Self-Anchored

Suspension Bridges

Part II:
Structural behaviour
&
Main span possibilities

David van Goolen

April 2007
 Self-anchored suspension bridges: Part II

Self-Anchored Suspension Bridges

Project: Master of Science Final Thesis

Start date: September 2006
Document: Part II: Main Study

Name: David van Goolen

Student nr.: 1040596

ii
 Self-anchored suspension bridges: Part II

Final Thesis Project

Topic: Self-Anchored Suspension bridges
Document: Part II: Main study

Name: David van Goolen

Student number: 1040596
University: Delft University of Technology
Faculty of Civil Engineering

Examining board: Prof. Ir. F.S.K. Bijlaard Delft University of Technology

Dr. A. Romeijn Delft University of Technology
Dr.ir. C. van der Veen Delft University of Technology
Ir. L.J.M. Houben Delft University of Technology
Ir. W.P.J. Langedijk Engineering office Iv-Infra B.V.

iii
 Self-anchored suspension bridges: Part II

Preface
The first phase of this final thesis project has been presented earlier by means of
part I: literature survey captured in a report:
‘Structural behaviour of self-anchored suspension bridges: Literature Survey
This partial document part II presents the second phase of this final thesis; the main study. In
the main study the structural behaviour is researched. First part is a parameter study. Second
part is increasing the span length and trying to find a limit. The third part is erection aspects of
a self-anchored suspension bridge at Nijmegen.
I would like to express my gratitude to Engineering office Iv-Infra, they offered me the
possibility to execute this study at their office and enabled me to fill in this study completely
according to my personal interpretations. Iv-Infra gave me the opportunity to make use of
their facilities and experience in bridge engineering. Especially I would like to thank my daily
supervisor at Iv-Infra, Mr. Walter Langedijk for providing me information, help and guidance
throughout the entire final thesis project. His experience helped me a lot with generating ideas
and tackling this thesis project.
Further more I would like to thank the other examining board members for offering guidance,
help and discussing the progress of my final thesis:
Prof. Ir. F.S.K. Bijlaard Delft University of Technology
Dr. A. Romeijn Delft University of Technology
Dr.ir. C. van der Veen Delft University of Technology
Ir. L.J.M. Houben Delft University of Technology

David van Goolen

Delft University of Technology
Faculty of Civil Engineering
Delft 2007
The Netherlands

iv
 Self-anchored suspension bridges: Part II

Summary
Since 1870, only about 25 highway bridges have been executed as a self-anchored suspension
bridge. The rise of the cable stayed bridge since 1955 made this suspension type an obsolete
alternative for a long period of time. The largest existing main span for a self-anchored
suspension bridge is 300 metres and dates from 1999. Main difficulties for this bridge type to
reach spans over 300 metres can be blamed on erection problems and the buckling stability of
the girder. Erecting the deck structure prior to the main cable makes this bridge technically
and economically less attractive than for instance the cable stayed bridge.
A dimensional inventory has shown that the deck slenderness is limited to about λ = 1/95 and
the sag ratio varies between 1/5-1/8. The deck slenderness is related to the required bending
stiffness to have sufficient resistance against buckling. Also the relatively high sag ratios,
compared to conventional suspension bridges, are mainly chosen to reduce the normal force
in the deck that is imposed by the main cable.
A parameter study into the structural behaviour has revealed that the most important bridge
parameters are the bending stiffness EIdeck of the deck and the axial stiffness EAmain cable of the
main cable. A well chosen ratio between the EIdeck and EAmain cable influences the maximum
bending moments and the deflections in the girder. In the pre-design process of a suspension
bridge type it is favourable to consider:
-A slender stiffening girder, to reduce the maximum bending moment in the girder
-A stiff main cable, to increase the global stiffness of the bridge and to reduce the
maximum bending moment in the girder
-A high sag to span ratio, to reduce the normal force in the deck and the maximum
bending moment in the deck.
A study to the static strength, stiffness, frequency behaviour and the buckling stability of the
box girder, revealed that a deck slenderness of the box girder of λ = 1/100 and even more
slender is very well feasible.
Exploring the main span possibilities of this bridge type, this study has shown that a span
length of 500 metres is very well possible and even beyond that. The on before hand
expected limitation on the global buckling stability of the girder has turned out to be feasible.
With an increasing main span the buckling phenomena does become more critical but still of
acceptable level.
A difference is visible in buckling of the main span and the side span. The upward buckling
mode of the side span is decisive over the downward buckling mode of the main span. But at
least up to 500 metres a deck slenderness of λ = 1/100 and beyond that is very well possible
regarding all important design criteria.
The most limiting factor for the self-anchored suspension bridge, to reach a large main span
and apply a very slender deck, is the erection stage. The number of temporary supports in the
main span determines a decisive stress condition for erecting the deck. Erecting with
temporary stays is an option but remains a laborious method.
It is almost inevitable that for the erection stage some significant provisions have to be made
in the cross section of the deck regarding the shear and bending conditions or else a much less
slender deck should be applied.

So it has been shown that it is structural feasible to reach more competitive main span lengths
up to at least 500 metres but that the erection stage can determine decisive conditions for
designing the deck.

v
 Self-anchored suspension bridges: Part II

Introduction
The total research of this M.Sc. study is presented in two parts to fulfil the requirements of the
degree of Master of Science obtained at Delft University, faculty of Civil Engineering.
This document presents part II; the main study of my final thesis project. The objective of
part II is to research the structural behaviour of a self-anchored suspension bridge. A
reference model is designed and used to investigate the influence of the main bridge
parameters on the structural behaviour of a self-anchored suspension bridge.
The first chapter presents the basic assumptions that are made to design a reference model for
the self-anchored suspension bridge. This reference design is modelled in a FEM-program to
investigate and calculate the force distribution and deformations.
Chapter 2 gives the verification of the reference design on the main design criteria like static
strength, stiffness and the buckling stability of the girder.
First results are given of a parameter study in chapter 3. The results of the parameter study are
used to determine an optimization of the reference design.
Chapter 4 explores and explains the effects of an increasing span length of a self-anchored
suspension bridge all important design aspects.

vi
 Self-anchored suspension bridges: Part II

Table of contents
PREFACE............................................................................................................................................................ IV

SUMMARY ...........................................................................................................................................................V

INTRODUCTION............................................................................................................................................... VI

TABLE OF CONTENTS.................................................................................................................................. VII

1 REFERENCE DESIGN .............................................................................................................................. 1

1.1 BASIC ASSUMPTIONS ............................................................................................................................. 1
1.1.1 City bridge of Nijmegen .................................................................................................................. 1
1.1.2 Stiffening girder .............................................................................................................................. 2
1.1.3 Loading scheme............................................................................................................................... 7
1.1.4 Dimensions...................................................................................................................................... 8
1.1.5 Cable configuration ........................................................................................................................ 8
1.1.6 Pylon frame ..................................................................................................................................... 8
1.1.7 Self weight ...................................................................................................................................... 9
1.2 THREE DIMENSIONAL MODELLING ...................................................................................................... 10
1.3 DIMENSIONING ................................................................................................................................... 14
1.3.1 Design stress ................................................................................................................................. 14
1.3.2 Estimation by hand calculation..................................................................................................... 16
1.3.3 Final dimensions reference model ................................................................................................ 21
2 VERIFICATION OF REFERENCE DESIGN ....................................................................................... 25
2.1 METHOD FOR VERIFICATION ON STATIC STRENGTH, STIFFNESS AND STABILITY .................................. 25
2.1.1 Static strength (ULS)..................................................................................................................... 25
2.1.2 Stiffness (SLS)................................................................................................................................ 27
2.1.3 Stability (ULS)............................................................................................................................... 28
2.1.4 Frequency analyses ....................................................................................................................... 31
2.2 VERIFICATION OF REFERENCE DESIGN ................................................................................................ 34
2.2.1 Verification reference design on strength ..................................................................................... 34
2.2.2 Verification reference design on stiffness...................................................................................... 37
2.2.3 Verification reference design on stability girder........................................................................... 38
2.2.4 Frequency behaviour..................................................................................................................... 42
3 PARAMETER STUDY INTO THE STRUCTURAL BEHAVIOUR................................................... 43
3.1 GIRDER INFLUENCE............................................................................................................................. 44
3.1.1 Stress level in box girder............................................................................................................... 45
3.1.2 Stability of the stiffening girder..................................................................................................... 46
3.1.3 Number of hinges .......................................................................................................................... 49
3.2 MAIN CABLE INFLUENCE .................................................................................................................... 50
3.2.1 Cable type...................................................................................................................................... 50
3.2.2 Cable cross sectional area ............................................................................................................ 51
3.3 SAG INFLUENCE .................................................................................................................................. 52
3.4 PYLON INFLUENCE .............................................................................................................................. 53
3.5 ASYMMETRIC TRAFFIC LOADING......................................................................................................... 54
3.5.1 Full loaded deck ............................................................................................................................ 54
3.5.2 Half loaded deck ........................................................................................................................... 54
3.6 CONCLUSION RESULTS OF THE PARAMETER STUDY............................................................................. 57
3.7 EVALUATION OF THE RESULTS ............................................................................................................ 57
3.7.1 Criteria.......................................................................................................................................... 57
3.7.2 Conclusions for optimized reference design.................................................................................. 58
4 INCREASING SPAN LENGTH .............................................................................................................. 61
4.1 BASIC ASSUMPTIONS ........................................................................................................................... 61
4.2 EFFECTS OF INCREASING THE SPAN BY SCALING ................................................................................. 63

vii
 Self-anchored suspension bridges: Part II

4.2.1 Developments on static strength.................................................................................................... 64

4.2.2 Developments on buckling stability of the stiffening girder .......................................................... 67
4.2.3 Frequency behaviour..................................................................................................................... 69
4.2.4 Developments on reaction forces .................................................................................................. 70
4.2.5 Material use .................................................................................................................................. 71
4.2.6 Effects on the erection ................................................................................................................... 73
4.3 EVALUATION AND CONCLUSION FOR AN INCREASING SPAN LENGTH................................................... 75
5 ERECTION ASPECTS OF THE NEW CITY BRIDGE AT NIJMEGEN .......................................... 78
5.1 CRITERIA NIJMEGEN CITY .................................................................................................................. 78
5.2 EXPLORATION OF ERECTION METHODS ............................................................................................... 79
5.2.1 Existing methods ........................................................................................................................... 79
5.2.2 Alternative methods....................................................................................................................... 80
5.3 CONSTRUCTING WITH TEMPORARY PIERS IN THE WAAL RIVER ........................................................... 82
5.3.1 Simply supported between temporary supports............................................................................. 82
5.3.2 Erecting as a continuous beam ..................................................................................................... 83
5.4 CONSTRUCTING WITHOUT TEMPORARY PIERS IN THE WAAL RIVER .................................................... 87
5.5 CONCLUSIONS REGARDING ERECTION ................................................................................................. 88
6 OVERALL CONCLUSIONS AND RECOMMENDATIONS .............................................................. 90
6.1 CONCLUSIONS..................................................................................................................................... 90
6.2 RECOMMENDATIONS........................................................................................................................... 91
LIST OF LITERATURE.................................................................................................................................... 93

LIST OF FIGURES AND TABLES .................................................................................................................. 94

APPENDICES ..................................................................................................................................................... 97
APPENDIX 1 CABLE MODELLING ....................................................................................................................... 98
APPENDIX 2A EFFECTIVE WIDTH AT MAIN AND SIDE SPAN ................................................................................ 99
APPENDIX 2B EFFECTIVE WIDTH AT SUPPORT ................................................................................................. 101
APPENDIX 3A GIRDER INFLUENCE ................................................................................................................... 103
APPENDIX 3B CABLE INFLUENCE .................................................................................................................... 104
APPENDIX 3C SAG OVER SPAN LENGTH INFLUENCE ......................................................................................... 105
APPENDIX 3D PYLON INFLUENCE .................................................................................................................... 105
APPENDIX 3E NUMBER OF HINGES IN GIRDER .................................................................................................. 105
APPENDIX 4 COMPARABLE BOX GIRDERS ........................................................................................................ 106
APPENDIX 5A STABILITY CHECK REFERENCE MODEL ...................................................................................... 107
APPENDIX 5B STABILITY CHECK OPTIMIZED MODEL ....................................................................................... 108
APPENDIX 6 DESIGN ORTHOTROPIC STEEL BOX OF THE NEW CARQUINEZ BRIDGE.......................................... 109
APPENDIX 7 STRESS CALCULATION IN REFERENCE MODEL.............................................................................. 110
APPENDIX 8A ERECTION OF KONOHANA BRIDGE............................................................................................ 112
APPENDIX 8B ERECTION OF YEOUNGJONG GRAND BRIDGE ............................................................................ 113
APPENDIX 8C ERECTION OF NESCIO BRIDGE ................................................................................................... 115

viii
 Self-anchored suspension bridges: Part II

1 Reference design
To be able to research a bridge behaviour, by change of the main bridge parameters, a
reference model is made first. For this reference model the main characteristics of a self-
anchored suspension bridge are captured. This part will explain the basic assumption for this
model, determination of the main bridge structural components like stiffening girder, main
cable, hangers and pylon. Also a description is given of the method of FE-modelling of the
bridge structure..

1.1 Basic assumptions

This part gives the basic assumptions that are made for the pre-design process in this research.

1.1.1 City bridge of Nijmegen

In November 2006, the city council of Nijmegen asked for a feasibility study for a suspension
bridge to cross the river Waal. This feasibility study should at least include cost estimation,
risk analyses, esthetical aspects, technical and economical considerations and is contracted out
to the engineering office Iv-Infra. This report intends to cover a part of this total study,
namely technical conditions of a self-anchored suspension bridge but with a general approach.
Figure 1 gives an overview of the new trace of the new city bridge in Nijmegen city.

Figure 1 Trace city bridge Nijmegen

The total width of the river and river banks is succeeding 1000 metres. The width of the river
is an average of 325 metres, see Figure 2.

Figure 2 Cross section Waal river near Nijmegen

As explained before this report focuses on the structural behaviour and exploring the span
possibilities of a self-anchored suspension bridge. The most important design demands for
this city bridge are incorporated for the research under consideration. In that way the results

1
 Self-anchored suspension bridges: Part II

of this MSc study can be used and interpreted for the city bridge of Nijmegen and also for
self-anchored suspension bridges in general.
Most important design criteria are:
• Vertical navigation clearance underneath the bridge deck is 9.10 metres
• Horizontal navigation clearance is 265 metres
• Traffic on the bridge consists of:
ƒ Local highway traffic with a design speed of 70
km/h
ƒ public transport like busses
ƒ bicycles
ƒ pedestrians
• Highway traffic lanes 2*2. And future expansion to 2*3 lanes
• Bicycles and pedestrian lanes in both ways and physically separated from the highway
traffic.
• Design life of the bridge is 100 years.

The number of lanes and type of traffic is determining the loading scheme and the severity of
the loads. It also determines the required deck width. This will be explained later on.
The horizontally required clearance determines the smallest free span length. This design
aspect will be used in the exploration of the span possibilities, and will be dealt with further
on in this report.
Figure 3 shows a simplification of the situation1 at the Waal in Nijmegen. From that given
situation a level of the bridge deck is chosen of 15 meters above ground level.

Figure 3 Situation at Waal River

1.1.2 Stiffening girder

Type of girder
For the deck, a steel box girder is used. The box will have a certain width w, height h and
thicknesses of web tw and flanges tf, see Figure 4

Figure 4 Box girder with equivalent flange thicknesses to account for longitudinal stiffeners

1
Design criteria of City bridge Nijmegen, June 2006.

2
 Self-anchored suspension bridges: Part II

In reality a box girder contains more structural details like:

• Longitudinal stiffeners:
Are needed on the flanges and on the webs. They are required to improve the stability
of the relatively thin steel plates to prevent local buckling of the plate elements.
Besides that the longitudinal stiffeners are required to transfer and resist the
concentrated wheel loading on the deck.
• Transverse stiffeners:
Are located at intermediate cross frames to improve the cross section against torsion
and distortion and to support the longitudinal stiffeners of the deck.
• Cross beams:
The cross beams support the longitudinal stiffeners and reduce the span length of these
stiffeners. The combination of the deck and longitudinal stiffeners act as a beam under
bending between two cross beams.
• Cross bracings/frames/diaphragms:
Prevent the cross section to distort and are designed to resist wind loads, to brace the
compression flanges and distribute vertical self weight and live loads.

Only the first detail, longitudinal stiffeners, will be taken into account in the modelling of the
box girder because only these elements contribute to the bending stiffness of the cross
section.

System for stiffening girder

As for the largest built self-anchored suspension bridges like the Konohana and Yeong Grand
bridge, see literature survey §2.6, a continuous stiffening girder is chosen for the reference
model in this research. No hinges at the pylon or at mid span are therefore applied. A system
without hinges is in general more stiff than a system with one or more hinges. To achieve the
largest span possible, a system should be chosen with the best stiffness behaviour: a
continuous stiffening girder.

Material
For the structural elements like the deck and pylon, a steel grade S355 is used in this research
which has the following properties:
Yield strength fy = 355 N/mm2
Modulus of elasticity E = 210000 N/mm2

Plate thicknesses
The longitudinal stiffeners on the deck plate and bottom flange of the box girder contribute to
the bending stiffness of the cross section, but not seriously to the torsional stiffness and shear
resistance of the cross section. The longitudinal stiffeners of the top flange, bottom flange and
the web are taken into consideration by assuming them as part of the area of the flange and
web. So an equivalent plate thickness will be used including the area of those stiffeners. The
following designation is used is this research:
t plate : for plate thickness which does not include required plate area for
longitudinal stiffeners.
t eq plate : for equivalent plate thicknesses including the area for longitudinal
stiffeners.
For instance for the calculation of shear stresses and effective width of the cross section, the
net plate thickness t plate is used.

3
 Self-anchored suspension bridges: Part II

For the minimum required plate thicknesses the NEN-EN 1993-2:2003 part c1.2.2 states:
Deck plate thickness in the carriageway t ≥ 16 for an asphalt layer > 40 mm.
Thickness of stiffeners: t stiff ≥ 6 mm.

Furthermore it is stated in the NEN-EN 1993-2:2003 that the minimal required plate
thicknesses for deck plates and stiffeners are:

So with an assumed asphalt layer of 50 mm the following minimum required plate thicknesses
are used:
t flange top ≥ 16 mm
t flange bottom ≥ 10 mm
t web, crossbeam ≥ 10 mm
t stiffeners ≥ 6 mm

To account for the required longitudinal stiffeners, the area of the stiffeners Astiff is estimated2
by a percentage of the area Aflange of the flanges and web Aweb:
A stiff. top flange = 65% * A top flange
A stiff . bottom flange = 35% * A bottom flange
A stiff. web = 15 % * A web
These ratio are determined from an existing box girder applied in a recent suspension bridge,
see Appendix 6 Design orthotropic steel box of the New Carquinez Bridge.
It clearly shows that the top compression flange needs much more stiffeners for it has to
account for local buckling effects and to be able to resist the local wheel loading of the traffic.
For the bottom flange less it is visible that less stiffeners are applied because no local wheel
loading can occur on this flange. These findings are correspond to the expectation that the
deck plate requires more stiffeners than the bottom flange and the estimated percentages will
be used from now on to determine the area for longitudinal stiffeners.
The area of the longitudinal stiffeners of web and flanges are taken into account as an
additional thickness to the web and flanges, called t equivalent. So:

t eq. flange top = 1.65 * t flange top = 1.65 * 16 = 26.4 mm

t eq. flange bottom = 1.35 * t flange bottom = 1.35 * 10 = 13.5 mm
t eq. web = 1.15 * t web = 1.15 * 10 = 11.5 mm

For ease of calculation a simple ratio between thickness of top- and bottom flange is chosen:
t flange top = 2 * t flange bottom

2
This estimation is retreived from a similar box girder of the same dimensions: Thimmardy.E. et.al., New
Carquinez bridge. North America’s Newest suspension bridge. Steel bridge 2004 Millau.

4
 Self-anchored suspension bridges: Part II

For a conservative approach, the following equivalent plate thicknesses are chosen:
Chosen equivalent thicknesses are :
t eq. flange top =40 mm
t eq. flange bottom =20 mm
t eq. web =15 mm

Cross section classification

The Eurocode NEN-EN-1993-1-1 distincts several classifications for steel cross sections. A
box girder, with dimensions given in this research, is likely to be classified as class 3 or 4 on
which the Eurocode states a definition:
-Class 3 cross-sections are those in which the stress in the extreme compression fibre of the
steel member assuming an elastic distribution of stresses can reach the yield strength, but
local buckling is liable to prevent development of the plastic moment resistance.
-Class 4 cross-sections are those in which local buckling will occur before the attainment of
yield stress in one or more parts of the cross-section.

Classification 4 is not applicable because the flanges and webs of box girder are locally
stabilized by longitudinal stiffeners. The box girder is assumed to have a cross section
classification 3.
So assumed is that the normal force in box girder is resisted by the total area of the cross
section, no effective cross section is taken into account for the normal force.

Bearing system of girder

The bearing system of the stiffening girder should ensure vertical, horizontal (transverse and
longitudinal) fixation. For simplification of the bearing system no temperature influence is
considered because it is assumed that the extra provisions needed for temperature influence
does not influence the considered aspects of the global structural behaviour of the bridge in
this research scope. Therefore a simple symmetrical bearing system is chosen. The total
bearing system consists of several slide- and fixed bearings.
In transverse direction all bearings are horizontally fixed. In longitudinal direction only one
pair is horizontally fixed, the rest allows for horizontal movement. The deck is horizontally
fixed at one of the pylons, as is generally3 done for suspension bridges.
In vertical direction all bearings are fixed. Next figure shows a scheme of the bearing system.

Figure 5 Bearing system

3
Gimsing, N.J., Cable supported Bridges, Wiley&Sons, 1998

5
 Self-anchored suspension bridges: Part II

Deck width
The deck width is depending on the required number of traffic lanes for the city bridge of
Nijmegen. This bridge has to accommodate:
• 2*2 traffic lanes. And a future expansion for 2*3 traffic lanes
• Pedestrian and bicycle lanes
Vo = 120 Vo 90
The Dutch ROA (guidelines design highways) km/h km/h
states the following requirements for clearances a. Lane 3,5 3,25
and width of lanes: b. dividing line 0,15 0,15
c. side line 0,2 0,2
d. safety strip 0,6 0,3
e. hard shoulder 3,25 3,25
l. side strip 0,5 0,5
m. object distance 1,5 1
Figure 6 Width of lanes [m]

Applying this, results in a total deck width of 35.6 metres.

Figure 7 Total deck width

For the deck system, a square box girder is chosen and the width is chosen according the
required deck width as shown in Figure 7. In this case a width of 35.6 metres is required.

Figure 8 Box girder

In reality a box girder would have a shape similar to that of in Figure 9. In here the webs are
placed inclined for aero dynamical reasons. Assuming a square box girder, like in Figure 8, is
a valid procedure regarding the mechanical properties like axial, bending and torsional
stiffness.

Figure 9 Box girder shape

6
 Self-anchored suspension bridges: Part II

1.1.3 Loading scheme

For analysing the structural behaviour and exploring the span possibilities only the main
loadings are taken into account. These are vertical loadings such as:
• Permanent loading
o Self weight (according to EN-1991-1-1:2002)
γ steel = 78.5 kN/m3
o Asphalt layer of 40 mm
γ hot rolled asphalt = 23 kN/m3
Assumed application of hot rolled asphalt, the permanent load caused by asphalt
load is: 0.040 * 23 = 0.92 kN/m2.
• Variable loading
o Traffic loading uniform distributed (highway traffic, pedestrian, bicycles)
o Traffic loading concentrated axle loads

According to NEN-EN 1991-3 load

model 1 is used to define the vertical
loads of traffic. The next table defines
the concentrated loads and the uniform
distributed loads.

Table 1 Concentrated and uniform distributed loads

The given values for Qik and qrk include a dynamic amplification. For pedestrian and cycle
actions the density of the uniformly distributed load is: qfk = 5 kN/m2
This value for pedestrian load is conservative because NEN-EN-1991-3 part 5.3.2 states that
a reduced value can be applied for bridges with individuals exceeding 10 metres.

Furthermore NEN-EN 1991-3 part 4.3.2 states that:

No more than one tandem system should be used considered per lane; only complete tandem
system shall be considered.
In Figure 10 the total loading scheme is illustrated for the complete width of the girder. The
load model shown in Figure 10 represents the situation to be used for the design of the bridge
components like main cable, hangers, girder and pylon. This because the largest reaction
forces exist for the cable plane.

Figure 10 Traffic loading scheme in transverse direction of the deck

7
 Self-anchored suspension bridges: Part II

1.1.4 Dimensions
A span length of 150 metres is chosen as a starting point for the so called reference model.
With this chosen main span of 150 metres it is possible to investigate later on the influence of
an increasing span length to about 500 metres on the mechanical properties of the stiffening
girder. Based on the literature survey, see the appendix of the literature survey, on self-
anchored suspension bridges, a few dimensional ratio’s are kept fixed:
• Main span to side span = 2.4
• Main span to hanger distance = 24
Other ratio’s like girder slenderness, and sag to span ratio are part of the research and which
will be varied.
The required vertical clearance of 9.10 metres remains fixed. So the pylon height under the
deck remains fixed on 15 metres.

1.1.5 Cable configuration

For the configuration of the main cable, a
parabolic shape is chosen. In reality a cable has
a catenary shape when it is loaded by its self
weight. The catenary shape is well
approximated with a parabolic line. Therefore a
parabolic shape is assumed for the main cable
in main- and side span.
Figure 11 Parabolic shape cable
2
4 l2
The sag of the cable in the mid of the side span f2 is determined by : f 2 = f 1 * 2
l1
The literature study has shown that that sag of the main span varies between 1/5 to 1/8 of the
main span.
Furthermore a vertical hanger configuration is chosen. Until now, only one self-anchored
suspension bridge has been built with inclined hangers in the longitudinal direction of the
bridge. So a vertical configuration is considered more as standard design.
Longitudinal inclined hangers have some advantages and disadvantages. The global stiffness
of the bridge is increased but more problem are expected with respect to erection of the bridge
and fatigue in the hangers.

1.1.6 Pylon frame

The pylon frame can have different appearances. Similar
to the cable stayed- and conventional suspension
bridges, H- and A- frames can be used in self-anchored
suspension bridges. In this research not much attention
goes out to the design of a pylon frame, so a standard H-
frame is chosen with a steel box cross section. In that
way the relevant mechanical properties as bending- and
axial stiffness can be assigned easily.

Figure 12 H- frame pylon Figure 13 Cross section pylon

4
Ulstrup, C., Rating and preliminary analysis of suspension bridges. Journal of structural engineering, Vol. 119,
No.9. September 1993.

8
 Self-anchored suspension bridges: Part II

1.1.7 Self weight

A suspension bridge is modelled according to its final desired geometry under self weight. In
many cases the main cable is given a pretension so that under dead load the bridge adopt it
final desired shape.
So the ideal FE model5 of a suspension bridge should represent this situation that on
application of the self weight load, the geometry of the bridge does not deviate from the
desired final shape of the bridge.
Furthermore, a general assumption in suspension bridge design is to have a reduced global
bending moment to about zero under self weight loading. This means that one wants to
achieve that the self weight load is completely supported by the main cable. This can be
approximately achieved by manipulating the initial tensile force in de the main cable. The
initial tensile force in the main cable can be found by trial and error until a situation is created
with minimum deck deflection and minimum bending stresses caused by the global bending
moment in the stiffening girder.
In the FEM program this initial tensile force on the main cable is done by applying a
temperature load that causes the cable to become shorter which is just a modelling tool to
apply a pretension on a structural member. Figure 14 shows the deflection due to self weight
only (deflection in mm). When the main cable is given a certain amount of pretensioning
(determined iteratively), the deflection of the girder is reduced to nearly zero, see Figure 15.

δmax = 641 mm
Figure 14 Deflection of girder [mm] due to self weight only

δmax = 1 mm

Figure 15 Deflection of girder [mm] due to self weight after pretensioning the main cable

Also the bending moments reduce to nearly zero. Figure 16 shows the bending moment
distribution due to self weight only and Figure 17 shows the bending moments after
pretensioning of the main cable. Figure 17 clearly shows the reduced global bending moment
to zero and the resulting small local bending moment between the hangers.

5
Ren, W. Roebling suspension bridge. 1:Finite element model and free vibration response, Journal of bridge
engineering, March/April 2004, pp 110-118.

9
 Self-anchored suspension bridges: Part II

Because the global bending moment is reduced to nearly zero, the assumption that the
bending stresses in the girder are almost reduced to zero under self weight loading, is hereby
verified.

Mmax = 148370 kNm

Figure 16 Global bending moments [kNm] in girder due to self weight

Mmax = 6653 kNm

Figure 17 Bending moments [kNm] in girder due to self weight after pretensioning main cable

1.2 Three dimensional modelling

For modelling the bridge, the FEM program SCIA ESA-PT 6.0.185 version is used. Such a
program enables the designer to model a structure and to apply certain loads and loading
combinations from which the effects like member forces and deflections can be calculated.
There are several ways to model a structure, depending on the type of structure and design
phase. Structures and parts of structures can be modelled with beam-, plate-, solid and cable
elements, etcetera. With beam elements the model is built up with one dimensional line
elements. Plate elements are two dimensional and solid elements are three dimensional
elements.
With respect to the scope of this research, the bridge will be modelled with beam elements.
This enables to model the total bridge structure and calculate member forces due to certain
load cases and combinations.
The scope of the model is to be able to analyse the model statically and dynamically in a three
dimensional way and to be able to analyze the effects of symmetric and asymmetric loading
statically. Also an assessment will be made with respect to the geometric non linear effects of
a cable supported bridge, the so-called second order effects.
Figure 18 shows the FE-model of the bridge which will be used for this study.

10
 Self-anchored suspension bridges: Part II

Figure 18 Model with beam elements

Pylon
The pylon is modelled as a simple portal frame. The cross section consists of a steel box. At
the basement the pylon is rotational fixed to provide for longitudinal stiffness to structure. The
pylon is loaded by an axial compression force
and a bending moment due to horizontal force on
the pylon caused by the tensile force in the main
cables.
The pylon is fixed supported in transverse and
longitudinal direction of the bridge. For ease of
modelling, the deck, represented by the stiffening
girder, is vertically supported on the ‘outside
world’ at the pylon. One rotation fixed support
creates the effect of two supports on the
stiffening girder. More detailed information is
given further on in this paragraph which explains
the modelling of the stiffening girder.
Modelling the pylon and bridge deck this way,
cancels out any influence of the girder support on
the pylon. This is assumed to be negligible.
Figure 19 Pylon frame
Main cable
Two different ways of modelling the main cable and hangers have been explored: cable
elements, and hinged truss elements (like a chain where every link is hinged connected). See
appendix 1 for the differences between these types of models for the main cable structure.
Based on the findings that no significant differences on mechanical behaviour for the two
alternatives where visible, one method to model the main cable is used from now on;
modelling with cable elements.
The main cable is modelled with cable elements. These are beam elements with a very low
bending stiffness. Also no shear forces exist for the cable. The cable element is subjected to
its own weight and accounts for the slackening effects in cables under self weight load.
An example is given to illustrate the effect of a cable element. Figure 20 shows on the top a
beam element with a certain span, on the bottom a cable element with the same span and
mechanical properties and self weight loading. The cable element displays a larger
deformation.

11
 Self-anchored suspension bridges: Part II

Beam element

Cable element

Figure 20 Deformation beam- and cable element

Another effect of a cable element which can be distinguished, is that the slack causes a
tension in the cable and therefore a horizontal reaction on the supports.

For modelling the cable, an equivalent modulus of elasticity has to be used to account for
elastic stretch and lengthening of the cable due to geometry change. These two effects reduce
the modulus of elasticity. The equivalent modulus of elasticity can be determined using the
formulae developed by H.J. Ernst. Euro code EN-19931-11 states for the effective modulus of
elasticity:
E is the modulus of elasticity of the cable
E w is the unit weight
Et = in which
w2l 2 E l is the horizontal span of the cable
1+ σ is the stress in the cable due to self weight and
12σ 3 permanent loading

Due to the relative small center to center distance of the hangers, the effect of elastic stretch
and lengthening due to change of geometry can be neglected. The cable spans a very short
distance between each hanger. So in this research the modulus of elasticity for the different
cable types is equal to the given modulus of elasticity for the several available cable systems,
see Figure 21.

Figure 21 Modulus of elasticity of cable types

For the reference model a bundle of parallel wires is chosen on the first hand.

Cable type
For the reference design the same cable type is chosen for the main cable as well as for the
hangers. A cable fabricated with a bundle of parallel strands has the largest modulus of
elasticity compared to other cable types and is therefore chosen. And also for increasing span
lengths it becomes impossible to apply prefabricated locked coil.

12
 Self-anchored suspension bridges: Part II

Stiffening girder
A single beam element is used to model the stiffening girder. In that way the mechanical
properties can easily be adopted. The girder is located in the middle of the two cable planes
and is connected to hangers by means of ‘rigid arms’.
A rigid arm is a connection between nodes with infinite stiffness which transfers all
deformations from one node to the other node.
An example is given in Figure 22 to illustrate the behaviour of a rigid arm. A simple model is
given of two horizontal beams that are connected with a vertical rigid arm. The top beam is
simply supported and the bottom beam in unsupported and connected with a rigid arm to the
top beam. The top beam is loaded with two load cases; a concentrated load and a torque at
mid span. For both situations it is shown that the rigid arm transfers the deformation to the
lower beam.

Figure 22 Rigid arm

On the left side is illustrated that the concentrated nodal force deflects the top beam and the
bottom beam follows the same vertical translation. On the right side the beam is loaded with a
bending moment causing the top beam to bend as illustrated. The unsupported bottom beam
follows this deflection by means of a rotation and translation.

So the rigid connection between the cable plane and the stiffening girder transfers the
deformation of the cable and hangers to the stiffening girder. For the analyses of the required
mechanical properties of the stiffening girder, reference is made to Figure 18 in which
application of rigid arms is visible.

Supports
Along the length of the bridge, the girder is
vertically supported on four locations: end
supports and at the pylons. All the supports
are vertically fixed and also rotation fixed
around the longitudinal direction of the
bridge deck to create a support reaction
similar to a system of two supports.
The stiffening girder is vertically supported
on bearings at the pylon. The cross section of
the pylon will be a simple square box section.
In that the mechanical properties can be
easily assigned in the model.
Figure 23 Modelling support of stiffening girder at pylon

13
 Self-anchored suspension bridges: Part II

Loads
To apply the before mentioned traffic loads in the model, it is translated to a resulting line
load which acts as a distributed line load along the longitudinal direction of the stiffening
girder. Because of the asymmetry of the traffic loads, the resulting line load has a certain
eccentricity to the gravity centre of the box girder and a value of qres,traffic = 131.5 kN/m based
on the information given in Figure 10, see also . Figure 24
Load Area
qfk = 5 kN/m2
Pedestrian and cyclist area
qrk1 = 9 kN/m2
Lane 1
Other lanes and remaining
qrk = 2,5 kN/m2 area
131,5 kN/m1 Resulting Line load
Figure 24 Resulting line load on box girder from distributed traffic load

Also the axle three loads are reduced to a resulting concentrated load Fres = 600 + 400 + 200 =
1200 kN with an eccentricity of 7.7metres. For pre-design reasons the concentrated loads are
left out of consideration because there global influence is not significant. For local design of
the orthotropic deck it becomes important to consider these local loading conditions for
design of longitudinal and transverse stiffeners.

1.3 Dimensioning
This part makes a first estimation of the required dimensions for the main bridge parts; girder,
pylon, main cable and hangers.

1.3.1 Design stress

Transverse stresses
For the stress verification a distinction should be made between the upper and lower flange.
The lower flange is loaded only in length longitudinal direction of the bridge and consists of
normal stresses caused by bending moments and axial forces. The upper flange is extra loaded
due to local traffic loading. This means there exists a number of stresses longitudinal as well
as in transverse direction. According to annex E of the Eurocode NEN-EN-1993-2 the
following combination needs to be taken into account.

14
 Self-anchored suspension bridges: Part II

In which the combination factor ψ = 0.7 for main span over 40 metres or determined on the
basis of the weight distributions of several lorries.

In this study the calculation of stress is limited to an assessment of the global stresses.
Because the local stresses caused by wheel loading on the deck have to be taken into account,
a certain design stress level is chosen for the global stresses.

To account for longitudinal-, transverse stresses and fatigue, which is most of times the
governing design aspect on detail level, a design stress level of 200 N/mm2 is assumed for the
structural steel used for the girder. This design stress level is used for determining the
required cross section of the structural members in the bridge. This design value for the stress
is used for determination of cross sections under static actions. Although a fatigue assessment
is still necessary, it is left out of consideration in this research.

Stress distribution
The stress distribution is assumed to develop linearly over web of the cross section with a
maximum stress in the outer fibres of the cross section.
The normal stresses are caused by two components; the normal stresses caused by a normal
force and the normal stresses caused by a bending moment. These two components can be
superposed to determine the
total normal stress that acts in
the outer fibres. This
maximum normal stress
(caused by the design value of
the normal force and bending
moment) may not exceed the
design value of the yield
strength fy.

Figure 25 Superposition of linear stress distribution over the web of the cross section

15
 Self-anchored suspension bridges: Part II

The bending stress distribution over the top and bottom flanges will be calculated according
to the effective width
method in which a reduced
cross section is determined,
Aeff. The normal stresses can
be linearly distributed over
this reduced area of the cross
section of the top and
bottom flanges, see Figure
26. The effective width be is
determined according the
Euro code.
Figure 26 Distribution stresses in flanges

So the stress distribution over the cross section of the box girder is:
-Normal force in the deck NEd A total
-Bending moments in the deck My,Ed A effective

1.3.2 Estimation by hand calculation

Stiffening girder
The bending stiffness EI of the girder is one of the most important design factor. For the first
estimation of the dimension of the cross section of the box girder an average height
slenderness is chosen 1/70 * main span (based on data given in the literature survey of this
research). So the box girder height is: 1/70 * 150 = 2.1 metres
The girder is chosen to be of a continuous type which means that there are no hinges in the
girder.

Used plate thickness in for instance the girder of the Konohana bridge varied between 12-20
mm.
Chosen equivalent thicknesses for the reference model in this research (so including an
additional needed area for the application of longitudinal stiffeners on deck plate and bottom
flange) are as mentioned earlier, see §1.1.2:
t eq. flange top =40 mm
t eq. flange bottom =20 mm
t eq. web =15 mm

Figure 27 Cross section box girder

The mechanical properties of the box girder with these characteristics are given in Table 2.

Table 2 Mechanical properties stiffening girder

A totaal 2.2 m^2
Iy 2.12 m^4 * = The St. Venant torsional constant It is based on a cross section with the area
of longitudinal stiffeners excluded. Because the longitudinal stiffeners have a
Iz 245 m^4 negligible contribution to the torsional stiffness of a cross section.
It * 5.33 m^4
Self weight 173 kN/m

16
 Self-anchored suspension bridges: Part II

Compared to the applied box girder in for instance the Jiang Yin suspension bridge6 the
presented mechanical properties in Table 2 have realistic values, see Figure 28 for the
mechanical properties of this stiffening girder.

Figure 28 Mechanical properties box girder Jiang Yin bridge

In the Appendix 4 Comparable box girders some more results on comparable bridge decks are
presented.

Cable
The main properties are expressed in the modulus of elasticity for different existing cable
types. Together with the cross sectional area A of the cable, these two properties determine
the axial stiffness EA of the cable.
An approximation of the cross section is determined by the largest normal force in the cable.
The horizontal component of the tension H in the cable, is constant along the main cable. In
the mid of the main span, only the H is acting in the cable because the cable is horizontal on
this location. The largest tensile force in cable acts directly on the side span side of the pylon:
the cable on this location has the largest angle with the horizontal. See Figure 29.

α1 side span = 41.9˚ α2 main span = 37.5˚

Figure 29 Angle of main cable in side and main span

The horizontal component H on location is the same and also a vertical component V acts in
the cable. The largest normal force Ncable in cable is therefore determined by :
N cable = H 2 + V 2 and
qG + Q * l 2
H = in which
8* f1
qG+Q = uniformly distributed load
G = permanent load

6
Cheng, J. et. al., Nonlinear aerostatic stability analysis of Jiang Yin suspension bridge. Engineering structures
24, 2002, pp 773-78.

17
 Self-anchored suspension bridges: Part II

Q = variable load
l = length of the main span
f1 = sag of the cable in main span

For the first approximation of the dimension of the cable the fatigue strength criterion is used.

The following steps are taken to determine the cable dimensions:

η = ratio between variable load and self weight.

q = variable load
g = self weight of the structure + permanent loading

Figure 30 Allowable cable stress

1:
According to NEN-EN 1991-3 part 4.6.2, in fatigue load model 1 the following values for the
axle loads Qik and uniformly distributed loads qik have to be used:
0.7 * Qik
0.3 * qik
But the contribution of the concentrated axle force of 1200 kN is due to the length of the
bridge relatively small and therefore not included in this pre-design. Also the influence of the

18
 Self-anchored suspension bridges: Part II

variable load Qik is very small for the main cable, the hangers are more sensitive for these
loadings.

Variable load q = 0.3 *131.5 kN/m traffic load

Self weight g = 173 kN/m box girder
6 kN/m estimation cable weight, diameter d = 300mm
32.8 kN/m asphalt layer of 40mm
0.3 * 131.5
η= = 0.19
Ratio 173 + 6 + 32.8

2:
Maximum level of Δσ = 200 N/mm2 is assumed, see Figure 30.

3:
Maximum allowable stress caused by self weight and permanent loading σg is approximately
σg = 600 N/mm2. (determined by use of the design graph in Figure 30)

4:
Total permanent load design value Gd = γG * (173 + 6 +32.8) = 1.35 * 211.8 = 285.9 kN/m
Design value of the horizontal component of tension in the cable
qG * l 2 285.9 * 150 2
H d ,G = = = 26803 kN , so per cable acts 13402 kN
8* f1 8 * 30
Maximum design value of the axial tension force in the main cable
N d ,G , cable = H d ,G + Vd ,G = H d ,G + (tan α1 * H d ,G ) 2 = 18006 kN
2 2 2

N d ,G
The required effective area is therefore Aeffective;required = = 30000 mm2
σg
The effective required diameter of the cable is therefore: d cable required= 200 mm (leaving the
fill factor of the cable out of consideration). To be a bit conservative an effective diameter d
cable = 240 mm is chosen.

Hangers
The hanger distance is 6.25 metres. To determine the required cross section, it is assumed that
the self weight and permanent load of the girder is uniformly distributed over the hangers. So
each hanger carries 6.25 metres of the girder. As the hangers are more sensitive for the
variable axle loading Qik, a distribution of 50 percent in each hangers is assumed for pre-
design estimation.
The following steps are taken to determine the hanger dimensions:
1:
In fatigue load model 1 the values for the axle loads 0.7 * Qik and uniformly distributed loads
0.7 * qik are used:
Variable load q = 0.3*131.5 kN/m traffic load
Q= 0.7*1200 kN axle loads
Self weight g = 173 kN/m box girder
6 kN/m estimation cable weight, with d = 200mm
32.8 kN/m asphalt
0.3 *131.5 + (0.7 * 0.5 * 1200) / 6.25
Ratio η = = 0.51
173 + 6 + 32.8

19
 Self-anchored suspension bridges: Part II

2:
Maximum level of Δσ = 200 N/mm2 is assumed

3:
Maximum allowable stress caused by self weight and permanent loading σg is approximately
σg = 350 N/mm2. (determined by use of the design graph in Figure 30)

4:
Total permanent load design value Gd = γG * (173 + 6 +32.8) = 1.35 * 211.8 = 285.9 kN/m
Design value of the vertical force in the hanger
N d ,G , hanger = 6.25 * 285.9 = 1787 kN per two hangers, so that is 894 kN per hanger
N d ,G
Aeffective;required = = 2556 mm2
σg
The required diameter of the hanger is therefore: d hanger required = 51mm (leaving the fill factor
of the cable out of consideration). To be a bit conservative a diameter of the hanger of
dhanger = 55 mm is chosen.

pylon:
The pylon should have an axial stiffness and a bending stiffness because it is loaded with a
large axial compression force caused by the vertical tension component of the main cable
And also a bending moment, caused by a horizontal deflection of the pylon, can act in the
pylon with a maximum at the base.

qG + Q * l 2 (1.35 * 211.8 + 1.5 *131.5) *1502

H d ,G + Q = = = 45298 kN, so that is 22649 kN per
8* f1 8 * 30
cable.
The normal force in the pylon is therefore:

N d ,G + Q , pylon = tan α1 * H d ,G + Q + tan α 2 * H d ,G + Q = 37700 kN

With an assumed design yield strength of σd = 200 N/mm2, the minimum required cross
section of the pylon is:

N d ,G + Q , pylon
Arequired , pylon = = 188505 mm2. To account for a possible additional bending
σd
moment, a cross section of 300000 mm2 is chosen.

A minimum plate thickness is assumed of tpylon = 25 mm.

The pylon is represented by a square cross section and mechanical properties are:
A total 0,3 m^2
Iy 0,45 m^4
Iz 0,45 m^4

Figure 31 Schematization of cross section of the pylon

20
 Self-anchored suspension bridges: Part II

This approach makes it easy to assign certain mechanical properties to the pylon and keeping
in mind that the same mechanical requirements can be achieved by a totally different cross
section, for instance with much more higher plate thickness and other dimensions. So the
emphasis lies on the mechanical requirements of the pylon and not the geometric properties.

1.3.3 Final dimensions reference model

Loading cases
• To determine the final dimensions of the structural components, the governing loading
conditions for strength and stiffness need to be determined. Several cross sections are
considered and the governing loading conditions for each cross will be determined.
For pre-design, only symmetric loading cases are considered where the traffic loading
is centrally positioned on the deck.

The three symmetrical load cases are considered:

• Load case 1: traffic load over full length

Figure 32 Load case 1

• Load case 2: traffic load over mid span

Figure 33 Load case 2

• Load case 3: traffic load over side span

Figure 34 Load case 3

• Several asymmetric traffic load cases will be discussed in §3.5, here an evaluation is
given of the consequences of asymmetric loading of the bridge. In case of asymmetric
loading, traffic load is eccentrically positioned on the deck, see Figure 24.

These load cases represent the different traffic conditions and will be combined with the self
weight of the bridge, permanent loading like asphalt layer, and the pretensioning of the main
cable.
Because of the little influence of the axle load on the global behaviour, these are left out of
consideration. The load is symmetrically applied, without eccentricity as a line load along the
length of the stiffening girder.

21
 Self-anchored suspension bridges: Part II

Governing cross section for static strength and stiffness

The three considered load cases have three different effects for the moment distribution and
deflections along the length of the bridge. Later on four cross section of the girder are verified
on the stresses if they meet the static strength criteria. Those four cross sections are:
1. End support
2. Mid of side span
3. Support at pylon
4. Mid of main span

Figure 35 Considered cross section 1-4

In the next figure an illustration is given of the moment distribution caused by the three
different traffic load cases and due to self weight of the bridge including pretension of the
main cable as explained in §1.1.7.
traffic load case 1
Bending moment along girder
traffic load case 2
150000
traffic load case 3
self weight + pretension
100000
Bending m om ent [kNm ]

50000

0
0 50 100 150 200 250 300

-50000

-100000

-150000
Position on bridge [m]

Figure 36 Bending moment along the deck

The following can be said about the bending moment distribution:

- Cross section 1 End support: The bending moment in the end support is zero, only the
normal force determines the level of stress in this cross section.

- Cross section 2 Mid of main span: Load case 2 traffic over the mid span is governing for
stress distribution in this cross section.

- Cross section 3 Support at pylon: Load case 1 traffic over the entire length of the bridge is
governing for this cross section and causes the largest bending moment along the length of the
bridge at the support location. Figure 36 shows that the hogging moment at the support
location of the girder near the pylon is quite large. In practice this kind of local problems at
supports can be solved by:
• Local application of additional plate thicknesses to reduce the stresses under the
design stress level.
• Adjustment of the internal forces by movement of the supports. For statically
undetermined systems, the load distribution due to traffic loading and therefore the

22
 Self-anchored suspension bridges: Part II

stress distribution can be greatly affected by a displacement of the intermediate or end

supports.

- Cross section 4 Mid of main span: Load case 2 traffic over the mid span is governing for the
stress distribution in this cross section.

For the deflection of the girder clearly shows that load case 2 causes the largest deflection at
the mid of the main span. Load case 3, traffic load on side spans, gives an upward deflection
of the girder. So when the stiffness criteria are checked, the deflection of the main span has to
be within a specified limit.
traffic load case 1
Deflections along girder traffic load case 2
200 traffic load case 3
self weight + pretension
100
Deflection [mm]

0
0 50 100 150 200 250 300
-100

-200

-300

-400

-500
Position on bridge [m]

Figure 37 Deflections along the deck

Stress distribution
From the previous figures it can easily be seen that the longitudinal stresses caused by
bending and axial force in girder at pylon support location, are decisive. For ease of designing
only one cross section is considered, namely the mid of the main span. Designing for the
stress conditions at this location, determines the cross section to be used along the whole
length of the girder.

Tabel 3 Dimensions reference design

Reference design Girder t_flange_top_eq 40 mm
For the mid of the main span, t_flange_bottom_eq 20 mm
a maximum allowable design t_web_eq 15 mm
stress of 200 N/mm2 is heigth 2100 mm
chosen for pre design width 35600 mm
purposes. After a hand
calculation of several bridge Main cable diameter 160 mm
components, the bridge Hangers diameter 55 mm
dimensions and mechanical
properties for the reference Pylon width 2000 mm
design are presented in Tabel depth 2000 mm
3. total height 50000 mm

23
 Self-anchored suspension bridges: Part II

This reference model will be checked on the design criteria like; static strength, stiffness,
stability, and frequency behaviour. See Figure 38 for the final dimensions of the reference
model for the research that is described in this report.

Figure 38 Final dimensions reference design

24
 Self-anchored suspension bridges: Part II

2 Verification of reference design

The reference design has to meet certain design criteria according to the Eurocode. This
chapter explains the method for the verification of the reference design on static strength,
stiffness and stability. Secondly the reference is verified according to the described method.
Verification on stresses, deformation and stability of the reference is given in this chapter.

2.1 Method for verification on static strength, stiffness and stability

The design should be verified on certain criteria to satisfy on static strength as well as fatigue
strength, stiffness and stability. A verification of the design according to the Eurocode will be
made. This part explains the method for checking the design for the mentioned criteria like
strength, stiffness and stability. Also an assessment has to be done for the frequency
behaviour of the bridge. Fatigue strength is not considered in this M.Sc. study because it is
assumed that this will not govern the global design of bridge.

2.1.1 Static strength (ULS)

When checking the design on static strength, the level of stresses may not exceed certain
material related design values. Stresses are caused by normal forces, bending moments and a
combination of these two. Shear forces will be left out of scope because these will not govern
the global design. According to the Euro code, strength criteria have to be checked in the
ultimate limit state (ULS).
The level of stresses will be checked in the decisive cross section of the:
• Pylon combination of bending moment and normal force
• Girder combination of bending moment and normal force
• Cable only a normal force
• Hanger only a normal force

Design values
According to the Eurocode the following design values for the yield strength of the steel
themselves can be used.

Figure 39 Design values steel

For this research the rather standard steel grade S355 is used.
For the main cable and hangers different steel grades can be used, so called high strength
steel. The nominal tensile strength for steel round wires is 1770 N/mm2.

25
 Self-anchored suspension bridges: Part II

In the ULS, design values have to be applied for the permanent and variable actions that act
on the structure. According to the Eurocode EN-1990 the following partial load factors for
determining the design values are used:
For self weight and permanent loading: γGj,sup = 1.35
For variable loading: γQi = 1.5

Effective width and shear lag

In the elementary beam theory the longitudinal normal stresses induced in the flanges are
assumed to be uniformly distributed across the flange width. In case of wide flanges this
assumption is not correct. The stresses in wide flanges are non-uniformly distributed due to
shear deformation of the flange plates. This phenomenon is called shear lag.

Figure 40 Left: stress distribution without shear deformation. Right: Stress distribution with shear
deformation

When checking the stress level of the girder, the effect of shear lag should be taken into
account (shear lag in the pylon will not be considered).
Eurocode NEN-EN 1993-1-5 part 3.1 states that shear lag effects in the flanges may be
neglected when it satisfies:
b0 < Le/20 for ultimate limit state
in which b0 is the deck width and Le is the length between points of zero bending moments
In this case the deck width of the box girder is 35.6 metres and the main span of the reference
model is 150 metres, so this criterion will never be satisfied. So shear lag effects in the flanges
of the box girder can not be neglected.
For the pylon the effective width effect with respect to shear lag is neglected and left out of
consideration.
The governing cross section depends on the effective width along the girder, the loading
conditions and the combination of normal force and bending.

Three cross sections (table 3.1 NEN-EN 1993-1-5) will be checked:

-sagging moment of main span β1
-sagging moment of side span β1
-hogging moment at support pylon β2
-end support β0

26
 Self-anchored suspension bridges: Part II

The effective width for each cross section is:

beff = β*b0

β is β1, β2 or β0 depending on the

considered cross section. β is called
the effective width factor and is
determined according to table 3.1:

So for the top and bottom flange, an

effective width is determined for
several cross sections (mid of the side
span, at support location at pylon, mid
of the main span). In combination
with the acting normal stresses this
will result in a governing cross
section of the stiffening girder which
has to satisfy certain unity check on
strength.

Regarding the resistance of the normal force in the deck NEd, the total area of the box girder
will contribute to this. So no effective width assessment is needed for the resistance of the
normal force in the deck

Verification
According tot NEN-EN-1993-1-1 in the ULS the strength verification in the box girder should
satisfy the following criteria:
For class 3 cross sections:
In absence of shear force, for class 3 cross –sections the maximum longitudinal stress shall
satisfy the criterion:
f
σ Ed ≤
y

γ M0

According to EN-1993-1-11 part 6 for cables, in the ULS, it shall be verified that:
FEd
≤1
FRd

2.1.2 Stiffness (SLS)

The serviceability limit state concerns the functioning of the structure, comfort of people and
the appearance of the construction works. Stiffness criteria are expressed in certain
deformation tolerances.
• Vertical allowable deflection of girder.
• Horizontal allowable deformation of pylon.
Besides the criteria on displacements, also criteria exist on passenger comfort which are
expressed in limitations for vertical acceleration. This criteria is assumed to satisfy for
medium span bridges considered in this M.Sc. study.

Further on the following maximum deflection are assumed:

• maximum vertical deflection of 1/350*Lmain span for the stiffening girder

27
 Self-anchored suspension bridges: Part II

• maximum horizontal deflection of 1/300*Lpylon for the pylon

No particular deflections limits could be found in the Eurocode, therefore the following
maximum deflections are assumed7 according to the bridge design of the Kanne self-anchored
suspension bridge.
So the vertical deflection of the mid of the main span may not exceed
1/350 * 150 = 0.43 m
The horizontal displacement of the pylon may not exceed
1/300 * 49.75 = 0.17 m

Design values SLS

Stiffness criteria are checked in the serviceability limit state (SLS) The partial factors for
determining the design value of the actions in SLS are equal to 1.

2.1.3 Stability (ULS)

Second order analyses (stability)
A second order analyses is performed because in cable supported structures geometrical non
linearity can be of importance. In general long span bridges such as cable stayed- and
suspension bridges exhibits geometric non linearity due to:
• The combination of axial compression forces and bending moments that act in the
stiffening girder and the pylon.
• The non linear behaviour caused by the cable. The relation between forces and the
resulting deformations are not linear. (e.g. an increased self weight load in the cable
results in a reduction of live load deflection 8,9. The tensile force in the cable produces
a geometrically non-linear stiffness of the cable.)
• Geometry changes in the bridge structure caused by large displacements.

To make an assessment of the geometrical non linearity’s in a cable supported bridge, the so
called ‘n-value’ gives an indication. The n-value gives information about the consequences of
geometrical non-linearity. In bridge design a general guidance can be used with respect to this
n-value:
1<n<2 wrong design
2<n<3 design problems to be expected
n>3 proper design, however, geometrical non-linearity should be taken into
account
n > 50 consequences of geometrical non linearity can be neglected

Also the Eurocode gives an guidance with respect to the n-value. Eurocode NEN-EN 1993-1-
1 part 6.3.1.2 states:
N Ed
When ≤ 0.04 , the buckling effects may be ignored and only cross sectional checks
N cr
apply.
Ncr = is the elastic critical force for the relevant buckling mode based on the gross cross
sectional properties.
NEd = design value of the axial force

7
These stiffness criteria were also used for the Self-anchored suspension bridge in Kanne in Belgium.
8
Gimsing, N.J., Cable supported Bridges, Wiley&Sons, 1998
9
Gasparini,D. V. Gautam. Geometrically Nonlinear Static behaviour of Cable Structures. Journal of Structural
Engineering, October 2002, pp 1317-1329.

28
 Self-anchored suspension bridges: Part II

Because N cr = n * N Ed , according to the Eurocode buckling can be ignored when the

n-value is: n ≥ 25
Determination of n-value and Euler buckling force Ncr
The amplification of the deflections can be determined by the difference between the
deflection calculated by a linear and a geometrical non-linear calculation.
n δ
The amplification factor is: = 2 in which
n − 1 δ1
δ 2 = the deflection determined by a second order analyses (geometric non-linear)
δ1 = the deflection determined by a first order analyses (linear).
Rewriting the formula for the amplification factor leads to formula to determine the
δ2
n-value: n =
δ 2 − δ1
With this n-value the Euler buckling force Ncr is determined by: N cr = n * N Ed
This means that the n-value represents the ratio between the design value of the axial force
and the Euler buckling force. An evaluation of the calculation of the Euler buckling force is
made further in this report when a stability check is done.
The next example explains and compares the calculation of the Euler buckling force with the
n-value and the theoretical buckling formula. As shown in Figure 41 both results correspond.
The second order deflections are calculated with the FE program ESA PT 6.0.185.
Column properties:
Iy = 1.56*10-5 m4
E = 210000 N/mm2
A = 2.7*10-3 m2
L = 10 m
π 2 EI
The theoretical Euler buckling force: N cr = 2
= 324kN
l buc
Buckling force according to the n-value:

δ 1 = 64 mm
δ 2 = 165 mm
δ2 165
n= = = 1.63
δ 2 − δ 1 165 − 64
N cr = n * N Ed = 1.63 * 200 = 327 kN
Figure 41 Example buckling force

According to the guidelines regarding the interpretation of the n-values, the presented
example on this page would mean an improper design because the n-value is < 3.
The same method is used to make an assessment of the second order effects in self-anchored
bridge in this M.Sc. study. Further on in this report a stability check is done for the stiffening
girder.

29
 Self-anchored suspension bridges: Part II

Stability check according to the Eurocode

Because the stiffening girder is subjected to a normal force and a bending moment, the
buckling resistance should be assessed for a member in bending and compression. The Euro
code 1993-2:2003 states that members under combined bending and axial bending should
satisfy:
N Ed β m ( M y , Ed + ΔM y , Ed )
+ ≤ 0.9 in which
χN cr M y , Rk
γ M1 γ M1
NEd is the design value of the compression force
My,Ed is the design value of the maximum moment about the y-y axis of the member
calculated with first order analyses and without using imperfections
ΔMy is the moment due to shift of the centroidal axis according to 6.2.10.3
βm is the equivalent moment factor, see table A.2 of EN-1993-1-1
χy is the reduction factors due to flexural buckling from 6.3.1

So for the stiffening girder (buckling of the pylon is not considered), only the global buckling
will be considered. Local buckling of the compression flanges is a matter of detailed design of
the box girder for the stiffeners longitudinal and transverse, cross beam, and cross frames. The
required cross sectional area for stiffeners is included in the equivalent plate thickness that is
chosen for the top- and bottom flanges of the box girder.

Buckling reduction factor χ

For the verification of the unity check the buckling reduction factor χ needs to be determined
by the following procedure:

Relative slenderness
Af y
λ=
N cr
In which
N cr = n * N Ed
Ncr = is the elastic critical force for the relevant buckling mode based on the gross
cross sectional properties.
NEd = design value of the axial force
n = the amplification determined from deflection in the first order (linear) and
second order (geometric non-linear) analysis of the bridge model.
δ2
n= in which δ 2 is the deflection determined by a second order analyses (geometric
δ 2 − δ1
non linear )and δ1 is the deflection determined by a first order analyses (linear).

30
 Self-anchored suspension bridges: Part II

And now the buckling resistance factor according to the Eurocode is determined by:

As mentioned earlier, the box girder is assumed to be a class 3 cross section. And according to
the Euro code 1993-1-1, the buckling resistance factor χ for welded box section in general has
to be determined by buckling curve b.

So the verification for buckling stability will be done according to the Euro code 1993-1-
1:2003 part 5.2 Global analysis and 6.3 buckling resistance of members and EN-1993-2:2003
section 6.3.

2.1.4 Frequency analyses

In order to determine the response of the bridge to dynamical loadings like wind- and traffic
loading, it is necessary to evaluate the frequency behaviour. The frequency behaviour of the a
bridge can be characterised by their natural frequencies.
Dynamic wind actions can cause several response phenomena of the bridge deck like
galloping, vortex shedding and flutter. Dynamic loading depends on many factors10 such as ,
weight of the bridge, dimensions, cross sectional shape, displacements, speed and acceleration
of bridge components etcetera. It is therefore a complex phenomena and in many cases of
possible bridge response to wind a wind tunnel research is required.
Because of the complexity of this topic, this M.Sc. study considers only the natural
frequencies of the bridge.

10
Romeijn, A. Examples of examination questions for Cable stayed bridges. December 2005

31
 Self-anchored suspension bridges: Part II

For the evaluation for risk for vortex shedding as well as for flutter response of the bridge, the
natural frequencies of bending and torsion frequencies are required data. The next formulas11
are used for determining the critical wind speed when vortex and flutter can be expected:
• Critical wind speed vortex shedding Vcr:
Vcr ;vortex = 12 * f * d 4 [m/s] for b / d 4 ≥ 10
f = smallest value of the bending frequency fb or torsional frequency ft
d4 = height of the girder
The bridge is stable when the next criteria is met:
Vcr > Vr
In which Vr is a reference wind speed according to BS 5400.

• Critical wind speed flutter Vf:

⎛ f ⎞ m*r
Vcr ; flutter = 4 * f t * b * ⎜⎜1 − b ⎟⎟ * [m/s]
⎝ ft ⎠ ρ * b3
m = mass of bridge deck per unit of length
Ip [kg/m]
In which r = ,
Agirder Ip = polar moment of inertia

The bridge is stable against flutter when:

Vf > 1.3*Vr
• Other instabilities mentioned by BS5400 like wake galloping and Stall flutter have to
be checked when:
b<4*d4 in which b = deck width and d4 = structural height of box girder.

The before mentioned formulas show that response analyses of the bridge requires the natural
bending- and torsional frequencies. These will be determined by the FEM using ESA PT
6.0.185

When an evaluation for flutter is made, a general rule of thumb is used that the ratio bending-
torsional frequency of 2.0 or more is recommended12. This is accepted as a sufficient
difference between the bending frequency and the torsional frequency which results in better
resistance against flutter, see Figure 42.

Figure 42 Critical wind speed flutter

11
Romeijn, A. Examples of examination questions for Cable stayed bridges. December 2005
12
Chen, W. L.Duan. Bridge Engineering Handbook. CRC Press 2000.

32
 Self-anchored suspension bridges: Part II

Determination of frequencies
For beams under bending a general formula13 is known to determine bending frequencies for
several modes. The natural frequency fn for bending modes can be calculated with:
C EI
fn = n
2π ml 3
The factor Cn depends on the support conditions and the considered frequency mode. For the
first bending frequency holds:

Cn = 9.87
Figure 43 First bending mode

Next a worked out example is given to compared the bending frequency of a simply supported
beam calculated by the before mentioned formula and the frequency calculated when the
beam is modelled in FE program ESA PT 6.0.185.
Beam properties HEA700A:
Iy = 1.43*10-3 m4
E = 210000 N/mm2
A = 1.91*10-2 m2
Figure 44 Frequency simply supported beam
L = 15 m
• The theoretical bending
frequency:
C EI
fn = n = 9.786 Hz
2π ml 3
m = mass of the beam [kg]

• Bending frequency according to

ESA PT: 9.78 Hz

This example shows that the frequency calculation by ESA is a well approximation compared
to the theory. So for further frequency calculation of the bridge deck, the results calculated by
ESA PT 6.0.185 are used.

For verification of these results for the bridge deck, a simple approximation method is
developed by Raleigh- to determine the first bending frequency of a bridge deck.
An estimation of the first bending frequency can be done with a method developed by
Raleigh14:
1.1 g 0.55
fb = = hertz
2π δ max δ max
So results of a frequency calculation by ESA PT for the bridge deck will be quickly verified
by this formula to confirm if the given values are realistic. Damping effects of the structure
are left out of consideration.

13
Overspannend staal, Rotterdam: Stichting Kennisoverdracht SG, Deel 3: Construeren B, 1996.
14
Romeijn, A. Examples examination question: topic cable stayed bridges.

33
 Self-anchored suspension bridges: Part II

2.2 Verification of reference design

This part gives the design verification according to the Eurocode as described in the previous
paragraph 2. This verification comprises static strength, stiffness and stability of the girder,
pylon, main cable and hangers.

2.2.1 Verification reference design on strength

Verification on strength is done for the ultimate limit state (ULS), meaning that the partial
factor for permanent loading is 1.35 and the partial factor for variable loading equals 1.5. The
total stress calculation is given in Appendix 7 Stress calculation in reference model.
Presented stresses are based on an effective cross section, see Appendix 2a Effective width at
main and side span and Appendix 2b Effective width at support for calculation of the effective
widths.

Girder
Four locations are checked for the girder.

• Mid of the main span:

Load case 2 (traffic over mid span) is governing for this cross section
The acting forces are in load case 2:
M_y,Ed,girder 164378 kNm
N_Ed 32827 kN
This combination of bending and axial force results in the following longitudinal
stresses:
Top flange 139 N/mm2 in compression.
σ 139 139
Unity check Ed = = = 0.7 ≤ 1 Satisfies
σ c , Rd 200 200
Bottom flange 197 N/mm2 in tension.

σ Ed 197 197
Unity check = = = 0.99 ≤ 1 Satisfies
σ t , Rd 200 200
• Mid of the side span:
Load case 2 (traffic over mid span) is governing for this cross section
The acting forces are in load case 2:
M_y,Ed,girder 111377 kNm
N_Ed 32827 kN
Resulting stresses:
Top flange 207 N/mm2 in tension.
σ 207 207
Unity check Ed = = = 1.04 ≤ 1 Does not satisfy
σ c , Rd 200 200
Bottom flange 330 N/mm2 in compression.
σ 330 330
Unity check Ed = = = 1.65 ≤ 1 Does not satisfy
σ t , Rd 200 200
The stress in the box girder at this location exceeds the assumed design stress of
200 N/mm2 but are still smaller than the yield strength of steel fy = 355 N/mm2. This
location require more attention in the final design, for example some additional plate
thicknesses at this location can reduce the stresses to an acceptable design level.

34
 Self-anchored suspension bridges: Part II

• Support at pylon
Load case 1 (traffic over full length) is governing for this cross section
The acting forces are in load case 1:
M_y,Ed,girder 193096 kNm
N_Ed 32491 kN
Resulting stresses:
Top flange 466 N/mm2 in tension.
σ 466 466
Unity check Ed = = = 2.33 ≤ 1 Does not satisfy
σ t , Rd 200 200
Bottom flange 783 N/mm2 in compression.

σ Ed 783 783
Unity check = = = 3.92 ≤ 1 Does not satisfy
σ c , Rd 200 200
This cross section does not meet the strength criteria. But this can be locally solved by
application of extra plate thicknesses and adjustment of the supports.

• End support
Load case 2 (traffic over mid span) is governing for this cross section
The acting forces are in load case 2:
M_y,Ed,girder 0 kNm
N_Ed 32827 kN
Resulting stresses:
Top flange 15 N/mm2 in compression.
σ 15 15
Unity check Ed = = = 0.08 ≤ 1 Satisfies
σ c , Rd 200 200
Bottom flange 15 N/mm2 in compression.

σ Ed 15 15
Unity check = = = 0.08 ≤ 1 Satisfies
σ c , Rd 200 200
One thing is notable about the stress distribution in the girder in the reference model,
compared to the stresses caused by bending the normal stresses caused by the compression
force are very low. From this it can be expected that buckling of the girder in this reference
design should not be an issue, see §2.2.3 for the stability check of the girder.

Cable
The governing acting normal force in the ULS is caused by loading case 2 (traffic over mid
span):
NEd = 21547 kN
A main cable = 20106 mm2
Unity check
FEd 21547 *10 3 2154710 3 21547 *10 3
= = = = 0.91 ≤ 1 Satisfies
FRd Fuk f u * Amain _ cable 1770 * 20106
1.5γ R 1.5 1.5

35
 Self-anchored suspension bridges: Part II

Hangers
The governing acting normal force in the ULS is caused by loading case 1(traffic over full
length):
NEd,hanger = 1254 kN
A hanger = 2376 mm2
Unity check
FEd 1254 *10 3 1254 *10 3 1254 *10 3
= = = = 0.45 ≤ 1 Satisfies
FRd Fuk f u * Amain _ cable 1770 * 2376
1.5γ R 1.5 1.5
Pylon
The acting normal force in the pylon is
N Ed, pylon = 28208 kN
M y, Ed,pylon = 25040 kNm (caused by the deflection of pylon towards the midspan)
Resulting stresses:
Top flange 239 N/mm2 in compression
σ 239 239
Unity check Ed = = = 0.67 ≤ 1 Satisfies
σ c , Rd fy 355
Bottom flange 40 N/mm2 in compression
σ 40 40
Unity check Ed = = = 0.11 ≤ 1 Satisfies
σ c , Rd f y 355

Buckling of pylon
Buckling of the pylon should be considered because a compressive force is acting in the
pylon. The pylon is assumed to be fixed at the basement and spring supported at the top of the
pylon because it is resisted in longitudinal direction by the main cable which is supported on
top of the pylon.

NEd

N Ed pylon [kN] 28208

H cable main span
[kN] 16748
H cable side span [kN] 16203
δ pylon [mm] 121
Length [m] 50
4
Iy [m ] 0.46

Table 4 Properties of pylon

Figure 45 Mechanical scheme pylon

Figure 46 Buckling force of fixed column and translation spring

ΔH
The spring stiffness15 is determined by k spring = In which ΔH is the difference between
δ pylon
the horizontal component of tensile force in main cable in the main and the side span.
15
Overspannend staal, Rotterdam: Stichting Kennisoverdracht SG, Deel 3: Construeren B, 1996.

36
 Self-anchored suspension bridges: Part II

k spring l pylon
With the beta factor β = the ratio between the Euler buckling length and the
3EI
system length can be found by using the graph presented in Figure 46.
With the forces on the pylon given by Table 4 the buckling force of the pylon can be
determined:
ΔH 725000
k spring = = = 5.99 * 10 6 N/m
δ pylon 0.121
k spring l pylon 5.99 *10 6 * 50
β= = = 2.45
3EI 3 * 210000 *10 6 * 0.47
From Figure 46 the ratio between the Euler buckling length and the system length can be
found:
2 2
N cr l sys l sys
= = 0.7 so lbuck = 59.8 metres
π 2 EI lbuck
2

The Euler buckling force:

π 2 EI
N cr = 2
= 270MN >>NEd = 28MN
l buc
So the pylon is stable against global buckling.

2.2.2 Verification reference design on stiffness

Stiffness criteria are checked in the serviceability limit state, meaning that that the partial
factor for the loading are equal to 1.
Two deflections are verified, the vertical deflection of the girder at the mid of the main span
and the horizontal; displacement of the tower.
Load case 2 is governing for the vertical deflection of girder and displacement of the pylon:
δ main span = 422 mm < 1/350 * L main span < 430 mm

Horizontal displacement of the pylon:

δ pylon = 121 mm < 1/300 * H pylon < 170 mm

On stiffness criteria the reference design meets the requirements.

37
 Self-anchored suspension bridges: Part II

2.2.3 Verification reference design on stability girder

A stability check has to be done for the stiffening girder because it is loaded with a large axial
compression force. A stiffening girder under Deflections caused by self weight and traffic
compression is prone to global buckling over full length first order deflections

second order
effects. To make an assessment of the
100
buckling risk of the girder, the Euler buckling 0
force has to be determined. -100
0 50 100 150 200 250 300

Deflection [mm]
N cr = n * N Ed -200

-300

δ2 -400

in which n = -500

δ 2 − δ1 -600

-700
The n-value represents a value that indicates Position on bridge [m]

the risk for global buckling of the bridge Deflections caused by self weight and traffic
deck. over main span first order deflections

second order

400
As mentioned earlier, an indication for the 200
stability of the bridge deck is the occurrence 0
Deflection [mm]

of second order effects regarding the -200

0 50 100 150 200 250 300

deflections of the bridge deck. -400

At first sight, the reference model revealed -600

-800
hardly any second order effects (which means -1000
that the deflection calculated by a linear and Position on bridge [m]

geometric non-linear analyses does not Deflections caused by self weight and traffic
over the two side spans
deviate significantly). There is no first order deflections

second order
amplification of the deflection of the
60
stiffening girder visible in the second order 40

analyses under the given loading conditions. 20

0
Deflection [mm]

Figure 47 shows that for the three considered -20 0

-40
50 100 150 200 250 300

load combinations the second order effects -60

-80
are hardly visible. This indicates that the -100
-120
reference bridge model behaves very stiff and -140
-160
that the deflections are relatively low to cause Position on bridge [m]

major second order effects.

Figure 47 1st and 2nd order deflection
due to the three considered loading
combinations including pretension of
main cable

These results are similar to the design calculations of the self-anchored suspension bridges
Kanne16 bridge and the Nescio17 bridge. Both calculation documents display the finding that
second order effects are hardly visible, both bridges have a similar main span of about
100 metres so is therefore comparable to the reference model in this research.

16
Alsemgeest,D. Rebuilding bridge Kanne, Suspension bridge-Static analyses- Check on strength, stiffness and
stability. Iv-Infra, October 2003
17
Ichimaru, Y., Design and engineering of ‘Nescio’ bridge-Amsterdam Rhine canal. Arup

38
 Self-anchored suspension bridges: Part II

Causes for the hardly visible second order effects could be that:
• The combination of axial forces and bending moments that act in the stiffening girder
and the pylon are not significant enough to cause visible second order effects in the
deflection of the stiffening girder. This reference model showed normal stresses in the
deck, caused by the deck compression force, of about 15 N/mm2. This is relatively
low.
• The stiffening effect of the girder. The illustration that is presented in Figure 48
indicates that regarding the reference model with a main span of 150 metres, the
stiffening effect18 of the bridge’s main girder is significantly large. Figure 48 shows
the relation between the non-dimensional maximum deflection v/l against l (v =
deflection and l = main span length). For different values of the girder bending
stiffness EI, a decreasing influence on
displacement is visible when the main
span(>2000 metres) is increased. For main
span smaller than 2000 metres, this figure
indicates that the stiffness of the girder has
a significant effect on the reduction of the
deflection in the bridge.
For a main span of 150 metres this would
indicate that the stiffness of the deck has
large influence on the reduction of the
deflections of the total bridge structure.
And therefore large geometry changes
(which is in many cases a cause for
geometric non linear behaviour) in the
bridge structure caused by large
displacements are not expected to exhibit in
relatively small spans like 150 metres of the
reference model.
Figure 48 Decreasing stiffening effect

So the combination of the relatively small deflections and stiff behaviour of the girder are
causes for the hardly visible second order effect for the reference model in this study.

Alternative approach to determine buckling force Ncr

With the present normal stresses in the girder of the reference model of about 15 N/mm2, no
second order effects are visible. When an additional normal force is imposed on the stiffening
girder of the reference model of the bridge, second order effects do become visible. Figure 49
illustrates the additional normal force that is applied on the girder, in this case an additional
force is applied of about ten times the actual normal force in the girder.

ΔF ΔF

Figure 49 Additional normal force

18
Clemente,P. G. Nicolosi, A. Raithel. Preliminary design of very long-span suspension bridges. Engineering
structures 22 (2000), 1699-1706.

39
 Self-anchored suspension bridges: Part II

In this case an additional force of ΔF=400000 kN is applied, about ten times the design value
of the normal force in the deck NEd. Now an assessment of the Euler buckling force can be
made with respect to the three considered load cases. A distinction is made between the main
and the side spans because the amplification of the deflections deviates from each other. From
this distinction the decisive Euler buckling force can be retrieved, the smallest buckling force
to cause buckling in the either the main span or the side span is the governing one.
For the reference model the design value of the normal force is NEd depends on loading
combination as presented below:
• Euler buckling force Ncr with traffic over full length
Second order deflection with additional F 1e order N deck = NEd + ΔF = 432420 kN
2e order
500 Main span:
0 δ1 [mm] δ2 [mm] n-value
0 50 100 150 200 250 300
790 1608 1,97
Deflection [mm]

-500
Ncr = 1.97 * 432420 = 851867 kN
-1000

Side span
-1500

δ1 [mm] δ2 [mm] n-value

-2000
position on bridge [m]
48 236 1.26
Ncr = 1.26 * 432420 = 544849 kN
Figure 50 1st and 2nd order deflection with traffic full length

• Euler buckling force Ncr with traffic over mid span

Second order deflection with additional F 1e order N deck = NEd + ΔF = 432524 kN
1000
2e order
Main span:
500 δ1 [mm] δ2 [mm] n-value
883 1882 1.88
Deflection [mm]

-500
0 50 100 150 200 250 300
Ncr = 1.88 * 432524 = 813145 kN
-1000
Side span:
-1500
δ1 [mm] δ2 [mm] n-value
-2000
position on bridge [m]
129 407 1,46
Ncr = 1.46 * 432524 = 631485 kN
Figure 51 1st and 2nd order deflections with traffic mid span

• Euler buckling force Ncr with traffic over the two side spans
Second order deflection with additional F 1e order N deck = NEd + ΔF = 423315 kN
0
2e order
n span:
0 50 100 150 200 250 300

-50
δ1 [mm] δ2 [mm] n-value
155 217 3.5
Deflection [mm]

-100
Ncr = 3.5 * 423315 = 1481603 kN
-150

Side span:
-200
δ1 [mm] δ2 [mm] n-value
-250
position on bridge [m]
84 112 4
Ncr = 4 * 423315 = 1693260 kN
Figure 52 1st and 2nd order deflections with traffic over side spans

40
 Self-anchored suspension bridges: Part II

Difference in buckling of the main and side span

Buckling of the side span occurs at a lower buckling force than buckling of the main span. For
full length- as well as mid span traffic loading the side span is most sensitive to second order
effect. The ratio of the Euler bucking force of the main span and side span is about:
Ncr _ side.span 544849
= = 0.64
Ncr _ main.span 851849

The difference can be explained by the fact that Figure 50 shows that the buckling of the main
span is a downward buckling mode and would therefore encounter upward resistance by the
main cables and hangers. Buckling of the side span is an upward buckling mode and would
therefore encounter no resistance by the main cable and hangers.
The upward buckling of the side span occurs therefore apparently at a lower buckling force
Ncr and is decisive over buckling of the main span.

Conclusion regarding buckling

Decisive is buckling of the side span caused by load case 1 (traffic over full length) occurring
at a Euler buckling force of :
Ncr = 544849 kN which is about 17 times larger than the actual normal force in the girder NEd
(representing an n-value of n = 17 which represents a proper design according to the given
guidelines in §2.1.3). So no buckling instability is expected in the reference model, the
buckling resistance of the stiffening girder is sufficient.

The stability check according to the Eurocode:

N Ed β m ( M y , Ed + ΔM y , Ed )
+ ≤ 0 .9
χN cr M y , Rk
γ M1 γ M1
N Ed β m M y , Ed
+ = 0.477 ≤ 0.9 Satisfies this criteria
χN cr M y , Rk
In which
NEd = 32420 kN
NRk = Ncr = 544849 kN
χ = 0.484 (see Appendix 5a Stability check reference model for calculation stability
check)
βm = 1.27
ΔMy,Ed = not applicable because the entire cross section is assumed to resist the acting
normal force in the cross section of the box girder.
γM1 =1
My,Rk = fy * W_eff box girder side span; top

Overall conclusion is that the bridge girder is satisfies the stability check according to an
alternative approach and the Eurocode check. Therefore the conclusion can be made that the
stiffening girder in the reference model is stable against buckling.

Further research on the stability phenomena of the stiffening girder will be presented in the
parameter study. The stability will be researched as function of the bending stiffness of the
box girder.
Also an exploration is done to the buckling behaviour for an increasing main span, see §4.2.2.

41
 Self-anchored suspension bridges: Part II

2.2.4 Frequency behaviour

An estimation of the first bending frequency can be done with a method developed by
Raleigh19:
1.1 g 0.55 0.55
fb = = = = 0.69 hertz
2π δ max δ max 0.641
In which
δmax = maximum deflection under self weight = 0.641 m, see Figure 14.
g = gravitational acceleration

The results of the calculation of the first bending frequency for the reference model are
determined with ESA PT and the Raleigh method and are given in Table 5.
Table 5 First bending frequency comparison
Raleigh ESA PT
1st Bending
frequency [Hz] 0,69 0,74

So the Raleigh method gives a good approximation of the results for the first bending
frequency given by ESA PT. The small difference between the two results are caused by the
fact that the Raleigh method is an quick approximation method and that the formula is based
on a simply supported beam. Because the bridge deck, in the bridge design under
consideration, is a continuous girder and is therefore more stiff than a simply supported
girder, the first bending frequency is in that case higher.
Further frequency calculations by ESA PT are assumed to be of realistic value.

Natural frequencies reference design

Table 6 gives the natural frequencies of the reference design.
Table 6 Natural frequencies reference design
1st fbending [Hz] 1st ftorsion [Hz]
0,74 5,2

Because the first bending and torsional frequency are clearly well separated, no problems are
expected regarding flutter instability for instance. The ratio between the first bending and
torsional frequency is well above two.

19
Romeijn, A. Examples examination question: topic cable stayed bridges.

42
 Self-anchored suspension bridges: Part II

3 Parameter study into the structural behaviour

Introduction
This chapter presents the results of a parameter study and the influences on the structural
behaviour of a self anchored bridge. Objective of this part of the study is to gain insight in the
behaviour of a cable supported bridge and to study the effect and sensitivity of the mechanical
properties of girder, pylon and cable on the bridge’s behaviour:
• EI girder
• EI pylon
• EA cable
• Sag to span ratio
This part of the final thesis will investigate the influence of before mentioned key bridge
parameters on the structural behaviour:

• Global stiffness (deflections)

• Reaction forces (normal force deck and vertical support reaction)
• Bending moments at support and in the main span of the stiffening girder.
• Stability of the girder (n-value deck): the n value of the deck gives information about
the consequences of geometrical non linearity. The following guidance can be used to
interpret the n-values:
• Frequency behaviour (lowest torsional and bending frequency)

The structural behaviour will be investigated under influence of the girder-, cable, pylon and
sag properties. Results are analysed and presented in tables and graphs to visualize the effects.

Load condition
Only load case 1(traffic over the entire length of the bridge) in combination with self weight,
permanent loading and pretensioning of the main cable is considered in this parametric study.

Results of the parameter study

Results of the parameter study are presented in graphs which illustrate the developments of:
• Bending moments in girder at support and main span
• Deflection of pylon and girder at mid of the main span.
• Frequencies, 1st bending and 1st torsional frequency
• The ratio of the bending moment carried by the deck and cable
These topics show the most significant effects under influence of the different design
parameters.

43
 Self-anchored suspension bridges: Part II

3.1 Girder influence

In this part the structural height of the box girder is varied which height Iy It
[mm] [m^4] [m^4]
changes the bending stiffness (and also the torsional stiffness).
1500 1,1085 2,78139
Based on realistic values of the slenderness of the girder, that has
1600 1,26 3,15281
been retrieved from literature survey, the structural height is varied
1700 1,4213 3,54602
from 1/50 up to 1/100 of the length of the main span. A slenderness 1800 1,5925 3,96076
of 1/50 means a structural height of 3 metres and 1/100 a height of 1900 1,7735 4,39681
1.5 metres. The next table shows the girder height with 2000 1,9643 4,85393
corresponding flexural stiffnesses and torsional stiffnesses. 2100* 2,1651 5,33188
2200 2,3758 5,83044
Increasing the bending stiffness has a significant effect on the 2300 2,5964 6,34938
moment distribution, see Figure 53. A larger stiffness of the girder 2400 2,8269 6,88849
means the bending moments increase significantly, with 2500 3,0674 7,44755
approximately 95 %, see Figure 53. Figure 54 clearly shows that 2600 3,3179 8,02633
with an increasing stiffness, the girder tends to carry a larger part of 2700 3,5785 8,62463
the total bending moment and smaller part is carried by the main 2800 3,849 9,24224
cable, reducing the normal force. 2900 4,1296 9,87896
3000 4,4203 10,5346
*=Reference model
Table 7 Girder mechanical properties
The total bending moment in the main span presented in Figure 54 is determined by:
My;Ed;total = 2*(Hmain cable;Ed* cable sag) + My;Ed;deck
Bending Moments Main span % of total moment carried by deck
300000 Support 20
18
250000
16
Bending moment[kNm]

200000 14
% My_total

12
150000 10
8
100000
6

50000 4
2
0 0
1000 1500 2000 2500 3000 3500 1000 1500 2000 2500 3000 3500
Height [mm] Height [mm]

Figure 53 Maximum bending moment in the girder

Figure 54 Contribution of the girder to bending moment

Looking at the stiffness, an increasing stiffness of the girder result in larger global stiffness
because the deflection reduces significantly. Figure 55 shows this tendency of a decreasing
deflection of the girder and displacement of the pylon. Girder and pylon deflection is reduced
with approximately 50%. An increasing global stiffness also results in higher bending and
torsional frequencies, see figure 40.
Deflection Deck Frequencies 1e frequency torsion
500 Pylon 1e frequency bending
450 6

400
5
350
Deflection [mm]

4
Frequency [Hz]

300
250
3
200

150 2

100
1
50
0 0
1000 1500 2000 2500 3000 3500 1000 1500 2000 2500 3000 3500
Height [mm] Height [mm]

Figure 55 Maximum deflection in the girder

Figure 56 Natural frequencies

44
 Self-anchored suspension bridges: Part II

The stiffness of the girder has a clear visible influence on the stiffness, strength and frequency
behaviour of the bridge.

Regarding the reaction force, the normal force in the deck and resulting vertical reaction at the
end support, Figure 57 shows that both maximum reaction forces decrease with a stiffer deck.
With a stiffer deck, a larger part of the bending moment is carried by the girder which leads
to decreasing normal force in the cable.
N deck
Reaction forces
40000 Vertical reaction
force at end support
35000
30000
Force [kN]

25000
20000
15000
10000
5000
0
1000 1500 2000 2500 3000 3500
Height [mm]

Figure 57 Reaction forces

3.1.1 Stress level in box girder

The stress level in the flanges of box girder is determined by:
N M y ; Ed
σ top = − Ed −
Abox Weff ;top
N Ed M y ; Ed
σ bottom = − +
Abox Weff ;bottom
So the stress level in the flanges depend on the normal force NEd in the girder, bending
moment My;Ed in the girder, the section modulus Weff and cross sectional area Abox of the
girder.

Figure 53 and Figure 57 show that choosing for a more slender girder results in significant
reduction of the bending moments My;Ed and an increasing normal force NEd in the girder. A
side effect is that the effective section modulus Weff also decreases with an increasing
slenderness of the girder. This development is expressed in the stress level in the flanges of
the box girder in the mid of the main span of the bridge, see Figure 58. This figure shows that
the development on the stresses as a function of the girder slenderness λ (height/length main
span) remain quite constant.
Stress top flange
Stress level flanges box girder main span main span;
compression
350 Stress level bottom
flange main span;
300 tension
Stress level [N/mm^2]

250

200

150

100

50

0
0 1000 2000 3000 4000
Height box girder [mm]

Figure 58 Stresses in flanges of the box girder

45
 Self-anchored suspension bridges: Part II

3.1.2 Stability of the stiffening girder

The resistance against buckling of the main girder is mainly determined by the bending
stiffness EI of the box girder. Two approaches will be set out in this part in order to compare
and verify the results of the calculation of the Euler buckling force of the stiffening girder.
Also an approach on the maximum girder slenderness will be given.

1. Additional normal force ΔF

First approach, as mentioned earlier, is that on the bridge deck an additional normal force
is applied to be able to analyze the second order deflections and determine the buckling
force, see Figure 59. The bridge model proved to be very stiff (see 2.2.3) and second order
effects only became visible by increasing the normal force on the deck, in this case by
applying an external additional force ΔF.

ΔF ΔF

Figure 59 Additional normal force applied on the girder

2. Spring model of bridge deck

The second approach is to model the bridge as supported by discrete springs with a certain
spring stiffness k [N/m], see Figure 60. This model is used to analyze the buckling effect
of the main span because the spring stiffness in the side is zero for an upward deflection.
In this model the girder is also loaded by a normal force NEd, similar to acting normal
force in the reference design; NEd = 32420 kN.
NEd NEd

Spring stiffness k

62.5 m 150 m 62.5 m

Figure 60 Spring supported bridge deck

According to Engesser’s20 formula, the Euler buckling force for a girder supported by
springs depends on the spring stiffness (in this case the springs represent the hangers on
which the girder is supported) and the bending stiffness of the girder.
Engesser formula: N cr = 2 cEI in which c is a bedding constant equal to the
spring stiffness divided by the individual distance between the springs (in this case the
c.t.c. distance between the hangers).

Spring stiffness k
A spring stiffness k is iteratively chosen for the springs that result in similar deflections of
the spring supported girder, under full length traffic loading of the main span, as in the
reference bridge model under the same loading condition. In case of an upward deflection
of the side span, the girder will not be resisted by the hangers, in that case the
spring stiffness k = 0 in the side span.

20
Overspannend staal, Rotterdam: Stichting Kennisoverdracht SG, Deel 3: Construeren B, 1996.

46
 Self-anchored suspension bridges: Part II

As mentioned earlier only a traffic loading over the entire bridge length is considered, as
shown in 2.2.3 this is decisive for buckling of the stiffening girder. Figure 61 presents a graph
of the relation between the Euler buckling force of the main span of the bridge deck and the
stiffness of the deck (moment of inertia of the box girder ranging from a girder height of
1500-3000 mm, representing a range of deck slenderness λ from 1/100 to 1/50) regarding the
two models. The deck slenderness is defined as λ = construction depth h of the girder/main
span length l.
Also the buckling force of the side span is plotted in the graph, which is governing over the
buckling force of the main span (see §2.2.3).
Euler buckling force Ncr main span; Additional F model
Ncr main span; Spring model
N Ed deck
1600000 Ncr side span; Additional F model

1400000
Euler buckling force [kN]

1200000

1000000

800000

600000

400000

200000

0
0 1 2 3 4 5
Moment of inertia Iy deck [m^4]

=Reference model,
λ = 1/100 λ = 1/70 λ = 1/50

See also Figure 50: NEd = 32420 kN

Ncr;main span = 851867 kN

Boundaries of considered deck slenderness λ =1/100 – 1/50

Figure 61 Euler buckling force as a function of the stiffness of the deck

Both approaches display a similar development with respect to the Euler buckling force of the
main span of the stiffening girder. The Euler buckling is in all cases well above the acting
normal force in the deck NEd. So also in case of a more slender deck with a
slenderness λ = 1/100, the resistance against buckling is still significant.

47
 Self-anchored suspension bridges: Part II

The buckling behaviour of the main and side span presented in Figure 61 can also be
expressed in the so called n-values, see Table 8 n-values for main and side span based on the
computed bridge model.

Table 8 n-values for main and side span related to the moment of inertia of the box girder
n-value n-value
Iy;girder [m^4] main span side span
1,11 15,2 12,7
1,26 17,0 13,3
1,42 18,9 13,9
1,59 20,6 14,6
1,77 22,4 15,3
1,96 24,3 16,0
2,17 26,2 16,3 = reference design
2,38 28,2 17,5
2,60 30,2 18,4
2,83 32,3 19,2
3,10 34,5 20,1
3,32 36,8 21,0
3,58 39,1 21,9
3,85 41,5 22,9
4,13 44,1 24,3
4,42 46,6 24,9

These n-values also clearly show that the side span exhibits more geometrical non-linear
effect, indicating that the side span is decisive for the buckling stability of the stiffening girder
and that the geometrical non-linearity’s decrease for both the main and the side span when the
stiffness of the deck is increasing.

Conclusion regarding the buckling stability

Figure 61 shows that the buckling stability of the girder requires attention in the design
process of a self-anchored suspension bridge. But a girder with a slenderness to about λ =
1/100 is expected to have enough resistance against buckling. Extrapolating this graph for
even more slender girders, meaning a higher slenderness λ =1/100..1/150 and further, a
limitation is expected with respect to the buckling resistance. Buckling than becomes critical
and a limiting design factor.
These results confirm the findings presented in the literature survey of this study which
showed that the slenderness of stiffening girders in existing self-anchored suspension bridges
is limited to about λ = 1/100 (e.g. Konohana bridge, Japan). Compare this to the applied
stiffening girders in conventional suspension bridges where no buckling risk of the girder is
present, here girder slenderness of λ = 1/200 – 1/300 are common.

48
 Self-anchored suspension bridges: Part II

3.1.3 Number of hinges

As mentioned earlier in the literature survey, the stiffening girder can either be executed as a
continuous girder or as a hinged girder. A continuous girder has no hinges along the length
and a hinged girder has two hinges, one at each support at the pylon, see Figure 62.

Figure 62 Hinged and unhinged girder

Figure 63 shows clearly that a continuous girder applied in the reference model. with no
hinges. is much stiffer. For deflection and frequency behaviour of the first torsional motion, a
continuous girder is much stiffer.

Figure 63 Maximum deflection in the girder

Figure 64 Natural frequencies

From now on only a continuous girder is chosen. Based on the given results and bridges
already built like the Konohana and Yeoungjong Grand bridge, a continuous girder is chosen.
Results have shown that a continuous girder is displays smaller deflections and offers a better
resistance for a torsional frequency motion.

49
 Self-anchored suspension bridges: Part II

3.2 Main Cable influence

Axial stiffness EA of the main cable is determined by two factors; the modulus of elasticity E
that changes with the different cable types, and the cross sectional area Acable of the cable.
The influence of the hangers is left out of consideration in this part of the research, it is
assumed to have little influence on the global strength and stiffness of the bridge.

3.2.1 Cable type

Four cable types can be applied in cable supported bridges:
• Parallel wires E = 205000 N/mm2
• Parallel strands E = 190000 N/mm2
• Full locked coil E = 150000 N/mm2
• Cable spiral strand E = 140000 N/mm2
The axial stiffness EA of the main cable is reduced with approximately 25% when choosing
for locked or spiral strands compared to parallel wires. A lower axial stiffness of the cable
means that larger bending moments will act in the girder, see Figure 65.The girder carries a
larger part of the total bending moment when the axial stiffness of the main cable is reduced,
see figure 43.
Main span % of total moment carried by deck
Bending moments
300000 Support 25
Bending moment [kNm]

250000
% o f to ta l m o m e n t

20
200000
15
150000
10
100000
5
50000

0 0
130000 150000 170000 190000 210000 130000 150000 170000 190000 210000
E [N/mm^2] E [N/mm^2]

Figure 65 Maximum bending moment in the girder

Figure 66 Contribution of the girder to bending moment

Compared to the main cable when it is composed of spiral strands, the parallel wired cable
reduces the deflection with approximately 14 %. Little effects are visible on the frequency
behaviour.
main span 1e bending
Deflection Frequencies
support 1e torsional
450
6
400
350 5
Deflection [mm]

Frequencie [Hz]

300 4
250
3
200
150 2
100
1
50
0 0
130000 140000 150000 160000 170000 180000 190000 200000 210000 130000 140000 150000 160000 170000 180000 190000 200000 210000
E cable [N/mm^2] E cable [N/mm^2]

Figure 67 Maximum deflection in the girder

Figure 68 Natural frequencies

50
 Self-anchored suspension bridges: Part II

3.2.2 Cable cross sectional area

Once a cable type is chosen, the axial stiffness of the main cable can be altered by means of
the cross sectional area Acable. The next table shows the cable diameters with the
corresponding effective area that are applied.

Table 9 Cable area

diameter
[mm] 100 120 140 150 160* 170 180 190 200 210 220 230 240
A eff.
main
cable
[mm^2] 7854 11310 15394 17671 20106 22698 25447 28353 31416 34636 38013 41548 45239

*=Reference model

Increasing the cable’s diameter displays significant effects on the maximum bending moment.
Increasing the diameter to from 160 mm to 240 mm, the bending moment in the deck reduces
to nearly zero, see Figure 69. This means that nearly a 100% of the total moment is carried by
the main cable, which is clearly visible in figure 47.

Main span
Bending moments % of total moment carried by deck
Support
450000
400000 40
Bending moment [knm]

350000 35
% of total moment

300000 30

250000 25
20
200000
15
150000
10
100000
5
50000
0
0
0 10000 20000 30000 40000 50000
0 10000 20000 30000 40000 50000
A eff main cable [mm^2]
A eff main cable [mm^2]

Figure 69 Maximum bending moment in the girder

Figure 70 Contribution of the girder to bending moment

Increasing the axial stiffness of the cable has favourable effects for the global stiffness, the
girder and pylon deflection both reduce. With respect to the frequency behaviour, the
increasing stiffness results in higher frequencies. Although there is a point where an
increment of the cable results in a decreasing torsional frequency. The reason for this can be
that the self weight of the cable rules out the stiffening effect of the cable.
Deflection girder 1e bending
Frequencies
600 pylon 1e torsional

6
500
5
400
deflection

Frequency [Hz]

4
300
3
200
2
100
1
0
0
0 10000 20000 30000 40000 50000 0 10000 20000 30000 40000 50000
A cable [mm^2] A cable [mm^2]

Figure 71 Maximum deflection in the girder

Figure 72 Natural frequencies

51
 Self-anchored suspension bridges: Part II

N deck
Reaction forces
Vertical reaction at end
50000 support

A larger contribution of the main cable to 45000

40000
the total bending moment leads to much 35000

Force [kN]
higher reaction forces. The normal force in 30000
25000
the deck and the vertical reaction force 20000
15000
increase significantly, see Figure 73. 10000
5000
0
0 10000 20000 30000 40000 50000
A cable [mm^2]

Figure 73 Reaction forces

3.3 Sag influence

As mentioned before the most common sag to span ratio for self-anchored suspension bridges
are 1/5 to 1/9. With an increasing sag ratio, the bending moment in girder reduces with
approximately 50%, see Figure 74. The cable carries a larger part of the total bending
moment when the sag ratio is increased, see figure 52.
main span
Bending moments % of total moment carried by girder
support
400000 35
350000
30
Bending moment [kNm]

300000
% of total moment

25
250000
20
200000
15
150000
10
100000
5
50000
0
0
0,1 0,12 0,14 0,16 0,18 0,2 0,22
0,1 0,12 0,14 0,16 0,18 0,2 0,22
Sag/L
Sag/L

Figure 74 Maximum bending moment in the girder

Figure 75 Contribution of the girder to bending moment

A larger sag ratio increases the stiffness of the bridge, the deflection decrease with about 28%.
The deflection of the pylon increases because the height of the pylon increases with a larger
sag ratio. So with an unchanged bending stiffness of the pylon, its deflection will increase
when the height is increased, see Figure 76.
The first torsional frequency motion is coupled with a longitudinal deflection of the pylon,
therefore the pylon’s stiffness will have effect on the torsional stiffness of the bridge. A larger
sag ratio means that the pylon becomes more flexible, so a larger sag ratio decreases the
torsional stiffness of the bridge which is visible in figure 54.
main span 1e bending
Deflection Freqencies
pylon 1e torsion

600 7

500 6
Deflection [mm]

5
Frequency [Hz]

400
4
300
3
200
2
100
1
0 0
0,1 0,12 0,14 0,16 0,18 0,2 0,22 0,1 0,12 0,14 0,16 0,18 0,2 0,22
Sag/L Sag/L

Figure 76 Maximum deflection in the girder

Figure 77 Natural frequencies

52
 Self-anchored suspension bridges: Part II

Regarding the reaction forces a larger sag Reaction forces N deck

ratio decreases the contribution of the cable 45000 Vertical reaction at end
support
to the total bending moment. The normal 40000
35000
force in the main cable decreases with a 30000

Force [kN]
larger sag ratio and so does the 25000
compression in force in the girder. 20000

Only the vertical reaction force at the end 15000

10000
support increases because the vertical 5000
component of the main cable’s normal 0
force increases with a larger pylon height. 0,1 0,12 0,14 0,16
Sag/L
0,18 0,2 0,22

Figure 78 Reaction forces

3.4 Pylon influence

This part presents the influence of the bending stiffness of the pylon in longitudinal direction
of the bridge. Looking at the presented results in Figure 79, no significant developments in
bending moments as function of the stiffness can be seen.
Bending moments main span
250000 % of moment carred by girder
support

20
Bendingmoment [kNm]

200000 18
16
% of total moment

150000 14
12
10
100000 8
6
50000 4
2
0
0 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9
0 0,2 0,4 0,6 0,8 1 Moment of inertia Iy [m^4]
Moment of inertia Iy [m^4]

Figure 79 Maximum bending moment in the girder

Figure 80 Contribution of the girder to bending moment

girder 1e bending
Deflection Frequencies
pylon 1e torsional

400 6
350
5
300
Deflection [mm]

Frequency [Hz]

4
250
200 3
150
2
100
1
50
0 0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9
Moment of inertia Iy [m^4] Moment of inertia Iy [m^4]

Figure 81 Maximum deflection in the girder

Figure 82 Natural frequencies

Only the frequency of the first torsional motion increases. A torsional motion of the girder
exhibits with longitudinal motion of the pylon, therefore a stiffer pylon has a positive effect
on the torsional stiffness of the bridge. In that way the pylon offers more resistance against a
torsional motion of the girder, see figure 59.

53
 Self-anchored suspension bridges: Part II

3.5 Asymmetric traffic loading

Until now only symmetric loading has been considered. This part evaluates the effects of an
asymmetric position of the traffic loading on the force distribution in the bridge. The traffic
loading is combined with the self weight of the bridge and other permanent loads on the
bridge including the pretensioning of the main cable.

3.5.1 Full loaded deck

First of a full loaded deck is considered which results in a certain eccentricity, see Figure 24.
This full loading conditions is in the model applied as line load of 131,5 kN/m with an
eccentricity of 1.48 metres. The effects of such a load condition is given in Table 10.
Table 10 Deck full loaded eccentricity

Traffic full
length
Symmetric Asymmetric
N deck [kN] 32420 32422
H cable 1 [kN] 16452 16743
H cable 2 [kN] 16542 16340
δ cable plane 1
main span[mm] 361 372
δ cable plane 2
main span[mm] 361 349
δ pylon [mm] 91 93
My main [kNm] 152065 152061
My sup [kNm] 193096 193093
My pylon [kNm] 21044 2152
M torsion kNm] 0 16194

Main effects of taking into account the asymmetric effect of a full loaded deck is a torsional
moment in the box girder at the support location at the pylon. Also the deck shows a rotation
of about 0.04 degrees.

3.5.2 Half loaded deck

The deck is in transverse direction loaded on one half of the bridge deck, see Figure 83

Figure 83 Half loaded bridge deck in transverse direction

This half loaded deck will be applied on the bridge model as a resulting line load of
75.5 kN/m with an eccentricity of 10.1 metres.
Three asymmetric loading positions will be considered (side span loading is not the decisive
load case concerning bending moments in-, and deflections of- the stiffening girder, see 1.3.3,
so no asymmetric loading of the side span is considered here). The traffic loading is combined
with the self weight of the bridge and other permanent loads on the bridge including the
pretensioning of the main cable.

54
 Self-anchored suspension bridges: Part II

• One sided full length traffic loading

Table 11 Asymmetric loading traffic full length
Traffic full
length
Symmetric Asymmetric
N deck [kN] 32420 28754
H cable 1 [kN] 16452 15310
H cable 2 [kN] 16542 13915
δ cable plane 1
main span[mm] 361 247
δ cable plane 2
main span[mm] 361 166
δ pylon [mm] 91 44
My main [kNm] 152065 99166
My sup [kNm] 193096 126525
My pylon [kNm] 21044 12555
M torsion kNm] 0 56067
Figure 84 One sided full length loading

Figure 84 presents the considered loading scheme of an asymmetric traffic load over the full
length of the bridge. An asymmetric loading has the effect that one cable plane is loaded more
than the other cable plane. The normal force in the loaded cable plane is therefore higher than
the normal force in the unloaded cable plane but still smaller than in full loading symmetric
condition.
The bending moments in the girder as well as in the pylon base, are in case of asymmetric
loading much less than in full loading symmetric condition. Also the deflections are smaller
than in full loading condition. One of the most adverse effect of asymmetric loading is a
torsional moment in the deck.
A second effect of asymmetric loading is the rotation of the deck. Table 11 shows a difference
is visible in the deformation of the deck on the loaded side and unloaded side. This results in a
very limited rotation of the deck of about 0.13 degrees.

• One sided mid span traffic loading

Table 12 Asymmetric loading traffic mid span
Traffic mid
span
Symmetric Asymmetric
N deck [kN] 32825 28986
H cable 1 [kN] 16820 15400
H cable 2 [kN] 16820 14154
δ cable plane
1[mm] 422 283
δ cable plane
2[mm] 422 202
δ pylon [mm] 121 58
My main [kNm] 164378 106269
My sup [kNm] 163167 109369
My pylon [kNm] 26040 14955
M torsion 0 57799

Figure 85 One side mid span loaded

55
 Self-anchored suspension bridges: Part II

Figure 85 presents the considered loading scheme of an asymmetric traffic load over the mid
span of the bridge. Again a redistribution of forces is visible, the loaded cable plane carries a
larger part of the vertical traffic loading and displays a higher normal force. The most adverse
effects are encountered in the torsional moment in the deck.
The rotation of the deck is due to the difference in deflection of the loaded and unloaded side
of the deck is about 0.13 degrees.

• Alternate full length traffic loading

Table 13 Alternate loading traffic full length
Traffic full
length
Symmetric Alternate
N deck [kN] 32420 28754
H cable 1 [kN] 16452 15169
H cable 2 [kN] 16542 14056
δ cable plane
1[mm] [mm] 361 250
δ cable plane
2[mm] 361 165
δ pylon [mm] 91 39
My main [kNm] 152065 99155
My sup [kNm] 193096 126529
My pylon [kNm] 21044 11631
M torsion 0 59695

Figure 86 Alternate loading full length

Figure 86 presents the considered loading scheme of an alternating traffic load over the entire
length of the bridge. This means that on the mid span the traffic load is situated one the
opposite cable plan than the traffic load on the side spans. Table 13 presents the member
forces of this loading conditions. A comparable redistribution of cable forces is visible and
torsional moment in the deck.
The rotation of the deck is due to the difference in deflection of the loaded and unloaded side
of the deck is about 0.14 degrees, also very limited.

Conclusion regarding asymmetrical loading

A consideration of an asymmetrical application of the vertical traffic loading reveals that
more attention has to go out for the design of the girder at the support location at pylon. At
this location occurs the largest torsional moment in the girder.
The resulting stresses caused by torsion can be distincted in the St. Venant shear flow and
restrained warping stresses:
-St. Venant shear stress (maximum occurs in the web because the plate thickness is the
smaller than in the flanges):
Mt 59695 * 106
τ= = = 30.7 N/mm2
2 Am t1 2 * 35600 * 2100 * 13

-Restrained warping stresses:

3 * ΔM t 3 * 59695 * 10 6
σw = = =52.9 N/mm2
It * A 12
5.33 * 10 * 2.15 * 10 6

56
 Self-anchored suspension bridges: Part II

These stresses require attention for the detailed design of the support conditions of the box
girder at the pylon location.
Asymmetrical loading causes also a rotation of the deck but is limited to about 0.14 degrees
which is assumed to be very limited and of acceptable level.

3.6 Conclusion results of the parameter study

This part has shown that many bridge parameters have a large influence on the force
distribution and deflections in a self-anchored suspension bridge.
It has been shown that the mechanical properties of the girder, cable and sag ratio govern the
design. These bridge parameters have large influence on the deflections and the maximum
bending moment in the stiffening girder. The stiffness of the pylon shows little influence on
the global behaviour.
Regarding the force distribution and deflection it is favourable to consider:
-A slender stiffening girder, to reduce the maximum bending moment in the girder. It
has been shown that even though the slenderness is reduced the stress level in the box
remain quite constant, see Figure 58.
-A stiff main cable, to increase the global stiffness of the bridge and to reduce the
maximum bending moment in the girder.
-A high sag to span ratio, to reduce the normal force in the deck and the maximum
bending moment in the deck.

3.7 Evaluation of the results

This part contains an evaluation and analyses of the results of the parameter study and
literature survey. It will give a motivation of the choices that are made to optimize the
reference design. The optimization is made with respect to the general performance of the
bridge but also with respect to the next topic of this research: increasing the span length.

3.7.1 Criteria
• Distribution of forces:
o Bending moments in girder
The sagging moment in the main span and hogging moment at support near
pylons, is determining the design of the girder. Designing for bending moments is
always more difficult and material consuming than for designing a cross section
for normal force only. So reducing bending moments in the girder is favourable.
o Normal force in cable.
The level of stresses determines the design of the cable and therefore the axial
stiffness.
• Stress conditions in the box girder:
The slenderness of the box girder influences the stiffness and therefore the bending
moment in the girder. Also the section modulus is determined by the slenderness which
determines the stress level in the cross section. Figure 58 shows the constant development
in stress with a decreasing slenderness of the deck. This development justifies the
possibility for choosing a more slender deck.

57
 Self-anchored suspension bridges: Part II

• Deflection of girder and pylon

The deflection of the pylon and girder have to be within limits to meet certain stiffness
criteria.
• Frequencies
It is important to consider the level of frequencies. Low torsional frequencies means that
the bridge structure becomes more susceptible for flutter risk. Also the ratio between the
first bending and the first torsional frequency illustrates the risk for flutter. A ratio close to
one is not desirable, in general a ratio of 2 or more is advisable.
The results show that the ratios of the 1st bending to 1st torsional frequencies are in all
cases well above 2. So in this stadium this criterion plays a secondary role.
• Girder stability
The compression loaded stiffening girder requires enough resistance against buckling. The
bending stiffness is dominating the buckling properties of the girder.
• Erection
For the erection of the main cable two methods are possible. Aerial spinning of cable
producing parallel wired cables and erection of a prefabricated locked coil or spiral strand
main cable. Especially with large span the last mentioned method becomes impossible.
• Material use/costs
The required mechanical properties of the main cable and girder can be translated to a
certain material use for the cross section. Designing cross section for bending moments is
always more material consuming that for normal forces. For the cross section of the girder
it is profitable to reduce the bending moments that act in the girder.
• Reaction forces
Regarding the reaction forces on the end support and compression force in deck it is
desirable to achieve an acceptable level in these forces.
• Examples from practice
Examples like the Konohana Bridge, Yeongjong Grand Bridge and Duisburg Bridge, with
a main span of respectively 300 metres, 300 metres and 285 metres, offer experience
which can contribute to certain choices in the design phase of a self-anchored suspension
bridge.

Optimization of the reference design is achieved by choosing mechanical requirements of the

girder, pylon, and cable, and a geometrical sag over main span length ratio which have the
most favourable effect on the before mentioned criteria.

3.7.2 Conclusions for optimized reference design

This part presents the chosen dimensions for all the structural elements and chosen ratio for
the optimized reference model. This model will be used for the next research topic in this
study, namely the exploration of the span length possibilities.

• Height of stiffening girder

Its is favourable for reducing the maximum bending, to choose a slender girder. Also for
the design of the cross section it results in less material use, larger bending moments
require heavier plate thicknesses and a larger amount of stiffeners. §3.1 has shown that
also the stress level does not increase and remain on an acceptable level when a more
slender girder is chosen. Therefore a slender girder is chosen of 1/95, this is similar to the
box girder of the Konohana bridge in Japan. The presented stability analyses in §3.1.2
shows that a more slender girder still possesses enough resistance against buckling.

58
 Self-anchored suspension bridges: Part II

Although the section modulus is hereby greatly reduced, as shown in §3.1.1 the stress
level in the box girder remain on an acceptable level.
• Sag to span ratio
§3.3 clearly shows the importance of the sag ratio on the moment distribution. A high sag
ratio showed to be favourable for the bending moment distribution, deflections of the deck
and the normal force in the deck. Examples like the Konohana bridge and the Yeongjong
grand bridge show that such pylon heights are feasible. Therefore a sag ratio of 1/5 is
chosen.
• Cross section pylon
The results in 3.4 show clearly that the influence of the pylon stiffness, in longitudinal
direction of the bridge on the global behaviour, is negligible. So no changes are made in
the mechanical properties of the pylon.
• Cable type
Although the different cable show a large influence on moment distribution, the erection
method is decisive for the cable type choice. For large spans it becomes impossible to use
prefabricated locked coil cables and spiral strand cables. With large spans only fabricating
a cable by means of aerial spinning is possible. Therefore parallel wires are chosen with a
modulus of elasticity E = 205000 N/mm2.
• Cross section cable axial (stiffness)
The cable cross sectional area can be altered to meet static strength criteria of the cable
and girder and the deflection criteria of the girder and pylon. The axial stiffness EA of the
cable is an important parameter to influence the force distribution and stiffness of the
bridge. Increasing EA of the cable increases the stiffness and reduces the bending
moments in the girder significantly.
For the optimized reference model a cable diameter of 210 mm is chosen, in that way it
meets the static strength and stiffness criteria at the main span:
σbottom flange 139,6 N/mm
2
< 200 N/mm2 the design stress at midspan

deflection midspan 422 mm < 430 mm maximum deflection of midspan

displacement pylon 132mm < 170 mm maximum deflection of pylon

Unity check stability according to Eurocode:
N Ed β m M y , Ed
+ = 0.515 ≤ 0.9 Satisfies. For calculation details see Appendix 5b
χN cr M y , Rk
Stability check Optimized model.

This optimized model is used to investigate the increasing span length of the bridge on the
structural behaviour. Central issue is monitoring the required mechanical properties of the
stiffening girder by an increment of the span length.

59
 Self-anchored suspension bridges: Part II

Conclusion optimized model

The optimized model for further research contains the following properties regarding
dimensions of the bridge components, see Table 14.
Table 14 Member dimensions optimized model
Girder t_flange_top_eq 40 mm
t_flange_bottom_eq 20 mm
t_web_eq 15 mm
heigth 1600 mm
width 35600 mm

Main cable diameter 210 mm

Hangers Diameter 55 mm

Pylon width 2000 mm

depth 2000 mm
Total height 50000 mm

With the above mentioned properties the bridge satisfies the design criteria concerning static
strength, stiffness and stability. But now the reduction of the bending moments in the girder is
quite significant, see Figure 87. The bending moments in the girder are nearly reduced by 50
percent. Making it much more efficient and cheaper to design and produce.
Reference model
Bendig moments in girder
Optimized model

250000

200000
Bending moment [kNm]

150000

100000
50000

0
-50000 0 50 100 150 200 250 300

-100000
-150000

-200000
Position on bridge [m]

Figure 87 Bending moments along the girder in reference model and optimized model

60
 Self-anchored suspension bridges: Part II

4 Increasing span length

Now the influence of an increasing span length will be investigated. This chapter tries to find
a certain span limit for the self-anchored suspension bridge.
One of the main concerns with an increasing span is the design of the stiffening girder. The
increasing axial compressive force, bending moments and second order effects are
determining factors in the design of the stiffening girder. Other important aspect is the
erection phase of a self-anchored suspension bridge. The distance between the temporary
supports can easily govern the girder design regarding the required bending stiffness. Much
attention is needed for the design of a stiffening girder to meet the different requirements in
finished- and erection phase of the bridge.
In this part of the research the central question :
What are the effects of an increasing main span length on the required mechanical properties
of the stiffening girder?
So the scope is trying to find a limit in span possibilities related to mechanical required
properties of the stiffening girder. The considered required properties of the stiffening girder
will be the bending stiffness EIgirder and the cross sectional area Abox.

4.1 Basic assumptions

To be able to increase the span in the reference model and analyzing the effect on the required
mechanical properties for the stiffening girder, all other dimensional and mechanical
properties should stay fixed. Only then a fair comparison of the results is allowed. Increasing
the span in the reference model is done by means of several scaling factors for cable sag, side
span, hanger distance, cable diameter, girder slenderness and pylon stiffness, which are fixed
in the following ratio:
• sag to span ratio f/l = 1/5
This ratio determines the horizontal component of the cable force and therefore the
compression force in the stiffening girder. Keeping this ratio fixed enables to discover the
influence of the increment of the compression force on the behaviour of the stiffening
girder.
• main span to side span ratio l/l1 =2.4
This ratio is kept fixed to rule out any influence of main span to side span ratio on the
behaviour of the stiffening girder.
• diameter of main cable to main span ratio d/l = 210/150 = 1.4
Theory shows that an increment of the span with l gives an exponential increase of the
horizontal cable component H. To maintain the same level of stresses in the cable, the
cross sectional area A should therefore be increased. Cross sectional area of a circular
cable is proportional to the square of the diameter. So an increment of the span length with
l gives a linear increase of cable diameter d (self-weight of the cable per unit of length
remains constant under a fixed sag to span ratio, when diameter is unchanged). A fixed
diameter to span ratio keeps the level of stresses due to self weight effects constant.
• hanger distance to main span ratio = 1/24
This ratio is kept fixed to rule out any influence of the hanger distance on the behaviour of
the stiffening girder. The cross sectional area of the hangers is linearly scaled with an
increasing core to cores distance of the hangers.
• deck height to main span ratio (deck slenderness) h/l = 1/95
Based on data in the literature survey, a realistic deck slenderness is chosen. As mentioned
earlier a deck slenderness of 1/95*L is chosen.

61
 Self-anchored suspension bridges: Part II

• The pylon height increases with a larger span

because the sag to span ratio is kept fixed. An
increasing pylon height requires more stiffness.
Therefore the longitudinal stiffness is scaled
according to the next schematisation and
formula:

The ratio l3/Iy pylon is kept fixed to have comparable stiffnesses of the pylon in the
longitudinal direction of the bridge for increasing span lengths.
• Step size 50 metres. The span length is increased with step sizes of 50 metres. With a
step size of 50 metres the span length is increased up to 500 metres. For each step an
evaluation is made on static strength, stiffness and stability criteria.
Based on the literature survey it is expected that a span length of more than 300 metres
is challenging for the self-anchored suspension bridge. Earlier presented research to
the buckling stability of the girder showed that this can be of importance. Secondly is
the erection phase. The distance between the temporary supports can also be a decisive
factor for the required mechanical properties of the girder and will we explored in this
part.
• Vertical clearance: The vertical clearance under the bridge deck is kept fixed to
15 metres. The pylon height under the bridge deck is therefore also fixed on 15 metres.

Load condition
Only load case 1(traffic over the entire length of the bridge) in combination with self weight,
permanent loading and pretensioning of the main cable is considered in this part of the study.

Monitoring developments
Increasing the span length of the bridge will cause several effects on static strength and
stiffness. Several effects are monitored to analyse the effects and to verify if the before
mentioned scaling assumptions are applicable and valid.

-Stresses in cable, girder and pylon

Ratio stress caused by bending to stress caused by normal force
-Second order effects
-deformation of girder and pylon
-reaction forces: vertical reaction force on end support and normal force in the deck of
the bridge.
-material use of cable, deck, pylon
-effects on erection of the bridge (number of temporary supports etc.)

62
 Self-anchored suspension bridges: Part II

4.2 Effects of increasing the span by scaling

Introduction
The following effects are visible when the span length of the optimized bridge model is
increased to 500 metres and bridge components are scaled according to the before mentioned
method, see Table 15 for the main results of member forces and deflections caused by self
weight, permanent loads and traffic over the full length of the bridge.
Table 15 Main results of increasing span
1st
L Ndeck Hcable Rz δmain pylon My;main My;support 1st fb 1st ft ftransverse
150 37682 19136 7679 336 87 83959 111176 0,72 4,43 5,21
200 54028 27361 11306 410 108 100200 139068 0,6 3,4 3,15
250 70944 35857 15048 468 127 115276 168634 0,52 2,81 2,1
300 88679 44773 18966 519 144 128073 198965 0,46 2,35 1,5
350 107413 54178 23091 567 161 139339 232275 0,4 2 1,12
400 127275 64153 27453 613 178 150087 270473 0,36 1,75 0,87
450 148451 74784 32089 660 196 159235 314120 0,32 1,53 0,7
500 171108 86155 37037 706 214 167839 365235 0,29 1,36 0,57

Horizontal equilibrium shows that 2*Hcable = Ndeck, but Table 15 shows a little deviation
caused by the fact a very small part of the horizontal component of the cable force Hcable is
resist as bending in the pylon’s base.

A quick stiffness check (δmax allowable = 1/350*Lmain span) reveals that the scaled bridges up to
500 metres performs very constant regarding the
maximum allowable deflection of the main span. Unity check on deflection main span

Figure 88 shows the constant performance 1,00

Unity check [max δ/allowable δ]

regarding deflection of the main span, the unity 0,90

0,80

check varies between 0.9-0.98. In this case the 0,70

0,60

unity check regarding stiffness is determined by: 0,50

0,40
UC = δmax main span / δmax allowable 0,30
0,20
So on stiffness criteria, increasing the main span 0,10

length up to 500 metres by means of the scaling 0,00

0 100 200 300 400 500

factors, satisfies and displays a constant Span length [m]

performance.
Figure 88 Development on Unity check deflection main span

Because the scaling of the bridge model up to 500m proved to be very constant regarding the
global stiffness of the bridge, no adjustments are in made in the previous mentioned scaling
assumptions. Now a comparison and analyses is made regarding the developments on all
other important design criteria and design aspects with an increasing main span:
• Static strength
• Stability of the stiffening girder
• Frequency behaviour
• Reaction forces
• Material use
• Effects on erection of the bridge

63
 Self-anchored suspension bridges: Part II

4.2.1 Developments on static strength

To give a total view on strength criteria, the development of the level of the stresses in the box
girder, the main cable, the hangers and pylon are presented in this paragraph.

Stresses in box girder

A fixed girder slenderness of λ = 1/95 is chosen for the several span lengths that are
considered. The mechanical properties of the box girder for each span length are given in
Table 16.
Table 16 Girder properties for the considered span lengths
Main Slenderness height Atotal
span [m] λ [m] [m] Iy [m4] Iz [m4] It [m4]
150 1/95 1,6 2,18 1,26 241 3,15
200 1/95 2,1 2,2 2,17 246 5,33
250 1/95 2,63 2,22 3,4 251 8,2
300 1/95 3,16 2,23 4,9 256 11,62
350 1/95 3,68 2,25 6,68 261 15,52
400 1/95 4,21 2,26 8,78 266 20,07
450 1/95 4,74 2,28 11,1 271 24,77
500 1/95 5,26 2,29 13,77 276 30,09

In order to determine cross sectional requirements (e.g. plate thicknesses) of a box girder in
the design process, the compressive and tensile stresses that are present in the top and bottom
flanges are calculated. The stresses are caused by two components:
• Bending stresses, caused by the global bending moment in the girder.
• Normal stresses, caused by introduction of the compressive force in the girder by the
main cable.
The sum of these two components give the total stress. The actual stresses in the top- and
bottom flanges are presented in Figure 89 and Figure 90. For a total overview method of the
calculation of these stresses, reference is made to Appendix 7 Stress calculation in reference
model.
Stresses in compression flanges compression top flange main span Stresses in tensile flanges tensile bottom flange main span

compression bottom flange support tensile top flange support

700,00
400,00
compression bottom flange side span tensile top flange side span
600,00 350,00
Compression stress [N/mm^2]

Tensile stress [N/mm^2]

500,00 300,00
250,00
400,00
200,00
300,00
150,00
200,00 100,00

100,00 50,00

0,00
0,00
0 100 200 300 400 500 600
0 100 200 300 400 500 600 -50,00
Main span [m] Main span [m]

Figure 89 Stresses in compression flanges

Figure 90 Stresses in tensile flanges

64
 Self-anchored suspension bridges: Part II

The stresses in the compression flanges of the side and main span of the girder have a nearly
constant development according to Figure 89. This indicates that the chosen girder
dimensions for each considered span length are properly chosen.
The significant decrease of the compression stresses at the support location is caused by the
fact that the hogging moment (but also the sagging moment at main span) does not increase
that rapidly with an increase of the main span. In case of an unsupported main span, for
instance a simply supported beam, an increase of the main span length l will result in an
increase of the maximum sagging moment by l2. As mentioned earlier the girder of a
suspension bridge is continuously stiff Sagging moment main span
supported by the hangers, so increasing Development of the max bending moment hogging moment support

the main span of the bridge with a 1400000

increase of moment by l^2

length l does not result in an increase of 1200000

Bending moment [kNm]

2
the global bending moments of l but 1000000

much less, see Figure 91 where the 800000

dotted line represents a development of 600000

the bending moment with an increase 400000

2
by l and the actual development of the 200000

0
maximum bending moments in the 0 100 200 300 400 500 600

continuous spring supported deck in Main span [m]

main span of the bridge.

Figure 91 Development of the maximum bending moment in the girder

Figure 90 shows that also the stresses in the tensile flanges of the box girder display an overall
decrease on all locations. To explain this development an analyses is made on the stress
calculation. The total stress in a flange is calculated by:
N My
σ top = − Ed −
Abox Weff ;top
N Ed My
σ bottom = − +
Abox Weff ;bottom
Besides the confined development of the bending moments other causes for this decrease in
tensile stresses in the flanges of the box girder are:
• The chosen girder properties for the considered span length up to 500 metres. For each
span the same slenderness λ = 1/95 of the box girder is chosen. So the height of the
box girder is with each increment of the main span linearly increased. Section
properties like the moment of inertia Iy and the section modulus Weff (for calculation
the stresses) increase also. The moment of inertia of a box girder is determined by:
I y ;total = I y ;eigen + I y ;steiner .
In which Isteiner is the dominating component which is proportionate to the square of
height h of the girder (see Table 16). This indicates that the section modulus of
Iy
elasticity Weff increases approximately linearly according to: Weff = . This
ey
combined with the fact that the bending moments do not increase that significant, the
bending stresses decrease with an increasing span length, see also Figure 90.

65
 Self-anchored suspension bridges: Part II

• The contribution of the normal force to the total stresses in box girder.
A graph is presented in Figure 92 Contribution of normal stresses compared to bending bending stresses
normal stresses
which clearly shows that with an 120
stresses

increasing span the stress caused by 100

Total stress [N/mm^2]

the compressive force in the girder
80
become dominant over the bending
60
stresses. Therefore a decreasing
development is visible in tensile 40

stresses presented in Figure 90. 20

Compressive stresses become 0

dominant in the cross section of the 150 200 250 300 350
Main span [m]
400 450 500

box girder for an increasing span.

Figure 92 Contribution of the normal compressive stress
in the top flange in the main span

Even a point can be reached where compression stresses can occur in the normally tensile
bottom flange of the box girder in the main span, see Figure 90, at a main span of 400 metres.
This will have effect on the design of the bottom flange in the mid of the main span, if
compression stresses occur also here, than local instabilities have to be checked and it is likely
that more stiffeners have to be applied (as is the case for the compressive bottom flange at
support location at the pylon).

Overall conclusion is that the level of stresses are of acceptable levels, when the span is
increased, and that the normal stresses become dominant over the bending stresses.

Stresses in main cable and hangers

The applied main cable and hanger properties for each span length are given in Table 17.
Table 17 Main cable and hanger properties
Main dcable Acable dhanger Ahanger Tensile stresses in main cable and hangers Tensile stress in main cable

span [m] [mm] [mm^2] [mm] [mm^2] 800 Tensile stress in hangers

150 210 34636 55 2376 700

T e n s ile s tr e s s le v e l

600
200 280 61575 64 3217
[N /m m ^ 2 ]

500
250 350 96211 71 3959
400
300 420 138544 78 4778 300
350 490 188574 84 5542 200
400 560 246301 90 6362 100

450 630 311725 95 7088 0

0 100 200 300 400 500 600
500 700 384845 100 7854 Main span [m]

Figure 93 Stress level in main cable and hangers

The level of the tensile stress in the decisive hanger stays approximately on the same level but
for the main cable a decreasing stress level is clearly visible in Figure 93. This result indicates
that stiffness is the governing design criteria for suspension bridges, even for relative short
spans up to 500 metres. The amount of material required for the main cable becomes with an
increasing span length less efficient on strength, the additional required main cable area is
needed to satisfy stiffness criteria, see also Figure 87.

66
 Self-anchored suspension bridges: Part II

Stresses in pylon
Without any provisions, an increasing main span would increase the bending moments at the
pylon drastically and can therefore become critical, see Figure 94.
In practice the bending moments can artificially
be reduced to zero by giving the pylon an Bening moment at pylon basement

outward displacement to counteract the inward 350000

displacement caused in finalized stage of the 300000

250000
bridge. This method can result in a completely

Bending moment
200000
vertical pylon with zero bending21 in the final 150000

self-weight loading condition. Therefore it has no 100000

sense to further analyze the stress conditions in 50000

the pylon. 0
0 100 200 300 400 500 600
Main span [m]

Figure 94 Bending moment at pylon base

4.2.2 Developments on buckling stability of the stiffening girder

With the same approach as described earlier, presented in Figure 95, the buckling force is
determined for the stiffening girder of the bridge. For each increment of the main span, with
step sizes of 50 metres, the buckling force is
determined. An additional force of
ΔF = 400000 kN has been applied in order to
obtain visible second order effect in the
deflection of the stiffening girder.
Figure 95 Additional F to determine buckling force

Figure 96 presents the results for buckling force calculation for each considered span length
up to 500 metres for a loading combination including self weight, permanent loading, traffic
load over the full length of the bridge and pretensioning of the main cable.
Ncr main span
Euler buckling force
Ncr side span
N Ed
5000000
4500000
Euler buckling force [kN]

4000000
3500000
3000000
2500000
2000000
1500000
1000000
500000
0
0 100 200 300 400 500 600
Main span length [m]

Figure 96 Development of the Euler buckling force deck as a function of an increasing main span

Again buckling of the side span remains governing over bucking of the main span of the
girder. The buckling force Ncr for the side span is well below the Ncr of the main span, even
for an increasing main span up to 500 metres.
21
Gimsing, N.J., Cable supported Bridges, Wiley&Sons, 1998

67
 Self-anchored suspension bridges: Part II

The reason for the increased Euler buckling force for an increasing span is the fact that height
of the box girder is linearly increased with the main span. This will increase the moment
inertia more than quadratic, see Table 16, and so the bending stiffness is increased with the
same proportion.
If the above presented graph is expressed in the so called n-value than it becomes visible that
buckling is getting more and more critical with an increasing span. When the n-value defined
as:
N N
nside _ span = cr;side span and nmain _ span = cr;main span
N Ed N Ed

The n-value for each considered span lengths are given in Table 18.
Table 18 n-values for main and side span
Main span Side span n-value side
[m] n-value main span [m] span
150 19,8 63 12,9
200 19,2 83 9,9
250 19,7 104 8,3
300 20,4 125 7,3
350 21,2 146 6,6
400 22,3 167 6,2
450 23,8 188 5,9
500 25,8 208 5,8

With the earlier given general used guidelines with respect to interpreting the n-values:
1<n<2 wrong design
2<n<3 design problems to be expected
n>3 proper design, however, geometrical non-linearity should be taken into
account
n > 50 consequences of geometrical non linearity can be neglected

The following conclusions can be made:

• Buckling of the main span is a downward buckling mode and would therefore
encounters upward resistance by the main cables and hangers. Buckling of the
side span is an upward buckling mode and would therefore encounter no
resistance by the main cable and hangers.
The upward buckling of the side span occurs at a much lower buckling force
Ncr and is decisive over buckling of the main span.
• The buckling stability of the side span girder becomes more and more a critical
design issue with an increasing span up to 500 metres. The so called n-values
given in Table 18 indicate that the consequences on geometrical non-linearity
have become more evident with an increasing main span. But given the
guideline regarding the interpretation of the n-value, the considered span
lengths are still considered to be of proper design because the n-value is in all
cases well above 3.
• The given approach to research the buckling stability of the girder presented in
this part has shown that buckling of the main span is not decisive. In this
approach a stiffening girder is chosen with the same slenderness along the
complete length of the bridge. With n-values of 20 or more (see Table 18),
regarding buckling of the main span, a much more slender girder could be
chosen for the main span. This can be of great contribution to the reduction of
material use.

68
 Self-anchored suspension bridges: Part II

So with respect to the buckling stability of the stiffening girder, a self-anchored suspension
bridge is possible up to a main span of 500 and maybe even beyond that. Assuming a
limitation for the girder slenderness of about λ = 1/100 regarding the buckling stability of the
deck is point of discussion.

4.2.3 Frequency behaviour

A good indication for the resistance against flutter is the development of the ratio between the
torsional and bending frequency.
With an increasing main span both bending and torsional frequency decrease significantly but
the ratio between these frequencies stay well above the general accepted level of 2, see Table
19. When this ratio is above 2 than the structure should have enough resistance against flutter
instability.

Table 19 Frequency ratio for each span length

Main 1st 1st
span fbending ftorsion Ratio * This main span length is the same as the Yeoungjong Grand Bridge.
[m] [Hz] [Hz] ft/fb The natural frequencies for this bridge are:
150 0,72 4,43 6,2 1st fbending = 0.454 Hz
1st ftorsion = 1.223 Hz
200 0,6 3,4 5,7 The frequencies calculated in this research are nearly the same, only
250 0,52 2,81 5,4 for the torsional frequency there is a deviation visisble. This is caused
300* 0,46 2,35 5,1 by the fact that in in the Yeoungjong bridge a truss girder was applied
350 0,4 2 5,0 which has in general a lower torsional stiffness than box girders (used
in this research).
400 0,36 1,75 4,9
450 0,32 1,53 4,8
500 0,29 1,36 4,7
Vcr;vortex
Critical wind speed flutter and vortex
Vcr;flutter
500
A decreasing ratio between the bending 450
Critical wind speed [m/s]

and torsional frequency indicates that the 400

350
bridge structure becomes more sensitive 300
for flutter. This is visualized in Figure 97 250
200
where the development of the critical 150
wind speed for flutter and vortex is 100
50
plotted for the given bending and 0

torsional frequencies in Table 19. 0 100 200 300 400 500 600
Main span length [m]

= V reference = 30 m/s
Figure 97 Critical wind speed for flutter and vortex as a
function of the main span

The critical wind speeds as presented Figure 97 are determined by:

• Vortex critical wind speed: Vcr ;vortex = 12 * f * d 4

For vortex it should be verified that Vcr;vortex > V reference

• Flutter critical wind speed: V ⎛ fb ⎞ m*r

⎜
cr ; flutter = 4 * f t * b * ⎜1 −
⎟⎟ *
⎝ ft ⎠ ρ * b3

69
 Self-anchored suspension bridges: Part II

For vortex it should be verified that Vcr;flutter > 1.3 * V reference

V reference is given also by the Eurocode NEN-EN-1991-2-4 Annex A. For the Netherlands
counts:
Vreference = 30 m/s

Figure 97 shows that the flutter phenomena becomes the governing dynamic response
phenomena when the main span length is getting larger. But for vortex shedding as well as
flutter, a self-anchored suspension bridge up to 500 metres should be stable, the critical wind
speed of both vibration motions is well above the reference wind speed Vreference.

Wake galloping and Stall flutter are not considered because the following criterion is not met:
b<4*d4 in which b = deck width and d4 = structural height of box girder.
So the bridges up to 500 metres considered in this study are not vulnerable for these vibration
motions.

4.2.4 Developments on reaction forces

The reaction forces at the end support have to be anchored into a:
• Vertical anchorage, which resist the vertical component of the cable force.
• Horizontal anchorage, which resists the horizontal component of the cable force.
As the horizontal reaction force is resisted in the deck, only a vertical reaction force has to be
resisted externally. As described earlier in the literature survey of this study, there are two
ways to anchor this uplifting vertical component of the main cable force:
- tie down cables/structure. The deck is vertically anchored and tied down to a
foundation which activates a ground mass to compensate the vertical action.
Also a big counter mass like a concrete block could be applied.
- a bracket at the location of the connection of the main bridge and approach
bridge. The weight of the approach bridge will balance the vertical force.
- a combination of these two mentioned methods.

Governing loading combination for the vertical reaction force in the end support is when the
side spans are unloaded, then no reducing effect of the traffic loading on the side span is
present. Figure 98 shows the increase of Reaction forces
Vertical reaction force

the vertical reaction force in the end Horizotal reaction force

supports for an increasing main span. 180000

160000
140000
Reaction force [kN]

The critical issue will be the horizontal 120000

anchorage of the main cable, see Figure 100000

80000
98. The combination of the rapidly 60000
increasing horizontal cable force and the 40000

much more complex nature of the 20000

0
horizontal anchorage makes it a 0 100 200 300 400 500 600

bottleneck in the design process. Main span [m]

Figure 98 Reaction forces in end supports

The horizontal anchorage is a much more complex structure due to the fact that the main
cable has to be splayed in many individual strands which all have to be anchored individually.
Many provisions like strands shoes, sockets, steel plates and stiffeners are required to
introduce the cable force in the bridge deck, see Figure 99.

70
 Self-anchored suspension bridges: Part II

Figure 99 Cable anchorage Konohana Bridge

The vertical anchorage is a relatively less complex structure then the horizontal cable
anchorage. A vertical counter force has to be activated which is much more easy to introduce
into a heavy concrete block or a bracket connection to the approach span.

4.2.5 Material use

This part gives an estimation of the amount of steel use in the self-anchored suspension
bridge. The main bridge components like deck, pylon, main cable and hangers are executed in
steel in this study. The total steel use is sum of these components.
For each span up to 500 metres the amount of steel has been calculated and a differentiation is
made to the steel use of each of the before mentioned bridge components.
Basic assumptions for the determination of the material use include:
• Girder: material use for the girder includes
-Top- and bottom flanges
-Webs
-Longitudinal stiffeners on top and bottom flanges and web
-An estimate of diaphragms and bracing with an intermediate
distance22 of 2.5 times the construction depth h of the box girder
• Cables: effective area of the main cable and hangers
• Pylon: estimation of the material use for the box section of the pylon includes
longitudinal stiffeners.

Figure 100 shows the total amount of steel used for a main span up to 500 metres. For an
increasing main span the required amount of steel grows almost linearly. If a differentiation
is made to the several before mentioned bridge component it becomes clear that the dominant
part in material use is the stiffening girder. Figure 101 shows that the girder consumes by far
the most amount of steel. But it also becomes visible that with an increasing span the required
amount of steel in the main cable is getting significant.

22
Romeijn, A. CT5125 Steel Bridges Part 1&2. Faculty of Civil engineering and Geosciences

71
 Self-anchored suspension bridges: Part II

Total amount of steel m^3 steel pylon

3500,00 Ste el use pe r cate gorie m^3 steel hangers
m^3 steel cable
m^3 steel girder
3000,00
100%
Amount of steel [m^3]

2500,00
80%

Amount of steel [m^3]

2000,00
60%
1500,00
40%
1000,00

500,00 20%

0,00 0%
0 100 200 300 400 500 600 150 200 250 300 350 400 450 500
Span length [m] Span length [m]

Figure 100 Total amount of steel

Figure 101 Steel use per category

The total steel use can also be expressed in the steel use per meter of length of the bridge.
Table 20 gives an estimation of the steel use per meter of bridge based on the assumption of
this research.
Table 20 Steel use per meter
Main span length
[m] 150 200 250 300 350 400 450 500
Steel use [ton/m] 19,1 19,7 20,5 21,4 22,4 23,6 24,9 26,3

Regarding these findings the following remarks have to made:

• As Figure 101 clearly shows is that much attention needs to go out for the
design of the stiffening girder. Saving material in this bridge component can
reduce the total amount of steel use significantly. Weight saving can be
achieved by adjusting the girder slenderness, reducing the structural height
improves the self weight. But this will have effect on the buckling resistance,
frequency behaviour and erection phase of the bridge.
• Figure 101 shows also a development of the relatively growth of the required
steel for the main cable. Although this material (high strength steel) is more
expensive than the structural steel used for the other bridge components, this
bridge element has the most efficient load transfer (because it is only tensile
loaded) and it is therefore profitable to put more material in the main cable in
order to be able to reduce the material use in the stiffening girder. Because the
stiffening girder is loaded with large bending moments, this bridge component
is much less efficient in material use because many provisions have to made
with respect to local stability within the plated box girder.
Also the presented parameter study has shown that increasing the axial
stiffness of the main cable (putting more material in the cable) reduces the
bending moments significantly in the stiffening girder.

So an optimization can be made with respect to the ratio of axial stiffness of the main cable
and the required material use for the box girder. Putting more material in the main cable can
reduce the required mechanical properties and material use of the stiffening girder. In that
way a significant reduction can be achieved in the total material use and therefore the costs.

72
 Self-anchored suspension bridges: Part II

4.2.6 Effects on the erection

There are two main erection methods to distinct nowadays:
• Cantilever method by means of the cable stayed method
• Erecting by use of temporary supports
For a more detailed description of both methods, reference is made to the literature survey of
this study, part 6.2.1.
Both methods have in common that erection of the deck has to take place prior to the erection
of the main suspension cable.
Great advantage of the first mentioned method is the elimination of temporary supports in the
navigation channel. But the fact is that this method implies building a cable stayed bridge
first, then erect the main cable and subsequently take away the cable stays. This will justify
technically and economically the choice for a cable stayed bridge, and not a self-anchored
suspension bridge. Only esthetical reasons remain as justification to build a self-anchored
suspension bridge.
The second mentioned method, erecting on temporary supports, is nowadays the most
accepted method to erect a self-anchored suspension bridge. Both Konohana bridge and
Yeoungjong Grand bridge (and many others) have been built using this method. The
economical and technical feasibility of this erection method depends on several factors like:
• The number of temporary supports
• The distance between the temporary supports
These two factors greatly influence the design of a self anchored suspension bridge in costs,
erection and structural properties of the stiffening girder. An optimum should be achieved in
the number of- and distance between the temporary supports in order to reduce the cost and
the need for excessive bending stiffness of the stiffening girder.

Number of- and distance between temporary supports

For the considered bridges in this study with a main span up to 500 metres, an exploration is
given now on the consequences for erection on temporary supports.
A calculation is made to determine the maximum free span length of the box girder based on
the assumptions that:
• The box girder is simply supported between two temporary supports. So for the
effective width calculation only the β1 factor is calculated for sagging bending.
• Only self weight of the steel box girder is taken into account, because no traffic
loading and permanent loading like asphalt occurs in this stadium of the bridge
erection.
• According to NEN-EN 1991-2-6 part 4.1.3, the safety factor for self weight during
erection γG;sup = 1.05
• The maximum bending moment in the box girder during this stage is determined by
the distance between the temporary supports according to:
1 2
M max;sagging = * q selfweight * l temp. sup .
8

73
 Self-anchored suspension bridges: Part II

Based on these assumptions the maximum achievable free span length of the box girders
(applied in the bridge models up to 500 metres) are given in Table 21.

Table 21 Achievable free span length and number of required temporary supports
Main Side Height Maximum Temporary Temporary
span span box free span supports supports Total
length length girder length in main in each temporary
[m] [m] [m] [m] span side span supports
150 63 1.6 56 2 1 4
200 83 2.1 72 2 1 4
250 104 2.6 95 2 1 4
300 125 3.2 122 2 1 4
350 146 3.7 144 2 1 4
400 167 4.2 164 2 1 4
450 188 4.7 180 2 1 4
500 208 5.3 196 2 1 4

This table shows that for each span length the box girder, with a slenderness of λ = 1/95, has
enough stiffness to span over 1/3 of the main span, meaning that 2 temporary supports are
sufficient for each considered span length.
Also a temporary support in the side span is necessary for all span lengths, making the total
required temporary supports 4. But Table 21 shows that the temporary supports in the side
span can be eliminated if:
-A less slender deck is chosen, in that way the section modulus Weff is increased and
therefore the maximum free span length during erection is bigger.
-The side span length is reduced. In that way the free span length of the box girder is
sufficient enough to span the complete side span in erection phase.
Elimination of the temporary supports in the side span reduces the total number of required
temporary supports to 2. But in general it is not a problem to situate temporary supports in the
side span, it is the main span where the temporary support will cause more problems.

Required lifting capacity

With 2 temporary supports in the main span, huge deck Deck
sections have to be lifted by a (floating) crane or strand Main section Crane
jack systems. An increasing span of the bridge has large span length capacity
[m] [m] [tons]
consequences on the required lift capacity of the crane.
150 50 945
Table 22 gives an indication of the required crane
200 67 1269
capacity to lift the huge deck section (approximately with 250 83 1597
a length of 1/3 of the main span) on two temporary 300 100 1931
supports in the main span. 350 117 2268
400 133 2611
Regarding erection cost an optimization has to be made 450 150 2958
with respect to the number of temporary supports and the 500 167 3309
lift capacity of the (floating) crane.
Table 22 Required lifting capacity

74
 Self-anchored suspension bridges: Part II

4.3 Evaluation and conclusion for an increasing span length

Several design issues have been evaluated in this chapter with respect to an increasing main
span of a self-anchored suspension bridge. This part will briefly discuss these design criteria
and mention the critical issues which require consideration in the design process of a self-
anchored suspension bridge in general.

• Stiffness
An important aspect to influence the force distribution in a (self-anchored) suspension bridge
is the ratio between the deck bending stiffness EIgirder and the axial stiffness EAcable. It is more
profitable to put more material in the main cable (increasing axial stiffness) in order increase
the stiffness of the bridge and to reduce the bending moments in the stiffening girder.
Designing and dimensioning structural bridge members under bending is always less effective
and more material consuming then that of members under tensile loading, such as the main
cable.
The stiffness criteria, expressed in allowable deflections, are easily met by choosing the
proper dimension for the main cable. For a bridge span length up to 500 metres, global
stiffness is not a critical design issue.

• Static strength
The stress levels in main cables have shown that stiffness is the governing design criteria over
static strength criteria of the main cable.
For the box girder a distinction can be made between tensile and compressive area for the top
and bottom flanges. Critical area is the support of the girder near the pylon. The large hogging
moment introduces high compressive stresses in the bottom flange at this location. But in
practice these stresses can be adjusted by means of support displacements in construction
phase of the bridge.
A second critical issue is that for an increasing span, the compressive stresses (introduced by
the main cable) in the girder become dominant. This can have consequences on the local
stability for the normally tensile bottom flange in the mid of the main span. More stiffeners
are required then on the bottom flange in the main span.
But overall for an increasing span length, stress levels in the box girder are of manageable
levels. Local additional plate thicknesses and stiffeners are required for some locations
depending on the span length.

• Buckling stability of the stiffening girder

An upward buckling of the side span is the decisive buckling mode of the stiffening girder.
Downward buckling of the main occurs at a much higher buckling force Ncr than for the side
span.
Research to the buckling phenomena of the side span has shows that with an increasing span,
the resistance against buckling of the side span reduces. Up to a main span length of 500
metres, buckling of the stiffening girder should be analyzed in the design process but it is
possible to reach such a span length.
The chosen slenderness λ = h/l main span of about 1/95 (until now the most slender deck ever
applied in a self-anchored suspension bridge; Konohana bridge in Japan) of the box girder is
sufficient to reach a span length of 500 metres. With respect to buckling of the main span of
girder, even a much more slender deck in the main span seems possible. Investigating the
possibility of a variation of the deck slenderness in longitudinal direction of the bridge
(application of stiffer deck in the side than in the main span) could greatly reduce the material
use.

75
 Self-anchored suspension bridges: Part II

• Frequency behaviour
Regarding the natural frequencies, the torsional and bending frequencies are well separated
for a span length up to 500 metres when a box girder is applied. The ration between fbending
and ftorsion is well above 2. So the self-anchored bridge type in this research has enough
resistance against flutter.
For an increasing span, flutter becomes more critical than vortex shedding of the deck. But the
critical wind speed for both vibration motions is well above the reference wind speed Vreference
indicating that bridges up to 500 have enough resistance against these two phenomena.
Attention has to go out to the aerodynamical shape of the girder. For cable supported bridges
it is generally advised to execute wind tunnel test to check if the bridge model has enough
resistance against vortex and flutter girder instabilities.
The natural frequencies obtained in this research give room for the possibility of choosing a
more slender deck than λ = 1/95 that is chosen in this research.

• Reaction forces
Due to the complex nature of the horizontal anchorage of the main cable, the horizontal
anchorage requires much attention. For an increasing main span, the horizontal cable force
increases rapidly and is has enormous consequences for the horizontal anchorage. The
introduction of the horizontal cable force requires many provisions like anchor shoes, plate
stiffeners. With an increasing main, the main cable diameter increases and contains more
strands to be anchored.
For the large span bridges an anchorage in which the cable is looped around the deck is worth
consideration (as applied in the East Bay bridge, see 5.5.4 of the literature survey)

• Material use
By far the biggest part of material use is required for the stiffening girder. For a bridge model
with a main span up to 500 metres, at least 70 percent of the total material use is taken by the
stiffening girder. And for an increasing main span also the contribution of the material use in
the main cable becomes significant. Up to 25 percent of the material use can be taken by the
main cables and hangers.
So the biggest cost reduction can be achieved by saving material in the stiffening girder. This
can be achieved by:
-Reducing the slenderness of the girder in the main span. As shown earlier, for
buckling resistance a more slender girder can be applied in the main span than in the
side span.
-Increasing the axial stiffness EAcable in order to reduce the bending moment in the
girder.
-Optimization of plate thicknesses along the length of the girder.

• Effects on erection of the bridge

With a slenderness of the girder of λ = 1/95, a free span length of more than one third of the
main span can be reached. This results in 2 required temporary supports in the main span and
one in each side span. The temporary supports in the side span can be eliminated by:
-Reducing the side span length
-Increasing the stiffness of the girder, keeping attention on the buckling stability of the
side span is necessary.
Erection of the deck on temporary supports results in a relatively high required stiffness of the
deck to reach a certain free span length. Not the stability phenomena of the deck but the
erection method is governing for the required slenderness of the deck.

76
 Self-anchored suspension bridges: Part II

Conclusion
This chapter has presented the exploration of the span possibilities of a self-anchored
suspension bridge with a main span up to 500 metres. It has been shown that it is technically
feasible to achieve a span length up to 500 metres. The expected problems regarding the
buckling stability of the main girder has turned out to be feasible for span lengths up to 500
metres. Main topics of attention for designing a self anchored suspension bridge with a large
span are:
-Support conditions of the girder at the pylon. Large compressive and tensile stresses
in the flanges of the flanges of box girder can occur. The hogging moment at this
location introduces high stresses in the flanges. Attention has to go out for support
displacement during erection of the bridge which can reduce the stresses in finalized
condition of the bridge.

-The possibility of compressive stresses in the bottom flange of the box girder in the
main span. This zone is in general an area where tensile stresses occur due to the
global bending moment. So in normal tensile stress condition in the bottom flange of
box girder in the mid of the main span, no local buckling stability has to be
considered. But for a larger getting main span of a self-anchored suspension bridge,
compressive stresses can occur in the bottom flange at the mid span, see §4.2.1 for a
main span length of about 400 meters and beyond that. This requires more longitudinal
stiffeners to resist local buckling instabilities which is more material consuming and
therefore more costs are involved.

-Required slenderness of the box girder is mainly dominated by the erection method
on temporary supports and not so much the buckling stability. Depending on the
circumstances during erection the slenderness of the box girder is can be even more
slender than λ = 1/95.

-Buckling of the stiffening girder. There is a difference in the buckling force Ncr of the
side- and main span. Based on the assumptions in this research, buckling of the girder
in the side span is decisive. A chosen girder slenderness of λ = 1/95 is sufficient to
resist buckling. For the main span even a more slender girder is possible because the
n-value is about 20-25 for a span length up to 500 metres. A slender girder can be of
great contribution of cost reduction, since at least 70 % of the total steel use in the
bridge is taken by the girder.

-Horizontal anchorage. Introducing the cable force in the deck requires attention for
splaying the and individually anchoring of the steel strands. With an increasing main
the axial compressive force increases quickly.

-Erection method. Stiffness of the deck and the related free span lengths determine the
required number of temporary supports during erection. Increasing the stiffness of the
deck also increases the maximum distance between the temporary supports but
requires more lifting capacity and is more material consuming.

-In this case a sag over span ratio has been chosen of 1/5 in order to reduce the
bending moments and normal force in the girder. Regarding the buckling stability of
the girder, which is sufficient with a girder slenderness of λ = 1/95, a smaller sag ratio
can be considered which increases the normal force the deck. Decreasing the sag to
span ration contributes to the reduction of the main cable length and pylon height.

77
 Self-anchored suspension bridges: Part II

5 Erection aspects of the New City bridge at Nijmegen

A successfully design of a self-anchored suspension depends also on the feasibility of the
erection method. As shown earlier the required mechanical properties of the girder can
depend greatly on the chosen construction method. So the way of construction should already
be considered early in the design process.
Choosing an erection method mainly depends on local conditions of the surrounding area of
the construction site of the bridge and criteria imposed by the client. Therefore it is hard to
conclude something in general for a construction method in case of self-anchored suspension
bridges. To make an assessment for the construction method, the situation of Nijmegen and
the Waal river is used again.
A brief exploration is made to point out the critical issues in erecting a deck prior to the main
cable.

5.1 Criteria Nijmegen City

Chosen design by engineering office Iv-Infra
The feasibility and design of a self-anchored anchored suspension bridge for Nijmegen was
researched by engineering office Iv-Infra Amsterdam. This design proposes a main span of
335 metres which would mean that the longest existing span is exceeded by more than
10 percent.
To explore and evaluate a construction method for a self-anchored suspension bridge at
Nijmegen presented in this chapter, focus is made on a main span of 350 metres. In that way
nearly the complete river is spanned.

Design criteria regarding erection

Several design criteria are defined by the city council of Nijmegen regarding the erection of
the New City Bridge. The following statements23 count for the Waal river near Nijmegen
where the construction site is located:

Part 3.1.3 Navigation clearance

ID Description
The minimal required vertical navigation clearance of 9.10 metres should be available for at least
11,2,4,2 80% of the of normal navigation width (265 metres)

Part 3.5.4 Erection

ID Description
Shipping traffic must be able to continue and encounter as little nuisance as possible during
11,6,5,3 erection of the bridge
11,6,5,4 No obstacles are allowed in the navigation channel
The minimal required navigational clearance (width and height) may be temporary altered during
11,6,5,5 erection
The minimal required vertical navigation clearance may not be smaller than that of the Railway
11,6,5,6 bridge that is in the vicinity of the New City Bridge
The normal navigation width (265 metres) should be maintained as much as possible during
11,6,5,7 erection of the bridge. A partial obstruction of the navigation channel is allowed incidentally

23
PvE Stadsbrug versie 1_0 20juni06.pdf, collected from the website:
http://www2.nijmegen.nl/mmbase/attachments/359936/PvE_Stadsbrug_versie_1_0_20juni06.pdf

78
 Self-anchored suspension bridges: Part II

So no clear restrictions are defined for the possibility of- and distance between temporary
supports in the Waal river. Temporary and partly blocking of the navigation channel is
allowed, enabling the erection on temporary supports in the Waal river for constructing a self-
anchored suspension bridge.

5.2 Exploration of erection methods

A short exploration is given on the erection methods of the two most significant self-anchored
suspension bridges; Konohana and Yeoungjong Grand bridge. And because of the unique
character and construction method, the construction method of the Nescio bridge in
Amsterdam is given.

5.2.1 Existing methods

An exploration of existing and most recent construction methods comprise the erection on
temporary supports and utilizing temporary cable stays.

Construction Konohana bridge

The main aspects in the construction of the Konohana Bridge are presented in Appendix 8a
Erection of Konohana Bridge. The main features are summarized here:

• Distance between temporary supports of 120 metres

• Heaviest section lifted by a floating crane of 2670 tons
• 2 temporary supports in the main span
• 1 temporary in one of the side spans. This side span was situated partly over land
where the crane could not reach to lift a deck section. There it needed temporary
support in one of the side span.
• The deck section in the middle was erected at the end a lifted on two ends of deck
which cantilevered out over the temporary supports.

Construction of Yeongjong Grand Bridge

The main aspects in the construction of the Konohana Bridge are presented in Appendix 8b
Erection of Yeoungjong Grand Bridge. The main features are summarized here:

• Distance between temporary supports of about 100 metres

• Heaviest section lifted by a floating crane of 3072 tons
• 2 temporary supports in the main span
• 1 temporary in each of the side spans. The deck section in the middle was erected at
the end a lifted on two ends of deck which cantilevered out over the temporary
supports.

Construction Nescio bridge Amsterdam

The main aspects in the construction of the Konohana Bridge are presented in Appendix 8c
Erection of Nescio bridge The main features are summarized here:

• Heaviest section lifted by a floating crane of 570 tons24

24
Habraken, A., Y. Ichimaru, Nesciobrug Amsterdam, Met een slinger over het kanaal, Bouwen met Staal 193,
December 2006

79
 Self-anchored suspension bridges: Part II

• The lifted deck had total a length of 200 metres.

• The deck was completely prefabricated and transported on a barge to Amsterdam,
travelling over more than a 100 km.
• Deck was lifted in one piece (of 200 metres) and supported on the permanent end
supports and by temporary stay cables attached to the pylon.

During the erection of the Nescio bridge temporary stay cables were used. Given the relative
light weight of the deck section of 570 compared to sections used for the Konohana and
Yeoungjong Grand bridge, it was possible to use only two stays in the main span to support
the deck.

Conclusion
This brief exploration shows that these three most recent self-anchored suspension bridges
exhibit two main similarities in erection, i.e.:
-lifting large and heavy deck section up to almost 3100 tons with lengths of 100-
200 metres.
-temporary support of the deck by means of a temporary piers for heavy deck sections
-temporary support of the deck by means of temporary stay cables for relatively light
weigh sections.

5.2.2 Alternative methods

Some possibilities are briefly discussed in order to explore the erection of the main cable prior
to the deck.

Possibility of erecting main cable prior to the girder

This method would enable a construction sequence that is similar to a conventional
suspension bridge where the main cable is erected first and the deck is lifted in sections and
connected to the hangers. In this way the temporary supports are eliminated. But with an
absence of an external anchorage the horizontal cable force has to be resisted by means of :

• Compressive struts in the side span

To resist the horizontal cable force, a compressive strut is needed in the side
span that reaches from the end support to the basement of the pylon.

Figure 102 Compressive strut in each side span

The compressive struts require an enormous length and therefore very prone to
buckling. Also the pylon’s basement will have to resist a horizontal component
which has consequences for the pile foundation that is mainly designed for
vertical loading and bending moments.

80
 Self-anchored suspension bridges: Part II

• Bending of the pylon

The earlier erected side spans (with or without the use of temporary supports)
could be utilized as a compressive strut to stabilize the horizontal cable force.

Figure 103 Utilizing the erected side span

The deck will have to be horizontally fixed at the pylon and introduce the
horizontal force in the pylon resulting in a bending moment. Also the
foundation of the pylon has to resist a horizontal force which is not a desirable
situation.

• Erecting a temporary external anchor block

The self weight condition of the bridge is so large compared to the traffic
loading condition, that the required dimensions for a temporary external
anchorage are almost that of an anchorage required for a finalized situation of
the bridge. This method will justify for a permanent external anchorage
resulting in a conventional suspension bridge and is therefore not logical to
consider.

Figure 104 Temporary external anchorage

To be able to erect the main cable prior to erection of the deck, requires drastic temporary
measures and will not be further considered.

Nijmegen city
Based on these findings a point of departure is defined to explore a construction method for a
Bridge in Nijmegen. Based on the heavy required box section for the city bridge at Nijmegen
the generally most accepted construction method will be briefly discussed in the next
paragraph: constructing with temporary piers in the Waal river.

Points of interest are evaluating the consequences for the:

-distance between temporary piers
-required dimensions of the deck
-required lifting capacity of the (floating) crane
-structural consequences for the deck
-distance of cantilevering over the temporary supports

Point of departure will be a bridge with a main span of 350 metres, similar to the design
proposal of engineering Iv-Infra offered to the city council of Nijmegen.

81
 Self-anchored suspension bridges: Part II

5.3 Constructing with temporary piers in the Waal river

First a brief exploration is given in the achievable maximum distance between the temporary
supports. Emphasis is made on the fact that many variances are possible on the distances and
number of temporary supports in the main- and side span. The brief exploration given in this
part focuses mainly on the recognition of general structural problems for the deck under self
weight loading that occur during this stage.

For navigational purposes it is desirable to maximize the distance between the temporary
supports. The basic assumption for this exploration are:
• The box girder is simply supported between two temporary supports or between zero
bending location in the bending moment line. So for the effective width calculation
the β1 factor for sagging and β2 factor for hogging moment are determined by the
effective length Leff between the temporary and permanent supports.
• Only self weight of the steel box girder is taken into account, because no traffic
loading and permanent loading like asphalt occurs in this stadium of the bridge
erection.
• According to NEN-EN 1991-2-6 part 4.1.3, the safety factor for self weight during
erection γG;sup = 1.05
• Temporary supports in the side will not be a critical issue and is assumed to be located
at the mid of the side span. Other configuration are possible but will not greatly
influence the critical main span conditions during erection.

5.3.1 Simply supported between temporary supports

As shown in §4.2.6, the free span length of a simply supported box girder with a slenderness
of λ = 1/95 is about 144 metres (with girder properties as shown in Table 16). This situation
means that the temporary supports would reduce the navigation width (265 metres) of the
Waal river significantly, see Figure 105 . To lift a deck section of 144 metres long requires a
lift capacity of about 2800 tons.

Figure 105 Temporary supports Waal river

Some measures to increase the distance between the temporary supports are for instance:
• Increasing the strength of the deck, for instance by means of choosing for a less
slender deck.
• Temporary pretensioning of the lifted deck section by an internal pretensioning cable.
This enhances the moment capacity of the deck enabling a larger free span length.
Furthermore it is desirable to have a zero difference in rotation of the location of the two deck
sections where they are welded together (at the location of zero bending at the supports), so a
camber should be applied on each lifted prefabricated deck section in order to ensure that the

82
 Self-anchored suspension bridges: Part II

deck section is in horizontal position after self weight deflection between the temporary
supports.
Another point of attention is the shear force during erection. The shear force in erection phase
is much higher than in the situation of the continuous supported deck by the hangers. The
maximum shear force in this case is determined by VEd = ½*qself weight * l1;distance temp.supports. The
maximum shear force capacity of the cross section is determined by NEN-EN-1993-1-5:
f h t 355*3685*13* 2
Vc , Rd = yw w = 19637kN
3γ M 1 3
The acting maximum shear force in the deck during this erection scheme is VEd = 13320 kN
So the shear force capacity is still sufficient in erection phase but remains an issue to be
checked for erection phase of the bridge.

Points of attention
The main points of attention for this method are:
-Camber of deck sections. A lifted deck section of 144 metres requires a camber of
about 780 mm in order to be in horizontal position after deflection and to be able to
weld this section to other sections.
-Welding conditions at the temporary and permanent supports to connect the deck
section.
-Reduction of the navigation channel to 55% involves risk of collision.
-Bending moment condition after hanger installation. Due to the simply supported
condition between the temporary supports the bending moment distribution after
hanger installation will deviate from the assumed condition as mentioned in §1.1.7.
-Prefabrication and transport of the large deck sections

5.3.2 Erecting as a continuous beam

In the finalized stage of the erection of the deck, when all sections are welded together, the
deck can be considered as a continuous beam on temporary- and permanent supports. The
sagging- and hogging bending moments due to self-weight in this situation determine the
stress condition in the flanges of the box girder. The location of the zero moments can be
utilized to divide the entire deck into several sections to be lifted. In that way the last section
in the mid can be lifted between to canti-levered deck sections over the temporary supports,
see Figure 106. The length of this section depends on the location of zero bending in the main
span, see Figure 107.

Figure 106 Canti levered deck sections

According to the given erection criteria given by the city council of Nijmegen, see §5.1, it is
desirable to span at least 80 percent of the navigation width of the channel. This results in a
required distance of the supports of approximately 210 metres.
The bending moments of a finalized erected deck under self weight loading on temporary
supports with a distance of 210 metres is given in Figure 107.

83
 Self-anchored suspension bridges: Part II

1 2
Figure 107 Bending moments [kNm] with temporary supports at 210m distance

This moment distribution reveals critical spots for the deck at two locations:
1. hogging moment at temporary supports in main span
2. sagging moment field location in the mid of the main span
Other locations along the length of the deck during erection will cause no problem due to the
limited bending moments compared to location 1 and 2 mentioned above.

With girder properties as shown in Table 16 the stresses at these locations are:
Location 1.:
Table 23 Stresses box girder erection phase, location 1
Effective width β1;top flange 0,179
β1;bottom flange 0,196
Plate thickness flanges t equivalent flange top 40 mm
location 1 t equivalent flange bottom 20 mm
Effective section
modulus Weff;top 953303462 mm3
Weff;bottom 604543040 mm3
Stresses flanges σtop flange 586 N/mm2
σbottom flange 923 N/mm2
Shear force VEd 19079 kN
Shear force resistance V c,Rd 19637 kN

Location 2.:
Table 24 Stresses box girder erection phase location 2
Effective width β1;top flange 0,434
β1;bottom flange 0,485
Plate thickness t equivalent flange top 40 mm
t equivalent flange bottom 20 mm
Effective section
modulus Weff;top 2265817108 mm3
Weff;bottom 1361830242 mm3
Stresses flanges σtop flange 196 N/mm2
σbottom flange 326 N/mm2

84
 Self-anchored suspension bridges: Part II

The stresses in the flanges of the box girder at the location 1, see Table 23, are well above the
acceptable yield strength. These peak stresses can be solved by an application of additional
plate thicknesses in the flanges over a certain length near the temporary supports.

Tabel 25 Stresses at location 1 with additional plate thickness

Additional plate
thickness t equivalent flange top 75 mm
location 1 t equivalent flange bottom 60 mm
Effective width β1;top flange 0,179
β1;bottom flange 0,196
Effective section
modulus Weff;top 1782264229 mm3
Weff;bottom 1593873380 mm3
Stresses flanges σtop flange 313 N/mm2
σbottom flange 350 N/mm2

Tabel 25 shows that when additional plate thicknesses are applied in the box girder at the
temporary support location, the stresses are reduced to acceptable levels. The additional plate
thickness should be applied in the area where the bending moment, given in Figure 107,
exceeds the moment capacity of the box girder with normal plate thicknesses i.e.
My,Rd;box girder = Weff;bottom * fy = 214613 kNm. The extra required plate thicknesses are
required over a length of about 53 metres near the temporary supports as shown in Figure 108.

Figure 108 Required length for additional plate thicknesses

Other possibilities for reducing the hogging moment at the temporary supports, like support
displacements of the temporary supports, are not discussed but can be worth considering.
Lowering a support reduces the hogging moment but increases the sagging moment in the
field.

85
 Self-anchored suspension bridges: Part II

Erection sequence of the deck

Based on the location of zero bending in each step, the deck can be divided in several section
which have to be lifted individually in a certain sequence. The deck can be welded together at
the zero bending location after each erection step as presented in Figure 109.

4
3 3
2 2
1 1

Step 4

Step 3

Step 2

Step 1

= location of zero bending in each erection step

Figure 109 Lifting sequence of deck sections & zero bending locations in M line

Figure 109 shows the following erection steps:

Step 1: Lifting the a section of 94 metres on the permanent end support and a temporary
support in the side span.
Step 2: Erecting a deck section of 70 metres on the canti levered part of section 1 and the
permanent support near the pylon.
Step 3: Erecting a deck section of 87 metres on the permanent support near the pylon and the
temporary support in the main span.
Step 4: Erecting the closing deck section of 140 metres between the two cantilever parts of
section 3. This step will introduce the decisive bending moments and stress conditions in the
deck during erection.

Points of attention
-Additional plate thicknesses in the flanges are required in the cross section of the box
girder at the location of the temporary supports in main span.
-Other option to reduce the hogging moment situation at the temporary support is to
apply a support displacement. This has both influence of the hogging moment at the
support and the sagging moment in the field.

86
 Self-anchored suspension bridges: Part II

-Shear force condition. Table 23 shows that the shear force after step 4 in the erection
sequence is becoming critical but still sufficient provided that enough longitudinal and
transverse stiffeners are applied at this location.
-Production and erection of the temporary supports
-Prefabrication and transport of the large deck sections
-Navigation clearance underneath the deck during erection
-Collision risk of the temporary supports by shipping traffic

Erecting the deck on temporary supports introduces many problems and decisive stress
conditions compared to the finalized stage of the bridge.

5.4 Constructing without temporary piers in the Waal river

Constructing without any temporary supports in the Waal river implies that temporary stay
cables have to be used to erect and temporary support the deck. Using temporary stay cables
can either be used without any- or in combination with some temporary supports. Some
erection possibilities with temporary stays in combination with temporary supports are given
in Figure 110 and Figure 111. Many variation to these configuration are possible.

Figure 110 Erecting with temporary stay cables

Figure 111 Erecting with temporary stays and temporary supports in the side span

Both methods imply the same bending moments in the deck, large hogging moments will
occur at the location of the stay cables and supports.
The elimination of the temporary supports in the main span is a big advantage of this method.
As shown in §5.3.2, large hogging moments require many provisions in the cross section of
the box girder. By using more than two temporary stay cables in the main span, the hogging
moment can be reduced which is favourable for the stress conditions in the box girder in this
phase.

87
 Self-anchored suspension bridges: Part II

Erection sequence of deck

This erection method is similar to that of the erection of a cable stayed bridge. Two main
erection sequences are known for this bridge type, see Figure 112 and Figure 113.

Figure 112 Erecting side span first

Figure 113 Cantilever process from pylon

Points of interest
Many issues are of importance for the evaluation of this method.
-Number of stay cables
-Properties of stay cable
-Vertical anchorage reaction at end support which can be decisive over the finalized
situation of a parabolic main cable.
-Shear force condition in the deck
-Anchorage of the stay cables to the deck
-Conditions at cable saddle on the top of the pylon, placement of stay cables and
parabolic main cable.
-Erection of the main cable and hangers and the removal of the temporary stays
afterwards.
-Combination with use of temporary supports in the side span and/or partly in the main
span
-Navigation clearance underneath the deck during erection

Many factors influence the feasibility of this erection method but in general when many stay
cable are used the structural consequences for the deck can be less severe compared to the
erection method solely on temporary supports as shown in §5.3.2. Another advantage is that
smaller deck sections can be lifted which reduces the required lifting capacity of the crane or
strand jacks.
A more detailed evaluation of this method is worth considering for the situation at Nijmegen
city where a busy navigation channel makes it not desirable to apply temporary support in the
main span. Reference is made to the erection methods used for cable stayed bridges.

5.5 Conclusions regarding erection

The brief exploration of an erection method for the new city bridge near Nijmegen city shows
that in general the erection phase can cause a decisive stress situation for the deck compared
to the finalized phase. During erection the deck is supported on a few temporary supports and
permanent supports introducing large shear forces and sagging and hogging moments on
locations which do not occur in a finalized situation of the deck where it is continuous
supported by many hangers. The stress conditions in the box girder in erection phase can
require a much stiffer deck than is required for the stress conditions and stability issues in
finalized conditions of the bridge.

For the situation of Nijmegen city a reduction of the navigation width of 265 metres can be
minimized to 80% (210 metres). But the consequences are that extra provisions have to be

88
 Self-anchored suspension bridges: Part II

made on the deck to be able to erect the deck on temporary supports with a distance of 210
metres.
It is clear that a further reduction of the distance between the temporary supports is desirable
because it leads to less severe structural consequences for the deck during erection. The
downside is off course that it gives more interference of the navigation channel.

The possibility of erecting without the use of temporary supports but stay cables is an
attractive alternative in order to eliminate interference of the navigation channel. Other
advantage is that the structural consequences can be less severe compared to the situation
described in §5.3.1. But this method remains laborious and gives a justification for the
application of a cable stayed bridge instead of a suspension bridge.

In general erecting a large self-anchored suspension bridge requires attention for:

-Erecting large prefabricated deck sections
-Required lifting capacity of a crane or strand jack are significant
-Shear force condition during erection
-Stress condition at temporary supports/ stay cables during erection in the box girder
can require extra provisions
-Welding procedure of the erected deck sections
-Number of and distance between the temporary supports
-Prefabrication and transport of the large deck sections
-Interference of the navigation channel

89
 Self-anchored suspension bridges: Part II

6 Overall conclusions and recommendations

Based on this M.Sc. study to the structural behaviour and main span possibilities of a self-
anchored suspension bridge, the following overall conclusions and recommendations can be
made.

6.1 Conclusions
Structural behaviour
A parameter study on the structural behaviour of a self-anchored suspension bridge has
revealed that for the global force distribution the ratio between the deck bending stiffness
EIgirder and the axial stiffness EAmain cable is an important aspect. The global stiffness of- and
the maximum bending moment in the girder are greatly influenced by these bridge
parameters. Regarding the global buckling of the stiffening girder it has been shown that the
side span is decisive, the n-value for the upward buckling of side span is much lower than for
the n-value of the downward buckling of the main span. A remark is made that these findings
are based on the geometry ratio, of for instance side span to main span, that are made in this
research. Adjustment of side span length can increase the resistance against buckling.
The overall conclusion of the parameter study into the structural behaviour is that on static
strength, stiffness, frequency behaviour and the buckling stability of the box girder, a deck
slenderness of the box girder of λ = 1/100 or even more slender is feasible.
Some local critical aspects in such a bridge are the support conditions of the girder at the
pylon and the horizontal anchorage at the end support. Both location are imposed by locally
high stresses which require attention for extra provisions in the cross section of the box girder.

Main span possibilities

The exploration of the span possibilities of a self-anchored suspension bridge has shown that
it is technically feasible to achieve a main span length up to 500 metres. The expected
problems regarding the buckling stability of the main girder has turned out to be feasible for
span lengths up to 500 metres. Regarding static strength, stiffness, frequency behaviour and
buckling stability, a girder slenderness of λ = 1/100 and even more slender is feasible. Even a
main span beyond 500 metres is in technical view possible. An increasing main span up to
500 metres does show that the normal stresses in the box girder become dominant over the
bending stresses, which can result in overall compressive stress conditions in the girder’s
flanges. Also the buckling stability of the girder becomes more critical for an increasing main
span but still feasible. The so called n-value decreases from n = 13 for a span length of
L = 150m to an n-value of n = 6 for a span length L = 500. This indicates that the buckling
resistance at 500 metres is still sufficient but requires close attention in the design process.
The most limiting factor for the slenderness of the girder is the erection phase of the bridge.
Depending on the number of- and distance between the temporary supports, the stress
conditions during erection can easily govern the design of the cross section of stiffening
girder. The choice for the construction method between erecting on temporary supports or
temporary stays has influence on the extent of the required provisions for the box girder
during erection.

90
 Self-anchored suspension bridges: Part II

Erection
The erection stage of a self-anchored suspension bridge remains the critical design issue. The
deck encounters decisive stress conditions in this stage compared to the finalized stage of the
bridge. An optimisation is to be found regarding the number of temporary supports allowed in
the navigation channel and the structural consequences for the deck. A deck slenderness of
about λ = 1/100 makes it possible to span at least 1/3 of the main during erection stage. In that
way only two temporary supports are required in the main span.
.

6.2 Recommendations
• In the pre-design process of a suspension bridge in general, the ratio between the
bending stiffness EIdeck of the deck and the axial stiffness of the main cable EAcable
should be evaluated. The maximum bending moment in the girder and global stiffness
of the entire bridge is mainly determined by these two bridge parameters.

• The support conditions of the girder near the pylon introduce peak stresses due to the
hogging moment and torsional effects by symmetric and asymmetric loading
conditions. A detailed design should asses the required number of-, location of- and
plate thicknesses of the longitudinal and transverse stiffeners to meet the local static
strength and local stability issues.

• A similar recommendation can be made with respect to the main cable’s anchorage to
the deck. The introduction of the individual cable strands requires a detailed design of
the box girder at this location. Many provisions like strand shoes, sockets, steel plates,
stiffeners are to be situated in the cross section of the box girder. Local static strength
and local stability issues of the steel plated elements should be assessed.
A system like in the San Francisco bay bridge where the main cable is looped around
the deck might me an interesting option to research into more detail.

• Regarding the material use, costs, structural behaviour and esthetical appearance of the
bridge, it might be worth while to research the possibility to apply a more slender deck
in the main than in the side span. On static strength, stability and dynamic issues the
analyses in this study shows that the possibility exists for a deck slenderness beyond
λ = 1/100.
Also the possibility of the application of a truss girder as an alternative for the box
girder might be interesting to investigate.

• As this research also has shown is that the erection stage causes static strength
problems for the deck. The pre-design of the bridge should include a close and early
assessment of the erection criteria determined by the local surrounding conditions and
the client. Applying more than two temporary supports in the main span will reduce
the decisive stress condition for the deck during erection. This is favourable for the
possibility of applying a slender deck.

• A further research into of the buckling behaviour of the side span and main span in
relation to the geometrical ratio like side span length to main span length can be
interesting. This could reveal more closely the sensitivity of- and difference between
the global buckling of the side span and main span and discover a more optimized
geometrical ratio.

91
 Self-anchored suspension bridges: Part II

• In technical view this research revealed the possibility of self-anchored suspension

bridge with a main span beyond 500 metres. The global research into the structural
behaviour and span possibilities of this research could therefore be extended to a span
range beyond 500 metres.

• No detailed assessment is made in this research into the fatigue strength issue. But as
for every bridge type and cable supported bridges as well, a fatigue assessment should
be made on bridge components like the deck, hanger connections to the deck, hangers,
anchorage of the main cable to the deck, etcetera. From the fatigue strength point of
view, the self-anchored suspension bridge has a similar approach as conventional
suspension bridges. So no real restrictions are to be expected for large span self-
anchored suspension bridges.

• The structural analyses of the self-anchored suspension bridge in this research has
been performed with the FEM program ESA PT 6.0.185. At first sight no second order
effects where visible in the geometrical non-linear analyses. But an alternative
approach to this phenomena, as shown in this research, revealed that it is an issue that
should be closely watched. The buckling stability of the deck is an issue that should be
considered.

92
 Self-anchored suspension bridges: Part II

List of literature
Alsemgeest,D. Rebuilding bridge Kanne, Suspension bridge-Static analyses- Check on
strength, stiffness and stability. Iv-Infra, October 2003

Chen, W. L.Duan. Bridge Engineering Handbook. CRC Press 2000

Cheng, J. et. al., Nonlinear aerostatic stability analysis of Jiang Yin suspension bridge.
Engineering structures 24, 2002, pp 773-78.

Clemente,P. G. Nicolosi, A. Raithel. Preliminary design of very long-span suspension

bridges. Engineering structures 22 (2000), 1699-1706.

Gasparini,D. V. Gautam. Geometrically Nonlinear Static behaviour of Cable Structures.

Journal of Structural Engineering, October 2002, pp 1317-1329.

Gimsing, N.J., Cable supported Bridges, Wiley&Sons, 1998

Habraken, A., Y. Ichimaru, Nesciobrug Amsterdam, Met een slinger over het kanaal, Bouwen
met Staal 193, December 2006

Ichimaru, Y., Design and engineering of ‘Nescio’ bridge-Amsterdam Rhine canal. Arup

Overspannend staal, Rotterdam: Stichting Kennisoverdracht SG, Deel 3: Construeren B,

1996.

PvE Stadsbrug versie 1_0 20juni06.pdf, collected from the website:

http://www2.nijmegen.nl/mmbase/attachments/359936/PvE_Stadsbrug_versie_1_0_20juni06.
pdf

Ren, W. Roebling suspension bridge. 1:Finite element model and free vibration response,
Journal of bridge engineering, March/April 2004, pp 110-118

Romeijn, A. Examples of examination questions for Cable stayed bridges. December 2005

Romeijn, A. CT5125 Steel Bridges Part 1&2. Faculty of Civil engineering and Geosciences

Thimmardy.E. et.al., New Carquinez bridge. North America’s Newest suspension bridge.
Steel bridge 2004 Millau.

Ulstrup, C., Rating and preliminary analysis of suspension bridges. Journal of structural
engineering, Vol. 119, No.9. September 1993.

93
 Self-anchored suspension bridges: Part II

List of figures and tables

FIGURE 1 TRACE CITY BRIDGE NIJMEGEN ................................................................................................................ 1
FIGURE 2 CROSS SECTION WAAL RIVER NEAR NIJMEGEN ........................................................................................ 1
FIGURE 3 SITUATION AT WAAL RIVER ..................................................................................................................... 2
FIGURE 4 BOX GIRDER WITH EQUIVALENT FLANGE THICKNESSES TO ACCOUNT FOR LONGITUDINAL STIFFENERS .... 2
FIGURE 5 BEARING SYSTEM ..................................................................................................................................... 5
FIGURE 6 WIDTH OF LANES [M] ............................................................................................................................... 6
FIGURE 7 TOTAL DECK WIDTH ................................................................................................................................. 6
FIGURE 8 BOX GIRDER ............................................................................................................................................. 6
FIGURE 9 BOX GIRDER SHAPE .................................................................................................................................. 6
FIGURE 10 TRAFFIC LOADING SCHEME IN TRANSVERSE DIRECTION OF THE DECK .................................................... 7
FIGURE 11 PARABOLIC SHAPE CABLE ...................................................................................................................... 8
FIGURE 12 H- FRAME PYLON FIGURE 13 CROSS SECTION PYLON ............................................................................ 8
FIGURE 14 DEFLECTION OF GIRDER [MM] DUE TO SELF WEIGHT ONLY .................................................................... 9
FIGURE 15 DEFLECTION OF GIRDER [MM] DUE TO SELF WEIGHT AFTER PRETENSIONING THE MAIN CABLE .............. 9
FIGURE 16 GLOBAL BENDING MOMENTS [KNM] IN GIRDER DUE TO SELF WEIGHT .................................................. 10
FIGURE 17 BENDING MOMENTS [KNM] IN GIRDER DUE TO SELF WEIGHT AFTER PRETENSIONING MAIN CABLE ...... 10
FIGURE 18 MODEL WITH BEAM ELEMENTS............................................................................................................. 11
FIGURE 19 PYLON FRAME ...................................................................................................................................... 11
FIGURE 20 DEFORMATION BEAM- AND CABLE ELEMENT........................................................................................ 12
FIGURE 21 MODULUS OF ELASTICITY OF CABLE TYPES .......................................................................................... 12
FIGURE 22 RIGID ARM............................................................................................................................................ 13
FIGURE 23 MODELLING SUPPORT OF STIFFENING GIRDER AT PYLON ...................................................................... 13
FIGURE 24 RESULTING LINE LOAD ON BOX GIRDER FROM DISTRIBUTED TRAFFIC LOAD ......................................... 14
FIGURE 25 SUPERPOSITION OF LINEAR STRESS DISTRIBUTION OVER THE WEB OF THE CROSS SECTION .................. 15
FIGURE 26 DISTRIBUTION STRESSES IN FLANGES ................................................................................................... 16
FIGURE 27 CROSS SECTION BOX GIRDER ................................................................................................................ 16
FIGURE 28 MECHANICAL PROPERTIES BOX GIRDER JIANG YIN BRIDGE.................................................................. 17
FIGURE 29 ANGLE OF MAIN CABLE IN SIDE AND MAIN SPAN .................................................................................. 17
FIGURE 30 ALLOWABLE CABLE STRESS ................................................................................................................. 18
FIGURE 31 SCHEMATIZATION OF CROSS SECTION OF THE PYLON ........................................................................... 20
FIGURE 32 LOAD CASE 1 ........................................................................................................................................ 21
FIGURE 33 LOAD CASE 2 ........................................................................................................................................ 21
FIGURE 34 LOAD CASE 3 ........................................................................................................................................ 21
FIGURE 35 CONSIDERED CROSS SECTION 1-4 ......................................................................................................... 22
FIGURE 36 BENDING MOMENT ALONG THE DECK ................................................................................................... 22
FIGURE 37 DEFLECTIONS ALONG THE DECK ........................................................................................................... 23
FIGURE 38 FINAL DIMENSIONS REFERENCE DESIGN ............................................................................................... 24
FIGURE 39 DESIGN VALUES STEEL ......................................................................................................................... 25
FIGURE 40 LEFT: STRESS DISTRIBUTION WITHOUT SHEAR DEFORMATION. RIGHT: STRESS DISTRIBUTION WITH
SHEAR DEFORMATION................................................................................................................................... 26
FIGURE 41 EXAMPLE BUCKLING FORCE ................................................................................................................ 29
FIGURE 42 CRITICAL WIND SPEED FLUTTER ........................................................................................................... 32
FIGURE 43 FIRST BENDING MODE ........................................................................................................................... 33
FIGURE 44 FREQUENCY SIMPLY SUPPORTED BEAM ................................................................................................ 33
FIGURE 45 MECHANICAL SCHEME PYLON .............................................................................................................. 36
FIGURE 46 BUCKLING FORCE OF FIXED COLUMN AND TRANSLATION SPRING ......................................................... 36
FIGURE 47 1ST AND 2ND ORDER DEFLECTION DUE TO THE THREE CONSIDERED LOADING COMBINATIONS
INCLUDING PRETENSION OF MAIN CABLE ...................................................................................................... 38
FIGURE 48 DECREASING STIFFENING EFFECT ......................................................................................................... 39
FIGURE 49 ADDITIONAL NORMAL FORCE ............................................................................................................... 39
FIGURE 50 1ST AND 2ND ORDER DEFLECTION WITH TRAFFIC FULL LENGTH ........................................................... 40
FIGURE 51 1ST AND 2ND ORDER DEFLECTIONS WITH TRAFFIC MID SPAN ................................................................ 40
FIGURE 52 1ST AND 2ND ORDER DEFLECTIONS WITH TRAFFIC OVER SIDE SPANS .................................................... 40
FIGURE 53 MAXIMUM BENDING MOMENT IN THE GIRDER ...................................................................................... 44
FIGURE 54 CONTRIBUTION OF THE GIRDER TO BENDING MOMENT ......................................................................... 44
FIGURE 55 MAXIMUM DEFLECTION IN THE GIRDER ................................................................................................ 44

94
 Self-anchored suspension bridges: Part II

FIGURE 56 NATURAL FREQUENCIES ....................................................................................................................... 44

FIGURE 57 REACTION FORCES................................................................................................................................ 45
FIGURE 58 STRESSES IN FLANGES OF THE BOX GIRDER .......................................................................................... 45
FIGURE 59 ADDITIONAL NORMAL FORCE APPLIED ON THE GIRDER ........................................................................ 46
FIGURE 60 SPRING SUPPORTED BRIDGE DECK ........................................................................................................ 46
FIGURE 61 EULER BUCKLING FORCE AS A FUNCTION OF THE STIFFNESS OF THE DECK ........................................... 47
FIGURE 62 HINGED AND UNHINGED GIRDER .......................................................................................................... 49
FIGURE 63 MAXIMUM DEFLECTION IN THE GIRDER ................................................................................................ 49
FIGURE 64 NATURAL FREQUENCIES ....................................................................................................................... 49
FIGURE 65 MAXIMUM BENDING MOMENT IN THE GIRDER ..................................................................................... 50
FIGURE 66 CONTRIBUTION OF THE GIRDER TO BENDING MOMENT ......................................................................... 50
FIGURE 67 MAXIMUM DEFLECTION IN THE GIRDER ................................................................................................ 50
FIGURE 68 NATURAL FREQUENCIES ....................................................................................................................... 50
FIGURE 69 MAXIMUM BENDING MOMENT IN THE GIRDER ...................................................................................... 51
FIGURE 70 CONTRIBUTION OF THE GIRDER TO BENDING MOMENT ......................................................................... 51
FIGURE 71 MAXIMUM DEFLECTION IN THE GIRDER ................................................................................................ 51
FIGURE 72 NATURAL FREQUENCIES ....................................................................................................................... 51
FIGURE 73 REACTION FORCES................................................................................................................................ 52
FIGURE 74 MAXIMUM BENDING MOMENT IN THE GIRDER ...................................................................................... 52
FIGURE 75 CONTRIBUTION OF THE GIRDER TO BENDING MOMENT ......................................................................... 52
FIGURE 76 MAXIMUM DEFLECTION IN THE GIRDER ................................................................................................ 52
FIGURE 77 NATURAL FREQUENCIES ....................................................................................................................... 52
FIGURE 78 REACTION FORCES................................................................................................................................ 53
FIGURE 79 MAXIMUM BENDING MOMENT IN THE GIRDER ...................................................................................... 53
FIGURE 80 CONTRIBUTION OF THE GIRDER TO BENDING MOMENT ......................................................................... 53
FIGURE 81 MAXIMUM DEFLECTION IN THE GIRDER ................................................................................................ 53
FIGURE 82 NATURAL FREQUENCIES ....................................................................................................................... 53
FIGURE 83 HALF LOADED BRIDGE DECK IN TRANSVERSE DIRECTION .................................................................... 54
FIGURE 84 ONE SIDED FULL LENGTH LOADING ...................................................................................................... 55
FIGURE 85 ONE SIDE MID SPAN LOADED ................................................................................................................ 55
FIGURE 86 ALTERNATE LOADING FULL LENGTH .................................................................................................... 56
FIGURE 87 BENDING MOMENTS ALONG THE GIRDER IN REFERENCE MODEL AND OPTIMIZED MODEL ..................... 60
FIGURE 88 DEVELOPMENT ON UNITY CHECK DEFLECTION MAIN SPAN .................................................................. 63
FIGURE 89 STRESSES IN COMPRESSION FLANGES ................................................................................................... 64
FIGURE 90 STRESSES IN TENSILE FLANGES ............................................................................................................. 64
FIGURE 91 DEVELOPMENT OF THE MAXIMUM BENDING MOMENT IN THE GIRDER .................................................. 65
FIGURE 92 CONTRIBUTION OF THE NORMAL COMPRESSIVE STRESS IN THE TOP FLANGE IN THE MAIN SPAN ........... 66
FIGURE 93 STRESS LEVEL IN MAIN CABLE AND HANGERS...................................................................................... 66
FIGURE 94 BENDING MOMENT AT PYLON BASE ...................................................................................................... 67
FIGURE 95 ADDITIONAL F TO DETERMINE BUCKLING FORCE ................................................................................. 67
FIGURE 96 DEVELOPMENT OF THE EULER BUCKLING FORCE DECK AS A FUNCTION OF AN INCREASING MAIN SPAN 67
FIGURE 97 CRITICAL WIND SPEED FOR FLUTTER AND VORTEX AS A FUNCTION OF THE MAIN SPAN ........................ 69
FIGURE 98 REACTION FORCES IN END SUPPORTS .................................................................................................... 70
FIGURE 99 CABLE ANCHORAGE KONOHANA BRIDGE ............................................................................................ 71
FIGURE 100 TOTAL AMOUNT OF STEEL .................................................................................................................. 72
FIGURE 101 STEEL USE PER CATEGORY ................................................................................................................. 72
FIGURE 102 COMPRESSIVE STRUT IN EACH SIDE SPAN ........................................................................................... 80
FIGURE 103 UTILIZING THE ERECTED SIDE SPAN .................................................................................................... 81
FIGURE 104 TEMPORARY EXTERNAL ANCHORAGE ................................................................................................. 81
FIGURE 105 TEMPORARY SUPPORTS WAAL RIVER ................................................................................................. 82
FIGURE 106 CANTI LEVERED DECK SECTIONS ........................................................................................................ 83
FIGURE 107 BENDING MOMENTS [KNM] WITH TEMPORARY SUPPORTS AT 210M DISTANCE ................................... 84
FIGURE 108 REQUIRED LENGTH FOR ADDITIONAL PLATE THICKNESSES ................................................................. 85
FIGURE 109 LIFTING SEQUENCE OF DECK SECTIONS & ZERO BENDING LOCATIONS IN M LINE ............................... 86
FIGURE 110 ERECTING WITH TEMPORARY STAY CABLES ....................................................................................... 87
FIGURE 111 ERECTING WITH TEMPORARY STAYS AND TEMPORARY SUPPORTS IN THE SIDE SPAN .......................... 87
FIGURE 112 ERECTING SIDE SPAN FIRST................................................................................................................. 88
FIGURE 113 CANTILEVER PROCESS FROM PYLON ................................................................................................... 88

95
 Self-anchored suspension bridges: Part II

Tables:
TABLE 1 CONCENTRATED AND UNIFORM DISTRIBUTED LOADS ................................................................................ 7
TABLE 2 MECHANICAL PROPERTIES STIFFENING GIRDER ....................................................................................... 16
TABEL 3 DIMENSIONS REFERENCE DESIGN............................................................................................................. 23
TABLE 4 PROPERTIES OF PYLON ............................................................................................................................. 36
TABLE 5 FIRST BENDING FREQUENCY COMPARISON .............................................................................................. 42
TABLE 6 NATURAL FREQUENCIES REFERENCE DESIGN ........................................................................................... 42
TABLE 7 GIRDER MECHANICAL PROPERTIES .......................................................................................................... 44
TABLE 8 N-VALUES FOR MAIN AND SIDE SPAN RELATED TO THE MOMENT OF INERTIA OF THE BOX GIRDER .......... 48
TABLE 9 CABLE AREA ............................................................................................................................................ 51
TABLE 10 DECK FULL LOADED ECCENTRICITY....................................................................................................... 54
TABLE 11 ASYMMETRIC LOADING TRAFFIC FULL LENGTH ..................................................................................... 55
TABLE 12 ASYMMETRIC LOADING TRAFFIC MID SPAN ........................................................................................... 55
TABLE 13 ALTERNATE LOADING TRAFFIC FULL LENGTH........................................................................................ 56
TABLE 14 MEMBER DIMENSIONS OPTIMIZED MODEL ............................................................................................. 60
TABLE 15 MAIN RESULTS OF INCREASING SPAN ..................................................................................................... 63
TABLE 16 GIRDER PROPERTIES FOR THE CONSIDERED SPAN LENGTHS ................................................................... 64
TABLE 17 MAIN CABLE AND HANGER PROPERTIES ................................................................................................. 66
TABLE 18 N-VALUES FOR MAIN AND SIDE SPAN ..................................................................................................... 68
TABLE 19 FREQUENCY RATIO FOR EACH SPAN LENGTH ......................................................................................... 69
TABLE 20 STEEL USE PER METER ........................................................................................................................... 72
TABLE 21 ACHIEVABLE FREE SPAN LENGTH AND NUMBER OF REQUIRED TEMPORARY SUPPORTS .......................... 74
TABLE 22 REQUIRED LIFTING CAPACITY ................................................................................................................ 74
TABLE 23 STRESSES BOX GIRDER ERECTION PHASE, LOCATION 1.......................................................................... 84
TABLE 24 STRESSES BOX GIRDER ERECTION PHASE LOCATION 2 ........................................................................... 84
TABEL 25 STRESSES AT LOCATION 1 WITH ADDITIONAL PLATE THICKNESS ........................................................... 85

96
 Self-anchored suspension bridges: Part II

Appendices

Content:

Appendix 1 Cable modelling

Appendix 2a Effective width at main and side span
Appendix 2b Effective width at support
Appendix 3a Girder influence
Appendix 3b Cable influence
Appendix 3c Sag over span length influence
Appendix 3d Pylon influence
Appendix 3e Number of hinges in girder
Appendix 4 Comparable box girders
Appendix 5a Stability check reference model
Appendix 5b Stability check Optimized model
Appendix 6 Design orthotropic steel box of the New Carquinez Bridge
Appendix 7 Stress calculation in reference model
Appendix 8a Erection of Konohana Bridge
Appendix 8b Erection of Yeoungjong Grand Bridge
Appendix 8c Erection of Nescio bridge

97
 Self-anchored suspension bridges: Part II

Appendix 1 Cable modelling

Deflections [mm] of model with cable elements

Deflections [mm] of model with a ‘chain’ as main cable

My;Ed [kNm] of model with cable elements

My;Ed [kNm] of model with a ‘chain’ as main cable

98
 Self-anchored suspension bridges: Part II

Appendix 2a Effective width at main and side span

Effective width of girder at main span

99
Self-anchored suspension bridges: Part II

100
 Self-anchored suspension bridges: Part II

Appendix 2b Effective width at support

Effective width of girder at support pylon

101
Self-anchored suspension bridges: Part II

102
 Self-anchored suspension bridges: Part II

Appendix 3a Girder influence

Iy δ δ My
box N H main pylon main My sup 1st 2nd 1st 2nd
[m^4] deck cable Rz 2nd 2nd 2nd 2nd fb fb ft ft
1,10 35502 18159 6975 474 120 104641 137291 0,63 1,1 4,58 3,97
1,26 34935 17863 6822 453 115 113262 147528 0,65 1,16 4,71 4,02
1,42 34388 17576 6673 433 110 121618 157416 0,67 1,23 4,83 4,07
1,59 33863 17300 6530 414 105 129681 166928 0,68 1,29 4,94 4,11
1,77 33359 17036 6392 395 100 137455 176048 0,7 1,35 5,04 4,14
1,96 32878 16783 6260 377 95 144910 184711 0,72 1,41 5,13 4,19
2,16 32420 16542 6134 361 91 152065 193096 0,74 1,46 5,2 4,19
2,37 31983 16312 6014 345 87 158907 201026 0,76 1,52 5,31 4,21
2,59 31568 16094 5899 330 83 165448 208572 0,77 1,58 5,37 4,23
2,82 31174 15886 5789 315 79 171690 215744 0,79 1,63 5,44 4,24
3,06 30801 15690 5685 302 76 177658 222556 0,81 1,68 5,49 4,26
3,31 30447 15503 5586 289 73 183345 229023 0,83 1,73 5,55 4,27
3,57 30111 15326 5491 277 69 188779 235159 0,84 1,75 5,59 4,28
3,85 29793 15158 5401 265 67 193962 240982 0,86 1,83 5,63 4,29
4,13 29492 14999 5316 254 64 198908 246507 0,88 1,87 5,67 4,3
4,42 29206 14849 5234 244 61 203631 251751 0,89 1,92 5,71 4,31

*=Reference model

103
 Self-anchored suspension bridges: Part II

Appendix 3b Cable influence

δ δ My
E cable N H main pylon main My sup 1st 2nd 2nd
[N/mm^2] deck cable Rz 2nd 2nd 2nd 2nd fb fb 1st ft ft
205000 32420 16542 6134 361 91 152065 193096 0,74 1,46 5,2 4,19
190000 31226 15977 5763 372 94 166998 211426 0,73 1,46 5,15 4,1
150000 27504 14219 4609 408 101 213485 268570 0,7 1,46 4,93 3,83
140000 26425 13711 4275 418 104 226931 285121 0,69 1,46 4,87 3,76

δ My My
A cable N H δ main pylon main sup 1st 2nd 2nd
[mm^2] deck cable Rz 2nd 2nd 2nd 2nd fb fb 1st ft ft
7854 18125 9806 1746 498 119 324316 408532 0,65 1,48 4,53 3,59
11310 23381 12274 3353 447 110 262660 328974 0,68 1,47 4,88 3,85
15394 28174 14533 4826 401 100 203917 256745 0,71 1,47 5,11 4,05
17671 30367 15559 5502 380 95 177113 223824 0,72 1,47 5,17 4,13
20106 32420 16542 6134 361 91 152065 193096 0,74 1,46 5,2 4,19
22698 34334 17451 6725 342 87 128715 164500 0,75 1,46 5,26 4,23
25447 36116 18299 7274 326 83 107047 137987 0,77 1,46 5,25 4,26
28353 37771 19088 7785 310 79 86964 113452 0,78 1,46 5,19 4,27
31416 39307 19822 8259 296 75 68391 90785 0,8 1,45 5,15 4,27
34636 40733 20505 8698 282 73 51232 69860 0,81 1,45 5,09 4,25
38013 42056 21139 9105 270 70 35380 50556 0,82 1,45 5,03 4,22

*=Reference model

104
 Self-anchored suspension bridges: Part II

Appendix 3c Sag over span length influence

δ δ My My
main pylon main sup 1st 2nd 1st 2nd
Sag/L N deck H cable Rz 2nd 2nd 2nd 2nd fb fb ft ft
1/5 32420 16542 6134 361 91 152065 193096 0,74 1,46 5,2 4,19
1/6 35545 18268 5739 400 83 186776 235639 0,71 1,47 5,82 4,74
1/7 37633 19491 5280 437 77 221178 277931 0,68 1,47 6,18 5,18
1/8 38934 20326 4808 471 73 253476 317643 0,66 1,47 6,4 5,52
1/9 39665 20871 4352 500 68 282937 353788 0,64 1,47 6,51 5,8

Appendix 3d Pylon influence

δ δ My
I pyloon N main pylon main My sup 1st 2nd 1st 2nd
[m^4] deck H cable Rz 2nd 2nd 2nd 2nd fb fb ft ft
0,238 32624 16428 6295 366 94 155668 196633 0,73 1,46 4,95 4,06
0,278 32593 16451 6269 365 94 154945 195894 0,73 1,46 5,03 4,09
0,321 32558 16474 6240 364 93 154225 195172 0,73 1,46 5,1 4,12
0,368 32518 16497 6207 363 92 153507 194465 0,74 1,46 5,15 4,15
0,419 32474 16519 6172 362 92 152780 193771 0,74 1,46 5,18 4,17
0,474 32420 16542 6134 361 91 152065 193096 0,74 1,46 5,2 4,19
0,533 32362 16564 6093 359 90 151328 192414 0,74 1,46 5,21 4,21
0,596 32299 16586 6050 358 89 150593 191747 0,74 1,46 5,28 4,23
0,664 32232 16608 6003 357 89 149845 191088 0,74 1,47 5,3 4,25
0,735 32162 16631 5955 355 88 149106 190432 0,74 1,47 5,31 4,26
0,812 32082 16653 5903 354 87 148354 189786 0,75 1,47 5,32 4,28

Appendix 3e Number of hinges in girder

δ δ My My
main pylon main sup 1st 2nd 1st 2nd
hinges N deck H cable Rz 2nd 2nd 2nd 2nd fb fb ft ft
0 32420 16542 6134 361 91 152065 193096 0,74 1,46 5,2 4,19
2 38278 19542 6269 472 120 162788 0 0,67 1,31 3,56 3,72

*=Reference model

105
 Self-anchored suspension bridges: Part II

Appendix 4 Comparable box girders

Bron: Dictaat Romeijn Stalen bruggen blz 84:

106
 Self-anchored suspension bridges: Part II

Appendix 5a Stability check reference model

107
 Self-anchored suspension bridges: Part II

Appendix 5b Stability check Optimized model

108
 Self-anchored suspension bridges: Part II

Appendix 6 Design orthotropic steel box of the New Carquinez Bridge

The stiffeners of this cross section can be expressed as a percentage of the area of the top
flange, bottom flange and web:
A stiff. top flange = 65% * A top flange
A stiff . bottom flange = 35% * A bottom flange
A stiff. web = 15 % * A web

This estimation is retrieved from a similar box girder of the same dimensions: Thimmardy.E. et.al., New
Carquinez bridge. North America’s Newest suspension bridge. Steel bridge 2004 Millau.

109
 Self-anchored suspension bridges: Part II

Appendix 7 Stress calculation in reference model

110
Self-anchored suspension bridges: Part II

111
 Self-anchored suspension bridges: Part II

Appendix 8a Erection of Konohana Bridge

112
 Self-anchored suspension bridges: Part II

Appendix 8b Erection of Yeoungjong Grand Bridge

113
Self-anchored suspension bridges: Part II

114
 Self-anchored suspension bridges: Part II

Appendix 8c Erection of Nescio bridge

115