Designers' Guide To en 1994-2 - Eurocode 4 - Design of Composite Steel and Concrete Structures

Delivered by ICEVirtualLibrary.
com to:
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:28:59
DESIGNERS GUIDES TO THE EUROCODES
DESIGNERS GUIDE TO EN 1994-1-1
EUROCODE 4: DESIGN OF COMPOSITE STEEL
AND CONCRETE STRUCTURES
PART 1.1: GENERAL RULES AND RULES
FOR BUILDINGS
1
Designers Guide to EN 1994-1-1
12 May 2004 11:52:25
Color profile: Disabled
Composite Default screen
Delivered by ICEVirtualLibrary.com to:
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:28:59
2
12 May 2004 11:52:25
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:28:59
EUROCODE 4: DESIGN OF COMPOSITE
STEEL AND CONCRETE STRUCTURES
PART 1.1: GENERAL RULES AND RULES
FOR BUILDINGS
R. P. JOHNSON and D. ANDERSON
3
12 May 2004 11:52:25
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:28:59
Published by Thomas Telford Publishing, Thomas Telford Ltd, 1 Heron Quay, London E14 4JD
URL: http://www.thomastelford.com
Distributors for Thomas Telford books are
USA: ASCE Press, 1801 Alexander Bell Drive, Reston, VA 20191-4400
Japan: Maruzen Co. Ltd, Book Department, 310 Nihonbashi 2-chome, Chuo-ku, Tokyo 103
Australia: DA Books and Journals, 648 Whitehorse Road, Mitcham 3132, Victoria
First published 2004
Also available from Thomas Telford Books
Designers Guide to EN 1990. Eurocode: Basis of Structural Design. H. Gulvanessian, J.-A.Calgaro
and M. Holick. ISBN 0 7277 3011 8
A catalogue record for this book is available from the British Library
ISBN: 0 7277 3151 3
The authors and Thomas Telford Limited 2004
All rights, including translation, reserved. Except as permitted by the Copyright, Designs and Patents
Act 1988, no part of this publication may be reproduced, stored in a retrieval system or transmitted in
any form or by any means, electronic, mechanical, photocopying or otherwise, without the prior
written permission of the Publishing Director, Thomas Telford Publishing, Thomas Telford Ltd, 1
Heron Quay, London E14 4JD
This book is published on the understanding that the authors are solely responsible for the statements
made and opinions expressed in it and that its publication does not necessarily imply that such
statements and/or opinions are or reflect the views or opinions of the publishers. While every effort has
been made to ensure that the statements made and the opinions expressed in this publication provide a
safe and accurate guide, no liability or responsibility can be accepted in this respect by the authors or
publishers
Typeset by Helius, Brighton and Rochester
Printed and bound in Great Britain by MPG Books, Bodmin
4
12 May 2004 11:52:25
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:28:59
Preface
EN 1994, also known as Eurocode 4, is one standard of the Eurocode suite and describes the
principles and requirements for safety, serviceability and durability of composite steel and
concrete structures. It is subdivided into three parts:
Part 1.1: General Rules and Rules for Buildings
Part 1.2: Structural Fire Design
Part 2: Bridges.
It is intended to be used in conjunction with EN 1990, Basis of Structural Design, EN 1991,
Actions on Structures, and the other design Eurocodes.
Aims and objectives of this guide
The principal aim of this book is to provide the user with guidance on the interpretation and
use of EN 1994-1-1 and to present worked examples. The guide explains the relationship
with the other Eurocode parts to which it refers and with the relevant British codes. It
also provides background information and references to enable users of Eurocode 4 to
understand the origin and objectives of its provisions.
Layout of this guide
EN 1994-1-1 has a foreword and nine sections, together with three annexes. This guide has
an introduction which corresponds to the foreword of EN 1994-1-1, and Chapters 1 to 9 of
the guide correspond to Sections 1 to 9 of the Eurocode. Chapters 10 and 11 correspond to
Annexes A and B of the Eurocode, respectively. Appendices A to C of this guide include
useful material from the draft Eurocode ENV 1994-1-1.
The numbering and titles of the sections in this guide also correspond to those of the
clauses of EN 1994-1-1. Some subsections are also numbered (e.g. 1.1.2). This implies
correspondence with the subclause in EN 1994-1-1 of the same number. Their titles also
correspond. There are extensive references to lower-level clause and paragraph numbers.
The first significant reference is in bold italic type (e.g. clause 1.1.1(2)). These are in strict
numerical sequence throughout the book, to help readers to find comments on particular
provisions of the code. Some comments on clauses are necessarily out of sequence, but use of
the index should enable these to be found.
All cross-references in this guide to sections, clauses, subclauses, paragraphs, annexes,
figures, tables and equations of EN 1994-1-1 are in italic type, which is also used where text
from a clause in EN 1994-1-1 has been directly reproduced (conversely, cross-references
to and quotations from other sources, including other Eurocodes, are in roman type).
5
12 May 2004 11:52:25
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:28:59
Expressions repeated from EN 1994-1-1 retain their number; other expressions have
numbers prefixed by D (for Designers Guide), e.g. equation (D6.1) in Chapter 6.
Acknowledgements
The authors are deeply indebted to the other members of the four project teams for
Eurocode 4 on which they have worked: Jean-Marie Aribert, Gerhard Hanswille, Bernt
Johansson, Basil Kolias, Jean-Paul Lebet, Henri Mathieu, Michel Mele, Joel Raoul,
Karl-Heinz Roik and Jan Stark; and also to the Liaison Engineers, National Technical
Contacts, and others who prepared national comments. They thank the University of
Warwick for the facilities provided for Eurocode work, and, especially, their wives Diana and
Linda for their unfailing support.
R. P. Johnson
D. Anderson
vi
6
12 May 2004 11:52:25
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:28:59
Contents
Preface v
Aims and objectives of this guide v
Layout of this guide v
Acknowledgements vi
Introduction 1
Chapter 1. General 3
1.1. Scope 3
1.1.1. Scope of Eurocode 4 3
1.1.2. Scope of Part 1.1 of Eurocode 4 3
1.2. Normative references 5
1.2.1. General reference standards 5
1.2.2. Other reference standards 5
1.3. Assumptions 5
1.4. Distinction between principles and application rules 5
1.5. Definitions 6
1.5.1. General 6
1.5.2. Additional terms and definitions 6
1.6. Symbols 6
Chapter 2. Basis of design 9
2.1. Requirements 9
2.2. Principles of limit states design 9
2.3. Basic variables 9
2.4. Verification by the partial factor method 10
2.4.1. Design values 10
2.4.2. Combination of actions 11
2.4.3. Verification of static equilibrium (EQU) 11
Chapter 3. Materials 13
3.1. Concrete 13
3.2. Reinforcing steel 15
3.3. Structural steel 16
3.4. Connecting devices 16
3.4.1. General 16
3.4.2. Stud shear connectors 17
3.5. Profiled steel sheeting for composite slabs in buildings 17
7
12 May 2004 11:52:25
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:28:59
Chapter 4. Durability 19
4.1. General 19
4.2. Profiled steel sheeting for composite slabs in buildings 19
Chapter 5. Structural analysis 21
5.1. Structural modelling for analysis 21
5.1.1. Structural modelling and basic assumptions 21
5.1.2. Joint modelling 21
5.2. Structural stability 22
5.2.1. Effects of deformed geometry of the structure 22
5.2.2. Methods of analysis for buildings 23
5.3. Imperfections 24
5.3.1. Basis 24
5.3.2. Imperfections in buildings 24
5.4. Calculation of action effects 27
5.4.1. Methods of global analysis 27
Example 5.1: effective width of concrete flange 29
5.4.2. Linear elastic analysis 29
5.4.3. Non-linear global analysis 33
5.4.4. Linear elastic analysis with limited redistribution for
buildings 34
5.4.5. Rigid plastic global analysis for buildings 36
5.5. Classification of cross-sections 37
Chapter 6. Ultimate limit states 41
6.1. Beams 41
6.1.1. Beams for buildings 41
6.1.2. Effective width for verification of cross-sections 43
6.2. Resistances of cross-sections of beams 43
6.2.1. Bending resistance 44
Example 6.1: resistance moment in hogging bending, with effective web 50
6.2.2. Resistance to vertical shear 54
Example 6.2: resistance to bending and vertical shear 55
6.3. Resistance of cross-sections of beams for buildings with partial
encasement 57
6.3.1. Scope 57
6.3.2. Resistance to bending 57
6.3.36.3.4. Resistance to vertical shear, and to bending and
vertical shear 57
6.4. Lateraltorsional buckling of composite beams 58
6.4.1. General 58
6.4.2. Verification of lateraltorsional buckling of continuous
composite beams with cross-sections in Class 1, 2 and 3
for buildings 58
6.4.3. Simplified verification for buildings without direct
calculation 61
Use of intermediate lateral bracing 63
Flow charts for continuous beam 64
Example 6.3: lateraltorsional buckling of two-span beam 66
6.5. Transverse forces on webs 66
6.6. Shear connection 67
6.6.1. General 67
Example 6.4: arrangement of shear connectors 69
viii
8
12 May 2004 11:52:25
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:28:59
6.6.2. Longitudinal shear force in beams for buildings 70
6.6.3. Headed stud connectors in solid slabs and concrete
encasement 70
6.6.4. Design resistance of headed studs used with profiled
steel sheeting in buildings 72
Example 6.5: reduction factors for transverse sheeting 76
6.6.5. Detailing of the shear connection and influence of
execution 76
6.6.6. Longitudinal shear in concrete slabs 81
Example 6.6: transverse reinforcement for longitudinal shear 82
Example 6.7: two-span beam with a composite slab ultimate limit
state 84
Example 6.8: partial shear connection with non-ductile connectors 100
Example 6.9: elastic resistance to bending, and influence of degree
of shear connection and type of connector on bending resistance 101
6.7. Composite columns and composite compression members 103
6.7.1. General 103
6.7.2. General method of design 105
6.7.3. Simplified method of design 105
6.7.4. Shear connection and load introduction 111
6.7.5. Detailing provisions 113
Example 6.10: composite column with bending about one or both axes 113
Example 6.11: longitudinal shear outside areas of load introduction,
for a composite column 118
6.8. Fatigue 119
6.8.1. General 119
6.8.2. Partial factors for fatigue assessment 119
6.8.3. Fatigue strength 120
6.8.4. Internal forces and fatigue loadings 120
6.8.5. Stresses 121
6.8.6. Stress ranges 122
6.8.7. Fatigue assessment based on nominal stress ranges 123
Example 6.12: fatigue in reinforcement and shear connection 124
Chapter 7. Serviceability limit states 127
7.1. General 127
7.2. Stresses 128
7.3. Deformations in buildings 128
7.3.1. Deflections 128
7.3.2. Vibration 130
7.4. Cracking of concrete 131
7.4.1. General 131
7.4.2. Minimum reinforcement 132
7.4.3. Control of cracking due to direct loading 134
General comments on clause 7.4 135
Example 7.1: two-span beam (continued) SLS 136
Chapter 8. Composite joints in frames for buildings 141
8.1. Scope 141
8.2. Analysis, including modelling and classification 142
8.3. Design methods 144
8.4. Resistance of components 145
Example 8.1: end-plate joints in a two-span beam in a braced frame 147
CONTENTS
ix
9
12 May 2004 11:52:25
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:28:59
Chapter 9. Composite slabs with profiled steel sheeting for buildings 161
9.1. General 161
9.2. Detailing provisions 162
9.3. Actions and action effects 162
9.4. Analysis for internal forces and moments 163
9.59.6. Verification of profiled steel sheeting as shuttering 164
9.7. Verification of composite slabs for the ultimate limit states 164
9.7.1. Design criterion 164
9.7.2. Flexure 164
9.7.3. Longitudinal shear for slabs without end anchorage 165
9.7.4. Longitudinal shear for slabs with end anchorage 167
9.7.5. Vertical shear 168
9.7.6. Punching shear 168
9.8. Verification of composite slabs for serviceability limit states 168
9.8.1. Cracking of concrete 168
9.8.2. Deflection 168
Example 9.1: two-span continuous composite slab 170
Chapter 10. Annex A (Informative). Stiffness of joint components in buildings 179
A.1. Scope 179
A.2. Stiffness coefficients 179
A.3. Deformation of the shear connection 181
Further comments on stiffness 181
Example 10.1: elastic stiffness of an end-plate joint 181
Chapter 11. Annex B (Informative). Standard tests 187
B.1. General 187
B.2. Tests on shear connectors 188
B.3. Testing of composite floor slabs 191
Example 11.1: mk tests on composite floor slabs 194
Example 11.2: the partial-interaction method 198
Appendix A. Lateraltorsional buckling of composite beams for buildings 203
Simplified expression for cracked flexural stiffness of a
composite slab 203
Flexural stiffness of beam with encased web 204
Maximum spacing of shear connectors for continuous U-frame
action 204
Top transverse reinforcement above an edge beam 206
Derivation of the simplified expression for
LT
206
Effect of web encasement on
LT
208
Factor C
4
for the distribution of bending moment 209
Criteria for verification of lateraltorsional stability without
direct calculation 209
Web encasement 210
Appendix B. The effect of slab thickness on resistance of composite slabs to
longitudinal shear 211
Summary 211
The model 211
x
10
12 May 2004 11:52:26
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:28:59
The mk method 212
The use of test results as predictors 212
Shape of function y(x) 213
Estimate of errors of prediction 213
Conclusion for the mk method 214
The partial-connection method 214
Conclusion for the partial-connection method 214
Appendix C. Simplified calculation method for the interaction curve for
resistance of composite column cross-sections to compression
and uniaxial bending 217
Scope and method 217
Resistance to compression 218
Position of neutral axis 219
Bending resistances 219
Interaction with transverse shear 219
Neutral axes and plastic section moduli of some cross-sections 219
General 219
Major-axis bending of encased I-sections 220
Minor-axis bending of encased I-sections 220
Concrete-filled circular and rectangular hollow sections 221
Example C.1: NM interaction polygon for a column cross-section 222
References 225
Index 231
CONTENTS
xi
11
12 May 2004 11:52:26
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:16
Introduction
The provisions of EN 1994-1-1
1
are preceded by a foreword, most of which is common to all
Eurocodes. This Foreword contains clauses on:
the background to the Eurocode programme
the status and field of application of the Eurocodes
national standards implementing Eurocodes
links between Eurocodes and harmonized technical specifications for products
additional information specific to EN 1994-1-1
National Annex for EN 1994-1-1.
Guidance on the common text is provided in the introduction to the Designers Guide to
EN1990, Eurocode: Basis of Structural Design,
2
and only background information essential to
users of EN 1994-1-1 is given here.
EN 1990
3
lists the following structural Eurocodes, each generally consisting of a number
of parts which are in different stages of development at present:
EN 1990 Eurocode: Basis of Structural Design
EN 1991 Eurocode 1: Actions on Structures
EN 1992 Eurocode 2: Design of Concrete Structures
EN 1993 Eurocode 3: Design of Steel Structures
EN 1994 Eurocode 4: Design of Composite Steel and Concrete Structures
EN 1995 Eurocode 5: Design of Timber Structures
EN 1996 Eurocode 6: Design of Masonry Structures
EN 1997 Eurocode 7: Geotechnical Design
EN 1998 Eurocode 8: Design of Structures for Earthquake Resistance
EN 1999 Eurocode 9: Design of Aluminium Structures
The information specific to EN 1994-1-1 emphasizes that this standard is to be used with
other Eurocodes. The standard includes many cross-references to particular clauses in
EN 1992
4
and EN 1993.
5
Similarly, this guide is one of a series on Eurocodes, and is for use
with the guide for EN 1992-1-1
6
and the guide for EN 1993-1-1.
7
It is the responsibility of each national standards body to implement each Eurocode part
as a national standard. This will comprise, without any alterations, the full text of the
Eurocode and its annexes as published by the European Committee for Standardization
(CEN). This will usually be preceded by a National Title Page and a National Foreword, and
may be followed by a National Annex.
Each Eurocode recognizes the right of national regulatory authorities to determine values
related to safety matters. Values, classes or methods to be chosen or determined at national
level are referred to as Nationally Determined Parameters (NDPs), and are listed in the
foreword to each Eurocode, in the clauses on National Annexes. NDPs are also indicated by
13
12 May 2004 11:52:26
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:16
notes immediately after relevant clauses. Each National Annex will give or cross-refer to the
NDPs to be used in the relevant country. Otherwise the National Annex may contain only the
following:
8
decisions on the application of informative annexes, and
references to non-contradictory complementary information to assist the user in applying
the Eurocode.
In EN 1994-1-1 the NDPs are principally the partial factors for material or product
properties peculiar to this standard; for example, for the resistance of headed stud shear
connectors, and of composite slabs to longitudinal shear. Other NDPs are values that may
depend on climate, such as the free shrinkage of concrete.
2
14
12 May 2004 11:52:26
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:20
CHAPTER 1
General
This chapter is concerned with the general aspects of EN 1994-1-1, Eurocode 4: Design of
Composite Steel and Concrete Structures, Part 1.1: General Rules and Rules for Buildings. The
material described in this chapter is covered in Section 1, in the following clauses:
Scope Clause 1.1
Normative references Clause 1.2
Assumptions Clause 1.3
Distinction between principles and application rules Clause 1.4
Definitions Clause 1.5
Symbols Clause 1.6
1.1. Scope
1.1.1. Scope of Eurocode 4
Clause 1.1.1
Clause 1.1.1(2)
The scope of EN1994 (all three parts) is outlined in clause 1.1.1. It is to be used with EN1990,
Eurocode: Basis of Structural Design, which is the head document of the Eurocode suite.
Clause 1.1.1(2) emphasizes that the Eurocodes are concerned with structural behaviour and
that other requirements, e.g. thermal and acoustic insulation, are not considered.
The basis for verification of safety and serviceability is the partial factor method. EN 1990
recommends values for load factors and gives various possibilities for combinations of
actions. The values and choice of combinations are to be set by the National Annex for the
country in which the structure is to be constructed.
Eurocode 4 is also to be used in conjunction with EN 1991, Eurocode 1: Actions on
Structures
9
and its National Annex, to determine characteristic or nominal loads. When a
composite structure is to be built in a seismic region, account needs to be taken of EN 1998,
Eurocode 8: Design of Structures for Earthquake Resistance.
10
Clause 1.1.1(3)
The Eurocodes are concerned with design and not execution, but minimum standards of
workmanship are required to ensure that the design assumptions are valid. For this reason,
clause 1.1.1(3) lists the European standards for the execution of steel structures and the
execution of concrete structures. The former includes some requirements for composite
construction, for example for the testing of welded stud shear connectors.
1.1.2. Scope of Part 1.1 of Eurocode 4
EN 1994-1-1 deals with aspects of design that are common to the principal types of
composite structure, buildings and bridges. This results from the CEN requirement that a
provision should not appear in more than one EN standard, as this can cause inconsistency
when one standard is revised before another. For example, if the same rules for resistance to
bending apply for a composite beam in a building as in a bridge (as most of them do), then
15
12 May 2004 11:52:26
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:20
those rules are general and must appear in EN 1994-1-1 and not in EN 1994-2 (on
bridges).
11
This has been done even where most applications occur in bridges. For example,
clause 6.8 (fatigue) is in Part 1.1, with a few additional provisions in Part 2.
In EN 1994-1-1, all rules that are for buildings only are preceded by a heading that
includes the word buildings, or, if an isolated paragraph, are placed at the end of the
relevant clause, e.g. clauses 5.3.2 and 5.4.2.3(5).
The coverage in this guide of the general clauses of Part 1.1 is relevant to both buildings
and bridges, except where noted otherwise. However, guidance provided by or related to the
worked examples may be relevant only to applications in buildings.
Clause 1.1.2(2) Clause 1.1.2(2) lists the titles of the sections of Part 1.1. Those for Sections 17 are the
same as in the other material-dependent Eurocodes. The contents of Sections 1 and 2
similarly follow an agreed model.
The provisions of Part 1.1 cover the design of the common composite members:
beams in which a steel section acts compositely with concrete
composite slabs formed with profiled steel sheeting
concrete-encased or filled composite columns
joints between composite beams and steel or composite columns.
Sections 5 and 8 concern connected members. Section 5, Structural analysis, is needed
particularly for a frame that is not of simple construction. Unbraced frames and sway
frames are within its scope. The provisions include the use of second-order global analysis
and prestress by imposed deformations, and define imperfections.
The scope of Part 1.1 extends to steel sections that are partially encased. The web of the
steel section is encased by reinforced concrete, and shear connection is provided between
the concrete and the steel. This is a well-established form of construction. The primary
reason for its choice is improved resistance in fire.
Fully encased composite beams are not included because:
no satisfactory model has been found for the ultimate strength in longitudinal shear of a
beam without shear connectors
it is not known to what extent some design rules (e.g. for momentshear interaction and
redistribution of moments) are applicable.
Afully encased beamwith shear connectors can usually be designed as if partly encased or
uncased, provided that care is taken to prevent premature spalling of encasement in
compression.
Part 2, Bridges, includes further provisions that may on occasion be useful for buildings,
such as those on:
composite plates (where the steel member is a flat steel plate, not a profiled section)
composite box girders
tapered or non-uniform composite members
structures that are prestressed by tendons.
The omission of application rules for a type of member or structure should not prevent its
use, where appropriate. Some omissions are deliberate, to encourage the use of innovative
design, based on specialized literature, the properties of materials, and the fundamentals of
equilibrium and compatibility; and following the principles given in the relevant Eurocodes.
This applies, for example, to:
large holes in webs of beams
types of shear connector other than welded studs
base plates beneath composite columns
shear heads in reinforced concrete framed structures, and
many aspects of mixed structures, as used in tall buildings.
4
16
12 May 2004 11:52:26
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:20
In addition to its nine normative sections, EN1994-1-1 includes three informative annexes:
Annex A, Stiffness of joint components in buildings
Annex B, Standard tests
Annex C, Shrinkage of concrete for composite structures for buildings.
The reasons for these annexes, additional to the normative provisions, are explained in the
relevant chapters of this guide.
1.2. Normative references
References are given only to other European standards, all of which are intended to be used
as a package. Formally, the Standards of the International Organization for Standardization
(ISO) apply only if given an EN ISO designation. National standards for design and for
products do not apply if they conflict with a relevant EN standard.
It is intended that, following a period of overlap, all competing national standards will
be withdrawn by around 2010. As Eurocodes may not cross-refer to national standards,
replacement of national standards for products by EN or ISO standards is in progress, with a
time-scale similar to that for the Eurocodes.
During the period of changeover to Eurocodes and EN standards it is likely that an EN
referred to, or its National Annex, may not be complete. Designers who then seek guidance
from national standards should take account of differences between the design philosophies
and safety factors in the two sets of documents.
1.2.1. General reference standards
Some references here, and also in clause 1.2.2, appear to repeat references in clause 1.1.1.
The difference is explained in clause 1.2. These dated references define the issue of the
standard that is referred to in detailed cross-references, given later in EN 1994-1-1.
1.2.2. Other reference standards
Eurocode 4 necessarily refers to EN1992-1-1, Eurocode 2: Design of Concrete Structures, Part
1.1: General Rules and Rules for Buildings, and to several parts of EN 1993, Eurocode 3:
Design of Steel Structures.
In its application to buildings, EN 1994-1-1 is based on the concept of the initial erection
of a steel frame, which may include prefabricated concrete-encased members. The placing of
profiled steel sheeting or other shuttering follows. The addition of reinforcement and in situ
concrete completes the composite structure. The presentation and content of EN 1994-1-1
therefore relate more closely to EN 1993-1-1 than to EN 1992-1-1.
1.3. Assumptions
The general assumptions are those of EN 1990, EN 1992 and EN 1993. Commentary on
them will be found in the relevant guides in this series.
1.4. Distinction between principles and application rules
Clauses in the Eurocodes are set out as either principles or application rules. As defined by
EN 1990:
Principles comprise general statements for which there is no alternative and requirements
and analytical models for which no alternative is permitted unless specifically stated
Principles are distinguished by the letter P following the paragraph number
Application Rules are generally recognised rules which comply with the principles and
satisfy their requirements.
5
CHAPTER 1. GENERAL
17
12 May 2004 11:52:26
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:20
There are relatively few principles. It has been recognized that a requirement or analytical
model for which no alternative is permitted unless specifically stated can rarely include a
numerical value, because most values are influenced by research and/or experience, and may
change over the years. (Even the specified elastic modulus for structural steel is an approximate
value.) Furthermore, a clause cannot be a principle if it requires the use of another clause that
is an application rule; effectively that clause also would become a principle.
It follows that, ideally, the principles in all the codes should forma consistent set, referring
only to each other, and intelligible if all the application rules were deleted. This over-riding
principle has strongly influenced the drafting of EN 1994.
1.5. Definitions
1.5.1. General
In accordance with the model for Section 1, reference is made to the definitions given
in clauses 1.5 of EN 1990, EN 1992-1-1, and EN 1993-1-1. Many types of analysis are defined
in clause 1.5.6 of EN 1990. It is important to note that an analysis based on the deformed
geometry of a structure or element under load is termed second order rather than
non-linear. The latter term refers to the treatment of material properties in structural
analysis. Thus, according to EN 1990 non-linear analysis includes rigid plastic. This
convention is not followed in EN 1994-1-1, where the heading Non-linear global analysis
(clause 5.4.3) does not include rigid plastic global analysis (clause 5.4.5).
Clause 1.5.1 References from clause 1.5.1 include clause 1.5.2 of EN 1992-1-1, which defines prestress
as an action caused by the stressing of tendons. This applies to EN 1994-2 but not to
EN 1994-1-1, as this type of prestress is outside its scope. Prestress by jacking at supports,
which is outside the scope of EN 1992-1-1, is within the scope of EN 1994-1-1.
The definitions in clauses 1.5.1 to 1.5.9 of EN1993-1-1 apply where they occur in clauses in
EN 1993 to which EN 1994 refers. None of them uses the word steel.
1.5.2. Additional terms and definitions
Clause 1.5.2 Most of the 13 definitions in clause 1.5.2 of EN 1994-1-1 include the word composite. The
definition of shear connection does not require the absence of separation or slip at the
interface between steel and concrete. Separation is always assumed to be negligible, but
explicit allowance may need to be made for effects of slip, e.g. in clauses 5.4.3, 7.2.1, 9.8.2(7)
and A.3.
The definition composite frame is relevant to the use of Section 5. Where the behaviour is
essentially that of a reinforced or prestressed concrete structure, with only a few composite
members, global analysis should generally be in accordance with Eurocode 2.
These lists of definitions are not exhaustive, because all the codes use terms with precise
meanings that can be inferred from their contexts.
Concerning use of words generally, there are significant differences from British codes.
These arose from the use of English as the base language for the drafting process, and the
need to improve precision of meaning and to facilitate translation into other European
languages. In particular:
action means a load and/or an imposed deformation
action effect (clause 5.4) and effect of action have the same meaning: any deformation
or internal force or moment that results from an action.
1.6. Symbols
The symbols in the Eurocodes are all based on ISO standard 3898: 1987.
12
Each code has its
own list, applicable within that code. Some symbols have more than one meaning, the
particular meaning being stated in the clause.
6
18
12 May 2004 11:52:27
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:20
There are a few important changes from previous practice in the UK. For example, an xx
axis is along a member, a yy axis is parallel to the flanges of a steel section (clause 1.7(2) of
EN 1993-1-1), and a section modulus is W, with subscripts to denote elastic or plastic
behaviour.
Wherever possible, definitions in EN 1994-1-1 have been aligned with those in EN 1990,
EN 1992 and EN 1993; but this should not be assumed without checking the list in clause 1.6.
Some quite minor differences are significant.
The symbol f
y
has different meanings in EN 1992-1-1 and EN 1993-1-1. It is retained in
EN 1994-1-1 for the nominal yield strength of structural steel, though the generic subscript
for that material is a, based on the French word for steel, acier. Subscript a is not used in
EN 1993-1-1, where the partial factor for steel is not
A
, but
M
; and this usage is followed in
EN 1994-1-1. The characteristic yield strength of reinforcement is f
sk
, with partial factor
S
.
7
CHAPTER 1. GENERAL
19
12 May 2004 11:52:27
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:25
CHAPTER 2
Basis of design
The material described in this chapter is covered in Section 2 of EN1994-1-1, in the following
clauses:
Requirements Clause 2.1
Principles of limit states design Clause 2.2
Basic variables Clause 2.3
Verification by the partial factor method Clause 2.4
The sequence follows that of EN 1990, Sections 24 and 6.
2.1. Requirements
Design is to be in accordance with the general requirements of EN 1990. The purpose of
Section 2 is to give supplementary provisions for composite structures.
Clause 2.1(3) Clause 2.1(3) reminds the user again that design is based on actions and combinations of
actions in accordance with EN 1991 and EN 1990, respectively. The use of partial safety
factors for actions and resistances (the partial factor method) is expected but is not a
requirement of Eurocodes. The method is presented in Section 6 of EN 1990 as one way of
satisfying the basic requirements set out in Section 2 of that standard. This is why use of the
partial factor method is given deemed to satisfy status in clause 2.1(3). To establish that a
design was in accordance with the Eurocodes, the user of any other method would normally
have to demonstrate, to the satisfaction of the regulatory authority and/or the client, that the
method satisfied the basic requirements of EN 1990.
2.2. Principles of limit states design
The clause provides a reminder that the influence of sequence of construction on action
effects must be considered. It does not affect the bending resistance of beams that are in
Class 1 or 2 (as defined in clause 5.5) or the resistance of a composite column, as these are
determined by rigid plastic theory, but it does affect the resistances of beams in Class 3 or 4.
2.3. Basic variables
Clause 2.3.3 The classification of effects of shrinkage and temperature in clause 2.3.3 into primary and
secondary will be familiar to designers of continuous beams, especially for bridges.
Secondary effects are to be treated as indirect actions, which are sets of imposed
deformations (clause 1.5.3.1 of EN 1990), not as action effects. This distinction appears to
have no consequences in practice, for the use of EN 1994-1-1.
21
12 May 2004 11:52:27
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:25
2.4. Verification by the partial factor method
2.4.1. Design values
Clause 2.4.1.1
Clause 2.4.1.2
Clauses 2.4.1.1 and 2.4.1.2 illustrate the treatment of partial factors. Recommended values
are given in Notes, in the hope of eventual convergence between the values for each partial
factor that will be specified in the National Annexes. This process was adopted because the
regulatory bodies in the member states of CEN, rather than CEN itself, are responsible for
setting safety levels. The Notes are informative, not normative (i.e., not part of the preceding
provision), so that there are no numerical values in the principles of clause 2.4.1.2, as
explained earlier.
The Notes also link the partial factors for concrete, reinforcing steel and structural steel to
those recommended in EN 1992 and EN 1993. Design would be more difficult if the factors
for these materials in composite structures differed from the values in reinforced concrete
and steel structures.
The remainder of EN 1994-1-1 normally refers to design strengths, rather than
characteristic or nominal values with partial factors. The design strength for concrete is
given by
f
cd
= f
ck
/
C
(2.1)
where f
ck
is the characteristic cylinder strength. This definition is stated algebraically because
it differs from that of EN 1992-1-1, in which an additional coefficient
cc
is applied:
f
cd
=
cc
f
ck
/
C
(D2.1)
The coefficient is explained by EN 1992-1-1 as taking account of long-term effects and of
unfavourable effects resulting from the way the load is applied. The recommended value is
1.0, but a different value could be chosen in a National Annex. This possibility is not
appropriate for EN 1994-1-1 because the coefficient has been taken as 1.0 in calibration of
composite elements.
Clause 2.4.1.3 Clause 2.4.1.3 refers to product standards hEN. The h stands for harmonized. This
term from the Construction Products Directive
13
is explained in the Designers Guide to
EN 1990.
2
Clause 2.4.1.4 Clause 2.4.1.4, on design resistances, refers to expressions (6.6a) and (6.6c) given in clause
6.3.5 of EN1990. Resistances in EN1994-1-1 often need more than one partial factor, and so
use expression (6.6a), which is
R
d
= R{(
i
X
k, i
/
M, i
); a
d
} i 1 (D2.2)
For example, clause 6.7.3.2(1) gives the plastic resistance to compression of a cross-section as
the sum of terms for the structural steel, concrete and reinforcement:
N
pl, Rd
= A
a
f
yd
+ 0.85A
c
f
cd
+ A
s
f
sd
(6.30)
In this case, there is no separate term a
d
based on geometrical data, because uncertainties in
areas of cross-sections are allowed for in the
M
factors.
In terms of characteristic strengths, from clause 2.4.1.2, equation (6.30) becomes:
N
pl, Rd
= A
a
f
y
/
M
+ 0.85A
c
f
ck
/
C
+ A
s
f
sk
/
S
(D2.3)
in which:
the characteristic material strengths X
k, i
are f
y
, f
ck
and f
sk
the conversion factors,
i
in EN 1990, are 1.0 for steel and reinforcement and 0.85 for
concrete
the partial factors
M, i
are
M
,
C
and
S
.
Expression (6.6c) of EN 1990 is R
d
= R
k
/
M
. It applies where characteristic properties and
a single partial factor can be used; for example, in expressions for the shear resistance of a
headed stud (clause 6.6.3.1).
10
22
12 May 2004 11:52:27
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:25
2.4.2. Combination of actions
No comment is necessary.
2.4.3. Verification of static equilibrium (EQU)
The abbreviation EQU appears in EN 1990, where four types of ultimate limit state are
defined in clause 6.4.1:
EQU, for loss of static equilibrium
FAT, for fatigue failure
GEO, for failure or excessive deformation of the ground
STR, for internal failure or excessive deformation of the structure.
This guide covers ultimate limit states only of types STRand FAT. Use of type GEOarises
in design of foundations to EN 1997.
14
11
CHAPTER 2. BASIS OF DESIGN
23
12 May 2004 11:52:27
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:42
CHAPTER 3
Materials
This chapter concerns the properties of materials needed for the design of composite
structures. It corresponds to Section 3, which has the following clauses:
Concrete Clause 3.1
Reinforcing steel Clause 3.2
Structural steel Clause 3.3
Connecting devices Clause 3.4
Profiled steel sheeting for composite slabs in buildings Clause 3.5
Rather than repeating information given elsewhere, Section 3 consists mainly of cross-
references to other Eurocodes and EN standards. The following comments relate to
provisions of particular significance for composite structures.
3.1. Concrete
Clause 3.1(1) Clause 3.1(1) refers to EN 1992-1-1 for the properties of concrete. For lightweight-aggregate
concrete, several properties are dependent on the oven-dry density, relative to 2200 kg/m
3
.
Complex sets of time-dependent properties are given in its clause 3.1 for normal concrete
and clause 11.3 for lightweight-aggregate concrete. For composite structures built unpropped,
with several stages of construction, simplification is essential. Specific properties are now
discussed. (For thermal expansion, see Section 3.3.)
Strength and stiffness
Strength and deformation characteristics are summarized in EN 1992-1-1, Table 3.1 for
normal concrete and Table 11.3.1 for lightweight-aggregate concrete.
Strength classes for normal concrete are defined as Cx/y, where x and y are respectively the
cylinder and cube compressive strengths in units of newtons per square millimetre. All
compressive strengths in design rules in Eurocodes are cylinder strengths, so an unsafe error
occurs if a specified cube strength is used in calculations. It should be replaced at the outset
by the equivalent cylinder strength, using the relationships given by the strength classes.
Classes for lightweight concrete are designated LCx/y. The relationships between cylinder
and cube strengths differ from those of normal concrete.
Except where prestressing by tendons is used (which is outside the scope of this guide), the
tensile strength of concrete is rarely used in design calculations for composite members. The
mean tensile strength f
ctm
appears in the definitions of cracked global analysis in clause
5.4.2.3(2), and in clause 7.4.2(1) on minimum reinforcement. Its value and the 5 and 95%
fractile values are given in Tables 3.1 and 11.3.1 of EN 1992-1-1. The appropriate fractile
value should be used in any limit state verification that relies on either an adverse or
beneficial effect of the tensile strength of concrete.
25
12 May 2004 11:52:27
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:42
Values of the modulus of elasticity are given in Tables 3.1 and 11.3.1. Clause 3.1.3 of
EN 1992-1-1 points out that these are indicative, for general applications. The short-term
elastic modulus E
cm
increases for ages greater than 28 days. The influence of this small
change on the effective modulus is negligible compared with the uncertainties in the
modelling of creep, so it should be ignored.
Stress/strain properties
The design compressive strength of concrete, f
cd
, is defined in clause 3.1.6(1)P of EN 1992-1-1
as
f
cd
=
cc
f
ck
/
C
where
cc
is the coefficient taking account of long term effects on the compressive strength and of
unfavourable effects resulting from the way the load is applied.
Note: The value of
cc
for use in a Country should lie between 0.8 and 1.0 and may be found in its
National Annex. The recommended value is 1.
The reference in clause 3.1(1) to EN 1992-1-1 for properties of concrete begins unless
otherwise given by Eurocode 4 . Resistances of composite members given in EN 1994-1-1 are
based on extensive calibration studies (e.g. see Johnson and Huang
15,16
). The numerical
coefficients given in resistance formulae are consistent with the value
cc
= 1.0 and the use
of either elastic theory or the stress block defined in clause 6.2.1.2. Therefore, there is no
reference in EN 1994-1-1 to a coefficient
cc
or to a choice to be made in a National Annex.
The symbol f
cd
always means f
ck
/
C
, and for beams and most columns is used with the
coefficient 0.85, as in equation (6.30) in clause 6.7.3.2(1). An exception, in that clause, is when
the value of 0.85 is replaced by 1.0 for concrete-filled column sections, based on calibration.
The approximation made to the shape of the stressstrain curve is also relevant. Those
given in clause 3.1 of EN1992-1-1 are mainly curved or bilinear, but in clause 3.1.7(3) there is
a simpler rectangular stress distribution, similar to the stress block given in the British
Standard for the structural use of concrete, BS 8110.
17
Its shape, for concrete strength classes
up to C50/60, and the corresponding strain distribution are shown in Fig. 3.1.
This stress block is inconvenient for use with composite cross-sections, because the region
near the neutral axis assumed to be unstressed is often occupied by a steel flange, and
algebraic expressions for resistance to bending become complex.
In composite sections, the contribution from the steel section to the bending resistance
reduces the significance of that from the concrete. It is thus possible
18
for EN 1994 to allow
the use of a rectangular stress block extending to the neutral axis, as shown in Fig. 3.1.
For a member of unit width, the moment about the neutral axis of the EN1992 stress block
ranges from 0.38f
ck
x
2
/
C
to 0.48f
ck
x
2
/
C
, depending on the value chosen for
cc
. The value for
beams in EN 1994-1-1 is 0.425f
ck
x
2
/
C
. Calibration studies have shown that this overestimates
14
x
Plastic
neutral axis
0 0.0035
Compressive strain
0.8 x
0
EN 1994-1-1:
0.85f
cd
, with f
cd
= f
ck
/g
C
Compressive stress
EN 1992-1-1:
f
cd
= a
cc
f
ck
/g
C
Fig. 3.1. Stress blocks for concrete at ultimate limit states
26
12 May 2004 11:52:28
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:42
the bending resistance of cross-sections of columns, so a correction factor
M
is given in
clause 6.7.3.6(1). See also the comments on clauses 6.2.1.2(2) and 6.7.3.6.
Clause 3.1(2) Clause 3.1(2) limits the scope of EN 1994-1-1 to the strength range C20/25 to C60/75 for
normal concrete and from LC20/22 to LC60/66 for lightweight concrete. These ranges
are narrower than those given in EN 1992-1-1 because there is limited knowledge and
experience of the behaviour of composite members with weak or very strong concrete. This
applies, for example, to the load/slip properties of shear connectors, the redistribution of
moments in continuous beams and the resistance of columns. The use of rectangular stress
blocks for resistance to bending (clause 6.2.1.2(d)) relies on the strain capacity of the
materials. The relevant property of concrete,
cu3
in Table 3.1 of EN 1992-1-1, is 0.0035 for
classes up to C50/60, but is only 0.0026 for class C90/105.
Shrinkage
Clause 3.1(3)
Clause 3.1(4)
The shrinkage of concrete referred to in clause 3.1(3) is the drying shrinkage that occurs
after setting. It does not include the plastic shrinkage that precedes setting, nor autogenous
shrinkage. The latter develops during hardening of the concrete (clause 3.1.4(6) of
EN 1992-1-1), and is that which occurs in enclosed or sealed concrete, as in a concrete-filled
tube, where no loss of moisture occurs. Clause 3.1(4) permits its effect on stresses and
deflections to be neglected, but does not refer to crack widths. It has little influence on
cracking due to direct loading, and the rules for initial cracking (clause 7.4.2) take account of
its effects.
The shrinkage strains given in clause 3.1.4(6) of EN 1992-1-1 are significantly higher than
those given in BS 8110. Taking grade C40/50 concrete as an example, with dry environment
(relative humidity 60%), the final drying shrinkage could be 400 10
6
, plus autogenous
shrinkage of 75 10
6
.
The value in ENV 1994-1-1 was 325 10
6
, based on practice and experience. In the
absence of adverse comment on the ENV, this value is repeated in Annex C (informative) of
EN 1994-1-1, with a Note below clause 3.1 that permits other values to be given in National
Annexes. In the absence of this Note, a design using a value from Annex C, confirmed in a
National Annex, would not be in accordance with the Eurocodes. This is because normative
clause 3.1.4(6) of EN1992-1-1 takes precedence over an informative National Annex, and all
variations in National Annexes have to be permitted in this way.
In typical environments in the UK, the influence of shrinkage of normal-weight concrete
on the design of composite structures for buildings is significant only in:
very tall structures
very long structures without movement joints
the prediction of deflections of beams with high span/depth ratios (clause 7.3.1 (8)).
There is further comment on shrinkage in Chapter 5.
Creep
The provisions of EN 1992-1-1 on creep of concrete can be simplified for composite
structures for buildings, as discussed in comments on clause 5.4.2.2.
3.2. Reinforcing steel
Clause 3.2(1) Clause 3.2(1) refers to EN1992-1-1, which states in clause 3.2.2(3)P that its rules are valid for
specified yield strengths f
yk
up to 600 N/mm
2
.
The scope of clause 3.2 of EN 1992-1-1, and hence of EN 1994-1-1, is limited to
reinforcement, including wire fabrics with a nominal bar diameter of 5 mm and above, that
is, ribbed (high bond) and weldable. There are three ductility classes, fromA(the lowest) to
C. The requirements include the characteristic strain at maximum force, rather than the
15
CHAPTER 3. MATERIALS
27
12 May 2004 11:52:28
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:42
elongation at fracture used in past British standards. Clause 5.5.1(5) of EN1994-1-1 excludes
the use of Class A reinforcement in any composite cross-section in Class 1 or 2.
The minimum ductility properties of wire fabric given in Table C.1 of EN 1992-1-1 may
not be sufficient to satisfy clause 5.5.1(6) of EN 1994-1-1, as this requires demonstration of
sufficient ductility to avoid fracture when built into a concrete slab.
19
It has been found in
tests on continuous composite beams with fabric in tension that the cross-wires initiate
cracks in concrete, so that tensile strain becomes concentrated at the locations of the welds
in the fabric.
Clause 3.2(2) For simplicity, clause 3.2(2) permits the modulus of elasticity of reinforcement to be taken
as 210 kN/mm
2
, the value given in EN 1993-1-1 for structural steel, rather than 200 kN/mm
2
,
the value in EN 1992-1-1.
3.3. Structural steel
Clause 3.3(1)
Clause 3.3(2)
Clause 3.3(1) refers to EN 1993-1-1. This lists in its Table 3.1 steel grades with nominal yield
strengths up to 460 N/mm
2
, and allows other steel products to be included in National
Annexes. Clause 3.3(2) sets an upper limit of 460 N/mm
2
for use with EN1994-1-1. There has
been extensive research
2023
on the use in composite members of structural steels with yield
strengths exceeding 355 N/mm
2
. It has been found that some design rules need modification
for use with steel grades higher than S355, to avoid premature crushing of concrete. This
applies to:
redistribution of moments (clause 5.4.4(6))
rotation capacity (clause 5.4.5(4a))
plastic resistance moment (clause 6.2.1.2(2))
resistance of columns (clause 6.7.3.6(1)).
Thermal expansion
For the coefficient of linear thermal expansion of steel, clause 3.2.6 of EN 1993-1-1 gives a
value of 12 10
6
per C (also written in Eurocodes as /Kor K
1
). This is followed by a Note
that for calculating the structural effects of unequal temperatures in composite structures,
the coefficient may be taken as 10 10
6
per C, which is the value given for normal-weight
concrete in clause 3.1.3(5) of EN 1992-1-1 unless more accurate information is available.
Thermal expansion of reinforcement is not mentioned in EN 1992-1-1, presumably
because it is assumed to be the same as that of normal-weight concrete. For reinforcement in
composite structures the coefficient should be taken as 10 10
6
K
1
. This was stated in
ENV 1994-1-1, but is not in the EN.
Coefficients of thermal expansion for lightweight-aggregate concretes can range from
4 10
6
to 14 10
6
K
1
. Clause 11.3.2(2) of EN 1992-1-1 states that
The differences between the coefficients of thermal expansion of steel and lightweight aggregate
concrete need not be considered in design,
but steel here means reinforcement, not structural steel. The effects of the difference from
10 10
6
K
1
should be considered in design of composite members for situations where the
temperatures of the concrete and the structural steel could differ significantly.
3.4. Connecting devices
3.4.1. General
Reference is made to EN 1993, Eurocode 3: Design of Steel Structures, Part 1.8: Design of
Joints,
24
for information relating to fasteners, such as bolts, and welding consumables.
Provisions for other types of mechanical fastener are given in clause 3.3.2 of EN 1993-1-3.
25
Commentary on joints is given in Chapters 8 and 10.
16
28
12 May 2004 11:52:28
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:42
3.4.2. Stud shear connectors
Headed studs are the only type of shear connector for which detailed provisions are given in
EN 1994-1-1, in clause 6.6.5.7. Any other method of connection must satisfy clause 6.6.1.1.
The use of adhesives on a steel flange is unlikely to be suitable.
Clause 3.4.2 Clause 3.4.2 refers to EN 13918, Welding Studs and Ceramic Ferrules for Arc Stud
Welding.
26
This gives minimum dimensions for weld collars. Other methods of attaching
studs, such as spinning, may not provide weld collars large enough for the resistances of studs
given in clause 6.6.3.1(1) to be applicable.
Shear connection between steel and concrete by bond or friction is permitted only in
accordance with clause 6.7.4, for columns, and clauses 9.1.2.1 and 9.7, for composite slabs.
3.5. Profiled steel sheeting for composite slabs in buildings
The title includes in buildings, as this clause and other provisions for composite slabs are
not applicable to composite bridges.
Clause 3.5 The materials for profiled steel sheeting must conformto the standards listed in clause 3.5.
There are at present no EN standards for the wide range of profiled sheets available. Such
standards should include tolerances on embossments and indentations, as these influence
resistance to longitudinal shear. Tolerances on embossments, given for test specimens in
clause B.3.3(2), provide guidance.
The minimum bare metal thickness has been controversial, and in EN 1994-1-1 is subject
to National Annexes, with a recommended minimum of 0.70 mm. The total thickness of zinc
coating in accordance with clause 4.2(3) is about 0.05 mm.
17
29
12 May 2004 11:52:28
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:46
CHAPTER 4
Durability
This chapter concerns the durability of composite structures. It corresponds to Section 4,
which has the following clauses:
General Clause 4.1
Profiled steel sheeting for composite slabs in buildings Clause 4.2
4.1. General
Almost all aspects of the durability of composite structures are covered by cross-references
to EN 1990, EN 1992 and EN 1993. The material-independent provisions, in clause
2.4 of EN 1990, require the designer to take into account 10 factors. These include the
foreseeable use of the structure, the expected environmental conditions, the design criteria,
the performance of the materials, the particular protective measures, the quality of
workmanship and the intended level of maintenance.
Clauses 4.2 and 4.4.1 of EN 1992-1-1 define exposure classes and cover to reinforcement.
A Note defines structural classes. These and the acceptable deviations (tolerances) for
cover may be modified in a National Annex. Clause 4.4.1.3 recommends an addition of
10 mm to the minimum cover to allow for the deviation.
As an example, a concrete floor of a multi-storey car park will be subject to the action of
chlorides in an environment consisting of cyclic wet and dry conditions. For these conditions
(designated class XD3) the recommended structural class is 4, giving a minimum cover for a
50 year service life of 45 mm plus a tolerance of 10 mm. This total of 55 mm can be reduced,
typically by 5 mm, where special quality assurance is in place.
Section 4 of EN 1993-1-1 refers to execution of protective treatments for steelwork.
If parts will be susceptible to corrosion, there is need for access for inspection and
maintenance. This will not be possible for shear connectors, and clause 4.1(2) of EN1994-1-1
refers to clause 6.6.5, which includes provisions for minimum cover.
4.2. Profiled steel sheeting for composite slabs in buildings
Clause 4.2(1)P
Clause 4.2(3)
For profiled steel sheeting, clause 4.2(1)P requires the corrosion protection to be adequate
for its environment. Zinc coating to clause 4.2(3) is sufficient for internal floors in a
non-aggressive environment. This implies that it may not provide sufficient durability for
use in a multi-storey car park or near the sea.
31
12 May 2004 11:52:28
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
CHAPTER 5
Structural analysis
Structural analysis may be performed at three levels: global analysis, member analysis, and
local analysis. This chapter concerns global analysis to determine deformations and internal
forces and moments in beams and framed structures. It corresponds to Section 5, which has
the following clauses:
Structural modelling for analysis Clause 5.1
Structural stability Clause 5.2
Imperfections Clause 5.3
Calculation of action effects Clause 5.4
Classification of cross-sections Clause 5.5
Wherever possible, analyses for serviceability and ultimate limit states use the same
methods. It is generally more convenient, therefore, to specify them together in a single
section, rather than to include them in Sections 6 and 7. For composite slabs, though, all
provisions, including those for global analysis, are given in Section 9.
The division of material between Section 5 and Section 6 (ultimate limit states) is not
always obvious. Calculation of vertical shear is clearly analysis, but longitudinal shear is in
Section 6. This is because its calculation for beams in buildings is dependent on the method
used to determine the resistance to bending. However, for composite columns, methods of
analysis and member imperfections are considered in clause 6.7.3.4. This separation
of imperfections in frames from those in columns requires care, and receives detailed
explanation after the comments on clause 5.4. The flow charts for global analysis (Fig. 5.1)
include relevant provisions from Section 6.
5.1. Structural modelling for analysis
5.1.1. Structural modelling and basic assumptions
General provisions are given in EN 1990. The clause referred to says, in effect, that models
shall be appropriate and based on established theory and practice and that the variables shall
be relevant.
Clause 5.1.1(2)
Composite members and joints are commonly used in conjunction with others of
structural steel. Clause 5.1.1(2) makes clear that this is the type of construction envisaged in
Section 5, which is aligned with and cross-refers to Section 5 of EN 1993-1-1 wherever
possible. Where there are significant differences between these two sections, they are
referred to here.
5.1.2. Joint modelling
Clause 5.1.2(2) The three simplified joint models listed in clause 5.1.2(2) simple, continuous and
semi-continuous are those given in EN 1993. The subject of joints in steelwork has its
33
12 May 2004 11:52:29
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
own Eurocode part, EN 1993-1-8.
24
For composite joints, its provisions are modified and
supplemented by Section 8 of EN 1994-1-1.
The first two joint models are those commonly used for beam-to-column joints in steel
frames. For each joint in the simple model, the location of the nominal pin relative to the
centre-line of the column, the nominal eccentricity, has to be chosen. This determines the
effective span of each beamand the bending moments in each column. Practice varies across
Europe, and neither EN 1993-1-1 nor EN 1994-1-1 gives values for nominal eccentricities.
Guidance may be given in a National Annex, or in other literature.
Clause 5.1.2(3)
In reality, most joints in buildings are neither simple (i.e. pinned) nor continuous. The
third model, semi-continuous, is appropriate for a wide range of joints with moment
rotation behaviours intermediate between simple and continuous. This model is rarely
applicable to bridges, so the cross-reference to EN 1993-1-8 in clause 5.1.2(3) is for
buildings. The provisions of EN 1993-1-8 are for joints subjected to predominantly static
loading (its clause 1.1(1)). They are applicable to wind loading on buildings, but not to
fatigue loading, which is covered in EN 1993-1-9 and in clause 6.8.
For composite beams, the need for continuity of slab reinforcement past the columns, to
control cracking, causes joints to transmit moments. For the joint to have no effect on the
analysis (from the definition of a continuous joint in clause 5.1.1(2) of EN 1993-1-8),
so much reinforcement and stiffening of steelwork are needed that the design becomes
uneconomic. Joints with some continuity are usually semi-continuous. Structural analysis
then requires prior calculation of the properties of joints, except where they can be treated as
simple or continuous on the basis of significant experience of previous satisfactory
performance in similar cases (clause 5.2.2.1(2) of EN 1993-1-8, referred to from clause
8.2.3(1)) or experimental evidence.
Clause 5.1.2(2) refers to clause 5.1.1 of EN 1993-1-8, which gives the terminology for the
semi-continuous joint model. For elastic analysis, the joint is semi-rigid. It has a rotational
stiffness, and a design resistance which may be partial strength or full strength, normally
meaning less than or greater than the bending resistance of the connected beam. If the
resistance of the joint is reached, then elasticplastic or rigid plastic global analysis is
required.
5.2. Structural stability
The following comments refer mainly to beam-and-column frames, and assume that the
global analyses will be based on elastic theory. The exceptions, in clauses 5.4.3, 5.4.4 and
5.4.5 are discussed later. All design methods must take account of errors in the initial
positions of joints (global imperfections) and in the initial geometry of members (member
imperfections); of the effects of cracking of concrete and of any semi-rigid joints; and of
residual stresses in compression members.
The stage at which each of these is considered or allowed for will depend on the software
being used, which leads to some complexity in clauses 5.2 to 5.4.
5.2.1. Effects of deformed geometry of the structure
Clause 5.2.1(2)P
Clause 5.2.1(3)
In its clause 1.5.6, EN 1990 defines types of analysis. First-order analysis is performed on
the initial geometry of the structure. Second-order analysis takes account of the deformations
of the structure, which are a function of its loading. Clearly, second-order analysis may
always be applied. With appropriate software increasingly available, second-order analysis
is the most straightforward approach. The criteria for neglect of second-order effects
given in clauses 5.2.1(2)P and 5.2.1(3) need not be considered. The analysis allowing for
second-order effects will usually be iterative but normally the iteration will take place within
the software. Methods for second-order analysis are described in textbooks such as that by
Trahair et al.
27
22
34
12 May 2004 11:52:29
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
Clause 5.2.1(3)
A disadvantage of second-order analysis is that, in general, the useful principle of
superposition does not apply. Clause 5.2.1(3) provides a basis for the use of first-order
analysis. The check is done for a particular load combination and arrangement. The
provisions in this clause are similar to those for elastic analysis in the corresponding clause in
EN 1993-1-1.
In an elastic frame, second-order effects are dependent on the nearness of the design
loads to the elastic critical load. This is the basis for expression (5.1), in which
cr
is defined as
the factor to cause elastic instability. This may be taken as the load factor at which
bifurcation of equilibriumoccurs. For a conventional beam-and-column frame, it is assumed
that the frame is perfect, and that only vertical loads are present, usually at their maximum
design values. These are replaced by a set of loads which produces the same set of member
axial forces without any bending. An eigenvalue analysis then gives the factor
cr
, applied to
the whole of the loading, at which the total frame stiffness vanishes, and elastic instability
occurs.
To sufficient accuracy,
cr
may also be determined by a second-order loaddeflection
analysis. The non-linear loaddeflection response approaches asymptotically to the elastic
critical value. Normally, though, it is pointless to use this method, as it is simpler to use the
same software to account for the second-order effects due to the design loads. Amore useful
method for
cr
is given in clause 5.2.2(1).
Unlike the corresponding clause in EN 1993-1-1, the check in clause 5.2.1(3) is not just
for a sway mode. This is because clause 5.2.1 is relevant not only to complete frames but also
to the design of individual composite columns (see clause 6.7.3.4). Such members may be
held in position against sway but still be subject to significant second-order effects due to
bowing.
Clause 5.2.1(4)P Clause 5.2.1(4)P is a reminder that the analysis needs to account for the reduction in
stiffness arising from cracking and creep of concrete and from possible non-linear behaviour
of the joints. Further remarks on how this should be done are made in the comments on
clauses 5.4.2.2, 5.4.2.3 and 8.2.2, and the procedures are illustrated in Fig. 5.1(b)(d). In
general, such effects are dependent on the internal moments and forces, and iteration is
therefore required. Manual intervention may be needed, to adjust stiffness values before
repeating the analysis. It is expected, though, that advanced software will be written for
EN 1994 to account automatically for these effects. The designer may of course make
assumptions, although care is needed to ensure these are conservative. For example,
assuming that joints have zero rotational stiffness (resulting in simply-supported composite
beams) could lead to neglect of the reduction in beam stiffness due to cracking. The overall
lateral stiffness would probably be a conservative value, but this is not certain. However, in a
frame with stiff bracing it will be worth first calculating
cr
, assuming joints are pinned and
beams are steel section only; it may well be found that this value of
cr
is sufficiently high for
first-order global analysis to be used.
Using elastic analysis, it is not considered necessary to account for slip (see clause
5.4.1.1(8)), provided that the shear connection is in accordance with clause 6.6.
5.2.2. Methods of analysis for buildings
Clause 5.2.2(1) Clause 5.2.2(1) refers to clause 5.2.1(4) of EN 1993-1-1 for a simpler check, applicable to
many structures for buildings. This requires calculation of sway deflections due to horizontal
loads only, and first-order analysis can be used to determine these deflections. It is assumed
that any significant second-order effects will arise only from interaction of column forces
with sway deflection. It follows that the check will only be valid if axial compression in beams
is not significant. Fig. 5.1(e) illustrates the procedure.
Clause 5.2.2(2) Even where second-order effects are significant, clause 5.2.2(2) allows these to be
determined by amplifying the results from a first-order analysis. No further information is
given, but clause 5.2.2(5) of EN 1993-1-1 describes a method for frames, provided that the
conditions in its clause 5.2.2(6) are satisfied.
23
CHAPTER 5. STRUCTURAL ANALYSIS
35
12 May 2004 11:52:29
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
Clause 5.2.2(3)
Clause 5.2.2(4)
Clause 5.2.2(5)
Clause 5.2.2(6)
Clause 5.2.2(7)
Clauses 5.2.2(3) to 5.2.2(7) concern the relationships between the analysis of the frame
and the stability of individual members. A number of possibilities are presented. If relevant
software is available, clause 5.2.2(3) provides a convenient route for composite columns,
because column design to clause 6.7 generally requires a second-order analysis. Usually,
though, the global analysis will not account for all local effects, and clause 5.2.2(4) describes
in general terms howthe designer should then proceed. Clause 5.2.2(5) refers to the methods
of EN 1994-1-1 for lateraltorsional buckling, which allow for member imperfections. This
applies also to local and shear buckling in beams, so imperfections in beams can usually be
omitted from global analyses.
In clause 5.2.2(6), compression members are referred to as well as columns, to include
composite members used in bracing systems and trusses. Further comments on clauses
5.2.2(3) to 5.2.2(7) are made in the sections of this guide dealing with clauses 5.5, 6.2.2.3, 6.4
and 6.7. Figure 5.1(a) illustrates how global and member analyses may be used, for a plane
frame including composite columns.
5.3. Imperfections
5.3.1. Basis
Clause 5.3.1(1)P Clause 5.3.1(1)P lists possible sources of imperfection. Subsequent clauses (and also clause
5.2) describe howthese should be allowed for. This may be by inclusion in the global analyses
or in methods of checking resistance, as explained above.
Clause 5.3.1(2) Clause 5.3.1(2) requires imperfections to be in the most unfavourable direction and form.
The most unfavourable geometric imperfection normally has the same shape as the lowest
buckling mode. This can sometimes be difficult to find; but it can be assumed that this
condition is satisfied by the Eurocode methods for checking resistance that include effects of
member imperfections (see comments on clause 5.2.2).
5.3.2. Imperfections in buildings
Clause 5.3.2.1(1)
Generally, an explicit treatment of geometric imperfections is required for composite
frames. In both EN 1993-1-1 and EN 1994-1-1 the values are equivalent rather than
measured values (clause 5.3.2.1(1)), because they allow for effects such as residual stresses,
in addition to imperfections of shape. The codes define both global sway imperfections for
frames and local bow imperfections of individual members (meaning a span of a beam or the
length of a column between storeys).
Clause 5.3.2.1(2)
The usual aim in global analysis is to determine the action effects at the ends of members.
If necessary, a member analysis is performed subsequently, as illustrated in Fig. 5.1(a); for
example to determine the local moments in a column due to transverse loading. Normally
the action effects at members ends are affected by the global sway imperfections but not
significantly by the local bow imperfections. In both EN 1993-1-1 and EN 1994-1-1 the effect
of a bow imperfection on the end moments and forces may be neglected in global analysis if
the design normal force N
Ed
does not exceed 25% of the Euler buckling load for the
pin-ended member (clause 5.3.2.1(2)).
Clause 5.3.2.1(3) Clause 5.3.2.1(3) is a reminder that an explicit treatment of bow imperfections is always
required for checking individual composite columns, because the resistance formulae are for
cross-sections only and do not allow for action effects caused by these imperfections.
Clause 5.3.2.1(4) The reference to EN 1993-1-1 in clause 5.3.2.1(4) leads to two alternative methods of
allowing for imperfections in steel columns. One method includes all imperfections in the
global analysis. Like the method just described for composite columns, no individual stability
check is then necessary.
The alternative approach is that familiar to most designers. Member imperfections are
not accounted for in the global analysis. The stability of each member is then checked using
end moments and forces from that analysis, with buckling formulae that take account of
imperfections.
24
36
12 May 2004 11:52:29
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
25
Do second-order global analysis
Note: These flow charts are for a
particular load combination and
arrangement for ultimate limit
states, for a beam-and-column
type plane frame in its own plane,
and for global analyses in which
allowances may be needed for
creep, cracking of concrete, and
the behaviour of joints.
EC3 means EN 1993-1-1
Yes
Yes
Check beams for lateraltorsional buckling, using resistance formulae that include member imperfections
(clauses 5.2.2(5) and 5.3.2.3(2))
Were member imperfections for columns included in the global analysis?
Is the member in axial compression only?
Verify column cross-sections to clause 6.7.3.6 or 6.7.3.7, from clause 5.2.2(6)
Do second-order analysis for
each column, with end action-
effects from the global
analysis, including member
imperfections, from
clause 5.2.2(6)
No
Use buckling curves
that account for
second-order effects
and member
imperfections to
check the member
(clause 6.7.3.5)
No
Yes
Do first-order global analysis
Determine frame imperfections as equivalent horizontal forces,
to clause 5.3.2.2, which refers to clause 5.3.2 of EC3. Neglect
member imperfections (clause 5.3.2.1(2))
No
Note: for columns,
more detail is given
in Fig. 6.36
Go to Fig. 5.1(e), on
methods of global analysis
For each column, estimate N
Ed
, find l to clause 5.3.2.1(2).
Determine member imperfection for each column (to clause
5.3.2.3) and where condition (2) of clause 5.3.2.1(2) is not
satisfied, include these these imperfections in second-order
analysis
Is second-order analysis
needed for global analysis?
Go to
Fig. 5.1(b),
on creep
Go to
Fig. 5.1(c),
on cracking
Determine appropriate stiffnesses, making allowance for
cracking and creep of concrete and for behaviour of joints
Go to
Fig. 5.1(d),
on joints
-
Fig. 5.1. Global analysis of a plane frame
(a) Flow chart, global analysis of a plane frame with composite columns
37
12 May 2004 11:52:30
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
26
Does clause 5.4.2.2(11), on use of a nominal modular ratio, apply?
No Yes
For composite beams, assume
an effective modulus (clause
5.4.2.2(11)) and determine a
nominal modular ratio, n
For composite beams, determine modular ratios n
0
for
short-term loading and n
L
for permanent loads. For a
combination of short-term and permanent loading,
estimate proportions of loading and determine a modular
ratio n from n
0
and n
L
For each composite column, estimate the proportion of permanent to total normal
force, determine effective modulus E
c, eff
(clause 6.7.3.3(4)), and hence the design
effective stiffness, (EI)
eff, II
, from clause 6.7.3.4(2)
Determine cracked stiffness for each composite column, to clause 6.7.3.4
Yes No
No
Yes
Yes
No
Is the frame braced? Assume uncracked beams
Make appropriate allowances for creep (clause 5.4.2.2) and
flexibility of joints (clause 8.2.2)
Analyse under characteristic combinations to determine internal
forces and moments (clause 5.4.2.3 (2)) and determine cracked
regions of beams
Do adjacent spans satisfy clause 5.4.2.3(3)?
Are internal joints rigid?
Assume cracked lengths for
beams (clause 5.4.2.3(3))
Assign appropriate stiffnesses for beams
No
Yes
Yes
No
Can the joint be classified on the basis of experimental evidence or significant experience of
previous performance in similar cases? (See EN 1993-1-8 (clause 5.2.2.1(2))
In the model for the frame, assign appropriate rotational stiffness to the joint
Determine rotational stiffness ((clause 8.2.2 and EN 1993-1-8 (clause 5.1.2))
Determine classification by stiffness
(clause 8.2.3 and EN 1993-1-8 (clause 5.2))
Calculate initial rotational stiffness, S
j, ini
(clause 8.3.3, Annex B and
EN 1993-1-8 (clause 6.3))
Is the joint nominally-pinned or rigid?
Fig. 5.1. (Contd)
(b) Supplementary flow chart, creep
(c) Supplementary flow chart, cracking of concrete
(d) Supplementary flow chart, stiffness of joints, for elastic global analysis only
38
12 May 2004 11:52:30
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
Global imperfections
Clause 5.3.2.2(1) Clause 5.3.2.2(1) refers to clause 5.3.2 of EN 1993-1-1. This gives values for the global sway
imperfections, describes howimperfections may be replaced by equivalent horizontal forces,
and permits these to be disregarded if the real horizontal forces (e.g. due to wind) are
significant relative to the design vertical load.
Member imperfections
Clause 5.3.2.3(1) Clause 5.3.2.3(1) refers to Table 6.5, which gives the amplitudes of the central bow of a
member designed as straight. It makes little difference whether the curve is assumed to be a
half sine wave or a circular arc. These single-curvature shapes are assumed irrespective of
the shape of the bending-moment diagramfor the member, but the designer has to decide on
which side of the member the bow is present.
Clause 5.3.2.3(2)
Clause 5.3.2.3(3)
Clauses 5.3.2.3(2) and 5.3.2.3(3) refer to member imperfections that need not be included
in global analyses. If they are, only cross-section properties are required for checking
resistances.
5.4. Calculation of action effects
5.4.1. Methods of global analysis
Clause 5.4.1.1
EN 1990 defines several types of analysis that may be appropriate for ultimate limit states.
For global analysis of buildings, EN 1994-1-1 gives four methods: linear elastic analysis (with
or without redistribution), non-linear analysis and rigid plastic analysis. Clause 5.4.1.1 gives
guidance on matters common to more than one method.
Clause 5.4.1.1(1)
For reasons of economy, plastic (rectangular stress block) theory is preferred for checking
the resistance of cross-sections. In such cases, clause 5.4.1.1(1) allows the action effects to
27
Is axial compression in the beams
not significant?
Determine a
cr
by use of appropriate software
or from the literature
Second-order effects to be
taken into account
Determine a
cr
for each storey by EN 1993-1-1 (clause 5.2.1(4)).
Determine the minimum value
First-order analysis is
acceptable
Yes No
Yes
Yes
No
Is the structure a beam-and-column plane frame?
Determine appropriate allowances for cracking and creep of concrete and for behaviour of joints,
clause 5.2.1(4). Assign appropriate stiffnesses to the structure. (See Figs 5.1(b) to (d))
Is a
cr
10?
Fig. 5.1. (Contd)
(e) Supplementary flow chart, choice between first-order and second-order global analysis
39
12 May 2004 11:52:31
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
be determined by elastic analysis; for composite structures this method has the widest
application.
Clause 5.4.1.1(2) Clause 5.4.1.1(2) makes clear that for serviceability limit states, elastic analysis should
be used. Linear elastic analysis is based on linear stress/strain laws, but for composite
structures, cracking of concrete needs to be considered (clause 5.4.2.3). Other possible
non-linear effects include the flexibility of semi-continuous joints (Section 8).
Clause 5.4.1.1(4)
Clause 5.4.1.1(5)
Clause 5.4.1.1(6)
Methods for satisfying the principle of clause 5.4.1.1(4) are given for local buckling in
clauses 5.4.1.1(5) and 5.4.1.1(6), and for shear lag in concrete in clause 5.4.1.2. Reference is
made to the classification of cross-sections. This is the established method of taking account
of local buckling of steel elements in compression. It determines the available methods of
global analysis and the basis for resistance to bending. The classification system is defined in
clause 5.5.
There are several reasons
28,29
why the apparent incompatibility between the methods used
for analysis and for resistance is accepted, as stated in clause 5.4.1.1(1). There is no such
incompatibility for Class 3 sections, as resistance is based on an elastic model. For Class 4
sections (those in which local buckling will occur before the attainment of yield), clause
5.4.1.1(6) refers to clause 2.2 of EN1993-1-1, which gives a general reference to EN1993-1-5
(plated structural elements).
30
This defines those situations in which the effects of shear lag
and local buckling in steel plating can be ignored in global analyses.
Clause 5.4.1.1(7) Clause 5.4.1.1(7) reflects a general concern about slip, shared by EN 1993-1-1. For
composite joints, clause A.3 gives a method to account for deformation of the adjacent shear
connectors.
Clause 5.4.1.1(8)
Composite beams have to be provided with shear connection in accordance with clause
6.6. Clause 5.4.1.1(8) therefore permits internal moments and forces to be determined
assuming full interaction. For composite columns, clause 6.7.3.4(2) gives an effective flexural
stiffness for use in global analysis.
Shear lag in concrete flanges, and effective width
Accurate values for effective width of an uncracked elastic flange can be determined by
numerical analysis. They are influenced by many parameters and vary significantly along
each span. They are increased both by inelasticity and by cracking of concrete. For the
bending resistance of a beam, underestimates are conservative, so values in codes have often
been based on elastic values.
Clause 5.4.1.2 The simplified values given in clause 5.4.1.2 of EN 1994-1-1 are very similar to those used
in BS 5950: Part 3.1:1990
31
and BS ENV 1994-1-1:1994. The values are generally lower than
those in EN 1992-1-1 for reinforced concrete T-beams. To adopt those would often increase
the number of shear connectors. Without evidence that the greater effective widths are any
more accurate, the established values for composite beams have mainly been retained.
The effective width is based on the distance between points of contraflexure. In
EN 1992-1-1, the sum of the distances for sagging and hogging regions equals the span of
the beam. In reality, points of contraflexure are dependent on the load arrangement.
EN 1994-1-1 therefore gives a larger effective width at an internal support, to reflect the
critical load arrangement for this cross-section. In sagging regions, the assumed distances
between points of contraflexure are the same in both codes.
Clause 5.4.1.2(4)
Although there are significant differences between effective widths for supports and
mid-span regions, it is possible to ignore this in elastic global analysis (clause 5.4.1.2(4)). This
is because shear lag has limited influence on the results.
Clause 5.4.1.2(5)
A small difference from earlier codes for buildings concerns the width of steel flange
occupied by the shear connectors. Clause 5.4.1.2(5) allows this width to be included within
the effective region. Alternatively, it may be ignored (clause 5.4.1.2(9)).
Clause 5.4.1.2(8) Clause 5.4.1.2(8) is a reminder that Fig. 5.1 is based on continuous beams. Although clause
8.4.2.1(1) refers to it, clause 5.4.1.2 does not define the effective flange width adjacent to an
external column. Figure 5.1 may be used as a guide, or the width may be taken as the width
occupied by slab reinforcement that is anchored to the column (see Fig. 8.2).
28
40
12 May 2004 11:52:31
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
Example 5.1: effective width of concrete flange
The notation and method used are those of clause 5.4.1.2. A continuous beam of uniform
section consists of two spans and a cantilever, as shown in Fig. 5.2. Values for b
eff
are
required for the mid-span regions AB and CD, for the support regions BC and DE, and
for the support at A.
The calculation is shown in Table 5.1. The result for support A is found from equations
(5.4) and (5.5), as follows:
b
eff
= 0.2 + [0.55 + (0.025 6.8/0.4)] 0.4 +
[0.55 + (0.025 6.8/0.85)] 0.85 = 1.23 m
Global analysis may be based on stiffness calculated using the results for AB and CD,
but the difference between them is so small that member ABCDE would be analysed as a
beam of uniform section.
5.4.2. Linear elastic analysis
The restrictions on the use of rigid plastic global analysis (plastic hinge analysis), in clause
5.4.5, are such that linearelastic global analysis will often be used for composite frames.
Creep and shrinkage
Clause 5.4.2.2 There are some differences in clause 5.4.2.2 from previous practice in the UK. The elastic
modulus for concrete under short-term loading, E
cm
, is a function of the grade and density of
the concrete. For normal-weight concrete, it ranges from 30 kN/mm
2
for grade C20/25
to 39 kN/mm
2
for grade C60/75. With E
a
for structural steel given as 210 kN/mm
2
, the
short-term modular ratio, given by n
0
= E
a
/E
cm
, thus ranges from 7 to 5.3.
Figure 5.1(b) illustrates the procedure to allow for creep in members of a composite
frame. The proportion of loading that is permanent could be obtained by a preliminary
global analysis, but in many cases this can be estimated by simpler calculations.
For composite beams in structures for buildings where first-order global analysis is
acceptable (the majority), clause 5.4.2.2(11) allows the modular ratio to be taken as 2n
0
for
both short-term and long-term loading an important simplification, not given in BS 5950.
The only exceptions are:
29
Table 5.1. Effective width of the concrete flange of a composite T-beam
Region
AB BC CD DE Support A
L
e
(from Fig. 5.2) (m) 6.80 4.5 7.0 4.0 6.80
L
e
/8 (m) 0.85 0.562 0.875 0.50
b
e1
(m) 0.40 0.4 0.4 0.4 0.40
b
e2
(m) 0.85 0.562 0.875 0.50 0.85
b
eff
(m) 1.45 1.162 1.475 1.10 1.23
0.4 0.2 1.4 8 10 2
A B C D E
Fig. 5.2. Worked example: effective width
41
12 May 2004 11:52:31
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
structures where second-order global analysis is required by clause 5.2
structures for buildings mainly intended for storage
structures prestressed by controlled imposed deformations this would apply, for
example, to the bending of steel beams by jacking before concrete is cast around one of
the flanges.
Where the conditions of clause 5.4.2.2(11) do not apply, the modular ratio for use in
analyses for the effects of long-term loading, n
L
, depends on the type of loading and the
creep coefficient
t
. This coefficient depends on both the age of the concrete on first loading,
t
0
, and on the age at the time considered in the analysis, which is normally taken as infinity.
The use of this method is excluded for members with both flanges composite; but as these
occur mainly in bridges, no alternative is given in Part 1.1.
Clause 5.4.2.2(3) Although clause 5.4.2.2(4) gives the age of loading by effects of shrinkage as 1 day, clause
5.4.2.2(3) allows one mean value of t
0
to be assumed. If, for example, unpropped
construction is used for floor slabs, this might be taken as the age at which they could be
subjected to non-trivial imposed loads. These are likely to be construction loads.
It makes quite a difference whether this age is assumed to be (for example) 2 weeks or
2 months. From clause 5.4.2.2(2), the values for normal-weight concrete are found from
clause 3.1.4 of EN 1992-1-1. Suppose that normal cement is used for grade C25/30 concrete,
that the building will be centrally heated, so inside conditions apply, and that composite
floor slabs with a mean concrete thickness of 100 mm are used. Only one side of the slabs is
exposed to drying, so the notional size is 200 mm. The increase in t
0
from 14 to 60 days
reduces the creep coefficient from 3.0 to about 2.1.
The effect of type of loading is introduced by the symbol
L
in the equation
n
L
= n
0
(1 +
L
t
) (5.6)
The reason for taking account of it is illustrated in Fig. 5.3. This shows three schematic curves
of the change of compressive stress in concrete with time. The top one, labelled S, is typical
of stress caused by the increase of shrinkage with time. Concrete is more susceptible to creep
when young, so there is less creep than for the more uniform stress caused by permanent
loads (line P). The effects of imposed deformations can be significantly reduced by creep
when the concrete is young, so the curve is of type ID. The creep multiplier
L
has the values
0.55, 1.1 and 1.5, respectively, for these three types of loading. The value for permanent
loading on reinforced concrete is 1.0. It is increased to 1.1 for composite members because
the steel component does not creep. Stress in concrete is reduced by creep less than it would
be in a reinforced member, so there is more creep.
These application rules are based mainly on extensive theoretical work on composite
beams of many sizes and proportions,
32
and find application more in design of composite
bridges, than in buildings.
Clause 5.4.2.2(6) The time-dependent secondary effects due to creep referred to in clause 5.4.2.2(6) are most
unlikely to be found in buildings. Their calculation is quite complex, and is explained, with an
example, in Johnson and Hanswille.
33
30
Time
S
P
ID
0
1.0
s
c
/s
c, 0
Fig. 5.3. Time-dependent compressive stress in concrete, for three types of loading
42
12 May 2004 11:52:32
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
For creep in columns, clause 5.4.2.2(9) refers to clause 6.7.3.4(2), which in turn refers to an
effective modulus for concrete given in clause 6.7.3.3(4). If separate analyses are to be made
for long-term and short-term effects, clause 6.7.3.3(4) can be used, assuming that the ratio of
permanent to total load is 1.0 and 0, respectively.
Shrinkage of concrete
For the determination of shrinkage strain, reference should be made to the commentary on
clause 3.1. The effects in columns are unimportant, except in very tall structures. In beams
with the slab above the steel member, shrinkage causes sagging curvature. This is its primary
effect, which is reduced almost to zero where the concrete slab is cracked through its
thickness.
In continuous beams, the primary curvature is incompatible with the levels of the
supports. It is counteracted by bending moments caused by changes in the support reactions,
which increase at internal supports and reduce at end supports. The moments and the
associated shear forces are the secondary effects of shrinkage.
Clause 5.4.2.2(7)
Clause 5.4.2.2(8)
Clause 5.4.2.2(7) allows both types of effect to be neglected at ultimate limit states in a
beam with all cross-sections in Class 1 or 2, unless its resistance to bending is reduced
by lateraltorsional buckling. This restriction can be significant. Clause 5.4.2.2(8) allows
the option of neglecting primary curvature in cracked regions.
34
This complicates the
determination of the secondary effects, because the extent of the cracked regions has to be
found, and the beam then has a non-uniform section. It may be simpler not to take the
option, even though the secondary hogging bending moments at internal supports are then
higher. These moments, being a permanent effect, enter into all load combinations, and may
influence the design of what is often a critical region.
The long-term effects of shrinkage are significantly reduced by creep. In the example
above, on creep of concrete,
t
= 3 for t
0
= 14 days. For shrinkage, with t
0
= 1 day, clause
3.1.4 of EN 1992-1-1 gives
t
= 5, and equation (5.6) gives the modular ratio as:
n = n
0
(1 + 0.55 5) = 3.7n
0
Where it is necessary to consider shrinkage effects within the first year or so after casting,
a value for the relevant free shrinkage strain can be obtained from clause 3.1.4(6) of
EN 1992-1-1.
The influence of shrinkage on serviceability verifications is dealt with in Chapter 7.
Effects of cracking of concrete
Clause 5.4.2.3 Clause 5.4.2.3 is applicable to both serviceability and ultimate limit states. Figure 5.1(c)
illustrates the procedure.
In conventional composite beams with the slab above the steel section, cracking of
concrete reduces the flexural stiffness in hogging moment regions, but not in sagging
regions. The change in relative stiffness needs to be taken into account in elastic global
analysis. This is unlike analysis of reinforced concrete structures, where cracking occurs
in both hogging and sagging bending, and uncracked cross-sections can be assumed
throughout.
Clause 5.4.2.3(2)
Clause 5.4.2.3(3)
EN 1994-1-1 provides several different methods to allow for cracking in beams. This is
because its scope is both general and buildings. Clause 5.4.2.3(2) provides a general
method. This is followed in clause 5.4.2.3(3) by a simplified approach of limited application.
For buildings, a further method is given separately in clause 5.4.4.
In the general method, the first step is to determine the expected extent of cracking in
beams. The envelope of moments and shears is calculated for characteristic combinations
of actions, assuming uncracked sections and including long-term effects. The section is
assumed to crack if the extreme-fibre tensile stress in concrete exceeds twice the mean value
of the axial tensile strength given by EN 1992-1-1. The cracked stiffness is then adopted for
such sections, and the structure re-analysed. This requires the beams with cracked regions to
be treated as beams of non-uniform section.
31
43
12 May 2004 11:52:32
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
The uncracked and cracked flexural stiffnesses E
a
I
1
and E
a
I
2
are defined in clause 1.5.2.
Steel reinforcement is normally neglected in the calculation of I
1
.
The reasons why stiffness is not reduced to the cracked value until an extreme-fibre stress
of twice the mean tensile strength of the concrete is reached, are as follows:
the concrete is likely to be stronger than specified
reaching f
ctm
at the surface may not cause the slab to crack right through, and even if it
does, the effects of tension stiffening are significant at the stage of initial cracking
until after yielding of the reinforcement, the stiffness of a cracked region is greater than
E
a
I
2
, because of tension stiffening between the cracks
the calculation uses an envelope of moments, for which regions of slab in tension are
more extensive than they are for any particular loading.
Clause 5.4.2.3(3)
Clause 5.4.2.3(4)
Clause 5.4.2.3(5)
Clauses 5.4.2.3(3) to 5.4.2.3(5) provide a non-iterative method, but one that is applicable
only to some situations. These include conventional continuous composite beams, and
beams in braced frames. The cracked regions could differ significantly from the assumed
values in a frame that resists wind loading by bending. Where the conditions are not satisfied,
the general method of clause 5.4.2.3 (2) should be used.
Cracking affects the stiffness of a frame, and therefore needs to be considered in the
criteria for use of first-order analysis (clauses 5.2.1(3) and 5.2.2(1)). For braced frames
within the scope of clause 5.4.2.3(3), the cracked regions in beams are of fixed extent, and the
effective stiffness of the columns is given by clause 6.7.3.4(2). The corresponding value of the
elastic critical factor
cr
can therefore be determined prior to analysis under the design loads.
It is then worth checking if second-order effects can be neglected.
For unbraced frames, the extent of the cracking can only be determined from analysis
under the design loads. This analysis therefore needs to be carried out before the criteria
can be checked. It is more straightforward to carry out a second-order analysis, without
attempting to prove whether or not it is strictly necessary. Where second-order analysis is
necessary, strictly the extent of cracking in beams should take account of the second-order
effects. However, as this extent is based on the envelope of internal forces and moments for
characteristic combinations, these effects may not be significant.
The encasement in clause 5.4.2.3(5) is a reference to the partially encased beams defined
in clause 6.1.1(1)P. Fully encased beams are outside the scope of EN 1994-1-1.
Temperature effects
Clause 5.4.2.5(2) Clause 5.4.2.5(2) states that temperature effects, specified in EN1991-1-5,
35
may normally be
neglected in analyses for certain situations. Its scope is narrow because it applies to all
composite structures, not buildings only. It provides a further incentive to select steel
sections for beams that are not weakened by lateraltorsional buckling.
Study of the factors of Annex A1 of EN 1990
36
for combinations of actions for buildings
will show, for many projects, that temperature effects do not influence design. This is
illustrated for the design action effects due to the combination of imposed load (Q) with
temperature (T), for a building with floors in category B, office areas. Similar comments
apply to other combinations of actions and types of building.
The combination factors recommended in clause A1.2.2(1) of EN 1990 are given in Table
5.2. It is permitted to modify these values in a National Annex. For ultimate limit states, the
combinations to be considered, in the usual notation and with the recommended
F
factors,
are
32
Table 5.2. Combination factors for imposed load and temperature
Action y
0
y
1
y
2
Imposed load, building in category B 0.7 0.5 0.3
Temperature (non-fire) in buildings 0.6 0.5 0
44
12 May 2004 11:52:32
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
1.35G
k
+ 1.5(Q
k
+ 0.6T
k
) and 1.35G
k
+ 1.5(T
k
+ 0.7Q
k
)
The second one, with T leading, governs only where T
k
> 0.75Q
k
. Normally, action effects
due to temperature are much smaller than those due to imposed load, and additional action
effects resulting from the inclusion of T in the first combination are not significant.
For serviceability limit states, much depends on the project. Note 2 to clause 3.4(1)P of
EN 1990 states: Usually the serviceability requirements are agreed for each individual
project. Similarly, clause A1.4.2(2) of Annex A1 of EN 1990 states, for buildings: The
serviceability criteria should be specified for each project and agreed with the client. Note:
The serviceability criteria may be defined in the National Annex.
There are three combinations of actions given in EN 1990 for serviceability limit states:
characteristic, frequent and quasi-permanent. The first of these uses the same combination
factors
0
as for ultimate limit states, and the comments made above therefore apply. The
quasi-permanent combination is normally used for long-term effects, and temperature is
therefore not included.
For the frequent combination, the alternatives are:
G
k
+ 0.5Q
k
and G
k
+ 0.5T
k
+ 0.3Q
k
The second one governs only where T
k
> 0.4Q
k
.
This example suggests that unless there are members for which temperature is the most
severe action, as can occur in some industrial structures, the effects T are unlikely to
influence verifications for buildings.
Prestressing by controlled imposed deformations
Clause 5.4.2.6(2) Clause 5.4.2.6(2) draws attention to the need to consider the effects of deviations of
deformations and stiffnesses from their intended or expected values. If the deformations are
controlled, clause 5.4.2.6(2) permits design values of internal forces and moments arising
fromthis formof prestressing to be calculated fromthe characteristic or nominal value of the
deformation, which will usually be the intended or measured value.
The nature of the control required is not specified. It should take account of the sensitivity
of the structure to any error in the deformation.
Prestressing by jacking of supports is rarely used in buildings, as the subsequent loss of
prestress can be high.
5.4.3. Non-linear global analysis
Clause 5.4.3 Clause 5.4.3 adds little to the corresponding clauses in EN 1992-1-1 and EN 1993-1-1, to
which it refers. These clauses give provisions, mainly principles, that apply to any method of
global analysis that does not conform to clause 5.4.2, 5.4.4 or 5.4.5. They are relevant, for
example, to the use of finite-element methods.
There is some inconsistency in the use of the term non-linear in the Eurocodes. The
notes to clauses 1.5.6.6 and 1.5.6.7 of EN 1990 make clear that all of the methods of
global analysis defined in clauses 1.5.6.6 to 1.5.6.11 (which include plastic methods) are
non-linear in Eurocode terminology. Non-linear in these clauses refers to the deformation
properties of the materials.
Moderate geometrical non-linearity, such as can occur in composite structures, is allowed
for by using analyses defined as second-order. The much larger deformations that can
occur, for example, in some cable-stayed structures, need special treatment.
In clause 5.4 of EN 1993-1-1, global analyses are either elastic or plastic, and plastic
includes several types of non-linear analysis. The choice between these alternative methods
should take account of the properties of composite joints given in Section 8 of EN1994-1-1.
Clause 5.4.3(1) Clause 5.7 in EN 1992-1-1 referred to from clause 5.4.3(1) is Non-linear analysis, which
adds little new information.
In EN1994-1-1, non-linear analysis, clause 5.4.3, and rigid plastic analysis, clause 5.4.5, are
treated as separate types of global analysis, so that clause 5.4.3 is not applicable where clause
33
45
12 May 2004 11:52:32
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
5.4.5 is being followed. The term non-linear is used also for a type of resistance, in clause
6.2.1.4.
5.4.4. Linear elastic analysis with limited redistribution for buildings
The concept of redistribution of moments calculated by linear-elastic theory is well
established in the design of concrete and composite framed structures. It makes limited
allowance for inelastic behaviour, and enables the size of a design envelope of bending
moments (from all relevant arrangements of variable loads) to be reduced. In composite
beams it is generally easier to provide resistance to bending in mid-span regions than at
internal supports. The flexural stiffness at mid-span is higher, sometimes much higher, than
at internal supports, so that uncracked global analyses overestimate hogging bending
moments in continuous beams. A flow chart for this clause is given in Fig. 5.4.
Clause 5.4.4(1) Clause 5.4.4(1) refers to redistribution in continuous beams and frames, but composite
columns are not mentioned. At a beamcolumn intersection in a frame, there are usually
bending moments in the column, arising from interaction with the beam. Redistribution for
the beam may be done by assuming it to be continuous over simple supports. If the hogging
moments are reduced, the bending moments in the column should be left unaltered. If
hogging moments are increased, those in the column should be increased in proportion.
The clause is applicable provided second-order effects are not significant. Inelastic
behaviour results in loss of stiffness, but EN 1994-1-1 does not require this to be taken into
account when determining whether the clause is applicable. Although there is considerable
experience in using expression (5.1) as a criterion for rigid plastic global analysis of steel
frames with full-strength joints, to allowfor non-linear material properties, clause 5.2.1(3) of
EN 1993-1-1 gives the more severe limit,
cr
15. This limit is a nationally determined
parameter. EN 1994-1-1 retains the limit
cr
10, but account should be taken of cracking
and creep of concrete and the behaviour of joints.
Clause 5.4.4(2) One of the requirements of clause 5.4.4(2) is that redistribution should take account of all
types of buckling. Where the shear resistance of a web is reduced to below the plastic value
V
pl, Rd
to allow for web buckling, and the cross-section is not in Class 4, it would be prudent
either to design it for the vertical shear before redistribution, or to treat it as if in Class 4.
Although the provisions of clause 5.4.4 appear similar to those of clause 5.2.3.1 of
BS 5950-3-1, there are important differences. Some of these arise because the scope of the
British standard is limited to conventional composite beams in normal building structures.
The Eurocode provisions are not applicable where:
(1) second-order global analysis is required
(2) a serviceability or fatigue limit state is being verified
(3) the structure is an unbraced frame
(4) semi-rigid or partial-strength joints are used
(5) beams are partially encased, unless the rotation capacity is sufficient or encasement in
compression is neglected
(6) the depth of a beam varies within a span
(7) a beam with steel of grade higher than S355 has cross-sections in Class 3 or 4
(8) the resistance of the beam is reduced to allow for lateraltorsional buckling.
The reasons for these exclusions are now briefly explained:
(1) redistribution arises frominelastic behaviour, which lessens the stiffness of the structure
and threatens stability
(2) fatigue verification is based on elastic analysis
(3) the amounts of redistribution given in Table 5.1 have been established considering
beams subjected only to gravity loading
(4) the amounts of redistribution given in Table 5.1 allow for inelastic behaviour in
composite beams, but not for the momentrotation characteristics of semi-rigid or
partial-strength joints
34
46
12 May 2004 11:52:33
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
Clause 5.4.4(3)
(5) crushing of the concrete encasement in compression may limit the rotation capacity
needed to achieve redistribution; limits to redistribution for partially encased beams can
be determined by using the rules for steel members or concrete members, whichever is
the more restrictive (clause 5.4.4(3))
(6) the amounts of redistribution given in Table 5.1 have been established only for beams of
uniform section
35
Is there a need for second-order
global analysis? (Clause 5.4.4(1))
No RoM from regions
influenced by buckling
No
Steel member. Follow
clause 5.4(2) of
EN 1993-1-1.
(Clause 5.4.4(3)
Do first-order elastic global analysis
No RoM
permitted
Concrete member. Follow
clause 5.5 of EN 1992-1-1.
(Clause 5.4.4(3))
Composite member. Are there any
partially encased beams without a
concrete or composite slab?
Are there any cross-sections or members where
resistance is influenced by any form of buckling
other than local buckling? (Clause 5.4.4(2))
Yes
Yes
Yes
Yes
Follow the more restrictive of
the rules for steel and for
concrete members.
(Clause 5.4.4(3))
Does the structure
analysed include
any composite
columns?
No RoM that reduces
moments in columns permitted,
unless rotation capacity has
been verified. (Inferred from
clause 5.4.5(5))
Is the beam partially encased?
Is the beam part of an unbraced
frame? (Clause 5.4.4(4))
Either: check rotation capacity, or:
ignore encasement in compression
when finding M
Rd
at sections where
BM reduced. (Clause 5.4.4(4))
Are the conditions of clause 5.4.4(4) on joints
and depth of member both satisfied?
No RoM
permitted.
(Clause 5.4.4(6))
(END)
Increase maximum hogging BMs by up to
10% for cracked analyses and 20% for
uncracked analyses. (Clause 5.4.4(5))
Is the grade of the steel
higher than S355?
No RoM
permitted
(END)
Treat Class 1 sections as
Class 2 unless rotation
capacity verified.
(Clause 5.4.4(6))
No RoM of BMs
applied to the steel
member. (Clause
5.4.4(7))
Maximum hogging BMs may be reduced, by amounts within the limits given in Table 5.1
unless verified rotation capacity permits a higher value. (Clause 5.4.4(5))
(END)
Are all cross-sections
in Class 1 or 2?
Are all cross-sections in Class 1 or 2?
Note: The
provisions from
here onwards
refer to beams
not to frames
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
Fig. 5.4. Flow chart for redistribution of moments. Note: this chart applies for verifications for limit
states other than fatigue, based on linear-elastic global analysis (RoM, redistribution of bending moments;
BM, bending moment)
47
12 May 2004 11:52:33
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
(7) the greater strains associated with higher grade steels may increase the rotation needed
to achieve redistribution
(8) lateraltorsional buckling may limit the rotation capacity available.
The last condition can be restrictive, as it may apply where > 0.2 (see clause 6.4.2(4)).
However, for composite beams with rolled or equivalent welded steel sections, the value
recommended in a note to clause 6.3.2.3(1) of EN 1993-1-1 is > 0.4.
The conditions above apply to the percentages given in Table 5.1, which applies only to
beams. It should not be inferred that no redistribution is permitted in structures that fail to
satisfy one or more of them. It would be necessary to show that any redistribution proposed
satisfied clause 5.4.4(2).
Under distributed loading, redistribution usually occurs from hogging to sagging regions
of a beam(except of course at an end adjacent to a cantilever). The limits to this redistribution
in Table 5.1 are based on extensive experience in the use of earlier codes and on research.
They have been checked by parametric studies, based on Eurocode 4, of beams in Class 2
28
and Class 3.
29
The limits given in Table 5.1 for uncracked analyses are the same as in BS 5950,
31
but are
more restrictive by 5% for cracked analyses of beams in Class 1 or 2. This reflects the
finding
37
that the difference caused by cracking, between moments calculated by elastic
theory in such beams, is nearer 15% than 10%.
No provision is made in EN 1994-1-1 for a non-reinforced subdivision of Class 1, for
which redistribution up to 50% is allowed in BS 5950-3-1. The use of such sections is not
prevented by EN 1994-1-1. The requirements for minimum reinforcement given in clause
5.5.1(5) are applicable only if the calculated resistance moment takes account of composite
action. The resistance of a non-reinforced cross-section is that of the steel member alone.
Clause 5.4.4(4)
The use of plastic resistance moments for action effects found by elastic global analysis
(clause 5.4.1.1(1)) implies redistribution of moments, usually from internal supports to
mid-span regions, additional to the degrees of redistribution permitted, but not required, by
clause 5.4.4(4).
38
Clause 5.4.4(5)
Where there are heavy point loads, and in particular for adjacent spans of unequal length,
there can be a need for redistribution frommid-span to supports. This is allowed, to a limited
extent, by clause 5.4.4(5).
Clause 5.4.4(7)
The effects of sequence of construction should be considered where unpropped construction
is used and the composite member is in Class 3 or 4. As Table 5.1 makes allowance for
inelastic behaviour in a composite beam, clause 5.4.4(7) limits redistribution of moments to
those arising after composite action is achieved. No such limitation applies to a cross-section
in Class 1 or Class 2, as the moment resistance is determined by plastic analysis and is
therefore independent of the loading sequence.
The reference in clause 5.4.4(3)(b) to redistribution in steel members is to those that do
not subsequently become composite.
5.4.5. Rigid plastic global analysis for buildings
Plastic hinge analysis, so well known in the UK, is referred to as rigid plastic analysis
because it is based on the assumption that the response of a member to bending moment is
either rigid (no deformation) or plastic (rotation at constant bending moment). Other types
of inelastic analysis defined in clause 1.5.6 of EN 1990 are covered in clause 5.4.3. Typical
momentcurvature curves are shown in Fig. 5.5. No application rules are given for them
because they require purpose-written computer programmes.
These other methods are potentially more accurate than rigid plastic analysis, but only if
the stressstrain curves are realistic and account is taken of longitudinal slip, as required by
clause 5.4.3(2)P.
Clause 5.4.5 Taking account of the cross-references in clause 5.4.5, the conditions under which use of
rigid plastic global analysis is allowed extend over two pages, which need not be summarized
here. Development of a collapse mechanism in a composite structure requires a greater
36
LT
LT
48
12 May 2004 11:52:33
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
degree of redistribution of elastic moments than in most steel structures, because beams are
usually much stronger at mid-span than at supports. The purpose of the conditions is to
ensure that this redistribution, and the large in-plane deformations associated with it, can
occur without loss of resistance caused by buckling, fracture of steel, or crushing of concrete.
Rigid plastic analysis is applicable only if second-order effects are not significant.
Comments on clause 5.4.1.1 referred to the use of expression (5.1) when material behaviour is
non-linear. Care needs to be taken if plastic hinges are expected to form in partial-strength
joints.
39
These may be substantially weaker than the connected members and the plastic
hinges may format relatively lowlevels of load. It would be prudent to neglect the stiffness of
such joints when determining
cr
.
Clause 5.4.5(1)
Clause 5.4.5(4)
Clause 5.4.5(1) requires cross-sections of steel members to satisfy clause 5.6 of
EN 1993-1-1. This applies during unpropped construction, but not to steel elements
of composite members, except as provided in clause 5.4.5(6). The provision on rotation
capacity in clause 5.6(2) of EN 1993-1-1 is replaced by the conditions of clause 5.4.5(4) of
EN 1994-1-1.
The rule on neutral axis depth in clause 5.4.5(4)(g) is discussed under clause 6.2.1.2.
5.5. Classification of cross-sections
Typical types of cross-section are shown in Fig. 6.1. The classification of cross-sections of
composite beams is the established method of taking account in design of local buckling of
steel elements in compression. It determines the available methods of global analysis and the
basis for resistance to bending, in the same way as for steel members. Unlike the method in
EN 1993-1-1, it does not apply to columns.
Clause 5.5 A flow diagram for the provisions of clause 5.5 is given in Fig. 5.6. The clause numbers
given are from EN 1994-1-1, unless noted otherwise.
Clause 5.5.1(1)P Clause 5.5.1(1)P refers to EN 1993-1-1 for definitions of the four classes and the
slendernesses that define the class boundaries. Classes 1 to 4 correspond respectively to the
terms plastic, compact, semi-compact and slender that were formerly used in British
codes. The limiting slendernesses are similar to those of BS 5950-3-1.
31
The numbers appear
different because the two definitions of flange breadth are different, and the coefficient that
takes account of yield strength, , is defined as (235/f
y
) in the Eurocodes, and as (275/f
y
) in
BS 5950.
The scope of EN 1994-1-1 includes members where the cross-section of the steel
component has no plane of symmetry parallel to the plane of its web (e.g. a channel section).
Asymmetry of the concrete slab or its reinforcement is also acceptable.
Clause 5.5.1(2)
The classifications are done separately for steel flanges in compression and steel webs, but
the methods interact, as described below. The class of the cross-section is the less favourable
of the classes so found (clause 5.5.1(2)), with three exceptions. One is where a web is assumed
37
0
Rigidplastic
Bending
moment
Elasticperfectly plastic
Elasto-plastic
Curvature
Fig. 5.5. Momentcurvature curves for various types of global analysis
49
12 May 2004 11:52:34
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
to resist shear forces only (clause 5.5.2(12) of EN1993-1-1). The others are the hole-in-web
option of clause 5.5.2(3) and the use of web encasement, both discussed later.
Reference is sometimes made to a beam in a certain class. This means that none of its
cross-sections is in a less favourable class than the one stated, and may imply a certain
distribution of bending moment. Clause 5.5.1(2) warns that the class of a composite section
depends on the sign of the bending moment (sagging or hogging), as it does for a steel section
that is not symmetrical about its neutral axis for bending.
Designers of structures for buildings normally select beams with steel sections such that
the composite sections are in Class 1 or 2, for the following reasons:
Rigid plastic global analysis is not excluded, provided that the sections at locations of
plastic hinges are in Class 1.
38
Is steel compression
flange restrained by
shear connectors to
clause 5.5.2(1)?
Is the web encased in
concrete to clause 5.5.3(2)?
Classify flange to Table 5.2 of
EN 1994-1-1
From clause 5.5.2(2), classify
flange to Table 5.2 of EN 1993-1-1
Note 1: Flange means steel
compression flange
No
No
Yes
Yes
Note: See Note 2, below
Web is Class 3
Class 1
Compression
flange is:
Locate the plastic neutral axis, allowing
for partial shear connection, if any
From clause 5.5.2(2), classify the web
using the plastic stress distribution, to
Table 5.2 of EN 1993-1-1
Locate the elastic neutral axis, assuming full shear
connection, taking account of sequence of construction,
creep, and shrinkage, to clause 5.5.1(4)
Web in Class 3
Class 2 Class 3 Class 4
Web not classified Web is Class 4
Is web encased to clause 5.5.3(2)?
Yes
No Yes
No
No
Yes
From clause 5.5.2(2), classify the web using the
elastic stress distribution, to Table 5.2 of EN 1993-1-1
Is the compression flange
in Class 1 or 2?
Is the compression
flange in Class 3?
Replace web by effective web
in Class 2, to clause 5.5.2(3)?
Web is Class 2 Effective web
is Class 2
Is the flange in Class 1?
Section is Class 1 Section is Class 2
Web is Class 1
Yes
Effective
section is
Class 2 Section is Class 3 Section is Class 4
No
Note 2: Where elastic global analysis will be used, and the web will be assumed to resist shear force only, clause
5.5.4(6) of EN 1993-1-1 permits the section to be designed as Class 2, 3 or 4, depending only on the class of the flange
Yes No
Fig. 5.6. Classification of a cross-section of a composite beam
50
12 May 2004 11:52:34
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
Bending resistances of beams can be found using plastic theory. For composite sections,
this gives resistances from 20 to 40% above the elastic resistance, whereas the increase
for steel sections is about 15%.
The limits to redistribution of moments are more favourable than for Classes 3 and 4.
Where composite floor slabs are used, it may be difficult to provide full shear
connection. Clause 6.6.1.1(14) permits partial connection, but only where all beam
cross-sections are in Class 1 or 2.
Clause 5.5.1(3)
Simply-supported composite beams in buildings are almost always in Class 1 or 2, because
the depth of web in compression (if any) is small, and the connection to the concrete slab
prevents local buckling of the adjacent steel flange. Clause 5.5.1(3) refers to this, and clause
5.5.2(1) refers to the more useful clause 6.6.5.5, which limits the spacing of the shear
connectors required.
Clause 5.5.1(4)
Since the class of a web depends on the level of the neutral axis, and this is different for
elastic and plastic bending, it is not obvious which stress distribution should be used for a
section near the boundary between Classes 2 and 3. Clause 5.5.1(4) provides the answer, the
plastic distribution. This is because the use of the elastic distribution could place a section in
Class 2, for which the bending resistance would be based on the plastic distribution, which in
turn could place the section in Class 3.
Clause 5.5.1(5)
Clause 5.5.1(6)
Clause 5.5.1(5), on the minimum area of reinforcement for a concrete flange, appears
here, rather than in Section 6, because it gives a further condition for a cross-section to be
placed in Class 1 or 2. The reason is that these sections must maintain their bending
resistance, without fracture of the reinforcement, while subjected to higher rotation than
those in Class 3 or 4. This is ensured by disallowing the use of bars in ductility Class A (the
lowest), and by requiring a minimumcross-sectional area, which depends on the tensile force
in the slab just before it cracks.
40
Clause 5.5.1(6), on welded mesh, has the same objective.
Clause 5.5.1(7) Clause 5.5.1(7) draws attention to the use of unpropped construction, during which both
the top flange and the web of a steel beam may be in a lower class until the member becomes
composite.
The hole-in-web method
Clause 5.5.2(3)
This useful device first appeared in BS 5930-3-1.
31
It is now in clause 6.2.2.4 of EN 1993-1-1,
which is referred to from clause 5.5.2(3).
In beams subjected to hogging bending, it often happens that the bottom flange is in Class
1 or 2, and the web is in Class 3. The initial effect of local buckling of the web would be a small
reduction in the bending resistance of the section. The assumption that a defined depth of
web, the hole, is not effective in bending enables the reduced section to be upgraded from
Class 3 to Class 2, with the advantages for design that are listed above. The method is
analogous to the use of effective areas for Class 4 sections, to allow for local buckling.
There is a limitation to its scope that is not evident from the wording in EN 1993-1-1:
The proportion of the web in compression should be replaced by a part of 20t
w
adjacent to the
compression flange, with another part of 20t
w
adjacent to the plastic neutral axis of the effective
cross-section.
It follows that for a design yield strength f
yd
the compressive force in the web is limited to
40t
w
f
yd
. For a composite beam in hogging bending, the tensile force in the longitudinal
reinforcement in the slab can exceed this value, especially where f
yd
is reduced to allow for
vertical shear. The method is then not applicable, because the second part of 20t
w
is not
adjacent to the plastic neutral axis, which lies within the top flange. The method, and this
limitation, are illustrated in Examples 6.1 and 6.2.
Partially encased cross-sections
Partially encased sections are defined in clause 6.1.1(1)P. Those illustrated there also have
concrete flanges. The web encasement improves the resistance of both the web and the other
39
51
12 May 2004 11:52:34
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:29:52
Clause 5.5.3(1) flange to local buckling. A concrete flange is not essential, as shown in clause 5.5.3(1), which
gives the increased slenderness ratios for compression flanges in Classes 2 and 3. The limit
for Class 1 is unaltered.
The rest of clause 5.5.3 specifies the encasement that enables a Class 3 web to be treated as
Class 2, without loss of cross-section. Conditions under which the encasement contributes to
the bending and shear resistance of the member are given in Section 6, where relevant
comments will be found.
40
52
12 May 2004 11:52:34
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
CHAPTER 6
Ultimate limit states
This chapter corresponds to Section 6 of EN 1994-1-1, which has the following clauses:
Beams Clause 6.1
Resistances of cross-sections of beams Clause 6.2
Resistance of cross-sections of beams for buildings with partial encasement Clause 6.3
Lateraltorsional buckling of composite beams Clause 6.4
Transverse forces on webs Clause 6.5
Shear connection Clause 6.6
Composite columns and composite compression members Clause 6.7
Fatigue Clause 6.8
Clauses 6.1 to 6.7 define resistances of cross-sections to static loading, for comparison with
action effects determined by the methods of Section 5. The ultimate limit state considered is
STR, defined in clause 6.4.1(1) of EN 1990 as:
Internal failure or excessive deformation of the structure or structural members, where the
strength of constructional materials of the structure governs.
For lateraltorsional buckling of beams and for columns, the resistance is influenced by
the properties of the whole member, and there is an implicit assumption that the member is
of uniform cross-section, apart from variations arising from cracking of concrete and from
detailing.
The self-contained clause 6.8, Fatigue, covers steel, concrete and reinforcement by
cross-reference to Eurocodes 2 and 3, and deals mainly with shear connection in beams.
Clause 6.1.1
Most of the provisions of Section 6 are applicable to both buildings and bridges, but a
number of clauses are headed for buildings, and are replaced by other clauses in EN1994-2.
Some of these differences arise from the different treatments of shear connection in the two
codes, which are compared in comments on clause 6.1.1.
6.1. Beams
6.1.1. Beams for buildings
Figure 6.1 shows typical examples of beams for buildings within the scope of EN 1994-1-1.
The details include web encasement, and profiled sheeting with spans at right angles to the
span of the beam, and continuous or discontinuous over the beam. The top right-hand
diagram represents a longitudinal haunch. Not shown (and not excluded) is the common
situation in which profiled sheeting spans are parallel to the span of the beam. A re-entrant
trough is shown in the bottom right-hand diagram. Sheeting with trapezoidal troughs is also
within the scope of the code.
53
12 May 2004 11:52:34
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
The steel cross-section may be a rolled I- or H-section or may be a doubly-symmetrical or
mono-symmetrical plate girder. Other possible types include any of those shown in sheet 1 of
Table 5.2 of EN 1993-1-1; for example, rectangular hollow sections. Channel and angle
sections should not be used unless the shear connection is designed to provide torsional
restraint. Stub girders are not within the scope of EN 1994-1-1. There is an extensive
literature on their design.
41
Shear connection
Clause 6.1.1(4)P
In buildings, composite cross-sections are usually in Class 1 or 2, and the bending resistance
is determined by plastic theory. At the plastic moment of resistance, the longitudinal force in
a concrete flange is easily found, so design of shear connection for buildings is often based on
the change in this force between two cross-sections where the force is known. This led to the
concepts of critical cross-sections (clause 6.1.1(4)P to clause 6.1.1(6)) and critical lengths
(clause 6.1.1(6)). These concepts are not used in bridge design. Cross-sections in Class 3 or 4
are common in bridges, and elastic methods are used. Longitudinal shear flows are therefore
found from the well-known result from elastic theory, v
L
= .
Points of contraflexure are not critical cross-sections, partly because their location is
different for each arrangement of variable load. A critical length in a continuous beam may
therefore include both a sagging and a hogging region. Where connectors are uniformly
spaced over this length, the number in the hogging region may not correspond to the force
that has to be transferred from the longitudinal slab reinforcement. This does not matter,
provided that the reinforcing bars are long enough to be anchored beyond the relevant
connectors. The need for consistency between the spacing of connectors and curtailment of
reinforcement is treated in clause 6.6.1.3(2)P.
Clause 6.1.1(5)
A sudden change in the cross-section of a member changes the longitudinal force in the
concrete flange, even where the vertical shear is zero. In theory, shear connection to provide
this change is needed. Clause 6.1.1(5) gives a criterion for deciding whether the change is
sudden enough to be allowed for, and will normally show that changes in reinforcement can
be ignored. Where the clause is applied, the new critical section has different forces in the
flange on each side of it. It may not be clear which one to use.
One method is to use the result that gives the greater change of force over the critical
length being considered. An alternative is to locate critical cross-sections on both sides of the
change point, not more than about two beamdepths apart. The shear connection in the short
42
Fig. 6.1. Typical cross-sections of composite beams
/ VAy I
54
12 May 2004 11:52:35
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
critical length between these two sections, based on the longitudinal forces at those sections,
needs to take account of the change of section.
The application of clause 6.1.1(5) is clearer for a beamthat is composite for only part of its
length. The end of the composite region is then a critical section.
A tapering member has a gradually changing cross-section. This can occur from variation
in the thickness or effective width of the concrete flange, as well as from non-uniformity in
the steel section. Where elastic theory is used, the equation v
L, Ed
= V
Ed
A /I should be
replaced by
(D6.1)
Clause 6.1.1(6)
where x is the coordinate along the member. For buildings, where resistances may be based on
plastic theory, clause 6.1.1(6) enables the effect to be allowed for by using additional critical
sections. It is applicable, for example, where the steel beam is haunched. The treatment of
vertical shear then requires care, as part of it is resisted by the sloping steel flange.
Clause 6.1.1(7)
Provisions for composite floor slabs, using profiled steel sheeting, are given only for
buildings. The space within the troughs available for the shear connection is often insufficient
for the connectors needed to develop the ultimate compressive force in the concrete flange,
and the resistance moment corresponding to that force is often more than is required,
because of other constraints on design. This has led to the use of partial shear connection,
which is defined in clause 6.1.1(7). It is applicable only where the critical cross-sections are in
Class 1 or 2. Thus, in buildings, bending resistances are often limited to what is needed, i.e. to
M
Ed
, with shear connection based on bending resistances; see clause 6.6.2.2.
Where bending resistances of cross-sections are based on an elastic model and limiting
stresses, longitudinal shear flows can be found from v
L, Ed
= V
Ed
A /I. They are related to
action effects, not to resistances. Shear connection designed in this way, which is usual in
bridges, is partial according to the definition in clause 6.1.1(7)P, because increasing it would
increase the bending resistances in the vicinity though not in a way that is easily calculated,
because inelastic behaviour and partial interaction are involved.
For these reasons, the concept partial shear connection is confusing in bridge design and
not relevant. Clauses in EN 1994-1-1 that refer to it are therefore labelled for buildings.
Effective cross-section of a beam with a composite slab
Where the span of a composite slab is at right angles to that of the beam, as in the lower half of
Fig. 6.1, the effective area of concrete does not include that within the ribs. Where the spans
are parallel ( = 0), the effective area includes the area within the depth of the ribs, but
usually this is neglected. For ribs that run at an angle to the beam, the effective area of
concrete within an effective width of flange may be taken as the full area above the ribs plus
cos
2
times the area of concrete within the ribs. Where >60, cos
2
should be taken as zero.
Service ducts in slabs can cause a significant loss of effective cross-section.
6.1.2. Effective width for verification of cross-sections
Clause 6.1.2(2)
The variation of effective width along a span, as given by clause 5.4.1.2, is too complex for
verification of cross-sections in beams for buildings. The simplification in clause 6.1.2(2)
often enables checks on the bending resistance of continuous beams to be limited to the
supports and the mid-span regions. This paragraph should not be confused with clause
5.4.1.2(4), which applies to global analysis.
6.2. Resistances of cross-sections of beams
This clause is for beams without partial or full encasement in concrete. Most of it is
applicable to both buildings and bridges. Partial encasement is treated for buildings only, in
clause 6.3. Full encasement is outside the scope of EN 1994.
43
CHAPTER 6. ULTIMATE LIMIT STATES
y
L, Ed Ed Ed
d( / )
/
d
Ay I
V Ay I M
x
= +
y
55
12 May 2004 11:52:35
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
No guidance is given in EN 1994-1-1, or in EN 1993, on the treatment of large holes in
steel webs, but specialized literature is available.
42,43
Bolt holes in steelwork should be
treated in accordance with EN 1993-1-1, particularly clauses 6.2.2 to 6.2.6.
6.2.1. Bending resistance
In clause 6.2.1.1, three different approaches are given, based on rigid plastic theory,
non-linear theory and elastic analysis. The non-linear theory is that given in clause 6.2.1.4.
This is not a reference to non-linear global analysis.
Clause 6.2.1.1(3) The assumption that composite cross-sections remain plane is always permitted by clause
6.2.1.1(3), where elastic and non-linear theory are used, because the conditions set will be
satisfied if the design is in accordance with EN 1994. The implication is that longitudinal slip
is negligible.
There is no requirement for slip to be determined. This would be difficult because the
stiffness of shear connectors is not known accurately, especially where the slab is cracked.
Wherever slip may not be negligible, the design methods of EN 1994-1-1 are intended to
allow for its effects.
Clause 6.2.1.1(5)
For beams with curvature in plan sufficiently sharp for torsional moments not to be
negligible, clause 6.2.1.1(5) gives no guidance on howto allowfor the effects of curvature. In
analysis from first principles, checks for beams in buildings can be made by assuming that
the changing direction of the longitudinal force in a flange (and a web, if significant) creates
a transverse load on that flange, which is then designed as a horizontal beam to resist that
load. A steel bottom flange may require horizontal restraint at points within the span of the
beam, and the shear connection should be designed for both longitudinal and transverse
forces.
Clause
6.2.1.2(1)(a)
Full interaction in clause 6.2.1.2(1)(a) means that no account need be taken of slip or
separation at the steelconcrete interface.
Reinforcement in compression
Clause
6.2.1.2(1)(c)
It is usual to neglect slab reinforcement in compression (clause 6.2.1.2(1)(c)). If it is included,
and the concrete cover is little greater than the bar diameter, consideration should be given
to possible buckling of the bars. Guidance is given in clause 9.6.3(1) of EN 1992-1-1 on
reinforcement in concrete walls. The meaning is that the reinforcement in compression
should not be the layer nearest to the free surface of the slab.
Small concrete flanges
Where the concrete slab is in compression, the method of clause 6.2.1.2 is based on the
assumption that the whole effective areas of steel and concrete can reach their design
strengths before the concrete begins to crush. This may not be so if the concrete flange is
small compared with the steel section. This lowers the plastic neutral axis, and so increases
the maximum compressive strain at the top of the slab, for a given tensile strain in the steel
bottom flange.
Clause 6.2.1.2(2)
A detailed study of the problem has been reported.
44
Laboratory tests on beams show that
strain hardening of steel usually occurs before crushing of concrete. The effect of this, and
the low probability that the strength of both the steel and the concrete will be only at the
design level, led to the conclusion that premature crushing can be neglected unless the grade
of the structural steel is higher than S355. Clause 6.2.1.2(2) specifies a reduction in M
pl, Rd
where the steel grade is S420 or S460 and the depth of the plastic neutral axis is high. This
problem also affects the rotation capacity of plastic hinges. Extensive research
45,46
led to the
upper limit to neutral-axis depth given in clause 5.4.5(4)(g).
For composite columns, the risk of premature crushing led to a reduction in the factor
M
,
given in clause 6.7.3.6(1), for S420 and S460 steels.
44
56
12 May 2004 11:52:36
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Ductility of reinforcement
Clause 6.2.1.2(3)
Reinforcement with insufficient ductility to satisfy clause 5.5.1(5), and welded mesh, should
not be included within the effective section of beams in Class 1 or 2 (clause 6.2.1.2(3)). This is
because laboratory tests on hogging moment regions have shown
19
that some reinforcing
bars, and most welded meshes, fracture before the momentrotation curve for a typical
double-cantilever specimen reaches a plateau. The problem with welded mesh is explained
in comments on clause 3.2(1).
Profiled steel sheeting
Clause 6.2.1.2(4)
Clause 6.2.1.2(5)
The contribution of profiled steel sheeting in compression to the plastic moment of
resistance of a beam is ignored (clause 6.2.1.2(4)) because at large strains its resistance can
be much reduced by local buckling. Profiled sheeting with troughs that are not parallel to the
span of a beam is ineffective in tension. This is because deformation could arise from change
in shape of the profile rather than strain resulting from stress. Where the troughs are
parallel, resistance to tension may still be difficult to achieve. For advantage to be taken
of clause 6.2.1.2(5), the sheeting needs to be continuous, and interaction with other
components of the cross-section has to be achieved.
Beams with partial shear connection in buildings
Clause 6.2.1.3(1)
Clause 6.2.1.3(2)
The background to the use of partial shear connection is explained in comments on clause
6.1.1. It is permitted only for the compressive force in the concrete slab (clause 6.2.1.3(1)).
Where the slab is in tension the shear connection must be sufficient to ensure yielding
(clause 6.2.1.3(2)) of the reinforcement within the effective section. Full shear connection is
required in hogging regions of composite beams for several reasons:
the bending moment may be larger than predicted because the concrete has not cracked
or, if it has, because of tension stiffening
the yield strength of the reinforcement exceeds f
sd
( = f
sk
/
S
)
tests show that at high curvatures, strain hardening occurs in the reinforcement
the design rules for lateraltorsional buckling do not allow for the effects of partial
interaction.
It could be inferred from the definition of full shear connection in clause 6.1.1(7) that
where the bending resistance is reduced below M
pl, Rd
by the effects of lateral buckling, shear
connection is required only for the reduced resistance. Clause 6.2.1.3(2) makes it clear that
the inference is incorrect. Thus, clause 6.2.1.2 on the plastic resistance moment M
pl, Rd
applies, amongst other cases, to all beams in Class 1 or 2 with tensile force in the slab.
The words hogging bending in clause 6.2.1.3 imply that the concrete slab is above the
steel beam. This is an assumption implicit in much of the drafting of the provisions for
buildings. In the general clauses, phrases such as regions where the slab is in tension are
used instead, because in bridges this can occur in regions of sagging curvature (Fig. 6.2).
The provisions referred to in clause 6.2.1.3(1) include clause 6.6.1.1(14), which begins If
all cross-sections are in Class 1 or Class 2 . This means all sections within the span
considered. In practice, the use of an effective web in Class 2 (clause 5.5.2(3)) ensures that
few Class 3 sections need be excluded.
45
Fig. 6.2. Example of a composite beam with the slab in tension at mid-span
57
12 May 2004 11:52:36
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Ductile connectors
Clause 6.2.1.3(3) Clause 6.2.1.3(3) refers to ductile connectors. The basic condition for the use of partial
shear connection is that the bending resistance must not fall below the design value until
after the curvature has reached the minimum value relied upon in the method of global
analysis used. Use of redistribution of moments, for example, relies upon curvatures beyond
the elastic range.
In other words:
slip required (i.e. relied on in design) slip available (D6.2)
It has been shown by extensive numerical analyses, checked against test results,
47,48
that
the slip required increases with the span of the beam and, of course, as the number of shear
connectors is reduced. The latter parameter is represented by the ratio of the number
of connectors provided within a critical length, n, to the number n
f
required for full
shear connection (defined in clause 6.1.1(7)P), i.e. the number that will transmit the force
N
c, f
to the slab (see Fig. 6.2). A reduced number, n, will transmit a reduced force, N
c
. Thus,
for connectors of a given shear strength, the degree of shear connection is
= N
c
/N
c, f
= n/n
f
(D6.3)
The application rules for general use are based on an available slip of 6 mm. Condition
(D6.2) was then applied by defining combinations of and span length such that the slip
required did not exceed 6 mm.
Headed studs regarded as ductile are defined in clause 6.6.1.2, where a flow chart (see
Fig. 6.11) and further comments are given. For partial shear connection with non-ductile
connectors, reference should be made to comments on clause 6.2.1.4.
Calculation models
The calculation model given in clause 6.2.1.3(3) can be explained as follows. For a given
cross-section, the force N
c, f
can be found using clause 6.2.1.2. For < 1, the concrete stress
block has a reduced depth, and a neutral axis at its lower edge. For longitudinal equilibrium,
part of the steel beam must also be in compression, so it too has a neutral axis. The model
assumes no separation of the slab from the beam, so their curvatures must be the same. The
strain distribution is thus as shown in Fig. 6.3, which is for the situation shown in Fig. 6.4 in
EN 1994-1-1. At the interface between steel and concrete there is slip strain, i.e. rate of
change of longitudinal slip. Neither the slip at this point nor the slip strain need be calculated
in practice.
Clause 6.2.1.3(4)
Clause 6.2.1.3(5)
Clauses 6.2.1.3(4) and 6.2.1.3(5) give two relationships between resistance moment M
Rd
and degree of shear connection. Calculations using the method above give curve AHC in the
upper part of Fig. 6.4(a), in which M
pl, a, Rd
is the plastic resistance of the steel section. The
line AC is a simpler and more conservative approximation to it. Their use is now illustrated.
The lower half of Fig. 6.4(a) shows the limits to the use of partial shear connection given in
clause 6.6.1.2.
46
0.85f
cd

N
c
= hN
c, f
f
yd
f
yd
M
a
M
Rd
N
a
0
Stresses
Strains
Fig. 6.3. Plastic stress and strain distributions under sagging bending for partial shear connection
58
12 May 2004 11:52:36
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Outline of a typical design procedure
The example is a simply-supported beam of span L, with a Class 1 section at mid-span. With
full shear connection the resistance is M
pl, Rd
. The steel section has flanges of equal area, and
the loading is uniformly distributed. Stud connectors are to be used.
(1) Find the minimumshear connection, (n/n
f
)
min
, for which the connectors are ductile, from
clause 6.6.1.2 (e.g. route DEF in Fig. 6.4(a)), and the corresponding resistance to
bending (route FMB).
(2) If that resistance exceeds the design moment M
Ed
, this degree of shear connection is
sufficient. Calculate the number of connectors for full shear connection, and then, from
(n/n
f
)
min
, the number required.
(3) If the resistance, point B, is much higher than M
Ed
, it may be possible to reduce the
number of connectors by using the method for non-ductile connectors, as explained in
Example 6.4 below.
(4) If the resistance, point B, is below M
Ed
, as shown in Fig. 6.4(a), the interpolation method
can be used (route GKN) to give the value of n/n
f
required. Alternatively, point Hcan be
determined, as shown below, and hence point J. The higher of the values of given by
points J and Fis the minimumdegree of shear connection, and, hence, n can be found.
(5) The spacing of the n connectors is now considered, along the length L
cr
between the two
relevant critical cross-sections (here, mid-span and a support). If the conditions of clause
6.6.1.3(3) are satisfied, as they are here if M
pl, Rd
2.5M
a, pl, Rd
, the spacing may be
47
25
20
10
0
0.4 0.6 0.8 1.0
L
e
(m)
n/n
f
Studs to clause 6.6.1.2(3),
A
b
= A
t
Other studs, A
b
= A
t
,
from clause 6.6.1.2(1)
f
y
= 275
f
y
= 460
P
Q
F
E
D
N
h (= N
c
/N
c, f
)
N
c, el
/N
c, f
N
c
/N
c, f
J
K
H
A
B M
G
C
0
0.4 1.0
M
pl, Rd
M
pl, a, Rd
M
Ed
1.0
C
Q
U
T
V
0
S
R
Other studs, A
b
= 3A
t
P
M
pl, Rd
M
el, Rd
M
Ed
M
a, Ed
Fig. 6.4. Design methods for partial shear connection
(a) Ductile connectors
(b) Non-ductile connectors
59
12 May 2004 11:52:37
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
uniform. Otherwise, an intermediate critical section must be chosen, clause 6.6.1.3(4), or
the spacing must be related to the elastic distribution of longitudinal shear, clause
6.6.1.3(5).
Determination of n for a given M
Ed
by the equilibrium method
The number of connectors needed to develop the moment M
Ed
is n = N
c
/P
Rd
, where P
Rd
is the
design resistance of a connector and N
c
is the force referred to in clause 6.2.1.3(3), for a
moment M
Ed
. Its calculation is tedious, but can be simplified a little, as shown here.
Figure 6.5 shows a cross-section of a beam in sagging bending, where the compressive
force in the concrete slab is N
c
, less than N
c, f
(equation (D6.3)), being limited by the strength
of the shear connection. Following clause 6.2.1.3(3), the plastic stress blocks are as shown in
Fig. 6.5. The neutral axis in the steel is at a depth x
a
below the interface. It is convenient to
retain the known force N
a
acting at the centre of area G of the steel section, and to take the
compressive strength over depth x
a
as 2f
yd
. This is because the stress in that area is to be
changed from yield in tension to yield in compression, providing a compressive force N
ac
in
the steel. We need to find out whether or not x
a
> t
f
.
In most beams with full shear connection, N
c, f
= N
a
, because the plastic neutral axis lies
within the slab. It is required that N
c
/N
c, f
0.4, from Fig. 6.4(a), so
N
c
0.4N
c, f
0.4N
a
(a)
From equilibrium,
N
c
+ N
ac
= N
a
(b)
so, from expression (a),
N
ac
0.6N
a
(c)
When the neutral axis is at the underside of the steel top flange, of area A
top
,
N
ac
/N
a
= 2A
top
/A
a
(d)
For most rolled I-sections, A
top
0.3A
a
, so when x
a
= t
f
, from equation (d),
N
ac
0.6N
a
(e)
From expressions (c) and (e), x
a
< t
f
when the preceding assumptions are valid. They usually
are, so only the case x
a
< t
f
is considered. Force N
ac
is then as shown in Fig. 6.5, and acts at a
depth h
c
+ h
p
+ 0.5x
a
below the top of the slab. For rolled sections, x
a
h
c
+ h
p
, so this
depth can be taken as h
c
+ h
p
. Taking moments about the top of the slab,
48
h
c
x
c
x
a
b
eff
h
p
h
g
t
f
b
G
(a)
N
a
= A
s
f
yd
N
c
= 0.85b
eff
x
c
f
cd
N
ac
= 2bx
a
f
yd
(b)
Fig. 6.5. Theory for force N
c
. (a) Cross-section. (b) Longitudinal stresses
60
12 May 2004 11:52:37
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
M
Ed
= N
a
(h
g
+ h
p
+ h
c
) x
c
N
c
N
ac
(h
c
+ h
p
) (D6.4)
Substitution for N
ac
from equation (b) and use of the expression for N
c
in Fig. 6.5 gives
M
Ed
= N
a
h
g
+ 0.85b
eff
x
c
f
cd
(h
c
+ h
p
x
c
)
This can be solved for x
c
, which gives N
c
, and then n from n = N
c
/P
Rd
, since P
Rd
is known for
given connectors.
In practice, it may be simpler to calculate N
a
, choose a convenient value for n, find N
c
and
then x
c
, calculate N
ac
fromequation (b), and see whether or not equation (D6.4) gives a value
that exceeds M
Ed
. If it does not, n is increased and the process repeated.
There is much interaction between clause 6.2.1.3 and clauses 6.6.1.2 and 6.6.1.3. The use of
partial shear connection is illustrated in Example 6.7, which follows the comments on clause
6.6, in Examples 6.8 and 6.9, which are based on the same data, and in Fig. 6.11.
Non-ductile connectors
Clause 6.2.1.4
These are connectors that do not satisfy the requirements for ductile connectors given in
clauses 6.6.1.1 and 6.6.1.2. Plastic behaviour of the shear connection can no longer be
assumed. Non-linear or elastic theory should now be used to determine resistance to
bending. Provisions are given in clause 6.2.1.4 and clause 6.2.1.5.
Clause 6.2.1.4(2)
The effect of slip at the steelconcrete interface is to increase curvature, and usually to
reduce longitudinal shear, for a given distribution of bending moment along a span. Where
connectors are not ductile, slip must be kept small, so it is rational to neglect slip when
calculating longitudinal shear. This is why clause 6.2.1.4(2) says that cross-sections should be
assumed to remain plane.
Non-linear resistance to bending
Clause 6.2.1.4(1)
Clause 6.2.1.4(2)
Clause 6.2.1.4(3)
Clause 6.2.1.4(4)
Clause 6.2.1.4(5)
There are two approaches, described in clause 6.2.1.4. With both, the calculations should be
done at the critical sections for the design bending moments. The first approach, given in
clause 6.2.1.4(1) to 6.2.1.4(5), enables the resistance of a section to be determined iteratively
from the stressstrain relationships of the materials. A strain distribution is assumed for the
cross-section, and the resulting stresses determined. Usually, the assumed strain distribution
will have to be revised, to ensure that the stresses correspond to zero axial force on the
section. Once this condition is satisfied, the bending moment is calculated from the stress
distribution. This may show that the design bending moment does not exceed the resistance,
in which case the calculation for bending resistance may be terminated. Otherwise, a general
increase in strain should be made and the calculations repeated. For concrete, EN 1992-1-1
gives ultimate strains for concrete and reinforcement which eventually limit the moment
resistance.
Clause 6.2.1.4(6)
Clearly, in practice this procedure requires the use of software. For sections in Class 1 or 2,
a simplified approach is given in clause 6.2.1.4(6). This is based on three points on the curve
relating longitudinal force in the slab, N
c
, to design bending moment M
Ed
that are easily
determined. With reference to Fig. 6.4(b), which is based on Fig. 6.6, these points are:
P, where the composite member resists no moment, so N
c
= 0
Q, which is defined by the results of an elastic analysis of the section
C, based on plastic analysis of the section.
Clause 6.2.1.4(7)
Accurate calculation shows QC to be a convex-upwards curve, so the straight line QC is a
conservative approximation. Clause 6.2.1.4(6) thus enables hand calculation to be used. For
buildings, clause 6.2.1.4(7) refers to a simplified treatment of creep.
Shear connection to clause 6.2.1.4
Computations based on the stressstrain curves referred to in clause 6.2.1.4(3) to 6.2.1.4(5)
lead to a complete moment-curvature curve for the cross-section, including a falling branch.
The definition of partial shear connection in clause 6.1.1(7)P is unhelpful, because the
49
61
12 May 2004 11:52:37
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
50
number of connectors for full shear connection can be found only fromthe stress distribution
corresponding to the maximum moment. This is why N
c, f
in Fig. 6.4(b) is based on plastic
analysis to clause 6.2.1.2. This figure is approximately to scale for a typical composite section
in sagging bending, for which
M
pl, Rd
/M
el, Rd
= 1.33
N
c, el
/N
c, f
= 0.6
The position of line PQ depends on the method of construction. It is here assumed that
the beam was unpropped, and that at the critical section for sagging bending, the moment
M
a, Ed
applied to the steel alone was 0.25M
pl, Rd
.
For a given design moment M
Ed
(which must include M
a, Ed
, as resistances of sections are
found by plastic theory), the required ratio n/n
f
is given by route TUV in Fig. 6.4(b).
No specific guidance is given for the spacing of shear connectors when moment resistance
is determined by non-linear theory. The theory is based on plane cross-sections (clause
6.2.1.4(2)), so the spacing of connectors should ideally correspond to the variation of the
force in the slab, N
c
, along the member. Where non-ductile connectors are used, this should
be done. For ductile connectors, clause 6.6.1.3(3), which permits uniform spacing, may be
assumed to apply.
Elastic resistance to bending
Clause 6.2.1.5(2)
Clause 6.2.1.4(6) includes, almost incidentally, a definition of M
el, Rd
that may seemstrange. It
is a peculiarity of composite structures that when unpropped construction is used, the elastic
resistance to bending depends on the proportion of the total load that is applied before the
member becomes composite. Let M
a, Ed
and M
Ed
be the design bending moments for the steel
and composite sections, respectively, for a section in Class 3. Their total is typically less than
the elastic resistance to bending, so to find M
el, Rd
, one or both of them must be increased
until one or more of the limiting stresses in clause 6.2.1.5(2) is reached. To enable a unique
result to be obtained, clause 6.2.1.4(6) says that M
Ed
is to be increased, and M
a, Ed
left
unchanged. This is because M
a, Ed
is mainly from permanent actions, which are less uncertain
than the variable actions whose effects comprise most of M
Ed
.
Unpropped construction normally proceeds by stages, which may have to be considered
individually in bridge design. For simply-supported spans in buildings, it is usually sufficiently
accurate to assume that the whole of the wet concrete is placed simultaneously on the bare
steelwork.
The weight of formwork is, in reality, applied to the steel structure and removed from the
composite structure. This process leaves self-equilibrated residual stresses in composite
cross-sections. For composite beams in buildings, these can usually be ignored in calculations
for the final situation.
Clause 6.2.1.5(5) One permanent action that influences M
Ed
is shrinkage of concrete. Clause 6.2.1.5(5)
enables the primary stresses to be neglected in cracked concrete, but the implication is that
they should be included where the slab is in compression. This provision, which affects M
el, Rd
,
should not be confused with clause 5.4.2.2(8), which concerns global analysis to determine
the secondary effects of shrinkage in statically-indeterminate structures. (Secondary effects
are defined in clause 2.3.3.)
These complications explain why, for ultimate limit states in buildings, design methods
based on elastic behaviour are best avoided, as far as possible.
Example 6.1: resistance moment in hogging bending, with effective web
A typical cross-section near an internal support of a continuous composite beam is shown
in Fig. 6.6(a). The plastic and elastic methods of calculation for the hogging moment of
resistance are illustrated by preparing a graph that shows changes in this resistance as the
effective area of longitudinal reinforcement in the slab, A
s
, is increased from zero to
62
12 May 2004 11:52:38
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
51
1800 mm
2
. The use of an effective web to clause 5.5.2(3), discussed in Chapter 5, is also
illustrated. The depth of the web, as used here, is the depth between flange fillets (or
welds), defined as c in Table 5.2 of EN 1993-1-1, not the clear depth between the flanges.
Also from EN 1993-1-1, c is the depth of the web in compression, for a section in Class 1
or 2, and is the correction factor for yield strength of the steel.
The definition of the hole-in-web model, given in Fig. 6.3 of EN 1993-1-1, omits the
depth of the hole and the location of the newplastic neutral axis (pna). They were given in
ENV 1994-1-1,
49
and are shown in Fig. 6.6(a). The depth of the hole, 2(c 40t),
includes a small approximation, which is now explained.
In principle, the hole should have zero depth when , c, t and are such that the web is
on the boundary between Class 2 and Class 3. When > 0.5 (as is usual), and from Table
5.2 of EN 1993-1-1, this is when
c/t = 456/(13 1)
As increases from0.5 to 1.0, the right-hand side of this equation reduces from41.4 to 38.
For < 0.5, it is 41.5. For simplicity, it is taken as 40 for all values of , so that the depth of
web in compression can be defined as 40t, in blocks of depth 20t above and below the
hole.
The original depth in compression, c, is reduced to 40t. For equilibrium, the depth in
tension must be reduced by c 40t, so the plastic neutral axis moves up by this amount,
as shown in Fig. 6.6(a), and the depth of the hole is thus 2(c 40t).
Other useful results for a symmetrical steel section are as follows. The depth of web
in compression, 40t, includes the depth needed to balance the tensile force in the
reinforcement, which is
h
r
= A
s
f
sd
/tf
yd
(D6.5)
The tension in the top flange, including web fillets, balances the compression in the
bottom flange, so for longitudinal equilibrium the depth of web in tension is
h
t
= 40t h
r
(D6.6)
The total depth of the web is c, so the depth of the hole is
h
h
= c 40t h
t
= c 80t + h
r
(D6.7)
26
199
286
116
116
B
0
0
A
C
D
A
s
80
70
New pna
ac 40te
ac
pna
142
8.6
19
c = 360
100
19
Hole
102
95
116
222
340
0
0
228
228 102 328
100
(a) (b) (c)
f
yd
f
yd
f
yd
6.3
h
t
= 152
h
h
= 4
20te
20te
2(ac 40te)
Fig. 6.6. Plastic resistance moment in hogging bending, for A
s
= 267 mm
2
(units: mm and kN).
(a) Cross-section of Class 3 beam with hole in web. (b) Stress blocks for M
pl, Rd
for Class 2 beam.
(c) Stress blocks in web for M
pl, Rd
for Class 3 beam
63
12 May 2004 11:52:38
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
52
Data and results
Data for the calculations, in addition to the dimensions shown in Fig. 6.6, are:
structural steel: f
y
= 355 N/mm
2
,
M
= 1.0, so f
yd
= 355 N/mm
2
reinforcement: f
sk
= 500 N/mm
2
,
S
= 1.15, so f
sd
= 435 N/mm
2
steel section: 406 140 UB39 with A

a
= 4940 mm
2
, I
a
= 124.5 10
6
mm
4
.
The upper limit chosen for A
s
corresponds to a reinforcement ratio of 1.5%, which is
quite high for a beam in a building, in a flange with b
eff
= 1.5 m. If there were no profiled
sheeting (which plays no part in these calculations) the ratio for the 150 mmslab would be
0.8%, and there would be two layers of bars. For simplicity, one layer of bars is assumed
here, with propped construction.
The full results are shown in Fig. 6.7. Typical calculations only are given here.
Classification of the cross-section
Clause 5.5.1(1)P refers to EN 1993-1-1, where Table 5.2 applies. For f
y
= 355 N/mm
2
, it
gives = 0.81.
For the bottom flange,
c = (142 6.3)/2 10.2 = 57.6 mm
c/t = 57.6/(8.6 0.81) = 8.3
This is less than 9, so the flange is in Class 1, irrespective of the area of slab reinforcement.
The web depth between fillets is c = 360 mm, so
c/t = 360/(6.3 0.81) = 70.5
300
220
200
240
260
280
340
320
0 500 1000 1500 2000
M (kN m)
Class 2

Class 3,
hole in web
bottom flange yields
B
C
A
J
F
H
G
E
A
s
(mm
2
)
M
el, Rd
M
el, Rd
M
pl, Rd
M
pl, Rd
reinforcement yields
Fig. 6.7. Influence of longitudinal reinforcement on hogging moments of resistance
64
12 May 2004 11:52:38
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
53
This is less than 72, so when A
s
= 0, the web is in Class 1. Addition of reinforcement
increases the depth of web in compression, so its class depends on A
s
. Here, > 0.5, and
from Table 5.2 in EN 1993-1-1 the limit for Class 2 is
c/t = 456/(13 1)
For c/t = 70.5 this gives = 0.574. It will be shown that this corresponds to
A
s
= 267 mm
2
, by calculating M
pl, Rd
by the methods for both a Class 2 and a Class 3
section.
M
pl, Rd
for Class 2 section with A
s
= 267 mm
2
The stress blocks for the web are shown in Fig. 6.6(b).
(1) Find M
pl, a, Rd
for the steel section. For a rolled section, the plastic section modulus is
usually found from tables. Here,
W
pl
= 0.721 10
6
mm
3
so
M
pl, a, Rd
= 0.721 355 = 256 kN m
(2) Find F
s
, the force in the reinforcement at yield:
F
s
= 267 0.435 = 116 kN
From equation (D6.5), the depth of web for this force is
h
r
= 116/(6.3 0.355) = 52 mm
To balance the force F
s
, the stress in a depth of web h
r
/2 changes from+f
yd
to f
yd
. This
is shown as ABCD in Fig. 6.6.
(3) The lever arm for the forces F
s
is 286 mm, so taking moments,
M
pl, Rd
= 256 + 116 0.286 = 289 kN m
M
pl, Rd
for Class 3 section with A
s
= 267 mm
2
The hole-in-web method is now used. Steps 1 and 2 are as above.
(3) The contribution of the web is deducted from M
pl, a, Rd
:
M
pl, a, flanges
= 256 0.36
2
6.3 355/4 = 183.5 kN m
(This value is not calculated directly because of the complex shape of each flange
which includes the web fillets, or, for a plate girder, a small depth of web.)
(4) From equation (D6.7), the depth of the hole is
h
h
= 360 408 + 52 = 4 mm
(5) From equation (D6.6), the depth of web in tension is
h
t
= 204 52 = 152 mm
and the force in it is
F
t
= 152 6.3 0.355 = 340 kN
(6) The stress blocks in the web are shown in Fig. 6.6(c). Taking moments about the
bottom of the slab,
M
pl, Rd
= 183.5 + 116 0.1 340 0.095 + 228(0.222 + 0.328) = 288 kN m
This agrees with the result for the Class 2 member, point A in Fig. 6.7, because this
section is at the class boundary. Hence, the depth of the hole is close to zero.
65
12 May 2004 11:52:39
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
54
M
pl, Rd
for A
s
> 267 mm
2
Similar calculations for higher values of A
s
give curve AB in Fig. 6.7, with increasing
depths of hole until, at A
s
= 1048 mm
2
, the new plastic neutral axis reaches the top of the
web, h
t
= 0, and the hole reaches its maximum depth. Further increase in A
s
, A
s
say,
causes changes of stress within the top flange only. Taking moments about the interface, it
is evident that the plastic bending resistance is increased by
M
pl, Rd
A
s
f
yd
h
s
where h
s
is the height of the reinforcement above the interface, 100 mm here. This is
shown by line BC in Fig. 6.7.
The bending resistance given by this method no longer approaches that given by elastic
theory (as it should, as the slenderness of the web in compression approaches the Class 3/4
boundary). Use of the method with the new plastic neutral axis in the top flange is
excluded by Eurocode 3, as explained in comments on clause 5.5.2(3). It is, in any case, not
recommended because:
the authors are not aware of any experimental validation for this situation (which is
uncommon in practice)
M
pl, Rd
is being calculated using a model where the compressive strain in the steel
bottom flange is so high that the rotation capacity associated with a Class 2 section
may not be available.
Equation (D6.6) shows that this restriction is equivalent to placing an upper limit of
40t on h
r
. Any slab reinforcement that increases h
r
above this value moves the section
back into Class 3. This can also be a consequence of vertical shear, as illustrated in
Example 6.2.
This point is relevant to the writing of software based on the code because software,
once written, tends to be used blindly.
For simplicity, M
el, Rd
has been calculated assuming propped construction. The stress in
the reinforcement governs until A
s
reaches 451 mm
2
(point J), after which M
el, Rd
is
determined by yield of the bottomflange. This is shown by curves EFand GHin Fig. 6.7.
Web in Class 4
The hole-in-web method is available only for webs in Class 3, so in principle a check
should be made that the web is not in Class 4, using the elastic stress distribution. AClass 4
web can occur in a plate girder, but is most unlikely in a rolled I- or H-section. In this
example, the Class 3/4 boundary is reached at A
s
= 3720 mm
2
.
6.2.2. Resistance to vertical shear
Clause 6.2.2 Clause 6.2.2 is for beams without web encasement. The whole of the vertical shear is usually
assumed to be resisted by the steel section, as in previous codes for composite beams. This
enables the design rules of EN1993-1-1, and EN1993-1-5
30
where necessary, to be used. The
assumption can be conservative where the slab is in compression. Even where it is in tension
and cracked in flexure, consideration of equilibrium shows that the slab must make some
contribution to shear resistance, except where the reinforcement has yielded. For solid slabs,
the effect is significant where the depth of the steel beam is only twice that of the slab,
50
but
diminishes as this ratio increases.
Clause 6.2.2.3(2)
In composite plate girders with vertical stiffeners, the concrete slab can contribute to the
anchorage of a tension field in the web,
51
but the shear connectors must then be designed for
vertical forces (clause 6.2.2.3(2)). The simpler alternative is to follow Eurocode 3, ignoring
both the interaction with the slab and vertical tension across the interface.
66
12 May 2004 11:52:39
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
55
Bending and vertical shear
Clause 6.2.2.4
Clause 6.2.2.4(1)
The methods of clause 6.2.2.4 are summarized in Fig. 6.8. Shear stress does not significantly
reduce bending resistance unless the shear is quite high. For this reason, the interaction may
be neglected until the shear force exceeds half of the shear resistance (clause 6.2.2.4(1)).
Clause 6.2.2.4(2) Both EN 1993-1-1 and EN 1994-1-1 use a parabolic interaction curve. In clause 6.2.2.4(2)
the reduction factor for the design yield strength of the web is (1 ), where
= [(2V
Ed
/V
Rd
) 1]
2
(6.5)
and V
Rd
is the resistance in shear. For a design shear force equal to V
Rd
, the bending resistance
is that provided by the flanges alone, denoted M
f, Rd
. This is calculated in Example 6.2.
The bending resistance at V
Ed
= 0 may be the elastic or the plastic value, depending on
the class of the cross-section. Where it is reduced to M
b, Rd
by lateraltorsional buckling,
interaction between bending and shear does not begin until a higher shear force than V
Rd
/2 is
present, as shown in Fig. 6.8(a).
Where the shear resistance V
Rd
is less than the plastic resistance to shear, V
pl, Rd
, because of
shear buckling, clause 6.2.2.4(2) replaces V
pl, Rd
by the shear buckling resistance V
b, Rd
.
Where the design yield strength of the web is reduced to allowfor vertical shear, the effect
on a Class 3 section in hogging bending is to increase the depth of web in compression. If the
change is small, the hole-in-web model can still be used, as shown in Example 6.2. For a
higher shear force, the new plastic neutral axis may be within the top flange, and the
hole-in-web method is inapplicable.
Clause 6.2.2.4(3) The section is then treated as Class 3 or 4, and clause 6.2.2.4(3) applies. It refers to
EN 1993-1-5. For beams, the rule given there is essentially
M
Ed
/M
Rd
+ (1 M
f, Rd
/M
pl, Rd
)(2V
Ed
/V
Rd
1)
2
1 (D6.8)
These symbols relate to the steel section only. For a composite section, longitudinal stresses
are found by elastic theory. These lead to values of M
Ed,
N
Ed
and V
Ed
acting on the steel
section, which is then checked to EN1993-1-5. It is fairly easy to check if a given combination
of these action effects can be resisted but calculation of bending resistance for a given
vertical shear is difficult.
Example 6.2: resistance to bending and vertical shear
Vertical shear is more likely to reduce resistance to bending in a continuous beam than in
a simply-supported one, and it is instructive to consider its influence on a beam with
bending resistance found by the hole-in-web method. Where the web is not susceptible to
shear buckling, the application of clause 6.2.2.4 is straightforward. This example is
therefore based on one of the fewUBsections where web buckling can occur, if S355 steel
Bending
resistance
0.5
V
Ed
/V
Rd
0
M
f, Rd
M
b, Rd
M
el, Rd
or M
pl, Rd
(a) (b)
132
180
100
48
102
102
108
43
326
118
100
213
213
328
f
yd
M
pl, a, fl
V
Rd
is the lesser of V
pl, Rd
and V
b, Rd
1.0
Fig. 6.8. Resistance to bending and vertical shear (dimensions in mm)
67
12 May 2004 11:52:39
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
56
is used. It is the section used in Example 6.1, 406 140 UB39, shown in Fig. 6.6, with
longitudinal reinforcement of area A
s
= 750 mm
2
.
The bending resistance will be calculated when the design vertical shear is V
Ed
= 300 kN.
All other data are as in Example 6.1.
FromEN1993-1-1, resistance to shear buckling must be checked if h
w
/t
w
> 72/, where
is a factor for which EN 1993-1-5 recommends the value 1.2. These clauses are usually
applied to plate girders, for which h
w
is the clear depth between the flanges. Ignoring the
corner fillets of this rolled section, h
w
= 381 mm, and for S355 steel, = 0.81, so
h
w
/t
w
= 381 1.2/(6.3 0.81) = 90.5
The resistance of this unstiffened web to shear buckling is found using clauses 5.2 and 5.3
of EN 1993-1-5, assuming a web of area h
w
t
w
, that there is no contribution from the
flanges, and that there are transverse stiffeners at the supports. The result is
V
b, Rd
= 475 kN
which is 85% of V
pl, Rd
, as found by the method of EN 1993-1-1 for a rolled I-section. From
clause 6.2.2.4(2),
= [(2V
Ed
/V
Rd
) 1]
2
= (600/475 1)
2
= 0.068
The reduced yield strength of the web is (1 0.068) 355 = 331 N/mm
2
. For hogging
bending, the tensile force in the reinforcement is
F
s
= 750 0.5/1.15 = 326 kN (D6.9)
From plastic theory, the depth of web in compression that is above the neutral axis of the
I-section is
326/(2 0.331 6.3) = 78 mm
This places the cross-section in Class 3, so the hole-in-web method is applied. FromFig.
6.6(a), the depth of the compressive stress-blocks in the web is 20t. The value for should
be based on the full yield strength of the web, not on the reduced yield strength, so
each block is 102 mm deep, as in Fig. 6.6(c). The use of the reduced yield strength
would increase and so reduce the depth of the hole in the web, which would be
unconservative. However, the force in each stress block should be found using the
reduced yield strength, and so is now
228 331/355 = 213 kN
The tensile force in the web is therefore
2 213 326 = 100 kN
and the longitudinal forces are as shown in Fig. 6.8(b).
From Example 6.1,
M
pl, a, flanges
= 183.5 kN m
Taking moments about the bottom of the slab,
M
pl, Rd
= 183.5 + 326 0.1 + 213(0.118 + 0.328) 100 0.043
M
pl, Rd
= 307 kN m (D6.10)
For this cross-section in bending only, the method of Example 6.1, with A
s
= 750 mm
2
,
gives
M
pl, Rd
= 314 kN m
The alternative to this method would be to use elastic theory. The result would then
depend on the method of construction.
68
12 May 2004 11:52:39
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
6.3. Resistance of cross-sections of beams for buildings with
partial encasement
Encasement in concrete of the webs of steel beams is normally done before erection, casting
one side at a time. This obviously increases the cost of fabrication, transport and erection,
but when it satisfies clause 5.5.3(2) it has many advantages for design:
it provides complete fire resistance for the web and, with longitudinal reinforcement,
compensation for the weakness of the bottom flange in fire, in accordance with
EN 1994-1-2
52
it enables a Class 3 web to be upgraded to Class 2, and the slenderness limit for a Class 2
compression flange to be increased by 40% (clause 5.5.3)
it widens significantly the range of steel sections that are not susceptible to lateral
torsional buckling (clause 6.4.3(1)(h))
it increases resistance to vertical shear, to clause 6.3.3(2)
it improves resistance to combined bending and shear, to clause 6.3.4(2)
it improves resistance to buckling in shear, clause 6.3.3(1).
6.3.1. Scope
Clause 6.3.1(2) To avoid shear buckling, clause 6.3.1(2) limits the slenderness of the encased web to
d/t
w
124. In practice, with encasement, steel sections in buildings are almost certain to be
in Class 1 or 2. Clause 6.3 is applicable only to these.
6.3.2. Resistance to bending
Clause 6.3.2
Clause 6.3.2(2)
The rules for resistance to bending, clause 6.3.2, correspond to those for uncased sections of
the same class, except that lateraltorsional buckling is not mentioned in clause 6.3.2(2).
Encasement greatly improves resistance to lateraltorsional buckling. Example 6.3, the
comments on clause 6.4.2(7), and pp. 153154 of Johnson and Molenstra,
44
are relevant.
However, it is possible for a beam within the scope of clause 6.3 to be susceptible; a
web-encased IPE 450 section in S420 steel is an example. Continuous beams that do not
satisfy clause 6.4.3 should therefore be checked.
Partial shear connection is permitted for a concrete flange, but not for web encasement.
The resistance to longitudinal shear provided by studs within the encasement is found in the
usual way, but no guidance is given on the contribution from bars that pass through holes in
the web or stirrups welded to the web, in accordance with clauses 5.5.3(2) and 6.3.3(2). They
are provided to ensure the integrity of the encased section.
The loadslip properties of different types of shear connection should be compatible
(clause 6.6.1.1(6)P). It is known from research on Perfobond shear connectors (longitudinal
flange plates projecting into the slab, with holes through which bars pass) that these bars
provide shear connection with good slip properties,
53,54
so a contribution from them could be
used here; but welds to stirrups may be too brittle.
6.3.36.3.4. Resistance to vertical shear, and to bending and vertical shear
Clause 6.3.3 The rules in clause 6.3.3 are based on the concept of superposition of resistances of
composite and reinforced concrete members. This concept has been used in Japan for
decades, in design of structures for earthquake resistance. The shear connection must
be designed to ensure that the shear force is shared between the steel web and the
concrete encasement. The references to EN 1992-1-1 are intended to ensure that the web
encasement retains its shear resistance at a shear strain sufficient to cause yielding of the
steel web.
Clause 6.3.3(1)
Clause 6.3.4
Clause 6.3.3(1) shows that no account need be taken of web buckling in shear. Moment
shear interaction is treated in clause 6.3.4 in a manner consistent with the rules for uncased
sections.
57
69
12 May 2004 11:52:40
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
6.4. Lateraltorsional buckling of composite beams
6.4.1. General
Clause 6.4.1(1)
It is assumed in this section that in completed structures for buildings, the steel top flanges
of all composite beams will be stabilized laterally by connection to a concrete or composite
slab (clause 6.4.1(1)). The rules on maximum spacing of connectors in clauses 6.6.5.5(1)
and 6.6.5.5(2) relate to the classification of the top flange, and thus only to local
buckling. For lateraltorsional buckling, the relevant rule, given in clause 6.6.5.5(3), is less
restrictive.
Clause 6.4.1(2)
Any steel top flange in compression that is not so stabilized should be checked for lateral
buckling (clause 6.4.1(2)) using clause 6.3.2 of EN 1993-1-1. This applies particularly during
unpropped construction. In a long span, it may be necessary to check a steel beam that is
composite along only part of its length. The general method of clause 6.4.2, based on the use
of a computed value of the elastic critical moment M
cr
, is applicable, but no detailed guidance
on the calculation of M
cr
is given in either EN 1993-1-1 or EN 1994-1-1. Buckling of
web-encased beams without a concrete flange has been studied.
55
However, for buildings,
the construction phase is rarely critical in practice, because the loading is so much less than
the design total load.
Steel bottom flanges are in compression only in cantilevers and continuous beams. The
length in compression may include most of the span, when that span is lightly loaded and
both adjacent spans are fully loaded. Bottom flanges in compression should always be
restrained laterally at supports (clause 6.4.3(1)(f) is relevant). It should not be assumed that
a point of contraflexure is equivalent to a lateral restraint.
In a composite beam, the concrete slab provides lateral restraint to the steel member, and
also restrains its rotation about a longitudinal axis. Lateral buckling is always associated with
distortion (change of shape) of the cross-section. Design methods for composite beams must
take account of the bending of the web, Fig. 6.9(b). They differ in detail from the method of
clause 6.3.2 of EN 1993-1-1, but the same imperfection factors and buckling curves are used,
in the absence of any better-established alternatives.
Clause 6.4.1(3) The reference in clause 6.4.1(3) to EN 1993-1-1 provides a general method for use where
neither of the methods of EN 1994-1-1 are applicable (e.g. for a Class 4 beam).
6.4.2. Verification of lateraltorsional buckling of continuous composite
beams with cross-sections in Class 1, 2 and 3 for buildings
This general method of design is written with distortional buckling of bottom flanges in
mind. It would not apply, for example, to a mid-span cross-section of a beam with the slab at
bottom-flange level (see Fig. 6.2). Although not stated, it is implied that the span concerned
is of uniform composite section, excluding minor changes such as reinforcement details and
effects of cracking of concrete. The use of this method for a two-span beam is illustrated in
Example 6.7.
Clause 6.4.2(1)
Clause 6.4.2(2)
Clause 6.4.2(3)
The method is based closely on clause 6.3.2 of EN 1993-1-1. There is correspondence in
the definitions of the reduction factor
LT
(clause 6.4.2(1)) and the relative slenderness,
(clause 6.4.2(4)). The reduction factor is applied to the design resistance moment M
Rd
, which
is defined in clauses 6.4.2(1) to 6.4.2(3). Expressions for M
Rd
include the design yield strength
f
yd
. The reference in these clauses to the use of
M1
is provided because this is a check on
instability. The recommended values for
M0
and
M1
are the same (1.0), but a National
Annex could define different values.
The determination of M
Rd
for a Class 3 section differs from that of M
el, Rd
in clause
6.2.1.4(6) only in that the limiting stress f
cd
for concrete in compression need not be
considered. It is necessary to take account of the method of construction.
The buckling resistance moment M
b, Rd
given by equation (6.6) must exceed the highest
applied moment M
Ed
within the unbraced length of compression flange considered.
58
LT
70
12 May 2004 11:52:40
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Lateral buckling for a Class 3 cross-section with unpropped construction
The influence of method of construction on the verification of a Class 3 composite section
for lateral buckling is as follows. From equation (6.4)
M
Rd
= M
el, Rd
= M
a, Ed
+ kM
c, Ed
(a)
where the subscript c indicates the action effect on the composite member.
From equation (6.6) the verification is
M
Ed
= M
a, Ed
+ M
c, Ed

LT
M
el, Rd
(b)
which is
LT
(M
a, Ed
+ M
c, Ed
)/M
el, Rd
= M
Ed
/M
el, Rd
(c)
The total hogging bending moment M
Ed
may be almost independent of the method of
construction. However, the stress limit that determines M
el, Rd
may be different for propped
and unpropped construction. If it is bottom-flange compression in both cases, then M
el, Rd
is
lower for unpropped construction, and the limit on
LT
from equation (c) is more severe.
Elastic critical buckling moment
Clause 6.4.2(4)
Clause 6.4.2(5)
Clause 6.4.2(4) requires the determination of the elastic critical buckling moment, taking
account of the relevant restraints, so their stiffnesses have to be calculated. The lateral
restraint from the slab can usually be assumed to be rigid. Where the structure is such that a
pair of steel beams and a concrete flange attached to them can be modelled as an inverted-U
frame (Fig. 6.11), continuous along the span, the rotational restraining stiffness at top-flange
level, k
s
, can be found from clauses 6.4.2(5) to 6.4.2(7).
Clause 6.4.2(6)
Clause 6.4.2(7)
Clause 6.4.2(5) gives conditions that define this frame. Analysis is based on its stiffness per
unit length, k
s
, given by the ratio F/, where is the lateral displacement caused by a force F
(Fig. 6.9(a)). The flexibility /F is the sum of flexibilities due to:
59
h
d
s
t
f
t
w
h
s
A B
F F d
q
0
d
a
(a)
C
D
(b)
L
L
1
L
2
0.8 L
2
/L
1
1.25
L
1
L
(c)
Fig. 6.9. U-frame action and distortional lateral buckling
71
12 May 2004 11:52:40
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
bending of the slab, which may not be negligible: 1/k
1
from equation (6.9)
bending of the steel web, which predominates: 1/k
2
from equation (6.10)
flexibility of the shear connection.
It has been found
56
that this last flexibility can be neglected in design to EN 1994-1-1. This
leads to equation (6.8) for stiffness k
s
.
This continuous U-frame concept has long been used in the design of steel bridges.
57
There is a similar discrete U-frame concept, which would be relevant to composite beams if
the steel sections had vertical web stiffeners. The shear connectors closest to those stiffeners
would then have to transmit almost the whole of the bending moment Fh (Fig. 6.9(a)), where
F is now a force on a discrete U-frame. The last of the three flexibilities listed above might
then not be negligible, nor is it certain that the shear connection and the adjacent slab would
be sufficiently strong.
58
Where stiffeners are present, the resistance of the connection above
each stiffener to repeated transverse bending should be established, as there is a risk of local
shear failure within the slab. There is at present no simple method of verification. The words
may be unstiffened in clause 6.4.3(1)(f) are misleading, as the resistance model is based on
both theory and research on unstiffened webs. It should, in the authors opinion, read should
be unstiffened. The problem is avoided in bridge design by using transverse steel members,
such as U- or H-frames (Fig. 6.10(a)).
The conditions in clause 6.4.3(1) referred to from clause 6.4.2(5) are commonly satisfied in
buildings by the beams that support composite slabs; but where these are secondary beams, the
method is not applicable to the primary beams because condition (e) is not satisfied. Bottom
flanges of primary beams can sometimes be stabilized by bracing from the secondary beams.
The calculation of k
s
is straightforward, apart fromfinding (EI)
2
, the cracked flexural stiffness
of a composite slab. An approximate method, used in Example 6.7, is derived in Appendix A.
Concrete-encased web
Clause 6.4.2(7)
Clause 6.4.2(9)
Clauses 6.4.2(7) and 6.4.2(9) allow for the additional stiffness provided by web encasement.
This is significant: for rolled steel sections, k
2
from equation (6.11) is from 10 to 40 times the
value from equation (6.10), depending on the ratio of flange breadth to web thickness.
Encasement will often remove any susceptibility to lateraltorsional buckling. The model
used for equation (6.11) is explained in Appendix A.
Theory for the continuous inverted-U frame model
Aformula for the elastic critical buckling moment for the U-frame model was given in Annex
B of ENV 1994-1-1,
49
but was removed from EN 1994-1-1, as it was considered to be
textbook material. However, it is sufficiently unfamiliar to be worth giving here.
Subject to conditions discussed below, the elastic critical buckling moment at an internal
support of a continuous beam is
M
cr
= (k
c
C
4
/L)[(G
a
I
at
+ k
s
L
2
/
2
)E
a
I
afz
]
1/2
(D6.11)
where: k
c
is a property of the composite section, given below,
C
4
is a property of the distribution of bending moment within length L,
G
a
is the shear modulus for steel (G
a
= E
a
/[2(1 + )] = 80.8 kN/mm
2
),
I
at
is the torsional moment of area of the steel section,
k
s
is the rotational stiffness defined in clause 6.4.2(6),
L is the length of the beam between points at which the bottom flange of the steel
member is laterally restrained (typically, the span length), and
I
afz
is the minor-axis second moment of area of the steel bottom flange.
Where the cross-section of the steel member is symmetrical about both axes, the factor k
c
is given by
k
c
= (h
s
I
y
/I
ay
)/[(h
s
2
/4 + i
x
2
)/e + h
s
] (D6.12)
with
e = AI
ay
/[A
a
z
c
(A A
a
)] (D6.13)
60
72
12 May 2004 11:52:41
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
where: h
s
is the distance between the centres of the flanges of the steel section,
I
y
is the second moment of area for major-axis bending of the cracked composite
section of area A,
I
ay
is the corresponding second moment of area of the steel section,
i
x
2
= (I
ay
+ I
az
)/A
a
, where I
az
and A
a
are properties of the steel section, and
z
c
is the distance between the centroid of the steel beamand the mid-depth of the slab.
Four of the conditions for the use of these formulae are given in paragraphs (c) to (f) of
clause 6.4.3(1). Three further conditions were given in ENV 1994-1-1. These related to the
resistance of the slab part of the U-frame to hogging transverse bending in the plane of the
U-frame, to its flexural stiffness, and to the spacing of shear connectors. It is now considered
that other requirements are such that these will, in practice, be satisfied. Further explanation
is given in Appendix A.
The coefficient C
4
was given in a set of tables, determined by numerical analyses, in which
its range is 6.247.6. These values are given in Appendix A (see Figs A.3 and A.4). The
coefficient accounts for the increased resistance to lateral buckling where the bending
moment is not uniform along the member. When checking lateral stability, the distribution
of bending moments corresponding to C
4
must be used as the action effects, and not an
equivalent uniform value. The calculation method is shown in Example 6.7.
Alternative theory for the elastic critical moment
There is an analogy between the differential equations for distortional lateral buckling,
taking account of restraint to warping, and those for a compressed member on an elastic
foundation. This has led
59
to an alternative expression for the elastic critical moment. Like
equation (D6.11), its use requires computed values that depend on the bending-moment
distribution and, for this method, also on the parameter
B
2
= k
s
L
4
/(E
a
I
D
)
where k
s
, Land E
a
are as above, and I
D
is the sectorial moment of inertia of the steel member
related to the centre of the restrained steel flange. Four graphs of these values are given in
Hanswille,
59
and a more general set in Hanswille et al.
60
Predictions of M
cr
by this method and by equation (D6.11) were compared with results of
finite-element analyses, for beams with IPE 500 and HEA 1000 rolled sections. This method
was found to agree with the finite-element results for both internal and external spans.
Equation (D6.11) was found to be satisfactory for internal spans, but to be less accurate
generally for external spans, and unconservative by over 30% in some cases. This suggests
that it needs further validation.
6.4.3. Simplified verification for buildings without direct calculation
Clause 6.4.3(1)
As calculations for the U-frame model are quite extensive, a simplified method has been
developed from it. Clause 6.4.3(1) defines continuous beams and cantilevers that may be
designed without lateral bracing to the bottom flange, except at supports. Its Table 6.1 gives
limits to the steel grade and overall depth of the steel member, provided that it is an IPE or
HE rolled section. The contribution from partial encasement is allowed for in paragraph (h)
of this clause.
These results are derived from equation (D6.11) for the elastic critical buckling moment,
making assumptions that further reduce the scope of the method. Accounts of its origin are
available in both English
61
and German.
62
It is outlined in Johnson and Fan,
56
and is similar to
that used in the treatment of lateraltorsional buckling of haunched beams.
63
The basis is that there shall be no reduction, due to lateral buckling, in the resistance of the
beam to hogging bending. It is assumed that this is achieved when 0.4. This value is
given in a Note to clause 6.3.2.3 of EN 1993-1-1, which can be modified by a National Annex.
Any National Annex that defines a lower limit should therefore also state if the method is
permitted.
61
LT
73
12 May 2004 11:52:41
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
The slenderness is a function of the variation of bending moment along the span. This
was studied using various loadings on continuous beams of the types shown in Fig. 6.9(c), and
on beams with cantilevers. The limitations in paragraphs (a) and (b) of clause 6.4.3(1) on
spans and loading result from this work.
Simplified expression for , and use of British UB rolled sections
Table 6.1 applies only to IPE and HE rolled sections. Criteria for other rolled I- and
H-sections are deduced in Appendix A. The basis for the method is as follows.
For uncased beams that satisfy the conditions that apply to equation (D6.11) for M
cr
, have
a double symmetrical steel section, and are not concrete encased, the slenderness ratio for a
Class 1 or Class 2 cross-section may conservatively be taken as
(D6.14)
The derivation from equation (D6.11) is given in Appendix A. Most of the terms in equation
(D6.14) define properties of the steel I-section; b
f
is the breadth of the bottom flange, and
other symbols are as in Fig. 6.9(a).
To check if a particular section qualifies for simplified verification, a section parameter F
is calculated. From equation (D6.14), it is
(D6.15)
Limiting values of F, F
lim
say, are given in Appendix A (see Fig. A.5) for the nominal steel
grades listed in Table 6.1. The horizontal S line at or next above the plotted point F gives the
highest grade of steel for which the method of clause 6.3.3 can be used for that section.
Some examples are given in Table 6.1, with the values of F
lim
in the column headings. Many
of the heavier wide-flange sections in S275 steel qualify for verification without direct
calculation, but few UB sections in S355 steel do so.
In ENV 1994-1-1, verification without direct calculation was permitted for hot-rolled
sections of similar shape to IPE and HE sections, that conformed to Table 6.1 and a
geometrical condition similar to the limit on F. In EN 1994-1-1 this has been replaced by a
reference to National Annexes.
Use of UB rolled sections with encased webs to clause 5.5.3(2)
It is shown in Appendix A(equation (DA.4)) that the effect of web encasement is to increase
F
lim
by at least 29%. All the sections shown in Table 6.1 now qualify for S275 steel, and all
62
LT
0.75 0.5 0.25
y
w s s f
LT
f f w f a 4
5.0 1
4
f
t h h t
b t t b E C

= +

0.75 0.25
w s s f
f f w f
1
4
t h h t
F
b t t b

= +

Table 6.1. Qualification of some UB rolled steel sections for verification of lateraltorsional stability,
in a composite beam, without direct calculation
Section
Right-hand
side of
expression
(D6.15)
S275 steel,
uncased
(13.9)
S355 steel,
uncased
(12.3)
S275 steel,
cased
(18.0)
S355 steel,
cased
(15.8)
457 152 UB52 16.4 No No Yes No
457 152 UB67 14.9 No No Yes Yes
457 191 UB67 13.6 Yes No Yes Yes
457 191 UB98 11.8 Yes Yes Yes Yes
533 210 UB82 14.4 No No Yes Yes
533 210 UB122 12.5 Yes No Yes Yes
610 229 UB125 14.1 No No Yes Yes
610 229 UB140 13.5 Yes No Yes Yes
LT
74
12 May 2004 11:52:41
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
except one do so for S355 steel. Web encasement is thus an effective option for improving
the lateral stability of a rolled steel section in a continuous composite beam.
The method of clause 6.4.2 should be used where the steel section does not qualify.
Use of intermediate lateral bracing
Where the buckling resistance moment M
b, Rd
as found by one of the preceding methods is
significantly less than the design resistance moment M
Rd
of the cross-sections concerned, it
may be cost-effective to provide discrete lateral restraint to the steel bottom flange. Where
the slab is composite, a steel cross-member may be needed (Fig. 6.10(a)), but for solid slabs,
other solutions are possible (e.g. Fig. 6.10(b)).
Clause 6.3.2.1(2) of EN 1993-1-1 refers to beams with sufficient restraint to the
compression flange, but does not define sufficient. A Note to clause 6.3.2.4(3) of
EN 1993-1-1 refers to Annex BB.3 in that EN standard for buckling of components of
building structures with restraints. Clause BB.3.2(1) gives the minimum stable length
between lateral restraints, but this is intended to apply to lateraltorsional buckling, and
may not be appropriate for distortional buckling. Further provisions for steel structures are
given in EN 1993-2.
64
There is no guidance in EN 1994-1-1 on the minimum strength or stiffness that a lateral
restraint must have. There is guidance in Lawson and Rackham,
63
based on BS 5950: Part 1,
clause 4.3.2.
65
This states that a discrete restraint should be designed for 2% of the maximum
compressive force in the flange. It is suggested
63
that where discrete and continuous restraints
act together, as in a composite beam, the design force can be reduced to 1% of the force in
the flange. Provision of discrete bracing to make up a deficiency in continuous restraint is
attractive in principle, but the relative stiffness of the two types of restraint must be such that
they are effective in parallel.
Another proposal is to relate the restraining force to the total compression in the flange
and the web at the cross-section where the bracing is provided, to take advantage of the steep
moment gradient in a region of hogging bending.
In tests at the University of Warwick,
66
bracings that could resist 1% of the total
compression, defined in this way, were found to be effective. The calculation of this
compressive force involves an elastic analysis of the section, not otherwise needed, and the
rule is unsafe near points of contraflexure. There is at present no simple design method
better than the 2% rule quoted above, which can be over-conservative. An elastic analysis
can be avoided by taking the stress in the flange as the yield stress.
The design methods based on equation (D6.11) for M
cr
work well for complete spans,
but become unsatisfactory (over-conservative) for short lengths of beam between lateral
bracings. This is because the correct values of the factors C
4
are functions of the length
between lateral restraints. For simplicity, only the minimum values of C
4
are given in the
figures in Appendix B of this guide. They are applicable where the half-wavelength of a
buckle is less than the length L in equation (D6.11). This is always so where L is a complete
span, but where L is part of a span, the value given may be over-conservative. This is
63
(a) (b)
Fig. 6.10. Laterally restrained bottom flanges
75
12 May 2004 11:52:42
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
illustrated at the end of Example 6.7, where the effect of providing lateral bracing is
examined.
Flow charts for continuous beam
The flow charts in Fig. 6.11 cover some aspects of design of an internal span of a continuous
composite beam. Their scope is limited to cross-sections in Class 1, 2 or 3, an uncased web, a
uniform steel section, and no flexural interaction with supporting members. It is assumed
that for lateraltorsional buckling, the simplified method of clause 6.4.3 is not applicable.
Figure 6.11(a) refers to the following notes:
64
Classify sections at A and C, to clause 5.5
Is section Class 1 or 2?
Class 1 or 2 Class 3
Use effective section in
Class 2, to clause 5.5.2(3)?
Find M
Rd
as M
pl, Rd
,

to
clause 6.2.1.2

Find M
a, Ed
at A and C by
elastic global analysis for
relevant arrangements of
actions on steel members;
no redistribution
Find M
el, Rd
at A
and C to clause 6.4.2(3)
No
No
Yes
Find l
LT
at A and C, to clause 6.4. Is steel element a rolled or equivalent welded section?
Yes
No
No
A C B
Elevation of beam
Is l
LT
> 0.4 for AB or BC? (Clause 6.3.2.3(1) of EN 1993-1-1) Is l
LT
> 0.2 for AB or BC?

Find c
LT
to clause 6.4,
and hence find M
b, Rd
Use elastic global analysis, with relevant modular ratio, for actions on steel and composite
members and relevant load arrangements to find M
Ed
(= M
a, Ed
+ M
c, Ed
) at A, B and C.
Is l
LT
> 0.2 (or 0.4) for either AB or BC? (See Notes (1) and (2), before Example 6.3)
Yes
Yes
No
Ignore shrinkage
(clause 5.4.2.2(7))
Yes
Classify cross-section B. Find M
Rd
at B for full shear connection. Is M
Ed
< M
Rd
at B?
Is M
Ed
less than M
Rd
or M
b, Rd
, as appropriate at both of sections A and C?
No
Yes
Redistribute hogging bending moments M
c, Ed
to clause 5.4.4, to find
new M
Ed
at sections A, B and C. See Fig. 5.4
No
Re-design is
required
(END)
Re-design is
required
(END)
Yes
No
Go to Fig. 6.11(b) for design of shear connection
Yes
Fig. 6.11. (a) Flow chart for design for ultimate limit state of an internal span of a continuous beam in a
building, with uniform steel section and no cross-sections in Class 4
76
12 May 2004 11:52:42
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
65
Are all critical sections in Class 1 or 2, to clause 6.6.1.1(14)?
Sagging region Hogging region
Yes
Find longitudinal shear flows,
ignoring tensile strength of
concrete and tension stiffening,
to clauses 6.6.2.1 and 6.6.2.2.
See comments on clause 6.6.2
No
Find h
min
to clause 6.6.1.2. Do you intend to use h

h
min
?
Has it been shown that
proposed connectors satisfy
clauses 6.6.1.1(4)P and
6.6.1.1(5)?
Are connectors studs to clause 6.6.1.2(1) or (3)?
Yes
(c) Choose h and
find M
Rd
such that
clause 6.6.1.1(3)P
is satisfied
Yes (advised)
Yes
(advised)
(b) Choose h
No
Find M
Rd
to
clauses 6.2.1.3(3)
to 6.2.1.3(5)
(a) Choose h

h
min

so that studs are
ductile

Is M
Rd
M
Ed
Increase h. Return to
(a), (b), or (c)
Yes
No
No
No
Use full shear connection
for N
c, f
found from M
Rd
, to
clause 6.6.2.1(1) or 6.6.2.2(1)
Are all critical cross-sections in Class 1 or 2?
Sagging region. Find n
s
= hN
c, f
/P
Rd
for region each side of B
Hogging region. Find n
h
= N
c, f
/P
Rd
for regions adjacent to A and C
Find P
Rd
for stud connectors to clause 6.6.3.1 or 6.6.4. For other connectors, find P
Rd
that satisfies clause
6.6.1.1 and determine if ductile
Find n = n
s
+ n
h
for critical lengths AB and BC. Are connectors ductile?
Space n connectors
uniformly over lengths AB
(and CD) to clause 6.6.1.3(3)
Space n connectors over AB (and CD)
in accordance with longitudinal shear
to clause 6.6.1.3(5)
Either Or
No
Yes
Is M
pl, Rd
> 2.5 M
pl, a, Rd
at A, B, or C ? Check spacing of connectors to clauses
6.6.5.5(3) and 6.6.5.7(4). Do they ensure
the stability of any part of the member?
Additional checks required to clause 6.6.1.3(4)
Yes
No
Check spacing to clause 6.6.5.5(2) and revise if necessary
Yes
(END)
No
Space connectors in accordance with
the shear flow, to clause 6.6.1.3(5)
(END)
Yes
No
Find M
Rd
to clause
6.2.1.4 or 6.2.1.5
Fig. 6.11. (Contd) (b) Flow chart for design of an internal span of a continuous beam in a building
shear connection
77
12 May 2004 11:52:42
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
(1) The elastic global analysis is simpler if the uncracked model is used. The limits
to redistribution of hogging moments include allowance for the effects of cracking,
but redistribution is not permitted where allowance for lateraltorsional buckling is
required (clause 5.4.4(4)). Cracked analysis may then be preferred, as it gives lower
hogging moments at internal supports.
(2) If fully propped construction is used, M
a, Ed
may be zero at all cross-sections.
Example 6.3: lateraltorsional buckling of two-span beam
The design method of clause 6.4.2, with application of the theory in Appendix A, is
illustrated in Example 6.7, which will be found after the comments on clause 6.6. The
length of the method is such that the simplified verification of clause 6.4.3 is used
wherever possible.
Limitations of the simplified method are now illustrated, with reference to the two-
span beam treated in Example 6.7. This uses an IPE 450 steel section in grade S355 steel,
and is continuous over two 12 m spans. Details are shown in Figs 6.236.25.
Relevant results for hogging bending of the composite section at support B in Fig.
6.23(c) are as follows:
M
pl, Rd
= 781 kN m = 0.43 M
b, Rd
= 687 kN m
The design ultimate loads per unit length of beam, from Table 6.2, are
permanent: 7.80 + 1.62 = 9.42 kN/m
variable: 26.25 kN/m
The steel section fails the condition in paragraph (g) of clause 6.4.3(1), which limits its
depth to 400 mm. The beam satisfies all the other conditions except that in paragraph (b),
for its ratio of permanent to total load is only 9.42/35.67 = 0.26, far below the specified
minimum of 0.4.
The simplest way to satisfy paragraph (g) would be to encase the web in concrete, which
increases the depth limit to 600 mm, and the permanent load to 11.9 kN/m.
The condition in paragraph (b) is quite severe. In this case, it is
11.9 0.4(11.9 + q
d
)
whence
q
d
17.8 kN/m
This corresponds to a characteristic floor loading of
17.8/(1.5 2.5) = 4.75 kN/m
2
which is a big reduction fromthe 7 kN/m
2
specified, even though the required reduction in
(to 0.4) is less than 10%.
This result illustrates a common feature of simplified methods. They have to cover so
wide a variety of situations that they are over-conservative for some of them.
6.5. Transverse forces on webs
Clause 6.5
The local resistance of an unstiffened and unencased web to forces (typically, vertical forces)
applied through a steel flange can be assumed to be the same in a composite member as in a
steel member, so clause 6.5 consists mainly of references to EN 1993-1-5. The provisions of
Section 8 of EN 1993-1-5 are not limited to rolled sections, so are applicable to webs where
the neutral axis is not at mid-depth, as is usual in composite beams.
In buildings, local yielding or buckling of a web may occur where a composite beam is
continuous over a steel beam that supports it. Rolled I-sections in Class 1 or 2 may not be
66
LT
LT
78
12 May 2004 11:52:42
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Clause 6.5.1(3)
susceptible, but a Class 3 web, treated as effective Class 2, should almost always be stiffened
(clause 6.5.1(3)).
A plate girder launched over roller supports cannot be stiffened at every section, so the
erection condition may be critical; but this situation is rare in buildings.
Clause 6.5.2 Flange-induced web buckling, clause 6.5.2, could occur where a large compression flange
is restrained from buckling out of its plane by a weak or slender web. This is prevented by
specifying a limit to the web slenderness, as a function of the ratio of flange area to web area.
The slenderness limit is reduced if the flange is curved in elevation, to ensure that the web
can resist the radial component of the force in the flange. This form of web buckling cannot
occur with straight rolled steel I-sections, and needs to be checked only for plate girders of
unusual proportions and for members sharply curved in elevation.
The effect of sharp curvature is illustrated by reference to an IPE 400 section of grade
S355 steel, which has been cold curved about its major axis, before erection. Assuming that
its plastic resistance moment is to be used, the minimum permitted radius of curvature given
by clause 8(2) of EN 1993-1-5 is 2.1 m. Flange-induced buckling is clearly a rare problem for
hot-rolled sections.
6.6. Shear connection
6.6.1. General
Basis of design
Clause 6.6.1.1(1) Clause 6.6 is applicable to shear connection in composite beams. Clause 6.6.1.1(1) refers also
to other types of composite member. Shear connection in composite columns is treated in
clause 6.7.4, but reference is made to clause 6.6.3.1 for the design resistance of headed stud
connectors. Similarly, headed studs used for end anchorage in composite slabs are treated in
clause 9.7.4, but some provisions in clause 6.6 are also applicable.
Clause 6.6.1.1(2)P Although the uncertain effects of bond are excluded by clause 6.6.1.1(2)P, friction is not
excluded. Its essential difference from bond is that there must be compressive force across
the relevant surfaces. This usually arises from wedging action. Provisions for shear
connection by friction are given in clauses 6.7.4.2(4) (columns) and 9.1.2.1 (composite slabs).
Clause 6.6.1.1(3)P
Clause 6.6.1.1(4)P
Clause 6.6.1.1(5)
Inelastic redistribution of shear (clause 6.6.1.1(3)P) is relied on in the many provisions
that permit uniform spacing of connectors. Clause 6.6.1.1(4)P uses the term ductile for
connectors that have deformation capacity sufficient to assume ideal plastic behaviour of the
shear connection. Clause 6.6.1.1(5) quantifies this as a characteristic slip capacity of 6 mm.
47
In practice, designers will not wish to calculate required and available slip capacities. Clause
6.6.1.2(1) enables such calculations to be avoided by limiting the extent of partial shear
connection and by specifying the type and range of shear connectors.
Clause 6.6.1.1(6)P The need for compatibility of load/slip properties, clause 6.6.1.1(6)P, is one reason why
neither bond nor adhesives can be used to supplement the shear resistance of studs. The
combined use of studs and block-and-hoop connectors has been discouraged for the same
reason, though there is little doubt that effectively rigid projections into the concrete slab,
such as bolt heads and ends of flange plates, contribute to shear connection.
Clause 6.6.1.1(7)P
Clause 6.6.1.1(8)
Separation, in clause 6.6.1.1(7)P, means separation sufficient for the curvatures of the
two elements to be different at a cross-section, or for there to be a risk of local corrosion.
None of the design methods in EN 1994-1-1 takes account of differences of curvature, which
can arise froma very small separation. Even where most of the load is applied by or above the
slab, as is usual, tests on beams with unheaded studs show separation, especially after
inelastic behaviour begins. This arises from local variations in the flexural stiffnesses of the
concrete and steel elements, and from the tendency of the slab to ride up on the weld collars.
The standard heads of stud connectors have been found to be large enough to control
separation, and the rule in clause 6.6.1.1(8) is intended to ensure that other types of
connector, with anchoring devices if necessary, can do so.
67
79
12 May 2004 11:52:43
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Resistance to uplift is much influenced by the reinforcement near the bottom of the slab,
so if the resistance of an anchor is to be checked by testing, reinforcement in accordance with
clause 6.6.6 should be provided in the test specimens. Anchors are inevitably subjected also
to shear.
Clause 6.6.1.1(9) Clause 6.6.1.1(9) refers to direct tension. Load from a travelling crane hanging from the
steel member is an example. In bridges, it can be caused by the differential deflexions of
adjacent beams under certain patterns of imposed load. Where it is present, its design
magnitude must be determined.
Clause 6.6.1.1(10)P Clause 6.6.1.1(10)P is a principle that has led to many application rules. The shear forces
are inevitably concentrated. One research study
67
found that 70%of the shear on a stud was
resisted by its weld collar, and that the local (triaxial) stress in the concrete was several times
its cube strength. Transverse reinforcement performs a dual role. It acts as horizontal shear
reinforcement for the concrete flanges, and controls and limits splitting. Its detailing is
critical where connectors are close to a free surface of the slab.
Larger concentrated forces occur where precast slabs are used, and connectors are placed
in groups in holes in the slabs. This influences the detailing of the reinforcement near these
holes, and is referred to in Section 8 of EN 1994-2.
Clause 6.6.1.1(12) Clause 6.6.1.1(12) is intended to permit the use of other types of connector. ENV1994-1-1
included provisions for many types of connector other than studs: block connectors, anchors,
hoops, angles, and friction-grip bolts. They have all been omitted because of their limited use
and to shorten the code.
Clause 6.6.1.1(13) Clause 6.6.1.1(13) says that connectors should resist at least the design shear force,
meaning the action effect. Their design for Class 1 and 2 sections is based on the design
bending resistances (see clause 6.6.2), and hence on a shear force that normally exceeds the
action effect.
Clause 6.6.1.1(14)P The principle on partial shear connection, clause 6.6.1.1(14)P, leads to application rules in
clause 6.2.1.3.
The flow chart in Fig. 6.11(b) is for a beam without web encasement, and may assist in
following the comments on clauses 6.2 and 6.6.
Limitation on the use of partial shear connection in beams for buildings
As noted in comments on clause 6.2.1.3, the rules for partial shear connection are based on
an available slip of 6 mm. Connectors defined as ductile are those that had been shown to
have (or were believed to have) a characteristic slip capacity (defined in clause B.2.5(4))
exceeding 6 mm.
Clause 6.6.1.2(1)
Prediction of slip capacity is difficult. Push tests on stud connectors have been reported in
scores of publications, but few tests were continued for slips exceeding 3 mm. Slip capacity
depends on the degree of containment of the connector by the concrete and its
reinforcement, and hence on the location of free surfaces (e.g. in haunches or edge beams)
as well as on the shape, size and spacing of the connectors. The information has been
summarized.
47,48
The conclusions led to the approval of certain stud connectors as ductile
(clause 6.6.1.2(1)), and also friction-grip bolts, which were within the scope of ENV1994-1-1.
These conclusions seem optimistic when compared with the results of some push tests
using solid slabs, but connectors behave much better in beams reinforced as required by the
Eurocode than in small push-test specimens, where splitting can cause premature failure.
One might expect a lower available slip fromstuds in very strong concrete, but the 6 mmlimit
has been confirmed
68
by four push tests with cylinder strengths f
cm
86 N/mm
2
.
Clause 6.6.1.2(3)
For certain types of profiled sheeting, available slips were found to be greater than
6 mm.
43,69
These results and other test data led to a relaxation of the limiting effective spans
at which low degrees of shear connection can be used, as shown in the lower part of Fig.
6.4(a). This applies only where the conditions in paragraphs (a)(e) of clause 6.6.1.2(3) are
satisfied, because these are the situations for which test data are available. There is no
validated theoretical model that includes all the many relevant variables, so this relaxation is
68
80
12 May 2004 11:52:43
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
69
allowed only where the force N
c
is determined by the more conservative of the two methods
given in clause 6.2.1.3, the interpolation method.
Research continues on the influence on slip capacity of profile shape and the detailing of
studs in troughs.
The limits to the use of partial shear connection in buildings are summarized in Fig. 6.4(a),
where L
e
is the effective span. The span limits given by the lines PQ to RS are from the
provisions of clause 6.6.1.2 for ductile connectors. The design of a long-span beam with a
low degree of shear connection is likely to be governed by the need to limit its deflection,
unless it is propped during construction or is continuous.
For composite beams in sagging bending, the steel top flange must be wide enough to
resist lateral buckling during erection, and for attachment of the connectors, but can often be
smaller than the bottom flange. A smaller flange lowers the plastic neutral axis of the
composite section, and increases the slip required by the model for partial-interaction
design. This is why clauses 6.6.1.2(1) and 6.6.1.2(2) give limits on the degree of shear
connection that are less liberal than those for beams with equal steel flanges.
Spacing of shear connectors in beams for buildings
Clause 6.6.1.3(1)P Clause 6.6.1.3(1)P extends clause 6.6.1.1(2)P a little, by referring to spacing of connectors
and an appropriate distribution of longitudinal shear. The interpretation of appropriate
can depend on the method of analysis used and the ductility of the connectors.
The principle may be assumed to be satisfied where connectors are spaced elastically to
clause 6.6.1.3(5), which has general applicability. The more convenient use of uniform
spacing requires the connectors to satisfy clause 6.6.1.3(3), which implies (but does not
require) the use of plastic resistance moments. The connectors must be ductile, as defined
in clauses 6.6.1.1(4)P and 6.6.1.1(5). This is normally achieved by satisfying clause 6.6.1.2.
Clauses 6.6.1.1(3)P to 6.6.1.1(5) provide an alternative, which enables research-based
evidence to be used, but its use is not appropriate for routine design.
Clause 6.6.1.3(2)P
In practice, it is possible to space connectors uniformly in most beams for buildings, so for
continuous beams clause 6.6.1.3(2)P requires the tension reinforcement to be curtailed to
suit the spacing of the shear connectors.
Clause 6.6.1.3(4) Clause 6.6.1.3(4) could apply to a simply-supported or a continuous beam with a large
concrete slab and a relatively small steel top flange. Connectors spaced uniformly along a
critical length might then have insufficient available slip. Use of an additional critical section
would lead to a more suitable distribution.
Example 6.4: arrangement of shear connectors
As an example of the use of these rules, a simply-supported beam of span 10 m is
considered. It has distributed loading, equal steel flanges, a uniformcross-section in Class
2, S355 steel, and stud connectors. At mid-span, M
Ed
is much less than M
pl, Rd
. The
cross-section is such that the required resistance to bending can be provided using 40% of
full shear connection (n/n
f
= 0.4).
Clause 6.6.1.2(1) gives n/n
f
0.55. However, if the slab is composite, and the other
conditions of clause 6.6.1.2(3) are satisfied, n/n
f
= 0.4 may be used.
Suppose now that the span of the beam is 12 m. The preceding limits to n/n
f
are
increased to 0.61 and 0.48, respectively. One can either design using these limits, or go to
clause 6.6.1.3(5), which refers to longitudinal shear calculated by elastic theory. This
presumably means using v
L
= V
Ed
A /I, where V
Ed
is the vertical shear on the composite
section. This gives a triangular distribution of longitudinal shear, or separate
distributions, to be superimposed, if creep is allowed for by using several modular ratios.
The force in the slab at mid-span now depends on the proportion of M
Ed
that is applied
to the composite member, and on the proportions of the cross-section. Connectors
corresponding to this force are then spaced accordingly, possibly with extra ones near
mid-span to satisfy the rule of clause 6.6.5.5(3) on maximum spacing of studs.
y
81
12 May 2004 11:52:43
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
70
Strictly, the envelope of vertical shear should be used for V
Ed
, which gives non-zero
shear at mid-span. For distributed loading this increases the shear connection needed by
only 2%, but the envelope should certainly be used for more complex variable loading.
For continuous beams where M
Ed
at mid-span is much less than M
pl, Rd
, the method
is more complex, as partial shear connection is permitted only where the slab is in
compression. The envelope of design vertical shear from the global analysis should be
used. For simplicity, longitudinal shear can be found using properties of the uncracked
cross-section throughout, because this gives an overestimate in cracked regions. Examples
6.7 and 6.8 (see below) are relevant.
It does not help, in the present example, to define additional critical sections within the
6 m shear span of the beam, because the limits of clause 6.6.1.2 are given in terms of
effective span, not critical length.
6.6.2. Longitudinal shear force in beams for buildings
Clause 6.6.2.1
Clause 6.6.2.2
Clauses 6.6.2.1 and 6.6.2.2 say, in effect, that the design longitudinal shear force should be
consistent with the bending resistances of the cross-sections at the ends of the critical length
considered, not with the design vertical shear forces (the action effects). This is done for two
reasons:
simplicity for the design bending moments often lie between the elastic and plastic
resistances, and calculation of longitudinal shear becomes complex
robustness for otherwise longitudinal shear failure, which may be more brittle than
flexural failure, could occur first.
Beam with Class 3 sections at supports and a Class 1 or 2 section at mid-span
Clause 6.6.2.1 applies because non-linear or elastic theory will have been applied to
cross-sections. The longitudinal forces in the slab at the Class 3 sections are then calculated
by elastic theory, based on the bending moments in the composite section. At mid-span, it is
not clear whether clause 6.6.2.2 applies, because its heading does not say resistance of all
cross-sections. The simpler and recommended method is to assume that it does apply, and
to calculate the longitudinal force at mid-span based on M
Rd
at that section, as that is
consistent with the model used for bending. The total shear flow between a support and
mid-span is the sum of the longitudinal forces at those points. The alternative would be to
find the longitudinal force at mid-span by elastic theory for the moments applied to the
composite section, even though the bending stresses could exceed the specified limits.
Clause 6.6.2.2(3) This absence of all from the heading is also relevant to the use of clause 6.6.2.2(3), on the
use of partial shear connection. The design for a beam with Class 3 sections at internal
supports limits the curvature of those regions, so the ultimate-load curvature at mid-span
will be too low for the full-interaction bending resistance to be reached. The use of partial
shear connection is then appropriate, with M
Rd
less than M
pl, Rd
.
6.6.3. Headed stud connectors in solid slabs and concrete encasement
Resistance to longitudinal shear
In BS 5950-3-1
31
and in earlier UK codes, the characteristic shear resistances of studs are
given in a table, applicable only when the stud material has particular properties. There was
no theoretical model for the shear resistance.
Clause 6.6.3.1(1)
The Eurocodes must be applicable to a wider range of products, so design equations are
essential. Those given in clause 6.6.3.1(1) are based on the model that a stud with a shank
diameter d and an ultimate strength f
u
, set in concrete with a characteristic strength f
ck
and a
mean secant modulus E
cm
, fails either in the steel alone or in the concrete alone.
The concrete failure is found in tests to be influenced by both the stiffness and the strength
of the concrete.
82
12 May 2004 11:52:43
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
This led to equations (6.18)(6.21), in which the numerical constants and partial safety
factor
V
have been deduced from analyses of test data. In situations where the resistances
from equations (6.18) and (6.19) are similar, tests show that interaction occurs between the
two assumed modes of failure. An equation based on analyses of test data, but not on a
defined model,
70
P
Rd
= k(d
2
/4)f
u
(E
cm
/E
a
)
0.4
(f
ck
/f
u
)
0.35
(D6.16)
gives a curve with a shape that approximates better both to test data and to values tabulated
in BS 5950.
In the statistical analyses done for EN 1994-1-1
71,72
both of these methods were studied.
Equation (D6.16) gave results with slightly less scatter, but the equations of clause
6.6.3.1(1) were preferred because of their clear basis and experience of their use in some
countries. Here, and elsewhere in Section 6, coefficients from such analyses were modified
slightly, to enable a single partial factor, denoted
V
(V for shear), 1.25, to be recommended
for all types of shear connection. This value has been used in draft Eurocodes for over
20 years.
It was concluded from this study
72
that the coefficient in equation (6.19) should be 0.26.
This result was based on push tests, where the mean number of studs per specimen was only
six, and where lateral restraint from the narrow test slabs was usually less stiff than in the
concrete flange of a composite beam. Strength of studs in many beams is also increased by
the presence of hogging transverse bending of the slab. For these reasons the coefficient was
increased from 0.26 to 0.29, a value that is supported by a subsequent calibration study
15
based on beams with partial shear connection.
Design resistances of 19 mm stud connectors in solid slabs, given by clause 6.6.3.1, are
shown in Fig. 6.12. It is assumed that the penalty for short studs, equation (6.20), does not
apply. For any given values of f
u
and f
ck,
the figure shows which failure mode governs. It can
be used for this purpose for studs of other diameters, provided that h/d 4.
The overall nominal height of a stud, used in equations (6.20) and (6.21), is about 5 mm
greater than the length after welding, a term which is also in use.
71
100
80
60
20 30 40 50
P
Rd
(kN)
Normal-density
concrete
Density class 1.8
f
u
= 500 N/mm
2
450
400
f
ck
(N/mm
2
)
f
cu
(N/mm
2
)
25 37 50 60
Fig. 6.12. Design shear resistances of 19 mm studs with h/d 4 in solid slabs
83
12 May 2004 11:52:44
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Weld collars
Clause 6.6.3.1(2) Clause 6.6.3.1(2) on weld collars refers to EN 13918,
26
which gives guide values for the
height and diameter of collars, with the note that these may vary in through-deck stud
welding. It is known that for studs with normal weld collars, a high proportion of the shear is
transmitted through the collar.
67
It should not be assumed that the shear resistances of clause
6.6.3.1 are applicable to studs without collars (e.g. where friction welding by high-speed
spinning is used). A normal collar should be fused to the shank of the stud. Typical collars in
the test specimens from which the design formulae were deduced had a diameter not less
than 1.25d and a minimumheight not less than 0.15d, where d is the diameter of the shank.
The collars of studs welded through profiled sheeting can be of different shape fromthose
for studs welded direct to steel flanges, and the shear strength may also depend on the
effective diameter of the weld between the sheeting and the flange, about which little is
known. The resistances of clause 6.6.3.1 are applicable where the welding is in accordance
with EN ISO 14555,
73
as is required.
Studs with a diameter exceeding 20 mm are rarely used in buildings, as welding through
sheeting becomes more difficult, and more powerful welding plant is required.
Splitting of the slab
Clause 6.6.3.1(3) Clause 6.6.3.1(3) refers to splitting forces in the direction of the slab thickness. These occur
where the axis of a stud lies in a plane parallel to that of the concrete slab; for example, if
studs are welded to the web of a steel T-section that projects into a concrete flange. These
are referred to as lying studs in published research
74
on the local reinforcement needed to
prevent or control splitting. There is an informative annex on this subject in EN 1994-2.
11
A
similar problem occurs in composite L-beams with studs close to a free edge of the slab. This
is addressed in clause 6.6.5.3(2).
Tension in studs
Clause 6.6.3.2(2)
Pressure under the head of a stud connector and friction on the shank normally causes the
stud weld to be subjected to vertical tension before shear failure is reached. This is why clause
6.6.1.1(8) requires shear connectors to have a resistance to tension that is at least 10% of the
shear resistance. Clause 6.6.3.2(2) therefore permits tensile forces that are less than this to be
neglected.
Resistance of studs to higher tensile forces has been found to depend on so many
variables, especially the layout of local reinforcement, that no simple design rules could be
given. It is usually possible to find other ways of resisting the vertical tension that occurs, e.g.
where a travelling crane is supported from the steel element of a composite beam.
6.6.4. Design resistance of headed studs used with profiled steel sheeting in
buildings
The loadslip behaviour of a stud connector in a trough (of sheeting) or rib (of concrete;
both terms are used) (Fig. 6.13) is more complex than in a solid slab. It is influenced by
the direction of the ribs relative to the span of the beam
their mean breadth b
0
and depth h
p
the diameter d and height h
sc
of the stud
the number n
r
of the studs in one trough, and their spacing
whether or not a stud is central within a trough, and, if not, by its eccentricity and the
direction of the shear.
The interactions between these parameters have been explored by testing, with sheeting
continuous across the beam. It is clear that the most significant are the ratios h
sc
/h
p
and b
0
/h
p
and, for ribs transverse to the supporting beams, n
r
and the eccentricity, if any. In
EN 1994-1-1, as in earlier codes, reduction factors k ( 1.0) are given, for application to the
design resistances of studs in solid slabs. They are based entirely on testing and experience.
72
84
12 May 2004 11:52:44
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Clause 6.6.4.1 Sheeting with ribs parallel to the supporting beams (clause 6.6.4.1)
There are two situations. The sheeting may be continuous across the beams its side walls
then provide lateral restraint to the concrete around the studs or it may be discontinuous,
as shown in Fig. 6.13, providing a haunch with a breadth that usually exceeds the breadth b
0
of a trough. Edge fixings provided for erection may not provide much lateral restraint. The
detailing rules for studs in unsheeted haunches, clause 6.6.5.4, then provide a guide to good
practice, and may be more conservative than the rules of clause 6.6.4.1. Alternatively, the
sheeting may be anchored to the beam. As practice varies, the means to achieve appropriate
anchorage is a matter for the National Annex.
A haunch that just complies with clause 6.6.5.4 is shown, to scale, in Fig. 6.13. The rules
specify:
the angle ( 45)
the concrete side cover to the connector ( 50 mm)
the depth of the transverse reinforcement below the underside of the head ( 40 mm).
The application of clause 6.6.4.1(2) to this haunch is nowconsidered. The reduction factor
is
k
= 0.6(b
0
/h
p
)[(h
sc
/h
p
) 1] 1 (6.22)
where h
sc
may not be taken as greater than h
p
+ 75 mm. The equation is from Grant et al.,
75
dating from1977 because there is little recent research on ribs parallel to the beam.
76
Here, it
gives
k
= 0.6(146/75)[(145/75) 1] = 1.09
so there is no reduction. There would be, k
= 0.78, if the height of the stud were reduced to,

say, 125 mm. In an unsheeted haunch it would then be necessary to provide transverse
reinforcement at a lower level than in Fig. 6.13, which might be impracticable. Reasonable
consistence is thus shown between clauses 6.6.4.1 and 6.6.5.4.
There is no penalty for off-centre studs in either of these clauses, or for more than one at a
cross-section. The rules of clause 6.6.4.1 for continuous or anchored sheeting, and of clause
6.6.5.4, for discontinuous un-anchored sheeting, should ensure good detailing.
73
26 26
60 60
40
146, or b
0
for continuous sheeting
h
sc
= 145
h
p
= 75
=
=
19
q
Fig. 6.13. Details of a haunch, with parallel sheeting (dimensions in mm)
85
12 May 2004 11:52:44
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Where sheeting is continuous across a beam and through-deck stud welding is used, there
may be shear transfer from the sheeting to the beam. No guidance is given on this complex
situation, which may be ignored in design.
Clause 6.6.4.2 Sheeting with ribs transverse to the supporting beams (clause 6.6.4.2)
In the preceding paragraphs, the likely mode of failure was loss of restraint to the base of the
stud due to bursting (lateral expansion) of the haunch. Arib transverse to the beam, shown in
Fig. 6.14 with typical dimensions, is more highly stressed, as it has to transfer much of the
shear from the base of the stud to the continuous slab above. Three modes of failure are
shown in Fig. 6.15(a)(c):
(a) failure surface above a stud that is too short (concrete pullout)
(b) haunch too slender, with plastic hinges in the stud
(c) eccentricity on the weak side of the centre, which reduces the effective breadth of the
haunch,
69
and causes rib punching failure.
77
In BS 5950-3-1,
31
the reduction formula from Grant et al.
75
is given. It is, essentially,
k
t
= 0.85(b
0
/h
p
)[(h
sc
/h
p
) 1]/n
r
0.5
k
t, max
(D6.17)
with k
t, max
falling from1 to 0.6 as n
r
, the number of studs in a rib, increases from1 to 3. Where
a single stud is placed on the weak side of the centre of a rib, the breadth b
0
is taken as 2e
(Fig. 6.15(c)).
In a study of this subject
61
it was concluded that equation (D6.17) was inconsistent and
could be unsafe. Anewformula by Lawson
78
gave better predictions, and was recommended.
For the Eurocode, a reviewof Mottramand Johnson
69
and Lawson,
78
and their application
to a wider range of profiles than is used in the UK, found that more test data were needed. It
was initially decided
72
to reduce the factor 0.85 in equation (D6.17) to 0.7, to eliminate
situations where it had been found to be unsafe, with the conditions: n
r
not to be taken as
greater than 2 in computations, b
0
h
p
, and h
p
85 mm, with k
t, max
as before. Where there is
one stud per rib, Fig. 6.16 shows that:
where h
sc
2h
p
and t > 1 mm, the reduction factor is usually 1.0
where h
sc
1.5h
p
and/or t 1 mm, reductions in resistance can be large.
Most of the test data had been from sheeting exceeding 1.0 mm in thickness, with
through-deck welding of studs up to 20 mm in diameter. Later work led to the reductions in
k
t, max
for other situations, given in Table 6.2, for thinner sheeting and studs welded through
74
h
c
= 75
h
sc
= 95
=
h
p
= 55
b
0
= 160
50
e = 30
=
Fig. 6.14. Details of a haunch and positions of stud connectors, for trapezoidal sheeting transverse to
the supporting beams (dimensions in mm)
86
12 May 2004 11:52:45
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
holes. Research also found that the use of a reduction factor for the strength of studs in a
solid slab is not appropriate where strong studs are placed in a relatively weak rib, so further
limitations of scope were added:
stud diameter not to exceed the limits given in Table 6.2
ultimate strength of studs in sheeting not to be taken as greater than 450 N/mm
2
.
75
(a) (b) (c)
e
(e) (d)
Fig. 6.15. Failure modes and placing of studs, for troughs of profiled sheeting. The alternate favourable
and unfavourable placing of pairs of studs is shown in (d), and the diagonal placing of pairs of studs in (e)
0.6
1.0
0.8
k
t
b
0
/h
p
1.0 2.0 3.0
h
sc
/h
p
= 1.5
h
sc
/h
p
= 2 n
r
= 2
n
r
= 2
n
r
= 1
n
r
= 1
t > 1 mm
t > 1 mm
t 1 mm
t 1 mm
Range common in practice
Fig. 6.16. Reduction factor k
t
for studs with d 20 mm and through-deck welding
87
12 May 2004 11:52:45
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Some trapezoidal sheetings have a small rib projecting above the main top surface; for
example, as shown in Fig. 9.4. It is clear fromclause 6.6.4.1(1) that in equation (6.22), h
p
is the
depth including the rib. The words overall depth are not repeated in clause 6.6.4.2, but
should be assumed to apply, unless there is extensive test evidence that equation (6.23) gives
safe predictions of k
t
when h
p
is taken as the depth excluding the rib.
Crowding studs into a rib reduces ductility
69
as well as strength. Ductility is needed most in
long spans where, fortunately, longitudinal shears are usually low(in buildings) in relation to
the width of steel top flange available for the placing of studs. This is relevant because the
failures sketched in Fig. 6.15 are in reality three-dimensional, and resistances depend in a
complex way on the arrangement of pairs of studs within a rib.
Where sheeting has a small stiffening rib that prevents studs from being placed centrally
within each trough, and only one stud per trough is required, application of the rule given in
clause 6.6.5.8(3) is straightforward. Comments on that clause are given later.
It is not clear from EN 1994-1-1 how two studs per trough should then be arranged. If
there were no central rib, then two studs in line, spaced 5d apart (clause 6.6.5.7(4)) would
be permitted. Small stiffening ribs, as shown in Fig. 6.14, should have little effect on shear
resistance, so two-in-line is an acceptable layout. If the troughs are too narrowto permit this,
then two studs side-by-side spaced 4d apart (clause 6.6.5.7(4)) fall within the scope of
clause 6.6.5.8(3), and can be arranged as shown in plan in Fig. 6.15(d). There is limited
evidence from tests
76
that the diagonal layout of Fig. 6.15(e), to which the code does not
refer, may be weaker than two studs in line.
Example 6.5: reduction factors for transverse sheeting
The sheeting is as shown in Fig. 6.14, with an assumed thickness t = 0.9 mm, including
zinc coating. The overall depth is 55 mm. For one stud in the central location (shown by
dashed lines in Fig. 6.14), b
0
= 160 mm, h
sc
= 95 mm, and n
r
= 1. Equation (6.23) and
Table 6.2 give
k
t
= (0.7 160/55)(95/55 1) = 1.48 (but 0.85)
For two studs per trough, placed centrally and side by side,
k
t
= 1.48/1.41 = 1.05 (but 0.7)
Further calculations of reduction factors are included in Example 6.7.
Biaxial loading of shear connectors
Clause 6.6.4.3 The biaxial horizontal loading referred to in clause 6.6.4.3 occurs where stud connectors are
used to provide end anchorage for composite slabs, as shown in Fig. 9.1(c). Clause 9.7.4(3)
gives the design anchorage resistance as the lesser of that given by clause 6.6 and P
pb, Rd
, the
bearing resistance of the sheet, from equation (9.10). Even where 16 mm studs are used in
lightweight concrete, the bearing resistance (typically about 14 kN) will be the lower of the
two.
These studs resist horizontal shear from both the slab and the beam. The interaction
equation of clause 6.6.4.3(1) is based on vectorial addition of the two shear forces.
6.6.5. Detailing of the shear connection and influence of execution
It is rarely possible to prove the general validity of application rules for detailing, because
they apply to so great a variety of situations. They are based partly on previous practice. An
adverse experience causes the relevant rule to be made more restrictive. In research, existing
rules are often violated when test specimens are designed, in the hope that extensive good
experience may enable existing rules to be relaxed.
Rules are often expressed in the formof limiting dimensions, even though most behaviour
(excluding corrosion) is more influenced by ratios of dimensions than by a single value.
Minimum dimensions that would be appropriate for an unusually large structural member
76
88
12 May 2004 11:52:46
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
could exceed those given in the code. Similarly, code maxima may be too large for use in a
small member. Designers are unwise to follow detailing rules blindly, because no set of rules
can be comprehensive.
Resistance to separation
Clause 6.6.5.1(1) The object of clause 6.6.5.1(1) on resistance to separation is to ensure that failure surfaces in
the concrete cannot pass above the connectors and below the reinforcement, intersecting
neither. Tests have found that these surfaces may not be plane; the problem is three-
dimensional. A longitudinal section through a possible failure surface ABC is shown in Fig.
6.17. The studs are at the maximum spacing allowed by clause 6.6.5.5(3).
Clause 6.6.5.1 defines only the highest level for the bottom reinforcement. Ideally, its
longitudinal location relative to the studs should also be defined, because the objective is to
prevent failure surfaces where the angle (Fig. 6.17) is small. It is impracticable to specify a
minimum angle , or to link detailing rules for reinforcement with those for connectors.
The angle obviously depends on the spacing of the bottom bars, assumed to be 800 mm
in Fig. 6.17. What is the maximum permitted value for this spacing? The answer is quite
complex.
Clause 6.6.6.3(1) refers to clause 9.2.2(5) in EN 1992-1-1, where a Note recommends a
minimum reinforcement ratio. For f
ck
= 30 N/mm
2
and f
sk
= 500 N/mm
2
this gives 0.088%,
or 131 mm
2
/m for this 150 mm slab. However, if the slab is continuous across the beam, most
of this could be in the top of the slab.
The minimum bottom reinforcement depends on whether the slab is continuous or not. If
it is simply-supported on the beam, the bottom bars are principal reinforcement to clause
9.3.1.1(3) of EN 1992-1-1, where a Note gives their maximum spacing as 400 mm for this
example, and 450 mm for secondary reinforcement. The rules of clause 6.6.5.3(2) on
splitting may also apply.
If the slab is continuous over the beam, there may be no need for bottom flexural
reinforcement to EN 1992-1-1. Let us assume that there is a single row of 19 mm studs with a
design resistance, 91 kN per stud, which is possible in concrete of class C50/60 (Fig. 6.12).
The bottom transverse reinforcement will then be determined by the rules of clause 6.6.6.1
for shear surfaces of type bb or cc. These show that 12 mm bars at 750 mm spacing are
sufficient. There appears to be no rule that requires closer spacing, but it would be prudent
to treat these bars as secondary flexural bars to EN 1992-1-1, not least because it increases
the angle . Using its maximum spacing, 450 mm, 10 mm bars are sufficient.
Cover to connectors
Clause 6.6.5.2
Shear connectors must project significantly above profiled steel sheeting, because of the
term (h
sc
/h
p
1) in equations (6.22) and (6.23) and the rule in clause 6.6.5.8(1) for projection
of 2d above the top of the steel deck. The interpretation of this for profiles with a small
additional top rib is not clear. If 2d is measured fromthis rib, and the rules of EN1992-1-1 for
cover are applied to the top of the connector, the resulting minimum thickness of a
composite slab may govern its thickness. However, these slabs are normally used only in
dry environments, where the concessions on minimum cover given in clause 6.6.5.2 are
appropriate.
77
800
95
55
30
A
B
C
a
Fig. 6.17. Level of bottom transverse reinforcement (dimensions in mm)
89
12 May 2004 11:52:46
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Loading of shear connection during execution
Clause 6.6.5.2(4)
As almost all relevant provisions on execution appear in standards for either steel or
concrete structures, there is no section on execution in EN 1994-1-1. This is why clause
6.6.5.2(4) appears here.
The method of construction of composite beams and slabs (i.e. whether propped or
unpropped), and also the sequence of concreting, affects stresses calculated by elastic
theory, and the magnitude of deflections. It is significant, therefore, in verifications of
resistance of cross-sections in Class 3 or 4, and in serviceability checks.
Where propped construction is used, it is usual to retain the props until the concrete has
achieved a compressive strength of at least 75% of its design value. If this is not done,
verifications at removal of props should be based on a reduced compressive strength.
Clause 6.6.5.2(4) gives a lower limit for this reduction in concrete strength. It also relates
to the staged casting of a concrete flange for an unpropped composite beam, setting, in
effect, a minimum time interval between successive stages of casting.
Local reinforcement in the slab
Clause 6.6.5.3(1)
Clause 6.6.5.3(2)
Where shear connectors are close to a longitudinal edge of a concrete flange, use of U-bars is
almost the only way of providing the full anchorage required by clause 6.6.5.3(1). The
splitting referred to in clause 6.6.5.3(2) is a common mode of failure in push-test specimens
with narrow slabs (e.g. 300 mm, which has long been the standard width in British codes). It
was also found, in full-scale tests, to be the normal failure mode for composite L-beams
constructed with precast slabs.
79
Detailing rules are given in clause 6.6.5.3(2) for slabs where
the edge distance e in Fig. 6.18 is less than 300 mm. The required area of bottom transverse
reinforcement, A
b
, per unit length of beam, should be found using clause 6.6.6. In the
unhaunched slab shown in Fig. 6.18, failure surface bb will be critical (unless the slab is very
thick) because the shear on surface aa is low in an L-beam with an asymmetrical concrete
flange.
To ensure that the reinforcement is fully anchored to the left of the line aa, it is
recommended that U-bars be used. These can be in a horizontal plane or, where top
reinforcement is needed, in a vertical plane.
No rules are given for the effectiveness as transverse reinforcement of profiled sheeting
transverse to the supporting beams, with ribs that extend to a cantilever edge of the slab. The
length needed to develop the full tensile resistance of the sheeting will be known from the
design procedure for the composite slab. It is always greater than the minimum edge
distance of 6d, and usually greater than 300 mm. Where it exceeds the length e available,
bottom transverse reinforcement will be needed. The situation can be improved by placing
all the connectors near the inner edge of the steel flange. Sheeting with ribs parallel to the
free edge should not be assumed to resist longitudinal splitting of the slab.
78
b
b
a
a
d
e ( 6 d )
0.5 d
Fig. 6.18. Longitudinal shear reinforcement in an L-beam
90
12 May 2004 11:52:46
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Lying studs
There is also a risk of splitting where the shank of a stud (a lying stud) is parallel and close to
a free surface of the slab, as shown, for example, in Fig. 6.19. The U-bars should then be in a
vertical plane. Research has found
74
that the 6d rule for edge distance in clause 6.6.5.3(2) and
the provision of hoops may not be sufficient to ensure that the design shear resistance of the
stud is reached. The height of the stud and the longitudinal bars shown are also important.
The slip capacity may be less than 6 mm, so the shear connection may not be ductile. The
stud shown in Fig. 6.19(b) may be subjected to simultaneous axial tension and vertical and
longitudinal shear, so details of this type should be avoided.
Lying studs are outside the scope of EN1994-1-1. There is aninformative annex inEN1994-2.
Reinforcement at the end of a cantilever
Clause 6.6.5.3(3)P
At the end of a composite cantilever, the force on the concrete from the connectors acts
towards the nearest edge of the slab. The effects of shrinkage and temperature can add
further stresses
37
that tend to cause splitting in region B in Fig. 6.20, so reinforcement in this
region needs careful detailing. Clause 6.6.5.3(3)P can be satisfied by providing herring-bone
bottom reinforcement (ABC in Fig. 6.20) sufficient to anchor the force from the connectors
into the slab, and to ensure that the longitudinal bars provided to resist that force are
anchored beyond their intersection with ABC.
The situation where a column is supported at point Bis particularly critical. Incompatibility
between the vertical stiffnesses of the steel beam and the slab can cause local shear failure of
the slab, even where the steel beam alone is locally strong enough to carry the column.
80
Haunches
Clause 6.6.5.4 The detailing rules of clause 6.6.5.4 are based on limited test evidence, but are long-
established.
81
In regions of high longitudinal shear, deep haunches should be used with
caution because there may be little warning of failure.
Haunches formed by profiled sheeting are considered under clause 6.6.4.1.
79
6d
6d
(a) (b)
Fig. 6.19. Examples of details susceptible to longitudinal splitting
Studs
Steel
beam
A
C
B
Transverse reinforcement not shown
Slab
Fig. 6.20. Reinforcement at the end of a cantilever
91
12 May 2004 11:52:47
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Maximum spacing of connectors
Situations where the stability of a concrete slab is ensured by its connection to a steel beam
are unlikely to occur in buildings. The converse situation, stabilization of the steel flange, is
of interest only where the steel compression flange is not already in Class 1, so it rarely arises
where rolled I-sections are used.
Where the steel beam is a plate girder, there is unlikely to be any need for a top flange in
Class 1. Its proportions can usually be chosen such that it is in Class 2, unless a wide thin
flange is needed to avoid lateral buckling during construction.
Clause 6.6.5.5(2) Clause 6.6.5.5(2) is not restrictive in practice. As an example, a plate girder is considered,
in steel with f
y
= 355 N/mm
2
, where the top flange has t
f
= 20 mm, an overall breadth of
350 mm, and an outstand c of 165 mm. The ratio is 0.81 and the slenderness is
c/t
f
= 165/(20 0.81) = 10.2
so, from Table 5.2 of EN 1993-1-1, the flange is in Class 3. From clause 6.6.5.5(2), it can be
assumed to be in Class 1 if shear connectors are provided within 146 mmof each free edge, at
a longitudinal spacing not exceeding 356 mm, for a solid slab.
The ratio 22 in this clause is based on the assumption that the steel flange cannot buckle
towards the slab. Where there are transverse ribs (e.g. due to the use of profiled sheeting),
the assumption may not be correct, so the ratio is reduced to 15. The maximum spacing in
this example is then 243 mm.
The ratio 9 for edge distance, used in the formula 9t
f
(235/f
y
)
0.5
is the same as in
BS 5400: Part 5,
82
and the ratio 22 for longitudinal spacing is the same as the ratio for
staggered rows given in BS 5400.
Clause 6.6.5.5(3) The maximum longitudinal spacing in buildings, given in clause 6.6.5.5(3), 6h
c
, 800 mm,
is more liberal than the general rule of BS 5950-3-1,
31
though that code permits a conditional
increase to 8h
c
. The rule of EN 1994-1-1 is based on behaviour observed in tests, particularly
those where composite slabs have studs in alternate ribs only. Some uplift then occurs at
intermediate ribs. Spacing at 800 mm would in some situations allow shear connection
only in every third rib, and there was a requirement in ENV 1994-1-1 for anchorage,
but not necessarily shear connection, to be provided in every rib. It is expected that
this requirement will be included in the forthcoming EN 1090 for execution of steel
structures.
Detailing, for stud connectors
Clause 6.6.5.6
Clause 6.6.5.7
The rules of clause 6.6.5.6 are intended to prevent local overstress of a steel flange near a
shear connector and to avoid problems with stud welding. Application rules for minimum
flange thickness are given in clause 6.6.5.7. Clauses 6.6.5.7(1) and 6.6.5.7(2) are concerned
with resistance to uplift. Rules for resistance of studs, minimum cover and projection
of studs above bottom reinforcement usually lead to the use of studs of height greater
than 3d.
Clause 6.6.5.7(3) The limit 1.5 for the ratio d/t
f
in clause 6.6.5.7(3) influences the design of shear connection
for closed-top box girders in bridges. For buildings, the more liberal limit of clause 6.6.5.7(5)
normally applies. It has appeared in several earlier codes.
Clause 6.6.5.7(4) In clause 6.6.5.7(4), the minimum lateral spacing of studs in solid slabs has been reduced
to 2.5d, compared with the 4d of BS 5950-3-1. Although connection to precast slabs is outside
the scope of EN1994-1-1, this facilitates the use of large precast slabs supported on the edges
of the steel flanges, with projecting U-bars that loop over pairs of studs. There is much
experience, validated by tests,
83
of this form of construction, especially in multistorey car
parks. Closely spaced pairs of studs must be well confined laterally. The term solid slabs
should therefore be understood to exclude haunches.
Clause 6.6.5.8
Further detailing rules for studs placed within troughs of profiled sheeting are given in
clause 6.6.5.8. The first two concern resistance to uplift and compaction of concrete around
studs.
80
92
12 May 2004 11:52:47
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
81
Clause 6.6.5.8(3) Clause 6.6.5.8(3) relates to a common situation: where a small stiffening rib in sheeting
prevents the placing of studs centrally within a trough. It would be prudent to locate studs on
the favourable side (see Fig. 6.15(a)), which, for symmetrical loading on a simply-supported
span, is the side nearest the nearer support. This was not given as an application rule because
it is difficult to ensure that the favourable side would be correctly chosen on site. Instead,
alternate-side placing is recommended, with alternate studs on the unfavourable side (see
Fig. 6.15(c)). However, research has found
77
that the mean strength of pairs of studs placed
on the two sides of a trough is about 5% less than if both were central, with a greater
reduction for sheeting less than 1 mm thick. Sheeting profiles with troughs wide enough to
have off-centre stiffening ribs are more suitable for composite slabs than those with central
ribs. This clause does not refer to layouts that require two studs per trough. These are
discussed in comments on clause 6.6.4.2
6.6.6. Longitudinal shear in concrete slabs
Clause 6.6.6 The subject of clause 6.6.6 is the avoidance of local failure of a concrete flange near the
shear connection, by the provision of appropriate reinforcement. These bars enhance the
resistance of a thin concrete slab to in-plane shear in the same way that stirrups strengthen a
concrete web in vertical shear. Transverse reinforcement is also needed to control and limit
the longitudinal splitting of the slab that can be caused by local forces from individual
connectors. In this respect, the detailing problem is more acute than in the flanges of
concrete T-beams, where the shear from the web is applied more uniformly.
The principal change from earlier codes is that the equations for the required area
of transverse reinforcement have been replaced by cross-reference to EN 1992-1-1. Its
provisions are based on a truss analogy, as before, but a more general version of it, in which
the angle between members of the truss can be chosen by the designer. It is an application of
strut-and-tie modelling, which is widely used in EN 1992-1-1.
Clause 6.6.6.1(2)P The definitions of shear surfaces in clause 6.6.6.1(2)P and the basic design method are as
before. The method of presentation reflects the need to separate the general provisions
(clauses 6.6.6.1 to 6.6.6.3) from those restricted to buildings (clause 6.6.6.4).
Clause 6.6.6.1(4) Clause 6.6.6.1(4) requires the design longitudinal shear to be consistent with that used for
the design of the shear connectors. This means that the distribution along the beam of
resistance to in-plane shear in the slab should be the same as that assumed for the design of
the shear connection. For example, uniform resistance to longitudinal shear flow (v
L
) should
be provided where the connectors are uniformly spaced, even if the vertical shear over the
length considered is not constant. It does not mean, for example, that if, for reasons
concerning detailing, v
L, Rd
= 1.3v
L, Ed
for the connectors, the transverse reinforcement must
provide the same degree of over-strength.
Clause 6.6.6.1(5) In applying clause 6.6.6.1(5), it is sufficiently accurate to assume that longitudinal bending
stress in the concrete flange is constant across its effective width, and zero outside it. The
clause is relevant, for example, to finding the shear on the plane aa in the haunched beam
shown in Fig. 6.15, which, for a symmetrical flange, is less than half of the shear resisted by
the connectors.
Resistance of a concrete flange to longitudinal shear
Clause 6.6.6.2(1) Clause 6.6.6.2(1) refers to clause 6.2.4 of EN 1992-1-1, which is written for a design
longitudinal shear stress v
Ed
acting on a cross-section of thickness h
f
. The clause requires the
area of transverse reinforcement A
sf
at spacing s
f
to satisfy
A
sf
f
yd
/s
f
> v
Ed
h
f
/cot
f
(6.21 in EN 1992-1-1)
and the longitudinal shear stress to satisfy
v
Ed
< f
cd
sin
f
cos
f
(6.22 in EN 1992-1-1)
where = 0.6(1 f
ck
/250), with f
ck
in units of newtons per square millimetre. (The Greek
93
12 May 2004 11:52:47
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
82
letter used here in EN 1992-1-1 should not be confused with the Roman letter v, which is
used for shear stress.)
The angle
f
between the diagonal strut and the axis of the beam is chosen (within limits)
by the designer. The use of the method is illustrated in the following example.
Example 6.6: transverse reinforcement for longitudinal shear
Figure 6.21 shows a plan of an area ABCD of a concrete flange, assumed to be in
longitudinal compression, with shear stress v
Ed
and transverse reinforcing bars of area A
sf
at spacing s
f
. The shear force per transverse bar is
F
v
= v
Ed
h
f
s
f
acting on side AB of the rectangle shown. It is transferred to side CD by a concrete strut
AC at angle
f
to AB, and with edges that pass through the mid-points of AB, etc., as
shown, so that the width of the strut is s
f
sin
f
.
For equilibrium at A, the force in the strut is
F
c
= F
v
sec
f
(D6.18)
For equilibrium at C, the force in the transverse bar BC is
F
t
= F
c
sin
f
= F
v
tan
f
(D6.19)
For minimum area of transverse reinforcement,
f
should be chosen to be as small
as possible. For a flange in compression, the limits to
f
given in clause 6.2.4(4) of
EN 1992-1-1 are
45
f
26.5 (D6.20)
so the initial choice for
f
is 26.5. Then, from equation (D6.19),
F
t
= 0.5F
v
(D6.21)
From equation (6.22) above,
v
Ed
< 0.40f
cd
If this inequality is satisfied, then the value chosen for
f
is satisfactory. However, let us
assume that the concrete strut is over-stressed, because
v
Ed
= 0.48f
cd
To satisfy equation (6.22)
C D
A B
F
v
F
c
F
v
F
t
q
f
s
f
sin q
f
s
f
Fig. 6.21. Truss analogy for in-plane shear in a concrete flange
94
12 May 2004 11:52:48
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
83
sin
f
cos
f
0.48
whence
f
37
The designer increases
f
to 40, which satisfies expression (D6.20). From equation
(D6.19),
F
t
= F
v
tan 40 = 0.84F
v
From equation (D6.21), the change in
f
, made to limit the compressive stress in the
concrete strut AC, increases the required area of transverse reinforcement by 68%.
The lengths of the sides of the triangle ABC in Fig. 6.21 are proportional to the forces
F
v
, F
t
and F
c
. For given F
v
and s
f
, increasing
f
increases F
c
, but for
f
< 45 (the maximum
value permitted), the increase in the width s
f
sin
f
is greater, so the stress in the concrete
is reduced.
Shear planes
Clause 6.6.6.1(3) refers to shear surfaces. Those of types bb, cc, and dd in Fig. 6.15 are
different from the type aa surface because they resist almost the whole longitudinal shear,
not (typically) about half of it. The relevant reinforcement intersects themtwice, as shown by
the factor 2 in the table in Fig. 6.15.
For a surface of type cc in a beam with the steel section near one edge of the concrete
flange, it is clearly wrong to assume that half of the shear crosses half of the surface cc.
However, in this situation the shear on the adjacent plane of type aa will govern, so the
method is not unsafe.
Minimum transverse reinforcement
Clause 6.6.6.3 Clause 6.6.6.3 on this subject is discussed under clause 6.6.5.1 on resistance to separation.
Clause 6.6.6.4 Longitudinal shear in beams with composite slabs (clause 6.6.6.4)
Clause 6.6.6.4(1)
Design rules are given for sheeting with troughs that run either transverse to the span of the
steel beam, or parallel to it. Where they intersect a beam at some other angle, one can either
use the more adverse of the two sets of rules, or combine them in an appropriate way. The
rule for thickness of concrete in clause 6.6.6.4(1) is independent of the direction of the
sheeting.
Clause 6.6.6.4(2) In clause 6.6.6.4(2), the type bb shear surface is as shown in Fig. 6.16, in which the
labelling of the shear surfaces differs from that in Fig. 6.15.
Clause 6.6.6.4(4)
Clause 6.6.6.4(5)
The contribution made by sheeting to resistance in longitudinal shear depends not only on its
direction but also on whether the designer can determine the position of the ends of
individual sheets, and whether these ends are attached to the steel beam by through-deck
welding of stud connectors. If these decisions are made subsequently by the contractor, the
designer may be unable to use transverse sheeting as reinforcement for shear in the plane of
the slab. Its contribution to shear resistance is substantial where it is continuous across
the beam (clause 6.6.6.4(4)) and is useful where through-deck welding is used (clause
6.6.6.4(5)).This is applied in Example 6.7.
Sheets attached only by fixings used for erection and those with their span parallel to that
of the beam are ineffective as transverse reinforcement.
Through-deck welding of studs is also used to enhance the resistance of a composite slab
to longitudinal shear. EN 1994-1-1 does not state whether both procedures can be used at
once.
The question is discussed with reference to Fig. 6.22, which shows an exploded view of the
base of a stud welded to a steel flange through a layer of sheeting, which spans towards the
right-hand side of the diagram. The concrete in the mid-span region of the composite slab
95
12 May 2004 11:52:48
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
84
applies the force A to the upper part of the stud, which enables it to anchor a tension A in the
sheeting. The force B is due to the action of the steel top flange and the sheeting as
equivalent transverse reinforcement.
It is clear from Fig. 6.22 that forces A and B are not additive for the stud, because A exists
above the sheeting and B below it, but they are additive for the sheeting. The model in clause
9.7.4(3) for the calculation of resistance P
pb, Rd
is not shear failure of the stud but bearing
failure of the sheeting, which depends on the distance fromthe stud to the edge of the sheet.
This analysis shows that if the two procedures are used for the same stud, the available
resistance P
pb, Rd
should be split between them, in whatever ratio may be required.
Example 6.7: two-span beam with a composite slab ultimate limit state
A floor structure for a department store consists of composite beams of uniform
cross-section at 2.50 mspacing, fully continuous over two equal spans of 12.0 m. The floor
consists of a composite slab of overall thickness 130 mm, spanning between the beams.
The three supports for each beam may be treated as point supports, providing lateral and
vertical restraint. A design is required for an internal beam, subjected to vertical loading
only. The design service life is 50 years.
This particular concept is unlikely to be used in practice. The design will be found to be
governed by resistance at the internal support, and only a small part of the bending
resistance at mid-span can be used. However, it illustrates the use of many of the
provisions of EN 1994-1-1, and is chosen for that reason. The design of the composite
slab, the checking of the beam for serviceability limit states, and the influence of
semi-continuous beam-to-column connections are the subjects of subsequent worked
examples in this and later chapters of this guide.
It will be assumed initially that unpropped construction is used, and that the whole 24 m
length of concrete flange is cast before significant composite action is developed in any of
this length.
Loadings and materials
The characteristic imposed load, including partitions, is
q
k
= 7.0 kN/m
2
F
= 1.50
where
F
is a partial safety factor. The floor finish adds
g
k1
= 1.20 kN/m
2
F
= 1.35
These are applied to the composite structure. For simplicity, the floor finish will be
treated in global analyses as an imposed load, which is conservative.
Stud
A
A
A
B
B
B
A + B
Sheeting
Flange
Fig. 6.22. A through-deck welded stud acting as an end anchorage for a composite slab, and also
providing continuity for transverse reinforcement
96
12 May 2004 11:52:48
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
85
Preliminary studies have led to the choice of lightweight-aggregate concrete, of density
class 1.8 and strength class LC25/28. For an oven-dry density not exceeding 1800 kg/m
3
,
Table 11.1 in clause 11.3 of EN 1992-1-1 gives the design density of this reinforced
concrete as 1950 kg/m
3
, which is now assumed to include the sheeting.
The chosen steel section is IPE450, with the dimensions shown in Fig. 6.23. Its weight is
g
k3
= 0.76 kN/m, with
F
= 1.35. Table 6.2 gives the loadings for a beamspacing of 2.5 m.
At this stage, the shape for the profiled sheeting is assumed to be as in Fig. 6.23. The
mean thickness of the floor is 0.105 m, giving a characteristic load
g
k2
= 0.105 1.95 9.81 = 2.01 kN/m
2
with a partial safety factor
F
= 1.35.
Properties of materials
Structural steel: grade S355, with
M
= 1.0, so
Table 6.2. Loadings per unit length of beam
Load (kN/m)
Characteristic
Ultimate
(minimum)
Ultimate
(maximum)
Composite slab 5.02 6.78 6.78
Steel beam 0.76 1.02 1.02
Total, on steel beam 5.78 7.80 7.80
Floor finishes 1.20 1.62 1.62
Imposed load 17.50 0 26.25
Total, for composite beam 18.7 1.62 27.9
Total, for vertical shear 24.5 9.42 35.7
b
eff
= 1600
80
50
450
14.6
9.4
IPE 450
379
21
190
12 at 125
12 at 200
E
E
30
100 75 25
42
88
1.0
(b)
A
12 000 12 000
D
D
B C
(c)
(a)
Fig. 6.23. Elevation and cross-sections of the two-span beam (dimensions in mm). (a) Section
DD. (b) Section EE. (c) Elevation of beam
97
12 May 2004 11:52:49
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
86
f
y
= f
yd
= 355 N/mm
2
(D6.22)
and
= (235/255) = 0.81
Concrete: f
ck
= 25 N/mm
2
,
C
= 1.5,
0.85f
cd
= 14.2 N/mm
2
(D6.23)
From clause 11.3.1 of EN 1992-1-1,
1
= 0.4 + 0.6 1800/2200 = 0.891, so the mean
tensile strength is
f
lctm
= 0.891 2.6 = 2.32 N/mm
2
(D6.24)
From clause 11.3.2 of EN 1992-1-1, E
lcm
= 31 (18/22)
2
= 20.7 kN/mm
2
, so the short-
term modular ratio is
n
0
= 210/20.7 = 10.1
From informative Annex C, and a dry environment, the nominal total final free shrinkage
strain is
cs
= 500 10
6
Reinforcement: f
sk
= 500 N/mm
2
,
S
= 1.15,
f
sd
= 435 N/mm
2
Ductility: to be in Class B or C, from clause 5.5.1(5).
Shear connectors: it is assumed that 19 mm studs will be used, welded through the steel
sheeting, with ultimate tensile strength
f
u
= 500 N/mm
2
and
V
= 1.25.
Durability
The floor finish is assumed to be such that the top of the slab is exposed to low air
humidity. From clause 4.2 of EN 1992-1-1, the exposure class is XC1. The minimum
cover (clause 4.4.1) is then 15 mmfor a service life of 50 years, plus a tolerance of between
5 and 10 mm that depends on the quality assurance system, and is here taken as 9 mm. At
the internal support, the 12 mm longitudinal bars are placed above the transverse bars,
with a top cover of 24 mm, as shown in Fig. 6.23.
Properties of the IPE 450 cross-section
From section tables:
Area: A
a
= 9880 mm
2
fillet radius: r = 21 mm
Second moments of area: 10
6
I
ay
= 337.4 mm
4
10
6
I
az
= 16.8 mm
4
Torsional moment of area: 10
6
I
at
= 0.659 mm
4
Radii of gyration: i
y
= 185 mm i
z
= 41.2 mm i
x
= 190 mm
Section moduli: 10
6
W
ay
= 1.50 mm
3
10
6
W
az
= 0.176 mm
3
Plastic section modulus: 10
6
W
pl, a, y
= 1.702 mm
3
(D6.25)
The factor 10
6
is used to enable moments in kN m to be related to stresses in N/mm
2
without further adjustment, and because it gives numbers of convenient size.
98
12 May 2004 11:52:49
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
87
Effective widths of concrete flange
In Fig. 5.1 of clause 5.4.1.2, L
1
= L
2
= 12.0 m; L
e
for b
eff,1
= 10.2 m. Assume b
0
= 0.1 m;
then
b
1
= b
2
= 2.5/2 0.05 = 1.20 m
At mid-span,
b
eff
= 0.1 + 2 10.2/8 = 2.65 m (but 2.5 m)
so b
ei
= 2.4/2 = 1.2 m. At the internal support, L
e
for b
eff, 2
= 0.25 24 = 6 m;
b
eff
= 0.1 + 2 6/8 = 1.60 m
At an end support, b
ei
= 1.20 m;
i
= 0.55 + 0.025 10.2/1.20 = 0.762
b
eff
= 0.1 + 2 0.762 1.2 = 1.83 m
Classification of composite cross-section
The class of the web is quite sensitive to the area of longitudinal reinforcement in the slab
at the internal support. It is (inconveniently) necessary to assume a value for this before
the checks that govern it can be made. Large-diameter bars may not give sufficient control
of crack widths, so the reinforcement is assumed to be 12 mm bars at 125 mm, giving 13
bars within a 1.625 m width, so
A
s
= 1470 mm
2
The effective area of concrete slab is 1.6 m by 80 mm, so
A
s
/A
c
=
s
= 0.0113
The requirement of clause 5.5.1(5) for minimum
s
will be checked during the design for
crack-width control; see Example 7.1.
The force in these bars at yield is
F
s, y
= 1470 0.435 = 639 kN
Assuming a Class 2 section, the stress distribution for M
pl, Rd
is needed. Starting from
the stress distribution for M
pl, a, Rd
, the depth of steel web that changes from tension to
compression is
639/(9.4 2 0.355) = 96 mm
For classification, clause 5.5.1(1)P refers to clause 5.5.2 of EN 1993-1-1. Its Table 5.2
defines the depth of web, c, as that bounded by the root radii. Here, c = 379 mm, of which
the depth in compression is
379/2 + 96 = 285 mm
whence
= 285/379 = 0.75
This exceeds c/2, so from Table 5.2, for Class 2,
c/t 456 /(13 1) = 42.2
The actual c/t = 381/9.4 = 40.5; so at the internal support, the web is Class 2. At mid-span,
it is obviously in Class 1 or 2. For the compression flange, from Table 5.2,
c = (190 9.4)/23 21 = 69.3 mm
Hence
99
12 May 2004 11:52:49
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
88
c/t = 69.3/(14.6 0.81) = 5.86
The limit for a Class 1 flange is 9.0; so the bottom flange is Class 1.
The member is Class 2 at the internal support and Class 1 or 2 at mid-span.
It follows from clause 5.4.2.4(2) that for global analyses for ultimate limit states,
(provided that lateraltorsional buckling does not govern) the whole of the loading may
be assumed to act on the composite member; and fromclause 6.2.1.1(1)P that rigid-plastic
theory may be used for resistances to bending at all cross-sections of the beam.
The use of partial shear connection (clause 6.2.1.3(1)) is limited to sagging regions, in
accordance with clause 6.6.2.2, which is for beams in which plastic theory is used for
resistance of cross-sections (i.e. sections in Class 1 or 2).
Design is thus much simpler when there are no beams with cross-sections in Class 3 or 4.
This can usually be achieved in buildings, but rarely in multi-span bridges.
It is notable that if the steel top flange were in Class 3, its connection to the slab would
not enable it to be upgraded, because the condition of clause 6.6.5.5(2) is that the spacing
of shear connectors does not exceed 15t
f
, which is 15 14.6 0.81 = 177 mm. This
would be impracticable with the sheeting profile shown in Fig. 6.23.
Plastic resistance to bending (clause 6.2.1.2)
At the internal support, it has been found (above) that a 96 mm depth of the upper half of
the web is in compression. The design plastic resistance to hogging bending is that of the
steel section plus the effect of the reinforcing bars, shown in Fig. 6.24(a):
M
pl, a, Rd
= 1.702 0.355 = 604 kN m (hogging and sagging)
M
pl, Rd
= 604 + 639 0.277 = 781 kN m (hogging)
The characteristic plastic resistance is also needed. With
S
= 1 for the reinforcement, its
force at yield increases to 639 1.15 = 735 kN; the depth of web to change stress is
110 mm; and, by the method used for M
pl, Rd
,
M
pl, Rk
= 802 kN m (hogging)
All of these resistances may need to be reduced to allow for lateraltorsional buckling or
vertical shear.
For sagging bending at mid-span, reinforcement in compression is ignored, and the
available area of concrete is 2.5 m wide and 80 mm thick. If it is all stressed to 0.85f
cd
, the
compressive force is
F
c
= 14.2 2.5 80 = 2840 kN (D6.26)
If the whole steel section is at yield, the tensile force is
100
225
96
604 kN m
639 kN
639 kN
277
1600
5
3507 kN
2840 kN
93
667 kN
222
2500
90
225
(a) (b)
Fig. 6.24. Plastic moments of resistance in (a) hogging and (b) sagging bending (dimensions in mm)
100
12 May 2004 11:52:50
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
89
F
a
= 9880 0.355 = 3507 kN
Assuming full shear connection, with the plastic neutral axis in the steel top flange, the
thickness of flange to change from tension to compression is
t
f, c
= (3507 2840)/(2 0.355 190) = 5.0 mm
The longitudinal forces are then as shown in Fig. 6.24(b), and
M
pl, Rd
= 2840 0.315 + 667 0.222 = 1043 kN m (sagging) (D6.27)
This resistance will be reduced later by the use of partial shear connection.
Plastic resistance to vertical shear
Clause 6.2.2.2 refers to clause 6.2.6 of EN 1993-1-1. This defines the shear area of a rolled
I-section as
A
v
= A 2bt
f
+ (t
w
+ 2r)t
f
= 9880 380 14.6 + 51.4 14.6 = 5082 mm
2
and gives the design plastic shear resistance as
V
pl, Rd
= A
v
( f
y
/3)/
M0
= 5082 0.355/3 = 1042 kN
For shear buckling, clause 6.2.2.3 refers to Section 5 of EN 1993-1-5. No buckling check is
required, because h
w
/t for the steel web, based on the depth between the flanges, is 45,
below the limit of 48.6.
Flexural properties of elastic cross-sections
Several sets of elastic properties are needed, even where the steel beam is of uniform
section, because of changes of modular ratio and effective width, and the use of cracked
and uncracked sections. Here, slab reinforcement is ignored in the uncracked analyses.
The error is conservative (except for the shear connection) and usually very small. It is
convenient to calculate all these properties at the outset (Table 6.3).
From clause 5.4.2.2(11), creep may be allowed for by using a modular ratio n = 2n
0
=
20.2 for both short-term and long-term loadings. Results for n = n
0
are included in Table
6.3 for use in Chapter 7. Effects of shrinkage are unusually high in this example. For these,
it helps to use the more accurate modular ratio 28.7, as explained later.
The calculation for the uncracked properties at the internal support, with n = 10.1, is
nowgiven, as an example. In steel units, the width of the slab is 1.6/10.1 = 0.158 m, so the
composite section is as shown in Fig. 6.25. Its properties are:
Area:
A = 9880 + 158 80 = 9880 + 12640 = 22 520 mm
2
Height of neutral axis above centre of steel section:
z
na
= 12 640 315/22 520 = 177 mm
Table 6.3. Elastic section properties of the composite cross-section
Cross-section
Modular
ratio
b
eff
(m)
Neutral
axis (mm)
10
-6
I
y
(mm
4
)
10
-6
W
c, top
(mm
3
)
(1) Support, cracked, reinforced 1.6 42 467
(2) Support, uncracked 10.1 1.6 177 894 50.7
(3) Support, uncracked 20.2 1.6 123 718 62.5
(4) Mid-span, uncracked 10.1 2.5 210 996 69.5
101
12 May 2004 11:52:50
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
90
Second moment of area:
10
6
I
y
= 337.4 + 9880 0.177
2
+ 12 640 0.138
2
= 894 mm
4
Flexural stiffness:
10
6
E
a
I
y
= 210 894 = 187 700 kN mm
2
Section modulus, top of slab, in concrete units:
10
6
W
c, top
= 894 10.1/178 = 50.7 mm
3
Similar calculations were done for the other elastic section properties required. The
results are given in Table 6.3.
Global analysis
The resistance to lateraltorsional buckling of the beam near the internal support, M
b, Rd
,
depends on the bending-moment distribution, so global analysis is done next.
Wherever possible, the rules for global analysis in Section 5 apply to both ultimate and
serviceability limit states. Where alternatives are permitted, both types of limit state
should be considered before making a choice.
For this beam, there is no need to take account of the flexibility of joints (clause 5.1.2).
First-order elastic theory may be used (clauses 5.2.2(1) and 5.4.1.1(1)).
Clause 5.2.2(4) on imperfections is satisfied, because lateraltorsional buckling is the
only type of instability that need be considered, and its resistance formulae take account
of imperfections.
The simplest method of allowing for cracking, in clause 5.4.2.3(3), is applicable here,
and will be used. Cracking almost always occurs in continuous beams. In this example,
calculations for uncracked sections, using the longer method of clause 5.4.2.3(2), found
that for the ultimate imposed loading on both spans, the tensile stress in the concrete at
the internal support exceeded three times its tensile strength. This ignored shrinkage,
which further increases the tension.
For this beam, clause 5.4.2.3(3) requires the use of cracked section properties over a
length of 1.8 m each side of the internal support. For global analysis, clause 5.4.1.2(4)
permits the use of the mid-span effective width for the whole span; but here b
eff
for
the cracked region is taken as 1.6 m, because resistances are based on this width and
reinforcement outside it may be quite light.
The proposed use of n = 20.2 for the uncracked region merits discussion. Most of the
permanent load is applied to the steel beam, which does not creep, but resistances are
based on the whole of the load acting on the composite beam. Bending moment at the
internal support governs. Creep increases this, and the long-term effects of shrinkage are
so significant that the case t is more critical than t 0.
Neutral axis
138
9880
177
158
90
225
12 640
Fig. 6.25. Uncracked composite section at internal support, with n
0
= 10.1 (dimensions in mm)
102
12 May 2004 11:52:50
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
91
For the quasi-permanent combination, the coefficient
2
for variable load in a department
store is given in clause A1.2.2(1) of EN 1990 as 0.6, and 0.6q
k
is 40% of 1.5q
k
, so some
creep of the composite member is likely.
In this example, n = 2n
0
= 20.2 will be used for all action effects except shrinkage, for
which a more accurate value is determined, as follows.
Modular ratio for the effects of shrinkage
From clause 5.4.2.2(4), the age of loading can be assumed to be 1 day. Creep coefficients
for normal-weight concrete are given in Fig. 3.1 of EN1992-1-1 in terms of h
0
, the notional
size of the cross-section. For a slab with both surfaces exposed, this equals the slab
thickness; but the slab here has one sealed surface, and h
0
is twice its thickness. The mean
thickness (see Fig. 6.23) is 105 mm, so h
0
= 210 mm.
For normally hardening cement and inside conditions, Fig. 3.1 of EN 1992-1-1
gives the creep coefficient (, t
0
) as 5.0. For lightweight concrete, clause 11.3.3(1) of
EN 1992-1-1 gives a correction factor, which in this case is (18/22)
2
, giving = 3.35. The
creep multiplier
L
in clause 5.4.2.2(2) takes account of the shape of the stresstime curve
for the effect considered, and is 0.55 for shrinkage. The modular ratio for shrinkage
effects is
n = n
0
(1 +
L
t
) = 10.1(1 + 0.55 3.35) = 28.7
Bending moments
Although, through cracking, the beam is of non-uniform section, calculation of bending
moments algebraically is straightforward, because the two spans are equal. When both
spans are fully loaded, only a propped cantilever need be considered (Fig. 6.26).
For distributed loading w per unit length, and ratio of flexural stiffnesses , as shown,
an equation for the elastic bending moment M
Ed
at B is derived in the first edition of
Johnson and Buckby
37
(p. 375). It is
M
Ed, B
= (wL
2
/4)(0.110 + 0.890)/(0.772 + 1.228)
The results are given in Table 6.4.
Hence, the design bending moment for the internal support, excluding shrinkage
effects, with n = 20.2, is
M
Ed, B
= 394 + 142 = 536 kN m
The vertical shear at the internal support is
M
Ed, B
0.85L
lI
I
B C w
0.15L
Fig. 6.26. Elastic propped cantilever with change of section at 0.15L
Table 6.4. Design ultimate bending moments at the fixed end of the propped cantilever
Loading w (kN/m) n 10
6
I
y
(mm
4
) M
Ed, B
(kN m)
Permanent 9.42 10.1 996 2.13 133
Permanent 9.42 20.2 828 1.77 142
Variable 26.25 10.1 996 2.13 370
Variable 26.25 20.2 828 1.77 394
103
12 May 2004 11:52:51
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
92
V
Ed, B
= (9.42 + 26.25) 6 + 536/12 = 259 kN
The plastic shear resistance of the IPE450 section, V
pl, Rd
, was found earlier to be 1042 kN.
From clause 6.2.2.4(1), bending resistance is not reduced by shear until V
Ed
> V
pl, Rd
/2, so
there is no reduction here.
Redistribution of moments is not used because clause 5.4.4(4) does not permit it if
allowance for lateraltorsional buckling is required.
Secondary effects of shrinkage
Shrinkage of the slab causes sagging curvature and shortening of the composite member,
the primary effects. In a continuous beam, the curvature causes bending moments and
shear forces, the secondary effects. In regions assumed to be cracked, both the curvature
and the stresses fromthe primary effects are neglected (clauses 5.4.2.2(8) and 6.2.1.5(5)).
The important secondary effect in this beam, a hogging bending moment at the internal
support, is now calculated. Shrinkage is a permanent action, and so is not reduced by a
combination factor
0
.
The slab is imagined to be separated fromthe steel beam, Fig. 6.27(a). Its area A
c
is that
of the concrete above the sheeting. It shrinks. A force is applied to extend it to its original
length. It is
F = A
c
(E
a
/n)|
cs
| (D6.28)
This acts at the centre of the slab, at a distance z
sh
above the centroid of the composite
section. The parts of the beamare reconnected. To restore equilibrium, an opposite force
F and a sagging moment Fz
sh
are applied to the composite section.
The radius of curvature of the uncracked part of the beam is
R = E
a
I
y
/Fz
sh
(D6.29)
If the centre support is removed, the deflection at that point is
= (0.85L)
2
/2R (D6.30)
from the geometry of the circle (Fig. 6.27(b)).
F
z
sh
Fz
sh
F
R
0.85L 0.85L 0.30L
(a) (b)
P/2
M
sh, B
0.85L 0.15L
P
(d)
(c)
P/2 P/2
d
d
d
Fig. 6.27. Secondary effects of shrinkage
104
12 May 2004 11:52:51
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
93
It remains to calculate the force P, applied at that point, to reduce the deflection to
zero, so that the centre support can be replaced (Fig. 6.27(d)). The secondary hogging
bending moment at B is
M
Ed, sh, B
= PL/2 (D6.31)
and the vertical shear is P/2.
Using slope and deflection coefficients for a cantilever (Fig. 6.27(c)), P can be found
from the result
= (P/2)L
3
(0.13 + 0.20)/(E
a
I
y
) (D6.32)
The calculation is as follows:
A
c
= 2.5 0.08 = 0.20 m
2
E
a
= 210 kN/mm
2
, n = 28.7 and
cs
= 500 10
6
; so, fromequation (D6.28), F = 732 kN.
From Table 6.3,
I
y
= 741 10
6
mm
4
z
sh
= 225 40 = 185 mm
so, from equation (D6.29), R = 1149 m, and, from equation (D6.30) with L = 12 m,
= 45.3 mm. From Table 6.3,
I
y
= 467 10
6
mm
4
so
= 741/467 = 1.587
so, from equation (D6.32),
P/2 = 10.0 kN
and
M
Ed, sh, B
= 10 12 = 120 kN m
Should shrinkage be neglected at ultimate load?
This unusually high value, 120 kN m, results from the use of a material with high shrinkage
and a continuous beam with two equal spans. It increases the ultimate design bending
moment at B by 22%. Clause 5.4.2.2(7) permits this to be neglected if resistance is not
influenced by lateraltorsional buckling.
The reasoning is that as the bending resistances of the sections are determined by
plastic theory, the ultimate condition approaches a collapse mechanism, in which elastic
deformations (e.g. those from shrinkage) become negligible in comparison with total
deformations.
However, if the resistance at the internal support is governed by lateraltorsional
buckling, and if the buckling moment (to be calculated) is far below the plastic moment,
the inelastic behaviour may not be sufficient for the shrinkage effects to become negligible,
before failure occurs at the internal support. Hence, the secondary shrinkage moment is
not neglected at this stage, even though this beam happens to have a large reserve of
strength at mid-span, and would not fail until the support section was far into the
post-buckling phase.
Resistance to lateraltorsional buckling
The top flange of the steel beam is restrained in both position and direction by the
composite slab. Lateral buckling of the bottom flange near the internal support is
accompanied by bending of the web, so the problem here is distortional lateral buckling.
The provisions in EN 1994-1-1 headed lateraltorsional buckling (clause 6.4), are in
fact for distortional buckling. Clause 6.4.1(3) permits, as an alternative, use of the
provisions in EN 1993-1-1 for steel beams. The method of clause 6.4.2 is used here. The
105
12 May 2004 11:52:51
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
94
detailed comments on it, both in the main text and in Appendix A, should be referred to as
required. The method requires the calculation of the elastic critical buckling moment at
the internal support, M
cr
, for which information is given in tables in ENV 1994-1-1
(reproduced as graphs in Appendix A (Figs A.3 and A.4)). The simple method of clause
6.4.3 is not available, as the loading does not conformto paragraph (b) of clause 6.4.3(1).
Buckling near an internal support is often most critical in a span with minimum load
that is adjacent to a fully loaded span. In this beamit was found that although the buckling
moment M
b, Rd
is increased when both spans are loaded (because the length of bottom
flange in compression is reduced), the increase in the applied moment M
Ed
is greater, so
the both-spans-loaded case is more critical. This case is now considered, with n = 20.2
and all load assumed to act on the composite member.
From Table 6.4, the bending moments in the beam are as shown in Fig. 6.28. Shrinkage
is considered separately. The simply-supported moment M
0
is 35.7 12
2
/8 = 643 kN m,
so, from Fig. A.3 for C
4
(see Appendix A),
= M
B
/M
0
= 536/643 = 0.834 and C
4
= 28.3
The elastic critical buckling moment was given earlier as
M
cr
= (k
c
C
4
/L)[(G
a
I
at
+ k
s
L
2
/
2
)E
a
I
afz
]
1/2
(D6.11)
where k
c
is a property of the composite section, given in Section 6.4.2, G
a
is the shear
modulus for steel,
G
a
= E
a
/[2(1 +
a
)] = 80.8 kN/mm
2
I
at
is the torsional moment of area of the steel section,
I
at
= 0.659 10
6
mm
4
k
s
is defined in clause 6.4.2(6), and I
afz
is the minor-axis second moment of area of the steel
bottom flange,
10
6
I
afz
= 1.90
3
14.6/12 = 8.345 mm
4
The stiffness k
s
is now found. It depends on the lesser of the cracked flexural stiffnesses
of the composite slab at a support and at mid-span, (EI)
2
. The value at the support
governs. An approximation for this, derived in Appendix A, is
(EI)
2
= E
a
[A
s
A
e
z
2
/(A
s
+ A
e
) + A
e
h
p
2
/12]
where A
e
is the equivalent transformed area per unit width of concrete in compression,
A
e
= b
0
h
p
/nb
s
536
339
403
494
B
0
A C
4.75 m
5.26 m
Full load on both spans
Variable load on span AB only
D
7.25 m
Fig. 6.28. Bending-moment distributions for ultimate limit state, excluding shrinkage
106
12 May 2004 11:52:51
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
95
where b
0
is the mean width of the troughs, b
s
is the spacing of the troughs, h
p
is the depth of
the sheeting, A
s
is the area of top reinforcement per unit width of slab and z is the lever
arm, as shown in Fig. A.1 of Appendix A.
Calculation of (EI)
2
for the composite slab, and k
s
It is assumed that the transverse reinforcement above the steel beam will be below the
longitudinal bars and not less than 12 mm bars at 200 mm, giving A
s
= 565 mm
2
/m and
d
s
= 42 mm, whence z = 63 mm (Fig. 6.29).
Assuming that buckling is caused by a short-termoverload, n is taken as 10.1. FromFig.
6.23, b
0
/b
s
= 0.5; h
p
= 50 mm; so A
e
= 2475 mm
2
/m. Hence,
(EI)
2
= 210[0.565 2.475 63
2
/3040 + 2.475 50
2
/12 000] = 491 kN m
2
/m
From clause 6.4.2(6), for unit width of a slab continuous across the steel beams at spacing
a, and assuming that the conditions for using = 4 apply,
k
1
= 4(EI)
2
/a = 4 491/2.5 = 786 kN m/rad
per metre width, and
k
2
= E
a
t
w
3
/[4h
s
(1
a
2
)]
where h
s
is the distance between the centres of the flanges of the IPE450 section, 435 mm.
Thus,
k
2
= 210 9.4
3
/(4 435 0.91) = 110 kN/rad
and
k
s
= k
1
k
2
/(k
1
+ k
2
) = 786 110/896 = 96.4 kN/rad
Calculation of k
c
For a doubly symmetrical steel section, from equations (D6.12) and (D6.13),
k
c
= (h
s
I
y
/I
ay
)/[(h
s
2
/4 + i
x
2
)/e + h
s
]
with
e = AI
ay
/[A
a
z
c
(A A
a
)]
In these expressions, the symbols are properties of the steel section, given earlier, except
that A is the area of the cracked composite section,
A = A
a
+ A
s
= 11 350 mm
2
b
p
b
0
d
s
A
s
h
p
/2
z
Fig. 6.29. Cross-section of the composite slab
107
12 May 2004 11:52:52
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
96
and z
c
is the distance between the centroid of the steel beam and mid-depth of the slab.
Here, slab means the 130 mm-deep composite slab, not the 80 mm depth of concrete
that contributes to the composite section. It is the stiffness of the composite slab in the
transverse direction that prevents rotation of the steel top flange; so
z
c
= 225 + 130/2 = 290 mm
Hence,
e = 11 350 337 10
6
/(9880 290 1470) = 909 mm
and
k
c
= (435 467/337)/[(218
2
+ 190
2
)/909 + 435] = 1.15
Calculation of M
cr
and M
b, Rd
From equation (D6.11) for M
cr
:
M
cr
= (1.15 28.3/12)[(80.8 0.659 + 96.4 12
2
/
2
) 210 8.345]
1/2
= 4340 kN m
From clause 6.4.2(4), the relative slenderness depends also on the characteristic resistance
moment, calculated earlier. The slenderness is
= (M
Rk
/M
cr
) = (802/4340)
1/2
= 0.430
For the reduction factor
LT
for a rolled section, clause 6.4.2(1) refers to clause 6.3.2.3 of
EN 1993-1-1, where buckling curve c is specified for this IPE section. This could be taken
to mean the curve c plotted in Fig. 6.4 of EN 1993-1-1; but that curve is inconsistent with
the equation for
LT
in clause 6.3.2.3. This is understood to mean that the value of
LT
given for curve c in Table 6.3 of EN 1993-1-1 should be used in calculating
LT
. Its value
depends on the parameters and , which can be given in a National Annex. The
recommended values, 0.4 and 0.75 respectively, are used here.
The calculation is
(The use of Fig. 6.4 gives the much lower result
LT
= 0.88.)
The buckling resistance is
M
b, Rd
=
LT
M
pl, Rd
= 0.983 781 = 767 kN m
This is well above the design ultimate moment with shrinkage included, M
Ed
= 656 kN m.
The result is quite sensitive to the values specified for and .
Clause 6.3.2.2 of EN 1993-1-1 includes a rule that if M
Ed
/M
cr
0.16, lateraltorsional
buckling effects may be ignored. This applies here: M
Ed
= 0.153M
cr
. The rule appears to
be linked to the use of = 0.4. It is not clear whether it applies if a National Annex
specifies a lower value.
The provision of bracing to the bottomflange is considered at the end of this example.
Design for sagging bending
The maximum sagging bending occurs in a span when the other span carries minimum
load, and is reduced by creep, so n = 10.1 is assumed. Removal of variable load from one
span halves the bending moment it causes at the internal support, so from Table 6.4 the
bending moment at the internal support is
M
Ed, B
= 133 + 370/2 = 318 kN m
For the span with loading 35.67 kN/m, the end reaction is
LT
2
LT LT LT LT, 0 LT
0.5[1 ( ) ] 0.577 = + - + =
LT
2 2
LT LT LT
1
0.983
[ ( )]

= =
+ -
LT, 0
LT, 0
LT, 0
108
12 May 2004 11:52:52
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
97
V
Ed, A
= 35.67 6 318/12 = 188 kN
so the point of maximum moment is at a distance 188/35.67 = 5.26 m from the support,
and the maximum sagging moment is
M
Ed
= 188 5.26/2 = 494 kN m
This is so far belowM
pl, Rd
, 1043 kN m, that the least permitted degree of shear connection
will be used. M
Ed
is even below M
pl, a, Rd
, which is 604 kN m.
From clause 6.6.1.2(1) the span length in sagging bending may be taken as 0.85L, or
10.2 m here. From equation (6.12) with f
y
= 355 N/mm
2
, the minimum degree of shear
connection is
= n/n
f
= 1 (0.75 0.03 10.2) = 0.56
Clause 6.6.1.2(3) permits a lower value, subject to some conditions. It was found that
one of these that there should be only one stud connector per rib of sheeting could not
be satisfied. The minimum number of connectors in each half of the sagging region is
therefore 0.56n
f
, where n
f
is the number for full shear connection. From Fig. 6.24(b), the
compressive force in the slab is then not less than 2840 0.56 = 1590 kN. Recalculation
of M
pl, Rd
, by the method used before gives
M
pl, Rd
= 946 kN m
which is almost twice the resistance required.
It does not follow that the design of this beam as non-composite would be satisfactory.
Example 7.1 shows that its deflection would probably be excessive.
Design of the shear connection
The degree of shear connection used here enables the studs to be treated as ductile. An
alternative design, using non-ductile connectors, is given in Example 6.8.
From clause 6.6.5.8(1) the height of the 19 mm studs must be at least 50 + 2 19 =
88 mm. The standard height of 95 mm, after welding, satisfies this rule.
The design shear resistance per stud is governed by equation (6.19) of clause 6.6.3.1(1):
P
Rd
= 0.29d
2
(f
ck
E
cm
)
1/2
/
V
= 0.29 19
2
(25 20 700)
1/2
/(1000 1.25) = 60.25 kN
This result is modified by a factor k
t
given in clause 6.6.4.2. It depends on the height of the
stud, h
sc
, the dimensions of the trough in the sheeting (see Fig. 6.23), the thickness of the
sheeting (assumed to be 1.0 mm) and the number of studs per trough, n
r
.
For n
r
= 1: k
t
= 0.7 100/50 (95/50 1) = 1.26, but 0.85
For n
r
= 2: k
t
= 1.26/2= 0.89, but 0.70
Provided that the studs are not also required to anchor the profiled sheeting, the
resistances are therefore
P
Rd, 1
= 0.85 60.25 = 51.2 kN (D6.33)
P
Rd, 2
= 0.7 60.25 = 42.2 kN (D6.34)
so that a trough with two studs provides the equivalent of 2 42.2/51.2 = 1.65 single
studs.
From Fig. 6.28, the studs for maximum sagging bending have to be provided within a
length of 5.26 m. The troughs are spaced at 0.2 m, so 26 are available. The design
compressive force in the slab is 1590 kN, so the design shear flow is
1590/5.2 = 306 kN/m
The number of single studs required is
109
12 May 2004 11:52:52
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
98
n
s
= 1590/51.2 = 31
Hence, two studs per trough have to be used over part of the span.
For a shear span in a building, EN 1994-1-1 does not specify how non-uniform shear
connection should be arranged. Slip is minimized when the density of the connection is
related to the shear per unit length, so the regions with pairs of studs should be adjacent to
the supports.
If the minimum number of troughs with two studs is n
2s
,
1.65n
2s
+ 26 n
2s
= 31
whence
n
2s
7.8
For the cracked section in hogging bending, A
s
f
sd
= 639 kN, which requires 12.5 single
studs. When the hogging moment is a maximum, the distance from an internal support to
the cross-section of maximum sagging moment is 7.25 m (Fig. 6.28), so 36 troughs are
available for a total of 31 + 12.5 = 43.5 single studs.
If the minimum number of troughs with two studs is n
2h
,
1.65n
2h
+ 36 n
2h
= 43.5
whence
n
2h
11.6
The design shear flow is
(1590 + 639)/7.2 = 310 kN/m
The arrangement of studs shown in Fig. 6.30 provides the equivalent of 31.2 and 43.8
studs, within the lengths 5.2 and 7.2 m, respectively.
The maximum sagging and hogging bending moments are caused by different
arrangements of imposed loading. This method of calculation takes account of this, with
the result that two studs near mid-span are effective for both sagging and hogging
resistance.
It is quicker to assume that the maximum sagging and hogging bending moments are
caused by a single loading. The resulting increase in the total number of studs is negligible.
The disadvantage is that it is not clear how many of the troughs with two studs per trough
should be near each end of the span.
256
422
100
n
L
(kN/m)
n
L, Rd
A B
1.6 4.8 9.6
Number of
troughs
8
22 19.8
Equivalent number of single studs
306
5.2 12.0
Distance from
support A (m)
16 2 22 12
16 2 13.2
310
n
L, Ed
for maximum
n
L, Ed
for maximum
sagging BM
hogging BM
Fig. 6.30. Arrangement of stud connectors in one 12 m span
110
12 May 2004 11:52:53
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
99
Design of the transverse reinforcement
Clause 6.6.6.1(4) says that the design longitudinal shear for the concrete slab should be
consistent with the design and spacing of the shear connectors. This is taken to mean that
the resistance of the shear connection, rather than the design loading, determines the
longitudinal shear. Its maximum occurs where there are two studs per trough, and is
v
L, Ed
= 10 42.2 = 422 kN/m
From clause 6.6.6.4(2), shear surfaces that pass closely around a stud need not be
considered. The critical situation is thus where sheeting is not continuous across a beam.
It is assumed to be anchored by a stud, as shown in Fig. 6.31. From symmetry, the critical
shear plane, labelled aa, is to be designed to resist 211 kN/m.
The shear resisted by the sheeting is given in clause 6.6.6.4(5). For the design bearing
resistance of the sheeting it refers to clause 9.7.4. For the detailing shown in Fig. 6.31, the
end distance a is 40 mm. The diameter of the weld collar is taken as
1.1 19 = 20.9 mm
whence k
in clause 9.7.4(3) is
k
= 1 + 40/20.9 = 2.91
The thickness of the profiled sheeting is shown in Fig. 6.23 as 1.0 mm; but the composite
slab has not yet been designed. Here, it is assumed to be at least 0.9 mm thick, with a yield
strength of 350 N/mm
2
and
M
= 1.0. From equation (9.10),
P
pb, Rd
= k
d
d0
tf
yp, d
= 2.91 20.9 0.9 0.35 = 19.1 kN/stud
From clause 6.6.6.4(5), with a stud spacing of 200 mm, the shear resistance provided by
the sheeting is
v
L, pd, Rd
= 19.1/0.2 = 95 kN/m
This must not exceed the yield strength of the sheeting, A
p
f
yp, d
, which for this sheeting is
over 400 kN/m.
The design shear for the concrete slab, of thickness 80 mm, is
v
L, Ed
= 211 95 = 116 kN/m
Clause 6.6.6.2(1) refers to clause 6.2.4 of EN 1992-1-1, for shear between web and
flanges of T-sections. The method is a truss analogy, with some choice provided for the
angle
f
between the concrete diagonals of the truss and the longitudinal direction. The
35 40 20
50
45
a
a
Fig. 6.31. Detail of a stud welded through discontinuous profiled sheeting (dimensions in mm)
111
12 May 2004 11:52:53
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
100
reinforcement required increases with the angle
f
. For simplicity, the minimum angle
given in clause 6.2.4(4) of EN 1992-1-1 for a tension flange, 38.6, will be used for the
whole length of the span.
The area of transverse reinforcement per unit length is given by
A
sf
> v
L, Ed
/(f
sd
cot
f
) = 116/(0. 435 1.25) = 213 mm
2
/m
This is much less than the amount assumed in design for lateral buckling (565 mm
2
/m). It
may also be affected by the need for crack control above the beam, which arises in the
design of the composite slab.
Design of bracing to bottom flanges near the internal support
It is nowshown that the values of factor C
4
given in Appendix Aare not suitable for design
based on M
cr
where intermediate lateral bracing is provided.
From Fig. 6.28, with full loading on both spans, the distance of a point of contraflexure
from support B is 12 2 4.75 = 2.5 m. Let us suppose that lateral bracing is to be
provided at this point, so that L in equation (D6.11) is reduced from 12 to 2.5 m.
Assuming that the bending-moment distribution over this 2.5 m length is linear, Fig. A.4
in Appendix A gives C
4
= 11.1. Substituting these values into equation (D.6.11), with
other values unchanged, gives M
cr
= 2285 kN m, which is much lower than the previous
value, 4340 kN m.
For this situation, it is necessary to find M
cr
from elastic critical analysis by computer.
Summary of Example 6.7
All important aspects of the design of this beam for persistent situations for ultimate limit
states have nowbeen considered. Serviceability checks are given in the worked example at
the end of Chapter 7. The same beam with semi-continuous joints at support B is studied
in Examples 8.1 and 10.1.
Example 6.8: partial shear connection with non-ductile connectors
The design bending-moment diagram for sagging bending in the preceding worked
example is shown in Fig. 6.28. The shear connection for the length AD of span AB was
designed for connectors that satisfied the definition of ductile in clause 6.6.1.2. The
result is shown in Fig. 6.30.
This work is now repeated, using the same data, except that the proposed connectors
are not ductile, to illustrate the use of clause 6.6.1.3(5). This requires the calculation of
the shear flow v
L, Ed
by elastic theory. No inelastic redistribution of shear is required, so
clause 6.6.1.1(3)P on deformation capacity does not apply.
Calculation of the resistance moment M
Rd
according to clause 6.2.1.3(3) is not permitted,
so stresses are calculated by elastic theory and checked against the limits in clause
6.2.1.5(2). From equations (D6.22) and (D6.23), these are
f
cd
= f
ck
/
C
= 25/1.5 = 16.7 N/mm
2
(D6.35)
and
f
yd
= f
yk
= 355 N/mm
2
(D6.36)
In this continuous beam, creep reduces stiffness at mid-span more than at the internal
support, where the concrete is cracked, so the sagging bending moment and longitudinal
shear (for constant loading) reduce over time. The short-term modular ratio, 10.1, is
therefore used. Taking account of the use of unpropped construction, it is found for the
loads in Table 6.2 that the maximum sagging bending moment acting on the composite
section, M
c, Ed
, occurs at 5.4 m from support A (Fig. 6.28), and is
M
c, Ed
= 404 kN m with M
a, Ed
= 110 kN m
112
12 May 2004 11:52:53
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
101
Shrinkage of concrete in this beam reduces both the mid-span moments and the
compressive stress in concrete. For simplicity and safety, its effects are ignored.
Using the elastic section properties fromrow4 of Table 6.3, the stresses are found to be
well below the limits given above. The mean stress in the 80 mm thickness of the concrete
slab, with b
eff
= 2.5 m, is 4.20 N/mm
2
, giving the longitudinal force in the slab as
N
c
= 4.2 2.5 80 = 840 kN
From equation (D6.26), the force for full shear connection is
N
c, f
= F
c
= 2840 kN
so the degree of shear connection needed is
= N
c
/N
c, f
= 840/2840 = 0.30
The elastic shear flow diagram is triangular, so the shear flow at support A is
v
L, Ed
= 2 840/5.4 = 311 kN/m
Stud connectors will be used, as before. They are not ductile at this lowdegree of shear
connection. Their resistances are given by equations (D6.33) and (D6.34). For the
profiled sheeting used (Fig. 6.23(b)), there are five troughs per metre. One stud per
trough (v
L, Rd
= 256 kN/m) is not sufficient near support A. Two studs per trough
provide 422 kN/m. Near mid-span, it is possible to use one stud every other trough,
(v
L, Rd
= 128 kN/m) as their spacing, 400 mm, is less than the limit set in clause 6.6.5.5(3).
Details of a possible layout of studs are shown in Fig. 6.32. This also shows the
resistance provided over this 5.4 m length from Fig. 6.30. That is higher because in the
previous example, = 0.56, not 0.30, and shear connection is provided for the whole load,
not just that applied to the composite member.
This is not a typical result, because the design sagging moment here is unusually low, in
relation to the plastic resistance to sagging bending. It may become more typical in design
to the Eurocodes, because of the influence of lateraltorsional buckling on accounting for
shrinkage and the restrictions on redistribution of moments.
Example 6.9: elastic resistance to bending, and influence of degree of shear
connection and type of connector on bending resistance
This example makes use of the properties of materials and cross-sections found in
Example 6.7 on the two-span beam, and of the results from Example 6.8 on non-ductile
256
311
422
128
1.2 3.8 5.4
0
n
L
(kN/m)
For ductile studs,
from Example 6.7
For elastic design,
from Example 6.8
No. of troughs 6 8 13
Distance from
support A (m)
n
L, Ed
n
L, Rd
Fig. 6.32. Longitudinal shear and shear resistance for length AD of the beam in Fig. 6.28
113
12 May 2004 11:52:53
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
102
connectors. The cross-section of maximum sagging bending moment, point D, was then
found to be 5.4 m from support A in Fig. 6.28.
For simplicity, the beam is now assumed to be simply-supported and of span 10.8 m, so
that this cross-section is at mid-span, with unpropped construction as before, so that the
bending moment in the steel beam at this section is M
a, Ed
= 110 kN m. Shrinkage effects
are beneficial, and are neglected.
The purpose is to find the relationship between the degree of shear connection, , and
the bending resistance of the beamat mid-span, in accordance with clauses 6.2.1.36.2.1.5,
6.6.1.2 and 6.6.1.3. The result is shown in Fig. 6.33.
At low degrees of shear connection, only elastic design is permitted, so the elastic
bending resistance M
el, Rd
is required, to clause 6.2.1.4(6). It depends on the modular ratio.
It is found that the limiting stress (equations (D6.35) and (D6.36)) is reached first in the
steel bottom flange, and is increased by creep, so n = 20.2 is assumed.
For M
a, Ed
= 110 kN m, the maximum stress in the steel beam is 73 N/mm
2
, leaving
355 73 = 282 N/mm
2
for loading on the composite beam. Using the elastic section
properties from row 5 of Table 6.3, the steel reaches yield when the bending moment on
the composite section is 610 kN m, so
M
el, Rd
= 110 + 610 = 720 kN m
The compressive force in the concrete slab is then N
c, el
= 1148 kN, so that
= N
c, el
/N
c, f
= 1148/2840 = 0.404
Figure 6.33 is based on Figs 6.5 and 6.6(b) of EN 1994-1-1. The results above enable
point B to be plotted.
When M
Ed
= 110 kN m, N
c
= 0. From equation (6.2) in clause 6.2.1.4(6), line BE is
drawn towards the point (0, 110) as shown. No lower limit to is defined. It will in practice
be determined by the detailing rules for the shear connection.
For full shear connection,
M
Rd
= 950
500

M
a, Ed
= 110
0 1.0
M
pl, a, Rd
= 604
M
pl, Rd
= 1043
0.404 0.574 0.2 0.8
M
el, Rd
= 720
E
B
G
D
C
A
Bending moment (kN m)
h = N
c
/N
c, f
Fig. 6.33. Design methods for partial shear connection
114
12 May 2004 11:52:54
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
103
M
pl, Rd
= 1043 kN m
fromequation (D6.27). This fixes point Cin Fig. 6.33, and line BC(fromequation (6.3)) is
drawn. Line EBC applies whether the connection is ductile or not.
For this beam, with f
y
= 355 N/mm
2
and equal steel flanges, equation (6.12) in clause
6.6.1.2(1) gives
1 (0.75 0.03 10.8) = 0.574
Hence,
N
c
0.574 2840 = 1630 kN
Using the method of clause 6.2.1.3(3) for ductile connectors, illustrated in Fig. 6.3, the
plastic bending resistance is
M
Rd
= 950 kN m
This gives point D in Fig. 6.33.
When = 0, the plastic resistance is M
pl, a, Rd
. From equation (D6.25) this is
M
pl, a, Rd
= 1.702 355 = 604 kN m
which is point A in Fig. 6.33.
Similar calculations for assumed degrees of shear connection give the curve ADC,
which can, for simplicity, be replaced by the line AC (clause 6.2.1.3(5)). Both the curve
and the line are valid only where is high enough for the connection to be ductile.
The design bending resistances for this example are therefore given by EBGDC in Fig.
6.33. Line BE gives the least shear connection that may be used when M
Ed
< M
el, Rd
,
without restriction on the type of connector.
For higher values of M
Ed
, the required degree of shear connection for non-ductile
connectors is given by the line BC. The bonus for using ductile connectors (as defined in
clause 6.6.1.2) is the area GDC, where the position of the line GD is determined by the
span of the beam, and moves to the right as the span increases.
For the beam analysed here, with M
a, Ed
= 110 kN m, a total bending moment
M
Ed
= 1000 kN m(for example) requires shear connection for about 2100 kN( 0.74) if
headed studs to clause 6.6.1.2 are used, but this increases to over 2600 kN for non-ductile
connectors.
6.7. Composite columns and composite compression
members
6.7.1. General
Scope
A composite column is defined in clause 1.5.2.5 as a composite member subjected mainly to
compression or to compression and bending. The title of clause 6.7 includes compression
members, to make clear that its scope is not limited to vertical members but includes, for
example, composite members in triangulated or Vierendeel girders. These girders may also
have composite tension members, for which provisions are given in EN 1994-2.
In this guide, references to columns includes other composite compression members,
unless noted otherwise, and for buildings, column means a length of column between
adjacent lateral restraints; typically, a storey height.
Design rules for columns sometimes refer to effective length. That term is not generally
used in clause 6.7. Instead, the relative slenderness is defined, in clause 6.7.3.3(2), in terms
of N
cr
, the elastic critical normal force for the relevant buckling mode.
This use of N
cr
is explained in the comments on clause 6.7.3.3.
Clause 6.7.1(1)P Clause 6.7.1(1)P refers to Fig. 6.17, in which all the sections shown have double symmetry;
but clause 6.7.1(6) makes clear that the scope of the general method of clause 6.7.2 includes
members of non-symmetrical section.
84
115
12 May 2004 11:52:54
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
The bending moment in a column depends on the assumed location of the line of action
of the axial force, N. Where the cross-section has double symmetry, this is the intersection of
the axes of symmetry. In other cases the choice, made in the modelling for global analysis,
should be retained for the analysis of the cross-sections. A small degree of asymmetry (e.g.
due to an embedded pipe) can be allowed for by ignoring in calculations concrete areas
elsewhere, such that symmetry is restored.
No shear connectors are shown in the cross-sections in Fig. 6.17, because within a column
length the longitudinal shear is normally much lower than in a beam, and sufficient interaction
may be provided by bond or friction. Shear connectors should be provided for load introduction,
following clause 6.7.4.
The minimum compression for a member to be regarded as a column, rather than a beam,
is not stated. As shown in Example 6.11, the use of the cross-sections in Fig. 6.17 as beams
without shear connectors is usually prevented by the low design shear strengths due to bond
and friction (clause 6.7.4.3).
Clause 6.7.1(2)P The strengths of materials in clause 6.7.1(2)P are as for beams, except that class C60/75
and lightweight-aggregate concretes are excluded. For these, additional provisions (e.g. for
creep, shrinkage, and strain capacity) would be required.
85,86
Clause 6.7.1(3) Clause 6.7.1(3) and clause 5.1.1(2) both concern the scope of EN 1994-1-1. They appear to
exclude composite columns in high-rise buildings with a reinforced concrete core. For these
mixed structures, additional consideration in the global analysis of the effects of shrinkage,
creep, and column shortening may be needed.
Clause 6.7.1(4) The steel contribution ratio (clause 6.7.1(4)), is essentially the proportion of the squash
load of the section that is provided by the structural steel member. If it is outside the limits
given, the member should be treated as reinforced concrete or as structural steel, as
appropriate.
Independent action effects
Clause 6.7.1(7)
The interaction curve for the resistance of a column cross-section to combined axial force N
and uniaxial bending M is shown in Fig. 6.19, and, as a polygon, in Fig. 6.38 of this guide. It
has a region BD where an increase in N
Ed
increases M
Rd
. Clause 6.7.1(7) refers to a situation
where at ultimate load the factored bending moment,
F
M
Ek
, could co-exist with an
independent axial force that was less than its design value,
F
N
Ek
. It says that verification
should be based on the lower value, 0.8
F
N
Ek
.
Discussion of this rule is illustrated in Fig. 6.34, which shows region BDC of the interaction
curve in Fig. 6.19, which is symmetrical about the line AD. If
104
N
pm, Rd
/2
g
F
N
Ek
0.8g
F
N
Ek
B
0
D
C
N
M
E
M
Rd
M
pl, Rd
A
Fig. 6.34. Independent bending moment and normal force (not to scale)
116
12 May 2004 11:52:55
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
F
N
Ek
< N
pm, Rd
/2 (D6.37)
then M
Rd
should be found for an axial force of 0.8
F
N
Ek
, as shown by point E. The reduction
in M
Rd
is usually small.
Asimpler but more conservative rule is given in ENV1994-1-1, with a clearer definition of
independent actions: if M
Rd
corresponding to
F
N
Ek
is found to exceed M
pl, Rd
, M
Rd
should
be taken as M
pl, Rd
. It is applicable unless the bending moment M
Ed
is due solely to the
eccentricity of the force N
Ed
. Its effect is to replace the curve BDC in Fig. 6.34 with the line
BC.
Local buckling
Clause 6.7.1(8)P The principle of clause 6.7.1(8)P is followed by its application rules. They ensure that the
concrete (which will be reinforced in accordance with EN 1992-1-1) restrains the steel and
prevents it from buckling even when yielded.
For partly encased sections, the encasement prevents local buckling of the steel web, and
prevents rotation of the steel flange at its junction with the web, so that a higher b
f
/t ratio may
be used than for a bare steel section. Table 6.3 gives the limit as 44, compared with about 22
(fromEN1993-1-1) for a Class 2 flange. (In EN1994, as in EN1993, = (235/f
y
), in units of
newtons per square millimetre.)
For concrete-filled rectangular hollow steel sections (RHS), the limit of 52 compares
with about 41 for a steel RHS. For a concrete-filled circular hollow section, the limiting d/t
of 90
2
compares with 70
2
for Class 2 in EN 1993-1-1.
6.7.2. General method of design
Clause 6.7.2
Designers of composite columns will normally ensure that they fall within the scope of the
simplified method of clause 6.7.3; but occasionally the need arises for a non-uniform or
asymmetric member. The general method of clause 6.7.2 is provided both for this reason,
and to enable advanced software-based methods to be used.
Clause 6.7.2(3)P
Clause 6.7.2 is more a set of principles than a design method. Development of software
that satisfies these principles is a complex task. Clause 6.7.2(3)P refers to internal forces.
These are the action effects within the column length, found from those acting on its ends,
determined by global analysis to Section 5. At present, such an analysis is likely to exclude
member imperfections and second-order effects within members, but comprehensive software
may become available.
Clause 6.7.2(3) also refers to elastic-plastic analysis. This is defined in clause 1.5.6.10 of
EN 1990 as structural analysis that uses stress/strain or moment/curvature relationships
consisting of a linear elastic part followed by a plastic part with or without hardening.
As the three materials in a composite section follow different non-linear relationships,
direct analysis of cross-sections is not possible. One has first to assume the dimensions and
materials of the member, and then determine the axial force N and bending moment M at a
cross-section from assumed values of axial strain and curvature , using the relevant
material properties. The MN relationship for each section can be found from many such
calculations. This becomes even more complex where biaxial bending is present.
87
Integration along the length of the column then leads to a non-linear member stiffness
matrix that relates axial force and end moments to the axial change of length and end
rotations.
6.7.3. Simplified method of design
Scope of the simplified method
Clause 6.7.3.1
Clause 6.7.3.1(1)
The method has been calibrated by comparison with test results.
88
Its scope (clause 6.7.3.1) is
limited mainly by the range of results available, which leads to the restriction 2 in clause
6.7.3.1(1). For most columns, the method requires second-order analysis in which explicit
account is taken of imperfections. The use of strut curves is limited to axially loaded
members.
105
117
12 May 2004 11:52:55
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Clause 6.7.3.1(2)
The restriction on unconnected steel sections in clause 6.7.3.1(1) is to prevent loss of
stiffness due to slip, which would invalidate the formulae for EI of the column cross-section.
The limits to concrete cover in clause 6.7.3.1(2) arise from concern over strain softening of
concrete invalidating the interaction diagram (Fig. 6.19), and from the limited test data for
columns with thicker covers. These provisions normally ensure that for each axis of bending,
the flexural stiffness of the steel section makes a significant contribution to the total stiffness.
Greater cover can be used by ignoring in calculation the concrete that exceeds the stated
limits.
Clause 6.7.3.1(3) The limit of 6%in clause 6.7.3.1(3) on the reinforcement used in calculation is more liberal
than the 4% (except at laps) recommended in EN 1992-1-1. This limit and that on maximum
slenderness are unlikely to be restrictive in practice.
Clause 6.7.3.1(4) Clause 6.7.3.1(4) is intended to prevent the use of sections susceptible to lateraltorsional
buckling. The reference to h
c
< b
c
arises because h
c
is defined as the overall depth in the
direction normal to the major axis of the steel section (Fig. 6.17). The termmajor axis can be
misleading, because some column sections have I
z
> I
y
, even though I
a, y
> I
a, z
.
Resistance of cross-sections
Clause 6.7.3.2(1)
Calculations for composite sections, with three materials, are potentially more complex than
for reinforced concrete, so simplifications to some provisions of EN 1992-1-1 are made in
EN1994-1-1. Reference to the partial safety factors for the materials is avoided by specifying
resistances in terms of design values for strength, rather than characteristic values; for
example in equation (6.30) for plastic resistance to compression in clause 6.7.3.2(1). This
resistance, N
pl, Rd
, is the ultimate axial load that a short column can sustain, assuming that the
structural steel and reinforcement are yielding and the concrete is crushing.
For concrete-encased sections, the crushing stress is taken as 85% of the design cylinder
strength, as explained in the comments on clause 3.1. For concrete-filled sections, the
concrete component develops a higher strength because of the confinement from the steel
section, and the 15% reduction is not made; see also the comments on clause 6.7.3.2(6).
Resistance to combined compression and bending
Clause 6.7.3.2(2)
The bending resistance of a column cross-section, M
pl, Rd
, is calculated as for a composite
beamin Class 1 or 2 (clause 6.7.3.2(2)). Points on the interaction curve shown in Figs 6.18 and
6.19 represent limiting combinations of compressive axial load N and moment M which
correspond to the plastic resistance of the cross-section.
The resistance is found using rectangular stress blocks. For simplicity, that for the
concrete extends to the neutral axis, as shown in Fig. 6.35 for resistance to bending (point B
in Fig. 6.19 and Fig. 6.38). As explained in the comments on clause 3.1(1), this simplification
is unconservative in comparison with stress/strain curves for concrete and the rules of
EN1992-1-1. To compensate for this, the plastic resistance moment for the column section is
reduced by a factor
M
in clause 6.7.3.6(1).
106
0.85f
cd

f
yd
f
yd
f
sd
f
sd
M
pl, Rd
+
+
Reinforcement Steel Concrete
Fig. 6.35. Stress distributions for resistance in bending (tension positive)
118
12 May 2004 11:52:55
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
As axial compression increases, the neutral axis moves; for example, towards the lower
edge of the section shown in Fig. 6.35, and then outside the section. The interaction curve is
therefore determined by moving the neutral axis in increments across the section, and
finding pairs of values of M and N from the corresponding stress blocks. This requires a
computer program, unless the simplification given in clause 6.7.3.2(5) is used. Simplified
expressions for the coordinates of points B, C and D on the interaction curve in Fig. 6.34 are
given in Appendix C. Further comment is given in Examples 6.10 and C.1.
Influence of transverse shear
Clause 6.7.3.2(3)
Clause 6.7.3.2(4)
Clauses 6.7.3.2(3) and 6.7.3.2(4), on the influence of transverse shear on the interaction
curve, are generally the same as clause 6.2.2.4 on momentshear interaction in beams. One
assumes first that the shear V
Ed
acts on the structural steel section alone. If it is less than
0.5V
pl, a, Rd
, it has no effect. If it is greater, there is an option of sharing it between the steel and
reinforced concrete sections, which may reduce that acting on the steel to below0.5V
pl, a, Rd
. If
it does not, then a reduced design yield strength is used for the shear area, as for the web of a
beam. In a column, however, the shear area depends on the plane of bending considered,
and may consist of the flanges of the steel section. It is assumed that shear buckling does not
occur.
Simplified interaction curve
Clause 6.7.3.2(5) Clause 6.7.3.2(5) explains the use of the polygonal diagram BDCA in Fig. 6.19 as an
approximation to the interaction curve, suitable for hand calculation. The method applies to
any cross-section with biaxial symmetry, not just to encased I-sections.
First, the location of the neutral axis for pure bending is found, by equating the
longitudinal forces from the stress blocks on either side of it. Let this be at distance h
n
from
the centroid of the uncracked section, as shown in Fig. 6.19(B) and Fig. C.2 in Appendix C. It
is shown in Appendix C that the neutral axis for point C on the interaction diagram is at
distance h
n
on the other side of the centroid, and the neutral axis for point D passes through
the centroid. The values of M and N at each point are easily found from the stress blocks
shown in Fig. 6.19. For concrete-filled sections the factor 0.85 may be omitted.
Concrete-filled tubes of circular or rectangular cross-section
Clause 6.7.3.2(6) Clause 6.7.3.2(6) is based on the lateral expansion that occurs in concrete under axial
compression. This causes circumferential tension in the steel tube and triaxial compression
in the concrete. This increases the crushing strength of the concrete
88
to an extent that
outweighs the reduction in the effective yield strength of the steel in vertical compression.
The coefficients
a
and
c
given in this clause allow for these effects.
This containment effect is not present to the same extent in concrete-filled rectangular
tubes because less circumferential tension can be developed. In all tubes the effects of
containment reduce as bending moments are applied; this is because the mean compressive
strain in the concrete and the associated lateral expansion are reduced. With increasing
slenderness, bowing of the member under load increases the bending moment, and therefore
the effectiveness of containment is further reduced. For these reasons,
a
and
c
are
dependent on the eccentricity of loading and on the slenderness of the member.
Properties of the column
For columns in a frame, some properties of each column length are needed before or during
global analysis of the frame:
Clause 6.7.3.3(1) the steel contribution ratio (clause 6.7.3.3(1))
Clause 6.7.3.3(2) the relative slenderness (clause 6.7.3.3(2))
Clause 6.7.3.3(3) the effective flexural stiffnesses (clauses 6.7.3.3(3) and 6.7.3.4(2))
Clause 6.7.3.3(4) the creep coefficient and effective modulus for concrete (clause 6.7.3.3(4)).
The steel contribution ratio is explained in the comments on clause 6.7.1(4).
107
119
12 May 2004 11:52:55
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
The relative slenderness is needed to check that the column is within the scope of the
simplified method, clause 6.7.3.1(1). is calculated using characteristic values and the
appropriate flexural stiffness is that given in clauses 6.7.3.3(3) and 6.7.3.3(4). The correction
factor K
e
is to allow for cracking.
As depends on the elastic critical normal force for the relevant buckling mode, the
behaviour of the surrounding members needs to be taken into account. This could require a
calculation of the load factor
cr
for elastic instability, following the procedure shown in Fig.
5.1(e). It will often be possible though to make simplifying assumptions to show that a
proposed column is within the scope for the method. For example, in a frame with a high
stiffness against sway it would be reasonable to calculate N
cr
assuming the member to be
pin-ended. In an unbraced continuous frame, the stiffness of each beam could be taken as
that of the steel section alone, permitting N
cr
to be determined from effective length charts
that assume a beam to be of uniform stiffness. In any case, the upper limit on is somewhat
arbitrary and does not justify great precision in N
cr
.
The creep coefficient
t
influences the effective modulus E
c, eff
(clause 6.7.3.3(4)), and
hence the flexural stiffness of each column. It depends on the age at which concrete is
stressed and the duration of the load. These will not be the same for all the columns in a
frame. The effective modulus depends also on the proportion of the design axial load that is
permanent. The design of a column is rarely sensitive to the influence of the creep coefficient
on E
c, eff
, so conservative assumptions can be made about uncertainties. Normally, a single
value of effective modulus can be used for all the columns in the frame.
Verification of a column
The flow chart of Fig. 6.36 shows a possible calculation route, intended to minimize
iteration, for a column as part of a frame. It is used in Example 6.10. It is assumed that the
column has cross-section details that can satisfy clauses 6.7.1(9), 6.7.3.1(2) to 6.7.3.1(4) and
6.7.5.2(1), and has already been shown to have 2, so that it is within the scope of the
simplified method of clause 6.7.3.
The relationship between the analysis of a frame and the stability of individual members is
discussed in the comments on clauses 5.2.2(3) to 5.2.2(7). Conventionally, the stability of
members is checked by analysis of individual members, using end moments and forces from
global analysis of the frame. Figure 6.36 follows this procedure, giving in more detail the
procedures outlined in the lower part of the flow chart for global analysis (see Fig. 5.1(a)). It
is assumed that the slenderness of the column, determined according to clause 5.3.2.1(2), is
such that member imperfections have been neglected in the global analysis. Comments are
now given in the sequence of Fig. 6.36, rather than clause sequence. If bending is biaxial, the
chart is followed for each axis in turn, as noted. It is assumed that loading is applied to the
column only at its ends.
The starting point is the output from the global analysis, listed at the top of Fig. 6.36. The
design axial compression is the sum of the forces from the two frames of which the column is
assumed to be a member. If, at an end of the column (e.g. the top) the joints to beams in these
frames are at a different level, they could conservatively be assumed both to be at the upper
level, if the difference is small. The flow chart does not cover situations where the difference
is large (e.g. a storey height). The axial force N
Ed
is normally almost constant along the
column length. Where it varies, its maximum value can conservatively be assumed to be
applied at the upper end.
Clause 6.7.3.4
For most columns, the method requires a second-order analysis in which explicit account
is taken of imperfections (clause 6.7.3.4). However, for a member subject only to end
compression, clause 6.7.3.5(2) enables buckling curves to be used. For columns that qualify,
this is a useful simplification because these curves allow for member imperfections. The
reduction factor depends on the non-dimensional slenderness . The buckling curves are
also useful as a preliminary check for columns with end moment; if the resistance to the
normal force N
Ed
is not sufficient, the proposed column is clearly inadequate.
108
120
12 May 2004 11:52:56
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
109
Known, for both the y and z axes: N
Ed
, N
pl, Rd
, L, (EI)
eff, II
and
the bending moments at both ends of the column. Find the
axial compression: N
Ed
= N
Ed, y
+ N
Ed, z

Is the column in axial compression
only? (Clause 6.7.3.5(2))
Note 1: Until Note 2, the chart
is for both the y and z planes,
separately, and the y or z
subscript is not given
Find l to clause 6.7.3.3(2),
then c to clause 6.7.3.5(2).
Is N
Ed
cN
pl, Rd
?
Find V
pl, a, Rd
to clause 6.2.2.2(2).
Is V
Ed
> 0.5V
pl, a, Rd
? (Clause
6.7.3.2(3))
Find M
pl, a, Rd
and M
pl, Rd
, and hence V
a, Ed
and
V
c, Ed
from clause 6.7.3.2(4). Is V
a, Ed
> 0.5V
pl, a, Rd
?
Find r from clause 6.2.2.4(2) and
hence reduced f
yd
Find interaction curve or polygon for the
cross-section (clauses 6.7.3.2(2) and 6.7.3.2(5))
From clause 6.7.3.4(5): find
N
cr, eff
= p
2
(EI)
eff, II
/L
2

find b for end moments M
Ed, top

and M
Ed, bot
to Table 6.4, and hence
k (= k
1
); find k
2
for b = 1; find the
design moment for the column,
M
Ed, max
= k
1
M
Ed
+ k
2
N
Ed
e
0

M
Ed, top

M
Ed, bot
Apply bow imperfection to
clause 6.7.3.4(4); max. bow e
0
Can first-order member analysis be used, because the member satisfies clause
5.2.1(3), based on stiffness to clause 6.7.3.4?
Find M
Ed
, the maximum first-
order bending moment within
the column length. If M
Ed, 1
=
M
Ed, 2
it is
M
Ed, max
= M
Ed, 1
+ N
Ed
e
0
No
Find M
Ed, max
by second-
order analysis of the
pin-ended column length
with force N
Ed
and end
moments M
Ed, 1
and M
Ed, 2

Either Or
Note 2: For biaxial bending,
repeat steps since Note 1
for the other axis
From N
Ed
and the interaction diagrams, find m
dy
and m
dz
from clause 6.7.3.7(1). Check that the cross-section
can resist M
y, Ed, max
and M
z, Ed, max
(END)
Yes
No
Yes
Yes
No
Column not strong
enough
No
Column
verified
(END)
(END)
Yes
No
Yes
Fig. 6.36. Flow chart for verification of a column length

121
12 May 2004 11:52:56
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Comment has already been made on the calculation of when used to check the scope of
the method. When is used as the basis for resistance, calculation of this parameter may still
be simplified, provided the result is conservative.
Columns without end moments are exceptional, and for most members the design process
continues, as shown in Fig. 6.36. Columns with transverse shear exceeding half of the shear
resistance of the steel element are rare; but shear is checked next, because if it is high the
interaction curve for the cross-section may be affected. Comment has been given earlier on
clauses 6.7.3.2(3) and 6.7.3.2(4), on shear, and on the interaction curve or polygon.
Much of the remainder of the flow chart is concerned with finding the maximum bending
moment to which the column will be subjected. In general, two calculations are necessary
(clause 6.7.3.6(1)). The maximum bending moment may occur at one end, when the design
moment equals to the larger of the two end moments; or the maximum bending moment
may occur at an intermediate point along the member. This is because second-order effects
and lateral load within the length of the member and initial bowing affect the bending
moment.
Clause 6.7.3.4(4)
If member imperfections have been neglected in the global analysis, it is necessary to
include themin the analysis of the column. Clause 6.7.3.4(4) gives the member imperfections
in Table 6.5 as proportional to the length L of the column between lateral restraints. The
imperfection is the lateral departure at mid-height of the column of its axis of symmetry from
the line joining the centres of symmetry at the ends of the column. The values account
principally for truly geometric imperfections and for the effects of residual stresses. They do
not depend on the distribution of bending moment along the column. The curved shape is
usually assumed to be sinusoidal, but use of a circular arc would be acceptable. The curve is
assumed initially to lie in the plane of the frame being analysed.
The next step is to ascertain whether second-order effects need to be considered within
the member length. Clause 6.7.3.4(3) refers to clause 5.2.1(3). Second-order effects can
therefore be neglected if the load factor
cr
for elastic instability of the member exceeds 10.
Possible sway effects will have been determined by global analysis, and will already be
included within the values for the end moments and forces. To calculate
cr
, the ends of the
column are assumed to be pinned, and
cr
is found using the Euler formula N
cr, eff
=
2
EI/L
2
,
L being the physical length of the column. The flexural stiffness to use is (EI)
eff, II
(clause
6.7.3.4(2)), with the modulus of elasticity for concrete modified to take account of long-term
effects (clause 6.7.3.3(4)). This flexural stiffness is lower than that defined in clause 6.7.3.3(3)
because it is essentially a design value for ultimate limit states. The factor K
e, II
allows for
cracking. The factor K
o
is from research-based calibration studies.
The neglect of second-order effects does not mean that increase in bending moment
caused by the member imperfection can also be ignored. The next box, on finding M
Ed, max
,
gives the example of a column in uniform single-curvature bending. If the end moments are
dissimilar or of opposite sign, the maximummoment in the column, M
Ed, max
, is likely to be the
greater end moment.
Clause 6.7.3.4(5)
In practice, most columns are relatively slender, and second-order effects will usually need
to be included. This can be done by second-order analysis of the member, treated as
pin-ended but subject to the end moments and forces given by the global analysis. Any
intermediate loads would also be applied. The analysis is to obtain the maximum moment in
the column, which is taken as the design moment, M
Ed, max
. Formulae may be obtained from
the literature. Alternatively, use may be made of the factor k given by clause 6.7.3.4(5).
It is assumed in Fig. 6.36 that the column is free from intermediate lateral loads. Two
factors are used, written as k
1
and k
2
, because two moment distributions must be considered.
The first gives the equivalent moment k
1
M
Ed
in the perfect column, where M
Ed
is the larger
end moment given by the global analysis. The definitions of M
Ed
in clause 6.7.3.4(5) and Table
6.4 may appear contradictory. In the text before equation (6.43), M
Ed
is referred to as a
first-order moment. This is because it does not include second-order effects arising within
the column length. However, Table 6.4 makes clear that M
Ed
is to be determined by either
first-order or second-order global analysis; the choice would depend on clause 5.2.
110
122
12 May 2004 11:52:56
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
The multiplier (from Table 6.4) allows for the shape of the bending-moment diagram.
The condition 0.44 in the table is to ensure sufficient protection against snap-through
buckling in double-curvature bending.
The first-order moment from the imperfection, N
Ed
e
0
, has a distribution such that = 1,
from Table 6.4, so k
2
normally differs from k
1
. The imperfection can be in any direction, so
the equivalent moment k
2
N
Ed
e
0
always has the same sign as k
1
M
Ed
, when the two are
combined.
Equation (6.43) states that k must be greater than or equal to unity, and this is correct for a
single distribution of bending moment. However, for a combination of end moments and
member imperfection, it could be conservative to limit all values of k in this way. At
mid-length the component due to end moments depends on their ratio, r, and therefore
could be small. The appropriate component is therefore k
1
M
Ed
without the limit k 1.0, and
the design moment within the column length is (k
1
M
Ed
+ k
2
N
Ed
e
0
). In biaxial bending, the
initial member imperfection may be neglected in the less critical plane (clause 6.7.3.7(1)).
The limit k 1.0 is intended to ensure that the design moment is at least the larger end
moment M
Ed
.
Clause 6.7.3.6(1)
For uniaxial bending, the final step is to check that the cross-section can resist M
Ed, max
with
compression N
Ed
. The interaction diagram gives a resistance
d
M
pl, Rd
with axial load N
Ed
, as
shown in Fig. 6.18. This is unconservative, being based on rectangular stress blocks, as
explained in the comment on clause 3.1(1), so in clause 6.7.3.6(1) it is reduced, by the use of a
factor
M
that depends on the grade of structural steel. This factor allows for the increase in
the compressive strain in the cross-section at yield of the steel (which is adverse for the
concrete), when the yield strength of the steel is increased.
Biaxial bending
Clause 6.7.3.7 Where values of M
Ed, max
have been found for both axes, clause 6.7.3.7 applies, in which they
are written as M
y, Ed
and M
z, Ed
. If one is much greater than the other, the relevant check for
uniaxial bending, equation (6.46) will govern. Otherwise, the linear interaction given by
equation (6.47) applies. If the member fails this biaxial condition by a small margin, it may be
helpful to recalculate the less critical bending moment, omitting the member imperfection,
as permitted by clause 6.7.3.7(1).
6.7.4. Shear connection and load introduction
Load introduction
Clause 6.7.4.1(1)P
Clause 6.7.4.1(2)P
The provisions for the resistance of cross-sections of columns assume that no significant slip
occurs at the interface between the concrete and structural steel components. Clauses
6.7.4.1(1)P and 6.7.4.1(2)P give the principles for limiting slip to an insignificant level in the
critical regions: those where axial load and/or bending moments are applied to the column.
Clause 6.7.4.1(3)
For any assumed clearly defined load path it is possible to estimate stresses, including
shear at the interface. In regions of load introduction, shear stress is likely to exceed the
design shear strength from clause 6.7.4.3, and shear connection is then required (clause
6.7.4.2(1)). It is unlikely to be needed elsewhere, unless the shear strength
Rd
from Table 6.6
is very low, or the member is also acting as a beam, or has a high degree of double-curvature
bending. Clause 6.7.4.1(3) refers to the special case of an axially loaded column.
Few shear connectors reach their design shear strength until the slip is at least 1 mm; but
this is not significant slip for a resistance model based on plastic behaviour and rectangular
stress blocks. However, a long load path implies greater slip, so the assumed path should not
extend beyond the introduction length given in clause 6.7.4.2(2).
Where axial load is applied through a joint attached only to the steel component, the force
to be transferred to the concrete can be estimated from the relative axial loads in the two
materials given by the resistance model. Accurate calculation is rarely practicable where
the cross-section concerned does not govern the design of the column. In this partly
plastic situation, the more adverse of the elastic and fully plastic models gives a safe result
111
123
12 May 2004 11:52:57
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Clause 6.7.4.2(1) (clause 6.7.4.2(1), last line). In practice, it may be simpler to provide shear connection based
on a conservative (high) estimate of the force to be transferred.
Where axial force is applied by a plate bearing on both materials or on concrete only, the
proportion of the force resisted by the concrete gradually decreases, due to creep and
shrinkage. It could be inferred from clause 6.7.4.2(1) that shear connection should be
provided for a high proportion of the force applied. However, models based on elastic theory
are over-conservative in this inherently stable situation, where large strains are acceptable.
The application rules that follow are based mainly on tests.
Clause 6.7.4.2(3)
Clause 6.7.4.2(4)
In a concrete-filled tube, shrinkage effects are low, for only the autogenous shrinkage
strain occurs, with a long-term value below 10
4
, from clause 3.1.4(6) of EN 1992-1-1. Radial
shrinkage is outweighed by the lateral expansion of concrete in compression, for its inelastic
Poissons ratio increases at high compressive stress. Friction then provides significant shear
connection (clause 6.7.4.2(3)). Friction is also the basis for the enhanced resistance of stud
connectors given in clause 6.7.4.2(4).
Clause 6.7.4.2(5)
Clause 6.7.4.2(6)
Detailing at points of load introduction or change of cross-section is assisted by the high
bearing stresses given in clauses 6.7.4.2(5) and 6.7.4.2(6). As an example, the following data
are assumed for the detail shown in Fig. 6.22(B), with axial loading:
steel tube with external diameter 300 mm and wall thickness 10 mm
bearing plate 15 mm thick, with strength f
y
= f
yd
= 355 N/mm
2
concrete with f
ck
= 45 N/mm
2
and
C
= 1.5.
Then, A
c
= 140
2
= 61 600 mm
2
; A
1
= 15 280 = 4200 mm
2
. From equation (6.48),
c, Rd
/f
cd
= [1 + (4.9 10/300)(355/45)](14.7)
0.5
= 8.8
and
c, Rd
= 8.8 30 = 260 N/mm
2
This bearing stress is so high that the fin plate would need to be at least 180 mmdeep to have
sufficient resistance to vertical shear.
Clause 6.7.4.2(9) Figure 6.23 illustrates the requirement of clause 6.7.4.2(9) for transverse reinforcement,
which must have a resistance equal to the force N
c1
. If longitudinal reinforcement is ignored,
this is given by
N
c1
= A
c2
/2nA
where Ais the transformed area of the cross-section 11 of the column in Fig. 6.23, given by
A = A
s
+ (A
c1
+ A
c2
)/n
and A
c1
and A
c2
are the unshaded and shaded areas of concrete, respectively, in section 11.
Transverse shear
Clause 6.7.4.3 Clause 6.7.4.3 gives application rules (used in Example 6.11) relevant to the principle of
clause 6.7.4.1(2), for columns with the longitudinal shear that arises from transverse shear.
The design shear strengths
Rd
in Table 6.6 are far lower than the tensile strength of concrete.
They rely on friction, not bond, and are related to the extent to which separation at the
interface is prevented. For example, in partially encased I-sections, lateral expansion of
the concrete creates pressure on the flanges, but not on the web, for which
Rd
= 0; and the
highest shear strengths are for concrete within steel tubes.
Clause 6.7.4.3(4)
Where small steel I-sections are provided, mainly for erection, and the column is mainly
concrete, clause 6.7.4.3(4) provides a useful increase to
Rd
, for covers up to 115 mm, more
simply presented as
c
= 0.2 + c
z
/50 2.5
Clause 6.7.4.3(5)
Concern about the attachment of concrete to steel in partially encased I-sections appears
again in clause 6.7.4.3(5), because under weak-axis bending, separation tends to develop
between the encasement and the web.
112
124
12 May 2004 11:52:57
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
113
6.7.5. Detailing provisions
Clause 6.7.5.1(2)
If a steel I-section in an environment in class X0 to EN 1992-1-1 has links in contact with its
flange (permitted by clause 6.7.5.2(3)), the cover to the steel section could be as low as
25 mm. For a wide steel flange, this thin layer of concrete would have little resistance to
buckling outwards, so the minimum thickness is increased to 40 mm in clause 6.7.5.1(2).
Clause 6.7.5.2(1) Minimum longitudinal reinforcement (clause 6.7.5.2(1)), is needed to control the width of
cracks, which can be caused by shrinkage even in columns with concrete nominally in
compression.
Clause 6.7.5.2(4) Clause 6.7.5.2(4) refers to exposure class X0 of EN 1992-1-1. This is a very dry
environment, with no risk of corrosion or attack. Buildings with very low air humidity are
given as an example, so some buildings, or spaces in buildings, would not qualify. The
minimum reinforcement provides robustness during construction.
Example 6.10: composite column with bending about one or both axes
A composite column of length 7.0 m has the uniform cross-section shown in Fig. 6.37. Its
resistance to given action effects will be found. After checking that the column is within
the scope of the simplified method, the calculations follow the sequence of the flow chart
in Fig. 6.36. The properties of the steel member, 254 254 UC89, are taken from section
tables, and given here in Eurocode notation. The properties of the materials, in the usual
notation, are as follows:
Structural steel: grade S355, f
y
= f
yd
= 355 N/mm
2
, E
a
= 210 kN/mm
2
.
Concrete: C25/30; f
ck
= 25 N/mm
2
, f
cd
= 25/1.5 = 16.7 N/mm
2
,
0.85f
cd
= 14.2 N/mm
2
, E
cm
= 31 kN/mm
2
, n
0
= 210/31 = 6.77.
Reinforcement: ribbed bars, f
sk
= 500 N/mm
2
, f
sd
= 500/1.15 = 435 N/mm
2
.
Geometrical properties of the cross-section
In the notation of Fig. 6.17(a),
b
c
= h
c
= 400 mm b = 256 mm h = 260 mm
c
y
= 200 128 = 72 mm c
z
= 200 130 = 70 mm
These satisfy the conditions of clauses 6.7.3.1(2), 6.7.3.1(4) and 6.7.5.1(2), so all the
concrete casing is included in the calculations.
Area of reinforcement = 4 36 = 446 mm
2
Area of concrete = 400
2
11 400 446 = 148 150 mm
2
260
Steel section: 254 254 UC89
A
a
= 11 400 mm
2
10
6
I
a, y
= 143.1 mm
4
10
6
I
a, z
= 48.5 mm
4
10
6
W
pa, y
= 1.228 mm
3
10
6
W
pa, z
= 0.575 mm
3
70
256
400
17.3
72
10.5
400
12
Fig. 6.37. Cross-section and properties of a composite column (dimensions in mm)
125
12 May 2004 11:52:57
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
114
The reinforcement has area 0.301% of the concrete area, so clause 6.7.5.2(1) permits it to
be included in calculations. For simplicity, its small contribution will be ignored, so the
values are
A
a
= 11 400 mm
2
A
c
= 400
2
11 400 = 148 600 mm
2
A
s
= 0
For the steel section:
10
6
I
a, y
= 143.1 mm
4
10
6
I
a, z
= 48.5 mm
4
Design action effects, ultimate limit state
For the most critical load arrangement, global analysis gives these values:
N
Ed
= 1800 kN, of which N
G, Ed
= 1200 kN
M
y, Ed, top
= 380 kN m; M
z, Ed, top
= 0
The bending moments at the lower end of the column and the lateral loading are zero.
Later, the effect of adding M
z, Ed, top
= 50 kN m is determined.
Properties of the column length
From clause 6.7.3.2(1),
N
pl, Rd
= A
a
f
yd
+ 0.85A
c
f
cd
= 11.4 355 + 148.6 14.2 = 4047 + 2109 = 6156 kN
From clause 6.7.3.3(1), the steel contribution ratio is
= 4047/6156 = 0.657
which is within the limits of clause 6.7.1(4).
For to clause 6.7.3.3(2), with
C
= 1.5,
N
pl, Rk
= 4047 + 1.5 2109 = 7210 kN
Creep coefficient
E
c
= E
cm
/[1 + (N
G, Ed
/N
Ed
)
t
] (6.41)
The creep coefficient
t
is (t, t
0
) to clause 5.4.2.2. The time t
0
is taken as 30 days, and t is
taken as infinity, as creep reduces the stiffness, and hence the stability, of a column.
From clause 3.1.4(5) of EN 1992-1-1, the perimeter exposed to drying is
u = 2(b
c
+ h
c
) = 1600 mm
so
h
0
= 2A
c
/u = 297 200/1600 = 186 mm
Assuming inside conditions and the use of normal cement, the graphs in Fig. 3.1 of
EN 1992-1-1 give
(, 30) = 2.7 =
t
The assumed age at first loading has little influence on the result if it exceeds about
20 days. If, however, significant load were applied at age 10 days,
t
would be increased to
about 3.3.
From equation (6.41),
E
c, eff
= 31/[1 + 2.7(1200/1800)] = 11.1 kN/mm
2
Elastic critical load, with characteristic stiffness
The minor axis is the more critical, so is needed. From clause 6.7.3.3(3),
126
12 May 2004 11:52:58
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
115
(EI)
eff
= E
a
I
a
+ K
e
E
c, eff
I
c
(6.40)
For the concrete,
10
6
I
c, z
= 0.4
2
400
2
/12 48.5 = 2085 mm
4
10
6
(EI)
eff, z
= 210 48.5 + 0.6 11.1 2085 = 24 070 kN mm
2
In this example, the end conditions for the column are assumed to be such that lateral
restraint is provided, but no elastic rotational restraint, so the effective length is the actual
length, 7.0 m, and
N
cr, z
=
2
(EI)
eff, z
/L
2
= 24 070
2
/49 = 4848 kN
= (N
pl, Rk
/N
cr, z
)
0.5
= (7210/4848)
0.5
= 1.22
Similar calculations for the y-axis give
10
6
I
c, y
= 1990 mm
4
10
6
(EI)
eff, y
= 43 270 kN mm
2
N
cr, y
= 8715 kN = 0.91
The non-dimensional slenderness does not exceed 2.0, so clause 6.7.3.1(1) is satisfied.
Resistance to axial load, z-axis
Fromclause 6.7.3.5(2), buckling curve (c) is applicable. FromFig. 6.4 of EN1993-1-1, for
= 1.22,
z
= 0.43
N
Ed

z
N
pl, Rd
= 0.43 6156 = 2647 kN
This condition is satisfied.
Transverse shear
For M
y, Ed, top
= 380 kN m, the transverse shear is
V
z, Ed
= 380/7 = 54 kN
This is obviously less than 0.5V
pl, a, Rd
, so clause 6.7.3.2(3) does not apply.
Interaction curves
The interaction polygons corresponding to Fig. 6.19 are determined in Appendix C (see
Example C.1), and reproduced in Fig. 6.38. Clause 6.7.3.2(5) permits them to be used as
approximations to the NM interaction curves for the cross-section.
First-order bending moments, y-axis
The distribution of the external bending moment is shown in Fig. 6.39(a). From clause
6.7.3.4(4), the equivalent member imperfection is
e
0, z
= L/200 = 35 mm
The mid-length bending moment due to N
Ed
is
N
Ed
e
0, z
= 1800 0.035 = 63 kN m
Its distribution is shown in Fig. 6.39(b).
z
127
12 May 2004 11:52:58
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
116
To check whether second-order moments can be neglected, a reduced value of N
cr
is
required, to clause 6.7.3.4(3). From equation (6.42),
(EI)
y, eff, II
= 0.9(E
a
I
a
+ 0.5E
c
I
c
) = 0.9 10
6
(210 143.1 + 0.5 11.1 1990)
= 3.70 10
10
kN mm
2
Hence,
N
cr, y, eff
= 37 000
2
/7
2
= 7450 kN
This is less than 10N
Ed
, so second-order effects must be considered.
Second-order bending moments, y-axis
Fromclause 6.7.3.4(5), Table 6.4, for the end moments, r = 0, = 0.66, and fromequation
(6.43),
k
1
= /(1 N
Ed
/N
cr, y, eff
) = 0.66/(1 1800/7450) = 0.87
N
Rd
(MN)
6
5
3
2
0
600 400 200
4
1
1.80
2.482
M
Rd
(kN m)
559
504
1.241
328
6.156
333
A
Minor axis
Major axis
B
C
D
414
534
Fig. 6.38. Interaction polygons for bending about the major and minor axes
M
y, Ed
(kN m)
380
0 7.0 3.5
x (m) x (m)
k
1
= 0.87

(a)
M from initial
bow (kN m)
63
3.5 7.0
(b)
83
k
2
= 1.32
0
First order
0.66 380
0.87 380
Fig. 6.39. Design second-order bending moments for a column of length 7.0 m
128
12 May 2004 11:52:58
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
117
This is not increased to 1.0, as it will be combined with the effect of imperfections, so the
major-axis bending moments are as shown in Fig. 6.39(a).
For the bending moment from the member imperfection, = 1.0. From equation
(6.43),
k
2
= 1/(1 1800/7450) = 1.32
thus increasing N
ed
e
0, z
to 83 kN m (Fig. 6.39(b)).
The total mid-length bending moment is
331 + 83 = 414 kN m
This exceeds the greater end moment, 380 kN m, and so governs.
The point (N
Ed
, M
y, Ed, max
) is (1800, 414) on Fig. 6.38. From the values shown in the
figure,
M
y, Rd
= 504 + (2482 1800) 55/(2482 1241) = 534 kN m
This exceeds M
pl, y, Rd
, so clause 6.7.3.6(2) is relevant.
In this case, it makes no difference whether N
Ed
and M
Ed
are from independent actions
or not, because the point M
y, Rd
lies on line CD in Fig. 6.38, not on line BD, so the
additional verification to clause 6.7.1(7) would not alter the result.
The ratio
M
y, Ed, max
/
d, y
M
pl, y, Rd
= 414/534 = 0.78
This is below 0.9, so clause 6.7.3.6(1) is satisfied. The column is strong enough.
Biaxial bending
The effect of adding a minor-axis bending moment M
z, Ed, top
= 50 kN m is as follows. It is
much smaller than M
y, Ed
, so major-axis failure is assumed. From clause 6.7.3.7(1), there is
assumed to be no bow imperfection in the xy plane; but second-order effects have to be
considered.
(EI)
z, eff, II
= 0.9 10
6
(210 48.5 + 0.5 11.1 2085) = 1.96 10
10
kN mm
2
Hence,
N
cr, z, eff
= 19 600
2
/49 = 3948 kN m
As before, = 0.66, and from equation (6.43),
k
1
= 0.66/(1 1800/3948) = 1.21
so
M
z, Ed, max
= 1.21 50 = 60.5 kN m
From Fig. 6.38, for N
Ed
= 1800 kN,
M
z, Rd
=
d, z
M
pl, z, Rd
= 330 kN m
M
y, Ed, max
/0.9M
y, Rd
+ M
z, Ed, max
/0.9M
z, Rd
= 414/(0.9 534) + 60.5/(0.9 330)
= 0.861 + 0.204
= 1.07
This exceeds 1.0, so the check is thus not satisfied, and the column cannot resist the
additional bending moment.
129
12 May 2004 11:52:58
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
118
Example 6.11: longitudinal shear outside areas of load introduction, for a
composite column
All of the data for this example are given in Example 6.10 and Fig. 6.37, and are not
repeated here. The design transverse shear was found to be
V
z, Ed
= 54 kN
The maximum shear that can be resisted without provision of shear connection is now
calculated. The critical cross-section is BB in Fig. 6.40.
Clause 6.7.4.3(2) permits the use of elastic analysis. Creep and cracking should be
considered. Creep reduces the shear stress on plane BB, so the modular ratio n
0
= 6.77 is
used. Uncracked section properties are used, and cracking is considered later.
The cover c
z
= 70 mm, so from equation (6.49) in clause 6.7.4.3(4),
c
= 1.60. From
Table 6.6,
Rd
= 0.30 1.6 = 0.48 N/mm
2
From Example 6.10,
10
6
I
a, y
= 143.1 mm
4
10
6
I
c, y
= 1990 mm
4
so for the uncracked section in concrete units,
10
6
I
y
= 1990 + 143.1 6.77 = 2959 mm
4
The excluded area
A
ex
= 400 70 = 28 000 mm
2
and its centre of area is
z = 200 35 = 165 mm
from G in Fig. 6.40. Hence,
V
z, Rd
=
Rd
I
y
b
c
/(A
ex
z) = 0.48 2959 400/(28 000 0.165) = 123 kN
This exceeds V
z, Ed
, so no shear connection is needed. The margin is so great that there is
no need to consider the effect of cracking.
Column section used as a beam
The shear strengths given in Table 6.6 are unlikely to be high enough to permit omission of
shear connection from a column member with significant transverse loading. As an
example, it is assumed that the column section of Fig. 6.40 is used as a simply-supported
beam of span 8.0 m, with uniformly distributed loading.
B

c
z
= 70
400
G
400
B
Fig. 6.40 Longitudinal shear on plane BB in a column cross-section (dimensions in mm)
130
12 May 2004 11:52:59
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
119
Using elastic analysis of the cracked cross-section with n
0
= 6.77 as before, but no axial
load, the results are
10
6
I
y, cracked
= 1441 mm
4
V
z, Rd
= 79 kN
This corresponds to a design ultimate load of 19.7 kN/m, which is only 4.4 times the
unfactored weight of the member.
6.8. Fatigue
6.8.1. General
Although fatigue verification is mainly needed for bridges, these general provisions find
application in some buildings; for example, where travelling cranes or fork-lift trucks are
used. They refer extensively to EN 1993-1-9,
89
Fatigue Strength, which is also general. The
Eurocode methods for fatigue are quite complex. There are supplementary provisions in
EN 1994-2, Composite Bridges.
The only complete set of provisions on fatigue in EN1994-1-1 is for stud shear connectors.
Fatigue in reinforcement, concrete, and structural steel is covered mainly by cross-reference
to EN 1992 and EN 1993. Commentary will be found in the guides to those codes.
6,7
Comments here are limited to design for a single cyclic loading: a defined number of
cycles, N
E
, of a loading event for which can be calculated, at a given point, either:
a single range of stress,
E
(N
E
) or
E
(N
E
), or
several stress ranges, that can be represented as N* cycles of a single damage equivalent
stress range (e.g.
E, equ
(N*)) by using the PalmgrenMiner rule for summation of
fatigue damage.
The term equivalent constant-amplitude stress range, defined in clause 1.2.2.11 of
EN 1993-1-9, has the same meaning as damage equivalent stress range, used here and in
clause 6.8.5 of EN 1992-1-1.
Damage equivalent factors (typically , as used in bridge design) are not considered here.
Reference may be made to guides in this series to Part 2 of Eurocodes 2, 3 and 4 (e.g.
Designers Guide to EN 1994
90
).
Clause 6.8.1(3)
Fatigue damage is related mainly to the number and amplitude of the stress ranges. The
peak of the stress range has a secondary influence that can be, and usually is, ignored in
practice for peak stresses below about 60% of the characteristic strength. Ultimate loads are
higher than peak fatigue loads, and the use of partial safety factors for ultimate-load design
normally ensures that peak fatigue stresses are below this limit. This may not be so for
buildings with partial shear connection, so clause 6.8.1(3) limits the force per stud to 0.75P
Rd
,
or 0.6P
Rk
for
V
= 1.25.
Clause 6.8.1(4) Clause 6.8.1(4) gives guidance on the types of building where fatigue assessment may
be required. By reference to EN 1993-1-1, these include buildings with members subject
to wind-induced or crowd-induced oscillations. Repeated stress cycles from vibrating
machinery are also listed, but these should in practice be kept out of the structure by
appropriate mountings.
6.8.2. Partial factors for fatigue assessment
Clause 6.8.2(1)
Resistance factors
Mf
may be given in National Annexes, so only the recommended values
can be discussed here. For fatigue strength of concrete and reinforcement, clause 6.8.2(1)
refers to EN 1992-1-1, which recommends the partial factors 1.5 and 1.15, respectively. For
structural steel, EN 1993-1-9 recommends values ranging from 1.0 to 1.35, depending on the
design concept and consequence of failure. These apply, as appropriate, for a fatigue failure
of a steel flange caused by a stud weld.
131
12 May 2004 11:52:59
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Fatigue failure of a stud shear connector, not involving the flange, is covered by
EN 1994-1-1. The value recommended in a note to clause 2.4.1.2(7)P,
Mf, s
= 1.0, corresponds
to the value in EN 1993-1-9 for damage tolerant design concept with low consequence of
failure. From clause 3(2) of EN 1993-1-9, the use of the damage tolerant method should be
satisfactory, provided that a prescribed inspection and maintenance regime for detecting
and correcting fatigue damage is implemented.... A note to this clause states that the
damage tolerant method may be applied where in the event of fatigue damage occurring a
load redistribution between components of structural elements can occur.
The second condition applies to stud connectors, but the first does not, for lack of access
prevents detection of small cracks by any simple method of inspection. The recommendation
of EN 1994-1-1 is based on other considerations, as follows.
Fatigue failure results froma complex interaction between steel and concrete, commencing
with powdering of the highly stressed concrete adjacent to the weld collar. This displaces
upwards the line of action of the shear force, increasing the bending and shear in the shank just
above the weld collar, and probably also altering the tension. Initial fatigue cracking further
alters the relative stiffnesses and the local stresses. Research has found that the exponent that
relates the cumulative damage to the stress range may be higher than the value, five, for other
welds in shear. The value chosen for EN 1994-1-1, eight, is controversial, as discussed later.
As may be expected from the involvement of a tiny volume of concrete, tests show a wide
scatter in fatigue lives, which is allowed for in the design resistances. Studs are provided in
large numbers, and are well able to redistribute shear between themselves.
The strongest reason for not recommending a value more conservative than 1.0 comes
from experience with bridges, where stud connectors have been used for almost 50 years.
Whenever occasion has arisen in print or at a conference, the first author has stated that
there is no known instance of fatigue failure of a stud in a bridge, other than a few clearly
attributable to errors in design. This has not been challenged. Research has identified, but
not yet quantified, many reasons for this remarkable experience.
91,92
Most of them (e.g. slip,
shear lag, permanent set, partial interaction, adventitious connection from bolt heads, and
friction) lead to predicted stress ranges on studs lower than those assumed in design. With an
eighth-power law, a 10% reduction in stress range more than doubles the fatigue life.
6.8.3. Fatigue strength
Clause 6.8.3(3) The format of clause 6.8.3(3), as in EN 1993-1-9, uses a reference value of range of shear
stress at 2 million cycles,
C
= 90 N/mm
2
. It defines the slope m of the line through this
point on the loglog plot of range of stress
R
against number of constant-range stress
cycles to failure, N
R
(Fig. 6.25).
It is a complex matter to deduce a value for m from the mass of test data, which are often
inconsistent. Many types of test specimen have been used, and the resulting scatter of results
must be disentangled from that due to inherent variability. Values for m recommended in
the literature range from 5 to 12, mostly based on linear-regression analyses. The method of
regression used (x on y, or y on x) can alter the value found by up to three.
91
The value eight, which was also used in BS 5400: Part 10, may be too high. In design for a
loading spectrum, its practical effect is that the cumulative damage is governed by the
highest-range components of the spectrum (e.g. by the small number of maximum-weight
lifts made by a crane). Alower value, such as five, would give more weight to the much higher
number of average-range components.
While fatigue design methods for stud connectors continue to be conservative (for bridges
and probably for buildings too) the precise value for m is of academic interest. Any future
proposals for more accurate methods for prediction of stress ranges should be associated
with re-examination of the value for m.
6.8.4. Internal forces and fatigue loadings
The object of a calculation is usually to find the range or ranges of stress in a given material at
a chosen cross-section, caused by a defined event; for example, the passage of a vehicle along
120
132
12 May 2004 11:52:59
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Clause 6.8.4(1)
a beam. Loading other than the vehicle influences the extent of cracking in the concrete, and,
hence, the stiffnesses of members. Cracking depends mainly on the heaviest previous
loading, and so tends to increase with time. Clause 6.8.4(1) refers to a relevant clause in
Eurocode 2. This defines the non-cyclic loading assumed to co-exist with the design value of
the cyclic load, Q
fat
: it is the frequent combination, represented by
where the Qs are non-cyclic variable actions. Often, only one G and one Q are relevant, and
there is no prestress. The design combination is then
G
k
+
1
Q
k
+ Q
fat
(D6.38)
Clause 6.8.4(2) Clause 6.8.4(2) defines symbols that are used for bending moments in clause 6.8.5.4. The
sign convention is evident from Fig. 6.26, which shows that M
Ed, max, f
is the bending moment
that causes the greatest tension in the slab, and is positive. Clause 6.8.4(2) also refers to
internal forces, but does not give symbols. Analogous use of calculated tensile forces in a
concrete slab (e.g. N
Ed, max, f
, etc.) may sometimes be necessary.
6.8.5. Stresses
Clause 6.8.5.1(1) Clause 6.8.5.1(1) refers to a list of action effects in clause 7.2.1(1)P to be taken into account
where relevant. They are all relevant, in theory, to the extent of cracking. However, this can
usually be represented by the same simplified model, chosen from clause 5.4.2.3, that is used
for other global analyses. They also influence the maximum value of the fatigue stress range,
which is limited for each material (e.g. the limit for shear connectors in clause 6.8.1(3)). It is
unusual for any of these limits to be reached in design for a building; but if there are highly
stressed cross-sections where most of the variable action is cyclic, the maximumvalue should
be checked.
The provisions for fatigue are based on the assumption that the stress range caused by a
given fluctuation of loading, such as the passage of a vehicle of known weight, remains
approximately constant after an initial shakedown period. Shakedown here includes the
changes due to cracking, shrinkage and creep of concrete that occur mainly within the first
year or two.
Most fatigue verifications are for load cycles with more than 10
4
repetitions. For this
number in a 50 year life, the mean cycle time is less than 2 days. Thus, load fluctuations slow
enough to cause creep (e.g. from the use of a tank for storing fuel oil) are unlikely to be
numerous enough to cause fatigue damage. This may not be so for industrial processes with
dynamic effects, such as forging, or for the charging floor for a blast furnace, but other
uncertainties are likely to outweigh those from creep.
Clause 6.8.5.1(2)P
The short-term modular ratio should therefore be used when finding stress ranges from
the cyclic action Q
fat
. Where a peak stress is being checked, creep from permanent loading
should be allowed for, if it increases the relevant stress. This would apply, for example, for
most verifications for the structural steel in a composite beam. The effect of tension
stiffening (clauses 6.8.5.1(2)P and 6.8.5.1(3)), is illustrated in Example 6.12, below.
Clause 6.8.5.1(3) The designer chooses a location where fatigue is most likely to govern. For a vehicle
travelling along a continuous beam, the critical section for shear connection may be near
mid-span; for reinforcement, it is near an internal support.
The extent of a continuous structure that needs to be analysed depends on what is being
checked. For a fatigue check on reinforcement it would be necessary to include at least two
spans of the beam, and perhaps the two adjacent column lengths; but ranges of vertical shear,
as the vehicle passes, are barely influenced by the rest of the structure, so for the shear
connection it may be possible to consider the beam in isolation.
For analysis, the linear-elastic method of Section 5 is used, from clause 6.8.4(1).
Redistribution of moments is not permitted, clause 5.4.4(1). Calculation of range of stress, or
of shear flow, fromthe action effects is based entirely on elastic theory, following clause 7.2.1.
121
k, 1, 1 k, 1 2, k,
1 1
j i i
j i
G P Q Q
>
+ + +

133
12 May 2004 11:53:00
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
Clause 6.8.5.1(4)
For stresses in structural steel, the effects of tension stiffening may be included or
neglected (clause 6.8.5.1(4)).
Concrete
Clause 6.8.5.2(1) For concrete, clause 6.8.5.2(1) refers to clause 6.8 of EN 1992-1-1, which provides (in clause
6.8.7) for concrete subjected to a damage equivalent stress range. For a building, fatigue of
concrete is unlikely to influence design, so this clause is not discussed here.
Structural steel
Clause 6.8.5.3(1)
Clause 6.8.5.3(2)
Clause 6.8.5.3(1) repeats, in effect, the concession in clause 6.8.5.1(4). Where the words or
only M
Ed, min, f
in clause 6.8.5.3(2) apply, M
Ed, max, f
causes tension in the slab. The use of the
uncracked section for M
Ed, max, f
could then under-estimate the stress ranges in steel flanges.
Reinforcement
For reinforcement, clause 6.8.3(2) refers to EN 1992-1-1, where clause 6.8.4 gives the
verification procedure. Its recommended value N* for straight bars is 10
6
. This should not be
confused with the corresponding value for structural steel in EN 1993-1-9, 2 10
6
, denoted
N
C
, which is used also for shear connectors (clause 6.8.6.2(1)).
Using the values recommended in EN 1992-1-1, its equation (6.71) for verification of
reinforcement becomes
E, equ
(N*)
Rsk
(N*)/1.15 (D6.39)
with
Rsk
= 162.5 N/mm
2
for N* = 10
6
, from Table 6.4N.
Where a range
E
(N
E
) has been determined, the resistance
Rsk
(N
E
) can be found from
the SN curve for reinforcement, and the verification is
E
(N
E
)
Rsk
(N
E
)/1.15 (D6.40)
Clause 6.8.5.4(1) Clause 6.8.5.4(1) permits the use of the approximation to the effects of tension stiffening
that is used for other limit states. It consists of adding to the maximum tensile stress in the
fully cracked section,
s, 0
, an amount
s
that is independent of
s, 0
and of the limit state.
Clause 6.8.5.4(2)
Clause 6.8.5.4(3)
Clauses 6.8.5.4(2) and 6.8.5.4(3) give simplified rules for calculating stresses, with
reference to Fig. 6.26, which is discussed using Fig. 6.41. This has the same axes, and also
shows a minimum bending moment that causes compression in the slab. A calculated value
for the stress
s
in reinforcement, that assumes concrete to be effective, would lie on line
AOD. On initial cracking, the stress
s
jumps from B to point E. Lines OBE are not shown in
Fig. 6.26 because clause 7.2.1(5)P requires the tensile strength of concrete to be neglected in
calculations for
s
. This gives line OE. For moments exceeding M
cr
, the stress
s
follows route
EFG on first loading. Calculation of
s
using section property I
2
gives line OC. At bending
moment M
Ed, max, f
the stress
s, 0
thus found is increased by
s
, from equation (7.5), as shown
by line HJ.
Tension stiffening tends to diminish with repeated loading,
93
so clause 6.8.5.4 defines the
unloading route from point J as JOA, on which the stress
s, min, f
lies. Points K and L give
two examples, for M
Ed, min, f
causing tension and compression, respectively, in the slab. The
fatigue stress ranges
s, f
for these two cases are shown.
Shear connection
Clause 6.8.5.5(1)P
Clause 6.8.5.5(2)
The interpretation of clause 6.8.5.5(1)P is complex when tension stiffening is allowed for.
Spacing of shear connectors near internal supports is unlikely to be governed by fatigue, so it
is simplest to use uncracked section properties when calculating the range of shear flowfrom
the range of vertical shear (clause 6.8.5.5(2)). These points are illustrated in Example 6.12.
6.8.6. Stress ranges
Clause 6.8.6.1 Clause 6.8.6.1 is more relevant to the complex cyclic loadings that occur in bridges than to
buildings, and relates to the provisions of EN 1992 and EN 1993. Where a spectrum
122
134
12 May 2004 11:53:00
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
of loading is specified, the maximum and minimum stresses, as discussed above for
reinforcement, are modified by the damage equivalent factor , a property of the spectrum
and the exponent m (clause 6.8.3(3)). Comment and guidance will be found in the relevant
guides in this series.
Clause 6.8.6.1(3) The need to combine global and local fatigue loading events (clause 6.8.6.1(3)), rarely
occurs in buildings, and is outside of the scope of this guide.
Clause 6.8.6.2
Where the design cyclic loading consists of a single load cycle repeated N
E
times, the
damage equivalent factor
v
used in clause 6.8.6.2 on shear connection can be found using the
PalmgrenMiner rule, as follows.
Let the load cycle cause a shear stress range in a stud connector, for which m = 8.
Then,
(
E
)
8
N
E
= (
E, 2
)
8
N
C
where N
C
= 2 10
6
cycles. Hence,
E, 2
/
E
=
v
= (N
E
/N
C
)
1/8
(D6.41)
6.8.7. Fatigue assessment based on nominal stress ranges
Clause 6.8.7.1 Comment on the methods referred to from clause 6.8.7.1 will be found in other guides in this
series.
Clause 6.8.7.2(1) For shear connectors, clause 6.8.7.2(1) introduces the partial factors. The recommended
value of
Mf, s
is 1.0 (clause 6.8.2(1)). For
Ff
, EN 1990 refers to the other Eurocodes. The
recommended value in EN 1992-1-1, clause 6.8.4(1), is 1.0. No value has been found in
EN 1993-1-1 or EN 1993-1-9. Clause 9.3(1) of EN 1993-2 recommends 1.0 for bridges.
Clause 6.8.7.2(2) Clause 6.8.7.2(2) covers interaction between the fatigue failures of a stud and of the steel
flange to which it is welded, when the flange is in tension. The first of expressions (6.57) is the
verification for the flange, from clause 8(1) of EN 1993-1-9, and the second is for the stud,
copied from equation (6.55). The linear interaction condition is given in expression (6.56).
It is necessary to calculate the longitudinal stress range in the steel flange that coexists
with the stress range for the connectors. The load cycle that gives the maximum value of
E, 2
in the flange will not, in general, be that which gives the maximum value of
E, 2
in a
123
s
s
(tension)
D
C
G
0
L
Ds
s
s
s, 0
s
s, max, f
s
s, min, f
B M
J
A
E
M
Ed, min, f
M
Ed, min, f
M
Ed, max, f
F
Ds
s, f
K
H
M
cr
Fig. 6.41. Stress ranges in reinforcement in cracked regions
135
12 May 2004 11:53:00
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
124
shear connector, because the first is caused by flexure and the second by shear. Also, both
E, 2
and
E, 2
may be influenced by whether the concrete is cracked, or not.
It thus appears that expression (6.56) may have to be checked four times. In practice, it is
best to check first the conditions in expression (6.57). It should be obvious, for these, whether
the cracked or the uncracked model is the more adverse. Usually, one or both of the
left-hand sides is so far below 1.0 that no check to expression (6.56) is needed.
Example 6.12: fatigue in reinforcement and shear connection
It is assumed that the imposed floor load of 7.0 kN/m
2
for the two-span beam in Examples
6.7 and 7.1 is partly replaced by a cyclic load. The resistance to fatigue of the
reinforcement at the internal support, point B in Figs 6.23 and 6.28, and of the shear
connection near the cyclic load are checked. All other data are as before.
Loading and global analysis
The cyclic load is a four-wheeled vehicle with two characteristic axle loads of 35 kN each.
It travels at right angles to beam ABC, on a fixed path that is 2.0 m wide and free from
other variable loads. The axle spacing exceeds the beam spacing of 2.5 m, so each
passage can be represented by two cycles of point load, 0350 kN, applied at point D
in Fig. 6.42(a). For a 25 year design life, 20 passages per hour for 5000 h/year gives
N
Ed
= 2.5 10
6
cycles of each point load. The partial factor
Ff
is taken as 1.0.
In comparison with Example 6.7, the reduction in static characteristic imposed load is
7 2.5 2 = 35 kN, the same as the additional axle load, so previous global analyses for
the characteristic combination can be used. These led to the bending moments M
Ek
at
support B given in the four rows of Table 7.2. Those in rows 2 and 4 are unchanged, and
M
Ek, B
= 263 kN m for loading q
k
. Analysis for the load Q
fat
alone, with 15% of each span
cracked, gave the results in Fig. 6.42(b), with 31 kN m at support B.
The frequent combination of non-cyclic imposed load is specified, for which
1
= 0.7.
From Table 6.2, q
k
= 17.5 kN/m. Therefore,
1
q
k
= 0.7 17.5 = 12.25 kN/m, acting on
span AB and on 10 m only of span BC, giving
M
Ek, B
= 0.7 (263 31) = 162 kN m
Table 6.5 gives
M
Ed, min, f
= 18 + 162 + 120 = 300 kN m
12 4 1 1 6
Q
fat
= 35 kN
B A
0.7q
k
D C
31
2.6 12
2.6 + 23
84
35
(b)
(a)
(c)
10.0 10.0
120
20.0
Fig. 6.42. Fatigue checks for a two-span beam. (a) Variable static and cyclic loads. (b) Design action
effects, cyclic load. (c) Design action effects, shrinkage
136
12 May 2004 11:53:01
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
125
From Fig. 6.42(b),
M
Ed, max, f
= 300 + 31 = 331 kN m
Verification for reinforcement at cross-section B
From equation (D7.5), the allowance for tension stiffening is
s
= 52 N/mm
2
. From
s, 0
given in Table 6.5,
s, max, f
= 201 + 52 = 253 N/mm
2
From Fig. 6.41,
s, min, f
= 253 300/331 = 229 N/mm
2
so
s, f
= 253 229 = 24 N/mm
2
From clause 6.8.4 of EN 1992-1-1, for reinforcement,
m = 9 for N
E
> 10
6
N
*
= 10
6
Rsk
(N*) = 162.5 N/mm
2
By analogy with equation (D6.41), the damage equivalent stress range for N
Ed
= 5 10
6
cycles of stress range 24 N/mm
2
is given by
24
9
5 10
6
= (
E, equ
)
9
10
6
whence
E, equ
= 29 N/mm
2
Using equation (D6.40),
29 162.5/1.15
so the reinforcement is verified.
If the axle loads had been unequal, say 35 and 30 kN, the stress range for the lighter axle
would be a little higher than 24 30/35 = 20.6 N/mm
2
, because its line OJ in Fig. 6.41
would be steeper. Assuming this stress range to be 21 N/mm
2
, the cumulative damage
check, for
Mf
= 1.15, would be
2.5 10
6
(24
9
+ 21
9
) (162.5/1.15)
9
10
6
which is
8.6 10
18
2.25 10
25
Verification for shear connection near point D
Vertical shear is higher on the left of point D in Fig. 6.42, than on the right. From Fig.
6.42(b), V
Ed, f
= 23 kN for each axle load.
Table 6.5. Stresses in longitudinal reinforcement at support B
Action n
Load
(kN/m)
M
Ek
(kN m)
10
6
W
s, cr
(mm
3
)
s, 0
(N/mm
2
)
Permanent, composite 20.2 1.2 18 1.65 11
Variable, static (
1
= 0.7) 20.2 12.25 162 1.65 98
Shrinkage 28.7 120 1.65 73
Cyclic load 20.2 31 1.65 19
Totals 331 201
137
12 May 2004 11:53:01
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:11
126
The maximum vertical shear at this point, including 10.0 kN from the secondary effect
of shrinkage, Fig. 6.42(c), is 59.5 kN, Table 6.6. The shear forces V
Ed
are found from the
loads and values M
Ek
in Table 6.5. The resulting maximum longitudinal shear flow, for the
uncracked unreinforced composite section, is 110 kN/m, of which the cyclic part is
43.2 kN/m.
Clause 6.8.1(3) limits the shear per connector under the characteristic combination to
0.75P
Rd
. For this combination, the shear flowfromthe non-cyclic variable action increases
by 19 from 44.7 to 44.7/0.7 = 63.9 kN/m, so the new total is 110 + 19 = 129 kN/m. The
shear connection (see Fig. 6.30), is 5 studs/m, with P
Rd
= 51.2 kN/stud. Hence,
P
Ek
/P
Rd
= 129/(5 51.2) = 0.50
which is below the limit 0.75 in clause 6.8.1(3).
The range of shear stress is
E
= 43 200/(5 9.5
2
) = 30.5 N/mm
2
The concrete is in density class 1.8. From clause 6.8.3(4),
c
= 90 (1.8/2.2)
2
= 60 N/mm
2
From equation (D6.41),
E, 2
= 30.5[5 10
6
/(2 10
6
)]
1/8
= 34.2 N/mm
2
As
Mf, s
= 1.0,
c, d
= 60 N/mm
2
so the shear connection is verified.
Table 6.6. Fatigue of shear connectors near cross-section D
Action
10
6
I
y
(mm
4
)
10
3
A
c
/n
(mm
2
) z (mm)
V
Ed
(kN)
V
Ed
A
c
z/nI
y
(kN/m)
Permanent, composite 828 9.90 157 2.7 5.0
Variable, static (
1
= 0.7) 828 9.90 157 23.8 44.7
Shrinkage 741 6.97 185 10.0 17.4
Cyclic load 828 9.90 157 23.0 43.2
Totals 59.5 110
138
12 May 2004 11:53:01
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:14
CHAPTER 7
Serviceability limit states
General Clause 7.1
Stresses Clause 7.2
Deformations in buildings Clause 7.3
Cracking of concrete Clause 7.4
7.1. General
Section 7 of EN 1994-1-1 is limited to provisions on serviceability that are specific to
composite structures and are not in Sections 1, 2, 4, 5 (for global analysis) or 9 (for composite
slabs), or in Eurocodes 1990, 1991, 1992 or 1993. Some of these other, more general
provisions are briefly referred to here. Further comments on them are in other chapters of
this book, or in other handbooks in this series.
The initial design of a structure is usually based on the requirements for ultimate limit
states, which are specific and leave little to the judgement of the designer. Serviceability is
then checked. The consequences of unserviceability are less serious than those of reaching
an ultimate limit state, and its occurrence is less easily defined. For example, a beam with an
imposed-load deflection of span/300 may be acceptable in some situations, but in others the
client may prefer to spend more on a stiffer beam.
The drafting of the serviceability provisions of the EN Eurocodes is intended to give
designers and clients greater freedomto take account of factors such as the intended use of a
building and the nature of its finishes.
The content of Section 7 was also influenced by the need to minimize calculations. Results
already obtained for ultimate limit states are scaled or re-used wherever possible. Experienced
designers know that many structural elements satisfy serviceability criteria by wide margins.
For these, design checks must be simple, and it does not matter if they are conservative. For
other elements, a longer but more accurate calculation may be justified. Some application
rules therefore include alternative methods.
Clause 7.1(1)P
Clause 7.1(2)
Clauses 7.1(1)P and 7.1(2) refer to clause 3.4 of EN 1990. This gives criteria for placing
a limit state within the serviceability group, with reference to deformations (including
vibration), durability, and the functioning of the structure.
Serviceability verification and criteria
The requirement for a serviceability verification is given in clause 6.5.1(1)P of EN 1990 as
E
d
C
d
139
12 May 2004 11:53:01
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:14
where E
d
is the design value of the effects of the specified actions and the relevant
combination, and C
d
is the limiting design value of the relevant criterion.
From clause 6.5.3 of EN 1990, the relevant combination is normally the characteristic,
frequent, or quasi-permanent combination, for serviceability limit states that are, respectively,
irreversible, reversible or a consequence of long-term effects. The quasi-permanent
combination is also relevant for the appearance of the structure.
For buildings, these combinations are used with the partial safety factor 1.0, from clause
A1.4.1 of EN1990, unless differently specified in another Eurocode. There are no departures
from1.0 in EN1994-1-1. The same provision, with value 1.0, is given for partial safety factors
for properties of materials, in clause 6.5.4(1) of EN 1990.
Clause A1.4.2 of EN1990 refers to serviceability criteria relevant for buildings. These may
be defined in National Annexes, and should be specified for each project and agreed with the
client.
Clause A1.4.4 of EN 1990 says that possible sources of vibration and relevant aspects of
vibration behaviour should be considered for each project and agreed with the client and/or
the relevant authority. Further guidance may be found in the relevant Eurocode Part 2
(bridges) and in specialized literature.
Comments on limits to crack width are given under clause 7.4.
No serviceability limit state of excessive slip of shear connection has been defined, but
the effect of slip is recognized in clause 7.3.1(4) on deflection of beams. Generally, it is
assumed that design of shear connection for ultimate limit states ensures satisfactory
performance in service, but composite slabs can be an exception. Relevant rules are given in
clause 9.8.2.
No serviceability criteria are specified for composite columns, so, from here on, this
chapter is referring to composite beams or, in some places, to composite frames.
7.2. Stresses
Clause 7.2.1
Clause 7.2.2
Excessive stress is not itself a serviceability limit state, though stress calculations to clause
7.2.1 are required for some verifications for deformation and cracking. For most buildings,
no checks on stresses are required, clause 7.2.2. No stress limits for buildings are given in
the Eurocodes for concrete and steel structures, other than warnings in clause 7.2 of
EN1992-1-1, with recommended stress limits in notes. The bridge parts of these Eurocodes
include stress limits, which may be applicable for buildings that have prestressing or fatigue
loading.
7.3. Deformations in buildings
7.3.1. Deflections
Global analysis
Deflections are influenced by the method of construction, and may govern design, especially
where beams designed as simpl- supported are built unpropped. For propped construction,
props to beams should not be removed until all of the concrete that would then be stressed
has reached a strength equivalent to grade C20/25, from clause 6.6.5.2(4). Then, elastic
global analysis to Section 5 is sufficient (clause 7.3.1(2)).
Clause 7.3.1(1)
Where unpropped construction is used and beams are not designed as simply supported,
the analysis may be more complex than is revealed by the reference to EN 1993 in clause
7.3.1(1). In a continuous beamor a frame, the deflection of a beamdepends on how much of
the structure is already composite when the slab for each span is cast. A simple and usually
conservative method is to assume that the whole of the steel frame is erected first. Then, all
of the concrete for the composite members is cast at once, its whole weight being carried by
the steelwork; but more realistic multi-stage analyses may be needed for a high-rise structure
and for long-span beams.
128
140
12 May 2004 11:53:02
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:14
Where falsework or re-usable formwork is supported from the steel beam, it will be
removed after the member becomes composite. The small locked-in stresses that result can
usually be ignored in buildings, but not always in bridges.
Where first-order elastic global analysis was used for ultimate limit states (ULS), it may
be possible to obtain some of the results needed for serviceability limit states (SLS) by
simple scaling by the ratio of the relevant loads. This ratio will depend on the method of
construction, and also on which of the three serviceability load combinations is being used
for the limit state considered.
As an example, suppose that for an unbraced frame at ULS,
cr
= 8, so that second-order
global analysis was used, from clause 5.2.1(3). If most of the load on columns is from
suspended floors, and these loads for SLS are 60% of those for ULS, the elastic critical load
will be little altered, so for SLS,
cr
8/0.6 = 13. This exceeds 10, so first-order analysis can
be used.
Redistribution of moments is permitted for most framed structures at SLS by clause
5.4.4(1), but the details in paragraphs (4) to (7) apply only to ULS. The relevant provisions in
Section 7 are in clauses 7.3.1(6) and 7.3.1(7), discussed below.
Limits to deflection of beams
The specification of a deflection limit for a long-span beam needs care, especially where
construction is unpropped and/or the steel beam is pre-cambered. Reference should be
made to the three components of deflection defined in clause A1.4.3 of EN 1990.
Clause 7.3.1(3)
Depending on circumstances, it may be necessary to set limits to any one of them, or to
more than one, related to a defined load level. Prediction of long-term values should take
account of creep of concrete, based on the quasi-permanent combination, and may need to
allow for shrinkage. Where precast floor units are used, it must be decided whether they
should be cambered to compensate for creep. Clause 7.3.1(3) relates to the use of false
ceilings, which conceal the sagging of a beam due to dead loading.
Longitudinal slip
Clause 7.3.1(4) Clause 7.3.1(4) refers to the additional deflection caused by slip at the interface between
steel and concrete. Its three conditions all apply. Condition (b) relates to the minimumvalue
of the degree of shear connection, , given as 0.4 in clause 6.6.1.2(1), and gives a higher limit,
0.5.
For use where the design is such that 0.4 < 0.5, ENV 1994-1-1
49
gave the following
equation for the additional deflection due to partial interaction:
=
c
+ (
a

c
)(1 ) (D7.1)
where = 0.5 for propped construction and 0.3 for unpropped construction,
a
is the
deflection of the steel beamacting alone, and
c
is the deflection for the composite beamwith
complete interaction; both
a
and
c
are calculated for the design loading for the composite
member. The method comes from a summary of pre-1975 research on this subject,
94
which
also gives results of relevant tests and parametric studies. Other methods are also available.
95
Cracking in global analysis
Clause 7.3.1(5)
Apart from the different loading, global analysis for serviceability differs little from that for
an ultimate limit state. Clause 5.4.1.1(2) requires appropriate corrections for cracking of
concrete, and clause 7.3.1(5) says that clause 5.4.2.3 applies. Clause 5.4.2.3(2) permits the use
of the same distribution of beam stiffnesses at SLS as for ULS. Clauses 5.4.2.3(3) to
5.4.2.3(5) also apply, including the reference in clause 5.4.2.3(4) to a method given in Section
6 for the effect of cracking on the stiffness of composite columns.
Clause 7.3.1(6)
In the absence of cracking, continuous beams in buildings can often be assumed to be of
uniformsection within each span, which simplifies global analysis. Cracking reduces bending
moments at internal supports to an extent that can be estimated by the method of clause
129
CHAPTER 7. SERVICEABILITY LIMIT STATES
141
12 May 2004 11:53:02
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:14
7.3.1(6), based on Stark and van Hove.
95
The maximum deflection of a given span normally
occurs when no imposed load acts on adjacent spans. The conditions for the use of curve Ain
Fig. 7.1 are then not satisfied, and the method consists simply of reducing all uncracked
moments at internal supports by 40%.
Using the new end moments, M
h1
and M
h2
, say, the maximum deflection can be found
either by elastic theory for the span, of uncracked flexural stiffness E
a
I
1
, or by an approximate
method given in BS 5950-3-1.
31
This consists of multiplying the deflection for the simply-
supported span by the factor
1 0.6(M
h1
+ M
h2
)/M
0
(D7.2)
where M
0
is the maximum sagging moment in the beam when it is simply supported.
Yielding of steel
Clause 7.3.1(7)
In continuous beams built unpropped, with steel beams in Class 1 or 2, it is possible that
serviceability loading may cause yielding at internal supports. This is permitted for beams in
buildings, but it causes additional deflection, which should be allowed for. Clause 7.3.1(7)
provides a method. The bending moments at internal supports are found by elastic analysis,
with allowance for effects of cracking. The two values given in the clause for factors f
2
correspond to different checks. The first is for dead load only: wet concrete on a steel beam.
According to the UKs draft National Annex to EN 1990,
96
the load combination to be
used for the second check depends on the functioning of the structure. It may be the
characteristic, frequent or quasi-permanent combination, with the load additional to that for
the first check acting on the composite beam. For each analysis, appropriate assumptions are
needed for the adjacent spans, on their loading and on the state of construction.
Local buckling
This does not influence stiffnesses for elastic analysis except for Class 4 sections. For these,
clause 5.4.1.1(6) refers to clause 2.2 in EN 1993-1-5, which gives a design rule.
Shrinkage
Clause 7.3.1(8)
In principle, shrinkage effects appear in all load combinations. For SLS, clause 5.4.2.2(7)
refers to Section 7, where clause 7.3.1(8) enables effects of shrinkage on deflections of beams
to be ignored for span/depth ratios up to 20. In more slender beams, shrinkage deflections
are significantly reduced by provision of continuity at supports.
Temperature
Clause 5.4.2.5(2), on neglect of temperature effects, does not apply. For buildings, neither
0
nor
1
is given as zero in clause A1.2.2 of EN 1990 (nor in the UKs draft National Annex to
BS EN 1990
96
), so if temperature effects are relevant at ULS, they should be included in all
SLS combinations except quasi-permanent.
Welded mesh
Clause 5.5.1(6) gives conditions for the inclusion of welded mesh in the effective section,
within the rules for classification of sections.
7.3.2. Vibration
Clause 7.3.2(1)
Limits to vibration in buildings are material-independent, and vibration is in clause A1.4.4
of EN 1990, not in EN 1994. Composite floor systems are lighter and have less inherent
damping than their equivalents in reinforced concrete. During their design, dynamic
behaviour should be checked against the criteria in EN1990 referred to fromclause 7.3.2(1).
These are general, and advise that the lowest natural frequency of vibration of the structure
or member should be kept above a value to be agreed with the client and/or the relevant
authority. No values are given for either limiting frequencies or damping coefficients.
130
142
12 May 2004 11:53:02
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:14
More specific guidance can be found in EN 1991-1-1 and the extensive literature on this
subject.
97,98
These sources refer to several criteria that are likely to be specific to the
individual project, and, with other aspects, should be agreed with the client. Anote to clause
7.2.3 of EN 1993-1-1 says that limits to vibration of floors may be specified in a National
Annex.
7.4. Cracking of concrete
7.4.1. General
In the early 1980s it was found
44,99
that for composite beams in hogging bending, the
long-established British methods for control of crack width were unreliable for initial cracks,
which were wider than predicted. Before this, it had been found for reinforced concrete that
the appropriate theoretical model for cracking caused by restraint of imposed deformation
was different fromthat for cracking caused by applied loading. This has led to the presentation
of design rules for control of cracking as two distinct procedures:
for minimumreinforcement, in clause 7.4.2, for all cross-sections that could be subjected
to significant tension by imposed deformations (e.g. by effects of shrinkage, which cause
higher stresses than in reinforced concrete, because of restraint from the steel beam)
for reinforcement to control cracking due to direct loading (clause 7.4.3).
The rules given in EN 1994-1-1 are based on an extensive and quite complex theory,
supported by testing on hogging regions of composite beams.
99,100
Much of the original
literature is in German, so a detailed account of the theory has recently been published
in English,
101
with comparisons with results of tests on composite beams, additional to
those used originally. The paper includes derivations of the equations given in clause 7.4,
comments on their scope and underlying assumptions, and procedures for estimating the
mean width and spacing of cracks. These are tedious, and so are not in EN 1994-1-1. Its
methods are simple: Tables 7.1 and 7.2 give maximum diameters and spacings of reinforcing
bars for three design crack widths: 0.2, 0.3 and 0.4 mm.
Tables 7.1 and 7.2 are for high-bond bars only. This means ribbed bars with properties as
in clause 3.2.2(2)P of EN1992-1-1. The use of reinforcement other than ribbed is outside the
scope of the Eurocodes.
Clause 7.4.1(1) The references to EN 1992-1-1 in clause 7.4.1(1) give the surface crack-width limits
required for design. Concrete in tension in a composite beam or slab for a building will
usually be in exposure class XC3, for which the recommended limit is 0.3 mm; however, for
spaces with low or very low air humidity, Tables 4.1 and 7.1N of EN 1992-1-1 recommend a
limit of 0.4 mm. The limits are more severe for prestressed members, which are not discussed
further. The severe environment for a floor of a multistorey car park is discussed in Chapter
4. All these limits may be modified in a National Annex.
Clause 7.4.1(2) Clause 7.4.1(2) refers to estimation of crack width, using EN 1992-1-1. This rather long
procedure is rarely needed, and does not take full account of the following differences
between the behaviours of composite beams and reinforced concrete T-beams. The steel
member in a composite beam does not shrink or creep and has much greater flexural
stiffness than the reinforcement in a concrete beam. Also, the steel member is attached
to the concrete flange only by discrete connectors that are not effective until there is
longitudinal slip, whereas in reinforced concrete there is monolithic connection.
Clause 7.4.1(3) Clause 7.4.1(3) refers to the methods developed for composite members, which are easier
to apply than the methods for reinforced concrete members.
Uncontrolled cracking
Clause 7.3.1(4) of EN 1992-1-1 (referred to from clause 7.4.1(1)) permits cracking
of uncontrolled width in some circumstances; for example, beams designed as simply
supported, with a concrete top flange that is continuous over simple beam-to-column
131
143
12 May 2004 11:53:02
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:14
Clause 7.4.1(4)
connections. These are flexible in bending, and rotate about a point that cannot be
predicted, as its position depends on tolerances and methods of erection of the steelwork. It
may then be impossible to predict the widths of cracks. Where the environment is dry and the
concrete surface is concealed by a flexible finish, such as carpeting, crack widths exceeding
0.4 mm may be acceptable. Even so, the minimum reinforcement required (for other
reasons) by EN 1992-1-1 may be inadequate to prevent the fracture of small-diameter bars
near internal supports, or the formation of very wide cracks. Minimum areas greater than
those of EN 1992-1-1 are therefore specified in clause 7.4.1(4) and, for composite slabs, in
clause 9.8.1(2).
The maximum thicknesses of slab that can be reinforced by one layer of standard welded
fabric, according to these rules, are given in Table 7.1. For composite slabs, the relevant
thickness is that above the profiled steel sheeting.
The maximum spacing of flexural reinforcement permitted by clause 9.3.1.1(3) of
EN1992-1-1 is 3h, but not exceeding 400 mm, where h is the total depth of the slab. This rule
is for solid slabs. It is not intended for slabs formed with profiled steel sheeting, for which a
more appropriate rule is that given in clause 9.2.1(5): spacing not exceeding 2h (and
350 mm) in both directions, where h is the overall thickness of the slab, including ribs of
composite slabs.
7.4.2. Minimum reinforcement
Clause 7.4.2(1)
The only data needed when using Tables 7.1 and 7.2 are the tensile stresses in the
reinforcement,
s
. For minimum reinforcement,
s
is the stress immediately after initial
cracking. It is assumed that the curvature of the steel beam does not change, so all of the
tensile force in the concrete just before cracking is transferred to the reinforcement, of area
A
s
. If the slab were in uniform tension, equation (7.1) in clause 7.4.2(1) would be
A
s
s
= A
ct
f
ct, eff
The three correction factors in equation (7.1) are based on calibration work.
102
These allow
for the non-uniform stress distribution in the area A
ct
of concrete assumed to crack.
Non-uniform self-equilibrating stresses arise from primary shrinkage and temperature
effects, which cause curvature of the composite member. Slip of the shear connection also
causes curvature and reduces the tensile force in the slab.
The magnitude of these effects depends on the geometry of the uncracked composite
section, as given by equation (7.2). With experience, calculation of k
c
can often be omitted,
because it is less than 1.0 only where z
0
< 1.2h
c
. Especially for beams supporting composite
slabs, the depth of the uncracked neutral axis below the bottom of the slab (excluding ribs)
normally exceeds about 70% of the slab thickness, and then, k
c
= 1.
For design, the design crack width and thickness of the slab, h
c
will be known. It will be
evident whether there should be one layer of reinforcement or two. Two layers will often
consist of bars of the same size and spacing, which satisfies clause 7.4.2(3). For a chosen bar
diameter , Table 7.1 gives
s
, and equation (7.1) gives the bar spacing. If this is too high or
low, is changed.
132
Table 7.1. Use of steel fabric as minimum reinforcement, to clause 7.4.1(4)
Bar size and spacing
Cross-sectional area
(mm
2
per m width)
Maximum thickness of slab (mm)
Unpropped, 0.2% Propped, 0.4%
6 mm, 200 mm 142 71
7 mm, 200 mm 193 96 48
8 mm, 200 mm 252 126 63
144
12 May 2004 11:53:03
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:14
Atypical relationship between slab thickness h
c
, bar spacing s and bar diameter is shown
in Fig. 7.1. It is for two similar layers of bars, with k
c
= 1 and f
ct, eff
= 3.0 N/mm
2
. Equation
(7.1) then gives, for a slab of breadth b,
(
2
/4)(2b/s) = 0.72 3bh
c
/
s
Hence,
h
c
s = 0.727
2
s
(D7.3)
For each bar diameter and a given crack width, Table 7.1 gives
2
s
, so the product h
c
s is
known. This is plotted in Fig. 7.1, for w
k
= 0.3 mm, as curves of bar spacing for four slab
thicknesses, which can of course also be read as slab thicknesses for four bar spacings. The
shape of the curves results partly from the use of rounded values of
s
in Table 7.1. The
optional correction to minimumreinforcement given in clause 7.4.2(2) is negligible here, and
has not been made. Figure 7.1 can be used for slabs with one layer of bars by halving the slab
thickness.
The weight of minimumreinforcement, per unit area of slab, is proportional to
2
/s, which
is proportional to
s
1
, from equation (D7.3). This increases with bar diameter, from Table
7.1, so the use of smaller bars reduces the weight of minimumreinforcement. This is because
their greater surface area provides more bond strength.
Clause 7.4.2(2)
The method of clause 7.4.2(1) is not intended for the control of early thermal cracking,
which can occur in concrete a few days old, if the temperature rise caused by the heat of
hydration is excessive. The flanges of composite beams are usually too thin for this to occur.
It would not be correct, therefore, to assume a very low value for f
ct, eff
. The suggested value,
3 N/mm
2
, was probably rounded from the mean 28 day tensile strength of grade C30/37
concrete, given in EN 1992-1-1 as 2.9 N/mm
2
the value used as the basis for the optional
correction given in clause 7.4.2(2). The difference between 2.9 and 3.0 is obviously negligible.
If there is good reason to assume a value for f
ct, eff
such that the correction is not negligible, it
is best used by assuming a standard bar diameter , calculating *, and then finding
s
by
interpolation in Table 7.1.
133
100
200
300
400
0
h
c
= 100 mm
150
200
300
5 6 8 10 12 16 20
f (mm)
s (mm)
Fig. 7.1. Bar diameter and spacing for minimum reinforcement in two equal layers, for w
k
= 0.3 mm
and f
ct, eff
= 3.0 N/mm
2
145
12 May 2004 11:53:03
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:14
Reinforcement for encasement of a steel web
Clause 7.4.2(6) Clause 7.4.2(6) gives a minimum value of A
s
/A
ct
for encasement of the type shown in Fig. 6.1.
The maximumbar size is not specified. Detailing of this reinforcement is usually determined
by the requirements of clause 5.5.3(2) and of EN 1994-1-2, for fire resistance.
7.4.3. Control of cracking due to direct loading
Clause 7.4.3(2) Clause 7.4.3(2) specifies elastic global analysis to Section 5, allowing for the effects of
cracking. The preceding comments on global analysis for deformations apply also to this
analysis for bending moments in regions with concrete in tension.
Clause 7.4.3(3)
Paragraph (4) on loading should come next, but it is placed last in clause 7.4.3 because of
the drafting rule that general paragraphs precede those for buildings. It specifies the
quasi-permanent combination. Except for storage areas, the values of factors
2
for floor
loads in buildings are typically 0.3 or 0.6. The bending moments will then be much less than
for the ultimate limit state, especially for cross-sections in Class 1 or 2 in beams built
unpropped. There is no need to reduce the extent of the cracked regions belowthat assumed
for global analysis, so the newbending moments for the composite members can be found by
scaling values found for ultimate loadings. At each cross-section, the area of reinforcement
will be already known: that required for ultimate loading or the specified minimum, if
greater; so the stresses
s, 0
(clause 7.4.3(3)) can be found.
Tension stiffening
A correction for tension stiffening is now required. At one time, these effects were not well
understood. It was thought that, for a given tensile strain at the level of the reinforcement,
the total extension must be the extension of the concrete plus the width of the cracks, so that
allowing for the former reduced the latter. The true behaviour is more complex.
The upper part of Fig. 7.2 shows a single crack in a concrete member with a central
reinforcing bar. At the crack, the external tensile force N causes strain
s2
= N/A
s
E
a
in the
bar, and the strain in the concrete is the free shrinkage strain
cs
, which is negative, as shown.
There is a transmission length L
e
each side of the crack, within which there is transfer of
shear between the bar and the concrete. Outside this length, the strain in both the steel and
the concrete is
s1
, and the stress in the concrete is fractionally below its tensile strength.
134
L
e
L
e
N N
e
s2
e
sm
e
c
(x)
e
s
(x)
x
e
s1
e
cm
e
cs
Tensile strain
0
Fig. 7.2. Strain distributions near a crack in a reinforced concrete tension member
146
12 May 2004 11:53:03
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:14
Within the length 2L
e
, the curves
s
(x) and
c
(x) give the strains in the two materials, with
mean strains
sm
in the bar and
cm
in the concrete.
It is now assumed that the graph represents the typical behaviour of a reinforcing bar in a
cracked concrete flange of a composite beam, in a region of constant bending moment such
that the crack spacing is 2L
e
. The curvature of the steel beam is determined by the mean
stiffness of the slab, not the fully cracked stiffness, and is compatible with the mean
longitudinal strain in the reinforcement,
sm
.
Midway between the cracks, the strain is the cracking strain of the concrete,
corresponding to a stress less than 30 N/mm
2
in the bar. Its peak strain, at the crack, is much
greater than
sm
, but less than the yield strain, if crack widths are not to exceed 0.4 mm. The
crack width corresponds to this higher strain, not to the strain
sm
that is compatible with the
curvature, so a correction to the strain is needed. It is presented in clause 7.4.3(3) as a
correction to the stress
s, 0
because that is easily calculated, and Tables 7.1 and 7.2 are based
on stress. The strain correction cannot be shown in Fig. 7.1 because the stress
s, 0
is
calculated using the fully cracked stiffness, and so relates to a curvature greater than the
true curvature. The derivation of the correction
101
takes account of crack spacings less than
2L
e
, the bond properties of reinforcement, and other factors omitted from this simplified
outline.
The section properties needed for the calculation of the correction
s
will usually be
known. For the composite section, A is needed to find I, which is used in calculating
s, 0
, and
A
a
and I
a
are standard properties of the steel section. The result is independent of the
modular ratio. For simplicity,
st
may conservatively be taken as 1.0, because AI > A
a
I
a
.
When the stress
s
at a crack has been found, the maximumbar diameter or the maximum
spacing are found from Tables 7.1 and 7.2. Only one of these is needed, as the known area of
reinforcement then gives the other. The correction of clause 7.4.2(2) does not apply.
Influence of profiled sheeting on the control of cracking
The only references to profiled steel sheeting in clause 7.4 are in clause 7.4.1(4), no
account should be taken of any profiled steel sheeting, and in the definition of h
c
in clause
7.4.2(1), thickness excluding any haunch or ribs.
The effects of the use of profiled sheeting for a slab that forms the top flange of a
continuous composite beam are as follows:
there is no need for control of crack widths at the lower surface of the slab
where the sheeting spans in the transverse direction, there is at present no evidence that
it contributes to the control of transverse cracks at the top surface of the slab
where the sheeting spans parallel to the beam, it probably contributes to crack control,
but no research on this subject is known to the authors.
For design, the definition of effective tension area in clause 7.3.4(2) of EN 1992-1-1
should be noted. Alayer of reinforcement at depth c + /2 below the top surface of the slab,
where c is the cover, may be assumed to influence cracking over a depth 2.5(c + /2) of the
slab. If the depth of the concrete above the top of the sheeting, h
c
, is greater than this, it
would be reasonable to use the lower value, when calculating A
s
from equation (7.1). This
recognizes the ability of the sheeting to control cracking in the lower half of the slab, and has
the effect of reducing the minimum amount of reinforcement required, for the thicker
composite slabs
General comments on clause 7.4
In regions where tension in concrete may arise from shrinkage or temperature effects, but
not from other actions, the minimum reinforcement required may exceed that provided in
previous practice.
Where unpropped construction is used for a continuous beam, the design loading for
checking cracking is usually much less than that for the ultimate limit state, so that the
quantity of reinforcement provided for resistance to load should be sufficient to control
135
147
12 May 2004 11:53:04
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:14
cracking. The main use of clause 7.4.3 is then to check that the spacing of the bars is not too
great.
Where propped construction is used, the disparity between the design loadings for the two
limit states is smaller. If cracks are to be controlled to 0.3 mm, a check to clause 7.4.3 is more
likely to influence the reinforcement required.
For beams in frames, the preceding comments apply where semi-rigid or rigid connections
are used. Where floors have brittle finishes or an adverse environment, simple beam-
to-column joints should not be used, because effective control of crack width may not
be possible.
Example 7.1: two-span beam (continued) SLS
Details of this beam are shown in Fig. 6.23. All of the design data and calculations for the
ultimate limit state are given in Examples 6.7 to 6.12. For data and results required here,
reference should be made to:
Table 6.2, for characteristic loads per unit length
Table 6.3, for elastic properties of the cross-sections at the internal support (B in
Fig. 6.23(c)) and at mid-span
Table 6.4, for bending moments at support B for uniform loading on both spans
Fig. 6.28, for bending-moment diagrams for design ultimate loadings, excluding the
effects of shrinkage.
The secondary effects of shrinkage are significant in this beam, and cause a hogging
bending moment at support Bof 120 kN m(Example 6.7). Clause 7.3.1(8) does not permit
shrinkage to be ignored for serviceability checks on this beam, because it does not refer to
lightweight-aggregate concrete, which is used here.
Stresses
Fromclause 7.2.2(1), there are no limitations on stress; but stresses in the steel beamneed
to be calculated, because if yielding occurs under service loads, account should be taken of
the resulting increase in deflections, from clause 7.3.1(7).
Yielding is irreversible, so, from a note to clause 6.5.3(2) of EN 1990, it should be
checked for the characteristic load combination. However, the loading for checking
deflections depends on the serviceability requirement.
96
The maximumstress in the steel beamoccurs in the bottomflange at support B. Results
for the characteristic combination with variable load on both spans and 15% of each span
cracked are given in Table 7.2. The permanent load, other than floor finishes, is assumed
to act on the steel beam alone. Following clause 5.4.2.2(11), the modular ratio is taken as
20.2 for all of the loading except shrinkage.
136
Table 7.2. Hogging bending moments at support B and stresses in the steel bottom flange, for the
characteristic load combination
Loading
w
(kN/m)
Modular
ratio
10
6
I
y, B
(mm
4
)
M
Ek, B
(kN m)
10
6
W
a
,
bot
(mm
3
)
a, bot
(N/mm
2
)
(1) Permanent (on
steel beam)
5.78 337 104 1.50 69
(2) Permanent (on
composite beam)
1.2 20.2 467 18 1.75 10
(3) Variable 17.5 20.2 467 263 1.75 150
(4) Shrinkage 28.7 467 120 1.75 69
148
12 May 2004 11:53:04
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:14
The total bottom-flange compressive stress is
Ek, bot, a
= 298 N/mm
2
(= 0.84f
y
)
so no allowance is needed for yielding.
Deflections
The maximum deflection of span AB of the beam will occur at about 4.8 m from A (40%
of the span), when variable load acts on span ABonly. The additional deflection caused by
slip of the shear connection is ignored, as clause 7.3.1(4)(a) is satisfied.
Calculated deflections at this point, with 15% of each span assumed to be cracked, are
given in Table 7.3. The frequent combination is used, for which
1
= 0.7, so the variable
loading is
0.7 17.5 = 12.3 kN/m
The following method was used for the shrinkage deflections. From Example 6.7 and
Fig. 6.27, the primary effect is uniform sagging curvature at radius R = 1149 m, with
deflection = 45.3 mm at support B. The secondary reaction at B is 20 kN. From the
geometry of the circle, the primary deflection at point E in Fig. 7.3(a) is
1, E
= 45.3 5.4
2
1000 (2 1149) = 33 mm
The upwards displacement at E caused by the 20 kN reaction at B that moves point B
back to Bwas found to be 26 mmby elastic analysis of the model shown in Fig. 7.3(b), with
15% of each span cracked. The total shrinkage deflection is only 7 mm, despite the high
free shrinkage strain, but would not be negligible in a simply-supported span.
The total deflection, 31 mm, is span/390. This ratio appears not to be excessive.
However, the functioning of the floor may depend on its maximum deflection relative to
the supporting columns. It is found in Example 9.1 that the deflection of the composite
slab, if cast unpropped, is 15 mm for the frequent combination. This is relative to the
supporting beams, so the maximum floor deflection is
137
(a)
R
45.3
5.4
33
E
1.8 10.2
B'
B
A
A
26 mm
20 kN
10 kN 10 kN
45.3 mm
B E
(b)
Fig. 7.3. Sagging deflection at point E caused by shrinkage
Table 7.3. Deflections at 4.8 m from support A, for the frequent combination
Load Modular ratio Deflection (mm)
Dead, on steel beam 9
Dead, on composite beam 20.2 1
Imposed, on composite beam 20.2 14
Primary shrinkage 27.9 33
Secondary shrinkage 27.9 26
149
12 May 2004 11:53:04
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:14
31 + 15 = 46 mm
or span/260. For the characteristic combination, this increases to
37 + 16 = 53 mm
or span/230.
The limiting deflections given in the UKs draft National Annex to EN 1990
96
depend
on the serviceability requirement. For floors with partitions, they range from span/300
to span/500. For this floor, some combination of using propped construction for the
composite slabs and/or the beams, and cambering the beams, will be necessary.
Reducing the modular ratio for imposed loading to 10.1 makes little difference: the
value 14 mm in Table 7.3 becomes 12 mm. The extensive calculations for shrinkage lead
to a net deflection of only 7 mm, because the secondary effect cancels out most of the
primary effect. This benefit would not occur, of course, in a simply-supported span.
Control of crack width
Clause 7.4 applies to reinforced concrete that forms part of a composite member. In
the beam considered here, the relevant cracks are those near support B caused by
hogging bending of the beam, and cracks along the beam caused by hogging bending of
the composite slab that the beam supports. The latter are treated in Example 9.1 on
a composite slab.
Clause 7.4.1(1) refers to exposure classes. From clause 4.2(2) of EN 1992-1-1, Class
XC3 is appropriate for concrete inside buildings with moderate humidity. For this class,
a note to clause 7.3.1(5) of EN 1992-1-1 gives the design crack width as 0.3 mm. The
method of clause 7.4.1(3) is followed, as clause 7.4.1(4) does not apply.
Minimum reinforcement
The relevant cross-section of the concrete flange is as shown in Fig. 6.23(a), except that
effective widths up to 2.5 m should be considered.
From clause 5.4.1.2, the effective width is assumed to increase from 1.6 m at support B
to 2.5 m at sections more than 3 m from B. It may be difficult to show that sections 3 m
from B are never subjected to significant tension (clause 7.4.2(1)). Calculations are
therefore done for both effective widths, assuming uncracked unreinforced concrete.
From the definition of z
0
in that clause, n
0
= 10.1.
It is found for both of these flange widths that z
0
is such that k
c
> 1, so, from equation
(7.2), it is taken as 1.0.
A value is required for the strength of the concrete when cracks first occur. As
unpropped construction is used, there is at first little load on the composite member,
so from clause 7.4.2(1), conservatively, f
ct, eff
= 3.0 N/mm
2
. Assuming that 10 mm bars
are used for the minimum reinforcement, Table 7.1 gives
s
= 320 N/mm
2
. Then, from
equation (7.1),
100A
s
/A
ct
= 100 0.9 1 0.8 3.0/320 = 0.675%
However, clause 5.5.1(5) also sets a limit, as a condition for the use of plastic resistance
moments. For this concrete, f
lctm
= 2.32 N/mm
2
, and f
sk
= 500 N/mm
2
. Hence, from
equation (5.8) with k
c
= 1.0,
100
s
= 100 (355/235)(2.32/500) = 0.70% (D7.4)
This limit governs; so for a slab 80 mm thick above the sheeting the minimum
reinforcement is
A
s, min
= 7 80 = 560 mm
2
/m
One layer of 10 mm bars at 125 mm spacing provides 628 mm
2
/m.
138
150
12 May 2004 11:53:04
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:14
Cracking due to direct loading
Only the most critical cross-section, at support B, will be considered. Clause 7.4.3(4)
permits the use of the quasi-permanent combination, for which the variable loading is
2
q
k
, with
2
= 0.6, from clause A1.2.2(1) of EN 1990.
From Table 7.2, the bending moment at B that stresses the reinforcement is
M
E, qp, B
= 18 + 263 0.6 + 127 = 303 kN m
The neutral axis for the cracked section is 313 mm below the top of the slab (Table
6.12), so the section modulus for reinforcement at depth 30 mm is
10
6
W
s
= 467/(313 30) = 1.65 mm
3
Hence, from clause 7.4.3(3),
s, 0
= 303/1.65 = 184 N/mm
2
The correction for tension stiffening, equation (7.5), is now calculated, assuming
that the reinforcement used in Example 6.7, 12 mm bars at 125 mm spacing, will be
satisfactory. This gives
s
= 0.0113.
Using values obtained earlier,
st
= AI/A
a
I
a
= 11 350 467/(9880 337) = 1.59
s
= 0.4 2.32/(1.59 0.0113) = 52 N/mm
2
(D7.5)
s
= 184 + 52 = 236 N/mm
2
From Table 7.1,
s
16 mm. From Table 7.2, the bar spacing 200 mm.
The use of 12 mm bars at 125 mm spacing at support B satisfies both conditions.
Finding the cross-sections of the beam at which this reinforcement can be reduced to the
minimumfound above may require consideration of the bending-moment envelopes both
for ultimate loads and for the quasi-permanent combination.
139
151
12 May 2004 11:53:04
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
CHAPTER 8
Composite joints in frames for
buildings
Scope Clause 8.1
Analysis, modelling and classification Clause 8.2
Design methods Clause 8.3
Resistance of components Clause 8.4
8.1. Scope
Clause 8.1(1)
Section 8 is based on relatively recent research on beam-to-column and beam-to-beam joints
of the types used in steel and composite frames for buildings, so its scope has been limited to
frames for buildings. The definition of composite joints to which clause 8.1(1) refers
includes joints with reinforced concrete members. These could occur, for example, in a tower
block with a concrete core and composite floors. However, no application rules are given for
such joints.
Clause 8.1(2) As stated in clause 8.1(2), both Section 8 and Annex A are essentially extensions to the
Eurocode for joints between steel members, EN 1993-1-8.
24
It is assumed that a user will be
familiar with this code, especially its Sections 5 and 6.
The only steel members considered in detail are I- and H-sections, which may have
concrete-encased webs. Plate girders are not excluded.
The application rules of EN 1994-1-1 are limited to composite joints in which
reinforcement is in tension and the lower part of the steel section is in compression (Fig. 8.1
and clause 8.4.1(1)). There are no application rules for joints where the axes of the members
connected do not intersect, or do so at angles other than 90; but the basic approach is more
general than the procedures prescribed in detail, and is capable of application in a wider
range of situations.
Many types of joint are in use in steelwork, so that EN 1993-1-8 is around 130 pages long.
The majority of the calculations needed for composite joints are specified there, and
explained in the relevant guide in this series.
103
The worked examples and much of the
comment in the present guide are limited to a single type of joint the double-sided
configuration shown in Fig. 8.1 but with an end plate, not a contact plate, and an uncased
column, as shown on the left of Fig. 8.1 and in Fig. 8.8.
Commentary on the design of this joint will be found, as appropriate, in this chapter and its
examples, and in Chapter 10 on Annex A.
Before the Eurocodes come into regular use, it is expected that tables of resistances and
stiffnesses of a wide range of steel and composite joints will become available, based on
153
12 May 2004 11:53:05
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
EN 1993 and EN 1994. The extensive calculations given here will rarely be needed. Many of
them serve to show that a particular property of a joint does not govern its resistance.
Experience will enable such checks to be omitted.
Detailed guidance on the Eurocode methods for composite joints appeared in 1998, in the
context of design to British codes.
104
An explanation of the provisions and approximations in
the Eurocodes, with worked examples, was then prepared, mainly by those who drafted the
codes. Its first edition
39
refers to the draft codes as they were in 1998, so some differences,
mainly in symbols, will be found between it and the published EN Eurocodes. With over 200
pages, it provides much broader coverage than is possible here.
8.2. Analysis, including modelling and classification
Clause 8.2.1(1) Clause 8.2.1(1) refers to Section 5 of EN 1993-1-8, which covers the same subjects as clause
8.2. Table 5.1 in clause 5.1 of EN 1993-1-8 defines the links between the three types of global
analysis, elastic, rigid plastic, and elasticplastic, and the types of models used for joints. This
enables the designer to determine whether the stiffness of the joint, its resistance, or both
properties, are relevant to the analysis.
Joints are classified in Section 5 by stiffness, as rigid, nominally pinned, or semi-rigid; and
by strength as full-strength, nominally pinned, or partial-strength. This classification relates
the property of the joint (stiffness or resistance) to that of the connected member, normally
taken as the beam.
Clause 8.2.2(1) This applies also to composite joints. The only modification, in clause 8.2.2(1), concerns
the rotational stiffness of a joint, S
j
. This is bending moment per unit rotation, shown in Fig.
8.2(b). The symbol is used for rotation, as well as for bar diameter.
The initial elastic stiffness, S
j, ini
, is reduced at high bending moments to allow for inelastic
behaviour. For global analysis, it is divided by , values of which, between 3.0 and 3.5, are
tabulated in clause 5.1.2 of EN 1993-1-8 for various types of steel joint. These apply where
the joint is composite. Clause 8.2.2(1) provides a further value, for contact-plate joints, as
shown on the right of Fig. 8.1 and in Fig. 8.4.
Clause 8.2.3(2)
The classification of a composite joint may depend on the direction of the bending
moment (e.g. sagging or hogging). This is unlikely in a steelwork joint, and so is referred to in
clause 8.2.3(2).
Clause 8.2.3(3) The reference in clause 8.2.3(3) to neglect of cracking and creep applies only to the
classification of the joint according to stiffness. Its initial stiffness is to be compared with that
of the connected beam, using Fig. 5.4 of EN 1993-1-8. The stiffer the beam the less likely it is
that the joint can be classified as rigid.
A more precise calculation of beam stiffness is permitted by the use of may in clause
8.2.3(3). For example, a representative value of modular ratio, to clause 5.4.2.2(11), may be
used. Account could also be taken of cracked and uncracked lengths within the beam
142
C B A
L
b
Nominally pinned joint
Semi-rigid joint
Fig. 8.1. Model for a two-span beam in a frame
154
12 May 2004 11:53:05
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
in accordance with clause 5.4.2.3, but the additional calculation would not normally be
worthwhile.
Outline of modelling of joints for global analysis
In global analysis, nominally pinned joints are represented by pins, and semi-rigid joints by
rotational springs, as shown in Fig. 8.1 for a two-span beam of uniform depth, supported by
three columns in a braced frame. Joints to external columns are usually designed as
nominally pinned, to reduce bending moments in the columns. The use of partial-strength
semi-rigid joints at point B, rather than nominally pinned joints, has advantages in design:
possible reduction in the section sizes for beams
reduction in the deflection of beams
reduction in crack widths near support B.
In comparison with full-strength rigid joints, the advantages are:
beams less susceptible to lateraltorsional buckling
simpler construction and significant reduction in cost
lower bending moments in columns.
The stiffness of a rotational spring, S
j
, is the slope of the momentrotation relationship for
the joint (Fig. 8.2(a)). The stiffness class is determined by the ratio of the initial slope, S
j, ini
, to
the stiffness E
a
I
b
/L
b
of the beam adjacent to the joint, as shown.
The initial stiffness of a joint is assembled from the stiffnesses of its components,
represented by elastic springs. Those for an end-plate joint with a single row of bolts in
tension, between beams of equal depth are shown in Fig. 8.3, in which all elements except
springs and pins are rigid. The notation for the spring stiffnesses k
i
is as in EN1993-1-8 and in
Examples 8.1 and 10.1, as follows:
k
1
shear in column web
k
2
compression of column web
k
3
extension of column web
k
4
bending of column flange, caused by tension from a single row of bolts
k
5
bending of end plate, caused by tension from a single row of bolts
k
10
extension of bolts, for a single row of bolts.
Stiffnesses in EN 1994-1-1, but not in EN 1993-1-8, are:
k
s, r
extension of reinforcement (denoted k
13
by ECCS TC11
39
)
K
sc
/E
s
slip of shear connection.
Each spring has a finite strength, governed by yield or buckling of the steel. The design
method ensures that non-ductile modes, such as fracture of bolts, do not govern.
143
CHAPTER 8. COMPOSITE JOINTS IN FRAMES FOR BUILDINGS
(1) S
j, ini

8EI
b
/L
b
(3) S
j, ini

0.5EI
b
/L
b
(2) Semi-rigid

0
tan
1
S
j, ini
M
j
0
M
j
M
j, Ed
M
j, Rd
2M
j, Rd
/3
f
Cd
(b) (a)
(3) Nominal pin
(1) Rigid
tan
1
S
j
(S
j
= S
j, ini
/m)
(2)
f f
Fig. 8.2. Momentrotation relationships for joints
155
12 May 2004 11:53:05
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
For the tension region, the weakest of the springs numbered 3, 4, 5 and 10, and of the
tension reinforcement, is found. This resistance is compared with the compressive resistance
of spring 2. The product of the lower of these resistances and the effective lever armgives the
plastic bending resistance of the joint. The resistance can be increased by strengthening the
weakest link; for example, by the addition of column-web stiffeners.
Where the beams are of unequal depth, or M
Ed, 1
M
Ed, 2
(Fig. 8.3), rotation at the joint is
increased by shearing deformation of the column web. For beams of equal depth, this is the
area ABCD in Fig. 8.3. Its deformation is resisted by the spring of stiffness k
1
. Depending
on the out-of-balance moment |M
Ed, 1
M
Ed, 2
|, the column web panel may govern the
resistance of the joint.
8.3. Design methods
Clause 8.3.1(1) Clause 8.3.1(1) refers to Section 6 of EN1993-1-8, which is 40 pages long. It defines the basic
components of a steelwork joint, their strengths and their elastic stiffnesses. It is shown how
these are assembled to obtain the resistances, rotational stiffness and rotation capacity of
complete joints.
A composite joint has these additional components:
longitudinal slab reinforcement in tension
concrete encasement, where present, of the column web
steel contact plates, if used (not covered in EN 1993-1-8).
In addition, account is taken of the slip of shear connection, by modifying the stiffness of the
reinforcement (Fig. 8.3).
Clause 8.3.1(2)
Clause 8.3.1(3)
All the properties of components given in, or cross-referenced from, EN 1994-1-1 satisfy
the condition of clause 8.3.1(2). The application of clause 8.3.1(3) to reinforcing bars is
illustrated in Examples 8.1 and 10.1.
Clause 8.3.2(1)
None of the additional components listed above influences resistance to vertical shear, so
this aspect of design is fully covered by EN 1993-1-8 (clause 8.3.2(1)).
144
k
1
k
2
k
3
k
4, 5, 10
k
s, r
K
sc
/E
s
D
C
B A
M
Ed, 2
M
Ed, 1
Fig. 8.3. Model for an internal beam-to-column joint
156
12 May 2004 11:53:06
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
Clause 8.3.2(2)
Composite joints in framed structures for buildings are almost always in regions of
hogging bending, for which full shear connection is normally required, to clauses 6.2.1.3(2)
and 8.4.2.1(2). The reference to shear connection in clause 8.3.2(2) reminds the user that no
provisions are given for composite joints in regions with partial shear connection.
Clause 8.3.3(1) The use of Annex A (informative) for finding rotational stiffnesses satisfies clause 8.3.3(1),
and is illustrated in Example 10.1.
Clause 8.3.3(2) The coefficient , referred to in clause 8.3.3(2), is used in clause 6.3.1(6) of EN1993-1-8 to
define the shape of the moment-rotation curve for a joint at bending moments M
j, Ed
that exceed 2M
j, Rd
/3, as follows. Let S
j, ini
be the stiffness at low bending moments. For
2M
j, Rd
/3 < M
j, Ed
M
j, Rd
, the stiffness is
S
j
= S
j, ini
/ (D8.1)
(Fig. 8.2(b)), where
= (1.5M
j, Ed
/M
j, Rd
)
(D8.2)
This clause gives the value for for a type of joint not included in Table 6.8 of EN 1993-1-8.
The table is applicable to other types of composite joint.
The rotation capacity of composite joints,
Cd
in Fig. 8.2(b), has been extensively
researched.
105
There are many relevant parameters. Analytical prediction is still difficult,
and there are as yet no design rules sufficiently well established to be included in
EN 1994-1-1.
Clause 8.3.4(2)
So-called simple joints have been widely used in composite structures. Some of them will
be found to qualify as partial-strength when Eurocode methods are used. The experience
referred to in clause 8.3.4(2) is then available. It is rarely necessary in design to calculate
either the available rotation capacity or the rotation required of a composite joint. Further
guidance is given by ECCS TC11.
39
8.4. Resistance of components
Clause 8.4.2.1(1)
This clause supplements clause 6.2 of EN 1993-1-8. The effective width of concrete flange in
tension is the same at a joint as for the adjacent beam (clause 8.4.2.1(1)). Longitudinal bars
above the beam should pass either side of the column.
Clause 8.4.2.1(4) Clause 8.4.2.1(4) applies at an external column with a partial- or full-strength joint. The
tensile force in the bars must be transferred to the column; for example, by being looped
round it. This applies also at internal columns where there is a change in the tension in the
bars (Fig. 8.2, clause 8.4.2.1(3) and Example 8.2).
Clause 8.4.2.2(1)
Clause 8.4.3(1)
Clauses 8.4.2.2(1) and 8.4.3(1) permit the same 45 spread of force in a contact plate
as used in EN 1993-1-8 for an end plate. The force is assumed in EN 1993-1-8 to spread at
tan
1
2.5 (68) through the flange and root radius of the column. Where the compressive
force relied on in design exceeds the resistance of the steel bottom flange, the length of the
contact plate should allow for this (Fig. 8.4).
In EN 1994-1-1, the word connection appears only in clause 8.4.3(1), clause A.2.3.2 and
Table A.1. It means the set of components that connect a member to another member; for
example, an end plate, its bolts and a column flange. Thus, a connection is part of a joint.
Clause 8.4.4.1(2) The model used in clause 8.4.4.1(2) is illustrated in Fig. 8.5. This figure shows an elevation
of the concrete encasement of width h 2t
f
(column depth less flange thicknesses) and
depth z, the lever arm between the resultant horizontal forces from the beam. A shear
force V is transferred through the encasement, which is of thickness b
c
t
w
(column
width less web thickness). The concrete strut ABDEFG has width 0.8(h 2t
f
)cos , where
tan = (h 2t
f
)/z, so its area is
A
c
= 0.8(h 2t
f
)cos (b
c
t
w
) (8.2)
145
157
12 May 2004 11:53:06
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
146
45
68
Contact plate
Column
Beam
Fig. 8.4. Detail of a contact plate between a beam bottom flange and a column
B
G
F E
D
q
q
h 2t
f
0.4(h 2t
f
)
0.8(h 2t
f
)cos q
V
V
A
z
N
Ed
C
Fig. 8.5. Strut model for the shear resistance of the concrete encasement to a column web
t
eff, c
t
f, c
t
p
t
f, b
0.212
0
1.3
2.0
k
wc, c
s
com, c, Ed
/f
cd

(b)
45
68
End plate
Beam
Column
r
(a)
Fig. 8.6. Model for resistance to compression of the concrete encasement to a column web
158
12 May 2004 11:53:07
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
Its compressive strength is 0.85f
cd
, giving the force C in Fig. 8.5. For horizontal equilibrium
at B and F, C sin = V. These are equations (8.1) to (8.3) in clause 8.4.4.1.
Clause 8.4.4.1(3)
The shear strength of concrete is increased by compression. This is allowed for by the
factor in clause 8.4.4.1(3), which ranges from 0.55 for zero axial compression to 1.1 for
N
Ed
0.55N
pl, Rd
.
Clause 8.4.4.2
The contribution of concrete encasement to the resistance of a column web to horizontal
compression is given in clause 8.4.4.2. For an end-plate joint to a column flange, the depth of
encasement assumed to resist compression, t
eff, c
, is shown in Fig. 8.6(a), with the 2.5:1
dispersion, referred to above, extending through the root radius r.
The horizontal compressive strength of the concrete is 0.85 k
wc, c
f
cd
, where k
wc, c
depends on
the vertical compressive stress in the column,
com, c, Ed
, as shown in Fig. 8.6(b).
Example 8.1: end-plate joints in a two-span beam in a braced frame
In development work that followed the publication of ENV 1994-1-1, a set of application
rules for composite joints was prepared, more detailed than those now given in Section 8
and Annex A of EN 1994-1-1. These are published by ECCS
39
as a model annex J. They
provide useful guidance in this example, and are referred to, for example, as clause J.1.1
of ECCS TC11.
Data
The subject of Examples 6.7 and 7.1 is a two-span beam ABC continuous over its central
support (see Figs 6.236.28). It is now assumed that this beam is one of several similar
beams in a multistorey braced frame (Fig. 8.7). Its joints with the external columns are
nominal pins. The spans of the composite-slab floors are 2.5 m, as before. For simplicity, in
the work on beams ABand BC, column EBFwill be treated as fixed at nodes Eand F. These
beams are attached to the column at Bby the end-plate connections shown in Fig. 8.8, which
also gives dimensions of the column section. Its other properties are as follows: HEB 240
cross-section, A
a
= 10 600 mm
2
, f
y
= 355 N/mm
2
, 10
6
I
y
= 112.6 mm
4
, 10
6
I
z
= 39.23 mm
4
.
The end plates are of mild steel, f
y
= 275 N/mm
2
, and relatively thin, 12 mm, to provide
the plastic behaviour required. They are attached to the beamby 10 mmfillet welds to the
flanges, and 8 mm welds to the web. They are each attached to the column by four Grade
8.8 M20 bolts with properties: f
ub
= 800 N/mm
2
, f
yb
= 640 N/mm
2
, net area at root of
thread A
s, b
= 245 mm
2
per bolt.
The only other change from the data used in Examples 6.7 and 7.1 on geometry,
materials and loadings concerns the reinforcement in the slab.
Longitudinal reinforcement at support B
These partial-strength joints need rotation capacity. Its value cannot be found at this
stage, but it is known to increase with both the diameter of the reinforcing bars in the slab
147
C B A
12
3
3
12
D
F
E
Fig. 8.7. Model for a two-span beam ABC, with an internal column EBF
159
12 May 2004 11:53:07
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
and the area of reinforcement provided. However, the amount of top reinforcement
should be limited, so that the whole of the compressive force across the joint can be
resisted by the beam bottom flange and the unstiffened column web.
Detailed guidance is given in Couchman and Way.
104
For a steel beam of depth 450 mm
in S355 steel, the recommended minimum areas are 3000 mm
2
for bars with 5%
elongation and 860 mm
2
for bars with 10% elongation. The recommended maximum
amount depends on the size of the column and the details of the bolts in tension, and is
about 1200 mm
2
for this example. The recommended bar diameters are 16 and 20 mm.
For these reasons, the previous reinforcement (13 No. 12 mm bars, A
s
= 1470 mm
2
) is
replaced by six No. 16 mm hot-formed bars (minimum elongation 10%): A
s
= 1206 mm
2
,
f
sk
= 500 N/mm
2
.
Classification of the joints
It is assumed initially that flexural failure of a joint will occur in a ductile manner, by
yielding of the reinforcement in tension and the end-plate or column flange in bending;
148
160
383
m
2
= 35.4 + 2
240
14.6
10 12
450
25
200
60
25
60
90
= =
A
9.4
A
230 270 130
800
240
17
10
21
100 mm
16 dia.
30
100
Fig. 8.8. Details of the beam-to-column end-plate connections
(a) Elevation (b) Section AA
(c) Column section (d) Slab reinforcement
160
12 May 2004 11:53:07
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
and that at bottom-flange level the compressive resistance of the column web will
be sufficient. As the spans are equal, it is unlikely that shear of the column web will be
critical.
The joint is expected to be partial-strength. This can be checked by comparing the
tension resistance of the top two bolts, from Table 3.4 of EN 1993-1-8, with the force to
yield the beam top flange:
F
T, Rd, bolts
= 2(k
2
f
ub
A
s
/
M2
) = 2 0.9 0.8 245/1.25 = 282 kN (D8.3)
F
Rd, flange
= b
f
t
f
f
yd
= 190 14.6 0.355 = 985 kN
Thus, the resistance moment M
j, Rd
for the joint will be much less than M
pl, Rd
for the beam,
and lateral buckling will be less critical than before.
There is no need to find the stiffness of the joint at this stage, because it is clearly either
rigid or semi-rigid. Either type may be treated as semi-rigid.
Approximate global analysis
Tables in Appendix B of Couchman and Way
104
enable a rough check to be made on this
initial design, without much calculation. They give resistances M
j, Rd
in terms of the
cross-section of the steel beam, its yield strength, the thickness and grade of the end plate,
the number and size of bolts in tension, and the area of reinforcement. Even though the
beam used here is an IPE section, it can be deduced that M
j, Rd
is about 400 kN m.
For both spans fully loaded, it was found in Example 6.7 that M
Ed
at B was 536 kN m
from loading (see Fig. 6.28) plus 120 kN m from shrinkage. The flexural stiffness of the
joint is not yet known, but it will be between zero and fully rigid. If fully rigid, the joint
will obviously be plastic under ultimate loading, and there will then be no secondary
shrinkage moment. At mid-span, for the total load of 35.7 kN/m (see Table 6.2), the
sagging bending moment is then
357 12
2
/8 400/2 = 443 kN m
If the joint acts as a pin, the mid-span moment is
443 + 200 = 643 kN m
It is recommended in Couchman and Way
104
that mid-span resistances should be taken as
0.85M
pl, Rd
, to limit the rotation required at the joints. From Example 6.7, M
pl, Rd
with full
shear connection is 1043 kN m, so the bending resistance of the beam is obviously
sufficient.
Vertical shear
For M
Ed
= 400 kN m at B, the vertical shear at B is
F
v, Ed, B
= 35.7 6 + 400/12 = 247 kN
The shear resistance of the four M20 bolts is now found, using Table 3.4 of EN 1993-1-8.
Two of the bolts may be at yield in tension. The shear applied to these bolts must satisfy
F
v, Ed
/F
v, Rd
+ F
t, Ed
/(1.4F
t, Rd
) 1.0 (D8.4)
The net shear area of each bolt is A
s, b
= 245 mm
2
, so from Table 3.4 of EN 1993-1-8,
F
v, Rd
= 0.6f
ub
A
s, b
/
M2
= 0.6 800 0.245/1.25 = 94.1 kN
From equation (D8.3) with F
t, Ed
= F
t, Rd
,
F
v, Ed
(1 1/1.4)F
v, Rd
= 27 kN
For four bolts,
F
v, Rd
= 2 (94.1 + 27) = 242 kN (D8.5)
149
161
12 May 2004 11:53:08
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
This shows that it may be necessary to add a second pair of bolts in the compression region
of the joint.
Bending resistance of the joint, excluding reinforcement
Unpropped construction was used in Example 6.7. From Table 6.2, the design ultimate
load for the steel beam is 7.8 kN/m. For the construction phase, this is increased to
9.15 kN/m, to allow for the higher density of fresh concrete and the construction imposed
loading. For rigid joints at the internal support between two 12 m spans,
M
Ed, B
= wL
2
/8 = 9.15 12
2
/8 = 165 kN m
The plastic resistance of the joint during construction is required. An upper limit is easily
obtained. The lever arm from the top bolts to the centre of the bottom flange is
z
bolts
= 450 60 7.3 = 383 mm (D8.6)
The resistance cannot exceed
F
T, Rd, bolts
z
bolts
= 282 0.383 = 108 kN m (D8.7)
so the previous hogging bending moment of 165 kN m cannot be reached. It is assumed
that the stiffness of the joint is sufficient for its plastic resistance, found later to be
83 kN m, to be reached under the factored construction loading.
Resistance of T-stubs and bolts in tension
The calculation of the bending resistance consists of finding the weakest links in both
tension and compression. In tension, the column flange and the end plate are each
modelled as T-stubs, and prying action may occur. Some of the dimensions required are
shown in Fig. 8.9. FromFig. 6.8 of EN1993-1-8, the dimensions moverlap with 20%of the
corner fillet or weld. Thus, in Fig. 8.9(a), for the end plate:
m = 45 4.7 0.8 8 = 33.9 mm (D.8.8)
It is evident from the geometry shown in Fig. 8.9 that the end plate is weaker than the
column flange, so its resistance is now found.
150
t
p
= 12
t
f
= 17
0.8r
c
= 16.8
m = 23.2
e = 75
r
c
= 21
m = 33.9
55
8
m
2
= 37.4
55 33.9
Fig. 8.9. Dimensions of T-stubs, and the yield line pattern
(a) Plan details of T-stubs
(b) Yield line pattern in end plate
162
12 May 2004 11:53:08
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
Clause 6.2.4.1 of EN 1993-1-8 gives three possible failure modes:
(1) yielding of the plate
(2) a combination of (1) and (3)
(3) failure of the bolts in tension.
Yield line theory is used for bending of the plate. The critical mechanism in this case
will be either that shown in Fig. 8.9(b) or a circular fan, for which the perimeter is
eff, cp
= 2m
from Table 6.6 in clause 6.2.6.5 of EN 1993-1-8.
For the non-circular pattern, dimension m
2
in Fig. 8.8(a) is also relevant, and
eff, nc
= m 2m
where is given by Fig. 6.11 in EN1993-1-8 or in Fig. 4.9 of Couchman and Way.
104
In this
case, = 6.8, so the circular pattern governs for mode (1), and
eff, 1
= 2m = 6.28 33.9 = 213 mm
The plastic resistance per unit length of plate is
m
pl, Rd
= 0.25t
f
2
f
y
/
M0
= 0.25 12
2
0.275/1.0 = 9.90 kN m/m (D8.9)
From equation (D8.3), the tensile resistance of a pair of bolts is 282 kN.
Mode 1. For mode 1, yielding is confined to the plate. FromTable 6.2 of EN1993-1-8, the
equation for this mode is
F
T, 1, Rd
= 4M
pl, 1, Rd
/m
with
M
pl, 1, Rd
= 0.25
eff, 1
t
f
2
f
y
/
M0
=
eff, 1
m
pl, Rd
= 0.213 9.9 = 2.11 kN m
From equation (D8.8) for m,
F
T, 1, Rd
= 4 2.11/0.0339 = 249 kN
Mode 2. This mode is more complex. The equation for the tension resistance F
T, 2, Rd
is
now explained. The effective length of the perimeter of the mechanism is
eff, 2
, and the
work done for a rotation at its perimeter is 2m
pl, Rd
eff, 2
, from yield line theory. With
each bolt failing in tension, the work equation is
151
F
t, Rd
F
T, 2, Rd
Q/2
m n
q
Fig. 8.10. Plan of a T-stub, showing failure mode 2
163
12 May 2004 11:53:08
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
F
T, 2, Rd
(m + n) = n
S
F
t, Rd
+ 2m
pl, Rd
eff, 2
(D8.10)
where n is shown in Fig. 8.10.
In Table 6.2 of EN 1993-1-8 there is the further condition that
n 1.25m = 1.25 33.9 = 42.4 mm
The effective length
eff, 2
= m = 6.8 33.9 = 231 mm
from Table 6.6 of EN 1993-1-8. For two bolts,
S
F
t, Rd
= 282 kN, from equation (D8.3).
Substituting in equation (D8.10):
F
T, 2, Rd
= (2m
pl, Rd
eff, 2
+ n
S
F
t, Rd
)/(m + n)
= (2 9.9 231 + 42.4 282)/76.3 = 217 kN (D8.11)
Mode 3. Failure of the bolts mode 3 has
F
T, 3, Rd
= 282 kN
from equation (D8.3), so mode 2 governs. From Fig. 8.10, the prying force is
Q = 282 217 = 65 kN
Beam web in tension
The equivalent T-stub in Fig. 8.10 applies a tensile force of 217 kNto the web of the beam.
Its resistance is given in clause 6.2.6.8 of EN 1993-1-8 as
F
t, wb, Rd
= b
eff, t, wb
t
wb
f
y, wb
/
M0
(equation (6.22) of EN1993-1-8), and b
eff, t, wb
is taken as the effective length of the T-stub,
eff, 2
= 231 mm. Hence,
F
t, wb, Rd
= 231 9.4 0.355/1.0 = 771 kN (D8.12)
so this does not govern.
Column web in tension
The effective width of the column web in tension, to clause 6.2.6.3 of EN 1993-1-8, is the
length of the T-stub representing the column flange. The resistance is
F
t, wc, Rd
= b
eff, t, wc
t
wc
f
y, wc
/
M0
where is a reduction factor to allowfor shear in the column web. In this case, the shear is
zero, and = 1. The column web is thicker than the beam web, so from result (D8.12), its
resistance does not govern.
Column web in transverse compression
The resistance is given in clause 6.2.6.2 of EN 1993-1-8. It depends on the plate slenderness
p
and the width of the column web in compression, which is
b
eff, c, wc
= t
f, b
+ 22a
p
+ 5(t
fc
+ s) + s
p
(equation (6.11) of EN 1993-1-8), where a
p
is the throat thickness of the bottom-flange
welds, so 2a
p
= 10 mm here; s
p
allows for 45 dispersion through the end plate, and is
24 mm here; and s is the root radius of the column section (s = r
c
= 21 mm). Hence,
b
eff, c, wc
= 14.6 + 20 + 5 (17 + 21) + 24 = 248 mm
For web buckling, the effective compressed length is
d
wc
= h
c
2(t
fc
+ r
c
) = 240 2 (17 + 21) = 164 mm
152
164
12 May 2004 11:53:09
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
The plate slenderness is
= 0.932(b
eff, c, wc
d
wc
f
y, wc
/E
a
t
wc
2
)
0.5
= 0.932 [248 164 0.355/(210 100)]
0.5
= 0.773
The reduction factor for plate buckling is
= ( 0.2)/
2
= 0.573/0.773
2
= 0.96
The factor for web shear is 1.0, as before.
It is assumed that the maximum longitudinal compressive stress in the column is less
than 0.7f
y, wc
, so from clause 6.2.6.2(2) of EN 1993-1-8, the reduction factor for this, k
wc
, is
1.0. From equation (6.9) in EN 1993-1-8,
F
c, wc, Rd
= k
wc
b
eff, c, wc
t
wc
f
y, wc
/
M1
= 0.96 248 10 0.355/1.0 = 845 kN (D8.13)
Clearly, the tensile force of 217 kN governs the resistance of the steel connection.
Bending resistance of the steel joint, for both beams fully loaded
From equation (D8.5), the lever arm is 383 mm, so the resistance, excluding the
reinforcement, is
M
j, Rd, steel
= 217 0.383 = 83 kN m (D8.14)
governed by bending of the end plate. The critical mode 2 includes failure of the top row
of bolts in tension. However, the joint is closely based on a type given in Couchman and
Way,
104
which is confirmed by ECCS TC11
39
as having ductile behaviour.
From Example 6.7, the plastic bending resistance of the steel beam, an IPE 450 section,
is
M
pl, a, Rd
= 1.702 355 = 604 kN m
This exceeds four times M
j, Rd
, so clause 5.2.3.2(3) of EN 1993-1-8 permits this joint to be
classified as nominally pinned for the construction stage.
Resistance of the composite joint
For the composite joint, the reinforcement is at yield in tension. Its resistance is
F
t, s, Rd
= 1206 0.500/1.15 = 524 kN
This increases the total compressive force to
F
c
= 217 + 524 = 741 kN (D8.15)
This is less than the compressive resistance of 845 kN, found above. The bars act at a lever
arm of 543 mm (Fig. 8.8(a)), so the bending resistance of the composite joint is
M
j, Rd, comp
= 83 + 524 0.543 = 83 + 284 = 367 kN m (D8.16)
Check on vertical shear
For the maximum design beam load of 35.7 kN/m and a hogging resistance moment at B
of 367 kN m, the vertical shear in each beam at B is 244 kN, which just exceeds the shear
resistance found earlier, 242 kN. It will probably be found from elasticplastic global
analysis that the vertical shear at B is reduced by the flexibility of the joints. If necessary,
two extra M20 bolts can be added in the lower half of each end plate. This has no effect on
the preceding results for resistance to bending.
Maximum load on span BC, with minimum load on span AB
This loading causes maximum shear in the column web. There is an abrupt change in the
tension in the slab reinforcement at B. The load acting on the steel members is equal for
153
p
165
12 May 2004 11:53:09
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
the two spans, and is assumed to cause a hogging bending moment at node B equal to the
resistance of the joints, 83 kN m from equation (D8.11). The ultimate loads on the
composite member are 1.62 kN/m on AB and 27.9 kN/m on BC (see Table 6.2).
The flexibility of the joints and cracking of concrete both reduce hogging bending
moments, so both are neglected in these checks on shear in the column web and
anchorage of the reinforcement. The moment on the composite joint at B in span BC is
taken as the additional resistance provided by the slab reinforcement, which is 284 kN m
(equation (D8.16)). For the other three members meeting at node B, elastic analysis
gives the bending moments shown in Fig. 8.11(a). The total bending moments at B,
including construction, are shown in Fig. 8.11(b). The shear forces in columns DB and
BE are
(75 + 37.5)/3 = 37.5 kN
If the end plate in span AB is plastic under ultimate construction loading, the whole of
the difference between the beam moments shown is caused by change of tension in the
reinforcement. Thus, the relevant lever arm, z, is 543 mm.
From clause 5.3(3) of EN 1993-1-8, the shear force on the web panel is
V
wp, Ed
= (M
b1, Ed
M
b2, Ed
)/z (V
c1, Ed
V
c2, Ed
)/2
= (367 217)/0.543 [37.5 (37.5)]/2 = 276.2 37.5 = 239 kN
The sign convention used here is given in Fig. 5.6 of EN 1993-1-8. It is evident from Fig.
8.11(b) and the equation above that the web shear from the beams, 276 kN, is reduced by
the shear forces in the column. The change of force in the reinforcement is 276 kN.
Shear resistance of the column web
From clause 6.2.6.1(1) of EN 1993-1-8, the shear resistance of an unstiffened column web
panel is
V
wp, Rd
= 0.9f
y, wc
A
vc
/(3
M0
)
where A
vc
is the shear area of the column web. This is given in EN 1993-1-1, and is
3324 mm
2
here. Hence,
V
wp, Rd
= 0.9 0.355 3324/(3 1.0) = 613 kN (> V
wp, Ed
)
This load arrangement therefore does not govern the design of the joint.
154
z
134 + 83
37.5
75
D
C
B
A
284
37.5
75
E
12
3
134
12
37.5
75
284 + 83
1.62 kN/m 27.9 kN/m
Fig. 8.11. Analyses for unequal design loadings (ultimate limit state) on spans AB and BC
(a) Bending-moment diagram (kN m) (b) Action affects on joint (kN and kN m)
166
12 May 2004 11:53:09
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
Anchorage of the force from the reinforcement
The force of 276 kN (above) has to be anchored in the column. The strut-and-tie model
shown in Fig. 8.2 requires transverse reinforcement to resist the force F
tq
shown in the
figure, and the force depends on the directions chosen for the struts.
The mean distance of the three 16 mm bars shown in Fig. 8.8(d) from the centre-line of
the column is 420 mm. The two concrete struts AB and AD shown in Fig. 8.12 can, for
calculation, be replaced by line AC. Resolution of forces at point Agives the strut force as
184 kN. The depth of concrete available is 80 mm. With f
ck
= 25 N/mm
2
, the total width of
the struts is
b
c
= (184 1.5)/(0.08 0.85 25) = 163 mm
This width is shown to scale in Fig. 8.12, and is obviously available.
The existing transverse reinforcement (Example 6.7) is 12 mm bars at 200 mm spacing.
Insertion of three more bars (A
s
= 339 mm
2
) at 200 mm spacing provides an extra
resistance
T
Rd
= 339 0.5/1.15 = 147 kN
which exceeds the tie force of 122 kN shown in Fig. 8.12. Three extra bars are provided on
each side of the column.
The available area of column flange to resist bearing stress is
80 (120 + 115) = 18 800 mm
2
so the mean stress is
155
120
420
50
A
300 120
B
C
184 kN
163
138 kN
D
3 No. 12 bars at 200 mm
122 kN
Fig. 8.12. Strut-and-tie model for anchorage of unbalanced tension in slab reinforcement
Table 8.1. Initial stiffnesses of joints, S
j, ini
(kN m/mrad)
No shear
Joint BA,
with shear
Joint BC,
with shear
Steel joint 61.1
Composite joint, elastic 146 118 48.8
Composite joint, no top bolts 110 118 38.6
167
12 May 2004 11:53:09
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
(276/2)/18.8 = 7.34 N/mm
2
well below the design compressive strength of the concrete.
Stiffness of the joints and rotation capacity, ultimate limit state
The initial stiffnesses of these joints are calculated in Example 10.1, and are given in Table
8.1. In the preceding global analyses, the joints were assumed to be rigid until their
resistance M
j, Rd
was reached, and to act as hinges for further loading. This neglect of the
elastic rotation of the joints at moments below M
j, Rd
leads to overestimation of the
hogging bending moments at the joints. The mid-span moments therefore exceed those
calculated. This method is safe for the verification of the joints, and is appropriate where
there is ample bending resistance at mid-span, as in this example.
There is little inelastic curvature in the regions of sagging moment, so the rotation at
the joints is much less than that required for the development of mid-span plastic hinges.
The typical joint details given in Couchman and Way,
104
which were used here, were
shown to have adequate rotation capacity by a calculation method
19
supported by tests
(clause 8.3.4(3)) or are those which experience has proved have adequate properties (clause
8.3.4(2)), so no further verification of rotation capacity is required.
Serviceability checks
The preceding analyses are inadequate for serviceability checks on deflections or crack
width. Account must be taken of the flexibility of each joint, as given by equations
(D8.1) and (D8.2). From clause 6.3.1(6) of EN 1993-1-8 or clause J.4.1(5) of ECCS
TC11,
39
= 2.7 for bolted end-plate joints. Thus, where M
j, Ed
= M
j, Rd
, from equation
(D8.2),
= 1.5
2.7
= 2.99 (D8.17)
This dependence of the joint stiffness S
j
on M
j, Ed
leads to iterative analysis, so
simplifications are given in clause 5.1.2 of EN 1993-1-8 and in section 9.5 of ECCS TC11
39
for use in elastic frame analysis, as follows:
for a composite beam-to-column joint with a flush end-plate connection, and usual
cases, a nominal stiffness S
j
= S
j, ini
/2 may be used, for moments up to M
j, Rd
(i.e.
= 2)
where joints are required to behave within their elastic range, the stiffness S
j, ini
should
be used, and M
j, Ed
should not exceed 2M
j, Rd
/3.
Here, the joints are included in conventional elastic analyses as follows. FromFig. 8.8, a
pair of joints (see Fig. 8.1) has an overall length of 240 + 2 12 = 264 mm, so each joint
is represented by a beam-type member of length L
j
= 132 mm and second moment of
area I
j
. For a bending moment M
j
, the rotation is
= M
j
/S
j
For the beam, it is
= M
j
L
j
/E
a
I
j
Eliminating /M
j
,
I
j
= (L
j
/E
a
)S
j
Hence,
10
6
I
j
= (132/210)S
j
= 0.63S
j
(D8.18)
with I
j
in units of mm
4
and S
j
in units of kN m/mrad.
156
168
12 May 2004 11:53:10
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
Maximum deflection, with imposed load on span BC only
Separate calculations are required for the steel joints and the composite joints. The load
at the end of construction, 5.78 kN/m, is applied to both spans. From Table 8.1 and
equation (D8.18), with = 2,
10
6
I
j
= 0.63 61.1/2 = 19.2 mm
4
For the beam,
10
6
I
ay
= 337.4 mm
4
so the calculation model for the steel joints is as shown in Fig. 8.13, with I in units of mm
4
.
The results are:
bending moment in the connection in span BC: 67.7 kN m, which is 82%of M

j, Rd, steel
maximum deflection of span BC: 13.8 mm.

The calculation model for the composite phase takes account of cracking, and uses the
modular ratio n = 20.2. The floor finishes, 1.2 kN/m, act on both spans, and the frequent
value of the imposed load, 0.7 17.5 = 12.3 kN/m, acts on span BC only.
For the joint in span BC, = 2. There is a little unused tensile resistance from
the steel connection. If this is neglected, then for the composite joint in span BC,
S
j, ini
= 39.6 kN m/mrad (Table 8.1) and
10
6
I
j, BC
= 0.63 39.6/2 = 12.5 mm
4
(The effect of including the full stiffness of the steel joint (S
j, ini
= 48.8 kN m/mrad, not
38.6 kN m/mrad) is small; it reduced the deflection by less than 1 mm.)
The moment in the connection in span BA is low, so = 1. From Table 8.1,
10
6
I
j, BA
= 0.63 118 = 74 mm
4
For the column,
10
6
I
y
= 113 mm
4
The calculation model is shown in Fig. 8.14, with I in units of mm
4
. The results are:
157
M
j,Ek
132
337.4 19.2
5.78 kN/m
11 868
C
B
10
6
I
Fig. 8.13. Analysis for steel beams, serviceability limit state
132
828 12.2
13.5 kN/m
10 200
C A
1668
467
828
467 74
1.2 kN/m
12 000
10
6
I
10
6
I
113
Fig. 8.14. Analysis for composite beams, serviceability limit state
169
12 May 2004 11:53:10
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
bending moment at B in span BC: 70.2 kN m
maximum deflection of span BC: 17.9 mm.

The total deflection of span BC is therefore
13.8 + 17.9 = 31.7 mm
or span/380. These two components exceed the values found for the fully continuous
beam (see Table 7.3) by 4.8 and 2.9 mm, respectively.
The imposed-load deflection is about 17 mm, or span/700, for the frequent loading.
The modular ratio used allows for some creep. The additional deflection of the composite
slabs, relative to that of the beams, is discussed in Example 7.1.
The effect of shrinkage of concrete.
Accurate calculation is difficult where semi-rigid joints are used, but estimates can be
made, as follows. The primary shrinkage deformations for the 24 m length of beam
are shown in Fig. 6.27(b). Calculations in Example 6.7 found that R = 1149 m and
= 45.3 mm. The most conservative assumption is that the flexibility of the joints reduces
the secondary shrinkage moment (which reduces deflections) to zero. Fig. 8.15 shows the
same primary curvature as in Fig. 6.27(b), but with compatibility restored by rotation of
the joint at B, rather than by secondary bending of the beam. From the geometry of Fig.
8.15, this rotation is found to be 3.8 mrad for each joint. The mid-span shrinkage
deflection is then about 15 mm. The direction shown for is consistent with that in Fig.
6.1(b), and is not important, as these angles of slope are very small.
The preceding calculation for the composite joint found M
j, Ek
= 70 kN m, and used
S
j
= 38.6/2 = 19.3 kN m/mrad, giving a rotation of 70/19.3 = 3.6 mrad. Thus, shrinkage
imposes a significant increase of rotation on each joint, and the resulting increase of
hogging moment at B, while far less than the 120 kN m found for the fully continuous
beam, will decrease the 15 mm deflection found above. The result for the fully
continuous beam was 7 mm (see Table 7.3). It is concluded that the shrinkage deflection
lies in the range 1012 mm, additional to the 31.7 mm found above.
Cracking of concrete
Elastic analysis of the composite frame for service (frequent) loading acting on both
spans, otherwise similar to those outlined above, found the hogging bending moment at B
to be 133 kN m. The lever arm for the reinforcement, of area 1206 mm
2
, is 543 mm (Fig.
8.8(a)), so the tensile stress is
s
= 133/(0.543 1.206) = 203 N/mm
2
158
1800
10 200
R
C
d
B
3.8 mrad
15 mm
Fig. 8.15. Primary shrinkage deformation of span BC
170
12 May 2004 11:53:10
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:18
The mean spacing of the 16 mm bars is about 250 mm (see Fig. 8.8(d)), which happens to
be the limiting value for a crack width of 0.3 mm given in Table 7.2 of EN 1994-1-1. The
alternative condition in Table 7.1 is satisfied by a wide margin.
However, the strain field in the slab is disturbed locally by the column and by the
concentrated rotations associated with the joints. The values in Tables 7.1 and 7.2 take no
account of this situation. It can be concluded that the top reinforcement is unlikely to
yield in service, and that very wide cracks will not occur.
159
171
12 May 2004 11:53:10
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
CHAPTER 9
Composite slabs with profiled
steel sheeting for buildings
General Clause 9.1
Detailing provisions Clause 9.2
Actions and action effects Clause 9.3
Analysis for internal forces and moments Clause 9.4
Verification of profiled steel sheeting as shuttering for ultimate limit states Clause 9.5
Verification of profiled steel sheeting as shuttering for serviceability limit
states Clause 9.6
Verification of composite slabs for ultimate limit states Clause 9.7
Verification of composite slabs for serviceability limit states Clause 9.8
9.1. General
Scope
Clause 9.1.1 The form of construction and the scope of Section 9 are defined in clause 9.1.1. The shape of
the steel profile, with ribs running in one direction, and its action as tensile reinforcement for
the finished floor, result in a system that effectively spans in one direction only. The slab can
also act as the concrete flange of a composite beam spanning in any direction relative to that
of the ribs. Provision is made for this in the clauses on design of beams in Sections 5, 6 and 7.
Clause 9.1.1(2)P The ratio of the gap between webs to the web spacing, b
r
/b
s
in clause 9.1.1(2)P, is an
important property of a composite slab. This notation is as in Fig. 9.2 and Fig. A.1 in
Appendix A. If the troughs are too narrow, the shear strength of stud connectors placed
within them is reduced (clause 6.6.4), and there may be insufficient resistance to vertical
shear. If the web spacing is too wide, the ability of the slab to spread loads across several webs
may be inadequate, especially if the thickness of the slab above the sheeting is minimized, to
save weight.
Such a wide range of profiles is in use that it was necessary to permit the limit to b
r
/b
s
to be
determined nationally. It should probably be a function of the thickness of the slab above the
sheeting. As a guide, it should normally be less than about 0.6.
No account is taken of any contribution fromthe top flange of the sheeting to resistance to
transverse bending.
The design methods for composite slabs given in Section 9 are based on test procedures
described in clause B.3. Although the initial loading is cyclic, the test to failure is under static
loading. Thus, if dynamic effects are expected, the detailed design for the particular project
173
12 May 2004 11:53:11
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
Clause 9.1.1(3)P
Clause 9.1.1(4)P
must ensure that the integrity of the composite action is maintained (clauses 9.1.1(3)P and
9.1.1(4)P).
Clause 9.1.1(5) Guidance on the degree of lateral restraint provided to steel beams (clause 9.1.1(5)) is
available in EN 1993-1-1 and elsewhere.
106
Inverted U-frame action relies also on flexural
restraint. This subject is covered in comments on clause 6.4.2.
Because of the wide range of profiles used, resistance to longitudinal shear has always
been based on tests. Slabs made with some profiles have a brittle mode of failure, which is
penalized in clause B.3.5(1).
Types of shear connection
Clause 9.1.2.1 As for other types of composite member, bond is not accepted in clause 9.1.2.1 as a reliable
method of shear connection. Sheeting without local deformations of profile is permitted
where the profile is such that some lateral pressure will arise from shrinkage of the concrete
(Fig. 9.1(b)). Here, the distinction between frictional interlock and bond is, in effect, that
the former is what remains after the 5000 cycles of loading specified in clause B.3.4.
The quality of mechanical interlock is sensitive to the height or depth of the small local
deformations of the sheeting, so tight tolerances (clause B.3.3(2)) should be maintained on
these during manufacture, with occasional checking on site.
Clause 9.1.2.2
These two standard forms of interlock are sometimes insufficient to provide full shear
connection, as defined in clause 9.1.2.2. They can be augmented by anchorages at the ends of
each sheet, as shown in Fig. 9.1, or design can be based on partial shear connection.
9.2. Detailing provisions
Clause 9.2.1.(1)P
Clause 9.2.1(2)P
The limits to thickness given in clauses 9.2.1(1)P and 9.2.1(2)P are based on satisfactory
experience of floors with these dimensions. No limits are given for the depth of the profiled
sheeting. Its minimum depth will be governed by deflection. For a slab acting compositely
with a beam, the minimum depths are increased (clause 9.2.1(2)P) to suit the detailing rules
for stud connectors, such as the length of stud that extends above the sheeting and the
concrete cover. A slab used as a diaphragm is treated similarly.
Clause 9.2.1(4)
Where a slab spans onto a hogging moment region of a composite beam, the minimum
reinforcement transverse to its span is governed by the rules for the flange of the beam (e.g.
Table 7.1), not by the lower amount given in clause 9.2.1(4).
Clause 9.2.3 The minimum bearing lengths (clause 9.2.3) are based on accepted good practice. The
lengths for bearing onto steel or concrete are identical to those given in BS 5950: Part 4.
107
9.3. Actions and action effects
Profiled sheeting
Clause 9.3.1(2)P Where props are used for profiled sheeting (clause 9.3.1(2)P), care should be taken to set
these at the correct level, taking account of any expected deflection of the surface that
supports them. If verification relies on the redistribution of moments in the sheeting due to
local buckling or yielding, this must be allowed for in the subsequent check on deflection of
the completed floor; but this is, of course, less likely to be critical where propping is used.
Clause 9.3.2(1) For the loading on the profiled sheeting, clause 9.3.2(1) refers to clause 4.11 of
EN 1991-1-6.
108
For working personnel and small site equipment, a note to clause 4.11.1(3)
proposes a characteristic distributed load of 1 kN/m
2
. Further guidance may be given in the
National Annex.
For the weight density of normal-weight concrete, Annex A of EN 1991-1-1
9
recommends
24 kN/m
3
, increased by 1 kN/m
3
for normal reinforcement and by another 1 kN/m
3
for
unhardened concrete. In addition to self-weight, clause 4.11 of EN 1991-1-6 specifies an
imposed load of 10% of the weight of the concrete, but not less than 0.75 kN/m
2
(which
162
174
12 May 2004 11:53:11
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
usually governs), applied to a working area of 3 3 m, and 0.75 kN/m
2
outside this area.
This corresponds to a layer of normal-weight concrete about 35 mm thick, to allow for the
mounding that occurs during delivery of fresh concrete. Guidance on the avoidance of
overload during construction is available elsewhere.
109
Partial factors for ultimate limit states are recommended in Table A1.2(B) of EN 1990, as
1.35 for permanent actions and 1.5 for variable actions. It would be reasonable to use 1.35 for
the whole of the weight density of 26 kN/m
3
, explained above, even though the extra 1 kN/m
3
for unhardened concrete is not strictly permanent.
Sometimes, to increase the speed of construction, the profiled sheeting is not propped. It
then carries all these loads. This condition, or the check on the deflection of the finished
floor, normally governs its design.
For the serviceability limit state, the deflection of the sheeting when the concrete hardens
is important, for use when checking the total deflection of the floor in service. The
construction load and the extra loading from mounding are not present at this time, so the
deflection is from permanent load only, and the factors for serviceability, given in Table
A.1 of EN 1991-1-6, are not required.
Clause 9.3.2(2) Clause 9.3.2(1) refers to ponding, and clause 9.3.2(2) gives a condition for its effects to be
ignored. Where profiled sheeting is continuous over several supports, this check should be
made using the most critical arrangement of imposed load.
Composite slab
Clause 9.3.3(2)
The resistances of composite slabs are determined by plastic theory or by empirical factors
based on tests in which all of the loading is resisted by the composite section (clause
B.3.3(6)). This permits design checks for the ultimate limit state to be made under the whole
of the loading (clause 9.3.3(2)).
9.4. Analysis for internal forces and moments
Profiled steel sheeting
Clause 9.4.1(1)
Clause 9.4.1(2)
Clause 9.4.1(1) refers to EN 1993-1-3,
25
which gives no guidance on global analysis of
continuous members of light-gauge steel. Clause 9.4.1(2) rules out plastic redistribution
where propping is used, but not where the sheeting extends over more than one span, as
is usual. Subsequent flexure over a permanent support will be in the same direction
(hogging) as during construction, whereas at the location of a prop it will be in the opposite
direction.
Elastic global analysis can be used, because a safe lower bound to the ultimate resistance is
obtained. Elastic moments calculated for uniform stiffness are normally greatest at internal
supports, as shown in Fig. 9.1 for a two-span slab under distributed loading. The reduction in
stiffness due to parts of the cross-section yielding in compression will be greatest in these
regions, which will cause redistribution of moment from the supports to mid-span. In a
technical note from 1984,
110
and in a note to clause 5.2 of BS 5950-4,
107
the redistribution is
given as between 5 and 15%. This suggests that redistribution exceeding about 10% should
not be used in absence of supporting evidence from tests.
163
CHAPTER 9. COMPOSITE SLABS WITH PROFILED STEEL SHEETING FOR BUILDINGS
0.125wL
2
0.070wL
2
L L
Fig. 9.1. Bending moments for a two-span beam or slab for uniform loading; elastic theory without
redistribution
175
12 May 2004 11:53:11
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
Composite slab
Clause 9.4.2(3)
As the steel sheets are normally continuous over more than one span, and the concrete is cast
over this length without joints, the composite slab is in reality continuous. If elastic global
analysis is used based on the uncracked stiffness, the resulting moments at internal supports
are high, as in the example in Fig. 9.1. To resist these moments may require heavy
reinforcement. This can be avoided by designing the slab as a series of simply-supported
spans (clause 9.4.2(5)), provided that crack-width control is not a problem. Other approaches
that reduce the quantity of hogging reinforcement needed are the use of redistribution of
moments (clause 9.4.2(3)), and of plastic analysis (clause 9.4.2(4)).
Clause 9.4.2(4) Numerical and experimental research on continuous slabs has been reported.
111
With
typical relative values of moment resistance at internal supports and at mid-span, the
maximum design loads calculated by elastic analysis with limited redistribution were found
to be less than those obtained by treating each span as simply supported. This arises because
the large resistance to sagging moment is not fully utilized.
If the slab is to be treated as continuous, plastic analysis is more advantageous. The studies
showed that no check on rotation capacity need be made provided the conditions given in
clause 9.4.2(4) are satisfied.
Effective width for concentrated point and line loads
Clause 9.4.3
The ability of composite slabs to carry masonry walls or other heavy local loads is limited.
The rules of clause 9.4.3 for the effective widths b
m
, b
em
and b
ev
are important in practice.
They are based on a mixture of simplified analysis, test data and experience,
107
and are
further discussed, with a worked example, in Johnson.
81
The effective width depends on the
ratio between the longitudinal and transverse flexural stiffnesses of the slab. The nature of
these slabs results in effective widths narrower than those given in BS 8110
17
for solid
reinforced concrete slabs.
Clause 9.4.3(5) The nominal transverse reinforcement given in clause 9.4.3(5) is not generous for a point
load of 7.5 kN, and should not be assumed to apply for the largely repetitive loads to which
clause 9.1.1(3)P refers.
9.59.6. Verification of profiled steel sheeting as shuttering
Clause 9.5(1) The design checks before composite action is established are done to EN 1993-1-3. Clause
9.5(1) refers to the loss of effective cross-section that may be caused by deep deformations of
the sheeting. This loss and the effects of local buckling are both difficult to determine
theoretically. Design recommendations provided by manufacturers are based in part on the
results of loading tests on the sheeting concerned.
Clause 9.6(2) The maximum deflection of L/180 given in the note to clause 9.6(2) is accepted good
practice. Deflection from ponding of wet concrete is covered in clause 9.3.2(2).
9.7. Verification of composite slabs for the ultimate limit
states
9.7.1. Design criterion
No comment is needed.
9.7.2. Flexure
Clause 9.7.2(3) The rules in clause 9.7.2 are based on research reported in Stark and Brekelmans.
111
Clause
9.7.2(3) says that deformed areas of sheeting should be ignored in calculations of section
properties, unless tests showotherwise. No guidance on relevant testing is given. Test results
are also influenced by local buckling within the flat parts of the steel profile, and by the
enhanced yield strength at cold-formed corners.
164
176
12 May 2004 11:53:11
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
For a composite slab in sagging bending, tests can be done in which the shear span is long
enough, or the end anchorage is sufficient, for flexural failure to occur. If the strengths of the
materials are known, the effective area of the sheeting, when in tension, can be calculated
from the moment resisted, and may be close to the gross area of the sheeting.
For a composite slab in hogging bending, the contribution from the sheeting is usually
ignored, because it may not be continuous. Where it is continuous, the area of tensile
reinforcement is usually small compared with the effective area of the sheeting, so that a
conservative estimate of the latter (e.g. excluding embossed areas) may reduce only slightly
the calculated resistance to bending. Alternatively, a value found from a bending test on the
sheeting alone could be used.
Clause 9.7.2(4) The effective widths in clause 9.7.2(4) for local buckling take account of the restraint
provided to one side of the sheeting by the concrete.
Clause 9.7.2(5) Bending resistances of composite slabs are based on rectangular stress blocks (clauses
9.7.2(5) to 9.7.2(7)). In design of reinforced concrete beams, the compressive strain in
concrete is limited, to prevent premature crushing of the concrete before the reinforcement
yields. There is no similar restriction for composite slabs. The design yield strength of the
profiled sheeting, typically between 280 and 420 N/mm
2
(lower than that of reinforcement),
and its own bending resistance make composite slabs less sensitive to premature crushing of
concrete. However, it could be a problem where stronger sheeting is used.
Clause 9.7.2(6)
For stress in concrete, the 0.85 factor is included, as discussed in comments on clause
3.1(1). When the neutral axis is within the sheeting, theory based on stress blocks as in Fig.
9.6 becomes very complex for some profiles so simplified equations are given in clause
9.7.2(6). Their derivation is on record.
111
As shown in Fig. 9.6, the bending resistance is
M
Rd
= M
pr
+ N
c, f
z (D9.1)
where z and M
pr
are given by equations (9.5) and (9.6).
The concrete in compression within the trough is neglected. When
N
c, f
= A
pe
f
yp, d
and the sheeting is entirely in tension, the neutral axis is at its top edge. Equation (9.5) then
correctly gives the lever arm as
z = h h
c
/2 e.
9.7.3. Longitudinal shear for slabs without end anchorage
Clause 9.7.3(2)
Design of composite slabs for longitudinal shear is based on the results of tests. The
specification for these involves a compromise between exploring interactions between the
many relevant parameters and limiting the cost of testing to a level such that the use of new
profiles is not prevented. The tests on composite slabs are defined in clause B.3, on which
comments are given in Chapter 11 of this guide. The tests are suitable for finding the design
resistance to longitudinal shear by either of the methods referred to in clause 9.7.3(2).
The mk method, and existing tests
An empirical mk method has long been established. It is difficult to predict the effect of
changes from test conditions using this method, because of the lack of analytical models,
especially for slabs with non-ductile behaviour. A model for ductile behaviour is given in
Appendix B. The mk test is included in EN 1994-1-1 in a modified form, to provide
continuity with earlier practice; but values of m and k, determined in accordance with codes
such as BS 5950: Part 4,
107
cannot be used in design to EN 1994-1-1, as explained in
comments on clause B.3.5. Subject to sufficient test data being available, it may be possible to
convert the former values to Eurocode values.
A study for the European Convention for Constructional Steelwork
112
found many
differences between methods of testing used. Conversion can be difficult, both for the mk
165
177
12 May 2004 11:53:12
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
method and for the partial-connection method, and the range of applicability of the values
found may be uncertain. Further details are given in Examples 11.1 and 11.2.
Clause 9.7.3(4) For the mk method, resistance to longitudinal shear is presented in clause 9.7.3(4) in
terms of vertical shear because of the way in which mand k are defined. Shearbond failure is
characterized by the formation of a major crack in the slab at between one-quarter and
one-third of the span from a support, as shown in Fig. 9.2. It is over the length AC from this
point to the end of the sheeting that significant slip occurs between the concrete and the
sheeting, so this is the length over which the resistance to longitudinal shear is mobilized.
Clause 9.7.3(5)
The location of point C is unknown, and the finite widths of both the applied load and the
support are further complications. The definition of shear span L
s
for use with the mk
method is treated in clause 9.7.3(5), which gives values for common load arrangements. For
the two-point loading in Fig. 9.2, it is the length BD. For the partial-interaction method,
calculation of the mean shear strength
u
is based on the length L
s
+ L
0
.
Clause 9.7.3(6)
In clauses 9.7.3(3) and 9.7.3(5), L is the span of a simply-supported test specimen, which
differs fromits use in clause 9.7.3(6). In clause 9.7.3(6), the isostatic span is the approximate
length between points of contraflexure in a continuous span of length L between supports.
The partial connection method
To avoid the risk of sudden failure, profiled sheeting should have ductile behaviour in
longitudinal shear. For plain sheeting, the ultimate shear resistance is not significantly
greater than that for initial slip. Such behaviour is not ductile to clause 9.7.3(3), and is
referred to as brittle in clause B.3.5(1). The partial shear connection method is not
applicable to slabs with this behaviour (clause 9.7.3(2)). The mk method may still be used,
but with an additional partial safety factor of 1.25, expressed by the reduction factor 0.8 in
clause B.3.5(1).
For profiles with deformations, the expected ultimate behaviour involves a combination
of friction and mechanical interlock after initial slip, giving a relationship between load and
deflection that should satisfy the definition of ductile in clause 9.7.3(3).
Clause 9.7.3(7) Design data from test results should be found by the partial connection method of clauses
9.7.3(7) to 9.7.3(10). The analytical model, now given, is similar to that used for composite
beams, clause 6.2.1.3, and has been verified for slabs by full-scale tests.
113,114
It is used in
Example 11.2.
For an assumed flexural failure at a cross-section at a distance L
x
fromthe nearest support,
the compressive force N
c
in a slab of breadth b is assumed to be given (equation (9.8)) by
N
c
=
u, Rd
bL
x
(D9.2)
where the design shear strength
u, Rd
is found by testing, to clause B.3. Its derivation takes
account of the difference between the perimeter of a cross-section of sheeting and its overall
breadth, b. By its definition in equation (9.8), force N
c
cannot exceed the force for full
interaction, N
c, f
. Hence, there are two neutral axes, one of which is within the steel profile.
In clause B.3.6(2) the degree of shear connection is defined as
166
L/4 L/4 L/2
A
B C D
L
0
L
s
Fig. 9.2. Shear spans for a composite slab with two-point loading
178
12 May 2004 11:53:12
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
= N
c
/N
c, f
(D9.3)
The longitudinal forces that determine the partial-interaction bending resistance M
Rd
are
all known, for given , but calculation of the resistance moment by the method used for
partial interaction in composite beams is difficult. The line of action of the longitudinal force
in the sheeting (shown in Fig. 11.5) depends on its complex geometry.
Clause 9.7.3(8) A simplified method is given in clause 9.7.3(8). It consists of using equation (9.5) of clause
9.7.2(6), with N
c, f
replaced by N
c, f
and 0.5h
c
replaced by 0.5x
pl
, to determine the lever armz:
z = h 0.5x
pl
e
p
+ (e
p
e)N
c, f
/A
pe
f
yp, d
(9.9)
The reason for these changes is that where x
pl
is much less than h
c
, the method gives too low a
value for M
Rd
, because equation (9.5) assumes that the line of action of the force N
c
in the slab is at
depth h
c
/2. The correct value is x
pl
/2, where x
pl
is the full-interaction value as shown in Fig. 9.5.
It is not clear in EN 1994-1-1 whether the symbol x
pl
in equation (9.9) means the
full-interaction value, or the reduced value, which in this guide is written x
pl
. The reduced
value corresponds to the model used, gives the higher value for z, and is recommended.
The value of the last two terms in equation (9.9) increases from e
p
to e as is increased
from 0 to 1. For profiled sheeting, the plastic neutral axis is usually above the centroidal axis
(i.e. e
p
> e), and (e
p
e) z. It can then be assumed, for simplicity, that the force N
c
in the
sheeting acts at height e
p
above its bottom fibre, giving a lever arm that is correct for = 0,
and slightly too low for < 1.
With these changes, equation (9.9) becomes
z = h 0.5x
pl
e
p
(D9.4)
The bending resistance M
Rd
of the slab, not stated in clause 9.7.3, can be deduced fromFig.
9.6 (and Fig. 11.5 in this guide). It is
M
Rd
= M
pr
+ N
c
z (D9.5)
Calculations for a range of values of L
x
thus give the curve relating resistance M
Rd
to the
distance to the nearest support. The design is satisfactory for longitudinal shear if the
corresponding curve for M
Ed
lies entirely within the one for M
Rd
, as shown in Example 9.1 and
Fig. 9.12.
Clause 9.7.3(9)
In tests, the resistance to longitudinal shear is increased by the friction associated with the
reaction at the adjacent end support. If this is allowed for when calculating
u, Rd
, a lower
value is obtained. Clause 9.7.3(9) provides compensation by using the same effect in the
structure being designed, R
Ed
, to contribute to the shear resistance required. The value
recommended for the coefficient of friction is based on tests.
Additional reinforcement
Clause 9.7.3(10)
Reinforcing bars may be provided in the troughs of the profiled sheeting, and this
reinforcement may be taken into account when calculating the resistance of the slab by the
partial connection method, clause 9.7.3(10).
The analytical model assumes that the total resistance is that fromcomposite action of the
concrete with both the sheeting and the bars, as for reinforced concrete, determined by
plastic analysis of the cross-section. The value of
u, Rd
is obtained, as before, by testing of
specimens without additional reinforcement, clause B.3.2(7).
If the mk method were to be used, reinforcement in troughs would be an additional
variable, which would require a separate test series (clause B.3.1(3)), with evaluation based
on measured strength of the reinforcement.
9.7.4. Longitudinal shear for slabs with end anchorage
Clause 9.7.4 Clause 9.7.4 refers to the two types of end anchorage defined in clause 9.1.2.1. The anchorage
provided by studs is ductile. It is preferable to that provided by deformed ribs of re-entrant
profiles, where poor compaction of concrete may occur.
167
179
12 May 2004 11:53:12
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
Clause 9.7.4(2)
Clause 9.7.4(3)
In the partial-connection method, the anchorage force is used as a contribution to the
total force N
c
(clause 9.7.4(2)). For through-deck-welded studs it can be calculated, to clause
9.7.4(3). The model is based on the weld collar pulling through the end of the sheeting.
Further comments are given in clause 6.6.6.4. It is not clear how a contribution from
deformed ribs should be determined, as end anchorage is not used in the test specimens
(clause B.3.2(7)).
For the mk method, end anchorage of either type may be included in the test specimens,
but each type is an additional variable (clause B.3.1(3)), and would require a separate test
series.
9.7.5. Vertical shear
Clause 9.7.5 Clause 9.7.5 refers to EN 1992-1-1, where resistance to vertical shear depends on the
effective depth d of the section. In a composite slab, where the sheeting is the reinforcement,
d is the distance d
p
to the centroid of the profile, shown in Fig. 9.6.
9.7.6. Punching shear
Clause 9.7.6(1) The critical perimeter for punching shear (clause 9.7.6(1)) has rounded corners, as does
that used in EN 1992-1-1. It is therefore shorter than the rectangular perimeter used in
BS 5950: Part 4. It is based on dispersion at 45 to the centroidal axis of the sheeting in the
direction parallel to the ribs, but only to the top of the sheeting in the less stiff transverse
direction.
Clause 6.4.4 of EN1992-1-1 gives the shear resistance as a stress, so one needs to knowthe
depth of slab on which this stress is assumed to act. For a concrete slab this would be the
appropriate effective depth in each direction. For a composite slab, no guidance is given.
BS 5950-4
107
takes the effective depth in both directions as that of the concrete above the top
of the profiled steel sheeting, to be used in conjunction with BS 8110.
17
9.8. Verification of composite slabs for serviceability limit
states
9.8.1. Cracking of concrete
Clause 9.8.1(1) Clause 9.8.1(1) refers to EN 1992-1-1, where, from clause 7.3.1(4),
Cracks may be permitted to form without any attempt to control their width, provided that they
do not impair the functioning of the structure.
Clause 9.8.1(2)
For this situation, the provisions for minimum reinforcement above internal supports of
composite slabs (clause 9.8.1(2)) are the same as those for beams in clause 7.4.1(4), where
further comment is given. Crack widths should always be controlled above supports of slabs
subjected to travelling loads. The methods of EN 1992-1-1 are applicable, neglecting the
presence of the sheeting.
9.8.2. Deflection
Clause 9.8.2(1) Clause 9.8.2(1) refers to EN 1990, which lists basic criteria for the verification of
deformations.
Clause 9.8.2(2) For the construction phase, clause 9.8.2(2) refers to EN 1993-1-3, where clause 7.3(2) says
that elastic theory should be used, with the characteristic load combination (clause 7.1(3)).
This corresponds to an irreversible limit state, which is appropriate for the deflection of
sheeting due to the weight of the finished slab. Where it can be assumed that when the slab
hardens, the imposed load on it is negligible, the remaining deflection of the sheeting due to
construction loads should also be negligible; but this assumption, if made, should be based
on relevant experience.
The prediction of construction loads may also be difficult. If in doubt, the deflection of the
sheeting at this time should be determined for the characteristic combination, even though
the less adverse frequent combination is permitted for reversible deflections.
168
180
12 May 2004 11:53:13
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
Clause 7.3.(4) of EN 1993-1-3 refers to a clause in EN 1993-1-1 which says that limits to
deflection of sheeting should be agreed, and may be given in a National Annex. Further
comment is given under clause 9.3.2, and in Examples 7.1 and 9.1.
Clause 9.8.2(3) For the composite member, clause 9.8.2(3) refers to Section 5 for global analysis, where the
comments made on continuous beams apply. Restraint of bending of slabs fromthe torsional
stiffness of supporting members is usually ignored, but situations arise where it could cause
cracking.
Clause 9.8.2(4) The rules in clause 9.8.2(4) permit calculation of deflections to be omitted if two
conditions are satisfied. The first refers to limits to the ratio of span to effective depth given
in EN 1992-1-1. These are 20 for a simply-supported slab, 26 for an external span of a
continuous slab and 30 for an internal span. The effective depth should be that given in the
comment on clause 9.7.5.
Clause 9.8.2(6) The second condition, clause 9.8.2(6), applies to external spans only, and relates to the
initial slip load found in tests information which may not be available to the designer.
Two-point loading is used for testing, so the design service load must be converted to a
point load. This can be done by assuming that each point load equals the end reaction for a
simply-supported span under service loading.
Clause 9.8.2(7) Where the initial slip load in the tests is below the limit given in clause 9.8.2(6), clause
9.8.2(7) provides a choice. Either the slip should be allowed for, which means estimating the
deflection from the test results, or end anchors should be provided. There is no guidance on
how much end anchorage is required. Its effectiveness would have to be found by testing and
analysis by the mk method, because clause B.3.2(7) does not permit end anchors to be used
in tests for the partial-connection method.
Tied-arch model
Clause 9.8.2(8) The situation that is covered by clause 9.8.2(8) seems likely to arise only where:
a high proportion of the shear connection is provided by welded studs
its amount is established by calculation to clause 9.7.4
no test data are available.
If the end anchorage is provided by deformed ribs, then experimental verification is
essential, and the slip behaviour will be known.
In the tied-arch model proposed, the whole of the shear connection is provided by end
anchorage. Accurate calculation of deflection is difficult, but the arch will be so shallow that
the following simplified method is quite accurate.
The tie shown in Fig. 9.3(a) consists of the effective area of the steel sheeting. Line AB
denotes its centroid. The thickness h
c
of the arch rib must be such that at the relevant
bending moment its compressive stress at mid-span is not excessive. There is interaction
between its assumed thickness h
c
and the lever arm h
a
, so a little trial-and-error is needed.
Longitudinal forces at mid-span are now known, and the strains
c
and
t
in the concrete and
steel members can be found, taking appropriate account of creep. In Fig. 9.3(b), curve CB
represents the arch rib. The ratio L/h
a
is unlikely to be less than 20, so both the curve length
and the chord length CB can be taken as L/2, and
cosec cot = L/2h
a
As shown in the figure, the changes in length of the members are
e
t
=
t
L/2
and
e
c
=
c
L/2
so the deflection is given by
169
181
12 May 2004 11:53:13
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
/L = (
t
+
c
)(L/4h
a
) (D9.6)
with both strains taken as positive numbers.
Example 9.1: two-span continuous composite slab
Data
Details of the geometry assumed for this composite slab are shown in Fig. 9.4(a). The
properties of the concrete and reinforcement are as used in the worked example on a
two-span beam in Chapters 6 and 7, where further information on them is given. The
spacing of the supporting composite beams is different, 3.0 m rather than 2.5 m. Their
top-flange width is 190 mm.
The cross-section assumed for the sheeting satisfies the condition of clause 9.1.1 for
narrowly spaced webs. It provides shear connection by embossments, in accordance with
clause 9.1.2.1(a). Its dimensions h
c
and h
p
in Fig. 9.2 are defined in clause 1.6. From Fig.
9.4(a), they are h
c
= 75 mm (concrete above the main flat surface of the sheeting) and
h
p
= 70 mm (overall depth of the sheeting). Thus, for profiles of this type, the thickness
of the slab, 130 mm here, is less than h
c
+ h
p
. For flexure of the composite slab, h
c
is the
appropriate thickness, but for flexure of the composite beam supporting the sheeting, or
for in-plane shear in the slab, the relevant thickness is, for this sheeting, h
c
15 mm. Some
of the design data used here have been taken fromthe relevant manufacturers brochure.
For simplicity, it is assumed that the reinforcement above the supporting beams, which
affects the properties of the composite slab in hogging bending, is not less than that
determined in Example 6.7 on the two-span beam, as follows:
in regions where the beam resists hogging bending, as determined by resistance to

distortional lateral buckling: A
s
= 565 mm
2
/m and 12 mm bars at 200 mm spacing
in other regions, as determined by resistance to longitudinal shear: A

s
213 mm
2
/m.
It is assumed that A252 mesh is provided (A
s
= 252 mm
2
/m), resting on the sheeting. This
has 8 mm bars at 200 mm spacing, both ways.
These details satisfy all the requirements of clause 9.2.1.
At the outset, two assumptions have to be made that affect the verification of a
composite slab:
whether construction will be unpropped or propped
whether the spans are modelled as simply supported or continuous.

Here, it will be assumed that unpropped construction will be used wherever possible. The
sheeting is provided in 6 m lengths, so no 3 m span can have sheeting continuous at both
ends. For a building 9 m wide, there will also be spans where 3 m lengths of sheeting are
used, so several end-of-span conditions will be considered.
170
C
A B
h
c
h
a
e
t
cot a
e
t
cosec a
C
A
a
d
e
c
e
t
L/2
B
(b) (a)
Fig. 9.3. Tied-arch model for deflection of a composite slab with an end anchorage
182
12 May 2004 11:53:13
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
Some of the design checks on a composite slab are usually satisfied by wide margins.
These checks will be simplified here by making conservative assumptions.
Properties of materials and profiled sheeting
Lightweight-aggregate concrete: grade LC25/28; 1800 kg/m
3
; 0.85f
cd
= 14.2 N/mm
2
;
E
lctm
= 20.7 kN/mm
2
; n
0
= 10.1; f
lctm
= 2.32 N/mm
2
.
Creep is allowed for by using n = 20.2 for all loading (clause 5.4.2.2(11)).
Reinforcement: f
sk
= 500 N/mm
2
, f
sd
= 435 N/mm
2
.
Profiled sheeting: nominal thickness including zinc coating, 0.9 mm; bare-metal
thickness, 0.86 mm; area, A
p
= 1178 mm
2
/m; weight, 0.10 kN/m
2
; second moment of area,
10
6
I
y, p
= 0.548 mm
4
/m; plastic neutral axis, e
p
= 33 mm above bottom of section (see Fig.
9.6); centre of area, e = 30.3 mm above bottom of section; E
a
= 210 kN/mm
2
; yield
strength, f
yk, p
= 350 N/mm
2
,
M, p
= 1.0; plastic moment of resistance in hogging and
sagging bending, M
pa
= 6.18 kN m/m this value is assumed to take account of the effect
of embossments (clause 9.5(1)).
Composite slab: 130 mm thick; volume of concrete, 0.105 m
3
/m
2
.
Loading for profiled sheeting
From clause 11.3 of EN 1992-1-1, the design density of the reinforced concrete is
1950 kg/m
3
, which is assumed to include the sheeting. The dead weight of the floor is
g
k1
= 1.95 9.81 0.105 = 2.01 kN/m
2
From Note 2 to clause 4.11.1(7) of EN 1991-1-6:
108
for fresh concrete the weight density
should be increased by 1 kN/m
3
, so for initial loading on the sheeting,
g
k1
= 2.01 20.5/19.5 = 2.11 kN/m
2
171
30.3
412 kN
0.85f
cd
29
(a) (b)
1178
x
51
49
100

e = 30
(c)
48
27
28
27
51.4
x
27
162
252
(d)
200
48 60
55
15
26
26
164 136
A252 mesh
0.9
(nom.)
Fig. 9.4. Composite slab. (a) Dimensions. (b) M
pl, Rd
, sagging. (c) Second moment of area, sagging.
(d) Second moment of area, hogging
183
12 May 2004 11:53:14
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
For construction loading, clause 9.3.2(1) refers to clause 4.11.2 of EN 1991-1-6. Clause
4.11.2(1) of EN 1991-1-6 gives the actions from personnel and equipment, q
k
, as 10% of
the self weight of the concrete, but not less than 0.75 and not more than 1.5 kN/m
2
.
EN 12812
115
may also be relevant.
The mounding of concrete during placing is referred to in clause 9.3.2(1), but not in
EN 1991-1-6, although its clause 4.11.1(2) refers to storage of materials. Here, to
demonstrate the method of calculation, q
k
will be taken as 1.0 kN/m
2
. The National Annex
may specify a different value.
Loading for composite slab
Floor finish (permanent): g
k2
= 0.48 kN/m
2
.
Variable load (including partitions, services, etc.): q
k
= 7.0 kN/m
2
.
The loads per unit area are summarized in Table 9.1, using
F, g
= 1.35 and
F, q
= 1.5.
Verification of sheeting
Ultimate limit state
Fromclause 9.2.3(2), the minimumwidth of bearing of the sheeting on a steel top flange is
50 mm. Assuming an effective support at the centre of this width, and a 190 mm steel
flange, the effective length of a simply-supported span is
3.0 0.19 + 0.05 = 2.86 m
Hence,
M
Ed
= 4.35 2.86
2
/8 = 4.45 kN m/m
This is only 72% of M
pl, Rd
, so there is no need to consider bending moments in the
continuous slab, or to check rotation capacity to clause 9.4.2.1(1)(b).
Excluding effects of continuity, for vertical shear:
V
Ed
= 4.35 1.43 = 6.22 kN/m
There are 6.7 webs of depth 61 mm per metre width of sheeting, at an angle cos
1
55/61 to
the vertical. In the absence of buckling, their shear resistance is given by EN 1993-1-3 as
(61/55)V
Rd
= 6.7 61 0.9 0.350/3
whence
V
Rd
= 74.3 kN/m
172
Table 9.1. Loadings per unit area of composite slab (kN/m
2
)
Type of load
Characteristic,
maximum
Characteristic,
minimum
Ultimate,
maximum
Ultimate,
minimum
During concreting
Sheeting and concrete 2.11 0.10 2.85 0.13
Imposed load 1.00 0 1.50 0
Total 3.11 0.10 4.35 0.13
For composite slab
Slab and floor finish 2.01 + 0.48 2.49 3.36 3.36
Imposed load 7.00 0 10.5 0
Total 9.49 2.49 13.9 3.36
For deflection of
composite slab
(frequent)
0.7 7 + 0.48 = 5.38 0.48 NA NA
184
12 May 2004 11:53:14
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
The slenderness of each web is 61/0.9 = 68, which is close to the limit at which buckling
must be considered; but V
Ed
is so far below V
Rd
, as is usual for the construction stage, that
no calculation is needed.
Deflection
A note to clause 9.6(2) recommends that the deflection should not exceed span/180. The
worst case is where a span is simply-supported. Then,
= 5wL
4
/(384E
a
I
y
) = 5 2.01 2.86
4
1000/(384 210 0.548) = 15.2 mm
For fresh concrete, this is increased to
15.2 20.5/19.5 = 16.0 mm
From clause 9.3.2(1), allowance should be made for ponding if the deflection exceeds
1/10th of the slab thickness (13 mm), as it does here. The specified thickness of the
additional concrete is 0.7. Its weight, for fresh concrete, is
0.7 0.016 20.5 = 0.23 kN/m
2
This increases the deflection to
16 2.34/2.01 = 17.7 mm
which is span/161, and so exceeds span/180.
It follows that where the sheeting is not continuous at either end of a 3 m span,
propping should be used during construction.
The effects of continuity at one end of a span are now considered. The most adverse
condition occurs when the concrete in one span hardens (with no construction load
present) before the other span is cast. The loadings are then 2.01 kN/m on one span and
0.10 kN/m on the other. The spans should be taken as slightly longer than 2.84 m, to allow
for the hogging curvature over the width of the central support. The appropriate length is
(2.86 + 3.0)/2 = 2.93 m
It can be shown by elastic analysis of a continuous beam of uniform section, with
uniformly distributed loading, that the deflection at the centre of a span is
=
0
[1 0.6 (M
1
+ M
2
)/M
0
)]
where the hogging end moments are respectively M
1
and M
2
, and
0
and M
0
are the
deflection and mid-span moment of the span when the end moments are zero.
From elastic analysis for a span of 2.93 m, M
0
= 2.16 kN m, M
1
= 1.13 kN m, M
2
= 0
and
0
= 16.7 mm. Hence,
= 16.7[1 0.6 (1.13/2.16)] = 11.5 mm
When the other span is cast, this deflection is reduced; but if the concrete in the first span
has already hardened, the reduction is small. The deflection is less than span/180.
Properties of the composite slab
Plastic resistance moment
For sagging bending, the plastic neutral axis is likely to be above the sheeting, so clause
9.7.2(5) applies. The tensile force in the sheeting, when at yield, is
F
y, p
= A
p
f
yp, d
= 1178 0.35 = 412 kN/m (D9.7)
The depth of slab in compression is
412/14.2 = 29 mm (D9.8)
The lever arm is
173
185
12 May 2004 11:53:14
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
130 30.3 29/2 = 85 mm
(Fig. 9.4(b)), so
M
pl, Rd
= 412 0.085 = 35 kN m/m (D9.9)
Second moments of area
For sagging bending, the transformed width of concrete is 1000/20.2 = 49 mm/m. For an
elastic neutral axis at depth x (Fig. 9.4(c)), first moments of area give
1178(100 x) = 49x
2
/2
whence x = 49 mm. Hence
10
6
I
y
= 0.548 + 1178 0.051
2
+ 49 0.49
2
/3= 5.53 mm
4
/m
For hogging bending with sheeting present, each trough in the sheeting is replaced by a
rectangle of width 162 mm. There is one trough per 300 mm, so the transformed width of
concrete is
49 162/300 = 27 mm/m
For a neutral axis at height x above the bottom of the slab (Fig. 9.4(d)),
252(82 x) = 1178(x 30) + 27 x (x/2)
whence x = 30.6 mm. Thus, the neutral axis almost coincides with the centre of area of the
sheeting, and
10
6
I
y
= 0.548 + (252 51.4
2
+ 27 30.6
3
/3) 10
6
= 1.47 mm
4
/m
Verification of the composite slab
Deflection
This is considered first, as it sometimes governs the design. Clause 9.8.2(4) gives
conditions under which a check on deflection may be omitted. It refers to clause 7.4 of
EN 1992-1-1. Table 7.4 in that clause gives the limiting ratio of span to effective depth for
an end span as 26. The depth to the centroid of the sheeting is 100 mm, so the ratio is
2.93/0.1 = 29.3, and the condition is not satisfied.
Deflection is a reversible limit state, for which a note to clause 6.5.3(2) of EN 1990
recommends use of the frequent combination. The
1
factor for this combination depends
on the floor loading category. From Table A1.1 in EN 1990, it ranges from 0.5 to 0.9, and
is here taken as 0.7, the value for shopping or congregation areas. From Table 9.1,
the maximum and minimum loadings are 5.38 and 0.48 kN/m, respectively. For a
simply-supported loaded span of 2.93 m, with I
y
= 5.53 10
6
mm
4
/m, the deflection is
4.4 mm. The region above the central support is likely to be cracked. A more accurate
calculation for the two-span slab, assuming 15% of each span to be cracked, and with
0.48 kN/m on the other span, gives the deflection of the fully loaded span as 3.5 mm.
Hence, the total deflection of the slab, if cast unpropped, is
total
= 11.5 + 3.5 = 15 mm
which is span/195. This value is not the total deflection, as it is relative to the levels of the
supporting beams. In considering whether it is acceptable, account should also be taken of
their deflection (Example 7.1). If it is found to be excessive, propped construction should
be used for the composite slab.
Ultimate limit states: flexure
From Table 9.1, the maximum loading is 13.9 kN/m. For a simply-supported span,
M
Ed
= 13.9 2.93
2
/8 = 14.9 kN m
174
186
12 May 2004 11:53:14
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
This is so far belowthe plastic moment of resistance, 35 kN m/m, that there is no need, for
this check, to consider continuity or the resistance to hogging bending.
Where it is necessary to consider continuity, clause 9.4.2 on global analysis and clause
9.7.2(4) on resistance to hogging bending are applicable.
Longitudinal shear by the mk method
For longitudinal shear, it is assumed that there is no end anchorage, so that clause 9.7.3
is applicable. The shear properties of this sheeting are determined and discussed in
Examples 11.1 and 11.2 and Appendix B.
It is assumed that tests have shown that the sheeting provides ductile shear connection
to clause 9.7.3(3), and that the values for use in the mk method are m = 184 N/mm
2
and
k = 0.0530 N/mm
2
. From clause 9.7.3(4), the design shear resistance is
V
l, Rd
= bd
p
/
VS
[(mA
p
/bL
s
) + k]
where d
p
is the depth to the centroidal axis of the sheeting, 100 mm; A
p
is the
cross-sectional area of breadth b of the sheeting, 1178 mm
2
/m; L
s
is span/4, or 0.73 m,
fromclause 9.7.3(5); and
VS
is the partial safety factor, with a recommended value of 1.25.
These values give: V
l, Rd
= 28.0 kN/m, which must not be exceeded by the vertical shear in
the slab.
For a simply-supported span,
V
Ed
1.5 13.9 = 20.9 kN/m
For the two-span layout, it will be a little higher at the internal support, but clearly will not
exceed 28 kN/m.
Vertical shear
Clause 9.7.5 refers to clause 6.2.2 of EN1992-1-1. This gives a formula for the resistance in
terms of the area of tensile reinforcement, which is required to extend a certain distance
beyond the section considered. The sheeting is unlikely to satisfy this condition at an end
support, but its anchorage has already been confirmed by the check on longitudinal shear.
Treating the sheeting as the reinforcement, the clause gives the resistance to vertical
shear as 49 kN/m, which far exceeds V
Ed
, found above.
Serviceability limit state cracking
Clause 9.8.1(1) is for continuous slabs. This slab can be assumed to satisfy the condition
of clause 9.8.1(2): to have been designed as simply supported in accordance with clause
9.4.2(5). To control cracking above intermediate supports, that clause requires the
provision of reinforcement to clause 9.8.1.
The amount required, for unpropped construction, is 0.2% of the area of concrete on
top of the steel sheet. For this purpose, the mean concrete thickness is relevant. This is
close to 75 mm, so the area required is 150 mm
2
/m, and the A252 mesh used here is
sufficient.
However, if any spans are constructed propped, to reduce deflections, then the required
area is doubled, and A252 mesh is not sufficient.
Longitudinal shear resistance by the partial-connection method
In Example 11.2 it is deduced from tests on slabs with the sheeting used here that its
design shear strength was
u, Rd
= 0.144 N/mm
2
(D9.10)
Other conditions for the use of this result in design to EN 1994-1-1, based on clause
B.3.1(4), are now compared with the data for this example, which are given in
parentheses, with indicating compliance:
175
187
12 May 2004 11:53:14
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
(a) thickness of sheeting 0.9 mm, including coating (t = 0.86 mm, plus coating; )
(b) concrete strength, f
ck
0.8f
cm
= 0.8 29.8 = 23.8 N/mm
2
(C25/30 concrete; )
(c) steel yield strength, f
yp
0.8 f
yp, m
= 0.8 376 = 301 N/mm
2
(f
yp
= 350 N/mm
2
; )
(d) concrete density, measured 1.5 h after mixing: 1944 kg/m
3
(design density
1800 kg/m
3
; ?)
(e) slab thickness, h = 170 mm, as in tests 1 to 4 (h = 130 mm; #).
Clause B.3.1(3) defines concrete density and slab thickness as variables to be
investigated. Section 9 gives no guidance on what allowance should be made, if any, for
differences between test and design values of these variables.
The density, measured at an early age, takes no account of subsequent loss of moisture.
The concrete strength, item (b), is acceptable, and the difference of density, item (d), is
small, so its effects can be ignored.
The difference in slab thickness, item (e), is significant. Its effect is discussed more
generally in Appendix B. The test results on thinner slabs led to a higher value of
u, Rd
, for
reasons explained in Example 11.2. It is assumed that the value for
u, Rd
given in equation
(D9.7) can be used here.
Design partial-interaction diagram
To satisfy clause 9.7.3(7) it is necessary to show that throughout the span (coordinate x)
the curve of design bending moment, M
Ed
(x), nowhere lies above the curve of design
resistance, M
Rd
(x), which is a function of the degree of shear connection, (x). These two
curves are now constructed.
From Table 9.1, the design ultimate load for the slab is 13.9 kN/m
2
, assumed to act on
the composite member. The verification for flexure assumed simply-supported spans of
2.93 m, which gives M
Ed, max
as 14.9 kN m/m, at mid-span. The parabolic bending-moment
diagram is plotted, for a half span, in Fig. 9.5.
From equation (9.8) in clause 9.7.3(8), the compressive force in the slab at distance x m
from an end support (i.e. with L
x
= x) is
N
c
=
u, Rd
bx = 0.144 1000x = 144x kN/m
N
c, f
= A
p
f
yp, d
= 412 kN/m
The length of shear span needed for full interaction is therefore
176
0
20
10
0.5 1.0 1.5
E
B
D
x (m)
M (kN m/m)
M
Ed
A
C
M
Rd
M
pa
= 6.18 1.27M
Ed
Fig. 9.5. Design partial-interaction diagram
188
12 May 2004 11:53:15
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:37
L
s, f
= 412/144 = 2.86 m
This exceeds span/2 (1.47 m), so full interaction is not achieved in a span of this length.
From equation (D9.5), the depth of the full-interaction stress block in the slab is
x
pl
= 29 mm. With partial interaction, this value gives a slightly conservative result for z,
and is used here for simplicity.
With h = 130 mm, e
p
= 33 mm and e = 30.3 mm, as before, equation (9.9) for the lever
arm gives
z = 130 29/2 33 + (33 30.3) 144x/412 = 82 + 1.05x mm
The reduced bending resistance of the composite slab is given by equation (9.6) with N
c, f
replaced by N
c
and M
pa
= 6.18 kN m/m, as before:
M
pr
= 1.25 6.18 (1 144x/412) = 7.72 2.7x 6.18
so
x 1.54/2.7 = 0.570 m
From Fig. 9.6, the plastic resistance is
M
Rd
= N
c
z + M
pr
= 0.144x (82 + 1.05x) + 7.72 2.7x
= 7.72 + 9.11x + 0.151x
2
for 0.57 x 1.47 m
For x < 0.57 m,
M
Rd
= 0.144x (82 + 1.05x) + 6.18 = 6.18 + 11.8x + 0.151x
2
The curve M
Rd
(x) is plotted as AB in Fig. 9.5. It lies above the curve 0C for M
Ed
at all
cross-sections, showing that there is sufficient resistance to longitudinal shear.
This result can be compared with that from the mk method, as follows. Curve 0C is
scaled up until it touches curve AB. The scale factor is found to be 1.27 (curve 0DE), with
contact at x = 1.0 m. Shear failure thus occurs along a length of 1.0 m adjacent to an end
support. The vertical reaction at that support is then
V
Ed
= 1.27 13.9 2.93/2 = 25.9 kN/m
This is 8% lower than the 28 kN/m found by the mk method. It is concluded in Example
11.2 that its result for
u, Rd
is probably too low because the test span was too long. The two
methods therefore give consistent results, for this example. No general comparison of
them is possible, because the partial shear-connection method involves the bending-
moment distribution, while the mk method does not.
The calculations summarized in this example illustrate provisions of EN 1994-1-1 that
are unlikely to be needed for routine design. They are relevant to the preparation of
design charts or tables for composite slabs using sheeting of a particular thickness
and profile. These are normally prepared by specialists working on behalf of the
manufacturer.
177
189
12 May 2004 11:53:15
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:40
CHAPTER 10
Annex A (Informative). Stiffness
of joint components in buildings
This chapter corresponds to Annex A in EN 1994-1-1, which has the following clauses:
Scope Clause A.1
Stiffness coefficients Clause A.2
Deformation of the shear connection Clause A.3
Annex A is needed for the application of clause 8. It is informative because the
component approach to the design of steel and composite joints continues to be developed.
Its content is based on the best available research, much of which is recent. It is informed by
and generally consistent with two reports prepared on behalf of the steel industry.
39,104
A.1. Scope
Clause A.1(1)
Clause A.1(2)
This annex supplements the provisions on stiffness of steel joints in clause 6.3 of
EN 1993-1-8.
24
Its scope is limited (clause A.1(1)). It covers conventional joints in regions
where the longitudinal slab reinforcement is in tension, and the use of steel contact plates in
compression. As in EN 1993-1-8, stiffness coefficients k
i
are determined (clause A.1(2)), with
dimensions of length, such that when multiplied by Youngs elastic modulus for steel, the
result is a conventional stiffness (force per unit extension or compression). For a known
lever arm z between the tensile and compressive forces across the joint, the rotational
stiffness is easily found from these coefficients.
As in EN1993-1-8, stiffnesses of components are combined in the usual way, for example:
k = k
1
+ k
2
(for two components in parallel)
1/k = 1/k
1
+ 1/k
2
(for two components in series)
(D10.1)
Clause A.1(3)
Stiffness coefficients k
1
to k
16
are defined in Table 6.11 of EN 1993-1-8. Those relevant to
composite joints are listed in Section 8.2 of this guide. Of these, only k
1
and k
2
are modified
here, to allow for steel contact plates and for the encasement of a column web in concrete
(clause A.1(3)).
A.2. Stiffness coefficients
Clause A.2.1.1(1)
The background to this clause is available.
116
The coefficients have been calibrated against
test results. For longitudinal reinforcement in tension, clause A.2.1.1(1) gives formulae for
the coefficient k
s, r
in terms of the bending moments M
Ed, 1
and M
Ed, 2
, shown in Fig. A.1 and
191
12 May 2004 11:53:15
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:40
Fig. 8.3. Subscripts 1 and 2 are used here also for the properties of the connections on which
these moments act. The concrete slab is assumed to be fully cracked.
The transformation parameters
1
and
2
allow for the effects of unequal bending
moments applied to the pair of connections on either side of a column. They are given
in clause 5.3(9) of EN 1993-1-8, with simplified values in clause 5.3(8). The latter are
discontinuous functions of M
Ed, 2
/M
Ed, 1
, as shown in Fig. 10.1(a). The sign convention in
EN 1993-1-8 is that both moments are positive when hogging, with M
Ed, 1
M
Ed, 2
. The range
covered by EN 1993-1-8 is
1 M
Ed, 2
/M
Ed, 1
1
However, the stiffness coefficients k
s, r
in Table A.1 are based on reinforcement in tension,
and composite joints with M
Ed, 2
< 0 are outside the scope of EN 1994-1-1.
A typical stiffness coefficient is
k
s, r
= A
s, r
/h (D10.2)
where h is the depth of the steel section of the column, and
1
and
2
(for the connections on
sides 1 and 2 of the column) are functions of
1
and
2
, respectively. Based on both the
precise and the simplified values for the s given in EN1993-1-8, they are plotted against the
moment ratio M
2
/M
1
(using this notation for M
Ed, 2
and M
Ed, 1
) in Fig. 10.1(b). For the
simplified s, they are discontinuous at M
2
/M
1
= 0 and 1.
It is the extension of a length h of reinforcement that is assumed to contribute to the
flexibility of the joint. The figure shows that when M
2
= M
1
, this length for connection 1 is
h/2, increasing to 3.6h when M
2
= 0. This is the value in Table A.1 for a single-sided joint. The
flexibility 1/k
s, r
for connection 2 becomes negative for M
2
< 0.5M
1
.
Clause A.2.1.2 The infinite stiffness of a steel contact plate (clause A.2.1.2) simply means that one term
in an equation of type (D10.1) is zero.
Clause A.2.2.1 In clause A.2.2.1, the stiffness of a column web panel in shear is reduced below the value in
EN 1993-1-8 because the force applied by a contact plate may be more concentrated than
would occur with other types of end-plate connection.
Clause A.2.2.2 Similarly, the stiffness for a web in transverse compression has been reduced in clause
A.2.2.2, where the value 0.2 replaces 0.7 in EN 1993-1-8.
Concrete encasement
Encasement in concrete increases the stiffness of the column web in shear, which is given for
an uncased web in EN 1993-1-8 as
k
1
= 0.38A
vc
/z
180
2
1
1 0 1
M
2
/M
1
M
2
/M
1
(b) (a)
b
1
and b
2
,
b
1
and b
2
,
approx.
precise
For precise b
For approx. b
l
2
5
1 0
0.5
3.6
l
1
l
b
Fig. 10.1. Flexibility of reinforcement. (a) Transformation parameter, . (b) Flexibility of
reinforcement, represented by
1
and
2
, for M
1
M
2
192
12 May 2004 11:53:16
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:40
Clause A.2.3.1(1) The addition to k
1
given in clause A.2.3.1(1) has a similar form:
k
1, c
= 0.06(E
cm
/E
a
)b
c
h
c
/z
where E
cm
/E
a
is the modular ratio and b
c
h
c
is the area of concrete.
Clause A.2.3.2 For the column web in compression (clause A.2.3.2), the relationship with the stiffness of
the steel web is similar to that for the web in shear. The coefficient in the additional stiffness
k
2, c
is 0.5 for an end plate, but reduces to 0.13 for a contact plate, because of the more
concentrated force.
A.3. Deformation of the shear connection
Clause A.3 The background to the rather complex provisions of clause A.3 is given in Appendix 3 of
ECCS TC11
39
and in COST-C1.
117
They are based on linear partial-interaction theory for the
shear connection. Equations (A.6) to (A.8) are derived as equations (7.26) and (7.28) in
Aribert.
118
Equation (A.5) in clause A.3(2) can be rearranged as follows:
1/k
slip
k
s, r
= (K
sc
+ E
s
k
s, r
)/K
sc
k
s, r
= E
s
/K
sc
+ 1/k
s, r
This format shows that the flexibility E
s
/K
sc
has been added to that of the reinforcement,
1/k
s, r
, to give the combined stiffness, k
slip
k
s, r
. It also shows that, unlike the k
i
, K
sc
is a
conventional stiffness, force per unit extension.
Clause A.3(3) The definition of the stiffness of a shear connector in clause A.3(3) assumes that the mean
load per connector will be a little belowthe design strength, P
Rk
/
VS
(typically 0.8P
Rk
). Where
the slab is composite, the tests should ideally be specific push tests to clause B.2.2(3). In
practice, the approximate value given in clause A.3(4) for 19 mm studs, 100 kN/mm, may be
preferred. Its use is limited to slabs in which the reduction factor k
t
(clause 6.6.4.2) is unity. It
may not apply, therefore, to pairs of studs in each trough, as used in Example 6.7.
In fact, this stiffness, the connector modulus, can vary widely. Johnson and Buckby
37
refer
to a range from 60 kN/mm for 16 mm studs to 700 kN/mm for 25 mm square bar connectors;
and an example in Johnson
81
uses 150 kN/mm for 19 mm studs. The value 100 kN/mm is of
the correct magnitude, but designs that are sensitive to its accuracy should be avoided.
Further comments on stiffness
These are found in Chapter 8 and in the following example.
Example 10.1: elastic stiffness of an end-plate joint
The rotational stiffness S
j, ini
, modified to give S
j
, is needed for elastic or elasticplastic
global analysis, both for finding the required rotation of the joint, and for checking
deflection of the beams. The stiffnesses k
i
shown in Fig. 8.3 are now calculated. Formulae
for the steel components are given in Table 6.11 of EN1993-1-8. Dimensions are shown in
Figs 8.8 and 8.9.
Column web in shear
From Table 6.11 of EN 1993-1-8,
k
1
= 0.38A
vc
/z
It was shown in Example 8.1 that the relevant lever arm, z, is 543 mm. This stiffness is
relevant only for unequal beam loading, for which the transformation parameter is 1.0
(Fig. 10.1). The shear area of the column web is 3324 mm
2
, so
k
1
= 0.38 3324/543 = 2.33 mm (D10.3)
181
CHAPTER 10. ANNEX A (INFORMATIVE). STIFFNESS OF JOINT COMPONENTS IN BUILDINGS
193
12 May 2004 11:53:16
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:40
Column web in compression, unstiffened
k
2
= 0.7b
eff, c, wc
t
wc
/d
c
From Example 8.1, the width of the web, b
eff, c, wc
, is 248 mm; t
wc
= 10 mm, and
d
c
= 240 2 (21 + 17) = 164 mm
so
k
2
= 0.7 248 10/164 = 10.6 mm
Column web in tension, unstiffened
k
3
= 0.7b
eff, t, wc
t
wc
/d
c
where b
eff, t, wc
is the smallest of the effective lengths
eff
found for the column T-stub. These
are given in Table 6.4 of EN 1993-1-8. The smallest is 2m, with m = 23.2 mm (Fig. 8.9),
so
eff
= 146 mm. Hence,
k
3
= 0.7 146 10/164 = 6.22 mm
Column flange in bending
k
4
= 0.9
eff
(t
fc
/m)
3
where
eff
= 2m as above, so
k
4
= 0.9 146 (17/23.2)
3
= 51.7 mm
End plate in bending
k
5
= 0.9
eff
(t
p
/m)
3
The smallest
eff
for the end plate was 213 mm (Example 8.1), and from Fig. 8.9,
m = 33.9 mm. Hence,
k
5
= 0.9 213 (12/33.9)
3
= 8.50 mm
Bolt in tension
k
10
= 1.6A
s
/L
b
where L
b
is the grip length (29 mm) plus an allowance for the bolt head and nut; total
44 mm. The tensile stress area is 245 mm
2
, so
k
10
= 1.6 245/44 = 8.91 mm
per row of two bolts.
Slab reinforcement in tension
Let h be the length of reinforcing bar assumed to contribute to the flexibility of the joint.
There are two cases:
(a) beams equally loaded, for which
1
=
2
= 0.5, from Fig. 10.1
(b) beams unequally loaded, for which
1
= 3.6 and
2
= 0.
182
194
12 May 2004 11:53:16
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:40
k
s, r
= A
s, r
/h
where A
s, r
= 1206 mm
2
and h = 240 mm. It follows that
k
s, r(a)
= 1206/(0.5 240) = 10.05 mm
for both joints,
k
s, r(b)
= 1206/(3.6 240) = 1.40 mm
for the joint in span BC and
k
s, r(b)

for the joint in span AB. This result is revised later.
Deformation of the shear connection
In the notation of clause A.3:
hogging length of beam: = 0.15 12 = 1.80 m
distance of bars above centre of compression: h

s
= 543 mm from Fig. 8.8(a)
distance of bars above centroid of steel beam: d

s
= 325 mm from Fig. 8.8
second moment of area of steel beam: 10

6
I
a
= 337 mm
4
.
Allowance will be made for the possible reduced stiffness of pairs of studs in a trough of
sheeting by taking N, the number in length , as the equivalent number of single studs. In
Example 6.7, 19.8 equivalent studs were spread along a 2.4 m length of beam each side of
support B (see Fig. 6.30). In Example 8.1, the area of tension reinforcement was reduced
from 1470 to 1206 mm
2
, but the shear connection is now assumed to be as before.
For = 1.8 m,
N = 19.8 1.8/2.4 = 14.85 studs
From equation (A.8),
= E
a
I
a
/d
s
2
E
s
A
s
= 210 337/(0.325
2
200 1206) = 2.778
= [(1 + )Nk
sc
d
s
2
/E
a
I
a
]
1/2
= [3.778 14.85 0.100 1.8 325
2
/(210 337)]
1/2
= 3.88
K
sc
= Nk
sc
/[ ( 1)(h
s
/d
s
)/(1 + )]
= 14.85 100/[3.88 2.88 (543/325)/3.778] = 570 kN/mm
From equation (A.5), the reduction factor to be applied to k
s, r
is
k
slip
= 1/(1 + E
s
k
s, r
/K
sc
) = 1/(1 + 210k
s, r
/570)
The symbol k
s, red
is used for the reduced value. For k
s, r
= 10.05 mm, k
slip
= 0.213, whence
k
s, red
= k
s, r
k
slip
= 2.14 mm
for beams equally loaded. For k
s, r
= 1.40 mm, k
slip
= 0.660, whence
k
s, red
= 0.924 mm (D10.4)
for joint 1, unequal loading. For k
s, r
, k
slip
= 0. Hence, k
s, red
is indeterminate. This is an
anomaly in the code. Research has found that the steel tension zone of the joint that
resists the lower bending moment should be treated as rigid, so in this case, for joint BA,
k
s, red
(D10.5)
183
195
12 May 2004 11:53:16
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:40
Stiffness of joints, for both beams fully loaded
The rules for assembly of stiffnesses in Table 6.10 of EN 1993-1-8 are extended on p. B3.7
of ECCS TC11
39
to allow for the slab reinforcement.
Stiffness in tension
Stiffnesses 3, 4, 5 and 10 are in series (see Fig. 8.3). For these,
1/k
t
= 1/6.22 + 1/51.7 + 1/8.5 + 1/8.91 = 0.410 mm
1
k
t
= 2.44 mm
The lever arm for k
t
is z
2
= 0.383 m. The lever arm for the reinforcement is z
1
= 0.543 m,
and k
s, red
= 2.14 mm. The equivalent lever arm is
z
eq
= (k
s, red
z
1
2
+ k
t
z
2
2
)/(k
s, red
z
1
+ k
t
z
2
)
= 0.989/2.096 = 0.472 m (D10.6)
The equivalent stiffness in tension is
k
eq
= (k
s, red
z
1
+ k
t
z
2
)/z
eq
= 2.096/0.472 = 4.44 mm (D10.7)
Stiffness in compression
Only the stiffness for the column web is required: k
2
= 10.6 mm.
Stiffness of joints
From clause 6.3.1 of EN 1993-1-8, the initial stiffness of each composite joint is
S
j, ini
= E
a
z
eq
2
/(1/k
eq
+ 1/k
2
)
= 210 0.472
2
/(1/4.44 + 1/10.6) = 146 kN m/mrad (D10.8)
If, during construction, the end plates yield in tension, the stiffness represented by k
t
is
ineffective for actions on the composite member. Then, for the stiffness of the composite
joint, k
t
= 0, z
eq
= z
1
and k
eq
= k
s, red
= 2.14 mm, and equation (D10.8) becomes
S
j, ini
= E
a
z
1
2
/(1/k
s, red
+ 1/k
2
)
= 210 0.543
2
/(1/2.14 + 1/10.6) = 110 kN m/mrad (D10.9)
For each joint during construction, k
s, red
= 0, z = z
2
and k
eq
= k
t
, so
S
j, ini, steel
= 210 0.383
2
/(1/2.44 + 1/10.6) = 61.1 kN m/mrad (D10.10)
Stiffness of joints, for imposed load on span BC only
The flexibility of the column web in shear, 1/k
1
(equation (D10.3)), must now be included,
and the values of k
s, red
are different for the two joints. There are two cases: either the end
plate is elastic, or it has yielded in the tension zone, so that k
t
= 0.
For the connection in span BC, with z
1
and z
2
as above, from equations (D10.6) and
(D10.7),
z
eq
= (0.924 0.543
2
+ 2.44 0.383
2
)/(0.924 0.543 + 2.44 0.383)
= 0.630/1.436 = 0.439 m
k
eq
= 1.436/0.439 = 3.27 mm
Including 1/k
1
in equation (D10.8),
S
j, ini, BC
= 210 0.439
2
/(1/2.33 + 1/10.6 +1/3.27) = 48.8 kN m/mrad (D10.11)
If k
t
= 0, then fromequations (D10.6) and (D10.7), z
eq
= z
1
, and k
eq
= k
s, red
=0.924 mm.
S
j, ini, BC
= 210 0.543
2
/(1/2.33 + 1/10.6 +1/0.924) = 38.6 kN m/mrad (D10.12)
184
196
12 May 2004 11:53:17
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:40
For the joint in span AB, with k
s, red
, then z
eq
= z
1
and k
eq
, and equation
(D10.8) gives
S
j, ini, BA
= 210 0.543
2
/(1/2.33 + 1/10.6) = 118 kN m/mrad (D10.13)
This use of equation (D10.8) where there is shear in the web panel includes an
approximation that leads to a small overestimate of the deflection of span BC. The shear
deformation of the column web panel, of stiffness k
1
(2.33 mm here), causes clockwise
rotation of end B of span AB, and hence increases the hogging bending moment at
this point. The effective stiffness S
j, ini, BA
therefore exceeds 118 kN m/rad. The resulting
increase in the hogging moment at Bin span BCis small, and so is the associated decrease
in the mid-span deflection.
Results (D10.8) to (D10.13) are repeated in Table 8.1, for use in Example 8.1.
185
197
12 May 2004 11:53:17
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
CHAPTER 11
Annex B (Informative). Standard
tests
B.1. General
Annex B is informative, not normative, because test procedures for products are strictly
outside the scope of a design code. From the note to clause B.1(1), they should be given in a
European standard or in guidelines for European technical approvals, which are not yet
available.
One of the objectives of standard tests is to provide guidance to designers in situations
where calculation models are not sufficient. This commonly occurs for two components of
composite structures: shear connectors and profiled steel sheeting. Existing design rules for
both shear connection and composite slabs are based mainly on test data obtained over many
decades using various procedures and types of test specimen, for which there has been no
international standard.
There are many national standards, and there is some international consensus on details
of the mk test for resistance of composite slabs to longitudinal shear. However, evidence
has accumulated over the past 20 years that both this test and the UKs version of the push
test for shear connectors have significant weaknesses. These restrict the development of new
products, typically by giving results that are over-conservative (push test) or misleading (mk
test). A full set of mk tests for a new profile is also expensive and time-consuming.
When the specification for an existing test is changed, past practice should, in principle, be
re-evaluated. This is one reason why the UKs push test has survived so long in its present
form. The new push test (clause B.2) has been in drafts of Eurocode 4 for 15 years, and was
based on research work in the preceding decade. Almost all non-commercial push tests since
that time have used slabs wider than the 300 mm of the specimen that is defined, for
example, in BS 5400: Part 5.
82
The new test generally gives higher results, and so does not raise questions about past
practice. It costs more, but gives results that are more consistent and relevant to the
behaviour of connectors in composite beams and columns.
For profiled sheeting, most research workers have concluded that the empirical mk test
procedure should be phased out. This method, as given in BS 5950: Part 4,
107
does not
distinguish sufficiently between profiles that fail in a ductile manner and those that fail
suddenly, and does not exploit the use of end anchorage or the ability of many modern
profiles to provide partial shear connection.
113,119
However, its use has been the principal
basis world-wide for the design of composite slabs for longitudinal shear. Re-testing of
the scores (if not hundreds) of types of composite slabs used in existing structures is
impracticable.
199
12 May 2004 11:53:17
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
Clause B.3 therefore sets out in detail a method of testing that can be used with what is
essentially the existing mk procedure, though with a tighter specification relating to the
mode of failure, and other changes based on recent experience of its use. The test method is
intended also for use with the more rational partial-interaction design of composite slabs,
which is to be preferred for reasons given in comments on clauses 9.7 and B.3.
Properties of materials
Ideally, the strengths of the materials in a test specimen should equal the characteristic
values specified for the application concerned. This is rarely possible in composite
test specimens that include three different materials. It is therefore necessary to adjust
resistances found by testing, or to limit the range of acceptable measured strengths of a
material so that any adjustment would be negligible. Relevant provisions are given within
clauses B.2 and B.3.
The influence of cracking of concrete may be assumed to be allowed for in the test
procedures. Tests that fully reproduce the effects of shrinkage and creep of concrete are
rarely practicable; but these effects can normally be predicted once the behaviour in a
short-term test has been established, and have little influence on ultimate strength except in
slender composite columns.
B.2. Tests on shear connectors
General
The property of a shear connector that is needed for design is a curve that relates
longitudinal slip, , to shear force per connector, P, of the type shown in Fig. B.2. No reliable
method has been found for deducing such curves from the results of tests on composite
beams, mainly because bending resistance is insensitive to the degree of shear connection, as
shown by curve CH in Fig. 6.4(a).
Clause B.2.2(1)
Clause B.2.2(2)
Almost all the load-slip curves on which current practice is based were obtained frompush
tests, which were first standardized in the UKin 1965, in CP117: Part 1. Ametricated version
of this test is given in BS 5400: Part 5,
82
and referred to in clause 5.4.3 of BS 5950: Part 3.1
31
without comment on the need to modify it when profiled sheeting is present. This test has
two variants, because the slab and reinforcement should be either as given in [the code] ... or
as in the beams for which the test is designed. This distinction is maintained in EN 1994-1-1
(clause B.2.2(1)). A standard specimen is specified in clause B.2.2(2) and Fig. B.1, and a
specimen for specific push tests is defined in more general terms in clause B.2.2(3). The
principal differences between the standard tests of EN 1994-1-1 and BS 5400 (the BS test)
are summarized below, with reference to Fig. 11.1, and reasons for the changes are given.
The standard test is intended for use
where the shear connectors are used in T-beams with a concrete slab of uniform thickness, or with
haunches complying with 6.6.5.4. In other cases, specific push tests should be used.
It can be inferred from clause B.2.2(1) that separate specific push tests should be done to
determine the resistance of connectors in columns and in L-beams, which commonly occur at
external walls of buildings and adjacent to large internal holes in floors. This is rarely, if ever,
done, although a connector very close to a free edge of a slab is likely to be weaker and have
less slip capacity than one in a T-beam.
79
This problem can be avoided by appropriate
detailing, and is the reason for the requirements of clauses 6.6.5.3 (on longitudinal splitting)
and 6.6.5.4 (on the dimensions of haunches).
Push tests to clause B.2, compared with the BS test
Welded headed studs are the only type of shear connector for which large numbers of tests
have been done in many countries, so all reported studies of push testing (e.g. see Johnson
and Oehlers,
67
Stark and van Hove
72
and Oehlers
120
) are based on these tests. It has been
188
200
12 May 2004 11:53:17
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
found that the results of the tests are widely scattered.
76
To obtain realistic characteristic
values, it is necessary to separate inherent variability from that due to differences in the test
specimens, the methods of casting and testing, and the ultimate tensile strength of the
connectors.
The BS specimen was probably designed to give results at the lower edge of the band
of uncertainty that existed 40 years ago, because it has very small slabs, prone to split
longitudinally because the mild steel reinforcement is light and poorly anchored. It has
connectors at only one level, which in effect prevents redistribution of load from one slab to
the other
92
and so gives the resistance of the weaker of the two pairs of connectors. The
changes from this test are as follows:
(1) The slabs have the same thickness, but are larger (650 600 mm, cf. 460 300 mm).
This enables reinforcement to be better anchored, and so avoids low results due to
splitting. The bond properties of the reinforcement are more important than the yield
strength, which has little influence on the result. Limits are given in Fig. B.1.
(2) The transverse reinforcement is 10 high-yield ribbed bars per slab, instead of four mild
steel bars of the same diameter, 10 mm, so the transverse stiffness provided by the bars is
at least 2.5 times the previous value. In T-beams, the transverse restraint from the
in-plane stiffness of the slab is greater than in a push specimen. The reinforcement is
intended to simulate this restraint, not to reproduce the reinforcement provided in a
beam.
(3) Shear connectors are placed at two levels in each slab. This enables redistribution of
load to occur, so that the test gives the mean resistance of eight stud connectors, and
better simulates the redistribution that occurs within the shear span of a beam.
(4) The flange of the steel section is wider (> 250 mm, cf. 146 mm), which enables wider
block or angle connectors to be tested; and the lateral spacing of pairs of studs is
standardized. The HE 260B section (Fig. B.1) is 260 260 mm, 93 kg/m.
Clause B.2.3(1)
(5) Each concrete slab must be cast in the horizontal position, as it would be in practice
(clause B.2.3(1)). In the past, many specimens were cast with the slabs vertical, with the
risk that the concrete just below the connectors would be poorly compacted.
189
CHAPTER 11. ANNEX B (INFORMATIVE). STANDARD TESTS
Cover 15
150
150
35
35 Recess optional
30
150
150
100
150 150 260
200 200 200
180 180 180
Bedded in mortar or gypsum
Reinforcement: ribbed bars, 10 mm, resulting
in a high bond with 450 f
sk
550 N/mm
2
Steel section: HE 260 B or 254 254 89 kg U.C.
250
250
600 100
P
Fig. 11.1. Test specimen for the standard push test (dimensions in mm)
201
12 May 2004 11:53:17
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
Clause B.2.3(3)
Clause B.2.3(4)
(6) Unlike the BS test, details of concrete curing are specified, in clauses B.2.3(3) and
B.2.3(4).
(7) The strength of the concrete measured at the time the push test is done must satisfy
0.6 f
cm
/f
ck
0.8 (D11.1)
where f
ck
is the specified strength in practice (clause B.2.3(5)). The corresponding rule
for the BS test is
0.86 f
cm
/f
cu
1.2
where f
cu
is the cube strength of the concrete in the beams. For both codes, the two
strength tests must be done using the same type of specimen, cylinder or cube.
Condition (D11.1) is now explained. It is essentially
f
cm
= 0.7f
ck
The resistance of a stud is usually found from equation (6.19):
P
Rd
= 0.29d
2
(f
ck
E
cm
)
0.5
/
V
In the push test, f
ck
is in effect replaced by 0.7f
ck
. Then,
P
Rd
= 0.29d
2
(0.7f
ck
E
cm
)
0.5
/
V
= 0.29d
2
(f
ck
E
cm
)
0.5
/1.5 (D11.2)
when
V
= 1.25. This shows that the purpose of condition (D11.1) is to compensate for
the use of a
V
factor of 1.25, lower than the value 1.5 normally used for concrete, and the
likelihood that the quality of the concrete in the laboratory may be higher than on site.
Clause B.2.4(1) (8) The loading is cycled 25 times between 5 and 40% of the expected failure load (clause
B.2.4(1)). The BS test does not require this. Stresses in concrete adjacent to shear
connectors are so high that, even at 40% of the failure load, significant local cracking
and inelastic behaviour could occur. This repeated loading ensures that if the connector
tested is susceptible to progressive slip, this will become evident.
Clause B.2.4(3)
Clause B.2.4(4)
(9) Longitudinal slip and transverse separation are measured (clauses B.2.4(3) and B.2.4(4)),
to enable the characteristic slip and uplift to be determined, as explained below. The BS
test does not require this.
Evaluation of results of push tests
Clause B.2.5(1)
Normally, three tests are conducted on nominally identical specimens to determine the
characteristic resistance P
Rk
for concrete and connector material of specified strengths f
ck
and f
u
, respectively. Let P
m
be the mean and P
min
the lowest of the three measured resistances
per connector, and f
ut
be the measured ultimate strength of the connector material. If all
three results are within 10% of P
m
, then, from clause B.2.5(1),
P
Rk
= 0.9P
min
(D11.3)
Clause B.2.5(2) refers to Annex D of EN 1990 (Informative) for the procedure to be
followed if the scatter of results exceeds the 10% limit.
Amethod to clause D.8 of EN1990 for the deduction of a characteristic value froma small
number of test results, which took no account of prior knowledge, would severely penalize a
three-test series. It is necessary to rely also on the extensive past experience of push testing.
Clause D.8.4 is relevant. For three tests it sets the condition that all results must be within
10%of the mean, P
m
. This appears in clause B.2.5(1). Clause D.8.4 then gives the characteristic
resistance as a function of P
m
and of V
r
, the maximum coefficient of variation observed in
previous tests, in which the 10% from the mean condition was satisfied.
Most of the previous results were fromresearch programmes, with many different types of
test specimen. The results for studs in profiled sheeting, for example, have been found to be
samples from seven different statistical populations.
76
It has not been possible to establish
the value of V
r
. The method of clause B.2.5(1), of reducing the lowest of the three results by
10%, is mainly based on previous practice. It can be deduced from clause D.8.4 that for a set
190
202
12 May 2004 11:53:18
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
of three results with the lowest 10% below the mean, the method of clause B.2 implies that
V
r
= 11%.
Clause B.2.5(1) gives a penalty that applies when f
ut
> f
u
. This is appropriate where the
resistance of a connector is governed by its own material, usually steel, but in practice
the resistance of a connector can depend mainly on the strength of the concrete, especially
where lightweight aggregate is used. The correction then seems over-conservative, because
f
u
is limited to 500 N/mm
2
by clause 6.6.3.1(1), and the strength of the material can exceed
600 N/mm
2
for studs.
In the BS test, a nominal strength P
u
is calculated from
P
u
= (f
ck
/f
c
)P
min
and, then,
P
Rd
= P
u
/1.4
It so happens that 1.25/0.9 = 1.4, so from equation (D11.3) the two methods give a similar
relationship between P
min
and P
Rd
, except that the Eurocode result is corrected for the
strength of the steel, and the BS result is corrected for the strength of the concrete. This is
probably because the results of the BS test are rarely governed by the strength of the steel, as
the slabs are so likely to split.
Clause B.2.5(3) Clause B.2.5(3) finds application for connectors such as blocks with hoops, where the
block resists most of the shear, and the hoop resists most of the uplift.
Clause B.2.5(4)
The classification of a connector as ductile (clause 6.6.1.1(5)) depends on its characteristic
slip capacity, which is defined in clause B.2.5(4). From the definition of P
Rk
(clause B.2.5(1)),
all three test specimens will have reached a higher load, so the slips
u
in Fig. B.2 are all taken
from the falling branches of the loadslip curves. It follows that a push test should not be
terminated as soon as the maximum load is reached.
B.3. Testing of composite floor slabs
General
The most usual mode of failure of a composite slab is by longitudinal shear, loss of interlock
occurring at the steelconcrete interface. Resistance to longitudinal shear is difficult to
predict theoretically. The pattern and height of indentations or embossments and the shape
of the sheeting profile all have significant effect. There is no established method to calculate
this resistance, so the methods of EN 1994-1-1 rely on testing.
Clause B.3.1(1)
Tests are needed for each new shape of profiled sheet. They are normally done by or for
the manufacturer, who will naturally be concerned to minimize their cost. Their purpose
(clause B.3.1(1)) is to provide values for either the factors m and k for the mk method, or
the longitudinal shear strength required for the partial shear connection method. These
procedures for verifying resistance to longitudinal shear are given in clauses 9.7.3 and 9.7.4.
Comments on them are relevant here.
Clause B.3.1(2) The tests also determine whether the shear connection is brittle or ductile (clause
B.3.1(2)). There is a 20% penalty for brittle behaviour (clause B.3.5(1)). In view of the
purpose of the tests, failure must be in longitudinal shear (clauses B.3.2(6) and B.3.2(7)).
Number of tests
Clause B.3.1(3)
Clause B.3.1(4)
The list of relevant variables in clause B.3.1(3) and the concessions in clause B.3.1(4) define
the number of tests required. As an example, it is assumed that a manufacturer seeks to
determine shear resistance for a new profile, as the basis for design data for a range of sheet
thicknesses, slab thicknesses and spans, and concrete strengths, with both lightweight and
normal-weight concrete. How many tests are required?
In view of the penalty for brittle behaviour (clause B.3.5(1)), it is assumed that the new
profile is found to satisfy clause 9.7.3(3) on ductility. The partial-connection method is more
191
203
12 May 2004 11:53:18
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
versatile than the mk method, and is recommended. Its calculations are straightforward on
a spreadsheet. From clause B.3.2(7), its tests are done in groups of four. The variables are
now considered in turn, to find the minimum number of values needed for each one, and,
hence, the number of tests needed for a full set.
(a) Thickness of sheeting: test the thinnest sheeting. As interlock is dependent on the local
bending of individual plate elements in the sheeting profile, the results may not be
applied to thinner or significantly weaker sheets, which would be more flexible . . . (1)
(b) Type of sheeting, meaning the profile, including any overlap details, and the specification
of embossments and their tolerances. Ensure that the embossments on sheets tested
satisfy clause B.3.3(2), and standardize the other details . . . . . . . . . . . . . . . . (1)
(c) Steel grade: test the highest and lowest grades to be used. The materials standards listed
in clause 3.5 include several nominal yield strengths . . . . . . . . . . . . . . . . . . (2)
(d) Coating: this should be standardized, if possible . . . . . . . . . . . . . . . . . . . . (1)
(e) Density of concrete: test the lowest and highest densities. . . . . . . . . . . . . . . . (2)
(f) Grade of concrete: test with a mean strength not exceeding 1.25 times the lowest value of
f
ck
to be specified (see clause B.3.1(4)). The results will be slightly conservative for
stronger concretes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (1)
(g) Slab thickness: test the thinnest and thickest slabs . . . . . . . . . . . . . . . . . . . (2)
The use of a single thickness is not permitted because the effectiveness of shear
connection may depend on the stiffness of the concrete component. Conclusions from a
theoretical model for the effect of slab thickness, given in Appendix B of this guide, are
summarized below.
(h) Shear span: account is taken of this in the provisions for use of the test results.
This gives a total of
4 1
4
2
3
= 32 tests
If it is suspected that the results for parameters (a) and (g) will be over-conservative for
thick sheets and strong concrete, respectively, even more tests would be needed. If an
alternative coating is to be offered, it should be possible to compare its performance with
that of the standard coating in a few tests, rather than another full set.
For the mk method, tests are in groups of six, so the number rises from 32 to 48.
The main conclusions from Appendix B of this guide are as follows;
for the partial-interaction method, interpolation between results from tests on slabs of
the same shear span and two thicknesses is valid for slabs of intermediate thickness
for the mk method, results from tests on two shear spans are applicable for shear spans
between those tested.
The status of Annex B, and use of fewer tests
Annex B is informative. From notes to clauses 9.7.3(4) and 9.7.3(8), its test methods may be
assumed to meet the basic requirements of the relevant design method for longitudinal shear.
These requirements are not defined; they have to be inferred, mainly from Annex B.
This implies that where the testing does not conform to the extensive scheme outlined
above, some independent body, such as the relevant regulatory authority or its nominee,
must be persuaded that the evidence presented does satisfy the basic requirements of one
or both of the two design methods. Where this is done, the design can presumably claimto be
in accordance with Eurocode 4, in this respect; but that would not apply internationally.
Annex B may eventually be superseded by a European standard on the determination
of the shear resistance of composite slabs. Until then, the situation is unsatisfactory.
Development of better theoretical models would help. At present, research workers cannot
validate these because manufacturers rarely release their detailed test results.
Where a new profile is a development from an existing range, it should be possible to use
the results of earlier tests to predict the influence of some of the parameters, and so reduce
the number of new tests required.
192
204
12 May 2004 11:53:18
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
Testing arrangement
Clause B.3.2(2)
Loading is applied symmetrically to a simply-supported slab of span L, at points distant L/4
from each support (clause B.3.2(2)). Crack inducers are placed beneath the loads (clause
B.3.3(3)), to reduce the effect of local variations in the tensile strength of the concrete. The
failure loading is much heavier than the slabs, so the shear spans L
s
(Fig. B.3) are subjected to
almost constant vertical shear. This differs from the test details in BS 5950: Part 4, in which
vertical shear is not constant over the length between a crack inducer and the nearer support.
Clause B.3.2(6)
As the shear span is a fixed proportion of the span, specimens for regions such as A and B
for the mk method (clause B.3.2(6) and Fig. B.4) are obtained by altering the span L. As this
method is empirical, it is good practice to ensure that the tests also encompass the range of
spans required for use in practice.
107
Clause B.3.2(7)
For sheeting with ductile behaviour and where design will use the partial-interaction
method, the number of tests in a series, for specimens of given thickness h
t
, can be reduced
from six to four (clause B.3.2(7)).
Clause B.3.3(1)
Clause B.3.3(2)
Clauses B.3.3(1) and B.3.3(2) are intended to minimize the differences between the
profiled sheeting used in the tests and that used in practice. The depth of embossments has
been found to have a significant effect on the resistance.
Clause B.3.3(6)
Propped construction increases the longitudinal shear. It is required for the test specimens
to enable the results to be used with or without propping (clause B.3.3(6)).
Clause B.3.3(8)
Clause B.3.3(9)
Fromclause B.3.3(8), the specimens for the determination of concrete strength, defined in
clause B.3.3(9), should be cured under the same conditions as the test slabs. This cannot of
course be the curing under water normally used for standard cubes and cylinders. When
deviation of strength from the mean is significant, the concrete strength is taken as the
maximum value (clause B.3.3(9)).This causes the applicability of the test results to be more
restricted (clause B.3.1(4)).
Clause B.3.3(10) The test procedure for the strength of the profiled sheeting (clause B.3.3(10)) is given
elsewhere.
121
Clause B.3.4(3)
Clause B.3.4(4)
The initial loading test is cyclic (clauses B.3.4(3) and B.3.4(4)), to destroy any chemical
bond between the sheeting and the concrete, so that the subsequent test to failure gives a
true indication of the long-termresistance to variations of longitudinal shear. The number of
cycles, 5000, is fewer than that required by BS 5940: Part 4, but has been judged to be
adequate for these purposes.
Design values for m and k
Clause B.3.5(1) Clause B.3.5(1) gives a design rule for the possibility that the two end reactions may differ
slightly, and applies an additional factor of safety of 1.25 to compensate for brittle behaviour,
in the form of a reduction factor of 0.8.
Clause B.3.5(2)
Clause B.3.5(3)
The method of clause B.3.5(2) is applicable to any set of six or more test results,
irrespective of their scatter. An appropriate statistical model will penalize both the scatter
and the number of results, if small, and may be that given in EN1990, to which clause B.3.5(3)
refers.
Where a series consists of six tests and the results are consistent, clause B.3.5(3) provides a
simple method for finding the design line shown in Fig. B.4, and hence values for m and k.
These are in units of N/mm
2
.
These methods differ from that of BS 5950: Part 4, both in the determination of the line
that gives the values of m and k (see Fig. 11.3) and in their definition. In BS 5950, k is
proportional to the square root of the concrete strength, which causes complications with
units, and the two sets of three results could be from slabs with different concrete strengths.
It has been found that the deliberate use of very different strengths for the specimens in
regions A and B in Fig. 11.3 can lead to unsafe applications of the method, when m and k are
defined as in BS 5950: Part 4. All the results in a diagram such as Fig. B.4 are required by
clause B.3.3(8) to be fromspecimens with nominally identical concrete, so there is no need to
include concrete strength in the functions plotted in this figure.
193
205
12 May 2004 11:53:18
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
The formula for vertical shear in clause 9.7.3(4) and the definitions of m and k in clause
B.3.5 are dimensionally correct, and can be used with any consistent set of units. However,
analyses of a given set of test results by the Eurocode method give values for m and k
different to those found by the BS method, and k even has different dimensions. The
conversion of BS values to Eurocode values is illustrated in Example 11.1. The applicability
of sets of test results not in accordance with Annex B are discussed in Appendix B of this
guide.
Design values for
u, Rd
Clause B.3.6(1) Clause B.3.6(1) refers to the partial-interaction curve shown in Fig. B.5. This is for sagging
bending, and is determined for a group of tests on specimens with nominally identical
cross-sections as follows:
(a) The measured values of the required dimensions and strengths of materials are
determined, and used to calculate the full-interaction plastic moment of resistance of a
test specimen, M
p, Rm
and the corresponding compressive force in the concrete slab, N
c, f
.
(b) A value is chosen for (= N
c
/N
c, f
), which determines a value N
c
, the partial-interaction
compressive force in the slab at the section where flexural failure is assumed to occur.
The corresponding value of the bending resistance M is then calculated from
M = M
pr
+ N
c
z
with M
pr
from equation (9.6), with N
c
replacing N
c, f
, and z from equation (9.9). This gives
a single point on the curve in Fig. B.5, which assumes that an undefined slip can occur at
the interface between the sheeting and the concrete. This necessitates ductile behaviour.
(c) By repeating step (b) with different values of , sufficient points are found to define the
partial-interaction curve.
Clause B.3.6(2) From clause B.3.6(2), a bending moment M is found from each test. This is M
test
in Fig. B.5,
and leads to a value
test
, and hence
u
from equation (B.2).
For the test arrangement shown in Fig. B.3, there will be an overhang L
0
beyond the shear
span L
s
, along which slip will occur. This is allowed for in equation (B.2), which assumes that
the shear strength is uniformalong the total length (L
s
+ L
0
). In reality, the strength includes
a contribution from friction at the interface between the sheeting and the concrete, arising
from the transmission of the vertical load across the interface to the support. In clause
B.3.6(2), this effect is included in the value found for
u
.
Clause B.3.6(3) A more accurate equation for
u
is given in clause B.3.6(3), for use with the alternative
method of clause 9.7.3(9), where relevant comment is given.
Clause B.3.6(4)
From clause B.3.2(7), a group of four tests on specimens of given span and slab thickness
gives three values of
u
, and evidence on ductility (clause B.3.2(7)). All the values of
u
are
used in a single calculation of the lower 5% fractile value, to clause B.3.6(4). This is divided
by
VS
to obtain the design value used in clause 9.7.3(8).
The shape of the partial-interaction curve depends on the slab thickness, so a separate one
is needed for each thickness. The need for tests at different thicknesses is discussed at the
end of Appendix B of this guide.
Example 11.1: mk tests on composite floor slabs
In this example, values of m and k are determined from a set of tests not in accordance
with clause B.3, Testing of composite floor slabs. These tests, done in accordance with the
relevant Netherlands standard, RSBV 1990, were similar to specific tests as specified in
clause 10.3.2 of ENV 1994-1-1, which has been omitted from EN 1994-1-1.
This set of tests on eight simply-supported composite slabs has been fully reported.
122
The cross-section of the profiled sheeting is shown in Fig. 9.4. The values of m and k
194
206
12 May 2004 11:53:19
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
determined here are used in Example 9.1, Two-span continuous composite slab, which
includes the results of tests on the non-composite sheeting.
Test specimens and procedure
All of the composite slabs had the same breadth, 915 mm, which satisfies clause B.3.3(5),
and were cast using the same mix of lightweight-aggregate concrete, with propped
construction (clause B.3.3(6)). (Where a clause number is given without comment, as
here, it means that the clause was complied with.) The tests to failure were done between
the ages of 27 and 43 days, when the strengths of test specimens stored with slabs (clause
B.3.3(8)) were as given in Table 11.1. Steel mesh with 6 mm bars at 200 mm spacing was
provided in each slab (clause B.3.3(7)).
For specimens 14, the overall thickness and span were h
t
= 170 mm and
L = 4500 mm, where the notation is as in Fig. B.3. For specimens 58, h
t
= 120 mm
and L = 2000 mm. The distance between the centre-line of each support and the adjacent
end of each slab was 100 mm (clause B.3.2(4)). The surfaces of the sheeting were not
degreased (clause B.3.3(1)).
The tensile strength and the yield strength of the sheeting were found fromcoupons cut
fromits top and bottomflanges (clause B.3.3(10)). The mean values were f
u
= 417 N/mm
2
and f
y, 0.2
= 376 N/mm
2
. The stress 376 N/mm
2
, measured at the 0.2% proof strain, is
significantly higher than the yield strength found in another series of tests,
123
which was
320 N/mm
2
. The nominal yield strength for this sheeting, now 350 N/mm
2
, was then
280 N/mm
2
, which is only 74% of 376 N/mm
2
, so clause B.3.1(4) is not complied with. This
is accepted, because small changes of yield strength have little influence on resistance to
longitudinal shear.
There is a similar 80% rule in clause B.3.1(4) for the strength of the concrete. Applying
it to the mean cube strength for this series, 34.4 N/mm
2
, gives a cube strength of
27.5 N/mm
2
, which is below the value used in Example 9.1, so the rule is satisfied.
The slabs were tested under four-point loading, as shown in Fig. 11.2, and crack
inducers were provided at distances L/8 each side of mid-span. This is not in accordance
with Fig. B.3; clause B.3.2(3) specifies two-point loading. An appropriate value of L
s
has
to be found, for use in the determination of m and k. A shearforce diagram for
two-point loading is found (the dashed line in Fig. 11.2) that has the same area as the
195
= = = = = = =
=
L
0
Vertical shear
due to applied
loading
L
s
W/2
W/4 W/4 W/4 W/4
Crack inducers
Fig. 11.2. Loading used in the tests on composite slabs
207
12 May 2004 11:53:19
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
actual shearforce diagram, and the same maximum vertical shear. Here, L
s
= L/4, as it
is in Annex B.
Cyclic loading
Clause B.3.4 specifies 5000 cycles of loading between 0.2W
t
and 0.6W
t
, where W
t
is a static
failure load. In these tests, W
t
for specimen 1 was 75.5 kN, and the range of loading in tests
24, for 10 000 cycles, was from 0.13W
t
to 0.40W
t
. For specimens 58, with a mean failure
load of 94.4 kN, the fatigue loading in tests 68 ranged from 0.19W
t
to 0.57W
t
, close to the
range specified. These divergences are not significant.
Results of tests to failure
To satisfy clause 9.7.3(3) on ductility, it is necessary to record the total load on the
specimen, including its weight, at a recorded end slip of 0.1 mm, at a deflection () of
span/50, and at maximum load. These are given in Table 11.1.
The failure load, as defined in clause 9.7.3(3), is for all these tests the value when L/50.
These loads all exceed the load at a slip of 0.1 mm by more than 10%, the least excess
being 13%. All failures are therefore ductile. From clause B.3.5(1), the representative
vertical shear force V
t
is taken as half the failure load.
Determination of m and k
In the original report
122
the axes used for plotting the results were similar to those in
BS 5950: Part 4 (Fig. 11.3(a)). They are
X = A
p
/[bL
s
(0.8 f
cu
)
0.5
] Y = V
t
/[bd
p
(0.8f
cu
)
0.5
] (D11.4)
where f
cu
is the measured cube strength. The other symbols are in Eurocode notation.
Values of X and Y were calculated from the results and are given in Table 11.1. From
these, the following values were determined:
m = 178 N/mm
2
k = 0.0125 N
0.5
mm (D11.5)
For m and k as defined in EN 1994-1-1, the relevant axes (Fig. 11.3(b)) are
x = A
p
/bL
s
y = V
t
/bd
p
so that
x = X(0.8f
cu
)
0.5
y = Y(0.8f
cu
)
0.5
(D11.6)
Approximate values for m and k to EN 1994-1-1 can be found by assuming that m to
BS 5950: Part 4 is unchanged, and k is (0.8f
cu, m
)
0.5
times the BS value, given above, where
f
cu, m
is the mean cube strength for the series. These values are
196
Table 11.1. Results of tests on composite floor slabs
Test
No.
f
cu
(N/mm
2
)
Load at
0.1 mm
slip (kN)
Load at
= L/50
(kN)
Maximum
load (kN)
10
3
X
((1/N
0.5
)mm)
10
3
Y
(N
0.5
mm)
10
3
x 10
3
y
(N/mm
2
)
1 31.4 63.3 75.5 75.5 0.222 58.6 1.115 294
2 30.4 65.3 75.5 75.5 0.226 59.6 1.115 294
3 32.1 64.3 73.9 73.9 0.220 56.7 1.115 287
4 34.2 66.8 75.4 75.4 0.213 56.1 1.115 293
5 35.3 52.2 94.0 94.2 0.472 108.2 2.51 575
6 37.4 54.2 90.9 94.0 0.459 104.9 2.51 574
7 37.2 52.2 91.7 93.9 0.460 105.0 2.51 573
8 36.9 56.2 94.7 95.4 0.462 107.2 2.51 582
208
12 May 2004 11:53:19
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
m = 178 N/mm
2
k = 0.066 N/mm
2
(D11.7)
These results are approximate because f
cu
is different for each test, and the procedure
of Annex B for finding characteristic values differs from the BS procedure. The correct
method is to calculate x and y for each test, plot a new diagram, and determine m and k
from it in accordance with clause B.3.5, as follows.
The values of x and y for these tests, from equations (D11.6), are given in Table 11.1.
The differences within each group of four are so small that, at the scale of Fig. 11.4, each
plots as a single point: A and B. Clause B.3.5(3), on variation within each group, is
satisfied. Using the simplified method of that clause, the characteristic line is taken to be
the line through points C and D, which have y coordinates 10% below the values for
specimens 3 and 7, respectively. This line gives the results
m = 184 N/mm
2
k = 0.0530 N/mm
2
(D11.8)
which are used in Example 9.1.
For these tests, the approximate method gives a small error in m, 3%, and a larger
error in k, +25%. From clause 9.7.3(4), the design shear resistance is
197
Shear-bond failure
regression line
Design line
(reduction of 15%)
A
p
/bL
s
V
t
/bd
p
A
p
/bL
s
f
cm
0.5
V
t
/bd
p
f
cm
0.5
Design line (minimum
values reduced by 10%)
(a) (b)
A
B
Fig. 11.3. Evaluation of the results of tests on composite slabs. (a) BS 5950: Part 4. (b) EN1994-1-1
0.6
0.4
0.2
0.001
1.0
m = 184 N/mm
2
B
D
C
A
A
p
/bL
s
V/bd
p
(N/mm
2
)
0.002 0.003
0
k = 0.053 N/mm
2
Fig. 11.4. Determination of m and k
209
12 May 2004 11:53:20
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
V
, Rd
= (bd
p
/
VS
)[(mA
p
/bL
s
) + k] (D11.9)
The second term in square brackets is much smaller than the first, and in this example the
errors in m and k almost cancel out. In Example 9.1, V
, Rd
was found to be 28.0 kN/m.
Using the approximate values of equations (D11.7) for m and k changes it only to
28.3 kN/m.
Similar comparisons are needed with other sets of test results, before any general
conclusion can be drawn about the accuracy of the approximate method of calculation.
Note on the partial-interaction method
A composite slab using sheeting of the type tested was designed in Example 9.1. The mk
method was used for the verification for longitudinal shear, with values of mand k calculated
from test results, as shown in Example 11.1.
To illustrate the partial-interaction method of clauses 9.7.3(7) to 9.7.3(9), an attempt will
be made in the following Example 11.2 to use it in an alternative verification of the same
composite slab, using the same set of test results.
122
It is assumed that the reader is familiar
with the two examples referred to. Example 11.2 illustrates potential problems in using this
method with existing test data, shows that a procedure given in ENV 1994-1-1,
49
and omitted
from EN 1994-1-1, can give unconservative results, and proposes a method to replace it.
Example 11.2: the partial-interaction method
It is shown in Example 11.1 that in the eight tests reported,
122
the behaviour of the slabs
was ductile to clause 9.7.3(3). For the partial-interaction method, clause B.3.2(7) then
requires a minimum of four tests on specimens of the same overall depth h
t
: three with a
long shear span, to determine
u
, and one with a short shear span, but not less than 3h
t
.
The tests available satisfy the 3h
t
condition, but not that for uniform depth. The four
long-span slabs were all thicker than the four short-span slabs. The purpose of the single
short-span test is to verify ductility, which is satisfactory here.
The maximum recorded end slips in the long-span tests (Nos. 1 to 4), only 0.3 mm,
reveal a problem. These tests were discontinued when the deflections reached span/50, so
it is unlikely that the maximum longitudinal shear was reached. The specimens had the
high span/depth ratio of 26.5, and probably failed in flexure, not longitudinal shear. This is
confirmed by the results that follow: the shear strength from tests 14 is about 30% lower
than that from the short-span tests 58, where the end slips at maximum load were from 1
to 2 mm and the span/depth ratio was 16.7.
Clause B.3.2(7) requires the shear strength to be determined from the results for the
long-span slabs. Its condition for a shear span as long as possible while still providing
failure in longitudinal shear was probably not satisfied here, so that the final design value
u, Rd
is lower than it would have been if a shorter span had been used for tests 14. In the
absence of guidance fromprevious tests, this condition is difficult to satisfy when planning
tests.
Other aspects of these tests are compared with the provisions of Annex B in Example
11.1.
The partial-interaction diagram
Measured cube strengths and maximum loads for the eight tests are given in Table 11.1.
The mean measured yield strength and cross-sectional area of the sheeting were
376 N/mm
2
and 1145 mm
2
/m, respectively.
122
For full shear connection, the plastic neutral
axes are above the sheeting, so clause 9.7.2(5) applies.
The longitudinal force for full interaction is
N
c, f
= A
p
f
yp
= 1145 0.376 = 431 kN/m (D11.10)
198
210
12 May 2004 11:53:20
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
and the stress blocks are as shown in Fig. 11.5(b). The mean cube strength for specimens
58, 36.7 N/mm
2
, corresponds to a cylinder strength f
cm
of 29.8 N/mm
2
, so that the depth of
the concrete stress block is
x
pl
= 431/(0.85 29.8) = 17.0 mm (D11.11)
With full interaction, the force N
c, f
in the sheeting acts at its centre of area, 30 mm above
its bottom surface, so the lever arm is
z = 120 17/2 30 = 81.5 mm
and
M
p, Rm
= 431 0.0815 = 35.1 kN m/m
The method of calculation for the partial-interaction diagram is now considered,
following clause B.3.6(1). The stress blocks in Fig. B.5 correspond to those used in clauses
9.7.2(5) and 9.7.2(6) modified by clause 9.7.3(8). It follows from equation (D11.11) that
for any degree of shear connection (= N
c
/N
c, f
), the stress block depth is 17 mm, with a
line of action 8.5 mm below the top of the slab (Fig. 11.5(c)).
For any assumed value for , the lever arm z is given by equation (9.9). For typical
trapezoidal sheeting, where the profile is such that e
p
> e (these symbols are shown in Fig.
9.6), the simplification given in equation (D9.4) should for this purpose be replaced by
z = h
t
0.5x
pl
e (D11.12)
where h
t
is the thickness of the slab tested. The use of e in place of e
p
is because an
approximation to the mean resistance M
Rm
should over-estimate it. This moves curve FG
in Fig. 11.6 upwards. For a given test resistance M, following the route ABC then gives a
lower value for
test
, and, hence, a lower predicted
u
, from equation (B.2).
Curve FG is found by calculations for a set of values for that covers the range of
M/M
p, Rm
found in the tests. The mean bending resistance is given by
M
Rm
= M
pr
+ N
c, f
z (D11.13)
(based on equation (D9.5)). The reduced plastic resistance of the sheeting, M
pr
, which
equals N
p
z
p
in Fig. 11.5(c), is found from equation (9.6).
Calculations for the partial-interaction diagram and
u
For 0.2, equation (9.6) and equation (D11.13) give
M
Rm
= N
c, f
z + 1.25M
p, a
(1 ) (D11.14)
Assuming = 0.7, for example, equation (D11.12) gives for specimens 58
z = 120 0.7 8.5 30 = 84.0 mm
199
N
cf
= 431
431
65
e = 30
81.5
55
8.5
N
c
= 431
8.5h
z
N
p
z
p
N
p
N
c
(b)
(c) (a)
Fig. 11.5. Stress blocks for bending resistance of composite slab with partial interaction (dimensions
in mm)
211
12 May 2004 11:53:21
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
Fromthe tests,
122
M
p, a
= 5.65 kN m/m, with e
p
= 33 mm, so e e
p
= 3 mm, much less than
z. From equation (D11.14),
M
Rm
= 0.7 431 0.084 + 1.25 5.65 0.3 = 27.48 kN m/m
and
M
Rm
/M
p, Rm
= 27.48/35.1 = 0.783
For specimens 14, h
t
= 170 mm and M
p, Rm
= 56.7 kN m/m. For = 0.7,
z = 170 0.7 8.5 30 = 134 mm
M
Rm
= 0.7 431 0.134 + 1.25 5.65 0.3 = 42.5 kN m/m
M
Rm
/M
p, Rm
= 42.5/56.7 = 0.750
Similar calculations for other degrees of shear connection give curves DE for the short-
span slabs 58 and FG for slabs 14.
In clause B.3.6(2), the bending moment M is defined as being at the cross-section under
the point load, on the assumption that two-point loading is used in the tests. Here,
four-point loading was used (see Fig. 11.2). At failure, there was significant slip throughout
the length of 3L/8 between each inner point load and the nearer support, so, for these
tests, M was determined at an inner point load, and L
s
was taken as 3L/8.
The calculation of M
test
for specimen 5 is now explained. From Table 11.1, the maximum
load was 94.2 kN. This included 2.2 kNthat was, in effect, applied to the composite slab by
the removal of the prop that was present at mid-span during concreting.
122
The loads on
the composite member were thus as shown inset on Fig. 11.6, and the bending moment at
point J was
M
test
= 47.1 0.75 23.0 0.5 = 23.83 kN m
This is for a slab of width 0.915 m, so that
M
test
/M
p, Rm
= 23.83/(0.915 35.1) = 0.742
From Fig.11.6,
test
= 0.646.
200
0.9
0.8
0.7
0.6
0.9 0.8 0.7 0.6
G
A
D
F
E
M/M
p, Rm

B
h
test
= 0.83
h = N
c
/N
cf
C
0.86
0.742
h
test
= 0.646
Tests 1 to 4
Tests 5 to 8
46.0
47.1
H
2.2
250 500 250
J
Tests 5 to 8
Fig. 11.6. Partial-interaction diagram from tests (units: mm and kN)
212
12 May 2004 11:53:21
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:30:45
Corresponding results for tests 14 are given in Table 11.2.
Clause B.3.6(4) defines
u, Rk
as the 5% lower fractile, based on the results for tests 14.
There may be other evidence on the variance of such results. Here, it is assumed that this
enables a value 10% below the mean to be used, as for the mk method.
The values of for tests 14 are so close that
u
can be found from their mean value,
0.83. From clause B.3.6(2),
u
= N
c, f
/[b(L
s
+ L
0
)] (D11.15)
For tests 14, L
s
= 3 4.5/8 = 1.69 m. For all tests, N
cf
= 431 kN/m, L
0
= 0.1 m, and b is
taken as 1.0 m. Hence,
u
= 0.83 431/1790 = 0.200 N/mm
2
From clause B.3.6(6), with
VS
taken as 1.25,
u, Rd
= 0.9 0.200/1.25 = 0.144 N/mm
2
The interaction curve DE for specimens 58 is slightly higher than FG in Fig. 11.6. Its
value at = 0, M
p, a
/M
p, Rm
, is higher because, for these thinner slabs (h
t
= 120 mm), M
p, Rm
is lower, at 35 kN m/m. Using the preceding method for these results gives
u, Rd
= 0.24 N/mm
2
. This much higher result confirms the suspicion, noted above, that
longitudinal shear failure was not reached in specimens 14.
Comments
Where the test data are in accordance with the specification in Annex B, determination of
u, Rd
is straightforward, as values of can be found by replacing the graphical method
(used here for illustration) by direct calculation. However, where tests are being planned,
or other data are being used, as here, the work requires understanding of the basis of the
provisions of Annex B. It is of particular importance to ensure that longitudinal shear
failures occur in the tests.
201
Table 11.2. Degree of shear connection, from tests on composite slabs
Test No. Maximum load (kN) M
test
(kN m) M
test
/M
p, Rm

1 75.5 44.8 0.863 0.835
2 75.5 44.8 0.863 0.835
3 73.9 43.9 0.846 0.814
4 75.4 44.7 0.862 0.833
5 94.2 23.83 0.742 0.646
213
12 May 2004 11:53:21
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:06
APPENDIX A
Lateraltorsional buckling of
composite beams for buildings
This appendix supplements the comments on clause 6.4.
Simplified expression for cracked flexural stiffness of a
composite slab
The cracked stiffness per unit width of a composite slab is defined in clause 6.4.2(6)
as the lower of the values at mid-span and at a support. The latter usually governs, because
the profiled sheeting may be discontinuous at a support. It is now determined for the
cross-section shown in Fig. A.1 with the sheeting neglected.
It is assumed that only the concrete within the troughs is in compression. Its transformed
area in steel units is
A
e
= b
0
h
p
/nb
s
(a)
where n is the modular ratio. The position of the elastic neutral axis is defined by the
dimensions a and c, so that
A
e
c = A
s
a and a + c = z (b)
where A
s
is the area of top reinforcement per unit width of slab, and
z = h d
s
h
p
/2 (c)
h
p
/2
d
s
b
0
b
r
b
s
A
s
z
h
c
a
Fig. A.1. Model for stiffness of a composite slab in hogging bending
215
12 May 2004 11:53:21
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:06
Assuming that each trough is rectangular, the second moment of area per unit width is
I = A
s
a
2
+ A
e
(c
2
+ h
p
2
/12) (d)
Using equations (b) to (d), the flexural stiffness is
(EI)
2
= E
a
[A
s
A
e
z
2
/(A
s
+ A
e
) + A
e
h
p
2
/12] (DA.1)
This result is used in Example 6.7.
Flexural stiffness of beam with encased web
For a partially encased beam the model used for the derivation of equation (6.11) for the
flexural stiffness k
2
is as shown in Fig. A.2(a). A lateral force F applied to the steel bottom
flange causes displacement . The rotation of line AB is = /h
s
. It is caused by a bending
moment Fh
s
acting about A. The stiffness is
k
2
= M/ = Fh
s
2
/
The force F is assumed to be resisted by vertical tension in the steel web and compression in a
concrete strut BC, of width b
c
/4. Elastic analysis gives equation (6.11).
Maximum spacing of shear connectors for continuous U-frame
action
A rule given in ENV 1994-1-1 is derived. It is assumed that stud connectors are provided at
spacing s in a single row along the centre of the steel top flange (Fig. A.2(b)). The tendency
of the bottomflange to buckle laterally causes a transverse moment M
t
per unit length, which
is resisted by a tensile force T in each stud. From Fig. A.2,
M
t
s = 0.4bT (a)
The initial inclination from the vertical of the web,
0
in Fig. 6.9(b), due to the tendency of
the bottom flange to buckle sideways, would be resisted by a moment k
s
0
, from the
definition of k
s
in clause 6.4.2(6). For a design longitudinal moment M
Ed
at the adjacent
internal support,
0
is assumed to be increased to
204
0.4b
T
T
M
t
C
A
B
d
F
h
s
b
c
/4
b
c
/4
(a) (b)
Fig. A.2. Resistance to transverse bending in an inverted U-frame. (a) Flexural stiffness of encased
web. (b) Spacing of shear connectors
216
12 May 2004 11:53:22
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:06
0
[(M
Ed
/M
cr
)/(1 M
Ed
/M
cr
)]
where M
cr
is the elastic critical buckling moment. This deformation causes a transverse
bending moment per unit length
M
t
= k
s
0
[(M
Ed
/M
cr
)/(1 M
Ed
/M
cr
)] (b)
where k
s
is the stiffness defined in clause 6.4.2(6).
The design procedure of clause 6.4.2(1) is such that M
Ed

LT
M
Rd
. Here, M
Rd
is taken as
approximately equal to the characteristic resistance M
Rk
. From clause 6.4.2(4),
= M
Rk
/M
cr
, so that equation (b) becomes
M
t
= k
s
0
[(
LT
)/(1
LT
)] (c)
It is assumed that the resistance of the studs to longitudinal shear, P
Rd
, must not be
reduced, and that this is achieved if
T 0.1P
Rd
(d)
The initial slope
0
is taken as L/400 h, where h is the depth of the steel section. A typical
L/h ratio is 20, giving
0
= 0.05. From these results,
(DA.2)
This upper limit to the spacing of studs reduces as the slenderness increases.
It can be evaluated where the conditions of clause 6.4.3 for simplified verification are
satisfied, because the value = 0.4 can be assumed. From Table 6.5 of EN 1993-1-1,
buckling curve c should be used for rolled I-sections with a depth/breadth ratio exceeding
2.0. For = 0.4 it gives
LT
= 0.90. For a typical 19 mm stud, the resistance P
Rd
is about
75 kN. The combined stiffness of the slab and the web, k
s
, depends mainly on the stiffness of
the web, and is here taken as 0.9k
2
where k
2
, the stiffness of the web, is given by equation
(6.10) as
k
2
= E
a
t
w
3
/[4(1
a
2
)h
s
] (DA.3)
For a typical I-section, h
s
0.97 h. With E
a
= 210 kN/mm
2
and = 0.3, substitution into
equation (DA.2) gives
s 6.66(b/t
w
)(h/t
w
)(1/t
w
) (e)
For rolled sections, the closest stud spacing is thus required for relatively thick webs.
Examples are given in Table A.1. For studs in two rows, these spacings can be doubled,
because the assumed lever arm for the moment M
t
would increase from 0.4b (Fig. A.2) to
about 0.8b. For web-encased beams, ENV 1994-1-1 required the maximum spacings to be
halved.
This check is not required by EN 1994-1-1. The results show that it would not govern in
normal practice, but could do so where there was a need for wide spacing of studs (e.g.
because precast concrete floor slabs were being used) on a beam with a relatively thick web,
or where web encasement was used.
205
APPENDIX A. LATERALTORSIONAL BUCKLING OF COMPOSITE BEAMS FOR BUILDINGS
2
LT
2
LT
2
LT
LT
LT
LT
2
Rd LT LT
2
t s LT LT
0.04 (1 ) 0.4
0.05
P s T
b M k

-
=
Table A.1. Maximum spacings for 19 mm studs, and minimum top reinforcement
Serial size Mass (kg/m) Web thickness (mm) s
max
(mm) 100A
s, min
/d
s
762 267 UB 197 15.6 362 0.06
610 305 UB 238 18.6 204 0.12
610 229 UB 101 10.6 767 0.02
IPE 600 122 12.0 509 0.03
HEA 700 204 14.5 452 0.05
217
12 May 2004 11:53:22
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:06
Top transverse reinforcement above an edge beam
Where the concrete flange of a beam is continuous on one side only, as in Fig. 6.9(a), top
transverse reinforcement (AB) is required to prevent lateral buckling by anticlockwise
rotation of the steel section, in the plane of the diagram. The preceding results for spacing of
stud connectors can be used to estimate the amount required.
The reinforcement will be light, so the lever arm for transverse bending can be taken as
0.9d
s
(notation as in Fig. 6.9(a)), even where concrete in the lower half of the slab is present
only in the troughs of sheeting. From expression (d) above, the force T per unit length is
0.1P
Rd
/s; so from equation (a) the transverse bending moment is
M
t
= 0.4bT/s = 0.04bP
Rd
/s = A
s
f
sd
(0.9d
s
)
where A
s
is the area of top transverse reinforcement per unit length along the beam, at its
design yield stress f
sd
. Using expression (e) for s,
0.9A
s
f
sd
d
s
0.0060P
Rd
t
w
3
/h
Assuming P
Rd
= 75 kN and f
sd
= 500/1.15 = 435 N/mm
2
:
100A
s
/d
s
115t
w
3
/d
s
2
h
The area A
s
is thus greatest for a thin slab, so an effective depth d
s
= 100 mm is assumed,
giving
100A
s
/d
s
0.0115t
w
3
/h (f)
These values are given in the last rowof Table A.1. They showthat although top reinforcement
is required for U-frame action, the amount is small. Clause 6.6.5.3, on local reinforcement in
the slab, does not refer to this subject, and could be satisfied by bottom reinforcement only.
The requirements for minimum reinforcement of clause 9.2.1(4) and of EN 1992-1-1 could
also be satisfied by bottom reinforcement, whereas some should be placed near the upper
surface.
Derivation of the simplified expression for
LT
The notation is as in the comments on clause 6.4 and in this appendix, and is not redefined
here.
Repeating equation (D6.11):
M
cr
= (k
c
C
4
/L)[(G
a
I
at
+ k
s
L
2
/
2
)E
a
I
afz
]
1/2
(D6.11)
From clause 6.4.2(4),
= (M
Rk
/M
cr
)
0.5
(a)
It is on the safe side to neglect the term G
a
I
at
in equation (D6.11), which in practice is usually
less than 0.1k
s
L
2
/
2
. Hence,
M
cr
= (k
c
C
4
/)(k
s
E
a
I
afz
)
1/2
(b)
It is assumed that the stiffness of the concrete slab k
1
is at least 2.3 times the stiffness of the
steel web, k
2
. The combined stiffness k
s
, given by equation (6.8) in clause 6.4.2, then always
exceeds 0.7k
2
, so k
s
in equation (b) can be replaced by 0.7k
2
. This replacement would not be
valid for an encased web, so these are excluded.
For a steel flange of breadth b
f
and thickness t
f
,
I
afz
= b
f
3
t
f
/12 (c)
The stiffness k
2
is given by equation (6.10) in clause 6.4.2. Using it and the equations above,
(d)
206
LT
2
2 2
4 s a Rk
LT 2 3 3 2
c a w f f 4
48 (1 )
0.7
h M
k E t b t C
-
=

218
12 May 2004 11:53:22
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:06
For sections in Class 1 or Class 2, M
Rk
= M
pl, Rk
. It can be shown that M
pl, Rk
is given
approximately by
M
pl, Rk
= k
c
M
pl, a, Rk
(1 + t
w
h
s
/4b
f
t
f
) (e)
For double-symmetrical steel I-sections, the plastic resistance to bending is given
approximately by
M
pl, a, Rk
= f
y
h
s
b
f
t
f
(1 + t
w
h
s
/4b
f
t
f
) (f)
From equations (d) to (f), with
a
= 0.3,
(D6.14)
as given in Annex B of ENV 1994-1-1.
207
0.75 0.5 0.25
y
w s s f
LT
f f w f a 4
5.0 1
4
f
t h h t
b t t b E C

= +

4
3.0
y
yM
0
yM
0
M
0
is the bending moment at midspan
when both ends are simply supported
yM
0
yM
0
yM
0
yM
0
yM
0
y
y
yM
0
0.5yM
0
0.75yM
0
2.0 1.0 0.4
30
20
10
C
4
Fig. A.3. Values of the factor C
4
for uniformly distributed and centre point loading
219
12 May 2004 11:53:23
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:06
Effect of web encasement on
LT
The reduction in relative slenderness achieved by encasing a steel web to clause 5.5.3(2)
can be estimated as follows. The subscript e is used for properties of the section after
encasement.
From equations (6.10) and (6.11),
The modular ratio n rarely exceeds 12, and b
f
/t
w
is at least 15 for rolled or welded I-sections.
With these values, and
a
= 0.3,
k
2, e
/k
2
= 12.2
Assuming, as above, that k
1
> 2.3k
2
and using equation (6.8), the change in k
s
is
It was found above that k
s
can be replaced by 0.7k
2
, so k
s, e
is now replaced by
2.78 0.7k
2
= 1.95k
2
. Thus, the divisor 0.7 in equation (d) above is replaced by 1.95. Hence,
208
2 2
2, e a f
2
2 w f w
(1 )
4(1 4 / )
k b
k nt b t
-
=
+
s, e 2, e
1 2
s 1 2, e 2
12.2 (2.3 1)
2.78
2.3 12.2
k k
k k
k k k k
+ +
= > =
+ +
4
y
Uniform loading
No transverse loading
L
c
/L = 0.25
L
c
/L = 0.50
L
c
/L = 1.00
L
c
/L = 0.75
L
c
L
yM
0
yM
yM
M
0
M
M
1.0 0.5 0
30
20
10
C
4
Fig. A.4. Values of the factor C
4
for cantilevers, and for spans without transverse loading
220
12 May 2004 11:53:23
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:06
= (0.7/1.95)
0.25
= 0.77 (DA.4)
Factor C
4
for the distribution of bending moment
The tables that were given in ENV1994
49
relate to distributions of bending moment between
points at which the steel bottom flange is laterally restrained, not necessarily to complete
spans. The more commonly used values for distributed loading on internal spans are plotted
in Fig. A.3. For values of exceeding 3.0, values corresponding to can conservatively
be used. These are also shown.
The dashed lines in Fig. A.3 are for point loads at mid-span. Two other sets of values are
plotted in Fig. A.4. The solid lines apply for lateral buckling of a cantilever of length L
c
,
where both it and the adjacent span of length L have the same intensity of distributed
loading. The dashed lines are for an unloaded span with one or both ends continuous.
Criteria for verification of lateraltorsional stability without
direct calculation
Unlike UB steel sections, the basic sets of IPE and HEA sections have only one size for each
overall depth, h. Plots of their section properties F, from equation (D6.15), against h lie on
straight lines, as shown in Fig. A.5. This enables limits on F to be presented in Table 6.1
of clause 6.4.3 as limits to overall depth. From equation (D6.14), F
lim
for given is
proportional to f
y
0.5
. From this, and the qualifying sections, it can be deduced that the values
of F
lim
used in EN 1994-1-1 for the various grades of steel are as given in Table A.2.
209
LT
LT, e LT
/
300 800 700 600 500 400
HEA sections
900
F
Depth, h (mm)
8
14
12
10
16
457 152 UB F
lim
for uncased section, S 235
S 420, S 460
S 355
S 275
IPE sections
457 191 UB
533 210 UB
610 305 UB
610 229 UB
Does not qualify for
grades S 420 and S 460
Fig. A.5. Property F (equation (D6.15)) for some IPE, HEA and UB steel sections
221
12 May 2004 11:53:24
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:06
The IPE and HEA sections shown qualify for all steel grades that have F
lim
above their
plotted value of F. The only exception is HEA 550, which plots just below the S420 and S460
line, but does not qualify according to Table 6.1.
For UB sections, crosses in Fig. A.5 represent the 10 sections listed in Table 6.1. The
entries Yes in that table correspond to the condition F F
lim
. It is not possible to give a
qualifying condition in terms of depth only; equation (D6.15) should be used.
Web encasement
From equation (DA.4), the effect of web encasement is to increase F
lim
by a factor of at least
1/0.77 = 1.29. These values are given in Tables 6.1 and A.2. The additional depths permitted
by clause 6.4.3(1)(h) are a little more conservative than this result.
210
Table A.2. Limiting section parameter F
lim
, for uncased and web-encased sections
Nominal steel grade S235 S275 S355 S420 and S460
F
lim
, uncased 15.1 13.9 12.3 10.8
F
lim
, encased web 19.5 18.0 15.8 13.9
222
12 May 2004 11:53:24
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:08
APPENDIX B
The effect of slab thickness on
resistance of composite slabs to
longitudinal shear
Summary
A mechanical model based on ductile shear connection has been applied to the mk and
partial connection methods of Section 9 for design of composite slabs for longitudinal shear.
For the mk method it was shown
124
that:
where the assumptions of the model apply, the two sets of tests from which m and k are
derived can be done on sets of slabs of different thickness but similar concrete strength
two widely different shear spans should be used
predictions by the mk method of EN 1994-1-1 for degrees of shear connection between
those corresponding to the shear spans tested, are conservative
predictions for degrees of shear connection outside this range are unconservative
the percentage errors can be estimated.
For the partial-connection method it was found
124
that:
where tests are done on slabs of one thickness only, the model gives no help in predicting
the effect of slab thickness on ultimate shear strength
u
slabs of at least two different thicknesses should be tested, preferably with the same
shear span.
The model
The notation and assumptions are generally those of clauses 9.7.3 and B.3, to which reference
should be made.
Figure B.1 shows the left-hand shear span of a composite slab of breadth b and effective
depth d
p
, at failure in a test in accordance with clause B.3. The self-weight of the slab is
neglected in comparison with V
t
, the value of each of the two point loads at failure.
The shear connection is assumed to be ductile, as defined in clause 9.7.3(3), with ultimate
shear strength
u
, as would be found by the procedure of clause B.3.6 (except that all values
here are mean values, with no partial safety factors).
The sheeting and the slab are shown separated in Fig. B.1, and the longitudinal shear force
between them is
223
12 May 2004 11:53:24
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:08
N
c, f
=
u
bL
s
(DB.1)
where is the degree of shear connection, 1. The value of
u
is assumed to be independent
of the shear span, so that
L
s
= L
sf
(DB.2)
where L
sf
is the length of shear span at which the longitudinal force N
c, f
equals the tensile
strength of the sheeting, N
pl
. In the absence of a partial safety factor for strength of concrete,
the rectangular stress block is quite shallow, and is assumed to lie within the concrete slab. Its
depth is x
pl
, where x
pl
is the depth for full shear connection.
Let the plastic bending resistance of the sheeting be M
p, a
, reduced to M
pr
in the presence of
an axial force N, as shown by the stress blocks in Fig. B.1. The resistance M
pr
is assumed to be
given by
M
pr
= (1
2
)M
p, a
(DB.3)
The bilinear relationship given in clause 9.7.2(6) is an approximation to this equation, which
is also approximate, but accurate enough for the present work.
For equilibrium of the length L
s
of composite slab,
V
t
L
s
= N
cf
(d
p
x
pl
/2) + M
pr
Hence,
V
t
= [N
cf
(d
p
x
pl
/2) + (1
2
)M
p, a
]/(L
sf
) (DB.4)
For a typical profiled sheeting, M
p, a
0.3h
p
N
pl
. This is assumed here. The conclusions do not
depend on the accuracy of the factor 0.3. Hence,
V
t
= (N
cf
/L
sf
)[d
p

2
x
pl
/2 + 0.3 h
p
(1
2
)] (DB.5)
For a particular sheeting and strength of concrete, it may be assumed that N
cf
, L
sf
, x
pl
and h
p
are constant. The independent variables are the slab thickness, represented by d
p
, and the
shear span in a test, represented by the degree of shear connection, . The dependent
variable is the vertical shear resistance, V
t
.
The mk method
The use of test results as predictors
The properties m and k are determined from the graph shown in Fig. B.2, by drawing a line
through two test results. The line is
y = mx + k
For a single result, (x
1
, y
1
), say,
y
1
= mx
1
+ k
212
d
p
h
p
V
t
V
t
hN
cf
hN
cf
hN
cf
d
p
hx
pl
/2
hx
pl
L = hL
sf
M
pr
e
Fig. B.1. Shear span of composite slab, and stress blocks at failure in longitudinal shear
224
12 May 2004 11:53:24
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:08
From the definitions of x and y, shown in Fig. B.2, and for this test result,
V
t1
= bd
p1
(mA
p
/bL
s1
+ k)
This is equation (9.7) in clause 9.7.3(4).Thus, the mk method predicts exactly the first test
result. This is so also for the second test result, even if the slab thickness and shear span are
different. Two test results predict themselves. The basic assumption of the mk method is
that other results can be predicted by assuming a straight line through the two known results.
It has been shown that it is in fact curved, so that other predictions are subject to error.
Shape of function y(x)
The slope of the curve y(x) was found for a set of tests done with different shear spans and
constant slab thickness. It was shown, by differentiation of equation (DB.5), that the curve
through the two points found in tests is convex upwards, as shown in Fig. B.2.
For some degree of shear connection between those used in two tests on slabs of the same
thickness, from which the mk line was predicted, the method gives the result V
pred
, shown in
Fig. B.2. This is less than the resistance given by equation (DB.5), V
true
, so the method is safe,
according to the model.
For a degree of shear connection outside this range, the mk method is unsafe.
Estimate of errors of prediction
As an example, suppose that for a set of tests with d
p
/h
p
= 2.0 the sheeting and concrete are
such that when the shear span is L
sf
, the depth x
pl
of the concrete stress block is given by
x
pl
/h
p
= 0.4. Equation (DB.5) then becomes
V
t
= (N
c, f
h
p
/L
sf
)(0.3/ + 2 0.5) (DB.6)
Suppose that four otherwise identical tests are done with shear spans such that = 0.4, 0.5,
0.7 and 1.0. The (assumed) true results for V
t
L
sf
/N
c, f
h
p
are calculated and plotted against
1/. They lie on a convex-upwards curve, as expected. By drawing lines through any two of the
points, values for the other two tests predicted by the mk method are obtained. Comparison
with the plotted points gives the error from the mk method, as a percentage. Typical results
are given in Table B.1, for which the mk line is drawn through the results for
1
and
2
, and
used to predict the shear resistance for a slab with =
3
. The values in columns 4 and 5 of
the table are proportional to V
t
, so the percentage values are correct.
213
APPENDIX B. SLAB THICKNESS AND RESISTANCE TO LONGITUDINAL SHEAR
y = V
t
/bd
p
x = A
p
/bhL
sf
V
pred
V
true
0
Fig. B.2. Determination of m and k from two sets of test results
225
12 May 2004 11:53:25
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:08
Conclusion for the mk method
Two sets of tests should be done, with shear spans as widely different as possible, subject to
obtaining the correct failure mode. The slab thickness should be roughly central within the
range of application, and can differ in the two sets; but the concrete strengths should be the
same. The mk results are applicable within the range of shear spans tested, and probably for
a short distance outside it.
The partial-connection method
From Fig. B.1, the ultimate bending moment at the end of the shear span is
M = V
t
L
sf
(DB.7)
The overhang L
0
(clause B.3.6(2) and Fig. 9.2) is much less than L
sf
. For simplicity it is
assumed that L
sf
+ L
0
L
sf
, so that equation (B.2) in clause B.3.6 becomes
u
= N
c, f
/bL
sf
(DB.8)
From equations (DB.5) and (DB.7),
M/M
p, Rm
= (N
c, f
/M
p, Rm
)[d
p

2
x
pl
/2 + 0.3h
p
(1
2
)] (DB.9)
where M
p, Rm
is the plastic resistance moment with full shear connection.
For any assumed value for , an ultimate bending moment M can be calculated from
equation (DB.9). Thus, an M curve (Fig. B.5 of EN1994-1-1) can be found. If safety factors
are omitted, the same curve is used to find
test
from a measured value M
test
, and hence
u
from equation (DB.8).
Equation (DB.9) is independent of shear span because ductile behaviour is assumed. It
gives no information on the rate of change of with slab thickness or shear span, so a single
group of four tests gives no basis for predicting
u
for slabs of different thickness from those
tested.
Assuming that the fourth test, with a short shear span (clause B.3.2(7)), shows ductile
behaviour, the effect of thickness can be deduced from results of a further group of three
tests. The specimens and shear span should be identical with those in tests 13 in the first
group, except for slab thickness. The thicknesses for the two groups should be near the ends
of the range to be used in practice.
Let the ratios d
p
/h
p
for these two series be denoted v
1
and v
2
, with v
2
> v
1
, leading to the
corresponding degrees of shear connection
1
and
2
. It is likely that
2
>
1
, because the
longitudinal strain across the depth of the embossments will be more uniform in the thicker
slabs; but the difference may be small.
Assuming that over this range the relationship between test bending resistance M and
ratio v is linear, it can be shown
124
that the v curve is convex upwards. Hence, interpolation
for is conservative for slab thicknesses between those tested, and may be unconservative
outside this range.
Conclusion for the partial-connection method
For this method, further tests at constant thickness but different shear span would only
provide a check on the presence of ductile behaviour. Information on the effect of slab
214
Table B.1. Errors in prediction of V
t
by the mk method
1

2

3
Prediction from
mk line Plotted value
Error of prediction
(%)
0.4 1.0 0.7 2.00 2.08 4
0.4 0.5 1.0 1.98 1.80 +10
0.7 1.0 0.4 2.78 2.55 + 9
226
12 May 2004 11:53:25
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:08
thickness is best obtained from tests at constant shear span and two thicknesses. It can
be shown that values of degree of shear connection for intermediate thicknesses can be
obtained by linear interpolation between the values for the thicknesses tested.
215
APPENDIX B. SLAB THICKNESS AND RESISTANCE TO LONGITUDINAL SHEAR
227
12 May 2004 11:53:25
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:13
APPENDIX C
Simplified calculation method
for the interaction curve for
resistance of composite column
cross-sections to compression
and uniaxial bending
Scope and method
Equations are given for the coordinates of points B, C and D in Fig. 6.19, also shown in Fig.
C.1. They are applicable to cross-sections of columns where the structural steel, concrete and
reinforcement are all doubly symmetric about a single pair of axes. The steel section should
be an I- or H-section or a rectangular or circular hollow section. Examples are shown in
Fig. 6.17.
0 M
pl, Rd
M
M
max, Rd
N
A
N
pm, Rd
/2
N
pm, Rd
N
pl, Rd
B
C
D
Fig. C.1. Polygonal interaction curve
229
12 May 2004 11:53:25
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:13
Plastic analysis is used, with rectangular stress blocks for structural steel, reinforcement,
and concrete in accordance with clauses 6.7.3.2(2) to 6.7.3.2(6). For filled tubes of circular
section, the coefficient
c
in clause 6.7.3.2(6) has conservatively been taken as zero.
In this annex, the compressive stress in concrete in a rectangular stress block is denoted f
cc
,
where generally, f
cc
= 0.85f
cd
. However, for concrete-filled steel sections, the coefficient 0.85
may be replaced by 1.0, following clause 6.7.3.2(1).
Resistance to compression
The plastic resistance N
pl, Rd
is given by clause 6.7.3.2. The resistance N
pm, Rd
is calculated as
follows.
Figure C.2(a) represents a generalized cross-section of structural steel and reinforcement
(shaded area), and of concrete, symmetrical about two axes through its centre of area G.
For bending only (point B) the neutral axis is line BB which defines region (1) of the
cross-section, within which concrete is in compression. The line CC at the same distance h
n
on the other side of G is the neutral axis for point C in Fig. C.1. This is because the areas of
structural steel, concrete and reinforcement in region (2) are all symmetrical about G, so
that the changes of stress when the axis moves from BB to CC add up to the resistance
N
pm, Rd
, and the bending resistance is unchanged. Using subscripts 1 to 3 to indicate
regions (1) to (3),
N
pm, Rd
= R
c2
+ 2|R
a2
| (C.1)
where R
c2
is the resistance of the concrete in region (2), and R
a2
is the resistance of the steel in
region (2).
In the notation of clause 6.7.3.2(1),
R
c2
= A
c2
f
cc
R
a2
= A
a2
f
yd
+ A
s2
f
sd
where compressive forces and strengths f
cc
, f
sd
and f
yd
are taken as positive.
From symmetry,
R
a1
= |R
a3
|
R
c1
= R
c3
(C.2)
218
D
B
C
h
n
h
n
D
(1)
B
C
(1)
(2)
(b)
b
h
e
y
t
w
t
f
c
y
c
y
c
z
c
z
b
c
h
c
e
z
z
y
(a)
G
Fig. C.2. Composite cross-sections symmetrical about two axes
230
12 May 2004 11:53:26
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:13
When the neutral axis is at BB, N = 0, so that
R
a1
+ R
c1
= |R
a2
| + |R
a3
| (C.3)
From eqs (C.2) and (C.3), |R
a2
| = R
c1
= R
c3
. Substituting in equation(C.l),
N
pm, Rd
= R
c2
+ R
c1
+ R
c3
= R
c
(C.4)
where R
c
is the compressive resistance of the whole area of concrete, which is easily
calculated.
Position of neutral axis
Equations for h
n
depend on the axis of bending, the type of cross section and the cross section
properties. The equations are derived from equations (C.1) and (C.4), and are given below
for some cross sections.
Bending resistances
The axial resistance at point D in Fig. C.1 is half that at point C, so the neutral axis for point
D is line DD in Fig. C.2(a).
The bending resistance at point D is
M
max, Rd
= W
pa
f
yd
+ W
ps
f
sd
+ W
pc
f
cc
/2 (C.5)
where W
pa
, W
ps
and W
pc
are the plastic section moduli for the structural steel, the reinforcement
and the concrete part of the section (for the calculation of W
pc
the concrete is assumed to
be uncracked), and f
yd
, f
sd
and f
cc
are the design strengths for the structural steel, the
reinforcement and the concrete.
The bending resistance at point B is
M
pl, Rd
= M
max, Rd
M
n, Rd
(C.6)
with
M
n, Rd
= W
pa, n
f
yd
+ W
ps, n
f
sd
+ W
pc, n
f
cc
/2 (C.7)
where W
pa, n
, W
ps, n
and W
pc, n
are the plastic section moduli for the structural steel, the
reinforcement and the concrete parts of the section within region (2) of Fig. C.2(a).
Equations for the plastic section moduli of some cross-sections are given below.
Interaction with transverse shear
If the shear force to be resisted by the structural steel is considered according to clause
6.7.3.2(4) the appropriate areas of steel should be assumed to resist shear alone. The method
given here can be applied using the remaining areas.
Neutral axes and plastic section moduli of some cross-sections
General
The compressive resistance of the whole area of concrete is
N
pm, Rd
= A
c
f
cc
(C.8)
The value of the plastic section modulus of the total reinforcement is given by
219
APPENDIX C. RESISTANCE OF COMPOSITE COLUMNS TO COMPRESSION AND UNIAXIAL BENDING
231
12 May 2004 11:53:26
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:13
(C.9)
where e
i
are the distances of the reinforcement bars of area A
s, i
to the relevant middle line
(y-axis or z-axis).
The equations for the position of the neutral axis h
n
are given for selected positions in the
cross-sections. The resulting value h
n
should lie within the limits of the assumed region.
Major-axis bending of encased I-sections
The plastic section modulus of the structural steel may be taken fromtables, or be calculated
from
(C.10)
and
(C.11)
For the different positions of the neutral axes, h
n
and W
pa, n
are given by:
(a) Neutral axis in the web, h
n
h/2 t
f
:
(C.12)
W
pa, n
= t
w
h
n
2
(C.13)
where A
sn
is the sum of the area of reinforcing bars within the region of depth 2h
n
.
(b) Neutral axis in the flange, h/2 t
f
< h
n
< h/2:
(C.14)
(C.15)
(c) Neutral axis outside the steel section, h/2 h
n
h
c
/2:
(C.16)
W
pa, n
= W
pa
(C.17)
The plastic modulus of the concrete in the region of depth from 2h
n
then results from
W
pc, n
= b
c
h
n
2
W
pa, n
W
ps, n
(C.18)
with
(C.19)
where A
sn, i
are the areas of reinforcing bars within the region of depth 2h
n
, and e
z, i
are the
distances from the middle line.
Minor-axis bending of encased I-sections
The notation is given in Fig. C.2(b).
The plastic section modulus of the structural steel may be taken from tables or be
calculated from
220
2
f w
pa f f
( 2 )
( )
4
h t t
W bt h t
-
= + -
2
c c
pc pa ps
4
b h
W W W = - -
pm, Rd sn sd cc
n
c cc w yd cc
(2 )
2 2 (2 )
N A f f
h
b f t f f
- -
=
+ -
pm, Rd sn sd cc w f yd cc
n
c cc yd cc
(2 ) ( )( 2 )(2 )
2 2 (2 )
N A f f b t h t f f
h
b f b f f
- - + - - -
=
+ -
2
2 w f
pa, n n
( )( 2 )
4
b t h t
W bh
- -
= -
pm, Rd sn sd cc a yd cc
n
c cc
(2 ) (2 )
2
N A f f A f f
h
b f
- - - -
=
ps, n sn, z,
1
| |
n
i i
i
W A e
=
=
ps s,
1
| |
n
i i
i
W A e
=
=
232
12 May 2004 11:53:26
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:13
(C.20)
and
(C.21)
For the different positions of the neutral axes, h
n
and W
pa, n
are given by:
(a) Neutral axis in the web, h
n
t
w
/2:
(C.22)
W
pa, n
= hh
n
2
(C.23)
(b) Neutral axis in the flanges, t
w
/2 < h
n
< b/2:
(C.24)
(C.25)
(c) Neutral axis outside the steel section, b/2 h
n
b
c
/2
(C.26)
W
pa, n
= W
pa
(C.27)
The plastic modulus of the concrete in the region of depth 2h
n
then results from
W
pc, n
= h
c
h
n
2
W
pa, n
W
ps, n
(C.28)
with W
ps, n
according to equation (C.19), changing the subscript z to y.
Concrete-filled circular and rectangular hollow sections
The following equations are derived for rectangular hollow sections with bending about the
y-axis of the section (see Fig. C.3). For bending about the z-axis the dimensions h and b are to
be exchanged as well as the subscripts z and y. Equations (C.29) to (C.33) may be used for
circular hollow sections with good approximation by substituting
h = b = d and r = d/2 t
(C.29)
with W
ps
according to equation (C.9).
W
pa
may be taken from tables, or be calculated from
(C.30)
(C.31)
W
pc, n
= (b 2t)h
n
2
W
ps, n
(C.32)
W
pa, n
= bh
n
2
W
pc, n
W
ps, n
(C.33)
with W
ps, n
according to equation (C.19).
221
pm, Rd sn sd cc
n
c cc yd cc
(2 )
2 2 (2 )
N A f f
h
h f h f f
- -
=
+ -
pm, Rd sn sd cc w f yd cc
n
c cc f yd cc
(2 ) (2 )(2 )
2 4 (2 )
N A f f t t h f f
h
h f t f f
- - + - -
=
+ -
2
3 2
pa pc ps
2
( ) ( ) (4 )(0.5 )
4 3
bh
W r t r t h t r W W = - + - + - - - - -
2
3 2
pc ps
( 2 )( 2 ) 2
(4 )(0.5 )
4 3
b t h t
W r r h t r W
- -
= - - - - - -
pm, Rd sn sd cc a yd cc
n
c cc
(2 ) (2 )
2
N A f f A f f
h
h f
- - - -
=
2
2 f w
pa, n f n
( 2 )
2
4
h t t
W t h
-
= +
pm, Rd sn sd cc
n
cc yd cc
(2 )
2 4 (2 )
N A f f
h
bf t f f
- -
=
+ -
2
c c
pc pa ps
4
h b
W W W = - -
2 2
f w f
pa
( 2 ) 2
4 4
h t t t b
W
-
= +
233
12 May 2004 11:53:27
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:13
Example C.1: NMinteraction polygon for a column cross-section
The method of Appendix C is used to obtain the interaction polygon given in Fig. 6.38 for
the concrete-encased H section shown in Fig. 6.37. The small area of longitudinal
reinforcement is neglected. The data and symbols are as in Example 6.10 and Figs 6.37,
C.1 and C.2.
Design strengths of the materials: f
yd
= 355 N/mm
2
; f
cd
= 16.7 N/mm
2
.
Other data: A
a
= 11 400 mm
2
; A
c
= 148 600 mm
2
; t
f
= 17.3 mm; t
w
= 10.5 mm; b
c
= h
c
=
400 mm; b = 256 mm; h = 260 mm; 10
6
W
pa, y
= 1.228 mm
3
; 10
6
W
pa, z
= 0.575 mm
3
; N
pl, Rd
=
6156 kN.
Major-axis bending
From equation (C.8),
N
pm, Rd
= 148.6 16.7 = 2482 kN
h
n
= 2482/[0.8 16.7 + 0.021 (710 16.7)] = 89 mm
so the neutral axis is in the web (Fig. C.4(a)), as assumed. From equation (C.11), the
plastic section modulus for the whole area of concrete is
10
6
W
pc
= 4
3
/4 1.228 = 14.77 mm
3
10
6
W
pa, n
= 10.5 0.089
2
= 0.083 mm
3
10
6
W
pc, n
= 400 0.089
2
0.083 = 3.085 mm
3
M
max, Rd
= 1.228 355 + 14.77 16.7/2 = 559 kN m
From equations (C.6) and (C.7),
M
pl, Rd
= 559 (0.083 355 + 3.085 16.7/2) = 504 kN m
The results shown above in bold type are plotted on Fig. 6.38.
222
r
b
(a)
h
y
z
e
y
e
z
e
z
e
y
t
(b)
y
z
t
Fig. C.3. Concrete-filled (a) rectangular and (b) circular hollow sections, with notation
234
12 May 2004 11:53:27
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:13
Minor-axis bending
From equation (C.4), N
pm, Rd
is the same for both axes of bending. Thus,
N
pm, Rd
= 2482 kN
Assuming that the neutral axis BB intersects the flanges, from equation (C.24),
h
n
= [2482 0.0105 (260 34.6) (710 16.7)]/[0.8 16.7
+ 0.0692 (710 16.7)] = 13.7 mm
so axis BB does intersect the flanges (Fig. C.4(b)). From equation (C.21),
10
6
W
pc
= 4
3
/4 0.575 = 15.42 mm
3
10
6
W
pa, n
= 34.6 0.0137
2
+ 0.0105
2
(260 34.6)/4 = 0.0127 mm
3
10
6
W
pc, n
= 400 0.0137
2
0.0127 = 0.0624 mm
3
M
max, Rd
= 0.575 355 + 15.42 16.7/2 = 333 kN m
M
n, Rd
= 0.0127 355 + 0.0624 16.7/2 = 5.03 kN m
M
pl, Rd
= 333 5 = 328 kN m
These results are plotted on Fig. 6.38, and used in Example 6.10.
223
89
B
D
B
89
C
D
C
(a)
B
D
B
13.7
C D
C
13.7
(b)
Fig. C.4. Neutral axes at points B, C and D on the interaction polygons
235
12 May 2004 11:53:27
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:32
References
1. British Standards Institution (2004) Design of Composite Steel and Concrete Structures.
Part 1.1: General Rules and Rules for Buildings. BSI, London, BS EN 1994 (in
preparation).
2. Gulvanessian, H., Calgaro, J.-A. and Holick, M. (2002) Designers Guide to EN 1990.
Eurocode: Basis of Structural Design. Thomas Telford, London.
3. British Standards Institution (2002) Eurocode: Basis of Structural Design. BSI, London,
BS EN 1990.
4. British Standards Institution (2004) Design of Concrete Structures. Part 1.1: General
Rules and Rules for Buildings. BSI, London, BS EN 1992 (in preparation).
5. British Standards Institution (2004) Design of Steel Structures. Part 1.1: General Rules
and Rules for Buildings. BSI, London, BS EN 1993 (in preparation).
6. Beeby, A. W. and Narayanan, R. S. (2004) Designers Guide to EN 1992. Eurocode 2:
Design of Concrete Structures. Part 1.1: General Rules and Rules for Buildings. Thomas
Telford, London (in preparation).
7. Nethercot, D. and Gardner, L. (2004) Designers Guide to EN 1993. Eurocode 3: Design
of Steel Structures. Part 1.1: General Rules and Rules for Buildings. Thomas Telford,
London (in preparation).
8. European Commission (2002) Guidance Paper L (Concerning the Construction Products
Directive 89/106/EEC). Application and Use of Eurocodes. EC, Brussels.
9. British Standards Institution (2002) Actions on Structures. Part 1.1: Densities, Self Weight
and Imposed Loads. BSI, London, BS EN1991 [Parts 1.2, 1.3, etc., specify other types of
action].
10. British Standards Institution (2004) Design of Structures for Earthquake Resistance. BSI,
London, BS EN 1998 (in preparation).
Part 2: Bridges. BSI, London, BS EN 1994 (in preparation).
12. International Organization for Standardization (1997) Basis of Design for Structures
Notation General Symbols. ISO, Geneva, ISO 3898.
13. European Commission (1989) Construction Products Directive 89/106/EEC. EC, Brussels,
OJEC No. L40, 11 Feb.
14. British Standards Institution (2004) Geotechnical Design. BSI, London, BS EN 1997 (in
preparation).
15. Johnson, R. P. and Huang, D. J. (1994) Calibration of safety factors
M
for composite
steel and concrete beams in bending. Proceedings of the Institution of Civil Engineers:
Structures and Buildings, 104, 193203.
16. Johnson, R. P. and Huang, D. J. (1997) Statistical calibration of safety factors for
encased composite columns. In: C. D. Buckner and B. M. Sharooz (eds), Composite
Construction in Steel and Concrete III. American Society of Civil Engineers, New York,
pp. 380391.
237
12 May 2004 11:53:27
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:32
17. British Standards Institution (1997) Structural Use of Concrete. Part 1: Code of Practice
for Design and Construction. BSI, London, BS 8110.
18. Stark, J. W. B. (1984) Rectangular stress block for concrete. Technical paper S16, June.
Drafting Committee for Eurocode 4 (unpublished).
19. Anderson, D., Aribert, J.-M., Bode, H. and Kronenburger, H. J. (2000) Design rotation
capacity of composite joints. Structural Engineer, 78, 2529.
20. Morino, S. (2002) Recent developments on concrete-filled steel tube members in
Japan. In: J. F. Hajjar, M. Hosain, W. S. Easterling and B. M. Shahrooz (eds),
Composite Construction in Steel and Concrete IV. American Society of Civil Engineers,
New York, pp. 644655.
21. Wakabayashi, M. and Minami, K. (1990) Application of high strength steel to composite
structures. Symposium on Mixed Structures, including New Materials, Brussels. IABSE,
Zurich. Reports, 60, 5964.
22. Hegger, J. and Dinghaus, P. (2002) High performance steel and high performance
concrete in composite structures. In: J. F. Hajjar, M. Hosain, W. S. Easterling and
B. M. Shahrooz (eds), Composite Construction in Steel and Concrete IV. American
Society of Civil Engineers, New York, pp. 891902.
23. Hoffmeister, B., Sedlacek, G., Mller, Ch. and Khn, B. (2002) High strength materials
in composite structures. In: J. F. Hajjar, M. Hosain, W. S. Easterling and B. M.
Shahrooz (eds), Composite Construction in Steel and Concrete IV. American Society of
Civil Engineers, New York, pp. 903914.
24. British Standards Institution (2004) Design of Steel Structures. Part 1-8: Design of Joints.
BSI, London, BS EN 1993 (in preparation).
25. British Standards Institution (2004) Design of Steel Structures. Part 1-3: Cold Formed
Thin Gauge Members and Sheeting. BSI, London, BS EN 1993 (in preparation).
26. British Standards Institution (1998) Welding Studs and Ceramic Ferrules for Arc Stud
Welding. BSI, London, BS EN 13918.
27. Trahair, N. S., Bradford, M. A. and Nethercot, D. A. (2001) The Behaviour and Design
of Steel Structures to BS 5950, 3rd edn. Spon, London.
28. Johnson, R. P. and Chen, S. (1991) Local buckling and moment redistribution in Class 2
composite beams. Structural Engineering International, 1, 2734.
29. Johnson, R. P. and Fan, C. K. R. (1988) Strength of continuous beams designed to
Eurocode 4. Proceedings of the IABSE. IABSE, Zurich, P-125/88, pp. 3344.
30. British Standards Institution (2004) Design of Steel Structures. Part 1-5: Plated Structural
Elements. BSI, London, BS EN 1993 (in preparation).
31. British Standards Institution (1990) Code of Practice for Design of Simple and Continuous
Composite Beams. BSI, London, BS 5950-3-1.
32. Haensel, J. (1975) Effects of creep and shrinkage in composite construction. Report
75-12. Institute for Structural Engineering, Ruhr-Universitt, Bochum.
33. Johnson, R. P. and Hanswille, G. (1998) Analyses for creep of continuous steel and
composite bridge beams, according to EC4: Part 2. Structural Engineer, 76, 294298.
34. Johnson, R. P. (1987) Shrinkage-induced curvature in cracked composite flanges of
composite beams. Structural Engineer, 65B, 7277.
35. British Standards Institution (2004) Actions on Structures. Part 1-5: Thermal Actions.
BSI, London, BS EN 1991 (in preparation).
36. British Standards Institution (2002) Eurocode: Basis of Structural Design. Annex A1:
Application for Buildings. BSI, London, BS EN 1990.
37. Johnson, R. P. and Buckby, R. J. (1986) Composite Structures of Steel and Concrete, 2nd
edn, Vol. 2. Bridges. Collins, London.
38. Johnson, R. P. and Huang, D. J. (1995) Composite bridge beams of mixed-class
cross-section. Structutural Engineering International, 5, 96101.
39. ECCS TC11 (1999) Design of Composite Joints for Buildings. European Convention for
Constructional Steelwork, Brussels. Reports, 109.
226
238
12 May 2004 11:53:28
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:32
40. Johnson, R. P. (2003) Cracking in concrete tension flanges of composite T-beams
tests and Eurocode 4. Structural Engineer, 81, 2934.
41. Ahmed, M. and Hosain, M. (1991) Recent research on stub-girder floor systems. In:
S. C. Lee (ed.), Composite Steel Structures. Elsevier, Amsterdam, pp. 123132.
42. Lawson, R. M. (1987) Design for Openings in the Webs of Composite Beams. Steel
Construction Institute, Ascot, Publication 068.
43. Lawson, R. M., Chung, K. F. and Price, A. M. (1992) Tests on composite beams with
large web openings. Structural Engineer, 70, 17.
44. Johnson, R. P. and Anderson, D. (1993) Designers Handbook to Eurocode 4. Part 1:
Basis of Design. Thomas Telford, London.
45. Ansourian, P. (1982) Plastic rotation of composite beams. Journal of the Structural
Division of the American Society of Civil Engineers, 108, 643659.
46. Johnson, R. P. and Hope-Gill, M. (1976) Applicability of simple plastic theory to
continuous composite beams. Proceedings of the Institution of Civil Engineers, Part 2, 61,
127143.
47. Johnson, R. P. and Molenstra, N. (1991) Partial shear connection in composite beams
for buildings. Proceedings of the Institution of Civil Engineers, Part 2, 91, 679704.
48. Aribert, J. M. (1990) Design of composite beams with a partial shear connection [in
French]. Symposium on Mixed Structures, including New Materials, Brussels. IABSE,
Zurich. Reports, 60, 215220.
Part 1.1: General Rules and Rules for Buildings. BSI, London, DD ENV 1994.
50. Johnson, R. P. and Willmington, R. T. (1972) Vertical shear in continuous composite
beams. Proceedings of the Institution of Civil Engineers, 53, Sept., 189205.
51. Allison, R. W., Johnson, R. P. and May, I. M. (1982) Tension-field action in composite
plate girders. Proceedings of the Institution of Civil Engineers, Part 2, 73, 255276.
Part 12: Structural Fire Design. BSI, London, BS EN 1994 (in preparation).
53. Andr, H.-P. (1990) Economical shear connection with high fatigue strength. Symposium
on Mixed Structures, including New Materials, Brussels. IABSE, Zurich. Reports, 60, 167
172.
54. Studnicka, J., Machacek, J., Krpata, A. and Svitakova, M. (2002) Perforated shear
connector for composite steel and concrete beams. In: J. F. Hajjar, M. Hosain, W. S.
Easterling and B. M. Shahrooz (eds), Composite Construction in Steel and Concrete IV.
American Society of Civil Engineers, New York, pp. 367378.
55. Lindner, J. and Budassis, N. (2002) Lateral distortional buckling of partially encased
composite beams without concrete slab. In: J. F. Hajjar, M. Hosain, W. S. Easterling
and B. M. Shahrooz (eds), Composite Construction in Steel and Concrete IV. American
56. Johnson, R. P. and Fan, C. K. R. (1991) Distortional lateral buckling of continuous
composite beams. Proceedings of the Institution of Civil Engineers, Part 2, 91, 131161.
57. British Standards Institution (2000) Design of Steel Bridges. BSI, London, BS 5400:
Part 3.
58. Johnson, R. P. and Molenstra, N. (1990) Strength and stiffness of shear connections for
discrete U-frame action in composite plate girders. Structural Engineer, 68, 386392.
59. Hanswille, G. (2002) Lateral torsional buckling of composite beams. Comparison of
more accurate methods with Eurocode 4. In: J. F. Hajjar, M. Hosain, W. S. Easterling
and B. M. Shahrooz (eds), Composite Construction in Steel and Concrete IV. American
60. Hanswille, G., Lindner, J. and Mnich, N. D. (1998) Zum Biegedrillknicken von
Stahlverbundtrgern. Stahlbau, 7.
61. Roik, K., Hanswille, G. and Kina, J. (1990) Background to Eurocode 4 clause 4.6.2 and
Annex B. University of Bochum, Bochum, Report RSII 2-674102-88.17.
227
REFERENCES
239
12 May 2004 11:53:28
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:32
62. Roik, K., Hanswille, G. and Kina, J. (1990) Solution for the lateral torsional buckling
problem of composite beams [in German]. Stahlbau, 59, 327332.
63. Lawson, M. and Rackham, J. W. (1989) Design of Haunched Composite Beams in
Buildings. Steel Construction Institute, Ascot, Publication 060.
64. British Standards Institution (2004) Design of Steel Structures. Part 2: Bridges. BSI,
London, BS EN 1993 (in preparation).
65. British Standards Institution (2000) Code of Practice for Design in Simple and Continuous
Construction: Hot Rolled Sections. BSI, London, BS 5950-1.
66. Johnson, R. P. and Chen, S. (1993) Stability of continuous composite plate girders with
U-frame action. Proceedings of the Institution of Civil Engineers: Structures and Buildings,
99, 187197.
67. Johnson, R. P. and Oehlers, D. J. (1981) Analysis and design for longitudinal shear in
composite T-beams. Proceedings of the Institution of Civil Engineers, Part 2, 71, 9891021.
68. Li, A. and Cederwall, K. (1991) Push Tests on Stud Connectors in Normal and High-
strength Concrete. Chalmers Institute of Technology, Gothenburg, Report 91:6.
69. Mottram, J. T. and Johnson, R. P. (1990) Push tests on studs welded through profiled
steel sheeting. Structural Engineer, 68, 187193.
70. Oehlers, D. J. and Johnson, R. P. (1987) The strength of stud shear connections in
composite beams. Structural Engineer, 65B, 4448.
71. Roik, K., Hanswille, G. and Cunze Oliveira Lanna, A. (1989) Eurocode 4, Clause 6.3.2:
Stud Connectors. University of Bochum, Bochum, Report EC4/8/88.
72. Stark, J. W. B. and van Hove, B. W. E. M. (1991) Statistical Analysis of Pushout Tests
on Stud Connectors in Composite Steel and Concrete Structures. TNO Building and
Construction Research, Delft, Report BI-91-163.
73. British Standards Institution (1998) Welding Arc Stud Welding of Metallic Materials.
BSI, London, BS EN ISO 14555.
74. Kuhlmann, U. and Breuninger, U. (2002) Behaviour of horizontally lying studs with
longitudinal shear force. In: J. F. Hajjar, M. Hosain, W. S. Easterling and B. M.
75. Grant, J. A., Fisher, J. W. and Slutter, R. G. (1977) Composite beams with formed
metal deck. Engineering Journal of the American Institute of Steel Construction, 1, 2742.
76. Johnson, R. P. and Yuan, H. (1998) Existing rules and new tests for studs in troughs
of profiled sheeting. Proceedings of the Institution of Civil Engineers: Structures and
Buildings, 128, 244251.
77. Johnson, R. P. and Yuan, H. (1998) Models and design rules for studs in troughs
of profiled sheeting. Proceedings of the Institution of Civil Engineers: Structures and
Buildings, 128, 252263.
78. Lawson, R. M. (1993) Shear connection in composite beams. In: W. S. Easterling and
W. M. K. Roddis (eds), Composite Construction in Steel and Concrete II. American
79. Johnson, R. P. and Oehlers, D. (1982) Design for longitudinal shear in composite
L-beams. Proceedings of the Institution of Civil Engineers, Part 2, 73, 147170.
80. Najafi, A. A. (1992) End-plate connections. PhDthesis, University of Warwick, Warwick.
81. Johnson, R. P. (2004) Composite Structures of Steel and Concrete: Beams, Columns,
Frames, and Applications in Building, 3rd edn. Blackwell, Oxford.
82. British Standards Institution (1987) Design of Composite Bridges. BSI, London, BS 5400:
Part 5.
83. Lam, D., Elliott, K. S. and Nethercot, D. (2000) Designing composite steel beams with
precast concrete hollow core slabs. Proceedings of the Institution of Civil Engineers:
Structures and Buildings, 140, 139149.
84. Roik, K. and Bergmann, R. (1990) Design methods for composite columns with
unsymmetrical cross-sections. Journal of Constructional Steelwork Research, 15, 153168.
228
240
12 May 2004 11:53:28
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:32
85. Wheeler, A. T. and Bridge, R. Q. (2002) Thin-walled steel tubes filled with high
strength concrete in bending. In: J. F. Hajjar, M. Hosain, W. S. Easterling and B. M.
86. Kilpatrick, A. and Rangan, V. (1999) Tests on high-strength concrete-filled tubular
steel columns. ACI Structural Journal, 96-S29, 268274.
87. May, I. M. and Johnson, R. P. (1978) Inelastic analysis of biaxially restrained columns.
Proceedings of the Institution of Civil Engineers, Part 2, 65, 323337.
88. Roik, K. and Bergmann, R. (1992) Composite columns. In: P. J. Dowling, J. L. Harding
and R. Bjorhovde (eds), Constructional Steel Design An International Guide. Elsevier,
London, pp. 443469.
89. British Standards Institution (2004) Design of Steel Structures. Part 1-9: Fatigue Strength
of Steel Structures. BSI, London, BS EN 1993 (in preparation).
90. Atkins, W. S. and Partners (2004) Designers Guide to EN 1994. Eurocode 4: Design of
Composite Structures. Part 2: Bridges. Thomas Telford, London (in preparation).
91. Johnson, R. P. (2000) Resistance of stud shear connectors to fatigue. Journal of
Constructional Steel Research, 56, 101116.
92. Oehlers, D. J. and Bradford, M. (1995) Composite Steel and Concrete Structural Members
Fundamental Behaviour. Elsevier, Oxford.
93. Gomez Navarro, M. (2002) Influence of concrete cracking on the serviceability limit
state design of steel-reinforced concrete composite bridges: tests and models. In: J.
Martinez Calzon (ed.), Composite Bridges Proceedings of the 3rd International Meeting.
Spanish Society of Civil Engineers, Madrid, pp. 261278.
94. Johnson, R. P. and May, I. M. (1975) Partial-interaction design of composite beams.
Structural Engineer, 53, 305311.
95. Stark, J. W. B. and van Hove, B. W. E. M. (1990) The Midspan Deflection of Composite
Steel-and-concrete Beams under Static Loading at Serviceability Limit State. TNO Building
and Construction Research, Delft, Report BI-90-033.
96. British Standards Institution (2004) Draft National Annex to BS EN 1990: Eurocode:
Basis of Structural Design. BSI, London (in preparation).
97. British Standards Institution (1992) Guide to Evaluation of Human Exposure to Vibration
in Buildings. BSI, London, BS 6472.
98. Wyatt, T. A. (1989) Design Guide on the Vibration of Floors. Steel Construction Institute,
Ascot, Publication 076.
99. Randl, E. and Johnson, R. P. (1982) Widths of initial cracks in concrete tension flanges
of composite beams. Proceedings of the IABSE, P-54/82, 6980.
100. Johnson, R. P. and Allison, R. W. (1983) Cracking in concrete tension flanges of
composite T-beams. Structural Engineer, 61B, 916.
101. Johnson, R. P. (2003) Cracking in concrete flanges of composite T-beams tests and
Eurocode 4. Structural Engineer, 81, 2934.
102. Roik, K., Hanswille, G. and Cunze Oliveira Lanna, A. (1989) Report on Eurocode 4,
Clause 5.3, Cracking of Concrete. University of Bochum, Bochum, Report EC4/4/88.
103. Moore, D. B. (2004) Designers Guide to EN1993. Eurocode 3: Design of Steel Structures.
Part 1-8: Design of Joints. Thomas Telford, London (in preparation).
104. Couchman, G. and Way, A. (1998) Joints in Steel Construction Composite Connections.
Steel Construction Institute, Ascot, Publication 213.
105. Bose, B. and Hughes, A. F. (1995) Verifying the performance of standard ductile
connections for semi-continuous steel frames. Proceedings of the Institution of Civil
Engineers: Structures and Buildings, 110, November, 441457.
106. Nethercot, D. A. and Lawson, M. (1992) Lateral Stability of Steel Beams and Columns
Common Cases of Restraint. Steel Construction Institute, Ascot, Publication 093.
107. British Standards Institution (1994) Code of Practice for Design of Floors with Profiled
Steel Sheeting. BSI, London, BS 5950-4.
229
REFERENCES
241
12 May 2004 11:53:28
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:32
108. British Standards Institution (2004) Actions on Structures. Part 1-6: Actions During
Execution. BSI, London, BS EN 1991 (in preparation).
109. Couchman, G. H. and Mullett, D. L. (2000) Composite Slabs and Beams Using Steel
Decking: Best Practice for Design and Construction. Steel Construction Institute, Ascot,
Publication 300.
110. Bryan, E. R. and Leach, P. (1984) Design of Profiled Sheeting as Permanent Formwork.
Construction Industry Research and Information Association, London, Technical
Note 116.
111. Stark, J. W. B. and Brekelmans, J. W. P. M. (1990) Plastic design of continuous
composite slabs. Journal of Constructional Steel Research, 15, 2347.
112. ECCS Working Group 7.6 (1998) Longitudinal Shear Resistance of Composite Slabs:
Evaluation of Existing Tests. European Convention for Constructional Steelwork,
Brussels. Reports, 106.
113. Bode, H. and Storck, I. (1990) Background Report to Eurocode 4 (Continuation of Report
EC4/7/88). Chapter 10 and Section 10.3: Composite Floors with Profiled Steel Sheet.
University of Kaiserslautern, Kaiserslautern.
114. Bode, H. and Sauerborn, I. (1991) Partial shear connection design of composite
slabs. In: Proceedings of the 3rd International Conference. Association for International
CooperationandResearchinSteelConcrete Composite Structures, Sydney, pp. 467472.
115. British Standards Institution (1997) Falsework. Performance Requirements and General
Design. Standard 97/102975DC. BSI, London, prEN 12812.
116. Huber, G. (1999) Non-linear Calculations of Composite Sections and Semi-continuous
Joints. Ernst, Berlin.
117. COST-C1 (1997) Composite Steelconcrete Joints in Braced Frames for Buildings. Report:
Semi-rigid Behaviour of Civil Engineering Structural Connections. Office for Official
Publications of the European Communities, Luxembourg.
118. Aribert, J. M. (1999) Theoretical solutions relating to partial shear connection of steel
concrete composite beams and joints. In: Proceedings of the International Conference
on Steel and Composite Structures. TNO Building and Construction Research, Delft,
7.1-7.16.
119. Patrick, M. (1990) A new partial shear connection strength model for composite slabs.
Steel Construction Journal, 24, 217.
120. Oehlers, D. J. (1989) Splitting induced by shear connectors in composite beams. Journal
of the Structural Division of the American Society of Civil Engineers, 115, 341362.
121. British Standards Institution (2001) Tensile Testing of Metallic Materials. Part 1: Method
of Test at Ambient Temperature. BSI, London, BS EN 10002.
122. van Hove, B. W. E. M. (1991) Experimental Research on the CF70/0.9 Composite Slab.
TNO Building and Construction Research, Delft, Report BI-91-106.
123. Elliott, J. S. and Nethercot, D. (1991) Non-composite Flexural and Shear Tests on
CF70 Decking. Department of Civil Engineering, University of Nottingham, Report
SR 91033.
124. Johnson, R. P. (2004) The mk and partial-interaction models for shear resistance of
composite slabs, and the use of non-standard test data. In: Composite Construction in
Steel and Concrete V [Proceedings of a Conference, Kruger National Park, 2004].
American Society of Civil Engineers, New York (in preparation).
230
242
12 May 2004 11:53:29
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:35
Index
Note: references to beams and to columns are to composite members
action effect see actions, effects of
actions 6
combinations of 323, 128, 130
characteristic 33, 1245, 136
frequent 33, 121, 124, 174
quasi-permanent 33, 91, 129, 134, 139
concentrated 36, 209
effects of 6
independent 104, 117
primary 9, 92
secondary 9, 92
second-order 223, 34
horizontal 27
indirect 9
see also fatigue
adhesives 17
analysis, elastic, of cross-sections
see beams; columns; etc.
analysis, global 6, 2140, 121, 156
cracked 13, 36
elastic 2936, 901, 121, 128, 134, 1567
elasticplastic 105
first-order 223, 90
for composite slabs 21, 164, 169
for profiled sheeting 163
non-linear 334
of frames 234, 149
rigid-plastic 334, 368
second-order 224, 30, 324, 105, 108, 110
uncracked 36
see also cracking of concrete
analysis, types of 6, 278
anchorage 76, 162, 1679, 187
see also reinforcement
Annex, National see National Annex
application rules 5
axes 7, 106
beams 4167
bending resistance of 39, 4454
hogging 45, 504, 88
sagging 4450, 889, 967
Class 1 or 2 9, 36, 389, 43, 45, 49, 68, 88, 130,
134, 207
Class 3 9, 36, 39, 50, 56, 59, 67, 70
Class 4 9, 28, 34, 36, 54, 58, 130
concrete-encased 4, 32, 34, 43, 57, 1801
cross-sections of 412, 44
asymmetric 378
classification of 28, 3740, 523, 878
critical 423, 69
effective 43, 90
elastic analysis of 44, 50, 54, 8990
non-uniform 43
sudden change in 42
see also slabs, composite; slabs, concrete
curved in plan 44
design procedure for 478
effective width of flanges of 289, 87, 138, 145
flexural stiffness of 32
haunched 734, 79, 81
in frames 32
L-section 72
of non-uniform section 4, 31, 34
shear connection for see shear connection
shear resistance of 89
stresses in 1367
see also analysis; buckling; cantilevers;
cracking of concrete; deflections; fire,
resistance to; imperfections; interaction;
shear ; vibration; webs
bending moments
elastic critical 5865, 946
in columns 34, 105, 11011, 115
redistribution of 15, 347, 39, 66, 129, 1634
243
12 May 2004 11:53:29
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:35
bolts, fracture of 143
bolts, holes for 44
bolts, stiffness of 182
bond see shear connection
box girders 4, 80
bracing to bottom flanges 60, 63, 100
breadth of flange, effective see beams; slabs,
composite
bridges, beams in 68, 120
British Standards 5
BS 5400 80, 120, 1878
BS 5950 34, 367, 39, 71, 74, 80, 130, 1623,
165, 168, 1878, 193, 196
BS 8110 1415, 164
buckling 224
in columns 23, 1056, 108, 113
lateraltorsional 24, 346, 41, 45, 55, 5766,
90, 936, 20310
local 24, 28, 37, 3940, 67, 80, 130, 164
see also beams, cross-sections of
of reinforcement 44
of webs in shear 24, 34, 557, 89
see also loads, elastic critical
buildings, high-rise 104, 128
calibration 1415, 71, 132
cantilevers 58, 79, 93, 209
CEN (Comit Europen Normalisation), 1, 3,
10
Class of section see Beams, Class
Codes of Practice see British Standards
columns 103119, 128
bi-axial bending in 105, 108, 111, 117
concrete-encased 4, 1056, 112, 147, 2203
concrete-filled 14, 1057, 112, 218, 221
creep in 31, 114, 118
cross-sections of
interaction diagram for 104, 1067, 115,
21723
non-symmetrical 1035
section moduli for 21921
design method for 10511
eccentricity of loading for 107
effective length of 103
effective stiffness of 28, 107, 110
high-strength steel in 44
load introduction in 104, 11112
momentshear interaction in 107, 110, 115
second-order effects in 110, 11617
shear in 112, 219
slenderness of 1078, 115
squash load of 106
steel contribution ratio for 104, 107, 114
transverse loading on 24, 118
see also buckling; bending moments;
imperfections; loads, elastic critical;
shear connection; stresses, residual
compression members see columns
concrete
lightweight-aggregate 13, 1516, 85, 91, 104,
136, 191, 195
precast 80, 129, 205
properties of 1316, 119, 147
strength classes for 13
strength of 10
stress block for 1415
see also cracking of concrete; creep of
concrete; elasticity, modulus of;
shrinkage of concrete
connections see joints
connector modulus see shear connection,
stiffness of
construction 3, 78, 168
loads 30, 1623, 168, 172
methods of 9
propped 78, 136, 138, 193
unpropped 13, 367, 39, 50, 59, 84, 128, 135
Construction Products Directive 10
contraflexure, points of 42
cover 19, 73, 77, 86, 106, 113
crack inducers 193, 195
cracking of concrete 121, 1316, 1589, 168
and global analysis 312, 90, 12930
control of
load-induced 1345, 139
restraint-induced 1323, 138
early thermal 133
uncontrolled 1312
creep of concrete 15, 2931, 49, 89, 91, 121
in columns 108
see also modular ratio; elasticity, modulus of
critical length 42
cross-sections see beams; columns
curvature of beams 49
decking, metal see sheeting, profiled steel
definitions 6
deflections 12830, 1378
due to shrinkage 130, 137
due to slip 137
limits to 129, 1378
of beams 128, 1578
of composite slabs 16870, 174
of profiled sheeting 1634, 173
deformation, imposed 9, 131
design, basis of 911
design, methods of see beams; columns; slabs
Designers Guides 1, 10
ductility see reinforcement, fracture of
durability 19, 86
effective length see columns
effective width see beams; slabs, composite
effect of action see actions, effects of
eigenvalue see load, elastic critical
elasticity, modulus of
for concrete 14, 29, 108
for steels 16, 29
232
244
12 May 2004 11:53:29
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:35
EN 1090 80
EN 1990 3, 91, 1279, 131, 174, 190, 193
EN 1991 3, 32, 131, 172
EN 1992-1-1 5
and detailing 77, 132
and materials 1316, 49, 91
and resistances 28, 812, 122, 128, 168, 175
and serviceability 131, 174
EN 1993-1-1 21, 23, 39, 58, 63, 205
EN 1993-1-3 16, 1634, 1689
EN 1993-1-5 28, 55
EN 1993-1-8 16, 22, 1415, 1504, 17980
EN 1993-1-9 11920, 122
EN 1993-2 123
EN 1994-1-1, scope of 35, 15
EN 1994-1-2 57, 134
EN 1994-2 4, 72, 103, 119
EN 1998 3
EN ISO 14555 72
ENV 1994-1-1 15, 602, 68, 80, 105, 194, 198,
204, 209
equilibrium, static 11
Eurocodes 1
European Standard 5
see also EN
examples
bending and vertical shear 556
composite beam, continuous 84100, 1369
composite column 1139, 2223
composite joint 14759, 1815
composite slab 1707, 194201
effective width 29
elastic resistance to bending 1013
fatigue 1246
lateraltorsional buckling 66
reduction factor for strength of stud 76
resistance to hogging bending 504
shear connection 6970, 1001
transverse reinforcement 82
execution see construction
exposure classes 19, 86, 113, 131
factors, combination 32
factors, conversion 10
factors, damage equivalent 123
factors, partial 2, 10, 11920, 163, 190
factors, reduction 58, 726, 96, 106, 108, 181
fatigue 11, 34, 11926
finite-element methods 33
fire, resistance to 57
flanges, concrete see beams; slabs
flow charts 257, 35, 38, 645, 109
forces, internal 105
formwork, re-usable 129
see also sheeting, profiled steel
foundations 11
frame, inverted-U, 5961, 162, 2046
frames, composite 6
braced 32, 147
unbraced 4, 32, 34, 108
see also analysis, global; buckling;
imperfections
geometrical data 10
see also imperfections
haunches see beams, haunched
hole-in-web method 389, 514
imperfections 212, 247, 90, 110, 115, 117
interaction, full 44
interaction, partial 43, 459, 129, 1767
see also shear connection
ISO standards 56
joints 16, 212, 28
beam-to-column 108, 14159
bending resistance of 1504
classification of 142, 1489
contact-plate 142, 1445, 180
end-plate 141, 143, 14759, 181
full-strength 22, 143, 145
modelling of 1437
nominally pinned 22, 136, 143, 145, 147, 153
partial-strength 2, 34, 37, 143, 145, 149
rigid 136, 142
rotational stiffness of 142, 1445, 156, 17985
rotation capacity of 1445, 147, 156
semi-continuous 22
semi-rigid 22, 34, 136, 143, 149
simple see nominally pinned
L-beams 78, 188
length, effective 103
limit states
serviceability 28, 33, 12739
ultimate 41126
loads
dynamic 1612
elastic critical 23, 32, 103, 108, 11415
imposed, for fatigue 120
see also actions
mk method 1659, 175, 177, 1914, 21114
see also slabs, composite, tests on
materials, properties of 1317
see also concrete; steel
mesh, welded see reinforcement, welded mesh
modular ratio 2931, 8991, 121
modulus of elasticity see elasticity, modulus of
moments see bending moments; torsion
nationally determined parameter 1, 34
National Annexes 13, 5, 19, 22, 73, 128
and actions 32, 130, 162, 172
and beams 61, 96
and materials 1417
233
INDEX
245
12 May 2004 11:53:29
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:35
and partial factors 10, 58, 119
and serviceability 131, 138, 169
notation see symbols
notes, in Eurocodes 2, 10
partial factors 2, 9, 128
M
, for materials and resistances 7, 10, 123,
175
F
, for actions 123
plastic theory see analysis, global, rigid-plastic
plate girders 80, 141
see also beams
plates, composite 4
prestressing 4, 6, 13, 30, 33
principles 56
propping see construction, methods of
prying 150, 152
push tests see shear connectors, tests on
redistribution see bending moments; shear,
longitudinal
reference standards 5
reinforcement 131, 1823
anchorage of 42, 789, 155
at joints 1478
fracture of, in joints 39, 45, 132
in beams 52, 139
minimum area of 13, 36, 39, 1315, 138
spacing of 1323
transverse 68, 77, 813, 99100, 162, 206
in columns 106, 11213
in compression 44
in haunches see beams, haunched
strain hardening in 45
welded mesh (fabric), 1516, 39, 45, 130
see also cover; fatigue; slabs, composite
reinforcing steel 1516, 119
resistances 10, 27, 34
see also beams, bending resistance of; etc.
rotation capacity 345, 37, 44, 164
see also joints
safety factors see partial factors
section modulus 7
sections see beams; columns; etc.
separation 6, 44, 678, 77, 80, 112, 190
serviceability see limit states
shakedown 121
shear see columns; shear, longitudinal; shear,
vertical; etc.
shearbond test see mk test
shear connection 423, 6781, 122, 183
and execution 78
by bond or friction 17, 67, 104, 112, 162,
1667, 1934
degree of 46, 69, 97, 1013, 129, 1667, 176
design of 44
detailing of 7681
for composite slabs 1658
full 45, 48, 145
in columns 104, 1112, 118, 188
partial 43, 4550, 57, 6870, 119, 145
equilibrium method for 489
interpolation method for 47
see also anchorage; fatigue; reinforcement, in
beams, transverse; shear connectors;
slab, composite; slip, longitudinal
shear connectors 17, 57
bi-axial loading of 76
ductile 46, 50, 679, 79, 97, 103, 191
flexibility of see stiffness of
non-ductile 47, 4950, 1001, 103
spacing of 478, 50, 58, 69, 77, 80, 98, 101,
2045
stiffness of 44, 60, 181
tests on 181, 18791
types of 68
see also studs, welded
shear flow 423, 81, 122
shear heads 4
shear lag see slabs, composite, effective width of
shear, longitudinal 21, 701, 814
in columns 112
see also shear connection; slabs, composite
shear, punching see slabs, composite
shear, vertical 21, 54, 912, 144, 149, 1534
and bending moment 557
see also buckling; slabs, composite
sheeting, profiled steel 17, 726, 85, 1612,
1713, 212
and cracking 135
and shear connection 68, 801, 192
as transverse reinforcement 834, 99
bearing length for 162, 172
depth of 162
design of 1624, 1723
effective area of 45
embossments on (dimples in), 17, 162, 164,
1913
fixing of 83
in compression 45
loading on 1623
properties of 171, 193, 195
see also buckling; deflection; durability; slabs,
composite
shrinkage of concrete 15
and cracking 15, 1312, 134
effects of 9, 301, 50, 8993, 101, 112, 1368,
158
see also deflections
slabs, composite 16177
as diaphragms 162
bending resistance of 165, 167, 1734, 177
brittle failure of 162, 1656, 187, 191, 193
concentrated loads on 164
cracking in 168, 175
effective thickness of 43, 170, 175
234
246
12 May 2004 11:53:29
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:31:35
effective width of 43, 1645
flexural stiffness of 2034
longitudinal shear in 1658, 191, 21115
partial-connection design of 1668, 1757,
188, 1914, 198
partial shear connection in 39, 191201,
21112, 21415
punching shear in 168
reinforcement in 132, 164, 1678
serviceability of 16870
shear span for 166
tests on 1657, 169, 1878, 1918, 201
thickness of 132, 162, 176, 21115
vertical shear in 166, 168, 1723, 175
see also anchorage; deflections; global
analysis; mk method; sheeting, profiled
steel
slabs, concrete
reinforcement in 68, 79
splitting in 72, 789, 81, 189
see also concrete, precast; slabs, composite
slabs, form-reinforced see slabs, composite
slenderness ratios, limiting see beams,
cross-sections of
slenderness, relative 58, 96, 103, 1078, 2069
slip, longitudinal 6, 23, 36, 44, 46, 49, 106, 111,
128, 144, 166, 188
and deflections 129, 132
available 46
capacity 679, 1901
slip strain 46
software for EN 1994 23, 49, 105
squash load see columns, squash load of
standards see British Standards; EN
standards, harmonized 10
steel see reinforcing steel; structural steels;
yielding of steel
steel contribution ratio see columns
steelwork, protection of see durability
stiffness coefficient 179185
stiffness, flexural see beams; columns; etc.
strength 10
see also resistance
stresses
excessive 128
fatigue 1214
residual 22, 24
see also beams, stresses in
stress range, damage equivalent 119, 122
stress resultant see actions, effects of
structural steels 16, 34, 44, 111
stub girders 42
studs, lying 79
studs, welded 46
length after welding 71
resistance of
in composite slabs 726, 81, 978, 161
in solid slabs 702
tension in 72
weld collar of 17, 68, 72, 99
see also fatigue; shear connection, detailing of;
shear connectors
subscripts 7
supports, friction at see shear connection
sway frames see frames, unbraced
symbols 67, 121, 143
temperature, effects of 9, 16, 323, 130, 132
tension stiffening 45, 122, 125, 1345, 139
testing see shear connectors; slabs, composite
through-deck welding see welding, through-deck
torsion 44, 169
trusses 24
tubes, steel see columns, concrete-filled
U-frame see frame, inverted-U
uplift see separation
vibration 128, 1301
webs 667, 1804
encased 4, 3840, 601, 66, 134, 141, 20410
holes in 4, 44
see also hole-in-web method; shear, vertical
web stiffeners 60
welding, through-deck 72, 74, 83, 168
width, effective see beams; slabs, composite
worked examples see examples
yielding of steel and deflections 130, 1367
yield line theory 1512
235
INDEX
247
12 May 2004 11:53:30
DESIGNERS GUIDE TO EN 1994-2
EUROCODE 4: DESIGN OF STEEL AND
COMPOSITE STRUCTURES
PART 2: GENERAL RULES AND RULES FOR BRIDGES
Eurocode Designers Guide Series
Designers Guide to EN 1990. Eurocode: Basis of Structural Design. H. Gulvanessian, J.-A. Calgaro and
M. Holicky. 0 7277 3011 8. Published 2002.
Designers Guide to EN 1994-1-1. Eurocode 4: Design of Composite Steel and Concrete Structures. Part 1.1:
General Rules and Rules for Buildings. R. P. Johnson and D. Anderson. 0 7277 3151 3. Published 2004.
Designers Guide to EN 1997-1. Eurocode 7: Geotechnical Design General Rules. R. Frank, C. Bauduin,
R. Driscoll, M. Kavvadas, N. Krebs Ovesen, T. Orr and B. Schuppener. 0 7277 3154 8. Published 2004.
Designers Guide to EN 1993-1-1. Eurocode 3: Design of Steel Structures. General Rules and Rules for Buildings.
L. Gardner and D. Nethercot. 0 7277 3163 7. Published 2004.
Designers Guide to EN 1992-1-1 and EN 1992-1-2. Eurocode 2: Design of Concrete Structures. General Rules
and Rules for Buildings and Structural Fire Design. A.W. Beeby and R. S. Narayanan. 0 7277 3105 X. Published
2005.
Designers Guide to EN 1998-1 and EN 1998-5. Eurocode 8: Design of Structures for Earthquake Resistance.
General Rules, Seismic Actions, Design Rules for Buildings, Foundations and Retaining Structures. M. Fardis,
E. Carvalho, A. Elnashai, E. Faccioli, P. Pinto and A. Plumier. 0 7277 3348 6. Published 2005.
Designers Guide to EN 1995-1-1. Eurocode 5: Design of Timber Structures. Common Rules and for Rules and
Buildings. C. Mettem. 0 7277 3162 9. Forthcoming: 2007 (provisional).
Designers Guide to EN 1991-4. Eurocode 1: Actions on Structures. Wind Actions. N. Cook. 0 7277 3152 1.
Forthcoming: 2007 (provisional).
Designers Guide to EN 1996. Eurocode 6: Part 1.1: Design of Masonry Structures. J. Morton. 0 7277 3155 6.
Forthcoming: 2007 (provisional).
Designers Guide to EN 1991-1-2, 1992-1-2, 1993-1-2 and EN 1994-1-2. Eurocode 1: Actions on Structures.
Eurocode 3: Design of Steel Structures. Eurocode 4: Design of Composite Steel and Concrete Structures. Fire
Engineering (Actions on Steel and Composite Structures). Y. Wang, C. Bailey, T. Lennon and D. Moore.
0 7277 3157 2. Forthcoming: 2007 (provisional).
Designers Guide to EN 1992-2. Eurocode 2: Design of Concrete Structures. Bridges. D. Smith and C. Hendy.
Designers Guide to EN 1993-2. Eurocode 3: Design of Steel Structures. Bridges. C. Murphy and C. Hendy.
Designers Guide to EN 1991-2, 1991-1-1, 1991-1-3 and 1991-1-5 to 1-7. Eurocode 1: Actions on Structures.
Trac Loads and Other Actions on Bridges. J.-A. Calgaro, M. Tschumi, H. Gulvanessian and N. Shetty.
Designers Guide to EN 1991-1-1, EN 1991-1-3 and 1991-1-5 to 1-7. Eurocode 1: Actions on Structures. General
Rules and Actions on Buildings (not Wind). H. Gulvanessian, J.-A. Calgaro, P. Formichi and G. Harding.
www.eurocodes.co.uk
EUROCODE 4: DESIGN OF STEEL AND
COMPOSITE STRUCTURES
PART 2: GENERAL RULES AND RULES
FOR BRIDGES
C. R. HENDY and R. P. JOHNSON
Published by Thomas Telford Publishing, Thomas Telford Ltd, 1 Heron Quay, London E14 4JD
URL: www.thomastelford.com
Distributors for Thomas Telford books are
USA: ASCE Press, 1801 Alexander Bell Drive, Reston, VA 20191-4400
Japan: Maruzen Co. Ltd, Book Department, 310 Nihonbashi 2-chome, Chuo-ku, Tokyo 103
Australia: DA Books and Journals, 648 Whitehorse Road, Mitcham 3132, Victoria
First published 2006
Eurocodes Expert
Structural Eurocodes oer the opportunity of harmonized design standards for the European
construction market and the rest of the world. To achieve this, the construction industry needs to
become acquainted with the Eurocodes so that the maximum advantage can be taken of these
opportunities
Eurocodes Expert is a new ICE and Thomas Telford initiative set up to assist in creating a greater
awareness of the impact and implementation of the Eurocodes within the UK construction industry
Eurocodes Expert provides a range of products and services to aid and support the transition to
Eurocodes. For comprehensive and useful information on the adoption of the Eurocodes and their
implementation process please visit our website or email eurocodes@thomastelford.com
A catalogue record for this book is available from the British Library
ISBN: 0 7277 3161 0
# The authors and Thomas Telford Limited 2006
All rights, including translation, reserved. Except as permitted by the Copyright, Designs and Patents
Act 1988, no part of this publication may be reproduced, stored in a retrieval system or transmitted in
any form or by any means, electronic, mechanical, photocopying or otherwise, without the prior
written permission of the Publishing Director, Thomas Telford Publishing, Thomas Telford Ltd,
1 Heron Quay, London E14 4JD.
This book is published on the understanding that the authors are solely responsible for the statements
made and opinions expressed in it and that its publication does not necessarily imply that such
statements and/or opinions are or reect the views or opinions of the publishers. While every eort
has been made to ensure that the statements made and the opinions expressed in this publication
provide a safe and accurate guide, no liability or responsibility can be accepted in this respect by the
authors or publishers.
Typeset by Academic Technical, Bristol
Printed and bound in Great Britain by MPG Books, Bodmin
Preface
EN 1994, also known as Eurocode 4 or EC4, is one standard of the Eurocode suite and
describes the principles and requirements for safety, serviceability and durability of compo-
site steel and concrete structures. It is subdivided into three parts:
.
Part 1.1: General Rules and Rules for Buildings
.
Part 1.2: Structural Fire Design
.
Part 2: General Rules and Rules for Bridges.
It is used in conjunction with EN 1990, Basis of Structural Design; EN 1991, Actions on
Structures; and the other design Eurocodes.
Aims and objectives of this guide
The principal aim of this book is to provide the user with guidance on the interpretation and
use of EN 1994-2 and to present worked examples. It covers topics that will be encountered
in typical steel and concrete composite bridge designs, and explains the relationship between
EN 1994-1-1, EN 1994-2 and the other Eurocodes. It refers extensively to EN 1992 (Design of
Concrete Structures) and EN 1993 (Design of Steel Structures), and includes the application
of their provisions in composite structures. Further guidance on these and other Eurocodes
will be found in other Guides in this series.
17
This book also provides background
information and references to enable users of Eurocode 4 to understand the origin and
objectives of its provisions.
The need to use many Eurocode parts can initially make it a daunting task to locate
information in the sequence required for a real design. To assist with this process, ow
charts are provided for selected topics. They are not intended to give detailed procedural
information for a specic design.
Layout of this guide
EN 1994-2 has a foreword, nine sections, and an annex. This guide has an introduction which
corresponds to the foreword of EN 1994-2, Chapters 1 to 9 which correspond to Sections 1 to
9 of the Eurocode, and Chapter 10 which refers to Annexes A and B of EN 1994-1-1 and
covers Annex C of EN 1994-2. Commentary on Annexes A and B is given in the Guide by
Johnson and Anderson.
5
The numbering and titles of the sections and second-level clauses in this guide also corre-
spond to those of the clauses of EN 1994-2. Some third-level clauses are also numbered (for
example, 1.1.2). This implies correspondence with the sub-clause in EN 1994-2 of the same
number. Their titles also correspond. There are extensive references to lower-level clause and
paragraph numbers. The rst signicant reference is in bold italic type (e.g. clause 1.1.1(2)).
These are in strict numerical sequence throughout the book, to help readers nd comments
on particular provisions of the code. Some comments on clauses are necessarily out of
sequence, but use of the index should enable these to be found.
All cross-references in this guide to sections, clauses, sub-clauses, paragraphs, annexes,
gures, tables and expressions of EN 1994-2 are in italic type, and do not include
EN 1994-2. Italic is also used where text from a clause in EN 1994-2 has been directly
reproduced.
Cross-references to, and quotations and expressions from, other Eurocodes are in roman
type. Clause references include the EN number; for example, clause 3.1.4 of EN 1992-1-1 (a
reference in clause 5.4.2.2(2)). All other quotations are in roman type. Expressions repeated
from EN 1994-2 retain their number. The authors expressions have numbers prexed by D
(for Designers Guide); for example, equation (D6.1) in Chapter 6.
Abbreviated terms are sometimes used for parts of Eurocodes (e.g. EC4-1-1 for EN 1994-
1-1
8
) and for limit states (e.g. ULS for ultimate limit state).
Acknowledgements
The rst author would like to thank his wife, Wendy, and two boys, Peter Edwin Hendy and
Matthew Philip Hendy, for their patience and tolerance of his pleas to nish just one more
paragraph. He thanks his employer, Atkins, for providing both facilities and time for the
production of this guide, and the members of BSI B525/10 Working Group 2 who provided
comment on many of the Eurocode clauses.
The second author is deeply indebted to the other members of the project and editorial
teams for Eurocode 4 on which he has worked: David Anderson, Gerhard Hanswille,
Bernt Johansson, Basil Kolias, Jean-Paul Lebet, Henri Mathieu, Michel Mele, Joel Raoul,
Karl-Heinz Roik and Jan Stark; and also to the Liaison Engineers, National Technical
Contacts, and others who prepared national comments. He thanks the University of
Warwick for facilities provided for Eurocode work, and, especially, his wife Diana for her
unfailing support.
Chris Hendy
Roger Johnson
vi
Contents
Preface v
Aims and objectives of this guide v
Layout of this guide v
Acknowledgements vi
Introduction 1
Additional information specic to EN 1994-2 2
Chapter 1. General 3
1.1. Scope 3
1.1.1. Scope of Eurocode 4 3
1.1.2. Scope of Part 1.1 of Eurocode 4 3
1.1.3. Scope of Part 2 of Eurocode 4 4
1.2. Normative references 5
1.3. Assumptions 7
1.4. Distinction between principles and application rules 7
1.5. Denitions 8
1.5.1. General 8
1.5.2. Additional terms and denitions 8
1.6. Symbols 8
Chapter 2. Basis of design 11
2.1. Requirements 11
2.2. Principles of limit states design 12
2.3. Basic variables 12
2.4. Verication by the partial factor method 12
2.4.1. Design values 12
2.4.2. Combination of actions 15
2.4.3. Verication of static equilibrium (EQU) 15
Chapter 3. Materials 17
3.1. Concrete 17
3.2. Reinforcing steel for bridges 19
3.3. Structural steel for bridges 21
3.4. Connecting devices 22
3.4.1. General 22
3.4.2. Headed stud shear connectors 22
3.5. Prestressing steel and devices 23
3.6. Tension components in steel 23
Chapter 4. Durability 25
4.1. General 25
4.2. Corrosion protection at the steelconcrete interface in bridges 27
Chapter 5. Structural analysis 29
5.1. Structural modelling for analysis 29
5.1.1. Structural modelling and basic assumptions 29
5.1.2. Joint modelling 30
5.1.3. Groundstructure interaction 30
5.2. Structural stability 30
5.2.1. Eects of deformed geometry of the structure 31
5.2.2. Methods of analysis for bridges 33
5.3. Imperfections 34
5.3.1. Basis 34
5.3.2. Imperfections for bridges 35
5.4. Calculation of action eects 36
5.4.1. Methods of global analysis 36
Example 5.1: eective widths of concrete ange for shear lag 41
5.4.2. Linear elastic analysis 42
Example 5.2: modular ratios for long-term loading and for shrinkage 53
Example 5.3: primary eects of shrinkage 54
5.4.3. Non-linear global analysis for bridges 56
5.4.4. Combination of global and local action eects 56
5.5. Classication of cross-sections 57
Example 5.4: classication of composite beam section in hogging bending 60
Flow charts for global analysis 62
Chapter 6. Ultimate limit states 67
6.1. Beams 67
6.1.1. Beams in bridges general 67
6.1.2. Eective width for verication of cross-sections 68
6.2. Resistances of cross-sections of beams 68
6.2.1. Bending resistance 69
Example 6.1: plastic resistance moment in sagging bending 72
Example 6.2: resistance to hogging bending at an internal support 73
Example 6.3: elastic bending resistance of a Class 4 cross-section 77
6.2.2. Resistance to vertical shear 79
Example 6.4: resistance of a Class 4 section to hogging bending and
vertical shear 85
Example 6.5: addition of axial compression to a Class 4 cross-section 86
6.3. Filler beam decks 89
6.3.1. Scope 89
6.3.2. General 90
6.3.3. Bending moments 90
6.3.4. Vertical shear 91
6.3.5. Resistance and stability of steel beams during execution 91
6.4. Lateraltorsional buckling of composite beams 91
6.4.1. General 91
6.4.2. Beams in bridges with uniform cross-sections in Class 1, 2
and 3 92
viii
6.4.3. General methods for buckling of members and frames 93
Example 6.6: bending and shear in a continuous composite beam 104
Example 6.7: stiness and required resistance of cross-bracing 111
6.5. Transverse forces on webs 113
6.6. Shear connection 114
6.6.1. General 114
Example 6.8: shear resistance of a block connector with a hoop 116
6.6.2. Longitudinal shear force in beams for bridges 118
6.6.3. Headed stud connectors in solid slabs and concrete
encasement 121
6.6.4. Headed studs that cause splitting in the direction of the
slab thickness 123
6.6.5. Detailing of the shear connection and inuence of
execution 124
6.6.6. Longitudinal shear in concrete slabs 127
Example 6.9: transverse reinforcement for longitudinal shear 130
Example 6.10: longitudinal shear checks 131
Example 6.11: inuence of in-plane shear in a compressed ange on
bending resistances of a beam 134
6.7. Composite columns and composite compression members 136
6.7.1. General 136
6.7.2. General method of design 137
6.7.3. Simplied method of design 138
6.7.4. Shear connection and load introduction 144
6.7.5. Detailing provisions 145
Example 6.12: concrete-lled tube of circular cross-section 145
6.8. Fatigue 150
6.8.1. General 150
6.8.2. Partial factors for fatigue assessment of bridges 151
6.8.3. Fatigue strength 152
6.8.4. Internal forces and fatigue loadings 152
6.8.5. Stresses 153
6.8.6. Stress ranges 155
6.8.7. Fatigue assessment based on nominal stress ranges 156
Example 6.13: fatigue verication of studs and reinforcement 157
6.9. Tension members in composite bridges 161
Chapter 7. Serviceability limit states 163
7.1. General 163
7.2. Stresses 164
7.3. Deformations in bridges 166
7.3.1. Deections 166
7.3.2. Vibrations 166
7.4. Cracking of concrete 167
7.4.1. General 167
7.4.2. Minimum reinforcement 168
7.4.3. Control of cracking due to direct loading 169
7.5. Filler beam decks 173
Example 7.1: checks on serviceability stresses, and control of
cracking 173
Chapter 8. Precast concrete slabs in composite bridges 179
8.1. General 179
8.2. Actions 180
ix
CONTENTS
8.3. Design, analysis and detailing of the bridge slab 180
8.4. Interface between steel beam and concrete slab 181
Chapter 9. Composite plates in bridges 183
9.1. General 183
9.2. Design for local eects 183
9.3. Design for global eects 184
9.4. Design of shear connectors 185
Example 9.1: design of shear connection for global eects at the
serviceability limit state 187
Chapter 10. Annex C (informative). Headed studs that cause splitting forces in
the direction of the slab thickness 189
C.1. Design resistance and detailing 190
C.2. Fatigue strength 191
Applicability of Annex C 191
Example 10.1: design of lying studs 192
References 195
Index 201
x
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:27:10
Introduction
The provisions of EN 1994-2
9
are preceded by a foreword, most of which is common to all
Eurocodes. This Foreword contains clauses on:
.
the background to the Eurocode programme
.
the status and eld of application of the Eurocodes
.
national standards implementing Eurocodes
.
links between Eurocodes and harmonized technical specications for products
.
additional information specic to EN 1994-2
.
National Annex for EN 1994-2.
Guidance on the common text is provided in the introduction to the Designers Guide to
EN 1990. Eurocode: Basis of Structural Design,
1
and only background information relevant
to users of EN 1994-2 is given here.
It is the responsibility of each national standards body to implement each Eurocode part
as a national standard. This will comprise, without any alterations, the full text of the Euro-
code and its annexes as published by the European Committee for Standardisation,
CEN (from its title in French). This will usually be preceded by a National Title Page and
a National Foreword, and may be followed by a National Annex.
Each Eurocode recognizes the right of national regulatory authorities to determine values
related to safety matters. Values, classes or methods to be chosen or determined at national
level are referred to as Nationally Determined Parameters (NDPs). Clauses in which these
occur are listed in the Foreword.
NDPs are also indicated by notes immediately after relevant clauses. These Notes give
recommended values. Many of the values in EN 1994-2 have been in the draft code for
over a decade. It is expected that most of the 28 Member States of CEN (listed in the Fore-
word) will specify the recommended values, as their use was assumed in the many calibration
studies done during drafting. They are used in this guide, as the National Annex for the UK
was not available at the time of writing.
Each National Annex will give or cross-refer to the NDPs to be used in the relevant
country. Otherwise the National Annex may contain only the following:
10
.
decisions on the use of informative annexes, and
.
references to non-contradictory complementary information to assist the user to apply
the Eurocode.
Each national standards body that is a member of CEN is required, as a condition of mem-
bership, towithdrawall conicting national standards by a givendate, that is at present March
2010. The Eurocodes will supersede the British bridge code, BS 5400,
11
which should therefore
be withdrawn. This will lead to extensive revision of many sets of supplementary design rules,
such as those published by the Highways Agency in the UK. Some countries have already
adopted Eurocode methods for bridge design; for example, Germany in 2003.
12
IP: 95.42.11.157
On: Sun, 11 Jul 2010 21:27:10
Additional information specic to EN 1994-2
The information specic to EN 1994-2 emphasises that this standard is to be used with other
Eurocodes. The standard includes many cross-references to particular clauses in EN 1990,
13
EN 1991,
14
EN 1992
15
and EN 1993.
16
Similarly, this guide is one of a series on Eurocodes,
and is for use with other guides, particularly those for EN 1991,
2
EN 1992-1-1,
6
EN 1993-1-
1,
7
EN 1992-2
3
and EN 1993-2.
4
The Foreword refers to a dierence between EN 1994-2 and the bridge parts of the other
Eurocodes. In Eurocode 4, the general provisions of Part 1-1 are repeated word for word in
Part 2, with identical numbering of clauses, paragraphs, equations, etc. Such repetition
breaks a rule of CEN, and was permitted, for this code only, to shorten chains of cross-
references, mainly to Eurocodes 2 and 3. This determined the numbering and location of
the provisions for bridges, and led to a few gaps in the sequences of numbers.
The same policy has been followed in the guides on Eurocode 4. Where material in the
5
is as relevant to bridges as to buildings, it is repeated
here, so this guide is self-contained, in respect of composite bridges, as is EN 1994-2.
A very few General clauses in EN 1994-1-1 are not applicable to bridges. They have
been replaced in EN 1994-2 by clearly labelled bridge clauses; for example, clause 3.2,
Reinforcing steel for bridges.
The Foreword lists the 15 clauses of EN 1994-2 in which national choice is permitted. Five
of these relate to values for partial factors, three to shear connection, and seven to provision
of further guidance. Elsewhere, there are cross-references to clauses with NDPs in other
codes; for example, partial factors for steel and concrete, and values that may depend on
climate, such as the free shrinkage of concrete.
Otherwise, the Normative rules in the code must be followed, if the design is to be in
accordance with the Eurocodes.
In EN 1994-2, Sections 1 to 9 are Normative. Only its Annex C is Informative, because it
is based on quite recent research. A National Annex may make it normative in the country
concerned, and is itself normative in that country, but not elsewhere. The non-contradictory
complementary information referred to above could include, for example, reference to a
document based on provisions of BS 5400 on matters not treated in the Eurocodes. Each
country can do this, so some aspects of the design of a bridge will continue to depend on
where it is to be built.
2
CHAPTER 1
General
This chapter is concerned with the general aspects of EN 1994-2, Eurocode 4: Design of
Composite Steel and Concrete Structures, Part 2: General Rules and Rules for Bridges. The
material described in this chapter is covered in Section 1, in the following clauses:
.
Scope Clause 1.1
.
Normative references Clause 1.2
.
Assumptions Clause 1.3
.
Distinction between principles and application rules Clause 1.4
.
Denitions Clause 1.5
.
Symbols Clause 1.6
1.1. Scope
1.1.1. Scope of Eurocode 4
The scope of EN1994 (all three Parts) is outlined in clause 1.1.1. It is to be used with EN1990,
Eurocode: Basis of Structural Design, which is the head document of the Eurocode suite, and
has an Annex A2, Application for bridges. Clause 1.1.1(2) emphasizes that the Eurocodes
are concerned with structural behaviour and that other requirements, e.g. thermal and
acoustic insulation, are not considered.
The basis for verication of safety and serviceability is the partial factor method. EN 1990
recommends values for load factors and gives various possibilities for combinations of
actions. The values and choice of combinations are set by the National Annex for the
country in which the structure is to be constructed.
Eurocode 4 is also to be used in conjunction with EN 1991, Eurocode 1: Actions on
Structures
14
and its National Annex, to determine characteristic or nominal loads. When
a composite structure is to be built in a seismic region, account needs to be taken of
EN 1998, Eurocode 8: Design of Structures for Earthquake Resistance.
17
Clause 1.1.1(3), as a statement of intention, gives undated references. It supplements the
normative rules on dated reference standards, given in clause 1.2, where the distinction
between dated and undated standards is explained.
The Eurocodes are concerned with design and not execution, but minimum standards of
workmanship are required to ensure that the design assumptions are valid. For this reason,
clause 1.1.1(3) lists the European standards for the execution of steel structures and the
execution of concrete structures. The standard for steel structures includes some requirements
for composite construction for example, for the testing of welded stud shear connectors.
1.1.2. Scope of Part 1.1 of Eurocode 4
The general rules referred to in clause 1.1.2(1) appear also in EN 1994-2, so there is (in
general) no need for it to cross-refer to Part 1-1, though it does refer (in clause 6.6.3.1(4))
Clause 1.1.1
Clause 1.1.1(2)
Clause 1.1.1(3)
Clause 1.1.2(1)
to Annex B of Part 1-1. The list of the titles of sections in clause 1.1.2(2) is identical to that
in clause 1.1.3, except for those of Sections 8 and 9. In Sections 17 of EN 1994-2, all for
buildings clauses of EN 1994-1-1 are omitted, and for bridges clauses are added.
1.1.3. Scope of Part 2 of Eurocode 4
Clause 1.1.3(1) refers to the partial coverage of design of cable-stayed bridges. This is the
only reference to them in EN 1994-2. It was considered here, and in EC2 and EC3, that
for this rapidly evolving type of bridge, it was premature to codify much more than the
design of their components (e.g. cables, in EN 1993-1-11), although EN 1993-1-11 does
contain some requirements for global analysis. Composite construction is attractive for
cable-stayed bridges, because the concrete deck is well able to resist longitudinal com-
pression. There is an elegant example in central Johannesburg.
18
Clause 1.1.3(2) lists the titles of the sections of Part 2. Those for Sections 17 are the same
as in all the other material-dependent Eurocodes. The contents of Sections 1 and 2 similarly
follow an agreed model.
The provisions of Part 2 cover the design of the following:
.
beams in which a steel section acts compositely with concrete
.
concrete-encased or concrete-lled composite columns
.
composite plates (where the steel member is a at steel plate, not a proled section)
.
composite box girders
.
tapered or non-uniform composite members
.
structures that are prestressed by imposed deformations or by tendons.
Joints in composite beams and between beams and steel or composite columns appear in
clause 5.1.2, Joint modelling, which refers to EN 1993-1-8.
19
There is little detailed coverage,
because the main clauses on joints in Part 1-1 are for buildings.
Section 5, Structural analysis concerns connected members and frames, both unbraced and
braced. The provisions dene their imperfections and include the use of second-order global
analysis and prestress by imposed deformations.
The scope of Part 2 includes double composite action, and also steel sections that are par-
tially encased. The web of the steel section is encased by reinforced concrete, and shear con-
nection is provided between the concrete and the steel. This is a well-established form of
construction in buildings. The primary reason for its choice is improved resistance in re.
Fully-encased composite beams are not included because:
.
no satisfactory model has been found for the ultimate strength in longitudinal shear of a
beam without shear connectors
.
it is not known to what extent some design rules (e.g. for momentshear interaction and
redistribution of moments) are applicable.
A fully-encased beam with shear connectors can usually be designed as if partly encased
or uncased, provided that care is taken to prevent premature spalling of encasement in
compression.
Prestressing of composite members by tendons is rarely used, and is not treated in detail.
Transverse prestress of a deck slab is covered in EN 1992-2.
3
The omission of application rules for a type of member or structure should not prevent its
use, where appropriate. Some omissions are deliberate, to encourage the use of innovative
design, based on specialised literature, the properties of materials, and the fundamentals
of equilibrium and compatibility. However, the principles given in the relevant Eurocodes
must still be followed. This applies, for example, to:
.
members of non-uniform section, or curved in plan
.
types of shear connector other than welded headed studs.
EN 1994-2 has a single Informative annex, considered in Chapter 10 of this book.
The three annexes in EN 1994-1-1 were not copied into EN 1994-2 because they are
Informative and, except for tests on shear connectors, are for buildings. They are:
Clause 1.1.2(2)
Clause 1.1.3(1)
Clause 1.1.3(2)
4
.
Annex A, Stiness of joint components in buildings
.
Annex B, Standard tests (for shear connectors and for composite slabs)
.
Annex C, Shrinkage of concrete for composite structures for buildings.
In ENV 1994-1-1,
20
design rules for many types of shear connector were given. All except
those for welded headed studs were omitted, clause 1.1.3(3), mainly in response to requests
for a shorter code. The Note to this clause enables national annexes to refer to rules for any
type of shear connector. In the UK, this is being done for block connectors with hoops and
for channels, and in France for angle connectors, based on the rules in ENV 1994-1-1.
Research on older types of connector and the development of new connectors continues.
2125
1.2. Normative references
References are given only to other European standards, all of which are intended to be used
as a package. Formally, the Standards of the International Organization for Standardization
(ISO) apply only if given an EN ISO designation. National standards for design and for pro-
ducts do not apply if they conict with a relevant EN standard.
As Eurocodes may not cross-refer to national standards, replacement of national stan-
dards for products by EN or ISO standards is in progress, with a timescale similar to that
for the Eurocodes.
During the period of changeover to Eurocodes and EN standards it is possible that an
EN referred to, or its national annex, may not be complete. Designers who then seek
guidance from national standards should take account of dierences between the design phi-
losophies and safety factors in the two sets of documents.
The lists in clause 1.2 are limited to standards referred to in the text of EN 1994-1-1 or
1994-2. The distinction between dated and undated references should be noted. Any relevant
provision of the general reference standards, clause 1.2.1, should be assumed to apply.
EN 1994-2 is based on the concept of the initial erection of structural steel members, which
may include prefabricated concrete-encased members. The placing of formwork (which may
or may not become part of the nished structure) follows. The addition of reinforcement and
in situ concrete completes the composite structure. The presentation and content of EN 1994-
2 therefore relate more closely to EN 1993 than to EN 1992. This may explain why this list
includes execution of steel structures, but not EN 13670, on execution of concrete structures,
which is listed in clause 1.1.1.
Clause 1.1.3(3)
Clause 1.2
Clause 1.2.1
Table 1.1. References to EN 1992, Eurocode 2: Design of Concrete Structures
Title of Part Subjects referred to from EN 1994-2
EN 1992-1-1,
General Rules and Rules for Buildings
Properties of concrete, reinforcement, and tendons
General design of reinforced and prestressed concrete
Partial factors
M
, including values for fatigue
Resistance of reinforced concrete cross-sections to bending and shear
Bond, anchorage, cover, and detailing of reinforcement
Minimum areas of reinforcement; crack widths in concrete
Limiting stresses in concrete, reinforcement and tendons
Combination of actions for global analysis for fatigue
Fatigue strengths of concrete, reinforcement and tendons
Reinforced concrete and composite tension members
Transverse reinforcement in composite columns
Vertical shear and second-order eects in composite plates
Eective areas for load introduction into concrete
EN 1992-2,
Rules for Bridges
Many subjects with references also to EN 1992-1-1 (above)
Environmental classes; exposure classes
Limitation of crack widths
Vertical shear in a concrete ange
Exemptions from fatigue assessment for reinforcement and concrete
Verication for fatigue; damage equivalent factors
5
CHAPTER 1. GENERAL
The other reference standards in clause 1.2.2 receive both general references, as in clause
2.3.2(1) (to EN 1992-1-1
15
), and specic references to clauses, as in clause 3.1(1), which
refers to EN 1992-1-1, 3.1. For composite bridges, further standards, of either type, are
listed in clause 1.2.3.
For actions, the main reference is in clause 2.3.1(1), to the relevant parts of EN 1991,
which include those for unit weights of materials, wind loads, snow loads, thermal
actions, and actions during execution. The only references in clause 1.2 are to EN 1991-2,
Trac loads on bridges,
26
and to Annex A2 of EN 1990, which gives combination rules
and recommended values for partial factors and combination factors for actions for
bridges. EN 1990 is also referred to for modelling of structures for analysis, and general
provisions on serviceability limit states and their verication.
Cross-references from EN 1994-2 to EN 1992 and EN 1993
The parts of EN 1992 and EN 1993 most likely to be referred to in the design of a steel and
concrete composite bridge are listed in Tables 1.1 and 1.2, with the relevant aspects of design.
Clause 1.2.2
Clause 1.2.3
Table 1.2. References to EN 1993, Eurocode 3: Design of Steel Structures
Title of Part Subjects referred to from EN 1994-2
EN 1993-1-1,
General Rules and Rules for Buildings
Stressstrain properties of steel;
M
for steel
General design of unstiened steelwork
Classication of cross-sections
Resistance of composite sections to vertical shear
Buckling of members and frames; column buckling curves
EN 1993-1-5,
Plated Structural Elements
Design of cross-sections in slenderness Class 3 or 4
Eects of shear lag in steel plate elements
Design of beams before a concrete ange hardens
Design where transverse, longitudinal, or bearing stieners are present
Transverse distribution of stresses in a wide ange
Shear buckling; ange-induced web buckling
In-plane transverse forces on webs
EN 1993-1-8,
Design of Joints
Modelling of exible joints in analysis
Design of joints and splices in steel and composite members
Design using structural hollow sections
Fasteners and welding consumables
EN 1993-1-9,
Fatigue Strength of Steel Structures
Fatigue loading
Classication of details into fatigue categories
Limiting stress ranges for damage-equivalent stress verication
Fatigue verication in welds and connectors
EN 1993-1-10,
Material Toughness and
Through-thickness Properties
For selection of steel grade (Charpy test, and Z quality)
EN 1993-1-11,
Design of Structures with Tension
Components
Design of bridges with external prestressing or cable support, such as
cable-stayed bridges
EN 1993-2,
Rules for Bridges
Global analysis; imperfections
Buckling of members and frames
Design of beams before a concrete ange hardens
Limiting slenderness of web plates
Distortion in box girders
M
for fatigue strength;
F
for fatigue loading
Damage equivalent factors
Limiting stresses in steel; fatigue in structural steel
Limits to deformations
Vibration
6
Many references to EN 1992-2
27
and EN 1993-2
28
lead to references from them to
EN 1992-1-1 and EN 1993-1-1, respectively. Unfortunately, the method of drafting of
these two bridge parts was not harmonised. For many subjects, some of the clauses
needed are general and so are located in Part 1-1, and others are for bridges and will be
found in Part 2. There are examples in clauses 3.2(1), 7.2.2(2) and 7.4.1(1).
Other Eurocode parts that may be applicable are:
EN 1993-1-7 Strength and Stability of Planar Plated Structures Transversely Loaded
EN 1993-1-12 Supplementary Rules for High Strength Steel
EN 1997 Geotechnical Design, Parts 1 and 2
EN 1998 Design of Structures for Earthquake Resistance
EN 1999 Design of Aluminium Structures.
1.3. Assumptions
It is assumed in EN 1994-2 that the general assumptions of ENs 1990, 1992, and 1993 will be
followed. Commentary on them will be found in the relevant Guides of this series.
Various clauses in EN 1994-2 assume that EN 1090 will be followed in the fabrication and
erection of the steelwork. This is important for the design of slender elements, where the
methods of analysis and buckling resistance formulae rely on imperfections from fabrication
and erection being limited to the levels in EN 1090. EN 1994-2 should therefore not be used
for design of bridges that will be fabricated or erected to specications other than EN 1090,
without careful comparison of the respective requirements for tolerances and workmanship.
Similarly, the requirements of EN 13670 for execution of concrete structures should be com-
plied with in the construction of reinforced or prestressed concrete elements.
1.4. Distinction between principles and application rules
Clauses in the Eurocodes are set out as either Principles or Application Rules. As dened by
EN 1990:
.
Principles comprise general statements for which there is no alternative and require-
ments and analytical models for which no alternative is permitted unless specically
stated.
.
Principles are distinguished by the letter P following the paragraph number.
.
Application Rules are generally recognised rules which comply with the principles and
satisfy their requirements.
There may be other ways to comply with the Principles, that are at least equivalent to the
Application Rules in respect of safety, serviceability, and durability. However, if these are
substituted, the design cannot be deemed to be fully in accordance with the Eurocodes.
Eurocodes 2, 3 and 4 are consistent in using the verbal form shall only for a Principle.
Application rules generally use should or may, but this is not fully consistent.
There are relatively few Principles in Parts 1.1 and 2 of ENs 1992 and 1994. Almost all of
those in EN 1993-1-1 and EN 1993-2 were replaced by Application Rules at a late stage of
drafting.
It has been recognized that a requirement or analytical model for which no alternative is
permitted unless specically stated can rarely include a numerical value, because most values
are inuenced by research and/or experience, and may change over the years. (Even the
specied elastic modulus for structural steel is an approximate value.) Furthermore, a
clause cannot be a Principle if it requires the use of another clause that is an Application
Rule; eectively that clause also would become a Principle.
It follows that, ideally, the Principles in all the codes should form a consistent set, referring
only to each other, and intelligible if all the Application Rules were deleted. This overriding
principle strongly inuenced the drafting of EN 1994.
7
CHAPTER 1. GENERAL
1.5. Denitions
1.5.1. General
In accordance with the model specied for Section 1, reference is made to the denitions
given in clauses 1.5 of EN 1990, EN 1992-1-1, and EN 1993-1-1. Many types of analysis
are dened in clause 1.5.6 of EN 1990. It should be noted that an analysis based on the
deformed geometry of a structure or element under load is termed second-order, rather
than non-linear. The latter term refers to the treatment of material properties in structural
analysis. Thus, according to EN 1990, non-linear analysis includes rigid-plastic. This con-
vention is not followed in EN 1994-2, where the heading Non-linear global analysis for
bridges (clause 5.4.3) does not include rigid-plastic global analysis. There is no provision
for use of the latter in bridges, so relevant rules are found in the buildings clause 5.4.5 of
EN 1994-1-1.
References from clause 1.5.1(1) include clause 1.5.2 of EN 1992-1-1, which denes pre-
stress as an action caused by the stressing of tendons. This is not sucient for EN 1994-2,
because prestress by jacking at supports, which is outside the scope of EN 1992-1-1, is
within the scope of EN 1994-2.
The denitions in clauses 1.5.1 to 1.5.9 of EN 1993-1-1 apply where they occur in clauses in
EN 1993 to which EN 1994 refers. None of them uses the word steel.
1.5.2. Additional terms and denitions
Most of the 15 denitions in clause 1.5.2 include the word composite. The denition of
shear connection does not require the absence of separation or slip at the interface
between steel and concrete. Separation is always assumed to be negligible, but explicit allow-
ance may need to be made for eects of slip, for example in clauses 5.4.3, 6.6.2.3 and 7.2.1.
The denition of composite frame is relevant to the use of Section 5. Where the behaviour
is essentially that of a reinforced or prestressed concrete structure, with only a few composite
members, global analysis should be generally in accordance with EN 1992.
These lists of denitions are not exhaustive, because all the codes use terms with precise
meanings that can be inferred from their contexts.
Concerning use of words generally, there are signicant dierences from British codes.
These arose from the use of English as the base language for the drafting process, and the
resulting need to improve precision of meaning, to facilitate translation into other European
languages. In particular:
.
action means a load and/or an imposed deformation
.
action eect (clause 5.4) and eect of action have the same meaning: any deformation
or internal force or moment that results from an action.
1.6. Symbols
The symbols in the Eurocodes are all based on ISO standard 3898.
29
Each code has its own
list, applicable within that code. Some symbols have more than one meaning, the particular
meaning being stated in the clause. A few rarely-used symbols are dened only in clauses
where they appear (e.g. A
c;eff
in 7.5.3(1)).
There are a few important changes from previous practice in the UK. For example, an xx
axis is along a member, a yy axis is parallel to the anges of a steel section (clause 1.7(2)
of EN 1993-1-1), and a section modulus is W, with subscripts to denote elastic or plastic
behaviour.
This convention for member axes is more compatible with most commercially available
analysis packages than that used in previous British bridge codes. The yy axis generally
represents the major principal axis, as shown in Fig. 1.1(a) and (b). Where this is not a
principal axis, the major and minor principal axes are denoted uu and vv, as shown in
Fig. 1.1(c). It is possible for the major axis of a composite cross-section to be the minor
axis of its structural steel component.
Clause 1.5.1(1)
Clause 1.5.2
8
Wherever possible, denitions in EN 1994-2 have been aligned with those in ENs 1990,
1992 and 1993; but this should not be assumed without checking the list in clause 1.6.
Some quite minor dierences are signicant.
The symbol f
y
has dierent meanings in ENs 1992 and 1993. It is retained in EN 1994-2 for
the nominal yield strength of structural steel, though the generic subscript for that material
is a, based on the French word for steel, acier. Subscript a is not used in EN 1993, where
the partial factor for steel is not
A
, but
M
. The symbol
M
is also used in EN 1994-2. The
characteristic yield strength of reinforcement is f
sk
, with partial factor
S
.
The use of upper-case subscripts for factors for materials implies that the values given
allow for two types of uncertainty: in the properties of the material and in the resistance
model used.
Clause 1.6
y y
z
z
v
y
y
z
z
u
u
v
z
z
y y
(a) (b) (c)
Fig. 1.1. Sign convention for axes of members
9
CHAPTER 1. GENERAL
CHAPTER 2
Basis of design
The material described in this chapter is covered in Section 2 of EN 1994-2, in the following
clauses:
.
Requirements Clause 2.1
.
Principles of limit states design Clause 2.2
.
Basic variables Clause 2.3
.
Verication by the partial factor method Clause 2.4
The sequence follows that of EN 1990, Sections 2 to 4 and 6.
2.1. Requirements
Design is to be in accordance with the general requirements of EN 1990. The purpose of
Section 2 is to give supplementary provisions for composite structures.
Clause 2.1(3) reminds the user again that design is based on actions in accordance with
EN 1991, combinations of actions and load factors at the various limit states in accordance
with EN 1990 (Annex A2), and the resistances, durability and serviceability provisions of
EN 1994 (through extensive references to EC2 and EC3).
The use of partial safety factors for actions and resistances (the partial factor method)
is expected but is not a requirement of Eurocodes. The method is presented in Section 6
of EN 1990 as one way of satisfying the basic requirements set out in Section 2 of that
standard. This is why use of the partial factor method is given deemed to satisfy status in
clause 2.1(3). To establish that a design was in accordance with the Eurocodes, the user
of any other method would normally have to demonstrate, to the satisfaction of the
regulatory authority and/or the client, that the method satised the basic requirements of
EN 1990.
The design working life for bridges and components of bridges is also given in EN 1990.
This predominantly aects calculations on fatigue. Temporary structures (that will not be
dismantled and reused) have an indicative design life of 10 years, while bearings have a
life of 1025 years and a permanent bridge has an indicative design life of 100 years. The
design lives of temporary bridges and permanent bridges can be varied in project speci-
cations and the National Annex respectively. For political reasons, the design life for per-
manent bridges in the UK may be maintained at 120 years.
To achieve the design working life, bridges and bridge components should be designed
against corrosion, fatigue and wear and should be regularly inspected and maintained.
Where components cannot be designed for the full working life of the bridge, they need to
be replaceable. Further detail is given in Chapter 4 of this guide.
Clause 2.1(3)
2.2. Principles of limit states design
The clause provides a reminder that it is important to check strength and stability through-
out all stages of construction in addition to the nal condition. The strength of bare steel
beams during pouring of the deck slab must be checked, as the restraint to the top ange pro-
vided by the completed deck slab is absent in this condition.
A beam that is in Class 1 or 2 when completed may be in Class 3 or 4 during construction,
if a greater depth of web is in compression. Its stresses must then be built up allowing for the
construction history. For cross-sections that are in Class 1 or 2 when completed, nal
verications of resistances can be based on accumulation of bending moments and shear
forces, rather than stresses, as plastic bending resistances can be used. The serviceability
checks would still necessitate consideration of the staged construction.
All resistance formulae for composite members assume that the specied requirements for
materials, such as ductility, fracture toughness and through-thickness properties, are met.
2.3. Basic variables
Clause 2.3.1 on actions refers only to EN 1991. Its Part 2, Trac loads on bridges, denes
load patterns and leaves clients, or designers, much choice over intensity of loading. Loads
during construction are specied in EN 1991-1-6, Actions during execution.
30
Actions include imposed deformations, such as settlement or jacking of supports, and
eects of temperature and shrinkage. Further information is given in comments on clause
2.3.3.
Clause 2.3.2(1) refers to EN 1992-1-1 for shrinkage and creep of concrete, where detailed
and quite complex rules are given for prediction of free shrinkage strain and creep
coecients. These are discussed in comments on clauses 3.1 and 5.4.2.2. Eects of creep of
concrete are not normally treated as imposed deformations. An exception arises in clause
5.4.2.2(6).
The classication of eects of shrinkage and temperature in clause 2.3.3 into primary and
secondary will be familiar to designers of continuous beams. Secondary eects are to be
treated as indirect actions, which are sets of imposed deformations (clause 1.5.3.1 of
EN 1990), not as action eects. This distinction is relevant in clause 5.4.2.2(7), where
indirect actions may be neglected in analyses for some verications of composite members
with all cross-sections in Class 1 or 2. This is because resistances are based on plastic
analysis and there is therefore adequate rotation capacity to permit the eects of imposed
deformations to be released.
2.4. Verication by the partial factor method
2.4.1. Design values
Clause 2.4.1 illustrates the treatment of partial factors. Recommended values are given in
Notes, in the hope of eventual convergence between the values for each partial factor that
will be specied in the national annexes. This process was adopted because the regulatory
bodies in the member states of CEN, rather than CEN itself, are responsible for setting
safety levels. The Notes are informative, not normative (i.e. not part of the preceding
provision), so that there are no numerical values in the principles, as explained earlier.
The Note below clause 2.4.1.1(1) recommends
P
1:0 (where subscript P represents
prestress) for controlled imposed deformations. Examples of these include jacking up at
supports or jacking down by the removal of packing plates. The latter might be done to
increase the reaction at an adjacent end support where there is a risk of uplift occurring.
The Notes to clause 2.4.1.2 link the partial factors for concrete, reinforcing steel and struc-
tural steel to those recommended in EN 1992-1-1 and EN 1993. Design would be more
dicult if the factors for these materials in composite structures diered from the values
in reinforced concrete and steel structures. The reference to EN 1993, as distinct from
EN 1993-1-1, is required because some
M
factors dier for bridges and buildings.
Clause 2.3.1
Clause 2.3.2(1)
Clause 2.3.3
Clause 2.4.1
Clause 2.4.1.1(1)
Clause 2.4.1.2
12
The remainder of EN 1994-2 normally refers to design strengths, rather than to character-
istic or nominal values with partial factors. Characteristic values are 5% lower fractiles for
an innite test series, predicted from experience and a smaller number of tests. Nominal
values (e.g. the yield strength of structural steel) are used where distributions of test
results cannot be predicted statistically. They are chosen to correspond to characteristic
values.
The design strength for concrete is given by:
f
cd
f
ck
=
C
2:1
where f
ck
is the characteristic cylinder strength. This denition is stated algebraically because
it diers from that of EN 1992-2, in which an additional coecient
cc
is applied:
f
cd

cc
f
ck
=
C
D2:1
The coecient is explained in EN 1992-2 as taking account of long-term eects and of
unfavourable eects resulting from the way the load is applied. The value for
cc
is to be
given in national annexes to EN 1992-2, and should lie between 0.80 and 1.00. The value
1.00 has been used in EN 1994-2, without permitting national choice, for several reasons:
.
The plastic stress block for use in resistance of composite sections, dened in clause
6.2.1.2, consists of a stress 0.85f
cd
extending to the neutral axis, as shown in Fig. 2.1.
The depth of the stress block in EN 1992-2 is only 80% of this distance. The factor
0.85 is not fully equivalent to
cc
; it allows also for the dierence between the stress
blocks.
.
Predictions using the stress block of EN 1994 have been veried against test results for
composite members conducted independently from verications for concrete bridges.
.
The EN 1994 block is easier to apply. The Eurocode 2 rule was not used in Eurocode 4
because resistance formulae become complex where the neutral axis is close to or within
the steel ange adjacent to the concrete slab.
.
Resistance formulae for composite elements given in EN 1994 are based on calibrations
using its stress block, with
cc
1:0.
The denition of f
cd
in equation (2.1) is applicable to verications of all composite cross-
sections, but not where the section is reinforced concrete only; for example, in-plane shear in
a concrete ange of a composite beam. For reinforced concrete, EN 1992-2 applies, with
cc
in equation (D2.1) as given in the National Annex. It is expected that the rules in the UKs
Annex will include:
cc
0:85 for flexure and axial compression
This is consistent with EN 1994-2, as the coecient 0.85 appears in the resistance formulae
in clauses 6.2.1.2 and 6.7.3.2. In these cases, the values 0.85f
cd
in EN 1994-2 and f
cd
in
EN 1992-2 are equal, so the values of symbols f
cd
are not equal. There is a risk of error
when switching between calculations for composite sections and for reinforced concrete
elements such as a deck slab both for this reason and because of the dierent depth of
stress block.
Plastic
neutral axis
Stress to
EN 1994
0.85f
ck
/
C

cc
f
ck
/
C
Stress to EN 1992,
for f
ck
50 N/mm
2
Strain
x
0.8x
cu3
f
yd
f
yd
Fig. 2.1. Rectangular stress blocks for concrete in compression at ultimate limit states
13
Care is needed also with symbols for steels. The design strengths in EN 1994 are f
yd
for
structural steel and f
sd
for reinforcement, but reinforcement in EN 1992 has f
yd
, not f
sd
.
The recommended partial factors given in EN 1992-2 (referring to EN 1992-1-1) for
materials for ultimate limit states other than fatigue are repeated in Table 2.1. For service-
ability limit states, the recommended value is generally 1.0, from clause 2.4.2.4(2).
The
M
values for structural steel are denoted
M0
to
M7
in clause 6.1 of EN 1993-2. Those
for ultimate limit states other than fatigue are given in Table 2.2. Further values are given in
clauses on fatigue. No distinction is made between persistent, transient, and accidental
design situations, though it could be, in a national annex.
For simplicity,
M
for resistances of shear connectors (denoted
V
), given in a Note to
clause 6.6.3.1(1), was standardised at 1.25, because this is the recommended value for
most joints in steelwork. Where calibration led to a dierent value, a coecient in the resis-
tance formula was modied to enable 1.25 to be used.
Clause 2.4.1.3 refers to product standards hEN and to nominal values. The h stands
for harmonised. This term from the Construction Products Directive
31
is explained in the
Designers Guide to EN 1990.
1
Generally, global analysis and resistances of cross-sections may be based on the nominal
values of dimensions, which are given on the project drawings or quoted in product stan-
dards. Geometrical tolerances as well as structural imperfections (such as welding residual
stresses) are accounted for in the methods specied for global analyses and for buckling
checks of individual structural elements. These subjects are discussed further in sections
5.2 and 5.3, respectively, of this guide.
Clause 2.4.1.4, on design resistances to particular action eects, refers to expressions (6.6a)
and (6.6c) given in clause 6.3.5 of EN 1990. Resistances in EN 1994-2 often need more than
one partial factor, and so use expression (6.6a) which is:
R
d
Rf
i
X
k;i
=
M;i
; a
d
g i ! 1 D2:2
Clause 2.4.1.3
Clause 2.4.1.4
Table 2.2. Partial factors from EN 1993-2 for materials, for ultimate limit states
Resistance type Factor
Recommended
value
Resistance of members and cross-sections
.
Resistance of cross-sections to excessive yielding including local buckling
M0
1.00
.
Resistance of members to instability assessed by member checks
M1
1.10
.
Resistance to fracture of cross-sections in tension
M2
1.25
Resistance of joints
.
Resistance of bolts, rivets, pins and welds
.
Resistance of plates in bearing
M2
M2
1.25
1.25
.
Slip resistance:
at an ultimate limit state
at a serviceability limit state
M3
M3;ser
1.25
1.10
.
Bearing resistance of an injection bolt
M4
1.10
.
Resistance of joints in hollow section lattice girders
M5
1.10
.
Resistance of pins at serviceability limit state
M6;ser
1.00
.
Pre-load of high-strength bolts
M7
1.10
Table 2.1. Partial factors from EN 1992-2 for materials, for ultimate limit states
Design situations
C
, for concrete
S
, reinforcing steel
S
, prestressing steel
Persistent and transient 1.5 1.15 1.15
Accidental 1.2 1.0 1.0
14
For example, clause 6.7.3.2(1) gives the plastic resistance to compression of a cross-
section as the sum of terms for the structural steel, concrete and reinforcement:
N
pl;Rd
A
a
f
yd
0:85A
c
f
cd
A
s
f
sd
6:30
In this case, there is no separate term a
d
for the inuence of geometrical data on resistance,
because uncertainties in areas of cross-sections are allowed for in the
M
factors.
In terms of characteristic strengths, from clause 2.4.1.2, equation (6.30) becomes:
N
pl;Rd
A
a
f
y
=
M
0:85A
c
f
ck
=
C
A
s
f
sk
=
S
D2:3
where:
the characteristic material strengths X
k;i
are f
y
, f
ck
and f
sk
;
the conversion factors,
i
in EN 1990, are 1.0 for steel and reinforcement and 0.85 for
concrete. These factors enable allowance to be made for the dierence between the
material property obtained from tests and its in situ contribution to the particular
resistance considered. In general, it is also permissible to allow for this eect in the
values of
M;i
;
the partial factors
M;i
are written
M
,
C
and
S
in EN 1994-2.
Expression (6.6c) of EN 1990 is:
R
d
R
k
=
M
It applies where characteristic properties and a single partial factor can be used; for example,
in expressions for the shear resistance of a headed stud (clause 6.6.3.1). It is widely used in
EN 1993, where only one material, steel, contributes to a resistance.
2.4.2. Combination of actions
Clause 2.4.2 refers to the combinations of actions given in EN 1990. As in current practice,
variable actions are included in a combination only in regions where they contribute to the
total action eect considered.
For permanent actions and ultimate limit states, the situation is more complex. Normally
the same factor
F
(favourable or unfavourable as appropriate) is applied throughout the
structure, irrespective of whether both favourable and unfavourable loading regions exist.
Additionally, the characteristic action is a mean (50% fractile) value. Exceptions are
covered by clause 6.4.3.1(4)P of EN 1990:
Where the results of a verication are very sensitive to variations of the magnitude of a permanent
action from place to place in the structure, the unfavourable and the favourable parts of this
action shall be considered as individual actions.
A design permanent action is then
Ed;min
G
k;min
in a favourable region, and
Ed;max
G
k;max
in an unfavourable region. Recommendations on the choice of these values and the
application of this principle are given in EN 1990, with guidance in the Designers Guide
to EN 1990.
1
2.4.3. Verication of static equilibrium (EQU)
The preceding quotation from EN 1990 evidently applies to checks on static equilibrium,
clause 2.4.3(1). It draws attention to the role of anchors and bearings in ensuring static
equilibrium.
The abbreviation EQU in this clause comes from EN 1990, where four types of ultimate
limit state are dened in clause 6.4.1:
.
EQU for loss of static equilibrium
.
FAT for fatigue failure
Clause 2.4.2
Clause 2.4.3(1)
15
.
GEO for failure or excessive deformation of the ground
.
STR for internal failure or excessive deformation of the structure.
As explained above, the main feature of EQU is that, unlike STR, the partial factor
F
for
permanent actions is not uniform over the whole structure. It is higher for destabilizing
actions than for those relied on for stability. This guide mainly covers ultimate limit states
of types STR and FAT. Use of type GEO arises in design of foundations to EN 1997.
32
16
CHAPTER 3
Materials
This chapter concerns the properties of materials needed for the design of composite
structures. It corresponds to Section 3, which has the following clauses:
.
Concrete Clause 3.1
.
Reinforcing steel for bridges Clause 3.2
.
Structural steel for bridges Clause 3.3
.
Connecting devices Clause 3.4
.
Prestressing steel and devices Clause 3.5
.
Tension components in steel Clause 3.6
Rather than repeating information given elsewhere, Section 3 consists mainly of
cross-references to other Eurocodes and EN standards. The following comments relate to
provisions of particular signicance for composite structures.
3.1. Concrete
Clause 3.1(1) refers to EN 1992-1-1 for the properties of concrete. For lightweight-aggregate
concrete, several properties are dependent on the oven-dry density, relative to 2200 kg/m
3
.
Comprehensive sets of time-dependent properties are given in its clause 3.1 for normal
concrete and clause 11.3 for lightweight-aggregate concrete. For composite structures built
unpropped, with several stages of construction, simplication may be needed. A simplica-
tion for considerations of creep is provided in clause 5.4.2.2(2). Specic properties are now
discussed. (For thermal expansion, see Section 3.3 below.)
Compressive strength
Strength and deformation characteristics are summarized in EN 1992-1-1, Table 3.1 for
normal concrete and Table 11.3.1 for lightweight-aggregate concrete.
Strength classes for normal concrete are dened as Cx/y, where x and y are respectively the
cylinder and cube compressive strengths in N/mm
2
units, determined at age 28 days. All
compressive strengths in design rules in Eurocodes are cylinder strengths, so an unsafe
error occurs if a specied cube strength is used in calculations. It should be replaced at the
outset by the equivalent cylinder strength, using the relationships given by the strength
classes.
Most cube strengths in Table 3.1 are rounded to 5 N/mm
2
. The ratios f
ck
,f
ck.cube
range
from 0.78 to 0.83, for grades up to C70/85.
Classes for lightweight concrete are designated LCx/y. The relationships between cylinder
and cube strengths dier from those of normal concrete; for example, C40/50 and LC40/44.
The ratios f
ck
,f
ck.cube
for the LC grades range from 0.89 to 0.92. Thus, cylinder strengths are
about 80% of cube strengths for normal-weight concrete and 90% for lightweight concrete.
Clause 3.1(1)
Comment on the design compressive strength, f
cd
f
ck
,
C
, is given at clause 2.4.1.2.
Tensile strength
EN 1992 denes concrete tensile strength as the highest stress reached under concentric
tensile loading. Values for the mean axial tensile strength of normal-weight concrete at 28
days, f
ctm
, are given in Table 3.1 of EN 1992-1-1. They are based on the following formulae,
in N/mm
2
units:
f
ctm
0.30 f
ck
2,3
. f
ck
C50,60 D3.1
f
ctm
2.12 ln1 f
cm
,10. f
ck
C50,60 D3.2
This table also gives the 5% and 95% fractile values for tensile strength. The appropriate
fractile value should be used in any limit state verication that relies on either an adverse or
benecial eect of the tensile strength of concrete. Tensile strengths for lightweight concrete
are given in Table 11.3.1 of EN 1992-1-1.
Mean tensile stress, f
ctm
, is used in several places in EN 1994-2 where the eects of tension
stiening are considered to be important. These include:
.
clause 5.4.2.3(2): rules on allowing for cracking in global analysis
.
clause 5.4.2.8(6): calculation of internal forces in concrete tension members in bowstring
arches
.
clause 5.5.1(5): minimum area of reinforcement required in concrete tension anges of
composite beams
.
clause 7.4.2(1): rules on minimum reinforcement to ensure that cracking does not cause
yielding of reinforcement in the cracked region
.
clause 7.4.3(3): rules on crack width calculation to allow for the increase in stress in re-
inforcement caused by tension stiening.
Elastic deformation
All properties of concrete are inuenced by its composition. The values for the mean short-
term modulus of elasticity in Tables 3.1 and 11.3.1 of EN 1992-1-1 are given with a warning
that they are indicative and should be specically assessed if the structure is likely to be
sensitive to deviations from these general values.
The values are for concrete with quartzite aggregates. Corrections for other types of
aggregate are given in EN 1992-1-1, clause 3.1.3(2). All these are secant values; typically,
0.4f
cm
/(strain at 0.4f
cm
), and so are slightly lower than the initial tangent modulus,
because stressstrain curves for concrete are non-linear from the origin.
Table 3.1 in EN 1992-1-1 gives the analytical relation:
E
cm
22 f
ck
8,10
0.3
with E
cm
in GPa or kN/mm
2
units, and f
ck
in N/mm
2
. For f
ck
30, this gives E
cm
32.8
kN/mm
2
, whereas the entry in the table is rounded to 33 kN/mm
2
.
A formula for the increase of E
cm
with time, in clause 3.1.3(3) of EN 1992-1-1, gives the
two-year value as 6% above E
cm
at 28 days. The inuence in a composite structure of so
small a change is likely to be negligible compared with the uncertainties in the modelling
of creep.
Clause 3.1(2) limits the scope of EN 1994-2 to the strength range C20/25 to C60/75
for normal concrete and from LC20/22 to LC60/66 for lightweight concrete. The upper
limits to these ranges are lower than that given in EN 1992-2 (C70/85) because there is
limited knowledge and experience of the behaviour of composite members with very
strong concrete. This applies, for example, to the load/slip properties of shear connectors,
the redistribution of moments in continuous beams and the resistance of columns. The use
of rectangular stress blocks for resistance to bending (clause 6.2.1.2(d)) relies on the
strain capacity of the materials. The relevant property of concrete in compression,
cu3
in
Table 3.1 of EN 1992-1-1, is 0.0035 for classes up to C50/60, but then falls, and is only
0.0026 for class C90/105.
Clause 3.1(2)
18
Shrinkage
The shrinkage of concrete referred to in clause 3.1(3) is (presumably) both the drying
shrinkage that occurs after setting and the autogenous shrinkage, but not the plastic
shrinkage that precedes setting.
Drying shrinkage is associated with movement of water through and out of the concrete
and therefore depends on relative humidity and eective section thickness as well as on
the concrete mix. It takes several years to be substantially complete. The mean drying
shrinkage strain (for unreinforced concrete) is given in clause 3.1.4(6) of EN 1992-1-1 as a
function of grade of concrete, ambient relative humidity, eective thickness of the concrete
cross-section, and elapsed time since the end of curing. It is stated that actual values have a
coecient of variation of about 30%. This implies a 16% probability that the shrinkage will
exceed the prediction by at least 30%.
A slightly better predictor is given in Annex B of EN 1992-1-1, as the type of cement is
included as an additional parameter.
Autogenous shrinkage develops during the hydration and hardening of concrete. It is that
which occurs in enclosed or sealed concrete, as in a concrete-lled steel tube, where no loss of
moisture occurs. This shrinkage strain depends only on the strength of the concrete, and is
substantially complete in a few months. It is given in clause 3.1.4(6) of EN 1992-1-1 as a
function of concrete grade and the age of the concrete in days. The time coecient given
is 1 exp0.2t
0.5
, so this shrinkage should be 90% complete at age 19 weeks. The
90% shrinkage strain for a grade C40/50 concrete is given as 67 10
6
. It has little inuence
on cracking due to direct loading, and the rules for minimum reinforcement (clause 7.4.2)
take account of its eects.
The rules in EN 1992-1-1 become less accurate at high concrete strengths, especially if the
mix includes silica fume. Data for shrinkage for concrete grades C55/67 and above are given
in informative Annex B of EN 1992-2.
Section 11 of EN 1992-2 gives supplementary requirements for lightweight concretes.
The shrinkage of reinforced concrete is lower than the free shrinkage, to an extent that
depends on the reinforcement ratio. The dierence is easily calculated by elastic theory, if
the concrete is in compression. In steelconcrete composite bridges, restraint of reinforced
concrete shrinkage by the structural steel leads to locked-in stresses in the composite
section. In indeterminate bridges, secondary moments and forces from restraint to the free
deections also occur. Shrinkage, being a permanent action, occurs in every combination
of actions. It increases hogging moments at internal supports, often a critical region, and
so can inuence design.
The specied shrinkage strains will typically be found to be greater than that used in
previous UK practice, but the recommended partial load factor, in clause 2.4.2.1 of
EN 1992-1-1, is
SH
1.0, lower than the value of 1.2 used in BS 5400.
There is further comment on shrinkage in Chapter 5.
Creep
In EN 1994-2, the eects of creep are generally accounted for using an eective modulus of
elasticity for the concrete, rather than by explicit calculation of creep deformation. However,
it is still necessary to determine the creep coecient ct. t
0
(denoted c
t
in EN 1994) from
clause 3.1.4 of EN 1992-1-1. Guidance on deriving modular ratios is given in section 5.4.2
of this guide.
3.2. Reinforcing steel for bridges
For properties of reinforcement, clause 3.2(1) refers to clause 3.2 of EN 1992-1-1, which in
turn refers to its normative Annex C for bond characteristics. EN 1992 allows the use of bars,
de-coiled rods and welded fabric as suitable reinforcement. Its rules are applicable to ribbed
and weldable reinforcement only, and therefore cannot be used for plain round bars. The
rules are valid for characteristic yield strengths between 400 N/mm
2
and 600 N/mm
2
. Wire
fabrics with nominal bar size 5 mm and above are included. Exceptions to the rules for
Clause 3.1(3)
Clause 3.2(1)
19
fatigue of reinforcement may be given in the National Annex, and could refer to the use of
wire fabric.
In this section 3.2, symbols f
yk
and f
yd
are used for the yield strengths of reinforcement, as
in EN 1992, although f
sk
and f
sd
are used in EN 1994, to distinguish reinforcement from
structural steel.
The grade of reinforcement denotes the specied characteristic yield strength, f
yk
. This
is obtained by dividing the characteristic yield load by the nominal cross-sectional area of
the bar. Alternatively, for products without a pronounced yield stress, the 0.2% proof
stress, f
0.2k
may be used in place of the yield stress.
Elastic deformation
For simplicity, clause 3.2(2) permits the modulus of elasticity of reinforcement to be taken as
210 kN/mm
2
, the value given in EN 1993-1-1 for structural steel, rather than 200 kN/mm
2
,
the value in EN 1992-1-1. This simplication means that it is not necessary to transform
reinforcement into structural steel or vice versa when calculating cracked section properties
of composite beams.
Ductility
Clause 3.2(3) refers to clause 3.2.4 of EN 1992-2; but provisions on ductility in Annex C of
EN 1992-1-1 also apply. Reinforcement shall have adequate ductility, dened by the ratio of
tensile strength to the yield stress, f
t
,f
y
k
, and the strain at maximum force,
uk
. The
requirements for the three classes for ductility are given in Table 3.1, from EN 1992-1-1.
Clause 3.2.4(101)P of EN 1992-2 recommends that Class A reinforcement is not used for
bridges, although this is subject to variation in the National Annex. The reason is that high
strain can occur in reinforcement in a reinforced concrete section in exure before the
concrete crushes. Clause 5.5.1(5) prohibits the use of Class A reinforcement in composite
beams which are designed as either Class 1 or 2 for a similar reason: namely, that very high
strains in reinforcement are possible due to plastication of the whole composite section.
Class 3 and 4 sections are limited to rst yield in the structural steel and so the reinforce-
ment strain is limited to a relatively low value. The recommendations of EN 1992 and
EN 1994 lead to some ambiguity with respect to ductility requirements for bars in reinforced
concrete deck slabs forming part of a composite bridge with Class 3 or 4 beams. Where main
longitudinal bars in the deck slab of a composite section are signicantly stressed by local
loading, it would be advisable to follow the recommendations of EN 1992 and not to use
Class A reinforcement.
Stressstrain curves
The characteristic stressstrain diagram and the two alternative design diagrams dened in
clause 3.2.7 of EN 1992-1-1 are shown in Fig. 3.1. The design diagrams (labelled B in Fig. 3.1)
have:
(a) an inclined top branch with a strain limit of
ud
and a maximum stress of kf
yk
,
S
at
uk
(for symbols k and
uk
, see Table 3.1), and
(b) a horizontal top branch without strain limit.
A value for
ud
may be found in the National Annex to EN 1992-1-1, and is recommended as
0.9
uk
.
Clause 3.2(2)
Clause 3.2(3)
Table 3.1. Ductility classes for reinforcement
Class
Characteristic strain at
maximum force,
uk
(%)
Minimum value
of k f
t
,f
y
k
A !2.5 !1.05
B !5 !1.08
C !7.5 !1.15, <1.35
20
From clause 6.2.1.4, reinforcement diagram (a) is only relevant when the non-linear
method for bending resistance is used. Elastic and plastic bending resistances assume that
the reinforcement stress is limited to the design yield strength.
The minimum ductility properties of wire fabric given in Table C.1 of EN 1992-1-1 may
not be sucient to satisfy clause 5.5.1(6), as this requires demonstration of sucient ducti-
lity to avoid fracture when built into a concrete slab. It has been found in tests on continuous
composite beams with fabric in tension that the cross-wires initiate cracks in concrete, so that
tensile strain becomes concentrated at the locations of the welds in the fabric.
33
3.3. Structural steel for bridges
Clause 3.3(1) refers to EN 1993-2, which in turn refers to EN 1993-1-1. This lists in its Table
3.1 steel grades with nominal yield strengths up to 460 N/mm
2
, and allows other steel
products to be included in national annexes. The nominal values of material properties
have to be adopted as characteristic values in all design calculations.
Two options for selecting material strength are provided. Either the yield strength and
ultimate strength should be obtained from the relevant product standard or the simplied
values provided in Table 3.1 of EN 1993-1-1 should be used. The National Annex for
EC3-1-1 may make this choice. In either case, the strength varies with thickness, and the
appropriate thickness must be used when determining the strength.
The elastic constants for steel, given in clause 3.2.6 of EN 1993-1-1, are familiar values. In
the notation of EN 1994, they are: E
a
210 kN/mm
2
, G
a
81 kN/mm
2
, and i
a
0.3.
Moduli of elasticity for tension rods and cables of dierent types are not covered by this
clause and are given in EN 1993-1-11.
Clause 3.3(2) sets the same upper limit to nominal yield strength as in EN 1993-1-1,
namely 460 N/mm
2
, for use in composite bridges. EN 1993-1-12 covers steels up to grade
S700. A comprehensive report on high-performance steels appeared in 2005,
34
and there
has been extensive research on the use in composite members of structural steels with
yield strengths exceeding 355 N/mm
2
.
3537
It was found that some design rules need modi-
cation for use with steel grades higher than S355, to avoid premature crushing of concrete.
This applies to:
.
plastic resistance moment (clause 6.2.1.2(2)), and
.
resistance of columns (clause 6.7.3.6(1)).
Ductility
Many design clauses in EN 1994 rely on the ductile behaviour of structural steel after yield.
Ductility is covered by the references in clause 3.3(1) to EN 1993.
The ductility characteristics required by clause 3.2.2 of EN 1993-1-1 are for a minimum
ratio f
u
,f
y
of the specied values; a minimum elongation; and a minimum strain at the
Clause 3.3(1)
Clause 3.3(2)
0
B
A idealised
B design
A
kf
yk
f
yk
kf
yk
kf
yk
/
S
f
yd
/E
s

ud

uk

f
yd
= f
yk
/
S
k = ( f
t
/ f
y
)
k
Fig. 3.1. Characteristic and design stressstrain diagrams for reinforcement (tension and compression)
21
specied ultimate tensile strength, f
u
. Recommended values are given, all of which can be
modied in the National Annex. The steel grades in Table 3.1 of EN 1993-1-1 all provide
the recommended level of ductility. It follows that the drafting of this part of a national
annex to EN 1993-1-1 should consider both steel and composite structures.
Thermal expansion
For the coecient of linear thermal expansion of steel, clause 3.2.6 of EN 1993-1-1 gives a
value of 12 10
6
per 8C (also written in Eurocodes as /K or K
1
). This is followed by
a Note that for calculating the structural eects of unequal temperatures in composite
structures, the coecient may be taken as 10 10
6
per 8C, which is the value given for
normal-weight concrete in clause 3.1.3(5) of EN 1992-1-1. This avoids the need to calculate
the internal restraint stresses from uniform temperature change, which would result from
dierent coecients of thermal expansion for steel and concrete. Movement due to
change of uniform temperature (or force due to restraint of movement) should however
be calculated using c 12 10
6
per 8C for all the structural materials (clause 5.4.2.5(3)).
Thermal expansion of reinforcement is not mentioned in EN 1992-1-1, presumably
because it is assumed to be the same as that of normal-weight concrete. For reinforcement
in composite members the coecient should be taken as 10 10
6
per 8C. This is not in
EN 1994.
Coecients of thermal expansion for lightweight-aggregate concretes can range from
4 10
6
to 14 10
6
per 8C. Clause 11.3.2(2) of EN 1992-1-1 states that: The dierences
between the coecients of thermal expansion of steel and lightweight aggregate concrete
need not be considered in design, but steel here means reinforcement, not structural
steel. The eects of the dierence from 10 10
6
per 8C should be considered in design of
composite members for situations where the temperatures of the concrete and the structural
steel could dier signicantly.
3.4. Connecting devices
3.4.1. General
Reference is made to EN 1993, Eurocode 3: Design of Steel Structures, Part 1-8: Design of
Joints
19
for information relating to fasteners, such as bolts, and welding consumables. Provi-
sions for other types of mechanical fastener are given in clause 3.3 of EN 1993-1-3.
38
Composite joints
Composite joints are dened in clause 1.5.2.8. In bridges, they are essentially steelwork joints
across which a reinforced or prestressed concrete slab is continuous, and cannot be ignored.
Composite joints are covered in Section 8 and Annex A of EN 1994-1-1, with extensive
reference to EN 1993-1-8. These clauses are written for buildings, and so are not copied
into EN 1994-2, though many of them are relevant. Commentary on them will be found
in Chapters 8 and 10 of the Designers Guide to EN 1994-1-1.
5
The joints classied as rigid or full-strength occur also in bridge construction. Where
bending resistances of beams in Class 1 or 2 are determined by plastic theory, joints in
regions of high bending moment must either have sucient rotation capacity, or be stronger
than the weaker of the members joined. The rotation capacity needed in bridges, where
elastic global analysis is always used, is lower than in buildings.
Tests, mainly on beam-to-column joints, have found that reinforcing bars of diameter up
to 12 mm may fracture. Clause 5.5.1 gives rules for minimum reinforcement that apply also
to joints, but does not exclude small-diameter bars.
3.4.2. Headed stud shear connectors
Headed studs are the only type of shear connector for which detailed provisions are given in
EN 1994-2, throughout clause 6.6. Their use is referred to elsewhere; for example, in clause
6.7.4.2(4). Their performance has been validated for diameters up to 25 mm.
39
Research on
22
larger studs is in progress. Studs attached to steel top anges present a hazard during
construction, and other types of connector are sometimes used.
23
These must satisfy
clause 6.6.1.1, which gives the basis of design for shear connection. Research on perforated
plate connectors (known initially as Perfobond) of S355 and S460 steel in grade C50/60
concrete has found slip capacities from 815 mm, which is better than the 6 mm found for
22-mm studs.
25
The use of adhesives on a steel ange is unlikely to be suitable. See also
the comment on clause 1.1.3(3).
Clause 3.4.2 refers to EN 13918 Welding Studs and Ceramic Ferrules for Arc Stud
Welding.
40
This gives minimum dimensions for weld collars. Other methods of attaching
studs, such as spinning, may not provide weld collars large enough for the resistances of
studs given in clause 6.6.3.1(1) to be applicable.
Shear connection between steel and concrete by bond or friction is permitted only in accor-
dance with clause 6.7.4, for columns.
3.5. Prestressing steel and devices
Properties of materials for prestressing tendons and requirements for anchorage and cou-
pling of tendons are covered in clauses 3.3 and 3.4, respectively, of EN 1992-1-1. Prestressing
by tendons is rarely used for steel and concrete composite members and is not discussed
further.
3.6. Tension components in steel
The scope of EN 1993-1-11 is limited to bridges with adjustable and replaceable steel tension
components. It identies three generic groups: tension rod systems, ropes, and bundles of
parallel wires or strands; and provides information on stiness and other material properties.
The analysis of cable-supported bridges, including treatment of load combinations and non-
linear eects, is also covered. These are not discussed further here but some discussion can be
found in the Designers Guide to EN 1993-2.
4
Clause 3.4.2
23
CHAPTER 4
Durability
This chapter corresponds to Section 4, which has the following clauses:
.
General Clause 4.1
.
Corrosion protection at the steelconcrete interface in bridges Clause 4.2
4.1. General
Almost all aspects of the durability of composite structures are covered by cross-references
in clause 4.1(1) to ENs 1990, 1992 and 1993. Bridges must be suciently durable to remain
serviceable throughout their design life. Clause 2.4 of EN 1990 lists ten factors to be taken
into account, and gives the following general requirement:
The structure shall be designed such that deterioration over its design working life does not
impair the performance of the structure below that intended, having due regard to its environment
and the anticipated level of maintenance.
The specic provisions given in EN 1992 and EN 1993 focus on corrosion protection to re-
inforcement, tendons and structural steel.
Reinforced concrete
The main durability provision in EN 1992 is the specication of concrete cover as a defence
against corrosion of reinforcement and tendons. The following outline of the procedure is for
reinforcement only. In addition to the durability aspect, adequate concrete cover is essential
for the transmission of bond forces and for providing sucient re resistance (which is of
less signicance for bridge design). The minimum cover c
min
to satisfy the durability
requirements is dened in clause 4.4.1.2 of EN 1992-1-1 by the following expression:
c
min
maxfc
min;b
; c
min;dur
c
dur;
c
dur;st
c
dur;add
; 10 mmg D4:1
where: c
min;b
is the minimum cover due to bond requirements and is dened in Table 4.2
of EN 1992-1-1. For aggregate sizes up to 32 mm it is equal to the bar
diameter (or equivalent bar diameter for bundled bars),
c
min;dur
is the minimum cover required for the environmental conditions,
c
dur;
is an additional safety element which EC2 recommends to be 0 mm,
c
dur;st
is a reduction of minimum cover for the use of stainless steel, which, if
adopted, should be applied to all design calculations, including bond. The
recommended value in EC2 without further specication is 0 mm,
c
dur;add
is a reduction of minimum cover for the use of additional protection. This
could cover coatings to the concrete surface or reinforcement (such as
epoxy coating). EC2 recommends taking a value of 0 mm.
Clause 4.1(1)
The minimum cover for durability requirements, c
min;dur
, depends on the relevant
exposure class taken from Table 4.1 of EN 1992-1-1.
There are 18 exposure classes, ranging from X0, no risk of corrosion, to XA3, highly
aggressive chemical environment. It should be noted that a particular element may have
more than one exposure class, e.g. XD3 and XF4. The XF and XA designations aect the
minimum required concrete grade (via EN 1992-1-1 Annex E) and the chemical composition
of the concrete. The XC and XD designations aect minimum cover and crack width
requirements, and XD, XF and XS aect a stress limit for concrete under the characteristic
combination, from clause 7.2(102) of EN 1992-2. The exposure classes most likely to be
appropriate for composite bridge decks are:
.
XC3 for a deck slab protected by waterproong (recommended in clause 4.2(105) of
EN 1992-2)
.
XC3 for a deck slab sot protected from the rain by adjacent girders
.
XC4 for other parts of the deck slab exposed to cyclic wetting and drying
.
XD3 for parapet edge beams in the splash zone of water contaminated with de-icing salts;
and also XF2 or XF4 if exposed to both freezethaw and de-icing agents (recommended
in clause 4.2(106) of EN 1992-2).
Informative Annex E of EN 1992-1-1 gives indicative strength classes (e.g. C30/37) for
each exposure class, for corrosion of reinforcement and for damage to concrete.
The cover c
min;dur
is given in Table 4.4N of EN 1992-1-1 in terms of the exposure class and
the structural class, and the structural class is found from Table 4.3N. These are reproduced
here as Tables 4.1 and 4.2, respectively. Table 4.2 gives modications to the initial structural
class, which is recommended (in a Note to clause 4.4.1.2(5) of EN 1992-1-1) to be class 4,
assuming a service life of 50 years and concrete of the indicative strength.
Taking exposure class XC4 as an example, the indicative strength class is C30/37. Starting
with Structural Class 4, and using Tables 4.1 and 4.2:
.
for 100-year life, increase by 2 to Class 6
.
for use of C40/50 concrete, reduce by 1 to Class 5
.
where the position of the reinforcement is not aected by the construction process, reduce
by 1 to Class 4.
Special quality control (Table 4.2) is not dened, but clues are given in the Notes to Table
4.3N of EN 1992-1-1. Assuming that it will not be provided, the Class is 4, and Table 4.1
gives c
min;dur
30 mm. Using the recommendations that follow equation (D4.1),
c
min
30 mm
The cover to be specied on the drawings, c
nom
, shall include a further allowance for devia-
tion (c
dev
) according to clause 4.4.1.3(1)P of EN 1992-1-1, such that:
c
nom
c
min
c
dev
Table 4.1. Minimum cover c
min;dur
for reinforcement. (Source: based on Table 4.4N of EN 1992-1-1
15
)
Environmental Requirements for c
min
(mm)
Exposure Class (from Table 4.1 of EN 1992-1-1)
Structural Class X0 XC1 XC2/XC3 XC4 XD1/XS1 XD2/XS2 XD3/XS3
1 10 10 10 15 20 25 30
2 10 10 15 20 25 30 35
3 10 10 20 25 30 35 40
4 10 15 25 30 35 40 45
5 15 20 30 35 40 45 50
6 20 25 35 40 45 50 55
26
The value of c
dev
for buildings and bridges is dened in the National Annex and is
recommended in clause 4.4.1.3(2) of EN 1992-1-1 to be taken as 10 mm. This value may
be reduced in situations where accurate measurements of cover achieved can be taken and
non-conforming elements rejected. This could apply to precast units.
Almost all the provisions on cover, but not the process to be followed, can be modied in
the National Annex to EN 1992-1-1.
Structural steel
The rules in Section 4 of EN 1993-1-1 cover the need for access for in-service inspection,
maintenance, and possible reconstruction of parts susceptible to corrosion, wear or
fatigue. Further provisions relevant to fatigue are given in Section 4 of EN 1993-2, and a
list is given of parts that may need to be replaceable. Corrosion allowances for inaccessible
surfaces may be given in the National Annex. Further discussion on durability of structural
steel is presented in the Designers Guide to EN 1993-2.
4
Access to shear connectors is not possible, so they must be protected from corrosion.
Clause 4.1(2) refers to clause 6.6.5, which includes relevant detailing rules, for cover and
for haunches.
4.2. Corrosion protection at the steelconcrete interface
in bridges
The side cover to stud connectors must be at least 50 mm (clause 6.6.5.4(2)). Clause 4.2(1)
requires provision of a minimum of 50 mm of corrosion protection to each edge of a steel
ange at an interface with concrete. This does not imply that the connectors must be pro-
tected.
For precast deck slabs, the reference to Section 8 is to clause 8.4.2, which requires greater
corrosion protection to a steel ange that supports a precast slab without bedding. Normal
UK practice when using Omnia planks has been to extend the corrosion protection a
minimum of 25 mm beyond the plank edge and its seating material, with due allowance
Clause 4.1(2)
Clause 4.2(1)
Table 4.2. Recommended structural classication. (Source: based on Table 4.3N of EN 1992-1-1
15
)
Structural Class
Exposure Class (from Table 4.1 of EN 1992-1-1)
Criterion X0 XC1 XC2/XC3 XC4 XD1 XD2/XS1 XD3/XS2/XS3
Service life of 100 years Increase
class by 2
Increase
class by 2
Increase
class by 2
Increase
class by 2
Increase
class by 2
Increase
class by 2
Increase
class by 2
Strength Class (see
notes 1 and 2)
!C30/37
Reduce
class by 1
!C30/37
Reduce
class by 1
!C35/45
Reduce
class by 1
!C40/50
Reduce
class by 1
!C40/50
Reduce
class by 1
!C40/50
Reduce
class by 1
!C45/55
Reduce
class by 1
Member with slab
geometry (position of
reinforcement not
aected by construction
process)
Reduce
class by 1
Reduce
class by 1
Reduce
class by 1
Reduce
class by 1
Reduce
class by 1
Reduce
class by 1
Reduce
class by 1
Special Quality Control
of the concrete ensured
Reduce
class by 1
Reduce
class by 1
Reduce
class by 1
Reduce
class by 1
Reduce
class by 1
Reduce
class by 1
Reduce
class by 1
Note 1: The strength class and water/cement ratio are considered to be related values. The relationship is subject to a
national code. A special composition (type of cement, w/c value, ne llers) with the intent to produce low permeability may
be considered.
Note 2: The limit may be reduced by one strength class if air entrainment of more than 4% is applied.
27
CHAPTER 4. DURABILITY
for placing tolerance. The connectors are not mentioned. They are usually surrounded by
in situ concrete, whether bedding is used (as is usual) or not. Corrosion protection to the
connectors is not normally required. It is possible that a thick coating could reduce their
stiness in shear.
28
CHAPTER 5
Structural analysis
This chapter corresponds to Section 5 of EN 1994-2, which has the following clauses:
.
Structural modelling for analysis Clause 5.1
.
Structural stability Clause 5.2
.
Imperfections Clause 5.3
.
Calculation of action eects Clause 5.4
.
Classication of cross-sections Clause 5.5
Structural analysis is performed at three levels: global analysis, member analysis and local
analysis. Section 5 of EN 1994-2 covers the structural idealization of bridges and the methods
of global analysis required in dierent situations to determine deformations and internal
forces and moments. It also covers classication of cross-sections of members, for use in
determining resistances by methods given in Sections 6 of EN 1993-2 and EN 1994-2.
Much reference has to be made to other parts of EC3, especially EN 1993-1-5
41
for the
eects of shear lag and plate buckling.
Wherever possible, analyses for serviceability and ultimate limit states use the same
methods. It is therefore more convenient to specify them in a single section, rather than to
include them in Sections 6 and 7.
The division of material between Section 5 and Section 6 (Ultimate limit states) is not
always obvious. Calculation of vertical shear is clearly analysis, but longitudinal shear is
in Section 6. For composite columns, Methods of analysis and member imperfections is
in clause 6.7.3.4. This separation of imperfections in frames from those in columns requires
care, and receives detailed explanation in the Designers Guide to EN 1994-1-1.
5
Two ow charts for global analysis, Figs 5.15 and 5.16, are given, with comments, at the
end of this chapter. They include relevant provisions from Section 6.
5.1. Structural modelling for analysis
5.1.1. Structural modelling and basic assumptions
The clause of EN 1990 referred to in clause 5.1.1(1)P says, in eect, that models shall
be appropriate and based on established theory and practice and that the variables shall
be relevant.
The basic requirement is that analysis should realistically model the expected behaviour of
the bridge and its constituent elements. For composite bridges, important factors in analysis
are the eects on stiness of shear lag and concrete cracking. For composite members,
dierent rules for shear lag apply for concrete anges and for the steel parts. The former
is dealt with in clause 5.4.1.2 and the latter in Section 3 of EN 1993-1-5. They are discussed
in this Guide under clause 5.4.1.2.
The eects of cracking of concrete can be taken into account either by using cracked
section properties in accordance with clause 5.4.2.3 or, for ller-beam decks only, by
Clause 5.1.1(1)P
redistributing the moments determined from an uncracked analysis away from the cracked
sections in accordance with clause 5.4.2.9. For Class 4 sections, plate buckling eects, which
have to be considered in accordance with clause 2.2 of EN 1993-1-5, can also lead to a reduc-
tion in stiness of cross-sections. This is discussed in this Guide under clause 5.4.1.1.
Global analysis can be signicantly aected by exibility at connections and by interaction
of the bridge structure with the soil, particularly in fully integral bridges. Guidance on
modelling joints and groundstructure interaction are given in clauses 5.1.2 and 5.1.3,
respectively.
Composite members and joints are commonly used in conjunction with others of structural
steel. Clause 5.1.1(2) makes clear that this is the type of construction envisaged in Section 5.
Signicant dierences between Sections 5 of EC3 and EC4 are referred to in this chapter.
5.1.2. Joint modelling
In analysis of bridges, it is generally possible to treat joints as either rigid or pinned, as
appropriate. Clause 5.1.2(1) refers to semi-continuous joints as an exception. They are
neither rigid nor pinned, and have sucient exibility to inuence the bending moment
transmitted. This could occur, for example, from the exure of thin end-plates in a bolted
end-plate connection.
The three simplied joint models listed in clause 5.1.2(2) simple, continuous and semi-
continuous are those given in EN 1993. Joints in steelwork have their own Eurocode part,
EN 1993-1-8.
19
Its design methods are for joints subjected to predominantly static loading
(its clause 1.1(1)). Resistance to fatigue is covered in EN 1993-1-9
42
and in clause 6.8.
Clause 5.1.2(3) prohibits the use of semi-continuous composite joints (dened in clause
1.5.2.8) in bridges. An example of such a prohibited joint might be a composite main
beam joined together through end-plate connections. Semi-continuous non-composite
joints should also be avoided where possible, so that fatigue can be assessed using the
detail categories in EN 1993-1-9.
Semi-continuous joints may, in some situations, be unavoidable, such as end-plate connec-
tions between composite cross-beams and main beam webs in some U-frame bridges, but
these would not be composite joints due to the lack of continuity of the slab reinforcement.
The exibility of such a joint would have to be considered in deriving the restraint provided
to the compression ange by the U-frame. Design rules are given in EN 1993-1-8 and in
EN 1994-1-1.
Another apparent exception to the above rule concerns the slip of bolts. This is discussed
under clause 5.4.1.1(7).
5.1.3. Groundstructure interaction
Clause 5.1.3(1)P refers to deformation of supports, so the stiness of the bearings, piers,
abutments and ground have to be taken into account in analysis. This also includes consid-
eration of stiness in determining eective lengths for buckling or resistance to buckling by
analysis. For further guidance on this, see Section 5.2 below.
The eects of dierential settlement must also be included in analysis, although from
clause 5.1.3(3) they may be neglected in ultimate limit state checks. Similar considerations
apply to other indirect actions, such as dierential temperature and dierential creep.
They are discussed in this Guide under clause 5.4.2.2(6).
5.2. Structural stability
The following comments refer to both entire bridges and isolated members. They assume
that the global analyses will be based on elastic theory. The exception in clause 5.4.3 is
discussed later. All design methods must take account of:
.
errors in the initial positions of joints (global geometric imperfections) and in the initial
geometry of members (member geometric imperfections)
Clause 5.1.1(2)
Clause 5.1.2(1)
Clause 5.1.2(2)
Clause 5.1.2(3)
Clause 5.1.3(1)P
Clause 5.1.3(3)
30
.
the eects of cracking of concrete and of any semi-rigid or nominally pinned joints
.
residual stresses in compression members (structural imperfections).
The stage at which each of these is considered or allowed for can be selected by the
designer, which leads to some complexity in clauses 5.2 to 5.4.
5.2.1. Eects of deformed geometry of the structure
In its clause 1.5.6, EN 1990 denes types of analysis. First-order analysis is performed on
the initial geometry of the structure. Second-order analysis takes account of the deforma-
tions of the structure, which are a function of its loading. Clearly, second-order analysis may
always be applied. With appropriate software increasingly available, second-order analysis is
now relatively straightforward to perform. The criteria for neglect of second-order eects
given in clauses 5.2.1(2)P and 5.2.1(3) need not then be considered. The analysis allowing
for second-order eects will usually be iterative but normally the iteration will take place
within the software. Methods for second-order analysis are described in text books such
as that by Trahair et al.
43
A disadvantage of second-order analysis is that the principle of superposition does not
apply and entire load combinations must be applied to the bridge model. In this case, the
critical load combinations can still rst be estimated using rst-order analysis, inuence
lines (or surfaces) and superposition of load cases.
Second-order eects apply to both in-plane and out-of-plane modes of buckling, including
lateraltorsional buckling. The latter behaviour is more complex and requires a nite-
element analysis using shell elements to model properly second-order eects and instability.
A method of checking beams for out-of-plane instability while modelling only in-plane
second-order eects is given in clause 6.3.4 of EN 1993-1-1 and discussed in section 6.4.3
of this guide. Out-of-plane second-order eects can only be neglected in bridge beams
where there is sucient lateral bracing present. In-plane second-order eects in the beams
will usually be negligible and lateraltorsional buckling may be checked using one of the
simplied methods permitted in clause 6.4. Integral bridges, with high axial load in the
beams caused by earth pressure, may be an exception.
Clause 5.2.1(3) provides a basis for the use of rst-order analysis. The check is done for a
particular load combination and arrangement. The provisions in this clause are similar to
those for elastic analysis in the corresponding clause in EN 1993-2. Clause 5.2.1(3) is not
just for a sway mode. This is because clause 5.2.1 is relevant not only to complete frames
but also to the design of individual columns (see clause 6.7.3.4 for composite columns,
and comments on it). Such members may be held in position against sway but still be
subject to signicant second-order eects due to bowing. Second-order eects in local and
global modes are illustrated in Fig. 5.1.
In an elastic frame, second-order eects are dependent on the proximity of the design loads
to the elastic critical buckling load. This is the basis for expression (5.1), in which c
cr
is
dened as the factor . . . to cause elastic instability. This may be taken as the load factor at
which bifurcation of equilibrium occurs. For a column or frame, it is assumed that there
are no member imperfections, and that only vertical loads are present, usually at their
maximum design values. These are replaced by a set of loads that produce the same set of
member axial forces without any bending. An eigenvalue analysis then gives the factor
Clause 5.2.1(2)P
Clause 5.2.1(3)
A
B C
D
E F
(a) (b)
Fig. 5.1. Examples of local and global instability: (a) local second-order eects; (b) global second-order
eects
31
c
cr
, applied to the whole of the loading, at which the system stiness vanishes and elastic
instability occurs.
To sucient accuracy, c
cr
may also be determined by a second-order loaddeection
analysis. The non-linear loaddeection response approaches asymptotically to the elastic
critical value. This may be useful as some software will perform this analysis but not an
elastic critical buckling analysis.
The use of expression (5.1) is one way of determining if rst-order analysis will suce.
Clause 5.2.1(3) also states that second-order eects may be ignored where the increases in
internal actions due to the deformations from rst-order analysis are less than 10%.
Hence, for members braced against lateral buckling:
M
I
,M
I
! 10 D5.1
where M
I
is the moment from rst-order analysis, including the eects of initial imperfec-
tions, and M
I
is the increase in bending moments calculated from the deections obtained
from rst-order analysis (the P.moments). By convention, the symbols . or c are used for
deformations. They should not be confused with , as used here in M
I
.
Application of this criterion, in principle, avoids the need for elastic critical buckling
analysis but its use has some problems as discussed below. For the case of a pin-ended
strut with sinusoidal bow of magnitude a
0
, expression (D5.1) is the same as expression
(5.1). This can be shown as follows.
The extra deection from a rst-order analysis can easily be shown to be given by:
a a
0
F
Ed
,F
cr
D5.2
where F
Ed
is the applied axial load and F
cr
is the elastic critical buckling load. It follows that
the extra moment from the rst-order deection is:
M
I
F
Ed
a
0
F
Ed
,F
cr
D5.3
Putting equation (D5.3) into equation (D5.1) gives expression (5.1):
M
I
,M
I

F
Ed
a
0
F
Ed
a
0
F
Ed
,F
cr
F
cr
F
Ed
c
cr
! 10
This direct equivalence is only valid for a pin-ended strut with a sinusoidal bow and hence
sinusoidal curvature but it generally remains suciently accurate. (Note: It is found for a
strut with equal end moments that:
M
I
,M
I

8
2
_
F
cr
F
Ed
_
For anything other than a pin-ended strut or statically determinate structure, it will not be
easy to determine M
I
from the deections found by rst-order analysis. This is because in
indeterminate structures, the extra moment cannot be calculated at all sections directly from
the local P. because of the need to maintain compatibility.
In the example shown in Fig. 5.2, it would be conservative to assume that at mid-height,
M
I
N.. (This is similar to secondary eects of prestressing in prestressed structures.) A
more accurate value could be found from a further rst-order analysis that models the rst-
order deected shape found by the previous analysis. To avoid the problem that low ratios
M
I
,M
I
can be obtained near points of contraexure, the condition M
I
,M
I
! 10 should
be applied only at the peak moment positions between each adjacent point of contraexure.
The maximum P. bending moment in the member can again be used as a conservative
estimate of M
I
.
Clause 5.2.1(4)P is a reminder that the analysis shall account for the reductions in stiness
arising from cracking and creep of concrete and from possible non-linear behaviour of
the joints. In general, such eects are dependent on the internal moments and forces, so
calculation is iterative. Simplied methods are therefore given in clauses 5.4.2.2 and
5.4.2.3, where further comment is given.
Manual intervention may be needed, to adjust stiness values before repeating an analysis.
It is expected however that advanced software will be written for EN 1994 to account
Clause 5.2.1(4)P
32
automatically for these eects. The designer may of course make assumptions, although care
is needed to ensure these are conservative. For example, assuming that joints have zero
rotational stiness (resulting in simply-supported composite beams) could lead to neglect
of the reduction in beam stiness due to cracking. The overall lateral stiness would
probably be a conservative value, but this is not certain.
Clause 6.7.3.4(2) gives an eective exural stiness for doubly symmetric columns which
may be used to determine c
cr
(clause 6.7.3.4(3)) and which makes allowance for the stiness
of the concrete, including the eects of cracking, and the reinforcement. The use of this
stiness in checking composite columns is discussed in section 6.7.3 of this guide.
For asymmetric composite compression members in general, such as a composite bridge
deck beam in an integral bridge, the eective stiness usually depends on the direction of
bowing of the member. This is inuenced by the initial camber and by the deection
under the loading considered. The deection under design ultimate load and after creep
usually exceeds the initial camber. The direction of bow is then downwards.
A conservative possibility for determining c
cr
is to ignore completely the contribution of
the concrete to the exural stiness, including reinforcement only; this is done in Example 6.6
to determine the elastic critical buckling load under axial force. An even more conservative
possibility is to base the exural stiness on the steel section alone. If second-order analysis is
necessary, this simplication will not usually be satisfactory as the use of cracked properties
throughout the structure, irrespective of the sign of the axial force in the concrete, would not
satisfy the requirements of clause 5.4.2.3 regarding cracking. Generally the results of a rst-
order analysis can be used to determine which areas of the structure are cracked and the
section properties for second-order analysis can then be modied as necessary. The stiness
of cracked areas can be based on the above simplication. The procedure can be iterative if
the extent of cracked zones is signicantly altered by the second-order analysis. An eective
modulus of elasticity for compressed concrete is also required to calculate the exural
stiness of uncracked areas. Clause 6.7.3.3(4) provides a formula.
5.2.2. Methods of analysis for bridges
Where it is necessary to take second-order eects and imperfections into account, EN 1993-2
clause 5.2.2 provides the following three alternative methods by reference to EN 1993-1-1
clause 5.2.2(3).
.
Use of second-order analysis including both global system imperfections and local
member imperfections as discussed in section 5.3 below. If this method is followed, no
individual checks of member stability are required and members are checked for cross-
section resistance only. An alternative method for bare steel members is discussed under
clause 5.3.1(2). For each composite member, it is necessary to use an appropriate exural
stiness covering the eects of cracking and creep as discussed in section 5.2.1 above.
(a) (b)
<N
(c)
N
N
N
N
Fig. 5.2. Extra bending moments from deection: (a) rst-order moment due to imperfections; (b) rst-
order deection; (c) additional moment from deection
33
If lateraltorsional buckling is to be covered totally by second-order analysis,
appropriate nite-element analysis capable of modelling the behaviour will be required.
.
Use of second-order analysis including global system imperfections only. For individual
bare steel members, stability checks are then required according to EN 1993-2 clause 6.3.
Since the member end forces and moments include second-order eects from global
behaviour, the eective length of individual members is then based on the member
length, rather than a greater eective length that includes the eects of global sway
deformations. Note that when clause 6.3.3 of EN 1993-1-1 is used for member checks
of bare steel members, the member moments will be further amplied by the k
ij
par-
ameters. Since the second-order analysis will already have amplied these moments,
providing sucient nodes have been included along the member in the analysis model,
this is conservative and it would be permissible to set the k
ij
parameters equal to
unity where they exceed unity. However, the imperfections within the members have
not been considered or amplied by the second-order analysis. These are included via
the rst term in the equations in this clause:
N
Ed
N
Rk
,
M1
For composite members in compression and bending, buckling resistance curves cannot
be used, and the moments from member imperfections in the member length should be
added. Second-order eects within the member are accounted for by magnifying the
resulting moments from the local imperfections within the length of the member according
to clauses 6.7.3.4(4) and 6.7.3.4(5) using an eective length based on the member length,
and then checking the resistance of cross-sections. Only the local member imperfections
need to be amplied if sucient nodes have been included along the member in the analysis
model, as all other moments will then have been amplied by the second-order global
analysis. Further comment and a ow chart are given under clause 6.7.3.4.
.
Use of rst-order analysis without modelled imperfections. For bare steel members, the
verication can be made using clause 6.3 of EN 1993-2 with appropriate eective lengths.
All second-order eects are then included in the relevant resistance formulae. This latter
method will be most familiar to bridge engineers in the UK, as tables of eective lengths
for members with varying end conditions of rotational and positional xity have
commonly been used. The use of eective lengths for this method is discussed in the
Designers Guide to EN 1993-2.
4
For composite compression members, this approach is generally not appropriate. The
method of clause 6.7.3 is based on calculation of second-order eects within members,
followed by checks on resistance of cross-sections. No buckling resistance curves are
provided. Composite beams in bending alone can however be checked for lateral
torsional buckling satisfactorily following this method.
Second-order analysis itself can be done either by direct computer analysis that accounts
for the deformed geometry or by amplication of the moments from a rst-order analysis
(including the eects of imperfections) using clause 5.2.2(5) of EN 1993-2. Where either
approach is used, it should only be performed by experienced engineers because the guidance
on the use of imperfections in terms of shapes, combinations and directions of application is
not comprehensive in EC3 and EC4 and judgement must be exercised.
5.3. Imperfections
5.3.1. Basis
Clause 5.3.1(1)P lists possible sources of imperfection. Subsequent clauses (and also
clause 5.2) describe how these should be allowed for. This may be by inclusion in the
global analyses or in methods of checking resistance, as explained above.
Imperfections comprise geometric imperfections and residual stresses. The term geometric
imperfection is used to describe departures from the intended centreline setting-out
Clause 5.3.1(1)P
34
dimensions found on drawings, which occur during fabrication and erection. This is
inevitable as construction work can only be executed to certain tolerances. Geometric
imperfections include lack of verticality, lack of straightness, lack of t and minor joint
eccentricities. The behaviour of members under load is also aected by residual stresses
within the members. Residual stresses can lead to yielding of steel occurring locally at
lower applied external load than predicted from stress analysis ignoring such eects. The
eects of residual stresses can be modelled by additional geometric imperfections. The
equivalent geometric imperfections given in EC3 and EC4 cover both geometric imperfec-
tions and residual stresses.
Clause 5.3.1(2) requires imperfections to be in the most unfavourable direction and form.
The most unfavourable geometric imperfection normally has the same shape as the lowest
buckling mode. This can sometimes be dicult to nd, but it can be assumed that this condi-
tion is satised by the Eurocode methods for checking resistance that include eects of
member imperfections (see comments on clause 5.2.2). Clause 5.3.2(11) of EN 1993-1-1
covers the use of a unique global and local system imperfection based on the lowest buckling
mode. This can generally only be used for bare steel members as the imperfection parameter
c is required and this is not provided for composite members. The method is discussed in the
4
5.3.2. Imperfections for bridges
Generally, an explicit treatment of geometric imperfections is required for composite frames.
In both EN 1993-1-1 and EN 1994-1-1 the values are equivalent rather than measured values
(clause 5.3.2(1)) because they allow for eects such as residual stresses, in addition to
imperfections of shape.
Clause 5.3.2(2) covers bracing design. In composite bridges, the deck slab acts as plan
bracing. Compression anges that require bracing occur in hogging regions of beam-and-
slab bridges and in sagging regions of half-through bridges, bowstring arches and similar
structures. The bracing of compression anges in sagging regions diers little from that in
all-steel bridges, and is discussed in the Designers Guide to EN 1993-2.
4
Steel bottom anges in hogging regions of composite bridges are usually restrained
laterally by continuous or discrete transverse frames. For deep main beams, plan bracing
at bottom-ange level may also be used. Where the main beams are rolled I-sections, their
webs may be sti enough to serve as the vertical members of continuous inverted-U
frames, which are completed by the shear connection and the deck slab. These systems are
discussed under clause 6.4.2.
Discrete U-frame bracing can be provided at the location of vertical web stieners. These
frames need transverse steel members. If these are provided just below the concrete deck,
they should be designed as composite. Otherwise, design for shrinkage and temperature
eects in the transverse direction becomes dicult. This problem is often avoided by
placing the steel cross-member at lower level, so creating an H-frame. Both types of frame
provide elastic lateral restraint at bottom-ange level, with a spring stiness that is easily
calculated.
The design transverse forces for these frames, or for plan bracing, arise from lateral
imperfections in the compressed anges. For these imperfections, clause 5.3.2(2) refers to
EN 1993-2, which in turn refers to clauses 5.3.2 to 5.3.4 of EN 1993-1-1. The design trans-
verse forces, F
Ed
, and a design method are given in clause 6.3.4.2 of EN 1993-2, though it
refers specically only to U-frame restraints. Comments on these clauses are in the relevant
Guides in this series.
4.7
The relevant imperfections for analysis of the bracing system are not necessarily the same
as those for the bridge beams themselves.
In hogging regions of continuous beam-and-slab bridges, distortional lateral buckling is
usually the critical mode. It should not be assumed that a point of contraexure is a
lateral restraint, for the buckling half-wavelength can exceed the length of ange in
compression.
44
Clause 5.3.1(2)
Clause 5.3.2(1)
Clause 5.3.2(2)
35
Where the restraint forces are to be transmitted to end supports by a system of plan
bracing, this system should be designed to resist the more onerous of the transverse forces
F
Ed
from each restraint within a length equal to the half wavelength of buckling, and the
forces generated by an overall ange bow in each ange according to clause 5.3.3 of
EN 1993-1-1.
For the latter case, the overall bow is given as e
0
c
m
L,500, where c
m
is the reduction
factor for the number of interconnected beams (c
m
0.866 for two beams), and L is the
span. The plan bracing may be designed for an equivalent uniformly-distributed force per
beam of 8N
Ed
e
0
c
q
,L, where c
q
is the deection of the bracing, and N
Ed
is the
maximum compressive force in the ange.
For very sti bracing, the total design lateral force for the bracing is:
8
N
Ed
,L
_ _
c
m
L,500
N
Ed
c
m
,62.5
Clause 5.3.2(2) should also be used for system imperfections for composite columns,
although its scope is given as stabilizing transverse frames. Its reference to clause 5.3 of
EN 1993-2 leads to relevant clauses in EN 1993-1-1, as follows.
Initial out-of-plumb of a column is given in clause 5.3.2(3) of EN 1993-1-1 which, although
worded for frames, is applicable to a single column or row of columns. Where a steel
column is very slender and has a moment-resisting joint at one or both ends, clause
5.3.2(6) of EN 1993-1-1 requires its local bow imperfection to be included in the second-
order global analysis used to determine the action eects at its ends. Very slender is
dened as:
"
`` 0.5
Af
y
,N
Ed
_
It is advised that this rule should be used also for composite columns, in the form c
cr
< 4,
with c
cr
as dened in clause 5.2.1(3). This is obtained by replacing Af
y
by N
pl
.
Clause 5.3.2(3) covers imperfections in composite columns and compression members
(e.g. in trusses), which must be considered explicitly. It refers to material in clause 6.7.3,
which appears to be limited, by clause 6.7.3(1), to uniform members of doubly symmetrical
cross-section. Clause 6.7.2(9), which is of general applicability, also refers to Table 6.5 of
clause 6.7.3 for member imperfections; but the table only covers typical cross-sections
of columns. Imperfections in compressed beams, which occur in integral bridges, appear
to be outside the scope of EN 1994.
The imperfections for buckling curve d in Table 5.1 of EN 1993-1-1 could conservatively
be used for second-order eects in the plane of bending. For composite bridges with the deck
slab on top of the main beams, lateral buckling eects can subsequently be included by a
check of the compression ange using the member resistance formulae in clause 6.3 of
EN 1993-1-1. Guidance on verifying beams in integral bridges in bending and axial load is
discussed in section 6.4 of this guide.
Clause 5.3.2(4) covers global and local imperfections in steel compression members, by
reference to EN 1993-2. Imperfections for arches are covered in Annex D of EN 1993-2.
5.4. Calculation of action eects
5.4.1. Methods of global analysis
EN 1990 denes several types of analysis that may be appropriate for ultimate limit states.
For global analysis of bridges, EN 1994-2 gives three methods: linear elastic analysis, with or
without corrections for cracking of concrete, and non-linear analysis. The latter is discussed
in section 5.4.3 below, and is rarely used in practice.
Clause 5.4.1.1(1) permits the use of elastic global analysis even where plastic (rectangular-
stress-block) theory is used for checking resistances of cross-sections. For resistance to
exure, these sections are in Class 1 or 2, and commonly occur in mid-span regions.
There are several reasons
4547
why the apparent incompatibility between the methods used
for analysis and for resistance is accepted. It is essentially consistent with UK practice, but
Clause 5.3.2(3)
Clause 5.3.2(4)
Clause 5.4.1.1(1)
36
care should be taken with mixing section classes within a bridge when elastic analysis is used.
An example is a continuous bridge, with a mid-span section designed in bending as Class 2
and the section at an internal support as Class 3. The Class 3 section may become over-
stressed by the elastic moments shed from mid-span while the plastic section resistance
develops there and stiness is lost.
There is no such incompatibility for Class 3 or 4 sections, as resistance is based on elastic
models.
Mixed-class design has rarely been found to be a problem, as the load cases producing
maximum moment at mid-span and at a support rarely coexist, except where adjacent
spans are very short compared to the span considered. A relevant design rule is given in
clause 6.2.1.3(2).
If redistribution is required to be checked, the conservative method illustrated in Fig. 5.3
may be used. In this example there is a Class 2 section at mid-span of the central span, and
the support sections are Class 3. A simplied load case that produces maximum sagging
moment is shown. Elastic analysis for the load P gives a bending moment at cross-section
C that exceeds the elastic resistance moment, M
el.C
. The excess moment is redistributed
from section C, giving the distribution shown by the dashed line. In reality, the moment
at C continues to increase, at a reduced rate, after the elastic value M
el.C
is reached, so the
true distribution lies between those shown in Fig. 5.3. The upper distribution therefore
provides a safe estimate of the moments at supports B and D, and can be used to check
that the elastic resistance moment is not exceeded at these points.
Elastic global analysis is required for serviceability limit states (clause 5.4.1.1(2)) to enable
yielding of steel to be avoided. Linear elastic analysis is based on linear stressstrain laws, so
for composite structures, appropriate corrections for . . . cracking of concrete are required.
These are given in clause 5.4.2.3, and apply also for ultimate limit states.
Clause 5.4.1.1(3) requires elastic analysis for fatigue, to enable realistic ranges of fatigue
stress to be predicted.
The eects of shear lag, local buckling of steel elements and slip of bolts must also be
considered where they signicantly inuence the global analysis. Shear lag and local buckling
eects can reduce member stiness, while slip in bolt holes causes a localized loss of stiness.
Shear lag is discussed under clause 5.4.1.2, and plate buckling and bolt slip are discussed
below.
Methods for satisfying the principle of clause 5.4.1.1(4)P are given for local buckling in
clauses 5.4.1.1(5) and (6). These refer to the classication of cross-sections, the established
method of allowing for local buckling of steel anges and webs in compression. It determines
the available methods of global analysis and the basis for resistance to bending. The
classication system is dened in clause 5.5.
Plate buckling
In Class 4 sections (those in which local buckling will occur before the attainment of yield),
plate buckling can lead to a reduction of stiness. The in-plane stiness of perfectly at plates
suddenly reduces when the elastic critical buckling load is reached. In real plates that have
imperfections, there is an immediate reduction in stiness from that expected from the gross
Clause 5.4.1.1(2)
Clause 5.4.1.1(3)
Clause 5.4.1.1(4)P
Clause 5.4.1.1(5)
Clause 5.4.1.1(6)
B
D
C
P
Moments with bending
moment at C reduced to M
el,C
Moments from elastic analysis
M
el,C
M
pl,C
Fig. 5.3. Eect of mixing section classes, and an approximate method for checking bending moments at
internal supports
37
plate area because of the growth of the bow imperfections under load. This stiness
continues to reduce with increasing load. This arises because non-uniform stress develops
across the width of the plate as shown in Fig. 5.4. The non-uniform stress arises because
the development of the buckle along the centre of the plate leads to a greater developed
length of the plate along its centreline than along its edges. Thus the shortening due to
membrane stress, and hence the membrane stress itself, is less along the centreline of the
plate.
This loss of stiness must be considered in the global analysis where signicant. It can
be represented by an eective area or width of plate, determined from clause 2.2 of
EN 1993-1-5. This area or width is greater than that used for resistance, which is given in
clause 4.3 of EN 1993-1-5.
The loss of stiness may be ignored when the ratio of eective area to gross cross-sectional
area exceeds a certain value. This ratio may be given in the National Annex. The recommended
value, given in a Note to clause 2.2(5) of EN 1993-1-5, is 0.5. This should ensure that plate
buckling eects rarely need to be considered in the global analysis. It is only likely to be of
relevance for the determination of pre-camber of box girders under self-weight and wet
concrete loads. After the deck slab has been cast, buckling of the steel ange plate will be
prevented by its connection to the concrete ange via the shear connection.
Eects of slip at bolt holes and shear connectors
Clause 5.4.1.1(7) requires consideration of slip in bolt holes and similar deformations of
connecting devices. This applies to both rst- and second-order analyses. There is a
similar rule in clause 5.2.1(6) of EN 1993-1-1. No specic guidance is given in EN 1993-2
or EN 1994-2. Generally, bolt slip will have little eect in global analysis. It has often
been practice in the UK to design bolts in main beam splices to slip at ultimate limit
states (Category B to clause 3.4.1 of EN 1993-1-8). Although slip could alter the moment
distribution in the beam, this is justiable. Splices are usually near to the point of contra-
exure, so that slip will not signicantly alter the distribution of bending moment. Also,
the loading that gives maximum moment at the splice will not be fully coexistent with that
for either the maximum hogging moment or maximum sagging moment in adjacent regions.
It is advised that bolt slip should be taken into account for bracing members in the analysis
of braced systems. This is because a sudden loss of stiness arising from bolt slip gives an
increase in deection of the main member and an increased force on the bracing member,
which could lead to overall failure. Ideally, therefore, bolts in bracing members should be
designed as non-slip at ultimate limit state (Category C to EN 1993-1-8).
The similar deformations quoted above could refer to slip at an interface between steel
and concrete, caused by the exibility of shear connectors. The provisions on shear connec-
tion in EN 1994 are intended to ensure that slip is too small to aect the results of elastic
global analysis or the resistance of cross-sections. Clause 5.4.1.1(8) therefore permits
Clause 5.4.1.1(7)
Clause 5.4.1.1(8)
Fig. 5.4. Stress distribution across width of slender plate
38
internal moments and forces to be determined assuming full interaction where shear con-
nection is provided in accordance with EN 1994.
Slip of shear connectors can also aect the exural stiness of a composite joint. A
relevant design method is given in clause A.3 of EN 1994-1-1. It is mainly applicable to
semi-continuous joints, and so is not included in EN 1994-2.
An exception to the rules on allowing for cracking of concrete is given in clause 5.4.1.1(9),
for the analysis of transient situations during erection stages. This permits uncracked global
analysis to be used, for simplicity.
Eective width of anges for shear lag
Shear lag is dened in clause 5.4.1.2(1) with reference to the exibility of anges due to
in-plane shear. Shear lag in wide anges causes the longitudinal bending stress adjacent
to the web to exceed that expected from analysis with gross cross-sections, while the stress
in the ange remote from the web is much lower than expected. This shear lag also leads
to an apparent loss of stiness of a section in bending which can be important in determining
realistic distributions of moments in analysis. The determination of the actual distribution of
stress is a complex problem.
The Eurocodes account for both the loss of stiness and localized increase in ange
stresses by the use of an eective width of ange which is less than the actual available
ange width. The eective ange width concept is articial but, when used with engineering
bending theory, leads to uniform stresses across the whole reduced ange width that are
equivalent to the peak values adjacent to the webs in the true situation.
The rules that follow provide eective widths for resistance of cross-sections, and simpler
values for use in global analyses. The rules use the word may because clause 5.4.1.2(1)
permits rigorous analysis as an alternative. This is not dened, but should take account
of the many relevant inuences, such as the cracking of concrete.
Steel anges
For steel plate elements clause 5.4.1.2(2) refers to EN 1993-1-1. This permits shear lag to be
neglected in rolled sections and welded sections with similar dimensions, and refers to
EN 1993-1-5 for more slender anges. In these, the stress distribution depends on the stif-
fening to the anges and any plasticity occurring for ultimate limit state behaviour. The
elastic stress distribution can be modelled using nite-element analysis with appropriate
shell elements.
The rules in EN 1993-1-5 are not discussed further in this guide but are covered in the
4
Dierent values of eective width apply for cross-section
design for serviceability and ultimate limit states, and the value appropriate to the location
of the section along the beam should be used. Simplied eective widths, taken as constant
throughout a span, are allowed in the global analysis.
Concrete anges
Eective width of concrete anges is covered in clauses 5.4.1.2(3) to (7). The behaviour is
complex, being inuenced by the loading conguration, and by the extent of cracking and of
yielding of the longitudinal reinforcement, both of which help to redistribute the stress across
the cross-section. The ability of the transverse reinforcement to distribute the forces is also
relevant. The ultimate behaviour in shear of wide anges is modelled by a truss analogy
similar to that for the web of a deep concrete beam.
The values for eective width given in this clause are simpler than those in BS 5400:Part 5,
and similar to those in BS 5950:Part 3.1:1990.
48
The eective width at mid-span and internal
supports is given by equation (5.3):
b
eff
b
0
b
ei
5.3
where b
0
is the distance between the outer shear connectors and b
ei
is either b
e1
or b
e2
, as
shown in Fig. 5.5, or the available width b
1
or b
2
, if lower.
Clause 5.4.1.1(9)
Clause 5.4.1.2(1)
Clause 5.4.1.2(2)
Clauses 5.4.1.2(3)
to (7)
39
Each width b
ei
is limited to L
e
,8, where L
e
is the assumed distance between points of zero
bending moment. It depends on the region of the beam considered and on whether the
bending moment is hogging or sagging. This is shown in Fig. 5.5, which is based on Fig. 5.1.
The values are generally lower than those in EN 1992-1-1 for reinforced concrete T-beams.
To adopt those would often increase the number of shear connectors. Without evidence that
the greater eective widths are any more accurate, the established values for composite
beams have mainly been retained.
In EN 1992-1-1, the sum of the lengths L
e
for sagging and hogging regions equals the span
of the beam. In reality, points of contraexure are dependent on the load arrangement.
EN 1994, like EN 1993, therefore gives a larger eective width at an internal support. In
sagging regions, the assumed distances between points of contraexure are the same in all
three codes.
Although there are signicant dierences between eective widths for supports and mid-
span regions, it is possible to ignore this in elastic global analysis (clause 5.4.1.2(4)). This is
because shear lag has limited inuence on the results. There can however be some small
advantage to be gained by modelling in analysis the distribution of eective width along
the members given in Fig. 5.5 or Fig. 5.1, as this will tend to shed some moment from the
hogging regions into the span. It would also be appropriate to model the distribution of
eective widths more accurately in cable-stayed structures, but Fig. 5.1 does not cover
these. Example 5.1 below illustrates the calculation of eective width.
Some limitations on span length ratios when using Fig. 5.1 should be made so that the
bending-moment distribution within a span conforms with the assumptions in the gure.
It is suggested that the limitations given in EN 1992 and EN 1993 are adopted. These
limit the use to cases where adjacent spans do not dier by more than 50% of the shorter
span and a cantilever is not longer than half the adjacent span. For other span ratios or
moment distributions, the distance between points of zero bending moment, L
e
, should be
calculated from the moment distribution found from an initial analysis.
Where it is necessary to determine a more realistic distribution of longitudinal stress across
the width of the ange, clause 5.4.1.2(8) refers to clause 3.2.2 of EN 1993-1-5. This might be
necessary, for example, in checking a deck slab at a transverse diaphragm between main
Clause 5.4.1.2(8)
L
1
L
e
= 0.85L
1
for b
eff,1
L
e
= 0.70L
2
for b
eff,1
L
e
= 0.25(L
1
+ L
2
)
for b
eff,2
L
e
= 2L
3
for b
eff,2
L
1
/4 L
1
/4 L
2
/4 L
2
/2 L
2
/4 L
1
/2
L
2
L
3
Centreline
b
eff,0
b
eff,1
b
e1
b
e2
b
0
b
1
b
2
b
eff,1 b
eff,2
b
eff,2
Fig. 5.5. Symbols and equivalent spans, for eective width of concrete ange (Source: based on Fig. 5.1 of
EN 1994-2)
40
beams at a support, where the deck slab is in tension under global bending and also subjected
to a local hogging moment from wheel loads. The use of EN 1993-1-5 can be benecial here,
as often the greatest local eects in a slab occur in the middle of the slab between webs where
the global longitudinal stresses are lowest.
Composite plate anges
Clause 5.4.1.2(8) recommends the use of its stress distribution for both concrete and steel
anges. Where the ange is a composite plate, shear connection is usually concentrated
near the webs, so this stress distribution is applicable. Eective widths of composite
plates in bridges are based on clause 5.4.1.2, but with a dierent denition of b
0
, given in
clause 9.1(3).
Composite trusses
Clause 5.4.1.2(9) applies where a longitudinal composite beam is also a component of a
larger structural system, such as a composite truss. For loading applied to it, the beam is
continuous over spans equal to the spacing of the nodes of the truss. For the axial force
in the beam, the relevant span is that of the truss.
Clause 5.4.1.2(9)
Example 5.1: eective widths of concrete ange for shear lag
A composite bridge has the span layout and cross-section shown in Fig. 5.6. The eective
width of top ange for external and internal beams is determined for the mid-span regions
BC and DE, and the region CD above an internal support.
L
1
= 19.000 m L
3
= 19.000 m L
2
= 31.000 m
b
1
b
2
b
e1
b
e2
C/L
250
C/L
3100
3100 2000 3100 3100 2000
250
A B C D E
Fig. 5.6. Elevation and typical cross-section of bridge for Example 5.1
The eective spans L
e
, from Fig. 5.5, and the lengths L
e
,8 are shown in Table 5.1. For
an external beam, the available widths on each side of the shear connection are:
b
1
1.875 m. b
2
1.425 m.
Table 5.1. Eective width of concrete ange of composite T-beam
External beam Internal beam
Region BC CD DE BC CD DE
L
e
(m) 16.15 12.50 21.70 16.15 12.50 21.70
L
e
,8 (m) 2.019 1.563 2.712 2.019 1.563 2.712
b
eff
(m) 3.550 3.238 3.550 3.100 3.100 3.100
41
5.4.2. Linear elastic analysis
Cracking, creep, shrinkage, sequence of construction, and prestressing, listed in clause
5.4.2.1(1), can all aect the distribution of action aects in continuous beams and frames.
This is always important for serviceability limit states, but can in some situations be
ignored at ultimate limit states, as discussed under clause 5.4.2.2(6). Cracking of concrete
is covered in clause 5.4.2.3.
Creep and shrinkage of concrete
The rules provided in clause 5.4.2.2 allow creep to be taken into account using a modular
ratio n
L
, that depends on the type of loading, and on the concrete composition and age at
loading. This modular ratio is used both for global analysis and for elastic section analysis.
It is dened in clause 5.4.2.2(2) by:
n
L
n
0
1
L
ct. t
0
5.6
where n
0
is the modular ratio for short-term loading, E
a
,E
cm
. The concrete modulus E
cm
is
obtained from EN 1992 as discussed in section 3.1 of this guide. The creep coecient ct. t
0
is also obtained from EN 1992.

The creep multiplier
L
takes account of the type of loading. Its values are given in clause
5.4.2.2(2) as follows:
.
for permanent load,
L
1.1
.
for the primary and secondary eects of shrinkage (and also the secondary eects of
creep, clause 5.4.2.2(6)),
L
0.55
.
for imposed deformations,
L
1.5.
The reason for the factor
L
is illustrated in Fig. 5.7. This shows three schematic curves of
the change of compressive stress in concrete with time. The top one, labelled S, is typical of
stress caused by the increase of shrinkage with time. Concrete is more susceptible to creep
when young, so there is less creep (
L
0.55) than for the more uniform stress caused by
permanent loads (line P). The eects of imposed deformations can be signicantly reduced
by creep when the concrete is young, so the curve is of type ID, with
L
1.5. The value
for permanent loading on reinforced concrete is 1.0. It is increased to 1.1 for composite
Clause 5.4.2.1(1)
Clause 5.4.2.2
Clause 5.4.2.2(2)
The eective widths are the lower of these values and L
e
,8, plus the width of the shear
connection, as follows:
for BC and DE. b
eff
1.875 1.425 0.25 3.550 m
for CD. b
eff
1.563 1.425 0.25 3.238 m
For an internal beam, the available widths on each side of the shear connection are:
b
1
b
2
1.425 m
These available widths are both less than L
e
,8, and so govern. The eective widths are:
for BC, CD, and DE, b
eff
2 1.425 0.25 3.100 m
Time
S
P
ID
0
1.0
c
/
c,0
Fig. 5.7. Time-dependent compressive stress in concrete, for three types of loading
42
members because the steel component does not creep. Stress in concrete is reduced by creep
less than it would be in a reinforced member, so there is more creep.
These values are based mainly on extensive theoretical work on composite beams of many
sizes and proportions.
49
The factor
L
performs a similar function to the ageing coecient found in Annex KK of
EN 1992-2 and in the calculation for loss of prestress in clause 5.10.6 of EN 1992-1-1.
The creep factor ct. t
0
depends on the age of the concrete, t, at which the modular ratio is
being calculated (usually taken as innity) and the age of the concrete at rst loading, t
0
. For
age t
0,
clauses 5.4.2.2(3) and (4) make recommendations for permanent load and shrinkage,
respectively. Since most bridges will follow a concrete pour sequence rather than have all
the concrete placed in one go, this age at rst loading could vary throughout the bridge.
Clause 5.4.2.2(3) permits an assumed mean value of t
0
to be used throughout. This simpli-
cation is almost a necessity as it is rare for the designer to have sucient knowledge of the
construction phasing at the design stage to be more accurate than this, but some estimate of
the expected timings is still required.
First loading could occur at an age as low as a week, for example, from erection of
precast parapets, but the mean age for a multi-span bridge is unlikely to be less than a month.
The creep coecient depends also on the eective thickness of the concrete element consid-
ered, h
0
. There is no moisture loss through sealed surfaces, so these are assumed to be at mid-
thickness of the member. After striking of formwork, a deck slab of thickness, say, 250 mm,
has two free surfaces, and an eective thickness of 250 mm. The application of waterproong
to the top surface increases this thickness to 500 mm, which reduces subsequent creep. The
designer will not know the age(s) of the deck when waterproofed, and so must make assump-
tions on the safe side.
Fortunately, the modular ratio is not sensitive to either the age of loading or the eective
thickness. As resistances are checked for the structure at an early age, it is on the safe side for
the long-term checks to overestimate creep.
As an example, let us suppose that the short-term modular ratio is n
0
6.36 (as found in a
subsequent example), and that a concrete deck slab has a mean thickness of 250 mm, with
waterproong on one surface. The long-term modular ratio is calculated for t
0
7 days
and 28 days, and for h
0
250 mm and 500 mm. For outdoor conditions with relative
humidity 70%, the values of c1. t
0
given by Annex B of EN 1992-1-1 with
L
1.1 are
as shown in Table 5.2.
The resulting range of values of the modular ratio n
L
is from 18.8 to 23.7. A dierence of
this size has little eect on the results of a global analysis of continuous beams with all spans
composite, and far less than the eect of the dierence between n 6.4 for imposed load and
around 20 for permanent load.
For stresses at cross-sections of slab-on-top decks, the modular ratio has no inuence in
regions where the slab is in tension. In mid-span regions, compression in concrete is rarely
critical, and maximum values occur at a low age, where creep is irrelevant. In steel,
bottom-ange tension is the important outcome, and is increased by creep. From Table
5.2, h
0
has little eect, and the choice of the low value of 7 days for age at rst loading is
on the safe side.
Modular ratios are calculated in Example 5.2 below.
Shrinkage modied by creep
For shrinkage, the advice in clause 5.4.2.2(4) to assume t
0
1 day rarely leads to a modular
ratio higher than that for permanent actions, because of the factor
L
0.55. Both the
Clause 5.4.2.2(3)
Clause 5.4.2.2(4)
Table 5.2. Values of c
0
c1. t
0
and modular ratio n
L
h
0
250 mm h
0
500 mm
t
0
7 days 2.48, 23.7 2.30, 22.4
t
0
28 days 1.90, 19.6 1.78, 18.8
43
long-term shrinkage strain and the creep coecient are inuenced by the assumed eective
thickness h
0
.
For the preceding example, the 1-day rule gives the values in rows 1 and 2 of Table 5.3. It
shows that doubling h
0
has negligible eect on n
L
, but reduces shrinkage strain by 10%.
Increasing the assumed mean relative humidity (RH) by only 5% has the same eect on
shrinkage strain as doubling h
0
, and negligible eect on n
L
. The error in an assumed RH
may well exceed 5%.
For this example, the safe choices for shrinkage eects are h
0
250 mm, and an estimate
for RH on the low side. As the concession in clause 5.4.2.2(3) (the use of a single time t
0
for
all creep coecients) refers to loads, not to actions, it is not clear if shrinkage may be
included. It is conservative to do so, because when t
0
is assumed to exceed 1 day, the relief
of shrinkage eects by creep is reduced. Hence, a single value of n
L
1. t
0
may usually be
used in analyses for permanent actions, except perhaps in special situations, to which
clause 5.4.2.2(5) refers.
Secondary eects of creep
Where creep deections cause a change in the support reactions, this leads to the
development of secondary moments. This might occur, for example, where there are
mixtures of reinforced concrete and steelcomposite spans in a continuous structure. The
redistribution arises because the free creep deections are not proportional everywhere to
the initial elastic deections and therefore the free creep deection would lead to some
non-zero deection at the supports. Other construction sequences could produce a similar
eect but this does not aect normal steelcomposite bridges to any signicant extent.
Clause 5.4.2.2(6) is however a prompt that the eects should be considered in the more
unusual situations.
Calculation of creep redistribution is more complex than for purely concrete structures,
and is explained, with an example, in Ref. 50. The redistribution eects develop slowly
with time, so
L
0.55.
Cross-sections in Class 1 or 2
Clause 5.4.2.2(6) is one of several places in EN 1994-2 where, in certain global analyses,
various indirect actions, that impose displacements and/or rotations, are permitted to be
ignored where all cross-sections are either Class 1 or 2. Large plastic strains are possible
for beams where cross-sections are Class 1. Class 2 sections exhibit sucient plastic strain
to attain the plastic section capacity but have limited rotation capacity beyond this point.
This is however normally considered adequate to relieve the eects of imposed deformations
derived from elastic analysis, and EN 1994 therefore permits such relief to be taken. The
corresponding clause 5.4.2(2) in EN 1993-2 only permits the eects of imposed deformations
to be ignored where all sections are Class 1, so there is an inconsistency at present.
In EN 1994, the eects which can be neglected in analyses for ultimate limit states other
than fatigue, provided that all sections are Class 1 or 2, are as follows:
.
dierential settlement: clause 5.1.3(3)
.
secondary creep redistribution of moments: clause 5.4.2.2(6)
.
primary and secondary shrinkage and creep: clause 5.4.2.2(7)
.
eects of staged construction: clause 5.4.2.4(2)
.
dierential temperature: clause 5.4.2.5(2).
Clause 5.4.2.2(5)
Clause 5.4.2.2(6)
Table 5.3. Eects of shrinkage
h
0
(mm) RH (%) 10
6
sh
n
L
250 70 340 18.8
500 70 304 18.0
250 75 305 19.2
44
The further condition that there should not be any reduction of resistance due to lateral
torsional buckling is imposed in all of these clauses, and is discussed under clause 6.4.2(1).
Primary and secondary eects of creep and shrinkage
Clause 5.4.2.2(7) requires appropriate account to be taken of both the primary and
secondary eects of creep and shrinkage of the concrete. The recommended partial
factor for shrinkage eects at ultimate limit states is
SH
1, from clause 2.4.2.1 of
EN 1992-1-1.
In a fully-restrained member with the slab above the steel beam, shrinkage eects can be
split into a hogging bending moment, an axial tensile force, and a set of self-equilibrated
longitudinal stresses, as shown in Example 5.3 below.
Where bearings permit axial shortening, there is no tensile force. In a statically deter-
minate system, the hogging bending moment is released, causing sagging curvature. These,
with the locked-in stresses, are the primary eects. They are reduced almost to zero where
the concrete slab is cracked through its thickness.
In a statically indeterminate system, such as a continuous beam, the primary shrinkage
curvature is incompatible with the levels of the supports. It is counteracted by bending
moments caused by changes in the support reactions, which increase at internal supports
and reduce at end supports. The moments and the associated shear forces are the secondary
eects of shrinkage.
Clause 5.4.2.2(7) permits both types of eect to be neglected in some checks for ultimate
limit states. This is discussed under clause 5.4.2.2(6).
Clause 5.4.2.2(8) allows the option of neglecting primary shrinkage curvature in cracked
regions.
51
It would be reasonable to base this cracked zone on the same 15% of the span
allowed by clause 5.4.2.3(3), where this is applicable. The use of this option reduces the
secondary hogging bending at supports. These moments, being a permanent eect, enter
into all load combinations, and may inuence design of what is often a critical region.
The long-term eects of shrinkage are signicantly reduced by creep, as illustrated in
clause 5.4.2.2(4). Where it is necessary to consider shrinkage eects within the rst year
or so after casting, a value for the relevant free shrinkage strain can be obtained from
clause 3.1.4(6) of EN 1992-1-1.
Primary eects of shrinkage are calculated in Example 5.3 below.
The inuence of shrinkage on serviceability verications is dealt with in Chapter 7.
For creep in columns, clause 5.4.2.2(9) refers to clause 6.7.3.4(2), which in turn refers to
an eective modulus for concrete given in clause 6.7.3.3(4). If separate analyses are to be
made for long-term and short-term eects, clause 6.7.3.3(4) can be used assuming ratios
of permanent to total load of 1.0 and zero, respectively.
Shrinkage eects in columns are unimportant, except in very tall structures.
Clause 5.4.2.2(10) excludes the use of the preceding simplied methods for members
with both anges composite and uncracked. The uncracked condition is omitted from
clause 5.4.2.2(2), which is probably an oversight. This exclusion is not very restrictive, as
new designs of this type are unusual in the UK. It may occur in strengthening schemes
where the resistance of the compression ange is increased by making it composite over a
short length.
Torsional stiness of box girders
For box girders with a composite top ange or with a concrete ange closing the top of an
open U-section, the torsional stiness is usually calculated by reducing the thickness of the
gross concrete ange on the basis of the appropriate long- or short-term modular ratio, and
maintaining the centroid of the transformed ange in the same position as that of the gross
concrete ange. From clause 5.4.2.2(11), the short-term modular ratio should be based on
the ratio of shear moduli, n
0.G
G
a
,G
cm
, where for steel, G
a
81.0 kN/mm
2
from clause
3.2.6 of EN 1993-1-1 and, for concrete:
G
cm
E
cm
,21 i
c
Clause 5.4.2.2(7)
Clause 5.4.2.2(8)
Clause 5.4.2.2(9)
Clause 5.4.2.2(10)
Clause 5.4.2.2(11)
45
Clause 3.1.3(4) of EN 1992-1-1 gives Poissons ratio (i
c
) as 0.2 or zero, depending on whether
the concrete is uncracked or cracked. For this application it is accurate enough to assume
i
c
0.2 everywhere. The method of clause 5.4.2.2(2) should be used for the modular ratio:
n
L.G
n
0.G
1
L
ct. t
0
The calculation of the torsional second moment of area (in steel units) then follows the
usual procedure such that:
I
T

4A
2
0
_
ds
ts
where A
0
is the area enclosed by the torsional perimeter running through the centreline of
the box walls. This is shown in Fig. 5.8. For closed steel boxes, the location of the centroid
of the composite ange can, for simplicity, be located on the basis of rst moment of area.
The integral
_
ds,ts is the summation of the lengths of each part of the perimeter divided
by their respective thicknesses. It is usual to treat the parts of the web projection into the
ange as having the thickness of the steel web.
The torsional stiness is also inuenced by exural cracking, which can cause a signicant
reduction in the in-plane shear stiness of the concrete ange. To allow for this in regions
where the slab is assumed to be cracked, clause 5.4.2.3(6) recommends a 50% reduction
in the eective thickness of the ange.
Eects of cracking of concrete
Clause 5.4.2.3 is applicable to beams, at both serviceability and ultimate limit states. The ow
chart of Fig. 5.15 below illustrates the procedure.
In conventional composite beams with the slab above the steel section, cracking of
concrete reduces the exural stiness in hogging moment regions, but not in sagging
regions. The change in relative stiness needs to be taken into account in elastic global
analysis. This is unlike analysis of reinforced concrete beams, where cracking occurs in
both hogging and sagging bending, and uncracked cross-sections can be assumed
throughout.
A draft of EN 1994-2 permitted allowance for cracking by redistribution of hogging
moments from uncracked analysis by up to 10%. Following detailed examination of its
eects,
52
this provision was deleted.
Clause 5.4.2.3(2) provides a general method. This is followed in clause 5.4.2.3(3) by a
simplied approach of limited application. Both methods refer to the uncracked and
cracked exural stinesses E
a
I
1
and E
a
I
2
, which are dened in clause 1.5.2. The exural
rigidity E
a
I
1
can usually be based on the gross concrete area excluding reinforcement with
acceptable accuracy. In the general method, the rst step is to determine the expected
extent of cracking in beams. The envelope of moments and shears is calculated for charac-
teristic combinations of actions, assuming uncracked sections and including long-term
eects. The section is assumed to crack if the extreme-bre tensile stress in concrete
exceeds twice the mean value of the axial tensile strength given by EN 1992-1-1.
Clause 5.4.2.3
Clause 5.4.2.3(2)
Torsional perimeter
Transformed thickness:
uncracked: h
c
/n
L,G
cracked: 0.5h
c
/n
L,G
h
c
Fig. 5.8. Torsional stiness of composite box girder
46
The reasons for twice in this assumption are as follows:
.
The concrete is likely to be stronger than specied, although this is partly catered for by
the use of a mean rather than characteristic tensile strength.
.
Test results for tensile strength show a wide scatter when plotted against compressive
strength.
.
Reaching f
ctm
at the surface may not cause the slab to crack right through, and even if it
does, the eects of tension stiening are signicant at the stage of initial cracking.
.
Until after yielding of the reinforcement, the stiness of a cracked region is greater than
E
a
I
2
, because of tension stiening between the cracks.
.
The calculation uses an envelope of moments, for which regions of slab in tension are
more extensive than they are for any particular loading.
The global model is then modied to reduce the beam stiness to the cracked exural
rigidity, E
a
I
2
, over this region, and the structure is reanalysed.
Clause 5.4.2.3(3) provides a non-iterative method, but one that is applicable only to some
situations. These include conventional continuous composite beams, and beams in braced
frames. The cracked regions could dier signicantly from the assumed values in a bridge
with highly unequal span lengths. Where the conditions are not satised, the general
method of clause 5.4.2.3(2) should be used.
The inuence of cracking on the analysis of braced and unbraced frames is discussed in the
Designers Guide to EN 1994-1-1.
5
For composite columns, clause 5.4.2.3(4) makes reference to clause 6.7.3.4 for the
calculation of cracked stiness. The scope of the latter clause is limited (to double symmetry,
etc.) by clause 6.7.3.1(1), where further comment is given. The reduced value of EI referred
to here is intended for verications for ultimate limit states, and may be inappropriate for
analyses for serviceability.
For column cross-sections without double symmetry, cracking and tension in columns are
referred to in clause 6.7.2(1)P and clause 6.7.2(5)P, respectively; but there is no guidance on
the extent of cracking to be assumed in global analysis.
The assumption in clause 5.4.2.3(5), that eects of cracking in transverse composite
members may be neglected, does not extend to decks with only two main beams. Their
behaviour is inuenced by the length of cantilever cross-beams, if any, and the torsional
stiness of the main beams. The method of clause 5.4.2.3(2) is applicable.
Clause 5.4.2.3(6) supplements clause 5.4.2.2(11), where comment on clause 5.4.2.3(6) is
given.
For the eects of cracking on the design longitudinal shear for the shear connection at
ultimate limit states, clause 5.4.2.3(7) refers to clause 6.6.2.1(2). This requires uncracked
section properties to be used for uncracked members and for members assumed to be
cracked in exure where the eects of tension stiening have been ignored in global analysis.
Where tension stiening and possible over-strength of the concrete (using upper character-
istic values of the tensile strength) have been explicitly considered in the global analysis, then
the same assumptions may be made in the determination of longitudinal shear ow. The
reason for this is that tension stiening can lead to a greater force being attracted to the
shear connection than would be found from a fully cracked section analysis.
The simplest and most conservative way to consider this eect is to determine the longi-
tudinal shear with an uncracked concrete ange. The same approach is required for
fatigue where tension stiening could again elevate fatigue loads on the studs according to
clauses 6.8.5.4(1) and 6.8.5.5(2).
For longitudinal shear at serviceability limit states, clause 5.4.2.3(8) gives, in eect, the
same rules as for ultimate limit states, explained above.
Stages and sequence of construction
The need to consider staged construction is discussed in section 2.2 of this guide. The reason
for allowing staged construction to be ignored at the ultimate limit state, if the conditions of
clause 5.4.2.4(2) are met, is discussed under clause 5.4.2.2(6). However, it would not be
Clause 5.4.2.3(3)
Clause 5.4.2.3(4)
Clause 5.4.2.3(5)
Clause 5.4.2.3(6)
Clause 5.4.2.3(7)
Clause 5.4.2.3(8)
Clause 5.4.2.4(2)
47
common to do this, as a separate analysis considering the staged construction would then be
required for the serviceability limit state.
Temperature eects
Clause 5.4.2.5(1) refers to EN 1991-1-5
53
for temperature actions. These are uniform
temperature change and temperature gradient through a beam, often referred to as dieren-
tial temperature. Dierential temperature produces primary and secondary eects in a
similar way to shrinkage. The reason for allowing temperature to be ignored at the ultimate
limit state, if the conditions of clause 5.4.2.5(2) are met, is discussed under clause 5.4.2.2(6).
Recommended combination factors for temperature eects are given in Tables A2.1 to
A2.4 of Annex A2 of EN 1990. If they are conrmed in the National Annex (as the
further comments assume), temperature will be included in all combinations of actions for
persistent and transient design situations. In this respect, design to Eurocodes will dier
from previous practice in the UK. However, the tables for road bridges and footbridges
have a Note which recommends that
0
for thermal actions may in most cases be reduced
to zero for ultimate limit states EQU, STR and GEO. Only FAT (fatigue) is omitted. It
is unlikely that temperature will have much inuence on fatigue life. The table for railway
bridges refers to EN 1991-1-5. The purpose may be to draw attention to its rules for
simultaneity of uniform and temperature dierence components, and the need to consider
dierences of temperature between the deck and the rails.
With
0
0, temperature eects appear in the ultimate and characteristic combinations
only where temperature is the leading variable action. It will usually be evident which
members and cross-sections need to be checked for these combinations.
The factors
1
and
2
are required for the frequent and quasi-permanent combinations
used for certain serviceability verications. The recommended values are
1
0.6,
2
0.5. Temperature will rarely be the leading variable action, as the following example
shows.
For the eects of dierential temperature, EN 1991-1-5 gives two approaches, from which
the National Annex can select. The Normal Procedure in Approach 2 is equivalent to the
procedure in BS 5400.
11
If this is used, both heating and cooling dierential temperature
cases tend to produce secondary sagging moments at internal supports where crack widths
are checked in continuous beams. These eects of temperature will not normally add to
other eects.
Combinations of actions that involve temperature
A cross-section is considered where the characteristic temperature action eect is T
k
, and the
action eect from trac load model 1 (LM1) for road bridges is Q
k
. The recommended
combination factors are given in Table 5.4.
The frequent combinations of variable action eects are:
.
with load model 1 leading:
1
Q
k
2
T
k
0.75 or 0.40Q
k
0.5T
k
.
with temperature leading:
1
T
k
2
Q
k
0.6T
k
The second of these governs only where 0.1T
k
(0.75 or 0.4)Q
k
.
Thus, temperature should be taken as the leading variable action only where its action
eect is at least 7.5 times (for TS) or 4 times (for UD) that from trac load model 1.
Clause 5.4.2.5(1)
Clause 5.4.2.5(2)
Table 5.4. Recommended combination factors for trac load and
temperature according to EN 1990 Annex A2
Action eect from:
0
2
Tandem system (TS), from LM1 0.75 0.75 0
Uniform loading (UD), from LM1 0.40 0.40 0
Temperature (non-re) 0.6 or zero 0.6 0.5
48
The uncertainty about which variable action leads does not arise in quasi-permanent
combinations, because the combination factor is always
2
. Temperature and construction
loads are the only variable actions for which
2
0 is recommended. Both the values
and the combination to be used can be changed in the National Annex.
Prestressing by controlled imposed deformations
As a principle, clause 5.4.2.6(1) requires that possible deviations from the intended amount
of imposed deection be considered, such as might occur due to the tolerance achievable with
the specied jacking equipment. The eect of variations in material properties on the action
eects developed must also be considered. However, clause 5.4.2.6(2) permits these eects to
be determined using characteristic or nominal values, if the imposed deformations are
controlled. The nature of the control required is not specied. It should take account of
the sensitivity of the structure to any error in the deformation.
At the ultimate limit state, clause 2.4.1.1 recommends a load factor of 1.0 for imposed
deformations, regardless of whether eects are favourable or unfavourable. It is recom-
mended here that where a structure is particularly sensitive to departures from the intended
amount of imposed deformation, tolerances should be determined for the proposed method
of applying these deformations and upper and lower bound values considered in the analysis.
Prestressing by tendons
Prestressing composite bridges of steel and concrete is uncommon in the UK and is therefore
not covered in detail here. Clause 5.4.2.6(1) refers to EN 1992 for the treatment of prestress
forces in analysis. This is generally sucient, although EN 1994 itself emphasises, in
clause 5.4.2.6(2), the distinction between bonded and unbonded tendons. Essentially, this
is that while the force in bonded tendons increases everywhere in proportion to the local
increase of strain in the adjacent concrete, the force in unbonded tendons changes in accor-
dance with the overall deformation of the structure; that is, the change of strain in the adja-
cent concrete averaged over the length of the tendon.
Tension members in composite bridges
The purpose of the denitions (a) and (b) in clause 5.4.2.8(1) is to distinguish between the
two types of structure shown in Fig. 5.9, and to dene the terms in italic print.
In Fig. 5.9(a), the concrete tension member AB is shear-connected to the steel structure,
represented by member CD, only at its ends (concrete here means reinforced concrete).
No design rules are given for a concrete member where cracking is prevented by prestressing.
Figure 5.9(b) shows a composite tension member, which has normal shear connection. Its
concrete ange is the concrete tension member. In both cases, there is a tensile force N from
the rest of the structure, shared between the concrete and steel components.
A member spanning between nodes in a sagging region of a composite truss with a deck at
bottom-chord level could be of either type. The dierence between members of types (a) and
(b) is similar to that between unbonded and bonded tendons in prestressed concrete.
Clause 5.4.2.8(2) lists the properties of concrete that should be considered in global
analyses. These inuence the stiness of the concrete component, and hence the magnitude
Clause 5.4.2.6(1)
Clause 5.4.2.6(2)
Clause 5.4.2.8(1)
Clause 5.4.2.8(2)
(a) (b) (c)
A B
C D
N
N
d
N
M
A
a
A
c
A
s
z
s
M
s
M
a
N
s
N
a
z
a
Fig. 5.9. Two types of tension member, and forces in the steel and concrete parts: (a) concrete tension
member; (b) composite tension member; (c) action eects equivalent to N and M
49
of the force N, and the proportions of it resisted by the two components. Force N is assumed
to be a signicant action eect. There will normally be others, arising from transverse loading
on the member.
The distribution of tension between the steel and concrete parts is greatly inuenced by
tension stiening in the concrete (which is in turn aected by over-strength of the concrete).
It is therefore important that an accurate representation of stiness is made. This clause
allows a rigorous non-linear method to be used. It could be based on Annex L of ENV
1994-2, Eects of tension stiening in composite bridges.
54
This annex was omitted from
EN 1994-2, as being text-book material. Further information on the theory of tension
stiening and its basis in tests is given in Ref. 55, in its references, and below.
The eects of over-strength of concrete in tension can in principle be allowed for by using
the upper 5% fractile of tensile strength, f
ctk.0.95
. This is given in Table 3.1 of EN 1992-1-1 as
30% above the mean value, f
ctm
. However, tension is caused by shrinkage, transverse
loading, etc., as well as by force N, so simplied rules are given in clauses 5.4.2.8(5) to (7).
Clause 5.4.2.8(3) requires eects of shrinkage to be included in calculations of the
internal forces and moments in a cracked concrete tension member. This means the axial
force and bending moment, which are shown as N
s
and M
s
in Fig. 5.9(c). The simplication
given here overestimates the mean shrinkage strain, and should be used for the secondary
eects. This clause is an exception to clause 5.4.2.2(8), which permits shrinkage in cracked
regions to be ignored.
Clause 5.4.2.8(4) refers to simplied methods. The simplest of these, clause 5.4.2.8(5),
which requires both uncracked and cracked global analyses, can be quite conservative.
Clause 5.4.2.8(6) gives a more accurate method for members of type (a) in Fig. 5.9. The
longitudinal stiness of the concrete tension member for use in global analysis is given by
equation (5.6-1):
EA
s
eff
E
s
A
s
,1 0.35,1 n
0
,
s
5.6-1
where: A
s
is the reinforcement in the tension member,
A
c
is the eective cross-sectional area of the concrete, ,
s
A
s
,A
c
, and
n
0
is the short-term modular ratio.
This equation is derived from the model of Annex L of ENV 1994-2 for tension stiening,
shown in Fig. 5.10. The gure relates mean tensile strain, , to tensile force N, in a concrete
tension member with properties A
s
, A
c
, ,
s
and n
0
, dened above. Lines 0A and 0B represent
uncracked and fully cracked behaviour, respectively.
Cracking rst occurs at force N
c.cr
, when the strain is
sr1
. The strain at the crack at
once increases to
sr2
, but the mean strain hardly changes. As further cracks occur, the
mean strain follows the line CD. If the local variations in the tensile strength of concrete
are neglected, this becomes line CE. The eective stiness within this stage of single
cracking is the slope of a line from 0 to some point within CE. After cracking has stabilized,
the stiness is given by a line such as 0F. The strain dierence u
sr
remains constant until
Clause 5.4.2.8(3)
Clause 5.4.2.8(4)
Clause 5.4.2.8(5)
Clause 5.4.2.8(6)
B
A
N
N
c,cr
(EA
s
)
eff
E
D
C
F
0

sr
sr
sr2
sr1

cr
Fig. 5.10. Normal force and mean strain for a reinforced concrete tension member
50
the reinforcement yields. It represents tension stiening, the term used for the stiness of the
concrete between the cracks.
It has been found that in bridges, the post-cracking stiness is given with sucient
accuracy by the slope of line 0E, with u % 0.35. This slope, equation (5.6-1), can be
derived using Fig. 5.10, as follows.
At a force of N
c.cr
the following strains are obtained:
fully cracked strain:
sr2
N
c.cr
,E
s
A
s
uncracked strain:
sr1
N
c.cr
,E
s
A
s
E
c
A
c
Introducing ,
s
A
s
,A
c
and the short-term modular ratio n
0
gives:
sr1
N
c.cr
n
0
,
s
,E
s
A
s
1 n
0
,
s
From Fig. 5.10, the strain at point E is:
cr

sr2
u
sr2
sr1

_
N
c.cr
E
s
A
s
__
1 u
_
1
_
n
0
,
s
1 n
0
,
s
___
This can also be expressed in terms of eective stiness as
cr
N
c.cr
,E
s
A
s
eff
Eliminating
cr
from the last two equations and dividing by N
c.cr
gives:
1,E
s
A
s
eff
1,E
s
A
s
u1 n
0
,
s
,1 n
0
,
s
,E
s
A
s
1 0.35,1 n
0
,
s
,E
s
A
s
which is equation (5.6-1).

In Ref. 56, a study was made of the forces predicted in the tension members of a truss using
a very similar factor to that in equation (5.6-1). Comparison was made against predictions
from a non-linear analysis using the tension eld model proposed in Annex L of ENV 1994-
2.
54
The two methods generally gave good agreement, with most results being closer to those
from a fully cracked analysis than from an uncracked analysis.
The forces given by global analysis using stiness EA
s
eff
are used for the design of the
steel structure, but not the concrete tension member. The tension in the latter is usually
highest just before cracking. For an axially loaded member it is, in theory:
N
Ed
A
c
f
ct.eff
1 n
0
,
s
D5.4
with f
ct.eff
being the tensile strength of the concrete when it cracks and other notation as
above. Usually, there is also tensile stress in the member from local loading or shrinkage.
This is allowed for by the assumption that f
ct.eff
0.7f
ctm
, given in clause 5.4.2.8(6).
Equation (D5.4) is given in this clause with partial factors 1.15 and 1.45 for serviceability
and ultimate limit states, respectively. These allow for approximations in the method.
Thus, for ultimate limit states:
N
Ed.ult
1.45A
c
0.7f
ctm
1 n
0
,
s
1.02f
ctm
A
c
E
s
,E
c
A
s
D5.5
which is the design tensile force at cracking at stress f
ctm
.
Clause 5.4.2.8(7) covers composite members of type (b) in Fig. 5.9. The cross-section
properties are found using equation (5.6-1) for the stiness of the cracked concrete ange,
and are used in global analyses. As an example, it is assumed that an analysis for an ultimate
limit state gives a tensile force N and a sagging moment M, as shown in Fig. 5.9(b). These are
equivalent to action eects N
a
and M
a
in the steel component plus N
s
and M
s
in the concrete
component, as shown. This clause requires the normal force N
s
to be calculated.
Equations for this de-composition of N and M can be derived from elastic section analysis,
neglecting slip, as follows.
The crosses in Fig. 5.9(b) indicate the centres of area of the cross-sections of the concrete
ange (or the reinforcement, for a cracked ange) and the structural steel section. Let z
s
,
Clause 5.4.2.8(7)
51
z
a
and d be as shown in Fig. 5.9(c) and I, I
a
and I
s
be the second moments of area of the
composite section, the steel component, and the concrete ange, respectively.
For M 0. N
a.N
N
s.N
N. and N
s.N
z
s
N
a.N
z
a
. whence N
s.N
Nz
a
,d a
For N 0. N
s.M
N
a.M
b
Equating curvatures. M,I M
s
,I
s
M
a
,I
a
c
For equilibrium. M M
s
M
a
N
a.M
z
a
N
s.M
z
s
d
From equations (b) to (d), with z
a
z
s
d,
M MI
s
,I I
a
,I N
s.M
d. whence N
s.M
d MI
s
I
a
I,I e
For N and M together, from equations (a), (c) and (e):
N
s
Nz
a
,d MI I
a
I
s
,Id D5.6
M
s
MI
s
,I D5.7
In practice, I
s
(I
a
, so I
s
, and hence M
s
, can often be taken as zero. The area of reinforce-
ment in the concrete tension member (A
s,
not A
s.eff
) must be sucient to resist the greater of
force N
s
(plus M
s
, if not negligible) and the force N
Ed
given by equation (D5.5).
Filler beam decks for bridges
There are a great number of geometric, material and workmanship-related restrictions which
have to be met in order to use the application rules for the design of ller beams. These
restrictions are discussed in section 6.3, which deals with the resistances of ller beams,
and are necessary because these clauses are based mainly in existing practice in the UK.
There is very little relevant research.
The same restrictions apply in the use of clause 5.4.2.9(1), which allows the eects of slip
at the concretesteel interface and shear lag to be neglected in global analysis only if these
conditions are met. One signicant dierence from previous practice in the UK is that
fully-encased ller beams are not covered by EN 1994. This is because there are no widely
accepted design rules for longitudinal shear in fully-encased beams without shear connectors.
Clause 5.4.2.9(2) covers the transverse distribution of imposed loading. Its option of
assuming rigid behaviour in the transverse direction may be applicable to a small single-
track railway bridge, but generally, one of the methods of clause 5.4.2.9(3) will be used.
These assume that there are no transverse steel members within the span. It is therefore
essential that continuous transverse reinforcement in both top and bottom faces of the
concrete is provided in accordance with the requirements of clause 6.3.
Clause 5.4.2.9(3) permits global analysis by non-linear methods to clause 5.4.3, but
normally orthotropic plate or grillage analysis will be used. For the longitudinal exural
stiness, smearing of the steel beams involves calculating the stiness of the whole width
of the deck, and hence nding a mean stiness per unit width. It is inferred from clause
5.4.2.9(4) that cracking may be neglected, though clause 5.4.2.9(7) provides an alternative
for some analyses for serviceability.
The exural stiness per unit width in the transverse direction is calculated for the
uncracked concrete slab, neglecting reinforcement. The result is a plate with dierent
properties in orthogonal directions, i.e. orthogonally anisotropic or orthotropic for short.
For grillage analysis, uncracked section properties should generally be used (as required by
clause 5.4.2.9(4)) but it is permissible to account for the loss of stiness in the transverse
direction caused by cracking, by reducing the torsional and exural stinesses of the
transverse concrete members by 50%. This can be advantageous, as it reduces the transverse
moments, and hence the stresses in the reinforcement.
The longitudinal moments obtained from elastic analysis of an orthotropic slab or grillage
may not be redistributed to allow for cracking. This is because cracking can occur in both
hogging and sagging regions, and there is insucient test evidence on which to base
Clause 5.4.2.9(1)
Clause 5.4.2.9(2)
Clause 5.4.2.9(3)
Clause 5.4.2.9(4)
52
design rules. However, clause 5.4.2.9(5) permits, for some analyses, up to 15% redistribution
of hogging moments for beams in Class 1 at internal supports. This is less liberal than it may
appear, because clause 5.5.3, which covers classication, does not relax the normal rules for
Class 1 webs or anges to allow for restraint from encasement. The concrete does, however,
reduce the depth of web in compression.
There are no provisions for creep of concrete at ultimate limit states, so clause 5.4.2.2
applies in the longitudinal direction. In the transverse direction, clause 3.1 of EN 1992-1-1
presumably applies. The modular ratios in the two directions may be found to be dierent,
because of the
L
factors in EN 1994.
These comments on creep also apply for deformations, from clause 5.4.2.9(6). Shrinkage
can be neglected, because it causes little curvature where there is little dierence between the
levels of the centroids of the steel and concrete cross-sections.
Clause 5.4.2.9(7) gives a simplied rule for the eects of cracking of concrete on deec-
tions and camber. Clause 5.4.2.9(8) permits temperature eects to be ignored, except in
certain railway bridges.
Clause 5.4.2.9(5)
Clause 5.4.2.9(6)
Clause 5.4.2.9(7)
Clause 5.4.2.9(8)
Example 5.2: modular ratios for long-term loading and for shrinkage
For the bridge of Example 5.1 (Fig. 5.6), modular ratios are calculated for:
.
imposed load (short-term loading)
.
superimposed dead load (long-term loading)
.
eects of shrinkage (long-term).
Modular ratios for long-term loading depend on an assumed eective thickness for
the concrete deck slab and a mean age at rst loading. The choice of these values is
discussed in comments on clause 5.4.2.2. Here, it is assumed that superimposed dead
load is applied when the concrete has an average age of 7 days, and that the eective
thickness of the deck slab is the value before application of waterproong, which is its
actual thickness, 250 mm. The deck concrete is grade C30/37 and the relative humidity
is 70%.
Live load
From Table 3.1 of EN 1992-1-1, E
cm
33 kN/mm
2
From clause 3.2.6 of EN 1993-1-1, E
a
210 N/mm
2
for structural steel,
The short-term modular ratio, n
0
E
a
,E
cm
210,33 6.36
Superimposed dead load (long term), from Annex B of EN 1992-1-1
From equation (B.1) of EN 1992-1-1, the creep coecient ct. t
0
c
0
u
c
t. t
0
, where
u
c
t. t
0
is a factor that describes the amount of creep that occurs at time t. When
t !1, u
c
1. The total creep is therefore given by:
c
0
c
RH
u f
cm
ut
0
(B.2) in EN 1992-1-1
where c
RH
is a factor to allow for the eect of relative humidity on the notional creep co-
ecient. Two expressions are given, depending on the size of f
cm
.
From Table 3.1 of EN 1992-1-1, f
cm
f
ck
8 30 8 38 N/mm
2
For f
cm
35 MPa. c
RH

_
1
1 RH,100
0.1
h
0
3
_ c
1
_
c
2
(B.3b) in EN 1992-1-1
RH is the relative humidity of the ambient environment in percentage terms; here, 70%.
The factors c
1
and c
2
allow for the inuence of the concrete strength:
c
1

_
35
f
cm
_
0.7
_
35
38
_
0.7
0.944
53
c
2

_
35
f
cm
_
0.2
_
35
38
_
0.2
0.984 (B.8c) in EN 1992-1-1
From equation (B.3b) of EN 1992-1-1:
c
RH

_
1
1 70,100
0.1
250
3
p 0.944
_
0.984 1.43
From equation (B.4) of EN 1992-1-1, the factor u f
cm
, that allows for the eect of
concrete strength on the notional creep coecient, is:
u f
cm

16.8
f
cm
_
16.8
38
p 2.73
The eect of the age of the concrete at rst loading on the notional creep coecient
is given by the factor ut
0
according to equation (B.5) of EN 1992-1-1. For loading at
7 days this gives:
ut
0

1
0.1 t
0.20
0

1
0.1 7
0.20
0.63
This expression is only valid as written for normal or rapid-hardening cements.
The nal creep coecient from equation (B.2) of EN 1992-1-1 is then:
c1. t
0
c
0
c
RH
u f
cm
ut
0
1.43 2.73 0.63 2.48
From equation (5.6), the modular ratio is given by:
n
L
n
0
1
L
ct. t
0
6.361 1.1 2.48 23.7
where
L
1.1 for permanent load.
Eects of shrinkage, from Annex B of EN 1992-1-1
The calculation of the creep factor is as above, but the age at loading is assumed to be one
day, from clause 5.4.2.2. For the factor ut
0
this gives:
ut
0

1
0.1 t
0.20
0

1
0.1 1
0.20
0.91
The nal creep coecient from equation (B.2) of EN 1992-1-1 is then:
c1. t
0
c
0
c
RH
u f
cm
ut
0
1.43 2.73 0.91 3.55
From equation (5.6), the modular ratio is given by:
n
L
n
0
1
L
ct. t
0
6.361 0.55 3.55 18.8
where
L
0.55 for the primary and secondary eects of shrinkage.
Example 5.3: primary eects of shrinkage
For the bridge in Example 5.1 (Fig. 5.6) the plate thicknesses of an internal beam at mid-
span of the main span are as shown in Fig. 5.11. The primary eects of shrinkage at this
cross-section are calculated. The deck concrete is grade C30/37 and the relative humidity
is 70%.
It is assumed that, for the majority of the shrinkage, the length of continuous concrete
deck is such that shear lag eects are negligible. The eective area of the concrete ange is
taken as the actual area, for both the shrinkage and its primary eects. The secondary
eects arise from changes in the reactions at the supports. For these, the eective
widths should be those used for the other permanent actions.
The free shrinkage strain is found rst, from clause 3.1.4 of EN 1992-1-1. By inter-
polation, Table 3.2 of EN 1992-1-1 gives the drying shrinkage as
cd.0
352 10
6
.
54
The factor k
h
allows for the inuence of the shape and size of the concrete cross-section.
From Example 5.2, h
0
250 mm. From Table 3.3 of EN 1992-1-1, k
h
0.80.
The long-term drying shrinkage strain is 0.80 352 10
6
282 10
6
.
The long-term autogenous shrinkage strain is:
ca
1 2.5 f
ck
10 10
6
2.5 30 10 10
6
50 10
6
250
1225
400 30
400 20
1175 12.5
25 haunch
3100
998
NA
0.35
12
T
T
C
1.08
49
Fig. 5.11. Cross-section of beam for Example 5.3, and primary shrinkage stresses with n
L
18.8
(T: tension; C: compression)
From clause 2.4.2.1 of EN 1992-1-1, the recommended partial factor for shrinkage is
SH
1.0, so the design shrinkage strain, for both serviceability and ultimate limit
states, is:
e
sh
1.0 282 50 10
6
332 10
6
From Example 5.2, the modular ratio for shrinkage is n
L
18.8.
The tensile force F
c
to restore the slab to its length before shrinkage applies to the
concrete a tensile stress:
332 10
6
210 10
3
,18.8 3.71 N,mm
2
The area of the concrete cross-section is:
A
c
3.1 0.25 0.4 0.025 0.785 m
2
therefore
F
c

sh
A
c
E
a
,n
L
332 0.785 210,18.8 2913 kN
For n
L
18.8, the location of the neutral axis of the uncracked unreinforced section is
as shown in Fig. 5.11. This is 375 mm below the centroid of the concrete area. Relevant
properties of this cross-section are:
I 21 890 10
6
mm
4
and A 76 443 mm
2
The total external force is zero, so force F
c
is balanced by applying a compressive force
of 2913 kN and a sagging moment of 2913 0.375 1092 kNm to the composite section.
The long-term primary shrinkage stresses in the cross-section are as follows, with
compression positive:
at the top of the slab,
o 3.71
_
2913 10
3
76 443

1092 502
21 890
_
1
18.8
0.35 N,mm
2
at the interface, in concrete,
o 3.71
_
2913 10
3
76 443

1092 227
21 890
_
1
18.8
1.08 N,mm
2
55
5.4.3. Non-linear global analysis for bridges
The provisions of EN 1992 and EN 1993 on non-linear analysis are clearly relevant, but are
not referred to from clause 5.4.3. It gives Principles, but no Application Rules. It is stated in
EN 1992-2 that clause 5.7(4)P of EN 1992-1-1 applies. It requires stinesses to be represented
in a realistic way taking account of the uncertainties of failure, and concludes Only those
design formats which are valid within the relevant elds of application shall be used.
Clause 5.7 of EN 1992-2, Non-linear analysis, consists mainly of Notes that give
recommendations to national annexes. The majority of the provisions can be varied in the
National Annex as agreement could not be obtained at the time of drafting over the use
of the safety format proposed. The properties of materials specied in the Notes have
been derived so that a single safety factor can be applied to all materials in the verication.
Further comment is given under clause 6.7.2.8.
Clause 5.4.1 of EN 1993-2 requires the use of an elastic analysis for all persistent and
transient design situations. It has a Note that refers to the use of plastic global analysis
for accidental design situations, and to the relevant provisions of EN 1993-1-1. These
include clause 5.4, which denes three types of non-linear analysis, all of which refer to
plastic behaviour. One of them, rigid-plastic analysis, should not be considered for
composite bridge structures. This is evident from the omission from EN 1994-2 of a clause
corresponding to clause 5.4.5 of EN 1994-1-1, Rigid plastic global analysis for buildings.
This method of drafting arises from Notes to clauses 1.5.6.6 and 1.5.6.7 of EN 1990, which
make clear that all of the methods of global analysis dened in clauses 1.5.6.6 to 1.5.6.11
(which include plastic methods) are non-linear in Eurocode terminology. Non-linear in
these clauses of EN 1990 refers to the deformation properties of the materials, and not to
geometrical non-linearity (second-order eects), although these have to be considered
when signicant, as discussed in section 5.2 above.
Non-linear analysis must satisfy both equilibrium and compatibility of deformations when
using non-linear material properties. These broad requirements are given to enable methods
more advanced than linear-elastic analysis to be developed and used within the scope of the
Eurocodes.
Unlike clause 5.4.1 of EN 1993-2, EN 1994-2 makes no reference to the use of plastic
analysis for accidental situations, such as vehicular impact on a bridge pier or impact on a
parapet. The National Annex to EN 1993-2 will give guidance. It is recommended that
this be followed also for composite design.
Further guidance on non-linear analysis is given in EN 1993-1-5, Annex C, on nite-
element modelling of steel plates.
5.4.4. Combination of global and local action eects
A typical local action is a wheel load on a highway bridge. It is expected that the National
Annex for the UK will require the eects of such loads to be combined with the global
eects of coexisting actions for serviceability verications, but not for checks for ultimate
limit states. This is consistent with current practice to BS 5400.
11
The Note to clause 5.4.4(1) refers to Normative Annex E of EN 1993-2. This annex was
written for all-steel decks, where local stresses in welds can be signicant and where local and
global stresses always combine unfavourably. It recommends a combination factor for
local and global eects that depends on the span and ranges from 0.7 to 1.0. The application
Clause 5.4.3
Clause 5.4.4(1)
at the interface, in steel,
o
2913 10
3
76 443

1092 227
21 890
49.4 N,mm
2
at the bottom of the steel beam,
o
2913 10
3
76 443

1092 998
21 890
11.7 N,mm
2
56
of this rule to reinforced concrete decks that satisfy the serviceability requirements for the
combined eects is believed to be over-conservative, because of the benecial local eects
of membrane and arching action. By contrast, if the EN 1993 rules are adopted, global
compression in the slab is usually favourable so consideration of 70% of the maximum
compressive global stress when checking local eects may actually be unconservative.
5.5. Classication of cross-sections
The classication of cross-sections of composite beams is the established method of taking
account in design of local buckling of plane steel elements in compression. It determines the
available methods of global analysis and the basis for resistance to bending, in the same way
as for steel members. Unlike the method in EN 1993-1-1, it does not apply to columns. The
Class of a steel element (a ange or a web) depends on its edge support conditions, b/t ratio,
distribution of longitudinal stress across its width, yield strength, and in composite sections,
the restraint provided against buckling by any attached concrete or concrete encasement.
A ow diagram for the provisions of clause 5.5 is given in Fig. 5.12. The clause numbers
given are from EN 1994-2, unless noted otherwise.
Clause 5.5
Note 1: Flange means
steel compression flange
No
No
Yes
Yes
Compression
flange is:
Yes
No Yes
No
No
Yes
From clause 5.5.2(2), classify the web using the
elastic stress distribution, to Table 5.2 of EC3-1-1
Section is Class 1
Section is Class 2
Yes
Effective
section is
equivalent
Class 2
Section is Class 3
Yes
Section is
Class 4
(see Note 2)
No
Note 2. Where the stress levels in a Class 4 section satisfy the limit given in
EC3-1-1/5.5.2(9), it can be treated as a Class 3 section, from clause 5.5.1(1)P.
No
Is steel compression
flange restrained by
shear connectors to
clause 5.5.2(1)?
Is the section a filler
beam to clause 6.3?
From clause 5.5.2(2), classify
flange to Table 5.2 of EC3-1-1
Classify flange to Table 5.2
of EC4-2
Class 1 Class 4 Class 2 Class 3
Locate the elastic neutral axis, taking
account of sequence of construction,
creep, and shrinkage, to clause 5.5.1(4)
Locate the plastic neutral axis, using gross
web and effective flanges, to clause 5.5.1(4)
Web is Class 3 Web not classified
Is web encased to clause 5.5.3(2)?
Is the compression flange
in Class 1 or 2?
Replace web by effective web
in Class 2, to clause 5.5.2(3)?
Web is Class 1
Is the flange in Class 1?
Web is Class 2
Web is Class 3
Is the compression
flange in Class 3?
Web is Class 4
Effective web
is Class 2
From clause 5.5.2(2), classify the web
using the plastic stress distribution, to
Table 5.2 of EC3-1-1
Fig. 5.12. Classication of a cross-section of a composite beam
57
Clause 5.5.1(1)P refers to EN 1993-1-1 for denitions of the four Classes and the
slendernesses that dene the Class boundaries. Classes 1 to 4 correspond respectively to
the terms plastic, compact, semi-compact and slender that were formerly used in
BS 5950.
48
The classications are done separately for steel anges in compression and
steel webs. The Class of the cross-section is the less favourable of the Classes so found,
clause 5.5.1(2), with one exception: the hole-in-web option of clause 5.5.2(3).
Idealized momentrotation curves for members in the four Classes are shown in Fig. 5.13.
In reality, curves for sections in Class 1 or 2 depart from linearity as soon as (or even before)
the yield moment is reached, and strain-hardening leads to a peak bending moment higher
than M
pl
, as shown.
The following notes supplement the denitions given in clause 5.5.2(1) of EN 1993-1-1:
.
Class 1 cross-sections can form a plastic hinge and tolerate a large plastic rotation
without loss of resistance. It is a requirement of EN 1993-1-1 for the use of rigid-
plastic global analysis that the cross-sections at all plastic hinges are in Class 1. For
composite bridges, EN 1994-2 does not permit rigid-plastic analysis. A Note to clause
5.4.1(1) of EN 1993-2 enables its use to be permitted, in a National Annex, for certain
accidental design situations for steel bridges.
.
Class 2 cross-sections can develop their plastic moment resistance, M
pl.Rd
, but have
limited rotation capacity after reaching it because of local buckling. Regions of
sagging bending in composite beams are usually in Class 1 or 2. The resistance M
pl.Rd
exceeds the resistance at rst yield, M
el.Rd
, by between 20% and 40%, compared with
about 15% for steel beams. Some restrictions are necessary on the use of M
pl.Rd
in combi-
nation with elastic global analysis, to limit the post-yield shedding of bending moment to
adjacent cross-sections in Class 3 or 4. These are given in clauses 6.2.1.2(2) and
6.2.1.3(2).
.
Class 3 cross-sections become susceptible to local buckling before development of the
plastic moment of resistance. In clause 6.2.1.5(2) their bending resistance is dened as
the elastic resistance, governed by stress limits for all three materials. A limit may be
reached when the compressive stress in all restrained steel elements is below yield.
Some rotation capacity then remains, but it is impracticable to take advantage of it in
design.
.
Class 4 cross-sections are those in which local buckling will occur before the attainment
of yield stress in one or more parts of the cross-section. This is assumed in EN 1993 and
EN 1994 to be an ultimate limit state. The eective cross-section should be derived in
accordance with EN 1993-1-5. Guidance is given in comments on clause 6.2.1.5(7),
which denes the procedure, and in the Designers Guide to EN 1993-2.
4
Clause 5.5.1(1)P
Clause 5.5.1(2)
M
pl
M
el
M
pl
M
el
M
pl
M
el
M
pl
M
el
M

Class 1 Class 2
M
M
Class 3
M
Class 4
Typical curve
from a test
Fig. 5.13. Idealized momentrotation relationships for sections in Classes 1 to 4
58
The Class of a cross-section is determined from the width-to-thickness ratios given in
Table 5.2 of EN 1993-1-1 for webs and anges in compression. The numbers appear dierent
from those in BS 5400:Part 3:2000
11
because the coecient that takes account of yield
strength, , is dened as
p
235,f
y
in the Eurocodes, and as
p
355,f
y
in BS 5400. After
allowing for this, the limits for webs at the Class 2/3 boundary agree closely with those in
BS 5400, but there are dierences for anges. For outstand anges, EN 1993 is more
liberal at the Class 2/3 boundary, and slightly more severe at the Class 3/4 boundary. For
internal anges of boxes, EN 1993 is considerably more liberal for all Classes.
Reference is sometimes made to a beam in a certain Class. This may imply a certain
distribution of bending moment. Clause 5.5.1(2) warns that the Class of a composite
section depends on the sign of the bending moment (sagging or hogging), as it does for a
steel section that is not symmetrical about its neutral axis for bending.
Clause 5.5.1(3) permits account to be taken of restraint from concrete in determining the
classication of elements, providing that the benet has been established. Further comment
is given at clause 5.5.2(1) on spacing of shear connectors.
Since the Class of a web depends on the level of the neutral axis, which is dierent for
elastic and plastic bending, it may not be obvious which stress distribution should be used
for a section near the boundary between Classes 2 and 3. Clause 5.5.1(4) provides the
answer: the plastic distribution. This is because the use of the elastic distribution could
place a section in Class 2, for which the bending resistance would be based on the plastic
distribution, which in turn could place the section in Class 3.
Elastic stress distributions should be built up by taking the construction sequence into
account, together with the eects of creep (generally through the use of dierent modular
ratios for the dierent load types) and shrinkage.
Where a steel element is longitudinally stiened, it should be placed in Class 4 unless it can
be classied in a higher Class by ignoring the longitudinal stieners.
Where both axial load and moment are present, these should be combined when deriving
the plastic stress block. Alternatively, the web Class can conservatively be determined on the
basis of compressive axial load alone.
Clause 5.5.1(5), on the minimum area of reinforcement for a concrete ange, appears
here, rather than in Section 6, because it gives a further condition for a cross-section to be
placed in Class 1 or 2. The reason is that these sections must maintain their bending
resistance, without fracture of the reinforcement, while subjected to higher rotation than
those in Class 3 or 4. This is ensured by disallowing the use of bars in ductility Class A
(the lowest), and by requiring a minimum cross-sectional area, which depends on the
tensile force in the slab just before it cracks.
55
Clause 5.5.1(6), on welded mesh, has
the same objective. Clause 3.2.4 of EN 1992-2 does not recommend the use of Class A re-
inforcement for bridges in any case, but this recommendation can be modied in a National
Annex.
During the construction of a composite bridge, it is quite likely that a beam will change its
section Class, because the addition of the deck slab both prevents local buckling of the top
ange and signicantly shifts the neutral axis of the section. Typically, a mid-span section
could be in Class 1 or 2 after casting the slab but in Class 3 or 4 prior to this. Clause
5.5.1(7) requires strength checks at intermediate stages of construction to be based on the
relevant classication at the stage being checked.
The words without concrete encasement in the title of clause 5.5.2 are there because
this clause is copied from EN 1994-1-1, where it is followed by a clause on beams with
web encasement. These are outside the scope of EN 1994-2.
Clause 5.5.2(1) is an application of clause 5.5.1(3). The spacing rules to which it refers
may be restrictive where full-thickness precast deck slabs are used. Clause 5.5.2(2) adds
little to clause 5.5.1.
The hole-in-web method
This useful device rst appeared in BS 5950-3-1.
48
It is now in clause 6.2.2.4 of EN 1993-1-1,
which is referred to from clause 5.5.2(3).
Clause 5.5.1(3)
Clause 5.5.1(4)
Clause 5.5.1(5)
Clause 5.5.1(6)
Clause 5.5.1(7)
Clause 5.5.2
Clause 5.5.2(1)
Clause 5.5.2(2)
Clause 5.5.2(3)
59
In beams subjected to hogging bending, it often happens that the bottom ange is in Class
1 or 2, and the web is in Class 3. The initial eect of local buckling of the web would be a
small reduction in the bending resistance of the section. The assumption that a dened
depth of web, the hole, is not eective in bending enables the reduced section to be upgraded
from Class 3 to Class 2, and removes the sudden change in the bending resistance that would
otherwise occur. The method is analogous to the use of eective areas for Class 4 sections, to
allow for local buckling.
There is a limitation to its scope that is not evident in the following wording, from
EN 1993-1-1:
The proportion of the web in compression should be replaced by a part of 20t
w
adjacent to the
compression ange, with another part of 20t
w
adjacent to the plastic neutral axis of the eective
cross-section.
It follows that for a design yield strength f
yd
, the compressive force in the web is limited
to 40t
w
f
yd
. For a composite beam in hogging bending, the tensile force in the longitudinal
reinforcement in the slab can exceed this value, especially where f
yd
is reduced to allow for
vertical shear. The method is then not applicable, because the second element of 20t
w
is
not adjacent to the plastic neutral axis, which lies within the top ange. The method, and
this limitation, are illustrated in Examples in the Designers Guide to EN 1994-1-1.
5
It should be noted that if a Class 3 cross-section is treated as an equivalent Class 2 cross-
section for section design, it should still be treated as Class 3 when considering the actions to
consider in its design. Indirect actions, such as dierential settlement, which may be
neglected for true Class 2 sections, should not be ignored for eective Class 2 sections.
The primary self-equilibrating stresses could reasonably be neglected, but not the secondary
eects.
Clause 5.5.3(2) and Table 5.2 give allowable width-to-thickness ratios for the outstands of
exposed anges of ller beams. Those for Class 2 and 3 are greater than those from Table 5.2
of EN 1993-1-1. This is because even though a ange outstand can buckle away from
the concrete, rotation of the ange at the junction with the web is prevented (or at least
the rotational stiness is greatly increased) by the presence of the concrete.
Clause 5.5.3(2)
Example 5.4: classication of composite beam section in hogging bending
The classication of the cross-section shown in Fig. 5.14 is determined for hogging
bending moments. The eective ange width is 3.1 m. The top layer of reinforcement
comprises pairs of 20 mm bars at 150 mm centres. The bottom layer comprises single
20 mm bars at 150 mm centres. All reinforcement has f
sk
500 N/mm
2
and
S
1.15.
These bars are shown in assumed locations, which would in practice depend on the speci-
ed covers and the diameter of the transverse bars.
1225
400 40
400 25
1160 25
775
Plastic NA
209
Elastic
NA
60
70
8.47 MN
3.45 MN
10.01 MN
5.52 MN
250
25
8.21 MN
1.80 MN
3100
Fig. 5.14. Cross-section of beam for Example 5.4
The yield strength of structural steel is thickness-dependent and is 345 N/mm
2
from
EN 10025 for steel between 16 mm and 40 mm thick. (If Table 3.1 of EN 1993-1-1 is
60
used, the yield strength can be taken as 355 N/mm
2
for steel up to 40 mm thick. The choice
of method can be specied in the National Annex. It is likely that this will require the
value from the relevant product standard to be used.)
Hence,
235,345
_
0.825.
Steel bottom ange
Ignoring the web-to-ange welds, the ange outstand c 400 25,2 187.5 mm.
From Table 5.2 of EN 1993-1-1, the condition for Class 1 is:
c,t < 9 9 0.825 7.43
For the ange, c,t 187.5,40 4.7, so the ange is Class 1.
Steel web
To check if the web is in Class 1 or 2, the plastic neutral axis must be determined, using
design material strengths.
The total area of reinforcement is:
A
s
3 100 3100,150 19 480 mm
2
so its tensile force at yield is:
19.48 0.5,1.15 8.47 MN
Similarly, the forces in the structural steel elements at yield are found to be as shown in
Fig. 5.14. The total longitudinal force is 27.45 MN, so at M
pl.Rd
the compressive force is
27.45,2 13.73 MN. The plastic neutral axis is evidently within the web. The depth of
web in compression is:
1160 13.73 5.52,10.01 951 mm
From Table 5.2 of EN 1993-1-1, for a part subject to bending and compression, and
ignoring the depth of the web-to-ange welds:
c 951,1160 0.82. 0.5
The web is in Class 2 if:
c,t 456,13c 1 456 0.825,9.66 39.0
Its ratio c,t 1160,25 46.4, so it is not in Class 2. By inspection, it is in Class 3, but
the check at the Class 3/4 boundary is given, as an example. It should be based on the
built-up elastic stresses, which may not be available. It is conservative to assume that
all stresses are applied to the composite section as this gives the greatest depth of web
in compression.
The location of the elastic neutral axis in hogging bending must be found, for the
cracked reinforced composite section. There is no need to use a modular ratio for re-
inforcement as its modulus may be taken equal to that for structural steel, according to
clause 3.2(2). The usual rst moment of area calculation nds the neutral axis to be
as shown in Fig. 5.14, so the depth of web in compression is 775 40 735 mm.
Again neglecting the welds, from Table 5.2 of EN 1993-1-1 the stress ratio is
1160 735,735 0.58
and the condition for a Class 3 web is:
c,t 42,0.67 0.33 42 0.825,0.67 0.19 72.4
The actual c,t 1160,25 46.4, so the composite section is in Class 3 for hogging
bending.
61
Flow charts for global analysis
The ow charts given in Figs 5.15 and 5.16 are for a bridge with the general layout shown in
Fig. 5.1. Figure 5.15, for the superstructure, provides design forces and displacements for the
beams and at the four sets of bearings shown. Figure 5.16, for the columns, takes account of
system instability shown in Fig. 5.1(b). Instability of members in compression is covered in
comments on clause 6.7.3.4.
For simplicity, the scope of these charts is limited by assumptions, as follows:
.
Fatigue, vibration, and settlement are excluded.
.
Axial force in the superstructure (e.g. from friction at bearings) is negligible.
.
The main imposed loading is trac Load Model 1, from EN 1991-2.
.
Only persistent design situations are included.
.
The limit states considered are ULS (STR) and SLS (deformation and crack width).
.
The superstructure consists of several parallel continuous non-hybrid plate girders
without longitudinal stieners, composite with a reinforced normal-density-concrete
deck slab.
.
There are no structural steel transverse members at deck-slab level.
.
The only steel cross-sections that may be in Class 4 are the webs near internal supports.
The depth of web in compression is inuenced by the ratio of non-composite to compo-
site bending moment and the area of reinforcement in the slab. The Class is therefore
dicult to predict until some analyses have been done.
.
The deck is constructed unpropped, and all structural deck concrete is assumed to be in
place before any of the members become composite.
.
The formwork is structurally participating precast concrete planks. They are assumed
(for simplicity here) to have the same creep and shrinkage properties as the in situ
concrete of the deck.
.
All joints except bearings are assumed to be continuous (clause 5.1.2).
.
Bearings are simple joints, with or without longitudinal sliding, as shown in Fig. 5.1.
Transverse sliding cannot occur.
In the charts, creep and shrinkage eects are considered only as long-term values
(t !1). The values of all Nationally Determined Parameters, such as and factors,
are assumed to be those recommended in the Notes in the Eurocodes.
The following data are assumed to be available, based on preliminary analyses and the
strengths of the materials to be used, f
y
, f
sk
and f
ck
(converted from an assumed f
cu
):
.
dimensions of the anges and webs of the plate girders
.
dimensions of the cross-sections of the concrete deck and the two supporting systems, BE
and CF in Fig. 5.1(a)
.
details and weight of the superimposed dead load (nishes, parapets, etc.)
.
estimated areas of longitudinal slab reinforcement above internal supports.
Assumptions relevant to out-of-plane system instability are as follows. The deck transmits
most of the lateral wind loading to supports A and D, with negligible restraint from the two
sets of internal supports. The lateral deections of nodes B and C inuence the design of the
columns, but stinesses are such that wind-induced system instability is not possible.
The following abbreviations are used:
.
EC2 means EN 1992-1-1 and/or EN 1992-2; similarly for EC3.
.
A clause in EN 1994-2 is referred to as, for example, 5.4.2.2.
.
Symbols g
k1
and g
k2
are used for characteristic dead loads on the steelwork, and on
composite members, respectively. Superimposed dead load is g
k3
. Shrinkage is g
sh
.
.
Characteristic imposed loads are denoted q
k
(trac), w
k
(wind) and t
k
(temperature).
There is not space to list on Fig. 5.15 the combinations of actions required. The notation in
the lists that follow is that each symbol, such as g
k2
, represents the sets of action eects (M
Ed
,
V
Ed
, deformations, etc.) resulting from the application of the arrangement of the action g
k2
that is most adverse for the action eect considered.
62
For the variable actions q
k
and w
k
, dierent arrangements govern at dierent cross-
sections, so envelopes are required. This may apply also for t
k
, as several sets of temperature
actions are specied.
For nding the cracked regions of longitudinal members, it is assumed that the short-
term values are critical, because creep may reduce tensile stress in concrete more than
Determine elastic properties of materials: E
a
,E
s
(taken as E
a
), E
cm
, etc., from 3.1 to 3.3.
Assume a mean value t
0
for permanent actions, to 5.4.2.2(3).
Determine creep coefficients (, t
0
); hence find modular ratios n
0
, n
L,s
, and n
L,p
for
short-term, shrinkage, and permanent actions, respectively (5.4.2.2(2)).
Determine free shrinkage strain,
cs
(, 1 day), to 3.1(3) and 5.4.2.2(4)
Stability. If no buckling mode can be envisaged, as here for the superstructure, then
cr
10 can be assumed, and global analyses can be first-order, to 5.2.1(3)
Select exposure (environmental) class(es) for concrete surfaces, to 7.1(3). Hence find minimum
covers to concrete from EC2/4.4, and the locations of main tensile reinforcement in composite
beams. These are required for cracked section properties
Determine shear lag effective widths for global analyses:
for concrete flanges, at mid-span and internal supports, to 5.4.1.2
for steel flanges, find b
eff
for SLS and ULS, to EC3-1-5/3.1 to 3.3
Imperfections. Determine system imperfections to 5.3.2(1). Out-of-plumb of supports BE and
CF are relevant for these members, but are assumed not to affect this flow chart.
For imperfections of column members, see flow chart of Fig. 6.44
Estimate distributions of longitudinal stress in steel webs at internal supports, and classify
them, to 5.5. If any are in Class 4, thicken them or find effective properties to EC3-1-5/2.2
Flexural stiffnesses of cross-sections. Determine E
a
I
1
for all uncracked composite and
concrete sections, for modular ratios n
0
, n
L,s
and n
L,p
, using effective widths at mid-span, or
at supports for cantilevers (1.5.2.11). Reinforcement may be included or omitted.
Determine E
a
I
2
for cracked reinforced longitudinal composite sections in hogging bending, with
effective widths as above (1.5.2.12 ). Represent bearings by appropriate degrees of freedom
Global analyses. Are
all span ratios 0.6?
Use stiffnesses to
5.4.2.3(3) (15% rule)
for global analyses
Do global analyses for the load cases g
k1
(on the steel structure), g
k2
, g
k3
, g
sh
, q
k
, w
k
and t
k
using both short-term and long-term modular ratios for the variable actions and for g
k2
, g
k3
and long-term for g
sh
. Refer to expressions (D5.9) etc., for the combinations required. (END)
5.4.2.3(2) applies. Use uncracked stiffnesses and modular
ratios n
0
, n
L,s
and n
L,p
. Analyse for load cases: g
k2
, g
k3
, g
sh
.
Find moment and shear envelopes for q
k
and t
k
.
Find the highest extreme-fibre tensile stresses in concrete,
f
ct,max
, for the combinations listed in expression D(5.8),
with effective widths to 5.4.1.2(7)
Find regions of longitudinal members where f
ct,max
> 2f
ctm
,
and reduce stiffnesses of these regions to E
a
I
2
Yes
No
Fig. 5.15. Flow chart for global analysis for superstructure of three-span bridge
63
shrinkage increases it. From clause 5.4.2.3(2), the following characteristic combinations are
required for nding cracked regions:
.
with trac leading: g
k2
g
k3
q
k
0.w
w
k
.
with wind leading: g
k2
g
k3
0.q
q
k
w
k
.
with temperature leading: g
k2
g
k3
0.q
q
k
0.w
w
k
t
k
(D5.8)
In practice, of course, it will usually be evident which combination governs. Then, only
regions in tension corresponding to that combination need be determined.
For nding the most adverse action eects for the limit state ULS (STR), all combinations
include the design permanent action eects:
G
g
k1
g
k2
g
k3
g
sh
using the more adverse of the long-term or short-term values. To these are added, in turn, the
following combinations of variable action eects:
.
with trac leading: 1.35q
k
1.5
0.w
w
k
.
with wind leading: 1.35
0.q
q
k
1.5w
k
.
with temperature leading: 1.35
0.q
q
k
1.5
0.w
w
k
t
k
(D5.9)
For serviceability limit states, deformation is checked for frequent combinations. The
combination for crack width is for national choice, and frequent is assumed here. These
combinations all include the permanent action eects as follows, again using the more
adverse of short-term and long-term values:
g
k1
g
k2
g
k3
g
sh
To these are added, in turn, the following combinations of variable action eects:
.
with trac leading:
1.q
q
k
2.t
t
k
because
2.w
0
.
with wind leading:
1.w
w
k
2.t
t
k
because
2.q
0
.
with temperature leading:
1.t
t
k
(D5.10)
As before, it will usually be evident, for each action eect and location, which combination
governs.
Flow chart for supporting systems at internal supports
At each of points B and C in Fig. 5.1, it is assumed that the plate girders are supported on at
least two bearings, mounted on a cross-head that is supported by a composite frame or by
two or more composite columns, xed at points E and F. Each bearing acts as a spherical
pin. Design action eects and displacements (six per bearing) are known, for each limit
state, from analyses of the superstructure.
Preliminary cross-sections for all the members have been chosen. Composite columns are
assumed to be within the scope of clause 6.7.3 (doubly symmetrical, uniform, etc.). The ow
chart of Fig. 5.16 is for a single composite column, and is applicable to composite columns
generally. For ultimate limit states, only long-term behaviour is considered, as this usually
governs.
Notes on Fig. 5.16
(1) For the elastic critical buckling force N
cr
, the eective length for an unbraced column, as
in Fig. 5.1(b), is at least 2L, where L is the actual length. If the foundation cannot be
assumed to be rigid, its rotational stiness should be included in an elastic critical
analysis, as the eective length then exceeds 2L.
In many cases,
"
`` will be much less than 2, and c
cr
will far exceed 10. These checks can
then be done approximately, by simple hand calculation. Other methods of checking if
second-order global analysis is required are discussed under clause 5.2.1.
Here, it is assumed that for the transverse direction, c
cr
10. No assumption is made
for the plane shown in Fig. 5.1. The ow chart of Fig. 5.16, which is for a single column,
includes second-order system eects in this plane.
64
(2) Out-of-plumb of columns, a system imperfection, should be allowed for as follows.
Figure 5.17(a) shows a nominally vertical column of length L, with design action
eects M, N and V from a preliminary global analysis. The top-end moment M could
represent an o-centre bearing. The design out-of-plumb angle for the column, c, is
found from clause 5.3.2(3) of EN 1993-1-1. Notional horizontal forces Nc are
applied, as shown in Fig. 5.17(b).
Second-order global analysis for the whole structure, including base exibility, if any,
then gives the deformations of the ends of the column and the action eects N
Ed
, etc.,
Yes
Determine elastic properties of materials, mean value t
0
, creep coefficients (, t
0
),
modular ratios n
0
, n
L,s
and n
L,p
, and shrinkage strain,
cs
(, 1 day), as in Fig. 5.15.
Select exposure class(es) for concrete, and find minimum covers, all as in Fig. 5.15.
All effective widths are assumed to be actual widths. Check that concrete covers satisfy
6.7.5.1, and reinforcement satisfies 6.7.5.2. Modify if necessary
For a concrete-encased section, find concrete cross-section for use in calculations from
6.7.3.1(2) on excessive cover. Check that 6.7.3.1(3) on longitudinal reinforcement and
6.7.3.1(4) on shape of section are satisfied
Find N
pl,Rd
from eq. (6.30), steel contribution
ratio from eq. (6.38), and check that
0.2 0.9, to 6.7.1(4). Find N
pl,Rk
to 6.7.3.3(2)
Determine N
cr
and then , to 6.7.3.3(2). Check
that 2, to 6.7.3.1(1). See Note 1 in main text
Estimate ratio of permanent to total design normal force (axial compression) at ULS
Hence find E
c,eff
from eq. (6.41), to allow for creep. Find the characteristic and design
flexural stiffnesses, (EI )
eff
and (EI )
eff,II
, from eqs (6.40) and (6.42)
Note: Until the final box, below,
the chart is for both y and z
planes, separately, and the
y or z subscript is not given
First-order global
analysis permitted
by5.2.1(3), with
system imperfections
Use second-order global analysis,
including both system and member
imperfections
5.3.2(2) EC3-2 EC3-1-1/5.3.2(6)
No
Use method (i) or (ii)
(see comment on
clause 5.2.2) for
global analysis
Yes
Perform global analysis, to find
action effects N
Ed
, M
Ed
and V
Ed
at each end of column
Repeat for the other plane of bending. Include
member imperfection only in the plane where
its effect is more adverse; 6.7.3.7(1) (END)
Find system imperfection to EC3-1-1/5.3.2(3) and hence notional transverse forces
N
Ed
at ends of column, and add to coexisting transverse forces (e.g. from wind loading)
See Note 2 in the main text
No
Note: For verification of the column length, see flow chart of Fig. 6.44
Estimate maximum
design axial force, N
Ed
.
Find
cr
= N
cr
/N
Ed
.
Is
cr
10? See Note 1
Is
cr
< 4?
Fig. 5.16. Flow chart for global analysis of a composite column
65
needed for verication of the column length. In Fig. 5.17(c), c
0
is the rotation of the
column base, and w
Ed
represents the transverse loading, which may be negligible. For
the determination of the bending moments within a column length and its verication,
reference should be made to the ow chart of Fig. 6.36 of Ref. 5.
L
M
w
Ed
M
Ed,1
M
Ed,2
V
Ed,1
V
Ed,2
N
Ed
N
Ed
M
N N
(a) (b)
(c)
V + N
V
0
Fig. 5.17. System imperfection and global analysis for a column
66
CHAPTER 6
Ultimate limit states
.
Beams Clause 6.1
.
Resistances of cross-sections of beams Clause 6.2
.
Filler beam decks Clause 6.3
.
Lateraltorsional buckling of composite beams Clause 6.4
.
Transverse forces on webs Clause 6.5
.
Shear connection Clause 6.6
.
Composite columns and composite compression members Clause 6.7
.
Fatigue Clause 6.8
.
Tension members in composite bridges Clause 6.9
Clauses 6.1 to 6.7 dene resistances of cross-sections to static loading, for comparison with
action eects determined by the methods of Section 5. The ultimate limit state considered is
STR, dened in clause 6.4.1(1) of EN 1990 as:
Internal failure or excessive deformation of the structure or structural members . . . where the
strength of constructional materials of the structure governs.
The self-contained clause 6.8, Fatigue, covers steel, concrete, and reinforcement by cross-
reference to Eurocodes 2 and 3. Requirements are given for shear connection.
Clause 6.9 does not appear in EN 1994-1-1 and has been added in EN 1994-2 to cover
concrete and composite tension members such as may be found in tied arch bridges and
truss bridges.
6.1. Beams
6.1.1. Beams in bridges general
Clause 6.1.1(1) serves as a summary of the checks that should be performed on the beams
themselves (excluding related elements such as bracing and diaphragms). The checks listed
are as follows:
.
Resistance of cross-sections to bending and shear clauses 6.2 and 6.3. In the Eurocodes,
local buckling in Class 4 members, due to direct stress, is covered under the heading of
cross-section resistance, even though this buckling resistance is derived considering a
nite length of the beam. In Eurocode 3, shear buckling is similarly covered under the
heading of cross-section resistance, but this is separately itemized below. A check of
the interaction between shear and bending is required in clause 6.2.2.4.
.
Resistance to lateraltorsional buckling clause 6.4. For lateraltorsional buckling, the
resistance is inuenced by the properties of the whole member. The rules of Eurocode 4
assume that the member is of uniform cross-section, apart from variations arising from
Clause 6.1.1(1)
cracking of concrete and from detailing. The resistance of non-uniform members is
covered in clause 6.3.4 of EN 1993-2.
.
Resistance to shear buckling and in-plane forces applied to webs clauses 6.2.2 and 6.5
respectively. As discussed above, shear buckling resistance is treated as a property of a
cross-section.
.
Resistance to longitudinal shear clause 6.6. According to clause 1.1.3(3), provisions for
shear connection are given only for welded headed studs. This is misleading, for much of
clause 6.6 is more widely applicable, as discussed under clause 6.6.1.
.
Resistance to fatigue clause 6.8.
The above checks are not exhaustive. Further checks that may be required include the
following:
.
Interaction with axial force. Axial force is not included in the checks above as clause
1.5.2.4 denes a composite beam as a composite member subjected mainly to bending.
Axial force does however occur in the beams of composite integral bridges.
57
This is
discussed in section 6.4 of this guide.
.
Addition of stresses in webs and anges generated from plan curvature, although this is
identied in clause 6.2.1.1(5). No method of combining (or calculating) these eects is
provided in Eurocodes 3 or 4. The Designers Guide to EN 1993-2
4
provides some
guidance, as do the comments on clause 6.2.1.1(5).
.
Flange-induced buckling of the web clause 6.5.2 refers.
.
Torsion in box girders, which adds to the shear in the webs and necessitates a further
check on the ange Section 7 of EN 1993-1-5 refers. The need to consider combinations
of torsion and bending is mentioned in clause 6.2.1.3(1).
.
Distortion of box girders, which causes both in-plane and out-of-plane bending in the
box walls clause 6.2.7 of EN 1993-2 refers and the Designers Guide to EN 1993-2 pro-
vides some guidance.
.
Torsion of bare steel beams during construction, which often arises with the use of
cantilever forms to construct the deck edge cantilevers. This usually involves a considera-
tion of both St Venant torsion and warping torsion.
.
Design of transverse stieners Section 9 of EN 1993-1-5 refers.
Steel cross-sections may be rolled I- or H-section or doubly-symmetrical or mono-
symmetrical plate girders. Other possible types include any of those shown in sheet 1 of
Table 5.2 of EN 1993-1-1; this includes box girders. Channel and angle sections should
not be used unless the shear connection is designed to provide torsional restraint or there
is adequate torsional bracing between beams.
6.1.2. Eective width for verication of cross-sections
Eective widths for shear lag are discussed in section 5.4.1.2 of this guide. Unlike in global
analysis, the eective width appropriate to the cross-section under consideration must be
used in calculation of resistance to bending. Distributions of eective width along a span
are given in Figure 5.1.
6.2. Resistances of cross-sections of beams
This clause is for beams without partial or full encasement in concrete. Filler beams with
partial encasement are treated in clause 6.3. Full encasement is outside the scope of
EN 1994.
No guidance is given in EN 1994, or in EN 1993, on the treatment of large holes in steel
webs without recourse to nite-element modelling (following the requirements of EN1993-1-
5), but specialized literature is available.
58.59
Bolt holes in steelwork should be treated in
accordance with EN 1993-1-1, particularly clauses 6.2.2 to 6.2.6.
68
6.2.1. Bending resistance
6.2.1.1. General
In clause 6.2.1.1, three dierent approaches are given, based on rigid-plastic theory, non-
linear theory, and elastic theory. The non-linear theory is that given in clause 6.2.1.4.
This is not a reference to non-linear global analysis.
Clause 6.2.1.1(1) only permits rigid-plastic theory to be used where cross-sections are in
Class 1 or 2 and where prestressing by tendons is not used. This is because no explicit
check of yielding of bonded tendons is given and therefore non-linear resistance calculation
is more appropriate. Comment on this use of plastic resistance with elastic analysis is given
under clause 5.4.1.1(1).
Clause 6.2.1.1(2) permits non-linear theory and elastic theory to be used for all cross-
sections. If unbonded tendons are used, the tendon forces used in section analysis should
however be derived in accordance with clause 5.4.2.7(2).
The assumption that composite cross-sections remain plane is always permitted by clause
6.2.1.1(3) where elastic and non-linear theory are used, because the conditions set will be
satised if the design is in accordance with EN 1994. The implication is that longitudinal
slip is negligible.
There is no requirement for slip to be determined. This would be dicult because the
stiness of shear connectors is not known accurately, especially where the slab is cracked.
Wherever slip may not be negligible, the design methods of EN 1994-2 are intended to
allow for its eects.
For beams with curvature in plan, clause 6.2.1.1(5) gives no guidance on how to allow for
the torsional moments induced or how to assess their signicance. Normal practice is to treat
the changing direction of the longitudinal force in a ange (and a web, if signicant) as a
transverse load applied to that ange, which is then designed as a horizontal beam spanning
between transverse restraints. It is common to use elastic section resistance in such circum-
stances to avoid the complexity of producing a plastic stress block for the combined local and
global loading. The shear connection and bracing system should be designed for the addi-
tional transverse forces.
Similar calculation should be carried out where curvature is achieved using a series of
straight sections, except that the transverse forces will be concentrated at the splices
between adjacent lengths. Particular care is needed with detailing the splices. Transverse
stieners and bracing will usually be needed close to each splice to limit the bending in the
ange. In box girders, the torsion from curvature will also tend to produce distortion of
the box. This must be considered in the design of both the cross-section of the box and its
internal restraints.
Bending in transverse planes can also be induced in anges by curvature of the ange in a
vertical plane, and should be considered. Clause 6.5 covers the transverse forces on webs that
this causes, but not the transverse bending in the ange. The latter is covered in the
4
6.2.1.2. Plastic resistance moment M
pl.Rd
of a composite cross-section
Full interaction in clause 6.2.1.2(1)(a) means that no account need be taken of slip or
separation at the steelconcrete interface.
Full interaction should not be confused with full shear connection. That concept is used
only in the rules for buildings, and is explained in clause 6.1.1(7)P of EN 1994-1-1 as
follows:
A span of a beam. . . has full shear connection when increase in the number of shear connectors
would not increase the design bending resistance of the member.
This link of shear connection to bending resistance diers from the method of EN 1994-2,
where shear connection is related to action eects, both static and fatigue. Shear connection
to Part 2 is not necessarily full according to the above denition (which should strictly read
. . . number of shear connectors within a critical length . . .). It would be confusing to refer to
it as partial, so this term is never used in Part 2.
Clause 6.2.1.1(1)
Clause 6.2.1.1(2)
Clause 6.2.1.1(3)
Clause 6.2.1.1(5)
Clause
6.2.1.2(1)(a)
69
Reinforcement in compression
It is usual to neglect slab reinforcement in compression (clause 6.2.1.2(1)(c)). Its eect on
the bending resistance of the composite section is negligible unless the slab is unusually
small. If it is included, and the concrete cover is little greater than the bar diameter, consid-
eration should be given to possible buckling of the bars.
Guidance on detailing is given in clauses 9.5.3(6) and 9.6.3(1) of EN 1992-1-1 for
reinforcement in concrete columns and walls respectively. The former requires that no bar
within a compression zone should be further than 150 mm from a restrained bar, but
restrained is not dened. This could be interpreted as requiring all compression bars in
an outer layer to be within 150 mm of a bar held in place by transverse reinforcement.
This would usually require link reinforcement in the ange. This interpretation was used
in BS 5400 Part 4
11
for compression bars assumed to contribute to the resistance of the
section. If the compression ange is classed as a wall, clause 9.6.3 of EN 1992-1-1 requires
only that the longitudinal bars are placed inside horizontal (i.e. transverse) reinforcement
unless the reinforcement in compression exceeds 2% of the gross concrete area. In the
latter case, transverse reinforcement must be provided in accordance with the column rules.
Stress/strain properties for concrete
The design compressive strength of concrete, f
cd
, is dened in clause 3.1.6(1)P of EN 1992-1-1
as:
f
cd
c
cc
f
ck
,
C
where:
c
cc
is the coecient taking account of long termeects on the compressive strength and of
unfavourable eects resulting from the way the load is applied.
Note: The value of c
cc
for use in a country should lie between 0.8 and 1.0 and may be
found in its National Annex. The recommended value is 1.
The reference in clause 3.1(1) to EN 1992-1-1 for properties of concrete begins unless
otherwise given by Eurocode 4. Resistances of composite members given in EN 1994-2 are
based on extensive calibration studies (e.g. Refs 60, 61). The numerical coecients given
in resistance formulae are consistent with the value c
cc
1.0 and the use of either elastic
theory or the stress block dened in clause 6.2.1.2. Therefore, there is no reference in
EN 1994-2 to a coecient c
cc
or to a choice to be made in a National Annex. The symbol
f
cd
always means f
ck
,
C
, and for beams and most columns is used with the coecient
0.85, as in equation (6.30) in clause 6.7.3.2(1). An exception, in that clause, is that the
0.85 is replaced by 1.0 for concrete-lled column sections, based on calibration.
The approximation made to the shape of the stressstrain curve is also relevant. Those
given in clause 3.1 of EN 1992-1-1 are mainly curved or bilinear, but in clause 3.1.7(3)
there is a simpler rectangular stress distribution, similar to the stress block given in the
British Standard for the structural use of concrete, BS 8110.
62
Its shape, for concrete strength
classes up to C50/60, and the corresponding strain distribution are shown in Fig. 6.1 below.
This stress block is inconvenient for use with composite cross-sections, because the region
near the neutral axis assumed to be unstressed is often occupied by a steel ange, and
algebraic expressions for resistance to bending become complex.
Clause
6.2.1.2(1)(c)
x
Plastic
neutral axis
0 0.0035
Compressive strain
0.8x
0
EN 1994-1-1:
0.85f
cd
, with f
cd
= f
ck
/
C
Compressive stress
EN 1992-1-1:
f
cd
=
cc
f
ck
/
C
Fig. 6.1. Stress blocks for concrete at ultimate limit states
70
In composite sections, the contribution from the steel section to the bending resistance
reduces the signicance of that from the concrete. It is thus possible
63
for EN 1994 to
allow the use of a rectangular stress block extending to the neutral axis, as shown in Fig. 6.1.
For a member of unit width, the moment about the neutral axis of the EN 1992 stress
block ranges from 0.38f
ck
x
2
,
C
to 0.48f
ck
x
2
,
C
, depending on the value chosen for c
cc
.
The value for beams in EN 1994-2 is 0.425f
ck
x
2
,
C
. Calibration studies have shown that
this overestimates the bending resistance of cross-sections of columns, so a correction
factor c
M
is given in clause 6.7.3.6(1). See also the comments on clause 6.7.3.6.
Small concrete anges
Where the concrete slab is in compression, the method of clause 6.2.1.2 is based on the
assumption that the whole eective areas of steel and concrete can reach their design
strengths before the concrete begins to crush. This may not be so if the concrete ange is
small compared with the steel section. This lowers the plastic neutral axis, and so increases
the maximum compressive strain at the top of the slab, for a given tensile strain in the steel
bottom ange.
A detailed study of the problem has been reported.
64
Laboratory tests on beams show
that strain hardening of steel usually occurs before crushing of concrete. The eect of this,
and the low probability that the strength of both the steel and the concrete will be only at
the design level, led to the conclusion that premature crushing can be neglected unless the
grade of the structural steel is higher than S355. Clause 6.2.1.2(2) species a reduction
in M
pl.Rd
where the steel grade is S420 or S460 and the depth of the plastic neutral axis
is high.
For composite columns, the risk of premature crushing led to a reduction in the factor c
M
,
given in clause 6.7.3.6(1), for S420 and S460 steels.
Ductility of reinforcement
Reinforcement with insucient ductility to satisfy clause 5.5.1(5), and welded mesh, should
not be included within the eective section of beams in Class 1 or 2 (clause 6.2.1.2(3)). This is
because laboratory tests on hogging moment regions have shown
33
that some reinforcing
bars, and most welded meshes, fracture before the momentrotation curve for a typical
double-cantilever specimen reaches a plateau. The problem with welded mesh is explained
in comments on clause 3.2(3).
6.2.1.3. Additional rules for beams in bridges
Clause 6.2.1.3(1) is a reminder that composite beams need to be checked for possible com-
binations of internal actions that are not specically covered in EN 1994. The combinations
given are biaxial bending, bending and torsion and local and global eects, with a reference
to clause 6.2.1(5) of EN 1993-1-1.
Signicant bending about a vertical axis is rare in composite bridges so biaxial bending is
rarely a concern. Despite the reference to EN 1993-1-1, there is little of direct relevance for
biaxial bending in composite beams therein. For Class 1 and 2 cross-sections, the interaction
of EN 1993-1-1 expression (6.2) could be used for resistance of cross-sections. Rather than
computing the resultant plastic stress block for axial load and biaxial bending, a linear inter-
action is provided:
N
Ed
N
Rd
M
y.Ed
M
y.Rd
M
z.Ed
M
z.Rd
1.0
where N
Rd
, M
y.Rd
and M
z.Rd
are the design resistances for each eect acting individually,
with reductions for shear where the shear force is suciently large. In theory, it is still
necessary to derive the resultant plastic stress block to check whether the cross-section is
either Class 1 or 2. This complexity can be avoided by performing the classication under
axial compression only. The same expression can be applied to Class 3 and 4 cross-sections
or the stresses can be summed using elastic section analysis. Care is needed where the sign of
the stress in the slab is dierent for each constituent action.
Clause 6.2.1.2(2)
Clause 6.2.1.2(3)
Clause 6.2.1.3(1)
71
In hogging zones of integral bridges, where there is usually a moderate coexistent axial
load induced by temperature or soil pressure, it is common to do calculations on the basis
of the fully cracked section. The non-linear method of clause 6.2.1.4 could also be used
but this is likely to require the use of computer software. Buckling needs to be checked
separately see section 6.4.
Signicant torsion is unlikely to be encountered in most composite I-girder bridges due to
the low St Venant torsional stiness of the steel beams. There are some exceptions including:
.
torsion in curved beams as discussed in comments on clause 6.2.1.1(5)
.
torsion in skew decks at end trimmers
.
torsion of bare steel beams where formwork for deck cantilevers is clamped to the outer
girders.
EN1993-1-1 clause 6.2.7(7) permits St Venant torsion to be ignored at ultimate limit states
provided that all the torsion is carried by resistance to warping. This is usually the most
ecient model and avoids a further interaction with shear stress from vertical shear in the
web. If the torque is resisted by opposing bending in the anges, they can be designed for
this bending combined with their axial force. If the length between restraints should be
long, then the warping bending stresses would become large and the section would try to
resist the torsion predominantly through St Venant shear ow. In that case it might be
better to derive the separate contributions from St Venant and warping torsion. Further
guidance on shear, torsion and bending is provided in the Designers Guide to EN 1993-2.
4
Pure torsion in box beams is treated simply by a modication to the shear stress in the
webs and anges, and the design is checked using clause 7.1 of EN 1993-1-5. Pure torsion
is however rare and most boxes will also suer some distortion. This leads to both in-
plane warping and out-of-plane bending of the box walls as discussed in Ref. 4.
The reference in clause 6.2.1.3(1) to combined local and global eects relates to the steel
beamonly, because this combination in a concrete deck is a matter for Eurocode 2, unless the
deck is a composite plate, when clause 9.3 applies. Such combinations include bending, shear
and transverse load (fromwheel loads) according to clause 6.2.8(6) of EN1993-1-1 and other
combinations of local and global load. The Von Mises equivalent stress criterion of EN1993-
1-1 expression (6.1) should be used in the absence of test-based interaction equations for
resistances.
Clause 6.2.1.3(2) relates to the use of plastic resistances in bending, which implies
shedding of bending moments, typically from mid-span regions to adjacent supports.
Non-linear global analysis allows for this, but linear-elastic analysis does not. The reasons
for permitting linear analysis, and for the limitations given in the present clause, are
explained in comments on clause 5.4.1.1(1). Amethod for making use of the limited ductility
of support regions has been proposed.
65
Clause 6.2.1.3(2)
Example 6.1: plastic resistance moment in sagging bending
For the bridge in Example 5.1 (Fig. 5.6), the resistance moment for an internal beam at
mid-span with the cross-section in Fig. 5.11 is determined. The deck concrete is Grade
C30/37 and the structural steel is S355 J2 G3. The cross-section is shown in Fig. 6.2
with relevant dimensions, stress blocks and longitudinal forces. The notation is as in
Fig. 6.2 and the concrete stress block as in Fig. 6.1.
Using the partial factors recommended in EN 1992-2 and EN 1993-2, the design
strengths are:
f
yd
345,1.0 345 N,mm
2
. f
cd
30,1.5 20 N,mm
2
The steel yield stress has been taken here as 345 N/mm
2
throughout. This is given in
EN 10025 for thicknesses between 16 mm and 40 mm. For the web, 355 N/mm
2
could
have been used.
72
250
1225
400 20
400 30
1175 12.5
25 haunch
3100
227
NA
172
769
1372
4.14 MN
2.76 MN
5.07 MN
11.97 MN
0.85f
cd
f
yd
Fig. 6.2. Plastic resistance of cross-section to sagging bending, for Example 6.1
Ignoring the haunch and any slab reinforcement, the available compressive force in the
concrete is:
N
c.f
3.1 0.25 0.85 20 13.18 MN
The available forces in the steel beam are:
in the top ange, N
a.top
0.4 0.02 345 2.76 MN
in the web, N
a.web
1.175 0.0125 345 5.07 MN
in the bottom ange, N
a.bot
0.4 0.03 345 4.14 MN
The total is 11.97 MN. As this is less than 13.18 MN, the neutral axis lies in the concrete
slab, at a depth:
z
na
25011.97,13.18 227 mm
The distances below force N
c
of the lines of action of the three forces N
a
are shown in
Fig. 6.2. Hence:
M
pl.Rd
2.76 0.172 5.07 0.769 4.14 1.372 10.05 MNm
The cross-section at the adjacent support is in Class 3, so potentially clause 6.2.1.3(2)
applies. Since the ratio of adjacent spans is 0.61, greater than the limit of 0.6, there is no
need to restrict the bending resistance at mid-span to 0.90M
pl.Rd
.
Example 6.2: resistance to hogging bending at an internal support
For the bridge shown in Fig. 5.6, and materials as in Example 6.1, the resistance to
hogging bending of the cross-section shown in Fig. 5.14 is studied. It was found in
Example 5.4 that the anges are in Class 1 and the web is in Class 3.
It appears from clause 5.5.2(3) that an eective section in Class 2 could be used, to
clause 6.2.2.4 of EN 1993-1-1. However, the wording of that clause implies, and its
Fig. 6.3 shows, that the plastic neutral axis of the eective section should lie within the
web. It is concluded in the Designers Guide to EN 1994-1-1
5
that the hole-in-web approx-
imation should only be used when this is so. As the area of longitudinal reinforcement in
the present cross-section is quite high, this condition is checked rst.
From clause 6.2.2.4 of EN 1993-1-1, the eective area of web in compression is 40t
2
w
.
Using rectangular stress blocks, the condition is:
A
s
f
sd
A
a.top
f
yd
< 40t
2
w
f
yd
A
a.bot
f
yd
Hence:
A
s
< 40t
2
w
A
a.bot
A
a.top
f
yd
,f
sd
D6.1
73
6.2.1.4. Non-linear resistance to bending
There are two approaches, described in clause 6.2.1.4. With both, the calculations should be
done at the critical sections for the design bending moments. The rst approach, given in
clause 6.2.1.4(1) to (5), enables the resistance of a section to be determined iteratively
from the stressstrain relationships of the materials.
Clause 6.2.1.4(1)
to (5)
From Example 5.4:
A
s
19 480 mm
2
. and
235,345
p
0.825
From Fig. 5.14:
A
a.bot
A
a.top
400 15 6000 mm
2
. t
w
25 mm
For the reinforcement:
f
sd
500,1.15 435 N,mm
2
From expression (D6.1):
A
s
< 40 25
2
0.825 6000 345,435 21 120 mm
2
Thus, expression (D6.1) is satised, so M
pl.Rd
will be found by the hole-in-web method.
The longitudinal forces in the reinforcement and the two steel anges are found as in
Example 6.1, and are shown in Fig. 6.3.
2.82
120
3.45
0.715
3.56
3.56
1420
1194
5.52
5.645
1160
25
412
Plastic NA
253
250
412
281.5
529
227.5
Forces in MN units
Hole in web
z
web
= 83
400 25
400 40
Fig. 6.3. Plastic resistance of cross-section to hogging bending, for Example 6.2
The depth of each of the two compressive stress blocks in the web is:
20t
w
20 25 0.825 412 mm
The force in each of them is 20t
2
w
f
yd
0.412 0.025 345 3.56 MN
The location of the plastic neutral axis is found from longitudinal equilibrium. The
tensile force in the web is:
T
w
3.56 2 5.52 5.645 2.82 3.45 0.715 MN 715 kN
This force requires a depth of web:
z
w
715,25 0.345 83 mm
This leads to the depth of the hole in the web, 253 mm, and the distances of the various
forces below the level of the top reinforcement, shown in Fig. 6.3. Taking moments about
this level gives the bending resistance, which is:
M
pl.Rd
12.64 MNm
74
A curvature (strain gradient) and neutral-axis position are assumed, the stresses deter-
mined from the strains, and the neutral axis moved until the stresses correspond to the
external longitudinal force, if any. The assumed strain distributions should allow for the
shrinkage strain of the concrete and any strain and/or dierence of curvature between
steel and concrete caused by temperature. The bending resistance is calculated from this
stress distribution. If it exceeds the external moment M
Ed
, the calculation is terminated. If
not, the assumed curvature is increased and the process repeated until a value M
Rd
is
found that exceeds M
Ed
. If one of the ultimate strains given in EN 1992-1-1 for concrete
and reinforcement is reached rst, the cross-section has insucient resistance. For Class 3
cross-sections or Class 4 eective cross-sections, the compressive strain in the structural
steel must not exceed that at rst yield.
Clearly, in practice this procedure requires the use of software. For sections in Class 1 or
2, a simplied approach is given in clause 6.2.1.4(6). This is based on three points on
the curve relating longitudinal force in the slab, N
c
, to design bending moment M
Ed
that are easily determined. With reference to Fig. 6.4, which is based on Fig. 6.6, these
points are:
.
A, where the composite member resists no moment, so N
c
0
.
B, which is dened by the results of an elastic analysis of the section, and
.
C, based on plastic analysis of the section.
Accurate calculation shows BC to be a convex-upwards curve, so the straight line BC is a
conservative approximation. Clause 6.2.1.4(6) thus enables hand calculation to be used.
The elastic analysis gives the resistance M
el.Rd
, which is calculated according to equation
(6.4). The moment acting on the composite section will generally comprise both short-
term and permanent actions and in calculating the stresses from these, appropriate
modular ratios should be used in accordance with clause 5.4.2.2(2).
Clause 6.2.1.4(7) makes reference to EN 1993-1-1 for the stressstrain relationship to be
used for prestressing steel. The prestrain (which is the initial tendon strain after all losses,
calculated in accordance with EN 1992-1-1 clause 5.10.8) must be taken into account in
the section design. For bonded tendons, this can be done by displacing the origin of the
stressstrain curve along the strain axis by an amount equal to the design prestrain and
assuming that the strain change in the tendon is the same as that in the surrounding concrete.
For unbonded tendons, the prestress should be treated as a constant force equal to the
applied force after all losses. The general method of section analysis for composite
columns in clause 6.7.2 would then be more appropriate.
6.2.1.5. Elastic resistance to bending
Clause 6.2.1.4(6) includes, almost incidentally, a denition of M
el.Rd
that may seem strange.
It is a peculiarity of composite structures that when unpropped construction is used, the
elastic resistance to bending depends on the proportion of the total load that is applied
before the member becomes composite. Let M
a.Ed
and M
c.Ed
be the design bending
Clause 6.2.1.4(6)
Clause 6.2.1.4(7)
C
1.0
M
pl,Rd
N
c
/N
c,f
N
c,el
/N
c,f
M
el,Rd
M
a,Ed
B
0
A
Fig. 6.4. Non-linear resistance to bending for Class 1 and 2 cross-sections
75
moments for the steel and composite sections, respectively, for a section in Class 3. Their
total is typically less than the elastic resistance to bending, so to nd M
el.Rd
, one or both
of them must be increased until one or more of the limiting stresses in clause 6.2.1.5(2) is
reached. To enable a unique result to be obtained, clause 6.2.1.4(6) says that M
c.Ed
is to
be increased, and M
a.Ed
left unchanged. This is because M
a.Ed
is mainly from permanent
actions, which are less uncertain than the variable actions whose eects comprise most of
M
c.Ed
.
Unpropped construction normally proceeds by stages, which may have to be considered
individually in bridge design. While the sequence of erection of the beams is often known
in the design stage, the concrete pour sequence is rarely known. Typically, either a range
of possible pour sequences is considered or it is assumed that the whole of the wet concrete
is placed simultaneously on the bare steelwork, and the resulting design is rechecked when
the pour sequence is known.
The weight of formwork is, in reality, applied to the steel structure and removed from the
composite structure. This process leaves self-equilibrated residual stresses in composite
cross-sections. Whether or not this is considered in the nal situation is a matter for judge-
ment, depending on the signicance of the weight of the formwork.
One permanent action that inuences M
el.Rd
is shrinkage of concrete. Clause 6.2.1.5(5)
enables the primary stresses to be neglected in cracked concrete, but the implication is
that they should be included where the slab is in compression. This provision should not
be confused with clause 5.4.2.2(8), although it is consistent with it. The self-equilibrating
stresses from the primary eects of shrinkage do not cause any moment but they can give
rise to stress. In checking the beam section, if these stresses are adverse, they should be
added to those from M
a.Ed
and M
c.Ed
when verifying stresses against the limits in clause
6.2.1.5(2). If it is necessary to determine the actual elastic resistance moment, M
el.Rd
, the
shrinkage stresses should be added to the stresses from M
a.Ed
and kM
c.Ed
when determining
k and hence M
el.Rd
. If this addition increases M
el.Rd
, it could be omitted, but this is not a
requirement, because shrinkage is classied as a permanent action.
Clause 6.2.1.5(6) is a reminder that lateraltorsional buckling should also be checked,
which applies equally to the other methods of cross-section design. The calculation of
M
el.Rd
is relevant for Class 3 cross-sections if the method of clause 6.4.2 is used, but the
above problem with shrinkage does not occur as the slab will be in tension in the critical
region.
Additional guidance is required for Class 4 cross-sections since the eectiveness of the
Class 4 elements (usually only the web for composite I-beams) depends on the stress distri-
butions within them. The loss of eectiveness for local buckling is dealt with by the use of
eective widths according to EN 1993-1-5. For staged construction, there is the additional
problem that the stress distribution changes during construction and therefore the size
and location of the eective part of the element also change at each stage.
To avoid the complexity of summing stresses from dierent eective cross-sections, clause
6.2.1.5(7) provides a simplied pragmatic rule. This requires that the stress distribution at
any stage is built up using gross-section properties. The reference to gross sections is not
intended to mean that shear lag can be neglected; it refers only to the neglect of plate
buckling. The stress distribution so derived is used to determine an eective web which is
then used to determine section properties and stresses at all stages up to the one considered.
The Note to clause 4.4(3) of EN 1993-1-5 provides almost identical guidance, but claries
that an eective ange should be used together with the gross web to determine the initial
stress distribution. Eective in this sense includes the eects of both shear lag and plate
buckling. Plate buckling for anges is likely to be relevant only for box girders. Example
6.3 illustrates the method.
Clause 6.2.1.5(7) refers to some clauses in EN 1993-1-5 that permit mid-plane stresses in
steel plates to be used in verications. For compression parts in Class 3, EN 1993-2 follows
clause 6.2.1(9) of EN 1993-1-1. This says:
Compressive stresses should be limited to the yield strength at the extreme bres.
Clause 6.2.1.5(2)
Clause 6.2.1.5(5)
Clause 6.2.1.5(6)
Clause 6.2.1.5(7)
76
It is followed by the Note:
The extreme bres may be assumed at the midplane of the anges for ULS checks. For fatigue see
EN 1993-1-9.
This concession can be assumed to apply also to composite beams.
An assumption in the eective section method is that there is sucient post-buckling
strength to achieve the necessary redistribution of stress to allow all components to be
stressed to their individual resistances. This approach is therefore not permitted (and is
not appropriate) in a number of situations where there may not be sucient post-buckling
strength or where the geometry of the member is outside prescribed limits. These exceptions
are given in EN 1993-1-5, clause 2.3(1).
Where prestressing is used, clause 6.2.1.5(8) limits the stress in tendons to the elastic range
and makes reference to clause 5.10.8 of EN 1992-1-1 for guidance on initial prestrain. The
latter covers both bonded and unbonded prestress.
Clause 6.2.1.5(9) provides an alternative method of treating Class 4 cross-sections using
Section 10 of EN 1993-1-5. This method can be used where the conditions of EN 1993-1-5
clause 2.3(1) are not met. Section 10 requires that all stresses are calculated on gross sections
and buckling checks are then carried out on the component plates of the cross-section. There
is usually economic disadvantage in using this method because the benecial load shedding
of stress around the cross-section implicit in the eective section method does not occur.
Additionally, the benet of using test-based interactions between shear and bending is lost.
If the whole member is prone to overall buckling instability, such as exural or lateral
torsional buckling, these eects must either be calculated by second-order analysis and the
additional stresses included when checking panels or by using a limiting stress o
limit
in
member buckling checks. For exural buckling, o
limit
can be calculated based on the
lowest compressive value of axial stress o
x.Ed
acting on its own, required to cause buckling
failure in the weakest sub-panel or an entire panel, according to the verication formula
in Section 10 of EN 1993-1-5. This value of o
limit
is then used to replace f
y
in the member
buckling check. It is conservative, particularly when the critical panel used to determine
o
limit
is not at the extreme compression bre of the section where the greatest stress increase
during buckling occurs. For lateraltorsional buckling, o
limit
can be determined as the
bending stress at the extreme compression bre needed to cause buckling in the weakest
panel. This would however again be very conservative where o
limit
was determined from
buckling of a web panel which was not at the extreme bre, as the direct stress in a web
panel would not increase much during lateraltorsional buckling.
A detailed discussion of the use of Section 10 of EN 1993-1-5 is given in the Designers
Guide to EN 1993-2.
4
Clause 6.2.1.5(8)
Clause 6.2.1.5(9)
Example 6.3: elastic bending resistance of a Class 4 cross-section
For the bridge in Example 5.1 (Fig. 5.6), the mid-span section of the internal beam in
Fig. 5.11 continues to the splice adjacent to each pier. The top and bottom layers of
reinforcement comprise 16 mm bars at 150 mm centres. There are 20 mm transverse
bars, with top and bottom covers of 40 mm and 45 mm respectively, so the locations of
the 16 mm bars are as shown in Fig. 6.5. All reinforcement has f
sk
500 N/mm
2
and
S
1.15. The steel yield stress is taken as 345 N/mm
2
throughout.
The cross-section is checked for the ultimate limit state hogging moments adjacent to
the splice, which are as follows:
steel beam only: M
a.Ed
150 kNm
cracked composite beam: M
c.Ed
2600 kNm (including secondary eects of
shrinkage)
By inspection, the cross-section is not in Class 1 or 2 so its classication is checked at the
Class 3/4 boundary using elastic stresses.
77
Clause 3.2(2)
47
247
68
73
1225
25
250
400 20
400 30
1175 12.5
4155 mm
2
4155 mm
2
389
Elastic neutral axis,
effective composite section
109
Fig. 6.5. Eective section and reinforcement for Example 6.3
Steel bottom ange
Ignoring the web-to-ange welds, the ange outstand c 400 12.5,2 193.8 mm.
From Table 5.2 of EN 1993-1-1, the condition for Class 1 is:
c,t < 9 9 0.825 7.43
For the ange, c,t 193.8,30 6.46, so the ange is Class 1.
Steel web
The area of each layer of reinforcement is A
s
64 3100,150 4155 mm
2
. It would
be conservative to assume that all stresses are applied to the composite section as this gives
the greatest depth of web in compression. The stresses below are however based on the
built-up elastic stresses. The elastic modulus for the reinforcement is taken as equal to
that for structural steel, from clause 3.2(2).
The elastic section moduli for the gross cross-section are given in rows 1 and 3 of
Table 6.1. The extreme-bre stresses for the steel section are:
o
a.top
150,12.87 2600,25.85 112.2 N/mm
2
tension
o
a.bot
150,15.96 2600,18.91 146.9 N/mm
2
compression
Table 6.1. Section moduli for hogging bending of the cross-section of Fig. 6.5, in 10
6
mm
3
units, and
height of neutral axis above bottom of section
Section modulus
Height of
Top layer
of bars
Top of steel
section
Bottom of
steel section
neutral axis
(mm)
Gross steel section 12.87 15.96 547
Eective steel section 12.90 15.77 551
Gross composite section 18.47 25.85 18.91 707
Eective composite section 18.47 25.94 18.63 713
Primary shrinkage stresses are neglected because the deck slab is assumed to be cracked.
Using the stresses at the extreme bres of the web, from Table 5.2 of EN 1993-1-1, the
stress ratio is:
108,140.6 0.768
and the condition for a Class 3 web is:
c,t 42,0.67 0.33 42 0.825,0.67 0.253 83.1
Neglecting the widths of the llet welds, the actual c,t 1175,12.5 94, so the com-
posite section is in Class 4 for hogging bending. An eective section must therefore be
78
6.2.2. Resistance to vertical shear
Clause 6.2.2 is for beams without web encasement. The whole of the vertical shear is usually
assumed to be resisted by the steel section, as in previous codes for composite beams. This
enables the design rules of EN 1993-1-1 and EN 1993-1-5 to be used. The assumption can
be conservative where the slab is in compression. Even where it is in tension and cracked
in exure, consideration of equilibrium shows that the slab must make some contribution
to shear resistance, except where the reinforcement has yielded. For solid slabs, the eect
is signicant where the depth of the steel beam is only twice that of the slab,
66
but diminishes
as this ratio increases.
In composite plate girders with vertical stieners, the concrete slab can contribute to the
anchorage of a tension eld in the web,
67
but the shear connectors must then be designed for
vertical forces (clause 6.2.2.3(2)). The tension eld model used in EN1993-1-5 is discussed in
the Designers Guide to EN 1993-2.
4
Since the additional tension eld supported by the
anges must be anchored at both upper and lower surfaces of the web, the weaker ange
Clause 6.2.2
Clause 6.2.2.3(2)
derived for the web in accordance with clause 6.2.1.5(7) using the built-up stresses
calculated on the gross cross-section above.
From Table 4.1 of EN 1993-1-5:
k
o
7.81 6.29 9.78
2
18.4 for 0.768.
From clause 4.4(2) of EN 1993-1-5:
`
p

b,t
28.4
k
o
p
1175,12.5
28.4 0.825
18.4
p 0.935 0.673
so the reduction factor is:
,
`
p
0.055 3
`
2
p
0.935 0.055 3 0.768

0.935
2
0.929
b
eff
,
"
bb,1 617 mm. b
e1
0.4 617 247 mm. b
e2
0.6 617 370 mm
Including the hole, the depth of web in compression is b
eff
,,, which is 664 mm, so the
width of the hole is 664 617 47 mm. The stress ratio for the web now diers from that
for the gross section, but the eect of this on the properties of the net section can be
neglected. It is clear from clause 4.4(3) of EN 1993-1-5 that (and hence , and b
eff
)
need not be recalculated. The level of the elastic neutral axis for this net section is
found to be as shown in Table 6.1; consequently, the depth b
e2
is in fact 389 mm, not
370 mm. The new section moduli are given in rows 2 and 4 of Table 6.1.
The eective section is as shown in Fig. 6.5. The nal stresses are as follows:
o
a.top
150,12.90 2600,25.94 111.8 N/mm
2
tension < 345 N/mm
2
o
a.bot
150,15.77 2600,18.63 149.1 N/mm
2
compression < 345 N/mm
2
o
s.top
2600,18.47 140.8 N/mm
2
< 500,1.15 435 N/mm
2
The stress change caused by the small reduction in web area is negligible in this case.
Fromclause 6.2.1.4(6), M
el.Rd
is found by scaling up M
c.Ed
by a factor k until a stress limit
is reached. By inspection of the nal stresses, the bottom ange will probably govern. In
fact, it does, and:
150,15.77 2600,18.63k 345
whence k 2.40, and the elastic resistance moment is:
M
el.Rd
150 2.40 2600 6390 kNm (provided that M
a.Ed
150 kNm)
79
will govern the contribution of the anges, V
bf.Rd
, to shear resistance. Comment given later
on clause 6.2.2.5(1) is relevant here.
Bending and vertical shear beams in Class 1 or 2
Shear stress does not signicantly reduce bending resistance unless the shear is quite high.
For this reason, the interaction may be neglected until the shear force exceeds half the
shear resistance (clause 6.2.2.4(1)).
Both EN 1993-1-1 and EN 1994-2 use a parabolic interaction curve. Clause 6.2.2.4(2)
covers the case of Class 1 or 2 cross-sections where the reduction factor for the design
yield strength of the web is 1 ,, where:
, 2V
Ed
,V
Rd
1
2
6.5
and V
Rd
is the resistance in shear (which is either the plastic shear resistance or the shear
buckling resistance if lower). The interactions for Class 1 and 2 cross-sections with and
without shear buckling are shown in Fig. 6.6.
For a web where the shear buckling resistance is less than the plastic shear resistance and
M
Ed
< M
f.Rd
, the anges may make a contribution V
bf.Rd
to the shear resistance according to
EN 1993-1-5 clause 5.4(1). For moments exceeding M
f.Rd
(the plastic bending resistance
ignoring the web), this contribution is zero as at least one ange is fully utilized for
bending. V
Rd
is then equal to V
bw.Rd
. For moments less than M
f.Rd
, V
Rd
is equal to
V
bw.Rd
V
bf.Rd
.
This denition of V
Rd
leads to some inconsistency in clause 6.2.2.4(2) as the resistance in
bending produced therein can never be less than M
f.Rd
. Where there is shear buckling there-
fore, it is best to consider that the interaction with bending and shear according to clause
6.2.2.4(2) is valid for moments in excess of M
f.Rd
only. For lower moments, the interaction
with shear is covered entirely by the shear check to EN 1993-1-5 clause 5.4(1).
Where a Class 3 cross-section is treated as an equivalent Class 2 section and the design
yield strength of the web is reduced to allow for vertical shear, the eect on a section in
hogging bending is to increase the depth of web in compression. If the change is small, the
hole-in-web model can still be used. For a higher shear force, the new plastic neutral axis
may be within the top ange, and the hole-in-web method is inapplicable. The section
should then be treated as a Class 3 section.
Bending and vertical shear beams in Class 3 or 4
If the cross-section is either Class 3 or Class 4, then clause 6.2.2.4(3) applies and the inter-
action should be checked using EN1993-1-5 clause 7.1. This clause is similar to that for Class
1 and 2 sections but an interaction equation is provided. This allows the designer to neglect
the interaction between shear and bending moment when the design shear force is less than
50% of the shear buckling resistance based on the web contribution alone. Where the design
Clause 6.2.2.4(1)
Clause 6.2.2.4(2)
Clause 6.2.2.4(3)
M
pl,Rd
0.5V
pl,Rd
V
pl,Rd
0.5V
bw,Rd
V
bw,Rd
V
Ed
M
f,Rd
M
pl,Rd
M
f,Rd
M
Ed
V
Ed
M
Ed
Interaction to
clause 6.2.2.4(2)
Shear resistance
V
b,Rd
to EN 1993-1-5/5
(a) (b)
Fig. 6.6. Shearmoment interaction for Class 1 and 2 cross-sections (a) with shear buckling and
(b) without shear buckling
80
shear force exceeds this value and M
Ed
! M
f.Rd
, the condition to be satised is:
j
1
1
M
f.Rd
M
pl.Rd
!
2j
3
1
2
1.0 7.1 in EN 1993-1-5
where j
3
is the ratio V
Ed
,V
bw.Rd
and j
1
is a usage factor for bending, M
Ed
,M
pl.Rd
, based on
the plastic moment resistance of the section. M
f.Rd
is the design plastic bending resistance
based on a section comprising the anges only. The denition of M
f.Rd
is discussed under
clause 6.2.2.5(2) below.
For Class 4 sections, the calculation of M
f.Rd
and M
pl.Rd
must consider eective widths for
anges, allowing for plate buckling. M
pl.Rd
is however calculated using the gross web, regard-
less of any reduction that might be required for local buckling under direct stress. If axial
force is present, EN 1993-1-5 clause 7.1(4) requires appropriate reduction to be made to
M
f.Rd
and M
pl.Rd
. Discussion of axial force is given before Example 6.4.
The interaction for Class 3 and 4 beams is illustrated in Fig. 6.7. The full contribution to
the shear resistance from the web, V
bw.Rd
, is obtained at a moment of M
f.Rd
. For smaller
moments, the coexisting shear can increase further due to the ange shear contribution,
V
bf.Rd
, from clause 5.4 of EN 1993-1-5, provided that the web contribution is less than the
plastic resistance. The applied bending moment must additionally not exceed the elastic
bending resistance; that is, the accumulated stress must not exceed one of the limits in
clause 6.2.1.5(2). This truncates the interaction diagram in Fig. 6.7 at a moment of
M
el.Rd
. The moment must also not exceed that for lateraltorsional buckling.
The value of M
Ed
for use in the interaction with Class 3 and 4 cross-sections is not clearly
dened. Clause 6.2.2.4(3) states only that EN 1993-1-5 clause 7.1 is applicable using the
calculated stresses of the composite section. These stresses are dependent on the sequence
of construction and can include self-equilibrating stresses such as those from shrinkage
which contribute no net moment. There was no problem with interpretation in earlier
drafts as j
1
, the accumulated stress divided by the appropriate stress limit, was used in the
interaction rather than j
1
.
For compatibility with the use of M
pl.Rd
in the interaction expression (based on the cross-
section at the time considered) it is recommended here that M
Ed
is taken as the greatest value
of (o
i
W, where o
i
is the total accumulated stress at an extreme bre and W is the elastic
modulus of the eective section at the same bre at the time considered. This bending
moment, when applied to the cross-section at the time considered, produces stresses at the
extreme bres which are at least as great as those accumulated.
The reason for the use of plastic bending properties in the interaction for Class 3 and Class
4 beams needs some explanation. Test results on symmetric bare steel beams with Class 3 and
Class 4 webs
68
and also computer simulations on composite bridge beams with unequal
anges
69
showed very weak interaction with shear. The former physical tests showed
Interaction to
EC3-1-5/7.1
(1100)
(2900)
A
B
(2900), etc. Values
from Example 6.4
M
el,Rd
(6390)
M
pl,Rd
(8089)
V
bw,Rd
/2
V
bw,Rd
(1834)
V
Ed
M
f,Rd
(5568)
M
Ed
Shear resistance
V
b,Rd
to EC3-1-5/5
Fig. 6.7. Shearmoment interaction for Class 3 and 4 cross-sections to clause 7.1 of EN 1993-1-5
81
virtually no interaction at all and the latter typically showed some minor interaction only
after 80% of the shear resistance had been reached. The use of a plastic resistance
moment in the interaction helps to force this observed behaviour as seen in Fig. 6.7.
No distinction is made for beams with longitudinally stiened webs, which can have less
post-buckling strength when overall web panel buckling is critical. There are limited test
results for such beams and the approach leads to an interaction with shear only at very
high percentages of the web shear resistance. A safe option is to replace j
1
by j
1
in the
interaction expression. For composite beams with longitudinally stiened webs, j
1
can be
interpreted as the usage factor based on accumulated stress and the stress limits in clause
6.2.1.5(2).
Various theories for post-critical behaviour in shear of webs in Class 3 or 4 under com-
bined bending and vertical shear have been compared with 22 test results from composite
beams.
69
It was found that the method of EN 1993-1-5 gives good predictions for web
panels of width/depth ratio exceeding 1.5, and is conservative for shorter panels.
Checks of bare steel anges of box girders are covered in the Designers Guide to EN 1993-2.
4
For open steel boxes, clause 7.1(5) of EN 1993-1-5 clearly does not apply to the reinforced
concrete top ange. For composite anges, this clause should be applied to the steel part of
the composite ange, but the eective area of the steel part may be taken as the gross area
(reduced for shear lag if applicable) for all loads applied after the concrete ange has been
cast, provided that the shear connectors are spaced in accordance with Table 9.1. Shear
buckling need not be considered in the calculation of " jj
3
. Since most continuous box-girder
bridges will be in Class 3 or 4 at supports, the restriction to elastic bending resistance forced
by clause 7.1(5) of EN 1993-1-5 should not be unduly conservative. The use of elastic analysis
also facilitates addition of any distortional warping and transverse distortional bending
stresses developed.
Bending and vertical shear all Classes
Clause 6.2.2.4(4) conrms that when the depth of web in compression is increased to allow
for shear, the resulting change in the plastic neutral axis should be ignored when classifying
the web. The reduction of steel strength to represent the eect on bending resistance of shear
is only a model to match test results. To add the sophistication of reclassifying the cross-
section would be an unjustied complexity. The scatter of data for section classication
further makes reclassication unjustied. The issue of reclassication does not arise when
using EN 1993-1-5 clause 7.1 as the interaction with shear is given by an interaction expres-
sion. The movement of the neutral axis is never determined.
Clause 6.2.2.5(1) refers to the contribution of anges to the resistance of the web to buck-
ling in shear. It permits the contribution of the ange in EN 1993-1-5 clause 5.4(1) to be
based on the bare steel ange even if it has the larger plastic moment resistance. It implies
that where this is done, the weaker ange is being assisted by the concrete slab in anchoring
the tension eld. From clause 6.2.2.3(2), the shear connection should then be designed for
the relevant vertical force. This additional check can be avoided by neglecting the concrete
contribution in calculating V
bf.Rd
.
The plastic bending resistance of the anges, M
f.Rd
, is dened in clause 6.2.2.5(2) for
composite sections as the design plastic resistance of the eective section excluding the
steel web. This implies a plastic neutral axis within the stronger ange (usually the composite
one). Clause 7.1(3) of EN 1993-1-5 allows M
f.Rd
to be taken as the product of the strength of
the weaker ange and the distance between the centroids of the anges. This gives a slightly
lower result for a composite beamthan application of the rule in EN1994-2. The denition in
EN 1994-2 is in fact also used in EN 1993-1-5, clauses 5.4(2) and 7.1(1).
It is stated in clause 7.1(1) of EN 1993-1-5 that the interaction expression for bending and
shear is valid only where j
1
! M
f.Rd
,M
pl.Rd
. From the denition of j
1
, this condition is
M
Ed
! M
f.Rd
. Where it is not satised (as in Example 6.5), the bending moment M
Ed
can
be resisted entirely by the anges of the section. The web is not involved, so there is no
interaction between bending and shear unless the shear resistance is to be enhanced by the
ange contribution in EN 1993-1-5, clause 5.4. In such cases, the check on interaction
Clause 6.2.2.4(4)
Clause 6.2.2.5(1)
Clause 6.2.2.5(2)
82
between bending and shear is eectively carried out using that clause as illustrated in Figs 6.6
and 6.7. No such condition is stated for j
1
, so it should not be applied when j
1
is replaced by
j
1
when required by clause 7.1(5) of EN 1993-1-5.
Eect of compressive axial force
Clause 6.2.2.5(1) makes clear that axial force means a force N
Ed
acting on the composite
cross-section, or an axial force N
a.Ed
applied to the steel element before the member
becomes composite. It is not the axial force in the steel element that contributes to the
bending resistance of a composite beam.
For Class 1 or 2 cross-sections, the resistance to bending, shear and axial force should be
determined by rst reducing the design yield strength of the web in accordance with clause
6.2.2.4(2) and then checking the resulting cross-section under bending and axial force.
For Class 3 or 4 cross-sections, clause 7.1(4) of EN 1993-1-5 is also relevant. This
eectively requires the plastic bending resistance M
pl.Rd
in the interaction expression of
EN 1993-1-5 clause 7.1(1) to be reduced to M
pl.N.Rd
(using the notation of clause 6.7.3.6)
where axial force is present. The resistance M
f.Rd
should be reduced by the factor in clause
5.4(2) of EN 1993-1-5, which is as follows:
1
N
Ed
A
f1
A
f2
f
yf
,
M0
!
5.9 in EN 1993-1-5
This is written for bare steel beams and A
f1
and A
f2
are the areas of the steel anges. These are
assumed here to be resisting the whole of the force N
Ed
, presumably because in this tension-
eld model, the web is fully used already.
For composite beams in hogging zones, equation (5.9) above could be replaced by:
1
N
Ed
A
f1
A
f2
f
yf
,
M0
A
s
f
sd
!
D6.2
where A
s
is the area of the longitudinal reinforcement in the top slab.
For sagging bending, the shear force is unlikely to be high enough to reduce the resistance
to axial force and bending. On the assumption that the axial force is applied only to the
composite section, the value of M
Ed
to use in the interaction expression can be derived
from the accumulated stresses as suggested above for checking combined bending and
shear, but the uniform stress component from the axial force should not be considered in
calculating o
i
. If the axial force, determined as acting at the elastic centroid of the compo-
site section, acts at another level in the model used for the resistance, the moment arising
from this change in its line of action should be included in M
Ed
. This is illustrated in
Example 6.5.
Clause 7.1(4) and (5) of EN1993-1-5 requires that where axial force is present such that the
whole web is in compression, M
f.Rd
should be taken as zero in the interaction expression, and
j
1
should be replaced by j
1
(which is dened in EN 1993-1-5 clause 4.6). This leads to the
interaction diagram shown in Fig. 6.8. The limit j
1
! M
f.Rd
,M
pl.Rd
for validity of expression
1.0
1.0
0.5
1
Fig. 6.8. Shearmoment interaction for Class 3 and 4 cross-sections with webs fully in compression
83
(7.1) given in EN 1993-1-5 clause 7.1(1) is not applicable in this case. The expression is
applicable where j
1
! 0.
The application of this requirement is unclear for beams built in stages. These could have
axial load applied separately to the bare steel section and to the composite section. A safe
interpretation, given the relatively small amount of testing on asymmetric sections, would
be to take M
f.Rd
as zero wherever the whole web is in compression under the built-up stresses.
For composite bridges, j
1
can be interpreted as the usage factor based on accumulated stress
and the stress limits in clause 6.2.1.5(2). However, this is likely to be conservative at high
shear, given the weak interaction between bending and shear found in the tests on composite
beams discussed above.
Vertical shear in a concrete ange
Clause 6.2.2.5(3) gives the resistance to vertical shear in a concrete ange of a composite
beam (represented here by a design shear strength, v
Rd.c
) by reference to clause 6.2.2 of
EN 1992-2. That clause is intended mainly to enable higher shear strengths to be used in
the presence of in-plane prestress. A Note in EN 1992-2 recommends values for its three
nationally determined parameters (NDPs). Where the ange is in tension, as in a continuous
composite beam, the reduced strengths obtained can be over-conservative. In EN1994-2, the
Note recommends dierent NDPs, based on recent research.
70
With these values, and for
eective slab depths d of at least 200 mm and
C
1.5, the rules are:
v
Rd.c
0.10 1
200
d
r
!
100, f
ck
1,3
0.12o
cp
a
and
v
Rd.c
! 0.035 1
200
d
r
!
3,2
f
1,2
ck
0.12o
cp
b
where: , A
s
,bd 0.02
o
cp
N
Ed
,A
c
< 0.2f
cd
(compression positive) c
and N
Ed
is the in-plane axial force (negative if tensile) in the slab of breadth b and with tensile
reinforcement A
s
, and f
ck
is in N/mm
2
units.
It can be inferred from Fig. 6.3 in EN 1992-2 that A
s
is the reinforcement in tension under
the loading which produces the shear force considered (a wording that is used in clause 5.3.3.2
of BS 5400-4). Thus, for shear from a wheel load, only one layer of reinforcement (top or
bottom, as appropriate) is relevant, even though both layers may be resisting global tension.
It thus appears fromequation (a) that the shear strength depends on the tensile force in the
slab. This awkward interaction is usually avoided, because EN 1994-2 gives a further
research-based recommendation, that where o
cp
is tensile, it should not be taken as
greater than 1.85 N/mm
2
. The eect of this is now illustrated, with d 200 mm.
Let the reinforcement ratios be ,
1
0.010 for the tensile reinforcement, ,
2
0.005 for
the other layer, with f
ck
40 N/mm
2
, and o
cp
1.85 N/mm
2
. From equation (a) above:
v
Rd.c
0.1 21.0 40
1,3
0.12 1.85 0.68 0.22 0.46 N,mm
2
and equation (b) does not govern. Fromequation (c) with values of o
cp
, N
Ed
and f
s
all negative:
o
cp
N
Ed
,A
c
f
s
A
s
,A
c
f
s
0.01 0.005bd,bh 0.015f
s
d,h
where the summation is for both layers of reinforcement, because o
cp
is the mean tensile
stress in the slab if uncracked and unreinforced.
For h 250 mm, the stress o
cp
then reaches 1.85 N/mm
2
when the mean tensile stress in
the reinforcement is 154 N/mm
2
, which is a low value in practice. At higher values, v
Rd.c
is
independent of the tensile force in the slab, though the resulting shear strength is usually
lower than that from BS 5400-4.
In the transverse direction, N
Ed
is zero unless there is composite action in both directions,
so for checking punching shear, two dierent shear strengths may be relevant.
Clause 6.2.2.5(3)
84
Example 6.4: resistance of a Class 4 section to hogging bending and vertical
shear
The cross-section in Example 6.3 (Fig. 6.5) is checked for resistance to a vertical shear
force of 1100 kN, combined with bending moments M
a.Ed
150 kNm and
M
c.Ed
2600 kNm
From clause 6.2.6(6) of EN 1993-1-1, resistance to shear buckling should be checked if:
h
w
,t
w
72,j
where j is a factor for which a Note to clause 5.1.2 of EN 1993-1-5 recommends the value
1.2. For S355 steel and a 12.5 mm thick web plate, 0.81, so:
h
w
j,t
w
1175 1.2,12.5 0.81 139 72
The resistance of this unstiened web to shear buckling is found using clauses 5.2 and
5.3 of EN 1993-1-5. The transverse stieners provided at the cross-bracings are conserva-
tively ignored, as is the contribution from the anges. For stieners at supports only, the
slenderness is obtained from EN 1993-1-5 equation (5.5):
`
w

h
w
86.4t
1175
86.4 12.5 0.81
1.343
Away from an end support, the column rigid end post in Table 5.1 of EN 1993-1-5
applies, so:
w

1.37
0.7 `
w
1.37
0.7 1.343
0.67
From EN 1993-1-5 equation (5.2):
V
bw.Rd

w
f
yw
h
w
t
3
p

M1
0.67 355 1175 12.5
3
p
1.1
1834 kN
The shear ratio j
3
1100,1834 0.60. This exceeds 0.5 so interaction with the hogging
bending moment must be considered using EN 1993-1-5, clause 7.1. The resistances M
f.Rd
and M
pl.Rd
are rst determined for the composite section. Clause 6.2.2.5(2) requires M
f.Rd
to be calculated for the composite section neglecting the web. M
pl.Rd
is calculated using the
gross web, regardless of the reduction for local buckling under direct stress. From section
analysis, using f
yd
345 N/mm
2
throughout:
M
f.Rd
5568 kNm
M
pl.Rd
8089 kNm
The applied bending moment M
Ed
is taken as the greatest value of (o
i
)W, as explained
earlier. Using stresses from Example 6.3 and section moduli from Table 6.1, the values of
M
Ed
for the extreme bres of the steel beam are as follows:
top ange: M
Ed
111.8 25.94 2900 kNm, which governs
bottom ange: M
Ed
149.1 18.63 2778 kNm
The bending ratio j
1
2900,8089 0.359.
This is less than the ratio M
f.Rd
,M
pl.Rd
, which is 0.69, so there is no interaction between
bending and shear.
To illustrate the use of interaction expression (7.1) of EN 1993-1-5, let us assume that
M
Ed
is increased, such that j
1
0.75, with j
3
0.60 as before. Then:
j
1
1
M
f.Rd
M
pl.Rd
!
2j
3
1
2
0.75 1
5568
8089
!
2 0.60 1
2
0.762 D6.3
The original action eects are shown as point A in Fig. 6.7. This lies below point B,
showing that in this case, the vertical shear does not reduce the resistance to bending.
85
The check of j
1
also required by the reference in clause 7.1(1) of EN 1993-1-5 to its
clause 4.6 is covered in Example 6.3. The ratio j
1
is the greatest usage based on accumu-
lated stress, which is 149.5,345 0.43 at the bottom ange. If the momentshear interac-
tion above is checked conservatively using j
1
, equation (D6.3) becomes:
M
Ed
M
el.Rd
1
5568
8089
!
2 0.60 1
2
1. or M
Ed
0.988M
el.Rd
thus giving a slight reduction to the bending resistance. This reduction will always occur
when " jj
3
0.5 is used in this interaction expression as can be seen from Fig. 6.8.
Example 6.5: addition of axial compression to a Class 4 cross-section
The eect of adding an axial compression N
Ed
4.8 MN to the composite cross-section
studied in Examples 6.3 and 6.4 is now calculated, using the method explained earlier.
It is assumed that in the global analysis, the beam was located at the level of the
long-term elastic neutral axis of the uncracked unreinforced section at mid-span, with
n
L
23.7. This is 951 mm (denoted z
elu
) above the bottom of the 1500 mm deep
section. Where the neutral axis of the cross-section being veried is at some lower level,
z, for example, with N
Ed
acting at that level, the coexisting hogging bending moment
should be reduced by N
Ed
(z
elu
z), to allow for the change in the level of N
Ed
.
The composite section, shown in Fig. 6.5, is also subjected, as before, to hogging
bending moments M
a.Ed
150 kNm and M
c.Ed
2600 kNm, and to vertical shear
V
Ed
1100 kN. The concrete slab is assumed to be fully cracked.
It was found in Example 6.4 that for the vertical shear, " jj
3
0.60, so the interaction
factor (2" jj
3
1
2
is only 0.04. The rst check is therefore on the elastic resistance of the
net section to N
Ed
plus M
Ed
.
If N
Ed
is assumed to act at the centroid of the net elastic Class 4 section, allowing for the
hole, its line of action, and hence M
Ed
, change at each iteration. It will be found that the
whole of the eective web is in compression, so that the interaction with shear should be
based on j
1
(accumulated stresses), not on " jj
1
(action eects and resistances). This enables
stresses from N
Ed
and M
Ed
to be added, and a non-iterative method to be used, based on
separate eective cross-sections for axial force and for bending. This method can always
be used as a conservative approach.
Stresses from axial compression
For the gross cracked cross-section, A 43 000 mm
2
, so:
o
a
4800,43.0 112 N/mm
2
compression
From Table 4.1 in EN 1993-1-5, 1 and k
o
4.0
From Example 6.3, for the web,
"
bb,t c,t 94
Assuming f
y
345 N/mm
2
, as in Example 6.1, then 0.825.
From EN 1993-1-5 clause 4.4(2) and Table 4.1:
`
p

b,t
28.4
k
o
p
94
28.4 0.825
4.0
p 2.01
,
`
p
0.055 3
`
2
p
2.01 0.055 3 1
2.01
2
0.444
b
eff
,
"
bb 0.444 1175 522 mm. so b
e1
b
e2
261 mm
The depth of the hole in the web is 1175 522 653 mm.
86
The net cross-section for axial compression is shown in Fig. 6.9(a). Its net area is:
A
eff.N
43 000 653 12.5 34 840 mm
2
The compressive stress from N
Ed
is o
a.N
4800,34.84 138 N/mm
2
.
The elastic neutral axis of the net section is 729 mm above the bottom, so the change in
neutral axis is z 951 729 222 mm
2
, and N
Ed
z 1066 kNm.
261
653
261
729
Elastic NA Elastic NA
109
98
247
396
12.5
493
712
39
(a) (b)
400 20
400 20
400 30
400 30
12.5
4155 mm
2
4155 mm
2
Fig. 6.9. Net cross-sections for (a) axial compression and (b) hogging bending
Stresses from bending moment
In Example 6.4, M
c.Ed
2600 kNm, hogging. The line of action of N
Ed
has been moved
downwards, so N
Ed
z is sagging, and now M
c.Ed
2600 1066 1534 kNm, with
M
a.Ed
150 kNm, as before.
Using section moduli for the gross cross-section from Example 6.3, the stresses at the
edge of the web are:
o
a.top
68.4 N/mm
2
tension, o
a.bot
86.5 N/mm
2
compression
Hence,
68.4,86.5 0.790
Proceeding as in Example 6.3, the results are as follows:
k
o
18.9,
"
``
p
0.923, , 0.941, b
eff
618 mm, b
e1
247 mm, b
e2
371 mm
The hole in the web is 39 mm deep. The eective cross-section is shown in Fig. 6.9(b).
The extreme-bre stresses from M
a.Ed
, M
c.Ed
and N
Ed
respectively are:
o
a.top
11.6 59.1 138 67 N,mm
2
o
a.bot
9.5 82.4 138 230 N,mm
2
It follows that there is some compression in the deck slab, which has been neglected, for
simplicity. The reinforcement is assumed here to carry all the compressive force in the slab.
Interaction with vertical shear
The whole of the web is in compression, so from clause 7.1(5) of EN 1993-1-5, M
f.Rd
0
and j
1
, not " jj
1
, is used. For the steel bottom ange, which governs,
j
1
230,345 0.67
From equation (7.1) of EN 1993-1-5, with " jj
3
= 0.60 and M
f.Rd
0,
j
1
2" jj
3
1
2
0.67 0.04 0.71
This is less than 1.0, so the cross-section is veried.
87
Method when the web is partly in tension
This example is now repeated with the axial compression reduced to N
Ed
2.5 MN and
all other data as before, to illustrate the method where " jj
1
is used and M
f.Rd
is not zero.
From clause 7.1(4) of EN 1993-1-5, M
pl.Rd
is reduced to allow for N
Ed
, as follows. For
the gross cross-section M
pl.Rd
8089 kNm, and the plastic neutral axis is 876 mm above
the bottom. Force N
Ed
is assumed to act at its level. The depth of web needed to resist it is
h
w.N
2500,12.5 0.345 580 mm. This depth is centred on the neutral axis as shown
in Fig. 6.10 and remains wholly within the web. Its contribution to M
pl.Rd
was:
f
yd
th
2
w.N
,4 345 12.5 0.58
2
,4 363 kNm
As its depth is centred on the plastic neutral axis for bending alone,
M
pl.N.Rd
8089 363 7726 kNm
12.5
556
290
0.17
Plastic
neutral
axis
2.40
N
Ed
4.14
1175 2.50
Forces in MN units
290
250
45
1.805
2.76
1.805
Tension
Compression
39
400 20
400 30
Fig. 6.10. Plastic resistance of Class 4 section to hogging bending and axial compression
From Example 6.4, M
f.Rd
5568 kNm. From equation (D6.2) and using forces from
Fig. 6.10, the reduction factor that allows for N
Ed
is:
1
N
Ed
A
f1
A
f2
f
yf
,
M0
A
s
f
sd
!
1
2.50
2.76 4.14 2 1.805
0.762
Hence,
M
f.N.Rd
0.762 5568 4243 kNm
The bending moment M
Ed
is found from accumulated stresses, using the Class 4 cross-
section for bending. Its elastic properties should strictly be determined with a value for
M
Ed
such that the line of action of N
Ed
is at the neutral axis of the eective cross-
section. This requires iteration. Finally, the line of action of N
Ed
should be moved to
the neutral axis of the plastic section used for the calculation of M
pl.Rd
. This requires a
further correction to M
Ed
.
The simpler and suciently accurate method used here is to nd the Class 4 section
properties ignoring any moment from the axial force, to use these to scale up M
a.Ed
,
and then to add the correction from N
Ed
. The change in the line of action of N
Ed
is:
951 876 75 mm, so M
c.Ed
2600 2500 0.075 2413 kNm
Calculation similar to that in the section Stresses from bending moment, above, nds the
depth of the hole in the web to be 322 mm. The top ange is found to govern, and
M
Ed
M
a.Ed
W
c.top
,W
a.top
M
c.Ed
150 26.6,13.0 2413 2720 kNm
This bending moment is less than M
f.N.Rd
, so it can be resisted entirely by the anges, and
there is no need to consider interaction with shear in accordance with clause 7.1(1) of
EN 1993-1-5.
88
6.3. Filler beam decks
6.3.1. Scope
The encasement of steel bridge beams in concrete provides several advantages for design:
.
It enables a Class 3 web to be upgraded to Class 2, and the slenderness limit for a Class 2
compression ange to be increased by 40% (clause 5.5.3).
.
It prevents lateraltorsional buckling.
.
It prevents shear buckling (clause 6.3.4(1)).
.
It greatly increases the resistance of the bridge deck to vehicular impact or terrorist attack.
These design advantages may not however lead to the most economic solution. The use of
longitudinal ller beams in new construction is not common at present.
There are a great number of geometric, material and workmanship-related restrictions
given in clauses 6.3.1(1) to (4) which have to be met in order to use the application rules
for the design of ller beams. These are necessary because the rules derive mainly from
existing practice in the UK and from clause 8 of BS 5400:Part 5.
11
No explicit check of
the shear connection between steel beams and concrete (provided by friction and bond
only) is required.
Clause 6.3.1(1) excludes fully-encased ller beams from the scope of clause 6.3. This is
because there are no widely-accepted design rules for longitudinal shear in fully-encased
beams without shear connectors.
Clause 6.3.1(2) requires the beams to be of uniform cross-section and to have a web depth
and ange width within the ranges found for rolled H- or I-sections. This is due to the lack of
existing examples of ller beams with cross-sections other than these. There is no require-
ment for the beams to be H- or I-sections, but hollow sections would be outside the scope
of clause 6.3.
Clause 6.3.1(3) permits spans to be either simply supported or continuous with square or
skew supports. This clarication is based on existing practice, and takes account of the many
other restrictions.
Clause 6.3.1(4) contains the majority of the restrictions which relate mainly to ensuring
the adequacy of the bond between steel beam and concrete, as follows.
.
Steel beams should not be curved in plan. This is because the torsion produced would
lead to additional bond stresses between the structural steel and concrete, for which no
application rules are available.
.
The deck skew should not exceed 308. This limits the magnitude of torsional moments,
which can become large with high skew.
.
The nominal depth, h, of the beam should lie between 210 mm and 1100 mm. This is
because anything less than 210 mm should be treated as reinforced concrete, and there
could in future be rolled sections deeper than 1100 mm.
.
A maximum spacing of the steel beams is set: the lesser of h,3 600 mm and 750 mm.
This reects existing practice and limits the longitudinal shear ow (and bond stresses)
between the concrete and the steel beam.
.
The minimum concrete cover to the top of the steel beams is restricted to 70 mm. Alarger
value may however be necessary to provide adequate cover to the reinforcement. The
Clause 6.3.1(1)
Clause 6.3.1(2)
Clause 6.3.1(3)
Clause 6.3.1(4)
Stability of the span in the vertical plane
Buckling of the whole span in the vertical plane is possible. Its elastic critical axial force is
nowestimated, treating it as pin-ended, and assuming sagging bending. The modular ratio
n
L
18.8 is used, intermediate between the values for short- and long-term loading. The
presence of a stier cross-section near the supports is ignored.
The result is N
cr
47 MN. The ratio c
cr
N
cr
,N
Ed
8.1 (<10), so from clause
5.2.1(3), second-order eects should be included in M
Ed
. There appears to be a sucient
margin for these, but this has not been checked.
89
maximum cover is limited to the lesser of 150 mm and h/3, based on existing practice and
to limit the longitudinal shear stress developed.
A further restriction is given such that the plastic neutral axis for sagging bending
remains below the level of the bottom of the top ange, since cracking of the concrete
in the vicinity of the top ange could reduce the bond stress developed. This rule could
only govern where the steel beams were unusually small. The side cover to the top
ange should be at least 80 mm.
.
The clear distance between top anges should not be less than 150 mm so that the
concrete can be adequately compacted. This is essential to ensure that the required
bond to the steel is obtained.
.
Bottom transverse reinforcement should be provided (through holes in the beam webs)
such that transverse moments developed can be carried. A minimum bar size and
maximum spacing are specied. Minimum reinforcement, here and elsewhere, should
also satisfy the requirements of EN 1992.
.
Normal-density concrete should be used. This is because there is little experience of
ller-beam construction with concrete other than normal-density, where the bond
characteristics could be aected.
.
The ange should be de-scaled. This again is to ensure good bond between the concrete
and the steel beam.
.
For road and railway bridges the holes in steel webs should be drilled. This is discussed
under clause 6.3.2(2).
6.3.2. General
Clause 6.3.2(1) refers to other clauses for the cross-section checks, which should be
conducted at ultimate and serviceability limit states. These references do not require a
check of torsion as discussed below.
Clause 6.3.2(2) requires beams with bolted connections or welding to be checked against
fatigue. The implication is that ller beams without these need not be checked for fatigue,
even though they will contain stress-raising holes through which the transverse reinforce-
ment passes. For road and railway bridges, where fatigue loading is signicant, clause
6.3.1(4) requires that all holes in webs are drilled (rather than punched), which improves
the fatigue category of the detail.
Clause 6.3.2(3) is a reminder to refer to the relaxations for cross-section Class in clause
5.5.3.
Mechanical shear connection need not be provided for ller beams (clause 6.3.2(4)). This
reliance on bond improves the relative economy of ller-beam construction but leads to
many of the restrictions noted above under clause 6.3.1.
6.3.3. Bending moments
The resistance of cross-sections to bending, clause 6.3.3(1), is determined in the same way as
for uncased sections of the same Class, with Class determined in accordance with clause
5.5.3. The relaxations in clause 5.5.3 should generally ensure that beams can be designed
plastically and thus imposed deformations generally need not be considered at ultimate
limit states (the comments made under clause 5.4.2.2(6) refer).
Lateraltorsional buckling is not mentioned in clause 6.3.3(1) because a ller-beam deck
is inherently stable against lateraltorsional buckling in its completed state due to its large
transverse stiness. The steel beams are likely to be susceptible during construction and
the title of clause 6.3.5 provides a warning.
For the inuence of vertical shear on resistance to bending, reference is made to the rules
for uncased beams. The shear resistance of ller-beam decks is high, so interaction is
unlikely, but it should be checked for continuous spans.
In the transverse direction, a ller-beam deck behaves as a reinforced concrete slab. Clause
6.3.3(2) therefore makes reference to EN 1992-2 for the bending resistance in the transverse
Clause 6.3.2(1)
Clause 6.3.2(2)
Clause 6.3.2(3)
Clause 6.3.2(4)
Clause 6.3.3(1)
Clause 6.3.3(2)
90
direction. A Note to clause 9.1(103) of EN 1992-2 makes minimum reinforcement a
nationally determined parameter.
No requirement is given for a check on torsion, which will be produced to some degree in
both longitudinal and transverse directions of the global analysis models allowed by clause
5.4.2.9(3). Neglect of torsion is justied by the limits imposed on geometry in clause
6.3.1(4), particularly the limit on skew angle, and by current UK practice.
6.3.4. Vertical shear
The simplest calculation of shear resistance involves basing the resistance on that of the steel
beam alone. Clause 6.3.4(1) indicates that this resistance can be calculated using the plastic
shear resistance and so ignoring shear buckling. The clause does permit a contribution from
the concrete to be taken. Clauses 6.3.4(2) and 6.3.4(3), respectively, cover a method of
determining the shear force that may be carried on the reinforced concrete section and the
determination of the resistance of this concrete section. Clause 6.3.4(3) applies also to
shear resistance in the transverse direction.
6.3.5. Resistance and stability of steel beams during execution
Clause 6.3.5(1) refers to EN 1993-1-1 and EN 1993-2 for the check of the bare steel beams.
This covers both cross-section resistance and lateraltorsional buckling. The latter is an
important consideration prior to hardening of the concrete.
6.4. Lateraltorsional buckling of composite beams
6.4.1. General
It is assumed in this section that in completed bridges, the steel top anges of all composite
beams will be stabilized laterally by connection to a concrete or composite slab (clause
6.4.1(1)). The rules on maximum spacing of connectors in clause 6.6.5.5(1) and (2)
relate to the classication of the top ange, and thus only to local buckling. For lateral
torsional buckling, the relevant rule, given in clause 6.6.5.5(3), is less restrictive.
Any steel top ange in compression that is not so stabilized should be checked for lateral
buckling (clause 6.4.1(2)) using clause 6.3.2 of EN 1993-1-1 to determine the reduction
factor for buckling. For completed bridges, this applies to the bottom ange adjacent to
intermediate supports in continuous construction. In a composite beam, the concrete slab
provides lateral restraint to the steel member, and also restrains its rotation about a longitu-
dinal axis. Lateral buckling is always associated with distortion (change of shape) of the
cross-section (Fig. 6.11(b)). This is not true lateraltorsional buckling and is often referred
to as distortional lateral buckling. This form of buckling is covered by clauses 6.4.2 and
6.4.3. The general method of clause 6.4.2, based on the use of a computed value of the
elastic critical moment M
cr
, is applicable, but no detailed guidance on the calculation of
M
cr
is given in either EN 1993-1-1 or EN 1994-2.
For completed bridges, the bottom ange may be in compression over most of a span
when that span is relatively short and lightly loaded and adjacent spans are fully loaded.
Bottom anges in compression should always be restrained laterally at supports. It should
not be assumed that a point of contraexure is equivalent to a lateral restraint.
Design methods for composite beams must take account of the bending of the web,
Fig. 6.11(b). They dier in detail from the method of clause 6.3.2 of EN 1993-1-1, but the
same imperfection factors and buckling curves are used, in the absence of any better-
established alternatives.
The reference in clause 6.4.1(3) to EN 1993-1-1 provides a general method for use where
the method in clause 6.4.2 is inapplicable (e.g. for a Class 4 beam). Clause 6.4.3 makes a
similar reference but adds a reference to a further method available in clause 6.3.4.2 of
EN 1993-2. During unpropped construction, prior to the presence of a hardened deck
slab, the buckling verication can be more complicated and often involves overall buckling
Clause 6.3.4(1)
Clause 6.3.4(2)
Clause 6.3.4(3)
Clause 6.3.5(1)
Clause 6.4.1(1)
Clause 6.4.1(2)
Clause 6.4.1(3)
91
of a braced pair of beams. This situation is discussed further at the end of section 6.4.3.2 of
this guide.
6.4.2. Beams in bridges with uniform cross-sections in Class 1, 2 and 3
This general method of design is written with distortional buckling of bottom anges in
mind. It would not apply, for example, to a mid-span cross-section of a beam with the
slab at bottom-ange level (Fig. 6.12). The reference to uniform cross-section in the title
of the clause is not intended to exclude minor changes such as reinforcement details and
eects of cracking of concrete. The method cannot be used for Class 4 cross-sections,
which is a signicant limitation for larger bridges, in which case the methods of clause
6.4.3 should be used. The latter methods are more general.
The method is based closely on clause 6.3.2 of EN1993-1-1. There is correspondence in the
denitions of the reduction factor
LT
, clause 6.4.2(1), and the relative slenderness,
"
``
LT
,
clause 6.4.2(4). The reduction factor is applied to the design resistance moment M
Rd
,
which is dened in clauses 6.4.2(2) and (3). Expressions for M
Rd
are given by references
to clause 6.2. It should be noted that these include the design yield strength f
yd
which
should, in this case, be calculated using
M1
rather than
M0
because this is a check of
instability. If the beam is found not to be susceptible to lateraltorsional buckling (i.e.
LT
1.0), it would be reasonable to replace
M1
with
M0
.
The determination of M
Rd
for a Class 3 section diers from that of M
el.Rd
in clause
6.2.1.4(6) only in that the limiting stress f
cd
for concrete in compression need not be consid-
ered. It is necessary to take account of the method of construction.
The buckling resistance moment M
b.Rd
given by equation (6.6) must exceed the highest
applied moment M
Ed
within the unbraced length of compression ange considered.
Clause 6.4.2(1)
Clause 6.4.2(2)
Clause 6.4.2(3)
(b)
(a)
h
F F
0
Fig. 6.11. (a) U-frame action and (b) distortional lateral buckling
Fig. 6.12. Example of a composite beam with the slab in tension at mid-span
92
Lateral buckling for a Class 3 cross-section with unpropped construction
The inuence of method of construction on verication of a Class 3 composite section for
lateral buckling is as follows. From equation (6.4),
M
Rd
M
el.Rd
M
a.Ed
kM
c.Ed
(a)
where subscript c is used for the action eect on the composite member.
From equation (6.6), the verication is:
M
Ed
M
a.Ed
M
c.Ed

LT
M
el.Rd
(b)
which is:
LT
! M
a.Ed
M
c.Ed
,M
el.Rd
M
Ed
,M
el.Rd
(c)
The total hogging bending moment M
Ed
may be almost independent of the method of
construction. However, the stress limit that determines M
el.Rd
may be dierent for propped
and unpropped construction. If it is bottom-ange compression in both cases, then M
el.Rd
is
lower for unpropped construction, and the limit on
LT
from equation (c) is more severe.
Elastic critical buckling moment
Clause 6.4.2(4) requires the determination of the elastic critical buckling moment, taking
account of the relevant restraints, so their stinesses have to be calculated. The lateral
restraint from the slab can usually be assumed to be rigid. Where the structure is such
that a pair of steel beams and a concrete ange attached to them can be modelled as an
inverted-U frame (clause 6.4.2(5) and Fig. 6.10), continuous along the span, the rotational
restraining stiness at top-ange level, k
s
, can be found from clause 6.4.2(6). In the deni-
tion of stiness k
s
, exibility arises from two sources:
.
bending of the slab, which may not be negligible: 1/k
1
from equation (6.9)
.
bending of the steel web, which predominates: 1/k
2
from equation (6.10).
A third source of exibility is potentially the shear connection but it has been found
71
that
this can be neglected providing the requirements of clause 6.4.2(5) are met.
There is a similar discrete U-frame concept, which appears to be relevant to composite
beams where the steel sections have vertical web stieners. The shear connectors closest to
those stieners would then have to transmit almost the whole of the bending moment Fh
(Fig. 6.11(a)), where F is now a force on a discrete U-frame. The exibility of the shear
connection may then not be negligible, nor is it certain that the shear connection and the
adjacent slab would be suciently strong.
72
Where stieners are present, the resistance of
the connection above each stiener to repeated transverse bending should be established,
as there is a risk of local shear failure within the slab. There is at present no simple
method of verication. This is the reason for the condition that the web should be unstiened
in clause 6.4.2(5)(b). The restriction need not apply if bracings (exible or rigid) are
attached to the stieners, but in this case the model referred to in clause 6.4.3.2 would be
used.
Clause 6.4.2(7) allows the St Venant torsional stiness to be included in the calculation.
This is often neglected in lateraltorsional buckling models based on buckling of the bottom
chord, such as that provided in EN 1993-2 clause 6.3.4.2.
No formula is provided for the elastic critical buckling moment for the U-frame model
described above. M
cr
could be determined from a nite-element model of the beam with a
lateral and torsional restraint as set out above. Alternatively, textbook solutions could be
used. One such method was given in Annex B of ENV1994-1-1
20
and is nowin the Designers
Guide to EN 1994-1-1.
5
6.4.3. General methods for buckling of members and frames
6.4.3.1. General method
Reference is made to EN 1993-2 clause 6.3.4 where the method of clause 6.4.2 for beams or
the non-linear method of clause 6.7 for columns does not apply.
Clause 6.4.2(4)
Clause 6.4.2(5)
Clause 6.4.2(6)
Clause 6.4.2(7)
93
EN 1993-1-1 clause 6.3.4 gives a general method of evaluating the combined eect of
axial load and mono-axial bending applied in the plane of the structure, without use of an
interaction expression. The method is valid for asymmetric and non-uniform members
and also for entire plane frames. In principle, this method is more realistic since the structure
or member does, in reality, buckle in a single mode with a single system slenderness.
Interaction formulae assume separate modes under each individual action with dierent
slendernesses that have to subsequently be combined to give an overall verication. The
disadvantage is that software capable of both elastic critical buckling analysis and second-
order analysis is required. Additionally, shell elements will be needed to determine elastic
critical modes resulting from exural loading.
An alternative method is to use second-order analysis with imperfections to cover both in-
plane and out-of-plane buckling eects as discussed in sections 5.2 and 5.3 of this guide, but
this has the same diculties as above.
The basic verication is performed by determining a single slenderness for out-of-plane
buckling, which can include combined lateral and lateraltorsional buckling. This slender-
ness is a slenderness for the whole system and applies to all members included within it. It
takes the usual Eurocode form as follows:
"
``
op

c
ult.k
c
cr.op
r
6.64 in EN 1993-1-1
where: c
ult.k
is the minimum load factor applied to the design loads required to reach the
characteristic resistance of the most critical cross-section ignoring out-of-plane
buckling but including moments from second-order eects and imperfections in-
plane, and
c
cr.op
is the minimum load factor applied to the design loads required to give elastic
critical buckling in an out-of-plane mode, ignoring in-plane buckling.
The rst stage of calculation requires an analysis to be performed to determine c
ult.k
. In-
plane second-order eects and imperfections must be included in the analysis because they
are not otherwise included in the resistance formula used in this method. If the structure
is not prone to signicant second-order eects as discussed in section 5.2 of this guide,
then rst-order analysis may be used. The exural stiness to be used is important in deter-
mining second-order eects and this is recognized by the text of clause 6.4.3.1(1). It will be
conservative to use the cracked stiness E
a
I
2
throughout if the bridge is modelled with beam
elements. If a nite-element shell model is used, the reinforcement can be modelled and the
concrete neglected so as to avoid an overestimation of stiness in cracked zones. Out-of-
plane second-order eects may need to be suppressed.
Each cross-section is veried using the interaction expression in clause 6.2 of EN 1993-1-1,
but using characteristic resistances. Eective cross-sections should be used for Class 4 sections.
The loads are all increased by a factor c
ult.k
until the characteristic resistance is reached. The
simple and conservative verication given in clause 6.2.1(7) of EN 1993-1-1 becomes:
N
Ed
N
Rk
M
y.Ed
M
y.Rk
1.0 D6.4
where N
Rk
and M
y.Rk
include allowance for any reduction necessary due to shear and
torsion, if separate checks of cross-section resistance are to be avoided in addition to the
buckling check being considered here. N
Ed
and M
y.Ed
are the axial forces and moments at
a cross-section resulting from the design loads. If rst-order analysis is allowable, the load
factor is determined from:
c
ult.k
N
Ed
N
Rk
M
y.Ed
M
y.Rk

1.0 D6.5
which is given in a Note to clause 6.3.4(4) of EN 1993-1-1.
If second-order analysis is necessary, c
ult.k
is found by increasing the imposed loads
progressively until one cross-section reaches failure according to expression (D6.4). This is
Clause 6.4.3.1(1)
94
necessary as the system is no longer linear and results from one analysis cannot simply be
factored up when the imposed load is increased.
The second stage is to determine the lowest load factor c
cr.op
to reach elastic critical
buckling in an out-of-plane mode but ignoring in-plane buckling modes. This will typically
require a nite-element model with shell elements to predict adequately the lateraltorsional
buckling behaviour. The reinforcement can be modelled and the concrete neglected so as to
avoid an overestimation of stiness in cracked zones. If the load factor can only be
determined separately for axial loads c
cr.N
and bending moments c
cr.M
, as might be the
case if standard textbook solutions are used, the overall load factor could be determined
from a simple interaction equation such as:
1
c
cr.op
1
c
cr.N
1
c
cr.M
Next, an overall slenderness is calculated for the entire systemaccording to equation (6.64)
of EN 1993-1-1. This slenderness refers only to out-of-plane eects as discussed above
because in-plane eects are separately included in the determination of action eects. A
reduction factor
op
for this slenderness is then determined. This reduction factor depends
in principle on whether the mode of buckling is predominantly exural or lateraltorsional
as the reduction curves can sometimes dier. The simplest solution is to take the lower of the
reduction factors for out-of-plane exural buckling, , and lateraltorsional buckling,
LT
,
from clauses 6.3.1 and 6.3.2, respectively, of EN 1993-2. For bridges, the recommended
reduction factors are the same but the National Annex could alter this. This reduction
factor is then applied to the cross-section check performed in stage 1, but this time using
design values of the material properties. If the cross-section is veried using the simple inter-
action expression (D6.4), then the verication taking lateral and lateraltorsional buckling
into account becomes:
N
Ed
N
Rk
,
M1
M
y.Ed
M
y.Rk
,
M1

op
D6.6
It follows from equation (D6.5) and expression (D6.6) that the verication is:
op
c
ult.k
M1
! 1.0 D6.7
Alternatively, separate reduction factors for axial load and
LT
for bending moment can
be determined for each eect separately using the same slenderness. If the cross-section is
veried using the simple interaction expression (D6.4), then the verication taking lateral
and lateraltorsional buckling into account becomes:
N
Ed
N
Rk
,
M1
M
y.Ed
LT
M
y.Rk
,
M1
1.0 D6.8
It should be noted that this procedure can be conservative where the element governing the
cross-section check is not itself signicantly aected by the out-of-plane deformations. The
method is illustrated in a qualitative example for a steel-only member in the Designers Guide
to EN 1993-2.
4
6.4.3.2. Simplied method
A simplied method is permitted for compression anges of composite beams and chords of
composite trusses by reference to EN 1993-2 clause 6.3.4.2. Its clause D2.4 provides the
stiness of U-frames in trusses (and plate girders by analogy). The method is based on
representing lateraltorsional buckling by lateral buckling of the compression ange. All
subsequent discussion refers to beam anges but is equally applicable to chords of trusses.
The method is primarily intended for U-frame-type bridges but can be used for other
types of exible bracing. It also applies to lengths between rigid restraints of a beamcompres-
sion ange, as is found in hogging zones in steel and concrete composite construction. The
use of the method for half-through bridges is discussed in the Designers Guide to EN 1993-2.
95
The method eectively ignores the torsional stiness of the beam. This may become signi-
cant for rolled steel sections but is generally not signicant for deeper fabricated girders.
A ow diagram for determining the slenderness
"
``
LT
for a length of beam of uniform depth
between rigid lateral supports is given in Fig. 6.13.
EN 1993-2 clause 6.3.4.2 allows the slenderness for lateral buckling to be determined from
an eigenvalue analysis of the compression chord. The ange (with an attached portion of web
No
Yes
either
or
END
Yes
Yes
Yes
No
No
Find
LT
from eq. (6.10) in EC3-2/6.3.4.2(4) (END)
Yes
No
No
No
No
Yes
Yes
Find c
c
. Then c = c
c
.
Find to
EC3-2/6.3.4.2(6)
Replace the bending
moment that causes
tension in the flange
by M
2
= 0
Scope: distortional lateral buckling of a steel flange of a span of an unhaunched composite
beam, with varying major-axis bending moment, rigid lateral supports at beam vertical
supports, intermediate lateral supports of stiffness C
d
(either rigid or flexible) at spacing
and/or continuous lateral restraint of stiffness c
c
per unit length
Is the flange susceptible to lateral buckling, to 6.3.2.4(1) of
EC3-1-1, via 6.4.3.2(1) of EC4-2 and 6.3.4.2(1) of EC3-2?
Neglect torsional
stiffness of the flange.
Simplified analysis to
6.4.3.2(1)
Elastic critical analysis as a
restrained strut to find N
crit
and
LT
to EC3-2/6.3.4.2(4)
Note 1: all subsequent clause
numbers refer to EN 1993-2
Restraint is flexible, of
stiffness C
d
. Find c
d
= C
d
/.
L is span length of beam
L, the distance between
rigid restraints =
Is there a continuous restraint? c = c
d
Classify discrete restraints.
Is C
d
> 4N
E
/L, to
EC3-2/6.3.4.2(6),
where L = ?
Find c
c
. Then c = c
d
+ c
c
Find to EC3-2/6.3.4.2(6)
Find N
crit
to eq. (D6.12) conservative unless n is an integral
number. Eqs (6.14) in EC3-2/6.3.4.2(8) are less conservative
This applies for uniform compressive force over length L, to EC3-2/6.3.4.2(6).
Is it required to take account of variation of bending moment over this length?
Find m from eqs (6.14) in the
Note to EC3-2/6.3.4.2(8)
Does the bending moment
change sign within length L?
Find N
crit
to
eq. (D6.13)
Find N
crit
= m
2
EI/L
2
but
2
EI/
2
Is there a
continuous
restraint?
c = = 0
N
crit
=
2
EI/L
2
Find n to eq. (D6.11).
Is n < 1?
Fig. 6.13. Flow diagram for slenderness for lateral buckling of a compressed ange
96
in the compression zone) is modelled as a strut with area A
eff
, supported by springs in the
lateral direction. These represent restraint from bracings (including discrete U-frames) and
from any continuous U-frame action which might be provided by the connection to the
deck slab. Buckling in the vertical direction is assumed to be prevented by the web in this
model but checks on ange-induced buckling according to Section 8 of EN 1993-1-5
should be made to conrm this assumption. Bracings can be exible, as is the case of
bracing by discrete U-frames, or can be rigid, as is likely to be the case for cross-bracing.
Other types of bracing, such as horizontal members at mid-height between beams together
with plan bracing or a deck slab, may be rigid or exible depending on their stiness as dis-
cussed below.
Elastic critical buckling analysis may be performed to calculate the critical buckling load,
N
crit
. The slenderness is then given by EN 1993-2 equation (6.10):
`
LT

A
eff
f
y
N
crit
s
where A
eff
A
f
A
wc
/3, as shown in Fig. 6.14. This approximate denition of A
eff
(greater
than the ange area) is necessary to ensure that the critical stress produced for the strut is the
same as that required to produce buckling in the beam under bending moment. For Class 4
cross-sections, A
eff
is determined making allowance for the reduction in area due to plate
buckling.
If smeared springs are used to model the stiness of discrete restraints such as discrete U-
frames, the buckling load should not be taken as larger than that corresponding to the Euler
load of a strut between discrete bracings. If computer analysis is used, there would be no
particular reason to use smeared springs for discrete restraints. This approximation is
generally only made when a mathematical approach is used based on the beam-on-elastic-
foundation analogy, which was used to derive the equations in EN 1993-2.
Spring stinesses for discrete U-frames and other restraints
Spring stinesses for discrete U-frames may be calculated using Table D.3 from Annex D of
EN 1993-2, where values of stiness, C
d
, can be calculated. (It is noted that the notation C
rather than C
d
is used in Table D.3.) A typical case covering a pair of plate girders with
stieners and cross-girders is shown in Fig. 6.15 for which the stiness (under the unit
applied forces shown) is:
C
d

EI
v
h
3
v
3

h
2
b
q
I
v
2I
q
D6.9
Section properties for stieners should be derived using an attached width of web plate in
accordance with Fig. 9.1 of EN 1993-1-5 (stiener width plus 30t
w
). If the cross-member is
composite, its second moment of area should be based on cracked section properties.
Equation (D6.9) also covers steel and concrete composite bridges without stieners and
cross-girders where the cross-member stiness is the short-term cracked stiness of the
deck slab and reinforcement, and the vertical-member stiness is based on the unstiened
web. For continuous U-frames, consideration of this stiness will have little eect in
A
f
A
wc
= t
w
h
wc
h
wc
Fig. 6.14. Denitions for eective compression zone for a Class 3 cross-section
97
raising the buckling resistance, unless the length between rigid restraints is large, and will
necessitate an additional check of the web for the U-frame moments induced. For multiple
girders, the restraint to internal girders may be derived by replacing 2I
q
by 3I
q
in the
expression for C
d
. Equation (D6.9) is then similar to equation (6.8). That diers only by
the inclusion of Poissons ratio in the stiness of the web plate and by the assumption
that the point of rotation of the compression ange is at the underside of the deck slab,
rather than some way within it.
The stiness of other restraints, such as a channel section placed between members at mid-
height, can be derived from a plane frame model of the bracing system. For braced pairs of
beams or multiple beams with a common system, it will generally be necessary to consider
unit forces applied to the compression anges such that the displacement of the ange is
maximized. For a paired U-frame, the maximum displacement occurs with forces in opposite
directions as in Fig. 6.15 but this will not always be the case. For paired beams braced by a
mid-height channel, forces in the same direction will probably give greater ange displacement.
A computer model is useful where, for example, the ange section changes or there is a
reversal of axial stress in the length of the ange being considered. In other simpler cases
the formulae provided in clause 6.3.4.2 of EN 1993-2 are applicable.
Elastic critical buckling load
The formula for N
crit
is derived from eigenvalue analysis with continuous springs. From
elastic theory (as set out, for example, in Refs 73 and 74), the critical load for buckling of
such a strut is:
N
crit
n
2

2
EI
L
2

cL
2
n
2
2
D6.10
where: I is the transverse second moment of area of the eective ange and web,
L is the length between rigid braces,
c is the stiness of the restraints smeared per unit length, and
n is the number of half waves in the buckled shape.
By dierentiation, this is a minimum when:
n
4
cL
4
4
EI
D6.11
which gives:
N
crit
2
cEI
p
D6.12
Equation (6.12) of EN 1993-2 is:
N
crit
mN
E
where:
N
E

2
EI
L
2
. m 2,
2

p
! 1.0.
cL
4
EI
and c C
d
,l
Neutral axis of
cross-girder
I
v
I
q
b
q
h
v
h
Fig. 6.15. Denitions of properties needed to calculate C
d
98
where C
d
is equal to the restraint stiness and l is the distance between restraints. When these
terms are substituted into equation (6.12) of EN 1993-2, equation (D6.12) is produced.
When
4
97, then c
4
EI,L
4
, and this model gives the results:
N
crit
2
2
EI,L
2
, n 1.0
It is not valid for lower values of c because then n<1, which implies a buckling half-
wavelength that exceeds the length L between rigid restraints, and a value of N
crit
lower
than that corresponding to a length L. In this case, the buckling load should be taken as:
N
crit

2
EI
L
2

cL
2
2
D6.13
Equation (D6.10) assumes that the end restraints that dene the length L are rigid. The
denition of rigid is discussed below. If intermediate bracings are not rigid, their stiness
can be taken to contribute to c but the length L is then dened by the length between
rigid bracings. Bracings at supports for typical composite bridges will usually be rigid due
to the need for them to provide torsional restraint to the beams.
Short lengths of beam between rigid bracings
Equation (D6.12) shows that the critical buckling load from equation (6.12) of EN 1993-2 is
independent of the length between rigid restraints. Equation (D6.13) is the basis for the rst
of equations (6.14) in EN 1993-2 for short lengths between rigid braces. The half wave length
of buckling is restricted to the length between braces, and any exible restraints included in
this length increase the buckling load from that for a pin-ended strut of length L. The
formulae provided also allow the eects of varying end moments and shears to be taken
into account, but they are not valid (and are actually unsafe) where the bending moment
reverses within length L. They are as follows:
m
1
1 0.441 j
1.5
3 2,350 50j or
m
2
1 0.441 j
1.5
0.195 0.05 j,100
0.5
if less than above
with:
j V
2
,V
1
and 21 M
2
,M
1
,1 j for M
2
< M
1
and V
2
< V
1
The subscripts in the symbols m
1
and m
2
correspond to the number of buckling half-waves
considered, n. Figure 6.16 enables the equation that gives the lower result to be found, by
0.0
0.2
0.4
0.6
0.8
1.0
0.4 0.2 0.0 0.2 0.4 0.6 0.8 1.0
500
450
350
400
300
260
280
240
M
2
/M
1
V
2
/V
1
Fig. 6.16. Values of (other than zero) at which both equations (6.14) in EN 1993-2 give the same value
for m. Below each curve, m
1
governs
99
giving values of at which m
1
m
2
. If the actual value of for a buckling length with ratios
V
2
,V
1
and M
2
,M
1
is lower than that shown in the gure, the equation for m
1
governs; if not,
m
2
governs.
Uniformly compressed ange
The benecial inuence of lateral restraint, represented by , is evident for the most adverse
case, a uniformly-compressed ange, for which j 1, 0. Then,
m
1
1.00 ,100. m
2
1.00 0.195
0.5
These ratios m are equal when
0.5
19.5, or 380, as shown by the point (1.0, 1.0) in
Fig. 6.16. The change from n 2 to n 3 can be found from equation (D6.10) which, in
terms of , is:
N
crit
,N
E
n
2
,n
2
This gives N
crit
for n 3 equal to that for n 2 when 3500 and N
crit
,N
E
13. The
equations for m
1
and m
2
are more complex than equation (D6.10) because their scope
includes non-uniform moment. Within the range of from 380 to 3000, the value m
2
for
uniform moment can be up to 10% higher than from equation (D6.10). This error is
small and is in part compensated for by the neglect of the torsional stiness of the beam
in this method. At 3500 it gives N
crit
,N
E
12.5, which is more conservative. It then
follows closely the results from equation (D6.10) as increases (and n also increases from
3 to 4). For up to 20 000, the values of m
2
dier from the predictions of (D6.10) by only
3.6%.
The shear ratio, j, in the equations for m
1
and m
2
, helps to describe the shape of
the bending moment diagram between points of restraint. It is linear if j 1.0. If j < 1.0,
the moments fall quicker than assumed from a linear distribution as shown in Fig. 6.17
and consequently the ange is less susceptible to buckling.
Change of sign of axial force within a length between rigid restraints
The lack of validity for moment reversal of equations (6.14) in EN 1993-2 is a problem for a
typical composite beam with cross-bracing adjacent to the internal supports. Where the most
distant brace from the pier is still in a hogging zone, the moment in the beam will reverse in
the span section between braces as shown in Fig. 6.18. In this region, m should not be
assumed to be 1.0 as this could lead to over-design of the beam or unnecessary provision
V
2
/V
1
< 1
V
2
/V
1
= 1
M
2
M
1
Fig. 6.17. Eect of shear ratio on the shape of the moment diagram
(6.14) not valid (6.14) not valid (6.14) valid
= bracing location
Fig. 6.18. Range of validity of equations (6.14) of EN 1993-2
100
of additional braces away from the pier, to ensure that the section between innermost braces
is entirely sagging and the bottomange is in tension. ANote to clause 6.3.4.2(7) of EN1993-
2 provides the option of assuming M
2
0. If benet fromthe restraining stiness of the deck
slab is ignored (i.e. c 0), and V
2
is conservatively taken equal to V
1
then this leads to
m 1.88.
Where the top ange is braced continuously by a deck, it may be possible to vary j to
produce a less conservative moment diagram. For the case in Fig. 6.19, the use of
V
2
,V
1
0, M
2
,M
1
0 achieves the same moment gradient at end 1 as the real set of
moments, and a distribution that lies everywhere else above the real moments and so is
still conservative. Equations (6.14) of EN1993-2 then give the value m 2.24, again ignoring
any U-frame restraint. Providing the top ange is continuously braced, the correct m would
be greater.
It is possible to include continuous U-frame action from an unstiened web between rigid
braces in the calculation of the spring stiness c. The benet is usually small for short lengths
between braces, and the web plate, slab and shear connection must be checked for the forces
implied by such action. Fig. 6.20 shows a graph of m against M
2
,M
1
with c 0, for varying
V
2
,V
1
.
It is possible to combine equations (6.10) and (6.12) of EN 1993-2 to produce a single
formula for slenderness, taking A
f
bt
f
for the ange area, as follows:
`
LT

A
eff
f
y
N
crit
s

A
f
A
wc
,3 f
y
L
2
m
2
EI
s
L
1 A
wc
,3A
f
f
y
,Em
2
12
b
3
t
f
bt
f
v
u
u
u
t
M
1
Conservative set of moments
with V
2
/V
1
= 0, M
2
/M
1
= 0
Real moments
M
2
= M
1
Fig. 6.19. Typical calculation of m where bending moment reverses
2.2
2.0
1.8
1.6
1.4
1.2
1.0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
m
= V
2
/V
1
= 1.0 0.75 0.50 0.25 0.00
M
2
M
1
M
2
/M
1
Fig. 6.20. Values of m ( N
crit
,N
E
) between rigid restraints with 0
101
so
`
LT
1.103
L
b
f
y
Em
r

1
A
wc
3A
f
s
D6.14
It is still necessary to evaluate N
crit
when checking the strength of the bracings.
The formulae in clause 6.3.4.2(7) of EN 1993-2 do not apply directly to haunched girders
as they assume that the ange force is distributed in the same way as the bending moment.
The general method of using an eigenvalue analysis based on the forces in the compression
chord is still applicable. Alternatively, the formulae provided could be applied using the least
value of the spring stiness c within the length considered. The ange force ratio F
2
,F
1
is
used instead of the moment ratio M
2
,M
1
, with V
2
,V
1
taken as 1.0, when applying equation
(6.14) of EN 1993-2.
EN 1993-2 clause 6.3.4.2(7) allows the buckling verication to be performed at a distance
of 0.25L
k
0.25L,

m
p
(i.e. 25% of the eective length) from the end with the larger
moment. (The symbols L
k
and l
k
are both used for eective length in 6.3.4.2.) This
appears to double-count the benet from moment shape derived in equations (6.14) of
EN 1993-2; but it does not do so. The check at 0.25L
k
reects the fact that the peak stress
from transverse buckling of the ange occurs some distance away from the rigid restraint
to the ange, whereas the peak stress from overall bending of the beam occurs at the
restraint. (In this model, the beam ange is assumed to be pin-ended at the rigid transverse
restraints.) Since these two peak stresses do not coexist they are not fully additive, and the
buckling verication can be performed at a design section somewhere between these two
locations. The cross-section resistance must still be veried at the point of maximum
moment.
There are clearly problems with applying clause 6.3.4.2(7) of EN1993-2 where the moment
reverses as the section 0.25L,

m
p
from an end may be a point of contraexure. In this situa-
tion, it is recommended here that the design section be taken as 25% of the distance from the
position of maximum moment to the position of zero moment. In addition, if benet is taken
of verication at the 0.25L
k
cross-section, the calculated slenderness above must be modied
so that it refers to this design section. The critical moment value will be less here and the
slenderness is therefore increased. This can be done by dening a new slenderness at the
0.25L
k
section such that
`
0.25Lk
`
LT
M
1
M
0.25Lk
s
where M
0.25Lk
is the moment at the 0.25L
k
section. This procedure is illustrated in Example
6.6 below. It should be noted that the k in M
0.25Lk
does not imply a characteristic value; this is
a design value.
Stiness of braces
The formulae in EN 1993-2 discussed above are only valid where the end restraints that
dene the length L are rigid. It is possible to equate N
crit
2
cEI
p
to
2
EI,L
2
to nd a
limiting stiness that gives an eective length equal to the distance between rigid restraints,
L, but this slightly underestimates the required stiness. This is because the formulae assume
that the restraints are continuously smeared when they are in fact discrete. The former
analysis gives a required value for C
d
of
4
EI,4L
3
, whereas the correct stiness is
4
2
EI,L
3
, which is 62% higher.
Interaction with compressive axial load
The interaction with axial load is covered in clause 6.4, in clause 6.4.3.1 only, via the general
method given in EN 1993-2. Axial load has several eects including:
.
magnication of the main bending moment about the horizontal axis of the beam (the
second-order eect)
102
.
increase of stress in the compression ange leading to an increased tendency for lateral
buckling.
Most bridge cross-sections are either Class 3 or 4 at supports so the stresses fromaxial load
can simply be assumed to be applied to the cracked composite section, and the elastic section
resistance can be used. At mid-span, beams are usually Class 1 or 2 and the calculation of a
modied plastic moment resistance in the presence of axial load is relatively simple. The
plastic neutral axis is so chosen that the total compressive force exceeds the total tensile
force by an amount equal to the axial load.
Care must however be taken to ensure that the bending resistance is obtained about an axis
at the height of the applied axial force assumed in the global analysis. This is important
for non-symmetric beams as the elastic and plastic neutral axes for bending alone do not
coincide, whereas they do for a symmetric section. Most of clause 6.7 is for doubly-symmetric
sections only, but the general method of clause 6.7.2 may be applied to beams provided that
compressive stresses do not exceed their relevant limiting values where Class 3 and 4 cross-
sections are involved.
Alternatively, the cross-section can be designed using a conservative interaction expression
such as that in clause 6.2.1(7) of EN 1993-1-1:
N
Ed
N
Rd
M
y.Ed
M
y.Rd
1.0
where N
Rd
and M
y.Rd
are the design resistances for axial force and moment acting individu-
ally but with reductions for shear where the shear force is suciently large. A similar inter-
action expression can be used for the buckling verication with the terms in the denominator
replaced by the relevant buckling resistances:
N
Ed
N
b.Rd
M
Ed
M
b.Rd
1.0 D6.15
The value for M
Ed
should include additional moments from in-plane second-order eects
(including from in-plane imperfections). Such second-order eects will normally be negligible.
The buckling resistance N
b.Rd
should be calculated on the basis of the axial stress required for
lateral buckling of the compression ange. This method is illustrated in Example 6.6 below.
Beams without plan bracing or decking during construction
During construction it is common to stabilize girders in pairs by connecting them with
torsional bracing. Such bracing reduces or prevents torsion of individual beams but does
not restrict lateral deection. Vertical torsional cross-bracing as shown in Fig. 6.21 has
been considered in the UK for many years to act as a rigid support to the compression
ange, thus restricting the eective length to the distance between braces. Opinion is now
somewhat divided on whether such bracing can be considered fully eective and BS
Centre of rotation
Torsional
bracing
(a) (b)
Fig. 6.21. Torsional bracing and shape of buckling mode, for paired beams: (a) plan on braced pair of
beams showing buckling mode shape; (b) cross-section through braced pair of beams showing buckling
mode shape
103
5400:Part 3:2000
11
introduced a clause to cover this situation which predicts that such
bracing is not fully eective.
This situation arises because equilibrium of the braced pair under torsion requires oppos-
ing vertical forces to be generated in the two girders. Consequently one girder moves up, one
moves down and some twist of the girder pair is generated, albeit much less than for an
unbraced pair. If the beam span-to-depth ratio is large, the deections and hence twists
can be signicant. The Designers Guide to EN 1993-2
4
suggests a method based on BS
5400 Part 3, but in some cases it may lead to the conclusion that plan bracing is necessary.
A better estimate of slenderness can be made using a nite-element analysis.
A nite-element model of a non-composite beam, using shell elements for the paired
main beams and beam elements to represent the bracings, can be set up relatively quickly
with modern commercially available software. Elastic critical buckling analysis can then
be performed and a value of M
cr
determined directly for use in slenderness calculation to
clause 6.3.2 of EN 1993-2. This approach usually demonstrates that the cross-bracing is
not fully eective in limiting the eective length of the ange to the distance between
bracings, but that it is more eective than is predicted by BS 5400. For simply-
supported paired girders, a typical lowest buckling mode under dead load is shown in
Fig. 6.21.
Example 6.6: bending and shear in a continuous composite beam
The locations of the bracings and splices for the bridge used in earlier examples are shown
in Fig. 6.22. The two internal beams are Class 3 at each internal support as found in
Example 5.4. The cross-sections of the central span of these beams, also shown, are as
in Examples 5.4 and 6.1 to 6.4. The neutral axes shown in the cross-section at the piers
are for hogging bending of the cracked composite section.
23 400 19 000
3800
19 000
Location of bracing
Location of splice, about 6.0 m from pier
E
A
B C F G
D
3800
400 40
400 25
1160 25
775
Plastic NA
209
Elastic NA
60
70
1225
250
25
3100
12 990 mm
2
6490 mm
2
Cross-section for lengths BD and EG
250
1225
400 30
400 20
1175 12.5
25 haunch
3100
998
NA
Cross-section for length DE
Fig. 6.22. Details for Examples 6.6 and 6.7
104
The design ultimate hogging moments at internal support B are 2213 kNm on the steel
section plus 4814 kNm on the composite section, so M
Ed
7027 kNm. The coexisting
hogging moment at the braced point C in the central span is 4222 kNm. The vertical
shear at point C is 70% of that at the pier B. The hogging bending moment at the
splice, where the beam cross-section changes, is 3000 kNm.
Lateraltorsional buckling adjacent to the pier and in the main span beyond the brace is
checked and the eect of a coexisting axial compression of 1000 kN applied to the
composite section is considered. This could arise in a semi-integral bridge with screen
walls at its ends.
Elastic resistance to bending at an internal support
The elastic section moduli for the cross-section at point B are given in Table 6.2. These are
based on the extreme bres but it would have been permissible to base them on the
centroids of the anges in accordance with EN 1993-1-1 clause 6.2.1(9). To nd M
el.Rd
,
the factor k (clause 6.2.1.4(6)) is found for the top and bottom surfaces of the steel
beam, as follows. The result will be used in checks on buckling, so f
yd
is found using
M1
(1.1). Primary shrinkage stresses are neglected because the deck slab is assumed
to be cracked.
Table 6.2. Section moduli at an internal support, in 10
6
mm
3
units
Top layer of bars Top of steel section Bottom of steel section
Gross steel section 18.28 22.20
Cracked composite section 34.05 50.31 29.25
For the top ange,
2213,18.28 kM
c.Ed
,50.31 345,1.1 so kM
c.Ed
9688 kNm
For the bottom ange,
2213,22.20 kM
c.Ed
,29.25 345,1.1 so kM
c.Ed
6258 kNm
The elastic resistance is governed by the bottom ange, so that:
M
el.Rd
2213 6258 8471 kNm
The maximum compressive stress in the bottom ange is:
o
a.bot
2213,22.2 4814,29.25 264 N/mm
2
Resistance of length BC (Fig. 6.22) to distortional lateral buckling
Strictly, the stiness of the bracing should rst be checked (or should later be designed) so
that the buckling length is conned to the length between braces. This is done in Example
6.7. The bending-moment distribution is shown in Fig. 6.23.
Where no vertical web stieners are required, the deck slab provides a small continuous
U-frame stiness. This could be included using Table D.3 of EN 1993-2, case 1a, to
calculate a stiness, c. This contribution has been ignored to avoid the complexities of
designing the deck slab and shear connection for the forces implied, and because any
need to stien the web has not yet been considered. Therefore from clause 6.3.4.2(6) of
EN 1993-2, cL
4
,EI 0.
The compression zone of the beam is designed as a pin-ended strut with continuous
vertical restraint, and lateral restraint at points 3.80 mapart (Fig. 6.22). Its cross-sectional
areas are:
.
ange: A
f
400 40 16 000 mm
2
105
.
web compression zone: A
wc
735 25 18 375 mm
2
. (Conservatively based on the
height of the neutral axis for the composite section. It could be based on the actual
depth for the accumulated stress prole, which is 633 mm.)
The second moment of area of the compressed area is calculated for the bottom ange,
as the contribution from the web is negligible:
I 40 400
3
,12 213.3 10
6
mm
4
The applied bending moments at each end of the equivalent strut are:
M
Ed.1
7027 kNm; M
Ed.2
4222 kNm
From EN 1993-2 clause 6.3.4.2(7):
M
2
,M
1
4222,7027 0.60 and j V
2
,V
1
0.70
21 M
2
,M
1
,1 j 21 0.60,1 0.70 0.46
When 0, the rst two of equations (6.14) in clause 6.3.4.2(8) of EN 1993-2 both give:
m 1 0.441 j
1.5
1 0.441 0.700.46
1.5
1.23
If the deck slab is considered to provide U-frame restraint, the value of mfor this length
of ange is found to be still only 1.26, so there is no real benet to lateral stability in con-
sidering U-frame action over such a short length of beam.
From equation (D6.14):
`
LT
1.1
L
b
f
y
Em
r

1
A
wc
3A
f
s
1.1
3800
400
345
210 10
3
1.23
r

1
18375
3 16000
r
0.45
This exceeds 0.2 so from clause 6.3.2.2(4) of EN 1993-1-1 this length of ange is prone to
lateraltorsional buckling.
The ratio h,b 1225,400 3.1 exceeds 2.0, so from Table 6.4 in clause 6.3.2.2 of
EN 1993-1-1 the relevant buckling curve is curve d. Hence, c
LT
0.76 from Table 6.3.
From equation (6.56) in EN 1993-1-1, clause 6.3.2.2:
LT
0.51 c
LT
`
LT
0.2

`
2
LT
0.5 1 0.760.45 0.2 0.45
2
0.696
LT

1
LT
2
LT
`
2
LT
q
1
0.696
0.696
2
0.45
2
p 0.81
Applying this reduction factor gives:
M
b.Rd

LT
M
el.Rd
0.81 8471 6862 kNm
At the internal support, M
Ed
7027 kNm (2% higher). However, clause 6.3.4.2(7) of
EN 1993-2 provides the option of making this check at a distance of 0.25L,

m
p
from
the support. This distance is:
0.25 3800,
1.23
p
857 mm
Using linear interpolation, Fig. 6.23, this gives M
Ed
6394 kNm. The modied slender-
ness is:
`
0.25Lk
`
LT
M
1
M
0.25Lk
s
0.45
7027
6394
r
0.47
This reduces
LT
from 0.81 to 0.80, so the new resistance is:
M
b.Rd

LT
M
el.Rd
0.80 8471 6777 kNm
This exceeds M
Ed
(6394 kNm), so this check on lateral buckling is satised.
106
The further condition that the elastic resistance (8471 kNm) is not exceeded at the point
of peak moment (7027 kNm) is satised. This cross-section should also be checked
for combined bending and shear, but in this case the shear is less than 50% of the
shear resistance and thus no interaction occurs.
Resistance with a coexisting axial compression of 1000 kN
An axial force of 1000 kN is applied at the height of the elastic centroidal axis of the
cracked cross-section at the internal support. Susceptibility to in-plane second-order
eects is rst checked. The second moment of area of the cracked composite section at
mid-span, 1.326 10
10
mm
4
, is conservatively used to determine the elastic critical
exural buckling load for major-axis buckling.
N
cr

2
EI
L
2

2
210 1.326 10
10
31 000
2
28 600 kN
From clause 5.2.1(3):
c
cr

28 600
1000
29 10
so in-plane second-order eects may be neglected.
At the internal support, the area of the cracked composite section is 74 480 mm
2
, so the
axial compressive stress is 1000 10
3
,74 480 13.4 N/mm
2
. This changes the total
stresses at the extreme bres of the steel beamto 204 N/mm
2
tension and 278 N/mm
2
com-
pression. This stress distribution is found to leave the cross-section at the pier still in Class
3. The section at the brace is also in Class 3.
The conservative linear interaction between axial and bending resistances of expression
(D6.15) will be used:
N
Ed
N
b.Rd
M
Ed
M
b.Rd
1.00
so N
b.Rd
must be found, based on buckling of the web and bottom ange under uniform
compression. Since the combined stress distribution leads to a Class 3 cross-section
throughout the buckling length, the gross section will be used in this analysis, even for
the calculation of the compression resistance, below.
If the section became Class 4, eective properties should be used. These could be
derived either separately for moment and axial force or as a unique eective section
under the combined stress eld. In either case, the moment of the axial force produced
by the shift of the neutral axis should be considered.
For axial force alone, the required areas of the cross-sections are:
A
f
400 40 16 000 mm
2
, A
wc
1160 25 29 000 mm
2
Neglecting U-frame restraint as before, equations (6.14) in clause 6.3.4.2(7) of EN1993-
2 with M
1
M
2
and 0 give m 1.00.
From equation (D6.14):
`
LT
1.1
L
b
f
y
Em
r

1
A
wc
3A
f
s
1.1
3800
400
345
210 10
3
1.00
r

1
29 000
3 16 000
r
0.54
From curve d on Fig. 6.4 of EN 1993-1-1, 0.74. The axial buckling resistance of the
cracked composite cross-section, based on buckling of the bottom ange, is:
N
b.Rd
Af
yd
0.74 74 480 345,1.1 17 285 kN
It should be noted that N
b.Rd
does not represent a real resistance to axial load alone as
the cross-section would then be in Class 4 and a reduction to the web area to allow for
local buckling in accordance with EN 1993-1-5 would be required. It is however valid
107
to use the gross cross-section in the interaction here as the cross-section is Class 3 under
the actual combination of axial force and bending moment. The use of a gross cross-
section also avoids the need to consider any additional moment produced by the shift
in centroidal axis that occurs when an eective web area is used.
From expression (D6.15),
1000,17 285 6394,6777 1.00
which is just satisfactory, with these conservative assumptions. A check of cross-section
resistance is also required at the end of the member, but this is satised by the check of
combined stresses above.
Buckling resistance of the mid-span region of the central span
An approximate check is carried out for the 23.4 m length between the two braced points
(C and F in Fig. 6.22), at rst using the cross-section at the pier throughout to derive the
reduction factor. This is slightly unconservative as the cross-section reduces at the splice.
An axial force is not considered in this part of the example.
3.80
4222
6394
7027
0
27.2 0.86
M
Ed
B
e
n
d
i
n
g

m
o
m
e
n
t

(
k
N
m
)
Distance from support (m)
Section:
M
2
= 0,V
2
= V
1
M
2
= 0,V
2
= 0
B F C
3000
D
Fig. 6.23. Bending action eects and resistances for an internal span
It is assumed that maximum imposed load acts on the two side spans and that only a
short length near mid-span is in sagging bending, as sketched in Fig. 6.23. Since the
bending moment reverses, equations (6.14) in EN 1993-2 are not directly applicable. If
the suggestion of clause 6.3.4.2(7) of EN 1993-2 is followed, and M
2
is taken as zero at
the other brace (cross-section F), the bending-moment distribution depends on the
value assumed for V
2
, the vertical shear at F. Two possibilities are shown in Fig. 6.23.
Their use does not follow directly from the discussion associated with Fig. 6.18, where
the moment was assumed to reverse only once in the length between rigid restraints. In
Fig. 6.23, the two ctitious sets of moments do not always lie above the real set and are
therefore not obviously conservative. However, the interaction of the hogging moment
at one end of the beam with the buckling behaviour at the other end is weak when the
moment reverses twice in this way.
BS 5400:Part 3
11
included a parameter j which was used to consider the eect of
moment shape on buckling resistance. For no reversal, m is in principle equivalent to
1,j
2
, although there is not complete numerical equivalence. Figure 6.24 gives comparative
values of m
0
1,j
2
. This shows that for the worst real moment distribution, where
the moment just remains entirely hogging, the value of m
0
is greater (less conservative)
than for the two possibilities in Fig. 6.23. This shows that the less conservative of these
possibilities (V
2
0) can be used. It gives m 2.24 from equations (6.14).
108
From equation (D6.14) with A
wc
25 735 18 375 mm
2
:
`
LT
1.1
L
b
f
y
Em
r

1
A
wc
3A
f
s
1.1
23 400
400
345
210 10
3
2.24
r

1
18 375
3 16 000
r
2.05 0.2
Using curve d in Fig. 6.4 of EN 1993-1-1,
LT
0.17.
If the slab reinforcement at cross-section C is the same as at the support,
M
el.Rd
8471 kNm, and M
b.Rd
0.17 8471 1440 kNm
This is far too low, so according to this method another brace would be required further
from the pier to reduce the buckling length. U-frame action is now considered.
m = 1.73 m = 2.5 m = 4.0
Fig. 6.24. Variation of m
0
with dierent bending-moment distributions, derived from BS 5400:Part 3
11
Use of continuous U-frame action simplied method of EN 1993-2, clause 6.3.4.2
The method of clause 6.4.2 is inapplicable because the cross-section is not uniform along
length CF, and a region near the splice is in Class 4 for hogging bending. The method of
clause 6.3.4.2 of EN 1993-2 is therefore used. Its spring stiness c is the lateral force per
unit length at bottom-ange level, for unit displacement of the ange. It is related to the
corresponding stiness k
s
in clause 6.4.2 by:
k
s
ch
2
s
D6.16
where h
s
is the distance between the centres of the steel anges (Fig. 6.10).
For most of the length CF the web is 12.5 mm thick so the cross-section for length DE
(Fig. 6.22) is used for this check, with h
s
1200 mm. For the bottom ange,
I
afz
30 400
3
,12 160 10
6
mm
4
D6.17
In comparison with this thin web, the slab component of the U-frame is very sti, so its
exibility is neglected here, as is the torsional stiness of the bottom ange. In general, the
exibility of the slab should be included. From equations (6.8) and (6.10) with k
1
)k
2
,
so that k
s
% k
2
, and using equation (D6.16),
c k
s
,h
2
s
E
a
t
w
h
s

3
,41 i
2
a
210 10
6
12.5,1200
3
,4 0.91 65.2 kN,m
2
As an alternative, equation (D6.9) could be used for the calculation of c. As discussed
in the main text, there are some minor dierences in the denition of the height terms,
h. Those in equation (D6.9) seem more appropriate where the slab exibility becomes
important.
From EN 1993-2 clause 6.3.4.2 and result (D6.17),
cL
4
,EI 65.2 23.4
4
,210 160 582
The less conservative assumption, V
2
0, is used with M
2
0. This gives j 0, 2.
From Fig. 6.16, the second of equations (6.14) in clause 6.3.4.2 of EN 1993-2 governs.
109
With j 0 it gives:
m 1 0.44
1.5
0.195 0.05
0.5
1 1.245 7.117 9.36
U-frame action (the nal term) is now a signicant contributor to m. (For uniform
moment and the same , m 5.70.)
The cross-section reduces at the splice position, approximately 6 m from the pier, so the
minimum cross-section is conservatively considered throughout. In equation (D6.14) for
"
``
LT
, the areas in compression are:
.
ange: A
f
400 30 12 000 mm
2
.
web compression zone: A
wc
683 47 12.5 7950 mm
2
. (Conservatively based
on the height of the neutral axis for the composite section shown in Fig. 6.5, which
is 683 mm above the bottom ange.)
Hence from equation (D6.14):
`
LT
1.1
L
b
f
y
Em
r

1
A
wc
3A
f
s
1.1
23 400
400
345
210 10
3
9.36
r

1
7950
3 12 000
r
0.942
Curve d of Fig. 6.4 in EN1993-1-1 gives
LT
0.49. With M
el.Rd
fromExample 6.3, the
buckling resistance
LT
M
el.Rd
is:
M
b.Rd
0.49 6390 3131 kNm
which is less than 4222 kNm (the moment at the brace). By inspection, a check at the
0.25L
k
design section will also not pass.
These checks are however conservative as they assume the minimum cross-section
throughout. If reference is made back to the expressions for slenderness given in equations
(6.10) of EN 1993-2 and (6.7), it is seen that:
`
LT

A
eff
f
y
N
crit
s

M
Rk
M
crit
s
so that M
crit
measured at the brace is eectively
M
crit

M
Rk
`
2
LT
6390 1.1
0.942
2
7921 kNm
For the length of the beam with the larger cross-section, M
Rk
is larger than assumed
above. The buckling moment M
crit
is however mainly inuenced by the long length of
smaller cross-section so that it will be similar to that found above, even if the short
lengths of stier end section are considered in the calculation.
Use of continuous U-frame action general method of EN 1993-1-1, clause 6.3.4
A revised slenderness can be determined using the method of EN 1993-1-1 clause 6.3.4.
Fromequation (D6.5) with N
Ed
0, within the span between braces, the minimumvalue of
c
ult.k

M
y.Rk
M
y.Ed
is
8471 1.1
4222
2.21
at the brace location, where:
M
y.Rd
8471 kNm
(For the weaker section at the splice location, M
y.Rk
,M
y.Ed
6390 1.1,3000 2.34.
110
The minimum load factor to cause lateraltorsional buckling is:
c
cr.op

M
crit
M
y.Ed
7921
4222
1.88
The system slenderness of the span between the braces, from equation (6.64) of EN 1993-
1-1, is:
"
``
op

c
ult.k
c
cr.op
r

2.21
1.88
r
1.08
and hence the reduction factor, fromcurve d of Fig. 6.4 in EN1993-1-1, is
op
0.43. The
verication is then performed according to equation (D6.7):
op
c
ult.k
M1
0.43 2.21
1.1
0.86 < 1.0
so the beam is still inadequate. This verication is equivalent to M
b.Rd
0.43 8471
3643 kNm at the brace, which is still less than the applied moment of 4222 kNm.
It would be possible to improve this verication further by determining a more accurate
value of M
crit
from a nite-element model. However, inclusion of U-frame action has the
disadvantage that the web and shear connection would have to be designed for the result-
ing eects. A better alternative could be the addition of another brace adjacent to the
splice location.
Example 6.7: stiness and required resistance of cross-bracing
The bracing of the continuous bridge beam in Example 6.6 comprises cross-bracing made
from 150 150 18 angle and attached to 100 20 stieners on a 25 mm thick web. The
bracings are checked for rigidity and the force in them arising from bracing the anges is
determined. The eects of the 1000 kN axial compressive force are included.
1 kN 1 kN
Stiffener effective
section
Deck slab
Bracing
1020
3100
Fig. 6.25. Cross-bracing for Examples 6.6 and 6.7
The stiness of the brace was rst calculated from the plane-frame model shown in
Fig. 6.25. From Fig. 9.1 of EN 1993-1-5, the eective section of each stiener includes
a width of web:
30t
w
t
st
30 0.81 25 20 628 mm
Hence,
A
st
628 25 100 20 17 700 mm
2
This leads to I
st
9.41 10
6
mm
4
.
The deck slab spans 3.1 m. From clause 5.4.1.2 its eective span is 3.1 0.7 2.17 m,
and its eective width for stiness is 0.25 2.17 0.542 m. Its stiness is conservatively
111
based on the cracked section, with the concrete modulus taken as E
cm
/2 to represent the
fact that some of the loading is short term and some is long term. Greater accuracy is not
warranted here as the concrete stiness has little inuence on the overall stiness of the
cracked section.
From elastic analysis for the forces shown in Fig. 6.25, the stiness is:
C
d
80 kN/mm
From clause 6.3.4.2(6) of EN 1993-2, the condition for a lateral support to a com-
pressed member to be rigid is:
C
d
! 4
2
EI,L
3
b
where, for length BC in Fig. 6.23,
L
b
3.80 m
and
I 213.3 10
6
mm
4
Hence,
C
d
! 4
2
210 213.3,3.8
3
1000 32.2 kN/mm
The support is rigid.
C
d
Fig. 6.26. Elastic stiness of bracing to a pin-ended strut
The formula in EN 1993-2 here assumes that supports are present every 3.8 m such that
the buckling length is restricted to 3.8 m each side of this brace and that the force in the
ange is constant each side of the brace. The limiting spring stiness for C
d
is analogous to
that required for equilibrium of a strut with a pin joint in it at the spring position, such
that buckling of the lengths each side of the pin joint occurs before buckling of the
whole strut into the brace see Fig. 6.26. A small displacement of the strut at the pin
joint produces a kink in the strut and a lateral force on the brace, which must be sti
enough to resist this force at the given displacement. For a given displacement, the
kink angle and thus the force on the spring is increased by reducing the length of the
strut each side of the spring. This kink force is also increased because the critical buckling
load for the lengths each side of the spring is increased by the reduction in length. Here,
the ange force is not the same each side of the brace and the length of unbraced ange is
greater than 3.8 m on one side of the brace. The kink force is therefore overestimated and
the calculated value for C
d
is conservative. This conrms the use in Example 6.6 of L as
the length between braces.
The design lateral force for the bracing is nowfound, using clause 6.3.4.2(5) of EN1993-
2. From Example 6.6,
"
``
LT
0.45, and the eective area of the compressed ange is:
A
eff
16 000 18 375,3 22 120 mm
2
From equations in EN 1993-2, clause 6.3.4.2,
k
EI,N
crit
1,2
EI
"
``
2
LT
,A
eff
f
y
1,2
0.45
210 213.3 1000
22 120 345

1,2
3.43 m
112
6.5. Transverse forces on webs
The local resistance of an unstiened and unencased web to forces (typically, vertical forces)
applied through a steel ange can be assumed to be the same in a composite member as in a
steel member, so clause 6.5 consists mainly of references to EN 1993-1-5. High transverse
loads are relatively uncommon in bridge design other than during launching operations or
from special vehicles or heavy construction loads, such as from a crane outrigger. Theoreti-
cally, wheel loads should be checked but are unlikely ever to be signicant.
The patch loading rules given in EN 1993-1-5 Section 6 make allowance for failure by
either plastic failure of the web, with associated plastic bending deformation of the ange,
or by buckling of the web. More detail on the derivation and use of the rules is given in
4
The rules for patch loading can only be used if the
geometric conditions in EN 1993-1-5 clause 2.3 are met; otherwise EN 1993-1-5 Section 10
should be used. Clause 6.1(1) of EN 1993-1-5 also requires that the compression ange is
adequately restrained laterally. It is not clear what this means in practice, but the restraint
requirement should be satised where the ange is continuously braced by, for example, a
deck slab or where there are sucient restraints to prevent lateraltorsional buckling.
Clause 6.5.1(1) states that the rules in EN 1993-1-5 Section 6 are applicable to the non-
composite ange of a composite beam. If load is applied to the composite ange, the rules
could still be used by ignoring the contribution of the reinforced concrete to the plastic
bending resistance of the ange. No testing is available to validate inclusion of any
contribution. A spread of load could be taken through the concrete ange to increase the
sti loaded length on the steel ange. There is limited guidance in EN 1992 on what angle
of spread to assume; clause 8.10.3 of EN 1992-1-1 recommends a dispersion angle of
tan
1
2,3, i.e. 348, for concentrated prestressing forces. It would be reasonable to use 458
here, which would be consistent with previous bridge design practice in the UK.
Clause 6.5.1(2) makes reference to EN1993-1-5 clause 7.2 for the interaction of transverse
force with axial force and bending. This gives:
j
2
0.8j
1
1.4
where:
j
2

o
z.Ed
f
yw
,
M1
F
Ed
f
yw
L
eff
t
w
,
M1
F
Ed
F
Rd
Clause 6.5
Clause 6.5.1(1)
Clause 6.5.1(2)
The distance between braced points is 3.8 m, so
k
< 1.2 .
(Since the brace has been found to be rigid, and from Example 6.6, m 1 so that
N
crit

2
EI,
2
,
k
is obviously less than , and this check was unnecessary.)
From clause 6.3.4.2(5), the lateral force applied by each bottom ange to the brace is:
F
Ed
N
Ed
/100
From Example 6.6, the greatest compressive stress in the bottom ange at the pier is
278 N/mm
2
.
Hence:
F
Ed

N
Ed
100
278 22 120, 100 1000 61.5 kN
The axial force in the bracing is then approximately:
61.5
cos tan
1
1020,3100
64.7 kN
There will also be some bending moment in the bracing members due to joint eccen-
tricities.
113
is the usage factor for transverse load acting alone, and
j
1

o
x.Ed
f
y
,
M0
N
Ed
f
y
A
eff
,
M0
M
Ed
N
Ed
e
N
f
y
W
eff
,
M0
is the usage factor for direct stress alone, calculated elastically. The calculation of j
1
should
take account of the construction sequence as discussed in section 6.2.1.5 of this guide. It can
be seen that this interaction expression does not allow for a plastic distribution of stress for
bending and axial force. Even if the cross-section is Class 1 or 2, this will not lead to any
discontinuity with the plastic bending resistance at low transverse load as only 80% of the
elastic bending stress is considered and the limiting value of the interaction is 1.4. The
ratio between the plastic and elastic resistances to bending for typical composite beams is
less than 1.4.
Clause 6.5.2 covers ange-induced buckling of webs by reference to EN1993-1-5 Section 8.
If the ange is suciently large and the web is very slender, it is possible for the whole ange
to buckle in the plane of the web by inducing buckling in the web itself. If the compression
ange is continuously curved in elevation, whether because of the sot prole or because the
whole beam is cambered, the continuous change in direction of the ange force causes a
radial force in the plane of the web. This force increases the likelihood of ange-induced
buckling into the web. Discussion on the use of Section 8 of EN 1993-1-5 is provided in
4
6.6. Shear connection
6.6.1. General
6.6.1.1. Basis of design
Clause 6.6 is applicable to shear connection in composite beams. Clause 6.6.1.1(1) refers also
to other types of composite member. Shear connection in composite columns is addressed in
clause 6.7.4, but reference is made to clause 6.6.3.1 for the design resistance of headed stud
connectors.
Although the uncertain eects of bond are excluded by clause 6.6.1.1(2)P, friction is not
excluded. Its essential dierence frombond is that there must be compressive force across the
relevant surfaces. This usually arises from wedging action. Provisions for shear connection
by friction are given in clause 6.7.4.2(4) for columns.
Inelastic redistribution of shear (clause 6.6.1.1(3)P) is most relevant to building design in
the provisions of EN1994-1-1 for partial shear connection. Inelastic redistribution of shear is
allowed in a number of places for bridges including:
.
clause 6.6.1.2 (which allows redistribution over lengths such that the design resistance is
not exceeded by more than 10%)
.
clause 6.6.2.2 (which permits assumptions about the distribution of the longitudinal shear
force within an inelastic length of a member)
.
clauses 6.6.2.3(3) and 6.6.2.4(3) for the distribution of shear in studs from concentrated
loads.
Clause 6.6.1.1(4)P uses the term ductile for connectors that have deformation capacity
sucient to assume ideal plastic behaviour of the shear connection. Clause 6.6.1.1(5)
quanties this as a characteristic slip capacity of 6 mm.
75
The need for compatibility of load/slip properties, clause 6.6.1.1(6)P, is one reason why
neither bond nor adhesives can be used to supplement the shear resistance of studs. The
combined use of studs and block-and-hoop connectors has been discouraged for the same
reason, though there is little doubt that eectively rigid projections into the concrete slab,
such as bolt heads and ends of ange plates, contribute to shear connection. This Principle
is particularly important in bridges, where the fatigue loading on individual connectors may
otherwise be underestimated. This applies also where bridges are to be strengthened by retro-
tting additional shear connectors.
Clause 6.5.2
Clause 6.6.1.1(1)
Clause 6.6.1.1(2)P
Clause 6.6.1.1(3)P
Clause 6.6.1.1(4)P
Clause 6.6.1.1(5)
Clause 6.6.1.1(6)P
114
Separation, in clause 6.6.1.1(7)P, means separation sucient for the curvatures of the
two elements to be dierent at a cross-section, or for there to be a risk of local corrosion.
None of the design methods in EN 1994-2 takes account of dierences of curvature,
which can arise from a very small separation. Even where most of the load is applied by
or above the slab, as is usual, tests on beams with unheaded studs show separation, especially
after inelastic behaviour begins. This arises from local variations in the exural stinesses of
the concrete and steel elements, and from the tendency of the slab to ride up on the weld
collars. The standard heads of stud connectors have been found to be large enough to
control separation, and the rule in clause 6.6.1.1(8) is intended to ensure that other types
of connector, with anchoring devices if necessary, can do so.
Resistance to uplift is much inuenced by the reinforcement near the bottomof the slab, so
if the resistance of an anchor is to be checked by testing, reinforcement in accordance with
clause 6.6.6 should be provided in the test specimens. Anchors are inevitably subjected also to
shear.
Clause 6.6.1.1(9) refers to direct tension. Load from a maintenance cradle hanging from
the steel member is an example of how tension may arise. It can also be caused by the
dierential deections of adjacent beams under certain patterns of imposed load, although
the resulting tensions are usually small. Greater tension can be produced near bracings as
identied by clause 6.6.1.1(13). Where tension is present in studs, its design magnitude
should be determined and checked in accordance with clause 6.6.3.2.
Clause 6.6.1.1(10)P is a principle that has led to many application rules. The shear forces
are inevitably concentrated . One research study
76
found that 70%of the shear on a stud was
resisted by its weld collar, and that the local (triaxial) stress in the concrete was several times
its cube strength. Transverse reinforcement performs a dual role. It acts as horizontal shear
reinforcement for the concrete anges, and controls and limits splitting. Its detailing is
particularly critical where connectors are close to a free surface of the slab or where they
are aligned so as to cause splitting in the direction of the slab thickness. To account for
the latter, clause 6.6.1.1(11) should also include a reference to clause 6.6.4 for design of
the transverse reinforcement.
Larger concentrated forces occur where precast slabs are used, and connectors are placed
in groups in holes in the slabs. This inuences the detailing of the reinforcement near these
holes, and is referred to in Section 8.
Clause 6.6.1.1(12) is intended to permit the use of other types of connector. ENV1994-1-1
20
included provisions for many types of connector other than studs: block connectors, anchors,
hoops, angles, and friction-grip bolts. They have all been omitted because of their limited use
and to shorten the code.
Clause 6.6.1.1(12) gives scope, for example, to develop ways to improve current detailing
practice at the ends of beams in fully integral bridges, where forces need to be transferred
abruptly from the composite beams into reinforced concrete piers and abutments. In
British practice, the use of bars with hoops is often favoured in these regions, and design
rules are given in BS 5400-5.
11
The word block rather than bar is used in this Guide to
avoid confusion with reinforcing bars. It is shown in Example 6.8 that the shear resistance
of a connector of this type can be determined in accordance with EN 1992 and EN 1993.
The height of the block should not exceed four times its thickness if the connector is
assumed to be rigid as in Example 6.8.
Clause 6.6.1.1(13) identies a problemthat occurs adjacent to cross-frames or diaphragms
between beams. For multi-beam decks, beams are often braced in pairs such that the bracing
is not continuous transversely across the deck. The presence of bracing locally signicantly
stiens the bridge transversely. Moments and shears in the deck slab are attracted out of the
concrete slab and into the bracing as shown in Fig. 6.27 via the transverse stieners. This
eect is not modelled in a conventional grillage analysis unless the increased stiness in
the location of bracings is included using a shear exible member with inertia and shear
area chosen to match the deections obtained from a plane frame analysis of the bracing
system. Three-dimensional space-frame or nite-element representations of the bridge can
be used to model these local eects more directly.
Clause 6.6.1.1(7)P
Clause 6.6.1.1(8)
Clause 6.6.1.1(9)
Clause
6.6.1.1(10)P
Clause 6.6.1.1(11)
Clause 6.6.1.1(12)
Clause 6.6.1.1(13)
115
The transfer of moment causes tension in the shear connectors on one side of the ange and
induces compression between concrete and ange on the other. Welds at tops of stieners
must also be designed for this moment, which often leads to throat sizes greater than a
nominal 6 mm.
In composite box girders, similar eects arise over the tops of the boxes, particularly at the
locations of ring frames, bracings or diaphragms.
6.6.1.2. Ultimate limit states other than fatigue
In detailing the size and spacing of shear connectors, clause 6.6.1.2 permits the design
longitudinal shear ow to be averaged over lengths such that the peak shear ow within
each length does not exceed the design longitudinal shear resistance per unit length by
more than 10%, and the total design longitudinal shear does not exceed the total design
longitudinal shear within this length. This is consistent with previous practice in the UK
and sometimes avoids the need to alter locally the number or spacings of shear connectors
adjacent to supports. Clause 6.6.1.2 has little relevance to the inelastic lengths in Class 1
and 2 members covered by clause 6.6.2.2(2) since the longitudinal shear is already averaged
over the inelastic zone in this method.
Clause 6.6.1.2
Fig. 6.27. Example of bending moments from a deck slab attracted into bracings
Example 6.8: shear resistance of a block connector with a hoop
Blocks of S235 steel, 250 mmlong and 40 mmsquare, are welded to a steel ange as shown
in Fig. 6.28, at a longitudinal spacing of 300 mm. The resistance to longitudinal shear,
P
Rd
, in a C30/37 concrete is determined. From clause 6.6.1.1(8) the resistance to uplift
should be at least 0.1P
Rd
. This is provided by the 16 mm reinforcing bars shown. None
of the modes of failure involve interaction between concrete and steel, so their own
M
factors are used, rather than 1.25, though the National Annex could decide otherwise.
The concrete is checked to EN 1992, so its denition of f
cd
is used, namely
f
cd
c
cc
f
ck
,
C
. Assuming that the National Annex gives c
cc
1.0 for this situation,
the design strengths of the materials are:
f
yd
! 225 N/mm
2
, f
sd
500,1.15 435 N/mm
2
, f
cd
1 30,1.5 20 N/mm
2
The blocks here are so sti that the longitudinal shear force can be assumed to be
resisted by a uniform stress, o
block
say, at the face of each block. Lateral restraint
enables this stress to exceed f
cd
to an extent given in clause 6.7 of EN 1992-1-1, Partially
loaded areas:
o
block
F
Rdu
,A
c0
f
cd
A
c1
,A
c0
p
3.0f
cd
6.63 in EN 1992-1-1
where A
c0
is the loaded area and A
c1
is the design distribution area of similar shape to
A
c0
, shown in Fig. 6.29 of EN 1992-1-1.
116
Clause 6.7 requires the line of action of the force to pass through the centres of both
areas, but in this application it can be assumed that the force from area b
2
d
2
/2 is resisted
by the face of the block, of area b
1
d
1
/2, because the blocks are designed also to resist uplift.
The dimensions of area A
c1
are xed by clause 6.7 of EN 1992-1-1 as follows:
b
2
b
1
h and b
2
3b
1
; d
2
d
1
h and d
2
3d
1
Here, from Fig. 6.28,
h 300 40 260 mm. b
1
2 40 80 mm. d
1
250 mm
where b
2
240 mm, d
2
510 mm.
40
45 45
120
d
2
= 510
h = 260
d
1
b
2
b
1
16 dia.
160
40
(a) (b)
Fig. 6.28. Block shear connector with hoop, for Example 6.8
A
c1
,A
c0
p

510 240,2,250 40
p
2.47 <3.0
and
F
Rdu
P
Rd
0.250 40 2.47 20 494 kN
The tensile stress in two 16 mm bars from an uplift force of 49.4 kN is 123 N/mm
2
. The
required anchorage length for a hooped bar is given in clause 8.4.4(2) of EN 1992-1-1. It is
proportional to the tensile stress in the bar and depends on its lateral containment, which
in this application is good. For o
sd
123 N/mm
2
the anchorage length is 115 mm, so
120 mm (Fig. 6.28(a)) is sucient.
The welds between the block and the steel ange are designed for the resulting shear,
tension, and bending moment in accordance with EN 1993-1-8. A separate check of
fatigue would be required using EN 1993-1-9 to determine the detail category and
stress range. The comments on clause 6.8.6.2(2) refer. The welds between the bar and
the block are designed for the uplift force.
The resistance given by this method is signicantly less than that from design to BS
5400-5, where the method is based mainly on tests.
77
The above method based on
clause 6.7 of EN 1992-1-1 would strictly require vertical reinforcement to control splitting
from the vertical load dispersal. However, as both push tests and practice have shown this
reinforcement to be unnecessary even when using the higher resistances to BS 5400-5,
vertical reinforcement need not be provided here.
The resistance of a block connector is much higher than that of a shear stud, so where
they are used in haunches, the detailing of reinforcement in the haunch needs attention.
EN 1992-1-1 appears to give no guidance on the serviceability stress limit in a region
where its clause 6.7 is applied. For shear connectors generally, clause 7.2.2(6) refers to
clause 6.8.1(3), where the recommended limit is 0.75P
Rd
. This agrees closely with the
corresponding ratio given in BS 5400-5.
11
117
6.6.2. Longitudinal shear force in beams for bridges
6.6.2.1. Beams in which elastic or non-linear theory is used for resistances of cross-sections
Clause 6.6.2.1(1) requires that the design longitudinal shear force per unit length (the shear
ow) at an interface between steel and concrete is determined from the rate of change of
force in the concrete or the steel. The second part of the clause states, as a consequence of
this, that where elastic bending resistance is used, the shear ow can be determined from
the transverse shear at the cross-section considered. To do this, it is implicit that the beam
is of uniform cross-section such that the usual expression for longitudinal shear ow,
v
L.Ed

V
c.Ed
Az
I
can be used, where:
A is the eective transformed area on the side of the plane concerned that does not include
the centroid of the section, sometimes named the excluded area;
" zz is the distance in the plane of bending from the member neutral axis to the centroid of
area A;
I is the second moment of area of the eective cross-section of the member.
The relevant shear V
c.Ed
is that acting on the composite section. Where the cross-section
varies along its length, the shear ow is no longer directly proportional to the shear on
the beam and the following expression should be used:
v
L.Ed

d
dx
M
c.Ed
Az
I

V
c.Ed
Az
I
M
c.Ed
d
dx
Az
I

D6.18
Equation (D6.18) does not directly cover step changes in the steel cross-section as often
occur at splices. In such situations, it would be reasonable to assume that the step change
occurs uniformly over a length of twice the eective depth of the cross-section when applying
equation (D6.18). Where there is a sudden change from bare steel to a composite section,
design for the concentrated longitudinal shear force from development of composite
action should follow clause 6.6.2.4.
The calculated elastic longitudinal shear ow is strongly dependent on whether or not the
concrete slab is considered to be cracked. In reality, the slab will be stier than predicted by a
fully cracked analysis due to tension stiening. Clause 6.6.2.1(2) claries that the slab should
therefore be considered to be fully uncracked unless tension stiening and over-strength of
concrete are considered in both global analysis and section design as discussed under clause
5.4.2.3(7).
Clause 6.6.2.1(3) requires account to be taken of longitudinal slip where concentrated
longitudinal forces are applied, and refers to clauses 6.6.2.3 and 6.6.2.4. In other cases,
clause 6.6.2.1(3) allows slip to be neglected for consistency with clause 5.4.1.1(8).
Composite box girders
For box girders with a composite ange, a shear ow across the shear connection can occur
due to shear from circulatory torsion, torsional warping and distortional warping. These
eects are discussed in the Designers Guide to EN 1993-2.
4
Clause 6.6.2.1(4) requires
them to be included if appropriate. This inuences the design longitudinal shear stress
(clause 6.6.6.1(5)), and hence the area of transverse reinforcement, to clause 6.6.6.2.
Shear lag and connector slip lead to a non-uniform distribution of force in the shear con-
nectors across the width of the ange. This is discussed in section 9.4 of this Guide.
6.6.2.2. Beams in bridges with cross-sections in Class 1 or 2
Where the bending resistance exceeds the elastic resistance and material behaviour is non-
linear, shear ows can similarly no longer be calculated from linear-elastic section analysis.
To do so using equation (D6.18) would underestimate the shear ow where elastic limits
are exceeded as the lever arm of the cross-section implicit in the calculation would be
overestimated and thus the element forces would be underestimated. Clause 6.6.2.2(1)
Clause 6.6.2.1(1)
Clause 6.6.2.1(2)
Clause 6.6.2.1(3)
Clause 6.6.2.1(4)
Clause 6.6.2.2(1)
118
therefore requires that account be taken of inelastic behaviour of the member and its com-
ponent parts in calculation of longitudinal shear. This diers from previous UK practice but
is more soundly based in theory.
Clause 6.6.2.2(2) gives specic guidance on how to comply with the above requirements
where the concrete slab is in compression. For the length of beam where the bending
moment exceeds M
el.Rd
, the longitudinal shear relationship should be determined from the
change in slab force. The relevant length is that between the points A and C in Fig. 6.29.
The longitudinal shear force in the length AB is determined as the dierence between
slab forces N
c.el
at point A and N
c.d
at point B. Appropriate shear connection to carry this
force is provided within this length. Similar calculation is performed for the length BC.
The spacing of the shear connectors within these lengths is left to the designer. Normally
for lengths AB and BC, and in the absence of heavy point loads, changes of cross-section,
etc., uniform spacing can be used. In a doubtful case, for example within AB, the slab force
N
c
at some point within AB should be determined from the bending moment and appropri-
ate numbers of connectors provided between A and D, say D, and between D and B.
The actual relationship between slab force, N
c
, and moment, M
Ed
, is shown in Fig. 6.30,
together with the approximate expression in Fig. 6.6 of clause 6.2.1.4(6) and the further
simplication of Fig. 6.11. It can be seen that both the approximations are safe for the
design of shear connection, because for a given bending moment, the predicted slab force
exceeds the real slab force.
For inelastic lengths where the slab is in tension, clause 6.6.2.2(3) requires the calculation
of longitudinal shear force to consider the eects of tension stiening and possible over-
strength of the concrete. Failure to do so could underestimate the force attracted to the
shear connection, resulting in excessive slip. As an alternative, clause 6.6.2.2(4) permits
the shear ow to be determined using elastic cross-section analysis based on the uncracked
cross-section. Elastic analysis can be justied in this instance because the conservative
Clause 6.6.2.2(2)
Clause 6.6.2.2(3)
Clause 6.6.2.2(4)
A B C
M
x
M
Ed
L
AB
L
BC
M
el,Rd
0
M
pl,Rd
Fig. 6.29. Denition of inelastic lengths for Class 1 and 2 cross-sections with the slab in compression
1.0
True behaviour
Idealisation in Fig. 6.6
Idealisation in Fig. 6.11
0 N
c,el
/N
c,f
M
a,Ed
/M
pl,Rd
M
Ed
/M
pl,Rd
M
el,Rd
/M
pl,Rd
1.0
N
c
/N
c,f
Fig. 6.30. Variation of longitudinal force in a concrete ange with bending moment
119
assumption of fully uncracked concrete osets the slightly unconservative neglect of
inelasticity.
6.6.2.3. Local eects of concentrated longitudinal shear force due to introduction of longitudinal forces
Where concentrated longitudinal forces are applied to a composite section, the stress state
can easily be determined some distance away from the point of application from considera-
tions of equilibrium and from the usual assumption that plane sections remain plane. Plane
sections do not, however, remain plane in the vicinity of the force, and accurate determina-
tion of the length over which longitudinal shear transfers between concrete slab and steel
ange together with the magnitude of the peak longitudinal shear ow requires complex
analysis. This clause is based on parametric nite-element analyses
78
and existing practice.
Clause 6.6.2.3 provides simple rules for the determination of the design shear ow between
steel and concrete where there is a concentrated longitudinal force, F
Ed
, applied to the
concrete slab. Clause 6.6.2.3(2) distinguishes between forces applied within the length of
the member and those applied at ends of the members. In the former case, the length over
which the force is distributed, L
v
, is equal to the eective width for global analysis, b
eff
,
plus e
d
, which is the loaded length plus twice the lateral distance fromthe point of application
of the force to the web centreline: L
v
b
eff
e
d
.
For forces applied at an end of a concrete ange, the distribution length is half of the
above. The reference to eective width for global analysis means that the simple provisions
of clause 5.4.1.2(4) can be used, rather than the eective width appropriate to the cross-
section where the force is applied.
The force cannot, in general, be transferred uniformly by the shear connection over the
above lengths and Fig. 6.12(a) and (b) shows the distribution to be used, leading to equa-
tions (6.12) and (6.13). Where stud shear connectors are used, these are suciently ductile
to permit a uniform distribution of shear ow over the above lengths at the ultimate limit
state. This leads to equations (6.14) and (6.15) in clause 6.6.2.3(3). For serviceability or
fatigue limit states, the distributions of equations (6.12) and (6.13) should always be used.
The shear force V
L.Ed
transferred to the shear connection is not F
Ed
, as can be seen from
Fig. 6.31 for the case of a force applied to the end of the concrete slab. Force V
L.Ed
is the
dierence between F
Ed
and the force N
c
in the concrete slab where dispersal of F
Ed
into
the cross-section is complete.
Clause 6.6.2.3(4) allows the dispersal of the force F
Ed
V
L.Ed
(which for load applied to
the concrete as shown in Fig. 6.31 is equal to N
c
) into either the concrete or steel element to
be based on an angle of spread of 2u where u is tan
1
2/3. This is the same spread angle used
in EN1992-1-1 clause 8.10.3 for dispersal of prestressing force into concrete. It is slightly less
than the dispersal allowed by clause 3.2.3 of EN 1993-1-5 for the spread through steel
elements.
6.6.2.4. Local eects of concentrated longitudinal shear forces at sudden change of cross-sections
Clause 6.6.2.4 provides simple rules for the determination of the design shear ow between
steel and concrete at ends of slabs where:
.
the primary eects of shrinkage or dierential temperature are developed (clause
6.6.2.4(1))
Clause 6.6.2.3
Clause 6.6.2.3(2)
Clause 6.6.2.3(3)
Clause 6.6.2.3(4)
Clause 6.6.2.4
Clause 6.6.2.4(1)
N
c
G
a
N
a
= V
L,Ed
F
Ed
V
L,Ed
= F
Ed
N
c
L
v
/2 = b
eff
/2 + e
d
/2
M
c
M
a
Fig. 6.31. Determination of V
L.Ed
for a concentrated force applied at an end of the slab
120
.
there is an abrupt change of cross-section (clause 6.6.2.4(2)), such as that shown in
Fig. 6.33.
The shear V
L.Ed
transferred across the concrete and steel interface due to shrinkage or
temperature may be assumed to be distributed over a length equal to b
eff
, as discussed in
section 6.6.2.3 above. Generally, clause 6.6.2.4(3) requires the distribution of this force to
be triangular as shown in Fig. 6.12(c), which leads to equation (6.16). Where stud shear
connectors are used, these are again suciently ductile to permit a uniform distribution of
shear ow over the length b
eff
. This leads to a design shear ow of V
L.Ed
,b
eff
.
A calculation of primary shrinkage stresses is given in Example 5.3 (Fig. 5.11). The force
N
c
is found from these. It equals the shear force V
L.Ed
, which is transferred as shown in
Fig. 6.32.
The determination of V
L.Ed
caused by bending at a sudden change in cross-section is
shown in Fig. 6.33 for the case of propped construction, in which the total moment at
cross-section BB is:
M
Ed.B
M
a
M
c
N
c
z
For unpropped construction, the stress would not vary linearly across the composite
section as shown in Fig. 6.33, but the calculation of longitudinal shear from the force in
the slab would follow the same procedure.
The length over which the force is distributed and the shape of distribution may be taken
according to clause 6.6.2.4(5) to be the same as that given in clause 6.6.2.4(3).
6.6.3. Headed stud connectors in solid slabs and concrete encasement
Resistance to longitudinal shear
In BS 5400 Part 5
11
and in earlier UK codes, the characteristic shear resistances of studs are
given in a table, applicable only when the stud material has particular properties. There was
no theoretical model for the shear resistance.
The Eurocodes must be applicable to a wider range of products, so design equations are
essential. Those given in clause 6.6.3.1(1) are based on the model that a stud with shank dia-
meter d and ultimate strength f
u
, set in concrete with characteristic strength f
ck
and mean
secant modulus E
cm
, fails either in the steel alone or in the concrete alone. The concrete
failure is found in tests to be inuenced by both the stiness and the strength of the concrete.
Clause 6.6.2.4(2)
Clause 6.6.2.4(3)
Clause 6.6.2.4(5)
Clause 6.6.3.1(1)
T
C
T
N
c
b
eff
M
c
M
a
G
a
N
a
= N
c
V
L,Ed
= N
c
L.Ed
for primary shrinkage at an end of a beam
B A
z
B
N
c
b
eff
M
c
M
Ed,A
M
a
G
a
N
a
= N
c
= V
L,Ed
M
Ed,B
V
L,Ed
= N
c
L.Ed
caused by bending moment at a sudden change of cross-section
121
This led to equations (6.18) to (6.21), in which the numerical constants and partial safety
factor
V
have been deduced from analyses of test data. In situations where the resistances
from equations (6.18) and (6.19) are similar, tests show that interaction occurs between
the two assumed modes of failure. An equation based on analyses of test data, but not on
a dened model
79
P
Rd
kd
2
,4 f
u
E
cm
,E
a
0.4
f
ck
,f
u
0.35
D6.19
gives a curve with a shape that approximates better both to test data and to values tabulated
in BS 5400.
In the statistical analyses done for EN 1994-1-1
80.81
both of these methods were studied.
Equation (D6.19) gave results with slightly less scatter, but the equations of clause
6.6.3.1(1) were preferred because of their clear basis and experience of their use in some
countries. Here, and elsewhere in Section 6, coecients from such analyses were modied
slightly, to enable a single partial factor, denoted
V
(V for shear), 1.25, to be recommended
for all types of shear connection. This value has been used in draft Eurocodes for over 20
years.
It was concluded from this study
81
that the coecient in equation (6.19) should be 0.26.
This result was based on push tests, where the mean number of studs per specimen was
only six, and where lateral restraint from the narrow test slabs was usually less sti than
in the concrete ange of a composite beam. Strength of studs in many beams is also increased
by the presence of hogging transverse bending of the slab. For these reasons the coecient
was increased from0.26 to 0.29, a value that is supported by a subsequent calibration study
60
based on beams with partial shear connection.
Design resistances of 19 mm stud connectors in solid slabs, given by clause 6.6.3.1, are
shown in Fig. 6.34. It is assumed that the penalty for short studs, equation (6.20), does
not apply. For any given values of f
u
and f
ck
, the gure shows which failure mode
governs. It can be used for this purpose for studs of other diameters, provided that
h,d ! 4. The reference to the slabs as solid means that they are not composite slabs cast
on proled steel sheeting. It does not normally exclude haunched slabs.
The overall nominal height of a stud, used in equations (6.20) and (6.21), is about 5 mm
greater than the length after welding, a term which is also in use.
Weld collars
Clause 6.6.3.1(2) on weld collars refers to EN 13918,
40
which gives guide values for the
height and diameter of collars, with the note that these may vary in through-deck stud
welding. It is known that for studs with normal weld collars, a high proportion of the
shear is transmitted through the collar.
76
It should not be assumed that the shear resistances
of clause 6.6.3.1 are applicable to studs without collars, as noted by clause 6.6.3.1(4) (e.g.
Clause 6.6.3.1(2)
100
80
60
20 30 40
50
Normal-density concrete
Density class 1.8
450
400
25 37 50
60
f
u
= 500 N/mm
2
f
ck
(N/mm
2
)
P
Rd
(kN)
f
cu
(N/mm
2
)
Fig. 6.34. Design shear resistances of 19 mm studs with h,d ! 4 in solid slabs
122
where friction welding by high-speed spinning is used). A normal collar should be fused to
the shank of the stud. Typical collars in the test specimens from which the design formulae
were deduced had a diameter not less than 1.25d and a minimum height not less than 0.15d,
where d is the diameter of the shank.
Splitting of the slab
Clause 6.6.3.1(3) refers to splitting forces in the direction of the slab thickness. These occur
where the axis of a stud lies in a plane parallel to that of the concrete slab; for example, if
studs are welded to the web of a steel T-section that projects into a concrete ange. These
are referred to as lying studs in published research
82
on the local reinforcement needed
to prevent or control splitting. Comment on the informative Annex C on this subject is
given in Chapter 10. A similar problem occurs in composite L-beams with studs close to a
free edge of the slab. This is addressed in clause 6.6.5.3(2).
Tension in studs
Pressure under the head of a stud connector and friction on the shank normally causes
the stud weld to be subjected to vertical tension before shear failure is reached. This is
why clause 6.6.1.1(8) requires shear connectors to have a resistance to tension that is at
least 10% of the shear resistance. Clause 6.6.3.2(2) therefore permits tensile forces that
are less than this to be neglected. (The symbol F
ten
in this clause means F
Ed.ten
.)
Resistance of studs to higher tensile forces has been found to depend on so many variables,
especially the layout of local reinforcement, that no simple design rules could be given.
Relevant evidence from about 60 tests on 19 mm and 22 mm studs is presented in Ref. 74,
which gives a best-t interaction curve. In design terms, this becomes
F
ten
,0.85P
Rd
5,3
P
Ed
,P
Rd
5,3
1 D6.20
Where the vertical tensile force F
ten
0.1P
Rd
, this gives P
Ed
0.93P
Rd
, which is plausible.
Expression (D6.20) should be used with caution, because some studs in these tests had ratios
h,d as high as 9; but on the conservative side, the concrete blocks were unreinforced.
6.6.4. Headed studs that cause splitting in the direction of the slab thickness
There is a risk of splitting of the concrete where the shank of a stud (a lying stud) is parallel
and close to a free surface of the slab, as shown, for example, in Fig. 6.35. Where the
conditions of clause 6.6.4(1) to (3) are met, the stud resistances of clause 6.6.3.1 may still
be used. The geometric requirements are shown in Fig. 6.35, in which d is the diameter of
the stud. A further restriction is that the stud must not also carry shear in a direction trans-
verse to the slab thickness. The example shown in Fig. 6.35 would not comply in this respect
unless the steel section were designed to be loaded on its bottom ange. Clause 6.6.4(3)
requires that the stirrups shown should be designed for a tensile force equal to 0.3P
Rd
per
stud connector. This is analogous with the design of bursting reinforcement at prestressing
anchorages. The true tensile force depends on the slab thickness and spacing of the studs
and the proposed value is conservative for a single row of studs. No recommendation is
given here, or in Annex C, on the design of stirrups where there are several rows of studs.
Clause 6.6.3.1(3)
Clause 6.6.3.2(2)
Clause 6.6.4(1)
to (3)
(b) (a)
s
A
A
View AA
P
Ed
e
v
6d
e
v
6d
v 14d
s, 18d
Fig. 6.35. Examples of details susceptible to longitudinal splitting
123
Some details which do not comply with clause 6.6.4 can be designed using the rules in the
informative Annex C, if its use is permitted by the National Annex. In Fig. 6.35, the eects of
local loading on the slab and of U-frame action will also cause moment at the shear connec-
tion which could cause stud tensions in excess of those allowed by clause 6.6.3.2. This detail is
therefore best avoided.
Planes of type aa such as section AA in Fig. 6.35 should be provided with longitudinal
shear reinforcement in accordance with clause 6.6.6.
6.6.5. Detailing of the shear connection and inuence of execution
It is rarely possible to prove the general validity of application rules for detailing, because
they apply to so great a variety of situations. They are based partly on previous practice.
An adverse experience causes the relevant rule to be made more restrictive. In research, exist-
ing rules are often violated when test specimens are designed, in the hope that extensive good
experience may enable existing rules to be relaxed.
Rules are often expressed in the form of limiting dimensions, even though most behaviour
(excluding corrosion) is more inuenced by ratios of dimensions than by a single value.
Minimum dimensions that would be appropriate for an unusually large structural member
could exceed those given in the code. Similarly, code maxima may be too large for use in
a small member. Designers are unwise to follow detailing rules blindly, because no set of
rules can be comprehensive.
Resistance to separation
The object of clause 6.6.5.1(1) on resistance to separation is to ensure that failure surfaces in
the concrete cannot pass above the connectors and below the reinforcement, intersecting
neither. Tests have found that these surfaces may not be plane; the problem is three-
dimensional. A longitudinal section through a possible failure surface ABC is shown in
Fig. 6.36. The studs are at the maximum spacing allowed by clause 6.6.5.5(3).
Clause 6.6.5.1 denes only the highest level for the bottom reinforcement. Ideally, its
longitudinal location relative to the studs should also be dened, because the objective is
to prevent failure surfaces where the angle c (Fig. 6.36) is small. It is impracticable to link
detailing rules for reinforcement with those for connectors, or to specify a minimum for
angle c. In Fig. 6.36, it is less than 88, which is much too low.
The angle cobviously depends onthe level of the bottombars, the height of the studs, andthe
spacing of both the bars and the studs. Studs ina bridge deck usually have a lengthafter welding
(LAW) that exceeds the 95mmshown. Assuming LAW 120 mm, maximumspacings of both
bars and studs of 450 mm, and a bottomcover of 50mmgives c ! 178, approximately, which is
suggested here as a minimum. Studs may need to be longer than 125 mm where permanent
formwork is used, as this raises the level of the bottom reinforcement.
Other work reached a similar conclusion in 2004.
83
Referring to failure surfaces as shown
in Fig. 6.36, it was recommended that angle c should be at least 158. In this paper the line AB
(Fig. 6.36) is tangential to the top of the bar at B, rather than the bottom, slightly reducing its
slope.
Concreting
Clause 6.6.5.2(1)P requires shear connectors to be detailed so that the concrete can be
adequately compacted around the base of the connector. This necessitates the avoidance
Clause 6.6.5.1(1)
Clause 6.6.5.2(1)P
800
55
95 30
A
B
C
Fig. 6.36. Level of bottom transverse reinforcement (dimensions in mm)
124
of excessively close spacings of connectors and the use of connector geometries that might
prevent adequate ow of the concrete around the connector. The former could be a consid-
eration at the ends of fully integral bridges where a very high shear ow has to be transferred
into the steel beam over a relatively short length. Since the resistances of connectors other
than studs are not covered by EN 1994-2, properties of other types of connector could be
referred to from a National Annex. The design of a block-and-hoop connector is illustrated
in Example 6.8. A novel type of connection could be investigated as part of the testing
requirements of clause 6.6.1.1(12).
Loading of shear connection during execution
Clause 6.6.5.2(3) particularly concerns the staged casting of concrete anges for typical
unpropped composite bridges. Partly matured concrete around shear connectors in a
recently cast length of beam could possibly be damaged by the eects of concreting
nearby. The recommended lower limit on concrete strength, 20 N/mm
2
, in eect sets a
minimum time interval between successive stages of casting. The rule begins Wherever
possible because there appears to be no evidence of damage from eects of early thermal
or shrinkage strains, which also apply longitudinal shear to young concrete.
In propped construction, it would be unusual to remove the props until the concrete had
achieved a compressive strength of at least 20 N/mm
2
, in order to avoid overstressing the
beam as a whole. Where the props are removed prior to the concrete attaining the specied
strength, verications at removal of props should be based on an appropriately reduced
compressive strength.
Local reinforcement in the slab
Where shear connectors are close to a longitudinal edge of a concrete ange, use of U-bars is
almost the only way of providing the full anchorage required by clause 6.6.5.3(1). The
splitting referred to in clause 6.6.5.3(2) is a common mode of failure in push-test specimens
with narrow slabs (e.g. 300 mm, which has long been the standard width in British codes). It
was also found, in full-scale tests, to be the normal failure mode for composite L-beams
constructed with precast slabs.
84
Detailing rules are given in clause 6.6.5.3(2) for slabs where
the edge distance e in Fig. 6.37 is less than 300 mm. The required area of bottom transverse re-
inforcement, A
b
per unit length of beam, should be found using clause 6.6.6. In the unhaunched
slab shown in Fig. 6.37, failure surface bb will be critical (unless the slab is very thick) because
the shear on surface aa is low in an L-beam with an asymmetrical concrete ange.
To ensure that the reinforcement is fully anchored to the left of the line aa, it is recom-
mended that U-bars be used. These can be in a horizontal plane or, where top reinforcement
is needed, in a vertical plane.
Reinforcement at the end of a cantilever
At the end of a composite cantilever, the force on the concrete from the connectors acts
towards the nearest edge of the slab. The eects of shrinkage and temperature can add
further stresses
74
that tend to cause splitting in region B in Fig. 6.38, so reinforcement in
this region needs careful detailing. Clause 6.6.5.3(3)P can be satised by providing herring-
bone bottom reinforcement (ABC in Fig. 6.38) sucient to anchor the force from the
connectors into the slab, and ensuring that the longitudinal bars provided to resist that
force are anchored beyond their intersection with ABC.
Clause 6.6.5.2(3)
Clause 6.6.5.3(1)
Clause 6.6.5.3(2)
Clause 6.6.5.3(3)P
b
b
a
a
d
e(6d )
0.5d
Fig. 6.37. Longitudinal shear reinforcement in an L-beam
125
Haunches
Haunches are sometimes provided in composite bridges to cater for drainage cross-falls so
that the thickness of the slab or deck surfacing need not be varied. The detailing rules of
clause 6.6.5.4 are based on limited test evidence, but are long-established.
85
In regions of
high longitudinal shear, deep haunches should be used with caution because there may be
little warning of failure.
Maximum spacing of connectors
Situations where the stability of a concrete slab is ensured by its connection to a steel beam
are unlikely to occur because a concrete slab that is adequate to resist local bridge loading is
unlikely to suer instability from membrane forces. The converse situation, stabilization of
the steel ange, is of interest only where the steel compression ange is not already in Class 1
or 2. Where the steel beam is a plate girder, its proportions will often be chosen such that it is
in Class 3 for the bare steel condition during construction. This maximises the lateral buck-
ling resistance for a ange of given cross-sectional area.
Clause 6.6.5.5(2) is not restrictive in practice. As an example, a plate girder is considered,
in steel with f
y
355 N/mm
2
, where the top ange has t
f
20 mm, an overall breadth of
350 mm, and an outstand c of 165 mm. The ratio is 0.81 and the slenderness is:
c,t
f
165,20 0.81 10.2
so from Table 5.2 of EN 1993-1-1, the ange is in Class 3. From clause 6.6.5.5(2), it can be
assumed to be in Class 1 if shear connectors are provided within 146 mm of each free edge, at
longitudinal spacing not exceeding 356 mm, for a solid slab.
The ratio 22 in this clause is based on the assumption that the steel ange cannot buckle
towards the slab. Where there are transverse ribs (e.g. due to the use of proled sheeting), the
assumption may not be correct, so the ratio is reduced to 15. The maximum spacing in this
example is then 243 mm.
Further requirements for composite plates in box girders are given in clause 9.4(7). These
also cover limitations on longitudinal and transverse spacings of connectors to ensure Class 3
behaviour. The rule on transverse spacing in Table 9.1 should be applied also to a wide
compression ange of a plate girder.
The maximum longitudinal spacing in bridges, given in clause 6.6.5.5(3), 4h
c
but
800 mm, is more liberal than the equivalent rule of BS 5400 Part 5.
11
It is based mainly
on behaviour observed in tests, and on practice with precast slabs in some countries.
Clause 6.6.5.5(4) allows the spacing rules for individual connectors to be relaxed if
connectors are placed in groups. This may facilitate the use of precast deck units with discrete
pockets for the shear connection (clause 8.4.3(3) refers) but many of the deemed-to-satisfy
rules elsewhere in EN 1994-2 then no longer apply. The designer should then explicitly con-
sider the relevant eects, which will make it dicult in practice to depart fromthe application
rules. The eects listed are as follows.
.
Non-uniform ow of longitudinal shear. If the spacing of the groups of connectors is
large compared to the distance between points of zero and maximum moment in the
beam, then the normal assumption of plane sections remaining plane will not apply
and the calculation of bending resistance to clause 6.2 will not be valid.
Clause 6.6.5.4
Clause 6.6.5.5(2)
Clause 6.6.5.5(3)
Clause 6.6.5.5(4)
Studs
Steel beam
A
C
B
Transverse reinforcement
not shown
Slab
Fig. 6.38. Reinforcement at the end of a cantilever
126
.
Greater risk of slip and vertical separation of concrete and steel. The latter carries a
corrosion risk for the steel ange which would be hard to quantify without testing.
.
Buckling of the steel ange. This can be considered by applying clause 4.4(2) of EN 1993-
1-5. The elastic critical buckling stress,
cr
, can be determined for the discrete supports
oered by the particular connection provided, either by nite-element analysis or from
standard texts such as Ref. 86. If the latter method is employed, account needs to be
taken of the benecial restraint provided by the concrete against buckling. In the
absence of this restraint, the ange would try to buckle in half wavelengths between
the studs, alternating towards and away from the concrete. Buckling into the concrete
is, in reality, prevented and therefore no rotation of the ange can occur along the line
of the studs. Discrete supports which clamp the plate at the stud locations may therefore
usually be assumed in determining the critical stress.
.
Local resistance of the slab to the concentrated force from the connectors. Groups of
studs apply a force analogous to that from an anchorage of a prestressing cable. The
region of transverse tension does not coincide with the location of the group. Both the
quantity and the location of the transverse reinforcement required may dier from
that given by clause 6.6.6.
Dimensions of the steel ange
The rules of clause 6.6.5.6 are intended to prevent local overstress of a steel ange near a
shear connector and to avoid problems with stud welding. Application rules for minimum
ange thickness are given in clause 6.6.5.7(3) and (5). The minimum edge distance for
connectors in clause 6.6.5.6(2) is consistent with the requirements of BS 5400:Part 5.
11
The limit is also necessary to avoid a reduction in the fatigue detail category for the ange
at the stud location, which from EN 1993-1-9 Table 8.4, is 80 for a compliant stud.
Headed stud connectors
Clause 6.6.5.7(1) and (2) is concerned with resistance to uplift. Rules for resistance of studs,
minimum cover and projection of studs above bottom reinforcement usually lead to the use
of studs of height greater than 3d.
The limit 1.5 for the ratio d=t
f
in clause 6.6.5.7(3) could, in principle, inuence the design
of shear connection for closed-top box girders in bridges, eectively requiring the thickness
of the top ange to be a minimum of 13 mm where 19 mm diameter studs are used. Studs
of this size are preferred by many UK fabricators as they can be welded manually. It is
unlikely that a composite box would have a ange this thin as it would then require
considerable longitudinal stiening to support the plate prior to setting of the concrete.
Clause 6.6.5.7(3) would also apply to plate girders adjacent to cross-braces or transverse
diaphragms, but practical sizes of anges for main beams will satisfy this criterion. It
could be a consideration for cross-girders.
In clause 6.6.5.7(4), the minimum lateral spacing of studs in solid slabs has been reduced
to 2.5d, compared with the 4d of BS 5950-3-1. This facilitates the use of precast slabs
supported on the edges of the steel anges, with projecting U-bars that loop over pairs of
studs. Closely spaced pairs of studs must be well conned laterally. The words solid
slabs should therefore be understood here to exclude haunches.
6.6.6. Longitudinal shear in concrete slabs
The subject of clause 6.6.6 is the avoidance of local failure of a concrete ange near the shear
connection, by the provision of appropriate reinforcement. These bars enhance the resistance
of a thin concrete slab to in-plane shear in the same way that stirrups strengthen a concrete
web in vertical shear. Transverse reinforcement is also needed to control and limit the long-
itudinal splitting of the slab that can be caused by local forces from individual connectors. In
this respect, the detailing problem is more acute than in the anges of concrete T-beams,
where the shear from the web is applied more uniformly.
The principal change from earlier codes is that the equations for the required area of
transverse reinforcement have been replaced by cross-reference to EN 1992-1-1. Its
Clause 6.6.5.6
Clause 6.6.5.6(2)
Clause 6.6.5.7(1)
Clause 6.6.5.7(2)
Clause 6.6.5.7(3)
Clause 6.6.5.7(4)
Clause 6.6.6
127
provisions are based on a truss analogy, as before, but a more general version of it, in which
the angle between members of the truss can be chosen by the designer. It is an application of
strut-and-tie modelling, which is widely used in EN 1992.
There is, however, a signicant dierence between the application of EN 1992 and
EN 1994. In the latter, the transverse reinforcement may be placed according to the distribu-
tion of vertical shear force envelope, or according to the stud forces for sections where the
elastic resistance moment is exceeded. In the former, the transverse reinforcement should
be placed according to the location of the web compression struts as they intersect the
anges and their subsequent continuation into the anges.
The denitions of shear surfaces in clause 6.6.6.1(2)P and the basic design method are as
before. The method of presentation reects the need to separate the general provisions,
clauses 6.6.6.1 to 3, from those restricted to buildings, in EN 1994-1-1 clause 6.6.6.4.
Clause 6.6.6.1(4) requires the design longitudinal shear to be consistent with that used for
the design of the shear connectors. This means that the distribution along the beam of
resistance to in-plane shear in the slab should be not less than that assumed for the design
of the shear connection. For example, uniform resistance to longitudinal shear ow (v
L
)
should be provided where the connectors are uniformly spaced, even if the vertical shear
over the length is not constant. It does not mean, for example, that if, for reasons concerning
detailing, v
L.Rd
1.3v
L.Ed
for the connectors, the transverse reinforcement must provide the
same degree of over-strength.
The reference to variation of longitudinal shear across the width of the concrete ange
means that transverse reinforcement could be reduced away from the beam centre-lines,
where the longitudinal shear reduces, if exural requirements permit.
In applying clause 6.6.6.1(5), it is suciently accurate to assume that longitudinal bending
stress in the concrete ange is constant across its eective width, and zero outside it. The
clause is relevant, for example, to nding the shear on plane aa in the haunched beam
shown in Fig. 6.15, which, for a symmetrical ange, is less than half the shear resisted by
the connectors.
Resistance of a concrete ange to longitudinal shear
Clause 6.6.6.2(1) refers to clause 6.2.4 of EN 1992-1-1, which is written for a design longi-
tudinal shear stress v
Ed
acting on a cross-section of thickness h
f
. This must be distinguished
from the design longitudinal shear ow v
L.Ed
used in EN 1994 which is equal to v
Ed
h
f
. The
clause requires the area of transverse reinforcement A
sf
at spacing s
f
to satisfy
A
sf
f
yd
,s
f
v
Ed
h
f
, cot 0
f
6.21 in EN 1992-1-1
and the longitudinal shear stress to satisfy
v
Ed
< if
cd
sin 0
f
cos 0
f
6.22 in EN 1992-1-1
where i 0.61 f
ck
,250, with f
ck
in N/mm
2
units. (The Greek letter i (nu) used here in
EN 1992-1-1 should not be confused with the Roman letter v (vee), which is used for
shear stress.)
The angle 0
f
between the diagonal strut and the axis of the beam is chosen (within limits)
by the designer. It should be noted that the recommended limits depend on whether the
ange is in tension or compression, and can be varied by a National Annex.
EN 1992-1-1 does not specify the distribution of the required transverse reinforcement
between the upper and lower layers in the slab. It was a requirement of early drafts of
EN 1992-2 that the transverse reinforcement provided should have the same centre of
resistance as the longitudinal force in the slab. This was removed, presumably because it
has been common practice to consider the shear resistance to be the sum of the resistances
from the two layers. Clause 6.6.6.2(3) refers to Fig. 6.15 which claries that the reinforce-
ment to be considered on plane aa for composite beams is the total of the two layers,
A
b
A
t
. It should be noted that application of Annex MM of EN 1992-2 would necessitate
provision of transverse reinforcement with the same centre of resistance as the longitudinal
force in the slab.
Clause 6.6.6.1(2)P
Clause 6.6.6.1(4)
Clause 6.6.6.1(5)
Clause 6.6.6.2(1)
128
The use of the method is illustrated in Example 6.9.
The reference in clause 6.6.6.2(1) is to the whole of clause 6.2.4 of EN 1992-1-1. This
includes clause 6.2.4(5) which requires a check of the interaction with transverse bending
in the concrete slab on planes of type aa. This requires the amount of transverse reinforce-
ment to be the greater of that required for longitudinal shear alone and half that required for
longitudinal shear plus that required for transverse bending.
This rule is illustrated in Fig. 6.39(b), in which A
req
.
d
is the total reinforcement required,
and subscripts s and b refer to the reinforcement required for shear and bending, respectively.
The rule is more onerous than that in BS 5400 Part 5, where no interaction is needed to be
considered on planes of type aa. This was because a transverse bending moment produces
no net axial force in the slab. Reinforcement in tension is then associated with concrete that
has a resistance to shear enhanced by the presence of transverse compression. Test evidence
supported this model.
Consideration of this interaction according to EN1992-1-1, rather than simple addition of
bending and shear requirements, is rational as the truss angle in the upper and lower layers of
the slab can be varied to account for the relative demands imposed on them of shear and
either direct tension or compression. Annex MM of EN 1992-2 reinforces this, although
its use suggests that the 50% of longitudinal shear rule is optimistic. The interaction on
surfaces around the studs diers slightly as discussed below under the heading Shear
planes and surfaces.
Clause 6.2.4(105) of EN 1992-2 adds an additional interaction condition for the concrete
struts, which was largely due to the strength of belief in its Project Team that the neglect of
this check was unsafe. It was less to do with any inherent dierence between the behaviour of
buildings and bridges. Considerations of equilibrium suggest that EN 1992-2 is correct. The
reference in EN 1994-2 to EN 1992-1-1 rather than to EN 1992-2 is therefore signicant as it
excludes the check. No check of the eect of moment on the concrete struts was required by
BS 5400 Part 5.
Clause 6.2.4(105) of EN 1992-2 refers to the compression from transverse bending. It is
however equally logical to consider the slab longitudinal compression in the check of the
concrete struts, although this was not intended. Once again, the use of Annex MM of
EN 1992-2 would require this to be considered. It is rare in practice for a concrete ange
to be subjected to a severe combination of longitudinal compression and in-plane shear. It
could occur near an internal support of a continuous half-through bridge. It is shown in
Example 6.11 that the reduction in resistance to bending would usually be negligible, if
this additional interaction were to be considered.
Neither EN 1992 nor EN 1994 deals with the case of longitudinal shear and coexistent
transverse tension in the slab. This can occur in the transverse beams of ladder decks near
the intersection with the main beams in hogging zones where the main beam reinforcement
is in global tension. In such cases, there is clearly a net tension in the slab and the reinforce-
ment requirements for this tension should be fully combined with that for longitudinal shear
on planes aa.
Shear planes and surfaces
Clause 6.6.6.1(3) refers to shear surfaces because some of them are not plane. Those of
types bb, cc and dd in Fig. 6.15 are dierent from the type aa surface because they
resist almost the whole longitudinal shear, not (typically) about half of it. The relevant
reinforcement intersects them twice, as shown by the factors 2 in the table in Fig. 6.15.
For a surface of type cc in a beam with the steel section near one edge of the concrete
ange, it is clearly wrong to assume that half the shear crosses half of the surface cc.
However, in this situation the shear on the adjacent plane of type aa will govern, so the
method is not unsafe.
Clause 6.6.6.2(2) refers to EN 1992-1-1 clause 6.2.4(4) for resistance to longitudinal
shear for surfaces passing around the shear studs. This does not include a check of the
interaction with transverse bending in the slab. This does not mean that such a check may
be ignored.
Clause 6.6.6.2(2)
129
Interaction of longitudinal shear and bending should be considered for the reinforcement
crossing shear surfaces around the connectors. Slab bottom reinforcement is particularly
important since most of the shear transferred by stud connectors is transferred over the
bottom part of the stud. The bottom reinforcement crossing surfaces passing around the
studs must control the localized splitting stresses generated by the high stud pressures. For
surfaces of type bb and cc in Fig. 6.15, it is clear that the reinforcement crossing the
surface must provide both the resistance to longitudinal shear and any transverse sagging
moment present. The reinforcement requirements for coexisting shear and sagging
moment should therefore be fully added.
For coexisting shear and hogging moment, there is transverse compression at the base of
the connectors. BS 5400 Part 5 permitted a corresponding reduction in the bottom reinforce-
ment. This could be the basis of a more accurate calculation permitted by clause 6.6.6.2(2).
For surfaces of type dd in Fig. 6.15, the haunch reinforcement crossed by the studs will
not be considered in the sagging bending resistance and therefore it need only be designed for
longitudinal shear. These recommendations are consistent with those in BS 5400 Part 5.
Minimum transverse reinforcement
Clause 6.6.6.3 on this subject is discussed under clause 6.6.5.1 on resistance to separation. Clause 6.6.6.3
Example 6.9: transverse reinforcement for longitudinal shear
Figure 6.39(a) shows a plan of an area ABCDof a concrete ange of thickness h
f
, assumed
to be in longitudinal compression, with shear stress v
Ed
and transverse reinforcement A
sf
at spacing s
f
. The shear force per transverse bar is:
F
v
v
Ed
h
f
s
f
acting on side AB of the rectangle shown. It is transferred to side CD by a concrete strut
AC at angle 0
f
to AB, and with edges that pass through the mid-points of AB, etc., as
shown, so that the width of the strut is s
f
sin 0
f
.
A B
C D
F
v
s
f
A
reqd
/A
s
A
b
/A
s
F
v
F
c
F
t
s
f
sin
f
f
0
2
1
1
0.5
0.5
(b) (a)
Fig. 6.39. In-plane shear in a concrete ange: (a) truss analogy; and (b) interaction of longitudinal
shear with bending
For equilibrium at A, the force in the strut is:
F
c
F
v
sec 0
f
D6.21
For equilibrium at C, the force in the transverse bar BC is:
F
t
F
c
sin 0
f
F
v
tan 0
f
D6.22
For minimum area of transverse reinforcement, 0
f
should be chosen to be as small as pos-
sible. For a ange in compression, the limits to 0
f
given in clause 6.2.4(4) of EN1992-1-1 are:
458 ! 0
f
! 26.58 D6.23
130
so the initial choice for 0
f
is 26.58. Then, from equation (D6.22),
F
t
0.5F
v
D6.24
From equation (6.22) in EN 1992-1-1, above,
v
Ed
< 0.40if
cd
If this inequality is satised, then the value chosen for 0
f
is satisfactory. However, let us
assume that the concrete strut is overstressed, because v
Ed
0.48if
cd
.
To satisfy equation (6.22), sin 0
f
cos 0
f
! 0.48, whence 0
f
! 378.
The designer increases 0
f
to 408, which satises expression (D6.23).
From equation (D6.22), F
t
F
v
tan 408 0.84F
v
.
From equation (D6.24), the change in 0
f
, made to limit the compressive stress in the
concrete strut AC, increases the required area of transverse reinforcement by 68%.
The lengths of the sides of the triangle ABC in Fig. 6.39(a) are proportional to the
forces F
v
, F
t
and F
c
. For given F
v
and s
f
, increasing 0
f
increases F
c
, but for 0
f
< 458
(the maximum value recommended), the increase in the width s
t
sin 0
f
is greater, so the
stress in the concrete is reduced.
Example 6.10: longitudinal shear checks
The bridge shown in Fig. 6.22 is checked for longitudinal shear, (i) adjacent to an internal
support, (ii) at mid-span of the main span and (iii) at an end support. Creep of concrete
reduces longitudinal shear, so for elastic theory, the short-term modular ratio is used.
(i) Adjacent to an internal support
The 19 150 mm stud connectors are assumed to be 145 mm high after welding. Their
arrangement and the transverse reinforcement are shown in Fig. 6.40 and the dimensions
of the cross-section in Fig. 6.22. The shear studs are checked rst. Although the concrete is
assumed to be cracked at the support in both global analysis and cross-section design,
the slab is considered to be uncracked for longitudinal shear design in accordance with
clause 6.6.2.1(2). The longitudinal shear is determined from the vertical shear using
elastic analysis in accordance with clause 6.6.2.1(1) since the bending resistance was
based on elastic analysis.
145
280
20 mm bars at
150 mm
85 85
250
Fig. 6.40. Shear studs and transverse reinforcement adjacent to an internal support
From Example 5.2, the modular ratio n
0
is 6.36. From Example 5.4, the area of
longitudinal reinforcement is A
s
19 480 mm
2
, which is 2.5% of the eective cross-
section of the concrete ange. Relevant elastic properties of the cross-section are given
in Table 6.3 for the uncracked unreinforced section (subscript U) and the uncracked
reinforced section (subscript UR). The eect on these values of including the reinforce-
ment is not negligible when the long-term modular ratio is used. However, the signicant
reduction in longitudinal shear caused by cracking is being ignored, so it is accurate
enough to use the unreinforced section when calculating longitudinal shear.
The height z to the neutral axis is measured from the bottom of the cross-section. The
property A" zz,I, for the whole of the eective concrete ange, is appropriate for checking
131
the shear resistance of the studs, and slightly conservative for the shear surface of type bb
shown in Fig. 6.40.
Table 6.3. Elastic properties of cross-section at internal support
Modular
ratio z
U
(mm) I
U
(m
2
mm
2
) A" zz,I
U
(m
1
) z
UR
(mm) I
UR
(m
2
mm
2
) A" zz,I
UR
(m
1
)
6.36 1120 38 530 0.8098 1146 39 890 0.8380
23.63 862 26 390 0.6438 959 31 200 0.7284
The calculation using n
0
is:
A" zz
I

3100
6.36
250

1375 1120
400
6.36
25

1238 1120
38 530 10
3
0.810 m
1
The design vertical shear forces on the composite section are: long-term, V
Ed
827 kN;
short-term, V
Ed
1076 kN; total, 1903 kN.
v
L.Ed
1903 0.810 1541 kN/m
Groups of four 19 mm diameter studs with f
u
500 N/mm
2
are provided at 150 mm
centres. Their resistance is governed by equation (6.19):
P
Rd

0.29cd
2

f
ck
E
cm
p
0.29 19
2

30 33 10
3
p
1.25 1000
83.3 kN,stud
Thus v
L.Rd
83.3 4,0.15 2222 kN/m which exceeds the 1541 kN/m applied.
Longitudinal shear in the concrete is checked next. The surface around the studs is
critical for the reinforcement check as it has the same reinforcement as the plane
through the deck (i.e. two 20 mm diameter bars) but almost twice the longitudinal
shear. It is assumed here that there is no signicant sagging bending over the top of the
main beams adjacent to the internal supports. From equation (6.21) in clause 6.2.4(4)
of EN 1992-1-1, assuming a 458 truss angle:
A
sf
f
yd
,s
f
v
Ed
h
f
,cot 0
f
where v
Ed
h
f
1541 kN,m
A
sf
! 1541 150, cot 458 500,1.15 532 mm
2
< 628 mm
2
from two 20 mm bars
The planes through the deck slab are critical for the concrete stress check since the
length of each concrete shear plane is 250 mm compared to 570 mm for the surface
around the studs, which carries twice the shear. The shear stress is:
v
Ed
1541,2 250 3.08 N,mm
2
From equations (6.6N) in clause 6.2.2(6) and (6.22) in clause 6.2.4(4) of EN 1992-1-1:
v
Ed
< if
cd
sin 0
f
cos 0
f
0.6 1 30,250 30,1.5 0.5 5.28 N,mm
2
so the shear stress is satisfactory.
It should be noted that there is another surface around the studs which intersects the
top of the haunch with length of 582 mm. It would be possible for such a plane to be
more critical.
(ii) Mid-span of the main span
The maximum spacing of groups of studs as shown in Fig. 6.41 is now found. The trans-
verse reinforcement is as before and the cross-sectional dimensions of the beamare shown
in Fig. 6.22. In the real bridge, the moment at mid-span was less than the elastic resistance
132
to bending and therefore a similar calculation to that in (i) could have been performed.
For the purpose of this example, the sagging bending moment at mid-span is considered
to be increased to 8.0 MNm, comprising 2.0 MNm on the bare steel section, 2.0 MNm on
the long-term composite section and 4.0 MNm live load on the short-term composite
section.
145
20 mm bars at
150 mm
250
280 85 85
a
a
Fig. 6.41. Shear studs and transverse reinforcement at mid-span and adjacent to an end support
The distance between the point of maximum moment and the point where the elastic
resistance to bending is just exceeded is assumed here to be 4.5 m. The longitudinal
shear force between these points is determined by rearranging equation (6.3) and substi-
tuting M
Ed
for M
Rd
so that:
V
L.Ed
N
c
N
c.el

N
c.f
N
c.el

M
Ed
M
el.Rd

M
pl.Rd
M
el.Rd

It is not obvious which modular ratio gives the more adverse result for this region, so
calculations are done using both n
0
and n
L
for the long-termloading. The elastic resistance
to bending is governed by the tensile stress in the bottom ange. From clause 6.2.1.4(6),
M
el.Rd
is found by scaling down M
c.Ed
by a factor k until a stress limit is reached. The
relevant section moduli in m
2
mm units are 15.96 for the steel section, and 23.39 short
term and 21.55 long term for the composite section. Hence,
2000,15.96 4000,23.39 2000,21.55k 345 11.7
where 11.7 N/mm
2
is the stress due to the primary eects of shrinkage from Example 5.3.
(The secondary eects of shrinkage have been ignored here as they are relieving.)
This gives k 0.788, and the elastic resistance moment is:
M
el.Rd
2000 0.788 4000 2000 6728 kNm
From Example 6.1, M
pl.Rd
10 050 kNm.
The compressive force N
c.el
in the composite slab at M
el.Rd
is nowrequired. The primary
shrinkage stress, although tensile, is included for consistency with the calculation of
M
el.Rd
. Its value at mid-depth of the 250 mm ange, from Example 5.3, is 0.68 N/mm
2
.
The eect of the stresses in the haunch is negligible here. Using the section moduli for
mid-depth of the ange, given in Table 6.4, the mean stress is:
o
c.mean
0.788 4000,969.1 2000,1143 0.68 3.95 N,mm
2
Table 6.4. Section moduli at mid-span for uncracked unreinforced section
Modular
ratio z
U
(mm) I
U
(m
2
mm
2
) A" zz,I
U
(m
1
)
W
c
, top
of slab
(m
2
mm)
W
c
, mid
slab
(m
2
mm)
W
c
, mid
haunch
(m
2
mm)
W
a
, bottom
ange
(m
2
mm)
6.36 1192 27 880 0.8024 575.9 969.1 3890 23.39
23.7 951 20 500 0.6843 882.7 1143 1692 21.55
133
Hence,
N
c.el
3.1 0.25 3.95 3.06 MN
The force for full interaction, N
c.f
, is based on rectangular stress blocks:
N
c.f
3.1 0.25 0.85 30,1.5 13.18 MN
From equation (6.3):
V
L.Ed
N
c
N
c.el
N
c.f
N
c.el
M
Ed
M
el.Rd
M
pl.Rd
M
el.Rd

13.18 3.068.0 6.728
10.05 6.728
3875 kN
The shear ow is therefore v
L.Ed
3875,4.5 861 kN/m
Similar calculations using n
0
6.36 for all loading give: M
el.Rd
6858 kNm, o
c.mean

4.33 N/mm
2
, and N
c.el
3.36 MN, from which v
L.Ed
781 kN/m, which does not govern.
From (i) above, P
Rd
83.3 kN/stud, so the required spacing of groups of three is:
s
v
83.3 3,861 0.290 m
Groups at 250 mm spacing could be used, unless a fatigue or serviceability check is found
to govern.
Longitudinal shear in the concrete is adequate by inspection, as the shear ow is less
than that at the internal support but both the thickness of the slab and the reinforcement
are the same.
(iii) End support
The shear stud arrangement and transverse reinforcement are as shown in Fig. 6.41 and
the cross-sectional dimensions are shown in Fig. 6.22. The studs are checked rst. The
longitudinal shear is determined from the vertical shear using elastic analysis in accor-
dance with clause 6.6.2.1(1) since the bending moment is less than the elastic resistance.
From similar calculations to those for the pier section, the longitudinal shear ow,
excluding that from shrinkage and temperature, is found to be v
L.Ed
810 kN/m. The
combination factor
0
for eects of temperature recommended in Annex A2 of
EN 1990 is 0.6, with a Note that it may in most cases be reduced to 0 for ultimate
limit states EQU, STR, and GEO. Temperature eects are not considered here.
The primary and secondary eects of shrinkage are both benecial, and so are not
considered. With P
Rd
83.3 kN/stud, the resistance of groups of three at 250 mm
spacing is 1000 kN/m, which is sucient.
Longitudinal shear in the concrete is adequate by inspection, as the shear ow is less
than that at the internal support but both concrete slab thickness and reinforcement
are the same.
The studs in all three cases above should also be checked at serviceability according to
clause 6.8.1(3). This has not been done here.
Example 6.11: inuence of in-plane shear in a compressed ange on bending
resistances of a beam
The mid-span region of an internal beam in the main span of the bridge is studied. Its
cross-section is shown in Fig. 6.22. The imposed sagging bending moments are as in
Example 6.10(ii): M
Ed.a
2.0 MNm, M
Ed.c
2.0 MNm short term plus 4.0 MNm long
term. Shear ow is reduced by creep, so the short-term modular ratio, n
0
6.36, is
used for all loading, and shrinkage eects are ignored. The section is in Class 2, but the
full plastic resistance moment, M
pl.Rd
, is unlikely to be needed. Properties found earlier
are as follows:
.
from Example 6.1, M
pl.Rd
10 050 kNm
134
.
from Fig. 6.2, at M
pl.Rd
, depth of slab in compression 227 mm
.
from Example 6.10, M
el.Rd
6858 kNm.
Elastic properties of the cross-section are given in Table 6.4.
Situations are studied in which a layer of thickness h
v
at the bottomof the slab is needed
for the diagonal struts of the truss model for in-plane shear. The bending resistance is
found using only the remaining thickness, h
m
250 h
v
(mm units). This is the depth
of compression considered in the bending assessment to which clause 6.2.4(105) of
EN 1992-2 could be assumed to refer. As discussed in the main text, it was only intended
to refer to transverse bending moment. The haunch is included in the elastic properties,
and neglected in calculations for M
pl.Rd
.
0
4
M
Rd
, MNm
h
m
(mm)
M
pl
,
Rd
M
el
,
Rd
8
6
100 200
10
Compression at the top
of the slab governs
Tension in the steel
bottom flange governs
50 150
250
Fig. 6.42. Inuence of thickness of slab in compression on resistance to bending
The eect on the bending resistances of reducing the ange thickness from 250 mm to
h
m
is shown in Fig. 6.42. Reduction in M
pl.Rd
does not begin until h
m
< 227 mm, and is
then gradual, until the plastic neutral axis moves into the web, at h
m
122 mm. When
h
m
100 mm, the reduction is still less than 10%.
The elastic resistance is governed by stress in the bottom ange, and at rst increases as
h
m
is reduced. Compressive stress in concrete does not govern until h
m
is less than 50 mm,
as shown. For h
m
100 mm, M
el.Rd
is greater than for 250 mm.
The vertical shear for which h
m
100 mm, h
v
150 mm is now found and compared
with the shear resistance of the steel web, V
bw.Rd
1834 kN from Example 6.4. The
method is explained in comments on clause 6.6.6.2(1). The optimum angle 0
f
for the
concrete struts, which is 458, is used. It is assumed that the transverse reinforcement
does not govern. The property A" zz,I should continue to be based on the full slab thickness,
not on h
m
, to provide a margin for over-strength materials, inelastic behaviour, etc.
Equation (6.22) in EN 1992-1-1 is v
Ed
< i f
cd
sin 0
f
cos 0
f
.
From Example 6.10(i) with f
ck
30 N/mm
2
, 0
f
458; this gives v
Ed
< 5.28 N/mm
2
.
From Fig. 6.41, the width of ange excluded by the shear plane aa is:
3100 450,2 1325 mm
The shear ow on this plane is:
v
L.Ed
v
Ed
h
v
1325,3100V
Ed
A" zz,I D6.25
From equation (D6.25) with h
v
150 mm and A" zz,I from Table 6.4,
V
Ed
< 5.28 150 3100,1325,0.8024 2309 kN
This limit is 26% above the shear resistance of the steel web. Evidently, interaction
between in-plane shear and bending resistance is negligible for this cross-section.
135
6.7. Composite columns and composite compression
members
6.7.1. General
Scope
A composite column is dened in clause 1.5.2.5 as a composite member subjected mainly to
compression or to compression and bending. The title of clause 6.7 includes compression
members to make clear that its scope is not limited to vertical members, but includes, for
example, composite members in trusses.
Composite columns are more widely used in buildings than in bridges, so their treatment
here is less detailed than in the Designers Guide to EN 1994-1-1.
5
Its Example 6.10 on a
concrete-encased I-section column is supplemented here by Example 6.12 on a concrete-
lled steel tube.
In this Guide, references to columns include other composite compression members,
unless noted otherwise, and column means a length of member between adjacent lateral
restraints. The concept of the eective length of a column is not used in clause 6.7.
Instead, the relative slenderness is dened, in clause 6.7.3.3(2), in terms of N
cr
, the
elastic critical normal force for the relevant buckling mode. This use of N
cr
is explained in
the comments on clause 6.7.3.3.
Clause 6.7.1(1)P refers to Fig. 6.17, in which all the cross-sections shown have double
symmetry; but clause 6.7.1(6) makes clear that the scope of the general method of
clause 6.7.2 includes members of non-symmetrical section.
87
Clause 6.7 is not intended for application to members subjected mainly to transverse
loading and also resisting longitudinal compression, such as longitudinal beams in an
integral bridge. These are treated in this Guide in the comments on beams.
The bending moment in a compression member depends on the assumed location of the
line of action of the axial force, N. Where the cross-section has double symmetry, as in
most columns, this is taken as the intersection of the axes of symmetry. In other cases the
choice, made in the modelling for global analysis, should be retained for the analysis of
the cross-sections. A small degree of asymmetry (e.g. due to an embedded pipe) can be
allowed for by ignoring in calculations concrete areas elsewhere, such that symmetry is
restored.
No shear connectors are shown in the cross-sections in Fig. 6.17, because within a column
length, the longitudinal shear is normally much lower than in a beam, and sucient inter-
action may be provided by bond or friction. Shear connectors are normally required for
load introduction, following clause 6.7.4.
Where the design axial compression is less than N
pm.Rd
, shown in Fig. 6.19 and Fig. 6.47, it
is on the safe side to ignore it in verication of cross-sections. Where there is moderate or
high transverse shear, shear connectors may be needed throughout the member. Example
6.11 of Ref. 5 is relevant.
The strengths of materials in clause 6.7.1(2)P are as for beams, except that class C60/75
and lightweight-aggregate concretes are excluded. For these, additional provisions (e.g. for
creep, shrinkage and strain capacity) would be required.
88.89
Clause 6.7.1(3) and clause 5.1.1(2) both concern the scope of EN 1994-2, as discussed
above.
The steel contribution ratio, clause 6.7.1(4), is the proportion of the squash load of the
section that is provided by the structural steel member. If it is outside the limits given, the
member should be treated as reinforced concrete or as structural steel, as appropriate.
Independent action eects
Clause 6.7.1(7) relates to the NM interaction curve for a cross-section of a column shown
in Fig. 6.19 and as a polygon in Fig. 6.47. It applies where the factored axial compression
F
N
Ek
is less than N
pm.Rd
/2, so that reduction in N
Ed
reduces M
Rd.
As this could be
unsafe where N
Ed
and M
Ed
result from independent actions, the factor
F
for N
Ek
is
reduced, as illustrated in Fig. 6.34 of Ref. 5.
Clause 6.7.1(1)P
Clause 6.7.1(2)P
Clause 6.7.1(3)
Clause 6.7.1(4)
Clause 6.7.1(7)
136
The bulge in the interaction curve is often tiny, as shown in Fig. 6.47. Asimpler and more
conservative rule, that ignores the bulge, was given in ENV 1994-1-1. It is that if M
Rd
corresponding to
F
N
Ek
is found to exceed M
pl.Rd
, M
Rd
should be taken as M
pl.Rd
. It is
applicable unless the bending moment M
Ed
is due solely to the eccentricity of the force N
Ed
.
It is doubtful if the 20% rule of clause 6.7.1(7) was intended to be combined with the
reduction of
F
from 1.35 to 1.0 for a permanent action with a relieving eect. Where that
is done, use of the simpler rule given above is recommended (e.g. in Fig. 6.47, to replace
boundary BDC by BC).
Local buckling
The principle of clause 6.7.1(8)P is followed by application rules in clause 6.7.1(9). They
ensure that the concrete (reinforced in accordance with clause 6.7.5) restrains the steel and
prevents it from buckling even when yielded. Columns are, in eect, treated in clause 6.7
as Class 2 sections. Restraint from the concrete enables the slenderness limits for Class 2
to be increased to the values given in Table 6.3. For example, the factor 90 given for a circular
hollow section replaces 70 in EN 1993-1-1. Members in Class 3 or 4 are outside the scope of
clause 6.7.
Fatigue
Verication of columns for fatigue will rarely be needed, but fatigue loading could occur in
composite members in a truss or in composite columns in integral bridges. Verication, if
required, should be to clause 6.8.
6.7.2. General method of design
The general method of clause 6.7.2 is provided for members outside the scope of the
simplied method of clause 6.7.3, and also to enable advanced software-based methods to
be used. It is more a set of principles than a design method. Writing software that satises
them is a task for specialists.
Much of clause 6.7.3 and the comments on it provide further guidance on design of com-
pression members that are outside its scope.
Clause 6.7.2(3)P refers to internal forces. These are the action eects within the com-
pression member, found from global analysis to Section 5 that includes global and local
imperfections and second-order eects.
Clause 6.7.2(3)P also refers to elasto-plastic analysis. This is dened in clause 1.5.6.10 of
EN 1990 as structural analysis that uses stress/strain or moment/curvature relationships
consisting of a linear elastic part followed by a plastic part with or without hardening.
As the three materials in a composite section follow dierent non-linear relationships,
direct analysis of cross-sections is not possible. One has rst to assume the dimensions
and materials of the member, and then determine the axial force N and bending moment
M at a cross-section from assumed values of axial strain and curvature c, using the relevant
material properties. The MNc relationship for each section can be found from many such
calculations. This becomes more complex where biaxial bending is present.
90
Integration along the length of the member then leads to a non-linear member stiness
matrix that relates axial force and end moments to the axial change of length and the end
rotations.
Clause 6.7.2(8) on stressstrain curves was drafted, as a General rule, before clause 5.7
of EN 1992-2 was available and refers only to the Parts 1-1 of Eurocodes 2 and 3. At that
time these rules appeared to be incompatible for use for composite structures. Hence, no
application rules on non-linear global analysis are given in clause 5.4.3, where further
comment is given.
In clause 5.7 of EN 1992-2, the intention is that realistic stinesses, not design stinesses,
should be used, on the basis that a small amount of material at the critical section with
design properties will not alter the overall response. For bridges, the recommended stress
strain curves are based on the characteristic strengths. This is consistent with Informative
Clause 6.7.1(8)P
Clause 6.7.1(9)
Clause 6.7.2
Clause 6.7.2(3)P
Clause 6.7.2(8)
137
Annex C of EN 1993-1-5, for structural steel. Both Eurocodes 2 and 3 refer to their national
annexes for this subject. In the absence of references in EN 1994-2 to these Parts of Eurocodes
2 and 3, guidance should be sought from the National Annex.
Where characteristic properties are used in non-linear global analysis, further checks on
cross-sections are required. An attractive proposition therefore is to use design values of
material properties throughout, so that the non-linear analysis itself becomes the verica-
tion, provided that the resistance found exceeds the factored loading. This approach is per-
mitted by clause 5.7 of EN 1992-2. However, it may not be conservative for serviceability
limit states if signicant internal forces arise from indirect actions such that greater stiness
attracts greater internal eects. There is a caveat to this eect in clause 5.7 of EN 1992-2.
6.7.3. Simplied method of design
Scope of the simplied method
The method has been calibrated by comparison with test results.
91.92
Its scope, clause 6.7.3.1,
is limited mainly by the range of results available, which leads to the restriction
"
`` 2 in
clause 6.7.3.1(1). For most columns, the method requires explicit account to be taken of
imperfections and second-order eects. The use of strut curves is limited in clause
6.7.3.5(2) to axially-loaded members.
The restriction on unconnected steel sections in paragraph (1) is to prevent loss of stiness
due to slip, that would invalidate the formulae for EI of the column cross-section. The limits
to concrete cover in clause 6.7.3.1(2) arise from concern over strain softening of concrete
invalidating the interaction diagram (Fig. 6.19), and from the limited test data for
columns with thicker covers. These provisions normally ensure that for each axis of
bending, the exural stiness of the steel section makes a signicant contribution to the
total stiness. Greater cover can be used by ignoring in calculation the concrete that
exceeds the stated limits.
The limit of 6% in clause 6.7.3.1(3) on the reinforcement used in calculation is more
liberal than the 4% (except at laps) recommended in EN 1992-1-1. This limit and that on
maximum slenderness are unlikely to be restrictive in practice.
Clause 6.7.3.1(4) is intended to prevent the use of sections susceptible to lateraltorsional
buckling.
Resistance of cross-sections
Reference to the partial safety factors for the materials is avoided by specifying resistances in
terms of design values for strength, rather than characteristic values; for example in equation
(6.30) for plastic resistance to compression in clause 6.7.3.2(1). This resistance, N
pl.Rd
, is the
design ultimate axial load for a short column, assuming that the structural steel and
reinforcement are yielding and the concrete is crushing.
For concrete-encased sections, the crushing stress is taken as 85% of the design cylinder
strength, as explained in the comments on clause 3.1. For concrete-lled sections, the
concrete component develops a higher strength because of the connement from the steel
section, and the 15% reduction is not made; see also the comments on clause 6.7.3.2(6).
Resistance to combined compression and bending
The bending resistance of a column cross-section, M
pl.Rd
, is calculated as for a composite
beam in Class 1 or 2, clause 6.7.3.2(2). Points on the interaction curve shown in Figs 6.18
and 6.19 represent limiting combinations of compressive axial load N and moment M
which correspond to the plastic resistance of the cross-section.
The resistance is found using rectangular stress blocks. For simplicity, that for the
concrete extends to the neutral axis, as shown in Fig. 6.43 for resistance to bending (point
B in Fig. 6.19 and Fig. 6.47). As explained in the comments on clause 3.1(1), this simplica-
tion is unconservative in comparison with stress/strain curves for concrete and the rules of
EN 1992-1-1. To compensate for this, the plastic resistance moment for the cross-section
is reduced by a factor c
M
in clause 6.7.3.6(1).
Clause 6.7.3.1
Clause 6.7.3.1(1)
Clause 6.7.3.1(2)
Clause 6.7.3.1(3)
Clause 6.7.3.1(4)
Clause 6.7.3.2(1)
Clause 6.7.3.2(2)
138
As axial compression increases, the neutral axis moves; for example, towards the lower
edge of the section shown in Fig. 6.43, and then outside the section. The interaction curve
is therefore determined by moving the neutral axis in increments across the section, and
nding pairs of values of M and N from the corresponding stress blocks. This requires a
computer program, unless the simplication given in clause 6.7.3.2(5) is used. Simplied
expressions for the coordinates of points B, C and D on the interaction curve are given in
Appendix C of Ref. 5. Further comment is given in Examples 6.10 and C.1 in that Guide
and in Example 6.12 here.
Inuence of transverse shear
Clauses 6.7.3.2(3) and (4), on the inuence of transverse shear on the interaction curve, are
generally the same as clause 6.2.2.4 on momentshear interaction in beams. One assumes rst
that the shear V
Ed
acts on the structural steel section alone. If it is less than 0.5V
pl.a.Rd
, it has
no eect on the curve. If it is greater, there is an option of sharing it between the steel and
reinforced concrete sections, which may reduce that acting on the steel to below
0.5V
pl.a.Rd
. If it does not, then a reduced design yield strength is used for the shear area,
as for the web of a beam. In a column the shear area depends on the plane of bending
considered, and may consist of the anges of the steel section. It is assumed that shear
buckling does not occur.
Shear high enough for V
c.Rd
to be relied on in design is unlikely to occur in a composite
column, so the code does not go into detail here. The reference in clause 6.7.3.2(3) to the
use of EN 1992 does not include EN 1992-2, where clause 6.2 was not drafted with
columns in mind. Equation (6.2.b) in EN 1992-1-1, which gives a minimum shear strength
for concrete regardless of reinforcement content, is not valid for unreinforced concrete as
it assumes that minimum reinforcement will be provided according to EN 1992 require-
ments. It can be inferred that for a concrete-lled tube with no longitudinal reinforcement
(permitted by clause 6.7.5.2(1)), the shear resistance V
c.Rd
according to EN 1992 should
be taken as zero.
Simplied interaction curve
Clause 6.7.3.2(5) explains the use of the polygonal diagramBDCAin Fig. 6.19 as an approx-
imation to the interaction curve, suitable for hand calculation. The method applies to any
cross-section with biaxial symmetry, not just to encased I-sections.
First, the location of the neutral axis for pure bending is found, by equating the longitu-
dinal forces from the stress blocks on either side of it. Let this be at distance h
n
from the
centroid of the uncracked section, as shown in Fig. 6.19(B). It is shown in Appendix C of
Ref. 5 that the neutral axis for point C on the interaction diagram is at distance h
n
on the
other side of the centroid, and the neutral axis for point D passes through the centroid.
The values of M and N at each point are easily found from the stress blocks shown in
Fig. 6.19. For concrete-lled steel tubes the factor 0.85 may be omitted.
Concrete-lled tubes of circular or rectangular cross-section
Clause 6.7.3.2(6) is based on the lateral expansion that occurs in concrete under axial
compression. This causes circumferential tension in the steel tube and triaxial compression
in the concrete. This increases the crushing strength of the concrete
91
to an extent that
Clause 6.7.3.2(3)
Clause 6.7.3.2(4)
Clause 6.7.3.2(5)
Clause 6.7.3.2(6)
0.85f
cd
+
+
Concrete Steel Reinforcement
f
yd
f
yd
f
sd
M
pl,Rd
f
sd
Fig. 6.43. Stress distributions for resistance in bending
139
outweighs the reduction in the eective yield strength of the steel in vertical compression. The
coecients j
a
and j
c
given in this clause allow for these eects.
This containment eect is not present to the same extent in concrete-lled rectangular
tubes because less circumferential tension can be developed. In all tubes the eects of
containment reduce as bending moments are applied, because the mean compressive strain
in the concrete and the associated lateral expansion are reduced. With increasing slenderness,
bowing of the member under load increases the bending moment, and therefore the eective-
ness of containment is further reduced. For these reasons, j
a
and j
c
are dependent on the
eccentricity of loading and on the slenderness of the member.
Properties of the column or compression member
For composite compression members in a frame, some properties of each member are needed
before or during global analysis of the frame:
.
the steel contribution ratio, clause 6.7.3.3(1)
.
the relative slenderness
"
``, clause 6.7.3.3(2)
.
the eective exural stinesses, clauses 6.7.3.3(3) and 6.7.3.4(2), and
.
the creep coecient and eective modulus for concrete, clause 6.7.3.3(4).
The steel contribution ratio is explained in the comments on clause 6.7.1(4).
The relative slenderness
"
`` is needed to check that the member is within the scope of the
simplied method, clause 6.7.3.1(1). Often it will be evident that
"
`` < 2. The calculation
can then be omitted, as
"
`` is not needed again unless the member resists axial load only.
The unfactored quantities E, I and L are used in the calculation of N
cr
, so
"
`` is calculated
using in equation (6.39) the characteristic (unfactored) value of the squash load, N
pl.Rk
, and
the characteristic exural stiness (EI)
eff
from clause 6.7.3.3(3). This is the only use of this
stiness in Section 6. The upper limit on
"
`` is somewhat arbitrary and does not justify great
precision in N
cr
.
Creep of concrete increases the lateral deformation of the member. This is allowed for by
replacing the elastic modulus E
cm
(in equation (6.40)) by a reduced value E
c.eff
from equation
(6.41). This depends on the creep coecient c
t
, which is a function of the age at which
concrete is stressed and the duration of the load. The eective modulus depends also on
the proportion of the design axial load that is permanent. The design of the member is
rarely sensitive to the inuence of the creep coecient on E
c.eff
, so conservative assumptions
can be made about uncertainties. Normally, a single value of eective modulus can be used
for all compression members in a structure. Further discussion is given under clause 5.4.2.2.
The correction factor K
e
is to allow for loss of stiness caused by possible cracking of
concrete.
The condition for ignoring second-order eects within the member is explained in com-
ments on clause 5.2.1(3). Where the ratio c
cr
( N
cr
/N
Ed
) is used, the critical load N
cr
is
the axial force in the member in the lowest buckling mode of the structure that involves
the member. In the rare cases where both ends of a column are detailed so as to behave as
pin-ended (as in Example 6.12), N
cr

2
EI
eff.II
/L
2
. The exural stiness (EI)
eff.II
is
obtained from clause 6.7.3.4(2).
In continuous construction, the critical buckling mode involves adjacent members, which
must be included in the elastic critical analysis.
Analyses for verication of a compression member
For the compression members of a frame or truss that are within the scope of clause 6.7.3, a
ow chart for calculation routes is given in Fig. 6.44.
The relationship between the analysis of a frame and the stability of individual members
is discussed both in the comments on clause 5.2.2 and below. If bending is biaxial, the
procedure in clause 6.7.3.4 is followed for each axis in turn.
Clause 6.7.3.4(1) requires the use of second-order linear-elastic global analysis except
where the option of clause 6.7.3.4(5) applies, or route (c) in Fig. 6.44 is chosen and
c
cr
10 in accordance with clause 6.7.3.4(3). The simplied method of clause 6.7.3.5(2)
Clause 6.7.3.3(1)
Clause 6.7.3.3(2)
Clause 6.7.3.3(3)
Clause 6.7.3.3(4)
Clause 6.7.3.4(1)
140
is rarely applicable in practice because some rst-order bending moment (other than from
imperfections) will usually be present; for example, due to friction at bearings.
Clause 6.7.3.4(2) gives the design exural stinesses for compression members, for use
in all analyses for ultimate limit states. The factor K
e.II
allows for cracking, as is required
by the reference in clause 5.4.2.3(4) to clause 6.7.3.4. The factor K
0
is from research-based
calibration studies. Long-term eects are allowed for, as before, by replacing E
cm
in equation
(6.42) by E
c.eff
from equation (6.41).
In clause 6.7.3.4(3), the elastic critical load refers to the frame at its lowest buckling
mode involving the member concerned; and second-order eects means those in the member
due to both its own imperfections and global imperfections. When deciding whether
second-order eects of member imperfections can be neglected, the eects of global imperfec-
tions can be neglected in an elastic critical analysis (route (c) in Fig. 6.44). A second-order
analysis for the asymptotic load, route (a), will give the same value for c
cr
whether global
imperfections are included or not. They are shown in Fig. 6.44 as included because the
same analysis can then give the design bending moments for the member concerned.
Clause 6.7.3.4(4) gives the equivalent member imperfection, for use in a global analysis, as
an initial bow. It is proportional to the length L of the member between lateral restraints and
is dened by e
0
, the lateral departure at mid-height of its axis of symmetry from the line
joining the centres of symmetry at its ends. The value accounts principally for truly geometric
imperfections and for the eects of residual stresses. It is independent of the distribution of
bending moment along the member. The curved shape is usually assumed to be sinusoidal,
but a circular arc is acceptable. The curve is assumed initially to lie in the plane normal to the
axes of the bending moments.
Clause 6.7.3.4(2)
Clause 6.7.3.4(3)
Clause 6.7.3.4(4)
Yes Yes No No Yes No
Note 1. Loading means a particular combination of actions, load case and load arrangement. In boxes (a) to (c)
the lowest
cr
for various loadings is found. The chosen loadings should include that for maximum
side-sway, and those that are expected to cause the greatest axial compression in each potentially critical
compression member.
Note 2. Analysis (a) includes both P effects from global imperfections and P effects from member
imperfections.
Note 3. For choice of loadings, see Note 1 and the comments on clause 5.2.1(3) and (4).
Note 4. No need to return to (a) or (b) where
cr
< 10 only in a local member mode (pin-ended conditions).
Then, do first-order g.a. with amplification to 6.7.3.4(5) and verify cross-sections.
Flow chart for global analysis (g.a.) and verification of a compression member in a composite frame, with reference
to global and member imperfections (g.imp and m.imp). This is for a member of doubly symmetrical and uniform
cross-section (6.7.3.1(1)) and for a particular loading. See Notes 1 and 2.
Analysis complete. Verify resistance of cross-sections of each compression
member, to 6.7.3.6 or 6.7.3.7 (END)
Go to box (a)
or (b), above,
except that
cr

need not be
found again.
See Note 4
Do 1st order
g.a. with g.imp,
and m.imp to
6.7.3.4(4)
Find
cr
to 5.2.1(3), using (EI )
eff,II
to 6.7.3.4(2) for all compression members by one of methods (a) to (c):
Do 1st order g.a.
with g.imp, and
m.imp to
6.7.3.4(4)
(b) As box (a)
but m.imp not
included.
Is
cr
10?
(c) Elastic critical analysis with
proportional loading and vertical
loads only, with no imp., to find
cr
.
See Note 3. Is
cr
10?
(a) 2nd order g.a. for proportional
loading with g.imp and m.imp. Find
ratio
cr
of asymptotic load to design
load. Is
cr
10?
Use 2nd order
analysis as
above. Allow
for 1st and 2nd
order m.imp
to 6.7.3.4(5)
Use results
from (a) for
the design
loading
1st order g.a. with
g.imp and m.imp is
permitted, but not
needed, as results
from (a) for the design
loading can be used.
(Use of 1st order could
be more economic)
Fig. 6.44. Flow chart for analysis and verication for a compression member
141
Where member imperfections are not included in the global analysis and c
cr
< 10,
clause 6.7.3.4(5) enables these imperfections to be allowed for. It is based on the critical
load N
cr.eff
for the isolated pin-ended member even where the critical buckling mode for
the frame involves sway, such that the eective length of the member exceeds its system
length. This is consistent with clause 5.2.2(7)(b) of EN 1993-1-1, which is referred to from
clause 5.2.2(1) via clause 5.2.2 of EN 1993-2. (This route also leads to clause 5.2.2(3) of
EN 1993-1-1, which sets out these options in detail and is consistent with Fig. 6.44.)
The reason for this denition for N
cr.eff
is that where necessary (e.g. where c
cr
< 10), the
eects of global imperfections and side-sway have been accounted for in the second-order
global analysis. This can be seen by using as an example a agpole-type column, with
both out-of-plumb and initial bow.
Determination of design bending moment for a compression member
It is now assumed that for a particular member, with or without lateral load within its length,
an analysis in which member imperfections were not included has provided a design axial force
N
Ed
and design end moments for one of the principal planes of bending. These are denoted M
1
and M
2
, with jM
1
j ! jM
2
j, as shown in Fig. 6.45(a). Biaxial bending is considered later. The
axial force is normally almost constant along the length of the member. Where it varies, its
maximum value can conservatively be assumed to be present throughout its length.
The factor k in clause 6.7.3.4(5) is proportional to u. For the end moments, from Table
6.4, u lies between 0.44 and 1.1, but for the member imperfection it is always 1.0. These
two values are denoted u
1
and u
2
. The calculation of u
1
is now explained, assuming that
the critical bending moment occurs either at the end where M
Ed
M
1
(where no member
imperfection or resulting second-order eect is assumed) or within the central 20% of the
length of the member.
Except where there is lateral loading, the maximumrst-order bending moment within this
central length is:
M
20%
M
1
0.6 0.4r
shown in Fig. 6.45(a). This is represented by an equivalent rst-order design value, given by
Table 6.4 as:
M
1st.Ed
M
1
0.66 0.44r
with a lower limit of 0.44M
1
. The ratio M
1st.Ed
,M
20%
is shown in Fig. 6.45(c). It is generally
1.1, but increases sharply where r < 0.5, which is where the lower limit of 0.44M
1
is
reached. This range of r represents signicant double-curvature bending. The increase pro-
vides protection against snap-through to single-curvature buckling.
The moment M
1st.Ed
is increased by the factor
1
1 N
Ed
,N
cr.eff
Clause 6.7.3.4(5)
0 M
1
M
20%
M
20%
/10
0.6L
2.0
1.0
(c) (b) (a)
0 1 0.5
0
M
1st,Ed
/M
20%
N
Ed
e
0
+ M
tr,Ed
M
2
= rM
1
r (=M
2
/M
1
)
0
Second-order effect
Second-order
effect
Fig. 6.45. Bending moments in a column: (a) end moments from global analysis; (b) initial imperfections
and transverse loading; and (c) equivalent rst-order bending moment
142
to allow for second-order eects. This factor is an approximation, as shown in Fig. 2.9 of
Ref. 93, which is allowed for by the use of the ratio 1.1 shown in Fig. 6.45(c).
Basing N
cr.eff
on pin-ended conditions can be conservative where the column is braced with
end rotational restraints that produce an eective length less than the column height.
Where there is lateral loading within the column length, the bending-moment distribution
should be treated as the sum of two distributions, one corresponding to each of the two parts
of Table 6.4. In the rst half of the table, M
Ed
is then the sum of contributions from member
imperfection and lateral load. These are not necessarily in the same direction, because the
member imperfection e
0
can be in any lateral direction and must be chosen to give the
most adverse overall result for the column.
Equation (6.43) states that k must be greater than or equal to unity, and this is correct for a
single distribution of bending moment. However, for a combination of two distributions, it
could be conservative to adjust both values of k in this way when the two sets of moments are
treated separately.
At mid-length the component due to end moments depends on their ratio, r, and therefore
could be small. The appropriate rst component is:
u
1
M
1
1 N
Ed
,N
cr.eff
without the condition that it should exceed M
1
.
The imperfection e
0
causes a rst-order bending moment N
Ed
e
0
at mid-length. The result-
ing contribution to M
Ed.max
is:
u
2
N
Ed
e
0
1 N
Ed
,N
cr.eff
. with u
2
1
This, plus the contribution fromany mid-length moment fromlateral loading, is added to the
rst component. The condition k ! 1 applies to the sum of the two components, and is
intended to ensure that the design moment is at least equal to the greatest rst-order moment.
In biaxial bending, the initial member imperfection may be neglected in the less critical
plane (clause 6.7.3.7(1)).
The denitions of M
Ed
in clause 6.7.3.4(5) and Table 6.4 may appear contradictory. In the
text before equation (6.43), M
Ed
is referred to as a rst-order moment. This is because it does
not include second-order eects arising within the member. However, Table 6.4 makes clear
that M
Ed
is to be determined by either rst-order or second-order global analysis, as shown
in Fig. 6.44.
The simplied method of clause 6.7.3.5 is rarely applicable, as explained in the comment
on clause 6.7.3.4(1), so no comment on it is given.
Verication of cross-sections
For uniaxial bending, the nal step is to check that the cross-section can resist M
max.Ed
with
compression N
Ed
. The interaction diagramgives a resistance j
d
M
pl.Rd
with axial load N
Ed
, as
shown in Fig. 6.18. This is unconservative, being based on rectangular stress-blocks, as
explained in the comment on clause 3.1(1), so in clause 6.7.3.6(1) it is reduced, by the use
of a factor c
M
that depends on the grade of structural steel. This factor allows for the
increase in the compressive strain in the cross-section at yield of the steel (which is adverse
for the concrete), when the yield strength of the steel is increased.
Where values of M
max.Ed
have been found for both axes, clause 6.7.3.7 on biaxial bending
applies, in which they are written as M
y.Ed
and M
z.Ed
. If one is much greater than the other,
the relevant check for uniaxial bending, equation (6.46), will govern. Otherwise, the linear
interaction given by equation (6.47) applies.
If the member fails this biaxial condition by a small margin, it may be helpful to recalculate
the less critical bending moment omitting the member imperfection, as permitted by
clause 6.7.3.7(1).
Clause 6.7.3.5
Clause 6.7.3.6(1)
Clause 6.7.3.7
143
6.7.4. Shear connection and load introduction
Load introduction
The provisions for the resistance of cross-sections of columns assume that no signicant slip
occurs at the interface between the concrete and structural steel components. Clauses
6.7.4.1(1)P and (2)P give the principles for limiting slip to an insignicant level in the
critical regions: those where axial load and/or bending moments are applied to the member.
For any assumed clearly dened load path it is possible to estimate stresses, including
shear at the interface. Shear connection is unlikely to be needed outside regions of load
introduction unless the shear strength t
Rd
from Table 6.6 is very low, or the member is
also acting as a beam, or has a high degree of double-curvature bending. Clause
6.7.4.1(3) refers to the special case of an axially loaded column.
Where axial load is applied through a joint attached only to the steel component, the force
to be transferred to the concrete can be estimated from the relative axial loads in the two
materials given by the resistance model. Accurate calculation is rarely practicable where
the cross-section concerned does not govern the design of the column. In this partly-
plastic situation, the more adverse of the elastic and fully-plastic models give a safe result
(clause 6.7.4.2(1), last line). In practice, it may be simpler to provide shear connection
based on a conservative (high) estimate of the force to be transferred.
Where axial force is applied by a plate bearing on both materials or on concrete only, the
proportion of the force resisted by the concrete gradually decreases, due to creep and
shrinkage. It could be inferred from clause 6.7.4.2(1) that shear connection should be
provided for a high proportion of the force applied. However, models based on elastic
theory are over-conservative in this inherently stable situation, where large strains are
acceptable. The application rules that follow are based mainly on tests.
Few shear connectors reach their design shear strength until the slip is at least 1 mm; but
this is not signicant slip for a resistance model based on plastic behaviour and rectangular
stress blocks. However, a long load path implies greater slip, so the assumed path should not
extend beyond the introduction length given in clause 6.7.4.2(2).
In a concrete-lled tube, shrinkage eects are low, for only the autogenous shrinkage
strain occurs, with a long-term value below 10
4
, from clause 3.1.4(6) of EN 1992-1-1.
Radial shrinkage is outweighed by the lateral expansion of concrete in compression, for
its inelastic Poissons ratio increases at high compressive stress, and eventually exceeds
0.5. Friction then provides signicant shear connection.
Concrete-lled tubular columns with bearings at both ends have found application in
bridge design. Where the whole load is applied to the concrete core through an end plate,
the conditions of clause 6.7.4.2(3) can be satised, and no shear connection is required.
This complete reliance on friction for shear transfer is supported by test evidence
94.95
and by inelastic theory. For columns of circular cross-section, no plausible failure mechan-
ism has been found for an end region that does not involve yielding of the steel casing in
hoop tension and vertical compression. For a large non-circular column, it would be
prudent to check behaviour by nite-element analysis. There is further discussion in
Example 6.12.
Friction is also the basis for the enhanced resistance of stud connectors given in
clause 6.7.4.2(4).
Detailing at points of load introduction or change of cross-section is assisted by the high
permissible bearing stresses given in clauses 6.7.4.2(5) and (6). An example is given in Ref. 5
in which the local design compressive strength of the concrete, o
c.Rd
in equation (6.48), is
found to be 260 N/mm
2
.
Clause 6.7.4.2(7) relates to load introduction to reinforcement in a concrete-lled tube.
This and some other concessions made at the ends of a column length are based mainly
on tests on columns of sizes typical of those used in buildings. Some caution should be
exercised in applying them to members with much larger cross-sections. Unless a column
is free to sway, a hinge forms in its central region before it fails. End regions that are slightly
weaker have little eect on the failure load, because at that stage their bending moments are
lower than at mid-length.
Clause 6.7.4.1(1)P
Clause 6.7.4.1(2)P
Clause 6.7.4.1(3)
Clause 6.7.4.2(1)
Clause 6.7.4.2(2)
Clause 6.7.4.2(3)
Clause 6.7.4.2(4)
Clause 6.7.4.2(5)
Clause 6.7.4.2(6)
Clause 6.7.4.2(7)
144
The absence from clause 6.7.4.2(8) of a reference to EN 1992-2 is deliberate, as its clause
9.5.3 gives a rule that is not required for composite columns.
Figure 6.23 illustrates the requirement of clause 6.7.4.2(9) for transverse reinforcement,
which must have a resistance to tension equal to the force N
c1
. If longitudinal reinforcement
is ignored, this is given by:
N
c1
A
c2
,2nA
where A is the transformed area of the cross-section 1-1 of the column in Fig. 6.23, given by:
A A
s
A
c1
A
c2
,n
and A
c1
and A
c2
are the unshaded and shaded areas of concrete, respectively, in section 1-1.
The modular ratio, n, should usually be taken as the short-term value.
Transverse shear
Clause 6.7.4.3 gives application rules (used in Example 6.11 in Ref. 5) relevant to the
principle of clause 6.7.4.1(2), for columns with the longitudinal shear that arises from
transverse shear. The design shear strengths t
Rd
in Table 6.6 are far lower than the tensile
strength of concrete. They rely on friction, not bond, and are related to the extent to
which separation at the interface is prevented. For example, in partially-encased I-sections,
lateral expansion of the concrete creates pressure on the anges, but not on the web, for
which t
Rd
0; and the highest shear strengths are for concrete within steel tubes.
Where small steel I-sections are present within a column that is mainly concrete, clause
6.7.4.3(4) provides a useful increase to t
Rd
, for covers c
z
up to 115 mm. The enhancement
factor is more simply presented as: u
c
0.2 c
z
,50 2.5.
Concern about the attachment of concrete to steel in partially-encased I-sections appears
again in clause 6.7.4.3(5), because under weak-axis bending, separation tends to develop
between the encasement and the web.
6.7.5. Detailing provisions
If a steel I-section in an environment in class X0 to EN 1992-1-1 has links in contact with its
ange (permitted by clause 6.7.5.2(3)), the cover to the steel section could be as low as
25 mm. For a wide steel ange, this thin layer of concrete would have little resistance to
buckling outwards, so the minimum thickness is increased to 40 mm in clause 6.7.5.1(2).
This is a nominal dimension.
Minimum longitudinal reinforcement, clause 6.7.5.2(1), is needed to control the width of
cracks, which can be caused by shrinkage even in columns with concrete nominally in com-
pression.
Clause 6.7.5.2(2) does not refer to EN1992-2 because its clause 9.5 introduces a nationally-
determined parameter, which is not needed for composite columns.
Clause 6.7.5.2(4) refers to exposure class X0 of EN 1992-1-1. This is a very dry environ-
ment, with no risk of corrosion or attack, so this clause is unlikely to be applicable in
bridges.
Clause 6.7.4.2(8)
Clause 6.7.4.2(9)
Clause 6.7.4.3
Clause 6.7.4.3(4)
Clause 6.7.4.3(5)
Clause 6.7.5.1(2)
Clause 6.7.5.2(1)
Clause 6.7.5.2(2)
Clause 6.7.5.2(4)
Example 6.12: concrete-lled tube of circular cross-section
This Example is based on columns used for a motorway interchange, as described in
Refs 94 and 96. The column shown in Fig. 6.46 is of circular cross-section. It has spherical
rocker bearings at each end, of radii 600 mm (internal) and 605 mm (external). Its
length between centres of bearings is 11.5 m. Its design will be checked for a design
ultimate load N
Ed
18.0 MN. The bearing at its lower end is xed in position. The
deformation of the superstructure at ultimate load causes a maximum eccentricity of
load of 75 mm.
Concerning
M
for structural steel, clause 5.2.2(7)(a) of EN 1993-1-1 says that no
individual stability check for the members according to 6.3 is necessary when
145
second-order analysis is used with member imperfections; and clause 6.1(1) of EN 1993-1-
1 then says that
M1
is used for instability assessed by member checks. This appears to
permit
M0
to be used when second-order analysis and cross-section checks are used, as
here, so
M0
1.0 is used.
The properties of the materials, in the usual notation, are as follows.
Structural steel: S355; f
y
f
yd
355 N/mm
2
, based on the use of Table 3.1 of EN1993-
1-1. (The UKs National Annex is likely to require the appropriate value to be taken from
EN 10025.) From EN 1993, E
a
210 kN/mm
2
.
Concrete: C40/50; f
ck
40 N/mm
2
; f
cd
40,1.5 26.7 N/mm
2
.
The coarse aggregate is limestone, so from clause 3.1.3(2) of EN 1992-1-1,
E
cm
0.9 35 31.5 kN/mm
2
; n
0
210,31.5 6.67.
600
35 35
680
6 mm cement-sand
bedding
11 500
Fig. 6.46. Cross-section at each end of concrete-lled tube of Example 6.12
Geometrical properties of the cross-section
In the notation of Fig. 6.17(e), d 750 mm, t 35 mm, giving d,t 21.4.
The limit, from Table 6.3, is d,t 90235,355 59.6.
From clause 6.7.5.2(1), normally no longitudinal reinforcement is necessary (unless
resistance to re is required), so A
s
0 is assumed.
For the concrete: area A
c
340
2
363 200 mm
2
I
c
340
4
,4 10 500 10
6
mm
4
For the steel tube: area A
a
375
2
363 200 78 620 mm
2
I
a
375
4
,4 10 500 10
6
5036 10
6
mm
4
Design action eects, ultimate limit state
For the most critical load arrangement, global analysis gives these values:
N
Ed
18.0 MN, of which N
G.Ed
13.0 MN (D6.26)
M
Ed.top
18 0.075 1.35 MNm (D6.27)
As the column has circular symmetry and its imperfection is assumed to lie in the plane of
bending, there is no need to consider biaxial bending. For a rocker bearing, the eective
length should be taken to the points of contact, and is:
L 11.5 2 0.6 12.7 m
Depending on the relationship between the rotation and the displacement of the
superstructure at the point of support, it can be shown that the points of contact
within the two bearings can be such that the ratio of end moments has any value
146
between 1 and 1. The worst case for the column, assumed here, is:
M
Ed.top
M
Ed.bot
(i.e., r 1 in Table 6.4). Double-curvature bending (r 1) causes higher transverse
shear force in the column, but it is too low to inuence the design.
Creep coecient
E
c
E
cm
,1 N
G.Ed
,N
Ed
c
t
(6.41)
The creep coecient c
t
is ct. t
0
to clause 5.4.2.2. The time t
0
is taken as 30 days, and t is
taken as innity, as creep reduces the stiness, and hence the stability, of a column.
From clause 3.1.4(5) of EN 1992-1-1, the perimeter exposed to drying is zero, so that
the notional size, h
0
!1. Assuming inside conditions and the use of normal cement, the
graphs in Fig. 3.1 of EN 1992-1-1 give:
c1. 30 1.4 c
t
The assumed age at rst loading has little inuence on the result if it exceeds about 20
days.
E
c.eff
31.5,1 1.413,18 15.7 kN/mm
2
(D6.28)
Shrinkage
When the upper bearing is placed, the concrete becomes eectively sealed, so its drying
shrinkage can be taken as zero. From EN 1992-1-1/3.1.4(6), the autogenous shrinkage is:
e
ca
11 2.5 f
ck
10 10
6
2.5 30 30 10
6
75 10
6
(D6.29)
Ignoring the restraint from the steel tube, this would cause the 11 m length of column to
shorten by only 0.8 mm. Its inuence on shear connection is discussed later.
Squash load, elastic critical load, and slenderness
N
pl.Rd
A
a
f
yd
A
c
f
cd
78.62 0.355 363.2 0.0267 27.9 9.70 37.6 MN D6.30
From clause 6.7.3.3(1), the steel contribution ratio is:
c 27.9,37.6 0.74
which is within the limits of clause 6.7.1(4).
For
"
`` to clause 6.7.3.3(2), with
C
1.5,
N
pl.Rk
27.9 1.5 9.70 42.46 MN (D6.31)
EI
eff
E
a
I
a
K
e
E
c.eff
I
c
210 5036 0.6 15.7 10 500 1.156 10
6
kNm
2
(6.40)
Hence,
N
cr

2
EI
eff
,L
2
1.156 10
3
2
,12.7
2
70.74 MN
"
`` N
pl.Rk
,N
cr
0.5
42.46,70.74
0.5
0.775
147
The non-dimensional slenderness does not exceed 2.0, so clause 6.7.3.1(1) is satised.
Clause 6.7.3.2(6), on the eect of containment on the strength of concrete, does not
apply, as
"
`` 0.5.
N
Rd
(MN)
M
Rd
(MNm)
N
Ed
= 18.0
N
pl,Rd
= 37.6
M
pl,Rd
= 6.884
N
pm,Rd
M
max,Ed
= 3149
20
0
4 2
30
10
A
4.836
6
4.85
9.700
7.057
B
C
D
Fig. 6.47. Interaction polygon for concrete-lled tube
Interaction polygon
The interaction polygon corresponding to Fig. 6.19 is determined using formulae that are
given and explained in Appendix C of Ref. 5. Clause 6.7.3.2(5) permits it to be used as an
approximation to the NM interaction curve for the cross-section.
The data are: h b d 750 mm; r d,2 t 375 35 340 mm.
From result (D6.30),
N
pl.Rd
27.9 9.70 37.6 MN
The equation numbers (C.) refer to those given in Ref. 5. From equation (C.29) with
W
ps
0,
W
pc

b 2t h 2t
2
4

2
3
r
3
680
3
,4 2 340
3
,3 52.4 10
6
mm
3
From equation (C.30):
W
pa

bh
2

4

2
3
r t
3
W
pc
750
3
,4 2 375
3
,3 52.4 10
6
17.91 10
6
mm
3
N
pm.Rd
A
c
f
cd
0.3632 26.7 9.70 MN
h
n

N
pm.Rd
A
sn
2 f
sd
f
cc
2b f
cd
4t2 f
yd
f
cc

9.70 10
6
2 750 26.7 1402 355 26.7
71.5 mm
W
pc.n
b 2th
2
n
750 70 71.5
2
3.474 10
6
mm
3
W
pa.n
bh
2
n
W
pc.n
750 71.5
2
3.474 10
6
0.358 10
6
mm
3
M
max.Rd
W
pa
f
yd
W
pc
f
cd
,2 17.91 355 52.4 26.7,2 7057 kNm
148
M
n.Rd
W
pa.n
f
yd
W
pc.n
f
cd
,2 0.358 355 3.474 26.7,2 173 kNm
M
pl.Rd
M
max.Rd
M
n.Rd
7057 173 6884 kNm
These results dene the interaction polygon, shown in Fig. 6.47.
Design maximum bending moment
From Table 6.4 with r 1, u 1, the equivalent rst-order bending moment is:
M
1st.Ed
1350 1.1 1485 kNm
From Table 6.5, the equivalent member imperfection is:
e
0
L,300 12 700,300 42.3 mm
For N
Ed
18 MN, the imperfection moment is 18 42.3 761 kNm
To check whether second-order moments can be neglected, an eective value of N
cr
is
required, to clause 6.7.3.4(3). From equation (6.42),
EI
eff.II
0.9E
a
I
a
0.5E
c
I
c
0.9 10
6
210 5036 0.5 15.7 10 500 1.026 10
12
kNmm
2
Hence,
N
cr.eff
1026
2
,12.7
2
62.8 MN
This is less than 10N
Ed
, so second-order eects must be allowed for. Table 6.4 gives u 1
for the distribution of bending moment due to the initial bowimperfection of the member,
so from clause 6.7.3.4(5), the second-order factor is:
1
1 N
Ed
,N
cr.eff
1
1 18,62.8
1.402
(The u factor for the end moments was accounted for in M
1st.Ed
.) Hence,
M
max.Ed
1.4021485 761 3149 kNm
Resistance of column
From Fig. 6.47 with N
Ed
18 MN, M
pl.N.Rd
4836 kNm.
From clause 6.7.3.6(1), the verication for uniaxial bending is:
M
Ed
,M
pl.N.Rd
0.9
Here, the ratio is 3149,4836 0.65, so the column is strong enough.
Shear connection and load introduction
This column is within the scope of clause 6.7.4.2(3), which permits shear connection to be
omitted. The signicance of this rule is now illustrated, using preceding results.
Creep increases shear transfer to the steel. Full-interaction elastic analysis with
n
L
n
0
1
L
c
t
17 (clause 5.4.2.2(2)) nds the action eects on the steel to be:
N
a.Ed
14.2 MN and M
a.Ed
1.2 MNm
based on M
Ed
1.35 MNm near an end of the column.
From the rules for shear connection, these transfers would require 90 25 mm studs for
the axial force plus 28 for the bending moment, assuming it can act about any horizontal
axis.
The signicance of friction is illustrated by the following elastic analysis assuming that
Poissons ratio for concrete is 0.5. The bearing stress on the concrete from N
Ed
18 MN
149
6.8. Fatigue
6.8.1. General
The only complete set of provisions on fatigue in EN 1994-2 is for stud shear connectors.
Fatigue in reinforcement, concrete and structural steel is covered mainly by cross-reference
to EN 1992 and EN 1993. Commentary will be found in the guides to those codes.
3.4
Further
cross-reference is necessary to EN1993-1-9,
42
Fatigue, which gives supplementary guidance
and fatigue detail classications which are not specic to bridges.
The fatigue life of steel components subjected to varying levels of repetitive stress can be
checked with the use of Miners summation. This is a linear cumulative damage calculation
for n stress ranges:
X
n
i 1
n
Ei
N
Ri
1.0 D6.32
where n
Ei
is the number of loading cycles of a particular stress range and N
Ri
is the number of
loading cycles to cause fatigue failure at that particular stress range. For most bridges, the
above is a complex calculation because the stress in each steel component usually varies
due to the random passage of vehicles from a spectrum. Details on a road or rail bridge
can be assessed using the above procedure if the loading regime is known at design. This
includes the weight and number of every type of vehicle that will use each lane or track of
the bridge throughout its design life, and the correlation between loading in each lane or
track. In general, this will produce a lengthy calculation.
As an alternative, clause 9.2 of EN 1993-2 allows the use of simplied Fatigue Load
Models 3 and 71, from EN 1991-2, for road and rail bridges respectively. This reduces the
complexity of the fatigue assessment calculation. It is assumed that the ctitious vehicle
(or train) alone causes the fatigue damage. The calculated stress from the vehicle is then
adjusted by factors to give a single stress range which, for N
cycles (2 million cycles for

structural steel), causes the same damage as the actual trac during the bridges lifetime.
This is called the damage equivalent stress range and is discussed in section 6.8.4 below.
Comments here are limited to the use of the damage equivalent stress method and, hence,
a single stress range.
is 49.6 N/mm
2
. Its resulting lateral expansion causes a hoop tensile stress in the steel of
225 N/mm
2
and a radial compression in the concrete of 23 N/mm
2
Fig. 6.48. Assuming
a coecient of friction of 0.4, the vertical frictional stress is:
t
Rd
23 0.4 9.2 N/mm
2
The shear transfer reduces the compressive stress in the concrete. Using a guessed mean
value of t
Rd
5 N/mm
2
leads to a transfer length for 14.2 MN of 1.33 m. This is less than
the introduction length of 2d ( 1.50 m here) permitted by clause 6.7.4.2(2).
These gures serve only to illustrate the type of behaviour to be expected. In practice, it
would be prudent to provide some shear connection; perhaps sucient for the bending
moment. Shrinkage eects are very small.
225 N/mm
2
225 N/mm
2
23 N/mm
2
23 N/mm
2
680
35
Fig. 6.48. Radial and hoop stresses near an end of a concrete-lled tube
150
The term equivalent constant-amplitude stress range, dened in clause 1.2.2.11 of
EN 1993-1-9, has the same meaning as damage equivalent stress range, used here and in
clause 6.8.5 of EN 1992-1-1 and clause 9.4.1 of EN 1993-2.
Fatigue damage is related mainly to the number and amplitude of the stress ranges as seen
in expression (D6.32). The peak of the stress range has a secondary inuence that can be, and
usually is, ignored in practice for peak stresses below about 60% of the characteristic
strength. Ultimate loads are higher than peak fatigue loads, and the use of partial safety
factors for ultimate-load design normally ensures that peak fatigue stresses are below this
limit. This may not be the case for long-span bridges with a high percentage of dead load,
so clause 6.8.1(3) species a limit to the longitudinal shear force per connector, with a
recommended value 0.75P
Rd
, or 0.6P
Rk
for
V
1.25. As fatigue damage to studs may
not be evident, some continental countries are understood to be specifying a lower limit,
0.6P
Rd
, in their national annexes. (For welded structural steel, the eect of peak stress is
eectively covered in the detail classications in EN 1993-1-9, where residual stresses from
welding, typically reaching yield locally, are catered for in the detail categories.)
Most bridges will require a fatigue assessment. Clauses 6.8.1(4) and (5) refer to EN 1993-2
and EN 1992-2 for guidance on the types of bridges and bridge elements where fatigue
assessment may not be required. Those relevant to composite bridge superstructures of steel
and concrete include:
(i) pedestrian footbridges not susceptible to pedestrian-induced vibration
(ii) bridges carrying canals
(iii) bridges which are predominantly statically loaded
(iv) parts of railway or road bridges that are neither stressed by trac loads nor likely to be
excited by wind loads
(v) prestressing and reinforcing steel in regions where, under the frequent combination of
actions and the characteristic prestress P
k
, only compressive stresses occur at the
extreme concrete bres. (The strain and hence the stress range in the steel is typically
small while the concrete remains in compression.)
Fatigue assessments are still required in the cases above (with the possible exception of (v)),
if bridges are found to be susceptible to wind-induced excitation. The main cause of
wind-induced fatigue, vortex shedding, is covered in EN 1991-1-4 and is not considered
further here.
6.8.2. Partial factors for fatigue assessment of bridges
Resistance factors
Mf
may be given in National Annexes, so only the recommended values
can be discussed here. For fatigue strength of concrete and reinforcement, clause 6.8.2(1)
refers to EN 1992-1-1, which recommends the partial factors 1.5 and 1.15, respectively,
for both persistent and transient design situations. For structural steel, EN 1993-1-9,
Table 3.1 recommends values ranging from 1.0 to 1.35, depending on the design concept
and consequence of failure. These apply, as appropriate, for a fatigue failure of a steel
ange caused by a stud weld. The choice of design concept and the uncertainties covered
by
Mf
are discussed in Ref. 4.
Fatigue failure of a stud shear connector, not involving the ange, is covered by EN 1994-2.
The recommended value of
Mf.s
for fatigue of headed studs is given as 1.0 in a Note to
clause 2.4.1.2(6) in the general rules of EN 1994. This is the value in EN 1993-1-9 for the
damage tolerant assessment method with low consequence of failure. From clause 3(2)
of EN 1993-1-9, the use of the damage tolerant method should be satisfactory, provided
that a prescribed inspection and maintenance regime for detecting and correcting fatigue
damage is implemented . . .. A Note to this clause states that the damage tolerant method
may be applied where in the event of fatigue damage occurring a load redistribution
between components of structural elements can occur.
The rst condition does not apply to stud connectors, as lack of access prevents detection
of small cracks by any simple method of inspection. For that situation, EN 1993-1-9
Clause 6.8.1(3)
Clause 6.8.1(4)
Clause 6.8.1(5)
Clause 6.8.2(1)
151
recommends use of the safe life method, with
Mf
1.15 for low consequence of failure.
The second condition does apply to stud connectors, and the value
Mf.s
1.0 is considered
to be appropriate for studs in bridges. Relevant considerations are as follows.
Fatigue failure results from a complex interaction between steel and concrete, commencing
with powdering of the highly stressed concrete adjacent to the weld collar. This displaces
upwards the line of action of the shear force, increasing the bending and shear in the shank
just above the weld collar, and probably also altering the tension. Initial fatigue cracking
further alters the relative stinesses and the local stresses. Research has found that the
exponent that relates the cumulative damage to the stress range may be higher than the
value, 5, for other welds in shear. The value chosen for EN 1994-2, 8, is controversial, as
discussed later.
As may be expected from the involvement of a tiny volume of concrete, tests show a wide
scatter in fatigue lives, which is allowed for in the design resistances. Studs are provided in
large numbers, and are well able to redistribute shear between themselves.
One reason for not recommending a partial factor more conservative than 1.0 comes from
experience with bridges, where stud connectors have been used for almost 50 years. When-
ever occasion has arisen in print or at a conference, the second author has stated that
there is no known instance of fatigue failure of a stud in a bridge, other than a few clearly
attributable to errors in design. This has not been challenged. Research has identied, but
not yet quantied, many reasons for this remarkable experience.
97.98
Most of them (e.g.
slip, shear lag, permanent set, partial interaction, adventitious connection from bolt
heads, friction) lead to predicted stress ranges on studs lower than those assumed in
design. With an eighth-power law, a 10% reduction in stress range more than doubles the
fatigue life.
6.8.3. Fatigue strength
The format of clause 6.8.3(3) for shear studs, as in EN 1993-1-9, uses a reference value of
range of shear stress at 2 million cycles, t
C
90 N/mm
2
. It denes the slope m of the
line through this point on the log-log plot of range of stress t
R
against number of
constant-range stress cycles to failure, N
R
, Fig. 6.25. Clause 6.8.3(4) modies the expression
slightly for lightweight-aggregate concrete.
It is a complex matter to deduce a value for m from the mass of test data, which are often
inconsistent. Many types of test specimen have been used, and the resulting scatter of results
must be disentangled from that due to inherent variability. Values for mrecommended in the
literature range from 5 to 12, mostly based on linear-regression analyses. The method of
regression used (x on y, or y on x) can alter the value found by up to 3.
97
The value 8, which was also used in BS 5400 Part 10, may be too high. In design for a
loading spectrum, its practical eect is that the cumulative damage is governed by the
highest-range components of the spectrum (e.g. by the small number of very heavy lorries
in the trac spectrum). A lower value, such as 5, would give more weight to the much
higher number of average-weight vehicles. This is illustrated in Example 6.13.
While fatigue design methods for stud connectors continue to be conservative (for bridges
and probably for buildings too) the precise value for m is of academic interest. Any future
proposals for more accurate methods for prediction of stress ranges should be associated
with re-examination of the value for m. Annex C gives a dierent design rule for fatigue of
lying studs, which is discussed in Chapter 10.
6.8.4. Internal forces and fatigue loadings
For fatigue assessment, it is necessary to nd the range or ranges of stress in a given material
at a chosen cross-section, caused by the passage of a vehicle along the bridge. Loading other
than the vehicle inuences the extent of cracking in the concrete, and hence, the stinesses of
members and the calculated stresses. Cracking depends mainly on the heaviest previous
loading, and so tends to increase with time. Clause 6.8.4(1) refers to clause 6.8.3 of
EN 1992-1-1. This denes the non-cyclic loading assumed to coexist with the design value
Clause 6.8.3(3)
Clause 6.8.3(4)
Clause 6.8.4(1)
152
of the cyclic load, Q
fat
: it is the frequent combination, represented by
X
j !1
G
k. j
P
1.1
Q
k.1
X
i 1
2.i
Q
k.i
where the Qs are non-cyclic variable actions.
Trac will usually be the leading non-cyclic action since the
2
value for trac recom-
mended in Annex A2 of EN 1990 is zero. With trac as the leading action, only thermal
actions have a non-zero value of
2
and therefore need to be considered.
The non-cyclic combination gives a mean stress level upon which the cyclic part of the
action eect is superimposed. The importance of mean stress is illustrated in Fig. 6.49 for
the calculation of stress range in reinforcement in concrete. It shows that the stress change
in the reinforcement for any part of the loading cycle that induces compression in the
concrete is much less than the stress change where the slab remains in tension throughout
the cycle.
Clause 6.8.4(2) denes symbols that are used for bending moments in clause 6.8.5.4. The
sign convention is evident from Fig. 6.26, which shows that M
Ed.max.f
is the bending moment
that causes the greatest tension in the slab, and is positive. Clause 6.8.4(2) also refers to
internal forces, but does not give symbols. Analogous use of calculated tensile forces in a
concrete slab (e.g. N
Ed.max.f
) may sometimes be necessary.
Clause 6.8.4(3) refers to Annex A.1 of EN 1993-1-9 for a general treatment of fatigue
based on summing the damage from a loading spectrum. As discussed in section 6.8.1
above, this would be a lengthy and complex calculation for most bridges and therefore
clauses 6.8.4(4) to (6) provide the option of using simpler load models from EN 1991-2.
The damage equivalent stress method for road bridges is based on Fatigue Load Model 3
dened in EN 1991-2 clause 4.6.4, while for rail bridges it is based on Load Model 71.
Clause 6.8.4(5) says that the additional factors given in EN 1992-2 clause NN.2.1 should
be applied to Load Model 3 where a road bridge is prestressed by tendons or imposed
deformations. As Annex NN is Informative, the situation is unclear in a country where
the National Annex does not make it available.
The load models and their application are discussed in the other guides in this series.
24
6.8.5. Stresses
Clause 6.8.5.1(1) refers to a list of action eects in clause 7.2.1(1)P to be taken into account
where relevant. They are all relevant, in theory, to the extent of cracking. However, this can
usually be represented by the same simplied model, chosen from clause 5.4.2.3, that is used
for other global analyses. They also inuence the maximum value of the fatigue stress range,
which is limited for each material (e.g. the limit for shear connectors in clause 6.8.1(3)).
The provisions for fatigue are based on the assumption that the stress range caused by a
given uctuation of loading, such as the passage of a vehicle of known weight, remains
approximately constant after an initial shakedown period. Shakedown here includes the
Clause 6.8.4(2)
Clause 6.8.4(3)
Clauses 6.8.4
to (6)
Clause 6.8.5.1(1)
Stress in
reinforcing bar
Tension
Time
Reduced stress range for
the same vehicle
Stress range from cyclic loads
(e.g. passage of fatigue load model)
Mean stress level
from non-cyclic loads
Compression
0
Fig. 6.49. Stress ranges for fatigue verication of reinforcement caused by the same cyclic action at
dierent mean stress levels
153
changes due to cracking, shrinkage, and creep of concrete, that occur mainly within the rst
year or two.
For bridges, most fatigue cycles occur over very short durations as the stress ranges are
produced either by the passage of vehicles or by wind-induced oscillations. Cycles of stress
from thermal actions also occur but over greater durations. The magnitude and small
number of these cycles do not generally cause any signicant fatigue damage. The short-
term modular ratio should therefore be used when nding stress ranges from the cyclic
action Q
fat
. Where a peak stress is being checked, creep from permanent loading should
be allowed for, if it increases the relevant stress.
The eect of tension stiening on the calculation of stress in reinforcement, clause
6.8.5.1(2)P and (3), is illustrated in Example 6.13 and discussed under clause 6.8.5.4
below. It is not conservative to neglect tension stiening in this calculation for a composite
beam as the increased stiness attracts more stress to the concrete slab and hence to the
reinforcement between cracks. For stresses in structural steel, the eects of tension stiening
may be included or neglected in accordance with clause 6.8.5.1(4). Tension stiening here
has a benecial eect in reducing the stresses in the structural steel. Tension stiening
should also be considered in deriving stresses for prestressing steel clause 6.8.5.1(5).
For analysis, the linear-elastic method of Section 5 is used, from clause 6.8.4(1). Clause
7.2.1(8) requires consideration of local and global eects in deck slabs. This is also reected
in clause 6.8.6.1(3). When checking fatigue, it is important to bear in mind that the most
critical areas for fatigue may not be the same as those for other ultimate limit state calcula-
tions. For example, the critical section for shear connection may be near mid-span, since its
provision is usually based on the static design, and the contribution to the static shear from
dead load is zero there.
Concrete
For concrete, clause 6.8.5.2(1) refers to clause 6.8 of EN 1992-1-1, where clause 6.8.5(2)
refers to EN 1992-2. EN 1992-2 clause 6.8.7(101) provides a damage equivalent stress
range method presented as for a spectrum. The method of its Annex NN is not applicable
to composite members. As a simpler alternative, EN 1992-1-1 clause 6.8.7(2) gives a conser-
vative verication based on the non-cyclic loading used for the static design. It will usually be
sucient to apply this verication to composite bridges as it is unlikely to govern design
other than possibly for short spans where most of the compressive force in concrete is pro-
duced by live load.
Structural steel
Clause 6.8.5.3(1) repeats, in eect, the concession in clause 6.8.5.1(4). Where the words or only
M
Ed.min.f
in clause 6.8.5.3(2) apply, M
Ed.max.f
causes tension in the slab. The use of the
uncracked section for M
Ed.max.f
would then underestimate the stress ranges in steel anges, so
that cracked section properties should be used for the calculation of this part of the stress range.
Reinforcement
For reinforcement, clause 6.8.3(2) refers to EN 1992-1-1, where clause 6.8.4 gives the
verication procedure. Its recommended value N
for straight bars is 10

6
. This should not
be confused with the corresponding value for structural steel in EN 1993-1-9, 2 10
6
,
denoted N
C
, which is used also for shear connectors, clause 6.8.6.2(1).
Using the values recommended in EN 1992-1-1, its expression (6.71) for verication of
reinforcement becomes:
o
E.equ
N
o
Rsk
N
,1.15 (D.6.33)
with o
Rsk
162.5 N/mm
2
for N
10
6
, from Table 6.3N.
Where a range o
E
(N
E
) has been determined, the resistance o
Rsk
(N
E
) can be found from
the SN curve for reinforcement, and the verication is:
o
E
N
E
o
Rsk
N
E
,1.15 (D6.34)
Clause 6.8.5.1(2)P
Clause 6.8.5.1(3)
Clause 6.8.5.1(4)
Clause 6.8.5.1(5)
Clause 6.8.5.2(1)
Clause 6.8.5.3(1)
Clause 6.8.5.3(2)
154
Clause 6.8.5.4(1) permits the use of the approximation to the eects of tension stiening
that is used for other limit states. It consists of adding to the maximum tensile stress in the
fully cracked section, o
s.o
, an amount o
s
that is independent of o
s.o
. The value of o
s
for
fatigue verication is modied by replacing the factor of 0.4 in equation (7.5) by 0.2. This is
to allow for the reduction in tension stiening caused by repeated cycles of tensile stress.
99
Clauses 6.8.5.4(2) and (3) give simplied rules for calculating stresses, with reference to
Fig. 6.26, which is discussed using Fig. 6.50. This has the same axes, and also shows a
minimum bending moment that causes compression in the slab. A calculated value for the
stress o
s
in reinforcement, that assumes concrete to be eective, would lie on line A0D.
On initial cracking, the stress o
s
jumps from B to point E. Lines 0BE are not shown in
Fig. 6.26 because clause 7.2.1(5)P requires the tensile strength of concrete to be neglected
in calculations for o
s
. This gives line 0E. For moments exceeding M
cr
, the stress o
s
follows
route EFG on rst loading. Calculation of o
s
using section property I
2
gives line 0C. At
bending moment M
Ed.max.f
the stress o
s.o
thus found is increased by o
s
, from equation
(7.5), as shown by line HJ.
Clause 6.8.5.4 denes the unloading route from point J as J0A, on which the stress o
s.min.f
lies. Points K and L give two examples, for M
Ed.min.f
causing tension and compression,
respectively, in the slab. The fatigue stress ranges o
s.f
for these two cases are shown.
Shear connection
The interpretation of clause 6.8.5.5(1)P is complex when tension stiening is allowed for.
Spacing of shear connectors near internal supports is unlikely to be governed by fatigue,
so it is simplest to use uncracked section properties when calculating range of shear ow
from range of vertical shear, clause 6.8.5.5(2). These points are illustrated in Example 6.13.
Reinforcement and prestressing steel in members prestressed by bonded tendons
Where bonded prestress is present, stresses should be determined in a similar manner to
the above for reinforcement alone, but account needs to be taken of the dierence of
bond behaviour between prestressing steel and reinforcement clause 6.8.5.6(1). Clause
6.8.5.6(2) makes reference to clause 7.4.3(4) which in turn refers to clause 7.3 of
EN 1992-1-1 for the calculation of stresses o
s
. This is a generic symbol here, and so
applies to the stress o
s.max.f
referred to in clause 6.8.5.6(2).
6.8.6. Stress ranges
Clause 6.8.6.1 is most relevant to the damage equivalent stress method where the complex
cyclic loadings from a spectrum of vehicles are condensed into one single stress range
which, for N
cycles, is intended to give the same damage during the bridges lifetime as
Clause 6.8.5.4(1)
Clause 6.8.5.4(2)
Clause 6.8.5.4(3)
Clause 6.8.5.5(1)P
Clause 6.8.5.5(2)
Clause 6.8.5.6(1)
Clause 6.8.5.6(2)
Clause 6.8.6.1
D
C
G
0
L
B M
Ed,min,f M
Ed,max,f
M
J
A
E
M
Ed,min,f
F
K
H
M
cr

s,max,f
s,f
s,o

s,min,f

s
(tension)
Fig. 6.50. Stress ranges in reinforcement in cracked regions
155
the real trac. This stress range is determined by applying the relevant fatigue load model
discussed in section 6.8.4 and by multiplying it by the damage equivalent factor `, according
to clause 6.8.6.1(2). The factor ` is a property of the spectrum and the exponent m, which is
the slope of the fatigue curve as noted in clause 6.8.6.1(4).
Deck slabs of composite bridge beams are usually subjected to combined global and local
fatigue loading events, due to the presence of local wheel loads. The eects of local and
global loading are particularly signicant in reinforcement design in slabs adjacent to
cross-beams supporting the deck slab, in zones where the slab is in global tension. Here,
wheel loads cause additional local hogging moments. Clause 6.8.6.1(3) provides a conserva-
tive interaction where the damage equivalent stress range is determined separately for the
global and local actions and then summed to give an overall damage equivalent stress range.
In combining the stress ranges in clause 6.8.6.1(3), it is important to consider the actual
transverse location being checked within the slab. The peak local eect usually occurs
some distance from the web of a main beam, while the global direct stress reduces away
from the web due to shear lag. The reduction may be determined using clause 5.4.1.2(8),
even though that clause refers to EN 1993-1-5, which is for steel anges.
A similar damage equivalent factor, `
v
, is used in clause 6.8.6.2(1) to convert the shear
stress range in the studs from the fatigue load model into a damage equivalent stress range.
For other types of shear connection clause 6.8.6.2(2) refers to Section 6 of EN 1993-1-9.
This requires the damage equivalent stress to be determined from its Annex A using the
actual trac spectrum and Miners summation. This approach could also be used for
shear studs as an alternative, provided that m is taken as 8, rather than 3.
For connectors other than studs, the authors recommend that the method of Annex A be
used only where the following conditions are satised:
.
the connectors are attached to the steel ange by welds that are within the scope of
EN 1993-1-9
.
the fatigue stress ranges in the welds can be determined realistically
.
the stresses applied to concrete by the connectors are not high enough for fatigue failure
of the concrete to inuence the fatigue life.
The exponent m should then have the value given in EN 1993-1-9; m 8 should not be used.
In other situations, fatigue damage to concrete could inuence the value of m. The
National Annex may refer to guidance, as permitted by the Note to clause 1.1.3(3).
Clauses 6.8.6.2(3) to (5) provide a method of calculating the damage equivalent factors
for studs. With the exception of `
v.1
, those for road bridges are based on those in EN 1993-2
clause 9.5.2, but with the exponents modied to 8 or
1
8
as discussed in section 6.8.3.
In EN 1993-2, an upper limit to ` is dened in clause 9.5.2, in paragraphs that EN 1994-2
does not refer to. This is because the upper limit is not required for stud shear connectors.
6.8.7. Fatigue assessment based on nominal stress ranges
Comment on the methods referred to from clause 6.8.7.1 will be found in other guides in this
series. The term nominal stress range in the heading of clause 6.8.7 is dened in Section 6 of
EN 1993-1-9 for structural steel. It is the stress range that can be compared directly with the
detail categories in EN 1993-1-9. It is not the stress range before the damage equivalent
factors are applied. It is intended to allow for all stress concentration factors implicit
within the particular detail category selected. If additional stress concentrating details
exist adjacent to the detail to be checked which are not present in the detail category selected
(e.g. a hole), these additional eects need to be included via an appropriate stress concentra-
tion factor. This factored stress range then becomes a modied nominal stress range as
dened in clause 6.3 of EN 1993-1-9.
For shear connectors, clause 6.8.7.2(1) introduces the partial factors. The recommended
value of
Mf.s
is 1.0 (clause 2.4.1.2(6)). For
Ff
, EN 1990 refers to the other Eurocodes. The
recommended value in EN 1992-1-1, clause 6.8.4(1), is 1.0. Clause 9.3(1) of EN 1993-2
recommends 1.0 for steel bridges.
Clause 6.8.6.1(2)
Clause 6.8.6.1(3)
Clause 6.8.6.2(1)
Clause 6.8.6.2(2)
Clauses 6.8.6.2(3)
to (5)
Clause 6.8.7.1
Clause 6.8.7.2(1)
156
Clause 6.8.7.2(2) covers interaction between the fatigue failures of a stud and of the steel
ange to which it is welded, where the ange is in tension. The rst of expressions (6.57) is
the verication for the ange, fromclause 8(2) of EN1993-1-9, and the second is for the stud,
copied from equation (6.55). The linear interaction condition is given in expression (6.56).
It is necessary to calculate the longitudinal stress range in the steel ange that coexists with
the stress range for the connectors. The load cycle that gives the maximum value of o
E.2
in
the ange will not, in general, be that which gives the maximum value of t
E.2
in a shear
connector, because the rst is caused by exure and the second by shear. Also, both o
E.2
and t
E.2
may be inuenced by whether the concrete is cracked, or not.
It thus appears that expression (6.56) may have to be checked four times. In practice, it is
best to check rst the conditions in expression (6.57). It should be obvious, for these,
whether the cracked or the uncracked model is the more adverse. Usually, one or both
of the left-hand sides is so far below 1.0 that no check to expression (6.56) is needed.
Clause 6.8.7.2(2)
Example 6.13: fatigue verication of studs and reinforcement
The bridge shown in Fig. 6.22 is checked for fatigue of the shear studs at an abutment and
of the top slab reinforcement at an internal support. The Client requires a design life of
120 years. Fatigue Load Model 3 of EN 1991-2 is used. The bridge will carry a road in
Trac Category 2 of Table 4.5(n) of EN 1991-2, roads with medium ow rates of
lorries. The table gives the indicative number of heavy vehicles expected per year and
per slow lane as 500 000, and this value is used. The safe life method (dened in
clause 3(7)(b) of EN 1993-1-9) is used, as this is likely to be recommended by the UKs
National Annex.
Studs at an abutment
The cross-section of an inner beam at the abutments is as shown for length DE in Fig.
6.22. Groups of three 19 mm studs are provided, Fig. 6.41, at 150 mm spacing. The
special vehicle of Load Model 3 is dened in clause 4.6.4 of EN 1991-2. For this
cross-section its passage produces maximum and minimum unfactored vertical shears
of 235 kN and 19 kN. Since the detail is adjacent to an expansion joint, these values
should be increased by a factor of 1.3 in accordance with EN 1991-2 Fig. 4.7, so the
shear range becomes 1.3 235 19 330 kN.
The short-termuncracked properties of the composite beamare used for the calculation
of shear ow. From Table 6.3 in Example 6.10, A" zz,I 0.810 m
1
. The range of shear
force per connector is:
0.810 330 0.150,3 13.4 kN
The shear stress range for the connector is:
t
13.4 10
3
19
2
,4
47.1 N,mm
2
To determine the damage equivalent stress range, the factor
`
v
`
v.1
`
v.2
`
v.3
`
v.4
should be calculated in accordance with clause 6.8.6.2(3). From clause 6.8.6.2(4),
`
v.1
1.55. The remaining factors are calculated from EN 1993-2 clause 9.5.2 using
exponents 8 and
1
8
in place of those given.
For `
v.2
it would be possible to use the recommended data for Load Model 4 in Tables
4.7 and 4.8 of EN 1991-2. However, the UKs National Annex to EN 1991-2 is likely to
replace these with the BS 5400 Part 10 data, which are given in Table 6.5.
From clause 9.5.2(3) of EN 1993-2:
Q
m1

P
n
i
Q
5
i
P
n
i
!
1,5
8.051 10
18
1.000 10
6
!
1,5
381.2 kN for checks on structural steel
157
Q
m1

P
n
i
Q
8
i
P
n
i
!
1,8
3.384 10
29
1.000 10
6
!
1,8
873.3 kN for checks on shear studs
It can be seen from the above that the contribution of the small number of very heavy
vehicles is much more signicant when the exponent 8 is used.
Table 6.5. Vehicle spectrum for fatigue verications
Vehicle
ref.
Weight (kN)
(Q
i
)
No. per million
vehicles (n
i
) (n
i
Q
5
i
) (n
i
Q
8
i
)
18GTH 3680 10 6.749 10
18
3.363 10
29
18GTM 1520 30 2.434 10
17
8.548 10
26
9TT-H 1610 20 2.164 10
17
9.029 10
26
9TT-M 750 40 9.492 10
15
4.005 10
24
7GT-H 1310 30 1.157 10
17
2.602 10
26
7GT-M 680 70 1.018 10
16
3.200 10
24
7A-H 790 20 6.154 10
15
3.034 10
24
5A-H 630 280 2.779 10
16
6.948 10
24
5A-M 360 14 500 8.768 10
16
4.091 10
24
5A-L 250 15 000 1.465 10
16
2.289 10
23
4A-H 335 90 000 3.797 10
17
1.428 10
25
4A-M 260 90 000 1.069 10
17
1.879 10
24
4A-L 145 90 000 5.769 10
15
1.759 10
22
4R-H 280 15 000 2.582 10
16
5.667 10
23
4R-M 240 15 000 1.194 10
16
1.651 10
23
4R-L 120 15 000 3.732 10
14
6.450 10
20
3A-H 215 30 000 1.378 10
16
1.370 10
23
3A-M 140 30 000 1.613 10
15
4.427 10
21
3A-L 90 30 000 1.771 10
14
1.291 10
20
3R-H 240 15 000 1.194 10
16
1.651 10
23
3R-M 195 15 000 4.229 10
15
3.136 10
22
3R-L 120 15 000 3.732 10
14
6.450 10
20
2R-H 135 170 000 7.623 10
15
1.876 10
22
2R-M 65 170 000 1.972 10
14
5.417 10
19
2R-L 30 180 000 4.374 10
12
1.181 10
17
Totals 1.000 10
6
8.051 10
18
3.384 10
29
For a two-lane road, trac category 2, N
Obs
0.5 10
6
. (The UKs National Annex to
EN 1991-2 may modify this value.)
Also from clause 9.5.2(3), N
0
0.5 10
6
Q
0
480 kN (weight of vehicle for Fatigue Load Model 3)
Therefore
`
v.2

Q
m1
Q
0
N
Obs
N
0

1,8
873.3
480
0.5 10
6
0.5 10
6
!
1,8
1.819
`
v.3
t
Ld
,100
1,8
1.023 for the 120-year design life.
`
v.4
1
N
2
N
1
j
2
Q
m2
j
1
Q
m1

8
N
3
N
1
j
3
Q
m3
j
1
Q
m1

8
. . .
N
k
N
1
j
k
Q
mk
j
1
Q
m1

8
" #
1,8
158
The inuence coecient from lane 2 is approximately 75% of that from lane 1. As both
lanes are slow lanes with N 0.5 10
6
vehicles per year,
`
v.4
1
N
2
N
1
j
2
Q
m2
j
1
Q
m1

8
" #
1,8
1
0.5 10
6
0.5 10
6
0.75
1.0

8
" #
1,8
1.012
`
v
`
v.1
`
v.2
`
v.3
`
v.4
1.55 1.819 1.023 1.012 2.92
From clause 6.8.6.2(1), t
E.2
`
v
t 2.92 47.1 138 N,mm
2
Ff
t
E.2
1.0 138 138 N,mm
2
From clause 6.8.3(3), t
c
90 N/mm
2
, so the fatigue resistance is:
t
c
,
Mf.s
90,1.0 90 N,mm
2
The shear studs are therefore not adequate and would need to be increased. There is no
need to check the interaction in clause 6.8.7.2(2) as the stress in the steel ange is small
and compressive at an abutment.
Fatigue of reinforcement, global eects
Note: throughout this Example, all cross-references commencing NN are to Annex NN
of EN 1992-2, Damage equivalent stresses for fatigue verication.
The cross-section at an intermediate support is shown in Fig. 6.22. For these cross-
sections, the axle loads of Fatigue Load Model 3 should be multiplied by 1.75 according
to clause NN.2.1(101). The maximum hogging moment from the fatigue vehicle was
1.75 593 1038 kNm and the minimum was 1.75 47 82 kNm.
In this calculation, the maximum and minimum moments occurred with the vehicle in
the same lane. Previous practice in the UK has been to calculate the stress range by allow-
ing the maximum and minimum eects from the vehicle to come from dierent lanes.
Clause 4.6.4(2) of EN 1991-2 however implies that the maximum stress range should be
calculated as the greatest stress range produced by the passage of the vehicle along any
one lane. The UKs draft National Annex currently requires the former interpretation
(the safer of the two) to be used, but there is no national provision in EN 1991-2 for
this to be done.
The maximum hogging moment on the composite section from the frequent combi-
nation was found to be 3607 kNm. This includes the eects of superimposed dead load,
secondary eects of shrinkage, settlement, thermal actions and trac load (load group
1a). Trac was taken as the leading variable action and hence the combination factors
applied were
1
for trac loading and
2
for thermal actions. Wind was not considered,
as the recommended value of
2
is zero from EN 1990 Table A2.1.
From clause 6.8.4(1), the minimum moment from the cyclic non-cyclic loading is:
M
Ed.min.f
3607 82 3525 kNm
and the maximum is:
M
Ed.max.f
3607 1038 4645 kNm
From clause 7.4.3(3) as modied by clause 6.8.5.4(1), the increase in stress in the re-
inforcement, due to tension stiening, above that calculated using a fully cracked analysis is:
o
s

0.2 f
ctm
c
st
,
s
where c
st

AI
A
a
I
a
74 478 22 660
55 000 12 280
2.50
The reinforcement ratio ,
s
0.025 and f
ctm
2.9 N/mm
2
and thus o
s
9.3 N/mm
2
.
From Example 6.6, the section modulus for the top layer of reinforcement is
34.05 10
6
mm
3
. Therefore the stress due to M
Ed.max.f
ignoring tension stiening is:
o
s.o
4645,34.05 136.4 N/mm
2
159
From equation (7.4), the stress including allowance for tension stiening is:
o
s.max.f
136.4 9.3 146 N/mm
2
o
s.min.f
136.4 3525,4645 9.3 113 N/mm
2
The damage equivalent parameters are next calculated from Annex NN. Figure NN.1
refers to the length of the inuence line. This length is intended to be the length of the
lobe creating the greatest stress range. EN 1993-2 clause 9.5.2 provides denitions of
the critical length of the inuence line for dierent situations and these can be referred
to. For bending moment at an internal support, the average length of the two adjacent
spans may be used, but here the length of the main span has been conservatively used.
From Fig. NN.1 for straight reinforcing bars (curve 3) and critical length of the inuence
line of 31 m, `
s.1
0.98.
From equation (NN.103):
`
s.2

"
QQ
N
obs
,2.0
k
2
p
with N
obs
in millions (which is not stated)
From Table 4.5(n) of EN 1991-2, N
obs
0.5 10
6
.
From EN 1992-1-1/Table 6.3N, k
2
9 for straight bars.
The factor for trac type,
"
QQ, is given in Table NN.1, but no guidance is given on its selec-
tion. Trac type is dened in Note 3 of EN 1991-2 clause 4.6.5(1). The denitions given
are not particularly helpful:
.
long distance means hundreds of kilometres
.
medium distance means 50100 km
.
local trac means distances less than 50 km.
Long distance will typically apply to motorways and trunk roads. The use of either of the
lower categories should be agreed with the Client as the trac using a road may not be
represented by a typical length of journey. Long distance trac is conservatively used
here, so from Table NN.1,
"
QQ 1.0. Thus,
`
s.2
1.0
0.5,2.0
9
p
0.86
From equation (NN.104): `
s.3

N
Years
,100
k
2
p

120,100
9
p
1.02
From equation (NN.105): `
s.4

P
N
obs.i
N
obs.1
k
2
s
.
Since both lanes are slow lanes, from Table 4.5 of EN 1991-2,
N
obs.1
N
obs.2
0.5 10
6
and therefore
`
s.4

9
0.5 0.5
0.5
r
1.08
From clause NN.2.1(108) and then EN 1991-2, Annex B, the damage equivalent impact
factor for surfaces of good roughness (i.e. regularly maintained surfaces) is
fat
1.2.
`
s

fat
`
s.1
`
s.2
`
s.3
`
s.4
1.20 0.98 0.86 1.02 1.08 1.11
From clause 6.8.6.1(2) and (7):
o
E
`c o
max.f
o
min.f

1.11 1.0 146 113 37 N/mm
2
There is an inconsistency here between EN1992-2, where `
s
includes the impact factor
fat
and EN 1994-2 where ` excludes this factor, which is written as c.
160
6.9. Tension members in composite bridges
The terms concrete tension member and composite tension member used in this clause are
dened in clause 5.4.2.8. Global analysis for action eects in these members and determina-
tion of longitudinal shear is discussed in comments on that clause.
Clause 6.9(1) concerns members that have tensile force introduced only near their ends. It
refers to their design to EN 1992, as does clause 6.9(2), with reference to simplications
Clause 6.9(1)
Clause 6.9(2)
From Table 6.3N of EN 1992-1-1, for straight bars, N
10
6
and o
Rsk
10
6

162.5 N/mm
2
.
The verication is carried out using expression (6.71) of EN 1992-1-1, taking
o
s.equ
N
o
E
above:
F.fat
o
s.equ
N
1.0 37 37 N,mm
2
o
Rsk
N
s.fat
162.5,1.15 141 N,mm
2
The reinforcement has adequate fatigue life under global loading.
Fatigue of reinforcement, local eects
Local bending moments are caused by hogging of the deck slab over the pier diaphragm.
The worst local eects are here conservatively added to the global eects according to
expression (6.53). A stress range of 44 N/mm
2
in the reinforcement was determined
from the maximum hogging moment caused by the passage of the factored fatigue
vehicle of EN 1992-2 Annex NN. (The moments were found using Puchers inuence
surfaces.
100
) Cracked section properties were used for the slab in accordance with
clause 6.8.2(1)P of EN 1992-1-1 since the slab remains in tension when the local eect
is added under the basic combination plus the cyclic action as dened in clause
6.8.3(3) of EN 1992-1-1.
The damage equivalent factors from EN 1992-2 Annex NN are the same as above with
the exception of `
s.1
. For local load, the critical length of the inuence line is the length
causing hogging moment each side of the pier diaphragm. From the Pucher chart used
to determine the local moment, the inuence of loads applied more than 6 m from the
pier diaphragm is approximately zero so the total inuence line length to consider is
approximately 12 m. From Fig. NN.1, `
s.1
0.91.
`
s
c
fat
`
s.1
`
s.2
`
s.3
`
s.4
1.20 0.91 0.86 1.02 1.08 1.03
From clause 6.8.6.1(2) and (7):
o
E
`c o
max.f
o
min.f

1.03 1.0 44 45 N,mm
2
Verifying as for the global loading:
F.fat
o
s.equ
N
1.0 45 45 N,mm
2
o
Rsk
N
s.fat
162.5,1.15 141 N,mm
2
The reinforcement has adequate fatigue life under local loading.
Fatigue of reinforcement, combined global and local eects
The simple interaction method of clause 6.8.6.1(3) is used. This entails summing the
damage equivalent stresses from the local and global loading, so that the total damage
equivalent stress o
E
37 45 82 N,mm
2
. In this case, the locations of peak global
stress in the reinforcement and peak local stress do coexist because there was no reduction
to the slab width for shear lag. The verication is now:
82 N/mm
2
141 N/mm
2
The reinforcement has adequate fatigue life under the combined loading.
161
given in clause 5.4.2.8. Clause 6.9(2) applies also to composite tension members, which have
shear connectors throughout their length. The subsequent paragraphs concern the distribu-
tion of the connectors along the member.
A plan of the steel members at deck level near one end of a bowstring arch bridge is shown
in Fig. 6.51(a). The arch applies concentrated forces T at points A and B. The force at A is
shared between the steel tie AC and the composite deck, shown shaded. The deck has steel
edge members such as DE, and spans longitudinally between composite cross-beams FG, etc.
The proportion of each force T that is resisted by the deck structure, T
d
say, depends on the
extent to which its stiness is reduced by cracking of the concrete. The force T
d
is applied to
the deck by diagonal members such as FH. The stiness of these members also inuences the
magnitude of T
d
. Details of bridges of this type are available elsewhere.
101
In some bridges, the deck is shear-connected directly to the main tie member, as shown in
Fig. 6.51(b). Clause 6.9(3) requires the shear connection for the force T
d
to be provided
within the lengths 1.5b shown.
The precise distribution of the connectors along a length such as JK has been studied,
using the rules of EN 1994-2 for tension stiening.
102
In this bridge, Newark Dyke, the
arch is the top chord of a truss of span 77 m with diagonals that apply longitudinal force
also at points such as L in Fig. 6.51(b). Neither paragraph (3) nor (6) denes the length
a over which shear connection near point L should be provided, but clause 6.6.2.3 provides
guidance.
The number of connectors to be provided over a length such as JK can be conservatively
found by assuming the deck to be uncracked. In this bridge, the design ultimate force T was
about 18 MN, and uncracked analysis gave T
d
% 9 MNat mid-span. Fully cracked analysis
gave this force as about 5 MN. Accurate analysis found the deck to be in a state of single
cracking (explained in comments on clause 5.4.2.8(6)), with a tensile force of about 8 MN.
Lower levels of shear connection are, of course, required along the whole length of the
deck for other combinations and arrangements of variable actions.
Clause 6.9(4)P, a principle that corresponds to clause 6.7.4.1(1)P for compression
members, is followed by application rules. For laterally loaded tension members, shear
connection within the length is related to the transverse shear in clause 6.9(5), exactly as
for composite beams.
Where axial tension is applied to the ends of a member through only one material, steel or
concrete, the length over which part of the tension should be transferred to the other material
(typically by shear connection), is limited by clause 6.9(6). This corresponds to clause
6.7.4.2(2) for compression members. Other provisions of clause 6.7.4.2 may be relevant here.
Clause 6.9(3)
Clause 6.9(4)P
Clause 6.9(5)
Clause 6.9(6)
T
T
T
T
2b
1.5b 1.5b
2b
A
G
F
E
D
C
B
J K
H
L
a
(b) (a)
Fig. 6.51. Shear connection to the deck of a bowstring arch bridge
162
CHAPTER 7
Serviceability limit states
.
General Clause 7.1
.
Stresses Clause 7.2
.
Deformations in bridges Clause 7.3
.
Cracking of concrete Clause 7.4
.
Filler beam decks Clause 7.5
7.1. General
Section 7 of EN1994-2 is limited to provisions on serviceability that are specic to composite
structures and are not in Sections 1, 2, 4 or 5 (for global analysis), or in Eurocodes 1990,
1991, 1992 or 1993. Some of these other provisions are briey referred to here. Further
comments on them are in other chapters of this book, or in other guides in this series.
The initial concept for a composite bridge is mainly inuenced by the intended method
of construction, durability, ease of maintenance, and the requirements for ultimate limit
states. Serviceability criteria that should be considered at an early stage are stress limits
in cross-sections in Class 1 or 2 and susceptibility to excessive vibration. It should not
however be assumed that Class 3 and 4 cross-sections require no checks of stress limits
at serviceability. For example, if torsional warping or St Venant torsional eects have
been neglected at ultimate limit state (ULS), as allowed by a reference in clause
6.2.7.2(1) of EN 1993-2, then the serviceability limit state (SLS) stresses should be
checked taking these torsional eects into account. Considerations of shear lag at SLS
may also cause unacceptable yielding as the eective widths of steel elements are greater
at ULS.
Control of crack width can usually be achieved by appropriate detailing of reinforcement.
Provision of re resistance and limiting of deformations have less inuence at this stage than
in structures for buildings. The important deformations are those caused by imposed load.
Limits to these inuence the design of railway bridges, but generally, stiness is governed
more by vibration criteria than by limits to deection.
The drafting of the serviceability provisions in the Eurocodes is less prescriptive than for
other limit states. It is intended to give designers and clients greater freedom to take account
of factors specic to the project.
The content of Section 7 was also inuenced by the need to minimize calculations. Results
already obtained for ultimate limit states are scaled or reused wherever possible. Experienced
designers know that many structural elements satisfy serviceability criteria by wide margins.
For these, design checks should be simple, and it does not matter if they are conservative. For
other elements, a longer but more accurate calculation may be justied. Some application
rules therefore include alternative methods.
Clause 7.1(1)P and (2) refer to clause 3.4 of EN1990. This gives criteria for placing a limit
state within the serviceability group, with reference to deformations (including vibration),
durability, and the functioning of the structure. The relevance of EN 1990 is not limited to
the clauses referred to, because clause 2.1(1)P requires design to be in accordance with the
general rules of EN 1990. This means all of it except annexes that are either informative or
not for bridges.
Serviceability verication and criteria
The requirement for a serviceability verication is given in clause 6.5.1(1)P of EN 1990 as:
E
d
C
d
where E
d
is the design value of the eects of the specied actions and the relevant combina-
tion, and C
d
is the limiting design value of the relevant criterion.
From clause 6.5.3 of EN 1990, the relevant combination is normally the characteristic,
frequent, or quasi-permanent combination, for serviceability limit states that are respectively
irreversible, reversible, or a consequence of long-term eects. The quasi-permanent
combination is also relevant for the appearance of the structure.
For bridges, rules on combinations of actions are given in clause A2.2 of EN 1990. Its
clause A2.2.2(1) denes a fourth combination, infrequent, for use for concrete bridges. It
is not used in EN 1994-2, but may be invoked by a reference to EN 1992, or found in a
National Annex.
Clause 7.1(3) refers to environmental classes. These are the exposure classes of EN1992,
and are discussed in Chapter 4. The exposure class inuences the cover to reinforcing bars,
and the choice of concrete grade and hence the stress limits.
Clause 7.1(4) on serviceability verication gives no detailed guidance on the extent to
which construction phases should be checked. The avoidance of excessive stress is one
example. Yielding of steel can cause irreversible deformation, and handling of precast
components can cause yielding of reinforcement or excessive crack width. Bridges can also
be more susceptible to aerodynamic oscillation during erection. In extreme cases, this can
lead to achievement of an ultimate limit state.
Clause 7.1(5) refers to the eight-page clause A2.4 of EN 1990. It covers partial factors,
serviceability criteria, design situations, comfort criteria, deformations of railway bridges
and criteria for the safety of rail trac. Few of its provisions are quantied. Recommended
values are given in Notes, as guidance for National Annexes.
The meaning of clause 7.1(6) on composite plates is that account should be taken of
Section 9 when applying Section 7. There are no serviceability provisions in Section 9.
No serviceability limit state of excessive slip of shear connection is dened. Generally, it
is assumed that clause 6.8.1(3), which limits the shear force per connector under the
characteristic combination, and other rules for ultimate limit states, will ensure satisfactory
performance in service.
No serviceability criteria are specied for composite columns, so from here on, this
chapter is referring to composite beams or plates or, in a few places, to composite
frames.
7.2. Stresses
Excessive stress is not itself a serviceability limit state. Stresses in bridges are limited to ensure
that under normal conditions of use, assumptions made in design models (e.g. linear-elastic
behaviour) remain valid, and to avoid deterioration such as the spalling of concrete or
disruption of the corrosion protection system.
The stress ranges in a composite structure caused by a particular level of imposed loading
take years to stabilise, mainly because of the cracking, shrinkage and creep of concrete.
Stress limits are also intended to ensure that after this initial period, live-load behaviour is
reversible.
Clause 7.1(1)P
Clause 7.1(2)
Clause 7.1(3)
Clause 7.1(4)
Clause 7.1(5)
Clause 7.1(6)
164
For the calculation of stresses, the principle of clause 7.2.1(1)P says, in eect, take
account of everything that is relevant. It is thus open to interpretation, subject to the
guidance in the rest of clause 7.2.1. Four of its paragraphs are worded may.
For persistent design situations, it is usual to check stresses soon after the opening of the
bridge to trac, ignoring creep, and also at a time when further eects of creep and shrinkage
have become negligible. Their values are usually found by letting t !1when applying the
data on creep and shrinkage in clause 3.1.4 and Annex B of EN 1992-1-1. Assuming that
t 10 years, for example, gives only about 90% of the long-term shrinkage strain and
creep coecient. It may be necessary to include part of the long-term shrinkage eects in
the rst check, because up to half of the long-term shrinkage can occur in the rst three
months after the end of curing of the concrete.
Clause 7.2.1(4) refers to the primary eects of shrinkage. These are calculated for
uncracked cross-sections (Example 5.3). After cracking, these eects remain in the concrete
between cracks, but have negligible inuence on stresses at the cracked cross-sections, at
which stresses are veried.
Clauses 7.2.1(6) and (7) refer to tension stiening. At a cross-section analysed as cracked,
its eect is to increase the tensile stress in the reinforcement, as discussed under clause 7.4.3.
It has negligible eect on the stress in the steel ange adjacent to the slab, and slightly reduces
the compressive stress in the other steel ange.
Clause 7.2.1(8) refers to the eects of local actions on the concrete slab, presumably a deck
slab. In highway bridges, these eects are mainly the sagging and hogging moments caused
by a single wheel, a pair of wheels, or a four-wheel tandem system, whichever is the most
adverse. In Load Model 1 these are combined with the eects of distributed loading and
the global eects in the plane of the slab. This combination is more adverse where the
slab spans longitudinally between cross-beams than for transverse spanning. Longitudinal
spanning can also occur at intermediate supports at the face of diaphragms. In combining
the stress ranges, it is important to consider the actual transverse location being
checked within the slab. The peak local eect usually occurs some distance from the
web of a main beam, while the global direct stress reduces away from the web due to
shear lag. The global stress distribution allowing for shear lag may be determined using
clause 5.4.1.2(8), even though this refers to EN 1993-1-5 which is for steel anges.
For serviceability stress limits, clause 7.2.2 refers to EN 1992 and EN 1993. Both codes
allow choice in the National Annex. EN 1992 does so by means of coecients k
i
, whereas
EN 1993 permits national values for a partial factor
M.ser
. If any National Annex uses
other than the recommended value, 1.0, this could be a source of error in practice,
because partial factors for serviceability checks are almost invariably 1.0, and so tend to
be forgotten.
Combinations of actions for serviceability checks
Clause 7.2 denes these combinations only by cross-reference, so the following summary is
given. The limiting stresses can be altered by National Annexes.
Clause 7.2.2(2) leads to the following recommendations for concrete.
.
Where creep coecients are based on clause 3.1.4(2) of EN1992-1-1, as is usual, its clause
7.2(3) gives the stress limit for the quasi-permanent combination as 0.45 f
ck
.
.
In areas where the exposure class is XD, XF or XS, clause 7.2(102) of EN1992-2 gives the
stress limit for the characteristic combination as 0.60 f
ck
. Comment on clause 4.1(1)
refers to the XD class in bridges.
From clause 7.2.2(4), the recommended limit for reinforcement is 0.8 f
sk
under the
characteristic combination, increased to 1.0 f
sk
for imposed deformations.
Clause 7.2.2(5) refers to clause 7.3 of EN1993-2, where the stress limits for structural steel
and the force limits for bolts are based on the characteristic combination, with a limit on
stress range under the frequent combination.
For service loading of shear connectors generally, clause 7.2.2(6) refers to clause 6.8.1(3).
That clause refers only to stud connectors under the characteristic combination, and uses a
Clause 7.2.1(1)P
Clause 7.2.1(4)
Clause 7.2.1(6)
Clause 7.2.1(7)
Clause 7.2.1(8)
Clause 7.2.2
Clause 7.2.2(2)
Clause 7.2.2(4)
Clause 7.2.2(5)
Clause 7.2.2(6)
165
factor k that can be chosen nationally. EN 1994-2 envisages the use of other types of
connector (for example, in clause 6.6.1.1(6)P). Rules for the use of these, which may be
given in a National Annex, from clause 1.1.3(3), should include a service load limit.
To sum up, most stress checks are based on characteristic combinations, as are the deter-
mination of cracked regions, clause 5.4.2.3(2), and the provision of minimumreinforcement,
clause 7.4.2(5). However, limiting crack widths are given, in clause 7.3.1(105) of EN 1992-2,
for the quasi-permanent combination.
Web breathing
Clause 7.2.3(1) refers to EN1993-2 for breathing of slender steel web plates. The eect on a
slender plate of in-plane shear or compressive stress is to magnify its initial out-of-plane
imperfection. This induces cyclic bending moments at its welded edges about axes parallel
to the welds. If excessive, it can lead to fatigue failure in these regions. Further comment
is given in the Guide to EN 1993-2.
4
7.3. Deformations in bridges
7.3.1. Deections
Clause 7.3.1(1) refers to clauses in EN 1993-2 that cover clearances, visual impression,
precambering, slip at connections, performance criteria and drainage. For precambering,
the eects of shear deformation . . . should be considered. This applies to vertical shear in
steel webs, not to the shear connection.
Clause 7.3.1(2) refers to Section 5 for calculation of deections. Rules for the eects of slip
are given in clause 5.4.1.1. They permit deformations caused by slip of shear connection to be
neglected, except in non-linear analysis. Clause 5.4.2.1(1) refers to the sequence of
construction, which aects deections. When the sequence is unknown, an estimate on the
high side can be obtained by assuming unpropped construction and that the adverse areas
of the inuence line, with respect to deection at the point being considered, are concreted
rst, followed by the relieving areas. Sucient accuracy should usually be obtained by
assuming that the whole of the concrete deck is cast at one time, on unpropped steelwork.
The casting of an area of deck slab may increase the curvature of adjacent beams where the
shear connectors are surrounded by concrete that is too young for full composite action to
occur. It is possible that subsequent performance of these connectors could be impaired by
what is, in eect, an imposed slip. Clause 7.3.1(3) refers to this, but not to the detailed
guidance given in clause 6.6.5.2(3), which follows.
Wherever possible, deformation should not be imposed on a shear connection until the concrete has
reached a cylinder strength of at least 20 N/mm
2
.
The words Wherever possible are necessary because shrinkage eects apply force to shear
connection from a very early age without, so far as is known, any adverse eect.
7.3.2. Vibrations
The limit state of vibration is covered in clause 7.3.2(1) by reference to other Eurocodes.
Composite bridges are referred to only in clause 6.4.6.3.1(3) of EN1991-2, which covers reso-
nance under railway loading. This gives lower bound values for damping that are the same
for composite bridges as for steel bridges, except that those for ller-beam decks are much
higher, and the same as for concrete bridges. Alternative values may be given in the National
Annex. The specialized literature generally gives damping values for composite oor or deck
systems that are between those for steel and for concrete members, as would be expected. In
railway bridges, the presence or absence of ballast is a relevant factor.
The reference to EN 1993-2 requires consideration of pedestrian discomfort and fatigue
under wind-induced motion, usually vortex shedding. The relevant reference is then to
EN 1991-1-4.
103
Its Annex E provides guidance on the calculation of amplitudes of oscilla-
tion while its Annex F provides guidance on the determination of natural frequencies and
Clause 7.2.3(1)
Clause 7.3.1(1)
Clause 7.3.1(2)
Clause 7.3.1(3)
Clause 7.3.2(1)
166
damping. The damping values for steel-composite bridges in its Table F.2 this time do lie
between the values for steel bridges and reinforced concrete bridges.
7.4. Cracking of concrete
7.4.1. General
In the early 1980s it was found
64.104
that for composite beams in hogging bending, the long-
established British methods for control of crack width were unreliable for initial cracks,
which were wider than predicted. Before this, it had been found for reinforced concrete
that the appropriate theoretical model for cracking caused by restraint of imposed defor-
mation was dierent from that for cracking caused by applied loading. This has led to the
presentation of design rules for control of cracking as two distinct procedures:
.
for minimum reinforcement, in clause 7.4.2, for all cross-sections that could be subjected
to signicant tension by imposed deformations (e.g. by eects of shrinkage, which
cause higher stresses than in reinforced concrete, because of restraint from the steel
beam)
.
for reinforcement to control cracking due to direct loading, clause 7.4.3.
The rules given in EN 1994-2 are based on an extensive and quite complex theory,
supported by testing on hogging regions of composite beams.
104.105
Much of the original
literature is either in German or not widely available
106
so a detailed account of the
theory has been published in English,
107
with comparisons with results of tests on composite
beams, additional to those used originally. The paper includes derivations of the equations
given in clause 7.4, comments on their scope and underlying assumptions, and procedures for
estimating the mean width and spacing of cracks. These are tedious, and so are not in
EN 1994-2. Its methods are simple: Tables 7.1 and 7.2 give maximum diameters and spacings
of reinforcing bars for three design crack widths: 0.2, 0.3 and 0.4 mm.
These tables are for high-bond bars only. This means ribbed bars with properties referred
to in clause 3.2.2(2)P of EN 1992-1-1. The use of reinforcement other than ribbed is outside
the scope of the Eurocodes.
The references to EN 1992 in clause 7.4.1(1) give the surface crack-width limits required
for design. Typical exposure classes for composite bridge decks are discussed in section 4.1 of
this guide.
Clause 7.4.1(2) refers to estimation of crack width, using EN 1992-1-1. This rather long
procedure is rarely needed, and does not take full account of the following dierences
between the behaviours of composite beams and reinforced concrete T-beams. The steel
member in a composite beam does not shrink or creep and has much greater exural stiness
than the reinforcement in a concrete beam. Also, the steel member is attached to the concrete
ange only by discrete connectors that are not eective until there is longitudinal slip,
whereas in reinforced concrete there is monolithic connection. There is no need here for a
reference to EN 1992-2.
Clause 7.4.1(3) refers to the methods developed for composite members, which are easier
to apply than the methods for reinforced concrete members.
Clause 7.4.1(4) refers to limiting calculated crack widths w
k
, with a Note on recommended
values. Those for all XC, XDand XS exposure classes are given in a Note to clause 7.3.1(105)
of EN 1992-2 as 0.3 mm. This is for the quasi-permanent load combination, and excludes
prestressed members with bonded tendons. Both the crack width and the load combination
may be changed in the National Annex. It is expected that the UKs National Annex to
EN 1992-2 will conrm these recommendations and give further guidance for combinations
that include temperature dierence.
Clause 7.4.1(5) and (6) draws attention to the need to control cracking caused by
early thermal shrinkage. The problem is that the heat of hydration causes expansion of
the concrete before it is sti enough for restraint from steel to cause much compressive
stress in it. When it cools, it is stier, so tension develops. This can occur in regions that
Clause 7.4.1(1)
Clause 7.4.1(2)
Clause 7.4.1(3)
Clause 7.4.1(4)
Clause 7.4.1(5)
Clause 7.4.1(6)
167
are in permanent compression in the nished bridge. They may require crack-control
reinforcement for this phase only.
The check is made assuming that the temperatures of the steel and the concrete are both
uniform. The concrete is colder, to an extent that may be given in the National Annex. This
causes tension, and possibly cracking. Further comment is given in Example 7.1.
7.4.2. Minimum reinforcement
The only data needed when using Tables 7.1 and 7.2 are the design crack width and the tensile
stress in the reinforcement,
s
. For minimum reinforcement,
s
is the stress immediately
after initial cracking. It is assumed that cracking does not change the curvature of the steel
beam, soall of the tensile force inthe concrete just before cracking is transferredtothe reinforce-
ment, of area A
s
. If the slab were in uniformtension, equation (7.1) in clause 7.4.2(1) would be:
A
s
s
A
ct
f
ct;eff
where f
ct;eff
is an estimate of the mean tensile strength of the concrete at the time of cracking.
The three correction factors in equation (7.1) are based on calibration work.
106
These
allow for the non-uniform stress distribution in the area A
ct
of concrete assumed to crack.
Non-uniform self-equilibrating stresses arise from primary shrinkage and temperature
eects, which cause curvature of the composite member. Slip of the shear connection also
causes curvature and reduces the tensile force in the slab.
The magnitude of these eects depends on the geometry of the uncracked composite
section, as given by equation (7.2). With experience, calculation of k
c
can often be
omitted, because it is less than 1.0 only where z
0
< 1:2h
c
. (These symbols are shown in
Fig. 7.5.) The depth of the uncracked neutral axis below the bottom of the slab normally
exceeds about 70% of the slab thickness, and then, k
c
1.
The method of clause 7.4.2(1) is not intended for the control of early thermal cracking,
which can occur in concrete a few days old, if the temperature rise caused by heat of
hydration is excessive. The anges of composite beams are usually too thin for this to
occur. It would not be correct, therefore, to assume a very low value for f
ct;eff
.
The suggested value of f
ct;eff
, 3 N/mm
2
, was probably rounded from the mean 28-day
tensile strength of grade C30/37 concrete, given in EN 1992-1-1 as 2.9 N/mm
2
the value
used as the basis for the optional correction given in clause 7.4.2(2). The maximum bar
diameter may be increased for stronger concrete because the higher bond strength of the
concrete compensates for the lower total perimeter of a set of bars with given area per
unit width of slab. The dierence between 2.9 and 3.0 is obviously negligible. It may be an
error in drafting, because in EN 1992, the value 2.9 N/mm
2
is used in both places.
If there is good reason to assume a value for f
ct;eff
such that the correction is not negligible,
a suitable procedure is to assume a standard bar diameter, , calculate
, and then nd
s
by
interpolation in Table 7.1.
The reinforcement in a deck slab will usually be in two layers in each direction, with at least
half of it adjacent to the surface of greater tensile strain. The relevant rule, in clause 7.4.2(3),
refers not to the actual reinforcement, but to the minimum required. The reference to local
depth in clause 7.4.2(4) means the depth at the cross-section considered.
The rule of clause 7.4.2(5) on placing of minimum reinforcement refers to its horizontal
extent, not to its depth within the slab. Analysis of the structure for ultimate load combina-
tions of variable actions will normally nd regions in tension that are more extensive than
those for the characteristic combination specied here. The regions so found may need to
be extended for early thermal eects (clause 7.4.1(5)).
Design of minimum reinforcement for a concrete slab
For design, the design crack width and thickness of the slab, h
c
, will be known. For a chosen
bar diameter , Table 7.1 gives
s
, the maximum permitted stress in the reinforcement, and
equation (7.1) allows the bar spacing to be determined. If this is too high or low, is
changed.
Clause 7.4.2(1)
Clause 7.4.2(2)
Clause 7.4.2(3)
Clause 7.4.2(4)
Clause 7.4.2(5)
168
Atypical relationship between slab thickness h
c
, bar spacing s and bar diameter c is shown
in Fig. 7.1. It is for two similar layers of bars, with k
c
1 and f
ct.eff
3.0 N/mm
2
. Equation
(7.1) then gives, for a fully cracked slab of breadth b:
(c
2
,42b,s 0.72 3bh
c
,o
s
Hence,
h
c
s 0.727c
2
o
s
(with o
s
in N/mm
2
units) (D7.1)
For each bar diameter and a given crack width, Table 7.1 gives c
2
o
s
, so the product h
c
s is
known. This is plotted in Fig. 7.1, for w
k
0.3 mm, as curves of bar spacing for four
given slab thicknesses, which can of course also be read as maximum slab thickness size
and for bar spacing. The shape of the curves results partly from the use of rounded values
of o
s
in Table 7.1. The correction to minimum reinforcement given in clause 7.4.2(2) is
negligible here, and has not been made.
The weight of minimumreinforcement, per unit area of slab, is proportional to c
2
,s, which
is proportional to o
1
s
, from equation (D7.1). The value of o
1
s
increases with bar diameter,
fromTable 7.1, so the use of smaller bars reduces the weight of minimumreinforcement. This
is because their greater surface area provides more bond strength.
7.4.3. Control of cracking due to direct loading
Clause 7.4.3(2) species elastic global analysis to Section 5, allowing for the eects of
cracking. The preceding comments on global analysis for deformations apply also to this
analysis for bending moments in regions with concrete in tension.
From clause 7.4.1(4), the combination of actions will be given in the National Annex.
There is no need to reduce the extent of the cracked regions below that assumed for
global analysis for ultimate limit states, so the new bending moments for the composite
members can be found by scaling values found previously. At each cross-section, the area
of reinforcement will already be known: that required for ultimate loading or the specied
minimum, if greater; so the stresses o
s.o
, clause 7.4.3(3), can be found.
Tension stiening
A correction for tension stiening is now required. At one time, these eects were not well
understood. It was thought that, for a given tensile strain at the level of the reinforcement,
the total extension must be the extension of the concrete plus the width of the cracks, so that
allowing for the former reduced the latter. The true behaviour is more complex.
The upper part of Fig. 7.2 shows a single crack in a concrete member with a central
reinforcing bar. At the crack, the external tensile force N causes strain
s2
N,A
s
E
a
in
the bar, and the strain in the concrete is the free shrinkage strain
cs
, which is shown as
Clause 7.4.3(2)
Clause 7.4.3(3)
100
200
300
400
0 5 6 8 10 12 16 20
h
c
= 100 mm
150
200
300
(mm)
s (mm)
Fig. 7.1. Bar diameter and spacing for minimum reinforcement in two equal layers, for w
k
0.3 mm and
f
ct.eff
3.0 N/mm
2
169
negative here. There is a transmission length L
e
each side of the crack, within which there is
transfer of shear between the bar and the concrete. Outside this length, the strain in both the
steel and the concrete is
s1
, and the stress in the concrete is fractionally below its tensile
strength. Within the length 2L
e
, the curves
s
(x) and
c
(x) give the strains in the two
materials, with mean strains
sm
in the bar and
cm
in the concrete.
It is now supposed that the graph represents the typical behaviour of a reinforcing bar in a
cracked concrete ange of a composite beam, in a region of constant bending moment such
that the crack spacing is 2L
e
. The curvature of the steel beam is determined by the mean
stiness of the slab, not the fully cracked stiness, and is compatible with the mean longitu-
dinal strain in the reinforcement,
sm
.
Midway between the cracks, the strain is the cracking strain of the concrete, corresponding
to a stress less than 30N/mm
2
in the bar. Its peak strain, at the crack, is much greater than
sm
,
but less than the yield strain of the reinforcement, if crack widths are not to exceed 0.3 mm. The
crack width corresponds to this higher strain, not to the strain
sm
that is compatible with the
curvature, so a correction to the strain is needed. It is presented in clause 7.4.3(3) as a correc-
tion to the stress o
s.o
because that is easily calculated, and Tables 7.1 and 7.2 are based on
stress. The strain correction cannot be shown in Fig. 7.2 because the stress o
s.o
is calculated
using the fully cracked stiness, and so relates to a curvature greater than the true curvature.
The derivation of the correction
107
takes account of crack spacings less than 2L
e
, the bond
properties of reinforcement, and other factors omitted from this simplied outline.
The section properties needed for the calculation of the correction o
s
will usually be
known. For the cracked composite cross-section, the transformed area A is needed to nd
I, which is used in calculating o
s.o
, and A
a
and I
a
are standard properties of the steel
section. The result is independent of the modular ratio. For simplicity, c
st
may conserva-
tively be taken as 1.0, because AI A
a
I
a
.
When the stress o
s
at a crack has been found, the maximum bar diameter or the maximum
spacing are found from Tables 7.1 and 7.2. Only one of these is needed, as the known area of
reinforcement then gives the other. The correction of clause 7.4.2(2) does not apply.
General comments on clause 7.4, and ow charts
The design actions for checking cracking will always be less than those for the ultimate limit
state due to the use of lower load factors. The dierence is greatest where unpropped con-
struction is used for a continuous beam with hogging regions in Class 1 or 2 and with
lateraltorsional buckling prevented. This is because the entire design hogging moment is
carried by the composite section for Class 1 and 2 composite sections at ULS, but at SLS,
reinforcement stresses are derived only from actions applied to the composite section in
the construction sequence. It is also permissible in such cases to neglect the eects of indirect
actions at ULS. The quantity of reinforcement provided for resistance to load should be
L
e
L
e
N
Tensile strain
0
x
N
c
(x)
s
(x)
s2
sm
cm
cs
s1
Fig. 7.2. Strain distributions near a crack in a reinforced concrete tension member
170
sucient to control cracking. The main use of clause 7.4.3 is then to check that the spacing of
the bars is not excessive.
Where propped construction is used, the disparity between the design loadings for the
two limit states is smaller. A check to clause 7.4.3 is then more likely to inuence the
reinforcement required.
Flow charts for crack-width control
The check to clause 7.4.3 is likely to be done rst, so its ow chart, Fig. 7.3, precedes Fig. 7.4
for minimum reinforcement, to clause 7.4.2. The regions where the slab is in tension depend
on the load combination, and three may be relevant, as follows.
.
Most reinforcement areas are found initially for the ultimate combination.
.
Load-induced cracking is checked for a combination to be specied in the National
Annex, to clause 7.4.1(4). It may be the quasi-permanent or frequent combination.
.
Minimum reinforcement is required in regions in tension under the characteristic
combination, clause 7.4.2(5).
Recommended only for cross-sections
with longitudinal prestress by tendons.
Outside the scope of this chart (END)
Yes
No
No
Reduce and s
s
Increase s
s
Go to flow chart for minimum reinforcement (Fig. 7.4)
See the Notes in the section Flow charts for crack-width control
Exposure classes. For each concrete surface in tension, find the exposure
class to clause 4.2 of EN 1992-1-1 and EN 1992-2 (referred to from 7.4.1(1))
Crack widths. Find the limiting crack widths w
k
and the combination of actions
for verification from the National Annex (from the Note to 7.4.1(4))
Do global analyses for this combination to find bending
moments in regions where the slab is in tension
For areas of reinforcement found previously, A
s
, determine tensile stresses
s,o
in
bars adjacent to surfaces to be checked, neglecting primary shrinkage, to 7.2.1(4).
Calculate tensile stresses
s
for tension stiffening, from 7.4.3(3). Determine the
tensile stress due to coexisting local actions,
s,loc
, and find
s,E
=
s,o
+
s
+
s,loc
, from 7.2.1(8)
Do you want to find crack widths
from EN 1992-1-1, 7.3.4?
(7.4.1(2))
From 7.4.1(3), use w
k
and
s,E
to find either max. bar spacing s
s
fromTable 7.2 and
calculate bar diameter from A
s
or (less convenient) find diameter * fromTable 7.1,
then from 7.4.2(2), and find bar spacing from A
s
Do you want to change or s
s
?
Only possible by increasing A
s
and
hence reducing
s,E
. This may permit a
small increase in . It is inefficient and
not recommended
Reduce , at constant A
s
.
Effect is to reduce bar spacing
Fig. 7.3. Flow chart for control of cracking due to direct loading
171
The following notes apply to these charts.
Note 1. Creep and shrinkage of concrete both increase stress in reinforcement at internal
supports of beams, so crack widths are usually veried using the long-term
modular ratio for permanent actions. This is assumed here.
Note 2. The ow charts apply to a tension ange of a continuous longitudinal beam. It is
assumed (for brevity) that:
.
areas of reinforcement required for ultimate limit states have been found
.
the exural stinesses E
a
I
1
are known for the uncracked cross-sections, using
relevant modular ratios (inclusion of reinforcement optional)
.
the cracked exural stinesses E
a
I
2
are known. For these, modular ratios are
usually irrelevant, unless double composite action is being used
.
the deck slab is above the steel beam, and at cross-sections considered,
maximum tensile strain occurs at the top surface of the slab
.
there is no double composite action
.
early thermal cracking does not govern
.
the symbols A
s
(area of reinforcement) are for a unit width of slab.
Appropriate for
sections prestressed
by tendons, 7.4.2(1).
Outside the scope of
this flow chart (END)
No Yes
The chosen bar size and spacing are satisfactory as minimum reinforcement,
but may not be sufficient to control cracking due to direct loading at the section
considered (END)
Calculate * = (2.9/f
ct,eff
)
No
Yes
Minimum reinforcement. Find all regions where concrete can be in flexural tension
under the characteristic combination of permanent and variable actions, 7.4.2(5),
taking account of shrinkage of concrete and effects of temperature and settlement,
if any, 7.4.2(1). These regions may be more extensive than those where crack control
for effects of loading is required
For each region, propose details of minimum longitudinal reinforcement: bar size
and spacing s
s
; usually in two layers, with at least half near the surface with the
greater tensile strain (7.4.2(3)). Figure 7.1 is useful where f
ct,eff
3 N/mm
2
For the design crack width w
k
, find
s,max
for * from Table 7.1, using
interpolation if necessary. (This route is used because must be a
standard bar diameter, but * need not be)
Do you wish to determine minimum reinforcement by a more accurate
method, following clause 7.3.2(1)P of EN 1992-1-1 (from 7.4.2(1))?
Calculate k
c
from eq. (7.2). Consider whether the
recommended values for k and k
s
are
appropriate, and change them if not (unlikely).
For unit width of the tensile zone considered,
find A
ct
. Find A
s,min
from eq. (7.1) using
s,max
For the proposed reinforcement,
find A
s
per unit width of slab.
Is A
s
A
s,min
?
Reduce s
s
or increase
(or combination of both)
Choose f
ct,eff
, defined in 7.4.2(1).
Use N/mm
2
units
Fig. 7.4. Flow chart for minimum reinforcement for control of cracking
172
Note 3. A second subscript, l for loading, is used in this note to indicate quantities found in
the check to clause 7.4.3. Area A
s.l
is usually that required for resistance to ultimate
loads. It is assumed, for simplicity, that minimum reinforcement consists of
two identical layers of bars, one near each surface of the slab. Area A
s.l
should
be compared with the minimum reinforcement area A
s
required when bars of
diameter c
l
are used. If A
s.l
< A
s
, minimum reinforcement governs.
7.5. Filler beam decks
This clause is applicable to simply-supported or continuous decks of the type shown in
Fig. 6.8, spanning longitudinally. The Note to clause 6.3.1(1) permits the use of transverse
ller beams according to the National Annex, which should refer also to serviceability
requirements, if any.
From clause 7.5.1(1), the methods of global analysis for serviceability limit states are the
same as for ultimate limit states, clause 5.4.2.9, except that no redistribution of moments is
permitted.
The word considered is used in clause 7.5.2(1) because some of the clauses may not be
applicable. For example, the thickness of the concrete in a ller-beam deck will exceed
that in a conventional composite beam, so the eects of heat of hydration will be greater.
The temperature dierence recommended in the Note to clause 7.4.1(6) may not be
appropriate. Tension stiening is also dierent (clause 7.5.4(2)).
The objective of clause 7.5.3(1) is to ensure that cracking of concrete does not cause its
reinforcement to yield. It applies above an internal support of a continuous ller-beam
deck, and can be illustrated as follows.
It is assumed that all the concrete above the top anges of the steel beams reaches its mean
tensile strength, f
ctm
, and then cracks. This releases a tensile force of A
c.eff
f
ctm
per beam,
where A
c.eff
s
w
c
st.
as in clause 7.5.3(1). The notation is shown in Fig. 6.8.
The required condition, with partial factors
M
taken as 1.0, is:
A
s.min
f
sk
! A
c.eff
f
ctm
(D7.2)
For concretes permitted by clause 3.1(2), f
ctm
4.4 N/mm
2
.
If, to satisfy expression (7.7), A
s.min
> 0.01A
c.eff
, then expression (D7.2) is satised if
f
sk
440 N/mm
2
. In design to EN 1994, normally f
sk
500 N/mm
2
, so the objective is met.
Clause 7.5.4(1) applies to longitudinal bottom reinforcement in mid-span regions. Widths
of cracks in the concrete sot between the steel beams should be controlled unless the
formwork used (shown in Fig. 6.8) provides permanent protection. Otherwise, wide
transverse cracks could form under the bottom transverse bars specied in clause 6.3.1(4),
putting them at risk of corrosion.
Clause 7.5.4(2) says, in eect, that the stress in the reinforcement may be taken as o
s.o
,
dened in clause 7.4.3(3). Thus, there is no need to consider tension stiening in ller-
beam decks.
Clause 7.5.1(1)
Clause 7.5.2(1)
Clause 7.5.3(1)
Clause 7.5.4(1)
Clause 7.5.4(2)
Example 7.1: checks on serviceability stresses, and control of cracking
The following serviceability checks are performed for the internal girders of the bridge in
Figs 5.6 and 6.22:
.
tensile stress in reinforcement and crack control at an intermediate support
.
minimum longitudinal reinforcement for regions in hogging bending
.
stresses in structural steel at an intermediate support
.
compressive stress in concrete at mid-span of the main span.
Tensile stress in reinforcement
The cross-section at an intermediate support is shown in Fig. 7.5. All reinforcement has
f
sk
500 N/mm
2
.
173
1225
400 40
400 25
1160 25
26
Including reinforcement
Elastic neutral axes
for n
0
= 6.36
60
70
25
3100
1120
Excluding reinforcement
20 mm bars at 75 mm 20 mm bars at 150 mm
h
c
= 250
z
0
125
Fig. 7.5. Elastic neutral axes of uncracked cross-section at internal support
From clause 7.2.1(8), stresses in reinforcement caused by simultaneous global and local
actions should be added. Load Model 1 is considered here. The tandem system (TS) and
UDL produce a smaller local eect than does Load Model 2, but Load Model 1 produces
the greatest combined local plus global eect. Local moments are caused here by hogging of
the deck slab over the pier diaphragm. Full local eects do not coexist with full global eects
since, for maximum global hogging moment, the axles are further from the pier. The
maximum local and global eects are here calculated independently and then combined.
Aless conservative approach, if permitted by the National Annex, is to use the combination
rule in Annex E of EN 1993-2 as noted below clause 5.4.4(1). This enables the maximum
global eect to be combined with a reduced local eect and vice versa.
The maximum global hogging moment acting on the composite section under the
characteristic combination was found to be 4400 kNm. Clause 7.2.1(4) allows the
primary eects of shrinkage to be ignored for cracked sections, but not the secondary
eects which are unfavourable here.
From Table 6.2 in Example 6.6, the section modulus for the top layer of reinforcement
in the cracked composite section is 34.05 10
6
mm
3
. The stress due to global eects in this
reinforcement, ignoring tension stiening, is:
o
s.o
4400,34.05 129 N/mm
2
From clause 7.2.1(6), the calculation of reinforcement stress should include the eects
of tension stiening. From clause 7.4.3(3), the increase in stress in the reinforcement, due
to tension stiening, from that calculated using a fully cracked analysis is:
o
s

0.4 f
ctm
c
st
,
s
where c
st

AI
A
a
I
a
74 478 2.266 10
10
55 000 1.228 10
10
2.50
The reinforcement ratio ,
s
0.025 and f
ctm
2.9 N/mm
2
and thus:
o
s

0.4 2.9
2.50 0.025
19 N,mm
2
The stress in the reinforcement from global eects is therefore:
129 19 148 N/mm
2
The stress in the reinforcement due to local load was found by determining the local
moment from an analysis using the appropriate Pucher chart.
100
This bending moment,
12.2 kNm/m hogging, is not sucient to cause compression in the slab, so it is resisted
solely by the two layers of reinforcement, and causes a tensile stress of 73 N/mm
2
in the
top layer. The total stress from global plus local eects is therefore:
o
s
148 73 221 N/mm
2
174
From clause 7.2.2(4) and from EN 1992-1-1 clause 7.2(5), the tensile stress in the
reinforcement should not exceed k
3
f
sk
400 N/mm
2
(recommended k
3
0.8, and
f
sk
500 N/mm
2
) so this check is satised.
Where the global tension is such that local eects cause net compression at one face of
the slab, iterative calculation can be avoided by taking the lever arm for local bending as
the lesser of the distance between the two layers of reinforcement and that derived from
considering the cracked section in exure alone.
Control of cracking due to direct loading persistent design situation
Creep of concrete in sagging regions causes hogging moments to increase, so this check is
done at t !1. Cracking during construction is considered later.
For concrete protected by waterproong, EN 1992-2 clause 4.2(105) recommends an
exposure class of XC3 for which Table 7.101N of EN 1992-2 recommends a limiting
crack width of 0.3 mm under the quasi-permanent load combination.
Crack widths are checked under this load combination for which the maximumhogging
moment acting on the composite section was found to be 2500 kNm. The secondary eects
of shrinkage were unfavourable here. The primary eects can be ignored because the deri-
vation of the simplied method (Tables 7.1 and 7.2) takes account of a typical amount of
shrinkage.
107
Using a section modulus from Table 6.2, the stress due to global eects in the top layer
of reinforcement, ignoring tension stiening, is:
o
s.o
2500,34.05 73 N/mm
2
The increase in stress in the reinforcement due to tension stiening is 19 N/mm
2
as
found above. From equation (7.4), the reinforcement stress is:
o
s
73 19 92 N/mm
2
For this stress, Table 7.1 gives a bar diameter above 32 mm, and Table 7.2 gives a bar
spacing exceeding 300 mm, so there is sucient crack control from the reinforcement pro-
vided, 20 mm bars at 75 mm spacing.
Control of cracking during construction
Construction is unpropped, so clause 7.4.1(5) is applicable. Eects of heat of hydration of
cement should be analysed to dene areas where tension is expected . These areas require
at least minimum reinforcement, from clause 7.4.2. The curvatures for determining
secondary eects should be based initially on uncracked cross-sections.
Clause 7.4.1(6) requires account to be taken of the eects of heat of hydration for
limitation of crack width unless specic measures are taken to limit its eects. The
National Annex may refer to specic measures.
Assuming that no specic measures have been taken, clause 7.4.1(6) applies. Its Note
recommends that the concrete slab should be assumed to be 20 K colder than the steel
member, and that the short-term modular ratio n
0
should be used. It is expected that
the National Annex for the UK will replace 20 K by 25 K, the value used here.
Referring to Fig. 5.6, it is assumed that the whole of the deck slab has been concreted
except for lengths such as CD, which extend for 6 m each side of each internal support.
These two 12 m lengths are now cast, over the whole width of the deck. When they
harden, it is assumed that the only longitudinal stress in their concrete arises from
shrinkage and the temperature dierence. The primary eects are uniform over the
whole 12 m length. The secondary eects are a hogging bending moment over the
whole length of the bridge, which is a maximum at the internal supports, and associated
shear forces.
It is reasonable to assume that the 12 m lengths are either cracked throughout, or
uncracked. In calculating secondary eects, primary eects due to shrinkage can be
neglected in cracked regions, from clause 5.4.2.2(8), so only eects of heat of hydration
175
need be considered. A similar conclusion is reached if the method of clause 5.4.2.3(3),
15% of the span cracked, is used. In any case, shrinkage is a minor eect, as shown
below.
FromExample 5.3, the autogenous and long-termshrinkage strains are
ca
50 10
6
and
cd
282 10
6
respectively. From clause 3.1.4(6) of EN 1992-1-1, drying shrinkage
may be assumed to commence at the end of curing.
Autogenous shrinkage is a function of the age of the concrete, t in days:
ca
t 1 exp0.2t
0.5
ca
1
Here, the temperature dierence is assumed to reach its peak at age seven days, and curing
is conservatively assumed to have ended at age three days. From EN 1992-1-1, less than
1% of the long-term drying shrinkage will have occurred in the next four days, so it is
neglected.
From the equation above,
ca
7 1 exp0.53 50 10
6
21 10
6
For temperature, clause 3.1.3(5) of EN 1992-1-1 gives the thermal coecient for
concrete as 10 10
6
, so a dierence of 25 K causes a strain of 250 10
6
.
The resulting sagging curvature of each 12 m length of deck is found by analysis of the
uncracked composite section with n
0
6.36 from Example 5.2. To nd whether cracking
occurs, the secondary eects are found from these curvatures by global analysis using
uncracked stinesses.
The denition of regions in tension is based on the characteristic combination, from
clause 5.4.2.3(2). It is suggested here that heat of hydration should be treated as a
permanent action, as shrinkage is. The leading variable action is likely to be construction
load elsewhere on the deck, in association with either temperature or wind. The
permanent actions are dead load and any drying shrinkage in the rest of the deck. The
action eects so found from uncracked analysis are added to the secondary eects,
above, to determine the regions in tension, in which at least minimum reinforcement
(to clause 7.4.2) should be provided. If tensile stresses in concrete are such that it
cracks, reanalysis using cracked sections and ignoring primary eects gives the tensile
stresses in reinforcement needed for crack-width control. Minimum reinforcement will
not be sucient at the internal supports, where reinforcement will have been designed
both for resistance to ultimate direct loading and for long-term crack-width control.
The seven-day situation studied here is unlikely to be more adverse.
Minimum reinforcement
The minimum reinforcement required by clause 7.4.2 is usually far less than that required
at an internal support. The rules apply to any region subjected to signicant tension and
can govern where the main longitudinal reinforcement is curtailed, or near a point of
contraexure.
Figure 7.1 gives the minimum reinforcement for this 250 mm slab as 20 mm bars at
260 mm spacing, top and bottom. The top layer has a cross-section only 29% of that of
the top layer provided at the pier. This result is now checked. From clause 7.4.2(1):
A
s.min
k
s
k
c
k f
ct.eff
A
ct
,o
s
(7.1)
with k
s
0.9 and k 0.8.
Assuming that the age of the concrete at cracking is likely to exceed 28 days,
f
ct.eff
3.0 N/mm
2
The only term in equation (7.1) that depends on the steel cross-section is k
c
, given by:
k
c

1
1 h
c
, 2z
0

0.3 1.0 7.2
176
where h
c
is the ange thickness (250 mm) and z
0
is the distance between the centres of
area of the uncracked composite section and the uncracked concrete ange, with
modular ratio n
0
(6.36 here). It makes little dierence whether reinforcement is included
or not. Excluding it (for simplicity) and ignoring the haunch, then from Fig. 7.5:
z
0
1500 1120 125 255 mm
and
k
c
1,1 250,510 0.3 0.97
If the area of longitudinal reinforcement at the internal support is included, k
c
is reduced
by only 0.02.
Ignoring the dierence between 2.9 and 3.0 N/mm
2
, Table 7.1 gives the maximum stress
for 20 mm bars as 220 N/mm
2
for a limiting crack width of 0.3 mm.
For a 1 m width of ange, from equation (7.1):
A
s.min
0.9 0.97 0.8 3.0 1000 250,220 2380 mm
2
/m
Clause 7.4.2(3) requires at least half of this reinforcement to be placed in the upper
layer. This gives 20 mm bars at 264 mm spacing; probably 250 mm in practice. A similar
calculation for 16 mm bars gives their spacing as 184 mm.
Stress limits for structural steel
The hogging moments under the characteristic load combination are found to be
1600 kNm acting on the steel beam and 4400 kNm acting on the composite section. The
relevant section moduli are given in Table 6.2 in Example 6.6. The stress in the bottom
ange was found to be critical, for which:
o
Ed.ser

1600
22.20
4400
29.25
222 N,mm
2
The coexisting shear force acting on the section was found to be 1900 kN, so:
t
Ed.ser

1900 10
3
25 1160
66 N,mm
2
if buckling is neglected at SLS.
From clause 7.2.2(5) and EC3-2 clause 7.3, stresses should be limited as follows:
o
Ed.ser
222
f
y
M.ser
345 N,mm
2
t
Ed.ser
66
f
y
3
p

M.ser
199 N,mm
2
o
2
Ed.ser
3t
2
Ed.ser
q
250
f
y
M.ser
345 N,mm
2
where
M.ser
1.00. The stresses in structural steel are all satisfactory.
Compressive stress in concrete
The cross-section at mid-span is shown in Fig. 6.2. The compressive stress in the slab is
checked under simultaneous global and local actions. Both the primary and secondary
eects of shrinkage are favourable and are conservatively ignored. The concrete is checked
shortly after opening the bridge, ignoring creep, as this produces the greatest concrete
compressive stress here, so the short-term modular ratio is used. The sagging moment on
the composite section from the characteristic combination was found to be 3500kNm.
The section modulus for the top of the concrete slab is 575.9 10
6
mm
3
from Table 6.4
in Example 6.10. The compressive stress at the top of the slab is:
o
c
3500,575.9 6.1 N/mm
2
177
The maximum longitudinal sagging moment from local loading was found to be
20 kNm/m. The short-term section modulus for the top surface of the uncracked
reinforced slab is W
c.top
10.7 10
6
mm
4
/m in concrete units, so the compressive
stress from local load is o
c.loc
20,10.7 1.9 N/mm
2
.
The total concrete stress, fully combining global and local eects, is:
o
c
6.1 1.9 8.0 N/mm
2
The global stress is sucient to keep the whole depth of the slab in compression so
this calculation of stresses from local moment using gross properties for the slab is
appropriate. If the local moment caused tension in the slab, this would no longer be
adequate. Either the stress from local moment could conservatively be calculated from
a cracked section analysis considering the local moment acting alone, or an iterative
calculation considering the local moment and axial force acting together would be
required. In this case, application of the combination rule of EN 1993-2 Annex E
would underestimate the combined eect as the global and local loading cases are the
same.
Fromclause 7.2.2(2) and from EN1992-1-1 clause 7.2 with the recommended value for
k
1
, concrete stresses should be limited to k
1
f
ck
0.6 30 18 N/mm
2
so the concrete
stress is acceptable.
178
CHAPTER 8
Precast concrete slabs in
composite bridges
This chapter corresponds to Section 8 in EN 1994-2, which has the following clauses:
.
General Clause 8.1
.
Actions Clause 8.2
.
Design, analysis and detailing of the bridge slab Clause 8.3
.
Interface between steel beam and concrete slab Clause 8.4
8.1. General
Clause 8.1(1) states the scope of Section 8: precast deck slabs of reinforced or prestressed
concrete which are either:
.
partial thickness, acting as permanent participating formwork to the in-situ concrete
topping, or
.
full thickness, where only a small quantity of concrete needs to be cast in situ to join the
precast units together. Figure 8.1 illustrates a typical deck of this type.
Precast slabs within the scope of Section 8 should be fully composite with the steel beam
clause 8.1(2). Non-participating permanent formwork is not covered, for it is dicult
both to prevent such formwork from being stressed by imposed loading, and to ensure its
durability.
Clause 8.1(1)
Clause 8.1(2)
Precast or in situ
edge beam
Precast or in situ
edge beam
Transverse joint
Pocket for shear
connectors
Precast deck slab
unit spanning
multiple girders
Projecting
reinforcement
not shown
Fig. 8.1. Typical full-thickness precast concrete deck
Clause 8.1(3) is a reminder that the designer should check the sensitivity of the detailing to
tolerances and specify stricter values than those required by EN 1992 (through EN 13670) if
necessary. Key issues to consider include:
.
detailing of the precast slabs at pockets to ensure that each pocket is correctly located
over the steel beam, that projecting transverse reinforcement will not clash with the
shear connection, and that there is sucient space for concreting (clause 8.4.3(2))
.
detailing of stitch reinforcement between adjacent precast slabs to ensure that bars do not
clash and to satisfy clause 8.3(1) on continuity
.
tolerances on overall geometry of each precast unit so that, where required, abutting
units are suciently parallel to each other to avoid the need for additional sealing
from underneath. The tolerances for steelwork are also important, and are referred to
in clause 8.4.1.
8.2. Actions
Clause 8.2(1) warns that the design of precast deck slabs should consider the actions
arising from the proposed construction method as well as the actions given in EN 1991-1-6.
30
8.3. Design, analysis and detailing of the bridge slab
Even where full-thickness slabs are used, some interaction with in-situ concrete occurs at
joints, so clause 8.3(1) is relevant to both types of precast concrete slab. Its requirement
for the deck to be designed as continuous in both directions applies to the nished structure.
It does not mean that the reinforcement in partial-thickness precast slabs or planks must be
continuous. That would exclude the use of Omnia-type planks, shown in Fig. 8.2. Precast
planks of this sort span simply-supported between adjacent steel beams and are joined with
in-situ concrete over the tops of the beams. The main reinforcement in the planks is not
continuous across these joints, but the reinforcement in the in-situ concrete is. In the other
direction, the planks abut as shown, so that only a small part of the thickness of the slab
is discontinuous in compression. Continuity of reinforcement is again achieved in the slab
but not in the planks. The resulting slab (part precast, part in situ) is continuous in both
directions.
EN 1992-1-1 clause 6.2.5 is relevant for the horizontal interface between the precast and
in situ concrete. Examples of bridges of this type are given in Ref. 108.
To allow precast slab units to be laid continuously across the steel beams, shear connection
usually needs to be concentrated in groups with appropriate positioning of pockets in the
precast slab as illustrated in Fig. 8.1. Clause 8.3(2) therefore refers to clause 6.6.5.5(4)
for the use of stud connectors in groups. Clause 8.3(3) makes reference to clause 6.6.1.2.
This allows some degree of averaging of the shear ow over a length, which facilitates
standardisation of the details of the shear connection and the pockets.
Clause 8.1(3)
Clause 8.2(1)
Clause 8.3(1)
Clause 8.3(2)
Clause 8.3(3)
Precast plank
In-situ concrete
Flange of steel beam
Fig. 8.2. Typical partial-thickness precast concrete planks
180
8.4. Interface between steel beam and concrete slab
Clause 8.4.1(1) refers to bedding, such as the placing of the slabs on a layer of mortar.
Sealing of the interface between steel beam and precast beam is needed both to protect the
steel ange from corrosion and to prevent leakage of grout when the pockets are concreted.
Where a precast unit is supported by more than two beams, bedding may also be needed to
ensure that load is shared between the beams as intended.
Bedding in clause 8.4.1(1) appears to mean a gap-lling material capable of transferring
vertical compression. Where it is intended not to use it, the clause requires special tolerances
to be specied for the steelwork to minimise the eects of uneven contact between slab and
steel ange.
This does not solve the problems of corrosion and grout leakage, for which a compressible
sealing strip could be applied to the edges of the ange and around the pocket. There would
then still be no direct protection of the top ange by in-situ concrete (other than at a pocket)
and so clause 8.4.2(1) requires that a top ange without bedding be given the same corrosion
protection as the rest of the beam, apart from the site-applied top coat.
If a non-loadbearing anti-corrosion bedding is provided, then the slab should be designed
for the transfer of vertical loads only at the positions of the pockets. It would be prudent also
to assume that clause 8.4.1(1) on special tolerances still applies.
Clause 8.4.3 gives provisions for the shear connection and transverse reinforcement, sup-
plementing Sections 6 and 7. Clause 8.4.3(2) emphasises the need for both suitable concrete
mix design and appropriate clearance between shear connectors and precast concrete, allow-
ing for tolerances, in order to enable in-situ concrete to be fully compacted. Clause 8.4.3(3)
highlights the need to detail reinforcement appropriately adjacent to groups of connectors.
This is discussed with the comments on clause 6.6.5.5(4).
EN 1994-2 gives no specic guidance on the detailing of the transverse and longitudinal
joints between precast deck units. Transverse joints between full-depth precast slabs at the
intermediate supports of continuous bridges are particularly critical. Here, the slab reinforce-
ment must transmit the tension caused by both global hogging moments and the bending
moment from local loading. To allow for full laps in the reinforcement, a clear gap
between units would need to be large and a problem arises as to how to form the sot to
the joint. One potential solution is to reduce the gap by using interlocking looped bars
protruding from each end of adjoining slab units. Such a splicing detail is not covered in
EN 1992-2, other than in the strut-and-tie rules. Experience has shown that even if
satisfactory ultimate performance can be established by calculation, tests may be required
to demonstrate acceptable performance at the serviceability limit state and under fatigue
loading.
The publication Precast Concrete Decks for Composite Highway Bridges
109
gives further
guidance on the detailing of longitudinal and transverse joints for a variety of bridge types.
Clause 8.4.1(1)
Clause 8.4.2(1)
Clause 8.4.3
Clause 8.4.3(2)
Clause 8.4.3(3)
181
CHAPTER 8. PRECAST CONCRETE SLABS IN COMPOSITE BRIDGES
CHAPTER 9
Composite plates in bridges
.
General Clause 9.1
.
Design for local eects Clause 9.2
.
Design for global eects Clause 9.3
.
Design of shear connectors Clause 9.4
9.1. General
A composite plate comprises a steel plate acting compositely with a concrete slab in both
longitudinal and transverse directions. The requirements of Section 9 apply to composite
top anges of box girders, which resist local wheel loads in addition to performing the func-
tion of a ange in the global system. Clause 9.1(1) claries that this section of EN 1994-2
does not cover composite plates with shear connectors other than headed studs, or sandwich
construction where the concrete is enclosed by a top and bottom steel plate. Composite plates
can also be used as bottom anges of box girders in hogging zones. This reduces the amount
of stiening required to prevent buckling. Composite bottom anges have been used both in
new bridges
110;111
and for strengthening older structures.
Clause 9.1(2) imposes a deection limit on the steel ange under the weight of wet
concrete, unless the additional weight of concrete due to the deection is included in the
calculation. In most bridges where this deection limit would be approached, the steel top
plate would probably require stiening to resist the global compression during construction.
Clause 9.1(3) gives a modied denition for b
0
in clause 5.4.1.2 on shear lag. Its eect
is that where the composite plate has no projection beyond an outer web, the value of
b
0
for that web is zero. For global analysis, the eects of staged construction, cracking,
creep and shrinkage, and shear lag all apply. Clause 9.1(4) therefore makes reference to
clause 5.4, together with clause 5.1 on structural modelling.
9.2. Design for local eects
Local eects arise from vertical loading, usually from wheels or ballast, acting on the
composite plate. For anges without longitudinal stieners, most of the load is usually
carried by transverse spanning between webs, but longitudinal spanning also occurs in the
vicinity of any cross-beams and diaphragms. For anges with longitudinal stieners, the
direction of spanning depends on the ange geometry and the relative stinesses of
the various components. It is important to consider local loading for the fatigue check of
the studs as the longitudinal shear from wheel loads can be as signicant as that from the
global loading in low-shear regions of the main member.
Clause 9.1(1)
Clause 9.1(2)
Clause 9.1(3)
Clause 9.1(4)
Clause 9.2(1) permits the local analysis to be carried out using elastic analysis with
uncracked concrete properties throughout. This is reasonable because the concrete is
likely to be cracked in exure regardless of the sign of the bending moment. There is therefore
no need to distinguish between uncracked and cracked behaviour, although where the steel
ange is in tension, the cracked stiness is likely to be signicantly higher for sagging
moments than for hogging moments. The same assumption is made in the design of
reinforced concrete and is justied at ultimate limit states by the lower-bound theorem of
plasticity. Clause 9.2(1) also claries that the provisions of Section 9 need not be applied
to the composite ange of a discrete steel I-girder, since the ange will not usually be wide
enough for signicant composite action to develop across its width.
A small amount of slip can be expected between the steel plate and concrete slab, as
discussed in the comments under clause 9.4(4), but as in beams its eect on composite
action is small. Clause 9.2(2) therefore allows slip to be ignored when determining
resistances. Excessive slip could however cause premature failure. This needs to be prevented
by following the applicable provisions of clause 6.6 on shear connection in conjunction with
clause 9.4.
Providing the shear studs are designed as above, the steel deck plate may be taken to act
fully compositely with the slab. Clause 9.2(3) then permits the section to be designed for
exure as if the steel ange plate were reinforcement. The requirements of EN 1992-2
clause 6.1 should then be followed. The shear resistance may similarly be derived by treating
the composite plate as a reinforced concrete section without links according to EN 1992-2
clause 6.2.2 (as modied by clause 6.2.2.5(3)), provided that the spacing of the studs trans-
versely and longitudinally is less than three times the thickness of the composite plate. The
studs should also be designed for the longitudinal shear ow from local loading for ultimate
limit states, other than fatigue, and for the shear ow from combined global and local eects
at serviceability and fatigue limit states.
Both punching and exural shear should be checked. Checks on exural shear for unstif-
fened parts of the composite plate should follow the usual procedures for reinforced concrete
design. An eective width of slab, similar to that shown below in Fig. 9.1, could be assumed
when determining the width of slab resisting exural shear. Checks on punching shear could
consider any support provided by longitudinal stieners, although this could conservatively
be ignored.
9.3. Design for global eects
Clause 9.3(1) requires the composite plate to be designed for the eects induced in it by axial
force, bending moment and torsion acting on the main girder. In the longitudinal direction,
the composite plate will therefore resist direct compression or tension. Most bridge box
girders will be in Class 3 or Class 4 and therefore the elastic stresses derived in the concrete
and steel elements should be limited to the values in clause 6.2.1.5 for ultimate limit states.
Torsion acting on the box will induce in-plane shear in both steel and concrete elements of
the ange. These shear ows can be determined using a transformed section for the concrete
as given in clauses 5.4.2.2(11) and 5.4.2.3(6). Checks of the steel ange under combined
direct stress and in-plane shear are discussed under the comments on clause 6.2.2.4(3).
The concrete ange should be checked for in-plane shear in accordance with EN 1992-2
clause 6.2.
Distortion of a box girder will cause warping of the box walls, and thus in-plane bending in
the composite plate. The direct stresses from warping will need to be added to those from
global bending and axial force. Distortion will also cause transverse bending of the compo-
site plate.
Once a steel ange in compression is connected to the concrete slab, it is usually assumed
that the steel ange panels are prevented from buckling (providing the shear studs are spaced
suciently closely clause 9.4(7) refers). It is still possible, although very unlikely, that the
composite plate might buckle as a whole. Clause 9.3(2) acknowledges this possibility and
Clause 9.2(1)
Clause 9.2(2)
Clause 9.2(3)
Clause 9.3(1)
Clause 9.3(2)
184
requires reference to be made to clause 5.8 of EN 1992-1-1 for the calculation of the second-
order eects. None of the simple methods of accounting for second-order eects in this
clause apply to plates so a general second-order non-linear analysis with imperfections
would be required in accordance with EN 1992-1-1 clause 5.8.6. No guidance is given in
EN 1992-1-1 on imperfections in plate elements. The imperfection shape could be based
on the elastic critical buckling mode shape for the composite plate. The magnitude of
imperfection could be estimated as the sum of the plate imperfection given in EN 1090
and the deection caused by wet concrete, less any specied camber of the plate.
EN 1992-1-1 clause 5.8.2(6) provides a criterion for ignoring second-order eects which
requires them rst be calculated. This is unhelpful. A simpler alternative would be to use
the criterion in clause 5.2.1(3) based on an elastic critical buckling analysis of the composite
plate.
Where account should be taken of signicant shear force acting on the studs in both
longitudinal and transverse directions simultaneously, clause 9.3(4) requires the force on
the connectors to be based on the vector sum. Hence,
P
Ed

P
2
l;Ed
P
2
t;Ed
q
where P
l;Ed
and P
t;Ed
are the shear forces per stud in the longitudinal and transverse
directions respectively. This can inuence the spacing of the studs nearest to the webs
because they are the most heavily stressed from global eects (see section 9.4 below) and
also tend to be the most heavily loaded from local eects in the transverse direction.
9.4. Design of shear connectors
The eect of clause 9.4(1) and clause 9.3(4) is that local and global eects need only be
combined in calculations for serviceability and fatigue limit states, but not for other ultimate
limit states. Several justications can be made for this concession. The main ones are as
follows.
.
The eects of local loading are usually high only over a relatively small width compared
with the total width providing the global resistance.
.
The composite plate will have signicant reserves in local bending resistance above that
obtained from elastic analysis.
.
Complete failure requires the deformation of a mechanism of yield lines, which is resisted
by arching action. This action can be developed almost everywhere in the longitudinal
direction and in many areas in the transverse direction also.
There is no similar relaxation in EN 1993-2 for bare steel anges so a steel ange should, in
principle, be checked for any local loading in combination with global loading. It will not
normally be dicult to satisfy this check, even using elastic analysis.
For serviceability calculations, elastic analysis is appropriate. This greatly simplies the
addition of global and local eects. For the serviceability limit state, the relevant limiting
force per connector is that in clause 6.8.1(3), referred to from clause 7.2.2(6). The force
per connector should be derived according to clause 9.3(4). When checking the Von
Mises equivalent stress in the steel ange plate, the weight of wet concrete carried by it
should be included.
No guidance is given on the calculation of stud shear ow from local wheel loads, other
than that in clause 9.2(1). Longitudinally stiened parts of composite plates can be designed
as beams spanning between transverse members, where present. The rules for eective width
of clause 5.4.1.2 would apply in determining the parts of the composite plate acting with each
longitudinal stiener. For unstiened parts of the composite plate, spanning transversely
between webs or longitudinally between cross-beams, a simplied calculation of shear
ow could be based on an equivalent simply-supported beam. A reasonable assumption
for beam width, as recommended in Ref. 112, would be the width of the load, w, plus
four-thirds of the distance to the nearest support, x, as shown in Fig. 9.1.
Clause 9.3(4)
Clause 9.4(1)
185
CHAPTER 9. COMPOSITE PLATES IN BRIDGES
At cross-sections where there is an abrupt change from composite plate to reinforced
concrete section, such as at the web of the box in Fig. 9.2, the method of clause 6.6.2.4
can be used to determine the transverse shear on the studs near the edge of the plate.
Although the composite section formed by a steel ange and concrete slab might provide
adequate strength against local sagging moments without additional transverse reinforce-
ment, transverse reinforcement is still required in the bottom of the slab to control cracking
and prevent splitting of the concrete ahead of the studs. Clause 9.4(2) requires a fairly
modest quantity of bottom reinforcement to be provided in two orthogonal directions. It
implies that in the absence of such reinforcement, the static design resistances of studs
given in clause 6.6.3.1(1) cannot be used as they assume that splitting is prevented. The limit-
ing fatigue stress range for studs provided in clause 6.8.3 is also inappropriate without some
transverse reinforcement as splitting will increase the exural stresses in the stud.
Clause 9.4(3) refers to the detailing rules of clause 6.6.5. The minimum steel ange
thickness in clause 6.6.5.7(3) is only likely to become relevant where the top ange is
heavily stiened as discussed in the comments on that clause.
The force on shear connectors in wide composite anges is inuenced both by shear lag in
the concrete and steel anges and also slip of the shear connection. At the serviceability limit
state, these lead to a non-uniform distribution of connector force across the ange width.
This distribution can be approximated by equation (9.1) in clause 9.4(4):
P
Ed

v
L;Ed
n
tot
3:85
n
w
n
tot
0:17
3
1
x
b
2
0:15
9:1
Equation (9.1) was derived from a nite-element study by Moat and Dowling.
113
The
study considered only simply-supported beams with ratios of ange half-breadth between
webs (b in equation (9.1)) to span in the range 0.05 and 0.20. The stud stiness was taken
as 400 kN/mm.
The studs nearest the web can pick up a signicantly greater force than that obtained by
dividing the total longitudinal shear by the total number of connectors. This is illustrated in
Example 9.1 and in Ref. 74. Connectors within a distance of the greater of 10t
f
and 200 mm
are assumed to carry the same shear force. This result is obtained by using x 0 in equation
(9.1) when calculating the stud force and it is necessary to avoid underestimating the force,
compared to the nite-element results, in the studs nearest the web. The rule is consistent
with practice for anges of plate girders, where all shear connectors at a cross-section are
assumed to be equally loaded.
The assumed value of stud stiness has a signicant eect on the transverse distribution of
stud force as greater slip leads to a more uniform distribution. Recent studies, such as that in
Ref. 98, have concluded that stud stinesses are signicantly lower than 400 kN/mm. The
same value of stiness is probably not appropriate for both fatigue calculation and service-
ability calculations under the characteristic load combination, due to the greater slip, and
Clause 9.4(2)
Clause 9.4(3)
Clause 9.4(4)
x
2
3
w 4x/3 + w
L
Fig. 9.1. Eective beam width for the determination of shear ow in a composite plate
186
therefore exibility, possible in the latter case. Nevertheless, the assumed stiness of 400 kN/
mm is an upper bound and therefore the transverse distribution is conservative.
Clause 9.4(5) permits a relaxation of the requirements of clause 9.4(4) for composite
bottom anges of box girders, provided that at least half of the shear connectors required
are concentrated near the webange junction. Near means either on the web or within
the dened adjacent width b
f
of the ange. The rule is based on extensive practice in
Germany, and assumes that there is no signicant local loading.
At the ultimate limit state, plasticity in the ange and increased slip lead to a much more
uniform distribution of stud forces across the box, which is allowed for in clause 9.4(6).
To prevent buckling of the steel compression ange in half waves between studs,
clause 9.4(7) refers to Table 9.1 for limiting stud spacings in both longitudinal and transverse
directions. These could, in principle, be relaxed if account is taken of any longitudinal
stiening provided to stabilise the compression ange prior to hardening of the concrete.
Most bridge box girders will have webs in Class 3 or Class 4, so it will usually only be neces-
sary to comply with the stud spacings for a Class 3 ange; there is however little dierence
between the spacing requirements for Class 2 and Class 3.
Clause 9.4(5)
Clause 9.4(6)
Clause 9.4(7)
Example 9.1: design of shear connection for global eects at the serviceability
limit state
The shear connection for the box girder shown in Fig. 9.2 is to be designed using 19 mm
stud connectors. For reasons to be explained, it may be governed by serviceability, for
which the longitudinal shear per web at SLS (determined from elastic analysis of the
cross-section making allowance for shear lag) was found to be 800 kN/m.
25 mm thick
200 580 580
100
3525
250 mm thick
1 2 3 4
605
x
Fig. 9.2. Box girder for Example 9.1
From clause 6.8.1(3) the force per connector at SLS is limited to 0.75P
Rd
. From
Example 6.10, this limit is 0:75 83:3 62:5 kN.
It will be found that longitudinal shear forces per stud decrease rapidly with distance
from the web. This leaves spare resistance for the transverse shear force and local
eects (e.g. from wheel loads), both of which can usually be neglected adjacent to the
web, but increase with distance from it.
113
The following method is further explained in
Ref. 74. It is based on nite-element analyses that included shear lag, and so is applied
to the whole width of composite plate associated with the web concerned, not just
to the eective width. In this case, both widths are 3525=2 1762 mm, denoted b in
Fig. 9.1 in EN 1994-2.
From clause 9.4(4), n
w
is the number of connectors within 250 mm of the web, because
10t
f
> 200 mm. The method is slightly iterative, as the rst step is to estimate the ratio
n
w
=n
tot
, here taken as 0.25. The longitudinal spacing of the studs is assumed to be
0.15 m, so the design shear per transverse row is v
L;Rd
800 0:15 120 kN.
187
CHAPTER 9. COMPOSITE PLATES IN BRIDGES
If all the studs were within 250 mm of the web, their x-coordinate in equation (9.1)
would be zero. The number of studs required is found from this equation by putting
P
Ed
0:75P
Rd
62:5 kN and x 0:
6:25
120
n
tot
3:850:25
0:17
3 0:15
whence n
tot
3:88. Design is therefore based on four studs per transverse row, of which
one will be within 250 mm of the web. To conform to the assumptions made in deriving
equation (9.1), the other studs will be equally spaced over the whole width b, not the
width b 250 mm. (This distinction only matters where n
tot
is low, as it is here.) In a
wider box, non-uniform lateral spacing may be more convenient, subject to the condition
given in the denition of n
tot
in clause 9.4(4).
For these studs, uniform spacing means locations at x b=6, b=2 and 5b=6 that is, at
x 294, 881 and 1468 mm. These are rounded to x 300, 880 and 1460 mm. The
maximum lateral spacing is midway between the webs: 3525 2 1460 605 mm.
This spacing satises the provisions on spacing of clause 9.2(3) (i.e.
<3 275 825 mm) and also those of Table 9.1 for Class 3 behaviour. It would not be
necessary to comply with the latter if the longitudinal stieners were close enough to
ensure that there is no reduction to the eective width of the plate for local sub-panel
buckling; this is not the case here.
The shear force P
Ed
per stud is found from equation (9.1), which is:
P
Ed

120
4
3:850:25
0:17
3
1
x
1762
2
0:15
Results are given in Table 9.1.

Table 9.1. Forces in studs from global eects
Stud No. 1 2 3 4 Total
x (mm) 100 300 880 1460
P
Ed
(kN) 60.7 43.2 18.6 6.2 129
The shear resisted by the four studs, 129 kN, exceeds 120 kN, and no stud is overloaded,
so the spacing is satisfactory. The result, 129 kN, diers from 120 kN because equation
(9.1) is an approximation. This ratio, 129/120, is a function, given in a graph in
Ref. 74, of n
w
=n
tot
and of the eective-width ratio b
eff
=b. Where the lateral spacing of
the n
tot
n
w
studs is uniform, as here, the ratio exceeds 1.0 provided that n
w
=n
tot
< 0:5
and b
eff
=b > 0:7. Its minimum value is about 0.93. If the design needs to be modied,
revision of the longitudinal spacing is a simple method, as the ratios between the forces
per stud do not then change.
From clause 9.4(6), at ULS the studs may be assumed to be equally loaded. Here, their
design resistance is:
n
tot
=0:15P
Rd
=
V
4=0:1583:3=1:25 1777 kN=m
This is more than twice the design shear for SLS, so ULS is unlikely to govern. Table 9.1
shows that for SLS, studs 2 to 4 have a reserve of resistance (cf. 62.5 kN) that should be
sucient for transverse shear and local eects.
188
CHAPTER 10
Annex C (Informative).
Headed studs that cause
splitting forces in the direction
of the slab thickness
This chapter corresponds to Annex C of EN 1994-2, which has the following clauses:
.
Design resistance and detailing Clause C.1
.
Fatigue strength Clause C.2
Annex A of EN 1994-1-1 is for buildings only. Annex B of EN 1994-1-1, Standard tests,
for shear connectors and composite oor slabs, is not repeated in EN 1994-2. Comment on
these annexes is given in Ref. 5.
Annex C gives a set of design rules for the detailing and resistance of shear studs that are
embedded in an edge of a concrete slab, as shown in Figs 6.13 and C.1 of EN 1994-2 and
in Fig. 6.35. Details of this type can occur at an edge of a composite deck in a tied arch
or half-through bridge, or where double composite action is used in a box girder. The
same problem, premature splitting, could occur in a steep-sided narrow haunch. The use
of such haunches is now discouraged by the 458 rule in clause 6.6.5.4(1).
The rules in Annex C were developed from research at the University of Stuttgart
that has been available in English only since 2001.
82.114.115
These extensive push tests
and nite-element analyses showed that to avoid premature failure by splitting of the
slab and to ensure ductile behaviour, special detailing rules are needed. Clause
6.6.3.1(3) therefore warns that the usual rules for resistance of studs do not apply.
The new rules, in Annex C, are necessarily of limited scope, because there are so
many relevant parameters. The rules are partly based on elaborate strut-and-tie model-
ling. It was not possible to nd rules that are dimensionally consistent, so the units to
be used are specied the only occasion in EN 1994 Parts 1-1 and 2 where this has
been necessary. For these reasons, Annex C is Informative, even though its guidance
is the best available. The simplied and generally more conservative rules given in
clause 6.6.4 do not cover interaction with transverse (e.g. vertical) shear or resistance to
fatigue.
It will be found that these lying studs have to be much longer than usual, and that the
minimum slab thickness to avoid a reduction in the shear resistance per stud can exceed
250 mm. The comments that follow are illustrated in Example 10.1, and in Fig. 10.1 where
the longitudinal shear acts normal to the plane of the gure and vertical shear acts down-
wards from the slab to the steel web.
C.1. Design resistance and detailing
Clause C.1(1) gives the static resistance of a stud to longitudinal shear in the absence of
vertical shear, which should not be taken as greater than that from clause 6.6.3.1(1). The
minimum length h of the stud and the reinforcement details are intended to be such that
splitting of the slab is followed by fracture or pulling out of the stud, giving a ductile
mode of failure. The important dimensions are a
0
r.o
and v, from the stud to the centre-lines
of the stirrup reinforcement, as shown in Fig. 10.1.
Equation (C.1), repeated in the Example, uses factor k
v
to distinguish between two
situations. The more favourable, where k
v
1.14, applies where the slab is connected to
both sides of the web and resists hogging bending a middle position. This requires re-
inforcement to pass continuously above the web, as shown in Fig. C.1. Some shear is then
transferred by friction at the face of the web. Where this does not occur, an edge position,
k
v
has the lower value 1.0. Details in bridge decks are usually edge positions, so further
comment is limited to these. The geometries considered in the Stuttgart tests, however,
covered composite girders where the steel top ange was omitted altogether, with the web
projecting into the slab.
The general symbol for distance from a stud to the nearest free surface is a
r
, but notation
a
r.o
is used for the upper surface, from the German oben, above. Its use is relevant where
there is vertical shear, acting downwards from the slab to the studs. Allowing for cover
and the stirrups, the important dimension is:
a
0
r.o
a
r.o
c
v
c
s
,2
If the lower free surface is closer to the stud, its dimension a
0
r
should be used in place of a
0
r.o
.
Clause 6.6.4 appears to cover only this edge position layout, and uses the symbol e
v
in
place of a
0
r.o
or a
0
r
.
Although f
ck
in equation (C.1) is dened as the strength at the age considered, the
specied 28-day value should be used, unless a check is being made at a younger age.
The longitudinal spacing of the stirrups, s, should be related to that of the studs, a, and
should ideally be uniform.
Clause C.1(1)
29
40
33
= 10
s
= 10
c
h
= 40
c
v
= 40
a
r,o
= 80
a
r,o
Centre-line
of slab
10
45 v = 136
h = 191
d = 19
Fig. 10.1. Notation and dimensions for lying studs in Example 10.1
190
For concretes of grade C35/45 and above, the resistance P
Rd
from clause 6.6.3.1 is
independent of the concrete grade. Then, equation (C.1) can be used to nd a minimum
value for a
0
r
such that P
Rd.L
! P
Rd
.
For example, let a s, d 19 mm, f
ck
35 N/mm
2
, k
v
1, and
V
1.25. Then,
P
Rd
90.7 kN from equation (6.18) and P
Rd.L
from equation (C.1) does not govern
unless a
0
r
< 89 mm, say 90 mm. With 40 mm cover and 10 mm stirrups, the minimum slab
thickness is then 290 40 5 270 mm if the studs are centrally placed, but greater if
they are o centre. If a thinner slab is required, the ratio a,s can be increased or more
studs provided, to compensate for a value P
Rd.L
< P
Rd
.
The limits on v in clause C.1(2) are more convenient for use in practice than the limits on
u, because the angle u is dened by the position of the longitudinal corner bar within the
bend of the stirrup (Fig. 10.1), which is dicult to control on site.
It follows from clause C.1(3) that the minimum stirrup diameter c
s
is roughly propor-
tional to the stud diameter d. Where a/s 1.0, c
s
% d/2.
The expression for interaction between longitudinal and vertical shear, clause C.1(4), is
only slightly convex. The vertical shear resistance given in Example 10.1, equation (C.4),
is typically less than 40% of the longitudinal shear resistance, being governed mainly by
the upper edge distance a
r.o
. Application of vertical shear to lying studs is best avoided,
and can be minimized by spanning the concrete slab longitudinally between cross-beams.
C.2. Fatigue strength
Equation (C.5) in clause C.2(1), P
R
m
N P
c
m
N
c
, diers from equation (6.50) in
clause 6.8.3(3) (for the fatigue strength of studs in a non-lying position) only in that
symbols P in equation (C.5) appear as t in equation (6.50). Both methods use m 8.
In Annex C, P
c
is a function of dimension a
0
r
(Fig. C.1, Fig. 10.1), and is 35.6 kN at
2 million cycles, for a
0
r
! 100 mm. With typical covers to reinforcement, this corresponds
to a lying stud at mid-depth of a slab at least 280 mm thick, where the splitting eect is
probably minimal. This value 35.6 kN should therefore correspond to the value given in
clause 6.8.3 for t at 2 million cycles, 90 N/mm
2
. It does so when the stud diameter is
22.4 mm. This is as expected, because in the fatigue tests, only 22 mm diameter studs were
used. However, the method of Annex C is provided for studs of diameter 19 mm to
25 mm, for which: numerous FE-calculations show, that the fatigue strength curve should
be based on the absolute range of shear force per stud rather than on the range of shear
stress (p. 9 of Ref. 115).
The value of P
R
given by the rule in clause 6.8.3 is proportional to the square of the
shank diameter of the stud. For 19 mm studs, it is only 25.6 kN at 2 10
6
cycles, which is
only 72% of the resistance 35.6 kN given in Table C.1. This is why clause C.2(1) requires
the lower of the two values from clause 6.8.3 and Annex C to be used.
Clause C.2(2) refers to the recommended upper limit for the longitudinal shear force
per connector, which is 0.75P
Rd
. The word longitudinal reveals that the rule for fatigue
resistance does not apply where vertical shear is present, which is not clear in clause
C.2(1). There were no fatigue tests in combined longitudinal and vertical shear another
reason why application of vertical shear to lying studs is best avoided.
Applicability of Annex C
The denition of a lying stud, given in the titles of clause 6.6.4 and Annex C, does not state
how far from a free surface a stud must be, for the normal rules for its resistance and the
detailing to apply.
Clause 6.6.5.3 refers to a longitudinal edge, not a top surface, but appears to deal with
the same problem of splitting parallel to a free surface nearby. Where the edge distance
(i.e. a
r.o
in Fig. 10.1) is less than 300 mm, it species U-bars (i.e. stirrups) of diameter
c
s
! 0.5d and an edge distance a
r
! 6d. These limits correspond closely with the results of
Clause C.1(2)
Clause C.1(3)
Clause C.1(4)
Clause C.2(1)
Clause C.2(2)
191
CHAPTER 10. ANNEX C (INFORMATIVE)
Example 10.1, c
s
0.53d and a
r
6.6d, but the stirrups are required to pass around the
studs, whereas in Annex C they pass between them. The dierence is that the surface parallel
to the plane of splitting, AB in Fig. 10.2, is normal to the plane of the slab in one case, not in
the other. The minimum height of the stud, about 90 mm to clause 6.6.5.1(1) and 191 mm in
Example 10.1, is less signicant in the detail in Fig. 10.2(a) than in detail (b). Details may
occur where clause 6.6.5.3 is also applicable, but it does not clarify the scope of Annex C.
Let us consider the options, as the local thickness h
c
of the slab in Fig. 10.1 is increased
(e.g. by the addition of an upstand haunch) without change to other details. If the top
cover is maintained, length a
0
r.o
must increase, so from clause C.1(2) for v, the studs have
to be longer. An alternative is to keep a
0
r.o
and v unchanged, by increasing the cover to the
legs of the stirrups. When h
c
exceeds 300 mm (using data from Example 10.1), two rows
of studs are possible, Fig. 10.2(b), because the minimum vertical spacing of studs is 2.5d,
from clause 6.6.5.7(4). In the absence of vertical shear, the shear resistance is doubled, as
is the potential splitting force. For two rows of lying studs, the force T
d
given by equation
(C.2) should be 0.3 times the sum of their resistances.
Limits to the applicability of Annex C are further discussed at the end of Example 10.1.
Example 10.1: design of lying studs
A bridge deck slab 250 mm thick is to be connected along its sides to steel webs. The nal
detail is shown to scale in Fig. 10.1. The design ultimate longitudinal shear at each edge is
V
L.Ed
600 kN/m. The vertical shear is at rst assumed to be negligible. The eect of
adding V
V.Ed
60 kN/m is then found. The strengths of the materials are:
f
yd
500,1.15 435 N/mm
2
, f
u
500 N/mm
2
and f
ck
40 N/mm
2
. Resistances in
Annex C are given in terms of f
ck
, not f
cd
, because the partial factor
V
allows for both
the stud material and the concrete. The minimum cover is 40 mm.
There is space only for a single row of studs at mid-depth, and 19 mm studs are pre-
ferred. Assuming the studs have a height greater than four times their shank diameter,
their shear resistance to equation (6.19) is:
P
Rd
0.29 19
2

40 35 000
p
,1.25 1000 99.1 kN,stud D10.1
and to equation (6.18) is:
P
Rd
0.8 500 19
2
,4 1.25 90.7 kN,stud D10.2
so equation (6.18) governs.
This gives a spacing of 90.7,600 0.151 m, so a spacing a 150 mm is assumed, with
stirrups also at 150 mm. Their required diameter c
s
can be estimated from clause C.1(3),
equation (C.2):
T
d
0.3P
Rd.L
Using P
Ed.L
as an approximation to P
Rd.L
:
T
d
c
2
s
,4 500,1.15 ! 0.3 600 1000 0.15
A
B
C D
125
50
125
125
D
C B A
50
125
125
(c) (b) (a)
50
Fig. 10.2. Cross-sections at an edge of a slab: (a) edge studs; (b) lying studs in an edge position; and
(c) three rows of lying studs
192
This gives c
s
! 8.9 mm, so 10 mm stirrups are assumed for nding a
0
r.o
.
With cover of 40 mm,
a
0
r.o
250 240 5,2 80 mm
From equation (C.1) with k
v
1.0 for an edge position and a s,
P
Rd.L
1.4k
v
f
ck
da
0
r
0.4
a,s
0.3
,
V
1.440 19 80
0.4
,1.25 91.7 kN,stud C.1
This is greater than result (D10.2), so the shear resistance is governed by equation (6.18).
Where this occurs, the term P
Rd.L
in equations (C.2) and (C.3) should be replaced by the
governing shear resistance.
For studs at 0.15 m spacing, V
Rd.L
90.7,0.15 605 kN/m which is sucient. This
result requires the diameter of the stirrups to be increased to 8.9 605,600 9.0 mm,
so 10 mm is still sucient. The length of the studs is governed by a
0
r
. From clause C.1(2):
.
for uncracked concrete, v ! maxf110. 1.7 80. 1.7 75g 136 mm
.
for cracked concrete, v ! maxf160. 2.4 80. 2.4 75g 192 mm
To these lengths must be added c
h
c
s
/2 (i.e. 45 mm) plus the thickness of the head of
the stud, 10 mm, giving lengths h after welding of 191 mm and 247 mm respectively. In
practice, these would be rounded up. Here, h is taken as 191 mm. Figure 10.1 shows
that the angle u can then be anywhere between 108 and 338. This is roughly consistent
with the alternative rule in clause C.1(2), u 308.
The simpler rules of clause 6.6.4 would require a
0
r.o
to be increased from 80 mm to
6 19 114 mm. It would then be necessary for the slab thickness to be increased
locally from 250 mm to at least 318 mm. Dimension v would be increased from 136 or
192 mm to 266 mm, requiring much longer studs.
In the absence of vertical shear, a check on fatigue to clause C.2(1) is straightforward.
Interaction with vertical shear
The design vertical shear is F
d.V
0.15 60 9 kN/stud.
Clause C.1(4) requires longitudinal reinforcement with c
l
16 mm, so this value is
used. From equation (C.4) with k
v
1.0 and a s,
P
Rd.V
0.012 f
ck
c
l
0.5
da,s
0.4
c
s
0.3
a
0
r.o
0.7
k
v
,
V
C.4
0.01240 16
0.5
19
0.4
10
0.3
80
0.7
1.0,1.25 33.8 kN,stud
From interaction expression (C.3):
600 0.15,90.7
1.2
9,33.8
1.2
1.20 (but 1.0)
which is too large. Changing the spacing of the studs and stirrups from 150 mm to 125 mm
would reduce result 1.20 to 0.96.
Two or more rows of lying studs
The detail shown in Fig. 10(c) is now considered, with materials and studs as before. The
deck spans longitudinally, so it can be assumed that no vertical shear is applied to the
three rows of lying studs. The preceding calculations apply to the top and bottom
rows. For the middle row, a
0
r
increases from 80 mm to 130 mm. From equation (C.1)
the resistance of its studs increases from 91.7 kN to 111 kN, so result (D10.2), 90.7 kN,
still governs. Assuming uncracked concrete, the minimum height of these studs increases
by 1.7 times the change in a
0
r
, from clause C.1(2); that is, from 191 mm to 276 mm. This
situation was not researched, and the increase may not be necessary; but the stirrups
should be designed for a force
T
d
0.33 90.7 82 kN per 150 mm length of beam,
if the full shear resistance of the studs is needed.
193
CHAPTER 10. ANNEX C (INFORMATIVE)
References
European Standards listed as EN. . . are being published in each Member State of the Comite
Europe en de Normalisation (CEN) by its National Standards organisation between 2002
and 2007. In the UK, publication is by the British Standards Institution, London, as BS
EN. . . .
The Eurocodes, EN 1990 to EN 1999, are accompanied by National Annexes. These
annexes for the UK are expected to be completed by the end of 2007.
1. Gulvanessian, H., Calgaro, J.-A. and Holicky , M. (2002) Designers Guide to EN 1990.
Eurocode: Basis of Structural Design. Thomas Telford, London.
2. Calgaro, J.-A., Tschumi, M., Gulvanessian, H. and Shetty, N. Designers Guide to EN
1991-1-1, 1991-1-3, 1991-1-5 to 1-7 and 1991-2, Eurocode 1: Actions on Structures.
(Trac loads and other actions on bridges). Thomas Telford, London (in preparation).
3. Smith, D. and Hendy, C. R. Designers Guide to EN 1992-2. Eurocode 2: Design of
Concrete Structures. Part 2: Bridges. Thomas Telford, London (in preparation).
4. Murphy, C. J. M. and Hendy, C. R. Designers Guide to EN 1993-2. Eurocode 3:
Design of Steel Structures. Part 2: Bridges. Thomas Telford, London (to be published,
2007).
5. Johnson, R. P. and Anderson, D. (2004) Designers Guide to EN 1994-1-1. Eurocode 4:
Design of Composite Steel and Concrete Structures. Part 1-1: General Rules and Rules
for Buildings. Thomas Telford, London.
6. Beeby, A. W. and Narayanan, R. S. (2005) Designers Guide to EN 1992. Eurocode 2,
Design of Concrete Structures. Part 1-1: General Rules and Rules for Buildings.
Thomas Telford, London.
7. Gardner, L. and Nethercot, D. (2005) Designers Guide to EN 1993. Eurocode 3,
Design of Steel Structures. Part 1-1: General Rules and Rules for Buildings. Thomas
Telford, London.
8. British Standards Institution. Design of Composite Steel and Concrete Structures. Part
1-1: General Rules and Rules for Buildings. BSI, London, EN 1994.
9. British Standards Institution. Design of Composite Steel and Concrete Structures. Part
2: General Rules and Rules for Bridges. BSI, London, EN 1994.
10. The European Commission (2002) Guidance Paper L (Concerning the Construction
Products Directive 89/106/EEC). Application and Use of Eurocodes. EC, Brussels.
11. British Standards Institution. Steel, Concrete and Composite Bridges. (In many Parts.)
BSI, London, BS 5400.
12. Hanswille, G. (2006) The new German design code for composite bridges. In: Leon,
R. T. and Lange, J. (eds), Composite Construction in Steel and Concrete V. American
13. British Standards Institution. Eurocode: Basis of Structural Design. (Including
Annexes for Buildings, Bridges, etc.). BSI, London, EN 1990.
14. British Standards Institution. Actions on Structures. BSI, London, EN 1991. (In many
Parts.) See also Refs 26, 30, 53 and 103.
15. British Standards Institution. Design of Concrete Structures. BSI, London, EN 1992.
(In several Parts.) See also Ref. 27.
16. British Standards Institution. Design of Steel Structures. BSI, London, EN 1993. (In
many Parts.) See also Refs 19, 28, 38, 41, and 42.
17. British Standards Institution. Design of Structures for Earthquake Resistance. BSI,
London, EN 1998. (In several Parts.)
18. Niehaus, H. and Jerling, W. (2006) The Nelson Mandela bridge as an example of the
use of composite materials in bridge construction in South Africa. In: Leon, R. T. and
Lange, J. (eds), Composite Construction in Steel and Concrete V. American Society of
19. British Standards Institution. Design of Steel Structures. Part 1-8: Design of Joints.
BSI, London, EN 1993.
20. British Standards Institution (1994) Design of Composite Steel and Concrete
Structures. Part 1-1, General Rules and Rules for Buildings. BSI, London, BS DD
ENV 1994.
21. Hosain, M. U. and Pashan, A. (2006) Channel shear connectors in composite
beams: push-out tests. In: Leon, R. T. and Lange, J. (eds), Composite
Construction in Steel and Concrete V. American Society of Civil Engineers, New
York, pp. 501510.
22. Veljkovic, M. and Johansson, B. (2006) Residual static resistance of welded stud shear
connectors. In: Leon, R. T. and Lange, J. (eds), Composite Construction in Steel and
Concrete V. American Society of Civil Engineers, New York, pp. 524533.
23. Andra , H.-P. (1990) Economical shear connection with high fatigue strength. Pro-
ceedings of a Symposium on Mixed Structures, including New Materials, Brussels.
IABSE, Zurich. Reports 60, 167172.
24. Marecek, J., Samec, J. and Studnicka, J. (2005) Perfobond shear connector behaviour.
In: Homeister, B. and Hechler, O. (eds), Eurosteel 2005, vol. B. Druck und
Verlagshaus Mainz, Aachen, pp. 4.3-1 to 4.3-8.
25. Hauke, B. (2005) Shear connectors for composite members of high strength
materials. In: Homeister, B. and Hechler, O. (eds), Eurosteel 2005, vol. B. Druck
und Verlagshaus Mainz, Aachen, pp. 4.2-57 to 4.2-64.
26. British Standards Institution. Actions on Structures. Part 2: Trac Loads on Bridges.
BSI, London, EN 1991.
27. British Standards Institution. Design of Concrete Structures. Part 2: Bridges. BSI,
London, EN 1992.
28. British Standards Institution. Design of Steel Structures. Part 2: Bridges. BSI,
London, EN 1993.
29. International Organisation for Standardization (1997) Basis of Design for Structures
Notation General Symbols. ISO, Geneva, ISO 3898.
30. British Standards Institution. Actions on Structures. Part 1-6: Actions during
Execution. BSI, London, EN 1991.
31. The European Commission (1989) Construction Products Directive 89/106/EEC,
OJEC No. L40 of 11 February. EC, Brussels.
32. British Standards Institution. Geotechnical Design. BSI, London, EN 1997. (In several
Parts.)
33. Anderson, D., Aribert, J.-M., Bode, H. and Kronenburger, H. J. (2000) Design
rotation capacity of composite joints. Structural Engineer, 78, No. 6, 2529.
34. Working Commission 2 (2005) Use and application of high-performance steels for
steel structures. Structural Engineering Documents 8, IABSE, Zurich.
35. Morino, S. (2002) Recent developments on concrete-lled steel tube members in
Japan. In: Hajjar, J. F., Hosain, M., Easterling, W. S. and Shahrooz, B. M. (eds),
Composite Construction in Steel and Concrete IV. American Society of Civil Engineers,
New York, pp. 644655.
196
36. Hegger, J. and Do inghaus, P. (2002) High performance steel and high performance
concrete in composite structures. In: Hajjar, J. F., Hosain, M., Easterling, W. S.
and Shahrooz, B. M. (eds), Composite Construction in Steel and Concrete IV.
37. Homeister, B., Sedlacek, G., Mu ller, C. and Ku hn, B. (2002) High strength materials
in composite structures. In: Hajjar, J. F., Hosain, M., Easterling, W. S. and Shahrooz,
B. M. (eds), Composite Construction in Steel and Concrete IV. American Society of
38. British Standards Institution. Design of Steel Structures. Part 1-3: Cold Formed Thin
Gauge Members and Sheeting. BSI, London, EN 1993.
39. Sedlacek, G. and Trumpf, H. (2006) Composite design for small and medium spans.
In: Leon, R. T. and Lange, J. (eds), Composite Construction in Steel and Concrete V.
40. British Standards Institution. (1998) Welding Studs and Ceramic Ferrules for Arc
Stud Welding. BSI, London, EN 13918.
41. British Standards Institution. Design of Steel Structures. Part 1-5: Plated Structural
Elements. BSI, London, EN 1993.
42. British Standards Institution. Design of Steel Structures. Part 1-9: Fatigue Strength of
Steel Structures. BSI, London, EN 1993.
43. Trahair, N. S., Bradford, M. A. and Nethercot, D. A. (2001) The Behaviour and
Design of Steel Structures to BS 5950, 3rd edn. Spon, London.
44. Johnson, R. P. and Cafolla, J. (1977) Stiness and strength of lateral restraints to
compressed anges. Journal of Constructional Steel Research, 42, No. 2, 7393.
45. Johnson, R. P. and Chen, S. (1991) Local buckling and moment redistribution in
Class 2 composite beams. Structural Engineering International, 1, No. 4, 2734.
46. Johnson, R. P. and Fan, C. K. R. (1988) Strength of continuous beams designed to
Eurocode 4. Proceedings of IABSE, Periodica 2/88, P-125/88, May, pp. 3344.
47. Johnson, R. P. and Huang, D. J. (1995) Composite bridge beams of mixed-class
cross-section. Structural Engineering International, 5, No. 2, 96101.
48. British Standards Institution (1990) Code of Practice for Design of Simple and Con-
tinuous Composite Beams. BSI, London, BS 5950-3-1.
49. Haensel, J. (1975) Eects of Creep and Shrinkage in Composite Construction. Institute
for Structural Engineering, Ruhr-Universita t, Bochum, Report 75-12.
50. Johnson, R. P. and Hanswille, G. (1998) Analyses for creep of continuous steel and
composite bridge beams, according to EC4:Part 2. Structural Engineer, 76, No. 15,
294298.
51. Johnson, R. P. (1987) Shrinkage-induced curvature in cracked concrete anges of
composite beams. Structural Engineer, 65B, Dec., 7277.
52. Guezouli, S. and Aribert, J.-M. (2006) Numerical investigation of moment redistribu-
tion in continuous beams of composite bridges. In: Leon, R. T. and Lange, J. (eds),
Composite Construction in Steel and Concrete V. American Society of Civil Engineers,
New York, pp. 4756.
53. British Standards Institution. Actions on Structures. Part 1-5: Thermal Actions. BSI,
London, EN 1991.
54. British Standards Institution (1997) Design of Composite Structures of Steel and
Concrete. Part 2: Bridges. BSI, London, BS DD ENV 1994.
55. Johnson, R. P. (2003) Cracking in concrete tension anges of composite T-beams
tests and Eurocode 4. Structural Engineer, 81, No. 4, Feb., 2934.
56. Johnson, R. P. (2003) Analyses of a composite bowstring truss with tension stiening.
Proceedings of the Institution of Civil Engineers, Bridge Engineering, 156, June,
6370.
57. Way, J. A. and Biddle, A. R. (1998) Integral Steel Bridges: Design of a Multi-span
Bridge Worked Example. Steel Construction Institute, Ascot, Publication 180.
58. Lawson, R. M. (1987) Design for Openings in the Webs of Composite Beams. Steel
Construction Institute, Ascot, Publication 068.
197
REFERENCES
59. Lawson, R. M., Chung, K. F. and Price, A. M. (1992) Tests on composite beams with
large web openings. Structural Engineer, 70, Jan., 17.
60. Johnson, R. P. and Huang, D. J. (1994) Calibration of safety factors
M
for composite
steel and concrete beams in bending. Proceedings of the Institution of Civil Engineers,
Structures and Buildings, 104, May, 193203.
61. Johnson, R. P. and Huang, D. J. (1997) Statistical calibration of safety factors for
encased composite columns. In: Buckner, C. D. and Sharooz, B. M. (eds), Composite
Construction in Steel and Concrete III, American Society of Civil Engineers, New
York, pp. 380391.
62. British Standards Institution (1997) Structural Use of Concrete. Part 1: Code of
Practice for Design and Construction. BSI, London, BS 8110.
63. Stark, J. W. B. (1984) Rectangular Stress Block for Concrete. Technical paper S16,
June. Drafting Committee for Eurocode 4 (unpublished).
64. Johnson, R. P. and Anderson, D. (1993) Designers Handbook to Eurocode 4. Thomas
Telford, London. [This handbook is for ENV 1994-1-1.]
65. La a ne, A. and Lebet, J.-P. (2005) Available rotation capacity of composite bridge
plate girders with negative moment and shear. Journal of Constructional Steelwork
Research, 61, 305327.
66. Johnson, R. P. and Willmington, R. T. (1972) Vertical shear in continuous composite
beams. Proceedings of the Institution of Civil Engineers, 53, Sept., 189205.
67. Allison, R. W., Johnson, R. P. and May, I. M. (1982) Tension-eld action in
composite plate girders. Proceedings of the Institution of Civil Engineers, Part 2,
Research and Theory, 73, June, 255276.
68. Veljkovic, M. and Johansson, B. (2001) Design for buckling of plates due to direct
stress. In: Ma kela inen, P., Kesti, J., Jutila, A. and Kaitila, O. (eds) Proceedings of
the 9th Nordic Steel Conference, Helsinki, 721729.
69. Lebet, J.-P. and La a ne, A. (2005) Comparison of shear resistance models with slender
composite beam test results. In: Homeister, B. and Hechler, O. (eds), Eurosteel 2005,
vol. B. Druck und Verlagshaus Mainz, Aachen, pp. 4.3-33 to 4.3-40.
70. Ehmann, J. and Kuhlmann, U. (2006) Shear resistance of concrete bridge decks in
tension. In: Leon, R. T. and Lange, J. (eds), Composite Construction in Steel and
Concrete V. American Society of Civil Engineers, New York, pp. 6776.
71. Johnson, R. P. and Fan, C. K. R. (1991) Distortional lateral buckling of continuous
composite beams. Proceedings of the Institution of Civil Engineers, Part 2, 91, Mar.,
131161.
72. Johnson, R. P. and Molenstra, N. (1990) Strength and stiness of shear connections
for discrete U-frame action in composite plate girders. Structural Engineer, 68, Oct.,
386392.
73. Trahair, N. S. (1993) Flexural Torsional Buckling of Structures. E & FN Spon, London.
74. Johnson, R. P. and Buckby, R. J. (1986) Composite Structures of Steel and Concrete,
Vol. 2, Bridges, 2nd edn. Collins, London.
75. Johnson, R. P. and Molenstra, N. (1991) Partial shear connection in composite beams
for buildings. Proceedings of the Institution of Civil Engineers, Part 2, Research and
Theory, 91, 679704.
76. Johnson, R. P. and Oehlers, D. J. (1981) Analysis and design for longitudinal shear in
composite T-beams. Proceedings of the Institution of Civil Engineers, Part 2, Research
and Theory, 71, Dec., 9891021.
77. Menzies, J. B. (1971) CP 117 and shear connectors in steelconcrete composite beams.
Structural Engineer, 49, March, 137153.
78. Johnson, R. P. and Ivanov, R. I. (2001) Local eects of concentrated longitudinal
shear in composite bridge beams. Structural Engineer, 79, No. 5, 1923.
79. Oehlers, D. J. and Johnson, R. P. (1987) The strength of stud shear connections in
composite beams. Structural Engineer, 65B, June, 4448.
80. Roik, K., Hanswille, G. and Cunze-O. Lanna, A. (1989) Eurocode 4, Clause 6.3.2:
Stud Connectors. University of Bochum, Report EC4/8/88, March.
198
81. Stark, J. W. B. and van Hove, B. W. E. M. (1991) Statistical Analysis of Pushout Tests
on Stud Connectors in Composite Steel and Concrete Structures. TNO Building and
Construction Research, Delft, Report BI-91-163, Sept.
82. Kuhlmann, U. and Breuninger, U. (2002) Behaviour of horizontally lying studs with
longitudinal shear force. In: Hajjar, J. F., Hosain, M., Easterling, W. S. and Shahrooz,
B. M. (eds), Composite Construction in Steel and Concrete IV. American Society of
83. Bridge, R. Q., Ernst, S., Patrick, M. and Wheeler, A. T. (2006) The behaviour and
design of haunches in composite beams and their reinforcement. In: Leon, R. T.
and Lange, J. (eds), Composite Construction in Steel and Concrete V. American
84. Johnson, R. P. and Oehlers, D. J. (1982) Design for longitudinal shear in composite
L-beams. Proceedings of the Institution of Civil Engineers, Part 2, Research and
Theory, 73, March, 147170.
85. Johnson, R. P. (2004) Composite Structures of Steel and Concrete, 3rd edn. Blackwell,
Oxford.
86. Bulson, P. S. (1970) The Stability of Flat Plates. Chatto & Windus, London.
87. Roik, K. and Bergmann, R. (1990) Design methods for composite columns with
unsymmetrical cross-sections. Journal of Constructional Steelwork Research, 15,
153168.
88. Wheeler, A. T. and Bridge, R. Q. (2002) Thin-walled steel tubes lled with high
strength concrete in bending. In: Hajjar, J. F., Hosain, M., Easterling, W. S. and
Shahrooz, B. M. (eds), Composite Construction in Steel and Concrete IV. American
89. Kilpatrick, A. and Rangan, V. (1999) Tests on high-strength concrete-lled tubular
steel columns. ACI Structural Journal, Mar.Apr., Title No. 96-S29, 268274. Amer-
ican Concrete Institute, Detroit.
90. May, I. M. and Johnson, R. P. (1978) Inelastic analysis of biaxially restrained
columns. Proceedings of the Institution of Civil Engineers, Part 2, Research and
Theory, 65, June, 323337.
91. Roik, K. and Bergmann, R. (1992) Composite columns. In: Dowling, P. J., Harding,
J. L. and Bjorhovde, R. (eds), Constructional Steel Design an International Guide.
Elsevier, London and New York, pp. 443469.
92. Bergmann, R. and Hanswille, G. (2006) New design method for composite columns
including high strength steel. In: Leon, R. T. and Lange, J. (eds), Composite Con-
struction in Steel and Concrete V. American Society of Civil Engineers, New York,
pp. 381389.
93. Chen, W. F. and Lui, E. M. (1991) Stability Design of Steel Frames. CRC Press, Boca
Raton, Florida.
94. Bondale, D. S. and Clark, P. J. (1967) Composite construction in the Almondsbury
interchange. Proceedings of a Conference on Structural Steelwork, British Construc-
tional Steelwork Association, London, pp. 91100.
95. Virdi, K. S. and Dowling, P. J. (1980) Bond strength in concrete-lled tubes. Proceed-
ings of IABSE, Periodica 3/80, P-33/80, Aug., 125139.
96. Kerensky, O. A. and Dallard, N. J. (1968) The four-level interchange between M4 and
M5 motorways at Almondsbury. Proceedings of the Institution of Civil Engineers, 40,
295321.
97. Johnson, R. P. (2000) Resistance of stud shear connectors to fatigue. Journal of
Constructional Steel Research, 56, 101116.
98. Oehlers, D. J. and Bradford, M. (1995) Composite Steel and Concrete Structural
Members Fundamental Behaviour. Elsevier Science, Oxford.
99. Gomez Navarro, M. (2002) Inuence of concrete cracking on the serviceability limit
state design of steel-reinforced concrete composite bridges: tests and models. In:
J. Martinez Calzon (ed.), Composite Bridges Proceedings of the 3rd International
Meeting, Spanish Society of Civil Engineers, Madrid, pp. 261278.
199
REFERENCES
100. Pucher, A. (1977) Inuence Surfaces of Elastic Plates. Springer-Verlag Wien, New
York.
101. Kuhlmann, U. (1997) Design, calculation and details of tied-arch bridges in composite
constructions. In: Buckner, C. D. and Sharooz, B. M. (eds), Composite Construction in
Steel and Concrete III, American Society of Civil Engineers, New York, pp. 359369.
102. Monnickendam, A. (2003) The design, construction and performance of Newark
Dyke railway bridge. Proceedings of a Symposium on Structures for High-speed
Railway Transportation, Antwerp. IABSE, Zurich. Reports, 87, 4243.
103. British Standards Institution. Actions on Structures. Part 1-4: General Actions Wind
actions. BSI, London, EN 1991.
104. Randl, E. and Johnson, R. P. (1982) Widths of initial cracks in concrete tension
anges of composite beams. Proceedings of IABSE, Periodica 4/82, P-54/82, Nov.,
6980.
105. Johnson, R. P. and Allison, R. W. (1983) Cracking in concrete tension anges of
composite T-beams. Structural Engineer, 61B, Mar., 916.
106. Roik, K., Hanswille, G. and Cunze-O. Lanna, A. (1989) Report on Eurocode 4, Clause
5.3, Cracking of Concrete. University of Bochum, Report EC4/4/88.
107. Johnson, R. P. (2003) Cracking in concrete anges of composite T-beams tests and
Eurocode 4. Structural Engineer, 81, No. 4, 2934.
108. Schmitt, V., Seidl, G. and Hever, M. (2005) Composite bridges with VFT-WIB
construction method. In: Homeister, B. and Hechler, O. (eds), Eurosteel 2005,
vol. B. Druck und Verlagshaus Mainz, Aachen, pp. 4.6-79 to 4.6-83.
109. Yandzio, E. and Iles, D. C. (2004) Precast Concrete Decks for Composite Highway
Bridges. Steel Construction Institute, Ascot, Publication 316.
110. Calzon, J. M. (2005). Practice in present-day steel and composite structures. In:
Homeister, B. and Hechler, O. (eds), Eurosteel 2005, vol. A. Druck und Verlagshaus
Mainz, Aachen, pp. 0-11 to 0-18.
111. Doeinghaus, P., Dudek, M. and Sprinke, P. (2004) Innovative hybrid double-
composite bridge with prestressing. In: Pre-Conference Proceedings, Composite
Construction in Steel and Concrete V, United Engineering Foundation, New York,
Session E4, paper 1.
112. Department of Transport (now Highways Agency) DoT (1987) Use of BS 5400:Part
5:1979. London, Departmental Standard BD 16/82.
113. Moat, K. R. and Dowling, P. J. (1978) The longitudinal bending behaviour of
composite box girder bridges having incomplete interaction. Structural Engineer,
56B, No. 3, 5360.
114. Kuhlmann, U. and Ku rschner, K. (2001) Behavior of lying shear studs in reinforced
concrete slabs. In: Eligehausen, R. (ed.), Connections between Steel and Concrete.
RILEM Publications S.A.R.L., Bagneux, France, pp. 10761085.
115. Kuhlmann, U. and Ku rschner, K. (2006) Structural behavior of horizontally lying
shear studs. In: Leon, R. T. and Lange, J. (eds), Composite Construction in Steel
and Concrete V. American Society of Civil Engineers, New York, pp. 534543.
200
Index
Notes: references to beams and to columns are to composite members; cross-references to EN1992
and EN1993 are too numerous to be indexed
action eect see actions, eects of
actions 6, 8
accidental 56
arrangement of 624
combinations of 3, 6, 11, 15, 31, 624
characteristic 46
for serviceability 1646, 171
frequent 48, 64, 153
infrequent 164
quasi-permanent 48
eects of 8
de-composition of 512
envelopes of 63
global with local
and fatigue 156, 161
and serviceability 165, 1745, 1778
at failure 567, 72
in composite plates 1845
independent 1367
local 161, 165, 1834
primary 12, 48, 168
secondary 12, 48
second-order 31, 334, 103, 107, 1412,
149, 185
indirect 12, 44, 60, 138, 1701
permanent 1516
temperature 48, 120
see also fatigue load models; forces,
concentrated; loading
analysis, elastic, of cross-sections see beams;
columns; etc.
analysis, global 8, 2966
cracked 501
elastic 30, 367, 40, 4253
elasto-plastic 137
nite-element 31, 34, 39, 93, 111
rst-order 313, 38
grillage 523
non-linear 8, 36, 56, 72, 138, 185
of ller-beam decks 523
of frames 29
rigid-plastic 8
second-order 8, 314, 38, 645, 94, 1401
uncracked 501
see also cracking of concrete; loading, elastic
critical
analysis, local 1834
analysis, rigorous 39
Annex, National see National Annex
annexes, informative 2, 45, 1912
application rules 7
arches see bridges, tied-arch
assumptions in Eurocodes 7
axes 89
beams
axial force in 81, 834, 869, 105, 111,
1612
hogging 734, 779, 83
sagging 723, 83
cantilever 68, 125
Class of 12, 5760
concrete-encased 4, 52, 59, 68, 89
concrete ange of 13, 71, 109
cross-sections of 6789
Class 1 or 2 11820
and axial force 83
and ller beams 53
and global analysis 36
and indirect actions 445
and reinforcement 20
and resistance to bending 69
and serviceability 163, 170
and vertical shear 80
Class 3 37, 108
Class 3 and 4 20, 802, 103, 163
Class 4 779, 859, 94, 97, 107
classication of 29, 37, 5761, 71, 82
elastic analysis of 589, 69, 757
beams cross-sections of (continued)
plastic analysis of 36
sudden change in 1201, 186
curved in plan 4, 689, 89
curved in elevation 69, 114
exural stiness of 46
haunched 102, 117, 122, 1267, 189
of non-uniform section 4
shear connection for see shear connection
shear resistance of 67, 7983
see also analysis; buckling; cantilevers;
cracking of concrete; deections; ller
beams; ange, eective width of;
imperfections; interaction; shear . . . ;
slabs, concrete; vibration; webs
bearings 11, 15, 62, 64, 141, 145
bedding see slabs, precast concrete
bending, bi-axial 71
bending moments
accumulation of 12
and axial force 59, 68, 1023
elastic critical 93
in columns 1423
redistribution of 18, 37, 53
bolts, holes for 68
bolts, stiness of 30, 378
bond see shear connection
box girders 82
distortion of 689
shear connection for 116, 118
torsion in 456, 72
see also composite plates
bracing, lateral 356, 69, 91, 1113, 115
and buckling 97
and slip of bolts 389
stiness of 969, 1024
breadth of ange, eective see ange, eective
width of
Bridge Code (BS 5400) 1, 39, 56, 59, 70, 89,
104
bridges
cable-supported 4, 23, 40
durability of see corrosion; durability
for pedestrians 151
integral 31, 33, 36, 68, 72, 105, 115
railway 151
strengthening of 114
tied-arch 35, 1612, 189
U-frame 30
see also box girders; ller beams
British Standards
BS 5400, see Bridge Code
BS 5950 39, 59
BS 8110 70
buckling
distortional lateral 35, 91, 1057
ange-induced 68, 114
exural 77, 89, 1845
in columns 34, 136, 1402
lateral 95
lateral-torsional 31, 34, 45, 67, 767, 90104,
111, 138
local 378, 579, 76, 127, 137
see also beams, Class of
of plates 29, 378, 1845
of webs in shear 68, 80, 82
see also bending moments, elastic critical;
ller beams
cables 4, 23
camber 166
cement, hydration of 1756
CEN(Comite Europe en Normalisation) 12
Class of section see beams, cross-sections of
class, structural 267
Codes of Practice, see British Standards;
EN. . .
columns 646, 13650, 164
analysis of 29, 1403
axially loaded 144
bi-axial bending in 140, 143, 146
concrete-encased 4, 138, 145
concrete-lled 4, 14450
cross-sections of
interaction diagram for 1389
non-symmetrical 33, 47, 136
design methods for 31, 13743
eective stiness of 33, 137, 140, 149
moment-shear interaction in 139, 148
out-of-plumb 656
second-order eects in 138
shear in 139, 145
squash load of 138, 140, 147
steel contribution ratio for 136, 140, 147
transverse loading on 143
see also buckling; bending moments; cracking
of concrete; creep of concrete; length,
eective; imperfections; loading, elastic
critical; load introduction;
reinforcement; shear connection;
slenderness, relative; stresses, residual
composite action, double 4, 183, 189
composite bridges, see bridges; Bridge Code
composite plates 4, 1838
compression members 13650
see also columns
concrete
compaction of 1245
lightweight-aggregate 17, 19, 22, 136, 152
over-strength of 47, 50, 1189
partial factors for 1314
precast 278, 62
properties of 1719
spalling of 4
strength classes for 1718, 26
strength of 13, 1718, 70
stress block for 18, 701, 138
thermal expansion of 22
202
see also cracking of concrete; creep of
concrete; elasticity, modulus of;
prestress; shrinkage of concrete; slabs
connecting devices 201
connections, see joints
connector modulus, see shear connectors,
stiness of
construction 3, 478, 1034, 164, 166
loads 6, 12
methods of 12, 180
propped 121, 171
unpropped 12, 76, 913, 121
see also erection of steelwork
Construction Products Directive 14
contraexure, points of 32, 35
corrosion 25, 27
at steel-concrete interface 278, 127, 181
of reinforcement 257
cover 257, 8990, 138, 145
cracking of concrete 46, 1524, 16773
and global analysis 29, 32, 36, 467, 50,
523
and longitudinal shear 47, 118
control of 163, 173
load-induced 16970, 175
restraint-induced 1689, 1757
early thermal 1678, 1723, 1756
in columns 47, 140, 141, 145
creep coecient 19, 423, 534, 140
creep multiplier 423
creep of concrete 12, 17, 19, 32, 425, 53
in columns 45, 140, 147
secondary eects of 44
see also modular ratio; elasticity, modulus of
cross-sections see beams, cross-sections of;
columns, cross-sections of
curves, buckling resistance 34
damage, cumulative 153, 1556
see also factors, damage equivalent
damping factor 1667
denitions 8
deections see deformations
deformations 166
limits to 163
deformation, imposed 12, 44, 49, 90
design, basis of 1116
design, methods of see beams; columns; slabs; etc.
design, mixed-class 367
Designers guides v, 2
to EN1990 14
to EN19932 35, 58, 689, 72, 77, 79, 82, 95,
104, 113, 114, 118
to EN1994-1-1 60, 73, 93, 136
diaphragms 1156
dimensions 14
dispersion, angle of 113, 120
distortion of cross sections 68, 72, 82, 184
ductility see reinforcement, fracture of;
structural steels
durability 258, 179
eective length see length, eective
eective width see beams; anges; slabs,
composite
eect of action see actions, eects of
eigenvalue see loading, elastic critical
elasticity, modulus of
for concrete 18, 33, 140
for shear 45
EN1090 7, 185
EN10025 60, 72, 146
EN13670 5, 7, 180
EN13918 23, 122
EN1990 v, 2, 6, 1415, 25, 29, 48, 56, 164
EN1991 v, 2, 6, 12, 48, 151, 166
EN1992 v, 2, 5
EN1993 v, 2, 6
EN1994-1-1 v, 2
EN1994-2 v, 2
EN1998 3, 7
ENV 1994-1-1 5, 137
environmental class see exposure class
equilibrium, static 1516
erection of steelwork 7, 39
European Standard see EN. . .
examples
bending and vertical shear 10411
block connector with hoop 1168
composite beam, continuous 602, 10411
composite column 136, 14550
concrete-lled tube 14550
control of crack width 1757
cross bracing 1113
distortional lateral buckling 1058
eective width 412
elastic resistance to bending 779
fatigue 15761
in-plane shear in a concrete ange 1301
longitudinal shear 1314
lying studs 1924
modular ratios 534
plastic resistance to bending 723
resistance to bending and shear 856, 10411
with axial compression 869, 1078
serviceability stresses 1735, 1778
shear connection for box girder 1878
shrinkage eects 546
transverse reinforcement 1301
execution see construction
exposure classes 267, 1645, 167, 175
factors, combination 6, 489, 567
factors, damage equivalent 156, 161
factors, partial see partial factors
factors, reduction 923, 95
fatigue 1516, 137, 15061
analysis for 37
load models for 150, 153, 15761
of joints 30
203
INDEX
fatigue (continued)
of reinforcement 1920, 1545, 15961
of shear connectors 47, 150, 1556, 183, 191
of structural steel 27, 90, 127, 1501, 154,
157
partial factors for 14, 1512
ller beams 29, 523, 60, 8991, 166, 173
nite-element methods 68, 934, 104, 115
see also analysis, global
re, resistance to 25, 163
anges
concrete see beams; slabs
eective width of 3941, 68, 111
plastic bending resistance of 82
steel 104, 186
ow charts v
for classication of sections 57
for compression members 141
for control of cracking 1712
for global analysis 46, 626
for lateral buckling 96
forces, concentrated 114, 1201
forces, internal 137
formwork, permanent 62, 17981
formwork, re-usable 76
foundations 7, 16
fracture toughness 12
frame, inverted-U 35, 93, 958, 101, 106,
10911
frames, composite 8, 35, 646, 141
braced 47
see also analysis, global; buckling;
imperfections
geometrical data 15
see also imperfections
girders see beams; box girders
ground-structure interaction 30
see also bridges, integral
haunches see beams, haunched
Highways Agency 1
hole-in-web method 58, 5960, 73, 80
impact factor 160
imperfections 7, 14, 29, 312, 336
and lateral buckling 91
in columns 65, 138, 141, 143
in plates 185
interaction, partial and full 69
ISO standards 5, 8
italic type, use of vvi
jacking, see prestress
joints 22, 30
between precast slabs 181
stiness of 38
length, eective 34, 64, 136, 143
see also slenderness, relative
limit states
serviceability 6, 37, 56, 16378
STR (structural failure) 67
ultimate 56, 67162
loading 12
arrangement of 31
construction 12
elastic critical 313
for beams 97102
for columns 140, 1478
for composite plates 185
wheel 567, 84, 165, 183, 185
see also actions
load introduction
in columns 136, 1445, 14950
in tension members 1612
lying studs see studs, lying
materials, properties of 1723
see also concrete; steel; etc.
mesh, welded see reinforcement, welded
mesh
modular ratio 423, 534
modulus of elasticity see elasticity, modulus
of
moment of area, torsional second 46
moments see bending moments; torsion
nationally determined parameter 12, 62, 84,
91
National Annexes 1, 3, 11, 56
and actions 48, 56, 153, 1589, 164, 169
and analysis, global 56, 58
and beams 38
and columns 138
and combination factors 48
and materials 13, 20, 22, 59, 61, 70, 146
and partial factors 116, 151
and resistances 95, 124, 128, 1567
and serviceability 27, 164, 1667, 1735
and shear connectors 125, 166
national choice 2
national standards 1
NDPs 12
normative rules 23
notation see symbols
notes, in Eurocodes 1, 12
partial factors 2, 3, 916
for fatigue 14, 1512, 156
F
, for actions 6, 15, 19, 45, 51, 55
M
, for materials and resistances 1315, 92,
165, 192
plastic theory see analysis, global, rigid-plastic;
beams, cross-sections of, Class 1 and 2
plate girders see beams
plates, buckling of see buckling
plates, composite 41, 72, 126, 164, 1838
plates, orthotropic 52
Poissons ratio 46, 144
204
prestress 4, 8
by jacking at supports 4, 8, 12, 49
by tendons 49, 77, 155, 167
transverse 4
principles 4, 7, 12
propping see construction, methods of
provisions, general 2
push tests see shear connectors, tests on
quality, control of 26
redistribution see bending moments; shear,
longitudinal
references, normative 57
reference standards 3, 57
regulatory bodies 12
reinforcement 9, 1921, 22
and lying studs 1901
ductility of 20, 59, 71
fracture of 212, 59
in beams
for crack control 16773
for shrinkage 19
minimum area of 59, 1689, 1767
transverse 115, 124, 12730
in composite plates 186
in compression 70
in ller-beam decks 90
in haunches see beams, haunched
yielding of 39
welded mesh (fabric) 1921, 59, 71
see also cover; fatigue
resistances 14
see also beams, bending resistance of; etc.
restraints, lateral see bracing, lateral
rotation capacity 12, 22, 58
see also joints
safety factors see partial factors
scope of EN1994-2 45, 36, 136, 138
section modulus 8
sections see beams; columns; etc.
separation 8, 115, 124
serviceability see limit states
settlement 12, 30
shakedown 1534
shear see columns, shear in; shear, longitudinal;
shear, vertical; etc.
shear connection 2, 8, 68, 11435
and execution 1245
and U-frame action 93, 111
by adhesives 23, 114
by bond or friction 23, 114, 136, 144, 14950,
190
design of 82, 155
detailing of 1247, 180, 18993
for box girders 1858
full or partial 69
see also fatigue; load introduction;
reinforcement, in beams, transverse;
shear connectors; slip, longitudinal
shear connectors 115, 156
and splitting see studs, lying
angle 5, 115
bi-axial loading of 185, 188
block with hoop 5, 114, 1167
channel 5
ductility of 114
fatigue strength of 151, 191
exibility of see stiness of
force limits for 151, 1656
in young concrete 166
partial factors for 14
perforated plate 22
spacing of 59, 91, 119, 1245, 162, 180, 1845,
1878
stiness of 18, 69, 164, 1867
tension in 115
tests on 5, 189
types of 5, 22
see also studs, welded
shear ow 116, 118
shear lag see width, eective
shear, longitudinal 47, 68, 114, 11821, 12730
see also columns, shear in; composite plates;
shear connection; shear ow
shear, punching 84, 184
shear ratio 100
shear, vertical 29
and bending moment 803, 87
and lying studs 191, 193
in deck slabs 84, 184
in ller-beam decks 91
see also buckling
shrinkage of concrete 19, 53
and cracking 16970, 1745
autogenous 19, 55, 144, 147
eects of 35, 45, 1656, 172
in tension members 50
modied by creep 434, 546
primary 12, 546, 76, 120, 1334
secondary 12, 134
situations, design 165
skew 89
slabs, concrete 53
reinforcement in 125
splitting in 115, 1234, 18993
see also plates, composite
slabs, precast concrete 115, 1257, 17981
slenderness, relative
for beams 92, 94, 97, 1012
for columns 136, 140, 1478
slip capacity 114
slip, longitudinal 8, 389, 69, 166
in columns 138, 144
in composite plates 184, 1867
software for EN1994 323, 75, 94
205
INDEX
splices 38, 118
squash load see columns, squash load of
stability see equilibrium, static
standards see British Standards; EN. . .
standards, harmonised 14
steel see reinforcing steel; structural steel;
yielding of steel
steel contribution ratio see columns
steelwork, protection of see durability
stieners, longitudinal 59, 82, 185
stieners, transverse web 35, 689, 93, 97, 116
stiness, eective, of reinforcement 51
stiness, exural see beams; columns; etc.
stiness, torsional 456
strength of a material 1314
characteristic 13, 15
see also resistance
stress block for concrete 13
stresses
accumulation of 12, 81, 84, 86
bearing 144
design, at serviceability limit state 1646
in concrete 26, 165, 1778
in reinforcement 165
in steel 165, 177
equivalent, in steel 72, 185
excessive 164
fatigue 1501
mid-plane, in steel 767
residual, in steel 31, 345, 141
shrinkage see shrinkage of concrete
temperature see temperature, eects of
stress range, damage equivalent 1557
stress resultant see actions, eects of
structural steels 202
partial factors for 9, 1415
thermal expansion of 22
strut, pin-ended 32
studs, lying 115, 1234, 18993
studs, welded 3, 223, 1213
detailing of 127, 18993
ductility of 120
length after welding 122
resistance of 1212, 189, 191
tension in 123
weld collar of 23, 115, 1223
see also fatigue; shear connection, detailing
of; shear connectors
subscripts 89
superposition, principle of 31
support, lateral see bracing, lateral
sway 142
symbols 89, 1314, 20
temperature, eects of 489, 53, 153
temporary structures 11
tendons see prestress
tension eld 7980
tension members 23, 4952, 1612
tension stiening
and cracking 47, 16970
and longitudinal shear 1189, 162
and stresses 1545, 160, 165, 1735
and tension members 501
testing see shear connectors
tolerances 14, 49, 1801
torsion 456, 68, 72, 91, 163, 184
trac, road, type of 160
truss analogy 128
trusses, members in 49, 1367, 140
and buckling 95
and eective widths 41
tubes, steel see columns, concrete-lled
U-frame see frame, inverted-U
units 189
uplift see separation
variables, basic 12
vibration 163, 1667
warping, resistance to 72, 118, 163, 184
webs
breathing of 166
eective area of 378
holes in 68, 90
transverse forces on 1134
see also hole-in-web method; shear,
vertical
web stieners see stieners, transverse web
width, eective 29, 1834
see also beams; anges, eective width of
worked examples see examples
yielding of steel 37
206

Designers' Guide To en 1994-2 - Eurocode 4 - Design of Composite Steel and Concrete Structures

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Designers' Guide To en 1994-2 - Eurocode 4 - Design of Composite Steel and Concrete Structures

Uploaded by

Copyright:

Available Formats

Delivered by ICEVirtualLibrary.

steel section: 406 140 UB39 with A

= 0.78, if the height of the stud were reduced to,

Fig. 6.36. Flow chart for verification of a column length

Table 6.2, for characteristic loads per unit length

bending moment in the connection in span BC: 67.7 kN m, which is 82%of M

maximum deflection of span BC: 13.8 mm.

bending moment at B in span BC: 70.2 kN m

maximum deflection of span BC: 17.9 mm.

in regions where the beam resists hogging bending, as determined by resistance to

in other regions, as determined by resistance to longitudinal shear: A

whether construction will be unpropped or propped

whether the spans are modelled as simply supported or continuous.

hogging length of beam: = 0.15 12 = 1.80 m

distance of bars above centre of compression: h

distance of bars above centroid of steel beam: d

second moment of area of steel beam: 10

is also obtained from EN 1992.

From Fig. 5.10, the strain at point E is:

which is equation (5.6-1).

0.935 0.055 3 0.768

0.67 355 1175 12.5

cycles (2 million cycles for

for straight bars is 10

Results are given in Table 9.1.

You might also like