values for beams subject to similar long-term loading, creep
and shrinkage, but in otherwise 'as new' condition. Unpropped conditions
assume 50% of service Ioads carried during repair; propped conditions
assume 10% of service Ioad at repair.

TABLE 1
Ratio of calculated deformations and Ioads for 'repaired' versus 'as new' sections:
influence of permanent Ioad
Permanent Propped/ Ratio for 'repaired'j'as new'
Ioad unpropped
Service Ioad First Ultimate
yield
Curvature Crack Reinforce- Stiffness
width ment
stress

DL+25% IL p 1·13 1·04 1·04 1·01 0·98 1·00

UP 1·35 1·10 1·09 1·04 0·96 1·00
DL+75% IL p 1·12 1·04 1·05 1·02 0·98 1·00
UP 1-40 1·12 1·11 1·06 0·95 1·00

Environment: temperate.
No crack wedging.
 Analysis of Reinforced Concrete Beams after Repair 651

Table 1 shows results for beams subject to varying proportians of

permanent Ioad. Dead Ioad, taken to be 50% of service Ioad throughout
these analyses, is always in place, but the proportion of imposed Ioad that
may be considered tobe permanent will depend on the nature and use ofthe
structure. The analysis used a creep factor of 1·0 and a shrinkage strain of
-110 x 10- 6 , values appropriate to outdoor conditions in a temperate
climate. Cracks were assumed free to close on unloading.
At the Serviceability Iimit state, Table 1 shows that the curvature, and
hence the deflection, will be greater following repair, particularly ifthe beam
was not propped. However, it may be seen that repaired beams are slightly
stiffer as the Ioad is increased from permanent to full dead + imposed Ioad. If
the Ioad is subsequently cycled between these Iimits, repaired and new beams
will have the same stiffness. The greater curvature of the repaired beam is
therefore attributable to deformations of the member in the deteriorated
state and not to a reduction in stiffness following repair. If sag is likely tobe
of concern visually, steps could generally be taken to repair the member to a
flatter profile.
Crack width is closely related to reinforcement stress, and both are greater
in the repaired beams. In the past, crack widths have been regarded as of
importance for durability, but this is now questioned. 5 Given the random
nature of cracking, the calculated reduction of around 5% gained from
propping during repair would not be noticeable. The increase in
reinforcement stress at service Ioad in the repaired beams is brought about
by movement of the centre of compression away from the compression face
of the member, and consequent reduction of the Iever arm between tension
reinforcement and the centre of compression. The movement is greater in
unpropped repairs, where new concrete is less highly stressed. The
movement also accounts for the reduction in the Ioad at which tension
reinforcement starts to yield. Load at first yield still comfortably exceeded
service Ioad, however.
As the Ioad on the beam increases towards the ultimate and stresses reach
a plateau, new concrete attains the same stress as the old, although it is still
less highly strained. Table 1 shows ultimate bending strength to be
unaffected by the repair.
A similar comparison is made in Table 2 for beams subjected to different
climates. A creep factor of 2·0 and shrinkage of- 350 x 10- 6 have been used
in the calculations for members in a 'hot and dry' climate. These values are
also reasonable for indoor conditions in the UK. The performance of
repaired beams relative to 'as new' beams is slightly worse in the less humid
environment. The difference is principally due to shrinkage of concrete used
for reinstatement. As mentioned earlier, ordinary concrete was assumed for
repair although shrinkage-compensated mixes would often be specified.
652 John Cairns

TABLE 2
Ratio of calculated de.formations and Ioads for 'repaired' versus 'as new' sections:
influence of creep and shrinkage

Climate Propped/ Ratio for 'repaired'j'as new'

unpropped
Service Ioad First Ultimate
yield
Curvaturc Crack Reinforce- Stiffness
width ment
stress

T p 1·13 1·03 1·04 1·01 0·98 1·00

UP 1·35 1·11 1·09 1·04 0·96 1·00
Hand D p 1·18 1·03 1·06 1·05 0·96 1·00
UP 1·53 1·14 1·13 1·07 0·94 1·00

Permanent Ioad: DL + 25% IL.

No crack wedging.
T, temperate; H and D, hot and dry.

Table 3 shows the effect of cracks becoming blocked and prevented from
full closure. The crack wedging value in the table represents the proportion
ofthe width of a crack at stage (g) that has filled with other material. A value
of 100% denotes a crack that has completely filled; a value of 0% denotes
perfect closure ofthe crack. Tension reinforeerneut is unable torelease some
of the stress carried as prop forces are applied if cracks are wedged open.

TABLE 3
Ratio o.f calculated deformations and Ioads .for 'repaired' versus 'as new' sections:
influence o.f crack wedging

Crack Propped/ Ratio for 'repaired'j'as new'

wedging unpropped
(%) Service Ioad First Ultimate
yield
Curvature Crack Reinforce- Stiffness
width ment
stress

0 p 1-13 1·03 1·04 1·01 0·98 1·00

25 p 1·13 1·03 1·04 1·01 0·98 1·00
50 p 1·19 1·05 1·04 1·03 0·98 1·00
75 p 1·21 1·06 1·05 1·05 0·98 1·00
100 p 1·22 1·07 1·05 1·07 0·98 1·00
0 UP 1·35 1·11 1·09 1·04 0·96 1·00

Permanent Ioad: DL + 25% IL.

Environment: tcmperatc.
 Analysis of Reinforced Concrete Beams after Repair 653

Stress in tension reinforeerneut at repair was calculated to increase from

25 N/mm 2 where 'perfect' crack closure would occur to 112 N/mm 2 where
the crackwas fully wedged open. Ratios for the beam repaired unpropped
are shown in the table for comparison. Crack wedging had no effect on
'unpropped' repairs at values of crack wedging below 75%, under the
loading regime used. The effect of crack wedging is to reduce the effectiveness
of propping as a means of destressing a member for repair.

CONCLUSIONS

An analysis ofthe behaviour ofrepaired members suggests that the ultimate

bending capacity of a beam will not be reduced by failure to relieve dead load
on the beam. Total deflection will increase if dead load is not relieved
although the stiffness of the repaired beam does not reduce. The bending
moment at onset of yield of reinforeerneut will reduce and crack widths will
mcrease.
Experimental work is being carried out to verify the analytical results
reported here.

REFERENCES

1. LEwrs, D. A. and BoAM, K. J., Cathodic protection of reinforced concrete. Proc.

Conf on Deterioration and Repair of Reinforced Concrete in the Arabian Gulf,
October 1987, Vol. 1, pp. 79-98 and discussion; Vol. 2, 1988, pp. 38-9.
2. HoDGKIESS, T. and ARTHUR, P. D., Fatigue and corrosion effects in reinforced
concrete beams partially submerged in seawater and subjected to reverse
bending. Offshore Technology Report, OTH 87242, HMSO, London.
3. CAIRNS, J., Analysis of structurally repaired reinforced concrete beams. Proc.
Conf on Deterioration and Repair of Reinforced Concrete in the Arabian Gulf,
October 1989.
4. BS 8110, The Structural Use ofConcrete. British Standards Institution, London,
1985.
5. BEEBY, A. W., Corrosion of reinforcing steel in concrete and its relation to
cracking. The Structural Engineer, 56A(3) (March 1978) 77-81.
 57
An Investigation into the Effectiveness of Silane for
Reducing Corrosion Activity in a Chloride-
Contaminated Reinforced Concrete
Bridge Structure

GUY P. J. DILL
HAMMERSLEY, MICHAEL
Laing Technology Group Limited,
Page Street, Mi!! Hili, London NW7 2ER, UK
and
J. DARBY
JOHN
Oxfordshire County Council,
Speedweil Hause, Speedweil Street, Oxford OXJJ NE, UK

ABSTRACT

An investigation was undertaken an the A34 Wolvercote Viaduct into the

effectiveness of silane for reducing chloride-induced corrosion activity.
Concrete patch repairs were undertaken in areas of cracking and spalling on
two crossbeams, but visibly intact chloride-contaminated concrete was left in
place. The upper surface of the beams was waterproofed and the sides and
soffits treated with silane. Permanent Ag/AgC! reference electrodes and
resistivity electrodes were installed to monitor the performance of the remedial
measures. Preliminary results after 1 year show a generat small increase in
resistivity and reduction in electrode potentials, indicating that some drying out
of the concrete has occurred.

INTRODUCTION

In concrete bridge structures suffering from reinforcement corrosion

induced by the penetration of deicing salt, the removal and replacement of
all severely chloride-contaminated concrete is often difficult, structurally or
technically undesirable and economically unviable. Of the alternatives, one
possible approach is to repair only cracked or spalled concrete and leave
655
656 Guy P. Hammersley, Michael J. Dill and lohn J. Darby

visibly undamaged but chloride-contaminated concrete in place, relying on a

surface treatment to maintain the concrete at a sufficiently low moisture
content to minimise corrosion activity. The treatment used should not only
prevent the ingress of further water and chlorides but should permit water
vapour to pass out, hence allowing the concrete to dry out. Water-repellant
surface impregnations such as. silane have been promoted as having the
desired qualities, but there is a lack of published data on their use in this
particular remedial application.
A trial was undertaken 1 with the objective of investigating the
effectiveness of this remedial treatment for reducing corrosion activity in a
typical chloride-contaminated highway bridge structure. The trial com-
prised three stages:
1. Survey.
2. Remedia! work and installation of instruments.
3. Monitoring.

DESCRIPTION OF STRUCTURE

The structure chosen was the Wolvercote Viaduct, which carries the A34
Oxford Ring Road over the A40, the Oxford Canal, a British Rail mainline
and a stream. The viaduct, completed in 1962, consists of two separate
structures carrying the north- and southbound carriageways, each

FIG. 1. General view of the Wolvercote Viaduct.

 Effect of Silane in Reducing Corrosion 657

compnsmg 12 spans consisting of six precast post-tensioned beams

supporting a cast in-situ reinforced concrete deck slab. The spans are
supported by trestle piers (Fig. 1). The trial was carried out on the
crossbeams of piers 7 and 8 beneath the northbound carriageway.

SURVEY

The survey was undertaken during May 1988 and consisted ofthe following:
1. Visual inspection and hammer sounding.
2. Cover survey.
3. Half-cell potential mapping using a silver/silver chloride (SSCE)
reference electrode.
4. Sampling by drilling and analysis of the dust for chloride content.
5. Carbonation testing.
The survey revealed the local deep penetration of chloride ions in areas of
the beams subjected to ponding or run-down ofwater leaking down the deck
joint above. The original design detail consisted of a small dam around the
top of the beams with a number of short drainage spouts. Inadequate falls
and blocked spouts resulted in considerable ponding of salt-contaminated

Equipotential contour map prior toremedial work (contours in mV SSCE)

Equipotential contour map nine months <~fter remedial werk

o
Instrumentation and repair patches

Mort<1r patch ~Concrete eAg/AgCI 0 Resistivity

~
patch reference electrodes
electrodes

FIG. 2. South face of picr 8 crossbeam.

FIG. 3. Large breakout on the south face of pier 8 crossbeam showing localised pitting
corrosion of the reinforcement.

FIG. 4. Small spalls over corroding reinforcing bars with low cover. Note the location ofthe
drainage spout above the spalls.
 Effect of Silane in Reducing Corrosion 659

water on top ofthe beams. Water emitted from spouts ran down the sides of
the beams. In affected areas chloride contents in the medium to high
category of BRE Digest 264 2 were frequently present at the depth of the
reinforcement (30--60 mm). The chloride contents reduced with depth into
the concrete, confirming the external source.
Half-cell potential mapping (Fig. 2) showed a wide variation in electrode
potential around the beams with a number of localised 'high spots'
characterised by steep potential gradients and electrode potentials reaching
- 540 mV (SSCE). These 'high spots' could be correlated with areas of
medium to high chloride contents and subsequent breakouts revealed
pitting corrosion (Fig. 3). In areas of lower cover (less than about 20 mm)
corrosion inevitably resulted in spalling (Fig. 4). However, where the cover
was higher the pitting corrosion did not always result in surface damage, and
was only revealed by breakouts.
The original quality of the concrete was judged to be reasonable, as
evidenced by generally low carbonation depths of less than about 5 mm.

REMEDIAL WORK

The remedial work was undertaken between June and August 1988, and
consisted of the following principal operations:
1. Cutting out cracked and spalled concrete, and grit blasting reinforce-
ment.
2. Repairing the above areas with proprietary cementitious repair mortar
or concrete appropriate to the size of the breakout
3. Cutting away of the dam around the top of the beams andremoval of
the existing drainage spouts.
4. Steam cleaning of concrete surfaces.
5. Application of silane to the sides and soffits of the beams.
6. Application of an acrylic waterproof coating to the upper surface ofthe
beams.
7. Fixing flashings, gutters and downpipes around the top of the beams.
All the materials used complied with Department of Transport Standard
BD 27/86 3 or the Department of Transport Specification for Highway
Works, 4 as appropriate. The materials are detailed in the Appendix.

INSTRUMENTATION

In order to monitor the eflectiveness of the remedial measures, the

instruments were installed in the two beams during the course of the
660 Guy P. Hammersley, Michael J. Dill and lohn J. Darby

TABLE 1
Instruments installed in the beams

Host material Total

Original Repair Repair

concrete concrete mortar

Pier 7 8 7 8 7 8
Silverjsilver chloride reference electrodes 5 10 1 2 2 20
Resistivity electrodes 3 5 1 1 10

60mm
Cable to data
Iogger

Polyester putty plug

Ag/AgCI Reference
electrode 3=1 fine sand /cement mortar
+ 3"1• Cl- by mass of cement
FIG. 5. Detail of haif-eeil potential reference electrode.

50mm 50mm 50mm

[]
IV 0
0
0
0

H i lti HSA 303 M6 X 40 anchor Cable to data Iogger

FIG. 6. Detail of rcsistivity clcctrodcs.

 Effect of Silane in Reducing Corrosion 661

FIG. 7. Part of the south face of pier 8 crossbeam showing the !arge concrete repair patch,
new drainage guttcr and permanent instrumentation consisting of 2Ag/ AgCI reference
electrodes and one set of resistivity electrodes (covered by the plastic trunking).

remedial work (Table 1). In addition, the air temperature and relative
humidity were monitored in the vicinity of the beams.
The distribution of the instruments on the south face of pier 8 is shown in
Fig. 2 and details of their installation are illustrated in Figs 5 and 6. Figure 7
shows part of the south face of pier 8 on completion of the remedial works
and instrumentation.
The instruments were monitared by two battery-powered data Ioggers,
one being dedicated to each beam. The Ioggers were located in traffic counter
boxes on top of the viaduct on the narrow pavement between the parapet
fence and Armco barrier.

HALF-CELL POTENTIAL SURFACE MAPPING

In addition to the fixed instruments, it was decided to take available

opportunities to carry out a half-cell potential mapping exercise on the
south face of pier 8. As it was uncertain whether half-cell potential
measurements could be taken satisfactorily through a silane impregnated
surface, measurements were made on a fixed grid shortly before and after the
silane treatment. This revealed that the silane itself did not significantly
affect the measurement of haif-eeil potentials.
662 Guy P. Hammersley, Michael J. Dill and lohn J. Darby

-100
~
~ ~~~ l:3-f%J ~
ü ~~
"'
.::_
-200

:r
I

"_..-!Ii
>
E -300 ~~r~ trAt, ~

.,öc ~~ /
+-lo.t-
.!!0 -400 ~- ~
n.

-500
/r
0 -Jul 01-Sep 01-Nov 01-Jon 01-Mar 01-May 01-Jul
-600
0 100 200 300 400

Days alter start ol monitaring (Date)

0 P14 + P15 <> P19
"' P20

FIG. 8. Average weekly electrochemical potential measurements for selected electrodes

(pier 8).

~----------------------------------------------------,

-100
" "'
ü
~
-200
r-~+
~w~
~~
~
~
~;.
~
"'
-1: ~

> -300
E

.,öc
G -400
+'
0
n.

-500

0 -Jul
1 01-Sep 01-Nov 01-Jan 01-Mar 01-Moy 01-Jul
-600
0 100 200 300 400
Days oller stort ol monitaring (Date)
0 P12 + P13 <> P17
"' P18

FIG. 9. Average weekly elcctrochemical potential measurements for selected electrodes

(pier 8).
 Effect of Silane in Reducing Corrosion 663

20
19
18
17

;lv~\u
16
15

f\ f-J··. ~
14
E
u 13
~~ \Ii~ ~~
" E 12
~ ~i!J
"Q+ h
.1:
0 11 s_ <I>

"'
I

10

..•
~
:~
ii
9
8
~~·
j
I!J
Ir: • 7
6 .._,;""~~~
5 ~~ ~ "l
4 -0--+
3
~
2
1 0 -Jul 01-Sep 01-Nov 01-Jan 01-Mar 01-May 01-Jul
0~~----r-~---.--~--.---~*T------~------~-----,~~~

0 100 200 300 400

Days after start of monitaring (Date)

0 ~ + ~ 0 ~

FIG. 10. Average wcekly resistivity mcasurements for sclected electrodes (pier 8).

100

90

80

~
70

E
'0 60-
'E
~
I 50

V 40-
f
.il 30
e
G
c. 20
E
~
10

0
0 -Jul 01-Sep 01-Nov 01-Jan 01-Mar 01-May 01-Jul
-10
0 100 200 300 400

Doys after start of monitaring (Date)

0 Tempereture ('C) + 7. RH

FIG. 11. Average weekly temperature and relative humidity.

664 Guy P. Hammersley, Michael J. Dill and lohn J. Darby

PRELIMINARY EVALUATION OF RESULTS

Equipotential Contour Mapping

In those areas where patch repairs were undertaken the electrode
potential in the new patch materials, as might be expected, were appreciably
lower than in the original concrete replaced. This can be clearly seen in Fig. 2
by comparing the equipotential contour maps before and after the remedial
work. Electrode potentials have generally reduced slightly (become less
negative) since completion of the remedial work. Some areas immediately
adjacent to repair patches are now local potential 'high spots' and
consequently may become future anodic corrosion sites, as is often observed
adjacent to patch repairs in chloride-contaminated concrete.

Permanent Instrumentation
A preliminary evaluation of the permanent instruments after approxi-
mately 1 year from completion of the remedial work has revealed a
reduction in electrode potentialsfor all but one ofthe reference electrodes of
between 50 and 150mV. Electrode P12, installed in a !arge concrete repair
patch on pier 8, has reduced from - 560 to -150mV. Plots for individual
instruments from the south face of pier 8 are shown in Figs 8 and 9.
Two of the ten resistivity electrode sets have failed. The remaining
instruments have revealed a general increase in resistivity of about 2-4
kohm/cm (Fig. 10). The data recovered from the resistivity electrodes have
generally exhibited greater ftuctuations between readings than those
obtained from the silverjsilver chloride reference electrodes.
Temperature and relative humidity data are presented in Fig. 11.

DISCUSSION

The observed increase in resistivity is indicative of drying out of the concrete

and the general reduction in electrode potentials is probably a consequence
ofthis. Figures 8-11 show that the above effects appear tobe independent of
the surrounding air temperature and relative humidity with a similar trend
during winter as well as summer months.
Although the observed changes in resistivity and electrode potential have
not been !arge and locally there is still a high risk of active corrosion, the
preliminary results do at least indicate a trend in the right direction. lt was
anticipated that any significant drying out of the concrete would only be
observed in the long term (years rather than months) and it is too soon to
draw any conclusions on the success of the remedial measures in reducing
corrosion risk. However, it can be concluded that some drying out of the
 Effect of Silane in Reducing Corrosion 665

concrete has already occurred. As a spin-off, useful experience has also been
gained in monitaring the effects of full-scale concrete repairs.

ACKNOWLEDGEMENT

The Wolvercote Viaduct is owned by the Department ofTransport, who are

acknowledged for funding the trial.

REFERENCES

1. LAING TECHNOLOGY GROUP LIMITED, A34 Wolvercote Viaduct, Oxford.

Investigation into the effectiveness of silane for the treatment of chloride-
contaminated concrete-report on survey, remedial work and instrumentation.
Report ref. RD88/74/LS, October 1988.
2. ßUILDING RESEARCH ESTABLISHMENT, The durability of steel in concrete.
Diagnosis and assessment of corrosion-cracked concrete. BRE Digest 264, 1982.
3. DEPARTMENT OF TRANSPORT, Materials for the repair of concrete highway
structures. Departmental Stan~ard BD 27/86, 1986.
4. DEPARTMENT OF TRANSPORT, Specification for highway works, 1986.

APPENDIX

Contract Details
Agent: Oxfordshire County Council
Remedia! design, Supervision and monitaring
Consultant: Laing Technology Group Limited
Remedia! design, survey instrumentation and monitaring
Contractor: Hertfordshire Concrete Repairs Limited
Remedia! works

Materials
Replacement concrete (for !arge repairs)
SBD Five Star Repair Concrete
Repair mortar (for small repairs)
SBD Mulsifix Acrylic Repair Mortar DG351
Water repellant (for sides and soffit)
SBD Silane
Waterproofing (for upper surface)
Stirling Lloyd Eliminator
666 Guy P. Hammersley, Michael J. Dill and lohn J. Darby

Instruments
Half-cell reference electrodes
Ag/AgCI general-purpose reference electrode manufactured by
Silvion
Resistivity electrodes
Wenner electrode arrangement using 6 mm diameter Hilti HSA
303M6 x 40 stainless steel anchors
Environmental monitoring
SKH 103 temperature and humidity probe supplied by Campbell
Scientific and located in a Gill radiation shield
Data logging
Campbell Scientific CRlO data Ioggers
 58
Bridge Strengthening U sing Load Relieving
Techniques

BRIAN PRITCHARD
W S. Atkins Consulting Limited, Transportalion Engineering Division,
Woodcote Graue, Ashley Road, Epsom, Surrey KT18 5BW, UK

ABSTRACT

The paper describes several new techniques for strengthening existing bridges
to withstand increased loading by imposing dead Ioad relief or Ioad sharing.
The techniques cover external prestressing, the installation of extra shear
connectors and the use of shock Iransmission units. They all benefit from
requiring minimum, if any, traffic disruption.

INTRODUCTION

Strengthening of the world's bridge stock is a growth industry. This is

inevitable as the years pass because existing bridges are expected to carry
traffic of increasing loading and intensity for which they were not originally
designed. The same passage of time also means that existing bridges are
increasingly subjected to weakening environmental hazards, ranging from
winter de-icing salt to polluted atmosphere carbonation.
Strengthening of an existing bridge may become necessary because of
increasingly apparent overloading or because major repairs are required and
the opportunity is taken to strengthen the bridge to higher standards while
traffic restrictions are in operation. The traffic restriction aspect is usually
dominant and often precludes straight bridge replacement. It also strongly
inftuences the method of strengthening and those methods which involve
little or no traffic restriction are strongly favoured.
Three recent techniques for strengthening existing bridges by load relief or
load sharing with minimum, if any, traffic disruption are described.
667
668 Brian Pritchard

DECK BENDING RELIEF BY EXTERNAL PRESTRESSING

General
Conventional prestressing of a bridge deck imposes a permanent direct
compression together with a bending moment which counters, or relieves,
the applied dead Ioad moments. The two effects can be most beneficial to
tension-weak concrete decks and together they allow the prestressed
concrete bridge deck to carry further superimposed dead and live Ioad
moments without exceeding the permissible bending stresses or Ioad factors.
The bending moment reduction effect of added prestressing can also be
used to advantage in relieving dead Ioad bending in existing overloaded
decks of reinforced concrete, steel or composite concrete deck/steel girder
structures. This dead Ioad bending relief can be sufficient to reduce the deck
bending under full dead and live loading to permissible Iimits. Alternatively,
a bridge deck can be upgraded to carry increased superimposed dead and/
or live loading.
In general, the direct compression effect of the added prestressing is not
helpful. Reinforced concrete allowable compressive stresses are usually
lower than with prestressed concrete and extra compression in steel
structures can Iead to plate stability problems. lt is therefore beneficial to
mobilise as much of the prestressing bending moment reduction as possible
and there is every advantage in locating the prestressing tendons at the beam
extremities, or even beyond.

External Prestressing Applied to an Existing Composite Deck

Rakewood Viaduct carries the M62 motorway between Lancashire and
Yorkshire across a 36m deep valley (Fig. 1). The 256m long six-span

FlG. I. Rakewood Viaduct.

 Bridge Strengthening Using Load Relieving Techniques 669

45·7m

Pairs ot 50 dia.
Mac Alloy bars

36·0m

Anchorage detail

B.M. due to prestress

FIG. 2. External prestressing.

continuous deck, completed in 1969, consists of ten 3m deep steel plate

girders carrying and composite with an in-situ reinforced concrete deck slab.
DTp (NWRO) required upgrading to cater for a proposed increase in traffic
lanes carried and the more onerous requirements of the newly introduced
BS 5400 bridge code. The main shortfall was identified as an approximate
40% overloading in the steel girder compression flanges over the piers.
Upgrading by 'unloading', using external prestressing, was found to provide
an economical strengthening procedure with minimal disruptions on this
heavily trafficked motorway. Figures 2 and 3 indicate the strengthening

FlG. 3. Prestressing anchorages.

670 Brian Pritchard

procedure, which first requires the attachment of fabricated steel anchors to

the locally stiffened underside of each steel beam bottom flange by HSFG
bolting. Three pairs of 50 or 36 mm diameter Macalloy prestressing bars of
overlapping lengths are then attached under each flange between piers.
Upon stressing, hogging bending is set up in the midspan regions of the
beam. However, it is the parasitic sagging moment over the piers, caused by
deck continuity, which performs the required 'unloading' to acceptable
stress Iimits in the bottom girder flanges over the piers.
The prestressing anchorages apply high local loading to the in-service
deck girders. They are therefore attached at non-critical global bending and
shear locations where the additional local stresses can be readily
accommodated. The dispersion ofthese high anchorage Ioads into the girder
flanges and webs and the associated local design had been examined using
three-dimensional finite element techniques. Special consideration has also
been given to the provision of anti-corrosion protection and intermediate
supports to prevent wind vibration ofthe stressing bars. The completed deck
strengthening is shown in Fig. 4.
lt so happens that a similar deck unloading procedure is being applied to
an understrength three-span composite girder viaduct in Iowa State, USA,
this year. Prior experimental work on large-scale models has already been
undertaken and covered in several recent papers by Professor F. W. Klaiber
and his colleagues at Iowa State University. lt has been proposed that both
parties undertake and compare monitaring of prestressing bar Ioads during
and after construction.

FIG. 4. Strengthened Rakewood Viaduct.

 Bridge Strengthening Using Load Relieving Techniques 671

FIG. 5. External prestressing of reinforced concrete bridge beam.

Externat Prestressing Applied to a Reinforced Concrete Deck

Figure 5 shows how a similar external prestressing technique was used to
'unload' the reetangula r beams of an understrength two-span continuous
reinforced concrete deck in South Wales. In this case prestressing was by
cables located on the sides ofthe beams and anchored and deftected by steel
assernblies attached by epoxy grouted bolts passing through the beams.

SUBSTRUCTURE TRACTION AND BRAKING LOAD RELIEF

USING SHOCK TRANSMISSION UNITS (STUs)

General
A !arge number of our existing stock of viaducts feature long sequences
of simply supported deck spans, often carried on a series of high and
substantial piers. This is particularly evident in major river crossings where
high navigation clearances require long approach viaducts (Fig. 6). The piers
under each simply supported span inevitably carry fixed bearings for one
span alongside free bearings for the adjacent spa n. This means that the
design longitudinal traction a nd braking must be individually applied to
each deck span throughout the viaduct. M ai n resistance is offered by the pier
carrying the fixed bearings ofthat particular span.
Current integrity assessments of a number ofthese viaducts often indicate
that the piers are understrength due to increases in the deck longitudinal
672 Brian Pritchard

Typical navigable river or canal crossing

Free bearing Deck traction or braking Ioad

45 tonnes +-----*
Fixed bearing
.........
Most of 45 tonnes
must b.e applied to
each p1er 1n turn v1a
the f1xed bearing

-
Enlarged detail of approach span
45tonnes

.__...
45 tonnes now greatly
reduced by m~ltl-span
local shar1ng 1nduced
by STUs

Approach spans with STUs

FIG. 6. Substructure strengthening using STUs.

loading since the original design, sometimes accompanied by darnage

generated by road salt, carbonation or ASR. A substructure of this type
with, say, ten equal height piers has a total resistance capacity of
approximately ten times the original deck design traction and braking
longitudinal Ioads. This total resistance capacity can be mobilised by
providing Ioad transfer STUs across the deck joints.

Shock Transmission Units (STUs)

STUs are mechanisms which are connected across movement JOmts
between structural elements. They transmit slow actingjoint movements like
temperature and shrinkage with negligible resistance and, when required,
transmit momentary impact forces like traction, braking and earthquake
with negligible movement.
A simple, economical and minimum maintenance bridge STU was
developed in the UK some years ago. Instead of oil the STU utilises the
peculiar properties of 'bouncing putty', a silicone compound which will
readily deform under slow pressure but becomes rigid under impact. The
unit consists of a steel cylinder containing a loose-fitting piston fixed to a
transmission rod, the void round the piston being filled with the silicone
putty. Under slow movement this putty is squeezed around the piston and
displaced from one end of the cylinder to the other (Fig. 8).
 Bridge Strengthening Using Load Relieving Techniques 673

I \
·t
'r
...

FIG. 7. Docklands Light Railway strengthening and Canary Wharf building in progress.

Load Relief Using STUs for Viaduct Piers of the London Docklands
Light Railway
The newly completed viaducts carrying London's Docklands Light
Railway (Fig. 7) were designed for a train service which, due to a breath-
taking increase in adjacent development at Canary Wharf, will now
require considerable expansion before 1990. This will mean heavier and

7 continuous spans , 7 continuous spans 7 continuous spans

STUs STUs

~~~~~~~~~~;A~II~intlrm;;iateli~er~s~~~~~~~~~~~
have Iaminated rubber
--'t'------,1--4-L-------'""'"'b::..:e=arings 'I'
Siiding bearings at
expansion joints (a)

Boron filled bi- methyl syloxane putty

Holding down bolts

in diamond drilled
hol es
(b)

F1u. 8. Detail of STU and mounting.

674 Brian Pritchard

morefrequent trains, which will add braking and traction effects in excess of
those originally catered for.
Figure 8 shows a typical as-built seven-span deck unit continuous
between expansion joints. Train traction and braking Ioads are currently
shared among the slender piers, which generally support the deck via rubber
bearings. STUs are being installed at raillevel betweenjoints such that, when
the new increased longitudinal traction and braking loading is applied to
one particular seven-span unit, load is beneficially transmitted and shared
with adjacent seven-span decks sufficient to require no pier and foundation
strengthening in any substructure. This simple procedure represents a
tremendous saving in cost and interference with the existing train service.

COMPOSITE DECK FATIGUE RELIEF USING ADDITIONAL

SHEAR CONNECTORS

General
Existing composite bridge decks, like the two described earlier, often
require strengthening or fatigue life enhancement of the shear connection
between the concrete deck slab and steel girders. This can be undertaken by
installing additional new shear connectors, ideally from the underside ofthe
slab/girder interface to minimise traffic interference.

Fatigue Life Enhancement of Viaduct Decks of the London Docklands

Light Railway
The existing new viaduct decks, completed in 1987, are generally of
continuous composite construction with an in-situ reinforced concrete deck
slab supported by and composite with twin steel universal or plate girders.
The original design of the decks to BS 5400 established that fatigue
considerations were a critical factor, particularly in the deck shear
connectors. As a result ofthe increase in weight and frequency oftrains after
1990, mentioned earlier, the fatigue life would suffer considerable reduction.
Strengthening measures to restore the fatigue life back to the originally
designed 120 years were required. Additional shear connectors installed
between the original 19 mm welded stud connectors would relieve the Ioads
on these connectors sufficiently to accomplish this.
Technically, the easiest solution was to drill out holes through the deck
slabs from above and install new stud connectors, in clusters ofthree or four,
to optimise the hole size and economies of installation. Practically, this
would be extremely disruptive and costly to undertake with a live service in
operation.
The provision of new shear connectors by drilling-in from under the top
 Bridge Strengthening Using Load Relieving Techniques 675

FIG. 9. 20mm spring steel pins.

ftange of the steel girders was also examined. Several types of connectors
were considered, including 20 mm diameterspring steel pin fasteners. These
offered the advantage of a readily achieved force fit into the hole drilled
through the steel ftange and lower section of the concrete deck slab with no
requirement for grouting, glueing or welding.
The benefit of causing no interference to the train service had to be
balanced against the unknown shear and fatigue parameters of such a
previously untried system. Strength and fatigue testing were carried out on
push-out samples by the Welding Institute at Cambridge. Two types of
20 mm spring pin were shown to have superior strength and fatigue
properties to the 19 mm studs. These are shown in Fig. 9.

FIG. 10. Jacking-in 'Spirol' spring pins.

676 Brian Pritchard

Both pins obtain their forcefit by jacking the lead-in chamfer into drilled
holes with slightly smaller diameters (Fig. 10). In one, the pin slot closes as
the pin diameter reduces during driving. In the other, the spring mechanism
is generated by the compression of a 2! turn spirally coiled strip of steel.
Good interface shear connection is established with a degree of pull-out
resistance afforded by the spring-loaded friction between the pin and the
hole face. In the event, the spiral pins were successfully installed with no
interruption to the train services.
 59
The Integrated Construction and Conversion of
Single and Multiple Span Bridges

MARTIN P. BURKE JR
Burgess & Niple Limited, 5085 Reed Road, Columbus, Ohio 43220, USA

ABSTRACT

In the United States and Canada, integrated bridge construction is becoming

one of the bridge engineer's primary responses to Joint related bridge darnage
caused by the use of deicing chemieals and the restrained growth of rigid
pavements. The relative success that has been experienced with integral
bridges-bridges without deck joints-is now being reflected, not only in the
increasing numbers of Ionger integral bridges but also in the integral
conversion of existing jointed bridges. As revealed in a recent survey made
during the development of the National Cooperative Highway Research
Program's Synthesis 'Bridge Deck Joints', the use of integral conversion
techniquesfär existing jointed bridges is generafing widespread interest but as
yet only a moderate number of'tentative applications. It appears that the initial
success ofsuch techniques lvill see an accelerated use of integrated conversion
as an effective alternative to bridge Joint rehabilitation.

INTRODUCTION

Integral bridge construction may be defined as the practice of constructing

bridges without deck joints. When using such construction to eliminate
intermediate joints in multiple span bridges, it is accepted that the continuity
achieved by such construction will subject superstructures to secondary
stresses. Thesestressesare due to the response of continuous superstructures
tothermal and maisture changes, and gradients, settlement of substructures,
post-tensioning, etc. When such construction is used to eliminate deckjoints
at abutments, it is likewise accepted that such structures will, in addition, be
677
678 Martin P. Burke Jr

subjected to secondary stresses due to restraint provided by abutment

foundations and backfill against the cyclic movement of bridge super-
structures. The justification for such construction is based on the recognition
that for small and medium span bridges of moderate lengths significantly
moredarnage and distress has been caused by the use of deck joints than by
the secondary stresses they were intended to prevent. In addition,
elimination of costly joints and bearings, and the details and procedures
necessary to permit their use, generally result in more economical bridges.
Consequently, more bridge engineers are now willing to relinquish some of
their control of secondary stresses, primarily to achieve simpler and less
expensive bridges with greater overall integrity and durability.

CONTINUOUS SUPERSTRUCTURES

Current design trends received their primary impetus and direction almost
six decades ago. In May 1930, a brief ten-page paper published in the
Proceedings of the American Society of Civil Engineers generated
considerable discussion in academia. lt also created a minor revolution in
the design and construction of small and medium span bridges. The paper,
'Analysis of continuous frames by distributing fixed-end moments' by Hardy
Cross/ presented a simple and quick method for the analysis ofintegral type
structures such as continuous beams and frames. The method was quickly
adopted by bridge engineers and the bridge practice ofmany transportation
departments began to change. Prior to Hardy Cross' moment distribution,
multiple span bridges were generally constructed as a series of simple spans.
F ollowing the introduction of moment distribution, bridge engineers began
eliminating troublesome deck joints at piers by providing continuous
superstructures.
Line A in Fig. 1 shows the beginning of the routine use of continuous
construction in the United States and Canada, and the per decade increase in
the number of transportation departments that have adopted the use of
continuous construction. As shown, 26 of 30 departments responding to a
recent mail survey, 2 or 87% of responding departments, now routinely use
continuous construction for short and medium span bridges.
Currently the state of Tennessee appears to be leading the way in
constructing long continuous bridges. For example, the Long Island Bridge
at Kingsport, Tennessee, was constructed in 1980 with 29 continuous spans
without a singleintermediate transverse joint. The totallength of this bridge
is about 2700 ft center to center of abutment bearings. Deck joints and
movable bearings have been furnished, but only at the two abutments. It has
been aptly named 'The Champ'.
 Construction and Conversion of Span Bridges 679

B Early use of mtegra I abutments

C Early conversion of
multiple s1mple spans
to contmuous spans

--+--
~
~

0
0
c
0
0
lL f--
0
0 1920 1930 1940 1950 1960 1970 1980 1990
z
YEAR
FrG. I. Design trends for continuous bridges.

Approximately 50% of transportation departments in the United States

and Canada Iimit maximum bridge length without intermediate transverse
deck joints to 600ft or less. However, almost half of the departments
responding to a recent mail survey 2 allow the unjointed bridge length to
exceed 1000 ft.

INTEGRAL BRIDGES

During the past 2-3 decades, many bridge engineers have become acutely
aware of the relative performance of bridges built with deck joints at
abutments and those built without them. In most respects, bridges without
joints, integral bridges, have performed more effectively since they remain in
service for Ionger periods of time with only moderate maintenance and
occasional repairs. Some of this experience was forced upon bridge
engineers by circumstances beyond their control.
Due to the growth and pressure generated by jointed rigid pavements,
many bridges built with deck joints at abutments have been and are being
severely damaged. Many abutment backwalls have been fractured. Other
abutments have been split from top to bottom. In Ionger bridges with
intermediate deck joints, piers have been cracked and fractured as weiL
In geographical areas with low seasonal temperatures and an abundance
of snow and freezing rain, the use of de-icing chemieals to maintain dry
pavements throughout the winter season has also had a significantly adverse
effect on the durability and integrity of bridges built with deck joints. Open
joints and sliding plate joints of shorter bridges and open finger joints of
Ionger bridges have allowed deck drainage, contaminated with deicing
680 Martin P. Burke Jr

chemicals, to penetrate below deck surfaces and wash over supporting

beams, bearings and bridge seats. The resulting corrosion and deterioration
have been so serious that some bridges have collapsed while others have had
tobe closed to traffic to prevent their collapse. Many bridges have required
extensive repair; and most of the bridges that have remained in service have
req uired almost continuous maintenance to counteract the adverse effects of
these chemicals. To help minimize or eliminate these corrective efforts, a
whole new industry was created.
Beginning in the early 1960s, the first elastomeric compression seals were
installed in bridges in the United States to sealdeck joints. Since these first
installations, numerous types of elastomeric joint seals have been developed
and improved in an attempt to achieve ajoint seal design that would be both
effective and durable. Mostdesigns have been disappointing. Many leaked.
Some required more maintenance than the original bridge built without
seals. By and large, the many disappointments associated with various types
of seals have caused bridge engineers to consider other options.
Costs of various types of bridges showed marked differences. For two
bridges built essentially the same, except with one provided with separate
abutments and deck joints and the other provided with integral abutments,
the jointed bridge was usually more expensive. In addition, bridges with
integral abutments suffered only minor darnage from pavement pressure,
were essentially unaffected by de-icing chemicals, and functioned for
extended periods without appreciable maintenance or repair. Consequently,
more bridge engineers began to appreciate the merits of integral bridges for
short tomoderate bridge lengths. Gradually design changes were made and
Ionger integral bridges were built and evaluated. In 1946, Ohio's initial
length Iimitation for its standard continuous concrete slab bridges was
175ft. In a 1973 study of integral construction, 3 four states responded that
they were using steel bridges and 15 states were using concrete bridges in the
201-300-ft range. In a 1982 study of integral abutment bridges, 4 even Ionger
bridges are reported:
Continuous steel bridges with integral abutments have performed
successfully for years in the 300-ftrangein such states as North Dakota,
South Dakota, and Tennessee. Continuous concrete structures 500 to
600 feet long with integral abutments have been constructed in Kansas,
California, Colorado, and Tennessee.
Currently 11 states are building continuous bridges with integral
abutments with lengths in the 300-ft range. Missouri and Tennessee report
even Ionger lengths. Missouri reports steel and concrete bridges in lengths of
500 and 600ft, respectively, while Tennessee reports lengths of 400 and 800ft
for similar bridges. Finally, line B of Fig. 1 shows that 20 of 30
 Construction and Conversion of Span Bridges 681

transportation departments, or 67% of survey responding departments, are

now using integral construction for continuous bridges.
The attributes of integral bridges have not been achieved without cost.
Parts ofthese bridges operate at very high stresses, stressesthat cannot easily
be quantified. These stresses are significantly above those permitted by
current design specifications. In this respect, bridge engineers have become
rather pragmatic. They would rather build eheaper integral bridges and
talerate these higher stresses than build the more expensive jointed bridges
with their vulnerability to destructive pavement pressures and deicing
chemical deterioration. In a 1985 magazine article, 5 Clellon Loveall, then
engineering director for the Tennessee Department of Transportation,
reflects this attitude when he writes:
In Tennessee DOT, a structural engineer can measure his ability by seeing
how long a bridge he can design without inserting an expansion joint. ...
Nearly all our newer (last 20 years) highway bridges up to several hundred
feet have been designed with no joints, even at the abutments. If the
structure is exceptionally long, we include joints at the abutment but only
there .... Jointsand bearings are costly to buy and install. Eventually they
are likely to allow waterandsalt to leak down onto the superstructure and
pier caps below. Many of our most costly maintenance problems
originated with leaky joints. So we go to great lengths to minimize them.
Even though bridge engineers have conditioned themselves to talerate
higher stress Ievels in integral bridges, occasionally their design control is not
sufficient to prevent these high stresses from resulting in structural distress
and structural fracture.

STRUCTURAL DISTRESS

Responses to an early survey about the construction of continuous bridges

with integral abutments indicated a rather widespread concern by bridge
engineers for the potentially high stresses that would be present in Ionger
bridges. 3 This concern, more than any other, appeared to be responsible for
the early Iack of enthusiasm for using integral abutments for Ionger
continuous bridges (line A, Fig. 1). Although the great majority of bridges
with integral abutments perform adequately, many of them operate at high
stress Ievels. F or instance, an abutment supported on a single row of piles is
considered flexible enough to accommodate longitudinal thermal cycling of
the superstructure and dynamic end rotations induced by the movement of
vehicular traffic. Nevertheless, the steel piles of such an abutment are
routinely subjected to axial and flexural stresses approaching, equaling or
682 M artin P. Burke Jr

exceeding yield stresses. 4 •6 Occasionally a combination of circumstances

results in visible distress.
Respanding to a 1973 survey, a nurober ofbridge engineers said that some
wingwalls oftheir integral abutments had minor cracks. 3 This problern was
corrected with the use of more generous wingwall reinforcing steel. Other
engineers reported pile cap cracking which appears to have been eliminated
by rotating steel H-piles to place the weak axis normal to the direction of
bridge movement. In a nurober of other instances, primarily in cast-in-
place construction, columns, stringers and bridge seals have cracked and
fractured when integral construction was used without proper allowance for
the restraint produced by rigid supports.
Currently precast concrete or prefabricated steel superstructures are
generally replacing small cast-in-place bridges in many states and provinces.
Consequently, problems associated with initial shrinkage are gradually
being eliminated. However, where cast-in-place construction continues tobe
used, flexibility of substructures remains a critical part of bridge design. F or
example, in Clellon Loveall's recent Civil Engineering article, 5 he says:
Structural analysis of our no joint bridges indicates that we should have
encountered problems, but we almost never have. Once we tied the stub
abutment of a bridge into rock, and the structure cracked near its end, but
we were able torepair the bridgeandinstaU [a] joint while the bridgewas
under traffic. The public never knew about it. That was one of few
problems.
Development of new forms of construction will be accompanied by
instances of structural distress, and this has certainly been true for
continuous bridges with integral abutments. However, as is shown by line A
of Fig. 1, the increased use of integral abutments suggests that 60% of
transportation departments are satisfied with the performance of integral
construction and are using such construction in one form or another for
Ionger and Ionger bridges. With continued care and consideration, the trend
shown by line A no doubt will continue.

INTEGRAL BRIDGE DETAILS

Illustrated in Figs 2 and 3 are integral abutment details for six

transportation departments. lt is probably not accidental that a fair amount
of similarity is evident in these designs since structural details from early
successful designs are adapted by other bridge engineers for use by their
departments. Even though there are similarities, there are also differences
which reftect the types ofbridges being built and the care and concern being
 Construction and Conversion of Span Bridges 683

given to the choice and development of specific details. lt should also be

realized that these sketches are 'bare bones' presentations. They do not
reflect other important design aspects such as skew, construction
procedures, etc., that are considered in the application of these details for
specific bridges. These aspects cannot be illustrated and properly described
in a paper as brief as this one. Nevertheless, since these aspects can have a
considerable effect on the performance, integrity and durability of integral
designs, it is appropriate to mention some of them for those engineers
considering such designs for the first time. Passive pressure, pile stresses and
cycle control joints are three topics under which these aspects may be
considered.

(A) Passive Pressure

To minimize passive pressure developed in abutment backfill by an
expanding integral bridge, design engineers have used a number of controls,
devices and procedures. Including but not limited to the following, they
have: limited bridge length, structure skew and the vertical penetration of
abutments into embankments; used select granular backfill and uncom-
pacted backfill; provided approach slabs to prevent vehicular compaction of
backfill or to permit the use of backfill voids behind abutments; used
embankment benches to shorten wingwalls and used suspended turn-back
wingwalls; and they have used semi-integral abutment designs to eliminate
passive pressure below bridge seats.

(B) Pile Stresses

Knowing that longitudinal forces in superstructures are somewhat
directly related to the resistance of abutment pile foundations to
longitudinal movement, design engineers have: limited the foundation of
integral bridges to a single row of slender vertical piles; limited the pile types;
oriented the weak axis of H -piles normal to the direction of movement; used
prebored holes filled with fine granular material for piles; provided an
abutment hinge to control pile flexure; limited structure skew; and used
semi-integral abutment designs for Ionger bridges to minimize foundation
restraint to longitudinal movement.

(C) Cycle Control Joints

Figure 4 illustrates design details being used by four states to provide for
the cyclic movement of integral bridges, and for the cyclic movement of
approach slabs where the slabs are attached to integral bridges. A number of
other states are using similar details, but most states do not appear to have
developed special approach pavement designs to accommodate the cyclic
movement of integral bridges.
684 Martin P. Burke Jr

Prestressed
concrete
beoms

LIMITATIONS
SKEW MAX "~'
s 45° 300'
~ 45° 150'

IOWA

Steel
girder

Varies
2'-d' Min

Brid e length "l'

Span
LI MI TATIONS
SKEW MAX "~'
45° 400'

TENNESSEE

OHIO
FIG. 2. Integral abutments.
 Construction and Conversion of Span Bridges 685

II" 2'-ff'

PPC 1-Bearn
or vr Bearn

ILLINOIS

Select
bockfill

.·.~,.

LIMITATIONS
SKEW MAX "~
o· 3r::~J'
Prestressed
30" 30d deck panels
NORTH DAKOTA

Prestressed
concrete
beoms

NOTE:
Turnbock w1ngs
eoch supported LIMITATIONS
by o steei-H pile. Not Estoblished

PENNSYLVANIA

Frc. 3. Integral abutments.

686 Martin P. Burke Jr

16' - 0" Concrete onci'Ol<

(a)

Cycte control
joint \

\ Bridge Approach slob \ 100'- 0" E1islin9

~ ~ ~rnem
~~~~==~~~F~-=~~-~~P.
~---~-~-=~~~-~.:~.,~-1~-.-:~-} .. ..
t. •... ~ . • .. .• •. r-'*""'
(b)
--41oo'-o" 3-o·
I~
Eaist. , pov't.

r -. -~~~---
~p,iä.;"; -;.:t
,ma
... ·· ·
Press.. e retiel joint

· · · ·• · · · :,;

SECTION

(c)
..•· . ''! : . . . •

Slidin11 , .. roce
5' -o·
Abutment

Appr~ch stob 15'- o• Anch slob

..
· (d)

FIG. 4. Integral bridge approaches.

 Construction and Conversion of Span Bridges 687

The purposes of all of the designs shown in Fig. 4 are the same although
specific details of each design are substantially different. Each recognizes
cyclic bridge movement and, where the approach slab is attached to the
bridge, cyclic approach slab movement as well. Each contains a reinforced
concrete approach slab for spanning the abutment backfill, and each
provides for the growth and pressure generated in rigid approach
pavements. All but one provides a sealing system for the cycle controljoints.
The initial appropriateness of these designs will be reftected in project
costs, and in the stress Ievels developed in response to cycling of bridges and
growth of approach pavements. Their actual effectiveness or success will
ultimately be measured by their ability to function for long periods under
traffic without periodic maintenance or major modifications. An evaluation
of these designs and a background forthat evaluation is given in Ref. 7. The
importance ofthe bridge/approach pavement interface design is emphasized
by a recent decision of the California Department of Transportation to
reconsider its use of Ionger integral bridges. Apparently cycle control joint
seal failures have been responsible for substantial erosion of abutment
backfill.
A number of questionnaires about integral bridge practices have been
circulated in recent years. The responses reftect the policies, attitudes and
opinions ofthose engineers responsible for bridge design policies. They also
show how some ofthose attitudes and opinions have changed during the last
decade. 2 ~ 4 • 8 · 9 References 4, 8 and 9 also contain valuable bibliographies for
those interested in a more in-depth study of current research on integral
bridge behavior and abutment piling performance.

INTEGRAL CONVERSIONS (RETROFIT)

Following the trend toward the use of continuous construction and the use
ofintegral abutments, as illustrated by lines A and B ofFig. 1, transportation
departments are also beginning to convert existing multiple span bridges
from simple to continuous spans. Line C shows that this effort began in the
1960s and has gathered strengthin the past two decades. Presently 11 of 30
departments, or about 37% of the transportation departments, have
converted one or more bridges from multiple simple spans to continuous
spans.
Although line C of Fig. 1 suggests considerable activity, it actually shows
only the relative number of departments that have madesuch conversions. It
is not indicative of the number of bridges that have been converted. For
example, when asked the question 'In recent years, have you converted any
bridges from multiple simple spans to continuous spans to eliminate
688 Martin P. Burke Jr

intermediate deck joints?', positive responses were received from only two
departments. The Ontario Ministry of Transportation and Communica-
tions responded:
We are modifying a few structures from simple spans to continuous spans,
eliminating the intermediate deck joints in the process ...
The Texas Department of Highwaysand Public Transportation responded:
In recent years, we have eliminated numerous intermediate joints.
Generally, this is done while replacing the slab. We simply place the slab
continuous across the bents. On a few occasions, we have removed only
the joint and surrounding deck area, added reinforcing, and replaced that
portion of the deck, thus tying the adjacent spans together.
The Tennessee Department of Transportation also has been actively
converting simple span bridges to continuous spans. To describe some of
this work, Edward Wasserman, the engineering director of structures,
presented a recent paper on 'Jointless bridges' in which he describes and
illustrates a number of such conversions. 10
To give this movement some direction, the Federal Highway Adminis-
tration has issued a Technical Advisory on the subject.U That advisory in
part recommends that a study of the bridge Iayout and existing joints be
made ' ... to determine which joints can be eliminated and what
modifications are necessary to revamp those that remain to provide an
adequate functional system.... [Where] feasible, develop continuity in the
deck slab. Remove concrete as necessary to eliminate existing armoring, and
add negative moment steel at the Ievel of existing top-deck steel sufficient to
resist transverse cracking [Figure 5a]'.
The detail shown in Fig. 5(a) refiects the procedure described by Texas.
Note that the detail shows only the slab portion of the deck is being made
continuous. The simply supported beams remain simply supported. For
such a construction it is important to ensure that one or both ofthe adjacent
bearings supporting the beams at a joint are capable of allowing horizontal
movement. Providing for such movement will prevent horizontal forces
from being imposed on bearings due to rotation ofthe beams and continuity
of the slab.
The state of Utah also has converted some simple span bridges to
continuous ones by using a design similar to the one shown in Fig. 5(b). For
deck slabs with a bituminous overlay, a waterproofing membrane can be
used to waterproof the new slab section over the piers. With a design like
this, it is understood that the deck slab would be exposed to longitudinal
fiexure due to rotation of the beam ends responding to the movement of
vehicular traffic. However, for short and medium span bridges, the deck
 Neaauve Koa>e~~t Rebaro

"\ \ ...., I Exiotinl lt.ebara to

Remove concrete as necessary to eli~inate

et
R..,..l%1 1!1 Place
(a) (b)
existing armoring, and add negative moment \_ Polyeihylene
steel at the level of existing top-deck band breoker ~
;:::
steel sufficient to resist transverse c.,
cracking. Generally reconstruct with ....
regular concrete to original grade. \.__..__r- Movoble 1:::
-
~
beorngs 5·
;:::
:::,
;:::
:::....
Comection ~
diopiYogm ;:::
4' - 0" <:::
<I>
Cast - in- ploce
;::;
- - Aspholt coocrete
1 deck slab weoring surfoce §"
t-:=4~"--:"'-:1 . .l ... l===!',.....,..._=?v=::j Beom reinf. Woterprooftng ~
membrone
i!lm!f!l ~
- Precost
§
Precost
1- beoms side -by-side ~
(c) box beoms ---+---l (d)
~
<I>
c.,
q_ Beorings I_ I•-~• I ..:.d
Elastomeric t. Beorings Prestressing
beorng pods stronds
a preformed Etostorneric
Iiner beoring pods
Pier

FIG. 5. 0\
Integral conversions at piers. 00
\0
690 Martin P. Burke Jr

cracking associated with such behavior is preferred by some over the long-
term adverse consequences associated with an open joint or a poorly
executed sealed joint.
In a new construction, conversion of simple spans to continuous spans is
rather commonplace. Figure 5(c) shows the design detail used by the state of
Wisconsin for prestressed I-beam bridges. A substantial concrete diaphragm
is placed at the piers between the ends of simply supported prestressed
beams of adjacent spans. I t extends transversely between parallel beam lines.
Then a reinforced concrete deck slab is placed to integrate the beams and
deck slab, thereby providing a fully composite continuous structure. This
type of prestressed I-beam construction appears to be standard for many
transportation departments.
Figure 5(d) shows the standard design detail used by the state of Ohio to
achieve continuity for simply supported prestressed box beams. These box
beams are placed side by side and then transversely bolted together. Finally,
continuity reinforeerneut is placed and the concrete closure placement is
made.
A 1969 paper by Clifford L. Freyermuth gives a rather complete
description ofthe considerations necessary to achieve continuity in a bridge
composed of a continuously reinforced concrete deck slab on simply
supported precast prestressed beams. 12 Conversion of existing bridgeseither
by a complete deck replacement or by replacing portions of the deck
adjacent to deck joints over piers can be accomplished by following the
procedures developed for new structures. Obviously, for existing bridges,
creep effects will be negligible. Shrinkage effects for other than complete deck
slab replacements should also be negligible. Such continuous conversion not
only eliminates troublesome deck joints but the continuity achieved also
results in a slightly higher bridge Ioad capacity since positive moments due
to live Ioad are reduced by continuous rather than simple beam behavior.
The details and methods described above provide either partial or
complete continuous behavior for live Ioads and superimposed dead load.lf
justified, continuity and composite behavior can be achieved for allloads by
providing temporary intermediate supports which are then removed after all
of the structural elements have been completed.
Although too recent to consider in terms of a design trend, conversion of
nonintegral to integral or semi-integral abutments for both single and
multiple span bridges has begun. Figure 6 illustrates design details used for
two recent conversions by the Ohio Department of Transportation.
Reconstruction of these abutments was made necessary by the substantial
darnage induced by pavement growth and pressure, by de-icing chemical
deterioration, or by both. lnstead of replacing backwalls and joints, and in
some cases bearings and bridge seats as well, it was decided to pattern the
 Construction and Conversion of Span Bridges 691

Slabs
reinf.
not shawn
t. Bearings -
.·

j..ii4-7ii''Y-'..,....,-,tt-l" PEJF
'-+.,........--1+-Steel trowel
finish and
2 layers af
grophite coated
sheet osbestos
pocking

BE FORE AFTER

4'-o" Slabs
1~5" 2'-1" 6" reinf.
not shown

BEFORE AFTER

FIG. 6. Integral convcrsions at abutments.

reconstruction after the design details used by the department for its new
integral bridges. In this way subsequent concern about the effects of
pavement pressure and de-icing chemical deterioration has been minimized.
Lastly, a number of transportation departments have begun to retrofit
multiple span steel beam or girder bridges constructed with intermediate
hingesunder unsealed deck joints. For one such example, end span hinges
692 Martin P. Burke Jr

and deck joints (originally intended to accommodate embankment

consolidation and abutment settlement) are being replaced with bolted
splices and a continuous concrete deck. Since this particular structure is 20
years old, the embankments are essentially fully consolidated and the
original justification for deck hinges no Ionger exists.

SUMMARY

As the trends shown in Fig. 1 continue, it appears that the use of continuous
construction for multiple span bridges will become standard for all
transportation departments in the very near future. lt also appears that the
use of integral abutments for single and multiple span bridges will increase
when comprehensive and conservative guidelines for their use become more
readily available, and when their long-term performance is more fully
documented.
Since design and construction of fully continuous bridges has become
routine, and continuous conversion of simple spans in new construction is
becoming more commonplace, it is surprising that similar conversion
techniques arenot used more often to convert existing bridges to continuous
bridges. Presumably the next decade or two will see a burgeoning in
retrofitting simple multiple span bridges to continuous bridges (line C, Fig.
1) and from nonintegral to integral abutments. When more information on
the operating stress Ievels of integral bridges is developed, and when more
fully described design details and procedures for integral conversions
become available, bridge engineers will be able to more fully justify their
consideration of such construction. Until then much intuition and prudent
judgement will continue to be used to ensure that integral construction and
conversion techniques will provide the service life needed to justify their
adoption and continued use.

REFERENCES
1. CRoss, H., Analysis of continuous frames by distributing fixed-end moments.
Proc. Am. Soc. Civ. Engrs (May 1930).
2. BURKE, M. P. JR, Bridge Deck Joints. National Cooperative Highway Research
Program Synthesis of Highway Practice, Transportation Research Board
(publication forthcoming).
3. EMANUAL, J. H., HULSEY, J. L., BEST, J. L., SENNE, J. H. and THOMPSON, L. E.,
Current Design Practice for Bridge Superstructures Connected to Flexible
Substructures. University of Missouri-Rolla, Rolla, Missouri, 1973.
4. WOLDE-TINSAE, A. M., GREIMANN, L. F. and YANG, P. S., Nonlinear Pile Behavior
in Integral Abutment Bridges. Iowa State University, Ames, Iowa, 1982.
 Con~truction and Conversion of Span Bridges 693

5. LovEALL, C. L., Jointless bridge decks. In Civil Engineering. American Society of

Civil Engineers, New York, November 1985.
6. JoRGENSON, J. L., Behavior of Ahutment Piles in an Integral Abutment Bridge.
Transportalion Research Record 903, Transportalion Research Board,
Washington, DC, 1983.
7. BuRKE, M. P. JR, Bridge Approach Pavements, Integral Bridges and Cycle
Contra! Joints. Transportalion Research Record 1113, Transportalion
Research Board, Washington, DC, 1987.
8. GREIMANN, L. F., WoLDE-TINSAE, A. M. and YANG, P. S., Skewed Bridges with
Integral Ahutments. Transportalion Research Record 903, Transportalion
Research Board, Washington, DC, 1983.
9. WoLDE-TINSAE, A. M. and KLINGER, J. E., Integral abutment bridge design and
construction. Report No. FHWA/MD-87/04, Maryland Department of
Transportation, 1987.
10. WASSERMAN, E., Jointless bridges. Engineering Journal, 24(3) (1987). American
Institute of Steel Construction, Inc.
11. Federal Highway Administration Technical Advisory T5140.16, Bridge Deck
Joint Rehabilitation ( Retrojit). US Department of Transportation, 1980.
12. FREYERMUTH, C. L., Design of continuous highway bridges with precast
prestressed concrete girders. AC! Journal, 14(2) (1969). Prestressed Concrete
Institute, Chicago, Illinois.
 60
Inspection and Rehabilitation of Steel Trusses for
Highway Bridges

A. G. LICHTENSTEIN
A. G. Lichtenstein and Associates, Inc., Consulting Engineers,
Fair Lawn, New Jersey, USA

ABSTRACT

Steel trusses played an important role in the deuelopment of highway

transportation in the United States between 1870 and 1940. In this period of
time, numerous patented structural steel ( and wrought iron) structures were
designed and constructed by engineers/contractors on major state highways
and rural county roads. Many ol these structures haue heen replaced over the
years but still a sufficient number remain today.
The suhject ol this paper is to describe the methods and procedures involved
in extending the usef'ul life and upgrading of these older truss bridges into
structures that can safely support modern and heavier loadings.
The author has selected three types ol trusses for a detailed presentation.

INTRODUCTION

Old bridges should not be allowed to die. They should be rehabilitated and
put back into service. This is particularly true for trusses on the county
roadway systems, where rural conditions still prevail and traffic has only
increased moderately over the years. Three examples will be presented as
follows:
1. A pin-connected Baltimore truss erected in 1899 by the Berlin Bridge
Company of Berlin, Connecticut. This bridge is located at Stuyvesant
Falls, Columbia County in the State of New York.
2. A pin-connected eye-bar deck truss bridge constructed in 1896 by the
Canton Bridge Company of Canton, Ohio, and first strengthened in
1920. This bridge is located in Boonton, Morris County, New Jersey.
695
696 A. G. Lichtenstein

3. A Phoenix truss with patented compression members built in 1894 by

the Stagg Company. It is located on Elm Street in Oradell, Bergen
County, New Jersey.

STUYVESANT FALLS BRIDGE

Two steel trusses 34ft (10·4 m) high, top to bottarn chord, and 19·4 ft (5·9 m)
CL CL carry a 17ft (5·2 m) wide roadway over Kinderhook Creek in one
simple span202ft (61·6 m) long. A 4·7 ft (1-4m) sidewalk is cantilevered off
the upstream truss. The trusses and fl.oorbeams were of the original
construction; however, longitudinal steel stringers and a grating deck had
replaced the original timher beams and fl.oor planks.
An in-depth inspection of the bridge was undertaken by our firm and
identified the following structural deficiencies:
1. The bottarn chord was so severely corroded at the south bearing that it
became discontinuous and the truss member and shoe moved outward
longitudinally, and the rocker tilted in a vertical direction.
2. The connections of three vertical members of the truss to the
fl.oorbeams, at the panel points, were extensively corroded and non-
functioning.
3. Many roadway stringers were rusted and cracked in the last panels at
the abutments, where salt and dirt accumulated araund them.
4. All floorbeams were in good condition but only capable of safely
carrying a 7-t (AASHTO H) live Ioad.
5. The trusses computed weil except for 16 subpanel verticals, whose
rating was 9 t.
6. Other elements of rehabilitation were required, such as restoration of
the twisted rocker, repairs to some sidewalk and railing areas, and
cleaning and painting of the entire bridge.
The county Superintendent of highways then decided to repair and
upgrade the bridge with county forces. The county had an abundance of steel
members stored in its yard and these could be tapped for use on this project.
The first order of business was to jack up the south end of the trusses
approximately 1 ft (0·3 m) and relieve the pressure on the twisted shoe. This
completed, the truss member and rocker were pushed horizontally by
another jack to a calculated point and a chord mending device was installed
to keep them there permanently.
The jacks and auxiliary beams were removed. Then the verticals and
floorbeams were reconnected with permanent type details, and the
subverticals ofthe two trusses reinforced with special 'horseshoe' collars (see
Fig. 1).
 Jacking and bottom chord Elevation view of truss V>
<see detail belowl ~
~
..._
Existing
Slotted holes for post
Temporary jacking new hanger rods
~
...---tw-: on assembly ~
~
Y.~ ~~ .
~" '0...,>
~ ~ ..,.. ::t:
o:Q•
~.t
I.
- . " ::::-
~
~~$
Bottom c hord
~
( ~
"' Truss ~
~
Bottom chord

2'-J'sploce rods
Existing
pin
.n Floor beam hangers Vertical member repair

Stuyvesant Falls Bridge

Jacking details and repair of bottom chord Columbia County, NY"
Typocal truss repaors
Not to scale
o-,
FIG. 1 '-0
-...1
698 A. G. Lichtenstein

Next in line was the strengthening of the floorbeams. Without removing

the grating (fair condition) and stringers, a platform was erected under each
floorbeam. Four angles were bolted onto the webs ofthe beam as close to the
ftanges as possible. The corroded stringers were left in place, but new beams
of similar depth were set alongside the bad ones and shimmed into place. The
railing and sidewalk were repaired in spots where needed. Cleaning and
painting, however, was deferred for next season. The bridge was reopened to
traffic with a capacity of 15 t (inventory), which is adequate to handle local
traffic including buses, fire engines, ambulances and other community type
vehicles. For the occasional runaway heavy truck, the operating rating ofthe
bridge is over 20 t. The time necessary to complete this work was 4 months.
The county solved the Stuyvesant Falls Bridge problern by doing most of
the work itself and gained a great deal of satisfaction at a reduced cost of
construction and restoration of a historic structure. The Superintendent and
Columbia County received awards from state and national organizations
commemorating their achievements on the Stuyvesant Falls Bridge.

THE BOONTON BRIDGE

The 90-year-old Boonton Bridge carries Washington Street over the

Rockaway River and the Jersey City Reservoir. In 1920 the bridge was
strengthened by the addition of intermediate chords and new deck, and
other structural work.
The original bridge consisted of four pin-connected deck truss spans
providing a total structure length of approximately 470ft (143-3 m). The
floor system was constructed of steel plate girder floorbeams and steel
stringers, which supported an asphalt-filled metal pan deck. The deck
provided a 23ft 6in (7·2m) roadway and was flanked by a 5ft (1·5m)
cantilever sidewalk. The substructure consisted of stone masonry abutments
and three non-reinforced concrete piers (Fig. 2).
A. G. Lichtenstein and Associates, Inc., made an in-depth inspection, and
prepared a rating and feasibility study for upgrading. An underwat~r
investigation by AGLAS found the piers in good condition. The
superstructure was determined to be substandard, however, and was posted
for 3 t loading. The study recommended that the bridge can be economically
rehabilitated to a 15 t capacity. The rehabilitation project included the
following:
1. Replacement of the asphalt pan deck with a concrete-filled grating
ftoor.
2. Upgrading the steel stringers and floorbeams.
3. Repair and strengthening of truss members, shoes and connections.
 Steel Trusses for Highway Bridges 699

(a)

(b)

FIG. 2

Total construction cost for the complete bridge and approaches was
approximately $2 600 000, which compared favorably with an estimated cost
for a new bridge of $7 000 000. Because of its historic nature, as well as its
location over a reservoir, the acquisition of permits from state and federal
agencies would have taken approximately 5 years. The obtaining ofpermits
700 A. G. Lichtenstein

and the preparation ofbid documents for the rehabilitation project required
a little more than 1 year. The actual construction was also completed in
about 1 year. It is interesting to note that the Federal Government
contributed 90% of the cost of construction, while the remaining portion
was jointly provided by the State of New Jersey and Morris County. The
Morris County engineer was Mr George E. Burke and the general
contractor was Karl Koch Erecting Company of Carteret, New Jersey. Our
firm prepared the bid documents and checked all the working and shop
drawings, while the New Jersey Department of Transportation supervised
the contractor's operations and provided a full-time resident engineer.
A modern bridge crossing over the Rockaway River for heavy interstate
traffic was constructed within 2 miles ofthe bridge site. The Boonton Bridge
services county type of traffic and will be providing the community with
excellent service for many years to come.

ELM STREBT BRIDGE

The Elm Street Bridge over the Hackensack River is a single span, simply
supported, half-through (pony) truss. The truss consists of five panels, each
15ft (4·6m) long for a totallength of75ft (22·9m). The trusses are spaced
23ft (7·0 m) between centerlines and are 10ft (3·0 m) high. The roadway
accommodates two lanes of traffic, one in each direction, and a 5 ft (1·5 m)
sidewalk was cantilevered off the west truss.
This structure is known as a Phoenix bridge, wherein the end posts,
verticals and upper chord members are composed of Phoenix columns of
wrought iron. These patented columns were fabricated utilizing four
channels which were turned by machine into quarter round sections, riveted
together, to result in a circular member. The open grating deck was found to
be in good condition with only light to moderate rust. The stringers
supporting the deck were found in generally good condition. However, they
rated 6 t. The floorbeams were found to have moderate rust but had a
computed live load capacity of 6 t. The truss appeared in fair condition with
some members showing evidence of impact from vehicles. However, the
capacity ofthe truss to carry liveloadwas only 2 t. The two abutments were
composed of ashlar stone masonry and were found to be in good condition,
except that many joints needed remortaring (Figs 3-5).
The assignment was to upgrade this 90-year-old bridge to a live load
capacity of at least 12 t, which would permit the safe crossing of school vans
and small buses. The deck and the stringers could economically be replaced.
The floorbeams, on the other hand, were strengthened by the addition of
four angles which were riveted just under the top and bottom flanges. The
 5 panels at 15'-0 =75'-0

3 rail railing 61i:i'0 OD phoenix

2-3"x3"x1!4"angles (truss) column section
w/lacing (typ)\ 114" thick Cl')
U1 U2 U41 ~
~
~
;:::

"'"'~
'0"
...,
12"0 san sewer and beams (15"1).L5
separate from truss ::X:
Conc pier to support san. sewer only ~
------=-._-=----------=-- -~---~- -.t=- -- -------=---=----
: I I
""
~
+ J ~ rr ~
River bed ' River bed
east elevation west elevation ~
(April 1980) (April1980) ~
Elevation

FIG. 3

--..)
0
-
 --:1
0
N

<t truss

~
I I
' I 14'-Q ~
2).._1 I
----"" 0
i~=" I:'-<
-- ;:::;·
I ;::-
......
! ~
--11-- - ;::
~I "'~
s·
~·~~'
IV:\~\~ i111F. to
'"f- -iF lifr~ _Jr=t __ .. liF·~ -~IF ~~F ~ t:;;\:1r v~tv~. ~F
<t truss ••
FIG. 4
 Steel Trusses for Highway Bridges 703

23'-0
,___----:
C""""t:- u s~s to C t r u s s
r -':

21'-3 curb to curb 5'-3

0_,
N

Floorbeam (4 typ.>
web pl ,".. x 30"
4 flange Ls
3" X 3" X 1/4" Sect•on A-A

(aJ

(h)

FIG. 5

main concern was how to reinforce the historic truss without dismantling it
or distorting its architectural appearance. The solution proposed was to add
two 36 in steel beams on the outside of the floorbeams under the trusses,
which beams would span from abutment to abutment. Each floorbeam was
then connected into these two auxiliary bea ms, creating a live Ioad system
wherein the old truss and the new beams were sharing the Ioad (Fig. 6).
To replace this historic Phoenix bridge would have required many years of
negotiations at a substantially increased construction cost for limited
serviceability. In this way Bergen County has retained and restored a historic
704 A. G. Lichtenstein

Exist1ng fence
to rema1n
0

Prop. bracing angle

Prop. side plates bolted to web // :>;/~~angle 6"x4"

/>'C
7 I
and1"0shoe
Floor beam repair /' Exist1ng support
East end f loorbeam no.1 to be removed
Ex1sting steel bolster sidewalk
stnngers to be removed

FIG. 6

property and, with normal maintenance and no overloading, the Elm Street
Bridge will provide this community with many years of customized usage.
The three examples described above show that, under favorable
circumstances, bridges can be retrofitted and placed into the transportation
systems with much success. The additional benefit of preserving the
American heritage of bridge construction is an unexpected bonus.
 61

The Renovation of a Victorian Swing Bridge

BRIAN SIMPSON and MICHAEL F. ßLYTH

Husband & Co., Consulting Engineers and Architects,
Alliance Hause, 12 Caxton Street, London SW1HOQP, UK

ABSTRACT

Cross Keys Bridge is an iron and steel swing bridge carrying the A17 Newark
to King's Lynn trunk road over the River Nene. Built in 1897 to carry a railway
line and single lane road, it was converted to two lanes ofhighway in 1963 after
the railway line was closed. In 1982 renovation of the old bridge for a limited
future life was the most economic optionfor providing a raute for the proposed
Lang Sutton and Sutton Bridgebypass over the River Nene. It also enabled the
bridge, by then a class //* listed structure, to be 'preserved' in working
condition.
Deterioration o.f some parts ofthe substructure and the mechanical wedging
system »'as so advanced that an early contract was Iet to deal with theseurgent
works. The subsequent main renovation contract included

-replacing deck to improve loading capacity,

-repairs and repainting of existing meta! work,
-increasing headroom .for road users,
-separating pedestrians from road traffic,
-modernising the mechanical drives and control systems, and
-fendering to reduce risk o{ shipping collision.

The paper describes these works and the repairs which were devised afier the
full state o{ corrosion darnage had been revealed. The repairs, including
replacement ofthe bridgedeck in stages, were undertaken without interruption
to the navigation requirements to swing the bridge and, with the exception o.f a
.few weckend closures, single way working for road traffic was maintained
throughout viiinter periods.
705
706 Brian Simpson and Michael F Blyth

HISTORY OF THE BRIDGE

In 1827 part of the work of improving the Fens included a new man-made
cut confining the River Nene at its mouth to provide a more direct and
reliable navigation channel northward to the Wash. This work included the
first Cross Keys Bridge.
In 1864 the expansion of railway ventures brought a rail track crossing
this channel at Sutton Bridge utilising an existing movable bridge, the
second such crossing at this location. The present structure was constructed
by the railway company in 1897 to carry a single track of railway. A roadway
and footpath were accommodated alongside the railway track and
separated from it by the middle of three longitudinal girders. Road users
were obliged to pay a toll and there is a commemorative plaque celebrating
the abolition of this toll in 1903. In 1963, after closure of the railway (in
1959), the bridge was redecked to accommodate two lanes of road traffic
separated by the central girder system. The present bridge was therefore
constructed with railway use and railway clearances in mind and then
converted to road use. The appearance of the control cabin for the swing
span located over the carriageways reflects its original additional function as
a signal box for the railway.
A feasibility study in 1963 for improving the A17 to bypass the villages of
Long Sutton and Sutton Bridge included a proposal to replace the swing
bridge with a new bascule bridge. This scheme did not proceed at the time
and the question of crossing the river as part ofSutton Bridgebypass was re-
examined in 1982. By this time there was increasing interest in industrial
archaeology and the bridge and the associated building nearby housing the
hydraulic accumulator were listedas dass li* buildings. U nder this category
a public enquiry would be necessary to obtain approval for demolition.

FIG. 1. Location of site.

 The Renovation of a Victorian Swing Bridge 707

OPTIONS FüR FUTURE ROAD TRAFFIC

Three options for the river crossing were considered.

1. A new single-leaf bascule bridge on the published line of the then
preferred route for the bypass.
2. A new swing bridge to provide a skew crossing adjacent to the existing
bridge.
3. Refurbish and strengthen the existing bridge, improving clearances
and separating pedestrians from vehicular traffic.
The preferred route for the bypass produced a crossing about 300m south
of the existing bridge and the navigation authority required the demolition
of the existing bridge as part of the scheme unless the new opening bridge
gave virtually no obstruction to navigation. This arose from the navigation
requirements peculiar to this site where, because of distances from the
mouth of the river and the berths at Wisbech and the presence of a sandbar
at the mouth of the river, ships' movements in both directions generally
occur at high flood tide. An additional navigation hazard is a small bend in
the channel at the bridge site.
lf the existing bridge were demolished, pedestrians would have a
considerably Ionger journey to cross the river and it may have been
necessary to provide an additional pedestrian bridge close to the existing
bridge site to meet this need.
A site for a new swing bridge immediately upstream ofthe old was found
which could be accommodated within the restrictions imposed by the
existing fendering system. With this option it would have been possible to

FIG. 2. Cross Keys Bridge before renovation.

708 Brian Simpson and Michael F Blyth

retain the old bridge in either an operational condition or swung clear ofthe
channel and kept for possible occasional use. With a new bridge close to the
site of the existing crossing, special provisions for pedestrians would not be
necessary.
The third optionwas torepair and improve the existing bridge to give it a
further, if limited, useful life.
The three options were compared and estimates of cost prepared. It was
decided to proceed with the refurbishment and strengthening of the existing
structure and this was executed in two stages. Priortothis consideration the
existing bridge had been inspected and shortfalls identified.

SHORTFALLS IDENTIFIED

The structure was inspected in 1982 and the following shortfalls identified:
1. Generally the structure was corroding and in need of structural repair
and repainting.
2. The overhead girders supporting the control cabin and the overhead
bracing trusses restricted vehicular headroom.
3. Through girder configuration of the swing and approach spans
physically restricted the road to two narrow single lane carriageways
with substandard side clearance.
4. Pedestrians were at risk from adjacent traffic lanes because of narrow
footpaths and shallow kerbs.
5. The water hydraulic machinery was old, worn and costly to maintain.
6. Electrical wiring and electrical components were obsolete.
7. Access and lighting for operators and maintenance staff were
unsatisfactory.
Following an assessment ofthe load capacity ofthe structure it was found
that road decks were not capable of carrying 45 units of type HB abnormal
loading required for this route.
A separate assessment of the risk of shipping collision found the existing
fendering systems inadequate to protect the structure from accidental
damage.

URGENT WORKS CONTRACT

The inspection identified the substructure subject to tidal action as in urgent

need of repair and it was decided to deal with this aspect ofthe renovation in
advance of the main strengthening and improvements. The wedging system
 The Renovation of a Victorian Swing Bridge 709

operated by a low pressure water hydraulic systemwas replaced by a new

high pressure hydraulic system. The rotational drive system remained on the
water system until the main renovation contract. To house the new
hydraulics and the associated electrical control system, a new pumphause
was installed below road deck Ievel between the turntable and the tail end of
the swing span. The main elements of the urgent works contract were
-repairs to the substructure,
-new pumphause and access,
-new wedge gear,
-new control system, and
-cleaning and painting substructure.
The turntable on which the bridge rotates is supported on nine piles cased
in cast iron and braced together. The floor of this turntable is comprised of
arched iron plates between an arrangement of radial and circumferential
box girders on which the roll er track of the turntable carriage is supported.
The arched plates had a concrete infill to give a reasonably level floor. At
high tides the water Ievel is higher than the underside of this turntable floor
and the iron work had suffered considerable corrosion. The lower part ofthe
main box girders had been inundated at high tides with similar
conseq uences.
The contract included removing the concrete infill and all the arched steel
floor plates to assess the extent of loss of material and decide on remedial
measures. Although the floor plates were corroded beyond repair it was
found that the supporting structural members werein reasonable condition
so the floor was reinstated using new arched floor plates connected to the
existing steel framework. The original floor plates had been riveted but the
replacements were bolted using grade 8·8 bolts. A bituminous sealing
compound was used to bed the plates and caulk the joints.

MAIN RENOVATION CONTRACT

The main renovation contract required the repair and repainting of the
structure, provision of new steel trough stiffened decks surfaced with epoxy
bauxite, a new footpath attached to the outside of the existing north truss
and plate girders, and the provision of greater headroom beneath the cross-
bracing and control cabin. Also included within the contract was abutment
strengthening, the completion of the refurbishment of the mechanical and
electrical drive system, and the provision of a new control cabin tagether
with major extensions to the existing feodering system.
The structure is on the main A 17 trunk road, and road traffic flow had to
710 Brian Simpson and Michael F Blyth

be maintained except for limited weekend closures. The bridge had to be

operational at all times to allow the passage of waterborne traffic on the
River Nene on their trips to and from Wisbech. The A17 is a busy route for
holiday traffic to and from the East Anglian coast with peak traffic flows
during summer, particularly at weekends. For this reason it was only
possible to close one lane of traffic during the winter months from October
to April. Because of the amount of work to be undertaken during such a
partial road closure period, the contract was programmed to run over two
winter periods with lane closures during consecutive winters. lt was also
necessary to let the contractor have complete possession of both
carriageways for a few weekend winter periods in order to carry out
operations that could not be done safely with traffic on the bridge. An
example ofthiswas raising the control cabin and its supporting framework
to increase vertical headroom over the carriageways.
During the first major closure period in the winter of 1987/88 the
contractor raised the control cabin and the overhead bracing, repaired and
part painted the north truss and part of the centre truss and approach span
plate girders along the north carriageway, and installed the new footpath. A
limited amount ofwork was carried out to the centre truss and the north face
of the centre plate girders. lt was not possible to complete all the work
planned for this period but it was possible to reopen the north carriageway
to road traffic in time for the summer holiday period.
Details of repairs to the existing corroded structure could not be
determined in advance ofletting the contract except in general terms because
the full extent of corrosion darnage could not be revealed until the existing
deck had been removed and the old metalwork blast cleaned. The provision
of access for dismantling, cleaning and repair was also essential for
inspection and assessment of damage. The areas of difficult access below
deck Ievel were those which had not had the benefit of routine painting and
were the most corroded. The existing road deck protected the centre sections
ofthe cross-girders and the intermediate longitudinal members but the ends
of these girders and the lower gusset plate connections on the trusses were
vulnerable to attack from road salted surface water and consequently were
in extremely poor condition.
The sequence of operations was dictated by the nature of the redecking
work. The old deck was taken up sequentially in panels starting from the
approach spans at the east end. Cleaning, inspection and repairs to the
structural members supporting the deck were then undertaken before the
new deck panels were introduced, again starting from the east end. Once
deck panels had been removed access to the area of activity was limited and
careful organisation was needed to ensure progressive repair and
reinstatement.
 The Renovation of a Victorian Swing Bridge 711

Work on the swing span required an additional degree of control to

maintain the swing span in a state of balance whilst deck panels were
removed and replaced. Furthermore, the addition of the footpath on the
north side of the structure meant progressive counterbalancing transversely
to keep the centre of gravity ofthe whole rotating structure reasonably close
to the centre of rotation at all times. This was necessary to ensure safe
operation and to maintain the loading pattern on the pile group under the
roller path.
Two methods ofcounterweighting were tried. For the replacement ofthe
first lane of road deck, water tanks were used but simple weights were
adopted for the second lane because the water tanks were found to impede
access.
To check the balance of the bridge, an attempt was made using strain
gauges but these were frequently damaged and the method eventually
employed entailed checking Ievels at the extreme ends of the swinging span,
enabling quite small variations in the state of balance to be monitored.
In the winter of 1988/89 work on the south truss and approach span plate
girders together with the centre truss and plate girders was completed. Work
on the centre truss was achieved by providing a timher sheet barrier bolted
to the north side of the truss, which allowed work to continue without
interference to traffic ftow. At the end of the second major lane closure
period outstanding work in the first (north) lane closure was completed.
Work not affecting the free passage ofvehicles was carried out throughout
the contract period. Complete road closures were allowed to enable the ends

FIG. 3. View illustrating restricted traffic lanes.

712 Brian Simpson and Michael F Blyth

of the swinging span to be repaired and painted, and to enable the

mechanical and electrical equipment to be tested.

UNUSUAL ASPECTS OF REPAIR WORK

Some unusual aspects of the renovation work relate to the standards for
health and safety introduced since the bridge was built. Reference has been
made to box girders with access manholes. Many of these are now regarded
as too small and exceptional measures had tobe taken to deal with internal
surfaces. It was first thought that because of these difficulties a sprayed
grease paint, which could be used without a high standard of surface
preparation, would be employed inside the boxes. However, access to the
inside of some box members was better than anticipated because large areas
ofweb and bottom flange had corroded completely away. Previous repairs in
the turntable area had compensated for this loss of material with additional
web and external flange plates.
Another difficulty related to the age of the structure was the uncertainty of
the material used in its construction and its weldability. Sampies of material
were taken from a number oflocations and identified by micro examination
and chemical analysis. This investigation by Messers Sandberg revealed a
mixture ofwrought iron and low carbon steel with sulphur and phosphorus
content higher than would be permitted to current standards for weldable
structural steels. As welding was a technique that was being considered as an
option for certain repairs, welding procedure trials were undertaken on the
samples which established satisfactory procedures for field use.
Although the composition of some of the existing paint system was
known, there was doubt about other paints. An investigation of paint
samples did not initially reveal the presence of Iead but when large-scale
paint removal started it was found that the existing paint system contained
lead and it was necessary to take the normal precautions to deal with this
potential health hazard. To minimise airborne lead pollution most of the
thicker coats of paint were removed by hand chipping prior to blasting to
clean metal.
The lower chords of the lattice girders and the lower gusset plated
connections had suffered quite badly from corrosion. As the deck was
removed from each carriageway these areas were cleaned and inspected
before a repair scheme for each connection was devised. These repairs were
rendered more difficult because the bridge had to be kept operational and
the Ioads in the members being connected were subject to variations due to
bridge openings and live load on the adjacent carriageway as well as
variations from thermal effects.
 The Renovation of a Victorian Swing Bridge 713

Wherever it was necessary to remove rivets these were replaced by high

strength fitted bolts and anti-vibration nuts. When gusset plates were
reduced in thickness, part ofthe plates were replaced and butt welded to the
remaining sound material. These welds were executed before bolting up the
repair.
These particular repairs were often a question of doing what was feasible
rather than what would be done for a new construction and considerable
ingenuity had to be used by all concerned to make an effective repair. F or
example, the general advice that welding should not be 'mixed' with riveting
or bolting had to be tempered to produce a repair sequence in which the
effects of welding (particularly welding shrinkage) could be accommodated
before bolting. lt was also found expedient to secure rivets which had
virtually lost their heads from corrosion but which were otherwise sound
and tight in their holes by placing a bead ofweld around the perimeter ofthe
exposed shank.

ACKNOWLEDGEMENTS

The work described in this paper was undertaken for the Director of
Transport, East Midlands Regional Office ofthe Department ofTransport,
and the Director of Highways and Planning of Lincolnshire County
Council. The authors gratefully acknowledge the permission of the Chief
Highway Engineer of the Department of Transport to publish this paper.
 62
Cost-Effective Strategies in Bridge Management

R. S. REEL and C. MURUGANANDAN

Bridge Management Section, Structural Office, Ministry of Transportation,
Ontario, Canada M3M JJ8

ABSTRACT

This paper assesses rehabilitation projects in an economic framework. Two

techniques are outlined. The first is the present value analysis which compares
alternatives at the project Ievel. The second is the incremental benefit/cost ratio
analysis which prioritizes the cost-effective rehabilitation alternatives at the
project and network Ievels. Various parameters requiredfor the analysis are
discussed. Computerprograms utilizing Lotus 1-2-3, Version 2.01, required
for the analysis are described. The application ofthe techniques is illustrated by
examples of bridge rehabilitations.

INTRODUCTION

An economic evaluation is an important step in the decision-making process

for work that involves major expenditures. At the project Ievel, the costs for
alternative Ievels of improvements to a bridge are compared to determine
the most economical option for the bridge based on either the present value
analysis or the incremental benefit/cost ratio analysis. At the network Ievel,
the costs for different Ievels of improvements for all the bridges in the
network are compared to prioritize work based on the incremental cost/
benefit ratio analysis.

PRINCIPLES OF COST-EFFECTIVE STRATEGIES

Theory of Present Value Analysis

The present value analysis involves the calculation of the cost of
alternative schemes in present-day monetary terms, i.e. the amount that is
715
716 R. S. Reet and C. Muruganandan

required in today's value to obtain goods and services at any future date. It
allows for the comparison of alternative schemes on an equitable basis.
The present value of an expenditure C in year n at a discount rate r is

PV= C
(1 + rt
The present value of a number of expenditures C1 n, where n = 1, 2, ... , N 1
for a period of N 1 years, is
N1

PV, _ \
L
C1n
1- (1 + rt
n=1

Incremental Benefit/Cost Ratio Analysis 1

The incremental benefitjcost ratio is the ratio of the additional benefits
realized in moving from one improvement alternative to another divided by
the corresponding difference in costs. This method not only optimizes the
selection of alternatives efficiently but also ranks the projects beginning with
the most net beneficial. It is used both at the project and network Ievels.

Total
Benefits

'
First
Costs($)
First
cost
1 2 3 4
Alternatives
FIG. l. Total benefits and first cost.

Figure 1 shows the total benefit and first cost curves plotted for the
various alternatives for a bridge. Initially the increment of benefit is higher
than the increment of cost. As costs increase the incremental benefits decline
and are less than the incremental costs. The slopes of these benefits and first
cost curves support the theory of diminishing returns. Fora partic..~Iar Ievel
of improvement there exist points on the benefit and cost curve" where the
slopes ofthe two curves are equal, i.e. IB = IC. At this Ievel ofimprovement
the net benefit is a maximum. This is evident from Fig. 2. Any option below
this Ievel where IB/IC > 1 is a desirable option.
The procedure is to Iist rehabilitation alternatives in the order of
increasing costs and calculate the incremental benefit/cost ratios.
Alternatives for which the incremental benefit/cost ratios fall below one are
discarded. U sually, as the Ievel of cost increases the incremental benefit/cost
 Cost-Effective Strategies in Bridge Management 717

Net
Banefit

1 2 3 4
Alternatives
FIG. 2. Net benefits.

ratio decreases. However, ifthe ratio IB/IC increases with increase in cost an
adjustment is made to that particular option. The incremental benefitjcost
ratio is calculated by considering the previous option but one instead ofthe
previous option. The options are sorted in descending order ofiB/IC. For an
unlimited budget the most net beneficial alternative is the one with the
largest initial cost and whose incremental benefit/cost ratio > 1. For a
limited budget the order of preference is the order from the highest to the
lowest incremental benefitjcost ratio.

PARAMETERS REQUIRED FOR COST-EFFECTIVE ANALYSIS

Parameters Required for Present Value Analysis 2

The following parameters are required to perform the present value
analysis for each alternative:
---capital cost,
-maintenance cost,
-life cycle,
-residual life, and
-discount rate.

Capital Cost
The following should be estimated for each alternative m constant
monetary terms:
-engineering design cost,
---construction cost, and
-miscellaneous costs such as demolition, right-of-way, approaches,
utilities, stream diversion, detours, etc.

Maintenance Costs
Costs associated with maintenance are the routine maintenance costs.
These would include minor repairs, maintenance, touch-up painting, etc.,
carried out on a regular basis.
718 R. S. Reel and C. Muruganandan

Life Cycle
The life cycles for the treatments should be estimated. Usually it is the time
between two successive replacements or rehabilitations.

Residual Life and Value

The various alternatives considered may have usefullives at the end ofthe
time frame. This is termed as the residuallife. There are no specific methods
of assessing this. A thorough knowledge of the performance of past
rehabilitations and experienced engineering judgement are probably the
best way of assessing the useful residual life. From the residual life the
residual value of the structure for the particular alternative can be
determined. There are several methods available for determining the
residual value. The method used here is the second cycle replacement
method.

TABLE 1
Residual value

Option Year of Replacement Residual Value at Differential Residual

replacement cost years year N 1 value value at
(2nd cycle) (option 1 as year 0
base)

1 c
2 c
r = discount rate
C = replacement cost
C1 =C C 2 = Cd(l + rtz-N 1
C01 = C 1 - C=O Coz = Cz- C
CRl =Ü eRZ= Coz/(1 + r)Nl

Table 1 shows the residual value calculations for options 1 and 2, whose
second cycle replacement will be in years N 1 and N 2 .

Discount Rate
The discount rate for government projects depends on several factors, 3
such as the magnitude of investment return, tax rates, capital market
conditions, preferences for current and future consumption, methods used
to finance projects, etc. A discountrate of 6% is recommended for projects
owned by government agencies. These rates may be different for other
agencies. Sensitivity analysis may be carried out by varying these rates.
 Cost-Effective Strategies in Bridge Management 719

Parameters -Required for Incremental Benefit/Cost Analysis

For the incremental benefit/cost analysis, the following additional
parameters are required:

-agency costs,
-user costs,
-agency benefits, and
-user benefits.

Agency costs
Agency costs are the same as for the present value analysis.

Agency benefits
Maintenance and various types of rehabilitationsextend the usefullife of
the bridge. These expenditures would postpone major expenditures for
replacement. The difference between the discounted future cost of a
rehabilitation option and that of a replacement option is the agency net
benefit. The agency net benefit plus the initial cost is the agency total benefit:

Agency benefit = PVR- PV1 + C 1

where C1 is the initial cost ofthe rehabilitation option, PV1 is the discounted
present value of costs associated with a rehabilitation option and P VR is the
discounted present value of costs associated with the replacement option.

User costs
User costs are costs incurred by the user due to deficiencies or substandard
conditions at the bridge. The following are the user costs.

Accident costs. Accident costs resulting from bridge width restrictions,

poor approach alignment, etc.

Functional restriction costs. Functional restnctwn costs due to load

restrictions and detours for certain classes of vehicles increase the travel
time, and hence operating costs. These vary for different locations and
countries.

User benefits
User benefits of a bridge rehabilitation option are the reduction in costs to
the users due to the rehabilitation. In determining user benefits it is assumed
that deficiencies will be eliminated when the bridge is replaced.
The reduction in the number of accidents due to a certain type of
improvement is used as a measure of user benefit for that type of
720 R. S. Reet and C. Muruganandan

improvement. The dollar value placed on different types of accidents is

crucial in estimating user benefits. These may vary for different countries:

Annual benefits = (change in accident rate) x (ADT) x (365) x (accident cost)

The change in accident rate is measured by the difference in the number of

accidents per million vehicles. The accident cost depends on the severity of
the accident. It is very difficult to place monetary values on accidents. Several
attempts have been made to quantify these. The two common methods are
(a) human capital approach and (b) willingness to pay approach.
The human capital approach takes into consideration the direct and
indirect costs. This approach does not consider the intangibles offered to the
society and the loss in the quality of life. The willingness to pay approach
includes the value of life in the estimates. As such the willingness to pay is
more conservative.

Costs
I
0 N
Years
FIG. 3. User benefits.

Figure 3 illustrates the method used in determining user benefits.

The user benefit is the present value ofthe annual benefits of( C 1 - C 2 ) for
years n1 to n 2 and C 1 from n 2 to N:

_ _ {1-(1+r)(nz-nJ)} 1
-(Cl C2) r (1+rtl

+ C {1- (1 + r)(N-nz)} 1
1 r (l+rtz
where N is the period considered for life cycle analysis, C 1 is the annual user
costs associated with the bridge, n1 is year of rehabilitation which would
extend the life, C 2 is user costs after rehabilitation and n 2 is year of
replacement.
 Cost-Effective Strategies in Bridge Management 721

Present Value Analysis Using PRVAL Program

PRVAL is a template overlay developed to perform financial analysis for
bridge rehabilitation projects using Lotus 1-2-3, Version 2.01, on a
worksheet format. There are four different options available to carry out the
financial analysis at different Ievels of sophistication.

Incremental Benefit/Cost Ratio Analysis Using COSBEN Program

COSBEN is a program developed to perform incremental benefit-cost
analysis for bridge rehabilitation projects using Lotus 1-2-3, Version 2.01,
on a worksheet format. At the project level the analysis can be carried out
with or without user costs. The output from the project level analysis is used
to perform the network analysis.

EXAMPLES OF COST-EFFECTIVE ANALYSIS

The application of the techniques at the project and network Ievels are
illustrated by the following examples (in $1000).

Example 1: Present Value Analysis at the Project Level

Select the most economical option from the three options given in Table 2
using the PRV AL pro gram.
(a) Using the present value analysis technique the PRVAL04 program
gives the output shown in Table 3.
The analysis in Table 3 shows option 3 is the preferred choice.

Example 2: lncremental Benefit-Cost Analysis at the Project Level

Select the most economical option from the three options given m
Example 1(a) using the COSBEN01 program.

TABLE 2
Cast data
--------- ---·-----------

Years Option 1 Option 2 Option 3

------- ---~-----~----

0 1000 200 300

5
15 100 1000
20 1000
30 200 100
35 100
45 200
 TABLE 3

Years Option 1 Option 2 Option 3

Cost PV Cost PV Cost PV

0 1000 1000 200 200 300 300

5
15 100 42 1000 417
20 1 000 312
30 200 35 100 17
35 100 13
45 200 15
Total PV 1077 649 625
Total PV* 1109 642 612

* The present values adjusted for residual value and uncertainty in costs.

TABLE 4
Total benefits

Options Net Cost Benefit

benefit (total)

1 0 1000 1000
2 458 200 658
3 489 300 789

TABLE 5
Incremental benefit-cost ratio

Options Net Cost Benefit IC IB IB/IC

benefit (total)

2 458 200 658 200 658 3·29

3 489 300 789 100 130 1·30
1 0 1000 1000 700 210 0·30

TABLE 6
Order of priority (with limited budget)

Priority Options Net Cost IB/IC

benefit

1 2 458 200 3·29

2 3 489 300 1-30
3 1 0 1000 0·30
 Cost-E.ffective Strategies in Bridge Management 723

TABLE 7
Input data

Bridge Option Net First cost IB/IC

number number benefit

1 2 950 100 10·5

1 4 1 550 500 2·5
1 5 2450 2000 1·2
2 2 1100 200 6·5
2 3 1 925 750 2·5
3 5 800 250 4·2
3 2 1400 500 3-4
3 4 1 580 800 1·6
4 2 1640 200 9·2
4 4 2660 500 4·4
4 5 2 810 1000 1·3
5 2 2300 200 12·5
5 3 3620 500 5-4
5 5 4520 750 4·6
6 3 2320 400 6·8
6 2 3020 600 4·5
6 4 3 710 900 3·3
7 2 1 725 250 7·9
7 3 2425 450 4·5
7 5 3 235 1200 1·3
8 2 4380 300 15·6
8 3 5 395 650 3·9
8 5 5 695 850 2·5
8 4 6 865 1600 1·2

TABLE 8
Output

Priority Bridge Option Net First cost IB/IC

number number benefit

1 5 5 4520 750 4·6

2 6 4 3 710 900 3-2
3 1 4 1550 500 2·5
4 8 5 5 695 850 2·5
5 2 3 1925 750 2·5
6 3 4 1 580 800 1·6
7 7 5 3 235 1200 1·2
8 4 5 2 810 1000 1·2

Budget spent = $6 750 000.

Balance= $250000.
724 R. S. Reel and C. Muruganandan

It is evident that options 2 and 3 are cost-effective options. lf there are no

limitations on the budget then option 3 would yield the maximum benefits.
The advantage of this method is that it ranks the projects.

Example 3: Network Analysis

Forabudgetlimitation of$7 000 000, prioritize the various options for the
preceding eight bridges using the COSBEN03 program.
The bridges are individually analysed to obtain the cost-effective options
as shown in Example 2 using either COSBENOl or 02 programs. These
results are used as input for COSBEN03. The program arranges them in the
order of decreasing incremental benefitjcost ratio and allocates the funds to
obtain the best cost-effective combination.

REFERENCES

1. FEDERAL HIGHWAY ADMINISTRATION, Bridgemanagement systems. Draft Report

No. FHWA-DP-71-01, Washington, DC, 1987.
2. MINISTRY OF TRANSPORTATION, Structural Financial Analysis Manual. Ontario,
1988.
3. HAVEMAN, R. and MARGOLIS, J., Public Expenditure and Policy Analysis, 3rd edn.
Roughton Miffiin Company Ltd, Boston, Chapter 12, 1983.
 63
Tension Drop in Cable-Band Bolts on Suspension
Bridges

YUJI KAGAWA
Bridge Technical Affairs Division, Honshu-Shikoku Bridge Authority,
45th Mori-Bldg, 1-5 Toranomon 5-Chmne, Minato-ku, Tokyo 105, Japan
and
AKIHIRO fUKUSHI

Bridge Engineering and Construction Department, Kobe Steel Ltd,

2-8 lwayanakamachi 4-Chame,Nada-ku, Kobe 657, Japan

ABSTRACT

The tension (axial force) in cahle-hand halts on suspension hridges decreases

with time. Conventianally holt tensian was measured periodically and holts
were retightened when the measured tension was lawer than that specified. The
authors undertook a survey of the holt tension on several suspension hridges
and analysed the data. The results indicated that the holt tension decrease has a
certain tendency and, in particular, tension drop decreases markedly after
retightening. This paper descrihes the method of investigation which the
authors used ta determine the tension in cahle-hand holts tagether with the
analysis of the data. In addition, proposals are given on how to determine the
optimum holt tightening time and an appropriate holt tension value to use as
criteria.

INTRODUCTION

Cable-bands play an important role in transmitting the tension of hanger

ropes to the main cable, and the cable-band is held to the main cable with the
tension of the cable-band bolts. The cable-band bolt tension generally
decreases with the passage of time. Consequently bolt tension control
becomes very important as it is necessary to retighten cable-band bolts
which have lost their original tension. To date, however, bolt tensions have
725
726 Yuji Kagawa and Akihiro Fukushi

been measured several times at optional intervals and retightened as

required.
The authors have carried out measurement, retightening and monitoring
of cable-band bolt tension on five suspension bridges under the control of
the Honshu-Shikoku Bridge Authority (Fig. 1 shows the twin bridges, which
are the Minami Bisan-Seto Bridge and Kita Bisan-Seto Bridge). Figure 2
shows a general view ofthe Minami Bisan-Seto Bridge, showing the overall
suspension bridge shape and the cable-band position and shape.
This paper reports the results of measuring the cable bolt tension, explains
the tendency of bolt tension to decrease with time, and proposes a concept
for determining the optimum bolt tension and a procedure for determining
the proper retightening time.

FrG. 1. Minami Bisan-Seto Bridge and Kita Bisan-Seto Bridge.

 Magnified figure
Cable band holt
~
;:::
"'6"
;:::
t"l
...,
Hanger rope .g
Truss s·
0
<::J-
~
'
Kojima Site
2'
;:::
,....,.,......_ ~

~
~
c;:::

~
{3
~
"'§"
Unit: m ~
~
~
FIG. 2. General vicw of Minami Bisan-Seto Bridge.

-..1
N
-..1
 -..1
N
00

Tension measuring system Principal specifications

Display
Meas~,o~red OOits 1~14 5 x 3. boh longth 980"""

Mu.suri!'CI l~ngch labout 980 mm

N
"" ;;:<
Un11 of measurfli l~h 1/ IOOOmm <:::::.
- ~tu.surn'IC stroke 5mm
-.
~
~
:;
$::)

§
~:)..

~
Measuring unit ;>;-
s:
::;·
0
=---t=
~:pccq]Jj~ I ~
;>;-

c-c Detail B
-.~
0
0
N

Steel ball ;2
 Tension Drop in Cable-Band Botts on Suspension Bridges 729

MEASURING METHODS AND EQUIPMENT

In the study cable bolt tension was obtained by measuring the elastic
extension ofthe bolts and converting this measured value into a tension. To
measure the elastic extension of the bolts, the electromagnetic digital
micrometer shown in Fig. 3 was used. The micrometer was equipped with a
portable microcomputer which can store data such as bolt number and
initiallength in memory on micro-cassette tapes. A length measuring flow
chart, criteria for accepting or rejecting the data and calculating equations
were input to the computer. If length is measured properly, bolt length is
immediately converted to tension and output.

FIG. 4. Cable inspection vehicle.

As a means oftransporting workers, the length measuring equipment and

other materials to the cable-band positions, the cable inspection vehicle
shown in Fig. 4 was used. 1 The cable inspection vehicle was developed by the
Honshu-Shikoku Bridge Authority and is a type of movable scaffold that
travels along the main cable using hand ropes as travelling rails. For the
drive method, in place of a conventional wire rope winching take-up system,
frictional force is generated by strongly squeezing the hand rope with two
pairs of caterpillars incorporating special rubber shoes. Travelling force is
then generated by rotating and driving the caterpillars, thereby ensuring
excellent mobility and travel stability.
730 Yuji Kagawa and Akihiro Fukushi

RESULTS OF MEASUREMENT

Tendency of Tension Drop

Table llists the bridges for which the drop in cable-band bolt tension with
time was investigated, together with the investigation time.
Figure 5 shows the relationship between the mean bolt tension and time
for each of the six bridges investigated. From Fig. 5 it can be seen that:

(1) The higher the tightening torque, the higher the residual tension.
(2) A similar tendency of tension drop with time is observed in all
bridges.
(3) The drop in cable-band bolt tension with time can apparently be
plottedas a straight line on a semi-logarithmic coordinate paper with
time as the abscissa.

Figure 6 illustrates the residual rate of bolt tension in one band of the
Innoshima Bridge. 4 This bridge, selected as an example, is a very typical case,
but virtually all bridges showed similar behaviour. The figure shows that in
one cable-band a smaller tension drop occurs with the upper bolt than with
the lower bolt, and with the edge bolt more than with the intermediate bolt.
For the Kanmon and Innoshima bridges, cable-band bolts were retightened
and later checked for bolt tension transition. Figure 7 shows the results
and proves that the rate of bolt tension drop after retightening markedly
decreases when compared with the rate before retightening.

Tightening Torque and Tension Drop Rate

Factors in the bolts' tension drop with time are thought to include:
(1) Decrease in the cable diameter caused by increase in the dead load of
the bridge.

TABLE 1
Measured bridges and investigation time ( elapsed time: days)

Name of bridge After bridge completion After retightening Year

1st 2nd 3rd 4th 1st 2nd 3rd 4th

Kanmon 2 •3 500 1200 10 800 1973

Innoshima 4 1000 1400 1800 2100 90 240 400 700 1983
Ohnaruto 1000 1 700 1985
Shimotsui-Seto 500 1988
} Not yet <etightened
Kita Bisan-Seto 600 1988
Minami Bisan-Seto 500 1988
 Tension Drop in Cable-Band Botts on Suspension Bridges 731

80

6 - · - 6 SHIMOTSUI·SETO
o--oKITA BISAN·SETO
•- -- -• MINAMI BISAN·SETO
10 •--- --• OHNARUTO
®----® INNOSHIMA
o--oKANMON

5 10 50 100 500 1,000 5,000

DAYS
F!G. 5. Decrease in mean bolt tension of each bridge (before retightening).

80 UPPER POSITIO

70~ i
60 • • ~
~ 50 •• • .CABLE·BAND
2:
~ 40
2:
w
I--
u.
0
w
1-- CABLE·BANO BOLT
<
0::
....!
:5 40 oAFTER 1.000 DA YS
9so
/Jl
w • AFTER 1.400 DAYS
0::60
70

80"--,'--;;;:;~==~~
LOWER POSmONS

FIG. 6. Comparison of residual tension rate in bolts of one band (Innoshima Bridge).
732 Yuji Kagawa and Akihiro Fukushi

80
WHEN THE BOLTS WERE RE-TIGHTENED
WHEN THE BRIDGE WAS COMPLETED

•
(/J
~ 60
--·----
•
•
g
Q
z 50
«:
o:l
w
...l
o:l
«:
u
""0z 30
0
üi
z
~ 20 ® - - ® INNOSHIMA (AFTER COMPLETED )
•---• ( RETIGHTENED)
o---o KANMON ( COMPLETED )
10 • - - • ( RETIGHTENED)

5 10 50 100 500 1,000 5,000

DAYS

FIG. 7. Reduction of mean tension in bolts before and after retightening.

AFTER RE- TIGHTENED

~
z
0
üi -----r-~--,.:=~ ..
z
~ ~~~~~~~~ "o~"-
f-. 60
.'--......
""0 A SHIMOTSU!-SETO
~
o KITA BISAN-SETO
!;;: 40
• MINAMI BISAN-SETO
c::
...l • OHNARUTO
«: ® INNOSHIMA
;:J
Q 20
o KANMON
üi
~
c::
5 10 50 100 500 1,000 5,000
(DAYSJ
0.5 1 2 3 5 10 20
(YEARSJ

FIG. 8. Residual rate of tension in cable-band bolts.

 Tension Drop in Cable-Band Botts on Suspension Bridges 733

(2) Rearrangement of the cable wires within the cable-band caused by

movement of the bridge from fluctuations in the live load.
(3) Compressed creep of the zinc layers of galvanised cable wires.
Replotting Figs 5 and 7, choosing the tension residual rate as the ordinate,
results in Fig. 8. This shows the following:
(1) Though some variations are found in each bridge, generally the bolt
tension decreases after bridge completion by about 25-35% in 500
days and 30-40% in 1000 days.
(2) The tension drop after retightening is noticeably lower than before
retightening, demonstrating that retightening has a great effect on
suppressing the tendency for tension drop to increase with time.

DISCUSSION

Effects of High Tightening Torque

Figure 8 indicates that the conditions of increasing tension drop fall into
one of two groups: firstly, those of the Kanmon Bridge and, secondly, all of
the other five bridges. Table 2 shows the differences in design. Figure 8 and
Table 2 show that tightening the bolts with high torque (stress) during both
construction and retightening effectively prevents an increase in tension
drop over a period of time.

Safety Factors and Retightening Time

The design standard of the Honshu-Shikoku Bridge Authority specifies
that the cable-band slip safety factor v must ensure v:;:::: 3 even if the bolt
tension reduces to 70% ofthat at the time of tightening. 5 •6 That is,

v = 0·7m,u 11T > 3

Th sin cp -
where .u = coefficient of friction,
11 = number of bolts,
T = force of designed bolt tension,
Th = vertical force created by banger rope, and
cp = inclination of cable at the cable-band installation point
Consequently it is evident that the proper retightening time should be
when v = 3. ludging from Fig. 8, the retightening time is about 1·5 years after
bridge completion (except for low design bolt tension, which is represented
by the Kanmon Bridge). However, a slip safety factor ofless than 3 does not
directly mean band slippage and, in reality, retightening is only required
734 Yuji Kagawa and Akihiro Fukushi

TABLE 2
Difference in holt design between Kanmon and other bridges

Name Diameter of bolts Designed tension of

(mm) bolts (t)

Kanmon 50 40
Others 45 65~74

within the period from 1·5 to 3 years after completion. lt can be seen from
Fig. 8 that after retightening any further retightening would not be required
for the time being.

CONCLUSION

The authors repeatedly measured the tension in cable-band bolts for several
suspension bridges and subsequently retightened bolts on parts of the
bridges depending on the results obtained. They also measured the bolt
tension after retightening. Analysing this measurement data indicated the
following:
(1) Retightening the bolts at the correct time definitely decreases the
tension drop rate thereafter, eliminating fears of bolt tension drop
indefinitely (for more than 10 years, or semi-permanently if
prevention of cable-band slip is the only requirement).
(2) It was also seen that tension drop is smaller when the bolt is tightened
to a high torque.
(3) With the foregoing description the authors suggest the following
methods in designing cable-band bolts and controlling bolt tension
in the long suspension bridges which will be constructed in the future:
(i) Design for as high a bolt tightening tension as possible.
(ii) Retighten bolts during the period from 1·5 to 3 years after
construction. During this period it is not at all necessary to
check bolt tension.
(iii) After retightening, check the tension at intervals of about 10
years and decide the necessity of further retightening from the
results.

REFERENCES

1. FUKUSHI, A., Cable inspection vehicle, KOBELCO Technology Review, No. 4,

Kobe Steel Ltd, Kobe, August 1988, p. 41 (in Japanese).
 Tension Drop in Cable-Band Bolts on Suspension Bridges 735

2. On Kanmon Bridge Cable-Band Retightening Works, Japan Highway Public

Corporation, Kobe Steel Ltd, March 1977 (in Japanese).
3. Investigation on cable-band tension ofKanmon Bridge, Japan Highway Public
Corporation, Kobe Steel Ltd, November 1978 (in Japanese).
4. MATSUI, T., HIRANO, S. and KANEKO, M., Shape measurement and axial force
research of cable-band bolts on Innoshima Bridge. Honshi Technical Report,
Vol. II, No. 44, Honshu-Shikoku Bridge Authority, Tokyo, October 1987,
pp. 26-34 (in Japanese).
5. Review on suspension bridge cable, Honshu-Shikoku Bridge Authority, March
1978, pp. 168-73 (in Japanese).
6. HIRAI, A., Steel Bridges (1!!), Gihodo, 1967, pp. 719-22 (in Japanese).
 64
The Design of a Flexible Surface Mix for U se at
Bridge Expansion Joints

A. R. WooosmE and W. D. H. WooowARD

Department of Civil Engineering and Transport, University of Ulster,
Newtownabbey, County Antrim BT37 OQB, UK

ABSTRACT

Leakage at bridge Joints causes corrosion and enhanced by rock-salts has

caused major structural problems in the infrastructure. Development of a
flexible surfacing material was considered to be a possible alternative to
restriet the ingress of water, i.e. the material could accommodate the changes in
movement experienced without cracking. This paper reports the work carried
out in designing a flexible surfacing mix known as 'H-MAC'. It initially
summarises the requirements of bridge Joints and the use of conventional
surfacing methods. Using the limitations of existing materials and the
structural requirements of the Joint, the development of the flexible mix is
outlined. Camparalive Iabaratory testing is discussed using various hinder
types and aggregate gradings. These resulted in a single sized 6·3 / 3·35 mm mix
using 23% Icosit Membrane H as a binder. Bridgejointtrials were carried out
using this prolotype mix and the results reported.

INTRODUCTION

In the UK roads network there are about 50 000 concrete bridges. These can
be considered as being of two groups~those built before World War Two
and those built after. The first are mostly 50-60 years old, reinforced with
plain mild steel and have simple structural forms. The second and more
recent group are constructed with high strength materials and many have
more complex structural forms demanded by !arger spans and more
ambitious requirements.
For the 90 years or so since they were introduced concrete bridges have
737
738 A. R. Woodside and W D. H. Woodward

performed well and there were few problems until the 1970s. During the
1960s the use of rock-salt was introduced for deicing. lt was not until some
years later than the deleterious effect of salting became apparent through the
development of corrosion ofthe steel reinforcement bars and carbonation of
the concrete.
These problems tend to occur more often with modern bridges. Studies
have shown that it is not due to modern concrete being less durable. Instead
modern bridges on motorways and trunk roads are designed with more
ambitious structural forms and so are plagued by leaking expansion joints,
resulting in piers and crossheads exposed to chlorides in a manner not
experienced in many of the older bridges.
The leakage of bridge joints causing corrosion and enhanced by deicing
salts has created a major problem. Considerable research is now being
carried out into early detection and preventive methods, and repair of
affected bridges.
Development of a flexible surfacing material was seen as a possible
alternative to restricting the ingress of water, i.e. the material could
accommodate the changes in movement experienced without cracking. This
paper reports the work carried out in developing such a surfacing and
known as 'H-MAC'.

FAlLURE OF SIMPLE EXPANSION JOINTS

Expansionjoints are used in bridges to accommodate movements due to

deformation of the bridge deck caused by changes in temperature, the
passage of traffic, and in concrete structures, creep and shrinkage.
The type ofbridge joint used depends on the structural materials as weil as
the span of the bridge. In Northern Irelandmost of the bridges requiring
expansionjoints are made ofreinforced concrete and have individual spans
of less than 20m. For this type of structure simple expansion joints can
accommodate the full range of expansion and contraction under normal
conditions.
A simple expansion joint consists of a vertical gap between the end of a
section of a bridge deck and an abutment, or between two sections of deck.
The end of the deck is therefore able to move horizontally and the gap will
open and close accordingly. The expansionjoint is usually covered by a layer
of paving material to provide a continuous smooth surface for vehicles and
pedestrians. It also prevents the ingress of water and dirt.
However, it has been found that normal bitumen-based surfacing
materials deteriorate and must be replaced within a period of 1-3 years. This
is caused by the constant movement across the joint and probably occurs
 Design of Flexible Surface Mix for Bridge Expansion Joints 739

FIG. 1. Typical highway bridge. (Detail shows expansion joint where pavement has
cracked.)

most rapidly at lower temperatures when the bitumen becomes stiff and
brittle.
Alternativematerials such as polymer modified bitumens have been used
instead of bitumen as the bioder but they areexpensive and must be applied
at carefully controlled temperatures.
Figure 1 illustrates a typical reinforced concrete bridge with expansion
joints at either end. lt can be seen how surface water, which may contain
corrosive materialssuch as de-icing salts, is able to penetrate down into the
structure so leading to unsightly staining and deterioration of the concrete
structure.

MOVEMENTS EXPERIENCED AT SIMPLE EXPANSION JOINTS

Movement at expansionjoints is mainly due to cyclic thermal expansion and

contraction, and occurs throughout the life of the structure. Shrinkage and
creep will also cause movement in concrete structures but 50% of the total
movement will probably occur within the first month before the joint is
covered. In addition, vertical and rotational deflections may be caused by
740 A. R. Woodside and W. D. H. Woodward

vehicles crossing the bridge but these are generally small enough to be
ignored.
Simple expansion joint systems may be designed for a 20-year return
period as they would have to be replaced if the highway was resurfaced.
The movement and corresponding temperature changes for a bridge
constructed over the M5 motorway in England was recorded. Movement
occurred in daily cycles; expansion during the day and contraction at night.
Daily movements of 2 mm were common and the surfacing had to
accommodate movements of up to 4 mm over a week. Also much larger
movements occurred over an annual cycle with 12-4 mm of movement
recorded between February and June.
In developing a flexible surfacing, movements such as these would have to
be accommodated if cracking is to be avoided.

DEVELOPMENT OF A FLEXIBLE SURFACING

In developing a flexible surfacing, the following criteria had to be met:

(i) Accommodate changes in expansion of ± 5 mm or greater.
(ii) Provide a durable impermeable surface.
(iii) Be relatively easy to manufacture and lay.
(iv) Be cost effective.
(v) Be mixed at ambient temperatures (5-25°C).

The investigation may be summarised as follows.

Materials
Aggregate-For all mixes Castlenavan Quarry gritstonewas used.
Binder-(a) 200 hinder supplied by Tennants Tar Distillers.
(b) Icosit Membrane H-a hand-applied two-pack solvent-
free fast-curing polyurethane resin combination used for
the production of elastic crack bridging coatings.

Types of Mix

Code Binder(%) Binder type Mix type

BS.l/234 5·1 200 pen 14 mm dense macadam
HS.l/234 5·1 Membrane H 14 mm dense macadam
H7.1/234 7·1 Membrane H 14 mm dense macadam
H7.1/235 7·1 Membrane H 14 mm open macadam
H23/3.35 23·0 Membrane H 6·3/3-35 mm single sized
 Design of Flexible Surface Mix for Bridge Expansion Joints 741

Mannfacture of Test Mixes

For testing purposes slabs 40 x 90 x 300 mm were made. This size of
mould is used for the wheel tracking test and was thought to be suitable to
simulate in the laboratory a model bridge joint.
All moulds manufactured were allowed to 'eure' for 14 days before testing.
An exception was the 6·3/3·35 mm single sized wheel tracking mould, which
was tracked after 23 h in order to simulate early trafficking.

Types of Test
Model Joint lest
This is a non-standard test in which the prepared specimen is bonded to
steel plates through which load is applied on an lnstron machine and pulled
apart at a constant rate. Test carried out at 20oC (Fig. 2).

Flexural lest
This is a non-standard test in which the specimen is clamped at both ends
and a vertical load is applied at the centre of the span by a Dartee until
failure. Test carried out at approximately 20oc (Fig. 3).

Wheel tracking lest

This is a simulation test in which a rubber wheel is repeatedly dragged

lO

e H23/3.35

4
•BS.l/234

e H7 .l/235
2 • H7.l/234
• HS .l/234
- Strength (N/mm) +

10 20 30 40
FIG. 2. Elongation plotted against strength for the modcl joint tcst.
742 A. R. Woodside and W D. H. Woodward

40
e H23/3 .35

t
30

I':
0
·..l
tm 20
• B5.1/234
.-1
.....
m
0

10
' H5.1/234 • H7.1/234

_ Strength (N/nun) +
~~--~--r-~--~~--~
2 4 6 8 10 12 14
FIG. 3. Deflection plotted against strength for the flexural test.

across the surface ofthe specimen contained in a water bathat 40oc to assess
susceptibility to concentrated trafficking or rutting.
Testing
The following tests were carried out:
Code Modeljoint test Flexural test Wheel tracking test
B5.1/234 j j j
H5.1/234 j j
H7.1/234 j j
I
H7.1/235 v -

H23/3.35 j j j
---------

Note: j denotes test carried out.

Results
Model Joint test
The model joint test results are shown below:
Elongation Stiffness
Strength (d uctili ty) (secant modulus)
Code (N/mm) (mm) (N/mm 2 )
---------

B5.1/234 0·6 (1·0) 3-4 (1·0) 0·18 (1·0)

HS.l/234 15·9 (26·5) 1·3 (0·38) 12·2 (67·9)
H7.1/234 27·6 (46·0) 1·8 (0·53) 15·3 (85·2)
H7.1/235 37·0 (61·7) 2·5 (0·74) 14·8 (82·5)
H23/3.35 22-4 (37-3) 8·9 (2·62) 2·52 (14·0)
-----
 Design of Flexible Surface Mix for Bridge Expansion Joints 743

Notes: The elongationisthat occurring between the two backing plates,

which were 50 mm apart.
The figures in parentheses have been normalised to results for
B5.1/234.
The following observations may be made:
(a) If the Membrane H-based macadam mixes are compared to the
200 pen/ 14 mm dense macadam mix they are 26-62 tim es stronger
and 68-83 times stiffer at 20uC (testing temperature).
(b) For Membrane H, elongation and strength are increased by 38% and
74% respectively if the hinder content is increased by 2% for the
14 mm dense macadam mixes.
(c) For Membrane H, elongation and strength are increased by 39% and
34% respectively when an open textured macadam (7·1% binder) is
used instead of a dense macadam (7·1% binder).
(d) The single sized 6·3/3-35 mm mix with a hinder content of 23% gave
an elongation of 8·9 mm, which is 2·6 tim es that of the 200 pen
macadam, a strength 37 times stronger and a stiffness 14 times
greater.

Flexural test
The ftexural test results are shown below:
--------------- ------- --------~

Stiffness
Strength Deftection (secant modulus)
Code (N/mm) (mm) (N/mm 2 )
~-------
-------

B5.1/234 0·74 (I) 17·9 (1) 0·041 (1)

H5.1/234 4·5 (6·08) 6·0 (0·34) 0·75 (18·3)
H7.1/234 9·82 (13·3) 6·6 (0·57) 1-49 (36·3)
H23/3.35 13·9 (18·8) 37·5 (2·09) 0·37 (9·02)
- ----------

Note: The figures in parentheses have been norma1ised to results for

B5.1/234.
The following observations may be made:
(a) If the Membrane H macadam mixes are compared to the 200 pen/
14mm dense macadam mixes they are at least 6-13 times stronger
and 18-36 times stiffer.
(b) For Membrane Ha 5·1% 14 mm dense macadam is 2-2 times stronger
and gives 10% more deftection than a 7·1% hinder mix.
(c) The 6·3/3·35 mm single sized mix with 23% hinder gave a deftection
of 37·5 mm before failure and was 18 times stronger than the 200 pen
hinder macadam.
744 A. R. Woodside and W D. H. Woodward

Wheel tracking lest

The 6·3/3·35 mm single sized 23% Membrane H mixwas compared to the
200 pen/14 mm dense macadam. The comparison test was carried out in its
most severe form, i.e. at 40°C with the tyre trailed across the surface. The
results are as follows:
Code Test assessment
BS.l/2.34 Rut 10-20mm resulted
H23/3.35 Rut < 1 mm-no noticeable deterioration
Note: The Membrane H mixwas tested after 23 h curing.

DISCUSSION OF DEVELOPMENTAL WORK

When Membrane H was used as the hinder substantial improvements in

strength and stiffness occurred in comparison to the 200 pen mix. This would
produce stronger bridge joints. However, a flexible bridge joint surfacing
must accommodate expansive changes andin this respect the Membrane H
macadams could not provide similar Ievels of elongation and deflection. In
practice, cracking of a traditional surfacing is more probable at low
temperatures due to embrittlement of the binder. As Membrane H is not
affected by temperature change then, despite the lower Ievels of elongation
and deflection, the stronger mix may give a Ionger Iasting bridge joint.
In combining strength, stiffness, elongation and deflection the single sized
6·3/3·35 mm mix by far out-performed the macadam type mixes. An
elongation of 8·9 mm occurred across the 50 mm distance between the test
backing plates with an elongation of approximately 50-lOOmm occurring
before complete separation.
Other notes of interest with this type of mix were:
(i) Single sized chippings, i.e. no grading problems.
(ii) Required minimal mixing.
(iii) Required no compaction other than levelling by hand.
(iv) Provided a high Ievel of texture depth.

ROAD TRIALS

Full-scale road trials were carried out during November 1988 as part of a
routine maintenance programme on the Maghery to Portadown bridge
crossing the M1 motorway in Northern Ireland.
Prior to laying the single sized 23% H-MAC material the bridgejointwas
sealed with ICOSIT Membrane H, a primer applied to improve the bond
 Design of Flexible Surface Mix for Bridge Expansion Joints 745

between the sealant and the H-MAC, and a debonding agent 25 mm wide
added to increase the movement across the joint.
H-MAC was mixed in batches, hand laid and levelled. A layer of fine
aggregate (3 mm gritstone) was added to improve skid resistance. After 2
days of curing the traffic was allowed over the joint.
Of the problems which arose the most apparent was that due to the
material's low viscosity, consequently there were considerable handling
problems associated with mixing and laying.
After 10 weeks the jointwas inspected and it was obvious that the material
had failed. In comparison with Iabaratory trials at 20oc the bridge joint trial
material showed poor cohesion properties. This illustrated the infiuence of
environmental factors in the 10-week period after the trial when the air
temperature rangewas + 6·9 to - 7·0°C, with the day ofthe trial bitterly cold
with a strong wind and high chill factor. lt is thus recommended that the
bridge be allowed to eure in temperatures exceeding + soc.
lt was obvious to the authors that the environmental conditions did not
allow adequate curing to occur. This was also evident from wheel tracking
tests which failed on material extracted from site and cured under similar
conditions.
Since the trials patehing of similar material has been carried out under
more favourable environmental conditions, and this appears to have cured
more effectively and consequently is performing perfectly on site.

CONCLUSIONS

In comparison with other binders (i.e. 200 pen in a 14 mm macadam)

Membrane H performs more effectively than other conventional flexible
surfacings. Developmental work has shown that this type of hinder may be
used to provide a superior flexible surfacing over bridge expansion joints.
However, the traditional prohibitive factor of cost must be viewed with
respect to the potential darnage to the bridge structure as a conventional
bituminous joint may Iead to corrosion problems at a later date and thus
shorten the effective life span of such a structure.

ACKNOWLEDGEMENTS

The authors wish to acknowledge the help and cooperation given to them by
Sika Ltd, the Department of the Environment (NI) Roads Service, Mr E. T.
Stewart and Mr P. Brown of Larsen Associates, and Mr Brian McDowell,
without whose help this research could not have been carried out and this
subsequent paper produced.
 65
The Repair of a Composite Concrete-Steel Bridge

P. H. M. WouTERS and C. WARNON

BESEM,
Ministry of Public Works,
Bridge Office, rue Cote d'Or 253-4200 Liege, Belgium
Bridge Office, WTC 3, Blvd S. Bolivar 30-1210 Bruxelles, Belgium
Roads Direction, Au. Gouverneur Bauesse 37/31-5100 Jambes, Belgium

ABSTRACT

Bridge 23 is an overpass over the E42 motorway on the raute from Dunkirk
(France) to Wurzburg (Germany) and is situated near the interchange of
Daussoulx in Belgium. The bridge, built in 1969, belongs to a standard series of
composite concrete-steel structures designed to pass above a motorway of
40 m wide. In 1984 the superstructure was crashed into by a trailer of a lorry,
which suddenly came loose and slightly rase from the carriageway, thus hitting
the girders of the bridge crossing the motorway. The darnage was considerable:
the.first two main girders were irreparably deformed. The Road Department,
which had commissioned the building of the structure, decided on a complete
repair of the bridge by replacing all damaged elements. This repair was in line
with what could be claimed from the insurance company: a restoration of the
bridge to its original state before the accident. Moreover, after replacing the
damaged girders, the entire steel construction was protected against corrosion
by using a method based on metallising and painting in situ.

DESCRIPTION OF THE DAMAGE (Fig. 1) AND CONCLUSIONS

AFTER THE ACCIDENT

Concrete Bridge Deck

After inspection ofthe bridge deck, no visible darnage to the concrete as a
result of the accident was found. In the corbellings there were many old
cracks through which water was seeping.

Main Girders
The first two main girders at the east side were considerably deformed
747
748 P. H. Besem, M. Wouters and C. Warnon

11 Cover plates
·A_
P5
I P4
P3
DG DG G P2
~G G+DF OG+DF
P1
Crack in the medium weld-'1 1..i ~Crack in the weid und er
of the cross-girder Deformed zone the cover plate

Legend
Cover plate + fasteni ng of
P4 p5 the cross girders
D =unsticking
G:sliding
DF =deformation
* = impact point
FIG. 1. Description of the damage.

from the pier onwards to the middle of the structure. The deformation was
both axial and vertical. The cover plates near the point of impact had slid
and loosened from the web of both girders, thus causing the cracking of the
welding seams situated at the lower part of the cover plates.

Cross Girders
The six cross girders between girders 1 and 2 and 2 and 3 were heavily
deformed. The second cross girder between girders 1 and 2 was broken near
its contact with girder No. 2. There were cracks in the welding seams in the
middle of the third cross girder between girders 1 and 2.

Fastening of the Cross Girders to the Main Girders

The cross girders were fastened to the main girders by means of T-irons
bolted on these girders. Most ofthese T-irons had slid away from the web of
the girders. Near the most deformed areas of the main girders these T-irons
were also seriously deformed and they had loosened from the web of the
girders.

Hearings on the Piersand on the Northern Abutment

The bearing of girder No.l on the northern pier had undergone a
displacement to the centre of the bridge as well as to the west side. The
bearing of girder No. 2 on the northern pier had undergone a horizontal
rotation towards the west. The corresponding bearings on the northern
abutment had undergone a differential settlement on their southern side but
there were no definite indications that these failings were due to the accident.
 The Repair of a Composite Concrete-Steel Bridge 749

CHARACTERISTICS OF THE STEEL STRUCTURE

Bridge 23 belongs to a series of composite concrete-steel bridges designed to

pass above motorways 40 m wide. A study was made of these structures in
order to standardise their elements as much as possible. Bridge 23 has five
continuous main solid-web girders, placed at 2·71 m intervals, resting on
four bearings, braced by lattice work and supporting a concrete slab of
18 cm thick contributing to the region of positive moments. The connection
is made by means ofpistol-welded dowels. The bearing distances are 22--41
and 22m for a total length of 85 m.
The height of the main girders varies linearly over 7 m on each side of the
piers. These main girders have two assembling joints situated in the centre
span at 13-50m on both sides of the axis of the structure. The bracing
consists of 14lattice cross girders, six in the centre span, two in each side span
and one on each bearing. Temporary wind bracing was placed when
mounting the structure. The building of the structure has taken place in six
successive phases:
1. Building of the infrastructure.
2. Mounting of the three sections of the main girders on the piers,
abutments and temporary support in the middle ofthe centre span, and
fitting of two bolted joints by means of high-tensile bolts (HT bolts).
3. Control ofthe stresses in the steel girders by a 20cm variation ofheight
of the intermediate supports.
4. Casting of the concrete slab.
5. Removal of the central support and the assembling wind braces.
6. Placing of a waterproofing layer-surface dressing--cut stones-
bridge rail.

STUDY OF THE REPAIR WORKS

When studying the repair works, two solutions were considered:

First solution: restoration of the structure to a state as close as possible to

its original state by replacing the damaged parts (main girders and cross
girders). To do this it was necessary to demolish and replace a considerable
part of the slab and the equipment (cut stones, bridge rail).

Secondsolution: reinforcement ofthe structure by bracing the lower girder

flange of the girders and by replacing the damaged cross girders only. The
first solution required large-scale and expensive works, necessitating a
temporary interruption ofthe motorway traffic during certain phases ofthe
750 P. H. Besem, M. Wouters and C. Warnon

work (dismantling and reassembling of the steel elements, for instance). The
second solution, on the other hand, would only have required traffic
interruptions in one direction. This solution, however, would have
amounted to reusing a structure containing damaged but repaired elements,
and therefore one had tobe absolutely sure that the remaining parts were
still in good condition (particularly the slab, dowels, weldings, assernblies
and bearings on the northern pier).
As the second solution involved more risks and because the consequences
of the accident were covered by the insurance, the department decided on the
first solution. This solution thus consisted in restoring the structure to a state
as close as possible to its original state (before the accident).
Therefore it was necessary to demolish the part ofthe structure containing
the slab at the east side over a width of about 5 m and to replace it by new or
reused elements. To do this the same scheme had to be followed as for the
original works. During the repairs certain stress regulations had to be
carried out as well.
The study ofthe repair works resulted in a working scheme containing the
following phases:

1. Building of a closed working platform above the carriageway with a

clearance 9·50m wide and 4·30m high for each traffic lane.
2. Removal of the equipment (bridge rail and cut stones which were to
be used again) over the whole length of the bridge and removal of the
expansion joints.
3. Removal of the surface pavement and the waterproofing layer over
the whole length of the bridge.
4. Unbolting andremoval of the cross girders between girders P2 and
P3.
5. Demolishing the slab over a width of 5·20 m at girders Pl and P2
over the whole length of the bridge, while taking care not to darnage
the dowels and the reinforcing bars coming out ofthe remaining part
of the slab.
6. Placing a temporary support under girders Pl and P2 (which were to
be replaced) in the longitudinal axis of the motorway and raising of
the girders Pl and P2 with 20 cm.
7. Unbolting ofthejoints ofthe web and boom plates ofthe girders Pl
and P2.
8. Removal of the central part of the Pl-P2 unit and removal of the
northern part (girders and cross girders) of the Pl-P2 unit.
9. In the workshop fabrication of the new girder sections and the new
cross girders. Control and restoration of the non-deformed cross
girders.
 The Repair of a Composite Concrete-Steel Bridge 751

10. Sub-assembly in the workshop of the central and side units of

girders P1 and P2, and the cross girders.
11. Placing of the central unit and bolting to the remaining southern
sections of P1 and P2.
Placing ofthe northern side unit. Connection by bolts ofthe several
elements. To assemble the different parts the girders were placed 20 cm
higher than their finallevels at the three piers (two definite piers and
one temporary pier placed in the central reserve ofthe motorway) and
afterwards those three supports were lowered by means oflong-travel
jacks installed on the temporary pier. A temporary wind bracing was
placed between P2 and P3.
12. Casting of the slab, after shuttering and reinforcement, while
providing a longitudinal joint in the middle of the slab between P2
and P3.
13. Removal of the central support after the concrete had hardened.
14. Mounting of the cross girders between P2 and P3.
15. Concreting of the longitudinal joint of the slab between P2 and P3.
16. Casting, placing and installation of the equipment: waterproofing
layer, protective layer, surface dressing, cut stones, bridge rail,
expansion joint, water diversion.
17. Removal of the working platform and putting the site back in order.

lt is worth mentioning that the composition of the new girder elements is

somewhat different from the original composition. This choice was based on
certain technical requirements. However, this has resulted in only a few
differences with regard to the stresses and in some minor adaptations with
regard to the cambers.
During the works, as the new concrete had not yet stopped creeping or
shrinking, a temporary overloading was placed on the new part before the
Iongitudinaljoint ofthe slab was cast, and before the cross girders were fixed
between P2 and P3 in order to compensate these effects.

EXECUTION OF THE WORK

Traftic Problems on the Motorway

When organising the repair works, specialattentionwas paid to the traffic
situation:

- The E42 is a very busy motorway with an average daily traffic of 17 900
vehicles, of which 22% are lorries.
- The bridge is situated at a few hundred metres of the motorway
752 P. H. Besem, M. Wouters and C. Warnon

interchange of Daussoulx between the E42 and the E411 Brussels-

Luxemburg. Near the bridge the traffic fiows of both motorways
converge together.
- The bridge is situated north of the town of Namur. The possible relief
routes guide the traffic either to rural roads with insufficient capacity or
through the centre of the town where the roads are already saturated.
The department therefore decided to maintain the traffic on 2 x 2 lanes
during the works, except during the dismounting of the damaged girders and
the mounting of the new girders. Tothat end the contractor built a closed
working platform preventing the traffic on the overpass and enabling the
works to be carried out.
This working platform was built as a walkway resting on scaffolds placed
in the central reserve and on the hard shoulder.
For the dismounting and mounting of the four sections of the main
girders, the traffic on the motorway was interrupted for 30 min under control
of the Gendarmerie (state police force). It would indeed have been too
dangeraus to transport girder elements of 8-10 t at the end of the jib of a
crane above moving vehicles. The traffic was interrupted during off-peak
hours.
Technological Conception of the New Steel Elements
Technological evolution
Nowadays the assembling methods which were applied when the initial
structure was built are not used any more. As one-third of the main girder
had stayed intact, the dimensioning of the new elementswas inftuenced by
the global dimensions and the thickness of the remaining parts.
The original main girder was made of a web plate in AE26 or AE36 steel,
according to the different parts and assembled by butt-joints to half-sections
in AE26 steel (upper fiange plate) and AE36 steel (lower ftange plate) which
were cut out of rolled steel HE beams.
The current conception is rather based on a connection by fillet weldings
of the web plate to the two ftange plates. The modern quality of the fillet
weldings made it possible to adopt this solution whereas butt-joints were
preferred before because of the possibility of examining the composition by
X-rays.
The plates also have a slightly different thickness. Because of the rather
small quantities required for the repair works it was decided to standardise
the products in order to Iimit the delivery period.
Figure 2 shows the concept of the connection between the remaining parts
and the new ones. The differences in thickness in the joints made it necessary
to use packing, sometimes at both sides of the joint, in order to maintain a
minimum packing thickness of 4 mm.
 The Repair of a Composite Concrete- Steel Bridge 753

112 HE 24A 11 2 Flange 240 x 12

-·++
- *~....
Existing part + +: .-
:t+
-.
I
-+
I
~ keee._ tf..:..j; Part to bu ild

I +~..
+- ... ;......
Flanges thickness
4mm
Web thickness
13mm
! !,.:+._
:+- : Web12mm
thickness

I +- ... ....... I
+ .. : .. .. r,
.... :....+
112 HE 32M t- + : t Flange
I r1
1
.+.:+~+ + 300x35
- .... ~ +
Packing thickness 6mm Packing thickness 11 mm
..!..L.&
FIG. 2. Connection between remaining and new parls.

Choice of the materials and assemblage concepts

All the welded steel which was used belonged to the qualities AE235D or
AE355D according to the Belgian NBN A21-101 standard.
All workshop assernblies were carried out by applying semi-automatic
and automatic welding methods involving the use of basic products. As
required by the regulations ofthe Ministry ofPublic Works, all welders and
operators were chartered in advance for the entire work. To be chartered it
was necessary to prepare standard test pieces while using the same steel types
and welding products as were to be used in the actual repair works, thus
allowing the department to check the exterior Iook of the joint and its
mechanical characteristics.
With regard to the automatic welding method, these test piecesalso served
to check the characteristics ofthe actualjoint. All steel types were tobe dried
or preheated before the welding when AE355 steel was used.
The basic products were kept in heated chambers until use. All the
assemblages in situ, those providing the connection with the remaining parts
as well as those between the new elements, were carried out according to the
new NBN E27 -071 to 073 standards by using HT bolts of the 10·9 quality. To
secure the bolts, the couple method u was applied. It is worth mentioning
that it was the first time in Belgium that HT bolts were used, of which the
nuts were covered in the factory by a lubricant based on teflon.
This method has the advantage of reducing considerably the value and the
dispersion of the friction coefficient k in the formula Ma = kdPv, enabling
determination of the screw-holding power Ma (Nm) according to the
diameter of the bolt, d (mm), and the tension force which will have to be
exerted, Pu (N). This force generally amounts to about 0·7 times the value of
754 P. H. Besem, M. Wouters and C. Warnon

the (lower) yield point of the bolt steel. Thanks to the teflon, the value goes
from 0·18 ± 0·02 to 0·10 ± 0·005.

Description of the Main Phases of the Repair Works

The works began with the removal ofthe equipment and the demolition of
the surface dressing and the waterproofing layer. To do this the bridge was
freed of all its dead Ioads and cut lengthwise into two separate decks. The
cross girders were unbolted, but because of the welding seams it was
impossible to remove them. Therefore the hindering seams were gouged.
The slab from the damaged part ofthe bridgewas demolished by means of
a concrete breaker. The parts of the slab which were to be demolished were
cut out in reetangular pieces, the reinforeerneut bars were burnt through
with a torch and the concrete blocks were removed. Near the upper
boom plates the remaining concrete was removed by means of a pick
hammer in order not to darnage the dowels of the girder elements, which
were to be used again.
The two deformed girders were jacked in the axis of the motorway, thus
bringing both elements to rest. After unbolting the cover plates, the girders
were lifted and placed on the remaining deck, then cut into transportable
blocks and disposed of. After having removed the girders, the contact
surfaces ofthe 20-year-old HT bolt assemblages were examined. lt appeared
that the contact surfaces had resisted well to corrosion. Little rust was
discovered, especially in the support zones of the washers (Figs 3 and 4).

FIG. 3. HT bolt assemblage. FIG. 4. Cross girder connection.

 The Repair of a Composite Concrete-Steel Bridge 755

Moreover, the contact surfaces had been painted before the cross girders
were connected. The behaviour ofthe paint appeared to have been excellent.
At the same time the new girders were fabricated in the steel workshops.
They were transported to the building site and placed on five supports before
being assembled as continuous girders by means of cover plates. Because of
the small weight of the structures, it only took 30 min to put the elements in
their right places and to assemble them. Next the centraljack was lowered by
20 cm, causing the supports on the pier to come down 20 cm as weiL This
prestressing by deformation ofthe girders caused stresses opposing the dead
Ioads in the critical sections.
The orthotropic slab was cast between the two girders, on the verge
corbels and by a corbelling between P2 and P3, in successive castings of
about 15m long and beginning at the edges.
Next the temporary central support was removed, the cross girders
between P2 and P3 were replaced, and the longitudinal joint of the slab was
concreted after partialloading of the bridge in order to adjust the Ievels of
the two decks. The existing concrete deck was repaired by means of a
modified hydraulic mortar reinforced with synthetic fibres. This repair
product is characterised by a good adhesion power to the existing concrete
and by an excellent compressive strength (50-60 MPa).
The whole bridgedeck was cleanedunder pressure with water in order to
remove the non-adherent particles and the remains of the waterproofing
layer. The deck wasmadewaterproof by means of an impregnating varnish
covered by a prefabricated asphalt membrane welded to the support by a
torch ftame. This membrane consists of a polyester reinforeerneut
impregnated with filled elastomer asphalt covered with an anti-adherent film
and with mineral aggregates. Protection of the waterproofing layer is by a
bituminous mix covered by a wearing course of the same type.
The repair works also affered the possibility to replace the neoprene
supports, of which the hoop reinforcements had deteriorated by corrosion,
by new supports completely encased with rubber; to renew the anchorages of
the bridge rail by galvanised all-thread rods sealed with modified hydraulic
mortar; to repair the spallings of the concrete on the borders of the slab and
to hide the repairs by means of an acrylic paint; and torepair the blue stone
boards under the bridge rail by glueing the broken elements by means of
epoxy gluc.
Finally, a waterproof expansion joint was placed at the mobile support
side and a 'Thorma' joint at the fixed support side of the bridge. All
equipment of the structure were then put back in place.

Anti-corrosive Profeetion of the Structure

Introduction
U ntil a few years ago all steel structures of the Ministry of Public Works
756 P. H. Besem, M. Wouters and C. Warnon

received, after an abrasive material blasting, an anti-corrosive protection

consisting of four layers of so-called 'formula paints'. 4
The composition of these paints was decided by the department, which
consequently guaranteed they would last well. Although those products had
been very satisfactory, more and more 'performance paints' are used
nowadays. In this system the composition and the performance are
guaranteed by the supplier.
The use ofthese paints is normally subject to an insurance policy covering
the damages on the structure for a specified period. 5 By this policy the anti-
corrosive behaviour, the adhesion, the non-chalking, the non-development
of cryptogamic substances as well as the preservation of the final hue must
be guaranteed.

Inspection of the existing structure

When the structure was built all steel surfaces were protected by a
painting method basedonred lead, basic lead carbonate (white lead) and ftax
seed oil.
After 18 years the existing paint had become completely porous, the
structure was covered with rust pits and large steel surfaces were completely
oxidised. This situation was discovered after an analysis by electronic
microscope scanning (EMS) of old paint samples. Semi-quantitative analysis
on a sample of existing paint indicates the nature of the discovered
substances and gives a close estimation of their weight percentages.
As there are no references in that field, the results should be considered as
an indication (Tables 1 and 2, Figs 5 and 6).

Study of the costs of a protection system

Before choosing between a 'formula' painting method and a 'perfor-

TABLE 1 TABLE 2
Paint face Steel face

Substance Electronic lay Wt% Substance Electronic lay Wt%

Na K 2·507 Al K 3·171
Al K 2·760 Si K 9·733
Si K 9·597 Pb M 22·137
p K 0·902 K K 0·417
Pb M 72·986 Ca K 0·520
K K 0·233 Ti K 0·802
Ca K 0·748 Mn K 0·313
Ti K 7·954 Fe K 31·537
Fe K 2·313 Zn K 0·463
Pb L 30·908
 The Repair of a Composite Concrete-Steel Bridge 757

FIG. 5 FIG. 6

mance' painting method it is, of course, necessary to compare the investment

cost as weil as the durability and the affered guarantees.
The investment cost of the considered performance method (total
thickness 240 J.lffi) amounts to about 150% of the cost of the method
consisting of applying four layers of formula paints (total thickness 160 J.lm).
The first method offers a guarantee of 20 years compared to a probable
durability of 5- 10 years before the first restoration for a traditional method.
Supposing that the traditional system requires (1) little maintenance
(touching up ofthe upper layer) every 5 years and a complete renewal after
15 years, (2) a repainting of the upper layer after 20 years with the
'performance painting' method and (3) an annual interest rate of 8% for the
invested sums, we conclude that after 10 years the method used in Daussoulx
becomes more economical.
A saving of 20% is made after 20 years. Moreover, the lower frequency of
the maintenance is a considerable advantage with regard to the comfort of
the road users.

Description of the chosen protection method

The following method for the maintenance of the structure was chosen:
- abrasive material blasting;
- metallising with Zn-Al (the minimum thickness of the metallayer is
80).lm);
758 P. H. Besem, M. Wouters and C. Warnon

--application of a 20 .um layer of epoxy paint diluted to 30% and

pigmented with micaceous iron oxide containing an amino hardener;
-application of a 60 .um layer of the same paint but non-diluted; and
-application of an 80 .um layer of polyurethane paint with two
components. (An acrylic polyurethane resin interwoven with an
aliphatic polyisocyanate resin used as a binder).
The choice of the method was based on the following arguments:
-taking into consideration the rusty state of the structure and the
conclusions of the above-mentioned EMS analysis, complete stripping
was necessary;
-the choice of the Zn-Al alloy was justified because of a better impact
resistance; and
-the aim of the application, very shortly after metallising, of the first
epoxy layer was to fill the porosities left after metallising in order to
prevent humidity inclusion; it also guaranteed good adhesion of the
first layer.
Considering these needs (filling and adhesion), the paint formula had to
comply with the following characteristics:
-an appropriate solvent balance;
-a well defined pigmentation with high volume concentration; and
-a sufficient binder percentage.
The result was good wettability of the metallised surface and excellent
filling power.
The first layer of 20 .um was covered with a second 60 .um layer of the same
paint in order to form an intermediate layer of 80 .um (dried coat) of epoxy
paint with a base of micaceous iron oxide, ensuring a good protective
coating. Indeed pigmentation with a base of micaceous iron oxide forms a
barrier against the permeation of humidity, which constitutes an essential
quality of an anti-corrosive painting method.
The finishing coat had, in addition to good chemical resistance, to comply
with aesthetic and mechanical standards, namely chalking resistance and
colour stability. Moreover, the chosen polyurethane formula allowed good
spray painting and brushing in situ, without blistering. lt also guaranteed
good coverage of the edges.
lt may be useful to repeat the main principles of the metal spraying
(metallising) method with pure zinc or Zn-Al alloy. This anti-corrosion
protection method, also known under the name of schoopage, consists of
melting the metal wires and spraying the melted metal by means of a fiame
spray gun; this method can be used in the workshop as well as on the
building site, regardless of the type of elements or their dimensions, and can
be applied on both new and old steel structures.
 The Repair of a Composite Concrete-Steel Bridge 759

Metallising is always preceded by abrasive material blasting (sand,

corundum, steel shot, cinders) of the steel surfaces, on the one hand, to
eliminate all traces of rust and impurities in order to obtain a high degree of
cleanliness and, on the other hand, to give them a degree of rugosity
enhancing the adhesion of the coating.
As metal wires are used for spraying, the zinc or the zinc alloy (85-15)
starts melting at the level of the spray gun nozzle under the influence of a gas
torch, the fuel being propane or acetylene and the combustible being oxygen.
Thus the melted metal is shot by compressed air from a distance of 15 cm
onto the steel target and the droplets solidify when they hit the target. This
way a structured and consistent coating is gradually formed by successive
sweepings along the surface which has to be coated. The Zn or Zn-Allayer
thus obtained after metallising is slightly porous and rough enough to allow
good adhesion of the paint layers.

Technical requirements of the Ministry of Public Works

Basic products: The instructions ofthe Ministry ofPublic Works 5 dictate
that the paint supplier must guarantee certain qualities for each layer as weil
as for the whole treatment. The method used is tested in advance for
confirmation by the department. The tests are carried out by the Bridge
Office of the Ministry of Public Works and they are confined to
technological tests on small plates covered either with one layer or with all
the layers of the protective treatment. Table 3 gives a survey of the results
obtained on test plates and of the results announced by the paint factory.
Application: Both the meta! spraying and the application of epoxy or
polyurethane paints are subjected to very strict conditions with regard to the
relative humidity, the ambient temperature when applying the paint and the
maximum time between two successive phases.
Because the whole treatment was carried out on the building site and
Belgium has a very unreliable climate, the following application conditions
were imposed:

-degree of sand blasting, SA 3;

-time between sand blasting and metallising, maximum 2 h;
-time between the metallising and the first paint layer, 1 h maximum;
--ambient temperature, higher than + 5°C; and
-steel temperature, sec higher than dewpoint.

Organisation of the building site for the painting

In order to observe the application requirements, the contractor had
organised his planning very meticulously. He had built his scaffolds athwart
of the structure; they were provided with three separate working platforms
of confined width.
760 P. H. Besem, M. Wouters and C. Warnon

TABLE 3
Camparisan between the announced data and the results of the tests

Tests Unit Epoxy Polyurethane System

GV MV GV MV GV MV

Cupping test mm 6 9·2 5 4 1·5 4

(ISO 1520)
Adhesion test by cross-cut 0-5 0 0 0 0
(IS02409)
Pull-off test for adhesion Njmm 2 60 54
(IS04624)
Pendulum damping test s 115 165 230 105
(ISO 1522)
Bend test on conical mandrei mm 15 10-12 5 58-70
(ISO 6860)
Rapid temperature variation Good N Good N
resistance
Resistance to 20% HCI Very N
(ISO 2812) good
Resistance to 20% NaOH Very N
(ISO 2812) good

GV = guaranteed value.
MV= measured value.
N = no degradation such as blistering, cracking, flaking, etc.

A wooden platform 3m wide and 13m long (corresponding to the width

of the bridge) had been fastened to a tubular scaffold travelling on rails.
Special protection by means of canvasses had also been provided in order to
isolate the workmen during the sand blasting, metallising and painting
operations.
The necessary time to set up the building site (protection, moving of the
scaffolds and collecting of the blasting material) took 249 working hours.
Two sandblasters arid a part-time helper (70% of his time) worked for
400 h to treat the 1661 m 2 steel surfaces of the bridge, which constituted a
blasting rate of 4m 2 /h. They used 50·5 kg/m 2 of abrasive material of the
'Vasilgrit' type, having a grain size between 0·5 and 2 mm and which was
stored in silos of about 20 t. The material was blasted at a pressure of 7 atm
(produced by a compressor) by means of a sand blast machine with a
capacity of 240 litres, of which the blast-pipewas made of carbide and had a
diameter of 8 mm (entry) and 12 mm (exit). In total, the compressor
consumed 2825 litres of diesei oil for the sand blasting and the metal
spraying operations.
A plan was worked out to observe the intervals between the sand blasting
 The Repair of a Composite Concrete-Steel Bridge 761

and metallising operations. Steel girder sections, each of 6 m long, were

successively stripped by the sandblasters so that metallising could take place
within 2 h after the sand blasting.
One metal sprayer and the helper (for the remaining 30% of his working
time) sprayed the Zn-Al on the successive girder sections of 6 m long, as
described in the plan.
In total, it took 184 working hours to spray a coating of minimum 80 Jlm
thick over the whole surface of the steel parts of the bridge (1661 m 2 ).
Therefore 1045 kg of Zinacor 850 wire with a diameter of 4·76 mm were
necessary, which corresponded to a covering rate of 9m 2 /h. In the flame
spray gun, the wire progressed at a speed of 2·64 m/min. This allowed the
operator to obtain the required coat thickness by two crossed sweepings.
The gas torch used acetylene as fuel and oxygen as combustible. The
complete metallising operation required the use of 18 acetylene bottles and
26 oxygen bottles.
Two painters had to work 285 h to treat the total surface ofthe metallised
structure. The first filling layer of 20 Jlm thick was applied with brushes
whereas the intermediate layer of 60 Jlm thick and the finishing polyurethane
layer (100 Jlm thick) were applied with a low-pressure spray gun (opening
1·5 mm). The following quantities were needed to paint the structure:
-epoxy, 707 litres;
-polyurethane, 365 litres;
-diluent, 350 litres; and
-cleanser, 230 litres.
During each working phase specific checks were carried out by the Bridge
Office:
- The condition ofthe stripped areas and the rugosity ofthe surfaces were
regularly checked. In cases of doubt visual control was completed by
rugosity measure. The required degree of cleanliness was based on the
Swedish Standard SIS 055900 B/SA3.
- The thickness of the metallayer was checked by means of an electronic
thickness meter (magnetic effect) of the Elcometer type.
The measured thickness always varied between 75 and 140 Jlm. Indeed the
required minimum thickness of 80 Jlm made it necessary for the metal
sprayer to aim at a much higher average of about 120 Jlm.
The adhesion of the metal layer was checked by means of cross-cut
adhesion tests carried out according to the Belgian Standard NBN T22-107:
-During the application ofthe intermediate paint layers, the thickness of
the layers was checked several times to make sure that the painter was
observing the required thickness limits.
762 P. H. Besem, M. Wouters and C. Warnon

80

70
-i
11)
::J
60 ~-
~
50
,....
Vl

"'• 40 11)
::J
....
lO

30 :::r
X
~ ~

"..
lO
20 ;:;
3
10 ~

0
0 50 100
Failure in the metallizing (Ofo)

FIG. 7. Relation between the adhesion and the percentage of disbonding of the metallayer.

Taking into account the thickness variations ofthe metallayer, it was

indeed very difficult to determine with accurate precision the thickness
of each paint layer.
--A final check of the entire coat showed that its thickness varied between
200 Jlm and more than 400 Jlm. The theoretical value being 240 Jlm,
some very restricted zones with a thickness between 200 and 240 Jlm
were tolerated.
The adhesion of the coat was checked by means of tensile tests according
to the International Standard ISO 4624/78. The values which were obtained
varied between 25 and 64 kgjcm 2 . The breaks occurred either in the metal
layer or in the polyurethane layer; the lowest values were obtained in the case
of debonding of the metallayer. From the results of those measurements it
was concluded that the average lower limiting value of the adhesion
amounted to 35 kgjcm 2 for the metal layer and 55 kg/cm 2 for the
polyurethane layer (Fig. 7).

TESTING OF THE STRUCTURE

As in the case of a new structure, a loading testwas carried out on the bridge
by means of 25-t lorries. The test was in four parts:
-centre span;
---centre span and one side span;
-two side spans; and
-asymmetrical loading.
 The Repair of a Composite Concrete-Steel Bridge 763

CONCLUSIONS

In spite of the considerable size of the elements which had to be replaced, a

thorough previous study of the construction phases had made it possible to
reduce the traffic interruption under the structure to a minimum.
The repair works have affered the opportunity to apply new techniques in
Belgium: high-tensile (HT) bolts covered with tefton, metallising with Zn-Al
of the steel structures on the building site, use of polyurethane paints, and
waterproofing layer consisting of a prefabricated elastomer asphalt
membrane.

PARTIES INVOLVED

Commissioning Authority
Ministry of Public Works Road Department- Direction of Namur
Avenue Gouverneur Bovesse 37, B-5100 Jambes

lnspection of the Works

Bridge Office-External Services, rue C6te d'Or 253, B-4200 Liege

Planning and Design

Bridge Office-Central Services, WTC-Tour 3, Boulevard Sirnon
Bolivar 30, B-1210 Brussels

Controlling Engineering Office

Bureau SECO, rue d'Arlon 53, B-1040 Brussels

Contractor
S.A. J. Richard, rue de Jemeppe 224, B-4431 Ans (Loncin)
Subcontractors
Steel construction: S.A. Poncin, Grand'rue 72, B-5292 Clavier
(Ocquier)
Paint works: Association Namotte S.A. and Appruzzese, rue du
Progres 10, B-4430 Alleur
Warking study: Bureau J. Hovelmann, Avenue du Luxembourg 7,
B-4020 Liege

REFERENCES

1. BESEM, P. H., DEHAN, E. and PIRAPREZ, E., Le serrage des boulons a haute
resistance dans les assemblages reels (Tightening of high-tensile bolts in actual
connections). CRIF-WTCM, July 1987.
764 P. H. Besem, M. Wouters and C. Warnon

a
2. LEROY, Y. and BESEM, P. H., Mise en oeuvre des boulons haute n!sistance.
Comparaison des methodes de serrage (Use ofhigh-tensile bolts. Comparison of
the tightening methods). Annales des Travaux Publies de Belgique, 4, 1979.
3. BESEM, P. H. and WARNON, C., La remise en etat du pont de Daussoulx (The repair
of the bridge of Daussoulx). Annales des Travaux Publies de Belgique, 1989.
4. Circulaire 576-56 du Ministere des Travaux Publics, 'Protection des metaux
ferreux contre la corrosion-Peintures a formule imposee' (Circular No. 576-56
of the Ministry of Public Works, 'Protection of ferrous metals against
corrosion-Formula paints').
5. Circulaire 576-63 du Ministere des Travaux Publics, 'Protection des metaux
ferreux contre la corrosion-Peintures a performances' (Circular No. 576-63 of
the Ministry of Public Works, 'Protection of ferrous metals against corrosion-
Performance paints').
 66

Hydrodemolition-A Modern Technique of Concrete

Removal in Bridge Repair

RENZO MEDEOT
FIP Industriale SpA, Via Scapacchio, 1-35030 Selvazzano (PD), ltaly

ABSTRACT

Given the ever-increasing traffic volume, repair of bridges as key elements of

road systems is a problern of growing importance. As is weil known, any
restoration work first involves the removal ofall traces ofdeteriorated concrete
only, while avoiding any darnage to .wund concrete and restoration materials.
The technique of hydrodemolition, developed at the end of the 1970s in Italy,
sarisfies these requirements, using high-speed water Jets without abrasives. This
report describes this technology, explaining its theoretical basis and giving
many examples ofpractical applications.

INTRODUCTION

Hydrodemolition is a new word that, borrowing two terms from ancient

Greek and Latin, has been created to describe a process as old as the earth
itself: the destruction of rock and materials harder than concrete by the
relentless force of falling and surging water. The wearing force of water is
well known. Over aeons of geological time the Colorado River carved the
Grand Canyon, and the Niagara River the famous falls. In practice the
hydraulic power of a waterfall of about tOOOOm, 150-300litresjmin, has
been harnessed to produce equipment to remove concrete. Hydrodemolition
technology, in essence, compresses time from centmies to seconds by
speeding water flow to real-time cutting force, to demolish the bonds uniting
the concrete aggregate. But Iet us pause here to give a history of the
discovery and development of hydrodemolition.
In the later 1970s it was realised that the problern of removing !arge areas
765
766 Renzo Medeot

of deteriorated concrete (for example bridge decks) was becoming

increasingly urgent. Apart from the specific repair techniques tobe adopted,
any restoration work first involves removal of deteriorated concrete. This
delicate and often difficult task requires
(a) total removal of all traces of deteriorated concrete;
(b) avoidance of any darnage to sound concrete and reinforcing steel;
and
(c) good bonding, e.g. a good support surface between existing concrete
and restoration materials.
Traditional methods, based essentially on the use of pneumatic hammers,
did not guarantee satisfactory results. In particular, the greatest problern
was operator difficulty in differentiating between poor quality and good
quality concrete, which led to either incomplete removal of poor concrete or
excess removal of good concrete. Serious research on possible alternative
methods revealed that several studies had been carried out on this subject,
thus confirming its importance. However, no practical results had been
achieved, although we may quote, for the sake of curiosity, heat treatment
methods (all based on producing rapid heating ofthe damaged area, e.g. by
using flamethrowers, plasma beams or even lasers); abrasive processes
(based on the use of rotating discs coated with industrial diamonds in a
metal matrix or carbonium bound with bakelite); electrical and chemical
processes (only applicable in special situations); and-the most curious new
method in the group-the use of microwaves.
Research revealed that although pressurised water could demolish
concrete no successful attempt had been made. In spite of this, we believed
that water jets represented the most promising path and, in spring 1979, we
decided to start a research programme aimed at producing equipment for
removing concrete by means of high-speed water jets.
The most important discovery made during this research was the
following: 'removing a layer of concrete is a process which differs radically
from boring and cutting'. To use a familiar example, it is like the difference
between sawing and planing a piece ofwood: the tools are different, and so
are the ways ofusing them. The failure of attempts all over the world was due
essentially to having used techniques which were more suitable for boring or
cutting. Strong in this knowledge, we needed a few months to prepare a
prototype which was successfully used on the Viadotto del Lago in
November 1979 in conjunction with the Italian Road Authorities. The first
commercial equipment was ready by spring 1980.
After a series of improvements to perfect the system, hydrodemolition
technology was introduced into other countries. The firstwas Sweden where,
in the summer of 1984, the equipment was used on many bridgesandwas
 H ydrodemolition 767

tested by the Swedish Road Authorities (Vägverket), with very fiattering

results. In spring 1984 it was presented at the World of Concrete in
Washington, DC. In the autumn of the same year it began working in
Toronto in the Manulife Parking Garage. In 1985 hydrodemolition
equipment was successfully used in the USA on the Memorial Bridge
Rehabilitation Project. Today hydrodemolition is unanimously accepted as
the best process for concrete removaL lt has become popular in many
countries and the new rehabilitation projects specify this technique at least
as an alternative to the traditional methods.

HYDRODEMOLITION
By 'hydrodemolition' we mean the process of selective removal of concrete
by means of one or more high-speed water jets. Although the term
'demolition' may recall its synonym, 'destruction', it should be clarified at
once that our technique deals with the selective removal of deteriorated
parts, aiming at static restoration of the structure and not at its total
destruction.

Hydrodemolition Mechanism
Concrete is an inhomogeneaus material made up of aggregates (sand and
gravel) and bonding agent (cement), with gaseous inclusions which make up
the so-called porosity. Porosity is generally undesirable, since it alone is an
effective weakening agent through which degradation takes place.
The water jet accomplishes its destructive action by means of three
separate mechanisms, i.e. direct impact, pressurisation of cracks and
cavitation. These three processes reach their maximum efficiency when the
water jet strikes the bonding agent. The nozzle is thus played rapidly and
continually over the area to be removed and excess water allowed to drain
away. However, jet efficiency is a maximum when the jet itself is stable, and
stability is infiuenced by the shape and configuration ofthe feeding pipe and
nozzle, exit speed of water, distance from point of impact, etc.
The conclusion is that an efficient and therefore economic removal
process by hydrodemolition may be obtained by carefully combining fiuido-
dynamic, geometric and kinetic parameters as a function of the existing
situation (strength of concrete, presence of reinforcing steel, cracks, etc.) and
the type of work required. Obviously satisfactory work requires highly
qualified and experienced personnel capable of optimising the equipment
and skills necessary for each single case. The equipment must be sufficiently
powerful, but sophisticated movements and electronic control systems are
also necessary, since without them one ofthe most important advantages of
hydrodemolition is lost, i.e. selective removaL
768 Renzo Medeot

Selective Removal
Clearly, operative conditions being equal, removal involves a greater
depth of degraded or generally weaker concrete than it does in the case of
sound and resistant concrete, but this is not selective removaL Referring to
hydrodemolition equipment, selective removal is defined as the capacity to
remove completely all and only the deteriorated concrete, independently of
the depth to which the darnage has penetrated. Deterioration in bridge decks
or parking areas may involve thicknesses which vary from point to point-
in practice from zero to the whole thickness (Fig. 1).

FIG. 1

However, selective removal may also be defined as the capacity to remove

only concrete with strength of less than a certain pre-established value,
avoiding removal of concrete which has been considered as acceptable by
the engineer. As weshall see later, the term 'strength' (commonly understood
as compression cubic strength) is incorrect in identifying the type of concrete
tobe removed. In order to understand the phenomenon of selective removal,
we must, albeit briefly, go back to theory.
lf gradually increasing force, e.g. compression, is exerted on a material, the
latter is deformed according to a curve called the stress- strain characteristic
up to breaking point (Fig. 2). By using specific units we have

where Er is specific breaking energy (kJ/m 3 ) and er is breaking strain.

Obviously, if the breaking stress is not reached or if the material receives
energy less than the breaking energy, it will remain intact. Clearly materials
 Hydrodemolition 769

E=(kJjm')
rf(N/mm')
100
50

80
Sec. AA

60

40 20

20 10

0 100 200 S(mm)

0 2 3 4 € ( %.)

FIG. 2 FIG. 3

at higher strength require greater energy, so that there is a law of

proportionality between the two parameters. We may thus conclude that

-all materials have a threshold energy value at breaking point and

- there is a law of proportionality between the above threshold energy
value and the strength of the same material.

Therefore, if we plot the trend of the strength of concrete as a function of

depth, as shown for example in section A-A in Fig. 1, the samediagram may
also represent the energy required to break the specimen on a suitable scale
(Fig. 3).
If we want to represent the trend of the strength (or breaking energy) in a
section in which deterioration extends to a greater depth, as in section B- B,
the trend is that shown in Fig. 4.
The power of a water jet of ftow rate q (m 3 /s) and velocity v (m/ s) is given
by the equation
W=!pqv 2 (W)
where p is the specific weight of water.
The energy developed by time interval t is
E=Wt (J)
Let us now presume that the operational parameters of the hydro-
demolition equipment (pressure or water velocity, ftow rate, and the other
geometric and kinetic parameters) have been fixed.
Wehave thus established the amount of energy which may be distributed
over one surface unit. Let us now define the trend of energy available in the
jet per unit of volume of concrete in increasingly deeper sections (Fig. 5). It
770 Renzo Medeot

E=(k J/ ni) E=~J Jrri)

100 100

80 80
Sec. B B

60 60

40 40

20 20

0 100 200 S (mm) 0 100 200 S (mm)

FIG. 4 FIG. 5

should be noted that the power of the jet decreases with distance from the
nozzle, not only due to dissipation in the concrete and water but also and
above all due to the instability of the jet itself, since it produces small drops
which rapidly lose their energy even in the air.
Ifwe superimpose the two curves (energy necessary to break the concrete
and energy available in the water jet), we see that their cross point defines the
thickness of concrete which will be removed (Fig. 6).
If deterioration in another part ofthe deck had reached a deeper Ievel (for
example section B-B of Fig. 1), deeper removal would automatically be
obtained (Fig. 7).
Naturally, with sound concrete, a constant thickness may be removed by
using a sufficiently powerful jet (Fig. 8).
As already explained, in order to change the curve of available energy at
the jet, various parameters may be changed:
E= Wt=Jpqvt

E (kJJ rri) E(kJJni)

100 100
Sec. A A ~,AA_
80
--.:._,:::
80 Sec B B

110 60

40 40

20 20

0 100 200 S (mm) 0 100 200 S (tnm)

FIG. 6 FIG. 7
 Hydrodemolition 771

E (kJ/ ni') E (kJfni')

100 100

80 80

60 60

40 40

20 20

0 100 200 S(mm) 0 100 200 s(mm)

FIG. 8 FIG. 9

However, the simplest method is that of varying time t by changing

kinematic and geometric parameters of the movement. According to the
above it seems that selective removal is an intrinsic characteristic of
hydrodemolition, and it may be achieved with any equipment capable of
controlling a water jet. This is not entirely accurate. In effect, selectivity may
be achieved only with equipment supplied with a nozzle-moving system and
electronic control for guaranteed constancy of selected parameters in time;
moreover, the water jets must be highly stable and powerful. The trend of an
unstable jet versus position is shown in Fig. 9.
By overlapping Figs 3, 4 and 9 we see that about the same thickness is
removed in each case, and that this thickness does not even change to any
great extent if the area in question has sound concrete (Fig. 10).
The experimental evidence of the capacity for selective removal of our
equipment was shown by two series oftests carried out by the Swedish Road

E (kJ/m')

100

80

60

40

20

0 100 200 S (mm)

FIG. 10
772 Renzo Medeot

Administration (Vägverket) in collaboration with the Royal Concrete

Institute of Stockholm. Concrete slabs with indentations of regular
geometry (squares and rectangles) and varying depths were prepared (Fig.
11). After emplacement of rebars, concrete of lesser strength was poured
over the slabs in order to simulate deteriorated concrete.
Hydrodemolition was carried out after curing. lt was noted that only the
'deteriorated' concrete was removed, leaving the sound concrete practically
intact, both in original geometry and depth of the indentations (Fig. 12).

Modes of Operation
The above shows that hydrodemolition with selective removal does not
require detailed testing of the bridgedeck in order to identify deteriorated
areas and their depth-operations which areexpensive and far from precise
with existing methods-but it is sufficient to calibrate the equipment
carefully and proceed to removaL Within this apparent simplicity there are
various modes of operation suiting many different situations. Two of the
main ones are described below.
The first case deals with quite widespread deterioration in terms ofsurface
area, with depths varying from zero to the whole thickness and with
potential involvement of rebars. Once the minimum thickness to which the
repair material can be applied (e.g. 50 mm) has been established, a few square
metres of sound concrete are identified. As a first step the strength of the
concrete is determined on samples or, more simply, in situ using non-
destructive methods or pull-out tests. Working parameters are then fixed
with the help of diagrams and tables, obtained from previous tests carried
out on a slab of predetermined strength.
An initial attempt is made on about 1m 2 and, ifnecessary, the parameters
are redefined. The equipment is then moved to the area of worst
deterioration (the ideal situation would be an area where deterioration
involves the entire thickness of the deck). Testing is considered successful
if, with the same parameters as before, all the deteriorated concrete is
removed.
lfthere is reason to suppose that the concrete is not homogeneous over the
entire deck, the minimum thickness of removed concrete should be checked
periodically. Sometimes all the rebars must be exposed and all the
deteriorated concrete removed at the same time. The mode of operation is in
any case the same.
The second case often encountered is that of decks of relatively good
condition but with insufficient cover, leading to delamination. In this case
hydroscarification (5-10 mm) of the entire surface is recommended. In this
way the upper part, possibly contaminated, is removed and, at the same time,
excellent roughness is ensured for good bonding of repair materials.
 Hydrodemolition 773

ol
~o oj
0 10

.c-=<:E::::::~::::::tJ
j
gj 0 "'

Seetion A A
"'

4000

0 rEJ-
~[\;d
i900
f- -+ 350 0
0
"'
- +--3--a--J I 1m
FrG. II

FrG. 12
774 Renzo Medeot

Hydroscarification also shows up possible areas which have undergone

some degree of deterioration, but where delamination has not yet occurred.
The bound areas of the deteriorated parts are marked with regular
geometrical shapes, if possible grouping several adjacent zones into a single
patch. Deep removal is then carried out. lt should be noted, however, that
patehing is never recommended. lt is done, for reasons of economy, only if
the deteriorated areas do not exceed 15-20% of the whole deck.

ADVANTAGES OF HYDRODEMOLITION

Understood as a new process in the field of removing concrete,

hydrodemolition generally offers many advantages over traditional
methods. Of course additional advantages also derive from the kind of
equipment used, its power, manoeuvrability, control system, etc., all of
which influence removal speed (or productivity) and quality of work-in
other words, the economic result.
From the technical viewpoint, the advantages ofhydrodemolition are the
following:

----constant, repeatable results, once Operating characteristics have been

established;
-guaranteed total removal of deteriorated concrete (see paragraph on
selective removal);
--no darnage caused to sound parts of concrete;
-possibility of working even in the presence of rebars, which are not
damaged; on the contrary, they are given a thorough cleaning and any
trace of corrosion is removed from even their lower parts, usually not
reached by other processes such as sand blasting;
----creation of a very rough surface ensuring excellent bonding to repair
materials, much higher than in the case of jacking or chipping hammers;
-no impacts of vibrations, thus on the one hand avoiding darnage to
reinforeerneut and on the other ensuring that noise is kept to an
acceptable level; moreover, some simultaneaus and otherwise imposs-
ible operations, such as casting in immediately adjacent areas, may be
carried out; and
--no dust or fumes (until now an inevitable accompaniment to concrete
removal works).

1t should also be noted that work may be done even in poor weather
conditions and sub-freezing temperatures.
 Hydrodemoli t ion 775

THE EQUIPMENT

As already noted, historically hydrodemolition came into being to solve the

problems ofbridge deck restoration. However, it has also been extended to
other similar applications, such as repair work on parking garages, airport
runways, concrete roads, etc. All these applications refer to horizontal or
almost horizontal surfaces.

FICi. 13

FIG. 14 FIG. 15
776 Renzo Medeot

Even for vertical surfaces and soffits a new demolition unit has been
designed and manufactured, using very complex mechanical and oleody-
namic systems requiring computerised control. Figure 13 shows a real robot
with a demolition head on the end of an articulated arm supported by a 360°
swivelling tower. It may be remote controlled for work even in inaccessible
areas (e.g. 'hot' areas of nuclear power plants during decommissioning).
For removal of concrete cladding in tunnels, a special truck-mounted
piece of equipment has been designed (Fig. 14).
Many accessories have been developed to fulfil special requirements. For
areas where access is very difficult, special manual equipment has been
designed-the 'Bazooka'. This is a kind ofthrust-compensated gun allowing
removal of concrete underneath bridge decks and girders (Fig. 15).

CONCLUSIONS

Experience world-wide has shown that hydrodemolition represents a

revolutionary but mature technique in concrete removaL It is not only a
great improvement over conventional systems but operations which were
once impossible may easily be carried out. Structures may be restored even
in cases of advanced deterioration. lt is a firm beliefthat to date only a few
applications have been explored. Therefore new R&D is tobe carried out
aimed at improving reliability and effi.ciency of existing machinery as weil as
designing new equipment.
 67
Aluminium Extrusion Bridge Rehabilitation System

LARS SVENSSON
Lars Svensson JADA AB, Brahegatan 56, S-114 37 Stockholm, Sweden
and
LARS PETTERSSON
The Royal Institute of Technology, Department of Structural Engineering,
S-100 44 Stockholm, Sweden

ABSTRACT

The deterioration of road bridges is becoming a serious problern in Sweden. An

increasing number ofbridge decks are in such poor condition that they must be
replaced. This is due mainly to a severe climate, the use of road salt in winter
time, and increasing traffic volume and Ioads. In certain cases major
reinforcement of bridge foundations or replacement of bridges has been called
for to guarantee safety, often at great cost. A system is describedfor replacing
damaged bridge decks. The system utilises an orthotropic plate that consists of
hollow aluminium extrusions .fitted together by tongue and groove. Major
reduction of dead Ioads is achieved which allows increased traffic Ioads without
reinforcing foundations or the structural system.

INTRODUCTION

Road bridges are long-term investments with an expected lifetime of at least

50 years. In many cases bridges are in such poor condition after a much
shorter timethat they need tobe repaired or replaced. A severe climate, road
salt and increasing traffic Ioads are three major factors causing deterioration
ofbridge structures. Deterioration of concrete decks is mainly due to the first
two factors. In areas with poor soil conditions increasing traffic Ioads have
caused darnage to foundations and supporting structures.
In cases of severe darnage to concrete decks the usual course of action is to
replace the deck with a new one. There are several methods for undertaking
777
778 Lars Svensson and Lars Pettersson

this. If the foundations are damaged, or if the bridge is to carry increased

Ioads, then reinforeerneut of selected areas is usually necessary, often at great
cost.
Bridgesare structures in which a !arge part ofthe Ioad-hearing capacity is
allocated to carrying its own weight. For a steel girder composite bridge
most of the dead Ioad lies in the concrete deck. By replacing the concrete
deck with a lighter one it is possible to increase the allowable live Ioad
without reinforcing the main structure or the foundations.
A lightweight system for replacing damaged concrete bridge decks has
been developed and used in Sweden. The basic concept is an orthotropic
plate of alumini um. Weight reduction has made it possible to use existing
foundations, supports and structural system for increased allowable live
Ioads on bridges that would have to be replaced otherwise.

BASIC CONCEPT AND STRUCTURAL SYSTEM

The basic concept is an orthotropic plate that is built up of hollow

aluminium extrusions. The extrusions are fitted tagether by means of a
tongue and groove in the upper flange, as shown in Fig. 1. This type of
connection transfers shear force from one extrusion to the other. At the
same time it allows each extrusion to rotate independently of the
neighbouring ones. The hollow section is used to create a high degree of
torsional stiffness in the extrusion.

FIG. 1. Section through aluminium bridge deck extrusion.

When subjected to concentrated Ioads, which is most often the case for
bridge decks, the resistance of the deck is due to a combination of the
bending and torsional stiffness in the extrusions. In this way the Ioad
distributionwill be considerable and concentrated Ioads will be carried by at
least seven extrusions at the same time.
The cross-section of the extrusion has the form of a truss. This is to ensure
resistance against local deformation due to point Ioads such as tyre Ioads
caused by vehicles.
The deck is mounted to a secondary structure of steel girders or, as in
certain cases, directly to the main structure. The distance between the
 Aluminium Extrusion Bridge Rehabilitation System 779

supporting girders varies between 1·2 and 3-0m depending on the size and
type of extrusions.
Fasteners of extruded aluminium are used to secure the deck to the
supporting structure. The fasteners hook on to grooves in the lower flanges
of the extrusions and are bolted to the supporting structure.
The surface of the deck is covered with an acrylic-based material called
Acrydur. This paving has been applied to bridges for many years and has
shown very high resistance to wear in existing bridges.
The weight of the deck system described lies between 50 and 70 kg/m 2 •
The weight of standard concrete decks is normally between 600 and 700
kg/m 2 . The aluminium deck is corrosion resistant even in a marine
environment. Construction time is short, which is favourable when repairing
bridges in areas with heavy traffic.

APPLICATIONS

The system has been applied to drawbridges, pontoon bridges and

stationary bridges with primary structures of steel. lt can also be applied to
new bridges, often with the added advantage of a reduction in the weight of
the primary structures and also in the size of foundations required.
In areas with poor ground conditions the weight reduction caused by
incorporating the deck system has allowed the use of existing foundations to
support increased Ioads without the need for additional reinforcement. In
piled foundations this often means fewer piles.
Studies by the senior author have shown that the reduction in weight due
to the use ofthe aluminiumdeck system increases with the size ofthe bridge.
This is largely due to the fact that the relative size of the dead Ioads increase
with increased span.

THEORETICAL AND EXPERIMENTAL INVESTIGATIONS

The system is designed to fulfil requirements given by the Swedish National

Road Authority. These include resistance to static and dynamic Ioads
according to the Swedish bridge and building codes.
A first analysis of the deck was undertaken by modeHing the structure by
the use of finite elements. A model of the deck was formulated using beam
finite elements. The extrusions were simulated by beam elements having the
flexural and torsional stiffness of the proposed section. The effect of
interaction between extrusions was simulated by using connecting beam
elements with almostinfinite flexural stiffness. The model is shown in Fig. 2.
780 Lars Svensson and Lars Pettersson

r r r
152.4x7 256.6x5 427.7x3 355.7x3 • [mm)
r 1
J

~ E~ ~~·

s s s s
S (:Support)
FIG. 2. Finite element model of aluminium bridge deck system.

The applied load represented a pair oftruck tyres. The load was placed in
the most unfavourable position to cause the largest possible deformation in
the structure, as shown in Fig. 3. The deformation reached a maximum of
6·5 mm at the right end of the extrusion subjected to the applied load. The
load in this case was 100 kN.
A full-scale model was tested 1 to confirm the results of the theoretical
investigations. These tests included static loading as described above and
dynamic loading in accordance with the requirements set by the Swedish
National Road Authority.
The results of the static loads showed very good agreement with the
results from the FEM analysis. A comparison between results is given in Fig.
4. The static loadwas increased to an ultimate value of 320 kN, which is well
beyond the dimensioning load of 100 kN. The maximum deftection of a deck
element under a load of 320 kN was 27 mm.
Dynamic loading was applied to investigate resistance to fatigue. The load
had an amplitude of96 kN and a maximum value of 100 kN. The frequency
was1-2Hz and a total oftwo million cycles was applied. No cracks or signs
offatigue were visible after the test and the maximumresidual deftection was
0-4mm.

t
-0
-1
-0 ..------+----.....,
e -1
e -2
..§. -3
.E -2 c
.Q -4
-~ -3 ö
-5
öE -4 E
.Q -6
a
Qj
-5 "'
0
-7
0 -6 -0.7 0 0.7 [m)
-7 .1...._~---+---~-'
-0.7 0.7 [m) FIG. 4. Comparison between deflections
obtained from the finite element analysis and
FIG. 3. Deformation of aluminium bridge full-scale tests. Load position according to
deck calculated by the finite element method. Fig. 3. ( - - ) FEM, (----) full-scale test.
 Aluminium Extrusion Bridge Rehabilitation System 781

Extruded deck elements were tested in bending and torsion to investigate

the influence ofthe surface material. These showed that the Acrydur surface
has negligible inftuence on the structural behaviour of the deck.
In conclusion, the experimental investigation confirmed the applicability
of the FEM analysis. The results were in good agreement with calculated
values and the structural resistance of the deck was well within the
req uiremen ts.

ECONOMIC ASPECTS

In many cases the alumini um bridgedeck can be a very competitive technical

and economic alternative to the conventional concrete bridge deck. Even
though the cost for the aluminium deck is approximately the same as for a
corresponding concrete deck, the aluminium deck gives a number of
advantages.
-Time for detail design is kept to a minimum.
-Very short time for construction works, which in many cases greatly
reduces the total investment costs. At present (June 1989) replacement
of a concrete bridge deck by an alumini um deck is under investigation.
The plan is to carry out the replacement injust 2 days in order to disturb
the traffic as little as possible. In addition, the costs for a temporary
bypass road can be avoided.
- The reduction of the dead Ioad can in many cases make it possible
to use existing bridge foundations and main girders. This is maybe the
greatest advantage and should be studied carefully for each particular
case.
-After completion the costs for maintenance arealso kept to a minimum
because the Acrydur paving is almostimpossible to wear down and, in
addition, by careful choice of a suitable alumini um alloy the corrosion
resistance will also be very good.

CASE STUDY

The new aluminium bridge deck system has been used in several cases in
Sweden, with the most recent case being the Tottnaes Bridge south of
Stockholm (Fig. 5).
The Tottnaes Bridge, a four-span steel girder bridge with one span
designed as a swingspan, was in immediate need of repair, primarily
resulting from the deterioration of the concrete bridge deck. In addition, a
study ofthe foundations (wood piles) showed that it would not be possible to
782 Lars Svensson and Lars Pettersson

FIG. 5. The Tottnaes Bridge south of Stockholm after renovation with the described
aluminiumbridgedeck system (June 1989).

FIG. 6. Detail of railing and edge beam of alumini um bridge deck system.
 Aluminium Extrusion Bridge Rehabilitation System 783

increase the allowable live load without reinforcement, which was required if
the bridge was to carry modern heavy vehicles.
The only possibility available to keep the old bridge foundations and at
the sametime increase the allowable liveloadwas to reduce the dead load of
the bridge. This was accomplished by the use of the new alumini um bridge
deck system. Compared with the old concrete deck a reduction in the dead
load of about 550 kg/m 2 was obtained, which was enough to maintain the
existing bridge foundations without reinforeerneut
Because the Tottnaes Bridge is the only connecting link between the
mainland and the island of Toroe, a short time for the restoration of the
bridgewas of great importance. Using the aluminiumbridgedeck elements
with the Acrydur surface applied to the aluminium extrusions, and by
preparing the support details and connections in advance, it was possible to
reduce the repair time to a minimum.
All other details such as railings and connections between railing and the
bridgedeck as weil as edge beams (Fig. 6) are readily available products well
tested and approved by the Swedish National Road Authority.

SUMMARY

A new bridge deck system is described which is built up from hollow

alumini um extrusions. The extrusions fit together by means of a tongue and
groove system. In this way the load distribution is considerable, which
makes it possible for this very light deck to carry heavy live Ioads. By
replacing old deteriorated concrete bridge decks with an aluminium deck
system a considerable reduction ofthe dead Ioad is achieved, which in many
cases means that existing foundations and main girders can be left without
additional reinforeerneut and still carry increased live Ioads. In addition, the
new bridge deck system offers a very short construction time and requires
very little maintenance.

REFERENCE

1. SALTIN, M. and McCARTHY, R., Brofarbana av aluminiumprofiler (Aluminium

extrusion bridge deck). Diploma work at the Royal Institute of Technology,
Department of Structural Engineering, Stockholm, 1988.
 Index of Contributors

Abdunur, C., 489 Dragotti, A., 155

Al-Mandil, M. Y., 549 Duchene, J.-L., 489
Azad, A. K., 549

Ellick, J. C. A., 585

Baluch, M. H., 549 Ellinas, C., 501
Barr, B., 349 El-Marasy, M., 89
Beales, C., 459
Beckett, D., 3
Berthelsen, F., 29 Flesch, R., 327
Besem, P. H., 747 Fukushi, A., 725
Birnstiel, C., 295
Blyth, M. F., 705
Bouabaz, M., 187 Göhler, B., 287
Brookes, C. L., 439
Bucak, Ö., 533
Buchner, S. H., 439 Halse, W. 1., 173
Buckland, P. G., 475, 575 Hammersley, G. P., 655
Burke, Jr, M. P., 677 Harvey, W. J., 515
Hoffman, G. L., 75
Holland, D. A., 135
Cairns, J., 619, 643 Horner, R. M. W., 187
Camomilla, G., 155
Cesare, M., 319
Cogswell, G., 39 Ingvarsson, H., 199
Cope, R. J., 429 Itzkovitch, M., 607
Cullimore, S., 205
Cullington, D., 447
Cuninghame, J. R., 459 Jackson, P. A., 429
Johnson, R. A., 501

Dahinter, K., 383

Darby, J. J., 655 Kagawa, Y., 725
Dawe, P. H., 135 Kähkönen, A., 101
Dill, M. J., 655 Keer, J. G., 233
785
786 Index of Contributors

Kernbichler, K., 327 Pettersson, L., 777

Kirkpatrick, J., 361 Pretlove, A. 1., 585
Pritchard, B., 667

Lebek, D., 563

Lee, D., 145 Ramezankhani, M., 349
Leeming, M. B., 243 Read, J. A., 267
Le Page, B. H., 233 Reel, R. S., 715
Lichtenstein, A. G., 695 Reij, A. W. F., 215
Lindbladh, L., 51 Romagnolo, M., 155
Low, A., 417 Rosenthal, 1., 607

Sharif, A. M., 549

Maguire, J. R., 595
Simpson, B., 705
Mang, F., 533
Sriskandan, K., 17
Manning, D. G., 305
Sloan, T. D., 361
Marshall, A. R., 63, 101
Smith, F. W., 515
Masliwec, T., 305
Smith, N. 1., 223
Maxwell, J. W. S., 113
Söderqvist, M.-K., 63
May, P. H., 121
Sorensen, A. B., 29
McClure, R. M., 75
Stephens, R. L. C., 173
Medeot, R., 765
Svensson, L., 777
Mehrkar-Asl, S., 439
Mehue, P., 633
Melbourne, C., 523
Tallin, A. G., 319
Middleton, C., 417
Thompson, A., 361
Miesseler, H.-J., 395
Mitchell-Baker, D., 205
Montgomery, F. R., 259
Moses, F., 405 Van der Toorn, A., 215
Vrahimis, S., 121
Murray, A. McC., 259
Muruganandan, C., 715
Waldron, P., 349
Warnon, C., 747
Nebbia, G., 155 Wenzel, H., 339
Wolff, R., 395
Wood, J. G. M., 501
Palmer, J., 39 Woodside, A. R., 737
Pearson-Kirk, D., 549 Woodward, W. D. H., 737
Petrangeli, M. P., 373 Wouters, M., 747
 Subject Index

Ageing mechanism of bridge structures, 219 Bascule bridges, 297-8

Al-Darb Bridge, 551-6 Belgo Bridge, 483-6
Aliasing, 597-8 Bending frequency factor (BFF), 591
Alkali aggregate reaction, 506-9 Berlin-Marienfelde Bridge, 400
Alkali-silica reaction, 24, 244 Berounka Rivcr bridge, 389-90
Aluminium extrusion bridge rchabilitation Blumberg Bridge, 541, 543
system, 777-83 Boonton Bridge, 698-700
applications, 779 Box girder bridges, glued segmental,
basic concept, 778-9 349-60
casc study, 781 3 Brenner motorway bridges, 327-38
economic aspects, 781 Bridge beams, hammer testing, 600-3
FEM analysis, 780-1 Bridge condition deficiencies, 83-4
structural systcm, 778-9 Bridge costs. See Finance
theoretical and experimental Bridge data information systems. See
invcstigations, 779-81 Computerised database and
Arch bridges information systems
analysis of multi-span, 515-22 Bridge decks
assessment of, 523-31 bending relief by external prestressing,
effect of defects, 523-31 668--71
hinge formation, 526-7 composite, fatigue relief, 674-6
indeterminacy of, 516-17 concrcte quality assessment, 552
maximum thrust case, 517 condition surveys, 305-17
minimum thrust case, 517 construction materials, 40
model tests, 524-5 darnage likelihood chart, 560
multiple redundancy, 519 21 deterioration, 306
released structurc, 517-18 major types of, 306
restoring redundancy, 518-19 structural and material damage, 549-61
Serviceability assessment, 585-94 Bridge design
six-metrc span tcst, 527-9 code models, 141
three-mctre span tests, 525-7 codes, 25
vibration tests, 585-94 criteria, 118
Arehing action, 430 Ioad concept, 568
Asphalt cracking, 557 regional inftuences on, 151
Asscssment code, 46, 138-9 strategies for improving, 149-51
Assessment methods, 24-5, 140, 489-99 vehicles, 553
A veraging, 598 see also Load; Loading
787
788 Subject Index

Bridge directory, 107-9 Cable-band bolt tension-contd.

entity attribute relationship, 109 tightening torque and tension drop rate,
Bridge furniture, 151 730--3
Bridge inventory, 20 Cable stayed bridges, 339-48
Bridge management construction and supervision, 343-5
basic topics, 216 design, 342
computer development, 43-4 inspection, 345-8
conceptual stage, 17-18 maintenance, 345-8
construction stage, 19-20 repair, 345-8
cost-effective strategies in, 715-24 Cable-supported bridges, 150-1
Cyprus, 121-32 CAN/CSA-S6-88, 575-84
design stage, 18-19 Capital costs, 717
DISK system, 89-99 Carbonation depth, 162
Europe, 29-38 Castigliano's theorem, 591
Far East, 29-38 Chloride contamination, 162, 246, 655-66
Finland, 63-74, 104-5, 109-11 Chloride profiles, 558, 559
highway bridges, 113-20 Clifton Suspension Bridge, 205-14
implementation of, 130--2 deflection measurements, 212
in-service stage, 20--6 early history, 207-8
logic diagram of, 127 inspection and testing, 211-12
organisation and budget arrangements, maintenance strategies, 210--11
123-4 operation, 210--11
overview, 17-27 policy of indefinite preservation, 205-7
Pennsylvania, 75-87 post-war period, 208-10
sequence of, 126-8 revised bolt assembly on suspender rod,
Surrey County Council, 39-50 209-10
Swedish National Road Administration, testing and analysis, 211-12
51-61 traffic statistics, 212
United States, 63-74 weight restriction, 208-10
use of term, 17 Code of Assessment, 46, 138-9
see also DANBRO; Maintenance Code of Practice for Bridges, 619
management Code of Practice for Fatigue, 619
Bridge procurement, 152 Coefficient ofthermal expansion, 354
Bridge structures Cogan Viaduct, 351, 352, 356--9
ageing mechanism of, 219 Coherence function, 598-9
deterioration, 23-5 Collapse analysis, 419-22
loading determination in, 504 Combined shear and bending failure,
reliability analysis, 501-13 420--1
structural analysis in, 504-5 Composite bridge decks, fatigue relief,
interventions on, 164-6 674-6
BRIDGET, 190 Composite concrete-steel bridges, repair
BRIDGIT, 43, 44 of, 747-64
Brittle fracture, 23 Composite steel bridges, 151
Bronx-Whitestone Bridge, 486 Compressive membrane action, 430
BS5400, 10,447-9,454,455,466,619,674 Computer-aided sketching of Ioad paths,
BS8110, 645-7 515-22
Computer analysis of post-tensioned
concrete bridges, 442-3
Cable-band bolt tension Computerised database and information
drop in, 725-35 systems, 54--6, 58-60, 69, 73, 77-8,
effects of high tightening torque, 733 116-18, 128-30,312-13,597-600
measuring methods and equipment, 729 Concrete
safety factors and retightening time, capillary flow or absorption
733-4 measurement, 252-3
 Subject Index 789

Concrete~contd. Crack-bridging coatings, 233-41

controlling movement of water in, 250-2 need for, 235-6
corrosion of reinforcement in, 245-7 parameters affecting selection, 235
curing/maturity of, 248 test methods for, 236-9
deterioration of, 244-7, 267-71 Cracks
main faults in, 270-1 curvature and stiffness redistribution,
remedial measures, 271-4 effect on, 494
rcplaccmcnt of mcmbcrs, 274-6 failure patterns, 451
water roJe in, 243-4 flyover viaducts, 634
see also FBECR mechanical behaviour, effect on, 489-99
gas diffusion in, 246 reinforced concrete beams, 644 ·
maintenance, 24 rib-to-deck plate junction, 635-7
maisture content of, 255 rib-to-floor beam junction, 639-40
quality assessment of bridge decks, 552 steel orthotropic decks, 633-43
removal in bridge repair, 765-76 welded joints, 637-9
surface treatments, 233-41, 250-1 Craibstone-Dyce link road, 282
water ingrcss into or through, 248-50 Creep-time relationship, 177-8
water roJe in, 247-8 Cross Keys Bridge
workabilityjcompaction of, 247 aspects of repair work, 712-13
Concrete bridges, 24 future road traffic options, 707-8
box girder, 349-60 history, 706
Czechoslovakia, 383-93 renovation, 705-13
inspection, 140 main contract, 709-12
performance of, 147 structural shortfalls, 708
rehabilitation programmes, 139-40 urgent works contract, 708-9
strength assessment, 429-38 Cyprus, bridgc management in, 121-32
structural and material damage, 549-61 Czechoslovakia, concrcte bridgcs, 383-93
traffic-induced strain in steel
reinforcement, 619-29
see also Prestressed concrete bridges;
Reinforced concrete bridges DANBRO
Conservation viewpoint, 171 bridgc managcment and maintenancc
Copper wire sensors, 399-400 system, 29-38
Corrosion, 559 bridge overview, 31
chloride-induced. See Chloride bridge rating, 34
contamination concept of systcm, 30-7
prestressing tendons, 141 hierarchic elemcnt structure, 31
protection of steel structures, 755-62 inspection and bearing capacity module,
reinforcement in concrete, 23, 25, 245-7, 32-4
267-84, 643 inventory module, 31
seealso FBECR main activitics covcred by system, 30
tendons, 24 objectives, 29
COSBEN Program, 721, 724 outlines for systcm, 36
Cost-effective analysis, 717-21 ranking and budgcting module, 34-6
examples of, 721-4 special inspection report, 33
Cost-effective strategies in bridge system modules, 31-6
management, 715-24 Danish Statc Railways, 37
Cost inventory file, 78 Darcy's law, 248
Cost model factor (CMF), 190, 192 DART (deck assessment by radar and
Cost-significant work packages (CSWPs), thermography), 306-16
189, 191-2 costs, 316
Costs. See Finance data acquisition/processing, 312-13
Cover plate terminus, reliability index, field Operationsand cxpcrience, 313-14
322-5 prototype vehiclc, 310-13
790 Subject Index

Data processing, 166-71 Elastic analyses, 515-16

dynamic testing, 597-600 Elastomeric compression seals, 680
see also Computerised database and Electrical components, 300-1
information systems Electrical power equipment and controls,
Databases. See Computerised database and 299-302
information systems; DISK Elm Street Bridge, 700-4
Debonding, 307, 309 Europe
Deicing problem, 146 bridge management, 29-38
Delaminations, 307, 310, 313, 557, 559 maintenance systems, 29-38
Department of Highways, Thailand, 37 Evaluation
Department of Transport, 135-43 Canadian Standard CAN/CSA-S6-88
Design. See Bridge design Clause 12, 575-84
Deteriorating bridge structures. See Bridge guidelines, 407
structures Expansion joints, 737-45
Deterioration curves over time, 164 failure of, 738-9
Deterioration rate modelling, 511 movements experienced at, 739-40
Digital spectral analysis, 597 Expert systems, maintenance management,
DIN4150, 587, 592 155-71
Discount rate for rehabilitation projects,
718
DISK system, 89-99 Far East
goals of, 90-1 bridge management, 29-38
historical information, 97 maintenance systems, 29-38
inspection, 92-5 Fast Fourier Transform (FFT), 596, 598
inventory and administrative Fatigue, code of practice for, 619
information, 91-2 Fatigue assessment of orthotropic steel
maintenance, 95-7 bridge decks, 459-73
management information, 98-9 Fatigue costs, 413
registration of information, 97-8 Fatigue cracking, 23
Dolsan Bridge, 340, 345-7 Fatigue fractures in railway bridges, 533-5
Dome effect, 430 Fatiguc lifc
Dornie Bridge, 282 calculation methods, 468-72
Dynamic analysis, 163 enhancement of, 674-6
Dynamic properties, hammer testing, Fatigue reliability, 319-26
595-604 definition, 321
Dynamic testing, 163 Fatigue tests on railway bridges, 539-47
analysis of measurements, 334 FBECR (fusion-bonded epoxy-coated
data processing, 597-600 reinforcement), 267-84
description of method, 332-7 changes necessary with, 276
dynamic calculations, 334 costs incurred using, 276-7
focus and limitations, 332-3 method of application, 277
highway bridges, 375-6 surface yreparation prior to coating,
history of application, 331-2 278-81
influence of traffic, 335-7 UK application, 281-3
main aims oC 327-9 Federal Highway Administration
main steps of, 328 (FHWA), 64-5
motorway bridges, 327-38 Federal Sufficiency Rating System (FSRS),
orthotropic steel bridge decks, 465-6 83
outline of method, 329 Fibre composite materials, 395-402
sensitivity investigations, 335 Figg water permeability test, 254
surfaced bridge decks, 469 Finance
systematic system identification, 334-5 bridge costs, 411-13
technique, 333 bridge management in Cyprus, 123-4
 Suhject Index 791

Finance-"--contd. Freeze/thaw damage, 244

capital costs, 717 Frequency distribution, 171
DART costs, 316 Frequency resolution, 597
maintenance costs, 13-14, 143, 145-53,
184, 187-97, 199-204,717
new-build costs, modelling, 188-90 Gänstorbrücke in Ulm, 331
operation costs, 199-204 Gas diffusion in concrete, 246
repair costs, 187-97 Germany, prestressed concrete bridges,
Swedish National Road Administration, 287-94
52-3 Glued segmental box girder bridges, 349-60
see also Cost differential temperature effects, 356-9
Finite element analysis, 552 instrumentation, 352-9
Finland programme objectives, 352
bridge management, 63-74, 104-5, short-term performance, 355
I09-11 time-dependent effects, 353-5
bridge statistics, 102 Golden Gate Bridge, 486
inspection, 73, 105-7 Governing failure mode, 422
Finnish Roads and Waterways Grangctown Viaduct, 351, 352, 355
Administration (R W A), 71-4 Gun railway, 537
First-order reliability methods (FORM),
319
Flexible surface mix, 737-45 Haeng Ju Bridge, 341, 342, 344-5, 348
development of, 740-A Hammer testing
Flexural stiffness, evaluation of, 490-2 bridge beams, 600--3
Flyover viaducts, cracks in, 634 dynamic properties, 595-604
Forth Rail Bridge typical setup, 599-600
ccntenary of, 3-15 Health and Safety at Work Act, 12
eonstruction, 8-9 High alumina cement, 245
contractors, 14-15 Highway bridges, 373-81
design, 4-8 design of, 576
engineers, 14-15 dynamic tests, 375-6
expansion/contraction lengths, 9 inspection, 374-8, 383-93
expansion joints, 10 management flow chart, 115
general arrangement of superstructure, 6 man2.gement systems, 113-20
historical background, 3-4 monitaring systems, 119
logistics, 14 policies and objectives, 114-16
maintenance, 12-14 proteelive measures, 259-65
maintenance eosts, 13-14 rehabilitation programmes, 135-43
painting and repainting, 12-13 repair, 378-80
principal dimensions, 5 Serviceability performance, 607-17
provision for movement, 9-10 steel trusses for, 695-704
structural principlc of, 5 strength evaluation, 383-93
test loading, I 0-11 structural and material damage, 549-61
workforce, 11-12 technical issues, 118-19
Foyle Bridge types and dimensions, 419
monitaring United States, 405
hardware, 364-7 vibration measurements, 607-17
in-service behaviour, 361 Hinge formation in arch bridges, 526-7
results, 368-70 H-MAC flexible surfacing mix, 738, 744-5
software, 367 Holland, maintenance strategies, 215-22
system requirements, 364 Hougomont, 12
Fracture mechanics crack growth analysis, Hydrodemolition
319-20 advantages of, 774
792 Subject Index

Hydrodemolition~contd. Inventory system in Cyprus, 128-3tl

equipment, 775-6 ISA T method, 254
mechanism of, 767 IS04624/78, 762
modes of operation, 772~ ltaly, highway bridges, 373-81
selective removal of concrete by, 768-72
technique, 765-76
Hydrophobie treatments, 239~0 Jindo Bridge, 340, 343-7
Hydroscarification, 772~

Kanmon Bridge, 730, 733

Impact device design, 588
Kings Bridge, Melbourne, 23
Incremental benefitjcost analysis, 719-20
Kishon River bridge, 607-17
Incremental benefit/cost ratio, 716-17, 721
Kita Bisan-Seto Bridge, 726
Information systems. See Computerised
Koblenz/Waldshut railway bridge, 535
database and information systems
Korea, cable stayed bridges, 339~8
Infrared thermography. See
Thermography
Initial surface absorption test (ISA T), 253,
256 Lavant Bridge, 331
lnnoshima Bridge, 730 Leakage
Inspection, 20-2 bridge joints, at, 737
cable stayed bridges, 345-8 problern of, 598
Clifton Suspension Bridge, 211-12 Level-of-service deficiencies, 83
complex diagnostic method, 384-5 Life cycles for rehabilitation treatments,
concrete bridges, 140 718
cost effectiveness of, 46 Limit state design (LSD), 407
Cyprus, 124-6 Limit state function, 423
Finland, 73, 105-7 Linear elastic fracture mechanics (LEFM),
highway bridges, 374-8, 383-93 319
movable bridge machinery, 302~ Lions' Gate Bridge, 478-83, 486
PennDOT, 78 Live Ioad factors, 576, 577, 582
prestressed concrete bridges, 287-94 Load and resistance factor design (LRFD),
reinforced concrete bridges, 261 407, 408
scheduling system, 67 Load capacity
special methods, 161 ~ classification, 59-60
steel bridges, 319-26 evaluation, 408-9
steel trusses, 695-704 Load factors, 579-81
Tamar Bridge, 177-8 Load paths
see also DISK system; DANBRO; analysis, 516
Surrey County Council; Swedish computer-aided sketching, 515-22
National Road Administration Load relieving techniques, 667-76
Instrumentation. See Monitaring Load simulation. See Traffic Ioad
Integrated bridge construction, 677-93 simulation programme
conversions to continuous span Loading assessment, 138
(retrofit), 687-92 Loading determination in deteriorating
cost comparisons, 680 bridge structures, 504
current design trends, 678-9 Loading increases, 26
cycle control joints, 683-7 London Docklands Light Railway, 673-6
integral abutment details, 682 LUSAS, 212
passive pressure, 683
performance comparisons, 679-81
pile stresses, 683 Magnesium phosphate concrete, 180
structural analysis, 682 Main-channel at Bamberg, 288-9
structural distress, 681-2 Maintenance cost index, 201-3
 Subject Index 793

Maintenance costs, 717 Midlands Links Motorway Viaducts, 148

comparative study of, 199-204 Milton Bridge, 620
Forth Rail Bridge, 13-14 Minami Bisan-Seto Bridge, 726
modeHing and predicting, 187-97 Miner rule, 537
Maintenance models, 216-22 Modal analysis, 596
application of, 218-21 Monitaring
Maintenance strategies glued segmental box girder bridges,
activity ranking, 79-80 349-60
activity urgency, 80 highway bridges, 119
approach to, 182-3 prestressed concrete structures, 395-402
bridge adequacy, 81 remote computer-aided bridge
bridge criticality, 80 performance, 361-71
cable stayed bridges, 345-8 traffic-induced strain in steel
Clifton Suspension Bridge, 210-11 reinforcement, 621-2
cost aspects. See Maintenance costs Wolvercote Viaduct, 659-61
current research, 228-9 Mortar renderings, 251
deficiency point assignment, 81-3 Motorway bridges, maintenance strategies,
design criteria, as, 150 155-71
economics of, 143 Movable bridge machinery, 295-304
Europe, 29-38 inspection, 302-4
existing situation, 225-6 rehabilitation, 304
expert systems, 155-71 Movable bridges, types of, 296-8
Far East, 29-38 Museum Railway Bridge, 539-40, 545
Forth Rail Bridge, 12-14
Holland, 215-22
implications for new bridges, 228 Natural frequencies, 590
management of, 22-3, 47-8, 226-8 Ness Viaduct, 282
motorway bridges, 155-71 Network analysis, 724
optimization, 101-11 New-build costs, modelling, 188-90
Noise, 598
past, present and future, 147-9
PennDOT, 78-83 Non-linear analysis, 433-4
prioritization procedure, 79 Northern Ireland, reinforced concrete
bridge protection, 259-65
qualitative decision tree for, 217
reinforced concrete bridges, 261-2 Northern Ireland Roads Service, 128
systema tics approach to, 215-22
Tamar Bridge, 176, 178-85 Obernberg Bridge, 331
UK, 223-30 Ohre River bridge, 387-9, 391-2
see also DANBRO; DISK system; Olympic Grand Bridge, 340-2, 344, 347-8
Rehabilitation; Repair Ontario Bridge Code, 616
Management systems. See Bridge Operation costs, comparative study of,
management 199-204
Marie d'Ivry metro-station, 401 Optical fibre sensors, 399
Masonry bridges, 190-6 Orthotropic steel bridge decks
see also Arch bridges cracks in, 633-43
Massachusetts Bay Transportation direct measurements under traffic
Authority (MBTA), 68-71 loading, 466
Massachusetts Department of Public dynamic testing, 465-6
Works (MDPW), 67 effect of surfacing, 471-2
M-beam type bridges, 419-22, 425-6, 430 fatigue assessment of, 459-73
Method of Measurement for Road and transverse position of traffic, 471-2
Bridge Works (MMRB), 191 Overloaded vehicle permits, 410-11
Metropoliran District Commission,
Boston, Massachusetts, 66-8
MEXE analysis, 517 Paint degradation, 23-4
794 Subject Index

Painting and repainting Radar-contd.

Forth Rail Bridge, 12-13 data acquisition/processing, 312-13
Tamar Bridge, 178 defect detection by, 307-10
Pardubice fiyover, 390-1 field operations and experience with, 314
Peel Green underbridges, 447-57 operational characteristics of, 315
Pelly River Bridge, 581 Radiography, 24
Pennsylvania Bridge Management System, Railway bridges
75-87 fatigue fractures in, 533-5
Permits for overloaded vehicles, 410-11 fatigue tests on, 539-4 7
Pig-tail railway, 537 residual service life of old structures,
Plastic collapse modes, 418 537
Poisson's ratio, 354 theoretical and experimental
Portal slab bridges investigations on, 533-48
collapse analysis, 424 Rakewood Viaduct, 668
failure modes, 420 Regional infiuences on bridge design, 151
loadcase governing collapse, 424 Rehabilitation programmes
reliability analysis, 424 background studics, 138-40
sensitivity analysis, 426 concretc bridges, 139-40
Pothole formation, 555-6 cost-effective strategies, 715-24
Present value analysis, 715-17, 721 discount rate for, 718
Prestressed concrete bridges, 150 15-year, 136-8
beam assessment, 447-57 highway bridges, 135-43
comparison of main defects, 292 managcmcnt of, 141-2
composite I-girder, 389-90 movable bridge machinery, 304
concrete frame, 392-3 prestressed concrete bridges, 293-4
inspection of, 287-94 rescarch and development, 140-1
rehabilitation, 293-4 steel trusses, 695-704
segmental box girder, 391-2 Tamar Bridge, 180-2
shear strength of beams, 448 seealso Aluminium extrusion bridge
stress assessment, 439-46 rehabilitation system
Prestressed concrctc flyover, 390 Rehabilitation/replacement prioritization,
Prcstressed M-beam bridges, 419 83-6
failure modes, 420-2 Reinforced concrete bridges, 190-6, 387-9
probability of failure, 425-6 chloride-induced corrosion, 655-66
Prestressing of bridge decks, 668 condition survey, 263-4
Probability of failure, 418, 422, 426, 578 continuous girder, 386-7
PROBAN reliability program, 321 cracks in beams, 644
PROINSP computer program, 321 deterioration and repair of beams,
Project Ievel system, 110 643-53
Protective measures, 165, 259-63 deterioration of, 269-70
PR VAL Pro gram, 721 expansion joints, 739
Pull-out tests, 162 inspcction, 261
Punching shear, 508-9 Iabaratory investigations, 263
Punching theory, 432-3 maintenance strategies, 261-2
portal slab, 419
protective measures, 259-63
QUADRO, 49 stress redistribution in beams, 643-53
QUADR04, 23 Reinforcing bars,
Quality assurance, 19, 119 corrosion of, 23, 25, 245-7, 267-84, 643
see also FBECR
Reliability analysis
Raach Bridge, 331 advanced Ievel II method, 423-4
Radar assessment of bridges, in, 417-28
bridge deck condition surveys, 305 basic variables, 422-3
 Subfeet Index 795

Rehability analysis-contd. Shock transmission units (STUs), 671-4

deteriorating bridge structures, 501-13 Silane for reduction of chloride-induced
FORM, 321 corrosion, 655-66
Rehability decrcase with time, 564 Siedgehammer for bridge beam testing,
Rcmaining bridge life assessment, 409-10 604
Repair, 165 S-N analysis, 319-20
cable stayed bridgcs, 345-8 Span drive machinery, 298-300
composite concrete-steel bridges, 747-64 Special tests, 161-4
concrete removal in, 765-76 Specification criteria, 118
costs, modelling and predicting, 187-97 Stabilizing machinery, 298-300
highway bridges, 378-80 Stahringen Bridge, 538
materials, 24 Standards, 24, 143, 575-84, 587, 592
Residual life assessment, 718 Steel beams, 151
Residual value assessment, 718 Steel bridges
Retractilc bridges, 298 inspection, 319-26
Richemont Bridge, 634-5, 639, 640, 642 sec also Orthotropic steel bridges
Rijkswaterstaat, 215 Steel rcinforcement, traffic-induced strain
Road and Waterways Administration in, 619-29
(RWA), 101-ll Steel trusses, inspection and rehabilitation,
Road bridges. See Highway bridges 695-704
Road data bank. 55 STREG, 43
Road Engineering Intelligence and Strength assessment, 138
Research, I 16 bridge decks, 552
Roads and Waterways Administration concrete bridges, 429-38
(RWA), 65 conventional approach, 431-2
Rolled sections, reliability indcx, 325 highway bridges, 383-93
Royal Border Bridge, 520 Strengthening methods for existing
bridges, 667-76
Stress assessment in post-tensioned
Safety aspects, 25, I 71, 406-8 concrete bridges, 439--46
Safety index, 578-81 Stress intensity functions, 320
Salt Stress rcdistribution
crystallisation in concrete, 245 concrete beams, assessment procedure,
scaling in concrete, 245 647-53
SAMOA program, 155-71 reinforced concrete beams, 643-53
aims of, 156 Stress state correlation with deterioration,
flow chart, 158 509-11
global assessment, 159 Structural analysis
implementation of, 157 bridge decks, 552
record forms, 159 deterioration bridge structures, 504-5
Saudi Arabia, highway bridges, 549-61 Structural assessment, 24-5
Sazava River bridge, 386-7 transversal cracks, 489-99
Scaling, 245, 309-10 Structural dynamics modification (SDM)
Scleromctcr tests, 162 software, 335
SCNALL, 367 Structural inventory records system
Sensors (SIRS), 77-81
coppcr wire, 399-400 STRUDL, 552
optical fi brc, 399 Stuttgart, Rosensteinbrücke, 291
Shear failure, 421-2 Stuyvesant Falls Bridge, 696-8
tests on prestressed concrete beams, Sulphate attack, 244
447-57 Sulphate profiles, 559
Shear strength Surface treatments
measurcd and calculatcd, 453 concretc, 233--41, 250-1
prestressed concrete beams, 448 tests on coatings, 253--4
796 Subject Index

Surrey County Council Thermography-contd.

assessment strategy, 46-7 data acquisition/processing, 312-13
bridge management, 39-50 defect detection by, 307
capital programme, 42-3 field operations and experience, 313-14
current allocation, 42-3 operational characteristics, 315
inspection strategy, 44-6 Thrust lines, 515-16
maintenance management, 47-8 Torpoint Ferry Joint Committee, 173
priorities, 49 Torridge Bridge, 351, 353, 356
Surveillance methods, 161 Total deficiency rating (TDR), 83-5
Suspension bridges Tottnaes Bridge, 781-3
aerodynamics, 486 Traffic
analysis and survey, 477-8 bridge repair problems, 751-2
assessment and rehabilitation, 475-87 control devices, 301-2
loading, 476-7 delay cost, 49
safety factors, 479 disruption, 143
temperature effects, 478 influence on dynamic testing, 335-7
tension drop in cable-band bolts, 725-35 loading on orthotropic steel bridge
see also Clifton Suspension Bridge decks, 466
Swedish Commission on Maintenance and redistribution analysis, 23
Costs (DKU), 199, 203 strain induced in steel reinforcement,
Swedish National Road Administration, 621-2
200, 202 transverse position of, 471-2
ADP-based system, 54 volume, mix and weight estimates, 118
bridge age distribution, 52 Traffic-induced strain in steel
bridge management, 51-61 reinforcement, 619-29
bridge stock, 51-2 test programme, 623-4
condition classifications, 59 test results, 624-8
finance, 52-3 theoretical calculation, 622-3
inspection system, 57-9 Traffic Ioad simulation programme,
load-carrying capacity, 52, 59-60 563-74
Swing bridges, 296-7 aim of, 567-8
renovation, 705-13 basis and requirements for development
of, 568-70
description of modules, 570--3
Tacoma Narrows Bridge, 483 Iayout of, 570-3
Tamar Bridge, 173-86 principles of, 568
construction details, 175 usage of, 573-4
inspection, 177-8 Traffic Ioads, C1ause 12:, 577-8
maintenance contract procedures, 183-4 Transport and Road Research Laboratory
maintenance strategies, 176, 178-85 (TRRL), 140-1
overview, 185-6 Transversal cracks, effect on mechanical
painting and repainting, 178 behaviour, 489-99
pre-construction, 174-5 Truck weight regulations, 411-13
rehabilitation, 180-2 Turner proposal, 411
statistics, 173-4
traffic conditions, 175-7
Target reliability index, 580--1, 583 Dienbergstrasse Bridge, 400
Tay Bridge, corrosion of reinforcement, 267 Ulm, Gänstorbrücke, 287-8
Testing. See lnspection and under specific Uhrasonie tests, 162
test methods United States
Thermal incompatibility of concrete bridge management, 63-74
components (TICC), 559 highway bridges, 405
Thermography Untermarchtal prestressed-concrete bridge,
bridge deck condition surveys, 305 289-91
 Subject Index 797

Vapour permeability measurement, 254-5 Windowing, 598

Vehicle braking efficiency, 119 Windsor probe, 162
Vertical Iift bridges, 297, 298 Wolvercote Viaduct, 655
Vibration contract details, 665-6
bridge decks, 623 description of structure, 656
highway bridges, 607-17 equipotential contour mapping, 664
modes, 163 evaluation of patch repairs, 664
tests on masonry arch bridges, 585-94 haif-eeil potential surface mapping, 661
velocity measurements, 590 monitoring, 659-61
permanent instrumentation, 664
remedial work, 659
survey, 657-9
Water in concrete. See Concrete
Waterjcement ratio, 247
Waterproofing problem, 146-7 Yellow Mill Pond Bridge, 323
Welded joints Yield-line analysis, 432
cracks in, 637-9 Young's modulus, 354, 600
orthotropic decks, 466-8 Yukon Territory, 578-82