2.b. COMBRI Design Manual Part II English

COMBRI DESIGN MANUAL
Part II: State-of-the-Art and Conceptual Design

of Steel and Composite Bridges
A project carried out with a financial grant from the Research Fund for Coal and Steel (RFCS)
of the European Community.
COMBRI DESIGN MANUAL
Part II: State-of-the-Art and Conceptual Design

of Steel and Composite Bridges
A project carried out with a financial grant from the Research Fund for Coal and Steel (RFCS)
of the European Community.
Although all care has been taken to ensure the integrity and quality of this publication and the
information herein, no liability is assumed by the project partners and the publisher for any damage to
property or persons as a result of the use of this publication and the information contained herein.
1st Edition
Copyright © 2008 by project partners
Reproduction for non-commercial purpose is authorised provided the source is acknowledged and
notice is given to the project coordinator. Publicly available distribution of this publication through other
sources than the web sites given below requires the prior permission of the project partners. Requests
should be addressed to the project coordinator:
Universität Stuttgart
Institut für Konstruktion und Entwurf / Institute for Structural Design
Pfaffenwaldring 7
70569 Stuttgart
Germany
Phone: +49-(0)711-685-66245
Fax: +49-(0)711-685-66236
E-mail: [email protected]
The present document and others related to the research project COMBRI+RFS-CR-03018
“Competitive Steel and Composite Bridges by Improved Steel Plated Structures” and the successive
dissemination project RFS2-CT-2007-00031 “Valorisation of Knowledge for Competitive Steel and
Composite Bridges”, which have been co-funded by the Research Fund for Coal and Steel (RFCS) of
the European Community, can be accessed for free on the following project partners’ web sites:
Belgium: www.argenco.ulg.ac.be
France: www.cticm.com
Germany: www.uni-stuttgart.de/ke, www.stb.rwth-aachen.de
Spain: www.labein.es, www.apta.org.es
Sweden: cee.project.ltu.se
Cover pictures (from left to right):

Valley bridge Haseltal near Suhl, Germany, 2006 (© KE)
Valley bridge Dambachtal near Suhl, Germany, 2005 (© KE)
Viaduct over Dordogne river near Souillac, France, 2000 (© Sétra)
2
Preface
This design manual is an outcome of the research project RFS-CR-03018 “Competitive Steel and
Composite Bridges by Improved Steel Plated Structures - COMBRI” [15] and the successive
Composite Bridges - COMBRI+” which have been funded by the Research Fund for Coal and Steel
(RFCS) of the European Community. Within the RFCS research project essential knowledge has been
acquired to enhance the competitiveness of steel and composite bridges and this has been incorporated
in the design manual at hand which has been also presented in the frame of several seminars and
workshops. The manual is subdivided into two parts to provide the reader with clearly arranged and
concise documents:
► Part I: Application of Eurocode rules
In the research project the different national background of each partner how to apply and interprete
Eurocode rules was brought together and a European melting pot of background information and
general knowledge has been created. In order to maintain this valuable information two composite
bridge structures - a twin-girder and a box-girder bridge - are covered in Part I of the COMBRI Design
Manual [16] on the basis of worked examples for which the knowledge is written down in a descriptive
manner. The examples include references to current Eurocode rules.
► Part II: State-of-the-Art and Conceptual Design of Steel and Composite Bridges
The national state-of-the-art in bridge design can be different so that firstly bridge types of the project
partners’ countries - Belgium, France, Germany, Spain and Sweden - are introduced. They reflect the
current practice in those countries and common bridge types as well as unusual bridges intended to
solve special problems and some solutions being part of development projects are presented in this part
of the COMBRI Design Manual. Also, improvements which can be provided to the design of steel and
composite bridges are discussed and the possibilities and restrictions given by the current Eurocode
rules are highlighted.
Moreover, the features of software EBPlate [26] developed in the research project to determine the
elastic critical buckling stresses are presented in its contributive application for bridge design.
Finally, the authors of this design manual gratefully acknowledge the support and financial grant of the
Research Fund for Coal and Steel (RFCS) of the European Community.
Ulrike Kuhlmann, Benjamin Braun

Universität Stuttgart, Institute for Structural Design / Institut für Konstruktion und Entwurf (KE)
Markus Feldmann, Johannes Naumes
RWTH Aachen University, Institute for Steel Structures
Pierre-Olivier Martin, Yvan Galéa
Centre Technique Industriel de la Construction Métallique (CTICM)
Bernt Johansson, Peter Collin, Jörgen Eriksen
Luleå University of Technology, Division of Steel Structures (LTU)
Hervé Degée, Nicolas Hausoul
Université de Liège, ArGEnCo Département
José Chica
Fundación LABEIN
Joël Raoul, Laurence Davaine, Aude Petel
Service d'études sur les transports, les routes et leurs aménagements (Sétra)
October 2008
Table of Contents
Table of Contents
Page
1 Introduction and scope .......................................................................................................................1
2 Bridge types .......................................................................................................................................3
2.1 General....................................................................................................................................3
2.2 Belgium...................................................................................................................................4
2.2.1 Introduction ...............................................................................................................4
2.2.2 Road bridges in Walloon Region...............................................................................5
2.2.3 Railway bridges in Belgium ......................................................................................6
2.2.4 Bridge examples ........................................................................................................8
2.2.4.1 Composite girder bridge.............................................................................8
2.2.4.2 Composite box-girder bridges..................................................................12
2.3 France ...................................................................................................................................12
2.3.1 Introduction .............................................................................................................12
2.3.2 Types of composite bridges in France.....................................................................14
2.3.2.1 General .....................................................................................................14
2.3.2.2 Twin-girder bridges..................................................................................14
2.3.2.3 Examples of composite box-girder bridges..............................................19
2.4 Germany................................................................................................................................20
2.4.1 Introduction .............................................................................................................20
2.4.2 Bridges with a one-piece composite superstructure ................................................23
2.4.2.1 General .....................................................................................................23
2.4.2.2 Influence of the cross-sectional layout.....................................................24
2.4.3 Bridges with airtight small-size box-girders ...........................................................29
2.4.4 Bridges with prefabricated components and in-situ concrete..................................31
2.4.4.1 General .....................................................................................................31
2.4.4.2 Steel girders and partial depth precast concrete elements........................32
2.4.4.3 Prefabricated composite girders...............................................................34
2.5 Spain .....................................................................................................................................36
2.5.1 Introduction .............................................................................................................36
2.5.2 Spanish regulatory frame for bridge design ............................................................38
2.5.3 Bridge examples ......................................................................................................40
2.6 Sweden..................................................................................................................................48
2.6.1 Introduction .............................................................................................................48
2.6.2 Road bridges............................................................................................................49
I
COMBRI Design Manual - Part II
2.6.3 Bridges with integral abutments ..............................................................................50

2.6.4 Bridges with full depth prefabricated deck slabs.....................................................50
2.6.5 Railway bridges .......................................................................................................52
2.6.6 Special bridges.........................................................................................................53
3 Steel grades ......................................................................................................................................57
3.1 Introduction...........................................................................................................................57
3.2 Discussions about the use of High Strength Steel (HSS) in bridge design...........................57
3.2.1 General ....................................................................................................................57
3.2.2 Influence of the deflection limitations.....................................................................58
3.2.3 Influence of buckling and possible use of hybrid girders........................................59
3.2.4 Conclusions .............................................................................................................61
3.3 Steel grades used in European countries...............................................................................61
3.3.1 Use in France...........................................................................................................61
3.3.2 Use in Germany.......................................................................................................62
3.3.3 Use in other European countries..............................................................................64
3.4 Through-thickness properties................................................................................................64
3.5 Application to the calculation example “box-girder bridge“ ................................................65
4 Flanges .............................................................................................................................................69
4.1 Introduction...........................................................................................................................69
4.2 I-girders.................................................................................................................................69
4.3 Bottom plate of box-girders ..................................................................................................70
4.3.1 General ....................................................................................................................70
4.3.2 French practice ........................................................................................................70
4.3.3 German practice.......................................................................................................71
4.3.4 Swedish practice......................................................................................................71
4.3.5 Conclusions and design recommendations..............................................................72
4.4 Double-composite action ......................................................................................................73
4.4.1 General ....................................................................................................................73
4.4.2 Use in France...........................................................................................................73
4.4.3 Use in Germany.......................................................................................................73
4.4.4 Conclusions and design recommendations..............................................................76
5 Webs.................................................................................................................................................79
5.1 Introduction...........................................................................................................................79
5.2 Transverse stiffeners .............................................................................................................79
5.3 Transition between unstiffened and longitudinally stiffened webs.......................................82
5.4 Type of longitudinal stiffener and continuity .......................................................................84
5.4.1 General ....................................................................................................................84
5.4.2 Single flat longitudinal stiffener ..............................................................................84
5.4.3 Closed shape longitudinal stiffener .........................................................................85
II
Table of Contents
5.4.4 Discontinuous longitudinal stiffeners......................................................................85

5.4.5 Exterior longitudinal stiffeners................................................................................86
6 Cross bracings and diaphragms........................................................................................................89
6.1 Introduction...........................................................................................................................89
6.2 Cross bracings in I-girders ....................................................................................................89
6.2.1 General ....................................................................................................................89
6.2.2 Standard design for the calculation example “twin-girder bridge” .........................91
6.2.2.1 Stiffness of the bracing frames.................................................................91
6.2.2.2 Spacing of the bracing frames..................................................................92
6.2.2.3 Verification against lateral torsional buckling .........................................92
6.2.3 Improved design for the calculation example “twin-girder bridge” ........................93
6.2.3.1 Cross-bracing characteristics ...................................................................93
6.2.3.2 Verification against lateral torsional buckling .........................................94
6.2.3.3 Verification against buckling of the members .........................................94
6.2.3.4 Choice of the reduction curve for lateral torsional buckling....................95
6.2.3.5 Cross-bracings in the sagging moment region .........................................96
6.3 Diaphragms in box-girders ...................................................................................................97
7 Launching of steel- and composite bridges......................................................................................99
7.1 Introduction...........................................................................................................................99
7.2 Local behaviour: introduction of the transverse load..........................................................100
7.2.1 General ..................................................................................................................100
7.2.2 Launching shoes ....................................................................................................100
7.2.2.1 General ...................................................................................................100
7.2.2.2 Wheels on balancing device...................................................................101
7.2.2.3 Wheels and cable ...................................................................................101
7.2.2.4 Balancers................................................................................................102
7.2.3 Sliding skates.........................................................................................................102
7.2.4 Other devices .........................................................................................................103
7.3 Global behaviour during launching ....................................................................................103
7.3.1 General ..................................................................................................................103
7.3.2 Launching nose......................................................................................................104
7.3.3 Temporary cross-bracings .....................................................................................104
7.4 Launching with a part of the concrete slab .........................................................................104
7.4.1 General ..................................................................................................................104
7.4.2 Application to the calculation example “twin-girder bridge”................................106
8 Summary ........................................................................................................................................109
References ..............................................................................................................................................111
List of figures .........................................................................................................................................117
List of tables ...........................................................................................................................................121
III
Introduction and scope
1 Introduction and scope

This design manual is based on results from the research project RFS-CR-03018 „Competitive Steel
and Composite Bridges by Improved Steel Plated Structures - COMBRI” [15] and the subsequent
Composite Bridges - COMBRI+”, both sponsored by the Research Fund for Coal and Steel (RFCS). It
is focused on the conceptual design of steel bridges and the steel parts of composite bridges and it is
based on the rules in EN 1993-1-5, EN 1993-2 and EN 1994-2. Design of steel bridges is a very wide
field and it can not be covered completely in this manual and a selection of topics has been made.
Part I of the COMBRI Design Manual [16] deals with the calculation of two composite bridges
according to the Eurocodes. This present Part II deals with the state of the art in several countries and
focus on some improvements that can be provided to the design of steel and composite bridges.
Chapter 2 gives an overview of bridge types in the countries participating in the project, Belgium,
France, Germany, Spain and Sweden. It reflects the current practice in those countries and presents
common bridge types as well as unusual bridges intended to solve special problems and some solutions
being parts development projects. There are notable differences between the practices of the countries
and the solutions presented may serve as inspiration.
Chapter 3 deals with the choice of steel grades. EN 1993-1-12 extends the range of permitted steel
grades of EN 1993 up to S700 but in most cases such high grades are not feasible. The problem is
usually that the fatigue requirements limit the full utilization of the strength. It is shown that hybrid
girders with higher strength in the flange than in the webs are economic.
Flanges are dealt with in Chapter 4 and the main topic is bottom flanges in box-girders. Such flanges
are in most cases stiffened and different types of stiffeners are discussed. The design rules of EN 1993-
1-5 may lead to unsafe results for very light stiffening and minimum stiffness of the stiffeners is
recommended. Further, double composite action with both top and bottom flanges being composite are
described and recommendations for design are given.
Webs are discussed in Chapter 5 and the main issue is to what extent stiffeners should be used. If the
method with effective cross section in EN 1993-1-5 is applied it is shown that longitudinal stiffeners are
not economical for web depths below 4 m.
Chapter 6 covers cross bracings and diaphragms for I-girder bridges and box-girders. Functional
requirements are described and ways to meet them are discussed. For economy it is important to
minimize the man hours for fabricating the bracings.
Launching has been studied in some detail in the COMBRI research project and it is dealt with in
Chapter 7. Particularly the resistance patch loading has been studied and the result is improved design
rules, which will be proposed for inclusion in EN 1993-1-5. The rules allow the utilisation of quite long
loaded lengths and accordingly quite high resistance. This makes it possible to launch bridges with the
concrete slab in place.
1
Bridge types
2 Bridge types
2.1 General
This chapter gives an overview of bridge types in the countries participating in the project, Belgium,
France, Germany, Spain and Sweden. It reflects the current practice in those countries and presents
common bridge types as well as unusual bridges intended to solve special problems and some solutions
being parts development projects. There are notable differences between the practices of the countries
and the solutions presented may serve as inspiration. There are however also similarities and some will
be mentioned here.
Pure steel bridges are unusual in all countries with some variations in the frequency. They are mainly
used as decks in suspension bridges, cable stayed bridges and movable bridges. The focus here will be
on composite bridges. In most cases a twin I-girder composite bridges is the most economical solution.
In Europe this design is deemed safe enough although the fact that if one girder fails the whole bridge
will fall down. In the US twin-girder are allowed but the safety factors are higher, which makes twin-
girder bridge less competitive. With equal safety factors a twin-girder bridge will be cheaper than a
multiple girder bridge, which needs more material and also more hours for fabrication and erection. An
exception is if the depth of the bridge is restricted, which may make multiple girders the best
alternative, especially if the span is so short that rolled beams can be used. If there are no restrictions on
the girder depth the span to depth ratio is normally chosen in the range 20 to 30. The higher end of the
range applies to interior spans in continuous bridges and the lower end to simply supported bridges and
end spans. If the support locations can be chosen without restrictions it is favourable to make the end
spans of a continuous bridge shorter 0,60 to 0,85 times the length of the interior spans.
The concrete deck of a twin-girder bridge can be reinforced up to a width of approximately 13 m and
for wider decks transverse prestressing is commonly used. Another solution for extending the use of
reinforced concrete is to reduce the span in the transverse direction by putting a small beam supported
by the cross braces in the middle between the main girders. The width of a reinforced concrete deck can
also be made wider by the use of I cross-girders connected to the slab and extended with cantilevers
outside the main girders, see Figure 2-1.
© Sétra © Sétra
Figure 2-1: Wide I-girder bridge with Figure 2-2: Box-girder bridge with edge beams
cantilevering cross girders (Bridge near supporting the deck. (Verrières viaduct near
Remoulins, France). Millau, France, 2002).
Box-girders have an advantage of very high torsional stiffness, which is useful in curved bridges. It
requires less depth than I-girders. Also for a straight bridge there is an advantage in that eccentric traffic
loads are carried by the whole cross section. This leads to less material than for I-girders but it is
3
normally counteracted by higher costs for fabrication and erection. If a box-girder is chosen it should
normally be a single cell box. Multiple cell boxes are sometimes used with the purpose to bring down
the transverse span of the deck but if that is needed it is a better idea to use a small longitudinal beam
supported on the cross braces. An alternative is to use beams supporting the cantilevering part of the
deck as shown in Figure 2-2. Box-girders are made of thin plates and they are sensitive to cross
sectional distortion. This has to be counteracted by diaphragms or cross braces. Cross braces are usually
the cheapest alternative.
The transportation of bridge parts from work shop to erection site sometimes put restrictions on the
design. Ideally the erection site and the work shop have access to navigable water and very large pieces
can be transported. One extreme example is the Öresund bridge between Denmark and Sweden where
140 m long spans with 11 m deep trusses and concrete deck were shipped from Spain. If transport on
road is necessary there are restrictions that vary from country to country and those restrictions may
influence the optimal design. It may be more economical to use a lower girder than to make a
longitudinal splice on the erection site.
Multiple span bridges can be made either continuous or as series of simply supported spans. The latter
may lead to lower initial investment but they require joints at all supports and the maintenance cost of
those joints leads to a higher life cycle cost. Further, it is less comfortable to drive on a series of simple
spans. The conclusion is that a continuous bridge is preferable.
2.2 Belgium
2.2.1 Introduction
In Belgium, the road and bridge facilities are managed or co-managed by regional administrations.
There are 3 regions in Belgium (see Figure 2-3):
• Region of Brussels (In 2008: 161,4 km², 1067162 inhabitants, 6601 inhab/km²)
• Flemish Region (In 2008: 13522 km², 6117440 inhabitants, 442 inhab/km²)
• Walloon Region (In 2008: 16844 km², 3435879 inhabitants, 202 inhab/km²)
Belgium Region of
Brussels
Flemish region Walloon region
Figure 2-3: Three regions in Belgium (Region of Brussels, Flemish and Walloon region).
This report contains data dealing only from the Walloon Region and provided by the MET
administration (Regional Ministry of Equipment and Transportation - MET). These data are not
exhaustive but give a good tendency on the set of bridges managed by Walloon Region. This tendency
can roughly be extrapolated to the whole Belgium.
4
Bridge types
Concerning the railway bridges in Belgium, the data are coming from results elaborated during a FP6
European project dealing with sustainable development, global change & ecosystems (Sustainable
Bridges – Assessment for Future Traffic Demands and Longer Lives). The data presented within this
report have been collected from the 17 European railway administrations, including the Belgian railway
administration NMBS-SNCB.
2.2.2 Road bridges in Walloon Region

In june 2002, the Walloon Region was managing, in all or partly, 3250 bridges. The most important part
of these bridges is linked to road and road-transport networks. Additionally, some bridges are managed
in sharing between MET and other organisms or institutions, such as for example SNCB (National
Railway Company of Belgium). These bridges present a great variety, by their mode of construction
(i.e. type of structure) and by the materials used. This diversity is illustrated by the following Figure 2-4
and Figure 2-5 which give the distribution in terms of types of structure and materials.
"Ponts Moveable
Special
hirondelles" 1.4%
1,6% Concrete
Tied-arch 2.0% Timber
Slab Brick 5.0%
1.8% 0.1%
8.9% 12.1%
True arch
2.4% Tee-beam Composite
Arched / Vault 19.2% Reinforced
4.9%
20.2% concrete
Steel 43.5%
Frame 9.6%
Box-girder
Portal / Inclined 36.8%
4.5%
pier frame Prestressed
1.2% concrete
24.8%
Figure 2-4: Distribution of existing structural types of Figure 2-5: Distribution of existing
road bridges in the Walloon region. bridge types of road bridges in the
Walloon region.
76 % of the bridges in the Walloon Region are frame bridges, arches bridges or beam bridges. The most
common material is concrete, either reinforced (44 %), prestressed (25%), unreinforced (5 %) or used in
combination with steel (composite bridges). Composite bridges represent 5 % of all the bridges in
Walloon region (around 160 composite bridges).
180
160
Number of built bridges
140
120
100
80
60
40
20
0
85
90
95
00
05
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
00
18
18
18
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
20
Year
Figure 2-6: Number of bridge openings per year in the Walloon region.
An important element in the long-term management of bridges is the age of bridges. The year of
construction is illustrated for the bridges currently in service on Figure 2-6. Figure 2-6 is interesting on
several aspects:
5
• A significant increase of bridges is observed from 1935. This is related with the appearance of
the first highways in Belgium.
• Between 1940 and 1945, the number of bridges bought is reduced due to 2nd world war is
observed.
• In the following years, a slow and regular increase of the number of bridges in service is
observed. A maximum is reached in years 1970-1975, followed by a slight fall probably due to
the crisis which followed the first oil crisis in 1973-1974.
• Since 1985, a very significant reduction of the number of bridges put into service can be
noticed. This fall finds its reason on one hand in the more severe economic context of these last
years, but also in a saturation of the road network.
• In present time, the building of a new bridge in Belgium is rare. The main activity is
replacement, repairing or reconstruction of old bridges. The number of newly bought bridges is
thus expected to remain very low for the next years.
Another interesting data are the overall length of bridges. This information gives an idea of the size
(and indirectly of the cost) of the bridges. The distribution is plot on Figure 2-7:
50%
25.6%
25%
16.5% 16.8% 15.6% 16.4%
6.0%
2.4%
0.6% 0.1%
0%
5
00
00
10
00
20
50
-5
-2
<
10
10
5-
-1
10
20
0-
0-
50
0-
>
10
20
50
Bridge length [m]
Figure 2-7: Distribution of bridge lengths of road bridges in the Walloon region.
25 % of Walloon bridges have a total length less than 5 m. 49 % of Walloon bridges have a total length
ranging between 5 and 50 m, which correspond to the most common bridges included in the road-
transport network. A rather significant number of bridges between 50 and 100 m is observed, essentially
on highways, for the crossing of secondary valleys. Finally, 9 % of Walloon bridges exceeds 100 m
long. This distribution of bridge lengths is rather representative on one hand of the relief of the country
(some big fluvial valleys, a significant number of secondary valleys and a big amount of small
watercourses), but also of the high density of the communication network, which implies frequent
crossings.
2.2.3 Railway bridges in Belgium

Figure 2-8, Figure 2-9 and Figure 2-10 are the summary of the investigation realized in the 6th
Framework project: "Sustainable Bridges – Assessment for Future Traffic Demands and Longer Lives".
The Belgian railway administration NMBS-SNCB manages a total of 5206 bridges.
6
Bridge types
Other
0.9% Reinforced
concrete
21.0%
Brick-vault
40.3%
Prestressed
concrete
Steel 13.4%
Stone-vault
10.8%
0.4% Composite
13.1%
Figure 2-8: Distribution of existing bridge types of railway bridges in Belgium.
50%
Concrete
Steel
33.3%
Composite
25.8%
Vault
25%
13.1%
8.6% 9.2% 8.3%
1.6%
0%
< 10 10-40 > 40
Span length [m]
Figure 2-9: Span profiles of existing railway bridges in Belgium.
50%
Concrete
39.5%
Steel
Composite
24.1% Vault
25%
10.4%
6.9% 8.7% 1.1%
3.5%
1.1% 2.0% 0.7%
2.1%
0%
< 20 20-50 50-100 > 100
Bridge age in [years]
Figure 2-10: Bridge age profiles of railway bridges in Belgium.
At the sight of Figure 2-8, Figure 2-9 and Figure 2-10, it can be observed that:
• 41 % of railway bridges in Belgium are arch bridges and are almost all masonry bridges. All
these arch bridges have a span less than 40 m and have been constructed more than 50 years
ago.
• 21 % of railway bridges in Belgium are reinforced concrete bridges, 13 % are prestressed
concrete and 13 % are composite or encased bridges.
7
• Only 11 % of railway bridges are in metal (steel). All railway bridges constructed in Belgium
and having a span higher than 40 m are in steel.
• All railway bridges having a span less than 10 m are mainly masonry arch bridges or concrete
bridges.
• All steel-concrete composite railway bridges have a span between 10 m and 40 m.
2.2.4 Bridge examples
2.2.4.1 Composite girder bridge
The "Eau Rouge" viaduct

The "Eau Rouge" viaduct is a viaduct of the Verviers-Prüm E42 (A26) highway. It has been constructed
during the years 1989-1993. The viaduct of Eau rouge has a total length of 652.5 m with a principal
span of 270 m constituted by two independent metal arches in shape of rectangular box. The composite
slab was partly constructed by launching and partly with a crane.
© photo-daylight.com
Figure 2-11: The "Eau Rouge" viaduct near Malmédy, Belgium, 1993.
8
Bridge types
The "Gueule" viaduct

The "Gueule" viaduct in Moresnet is a viaduct of the line 24 Glons-Aachen railway. The "Gueule"
viaduct has a total length of 1,108 m and is composed of 22 isostatic spans of 48 m each. The bridge is
constituted by two main steel truss girders with a height of 6.495 m linked by a 30 cm thick reinforced
concrete slab. The bridge was constructed in 1914-1917. In 2002-2005, the viaduct has been
strengthened by additional concrete shell on every pier and by the replacement of all truss elements of
the superstructure.
Figure 2-12: The "Gueule" viaduct near Moresnet, Belgium, 1917/2005.
9
The "Secheval" viaduct

The "Secheval" viaduct is a viaduct of the Maastricht-Liège-Arlon E25 (A26) highway. It was
constructed during the years 1975-1979. The "Secheval" viaduct has a total length of 294.7 m with 5
spans respectively of 39.35 m, 72 m, 72 m, 72 m and 39.35 m. It is constituted by two independent
parallel decks. The metal parts have been constructed on the bank then launched. Each bridge is
constituted by two main steel girders of a height of 3.6 m linked by a 24 cm thick reinforced concrete
slab.
Figure 2-13: The “Secheval” viaduct, Belgium, 1979.
10
Bridge types
The viaduct of Remouchamps

The viaduct of Remouchamps is a viaduct of the Maastricht-Liège-Arlon E25 (A26) highway just near
the viaduct of Secheval. It was constructed during the years 1975-1980. The viaduct of Remouchamps
has a total length of 939.10 m with spans respectively of 39.35 m, 94.50 m, 108.00 m, 2 x 117.00 m,
76.50 m, 67.50 m, 72.00 m, 63.00 m, 54.00 m and 44.75 m. The bridge is constituted by two main steel
girders with a height of 5.1 m linked with a reinforced concrete slab and truss cross-bracings every 9 m.
One peculiarity of this bridge is that it is curved, with a maximal transversal slope of 3 %.
Figure 2-14: The viaduct of Remouchamps, Belgium, 1980.
11
The “Croupets du Moulin” viaduct

The viaduct of "Croupets du Moulin" at Sart is a viaduct of the Verviers-Prüm E42 (A26) highway. It
was constructed approximately during the years 1975-1979. Its design is similar to the "Secheval"
viaduct described previously.
Figure 2-15: The “Croupets du Moulin” viaduct near Sart, Belgium, 1979.
2.2.4.2 Composite box-girder bridges
The viaduct of Polleur

The viaduct of Polleur is a viaduct of the Verviers-Prüm E42 (A27) highway. It was constructed
approximately during the years 1975-1979. The viaduct of Polleur has a total length of 463.9 m with
spans respectively of 93.35 m, 118.8 m, 2 x 93.6 m and 64.55 m. The bridge is a composite steel/light-
concrete box-girder bridge with inclined webs (5.2 m high) linked by a light-concrete reinforced slab
with a thickness that varies from 30 cm to 60 cm. The box comprises truss diaphragms every 3.6 m. Its
maximal longitudinal slope is equal to 6 %.
Figure 2-16: The viaduct of Polleur, Belgium.
2.3 France
2.3.1 Introduction
The success of composite bridges in France begins during the early 1980s and coincides with the
publication of new French design rules in 1981. This success has not been denied since then.
Considering all the new bridges built in France in 2004, the distributions presented in Figure 2-17 can
be drawn. The percentages concern 52 new railway bridges and 282 new road bridges (for main roads,
except highways). It is shown that the solution of a steel or composite bridge (including filler beam
12
Bridge types
deck) is quite often used for railway bridges, but represents only 16% of the new road bridges. In fact
these distributions include the very small bridges (for instance, frames) which are very numerous,
mainly in reinforced concrete, and not so important from an economical point of view. That is why it
makes more sense to study the distribution of bridges according to the main span length. Figure 2-18
shows that for spans between 40 m and 80 m, the twin-girder composite bridge is a very competitive
solution. For greater spans the new bridges are not enough numerous to justify statistical analysis, but it
can be remarked that the composite solution remains even competitive up to 130-m-long span.
Reinforced Filler beam

concrete Composite
9.9% 3.7%
23.1%
Steel
2.1%
Filler beam Reinforced
Prestressed concrete
44.2%
concrete Prestressed 55.0%
5.8% concrete
29.3%
Steel
Composite 11.5%
15.4%
a) Railway bridges (52 new bridges in 2004) b) Road bridges (282 new bridges in 2004)
Figure 2-17: Distribution of the French new bridges in 2004 according to the type of structure.
100%
Reinforced
concrete
75% Prestressed
concrete
Steel
50%
Composite
25% Filler beam
0%
0 10 20 30 40 50 60 70 80
Main span length [m]
Figure 2-18: Distribution of the French new (railway and road) bridges in 2004 according to the
main span length.
On the 10 000 km of roads which are managed by the French National Road Authority, the bridges are
classified between small and large bridges (i.e. main span greater than 40 m, or total deck surface
greater than 1200 m²). These large bridges represent 10 to 15% of all the new bridges built every year,
and 50% of the total new deck surface, with an average price of 1600 €/m² (in 2005). Nearly all the
corresponding 30 new large bridges per year have a steel or composite structure (whereas 30 years ago,
the situation was opposite with nearly all new large bridges in concrete).
This trend in materials is also observed for railway bridges. For instance, have a look at the history of
the high speed railway lines since 25 years:
south TGV line (Paris to Lyon, built in 1983) and west TGV line (Paris to Le Mans, 1990): no steel
bridge
• north TGV line (Paris to Lille, 1993): 13 000 tons of steel used for bridges (for 3600 m of the
line)
• south TGV line (Lyon to Marseille, 2001): 42 500 tons (for 9500 m)
13
• east TGV line (Paris to Nancy, 2007): all the large bridges are composite, 26 300 tons (for 5790
m)
2.3.2 Types of composite bridges in France
2.3.2.1 General
The majority of the large composite new bridges are twin-girder bridges with an upper reinforced
concrete slab. For instance, for the medium and large bridges, the east TGV line (300 km) has 20
composite bridges among which 13 twin-girder bridges, 4 trough composite bridges, two multi-girders
bridges, and one double box-girder bridge. The main reasons for the success of this two-girder design
are the minimising of the fabrication costs and of the construction time.
2.3.2.2 Twin-girder bridges

The most common type of structure consists of a reinforced concrete slab poured in-situ and connected
by studs (or angles) to the two main I-girders. These I-girders are cross-braced by transverse beams
every 6 to 10 m, see Figure 2-19.
© Sétra
Figure 2-19: Usual transverse cross-sections of a composite twin-girder bridge.
There are a large number of variations from Figure 2-19, the main ones are:
• Slab of constant depth
• Slab with transverse prestressing which can be used for wide bridges (example: "Charles de
Gaulle" bridge in Paris, in 1996).
• Slab connected to the cross bracing. In this case, the composite cross-girders are welded to the
upper surface of the main girders. The cross-girders are also more closely spaced as their
centre-to-centre distance is reduced to about 4 m in order to support a thinner slab (25 cm
deep). The upper flange of the cross-girder is butt-welded to the main girder flanges.
• Flat steel vertical stiffener used in low height short span bridges
• Cross-girders extending under the cantilever parts of the deck. See Figure 2-1 and Figure 2-20.
Figure 2-20: Transverse cross girders supporting the slab.
14
Bridge types
When concreting the slab in-situ, a sequence of construction of the different slab segments is usually
defined. In order to minimise the tension in the concrete around the internal supports, the corresponding
slab segments are concreted at the end. See Figure 2-21. If the closest concrete plant is too far from the
site or if the needed concrete volume is too small for justifying a concrete plant directly on site, the
design is performed with prefabricated slab segments. They are connected to the structural part in a
second phase by concreting rectangular openings and transverse joints. The studs can be welded in the
workshop or directly on site using a specific gun. See Figure 2-22.
© Sétra
Figure 2-21: Concreting on internal supports at the end.
© Sétra © Sétra
Figure 2-22: Use of prefabricated slab segments.
A halfway solution (rather used for bridges with multiple girders) consists in using very thin concrete
pre-slabs laid between the steel girders and used as a formwork for concreting the rest of the slab in-
situ. See Figure 2-23.
© Sétra
Figure 2-23: Use of pre-slab as formwork.
For railway bridges, a specific transverse diaphragm is usually used to have a better behaviour for
torsion. See also Section 6 for further details and Figure 6-4.
15
Discussions of the design from the economical point of view

The fabrication of the steel structure of a twin-girder bridge can be broken up into 4 steps : the
supplying of the workshop with the steel plates coming from the steel mill, the partial assembling made
in the workshop, the transport from the workshop to the site and the final assembling on site. Each step
imposes conditions which influence the composite bridge design and the competitiveness of the chosen
solution.
For being economical the steel plates should be ordered to the ironworks with respect to :
• the ironworks capacity : maximum thickness, length width or weight per plate.
• For instance, if the plate weight exceeds around 23 tons, the production process changes (from
continuous slab casting to top casting) and the price increases (+ 80 €/ton).
• the capacity of the transport by rail : maximum length and width
• the optimisation for cutting the maximum different plate elements (in the workshop before
welding) out of the same rectangular mother plate and with the minimum off cuts.
For a twin-girder bridge with the usual cross-bracing from Figure 2-19, the construction time in the
workshop can be broken up into the percentages indicated in the table below. It results in highlighting
the high time costs induced by the stiffening in comparison to the implemented steel weight. It should
be mentioned that the built-up of the I-girders can be automated (up to a depth equal to 5750 mm)
which is not the case of the stiffening, see Figure 2-24. The girder handling (for instance turning down
before welding on the other side) still remains one of the most time consuming part of the work.
Table 2-1: Indication for construction times of a twin-girder bridge.
Steps Time consumption Weight
Fabrication of the main girders

y Cutting the plate elements 10 %
y Butt welding 11 %
y Built-up of the I-girder 14 % 85 %
y Stiffening 48 % 06 %
y Studs 04 %
y Dispatch 05 %
Fabrication of the cross-bracing 5% 6%
Miscellaneous 3% 3%
The transport of the I-girder segments from the workshop to the bridge site can be performed by barges,
by trains or by trucks, according to the workshop facilities. The kind of transport influences the size of
the segments, the splices that have to be welded on site, and eventually the retained solution (I- or box-
girders). The table below gives some indications about the transport limitations in France. For instance
it results that a box-girder bridge wider than 6 m should be split into two pieces which have to be
longitudinally welded on site.
The final assembling on site depends on the clearance of the bridge and the availability of the areas
surrounding the abutments. If the clearance is not too high, a construction by crane (or an heavy lift
derrick on a barge) can be performed. If not, the bridge should be launched for one (or both) side. A
launching platform should be available in the bridge continuation. See Chapter 7 for further details
about launching devices and process.
16
Bridge types
© Sétra © Sétra
Figure 2-24: Built-up of a bridge I-girder and welding of transverse stiffeners.
Table 2-2: Transport limitations in France.
By trucks By rail
Exceptional convoy (with

specific authorisations
Normal convoy and escort police)
Weight 20 tons < p < 30 tons up to 100 tons 50 tons (up to 70 tons with
specific wagons)
Length 22 m 50 m 32 m
Width 4.50 m 6.00 m 3.00 m (and a concomitant

depth up to 1.5 m)
Depth 3.20 m 6m 3.05 m (and a concomitant

(only on specific itineraries) width up to 0.8 m)
17
Examples
La Risle viaduct is a typical twin-girder road bridge built for the highway A28 in Normandy in 2004.
General data of La Risle viaduct is as follows:
Bridge data:
Total length = 1320 m
Span distributions = 65 – 4 x 80 – 7 x 90 – 3 x 80 – 65 m
Girder height = 3.5 m with cross-bracing every 8 m
Reinforced concrete slab, 15-m-wide and around 30-cm-deep
5295 tons of steel, built by launching from both abutments
Clearance = 70 m
One lane in each direction with a central separator
© Sétra
Figure 2-25: La Risle viaduct near Brionne, France, 2004.
A new road composite viaduct is under construction in Avignon (south of France). The structure is a
twin-girder bridge with a wide concrete slab supporting 2 x 2 traffic lanes. The two girders are cross-
braced with I-girders connected to the slab and extending over the cantilever parts. General data is as
follows:
Bridge data:
Span distributions = 36 – 60 – 64 – 80 – 84 - 4 x 88 –
64 m
Girder height = 3.5 m with a clear transverse
distance of 12 m between the main girders
Cross-bracing every 4 m
Reinforced concrete slab, 21.5-m-wide and 24-cm-
deep
4500 tons of steel (S355 + S460), built by launching
© Sétra
Figure 2-26: LEO viaduct over Durance near Avignon, France, 2008.
This bridge has been designed using Eurocodes (EN final versions). For a 33-m-long zone of the main
steel girders surrounding each internal support, the steel grade S460 has been used for the flanges.
Compared to a design made of S355 only, the used steel weight is reduced by 8%.
The last example mentioned here is one of the composite railway bridges built for the new East TGV
line in 2007: the viaduct over the Ourcq valley. This example is representative for the specific cross-
bracing design generally used for railway bridge (steel diaphragm). The lower steel flanges are also
braced with prefabricated slab segments for reducing the noise induced by the TGVs, for a better
dynamic behaviour of the deck and for a better torsion behaviour.
18
Bridge types
Bridge data:
Span distributions = 45 – 6 x 60 – 45 m
Girder height = 3.9 m
Cross-bracing made of diaphragms every 4 m
Upper reinforced concrete slab, 12.6-m-wide and from 25-cm to
40-cm-deep, poured in-situ
1963 tons of S355, built by launching (with the lower concrete
slab)
See also Figure 6-4 for further details.
© Sétra
Figure 2-27: Viaduct over Ourcq valley, France, 2006.
2.3.2.3 Examples of composite box-girder bridges

The Verrières viaduct is a composite box-girder bridge for the highway A75 near Millau. The central
steel box is closed and its upper flange is connected to the concrete slab. The wide slab supports 2 x 2
traffic lanes. Cross-bracing connected to the slab, steel diagonals and a longitudinal beam help to
support the cantilever part of the slab, see Figure 2-2.
Bridge data:
Total length = 719.5 m
Span distributions: 96 – 136 – 144 – 136 – 128 - 80 m
Girder height = 4.5 m
Box-girder with vertical webs (spacing = 7 m)
Cross-bracing every 4 m
Reinforced concrete slab, 23.5-m-wide
6226 tons of steel
Longitudinal splice for assembling two half box-girders
Built by launching
2261 €/m² (2002)
© Sétra
© Sétra
© Sétra
Figure 2-28: Verrières viaduct near Millau, France, 2002.
19
Of course, for the same highway, the steel box-girder of the Millau viaduct (built in 2004) can also be
mentioned, see Figure 4-3. Another remarkable recent example is the Jaulny viaduct for the East TGV
line (2005). The structure is a double closed box-girder bridge with a main span of 73.80 m for a total
length of 478.70 m. The average depth of the steel box sections is 3.60 m, the flange stiffeners are
trapezoidal closed profile whereas the web stiffeners are flat strips.
Figure 2-29: Jaulny railway viaduct, France, 2005.
2.4 Germany
2.4.1 Introduction
In Germany a successful rise of the construction of steel-concrete composite bridge structures started in
the 1990s which is linked closely to a new political situation: the German reunification. At that time, the
infrastructure in Eastern Germany was in a very bad condition and the objective was to rebuild it as fast
as possible. During the conceptual design phase, innovative designs were developed e.g. in order to
reuse the existing foundations and substructures which were - in contrast to the superstructures - in a
rather good condition. As a result, the use of lightweight superstructures was often required, which is a
domain of the composite structures. The traditional design approach for long-span (motorway) bridges
is to have separate superstructures for each carriageway which are often built up from I- or box-girders.
As example, bridges near Cottbus and Schrotetal are shown in Figure 2-30 and Figure 2-31 which used
to be the main bridge types in Germany until the end of the 1990s.
Figure 2-30: River bridge Spree near Cottbus, Germany, 1994 [10].
20
Bridge types
Figure 2-31: Bridge Schrotetal near Magdeburg, Germany, 1997 [10].
However, the fast reconstruction of the infrastructure in Eastern Germany led to the development of
advanced design approaches on which in the following sections the focus is set. The bridge types which
were able to establish themselves since then are:
• Bridges with a one-piece composite superstructure, cf. Section 2.4.2
• Bridges with airtight small-sized box-girders, cf. Section 2.4.3
• Bridges with prefabricated components and in-situ concrete, cf. Section 2.4.4
In Germany a dense infrastructure network exists with over 231,000 km of roads and approximately
34,000 km of railway lines. The length of federal highways is 53,346 km (motorways: 12,363 km,
national roads: 40,983 km) [4] and the operated railway line length is 33,897 km [20]. The number of
bridges adds to 35,675 for federal highways [84] and 27,165 for railways [20]. Data on state and district
roads is not centrally recorded so that in the following only information on federal highways can be
given.
Figure 2-32 shows the number of bridges with regard to bridge length for road bridges and span length
for railway bridges. For road bridges, an evaluation of the database based on the span length is not
possible so that only the bridge length is referred here. However, in both cases it can be seen that
bridges in the small and medium span range dominate. Usually, the roads consist of two lanes in each
direction leading to bridges with total widths between 15.5 and 29.5 m depending on the type of
superstructure, two-piece or one piece.
48.9%
50%
25%
16.9% 15.4%
12.0%
6.7%
0%
2-5 5-30 30-50 50-100 > 100
Bridge length [m]
a) Road bridges (federal highways) [84].

Figure 2-32: Distribution of span and bridge lengths in Germany [84], [87].
21
75%
62.2%
50%
25.9%
25%
6.9%
2.1% 2.9%
0%
0-10 10-20 20-30 30-40 > 40
Span length [m]
b) Railway bridges [87].

Figure 2-32 (continued): Distribution of span and bridge lengths in Germany [84], [87].
An analysis of the existing bridge types in Germany shows that especially for road bridges the market
has been dominated by concrete bridges, cf. Figure 2-33. In contrast to this, the bridge types for railway
bridges are more balanced but in both cases it is obvious that steel-concrete composite bridges played
only a minor role so far.
Steel Composite
5.1% 2.1%
Filler beam Stone-vault
25.2% 28.1%
Reinforced
Prestressed
concrete Composite
concrete
54.4% 1.2%
38.4%
Steel Reinforced
24.8% concrete
Prestressed 16.9%
concrete
3.9%
a) Road bridges (federal highways) [84]. b) Railway bridges [87].

Figure 2-33: Distribution of existing bridge types in Germany.
Steel Filler beam

Reinforced Prestressed 1.4%
concrete concrete 3.4%
4.5% 2.8%
Prestressed
concrete Reinforced
Composite
36.4% concrete
54.6%
92.4%
Steel
4.5%
a) Road bridges (federal highways), b) Railway bridges.

Bridge lengths referenced between 30 and 60 m.
Figure 2-34: Distribution of current construction types in Germany [84].
22
Bridge types
Current construction types, see Figure 2-34, show that the composite bridges have gained a good market
share at least for road bridges. Reasons for this are given in the following sections.
2.4.2 Bridges with a one-piece composite superstructure
2.4.2.1 General
Until 1997 the former German Federal Ministry of Transport, Building and Housings (BMVBW)
required to design separate superstructures for each carriageway for motorway bridges made of concrete
and steel-concrete in order to be able to sustain the traffic during maintenance operations. This was
mainly related to necessary repairs of the concrete slab so that steel bridges with orthotropic decks were
excluded from this requirement from the beginning. However, in 1997 the BMVBW decided to allow
one-piece superstructures for steel-concrete composite bridges for several reasons. In the following
general part, reasons for and characteristics of one-piece composite superstructures are given. In Section
2.4.2.2 the implications of different cross-sectional layouts are discussed in detail with regard to
structural and fabrication-related aspects (see also e.g. [19]).
Starting point to mitigate the requirement of building two separate superstructures was the development
of motorways in the newly-formed German states after the reunification. Especially in Thuringia a large
number of deep valleys needed to be crossed and the objective was to encourage economic bridge
designs with a high aesthetical quality. Depending on the boundary conditions these criteria could be
fulfilled very well by one-piece composite superstructures due to the following reasons:
• Number of piers. With a one-piece superstructure the number of piers can be reduced by half
in comparison with separated superstructures. As a result, the perspective of the valley is not as
severely obstructed as if two parallel rows of piers are required and also the costs for piers and
foundations are reduced. In case of one-piece superstructures, a monetary breakeven is reached
for pier heights starting at 40 to 50 m and higher, cf. [41].
• Durability. In comparison with a pure concrete bridge the durability of the composite bridges
is higher because a renewal of the corrosion protection as well as replacement of the wearing
part “concrete slab” is possible. A prerequisite for the latter is that the replacement can take
place under traffic and that the resulting load effects are already considered at the design stage,
which has also influence on the characteristics of the bridge as described later on. The design
for a replacement of the concrete slab is decisive for the design of the superstructure so that it is
over-designed for the serviceability limit state which in turn, however, leads to an expected
robust and durable structure without endangering its competitiveness. The general design
considerations lead to certain characteristics of one-piece composite superstructures which are
introduced below. Figure 2-35 shows a typical layout of a one-piece composite superstructure.
• Cantilevering concrete slab. The wide concrete slab usually without pretensioning carries
both carriageways. Due to its width the slab is usually supported longitudinally at five locations
in the transverse direction. Two of them are the upper flanges of the box-girder which are
complemented by longitudinal beams in the middle and on the outside. The outer beams are
supported by lateral struts.
• Single steel box-girder with inclined struts and tension band. The steel box usually has a
wide bottom plate with inclined webs. The cantilevering concrete slab is supported by lateral
struts which often consist of hollow cross sections. In general, outer edge beams and a middle
beam are foreseen in the longitudinal direction. The deviating forces of the lateral struts are
counteracted by tension bands at the level of the upper flanges. These tension bands are usually
connected to the concrete slab by studs. The cross-bracings and inclined struts are arranged
with a distance between 4 to 5 m in a regular pattern.
• Replaceability of the concrete slab. It is a requirement that the concrete slab can be replaced
under traffic on the bridge. This means that the traffic is running on one side of the bridge only,
whereas on the other side the slab is removed in section of 10 to 15 m in the longitudinal
direction. In the transverse direction it is also possible to remove the slab not as one piece per
each side but to subdivide it here as well. As the traffic load itself gives a high torsional
23
moment to the bridge girder, counter weights should be used to reduce this effect. The
replacement of the concrete slab has to be already considered as a load case during the design
of the bridge. The proposed procedure of replacement has to be sufficiently well documented.
longitudinal
tension band middle beam concrete slab
steel
box
longitudinal
outer beam lateral strut
Figure 2-35: Layout of a one-piece composite superstructure.
Table 2-3 summarises the data of bridges with a one-piece composite superstructure, which have been
completed in Germany up to now. The costs indicated refer to the time of construction.
2.4.2.2 Influence of the cross-sectional layout

In the following implications of the choice of different cross-sectional layouts are given, which concern
mainly the steel structure. The following topics are discussed:
• Longitudinal middle and outer beams
• Tension band
• Concrete slab
• Influence of the cross-sectional layout on the fabrication of the concrete slab.
Longitudinal middle and outer beams

Longitudinal middle and outer beams are typically chosen because the construction phase is simplified.
The longitudinal beams e.g. often serve as runways for the formwork carriages. Besides that, in the final
stage, longitudinal beams avoid a point-support of the slab because a biaxial tensile state of stress is not
covered by the codes so that no experiences with regard to robustness and fatigue exist. If a middle
beam is chosen, it is often a closed cross-section in order to cope with the torsional stresses which are
induced by a possible replacement of the concrete slab. In case of a closed cross-section it is a airtight
section. However, longitudinal outer beams are not a necessary requirement as the bridge examples
Reichenbach, Steinbachtal and Nesenbachtal show.
Tension band
There are different possibilities where to place the tension band.
For a tension band in the concrete slab, the surrounding concrete offers some protection, however,
special care has to be taken in case of a decomposition of the concrete slab. Moreover, local load effects
of the slab have to be considered in their design. The connection to introduce the tensile forces into the
cross-bracings is usually rather complicated.
For a tension band underneath and connected to the concrete slab, a grid-like action can be accounted
for the support of the concrete slab which may result in a smaller slab thickness. The tension band acts
like a concrete beam with outside reinforcement. Due to this, it has to be taken into account that the
studs will also get some tension because they act like a stirrup.
24
Bridge Year of Total Spans Max. Deck Superstructure Costs
com- length Height area
pletion
Concrete Reinforcing / Constructional
prestressing steel steel
[m] [m] [m] [m2] [m3] [m3/m2] [t] [kg/m2] [t] [kg/m2] [€/m2]
Wilde Gera 2001 552 30-42 110 14,628 4,630 0.317 895 61 2,508 172 1,425
Albrechtsgraben 2002 770 45-70 80 21,945 8,500 0.387 1,500 68 5,100 232 1,267
Reichenbach 2002 1,000 40-105 60 28,500 9,700 0.340 2,144 75 6,323 222 1,140
Schwarza 2002 675 55-85 68 19,238 6,450 0.335 1,700 88 4,428 230 1,138
Seßlestal 2002 320 73-88 53 9,120 3,300 0.362 800 88 2,521 276 1,349
Steinbachtal 2002 372 48-78 30 10,974 n.a. n.a. n.a. n.a. n.a. n.a. n.a.
Dambachtal 2005 370 45-85 65 10,545 4,400 0.417 1,000 95 2,800 266 1,166
Elben 2005 432 40-80 52 12,312 4,545 0.369 870 71 2,950 240 1,218
Thyratal 2005 1,115 70-90 40 32,893 12,300 0.374 2,550 78 8,300 252 1,216
Table 2-3: Data of bridges with a one-piece composite superstructure in Germany.
Haseltal 2006 845 70-175 82 24,083 10,200 0.423 2,400 100 9,400 390 1,428
25
Bridge types
For a tension band underneath the concrete slab with a certain distance, the structural system is clearly
divided in the longitudinal and in the transverse direction. A disadvantage is the offset at the joint to the
slab which causes local bending. Due to different thermal expansion of the slab and the tension band
tensile stresses may occur in the slab, which has to be considered when reinforcing the slab.
Concrete slab
The concrete slab usually consists of reinforced concrete without pretensioning, which is erected with
the back-step method. The layout of the concrete slab can be as follows:
• Slab with a constant thickness, e.g. valley bridge Albrechtsgraben
• Slab with haunches at the outer longitudinal beams and the upper flanges of the box-girder
• Coffered slab which reduces the amount of concrete and thus the selfweight of the bridge is
also reduced. However, the layout of the reinforcement becomes more complex
As said in the introduction, it is required that the concrete slab can be replaced under traffic on the
bridge. As the traffic load itself gives a high torsional moment to the bridge girder it is recommended to
use ballast weights to reduce this effect. In addition to that, joints at the gusset plates for the horizontal
bracings in case of a slab exchange have to be foreseen as shown in Figure 2-36. Other design aspects
may include e.g. that the end cross girder is made of concrete in order to avoid lifting forces when
removing the concrete slab in the end span.
© KE
Figure 2-36: Joints foreseen to take the horizontal bracings in case of slab exchange.
Influence of the cross-sectional layout on the fabrication of the concrete slab

When several concreting steps are foreseen for the casting of the concrete slab a formwork carriage is
usually used, which has the task to resist all loads of the formwork, the fresh concrete and other
working loads. For the formwork and its carriage the following general characteristics apply:
• No ground-based falsework is possible so that the loads have to be introduced adequately into
the steel superstructure
• Easy to move taking into account the pass-by at the lateral struts
• Only few available support locations
• Compatibility of superstructure stiffness and falsework stiffness
• Lightweight
• Cost-efficient
Based on these requirements different layouts of the formwork carriage exist, which are dependent on
the cross-sectional layout. In the following formwork carriages “running on top” or “running
underneath” are presented.
26
Bridge types
Formwork carriages “running on top“

For formwork carriages running on top, stands are necessary to introduce the loads into the steel
superstructure. These stands are possible at a maximum of five locations in the axis of the cross-
bracings and struts in the transverse direction to be able to carry also horizontal forces into the
superstructure. Moreover, the stands need recesses in the concrete slab. After the casting procedure the
stands are cut below the slab’s surface and the recess is closed afterwards. These recesses are rather
undesirable because the quality of the slab suffers and also the layout of the reinforcement is
complicated in this area. It has to be considered that the steel superstructure has to be designed for these
concentrated high forces especially if the number of stands is reduced as far as possible. It has also to be
taken into account that the formwork has to pass by the outer inclined struts. A separate formwork
inside the box-girder is commonly provided which can be lifted or lowered independently from the
outer formwork.
For the valley bridge Reichenbach all five possible locations for stands in the transverse direction have
been used. In Figure 2-37 these stands are shown. It can be seen that outer longitudinal beams do not
exist. As a result the formwork of the slab area above the outer inclined struts and the cantilevering part
was provided by foldable formwork panels which were hooked on the outer edge of the formwork
carriage. The formwork of the slab area inside the box-girder was provided by formwork panels
travelling inside the box [70].
© KE
Figure 2-37: Small stands as support for formwork carriages “running on top“, valley bridge
Reichenbach near Ilmenau, Germany, 2002.
For the valley bridge Wilde Gera the number of stands in the transverse direction was reduced to two,
which are located at the joint of outer inclined struts, tension band and outer longitudinal beam. As a
result the recesses in the concrete slab could be reduced and thus the quality of the slab was improved.
For the cantilevering slab area foldable formwork panels were used, which were hung from the
formwork carriage. The formwork of the areas between the inclined struts were provided by formwork
panels, which were supported on temporary brackets located at the struts and the cross-bracings of the
steel box as shown in Figure 2-38 and [18].
Figure 2-38: Example of a formwork carriage “running on top“.
27
Formwork carriages running underneath

If an outer continuous longitudinal beam exists it is advantageous to have a formwork carriage attached
to this beam, which is running underneath. Besides the longitudinal beams it is laterally supported by
the lower flanges of the steel box. As a result no recesses in the concrete slab are required for the stands
of the formwork carriage. Although this leads to higher quality of the concrete slab, maybe additional
mounting parts and strengthening methods are necessary for the steel parts. Special care should be taken
of the corrosion protection.
For a number of valley bridges a formwork carriage running underneath was used, which was attached
to the outer longitudinal beams. Usually foldable formwork panels are used, which are connected to
carriages, hanging and travelling only on the lower flanges of the outer longitudinal beam. During
casting of the concrete, the formwork carriages are supported by the outstand of the bottom plate as
well. Inside the steel box formwork panels are used similar to formwork systems used with formwork
carriages running on top. Figure 2-39 shows the formwork carriage of the valley bridge Schwarza. This
method has also been used e.g. for the valley bridges Albrechtsgraben and Elben.
© KE
© KE
Figure 2-39: Foldable formwork panels, valley bridge Schwarza, Germany, 2002.
© KE
© KE
Figure 2-40: Valley bridge Nesenbachtal,

Stuttgart, Germany, 2000.
28
Bridge types
Use of partial depth precast composite slabs

Instead of in-situ cast concrete for the slab, an alternative solution is to use precast composite slabs,
which are placed on the steel structure and covered by a layer of in-situ cast concrete. An example for
such a solution is the bridge Nesenbachtal [6] as shown in Figure 2-40. Although it carries only one
carriageway comprising of two lanes it has the same characteristics as the one-piece composite
superstructures: it is a single steel box-girder with diagonally inclined struts and a tension band. The
concrete slab is built up with precast concrete components, which are placed on the transverse girder at
a distance of 4 m. Measurements showed that the precast concrete components are able to form a stable
unit with the in-situ concrete parts [48].
2.4.3 Bridges with airtight small-size box-girders

Since the mid-1990s bridges with small-size box-girders, which were airtight without corrosion-
protection in the inside started to spread all over Germany. This design was able to compete against
other types of composite and concrete structures in the span range between 30 to 80 m. A typical cross-
section consists of two main girders running in the longitudinal direction as shown in Figure 2-41.
© KE
a) Bridge Werratal near Einhausen, Germany, b) Typical cross-section with a cross girder.
2003.
Figure 2-41: Bridges with airtight small-size box-girders.
Originally strong doubts existed against airtight box-girders and the effectiveness of the airtightness to
function as corrosion protection. However, it was already widely used as inside protection e.g. for
inaccessible members of truss and arch systems with box sections as well as stiffeners with a closed
section shape. A change in mind came in the mid-1990s when the continuous steel box-girders of the
guideway at the Transrapid test facility were made accessible, cf. [83]. These girders were built in 1982
as airtight box-girders with a projected life-time of 12 years only. Their inspection showed that no
corrosion had taken place inside the airtight box. Finally, it was given way for a new bridge type, which
was further supported by improved manufacturing techniques for the steel parts and advancements in
computer technology, which facilitated the calculation of composite structures. A summary of the most
important bridges which were published in journals, books, etc. [11], [13], [14], [21], [22] is given in
Table 2-4 but many more bridges with airtight small-sized box-girders have been built so far, which
were not made known to the public very well.
The characteristics of the airtight box-girders can be summarised as follows, cf. [74]:
• Support of the concrete slab. The two webs of each main girder provide a favourable support
condition for the concrete slab.
• Mass. The small width of the steel bottom plate leads to a larger effective width in the hogging
bending moment areas. The close supports of the concrete slabs lead to slabs with a maximum
thickness of 35 cm, which is favourable for the overall self-weight of the bridge. A transverse
pre-stressing is usually not used.
29
Table 2-4: Data of bridges with airtight small-sized box-girders in Germany.
Bridge Year of Total Spans Slab No. of Deck Constr Costs

comple- length width girders area . Steel
tion
[m] [m] [m] [m2] [kg/m2] [€/m2]
Union 2000 253.7 25-35 19.1 2x2 9,704 n.a. 1,106
Werratal 2003 1194.0 37-85 14.25 2x2 34,029 n.a. 837
Müglitz 2003 310.0 45-55 14.25 2x2 8,835 n.a. n.a.
Schwentine 2003 296.0 20-52 11.5 2 3,404 241 1,704
Wakenitz 2004 294.5 32-55 14.9 2x2 8,781 247 1,651
Seidewitztal 2006 605.0 31-55 12.9 2x2 15,609 n.a. n.a.
Wiedersbach 2006 177.5 30-60 14.25 2x2 5,059 217 1,192
Schleuse 2007 680.0 40-80 14.25 2x2 19,380 260 1,028
• Fabrication. A prefabrication of the box-girders up to maximum transportation lengths is

possible so that only splices in the longitudinal direction have to be welded on site. The small
dimensions of the box-girders are easy to handle at the workshop but they also lead to difficult
welding positions of welds inside the box.
• Corrosion protection. In comparison with an I-girder equipped with stiffeners, there are only
few and plane outside surfaces in case of the box-girders. Moreover, inside there is no corrosion
protection needed at all.
• Transportation. The box-girders can be transported whole without the need to be assembled
on site in the transverse direction. The high torsional stiffness facilitates the transport.
• Erection. The box-girder is stable during erection and final stages due to its high torsional
stiffness. Usually the elements are lifted into their final position but also the launching
technique can be used as done for the Bridge Seidewitztal [27]. There is not much space to do
the welding on site and also the manholes have to be closed in the ends. There are only few
areas that need a final painting on the erection site.
• Erection time. The crane mounting and easy erection of large prefabricated sections leads to a
very short erection time without much influence on the traffic.
• Equipment. The equipment of walkable box-girders such as catwalk, lighting, etc. is not
required.
• Maintenance. Time and effort for maintenance of the steel structure are reduced because only
outer surfaces are observed. However, the maintenance of other equipment such as the drainage
on the outside requires more effort. Structural elements inside the box-girder are not easily
accessible. If an inspection inside the box-girder is necessary rail-bound devices can be used,
which have to be considered already at the design stage.
The design of small-size box-girders depends mainly on the slab width and the main span. For small
spans the main girders are closely spaced with a distance of 2 to 3 m. In the medium span range the
typical construction consists of two main girders, which support slabs with widths between 11 to 16 m.
As can be seen from Table 2-4 that this range of slab width is a very common. For larger widths either
30
Bridge types
more main girders or sometimes struts and cantilever arms are used are used. In this case the use of
prefabricated concrete slabs can be favourable [75]. Another not very common solution is the pre-
stressing of the slab in the transverse direction, which has been used e.g. for the Union Bridge, Dresden
[11]. Besides that the following facts should be kept in mind [72], [73]:
• In Germany the weld specifications of ZTV-ING T4 are sufficient for the manufacturing of
airtight cross sections.
• An extra amount in steel thickness to account for corrosion is not required.
• An examination of the airtightness is not necessary.
• A corrosion protection inside the box-girder is not required. Also parts, which are welded on
site do not need a protection although quite often a base coating is applied to these parts. The
box-girder section does not need to be sealed until erection.
• A load case pressure difference of ±15 K has to be considered. Thus, the closure of the box-
girders on site should be done with respect to the assumed temperatures in the calculations.
• Due to the width of the upper flange, the clamping of the slab has to be taken into account.
• An inspection of the inside of the box-girder is not required if the statically necessary welds can
be checked from the outside.
• To conclude, it can be said that the use of airtight small-size box-girders has become a state-of-
the-art building technique for medium span bridges in Germany.
2.4.4 Bridges with prefabricated components and in-situ concrete
2.4.4.1 General
Due to the dominance of concrete bridges for small and medium spans in Germany, it has been
recognised that the competitiveness of steel-concrete composite bridges can be decisively improved by
a high level of prefabrication. Starting at the beginning of the 1990s, Roik [76] introduced the idea to
use hot-rolled steel beams in composite road bridges for the small and medium span range together with
prefabricated concrete elements. The main advantages of this proposal are still valid today:
• Prefabrication. A high level of quality can be achieved, e.g. welding on-site is not necessary
and the corrosion protection can be fully applied in the workshop.
• Erection time. The high level of prefabrication in general leads to short erection times because
the traffic is blocked only during the installation of the girder. Only short-time blocking of
roads or rail tracks are required.
• Lightweight. If substructures exist and are in a good condition, they can be used further.
Since the end of the 1990s numerous composite bridges have been successfully built thanks to a high
level of prefabrication and they have been able to establish themselves as an economic solution in the
span range between 20 to 80 m. From the beginning, attempts have been undertaken to optimise not
only the steel structure but also the way how to fabricate the concrete slab. For multi-span valley
bridges a formwork carriage is often used, which is usually not an appropriate solution for bridges in
the small and medium span range. Thus, two innovative approaches evolved, which are both combined
with in-situ concrete.
• Steel girders and precast concrete elements, see Section 2.4.4.2
• Prefabricated composite girders, see Section 2.4.4.3
31
2.4.4.2 Steel girders and partial depth precast concrete elements

In 1989, the steel fabricator ARBED initiated a research programme to optimise the design and
construction of composite bridges when using hot-rolled steel beams with advanced characteristics such
as a girder height up to 1100 mm and a yield strength up to 460 MPa [49]. The characteristics of hot-
rolled beams in contrast to welded girders were:
• multi-girder layout with distances of 2.5 m
• main girder made of hot-rolled beams without longitudinal stiffeners or bracings
• limited construction height of 1100 mm
• reduced slab thicknesses of 27 cm in comparison to the usual 35 cm
The main outcomes of the research project were that due to the geometry of the hot-rolled beams the
risk of lateral-torsional buckling of the compression (bottom) flange at the supports is reduced because
the web provides a good elastic restraint. The whole system was already intended to be used with
precast concrete slab elements in order to reduce erection time because no additional formwork is
needed and the use of precast concrete slab elements as permanent formwork facilitates erection.
However, the cantilevering arms were planned to be made with in-situ concrete. Another drawback is
that hot-rolled beams are only economically transportable with lengths up to 30 m.
In 1997, standardised designs for bridges crossing German motorways were worked out because the
system requirements for the superstructures were the same in almost all the cases. The most frequently
road types which need to be bridged are access roads “WW” (farm track), two- or three-lane roads with
a RQ10.5 and RQ15.5.
6.50
WW
0.75 5.00 0.75
2.5%
12.00
RQ 10,5
2.00 0.50 3.50 3.50 0.50 2.00
2.5%
16.75
RQ 15,5
2.00 0.50 3.75 0.50 3.75 3.75 0.50 2.00
2.5%
Figure 2-42: Typical layout of composite bridges with prefabricated slabs. Only slab system
based on [71] is shown. Dimensions in [m].
Thus, the aim was to help the designers for typical cross sections of bridges with a single carriageway.
The different standardised types have been also published in [81] and [82]. In contrast to the precast
concrete elements which are commonly known for building construction, here so called “large area
formwork elements” have been developed. They have the following characteristics [81]:
32
Bridge types
• The shear connection between large area formwork element and in-situ concrete is realised by
garland-type shear reinforcement which have no upper reinforcement bar. As a result the
reinforcement can be easily placed without the need of a complicated thread.
• The large area formwork elements are fully effective in the final state. To realise the shear
connection, recesses are provided in the precast elements to fit the groups of studs, which are
cast later on in a first step.
• Elastomer strips are applied to the steel girders to compensate for tolerances and to seal the
possible gap between steel and concrete.
• The transverse joints are sealed in the workshop by foam rubber. They have a certain layout for
the following reasons: the slab thickness is reduced at the joints to provide a sufficient concrete
cover of 4.5 centimetres. The edges are broken to prevent flakings.
For the different construction types, distribution plans how many and where to place the large
formwork elements most efficiently were already provided.
0.20
WW b = 2.40
5
0.10
1.50 3.00 1.50

6.00
0.20
RQ 10,5 be = 2.40
5
0.10 5
bi = 2.23
4.25 2.80 2.80 1.45
0.20
WW l = 6.34
0.10
2.55 0.90 2.55

5
2.61
0.20
RQ 10,5 le = li = 6.34
0.06
5 0.10
0.90 0.90 0.90

5 5
1.71 2.78 2.78 1.71
2.10 3.55 3.55 2.10
11.30
Figure 2-43: Example cross-sections of large area formwork elements based on [81] for the
layouts of Figure 2-42. Dimensions in [m].
Figure 2-44 shows an example of a bridge with steel girders and precast concrete elements which has
been built near Ravensburg. It can be seen that studs are grouped in order to allow for the cantilevering
of the large area formwork elements.
33
© KE
Figure 2-44: Erection of a bridge with steel girders and precast concrete elements near
Ravensburg, Germany.
2.4.4.3 Prefabricated composite girders

Figure 2-45 shows the layout of a typical prefabricated composite girder. It consists of a rolled or
welded steel girder to which a concrete flange is added. The concrete flange is already prefabricated in
the workshop and the shear connection can be realised either by vertical or horizontal studs, in the latter
case omitting the top flange or by concrete dowels. On the building site this concrete flange serves then
as formwork for the completion of the deck with in-situ concrete.
cast-in-place concrete
prefabricated concrete flange
a) With vertical studs. b) With horizontal studs.

Figure 2-45: Typical cross-sections of prefabricated composite girder with in-situ concrete.
In the following, the characteristics of bridges with prefabricated composite girders and in-situ concrete
are presented:
• Mass. In traditional composite construction only the steel girder is prefabricated and
transported to the building site. Usually the steel girders support the formwork. In case of an
unpropped erection, high internal forces due to the weight of formwork and concrete arise,
which have to be resisted by the steel girder alone because composite action is not yet present.
This increases the amount of structural steel and lowers the competitiveness. When using
prefabricated composite girders instead, it has been shown in [24], [85] that savings in
structural steel can be up to 30% in comparison with a traditional composite construction
because the concrete flange as a structural element adds strength and stability to the steel girder.
• Fabrication. During fabrication of the composite girder no stresses are applied to the steel
section. The steel girder is manufactured according to a stress-free workshop shape and the
corrosion protection is already applied. During casting and curing of the concrete flange
hereafter, the steel girder is supported in such a way that no internal forces arise in the girder
and not intended deformations are avoided.
34
Bridge types
• Transportation. Due to their light weight, prefabricated composite girders can be transported
and dealt with more easily than prestressed concrete girders for which only a maximum length
of about 30 to 40 m due to their self-weight is feasible. In contrast to this, prefabricated
composite girders can be delivered in larger dimensions: transportation issues limit the girder
lengths to 60 m on the road and up to 100 m on the waterway according to [85].
• Erection. In traditional composite construction not only the steel girders had to take the internal
forces induced by the weight of the formwork and the concrete but they also had to be braced to
prevent them from overturning or lateral buckling. However, these bracings are difficult to
install and expensive. The use of prefabricated composite girders simplifies this work step
considerably because the composite girder itself provides strength and stability, which can be
even enhanced by a coupling of the flanges. At best, the prefabricated composite girders and
their arrangement are designed such that no additional formwork is needed for the casting of the
concrete slab. Joints between prefabricated composite girders and adjacent structural
components made of concrete can easily be designed structurally so that they can be used for all
types of structural systems: simple and continuous beams as well as frame systems.
• Erection time. The high level of prefabrication in general leads to short erection times because
the traffic is blocked only during the installation of the girder. No additional blocking of roads
or rail tracks is required if the prefabricated composite girders are arranged in such a way that
they provide both the working level and the formwork for the in-situ cast concrete slab.
© KE © KE
Figure 2-46: Erection of a prefabricated composite girder with horizontal studs near
Münsingen, Germany.
• Maintenance. The fabrication of the girders in the factory assures a high level of workmanship
e.g. with regard to the corrosion protection and the concrete quality such that little maintenance
works is expected. At a later stage, composite girders are easier to inspect due to their open
structural layout than prestressed concrete girders, which has become an important issue for
bridge owners and operators due to a high number of damage patterns at prestressed concrete
bridges in the 1990s.
Depending on the boundary conditions of the bridge location, the advantages of the prefabricated
composite girders become fully effective when they are consistently utilised. Thus, they become a
competitive solution in comparison with prestressed concrete girders. The following choices and
advantages should be consistently considered when designing a prefabricated composite girder:
• Use of high-strength steel grades
• Cheap foundations due to a light weight superstructure
• Few obstructions of traffic due to short erection times
• Well-defined pricing and cost calculation possible due to prefabrication
• Advanced structural systems: bridges with integral abutments which reduce the deformation
and vibration of the slender superstructure
35
• Advanced structural systems: bridges without middle piers so that the traffic is neither blocked
during erection nor during maintenance e.g. of the bearing at the middle pier
• Low maintenance costs for integral bridges without bearings and expansion joints
In Table 2-5, examples of published bridge data are given which were successfully accomplished.
However, they represent only a small part of the bridges built with prefabricated composite girders. In
general, a composite bridge becomes an economic alternative to concrete bridges when no ground-
based falsework is feasible for casting. It can be seen that all types of structural systems are built
although continuous beams and frame systems are more preferable than simple beams. An outstanding
example of a continuous twin-girder system is the bridge Oberhartmannsreuth [2], which is a long
multi-span valley bridge. For a number of bridges crossing e.g. motorways a frame system is often
chosen [7].
Table 2-5: Data of bridges with prefabricated composite girders and in-situ concrete in Germany.
Bridge Year of Total Spans Structural Deck Costs

comple- length system area
tion
[m] [m] [m2] [€/m2]
Schmuckerweg n.a. 33.5 33.5 simple 436 n.a.
Regen n.a. 107.6 33.4-40.8 continuous 1399 n.a.
Oberhartmannsreuth 2000 201.6 34.8-44.0 continuous 5860 n.a.
Unterhaching 2000 44.5 44.5 frame 614 1,873
Bergkirchen 2002 46.5 46.5 frame 465 ~ 2,000
Sulzemoos 2002 48.5 48.5 frame 465 2,120
Further development of the prefabricated composite girders aims at a reduction of the amount of
welding and simple shear connections. As a result, rolled steel girders are cut in the web in such a way
that the generated geometry can be used as concrete dowels. An example of this bridge type is the
bridge Pöcking described in [86].
2.5 Spain
2.5.1 Introduction
According to the available information from several sources by the Ministry of Public Works, it is
possible to provide some details with the aim to offer, at least, a qualitative analysis of the market share
of steel and composite bridges in the latest years.
The surface of the country is 505,000 km2 and Spain has more that 164,000 km of National Roads, that
figure does not take into consideration the city roads, managed by the city councils, which are about
490,000 km more. Moreover, the Spanish National Company of Railways (RENFE) has 15,700 km of
railways with a total amount of 6,401 bridges.
However, the information presented as follows has been obtained from the Spanish Road
Administration’s database of bridges built only in the National Roads. This database considers bridges
with span longer than 10 m and the public available information is up to1996:
36
Bridge types
• 8,500 bridges, by typologies, 420 of them being pedestrian bridges:

o 5,509 I-girder bridges,
o 506 box-girder bridges,
o 5 cable-stayed and
o 17 truss bridges.
• Regarding the structural material, the vast majority are made of concrete and the figures about
steel and composite bridges up to 1996 are as follows:
o 82 are steel bridges, see Table 2-6, and
o 159 are composite bridges
• Other information provided by the database about the structural typology is the following:
o Average maximum span: 23.66 m.
o Average length: 69 m.
And focusing on steel and steel and concrete composite bridges, the steel bridges have an average
maximum span of 41.11 m and an average length of 109.80 m, for the steel and concrete composite
bridges those values are 38.50 m and 102.80 m respectively.
Table 2-6: Steel bridges in Spain by typologies (Information up to 1996).
Structural typologies Number of existing

steel bridges
Girder Bridges 33
Box-girder Bridges 25
Cable-stayed Bridges 1
Truss Bridges 17
Other 6
Total 82
In addition to the information presented up to 1996, is it possible to have some information about the
activities related to the bridge construction in most recent years. The Ministry of Public Works, by
means of the Spanish Road Administration, build between 300 and 400 bridges per year in the National
Roads and about 10 to 20 of them are made of steel or steel and concrete composite structures. Those
data could help to make an estimation of the current market share of steel and steel and concrete
composite bridges, less that 3%.
In Spain is generally acknowledged that the market share of steel and steel and concrete composite
bridge is not representative of the technical capabilities of the civil engineers, steelwork companies and
general constructors of the country. There are some commonly agreed explanations to this situation,
[58], [60]:
• Although in Spain there are more than 40 years of experience in outstanding composite
structures, the engineering cost for the design of composite bridges is more expensive than for
the concrete based alternatives.
• The strict box type composite solutions, which have been more developed in Spain, provide
additional advantages such quality assurance due to the industrialised construction, reduction of
construction time, and indirect costs, thanks to the highly prefabricated components used,
37
minimisation of material employed but all these advantages are not easy to be taken into
consideration in the budget and, at least up to now, a competitive budget is the key issue.
• Finally, composite bridges are seen as the preferred option in urban areas, by their aesthetical
value, or in seismic areas or where tall piers or deep foundations are required, due to the notable
reduction in weigh through composite solutions. But, for more conventional medium span
bridges, composite bridges are rarely built Spain.
According to design experts and the Spanish independent steel promotion association, APTA [58], [60],
“our country’s experience, and that of our European neighbours shows that there must be a decided
and active impulse by the steel sector with ensuing support and backing by the relevant authorities, in
order to ensure the introduction composite bridges.
The French experience shows that through the systematic use and optimization of these solutions on
high-speed train lines, it has been possible to develop a technologically advanced and very competitive
“national” bridge type, which may be potentially exported in the near future for use in railway
infrastructure throughout East Asia, Eastern Europe or South America.”
2.5.2 Spanish regulatory frame for bridge design

The Spanish regulatory frame for steel and steel and concrete composite bridge design is based on the
following documents:
• To determine the actions, the compulsory document is the “Instrucción sobre las acciones a
considerar en el proyecto de puentes de carretera” [42].
• For steel bridges the design recommendations are covered by the “Recomendaciones para el
proyecto de puentes metálicos para carreteras” [77].
• In case of composite bridges, the relevant document is the “Recomendaciones para el proyecto
de puentes mixtos para carreteras” [78].
• Finally, the Ministry of Public Works edited a handbook to help civil engineers in the use of the
Spanish bridge design recommendation: “Manual de aplicación de las recomendaciones RPM-
RPX/95” [79].
The relationship between the Spanish design recommendations and the relevant Eurocodes for steel and
concrete composite bride design can be discussed in this Design Manual thanks to the background
provided by a study carried out by the Ministry of Public Works in collaboration with the Spanish
Structural Concrete Scientific and Technical Association, ACHE [1], [17].
In principle, the basis of design of the Spanish documents are harmonised with those settled in the
Eurocodes. In fact, the Spanish documents were drafted in the period of the transformation of the
Eurocodes from ENV version to EN version, and because of that reason it was possible to take benefit
of the draft versions of the Eurocodes, prEN, to develop the Spanish documents.
However, there exist differences between the Eurocodes and the Spanish recommendations, mainly
regarding the detailing of the bridge. To analyze the real impact of the final design due to the
differences between the Spanish recommendations and the Eurocodes, the case study showed in the
Figure 2-47 has been analysed in detail by the Ministry of Public Works and ACHE.
The main conclusions derived from this case study showed that the differences in the verification of the
limit states according to the Eurocodes or the Spanish recommendations result in a very limited impact
in the resulting final design, see Figure 2-48. The main differences detected in the case study are the
following:
• Basis of design: the Spanish recommendations are harmonised with the Eurocodes.
• Actions and combination of actions: the live load is much greater in EN 1991-2 [29], than in
IAP [42]. However, additional differences on the partial factor for actions and combination of
actions, ψ, and for materials, γ, diminish the effect of the bigger live load of the EN 1991-2 and
the ULS safety reserve is quite similar as consequence.
38
Bridge types
• Actions due to vehicles: EN 1991-2 [29] establishes an adjustment factor for heavy vehicles
and another for uniform live load, that approach offers the possibility to adapt the values of
these load models to the specific Spanish expected traffic. For the case study presented, it was
taken the default value: 1.0
• Section resistance and effective width: there are differences in several aspects related to the
formulations regarding the design of the cross section such as the effective widths, the
evaluation of creep and cracking, etc. However, the final results are quite similar in the case
study analysed and, in general, the methodology and approach has a common conceptual basis
in the Spanish recommendations and in the Eurocodes for bridge design.
• Bending moment resistance of the cross section: for this issue there exists a very important
difference between the Spanish recommendations and the Eurocodes. The Spanish
recommendations for composite bridge design RPX [78] provides the so called “elastoplastic
method (EP)” for which, as opposed to the Eurocodes, it is not necessary to classify the cross-
sections and the effective cross section is obtained from the strain distribution. By using this
approach, the discontinuities at the limits between different cross-section classes subsequently
disappear.
• Shear connection:
o The approach of the Spanish recommendations settled in the RPX is easier to use than the
provided in the EN 1994-2 [36] and, in addition, the RPX offers some coherence because
in the calculation of the resistance of the shear connection the bending resistance of the
cross sections is the value taken into account. The Eurocode, instead, uses the acting
moment. For the case studied, the RPX approach does not have a relevant additional cost
in the final result of the shear connection in comparison with the result of the Eurocode.
o However, the EN 1994-2 approach provides a better shear connectors distribution for the
ULS elastoplastic distribution of longitudinal shear per unit length, while the SLS of the
shear connection is assured by an additional condition. According to this conclusion, the
RPX procedure could be improved by the amendment of supplementary limits with regard
to the minimum number of shear connectors in the mid - span sections.
o With regard to the design of the sections submitted to hogging bending moments, the
method provided by EN 1994-2 [36], needs to be amended in order to clarify how the shear
connection has to be designed when the acting moment is larger than Mel,Rd.
Figure 2-47: Case study for comparison between Spanish recommendations and Eurocodes for
steel and steel and concrete composite bridge design [1], [17].
39
All the previous conclusions have been derived after the analysis in detail of the case study presented in
the Figure 2-48, obviously, in order to evaluate the consequences of applying the Eurocodes to
composite bridge design in Spain and, moreover, to establish the economic impacts derived from the
use of Spanish recommendations or the Eurocodes, it would be necessary to carry out further analysis
considering parameters such us span, bridge typology, detailing and fabrication costs of the structure…
But for the aim of this Design Manual, these conclusions are enough to detect and to underline the main
differences between the Spanish recommendations and the Eurocodes for bridge design.
Figure 2-48: Final design of the sections and arrangements of stiffeners and diaphragms [1],
[17].
2.5.3 Bridge examples

The preference in Spain for the box sections in steel and in steel and concrete composite bridges and in
addition the frequent practice of the double composite action in bridge structures is widely
acknowledged [5], [80]. The Spanish civil engineers Martinez Calzón and Fernández Ordoñez, , see
www.mc2.es and Figure 2-49, adopted 35 year ago innovative approaches for composite bridges.
© Millanes Mato, F; IDEAM, S.A.
a) Puente de Juan Bravo, Madrid, 1970 b) Puente del Diablo, Barcelona, 1972
(depth-to-span ratio L/h = 40 , weathering steel). (100 m span and twin composite triangular
boxes).
Figure 2-49: Innovative designs for bridges in Spain [60].
40
Bridge types
c) Puente sobre la Ría de Ciérvana, Bilbao, 1978

(first example worldwide of double composite action reported).
Figure 2-49 (continued): Innovative designs for bridges in Spain [60].
In order to explain in a more detailed way the use of the double composite action, the example of “The
Tina Menor viaduct on the Cantábrico highway” presents a complete case study to illustrate this
typology of composite bridge with a box section.
In Spain, an innovative approach has been developed for the double composite action that is the so
called “strict box composite bridge” [54], [59] consisting of a hybrid between the French twin-girder
system and the Spanish double composite action box type, which combines the response of the open
section at mid-span with a double composite closed section at the area of supports submitted to hogging
bending. An additional innovation consists of replacing the lower steel torsion bracing at mid spans by
prefabricated slabs, see the example of the “Arroyo de las Piedras viaduct on high speed railway
between Córdoba and Málaga” in Figure 2-51 where a case study to illustrate the use of the “strict box
composite bridge typology” is presented.
Figure 2-50: Puente Betxi Borriol, Valencia, Spain [60].
41
Although the contents of this Design Manual are mainly aimed to provide guidance about twin - girder
and box-girders, in the Figure 2-51 there are other examples of Spanish applications of composite
bridges: “Abaqus” system for variable depth launching, composite arch bridges and cable stayed
composite bridges.
The following sections, as stated before, are focused on those Spanish specificities and presented
examples aim to offer a vision of the state of the art and current practice on most widely built composite
bridges in Spain, with the limitation of this Design Manual able to address only a few case studies.
© Martinez Calzón, J;
MC2 Estudio de Ingeniería S.L.
a) Puente del Arenal, Cordoba, 1993, and scheme of the “Abaqus” system for the launching of bridges
with variable depth [55].
© Manterola Armisén, J; CARLOS FERNANDEZ CASADO, S.L.
b) Composite tubular arch: Puente sobre el Escudo, Unquera, 2001 [52].

© Pantaléon Prieto, M.J; APIA XXI, S.A.
c) Double composite box section: Puente sobre el río Tajo, Alcántara Reservoir, 2006 [51].
Figure 2-51: Further examples of composite bridges in Spain.
42
Bridge types
© Llombart Jaques, J.A; EIPSA
d) Cable stayed composite bridge: Viaducto de Escaleritas, Las Plamas de Gran Canaria, 2007 [66].
Figure 2-51 (continued): Further examples of composite bridges in Spain.
The Tina Menor viaduct on the Cantabrico highway

The viaduct across the Tina Menor river allows the Cantabrico Highway to cross the estuary at San
Vicente de la Barquera, a small tourist city close to Santander. The owner of the viaduct is the Ministry
of Public Works and the design was done by Javier Manterola Armisén, Antonio Martínez Cutillas y
Armando López Padilla from the engineering company Carlos Fernández Casado [53], [56].
The total length of the viaduct, see Figure 2-52, is 378.50 m with four spans of 64.25, 125.00, 125.00
and 64.25 m. It is important to note that the first two spans are in a circular alignment of 600 m. radius
and the last two spans are in a transition curve with a change in the curvature sign.
The bridge structure is a continuous beam with a steel and concrete composite box section. The main
dimensions of the box cross section, Figure 2-53, are the following: a constant depth of 6.50 m and a
constant width of 10.00 m. The concrete slab has a total thickness of 0.32 m and was built by using
prefabricated ribbed slabs with a minimum thickness of 0.06 m and a additional thickness of 0.26 m
with on site concrete.
To achieve the 30.00 m wide corresponding to the two roadways of the highway, the box section has a
lateral truss, see Figure 2-54. In addition, the lateral truss contributes significantly to the torsional
stiffness of the section, approximately 25% of that provided by the box section considered alone.
The thickness of the box-girder bottom flange ranges from 20 to 30 mm and the web varies for 15 to 30
mm. The significant reduction of the bottom flange thickness, and the requirement of stiffening, is
achieved thanks to the use of the double composite actions concept. The sections at supports have been
provided with a reinforced concrete slab, connected to the lower plate of the box, with a slab thickness
that range from 0.20 m to 0.70 m.
The whole deck is prestressed transversally and the construction was made with temporary supports in
the first two spans and half of the third span. The construction procedure for the other parts of the
structure used was incremental launching due to the difficulty for erection in some areas of the bridge
location, see Figure 2-55.
43
Figure 2-52: The viaduct of Tina Menor, elevation and plan view [56].
Figure 2-53: The viaduct of Tina Menor, cross section and detail of the concrete slab in the
bottom flange (light grey) to obtain the double composite action for the cross sections closer to
the piers [56].
© Martinez Cutillas, A;
CARLOS FERNANDEZ CASADO, S.L.
Figure 2-54: The viaduct of Tina Menor, lateral truss [56].
44
Bridge types
© Martinez Cutillas, A; CARLOS FERNANDEZ CASADO, S.L.
Figure 2-55: The viaduct of Tina Menor, propped construction [56].
The Arroyo de las Piedras viaduct on high speed railway between Córdoba and Málaga
The Arroyo de las Piedras viaduct offers an innovative solution in steel and concrete composite bridges
for high speed railway lines. The design was done by Francisco Millanes, Javier Pascual Santos and
Miguel Ortega Cornejo from IDEAM, [59], [61], and is the first composite high speed railway bridge in
Spain.
According to the “new Spanish tradition”, the cross section adopted is the “strict box”, the double
composite action in box sections commonly used in Spain for composite road bridges. But in this bridge
several innovations has been adopted that provides in fact a new design approach for the double
composite action in bridges .
Before presenting the details of the cross section, it is relevant to explain in depth its concept; the
starting point of the design is the typical twin plate girder solutions, frequently used in France and
whole Europe, modified to be improved according the strict box-girder concept but keeping the
construction advantages of former. In addition to the use of the double composite action in hogging
areas another innovation in this bridge is to use it along the whole length of the bridge to provide the
torsional stiffness required due to dynamic actions and eccentrically loading for trains operating along a
single track.
45
The structure is a continuous composite beam with spans of 50.40 m, 17 x 63.50 m, 44.00 m and 35.00
m. At the time of the design and building stages it was the longest span viaduct of its type for high
speed railway bridges, 0.50 m more than the Orgon viaduct on the French TGV Mediterranée; regarding
the piers, several of them exceed 93 m, see Figure 2-56.
Figure 2-56: The viaduct of the Arroyo de las Piedras [61].
The cross section of the deck is made of two I-girders of 3.85 m depth and a concrete slab connected to
the top flange of 14 m wide, whose thickness varies from 0.41 m in the longitudinal axis to 0.22 m in
the edge of the cross section. The resulting composite cross section, see Figure 2-57, has a constant total
depth of 4.26 m.
Without a more detailed description, the adopted cross section could be understand as somewhat similar
to the typical twin-girder solutions frequently used in France but there are several interesting
modifications that will be described as follows, [61]:
• Cross bracings are used instead of full web diaphragms with the same depth as the main beam.
This improvement facilitates on-site assembly, reduces the weight of steel and, as consequence,
the volume of welding required. These cross bracings are located every 8 m along the bridge.
• The bottom steel truss is replaced by prefabricated slabs, 2 m wide and 14 cm thick. At mid
span, the prefabricated slabs are not connected together and only one metre of each is
connected to the bottom flanges to transfer torsion shear flows between the main girders and the
slabs. With this solution, the required torsional stiffness is thus guaranteed, and is even higher
than with the typical bottom steel truss.
a) Mid span.
Figure 2-57: Cross section of the Arroyo de las Piedras viaduct [61].
46
Bridge types
b) Support.
Figure 2-57 (continued): Coss section of the Arroyo de las Piedras viaduct [61].
• Regarding the steel girders, they have been fabricated with two external triangular corner cells
instead of longitudinal web stiffeners, see Figure 2-58. The cells improve the stability of
flanges and webs submitted to compression during the launching and in service conditions. In
addition, the joining collaboration of the bottom external cell and the bottom composite flange,
resulting from the previously described solution, provide an excellent system to improve web
resistance to concentrated loads during launching. On the other hand, vertical stiffeners are
adopted every 4 m.
• Regarding the top flange, transversal IPN sections are connected to the upper slab
approximately every 2 m to obtain a steel and concrete composite deck, leading to a reduction
of the total weight of the bridge.
© Millanes Mato, F; IDEAM, S.A. © Millanes Mato, F; IDEAM, S.A.
a) Twin plate girders adopted for the cross b) Girder with external triangular cells.
section.
c) Prefabricated slabs in the top flange. d) Prefabricated slabs and on site concrete in the bottom
flanges.
Figure 2-58: Erection of the Arroyo de las Piedras viaduct [61].
47
As summary, the typical cross section in hogging areas is similar to the mid span cross section, but with
the difference that the double composite action is provided by concreting in situ over the bottom flange
to achieve the full collaboration of the precast slabs. The bottom slab extends 13.90 m on both sides of
the pier in the 63.50 m spans, and a bit less in the shorter spans. The thickness of the slab varies from 25
cm at the end to a maximum of 50 cm in the section located at the piers. The slab is connected with
studs and passive reinforcement to the twin-girders flanges.
Thanks to the double composite action, the maximum thickness of steel plates in the bridge is 40 mm,
much thinner than the classical twin-girder solution. The construction procedure was the launching of
the bridge from both abutments, see Figure 2-59.
a) Several stages of the launching.
b) Detail of the auxiliary launching c) Operations during pier passing and

structure and the launching cables nose beam raising.
and jacks.
Figure 2-59: Launching of the Arroyo de las Piedras viaduct [61].
The steel used in the main structure has been S-355 J2G2W Cor-Ten, weathering steel. This steel is
appropriate for the climatic conditions of the area where the bridge is located. For the internal truss
diaphragms, carbon steel has been used, which makes it possible to reduce the cost of the bridge, after
having taken into account the cost of maintaining the carbon steel, which are perfectly accessible for
maintenance and inspection.
2.6 Sweden
2.6.1 Introduction
Sweden is large country with an area equal to France but with only 9 million inhabitants. One
consequence is that it has long roads with little traffic. Most roads have two lanes only and motorways
have two lanes in each direction. In the latter case it is common to build one bridge for each direction.
This solution is obviously more expensive than one wide bridge with all four lanes. Although this, it is
mostly preferred by clients because of the increased flexibility for maintenance and repair. This means
that Swedish composite bridges are quite narrow with widths not exceeding 13 m and the common
48
Bridge types
solution is a twin I-girder bridge. In average 100 road bridges per year are built in Sweden and the
market share for composite bridges is about 40%.
For a long time rail bridges were only built as replacements of old bridges but now there is renaissance
for rail roads. Currently a new rail road with length 190 km is being finished along the northern coast
including 120 bridges of which many of the largest bridges are composite bridges. Furthermore, on the
Swedish West Coast 140 new road bridges will be built.
2.6.2 Road bridges

The bridge at Rångedala is a typical example of a modern Swedish motor way bridge. It is located on
national road E6 in south western Sweden. It is a motorway with four lanes and the bridge is split in
two, each with two lanes, see Figure 2-60 and Figure 2-61. General data of the bridge is as follows:
Spans: 54+4x70+54 m
Width of deck: 9.75 m
Type: Twin I-girders 2.6 m deep
Steel: S460 and S355 hybrid girders.
Concrete: C35/45 1.30 m2 per girder
Steel weight: 1480 tons or 196 kg/m2
Figure 2-60: Cross section of the bridge at Rångedala, Sweden.
Figure 2-61: Bridge at Rångedala, Sweden, during erection.
The bridge was designed according to the Swedish design code Bro 2004 [9]. The traffic loads are a bit
smaller than those of EN 1991. The code requires elastic global analysis and elastic resistance but
allows the use of post-critical resistance for buckling. For breathing there is a limitation of the stresses
49
related to the critical stresses, which usually governs the web thickness in mid span. Fatigue from traffic
does not usually govern the design, here the number of cycles is set to 100 000 cycles or 400 000 cycles
(major roads like E4, E6 or some bridges in urban areas like Stockholm).
2.6.3 Bridges with integral abutments

The cost of maintenance is an ever-growing problem for road administrations around the world, and
bridges are no exception to the rule. One way to reduce the need for future maintenance, as well as the
investment cost, is to make bridges without transition joints. It is a cost effective design but one
question is how long the bridge can be allowed to be. This is not only governed by the absence of
expansion joints but also by yielding of the piles, which is caused by imposed deformations from
temperature and traffic. In the US the yielding from imposed deformations is ignored and a bridge up to
at least over 300 m has been built with this design.
In the 1980´s a few bridges with integral abutments were built in Sweden. Most of the short to medium
span bridges in Sweden are semi-integral, meaning that there are no joints on the road surface but that
the bridge rests on bearings. In a project at Lulea University of Technology it was investigated if piles
with cruciform cross section were suitable for integral bridges, including full scale tests. Within the
project a bridge was built in the Swedish province of Västerbotten, completed in September 2000. The
bridge was a single span composite bridge with a span of 37.15 m.
Figure 2-62: Integral abutment bridge which is supported by steel piles under the back wall.
In order to minimize the bending stresses arising from the deflection of the bridge, the work was carried
out in the following way:
1. Eight piles, X180·24 mm, were used for each abutment. The piles were rotated 45 degrees from
the line of support, minimizing the bending stresses from the traffic load.
2. The steel girders were erected on steel bearings on top of the lower part of the back wall. On
safe-hand side, the girders were designed as simply supported girders, not taking the restraint
from the embankment into account.
3. The formwork for the side wings was removed, giving the steel piles a rotation in the opposite
direction of the one arising from the traffic. In other words, the piles were pre-stressed to
compensate for later rotations from dead load and traffic.
4. The upper parts of the back walls were cast together with the concrete deck of the bridge.
5. The embankment behind the back walls was filled up, and the surfacing as well as the side rails
were placed on the bridge.
2.6.4 Bridges with full depth prefabricated deck slabs

Several road bridges in Sweden have been built with full depth prefabricated decks. Most of those have
had transverse joints with a gap of 400 mm in which hair pin stirrups have overlapped. The shear
50
Bridge types
connectors have been concentrated in groups typically located at the joints between the deck elements
with a centre to centre spacing of 1.8 m. After placing the elements on the girders the joints and the
holes for the shear studs have been concreted. Based on reach at LTU a new solution that brings the
technique a further step forward will be described here.
In 2002, the Swedish road bridge AC 1684 was built as a railway crossing, in Norrfors, replacing an old
narrow bridge in bad condition. It is a single span composite bridge with a span of 28 m. The bridge
deck was designed to be prefabricated, in 16 concrete deck elements, which were assembled with dry
joints. The construction costs were presumed to be a bit higher than for a conventional concrete bridge,
but since the disturbance of the railway traffic could be minimized it was worth trying the new concept.
One challenge was the requirement that the bridge should be assembled in less than 24 hours. The time
limit was governed by how long the electricity for the railroad could be switched off. A plan of the
bridge is shown in Figure 2-63. As can be seen the bridge is curved and the deck has a single slope,
which makes the geometry complicated both for the concrete elements and the steel girders. High
requirements had to be fulfilled by all of the steel parts that were in contact with the concrete.
Figure 2-63: Plan and cross section of the prefabricated bridge. Elements 1 and 18 are the
prefabricated back walls/end screens.
Figure 2-64: Plan of a bridge deck element and cross section above girder.
Each deck element had the dimension of 1800x7500x280 mm, giving an element weight of about 10
tons. The elements were tapered in plan because of the curve, see Figure 2-64. The elements have a
tunnel above the girders, which was grouted with a special concrete mix through 100 mm holes, and the
tunnel sides have a pattern to ensure shear transfer between the grout and the elements. In the joints at
the edges and in the middle of the elements, there are tongues and groves for transfer of vertical shear
between the elements. In order to make the elements fit the next element was cast with the edge of the
previous one as form work (match casting). The accuracy of this procedure was good enough and a
51
small prestressing of 600 kN by external devices at the ends of the bridge using the girders as ties left
very small gaps between the elements.
Figure 2-65: Bridge deck element being lifted in place.
Figure 2-66: Illustration of the narrow tolerances.
Figure 2-65 shows an element being lifted in place and Figure 2-66 shows the narrow tolerances
necessary to make the pieces to fit. The project was evaluated in order to gather the experiences and the
opinions about this type of prefabricated element bridges. The evaluation states that the most important
experience is that all of the actors must be aware of the aim of the project and their responsibilities. It is
necessary that all actors realize the importance of the demanded precision, which is much stricter than
in normal construction. There is no time for corrections during the assembly of the bridge. Although
everything was not perfect, the bridge came into place in time. However, the construction industry has
to get used to stricter procedure and tolerances before this kind of industrialized concept can be used in
a larger scale.
2.6.5 Railway bridges

The Swedish national Rail Administration prefers concrete bridges but for medium and long spans over
water composite bridges are a competitive alternative. A recent example is the bridge over
Veckefjärden, which is a composite I-girder bridge on the new railway line mentioned in the
introduction. The total length is 490 m with the typical span 60 m. The girders were delivered in 24 m
long pieces with a shop splice in the middle. The 24 m long units had a weight of up to 65 tons.
In 2007 the Swedish Code for Railway Bridges changed the fatigue spectra from 106 to 2·106 cycles,
with the consequence that steel grades higher than S355 can not be fully utilized. This bridge was
however designed in 2004 and S420 was used for bottom flanges over support and for upper flanges in
mid span.
The spacing of the I-girders is 2.5 m, and the depth of the girders is L/20 = 3.0 m, which is more than
would have been used for a road bridge. Since the alignment of the rail track goes from a horizontal
52
Bridge types
radius of 3,220 m at one end of the bridge to a straight line at the other end, the main girders are straight
between the points 6 m from support. Since the bridge was push launched from both ends, a special
adjustable hinge with vertical axis made it possible to adjust for the alignment both with and without
radius. After launching the hinge was removed and the girders were welded together.
Figure 2-67: Elevation and cross section of the bridge over Veckefjärden. The cross bracings
are made of channels bolted to the web stiffeners.
Figure 2-68: Picture of bridge over Veckefjärden during wintertime.
Since the temperature wintertime can fall to -30 oC, it is very practical to cover the working area with
big tents, which not only makes it warmer, but also eliminate winds, rain and snow. To avoid too big
differential temperatures between the steel girders and the concrete, the underlying girders are also
preheated. If the steel is not heated the effect of the differential temperature will be the same as a
magnified shrinkage of the concrete.
2.6.6 Special bridges

Most bridges are designed without special concerns of aesthetics. The only formal requirement based on
aesthetics is that stiffeners should be placed on the inside of girders. Architects are rarely involved in
bridge design except for very big bridges. Occasionally bridge owners are prepared to pay a bit extra for
beauty. Beauty is in the eye of the beholder, as the saying goes, and clever choices of bridge colour and
form can yield an interesting result.
In 1993, the Swedish Steel Construction Institute published ”Steel Bridges”, proposing means of
making existing and planned bridges more aesthetically pleasing by means of well-planned steel
alternatives. One of the most effective ideas involved replacing the usual concrete 3 span bridge with
inclined legs acting as intermediate supports by a composite bridge with inclined steel legs. This type of
structure is very common for highway crossings. The architects leapt at the chance to try this concept in
the bridges of the Höga Kusten project.
53
Höga Kusten project
Figure 2-69: Bridge over E4 at Höga Kusten, Sweden.
The arch legs carry the load from the center of the bridge, and although they might as well be straight
from a static point of view, the curved arch structure undeniably lends a more appealing impression,
which is enhanced further by the slender dimensions of the steel. For transportation, the steel was
divided into one girder over mid span, two approach girders and two curved legs.
Vallsundet
The final case concerns the 1,500 m long bridge over Vallsundet in Sweden, connecting the island
Frösön with the mainland. This bridge has a box in composite action with the concrete deck. The bridge
is mainly characterized by the soft vertical radius and the triangular supports under the highest point. In
addition to being aesthetically pleasing, they facilitated the introduction of an extra wide span across the
navigation channel.
For composite boxes, the concrete is often cast between the webs by means of trapezoidally profiled
sheeting. The sheeting is placed transversely across the bridge, mounted on trestles that are braced
against the box floor. The roadway cantilevers are cast using a form carriage, which is moved along
rails placed on the upper flange. For shorter bridges, two casting sequences per span is adequate: first at
mid span, then at supports, to minimize tensile forces in the concrete over support. In the case of longer
spans, two or more form carriages can be used simultaneously.
Figure 2-70: View of the Vallsundet bridge, Sweden, and the triangular supports over the
navigation channel.
54
Bridge types
Figure 2-71: Casting of the concrete deck of the Vallsundet bridge, Sweden.
Figure 2-72: Cross section as well as details over supports for the bridge over Vallsundet,
Sweden.
As shown in Figure 2-71, for casting of the concrete deck for a composite box a form carriage is used.
In between the webs 45 mm deep trapezoidally profiled sheeting acts as formwork for the concrete. The
plywood put on top makes it possible to transport re-bars etc before the casting of the deck.
Figure 2-72 shows the cross section as well as details over supports for the bridge over Vallsundet. The
shear force and torsional moment is transmitted to the bearings by two diaphragms with t= 18 mm (pos
1). The trapezoidal stiffeners prevent the bottom flange from local buckling, and are also air ducts for
dehumidified air. The inside is not painted. The inclined stiffeners (pos 4) provide lifting points for
replacing the bearings.
55
Steel grades
3 Steel grades
3.1 Introduction
In Europe the most commonly used steel grade in bridge design is S355. The use of higher steel grades
(S460, S690) differs from a country to another and depends mainly on national rules and experience.
For instance, the steel grade S460 is quite commonly used in France and in Sweden it is the standard
grade whereas in Germany its use is much more ungenerous. Besides that, it has to be noticed that the
European standard EN 10025 [37] deals with steel grades up to S960 and EN 1993-1-12 [34] adds
specific rules to use steel grades up to S700 when designing a steel bridge according to the Eurocodes.
For composite bridge, EN 1994-2 limits the use of steel grades up to S460. Even in case of the use of
S460 the bending resistance may be reduced by the β factor (see EN 1994-2, 6.2.1.2(2)). The
experience is very limited for composite sections beyond grades S460 where the elastic bending
resistance should be used by the time being. Moreover it should be noticed that other branches (mobile
cranes, offshore structures, or shipbuilding industry) already use steel grades up to a yield stress of 1100
MPa.
In the following, the advantages and obstacles for using high strength steel (HSS) will be first discussed
from the design and economic points of view. Then national trends will be illustrated through European
bridge examples using such HSS. Accompanying the steel grades, quality and through thickness
properties should also be specified when designing a bridge. In the future this will be performed using
EN 1993-1-10 [33].
3.2 Discussions about the use of High Strength Steel (HSS) in

bridge design
3.2.1 General
The increase of the steel strength can lead to material savings, and then can reduce the fabrication costs
(time for welding, areas to be painted,…) and the erection costs of a bridge (less material to handle and
transport, reduced weight simplifying the erection, less costs for foundations,…). The structural
elements become lighter and more slender enabling special aesthetic and elegant structures.
Constructions with less steel are also in good agreement with the sustainability problematic and a
reduced consumption of the world's natural resources. It has been shown that HSS can exhibit not only
a higher strength but also an excellent toughness and superior welding properties, so that a high safety
both in fabrication and in structural design is ensured. The material savings also reduce the values of
internal forces and moments in the zones surrounding the intermediate supports of the bridge. This
finally leads to an increase in the competitiveness of a steel or composite bridge using HSS.
In order to study the economy of using high strength steel, an estimate of prices is needed, which is a
quite intricate question. The price of structural steel usually increases with the strength, which can be
seen from Figure 3-1 [44]. It shows relative prices for heavy plates from three leading European
producers of high strength steel in which S235 has been chosen as reference. Figure 3-1 also shows a
trend curve, which follows the square root of the yield strength. There is a substantial scatter in prices
from time to time due to the market situation and the marketing strategy of the producer. The
production cost increases mainly when the production process changes e. g. from TM to QT. Also the
number of grades that has to be produced influences the production cost and it is a matter of strategy
where to allocate these costs. An unusual example is that you can buy S355 cheaper than lower grades
57
in the US. Anyway, the trend curve in Figure 3-1 will be used in this study as an evaluation of probable
prices.
If the strength can be fully utilised the cost of material will be lowered as the strength is increased, see
Figure 3-2. The cost of a structure depends however more on costs for fabrication and erection than on
the price of the material but here only the cost of material will be studied.
3.0 1.2
tonne[-][-]
Relative material cost [-]

2.5 1.0
/ Tonne
2.0 0.8
per
Preis
1.5 0.6
price
Bezogener
1.0 0.4
Relative
0.5 0.2
0.0 0.0
200 400 600 800 1000 1200 200 400 600 800 1000 1200
Yield strength f y [MPa] Yield strength f y [MPa]
Figure 3-1: Approximate price per tonne of hot Figure 3-2: Approximate material cost
rolled steel normalised with price of S235 as normalised with the cost of S235 assuming
function of yield strength. that the strength can be fully utilised.
The conclusion from Figure 3-2 is that increasing the steel grade saves costs as long as the strength can
be utilized. Limitations of utilizing the strength may be buckling (stability phenomena), fatigue and
deflection limits. Thus it has been shown e.g. in [47] for composite road bridges that the benefits of
HSS in application are mainly determined by the following aspects:
• Stability. In case of stability, critical loads are independent of the material strength so that for
slender structures the use of high steel grades becomes uneconomical.
A solution could be the use of hybrid girders, see 3.2.2 below.
• Fatigue. When using high steel grades, fatigue often becomes decisive because it is almost
independent of the base material strength. Therefore, it is reasonable to use high steel grades in
cases where the influence of fatigue is small such as large spans and/or small traffic loads as
well as for areas e.g. close to interior supports. In order to increase the fatigue strength of
welded structures, post weld treatment methods could be applied [25].
It should be noticed that the fatigue loads will differ from country to country, and from road to
road, depending on traffic intensity. The fatigue aspects will not be dealt with in more details
within the scope of this Design Manual.
• Deflections. The reduction in girder dimensions reduces also the mechanical properties such as
bending stiffness. This in turn leads to higher deflections, which might become a decisive
design criterion. The deflection limitations vary very much between the countries, see 3.2.1.
For the typical span range of railway bridges, strict deflection limits but also fatigue issues can make
high steel grades not economic because they offer no advantages in comparison with steel grades S235
and S355 [88].
3.2.2 Influence of the deflection limitations

The deflection limitations shown in Table 3-1 vary very much from country to country. The purpose of
the limitation may also vary but it is believed that the comfort of drivers and pedestrians is the main
reason for road bridges. For railway bridges, EN 1990/A1, A2.4.4 [28] gives detailed rules to cover
comfort, vibration, deck twist, rail-wheel contact, …
58
Steel grades
The limitations are usually related to the deflection caused by the characteristic traffic load but in Spain
they use the frequent combination. France has no limitation at all for road bridges and Sweden has
L/400 for characteristic traffic load. This poses no problem for composite bridges with steel up to S690
for short to medium spans. Spain on the other hand has limitations that depend on the traffic intensity
and structural system such that simply supported spans have stricter limitations than continuous girders.
The strictest limitation L/4000 for multiple simply supported spans in a highway is so strict that it
makes this system impossible irrespective of steel grade. The system is on the other hand not attractive
because of the maintenance of the joints and for instance in Sweden it is not allowed at all.
For rail bridges the dynamic behaviour is of concern and especially for high speed trains it is common
that a dynamic analysis is required (see EN 1990/A1 and EN 1991-2). The deflection limits for
Germany and Spain in Table 3-1 are valid for normal speed trains. Spain uses deflection limitations
substituting a dynamic analysis. These limitations range from L/600 to L/2400 depending on train speed
and bridge span. For cases where a dynamic analysis is not required a deflection limitation can be seen
as a substitute. Normally the limitations are stricter than for road bridges. The allowable deflection in
Sweden is L/800, which is half of that for road bridges. Sometimes this puts a restriction on the use of
higher steel grades but the fatigue is a more common limitation.
Table 3-1: Summary of national requirements and praxis for bridges.
Country Road bridges Railway bridges Hybrid

girders
Highest steel Deflection Highest steel Deflection
grade limit grade limit
allowed but
Belgium S355 to S460 L/700 S355 L/900
not used
EN 1991-2,
France S460 n.a. S355 no
EN 1990/A1
S355
(higher steel allowed but
Germany n.a. S355 L/600 to L/800
grades only not used
with a “ZiE”*)
Spain S460 L/600 to L/4000 n.a. L/600 to L/900 no
Sweden S460 to S690 L/400 S355 to S420 L/800 yes
*„ZiE“ means „Zustimmung im Einzelfall“ (project-related expertise)
3.2.3 Influence of buckling and possible use of hybrid girders

The thickness of the web in a plate girder is governed by the required shear resistance. The shear
resistance is governed by buckling and this is taken according to EN 1993-1-5. The web is assumed to
be unstiffened except at support and the end stiffener is assumed to be non-rigid. The material cost is
taken from the curve in Figure 3-1 and the relative material cost for a web with a given shear resistance
is shown in Figure 3-3. There is a cost reduction for stocky webs but such are not used for plate girders.
For hw/tw >60 the cost is independent of the yield strength. This result indicates that a hybrid girder with
high strength steel only in the flanges may be cost effective. EN 1993-1-5 [31] gives rules for hybrid
girders and recommends that the yield strength of the flanges should not be larger than two times that of
the web. A summary of the design procedure including simplified formulae can be found in [89].
59
1.1
hw/tw > 60
hw/tw = 60
Relative material cost [-]
1.0
0.9
hw/tw = 50
0.8
0.7
0.6
0.5 hw/tw = 30
0.4
200 400 600 800 1000 1200
Yield strength f y [MPa]
Figure 3-3: Relative material cost for web with flexible end stiffener subject to shear. Reference
cost is for S235 [44].
The use of hybrid girders with this difference between the strength of flanges and web implies that the
requirement that yielding should not occur in SLS is interpreted such that it applies to the flanges but
not to direct stresses in the web. The local yielding in the web is limited by the elastic strains in the
flanges and after the first yielding the behaviour is reversible. This reasoning is not always accepted and
the interpretation may vary from country to country. For instance in Sweden the limit is set at a flange
strength 1,5 times the web strength and in France it will be required that the web should not yield in
SLS.
In general it saves costs to use hybrid girders and this can be demonstrated with the following example.
Consider the pier section in Figure 3-4. It is designed in S460 and we will compare with an alternative
with web in S355. To make the comparison fair we have to change the web in S460 with an area
17,1x2491 mm², which has a shear resistance of exactly the required 5,05 MN. A web with the same
shear resistance in S355 requires an area 18,4x2491 mm². The use of a lower steel grade results in
yielding of the web close to the flanges as shown in Figure 3-4. If the yield strength of the flange 430
MPa is reached in the centre of the flange the stress at the edge of the web would have been 426 MPa
but it can only be 355 MPa and 71 MPa is “missing”. There is a missing triangular stress block with
depth 208 mm representing a force:
ΔF = 71⋅ 0, 208 ⋅ 0, 0184 / 2 = 0,136 MN
fyf fyw
tf
tw ΔF
hw
0.6·beff
r
ΔF
0.4·beff
tf
Figure 3-4: Stress distribution in the web of a hybrid girder in hogging bending.
60
Steel grades
The part of the effective web close to the flange is 322 mm so the missing triangular stress block is
within the effective width. The forces represent a bending moment of
ΔM = 0,136 ⋅ (2, 491 − 2 ⋅ 0, 208 / 3) = 0,32 MNm
This has to be compensated by an increase of the flange thicknesses of
ΔM 0,32
Δt = = = 0,33 mm
f y b f h 2,54.430.0,9
This can be considered as negligible in this case.
The change in cost can be calculated considering costs for material and splices only and it is shown in
Table 3-2 for the 16 m long piece.
Table 3-2: Change in cost when the web is changed from S460 to S355.
Item Amount Unit Cost/Unit [SEK] Cost [SEK]
S460 - 5,376 kg 011.5 - 61,824.-
S355 - 5,866 kg 009.4 - 55,148.-
Splice 1.0 h 450.0 - 0,0450.-
Total [SEK] 0- 6,226.-
Total [€] 0,0- 685.-
The saving is not very big, 9 EUR/m2 deck area, but it is a clear result and it is cost efficient to use a
hybrid girder compared to a homogenous girder.
3.2.4 Conclusions
The above discussion highlights advantages and obstacles for using HSS in bridge design. It is shown
that increasing the steel grade saves costs as long as the strength can be utilised. Limitations of utilizing
the strength may be buckling, fatigue and deflection limits.
In order to answer the buckling limit, it has been shown that the use of hybrid girders with a web in
S355 and flanges in S460 can make HSS economical even for slender structures. Finally the highest
useful steel grade for bridges will vary from country to country due to traffic intensity (fatigue) and due
to deflection limitations. The first is a fact of life but the second is less rational and a review of the
requirements could save expenses in some countries.
3.3 Steel grades used in European countries
3.3.1 Use in France

In France, the most commonly used steel grade for bridges is S355. The risk of brittle fracture is
controlled by quality selection in accordance with French standard A36-010 and the Ministry of
Transport recommendation CCTG-F66. This standard defines:
• The basic grade S355N (Charpy energy of 40J guaranteed at –20°C) for thicknesses of less than
80 mm (S355 K2 is possible and generally used for thicknesses less than 30 mm),
• S355NL (Charpy energy of 27J guaranteed at –50°C) for thicknesses between 80 and 150 mm
61
These allowable thicknesses are those of the French regulations. In Eurocodes, they have been
completely recalibrated but they do not differ too much (see EN 1993-1-10 [33] and EN 1993-2 [35]).
For bridges, the grades S420 and S460 were already permitted according to the previous French
standard A36-201 (dated 1972), but only with the quality N/NL. At that time the maximum usable
thickness was 50 mm, but this value has been successively increased up to 100 mm in 1984. However
these steel grades were practically not used for bridges because of a very bad weldability (however, one
example: 293 tons of S460N in 1979 for the Mathilde Bridge in Rouen), but rather for offshore
construction. The change came in 1993 with the publication of EN 10113 and new fine grain thermo-
mechanical steels (M/ML). In 1997 a French guidance book [65] facilitates the use of thermo-
mechanical steel (for grades S420 and S460) for bridges in France. Table 3-4 below presents a non-
exhaustive list.
Today higher steel grades (S690 and more) are not yet used for bridges in France, but a working group
is currently studying all the relevant aspects, specially welding procedures, for the use of S690 and also
hybrid girders.
3.3.2 Use in Germany

In Germany, the most commonly used steel grade for bridges is S355. Although DIN-Fachbericht 103
[23] deals with steel grades up to S460, the use of any steel grade higher than S355 requires a project-
related expertise [3]. However, in few road bridges high steel grades up to S690 have been used already
in parts of the structure especially where certain requirements had to be fulfilled, as described in Table
3-3. In general, the application of steel grades higher than S355 or even hybrid girders is not common in
Germany. Table 3-3 summarises bridge examples in Germany where steel grades higher than S355 have
been used.
Table 3-3: Data of German bridges with steel grades higher than S355.
Bridge Year of Characteristics Steel grade(s) Masses

comple- [t]
tion
Nesenbachtal 2000 Five-span continuous girder bridge with S355 J2G3 1341
a maximum L/h-ratio of 30, which S690 QL 284
required the use of S690 in support
areas
Wilde Gera 2001 Arch bridge with composite box-girder S355 insgesamt:
S460 6323
Reichenbach 2002 One-piece composite cross section with S355 J2G3 insgesamt:
tension bands made of S460 S460 M 6000
Airport bridge 2002 Cable-stayed bridge with a main span of S355 7180
Ilverich 287.5 m and a reduced pylon height due S460 ML 520
to the nearby airport. V-shaped pylons
with a tension band made of steel grade
S460 are used (for which only S355 was
considered to achieve redundancy)
62
Steel grades
Table 3-4: Data of French bridges with steel grades higher than S355.
Bridge Year Characteristics Steel grade(s) Masses

[t]
Remoulins 1993 First bridge with TM-steel, twin-girder S 355 M 200

bridge -1994 bridge, main span 80 m, max. plate S 460 ML 180
(Figure 2-1) thickness 80 mm.
Highway A16 1993 Rolled beams (max. web depth about S 460 M 980
-1994 1 m) used for the main girders of many
small span bridges comprising two or
more girders
Normandy 1992 Cable-stayed bridge, max. span 856 m S355/460NL 4000

bridge -1994 (624 m in steel), max. plate thickness S 420 M 1800
for in-span cross-sections in S420M:
30 mm, use of S420M decreased the
structure weight for the main span
Jassans-Riottier 2000 Composite twin-girder bridge, max. S 460 M/ML k.A.

bridge over the span 130 m, S460M/ML around
Saône river internal supports
Bridge of 1998 Steel arch bridge, closed steel box- Bogen in 350
Europe in -2000 girder, main span 202 m S460 M/ML
Orléans
Garrigue viaduct 1999 Composite twin-girder bridge, max. S460 M/ML 290
on Highway A75 -2001 span 74 m, max. plate thickness: 120
mm in S460ML
Verrieres 1998 Closed steel box-girder, connected to S460 M/ML ca. 2000
viaduct on -2002 a concrete slab, with a main span of
Highway A75 144 m, flanges of the box-girder
(Figure 2-2; section around internal supports: from
Figure 2-28) 30 to 67 mm in S460
Millau viaduct 2001 Cable-stayed bridge, max. span 342 Deck: S355 23500
on Highway A75 -2004 m, the central part of the steel box- S460 12500
girder section is in S460 M/ML
Pylons: S355 3200
S460 1400
Provisional S355 3200
supports: S460 3200
New bridge over 2001 Closed steel box-girder connected to a S460 M 1250
the Rhone in -2004 concrete slab, max. span 125 m, max.
Valence plate thickness in S460M: 60 mm
63
3.3.3 Use in other European countries

In Belgium S460 is allowed but in practice it is hardly used. S355 is the common grade.
In Sweden it is usually possible to use the full strength of S690 and still satisfy the fatigue rules for road
bridges. The most common material is S460 and the girders are frequently hybrid with S355 in the
webs. For rail bridges the fatigue requirements are stricter and the practical limit is S355 or S420 for
long spans.
3.4 Through-thickness properties

The selection of materials for through-thickness properties is dealt with in Sec. 3, EN 1993-1-10 [33].
The purpose of this selection is to avoid lamellar tearing resulting from strains in the thickness
direction. These strains may be produced by external forces but also by shrinkage of welds.
The susceptibility to lamellar tearing is highly related to the sulphur content, the limit of which is
defined in EN 10025 [37]. It is worth noting that these limits (around 0.30%) are much higher than the
content that can be achieved in modern steel making plants. Depending on what is stated in the National
Annex to EN 1993-1-10 [33] and on the project specifications, two routes may be followed (separately
or in combination):
• Fabrication inspection (generally by ultrasonic method)
• Specification of a Z quality in accordance with EN 10164 [38]
The second point is developed in the following for two welding details:
• Detail 1 is the connection of the 19-mm-thick web to the 35-mm-thick lower flange by two
7 mm fillet weld (single run)
• Detail 2 is the connection of the flange of the vertical T stiffener to the upper flange by two
15 mm fillet weld (multiple runs)
Detail 2
95
a = 15 mm
30
19
Detail 1
35 a = 7 mm
Figure 3-5: Two welding details of a twin-girder bridge. Dimensions in [mm].
64
Steel grades
In the following the different Zi coefficients are calculated for each detail.
Detail 1 Detail 2
Za = 3 (aeff = 7 2 ) Za = 9 (aeff = 15 2 )
Zb = 0 Zb = 0
Zc = 8 (no compression) Zc = 15 (Note: this weld could be stressed in
tension by traffic loads, see EN 1994-2,
6.6.1.1(13))
Zd = 0 (when the web is welded to the flange, Zd = 3 (medium restraint due to welding between
there is no restraint) the flanges of the main girders)
Ze = 0 (assumption preheating less than 100°C) Ze = 0 (assumption preheating less than 100°C)
Finally, ZEd = ∑Z
i
i = 11 Finally, ZEd = ∑Z
i
i = 27
According to Table 3.2 of EN 1993-2, a quality According to Table 3.2 of EN 1993-2, a quality
Z15 is necessary. Z25 is necessary.
3.5 Application to the calculation example “box-girder bridge“

The box-girder bridge used as example in Part I of the COMBRI Design Manual [16] has been firstly
designed using S355 which lead to double upper flange plates around the intermediate supports. After
discussing the use of HSS (see paragraph 3.2), it becomes interesting to re-design it and to show how
the competitiveness of the bridge can be improved.
The re-design has been performed using hybrid girders with the web and the lower flange in S460. The
upper flange is also in S460 except around the internal supports where the steel grade S690 has been
used to avoid the additional upper plates. The quality M has been adopted for S460 whereas the quality
QL1 is needed for S690. The optimization of the steel distribution has been performed with regards to
normal stresses, shear resistance and interaction between shear and bending at ULS, for the sections on
internal supports and for the in-span sections.
Table 3-5 illustrates the corresponding material savings which is quite significant, representing 30% of
the steel quantities or 7% of the total bridge deck weight. Table 3-6 illustrates the increase in the span
deflections under the characteristic traffic loads (LM1 model from EN 1991-2). This increase (25%)
remains reasonable with values lower than those summarized in Table 3-1.
Table 3-5: Comparison of the steel quantities.
Advanced design Reference design

in S460/S690 in S355
Steel for the main box-

girder (without taking the
2,470 tons 3,540 tons
bracing frames into
account)
S355 - 3,540 tons
S460 2,301 tons -
S690 0,170 tons -
65
Table 3-6: Comparison of the deflections at mid-span.
Advanced design Reference design Difference

in S460/S690 in S355
mid-
6.7 cm = L / 1350 5.2 cm = L / 1730 + 29%
span 1
Deflection under
characteristic mid-
13 cm = L / 925 10.5 cm = L / 1140 + 24%
LM1 span 2
mid-
14 cm = L / 850 11.4 cm = L / 1050 + 23%
span 3
As already mentioned in the discussion from Paragraph 3.2, the benefit of the 30% material savings
could be counterweighted by the fact that the high strength steels are more expensive. It becomes then
interesting to compare the prices of both designs. This comparison is illustrated in Table 3-7 and has
been performed on the basis of the material costs only (without including the delivery, the welding, the
erection…). Usually, in France for instance, these material costs represent around 30% of the global
price of the steel structure. However, there is also a substantial saving in fabrication and erection due to
decreased weld volumes but this has not been investigated in this comparison.
Table 3-7: Comparison of material costs.
Upper Bottom
Box-girder Whole Webs
flanges flange
S355 2,520,000 € 921,000 € 943,000 € 656,000 €
S460/S690 2,020,000 € 518,000 € 749,000 € 756,000 €
Difference - 20% - 44% - 21% + 15%
0%
90 m 120 m 120 m 120 m 90 m

-5%
Cost reduction [€/m]
-10%
-15%
-20%
-25%
-30%
0
0
0
0
0
0
0
0
0
0
0
90
0
0
0
0
0
30
60
30
33
36
39
42
45
48
51
54
12
15
18
21
24
27
Bridge length x [m]
Figure 3-6: Cost reduction by re-design the box-girder bridge in S460/690 instead of S355.
66
Steel grades
It should be noticed that the S460 web is less economical than the S355 web (+15%) as the shear
buckling verification does not allow an important decrease of the web thickness. At the level of the
whole steel cross-section, the most significant savings (25%) are observed for the section located
around the internal supports, see Figure 3-6.
Note: This result is slightly different from the conclusions in paragraph 3.2.2 for the for the following
reasons:
• The cost of S460 in comparison to S355 is slightly higher (price 2006),
• Vertical and longitudinal stiffeners are welded to the web.
Finally this example shows that the use of HSS could be a very interesting option with cost savings of
around 20% for the steel material supplies. Other costs for fabrication and for erection on site would
also be influenced in a positive way by the use of less material.
67
Flanges
4 Flanges
4.1 Introduction
Flanges play the most important role in the bending resistance of I- and box-girders. I-girders usually
have stocky flanges so that can be fully utilised whereas in case of box-girders, their overall size
determines if the flanges are longitudinally stiffened or not. For small box-girders without longitudinal
stiffeners, the ultimate behaviour is basically the same as for I-girders. Therefore, Chapter 4 will focus
on wide flanges, i.e. longitudinally stiffened plates under compression. The ability of code rules to
evaluate accurately the carrying capacity of compression flanges has been studied in the COMBRI
project, accounting for the interaction of local instabilities of subpanels and global instability of the
stiffened plate. Particular attention is put on the consequences of using advanced software tools to
evaluate the critical buckling stress of the system.
In case of wide flanges the shear lag phenomenon may be important in the determination of the bending
resistance as it results in a non-uniform distribution of stresses across the width of the flange. For
slender compression steel flanges, this phenomenon may also interact with plate buckling. A distinction
between the effective width resulting from shear lag and from plate buckling is made as follows:
• “effectives” denotes the effect of shear lag
• “effectivep” denotes the effect of local and/or global plate buckling
• “effective” denotes the effects of shear lag and plate buckling
The effective width influences the mechanical properties of the cross section which has to be considered
when calculating the internal forces within the global analysis.
4.2 I-girders
Generally flanges in I-girder are chosen to be in Class 3 or lower. If lateral torsional buckling is
governing a wider flange might be favourable but the normal solution is to chose the slenderness b/t
close to the limit for Class 3 and adjust the distance between cross braces such that lateral torsional
buckling does not reduce the resistance.
The rules for shear lag in EC 3-1-5 may give a reduction for shear lag also for I-girder flanges. Such
reduction is most likely not needed and common practice is not to check shear lag at all for I girder
flanges.
For aesthetical reasons the bottom flange, which is visible, should be made with a constant width. It
does not matter if the bottom flange in the sagging region is more slender than the limit for Class 3
because it is in tension. The top flange can have a variable width because it does not disturb the
appearance. In the sagging region the size of the top flange will be small and governed by lateral
torsional buckling during casting. In Sweden it is common to use 20x400 mm as minimum size and it is
sufficient for avoiding lateral torsional buckling for medium size bridges with 8 m between the cross
braces. After casting of the bridge deck the additional stresses in the top flange are small and it is
usually sufficient to use S355 even if the bottom flange is of higher grade.
I-girder cross sections sometimes lead to problems caused by the wide outstanding bottom flanges when
pigeons and other birds settle and nest there. Looking at old bridges and railway stations in urban areas,
one can imagine that this is not only an aesthetical problem but also a matter of corrosion because the
birds’ dirt reduces the reliability of the corrosion protection significantly in the long run. Therefore, a
69
significant effort is undertaken for cross-sections today to protect these areas by steel gratings or
meshes. Another solution is to weld inclined plates between web and flanges as shown in Figure 4-1. As
these inclined plates become fully load-bearing, they are subject to the same execution standards for the
welds. However, at the detailing points the geometry is complex and difficult to weld. Moreover, the
inspection of welds is not possible anymore for the web-flange connection. For these reasons, steel
gratings and meshes should be preferred for protection from birds.
© KE
Figure 4-1: I-girder with inclined flange plates at the bottom flange.
4.3 Bottom plate of box-girders © KE
4.3.1 General
As already stated box-girders are only economic in special cases for instance in curved bridges. For
road bridges it is common to make the webs inclined. Beside aesthetic reasons this practice leads to a
more favourable cross section geometry of the box-girder in terms of the bottom plate widths. Whereas
for the bridge deck a wide distance between the top flanges is more favourable, it is usually the exact
opposite for the bottom flange. One reason is to avoid longitudinal splices in the bottom flange. For
continuous girders the bottom flange suffers local buckling at the piers and for that reason a narrower
flange gives a smaller width to thickness ratio. The effect of shear lag is another reason for choosing a
narrow flange. The inclination of the webs can be up to 30-40o and it is limited by the increase of the
width of the web plate.
In spite of a possible web inclination bottom flanges of box-girder are usually so wide that shear lag is
significant in the serviceability limit state according to rules given in EC3-1-5. For the ultimate limit
state limited plastic deformations may be allowed, which limits the effect of shear lag, see EC3-1-5
section 3.3.
4.3.2 French practice

In France, for the large bridges built since 1995, the box-girder bottom flange longitudinal stiffeners are
either flat stiffeners, or trapezoidal stiffeners, or T-shaped stiffeners. The stiffeners are continuous along
the whole bridge.
For the flat and T-shaped stiffeners, classical transverse spacing is around 0.6 m, but no more than 1.0
m. The lower transverse width of the trapezoidal stiffeners is in most cases between 400 and 600 mm,
like the clear distance between two trapezoidal stiffeners. The bracing frames longitudinal spacing is
usually between 4.0 and 5.5 m.
Two pictures of typical bottom flanges stiffeners are given in Figure 4-2 and 4-3:
70
Flanges
© Sétra © Sétra
Figure 4-2: Box-girder with open stiffeners, Lille, Figure 4-3: Box-girder with closed
France. stiffeners, Millau viaduct, France.
4.3.3 German practice

In Germany, bottom plates of steel and composite box-girders are designed according to the design
requirements given in the Eurocode-based “DIN-Fachbericht 103” [23]. Unlike for orthotropic decks,
the code gives no direct recommendation concerning the design and detailing of longitudinal stiffened
bottom plates. The design of the plate is only restricted in terms of the static requirements of the plate,
such as its buckling resistance. No specific recommendation concerning the size and number of
stiffeners is given.
Common practise in Germany for road bridges is the use of trapezoidal stiffeners with Class 1-3 cross-
sections with typical heights and thicknesses of about hst = 200 ÷ 300 mm and tst = 6 ÷ 10 mm. So far
the use of bigger stiffeners with Class-4 cross-sections is quite uncommon in Germany and only used
for very big bridges. Flat stiffeners with class-1 cross-sections are used as well, but mainly for rail way
bridges.
A sketch of a typical cross section of a German composite box-girder bridge is given in Figure 4-4. The
distance between cross-bracing frames usually is between 3 and 5 m.
55cm 40cm
2,5% 25cm
6.80m
4.30m 8.15m 4.30m
Figure 4-4: Typical cross-section of a German composite bridge with box-girder and
trapezoidal stiffeners [39].
4.3.4 Swedish practice

Box-girders are used occasionally in Sweden for road bridges as well as rail bridges. For road bridges
the trapezoidal shape shown in Figure 4-5 is common. The stiffeners on the bottom flange are usually
fairly large cold formed trapezoidal profiles. It is common not to paint the inside of the box but instead
71
to provide dehumidifying. The trapezoidal stiffeners can then be used as air ducts in order to spread the
dry air uniformly in the box. There is also a requirement that the concrete slab should be protected from
excessive drying. This can be done with a steel plate on the top of the box. It is usually a trapezoidally
corrugated sheet that also serves as lost form work and it creates a closed box during the erection of the
bridge. This is essential for preventing lateral torsional buckling. The shear centre of the open box is
well below the bottom of the box and a buckling mainly consisting of rotation around the shear centre
gives usually a very low critical load. Therefore the box has to be closed during erection or launching.
Note that the stiffness of such sheeting is essentially influenced by its connections and it is much
smaller than that of a solid steel plate with the same thickness. Another solution is to provide a lattice
between the top flanges but this is more expensive and it is only used when the strength of a
trapezoidally corrugated sheet is insufficient.
= = = = =
Figure 4-5 Box-girder with trapezoidal stiffeners on the bottom flange.
4.3.5 Conclusions and design recommendations

The common practice for design of longitudinal stiffened bottom plates is still quite different in
different European countries and is influenced by the specific construction techniques and traditions.
General recommendations concerning number and size of stiffeners are hard to give as they are
dependent on many factors, e.g. the distance between the diaphragms. Focusing on the static
requirements of the bottom plate, the following conclusions and recommendations based on the
Eurocode design rules and the outcome of the COMBRI research project can be given:
1. Stiffeners are almost always needed on the bottom flange of continuous bridges because the
drop-off in resistance with increasing slenderness is quite sharp.
2. For an adequate stiffening of the bottom plate with open stiffeners the stiffeners have to be
designed to avoid tripping. The rules for a general case are given in EN 1993-1-5 and they are
intended to ensure Class 3 behaviour. The rules are quite severe for flanged stiffeners as T or L
shapes. In order to fulfil them the critical torsional buckling stress usually has to be calculated
considering the restraint from the bottom flange plate. As an alternative closed stiffeners can be
used and they may even be Class 4 stiffeners if local buckling of the stiffeners is taken into
account.
3. In case of stiffened plates an appropriate numerical simulation leads to higher and more
accurate critical buckling stresses than hand-calculation methods. A convenient tool for the
determination of all kinds of critical buckling stresses of stiffened plates is EBPlate developed
within the scope of this project, see [26].
4. The torsional rigidity of the stiffeners may be taken into account, which leads to even more
beneficial results compared to hand calculation methods, especially in case of large trapezoidal
stiffeners. The mentioned program EBPlate gives the critical buckling stress considering the
torsional stiffness and cross sectional distortion of the stiffeners.
5. Due to the mechanical model behind the calculation procedure of the EC3-1-5 design rules, the
use of very weak stiffeners leads to unsafe results. The effect of unsafe results becomes less
72
Flanges
relevant with increasing number of stiffeners. Only stiffeners with a minimum rigidity γ L ≥ γ L*
should be used where γ L* is the relative stiffener rigidity, which raises the global plate buckling
stress σcr,p,global up to the level of the local buckling stress σcr,p,local, thus σcr,p,global ≥ σcr,p,local.. The
exact value has to be determined for every specific cross section. A good approximation for the
pre-design the required minimum stiffener rigidity can be assumed to be γ L* = 25 in case of
open stiffeners and γ L* = 50 in case of closed stiffeners.
6. The slenderness of the sub-panels at which reduction starts is b/t = 42 · ε. The reduction is quite
steep when this limit is passed and it is often favourable to stay below this limit. This can be
achieved conveniently with large trapezoidal stiffeners as shown in Figure 4-5.
7. Large stiffeners do have the advantage of allowing a larger distance between the cross braces or
diaphragms, if there are no restrictions from other cross-section members e.g. the top-flange.
4.4 Double-composite action
4.4.1 General
The expression double composite action refers to a girder with two composite flanges connected with
steel webs. The basic driver is that concrete is cheaper than steel when it comes to carry compression.
Starting with a normal composite bridge with a concrete slab on top of the girder serving as deck and
flange a concrete slab may be added to the lower flange in areas where it is in compression. In large
span box-girders this has been used in Germany. In addition to add area to the bottom flange the
concrete can also be used to prevent local buckling of the flange.
Another application of double composite action has been used in France, which can be described as a
prestressed concrete bridge with a steel web. In order to avoid that too much of the prestressing is lost
to the steel the web has to be flexible in the axial direction.
The use of double composite started quite recently and the technique is not yet fully developed.
Experiences from France and Germany are described below and finally recommendations for design
and erection are given.
4.4.2 Use in France

In France, for railway bridges, a lower concrete slab is sometimes added between the two lower flanges,
but this design is not current. This lower slab is connected to the main webs using horizontal studs, and
sometimes to the upper face of the lower flanges. However, the slab segments are only added in the
zones surrounding the intermediate supports, where they are in compression. Moreover, the transverse
joints are not concreted between the different slab segments, so that they don't really perform a double
composite action, see Figure 6-4.
This design ensures a better torsional behaviour of the bridge deck, a reduction in the noise emission
(particularly important for high speed train lines) and an improvement of the dynamic behaviour of the
bridge. It is also used to prevent the local buckling of the lower flange in compression in the zone
surrounding the internal supports.
4.4.3 Use in Germany

In Germany, some box-girder bridges have a concrete bottom slab. These bridges are often haunched in
the longitudinal direction and have a so-called double-composite cross section at the supports [46]. The
general characteristics of such bridges are:
• Stiffness. Internal forces are attracted by the double-composite cross section due to its stiffness
which leads to an increase of bending moment at the supports and a reduction at mid-span.
73
• Stability. The transfer of forces from the steel bottom flange to the concrete bottom slab
reduces the steel flange thickness and increases the buckling resistance of web and bottom
flange.
• Construction. The layout of the transition zone between steel and composite cross section is
complex especially with regard to longitudinal stiffeners, cross frames and reinforcement.
• Erection. The erection time is increased because the reinforcement has to be placed through the
steel cross frames and additional concreting phases are necessary near the supports.
As bridges with double-composite action represented a new constructional type in Germany in the mid-
1980s, accompanying research projects have been carried out with regard to three bridges which have
been built between 1987 and 1995: the Inn river bridge Wasserburg, the Elbe river bridge Torgau and
the Mosel river bridge Bernkastel-Kues. In the following their characteristics and the outcomes of the
research projects are summarised on a bridge-by-bridge basis, see also [62], [63].
Inn river bridge Wasserburg. This is a steel-box-girder bridge with a constant depth, which has a
concrete top slab prestressed in the transverse and longitudinal direction. The concrete bottom slab has
a length of 20.38 m at each side of the pier, which corresponds to 0.25 and 0.20-times of each span
length. The slab thickness starts at 20 cm and increases over a length of 16.65 m to 65 cm and then to
200 cm. The cross girders at the bearings are made out of concrete. An elevation is shown in Figure 4-6.
Elevation
83.0 m 104.5 m 104.5 m 83.0 m

375.0 m
Figure 4-6: Elevation of the Inn river bridge Wasserburg, Germany, 1987.
Strain measurements during construction showed a nonlinear stress distribution in the steel bottom
flange especially at the supports with high stress concentrations close to the webs. The deviation of the
measurement with regard to the calculated distribution was about 43% at the supports and diminished
with increasing distance from the pier axis. However, fatigue of the highly stressed studs was suspected.
In the frame of the research project an approximate method was developed in order to determine the
shear stress distribution in the composite interface. As a result the steel webs transfer 79% of the shear
directly into the concrete chord whereas only the remaining 21% are transferred by the steel bottom
flange. It was shown that the concrete chord dominates the shortening behaviour and thus influences the
steel bottom flange, which was opposite to the expected behaviour. This leads to this unexpected load
distribution of the stud shear connectors. However, the cross frames contribute to a load transfer in the
middle of the bottom flange. The cross frames and transverse stiffeners have the positive effect that they
transfer about 60% of the shear forces whereas only 40% are carried by the studs.
Besides that cracks were detected in the concrete cross beams at the interior supports during a load test
in 1989. It was suspected that a repeated traffic loading and changing of the crack width may lead to
fatigue and corrosion of the reinforcement. As a result, measurements were conducted in 1996 under
normal traffic in order to check the crack widths. Almost no increase could be observed, which was
related the fact that the loads were different (load test vs. normal traffic) and the crack widths increase
over-proportionally. The measured tensile strains were unexpected in an area were usually compression
forces prevail. However, due to the high shear and bending stiffness of the concrete bottom chord in
comparison with the steel webs, a partial bending moment in the concrete chord occurs, which is not
negligible any more.
Elbe river bridge Torgau. This is a haunched steel-box-girder bridge, which has a concrete top slab
without prestressing. The concrete bottom slab has a length of 21.25 m in the large span and 23.75 m in
the short span close to the pier with haunch which corresponds to 0.20 and 0.37-times of each span
74
Flanges
length. The slab thickness starts at about 50 cm and increases to 90 cm at the pier axis. The cross
girders at the bearings are made out of steel. An elevation is shown in Figure 4-7.
Elevation
53.0 m 106.0 m 65.0 m 3 x 46.0 m 36.0 m

509.0 m
Figure 4-7: Elevation of the Elbe river bridge Torgau, Germany, 1993.
For the Elbe river bridge Torgau, outcomes of the Inn river bridge Wasserburg were considered. Thus,
the offset of the neutral axis between steel and composite bottom flange has been taken into account
more precisely and the carding moments (bending moment due to the different locations of the neutral
axis of the steel section and the composite section) were resisted by a pair of forces, which is provided
by studs located at the adjacent webs of the cross frames.
First measurements on the bridge showed that the concrete bottom slab is fully effective. However, it
was surprising that the concrete top slab was not fully cracked as originally assumed so that this higher
stiffness added to the distribution of the internal forces. Measurements in 1995 showed that the concrete
bottom slab is still fully effective. It is observed that the hogging moment at the supports was not
reduced as expected but that the mean shear forces between steel and concrete bottom chord were
decisively smaller.
Mosel river bridge Bernkastel-Kues. This is a haunched steel-box-girder bridge with two boxes,
which has a concrete top slab prestressed in the longitudinal direction. The concrete bottom slab has a
total length of 28.6 m at the piers, which leads to a double-composite cross section of 0.42 and 0.18-
times of outer and inner span length. The slab thickness reaches to 50 cm at the pier axis. The cross
girders at the bearings are made out of steel. An elevation is shown in Figure 4-8.
Elevation
36.4 m 74.4 m 36.4 m

147.2 m
Figure 4-8: Elevation and longitudinal section of the Mosel river bridge Bernkastel-Kues,
Germany, 1995.
The aim of the accompanying research project was mainly to enhance the knowledge of the in plane
composite action to strongly haunched girders. The measurements at the bridge showed that the
haunches act positively with regard to shear and stud forces.
The outcomes of the research projects on the aforementioned bridges have greatly influenced the design
and construction of later bridges with double-composite action in Germany. A well-known example is
the Inn river bridge Neuötting with a main span of 154 m, which was completed in the year 2000, see
Figure 4-9 and [12], [50]. The haunched girder has a concrete top slab without prestressing. The
concrete bottom slab has a length of 50 m and 40 m above the left and the right river piers with a
variable thickness of 40 cm at the beginning of the concrete chord and up to 120 cm at the piers.
Table 4-1 gives an overview on the most prominent bridges with double-composite action which have
been built in Germany.
75
Longitudinal section Bottom concrete chord
95.0 m 154.0 m 95.0 m 68.0 m 58.0 m

470.0 m
Figure 4-9: Longitudinal section of the Inn river bridge Neuötting, Germany, 2000.
Table 4-1: Data of bridges with double-composite action in Germany.
Brücke Year of Total length Spans Deck area

completion [m] [m] [m2]
Inn river bridge

1987 375 83-104.5 5,156
Wasserburg
Elbe river bridge

1993 395 36-106 5,925
Torgau
Main river bridge

1994 374.4 83.2-208 5,354
Nantenbach
Mosel river bridge

1995 147.2 36.4-74.4 1,766
Bernkastel-Kues
Elbe river bridge

1997 330 36-108 6,943
Meißen
Inn river bridge

2000 470 58-154 13,865
Neuötting
Havel river bridge

2002 240 35-70 3,180
Plaue
4.4.4 Conclusions and design recommendations

In the following the state-of-the-art knowledge on double-composite action as a result of the various
research projects mainly in Germany is summarised:
• Conceptual design. The concrete bottom slab should be placed in the negative bending
moment region so that only compression forces occur. It should begin approximately at the zero
points of the moment distribution in order to reduce the forces that need to be transferred from
the steel plated elements to the concrete chord at the changing of the cross-sectional type.
• Global analysis. A double composite girder without longitudinal prestressing of the concrete
upper slab can be designed like a common composite girder considered upside-down. In case of
a longitudinal prestressing of the concrete upper slab, the influence of the concrete chords due
to creep and shrinkage has to be considered. In [43] the mutual influence of creep and shrinkage
in case of longitudinally prestressed composite upper slabs has been studied.
• Internal force distribution between steel box-girder and concrete bottom slab. Double-
composite structures cannot be calculated on the basis of Bernoulli’s hypothesis because partial
76
Flanges
interaction occurs in multi-layer beams in case of a flexible connection, which cannot be

neglected. As a result these girders show a nonlinear stress distribution over their depth, which
has to be taken into account in the calculations. In [40] these effects have been discussed and
enhanced to dynamic problems.
• Load transfer between steel girder and concrete bottom slab. It has to be taken into account
that due to the offset of the neutral axis carding moments occur, which should be resisted by
studs which are located at the webs of the cross frames.
• Shear stress distribution in plane composite action. For the connection between concrete
chord and steel plated elements usually stud shear connectors are used. Their placement is
characterised by two main areas. In the transition zone between steel bottom flange and
concrete chord, the studs are equally distributed over the whole bottom flange width and the
distribution of forces is determined according to axial stiffness of the steel and concrete
sections. In the other parts of the steel bottom flange, the studs are mainly provided at the
flange areas close to the web because a highly nonlinear stress distribution exists and the studs
which are located at the webs carry more load than those of the bottom flange. In [64] the shear
stress distribution in plane composite action has been studied. Design rules are given in Section
9 of EN 1994-2, which favours a concentration of the studs close to the webs.
• Stud shear connectors. In contrast to the common headed stud layout, the studs at the web,
which have a horizontal arrangement with a short distance to the concrete surface have a
reduced resistance, as well static as fatigue. In [8] these topics have been addressed and design
rules are given in Annex C of EN 1994-2 [36]. Because of the stress distribution the shear
connectors are concentrated to an area close to the webs. In the interior part of the flange, where
a restraint from the shear connectors is relied upon to prevent local buckling of the steel
element of a composite plate in compression, the centre-to-centre spacings of the connectors
should not exceed the limits given in paragraph 9.4(7) of EN 1994-2.
• Construction. For the dimensioning of the steel bottom plate the concreting of the chord is
decisive. Thus, the concrete is often applied in several layers to generate already a partial
composite action for the successive concreting layers. A general question arises whether the
steel plate is a very expensive formwork. Depending on the labour or material cost, one will
choose rather a thick plate with as few longitudinal stiffeners as possible or a slender plate,
which is longitudinally stiffened to resist the weight of the fresh concrete. If the bridge is
erected by launching and the concreting is done before launching the bottom plate can be
propped and longitudinal stiffeners can be avoided also for thin plates.
77
Webs
5 Webs
5.1 Introduction
In Section 4 the flanges as part of the cross section were discussed separately, because they represent a
substantial element to reach the required bending moment resistance. Webs are equally important, if it
concerns to shear force resistance. Besides that, they have the task to interconnect the flanges. The webs
thickness is mainly chosen on basis of the required shear force resistance. For steel plated structures
large heights and plate slenderness occur so that the stability behaviour of the web must be usually
considered.
For the stiffening of the web both longitudinal and transverse stiffener can be used. A transverse
stiffener has mainly influence on the shear force resistance of the web. This is however only the case if
the distance between the transverse stiffeners is small, otherwise the influence is low and it does not
justify the costs of the transverse stiffener. Different design aspects of transverse stiffeners are
discussed in Section 5.2.
Longitudinal stiffeners increase not only the bending moment resistance but also the shear force
resistance of the web. It is however interesting, from which web height on a stiffening should be used at
all. The economy of longitudinal stiffeners is studied in Section 5.3.
In Section 5.4 the possibilities of the cross sectional layout of longitudinal stiffeners are presented
briefly and their practical application in different countries is described. Aside the typical arrangements
there exist also specific layouts for longitudinal stiffener, e.g. discontinuous longitudinal stiffeners and
exterior longitudinal stiffeners where the transverse stiffener is located on the opposite (usually inner)
side of the web plate in order to avoid the complicated intersecting detail between longitudinal and
transverse stiffener.
5.2 Transverse stiffeners

Transverse stiffeners are usually placed at the locations of cross bracings or diaphragms. This gives the
stiffeners a double purpose of stiffening the web and to serve as brackets for cross bracings. The cross
bracings are needed to prevent lateral torsional buckling during erection and in the pier regions also
during service. The effect of the transverse stiffeners on the resistance of the girders is limited to an
increase of the shear buckling resistance if the web does not have longitudinal stiffeners. The increase
in shear resistance makes it possible to reduce the web thickness, which saves some cost. This is
however counteracted by an increase of the cost of stiffeners and comparisons show that there is no net
saving. The first conclusion is that it does not pay to add stiffeners between the cross bracings. In old
codes there were rules giving maximum spacing between vertical stiffeners. However, there is no
rational ground for such rules and there are no such restrictions in EN 1993-1-5 except that vertical
stiffeners are needed at the supports. Therefore the second conclusion is to take away redundant vertical
stiffeners completely and use small brackets for fixing the cross bracings as shown in Figure 6-13. This
may be suitable in the sagging region of small and medium span bridges. For large spans it may be
advisable to add a horizontal beam at the top as well.
Vertical stiffeners are either flat stiffeners or T-shaped stiffeners. When they are included in a bracing
frame, T-shaped stiffeners may be provided and the cross-girder is welded on the flange of the vertical
stiffener. If a vertical stiffener is added between two bracing frames, it is usually a single flat plate.
79
For the T-shaped stiffeners, the web and the flange of the stiffeners are welded on the main girders
upper flange. In span, the flange of the vertical T-shaped stiffeners has a V-shaped cut-out for fatigue
reasons, and so is not welded on the main girders lower flange, see Figure 5-1 . At supports, the vertical
T-shaped stiffeners are very often duplicated outside the main girder and the T section is entirely
welded on the lower flange.
Figure 5-1 T-shaped vertical stiffeners.
Vertical stiffeners are mostly present at each bracing frame location. These bracing frames are usually
spaced by 6.0 to 10.0 m for girder bridges, and by 4.0 to 5.5 m for box-girder bridges. Around the
intermediate supports, some vertical stiffeners may be added in order to limit the aspect ratio a/b of the
first web panel to better resist shear buckling. As mentioned above an alternative is to increase the web
thickness.
It is useful to note that EN 1993-1-5 requires that, in case of changes in the plate thickness of the web,
the transverse welded splice should be close enough to a transverse stiffener (if this requirement is not
fulfilled, effects of eccentricity need to be taken into account). This may have consequences on the
positioning of the stiffeners, in relation with the optimisation of the plate thicknesses.
In addition to plate buckling rules, EN 1993-1-5 gives also recommendations for the verification of the
stiffeners themselves. Transverse stiffeners should be able to carry deviation forces from the adjacent
compressed panels and be designed for both appropriate strength and stiffness. For the verification, the
stiffener should be considered as a simply supported beam with initial sinusoidal imperfection,
according to the static scheme given in Figure 5-2, assuming that the adjacent stiffeners are rigid and
straight.
Figure 5-2: Static scheme for the verification of stiffeners.
In principle, based on a second order elastic method analysis, both the following criteria should be
satisfied at the ultimate limit state:
- the maximum stress in the stiffener should not exceed fy/γM1
- the additional deflection should not exceed b/300
80
Webs
Any relevant load acting on the stiffener should be included, such as for example axial force in the
stiffener due to directly applied external forces or horizontal transverse loading of the stiffener due to
in-plane curvature of the girder. In the most general case, a transverse stiffener may be loaded with:
- Transverse deviation forces originated from longitudinal compressive force or bending moment of
the adjacent panel;
- External transverse loading in the horizontal direction;
- Axial force in the stiffener coming from vertical transverse loading on the girder;
- Axial force in the stiffener coming that may develop from buckling in shear, see 9.3.3(3) of EN
1993-1-5.
For some of these situations (transverse deviation forces with or without direct axial forces in the
stiffener), EN 1993-1-5 proposes equivalent rules based on inertia criteria or on equivalent linear
analysis EN 1993-1-5, 9.2.1 (1) to (7) [31].
Detailed information on these requirements can be found in reference [45], as well as a comprehensive
worked example of a plate girder, including the verification of the transverse stiffeners.
Moreover, EN 1993-1-5 requires that the torsional buckling of stiffeners is prevented. To this purpose,
two criteria are given for cases where the warping stiffness is considered or not. Chapter 3 of the
COMBRI background document [68] dealing with the "Design of bracing frames for a twin-girder
bridge" shows an example of the use of these criteria.
Besides acting as a bearing stiffener resisting the reaction force at the support, a rigid end post should
additionally be designed as a short beam, with a length being equal to the web depth, resisting the
longitudinal membrane stresses in the plane of the web. A rigid end post should comprise two double-
sided transverse stiffeners that form the flanges of the short beam, with a minimal cross-sectional area
and a maximal distance between both stiffeners. Alternatively, a rigid end post may be realised by
inserting a hot-rolled section. If the end post cannot be considered as rigid, a reduced shear resistance of
the end panel must be calculated according to Section 5 of EN 1993-1-5, see Figure 5-3
It has been showed in the COMBRI research project that, for webs with closed longitudinal stiffeners
(i.e. trapezoidal stiffeners) welded to the transverse stiffeners, the latter could be considered as rigid end
posts, even if not fulfilling the above conditions. This possibility is however not included in EN 1993-1-
5.
a) Rigid end post. b) Non rigid end post.

Figure 5-3 Rigid and non rigid end post.
Intermediate stiffeners that act as rigid supports at the boundary of inner web panels shall be checked
for strength and stiffness, following the procedure described previously. Further minimum stiffness
requirements are also given by EN 1993-1-5 for the intermediate transverse stiffeners to be considered
as rigid. If the relevant requirements are not met, transverse stiffeners are considered flexible and their
actual stiffness may be considered in the calculation of the shear buckling coefficient kτ. However, no
information is given in EN 1993-1-5. Thus appropriate software should be used, such as EBPlate,
developed in the frame of the present project.
81
Although EN 1993-1-5 is not clear on this point the rules presumes that the transverse stiffeners has a
function that increases the resistance. Concerning longitudinal compression this is the case for a web
with longitudinal stiffeners. For an unstiffened web the wave length of the buckles caused by direct
stress is usually so small that no increase in the resistance is taken into account. In such a case the above
criteria are not relevant for a web without longitudinal stiffeners. However, if the stiffener carries direct
or transverse load or if it causes an increase in the shear resistance the effect of the longitudinal
compression should be taken into account. These rules are in most cases over-conservative and they are
also a good reason for not using transverse stiffeners unless needed for a specific purpose of which
increasing the shear resistance is not a good one.
5.3 Transition between unstiffened and longitudinally stiffened

webs
In the COMBRI research project a 2.8 m deep I-girder bridge was studied and designs with and
without longitudinal stiffeners were compared [15]. The designs for one internal span are
shown in Figure 5-4.
Cross section Top flange Bottom flange Web
Type 1 51x900 58x900 18x2491
Type 2 26x700 40x700 16x2534
Type 3 20x400 23x700 13x2557
Type 4 53x900 56x900 15x2491
Type 5 30x700 36x700 14x2534
Figure 5-4: View of girder without (top) and with longitudinal stiffener (bottom), not to scale.
List of material with all plates in S460.
The two designs have the same cross section in the span and if the cost difference is attributed to the
parts that are different, it corresponds to a cost increase of 18 EUR/m2 of bridge deck for the alternative
with longitudinal stiffeners. It is less than 2% of the total cost of the bridge and about 4% of the cost for
the steel girders. Even if the difference is small it is a clear result in favour of the design without
longitudinal stiffeners. This is also consistent with the current Swedish practice not to use longitudinal
82
Webs
stiffeners for girders up to 3.2 m web depth. The figure is not exactly the break even, which may vary
from case to case. The actual reason for this limit is that this is the widest plate manufactured in Sweden
and Finland and in order to avoid a longitudinal splice this depth is kept even if a deeper girder would
require less material.
The above conclusion is related to the use of effective widths for the design of Class 4 sections. The
effective width method has been allowed in Sweden since 1988, however with stricter limitations for
web breathing than those in EN 1993-2. EN 1993-1-5 gives two methods for the design of Class 4
sections. One is the effective width method used in the above study and one based on stress limitations
without redistribution of stresses given in Section 10. If the latter is applied the alternative with
longitudinal stiffeners would be more or less all right (not checked in detail) but the alternative without
would be quite impossible. For the pier section the web thickness has to be roughly doubled (from 18 to
36 mm) in order to avoid reduction of the compression stresses because of web buckling. The flange
could then be made smaller but the increase in weight would still be substantial and the alternative
would be uneconomical.
In order to find a limiting depth at which unstiffened webs are no longer economical, the COMBRI
research project was extended, and another (imaginary) bridge with deeper spans has been designed.
The designs for one internal span are shown in Figure 5-5.
Cross section Top flange Bottom flange Web
Type 1 77x1150 86x1150 24x3837
Type 2 55x1000 57x1150 22x3888
Type 3 43x600 55x1150 17x3902
Type 4 80x1150 80x1150 21x3840
Type 5 55x1000 54x1150 19x3891
Figure 5-5: View of girder without (top) and with longitudinal stiffener (bottom), not to scale.
List of material with flanges in S460 and webs in S355.
This comparison is made for a 4 m deep I-girder, and the comparative analysis is calculated with the
same assumptions as for the 2.6 m deep I-girder. Two sections have the same cross section in the span
and the cost difference is attributed to the sections that are different. The cost comparison for the two
different solutions results in a negligible difference of about 1 EUR/m2 in favour of the unstiffened
alternative, which indicates that the limiting depth at which unstiffened webs are no longer economical
83
would be approximately 4 m. For both bridges, trapezoidal longitudinal stiffeners have been used in the
calculations, and the costs have been calculated by the same bridge contractor.
5.4 Type of longitudinal stiffener and continuity
5.4.1 General
Two main types of longitudinal stiffener may be used:
• Open stiffener, usually single flat stiffener
• Closed stiffener, usually trapezoidal one
In case of box-girder bridges, they are located inside the box section. In case of I girder bridges, they
are in general located between the girders. In a few cases longitudinal closed stiffeners have been put
outside the girder to solve the problem of the intersection with the vertical stiffeners.
5.4.2 Single flat longitudinal stiffener

Flat stiffeners are in general continuous through the length of the bridge and pass through the vertical
ones. In this case they are fully taken into account for global and section analysis. Their location in the
web depth results from a compromise between different verifications (shear, bending, patch loading).
As an exception in France, in order to avoid the problem of cutting the vertical stiffeners, the
longitudinal single flat are discontinuous (see Figure 5-6 for an illustration of this typical French
design). In this case, EN 1993-1-5, 9.2.2 (2) indicates the way to take them into account:
• Neglected in the global analysis and in the calculation of stresses,
• Considered in the calculation of the effectivep widths of web sub-panels and in the calculation
of the elastic critical stresses.
In order to improve the fatigue behaviour, these discontinuous longitudinal stiffeners have tapered ends,
see Figure 5-7 for the fatigue classification.
© Sétra © Sétra
Figure 5-6: Twin-girder bridge in Triel-sur-Seine, France, 2003.
84
Webs
nono radiustransition
radius transition
Δσc = 56 MPa
transitionwith
transition withchamfer
chamfer Δσc = 71 MPa
(angle not higher than 45°)
smooth
smooth radiustransition
radius transition Δσc = 80 MPa
r > r150 mm
> 150 mm (see additional requirements in EN 1993-1-9 [32],
Table 8-4, Detail 3)
Figure 5-7: Fatigue detail categories for a single flat stiffener.
5.4.3 Closed shape longitudinal stiffener

They should be continuous and pass through openings made in the transverse stiffeners. EN 1993-1-5,
9.2.4 (4) gives detailing about the dimensions of the openings in the transverse stiffeners (see Figure
5-8). After having been cut out, the transverse stiffener has a reduced resistance to shear which is taking
into account by EN 1993-1-5, 9.2.4(5) by adding a criterion for the shear resistance of the gross web
adjacent to the cut out. This detailing with openings in the transverse stiffeners is also better for fatigue
behaviour.
≤ 0.6hs
hs
Figure 5-8: Cut outs in transverse stiffener
5.4.4 Discontinuous longitudinal stiffeners

Current French regulations concerning plate buckling, which are based on elastic buckling theory, can
lead to the use of one or two longitudinal stiffeners (depending on the web depth) welded to the main
girders.
For I-girder bridges, these longitudinal stiffeners are usually single flat plates. Their location in the web
depth results from a compromise between different verifications (shear, bending, patch loading). In
order to improve the fatigue behaviour, the longitudinal stiffeners are discontinuous (no welding on the
vertical ones) with tapered ends (see Figure 5-6).
As they are not continuous, their steel section is not taken into account in the global analysis.
85
5.4.5 Exterior longitudinal stiffeners

In general, longitudinal stiffeners are oriented towards
the inside of the bridge structure. In this place, usually
other stiffening elements such as vertical stiffeners
and/or cross-frames are arranged as well. As a result,
constructional details between the stiffener running
through in the longitudinal direction and the vertical
stiffeners and/or cross-frames need to be solved. An
example of such a situation in a typical box-girder is
shown in Figure 5-9.
As an alternative solution the longitudinal stiffeners can
be applied on the opposite side than the vertical stiffeners
and/or cross-frames which would commonly correspond
to the exterior surface of the web plate. Thus, not only a
costly detailing and fabrication of the intersecting joint is © KE
avoided but also the number of notches is reduced with Figure 5-9: Constructional detail of the
regard to fatigue. intersection “longitudinal stiffener and
A well-cited example in literature of a bridge where the cross-frame”.
longitudinal stiffeners have been applied on the exterior
of the web is the river bridge Nordsteg across the Danube in Vienna, Austria [57], [67], cf. Figure 5-10.
The bridge has been completed in 1996 and firstly it was used as a road bridge to allow for
comprehensive maintenance works of a neighbouring bridge before it finally became a footbridge. For
the longitudinal stiffeners a triangular cross-sectional shape, see Figure 5-11 has been chosen and in
total four longitudinal stiffeners have been applied over the web height. The orientation on the outside
also facilitated the installation of two large pipes inside the box-girder.
© KE (Photographer:
Florian Stockert) © MCE Stahl- und Maschinenbau GmbH & Co
Figure 5-10: River bridge Nordsteg Figure 5-11: Cross-sectional view of the Nordsteg in
in Vienna, Austria, 1996. Vienna, Austria, 1996.
In Germany, a railway bridge with longitudinal stiffeners on the outside of the web was erected near
Riesa in 2005, cf. Figure 5-12. The bridge consists of a tied-arch bridge with approaching spans which
is erected by the incremental launching technique. The web of each main girder has two exterior
longitudinal stiffeners.
86
Webs
© Donges Steeltec GmbH
Figure 5-12: Launching of a railway bridge near Riesa, Germany, 2005.
In the following, the characteristics and implications of bridges with longitudinal stiffeners on the
outside of the web are summarised:
• Fabrication. As the complex constructional detail at the intersecting joint between longitudinal
stiffener and vertical stiffener and/or cross-frame disappears, it saves cost and time with regard
to the cutting, trimming and welding of the steel plates.
• Corrosion protection. In order to avoid the accumulation of dirt and standing water, a
sufficient inclination of stiffener flanges is recommended.
• Fatigue. Due to the disappearance of the intersecting joint between longitudinal stiffener and
vertical stiffener and/or cross-frame, the number of notches is reduced which is favourably with
regard to fatigue behaviour.
• Appearance. The application of exterior longitudinal stiffeners can be used to influence the
architectural quality of the bridge.
87
Cross bracings and diaphragms
6 Cross bracings and diaphragms
6.1 Introduction
This chapter deals only with intermediate cross bracings. Cross bracings or diaphragms serve the same
purpose and are used in bridge girders to provide the following functions:
• Prevent lateral torsional buckling during erection;
• Distribute loads between multiple girders (if more than two);
• Transfer lateral loads (wind) on the girders to the deck;
• Prevent lateral torsional buckling of a compressed bottom flange during service;
• Prevent cross sectional distortion of box-girders.
There is no requirement of cross bracings in the Eurocodes and accordingly they can be omitted if the
functional requirements listed above can fulfilled with other means. This may be the case for short span
bridges with rolled girders
6.2 Cross bracings in I-girders
6.2.1 General
Cross bracings in I-girder bridges can be designed as a frame like the one shown in Figure 6-1 or like a
lattice shown in Figure 6-2. The former relies on the bending stiffness of the transverse beam and the
stiffeners. The joint between those has to be able to transfer moments and may be welded or bolted. The
latter relies on the axial stiffness of the bars and if it is a complete lattice it is usually so stiff that a
separate check of its stiffness is not needed. The adequacy of the frame type cross brace for preventing
lateral torsional buckling of the top flange during erection and for the bottom flange during service has
to be checked. Simplified rules are given 6.3.4.2 of EN 1993-2 for the check of the stiffness as well as
the strength.
Figure 6-1: Cross bracing in the shape of a frame. Figure 6-2: Cross bracing as a full lattice.
89
The frame type cross bracing can be used to carry the deck slab and also be extended under the
cantilever part of the slab (see Figure 6-3).
For railway bridges, the cross bracing is strengthened due to heavier loads and dynamic effects induced
by high speed. A lower concrete slab (or a horizontal steel bracing) is added between the two lower
flanges to increase the torsional rigidity of the bridge. Figure 6-4 shows the horizontal studs, which
connect the lower concrete slab to the lower parts of the webs.
© Sétra © Sétra
Figure 6-3: Twin-girder bridge near Avignon, Figure 6-4: Railway twin-girder bridge (TGV
France, 2008. Est, Canal de l'Ourcq) with diaphragms,
France, 2006.
The temporary wind bracing used during construction before the concrete slab is placed, could be made
of rigid members assembled using bolts, or cables (see Figure 6-5 and Figure 6-6). This wind bracing is
designed with in-plane cross beams between the main girders. The wind bracing is located below the
top of the girders in order not to be in the way for the formwork for concreting the slab.
© Sétra © Sétra
Figure 6-5: Twin-girder bridge in Sens, France.
© Sétra
Figure 6-6: Railway twin-girder bridge (TGV Est, Pont à Mousson), France.
90
6.2.2 Standard design for the calculation example “twin-girder bridge”
6.2.2.1 Stiffness of the bracing frames

Typical bracing frames are modelled as illustrated in Figure 6-7, with the following hypotheses:
• the concrete slab transverse flexibility and its extensibility are neglected
• the web part acting together with the vertical frame post is 15ε tw
• the vertical frame post is assumed to make an hinge with the concrete slab
Lateral displacements under two load cases are studied (see Figure 6-8). The stiffness Cd of the bracing
frame is then given by:
⎛1 1 ⎞
C d = min⎜⎜ , ⎟⎟
⎝ δ1 δ 2 ⎠
Figure 6-7: Modelled transverse frame.
be
h m2
(m2 )
hm
(e)
h m1
(m1 )
F=1 F=1 F=1 F=1
δ1 δ1 δ2 δ2
a) Same direction forces (load case 1). b) Opposite direction forces (load case 2).
Figure 6-8: Load cases for the rigidity Cd calculation.
91
6.2.2.2 Spacing of the bracing frames

The initial proposed distribution of bracing frames in Part I of the COMBRI Design Manual [16] for the
twin-girder bridge was as follows :
• rigid transverse bracing on abutments and on internal supports which are considered to be fixed
lateral supports for the main girders,
• in-span bracing frames every 8.33 m in side spans (C0-P1 and P2-C3) and every 7.5 m in
central span (P1-P2), with a rigidity Cd = 24.2 MN/m (see Figure 6-9).
To justify lateral torsional buckling around internal supports (P1 and P2), some additional provisions
are proposed :
• additional bracing frames located at 3.5 m from P1 and P2 in the end spans,
• additional bracing frames located at 3.0 m from P1 and P2 in the central span,
• the rigidity of eight bracing frames (two on each side of P1, and two on each side of P2) is
increased.
The corresponding design is illustrated in Figure 6-10. The obtained stiffness reaches Cd = 46.6 MN/m
instead of the previous value.
263
220 17
19
300 20 400 30
753 11.5
600 12
300 340
15 30
Figure 6-9: Non-strengthened transverse Figure 6-10: Strengthened transverse cross-

cross-bracing in span. Dimensions in [mm]. bracing in span. Dimensions in [mm].
6.2.2.3 Verification against lateral torsional buckling

Table 6-1 illustrates the lateral displacements corresponding to the first three buckling modes of the
numerical model. The factor αcr,op is the factor by which the ULS applied load should be multiplied to
get the critical load for a given buckling mode.
αult,k = 1.168 and αcr,op = 15.676, so the reduced slenderness is equal to:
α ult ,k 1.168
λop = = = 0.273 ≥ 0.2
α cr ,op 15.676
1
( )
Φ = . ⎡⎢1 + α λ op − 0.2 + λ op ⎤⎥ = 0.565 and χ =
1
2
Then ≤ 1.0 is equal to 0.944.
2 ⎣ ⎦ 2
Φ + Φ 2 − λ op
The LTB criterion is then verified:
92
α ult ,k
χ op = 1.002 > 1.0
γ M1
Table 6-1: Transverse displacement of the first three elastic critical buckling modes.
Mode αcr,op Description of the observed transverse displacement
C0 P1 P2 C3
1st 15.676
Anti-symmetric waves around the support P2
C0 P1 P2 C3
2nd 17.716
Anti-symmetric waves around the support P1
C0 P1 P2 C3
3rd 27.111
Quasi-symmetric waves around the support P2, and a small wave around P1
6.2.3 Improved design for the calculation example “twin-girder bridge”
6.2.3.1 Cross-bracing characteristics

The twin-girder bridge designed in Part I of the COMBRI Design Manual [16] was provided with a
frame type cross bracing. An improved design of cross bracing for LTB is presented in the following
based on [68]. The spacing between bracing frames is the same as the spacing described in Section
6.2.2. The difference is the nature of the cross-bracing, made up of several bolted rolled members
instead of welded steel sections.
Figure 6-11: Alternative transverse cross-bracing in span.
The eight strengthened bracing frames (two on each side of P1, and two on each side of P2) are made
up of channels UPE 100 as diagonals (two channels UPE 100 back to back per diagonal) and of a HEA
180 as horizontal, see Figure 6-7. The frame posts are flat plates (dimensions 300 mm x 30 mm). The
stiffness of these bracing frames is Cd = 48 MN/m. This stiffness includes the flexibility introduced by
the eccentricity between the bottom flange and the horizontal HEA. The other in-span bracing frames
93
are assumed to be made up of channels UPE 80 as diagonals (two channels UPE 80 back to back per
diagonal) and of a HEA 160 as horizontal. The frame posts are flat plates (dimensions 300 mm x 30
mm). They are not analyzed here but later in 6.2.3.5.
6.2.3.2 Verification against lateral torsional buckling

The buckling modes have similar shape to the buckling modes presented in 2.6 of Part I of the Design
Manual, with close factors αcr,op equal to 16.0 for the first mode, 18.1 for the second mode, and 27.8 for
the third mode. The LTB criterion is then verified:
α ult ,k
χ op = 1.005 > 1.0
γ M1
6.2.3.3 Verification against buckling of the members

In the first in-span bracing frame closest to P2, at the abscissa x=107 m, there is a risk of buckling of
the diagonals and of the horizontal HEA.
The following horizontal load cases are applied to the level of the bottom flange:
• a lateral wind
• a force equal to 1/100 of the compressive force at ULS in the lower flange at the position of the
bracing.
The total gives a 0.24 MN transverse load. The corresponding normal force in the bracing frame are
presented in Figure 6-8, for the cases buckling towards each other and buckling in the same direction.
The normal force due to the wind only applies on one side, but as it is small compared to the normal
force due to NULS/100, it has been applied on both sides.
Figure 6-12: Normal force in alternative bracing frames, for the study of the members buckling.
For the buckling verification, the members are considered as hinged at both ends (this assumption is of
course very unfavourable).
Buckling of the diagonals

For each diagonal (2 UPE 100 back to back), the length of the member is about 3.6 m, so the critical
force is equal to 570 kN and the reduced slenderness is equal to 1.23. The diagonals are channels, so the
reduction curve is the curve c, with α = 0.49. The corresponding reduction factor χ is equal to 0.42,
N Ed
and the ratio is equal to 0.97.
N b , Rd
This value is inferior to 1.0, so there is no risk of buckling of the diagonal members.
Buckling of the horizontal member

The length of the horizontal member (HEA 180) between the frame posts is about 6.40 m, so the critical
force is equal to 0.47 MN and the reduced slenderness is equal to 1.83. The horizontal is a rolled
94
sections, so the reduction curve is the curve c for buckling in the weak direction, with α = 0.49. The
N Ed
corresponding reduction factor χ is equal to 0.23, and the ratio is equal to 0.85 < 1.0.
N b , Rd
So there is no risk of buckling of the horizontal member.
Comparison with the previous design

The two designs presented in Part 1 and here comprise the same number of bracing frames. Each design
consists of two different bracing frames, a non-strengthened one, and a strengthened one.
In the first design, the members are welded, whereas they are bolted in the second design, so the
fabrication of the second one is easier and cheaper.
Concerning the steel quantity of the different bracing frames designs, shows that the lattice type design
saves 52 % of steel.
Table 6-2: Steel quantities for cross bracings.
Weight of Number of Total weight Saving

one bracing bracing frames of the
frame on the whole bracing
[kg] bridge frames
[tons]
Frame type, non-strengthened 1,180 13

28.9
Figure 6-7
strengthened 1,690 8
Lattice type, non-strengthened 640 13

13.9 52 %
Figure 6-11
strengthened 700 8
6.2.3.4 Choice of the reduction curve for lateral torsional buckling

For the previous bracing frames designs, the curve d corresponding to α = 0.76 has been used,
according to EN 1993-1-1 [30]. Recent studies show that this curve d is very safe for LTB, and that the
curve b would be more appropriate. The LTB verification is thus more unfavourable with EN 1993-1-1
than with the previous national rules of many European countries.
Table 6-3: Steel saving on cross bracings from using reduction curve “b” instead of curve “d”
for lateral torsional buckling.
Total weight of the

Saving
bracing frames [tons]
Frame type, reduction curve d 28.9

Figure 6-7
reduction curve b 20.1 30%
Lattice type, reduction curve d 13.9

Figure 6-11
reduction curve b 10.8 22%
95
With the use of the reduction curve b, a spacing between bracing frames similar to the one described in
original design with bracing frames close to the non-strengthened would be sufficient for verification of
LTB, with an undoubted steel saving, see Table 6-3.
6.2.3.5 Cross-bracings in the sagging moment region

So far it has been assumed that the girders have a wind bracing during erection. It can be seen from for
instance Figure 6-5 and Figure 6-6 that such bracings are used in France and from Figure 2-61 that they
are not used in Sweden. The differences in practice may be caused by different rules for load
combination or differences in wind loads considered for temporary stages. If there is a wind bracing the
stability of the top flange during casting the slab can be checked as lateral buckling between the cross
braces. If there is no wind bracing the stability of the top flange has to be assured by the cross bracings.
If the cross bracing is of the lattice type discussed above the stiffness of the bracing is quite small and
relies mainly on the bending stiffness of the horizontal beam. Figure 6-9 shows a cross bracing without
vertical stiffeners, which is possible when the function is to prevent lateral buckling of the top flange
during erection. If it is intended to prevent buckling of the bottom flange during service the eccentricity
of the horizontal to the bottom flange has to be considered and vertical stiffeners will be required as
shown before.
In addition to being cost effective this cross bracing also minimizes the restraint for rotation of the top
flange. It is sometimes a concern that bending of the bridge deck induces tension in the shear studs if
the rotation of the top flange is restrained, for instance by vertical stiffeners. In 6.6.1.1(13) of EN 1994-
2 it is stated that this should be considered but no quantitative rules are given. The use of the cross
bracing in Figure 6-13 would be one way of satisfying this application rule.
Figure 6-13: Cross bracing without vertical stiffeners.
A verification of the top flange at the mid span of the I-girder bridge in Part I of the COMBRI Design
Manual [16] will be shown assuming that the horizontal beam is a HEA 180 in S355. This is one size
larger than in 6.5.3 but it will turn out to be needed. The spring stiffness is calculated assuming that the
axial stiffness of the bars is high and that the flexibility comes only from bending of the horizontal
beam. With a fictitious unit force in each top flange directed outwards we get a deflection upwards in
the middle of the beam of 1.87 m/MN and the corresponding outward displacement of the top flange
becomes 0,976 m/MN. The spring stiffness thus becomes Cd =1/0.976 = 1.02 MN/m. The distance
between the braces is 7.5 m and the distributed spring stiffness becomes c = 1.02/7.5 = 0.136 MN/m2.
For the calculation of the critical force it is assumed very conservatively that the top flange 35x 800 is
infinitely long with constant axial force.
N cr = 2 cEI = 2 0.136 ⋅ 314 = 13.1 MN
The area should be taken as the flange area plus one third of the compression zone which gives A=
0,0358 m2 and σcr = 366 MPa. The slenderness parameter and the reduction factor with curve d
becomes
345
λ LT = = 0.97
366
χ LT = 0.483
96
The resistance expressed as stress becomes

345
σ Rd = 0,483 = 151 MPa
1.1
The actual design stress during casting is 1.35x94=127 MPa and the stability is verified.
In addition the strength of the cross bracing should be verified according to 6.3.4.2(5) of EN 1993-2.
The axial force in the flange is taken as
N Ed = 127 ⋅ 0.0358 = 4.54 MN
First we have to check if second order effects have to be considered with
EI
lk = π = 15.4 m >1.2 l = 9.0 m
N cr
It means that second order effects has to be considered and the applicable formula for the lateral force is
l N Ed 1 7.5 4.54 1
FEd = = = 0.0423 MN
l k 80 N Ed 15.4 80 4,54
1− 1−
N cr 13.1
This force applied to each flange and outwards gives a bending moment in the beam 88 kNm to be
compared with the bending resistance of 105 kNm. The force FEd also gives a compression force in the
web of 25 kN, which should be resisted as opposite patch on the web. This is obviously no problem and
the calculation is omitted.
6.3 Diaphragms in box-girders

A box-girder needs diaphragms or cross braces in order to avoid cross sectional distortion caused by
eccentric traffic load. The most cost effective solution is usually cross braces, for example as shown in
Figure 4-5. The forces on the cross braces are fairly high compared to those in I-girder bridges. The
solution shown in Figure 4-5 without transverse web stiffeners requires that the transverse forces in the
web can be carried by the web alone. This has to be checked with the rules for opposite patch loading in
EN 1993-1-5.
There is a rule in 6.2.7.1(3) of EN 1993-2 [35] stating that an increase of the bending stresses in a box-
girder due cross sectional distortion of 10% may be neglected. This rule makes it possible to determine
a maximum distance between diaphragms. It is however hard to give general rules because there are
many parameters that influence the magnitude of the additional stresses due to cross sectional
distortion.
97
Launching
7 Launching of steel- and composite bridges
7.1 Introduction
The incremental launching is the most common method for building the structural steel part of a bridge.
The steel elements (coming from the workshop) are assembled on a devoted area behind the bridge
abutment. The steel part of the bridge is then pushed (from one side or from both ones) step by step to
reach its final position. In France the longest span (171 m) has been launched in 2003 for the Millau
viaduct. For a twin-girder bridge, the record is held in France by the Triel-sur-Seine bridge with 124 m
in 2003. Another noticeable example is the composite box-girder of the Verrieres viaduct with a
maximum launched span of 144 m in 2000. Different pictures have already been included in this
Manual, see Figure 2-25 or Figure 2-27 for instance.
It should be notice that the launching technique can be used for curved bridge with a constant radius, or
for girders with a variable height, see Figure 7-1. For each step of the launching process, the track of the
launched part should coincide with the final transverse position of the bridge.
© Sétra © Sétra
a) Launching of a curved bridge b) Launching of girder with a variable height near Pont-à-Mousson,
near Cannes, France. France.
Figure 7-1: Particular launching case.
From an economical point of view

The main advantages of this technique are as follows and result in a reduction of the construction time
and costs for the bridge:
• Cross the valley widely above the natural ground
• Reduction of risk for the workers who work directly on the ground in the launching area (and
not above the valley)
99
• Possibility of making the welds in a provisional shelter with all facilities for assembling the
steel segments arriving from the workshop
• Launching above railway tracks or roads without interrupting the traffic, and resulting in cost
savings
The main inconvenient is that an reserved area is needed behind the bridge abutment in order to
assemble the steel structure before launching. The scope of the launching process in terms of span
length and of girder weight is also closely linked to the capacities of the launching devices, see below.
From the designer point of view

The resistance verification should be performed for each launching step and then for each bridge cross-
section. For a given launching step, the in-span cross-sections are not usually designed by this transient
situation. Even if the structural steel part of the section resists alone (without steel-concrete composite
behaviour), the applied loads are twice (even more) smaller than in the final situation of the bridge
under traffic loads.
These verifications also include the buckling which could occur in the web panel located provisionally
on an internal support. This panel is submitted to transverse patch load in combination with the bending
moment and the shear force coming from the cantilever part of the bridge (which has already crossed
this internal support). The web panel is limited by two transverse vertical stiffeners and its buckling can
be justified by eventually adding longitudinal stiffeners. See Paragraph 7.2 below.
The global buckling of the launched steel girder should also be justified. This mainly concerns the
lateral torsional buckling which can occur in the cantilever steel girder (partially launched) with the
whole bridge section, or in the upper compressive flange of a box-girder (or of an I-girder) which is
located in a completely launched span. This lateral torsional buckling can be prevented by using a well-
designed provisional cross-bracing. See Paragraph 7.3 below.
7.2 Local behaviour: introduction of the transverse load
7.2.1 General
Different kinds of devices could be used for bridge launching. The two main ones are launching shoes
and sliding skates. The sliding skates are preferable if the weight is high. If the support reaction does
not exceed around 300 tons, the rolling shoes become more efficient because their use increases the
launching speed (up to 1.5 m per minute).
With the following devices, the classical web buckling verification (for instance, the Eurocodes)
considers that the transverse load is applied in the web plane without transversal eccentricity and that
the load intensity is uniform over the whole length of the device.
7.2.2 Launching shoes
7.2.2.1 General
The main problem of this device is that their length increases very quickly with the reactions on the
support. The reactions on support can be particularly high when the spans are long or when the bridge is
launched with part of its concrete slab. After launching, the transfer from the launching shoe to the
definitive support becomes also more difficult and more expensive.
The two techniques explained below insure that the vertical reactions are the same for each wheel of the
shoe. In the first case this is obtained by the balancing system and in the second case, by the use of the
cable.
100
Launching
7.2.2.2 Wheels on balancing device

These shoes are composed with elementary balancing devices of 2 steel wheels. The balancing device is
installed on a roller bearing that allows the launching of bridges with variable depth or large pre-
camber. The articulations of bearings keep the contact between the bridge girder and the wheels of the
launching shoe during launching phases. Different configurations could be drawn with 2, 3, 4 or 6
wheels in a launching shoe. The number of wheels is determined by the load calculation on supports
during launching. Each wheel corresponds to a maximum load of 30 to 55 tons, its thickness varies
from 60 to 180 mm and its diameter varies from 350 to 800 mm following the constructor.
© Sétra
Figure 7-2: Launching shoe with 6 wheels (maximum loads = 330 tons).
© Sétra / Baudin Chateauneuf Enterprise
Figure 7-3: Launching shoe with 4 wheels and 2 lateral wheels for longitudinal guidance (total
length = 1.25 m).
7.2.2.3 Wheels and cable

The steel wheels are set down on a cable that is anchored in the frame of the launching shoe. Each
wheel (4 or 5 per shoe) corresponds to a maximum load of around 40 tons.
This type of launching shoes is used for the launching of bridges with a variable curve in plan because
they could be installed on a roll-on plate that allows the rotation around a vertical axis. The
inconvenience is their large length that limits the allowed variations of the longitudinal section of the
bridge.
101
© Sétra
Figure 7-4: Launching shoe with wheels and cable.
7.2.2.4 Balancers
In order to provide a favourable long loading length and a small structural height of the launching shoe
at the same time, a balancer as shown in Figure 7-5 and Figure 7-6 can be used. An additional
elastomeric layer between the steel beam and the bridge girder helps to achieve a uniform-like load
distribution of transverse stresses. Besides that, the construction is able to account for the deformation
and the curvature of the launched bridge girder.
Cross-sectional view

Figure 7-5: Balancer of the valley bridge Elben near Siegen, Germany, 2005.
Figure 7-6: Balancer of the valley bridge Elben near Siegen, Germany, 2005.
7.2.3 Sliding skates

The sliding skates could be fixed to the launched girder or to the guidance rails.
In the first case, the load could be introduced in the web right on a transverse stiffener during the
complete launching phase (see Figure 7-7). This is of course favourable to limit the web buckling. The
skate faces are covered with Teflon (polytetrafluoro-ethylene or P.T.F.E.) and the guidance rails are
made of stainless steel. Lubricant could also be used (for instance black soap for Charles de Gaulle
bridge in Paris, in 1996).
102
Launching
In the second case, the bottom flange of the steel girder slides on steel plates covered by a Teflon film
and fixed to the support. During the launching phase, the load could then be introduced between two
vertical web stiffeners.
© Sétra
Figure 7-7: Sliding skate fixed under a vertical stiffener, Verrières viaduct near Millau, France,
2002.
In comparison with the launching shoes (see Paragraph 7.2.2), the use of sliding skates has two main
advantages. The first one is a better introduction of the load in the web plane (more uniformly
distributed), and the second one is the smaller size of the whole launching device combined with a
better load carrying capacity.
7.2.4 Other devices

The transverse load could also be introduced in the web plane through a continuous patch length,
without any sliding on the interface. The progression of the bridge is then obtained by a jack system (for
instance, the Millau viaduct, see Figure 7-8a) or by a track (Japanese technique), see Figure 7-8b).
© Sétra © Sétra
a) Jack system from the Millau viaduct b) Track system (Japan).

(France)
Figure 7-8: Other launching devices.
7.3 Global behaviour during launching
7.3.1 General
It mainly concerns the control of the vertical deflections and the risk of global instability (lateral
torsional buckling). The considered actions are the global bending moment of the bridge girder and
eventually the wind action.
This topic was not part of the COMBRI research project and does not deal with plate buckling, so it is
only quickly mentioned here.
103
7.3.2 Launching nose

In order to recover the deflection of the cantilever part of the structure when a support is reached, and to
lighten this cantilever part, a launching nose is generally used. In extreme cases, a launching pylon can
even be added, see Figure 2-28. The length of the launching nose varies from 20 to 30% of the main
span length.
© Sétra © Sétra
Figure 7-9: Launching nose for the Verrières viaduct near Millau, France, 2002.
7.3.3 Temporary cross-bracings

To avoid lateral torsional buckling during launching phases, a temporary horizontal wind bracing may
be added between the permanent transverse girders or diaphragms of the bridge if these alone are not
sufficient. This wind bracing can be performed with cables or rigid diagonals (tubes or angle sections).
Such devices can be seen in Figure 2-19, Figure 6-5 or Figure 6-6.
7.4 Launching with a part of the concrete slab
7.4.1 General
Three options can be considered:
• launching with the formwork and reinforcing steel bars only.
• launching with prefabricated slabs segments laid on the steel girders but not connected.
• launching with a part of the concrete deck poured in-situ (before launching) and connected to
the steel girders.
The third option leads to a very heavy structure. Then it becomes very difficult to use rolling shoes and
the only solution is to use sliding skates fixed on the steel structure just below a vertical stiffener (see
Figure 7-7). Even using these sliding skates, the maximum span length that could be launched is limited
to around 20 m. The main French experience is a two girder bridge in Cannes over the highway A8,
with a 23-m-long span. For longer span, a longitudinal prestressing of the concrete slab becomes
104
Launching
necessary to avoid excessive cracking in the slab. In any case the concrete slab should not be put on the
cantilever part of the steel girder to limit the negative bending moment on internal supports during
launching. Further concreting is then unavoidable after the launching is completed. This kind of
launching is not a very economical solution and should be kept for cases where the launched bridge
crosses railway lines or roads with heavy traffic that can not be interrupted to handle the concrete
formworks above. One example in France is the Croix Verte Viaduct near Avignon above the TGV
railway line in 1994. A 30-m-long segment has been poured before launching, for a total bridge length
of 138.50 m (see Figure 7-10).
© Sétra (Fotographer: J. Berthellemy)
Figure 7-10: Croix Verte viaduct, launching above train lines, Avignon, France, 1994.
Launching with prefabricated slab segments (laid on the steel girders and acting as a dead load without
any steel-concrete connection) offers some advantages among which the main ones are:
• the reduction of construction time for the bridge deck
• the reduction of risks
• and as a consequence, the reduction of costs
The reduction of risks mainly concerns the workers who act on the launched steel girders for concreting
the slab on site after launching. It is all the more so as the launching technique is chosen (instead of the
use of a crane) when the bridge deck crosses the valley widely above the natural soil. By the use of
precast slab segments, the part of the work performed on the launched steel structure is reduced.
The prefabricated slab segments act as a dead load for the steel girders during launching. This has some
consequences :
• It is not recommended to put the slab segments on the cantilever part of the steel girders,
whatever the considered launching step. It would increase the internal forces and moments and
the risk of lateral torsional buckling in the cantilever part of the steel girders. The vertical
deflection of this cantilever part would also be increased and the accosting on the next support
would become more difficult.
• The weight of the slab segments increases the bending moment in the steel girder and the web
buckling verifications should be performed carefully.
• The interaction between shear force and transverse load should be carefully verified for the web
panel located right above the last crossed internal support. In fact the cantilever steel part of the
bridge (on one side of the considered panel) resists only its self weight whereas on the other
side, the steel girder resists its self weight and the dead load of the slab segments.
105
7.4.2 Application to the calculation example “twin-girder bridge”

One of the outcomes from the COMBRI research project is a modification of the EN 1993-1-5 formulae
(Section 6) for verifying the web buckling above the launching device. This criterion has been proved to
be conservative for slender bridge webs [15]. Then it has been modified by deleting the factor m2 in the
yield resistance formula and by calibrating a new reduction curve. It becomes thus interesting to use this
new criterion for launching heavier structures, for instance the steel part of a composite bridge with
precast slab segments already laid on it.
The launching process has already been verified (according to Section 6 and 10 of EC3 Part 1-5) for a
twin-girder bridge example in Part I of this Design Manual [16]. On the basis of this example,
additional studies have been conducted to study the launching with parts of the concrete slab.
Launching with prefabricated slabs segments

Figure 7-11 shows the most unfavourable launching situation of the twin-girder bridge of Part I of the
COMBRI Design Manual [16]. The adding of concrete slab segments before launching will noticeably
increase the reactions at supports and the stresses in the steel girder. It depends on the spans, the slab
width and the steel girder dimensions, so that the length L has to be evaluated for each project. In this
case, the slab segments are put in place on the left side of the section A, along a length of L = 111.75 m,
avoiding the cantilever part.
For the launching situation A, as shown in Figure 7-11, the internal forces at pier P1 (x = 111.75 m)
become:
• Bending moment MEd = - 19.26 MNm
• Support reaction FEd = 2.67 MN
prefabricated slab segments x = 111.75 m
C0 P1 P2 C3
Figure 7-11: Launching with prefabricated slab segments (Position A).
In case the whole launching procedure is evaluated it can be shown that for a launching with
prefabricated slab segments, the governing launching situation is for section B, as shown in Figure
7-12. Thus, the internal forces at pier P1 (x = 51.50 m) become:
prefabricated slab elements x = 51.50 m
C0 P1 P2 C3
Figure 7-12: Launching with prefabricated slab segments (Position B).
Because the steel cross sections are the same for locations x = 51.50 m and x = 111.75 m, only
launching situation B is evaluated. The results are given in Table 7-1.
106
Launching
Launching with reinforcing steel bars

After the case of launching with some precast slab segments acting as a dead load for the steel structure,
this new paragraph deals with the case where only the reinforcement is put on the steel girders before
launching. In comparison to the previous case, the new launching dead loads will be really lower. This
new lighter dead load has a small influence on the lateral torsional buckling risk for the cantilever part
in a given launching step. So the reinforcement can be installed all along the steel girders before
launching, see Figure 7-13. In comparison to the launching with slab segments, the steel reinforcement
can not be considered as a provisional wind bracing for the steel girders so that a temporary wind
bracing needs to be installed additionally.
For the launching situation A, as shown in Figure 7-13, the internal forces at pier P1 (x = 111.75 m)
become:
reinforcing steel bars x = 111.75 m
C0 P1 P2 C3
Figure 7-13: Launching with reinforcing steel bars (Position A).
Comparison of results
In Table 7-1 all results are summarised and the observed benefits for alternative launchiung methods are
highlighted. It can be shown that due to the large margin the bridge launching of the steel beams with
prefabricated slab segments or with reinforcing steel bars can be justified without additional effort if
current Sec. 6, EN 1993-1-5 [31] is used. As expected, the use of Sec. 10, EN 1993-1-5, leads to
slightly lower resistances. For the launching of the steel beams with partial slab elements the patch
loading resistance cannot be justified when using Section 10. When the proposals from the COMBRI
research project for modifying Sec. 6, EN 1993-1-5, are used, the calculated resistances can be
increased which leads to a larger margin because the level of utilisation is lower. This can be
advantageous for other bridges, for which a verification without adjusting the cross sectional properties
would not be possible. In this example, however, this is not the case.
A detailed description of the calculations can be found in [69].
Table 7-1: Comparison of results for different launchings of a twin-girder bridge.
Launching with …
steel and steel and

steel
some slab elements reinforcing bars
studied section at x [m] 111.75 051.50 111.75
bending moment [MNm] - 19.26 - 23.93 - 24.93
support reaction [MN] 001.46 004.11 001.93
107
Table 7-1 (continued): Comparison of results for different launchings of a twin-girder bridge.
Launching with …
steel and steel and

steel
some slab elements reinforcing bars
Results with Sec. 6, EN 1993-1-5
bending (η1 ) 0.27 0.33 0.34
patch load (η2) 0.35 1.00 0.46
interaction (η2 + 0.8·η1) 0.56 1.26 0.74
margin 60% 10% 47%
Results with proposals from the COMBRI research project for modifiying Sec. 6, EN 1993-1-5
patch load (η2) 0.27 0.76 0.36
saving 22% 24% 22%
interaction (η2 + 0.8·η1) 0.48 1.02 0.63
saving 14% 19% 14%
margin 65% 27% 55%
Results with Sec. 10, EN 1993-1-5, with different buckling curves (Sec. 4)
patch load (=η2) 0.38 1.07 0.50
interaction (interpolated ρ) 0.46 1.001 0.60
margin 54% 0% 40%
Results with Sec. 10, EN 1993-1-5, with a single buckling curve (Ann. B)
pure patch (=η2) 0.46 1.29 0.61
interaction 1/1.73 = 0.58 1/0.81 = 1.23 1/1.32 = 0.76
margin 42% not verified! 24%
108
Summary
8 Summary
This publication is Part II of the Design Manual based on results from the research project „Competitive
Steel and Composite Bridges by Improved Steel Plated Structures - COMBRI“ [15] and the subsequent
dissemination project “Valorisation of Knowledge for Competitive Steel and Composite Structures -
COMBRI+“, both sponsored by RFCS. Part I is a separate publication, which shows the application of
Eurocodes on a composite I-girder bridge and a composite box-girder bridge [16]. This Part II focuses
on the conceptual design of steel bridges and the steel parts of composite bridges and it is based on the
rules in EN 1993-1-5, EN 1993-2 and EN 1994-2. Design of steel bridges is a very wide field and it has
not been covered completely in this manual but a selection of topics has been made and their main
conclusions are summarised below.
In Chapter 2 an overview of bridge types in the countries participating in the project was given:
Belgium, France, Germany, Spain and Sweden. It reflects the current practice in those countries and
presents common bridge types as well as unusual bridges intended to solve special problems and some
solutions being parts development projects. There are notable differences between the practices of the
countries and these differences are to some extent caused by differences between the national design
standards but more often they are caused by different traditions and praxis. Thus, the solutions
presented are intended to serve as inspiration for the conceptual design of new bridges.
In Chapter 3 the choice of steel grades has been discussed. The EN 1993-1-1 covers steel grades up to
and including S460 but EN 1993-1-12 extends the range of permitted steel grades up to S700. However,
in most cases such high grades are not feasible. The problem is usually that the fatigue requirements
limit the full utilization of the strength. The grade S460 seems to be the most suitable for normal road
bridges and S355 for normal rail bridges. It is also shown that hybrid girders with higher strength in the
flange than in the webs are economic in many applications. The box-girder from Part I of this Design
Manual was redesigned from S355 to a hybrid girder with S460 and S690 and it turned out that the cost
of the material was reduced by 10% in the spans and 25% at the piers. In addition, there will be a
reduction of the fabrication cost for sure as well but this has not been quantified.
Flanges are dealt with in Chapter 4 and the main topic has been bottom flanges in box-girders. Such
flanges are in most cases stiffened and different types of stiffeners are discussed. Large trapezoidal
stiffeners are favourable as they give two stiffened lines for the same welding effort as one open
stiffener. Further, their torsional stiffness increases the critical stress and this can be calculated with the
software EBPlate [26] which has been developed in the COMBRI research project. Another topic is the
double composite action with both top and bottom flanges being composite which has been used for
some large bridges in Germany and France. The top flange is as usual the bridge deck and the bottom
flange has a concrete slab at the piers where the bottom flange is in compression. The design of bridges
with double composite action is more complicated than the design of a normal composite bridge so that
past experience is summarised and recommendations for design are given.
Webs have been discussed in Chapter 5 with the focus on to what extent stiffeners should be used. It is
common that transverse stiffeners are used at the locations of the cross bracings of which the transverse
stiffeners form a part. The effect of the transverse stiffeners on the resistance of the web is an increase
in the shear buckling resistance. However, unless the distance between the transverse stiffeners is very
short this effect is small and it does not justify the cost of the stiffeners. The possibility of omitting the
transverse stiffeners is discussed. It should be noted that EN 1993-1-5 does not require any transverse
stiffeners except at the supports. Besides that, longitudinal stiffeners on webs increase the resistance for
bending as well as for shear. The economy of using longitudinal stiffeners has been studied and if the
method with effective cross section in EN 1993-1-5 is applied it is shown that longitudinal stiffeners are
not economical for web depths below ca. 4 m. The detailing of longitudinal stiffeners has been
discussed as well and the main point is the intersection with the transverse stiffeners. One solution is to
109
use discontinuous stiffeners and another is to put the transverse and the longitudinal stiffeners on
opposite sides of the web.
Chapter 6 covers cross bracings and diaphragms for I-girder bridges and box-girders. Functional
requirements are described and ways to meet them are discussed. The main functions are to prevent
lateral torsional buckling and to transfer lateral loads on the girders to the deck. Traditional cross
bracings can be of truss type or frame type including transverse stiffeners on the webs. The distance
between the cross bracings is typically up to 7 to 10 m in I-girder bridges. It is not much material used
for cross bracings but from an economical point of view it is important to minimize the man hours for
fabrication. This has been discussed in terms of eliminating parts and possibly also the transverse
stiffeners leading to straightforward solutions. For box-girders, the cross bracings or diaphragms also
have the function of preventing cross sectional distortion and in many cases they also support the bridge
deck. Therefore the distance between the cross bracings is rather small, typically 4 to 5 m.
Launching has been studied in some detail in the COMBRI research project and it is dealt with in
Chapter 7. The technique of launching bridges has been improved and the method is very popular. It is
described in some detail including the equipment that is used. At launching the resistance to patch
loading is of importance as very high support reactions have to be resisted in combination with high
bending moments. This has been studied in the project and it resulted in improved design rules which
will be finally proposed for inclusion in EN 1993-1-5. The rules allow the utilisation of quite long
loaded lengths and accordingly quite high resistance can be achieved. This may make it possible to
launch bridges with parts of the concrete slab or the reinforcement in place. For the twin-girder bridge
of Part I of this Design Manual, these two possibilities have been studied and the results are compared.
If it is useful to have the concrete slab or the reinforcement already in place, the outcomes of the
COMBRI research project are very helpful and may lead to more economic solutions.
110
References
References
[1] ACHE: Asociación Científico técnica del Hormigón Estructural: Comprobación de un Tabellero
mixto. Comisión 5, Grupoide Trabajo 5/3 “Puentes Mixtos”, Madrid, Spain, 2006.
[2] Allmeier, S.; Frenzel, J.; Schiefer, S.; Seidl, G.; Weber, J.: Innovation im Verbundbrückenbau -
Talbrücke Oberhartmannsreuth. Stahlbau 69 (2000), No. 9, pp. 707-713.
[3] ARS: Allgemeines Rundschreiben Straßenbau 12/03, 2003.
[4] BASt: Bundesanstalt für Straßenwesen (Federal Highway Research Institute): Summary Statistics
of Traffic Data in Germany, 2006.
[5] Bernabeu Larena, J.: Evolución tipológica y estética de los puentes mixtos en Europa. Tesis
Doctoral, Departamento de Mecánica de Medios Continuos y Teoría de Estructuras. Escuela
Técnica Superior de Ingenieros de Caminos, Canales y Puertos Universidad Politécnica de
Madrid, Madrid, Spain, 2004.
[6] Bornscheuer, B.; Eisele, S.: Die Nesenbachtalbrücke - Vom Wettbewerb bis zur Ausschreibung.
Stahlbau 67 (1998), No. 11, pp. 827-830.
[7] Braun, A.; Seidl, G.; Weizenegger, M.: Rahmentragwerke im Brückenbau - Konstruktion,
Berechnung und volkswirtschaftliche Betrachtung. Beton- und Stahlbetonbau 101 (2006), No. 3,
pp. 187-197.
[8] Breuninger, U.: Zum Tragverhalten liegender Kopfbolzendübel unter Längsschubbeanspruchung.
Dissertation, Mitteilung Nr. 2000-1 des Instituts für Konstruktion und Entwurf, Universität
Stuttgart, Germany, 2000.
[9] Bro 2004: Swedish design code for road bridges, 2004.
[10] Bundesministerium für Verkehr: Straßenbrücken in Stahl-Beton-Verbundbauweise.
Bundesministerium für Verkehr, Abteilung Straßenbau, Referat StB 25, Bonn 1997.
[11] Bundesministerium für Verkehr, Bau- und Wohnungswesen: Brücken und Tunnel der
Bundesfernstraßen 2001. Deutscher Bundes-Verlag, Köln, 2001.
[15] COMBRI: Competitive Steel and Composite Bridges by Improved Steel Plated Structures. Final
Report, RFCS research project RFS-CR-03018, 2007.
[16] COMBRI+: COMBRI Design Manual - Part I: Application of Eurocode rules. RFCS project
RFS2-CT-2007-00031, 2008.
[17] Crespo Rodríguez, P.: Verification of a composite bridge deck according to the Spanish RPX
code and Eurocode 4 - Comparative analysis. Third International Meeting on Composite Bridges,
Madrid, Spain, 2001, pp. 479-496.
[18] Denzer, G.; Gräßlin, W.; Hanswille, G.; Schmidtmann, W.: Die Talbrücke über die Wilde Gera -
Erfahrungen bei der Planung und Ausführung von Talbrücken mit einteiligen
Verbundquerschnitten. Stahlbau 69 (2000), No. 11, pp. 842-850.
111
[19] Denzer, G.; Schmackpfeffer, H.: Entwicklungen im Stahlverbundbrückenbau - Autobahnbrücken

mit einteiligen Verbundquerschnitten. Stahlbau 74 (2005), No. 9, pp. 649-656.
[20] Deutsche Bahn AG: Facts and Figures of the Annual Report 2007, 2008.
[21] Deutsche-Einheit-Fernstraßenplanungs- und -bau GmbH: Brückenbauwerke in den neuen
Bundesländern. Ernst & Sohn Verlag, Berlin, 2004.
[22] Deutsche-Einheit-Fernstraßenplanungs- und -bau GmbH: Acht Großbrücken an der A73. Hubert
von Brunn, Berlin, 2007.
[23] DIN-Fachbericht 103: Stahlbrücken. Edition March 2003.
[24] Doss, W.; Gebeshuber, A.; Friedrich, N.; Schmitt, V.; Seidl, G.; Weizenegger, M.: VFT-
Bauweise - Entwicklung von Verbundfertigteilträger im Brückenbau. Beton- und Stahlbetonbau
96 (2001), No. 4, pp. 171-180.
[25] Dürr, A.: Zur Ermüdungsfestigkeit von Schweißkonstruktionen aus höherfesten Baustählen bei
Anwendung von UIT-Nachbehandlung. Dissertation, Mitteilung Nr. 2006-3 des Instituts für
Konstruktion und Entwurf, Universität Stuttgart, Germany, 2006.
[26] EBPlate: A piece of software developed in the frame of the COMBRI research project [15]. Its
aim is to assess the elastic critical stresses of plates. EBPlate is free of charge and can be
downloaded from the web site of cticm: www.cticm.com
[27] Eilzer, W.; Reintjes, K.-H.; Pötzsch, D.; Schreiber, O.: Entwurf und Ausführung der
Seidewitztalbrücke im Zuge der BAB A 17. Stahlbau 75 (2006), No. 2, pp. 117–127.
[28] EN 1990/A1 Eurocode: Basis of structural design – Application for bridges, December 2005.
[29] EN 1991-2: Eurocode 1: Actions on structures – Part 2: Traffic loads on bridges, September 2003.
[30] EN 1993-1-1: Eurocode 3: Design of steel structures – Part 1-1: General rules and rules for
buildings, May 2005.
[31] EN 1993-1-5: Eurocode 3: Design of steel structures – Part 1-5: Plated structural elements,
October 2006.
[32] EN 1993-1-9: Eurocode 3: Design of steel structures – Part 1-9: Fatigue, May 2005.
[33] EN 1993-1-10: Eurocode 3: Design of steel structures – Part 1-10: Material toughness and
through-thickness properties, May 2005.
[34] EN 1993-1-12: Eurocode 3: Design of steel structures - Part 1-12: Additional rules for the
extension of EN 1993 up to steel grades S700, February 2007.
[35] EN 1993-2: Eurocode 3: Design of steel structures – Part 2: Steel Bridges, October 2006.
[36] EN 1994-2: Eurocode 4: Design of composite steel and concrete structures – Part 2: General rules
and rules for bridges, October 2005.
[37] EN 10025: Hot rolled products of structural steels, November 2004.
[38] EN 10164: Steel products with improved deformation properties perpendicular to the surface of
the product – Technical delivery conditions, December 2004.
[39] Hanswille, G., Sedlacek, G.: Steel and Composite Bridges in Germany – State of the Art.
Norwegian Steel Conference, Oslo, Norway, October 18th, 2007.
[40] Hertle, R.: Zur dynamischen Analyse von schubweich und diskret gekoppelten
Mehrschichtenträgern. Dissertation, Berichte aus dem Konstruktiven Ingenieurbau No. 2/1992,
TU München, Germany, 1992.
[41] Hilgendorff, K.-D.; Neumann, W.; Reitz, D.; Schmitz, C.: Talbrücke Elben - Eine
Stahlverbundbrücke mit einteiligem Querschnitt. Stahlbau 76 (2007), No. 5, pp. 320-335.
[42] IAP: Instrucción sobre las acciones a considerar en el proyecto de puentes de carretera. Ministerio
de Fomento, Dirección General de Carreteras, Madrid, Spain, 1998.
112
References
[43] Ibach, H.: Zum Kriechen und Schwinden von Verbundbrücken auf Grundlage der Eurocodes.
Dissertation, Berichte aus dem Konstruktiven Ingenieurbau No. 2/2001, TU München, Germany,
2001.
[44] Johansson, B.: Buckling resistance of structures of high strength steel. IABSE Structural
Engineering Documents 8, 2005, pp. 120-128.
[45] Johansson, B.; Maquoi, R.; Sedlacek, G.; Müller, C.; Beg, D.: Commentary and worked examples
to EN 1993-1-5 "Plated structural elements". Joint report JRC-ECCS, 2007.
[46] Kuhlmann, U.; Detzel, A.: Verbundbrücken. In: Mehlhorn, G. (Hrsg.): Handbuch Brücken,
Springer Verlag, Berlin, Germany, 2007, pp. 209-219.
[47] Kuhlmann, U.; Dürr, A.; Roos, F.: Ermüdungsnachweis für Straßen-Verbundbrücken nach DIN-
Fachbericht. Stahlbau 72 (2003), No. 9, pp. 679-682.
[48] Kuhlmann, U.; Ehmann, J.; Raichle, J.: Messungen an der Nesenbachtalbrücke. Final Report,
Mitteilung Nr. 2004-16X des Instituts für Konstruktion und Entwurf, Universität Stuttgart,
Germany, 2004.
[49] Kulka, H.; Schaumann, P.: Neue Entwicklungen im Verbundbrückenbau. Straßen und Tiefbau 45
(1991), No. 5, pp. 6-16.
[50] Langen, T.; Pfisterer, H.; Skusa, F.; Weber, J.: Größte Doppelverbund-Straßenbrücke
Deutschlands – Die Innbrücke Neuötting. Stahlbau 69 (2000), No. 11, pp. 833-841.
[51] Llombart Jaques, J.A.; Revoltós Fort, J.; Couto Wörner, S.: Puente sobre el río Tajo, en el
embalse de Alcántara (Arcos de Alconétar). Hormigón y Acero Nº 242, 4.er Trimestre (2006),
Madrid, Spain.
[52] Manterola Armisén, J.: Composite arch bridges. Proceedings of the 3rd International Meeting on
Composite Bridges (Hrsg.: Martínez Calzón, J.), Madrid, Spain, 2001, pp. 173-185.
[53] Manterola Armisén, J.: New proposals in steel bridges and footbridges. Proceedings of the 5th
International Symposium on Steel Bridges, Barcelona, Spain, March 5th-7th, 2003.
[54] Martínez Calzón, J.: Strict box composite bridges - A new design of the optimum use of
composite typology. Proceedings of the 12th Annual International Bridge Conference, Pittsburg,
USA, 1995.
[55] Martínez Calzón, J.: The abacus system for the launching of large span constant edge composite
bridges - Types and possibilities. Proceedings of the 3rd International Meeting on Composite
Bridges (Hrsg.: Martínez Calzón, J.), Madrid, Spain, 2001, pp. 219-231.
[56] Martínez Cutillas, A.: Aplicación de los métodos generales de cálculo no lineal al estudio de los
puentes mixtos. Proceedings of the 3rd International Meeting on Composite Bridges (Hrsg.:
Martínez Calzón, J.), Madrid, Spain, 2001, pp. 234-261.
[57] Ramberger, G.; Aigner, F.: Stahlbrücken. In: Mehlhorn, G. (Hrsg.): Handbuch Brücken, Springer
Verlag, Berlin, Germany, 2007, pp. 828-862.
[58] Millanes Mato, F.: Comparative analysis of double composite action launched solutions and
prestressed solutions in high speed railway viaduct. Proceedings of the 3rd International Meeting
on Composite Bridges (Hrsg.: Martínez Calzón, J.), Madrid, Spain, 2001, pp. 383-404.
[59] Millanes Mato F., Pascual, J.: The viaduct across the “Arroyo de las Piedras” in the high speed
line between Córdoba and Málaga - An innovative solution for the first high speed line steel
concrete composite bridge in Spain. Proceedings of the 5th International Symposium on Steel
Bridges, Barcelona, Spain, March 5th-7th, 2003.
[60] Millanes Mato, F.: Construir en acero los puentes de luces medias. Asociación para la Promoción
Técnica del Acero, APTA, Madrid, Spain, 2006 (http://www.apta.com.es/pdf
/Jornada%202004/francisco_millanes.pdf).
[61] Millanes Mato, F.; Pascual Santos, J.; Ortega Cornejo, M.: Viaducto “Arroyo las Piedras” -
Primer viaducto mixto de las Líneas de Alta Velocidad Españolas. Hormigón y Acero Nº 243, 1.er
Trimestre (2007), Madrid, Spain.
113
[62] Nather, F.: Stahlbrücken mit Doppelverbund in Deutschland - Überblick und Ergebnisse von
Forschungsaufträgen. Bauingenieur 72 (1997), pp. 131-141.
[63] Ndogmo, J.: Doppelverbundbrücken im Bau- und Endzustand. Stahlbau 75 (2006), No. 8, pp.
670-677.
[64] Neuner, F.: Zur Berechnung prismatoidisch gefalteter Flächentragwerke unter besonderer
Berücksichtigung elastischer Fugen. Dissertation, Berichte aus dem Konstruktiven Ingenieurbau
No. 1/93, TU München, Germany, 1993.
[65] OTUA 1997: Les aciers thermomécaniques. Documents scientifiques et techniques de l’AFPC
(Association Française Pour la Construction) et de l’OTUA (Office Technique pour l’Utilisation
de l’Acier). ISSN 0150-6900, May 1997.
[66] Pantaleón Prieto, M.J.; J. A., Ramos Gutiérrez, O. R.; Ortega Carreras. G.; Martínez García, J. M.:
Viaducto de Escaleritas en Las Palmas de Gran Canaria. Hormigón y Acero Nº 244, 2.er
Trimestre (2007), Madrid, Spain.
[67] Pauser, A.: Brücken in Wien. Springer Verlag, Wien, Austria, 2005.
[68] Petel, A.; Davaine, L.: Design of bracing frames for the twin-girder bridge. Background document
COMBRI+054, RFCS project RFS2-CT-2007-00031, 2008.
[69] Petel, A.; Davaine, L.; Braun, B.: Launching of the composite twin-girder bridge. Background
document COMBRI+037, RFCS project RFS2-CT-2007-00031, 2008.
[70] Peter, J.; Wetzel, R.; Bach, T.: Die Reichenbachtalbrücke der Thüringer Waldautobahn A 71.
Stahlbau 72 (2003), No. 3, pp. 127-138.
[71] RASQ: Richtlinie für die Anlage von Straßen - Teil: Querschnitte, 1996.
[72] Reintjes, K.-H.: Die Bauweise der dichten Kleinkästen im Verbundbrückenbau. Stahlbau 74
(2005), No. 9, pp. 657-663.
[73] Reintjes, K.-H.; Gebert, G.: Das Zügelgurt-Fachwerk der Muldebrücke Wurzen. Stahlbau 75
(2006), No. 8, pp. 613-623.
[74] Reintjes, K.-H.; Schreiber, O.: Die Stahlverbundbrücke über die Müglitz im Zuge der Autobahn
Dresden–Prag - Entwurf und Ausführung. Stahlbau 73 (2004), No. 1, pp. 4-13.
[75] Roesler, H.; Schmidt, D.: Die neue Leinetalbrücke bei Heiligenstadt. Stahlbau 72 (2003), No. 6,
pp. 422–425.
[76] Roik, K.-H.: Taking Steel Structures into the 20th Century. Symposium, Luxemburg, 1990.
[77] RPM. Recomendaciones para el proyecto de puentes metálicos para carreteras. Ministerio de
Fomento, Dirección General de Carreteras, Madrid, Spain, 1996.
[78] RPX. Recomendaciones para el proyecto de puentes mixtos para carreteras. Ministerio de
Fomento, Dirección General de Carreteras, Madrid, Spain, 1996.
[79] Manual de aplicación de las recomendaciones RPM-RPX/95. Ministerio de Fomento, Dirección
General de Carreteras, Madrid, Spain, 2000.
[80] Rui-Wamba, J.: The Spanish contribution to design and construction of steel bridges. Proceedings
of the 5th International Symposium on Steel Bridges, Barcelona, Spain, March 5th-7th, 2003.
[81] Schmackpfeffer, H.: Typenentwürfe für Brücken in Stahlverbundbauweise im mittleren
Stützweitenbereich (Teil I). Stahlbau 68 (1999), No. 7, pp. 264-276.
[82] Schmackpfeffer, H.: Typenentwürfe für Brücken in Stahlverbundbauweise im mittleren
Stützweitenbereich (Teil II). Stahlbau 70 (2001), No. 4, pp. 429-435.
[83] Schmidtmann, W.; Wagner, P.: Die Werratalbrücke Einhausen – neue Wege beim Bau einer
Autobahn-Verbundbrücke mit getrennten Überbauten. Stahlbau 74 (2005), No. 10, pp. 735-748.
[84] Schmitt, V.: Untersuchungen zum verstärkten Einsatz von Stahlverbund-konstruktionen bei
Brücken kleiner und mittlerer Stützweiten. Final report, FOSTA research project P629, 2005.
114
References
[85] Schmitt, V.; Seidl, G.: Verbundfertigteil-Bauweise im Brückenbau. Stahlbau 70 (2001), No. 8, pp.
546-553.
[86] Schmitt, V.; Seidl, G.; Hever, M.; Zapfe, C.: Verbundbrücke Pöcking – Innovative VFT-Träger
mit Betondübeln. Stahlbau 73 (2004), No. 6, pp. 387-393.
[87] SUSBRI: Sustainable Bridges: European Railway Bridge Demography. Document WP1-02-T-
040601-F D 1.2, EU research proejct TIP3-CT-2003-001653, 2004.
[88] Tschumi, M.; Grüter, R.; Ramondenc, P.: Fortschritte im Eisenbahn-Stahlbrückenbau Europas.
Stahlbau 67 (1998), No. 8, pp. 612-626.
[89] Veljkovic, M.; Johansson B.: Design of hybrid steel girders. Proceedings of the Third European
Conference on Steel Structures (Vol. 1), Coimbra, Portugal, September 19th-20th, 2002.
115
List of figures
List of figures
Figure 2-1: Wide I-girder bridge with cantilevering cross girders (Bridge near Remoulins, France). ..... 3
Figure 2-2: Box-girder bridge with edge beams supporting the deck. (Verrières viaduct near Millau,
France, 2002)........................................................................................................................ 3
Figure 2-3: Three regions in Belgium (Region of Brussels, Flemish and Walloon region)...................... 4
Figure 2-4: Distribution of existing structural types of road bridges in the Walloon region. ................... 5
Figure 2-5: Distribution of existing bridge types of road bridges in the Walloon region. ........................ 5
Figure 2-6: Number of bridge openings per year in the Walloon region. ................................................. 5
Figure 2-7: Distribution of bridge lengths of road bridges in the Walloon region.................................... 6
Figure 2-8: Distribution of existing bridge types of railway bridges in Belgium. .................................... 7
Figure 2-9: Span profiles of existing railway bridges in Belgium. ........................................................... 7
Figure 2-10: Bridge age profiles of railway bridges in Belgium............................................................... 7
Figure 2-11: The "Eau Rouge" viaduct near Malmédy, Belgium, 1993. .................................................. 8
Figure 2-12: The "Gueule" viaduct near Moresnet, Belgium, 1917/2005................................................. 9
Figure 2-13: The “Secheval” viaduct, Belgium, 1979............................................................................. 10
Figure 2-14: The viaduct of Remouchamps, Belgium, 1980. ................................................................. 11
Figure 2-15: The “Croupets du Moulin” viaduct near Sart, Belgium, 1979............................................ 12
Figure 2-16: The viaduct of Polleur, Belgium......................................................................................... 12
Figure 2-17: Distribution of the French new bridges in 2004 according to the type of structure. .......... 13
Figure 2-18: Distribution of the French new (railway and road) bridges in 2004 according to the
main span length................................................................................................................. 13
Figure 2-19: Usual transverse cross-sections of a composite twin-girder bridge.................................... 14
Figure 2-20: Transverse cross girders supporting the slab. ..................................................................... 14
Figure 2-21: Concreting on internal supports at the end. ........................................................................ 15
Figure 2-22: Use of prefabricated slab segments. ................................................................................... 15
Figure 2-23: Use of pre-slab as formwork. ............................................................................................. 15
Figure 2-24: Built-up of a bridge I-girder and welding of transverse stiffeners. .................................... 17
Figure 2-25: La Risle viaduct near Brionne, France, 2004. .................................................................... 18
Figure 2-26: LEO viaduct over Durance near Avignon, France, 2008. .................................................. 18
Figure 2-27: Viaduct over Ourcq valley, France, 2006........................................................................... 19
Figure 2-28: Verrières viaduct near Millau, France, 2002. ..................................................................... 19
Figure 2-29: Jaulny railway viaduct, France, 2005. ................................................................................ 20
Figure 2-30: River bridge Spree near Cottbus, Germany, 1994 [10]. ..................................................... 20
Figure 2-31: Bridge Schrotetal near Magdeburg, Germany, 1997 [10]. ................................................. 21
Figure 2-32: Distribution of span and bridge lengths in Germany [84], [87].......................................... 21
117
Figure 2-33: Distribution of existing bridge types in Germany. ............................................................. 22

Figure 2-34: Distribution of current construction types in Germany [84]............................................... 22
Figure 2-35: Layout of a one-piece composite superstructure. ............................................................... 24
Figure 2-36: Joints foreseen to take the horizontal bracings in case of slab exchange. .......................... 26
Figure 2-37: Small stands as support for formwork carriages “running on top“, valley bridge
Reichenbach near Ilmenau, Germany, 2002....................................................................... 27
Figure 2-38: Example of a formwork carriage “running on top“. ........................................................... 27
Figure 2-39: Foldable formwork panels, valley bridge Schwarza, Germany, 2002................................ 28
Figure 2-40: Valley bridge Nesenbachtal, Stuttgart, Germany, 2000. .................................................... 28
Figure 2-41: Bridges with airtight small-size box-girders....................................................................... 29
Figure 2-42: Typical layout of composite bridges with prefabricated slabs. Only slab system based
on [71] is shown. Dimensions in [m]. ................................................................................ 32
Figure 2-43: Example cross-sections of large area formwork elements based on [81] for the layouts
of Figure 2-42. Dimensions in [m]. .................................................................................... 33
Figure 2-44: Erection of a bridge with steel girders and precast concrete elements near Ravensburg,
Germany. ............................................................................................................................ 34
Figure 2-45: Typical cross-sections of prefabricated composite girder with in-situ concrete................. 34
Figure 2-46: Erection of a prefabricated composite girder with horizontal studs near Münsingen,
Germany. ............................................................................................................................ 35
Figure 2-47: Case study for comparison between Spanish recommendations and Eurocodes for steel
and steel and concrete composite bridge design [1], [17]. ................................................. 39
Figure 2-48: Final design of the sections and arrangements of stiffeners and diaphragms [1], [17]. ..... 40
Figure 2-49: Innovative designs for bridges in Spain [60]. ..................................................................... 40
Figure 2-50: Puente Betxi Borriol, Valencia, Spain [60]. ....................................................................... 41
Figure 2-51: Further examples of composite bridges in Spain................................................................ 42
Figure 2-52: The viaduct of Tina Menor, elevation and plan view [56]. ................................................ 44
Figure 2-53: The viaduct of Tina Menor, cross section and detail of the concrete slab in the bottom
flange (light grey) to obtain the double composite action for the cross sections closer
to the piers [56]. ................................................................................................................. 44
Figure 2-54: The viaduct of Tina Menor, lateral truss [56]..................................................................... 44
Figure 2-55: The viaduct of Tina Menor, propped construction [56]...................................................... 45
Figure 2-56: The viaduct of the Arroyo de las Piedras [61]. ................................................................... 46
Figure 2-57: Cross section of the Arroyo de las Piedras viaduct [61]..................................................... 46
Figure 2-58: Erection of the Arroyo de las Piedras viaduct [61]............................................................. 47
Figure 2-59: Launching of the Arroyo de las Piedras viaduct [61]. ........................................................ 48
Figure 2-60: Cross section of the bridge at Rångedala, Sweden. ............................................................ 49
Figure 2-61: Bridge at Rångedala, Sweden, during erection................................................................... 49
Figure 2-62: Integral abutment bridge which is supported by steel piles under the back wall. .............. 50
Figure 2-63: Plan and cross section of the prefabricated bridge. Elements 1 and 18 are the
prefabricated back walls/end screens. ................................................................................ 51
Figure 2-64: Plan of a bridge deck element and cross section above girder. .......................................... 51
Figure 2-65: Bridge deck element being lifted in place. ......................................................................... 52
118
List of figures
Figure 2-66: Illustration of the narrow tolerances. .................................................................................. 52

Figure 2-67: Elevation and cross section of the bridge over Veckefjärden. The cross bracings are
made of channels bolted to the web stiffeners.................................................................... 53
Figure 2-68: Picture of bridge over Veckefjärden during wintertime. .................................................... 53
Figure 2-69: Bridge over E4 at Höga Kusten, Sweden. .......................................................................... 54
Figure 2-70: View of the Vallsundet bridge, Sweden, and the triangular supports over the
navigation channel.............................................................................................................. 54
Figure 2-71: Casting of the concrete deck of the Vallsundet bridge, Sweden. ....................................... 55
Figure 2-72: Cross section as well as details over supports for the bridge over Vallsundet, Sweden..... 55
Figure 3-1: Approximate price per tonne of hot rolled steel normalised with price of S235 as
function of yield strength. .................................................................................................. 58
Figure 3-2: Approximate material cost normalised with the cost of S235 assuming that the strength
can be fully utilised. ........................................................................................................... 58
Figure 3-3: Relative material cost for web with flexible end stiffener subject to shear. Reference
cost is for S235 [44]. .......................................................................................................... 60
Figure 3-4: Stress distribution in the web of a hybrid girder in hogging bending................................... 60
Figure 3-5: Two welding details of a twin-girder bridge. Dimensions in [mm]. .................................... 64
Figure 3-6: Cost reduction by re-design the box-girder bridge in S460/690 instead of S355................. 66
Figure 4-1: I-girder with inclined flange plates at the bottom flange...................................................... 70
Figure 4-2: Box-girder with open stiffeners, Lille, France...................................................................... 71
Figure 4-3: Box-girder with closed stiffeners, Millau viaduct, France. .................................................. 71
Figure 4-4: Typical cross-section of a German composite bridge with box-girder and trapezoidal
stiffeners [39]. .................................................................................................................... 71
Figure 4-5 Box-girder with trapezoidal stiffeners on the bottom flange. ................................................ 72
Figure 4-6: Elevation of the Inn river bridge Wasserburg, Germany, 1987............................................ 74
Figure 4-7: Elevation of the Elbe river bridge Torgau, Germany, 1993. ................................................ 75
Figure 4-8: Elevation and longitudinal section of the Mosel river bridge Bernkastel-Kues,
Germany, 1995. .................................................................................................................. 75
Figure 4-9: Longitudinal section of the Inn river bridge Neuötting, Germany, 2000. ........................... 76
Figure 5-1 T-shaped vertical stiffeners.................................................................................................... 80
Figure 5-2: Static scheme for the verification of stiffeners. .................................................................... 80
Figure 5-3 Rigid and non rigid end post.................................................................................................. 81
Figure 5-4: View of girder without (top) and with longitudinal stiffener (bottom), not to scale. List
of material with all plates in S460...................................................................................... 82
Figure 5-5: View of girder without (top) and with longitudinal stiffener (bottom), not to scale. List
of material with flanges in S460 and webs in S355. .......................................................... 83
Figure 5-6: Twin-girder bridge in Triel-sur-Seine, France, 2003............................................................ 84
Figure 5-7: Fatigue detail categories for a single flat stiffener................................................................ 85
Figure 5-8: Cut outs in transverse stiffener ............................................................................................. 85
Figure 5-9: Constructional detail of the intersection “longitudinal stiffener and cross-frame”. ............. 86
Figure 5-10: River bridge Nordsteg in Vienna, Austria, 1996. ............................................................... 86
Figure 5-11: Cross-sectional view of the Nordsteg in Vienna, Austria, 1996......................................... 86
119
Figure 5-12: Launching of a railway bridge near Riesa, Germany, 2005. .............................................. 87
Figure 6-1: Cross bracing in the shape of a frame................................................................................... 89
Figure 6-2: Cross bracing as a full lattice................................................................................................ 89
Figure 6-3: Twin-girder bridge near Avignon, France, 2008. ................................................................. 90
Figure 6-4: Railway twin-girder bridge (TGV Est, Canal de l'Ourcq) with diaphragms, France,
2006.................................................................................................................................... 90
Figure 6-5: Twin-girder bridge in Sens, France. ..................................................................................... 90
Figure 6-6: Railway twin-girder bridge (TGV Est, Pont à Mousson), France. ....................................... 90
Figure 6-7: Modelled transverse frame. .................................................................................................. 91
Figure 6-8: Load cases for the rigidity Cd calculation............................................................................. 91
Figure 6-9: Non-strengthened transverse cross-bracing in span. Dimensions in [mm]. ......................... 92
Figure 6-10: Strengthened transverse cross-bracing in span. Dimensions in [mm]. ............................... 92
Figure 6-11: Alternative transverse cross-bracing in span. ..................................................................... 93
Figure 6-12: Normal force in alternative bracing frames, for the study of the members buckling. ........ 94
Figure 6-13: Cross bracing without vertical stiffeners. ........................................................................... 96
Figure 7-1: Particular launching case. ..................................................................................................... 99
Figure 7-2: Launching shoe with 6 wheels (maximum loads = 330 tons). ........................................... 101
Figure 7-3: Launching shoe with 4 wheels and 2 lateral wheels for longitudinal guidance (total
length = 1.25 m). .............................................................................................................. 101
Figure 7-4: Launching shoe with wheels and cable. ............................................................................. 102
Figure 7-5: Balancer of the valley bridge Elben near Siegen, Germany, 2005. .................................... 102
Figure 7-6: Balancer of the valley bridge Elben near Siegen, Germany, 2005. .................................... 102
Figure 7-7: Sliding skate fixed under a vertical stiffener, Verrières viaduct near Millau, France,
2002.................................................................................................................................. 103
Figure 7-8: Other launching devices. .................................................................................................... 103
Figure 7-9: Launching nose for the Verrières viaduct near Millau, France, 2002. ............................... 104
Figure 7-10: Croix Verte viaduct, launching above train lines, Avignon, France, 1994....................... 105
Figure 7-11: Launching with prefabricated slab segments (Position A). .............................................. 106
Figure 7-12: Launching with prefabricated slab segments (Position B). .............................................. 106
Figure 7-13: Launching with reinforcing steel bars (Position A).......................................................... 107
120
List of tables
List of tables
Table 2-1: Indication for construction times of a twin-girder bridge.. .....................................................16
Table 2-2: Transport limitations in France. ..............................................................................................17
Table 2-3: Data of bridges with a one-piece composite superstructure in Germany................................25
Table 2-4: Data of bridges with airtight small-sized box-girders in Germany.........................................30
Table 2-5: Data of bridges with prefabricated composite girders and in-situ concrete in Germany........36
Table 2-6: Steel bridges in Spain by typologies.......................................................................................37
Table 3-1: Summary of national requirements and praxis for bridges. ....................................................59
Table 3-2: Change in cost when the web is changed from S460 to S355. ...............................................61
Table 3-4: Data of French bridges with steel grades higher than S355....................................................63
Table 3-5: Comparison of the steel quantities. .........................................................................................65
Table 3-6: Comparison of the deflections at mid-span.............................................................................66
Table 3-7: Comparison of material costs..................................................................................................66
Table 4-1: Data of bridges with double-composite action in Germany....................................................76
Table 6-1: Transverse displacement of the first three elastic critical buckling modes.............................93
Table 6-2: Steel quantities for cross bracings...........................................................................................95
Table 6-3: Steel saving on cross bracings from using reduction curve “b” instead of curve “d” for
lateral torsional buckling. ......................................................................................................95
Table 7-1: Comparison of results for different launchings of a twin-girder bridge. ..............................107
121

2.b. COMBRI Design Manual Part II English

Uploaded by

Copyright:

Available Formats

2.b. COMBRI Design Manual Part II English

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2.b. COMBRI Design Manual Part II English

Uploaded by

Copyright:

Available Formats

What are some of the bridge types introduced in different countries?

What are some of the bridge types introduced in different countries?

What are some methods of bridge construction discussed in the document?

What are some methods of bridge construction discussed in the document?

COMBRI DESIGN MANUAL

Part II: State-of-the-Art and Conceptual Design

Part II: State-of-the-Art and Conceptual Design

Cover pictures (from left to right):

Ulrike Kuhlmann, Benjamin Braun

2.6.3 Bridges with integral abutments ..............................................................................50

5.4.4 Discontinuous longitudinal stiffeners......................................................................85

1 Introduction and scope

Flemish region Walloon region

2.2.2 Road bridges in Walloon Region

16.5% 16.8% 15.6% 16.4%

Bridge length [m]

2.2.3 Railway bridges in Belgium

Figure 2-8: Distribution of existing bridge types of railway bridges in Belgium.

Figure 2-9: Span profiles of existing railway bridges in Belgium.

Figure 2-10: Bridge age profiles of railway bridges in Belgium.

2.2.4 Bridge examples

2.2.4.1 Composite girder bridge

The "Eau Rouge" viaduct

The "Gueule" viaduct

Figure 2-12: The "Gueule" viaduct near Moresnet, Belgium, 1917/2005.

The "Secheval" viaduct

Figure 2-13: The “Secheval” viaduct, Belgium, 1979.

The viaduct of Remouchamps

Figure 2-14: The viaduct of Remouchamps, Belgium, 1980.

The “Croupets du Moulin” viaduct

2.2.4.2 Composite box-girder bridges

The viaduct of Polleur

Figure 2-16: The viaduct of Polleur, Belgium.

Reinforced Filler beam

25% Filler beam

2.3.2 Types of composite bridges in France

2.3.2.2 Twin-girder bridges

Figure 2-19: Usual transverse cross-sections of a composite twin-girder bridge.

Figure 2-20: Transverse cross girders supporting the slab.

Figure 2-21: Concreting on internal supports at the end.

Figure 2-22: Use of prefabricated slab segments.

Discussions of the design from the economical point of view

Table 2-1: Indication for construction times of a twin-girder bridge.

Steps Time consumption Weight

Fabrication of the main girders

Fabrication of the cross-bracing 5% 6%

Figure 2-24: Built-up of a bridge I-girder and welding of transverse stiffeners.

Table 2-2: Transport limitations in France.

Exceptional convoy (with

Width 4.50 m 6.00 m 3.00 m (and a concomitant

Depth 3.20 m 6m 3.05 m (and a concomitant

Figure 2-25: La Risle viaduct near Brionne, France, 2004.

Figure 2-27: Viaduct over Ourcq valley, France, 2006.

2.3.2.3 Examples of composite box-girder bridges

Figure 2-28: Verrières viaduct near Millau, France, 2002.

Figure 2-29: Jaulny railway viaduct, France, 2005.

Figure 2-31: Bridge Schrotetal near Magdeburg, Germany, 1997 [10].

a) Road bridges (federal highways) [84].

b) Railway bridges [87].

a) Road bridges (federal highways) [84]. b) Railway bridges [87].