Bridges are an integral part of the worldwide traffic infrastructure and long-span bridges

KI-NA-24175-EN-C
especially contribute to mobility and economy of time in travelling. Improvements of the
steel-plated cross-sections of steel and composite bridge structures help to enhance the
competitiveness of such bridges. Herein the aims are the valorisation and dissemina-
tion of the knowledge and results which have been acquired within the preceding RFCS
research project ‘Competitive steel and composite bridges by improved steel plated
structures — Combri’ for practitioners with regard to plate buckling verifications. The
outcome is the Combri design manual consisting of two parts which provide clearly
arranged and concise documents for daily use. Part I ‘Application of Eurocode rules’
covers two composite bridge structures — a twin-girder and a box-girder bridge — on
the basis of worked examples for which the knowledge is written down in a descriptive
manner and references are given to current Eurocode rules. Part II ‘State-of-the-art and
conceptual design of steel and composite bridges’ presents the current practice in sev-
eral European countries and common bridge types as well as unusual bridges for special
purposes or development projects. 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. In this report, proposals are also formu-
lated to implement the newly gained state-of-the-art knowledge into standardisation via
nationally determined parameters (NDP), non-contradictory complementary information
(NCCI) and suggestions for the next revision of the Eurocodes.

Combri
Valorisation of knowledge for competitive
steel and composite bridges

Price (excluding VAT) in Luxembourg: EUR 4

Interested in European research?

RTD info is our quarterly magazine keeping you in touch with main developments (results, How to obtain EU publications
programmes, events, etc.). It is available in English, French and German. A free sample copy Free publications:
or free subscription can be obtained from:
• via EU Bookshop (http://bookshop.europa.eu);
Directorate-General for Research • at the European Commission's representations or delegations.
Information and Communication Unit You can obtain their contact details by linking http://ec.europa.eu
European Commission or by sending a fax to +352 2929-42758.
B-1049 Brussels
Fax (32-2) 29-58220
Publications for sale:
E-mail: [email protected] • via EU Bookshop (http://bookshop.europa.eu);
Internet: http://ec.europa.eu/research/rtdinfo/index_en.html
• Priced subscriptions (Official Journal of the EU, Legal cases of the Court of
Justice as well as certain periodicals edited by the European Commission) can
be ordered from one of our sales agents.
You can obtain their contact details by linking http://bookshop.europa.eu, or by
sending a fax to +352 2929-42758.

EUROPEAN COMMISSION
Directorate-General for Research
Research Fund for Coal and Steel Unit

Contact: RFCS publications

Address: European Commission, CDMA 0/124, B-1049 Brussels
Fax (32-2) 29-65987; e-mail: [email protected]
 European Commission

Research Fund for Coal and Steel

Combri
Valorisation of knowledge for competitive steel
and composite bridges
U. Kuhlmann, B. Braun (1), M. Feldmann, J. Naumes (2), P.-O. Martin, Y. Galéa (3),
J. Raoul, L. Davaine, A. Petel (4), B. Johansson, P. Collin, J. Eriksen (5),
H. Degée, N. Hausoul (6), J. Chica, S. Meno (7)
(1) Institute for Structural Design, Universität Stuttgart — Pfaffenwaldring 7, 70569 Stuttgart, GERMANY
(2) Institute for Steel Structures, RWTH Aachen — Mies-van-der-Rohe-Straße 1, 52074 Aachen, GERMANY
(3) CTICM — Centre Technique Industriel de la Construction Métallique
Espace technologique, L’orme des merisiers, Immeuble Apollo, 91193 Saint-Aubin, FRANCE
(4) SETRA — Service d’Études sur les Transports, les Routes et leurs Aménagements,
46, avenue Aristide Briand, 92225 Bagneux Cedex, FRANCE
(5) Division of Steel Structures — Luleå University of Technology, SE-971 87 Luleå, SWEDEN
(6) Department ArGEnCo, Division of Structural Engineering, University of Liège —
Chemin des Chevreuils 1, 4000 Liège, BELGIUM
( ) Fundacion LABEIN, Geldo — Parque Tecnológico de Bizkaia, Edificio 700, 48160 Derio, SPAIN
7

Contract No RFS2-CT-2007-00031
1 July 2007 to 31 December 2008

Final report

Directorate-General for Research

2010 EUR 24175 EN

LEGAL NOTICE

Neither the European Commission nor any person acting on behalf of the Commission
is responsible for the use which might be made of the following information.

Europe Direct is a service to help you find answers

to your questions about the European Union

Freephone number (*):

00 800 6 7 8 9 10 11
(*) Certain mobile telephone operators do not allow access to 00 800 numbers or these calls may be billed.
A great deal of additional information on the European Union is available on the Internet.
It can be accessed through the Europa server (http://europa.eu).

Cataloguing data can be found at the end of this publication.

Luxembourg: Publications Office of the European Union, 2010

ISBN 978-92-79-14046-4

ISSN 1018-5593

doi: 10.2777/75697

© European Union, 2010

Reproduction is authorised provided the source is acknowledged.

Printed in Luxembourg

Printed on white chlorine-free paper

FINAL SUMMARY

Objectives of the project

The objectives of this project are the valorisation and dissemination of the knowledge and the results
which have been acquired within the related preceding RFCS research project “Competitive Steel and
Composite Bridges by Improved Steel Plated Structures – COMBRI” [COMBRI2007] with regard to
plate buckling verifications in order to promote and to encourage the wider use of steel plated structures
in bridges. To achieve this, a design manual is created particularly for practitioners. It consists of two
parts: in Part I “Application of Eurocode rules” [6] two composite bridge structures are covered on the
basis of worked examples incorporating the advantageous use of EBPlate software [12] developed in
the frame of the COMBRI research project. In Part II “State-of-the-Art and Conceptual Design of Steel
and Composite Bridges” [7] the current practice and possibilities as well as restrictions of current
Eurocode rules are presented. The aim of the design manual is not only to bring together the knowledge
and the practice from different European countries but also to give practitioners ideas for an out-of-the-
box thinking.
Another objective, and final conclusive step, is the implementation of the newly gained state-of-the-art
knowledge into standardisation via Nationally Determined Parameters (NDP), Non-Contradictory
Complementary Information (NCCI) and suggestions for the next revision of the Eurocodes to improve
the range of application and bring forward economic advantages. Then, practitioners will benefit in the
long run from this common European basis for the design of steel plated structures on an up-to-date
level.
Thus, the three major objectives of this project can be summarised as:
• Preparation of the COMBRI Design Manual
• Promotion and dissemination in seminars and workshops
• Formulation of proposals for implementation in standardisation

Comparison of initially planned activities and work accomplished

The work of this project closely followed the original work plan and the distribution of work according
to the different work packages was kept. During the coordination meetings minor adjustments to the
COMBRI Design Manual were thoroughly discussed and agreed on to better achieve the objectives of
the project.
Initially within WP 1 the design example of only one twin-girder composite bridge with a fully-detailed
calculation was planned. During the first coordination meeting the partners decided to focus especially
on the topics related to COMBRI research project, i.e. plate buckling verifications. For instance, the
study about buckling of stiffened plates in pure compression can be valorized only if a box-girder
bridge example is carried out. The level of detail of the different verifications to be performed has been
adjusted according to the main topics of the COMBRI research project. For instance, reasonable values
were chosen for other topics such as fatigue issues, justification of the reinforced concrete slab, weld
sizes, without detailed calculation. Finally, the partners decided to design two composite bridge
examples - a twin-girder and a box-girder bridge - instead of one as a basis for illustrating the practical
use of Eurocode rules and for allowing the quantification of the design options and new application
rules developed within COMBRI project.
During the third coordination meeting the partners agreed to change the originally planned title of WP 2
“Conceptual Design of Steel Bridges and Composite Bridges” into “State-of-the-art and Conceptual
Design of Steel and Composite Bridges” because a comparison with current practice supports better the
highlighting of possibilities and restrictions of current Eurocode rules. Thus, the contents of Part II of
the manual have been enhanced with examples on current practice in several European countries and
common bridge types as well as unusual bridges for special purposes or development projects in an
introductory section.
3
Description of activities and discussion
During the project runtime the necessary close cooperation between all partners was ensured by four
coordination meetings. The intermediate results were reported regularly in the technical reports. More
than 70 internal working documents were prepared as contributions to the COMBRI design manual and
to the proposals for an implementation in standardisation.
The preparation of the COMBRI Design Manual was accomplished hand in hand between all partners
and it allowed taking into account the national experience and knowledge in bridge design. The 18-
month runtime which has been applied for in the beginning was absolutely necessary and paid off in the
proper preparation of an English reference document and translation of the manual into French, German
and Spanish.
Mainly the national partners were responsible for the organisation and marketing of their national
dissemination activities at the end of 2008. In several seminars and workshops the design manual and
EBPlate software were disseminated and received a positive feedback throughout. Compliments were
received especially for the highlighting of the possibilities and restrictions of current Eurocode rules
and the overview on best national practices in the different European countries. The presentation of
EBPlate software was highly appreciated and suggestions for future development were given.
Based on the results from the COMBRI research project, the preparation of the manual and discussions
at seminars and workshops, proposals were formulated finally for an implementation of the project
developments into standardisation.

Conclusions
This section summarises in short the conclusions which have been drawn from the COMBRI Design
Manual and the proposals for implementation of the results into standardisation.
An overview of the bridge types in the participating partner’s countries - Belgium, France, Germany,
Spain and Sweden – show the current practice in those countries. It can be stated that 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 practice. Thus, the solutions presented are intended to serve as inspiration for the conceptual design
of new bridges.
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. It is shown that in most cases such high grades are not feasible. The
problem is usually that the fatigue requirements limit the full utilisation 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 large span box-girder from Part I of the manual was redesigned from S355 to a hybrid
girder with S460 and S690 and it turned out that the costs of the material was reduced by 10% in the
spans and 25% at the piers. In addition, there is a reduction of the fabrication costs.
Flanges as bottom flanges in box-girders are in most cases stiffened and different types of stiffeners are
discussed. It is shown that 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 EBPlate software 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 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.
With regard to webs the focus was 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. Besides that, longitudinal stiffeners on webs increase the resistance for bending

4
as well as for shear. The economy of using longitudinal stiffeners was 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 use
discontinuous stiffeners and another is to put the transverse and the longitudinal stiffeners on opposite
sides of the web.
Cross bracings and diaphragms for I-girder bridges and box-girders have the function 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. Although it is not much material used for
cross bracings, from an economical point of view it is important to minimize the man hours for
fabrication. This was 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.
The technique of launching bridges is very popular and verification methods have been improved in the
COMBRI research project. They allow now 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 the 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.
The proposals which have been formulated for an implementation into standardisation were classified
according to Nationally Determined Parameters (NDP), Non-Contradictory Complementary
Information (NCCI) and amendments to be used in the next revision of Eurocode in order to provide a
concise background and proposal scheme. The transfer of these proposals to CEN is ensured through
the collaborative work of the project members in that committee.

Exploitation and impact of the research results

The proper exploitation and impact of the COMBRI research project is the main objective and actual
origin of this valorisation and dissemination project in order to reach a wider audience and to ensure the
application of the outcomes of the COMBRI research project.
Within this project, the COMBRI Design Manual has been created, Part I and Part II being a 277-page
and 121-page document which is available in English, French, German and Spanish. In Part I
“Application of Eurocode rules” two composite bridge structures are covered on the basis of worked
examples and the advantageous use of EBPlate software developed in the frame of the COMBRI
research project is shown. In Part II “State-of-the-Art and Conceptual Design of Steel and Composite
Bridges” the current practice and possibilities as well as restrictions of current Eurocode rules are
presented. Thus, the design manual including worked examples, state-of-the-art and conceptual design
issues contributes to a better understanding and assessment of the knowledge gained along with the
possibility to highlight relevant topics.
The effective and wide dissemination of the design manual and its contents including EBPlate software
was carried out by the organisation of seminars and workshops in each participating partner’s country.
The dissemination activities took place according to Table 3-1. The feedback from the seminars and
workshops was positively throughout. Compliments were received especially for the highlighting of the
possibilities and restrictions of current Eurocode rules and the overview on best national practices in the
different European countries. The presentation of EBPlate software was highly appreciated and
suggestions for future development were given.
Within this project, proposals for an implementation of the results into standardisation via Nationally
Determined Parameters (NDP), Non-Contradictory Complementary Information (NCCI) and
suggestions for the next revision of Eurocode were also formulated to ensure the transfer of knowledge
and results in the long run so that a common basis is created from which practitioners can benefit from
in the future.

5
TABLE OF CONTENTS
Page

1  Introduction .................................................................................................................................... 11 

2  COMBRI Design Manual .............................................................................................................. 11 
2.1  General ................................................................................................................................ 11 
2.2  Part I.................................................................................................................................... 12 
2.3  Part II .................................................................................................................................. 13 
3  Dissemination Activities ................................................................................................................ 14 
3.1  General ................................................................................................................................ 14 
3.2  Belgium............................................................................................................................... 14 
3.3  France ................................................................................................................................. 16 
3.4  Germany ............................................................................................................................. 17 
3.5  Spain ................................................................................................................................... 18 
3.6  Sweden ................................................................................................................................ 19 
4  Implementation for Standardisation ............................................................................................... 21 
4.1  General ................................................................................................................................ 21 
4.2  Proposals for Nationally Determined Parameters ............................................................... 21 
4.2.1  Patch loading resistance of longitudinally stiffened girders .................................. 21 
4.2.2  Patch loading resistance according to Chapter 10 ................................................. 24 
4.3  Proposals for Non-Contradictory Complementary Information ......................................... 25 
4.3.1  Ultimate resistance of longitudinally stiffened plates under uniform
compression ........................................................................................................... 25 
4.3.2  Effective area of stiffened plates ........................................................................... 30 
4.3.3  Shear resistance of unstiffened and longitudinally stiffened girders ..................... 32 
4.3.4  Interaction between shear- and patch loading ....................................................... 35 
4.4  Proposals for revision ......................................................................................................... 37 
4.4.1  Ultimate resistance of longitudinally stiffened plates under uniform
compression ........................................................................................................... 37 
4.4.2  Effective area of stiffened plates ........................................................................... 40 
4.4.3  Shear resistance of unstiffened and longitudinally stiffened girders ..................... 44 
4.4.4  Patch loading resistance according to Chapter 6 ................................................... 44 
4.5  Conclusion .......................................................................................................................... 50 
5  Final Remarks ................................................................................................................................ 50 

Note: This report is accompanied by and refers to the COMBRI Design Manual Part I and Part II
in English, French, German and Spanish language.

6
LIST OF FIGURES
Figure 2-1: Eurocodes to be used in a composite bridge design ............................................................. 12 
Figure 3-1: Pictures of the workshop in Namur, Belgium ...................................................................... 15 
Figure 3-2: Promotional tools of the French workshop........................................................................... 17 
Figure 3-3: Participants of the workshop in Paris, France ...................................................................... 17 
Figure 3-4: Pictures of the workshop in Düsseldorf, Germany ............................................................... 18 
Figure 3-5: Pictures of the dissemination activities organised in Spain .................................................. 19 
Figure 3-6: Participants of the workshop in Stockholm, Sweden ........................................................... 20 
Figure 3-7: Participants at course for bridge engineers in Borlänge, Sweden......................................... 20 
Figure 4-1: Contradictory variation of the design resistance FRk according to EN 1993-1-5 and to Finite
Element calculations [8] .................................................................................................................. 22 
Figure 4-2: Buckling mode in the directly loaded subpanel .................................................................... 23 
Figure 4-3: Comparison of the current interpolation function according to EN 1993-1-5 and the required
interpolation function according to Seitz [26] ................................................................................. 24 
Figure 4-4: Weighting factor ξ in function of the aspect ratio α for pure compression (ψ = 1); Curve
parameter: Sum of the relative stiffness γ of the longitudinal stiffeners ......................................... 28 
Figure 4-5: Geometries of the plates considered in the numerical simulations ....................................... 30 
Figure 4-6: Comparisons between numerical simulations and EN 1993-1-5 formulas for the resistance
of plates stiffened by one stiffener – σcr,p calculated from Annex A ............................................... 31 
Figure 4-7: Comparisons between numerical simulations and EN 1993-1-5 formulas for the resistance
of plates stiffened by one stiffener – σcr,p calculated from EBPlate ................................................ 31 
Figure 4-8: Examples of FEM models at ULS (1 stiffener + Rigid end-post; 2 stiffeners + non-rigid
end-post) .......................................................................................................................................... 33 
Figure 4-9: Ultimate loads, comparison between application of EN1993-1-5 and FEM results (left: one
stiffener, right: two stiffeners) ......................................................................................................... 34 
Figure 4-10: Ultimate loads, comparison between application of amended EN1993-1-5 and FEM results
(left: one stiffener, right: two stiffeners) ......................................................................................... 34 
Figure 4-11: Subdivision into basic load cases ....................................................................................... 35 
Figure 4-12: Experimental and numerical results compared to proposed interaction equation based on
different resistance models (FR,unstiffened: Gozzi [19], FR,stiffened: Davaine [8] and VR: EN1993-1-5
[13]). ................................................................................................................................................ 36 
Figure 4-13: Shear force diagrams for construction stages during bridge launching. ............................. 36 
Figure 4.14: Stress distribution for elastic critical buckling and over-critical buckling ......................... 38 
Figure 4.15: Maximum possible difference Δρc for an imperfection factor α0 ≥ 0,34 ............................ 38 
Figure 4.16: Maximum possible deviation between both approaches for 0 ≤ ξ ≤ 1 ................................ 39 
Figure 4-17: Influence of the relative bending stiffness of the stiffener γ on the resistance of a plate
under compression ........................................................................................................................... 41 
Figure 4-18: Influence of the stiffness of a stiffener on the buckling of the plate .................................. 41 
Figure 4-19: Calibration of the proposal for a plate stiffened by one stiffener σcr,p calculated with Annex
A ...................................................................................................................................................... 42 

7
Figure 4-20: Calibration of the proposal for a plate stiffened by one stiffener σcr,p calculated with
EBPlate ............................................................................................................................................ 43 
Figure 4-21: Calibration of the simplified proposal for a plate stiffened by one stiffener σcr,p calculated
with Annex A .................................................................................................................................. 43 
Figure 4-22: Load-displacement curve (left) and vertical stress distribution in the web plate (right) for
FE-analysis. Material properties corresponding to a S355 steel. .................................................... 48 
Figure 4-23: Proposed reduction factor for patch loading resistance compared with test results with
small bending moment (less than 0,4 times resistance)................................................................... 49 
Figure 4-24: Test results (left) and computer simulations (right) for patch loading of girders with
longitudinal stiffeners related to the proposed characteristic resistance ......................................... 49 

List of tables
Table 3-1: Schedule of the dissemination activities ................................................................................ 15 
Table 4-1: Summary of the geometric configurations ............................................................................. 34 

List of references
[1] Boissonnade, N.; Degée, H.: Background Document to Work Package 1.2 – Numerical and code
aspects. Background document COMBRI-Report-ULIEGE-001, RFCS research project RFS-CR-
03018, 2006
[2] Boissonnade, N.; Degée, H.; Oppe M.; Naumes J.: Experimental and numerical investigations for
I-girders in bending and shear stiffened by trapezoidal stiffeners. Int. J. of Advanced Steel
Construction, Vol. 4 n°1, pp. 1-13, 2008
[3] Braun, B.: Combined Shear- and Patch Loading - Test Report. Background document COMBRI-
Report-LTU-002, RFCS research project RFS-CR-03018, 2006
[4] Clarin, M.: Plate Buckling Resistance - Patch loading of longitudinally stiffened webs and Local
buckling. Doctoral thesis 2007:31, Division of Steel Structures, Luleå University of Technology,
Sweden, 2007
[5] COMBRI: Competitive Steel and Composite Bridges by Improved Steel Plated Structures. Final
Report, RFCS research project RFS-CR-03018, 2007
[6] COMBRI+: COMBRI Design Manual - Part I: Application of Eurocode rules. RFCS project
RFS2-CT-2007-00031, 2008
[7] COMBRI+: COMBRI Design Manual - Part II: State-of-the-Art and Conceptual Design of Steel
and Composite Bridges. RFCS project RFS2-CT-2007-00031, 2008
[8] Davaine, L.: Formulation de la résistance au lancement d’une âme métallique de pont raidie
longitudinalement. Doctoral thesis D05-05, Institut National des Sciences Appliquées, Rennes,
France, 2005
[9] Degée, H.; Boissonnade, N.; Maquoi, R.: Numerical investigations on web panels stiffened by one
or two trapezoidal stiffeners. Proceedings of the 5th International Conference on Coupled
Instabilities in Metal Structures, pp. 324-332, 2008
[10] DIN 18800-Teil 3: Stahlbauten – Stabilitätsfälle, Plattenbeulen, German Code, DIN Deutsches
Institute für Normung e.V., November, 1990
[11] DIN-Fachbericht 103: Stahlbrücken. 3. Ausgabe, 2009
[12] EBPlate: A piece of software developed in the frame of the COMBRI research project, see [5]. 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
[13] EN 1993-1-5: Eurocode 3: Design of steel structures – Part 1-5: Plated structural elements,
October 2006

8
[14] EN 1993-2: Eurocode 3: Design of steel structures – Part 2: Steel Bridges, October 2006
[15] EN 1994-2: Eurocode 4: Design of composite steel and concrete structures – Part 2: General rules
and rules for bridges, October 2005
[16] European Recommendations for the Design of Longitudinally Stiffened Webs and of Stiffened
Compression Flanges, 1st Edition, 1990
[17] FineLG: Non-linear finite element analysis program “FINELg” (2002), User’s Manual, version
8.5, ArGEnCo Department, University of Liège, Greisch Info S.A., Liège, Belgium
[18] Galéa, Y.; Martin, P-O.: Plate type behaviour – Use of EBPlate to calculate σcr,p. COMBRI-
Report-CTICM-065, RFCS research project RFS-CR-03018, 2006
[19] Gozzi, J.: Patch Loading Resistance of Plated Girders - Ultimate and serviceability limit state.
Doctoral thesis 2007:30, Division of Steel Structures, Luleå University of Technology, Sweden,
2007
[20] Graciano, C.: Patch Loading – Resistance of Longitudinally Stiffened Steel Girder Webs.
Doctoral Thesis 2002:18, Division of Steel Structures, Luleå University of Technology, Sweden,
2002
[21] Kuhlmann, U.; Braun, B.: Combined Shear- and Patch Loading - Numerical Studies and
Development of an Interaction Equation. Background document COMBRI-Report-USTUTT-002,
RFCS research project RFS-CR-03018, 2006
[22] Lagerqvist, O.: Patch Loading – Resistance of steel girders subjected to concentrated forces,
Doctoral Thesis 1994:159D, Division of Steel Structures, Luleå University of Technology,
Sweden, 1994
[23] Martin, P-O.; Galéa, Y.: Combine use of EN 1993-1-5 and EBPlate – Effective area of stiffened
plates. COMBRI+-Report-CTICM-029, RFCS research project RFS2-CT-2007-00031, 2007
[24] Martin, P-O.; Galéa, Y.: Validation of EBPlate Effective width of stiffeners - COMBRI-Report-
CTICM-064, RFCS research project RFS-CR-03018, 2006
[25] Müller, C.: Zum Nachweis ebener Tragwerke aus Stahl gegen seitliches Ausweichen. Doctoral
Thesis, Lehrstuhl für Stahlbau, RWTH Aachen, Heft 47, 2003
[26] Seitz, M.: Tragverhalten längsversteifter Blechträger unter quergerichteter Krafteinleitung.
Doctoral Thesis, Institute of Structural Design, Universität Stuttgart, Mitteilungen des Instituts für
Konstruktion und Entwurf Nr. 2005-2, 2005

9
1 INTRODUCTION
The objectives of this project are the valorisation and dissemination of the knowledge and the results
which have been acquired within the related preceding RFCS research project “Competitive Steel and
Composite Bridges by Improved Steel Plated Structures – COMBRI” [5] with regard to plate buckling
verifications in order to promote and to encourage the wider use of steel plated structures in bridges.
Work has been performed on the following topics and is presented as follows:
• Preparation of the COMBRI Design Manual, see Chapter 2
• Promotion and dissemination in seminars and workshops, see Chapter 3
• Formulation of proposals for implementation in standardisation, see Chapter 4

2 COMBRI DESIGN MANUAL

2.1 General
The COMBRI Design Manual is an outcome of the research project RFS-CR-03018 “Competitive Steel
and Composite Bridges by Improved Steel Plated Structures - COMBRI” [5] and it has been prepared
in the frame of this successive dissemination project. The manual is available in English, French,
German and Spanish. The essential knowledge which has been acquired within the RFCS research
project to enhance the competitiveness of steel and composite bridges has been incorporated in the
COMBRI Design Manual. It was promoted and disseminated in the frame of several seminars and
workshops, see Chapter 3. The manual is subdivided into two parts to provide the reader with clearly
arranged and concise documents:
► Part I: Application of Eurocode rules, see Section 2.2
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 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, see Section 2.3
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 difficulties and some solutions being part of development projects are presented in Part II
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 [12] developed in the research project to determine the
elastic critical plate buckling stresses are presented in its contributive application for bridge design.
The authors of the COMBRI Design Manual are:
Ulrike Kuhlmann, Benjamin Braun
Universität Stuttgart, Institute for Structural Design / Institut für Konstruktion und Entwurf
Markus Feldmann, Johannes Naumes
RWTH Aachen University, Institute for Steel Structures

11
 Pierre-Olivier Martin, Yvan Galéa
Centre Technique Industriel de la Construction Métallique
Bernt Johansson, Peter Collin, Jörgen Eriksen
Luleå University of Technology, Division of Steel Structures
Hervé Degée, Nicolas Hausoul
Université de Liège, ArGEnCo Département
José Chica, Sandra Meno
Fundación LABEIN
Joël Raoul, Laurence Davaine, Aude Petel
Service d'études sur les transports, les routes et leurs aménagements

2.2 Part I
Part I of the COMBRI Design Manual [6] deals with the calculation of two composite bridges
according to the Eurocodes.
In the COMBRI research project [5] the different national background of each partner how to apply and
interprete Eurocode rules was brought together and a lot of background information and general
knowledge has been created. In order to facilitate the implementation of Eurocodes EN 1993-1-5
“Plated Structural Elements” [13], EN 1993-2 “Steel bridges” [14] and EN 1994-2 “Composite bridges”
[15] with regard to plate buckling verifications, it was decided to cover two steel-concrete composite
bridges - a twin-girder and a box-girder bridge - in order to present the knowledge with the help of
worked examples and in a very descriptive manner. As the examples focus in detail on the application
and interpretation of Eurocode rules which are related to plate buckling verifications, the overall view
on bridge design cannot be covered. In this context, Figure 2-1 shows how many standards can be
involved in the design of a composite bridge.
Part 1 - 11

- 10

re
ctu
s

fra
t1
Cable

-9
tle

Pa
Par

rt
it
Br

1
rt

-1 ue
Pa

Ge t ig -8
ner
al r Fa rt
1
ul e
s- Pa ts
c on
cre Join 5
te rt 1-
es St Pa
ri dg ee lates
eb EN lb ned p
ret - 2
19
rid
ge Stiffe
nc 92 93 s - 1
Co 19 - 2 Part 1
EN
EN 1994 - 2 General rules - steel

Part 1 - 1 Composite bridges

ctions EN
nent a 90 n
Perma Ac 91 1 9 es i g
tio 91
ns E N of d
is
1-4 Ba
s
Part An
Ap ne
d xA
i on

Win 5 ns
p li
ca 2
1-
ut

tio tio
ac
ec

rt n
s

Pa al fo
ex

rb
ctio

m
er -6
g

rid
Th
rin

1 g
7

rt
ta l a

es
du

Pa
1-
ns

2
iden
rt
tio

Pa r t

Traffic
Pa
Ac

Acc

Figure 2-1: Eurocodes to be used in a composite bridge design

In some parts, this design manual introduces general assumptions e.g. on actions without aiming to
present the theoretical background or the modelling in detail. In addition to that, it is assumed that the
reader is familiar with general design and modelling issues of bridges because this design manual gives
a detailed view on plate buckling topics but it does not cover all other topics related to the verification
of the design. For further information to the aforementioned topics, the reader is referred to relevant
literature.
In the document, the worked examples are presented in a double-sided layout with comments,
background information and interpretation issues on the left-hand side and the example calculation on
12
the right. All relevant references to current Eurocode rules are provided. As mentioned above, the
examples are a twin-girder and a box-girder bridge which allows looking at a design without and with
longitudinal stiffeners.
In chapter 2 the deck of the twin-girder and the box-girder bridge is described and the global analysis of
both bridges is introduced. For this purpose, an overview on the bridge geometry, material distribution
and construction sequences is given firstly. Secondly, a general section follows in which common data
such as material properties and actions as well as combinations thereof are given. Last but not least, the
global analysis is presented for both bridges and the relevant results - internal forces and moments - are
summarised for the verifications. Based on that, chapters 3 and 4 look at the verifications during the
final stage and the execution stage. Here, each chapter is subdivided into a part dealing with the
verifications of the twin-girder bridge or the box-girder bridge.

2.3 Part II
Part II of the COMBRI Design Manual [7] deals with the state-of-the-art in several countries and
addresses improvements that can be provided to the design of steel and composite bridges. It is focused
on the conceptual design of steel bridges and the steel parts of composite bridges. Design of steel
bridges is a very wide field which can not be covered completely in the manual so that a selection of
topics has been made. A short summary of the content and conclusions is given hereafter:
In chapter 2 of the manual an overview of bridge types in the countries participating in the project is
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 difficulties 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 of the manual the choice of steel grades is discussed. 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 large span box-girder from Part I of the
COMBRI Design Manual is redesigned from S355 to a hybrid girder with S460 and S690 and it turns
out that the cost of the material is reduced by 10% in the spans and 25% at the piers. In addition, there
will be a reduction of the fabrication cost as well.
Flanges are dealt with in chapter 4 of the manual and the main topic is 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 [12] 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 of the manual 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
13
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 use discontinuous stiffeners and another is to put the transverse and the longitudinal
stiffeners on opposite sides of the web.
Chapter 6 of the manual 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 is 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 detail in the COMBRI research project and it is dealt with in chapter 7 of
the manual. 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 the COMBRI 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.

3 DISSEMINATION ACTIVITIES

3.1 General
The effective and wide dissemination of the design manual and its contents including EBPlate software
was carried out by the organisation of seminars and workshops in each participating partner’s country.
The dissemination activities took place according to Table 3-1. A short report about the national
seminars and workshops can be found in Sections 3.2 to 3.6. The feedback from the seminars and
workshops was positively throughout. The presentation of EBPlate software was highly appreciated
and suggestions for future development were given.

3.2 Belgium
For the dissemination activity in Belgium, The University of Liège (ULg) obtained the support of the
“Centre Information Acier” (CIA-Infosteel) for the promotion and the organization of the seminar. The
one-day workshop was held centrally (Walloon region) at New Hotel de Lives near Namur. The
workshop was promoted by e-mail to the subscribers of the CIA and an announcement homepage on
CIA website: http://www.infosteel.be. The background of the participants stretched out over
administration offices, researchers, engineering offices and construction companies.

14
Table 3-1: Schedule of the dissemination activities
Country Date Place Number of
Participants

Belgium 30/10/2008 New Hôtel de Lives, Namur 24

France 11/12/2008 Centre de conférence BSA, Paris 55

Germany 07/11/2008 Stahl-Zentrum, Deutscher Stahlbau-Verband (DSTV), 48

Düsseldorf

Spain 30/06/2008 Sede de la Fundación LABEIN, Derio - Bizkaia 24

22/09/2008 Universidad de Burgos, Escuela Politécnica Superior 7
Ingeniería de Caminos Canales y Puertos, Burgos
21/10/2008 Plataforma Tecnológica Española del Acero, PLATEA, 32
sede de AITEMIN en Porriño - Vigo
28/10/2008 Colegio de Ingenieros de Caminos, Canales y Puertos, 144
Madrid
21/11/2008 Universidad de Burgos. Escuela Politécnica Superior 174
Ingeniería de Caminaos Canales y Puertos, Burgos

Sweden 16/10/2008 Kungliga Tekniska Högskolan, Stockholm 35

To begin the seminar, Jo Naessens (CIA) introduced the program of the day and the different activities
which CIA is in charge (as organisation of others RFCS project). First, Professor René Maquoi at the
University of Liège (ULg) spoke about “The objectives of Eurocodes”. Then, some general principles
on the conception on concrete-steel composite bridge were introduced by Hervé Degée (from ULg).
After that, two design examples were introduced during the seminar by Hervé Degée and Nicolas
Hausoul (from ULg): one concerned a composite steel-concrete twin-girder bridge and another one
concerned a composite steel-concrete box-girder bridge. A design manual including the 2 design
examples were distributed to the participants. To end the seminar, Professor René Maquoi presented the
software EBPlate with some application examples. During the all seminar, some aspects on Eurocode
was explained and discussed with the audience.

a) Presentation of one design example b) Participants

Figure 3-1: Pictures of the workshop in Namur, Belgium

15
 c) Hervé Degée d) Participants

Figure 3-1 (continued): Pictures of the workshop in Namur, Belgium

3.3 France
For the dissemination in France, the French seminar took place in Paris on the 11th of December 2008.
It had been previously promoted by a printed flyer, announcements in several issues of the CTICM
bulletin “Construction Métallique Informations”, which is distributed at about 7500 copies, a dedicated
webpage on the CTICM website (www.cticm.com) and a general emailing to French engineers involved
in steel construction (see Figure 3-2). Finally 55 participants attended the meeting, coming from
engineering schools, universities, engineering offices and steel construction companies. Each
participant received a printed copy of the two COMBRI+ Guides and a USB key containing the pdf
files of the Guides and the EBPlate software.
The seminar was divided into two parts. The morning session was dedicated to the state-of-the-art and
to the standard environment for the construction of steel and composite bridges. A panorama of the
most common types of bridges and constructional details used in France has been drawn, also
presenting some original solutions as adopted in several European countries. The normative context has
been described, showing the progressive introduction of Eurocodes in French standards. The problem of
plate buckling and its specific treatment in EN 1993-1-5 has been looked at, and finally the COMBRI
project and its outcomes were presented.
The afternoon session dealt with some specific issues in the design of steel and composite bridges and
with the solutions recommended by the COMBRI project. It began by a presentation of the EBPlate
software. It was followed by a talk about the use of high strength steel in the bridge construction. An
example of application illustrated the economical interest of hybrid girders. The three last lectures were
dedicated to important steps in the design of a bridge: the design of the bottom flange of box girders,
the launching during the erection and the resistance to lateral torsional buckling of the structure.
The speakers of the seminar were French engineers involved in the COMBRI+ project (L. Davaine, A.
Petel, J. Raoul and P-O. Martin) and two invited personalities (Dr. Eng. D. Bitar, head of the Bridges
Unit – CTICM and J.M. Vigo from ConstruirAcier, a French organization for the promotion of steel).
The seminar ended by fruitful exchanges between participants and speakers. This was the opportunity to
promote several tools available to help designers in the use of the Eurocodes, including the COMBRI+
Guides. An investigation showed that participants were very satisfied of this seminar and of the
outcomes of the COMBRI project (Guides, software…).

16
 a/ flyer
b/ emailing c/ Guides

Figure 3-2: Promotional tools of the French workshop

Figure 3-3: Participants of the workshop in Paris, France

3.4 Germany
For the dissemination activity in Germany, Universität Stuttgart obtained the support of Deutscher
Stahlbau-Verband (DSTV), namely Fachgemeinschaft Brückenbau, and the one-day workshop was
held centrally at the German Steel Centre in Düsseldorf. The workshop was promoted by mailing list,
printed flyer, an announcement homepage, several important websites in Germany and announcements
in the German journals “Stahlbau” and “Bauingenieur”. Eventually it could attract not only participants
from Germany but also from Austria. Amongst the participants, students, researchers, engineers from
engineering offices and construction companies could be found.
For the workshop presentations the support of Ministerialrat Dipl.-Ing Joachim Naumann of the
German Federal Ministry of Transport, Building and Urban Affairs was gained. He spoke about
“Eurocodes and their implementation in Germany”. Also Dr. Eng. Laurence Davaine of the Service
d'études sur les transports, les routes et leurs aménagements, Paris, was invited and reported about the
“French practice in steel and composite bridge design”.
In the end, compliments were received especially for the highlighting of the possibilities and restrictions
of current Eurocode rules and the overview on best national practices in the different European
countries.

17
a) Greetings of host Dr.-Ing. Volkmar Bergmann b) Invited speaker Dipl.-Ing. Joachim Naumann

c) Invited speaker Dr. Eng. Laurence Davaine d) Participants

Figure 3-4: Pictures of the workshop in Düsseldorf, Germany

3.5 Spain
LABEIN - Tecnalia has been supported by the Spanish Independent Steel Promotion Organisation,
Asociación para la Promoción Técnica del Acero (APTA), in addition, one of the workshops was
organised in cooperation with the Spanish Steel Technology Platform, Plataforma Tecnológica
Española del Acero (PLATEA) and one of the seminars was organized in cooperation with the
Technical University of Burgos.
All the dissemination activities were promoted between professionals related to civil works and related
to the steel construction. The contacts were done via e-mail and by Internet through the homepage of
APTA, PLATEA and LABEIN-Tecnalia.
The dissemination events organized by LABEIN-Tecnalia can be classified in two ways:
a. Workshops to discuss with practicing professionals about the contents developed in
COMBRI and disseminated through COMBRI+:
i. Competitiveness of the composite bridges attending to project, fabrication and
construction stages. Derio -Bizkaia.
ii. Implementation of the Eurocodes for bridge design. Burgos.
iii. R&D to increase the competitiveness of the steel construction sector for
bridges. Vigo.
b. Seminars to disseminate the outcomes of COMBRI and the Design Manual COMBRI:

18
 i. Dissemination to practicing civil engineers including lectures performed by
acknowledged Spanish civil engineers to demonstrate the competitiveness of
the composite bridges. Madrid.
ii. Dissemination to the academy and students of civil engineering including
lectures performed by acknowledged Spanish civil engineers to demonstrate
the competitiveness of the composite bridges. Burgos.
As main conclusion, the participants acknowledged the Design Manual COMBRI as a very useful tool
for the efficient implementation of the Eurocodes related to the bridge design.

a) Seminar organised at the Colegio de b) Lecture by Julio Martínez Calzón at the

Ingenieros de Caminos Canales y Puertos in seminar in Madrid. Other keylectures were
Madrid presented by Javier Manterola, Francisco
Millanes, José Romo and Guillermo Capellán

c) Seminar organised at the University of Burgos d) Workshop organised in cooperation with

PLATEA in Vigo

Figure 3-5: Pictures of the dissemination activities organised in Spain

3.6 Sweden
The dissemination of the COMBRI results in Sweden took place at the Royal Institute of Technology in
Stockholm. It was originally planned as a half day seminar but it was decided to make it a full day by
adding more general information on steel bridges. The seminar was arranged by professor Peter Collin,
Ramböll and LTU in cooperation with the Swedish Steel Construction Institute which used its large
address register to promote the seminar. The Design Manuals were given to the participants free of
charge.
A summary of the two bridge examples from the Part I of the Design Manual was presented by Mr.
Jörgen Eriksen from LTU. The content of Part II of the design manual was covered in three different
presentations. Professor Peter Collin talked about recent bridges in Sweden and professor Milan
19
Veljkovic presented what we can learn from other countries. Professor emeritus Bernt Johansson talked
about conceptual design of composite bridges with possibilities and restrictions given by the Eurocodes.
Mr. Anders Spåls, Ruukki gave a presentation on fabrication and erection of steel girders for composite
bridges with focus on limiting expensive details. After that followed a presentation by Dr. Lars
Pettersson, Skanska on the use of composite bridges from a general contractor’s view with focus on the
competiveness of the composite concept. Finally, Peter Collin informed about some ongoing research
projects on bridges including integral abutments using steel piles and full depth concrete elements for
bridge decks.

Figure 3-6: Participants of the workshop in Stockholm, Sweden

An additional promotion of the Combri results was undertaken at a course on Eurocodes for bridge
engineers held 21 October, 2008, in Borlänge, Sweden. The 30 participants were bridge engineers from
the National Road Administration and the National Rail Administration. The Design Manual was
handed out to the participants.

Figure 3-7: Participants at course for bridge engineers in Borlänge, Sweden

20
4 IMPLEMENTATION FOR STANDARDISATION

4.1 General
In the COMBRI research project [5] several aspects concerning an improvement of EN 1993-1-5 [13]
have been addressed. In this project, these topics have been prepared for an implementation for
standardisation. All topics are structured in such a manner that the affected section is addressed and a
point of discussion is given. After that, a proposal is formulated which is supported by a section with
background information. In most cases, this background information is available through the COMBRI
final report [5]and its background documents. However, in some cases new results became available
and this information has been included in this document. In the following the topics are summarized
based on their classification as Nationally Determined Parameter (NDP), Non-Contradictory
Complementary Information (NCCI) and amendments for the next revision of Eurocode.
The following topic is intended to be addressed as Nationally Determined Parameter (NDP):
• Patch loading resistance of longitudinally stiffened girders, see Section 4.2.1
An overview of the topics addressed in terms of Non-Contradictory Complementary Information
(NCCI) is given below:
• Ultimate resistance of longitudinally stiffened plates under uniform compression, see Section
4.3.1
• Effective area of stiffened plates, see Section 4.3.2
• Shear resistance of unstiffened and longitudinally stiffened girder, see Section 4.3.3
• Interaction between shear- and patch loading, see Section 4.3.4
Furthermore the following topics are also foreseen as amendment within a future revision:
• Ultimate resistance of longitudinally stiffened plates under uniform compression, see Section
4.4.1
• Effective area of stiffened plates, see Section 4.4.2
• Shear resistance of unstiffened and longitudinally stiffened girder, see Section 4.4.3
• Patch loading resistance of unstiffened girders, see Section 4.4.4

NOTE: References written in normal type in the boxed sections - e.g. equation numbers and section
numbers - refer to current EN 1993-1-5 [13]. However, references written in italic type in the boxed
sections correspond to a new numbering and they are used in the explanatory text.

4.2 Proposals for Nationally Determined Parameters

4.2.1 Patch loading resistance of longitudinally stiffened girders

Affected section
Section 6.4(2), EN 1993-1-5 [13]
"χ−λ approach"

21
Point of discussion
Section 6.4(2) of EN 1993-1-5 [13] deals with the calculation of the critical load Fcr for a longitudinally
stiffened web through a note (leaving choice to the National Annex) and with the reduction factor χF for
the patch loading resistance.
To be efficient for the patch loading resistance, the longitudinal stiffener should be as close as possible
to the loaded flange (the EN 1993-1-5 scope imposes b1/hw U 0.3). In the design of a launched bridge,
the typical location is b1/hw = 0.25. The failure mode is then a local buckling in the subpanel closest to
the loaded flange. However the kF formula in EN 1993-1-5 corresponds to the first buckling mode of
the web which is a buckle in the deepest web subpanel. Consequently if the longitudinal stiffener moves
away from the loaded flange, the depth of the non directly loaded web subpanel decreases, the critical
load according to EN 1993-1-5 increases and finally the design resistance increases. This is
contradictory to the fact that the web is less stiffened for patch loading (the subpanel closest to the
loaded flange is deeper). In [8] this contradictory behaviour has also been highlighted through Finite
Elements calculations by Davaine [8], see Figure 4-1.

Note:
Results for a 8-m-long and 3-m-deep
web panel with a single flat longitudinal
stiffener located at a distance h1 from the
patch load

Figure 4-1: Contradictory variation of the design resistance FRk according to EN 1993-1-5 and
to Finite Element calculations [8]

To solve this contradiction, a new way for calculating the critical load of a longitudinally stiffened web
submitted to a patch load has been proposed in [8]. The same study has also shown that the use of the
coefficient m2 for the plastic resistance Fy was not justified in comparison to FE simulations. This point
about Fy has been further studied and confirmed by Gozzi in [19].

If a change is introduced in one of the steps of the χ−λ approach (here the calculation of Fcr and of Fy ),
then a new reduction factor χF should be derived to ensure the safety level of the model as a whole. This
has been done in [8] by using the format proposed in Annex B of EN 1993-1-5.

Proposal
To 6.4 (2) Note

For webs with longitudinal stiffeners, the critical load Fcr may be determined by using the following
procedure :
ss + 2 tf
0 ,6 + 0 ,5
⎛ s + 2tf ⎞⎛ a ⎞ a
kF ,2 = ⎜ 0, 8 s + 0, 6 ⎟ ⎜ ⎟
⎝ a ⎠ ⎝ b1 ⎠

π 2E tw 3
Fcr ,2 = kF ,2
12 (1 − ν 2 ) b1

22
 Fcr ,1Fcr ,2
Fcr =
Fcr ,1 + Fcr ,2

where Fcr ,1 is given by Equation (6.5) using the buckling coefficient kF

according to Equations (6.6) and (6.7).
The reduction function χ F should be determined from the following equation :

χF =
1
(
≤ 1, 0 with ϕF = 0, 5 1 + 0, 21( λ F − 0, 8 ) + λ F )
ϕ F + ϕF 2 − λ F

To 6.5
For determination of the yield load Fy, the effective loaded length ℓy should be determined according
to Equation (6.10) with m2 = 0.

Background information
The interpolation between Fcr ,1 and Fcr ,2 proposed in [8] relies on the idea to superimpose the first
buckling mode as proposed in EN 1993-1-5 (i.e. a buckle in the deepest not directly loaded web
subpanel) and the buckling mode in the directly loaded web subpanel as shown in Figure 4-2. This
interpolation solves the contradictory behaviour observed in Figure 4-1.
ss
ss + 2tf

b1 45°

hw ss + 2tf + 2b1

π2E tw3
Fcr,2 = k F,2
12 (1 − ν 2 ) b1
a

Figure 4-2: Buckling mode in the directly loaded subpanel

An extensive database of numerical simulations can be found in [8] for scanning bridge dimensions and
completing the experimental database. These data have been used to calibrate the coefficients α F = 0,21
and λ 0 = 0,8 in the reduction factor. The format of this reduction factor results from a transition
between the plastic design of the web without buckling ( χ F = 1 ) and a web plate buckling ( χ F = 1 / λ
according to Von Karman hypothesis). By analogy with the European buckling curves, the reduction
function is then the solution of the following equation: (1 − χ F ) (1 − χ F λ ) = α F ( λ − λ 0 ) χ F where the
coefficients α F and λ 0 should be calibrated with the experimental and numerical databases.
In the subsequent studies at LTU by Gozzi and Clarin, the interpolation between Fcr ,1 and Fcr ,2 has been
improved by using min ( Fcr ,1; Fcr ,2 ) and the suppression of the coefficient m2 in the yield resistance has
been confirmed. This new change in calculating Fcr needs then a new calibration of the coefficients α F
and λ 0 . The new proposal is detailed in Section 4.4.4.

23
4.2.2 Patch loading resistance according to Chapter 10
Affected section
Section 10(5), EN 1993-1-5 [13]:

Point of discussion
In section (10.5) the reduction factor ρz for transverse patch stresses taking into account plate-like
behaviour may be determined according to section 4.5.4 which corresponds to the reduction factor ρx
for longitudinal stresses. However, in [26] it was shown that in case of column-like behavior the
interpolation function used for the determination of the reduction factor ρz overestimates the resistance,
see Figure 4-3. Thus, a transferability of the interpolation function from longitudinal stresses is not
given.

1.0
f_int (EN 1993-1-5)

f_int (Seitz)
4.5.4, EN 1993-1-5 [13]
0.8
Discrepancy Proposal Seitz [26]

0.6
f [-]

0.4

0.2

0.0
1.0 1.5 2.0 2.5 3.0 3.5
σcr,p/σcr,c [-]

Figure 4-3: Comparison of the current interpolation function according to EN 1993-1-5 and
the required interpolation function according to Seitz [26]

Proposal
To 10(5)
The reduction factor ρ may be determined using either of the following methods:
a) the minimum value of the following reduction factors:
[…]
ρz for transverse stresses from Annex B.1(3) taking into account column-like behavior
where relevant;
[…]

Background information
The discrepancy of the interpolation function can be solved either by defining a new interpolation
function or by choosing an appropriate reduction curve. During the revision of the German DIN-
Fachbericht 103 [11] it was decided to adopt the reduction curve for transverse stresses according to
Annex B, EN 1993-1-5.

24
4.3 Proposals for Non-Contradictory Complementary Information

4.3.1 Ultimate resistance of longitudinally stiffened plates under uniform

compression
Affected section
Section 4.5.4(1) of EN 1993-1-5 [13]:
“Interaction between plate and column buckling”

Point of discussion
The determination of the weighting factor ξ is an extensive calculation procedure as it requires the use
of several different sections of EN 1993-1-5 [13], namely Section 4.5.3, Annex A.1, Annex A.2.1 and
Annex A.2.2. Formulae for the direct calculation of the weighting factor ξ should be provided.
Particularly in combination with the proposal for revision given in section 4.4.1 of this report, the direct
calculation of ξ would lead to a strong simplification and consolidation of the calculation procedure.

Proposal
Two paragraphs are added to Section 4.5.4 of EN 1993-1-5 [13] to allow the direct determination of the
weighting factor ξ. The proposal for modification of Section 4.5.4 is as follows:

To 4.5.4
(1) The final reduction factor ρc should be obtained by interpolation between χc and ρ as follows:
ρ c = (ρ − χ c ) ξ (2 − ξ ) + χ c (4.13)

σ cr , p
where ξ= − 1 but 0 ≤ ξ ≤ 1
σ cr ,c

σcr,p is the elastic critical plate buckling stress, see Annex A.1(2) and A.2.2(1),
respectively;
σcr,c is the elastic critical column buckling stress according to 4.5.3(2) and (3),
respectively;
χc is the reduction factor due to column buckling.
ρ is the reduction factor due to plate buckling, see 4.4(1).
(2) For longitudinal stiffened plates the weighting factor ξ may be obtained directly from one of
the equations given in 4.5.4(3) and 4.5.4(4).
(3) For orthotropic plates with at least three stiffeners the weighting factor ξ may be obtained
from
bsl ,1 1 + δ
ξ = kσ , p ⋅ α 2 ⋅ ⋅ −1 for ψ ≠ 1 (4.14)
bc 1 + γ

and

ξ=
(1 + α ) 2 2
+γ
−1 for ψ = 1 (4.15)
1+ γ

25
 but 0 ≤ξ ≤1

Note: All parameters kσ,p, α, bsl ,1 , bc, δ and γ are specified in Annex A.1 and Figure A.1,
respectively.
(4) For plates with one or two stiffeners the weighting factor ξ may be obtained from

bsl ,1 Asl ,1 ⋅ t 2
ξ = kσ , p ⋅ α 2 ⋅ ⋅ − 1 , where kσ,p is known from relevant computer
bc I
sl ,1 ⋅ 12 ⋅ (1 −ν ) 2

simulations (4.16)
or

a4 ⋅ b ⋅ t3
ξ= , acc. to Annex A.2
( )
(4.17)
4 ⋅ π 4 1 − ν 2 ⋅ b12 ⋅ b22 ⋅ I sl ,1

but 0 ≤ ξ ≤1

Note: The parameters kσ,p, α, a, b, b1, b2, t, ν, Asl ,1 and I sl ,1 are specified in Annex A.2. The
geometric values bsl ,1 and bc, from the stress distribution are specified in Figure A.1.

Background information
The added formulae for the direct determination of the weighting factor ξ are derived by application of
Section 4.5.3, Annex A.1, Annex A.2.1 and Annex A.2.2, thus they are non-contradictory to EN 1993-
1-5 [13].
The explanation of the used symbols can be taken from EN1993-1-5 [13], if not specified hereafter. In
detail the derivations of the formulae are as follows:
Equation (4.14) and (4.15):

The elastic critical buckling stress of a plate is given by σ cr . p = kσ , p ⋅ σ e . The values for the buckling
coefficient kσ , p can be taken from buckling charts, relevant computer simulations (e.g. with EBPlate)
and Annex A.1, respectively. The Euler-stress σ e is given by

π 2 ⋅ E ⋅t2
σe =
(
12 ⋅ b 2 ⋅ 1 −ν 2 )
The elastic critical column buckling stress of the orthotropic plate can be derived by an equivalent
Euler-column with the geometric properties of the stiffener and its adjacent parts:
π 2 ⋅ E ⋅ I sl
σ c , sl = .
a2 ⋅ A
With
b ⋅t3
I sl = ⋅ (1 + γ ) ,
(
12 ⋅ 1 −ν 2 )
( )
A = A p + ∑ Asl = b ⋅ t ⋅ (1 + δ ) ,

a = α ⋅b
the elastic critical column buckling stress reads

26
 ⎛ σe ⎞ 1+ γ
σ cr , sl = ⎜ ⎟⋅ .
⎝α ⎠ 1+ δ
2

For the final comparison with the elastic critical plate buckling stress σ cr. p , the critical column
buckling stress σ cr.sl has to be extrapolated to the plate edge according to the stress distribution, cp.
NOTE of section 4.5.3(3), EN 1993-1-5 [13]. Hence the elastic critical column buckling stress at the
plate edge with the maximum compression is given by
bc
σ cr ,c = σ cr , sl ⋅
bsl ,1

which can be rewritten to

bc ⎛ σe ⎞ 1+ γ
σ cr ,c = ⎜ 2 ⎟⋅ .
bsl ,1 ⎝α ⎠ 1+ δ
Thus the resulting weighting factor ξ to obtain the final reduction factor ρ c reads:

bsl ,1 1 + δ
ξ = kσ , p ⋅ α 2 ⋅ ⋅ − 1 , but 0 ≤ ξ ≤ 1
bc 1 + γ

For uniform compression (ψ = 1 ) the buckling coefficient kσ , p is given by

kσ , p =
(1 + α ) 2 2
+γ
, see also NOTE 2
α ⋅ (1 + δ )
2

and the equation for the resulting weighting factor ξ can be simplified to

ξ=
(1 + α ) 2 2
+γ
− 1 , for (ψ = 1 )
1+ γ

NOTE 1: To get save sided results, the orthotropic plate parameters δ and γ for the determination of
the weighting factor ξ should be calculated with the same number of stiffeners, which has been used for
the calculation of kσ , p . Computer programs and buckling charts usually increase the number of
( )
stiffeners by one nx = nstiffeners + 1 for the computation of the orthotropic plate, as it leads to an higher
consistency between calculations with discrete and smeared stiffeners. This has be taken into account
for the determination of ξ.

NOTE 2: Equation (A.2) given in Annex A, EN 1993-1-5 [13], is equal to kσ , p = (1 + α2 )

2 2
+ γ −1
α ⋅ (1 + δ )
Comparative calculations show a discrepancy between Equation (A.2) and numerical results from
EBPlate. For ψ = 1 this discrepancy can be eliminated completely by erasing the dissenting term “-1”.

27
 ξ
1.0

0.9 ψ=1
0.8

γ=1
0.7 10 20
50
0.6
100
0.5
200

0.4
500
0.3
1000
0.2
2000
0.1
5000
10000
0.0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 3.25 3.50 3.75 4.00 4.25 4.50

α
Figure 4-4: Weighting factor ξ in function of the aspect ratio α for pure compression (ψ = 1);
Curve parameter: Sum of the relative stiffness γ of the longitudinal stiffeners

Equation (4.16):

The elastic critical buckling stress of a plate is given by σ cr . p = kσ , p ⋅ σ e . The values for the buckling
coefficient kσ , p can taken from buckling charts and relevant computer simulations (e.g. with EBPlate),
respectively. The Euler-stress σ e can whether be determined by

π 2 ⋅ E ⋅t2
σe =
(
12 ⋅ b 2 ⋅ 1 −ν 2 )
or directly be taken from the used computer program.
The elastic critical column buckling stress of the stiffener close to the edge with the highest
compression is given by:
π 2 ⋅ E ⋅ I sl,1
σ cr .sl ,1 =
a 2 ⋅ Asl ,1

where I sl ,1 is the second moment of area of the gross-section of the stiffener and the adjacent parts
of the plate, relative to the out-of-plane bending of the plate
Asl,1 is the cross-sectional area of the stiffener and the adjacent parts of the plate according to
Figure A.1, EN 1993-1-5 [13]
NOTE: For plates under bending the width of the adjacent parts has to be determined according to
Annex A.2.1.(2) , EN 1993-1-5 [13].
With a = α ⋅ b the elastic critical column buckling stress can be written as
π 2 ⋅ E ⋅ I sl,1
σ cr .sl ,1 =
α 2 ⋅ b 2 ⋅ Asl ,1

28
For the final comparison with the elastic critical plate buckling stress σ cr. p , the critical column
buckling stress σ cr .sl ,1 has to be extrapolated to the plate edge according to the stress distribution, cp.
NOTE of section 4.5.3(3), EN 1993-1-5 [13].
With the given equations
bc
σ cr ,c = σ cr , sl ,1 ⋅
bsl ,1

and
σ cr , p
ξ= −1
σ cr ,c

the weighting factor ξ to obtain the final reduction factor ρ c can be derived to:

bsl,1 Asl ,1 ⋅ t 2
ξ = kσ , p ⋅ α 2 ⋅ ⋅ − 1 , where 0 ≤ ξ ≤ 1
bc I (
sl ,1 ⋅ 12 ⋅ 1 −ν
2
)
Equation (4.17):
In case of a stiffened plate with one longitudinal stiffener located in the compression zone, the elastic
critical buckling stress σ cr, p can be calculated by a fictitious isolated strut supported on an elastic
foundation according to Annex A.2.2, whereas the stiffeners in the tension zone are ignored. For plates
with a length of a ≥ a c the weighting factor results to ξ = 1 , thus the decisive equation for the elastic
global plate buckling stress σ cr, p for a plate with one stiffener is given by

bsl ,1 bsl ,1 ⎛⎜ π 2 ⋅ EI sl ,1 E ⋅ t3 ⋅ b ⋅ a2 ⎞
⎟
σ cr , p = ⋅ σ cr , sl = ⋅⎜ +
bc bc ⎜ Asl ,1 ⋅ a 2
⎝ ( )
4 ⋅ π 4 1 − ν 2 ⋅ Asl ,1 ⋅ b12 ⋅ b22
⎟⎟
⎠
with the elastic critical column buckling stress according to Section 4.5.3(3), EN 1993-1-5 [13]
bsl ,1 bsl ,1 π 2 ⋅ EI sl ,1
σ cr ,c = ⋅ σ cr , sl = ⋅
bc bc Asl ,1 ⋅ a 2

and
σ cr , p
ξ= −1
σ cr ,c

the weighting factor ξ to obtain the final reduction factor ρ c can be derived to:

a4 ⋅ b ⋅t3
ξ= , where 0 ≤ ξ ≤ 1
(
4 ⋅ π 4 1 −ν 2 ) ⋅ b12 ⋅ b22 ⋅ I sl ,1

In case of a plate with two longitudinal stiffeners located in the compression zone the elastic critical
plate buckling stress should be taken as the lowest of those computed for the three cases b1 = b1∗ ,
b2 = b2∗ and b = B ∗ , cp. Figure A.3, EN 1993-1-5 [13].
CAUTION: Expression “ σ cr, sl ” is used twice in the Eurocode with different meaning: On the one
hand for the elastic critical column buckling stress according to Section 4.5.3(3), EN 1993-1-5 [13] and
on the other hand for the elastic critical plate buckling stress according to Annex A.2.2, EN 1993-1-5
[13].

29
4.3.2 Effective area of stiffened plates
Section 4.5, EN 1993-1-5 [13]
“Calculation of the effectivep area of stiffened plates with longitudinal stiffeners under direct stresses”

Point of discussion
Numerical simulations have shown that the equation (4.5) of EN 1993-1-5 may lead to over-evaluated
effectivep areas and so to unsafe results for plates with weak stiffeners. According to these studies, the
results become unsafe when the relative bending stiffness γ of the stiffeners is less than 25.

Proposal
To 4.5
Stiffened plates having weak longitudinal stiffeners should be considered as unstiffened plates
regarding their resistance to direct stresses and their effectivep area should be calculated according to
section 4.4 of EN 1993-1-5. Longitudinal stiffeners should be considered as weak stiffeners if their
relative bending stiffness γ is less than 25, where γ is defined by:
E Is
γ=
bD
where E is the Young’s modulus
b is the width of the plate
E t3
is the bending stiffness of the plate, defined by D =
D
(
12 1 − υ 2 )
t is the web thickness
ν is the Poisson’s coefficient
Is is the second moment of area of the stiffener for out-of-plane bending, its cross-section
including a participating width of web of 10 t each side of each stiffener-to-web
junction.

Background information
Numerical investigations have been carried out in the frame of WP1 of the COMBRI research project
[5], from which an important database has been constituted. Four sets of rectangular plates have been
simulated, with two different uniform thicknesses, several lengths and several values of the relative
bending stiffness γ of the stiffeners (see Figure 4-5).

thickness t Set 0 - unstiffened Set 1 - one stiffener

b

a Set 2 - two stiffeners Set 3 - four stiffeners

Figure 4-5: Geometries of the plates considered in the numerical simulations

For each case, the final resistance of the plate under pure normal stresses had been determined by
performing a non–linear Finite Elements analysis taking into account global and local imperfections

30
according to Annex C of EN 1993-1-5. This resistance has been compared to the resistance calculated
according to chapter 4.5 of EN 1993-1-5, more specifically with eq. (4.5).
The comparison shows that the EN 1993-1-5 method may be unsafe (see Figure 4-6) for weak
stiffeners. The threshold value of the relative bending stiffness γ of the stiffeners has been evaluated to
about 25: for γ > 25, the EN 1993-1-5 is always safe (except very short plates), and for γ < 25 it is
generally unsafe [23].
This outcome has been verified in the frame of two different ways for assessing σcr,p (cf. 4.5.2 of EN
1993-1-5): either Annex A of EN 1993-1-5 (see Figure 4-6) or the method proposed in the COMBRI
project [18] using EBPlate software [12] (see Figure 4-7).
It is to be noted that the calculation of the γ parameter requires taking into account a participating width
of the plate, equal to 10 t each side of the stiffener-to-web junction (see [24]).
An analysis of the Equation 4-5 of EN 1993-1-5 has been carried out [23] and it has been shown that
the model leading to this formulation considers in fact rigid stiffeners. This formulation gives correct
results for the resistance where this assumption is verified, but may lead to unsafe results for weak
stiffeners. See Section 4.4.2 for a proposal for the improvement of eq. (4.5).
1,4 Rk,FEM EN + Annex A
Rk,EN
1,2

1,0

γ=5
0,8
γ = 10
γ = 25
0,6
γ = 50
γ = 100
0,4
t = 20 mm
t = 8 mm
0,2

database ranked by γ values and increasing lengths

0,0

Figure 4-6: Comparisons between numerical simulations and EN 1993-1-5 formulas for the
resistance of plates stiffened by one stiffener – σcr,p calculated from Annex A

1,4
Rk,FEM EN + EBPlate
Rk,EN
1,2

1,0

0,8
γ=5
γ = 10
0,6 γ = 25
γ = 50
0,4 γ = 100

t = 20 mm
0,2 t = 8 mm

database ranked by γ values and increasing lengths

0,0

Figure 4-7: Comparisons between numerical simulations and EN 1993-1-5 formulas for the
resistance of plates stiffened by one stiffener – σcr,p calculated from EBPlate

31
4.3.3 Shear resistance of unstiffened and longitudinally stiffened girders
Affected section
Section 5.3(2) to 5.3(4), EN 1993-1-5 [EC3-1-5]

Point of discussion
The proposal deals with the resistance of web panels subjected to shear combined or not with bending
and stiffened by one or two longitudinal closed stiffeners. Indeed, though EC3-1-5 is one of the most
advanced standards regarding plate buckling, the beneficial effect of closed section longitudinal
stiffeners is not explicitly addressed, even if those have a significant positive impact on the stability of
sub-panels and on the actual rigidity of transverse stiffeners compared to a design using open section
stiffeners.
Experimental tests and numerical simulations performed during the COMBRI project and in
complementary studies have shown that the ultimate load provided by tests or simulations is
significantly higher than predicted by the codes as soon as shear is concerned (i.e. if the resistance is
governed either by shear or by bending-shear interaction).

Proposal
In order to take into account the beneficial effects of closed stiffeners, the following proposals for Non
Contradictory Complementary Information are made:

To 5.3(2)
For webs stiffened by closed-section longitudinal stiffeners connected to end posts and vertical
stiffeners, end post may always be considered as rigid.
To 5.3(3)
For the assessment of the plate critical buckling stress τcr of unstiffened webs or of webs stiffened by
open section longitudinal stiffeners, simply supported conditions should always be assumed, while
for webs stiffened by closed section longitudinal stiffeners, the torsional restraint brought by the
flanges and transverse stiffeners may be accounted for.
To 5.3(4)
Due to the high torsional stiffness of closed stiffeners, the reduction of the second moment of area of
the stiffeners to 1/3 of their actual value is not required for the calculation of the shear buckling
coefficient kτ.

Background information
This proposal relies on extensive parametric studies based on FE simulations of web-panels with one or
two longitudinal closed section stiffeners subjected to bending and shear. Comparisons with application
of Eurocode 3 Part 1.5 have been performed, leading to the code improvements proposed above. These
proposals are non contradictory since they are dealing exclusively with closed section stiffeners that are
not specifically addressed in Eurocode 3.
This study is an extension of investigations performed previously within the ComBri Project [1,2]. FE
calculations have been conducted with the software FINELg [17] using shell finite elements and
accounting for geometric and material non-linearities as well as initial imperfections. The modelled
stiffened panel is composed of two flanges, web, transverse and longitudinal stiffeners. The latter have
a trapezoidal cross-sectional shape. Configurations of transverse stiffeners that should be considered
either as rigid or non-rigid end-posts are also defined (see Figure 4-8). The constitutive law of the steel
material is taken elastic-plastic with strain-hardening. The influence of residual stresses is disregarded
since it is known to have only a slight influence on the ultimate strength when shear governs. The panel
is supposed to be half of a three-point bending girder situation and the support conditions and loading
conditions are adopted accordingly.

32
Figure 4-8: Examples of FEM models at ULS (1 stiffener + Rigid end-post; 2 stiffeners + non-
rigid end-post)

The different geometric configurations of the parametric study are summarized in Table 4-1. All results
provided by the numerical simulations have been compared to those resulting from the application of
EN1993-1-5 specifications. In accordance with the latter, the following resistance checks must be
performed: bending resistance at mid-span of the member with a gross elastic cross-section (“gross”),
bending resistance at a distance of min (0.4 a; 0.5 b) from mid-span (“bending”), shear buckling
resistance (“shear”), possible bending-shear interaction at a distance of 0.5 hw from mid-span
(“interaction”) and resistance of the transverse stiffeners (“transverse”). As far as Eurocode results are
of concern, the values reported in the following correspond to the minimum of those relative to above-
listed criteria.
Results obtained for plates respectively with one or two stiffeners are reported in Figure 4-9. It can be
observed that as soon as shear is concerned (i.e. when the governing EN1993 check is either "shear" or
"interaction"), the value of the ultimate load provided by the FEM simulation is significantly higher
than the value predicted by the code. This had already been evidenced in [1,2] for what regards
configurations with one single stiffener and is confirmed in [9] for configurations with 2 stiffeners, with
an even more important scattering for the "interaction" criterion.
In order to improve the results obtained with the normative approach, the 3 modifications of the
procedure already proposed in [1] for the 1-stiffener configuration have also been applied to the 2-
stiffeners configuration in [9]. These modifications can be summarized as follows:
- Using a dedicated software such as EBPlate [12] for the determination of the elastic critical
stress σcr,pl corresponding to the so-called “plate behaviour”. In contrast with EN1993-1-5, where
critical stresses are supposed to be assessed with the assumption of simply supported edges, this
software can account for restraining effects provided by both flanges and transverse stiffeners. While
the assumption of simple supports is not much disputed for unstiffened web panels or panels
stiffened by open cross-section stiffeners, it is found unduly conservative for panels stiffened by
closed stiffeners.
- Adopting the actual stiffener’s inertia for calculating the shear buckling stress of the stiffened
panel, instead of 1/3rd of its value as recommended in EN1993-1-5. This allows accounting for
the higher torsional stiffness of the closed trapezoidal stiffeners, since the reduction to 1/3rd proposed
by EN1993-1-5 is actually based on investigations on panels with open stiffeners. The proposal is
supported by the comparison of FEM results obtained in the presently summarized study [1,9]
considering open or close stiffeners.
- Using the upper shear resistance curve for web contribution (factor χw), normally dedicated to
rigid end-posts, even in the case of non-rigid end-posts, provided that the longitudinal
stiffeners are welded to the transverse stiffeners. This is justified by the fact that both webs of a
usual closed section stiffener stabilize the end-post over a non negligible part of its depth, in such a
way that it behaves as rigid.

33
Figure 4-10 presents a comparison between FEM results and normative assessments once the 3
amendments are applied. The proposed amendments result in a much better quality of the results
especially in terms of accuracy. One may also note that the proposed modifications for kτ are leading to
higher failure loads associated with 'interaction' and 'shear' criteria and are thus responsible for the
different distributions of failure modes observed when comparing Figure 4-9 and Figure 4-10 in
particular for the 2-stiffeners configuration. Indeed, some cases that were assumed to be governed by
shear or shear-bending interaction appear finally as governed either by bending or by the resistance of
the transverse stiffener.

Table 4-1: Summary of the geometric configurations

One stiffener – 1st set One stiffener – 2nd set Two stiffeners
hw [mm] 1000 and 2000 2000, 3000 and 4000 4000
tw [mm] 6 and 12 20 20, 30 and 40
α 1, 3 and 5 1, 1.5 and 2 1, 2 and 4
Stiffener geometry 2 (weak + strong) 2 (weak + strong) 1 (intermediate)
0.25 hw + 0.5 hw
Location of 0.3 hw 0.3 hw
0.125 hw + 0.375 hw
longitudinal stiffeners 0.5 hw 0.5 hw
1/3 hw + 1/3 hw
End post Rigid + non-rigid Rigid + non-rigid Rigid + non-rigid
Resulting number of
384 288 648
configurations

100 160

Shear Gross Interaction Gross Shear Trans.

140
80

120
60
100
FEM / EC3 [%]

FEM / EC3 [%]

40 80

20 60

40
0
0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 20

-20
0
0 10 20 30 40 50 60 70 80
-40 -20

N° test N° test

Figure 4-9: Ultimate loads, comparison between application of EN1993-1-5 and FEM results
(left: one stiffener, right: two stiffeners)

100 160

Shear Gross Interaction Gross Shear Transverse

140
80

120
60
FEM / Proposal [%]

FEM / Proposal [%]

100

40 80

20 60

40
0
0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 20

-20
0
0 10 20 30 40 50 60 70 80
-40 -20

N° test N° test

Figure 4-10: Ultimate loads, comparison between application of amended EN1993-1-5 and
FEM results (left: one stiffener, right: two stiffeners)

34
4.3.4 Interaction between shear- and patch loading
Affected section
Section 7, EN 1993-1-5 [13]
“Interaction”

Point of discussion
Currently no formulation for the interaction between transverse force and shear force is given in Section
7, EN 1993-1-5.

Proposal
To 7 Interaction between transverse force and shear force
If the girder is subjected to a concentrated transverse force acting on the compression flange in
conjunction with shear force, the resistance should be verified using 4.6, 5.5 and the following
interaction expression:
1.6
⎡ ⎛ FEd ⎞⎤
⎢η 3 ⋅ ⎜⎜1 − ⎟⎟⎥ + η2 ≤ 1 .0
⎣⎢ ⎝ 2 ⋅ VEd ⎠⎦⎥

Background information
Currently no formulation for the interaction between transverse force and shear force is given. All
studies on the interaction between shear and patch loading have in common that they subdivide the
combined loading into two basic load cases. Thus, the influence of shear stresses caused by the
transverse forces can be accounted for because it is already included in the patch loading model. Figure
4-11 shows the basic load cases “pure patch loading” and “pure shear force” which in combination
leads to the investigated type of interaction. This subdivision is also the basis on which the interaction
equation (7.3) is defined.

(i) pure patch loading (ii) pure shear loading

Figure 4-11: Subdivision into basic load cases

In the frame of the COMBRI research project, experimental and numerical studies on steel plated
girders have been conducted in order to review and to propose an interaction equation for combined
shear and patch loading. Based on the own test girders a Finite-Element model has been created and the
own experiments as well as tests from literature were recalculated successfully with the numerical
model. The numerical model was then used for parameter studies in which the most important
parameters with regard to bridge launching conditions were varied. The behaviour of longitudinally
stiffened girders, which may show global or local buckling, is found to be covered within the
interaction range of the unstiffened girder cases and a common interaction equation according to
Equation (7.3) has been developed,. The statistical evaluation of the proposed interaction equation has
been done with the shear resistance model according to Section 5, EN1993-1-5. For the patch load
resistance, several models exist which have been developed in order to improve the current formula of
Section 6, EN1993-1-5. Here, these models have been evaluated and it is shown that the interaction
equation is safe sided for these models not only for unstiffened but also for longitudinally stiffened
girders. Figure 4-12 shows this representatively for the improved patch loading resistance models for
unstiffened girders according to Gozzi [19] and for longitudinally stiffened girders according to

35
Davaine [8]. Another approach for longitudinally stiffened girders has been proposed by Clarin [4]. In
all cases the applicability of the current resistance models according to EN1993-1-5 and for improved
resistance models is proven.

unstiffened
stiffened (global failure)
stiffened (local failure)
Equation (7.3)

relevant for bridge

launching

Figure 4-12: Experimental and numerical results compared to proposed interaction equation
based on different resistance models (FR,unstiffened: Gozzi [19], FR,stiffened: Davaine [8] and VR:
EN1993-1-5 [13]).

a) Stage “cantilever”

b) Stage “pier reached”

Figure 4-13: Shear force diagrams for construction stages during bridge launching.

From the investigations, it is obvious that the interaction between shear and patch loading is not
negligible. Although the interaction equation and diagram appears severely at first sight, for bridge
launching the specific conditions have to be taken into account. For this reason, Figure 4-12 shows two
relevant construction stages: a) when the bridge girder is about to arrive at the support and a cantilever
36
exists; b) when the bridge girder has reached the pier. In stage a) the introduced patch load is almost
equally equilibrated resulting in a pure patch loading situation where the shear is already considered in
the patch load model. In stage b) the maximum internal shear force approximates the value of the
applied patch load which leads to an asymmetric patch loading condition. For this situation the
interaction becomes relevant, the average will result in reductions of around 10%, see Figure 4-12.
However, for the verification of the cross-section the interaction load case “patch loading and bending
moment” is decisive.Further background information is given in COMBRI documents [3], [5] and [21].

4.4 Proposals for revision

4.4.1 Ultimate resistance of longitudinally stiffened plates under uniform

compression
Affected section
Section 4.5.3(4) and 4.5.3(5), EN 1993-1-5 [13]
“Column type buckling behaviour”

Point of discussion
The use of two different slendernesses λ p and λc to determine the reduction factor for plate type
behaviour ρ and the reduction factor for column type behaviour χ c leads to a labour-intensive and non
consistent calculation procedure. It is furthermore contradictory to the procedure given in section 10 of
EN 1993-1-5 [13], where only one slenderness λ p is used for the determination of all reduction factors,
cp. Section 10(5) EN 1993-1-5 [13]. For more details see Background information below.

Proposal

Section 4.5.3(4) EN 1993-1-5 [13] is cancelled.

Section 4.5.3(5) EN 1993-1-5 [13] is modified in the following way:

(5) The reduction factor χ c should be obtained from 6.3.1.2 of EN 1993-1-1. The provided
slenderness λ is calculated from equation (4.7). For unstiffened plates α = 0,21 corresponding to
buckling curve a should be used. For stiffened plates its value should be increased to:
0,09
αe = α + (4.12)
ie

I sl ,1
with i =
Asl ,1

e = max (e1, e2) is the largest distance from the respective centroids of the plating and the
one-sided stiffener (or of the centroids of either set of stiffeners when present on both
sides) to the neutral axis of the effective column, see Figure A.1;
α = 0,34 (curve b) for closed section stiffeners;
= 0,49 (curve c) for open section stiffeners.

37
Background information
The ultimate resistance of a plate under compression is not necessarily limited to its elastic critical
buckling stress σcr, as it is the case for bar-like structures. Due to stress redistributions at the unloaded
edges, plates are able to exceed the elastic critical buckling stress limit up to an over-critical value σx,
cp. Figure 4.14.

Figure 4.14: Stress distribution for elastic critical buckling and over-critical buckling

The magnitude of this over-critical reserve depends on the capability of the plate to redistribute the load
to the supported edges and its initial imperfection. Very wide or strongly stiffened plates are only able
to redistribute a minor part of the load. These plates react more column-like and the reduction factor has
to be determined via an interpolation between Winter and column-buckling curve.
To determine whether a plate reacts more column or plate-like Section 4 of EN1993-1-5 [13] uses an
established procedure, which compares the existing elastic critical buckling stress σcr,p of the plate to
those of an equivalent column (σcr,c). The closer both values are together, the more column-like is the
behaviour of the plate and the column buckling curve has to be applied. Different to other calculation
rules (e.g. in Section 10 of EN1993-1-5 [13] and DIN 18800 Teil 3 [10]) Section 4 does not only use
the critical buckling stress σcr,c of the equivalent column to determine the weighting factor ξ for the
interpolation between the two limits ρ and χc but also for the determination of the lower limit χc itself.
Thus the interpolation does not take place between the two extreme values ρ (λ p ) and χ (λ p ) of the
real plate, which is the case for the direct method (cp. green line in Figure 4.15), but between one value
ρ (λc ) determined with the real plate properties and one value χ (λc ) determined with a not existing
fictitious column (red broken line).
χ, ρ 1.2

1.1
1.0
0.9
0.8
0.7
ρ(λp)
0.6
ρ
0.5
max. Δρc
0.4
χc(λc)
0.3 χ(λp)
χc
0.2
0.1 ⎯λp ⎯λc
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
⎯λ

Figure 4.15: Maximum possible difference Δρc for an imperfection factor α0 ≥ 0,34

38
This procedure is not only less consistent and contradictory to Section 10, but makes the calculation
procedure even more laborious. Furthermore it impairs the efficiency of other improvements as e.g. the
ones proposed in section 4.3.1 of this report.
Beside these facts it has to be mentioned that the total effect on the finally determined reduction factor
for column-type behaviour ρc is smaller than one might expect. The maximum possible difference
between the reduction factors ρc determined in the two different ways, described above, takes place for
a plate buckling slenderness of λ p = 1,475 in combination with an elastic critical column buckling
stress of σ cr ,c = 0,8 ⋅ σ cr , p . In this case the difference Δρ c between the both reduction factors is about
7,36 %. This situation is given in Figure 4.15. In any other case the difference in result between both
calculation procedures is smaller. For those cases where the weighting factor ξ is outside its limits
0 ≤ ξ ≤ 1 , the difference is 0,0 %.
Figure 4.16 displays the magnitude of the possible differences between both approaches, whereas for
each plate buckling slenderness λ p every possible combination of 1 ≤ σ cr , p σ cr ,c ≤ 2 has been
checked. Considering the same occurrence probability for all possible combinations ( λ p ; σ cr , p σ cr ,c ) a
mean deviation between both approaches of ≤ 2,62 % can be determined. With increasing imperfection
factor α0 the difference between both approaches is even getting smaller. A typical imperfection factor
for single sided stiffeners is about α0 ≈ 0,6.
10% 1 − ρc(χc)/ρc(χp) α0 = 0.34
10% 1 − ρc(χc)/ρc(χp) α0 = 0.49
9% 9%

8% max. possible deviation 8%

max. possible deviation
7% 7%

6% 6%
standard deviation
5% standard deviation
5%

4% mean 4% mean
total mean total mean
3% 3%

2% 2%

1% 1%
⎯λ ⎯λ
0% 0%
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

10% 1 − ρc(χc)/ρc(χp) α0 = 0.60

10% 1 − ρc(χc)/ρc(χp) α0 = 0.70
9% 9%

8% 8%

7% max. possible deviation 7% max. possible deviation

6% 6%

5% standard deviation 5% standard deviation

4% mean 4%
mean
total mean total mean
3% 3%

2% 2%

1% 1%
⎯λ ⎯λ
0% 0%
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

Figure 4.16: Maximum possible deviation between both approaches for 0 ≤ ξ ≤ 1

Considering all these aspects the clear conclusion has to be drawn that only one slenderness λ should
be used to make EN 1993-1-5 more consistent and user-friendly.
Particularly with regard to the NCCIs proposed in section 4.3.1 of this report, the use of one single
slenderness λ p instead of two different ones will lead to a strong simplification and consolidation of
the given design procedure.

39
4.4.2 Effective area of stiffened plates
Affected section
Section 4.5(4), EN 1993-1-5 [13]
“Calculation of the effectivep area of stiffened plates with longitudinal stiffeners under direct stresses”

Point of discussion
Numerical simulations have shown that the equation (4.5) of EN 1993-1-5 may lead to over-evaluated
effectivep areas and so to unsafe results for plates with weak stiffeners. According to these studies, the
results become unsafe when the relative bending stiffness γ of the stiffeners is less than 25.

Proposal

To Equation (4.5)

Ac,eff = ρc Ac,eff,loc + Σ ρedge bedge t (4.5)

where: ρedge is the reduction factor for the edge subpanel, calculated from equation (4.2) with the
relative slenderness λ p,edge defined as:

fy
λ p ,edge =
min(σ cr ,p ; σ cr ,sp )
σcr,p is the elastic critical buckling stress for the global buckling of the plate (see 4.5.2)
σcr,sp is the elastic critical buckling stress of the edge subpanel considered as isolated.

When σcr,p is calculated with Annex A, ρedge can be calculated by the following modified relation:
ρedge = min(ρedge,loc ; ρ)
where: ρedge,loc is the reduction factor for the edge subpanel calculated according to 4.4 (2)
ρ is the reduction factor for the global buckling of the plate, calculated according to 4.5 (2).

Background information
In the frame of the COMBRI project, equation (4.5) of EN 1993-1-5 has been proved to be unsafe (see
Section 4.3.1) for plates stiffened by weak stiffeners (γ < 25).
By studying the effect of the relative bending stiffness within the EN 1993-1-5 model of resistance, it
has been shown that the model for stiffened plates is not consistent with the model for unstiffened plates
when the relative bending stiffness γ tends to zero. In this latter condition, the resistance of the plate
does not tend to the resistance of the unstiffened plate, as it would have to, but to a higher value (see
Figure 4-17).

40
 0,60 Nult / Npl

0,50 Plate with one stiffener,

acc. to EN 1993-1-5,
σcrp acc. to EBPlate

0,40

Plate with one stiffener,

0,30 acc. to EN 1993-1-5,
σcrp acc. to Annex A

0,20
Unstiffened plate
Inconsistency
of models
0,10

Log(γ L)
0,00
0,1 1 10 100 1000

Figure 4-17: Influence of the relative bending stiffness of the stiffener γ on the resistance of a
plate under compression

This over-evaluation may be related to the fact that the effective width attached to the edge of the plate
(bedge,eff in eq (4.5) of EN 1993-1-5) is calculated assuming that the stiffeners are sufficiently rigid so
that no influence of the global buckling behaviour of the plate is expected to affect this effective width.
This can be accepted as far as the first critical mode of the plate is the buckling of the sub panels (i.e.
fully rigid stiffeners) but it can be clearly criticized if the first critical mode is global (i.e. flexible
stiffeners). This latter case implies an influence of this global mode of the edge effective width, the
global critical stress being lower than the critical stress for the elastic buckling of the sub-panel, see
Figure 4-18.
σcr

σcr,p
Global buckling
of the plate

Buckling
of the sub-panels

σcr,nat
σcr,sp

Global buckling
of the plate

σcr,0
log(γ)
Flexible stiffener Rigid stiffener

γ*

Figure 4-18: Influence of the stiffness of a stiffener on the buckling of the plate

The idea to correct the inaccuracy of the model is then to calculate the effective width bedge,eff of an
edge subpanel by considering the actual critical stress for the buckling which reduces the resistance of
this subpanel (see Figure 4-18):
- either the stiffener is sufficiently rigid so that the first buckling mode of the plate is a local buckling
of the subpanel; bedge,eff is then calculated using σcr,sp the elastic critical stress of the subpanel (as it is
done in the present formulation of EN 1993-1-5);

41
- or the stiffener is not sufficiently rigid so that a global buckling of the plate occurs before the local
buckling of the subpanel. In this case, the global buckling will affect the resistance of the edge panel
and it is proposed to use σcr,nat = σcr,p, where σcr,p < σcr,sp.
Finally, the proposal can be summarized as follows:
fy
σcr,nat = min(σcr,p, σcr,sp) and then λ p,edge =
σ cr ,nat

ρedge is calculated from λ p,edge with equation (4.2) of EN 1993-1-5

bedge,eff = ρedge bedge
This proposal has been evaluated and calibrated using the database constituted in the frame of WP1 of
the COMBRI and described in Section 4.3.1 (see [23]). The results are clearly improved compared to
the ones given by the present formulation of equation (4.5) and nearly all of them are on the safe side
(see Figure 4-19 when using Annex A or Figure 4-20 when using EBPlate [18] for the calculation of
σcr,p). It is to be noted that the proposal does not change the results for plates with rigid stiffeners.

1,4 Rk,FEM Proposal + Annex A

Rk,EN
1,2

1,0

0,8 γ=5
γ = 10
0,6 γ = 25
γ = 50
γ = 100
0,4
t = 20 mm
t = 8 mm
0,2

database ranked by γ values and increasing lengths

0,0

Figure 4-19: Calibration of the proposal for a plate stiffened by one stiffener
σcr,p calculated with Annex A

42
 1,4
Rk,FEM Proposal + EBPlate
Rk,EN
1,2

1,0

0,8
γ=5
γ = 10
0,6 γ = 25
γ = 50
γ = 100
0,4
t = 20 mm
t = 8 mm
0,2

database ranked by γ values and increasing lengths

0,0

Figure 4-20: Calibration of the proposal for a plate stiffened by one stiffener
σcr,p calculated with EBPlate

When Annex A is used to assess σcr,p, the following modified relation could be used instead of the
previous one: ρedge = min(ρedge,loc; ρ), where ρedge,loc is the reduction factor for the edge subpanel
calculated according to 4.4 (2) and ρ the reduction factor for the global buckling of the plate calculated
according to 4.5.2. The difference between the two methods is in the coefficient βA,c used for the
reduced slenderness of the equivalent plate in the equation (4.7). See Figure 4-21 for the calibration of
this simplified proposal. This simplification cannot be used when σcr,p is not assessed with Annex A,
excluding so the method with EBPlate [18].
1,4 Simplified proposal + Annexe A
R k,FEM
R k,EN
1,2

1,0

0,8 γ=5
γ = 10

0,6 γ = 25
γ = 50
γ = 100
0,4
t = 20 m m
t = 8 mm
0,2

datab ase rank ed b y γ values and increasing lengths

0,0

Figure 4-21: Calibration of the simplified proposal for a plate stiffened by one stiffener
σcr,p calculated with Annex A

43
4.4.3 Shear resistance of unstiffened and longitudinally stiffened girders
Affected section
Section 5.3(2) to 5.3(4), EN 1993-1-5 [13]

Point of discussion
See Section 4.3.3 of the present document. The Non-Contradictory Complentary Information (NCCI)
proposed regarding web panels stiffened by closed stiffeners should preferably be included in the main
text of EN 1993-1-5 in a next revision.

Proposal
Section 4.3.3.

Background information
Section 4.3.3.

4.4.4 Patch loading resistance according to Chapter 6

Affected section
Section 6, EN 1993-1-5 [13]

Point of discussion
The present rules for patch loading resistance are safe but sometimes too safe-sided and uneconomical.
The rules use the same format as for other plate buckling phenomena but there are quite large
differences between the reduction functions. A step towards harmonization is desirable.

Proposal
It is suggested that the following text replaces the present Section 6 of EN 1993-1-5:2005.

6 Resistance to transverse forces

6.1 Basis
(1) The design resistance of the webs of rolled beams and welded girders should be determined in
accordance with 6.2, provided that the compression flange is adequately restrained in the lateral
direction and for case (c) in 6.1(2) both flanges.
(2) The load is applied as follows:
a) through the flange and resisted by shear forces in the web, see Figure 6.1(a);
b) through one flange and transferred through the web directly to the other flange, see
Figure 6.1(b);
c) through one flange adjacent to an unstiffened end, see Figure 6.1(c) where the flanges are
restrained from relative lateral movements.
(3) For box girders with inclined webs the resistance of both the web and flange should be checked.
The internal forces to be taken into account are the components of the external load in the plane
of the web and flange respectively.
(4) The interaction of the transverse force, bending moment and axial force should be verified using
7.2.

44
 Type (a) Type (b) Type (c)

FS FS FS

ss ss c ss
V1,S V2 ,S hw VS
a

2 2
⎛h ⎞ ⎛h ⎞ ⎛s + c⎞
kF = 6 + 2 ⎜ w ⎟ k F = 3,5 + 2 ⎜ w ⎟ k F = 2 + 6 ⎜⎜ s ⎟⎟ ≤ 6
⎝ a ⎠ ⎝ a ⎠ ⎝ hw ⎠
Figure 6.1: Buckling coefficients for different types of load application

6.2 Design resistance

(1) For unstiffened or stiffened webs the design resistance to local buckling under transverse forces
should be taken as

χ F f yw l y t w
FRd = (6.1)
γ M1

where tw is the thickness of the web;

fyw is the yield strength of the web;

ly is the effective loaded length, see 6.5, appropriate to the length of stiff bearing ss,
see 6.3;

χF is the reduction factor due to local buckling, see 6.4(1).

6.3 Length of stiff bearing

(1) The length of the stiff bearing ss on the flange should be taken as the distance over which the
applied load is effectively distributed at a slope of 1:1, see Figure 6.2. However, ss should not be
taken as larger than hw.
(2) If several concentrated forces are closely spaced, the resistance should be checked for each
individual force as well as for the total load with ss as the distance between the outer loads.

45° FS FS FS FS
FS

ss ss ss ss tf Ss = 0

Figure 6.2: Length of stiff bearing

(3) If the bearing surface of the applied load rests at an angle to the flange surface, see Figure 6.2, ss
should be taken as zero.

45
6.4 Reduction factor χF for buckling resistance
(1) The reduction factor χF should be obtained from:
1
χF = ≤ 1 .2 (6.2)
ϕF + ϕ 2F − λF

l y t w f yw
where λ F = (6.3)
Fcr

ϕF =
1
2
(
1 + 0,5 ( λF − 0, 6 ) + λF ) (6.4)

(2) For webs without longitudinal stiffeners Fcr should be obtained from Equation (6.5) with kF
according to Figure 6.1.

t w3
Fcr = 0,9 k F E (6.5)
hw
NOTE 1 The values of kF in Figure 6.1 are based on the assumption that the load is introduced
by a device that prevents rotation of the flange.

(3) For webs with longitudinal stiffeners and loading type (a) Fcr may be taken as the smallest of
(6.9) and (6.10) with kF,1 and kF,2 according to Equations (6.6) and (6.7)
2
⎡h ⎤ ⎡ b ⎤
k F,1 = 6 + 2 ⎢ w ⎥ + ⎢5,44 1 − 0,21⎥ γ s (6.6)
⎣ a ⎦ ⎣ a ⎦

⎛ s +2tf ⎞
0, 6⎜ s ⎟ + 0 ,5
⎡ ⎛ s + 2t f ⎞ ⎤ ⎛ a ⎞ ⎝ a ⎠
k F, 2 = ⎢0,8⎜ s ⎟ + 0,6⎥ ⋅ ⎜⎜ ⎟⎟ (6.7)
⎣ ⎝ a ⎠ ⎦ ⎝ b1 ⎠
3
I ⎡a⎤ ⎡ b⎤
γ s = 10,9 sl13 ≤ 13 ⎢ ⎥ + 210⎢0,3 − 1 ⎥ (6.8)
hwt w ⎣ hw ⎦ ⎣ a⎦

where b1 is the depth of the loaded subpanel taken as the clear distance between the loaded
flange and the stiffener;

I sl1 is the second moments of area of the stiffener closest to the loaded flange
including contributing parts of the web according to Figure 9.1

b1 b
Equation (6.6) is valid for 0,05 ≤ ≤ 0,3 and 1 ≤ 0,3
hw a

t 3w
Fcr ,1 = 0,9 k F,1 E (6.9)
hw

46
 t 3w
Fcr , 2 = 0,9 k F, 2 E (6.10)
b1

(4) ly should be obtained from 6.5.

6.5 Effective loaded length

(1) The effective loaded length ly should be calculated as follows for type (a) and (b) in Figure 6.1:

⎛ f yf bf ⎞
ly = ss + 2tf ⎜1 + ⎟≤a (6.11)
⎜ f yw tw ⎟⎠
⎝
For box girders, bf in expression (6.11) should be limited to 15εtf on each side of the web.

(2) For type c) ly should be taken as the smallest value obtained from equations (6.11), (6.12) and
(6.13):
2
⎛l
f yf b f ⎞
l y = le + tf +⎜ e ⎟ (6.12)
2 f yw hw ⎜⎝ t f ⎟
⎠

f yf b f
l y = le + tf (6.13)
f yw hw

k F E t w2
where l e = ≤ ss + c (6.14
2 f yw hw

6.6 Verification
(1) The verification should be performed as follows:
FEd
η2 = ≤ 1,0 (6.15)
FRd

where FEd is the design transverse force;

FRd is the design resistance to transverse forces, see 6.2(1).

Background information
The present rules in EN 1993-1-5:2006 are based on [22] and represented a step forward compared to
previous rules. They gave a higher resistance particularly for long loaded lengths and they were written
in the same format as rules for other buckling phenomena. During the drafting of EN 1993-1-5 the
reduction function was modified such that the resistance was lowered especially for very slender webs.
The rules for webs with longitudinal stiffeners were based on [20] and they gave an increase of the

47
resistance based on the increase of the first critical load. It was later shown in [8] and [26] that the
resistance did not follow that simple assumption.
During the COMBRI research project the patch loading problem was studied further and the results are
summarized in [5] and more detailed information is given in [19] and [4]. A brief description of the
background to the suggested changes in the design rules will be given here.
The first change to be discussed is the definition of the plastic resistance. For the case of a concentrated
load transferred via a flange to the web there is no obvious definition because as the deformation
increases the resistance keeps increasing gradually, see Figure 4-22, which shows results from a
computer simulation from [19]. This is partially due to strain hardening but also an effect of increased
loaded length as the bending of the loaded flange increases.
1200
0.2
1000 0

800 -0.2
Load [kN]

-0.4

σw / fyw
600
-0.6
400 -0.8
ly,FE,norm=1,0
200
-1
ly,m +m ,norm=1,28
1 2
-1.2 ly,m ,norm=1,01
0 1

-1.4
0 2 4 6 8 10 12 0 0.2 0.4 0.6 0.8 1
Displacement [mm] Distance, x, 0,5 centre

Figure 4-22: Load-displacement curve (left) and vertical stress distribution in the web plate
(right) for FE-analysis. Material properties corresponding to a S355 steel.

In an attempt to harmonize the design rules for buckling Müller noted that the rules for patch loading
differed substantially from other plate buckling rules [25] such that the reduction function was much
lower. The reason was stated to be a too high starting value, the plastic resistance. Davaine noted the
same in her study of patch loading of longitudinally stiffened girders [8] and proposed that the effective
loaded length should be taken according to expression (6.9) in the proposed new text for EN 1993-1-5
above. This issue was further studied by Gozzi in [19] in which results from computer simulations of
several patch loading cases were presented. In the model the web was prevented from buckling. The
results in Figure 4-22 are for a girder with web 4x992 mm and flange 16x250 mm and a loaded length
of 200 mm. The left diagram shows the relation between load and deformation and the dot on the curve
marks the proposed plastic resistance. The right hand figure shows the vertical stresses in the web
normalized with the yield strength. The solid line is the FE result and the two rectangular blocks are the
old and the new definition of the plastic resistance.
The study in [19] included also a check of reversible behaviour for the case of a travelling load. In order
to guarantee a reversible behaviour the load in SLS should fulfil
FSLS ≤ (0,05 + 0,44λ F ) FR
where FR is the characteristic resistance in ULS.
The new definition of the plastic resistance and the critical load requires a new reduction function.
There is today a trend to harmonize the mathematical form of the reduction functions to the type given
in Annex B of EN 1993-1-5 and the expressions in (6.2) to (6.5) in the proposed new text are of this
type and were developed in [19]. It differs from the one proposed in Annex B in that it gives a lower
resistance for λ > 0,6 and a higher for smaller λ. The reduction function was evaluated from 186 tests
with small bending moments and the comparison is shown in Figure 4-23. The test data base is the same
as the one used by Lagerqvist [22] for drafting the present rules except the additional tests from the
COMBRI research project which are designated “Tests, author”. The statistical evaluation shows a
smaller scatter than the present rules in EN 1993-1-5 but actually higher than the original proposal of
Lagerqvist. It may still be seen as a step forward in harmonizing design rules for plate buckling. An
evaluation according to Annex D of EN 1990 resulted in a recommended value γM1 of 1,0.

48
For stocky plates it is clear from Figure 4-23 that the resistance exceeds the new plastic resistance. The
reasons have been discussed above. In order to partially compensate for this, the reduction function is
allowed to continue up to 1,2 in the same way as it does for the shear resistance and in fact for very
much the same reasons.

Figure 4-23: Proposed reduction factor for patch loading resistance compared with test
results with small bending moment (less than 0,4 times resistance)

Patch loading of girders with longitudinal stiffeners has been studied in two thesis’s since the EN 1993-
1-5 was written. Davaine made extensive computer simulations presented in [8], which showed that the
present rules in EN 1993-1-5 were oversimplified. The critical load for global buckling in EN 1993-1-5
was accepted but there was a need also to consider buckling of the loaded panel and an empirical
estimate of this critical load was developed, see expression (6.7) in the proposed new text. It was also
suggested that an interaction between the two buckling modes should be considered. In a later study by
Clarin [4] it was however shown that it was sufficient to use the smaller of the two critical loads in
combination with the same reduction function as for unstiffened webs. The validation of this proposal is
shown in Figure 4-24 for tests from literature and for computer simulations from [8]. The results are for
open stiffeners but also closed stiffeners were studied and it was shown that closed stiffeners gave a
larger increase of the resistance. A calibration according to Annex D of EN 1990 showed that γM1 = 1,0
can be recommended.

Figure 4-24: Test results (left) and computer simulations (right) for patch loading of girders
with longitudinal stiffeners related to the proposed characteristic resistance

49
4.5 Conclusion
This chapter has summarised a number of proposals which have been developed in the frame of the
COMBRI research project [5] and the COMBRI+ project. They are classified according to their
possibility of implementation as:
• Nationally Determined Parameter (NDP)
• Non-Contradictory Complementary Information (NCCI)
• Amendment for revision
All amendments clearly aim at an improved economy and easier harmonisation so that they are strongly
recommended to be implemented at due time. The same is valid for most of the NCCI. However, some
proposals for NCCI tackle questions of safety such as Section 4.2.2 so that a strong recommendation is
given to implement them already in the National Annexes drafted by the various countries.

5 FINAL REMARKS
The valorisation of the results of the COMBRI research project was successfully implemented in the
COMBRI Design Manual and its promotion and dissemination in several seminars and workshops at
national level. Finally the manual was made available in English, French, German and Spanish
language. The feedback for participants was positively throughout. Compliments were received
especially for highlighting of the possibilities and restrictions of current Eurocode rules and the
overview on best national practices in the different European countries. The presentation of EBPlate
software which was developed during the COMBRI research project was highly appreciated. In the
following the conclusions which have been drawn in the COMBRI Design Manual are summarised in
short:
An overview of the bridge types in the participating partner’s countries - Belgium, France, Germany,
Spain and Sweden – show the current practice in those countries. It can be stated that 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 practice. Thus, the solutions presented are intended to serve as inspiration for the conceptual design
of new bridges.
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. It is shown that in most cases such high grades are not feasible. The
problem is usually that the fatigue requirements limit the full utilisation 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 the manual was redesigned from S355 to a hybrid girder
with S460 and S690 and it turned out that the costs of the material was reduced by 10% in the spans and
25% at the piers. In addition, there is a reduction of the fabrication costs.
Flanges as bottom flanges in box-girders are in most cases stiffened and different types of stiffeners are
discussed. It is shown that 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 EBPlate software 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 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.
With regard to webs the focus was 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
50
buckling resistance. However, unless the distance between the transverse stiffeners is very short this
effect is small and it does not justify the costs of the stiffeners. The possibility of omitting the transverse
stiffeners is discussed. Besides that, longitudinal stiffeners on webs increase the resistance for bending
as well as for shear. The economy of using longitudinal stiffeners was 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 use
discontinuous stiffeners and another is to put the transverse and the longitudinal stiffeners on opposite
sides of the web.
Cross bracings and diaphragms for I-girder bridges and box-girders have the function 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. Although it is not much material used for
cross bracings, from an economical point of view it is important to minimize the man hours for
fabrication. This was 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.
The technique of launching bridges is very popular and verification methods have been improved in the
COMBRI research project. They allow now 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 the 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.
The proposals which have been formulated for an implementation into standardisation were classified
according to Nationally Determined Parameters (NDP), Non-Contradictory Complementary
Information (NCCI) and amendments to be used in the next revision of Eurocode in order to provide a
concise background and proposal scheme. The transfer of these proposals to CEN is ensured through
the collaborative work of the project members in that committee.

51
European Commission

EUR 24175 — C
 OMBRI — Valorisation of knowledge for competitive steel and composite
bridges

U. Kuhlmann, B. Braun, M. Feldmann, J. Naumes, P.-O. Martin, Y. Galéa, J. Raoul, L. Davaine,

A. Petel, B. Johansson, P. Collin, J. Eriksen, H. Degée, N. Hausoul, J. Chica, S. Meno

Luxembourg: Publications Office of the European Union

2010 — 51 pp. — 21 × 29.7 cm

Research Fund for Coal and Steel series

ISBN 978-92-79-14046-4

ISSN 1018-5593

doi: 10.2777/75697

Price (excluding VAT) in Luxembourg: EUR 4