Bridge Engineering

Module Handbook

Level 7
2014-15

School of Computing, Science and Engineering.

The module team:

Dr Laurence Weekes BEng(Hons) PhD

MSc Programme Leader and Bridge Engineering Module
Coordinator. Laurence specialises in pre and post tensioned
concrete design and advanced analysis of structures.

Dr Jinyan Wang BEng MSc PhD PGCAP FHEA MIMS

Jinyan specialises in the advanced study of masonry arch bridges,
particularly soil-structure interaction including FEA modelling and
new masonry arch bridge assessment methods.

Dr Wayne Wang BSc MSc PhD GMICE PGCAP FHEA MIMA

Wayne is a lecturer in construction materials. He has abundant
research experience in the fundamental physics of concrete
deterioration and structural durability.

Mr Jonathan Haynes BEng(Hons) MSc CEng FIStructE MICE PgC FHEA

Jonathan is a former moving bridge designer and is now a part-time
masonry arch bridge researcher and structural engineering lecturer.
He is interested in how structural engineers learn.

Each year the module leader may award The Melbourne Prize, which is awarded for
excellent academic performance in Bridge Engineering. Professor Clive Melbourne retired
from civil engineering at Salford in 2012. He began his career as a bridge designer and
devoted much effort to advancing knowledge of the masonry arch bridge, being
convenor of the first International Arch Bridge Conference. Amongst many duties, for
which he received the Lewis Kent Award, he chaired the Institution of Structural Engineers
examinations panel for over a decade, was chairman of the Lancashire & Cheshire
Branch and instigated the North West Structural Awards.

excellent educ ation s ince 1897

Institution of Structural Engineers Award for Excellence in Structural Engineering Education 2012
Bridge Engineering
CRN: 32900
Credit Rating : 30 Credits
Level 7, Semester 2

Date Version Revision

Jan-13 1.0 Initial print for new module
Jan-14 2.0 Assessments updated, sections 3, 5, 6, & 8 added
Jan-15 3.0 Section 7 added and section 3 extended

Module Leader:
Dr Laurence Weekes
Newton Building, LG14, University of Salford, Greater Manchester, M5 4WT
email: [email protected]

Technical contents © Mr Jonathan Haynes

Civil Engineering website: www.cse.salford.ac.uk/civilengineering

Blackboard website:
http://blackboard.salford.ac.uk - BridgeEngineering

Directorate of Civil Engineering

School of Computing, Science & Engineering
College of Science and Technology
32900 Bridge Engineering
MSc Structural Engineering

CONTENTS
Module Details ....................................................................................................................... vi
The Aims of the Module ............................................................................................................................... vi
The Learning Outcomes of the Module.................................................................................................... vi
The Learning and Teaching Strategies of the Module........................................................................... vi
The Assessment Regime ............................................................................................................................... vi
Key Threads for MSc Structural Engineering............................................................................................. vi
Outline Syllabus.............................................................................................................................................. vi
Essential Reading........................................................................................................................................... vi
Module Schedule.......................................................................................................................................... vi
Study Plan ....................................................................................................................................................... vi
Assignment 1 Case Study. Marking and Feedback Plan ...................................................................... vi
Assignment 2 Conceptual Design and Scheme Appraisal. Marking and Feedback Plan ............ vi
Assignment 3 Analysis and Detailed Design. Marking and Feedback Plan.................................. vi
Client’s Brief .................................................................................................................................................... vi
Typical Assessment Criteria.......................................................................................................................... vi
A Bridge Lifetime Flowchart......................................................................................................................... vi
A Bridge Design Checklist............................................................................................................................ vi
1. Bridge Concepts................................................................................................................. 6
1.1 A Brief History of Bridge Engineering..................................................................................................... 6
1.2 The Four Internal Forces and Four Structural Elements ..................................................................... 6
1.3 Bridge Structural Forms............................................................................................................................ 6
1.3.1 Stability Systems ................................................................................................................. 6
1.3.2 The Outline Construction Sequence................................................................................. 6
1.3.3 Crane Data ........................................................................................................................ 6
1.4 Choosing Appropriate Materials........................................................................................................... 6
1.5 Scale of Use............................................................................................................................................... 6
1.6 Aesthetic Appeal ..................................................................................................................................... 6
1.7 Capital and Carbon Costing using CESMM ....................................................................................... 6
1.8 The Principles of Risk Management...................................................................................................... 6
1.8.1 Assessing Risk...................................................................................................................... 6
1.8.2 Work Activity Risk Assessment............................................................................................ 6
1.8.3 Risk Management.............................................................................................................. 6
1.9 Comparison of Schemes ........................................................................................................................ 6
2. Bridge Deck Loading & Analysis ....................................................................................... 6
2.1 Bridge Deck Loading............................................................................................................................... 6
2.1.1 Traffic Actions..................................................................................................................... 6
2.1.2 Wind Actions ...................................................................................................................... 6
2.1.3 Thermal Actions ................................................................................................................. 6
2.1.4 Earthquake Actions ........................................................................................................... 6
2.1.5 Snow Actions...................................................................................................................... 6
2.2 Local Analysis Models of Bridge Decks................................................................................................ 6
2.2.1 Distribution of Actions on the Deck .................................................................................. 6
2.2.2 Design Forces in Simple Spans .......................................................................................... 6
2.2.2 Analysis of Moving Actions on Continuous Beams using Influence Lines ....................... 6
2.2.3 Analysis of Moving Actions on Slabs using Influence Surfaces........................................ 6
2.3 Global Analysis Models of Bridge Decks ............................................................................................. 6
2.3.1 Grillage Analysis................................................................................................................. 6
2.3.2 Finite Element Analysis of Bridge Decks............................................................................ 6
3. Structural Element Design .................................................................................................. 6
3.1 Reinforced Concrete Decks .................................................................................................................. 6
3.1.1 Bending .............................................................................................................................. 6
3.1.2 Shear................................................................................................................................... 6
3.2 Steelwork Elements .................................................................................................................................. 6
3.2.1 Plate Girders....................................................................................................................... 6
3.2.2 Columns and Bracing........................................................................................................ 6

© Jonathan Haynes version 3.0 Page iv

32900 Bridge Engineering
MSc Structural Engineering

3.2.3 Slenderness and Deflection .............................................................................................. 6

3.2.4 End Bearing Stiffeners ........................................................................................................ 6
3.2.5 Trusses ................................................................................................................................. 6
3.2.6 Composite Beams ............................................................................................................. 6
3.3 Abutments and Piers ............................................................................................................................... 6
3.3.1 Calculating Forces and Settlements in Foundations ....................................................... 6
3.3.2 Geotechnical Design ........................................................................................................ 6
3.3.3 Concrete Struts .................................................................................................................. 6
3.4 Post-tensioned Concrete Beams .......................................................................................................... 6
3.4.1 Materials and Equipment.................................................................................................. 6
3.4.2 Principles of Analysis .......................................................................................................... 6
3.4.3 Design Considerations....................................................................................................... 6
3.4.4 Design Checks ................................................................................................................... 6
3.5 Box Girder Decks...................................................................................................................................... 6
3.5.1 Historical Background........................................................................................................ 6
3.5.2 Stress Analysis of Box Girders ............................................................................................. 6
3.5.3 Failure Criteria .................................................................................................................... 6
3.5.4 Design of Plate Structures Loaded Transversely (out-of-plane) ...................................... 6
3.5.5 Calculation of Effective Compression Flange Width....................................................... 6
3.5.6 Calculation of Design Stresses Due to Torsion and Warping........................................... 6
3.5.7 Design of Intermediate Diaphragms ................................................................................ 6
3.5.5 Construction Sequence .................................................................................................... 6
3.6 Bridge Fittings ............................................................................................................................................ 6
3.6.1 Bearings.............................................................................................................................. 6
3.6.2 Concrete Corbels .............................................................................................................. 6
3.6.3 Parapets ............................................................................................................................. 6
3.6.4 Movement Joints ............................................................................................................... 6
4. Arch Bridges........................................................................................................................ 6
4.1 Arch Bridge Analysis and Design .......................................................................................................... 6
4.1.1 Steel Rib Design.................................................................................................................. 6
4.1.2 Concrete Vault Design...................................................................................................... 6
4.2 Masonry Arch Bridges.............................................................................................................................. 6
4.2.1 MEXE Assessment ............................................................................................................... 6
4.2.2 Analysis of an Elastic Rib.................................................................................................... 6
4.3 Masonry Arch Bridge Research............................................................................................................. 6
5. Integral Bridges................................................................................................................... 6
5.1 The Development of Integral Bridges .................................................................................................. 6
5.1.1 Integral Bridge Arrangements........................................................................................... 6
5.1.2 Semi-integral Bridge Arrangements.................................................................................. 6
5.1.3 Limitation of Application ................................................................................................... 6
5.1.4 Construction Methods....................................................................................................... 6
5.2 Soil-Structure Interaction......................................................................................................................... 6
5.2.1 Effective Stress and Earth Pressure.................................................................................... 6
5.2.2 Soil Movement at Integral Bridges .................................................................................... 6
5.3 Design of Integral Bridges....................................................................................................................... 6
5.3.1 Spread Footings Supporting Abutments........................................................................... 6
5.3.2 Earth Pressure Estimates at Integral Abutments ............................................................... 6
5.3.3 Laterally Loaded Piles........................................................................................................ 6
5.4 Structural Modelling................................................................................................................................. 6
5.4.1 Beam Element Models ...................................................................................................... 6
5.4.2 Grillage Analysis................................................................................................................. 6
6. Long Span Bridges .............................................................................................................. 6
6.1 Components of Cable Stayed Bridges................................................................................................ 6
6.1.1 Cables ................................................................................................................................ 6
6.1.2 Cable Arrangements......................................................................................................... 6
6.1.3 Stiffening Girder ................................................................................................................. 6
6.1.4 Tower .................................................................................................................................. 6

© Jonathan Haynes version 3.0 Page v

32900 Bridge Engineering
MSc Structural Engineering

6.1.5 Cable Connections ........................................................................................................... 6

6.2 Erection Methods..................................................................................................................................... 6
6.2.1 Cantilever Method ............................................................................................................ 6
6.2.2 Installation by Progressive Launching............................................................................... 6
6.2.3 Installation by Span-wise Launching ................................................................................ 6
6.2.4 Construction on Scaffolding ............................................................................................. 6
6.3 Structural Behaviour................................................................................................................................. 6
6.4 Preliminary Design of Cable-Stayed Bridges ...................................................................................... 6
6.4.1 Preliminary Calculations for Cable Stays.......................................................................... 6
6.4.2 Back Span to Main Span Ratio ......................................................................................... 6
6.4.3 Pylon Height ....................................................................................................................... 6
6.4.4 Deck ................................................................................................................................... 6
6.5 Suspension Bridges................................................................................................................................... 6
6.5.1 Structural Arrangement..................................................................................................... 6
6.5.2 Analysis of Suspension Bridges .......................................................................................... 6
6.5.3 Construction of Suspension Bridges .................................................................................. 6
6.6 Finite Element Analysis of Cables.......................................................................................................... 6
6.6.1 Dynamics............................................................................................................................ 6
6.6.2 Damping ............................................................................................................................ 6
6.6.3 Cable stiffness .................................................................................................................... 6
7. Durability of Bridges............................................................................................................ 6
7.1 Durability Issues for Bridges..................................................................................................................... 6
7.2 Fatigue ....................................................................................................................................................... 6
7.2.1 Concrete............................................................................................................................ 6
7.2.2 Steel .................................................................................................................................... 6
7.2.3 Steel Cable ........................................................................................................................ 6
7.3 Degradation Systems .............................................................................................................................. 6
7.3.1 Concrete............................................................................................................................ 6
7.3.2 Steel .................................................................................................................................... 6
7.3.3 Masonry.............................................................................................................................. 6
7.3.4 Steel Cable ........................................................................................................................ 6
7.4 Sustainability Issues................................................................................................................................... 6
7.4.1 Bearing Replacement ....................................................................................................... 6
7.4.2 Safe Demolition.................................................................................................................. 6
7.4.3 New Concrete ................................................................................................................... 6
7.4.4 Foundation Scour .............................................................................................................. 6
7.4.5 Masonry Arch Repair and Rehabilitation Systems ........................................................... 6
7.4.6 Fibre-reinforced Polymer Composites .............................................................................. 6
7.5 Bridge Condition Assessment ................................................................................................................ 6
7.5.1 Periodic Assessment .......................................................................................................... 6
7.5.2 Monitoring Systems ............................................................................................................ 6
8. Design Data......................................................................................................................... 6
8.1 Section Properties of Fundamental Shapes........................................................................................ 6
8.2 Standard Force and Deflection Formulae.......................................................................................... 6
8.3 Structural Mechanics Formulae ............................................................................................................ 6
8.4 Units and Conversion............................................................................................................................... 6
8.5 Values of Common Structural Material Properties ............................................................................ 6
8.6 Moment and Shear Coefficients for Continuous Slabs..................................................................... 6
8.7 Concrete Reinforcement Data............................................................................................................. 6
8.8 Approximate Methods of Analysis........................................................................................................ 6
8.8.1 Portal Method for Sway Frames and Vierendeel Trusses................................................. 6
8.8.2 Centroidal Distance Method for Sway Frames ................................................................ 6
8.8.3 Cantilever Method for Sway Frames ................................................................................ 6
8.8.4 Truss Analogy for Plate Girders.......................................................................................... 6
8.8.5 Simplified Truss Analysis and Beam Analogy for Deflection ............................................ 6
8.9 Plate Girder Section Properties.............................................................................................................. 6

© Jonathan Haynes version 3.0 Page vi

32900 Bridge Engineering
MSc Structural Engineering

Module Details
Bridge Engineering is a 30 credit module taken as a compulsory module by the MSc Structural
Engineering and MEng Civil & Architectural Engineering. A major part of this module is shared
with another module which is taught to the MSc Transport Engineering and Planning and MEng
Civil Engineering programmes. The module presents detailed design of bridge elements in the
major construction materials e.g. steel, concrete and timber, including issues of sustainability,
durability and risk management.

The Aims of the Module

To enable students to demonstrate:
• a comprehensive practical ability to undertake design and analysis of bridge structures with
regard to structural form, materials and specific loading,
• critical awareness of cutting edge technologies used in transport infrastructure combined
with application of original work to generate a comprehensive understanding of the
subject.
• an ability to deal with incomplete data and work autonomously at a professional level to
solve open-ended design problems.

The Learning Outcomes of the Module

On completion of this module students should be able to:
• undertake original design of structures specifically related to bridges (abutments, bearings,
cables, beams, trusses and decks).
• generate and critically evaluate the output from various structural analysis methods related
to bridge structures, using relevant IT resources.

The Learning and Teaching Strategies of the Module

• Learning will be promoted during lectures and design studio work, supported by tutorials
and computer laboratory classes. Visual aids and demonstrations will be used as
appropriate and students will be encouraged to ask questions and to take part in short
discussions within the lecture context. Emphasis will be placed upon students developing
their own learning strategies for self and group study.
• Assessment will be equally split between an unseen end examination and a design exercise.
• The examination will be an open book, technical design assessment.
• The design exercise will require small groups of students to produce conceptual options
which address a unique and challenging client brief, critically evaluate and detail design a
scheme. It will create an opportunity for students to synthesise knowledge from the entire
syllabus, develop interpersonal and time management skills and prepare themselves for a
career in industry.
• The final output may be a combination of written report, drawings, sketches, poster, oral
presentation, computer model or physical model.
• Students will be encouraged to develop designs under staff supervision in the Design Studio
environment.
• Students will require access to a laptop computer which runs a range of software, including
Linpro, MathCAD, ANSYS, ACES and Excel.

The Assessment Regime

There are two assessment elements in this module: the design exercise portfolio and the
traditional examination. An outline of the assessments follows.

Examination
The examination paper will comprise of questions on any topic area from the syllabus. The
examination has a weighting of 50% and lasts two and a half hours. The examination is open
book, and as a result is relatively conceptual in nature.

© Jonathan Haynes version 3.0 Page vii

32900 Bridge Engineering
MSc Structural Engineering

Design Exercise
This element comprises five separate tasks. Students will conduct detailed design of bridge
elements using several types of material used in construction (e.g. steel, concrete, timber and
masonry). The design includes abutments, bearings, cables, trusses, etc. using case study
examples. This portfolio has a weighting of 50%. The design exercise will be submitted
electronically using Turnitin, the work will be undertaken in small groups, and as such individual
members of groups may not invoke extensions to submission dates as there is a single submission
for each group (this includes students with individual support plans).
Students will undertake a peer assessment exercise for this element of assessment. This will
involve marking other team members, in a secret vote, for time keeping, team working and
commitment to the project. These marks are used to adjust the team marks to account for
individual effort. Individual marks for group submissions will be calculated as follows:
Individual mark = mark for group submission x total individual mark
group average mark
Peer assessment may be moderated by the lecturing team. Students will be informed of their
peer assessment mark and the group average mark.

1. Case study 2. Conceptual design 3. Analysis and detail

What
and scheme appraisal design
Where Submit on Turnitin Submit on Turnitin Submit on Turnitin
How Electronic Electronic Electronic
Who Group Group Group
Submit week 20 Submit week 24 Submit week 32
When
4pm 13/02/15 4pm 13/03/15 4pm 08/04/15
% of module
10% 20% 20%
mark
Workload
35 hours 70 hours 70 hours
guide
D


JBM
H



thread
S

Assignment 1 – Case Study

A study of a famous or technically interesting bridge structure (e.g. it could be in very poor
condition), presented as a technical paper. The template available on Blackboard must be
used for this submission.

Assignment 2 – Conceptual Designs and Scheme Appraisal

Multiple viable concepts to achieve a predefined set of clients requirements.
A comprehensive assessment of the schemes, considering durability, sustainability, construction
sequence, aesthetics, risk management (cost and time – a spreadsheet is available on
Blackboard).

Assignment 3 – Analysis and Detailed Design

Analysis of the bridge superstructure and foundations. Typically a computer based model
supported by hand calculations to provide validation. A third party check will be undertaken of
another groups work.
Code based checks of bridge elements, such as the deck, piers, foundations and any support
super structure.

© Jonathan Haynes version 3.0 Page viii

32900 Bridge Engineering
MSc Structural Engineering

Key Threads for MSc Structural Engineering

In accordance with the programme’s accreditation as satisfying the requirements for Further
Learning towards becoming a Chartered Engineer, students should have appropriate exposure
to threads in Design, Sustainability and Health and Safety.
Design
As well as covering the specific design elements and assumptions for different types of material
used in bridge construction, this module focuses on how the main elements of bridges are
designed in a systematic way (for example, using mathematical analysis and computer
software). Design is covered within the examined syllabus and the design portfolio.
Sustainability
There is consideration of sustainability and durability of material used for the different elements of
a bridge. The aspect of sustainability is assessed in the design portfolio.
Health and Safety
There are no field exercises or laboratory work associated with this module so students do not
have a direct health and safety exposure. However, as part of the syllabus, students are required
to provide outline construction method statements and designer’s risk assessments as part of the
examination. The design portfolio includes the application of risk management as a selection
and comparison tool, which must include carbon costing.

Outline Syllabus
The syllabus to be studied will include:
• Bridge conceptual design to suit client requirements.
• Bridge deck loading.
• Analysis methods relating to bridge structures, influence lines and surfaces, grillage analysis,
Finite Element Analysis of bridge decks and assemblies.
• Detailed design of bridge elements using the principal construction materials - steel,
concrete, timber and masonry.
• Design of abutments, bearings, cables, beams, trusses etc.
• Prestressed concrete determinate and indeterminate systems, arches, integral bridges,
cable stayed and suspension bridges.
• Issues of durability and sustainability, and their relation to risk management.
It is imperative that students understand that Bridge Engineering is a broad and varied topic with
international application. A 30 credit masters level module cannot hope to comprehensively
deal with all issues in the topic, this module will provide a clear introduction to the majority of
challenges in bridge design.

Essential Reading
Gottemoeller, F. (1998). Bridgescape: The Art of Designing Bridges. Chichester: Wiley.
Hambly, E.C. (1991). Bridge Deck Behaviour. 2nd Edition. London: E&FN Spon.
O'Brien, E. & Keogh, D.L. (1999). Bridge Deck Analysis. London: Spon.
Parke, G.A.R. & Hewson, N.R. (Eds). (2010). ICE Manual of Bridge Engineering. London: Thomas
Telford Publishing.
Ryall, M.J. (2010). Bridge Management. 2nd Edition. Oxford: Butterworth Heinmann.

The ICE journal, Bridge Engineering is available free of charge through the library search engine
when logged into the university network. It provides excellent quality papers on general bridge
engineering topics, and of particular use when writing a case study.
Previous examples of Assignment 1 submissions are available in the Salford Journal of Civil
Engineering.

© Jonathan Haynes version 3.0 Page ix

32900 Bridge Engineering
MSc Structural Engineering

Module Schedule
Week
Morning Afternoon
No
1
2
3
4
5
6
7
8 *a Assignment No 1 submission
9 *b Assignment No 2 submission
10 *c Assignment No 3 submission
11 *d Constructionarium week
12
13
14 Christmas vacation
15
16
Semester 1 Examinations
17
18 Inter-semester break
19 Bridge concepts JH Bridge concepts JH
20 *a Deck loading JH Bearings, parapets and services JH
21 Durability & sustainability WW Bridge Assessment and monitoring WW
22 Hand analysis of decks JH Grillage and FEA of decks JH
23 Steel elements JH Steel elements JH
24 *b Reinforced concrete elements JH Reinforced concrete elements JH
25 Integral bridges JH Abutments and SSI JH
26 *d Masonry arch bridges JW Masonry arch bridges JW
27
28 Easter vacation
29
30 Cable stayed bridges JH Suspension bridges JH
31 Steel box girder bridges JH Prestressed concrete box girders LW
32 *c Examination preparation (Bank Holiday Monday)
33
34 Semester 2 Examinations
35

Study Plan
It is important to devote sufficient time to studying at level 7 since the level of understanding and
critical review is significantly greater than at bachelor’s level. As a guide, Bridge Engineering
should be divided as follows:
Hours Mode
Assessed student led design, application of material from lectures
175 Design Exercise and background research. Contributes to knowledge and ability
required for examination.
60 Lecture / tutorial Contact with lecturers is only a fifth of the study time
6 Exam preparation Lecturer led guidance on passing the examination
Students should be spending an equal amount of time in lectures
59 Self study
and on self study (background reading, researching etc.)
300 this is the minimum expected study for a highly capable student

© Jonathan Haynes version 3.0 Page x

32900 Bridge Engineering
MSc Structural Engineering

Assignment 1 Case Study. Marking and Feedback Plan

20-39% Unsatisfactory
Mark breakdown

60-79% Very Good

80-100% Excellent
0-19% Very Poor
Possible Mark
Assessment breakdown

40-59% Fair
Identification of the case study subject.
Location.
Designer, constructor, owner, cost, construction period.
Type and use.
Age, environmental conditions. 50
Geometry and technical arrangement drawings including
foundations and details of interest.
Materials.

Structural form.
Assessment of aesthetics.
Loading regime and analysis assumptions.
20
Fatigue / dynamic behaviour.

Assessment of structural performance, durability and

sustainability issues - critical review of bridge.

20

Presentation in compliance with university requirements.

Referencing and citation using Harvard APA 6th
Maximum of 2000 words or 4 sheets A4 paper (writing must be
10
very concise), 10 point Times New Roman.

Mark awarded : %

Comments:

© Jonathan Haynes version 3.0 Page xi

32900 Bridge Engineering
MSc Structural Engineering

Assignment 2 Conceptual Design and Scheme Appraisal.

Marking and Feedback Plan

20-39% Unsatisfactory
Mark breakdown

60-79% Very Good

80-100% Excellent
0-19% Very Poor
Possible Mark
Assessment breakdown

40-59% Fair
One concept shall be provided by each student in the
group. Using annotated sketches, propose distinct and viable
structural forms for the superstructure and substructure of
each concept. Indicate clearly the functional framing, load 35
transfer and stability aspects of each concept. Include
scheme stage sizing to ensure the schemes are feasible as
proposed.
Giving reasons, identify the most appropriate materials for the
structural elements of each concept. 10
Provide a general arrangement sketch which identifies all
principal dimensions and structural elements (including
bearings, parapets, drainage and foundations) which will be 5
designed in Assignment 3. Drawings should be no larger than
A3.
Critical review of durability and sustainability issues in each
concept. 10
Critical review of the aesthetic appeal of each concept.
10
Annotated pictorial outline construction sequence for each
scheme including temporary works. Include identification of
significant construction health & safety risks. Identify the four 10
most significant risks in terms of time delay and cost over-run
during construction.
Estimate of capital and carbon cost for the whole life of the
bridge for each scheme, including temporary works. 10
Qualitative and quantitative assessment of the schemes, with
a recommendation of which one to take to detail design
stage (no amalgamations of schemes will be acceptable).
The assessment must consider, at least, aesthetics, capital 10
and carbon cost (permanent and temporary works),
buildability, construction period, environmental harm, whole
life cost, durability, structural efficiency.
Mark awarded : %

Comments:

© Jonathan Haynes version 3.0 Page xii

32900 Bridge Engineering
MSc Structural Engineering

Assignment 3 Analysis and Detailed Design.

Marking and Feedback Plan

20-39% Unsatisfactory
Mark breakdown

60-79% Very Good

80-100% Excellent
0-19% Very Poor
Possible Mark
Assessment breakdown

40-59% Fair
Deck loading calculations.
10

Establish an appropriate computer analysis model,

substantiating input data, including section and material
20
properties, restraints (bearings) and support conditions.

Independent verification of input data by another team.

10

Validation models using hand calculations.

10

Design of deck.
15

Design of support superstructure.

15

Structural design of piers and substructure.

15

Fully code referenced and internally referenced.

Hand written Salford calculation sheets only.
Neatly and logically presented. 5
Littered with supporting diagrams.

Mark awarded : %

Comments:

© Jonathan Haynes version 3.0 Page xiii

32900 Bridge Engineering
MSc Structural Engineering

Client’s Brief

1. A bridge is required to carry two new railway lines over an existing railway junction cutting.
Construction is to take place whilst the existing lines are live. The cutting is to remain unchanged. A
general section through the railway cutting is shown below.
33 m

+0.0 m
5.75 m

emax Medium stiff CLAY

emin Cu = 70 kN/m2 at -3.0m depth
increasing to
emin Cu = 100 kN/m2 at -6.0m depth.
I f 'I

2. A client requires a horse-shoe cantilever observation deck at a canyon edge. The deck must permit cars
to be driven (one-way) around the horse-shoe. The structure should impede visitors outward views as
little as possible. A general section through the site is shown below.
14.0 m

G.L +68.23 m

medium dense SAND, allowable

bearing pressure 100 kN/m2
+66.15 m

horizontally bedded
SANDSTONE
average allowable bearing
pressure 700 kN/m2

70m high
escarpment

one-way travel

© Jonathan Haynes version 3.0 Page xiv

32900 Bridge Engineering
MSc Structural Engineering

3. A client requires a wild cat observation facility within an existing zoo enclosure (the animals will be
removed whilst construction takes place). The enclosure is virtually level and 80m square. Members of
the public must have access to a 10m square deck at the centre of the enclosure, elevated 4m above
ground level. Any foundations for the deck must be located in the central 5m of the enclosure. There are
no restrictions on structure or foundations outside the enclosure.

46 m

fence around animal observation deck

enclosure 10 m

foundation zone
4m

+0.0 m

5m Ground water
level -1.25 m
medium dense SAND
N = 14 at -1.1m depth

-4.0 m

bedded SANDSTONE
average allowable bearing
pressure 1000 kN/m2

4. Salford City Council require a road bridge across the A6 at Pendlebury, where it passes through a
600mm thick reinforced concrete culvert. The bridge is expected to form a significant landmark and will
be the only access to a new retail park. It must carry at least two lanes of 3.6m wide road traffic and
segregated foot/cycle ways. The minimum clearance for vehicles is 5.0 m.

+48.73 m +47.92 m

+42.7 m

32.4 m

© Jonathan Haynes version 3.0 Page xv

32900 Bridge Engineering
MSc Structural Engineering

5. Salford City Council require a pedestrian crossing of the River Irwell between Elton Street and Saint
Simon Street (adjacent to Blackfriars Road). The footbridge should be architecturally striking. The soil
is alluvial clays of at least 12.0 m depth over red sandstone bedrock. The minimum clearance for river
users is 4.0 m.

+54.04 m
HWL +48.50 m
+51.35 m
LWL +46.10 m

4.6 m 28 m 11.2 m

6. A client requires a landmark footbridge across the Manchester Ship Canal at Salford Quays, adjacent
to a new television studio and broadcast facility. The canal is 38m wide and a section through the canal
wall is shown below. A space 6.0m x 6.0m is to be allocated at each quay wall for bridge support but they
are offset by 4.0m on plan. A clearance of 5.0m above dock water level is required for the intermittent
passage of ships, over the central half of the span.

4.0 m

6.0m x 6.0m
landing area

dock water -1.5m +0.0 m ground water level -1.5 m

dock infill SAND

N = 8 at -1.0m depth
-2.0 m

bed rock average allowable

bearing pressure 800 kN/m2

© Jonathan Haynes version 3.0 Page xvi

32900 Bridge Engineering
MSc Structural Engineering

7. A client is constructing a multi-storey car park on a 30.0m x 45.0m city centre site, which is fully
surrounded by 3.0m wide pedestrian footpaths and 7.2m wide roads. The client also owns car parks on the
roofs of adjacent buildings and wishes to link them together with bridges

+9.60 m
+8.00 m
8.0 m

10.0 m
+1.20 m
+0.00 m

ground water level -0.5 m

-5.1 m
horizontally bedded MUDSTONE
average allowable bearing pressure
600 kN/m2

8. A pedestrian bridge is required in a remote upland national park location. The structure is likely to be
lightly used but must be highly durable with very little maintenance. A general section through the site is
shown below.

54.0 m

2.0 m

vertically bedded
GRITSTONE
average allowable bearing
30.0 m pressure 400 kN/m2

Surface joints are friable and

loose

© Jonathan Haynes version 3.0 Page xvii

32900 Bridge Engineering
MSc Structural Engineering

9. The University of Salford has purchased the disused Police station on the A6, opposite the Maxwell
Upper Hall. A footbridge is required to link the second floor of both buildings. Permission has been
obtained to site supports in the central reservation of the A6 if desired. Sufficient clearance for HGV’s is
required. There is 4.2m of made ground, over a thin layer of sandy clay on bedrock, under the central
reservation of the A6. The bridge must be a landmark.

2nd Floor +48.13 m

+47.73 m 2nd Floor

+40.7 m

58 m

10. An electrified light rail system requires a new station with access to three platforms (staircase and a
lift). Public access is only available from one side of the running lines. A general section through the
proposed railway station is shown below. The trams are 30.0m long and the platforms are 42.0m long.

3.0 m 6.0 m 8.0 m 6.0 m 3.0 m

+52.10 m Public
access
+51.85 m

+48.10 m
+47.10 m

Medium dense SAND

Ground water level N = 25 at -2.0m depth
+45.25 m increasing to
N = 45 at -6.0m depth.

11. A new Metrolink line requires a bridge across a narrow lower Pennine valley. A general
section through the site is shown below.
100.0 m
+91.25 m

competent
unweathered
+84.6 m gritstone, overlain
by a superficial
+83.0 m layer of clay.

Ground water level -0.50 m

loose SAND and cobbles to depth
N = 13 at -3.0m depth.

© Jonathan Haynes version 3.0 Page xviii

32900 Bridge Engineering
MSc Structural Engineering

12. A pedestrian bridge is required to cross a dual-carriage way in Salford. A minimum clearance
envelope has been set by Salford Council, no structure of any kind may be placed within this zone. There
are public parks on each side of the highway, so there are no restrictions on the bridge outside the
clearance envelope. A general section through the site is shown below.
horizontal clearance
21.6 m
vertical clearance
4.8 m

+0.00 m

ground water level

-1.0 m
loose SAND
N=8
-8.5m
medium dense SAND
N = 22 -11.5 m

bedded SANDSTONE
average allowable bearing pressure 1000 kN/m2

13. A two line rail bridge is required across a narrow valley in open countryside. A general
section through the site is shown below.

80.5 m
+48.00 m

+42.00 m

Ground water level -1.30 m

3.50 m

loose SAND and cobbles

N = 12 at -2.0m depth.

very stiff CLAY

2.00 m

Cu = 500 kN/m2 at -4.0m depth.

Competent rock

© Jonathan Haynes version 3.0 Page xix

32900 Bridge Engineering
MSc Structural Engineering

14. A pedestrian bridge is required to link the communities of Prestolee and Little Lever to Darley Park,
by connecting three banks of the River Irwell. The bridge must fit into its rural setting. The site is prone
to seasonal flooding as the river banks are only 1.20m above mean water level, no river access headroom
is required. A general plan of the site is shown below.

The site investigation shows 1.20m of topsoil,

overlaying 6.1m of clay stiffening with depth, over
sandstone rock head.
Ground water was encountered at -0.85m
Cu = 80kN/m2 at -2.50m
Cu = 125kN/m2 at -3.50m
Cu = 190kN/m2 at -5.50m

15. A two lane road bridge with footpaths is required to cross a canal and unclassified road, in an
upland rural location. There must be at least 4.20m headroom over both canal and existing road.
A general section through the site is shown below.

38.0 m 12.5 m 22.5 m

+120.20 m

+113.25 m

+108.35 m

+102.00 m

Ground water level -0.30 m

2.25 m

dense SAND
N = 52 at -1.0m depth.

medium stiff CLAY

1.75 m

Cu = 100 kN/m2 at -3.0m depth.

Competent rock

© Jonathan Haynes version 3.0 Page xx

32900 Bridge Engineering
MSc Structural Engineering

16. A two lane road bridge is required to cross a deep ravine. The road alignment is cut through a series
of granite peaks, being 8.0m diameter bores through solid granite. The structure must be visually striking.
A general section through the site is shown below.

51.0 m

28.0 m
alignment straight on plan
but inclined at 4.0 degrees
vertically.

horizontally bedded Granite

average allowable bearing
pressure 40,000 kN/m2
48.0 m
Surface is solid and stable.

17. A two lane road bridge is required to cross an A road, on a flat site in the Cheshire Plain. The vertical
alignment is shown below and the roads are horizontally skewed at 30 degrees. Since the environment is
flat, it is required that the bridge forms a landmark for motorists.

minimum clearance 100m radius

envelope 4.2m x 16.0m

ground water level 1.20m 100m radius stiff CLAY

Cu = 70 kN/m2 at -3.0m depth
increasing to
Cu = 100 kN/m2 at -6.0m depth.

© Jonathan Haynes version 3.0 Page xxi

32900 Bridge Engineering
MSc Structural Engineering

18. A footbridge is required to span over a pedestrianised street, between the roofs of two buildings at
MediaCity. The two buildings overlap on plan by 2.4m but the bridge must be at least 4.0m wide and
include social space, seating and planters. No loading may be applied to the buildings but supports may
be positioned in the street, so long as there is adequate access for vehicles. The single carriageway road is
7.2m wide.

14 m
+40.5 m +40.7 m Site Investigation data:
Medium dense SAND
N = 25 at -1.0m depth
increasing to
+36.2 m N = 47 at -4.0m depth.
Water table at -1.2m.

2.4 m

19. A two lane road bridge is required to span across the River Irwell, it should include standard parking
bays along its length on both sides of the carriageway. At least four bays must be designated for disabled
use. Pedestrian footpaths are required on both sides of the carriageway. A dedicated cycle way, 3.0m
wide, is required on at least one side of the carriageway. Motorists should have to walk no further than
12m to a parking ticket machine. A minimum clearance of 3.6m must be maintained between water level
and the underside of the bridge structure.

36 m

+32.0 m
+28.4 m
+27.9 m

20. A two lane road bridge is required to cross a steep river valley. Separated provision for pedestrian
and cycle users is required. A general section through the site is shown below.
86.0 m

bottom of valley to be free

of structural support due to
42.0 m
existing road and river use
horizontally bedded limestone
average allowable bearing
pressure 800 kN/m2

surface 2.0m is peaty sand

© Jonathan Haynes version 3.0 Page xxii

32900 Bridge Engineering
MSc Structural Engineering

21. A bridge for pedestrian and cycle users is required. The site is a busy roundabout at the junction of
four 7.2m wide carriageways. A general section through the site and plan on the roundabout are shown
below. There must be clearance for standard HGV’s. Users must be able to leave the junction by any exit
without crossing the path of vehicular traffic.

2.0m wide cycle and footway in both

directions at each carriageway.

+120.0 m
22 m
ground water
level -1.2 m
very loose SAND
N=5
-1.75m
medium dense SAND
N = 26 -3.5 m

bedded MUDSTONE
average allowable bearing pressure 450 kN/m2

22. A pedestrian footbridge is required to cross a motorway cutting through a ridge. The footpath is part
of a long distance upland route and runs along the top of the horizontal ridge, which is formed in
saturated chalk and has an allowable bearing pressure of 100kN/m2. The prevailing wind direction is onto
the side of the proposed bridge. The chalk may be battered at a maximum angle of 70o.

chalk ridge which motorway

proposed footbridge access level is to pass through

proposed motorway profile requires

minimum headroom of 6.2m
+106.1 m
45o
40o

+87.2 m

25 m

© Jonathan Haynes version 3.0 Page xxiii

32900 Bridge Engineering
MSc Structural Engineering

Typical Assessment Criteria

Outstanding 90-100% Excellent 80-89% Very Good 70-79% Good 60-69% Satisfactory 50-59%
Technical content and its interpretation & evaluation
Scope Outstanding clarity of focus, Excellent clarity of focus, Clear focus. Very good Clear scope and focus, with Scope evident and
includes what is important, boundaries set with no setting of boundaries, some minor omissions or satisfactory but with some
and excludes irrelevant significant omissions or includes most of what is unnecessary information. omissions and unnecessary
information. All required unnecessary information. relevant. All required issues All required issues covered information. All required
issues addressed at least a All required issues addressed addressed at least a very at least a good level. issues covered at least an
very high level. at least a high level. good level. adequate level.
Understanding Outstanding with critical Excellent with critical Very good with critical Good with some awareness Basic with limited awareness
of subject awareness of relevance of awareness of relevance of awareness of relevance of of relevance of issues. Ideas of relevance of issues.
matter issues. Exceptional issues. Excellent expression issues. Very good expression are expressed well, with Limited but satisfactory
expression of ideas, and of ideas, some originality. of ideas, potential for some minor limitations. expression of ideas.
evidence of originality. originality.
Use of sources Outstanding collection of Information collection of Information collection of Good information Adequate engagement
pertinent information, very high standard, relevant high standard specific to collection, relevant to the with relevant information
almost exclusively from to assignment and mostly assignment and mostly from assignment, significant collection, reasonable
primary sources. from primary sources. primary sources. fraction from primary fraction from primary
sources. sources.
Critical analysis Outstanding analysis of Highly coherent analysis of Very good critical analysis Critical analysis of collected Analysis of collected
based on collected information, with collected information, with of collected information, information with some information evident but
evidence an understanding of bias. an understanding of bias. with an understanding of understanding of bias. uncritical. Reasonably
Compelling discussion. Very persuasive discussion. bias. Persuasive discussion. Convincing discussion. convincing discussion.
Report structure, presentation and clarity
Presentation Superbly presented. Crystal Very well presented. Very Well presented. Clear Satisfactory presentation Satisfactory presentation
clear throughout. clear throughout throughout and generally clear. but not always clear.
Spelling, Outstanding written Excellent written language, Very good written language Good written language. Acceptable written
grammar and language. Flawless. with only minor stylistic flaws. with few, very minor errors. Some minor errors, but none language. Some errors in
syntax affects clarity. punctuation, spelling,
sentence construction.
Referencing Referencing perfect Referencing almost perfect, Referencing very good, but Referencing good but with Acceptable attempt made
throughout. only very minor errors. with some errors. some errors and/or at proper referencing, with
omissions. a number of
errors/omissions.

© Jonathan Haynes version 3.0 Page xxiv

32900 Bridge Engineering
MSc Structural Engineering

Unsatisfactory 40-49% Inadequate 30-39% Poor 20-29% Very Poor 10-19% Extremely Poor 0-9%
Technical content and its interpretation
Scope Inadequately scoped, with Very vague definition of Extremely confused Information presented No awareness of scope of
significant omissions and topic with limited relevant perception of topic with the almost all irrelevant to topic or presentation of any
unnecessary information. information and much majority of the information assignment. Nearly all issues relevant information.
Some issues not addressed irrelevant information. being irrelevant to the either not addressed or at Essentially no issues
or all issues addressed but Several issues not addressed assignment. A majority of an unsatisfactory level. addressed.
some are at an inadequate or all issues addressed but a issues either not addressed
level. number are at an or at an unsatisfactory level.
unsatisfactory level.
Understanding Inadequate understanding Very shallow understanding Some significant Subject misunderstood in Total misunderstanding of
of subject with little awareness of with many relevant misunderstandings which the main, with significant subject.
matter relevance of issues. elements omitted. prevent coherent errors and omissions in
discussion. knowledge.
Use of sources Inadequate use of primary Insufficient collection of Poor collection of primary Unusable primary data, No evidence of collection
information for purposes of primary information. information. through inadequate of primary data.
assignment. collection or
methodological flaws.
Critical analysis Vague analysis displaying Very vague analysis with Extremely limited and No analysis or discussion No analysis or discussion.
based on lack of clarity or focus on apparent contradictions / largely unsuccessful beyond general
evidence purpose of the assignment. errors. Limited discussion attempt at analysis or speculation.
Limited discussion. discussion.
Report structure, presentation and clarity
Presentation Inadequate presentation. Poorly organized and presented with some information No attempt to present work in acceptable format.
and Information can be difficult to understand. Presentation hinders presentation of
communication followed and understood key themes.
only with effort.
Spelling, Enough errors in Significant errors in Coherence and structure of Almost complete lack of Assignment
grammar and punctuation, use of words, punctuation, use of words, argument is fundamentally comprehension with incomprehensible due to
syntax spelling and sentence spelling, sentence obscured due to poor use argument / information only levels of incorrect spelling,
construction, that the construction, making of language. vaguely understandable grammar and syntax.
meaning of the text is
arguments difficult to due to very poor use of
obscured.
understand. language.
Referencing Inadequate attempt made Unsatisfactory attempt Poor attempt made at Very poor attempt made at Essentially no attempt made
at proper referencing – made at proper referencing proper referencing – proper referencing – almost at proper referencing.
significant number of – large number of majority are missing or all are missing or incorrect.
errors/omissions. errors/omissions. incorrect.

© Jonathan Haynes version 3.0 Page xxv

32900 Bridge Engineering
MSc Structural Engineering

A Bridge Lifetime Flowchart

clients requirements
define the function of
the bridge

Design
Bridge
Design
Bridge

establish alignment with transport engineer

Function defines loading. Alignment, space

restrictions, proportion and scale identify
Scheme design stage
possible structural Forms. Select appropriate
materials. Arrange structure with due
account of aesthetic guidance.

identify a construction method and assess the

Documents: risks and costs. If necessary undertake a cost-
benefit analysis to select the best scheme for
detail design.
Drawings
Specifications
Outline construction method
Contract

Third Party design check

Risk register
Detail design stage

deck loading and superstructure wind loading

validated analysis model
element code compliant design

several contractors
tender to construct
the works – price, general arrangement drawings
programme, quality detail drawings

appoint
contractor

Documents:

As-built drawings and

Construct the works - Health & Safety file, used
hand over to client for subsequent periodic
assessments.

Periodic condition assessment

until end of life.

Demolition.

© Jonathan Haynes version 3.0 Page xxvi

32900 Bridge Engineering
MSc Structural Engineering

A Bridge Design Checklist

Site investigation Locate site, underlying geology, necessary soil tests, specify GWL -
ground water level
Alignment Straight or skew, vertical curve or gradient
Clearance envelope Required clearance may affect alignment or elevation
River levels Tidal or flood prone rivers will be subject to changes in water level.
If a support is being created in a river bed, there must be
consideration of hydraulics – will scour be a potential problem;
can a caisson be created safely ?
HWL - high water level, LWL - low water level
Construction sequence Many bridge forms will experience their most onerous loading, or
restraint conditions, during construction, this may include changes
in internal force distribution and the position of compression flange
restraints etc
Cranage The need for a mobile crane to access the site or reach a
particular location on the site may be critical to the construction
sequence
Loading regime The use of the bridge sets the loading regime to be considered
Bearings The inclusion (or not) and location of bearings with have a
significant effect upon the analysis results, use the simplest
arrangement and form of bearing where possible; different
bearings may be needed for the construction stages
Expansion joints All bridges need expansion joints. This may only be where the
bridge ends, but can also be within the bridge structure
Analysis model The analysis model must reflect the as-built bridge; its geometry,
lines of stiffness, rotational and positional restraints and loadings.
This will generally entail a computer model and a hand validation
model
Deck waterproofing It is necessary to provide a waterproofing layer across the entire
deck, below the wearing surface, and must reflect the flexibility of
the deck
Deck drainage Provision must be made to collect water from the deck and
discharge it safely and without detriment to the long term
durability of the bridge
Lighting There may be a need to provide street lighting on the bridge, or
decorative lighting on landmarks
Parapets / handrailing Falling from the edge of the deck (vehicles or people) must be
prevented, as must collision with oncoming vehicles on motorways
and dual carriageways
Corrosion protection Metallic parts must be protected against corrosion for a
reasonable period (not necessarily the life span of the bridge). In
some circumstances concrete reinforcement will need protection
Maintenance and repair Where regular maintenance is anticipated, provision must be
made to undertake it safely. Many large bridges include an
inspection gantry
Demolition Every bridges must eventually be demolished, so it is important to
consider how this may be done. Remember, every structure must
have a Health & Safety File which provides method statements for
maintenance and repair procedures, and a demolition sequence

© Jonathan Haynes version 3.0 Page xxvii

32900 Bridge Engineering
MSc Structural Engineering

1. Bridge Concepts
This section will introduce the process of identifying and developing feasible bridging options,
based upon the requirements of a client brief. It will present simplified structural options,
guidance for choosing the best structural material, and address the growing fields of carbon
costing and risk management.
You should also review: 1 Bridge Concepts.ppt on Blackboard
Principal References: Mainstone, R.J. Developments in structural form.
Gottemoeller, F. Bridgescape: The Art of Designing Bridges.

The design of bridges has been simplistically summarised as spanning space and is an area of
structural engineering which truly encompasses science and art. The identification and selection
of a bridge scheme is probably closer to art than science, which in part explains why many
iconic bridges involve the input of an Architect. There is no reason why the Engineer should not
undertake the full design of a bridge.

1.1 A Brief History of Bridge Engineering

Mankind’s first attempts at permanent bridging are likely to have been little
more than ropes slung between river banks, or felled trees.
The first modern bridging form to be developed was the masonry arch, largely
perfected by the Romans (see below).
Many timber forms were developed and used in the middle and far East, many
of which have remained unchanged in China and Japan.
An expansion of bridging form begins with the early industrial revolution and its
demand for better transportation links, fuelled by experimentation with new material
technologies.
The Iron Bridge at Coalbrookdale is probably the beginning of modern bridge construction,
completed in 1779. It is the first use of cast iron in a bridge structure. The form is a series of iron
arch ribs joined using timber construction techniques. Cast iron has low tensile strength so a
compression element is the best choice for this material.
Wrought iron replaced cast iron in situations which develop tension but irons proclivity to brittle
failure was only solved when Bessemer developed a system to make industrial quantities of steel
in 1865. By 1890 steel was the engineers construction material of choice for heavy loads and long
spans. Steel was joined by riveting until bolting was seriously developed after WW2. The first
welded bridge was not constructed until 1927 but welding was not metallurgically understood for
several decades afterward.
Though iron and steel were the new wonder materials, brick masonry was the Victorian railway
engineers preferred material and the segmental multi-ring arch was the ideal form to give long
spans with cost efficiency. Over 40% of European road and rail bridges are masonry arches, most
being around 150 years old.

Pont du Gard Castlefield lattice web girder and cross heads Liverpool St. rail bridge
Liverpool Street station in Manchester, was the first passenger railway terminal (now part of MOSI)
and still uses the original built-up plate girder overbridges which became synonymous with

© Jonathan Haynes version 3.0 Page 1 of 239

32900 Bridge Engineering
MSc Structural Engineering

Victorian railways. These, and lattice web girders were greatly developed in the late 1800’s,
eventually leading to the modern steel plate girder.
The development of long span suspension bridges begins with high level crossings such as
Telford’s 176m, 1826 Menai straits crossing and Brunel’s 214m, 1864 Clifton suspension bridge,
both of which use chain links. The first spun wire rope suspension bridge was Grand Pont
Suspendu, Fribourg in 1834 but at 40m span is less significant that the Niagara Falls crossing in
1848.
The story of suspension bridges is not complete until the construction of river caissons is
developed, firstly with the Brooklyn Bridge. The Forth Road Bridge opened in 1890 and was an
ambitious estuary crossing which adopted the revolutionary concept of cantilever piers. It was
the first UK bridge structure to be constructed in steel.
The main structural tubes were
produced on site by fabricating beam
elements from angles and plates,
which were riveted to plates to form
large diameter tubes.
Construction of the caissons used a
fabricated steel tube with a cutting
toe. The caisson was located on the
estuary bed, weighed down with
kentledge and the internal water
pumped out. Excavation of the estuary
bed could take place in the now dry caisson, which descended under its self weight. To prevent
ingress of water, the excavation chamber was pressurised with compressed air (to balance the
pressure of water outside). This is the same environment which deep sea divers work in and many
operatives suffered from the bends, some fatally.

The rise of concrete bridges is largely due to Eugène Freyssinet, who developed post-tensioning
systems to allow concrete to span significant distances.
The ubiquitous post-tensioned motorway junction bridge is perhaps the worst example of dull,
minimum cost infrastructure which civil engineers engage in. Although functional, few are
expected to reach their design life of 120 years.

The Iron Bridge Blackfriars Bridge a new M1 overbridge

© Jonathan Haynes version 3.0 Page 2 of 239
32900 Bridge Engineering
MSc Structural Engineering

The box girder bridge became the benchmark form for urban flyovers in the 1960’s, particularly
when launched segmentally. Three collapses led to greater understanding of its behaviour in the
1970’s.

Duke St. rail viaduct Clifton suspension bridge Mancunian Way

Learning from failure
Broughton Suspension Bridge was built in 1826 to span 44m across the River Irwell in Salford. It was
the first wire cable suspension bridge constructed in Britain. John Fitzgerald, the wealthy owner of
Castle Irwell House, built the bridge at his own expense, all users of the bridge were required to
pay a pontage to cross.
On 12th April 1831, the bridge collapsed, reportedly due to resonance induced by troops
marching over the bridge in step, as a result the British army issued an order that troops should
break step when crossing bridges. The bridge was rebuilt and strengthened after the collapse
but was propped with temporary piles whenever a large crowd was expected. In 1924, it was
replaced by a Pratt truss footbridge (painted green), which is still in use.
The Tacoma Narrows Bridge opened to traffic in July 1940, it was the third longest suspension
bridge in the world. Following deck construction, it would move vertically in windy conditions,
which led to the bridge acquiring the nickname Galloping Gertie. Several measures aimed at
stopping the motion were ineffective or destroyed by the bridge motion, so a 1:200 scale model
of the bridge was tested in a wind tunnel and the study concluded two possible solutions:
• drill holes in the deck beams to permit air flow through them (reducing wind lift forces)
• give a more aerodynamic shape to the deck by adding fairings or deflector vanes.
Five days after the study reported, the bridge dramatically collapsed in moderate 65 km/h wind
conditions. The cause of failure was what we now know as aeroelastic flutter.
The Hyatt Regency hotel walkway collapse occurred in July 1981 in Kansas City, where
suspended walkways at second and fourth floors collapsed killing 114 people. The designer’s
original detail for the fourth floor beams showed the hanger rod passing through and a nut
holding the beam in place; the rod then continued to the second floor to support a beam there
also. The steelwork contractor objected to the detail as the fourth floor nut would need to be
wound up two floors of thread to erect the beams. The detail was modified so that two separate
rods were used, each terminating or starting at the fourth floor. No-one redesigned the detail,
and consequently everyone missed the fact that the force resisted by the nut at the bottom of
the top hanger rod had doubled.
The flaws of the revised detail were compounded by the fact that both designs placed the rod
directly through a welded joint connecting two channels to make a box section, and a
significant localised shear force was introduced at this point. Once a single nut failed the entire
walkway suffered progressive collapse.

Broughton Suspension Bridge Tacoma Narrows HR Hotel (Kansas) walkway

© Jonathan Haynes version 3.0 Page 3 of 239

32900 Bridge Engineering
MSc Structural Engineering

The Quebec Bridge was part of a Canadian transcontinental railway project, and was similar in
form to the Forth rail bridge. Preliminary calculations were never properly checked when the
design was finalised or when the span was lengthened, resulting in the actual weight of the
bridge being far in excess of its carrying capacity. In the summer of 1907 distortions of key
structural elements in the steel cantilever truss towers were noticeable. The southern cantilever
and part of the central section of the bridge collapsed into the St. Lawrence River killing 75
construction workers.
After an inquiry, construction started on a second bridge. The new design used the same form
but was much more massive. In September 1916, when the central span was being raised into
position, a problem with the hoisting devices caused it to collapse into the river, killing 13
construction workers. Re-construction began almost immediately after the accident as the
design was deemed to be sound. Construction was ultimately completed in August 1919, its
centre span of 549m remains the longest cantilevered bridge span in the world.
The first Tay Bridge collapsed while a train was passing over it, during a violent storm in December
1879, killing all 75 people aboard. The bridge used lattice-web girders supported by cast iron
piers, and wrought iron cross-bracing. Expert wind loading advice had been sought during the
design (opinion at this time varied from virtual denial, to wildly excessive) and as a result no
explicit allowance had been made for wind in the design. There were other flaws in detailed
design, in maintenance and in quality control of iron castings.
Bedrock had been deeper than anticipated and the bridge was redesigned with fewer piers
and correspondingly longer span girders. The pier foundations were constructed by sinking brick-
lined wrought-iron caissons onto the riverbed, and filling these with concrete. To reduce the
loading on foundations, masonry piers were abandoned in favour of lattice iron piers. There were
13 high level girder spans but to permit thermal expansion only 3 had a fixed connection.
Painters reported that the bridge shook when a train was on it, and worse when going fast. Many
of the diagonal bracing elements had to be re-tensioned before the bridge opened and several
cast iron columns had cracked and been wrapped with wrought iron hoops. Future British bridge
designs had to allow for wind loadings of up to 2.7 kN/m2.

Firth of Tay rail bridge Quebec 1st collapse Quebec 2nd collapse
The Westgate Bridge in Melbourne, Australia was a box girder beam
spanning between reinforced concrete piers, crossing a bay. In October
1970, the 112m span box girder between piers 10 and 11 collapsed and fell
50m to ground level, killing 35 construction workers, some of whom were
working on and inside the girder when it fell. Failure of the bridge was
attributed to the design and an unusual method of construction.
On the day of the collapse, there was a difference in level of 114mm
between two half-box girders at the west end of the span which was to be
joined longitudinally by welding. The higher half-box was weighted down
with ten 8 Tonne concrete blocks. The weight of these blocks caused the
deck plate to buckle. The longitudinal joining of the half-box girders was
partially complete when an instruction was received to remove the buckle.
As bolts were removed, the bridge snapped back and the span collapsed.

© Jonathan Haynes version 3.0 Page 4 of 239

32900 Bridge Engineering
MSc Structural Engineering

1.2 The Four Internal Forces and Four Structural Elements

Just as there are only four structural forces (axial, shear, bending and torsion), there are only four
structural elements.

V T

V T
N N M M

Axial Shear Moment Torsion

Tie – an axial tension-only element, which means that the external force always acts to stretch
and straighten (hence avoid buckling of any kind). Elements may therefore by very slender. The
material will usually reach yield stress (or its equivalent) and maximum efficiency is obtained.
Strut – an element subject to axial compression which will suffer lateral buckling, often sized to limit
slenderness and hence will rarely reach yield stress due to axial load alone (lateral displacement
creates bending stress which uses up capacity). Struts make less efficient use of material than ties.
Beam – an element which transfers in-plane force by bending and shear; this usually results in a
tension face (which is laterally stable) and a compression face which will buckle laterally, like a
strut, if unrestrained. Beams make less efficient use of material than struts.
Plate – an element which transfers out-of-plane forces by bending and shear, in a similar manner
to beams but on two axes rather than one. Plate thickness is small in comparison to the other
dimensions and large deflections occur, resulting in a general need for more non-linear
geometry analysis. Very thin plates are referred to as shells, and their action is dominated by in-
plane forces. In most structures plates are also subject to axial forces.

All structures can be viewed as an assembly of these structural elements. The assembly may be in
2D or 3D, for example:
• straight ties in bridges are usually cables, the 2D form of a tie is a catenary cable, the 3D forms
include the cable net, tensegrity and fabric structures which is rarely used in bridges
• the 2D form of a strut is an (elemental) arch but the most common bridge is one of the 3D
forms of a strut – the vault. There are several other 3D struts such as the groin vault, dome and
geodesic dome
• struts and ties are commonly formed into 2D trusses but may also be constructed in 3D to form
space frames which can be used for very large clear spans, 3D trusses are essentially the
analysis model for pile caps
• flat plates may be formed into 2D multi-storey shear walls and further, into 3D shear cores
• 2D flat plates also form the main constituent of deck slabs and foundations – retaining walls,
rafts and spread foundations. Plates may be curved into 3D structures but these are rarely
used in bridges

© Jonathan Haynes version 3.0 Page 5 of 239

32900 Bridge Engineering
MSc Structural Engineering

1.3 Bridge Structural Forms

With the exception of stressed ribbon bridges, the bridge deck is invariably level and usually
straight. The arrangement of the structure which supports the deck may be of any form, the most
common in general use are:
Slab deck (plate). The bridge running deck forms the entire
bridge. This is restricted to short spans (<15m) and is invariably
formed in concrete. The simplest option is a solid reinforced
concrete slab, where this is insufficiently stiff voids are
introduced to increase depth without a proportional increase
in weight. There is a natural progression of voided decks to the
box girder form.
Deck (plate) on beams. Bridge decks are very often little more than a reinforced concrete slab,
spanning continuously across a grillage of beams. Steel plate decks are also common in steel
half-through (drop-girder) rail bridges. Reinforced concrete decks are often 250-300mm thick,
with steel plate girders positioned 3.0 to 3.5m apart. The deck should not cantilever more than
2.0m and preferably 1.5m. Cross bracing should be restricted to two adjacent beams (as shown)
to prevent the bracing forming a lateral stiff point and inducing fatigue stresses in the bridge.

Ladder deck steel bridges are usually an arrangement of two deep plate girders with smaller
cross beams which carry the concrete slab. Cross beams are spaced at 3.5m.
Steel beams carrying a compression flange deck should always be composite for efficiency and
proportioned for depth at about L/20 when simply supported or L/40 when continuous.
Concrete beams should be cast into an end block at supports.
A mass arch (strictly a vault, strut). Although concrete vaults have been constructed, the majority
of vault bridges are of masonry construction (stone or brick). The vault has high mass and is
therefore difficult to vibrate; the large amounts of material used also mean that stresses are
relatively low, which contributes to the long life of this type of bridge.

L
R
30
or
L
0.105 R

Elemental arch (strut). A true arch can be used to span long distances if properly restrained, and
may be used to support the bridge deck in several configurations. The arch foundations must be
able to prevent support spread, making the tied arch (bowstring) a particularly efficient form as
the deck also forms the tie.

Bowstring arch Network arch

© Jonathan Haynes version 3.0 Page 6 of 239

32900 Bridge Engineering
MSc Structural Engineering

Deck (plate) and truss (strut/tie). The bridge

deck may be carried on, in or under the truss.
The truss may take any of the common forms,
such as Warren and Pratt. It is important to
arrange the bridge truss such that the
compression chord is restrained (often by a
horizontal truss) but the restraint reaction is
typically carried to the abutment by a
portalised end frame.
Box girder (beam). The box
girder is a form of beam and
plate which self stabilises its
compression elements. The
box section often tapers in
proportion to the bending
moment, especially in
concrete bridges.

L L
Varying depth box, > 60m span: ≅ 33 − 50 ≅ 12 − 20 dsp
d sp ds ds

L L haunched box girder elevation

Constant depth box, < 90m span: ≅ 18 (r.c.) ≅ 25 (prestressed)
d d
df
a df
a d d f ≥ 200mm ≈2 0.2 ≤ ≤ 0.25
d d

Composites of steel box and concrete deck are common. This deck is very torsionally stiff in
comparison to all other decks and is often needed in cable stayed bridges which only have one
vertical plane of cables.
Frame (beam/plate). Formed by rigidly connected elements, usually portal frames or box
culverts.

Cable stayed (tie/beam). A system which supports a long span bridge deck using triangulated
cables. The cables may be configured in several ways, and asymmetrical arrangements are
possible. This is the most recent development in long span bridges and provides an economical
alternative to small-medium span suspension systems.
fan harp

Suspension (tie/beam). A system which supports a long span bridge deck using catenary cables.
Vertical deck hanger cables developed from the original use of suspension links. Suspension
bridges usually carry a horizontal deck but the stress ribbon bridge is a much shallower version

© Jonathan Haynes version 3.0 Page 7 of 239

32900 Bridge Engineering
MSc Structural Engineering

which carries the deck on the catenary cable.

Suspension cables are inherently flexible structures, and when combined with their primary role of
spanning long distances, this makes them wind sensitive. As flexibility increases, so too does
deflection but natural frequency drops; making flexible bridges potentially subject to resonance
problems.
Moving
Bridges may also be
complicated by being
moving mechanisms. The
most common moving
bridges either lift vertically(a
bascule) or swing
horizontally. It is very
important to consider the
normal in-use design load
case and intermittent
Elevation on a moving load case, as
bascule bridge Plan on swing bridge
inertial loads from
accelerating and braking
bridge decks can be onerous. The mechanical and electrical systems of the lifting or swinging
mechanism should be designed by an appropriate engineer. Most moving bridges are balanced
to make moving easy, so the in-use condition will require the deck to be wedged in place to
prevent unanticipated movement.

1.3.1 Stability Systems

Like any structure, a bridge must be stable in all three dimensions and there must be a clear load
path for all forces to reach the foundations, where the supporting soil must not be overloaded
The same principle options are available in bridges and buildings – truss (triangulated pin-jointed
framework), rigid frame (moment joints) or shear walls.
In a similar manner to building floor plates, bridge deck plates distribute vehicle loads to
supporting structure and act as a horizontal diaphragm for lateral loads.
Use is often made of retained earth to resist lateral loads at foundation level.
Vertical loading will predominantly be due to self weight and vehicles. There will be significant
lateral loading from wind and vehicles (braking and centripetal forces) to consider. These must
be transmitted through bearings, which form points of stress concentration in the structure.
It is not always obvious that elements such as stay cables perform a dual purpose as the primary
load path for vertical loads and as restraint to the pylons which support them.
Bridges often incorporate stability elements not seen in other structures, such as u-frame action,
which is used to stabilise a truss compression chord when no structure is provided in the buckling
plane.

© Jonathan Haynes version 3.0 Page 8 of 239

32900 Bridge Engineering
MSc Structural Engineering

1.3.2 The Outline Construction Sequence

There are many ways to arrange the fundamental bridging forms, and still further ways to
construct them.
Steel bridges are often erected piecemeal like a building frame, reinforced concrete decks are
often shuttered and poured like a building frame. However, there are many construction
(execution as the Eurocode refers to it) techniques which are peculiar to bridges:
• Launching – jacking a completed deck from an abutment until it reaches the far abutment or
another deck being jacked towards it.
• Balanced cantilever – incrementally building a multi-span bridge from a central pier in two
directions (so that it balances).
• Cable spinning – creating a large diameter suspension cable by pulling small diameter wires
repeatedly along the bridge cable and looping them over the anchorage at each end.
Key knowledge for a bridge designer is how the bridge will be constructed, as it often leads to
identifying one of the most onerous loading conditions. Temporary conditions may involve lower
loading levels than when the bridge is in use, but often relate to more onerous laterally
unrestrained lengths, or force analysis assumptions. A designer’s risk assessment for a bridge
should specify the assumed construction sequence. If this is not done, then the contractor may
proceed with a different sequence which generates forces which the designer did not envisage;
this may cause excessive deflections in the completed structure (immediate or long term creep)
or even collapse of the incomplete structure.
Annotated time-lapse sketches are the simplest way to convey the intended sequence of
events, as shown below for the Viaduc de Clermont Ferrand.

1. Excavate for thrust blocks on both banks. Fill

south bank approach to level. Pile south bank
and erect temporary steel pier.

2. Erect south bank shutter, reinforce and cast

inclined strut and box head. Pile north bank
and erect temporary steel pier. Construct
south bank abutment.

3. Deliver off-site precast concrete box girder

sections, progressively hoist over and tension
back to box head, forming symmetrical
cantilevers. Erect north bank shutter, reinforce
and cast inclined strut and box head.

4. Construct north bank abutment. Repeat

box girder construction from north bank.

5. Cast last 500mm section insitu and post-

tension both halves together to form
continuous bridge. Release hydraulic jack on
temporary steel piers and remove.

Clearly, the temporary works may constitute a significant proportion of the initial capital costs of
a bridge, and so must be considered when selecting a scheme.
© Jonathan Haynes version 3.0 Page 9 of 239
32900 Bridge Engineering
MSc Structural Engineering

1.3.3 Crane Data

All cranes are supplied with individual lift-reach data sheets. Most bridges are constructed using
mobile cranes which are available in sizes up to 300 Tonne lift capacity. However, capacity is
drastically reduced as the hook moves away from the crane body (called reach). Most mobile
cranes also have a telescopic jib, which is primarily to increase the possible height of lift. There is
likely to be a requirement to provide a counterweight (or kentledge) on the crane body, to
achieve a specified lift capacity.

Mobile cranes need space to manoeuvre, level ground to operate

from and stiff soil to stand on.
Stability outriggers should always be deployed if the jib will swing
across the longitudinal axis of the crane during the lift operation.
Every lift operation requires a lifting plan, which instructs the crane
driver what to do, in what order. A qualified banksman must direct
the driver, as the hook and lifted object may not always be in the
drivers view.

Always check whether the data sheet is given in Imperial Tons or metric Tonnes, as they are
slightly different values.

the jib length (or

A typical crane capacity chart main boom) is the head
shown here height is
shown here

the permissible lift

mass is shown here

the reach (or

radius) is Kentledge
shown here

Using this crane to lift 5 Tonnes at a reach of 18.0m:

Counterweight of 6 Tonnes + jib length of 29.8m + head height of 25.0m = lift capacity 5.4 Tonnes

If possible, avoid using more than one crane to lift an object as the interaction of forces can be
difficult to predict, and the consequences are often catastrophic.

Data sheets for Liebherr 40T and 150T mobile cranes may be found on Blackboard in the Deck Loading &
Analysis section.

© Jonathan Haynes version 3.0 Page 10 of 239

32900 Bridge Engineering
MSc Structural Engineering

1.4 Choosing Appropriate Materials

The vast majority of new bridges are constructed in either steel or concrete. The bridge deck is
almost invariably reinforced concrete, which serves to do little more than spread loading, tie the
structure and form a horizontal diaphragm. For composite bridges this is essential to efficient
design and therefore sets the deck at the level of any compression elements (particularly steel).
Reinforced concrete is also the natural choice for bridge foundations since it can be formed into
any shape (pile caps, retaining walls, footings and gravity blocks) and is highly durable in
aggressive environments if properly specified and constructed. Since foundations are usually
buried and forgotten, it is of great importance to durability that the concrete has the correct
cover and compaction.
When choosing a surface treatment it is worth considering that concrete is close to white when
first cast but will be a dark grey/black within ten years.
Concrete bridges (and prestressed in particular) can be difficult to modify after construction. This
is not necessarily true of steel bridges but site welding of steelwork is to be avoided where
possible due to the precautions necessary to ensure work quality.
Bolted connections in steel bridges should utilise HSFG or other similar fasteners, to avoid loss of
integrity due to vibration.
It is often appropriate to use brick or stone for arch bridges in
controlled rural settings such as National Parks. In such cases
care should be taken to match local material, and use local
sources where possible. Durability is a key feature of masonry
arch bridges, it is the only form which has working examples over
100 years old, many being several hundred years old. This
depends upon the appropriate choice of masonry quality, the
difference between common and engineering designation
brickwork is shown in this bridge parapet and approach panel.
The variety of masonry units available means that most
aesthetic demands can be accommodated. Masonry also has
excellent thermal loading characteristics but it must be borne in mind that masonry arches
accommodate thermal movement by cracking (the bond between unit and mortar) and
expansion joints are avoided.
Existing masonry arch bridges are usually founded on a stepped masonry footing but some will
have included piled foundations, often timber, in a marine environment. Timber is usually
perfectly preserved in air free environments (e.g. wholly submerged in water or mud).
Moving bridges will rarely use concrete for anything but
counterbalance, or towers of very large lifting structures (such as
the Rouen Lift Bridge). However, moving bridge decks are often
very light (to make movement possible) and utilise aluminium
trough decking with a very thin (but expensive) wearing course.
Wrenbury Lift Bridge (left) which crosses the Llangollen canal, has
a timber lifting structure with iron tie rods on the cantilever
beams, a lead weighted counterbalance box and iron lifting
chains. The deck is formed from two steel RHS beams (to mimic
the original timber beams) with lightweight aluminium cross
beams, topped with 8mm of bonded wearing course. The steel and aluminium are separated by
Denzo tape to prevent sacrificial corrosion.
Timber bridges are rare in the UK for anything but short span pedestrian bridges. Forms are
restricted to glulam beams or trusses. In rural environments, colour free preservatives should be
used to enhance the visual beauty of timber trusses.
Advanced materials such as Carbon Fibre Reinforced Plastics are seldom used for more than
adding tensile reinforcement to existing structures. However, at least one complete cellular CFRP

© Jonathan Haynes version 3.0 Page 11 of 239

32900 Bridge Engineering
MSc Structural Engineering

pedestrian bridge has been constructed, largely due to its ultra-light weight and ability to be
craned over a railway line during a short possession.
When designing a new short to medium span bridge, the choice of material is largely between
steel and concrete, or a combination of both. It is worthwhile considering the obvious
advantages and disadvantages of each, which are summarised as follows:
Steel:
Tends to be used for unusual designs (easier to design for combinations of all force) and moving
bridges (lighter), and is perhaps more versatile in terms of reducing construction depths (more
slender appearance)
Lower weight of deck is possible
Smaller or fewer foundations due to lower weight
Offsite fabrication is undertaken in a high quality controlled environment
Site connections are bolted, which improves end of life demolition ease or reuse
Erected piecemeal, must be small enough to be lifted by a (mobile) crane, so jib and reach may
be important
Usually no temporary propping of structure required during execution
Shorter execution time but longer lead time
Corrosion protection has life span < 120years, weathering steel an option if suitable
Relatively easy to widen structure by addition to existing structure
Relatively easy to repair patch damaged material
Concrete:
Usually forms robust (relatively massive) structure
Higher mass deck avoids most excitation problems but larger foundations necessary
Any shape desired can be formed if budget permits
Site quality control important if reinforcement corrosion / spalling is to be avoided
Can be designed to accommodate corrosive environments
Highly sustainable if local materials are sourced
Can be used for mass arch forms
Combinations of precast and insitu elements avoid need for formwork
Better at dealing with large thermal gradients
Should be able to reach 120 year life without additional protection
Often lower initial cost
Composite:
Makes best material use of steel and concrete
Forms robust and stiff structure by shear connection
Integral composites avoid the need for bearings or movement joints – minimising the most
common durability problem in short/medium span bridges

Many of these advantages and disadvantages will be pertinent to a particular site, others will
not. Therefore it is important to leave material choice until after the structural form and execution
method have been identified, and only then to select the most appropriate material, for each
part.

© Jonathan Haynes version 3.0 Page 12 of 239

32900 Bridge Engineering
MSc Structural Engineering

1.5 Scale of Use

It Is unhelpful to be prescriptive regarding the structural form to use for particular spans but the
chart below is intended to assist in eliminating uneconomic options at the scheme design stage.

Suspension
50-2000m
Akashi Kaikyo 1991m (truss), Humber 1410m (aerodeck)

Cable stayed
30-1100m
Riskky 1104m, Pont de Normandy 856m, Second Severn 456m

Masonry arch single span Cantilever truss 100-550m

3-20m solid, but up to 85m Cantilever box girder 50-150m
Stockport 650m (27 bays), Solkan 85m (open spandrel) Quebec 549m, Forth rail 521m, Oakland Bay 427m
Maidenhead Rail 39m (brick), Grosvenor 61m (stone)

Truss with deck under or over

20-200m
Tamar 139m (lenticular), Runcorn 482m (truss arch), Exhibiton Footbridge 44m (space truss)

Slab deck on haunched box girder

40-120m
Byker Metro 60m (concrete)

Vierendeel with deck under Slab deck on box girder

10-25m 30-100m
Hylton viaduct 46m (steel)

Slab deck on haunched beam

15-60m
Thelwall viaduct 1350m (max span 102m)

Slab deck on beam

Through, over or tied arch Under arch
5-30m
15-550m 40-900m
GMex 45m, Pentele 308m, Sydney Harbour 503m – Oporto 280m, New River Gorge 924m, Taf Fechan 70m

10 20 30 40 50 60 70 80 90 100 110 120

Span, L (m)

© Jonathan Haynes version 3.0 Page 13 of 239

32900 Bridge Engineering
MSc Structural Engineering

1.6 Aesthetic Appeal

Discussion of aesthetics is traditionally reserved for those who study the arts; however, in the case
of bridges, there is a fusion of art and science which requires Engineers to express opinion.
Nothing restricts artistic design more than rules, according to Bennett (1997), and the most
important rule is usually low construction cost. Attractive bridges often express creativity and
clarity of structural form but are not necessarily the most expensive option. Having inferred that
the bridge designer should be brave in their creation of schemes, there are some established
observations which can be used as guidance.
Appreciation of beauty is not only difficult to define but also subject to individual perception of
colour, shape, texture and setting. Some arrangements of shape and proportion are generally
more acceptable than others. This was first explored and recorded in ancient Greek culture.
The Greek classical style lasted around a millennia (600BC to 500AD) with the Roman classical
period being around 200BC to 500AD, whereupon the Roman empire fell and classical
architecture was essentially lost until the Renaissance (rebirth) when first Italian, then French
architecture adopted the style around 1420. Western architecture therefore follows rules of
proportion which can still be used to establish aesthetically pleasing proportions. Since the
Roman empire extended into what is now the middle east, the architectural books of Vitruvius
have also been used to proportion the great Islamic structures.
One of the simplest observations is that a well proportioned rectangle has sides of ratio 1:0.618 or
about 13:21 in whole numbers.

Proportions of the golden rectangle. Spiral of the golden rectangle.

Any classical building will use the golden proportions, as do Roman arches but many modern
bridges also follow the same rules.

acceptable acceptable

better better

Winning a bridge design competition may be attributed to little more than being able to convey
the designer’s ideas better than the other competitors. More than one vision of the bridge is
often needed, as the structural concept will need different explanation to the architectural
concept. An architectural concept sketch must convey environmental context, whereas the
structural concept must depict form and use.

© Jonathan Haynes version 3.0 Page 14 of 239

32900 Bridge Engineering
MSc Structural Engineering

An architectural concept sketch. A structural concept sketch.

The Highways Agency (1996) and Leonhardt (1982) both offer judgment (verging on rules) for the
aesthetic design of bridges. This is rare advice, and is presented in a way which engineers can
understand – there is limitless project specific architectural opinion on the topic but is of little
practical general use.
Setting
A bridge will be most dramatic if it springs from abutments set in near-vertical faces of valleys or
buildings, or when the abutment is masked by trees.
Where a new road cuts through a ridge, the overbridge deck should be at the ridge level.
Equally, the notch cut through the ridge should be as steep as possible to improve the drama of
the bridge setting.
Minor roads should bridge over major roads to reduce the potential costs and avoid a broad,
unsightly, darkened overbridge.
Where a new bridge is to be constructed adjacent to an existing, it should be parallel and in an
identical alignment, maintaining as large a gap as possible. The form should be similar but not
necessarily identical e.g. a concrete flat arch adjacent to a masonry arch. If similarity is not
possible then all aspects should be contrasted as far as possible.
For complex alignments which incorporate junctions, the best results are obtained using insitu
concrete (reinforced or post-tensioned) as it can accommodate all manner of shapes.
Where possible, drainage gulley’s should be avoided on bridges. All water should be channelled
off the bridge. This is not possible in long viaducts, where drainage pipes are incorporated in or
under the deck, particular care must be taken in waterproofing holes through decks.
Trusses
In truss bridges it is important to have order: never change truss type (e.g. between towers and
beams). If possible the chords should be parallel to each other and the road alignment. To
improve the view out of the bridge, the diagonal elements should be at about 60 degrees to the
vertical. This suggests that a Warren arrangement is visually best.
Trusses should be as visually light and transparent as possible. Concrete trusses were popular in
central Europe during the 1950’s but invariably appear too stocky to be elegant.
Of prime importance in trusses, is the detailing of connections. Welded hollow sections provide
the best visual appearance; conversely, oversized bolted gusset plate connections are worst
(though cheapest).
Where there is sufficient depth available, the highway should be arranged on top of the truss,
where the deck conveniently provides restraint to the truss compression chord. However, if the
bottom chord must carry the highway, the compression chord will need restraint from a
secondary system such as a U-frame, transverse bending stiffness of the chord or a further Warren
truss in the plane of the top chord.
Although not material efficient, it is best to use consistent cross-sections throughout the truss – do
not mix hollow sections, angles, beams and channels in the same plane.
Openings in trusses should either be blanked or open, never partially filled.
Care must be taken with scale when using a hollow section vierendeel for a pedestrian truss. If
the panel is too large the bridge appears to be a handrail for giants.

© Jonathan Haynes version 3.0 Page 15 of 239

32900 Bridge Engineering
MSc Structural Engineering

Since trusses are visually light, they can be painted in a colour which contrasts starkly with their
surroundings without jarring, often white in urban environments.
Continuous steel trussed arches should use the hogging rather than sagging form (haunch the
supports and run the top chord level through).
Truss joints should be noded to negate secondary bending forces.

a. Salford Quays Dock 9 – a cantilever truss swing bridge

b. GMex Metrolink, Gt Bridgewater St – a bow string tied arch
c. Trinity footbridge – a cylindrical 3D truss

d. Jackson’s Boat footbridge – a warren truss with hoop restraint ties, painted to blend in
e. Deansgate Stn Exhibition Footbridge – a pentagonal section 3D truss

f. Hungerford Bridge, London – a well detailed and painted truss

g. Princess Bridge – a highway bridge now restricted to pedestrian use
h. Princess Bridge – riveted connection detail

i. New Quay St bridge – a curved chord truss

j. Pomona Stand (Manchester ship canal) – cross braced truss footbridge with hoop restraint ties

Arches
Masonry arches should not be too thin at the crown (to avoid the illusion of instability) and should
therefore incorporate a parapet wall rather than handrailing.

© Jonathan Haynes version 3.0 Page 16 of 239

32900 Bridge Engineering
MSc Structural Engineering

Multi-span masonry arches should include relatively thick piers. The only exception is where a
viaduct has tall piers.
Multi-span masonry arches should not contain different arch shapes or gross changes in
adjacent spans. Where this is not possible, e.g. a larger navigable span, the pier between the
differing spans should be more substantial or incorporate an enlarged pier.
Embellishments in new masonry arches should be restricted to coping and string courses since
satisfactory craftsmen to undertake elaborate surface dressing are difficult to find.
Semi-circular masonry arches can appear massive (e.g. in deep cuttings), so it may be prudent
to express the arch voussoirs or ring and use a handrail rather than parapet wall, which will
reduce the area of spandrel wall. As the arch shape becomes flatter (segmental or elliptical) the
horizontal thrust at the abutment increases, resulting in a need for greater mass at the abutment
(a quarter to a fifth of the span). Central piers need not be massive where adjacent spans are
equal since the thrusts will balance.





Though masonry arches are commonly built with a span: rise ratio of 4, steel and concrete
arches should be proportioned closer to 7. Continuous concrete arches look best when
ambitiously flat and when clearance over the valley/river level is as small as possible. Curved
haunches giving a parabolic shape between intermediate supports are best when combined
with a curved vertical alignment centred on the deck summit: in this case a central span: depth
ratio of 50 is achievable.
Concrete and steel under-arches should be as visually simple as possible, usually consisting only
of an arch, deck and strutting (in order of importance). Concrete arches usually present the
strutting and arch in the same plane, with a cantilevering deck edge to create shadow. Steel rib
arches should be similarly arranged. A chord drawn between the arch springing points should be
parallel to the deck. If the arch and deck do not touch at the crown, there should not be a
central strut. If possible concrete arches should avoid a gap between the arch and deck;
preferably arranged so that the deck soffit and arch centreline are tangential.
Beams
Generally beam and slab bridge decks will appear more slender if the deck is expressed as a
small cantilever past the support beam. This avoids the side of the bridge appearing to be a flat
slab of concrete. Handrailing should be set back from the edge of a deck. Although most deck
edges are vertical, a sloping deck edge will reflect light differently, visually identifying the line of
the deck edge.
Though span: depth ratios give guidance on the rough proportions of elements (about 20-30
usually), spans shorter than about 6.0m will appear too shallow and a ratio of 6 may be
appropriate; whereas for long span continuous beams, deep beams look very heavy and a ratio
approaching 45 may be appropriate (though this may require support haunching). Support
haunches should not exceed 20% of the span and be inclined at about 1:8.
Curvature of any kind is not easily accommodated by simple beams, and when combined with
a horizontally curved deck, the resulting shadows on the beam web may suggest the beam is
sagging. Super-elevation of decks with parallel flange beams should be avoided as this exposes
an unsatisfactorily complex shadowed soffit view to road users.

© Jonathan Haynes version 3.0 Page 17 of 239

32900 Bridge Engineering
MSc Structural Engineering

If a beam is curved in elevation, then the beam bottom flange should never be straight. Where
the beam meets an abutment, handrailing should continue onto the embankment, rather than a
solid parapet wall extending onto the abutment.
L L L
35 40 18

L L
L L
15 20

L L
Common proportions for over-bridges.

Beam viaducts
Generally, it is best to use an odd number of spans in a viaduct. Closely spaced, slender piers will
emphasise height, widely spaced stocky piers emphasise width.
When crossing deep V-shaped valleys it is best to use three or five larger spans, avoiding a pier at
the lowest point of the valley if possible. This should present vertically tall rectangles formed by
the beams and piers.
When crossing wide valleys it is best to present horizontally wide rectangles formed by the beams
and piers, of proportions about 1.5 : 1. The most visually pleasing arrangement is to have varying
spans up the valley side, such that the diagonal of the rectangles maintains the same inclination.
Solid support pier breadth (across the viaduct) should be about span / 8 but if there are several
columns forming the pier they should amount to no more than span / 3 (positioned centrally).
Where possible, deck continuity should be adopted as this reduces depths and prevents water
and salts ingress which lead to staining.

a. GMex viaduct – a hideous concrete slab deck suffering weathering and staining
b. M60 – integral slab deck and secant pile wall under bridge
c. Water St – arch rib railway bridge

d. Hampson St – well proportioned modern half-through railway plate girder

e. Hampson St – end bearing and pin support
f. Deansgate - a pseudo-gothic, wrought iron arch rib railway bridge

© Jonathan Haynes version 3.0 Page 18 of 239

32900 Bridge Engineering
MSc Structural Engineering

g. Liverpool St – a very old and featureless built-up plate girder skew overbridge
h. M60 – an unsympathetic pier strengthening, note carbon strip fibre deck strengthening

Cable stays
Cable stayed bridge decks can become progressively slender as the distance between cable
supports decreases, as will the cable diameter.
Fan cable arrangements are generally most structurally efficient and cheaper to build (fewer
cable anchors) but are visually confusing when viewed obliquely. Whereas, harp arrangements
are least efficient (grossly different extensions under load) but are visually most attractive (cables
in two vertical planes remain parallel when viewed obliquely).
The slender continuity of a cable stayed deck can be ruined by externally located stay anchors.
Visual appearance is much improved by inboard anchor points, either behind a beam fascia
board or recessed into the deck slab.
Where there are two vertical planes of cables, the tower can be designed without a cross-head
or beams because the cables provide in-plane stability (though not at the temporary
construction stage) and the lateral wind forces from the cables are small. For large towers, or
where a single plane of cables is used, an A-frame is the preferable tower form.
Back-spans should be less than a third of the main span. Where back-spans are stayed to several
supports, the resulting bridge will be stiffer as there is no necessity to react against a flexible deck.
Sagging vertical curve alignment should be avoided since the bridge will appear to be
collapsing, this is exacerbated if there is a central tower.
Suspension
Suspension cables should be a parabolic shape, and hence are unusual for spans less than 300m
as there is insufficient cable sag to visually locate a well proportioned bridge deck.
At centre span the cable should either pass below the deck, or not pass below parapet level.
Back-spans should be less than half the main span, and are usually less than a third.
The open space under the bridge should be long and shallow, so bridges with high water
clearances should be of long span.
Where possible, the American tradition of suspending a truss should be avoided, as this loses the
light slenderness of the suspension system (and hugely increases aerodynamic loading). The
pylons need not be slender to enhance the visual effect of a slender bridge and for moderate
spans it is possible to use masonry piers.
The tops of suspension bridge pylons generally require connecting together for out-of-plane
lateral stability purposes, this is best done a short distance from the top.
Asymmetrical suspension bridges are rare but should generally benefit from a pylon which is
closer to the forms used for cable stayed bridges.
To improve views from the bridge, it is common to drop the level of the pedestrian walkway
below the road level. This should also afford some wind protection to pedestrians.
Footbridges
The combination of light loading and maximum potential for re-alignment, mean that
footbridges offer the greatest potential for artistic flair in design.
© Jonathan Haynes version 3.0 Page 19 of 239
32900 Bridge Engineering
MSc Structural Engineering

Pedestrian bridges often incorporate roofing

Ramps and steps are perhaps the most unsightly element of footbridges. It is best to cut ramps
and steps into the natural terrain when possible (or build a bund embankment), or else to make
support structures as visually light as possible. In this case the ramp structure should be of similar
proportions to the footbridge.
Access ramps should be no steeper than 1:12 when connected to a building (and are thus
covered by the Building Regulations). However, steeper ramps are permissible in open
countryside, and in this case steps would be preferable.
It may be possible to use vertical alignment curves to reduce access ramp heights, by extending
the bridge below its abutments.
Spiral ramps were popular in urban areas. If used they should be limited to one revolution as
stacked spirals appear very massive even if sparsely supported.
Sloping footbridge decks may be considered but should comply with Building Regulation
recommendations for disabled access.

a. Tensegrity systems should be approached with extreme caution

b. Humber Bridge – the longest single span aerofoil deck, massive anchor blocks
c. Very short suspension bridges must be efficiently designed to achieve a light visual appearance

d. Pont du Nord – a long cable stayed span with many back stay tie downs
e. Bristol canal – a cable stay with dysfunctional back stays
f. Altrincham Moss Lane – pedestrian overbridge with glass parapets

g. Merchants Bridge – a balanced torsion arch

h. Bristol canal – a counterbalanced swing bridge

© Jonathan Haynes version 3.0 Page 20 of 239

32900 Bridge Engineering
MSc Structural Engineering

1.7 Capital and Carbon Costing using CESMM

A bridge is an expensive but necessary item of infrastructure. It is important to know how much a
bridge will cost to construct because they are almost invariably funded from the public purse.
Cost is also an important comparator – it helps the client choose between scheme proposals.
Estimates of construction cost can be produced using the Civil Engineering Standard Method of
Measurement. CESMM is a standardised way of producing a bill of quantities which is normally
priced by contractors who are tendering for the works.
For the purposes of this module, a comparative estimate of construction cost and carbon cost
should be made using the Bridge capital and carbon pricing.xls Excel workbook which is
available on Blackboard. The workbook contains twelve spreadsheets which are populated with
standard entries from CESMM3 (section reference, description, unit prices and unit carbon cost),
requiring only the quantity from the bridge design to be entered. Automated calculations will
then generate the capital and carbon costs.

CESMM3 section reference

the quantity cells are the only

cells in which values may be
entered

outcome capital and carbon

costs for each item

tabs to each spreadsheet

The summary spreadsheet provides a collection of the prices for each class of quantity. Only the
bill classes which pertain to bridge construction have been included in this workbook (e.g.
cladding and general building works have been excluded).

values carried forward from the

billing spreadsheets

you will need to calculate the

inflationary increase in prices
since the spreadsheet values
were assessed in 2010

Risk contingency from your Risk

Management assessment (see
next section)

Final bridge construction cost

estimate

The final contract value produced should cover design fees, construction, insurance and risk
contingency. There is also a quantified carbon cost. Also include all necessary TEMPORARY
WORKS.
© Jonathan Haynes version 3.0 Page 21 of 239
32900 Bridge Engineering
MSc Structural Engineering

1.8 The Principles of Risk Management

Though the term risk assessment is generically applied across many industries, in construction
there are three identifiable levels of risk assessment which require slightly different responses:
Work activity risk assessment is concerned with the physical dangers of doing a job on site.
Designer’s risk assessment is concerned with the process of reviewing and amending a design
to avoid, substitute, segregate or protect operatives from danger. It happens before the
construction phase begins.
Risk management is a term which describes the process of assessing, limiting and managing
dangers to commercial success. This is usually about controlling cost and time overruns. It is a
topic which is developing quickly and only a brief over-view of its use in project comparison will
be undertaken here.
The English legal system is complex and hierarchical. There are four branches of English law but
health and safety issues are generally confined to statute and common law.
Statute law is the written laws of England which are set out in Acts of Parliament. Health & Safety
law is made under the Health & Safety at Work Act 1974. This gives the government power to
introduce Regulations without approval of Parliament, as and when necessary.
Employers are subject to vicarious liability under English Common Law, which has evolved
through centuries of rulings by judges in court. Everyone, including employers owe a general
duty of care to others.
The Health & Safety at Work Act 1974 (HASAWA) is enforced by the Health and Safety Executive
(HSE) and specifies many duties of interested parties, these are summarised below.
Employers must provide and maintain:
• Safe plant and systems of work.
• Safe handling, storage and transport of work articles and substances.
• A safe working environment with adequate welfare facilities.
Employees must:
• Take reasonable care of their health and safety, and that of others who may be affected by
their acts or omissions.
• Comply with employers to ensure health and safety.
• Not intentionally or recklessly interfere with or misuse equipment provided for health and
safety reasons.
In short, we all have the right to expect not to be injured at work.
A large number of Regulations are in force under HASAWA, many as a result of European Law.
Several are directly applicable to construction.
The Provision and Use of Work Equipment Regulations 1992 (UWED)
This sets out requirements for the provision of guards on moving devices, and requirements for
traceable records of approved testing.
The Manual Handling Operations Regulations 1992
This sets out a hierarchy of actions which eliminate manual handling if possible, or require the use
of mechanical assistance. It strictly applies to repetitive handling operations such as factory
work but has been applied to construction activities.
Personal Protective Equipment at Work Regulations 1992 (PPEWR)
PPE is equipment worn or held, whilst at work, to reduce the risk of an accident. PPE must be
suitable (appropriate to the risk, ergonomically designed and effective). The employer must
take steps to ensure it is used and the employee must wear it.
The Control of Substances Hazardous to Health Regulations 1994 (COSHH)
This sets out requirements to anticipate and deal with risks due to dangerous substances in the
work place.
The Reporting of Injuries, Diseases and Dangerous Occurrences Regulations 1985 (RIDDOR)
© Jonathan Haynes version 3.0 Page 22 of 239
32900 Bridge Engineering
MSc Structural Engineering

This sets out the requirement placed upon employers to notify the HSE of any serious accidents at
work (such as poisoning, loss of limb, death, radioactivity leak, air accident/near miss, death or
injury on an active railway).
The Construction (General Provisions) Regulations 1961
Most of the general site safety requirements are contained in this regulation, such as:
• Requirement to appoint a single site safety officer.
• Safety of excavations (>2m deep).
• Construction of caissons and cofferdams.
• Use of explosives.
• Transport safety.
• Demolition.
• Inspection record keeping.
The Construction (Lifting Operations) Regulations 1961
This regulation sets out the requirements for inspection of lifting equipment, planning of lifting
operations to avoid accidents, the use of hoists and mobile cranes.
The Construction (Working Places) Regulations 1966
This regulation sets out the detailed requirements for safety of working places in the construction
industry, with particular respect to scaffolding. This includes erection and inspection of static and
mobile scaffolds. It also limits the use of ladders for access platforms.
The Construction (Health and Welfare) Regulations 1966
This regulation sets out the detailed requirements for employees welfare facilities on construction
sites.
The Construction (Head Protection) Regulations 1989
This regulation sets out the requirements for the provision, maintenance and use of adequate
head protection (hard hats) on construction sites.
The Construction (Health, Safety and Welfare) Regulations 1996
These are general rules covering site safety, such as provision of safe places to work, falls, fragile
(roof) materials, falling objects, (temporary) structural stability, demolition, dismantling, explosives,
excavations, cofferdams, drowning, emergency procedures, fire, welfare facilities, lighting,
training and inspection.
The Construction (Design and Management) Regulations 2007 (CDM)
This regulation has had the most impact on construction in recent years. It imposes a
requirement to predict possible safety problems (Risk Assessment) and put safe systems of work in
place to address the risks (Method Statement). It also introduced new roles to the construction
team.
New definitions
Construction work now means building, civil engineering or construction work; but also includes
alteration, fit-out, repair or maintenance, site clearance, site investigation, demolition or
dismantling.
Contractor means anyone who carries out construction work.
Designer means anyone who undertakes design, such as drawing, detailing, specification or
production of bills of quantities.
Structure means any building or form in steelwork or reinforced concrete, railway line, dock,
harbour, viaduct, tunnel…cable, earthwork, mast…scaffold, formwork…or access that involves a
risk of falling more than 2m. Essentially anything involved in construction work!
This regulation applies to any construction work which last more than 30 days or, will involve more
than 500 person days or, involves demolition or, involves five or more people on site at any one
time. Essentially any construction work!

© Jonathan Haynes version 3.0 Page 23 of 239

32900 Bridge Engineering
MSc Structural Engineering

CDM places a number of new duties on parties to the construction contract. Five parties are
identified: the client, the designer(s), the Safety Coordinator, the Principal Contractor and
contractor(s).
Clients must:
• Appoint a Safety Coordinator as soon as the project begins. Appoint a Principal Contractor
before work on site begins.
• Ensure the Safety Coordinator and Principal Contractor are competent to undertake health
and safety tasks.
• Supply all necessary information about their structures.
• Keep the Health & Safety File for future reference.
The Safety Coordinator must:
• Notify the HSE that the project is beginning.
• Prepare a Health and Safety Plan, composed of information from the designers, which informs
the Principal Contractor of the anticipated construction risks.
• Hand the H&S Plan to the Principal Contractor at the commencement of construction work.
• Provide the Client with a completed H&S File for the structure, including as-built drawings, risk
assessments and method statements for construction, maintenance and demolition.
All designers must:
• Undertake a risk assessment on their design (consider construction, maintenance and
demolition).
• Take steps to design out risk if possible (consider construction, maintenance and demolition).
• Cooperate with the Safety Coordinator.
The Principal Contractor must:
• Ensure that construction work is undertaken in a safe manner (in accordance with method
statements).
• Take the H&S Plan and keep it up to date during the construction period. Handing all
information back to the Safety Coordinator at the end of construction.
Contractors must:
• Prepare method statements for their construction work, based upon the designers anticipated
risks.

© Jonathan Haynes version 3.0 Page 24 of 239

32900 Bridge Engineering
MSc Structural Engineering

1.8.1 Assessing Risk

The Health & safety at Work Act 1974 places a duty on employers to maintain a safe working
environment, and employees to comply with employers safety instructions. Part of the process is
to assess the risk inherent in work environments.
A hazard is something which has potential to do harm – for example, a tower crane is a hazard
because it could drop a concrete skip. A computer is a hazard because it is linked to mains
electricity and could electrocute you.
Risk is the likelihood that a hazard will do harm – for example, the risk of a tower crane dropping
a skip of concrete is low if the crane is well maintained, properly operated and only picks up a
concrete skip once a day. The risk of electrocution from a computer is low if it has a current PAT
certificate. Risk may therefore be assessed by examining the probability of an occurrence and
the severity of an occurrence
A simple matrix is usually employed to allow a categorisation of risk (high, medium or low), rather
than a numerical assessment.
As a rough guide:

Risk of Event Probability of occurrence Note :

High Medium Low It can be difficult to reduce

the severity of an
High H H M occurrence, even when
occurrence
Severity of

controls are applied to a

Medium M M L process.

Low M L L

Medium severity may involve a hospitalising injury

High severity may involve death
Low probability < 25% chance of happening
High probability > 75% chance of happening
It is common to tabulate risk assessment, as shown overleaf. The form is filled in from left to right:
• Identify the hazard.
• Identify who is at risk – staff and/or the public.
• Identify the probability and severity of the event happening.
• Calculate the Raw Risk from the Risk Matrix.
• If the risk is Low no further action is needed.
• If the risk is Medium of High identify a control measure which will reduce the risk of an event.
This will reduce either the severity or probability, or both. Then calculate the Residual Risk, this
should be Medium or Low.
Hierarchy of Protection Options
If a risk is identified which is unacceptable then it is necessary to put a control measure in place
before the activity begins. The protection measure should follow a hierarchy of:
• Avoid the risk by changing the design, or if this is not possible;
• Substitute by changing the hazardous process, or if this is not possible;
• Segregate by removing personnel from the hazardous workplace, or if this is not possible;
• Use of safe systems of work by imposing working methods which minimise risk, or if this is not
possible;
• Use of protective equipment such as respirators, hard hats and safety boots.
It is advisable to use these key words in the Comments and control measures section of the risk
assessment.
Hazards which are not assessed as LOW RESIDUAL RISK must be addressed by a Work Method
Statement which sets out a safe method of work.

© Jonathan Haynes version 3.0 Page 25 of 239

32900 Bridge Engineering
MSc Structural Engineering

Common Construction Hazards

The list of hazards below is not exhaustive but includes many of the most common which occur in
construction. There will be specific hazards which apply to most infrastructure development
schemes.
• Falling from height (during erection, during lifting operations, whilst accessing works)
• Drowning (proximity to water, inundation of excavation)
• Wet concrete (highly alkaline)
• Restricted access (confined spaces, interaction of traffic and pedestrians)
• Lifting (failure of lifting equipment, dropped payloads, slewing over public roads)
• Excavation (deep, in water, around buried services)
• Stability (partly erected buildings, slopes etc.)
• Substances hazardous to health (transportation, storage, use)
• Demolition (piece-wise removal, explosive, façade retention)
• Manual handling (activities which cannot be mechanised, repetitive handling)
• Electrocution (use of electrical equipment, inadvertent severing of power lines)
• Collision (between vehicles on constricted sites, or between pedestrians and vehicles in
traffic management systems)
Things to Remember
The Designer’s Risk Assessment is solely concerned with Health, Safety & Welfare of people at
work. It is a live document which is continually updated as design progresses, for example:
• Initial design includes a pier in a tidal river,
• high risk of drowning during installation avoided by changing design to clear span the river,
• high risk of crane toppling reduced by substituting concrete beams with launched steel truss.
The designer need only consider hazards which a competent contractor would not spot, so
need not list obvious, everyday problems such as regular assessment of scaffolding, or the
requirement to wear hard hats.
If something goes wrong and the designer has failed to comply with their obligations under the
Health & Safety at Work Act (undertake risk assessment for the design), the HSE will impose fines
or a jail sentence. A risk assessment which shows why the design developed into the final version
is a get out of jail free card.

© Jonathan Haynes version 3.0 Page 26 of 239

32900 Bridge Engineering
MSc Structural Engineering

1.8.2 Work Activity Risk Assessment

Project Title: A6 Underpass Reference Number : 89755/01
Raw Risk Residual Risk

Persons at Risk

Probability of

Probability of
Occurrence

Occurrence

Occurrence

Occurrence
Severity of

Severity of
Hazard Comments or Control Measures Specified by the Assessor

Risk

Risk
Crushing (soil slip) Staff + M H H Risk cannot be avoided or substituted. Soil must be L H M
public temporarily stored at site perimeter to avoid removal (at
limited times).
Apply for closure of near side footpath during excavation
For a work activity, sequence to segregate public from possible harm.
identify the thing which
can cause harm.

Who is at risk, could it Without considering any

be a member of the protection, quantify the
If the raw risk is medium or
public ? probability and severity, Now reassess the risk,
high you must think of a
and hence the risk remember that the
means of managing the risk
so that it is ALARP (as low severity is difficult to
as reasonably practical). change.
You must tell the contractor
what you need to be done. Is the final risk low or
medium ?
Can you do more to
reduce the risk ?

© Jonathan Haynes version 3.0 Page 27 of 239

32900 Bridge Engineering
MSc Structural Engineering

1.8.3 Risk Management

‘Success is about being able to deal with failure’.
Risk Management can be used in several different ways during construction projects. The most
common use is to avoid un-necessary cost and time over-runs. In civil engineering it is
particularly conspicuous when major public works take significantly longer to build or are
significantly more expensive than estimated. The media are particularly attracted to such stories
as they involve public money.
Several government initiatives have attempted to deal with this in recent years, particularly
Latham’s Build the Team (1994) and Egan’s Rethinking Construction (1998), which spawned the
Movement for Innovation (M4I).
The out-turn cost is often about: π x estimated cost (complex projects)
1.75 x estimated cost (intermediate projects)
1.2 x estimated cost (repetitive projects)
Reason for overruns include:
• Objectives unrealistic or changed during project
• Cost estimates optimistic
• Contingency underestimated
• Project brief incomplete or unclear at the beginning
• Design incomplete before construction begins
• Design poor – not buildable
• Risk allocation ambiguous
• Project Management inadequate
Many of the factors become easier to predict or control as the project progresses through
gateways – defined stages, such as:
• Business case
• Allocation of procurement strategy
• Scheme design
• Construction contract awarded
• Construction close-out
It is good practice for clients to obtain project funding which will cover the tender price plus a
contingency to allow payment for unforeseen events. Quantifying this contingency for time and
cost is the purpose of risk management. To do this we must assess the time and cost implications
of an event happening.
Identify the event
Produce a list of things which could go wrong during a project. Concentrate on those elements
which are outside the project teams direct control, e.g. availability of materials, weather,
changes to the clients requirements. Make an assessment of the cost and time implications of
the event. Sketching helps this process, as it is dependant upon knowing how the bridge will be
constructed and the likely sequence of activities which must be followed to do it safely.
Assess the probability an event will happen
It is inevitable that some of these events will occur during the project; equally some will not.
Make a qualitative assessment of the likelihood the event will occur.
Assess the severity of an event happening
Assuming the event happens during the project, make a qualitative assessment of the likely
consequences.
The risk matrix
All assessments of risk are subjective, and so too are the combinations of probability and severity.
The Salford Risk Management Matrix may be used to quantifying risk (other matrices are
available and may be equally valid). Using the event probability and severity already assessed,
obtain a risk factor for the event from the matrix.
© Jonathan Haynes version 3.0 Page 28 of 239
32900 Bridge Engineering
MSc Structural Engineering

The Salford Risk Management Matrix

Risk Factor Qualitative assessment of event probability
γrisk Very High High Medium Low Very Low
Very High 0.95 0.875 0.75 0.625 0.5
assessment of
Qualitative

event severity

High 0.875 0.75 0.625 0.5 0.4

Medium 0.75 0.625 0.5 0.4 0.3
Low 0.625 0.5 0.4 0.3 0.2
Very Low 0.5 0.4 0.3 0.2 0.05

Quantify the risk

To quantify the risk simply obtain the product of the event cost and risk factor. Similarly for event
time risk.
Risk prioritisation
Once the risks of all the possible events have been quantified, they can be ranked and actions
taken to mitigate the most probable events. These may include:
• AVOID the risk – change the design
• REMOVE the risk – omit parts of a project which you cannot control
• REDUCE the risk – apply control measures, such as using a specialist subcontractor
• TRANSFER the risk – insure against the event
• ACCEPT the risk – apply your own management of the event, if and when it happens
Remember, it is best to allocate ownership of a risk to the person, or organisation, which is best
placed to control (or manage) it. Cost over-runs due to poor weather cannot be controlled by
the contractor, so the client is best placed to deal with this event.
It is possible to undertake very complex and accurate risk analyses (Monte Carlo simulations),
which require in-depth knowledge of the project and probabilities involved. These are often run
on a computer which will simulate a million scenarios, and identify then rank the project risks.
Risk management may also be used to undertake ranking of scheme designs, by summating the
risk cost and time of each scheme. The results may then be compared and used in a holistic
assessment of scheme cost, time and quality.
Of course, the likelihood of all the possible detrimental events happening during a project is
small. For this reason it is customary to address the risk cost and risk time using one of several
statistical techniques. The most simplistic is to quantify the square root of the sum of the squares
(SRSS) for cost and time.
A Bridge Building Case Study
The following pages contain details of five bridge schemes, which will be assessed using risk
management techniques, in order to assist in the selection of the best design. A risk contingency
estimate has been made which addresses the key design issues associated with each scheme.
These are not restricted to health & safety but cover any issue which may affect the project
outcome. The results are:
Gross Risk SRSS
Bridge scheme
Cost (£) Time (days) Cost (£) Time (days)
1. Suspension 599,000 25 484,000 16
2. Symmetrical cable stayed 484,000 93 301,000 55
3. Asymmetrical cable stayed 171,000 98 129,000 49
4. Segmental pre-cast concrete 885,000 114 637,000 60
5. Launched pre-cast concrete 376,000 68 273,000 41

This assessment suggests that scheme 3 provides the best option in terms of cost. However,
scheme 1 is least risky in terms of time. When the cost and programme time of construction have

© Jonathan Haynes version 3.0 Page 29 of 239

32900 Bridge Engineering
MSc Structural Engineering

been calculated, the SRSS risk cost and time, of the selected scheme, should be added to
produce a tender price.
Project risk management is largely about finding solutions for problems before they happen; and
when there is no simple solution, ensuring there is sufficient funding and time to allow the
problem to be fixed without resorting to a dispute.
A contract is just a means of defining who should be responsible for risk, although the client may
not appreciate this, the party who is best placed to control a risk should be responsible for it in
the contract. In this way, the minimum amount of conflict should result during the execution of
the works.
Like all risk assessment techniques, the results are dependent upon personal opinion – so
engineers’ ability to use risk management techniques requires judgement, which improves with
experience.
The Risk Contingency Estimates.xls spreadsheet used to undertake the assessment is available on
Blackboard.

The spreadsheet requires descriptive text entries for the unforeseen event and estimates of its
likely cost in terms of time and monetary cost. Severity and probability of the event must be
selected from drop-down lists. Descriptive text entries for suggested control measures can also
be entered. One spreadsheet should be produced for each scheme design.
Visual representations of risk cost and risk time for each event are automatically generated to
aid understanding

© Jonathan Haynes version 3.0 Page 30 of 239

 32900 Bridge Engineering
MSc Structural Engineering

Scheme 1- Suspension Bridge

The process of assessing risk (at

any level), estimating costs,
construction programme
periods, structural form and
aesthetics are all underpinned
by outline construction
sequence sketches.

Sketching is a skill demanded

by the Joint Board of
Moderators and lauded by
employers.

Sketching is a basic
communication skill, and will
save time and effort over
written explanations of the
construction process /
sequence.

Risk Contingency Estimate

Project Title : Scheme 1 - Suspension Bridge
Date : Febrauary 2012
Event Cost Time Severity Probability Risk Factor Cost of Risk Time at Risk Comments or control measure put in place

estimated estimated qualitative qualitative

cost of the time of the assessment assessment of cost x risk time x risk
event event of event event γrisk taken from factor factor
(£) (days) severity probability risk matrix (£) (days)
North bank foundating level £35,000 4 Medium Low 0.400 £14,000 1.6 Site investigation limited to two boreholes,
must be 3m lower than both 10m from a probable pylon location.
expected
Poor weather extends £950,000 25 Medium Medium 0.500 £475,000 12.5 Unable to restrict cable spinning operations
cable spinning period to spring / summer periods.

Unavailability of cable and £300,000 30 Very High Very Low 0.300 £90,000 9.0 Can only place firm order for material when
hanger material in sufficient contract is signed.
quantities to meet
programme

Humber pilot dispute £65,000 5 High Very Low 0.300 £19,500 1.5 Possibility of using a private river pilot in the
prevents deck sections unlikely event of an industrial dispute.
being delivered by boat as
required to meet
programme
TOTAL RISK £598,500 25 days

SQUARE ROOT OF THE SUM OF THE SQUARES RISK £484,047 16 days

© Jonathan Haynes version 3.0 Page 31 of 239

 32900 Bridge Engineering
MSc Structural Engineering

Scheme 2 – Symmetrical Cable-stayed Bridge

This is a poor set of

diagrams, as they do
not:

1. show how the pylon is

lifted (temporary works);

4. show how the deck

cantilever is supported
whilst the cable is
installed and tensioned;

6. show any information

about the construction
of the central sections –
how is the deck
stabilised transversely
whilst it cantilevers ?

Risk Contingency Estimate

Project Title : Scheme 2 - Symmetrical Cable-stayed Bridge
Date : Febrauary 2012
Event Cost Time Severity Probability Risk Factor Cost of Risk Time at Risk Comments or control measure put in place

estimated estimated qualitative qualitative

cost of the time of the assessment assessment of cost x risk time x risk
event event of event event γrisk taken from factor factor
(£) (days) severity probability risk matrix (£) (days)
Unique bearing £250,000 30 High Medium 0.625 £156,250 18.8 Commission finite element anaylsis of
arrangement at back-stay proposed connection before final design to
fails in testing investigate likely failure modes.

Temporary lateral bracing £35,000 10 Medium Low 0.500 £17,500 5.0 Adjust programme to ensure deck is
required for deck sections in constructed in summer months.
high winds
Articulated deck edge £150,000 60 Very High Very Low 0.400 £60,000 24.0 System has been used successfully before.
beam / hanger erection Standby temporary supports on pontoons.
system fails

One of the two £500,000 90 High Low 0.500 £250,000 45.0 Redesign cable to use lower strength but
manufacturers of special more readily available material.
high tensile cable goes out
of business

TOTAL RISK £483,750 93 days

SQUARE ROOT OF THE SUM OF THE SQUARES RISK £301,364 55 days

© Jonathan Haynes version 3.0 Page 32 of 239

 32900 Bridge Engineering
MSc Structural Engineering

Scheme 3 – Asymmetrical Cable-stayed Bridge

If you were tasked

with constructing this
bridge from this
construction
sequence, what
questions would you
ask the designer ?

Risk Contingency Estimate

Project Title : Scheme 3 - Asymmetrical Cable-stayed Bridge
Date : Febrauary 2012
Event Cost Time Severity Probability Risk Factor Cost of Risk Time at Risk Comments or control measure put in place

estimated estimated qualitative qualitative

cost of the time of the assessment assessment of cost x risk time x risk
event event of event event γrisk taken from factor factor
(£) (days) severity probability risk matrix (£) (days)
River bed material weaker £165,000 35 Very High Medium 0.750 £123,750 26.3 Site investigation limited to two boreholes,
than expected, requiring both on banks of river.
deeper pylon foundation

Poor weather extends £6,000 30 Low Medium 0.750 £4,500 22.5 Caisson construction not on critical path for
caisson construction period completion.

Poor weather delays £8,000 30 Very High High 0.875 £7,000 26.3 Deck construction forms critical path for
deployment of pontoons to completion for 8 weeks.
temporarily support deck

Construction of caisson in £72,000 45 High Very Low 0.500 £36,000 22.5 Include anti-scour rock armour in design.
river bed generates scour
significantly worse than
expected

TOTAL RISK £171,250 98 days

SQUARE ROOT OF THE SUM OF THE SQUARES RISK £129,148 49 days

© Jonathan Haynes version 3.0 Page 33 of 239

 32900 Bridge Engineering
MSc Structural Engineering

Scheme 4 – Segmental Pre-cast Concrete Bridge

Risk Contingency Estimate

Project Title : Scheme 4 - Segmental Precast Concrete Bridge
Date : Febrauary 2012
Event Cost Time Severity Probability Risk Factor Cost of Risk Time at Risk Comments or control measure put in place

estimated qualitative qualitative

estimated cost time of the assessment assessment of cost x risk time x risk
of the event event of event event γrisk taken from factor factor
(£) (days) severity probability risk matrix (£) (days)
Insufficient space on site to £150,000 60 High Low 0.500 £75,000 30.0 This construction technique requires previous
precast deck sections section to be used as a shutter (the surfaces
will be mated later). A large space is
required to undertake this operation.
Investigate temporary used of adjacent
waste land for this purpose.
Failure of temporary post- £500,000 90 Very High Very Low 0.400 £200,000 36.0 Adjacent units are resin bonded and post-
tensioing tensioned in temporary case. Use personnel
and materials which have been successful in
the past.
Unsatisfactory durability of £1,500,000 90 Very High Very Low 0.400 £600,000 36.0 Possible insurance claim / corrective works if
deck joints deck joints are not sealed. Use personnel
and materials which have been successful in
the past.
Fall from falsework £25,000 30 Very High Very Low 0.400 £10,000 12.0 Use competent personnel. Apply rigorous
safety procedures.

TOTAL RISK £885,000 114 days

SQUARE ROOT OF THE SUM OF THE SQUARES RISK £636,965 60 days

© Jonathan Haynes version 3.0 Page 34 of 239

 32900 Bridge Engineering
MSc Structural Engineering

Scheme 5 – Launched Bridge

4. is it really possible to
launch with the sort of
arrangement shown ?

Risk Contingency Estimate

Project Title : Scheme 5 - Launched Precast Concrete Bridge
Date : Febrauary 2012
Event Cost Time Severity Probability Risk Factor Cost of Risk Time at Risk Comments or control measure put in place

estimated estimated qualitative qualitative

cost of the time of the assessment assessment of cost x risk time x risk
event event of event event γrisk taken from factor factor
(£) (days) severity probability risk matrix (£) (days)
Compelled set up concrete £40,000 20 Medium Very High 0.750 £30,000 15.0 Remote location precludes use of ready
batching plant mixed concrete delivered to site.

Launching jack or bearing £125,000 12 Very High High 0.625 £78,125 7.5 Jacks and bearing under continual stress.
failures Provide spares and standby fitting team.

Loss of stability during £650,000 90 Very High Low 0.400 £260,000 36.0 Positioning and quantity of kentledge very
launching important during movement of deck.
Rehersal of sequences and backup for
computer controls essential.

Fall from deck £25,000 30 Very High Very Low 0.300 £7,500 9.0 Workforce only required to work over water
after deck is complete.

TOTAL RISK £375,625 68 days

SQUARE ROOT OF THE SUM OF THE SQUARES RISK £273,239 41 days

© Jonathan Haynes version 3.0 Page 35 of 239

32900 Bridge Engineering
MSc Structural Engineering

1.9 Comparison of Schemes

The comparison of scheme designs, for the purposes of recommending a scheme to a client, is
partly subjective but must present a reasoned quantitative decision. Commonly used
comparators include:
• aesthetic appearance
• environmental suitability
• potential durability problems, such as bearings or river piers
• initial construction cost, including risk cost
• initial construction time and risk time (particularly on railway bridges)
• maintenance costs, including the delay cost to bridge users
• alignment quality
• buildability
• carbon cost and other sustainability issues
It is conceivable that some comparators are more important than others and will therefore
require greater weighting, the relative values are a decision for the student but must be justified.
A gross summary of scheme comparison should be presented, this may take any form but the
following table is an example.
1 2 3 4
Weighting (%)

Comparator

Splayed tubular arches and The inclined pylons and Structurally efficient Traditional masonry form
simple hanger arrangement asymmetric cable parabolic bottom chord in fits well with environment
give a pleasing visual arrangement are poor. steel cable gives modern but is too visually
Aesthetic appeal 5 appearance from all appearance but lacks intrusive (massive).
approaches. drama.
Mark 9/10 Weighted Mark 4/10 Weighted Mark 7/10 Weighted Mark 8/10 Weighted
0.45 0.20 0.35 0.40
£6,750,000 main risks £8,250,000 main risks in £4,500,000 main risks in £5,000,000 main risks
single lift of arches. temporary stability of post-tensioning cable to involve finding arch
Construction cost 25 pylons. high level. masons.
Mark 7/10 Weighted Mark 5/10 Weighted Mark 10/10 Weighted Mark 9/10 Weighted
1.75 1.25 2.50 2.25

Σ 100 12.55 6.30 15.45 13.65

In this example, scheme 3 is the preferred option.

© Jonathan Haynes version 3.0 Page 36 of 239

32900 Bridge Engineering
MSc Structural Engineering

Bridge Concepts - Conclusion

Things to do now :
1. Review the Powerpoint lectures on Blackboard.
2. Attempt the Self Assessment Exercises below. Then check your answers against the
solutions on Blackboard.
3. Have a rest.

4. Prepare your group submission for Assignments 1 and 2.

Self Assessment Exercises

1. Draw a transverse section through a three lane motorway and a two track railway, include the
minimum vehicle clearances.
2. Sketch three viaduct arrangements for the valley below, using different structural forms.

Scale 1:2000
0 30 60 90 120 150 180 210 240 270 300

Scale 1:2000
0 30 60 90 120 150 180 210 240 270 300

Scale 1:2000
0 30 60 90 120 150 180 210 240 270 300

3. Sketch four arrangements for the motorway overbridge below, using different structural forms.

0 5 10 15 20 25 30
Scale 1:500

© Jonathan Haynes version 3.0 Page 37 of 239

32900 Bridge Engineering
MSc Structural Engineering

4. Sketch two arrangements for the gorge below, using different structural forms.

0 5 10 15 20 25 30
Scale 1:500

0 5 10 15 20 25 30
Scale 1:500

5. Identify the advantages and disadvantages of the various structures developed above and
suggest the best materials to use in each.
Further Reading
Agrawal, R. (2009). Stress ribbon bridges. The Structural Engineer, 87: 22, pp22-27.
Bennett, D. (1997). The Architecture of Bridge Design. London: Thomas Telford. [ISBN 0-7277-2529-7]
Collings, D. (2006). An environmental comparison of bridge forms. Bridge Engineering. 159: BE4,
pp 163–168.
Engel, H. (2007). Structure Systems. 3rd Edition. Ostfildern: Hatje Cantz. [624.1771/ENG]
Gottemoeller, F. (1998). Bridgescape: The Art of Designing Bridges. Chichester: Wiley.
Leonhardt, F. (1982). Bridges. Aesthetics and Design. London: Architectural Press.
Mainstone, R.J. (2001). Developments in structural form. Oxford: Architectural Press. [FOLIO
721/ROW and info4education]
Schaich, J. & Scheef, H. (1982). Concrete box-girder bridges. Zurich: International Association for
Bridge and Structural Engineering. [ISBN 3-85748-0319]
Spencer, P.C., Hendry, C.R. & Petty, R. (2012). Quantification of sustainability principles for bridge
projects. Bridge Engineering. 165: BE2, pp 81–89.
The Institution of Civil Engineers (2010). Civil Engineering Standard Method of Measurement 3:
Carbon and Price Book 2011. London: Thomas Telford.
The Highways Agency. (1996). The appearance of bridges and other highway structures. London:
HMSO. [624.2563/HAM]
The Highways Agency. (2005). TD 27/05. Design Manual for Roads and Bridges. Volume 6 Road
Geometry. Section 1 Links. Part 2 Cross Sections and Headrooms. Norwich: H.M.S.O.

© Jonathan Haynes version 3.0 Page 38 of 239

32900 Bridge Engineering
MSc Structural Engineering

2. Bridge Deck Loading & Analysis

This section will introduce Eurocode bridge deck loading rules and some of the techniques
which can be used to create computer models of bridge decks subject to vehicle loads. Simple
hand analysis methods for beam and slab decks will be explored in order to offer a means of
validation for more complicated models.

You should also review: 2 Bridge Deck Loading.ppt , 4a Influence lines and surfaces.ppt and
4b Grillage analysis and FEA.ppt on Blackboard
Principal References: Hambly, E.C. Bridge Deck Behaviour.
O'Brien, E. & Keogh, D.L. Bridge Deck Analysis.

The bridge deck is the primary load application point on a bridge. Many countries have their
own code based rules for calculating actions on bridge decks. Eurocode rules are contained in
EN1991-2 Traffic loads on bridges, EN1991-1-4 Section 8 - Wind loads on bridges, EN1991-1-5
Section 6 – Temperature changes in bridges and EN1990 A.2 Special combination rules for
bridges.

2.1 Bridge Deck Loading

Like any structure designed to ultimate limit state principles, the most onerous combination of
load cases must be identified. This is complicated by the fact that the principal loading moves.
Bridges are designated as Design Working Life Category 5 structures – 120 year life (in the UK).
Loading is split into three categories of action;
• Permanent Actions (G) include self weight, fixings, road surfacing, and indirect actions due to
shrinkage, prestress and differential settlement.
• Variable actions (Q) include wind, snow, water and vehicle loads.
• Accidental actions (A) include vehicle impact and earthquake loads.
There are several limit states to consider. The following Ultimate Limit States may be relevant to a
bridge design:
• EQU - loss of static equilibrium
• STR - internal failure or excessive deformation of a structural member
• GEO - failure or excessive deformation of the ground
• FAT - fatigue failure of a structural member
• UPL - loss of equilibrium due to buoyancy effects
• HYD - loss of equilibrium due to hydraulic heave, erosion or piping in the ground
The key differentiator between limit state checks are the values of the partial safety factors for
actions.
ULS, combination values may be calculated from, ∑γ G Gk + γ Q Qk 1 + γ Qψ 0 ,i Qki + γ P P
The Serviceability Limit State will be consideration of deformation and vibration.
SLS, frequent values may be calculated from, ∑G k +ψ1,1Qk1 +ψ1,i Qki + P
where Gk and Qk are the characteristic permanent and variable actions respectively, P is the
prestress, γG is the partial safety factor for action, and ψo is the sensitivity factor for secondary
variable actions. All conceivable actions should be applied to the deck simultaneously, where
this involves more than one variable action it is necessary to consider each in turn as the leading
action with probabilistically reduced accompanying actions.

Normal combination exclusions in the UK:

For road, rail and footbridges, snow is not an onerous action (except roofed footbridges) and the
combination of wind with thermal is generally ignored.

© Jonathan Haynes version 3.0 Page 39 of 239

32900 Bridge Engineering
MSc Structural Engineering

EN1991-2 and the NA set out combinations of variable action groups to be considered for bridges.

Road Bridges
Vertical Forces Horizontal Forces
Group
LM1 LM2 LM3 LM4 Braking Transverse
1a γQ ψ0γQ
1b γQ
2 ψ1γQ γQ γQ
3 γQ
5 ψ1γQ γQ
6 γQ γQ γQ
Pedestrian Bridges
Vertical Forces Horizontal Forces
Group
LM1 LM2 LM3 LM4 Braking Transverse
4 γQ
Rail Bridges
Vertical
Horizontal Forces
Group Forces Groups 1-4 relate to
LM71 or SW/0 Braking Centrifugal Nosing single tracks, groups 21-24
1 γQ relate to twin tracks. All
2 γQ tracks must be loaded for
3 ψ0γQ γQ ψ1γQ ψ1γQ analysis.
For three or more tracks,
4 ψ0γQ ψ1γQ γQ γQ any two tracks must be
21 γQ ψ0γQ ψ1γQ ψ1γQ fully loaded and the
22 γQ ψ1γQ ψ0γQ ψ0γQ remaining tracks
23 ψ1γQ γQ ψ0γQ ψ0γQ unloaded.
24 ψ1γQ ψ0γQ γQ γQ
Sensitivity factors ψ for Road Bridges.
Action Group Component ψ0 ψ1 ψ2
Tandem axle system 0.75 0.75 0
Gr 1a UDL 0.75 0.75 0
Pedestrian 0.4 0.4 0
Traffic
Gr 1b Single axle 0 0.75 0
Gr 2 Horizontal loads 0 0 0
Gr 3 Pedestrian loads 0 0.4 0
Wind Fwk 0.5 0.2 0
Thermal Tk 0.6 0.6 0.5
Snow Qsn,k 0.8 0 0
Sensitivity factors ψ for Footbridges.
Action Group Component ψ0 ψ1 ψ2
Traffic Gr 4 Crowd 0.4 0.4 0
Wind Fwk 0.5 0.2 0
Thermal Tk 0.6 0.6 0.5
Snow Qsn,k 0.8 0 0
Sensitivity factors ψ for Railway Bridges.
Action Group Component ψ0 ψ1 ψ2
Tandem axle system 0.8 0.8 0
Traffic Gr 1, 2, 3, 4, 21, 22, 23, 24
UDL 0.8 0.7 0
Wind Fwk 0.75 0.5 0
Thermal Tk 0.6 0.6 0.5
Snow Qsn,k 0.8 0 0

© Jonathan Haynes version 3.0 Page 40 of 239

32900 Bridge Engineering
MSc Structural Engineering

Action partial safety factors, γf STR / GEO Limit State EQU Limit State
Gk Qk Gk Qk
Unfav Fav Unfav Fav Unfav Fav Unfav Fav
Concrete, ballast, soil 1.35
0.95 1.05 0.95
Steel, road surfacing 1.20
Road traffic
1.35 1.35
Pedestrian
Wind 1.70 1.70
Thermal 1.55 0 1.55 0
Rail – LM71, SW/0, HSLM 1.45 1.35
Rail – SW/2 1.40 1.40
Rail - Real 1.70 1.70
Note that all SLS partial safety factors for actions are set at 1.0

Unit weights of common building materials (EN1991-1-1)

Material (kN/m3) Material (kN/m3)
asphalt hot rolled 23 softwood timber 5
ballast 20 soil (sand) 16-18
clay masonry facing brickwork 22 soil (clay) 20-22
glass 25 steel 77
hardwood timber 7 stone (medium density granite) 29
reinforced concrete 25 water 10

2.1.1 Traffic Actions

On road bridges the carriageway is divided into notional lanes, each of 3.0m width. This ignores
any road markings which may be intended to limit the number of lanes of traffic. Any area
separated from the carriageway by barriers may be excluded from consideration. For a
carriageway of width w between kerbs, the number of full loaded lanes is, w any further
3
carriageway is referred to as remaining area. The inside lane is designated lane 1, this is where
special vehicles (abnormal loads) are assumed to be located.
Road traffic loads are predicted using one of the prescribed bridge loading models:
Loading Model 1 – General lorry and car loading double axle (tandem system TS).
General traffic loading is represented by a uniformly distributed loading of 2.5kN/m2, and if a
carriageway has remaining area then this is the only loading applied. To account for the higher
probability of trucks on the inside lane the uniformly distributed loading is increased to 9.0kN/m2
(or 27kN/m). In addition a tandem axle, which represents a truck cab, is placed in the first three
lanes; 600kN in lane 1, 400kN in lane 2 and 200kN in lane 3. The wheel contact area is 400mm
square. The TS axles must be arranged to produce the most onerous effect (usually lined up at
the influence line peak).

1.20m
3.0m
2.0m

u.d.l. 9.0 kN/m2 Lane 1. 300kN TS

300kN

300kN

direction of travel
2.0m

3.0m

u.d.l. 2.5 kN/m2 Lane 2. 200kN TS

200kN

200kN

2.0m

3.0m

u.d.l. 2.5 kN/m2 Lane 3. 100kN TS

100kN

100kN

u.d.l. 2.5 kN/m2 Remaining area

© Jonathan Haynes version 3.0 Page 41 of 239

32900 Bridge Engineering
MSc Structural Engineering

Loading Model 2 – Single axle system.

Used for very short structural elements (3-7m) or investigation of local effects.

2.0m
This model is not intended for global analysis consideration, it constitutes a
single axle of 400kN with wheels spaced at 2.0m as specified in LM1. This axle

400kN
load should not be combined with uniform loading. A single wheel of 200kN
may be considered separately when more onerous forces are generated. The
wheel contact area is 400mm square.
Loading Model 3 – Special Vehicles.
There are a range of special vehicles under the Special Type General Order (STGO) and Special
Order (SO) Regulations. Common STGO vehicles weigh 80, 100 and 196 Tonnes (SV vehicles) and
250, 350, 450 and 600 Tonnes (SOV vehicles).
An SV100 vehicle will be detailed here (category 3 vehicle of gross weight 100 Tonnes) six axle
loads of 165 kN, each axle having two wheels with contact area of 350mm square. Details of
other vehicles are contained in EN1991-2 UKNA 2.16.1/2.
Load model 3 should be combined with LM1 loading with the following restrictions:
• Only one special vehicle should be applied to a bridge deck analysis
• Special vehicle axles loads must be multiplied by the relevant Dynamic Amplification Factor
• The LM1 loading is accompanying variable action and is reduced to frequent action level (ψ1)
• 5.0m of unloaded deck should be left between LM1 TS and LM3 axles, front and rear
• When applied to an influence line the whole SV must be used (no truncated loadings)

3.0m
3.0m

ψ1 9.0 kN/m2 Lane 1. SV + ψ1300kN TS

5.0m 1.2m most onerous of 5.0m

3.0m
ψ1 2.5 kN/m2 Lane 2. ψ1200kN TS
1.2, 5.0 or 9.0m

direction of travel
3.0m

ψ1 2.5 kN/m2 Lane 3. ψ1100kN TS

Special Vehicle axle Dynamic Amplification Factors (DAF).

Axle load (kN) 100 130 165 180 225
DAF 1.20 1.16 1.12 1.10 1.07

Loading Model 4 – Crowd Loading. A uniformly distributed loading of 5.0kN/m2 is applied to the
entire deck of footbridges. Generally this should also be applied to road and rail bridge footways
but if required a reduced pedestrian allowance may be used:
120
2.50kN/m 2 ≤ q fk = 2 + ≤ 5.0kN/m 2 where L is the loaded length
L + 10
On railway bridges loadings are defined for a pair of running rails. Like road bridges, there are
several load models, which include:
• LM71 – the normal rail traffic load train
• SW/0 - the normal rail traffic load train for continuous bridges only
• SW/2 - the heavy rail traffic load train (not for use in the UK)
• HSLM – the high speed (>200km/hr) passenger load train
• Unloaded train.
Only LM71 and SW/0 will be detailed here.

© Jonathan Haynes version 3.0 Page 42 of 239

32900 Bridge Engineering
MSc Structural Engineering

LM71 Qk train envelope

4 x 250kN approx.
80kN/m 80kN/m 2.8m wide x
3.8m high
u.d.l of unlimited length 3 x 1.6m 0.8m QkL
QkR
e

SW/0 375
133kN/m 133kN/m
r = 1.435m 4 300
1
5.3m 15m
r QkL
Centripedal effects, e ≤ and ≤ 1.25
18 QkR

• Where a bridge carries more than one set of running rails, the load train shall be applied
simultaneously for each pair of running rails. The minimum distance between centrelines of
running rail pairs is 3.400m
• The load train values must be increased by a factor α, which is 1.10 in the UK
• Eccentricity of vertical loading should be accounted for by redistributing the loading
between rails by up to a proportion of 1.25:1 but limiting the eccentricity considered to one
eighteenth of the rail centres. This is achieved by loading each of the rails with 56% and 44% of
the loading model variable action respectively
• Rail loads may be spread through the ballast at 4:1 down to the bridge deck structure.
Without precamber, the minimum dimensions for the rail and sleeper, and ballast are 375mm
and 300mm respectively.

Horizontal Deck Actions

A braking force Qlk is applied as a longitudinal

Qlk horizontal force at the surfacing level of the
carriageway. It is limited to 900kN on any deck.
Qtrk A centrifugal force Qtk is applied radially as a
horizontal force, as a point load at any deck
cross-section on a plan curve.

For LM1 the force is a fraction of the total vertical TS and u.d. loads and applied in each lane
Qlk = 0.6(2TS ) + 0.1qwL where TS is the tandem system axle load, q is the uniform load, w is the
lane width and L is the span length being considered.
TS axle (kN) u.d.l (kN/m2) Qlk (kN) Qtrk (kN)
300 9 360+2.7L 184.1
200 2.5 240+0.75L 121.1
100 2.5 120+0.75L 90.0
For LM3 force is also applied in each lane. Acceleration forces are considered to be 10% of the
vehicle weight (less onerous than braking forces).
Qlk , s = δ ⋅ W where δ is the deceleration factor and W is the basic axle load of the SV
For LM1 a skew braking or skidding force Qtrk should be applied laterally in combination with Qlk
Qtrk = 0.5Qlk for loaded lengths up to 120m, or 280kN for loaded lengths >120m
For LM3 a skew braking or skidding force Qtk,s should be applied laterally in combination with Qlk,s
Wv 2
Qtk , s = where W is the weight of the special vehicle, v is its velocity (related to the type of
gr
vehicle), g is acceleration due to gravity and r is the radius of the carriageway horizontal curve.

© Jonathan Haynes version 3.0 Page 43 of 239

32900 Bridge Engineering
MSc Structural Engineering

Qtk,s (kN) for increasing values of r (m)

Vehicle δ Qlk,s (kN)
r ≤ 200 250 500 750 1000 1500
SV80 0.5 392.4 266.3 213.0 106.5 71.0 53.3 35.5
SV100 0.4 392.4 266.8 213.4 106.7 71.1 53.4 35.6
SV196 0.25 480.7 266.8 213.4 106.7 71.1 53.4 35.6
SOV250 0.2 490.5 189.1 151.3 75.6 50.4 37.8 25.2
SOV350 0.2 686.7 204.1 163.3 81.6 54.4 40.8 27.2
SOV450 0.2 882.9 220.5 176.4 88.2 58.8 44.1 29.4
SOV600 0.2 900 243.0 194.4 97.2 64.8 48.6 32.4

Horizontal loading on rail bridges is complex, and centrifugal force calculation is not included
here. Since all rail vehicles will weave between the rails whilst moving forwards, a nosing force of
100kN must be applied laterally to one rail in combination with vertical forces.
For LM71 and SW/0 load trains the following accelerating and braking forces apply longitudinally
along the rail, in combination with vertical forces:
Traction (acceleration) force, Qlak = 33 kN/m ≤ 1000 kN

Braking force, Qlak = 20 kN/m ≤ 6000 kN

2.1.2 Wind Actions

EN1991-1-4 Section 8 deals with wind loading on bridges. These notes will detail the simplified
method of obtaining wind loads for single decks which are of constant depth. When considering
load combinations of wind and traffic, the area of vehicles should be included in deck depth.
Use of the simplified method for wind loads is dependant upon the assumptions that:
• bridges which span less than 200m need no specific allowance for structural response
• dynamic amplification of vertical response may be ignored if the natural frequency in both
bending and torsion is less than 1.0Hz
The natural frequency may be obtained from the Eigen values of an adequate stiffness analysis
of the deck. Alternatively, EN1991-1-4 Annex F provides methods to obtain the bending and
torsional natural frequencies using hand calculations.
The UK National Annex provides a design based alternative check;
2
 vm   ρb 2  σ fm b 
     ≤ 1.0 where vm is the site mean wind speed, fN is the deck natural
 fNb   m  σ c 
frequency, b is the deck width, ρ is the density of air 1.226kg/m3, m is the mass per unit length of
deck, σfm is the peak stress per unit deflection in first mode vibration, and σc is a reference stress
taken as 600 N/mm2 for beams, 750 N/mm2 for trusses and 80 N/mm2 for concrete structures.

The site mean wind speed, Vm = Vb cr and cr may be taken from EN1991-1-4 figure NA.3

The fundamental natural frequency may be estimated from first principles using an analysis
1 K
model and f N = where K is the structure stiffness (N/m) and m is the participating mass
2π m
(kg), alternatively the method given in Annex F(5) may be used.
Due to the level of arrangement information required to complete this check, it must be
undertaken after the design is finalised.
EN1991-1-4 NA2.49.3 offers a method to check aeroelastic stability of bridges.

It is likely that topography may be onerous for many bridges. EN1991-1-4 figure NA.2 can be used
to determine whether the site is located sufficiently close to the summit of a hill or escarpment to
require enhancement to the wind speed.

© Jonathan Haynes version 3.0 Page 44 of 239

32900 Bridge Engineering
MSc Structural Engineering

z b is the width of the deck

L is the overall length of the deck
y d is the depth of the deck structure which is solid.
If the parapet is solid, its full height should be
included. For open parapets and vehicle barriers
L 0.30m height of wind should be assumed for each.
d Structures like trusses allow the wind to blow
x through and may be assessed as the net area of the
front face.
b

Calculation of simplified wind pressure across the deck in the x-direction.

The basic 10 minute mean wind velocity, Vb = Vb ,0 c dir c season c alt c prob
where Vb,0 is the characteristic 10 minute mean wind velocity (for category II terrain taken from
the UK Isotach), the altitude factor may conservatively be taken as c alt = 1 + 0.001A , where A is
the site altitude above Ordnance Datum. The correction factors cdir and cseason should be taken as
1.0. The probability factor accounts for the increased chance of a strong wind as the life of a
structure increases. Cprob = 1.047 may be used for a return period of 120 years with 5% chance of
exceedance.
qb = 0.613(Vb ,0 c alt c prob )
2
The basic velocity pressure,

The peak velocity pressure, q p = q b c e ≤ 0.75 kN/m 2 road bridges or ≤ 0.98 kN/m 2 rail bridges,
where ce is the exposure correction factor, obtained from a log-log chart which relates distance
upwind from the shoreline to the exposed height of the structure (z-hdis) which may be taken as
the height to the top of the bridge deck.
Z – hdis (m)

distance upwind to shoreline (km)

Characteristic 10 minute mean wind velocity, Exposure factor, ce
Vb,0 (m/s)

Simplified x-direction wind force, Fwx = dLq p c fx

c fx = 1.30 b ≥ 4 .0
where d

d
( )
c fx = −0.3 b − 4 + 1.3 1≤ b < 4
d
© Jonathan Haynes version 3.0 Page 45 of 239
32900 Bridge Engineering
MSc Structural Engineering

Calculation of simplified wind pressure on the deck in the z-direction.

Simplified z-direction wind force, Fwz = bLq p c fz where c fz = ±0.9

Calculation of simplified wind pressure along the deck in the y-direction requires the
identification of the most onerous of wind on the superstructure alone, or wind on the
superstructure + wind on the traffic derived separately. Generally, this may be achieved by
summing the elements:
Deck y-direction wind force, FDy = kbdq p c fx where k=0.25 solid web or k=0.50 trusses.

Traffic y-direction wind force, FLy = kq p At c fx where k=0.50, cfx=1.45 and At is the area of traffic
loading perpendicular to the x-direction wind (road traffic is assumed to be 2.0m high and rail
traffic is assumed to be 4.0m high).
Pier y-direction wind force, FPy = bhq p c fp where h is the pier height, b is the pier width, t is the
pier thickness and cfp is the pier force coefficient.

Pier aspect ratio, h/b

Pier force
coefficient, cfp 1 2 4 10 20 40
≤ 0.25 1.3 1.4 1.4 1.6 1.8 1.9
Rectangular

0.5 1.4 1.5 1.6 1.7 1.9 2.1

plan, t / b

h 1.0 1.4 1.4 1.5 1.5 1.8 2.0

2.0 1.0 1.1 1.2 1.3 1.4 1.6
≥ 4.0 0.8 0.8 0.8 1.0 1.0 1.1
Circular 0.75 0.78 0.82 0.91 1.0 1.0
b

2.1.3 Thermal Actions

Bridges are subject to changes in daily and seasonal air temperature which result in a distribution
of temperature within structural elements. The magnitude of the temperature change is largely
governed by air temperature, solar orientation, mass and structure finishes (black material like
carriageway surfacing absorbs more heat than concrete). Thermal change forces geometrical
change on the bridge, as defined by the coefficient of thermal expansion of the materials used.
Uniform geometrical change caused by seasonal temperature variation induces axial strain over
the entire section; whereas daily temperature variation induces localised geometrical change
and hence bending stresses.

z
y
x

∆TM
∆TMz non-linear
∆TMy
∆TN thermal gradient

The daily temperature change stresses generated are self-equilibrating – internal stresses
balance the forced geometrical change. This is usually assessed as an isolated load case.
The seasonal variations are applied to a global bridge analysis as a temperature change load
case.
EN1991-1-5 Section 6 sets out the requirements for thermal actions on bridge structures. Two
approaches are detailed in the code but in the UK approach 2 is required. This is the

© Jonathan Haynes version 3.0 Page 46 of 239

32900 Bridge Engineering
MSc Structural Engineering

consideration of a vertical temperature component with non-linear thermal gradient through the
deck. Bridges are split into three types to grossly account for differing thermal mass:
1. Steel
2. Composites (using a short-term modular ratio)
3. Concrete
Structural elements which are buried by at least 0.6m of cover material are considered to be
protected from thermal change.
Uniform temperature component, ∆TN is a general change in the bridge temperature which
governs the expansion or contraction of elements and generates uniform axial stress.
For the UK, values of the minimum and maximum shade air temperatures (Tmax and Tmin) are
obtained from Isotherms. These are converted to maximum and minimum bridge uniform
temperature components (Te max and Te min).

Type Te min Te max

1 Tmin - 3 Tmax + 16

2 Tmin + 4 Tmax + 4
3 Tmin + 8 Tmax + 2

Assumed surfacing thickness

1 40mm

2 100mm
3 100mm

Minimum and Maximum UK Isotherms (Tmin and Tmax)

The values of Te should be further modified for actual surfacing thickness.

Type 1 Type 2 Type 3
Temperature range adjustment (°C)
Te min Te max Te min Te max Te min Te max
Unsurfaced 0 +4 -3 0 -1 0
Waterproofed 0 +4 -3 +4 -1 +2
40mm surfacing 0 0 -2 +2 -1 +1
100mm surfacing N/A N/A 0 0 0 0
200mm surfacing N/A N/A +3 -4 +1 -2
The temperature at which the deck or other bridge element is fixed in position (T0) and restrained
is important, since this is the temperature from which all variations are measured. If this is not
specified by the designer, then values of +0 and +20 should be used for expansion and
contraction calculations respectively.
The maximum expansion thermal range, ∆TN exp = Te max − T0
The maximum contraction thermal range, ∆TN con = T0 − Te min
For bearings, the maximum expansion thermal range, ∆TN exp + 20
and the maximum contraction thermal range, ∆TN con + 20
The overall range uniform bridge temperature component, ∆TN = Te max − Te min

© Jonathan Haynes version 3.0 Page 47 of 239

32900 Bridge Engineering
MSc Structural Engineering

In structures where a temperature difference between major components would create a

credible adverse load case (such as between the tie and arch elements in a bow-string truss) the
minimum temperature difference should be 15°C.
Vertical temperature difference component, ∆TM accounts for solar heating of the bridge deck
which generates bending stresses, these are more significant in continuous structures. Values for
the temperature range to be considered depend upon the bridge type, the thickness of
surfacing and whether the bridge is in a cooling or heating cycle.

∆TM,heat (°C) ∆TM,cool (°C)

40mm surfacing Type 1 -6
0.1h 24
0.2h 8 14 0.5h
0.3h
4

40mm surfacing Type 1

21 -5
0.5h 0.1h

100mm surfacing Type 2 100mm surfacing Type 2

10 -10

100mm surfacing Type 3 100mm surfacing Type 3

∆T1 ∆T4
0.3h≤0.15m 0.2h≤0.25m
∆T2 ∆T5
0.1m≤0.3h≤0.25m 0.25h≤0.20m

0.25h≤0.20m ∆T6
0.3h≤0.11m 0.2h≤0.25m
∆T3 ∆T7

∆T1 ∆T2 ∆T3 h (m) ∆T4 ∆T5 ∆T6 ∆T7

8.5 3.5 0.5 ≤0.20 -2 -0.5 -0.5 -1.5
12 3 1.5 0.40 -4.5 -1.4 -1 -3.5
13 3 2.0 0.60 -6.5 -1.8 -1.5 -5
13.5 3 2.5 0.80 -7.6 -1.7 -1.5 -6
1.00 -8 -1.5 -1.5 -6.3
≥1.50 -8.4 -0.5 -1 -6.5
The values of ∆TM,heat and ∆TM,cool may be modified for actual surfacing thickness (other than
40mm type 1 and 100mm types 2 and 3) by reference to EN1991-1-5 Annex B.

To calculate the self-equilibrating stresses which daily thermal changes generate:

• divide the deck vertical section into suitable elements which relate to the steps in the vertical
temperature difference component profile
• calculate the axial restraint forces for each step in the temperature profile, Fi = E ⋅ α ⋅ ∆T ⋅ Ai
• calculate the axial restraint moments about the deck neutral axis for each step in the
temperature profile, M = Fi ⋅ hi
• calculate the axial restraint stresses for each step in the temperature profile, f R = E ⋅ α ⋅ ∆T

• calculate the balancing axial stress, f N =

∑F i

A
Mz
• calculate the balancing bending stresses for each step in the temperature profile, f M =
I
• calculate the final stresses by summing the stress components, f s = f R + f N + f M

© Jonathan Haynes version 3.0 Page 48 of 239

32900 Bridge Engineering
MSc Structural Engineering

Calculate the self equilibrating stresses in a concrete box girder due to heating phase vertical
differential temperature.
0 .3 0 .3
 f + 8
A=1530x103mm3, Iyy=477672x106mm3, α =10x10-6/°C, C32/40 normal weight concrete) Ecm = 22 ck
 32 + 8 
= 22 ×  = 33kN / mm2
 10   10 
5000
100 surfacing 13.5 1.43
3.0 -4.50 1.27 -1.80
150 FB FA 1.35
z = 551 250 FC
hC z1 -1.00 1.78
200 200 Elastic Neutral Axis
1350 949 839 + + =
-0.50
150 110 FD
2.5 -0.83 -2.18 -1.58
2000
∆ TM,heat fR fN fM fS
(N/mm2)

Calculate restraint forces,

Area, Ai (mm2) ∆T (°C) Fi = Eα∆TAi (kN) hi (mm) M = Fi hi (kNm)
A 5000x150=750000 (13.5-3)/2=5.25 33x10x10-6x5.25x750000=1300 551-(150/3)=501 1300x0.501=651
B 5000x150=750000 3.0 743 476 353
C 2x200x250=100000 (3.0-0)/2=1.5 50 318 16
D 2000x110=220000 (2.5-0)/2=1.25 91 912 -83
ΣFi = 2184 ΣMi = 937

Calculate balancing stresses,

z (mm) ∆T (°C) f = ΣFi f R = E ⋅ α ⋅ ∆T zi (mm) f M = Mz fs = f R + f N + f M
N A I yy
1 551 13.5 33x10x10-6x13.5=-4.5 551 937x551/447.672=1.27 -4.5+1.43+1.27=-1.80
2 401 3.0 2184 × 103 33x10x10-6x3.0=-1.0 551-150=401 937x401/447.672=0.92 -1.0+1.43+0.92=1.35
=
3 151 0 1530 ×10 3 0 551-400=151 937x151/447.672=0.35 0+1.43+0.35=1.78
4 839 0 = 1.43 0 949-110=839 937x839/447.672=-1.93 0+1.43-1.93=-0.50
5 949 2.5 33x10x10-6x2.5=-0.83 949 937x949/447.672=-2.18 -0.83+1.43-2.18=-1.58

Where a frame structure is analysed, both the uniform temperature component and vertical
temperature difference should be applied in the analysis. In which case the most onerous
combination should be applied from:
∆TM ,heat + ∆TN exp or ∆TM ,cool + ∆TNcon
Coefficient of thermal
Material
expansion, α x10-6/°C
Aluminium 24
Normal weight concrete 10
Light weight concrete 7
Iron and Steel 12
Stainless steel 16
Masonry 6-10

2.1.4 Earthquake Actions

Earthquake actions on bridges are addressed in EN1998-2. However, consideration of
earthquake loading is not covered by this module and will not be expected in the assessments.

2.1.5 Snow Actions

Snow action is addressed in EN1991-1-3. The magnitude of snow load in the UK is unlikely to
approach the magnitude of primary traffic uniform loading or deck crowd loading, and so
would only be combined as a less significant secondary variable action. It is also probabilistically
unlikely that heavy snow fall would combine with full traffic loading. For these reasons, snow
loading is not considered here.

© Jonathan Haynes version 3.0 Page 49 of 239

32900 Bridge Engineering
MSc Structural Engineering

2.2 Local Analysis Models of Bridge Decks

Computer models for forces in bridge decks will be addressed in the next section. However, it is
important to be able to validate computer models; hand analysis techniques provide a quick
and simple way to do this. This section will therefore deal with how and why vehicle actions are
distributed laterally on a bridge deck, and address the practical use of influence lines for beam
analysis and influence surfaces for slab analysis.

2.2.1 Distribution of Actions on the Deck

The relevant load train must be positioned on the deck to obtain the most onerous design forces.
The position will be different for shear force and bending moment considerations; furthermore,
the position will differ for beam and slab design.
The most simplistic load distribution model is to assume that the deck slab is a wide simply
supported beam, spanning transversely between the longitudinal beams. Clearly this will be as
erroneous as assuming the slab is rigidly fixed between longitudinal beams but it will provide a
quick and pessimistic load distribution model.
The problem is that a bridge deck is a flexible structure in which a deflecting element will shed
some load to adjacent elements, and establishing how much is difficult.

Only whole deck models will predict the interaction of transverse deck and longitudinal beam
stiffness, which controls the real distribution of forces. The problem is further complicated when a
box girder deck is used as significant torsion must be resisted (by circulating shear stresses) when
loading is asymmetrical.
It is possible to make simple 2D linear elastic models for initial design which use vertical springs to
represent longitudinal beams. The stiffness of these springs are established from 2D linear elastic
models of simply supported or continuous beams.

2.2.2 Design Forces in Simple Spans

Where a bridge is formed from single span beams, the maximum bending moments in the beams
may be established by positioning the load train such that one set of axles and the centroid of
the load train are equidistant from the beam midspan (several configurations are possible).

ΣP
P1 P1
P2 P2

= =
L
2 L

The maximum shear force in the beams may be established by positioning one axle over the
reaction and the remaining axles in the span.
Due to long term problems with water/salt leakage at bearings of multi-span bridges, the use of
continuous decks is now preferred.

© Jonathan Haynes version 3.0 Page 50 of 239

32900 Bridge Engineering
MSc Structural Engineering

2.2.2 Analysis of Moving Actions on Continuous Beams using Influence Lines

Since vehicle actions are moving loads, techniques to calculate the forces in deck beams would
require a series of analyses for incrementally moved wheel loads, until the most onerous case of
bending moment, shear and deflection are identified. This would be particularly tedious for
continuous structures.
The influence line is an analysis method which addresses this problem. Although influence lines
can be calculated for individual structures, many texts contain pre-prepared influence line
tables which are suited to the practical analysis of bridge deck beams. Tables for two and three
span beams are included in this text.
The shape of the influence line indicates where to place wheel loads to generate the most
onerous forces (+ positive regions increase the force and – negative regions reduce it). The
influence line only gives results at one point in the beam. Where spans are not identical,
coefficients are given for particular span ratios.
The force, in this case bending moment, is a function of the shorter span. The bending moment
due to a wheel load P, positioned over an ordinate k is given by, M = kL1 F and if there are
multiple equal wheel loads, M = L1 F ∑k i

I.L. for span moments

at this point only

A C C’ A’
L1 L2 L1

loads in this span reduce M loads in this span increase M

k2
k1 +
_ +
loads in this span increase M

For a three span (10:15:10m) continuous beam carrying two 165kN (distributed LM2) wheel loads,
calculate the short span bending moment.

165kN 165kN span moment in outer span

1.20 5.00
L1 = 10m L2=15m L3=10m

0.209
0.154
+ +
a b c
1.667
d e _
The LM2 wheel train has two axles at 1.2m centres. The diagram shows the ordinates for the three
span influence line using a span ratio of 1:1.5:1, taken for end span AC (ordinates a to e) and

© Jonathan Haynes version 3.0 Page 51 of 239

32900 Bridge Engineering
MSc Structural Engineering

midspan of endspan AC. To obtain the most onerous bending moment, place a wheel at the
maximum ordinate in span AC, where the ordinate is 0.209. The second wheel may be placed to
the left or right of the first wheel but the ordinates to the left are larger, so the second wheel is
positioned as shown. Because the ordinates given are at span/6 spacing, it is necessary to
interpolate between ordinates b and c to obtain the ordinate under the second wheel of 0.154.

The short end span bending moment is, M span ,1 = 10 × 165 × (0.209 + 0.154) = 599kNm

Similarly, the long central span bending moment is, M span, 2 = 10 × 165 × (0.245 + 0.192 ) = 721kNm
note the short span L of 10m is used and the interval between tabulated coefficients is 15m/6 or
2.50m.
To calculate the hogging moment at a support, the influence line is shaped as shown below.
Any wheel placed on the spans adjacent to the support being investigated will increase the
bending moment.

I.L. for support moments

at this point only

A C C’ A’
L1 L2 L1
loads in this span increase M loads in this span increase M
_
+ k1 k2 +
loads in this span reduce M

Continuing with the same example, using ordinates for central span CC’ (ordinates f to k) and
interior support. Place a wheel at the maximum ordinate in span CC’ where the ordinate is 0.139.
The second wheel is placed to the right where the ordinates are larger, and interpolation is
required between ordinates g and h to obtain the ordinate under the second wheel of 0.135.

The support bending moment may be calculated as, M sup = 10 × 165 × (0.139 + 0.135) = 452kNm

support moment 165kN 165kN

1.20m

L1 = 10m L2=15m L3=10m

2.50m
f g h j k _
+ 0.139 0.135 +

© Jonathan Haynes version 3.0 Page 52 of 239

32900 Bridge Engineering
MSc Structural Engineering

Two-span Continuous Beam – Bending Moment Influence Lines

Influence Line Coefficients for Midspan Moment

0.4
moment in moment in
0.3 span AC span CE

0.2

0.1

0
A C E
a b c d e f g h j k
-0.1

-0.2

Influence Line Coefficients for Central Support Moment

0.05
A C E
0
a b c d e f g h j k
-0.05

-0.1

-0.15

-0.2

-0.25

-0.3
Span ratio
midspan Section

Ordinates
L 1 : L2

Short Span AC Long Span CE

a b c d e f g h j k
Shortspan

1:1 0.063 0.130 0.203 0.121 0.052 0.032 0.046 0.047 0.037 0.020
AC

1:1.5 0.067 0.137 0.213 0.130 0.058 0.058 0.083 0.084 0.067 0.037
1:2 0.070 0.142 0.219 0.136 0.062 0.085 0.124 0.125 0.099 0.054
support C

1:1 0.041 0.074 0.094 0.093 0.064 0.064 0.093 0.094 0.074 0.041
Central

1:1.5 0.032 0.059 0.075 0.074 0.051 0.115 0.167 0.169 0.133 0.073
1:2 0.027 0.049 0.063 0.062 0.042 0.170 0.247 0.250 0.198 0.108
Longspan
midspan

1:1 0.020 0.037 0.047 0.046 0.032 0.052 0.121 0.203 0.130 0.063
CE

1:1.5 0.016 0.030 0.038 0.037 0.025 0.067 0.167 0.291 0.183 0.088

1:2 0.014 0.025 0.031 0.031 0.021 0.082 0.210 0.375 0.235 0.113

© Jonathan Haynes version 3.0 Page 53 of 239

32900 Bridge Engineering
MSc Structural Engineering

Three-span Continuous Beam – Bending Moment Influence Lines

Influence line for bending moments at midspan of end span AC

and midspan of central span CC'
0.35
0.3 moment in moment in
0.25 span AC span CC'
0.2
0.15
0.1
0.05
0
A C C' A'
-0.05 a b c d e f g h j k l m n
-0.1
-0.15

Influence line for bending moments at interior support C

0.05
A C C' A'
0
a b c d e f g h j k l m n
-0.05

-0.1

-0.15

-0.2

-0.25
Span ratio

Ordinates
L1 : L2 : L1
Section

End Span AC Central Span CC' End Span C'A'

a b c d e f g h j k l m n
1:1:1 0.062 0.127 0.200 0.117 0.050 0.029 0.040 0.038 0.027 0.013 0.012 0.013 0.010
endspan
Midspan

AC
of

1:1.5:1 0.066 0.134 0.209 0.126 0.056 0.051 0.070 0.065 0.046 0.021 0.012 0.012 0.010
1:2:1 0.068 0.139 0.215 0.132 0.060 0.075 0.102 0.094 0.065 0.029 0.012 0.012 0.009
support C

1:1:1 0.043 0.079 0.100 0.099 0.068 0.057 0.079 0.075 0.054 0.026 0.025 0.025 0.020
Interior

1:1.5:1 0.036 0.065 0.082 0.081 0.056 0.102 0.139 0.130 0.092 0.042 0.024 0.025 0.020
1:2:1 0.030 0.056 0.070 0.069 0.048 0.151 0.204 0.188 0.129 0.058 0.023 0.023 0.019
of central
span CC'

1:1:1 0.016 0.030 0.038 0.037 0.025 0.042 0.100 0.175 0.100 0.042 0.037 0.038 0.030
Midspan

1:1.5:1 0.013 0.023 0.029 0.028 0.020 0.053 0.135 0.245 0.135 0.053 0.028 0.029 0.023
1:2:1 0.010 0.019 0.023 0.023 0.016 0.063 0.167 0.313 0.167 0.063 0.023 0.023 0.019

© Jonathan Haynes version 3.0 Page 54 of 239

32900 Bridge Engineering
MSc Structural Engineering

2.2.3 Analysis of Moving Actions on Slabs using Influence Surfaces

Westergaard was the first to provide a working solution to the problem of analysing plates for
moving loads. His influence surface is split into four quarters, each addressing major axis or minor
axis bending or torsion.
Pucher (1973) provides a complete solution for forces in an elastic plate, using a series of
scalable graphs relating to pinned or fixed edge conditions.
Since we know that a reinforced concrete bridge deck slab will respond to loading as a function
of the deck transverse stiffness and beam longitudinal stiffness but the exact stiffness cannot be
known with certainty, then it is wise to assess the forces for extremes of fixity (pinned and fixed
edges) and reinforce the deck accordingly.
Deck bending moments using influence surfaces are calculated in a similar way to beam
bending moments from influence lines. Wheel areas are positioned to obtain the most onerous
F
effect, and coefficients are read from the chart to use in the equation, M =
8π
∑k j
where F is

the wheel load and kj are the coefficients. However, when using Pucher’s charts it is best to
spread the wheel load to the deck reinforcement (or neutral axis if the deck is prestressed). This is
a better representation of the actual loaded area of the deck and it is therefore customary to
split the loaded area into at least quarters to obtain the influence coefficients.

w
2
1
The contact pressure at the reinforcement is now,
tyre 1
1 F
t surfacing q=
(w + t + 2h)2
concrete
h and the deck bending moment is now,
q deck
1
reinforcement M=
8π
∑ q. A.k j

L=w+t+2h

Seven of Pucher’s charts are included in this handbook;

1 & 2 cover major and minor axis sagging moments at the centre of a simply supported plate.
3 & 4 cover major and minor axis sagging moments at the centre of a fixed plate.
5 covers major axis hogging moment at the edge of a fixed plate.
6 & 7 cover major axis hogging moment at the edge of a cantilever plate.

M2 Charts 3, 4 and 5 M4
Charts 1 and 2
idealise the deck idealise the deck
between beams as a between beams as an
simply supported encastre plate along
plate along the two the two long edges.
long edges.
direction of travel

M1 M5 M3

Charts 6 (midspan) and 7 (abutment) idealise a cantilever

deck as a rigid cantilever.

M6 or M7
BEAM & SLAB BRIDGE DECK

The use of influence surfaces is best demonstrated by an example.

© Jonathan Haynes version 3.0 Page 55 of 239

32900 Bridge Engineering
MSc Structural Engineering

300kN 300kN
A 250mm thick reinforced
concrete bridge deck slab
spans across three beams,
spaced at 2.50m centres. The
0.150m deck reinforcement is 20mm
0.190m diameter and the cover is
50mm.
Obtain the maximum sagging
bending moment in the slab
when subject to an EC1 LM1
wheel train.

2.50m

1. To set the chart scale for all measurements,

pitch a scale rule so that the distance across the
chart is the distance between the deck beams.
2. The length of contact at the reinforcement is,
L = w + t + 2h =
0.93m L = 400 + 150 + 2 × 190 = 930mm
3. The wheel loads must be positioned on the
influence surface to obtain the most onerous
a b effect. It makes sense to begin by locating a
wheel on the origin point, as the coefficient is
0.93m

greatest at that point. Other wheels are located by

scaling 1.20m along the y axis and 2.0m along
the x axis.
c d 4. Break each of these contact areas into four
1.20m

equal areas of side 930/2 = 465mm, scale this

e f onto the chart. Draw a cross inside each quadrant
to locate the centre, and read off the coefficient.
5. Sum the moment coefficients in a table;
Quadrant kj
a 1.8
g h b 1.8
2.00m c 2.67
d 2.67
e 4.0
f 4.0
g 4.0
h 4.0
Σ 24.94
2.50m 8. The contact pressure at the reinforcement is,
F 300
2.00m
25 q= = = 347kN / m 2
L2 0.93 2
web of left beam

web of right beam

20
9. The deck bending moment is,
1.20m
1
0.465m
15
M=
8π
∑ q. A.k j
10
347 × 0.4652 × 24.94
5 M1 = = 74.5kNm
8×π

The process is then repeated for other wheel positions until the most onerous sagging moment is
identified. Charts 2-7 may be used to obtain the hogging moments in the slab.

© Jonathan Haynes version 3.0 Page 56 of 239

32900 Bridge Engineering
MSc Structural Engineering

Influence surface for major axis sagging moment mx at the centre of a plate simply supported on two long edges. Pucher 1

Influence surface for minor axis sagging moment my at the centre of a plate simply supported on two long edges. Pucher 2

© Jonathan Haynes version 3.0 Page 57 of 239

32900 Bridge Engineering
MSc Structural Engineering

Influence surface for major axis sagging moment mx at the centre of a plate encastre on two long edges. Pucher 3

Influence surface for minor axis sagging moment my at the centre of a plate encastre on two long edges. Pucher 4

Influence surface for major axis hogging moment mx at the support of a plate encastre on two long edges. Pucher 5

© Jonathan Haynes version 3.0 Page 58 of 239

 32900 Bridge Engineering
MSc Structural Engineering

Influence surface for major axis hogging moment mx of a plate cantilevering from a long encastre edge (-1.3L< y <+1.3L). Pucher 6

Influence surface for major axis hogging moment mx of a plate cantilevering from a long encastre edge (0< y <+2.6L). Pucher 7

© Jonathan Haynes version 3.0 Page 59 of 239

32900 Bridge Engineering
MSc Structural Engineering

2.3 Global Analysis Models of Bridge Decks

Systems of analysis which consider the whole bridge deck as a single model are currently
standard practice. This method first emerged with the development of 3D stiffness analysis
software in the late 1970’s but only became widely available in the mid-1980’s, when hardware
became commonplace which could solve medium span decks in reasonable periods of time.

2.3.1 Grillage Analysis

The first stiffness method model for bridge decks was the grillage analysis. This is essentially a 2D
structural model subject to out-of-plane loading (in the third dimension). A grillage model
represents all elements of the deck as rigidly connected beams. In a beam and slab deck it is
easy to visualise the downstand beams (beam elements positioned at the downstand beam
centroids). The slab which spans across the downstand beams must also be idealised as beam
elements, often by making one beam element per metre width of slab but this depends upon
the deck form.

Real deck with slab and down-stand beams Grillage of orthogonal beams

Although it is possible to create a series of analysis models in which the vehicle loading is
progressively moved across the bridge to create an envelope of design forces, it is adequate to
use an influence line to position the vehicle in the most onerous location for each design force.
2D grillage models are crude representations of reality but are simple to construct and interpret,
and provide answers which agree well with more rigorous methods of analysis.
Uniform traffic loads are generally converted to point loads which are applied to the model.
Slab decks
In slab decks there is no dominant line of longitudinal stiffness, so there will be significant torsion
as well as bending in the slab. The grillage bending elements should therefore be approximately
equal in stiffness in both directions to represent an isotropic slab. Spans are limited to about 15m.
The deck will behave like a plate.
A beam element should be positioned about 0.3h from the side edge of the slab.
Transverse elements should be perpendicular to longitudinal elements, even in skew decks.

0.3h
LE=0.85L

LE=0.3L

beam elements
at level of slab
neutral axis
3h ≤ 0.25LE

© Jonathan Haynes version 3.0 Page 60 of 239

 32900 Bridge Engineering
MSc Structural Engineering

The mesh of beam elements must be fine enough so that a wheel load dispersed through the
surfacing (at 2:1) and slab concrete (at 1:1) to the slab neutral axis can lie on one beam
element. If this is adhered to then the grillage should give a reasonable representation of local
bending in the slab. However, there is little advantage in positioning elements closer than 3h
centres, or 0.25LE (this may be decreased close to support regions where peak hogging moments
occur).
The grillage properties should represent reality – reinforced concrete elements must crack and
will require cracked section properties, whereas prestressed elements should use uncracked
properties. In solid reinforced concrete isotropic decks (approximately equal reinforcement in
both directions) then the use of uncracked properties will make little difference in the distribution
of force.
Skew decks behave globally like a plate. The physical effects of this are that shear and high
reactions are localised at the obtuse corners and the acute corners can suffer uplift. This can
lead to fatigue problems, but can be ameliorated by the use of soft bearings (rubber) which
allow some redistribution of peak stresses but at the expense of increased sagging moments in
the deck.
≥ 0.05L
≥ 0.05L

≥ 0.09L

≥ 0.09L
L

≥ 0.033L

< 20° ≥ 20°

For skew angles below 20 degrees the grillage can include skew transverse elements (as this can
make reinforcement detailing easier), otherwise transverse elements musts be orthogonal.
Beam and slab decks
These are orthotropic decks in which there are identifiable lines of longitudinal stiffness. The deck
will behave like a series of beams with a flexible cross slab. Consequently, there will be little
torsion in the slab (with the exception of asymmetrically loaded box girders).
Beam elements should be positioned at the centroid of the longitudinal beams, with transverse
beam elements to represent the deck slab at the same centroidal level. Steel beams should be
represented by single beam elements. Contiguous concrete beams should be represented by
one beam element per two real beams. Wide flange concrete beams should be represented by
two beam elements.

τ T
τ
beam elements at level of:
• steel beam centroid
τ
• contiguous beam centroid
• full depth wide flange beam centroid
• full box girder centroid

© Jonathan Haynes version 3.0 Page 61 of 239

32900 Bridge Engineering
MSc Structural Engineering

Transverse elements should be spaced at 0.125-0.25LE and have section properties to represent a
cracked reinforced concrete slab.
Where there is a support diaphragm, a beam element should be included, particularly if it is
skew. Intermediate compression flange cross bracing which only connects two beams may be
ignored in the analysis but if it connects more than two beams, there may be a line of transverse
stiffness which will require more rigorous analysis.
Beam and slab decks often include edge cantilevers which deflect significantly. This can be
reduced by including an edge beam or upstand.
Composite systems require care when assessing section properties. A composite deck is formed
from separate beams and slab, which are only attached by shear connectors at the beam
flange. When the beam is forced to deflect by traffic loading, the slab directly above the beam
must deflect by an equal amount. However, the slab remote from the beam will not deflect as
much – the plane section of the slab does not remain plane during bending, and is said to lag
behind. This means that the effective width of the compression flange (slab) changes with the
deformation of the beam in a phenomena known as shear lag. The shape of the beam
deformation is conveniently described by the effective length (the distance between points of
contraflexure).
The compression flange width cannot exceed the beam spacing, b. Where the spacing of
beams exceeds 0.167LE or 0.083LE at a cantilever, then the effects of shear lag will reduce the
width of the composite flange. This is worse when the loading is not uniformly distributed (traffic
wheel loads), in which case the effective flange width, be should be limited to 0.1LE.
In addition, the effective length of continuous beams will vary along their length and depend
upon the fixity of supports. It is customary to simplify this by restricting consideration to sagging
regions, where LE=0.85L and hogging regions, where LE=0.3L.
there is longitudinal shear stress between
effective flange width in continuous beam

the effective slab flanges of each beam

bending moment diagram for continuous beam

be ≈ 0.1LE LE=0.85L

be ≈ 0.2 LE LE=0.3L

be
slab section does not remain a
rectangle during bending b

Composite sections should be converted to a single material element using the modular ratio.
The short term modulus of concrete should be used.
Multi-cellular decks
The cell structure can be idealised as a series of I beams and represented as longitudinal beam
elements with transverse elements which represent the running slab of the cellular structure.
Multi-cellular decks are torsionally stiff, which is achieved by the free flow of shear stresses around
the cell walls resisting the torsional effects of asymmetric loading. There will also be significant
shear stresses between the webs and flanges of the cells which are due to overall bending of the
deck.

© Jonathan Haynes version 3.0 Page 62 of 239

32900 Bridge Engineering
MSc Structural Engineering

τ τ
τ T τ
τ
the cellular deck may be split into I and T
beams and the grillage beam elements set at
the full deck centroid

The shape of the cells can have a significant effect upon the deflection of the deck. Circular cell
structures of equal area to square cell structures may be twice as stiff, this is because the
vierendeel bending effect is most pronounced in the thinner square cell walls.
Foundation stiffness
Most analysis models will assume the supports are rigidly positioned in space – the classical
assumptions of pinned or fixed supports. This is clearly grossly inaccurate as all soil and rock is
compressible and will be exacerbated by the transient nature of vehicle loadings and the
flexibility of superstructure connections. It is therefore worthwhile exploring the effects of soil
stiffness by including a crude representation of soil as an elastic spring and undertaking a series
of increasingly complex analyses which model:
• the structure with pinned or fixed supports
• the structure with differential bearing compression
• the structure supported on soil springs
The design may not adopt design forces from the second and third analyses but the trend of
changing force distribution can inform decisions about whether to doubly reinforce concrete
elements or extend the regions of compression restraint stays in steel beams.

a) Real Foundation A 2D analysis model of a spread foundation of width w and breadth

b (into the page) subject to vertical, horizontal and bending forces,
may be modelled with three springs as shown.
0.25 0.75
distance between spring couple, l = 0.82b w
h
bh 3
foundation element stiffness, I = cross sectional area, A = bh
12
w
1.25G(bw)
0.5

vertical stiffness, kv =
1 −ν
b) Structural Model horizontal stiffness, kh = 2G(1 + ν )(bw)
0.5

E
shear modulus, G =
kh 2(1 + ν )
element represents bridge pier,
kv with release if to be pinned base

Values for Poisson’s ratio for soil generally lie in the range 0.3-0.5. Material property values for soils
are related to shear or compression strength, suggested values are given in Section 8.4.

A functioning demonstration version of ACES6 grillage analysis software is available at:

http://www.aces-systems.com/downloads.html

Limited licence numbers of the SAM bridge modelling suite are available in Newton 155 & 156.

© Jonathan Haynes version 3.0 Page 63 of 239

32900 Bridge Engineering
MSc Structural Engineering

2.3.2 Finite Element Analysis of Bridge Decks

Engineering judgement must be exercised when using computer analysis output since
calculated elastic forces are often reduced in reality by redistribution, shakedown and
unexpected joint fixity. Conversely, elastic deformations are often exceeded in reality due to
creep, local plasticity and unexpected connection flexibility. The limitations of representing a
bridge deck using a two dimensional grillage are obvious. Deck forms which have significant
depth, such as box girders, or incorporate flexible supports, such as cable stays, deserve a more
rigorous analysis.
It is possible to use 3D (space frame) analysis to accurately represent box girder decks and
trusses, even with curved alignments, using only axial or bending elements. However, as stiffness
representations of plate theory has developed, so too has the availability of (relatively) user
friendly finite element software.
Finite Element analysis is merely stiffness analysis with a more versatile array of element types than
plane or space frame analysis. The technique seeks to solve an elastic model for joint
deformations (rotations and displacements) and from these calculate the internal forces from
the standard stiffness matrix.
Complex elements such as cables, multi-cell box girders and moving joints can be modelled
using the more advanced finite elements and restraint systems. However, most bridge structure
can be represented by combinations of the more mundane axial bar, beam and plate
elements.
It is important to note that FE models are not continuums but a series of nodally connected
triangles or quadrilaterals. Up to a limit, accuracy is improved by increasing the number of
elements (or mesh density) but at the expense of computational demand. The deformation
solutions produced are approximations, and as such any stress plots are smoothed by post-
processing to give the illusion of gradual changes across the model.
Standard Finite Elements and their uses
Axial bar or spar elements can be used to represent ties and struts where only shear and axial
forces are to be transmitted. They can be used to form a pinned rigid link if the Young’s modulus
of elasticity is set very high.
Beam elements are the only elements usually available in 2D and 3D linear elastic frame analysis
programs. A beam element is essentially a one dimensional bar which can transmit axial, shear
and bending in three dimensions (bending about the longitudinal axis is strictly torsion).

N
V

V
N
Axial bar M
Beam

Modelling with line elements (beams and spars) affords the benefit that they do not suffer from
mesh refinement when used for general elastic problems, and they provide moment and force
output at nodes which is ideal for assessment and design. They require section properties such as
cross sectional area, second moment of area and appropriate material properties such as
Young’s modulus and Poisson’s ratio. The length of the element is simply controlled by its
connectivity to nodes (the only exception being if material or geometric nonlinearity is required).
Plane stress elements can be used to investigate the behaviour of 2D structure subject to in-
plane forces, such as box girder webs. The finite element can only be stretched or squashed in its
plane.
Plate bending elements should be capable of distortion due to in-plane and out-of-plane forces.
This is because they will be used to build up three dimensional models of structures, in which
plates will be connected at nodes in different planes. In-plane forces in one plane become out-

© Laurence Weekes & Jonathan Haynes version 2.0 Page 64 of 239

32900 Bridge Engineering
MSc Structural Engineering

of-plane forces in another plane. The finite element can be stretched, squashed and bent out-
of-plane.

Plate bending Plane stress

Shell elements are essentially a combination of plane stress and plate elements, so they can
model both membrane and out-of-plane bending. They tend to be employed for curved
surfaces or in general where both of these actions are desirable. For example arch systems (such
as jack arches) or deck profile changes, these elements may be desirable. Sections properties
such as thickness are required, from which properties such as second moment of area are
automatically derived. Many FE codes implement facilities to incorporate elastic foundation
stiffnesses and allow input of ratios to ramp up/down section properties. This is particularly useful
when considering profiled sheeting for example.
When modelling with line elements the stiffness formulation is such that the number of elements
tends not to affect the global structural behaviour but refinement may be exercised to provide
nodes at locations where forces are required for design and/or assessment.
For 2D elements, the formulation is such that mesh refinement is usually required. The emphasis
for 2D element output is stress rather than stress resultants (forces). However, it is possible to
extract shell section forces whereby a (node weighted) moment per unit width is output at the
centroid of any given element. This should be used for activities such as designing reinforcement
for a deck modelled using plates.
Hence it is usual to conduct a mesh sensitivity study to determine a mesh density which is fit for
purpose in terms of sufficient accuracy. This normally forms part of any verification and validation
work during the modelling process. As a rule of thumb, when modelling a wall or slab it is
advisable to provide a minimum of four elements per span to capture bending effects which are
of acceptable accuracy. If capturing natural frequencies through a modal (Eigen value)
analysis then more may be required depending on the number of modes required.
Solid elements are simply a 3D extension of 2D elements and are generally termed brick
elements. They have no bending formulation associated within their interior, and rely on
translational fixity in the three Cartesian orthogonal directions. They also suffer from mesh
refinement but do not require section properties to be input. At the outset it may appear that this
is the answer to modelling anything, however more caution is required. When capturing bending
and shearing effects, sufficient elements are required over the depth of the structure to capture
sufficiently accurate shearing stresses. This has in the past been the culprit of physical failures
such as the Sleipner A oil rig (Harris et al, 2002). Again in terms of output it is usual to examine
stress as forces are usually harder to post process.
Common problems encountered in FE models
As with any analysis or design, whether by hand or computer, the engineer must have
confidence that the assumptions upon which the analysis is based are sufficient for the structural
problem at hand. Therefore verification and validation are required, especially in the case of
computer modelling where the structural system is complex. In the first instance there is much
data input and this can lead to simple verification errors. Typically the engineer should be
answering the following in a structured debugging regime:
• Are the units used consistent throughout all of the modelling? If units for materials are in N
and m, then the loading must be applied in N and the lengths must be in m.
• Have boundary conditions been applied correctly? This is usually addressed by checking the
reactions for equilibrium compliance.
Loading can be applied either directly through forces or pressures, or by specifying a mass and
accelerations. Again, reactions can be checked for equilibrium. In fact for the latter method, it is
© Laurence Weekes & Jonathan Haynes version 2.0 Page 65 of 239
32900 Bridge Engineering
MSc Structural Engineering

usual to apply accelerations in all three orthogonal Cartesian directions as a check to ensure
that all parts of the model are connected as desired.
Care should be exercised with regards stiffness distribution in the model. Ill-conditioning of the
stiffness matrix can arise as a result of extreme disparate distributions of stiffness. Solvers that
perform a Gauss elimination will be required to manipulate very small and very large numbers
which can lead to significant errors.
When using 2D or 3D elements, mesh sensitivity studies on small sub-models are useful to home in
on a mesh which is sensible (not overly refined). For elements which exhibit nonlinearity such as
cracking or contact elements, or when plasticity is implemented through nonlinear stress strain
curves, over refinement can lead to non-convergence. It is best to use quadrilateral or cuboid
elements rather than triangles or tetrahedron’s for reasons of true element formulation.
Sometimes it is unavoidable in complex shapes, so if the aim is to extract forces from the model
for post-processing, then as a rule 95% or more of the model should comprise cuboid or
quadrilateral elements.
This list of issues is not exhaustive. Checking models are an important tool for debugging and
providing confidence in the FE model results. However the engineer needs to be able to interpret
the assumptions in the models and therefore develop an understanding of the reasons for any
discrepancy in the models, thereby justifying the FE analysis results.
Modelling bearings
Bridge bearings can be modelled in a variety of ways. The simplest representation is to release
the degree of freedom (if the FE code allows) which connects the deck to its support (assign zero
horizontal stiffness). This representation should allow the global model to behave as desired. If
rotation is required to a certain degree then either the rotational degree of freedom can be
released (and the amount of displacement monitored), or alternatively, rotational springs can be
employed to model stiffness. These can be modelled as nonlinear if the bearings have limitations
in their rotational movement. If a more detailed model is required (i.e. when the performance of
bearings is to be investigated), then the bearings can be modelled with systems of springs,
laminar solids (for elastomeric bearings), or specialised joint elements.

© Laurence Weekes & Jonathan Haynes version 2.0 Page 66 of 239

32900 Bridge Engineering
MSc Structural Engineering

Bridge Deck Loading & Analysis - Conclusion

Things to do now :
1. Review the Powerpoint lectures on Blackboard.
2. Attempt the Self Assessment Exercises below. Then check your answers against the
solutions on Blackboard.
3. Have a rest.

4. Prepare your group submission for Assignment 3.

Self Assessment Exercises

A two lane, two span highway bridge is formed from a solid reinforced concrete slab and two
plate girder beams. A pedestrian footpath is provided in both directions. There is no central
reservation. The bridge is supported on a central pier and retaining wall abutments.
5.4m 1.0m open steel
parapets
0.35m

4.5m

300mm thick deck slab tapering to 20.0m

150mm at edge. 100mm asphalt
surfacing over waterproofing layer, 5.0m
with 175mm concrete footpath.
Deck altitude 65m AoD, 50km 20.0m G.L.
from the shoreline in Salford.

steel plate girders reinforced concrete pier

1.0m deep x 0.3m wide 6.0m x 5.0m high x 0.3m

1. Calculate the self weight of the bridge deck.

2. Calculate the variable action traffic loads.
3. Calculate the variable action wind loads.
4. Calculate the variable action thermal loads.
5. Calculate the design combination loadings.
6. Using influence lines, estimate the design forces in a beam.
7. Using influence surfaces, estimate the design forces in the deck slab.
8. Using suitable software, create a grillage analysis model of the entire bridge deck and
compare the design forces to those estimated previously.

Further Reading
Hambly, E.C. (1991). Bridge Deck Behaviour. 2nd Edition. London: E&FN Spon.
Hambly, E. (1994). Structural Analysis by Example. Birkhamstead: Archimedes.
Harris, P.R., MacLeod, I.A., Bond, A.J., Gardner, P.J., Harvey, W.J., Knowles, N.C. & Beale, B.S.
(2002). The use of computers for engineering calculations. London: The Institution of Structural
Engineers: London.
O'Brien, E. & Keogh, D.L. (1999). Bridge Deck Analysis. London: Spon.
O’Connor, C. (1971). Design of bridge superstructures. New York: Wiley.
O’Connor, C. & Shaw, P.A. (2000). Bridge Loads. London: Spon.
Pucher, A. & Juhl, H. (1973). Influence surfaces of elastic plates. New York: Springer-Verlag.
Rourke & Young (2002). Formulas for Stress and Strain. 7th Edition. London: McGraw-Hill.
Steel Construction Institute. (1998). Steel Designers Manual, 5th Edition. Oxford: Blackwell.

© Jonathan Haynes version 3.0 Page 67 of 239

32900 Bridge Engineering
MSc Structural Engineering

3. Structural Element Design

This section will address the code design rules for structural elements such as deck slabs, plate
girders, post-tensioned beams, abutments, piers and foundations. Though it is not intended to
be exhaustive, the rules presented will provide sufficient depth to ensure strength and stiffness
are adequately catered for.
You should also review:
3 Bearings, parapets and services.ppt
5 Steel Elementss.ppt
6 Concrete Elements.ppt
7 Stress and box girder bridges.ppt
9b SSI and abutments.ppt on Blackboard
Principal References:
Brooker, O. et al Concise Eurocode 2 for Bridges. CCIP-038.
SCI. Bridge design to the Eurocodes. Simplified rules for use in student projects.
Benaim, R. The design of prestressed concrete bridges. Concepts and principles.

Many bridge designers perceive structural element design as the key step in the process.
However, a bridge which has been conceived with good proportions and scale should pass
most code clause checks by default.
Before embarking upon lengthy code checking it is worth spending time identifying opportunities
for repetition and standardisation of structural element sizes.
All structural elements must provide sufficient stiffness (limited deflection / vibration) and strength
(resistance to the four internal forces).
There is insufficient time in this module to address the detailed design of all possible structural
forms and materials, so effort will be concentrated upon reinforced concrete slabs, post-
tensioned concrete beams, steel plate girders and trusses, and reinforced concrete abutments,
piers and foundations.
There is extensive literature available to guide students in the detailed design of steel, reinforced
concrete, post-tensioned concrete and composite construction bridge elements. Information on
secondary items such as parapets and bearings is best sources directly from manufacturers of
proprietary products.

© Jonathan Haynes version 3.0 Page 68 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.1 Reinforced Concrete Decks

Concrete is a construction material formed from coarse and fine aggregates, cement and
water. The wet concrete is poured into a shutter, where a chemical reaction (hydration) takes
place and the concrete sets solid. Concrete is highly alkaline when wet, and causes severe burns
when in prolonged contact with skin.
Material Safety Factors
For concrete, γc = 1.5, for reinforcing steel, γs = 1.15
Concrete Strength
Eurocodes primarily specify 28 day concrete cylinder strength, fck. Concrete strength may also be
specified using compressive cube strength, fck,cube.
fck,cube = 1.2 x fck and concrete strength is specified in the form C25/30 (normal weight
concrete/cylinder strength/cube strength). Note that lightweight concrete is designated LC.
Reinforcement Strength
UK reinforcement production is now based upon high yield ribbed round bar steel reinforcement
in three ductility classes (A, B & C). Where redistribution of 20% or more, or plastic analysis is used,
grade A is not permissible. Generally, fyk = 500 N/mm2. Class A reinforcement MUST NOT be used in
UK bridges.
Material Properties of Concrete
0. 3
 f + 8
Ecm = 22  ck  kN / mm 2
 10 
The (Young’s) modulus of elasticity of reinforcing steel; E = 200 kN/mm2
The coefficient of thermal expansion; α = 10x10-6 /0C
Poisson’s ratio; ν = 0.2 (but zero for cracked sections)
Concrete cover
The minimum thickness of concrete around the reinforcement defines the fire resistance and
durability of the element. A 120 year design life is standard for bridges. The cover depends upon
environmental exposure (dry / wet / exposed to sea water etc.).
Cover is made up of a number of components:
The axis dimension (Caxis) is used in fire resistance assessments, it is the distance from
Cmin ∆c
the outer concrete face to the centroid of the main reinforcement.
The nominal cover (Cnom) to will be quoted on drawings, it is the distance from the
outer concrete face to the outermost point of the reinforcement (often shear links). Cnom
The nominal cover is made up of two sub-dimensions, the minimum cover (Cmin) and
the deviation (∆c). ∆c = 10mm for normal construction but may be reduced to 5mm
when a QA system is used to guarantee actual cover is compliant (well controlled site Caxiz
construction). If the actual cover can be checked and inadequate construction
rejected, then ∆c may be set to zero but this is only likely to happen in a precasting
factory.
BS8500 and The National Structural Concrete Specification deal extensively with the specification
of prescribed, standard and proprietary mixes. Minimum cover in foundations shall be 40mm
when cast on blinding concrete or 65mm otherwise. Cover should not be less than the bar
diameter.
Concrete surfaces within 10m horizontally or 5m vertically of carriageway surfaces affected by
de-icing salt must be specified for exposure condition XD3, XF2 or XF4.

© Jonathan Haynes version 3.0 Page 69 of 239

 Normal weight concrete strength (cylinder / cube) , maximum water cement ratio
Cement and minimum cement content (kg/m3) for nominal concrete cover.
Exposure conditions
type
15+∆c 25+∆c 30+∆c 35+∆c 40+∆c 45+∆c 50+∆c 60+∆c

C20/25
XC1 Dry or permanently wet OPC

© Jonathan Haynes
0.7 , 240
C25/30
XC2 Wet, rarely dry OPC
0.65,260
32900 Bridge Engineering
MSc Structural Engineering

corrosion
XC3 Moderate humidity OPC
C40/50 C35/45 C30/37
0.45,340 0.50,320 0.55,300
XC4 Cyclic wet and dry OPC

Carbonation induced
C45/55 C40/50 C35/45 C32/40
XD1 Moderate humidity OPC
0.40,380 0.45,360 0.50,340 0.55,320
C35/45 C32/40 C28/35
XD2 Wet, rarely dry SRPC
0.45,360 0.50,340 0.55,320

version 3.0
chloride
Airborne

corrosion
C35/45 C28/35
XD3 Cyclic wet and dry SRPC
0.40,380 0.50,340
C35/45 C32/40 C28/35
XS1 Airborne salts, no contact IIB-V
0.40,380 0.45,360 0.55,320
C35/45 C32/40 C28/35
XS2 Wet, rarely dry OPC
0.45,360 0.50,340 0.55,320

Seawater
C35/45 C28/35
XS3 Tidal, splash, spray zones IIB-V

induced corrosion
0.40,380 0.50,340
Concrete specification for 120 year life using 20mm maximum aggregate.

XF1 Wet no de-icing salt IIB-V C35/45 C32/40 C28/35

0.40,380 0.45,360 0.55,320
XF2 Wet + de-icing salt IIB-V C35/45 C28/35
0.40,380 0.50,340
XF3 Saturation no de-icing salt OPC IIB-V is OPC with 21-35% fly ash *C40/50
0.40,380

Page 70 of 239
XF4 Saturation + de-icing salt IIB-V * requires freeze thaw resistant aggregates *C40/50 *C40/50

Freeze / thaw attack

0.40,380 0.40,380
32900 Bridge Engineering
MSc Structural Engineering

3.1.1 Bending
x
, the redistribution ratio must not exceed 30%, δ = 0.4 + d ≤ 0.7
u
In solid slabs δ
Rectangular section Strain block Stress block and forces
b
d2 εc fcd

As2 εsc 0.8x Fsc

x Fc

h d neutral axis z

As εs Fst

M Ed
K=
bd 2 f ck

K ' = 0.6δ − 0.18δ 2 − 0.21 ≤ 1.0

K ≤ K' K > K'

no compression reinforcement compression reinforcement required

z=
d
2
[
1 + 1 − 3.53K ≤ 0.95d ] z=
d
2
[
1 + 1 − 3.53K ' ]
x = 2.5(d − z )
f yk  x − d2 
f yd = f sc = 700  ≤ f yd
γs  x 
( K − K ' ) f ck bd 2
As 2 =
f sc (d − d 2 )
M Ed K ' f ck bd 2 As 2 f sc
As = As = +
f yd z f yd z f yd

for flanged beams check neutral axis is within flange, x ≤ 1.25h f

check minimum reinforcement

As 0.26 f ctm 2
≥ ≥ 0.0013bd where, f ctm = 0.3 f ck 3
bd f yk

check maximum reinforcement

As
≤ 0.04
Ac

© Jonathan Haynes version 3.0 Page 71 of 239

32900 Bridge Engineering
MSc Structural Engineering

EC2 singly reinforced beam design chart, f y =500N/mm 2

0.25

C25/30 C28/35 C32/40 C40/50

no redistribution k ≤0.21
0.20

0.15
M/bd2fck

30% redistribution k ≤0.122

0.10

0.05

0.00
0.00 0.50 1.00 1.50 2.00 2.50 3.00

100A s /bd

Bending reinforcement for singly reinforced slabs and beams.

EC2 doubly reinforced beam design chart, f y =500N/mm 2

0.45
compression reinforcement tension reinforcement

0.40
δ =1.0 δ =0.7 C25/30 C28/35 C32/40

0.35

0.30

0.25
M/bd fck
2

0.20 no redistribution k ≤k' =0.21 INSTRUCTIONS

1. choose f ck
2. calculate k
0.15
3. choose redistribution

30% redistribution k ≤k'=0.122 30% redistribution, δ=0.7

0.10 no redistribution, δ=1.0

4. if k ≤ k ' read 100A s /bd

read As2 read As
0.05
5. if k > k' read 100A s /bd
read As only and 100A s2 /bd

0.00
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00

100A s /bd

Bending reinforcement for doubly reinforced slabs and beams.

© Jonathan Haynes version 3.0 Page 72 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.1.2 Shear
Although floor slabs are usually designed without shear reinforcement, all other elements must
contain shear reinforcement (usually in the form of links or stirrups). Ultimate shear is
characterised by highly undesirable brittle failure. Shear is usually checked at stress level. vEd is
the applied design shear stress at d from the support, vRd is the design shear resistance.
For sections without shear reinforcement (slabs), shear is acceptable if,
VEd 0.18 1 3 1
v Ed = ≤ v Rd ,max = k (100 ρ1 f ck ) 3 ≥ 0.035k 2 f ck 2
bz γc
As 200
where ρ1 = ≤ 0.02 and k = 1 + ≤2
bd d

For all other elements, shear reinforcement must be designed using the variable strut inclination
method, as follows:
VEd cot θ
v Ed = ≤ v Rd ,max = vf cd
bz 1 + cot 2 θ

 f ck  α cc f ck
where, ν = 0.61 − and f = αcc = 0.85 for bending and axial, 1.0 for shear
250 
cd
 γc
1.0 ≤ cotθ ≤ 2.5 (if shear is adequate using cotθ ≤ 2.5 no further checks are required).

Where shear reinforcement is required the variable strut inclination method must be used,
Asw γ sVEd 0.08b fck
≥ ≥
s zf ywk cot θ f ywk
where s is the link spacing and fywk is the shear link strength.
Loading within d of the support will
not contribute to the shear force

av the variable strut

inclination method permits
22o< θ <45o q

compression
in steel
tension

d z=0.95d
θ
Tension in steel
compression in
VEd concrete

Shear enhancement close to supports may be adopted by designing for the shear force d from
the face of the support.

© Jonathan Haynes version 3.0 Page 73 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.2 Steelwork Elements

Design rules for steel bridges are contained in BS EN 1993-2 and covers all steel bridges and steel
parts of composite bridges (see also, PD 6695-2 Design of steel bridges). It references clauses
given in BS EN 1993-1 General rules and rules for buildings. Under Eurocode 3 the major axis is Y-Y,
the minor axis is Z-Z, leaving X-X to run along the longitudinal axis of the element.

3.2.1 Plate Girders

b
Z a
tf

Y Y
df=h-tf

h ho ts ho
hw tw d
r ecc ≤5tw

Z full depth load a Intermediate transverse

bearing stiffener 1≤ <3 stiffener is required to
is required at
hw prevent shear buckling
each support ho
h the load bearing stiffener design ≤ 10.5
ho ≤ w must allow for construction ts
10 tolerance and thermal expansion
Member strength checks are based upon plastic stress distributions. However, steel sections are
formed from thin elements, which may be quite slender. If local buckling sets in before yield
stress is reached the section will not achieve plastic capacities, it is therefore necessary to check
section classification. Sections which can reach plastic capacities are designed using plastic
section properties, otherwise elastic section properties are appropriate.
Section Classification
There are two elements of any steel section; an outstand compression element usually the part
of the flange which extends from the web, and the internal compression element usually the
part of the web between root fillets or welds.
c
Ratio’s for are calculated for each element and must be less than the limit value.
t
Classification limits Class 2 Class 3
Flange element ≤ 10ε ≤ 14ε
UB or UC sections in
pure bending Web element ≤ 83ε ≤ 124ε h
UB or UC sections in Flange element ≤ 10ε ≤ 14ε
pure compression Web element ≤ 38ε ≤ 42ε
Hollow sections in pure compression ≤ 70ε2 ≤ 90ε2
t
Angle sections in pure compression ≤ 15ε or (b+h)/2t ≤ 11.5ε
c b − t w − (2 × r ) c d c h
Flange element, = Web element, = Hollow section, =
t 2×tf t tw t t

© Jonathan Haynes version 3.0 Page 74 of 239

32900 Bridge Engineering
MSc Structural Engineering

Capacities are based upon grade S235 steel, since higher grades are more commonly used for
235
bridges, section classification limits are adjusted using the factor, ε= where fy is the design
fy
strength of the steel (eg. 355 N/mm2).
Steel Grades
EN 10 025 specifies steel strengths (or grades) at yield and ultimate levels, these depend upon
the steel thickness.
Thickness S 275 S 355
t (mm) Yield strength Ultimate strength Yield strength Ultimate strength
fy (N/mm2) fu (N/mm2) fy (N/mm2) fu (N/mm2)
16 275 410 355 470
40 265 410 345 470
63 255 410 335 470
80 245 410 325 470
Higher strength is achieved at the expense of ductility, and to account for the greater probability
of inclusions in thicker steel, subgrades are specified by consideration of toughness (strength,
ductility and weldability) which is tested using the Charpy Impact test (an assessment of material
quality). Several subgrades are available, they relate to minimum operating temperature. BS EN
1991-10 Table 2.1 provides a range of subgrades available for bridges. Subgrade is selected by
calculating the reference temperature and identifying the subgrade which corresponds to a
maximum thickness greater than the steel element flange.
Maximum steel thickness (mm) Reference temperature, TEd (°C)
Sub-grade T (°C) Jmin (J) -20 -30 -40 -50
JR 20 27 20 15 15 10
J0 35 27 35 25 20 15
J2 50 27 50 40 35 25
K2, M, N 60 40 60 50 40 35
ML, NL 90 27 90 75 60 50
An example of a full specification for structural steel open sections: EN10 025 – 2 : S 355 JR
Alternatively, open sections produced by TATA in the UK may be specified using their trade mark
sections, which are CE mark (European Union, 2006 Construction Product Directive) compliant:
ADVANCE : S 355 JR
Specific rules apply to hollow sections. Tubes are manufactured by bending plates and welding
the longitudinal seam. Specification is much more important for tubes as they may be hot or
cold formed, which has a significant effect upon their properties. Hot formed tubes are annealed
after forming, which relieves stresses and permits a tight root radius to be used, thus hot formed
sections display more ductility and better compression characteristics but are more expensive.
Cold formed sections must have larger radius corners to prevent cracking inside the root, and
can brittle fracture if welded. All tubes are manufactured in S355 material.
An example of a full specification for structural steel hollow sections would be:
EN10 210 : S 355 JR for hot formed sections
EN10 219 : S 355 JR for cold formed sections
Alternatively, hollow sections produced by TATA in the UK may be specified:
CELCIUS : S 355 JR for hot formed sections
HYBOX : S 355 JR for cold formed sections
Thick elements subject to through thickness tension may need a special HiZed grade specified to
BS EN 10164 (see also PD 6695-1-10). Ultrasonic lamination checks are used to identify any
sulphide inclusions (which cause lamina tears). Three grades are available: Z15, Z25 and Z35 - the
number being the average percentage reduction in area under a standard tensile test.
Material Properties
The (Young’s) modulus of elasticity of steel; E = 210 kN/mm2

© Jonathan Haynes version 3.0 Page 75 of 239

32900 Bridge Engineering
MSc Structural Engineering

The coefficient of thermal expansion; α = 12x10-6 /0C

Poisson’s ratio; ν = 0.3
Material Safety Factor
Material partial safety factors (γM ) are used to account for the change of reliability with different
failure modes. For bridge design the material factors are:
Local failure (yielding) γm0 1.00
Member instability (buckling) γm1 1.10
Tension fracture γm2 1.25
Joint resistance γm2 1.25
Bolt friction SLS γm3,ser 1.10
Bolt friction ULS γm3 1.25
Joints in hollow section trusses γm5 1.10
Pins at serviceability limit state γm6,ser 1.00
Shear Capacity, Vb,Rd
The simplified model of steel beams assumes shear is resisted by the web alone.
hw ε
Shear buckling need only be checked if > 72 where η=1.0 but generally plate girders are
tw η
of sufficient depth to warrant the need for intermediate web stiffeners to prevent panel shear.
f ywhwt w
VEd ≤ Vb ,Rd = χ w where Vb,Rd is the shear capacity of the web alone.
3γ m1
λ w ≤ 0.83 χ w = 1.0
h 
2
a hw 0.83
kτ = 5.34 + 4 w  , when ≥1 S275 λ w = 0.83 ≤ λ w ≤ 1.08 χ w =
 a  hw 34.4t w kτ λw
h 
2
a hw 1.37
kτ = 4 + 5.34 w  , when <1 S355 λ w = λ w > 1.08 χ w = 0.7 + λ
 a  hw 30.3t w kτ w

Bending Capacity, Mc,Rd (full plastic moment capacity)

When the compression flange of the beam is fully laterally restrained by positive connection to a
stability system through a horizontal diaphragm (deck slab), then the section will develop full
plastic moment capacity.
W pl f y
M Ed ≤ M pl , Rd = for Class 1 and 2 sections, where Wpl is the plastic section modulus.
γM0
Wel f y
M Ed ≤ M el , Rd = for Class 3 sections, where Wel is the elastic section modulus.
γ M0
VEd A f d
If ≤ 0.5 and M Ed ≤ M f ,Rd = f y f no shear and moment interaction check is needed.
Vbw, Rd γ m0
Bending Capacity, Mb,Rd (buckling moment capacity)
When the compression flange of the beam is not fully restrained by positive connection to the
stability system, then the compression flange acts as a strut and buckles, becoming less stable.
The tension flange acts as a tie, therefore becoming more stable as it is loaded. As the two
flanges are connected together, their interaction is complicated and a compression flange local
buckling phenomena called lateral torsional buckling (LTB) develops. The section must be
designed for a reduced bending stress, leading to the buckling moment capacity.
As the compression flange attempts to displace sideways, there is a lateral force in the flange
which must be transferred into the floor diaphragm at points of intermediate lateral restraint.
χ LT W y f y
M Ed ≤ M b, Rd = where χLT is the LTB capacity reduction factor.
γ M1
© Jonathan Haynes version 3.0 Page 76 of 239
32900 Bridge Engineering
MSc Structural Engineering

For class 1 and 2 sections Wy is Wpl,y the Y-Y axis plastic section modulus.
For class 3 sections Wy is Wel,y the Y-Y axis elastic section modulus.
1 1
χ LT = ≤ 1.0 and ≤ 2
where λ LT , 0 = 0.4 and β = 0.75
2
φ LT + φ 2
LT − βλ LT
λ LT
and { ( ) 2
}
φ LT = 0.5 1 + α LT λ LT − λ LT ,0 + β λ LT where αLT is an imperfection factor
Wy f y
λ LT = however, calculation of Mcr is complex, alternatively the following may be used,
M cr
Section
S275 S355 λzf is calculated using the radius of gyration of the
Class
compression flange and one third of the web in
λ zf λ zf compression
2 λ LT = λ LT =
87 76
λ zf λ zf hw / 6
3 λ LT = λ LT =
100 88
LTB curves of χLT against λ LT may be plotted.
Buckling curve a b c d
Imperfection factor, αLT 0.21 0.34 0.49 0.76

Buckling curve
Aspect
ratio Rolled
General
sections
h
≤2 c b
b
h
>2 d c
b
Continuous Plate Girder BMD

compression flange at top compression flange at top

compression flange
at bottom
LE

* * * *
if compression flange is NOT *
restrained by bridge deck

if compression flange is restrained LE LE

by bridge deck
Plan on plate girders

deck provides horizontal

diaphragm action

bracing provides torsional restraint

to compression flange *

Effective Length of Plate Girders

© Jonathan Haynes version 3.0 Page 77 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.2.2 Columns and Bracing

Axial Flexural Buckling Compression Capacity, Nb,Rd (strut capacity)
Most axially compressed elements are proportioned such that flexural buckling (mid-length
lateral displacement) controls the capacity. For strut elements of pin-jointed frameworks (trusses)
this is the only check required.
χAf y
N Ed ≤ N b, Rd = where χ is a capacity reduction factor.
γ M1
Af y λ E
λ= = where λL = π is the limiting slenderness
N cr λL fy
Strut curves of χ against λ may be plotted to simplify the design process.
The maximum slenderness about either y-y or z-z must be used in design.
π 2 EI
Alternatively N cr = 2
may be used (Euler buckling load).
LE

{ (
φ = 0.5 1 + α λ − 0.2 + λ ) 2
} where α is an imperfection factor and χ=
1
2 2
≤ 1.0
φ + φ −λ
For rolled sections the following may be used to select buckling curves and axial buckling
imperfection factors.
Buckling curve a0 a b c d
Imperfection factor, α 0.13 0.21 0.34 0.49 0.76

Flexural buckling curve selection, rolled sections.

Buckling curve
Aspect Buckling
Material thickness
ratio axis S235, S275,
S460
S355, S420
t f ≤ 40mm y-y a a0
h z-z b a0
> 1.2 40mm < t f ≤ 100mm y-y b a
b z-z c a
t f ≤ 100mm y-y b a
h z-z c a
≤ 1. 2 t f > 100mm y-y d c
b z-z d c

Flexural buckling curve selection, other sections.

Buckling curve
Buckling
Section type Description
axis S235, S275,
S460
S355, S420
Hot finished any a a0
Hollow Section
Cold finished any c c

Welded box Weld < 0.5tf any b b

section Weld > 0.5tf any c c
Channel, Tee, any c c
plates and bar
Angle any b b

© Jonathan Haynes version 3.0 Page 78 of 239

32900 Bridge Engineering
MSc Structural Engineering

Design of Beam-Columns (full interaction check)

Steel elements subject to biaxial bending and compression exhibit complex structural behaviour.
Bending displacement is amplified (non-linear geometry) by applied axial loads, which
constitutes a second order moment effect. This is taken into account by using either enhanced
end moments or using appropriate effective lengths.
For class 1, 2 and 3 sections, there are two interactions to satisfy:
N E,d M y , Ed M z , Ed
+ k yy + k yz ≤ 1.0
N b , y , Rd M b, y , Rd M pl , z , Rd

N E ,d M y , Ed M z , Ed
+ k zy + k zz ≤ 1.0
N b , z , Rd M b , y , Rd M pl , z , Rd
Calculation of k factors is complex but may be conservatively taken as 1.0

Section classification for bending & compression.

For UB or UC sections
in Y-Y bending (only) Class 1 Class 2 Class 3
+ compression
c c c
Flange element ≤ 9ε ≤ 10ε ≤ 14ε
t t t
c 396ε c 456ε If, ψ > -1
if , α > 0.5, ≤ if , α > 0.5, ≤
t 13α − 1 t 13α − 1
Web element c 42ε
c 36ε c 41.5ε ≤
if , α ≤ 0.5, ≤ if , α ≤ 0.5, ≤ t 0.67 + 0.33ψ
t α t α
1 N Ed   2 N Ed 
Depth of compression zone, α = 1 + Proportion of tensile stress, ψ =   −1
2  f y t w d   Af 
 y 

elastomeric bearing
LE=1.5L
LE=0.85L

LE=1.3L

LE=0.7L
LE=1.0L

L
LE=2.3L

pinned support virtual pin at point of fixed support

contraflexure

non-sway struts sway struts

Effective lengths of struts LE

© Jonathan Haynes version 3.0 Page 79 of 239

32900 Bridge Engineering
MSc Structural Engineering

Lateral Torsional Buckling Curve

1.1

0.9

0.8
a
0.7

0.6
χ LT d
0.5

0.4

0.3

0.2

0.1

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3
λ LT

Flexural Buckling Curve

1.1

0.9
a
0.8

0.7

0.6
χ d
0.5

0.4

0.3

0.2

0.1

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3
λ

© Jonathan Haynes version 3.0 Page 80 of 239

32900 Bridge Engineering
MSc Structural Engineering

Tension Capacity Nt,Rd

Af y 0.9 Anet f u
Elements in pure tension must satisfy, N Ed ≤ N pl ,Rd = ≤ N u , Rd =
γ M0 γM2
where A is the gross cross sectional area and Anet is the net c-s-a after deducting for holes.
It may be necessary to check several failure paths to obtain the smallest net area:

Possible failure paths of a bolted tension element.

3.2.3 Slenderness and Deflection

The Eurocodes do not offer specific advice on the limitations which should apply to slenderness
and deflection in bridge structures. It is worth noting that the ‘approving authority’ is required to
set limits and in the UK the DoT generally accepts:

General compression elements, λ ≤ 180

Compression elements subject to wind load only λ ≤ 250

Tension only elements which may be subject to reversal λ ≤ 350

Vertical deflection of bridges or bridge elements ∆≤L

500

Cantilevers parts of bridges ∆≤L

300

Vertical deflection natural frequency 5 Hz

Horizontal deflection natural frequency 1.5 Hz

A simplified method of calculating natural frequency is contained in BD 37/01 Appendix B.

It is normal practice to provide precamber in primary steel beams and trusses, which should
remove permanent action deflection.

© Jonathan Haynes version 3.0 Page 81 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.2.4 End Bearing Stiffeners

Web Buckling and Bearing, Fw,Rd
Buckling of a beam web may occur due to an excessive vertical action being applied to a
beam flange which forces the web to buckle like a slender vertical strut. Bearing is similar to web
buckling but the web fails by yielding close to the root radius, like a stocky strut.

critical
section

length of web
NEd
yielding at bearing
area of web NEd
o acting as a strut
45

ss
ss d

leff

NEd

Web buckling and bearing.

f yw Leff t w
Web buckling and bearing are checked using a single design equality, N Ed ≤ FRd =
γ M1
where Leff is effective length, fyw is the web yield strength and tw is the web thickness.
0.5 l y tw f y
Leff = χ F l y where, χ F = ≤ 1.0 and λ F =
λF Fcr
t w3
Fcr is the web critical buckling force given by, Fcr = 0.9k F E
d
l y1 = s s + 2t f 1 + m1 + m2 ( )
2
m1  l e 
The effective loaded length ly is the minimum of l y 2 = le + t f + + m2
2  t f 

l y 3 = le + t f m1 + m2
where tf is the flange thickness and ss is the horizontal stiff bearing length of web over which the
transverse action is applied, other terms may be calculated from:
f yf b
m1 = where fyf is the flange yield strength and b is the breadth of the flange
f yw t w
k F Et w2
le = ≤ s s + c where E is the modulus of elasticity and d is the depth of the web between
2 f yw d
root fillets.

© Jonathan Haynes version 3.0 Page 82 of 239

32900 Bridge Engineering
MSc Structural Engineering

NEd NEd

c ss d hw

A A
a Ls Ls
tw
Unstiffened and stiffened web factors.
2
 ss + c  d
For unstiffened webs, k F = 2 + 6 ≤6 For stiffened webs, k F = 6 + 2 
 d  a
where c is the length of web beyond the stiff bearing, a is the distance between stiffeners.
2
d 
If λ F > 0.5 then m2 = 0.02  otherwise, if λ F ≤ 0.5 then m2 = 0 .
t 
 f 
When a beam web is not adequate to resist an applied transverse action then a web stiffener
must be added.
To calculate the axial capacity of the stiffened web, a cruciform section should be assessed. This
is formed from part of the beam web and the effective area of the stiffeners (since long
outstands of stiffening plates will tend to buckle locally).

length of stiffener Ls ≤ 14t s ε for class 3 outstand Ls tw

area of a stiffener As = Ls t s
[
effective area of a stiffener Aeff , s = 2 As + (2 × 15twε ) + ts tw ] 15twε
t s × (2 Ls + t w )
3

stiffener second moment of area I eff , s = ts

12
I eff , s 15twε
radius of gyration of stiffener ieff , s =
Aeff , s
0.75hw
dimensionless slenderness ratio of stiffener λ=
ieff , s × 93.9ε
Section A-A
χA f
Axial capacity N b , Rd = eff , s y where χ is obtained from axial buckling curve c.
γ m1

© Jonathan Haynes version 3.0 Page 83 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.2.5 Trusses
Truss frameworks are specifically addressed in BS EN 1993-2 Annex D.2 and PD 6695-2.
Basic rules for setting out trusses are:
• Split the structure into triangles,
• Element centre lines should join at nodes to avoid eccentricity moments,
• Provide a complete load path to ground,
• Keep bracing angles close to 45o for maximum efficiency,
• All joints are nominally pinned,
• All loads are applied at joints.
A truss bridge must be stable in three dimensions, with a truss forming each side face, also:
• Bottom face - it is common to use the deck as a horizontal diaphragm but ensure there is a
direct load path for horizontal forces from the truss node into the deck. When the deck is
mounted on the top chord of the truss, the bottom face will need direct bracing for stability.
• Top face – a horizontal truss is preferable but with through (or half-through) forms, there must
be a secondary stability system such as U-frames.
To simplify the analysis and subsequent fabrication it is common to rationalise the serial sizes used
to three or four (top / bottom chords and internal strut / tie).
Dependant upon the magnitude of actions, trusses may be subject to reversal of stress due to
wind uplift or continuity of structure, so the location of the compression chord may change.
The analysis model must reflect the continuity of the actual connections in the finished truss. The
magnitude of forces in bridge structures generally means that pinned conditions are rarely met.
The design of truss elements should comply with the rules for either tension elements or
compression elements. The compression capacity of any truss strut may be computed using:
2
 π 
N cr =   EI where Ι is second moment of area, L is the element length and β is an effective
 βL 
length factor. Buckling will be governed by the most slender axis.
Effective length of truss internal elements, β
For any element with nominally fixed ends:
In-plane buckling, β = 0.90
Out-of-plane buckling, β = 1.00

Out-of-plane buckling of truss vertical elements, which are also part of a transverse rigid frame.

h hr

β b

bEc I c for portal column

η= for portal beam
hEb I b

h
hr

© Jonathan Haynes version 3.0 Page 84 of 239

32900 Bridge Engineering
MSc Structural Engineering

In-plane buckling of diagonal elements with an intermediate elastic support.

L l

3kL 4l
β = 1− where, k≈
16l L

Effective length of truss chord elements, β

In a full box truss, the chord effective length may be taken as the distance between nodes of
restraint offered by the lateral truss. To be an effective restraint, the node must be able to resist
1% of the design axial force in the chord at that point.
Where U-frame action is needed to restrain a compression chord, two further analysis models are
required:
1. A plane frame model of a U-frame which reflects the actual stiffness of the beam and
column elements. Unit loads are applied to the compression chord locations and the
deflection obtained. The stiffness of the U-frame is obtained by dividing the load by
deflection.
2. A plane frame model of the compression chord subject to the design axial action NEd and
restrained intermittently by springs of stiffness ku.
The out-of-plane effective length of the chord is very large (length of the bridge) if the spring
stiffness is low. As the U-frame becomes stiffer the effective length will reduce. It is common
practice to use a UB section (web horizontal) to counter the large out-of-plane slenderness of
the compression chord.

Chord analysis model

NEd
1. 0
ku =
δu

δu

1 kN 1 kN

b
Bridge Arrangement
U-Frame analysis model

EI c I b
The U-frame stiffness may be estimated from, k u =
( )
3 h BI c + 2 I b h 3
2

The in-plane effective length of the chord may conservatively be taken as the maximum
distance between vertical supports (bay spacing).

There are particular rules for effective lengths of angles in compression since they can buckle
about any of four axes. Where angles form web elements of trusses, and two or more bolts are
used to connect each end of the angle, the slenderness may be calculated from:

© Jonathan Haynes version 3.0 Page 85 of 239

32900 Bridge Engineering
MSc Structural Engineering

For buckling about v-v axis, λ eff , v = 0.35 + 0.7λ v

Leff
For buckling about y-y axis, λ eff , y = 0.50 + 0.7λ y where, λ=
i × 93.9ε
For buckling about z-z axis, λ eff , z = 0.50 + 0.7λ z

Leff is taken as the distance between the intersection of web and chord element centroids.
centroidal axes cannot coincide at a point
unless the angle leg is cut, so the design must
include an eccentricity moment

Leff

centroidal axes of all elements

coincide at a point, so there is
no eccentricity

© Jonathan Haynes version 3.0 Page 86 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.2.6 Composite Beams

A popular and efficient way to achieve long span deck beams using steel is the composite
beam. The concrete deck is utilised to make an effectively deeper and stiffer beam than the
steelwork provides alone. The secret is to provide welded studs which prevent the steel beam
and concrete slab slipping along the longitudinal joint between them. This is called shear
connection.

beff

hf
h
d hsc hp
0.5hp Beam and slab without shear
connection – acts as two separate
bo sections.
bd 2 bd 2
For rectangles, Wel = 2 =
6 3
ha

Beam and slab with shear connection –

tf
acts as one deeper section.
b b(2d )
2
2bd 2
For rectangles, Wel = =
6 3

Shear and moment capacities at the construction stage may be taken for the steel beam alone.
The profiled steel decking should be shot-fired to the unpainted beam top flange. It is not until
shear connectors have been welded and the insitu concrete has hardened that composite
action between concrete and steel can be achieved.
Av f y
Vs = where Av is the shear area, conservatively Av = ht w
3γ m 0
W pl , y f y
Ms = for Class 1 and 2 sections, where Wpl is the plastic section modulus.
γM0

Shear and moment capacities at the composite stage may initially be calculated assuming the
position of the neutral axis.
Resistance of concrete flange, Rcf = 0.567 f ck beff h − h p ( )
Resistance of the steel section, R s = f y Aa
Resistance of the steel flange, R sf = f y bt f
Resistance of overall web depth, R w = Rs − 2 R sf
Resistance of clear web depth, Rv = f y dt w
Resistance of concrete above neutral axis, Rcx = 0.567 f ck beff x
Resistance of steel flange above neutral axis, R sx = f y bx1
Resistance of the web over distance x2, R wx = f y t w x 2

© Jonathan Haynes version 3.0 Page 87 of 239

32900 Bridge Engineering
MSc Structural Engineering

L
Effective breadth of concrete flange, be = ≤ bL where bL is the distance between adjacent
4
parallel beams and L is the span.

x < h : Rcf > Rs

For N.A. in concrete flange, h Rs (h − h p )
M c = Rs  a + h − 
 2 2 Rcf 

h < x < h + t f : R s > Rd > R w

Rc (h + h p ) (Rs − Rcf ) t f
2
For N.A. in steel flange, Rh
Mc = s a + −
2 2 4 Rsf

x > h + t f : Rcf < Rw

Rcf (ha + h + h p )
2
For N.A. in steel web, Rcf d
Mc = Ms + −
2 4 Rv

Shear connectors are needed between the steel beam and concrete flange to prevent
longitudinal slip. The proportion, or degree, of shear connection controls the proportion of extra
moment capacity generated (from a minimum of steel beam alone to full mobilisation of a
completely and rigidly shear connected section).
2
0.8 f u πd 2 k1 0.29αd k1 f ck E cm
Resistance of a headed stud, PRd = ≤
4γ mv γ mv
hsc h  h
for 3 ≤ ≤ 4 : α = 0.2 sc + 1 or for sc > 4 : α = 1
d  d  d
where hsc is the height of the stud, d is the stud diameter, γmv=1.25, fck is the concrete cylinder
0.3
 f + 8
strength and Ecm is the concrete secant modulus Ecm = 22 ck  (kN/mm ).
2

 10 
When profiled steel deck ribs are parallel to the supporting beam, fu=500N/mm2 and
b0  hsc 
k1 = 0.6  − 1 ≤ 1.0
hp h 
 p 
When profiled steel deck ribs are perpendicular to the supporting beam, fu=450N/mm2 and
0.7b0  hsc 
k1 = − 1 ≤ 1.0
nr h p  h p 

where nr is the number of stud connectors in one rib (must be ≤ 2.0 ).
Rc Rs
Number of studs for full shear connection, n f = ≤
k1 PRd k1 PRd
n
Degree of shear connection, η = where n is the actual number of studs provided.
nf
Studs are usually 19mm diameter x 95mm long after welding, and for beams of span less than
 355 
25m, η ≥ 1 −  (1.0 − 0.04 L ) ≥ 0.40
 f y 
Design moment capacity of composite beam, M p = M s + η (M c − M s )

© Jonathan Haynes version 3.0 Page 88 of 239

32900 Bridge Engineering
MSc Structural Engineering

Transverse Reinforcement
To prevent vertical shear failure of the concrete slab adjacent to the steel beam flange, steel
reinforcement must be included in the concrete if there is insufficient shear resistance.
fy
Design strength of steel reinforcement, f yd = where γms = 1.15 and fy =500N/mm2
γ ms
R L
Longitudinal shear stress, V Edl = where ∆ x =
2(h − h p )∆ x 2
Asf V Edl (h − h p )
Shear reinforcement, = where s is the spacing and θ may be taken as 26.50
s f yd cot θ

Deflection checks should be based upon the composite section properties. First mode natural
frequency should be checked assuming full variable action is applied.

© Jonathan Haynes version 3.0 Page 89 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.3 Abutments and Piers

Bridge substructures carry the deck or superstructure loading to the foundation soil, and retain
earth if needed.
Abutments are the supports at each end of a bridge. Many arrangements are possible, using
masonry, reinforced concrete (solid or cellular) or steel sheet piles. Abutments often incorporate
wing walls which function purely to retain earth.
When a tall abutment is not needed because of the approach geometry or earthworks, a
smaller version called a bank seat may be used.
Piers provide support at intermediate points in the span of the bridge. Invariably piers are
columns or walls of steel or reinforced concrete construction.

abutment pier bank seat

sheet pile
capping
beam
joint

bearing
cantilever retaining
wall, possible
counterfort’s at
about 3.0m centres
possible wall or columns on
shear key footing or pile cap

wing wall

high
modulus
sheet pile
wall and
wing wall

Care must be taken when selecting load combinations for design of foundations. In the case of
an abutment retaining wall, maximum vertical deck loading will enhance sliding resistance but
create the most onerous stem axial force. It is prudent to assess at least the maximum and
minimum vertical loading and maximum horizontal loading combinations. Remember there are
separate limit states for structural design (STR) and geotechnical soil checks (GEO).

© Jonathan Haynes version 3.0 Page 90 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.3.1 Calculating Forces and Settlements in Foundations

Like any structure, bridges may be supported on shallow or deep foundations. Shallow, (spread)
foundations will generally be the more economic option but are only suitable on stronger soils.
Pile foundations are common in bridge design since vehicle loads induce high local stresses in
most parts of the bridge superstructure and settlements must be carefully limited to avoid
permanent changes in level at movement joints.
All soil, including rock, will deform when loaded. This deformation will affect the forces in the
bridge structure loading the soil. This will have greatest effect when the structure is continuous.
This interdependence is termed soil-structure interaction.
Soil is the least predictable (or reliable) engineering material, and is usually the least tested, so
predictions of capacity must be conservative to ensure safety. For this reason, the use of
complex analysis for foundations has traditionally been avoided. The simplest and most popular
analysis technique is termed the rigid soil analogy, which treats soil like any other solid
engineering material (like the foundation concrete), so rigid elastic stress resultants are assumed.
Rigid Soil Analogy
The rule of the middle third states that tension will not be developed until the action is placed
outside the middle third of the section.

1. Centroid of action 2. Centroid of action within 3. Centroid of action outside

coincides with centroid of middle third of foundation. middle third of foundation.
foundation. No eccentricity. Small eccentricity. B is the Large eccentricity.
breadth of the section into
the paper.
L L
P e≤ P e> P
6 6

L
R
z
P P M P Pe 3z
f max = f max = + = + 2 L
A A Wel BL BL z= −e
6 2
P  6e  2R
f max = 1 +  f max =
BL  L 3Bz
P M P  6e  f min = 0
f min = − = 1 − 
A Wel BL  L

It is clear that the interface between soil and a reinforced concrete foundation is not similar to
the stress distribution in a steel beam (which is what the rigid soil analogy assumes). More
advanced analogies were developed in the middle of the twentieth century.
In real soils, stress measurements show that different soils react to loading in very divergent ways.
We commonly classify soils as cohesive or cohesionless (clay or sand) this is largely because these
materials exhibit the two extremes of soil behaviour. Further, soil is a three phase material (air,
solids and water) and it is the water content which has the greatest effect upon capacity. The
total stress in a soil is the effective stress minus pore water pressure (Smith, 1982):
τ ′ = σ ′ + Tanφ
© Jonathan Haynes version 3.0 Page 91 of 239
32900 Bridge Engineering
MSc Structural Engineering

Considering the deformed shape of a uniformly loaded flexible foundation:

dry sand
saturated clay

The drained stiffness of sand depends on the The undrained stiffness of clay is
confining stress, which is zero at the edges independent of confining stress, so the
and maximum at the centre. Settlement is settlement takes a concave shape. Settlement
minimum at the centre. is maximum at the centre.

Considering the contact pressure under a uniformly loaded rigid foundation:

dry sand
saturated clay

Since the drained stiffness of sand is zero at Since the foundation settles at the centre,
the edges, the contact stress will also be zero stress is redistributed and minimised.
at the edges. Contact stress is maximum at However, at the edges there is little
the centre. settlement and therefore maximum stress. As
the soil approaches its capacity, yielding
occurs at the edges.

Beam on Elastic Foundation

Soil is not homogeneous (same stiffness with depth) or isotropic (same material properties across
the section). Any mathematical model must be an approximation of reality. The first successful
attempt to model soil and structure separately was provided by Terzaghi (1955) based upon the
Winkler spring stiffness, kw (or Modulus of Subgrade Reaction). This assumes that soil is a series of
discrete springs which support a foundation
plate.
q stress q
h For a plate, stiffness k w = =
settlement δ
δ
Timoshenko & Goodier (1969)suggest that,
kw
δ=
(
0.95qh 2 1 − ν 2 )
E A
E
∴ kw =
(
0.95h 1 − ν 2 )
load Winkler spring stiffness is not a fundamental
ultimate
capacity
property of a soil since it is dependant upon
geometry, it is obtained from a 12 inch plate
load test. A steel plate is jacked into the soil
whilst measuring settlement and load. From a
one third ultimate capacity load – deflection graph the ultimate capacity is
P
identified, stiffness is then calculated at one third
of ultimate capacity.
δ settlement
P
kw =
δ

© Jonathan Haynes version 3.0 Page 92 of 239

32900 Bridge Engineering
MSc Structural Engineering

Timoshenko and Woinowsky-Kreiger (1959) provided theory of internal forces in an edge

supported plate subject to out-of-plane loading. For deflection, w and plate thickness, t

Qx q Qy ∂ 4 w 2∂ 4 w ∂ 4 w
+ + = q
(
12 1 − ν 2 )
∂x 4 ∂x 2 ∂y 2 ∂y 4 Et 3
Solutions of three dimensional problems are
tedious without a computer but this can be
Mx combined with beam theory to arrive at a
My
two dimensional solution called the beam
Myx Mxy
on elastic foundation model.
M σ E
Simple bending, = =
I y R
d2y
q (kN/m )2 Differential equation of flexure, EI = −M
dx 2
d
dQ d 2 M
x Load-shear-moment, − q = =
dx dx 2
kw
y d4y
So, q = EI
dx 4
If Winkler spring, plate bending and beam bending theories are combined, the governing
equation for a beam on elastic foundation may be obtained,
d 4 y Bk Bq
4
+ y=
dx EI EI
Hetenyi (1946) provided a general solution to this equation, considering an infinitely long beam
which has breadth, B (into the page), depth d and Young’s modulus of elasticity E.
Pλ
Settlement, y = (cos λx + sin λx )e −λx kwB
2k for a point load P and λ=4 Hetenyi presents
4EI
P
Moment, M= (cos λx − sin λx )e −λx numerical solutions for the exponential functions
4λ
(see graph) to permit hand solutions. Simple cases
P
Shear, S = − (cos λx )e −λx are depicted below:
2

q P M

As practical problems rarely involve an infinitely long beam and a single loading, it is necessary
to adjust solutions by adding results from multiple loads by applying the Principle of Superposition
and correcting for bending moments to the foundation ends, by applying balancing fictitious
bending moments to the foundation ends.
© Jonathan Haynes version 3.0 Page 93 of 239
 32900 Bridge Engineering
MSc Structural Engineering

1. create a solution for a beam of infinite

length, using the load position as zero x,
2. if there are moments at the foundation
infinite beam bending
moment solution
-x +x ends, create a solution for twice that
moment (opposite rotation direction) using
MA the load position as zero x,
MC
+ 3. algebraically summate all the solutions to
end A correction -2MA obtain a final solution. There should be zero
-MA moments at the foundation ends, a second
+x + -2MC cycle or correction may be needed.
end B correction -MC
= +x Go to Blackboard and download the spreadsheet
BeamOnElasticFoundation3.xls
final solution
and have a go at comparing rigid body statics
with beam on elastic foundation solutions.

Hetenyi's coefficients for beam on elastic foundation

1.0
Point Load
Pλ
P
y= φ λx
2k
P x
M= ψ λx
0.8 4λ o
−P
φ S=
2
θ λx

Moment
θ Mλ2
0.5 y= β λx M
Coefficient

k x
ψ M
M = θ λx o
2
− Mλ
S= φλ x
0.3 2

β
λx
0.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0

-0.3

Geotechnical Finite Element Analysis

Generally FEA uses shape functions to ensure compatibility of displacements between elements.
Stresses may not be in equilibrium between adjacent elements because FEA only provides
approximate solutions to governing differential equations and equilibrium is satisfied on average
across the model. There is rarely sufficient site investigation data to justify confidence in
geotechnical FEA output. Therefore, it is best to perceive geotechnical finite element analysis as
a tool to investigate the relative magnitude of settlements in soil, as it is not possible to
accurately calibrate output to the extent which can be achieved with structural finite element
analysis – the need for sensitivity studies is more acute with soil analysis.

© Jonathan Haynes version 3.0 Page 94 of 239

32900 Bridge Engineering
MSc Structural Engineering

Two dimensional geotechnical FEA is becoming commonly used but it requires greater
consideration of fundamental physics than structural FEA. The common steps in a Geotechnical
FEA may be outlined as:

© Jonathan Haynes version 3.0 Page 95 of 239

32900 Bridge Engineering
MSc Structural Engineering

Stage 1 – Mesh soil continuum

GeoFE models use a large elastic half
space (a) which allows strain to dissipate
broadly. Apply constraints at edges (b),
usually roller pin supports. Mesh size varies a
with distance from load application.
Largest displacements are closest to load,
so need the finest mesh (called grading).
Most common FE’s are the 6 node linear
strain triangle (LST) and 8 node linear strain
quadrilateral (LSQ). Although the LSQ will
offer greater accuracy, it is more difficult to
use and requires greater computing power
to solve.
Stage 2 – Assign Material Properties
The foundation requires concrete b
properties. The soil requires a constitutive
model (like Hookes Law) but developed for
a particular soil. This defines:
Yield Function – defines the surface that
separates elastic and plastic behaviour, σ Strain hardening
commonly Mohr-Coulomb or Tresca.
c
Flow Rule - defines the direction of plastic Perfect plastic
strain at critical failure (d). σy
Hardening Law (c) - defines pre and post
yield stiffness. Most soils have elasto-plastic Strain softening
models.
The drained (effective elastic) properties of
the soil must be defined. ε
Bulk modulus of water, Kw = 0 for drained soil tangent to
failure
(no shear stiffness). For saturated soil, Kw =
100xγ soil drained to avoid convergence σ2 d
direction of
problems in the matrix analysis. σy plastic
Stage 3 – Initial Stress State strain
failure point progressio
Unlike Structural FE, GeoFE models must be σy
created with an initial stress state.
σy σ1
• Define bulk density above and below
the water table, Von Mises
Yield Loci
• Make an initial time increment, when σy
gravity is applied to the model, this
creates a horizontal effective stress,
• Excavate soil to formation level by
removing elements of soil,
• Construct foundation by adding e
elements of concrete,
• Load foundation.
Stage 4 – Analysis
As with Structural FE - solve for node
displacements (e) then backsubstitute for
node stresses. f
Output is usually in the form of a contour
stress plot (f) this can be Principal stress or
von Mises stress, which permit checking
against design limits.

© Jonathan Haynes version 3.0 Page 96 of 239

32900 Bridge Engineering
MSc Structural Engineering

Plaxis Geotechnical FEA is available in Newton 155 and 156. It has a full help system but
advice is available from Dr Swift or Dr Toma-Sabbagh.

The choice of which analysis option to adopt should reflect the necessary accuracy and extent
of site investigation data available.
Analysis Disadvantages Advantages
method
Output for foundation only. Site
Very fast but grossly
Investigation merely needs to establish
Rigid soil unrepresentative of reality.
Allowable Bearing Pressure (typically
analogy Underestimates B-P and
unconfined compression Cu or Standard
overestimates bending moment.
Penetration Test blow count N)
Output for foundation only but can
Beam on Simple to operate or automate and
include settlement, shear force and
elastic better representation of reality but
bending moment. Needs a plate load
foundation still poor.
test to establish kw.
Output for soil and foundation but vast
amount of data will need interpretation.
Requires wide range of complex soil Best representation of reality
tests. Unlikely that actual predicted available but will not provide
Geotechnical settlements will reflect reality – needs a reliable data (insufficient SI data),
FEA sensitivity study for calibration. Model used for comparative analysis
obeys equilibrium averaged across (sensitivity study). Gives stress and
model, so stress may not be in settlement values in all elements.
equilibrium across adjacent elements –
use shape functions.

3.3.2 Geotechnical Design

A geotechnical action is the load imparted on a structure by the ground.
Ground is soil, rock or fill in place before the works are executed.
Structure is the material introduced to the site during the works, this may include imported
ground.
EC7 permits design by calculation, prescriptive measures, load testing and the ‘observational’
method. Only design by calculation will be pursued here. EC7 prescribes parameters to be
considered in a geotechnical investigation, data and results being required to form part of a
Geotechnical Design Report (GDR) which should also record the assumptions, method of
calculation and results used to produce a safe structure.
Three geotechnical categories are specified in EC7. The first covers small or simple structures,
where there is negligible risk. The second covers conventional structures (spread foundations)
with no difficult ground conditions, where there is no exceptional risk. The third covers all other
structures or those on difficult ground which would create abnormal risks. Foundations in this
category are required to be assessed by a competent geotechnical engineer.
The structural design of foundations should consider the EQU and STR limit states as normal.
However, the geotechnical soil checks must also be undertaken to limit state theory.
Characteristic soil strength parameters (taken from the site investigation and testing report) must
Xk
therefore be reduced by the relevant material partial safety factor, or mathematically X d =
γm

© Jonathan Haynes version 3.0 Page 97 of 239

32900 Bridge Engineering
MSc Structural Engineering

Geotechnical ultimate limit states

consideration of static equilibrium considering the structure as
EQU Equilibrium
a rigid body
failure or intolerable deformation of the structure where
STR Structure
strength of the material provides resistance
failure or intolerable deformation of the ground where
GEO Geotechnical
strength of the ground provides resistance
UPL Uplift loss of equilibrium due to buoyancy
HYD Hydraulic heave Erosion or piping in the ground due to hydraulic gradient

There are four general failure mechanisms which may occur in foundations:

W W W
F F
h

C
a

C
Overturning Failure Structural Failure Sliding Failure Soil Shear Failure
EQU FoS against overturning STR FoS against bending/shear GEO FoS against sliding GEO FoS against slip circle

Rk Ck Ck
Wk γ f a γ ms γ mc ' γ mc '
≥ 1.00 ≥ 1.00 ≥ 1.00 ≥ 1.00
Fk
h Ek γ f Fk γ f Wk γ f
γm
Ground parameter material partial safety factors.
EQU STR & GEO UPL HYD
Parameter γm
M1 M2
Angle of shear resistance γφ’ 1.1 1.0 1.25 1.25 -
Effective cohesion γc’ 1.1 1.0 1.25 1.25 -
Undrained shear strength γcu 1.2 1.0 1.4 1.4 -
Unconfined strength γqu 1.2 1.0 1.4 1.4 -
Unit weight γγ 1.0 1.0 1.0 - -
Tensile pile resistance γst - - - 1.4 -
Anchorage γR - - - 1.4 -

Generally, the STR limit state will govern structural design of the foundation (reinforcement
design) and GEO limit state will govern the geotechnical design (bearing pressure, sliding and
settlement) but the EQU limit state is intended for overturning checks. The calculation of these
checks has therefore become rather tedious and is ideally undertaken on a spreadsheet or other
automatic calculation medium.
Spread Footings
The pad footings are the simplest engineered foundation. For economic reasons, foundations are
often thin and reinforced to prevent cracking. As a rough guide, the thickness to outstand ratio
should be:
1
h  p2  4

≥ 0.15 
 ≈ 0.5 where p is the max SLS bearing pressure, h is overall depth of the footing.
a f
 ck , cube 
© Jonathan Haynes version 3.0 Page 98 of 239
32900 Bridge Engineering
MSc Structural Engineering

Minimum depths of foundations in the U.K. are governed by freeze-thaw depths and covered by
the Building Regulations. This usually sets underside of foundation at -1.0m or deeper. For
durability, concrete specification would normally exceed C28/25, using cover of at least 50mm
where blinding concrete is used or 75mm where concrete is cast against the ground.
Bending moment is calculated assuming the pad cantilevers out from the face of the column in
all directions. Bending reinforcement is designed in the same manner as a beam. The
reinforcement is placed in both directions, in the bottom of the pad.
N Ed pmax, Ed × a 2
If the factored contact pressure is pmax, Ed = then M Ed = this moment will be
BL 2
per metre run of the foundation breadth (B).

Shear failure may occur by several modes:

L

2.0d
1.0d
a B

NEd
NEd NEd
NEd

Bending failure Column face shear Punching shear Line shear

1. Column face shear, where the column simply pushes through the foundation.
vf cd ud ave
shear capacity of pad, V Rd ,max = ≤ N Ed
2
 f ck 
where v = 0.6 1 − , u is the perimeter of the column, dave is the average effective depth and
 250 
α cc f ck
f cd = where αcc = 1.0 for shear and γc = 1.5.
γc
2. Punching shear, where the load spreads out into the foundation (to the punching shear
perimeter 2.0d from the column face) but still pushes through.

length of the critical perimeter, L2 = u + 4πd

area within critical perimeter, A2 = c + 4d ( 2
) − (4 − π )(2d ) 2

action within critical perimeter, V Ed , 2 = p max, Ed A2

VEd , 2
shear stress at critical perimeter, v Ed = ≤ v Rd , c
L2 d ave
cot θ
shear capacity, v Rd , max = vf cd 1.0 ≤ cotθ ≤ 2.5 (if shear is adequate using cotθ ≤ 2.5
1 + cot 2 θ
no further checks are required).
3. Line shear, where a strip of the foundation shears off, 1.0d from the column face.
© Jonathan Haynes version 3.0 Page 99 of 239
32900 Bridge Engineering
MSc Structural Engineering

action outside line, V Ed ,3 = p max, Ed BL3

VEd , 3
shear stress at line, v Ed = ≤ v Rd ,c
Bd ave
It is normal practice to avoid shear reinforcement in simple foundations by increasing the pad
thickness until sufficient shear strength is achieved.
Combined Footings
If it is not possible to provide individual foundations for each column or wall, a combined footing
may be used. If a spread footing is the preferred option, the procedure for design is the same as
previously outlined for individual pad footings but the eccentricity must be established by
locating the centroid of loading.

y Column 2 Taking moments of load, (x-x)

y=
∑ Py = P y 1 1

centroid of load
∑P P y + P y
1 1 2 2

y2

y Taking moments of load, (y-y)

Column 1
x x
x=
∑ Px = P x 1 1
x ∑P Px + P x
1 1 2 2

x2
y

The column loads are now combined into a resultant positioned at the centroid. If the
foundation can be placed to coincide with the loading centroid there is (nominally) no
eccentricity, otherwise possible bi-axial moments should be considered. Practically, it is advisable
to allow for construction tolerance, generally taken to be at least 75mm about both axes. A
combined footing should be reinforced on both faces to account for all possible load
combinations.
Pile Foundations
When shallow foundation solutions are inadequate with respect to soil bearing capacity or
settlement, then pile foundations are necessary. A pile is essentially a column buried in the
ground. Piles are normally installed in groups and connected together at ground level by a cap.
The action from the structure is applied directly to the cap, which is designed to distribute the
action equally between all of the piles, thus it must be very stiff.
Piles are most efficient when resisting compressive forces, although they can be designed to
resist tensile forces. Until recently horizontal forces were resisted by introducing raking piles to the
group, however it is now possible to design for horizontal action using Brohm’s method.
The capacity of a pile is a geotechnical consideration and is dependant upon many factors,
including :
• Soil type and properties
• Method of pile construction
• Degree of interaction with other piles in a group
• Nature of the loading
There follow some basic notes on scheme design of piles for structural capacity only. Piles may
be categorised as end bearing (transferring load to a rock layer) or frictional (transferring load
into the surrounding soil), and by interacting with cohesive (clayey) or cohesionless (sandy) soil.

© Jonathan Haynes version 3.0 Page 100 of 239

32900 Bridge Engineering
MSc Structural Engineering

Qshaft
The pile capacity is :
D
Qshaft Q shaft Qbase Q shaft + Qbase
N Ed = + ≤
2D
1 .5 3 2.5
2D
2D
Qbase
Qbase

END BEARING FRICTIONAL UNDER-REAMED

Different capacity formulae are used for cohesive and cohesionless soils. Piles in cohesionless soils
are usually driven to a set – hammered into the ground until the deflection is reduced to an
acceptable level. Piles in cohesive soils are usually designed, installed and load tested to prove
capacity.
Bored Piles in Clay
Piles in cohesive soils are usually bored – a drilled hole is filled with concrete.
capacity of shaft, Qshaft = α cu As
where, surface area of the pile shaft, As = πDL , average unconfined compression strength over
shaft, cu and adhesion factor, α.
capacity of base, Qbase = 9cub Ab

where, base surface area, Ab =

πD 2 , average unconfined compression strength over shaft, c
ub
4
The adhesion factor accounts for the fact that soft clay are stickier than hard clays and bond
better to wet concrete. For this reason the value of α may be assumed to vary linearly between
1.0 and 0.4 for unconfined compression strengths of 50kN/m2 and 200kN/m2 respectively. The top
1.0m of clay should be ignored when calculating shaft capacity as it is likely to be disturbed in the
boring process. When assessing an under-ream it is advisable to ignore shaft capacity from the
bottom 4D as this section is unlikely to be able to displace sufficiently to contribute to capacity.
A bored pile must be reinforced with at least six longitudinal bars of 16mm diameter, of a
maximum circumferential spacing of 200mm
Driven Piles in Sand
Piles in cohesionless soils are usually driven to a set – hammered into the ground until the
deflection is reduced to an acceptable level. Since the piles are precast they are usually square.
γLk s As Tanδ
capacity of shaft, Qshaft =
2
where, surface area of shaft, As = 4 DL , angle of friction between soil and pile, δ = 0.75φ
unit weight of soil, γ angle of internal shear friction, φ shaft adhesion factor, ks
capacity of base, Qbase = γLN q Ab
2
where, surface area of base, Ab = D and Terzaghi’s coefficient, Nq
Design values for dry sand.
Loose Medium Dense
φ (degrees) 30 35 40 45
Nq 22.5 41.4 81.3 173
γ (kN/m2) 12 14 16 18
ks (degrees) 1.0 1.0 1.5 2.0
© Jonathan Haynes version 3.0 Page 101 of 239
 32900 Bridge Engineering
MSc Structural Engineering

Qshaft
Do not exceed a shaft resistance of = 100kN / m 2
As
Q 2
Do not exceed a base resistance of base = 10,000kN / m
Ab
If a pile is driven through a soft soil into a harder soil then it is advisable to assume the soft soil will
be driven 3D into the top of the hard layer. A pile hammer which weighs less than the pile will
bounce. A driven pile will achieve peak capacity at a set of approximately 10mm per blow.
A driven pile densifies an area around the pile of approximately 5D diameter.
Pile Group Capacity
Groups of piles do not achieve a capacity equal to the sum
of individual pile capacities, this is because of multiple
5D D loading of the soil zones between piles. Piles in groups are
normally spaced at 3D centres.
For two piles at 3D spacing, the loss in capacity is around 25%,
3D for four pile groups this grows to around 40%.
For a more exact assessment, the group capacity factor may
be calculated from,
  m(n − 1) + n( m − 1)  
H
(
f g = 1 −  Tan −1 D 
S 
) 

 90nm
for n rows x m columns of piles
In clay, a large group of piles may exhibit block failure in
L which the whole group moves together, and capacity is
limited by group base capacity.
Group capacity is therefore,
B N Ed , g = N Ed nmf g ≤ N c cub LB

60o
(
where, N c = 5 1 + 0.2 H
B
)(1 + 0.2 B L )
s=3D Arrangements for three and four pile caps are shown. Where
more piles are needed these may be combined.
The outer edge of the cap should overhang beyond the
outer pile perimeter by at least 150mm.
s=3D s=3D
Eurocodes permit pile cap analysis by beam or truss
analogy. Truss analogy should be used since this is both most
economical and closer to the real action of the cap.
where d is the effective depth of the tension
reinforcement.
F If we consider a two pile cap subject to an
applied action F, the reaction at each pile is
F
d
compression 0.5F, From the truss geometry, 2 = Tanθ = d
tension T S
2
FS
and hence, T = (tensile force across cap)
s=3D 150 4d
the required area of tension reinforcement is
F F
T
2 2 given by, As =
f yd

© Jonathan Haynes version 3.0 Page 102 of 239

32900 Bridge Engineering
MSc Structural Engineering

Tensile force formulae for pile caps.

No of piles 2 3 4 5 6
FS FS FS FS FS
Tensile force, T (kN)
4d 9d 8d 10d 12d
The reinforcement should be evenly distributed across the pile cap. Where pile spacing is
greater than 3d tension reinforcement should be placed in the zone 1.5d each side of the pile.
Shear strength must be checked in pile caps of any arrangement. There are several possibilities
for the location of the critical vertical section. The applied shear force is due to all piles outside
V Ed
the critical section. Generally, for a satisfactory section, v Ed = ≤ v Rd ,c .
bd
D D D
5 5 5
D D D
5 5 5
1 ≤ 2d 4 1 ≤ 2d 4 1 ≤ 2d 4

≤ 2d ≤ 2d ≤ 2d
2 3 2 3 2 3

Shear force from piles 1 & 2 Shear force from piles 2 & 3 Shear force from all piles

If pile caps are sized such that

av
minimum two rows ≈ 1.0 then deflection and
av circumferential
h
links of 12 φ bar cracking will not be
onerous design checks.
Since there are often large forces
h>2D+100

in the tension reinforcement of pile

< 300mm

caps and the ends of the

reinforcement must be anchored
into the concrete (beyond the
≥ 100 mm
point where it ceases to be
≥ 150mm required) then there is usually the
s=3D need for a slow bend to prevent
bursting of the concrete inside the
bend.

bar stress reduces to zero

at full anchorage length Fbt 1 1 
compression in
concrete
φm , min ≥  + 
f cd  ab 2φ 
where ab is half the bar
φm,min
spacing, Fbt is the tension in
Fbt the bar, φ is the bar
diameter, φm,min is the
minimum mandrel bending
anchorage lap length stress in bar maximum at diameter, ≤ 16φ ⇒ 4φ
pile centreline
> 16φ ⇒ 7φ
C
L Pile

© Jonathan Haynes version 3.0 Page 103 of 239

32900 Bridge Engineering
MSc Structural Engineering

Eccentrically Loaded Pile Caps

All pile caps should be designed for a minimum load eccentricity of 75mm to account for setting
out tolerances. It may also be necessary to design for deliberate eccentricity, in which case the
principles of solid mechanics may be used :
x Y
x2
Centroid of pile group

2 3 Centroid of column or pier

a2 a3 Position of pile
y2 b2

X X
ey F, action on column
W, self weight of pile cap
b1 N, number of piles
y ex
4 Pn, action on pile n
1
x4

Calculate position of action centroid, x =

∑P x n n
and y =
∑P y
n n

∑P ∑P
Calculate eccentricities and eccentricity moments, e x = xcol − x and M y = ∑ Pe x

F +W
Calculate the actions on individual piles, which are comprised of a direct axial element
N
a b
and bending elements ±My and ± M x
∑ a2 ∑ b2
Retaining Walls
Structures used to prevent the collapse of earth (or other loose material) are termed retaining
walls. The soil mechanics of retained earth is covered under Integral Bridges.
The following notes deal with the design of a cantilever retaining wall, since this is the most
common structural form. A wall must displace and translate before the earth pressure changes
from earth pressure at rest (Po) to active earth pressure (Pa).
For passive pressure to develop, displacements approximately three times larger than those for
active pressure, are required, these would usually be of an unacceptable magnitude.
All retaining walls should be designed for hydrostatic pressure equivalent to ground water level
at 0.75 wall height. This applies even if weep holes are provided.

Po = koγh 2 k aγh 2
h Pa =
2
h 0.5o
h
2 3
koγh h kaγh
1000
AT REST TRANSLATED ROTATED
Stability checks are required, which should take account of partial safety factors (action and
material). The most onerous combination of adverse and beneficial factors must be used. The
© Jonathan Haynes version 3.0 Page 104 of 239
32900 Bridge Engineering
MSc Structural Engineering

extremes of loading on the abutment will often be a STR combination check with full adverse
variable action on the deck, and an EQU check with minimum deck action and maximum
lateral pressures.
Wb q

γ w = 10 kN / m 3
hw ≥ 0.75 h Ws Pq = ka qh
xs kaγh 2
L ≈ 0.67 h Pa = 2
2 γ w hw

hw
Pw =

h
Ww ya
2 yq
2 yw
k pγ h
Pp =
2
hp

yp
L where, Wb is the bridge deck reaction, Ww is the permanent
action of the retaining wall, Ws is the permanent action of
BPmin the soil over the base slab, Pa is the active earth pressure,
Pw is the water pressure and Pq is the surcharge pressure.
BPmax

A shear key may be needed to avoid sliding. Since displacements to develop passive pressure
(Pp) can be three times those for active pressure, it may be wise to ignore the assistance of
passive resistance.

ΣW obtain the location of load centroids

h
stem thickness ≈ using moments of force,
12

x=∑ y=∑
Wx Py
x
∑W ∑P
base outstand ≈
h ΣP
~
x = x− y
∑P e=
L ~
−x
8 y ∑W 2

R h
~
x e base thickness ≈
10

Factor of safety against overturning, FoS =

∑W x ≥ 1.0
∑Py
Factor of safety against sliding, FoS =
∑R ≥ 1. 0
∑ Pγ f

C φ
on cohesive soils, ∑R = γ A on cohesionless soil, ∑ R = WTan γ
mc' mc '

Bearing pressure, BP =
∑W γ f
±
∑ Pyγ f + ∑Weγ f
if this is negative, tension exists under the
A Z
base and maximum pressure should be reassessed ignoring the area subject to tension.

3.3.3 Concrete Struts

Reinforced concrete piers will generally be solid walls but may be individual columns with
capping beams. Abutments are almost invariably solid walls. Both piers and abutments are
subject to predominantly axial loads with out-of-plane uniaxial moments.

© Jonathan Haynes version 3.0 Page 105 of 239

32900 Bridge Engineering
MSc Structural Engineering

Out-of-plane moments are may be due to horizontal load cases such as longitudinal braking
forces. Biaxial moments may occur when in-plane moments develop, due to load cases such as
centrifugal braking forces or lateral wind loading. Vertical deck reactions will create out-of-plane
moments if the bearings are supported on a corbel.
EC2 classifies any part of the structure which contributes to stability (columns or walls) as
unbraced, and these elements are likely to have effective length factors greater than unity.
Conversely, braced elements are those which do not contribute to stability, these elements are
likely to have effective length factors smaller than unity. Compression elements are further
classified as stocky or slender depending upon slenderness ratio. Slender elements are defined
as those where second order moments (P-∆ effect) are significant.

central pier where no horizontal forces are

resisted, elastomeric bearings provide little
restraint to horizontal movement of abutment where horizontal
continuous beams forces are resisted

l0=2.3L
l0=1.3L

base of abutment connected rigidly to wall,

but joint assumed to suffer some rotation
l0 > 1.0 L
l 0 > 1. 0 L due to foundation settlement / rotation

base of central pier connected rigidly to wall,

but joint assumed to suffer some rotation due
to foundation settlement / rotation Effective lengths of piers and abutments

Effective lengths must be checked on both axes, since the most slender will govern buckling. A
concrete strut may conservatively be size to avoid secondary moments by ensuring it is stocky,
lo 15.4C f
λ= ≤ λlim = where lo is the effective length, f cd = ck and C =1.7- rm
i N Ed γc
Ac f cd

Mo1 Mo1 Mo1

the factor C depends upon the
shape of the elastic bending
moment diagram,
Mo2 Mo2 Mo2 M o1
rm = -ive rm = 0 rm = +ive rm = and Mo2 > Mo1
M o2

lo h
the design bending moment, M Ed = M o 2 + eN Ed where e = ≥ ≥ 20mm
400 30
Where columns are bent about two axes, uniaxial checks will suffice if the following are satisfied:
ey ez
λy λz h ≤ 0.2 or b ≤ 0.2 where e = M Ed
≤2 and ≤2 and,
λz λy ez ey N Ed
b h
Otherwise the following interaction must be satisfied:
a a
 M Edz   M Edy 
  +   ≤ 1.0 MRdz and MRdy are the respective moment capacities
 
 M Rdz   M Rdy 
© Jonathan Haynes version 3.0 Page 106 of 239
32900 Bridge Engineering
MSc Structural Engineering

Values of biaxial bending power, a.

square and N Ed
N Rd 0.1 0.7 1.0
rectangular
columns where, N Rd = Ac f cd + As f yd
a 1.0 1.5 2.0
circular a 2.0
columns
The process for biaxial bending requires the column size and reinforcement to be known first. For
direct design, the BS 8110 method of calculating an equivalent uniaxial moment may be used:
b
My Mz 1 b' b’
If ≥ then M y = M y + β M z
b h h' y

My Mz 1 h'
If < then M z = M z + β M y
b h b' z z
h h’ Mz
N Ed
where β =1−
bhf ck
y
Further checks for columns and walls,
As My
maximum reinforcement percentage, ≤ 0.04
Ac
0.1N Ed
minimum reinforcement percentage, As min = ≥ 0.002 Ac
f yd

link diameter, φlink ≥ φ 4 ≥ 6mm and vertical spacing, s ≤ 12φ ≤ 0.6b ≤ 240mm

© Jonathan Haynes version 3.0 Page 107 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.4 Post-tensioned Concrete Beams

There are severe limitations on the ability of reinforced concrete to span the long distances
commonly associated with bridges. It is therefore common for elements like concrete beams to
be stressed in some way, to utilise the inherent high compressive strength of concrete, and
obviate its poor tensile strength.
Precast concrete systems are very popular since they provide factory quality construction and
permit shorter on-site construction periods. Although precast prestressed elements are very
common in buildings (particularly floor planks) they are rarely used in bridges. Prestressed planks
are cast in long beds which used hydraulic jacks to apply tendon stress.
Precast post-tensioned elements are very common in bridge construction. Deep beam sections
are often precast in outdoor yards. Carefully positioned ducts for stressing tendons and bending
(lifting) and shear reinforcement are cast in. Tensioning of tendons is undertaken insitu and may
undertaken in several steps as part of the construction sequence.
Only stressing for major axis bending moment will be considered here.

3.4.1 Materials and Equipment

Since the compressive strength of concrete is being exploited, high strength concrete is
frequently used to create efficient designs. Strengths of 60N/mm2 or more are common.
5mm wire strand, similar to the material used in cable wire, is the basic stressing element. This is
wound together to form tendons, which is often seven strands. Solid steel bars and reinforcing
bar are also used but this is declining.
Strand ducts are galvanised spiral products which must be flexible enough to be bend to large
radii and wired into position in a shutter. If the tendon is to be grouted in after stressing, there
must also be grout ducts at regular intervals along the beam.
The stressing tendon must apply an axial compression to the concrete. This is achieved by
casting an endblock (where the stress is applied) and a stopend into the ends of the beam.
These anchorages are proprietary products designed to distribute stress across the beam ends.
The tendons are locked at the stopend and a hydraulic jack is fitted to the endblock, the tendon
is then stretched a predetermined distance and wedges fitted to lock the arrangement in place.

Lifting precast prestressed floor planks; stressing ducts in a post-tensioned flat slab; prestressed plank long
bed manufacturing process using hydraulic jacks.

Stressing strand, tendons and bar; a strand endblock with 17 wedges; stopends and a stressing point in a
post-tensioned flat slab, showing blue grouting ducts.

© Jonathan Haynes version 3.0 Page 108 of 239

 32900 Bridge Engineering
MSc Structural Engineering

3.4.2 Principles of Analysis

Reinforced concrete is a full composite material, where there must be strain compatibility
between the steel reinforcement and concrete. The analysis of RC elements in flexure is therefore
complex. RC elements will crack at very low loads and are designed to resist ULS actions.
Post-tensioned concrete follows much simpler principles (and is older technology). Since
cracking is eliminated at SLS loads, simple elastic bending theory may be applied to obtain
direct stresses.

w kN/m
HA A B When the prestress force, P is applied at the
P
element centroid, the stress will be uniformly
distributed across the cross-sectional area. If the
VB axial force is applied above or below the neutral
VA
L axis there will be a secondary bending moment
created. The product of prestress and tendon
wl eccentricity e is the secondary moment Me=Pe. This is
V= used to counteract the major axis bending moment
S.F.D. (kN)

2 due to loading, the object being to have little or no

tension in the concrete and hence no cracks.
Uncracked concrete sections are stiffer and less
prone to durability problems.

Bending stress, fm is calculated using the equation of

B.M.D. (kNm)

Mz
fm = simple bending. z is the vertical distance from the
I yy neutral axis to the plane where stress is being
calculated, usually the top or bottom. Iyy is the major
wl 2 axis second moment of area.
M=
8
A.F.D. (kN)

P where P is the prestressing force and A is the

fn = element cross sectional area.
A
P

Mzt P Pezt
f mt = fn = f met =
I yy A I yy fmax

zt
+ + =
zb
Mzb Pezb fmin
f mb = f meb =
I yy I yy
primary bending stress axial prestress secondary bending stress combined direct stress
stress

There are several complications to consider, the maximum (compressive) stress should never
exceed 0.6fck and the minimum (tensile) force should be zero in class 1 structures, or almost zero in
class 2 structures (but still no cracking).

© Jonathan Haynes version 3.0 Page 109 of 239

32900 Bridge Engineering
MSc Structural Engineering

Post-tensioned concrete is designed by checking stresses at SLS loading. This would include

• final Frequent action level using combination action ∑G k +ψ 1Qk1 +ψ 2Qk 2 + ... + P
• transfer condition (after stressing operations) when there is minimal permanent action and
may be supporting only the units self weight but under full prestress ∑G k ,transfer +P

• final Quasi-permanent action level using combination action ∑G k + ψ 2 Qk1 +ψ 2 Qk 2 + ... + P

but since only construction loads and thermal loads have a non-zero value of ψ2 the variable
actions may be ignored, reducing the combination action to ∑G k +P
In most cases the quasi-permanent condition would not be an onerous stress check (since
transfer should be worse) but would be the loading at which long term deflection is checked.

3.4.3 Design Considerations

The force developed in a steel tendon will reduce with time, for several reasons. The net result is
that in the long term, the prestress will reduce to KP, where K is between 0 and 1.0 so it must be
incorporated into long term design calculations. When all losses are expressed as percentages,
K = 1 − ∑ losses
This is the inward movement of the jack wedges after release, it only
affects the required jack force and not the losses. There is a zone of
reverse friction extending from the wedges, which should not extend to
midspan (assuming that is the location of maximum flexure.
Draw-in of end
The additional tendon extension required is given by; eadd =
(Po + Px )S
blocks
2 Aps E s
where, Po and Px are the jacking force at end and prestress force at
midspan respectively, S is half span, Aps is the area of prestressing steel, Es is
the Young’s modulus of elasticity of the steel.
This is the friction which must be overcome to shorten a tendon in a
curved duct, it only affects the required jack force and not the losses.
There are two contributing factors – the tendon profile curvature and
variation from the intended profile.
Short-term

P x
Duct friction The additional tendon force required is given by; Po =
 µx 
−  + kx 
 R 
e
where, 1/R is the radius of curvature of the tendon profile; k and µ are the
tendon manufacturers coefficients for variation from profile and friction
respectively; x is the distance along the beam, typically at midspan.
This is the axial shortening of the concrete when subject to stressing, which
reduces the stretching of steel tendon and reduces the force in the
tendon.
Po
Concrete Remaining prestress, P ' =
elastic Aps  e 2 Ac 
1+ αe 1 + 
shortening Ac  I yy 

where, P’ is the prestress force after friction and draw-in losses, Ac is the area
of concrete, Ecm is the Young’s modulus of elasticity of concrete at transfer,
αe is the modular ratio Es/Ecm and e is the eccentricity of the tendon.

© Jonathan Haynes version 3.0 Page 110 of 239

32900 Bridge Engineering
MSc Structural Engineering

This is the short term contraction of the concrete whilst curing and is based
upon empirical coefficients for shrinkage per unit length, dependant upon
Concrete
humidity and age at transfer.
shrinkage
Loss, ∆P = ε s E s A ps

This is the long term contraction of the concrete whilst curing and is based
upon empirical coefficients for shrinkage per unit length, dependant upon
relative humidity (R.H)and age at transfer.
Concrete
E s A ps  
creep 1 + e 2 Ac  P'
Long-term

Loss, ∆P = ϕ
1.05 Ac Ecm  I yy 

where, φ is the creep coefficient.
A stress relieving heat treatment process in tendon manufacture is the
cause of a reduction in the prestress force over time. Manufacturers
therefore provide relaxation values from a 1000 hour load test; for class 2:
5% low relaxation tendons the loss in prestress will be no more than 5%.
Steel tendon
relaxation Loss, ∆P = γ relax k 70 k P '

where, γ relax is the relaxation coefficient calculated from EC2or supplied by

the manufacturer; k70k is the manufacturers specified loss percentage at
70% of the characteristic strength.

Values of steel strain s for N.W.C at ≤50% R.H.

2 Ac
u
100 200 300 500
550 x10-6 470 x10-6 410 x10-6 390 x10-6
Ac is the cross sectional area of concrete, u is the exposed perimeter
of cross section.
Values of creep coefficient ϕ for N.W.C at ≤50% R.H., loaded to ≤0.45fck at age

Age at
2 Ac
loading (days)
u
100 200 300 500
3 4.6 4.0 3.8 3.6
7 3.8 3.5 3.2 2.9
28 3.0 2.8 2.6 2.3
100 2.7 2.5 2.3 2.1
Ac is the cross sectional area of concrete, u is the exposed perimeter of cross section.

The calculation of stresses may therefore be summarised as follows:

Mzt KPezt Mzt KP KPezt

ft = + + − ≤ 0.6 f ck
Final frequent action

-Tension +Compression
I yy I yy I yy A I yy

+ + =
Mzb KP KPezb
Mzb KP KPezb
I yy A I yy fb = − + − ≤ f ctm
primary bending stress axial prestress secondary bending stress
I yy A I yy

© Jonathan Haynes version 3.0 Page 111 of 239

 32900 Bridge Engineering
MSc Structural Engineering

Mz t P Pez t
Mzt Pezt f 't = + + − ≤ f ctm
-Tension +Compression I yy A I yy
I yy I yy
Transfer action

+ + =
Mzb P Pezb
Mz b KP KPez b
I yy A I yy fb ' = − + + ≤ 0.6 f ck
primary bending stress axial prestress secondary bending stress I yy A I yy

Post-tensioned beams are rarely, if ever, subject to the same bending moment over their full
length. For this reason the eccentricity of the tendon should change – no eccentricity where
there is zero bending moment, and maximum eccentricity at maximum bending moment. The
need to limit compressive and tensile stresses complicates the selection of eccentricity and
prestressing force, so a Magnel diagram is used. This plots the possible combinations graphically.

 fI  e
 M −  t yy   4 3
I  zt  1
e ≥ yy +   …1
Azt  K P 2
  zone of possible solutions

 f I 
 M +  b yy  
I yy  z b 1
e≥− +  …2
Azb  K P emax
  1

  f 'I  I yy
 M −  t yy  
I  zt  1 Azt
e ≤ yy +   …3
1
Azt  K P
 
I yy
P
  f 'I  − possible range of 1 / prestress
M −  b yy  
I yy   z b 1
Azb
e≤− +  …4
Azb  K P
 

Tendon prestress should be roughly 60% GUTS at preliminary design stage and never exceed 75%.
φd
The maximum tendon eccentricity at midspan is, e max = y b − c − φ s − where c is the concrete
2
cover, φs is the diameter of the shear reinforcement and φd is the diameter of the tendon duct.
The tendon profile should follow the bending moment diagram shape. P and e are known at the
point of maximum bending moment, and P is constant along the beam so e must be varied to
suit the eccentricity equations 1-4. The equations can be simplified for the points where bending
moment is zero (simple supports), at midspan equations 1-4 apply. Once the maximum and
minimum limits at the end and midspan have been found the permissible tendon zone can be
drawn:

© Jonathan Haynes version 3.0 Page 112 of 239

32900 Bridge Engineering
MSc Structural Engineering

I yy f t I yy
e≥ + …1a
Azt KPz t
I yy f b I yy emin
e≥− + …2a
Azb KPz b neutral axis

I yy f t ' I yy emin eend

drape
e≤ − …3a
Azt KPz t emax
eCL
I yy f b ' I yy
e≤ + …4a emax
Azb KPz b

The eccentricity at the midspan eCL is known, and the eccentricity at the end eend is then chosen
considering It is best to have the tendon at a about mid-depth to help with the design of the
anchor blocks but as high as possible for shear resistance.
Rather than calculate the profile of the zones it is usual to assume a simple quadratic function
between the limits, of the form y = ax2
 e − eCL  2
y =  end x

 (0.5 L )
2

where x is the distance along the beam, L is the span and y is the drape of the duct profile from
the end eccentricity (not the neutral axis).

3.4.4 Design Checks

Although SLS stress checks are normally onerous for flexure at in-service action levels, lifting
reinforcement must be designed for ULS.
The ULS moment and shear capacity must be checked as follows:
Flexural capacity of a stressed concrete member depends on whether the tendon is grouted
after tensioning. Unbonded Tendons cannot create composite action with the concrete, so EC2
1-5 states that the increase in stress in an unbonded tendon at ULS should be taken as 100
N/mm2. Bonded Tendons are more complex and analysis is based on strain compatibility, in a
similar fashion to RC elements. The following equations are obtained from a simplified stress block
which limits concrete stress to 0.57fck over 80% of the compression zone, and steel stress to 0.78fpk.

Moment of resistance, M Ed = 0.456 xzf ck b = 0.78 zf pk A ps

0.78 f pk Aps
Lever arm, z = d − 0.4 x where x = and d is the effective depth to the steel tendons.
0.456 f ck b
Bridon 0.1% Tensioning loads (kN)
standard Area, Tensile proof
prestress Dia. Aps strength GUTS load
strand (mm) (mm2 ) (N/mm2) (kN) Fp0.1(kN) 0.65GUTS 0.75GUTS 0.75GUTS 0.80GUTS
7-wire strand 15.7 150 1770 265 228
single strand 7 38.5 1670 64.3 73.9 41.8 45.0 48.2 51.4
single strand 6 28.3 1770 50.1 44.1 32.6 35.1 37.6 40.1
single strand 5 19.6 1770 34.7 30.5 22.6 24.3 26.0 27.8

Maximum relaxation after 1000 hours : 60%GUTS=1%, 70%GUTS=2.5%, 80%GUTS=4.5%

Shear is dealt with using the same principles outlined under reinforced concrete elements. The
reaction locations in post-tensioned beams would normally be enlarged to create solid end
blocks. This allows rotationally stiff seating on bearings and enhances shear capacity. Shear
should be checked using ULS action combinations.
© Jonathan Haynes version 3.0 Page 113 of 239
32900 Bridge Engineering
MSc Structural Engineering

( [ ]
The basic shear capacity is given by, VRd 1 = k1τ Rd 1.2 + 40 ρ1 + 0.15σ cp bw d )
Aps + As N Ed
k1 = 1.6 − d ≤ 0.6 , ρ1 = ≤ 0.02 and σ cp = ≤ 0.2 f cd
bw d Ac
where τRd is the basic shear strength, bw is the web thickness, As is the area of untensioned
reinforcement, σcp is the axial compressive stress in the concrete due to prestress.
Concrete Grade C40/50 C45/55 C50/60
τRd (N/mm2) 0.41 0.44 0.48
Shear enhancement may invoked by taking the design shear force VEd at d from the support.
Account may be taken of the shear enhancement due to prestress by deducting Psinθ , the
vertical component of prestress;
Vsd = VEd − γ p P sin θ where γp = 0.9 and θ is the small angle between horizontal and the
inclination of the tendon.
As
If Vsd < VRd1 then provide minimum shear reinforcement given by, = ρ wbw sin α
s
For high yield steel ρw = 0.013 for fck ≤ 35N/mm2 or ρw = 0.016 for fck ≥ 40N/mm2
If Vsd ≥ VRd1 then provide shear reinforcement designed for Vsd − VRd 1 in accordance with the
rules for RC elements.
f ck
The maximum shear capacity is given by, VRd 2 = 0.3k 2νf ck bw d where ν = 0.7 − ≥ 0.50
200
 1.5σ cp 
If σ cp ≥ 0.27 f ck then k 2 = 1.671 −  otherwise k2=1.0
 f ck 

If Vsd ≥ V Rd 2 then a larger section size is needed.

VRd 2
If Vsd <
then s ≤ 0.8d or 300mm
5
V 2VRd 2
Maximum link spacing is given by; If Rd 2 ≤ Vsd < then s ≤ 0.6d or 300mm
5 3
2V
If Vsd ≥ Rd 2 then s ≤ 0.3d or 200mm
3
When checking deflection of post-tensioned sections consideration should be given to unloaded
and loaded conditions.

© Jonathan Haynes version 3.0 Page 114 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.5 Box Girder Decks

The box girder bridge has become a common structural form, offering both graceful lines and
often being the only deck capable of spanning long distances with adequate stiffness. The
major advantages are:
• Economy of material since the form places material at great distance from the centroid
• Reduced foundation and pier requirement because longer spans are possible
• High torsional stiffness due to the closed box form which can be easily curved on plan and
elevation
• Many construction options – progressive launching, span-wise launching and balanced
cantilever are the most common
• Easy to thread through or over existing structures, making it popular in crowded city centres
• Popular for use in suspension and cable stay bridge decks because of high torsional stiffness
• Aerodynamically stable when formed into an inverted aerofoil profile.

3.5.1 Historical Background

Wrought iron plate became readily available in the UK by 1835 but quality improved
dramatically in 1856 when Bessemer produced his iron converter, and yet again in 1865 when
the Seimen’s process was introduced.
Box girder bridges developed because Fairbairn and Telford needed to get heavy steam trains
to Holyhead, via the Menai Straits between north Wales and Anglesey. Link suspension bridges
had been developed for the coach road to Holyhead but the weight difference was significant
and a stiffer form was required. After experimentation and many failures, the wrought iron box
girder of the Conwy and Jubilee bridges was used successfully. What they did not realise was
that their bridges were only designed for concentric forces but luckily the railways they carried
could impose little or no torsion du to their fixed position.

Conwy Railway Bridge, (Fairbairn) 1849 Britannia Bridge, Menai Straits (R. Stephenson) 1850

There was no further notable development until after World War 2 when Germany was short of
construction material. Welding only became a reliable process in the 1940’s, so when Leonhardt
was designing Rhine river crossings (which reused earlier bridge piers) he utilised riveted box
girder decks (with external web stiffeners). In the 1950’s welding had developed far enough to
be used reliably in bridges and by the 1960’s stiffeners had moved inside the box. Understanding
of the torsion developed by eccentric loading was still far from adequate.

Cologne-Deutz Bridge (Leonhardt) 1948 Theodor-Heuss Bridge (Leonhardt) 1957

Telford and Fairbairn designed for bending stress. Leonhardt also accounted for torsion since he
provided diaphragms at close centres – as he would for a plate girder. Into the 1960’s engineers
produced more economical designs and developed better construction methods. Eventually
© Jonathan Haynes version 3.0 Page 115 of 239
32900 Bridge Engineering
MSc Structural Engineering

fate caught up with efficient box girder designs and in the early 1970’s a series of failures
prompted a moratorium on the use of box girders until research had filled the knowledge gap.
This culminated in Professor Horne’s CIRIA 3 guide. The problem was that boxes with insufficient
transverse diaphragms also suffer distortional warping which grossly increases the shear stresses –
plane sections do not remain plane.

3.5.2 Stress Analysis of Box Girders

A box girder which contains regular transverse diaphragms will behave in a similar manner to
simple tube subject to torsion – there will be a mixture of pure bending and pure torsion, which
can be analysed using simple bending, shear flow and simple torsion theories. This is classically
done by dividing the torsional moment into components of vertical loading which have the
same net effect.

0.5 0.5P 0.25P 0.25P 0.25P 0.25P

P
0.25P 0.25P

0.25P

0.25P

0.25P
0.25P
+ = + + + + +

0.25P 0.25P
eccentrically loaded warping shear
bending torsional shear
box girder
Mz Tr d 3φ
fm = τt = τ w = − EC m
I yy I xx dx 3

VA ' z '
τs = frequent stiff few or no stiff
I yy t diaphragms diaphragms

The symmetric bending of box girders is adequately described by simple bending theory and the
equation of shear flow, for direct and shear stresses respectively. However, loadings which do not
act through the shear centre of the box girder will induce torsion. If transverse panel stiffeners are
provided inside the box at close centres, it is possible to calculate the torsional shear stress using
(St. Venant) simple torsion theory only. If the distance between transverse panel stiffeners is large,
or they are only provided at supports then there will be additional torsional warping stresses
which will distort the cross section shape (plane sections do not remain plane). This will manifest
as additional direct bending and shear stresses.
Box girders which include wide flanges will also develop deck plate warping which results in
additional longitudinal bending stress in the top flange, which peak at the unrestrained flange
edges. This is due to differences between the shear strain at the web-flange joint and the flange
edge regions (the flexible flange edge sheds load when buckling). This can be accounted for in
design using the effective width concept, be which converts the actual flange width into an
equivalent design width. The effective width is maximised where plane sections remain plane (at
midspan in a simply supported beam) and minimised where the greatest warping occurs (at the
rotating support in a simply supported beam). The effective width is also related to the
concentration of the loading.

© Jonathan Haynes version 3.0 Page 116 of 239

32900 Bridge Engineering
MSc Structural Engineering

0.5P 0.5P

1.0

Effective width, be
B
+
uniform loads

top flange curls 0.5

concentrically
loaded box girder B
point loads

effective width
of top flange, be 0 0.5 1.0 1.5
Aspect ratio, B
L
longitudinal bending
stress in top flange

L top flange distorts

in shear
B

The direct and shear stresses due to eccentric vertical loading and horizontal deck loading may
be conveniently separated for the purposes of simplistic design.

ev
P P

H H
eh
+ + + = T
+

torsion about the shear centre

My y T = Pe v + He h
fm =
I zz
z Mz
fm = direct bending stress
I yy

T
τs = A’ area of section above
2 BDt ft level under investigation

z’ VA ' z '
τs = shear stress
D I yy t w
T
τs =
B 2 BDt fb

3.5.3 Failure Criteria

Once direct and shear stresses have been calculated, it is necessary to determine whether the
material is satisfactory. This is achieved using principal stresses. The principal stress is the maximum
direct stress at the point in a structure being assessed. This will always occur in the direction
where shear stress is zero. The maximum shear stress will occur in a direction at 45 degrees to the
principal stress.

© Jonathan Haynes version 3.0 Page 117 of 239

32900 Bridge Engineering
MSc Structural Engineering

shear stresses act

σz
perpendicular to
longitudinal (x-x) axis τ xz
the stress element locates the
σx σx
point being investigated
direct (axial + bending τ xz
stresses act parallel to
longitudinal (x-x) axis σz

This is best shown graphically by plotting Mohr’s circle of stress, which plots direct stress against
shear stress. The principal stresses correspond to the direct stresses where the circle intersects the
abscissa.

σz shear
Max shear element stress τ
τ xz Stress element τ max
σ2 τ
σ1
τ max
σz
σx σx σ
θ σ2 σx σ1 direct
45o
σ1 stress
τ max −τ
σ2
τ xz
σz Principal element Mohr’s Circle of stress

Mathematically, the principal stresses may be calculated using the following:

1 σ x +σ z
maximum principal stress, σ 1 = + (σ x − σ z )2 + 4τ xz 2
2 2
σx +σz 1
minimum principal stress, σ 2 = − (σ x − σ z )2 + 4τ xz 2
2 2
2τ xz
principal stress element rotation angle, Tan 2θ =
σ x −σ z
σ1 − σ 2
maximum shear stress, τ max =
2
It was found that elastic materials do not fail at the maximum principal stress but in accordance
with a more complex limiting value. Several investigators have postulated elastic failure criteria
and the two most commonly used are credited to:
Tresca & Guest who suggest there is a maximum shear stress yield criterion. Failure occurs when
maximum shear stress equals the maximum shear stress at failure in pure tension, or
f y ≤ max .{σ 2 − σ 1 , σ 1 , σ 2 }

von Mises, Hencky and Maxwell (though Hencky and Maxwell are commonly forgotten about)
who suggest there is a maximum shear strain energy yield criterion. Failure occurs when
maximum shear strain equals the maximum shear strain at failure in pure tension, or

f y ≤ σ 12 + σ 22 − σ 1σ 2

Most FEA packages will automatically output von Mises stress plots so that the values may be
compared to a yield stress limit for the material.

© Jonathan Haynes version 3.0 Page 118 of 239

32900 Bridge Engineering
MSc Structural Engineering

tangent to
σ2 failure point

direction of
σy plastic strain
Von Mises
yield Loci progression
failure point

σy
σ1
σy
σ
Strain hardening
σy Perfect plastic section is plastic – calculate non-
σy linear deformations
Strain softening
section is elastic – calculate
ε stresses from deformations

Von Mises failure criterion will also be used in non-linear material FE analysis, whereby the
programme will take a stress-strain relationship for the material and load step the structure until
the yield loci is reached, then determine the plastic strain trajectory from the loci and continue
the analysis load steps along a post-yield relationship. The most common post-yield relationship is
perfect plastic (a simplification of the mild steel stress-strain relationship) however, soils will often
follow a strain softening relationship.

3.5.4 Design of Plate Structures Loaded Transversely (out-of-plane)

The global (longitudinal) effects of wheel loads can be considered by using influence lines. The
local (transverse) effects of wheel loads can be considered by using influence surfaces.
Alternatively, grillage or whole bridge FEA may be used.
It is important to remember that plates loaded out-of-plane which deform more than half their
thickness will not be accurately represented by small deflection theory. So non-linear geometry
must be included in the analysis process.
A laterally loaded plate which is continuous with adjacent panels which deflects out-of-plane
will develop an in-plane membrane stress which stiffens it against further deflection, until the
critical buckling load is reached and crumpling occurs quickly.
Axial buckling of plates
Since box girders are formed from plate elements, they must be designed in accordance with
different rules to beam elements. The characteristic behaviour of struts and ties still apply, so
tension elements may be permitted to reach yield stress and compression elements will buckle
before reaching yield stress.

axial stress distribution in

buckled shape of
compression flange
compression flange

design axial stress distribution

beff in compression flange

© Jonathan Haynes version 3.0 Page 119 of 239

32900 Bridge Engineering
MSc Structural Engineering

The compression face of a box girder predominantly carries in-plane axial compression and will
buckle by forming surface ripples. The lowest mode of buckling will form a half wave across the
box width and at least one full wavelength along the box length between diaphragm stiffeners.
The distribution of axial stress across the compression flange is not uniform because the flange-
web joint is much stiffer than the mid-flange region (which displaces during buckling and sheds
load). To account for this, the design of plate struts uses the effective width concept and a
uniform design stress distribution.

σy an axially compressed slender

plate behaves in the same
manner as a slender strut, it will
exhibit elastic response until a
critical load is reached then
σcr yielding will set in (due to
bending stresses caused by
lateral deflection) so that
capacity reduces with
increasing lateral deflection
εy

Buckling of a flange plate may be checked using the following interaction equation;

N Ed M y , Ed + N Ed e y M z , Ed + N Ed e z  σ m, Ed σ cr ,c 
η1,mod = + + +  ≤ 1.0
 f 
Aeff f yd Weff , y f yd Weff , z f yd  yd σ cr ,c − σ Ed 
π 2 Et 2
where σ cr ,c = is the elastic critical column buckling stress for unstiffened plates and a
12(1 −ν 2 )a 2
is the shorter panel dimension. Aeff is obtained assuming axial stresses only and Weff is obtained
assuming bending stresses only.

ineffective internal element ineffective outstand element

centroid of gross section

ey movement of centroid
centroid of effective section

Yielding of a flange plate may be checked using the following interaction equation;
2 2 2
 σ x , Ed   σ z , Ed  σ x , Ed   σ z , Ed  τ Ed 
  +  −   + 3  ≤ 1. 0
 f yd   f yd   f yd   f yd   f yd 
where, σx is the longitudinal direct stress in the flange plate calculated using effective section
properties, allowing for plate buckling and shear lag; σz is the transverse direct stress in the flange
plate; and τ is the in-plane shear stress in the flange, taken as the torsional shear stress plus half
the shear stress at the web-flange joint.
Shear buckling of plates
Buckling behaviour in thin plates is not confined to axial compression. Compression zones will
develop due to bending and shear action also, all of which manifest as out-of-plane
deformation. This is usually confined by adding stiffeners and therefore reducing panel size and
hence reducing the size of deformations – increasing the buckling capacity. Simplistically the
shear buckling capacity of a web plate is τ cr td

© Jonathan Haynes version 3.0 Page 120 of 239

32900 Bridge Engineering
MSc Structural Engineering

2
Kπ 2 E  t 
where τ cr =   is the elastic critical shear buckling stress for unstiffened plates and d
( )
12 1 − ν 2  d 
is the depth of the web panel. K is a coefficient which relates to the edge restraint conditions;
K=5.34 for panels where all four edges are simply supported and K=8.98 for panels where all four
edges are fixed. The design rules for plate girder webs may be applied to box girder web plates.

a web panels divided by

fa
intermediate stiffeners (like web
stiffeners in plate girders) behave
as a series of sub panels
when subject to axial compression
the whole sub panel will buckle
out of plane

fb
when subject to bending only the
compression zone will buckle out
of plane

when subject to shear the plate

will develop an elastic shear
τ
buckling mode when loaded to
the critical shear stress τcr (as
shown) but when overloaded will
form a diagonal tension field
which provides post-buckling
reserves of strength

3.5.5 Calculation of Effective Compression Flange Width

For compression flanges which are not enhanced with stiffeners, the following procedure may be
used to establish the effective flange width, which should be used in the global analysis. This will
account for the effects of shear lag in the compression flange.

Bc Bt
Bt beff = β bo where, bo = or Bc
2

L3
β2
LE=2L3
L2 0.5L

β1
LE=0.7L2
L1 0.5L 
B effective LE=0.25(L1+L2) 
flange
width, β BMD LE=0.85L1  
deflected shape & 
effective length 
compression flange direct
 uniform compression throughout stress distribution options
 linearly varying compression throughout
 linearly varying tension and compression in internal element
 predominantly tension, some compression in outstand element
predominantly compression, some tension in outstand element

© Jonathan Haynes version 3.0 Page 121 of 239

 32900 Bridge Engineering
MSc Structural Engineering

σ2 
σ1

σ1
σ2
 
 σ1 σ2 σ1
σ2 σ1 σ2

σ1
σ2 σ1 σ2 σ1
σ2


Value of ψ = σ 2
Value of kσ Value of kσ
σ1
Outstand compression elements
Internal compression elements

ψ =1 4.0 0.43

8.2 0.578 σ 2 ≥ σ1
1 ≥ψ > 0
1.05 + ψ 0.34 + ψ Compression +ive
Tension -ive
ψ =0 7.81 1.70
2
0 ≥ ψ > −1 7.81 − 6.29ψ + 9.78ψ
1.7 − 5ψ + 17.1ψ 2
ψ = −1 23.9 23.8
− 1 ≥ ψ > −3 5.89(1 −ψ )
2

σy b
t 235
dimensionless slenderness ratio, λp = = where ε=
σ cr 28.4ε kσ fy

effective cross-sectional area, Ac,eff = ρAc where Ac is the gross cross-sectional area of the plate,
and ρ is the plate buckling reduction factor, taken from the table below.
Internal compression elements (Bt) Outstand compression elements (Bc)

if λ p ≤ 0.5 + 0.085 − 0.055ψ then ρ = 1.0 if λ p ≤ 0.748 then ρ = 1.0

λ p − 0.055(3 +ψ ) λ p − 0.188
if λ p > 0.5 + 0.085 − 0.055ψ then ρ= if λ p > 0.748 then ρ=
λ p2 λ p2

Aeff = Ac ,eff β where Ac,eff is the effective area of the compression flange due to plate buckling
and β is an effective flange width reduction factor, taken from the table below.

© Jonathan Haynes version 3.0 Page 122 of 239

32900 Bridge Engineering
MSc Structural Engineering

Value of κ Value of β
κ < 0.02 β =1
1
sagging β = β1 =
1 + 6.4κ 2
0.02 < κ ≤ 0.70 1
β = β2 =
hogging  1  2 α 0*bo
1 + 6 κ −  + 1.6κ where, κ=
 2500κ  LE
1
sagging β = β1 =
5.9κ
κ > 0.70 Ac ,eff
1 α 0* =
hogging β = β2 = and
8.6κ bot f
 0.025 
Any simple end β =  0.55 +  β ≤ β1
 κ  1
Any cantilever β = β2

EN 1993-1-5 contains further rules for longitudinally stiffened plates in section 4.

3.5.6 Calculation of Design Stresses Due to Torsion and Warping

For design purposes, the shear stresses due to torsion may be approximated using the following
equations which are specific to the joint between web and flange, the stresses at an outstand of
the top flange may be linearly extrapolated.

Stresses between flanges and web

DT
Bc Bt
Torsional shear stress at joint of bottom flange and web, τ Twb =
IT
4 A02 Bt + Bb
Torsional constant, I T = where, A0 = D
∑b 2
t
D + 2
T B  DT
Torsional shear stress at joint of top flange and web, τ Twt =  b  3
 Bt   2B 
I T 1 + c 
Bb  Bt 

For design purposes, the direct stresses due to warping may be approximated using the following
equations which are specific to the joint between web and flange, the stresses at an outstand of
the top flange may be linearly extrapolated.

Stresses between flanges and web Warping direct stress at joint of bottom flange and web, due to a
concentrated torque;
Bc Bt T zLD T zLD
σ Dw = if β LD ≤ 1.0 σ Dw = if β LD > 1.0
Bt I yy β LD Bt I yy
zt  KL4 
0.25
 Kx 4 
0.25
+ 24 DYt RD
D T β LD =  D  K= β x =  

 EI yy  Bt3  EI yy 
Where, x is the distance along the longitudinal axis; LD is the distance between
diaphragms.
Bb The effect of multiple concentrated torques is calculated using a beam on
elastic foundation factor;

∑ σ Dw = σ Dw ∑ Pn (cos β x − sin β x )e − βx

© Jonathan Haynes version 3.0 Page 123 of 239

32900 Bridge Engineering
MSc Structural Engineering

Bb + Bt
RD =
 Bb  2 DYt d   B  D d 
Bb   + 1 − BbVD  2 + b  Yt + 1
 Bb + Bt  DYw Bt   Bt  DYw Bt 
Where, d is the depth of the web in its plane; DYt is the flexural rigidity of the top flange per unit length of span; DYw is
the flexural rigidity of the web per unit length of span; VD is given below.
 DYt d  B  
  2 + b  + 1
VD =  DYw Bt  Bt  

 
3
 Bt  D d  Bb  Bb    DYt  Bb  
2

 + 1 1 + 2  Yt  1+ +  +   
 Bb    DYw Bt  Bt  Bt    DYw  Bt  
  

3.5.7 Design of Intermediate Diaphragms

The design of intermediate diaphragms may be approximated using the following equations for
plate shear or bracing force. Additionally, the diaphragm must be stiff enough to resist in-plane
distortion, this should be adequate if the calculated stiffness, S exceeds the tabulated minimum
value.

Diaphragm Strength and Stiffness

Bc Bt B B

D D
D 2 2
Lb
Lp
Bb
plate thickness, tp

shear stress in plate diaphragm, force in X-brace, force in V-brace,

T TL p Bb TLb Bb
τD = FB = FB =
2 BDt p 4 BDBt 2 BDBt
stiffness of plate diaphragm, stiffness of X-bracing, stiffness of V-bracing,
2
2 2
Gt p L δ K EA δ Kb b EAb L2pδ b2 K
S=
p b S= S=
2 Ap LD L p LD 4 L3b LD

Bt + Bb 4 DB where Ab is the cross-sectional area of the

B= Lp = D2 + B2 δb = bracing element, and LD is the distance
2 KBb L p between diaphragms.

Minimum stiffness, S
βLD 3.0 2.0 1.5 1.0 0.8 0.5 0.3
Point torque, σDw 0 0 2 20 50 500 2000
Uniform torque, σDw 0 0 10 100 200 200 200

© Jonathan Haynes version 3.0 Page 124 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.5.5 Construction Sequence

Box girder bridges offer a high degree of stability both laterally and longitudinally, particularly
during construction if the construction sequence is well planned. This was one of the reasons that
they were the form of choice when constructing inner city fly-over’s or bridging inaccessible
voids. There are many possible construction sequences but the three most popular are
progressive launching, span-wise launching and balanced cantilever methods.
Progressive Launching

Stage 1: Construction of piers, caissons and temporary piers

The off-shore towers require sheet piled cofferdams to create dry areas for the foundation works. Piled pontoons are
required in the spans, as dictated by the cantilever capacity of the box girder.

Stage 2: Jacking of first section

The worst temporary case is likely to be just before landing at the first support. There may be a number of back spans
under uplift conditions (kentledge may be needed). The launching nose may need a hydraulic section to lift the
sagging cantilever beam.

Stage 3: Box girder becomes a continuous beam

As additional box girder sections are constructed and added, they are then hydraulically jacked into the span.
Special sliding surfaces may be needed (such as PTFE) or include simultaneous vertical jacking

Stage 4: Completed bridge subject to traffic loading

The final arrangement may include the most onerous loading during the bridges design life but the
location of maximum tension may have also been a location of maximum compression during erection

There are a range of stress regimes during the launching process.

When prestressed precast sections are used, there will normally be a serrated shear key between
sections. Stressing strand for launching forces will normally be removable (in accessible ducts).
When steel sections are used they are normally fully welded on site. This requires tenting to
produce an acceptable welding environment, and subsequent full non-destructive testing of
welds.

© Jonathan Haynes version 3.0 Page 125 of 239

32900 Bridge Engineering
MSc Structural Engineering

Balanced Cantilever

Stage 1: Pier construction

Foundations, piers and temporary shuttering for pier head box girder. A climbing crane is often used so
that it can be used to lift subsequent sections into place.

Stage 2: Cantilever box girder construction

Crane in sections of box girder – match cast with shear connection. Stress concrete sections back to
pier head, or butt weld steel plate sections. The pier must resist the out-of-balance moment from one
section. The stressing cable is set to suit a cantilever bending moment.

Stage 3: Completion of central span

The bridge becomes a continuous cantilever. Need to transfer stressed cables to suit new bending
moment profile.

Stage 4: Complete bridge

End spans now need transfer of stressing cables to suit new bending moment profile. End spans will sit
on bearings.

Clearly, box girders require significant consideration of the many temporary loading conditions
which occur before the permanent works are complete. This, and the complex internal stress
regime are the main reasons why there were a number of box girder bridge collapses in the late
1960’s and early 1970’s.
Very often the most onerous design cases will be during launching, lifting or load transfer
operations. This is also when effective lengths can be difficult to evaluate due to possible stress
reversal and achieving lateral restraint during construction can require special bearings which
are replaced for permanent condition.
Box girder design is complex but often offers the best options for overall economy and minimal
temporary works cost, compared to other bridge forms of equal spans.

© Jonathan Haynes version 3.0 Page 126 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.6 Bridge Fittings

3.6.1 Bearings
A bearing is a component used to transfer localised forces between bridge parts, usually
between the deck and abutment or, pier. They are expensive items, both in capital cost and
construction detailing. They are a common source of problems – loss of articulation induces
undesirable forces in other bridge components; a design which accommodates ease of
replacement is therefore desirable.
The majority of force transferred through the bearing will be due to traffic loading. Generally this
will be vertical shear force but there will also be horizontal braking effects. Under the normal
working thermal range, expansion or contraction must be accommodated by the deck.
The possible range of movement to be accommodated by the bearing must be calculated from
thermal expansion of the deck, horizontal movements of pier head and construction tolerances.
Elastomeric bearings are formed from stiff rubberised compounds such as EPDM, which can
distort anticlastically without reducing in volume. When reinforced with horizontal mild steel
plates or bars, they form laminated bearings. These bearings accommodate deck translation
and rotation by distortion. Plane sliding bearings can only accommodate sliding in the horizontal
plane. They are usually formed of a steel body with a low friction polymer sliding surface, such as
PTFE. Guides may be provided to restrict movement to one direction, forming a pot bearing.
Roller bearings may be as simple as a cylinder of special steel contacting steel plates cast into
the deck and abutment. They can be used to form the fixed bearing (fixed location, not fixed
against rotation).
Refer to BS EN1337: Structural bearings and PD 6703: 2009 Structural bearings – guidance on their
use, for technical performance specifications.

position fixed on plan in position not fixed on plan in

longitudinal axis only, expansion either axis, expansion permitted
permitted across deck along and across deck

position fixed on plan in

position fixed on plan in both
transverse axis only, expansion
axes
permitted along deck

© Jonathan Haynes version 3.0 Page 127 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.6.2 Concrete Corbels

Once a bearing has been chosen, it is important to size the supporting structure to avoid
overstress. In steel structures this will normally involve sizing stiffeners and welds. In concrete
structures there must be adequate edge distances to avoid shear or bearing failure. This is
particularly important if a corbel is used (commonly in abutments), general dimensional
guidance is detailed below.

FEd × ac + H Ed × ah
force in tension reinforcement / concrete compression zone, Ftd = Fcd =
d
Ftd γ ms  f  f
reinforcement area required, Ast = concrete stress limited, FEd ≤ 0.425bd 1 − ck  ck
f yk  250  γ mc

a3 a1 a2+∆2
bearing plan dimensions
a1 x b1
ac bearing contact stress,
FEd FEd
σ Ed =
a1b1
HEd

ah
Ftd
1.0 ≤ Tan θ ≤ 2.5
check reinforcement anchorage length

θ 45 o ≤ θ ≤ 68 o
r

Fcd
horizontal links,
cos θ d
As , h,link ≥ 0.5 Ast

vertical links,
0.5 FEd
As ,v ,link ≥
f yd

FEd

σEd / fcd < 0.15 0.15-0.40 > 0.40

a1 90 110 140
Steel a2 5 10 15
Concrete a2 15 25 35
a3 15mm ≥ cover
check radius of main tension reinforcement ∆2 10mm ≤ L / 1200 ≤ 30mm
to limit concrete bearing stress ≤ fcd L is the slab or beam span

A copy of the CCL Stressing Bridge Bearings brochure is available on Blackboard.

© Jonathan Haynes version 3.0 Page 128 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.6.3 Parapets
The purpose of a parapet wall is to prevent bridge users from inadvertently leaving the bridge by
a route not intended by the designer. On some bridges they also prevent head-on collisions of
vehicles travelling in opposite directions.
In the past the DoT specified the form and design of bridge parapets (P1 to P6 designations) but
now any proven design is permitted (see TD 16/06 DMRB Vol 2 Section 2 Part 8 for test
requirements).
EN 1317 Parts 1- 6 : Road restraint systems.
Containment EC Vehicle speed
Old Designation Test vehicle
Level Designation (km/h)
80
Normal N1 1.5T car

P2 (80) 110
Normal N2 1.5T car

Higher H2 P1, P2(113), P5 13T bus 70

Very high H4a P6 30T rigid HGV 65

A parapet wall must be provided at the free edges of the bridge deck, and at a kerbed central
reservation such as on motorway or dual carriageways.
A parapet wall may be solid, such as masonry or reinforced concrete, or elemental such as
aluminium or steel railings. The base connection of all parapets must be capable of withstanding
the impact of a moving test vehicle.

Parapet walls must extend beyond the bridge deck onto the approach far enough to prevent
vehicles leaving the overbridge and descending down any embankment. This is of particular
importance when a vehicle may descend onto a live railway line.

© Jonathan Haynes version 3.0 Page 129 of 239

32900 Bridge Engineering
MSc Structural Engineering

3.6.4 Movement Joints

Movement joints in bridge decks are similar in principle to their counterparts in ground bearing
floor slabs for buildings. Though there are many proprietary products available, there are two
basic types:
1. Deck construction joint – to prevent cracking of the deck waterproofing at a concrete deck
slab construction joint (concrete will contract after casting).
2. Full deck expansion joint – to allow unrestrained thermal movement in the deck slab, without
compromising the waterproofing system or the running surface of the bridge.
The CIRIA guide 543 provides several generic details used for bridge expansion and construction
joints in the UK, depending upon the required movement range. DoT requirements are specified
in documents BD 33/94 and BA 26/94.
For expansion ranges greater than 40mm, proprietary engineered designs are available.

multiple elastomeric element on

cantilever finger joint

metal runner joint

The joint must permit movement of adjacent bridge parts without over-spanning (steel) or over-
stretching (rubber) the joint bridging material. There must also be security of the waterproofing
layer, particularly at edges of concrete. It is important not to apply waterproofing to sharp
concrete edges, so fillets must be either added to shuttering or added with post-cast mortar.

Waterproofing tuck

Commonly available waterproofing systems may be:

Sheet membrane - bituminised fabric, polymer membrane which provides a factory quality
product, it must be heated and overlapped when installed on site, is the cheapest option and
probably easiest to install competently.
Liquid - bitumen based product, which is workmanship quality dependant and therefore least
likely to be applied in a uniform thickness.

© Jonathan Haynes version 3.0 Page 130 of 239

32900 Bridge Engineering
MSc Structural Engineering

Sprayed liquid - usually two-part epoxy or polyurethane, which is very fast to apply and bonds
best to most materials, it is also the most expensive.

Contraction range 0 – 3mm Expansion range 0-10mm

Expansion range 10-20mm Expansion range 20-40mm

Expansion range >40mm

© Jonathan Haynes version 3.0 Page 131 of 239

32900 Bridge Engineering
MSc Structural Engineering

Structural Element Design - Conclusion

Things to do now :
1. Review the Powerpoint lectures on Blackboard.
2. Attempt the Self Assessment Exercises below. Then check your answers against the
solutions on Blackboard.
3. Have a rest.

4. Prepare your group submission for Assignment 3.

Self Assessment Exercises

Using the loading calculated in section 2.0 for a 40.0m long, two span highway bridge;

5.4m 1.0m abutment and

pier head beam 1.00m
0.35m 5.50m wide
0.15m
5.00m
0.30m

4.5m 1.00m
2.25m 1.00m
SECTION ACROSS DECK
steel plate 7.00m 7.50m
girder
0.50m 2.5m ELEVATION ON PIER
0.70m

5.00m

SECTION ON ABUTMENT

20.0m 20.0m

0.40m thick simple support

5.0m

R.C. pier
G.L.

3.0m
ELEVATION ON BRIDGE

1. Produce a bearing layout, justifying your choice of bearing.

2. Select a grade of concrete and design any necessary reinforcement for the deck slab.
3. Using S355 steel, estimate the plate girder size and check it is adequate.
4. Assuming the central pier is rigidly connected to its foundation, design the pier wall.
5. Check the abutment for overturning, sliding and bearing pressure. Design the stem, base and
corbel reinforcement. Assume soil design data to be C=50 kN/m2, φ=350, γ =22 kN/m3.
6. Use beam on elastic foundation theory to determine the pier foundation settlement, shear
force and bending moment diagrams. The results of a 12 inch plate load test at foundation level,
were as follows:
Load (kN) 0 5.0 10.0 15.0 20.0 25.0 29.6 29.7 29.8
Settlement (mm) 0 0.06 0.114 0.40 1.01 2.20 4.50 9.7 19.5

© Jonathan Haynes version 2.0 Page 132 of 239

32900 Bridge Engineering
MSc Structural Engineering

Further Reading
BCSA Publication No 51/10. Steel Bridges. (2010). A practical approach to design for efficient
fabrication and construction. 3rd edition. London: British Constructional Steelwork Association.
BCSA Publication No 41/05. Steel Bridges. (2005). Steel Details. London: British Constructional
Steelwork Association.
BSI. Eurocode 3 – Design of steel structures. Part 1- 1: General rules and rules for buildings. BS EN
1993-1-1:2005. Corrigendum February 2006 and April 2009.
BSI. Eurocode 3 – Design of steel structures. Part 1- 5: Plated structural elements. BS EN 1993-1-
5:2006. Corrigendum April 2009.
BSI. UK National Annex to Eurocode 3: Design of steel structures. Part 1- 5: Plated structural
elements. NA to BS EN 1993-1-5:2006
Benaim, R. (2008). The design of prestressed concrete bridges. Concepts and principles. London:
Taylor & Francis.
Broms, B.B. (1964a) Lateral resistance of piles in cohesive soils. J.ASCE, 90, SM2, 27-64.
Broms, B.B. (1964b) Lateral resistance of piles in cohesionless soils. J.ASCE, 90, SM3, 123-156.
Brooker, O., Jackson, P.A. & Salim, S.W. (2009). Concise Eurocode 2 for Bridges. CCIP-038.
Camberley: The Concrete Centre.
Chapman, J.C., Dowling P.J., Lim, P.T.K. & Billington, C.J. (1971). The structural behaviour of steel
and concrete box girder bridges. The Structural Engineer, 49 3.
Concrete Bridge Development Group (2006). An introduction to concrete bridges. Camberley:
The Concrete Society.
Dowling, P.J. & Moffatt, K.R. Shear lag in steel box girder bridges. (1975). The Structural Engineer,
53 10.
Fryba, L. (1996). Dynamics of railway bridges. London: Thomas Telford.
Hamill, L. (1999). Bridge hydraulics. London: E&F.N. Spon.
Hendry, C.R. & Iles, D.C. (2010). Steel bridge group: guidance notes on best practice in steel
bridge construction. SCI Publication 185. 5th edition. Ascot: The Steel Construction Institute.
Hetenyi, M. (1946) Beams on Elastic Foundation: Theory with Applications in the Fields of Civil and
Mechanical Engineering. US: University of Michigan Press.
Highways_Agency. (2001b). BD 21/01. Design Manual for Roads and Bridges. Volume 3 Highway
Structures - Inspection and Maintenance. Section 4 Assessment. Part 3 The Assessment of
Highway Bridges and Structures. Norwich: H.M.S.O.
Horne, M.R. (1977). Structural actions in steel box girders. CIRIA Guide 3.
Iles, D.C. (2004). Design guide for steel railway bridges. SCI Publication 318. Ascot: The Steel
Construction Institute.
Iles, D.C. (2004). Design guide for box girder bridges. SCI Publication 140. Ascot: The Steel
Construction Institute.
Iles, D.C. (2010a). Composite highway bridge design. In accordance with Eurocodes and the UK
National Annexes. SCI Publication 356. Ascot: The Steel Construction Institute.
Iles, D.C. (2010b). Composite highway bridge design: worked examples. In accordance with
Eurocodes and the UK National Annexes. SCI Publication 357. Ascot: The Steel Construction
Institute.
Iles, D.C. (2012). Design of composite highway bridges curved on plan. SCI Publication 393.
Ascot: The Steel Construction Institute.
Maisel, B.I., Rowe, R.E. & Swann, R.A. Concrete box girder bridges. (1973). The Structural Engineer,
51 10.
Roads and Transport Association of Canada. (2004). Guide to bridge hydraulics. 2nd Edition.
London: Thomas Telford.
SCI. (1995). Bridge design to the Eurocodes. Simplified rules for use in student projects. SCI
Publication RT1156. Ascot: The Steel Construction Institute.

© Jonathan Haynes version 2.0 Page 133 of 239

32900 Bridge Engineering
MSc Structural Engineering

Smith, G.N. (1982). Elements of soil mechanics for civil and mining engineers. 5th edition. Collins
Professional Books: Letchworth.
Terzaghi, K. (1995). Evaluation of coefficients of subgrade reaction. Geotechnique, 5(4), 41-50.
The British Standards Institute. PP1990: 2010 Extracts from Structural Eurocodes for students of
structural design. London: HMSO.
Thorburn, S., Burland, J.B., Cooke, R.W., Gould, H.B., Larnach, W.J. & Wex, B.P. (1989). Soil-structure
interaction. The real behaviour of structures. London: The Institution of Structural Engineers.
Tiller, R.M. (1973). Concrete footbridges. London: Cement & Concrete Association.
Timoshenko, S.P., & S. Woinowsky-Krieger. (1959). Theory of Plates and Shells. 2nd Ed. New York:
McGraw-Hill.

© Jonathan Haynes version 2.0 Page 134 of 239

32900 Bridge Engineering
MSc Structural Engineering

4. Arch Bridges
This section will deal with the ancient and enigmatic arch bridge. It will introduce the difficulties
of analysing a complex, indeterminate structure and the methods available to assess existing
arch bridges and design new ones. Some leading research on the topic of brickwork masonry
arch bridges will also be explored.
You should also review: 4 Arch Bridges.ppt on Blackboard
Principal References: Heyman, J. The masonry arch.

The arch vault is the oldest true bridging form, being a 3D curved strut. Although its origins are in
antiquity, it was Roman engineering which delivered the technology used extensively throughout
the industrial revolution, to construct the most prolific medium span bridge in history.

The arch bridge can now be constructed in any material, and to astonishing spans like the
concrete New River Gorge arch which reaches 924m. Although concrete and masonry arches
are invariably barrel vaults, timber and steel arches are normally individual ribs or even arched
trusses.

4.1 Arch Bridge Analysis and Design

Whatever the arch form used, the structural behaviour is always that of a strut, which implies that
capacity will be governed by buckling. This may be global or local buckling of elements, which
will not be evident from a simple elastic analysis of the structure.
A strut is a structural element subject to compression. Short (or stocky) struts may be compressed
until yield stress is reached, this is called the Squash Load and is largely independent of cross-
sectional shape. However, useful struts are relatively long in comparison to their cross-sectional
area and suffer buckling before achieving yield stress. Buckling is a phenomenon caused by
lateral displacement of an axially loaded strut. Long (or slender) strut buckling capacity is highly
dependant upon length and cross-sectional shape.
Effective length may be defined as the distance between points of contraflexure (the point
where slope changes direction), which is largely dependant upon end fixity.
P P P P
Z

LE=0.7L
L Y
LE=1.0L LE=0.5L Y

LE=2L
Z

pinned virtual pin at fixed

support point of support
contraflexure

© Jonathan Haynes version 3.0 Page 135 of 239

32900 Bridge Engineering
MSc Structural Engineering

Due to the problems ensuring fixity at a joint, it is customary to use design effective lengths of 0.85
for a pin-fix strut and 0.7 for a fix-fix strut.
Slenderness is a measure of a structural element’s propensity to buckle (or displace sideways). It
may be defined mathematically as the quotient of effective length and radius of gyration, or
LE I
λ= where, i =
i A
Since struts may be made of non-symmetrical sections, the effective length and radius of gyration
may be different about the principal axes (y-y and z-z). It is usually necessary to calculate
slenderness about each axis and select the largest slenderness ratio for design.
A strut buckles about the axis which has the largest slenderness ratio. However, because an arch
is curved in one plane, the direction of buckling displacement can be anticipated and the
element orientation chosen to suit this.

strut displaces
laterally when 3m
buckling P
PE
LEz = 3m
PE
4m

LEz = 4m LEy = 4m

if λz>λy then an I-section two storey strut, if λy>λz then

minor axis restrained against lateral major axis
buckling will displacement on the minor buckling will
happen first axis at first floor level. happen first
There are several buckling theory solutions which provide formulae for strut capacity. Notably,

π 2 EI fy A
Euler, PE = 2
and Rankine, PR = where a = 0.0001
LE 1 + a λ2
1 1  π 2E
Perry-Robertson, PPR = A 
2
[f y ]
+ f cr (1 + η ) −
4
[ ]
f y + f cr (1 + η ) − f y f cr  where, f cr = 2
2

λ
 
fy is the material yield stress and η = 0.003λ .
A graph of these three solutions for various slenderness ratios reveals that the (earliest) Euler
theory is unsafe for stocky struts but Rankine and Perry-Robertson provide very similar capacities.
The Perry-Robertson solution is generally embodied in modern design codes.
Elemental arches, of discrete steel or timber struts must be assessed for buckling as described.
Lateral restraint bracing must be positioned to ensure the effective length assumed is achieved
in the finished structure. Plate arches, of masonry of concrete vaults will generally be too stocky
to buckle out-of-plane. This leads to very efficient solutions since the vault must fail by buckling
about its major axis.

Elemental rib elements will buckle in the manner of Plate vaults are enormously stiff about their minor
a multistorey column and therefore require bracing. axis. Buckling is usually about the major axis.

© Jonathan Haynes version 3.0 Page 136 of 239

32900 Bridge Engineering
MSc Structural Engineering

Strut Buckling Curves

350
The three buckling
300
curves shown are for a
fy mild steel section,
250 Rankine constant =
Stress f (N/mm )

Euler curve
2

0.0001 and Robertson

200 constant = 0.003.

150
Note that the Euler
Rankine curve solution diverges
100
significantly from reality
50 for λ < 120 and must be
Perry-Robertson curve limited to yield stress for
0 λ < 85
0 50 100 150 200 250
Sle nde rne ss λ
The analysis of arches depends largely upon the support assumptions made by the designer.
Engineers must be pragmatic when considering arch supports, since it is likely that the bridge will
be founded on soil which will settle under loading. This means that the assumptions of a fixed
support is unlikely to ever be reflected in reality; further, some allowance must be made for
differential settlement which may occur vertically and horizontally.
Shape can have a profound affect upon the reactions from an arch. Semi-circular and elliptical
arch shapes will (theoretically) produce only vertical reactions. Whilst segmental arches will
produce significant horizontal reactions, which increase as the arch profile flattens. The
segmental arch was particularly popular amongst Victorian engineers because it often minimises
overall construction depth.

semi-circular segmental elliptical gothic

The statically determinate three-pinned arch may be analysed using equilibrium alone. This
model should only be used if a specifically designed pin is to be included in the structure. It has
advantages for short to medium span structures because the pin ended elements can be
premanufactured and assembled on site without moment joints. The gothic arch geometry is
particularly well suited to three-pinned arch ribs.
The two-pinned arch is statically indeterminate, and is best analysed using a plane frame analysis
since temperature and settlement can be readily incorporated as load cases. As masonry
arches remain at low levels of stress during their life and have relatively large self weight, there is
a moderate level of continuity at the supports. However, as collapse approaches the arch
buckles sideways and hinges form in the bedjoints. It is therefore conservative to assume pinned
support conditions.

© Jonathan Haynes version 3.0 Page 137 of 239

32900 Bridge Engineering
MSc Structural Engineering

4.1.1 Steel Rib Design

According to Roark & Young (2002), the use of simple bending theory for curved ribs will be
acceptable if the radius of curvature exceeds ten times the arch thickness (there will be a linear
distribution of bending stress across the section). EN1993-2 D.3.1 offers design rules for the
practical analysis and design of steel rib arches.
Arch rib critical axial flexural buckling capacity, Ncr may be calculated using standard Eurocode
capacity formulae for steel elements but using effective length factors specific to the form of rib
being designed.
2
π 
In-plane buckling, N cr =   EI y where s is half length of arch, Ι is second moment of area, β is
 βs 
the effective length factor.
2
π 
Out-of-plane buckling, N cr =   EI z where L is projected length of arch.
 βL 
EN1993-2 6.3.3(1) offers a simplified overall buckling check approach for bridge elements.

N E ,d
+ Cm
(M y , Ed + ∆M y , Ed )
≤ 0.9
N b , Rd M b , Rd
where My,Ed is the first order moment, ∆My,Ed is the moment due to shift in the centroidal axis due to
coincident axial and bending, Cm is the equivalent uniform moment factor.

0 -41 0 -10 0 -51

Neutral + =
ecc =Shift in NA
Axis

∆M y ,Ed = N × ecc
41 0 31
0 0
Axial Stress (N/mm2 ) Bending Stress (N/mm2) Combined Stress (N/mm2)

The effective length factor (β) of steel arch ribs:

In-plane buckling Tied Arch.

For parabolic forms subject to vertical loading

β
s
L
r m= −1
p

p
r
L
L

© Jonathan Haynes version 3.0 Page 138 of 239

32900 Bridge Engineering
MSc Structural Engineering

Out-of-plane buckling of free standing arch β = β1 β 2

deck on r/L
β2 = 1.0 β1
top of 0.05 0.10 0.20 0.30 0.40
arch
Constant Iz 0.5 0.54 0.65 0.82 1.07
deck on qh Varying Iz 0.5 0.52 0.59 0.71 0.86
β 2 = 1 − 0.35
hangers q

qp
deck on β 2 = 1 + 0.45
posts q

q is the total load, qh is the load in the hangers, qp is the load in the posts

Out-of-plane buckling of circular arch with radial loads

L π 2 + α 2k r
β =π α for 0 ≤ α ≤ π
(
2 L π 2 −α 2 ) α

2(1 + ν )I z
k= L
IT

In-plane buckling of arches Two pinned arch

For parabolic forms subject to vertical loading
β
r
Three pinned arch 0,1 0,2 0,3 0,4 0,5 L

r
0,1 0,2 0,3 0,4 0,5 L
Rigid arch

β
r
0,1 0,2 0,3 0,4 0,5 L

© Jonathan Haynes version 3.0 Page 139 of 239

32900 Bridge Engineering
MSc Structural Engineering

Out-of-plane buckling arch with bracing & end portals

β h

α
bE I for end portal column
η= c c for end portal beam b
hEb I b

h
1
hr hr is the average hanger length ×
sin α

To prevent snap-through of the arch rib, EN1993:2 D3.2(3), offers the following stability check:

K is taken from Table D5 (below)

r EA I is y-y second moment of area
K≤L
12 EI y A is the cross-sectional area
L is projected length (span) of arch

Values of K r/L
0.05 0.075 0.1 0.15 0.20
Pin supports 35 23 17 10 8
Fixed supports 319 97 42 13 6

4.1.2 Concrete Vault Design

There are no direct Eurocode rules for vault design.
Analysis should model the arch rib with curvature tolerance set out in the National Structural
Concrete Specification. Ideally non-linear geometrical effects should be considered.
EN1992-2 offers design rules for the practical analysis and design of concrete sections. An arch
rib should be considered in the same way as a columns subject to axial compression, bending
moment and shear. Consideration of effective length may follow the guidance for steel ribs.

© Jonathan Haynes version 3.0 Page 140 of 239

32900 Bridge Engineering
MSc Structural Engineering

4.2 Masonry Arch Bridges

Masonry arches are traditionally sized to contain the line of thrust (centroid of compression)
within the middle third of the arch barrel, this ensures that tension is not developed and cracking
is minimised.
parapet wall above surfacing spandrel wall built on barrel edge waterproofing on extrados

wing wall beyond springing coping stone road surfacing arch barrel

fill material mass masonry

abutment formed
in brickwork with
footing
extrados

brickwork rings intrados

form arch barrel

abutment

springing or skewback

Sectioned view of the common elements of a masonry arch bridge

4.2.1 MEXE Assessment

Analysis of an indeterminate arch is a complex process, so a simplified (and fast) capacity
assessment method was developed following WW2. It was based upon the assumptions of
Pippard’s elastic model:
• parabolic arch shape with pin supported unyielding abutments
• the fill and barrel are of equal density
• the strengthening effect of fill and spandrel walls are ignored – all capacity is from the arch
barrel
• at the provisional axle load, a limiting compressive stress of 1.4N/mm2 exists in the crown masonry

The Military Engineering Experimental Establishment (MEXE) method comprises the calculation of
a provisional axle load (PAL) that relates to the performance of a standard arch barrel using
either a nomogram given in BA16/97 or the equation:

PAL = 740
(d + h)2 ≤ 70
L1.3
where PAL is measured in Tonnes, L is the span (in metres), d is the thickness of the arch barrel
adjacent to the keystone (in metres), and h is the average depth of fill (in metres) at the quarter
points of the transverse road profile, between the road surface and the arch barrel at the crown,
including any road surfacing. It may be unsafe to assume h > d, even if the condition exists in the
real bridge.
load width 305mm

h
152mm

heff, fill effective

thickness at crown
h
45o rq rc

d, arch barrel L L L
thickness 4 4 2

© Jonathan Haynes version 3.0 Page 141 of 239

32900 Bridge Engineering
MSc Structural Engineering

To account for the significant variation in masonry arch construction, this value is adjusted to give
a modified axle load (MAL) that relates to the specific bridge,

MAL = PAL ⋅ Fsr Fp FM Fj FcM

where the modification factors are:
−0.7091 0.6
L
Fsr = 2.6727 
 rc − rq 
F p = 2.3 (Fb d ) + (F f h)
 Fm = F j = Fw Fd Fmo
 rc   rc  d +h
Span:rise factor (Fsr) adjusts for the weakness of relatively flat rings.
Profile factor (Fp) accounts for profiles which are not parabolic, where rq = 0.75rc
Material factor (Fm) where Fb is the barrel factor taken as 0.7 for brickwork masonry and Ff is the fill
factor taken as 0.7 or 0.5 for well compacted for weak material respectively.
Joint factor (Fj) where Fw is the width factor taken as unity for 6 mm bedjoints and reducing to 0.8
for joints thicker than 12.5mm, Fd is the depth factor which is unity for good condition filled joints
reducing to 0.8 for insufficiently filled joints, Fmo is unity or 0.9 for mortar in good or friable condition
respectively.
Condition factor (FcM) is a qualitative factor between 0.0 and 1.0 which is based upon
engineering judgement. It is intended to give the assessing engineer an opportunity to consider
factors which have not been considered by the standard assessment, such as impact damage
or ring separation. Ring separation has reduced capacity by 25-40% in controlled tests.
An Allowable Axle Load (AAL) is then calculated that accounts for the possibility of multiple axled
bogies loading the bridge. There is also a possible lift-off case, where multiple axled bogies do
not share the vehicle weight equally (e.g. at the crown of hump-backed bridges or bridges on a
tight horizontal curve).
AAL = MAL ⋅ A f
For a two axle bogie which cannot lift-off, Af is unity.
For a two axle bogie which could lift-off,
A f = 0.66 0 ≤ L ≤ 4. 0m
A f = 0.0181(L − 4) + 0.66 4.0 ≤ L ≤ 20.0m
For a single axle bogie which cannot lift-off,
A f = 1.0 0 ≤ L ≤ 4.0m
A f = 0.143(L − 4 ) + 1.0 4.0 ≤ L ≤ 7.5m
A f = 0.0385(L − 7.5) + 1.5 7.5 ≤ L ≤ 14.0m
A f = 1.75 14.0 ≤ L ≤ 20.0m
Once the AAL is calculated, the permissible gross vehicle weight can be established for the
bridge.
AAL (T) per axle Maximum gross Type of vehicle
vehicle weight (T)
Single bogie Double bogie
11.5 10 40 HGV 5 axle
11.5 9.5 32 HGV 4 axle
11.5 9.5 26 HGV 3 axle
11.5 18 HGV 2 axle
9.0 12.5 HGV 2 axle
7.0 10 HGV 2 axle
5.5 7.5 LGV
2.0 3 Van & car

© Jonathan Haynes version 3.0 Page 142 of 239

32900 Bridge Engineering
MSc Structural Engineering

4.2.2 Analysis of an Elastic Rib

Should a masonry arch bridge fail a modified MEXE assessment, it should be assessed using a
more advanced technique. Like all structures, masonry arches may be analysed using complex
FEA models if desired but this often raises more questions than are answered. It is preferable to
firstly resort to one of the following:
• An ultimate limit state analysis using ARCHIE M or RING, both of which are available free for
academic purposes (you need to be on the university network). These are proprietary
computer programmes which apply Heyman’s collapse mechanism method.
• Apply the MEXE method’s condition factors to a plane frame analysis of an elastic rib which
represents a unit width of the arch barrel.
The section properties of the rib should represent solid uncracked masonry. The rib should be
pinned at each springing and the geometry should represent the barrel centreline. There should
be at least twelve elements in the rib, including joints at quarter and third span.
For short span bridges, where only one axle of an HGV can fit on the bridge, the axle should be
positioned at one third span. Otherwise, a single span influence line can be used to identify the
most onerous axle locations.

h
2 d
1

load width
300mm

rc

h
L
3

b=h+1.5
effective load width

Axle loads should be spread transversely over a distance of h+1.5 (m) and longitudinally through
the surfacing and fill at a slope of 2:1.
Since the object of the assessment is to establish the allowable wheel load, the permanent
action (self weight of surfacing, fill and arch) should be applied as a separate load case, as a
variable intensity uniform loading. The variable action is then applied as an isolated uniform load.
The variable action should be progressively increased until the maximum compressive direct
stress (axial+bending) due to combined permanent and variable actions, equals the stress limit
of the masonry. The wheel load corresponding to this stress is the Limiting Variable Action (LVA).

LVA ⋅ F j FcM A f
The Allowable Wheel Load, AWL =
γ fL
where γfL = 3.4 for the first axle and 1.9 for subsequent axles (BD21/01).
It is difficult to establish the strength of old masonry but the table below indicates some
commonly accepted values.

© Jonathan Haynes version 3.0 Page 143 of 239

32900 Bridge Engineering
MSc Structural Engineering

Masonry type Masonry strength (N/mm2) 1:2:9 mortar No rings 1 2 3 4

Brick unit strength Masonry unit strength, fk d (mm) 102 215 327 440
London stocks 10 3.5
Fletons 20 5
Engineering B 50 9
Engineering A 70 11
Sandstone 140 14
Granite 220 16

To use the influence line, place the scaled vehicle axle

positions on the diagram and scale the coefficients.
Iterate until the maximum sum of coefficients is found.

Wheels in the + region will increase the bending moment,

+ whereas wheels in the - region reduce it.

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

_
0.8 0.9 1
The numbers 0 to 1 are proportions of the arch span.

4.3 Masonry Arch Bridge Research

Although the masonry arch is the oldest bridging form in use today, it is also the least well
understood. The analysis and design of masonry arches has never been codified in the UK, this
being a reflection of the level of dispute on the topic between engineers during the period of
extensive arch construction (Tilly, 2002) between 1700 and 1900. Ruddock (1979) presents a
historical treatise which gives considerable insight into the early field of masonry arch bridges, in
particular he recants the famous assertion of the polymath Hooke on the design of arches to be
little more than one of many single sentence solutions he gave for mechanics problems. Originally
in Latin, it translates as hangs the flexible line, so but inverted will stand the rigid arch. This of course
found popularity because of its visual clarity but few bridges were designed on this basis because it
had already become practice to construct much shallower and more visually appealing arches
than an inverted line or chain would suggest. Additionally, the masonry arch is not a flexible line
and must support more than its self-weight, so the analogy is of limited design use.
Some twenty years after Hookes observation, Gregory derived the equation of the hanging
flexible line, which he called the catenaria and we now know as the catenary. Further, he states
that an arch of any shape but the catenary will only stand if a catenary can be inscribed within
its thickness. McKribbins et al (2006) further relates that strictly the arch is an antifunicular
geometry of a particular set of loads and that the arch resists loading which distorts this
geometry by gravitational prestress.

where the line of thrust touches the extrados, a hinge will form in the barrel

a chain hanging under its self weight forms a classical three and four hinge collapse mechanisms are common
catenary, when loaded it is easily distorted.

© Jonathan Haynes version 3.0 Page 144 of 239

32900 Bridge Engineering
MSc Structural Engineering

In the eighteenth century Couplet and La Hire both further developed arch theory with
experimental tests. The outcome was what we would now recognise as limit or plastic analysis,
from which the line of thrust concept and collapse mechanisms originate. In the nineteenth
century Navier demonstrated that for an elastic section where plane sections remain plane,
tension may be avoided by containing the thrust within the middle third of the section – now
commonly know as the middle third rule. Twenty years later Barlow demonstrated that there is no
unique thrust line associated with a stable arch (there are many solutions). Some thirty years later
still, Castigliano reported that his theorem of minimum strain energy applied to arch structures, so
long as there is a linear relationship between load and displacement; this was demonstrated
experimentally by Pippard & Baker in 1957 (up to the point where hinges develop). Later Heyman
would demonstrate the acceptable assumption of the middle third rule by independent means.
Pippard (1952) originated the rigid block collapse analysis method (using Castigliano’s theorem)
which led to the MEXE method of masonry arch assessment. Amendments have been made to
the original formulation and further restrictions are proposed for particular types of bridge. Wang
& Melbourne (2010) and Wang, Melbourne & Tomor (2010) derive improved relationships for short
span or high span:rise geometries and identify potentially unconservative assessment outcomes
using the existing MEXE formulations.
Heyman (1966 & 1982) revived what we now know as mechanism analysis, being an elegant
and easily understood method. The assumptions made are:
• Sliding failure cannot occur since friction between the units is high enough to avoid slippage
• Masonry has no tensile strength. Although in reality there is a small tensile capacity, it is safe or
conservative to ignore it
• Masonry has infinite compressive strength. Forces are assumed to be transferred through the
masonry units rather than distorted by them. This is not true so it is unsafe or unconservative.
Conditions to satisfy are:
• External and internal energy must be in equilibrium
• A collapse mechanism must form
• Yield must not be exceeded.
The attraction of the mechanism method is that it is essentially a collapse, or plastic analysis. This
is advantageous (in comparison to elastic analysis) because it does not depend upon detailed
knowledge of the arch behaviour and is largely insensitive to geometrical changes. This is
supported in later work such as (Heyman, 1998) where assertions such as the cracked state is the
normal state of masonry and there is no unique equilibrium state for masonry are substantiated.
Many subsequent researchers have proposed amendments and refinements to Heyman’s work.
De Rubeis (1998) develops expressions for the geometrical safety factor, being the ratio of ring
thickness to minimum thickness for arch stability. Further, the author relates the geometrical
safety factor to a mechanical safety factor, being the ratio of minimum point load to total arch
weight.
Several modes of failure have been identified in masonry arches:
• Four hinge mechanism, progressive opening of alternately rotating cracks at positions of
sufficient eccentricity in the thrust line. Four hinges are required for collapse to occur. It may be
possible for a collapse mechanism to form with three arch hinges in combination with
considerable abutment movement (translation, rotation or a combination of the two) which
constitutes the fourth hinge
• Local instability (crushing failure) at a point of highly concentrated loading can precipitate
global failure. This type of failure is undesirable since it is effectively brittle – as crushing
progresses, the resisting area of masonry diminishes and instigates a progressive collapse
• Sliding blocks may develop between masonry units if insufficient friction is developed by the
mortar bond
• Snap-through (in the manner of a multi-span portal frame) is possible, especially in centrally
loaded arches of high span:rise ratio.
Melbourne et al (2007b) and Melbourne, Tomor & Wang (2005) have shown that multi-ring
segmental brickwork masonry arches are likely to fail prematurely by ring separation
(delamination) under certain cyclic loading conditions. This is the case when the arch rings are

© Jonathan Haynes version 3.0 Page 145 of 239

32900 Bridge Engineering
MSc Structural Engineering

constructed without headers – bricks which pass through the inter-ring mortar joint. As the
maximum vehicle axle weight is increasing, it is possible that many arch bridges are nearing their
endurance limit and a predictable means of repair is needed.

masonry block sliding local crushing

snap through

brittle hinges form in the masonry to

permit gross rotation of blocks in a
ring separation plastic collapse mechanism

According to Melbourne et al (2007a) there has been an extensive programme of research over
the past decades, which considered some aspects of masonry arch behaviour. Early full scale
tests undertaken for TRRL (Melbourne & Walker, 1990) were laboratory based full size, complete
bridges, loaded statically (monotonic) to collapse. However, these tests were of limited used for
code based guidance (it is difficult to identify the proportion of capacity attributable to each
part of the bridge) and later testing programmes have concentrated upon structural elements of
the bridge such as the barrel. Investigation has also concentrated upon the long-term effect of
traffic (cyclic) loading and the effect of deteriorated masonry on the fatigue life of the bridge. A
second generation test programme is underway at UoS which seeks to investigate the behaviour
of the arch barrel and backfill material under laboratory test conditions.
tests with monotonic loading then cyclic loading
establish the capacity and fatigue life of the barrel
only, the series of blue box tests seek to establish full scale test results are used to validate many small scale
the capacity and fatigue life of the barrel with (cheaper) tests which establish the effect of individual capacity
compacted backfill material (cohesive or or fatigue parameters, this data is used to validate FEA models
cohesionless) of a parametric study.
All of this can be used to predict the capacity of a real bridge
with defects such as mortar loss, masonry unit deterioration,
abutment displacement and ring separation.

3T

few full scale tests validate many small scale tests validate FEA models validate capacity predictions

Future testing programmes are likely to address the contribution of the spandrel walls, which
generally aid stability and offer restraint to the barrel free edge.

© Jonathan Haynes version 3.0 Page 146 of 239

32900 Bridge Engineering
MSc Structural Engineering

Backfill over the barrel extrados will generally increase the dead loading and therefore improve
stability. As the arch displaces sideways under non-concentric loading the backfill at the side
remote from the loading will offer restraint against arch movement (passive earth pressure will be
mobilised). However, at the loading side the backfill will contribute to the side-sway (active earth
pressure will be mobilised). Obviously there is a magnitude of difference in the displacement
required to mobilise these two earth pressures, so that the analysis is complex and there would
be little confidence in adopting the results generated in a structure where little is likely to be
known, with confidence, about its construction.
Interaction between the spandrel walls and backfill are likely, though this is also difficult to
quantify and further exacerbated by the possibility of lateral displacement of the spandrel wall.
Where present, wing walls are likely to increase the in-plane stiffness of spandrel walls by
restricting the base rotation of the spandrels.
Van der Cruyssen & Beeby (2000) identify developments of Heyman’s mechanism method
incorporating the effects of lateral earth pressure due to backfill over the arch barrel. Under long
term loading the factor of safety can be dramatically increased for low span: rise geometries
and reduced for high span: rise geometries. Design values for earth pressure at rest are given. Ng
& Fairfield (2004) also offer a modified mechanism analysis using backfill lateral earth pressure.

a four hinge collapse mechanism with fill

interaction

as the ring deforms under vehicle loading,

earth pressure on the approach side increases
the displacement

sway of the other side of the ring is restrained

by earth pressure

Masonry is known to behave differently to most engineering materials in terms of fatigue. It has
long been known that high cycle loading reduces the load value at first crack (hinge formation)
but that the final collapse loading is virtually unaffected. McKribbins et al (2006) suggest this is
due to low stresses being developed until the later stages of collapse mechanism formation and
the fact that material strength has a less significant role in the ultimate strength of arches than in
other common structural forms (geometry assists in load resistance). Current arch bridge
assessment documents simply indicate that it will be prudent to limit the regularly applied loads
to half the ultimate failure load.
Stresses in masonry can be redistributed due to creep. Under quasi-static loading the
compressive strength of brick masonry depends upon:
• the compressive strengths of the bricks and the mortar
• nature of the applied loading (static or dynamic)
• environmental conditions (wet or dry).
Under concentric loading quasi-static compressive strength is well documented in codes of
practice, under non-uniform loading the strength may be significantly greater (Hendry, 1990)
typically up to 20%. This may be accounted for by assuming plastic, parabolic or rectangular
stress distributions to determine the maximum induced stress.
Little information is available concerning the high cycle fatigue strength of brick masonry but a
value of approximately 50% of its quasi-static compressive strength is commonly taken. A lower
bound high cycle fatigue strength of dry brick masonry is given by the relationship:
(∆S × S max )0.5 = 0.7 − 0.05 log N
Su
where ∆S and Smax are the induced stress range and maximum induced stress respectively, Su is
the quasi-static compressive strength under similar loading conditions and N is the number of
constant amplitude load cycles.

© Jonathan Haynes version 3.0 Page 147 of 239

32900 Bridge Engineering
MSc Structural Engineering

Little information is available concerning the influence of moisture content on the strength and
stiffness of brick masonry. McKribbins et al (2006) report experimental work which indicated that
the presence of moisture can reduce the quasi-static compressive strength by 20%.
Singh-Curry (2006) summarises work undertaken at the UoS in producing a relationship to
describe the endurance limit of an arch:
E = 10 H R − m
where E is the cyclic endurance limit, R is the range of load which generates tensile stresses, H is
the vertical intersection on logN (cycles) ordinate and m is the slope of the best fit straight line on
the logN against logW (load) graph.
Humphrey (2007) reports that when arches are strained, axial stresses are induced in the
brickwork but since the rings are of different lengths the axial shortening will vary between the
rings. This difference in displacement between the rings induces longitudinal shear stresses so that
composite action can develop. Multi-ring arches without headers must transfer longitudinal shear
for composite action through the mortar bed only, so there is more propensity to ring separation
or delamination occurring before a collapse mechanism can develop. In arches constructed of
strong bricks and weak mortar, failure must occur in the mortar first but the bond between brick
and mortar is weaker still and cracking is therefore the most commonly observed failure. Since
shear is a shape distortional stress, greatest shear stresses accompany the greatest
displacements of the arch barrel. Logically, as permissible axle loads increase, so too will
displacements and shear stresses.
Under cyclic loading there is an endurance limit associated with each load level. Endurance
appears to drop exponentially as the applied load approaches the multi-ring arch capacity.

% static load

100

50

load cycle
endurance limit
100000
400000
50
mortar bond No load cycles
endurance limit 0

an endurance surface can be constructed for any particular arch

% mortar bond

Once ring separation occurs the composite arch barrel must resist loading as separate rings,
which will have a reduced direct and longitudinal shear capacity. This gives way to the
observation that high stress-low cycle loading is likely to generate a hinged collapse mechanism
failure but low stress-high cycle loading is likely to generate a ring separation failure in multi-ring
masonry arches.
Gilbert & Melbourne (1994) describe the development of a rigid-block method of analysis that
can be used by the analyst to determine the plastic collapse loads of masonry block structures.
In the case of most masonry arch bridges the lack of knowledge regarding material properties,
internal construction details and initial stress state indicates that the use of sensitive,
computationally-intensive finite-element methods may not be justified. It is in these cases
particularly that the relative simplicity and computational efficiency of the method described will
prove particularly useful. The method is now embodied in the Limit State Ring computer
programme.
Harvey (1991) explains the concept of the zone of thrust has long been used in assessing arch
bridges but the value of the idea is not well understood. The implications of strength and

© Jonathan Haynes version 3.0 Page 148 of 239

32900 Bridge Engineering
MSc Structural Engineering

elasticity in conjunction with stability are discussed in some detail. Each is attached to the idea
of the zone of thrust, and an approach to application is presented using simple, statically
determinate examples. He concludes that the zone of thrust is a flexible and discerning
analytical tool, that engineering judgment is always necessary and is more readily applied to the
results of multiple graphical analyses than to the input data for a FEA programme and finally that
the application of suitable factors will allow the approach to be used for both ultimate and
serviceability limit states. These principles are embodied in the ARCHIE-M computer programme.
Hogg & Choo (2000) report it is possible to obtain a full-scale collapse capacity from model
testing but deflections cannot be accurately scaled onto real structures from masonry models.
Melbourne & Alnuaimi (2001) investigate capacity change due to ring separation in cyclically
loaded multi-ring masonry arches. Fatigue load is defined as that which when repeated induces
material deterioration and reduction in load carrying capacity. The S-N (stress-number of cycles)
fatigue failure surface is introduced. Initially a test arch was loaded to first hinge formation and
the cyclic fatigue load set at half this value. A cycle frequency was set at 2Hz to avoid dynamic
inertial effects and simulate vehicle passage. The general conclusions were that cyclic loading
below 50% first crack load produced no deleterious effects; increasing the load above the first
crack load induced ring separation; arches which had ring separated suffered further
propagation even when the load was below first hinge load. This suggests that any multi-ring
brickwork masonry arch which does not contain headers and has developed at least one hinge
in its life will fail prematurely (compared to a four hinge collapse mechanism) by ring separation.

a fatigue failure surface can be constructed for any particular arch

low cycle, high stress loading

– four hinge collapse failure
logN
high cycle, low stress loading
– ring separation failure

logS

Melbourne & Wang (2005) examine the 3D behaviour of skew multi-ring brick arches subject to
point loading. Physical testing also explored strong and weak bricks and the effect of in-span
supports. The conclusions were that deflection is proportional to Young’s modulus of elasticity;
arch strength is not directly proportional to brick strength; point loads are distributed widely
across an arch barrel; FEA models will accurately predict failure loads but are sensitive to the
value of tensile strength used, which should be about 2% of compressive strength.
Melbourne, Wang, & Tomor (2007c) present an assessment procedure (SMART) which differs
philosophically from all previous load assessment techniques in as much as it takes a holistic
approach in considering all possible modes of failure that the structure may experience under
any given loading regime. Limit states are discussed and a new permissible limit state that is
specific for masonry is proposed. The method advocates consideration of the permissible
working loads, long-term behaviour and residual life of the bridge. This enables the assessing
engineer to prioritise conflicting maintenance demands on limited budgets. The method is based
upon recent research related to the long-term fatigue performance of masonry arch bridges
subjected to cyclic loading. It is important to note that once a fatigue crack has been
established, it will continue to grow under cyclic loading even when it is well below the load
which created the crack.

© Jonathan Haynes version 3.0 Page 149 of 239

32900 Bridge Engineering
MSc Structural Engineering

Masonry Research Papers and Texts

De Rubeis, A. (1998). On the definition of the geometrical safety factor of masonry arches. Paper
presented at the Second International Arch Bridge Conference. Venice. 6th-9th October 1998.
Gilbert, M., & Melbourne, C. (1994). Rigid-block analysis of masonry structures. The Structural
Engineer. 72(21), 356-361.
Harvey, W. J. (1991). Stability, strength, elasticity and thrust lines in masonry structures. The
Structural Engineer. 69(9), 181-184.
Hendry, A. W. (1990). Masonry properties for assessing arch bridges. Crowthorne: TRRL.
Heyman, J. (1966). The stone skeleton. International Journal of Solids and Structures, 2(2), 249-279.
Heyman, J. (1982). The masonry arch. Chichester: Chichester : Horwood.
Heyman, J. (1998). The assessment of strength of masonry arches. Paper presented at the
Second International Arch Bridge Conference. Venice. 6th-9th October 1998.
Hogg, V., & Choo, B. S. (2000). A study of scale effects in masonry arch bridges: is testing of large-
scale structures still necessary? The Structural Engineer. 78(5), 24-29.
Humphrey, D. G. (2007). Assessment of radial tying as a technique to rehabilitate arch bridges
exhibiting ring separation. M.Sc. Dissertation, University of Salford.
McKribbins, L. D., Melbourne, C., Nisar, S., & Gaillard, C. S. (2006). Masonry arch bridges: condition
appraisal and remedial treatment. London: C.I.R.I.A.
Melbourne, C., & Alnuaimi, M. M. (2001). The behaviour of multi-ring brickwork arches subjected
to cyclic loading. Paper presented at the 3rd International Arch Bridge Conference, Paris.
Melbourne, C., Tomor, A. K., & Wang, J. (2005). Modes of failure of multi-ring arches under fatigue
loading. Paper presented at the 5th International Conference on Bridge Management. The
University of Surrey. 11-13 April 2005.
Melbourne, C., & Walker, P. (1990). Load test to collapse on a full scale model six metre span
brick arch bridge. Crowthorne: TRRL.
Melbourne, C., & Wang, J. (2005). The 3D behaviour of masonry arches. Paper presented at the
5th International Conference on Bridge Management. University of Surrey. 11-13 April 2005.
Melbourne, C., Wang, J., & Tomor, A. (2007a). Structural Assessment of Masonry Arch Bridges Sixth
Framework Programme. Priority 6. Sustainable Bridges - Assessment for Future Traffic Demands
and Longer Lives (Vol. D4.7.1 Background Document).
Melbourne, C., Wang, J., & Tomor, A. (2007b). The analysis and assessment of masonry arch
bridges. In J. Bien, L. Elfgren & J. Olofsson (Eds.), Sustainable bridges - Assessment for future traffic
demands and longer lives. European 6th RTD Framework Programme. Wroclaw, Poland:
Dolnoslaskie Wydawnictwo Naukowa PAN.
Melbourne, C., Wang, J., & Tomor, A. K. (2007c). A new masonry arch bridge assessment strategy
(SMART). Proceedings of the ICE - Bridge Engineering. 160(2), 81-87.
Ng, K. H., & Fairfield, C. A. (2004). Modifying the mechanism method of masonry arch bridge
analysis. Construction and Building Materials. 18(2), 91-97.
Pippard, A. J. S. (1952). Studies in elastic structures. London: Edward Arnold.
Ruddock, T. (1979). Arch bridges and their builders 1735-1835. Cambridge: Cambridge University
Press.
Singh-Curry, D. (2006). The investigation of masonry arch bridges under monotonic and cyclic
loading. M.Sc Dissertation, University of Salford.
Tilly, G. (2002). Conservation of bridges. London: Gifford & Partners, Highways Agency, Spon
Press.
Van der Cruyssen, D., & Beeby, A. (2000). Lateral loading on masonry arch bridges. The Structural
Engineer. 78(12), 14-17.
Wang, J., & Melbourne, C. (2010). Mechanics behind the MEXE method for masonry arch
assessment. Proceedings of the ICE - Engineering and Computational Mechanics. 163(3), 187-
202.
Wang, J., Melbourne, C., & Tomor, A. (2010). Development of Pippard’s elastic method for the
assessment of short span masonry arch bridges. Paper presented at the 6th International
Conference on Arch Bridges, Fuzhou, China.
Wang, J., Haynes, B.J., & Melbourne, C. (2013). A comparison between the MEXE and Pippard’s
methods of assessing the load carrying capacity of masonry arch bridges. Paper presented at
the 7th International Conference on Arch Bridges, Trogir-Split, Croatia.

© Jonathan Haynes version 3.0 Page 150 of 239

32900 Bridge Engineering
MSc Structural Engineering

Arch Bridges - Conclusion

Things to do now :
1. Review the Powerpoint lectures on Blackboard.
2. Attempt the Self Assessment Exercises below. Then check your answers against the
solutions on Blackboard (where you will find the spreadsheet ArchCoordinates.xls for
arch coordinates and capacity assessments etc).
3. Have a rest.

4. Prepare your group submission for Assignments 2 & 3.

Self Assessment Exercises

1. Using the modified MEXE method, assess the load carrying capacity of a 6m wide single span
three ring engineering brickwork masonry arch which has the following geometry and condition:

Spandrel walls in poor condition due to

vehicle collisions. Some bedjoint shearing
evident. 300mm
Arch ring brickwork in good condition, all
joints filled but generally 14mm thick in
outer ring.
Approximately 1.0m rise in ‘hump-back’, 1.5m
some evidence of vehicle grounding on
carriageway and tyre rutting in asphalt. 6.0m

2. Prepare an elastic plane frame analysis model for a steel arch pedestrian bridge:

There are two separate arch ribs, 2.0m apart.

The deck is a 2.0m wide x 10mm thick mild
10.0m
steel plate with anti-slip bitumen coating.
Each strut onto the arch supports a deck
transverse beam and out-of-plane cross
40.0m
bracing.

Compare your results with the LinPro file on Blackboard. Create some feasible EC load
combinations and establish which is the most onerous for axial compression in the arch rib. Run
the file with a horizontal roller support at one arch springing and review the change in results.

Further Reading
Highways_Agency. (2004). BD 91/04. Design Manual for Roads and Bridges. Volume 2 Highway
Structures: Design (Sub-structures and Special Structures) Materials. Section 2 Special Structures.
Part 14 Unreinforced Masonry Arch Bridges. Norwich: H.M.S.O.
Highways_Agency. (2001). BD 16/97. Design Manual for Roads and Bridges. Volume 3 Highway
Structures: Inspection and Maintenance. Section 4 Assessment. Part 4 The Assessment of Highway
Bridges and Structures. Norwich: H.M.S.O.

© Jonathan Haynes version 3.0 Page 151 of 239

32900 Bridge Engineering
MSc Structural Engineering

5. Integral Bridges
This section will focus on those features that are of significance for the design of integral and
semi-integral bridges. The many positive attributes and few significant limitations will be
elaborated and discussed; in particular, points for consideration in evaluating the suitability of
integral bridges, their more complex analysis requirements and detailing will be covered.
You should also review: 5 Integral Bridges.ppt on Blackboard
Principal References: Burke, M.P. Integral and Semi-Integral Bridges.

5.1 The Development of Integral Bridges

The most common design concept for a road bridge traditionally consists of some type of
superstructure resting on an abutment at each end.
Due to natural, seasonal variations in air temperature, the bridge superstructure will change
dimension in its longitudinal and lateral directions. Lateral change is relatively small and can be
discounted. To accommodate this relative movement between superstructure and abutments
and prevent temperature-induced stresses from developing within the superstructure, the
traditional solution has been to provide expansion joints and bearings at each end of the
superstructure.
Although the traditional design works well in concept, experience indicates that expansion joint
and bearing details can be the cause of significant post-construction maintenance expense
during the service life of a bridge. Therefore, a structural form was developed to physically and
structurally connect the superstructure and abutments to create an integral bridge.

expansion joint to accommodate thermal movement approach slab

bearing
moment transfer capping beam

traditional bridge form integral bridge form

Clearly the fundamental concept of a jointless bridge is not new, since masonry arch bridges are
composed of an arch ring integral with masonry abutments. However, modern integral bridges
are a solution which seeks to use current economical construction methods combined with the
advantages of ductile materials.

5.1.1 Integral Bridge Arrangements

Integral bridges can be defined as bridges without joints. They span from one abutment, over
intermediate supports to the other abutment without movement joints in the deck. The
advantages of integral construction are greater durability and lower maintenance costs when
compared with jointed bridges.
A survey of the performance of 200 highway bridges in 1989 confirmed that most bridge joints
leak and consequently areas of the abutments, piers and deck soffits become stained and
contaminated with chlorides from road salts. The Highways Agency recommends that integral
construction should therefore be the first option for all bridge decks shorter that about 60m. There
are two main types of integral abutment bridge:
• Those with short stub-type abutments that sit on piles and support the deck beams or slab.

© Jonathan Haynes version 3.0 Page 152 of 239

32900 Bridge Engineering
MSc Structural Engineering

• Those with full abutment walls, sitting on piles, that retain the ground behind the wall as well
as support the deck beams or bridge slab.

single span with wall-type abutments multiple span with stub-type abutments

Typically these bridges have stub-type abutments supported on piles and a continuous bridge
deck from one embankment to the other. Foundations are usually designed to be small and
flexible to facilitate horizontal movement or rocking of the support.
With integral bridges thermal deck movements are accommodated by soil structure interaction
between the supporting piles and the surrounding strata. Deck loading is also affected by the soil
which acts as both load and support system to the piles upon which the structures are founded.
Specifying a series of spring supports along a pile to approximate soil behaviour is a commonly
used modelling method when the structural load effects are the main item of interest. When the
soil movement is of interest elastic continuum models must be used.
Integral bridges present a challenge for load distribution calculations because the bridge deck,
piers, abutments, embankments and soil must all be considered as a single compliant system.

5.1.2 Semi-integral Bridge Arrangements

Semi-integral bridges have no deck joints and thus have the advantages of full integral deck
construction. However, bearings are introduced below the deck and an end wall carries the
horizontal forces to the soil. The connection between the deck structure and abutment does not
transfer moment to the abutment; rather, the bridge deck is integral with any retaining structure.

approach slab

bearing which permits rocking

integral end screen wall, (beam articulation) and sliding
sealed against abutment (thermal expansion)

The use of bearings in semi-integral bridges must balance the reduced possibility of ingress by
degrading agents such as road salt against the long term maintenance problems of bridge
bearings. It is therefore imperative to design the bearings for ease or inspection and
replacement.

5.1.3 Limitation of Application

The superior economy of integral bridges is due to their ability, within a limited application range,
to satisfy all functional requirements with safety, durability and optimal economy. They are not
broadly applicable as their length is limited, to:
• minimise passive pressure effects
• limit bridge movements to those appropriate to articulation of approach slab joints.
They should not be used where:
• skew is greater than 30 degrees
• abutment piles cannot penetrate at least 3-4.5m of soil
• stability of sub-soils is uncertain or settlement may be significant.

© Jonathan Haynes version 3.0 Page 153 of 239

32900 Bridge Engineering
MSc Structural Engineering

Integral bridge design depends on idealisation and simplification of soil-structure interaction,

which does not accurately reflect the true pressure on abutments and piles, especially the effect
of long term pressure together with cyclic temperature changes.

5.1.4 Construction Methods

The two most common structural arrangements for integral bridges in the UK are composite steel
plate girders combined with high modulus steel sheet piling, or post-tensioned concrete beams
combined with insitu retaining piles (secant pile or diaphragm wall). Both configurations achieve
the same result and may employ essentially the same construction sequence.
The integral bridge is particularly well suited to situations where a new route may pass under an
existing route, by creating a cutting.

a) installation of piles to form retaining wall abutments b) deck beams and moment transfer cap beam

c) casting deck slab and excavation of underpass d) completion of approach slabs and highway finishes

This is peculiar in bridge construction since it involves little working at height.

5.2 Soil-Structure Interaction

The detailed study of soil-structure interaction is demanding for the relatively simple cases of
structure loaded soils. Integral bridges are fundamentally different since structural movements
accumulate to convert the as-built bridge into a soil loaded structure. It is worthwhile reviewing
some basic soil mechanics before exploring soil movement in integral bridges.

5.2.1 Effective Stress and Earth Pressure

Soil is a three phase material (it may contain solid particles, gas
and liquid water). Any gas present is not capable of
transferring stress. Water can transfer stress in all directions, and
is called the pore water pressure. Solid particles can transfer
stress where they touch each other; since this is the residual
capacity of soil when the water drains away (or when the soil is
u subject to shear, as water has no shear strength) it is referred to
as the effective stress.
The vertical stress which can be transmitted through a soil is the
sum of vertical components of the particulate stress (effective
stress) plus the pore water pressure, or σ = σ '+u .
A general categorisation of engineering soils is to consider them to be cohesionless or cohesive
in behaviour. Cohesionless soils such as sands and gravels may be round or angular in shape and
of particle size > 0.06mm. Cohesive soils such as clays are shaped like thin plates and of particle
size < 0.002mm. Silt is generally considered to be a cohesive soil but spans the particle size
between clays and sands, and exhibits the worst behaviours of both types of soil.
© Jonathan Haynes version 3.0 Page 154 of 239
32900 Bridge Engineering
MSc Structural Engineering

Behaviour of cohesionless soils is governed by body forces between discrete particles. When the
particles are loosely packed they can move over each other under load by rolling, which is only
restrained by frictional forces. When this happens the size of the voids reduces and the sand
densifies. When the particles are densely packed they must overcome their physical mechanical
interlock before movement can occur. When this happens the size of the voids will initially
expand, followed by a return to the original void ratio. The ability of the particles to move is also
dependant upon the vertical stress (or normal stress, usually composed of the weight of soil
above).

LOOSE sand void reduction shear stress, τ

mechanical
interlock
DENSE sand

DENSE sand temporary void expansion

sliding and
rolling friction
LOOSE sand

0.02
strain, ε

Since loose sand experiences void reduction under load there must be an increase in pore water
pressure, hence a reduction in effective stress and the ultimate shear capacity of the soil. Dense
sand will initially exhibit the same behaviour, however once mechanical interlock between the
particles is overcome there is a void expansion under load, reduction in pore water pressure,
hence an increase effective stress and the ultimate shear capacity of the soil.
For these behaviours to be exhibited the water in the soil must be confined (if it can leak away
quickly the pore water pressure will drop to one atmosphere). This is a function of soil
permeability, which generally increases with particle size (water flows through gravel easily).
Behaviour of cohesive soils is governed by inter-molecular forces. Clay is composed of
microscopic plate-like particles and one or more of the three clay minerals – kaolinite, illite and
montmorillonite. The particles behave as if they have an electric charge, with the long edge
negatively charged and the short edge positively charged.
When clays are formed, the particles are
_ __ __ _
carried by geological weathering processes
+ + + long thin clay particle into bodies of water such as lakes, where
- - - - - - - - - +
+
-+- - - - - - - -
with +ive charged ends they are slowly deposited as bed material.
+ + Water is therefore bound into the structure of
H2O butterfly molecule
_ _ _ __ _ clays. Salt water lakes contain dissolved salt
with –ive charged wings
ions, which exist at far lower concentrations
in fresh water lakes. As the particles settle in
the water, polar electric charges force particular orientations of the water molecules and clay
particles (like charges repel), so that a layer of water is trapped between the clay particles (the
Gouy-Chapman layer).
Sea water has a higher ion concentration
(negative charge) than fresh water, so the
water layer is thinner as less water is needed to
neutralise the negative charge at the clay
particle surface. Van de Waals forces are the
atomic attraction forces generated by
spinning electrons, which are stronger than the
Marine clay Lacustrine clay
forces of polar electric repulsion. So when two
particles approach closely enough they
adhere, giving clay its cohesive properties.
Clays deposited in sea water (marine clays) include thin water layers and Van der Waals forces
govern, so the particles are arranged in an edge-to-face structure. Clays deposited in fresh
© Jonathan Haynes version 3.0 Page 155 of 239
32900 Bridge Engineering
MSc Structural Engineering

water (lacustrine clays) include thick water layers and polar electric repulsion forces govern, so
the particles are arranged in a random dispersed structure.
As clay is deposited the material is progressively compressed (consolidated), which can force
more particles to come into contact, reduce the water layer thickness and possibly rearrange
the particles of clay. When the compression is removed or reduced only the reduction in
thickness of the water layer is reversible. Generally the shear resistance of clays improves as the
void ratio decreases (water has no shear strength) but if the clay particle is broken all associated
shear resistance is lost and cannot be re-established.
Since water is a constituent part of clay structure, it has no apparent shear strength when dry
(zero moisture content) as the particles are not bound together. This is why long term clay
strengths can drop to zero when not subject to any normal stress. Clays which have been subject
to low levels of consolidation have high moisture contents and low strengths; conversely, high
strength clays have low moisture contents and high levels of consolidation. The shear strength of
cohesive soil is therefore related to moisture content and effective stress.
If the effective stress in a clay is significantly reduced (e.g. when a glacial sheet retreats at the
end of an ice age) the moisture content will increase and strength will drop. Such clays are
referred to a over-consolidated. Normally consolidated clays exist at similar levels of effective
stress to those which prevailed when the material was formed.

over-consolidated strength envelope

soil at small strains c = σ (0.11 + 0.037 PI )
shear stress τ

where,
c is the undrained shear strength
residual strength σ is the effective stress
at large strains PI is the plasticity index (liquid
c limit – plastic limit)
φr
normal stress σ

short term strength

shear stress τ

peak strength
shear stress τ

φ long term residual strength

strength
c
φr
normal stress σ strain ε

The relationship between shear stress and strain for clay is generally linear up to a peak strain
which drops to a residual strength at large strain values. The linear region relates to a state where
intermolecular bonds are being broken and re-established as particles slide over each other. The
residual strength region relates to a state where clay particles are being broken.
The rate of strain applied to a clay will have a significant effect upon the behaviour exhibited.
Short term strengths relate to confined pore water pressures, whereas long term strengths relate
to a clay with reduced moisture content (no apparent cohesion). It is very important to consider
the strength of clay over the likely life of the structure being designed, and use appropriate
strength values.
When soil is retained by a structure, the soil has not moved horizontally and is initially at rest – the
relationship between vertical and horizontal stress is defined as ko. Due to the imbalance of
loading the wall is pushed forward by a wedge of soil which cannot restrain its own movement. If
the wall translates and rotates sufficiently the horizontal stress in the wedge of soil decreases to
an active level, ka.

© Jonathan Haynes version 3.0 Page 156 of 239

32900 Bridge Engineering
MSc Structural Engineering

The horizontal strain required to reach active pressure is generally less than 0.5%. As the wall
translates, the soil in front of the wall is compressed and will eventually arrest the wall movement;
the ratio of horizontal to vertical stress must be around ten times greater than those required for
active stresses. The principal stresses for active and passive pressures are related by the soil angle
of internal shear friction, φ.

σv=γh1 σh
σv
σha

6
σv=γh2
kp

5
σhp φ
45 + 5% strain for
2
dense sand

4
15% strain for
loose sand

3
Mohr’s circle for soil elements in the active and
passive soil pressure zones

2
< 0.5% strain
φ ka
σ

1
ko

σha=kaσv σhp=kpσv
-5 0 +5
horizontal strain, ε (%)

For cohesionless soils,

k o = 1 − sin φ
1 − sin φ 1
ka = and k p =
1 + sin φ ka
Generally soil will be saturated with water and effective soil parameters are used in calculations.
For cohesive soils, the active pressure at depth is reduced by the soils adhesive nature, so

σ ha = k a γh − 2c k a

 φ
The horizontal strain required to develop full passive pressure is, εh tan 45 +  which is
 2
approximately 0.20m for a 2.0m high wall in dense sand where φ=35o and the ultimate strain is
5%. This order of movement is never likely to be acceptable and so any reliance on passive
pressure for stability should be avoided.

5.2.2 Soil Movement at Integral Bridges

Although the integral bridge concept has proven to be economical in initial construction for a
wide range of span lengths as well as technically successful in eliminating expansion joint and
bearing problems, it is not without in-service problems. Fundamentally, these problems are due
to a complex soil structure interaction mechanism involving relative movement between the
bridge abutments and adjacent retained soil. As this movement is the result of natural, seasonal
thermal variations, it is inherent in all integral bridges. Although these problems turn out to be
© Jonathan Haynes version 3.0 Page 157 of 239
32900 Bridge Engineering
MSc Structural Engineering

primarily geotechnical in their cause, they can result in significant damage to structural
components of the bridge. In addition, on the approach to an integral bridge there can be
pavement ride quality problems for motor vehicles due to settlement of the supporting soil at the
end of the bridge - the primary cause of post-construction, in-service problems for Integral
bridges.
As the bridge superstructure goes through its seasonal length changes, it causes the structurally
connected abutments to move away from the soil they retain in the winter (the deck contracts)
and into the soil during the summer (the deck expands). The mode of abutment movement is
primarily rotation about their bottom although there is a component of translation (horizontal
displacement) as well. The total horizontal displacements are greatest at the top of each
abutment and can have a maximum magnitude of the order of several tens of millimetres.
At the end of each annual thermal cycle, there is often a net movement of each abutment
away from the retained soil. This net movement is exacerbated when the bridge superstructure is
composed primarily of Portland cement concrete because it undergoes inherent post-
construction shrinkage that causes the bridge superstructure to shorten permanently and pull the
abutments away from the retained soil in the process.

winter long term

position position
potential backfill void due to long term
summer
position abutment shake-down, if approach slab
used; otherwise significant step in road
construction

net effect of annual thermal changes on abutment movement

There are at least two important consequences of this annual thermal movement cycle in
integral bridges:
• Large horizontal earth stresses (lateral earth pressures) develop between the back of each
abutment and the retained soil during the annual summer expansion of the superstructure.
As each abutment is rotated and translated into the soil mass that it retains, the lateral earth
pressures can approach the theoretical passive state, especially along the upper portion of
the abutment where horizontal displacements are largest. This means that the resultant
horizontal earth force on each abutment can be more than an order of magnitude greater
than that for which an abutment would typically be designed. This far exceeds any normal
structural safety factors and thus can result in structural failure of an abutment.
• Due to the net inward movement of the abutments a settlement trough develops adjacent
to each abutment. This is the result of the soil wedge slumping downward and toward the
back of each abutment. The consequences of this settlement depend on whether or not an
approach slab was constructed as part of the bridge. If there is no slab, there will be a
difference in road surface elevation occurring over a short distance creating a step at the
end of the bridge. If there is a slab, initially it will span over the void created underneath it by
the settled soil. However, with time and traffic loading the slab is likely to fail in flexure.
England et al (2000) suggest a maximum slump of the approach slab to be 110mm for a 60m
span bridge over its 120 year design life. They also conclude that long-term soil stresses on the
abutment are unaffected by the initial backfill density or the season of construction, however the
initial rate of stress increase is highest for bridges completed in winter.
There is also the possibility of a similar but less significant movement effect due to the daily
thermal cycle, though the abutment movements created are only between a quarter to a tenth
of the seasonal movements. When daily and seasonal movements are combined (e.g. at the
solstice’s) the seasonal movements may be increased by up to 100%.
The seasonal movements of the abutment wall create volume changes in the retained soil,
which are greatest at the top of the wall. As the bridge deck contracts in winter the abutment

© Jonathan Haynes version 3.0 Page 158 of 239

32900 Bridge Engineering
MSc Structural Engineering

wall moves away from the retained soil (which slumps into the space) and active earth pressure
states exist. When the bridge deck expands in summer the abutment wall moves into the
retained soil and passive earth pressure states exist. However, when the winter slump occurs,
there is a compaction and dilation of the soil granules due to residual shear strains which
accumulate with each cycle of movement, densifying the soil. The net result is a ratcheting of soil
strain close to the wall, which reduces with distance from the wall and with depth. The physical
manifestation of this strain is granular flow of soil down the wall (resulting in the soil wedge slump),
away from the bottom of the wall and upward at the active shear failure plane.
A rigid bridge structure founded in clay soil will behave slightly differently. Cyclic horizontal
displacements will cause a progressive reduction in shear strength to a residual value where
φr=10-15o. This is likely to result in bodily abutment slip but no significant settlement.
A flexible bridge founded in clay may behave like a rigid bridge but is more likely to exhibit
movements like a granular soil and absorb these movements by elastic deformation, which may
lead to later shear failure of the soil.

5.3 Design of Integral Bridges

The design of integral bridge elements is no different to other beam and slab bridges, however
there are more complex considerations relating to soil-structure interaction to consider. The
structure is essentially a continuous framed structure but the supporting soil also forms one of the
principal actions.
Design calculations are based upon an effective bridge temperature (EBT) which is based upon
extreme shade temperature. EBT is related to location (in the UK only) and the type of bridge
deck. EBT of a solid concrete deck is related to the average environmental temperature over the
past two days, whilst for steel composite decks it is related to a period of the last eight hours. The
steel composite deck will suffer a seasonal movement about 20% greater than the solid concrete
deck.
Maximum EBT Annual EBT Variation Minimum EBT
Location
Concrete Composite Steel Concrete Composite Steel Concrete Composite Steel

Birmingham 32 36 44 42 51 64 -10 -15 -20

Newcastle 29 34 42 35 44 56 -6 -10 -14
London 34 39 45 38 46 55 -4 -7 -10

EBT (oC)
Variation of EBT with time for concrete and composite deck construction

Max EBT
annual variation for composite deck
annual variation for concrete deck

1 6 12 months

Min EBT daily variation for daily variation for

composite deck concrete deck

© Jonathan Haynes version 3.0 Page 159 of 239

32900 Bridge Engineering
MSc Structural Engineering

5.3.1 Spread Footings Supporting Abutments

The cyclic nature of the abutment wall movements tends to rock spread footings on cohesionless
soil. This will usually mean that one edge of the footing is subject to tension and over a long
period of time a central core of the footing resists the vertical loadings, as if it were a footing
subject to axial and in-plane moment.

d
soil wedge slump and heave

Lα∆T
compaction, dilation annual displacement of wall head, d=
and movement of 2
θ granular soil d
seasonal wall rotation, θ ≈
h h
strain at k* conditions, ε * = 0.05%
2.5bε *
b
φ elastic core of rocking footing, bc =
θ
where,
plan on elastic L is the deck span, α is the deck coefficient of thermal
core of rocking expansion, ∆T is the EBT
bc footing
2.5b

bulb of pressure

5.3.2 Earth Pressure Estimates at Integral Abutments

The design recommendations for the magnitude of the lateral earth pressures in BA 42/96 are
largely based on the findings of centrifuge and analytical studies reported by Springman et al.
(1996). The report recognises the potential for stress escalation with time and proposes earth
pressure distributions for three principal structural forms.
These pressure distributions are expressed in terms of ko and k*, where k* is defined in terms of the
retained height h, thermal displacement of the top of the abutment d, wall friction δ = φ ' , φ' is
2
the effective angle of shearing resistance and kp is the passive lateral earth pressure coefficient.
0 .4
 d  kp
k* = kp   ≥ k0 = 1 − sin φ ' ≥
 0.05h  3
0.6
England et al (2000) offer an updated formula, k * = k 0 + k p 
 d  based upon field testing

 0.03h 
and FEA modelling.
Clearly, the detailed design of an integral bridge will rely upon a detailed site investigation and
extensive soil testing to determine reliable values of all geotechnical parameters. Additionally, it
is known that geotechnical FEA is less reliable than structural FEA in predicting actual strain in the
material modelled. This suggests that structural design forces should be derived from a sensitivity
study of the enveloped possibilities of soil performance.

© Jonathan Haynes version 3.0 Page 160 of 239

32900 Bridge Engineering
MSc Structural Engineering

assumed to mobilise full

passive pressures

Shallow height bank pad and end screen abutments

k*
earth pressure
based on k*
0.5h

earth pressure
fill based on k0
material

ko
earth pressure coefficient earth pressure distribution
Full height frame abutments
k*
earth pressure
based on k*
0.677h

earth pressure
virgin based on ko
soil

ko
Full height embedded wall abutments earth pressure coefficient earth pressure distribution

5.3.3 Laterally Loaded Piles

For most piles, vertical loading dominates and consideration of small horizontal forces is not
necessary for design purposes. However, in the case of integral bridges the piled abutments are
required to resist considerable horizontal forces and bending moments from the bridge deck.
The lateral load capacity has only been studied since the 1960’s (Broms, 1964). Research and
field observation show that maximum bending moments are experience near the pile head
especially in long flexible piles. In slender piles, the bending moment capacity of the pile governs
design, whereas in stocky piles the shear capacity of the soil governs.
For over-consolidated clays, the soil horizontal spring stiffness kh is assumed to be constant with
depth,
0.25
k k D
kh = w and β =  h 
5D  EI 
where, D E and I relate to the pile and kw is the Winkler spring stiffness from a 12” plate load test.
For normally-consolidated clays and sands, the soil horizontal spring stiffness kh is assumed to
increase linearly with depth,
0.20
ηh x η D 
kh = and β =  h 
D  EI 
where, D E and I relate to the pile and ηh is the soil Coefficient of Modulus Variation
If βL ≤ 2 then a pile of length L is short.
If βL ≥ 3.5 then a pile of length L is long.
Terzaghi (1955) recommends values for the Coefficient of Modulus Variation,
ηh (kN/m3) Loose Medium Dense
Dry or moist clay or sand 2500 7500 20000
Submerged clay or sand 1400 5000 12000
Soft organic silt 150

© Jonathan Haynes version 3.0 Page 161 of 239

32900 Bridge Engineering
MSc Structural Engineering

Lateral deflection of the pile is largely limited to a ten diameter section at the pile head. This
leads to a crude pile model which is a cantilever that has an effective length of 10D. However,
better results may be obtained from a 2D linear elastic model supported by lateral springs which
represent the soil stiffness (and may increase in stiffness with depth).
The discrete spring relationship is, p = − k h y and its governing differential equation is essentially
the same as that used for Beam on Elastic Foundation problems,
d4y
EI + Dk h y = 0
dx 4
Hβ η L−0.88
Pile head deflection in cohesive soils, y = Pile head deflection in cohesive soils, y = 0.h4
kh D η h ( EI )0.6

Deflection Soil Bending Deflection Soil Bending

Reaction Moment Reaction Moment

1.5D

L L

9CuD 3DγLkp
Short fixed head pile in cohesive soil

Deflection Soil Bending Deflection Soil Bending

Reaction Moment Reaction Moment

M M
1.5D f

f
M M
L L

9CuD 3Dγfkp

Long fixed head pile in cohesive soil Long fixed head pile in cohesionless soil

Pile group behaviour is complex when subject to lateral loading. However, if all the piles are
vertical and of the same cross section and length, the ground conditions are uniform and the
pile cap is rigid then the group capacity may be assumed to be the summation of individual
capacities for horizontal loading. Though it would be preferable to use stiffness analysis to
distribute combinations of vertical and horizontal pile cap loading to the piles.

© Jonathan Haynes version 3.0 Page 162 of 239

32900 Bridge Engineering
MSc Structural Engineering

5.4 Structural Modelling

5.4.1 Beam Element Models
Conventionally, an integral abutment bridge is idealised as a continuous beam frame structure.
If no soil reactions were considered, the problem could be solved statically using moment
distribution (or the stiffness method). In this way, the continuous frame is solved only for the design
of deck-abutment joints. However, the structural model fails to reflect the actual behaviour of
the bridge with the effects of several loading conditions and the soil-structure interactions.

beam elements
abutment
pier support

pile

structural model for deck design structural model for deck-abutment joint design

5.4.2 Grillage Analysis

This method is usually used for analysis of bridges based on the consideration of the bridge deck
as an elastic continuum in the form of an orthogonally anisotropic plate. Using the stiffness
method of structural analysis, it is possible to analyze the bridge deck structure as an assembly of
elastic beam elements connected together at discrete nodes. It may also be applied to
retaining elements of an integral bridge. The following example will use simplifications of the
foundation stiffness relationships given in Section 2.3.1:

24.00m
16.00m
k v = 0.75 E (bw)
0.5
kh

k h = E (bw)
0.5
kv
0.25 0.75
l = 0.82b w l
4.00m

The footings for the portal frame bridge shown above have w=4.0m (parallel to span) and
b=16.0m wide. If E=100000 kN/m2 then vertical and horizontal stiffnesses are:

l = 0.82b 0.25 w 0.75 = 0.82 × 16 0.25 × 4 0.75 = 4.64m

k v = 0.75 E (bw) = 0.75 × 100000 × (16 × 4 )
0.5 0.5
= 600000kN / m
k h = E (bw ) = 100000 × (16 × 4)
0.5 0.5
= 800000kN / m
It should be noted that because of the square root in the equations these stiffnesses have to be
calculated for the full foundation before being reduced to stiffness per unit width,
k v 600
kv = = = 38000 kN/m per metre run
b 16
k 800
kh = h = = 50000 kN/m per metre run
b 16

© Jonathan Haynes version 3.0 Page 163 of 239

32900 Bridge Engineering
MSc Structural Engineering

Integral Bridges - Conclusion

Things to do now :
1. Review the Powerpoint lectures on Blackboard.
2. Attempt the Self Assessment Exercises below. Then check your answers against the
solutions on Blackboard.
3. Have a rest.

4. Prepare your group submission for Assignments 2 & 3.

Self Assessment Exercises

1. A 50 m long integral bridge has deep wall abutments which retain 6 m of well compacted
granular fill. The peak angle of friction of the fill is 45° and its dry density is 1900 kg/m3. The
design extreme event for the determination of maximum abutment pressures is a 40° increase
in temperature. Assume α for the deck is 12x10−6 /°C and Kp =17.5 for φ’=45°. Determine design
abutment earth pressures.
2. For the elastic plane frame analysis model prepared for the pedestrian arch bridge self
assessment (remember, there is a LinPro file on Blackboard).

10.0m

40.0m

Explore the effects of soil stiffness by including a representation of soil as elastic springs at
each support. Each arch is supported on two spring supports and restrained horizontally by a
lateral spring. Assume soil properties as follows:
E = 236000 kN/m2, vertical stiffness kv = 94000 kN/m/m run and horizontal stiffness kh = 118000
kN/m/m run.
Further Reading
Arsoy, S., Baker, R.M. & Duncan, J.M. (1999). The behaviour of integral abutment bridges. Report
FHWA/VTRC00-CR3, Virginia Transportation Research Council, Charlottesville, USA.
Burke, M.P. (2009). Integral and Semi-Integral Bridges. UK: Wiley Blackwell.
Concrete Bridge Development Group. (2010). Integral concrete bridges to Eurocode 2.
Camberley: The Concrete Society.
England, G.L., Bush, D.I., Tsang, N.C.M. (2000). Integral Bridges: A Fundamental Approach to the
Time-temperature Loading problem. London: Thomas Telford
Fang, H-Y. (1991). Foundation engineering handbook. U.K: Springer
Nicholson, B. (1998). Integral abutments for prestressed beam bridges. Leicester: British Precast
Concrete Federation.
Bolton, M.D., Springman, S.M., Sun, H.W. (1989). The Behaviour of Bridge Abutments on Clay. UK:
University of Cambridge, Department of Engineering.
Springman, S.M., Norrish, A.R.M. & Ng, C.W.W. (1996). Cyclic loading of sand behind integral
bridge abutments. TRL Project Report 146. Crowthorne: Transport Research Laboratory.
Wallbank, E. (1989). The performance of concrete in bridges – a survey of 200 highway bridges.
London: HMSO.
Way, J.A. & Yandzio, E. (1997). Integral steel bridges: design of a single span bridge – worked
example. SCI Publication 180. Ascot: The Steel Construction Institute.

© Jonathan Haynes version 3.0 Page 164 of 239

32900 Bridge Engineering
MSc Structural Engineering

6. Long Span Bridges

This section will introduce basic principles of design for long span bridges. It will describe forms
of construction that have been adopted for these bridges, focusing in turn on the cable
system, the pylon and the deck.
You should also review: 6 Long Span Bridges.ppt on Blackboard
Principal References: Pugsley, A. The theory of suspension bridges.
Walther, R. et al. Cable stayed bridges.

Cable structures are ideal for long spans because of their efficient use of tensile resistance. Their
popularity has increased in the last few decades because of increasing use of lightweight
construction. High tensile steel cable allows a high concentration of stress in its cross section. In
this respect, it is not equalled by any other type of light weight structural member. The span
range for cable supported bridges is between 200m and 2000m.
The structural system of cable assisted bridges comprises of the following four main components:
• The stiffening girder with the bridge deck
• The cable system supporting the stiffening girder
• The towers
• The anchor blocks
There are different types of cable supported bridges which are characterized by the
arrangement of the cable systems i.e. suspension system and cable stayed system.
The suspension system consists of a parabolic main cable and vertical or inclined hanger cables
connecting the stiffening girder to the main cable. Where as, the cable stayed system contains
straight cables connecting the stiffening girder to the towers.

tower saddle main cable

cable system
tower anchor block
hanger tower
stiffening girder
cable

Although there are common features in the behaviour of suspension bridges and cable stayed
bridges since cables are basic elements in both bridges, there are also significant differences. In
suspension bridges, the cables and towers are primary load carrying elements; the deck
essentially provides a platform for the traffic and distributes the effect of concentrated loads and
also keeps the deformation of the main cable due to live load to an acceptable limit. The deck
strength and stiffness are generally determined by this later function. Also the deck bending
moments are normally zero in the dead load condition. In cable stayed bridges, the cables
merely assist the deck in supporting the load and deck dead load moments are non-zero.

6.1 Components of Cable Stayed Bridges

6.1.1 Cables
The cables used in cable stayed bridges may be composed of helically wound strand, parallel
wire strand or locked coil wire ropes.
Parallel wire strands consist of galvanized round wires laid up in a hexagonal form, with a long
helix. The product is then encased in a tight fitting high density polyethylene (HDPE) tube. Parallel
wire bundles are ideal for cable stayed bridges, offering high axial stiffness and an especially
high modulus.

© Jonathan Haynes version 3.0 Page 165 of 239

32900 Bridge Engineering
MSc Structural Engineering

Cable type Shape Structure Bridge

Wires are hexagonally bundled in Brooklyn

Parallel Wire Strand parallel. Humber
Akashi Kaikyo

Six strands made of several wires are

Strand Rope closed around a core strand. St. John

Deformed wires are used for the Little Belt

Spiral Rope outside layers. Wakato

Wires are stranded in several layers Kvalsund

Locked Coil Rope mainly in opposite lay direction. Emmerich

Locked coil strands consist of a centre of one or more layers of large diameter galvanized round
wires helically spun together. Stranded on top of this centre are one or more layers of large
diameter galvanized shaped wires, mainly in opposite directions, to achieve the required
diameter. The closed construction and smooth outer layers offer high resistance to deformation
and specific pressures.

6.1.2 Cable Arrangements

There are many possible arrangements for the cables in cable stayed bridges. In the transverse
direction the cables may lie in either a single vertical plane, double vertical planes or inclined
planes. In the longitudinal direction, the most common arrangements are fan and harp but the
star arrangement is also possible.

Cable-stayed bridges with two pylons and single counterbalanced pylon.

harp fan

star it may be necessary to improve bending and

torsional stiffness by providing tie-downs in the
back spans, either by attaching a frame to a
foundation or by connecting directly into a
counterbalance block

The harp system has cables parallel and equidistant from each other. This system is popular for
aesthetic reasons. However, it causes bending moments in the tower. Variation in cable stiffness
is only related to length.
The fan system has cables equidistant from each other at the ends but different spacing at the
tower. The result is a reduced bending moment in the tower but more cable, although the
© Jonathan Haynes version 3.0 Page 166 of 239
32900 Bridge Engineering
MSc Structural Engineering

arrangement is most structurally efficient because cable inclination is greatest. The top cable
connections are often complex but cheaper than the numerous connections required for the
harp arrangement. Variation in cable stiffness is related to length and inclination angle.

inclined plane cables

double plane cables
single plane cables

The aerodeck uses an inverted aerofoil shape to reduce flutter and ensure that the net wind induced deflection of
the deck is downward, thus increasing tension in the hanger cables (avoiding compression and cable instability).

In the star arrangement, the cables intersect the tower at different levels (like the harp
arrangement) but converge on the deck at a common point. This is only acceptable in relatively
short spans.

6.1.3 Stiffening Girder

In cable stayed bridges, the stiffening girders are designed to sustain bending, torsion and also
the axial force component induced by the cables. The choice of the cross-section for the
stiffening girder plays an important role in the process of designing a cable stayed bridge.
In the single plane cable arrangements, the cables only assist the bridge deck in bending, hence
torsion has to be resisted by the stiffening girder alone. Torsionally stiff single box girders and
multiple web box girders provide the best solution in this case. The box girder has sufficient
torsional stiffness to transmit any twisting moment from a load with an eccentric resultant such as
traffic load in only one carriageway.
With two vertical cable planes attached along the girder, both vertical and torsional support is
provided by the cable system and it is therefore not required that the girder in itself possesses
torsional stiffness. The girder can simply consist of two I-shaped plate girders directly under the
cable planes.
With two inclined cable planes intersecting at the top of the pylon, the girder receives the same
cable support as with two vertical cable planes. In this case a girder with torsional stiffness is also
not required.
In cable stayed bridges with very long spans, where the torsional stiffness becomes essential to
achieve aerodynamic stability, it is often advantageous to use a box girder combined with two
cable planes, and also to give the girder a favourable streamlined shape. It should however, be
emphasised that an aerofoil deck is only required for very long spans (above 500m) or for width /
span ratios below 1/25 (Troitsky, 1988).
It is important to examine pattern load cases in all long span bridges, to identify maximum deck
torsion effects, maximum tower bending/axial, maximum cable axial effects and any possible
uplift conditions.

© Jonathan Haynes version 3.0 Page 167 of 239

32900 Bridge Engineering
MSc Structural Engineering

6.1.4 Tower
The form and proportions of the towers in a cable stayed bridge are very important because the
towers provide the main visual impact of the bridge. The appropriate form depends on
balancing structural, maintenance, geometrical and aesthetic factors. Towers may be
constructed in concrete or in steel as they sustain compressive forces and bending. The form of
the tower depends on the cable arrangement and its height depends on the inclination of the
cable.

a) H frames b) A frame c) Inverted Y frame

d) central cable plane e) cables in parallel planes f) cables in twin inclined planes

In the case of a single plane cable arrangement, a single tower is required at the central
longitudinal axis of the bridge deck. The towers are simply cantilevers or props which may be
pinned to the girder in the longitudinal direction, or may be fixed to the girder.
With the double plane arrangement in which the towers are vertical, there should ideally be a
cross member connected to the towers to form a portal frame.
With the inclined plane arrangements, A-frame towers are normally used. This type of tower has
proven the optimal solution for appearance and stability in strong winds. It provides greatest
torsional stiffness to the deck.

6.1.5 Cable Connections

The joint details between the cable, girder and tower components present some important
design features. In cable stayed bridges, special connections are required to allow the
transmission of the cable forces to the girder and the tower (Walther et al, 1988).
Due to the fact that the high strength of the wires is achieved by carbon content approximately
five times larger than in normal structural steel, the wires cannot be welded. Instead sockets are
fixed to the ends of the wire bundle constituting the stay cable.
The sockets are made of cast steel in the form of a short cylinder
with a conical cavity. Inside this cavity the strand is broomed and
subsequently the space is filled with a metallic zinc alloy or a
mixture of epoxy resin, zinc dust and steel balls.
Alternatively, strand developed from post-tensioning systems (such
as Freyssinet) may be used. This requires the strand to be locked in
an end block using wedges.

© Jonathan Haynes version 3.0 Page 168 of 239

32900 Bridge Engineering
MSc Structural Engineering

15mm φ galvanised
case hardened alloy steel strand wax
cable
jaws for fatigue strength
Injection cap

strand overlength bearing surface

anchorage tube HDPE casing

wax fill guide-stuffing box

anchor block

The drawn wire strand is galvanised then bundled and encapsulated in a high density
polyethylene casing, the inner gaps of which are pressure filled with petroleum wax. Each strand
is therefore individually corrosion protected.
Parallel wire
Strand rope Locked coil
strand
Diameter (mm) 15.7 15.2 75 85 100 110
2
C.S.A. (mm ) 150 170 3821 4908 6793 8220
2
UTS (N/mm ) 1770 1570 1250
2
0.45UTS (N/mm ) 797 707 565
2
E (kN/mm ) 205 200 165
Strand rope is usually arranged in 19, 31, 37, 61, 73 or 93 strands.
Parallel wires are normally arranged in 21, 37, 45, 57, 69, 77 or 81

a tie-down a strand end-block a broomed end-block a suspension cable hanger

saddle

© Jonathan Haynes version 3.0 Page 169 of 239

32900 Bridge Engineering
MSc Structural Engineering

6.2 Erection Methods

The erection of the superstructure of cable stayed bridges can be broadly considered by three
methods, namely the cantilever method, by launching and construction on scaffolding.

6.2.1 Cantilever Method

The cantilever method is very popular for the erection of long spans. However this method is
costly for medium span cable stayed bridges.
The cantilever method involves free construction from a support. It can be executed from one
end of the bridge, from both ends towards the centre and also from a central pier towards the
ends. It is characterised by large deflections at its free ends. Free cantilevering is most suitable for
deep terrain or for rivers having busy navigation.

Stage 1: Construction of piers, back-spans and caissons

The off-shore towers require

sheet piled cofferdam to
create dry areas for the
foundation works. A piled
pontoon or floating crane is
required
Stage 2: Construction of work stations on off-shore towers

Tower must be designed for

bending due to load imbalance
or be guyed to adjacent piers

Stage 3: Construction of first off-shore cantilever sections

Deck sections erected and

cantilever from stayed deck.
New sections delivered to
central pier by boat. Strand
fitted and tensioned to support
new section

Stage 4: Construction of central section

The deck is potentially least

stable just before the spans are
joined as bending is greatest
but lateral restraint is least

Stage 5: Completion of central spans

Sheet piling removed and final

cable tensioning to level deck
under permanent actions

© Jonathan Haynes version 3.0 Page 170 of 239

32900 Bridge Engineering
MSc Structural Engineering

The intermediate piers in the side spans help the cantilever stability during construction. In this
case the side spans are to be built first, later the long cantilevers in the main span are
constructed on stabilised back-spans.
The cantilever method can create unfavourable temporary bending moments in the deck. For
example for the Arade bridge in Portugal temporary diwidag bars had to be installed above the
top slab, anchored to the top slab with the help of steel anchorage blocks bolted to the slab, as
the final tendons in the deck were not able to balance all bending forces during the
construction phase. These temporary bars were de-tensioned when possible and reused further
in the cantilever (Troitsky, 1988).

6.2.2 Installation by Progressive Launching

Launching involves the progressive hydraulic jacking of deck sections across the span.

Stage 1: Construction of piers, caissons and temporary pier

The off-shore towers require sheet

piled cofferdam to create dry
areas for the foundation works.
Piled pontoons are required in the
spans, as dictated by the
cantilever capacity of the box
girder

Stage 2: Jacking of first section

Towers must be designed for
bending due to load imbalance.
There must be sufficient back-
span box girder to
counterbalance the advancing
front of the deck (or add
kentledege until there is). A
launching nose permits the
deflecting deck to reach the next
bearing
Stage 3: Construction of first off-shore cantilever sections

The deck sections are

constructed / assembled at the
abutment then hydraulically
jacked into the span

Stage 4: Construction of central section

The bearings used for launching

will be designed for very low
horizontal friction, or include
simultaneous vertical jacking

Stage 5: Completion of central spans

Sheet piling and temporary piers

removed and final cable
tensioning to level deck under
permanent actions, launching
bearings replaced by permanent
bearings

© Jonathan Haynes version 3.0 Page 171 of 239

32900 Bridge Engineering
MSc Structural Engineering

It is common in multi-span bridges, to jack from several points – often all the piers. The bearings
used for launching are often horizontal rollers, so they are replaced by permanent bearings
when the launching process is completed.

6.2.3 Installation by Span-wise Launching

The launching technique consists of complete deck assembly behind the abutments, launching
across the spans and finally lowering down of the full length of the deck of the bridge onto the
piers. Supporting devices such as rollers or sliding bearings are necessary. This requires temporary
structures which can span the same distance as the finished bridge spans (usually lightweight
trusses) and is very efficient for multi-span viaducts where there is significant repetition. In some
systems the deck is constructed on a hydraulically moveable shutter.

Stage 1: Construction of piers and caissons

The off-shore towers require
sheet piled cofferdam to
create dry areas for the
foundation works. A
temporary truss of sufficient
length to cantilever the
largest span is launched

Stage 2: Jacking of launching truss

The truss may contain

hydraulic shuttering to
permit a whole span to be
cast insitu, or a precast
deck span may be craned
down the truss and
positioned on bearings

Stage 3: Construction of first off-shore deck sections

The truss is then advanced

and the process repeated;
this procedure is best
applied to repetitive
identical spans. The spans
should hogg under self
weight only.

Stage 4: Construction of towers

The towers are erected and

cables tensioned to
removed permanent action
sag in the deck sections

Stage 5: Completion of cable stays

Sheet piling removed and

final cable tensioning
cycle, launching bearings
replaced by permanent
bearings

© Jonathan Haynes version 3.0 Page 172 of 239

32900 Bridge Engineering
MSc Structural Engineering

6.2.4 Construction on Scaffolding

This method offers the advantage of constructing deck continuously from one end to the other
allowing the transportation of men, equipment and material on the completed part of the deck.
It also leads to an efficient control of deck geometry and cable tension.
In many cases clearance requirements will exclude the installation of temporary supports and
sometimes the cost of erecting the temporary supports over a large water depth in the main
span may make this procedure unsuitable.

6.3 Structural Behaviour

Cable-stayed bridges are structural systems in which inclined stay cables act as tension members
supporting the bridge deck directly. A large number of stays is in general the best solution. A
small number of stays results in large cable forces. These require complicated anchorage systems
and detailing. A relatively rigid deck is required to span the distance between the anchorage
points.
The cables provide intermediate elastic supports for the bridge deck and introduce compression
in the pylons and deck.
Under permanent actions, the cable forces must be adjusted in order to obtain the required
bridge profile and optimum distribution of the internal forces. Under traffic loadings, the deck
distributes the loads between the stays, which work as elastic supports. The distribution of internal
forces is dependent upon the relative stiffness of the various elements. Particular attention must
be given to the stiffness distribution adopted for the structural elements.

Outline of an initial analysis for a cable-

stayed bridge.

Moment due to permanent and

variable action

Moment due to cable extension due to

variable action

Axial due to cable triangulation

The cantilever method is particularly suitable for cable-stayed bridge erection but geometric
control during construction is essential. Particular attention must be given to the control of
permanent actions during the erection phase. The correlation between analysis and
construction is highly dependent on the accuracy of the permanent action values.
Long span cable-stayed bridges may be sensitive to aerodynamic effects. In such cases, local
wind conditions must be carefully evaluated. This often involves wind tunnel testing.
At the end supports, vertical uplift reactions must be anchored. These forces depend primarily on
the ratio between central and side spans. A similar situation occurs when backstay cables are
anchored in a counterweight or even directly anchored in the ground. The uplift reactions can
increase significantly if cracking occurs at the base of the pylons (permitting rotations).

6.4 Preliminary Design of Cable-Stayed Bridges

A proper balance of side-span length to main-span length must be established if uplift at the
abutments is to be avoided. Wide box girders are mandatory as stayed girders for single-plane
systems, to resist the torsion of eccentric loads. Box girders, even narrow ones, are also desirable
for double-plane systems to enable cable connections to be made without eccentricity. Single-
web girders, however, if properly braced, may be used.

© Jonathan Haynes version 3.0 Page 173 of 239

32900 Bridge Engineering
MSc Structural Engineering

6.4.1 Preliminary Calculations for Cable Stays

The manual procedure is not precise but does provide first-trial cable-stay areas. With the
analogy of a continuous, elastically supported beam, influence lines for stay forces and bending
moments in the stayed girder can be readily determined.
If the permanent action cable forces reduce deformations in the girder and pylon at supports to
zero, the girder acts as a beam continuous over rigid supports, and the reactions can be
computed for the continuous beam.
If, in a first-trial approximation, variable action is applied to the same system, the forces in the
stays under the total load can be computed from:

Ri
Pi =
Pi sin α i
w where Ri is the sum of
α
permanent and variable
i-1 i i+1
action reactions.
s
Ri

Since stay cables are usually designed for SLS actions, the cross-sectional area of stay i may be
determined from:
Ri
Areq = where unit stress for the cable steel, fpu = 1870 N/mm2
0.45 f pu sin α i

Fh

α cables induce a horizontal

i-1 i i+1 force Fh at top of pylon.
Ri ’ Ri

If a cable-stayed girder is supported by cable force Pi , at ith point of cable attachment. Ri is the
vertical component of Pi .
The reactions may be taken as Ri = w x s, where w is the uniform load in kN per metre, and s is the
distance between stays. Determination of the force Po acting on the back-stay cable connected
to the abutment requires that the horizontal force Fh at the top of the pylon be computed first.
Maximum force on that cable occurs with permanent + variable action on the centre span and
permanent action only on the side span. If the pylon top is assumed immovable, Fh can be
determined from the sum of the forces from all the stays, except the back stay:
Ri Ri '
Fh = ∑ −∑ where Ri , Ri‘ are the vertical component of force in the ith stay in the
Tanα i Tanα i '
main span and side span respectively.

6.4.2 Back Span to Main Span Ratio

When establishing the conceptual arrangement of the bridge it is important that the ratio
between back span and the main span be less than 1:2. This ratio gives a clear visual emphasis
to the main span. Also it influences the uplift forces at the anchor pier and the range of load
within back stay cables supporting the top of the pylon.

© Jonathan Haynes version 3.0 Page 174 of 239

32900 Bridge Engineering
MSc Structural Engineering

6.4.3 Pylon Height

The height of the pylon will determine the overall stiffness of the structure. As the stay angle
increases, the required stay size will decrease and the height of pylon will increase. However, the
deflection of the deck will increase as each stay becomes longer. Both the weight of stay and
the deflection of the deck become minimum when the expression 1/(Sinα x Cosα) is also a
minimum. Where α is stay inclination angle. Therefore the most efficient stay inclination angle is
45°. An optimum ratio of pylon height above the deck H to main span L is between 0.2 and 0.25.

6.4.4 Deck
It is possible, by tuning the forces in the stays, to reduce the moments in the deck, under
permanent action.
For initial design purposes a minimum deck section can be assumed and using this section the
dead weights and section properties are calculated. The structural system is then analysed
incorporating a preliminary distribution of permanent action moment and the sections checked
against the distribution of total moments and normal forces. The sections where stresses exceed
the permissible limits are then modified.

6.5 Suspension Bridges

6.5.1 Structural Arrangement
The simplest type of suspension bridge is one that hangs under self weight in the shape of a
catenary. These bridges have a roadway that hangs from steel cables supported by two tall
towers.

L
cable sag h ≈
11
three span suspended
L

inclined hangers

straight back-stays

The difference between suspension bridges and cable-stayed bridges is that suspension bridge
cables are not directly anchored to the towers, they pass over the top of the towers on a saddle.
A suspension bridge has at least two main cables. These cables extend from one end of the
bridge to the other. Suspender cables (or hangers) connect these main cables to the roadway.
Towers must be very tall to accommodate a workable cable sag. They must have lateral stability,
so are often braced in some way. Longitudinal stability often relies upon main cable restraint.
Any deck form may be used but for long spans the choice is often between an aerofoil box
girder or a stiffened (trussed) girder.
Anchorages are preferably tied to rock but may be mass concrete in shorter spans.
The first suspension bridges were constructed before the development of wire rope (and hence
cables). The Menai Suspension Bridge was designed by Thomas Telford (the first President of ICE)
and was completed in 1826. It used a system of wrought iron chain links to form the suspension
system. The Humber Bridge was designed by John Hyatt, Douglas Strachan (and others from
Freeman Fox) opened in 1981, being the longest single span suspension bridges for almost 20
© Jonathan Haynes version 3.0 Page 175 of 239
32900 Bridge Engineering
MSc Structural Engineering

years. Both employ essentially the same arrangement although the masonry towers are replaced
by reinforced concrete and the wrought iron links are replaced by spun high-tensile steel cables.

Chronologically opposite UK long span suspension bridges.

Suspended cables resist applied loads primarily by adjusting their geometry to accord with the
applied loading and as a result their structural behaviour differs considerably from that of beam-
type structures. In the absence of applied loads, the geometry of the cable will correspond to
the funicular curve of its self-weight loading but the addition of applied loading will cause the
cable to deflect due to a combination of geometrical change and the elastic extension of the
cable, with the relative importance of these effects being dependent on the initial tension of the
cable and the nature of the applied loading.
For example the application of a uniformly distributed load over the whole span will result in an
increase in horizontal component, H and hence the cable tension, T at all points, so that at each
point the change in T will be proportional to the applied 1oad, and no change of shape is
required to maintain equilibrium. However, the increase in tension will result in a lengthening of
the cable, so that all points will move downwards, with the deflection being entirely due to
elastic extension with a maximum value at mid-span.
The effect of a concentrated or part span load is rather different. Since H must be constant, the
tension is increased over the whole length of the cable and, to preserve vertical equilibrium, the
cable profile flattens in the parts of the cable which do not have additional loading. As a result,
the cable profile has to change, with those parts remote from the additional loading deflecting
upwards.

Cable distortion due

to vertical loading.

Additional deflection will occur due to the cable extension, but in most cases this will be small
relative to that due to the geometrical shape adjustment, and in the case of a concentrated or
short length of distributed load, the maximum deflection will be almost entirely due to profile
change rather than cable extension. The resistance of a heavy cable to its displacement by
additional loads is referred to as its gravity stiffness.
For a given main span length, the principal geometric variable is the sag of the main span cable
at the centre of that span.
For the three span bridge of figure 6.13a, ignoring the relatively small contribution of the hangers:

© Jonathan Haynes version 3.0 Page 176 of 239

32900 Bridge Engineering
MSc Structural Engineering

y
2h Sw
cable
cable span, L= Sinh−1
w 2H
T tension at p
V w
H=
length of
p θ (
8h S − 4 h 2
2
)
T’ cable op = x H
T
V’ W
Sw
θ V=
H’ o 2
T0
tension at o T0
deck
x
2
 Sw 
T = H + 
2

W
weight of hatched  2 
section of deck

where, S = total length of the cable, L = total length of the span, h = cable sag, w = weight per
unit length, T = tension in cable, in direction of the derivative dy/dx, H is constant at all points of
the chain, since the weight of the chain acts vertically downward. Therefore, the vertical
component of the tension in the chain is V = H dy/dx.
At the cable anchorage, the individual strands are separated and looped over long anchors.
The anchors are cast into a substantial concrete block which provides resistance to the cable
tension.

T ⋅a T
FoS = ≤ 1. 0
steel plate anchors, held on a W ⋅b
trussed frame and encased in separated strand cable
reinforced concrete

saddle

anchor block attempts to

rotate about this point
W b

a) Forth Road Bridge anchorage b) and c) Humber Bridge anchorage

© Jonathan Haynes version 3.0 Page 177 of 239

32900 Bridge Engineering
MSc Structural Engineering

6.5.2 Analysis of Suspension Bridges

The first analysis of the complete suspension bridge system was made by Rankine, and was
based on the assumption that the cable profile under dead load was parabolic, and the
stiffening girder was sufficiently stiff to distribute any imposed loading so that this profile remained
parabolic. This in effect assumed that the loading in the hangers remained uniform for any given
imposed loading, and it was further assumed that the hanger loads were equal to the total load
divided by the span length. With these assumptions, the increase in cable tension and the
bending moments and shears in the stiffening girder could be derived.
An improvement on this is the elastic theory, usually ascribed to Navier, which retains the
assumptions of a parabolic cable profile and uniform hanger loading but uses a strain energy
method to derive a more rational hanger loading due to imposed loads.
Both the Rankine and elastic theories implicitly assume that cable displacements due to the
imposed load are small compared to the initial cable shape, and this approximation introduces
considerable errors in the stiffening girder bending moments, particularly for long spans. To
eliminate these errors, the deflection theory was developed by Melan (1888).
Solutions to the full non-linear deflection theory equations for complex loadings can be
developed using a Fourier series representation of the deflections.
It will generally be appropriate for an initial global analysis to use a 2D stifness model of the main
cable, hangers, towers and stiffening girder to determine the structure geometry for permanent
loads, and the deck girder bending moments, shear forces and deflections due to traffic loads
and temperature variations.
Once initial sizing of members has been completed, the model can be expanded into a full
three-dimensional representation of the structure to analyse wind loading and differential
temperature effects, torsional moments in the deck, hanger and bearing loads.
Dynamic loading from interaction with the wind is regarded as the most aggressive external
excitation for long-span flexible structures in terms of displacements and internal actions.
Aerodynamic excitation of the superstructure of any type of long-span bridge but particularly of
suspension bridges, may cause unacceptable oscillations (think of Tacoma Narrows). Five distinct
forms of excitation may occur:
• Vortex excitation
• Galloping
• Classical flutter
• Stall flutter
• Gust response
In addition, a quasi-static aerodynamic instability known as divergence may occur.
Design against these effects requires specialised expertise and cannot be considered fully in this
text. However some of them will be described briefly to give some knowledge of what has to be
considered.
Vortex Excitation
When wind flows past a bridge deck, vortices are shed alternately from the upper and lower
surface thus creating an alternating differential pressure and hence force on the bridge. The
frequency of vortex shedding is proportional to the wind speed, and the strength and regularity
depend on the cross-section shape. If the frequency of shedding coincides with a natural
frequency of the bridge, there is a risk of a resonant oscillation occurring.
vortex

Vortex shedding around a cable.

Irregular sections, such as trusses, are rarely prone to vortex excitation. The excitation even of
regular sections is very seldom strong enough to cause large amplitudes. The amplitude is

© Jonathan Haynes version 3.0 Page 178 of 239

32900 Bridge Engineering
MSc Structural Engineering

inversely proportional to the structural damping, and thus adding damping can always be a
cure.
Classical Flutter
Classical flutter is another serious aerodynamic phenomenon, in which vertical and torsional
oscillations are coupled and the lift moment on a moving cross-section reinforces the movement.
It is a well known phenomenon of flight control surfaces in aircraft. It is amenable to
mathematical analysis for plate-like structures (the streamlined boxes of the Severn Bridge type
are nearly plate-like).
It is probable that any section whose torsional frequency is higher than its bending frequency will
eventually flutter. The objective is to ensure that this effect occurs at a wind speed substantially
higher than expected to occur in the actual location. The further apart the bending and
torsional frequencies, the higher the wind speed causing flutter. This is normally achieved by
ensuring the bending and torsional natural frequencies of the deck are grossly different.

Plate girder decks are particularly prone to flutter but can be The performance of aero decks can be greatly improved
greatly improved by adding aerofoil skirts to the leading by adding wind flow deflectors to the outside corners.
edges.
stiffness along the deck lead to
different vertical deflections -
variations in bending support
a lack of torsional stiffness – the
across the deck, combined with
turbulence creates unequal lift

f Nt
≥ 2.5
f Nb
galloping
deck twists

b 1. 1 g
f Nt = f Nb f Nb =
2i 2π ∆ max
b

time

b is the distance between cable hangers and i is the radius of gyration of the deck

Normally it is the fundamental modes which are coupled. However in the first Tacoma Narrows
Bridge, there was a coupling of higher modes in a flutter-like oscillation. Sometimes asymmetric
modes can be suppressed by using a central tie between the cable and deck at midspan. This
tie stops the longitudinal cable movement associated with such modes. However very large
forces occur in the tie, so special care has to be taken when designing the connections.
Truss-type suspension bridges are not necessarily safe from classical flutter, since the roadway
deck is like a plate. They can be improved by leaving open slots between the carriageways to
allow air to pass through, or by having permeable grillages within the carriageways themselves.
Hangers
The hangers to the suspended deck are particularly prone to vibrations induced by periodic
vortex shedding, as their dead load tensions, size and construction type do not vary significantly
along the span but have lengths that can vary from a few metres at mid-span to as much a
several hundred metres near the towers, resulting in a large range of natural frequencies, with a
high probability that at least some hangers will be susceptible to this form of induced vibration,
including the effects of rain–wind interaction.
Possible measures to deal with this included the attachment of dampers, linking the hangers with
secondary stabilising cables, and the attachment of dampers at the hanger–deck connection.
© Jonathan Haynes version 3.0 Page 179 of 239
32900 Bridge Engineering
MSc Structural Engineering

Towers
In the permanent condition, the towers are elastically supported at saddle level by the main
cables and are not susceptible to wind-induced vibrations, but before erection of the main
cables the towers are longitudinally freestanding cantilevers and are much more flexible so that
aerodynamic effects, which are dependent on the structural properties of the tower and its
cross-sectional shape, require consideration.
Concrete towers have thick reinforced walls which for a given compressive strength produce a
high structure mass and, although generally of rectangular shape, usually have rounded corners
that decrease to some extent the intensity of any shed vortices. Concrete towers are therefore
unlikely to suffer from vortex-induced vibrations and although galloping excitation could occur,
this will only be at very high wind speeds. By contrast, the rectangular cross-section of steel
towers is made up from welded stiffened plate and for a comparable compressive strength has
a much lower mass (of the order of one-fifth to one-seventh of a concrete structure), has low
internal damping and the sharp corners of the cross-section can produce strong shed vortices.
It is possible therefore that a steel tower, either part completed, or at full height, will have a
natural frequency coinciding with that of vortex shedding at wind speeds likely to occur during
the construction period, and if analysis and wind tunnel model testing predicts this, additional
damping must be provided either by an external friction damper or an internally mounted tuned
mass damper.

The effective length of the tower

must be assessed in both directions
and at several stages of
construction to identify the most
onerous slenderness ratio for
design.
Although the tower will be
restrained by the suspension
cables in the longitudinal direction
in the final state, they will
cantilever before cable spinning is
complete.
In the lateral direction it is
customary to create sway frames
or braced sections.

6.5.3 Construction of Suspension Bridges

Since the adoption of wire cables, suspension bridge cables have been constructed by spinning.
This process involves producing a continuous multi-strand wire, draped over the tower saddles
and looped around strand shoes at each anchorage.
Reels of coiled wire are delivered to site and laid out to the cable profile using a spinning wheel.
Coils are joined using mechanical connectors.
Hanger rods or cables are attached to the main cable by clamping a saddle around the
finished cable.
The sequence of deck section erection must be carefully planned to avoid excessive
unsymmetrical distortion of the cable. In longer spans it is also necessary to construct the towers
leaning outwards so that they are vertical in the final condition under self weight loading only.

© Jonathan Haynes version 3.0 Page 180 of 239

 32900 Bridge Engineering
MSc Structural Engineering

temporary spinning cable spinning wheel cable reel

strand shoe

Stage 1: Tower and anchorage towers are constructed leaning outwards

construction

Stage 2: Cable spinning individual wires are combined into parallel strands, placed in a hexagonal
shape which is later repacked into a circle and held in shape by steel bands

Stage 3: Back-span deck erection whilst maintaining a symmetrical construction sequence, the back-span
deck sections are transported to site and erected by strand jacking

Stage 4: Central span deck erection uplift of the main cable is restricted by balancing back-span dead load
in the central span
possible expansion
joint locations

Stage 5: Central span deck erection the last sections to be erected are normally either adjacent to the piers
or at centre span

Stage 6: Completion of handrailing, lighting, corrosion protection and surfacing

© Jonathan Haynes version 3.0 Page 181 of 239

32900 Bridge Engineering
MSc Structural Engineering

6.6 Finite Element Analysis of Cables

Cable elements. When modelling tension structures such as cable stayed and suspension bridges
using FE, more thought is required. For other bridge structures modelled using any of the
aforementioned elements, it is not usually required to implement any kind of initial conditions. For
example the model begins its life in an unstressed state, then loads are applied such as gravity
and vehicle loading, causing internal forces and deflections.
However, during construction of cable supported structures the cables are effectively loaded in
sequence and progressively stressed until the deck reaches its design profile. The FE model must
reflect this. Of course it is not usually feasible to construct the FE model in this way, we really only
want to examine the final state and possibly some intermediate states during construction. If we
apply no stress to the cables and apply only gravity then the deck will tend to sag and drape. If
we attempt to apply a pre-calculated stress to the cables based on some hand calculated
model then we may see excessive hogging. This latter approach tends to be acceptable for 2D
cable stayed bridge models and the solution may close to the real final form. Regardless, the
program will require the cable forces to be iterated until the final deck form is reproduced within
a prescribed tolerance. This process is known as form finding and is a typical requirement of all
tension structures. More sophisticated FE codes will have optimization and form finding algorithms
implemented in them.
Analysis with no prestrain Analysis with prestrain
1. Model created 1. Model created
using perfect using geometry with
centreline geometry. shortened
(prestrained) cables.

2. Self weight
2. Self weight applied and gravity
applied and gravity ‘turned on’. Further
‘turned on’. stressing of cables.
Stressing of cables. Deck profile
virtually straight.

3. Variable actions
3. Variable actions applied to deck.
applied to deck. Some deck sag.
Unrealistic deck sag
and cable tension.

Depending upon the FE code used, line elements will usually allow a prestrain to be input so that
this will obviously induce a stress in the element. However if this is not available then prestrain can
also be implemented using thermal control. By specifying a suitable coefficient of thermal
expansion for the material and effectively reducing the temperature, the desired contraction
can be obtained.
Contraction prestrain, e = − ∆T ⋅ L ⋅ α where ∆T is the change in temperature, L is the length of
the cable and α is the coefficient of thermal expansion.

6.6.1 Dynamics
If an FE model of a bridge needs to be examined for dynamic behaviour then there are a
number of issues which need to be considered in the analysis. All normal vertical loading initiates
through gravitational effects (the product of mass multiplied by the acceleration due to gravity).
Dynamic behaviour (whether this be through wind, traffic, or seismic loading) is usually modelled

© Laurence Weekes & Jonathan Haynes version 2.0 Page 182 of 239
32900 Bridge Engineering
MSc Structural Engineering

as an applied acceleration in a translational coordinate direction which varies with time. Inertia
effects are all important (essentially mass x acceleration) therefore a representative loading must
be present in the model, as mass. Following a static analysis (which is usually carried out for
verification purposes), a modal analysis is required to extract natural frequencies. This analysis
should highlight not only the frequencies where resonance might occur, but provide a measure
of the importance of the modes by providing the effective mass. Subsequent time history or
frequency domain (response spectra analyses) will usually make use of the results from the
modal analysis.

6.6.2 Damping
There is a tendency for long cables to vibrate at low frequencies, particularly as self weight sag
and length increases (when inclination angle α is low). This may be induced by traffic or
pedestrian movement, wind or rain rivulets, even when cable tension is high. This is a basic
resonance issue and manifests as a standing wave of displacement in the cable between points
of restraint (usually the deck and pylon). Although complex solutions are available, using viscous,
elastomeric and friction dampers, it is most common to deploy tuned mass dampers directly on
the cables.

Stockbridge dogbone dampers, which have a target frequency range of 3~4 Hz, can be
mounted on cables. These can be placed strategically on the cable to change the length of
any standing wave, making it less likely to form. It may be necessary to mount them orthogonally
if cable displacement can be on more than one axis.

6.6.3 Cable stiffness

According to Walther et al (1999) the stiffness of cable stays is related to the strain in the cable.
The cable only displays its full Young’s modulus, E when it carries a stress greater than 50% of its
ultimate tensile stress. This is because at low levels of strain the cable sags under its self weight
and additional force must be applied to straighten it before its full axial stiffness is realised.
For a cable which is subject to an increase in axial tension,
such that it extends ∆s , the apparent Young’s modulus of
elasticity Ea is given by;
s+∆s
Ea 1
=
E (γL )
2
E
1+
12σ 3
δ2<δ1
where, γ is the material density of steel, and σ is the SLS
δ1
tensile stress in the cable.
s
This is based upon a number of assumptions, principal of
which is the shape of the cable catenary being
L approximated to a parabola, or δ ≤1 .
s 12

© Laurence Weekes & Jonathan Haynes version 2.0 Page 183 of 239
32900 Bridge Engineering
MSc Structural Engineering

Long Span Bridges - Conclusion

Things to do now :
1. Review the Powerpoint lectures on Blackboard.
2. Attempt the Self Assessment Exercises below. Then check your answers against the
solutions on Blackboard.
3. Have a rest.

4. Prepare your group submission for Assignments 2 & 3.

Self Assessment Exercises

1. Consider a cable stayed pedestrian bridge as shown below. The design assumptions are:
• The deck is 4.00 m wide.
• All structural steelwork has a characteristic yield strength of 355 N/mm2.
• The deck is 175mm thick solid reinforced concrete and spans one-way between edge beams.
• The uniformly distributed variable action on the deck is 5 kN/m2.
• The pylon is a 10mm thick, 350mm outside diameter CHS.
• The span cables are 15mm diameter. The back-stay cable is 25mm diameter. Cables have a
characteristic yield strength of 1250 N/mm2.
• The deck edge beams are asymmetric beams formed by welding together a 150x15 top
flange, 300x8 web and 300x15 bottom flange.
• The reinforced concrete tie-down at the back-stay (joint E) is 5.0m x 2.5m x 2.0m.

You are required to:

1. Calculate the ultimate limit state gravity action on the bridge deck, as a uniformly
distributed action on the deck beams.
2. Analyse beam, pylon and stay elements for forces. Assume the deck beams are simply
supported at the stay cables.
3. Check whether the concrete tie-down is large enough to resist uplift in the EQU limit state.
4. Using a linear elastic analysis model, reanalyse the structure assuming the beams are
continuous and fully elastic (this should be a separate beam model).
5. Now create a Plane Frame model including the pylon and stay cables. Extract the
Permanent Action deflection of joints B, C and F from the model. Use them to calculate the
pre-shortening of the cables required to obtain zero deflection at the support joints under
unfactored permanent action. Using a thermal load case, apply cable shortening. Cycle
the calculation until zero deflection is achieved.
6. Compare the deck forces in 2, 4 and 5. Explain the differences.

© Jonathan Haynes version 2.0 Page 184 of 239

32900 Bridge Engineering
MSc Structural Engineering

Further Reading
Bangash, M.Y.H. (1999). Prototype Bridge Structures-Analysis and Design. London: Thomas Telford.
Broughton, P. & Paul, N. (1994). The Analysis of Cable and Catenary Structures. London: Thomas
Telford.
Burden A.R. (1990). Japanese cable-stayed bridge design. Proc. ICE, Part 1, v90, pp. 1021-1051.
Chen, W-F. & Lian, D. (1999). Bridge Engineering Handbook. Florida, US: CRC Press Inc.
Fleming J.F. (1979). Nonlinear static analysis of cable stayed bridge Structures. J. Computers and
Structures. v10 (4), pp. 621-635.
Gimsing N.J. (1997). Cable Supported Bridges: Concept and Design. 2nd ed. Chichester: Wiley.
Gimging, N.J. Anchored or Partially Anchored Stayed Bridges. (1966). Proc. Symposium on
Suspension Bridges. Lisbon.
Hambly. E.C. (1990). Bridge Deck Behaviour. 2nd ed. Oxford,UK: Taylor & Francis.
Hardesty S. and Wessman H. E. (1938). Preliminary design of suspension bridges. ASCE
Proceedings, Paper 2029, pp. 69–95.
Hutson, G.F. and Freeman, R. A. (1987). The Bangkok Cable Stayed Bridge. Thailand: Aksorn
Sampan.
Ito, M., Fujino, Y., Miyata, T., Narita, N. (1991). Cable-Stayed Bridges - Recent developments and
their future. Yokohama, Japan: Elsevier Ltd.
Iwaya K., Tsutsumi Y. and Fukushi A. (1993). Tension drop in cable-band bolts on suspension
bridges, in Bridge Management 2. London: Thomas Telford.
Jennings A. (1962). The free cable. The Engineer, December.
Jensen G. & Petersen A. (1994). Erection of suspension bridges. Proceedings of an International
Conference, Deauville, 2, 351–362.
Kuzmanovic, B.O. History of the Theory of Bridge Structures. (1977). J.Struc. Div., Proceeding ASCE.
v103, May.
Melan J. (1888). Theorie der Eisernen Bogenbrucken und der Hangebrucken, 2nd Edition. Leipzig:
Engelmann.
New Civil Engineer. Humber Bridge. May 1981.
O’Brien T. (1967). General solution of suspended cable problems. Journal of the Structural
Division, ASCE, 5085.
Podolny, W. (1971). Static Analysis of Cable Stayed Bridges. Ph.D. Thesis, University of Pittsburgh.
Podolny, W. & Fleming, J.F. (1972). Historical Development of Cable-stayed Bridges. J. Struc. Div.,
Proceeding ASCE. v98, September.
Pugsley, A. (1968). The theory of suspension bridges. London: Edward Arnold.
Steel Construction Institute. (1998). Steel Designers Manual, 5th Edition. Oxford: Blackwell.
Stasky, J. (2005). Stress ribbon and cable-supported pedestrian bridges. London: Thomas Telford.
Troitsky, M.S. (1998). Cable Stayed Bridges: Theory and Design. 2nd edition. Oxford: BSP
Professional Books.
Troyano, L.F. (2003). Bridge Engineering - A Global Perspective. London: Thomas Telford.
Walther, R. Houriet, B. Isle, W. Moia, P. (1988). Cable stayed bridges. (2nd Edition) London: Thomas
Telford.
Wenzel, H. (1998). Cable Stayed Bridges History Design Application. Germany: Northern Gate
Book Co. Ltd.

© Jonathan Haynes version 2.0 Page 185 of 239

32900 Bridge Engineering
MSc Structural Engineering

7. Durability of Bridges
This section will identify some of the major durability problems encountered in bridge
engineering, and provide insight into the material specific mechanics which lead to problems.
Common protection and remediation systems will also be presented.

You should also review: 3 Bearings parapets & services.ppt, Bridge Durability.ppt and Bridge
Inspection, Monitoring and Assessment.ppt on Blackboard
Principal References: SCI P154. Design of steel bridges for durability. Concrete Bridge
Development Group. Current Practice Sheet No 2 - Bridge durability.

7.1 Durability Issues for Bridges

Bridges are expensive items of infrastructure and their indirect contribution to the economy of a
country is significant (consider the time cost and environmental damage of diverting all the
traffic on a motorway for one week).
All bridges must be designed to last more than a century, in which time there is the possibility of
many changes in:
• Loading magnitude – will larger / faster vehicles be permitted
• Environmental conditions and chemistry – will harsher storms be more frequent, or road
salting be required more frequently, or routine maintenance be less frequent
• Supporting soil condition – will soil become saturated or removed, or adjacent structures
changed which increase soil stress
• Construction material integrity – will the bridge material become weaker or less stiff with
time, or could matrix integrity break down due to chemical attack
Many long term problems are created at the design stage by failing to consider movement of
water in the finished bridge - where does water from the deck go ? How does water which
penetrates expansion joints escape ?
Avoid creating sharp external corners in reinforced concrete (include a 450 chamfer).
Avoid welding steel elements which bend in orthogonal directions.
Design bearing shelves to allow for jacking and bearing replacement.

7.2 Fatigue
Bridges are by their nature subject to moving actions. This induces stress reversal (from tension to
compression and vice versa) producing the perfect conditions for fatigue crack growth. The
consequences of fatigue failure are severe - collapse without warning, and so must be
addressed as a separate limit state.
Some bridge types are less likely to be susceptible to fatigue, these include:
• pedestrian bridges because the magnitude of actions is small compared to self weight
• buried culverts because the dead weight of overburden prevents stress reversal

7.2.1 Concrete
Reinforced concrete decks are the elements likely to be most susceptible to fatigue conditions in
concrete bridges. However, physical testing shows that real stress ranges are much lower than
predicted by elastic analysis; for this reason EN1992-2 6.8.1(102) lists situations where fatigue
checks may be conservatively circumvented:
• Non wind-sensitive footbridges
• Buried frame structures
• Foundations
• Piers, abutments and columns not rigidly connected to superstructure
© Jonathan Haynes version 3.0 Page 186 of 239
32900 Bridge Engineering
MSc Structural Engineering

Further, when considering wheel loads applied directly to decks, fatigue verification may be
avoided when the slab contains no couplers or welded reinforcement, is proportioned such that
L/d < 18, the slab acts compositely with beams and the slab width is at least three times its clear
span.

7.2.2 Steel
Methods to address fatigue in steel bridges are detailed in BS EN 1993-1-9 but BS EN 1993-2 Annex
C gives specific detailing procedures for steel bridge decks containing stiffeners.
Stress raisers may be created in steel structures when the flow of forces through a structure are
interrupted or diverted. The principal parameters are the stress range and number of stress
cycles. Generally, sharp corners should be avoided. It should also be noted that welding will lock
zones of elevated stress into sections, particularly if the weld joints two dissimilar thicknesses of
plate.
The growth of a fatigue crack in steel may be arrested by drilling a hole at the ends of the crack.
This simply removes the stress raiser at the end of the fatigue crack. An assessment of the
structure containing the crack must be undertaken to determine whether a repair is needed,
and its configuration.

7.2.3 Steel Cable

Steel wire is drawn (cold worked) which work hardens the material (reducing ductility), and it is
reeled after forming which induces tensile stress when straightened into its final position. The
cable microenvironment contains atmospheric contaminants, zinc and water. Being rich in
hydrogen molecules, they can initiate crack development in steel.
Detailing of cables connections can be important. Hanger connections must
allow for cable displacement on two axes or unexpected bending stress is
induced, which increases the tensile stress in the cable.
Corrosion fatigue is a phenomena which produces cracking as a result of
fluctuating tensile stress in a corrosive environment, although no corroding
electrolyte is required. This reduces the fatigue resistance of the material and
is often initiated at corrosion pits.

7.3 Degradation Systems

Each construction material is susceptible to degradation by different processes under varying
environmental conditions.

7.3.1 Concrete
Concrete durability can be defined as its ability to resist attack from the environment during its
service life. The nature of the attack can take two basic forms:
• Physical (abrasion, impact, salt crystallisation, ice growth, permeation/diffusion)
• Chemical (sulphates, chlorides, carbon dioxide, alkali-silica/carbonate reaction)
The durability of concrete is primarily related to its permeability and diffusion to liquids, gases and
ions. The intrinsic properties of concrete governing permeability and diffusion are both the total
pore volume and the relative sizes of the pores existing in the hardened cement paste. The size
of ions or gas molecules are an order of magnitude smaller than those of typical gel pores.
High water/cement ratios lead to an increase in pore volume and a corresponding increase in
permeability. The use of blended cements containing pulverised fuel ash (pfa), ground
granulated blast furnace slag (ggbs) or silica fume can result in a potentially finer pore structure
and therefore a less permeable paste. Such cements are, however, more sensitive to poor curing
than Ordinary Portland Cement (OPC).
© Jonathan Haynes version 3.0 Page 187 of 239
32900 Bridge Engineering
MSc Structural Engineering

Providing concrete is adequately cured, the hydration process will continue with age leading to
a finer pore structure and reduced permeability. The concrete near the surface of a structure is
far more susceptible to lack of curing than that remote from the surface. Lack of curing,
particularly during the first three days, will result in a highly permeable surface layer but will have
little effect on the zone more than about 50 mm beneath the surface. It is this surface layer
which is most important in terms of durability. Inadequate compaction will result in an increase in
porosity and a reduction in durability. The formation of plastic cracking as a result of poor curing
will aid the ingress of deleterious materials leading to a reduction in durability. Permeability is
even more sensitive than compressive strength to water/cement ratio and degree of hydration.
There is therefore, no definitive relationship between permeability and strength and to make an
assessment on durability purely based on compressive strength can be misleading.
During reinforcement corrosion, the alkalinity of the concrete is lowered below a critical pH of
10.5, due to the ingress of chlorides or carbon dioxide (carbonation). The pH of good quality
concrete is about 12.8. The rate at which carbon dioxide penetrates into the concrete
decreases with time, because of the formation of calcium carbonate which tends to block the
pores and reduce the permeability of the surface concrete.
Sulphates are found in soils (ground water), particularly clay, sea water, sewage and in some
aggregates. They can react with the calcium aluminate hydrate and in some cases the calcium
hydroxide in the cement paste, both reactions result in volume expansion and eventual
disruption of the concrete matrix:
• Ca3A reaction, all forms of sulphate react with calcium aluminate hydrate producing
ettringite, the volume expansion factor is about five.
• Ca(OH)2 reaction, takes place between Ca(OH)2 and sodium or magnesium sulphate; the
volume expansion factor is about two.
The risk of sulphate attack can be minimised by use of low permeability concrete and good
material selection:
• use cements with low Ca3A content (≤ 3.5%) i.e. sulphate resisting Portland cement or
cements containing 25% - 40% pfa or 70% - 90% ggbs
• aggregates contaminated with sulphates should not be used, a maximum limit of 4.0% SO3
content by weight of cement is recommended
• for very severe exposure conditions protective coatings must be applied to the concrete as
well as using SRPC
The highly alkaline nature of OPC concrete makes it very prone to attack by Acids; as a general
rule attack will occur at a pH below 6.5. No Portland cement is resistant to acid attack and the
only effective solution is to use some form of physical barrier such as bituminous or resin coatings,
or to line with stainless steel.
Concrete can deteriorate as a result of the reaction between alkaline pore fluids and reactive
minerals in certain types of aggregates. The mechanism of deterioration is known as Alkali-
Aggregate Reaction and can occur in a number of forms - alkali-silica, alkali-silicate and alkali-
carbonate. The most common of these is the alkali-silica reaction (ASR).
The alkalis taking part in the alkali-silica reaction are those of Calcium (Ca), Potassium (K) and
Sodium (Na) coming primarily from the cement. Certain aggregate types contain forms of
amorphous silica which react with these alkalis in the pore fluid to form an alkali-silica gel. In time,
as this gel takes in water, it expands causing disruption of the concrete.
The risk of ASR can be minimised by using materials of low alkali content. It is recommended that
the total available alkalis in the concrete should be below 3.0 kg/m3. This may be achieved by
using low alkali Portland cements (Na2 0 equivalent < 0.6%) or by using blended cements
containing pfa, ggbs, silica fume or other natural pozzolana. The reaction can only take place in
the presence of water, so structures that are kept in a dry environment are unlikely to be at risk.
The mechanism of Frost Attack in hardened concrete is a combination of hydraulic pressure (as
the water in the cement pores freezes there is a 9% increase in volume, creating disruptive forces
within the hardened cement paste) and osmotic pressure (pore water within the cement paste
contains dissolved salts in varying concentrations which means that not all the water will freeze

© Jonathan Haynes version 3.0 Page 188 of 239

32900 Bridge Engineering
MSc Structural Engineering

at the same time). In addition the differences in ionic concentrations within the unfrozen pore
water leads to the creation of disruptive osmotic pressures.
The risk of frost attack is minimal in low permeability concrete, as the degree of damage
depends to a large extent on the degree of saturation of the pores within the cement paste. If
the pores are only partially saturated there will be room for expansion as the ice forms and the
damage caused will be minimal.

7.3.2 Steel
Steel degrades by corrosion which can manifest by several different mechanisms:
Atmospheric corrosion occurs in the presence of oxygen and hydrogen and is characterised by
a uniform loss of material over the entire surface. Contact with an aqueous solution is generally
required and a thick layer of red/brown rust Fe2O3 is normally the corrosion product. The rate of
corrosion depends upon atmospheric conditions such as wet/dry cycle time, humidity and
temperature. In highly alkaline conditions (pH>10) corrosion will be very slow.
Pitting corrosion attacks the wire locally at locations of uneven protection such as notches in
galvanising. The rate of corrosion is usually largest perpendicular to the metal surface.
Crevice corrosion occurs at cracks when there are differing levels of oxygen in the corroding
electrolyte from the parent body. Once the oxygen at the crack is depleted corrosion is locally
concentrated to produce a result similar to pitting corrosion.
Stress corrosion cracking can happen when corrosive attack occurs at the same time as
sustained tensile stress, and is most prevalent in highly stressed high strength steel. It produces
brittle fracture in stress ranges where the steel is expected to be ductile. Failure is characterised
by transverse branching stress cracks, and is believed to be caused by local absorption of
corroding compounds at points of high stress (where cracks are most likely to initiate).

pitting corrosion in a galvanised steel wire

stress corrosion cracking showing the

characteristic transverse cracks

Hydrogen cracking is closely associated with stress corrosion cracking and is prevalent in stressed
high strength steels subject to a hydrogen rich atmosphere such as high relative humidity or
some acidic environments. This failure is common in martensitic steels (a highly strained body
centred cubic form of ferrite that is supersaturated with carbon) which have been subject to
high temperatures in the presence of hydrogen; conditions which high strength steel wires are
subject to in the galvanising process.
Electrolytic corrosion may occur when steelwork becomes part of a grounded electrical circuit
where the steel forms an anode and corrodes. Cathodic protection may be used to prevent or
stop the corrosion, usually be attaching a sacrificial anode such as a zinc block.
Weathering steel is an alloy which suffers a lower corrosion rate than standard mild steel by
forming a protecting surface patina of rust. It is often known by its trade name CorTen. It is
intended to be unpainted in external applications and is easily identified by its rich brown colour.
Alloying elements include carbon, silicon, manganese, phosphorous, sulphur, chromium, copper,
vanadium and nickel. It can corrode at different rates so detailing must eliminate pools of
standing water or dirt traps.
The corrosion rate is so low that bridges formed in unpainted weathering steel can achieve a 120
year design life with little maintenance. A weathering steel bridge in an appropriate environment
provides an attractive, very low maintenance, economic solution.

© Jonathan Haynes version 3.0 Page 189 of 239

32900 Bridge Engineering
MSc Structural Engineering

Weathering steel should not be used in atmospheres where there are high concentrations of
corrosive chemicals such as SO2 (> 200mg/m2/day). It should not be used in situation where the
steel will be permanently wet, such as underwater or buried in the ground, as this prevents the
surface patina from forming and stabilising.
Corrosion protection
Structural steelwork can be protected by providing a moisture proof barrier such as painting or a
sacrificial surface layer such as galvanising or aluminium spray. The Highways Agency (Manual of
Contract Documents for Highway Works) and Railtrack (NR/L3/GN/039) provide different
specifications, such as the excerpts below:
Minimum
Preparation Metal coat Paint coats Finish coat total dry film
thickness

Micaceous iron
Zinc phosphate Polyurethane 2
oxide high build
high build quick pack finish or
HAi Sa 2½ quick drying 300µ m
drying epoxy 2 moisture cured
epoxy 2 pack
pack primer polyurethane finish
undercoat

Aluminium Micaceous iron

Zinc phosphate Polyurethane 2
metal spray oxide high build
high build quick pack finish or
HAii Sa 2½ 100µ m and quick drying 300µ m
drying epoxy 2 moisture cured
Aluminium epoxy 2 pack
pack primer polyurethane finish
epoxy sealer undercoat

Thermally
sprayed High solids epoxy Anti-graffiti paint –
N1 Sa3 Epoxy sealer polyurethane finish 300µ m
Aluminium or primer 150µm
Zinc 100µ m 50µm or acrylic
urethane 50µ m or
Epoxy blast primer Epoxy glass flake polysiloxane 50µ m
N2 Sa 2½ 475µ m
25µm 400µ m

The systems are intended to be maintenance free for twelve years, with major maintenance
from 20 years as identified by bridge inspections.

7.3.3 Masonry
Masonry is the only bridging material known to have survived for several millennia. It is however
susceptible to basic mechanical degradation such as freeze-thaw and it is important to specify
frost resistant bricks in all civil engineering applications (combined with deck waterproofing).
However, masonry is the least reliable construction material and relies upon adequate
workmanship for durability.
Masonry brickwork subject to cycled stress, such as arch bridge rings, is known to suffer
degradation in capacity. The number of variables involved is great and it is therefore difficult to
obtain quantified advice; however, in comparison to ULS static load tests:
• Weak bricks reduce fatigue strength by up to 20%
• Strong bricks reduce fatigue capacity by up to 57%
• Saturated bricks reduce fatigue strength by up to 20%
• Saturated bricks reduce fatigue stiffness by up to 10%
Masonry exhibits different strengths and stiffnesses depending upon the type of loading it is
resisting. Although masonry is generally regarded as a brittle material, it will exhibit some plasticity
when also subject to a constant compressive forces on the bedjoints. This is why structural
masonry should never be built in a stack-bond arrangement (the complimentary shear is only
resisted by mortar).

© Jonathan Haynes version 3.0 Page 190 of 239

32900 Bridge Engineering
MSc Structural Engineering

σ2
σ2

masonry loaded parallel and perpendicular

to the bed joints can sustain almost equal
σ1 σ1 stresses in both directions but is prone to
failure under uniaxial load due to lack of
confinement stress

σ2 σ1
σ2
σ1 σ2
when the direction of stress is rotated
(introducing shear) the range of acceptable
stress interaction is considerably reduced

these two relationships can be used to

explain the behaviour of real masonry

σ2 σ1
σ1

σ σ

δ δ
Tension – brittle loss of integrity Compression – softening ductile
loss of strength

Shear without precompression Shear with precompression – when subject to tension (no confining stress)
τ – brittle loss of integrity softening ductile loss of strength there is an elastic response until brittle failure
when subject to compression (no confining
stress) there is a near elastic response until
brittle failure but since the units remain in
contact the behaviour is apparently ductile
with a softening stiffness
when subject to shear (no confining stress)
the behaviour is essentially tensile (brittle);
however when confining stress is added the
behaviour changes to that of compression
δ (ductile with softening stiffness)

7.3.4 Steel Cable

Wire is manufacture in the same way as structural mild steel to a chemical composition which
limits elements other than iron (carbon 0.75 - 0.85%, manganese 0.55 – 0.75%, silicon 0.15 – 0.30%,
phosphorous <0.04%, sulphur <0.04%). The wire coil is then passed through a lead bath, pickled,
drawn to the required diameter, cleaned, hot dip galvanised and reeled.
The most common form of suspension cable is insitu spinning. This involves unreeling wire into the
final suspension cable profile using a spinning wheel; the wire being looped over strand shoes at
the anchor blocks to create a cable which is essentially one continuous wire. The individual wires

© Jonathan Haynes version 3.0 Page 191 of 239

32900 Bridge Engineering
MSc Structural Engineering

are combined into parallel strands, placed in a hexagonal shape which is later repacked into a
circle and held in shape by steel bands. The final cable contains approximately 20% voids.
A layer of corrosion protection is then applied to the perimeter of the cable, this was traditionally
red lead paste. A further layer of circumferential wire wrapping is applied with a waterproof
outer coating (such as neoprene, fibreglass or paint).
It is now understood that red lead paint dries out completely (even in marine environments)
allowing the passage of salts and other corrosive compounds to the steel wire below.
Since corrosion of steel requires oxygen and an electrolyte, it is easiest to eliminate the
electrolyte, thus preventing continuation of corrosion. This is achieved by de-humidifying the
cable. The cable must be sealed using an elastomeric sheet, usually two overlapped layers. Dry
air (RH≤10%) is then pumped into, and vented out of the cable at intervals. If the cable internal
environment is held below 40% relative humidity, corrosion should not begin, or continue.
Corrosion of wires will only progress once the galvanic zinc coating has preferentially
deteriorated and iron is exposed to an electrolyte such as water.

a) cable coated in red lead paint then wound in circumferential wire, b) cable wedged open for condition inspection, c)
cable repainted and sealed in double overlapping elastomeric sheet as part of a dehumidification system, d) an exhaust
port on a cable dehumidification system

Forth Road Bridge cable dehumidification system, 10% RH air is pumped in at green triangles and exhausted at red
triangles. After approximately a year the exhaust air RH was consistently below 40%.

7.4 Sustainability Issues

7.4.1 Bearing Replacement
One of the most common problems encountered in post WW2 bridges in the UK is bearing failure.
This is not a significant problem in itself as bearings are expensive but small and easily obtained.
Physically removing and replacing the bearings can be very difficult if the bridge was not
designed with consideration of replacement.
Design for bearing replacement would include provision of
space adjacent to each bearing where a hydraulic jack can
be installed. Any reinforcement specifically provided at the
bearing to deal with local stresses should also be included at
jacking points. This will permit the bearings to be replaced
whilst the bridge is in use.

© Jonathan Haynes version 3.0 Page 192 of 239

32900 Bridge Engineering
MSc Structural Engineering

7.4.2 Safe Demolition

Demolition is one of the most challenging and dangerous activities in construction. It requires as
much, if not more, consideration than planning a construction sequence. The use of temporary
works will inevitably be necessary which require careful design.
Ideally, a bridge should be demolished using the reversal of its construction sequence. However,
even in simple demolition sequences there will be problems such as corroded fastenings,
jammed movement joints and bonded bearings. Risk management principles must be applied,
together with contingencies and alternative procedures.
Some points worthy of consideration:
• Competence of the contractor – previous experience or specialist advice
• Stability of cranes or trestles (and supporting ground) subject to eccentric or impacting loads
from swinging structure or unsticking or frictional forces
• Protection of services such as railway lines or sewer pipes
• Sufficient storage or lay-down space if material is not to be immediately removed during
demolition
• Possibility of disproportionate collapse in continuous structures such as multi-span arches
Prestressed and post-tensioned concrete structures are generally regarded as dangerous to
demolish. However, this can be overcome by careful de-construction (balancing cutting out
strand with removal of loading, so that concrete compressive strength are never exceeded) so
that an explosive response is avoided.
There is great potential for demolition activities to release contaminated material into the
environment, so there must be a Health, Safety & Environmental Plan in place before works
begin. Permits may be required from the Environment Agency and DoT. The two major
considerations are airborne contaminants, which are likely to affect site operatives; and water
borne contaminants which are likely to affect adjacent populations. Of particular concern is
heavy metal contamination of drinking water sources (old paint systems often contain
dangerous levels of lead and iron will leach from corrosion products).

7.4.3 New Concrete

Concrete is known to have a high carbon footprint (Portland cement 930kg/T) but can be
drastically reduced by the specification of cement replacements such as GGBS (52kg/T) or Fly
ash (4kg/T). There may be greater transport costs associated with using many constituents but
there is also a reduction in land fill as fly ash and ggbs are waste products.
BS8500 limits replacement and lists acceptable mixes but 50% replacement is easily attainable
with no loss of durability (the workability can reduce with high proportions of replacement as the
material is finer and can set faster). Increasing cement replacement will reduce early age
strength but by 7 days, 70% ggbs replacement is only 5% lower than Portland cement and by
28day s the difference is negligible.

7.4.4 Foundation Scour

Bridge piers founded in rivers or estuaries may be subject to channel bed scour, which will
potentially undermine and destabilise the pier. This may only happen during flood events, and so
can take many years to become apparent.
The principal cause of scour is the hydraulic jump – where supercritical flow is forced to change
to subcritical flow. This will naturally occur when the channel gradient is reduced or where the
channel temporarily narrows; the second option may be artificially created when piers are
constructed in a channel. Ponding of water will be observed upstream of the pier (subcritical
flow), with drawdown of the water level at the pier (supercritical flow). The changed to
subcritical flow accompanies a loss of energy, which manifests as turbulence that can disturb
and transport the channel bed material – this is scour.
As the channel bed profile changes, the hydraulic jump migrates upstream until it reaches a
critical point which scours material from the leading edge of the pier and deposits it at the
leeward edge.
© Jonathan Haynes version 3.0 Page 193 of 239
32900 Bridge Engineering
MSc Structural Engineering

subcritical flow supercritical flow subcritical flow ds

h
h < 5w h ≥ 5w

w
ds 2.0w 3.0w
L

1.2w 1.8w

1.5w 2.25w

sloping leading edge

1.5w 2.25w

≥ 20 o
1.0w 1.5w

values of ds for skew pier, h<5w

α L
w
4 8 12
α
15 1.5w 2w 2.5w
30 2w 2.5w 3.5w
45 2.5w 3.5w 4.5w

Scour in cohesionless soil is dependent upon grain size and density, whereas scour in cohesive
soils is dependent upon a wide range of characteristics such as particle cohesion and water pH.
Piled foundations are significantly better at resisting scour than spread foundations.
The most common remedy or preventative measure is the inclusion of a rip-rap apron, which is a
substantial layer of large rocks around the pier. It may be placed at the surface as a remedial
measure, or placed at the maximum scour depth with the natural channel material replaced on
top to maintain flow characteristics.
>4w

flow extent of rip-rap on plan shown dotted

thickness to be at least twice the nominal size of

stone used to form rip-rap
>1.5w

With advice on the specific hydraulic characteristics of the channel it is possible to predict the
depth of general scour in the channel and local scour at the pier. It is then possible to construct
the pier and foundation at sufficient depth to avoid probable future undermining.

high water level

scoured channel
low water level bed profile
overall depth of scour
local scour

general scour depth

© Jonathan Haynes version 3.0 Page 194 of 239

 32900 Bridge Engineering
MSc Structural Engineering

7.4.5 Masonry Arch Repair and Rehabilitation Systems

A full treaties on possible masonry arch bridge repair systems is provided in CIRIA C656. A brief
summary of the most popular is provided below
System Cost Problem Description
Grouting 1 Voids in backfill caused Pressure grout through holes drilled in intrados.
by water percolation Care needed in pressure control

Concrete saddle 3 Inadequate load Remove surfacing and fill material, replace
capacity with RC ‘saddle’ composite with arch barrel
and spandrel walls
Parapet 2 Parapet has insufficient Reconstruction with RC cavity or reinforced
reinforcement impact capacity with retrofitted reinforcement

Relieving slab 3 Inadequate load Remove surfacing and part of fill material and
capacity replace with horizontal RC slab which spans
b/n abutments
Retro reinforcement 2 Inadequate load Small diameter bars are grouted in slots cut in
capacity or ‘plasticity’ the intrados, to add moment capacity at
problems hinges
Surface thickening 2 Marginally inadequate Additional layer of surfacing added to
load capacity increase load spread through fill material

Radial pinning 1 Ring separation Metal pins are grouted through the barrel
thickness to reinstate composite action
between rings

It is of great importance to identify the cause of the structural problem before embarking upon a
costly repair. However, a large proportion of repairs are undertaken to improve load carrying
capacity following periodic assessment.

Fixing reinforcement to apply a sprayed Slotting an arch to receive retro-

concrete inner ring reinforcement to the intrados

Patras plates fixed across the bridge to prevent Removal of fill from the arch barrel to
further spreading of spandrel walls facilitate construction of a relieving slab

© Jonathan Haynes version 3.0 Page 195 of 239

 32900 Bridge Engineering
MSc Structural Engineering

The process of assessing load carrying capacity can be complex for a masonry arch bridge, so
the effort expended should be proportional to the type of bridge and scale of load capacity
required. The SMART assessment system seeks to quantify where to target the effort and identify
the likely problems in the stock of old and possibly overloaded masonry arch bridges.
Simple bridges may be assessed with straight forward systems like MEXE, whilst most challenging
bridges will require greater, and more specialist knowledge to create probabilistic models.
Initial Intermediate Advanced
Square, segmental, < 6m Square, elliptical, < 15m Skew, multi-span, >15m span
span span
11 Poor condition
Good condition Moderate condition

Geometry, construction + exploratory + geometry and construction

2 inspection and condition investigation survey to extents of influence
survey
Determine loads from Determine loads from Determine loads from intended
3 intended future use intended future use future use

Identify materials used Determine material Determine material properties of all

4 properties of critical elements
elements
Elastic rib or MEXE 2D rigid block using Sophisticated 3D FEA or Discrete
5 actual material and Element (DE) analysis using soil-
section properties structure interaction
Determine ULS load Determine ULS load Determine ULS load accounting for
accounting for other failure accounting for other all possible failure modes. Apply
6
modes such as ring failure modes such as stress cycle behaviour theory
separation ring separation
Determine PLS working load Determine PLS working Determine PLS working load
capacity (apply load load capacity (check capacity (check stress range using
7
factor) stress range on elastic 3D model and parametric study)
model)
Apply safety / condition Ensure residual life by Ensure residual life by probabilistic
8
factors stress control stress control

7.4.6 Fibre-reinforced Polymer Composites

Fibre-reinforced polymer (FRP) composite can be used to repair or strengthening steel and
reinforced concrete bridges. Typically they are bonded to the surface of the structure to
enhance its tensile strength. There are three components which comprise a FRP composite
strengthening system:
• high strength fibres such as carbon, aramid or glass
• a polymer matrix which binds the fibres together
• an adhesive which bonds the composite material to the structure
FRP composites have good corrosion and fatigue resistance and have a high strength to weight
ratio. They can be rapidly installed and the requirements for temporary works may be
significantly reduced. Techniques used for FRP strengthening are:
• prefabricated plates, manufactured either by pultrusion or preformed pre-impregnated
sheets bonded with a two-part epoxy adhesive system that cures at ambient temperatures,
most structural strengthening uses this system
• wet lay-up systems, made from pre-impregnated sheets or woven fabrics
• vacuum infusion
• filament winding such as automated wrapping of columns
© Jonathan Haynes version 3.0 Page 196 of 239
32900 Bridge Engineering
MSc Structural Engineering

Strengthening with FRP composites can increase the axial tension, flexural tension and (to a
limited extent) shear capacity of a section. This can lead to improved stiffness and fatigue life.

Shear strengthening of a Axial strengthening of a Flexural strengthening of a

beam reaction column by preventing bursting deck soffit

Three FRPs are commonly used; glass fibre reinforced polymers (GFRP), carbon fibre reinforced
polymers (CFRP) and aramid fibre reinforced polymers (AFRP). The mechanical and thermal
properties of a composite depend on the properties of the fibres, the properties of the matrix, the
amount and the orientation of fibres.
Mechanical Tensile Flexural Coefficient of Failure
Properties of FRP UTS E UTS E thermal expansion, strain
(N/mm2) (N/mm2) (N/mm2) (N/mm2) α (10-6/°C) (%)

Expoxy adhesive 100 4 50

Glass S-2 1700 52 5.7
Carbon PAN IM 2500 150 1590 110 1.7
Aramid 49 2750 125 2.4
FRP can be used to increase the stiffness of cast and wrought iron but it is most commonly used
to act as additional tensile reinforcement for steel and reinforced concrete structures. Steel is a
ductile material exhibiting significant plastic flow at yield so that significant redistribution of
stresses can occur. Lower modulus but higher strength CFRP composites are commonly used to
strengthen steel bridges.
FRP strengthening relies heavily on the quality of workmanship, so quality testing must be
undertaken to ensure repairs will be successful. To strengthening metal bridges, the adhesive joint
is critical. so all loose material must be removed and metal surface thoroughly cleaned, followed
by abrasion. Two-part epoxy adhesives will not cure at temperatures below 5°C so heating may
be needed during night possessions or winter working.
It should be noted that FRP composites have anisotropic material properties with significantly
greater strength in the direction of the fibres (warp and weft).
The design concept of externally bonding FRP to metallic structures is like any other composite –
the adhesive joint must prevent slip - transfer high longitudinal shear stresses through the
adhesive. Any discontinuities in the structure will develop stress raisers in the adhesive, creating
high local shear and peel stresses. These discontinuities are typically at the end or edge of the
strengthening material, at bond defects or where cracks or joints occur in the metal.
Although elastic analysis is used to evaluate critical stresses at discontinuities, plasticity of the
adhesive will relieve peak stresses, so that elastic analysis will overestimate the magnitude of
stresses.
The coefficients of thermal expansion of the FRP composite and the metal will be significantly
different, and CFRP may have a negative coefficient of expansion so that it contracts when
heated. However the typical resins used within the composite matrix have high coefficients of
expansion so the overall coefficient for the composite material is typically positive. The difference
in the coefficients of thermal expansion can cause significant shear stresses in the adhesive joint.
Stresses in the substrate material should be checked - cast iron elements may be overstresses just
by adding CFRP.
Many of the FRP composite systems use ambient cured two-part epoxy resins. These provide a
good and durable bond of the FRP composite to the structure. However these materials have a
glass transition temperature which is typically between 55°C to 70°C. As the temperature of a

© Jonathan Haynes version 3.0 Page 197 of 239

32900 Bridge Engineering
MSc Structural Engineering

polymer increases towards its glass transition temperature the polymer softens and the shear
strength is dramatically reduced. It is important that the in-service temperature is considerably
below the glass transition temperature.
Galvanic corrosion can occur when metals and conducting non-metals contact. The risk of
corrosion depends on where the materials are ranked on the electro-potential series and as
carbon is widely separated from both steel and iron, the risk of galvanic corrosion is high. While it
is possible that the adhesive will provide an adequate layer of insulation, it is recommended that
a glass fibre (or other insulating fibre) layer is placed between the metal and the CFRP. When
strengthening weathered iron or steel, corroded areas should be grit-blasting and surface
ground. Where possible, bonding to corroded areas should be avoided as corrosion can
sometimes reactivate.
FRP can be used to construct entire footbridges. Some of the advantages and disadvantages
are:
• ultra-lightweight e.g. 2 Tonnes for a 12 metre span (88-171 kg/m2 without a wearing surface)
• quicker and easier to install minimising possession times and reducing crane requirements
• less substantial foundations, piers and temporary works
• no corrosion, reducing maintenance costs
• modular applications particularly desirable
• high strength to weight ratio
• chemically inert
• CFRP stay cables can be used rather than high strength steel
• Young’s modulus of elasticity GFRP composites are one-fifth of steel, so likely to be deflection
limited
The Aberfeldy Footbridge was completed in 1992 to connect two parts
of a golf course over the River Tay, this was the world's first FRP
footbridge. It is a three-span cable stayed footbridge of 25m, 63m and
25m span. The towers are 17m tall A-frames, giving an inclined cable
arrangement which helps stabilise the bridge deck laterally.

A standard deck arrangement is constructed from GFRP components.

The cables are Parafil Kevlar composites. Kevlar has a good strength to
weight ratio but has a lower stiffness than steel which means
serviceability dominates the design – the deck is so light it must be
counterbalanced with concrete to achieve acceptable dynamic performance in cross winds.

7.5 Bridge Condition Assessment

Although a bridge will be designed for a specific life span and load carrying capacity, there
may be changes to some design parameters during this period:
• Increased variable action or load width requirement
• Loss or change in ground conditions at the supports
• Loss of material due to degradation or impact damage
A bridge, like any structure, must have a Safety File which contains the design information and
as-constructed drawings which the bridge owner retains. It is a statutory requirement to
periodically determine the adequate condition of publically owned bridges in the UK.
Once a bridge has been assessed and problems have been identified, they must be
immediately addressed (repaired), or if not life threatening they may be monitored. This is
sometimes periodic (one, two or six year intervals) or if failure could be more imminent,
continuous monitoring may be preferable. Structures which have been significantly repaired are
often continuously monitored to determine if the repair has worked (e.g. cathodic protection) or
to monitor environmental conditions which the repair depends upon (e.g. cable
dehumidification).

© Jonathan Haynes version 3.0 Page 198 of 239

 32900 Bridge Engineering
MSc Structural Engineering

7.5.1 Periodic Assessment

The UK Bridges Board (Highways Agency) publishes The Inspection Manual for Highway Structures
Volume 1: Reference Manual which contains background information on bridge structures and
comprehensive advice on the procedures and scope of periodic inspections, including health
and safety considerations.
Some common defects to look out for during a visual inspection:
• Paint system failure on steelwork
• Corrosion staining on concrete surfaces
• Collision damage, particularly on thin element steel beam bridges
• Shear (diagonal) cracking around bearings or half-joints in concrete bridges
• Surface cracking in concrete plates, which does not follow reinforcement lines, this may
indicate sulphate or chloride problems due to high surface porosity
• Local or global buckling of steel elements, particularly secondary elements such as bracing
• Adequate function of surface drainage system
• Water leaching through solid structures such as masonry retaining walls or barrel soffits.

buckled flange of a steel plate girder following a ship strike; local buckling of longitudinal stiffeners in a steel box
girder; advanced corrosion of deck slab soffit reinforcement which has led to complete loss of cover; paint
protection failure at a the bearing of a riveted steel truss leading to significant corrosion and section loss.

If a visual inspection reveals problems, it may be necessary to remove material samples for
strength testing. This would often include coring concrete cylinders (from are not subject to high
compression forces), or removing strand from post-tensioned concrete beams. Material testing
and data interpretation is a specialist activity.

7.5.2 Monitoring Systems

Many levels of complexity are available for monitoring systems, from crack gauges to full real-
time logged instrumentation. Some of the more common systems are:
Avonguard tell-tale crack monitoring gauges which are glue to each side of a crack (masonry or
concrete) and read periodically to check the growth and size of a surface crack.
Scour (removal of foundation material) and flood level monitoring, as many bridge collapses are
caused by hydraulic movement.
Displacement monitoring using dial gauges or electrical transducers. This may be vertical or
rotational. Accelerometers may also be used to determine dynamic sensitivity and natural
frequency under loading. This may be under normal working conditions (in-service) or load test
conditions, where SLS and ULS design loads may be applied to the bridge deck to assess load
carrying capacity. Where such systems are permanently left in place, they are referred to as
condition or health monitoring systems.

© Jonathan Haynes version 3.0 Page 199 of 239

32900 Bridge Engineering
MSc Structural Engineering

Durability & Sustainability of Bridges - Conclusion

Things to do now :
1. Review the Powerpoint lectures on Blackboard.
2. Attempt the Self Assessment Exercises below. Then check your answers against the
solutions on Blackboard.
3. Have a rest.

4. Prepare your group submission for Assignment 2 & 3.

Self Assessment Exercises

1. For the bridge designed in sections 2 and 3, identify grades of concrete and steel which will be
adequate for the environment in which the bridge is to be sited. Explain your reasons.
2. With enhanced durability as the primary goal, suggest best practice details you would adopt
for bolted steelwork connections, bearing fixings and expansion joints.
3. Detail the systems you would adopt to prevent corrosion of steelwork beams and foundation
reinforced concrete.
4. Suggest an outline specification which could be used to repair a deck slab soffit which has
suffered localised corrosion expansion around main reinforcing bars, leading to some loss of
cover concrete. Include the surface preparation, repair products and possible testing required.

Further Reading
Brown, C.W. & Iles, D.C. (1995). Design of steel bridge for durability. SCI Publication 154. Ascot:
The Steel Construction Institute.
Davies, R.D. & Buenfeld, N.R. (2007). Automated monitoring of the deterioration of concrete
structures. London: Department of Trade & Industry.
Highways_Agency. (2006). BA 86/06. Design Manual for Roads and Bridges. Volume 3 Highway
Structures - Inspection and Maintenance. Section 1 Inspection. Part 7 Advice Notes on the Non-
Destructive Testing of Highway Structures. Norwich: H.M.S.O.
Illston, J.M. and Domone, P.L.J. (Eds.) (2001). Construction Materials: their nature and behaviour.
3rd Edition. London: Spon.
Page, J. (1996). A guide to repair and strengthening of masonry arch highway bridges.
Crowthorne: Transport Research Laboratory.
McKribbins, L.D., Melbourne, C., Nisar, S., & Gaillard, C.S. (2006). C656 Masonry arch bridges:
condition appraisal and remedial treatment. London: C.I.R.I.A.
Ryall, M.J. (2010). Bridge Management. 2nd Edition. Oxford: Butterworth Heinmann.
Spencer, P.C., Hendry, C.R. & Petty, R. (2012). Quantification of sustainability principles for bridge
projects. Bridge Engineering. 165: BE2, pp 81–89.
Stahl, F.L. & Gagnon, C.P. (1996). Cable corrosion in bridges and other structures. New York: ASCE
Press.
Tilly, G.P., Matthews, S.J., Deacon, D., De Voy, J. & Jackson, P.A. (2008). C664 Iron and steel
bridges: condition appraisal and remedial treatment. London: C.I.R.I.A.
Transport Association of Canada. (2004). Guide to Bridge Hydraulics. 2nd Edition. London: Thomas
Telford.

© Jonathan Haynes version 3.0 Page 200 of 239

 32900 Bridge Engineering
MSc Structural Engineering

8. Design Data
8.1 Section Properties of Fundamental Shapes
Shear Area,
Shape A Iyy Izz Ixx As
b
2
3
bd 3 db
bd 0.83 A
d 12 12
2

b
3
bd bd 3
2 36
d
3

πD 2 πD 4
D 1 .1 A
R 4 64

d π(D2 −d2 ) π(D4 − d4 )

D 0 .5 A
4 64

b
2
tf
dt w
d
2
tw
b
2
tf
0.85ht w
d

tw

© Jonathan Haynes version 3.0 Page 201 of 239

 32900 Bridge Engineering
MSc Structural Engineering

8.2 Standard Force and Deflection Formulae

Shear
Arrangement Shear Force Bending Moment Bending deflection
deflection
P
P PL PL3 PL
∆m = ∆s =
3EI AG
L

w (per metre)

wL wL2 wL4 wL 2
∆m = ∆s =
2 8 EI 2 AG
L

w (per metre)
L
w
2 5wL4 wL2
∆m = ∆s =
L 384 EI 8 AG
wL2
8

P
P
2 PL3 PL
∆m = ∆s =
L L 48 EI 4 AG
P
2 2
PL
2
4

P
Pa
PL3  4a a 
2
L ∆m =  − 3  
Pb
48EI  L  L  
a b Pab
L L

w (per metre) L2
3wL w
8 wL4 wL2
8 ∆m = ∆s =
9 185 EI 8 AG
wL2
L 128

P 3PL
5P
16 16 7 PL3 PL
∆m = ∆s =
L L 5PL 768 EI 4 AG
2 2 32

P Pb(L2 − b 2 )
Pa 2 2
(b + 2L) 2L 2 Pa 3 b 2
2 L3 ∆m =
3 EI (L + 2 a )
2

a b Pb  3b b3 
 2 − + 3 
2  L L 
w (per metre) L wL2
w 12
2 wL4 wL2
∆m = ∆s =
wL2 384 EI 8 AG
L 24

© Jonathan Haynes version 3.0 Page 202 of 239

 32900 Bridge Engineering
MSc Structural Engineering

P
P PL
2 8 PL3 PL
∆m = ∆s =
L L 192 EI 4 AG
PL
2 2 8
2
P
2
a  b  2 Pa b 2 Pab 2
P    1 + 2  L2 L2
L  L Pa 3b 3
∆m =
3EIL3
2 Pa 2b2
a b
L3

P Pl
P PL2 (L + l )
Pl ∆m =
3 EI
L
L l

8.3 Structural Mechanics Formulae

M f E
Simple bending: = =
I z R
VA' z '
Shear flow : τ =
Ib
T Gθ τ
Simple torsion: = =
r L Ip
L
M ∂U FL ∂U
Castigliano’s Theorem : δ = ∫ EI . ∂W = ∑ AE . ∂W
0

d 2z
Differential equation of flexure : EI = −M
dx 2
∑ Az
Theorem of the Parallel Axis : I NA yy = ∑I + ∑ Ah 2 where z =
∑A
yy

2τ xz σ −σ2
Principal Stress : Tan 2θ = , τ max = 1
σ x −σ z 2
σ +σ z 1 σ +σ
σ1 = x + (σ x − σ z )2 + 4τ xz 2 and σ 2 = x z − 1 (σ x − σ z )2 + 4τ xz 2
2 2 2 2

© Jonathan Haynes version 3.0 Page 203 of 239

 32900 Bridge Engineering
MSc Structural Engineering

8.4 Units and Conversion

With a few exceptions, all measurements are made in specific units. To avoid confusion in later
use, units must be quoted correctly and carried through calculations.

Fundamental Units
Length metre m
Mass kilogram kg
Time second s

Base Units
Force Newton N kg.m/s2
Energy Joule J N.m
Power Watt W kg.m/s
Pressure Pascal Pa N/m2
Angle Radian c or Rad. 1800 = πc
Acceleration m/s2
Volume Litre L m3

Metric Conversion
micro µ 10-6 eg. 1 µm = 1 x 10-6 m = 0.000001 m
milli m 10-3 eg. 1 mm = 1 x 10-3 m = 0.001 m
kilo k 103 eg. 1 kJ = 1 x 103 J = 1000 J
mega M 106 eg. 1 Mg = 1 x 106 g = 1000 kg
giga G 109 eg. 1 GPa = 1 x 109 Pa = 1000000 kN/m2

the mass of 1L of water is 1kg, 1 m3 of water weighs 10 kN

acceleration due to gravity, g = 9.80665 m/s2 at sea level

Imperial to Metric Conversion

1 inch = 25.4mm
1 foot = 12 inches = 304.8 mm
1 yard = 3 foot = 914.4 mm
1 mile = 1760 yards = 1609.344m = 1.609344 km
1 pound (lb) = 0.4536 kg
1 Ton = 2240 lb = 1.016047 Tonne
1 atmosphere = 101.325 kPa (N/mm2)
1 bar = 100 kPa (N/mm2)
1 lb/inch2 (PSI) = 6.895 kPa (N/mm2)
100 lbft = 0.4448 kN
0 Kelvin = -273 °C (Kelvin and Celcius are of equal graduation)
-50 °C = -58 °F, 0 °C = 32 °F, 100 °C = 212 °F, 150 °C = 302 °F
1 mph = 1.6093 km/h

© Jonathan Haynes version 3.0 Page 204 of 239

 32900 Bridge Engineering
MSc Structural Engineering

8.5 Values of Common Structural Material Properties

Elongation at failure (%)

expansion, α (x10-6/oC)
Young’s Modulus,

Coefficient thermal
Property

Poisson’s ratio, υ
Shear Modulus,

Ultimate Stress,
Unit weight,
G (kN/mm2)
E (kN/mm2)

Yield Stress,

σ u (N/mm2)
σ y (N/mm2)
γ (kN/m3)
Material

Mild steel 210 79 0.3 12 77 35 250 420

High tensile steel 200 79 0.3 12 77 18 400 650
304 Stainless steel 180 76.9 0.3 18 80 45 210 520
Concrete (30 N/mm2) 26 11.5 0.13 10 24
Reinforced concrete 25
Aluminium 70 26.3 0.33 13 27 240 260
Timber (D40) 11 4 0.31 30 7 20
Softwood 5
Chipboard / plywood 7
Asphalt 22
Float Glass 70 28.7 0.22 9 25 1000
Brickwork Masonry 5.22 1.96 0.33 22 20
Bronze 100 40 0.34 20 83 5-60 82 200
Wrought Iron 190 75 0.3 12 75 35 210 340
Carbon Fibre 240 20 1.5 3500
Cast iron 0.3 10 73
Wrought iron 170 0.3 12 71
Lead 16 0.44 28 111
Roof slates 29
Clay roof tiles 20
Water 10
clay masonry facing brickwork 22
medium density conc. blockwork 18
insulation board 3
soil (sand) 16-18
soil (clay) 20-22
plaster 17
stone (medium density granite) 29
plastic 21

Coefficient of Friction, µ between surfaces

Metal - metal 0.15–0.60
Metal - hardwood 0.20–0.60
Wood - wood 0.25–0.50
Rubber - paving 0.70–0.90
Nylon - steel 0.30–0.50
PTFE - steel 0.05–0.20
Masonry - masonry 0.60–0.70
Masonry - earth 0.50

© Jonathan Haynes version 3.0 Page 205 of 239

 32900 Bridge Engineering
MSc Structural Engineering

Soil Structural Properties

Young’s Modulus of
Winkler spring stiffness,
Soil description Elasticity, Poisson’s ratio, υ
kw (kN/m3)
E (kN/m2)
Loose SAND
10000-20000 4800-16000
φ = 30°
Medium dense SAND
20000-50000 9600-80000
φ = 35°
Dense SAND 0.15 coarse grain
50000-80000 64000-128000
φ = 40° 0.25 fine grain
Loose SAND & GRAVEL 50000-150000
Dense SAND & GRAVEL 100000-200000

Shale 120000-140000 0.10-0.40

Silt 2000-10000 0.30-0.35

Soft CLAY 5000-15000 12000-24000
Medium stiff CLAY 15000-50000 0.40-0.50 saturated, 24000-48000
50000-75000 0.10-0.30
Stiff CLAY 48000-60000
or 300Cu unsaturated
75000-100000
Very stiff CLAY
or 500Cu
Long term stiffness of draining clays may be estimated as one third of the short term values given above.

If soil testing has been undertaken but no geotechnical interpretative report is available,
the following rules of thumb may be applied:
• In clay soils, ABP = 2Cu where Cu is the unconfined compression strength.
• In sandy soils, ABP = 10N where N is the number of blows in a SPT test.
In saturated soils, where the water table is within foundation width B of the bottom of the
foundation, the values obtained should be halved.

It should be noted that there are three bearing pressures:

• Ultimate Bearing Pressure – the capacity at which shear failure occurs in the soil,
• Safe Bearing Pressure – the capacity at which failure is statistically avoided,
nominally equal to the safe bearing capacity divided by a factor of safety,
• Allowable Bearing Pressure – the capacity of the soil accounting for defactored
shear strength and permissible levels of settlement.

© Jonathan Haynes version 3.0 Page 206 of 239

 32900 Bridge Engineering
MSc Structural Engineering

8.6 Moment and Shear Coefficients for Continuous Slabs

Rectangular panels supported on all four edges with provision for torsion at the corners (BS8110-1).

Moment βmy βmz

Span ratio, lz / ly
In a panel of shorter span ly
Coefficients and longer span lz , with q
1.0 1.1 1.2 1.3 1.4 1.5 1.75 2.0
uniform load per unit area:
edge 0.03 0.04 0.04 0.05 0.05 0.05 0.06 0.06 0.03
1
midspan 0.02 0.03 0.03 0.04 0.04 0.04 0.04 0.05 0.02 Bending moments,

2
edge 0.04 0.04 0.05 0.05 0.06 0.06 0.06 0.07 0.04 M y = β my ⋅ q ⋅ l y2
midspan 0.03 0.03 0.04 0.04 0.04 0.04 0.05 0.05 0.03
edge 0.04 0.05 0.06 0.06 0.07 0.07 0.08 0.09 0.04 M z = β mz ⋅ q ⋅ l y2
3
midspan 0.03 0.04 0.04 0.05 0.05 0.06 0.06 0.07 0.03
Shear forces,
edge 0.05 0.06 0.06 0.07 0.07 0.08 0.09 0.09 0.05
4 Vy = β vy ⋅ q ⋅ l y
midspan 0.04 0.04 0.05 0.05 0.06 0.06 0.07 0.07 0.03
edge 0.05 0.05 0.05 0.06 0.06 0.06 0.07 0.07 0 Vz = β vz ⋅ q ⋅ l y
5
midspan 0.03 0.04 0.04 0.04 0.05 0.05 0.05 0.05 0.03
edge 0 0 0 0 0 0 0 0 0.05
6
midspan 0.03 0.05 0.06 0.07 0.07 0.08 0.09 0.1 0.03
edge 0.06 0.07 0.07 0.08 0.08 0.08 0.09 0.1 0
7
midspan 0.04 0.05 0.05 0.06 0.06 0.06 0.07 0.07 0.04
edge 0 0 0 0 0 0 0 0 0.06
8
midspan 0.04 0.05 0.06 0.07 0.08 0.08 0.1 0.11 0.04

9 edge 0 0 0 0 0 0 0 0 0
midspan 0.06 0.07 0.07 0.08 0.09 0.09 0.1 0.11 0.06

Shear βvy βvz Key Plan

Span ratio, lz / ly long edge continuous
Coefficients
1.0 1.1 1.2 1.3 1.4 1.5 1.75 2.0

1 continuous 0.33 0.36 0.39 0.41 0.43 0.45 0.48 0.5 0.33 4 2
discontinuous 0 0 0 0 0 0 0 0 0
2 continuous 0.36 0.39 0.42 0.44 0.45 0.47 0.5 0.52 0.36
3 1
discontinuous 0 0 0 0 0 0 0 0 0.24
continuous 0.36 0.4 0.44 0.47 0.49 0.51 0.55 0.59 0.36
3
discontinuous 0.24 0.27 0.29 0.31 0.32 0.34 0.36 0.38 0
continuous 0.4 0.44 0.47 0.5 0.52 0.54 0.57 0.6 0.4
4 short edge continuous
discontinuous 0.26 0.29 0.31 0.33 0.34 0.35 0.38 0.4 0.26
5 continuous 0.4 0.43 0.45 0.47 0.48 0.49 0.52 0.54 0
7 5
discontinuous 0 0 0 0 0 0 0 0 0.26
continuous 0 0 0 0 0 0 0 0 0.4
6
discontinuous 0.26 0.3 0.33 0.36 0.38 0.4 0.44 0.47 0 6
continuous 0.45 0.48 0.51 0.53 0.55 0.57 0.6 0.63 0
7
discontinuous 0.3 0.32 0.34 0.35 0.36 0.37 0.39 0.41 0.29 9 8

8 continuous 0 0 0 0 0 0 0 0 0.45
discontinuous 0.29 0.33 0.36 0.38 0.4 0.42 0.45 0.48 0.3 long edge discontinuous
continuous 0 0 0 0 0 0 0 0 0
9 short edge discontinuous
discontinuous 0.33 0.36 0.39 0.41 0.43 0.45 0.48 0.5 0.33

© Jonathan Haynes version 3.0 Page 207 of 239

 32900 Bridge Engineering
MSc Structural Engineering

8.7 Concrete Reinforcement Data

Cross sectional areas of bars
Number of bars
1 2 3 4 5 6 7 8 9 10
8 50 101 151 201 251 302 352 402 452 503
Diameter of bar

10 79 157 236 314 393 471 550 628 707 785

12 113 226 339 452 565 679 792 905 1018 1131
(mm)

16 201 402 603 804 1005 1206 1407 1608 1810 2011
20 314 628 942 1257 1571 1885 2199 2513 2827 3142
25 491 982 1473 1963 2454 2945 3436 3927 4418 4909
32 804 1608 2413 3217 4021 4825 5630 6434 7238 8042
Cross sectional areas of bars at defined spacings
Spacing of bars (mm)
75 100 125 150 175 200 225 250 275 300
8 670 503 402 335 287 251 223 201 183 168
Diameter of bar

10 1047 785 628 524 449 393 349 314 286 262
12 1508 1131 905 754 646 565 503 452 411 377
(mm)

16 2681 2011 1608 1340 1149 1005 894 804 731 670
20 4189 3142 2513 2094 1795 1571 1396 1257 1142 1047
25 6545 4909 3927 3272 2805 2454 2182 1963 1785 1636
32 10723 8042 6434 5362 4596 4021 3574 3217 2925 2681
Area ratios of links at defined spacings
Ratio Asw / S for various link diameters and spacings - VALUES FOR A SINGLE LEG
Link dia. Link spacing (mm)
(mm) 75 100 125 150 175 200 225 250 275 300
8 0.670 0.503 0.402 0.335 0.287 0.251 0.223 0.201 0.183 0.168
10 1.047 0.785 0.628 0.524 0.449 0.393 0.349 0.314 0.286 0.262
12 1.508 1.131 0.905 0.754 0.646 0.565 0.503 0.452 0.411 0.377
16 2.681 2.011 1.608 1.340 1.149 1.005 0.894 0.804 0.731 0.670
Standard pre-welded reinforcement sheets (4.8m x 2.4m fabric)
Longitudinal wires Cross wires
Fabric reference
Bar dia. Pitch Bar dia. Pitch
(main bar area)
(mm) (mm) (mm) (mm)
A393 10 10
square

A252 8 8
200 200
A193 7 7
A142 6 6
B1131 12 8
structural

B785 10 8
B503 8 100 8 200
B385 7 7
B283 6 7
C785 10 6
C636 9 6
long

100 400
C503 8 5
C385 7 5
wrapping

D98 5 200 10 200

D49 2.5 100 8 100

© Jonathan Haynes version 3.0 Page 208 of 239

 32900 Bridge Engineering
MSc Structural Engineering

8.8 Approximate Methods of Analysis

It is vitally important to be able to check the output from computer analysis programmes are
correct. The most productive method is to perform an analysis using another method, preferably
quick and intuitive. Any hand analysis method can be used. The following techniques may be
applied to specific structural forms. Always check equilibrium – do the total vertical and horizontal
reactions match the total vertical and horizontal actions ?

Continuous Frames
Assume points of contraflexure are ‘pins’. I column small compared to I deck so assume
Chop into statically determinate structures. columns are roller supports (column moment is small).
Central span is simply supported at ‘pins’. Back span cantilevers to pick up central span.

8.8.1 Portal Method for Sway Frames and Vierendeel Trusses

Ma
F h1
F Ma = ×
F 2 2

Mc
F Mb F/2 F/2
Ma

2F h
Mb = × 2
F 2F 2 2
Mb
Mc

2F/2 2F/2

beam moment balances

column moments, -Mc=Ma+Mb

Sway Moments in Multi-storey Frames Good for frames up to 10 storeys

Assume points of contraflexure are ‘pins’ at mid-height / length. but base moment drifts off.
Divide multi-bays into single bays and split lateral load in proportion Combine gravity analysis
to span. For pin based frames assume contraflexure at two thirds assuming beams are fixed.
height of first storey.

© Jonathan Haynes version 3.0 Page 209 of 239

 32900 Bridge Engineering
MSc Structural Engineering

8.8.2 Centroidal Distance Method for Sway Frames

Fa
d 1V2 d 2V2
V1 = V3 =
d2 d3
4 5
Fb
moments about reaction 3,
I1 I2 I3
moments to left of hinge 4,

moments to left of hinge 5,

d1 d2
d3
V1 V2 V3

The centroidal distance method distributes 1. Obtain the vertical reactions as a function of V
vertical reactions to multiple bays in proportion to and d
their distance from the centroid of the columns. 2.Take moments about one reaction
3. Assuming pins at beam contraflexure points,
V1 V2 V3 examine equilibrium in frame parts to obtain
= =
d d d horizontal reactions
1 2 3

8.8.3 Cantilever Method for Sway Frames

F F

F F

F F

M
M
M = F × h1 + F × h2 + F × h3 ML =
ns nc
Cantilever Base Moment ns is number of storeys above
Divide multi-bays into single bays and split lateral nc is number of columns
load in proportion to span.
Calculate base moment as a single cantilever.

© Jonathan Haynes version 3.0 Page 210 of 239

 32900 Bridge Engineering
MSc Structural Engineering

8.8.4 Truss Analogy for Plate Girders

Plate Girders Top flange force ≡ Top chord force

An approximate assessment of forces in the Web stiffener force ≡ Internal strut force
flanges, web stiffeners and web of a plate End post force ≡ End strut force
girder may be obtained using a truss analogy. Web tension field force ≡ End tie force

8.8.5 Simplified Truss Analysis and Beam Analogy for Deflection

M
Vb − Fc =
F= h
h
b
h
V

V M wL2 a
Ft = M=
h 8

top chord

z
h
Neutral Axis
2
(
I NA = I yy ,top + Atop z + I zz,btm + Abtm h − z )
2

bottom chord

Pin-jointed Frameworks Rework the node point loads into an equivalent

Chord forces are greatest at midspan. Internal uniform loading to calculate M.
strut and tie forces are greatest at the reactions. Approximate deflection using standard beam
M is the maximum free bending moment. bending formulae and compound stiffness of chords.

© Jonathan Haynes version 3.0 Page 211 of 239

 32900 Bridge Engineering
MSc Structural Engineering

8.9 Plate Girder Section Properties Z

h Y Y

Thickn Thickn Web

SYMMETRICAL PLATE Depth of Width of Web Web Second Moment of Radius of Shear Moment Area of
ess of ess of Panel Plastic Modulus
GIRDER S355 Section Section Depth Aspect Area Gyration Capacity Capacity Section
Web Flange Width

Serial size mass h b t t h a I I i i W W V M A

w f w a/h yy zz y z pl, y pl, z b,Rd pl,y
w
kg/m mm mm mm mm mm mm cm4 cm4 cm cm cm3 cm3 kN kNm cm2

1500 x 300 x 25 232 1500 300 10 25 1450 3000 2.069 1069990 11262 60.2 6.2 11425 774 2158 3942 295

1500 x 300 x 40 300 1500 300 10 40 1420 3000 2.113 1517887 18012 63.0 6.9 18088 1224 2130 6059 382

1500 x 300 x 50 367 1500 300 12 50 1400 3000 2.143 1851900 22520 62.9 6.9 22590 1534 2954 7568 468

1500 x 300 x 75 512 1500 300 15 75 1350 3000 2.222 2594109 33788 63.1 7.2 33581 2301 3454 10914 653

1500 x 500 x 25 310 1500 500 10 25 1450 3000 2.069 1613948 52095 63.9 11.5 18800 2108 2158 6486 395

1500 x 500 x 40 425 1500 500 10 40 1420 3000 2.113 2370741 83345 66.1 12.4 29768 3357 2130 9972 542

1500 x 500 x 50 524 1500 500 12 50 1400 3000 2.143 2903567 104187 65.9 12.5 37090 4200 2954 12425 668

1500 x 500 x 75 748 1500 500 15 75 1350 3000 2.222 4118484 156288 65.8 12.8 54956 6301 3454 17861 953

2000 x 500 x 50 542 2000 500 10 50 1900 4000 2.105 5325750 104183 87.9 12.3 49700 4198 2282 16650 690

2000 x 500 x 60 648 2000 500 12 60 1880 4000 2.128 6311667 125027 87.4 12.3 59554 5045 3083 19950 826

2000 x 500 x 75 807 2000 500 15 75 1850 4000 2.162 7743016 156302 86.8 12.3 74269 6319 4733 24137 1028

2000 x 750 x 50 738 2000 750 10 50 1900 4000 2.105 7702833 351578 90.5 19.3 74075 9407 2282 24815 940

2000 x 750 x 60 884 2000 750 12 60 1880 4000 2.128 9135267 421902 90.1 19.4 88654 11295 3083 29699 1126

2000 x 750 x 75 1101 2000 750 15 75 1850 4000 2.162 11218797 527396 89.4 19.4 110363 14132 4733 35868 1403

2500 x 750 x 50 871 2500 750 15 50 2400 4000 1.667 12984250 351630 108.2 17.8 93675 9465 4332 31381 1110

2500 x 750 x 60 987 2500 750 15 60 2380 4000 1.681 15083459 421942 109.5 18.3 111942 11339 4335 37501 1257

2500 x 750 x 75 1160 2500 750 15 75 2350 4000 1.702 18166766 527410 110.9 18.9 139050 14151 4210 45191 1478

3000 x 750 x 50 930 3000 750 15 50 2900 5000 1.724 19367375 351644 127.8 17.2 112800 9484 4684 37788 1185

3000 x 750 x 60 1046 3000 750 15 60 2880 5000 1.736 22436784 421956 129.8 17.8 134892 11358 4689 45189 1332

3000 x 750 x 75 1219 3000 750 15 75 2850 5000 1.754 26961609 527424 131.8 18.4 167738 14169 4557 54515 1553

© Jonathan Haynes version 3.0 Page 212 of 239