Railway Bridges 2

P318: Design Guide for Steel Railway Bridges
Created on 27 March 2006

This material is copyright - all rights reserved. Use of this document is subject to the terms and conditions of the Steelbiz Licence Agreement
SCI PUBLICATION P318
Design Guide for Steel

Railway Bridges
D C ILES MSc ACGI DIC CEng MICE
Published by:
The Steel Construction Institute
Silwood Park
Ascot
Berkshire SL5 7QN
Tel:
Fax:
01344 623345
01344 622944

2004 The Steel Construction Institute

Apart from any fair dealing for the purposes of research or private study or criticism or review, as permitted under the
Copyright Designs and Patents Act, 1988, this publication may not be reproduced, stored or transmitted, in any form or by
any means, without the prior permission in writing of the publishers, or in the case of reprographic reproduction only in
accordance with the terms of the licences issued by the UK Copyright Licensing Agency, or in accordance with the terms
of licences issued by the appropriate Reproduction Rights Organisation outside the UK.
Enquiries concerning reproduction outside the terms stated here should be sent to the publishers, The Steel Construction
Institute, at the address given on the title page.
Although care has been taken to ensure, to the best of our knowledge, that all data and information contained herein are
accurate to the extent that they relate to either matters of fact or accepted practice or matters of opinion at the time of
publication, The Steel Construction Institute, the authors and the reviewers assume no responsibility for any errors in or
misinterpretations of such data and/or information or any loss or damage arising from or related to their use.
Publications supplied to the Members of the Institute at a discount are not for resale by them.
Publication Number: SCI P318
ISBN 1 85942 150 4
British Library Cataloguing-in-Publication Data.
A catalogue record for this book is available from the British Library.
P:\PUB\PUB800\SIGN_OFF\P318\P318v01D21.doc
ii
FOREWORD
This publication has been prepared to provide a comprehensive yet relatively concise
introduction to the design of steel and composite railway bridges, with particular
reference to design in accordance with BS 5400 Design of steel concrete and composite
bridges.
The publication has been prepared by Mr D C Iles (The Steel Construction Institute),
based on a number of substantial contributions from practising senior engineers and with
the assistance of an Editorial Advisory Group.
The main contributions were made by the following companies:

Atkins Rail
Corus Rail Consultancy
Scott Wilson Railways
Network Rail
Fairfield Mabey Ltd
Guidance in drafting Sections 6, 7 and 8 was provided by Mott MacDonald. The worked
example of a U-frame bridge is based on an example provided by Cass Hayward and
Partners.
The SCI is very grateful to these companies for their contributions.
The Editorial Advisory Group comprised the following members:
Mr
Mr
Mr
Mr
Mr
J Bleasdale
I K Bucknall
A C G Hayward
W Heggie
R Molloy
Metronet Rail SSL Ltd

Network Rail
Cass Hayward and Partners
Scott Wilson Railways
Atkins Rail
The SCI wishes to express thanks to the members of the editorial group for their
assistance and advice in the preparation of this publication. Thanks are also expressed to
Mr D Merrett, Corus Rail Consultancy, who provided valuable comment in the latter
stages of drafting.
The work leading to this publication was funded by Corus Construction and Industrial
and by Network Rail.
iii
CONTENTS

Page No.
FOREWORD
iii
SUMMARY
vii
INTRODUCTION
FUNCTIONAL REQUIREMENTS
2.1
Key functional requirements
2.2
The railway infrastructure
2.3
Strength and fatigue endurance
2.4
Bridge deformation
2.5
Robustness
2.6
Durability
2.7
Clearances from the railway
2.8
Underline clearances
3
3
3
9
10
11
12
13
21
CONSTRUCTION REQUIREMENTS
3.1
General
3.2
Minimising disruption to the operational railway
3.3
General site constraints
3.4
Third party issues
3.5
Methods of bridge erection
3.6
Replacement of the permanent way
3.7
Buildability
22
22
22
24
25
25
30
30
FORMS OF CONSTRUCTION
4.1
Influences on form of construction
4.2
Shallow deck-type bridges
4.3
Half through plate girder bridges
4.4
Standard box girder bridges
4.5
Slab-on-beam composite girder bridges
4.6
Truss girder bridges
4.7
Direct fastening construction
4.8
Integral construction
33
33
34
36
42
44
45
46
47
DESIGN STANDARDS
5.1
HMRI principles and guidance
5.2
Railway standards
5.3
Technical approval
5.4
Particular railway standards
5.5
BS 5400
5.6
Design Manual for Roads and Bridges
5.7
Matters not covered by standards
5.8
Development of Standards
49
49
49
50
50
51
53
53
54
DESIGN PROCEDURES
6.1
General
6.2
Design basis
6.3
Loading
6.4
Initial design
6.5
Global analysis
55
55
56
57
63
66

6.6
6.7
Detailed design
Design of half through bridges
70
70
DESIGN FOR STRENGTH

7.1
General
7.2
Design resistance of beams without longitudinal stiffeners
7.3
Filler beams
7.4
Restraint systems
7.5
Material strength and notch toughness
7.6
Slab design
7.7
Interfaces with substructures
7.8
Bearings
72
72
78
82
83
85
87
88
88
DESIGN FOR FATIGUE ENDURANCE

8.1
General
8.2
Methods of assessment
8.3
Fatigue loading
8.4
Classification of details
8.5
Limiting stress range
8.6
Considerations for particular details
90
90
90
91
91
92
92
DESIGN FOR DEFORMATION PERFORMANCE

9.1
General
9.2
Deflection under permanent loads
9.3
Deformation due to rail loading
9.4
Deformation due to rail loading and temperature
9.5
Lateral deformation
9.6
Effect of deformation on clearances
95
95
95
95
97
98
98
10
BRIDGE DETAILS
10.1 Shallow deck-type bridges
10.2 Half through plate girder bridges
10.3 Standard box girder bridges
10.4 Slab-on-beam composite girder bridges
10.5 Bearing details
10.6 Deck end waterproofing details
10.7 Footways
99
99
100
116
116
116
117
118
11
REFERENCES
120
Appendix A
STEEL BRIDGE GROUP GUIDANCE NOTES
125
Appendix B
WORKED EXAMPLE
127
vi
SUMMARY
This guide offers an introduction to the design of steel and composite railway bridges. It
presents a review of the particular requirements for carrying railway traffic and
interfacing with railway infrastructure and, since construction issues are particularly acute
for replacement bridges, describes the constraints and options for construction. Typical
forms of construction for short to medium span bridges are described and simplified
cross sectional arrangements are illustrated. A key consideration for the selection of
railway bridges is a safe and reliable structure that can be constructed and maintained
with minimal disruption to railway passengers and traffic steel railway bridges meet
these requirements particularly well.
The guide explains that design of railway bridges in the UK is governed not only by
BS 5400 but also by the comprehensive additional requirements of the railway authorities,
notably Network Rail. Design procedures are described and detailed aspects of design
for strength, for fatigue endurance and for deformation performance are discussed. A
range of typical details is illustrated, with comments about the factors that need to be
considered when selecting and designing them.

A worked example of a single span half through bridge with a composite deck carrying
two tracks is presented. It illustrates the main features in determining the appropriate
bridge details and in verifying the adequacy of the bridge.
vii
INTRODUCTION
Steel has been used in the construction of railway bridges since Victorian times,
both in major long-span structures and in more modest spans, such as over local
highways. Many of these older steel bridges are still in service, thanks to an
on-going programme of maintenance, refurbishment and upgrading to suit
changing requirements. Today, most railway bridges are built as replacements
for earlier structures, although some entirely new structures are being built on
new alignments or routes, most notably for the Channel Tunnel Rail Link.
For replacement bridges, steel construction can achieve shallow construction
depths, which is essential where the track level is fixed yet sufficient underline
clearance is needed for highways or other services below the bridge. Steel
construction is suited to prefabrication and preassembly and, because of its
relatively modest self weight, can be transported or lifted into position in the
very short periods during which the railway can be closed to traffic.

For bridges on new alignments, where there is greater freedom with the vertical
profile of the route, greater construction depths can be used and this opens up
the possibility of the use of slab-on-beam composite construction, in a form
similar to that favoured for highway bridges.
This guide offers an introduction to the design of steel and composite railway
bridges, for readers who are generally familiar with the principles of limit state
design. Some experience of bridge design, perhaps of the design of highway
bridges, would be advantageous in understanding the detailed design procedures.
The guide begins with a review of the particular requirements for carrying
railway traffic and interfacing with railway infrastructure. Since construction
issues are particularly acute for replacement bridges, because of the need to
minimise disruption to users of the railway, guidance is given on the constraints
and options for construction.
So-called standard bridge types were developed by British Rail from the 1950s
onwards and although there is currently only one true standard bridge type in
common use, features of many of the others are still appropriate. Section 4
presents a brief review of all the main forms of construction.
Whilst BS 5400 is currently the principal structural design standard, there are
supplementary requirements that have been developed especially for railway
bridges. A wide range of other standards that relate to track formation,
clearances to railway traffic, etc. have been issued by Her Majestys Railway
Inspectorate, Network Rail and London Underground. Section 5 catalogues
these standards.
Design procedures are described in Section 6; there are three principal criteria
for the design to meet adequacy of fatigue life (railway bridges are subject to
particularly onerous fatigue loading), adequacy of strength (i.e. at the Ultimate
Limit State), and adequacy of the deformation response (which can affect the
safety of rail traffic and the comfort of passengers). These procedures are
described in Sections 7, 8 and 9.
Success or failure of a design often depends on the quality of the detailing and
durability often depends on the provision of features that allow proper
inspection and maintenance. Section 10 provides a collection of good practice
details for the readers guidance.
Finally, Appendix B presents a worked example of design calculations for a
typical half through two-track railway bridge. The example illustrates many
aspects of the design procedures described in the earlier Sections and should
prove a valuable insight into the application of BS 5400 to the design of a steel
railway bridge.
Most of the detail in the guide relates to (standard gauge) mainline railways and
thus to the standard mainline loading adopted throughout Europe since 1973,
following the demise of steam traction. The principles are, however, equally
applicable to light rail applications, although the different balance between
static and live loading for such structures will affect design.

The guidance in this publication is intended to assist designers with the design
of conventional rail bridges. Unusual forms of construction and temporary steel
bridges are outside the scope of this publication
For a detailed commentary on BS 5400-3:2000, see the separate SCI publication
Commentary on BS 5400-3:2000, Code of practice for the design of steel
bridges[46]. References are made in the text to further general advice in
Guidance Notes. These are a series of notes, published by The Steel
Construction Institute[47], that give concise advice on best practice in steel bridge
design, from the members of the Steel Bridge Group, a technical group of
experienced designers, fabricators and clients. References to specific Notes in
that document are given in the form GN 1.02. A full list of Guidance Notes
is given in Appendix A.
FUNCTIONAL REQUIREMENTS
2.1
Key functional requirements
There are two key functional requirements for a bridge carrying a railway:
i)
Provision of satisfactory support to the railway traffic and infrastructure

throughout the life of the bridge.
ii)
Provision of adequate clearances between the structure and the traffic on

and beneath the bridge.
The first requirement can be expressed in terms of requirements for:
Limiting the bridge deformation
Robustness
Durability

The second requirement is expressed in terms of various clearance gauges that

are defined by the railway and highway authorities.
To verify that the requirements continue to be met throughout the life of the
structure (i.e. to ensure on-going serviceability), there is also a need to make
provision for access to inspect and maintain the elements of the structure in a
safe and convenient way.
To aid the appreciation of the functional requirements, Section 2.2 presents a
simple overview of the nature of the railway infrastructure. The requirements
for strength/fatigue endurance, deformation. robustness, durability and provision
of adequate clearances are discussed in general terms in Sections 2.3 to 2.8.
2.2
The railway infrastructure
The railway infrastructure comprises the permanent way, which carries the
railway traffic, access ways beside the track, and the associated plant,
equipment and services that enable the railway to function.
2.2.1 Permanent way (track)

Permanent way, generally referred to as track, can be defined as comprising the
rails, sleepers and ballast that provide support and guidance to rail traffic.
The great majority of the track on UK railways is of the conventional ballasted,
cross-sleepered type. The rails are fixed to transverse sleepers (of either
timber, steel or pre-stressed concrete), which are set in a bed of crushed stone
ballast. New or replacement bridges are usually designed to accommodate this
type of track, and the weight of the ballast usually adds substantially to the
superimposed dead load. On many older steel bridges, the rails are fixed to
longitudinal timbers, which are fixed directly to the bridge structure without any
ballast. However, such construction details are now rarely used for new bridges
on main line railways because of track maintenance problems, especially the
run-on/run-off effects and the lack of flexibility in track location. See page 1
for further discussion on direct fastening.
The standard gauge (i.e. the width between the inner faces or running edges
of the rails) on the UK national railway network is nominally 1435 mm
(4ft 8 in). Since the railhead is typically 70 mm wide, the centrelines of the
rails are 1505 mm apart, usually taken as 1.5 m when applying wheel loads in
design calculations. Where there are two tracks, the minimum distance between
the centrelines of the tracks is 3400 mm (see RT/CE/S/049[16], sheets A8.2 and
A8.2). The space between tracks is commonly referred to as the six-foot.
Rails
In the UK, until the 1950s, rails were mostly of the bullhead type, weighing
typically 95 lb/yd. These are still widely found on London Underground, but
are being replaced by the flat-bottom type. On the national network, bullhead
rails have now been almost entirely superseded by flat bottom rails, weighing
typically 113 lb/yd (56 kg/m). Heavier and deeper RT60 rails (also known as
CEN60 or UIC60 rails), weighing 60 kg/m, have recently been introduced.
For new work, London Underground generally uses 113 lb/yd flat bottom rail
running rails, plus third and fourth current rails of 150 lb/yd maximum.

Continuous welded rail

In the past, rails were laid with bolted fish-plated joints, typically at 18.3 m
(60 ft) intervals. Whilst some jointed track remains, all main lines and most
secondary routes are now laid with continuous welded rail (CWR), which has
no fish-plated joints. The rails are pre-tensioned on installation so that they are
stress free at a nominal temperature of 27C. Any tendency for the track to
buckle, owing to compressive stresses in the rails at higher temperatures, is
resisted by the ballast and the weight of the track. Substantial shoulders of
ballast are provided at the ends of the sleepers to ensure stability. Modern
points and crossings also incorporate welded rails and are reinforced to
accommodate the thermal stresses from the adjoining CWR plain line.
Where a length of CWR adjoins jointed track, special expansion joints (known
as adjustment switches, expansion switches or breather switches) are
provided. Similar joints are provided at the ends of longer bridges to
accommodate thermal expansion/contraction of the bridge. On shorter bridges,
no expansion joints are provided in the rails and the ballast accommodates any
slight differential movement between track and bridge. The presence, or
otherwise, of joints in the track will affect the way longitudinal traction and
braking forces are transferred to the bridge structure (see Section 6.5.8).
On London Underground, the stress-free temperature for CWR is specified in
Engineering Standard TE MTS 0314[29] as within the range 21C to 27C.
Sleepers and base-plates
Most sleepers are of pre-stressed concrete, which is preferred for heavily used
high-speed routes, but some timber sleepers are in use, together with increasing
numbers of steel sleepers. The rails are clipped directly onto sleepers, with a
resilient elastomeric pad under the rail foot to prevent abrasion. On timber
sleepers, the rail is clipped to a cast iron base-plate that is screwed or spiked to
the sleeper. Standard concrete sleepers are deeper than timber or steel
alternatives, but special shallow depth types are available for use on Network
Rail bridges where construction depth is very limited. (These shallow depth
concrete sleepers are not available for London Underground tracks.)
Ballast
Ballast usually consists of hard angular pieces of crushed stone of about 50
65 mm size. It provides a resilient bed for the sleeper, distributes the load onto
the bridge deck, and allows for drainage. It also provides a means of adjusting
the track alignment and level. Normal practice is to provide at least 300 mm of
ballast under the sleeper (230 mm for London Underground track), but it is
sometimes necessary to provide less than this where construction depth* is
limited. A minimum of 200 mm depth of ballast is necessary to prevent
damage to the bridge waterproofing by track maintenance machines and to
ensure the satisfactory distribution of wheel loads. (This minimum dimension is
recommended in GC/RT5510, clause 6.3.6, regardless of sleeper type, although
RT/CE/S/102[18] allows 150 mm below steel sleepers for ordinary track
construction.) However, where construction depth is severely restricted, ballast
depths of 150 mm (minimum) may be preferable to direct fastening, provided
that the effect on load distribution is considered and appropriate precautions are
taken to protect the waterproofing.

Where track is canted, the depth of ballast below the sleeper is usually
measured under the low end of the sleeper but it may be acceptable in some
circumstances to measure it under the low rail.
During normal track maintenance, the ballast is mechanically compacted under
the sleeper. The compaction is not uniform under the length of the sleeper but
is concentrated in the areas under the rails. The rules for distribution of load
through the ballast[40] take this into account.
Track types, track depth
Normally the type of track to be provided, over a new or reconstructed bridge,
will be specified by the owner of the infrastructure. Guidance on the types of
track appropriate to different categories of lines on the national network is given
in RT/CE/S/102[18].
For London Underground track types, refer to Engineering Standard E8001[25].
A summary of track types and their depths is given in Table 2.1.
Details of track components may be found in the Track design handbook[16].
Construction depth is measured from tops of rails to bridge soffit, see Section 4.1.
Table 2.1
Types of track commonly found on the UK rail network
Track type
Track depth*
Track for Network Rail

CEN 60 FB rail on G44 concrete sleeper
377 mm
CEN 60 FB rail on type 600 steel sleeper
310 mm
BS113A FB rail on F27 or F40 concrete sleeper
368 mm
BS113A on EF29 shallow concrete sleeper
330 mm
BS113A on timber sleeper with baseplate
330 mm
BS113A on type 500 or 560D steel sleeper
281 mm
Track for London Underground

BS113A rail on NTF 504 concrete sleeper
329 mm
Bull Head rail on timber sleeper
315 mm
* Track depth is measured from top of rail to underside of sleeper
Ballast retention

Bridges carrying ballasted track are constructed in the form of a trough with the
sides raised sufficiently to contain the ballast, with some allowance for future
track-raising during maintenance. Where tracks are canted, bridge structures
are sometimes tilted transversely to avoid excessive ballast depths. However,
this should be limited, as a tilted deck is likely to encourage migration of the
ballast. A limit of 1 in 15 is usually observed for ordinary mainline traffic.
It should be noted that track maintenance methods tend to raise the track over
time, since it is far easier to adjust the longitudinal profile by raising the dips
than by lowering high spots. Furthermore, the tracks in many places were
raised when CWR was introduced, to improve drainage, and wider ballast
shoulders were provided. For these reasons many existing bridges have
inadequate containment of the ballast (to meet requirements for new
construction), particularly on the approaches at either end, where the original
formation may be too narrow or the wing walls too low to support a ballast
shoulder and cess walkway to current construction standards.
In designing new or replacement bridges, care should be taken to contain and
support the ballast, particularly at the ends of the deck. Three-dimensional
sketches or drawings will help in visualising the often-complex geometry and
will also help in detailing the waterproofing, drainage, cess walkways, cable
runs etc. in these areas.
Direct fastening
As discussed above, wherever possible new or replacement bridges are designed
to accommodate conventional cross-sleepered, ballasted track. However, in
order to achieve the minimum possible construction depth, or to minimise dead
weight of a bridge, it may be necessary in extreme cases to omit the sleepers
and ballast and to fasten the rails directly to the bridge deck. This is known as
direct fastening. Particular considerations are:
It is difficult in practice to arrange and maintain a satisfactory transition

from the directly fastened track to the ballasted track at the ends of the
bridge. This is particularly true on high-speed routes, owing to the small
permitted tolerances on track geometry, or where there is horizontal or
vertical curvature.
Because track maintenance is highly mechanised, the introduction of short

lengths of non-standard track, which require hand maintenance, increases
costs.
Successful installation and maintenance of a direct fastening system requires

a high degree of care and precision that may not be achievable in the
limited time and difficult conditions of a railway possession (for a
discussion of possessions, see Section 3.2.2.
For bridges with direct fastened track, consideration of noise levels may be
required; special precautions can be taken if necessary. (See Reducing
noise emission from steel railway bridges[44] and GN 1.06.
The reduced dead load, compared to a ballasted bridge, may have an

adverse effect on the dynamic response of the bridge, particularly for short
spans, and a check should be made that the natural frequency is within the
prescribed limits. Very light bridges can also be prone to hammering of
the bearings under load, and are less resistant to accidental vehicle impact
where they span over highways.

In general, a reduced ballast depth is often preferable to direct fastening from

the point of view of track maintenance.
Recent strain gauge tests on rails in service indicate that local deck elements of
direct fastening bridges need to be designed for much higher wheel loads than is
the case for ballasted track, to take account of the effects of wheel flats (because
there is no ballast to cushion such local effects).
Special baseplates are available that have been designed for use in direct
fastening applications. These provide for limited lateral and vertical adjustment
of the rail and also incorporate additional resilience to control bending stresses
in the rail. Some types include noise-insulating material. Specialist advice
should be sought from the railway authority for the design of transition
arrangements at either end of the bridge.
The comments above are applicable to new or reconstructed bridges,
particularly short span bridges, on the existing UK national rail network. Some
of the disadvantages of direct fastening do not apply to completely new
railways, or to rapid transit systems, and direct fastening is increasingly being
used worldwide as new developments are made. Obviously, direct fastening is
appropriate where slab track is in use instead of conventional ballasted track.
(Slab track, consisting of a jointless reinforced concrete slab, to which the rails
are directly fastened or are bedded into grooves, has been used in some
countries but has not yet been used widely in the UK, except for tramways.)
2.2.2 Access along the railway

For the UK national rail network, document GC/RT5203[8] gives requirements
for two types of access along the railway cess walkways for general access
by permitted people and continuous positions of safety immediately adjacent to
the track, where people who have access to the line can seek refuge when a
train comes. Minimum dimensions and clearances from the track are specified
in the document. As a minimum, there should be a continuous position of
safety on each side. A cess walkway (which is wider) can be used as a
continuous position of safety, provided that the necessary clearances are
observed; ideally such a walkway should be provided on both sides but it may
be acceptable to have a cess walkway on one side and a continuous position of
safety on the other. The railway authority will determine what is acceptable for
any particular bridge.
2.2.3 Other infrastructure

Apart from the permanent way and access ways beside the track (see
Section 2.7.1), the elements of infrastructure that may need to be supported
include:
Traction power systems
Signalling cables and equipment
Telecommunications cables and equipment
Power cables
Third party cables and pipes
Mechanical and electrical equipment and plant.

In some cases, these elements may influence the structural form that is selected
for a particular bridge.
Prior to detailed design work commencing, a survey needs to be undertaken by
competent persons to identify the equipment present and develop proposals for
its temporary and/or permanent relocation, if affected. These proposals may
require additional possessions in advance of the main installation to minimise
the work and the risk of unforeseen difficulties during the main installation
possession.
Traction power systems
Traction power systems generally fall into two types:
Overhead line electrification (OLE)
Third and fourth rail systems, which are predominantly used in the southeastern part of the UK and on London Underground.
Slewing or dismantling of existing OLE equipment should be avoided where

possible and avoidance of such activities can have significant influence on the
form of construction.
Steel bridges need to be electrically bonded to the traction current return rail or
earth wire. Individual components may need to be electrically bonded together,
if they are not adequately connected by bolting or welding. Longer bridges may
carry masts supporting overhead electrification wires and the bridge will need to
be designed to accommodate the forces imposed, including consideration of
unbalanced forces if cables should break. Specialist advice should be sought as
to how OLE supporting structures can either be relocated off the new structure
or incorporated into the design.
The third and fourth rail systems have little effect on the design of the bridge,
other than a slight increase in width in some cases, to provide clearance
between electrical rails and structure.
Cables
There is often significant cabling infrastructure running parallel to the tracks.
The cables can be for signalling, telecommunications, power, third parties and
other M&E functions. The cables are usually located in concrete toughing on
UK main line railways and on brackets connected to parapet steelwork on
London Underground. The cables can vary in age and condition; the latter will
dictate how they are managed. It is quite common to be able to move the
cables on to an adjacent temporary cable bridge, should there be adequate slack
in the cables. After the bridgeworks, the cables are moved back, and tested
before possession hand-back. Occasionally it is necessary to replace a section
of cable in advance of the works, if its condition will not permit it to be moved
or if there is insufficient slack for the cable to be relocated to its temporary
and/or new alignment. Access to cables is controlled in a similar manner to
that for the railway as a whole, and it is necessary to plan work well in
advance, so that any permitted outages for the cable services (i.e. when the
service is interrupted) can align with the available enabling works possessions.

Other equipment
Other equipment such as signals and equipment cabinets are not usually located
on or near structures. If they are, then specialist advice should be sought. In
the case of a new bridge, if possible the site of the structure should be selected
so as not to impact on these infrastructure elements, as these will increase the
disruption to the railway by increasing the enabling works and requiring
additional interfaces to be managed.
2.2.4 Services ducts

Provided that it does not obstruct the walkways, the usual precast concrete cable
troughing can be continued over the bridge. Otherwise a special conduit may
be incorporated in the structure, under or adjacent to the walkway, with
removable access covers, and provision for drainage. Some thought should be
given to the detailing at the ends of the bridge to ensure a suitable alignment
and support to the cables (ensuring that any bends are large radius) and to
provide a clear walking route without tripping hazards.
On London Underground, it is not usual for cables to be placed in concrete
troughing. London Underground normally requires the provision of permanent
cable brackets and supports on bridges.
2.3
Clearly, a bridge must have sufficient strength to carry the expected loading
upon it. The loading depends on the type of traffic, speed of the trains and the
number of tracks carried by the bridge. The loading and internal design forces
for the structural members also depend on the geometry of the track (whether
straight or on a curve) and of the bridge (particularly when the bridge is skew).
The loading on railway bridges is significantly affected by the dynamic effects
of traffic actions, both due to the nature of the traffic and to the response of the
structure. Particular aspects that give rise to dynamic effects are:
a) The rapid rate of loading, due to the speed of the traffic coming onto the
structure, and the inertial response of the structure.
b)
The passage of successive axles or bogies, with approximately uniform

spacing, which can excite the structure and, under certain circumstances,
can create resonance.
c)
Variations in wheel loads resulting from track or vehicle imperfections

(including wheel irregularities).
d)
Centrifugal effects on curved track.
e)
Lateral wheel/rail contact forces.
f)
Lurching of rail vehicles, which transfers some of the vertical load from
one rail to the other.
Bridges must also be able to endure the repeated application of loading

throughout the design life they must be designed for fatigue loading. For
railway bridges, fatigue loading is particularly onerous and may well govern
many aspects of detail, the size of the principal members and even the form of
construction. The fatigue loading depends, in addition to the above factors, on
the annual tonnage of traffic on the line.

Requirements for strength and fatigue endurance are both ultimate limit state
considerations. Guidance on structural design is given in Sections 6, 7 and 8.
Guidance on loading, including the manner of allowing for dynamic effects in
standard loading, is given in Section 6.3.4.
2.4
Bridge deformation
The following is a general discussion of functional requirements.

limits on deformations are discussed in Section 9.
Detailed
2.4.1 General
An allowance should be made for deformations and any pre-camber in the
bridge deck when calculating clearances, headroom and construction depth.
Detailed requirements for deformation limits are given in GC/RC5510 and UIC
Leaflet 776-3R[35]. The limits relate principally to vertical deflection under live
load. There is an overall requirement that total deflection does not encroach on
any clearances.
2.4.2 Dead load deflection

Consideration should be given to pre-cambering main girders to counteract dead
load deflections (and to counter the optical illusion that a flat soffit appears to
be sagging). However, on short spans this can cause the fabricator more work
than is justified by the marginal aesthetic improvement. Recommendations are
given in GC/RC5510.
Dead load deflection is an indicator of the natural frequency. Light, stiff
bridges (high frequency) or heavy flexible spans (low frequency) can fall outside
the limits of applicability of the design rules. Documents GC/RC5510 and UIC
776-3R give upper and lower limits on dead load deflection. It is a valuable
simple check to evaluate dead load deflection at an early stage and thus ensure
that the dynamic live load effects will be adequately covered by the dynamic
factors included in the loading specification (see Section 6.3.4).
10
2.4.3 Live load deflection

Limits on live load deflections are set in GC/RC5510 and UIC 776-3R[35] to
ensure the safety of rail traffic and passenger comfort, to control rail forces and
to ensure the track geometry remains within the specified limits. The limits
imposed are onerous, particularly on high-speed lines and on multi-span noncontinuous bridges , and should be checked at an early stage in the design.
For London Underground bridges, maximum train speeds are less than
100 km/hr and train weights are less; consequently, deformation limits are less
critical.

Experience indicates that satisfying deformation criteria can dictate either

structural form or the size of main structural members. Deformation limits that
may have to be taken into account in the design relate to:
Maximum vertical deflection in a span
Twist (rate of change of cant along the track)
Uplift at the end of a deck (kick-up)
Change of angle at the end of a deck
Horizontal deflection of a deck (change of angle at the end of the deck and
lateral displacement in the span)
Vertical deck acceleration (on high speed lines).
2.5
Robustness
Redundancy
Bridges should be designed such that they have sufficient robustness not to
suffer damage due to accidental events or vandalism, to an extent that would be
disproportionate to the severity of the cause.
Consideration should also be given to incorporating structural redundancy within
the design, so that alternative load paths are available in the event of unforeseen
failure of part of the structure. Generally, the structure should be designed so
that the critical failure modes are those that give advance warning of failure
(e.g. bending rather than shear).
General detailing
Railway bridges should be designed to withstand normal track maintenance,
relaying and reballasting operations without damage. Waterproofing should be
protected from ballast abrasion and damage by rail-mounted ballast tamping
machines, tampers or hand tools by protective layers and adequate ballast
depths. Girder flanges and any stiffeners adjacent to the track should be
sufficiently robust to avoid damage during relaying operations or minor
derailments. For this reason, and also to simplify waterproofing and inspection,
web stiffeners of half through bridges are usually placed on the outside of the
bridge.
Derailed trains
To comply with HMRI requirements[49], railway underline bridges should be
provided with a robust kerb to contain derailed trains (see also GC/RC5510
Clause 7.3.3). For half through bridges, the main girders may be deemed to
11
act as the robust kerb, provided that they are at least 300 mm above rail level,
otherwise a separate upstand must be provided to fulfil this function.
GC/RC5510 recommends that the kerb should preferably be set at least
1500 mm from the adjacent rail, although it is accepted that this is impractical
in many cases, particularly in reconstruction of existing bridges. (GC/RC5510
also recommends that kerbs should be preferably 350 mm high, to allow for
possible future track lifting.)
The most common type of derailment is where only a few wheelsets in a train
are off the rails, each with one line of wheels between the rails and one running
on the sleeper ends, and with the derailed vehicles upright and in line with the
rest of the train. Such a partially derailed train can run for a considerable
distance before it is brought to a stand and underline bridges should not
introduce obstructions that would make the consequences of derailment worse.
The area next to the track needs to be able to carry loads from derailed trains
(see Section 6.3.5). Consequently, open type flooring is no longer acceptable
for new or reconstructed mainline railway bridges.
For London Underground underbridges, reference to Engineering Standard
E8001[25] and consultation with the Track Engineer is required to determine the
provision of derailment containment.

Impact from traffic under the bridge

Bridges liable to impact from road vehicles (or impact from floating debris or
vessels) should be designed to resist the appropriate loads (see Section 6.3.5).
Elements liable to local damage in these circumstances, such as bottom flanges
of girders, should be robustly proportioned and bearings may be required to
provide resistance to uplift and lateral displacement.
2.6
Durability
Owing to the need for track possessions, road closures, or access through land
belonging to third parties, it is usually difficult and expensive to gain access to
railway bridges for the purpose of inspection or maintenance. Particularly for
smaller bridges, the cost of access and inspection can be large in proportion to
the value of the structure, or to the cost of any physical repairs. Minor defects
can generate significant administrative costs in the form of reports,
correspondence etc. especially where members of the public express concern,
even though the defects may be of no structural significance.
From the maintenance engineers point of view, the ideal steel bridge has a
minimum number of parts to inspect, all of which are easily accessible for
inspection, blast cleaning and repainting, without places where dirt and moisture
can collect. From the assessment engineers point of view, the ideal steel
bridge will have all its main structural elements visible and measurable.
Important components and connections that can only be inspected by removing
the track and ballast should be avoided. Detailed bridge examinations take
place on a different (and more frequent) cycle from that for track renewals and
in practice it is difficult to co-ordinate the two. Track removal purely for the
purpose of bridge examination is rarely justifiable, because of the disruption and
costs incurred.
Protective coatings for steelwork need to be high quality and with a long life
between maintenance, because of the difficulties and costs of access to the
12

surfaces. Code of Practice RT/CE/C/002[11] gives recommendations for the

selection and application of protective treatment and Specification
RT/CE/S/039[15] gives the actual requirements for the treatments. Instead of
surface coatings, in some circumstances (for example, deck type structural
forms) the use of uncoated weathering steel may be appropriate and this can
eliminate the need for any maintenance of the exposed surfaces during the life
of the bridge (although periodic inspection will still be needed). General advice
on the use of weathering steel is given in GN 1.07[47], in ECCS publication No.
81[38] and in the Corus publication Weathering steel bridges[39].
Steelwork should not be in direct contact with track ballast, because of the risk
of damage and corrosion. Steel surfaces buried in ballast should be protected
by waterproofing and protective layers or by concrete. Ingress of water behind
protective concrete should be prevented by fully effective sealing. This is often
achieved by the use of weather flats (see detail in Section 10.2.5). For all
main steelwork, other than fully encased beams, at least one face of each
element should be accessible, so that the thickness can be checked with an
ultrasonic gauge. For this and other reasons, half through bridges with a
central plate girder located in the space between two tracks are undesirable,
since the web of the centre girder may be inaccessible. If it is essential to
provide a girder in this location, because of construction depth constraints, then
it is better to provide two webs, with access between them, by using a box
girder or a pair of Z-section plate girders. (Alternatively, provide ballast
retention plates either side of the central girder, separated from the girder by an
air gap sufficient to permit inspection and maintenance but detailed to prevent
ingress of debris.)
It is considered good practice on smaller spans to use fabricated bearings as
these can be more robust and require less maintenance than proprietary
bearings; they are also generally more compact, which eases geometric
constraints. Simple robust fabricated steel bearings can reasonably be expected
to last for the design life of the superstructure. See Section 7.8 for guidance on
design of bearings.
General advice on good detailing is given in the SCI publication Design of steel
bridges for durability[45].
2.7
Clearances from the railway
2.7.1 General principles

Clearances between passing trains and the bridge structure, or any fixtures
attached to it (e.g. cables supported off the parapet, handles on removable
ballast plates, etc.), must be sufficient to ensure that:
Trains and their occupants may pass safely (including passengers or crew
who may be leaning out of windows).
Staff may work on the lineside safely.
Aerodynamic forces on the structure, trains and passengers are acceptably

low.
Where the railway is electrified, electrical arcing does not occur between
the structure and the pantographs, collector shoes or other electrical
equipment of passing trains.
13
The boundary enclosing the necessary clearances (outside which all structures
must be located) is known as the structure gauge. The clearances required
depend on the permitted speed of trains and track geometry. For new or
reconstructed bridges, a simplified structure gauge is used, which is based on
conservative assumptions for the sizes of trains and is defined by dimensions
from the rails, rather than by clearances from trains. The requirements have
evolved over the years to accommodate developments such as increased speeds,
new rolling stock and larger freight vehicles.
The structure gauge and clearances are measured from the running edges (or
inner faces) and the tops of the rails. Where the track is canted, dimensions are
measured parallel and normal to the plane of the top of the rails (see
Section 2.7.4).

Minimum clearances are given in the HMRI Railway safety principles and
guidance[49]. Part 2A, Chapter 6 gives the definitions and principles that are
used to define a structure gauge. Further requirements and recommendations,
based on the HMRI guidance, are given in the relevant Railway Group
For a full understanding, these documents should be
Standards[5,8,9,10].
consulted.
The following Sections illustrate the application of current
requirements to typical steel underline bridges.
The clearance requirements for London Underground tracks are given in
Engineering Standard E8013[26]. London Underground operates two distinct
sizes of rolling stock Subsurface stock, which is very similar in section to
national railways stock, and the smaller Tube stock. Some open sections of
Tube lines were originally constructed for subsurface stock; hence, they can still
be used for new stock delivery, transfers between lines etc. Also, track
maintenance contractors are increasingly using main-line rail-mounted machines
on the Tube open sections where clearances permit. Consequently, clearances
depend on the Line and the permitted rolling stock.
Outside the structure gauge, there needs to be space for access along the railway
(see Section 2.2.2) and this requirement contributes to the total clearance
necessary between track and structure.
2.7.2 Lateral clearances to parapets

For a simple composite deck type bridge, the lateral clearance between the track
and the structure is as shown in Figure 2.1. The required clearance is shown as
the sum of two components, one the clearance required to the near edge of the
lineside access way and the other for the width of access way beside the track.
The Figure shows the provision for a cess walkway, which would normally be
provided on both sides of the line, but it may be acceptable to provide a cess
walkway on one side and a (narrower) position of safety on the other. In this
arrangement, the parapet upstand acts as a robust kerb (see Section 2.5)
14
X+Y
is the width of cess walkway (700 mm minimum)
is the clearance from running edge of nearest rail to the walkway or place of
safety (dimension depends on line speed)
(These dimensions are defined in GC/RT5203)
Figure 2.1
Lateral clearances to parapets

For London Underground tracks, Engineering Standard E8013[26] gives the

required clearances and defines the installation of Limited Clearance Warning
Boards where required clearances cannot be achieved
Walkway requirements for tracks are given in Engineering Standard E8052[27].
2.7.3 Clearance to girders of half through bridges

Half through construction is extensively used for steel railway underline bridges
in order to minimise construction depth. The main girders are positioned either
side of the track and the track is supported by a shallow floor spanning
transversely between the main girders, near the level of their bottom flanges.
(See descriptions of various forms of half through construction in Section 4.3.)
To minimise the depth of the floor, the main girders are placed as close to the
track as possible. In such cases, the positions of the girders, and the amount of
clearance required between them and the nearest rails, depends on the height of
the top of the girder above rail level. There are three basic cases, as described
below:
15
(i) Top of main girder more than 915 mm above rail level
Sufficient space must be provided for a walkway or place of safety between the
track and the girder and this determines the spacing of the main girders. See
Figure 2.2. The top flange of the girder acts as a robust kerb, as described in
Section 2.5.
X+Y
X
Y
is the width of cess walkway (700 mm minimum) or the width of a continuous

position of safety (400 mm minimum)

Figure 2.2
Half through bridge with top of main girder more than

915 mm above rail level and with an access way inside the
line of the girders
Where the width of a replacement bridge is constrained by the width of the

existing abutments and there is insufficient width for cess walkways, separate
walkways can be provided on the outer faces of the girders. There should still
be positions of continuous safety alongside the track or a walkway outside the
girders. See Figure 2.3
X+Y
X
Y
is the width of a continuous position of safety (400 mm minimum)

Figure 2.3
Half through bridges with the top of the main girder more
than 915 mm above rail level and with a walkway outside
the line of the girders
16
(ii) Top of main girder between 110 mm and 915 mm above rail level
When the top of the girder is between 110 mm and 915 mm above rail level, a
walkway/place of safety is usually provided at the level of the top flange of the
girder. The girder spacing is then governed by the lower sector structure gauge
(commonly known as the platform gauge, since it is occupied by station
platforms) given in GC/RT5212[9] Appendix 1. A typical arrangement is shown
in Figure 2.4. In this Figure, a possible location for a third main girder is
shown; it is located within the six foot, the space between tracks, and within
the limits of the lower sector structure gauge. (See comments on durability
aspects of the option in Section 2.6.)
X+Y
See Note 2

915
See Note 1
X and Y are as defined in Figure 2.1

Note 1: The lower sector structure gauge (or platform gauge) as defined in GC/RT5212
Appendix I (note that the horizontal dimension depends on line speed and track
curvature)
Note 2: The space available for a centre girder (not preferred)
Figure 2.4
Half through bridge with top of main girder between

110 mm and 915 mm above rail level
Where the top of the girder is more than 500 mm above the tops of the sleepers
then access by steps or similar must be provided at intervals to give access from
the track (See GC/RT5203[8]). Note that if the girder is less than 300 mm above
rail level it cannot be considered to provide a robust kerb (see Section 2.5) to
contain derailed trains. Consequently, the girder will need to be designed to
carry vertical loads from derailed trains and a suitable parapet that is capable of
acting as a robust kerb should be provided.
(iii) Top of main girder not more than 110 mm above rail level
When the top of the main girder is 110 mm or less above rail level, the girder
can, subject to a specific derogation from Network Rail, be located within the
outer part of the "area for items intended to come in close proximity to trains"
given in the GC/RT5512 diagram for the lower sector gauge. This is shown in
Figure 2.5. (Location of fixed infrastructure in this area was previously
allowed by British Rail; Network Rail are currently considering altering the
diagram.) In practice, the actual clearance from the running edge of the rail to
the girder is limited by the length of the sleeper end and the need to leave space
to pack ballast between the sleeper end and the structure (typically 120 mm
minimum). Where there is third rail electrification, electrical clearance to the
conductor rail may be the limiting factor (see the Track design handbook[16],
Sheet A.8.7). Note that since, the girder is less than 300 mm above rail level,
17
it cannot be considered to provide a robust kerb; the walkway must therefore

be designed to support vertical derailment loads and be provided with a separate
robust kerb, as shown on the left hand side of Figure 2.5. An alternative
arrangement is illustrated on the right hand side of the Figure, where the outer
girder has been raised to provide a robust kerb, allowing a cantilevered
walkway to be used.
X+Y
Z
See Note 1
See Note 2
110 max.
X and Y are as defined in Figure 2.1

Z = (Sleeper length gauge / 2, plus sufficient gap to enable ballast to be placed at
the sleeper end, plus an allowance for track maintenance and construction
tolerances. For third and fourth rail electrification, additional clearance may be
needed to the conductor rails.

Note 1: The lower sector structure gauge for areas close to the plane of the rails, as
defined in GC/RT5212 Appendix 1. See further comment in text
Note 2: An alternative arrangement, with the outer girder raised to provide a robust
kerb, permitting the use of a lightweight walkway
Figure 2.5
Half through bridge with top of main girder not more than
110 mm above rail level
On London Underground tracks, the relevance of the height of the top of girder
above rail level depends upon the permitted rolling stock for a particular line,
therefore reference must always be made to Engineering Standard E8013[26] and
to its structure gauge and structure profile diagrams.
2.7.4 Effects of curvature and cant

End and centre throw
When a rail vehicle is on curved track, its corners overhang the rail to the
outside of the curve by an increased amount. This increase is known as the
end throw. To the inside of the curve, the middle of the vehicle has an
increased throw, and this is known as the centre throw. Structure gauges are
normally specified for straight track and the lateral clearances need to be
increased on curves to allow for throw. Guidance on the calculation of throw
may be found in the Track design handbook[16] and in document GM/RT2149[10].
The amount of additional width required to accommodate throw is small on
normal radius curves and it is usual to provide a generous nominal amount. In
the case of the structure gauge for the area up to 1100 mm above the plane of
the rails, defined in GC/RT5212[9], Appendix 1, specific requirements are given
for increasing the horizontal clearance to allow for throw.
18
End throw
Centre throw
Figure 2.6
End and centre throws for a vehicle on curved track
Cant
Railway track on curves is usually superelevated to compensate for the effects
of centrifugal forces. The amount by which the outer rail is raised above the
inner rail is known as the cant. The maximum amount of cant permitted is
normally 150 mm, although in certain circumstances this can be exceeded. The
effects of cant on clearances are usually dealt with by tilting the structure gauge,
i.e. all clearances are measured relative to the plane of the canted track, rather
than horizontally and vertically. This is illustrated, (applied to the half through
arrangement illustrated in Figure 2.4) in Figure 2.7.
Ideally, for track
maintenance reasons, canted tracks are arranged coplanar but sometimes the
track planes (top of rails) are staggered, as shown in Figure 2.7.
X+Y
X+Y

See Note 1
90
See Note 1
Track stagger
X
is as defined in Figure 2.1
Y
is as defined in Figure 2.1 plus an allowance for throw
Note 1: The lower sector structure gauge for area up to 1100 mm above the plane of
the rails, defined in GC/RT5212 Appendix I (note that the horizontal dimension
depends on line speed and track curvature)
Note 2: Any track stagger will be limited by the railway authority, to ensure adequate
ballast retention
Figure 2.7
Half through bridge (tops of main girders between 110 mm

and 915 mm above plane of rails) showing effects of cant
on required clearances
This approach is not entirely appropriate when considering lateral clearances to

parapets based on considerations of staff safety, and the preferred method is
illustrated in Figure 2.8.
19
See Note 1
2000
See Note 1
90
X
Y
Track stagger
is as defined in Figure 2.1

is as defined in Figure 2.1 plus an allowance for throw
Note 1: The lower sector structure gauge for area up to 1100 mm above the plane of
the rails, defined in GC/RT5212 Appendix I (note that the horizontal dimension
depends on line speed and track curvature)
Note 2: Any track stagger will be limited by the railway authority, to ensure adequate
ballast retention

Figure 2.8
Half through bridge (main girder between 110 mm and

1100 mm above plane of rails) showing effects of cant
(preferred method)
On London Underground, the calculation of the effect of curvature (i.e. the end
throw, centre throw and cant) on clearances is given in Engineering Standard
E8013. This standard also covers the clearance approval procedure.
2.7.5 Track position on the bridge

The position of the track is determined by the railway authority. In determining
the geometry of the bridge, the designer should take into account:
The specified nominal position of the track
The tolerance in track position permitted by the relevant railway standards

(and whether such tolerance is already included in the clearance gauge)
Any specified allowance for planned track realignment in the future
Installation tolerance (typically 25 mm on the plan position of the

structure, see Section 3.7)
If space permits, an allowance for other unforeseen adjustments (typically

up to 100 mm)
Any proposals to change line or level of the track to suit the structure (even to
meet current requirements for clearances, etc.) must be agreed with the railway
authority.
20
2.8
Underline clearances
2.8.1 Headroom
In the case of an existing underline bridge, the owners legal liability for span
and headroom is usually specified by the Act of Parliament under which that
particular line was constructed. Before commencing design of a scheme for
reconstruction, the legal status of the bridge should be established.
Many existing railway bridges over highways are of substandard headroom and
are vulnerable to accidental impact from road vehicles. Where such a bridge is
to be reconstructed consideration should be given to reducing the risk by
increasing the headroom, by use of robust construction (see Section 2.5) and/or
by providing collision protection beams or other measures, such as improved
signing or traffic control. Detailed guidance is given in GC/RC5510[5].
Guidance on and requirements for the design of collision protection beams is
given in BD 65/97[41] (part of the Design Manual for Roads in Bridges).
2.8.2 Construction depth

The construction depth of a railway bridge is the vertical dimension between

the tops of the rails and the bridge soffit.
Because the lowest acceptable soffit level is usually constrained by the existing
roadway and its required headroom, and the level of the railway is very tightly
constrained by track geometry, the available construction depth is a very
important design parameter. It has a strong influence on the form of
construction that can be adopted (see Section 4).
There are three components that add up to the construction depth of the bridge:
the track depth (top of rail to underside of sleeper), the ballast depth beneath
the sleepers and the structure depth (top of the deck or ballast tray to the
underside of the structure and including the thickness of the waterproofing). Of
these components, the track depth is fixed by the choice of track and the ballast
depth is usually 300 mm (see discussion on page 1). The structure depth is
therefore usually constrained by the difference between the available
construction depth and the sum of track depth and ballast depth.
21
CONSTRUCTION REQUIREMENTS
3.1
General
The design of railway bridges has always required the design engineer to give
detailed consideration to the possible methods of construction that might be
available at a particular site. This is considered a fundamental requirement in
order to produce a design solution that can be translated into reality within the
very short periods usually available for such activities. This is particularly so in
the case of underline bridges because they are required to be capable of
supporting the imposed railway loads by the time the structure is ready for
reinstatement of the track.

With the introduction of the Construction, (Design and Management)

Regulations in 1995, the need for the designer to consider carefully the effect of
his proposals, from a safety point of view, was formalised. The designer is
required to undertake design risk assessments and thereby develop a solution
that minimises the health and safety risks to the public and the personnel who
will build and maintain the bridge. For general guidance on the effect of CDM
on steel bridge construction, see GN 9.01[47].
When the requirements of the CDM Regulations are considered together with
the clients operational requirements, it becomes obvious that knowledge of
methods of construction, and the suitability of each, play an important part in
selecting the appropriate design solution for the particular site.
3.2
Minimising disruption to the operational

railway
Minimising disruption to the railway is one of the key criteria to be understood

by the designer in developing a successful solution. Due to the nature of the
railway, access on or near the railway is restricted and generally personnel or
operations that can lead to objects coming within a horizontal distance of 3 m of
the nearest operational rail can only be undertaken during closures of the
railway, called possessions. (The restriction applies also to plant that could
potentially fall and land within this distance.) Even work outside the 3 m limit
(such as preparatory works) may be subject to speed restrictions being placed on
the line and this clearly is another disruption to railway operations.
Therefore, prior to commencing the design of a new or replacement railway
structure, consideration must be given to the level of disruption that would be
acceptable on the railway line(s) that would be affected by the works. For
example, on a little used siding, a closure in the order of weeks may be
acceptable but on a busy suburban or main line, a closure of only one or two
days may be the maximum available for the proposed works.
Disruption can be considered in two main categories, Speed restrictions and
Possessions.
3.2.1 Speed restrictions

Any works that are likely to affect the stability of the track and thereby increase
the risk of a derailment occurring will require a Temporary Speed Restriction
22
(TSR) to be imposed. Installation of the track and ballast on to a newly

constructed bridge will require the imposition of a TSR. The severity and
duration of the TSR will be governed by the amount of time available for the
trackwork contractor to lay and bed down the track. Only a certain level of
disruption is permitted on a particular stretch of railway line at a given time;
this may dictate when the speed restriction can be available for particular
bridgeworks.
3.2.2 Possessions
Most bridgeworks are undertaken during possessions. A possession can be
defined as the closure of a section of the railway to normal rail traffic
The availability of possessions is classified as either Rules of Route, which are
those available for the day to day maintenance of the railway or Outside Rules
of the Route (also known as abnormal), which are special possessions usually
of longer duration and booked for specific activities.

Rules of Route Possessions, available for regular maintenance of the railway,

usually vary between 4 and 29 hours depending on the lines affected. These
possessions are generally booked three months in advance. Sometimes it is
possible to reduce this booking period by undertaking works on the back of a
possession provided for other work, as long as it does not adversely affect the
works the possession was originally booked for.
Outside Rules of the Route (abnormal) possessions are the ones most commonly
used for bridgeworks because they generally offer a longer duration for
construction activity. These possessions are usually booked up to two years in
advance and are commonly up to two days duration or, in exceptional
circumstances (such as over periods with public holidays), longer. Occasionally
the abnormal possession duration for the main bridgeworks will have been
agreed in advance of the design commencing, thus representing a constraint on
the solution to be designed.
The possession duration includes for all activities, from taking the possession to
handing it back. Depending on the complexity of the railway infrastructure,
(such as the track layout, traction power and signalling at the particular
location, most of which has to be removed and replaced to allow the
bridgeworks to proceed), there is often little more than half the total length of
the possession available for the bridgeworks. Therefore, this often leads to
prefabrication being maximised and any connection details being robust and
fully proven in advance of the installation possession.
Because of the importance of handing back the possession on time for train
operations to recommence, the designer needs to produce a viable scheme that
can be constructed within the available possession period, during any reasonably
foreseeable inclement weather and with consideration given to the robustness of
every detail. Specifying that all components are trial erected, and ensuring
there are adequate tolerances in the design for fit up, will significantly help to
achieve this.
It should also be remembered that any closure of a road below the railway will
require the permission of the highway authority, although the requirements are
not usually as onerous as those for railway possessions.
23
3.2.3 London Underground

The situation is similar for London Underground, except that work can also be
carried out during nightly engineering possessions. The period between traction
current 'off' and 'on' is termed Engineering hours and it is often possible to
carry out some structural work within this short period. Where the work would
extend into normal traffic hours, a possession for the full duration of the works
must be applied for.
3.3
General site constraints
For most structures, it is necessary to understand the impact of site constraints

before developing the design solution, as these factors can dictate the method of
erection and the form of the structure.
3.3.1 Site Access

This is the single most important consideration because access by road to the
bridge site is not always available, particularly if the bridge spans over an
obstruction other than a road (e.g. a canal, river, or flood plain). In such cases
careful thought needs to be given to researching the types and quality of access
that might be arranged to enable the particular design solution being considered.
Examples of types of access often considered are:
The construction of temporary road across an adjacent landowners field or

flood plain after appropriate negotiations have been conducted. Access
across a flood plain is risky for obvious reasons and is subject to consent
from the Environment Agency.
Accessing the bridge site along the railway from one end using road plant
if convenient access to the railway is available nearby.
Accessing the bridge site along the railway from one or both ends using rail
mounted plant or road-rail plant.
The quality of site access will determine the type and size of bridge elements
(and the type and size of plant) that can be brought to site. Particular care
needs to be taken to make sure that the load carrying capacity of any bridges
supporting the access road are adequate for the weight of plant and bridge
elements being considered.
3.3.2 Available working space

Many rail sites are in heavily built up areas and often buildings have been
erected adjacent to the railway after the line was constructed. These may
preclude various methods of erection and significantly constrain others. Apart
from the obvious physical obstructions, there are the issues of complex logistical
planning necessary to be able to work within a very small site that may consist
of a single street. This may affect the required possession duration, with each
bridge element needing to be delivered to the crane in a specific order in
accordance with the programme.
Other issues include those of air rights when either lifting elements or working
over others property. Consideration should be given to these by the designer
to try to minimise and eliminate these requirements where possible.
24
3.3.3 Services
Services include all statutory undertakers plant, from sewers and fibre optic
telecommunications cables buried in the road through to overhead power lines.
As part of the initial option development, the location of utilities services need
to be confirmed and their impact on the proposed scheme identified. This will
need to cover the viability of any diversionary works together with an order of
costs. These services can have a variety of impacts for example:
Sewers and other older gas and water services with shallow cover may
preclude access or use of heavy plant such as transporters or large cranes,
unless additional protection and or strengthening measures can be
implemented.
Fibre optic cables may require diversion for strengthening of existing

substructures, which may cost more to divert than the rest of the scheme.
Overhead power lines may require diversion or isolation during the

installation possession if a cranage scheme is being proposed.

3.4
Third party issues
Planning approval for the reconstruction of existing railway bridges is usually

sought by the owner under a permitted development order. Usually the
designer will be asked to prepare a drawing showing comparative elevations of
the bridge before and after reconstruction, including details of the colour of the
steelwork.
It is usually necessary to consider the effect of the works on statutory service
authorities. Many railway bridges carry telecommunications cables owned by
non-railway authorities, in addition to the railway-owned signalling, electrical
supply and telecommunication cables.
Liaison with the local authority regarding the effects of the works on street
lighting and road signs may be required. There may be constraints on the
availability of lane/road closures owing to the lack of suitable diversionary
routes, or the needs of emergency services or bus operators.
For waterway crossings, the appropriate river/canal authority must be consulted.
Local or residential noise and vibration restrictions may affect the design of the
bridge or track form required, particularly on London Underground lines.
There are various types of noise- and vibration-reducing track forms available,
with direct and indirect fixings. The bridge construction may have to be
designed to accommodate a particular specialised track form.
3.5
Methods of bridge erection
The erection of a new steel railway bridge will involve different activities
depending on whether it is a completely new structure or the reconstruction of
an existing superstructure.
The main methods of erection are as follows:
Lifting (by crane)
Rolling
25
Sliding
Transporting.
The most commonly used method of erecting a steel railway bridge is that of
lifting by crane, either piecemeal or in a single lift. However, in some
situations bridge elements cannot be lifted into their final position because there
is insufficient possession time available (for erection of pieces or for slewing of
an OLE system) or because of limitations on crane reach. Under such
circumstances, consideration needs to be given to designing a bridge that can be
erected off-line and then, during a possession, be moved to its final position. In
these situations, the other methods of erection such as rolling, sliding and
transporting need to be evaluated.
Diversion of the railway to facilitate construction of a replacement bridge is
unlikely to be feasible except in extremely rare cases where it is part of a major
scheme and, even then, it would have to demonstrate considerable advantages
over other methods. The disruption and costs of diverting a railway line,
especially the specialist signalling, electrification and permanent way elements,
are prohibitive and would probably outweigh the cost of rest of the scheme.

3.5.1 Lifting
Traditionally, the most commonly used method to erect a bridge is piecemeal by
use of one or more cranes. An all-steel bridge fully fabricated in a workshop
and trial erected to prove acceptable fit-up of the components can be speedily
erected on prepared substructure. The actual amount of time for the erection
depends on the overall size and type of the structure, particularly on the type of
connections to be made, and on the type of crane to be used.
The cranes used are usually road-mobile but can also be either rail-mounted or
on a floating vessel. Developments in both road-mobile and rail-mounted
cranes have significantly increased the size of elements that can be installed.
Some examples of typical scenarios are as follows:
Short span bridges (up to about 17 metre span) can generally be designed,
fabricated, and transported to site in such a way that complete bridge
decks, each capable of supporting one track, can be lifted into position in a
single operation. Two independent half through decks, such as the Z or
U Types (see Section 4), can carry a double track railway. For a 10 m
span, a single-track all-steel or steel/concrete composite Z or U type
unit would weigh between 35 and 70 tonnes.
If site constraints do not allow lifts of such magnitude, such as where OLE
is present and there is not enough time to allow it to be slewed, then a
short span bridge may have to be delivered to site as girders and short
lengths of transversely stiffened deck. Having previously trial erected the
deck complete with sill units at the fabricators works, the process of
bolting together the components on site can be rapid.
Medium span bridges, from say 15 m to 40 m, and continuous multi-span

bridges of a similar configuration are usually too heavy or bulky to be
lifted in a single lift and are therefore most easily dealt with piecemeal, as
long as the construction is fully prefabricated, either wholly in steel or
steelwork with precast concrete units.
In assessing the viability of cranage proposals, advice should be sought from

specialist cranage contractors, especially for larger and more complex lifts using
26
more than one crane or using vessels. This advice should also be sought to
ensure a workable cranage scheme is developed that takes account of the
physical dimensions of the proposed crane, access for the crane (including for
rigging), the suitability of the ground to support the crane out rigger loads and
the presence of services.
Before the final selection of cranage as the erection method, consideration
should also be given to the following:
Access to site for crane and bridge elements
Overhead power lines
Overhead electrification equipment
Underground services
Exposure to wind or flooding
Available possession time
Availability and locality of back up plant, spares and fitters.

Further general guidance is available from documents such as the Network Rail
Model clauses that cover such works.
3.5.2 Sliding/rolling
As with the other techniques, there is little technical literature on this subject
therefore detailed information on systems can only generally be obtained from
specialist contractors and manufacturers of such systems.
The bridge designer does not need to detail the system (although, to comply
with CDM at least one specific system should be shown to be viable) but
information on the techniques is given here to aid an appreciation of the benefits
and restrictions of such systems.
Although the principle behind the two methods is the same, sliding and rolling
are different techniques.
Sliding consists of sliding the structure (usually heavier structures) on low
friction surfaces. A large number of systems are available using different
combinations of materials at the sliding interfaces such as phosphor/bronze or a
PTFE sledge on stainless steel. The actual coefficients of friction will depend
on the characteristics of the particular materials used. Typically, the coefficient
of friction at breakout is likely to vary between a minimum of 5% and a
maximum of 12%. During sliding, the coefficient generally reduces to between
2% and 8%.
Rolling consists of either rolling the structure (usually lighter structures) on ball
bearings constrained in a channel or on proprietary rollers supported on rails or
on proprietary skates. The comparable values for friction for rolling on 75 mm
diameter steel balls is about 10% at breakout and 2.5% once rolling. For
proprietary roller units these values can be as low as 2.5% for both breakout
and rolling resistance.
Both techniques potentially carry higher risks and may be more expensive than
erection by crane. The risks are generally associated with the installation of the
temporary works to support the slide paths or methods of controlling the
27
movement so that the structure does not crab and/or seize during the possession.
However, these methods offer speed advantages and in certain circumstances
offer the only viable method of erection. The principle of building the bridge in
one place and moving it to its final position is particularly suitable for the
erection of medium to long span and multi-span steel, steel/concrete composite
bridges. Such bridges would take too long to erect in situ, unless a special long
possession can be obtained.

Both methods require the construction of temporary works adjacent to and

beneath the track to support the slide path on which the new bridge will be
erected. The rails, rolling paths or sliding paths are usually built into the
temporary works and generally run parallel to the lines of the substructure
elements. They can in some cases form part of the permanent works. The
bridge is then constructed parallel to its final position and slid /rolled into
position, usually at a higher level and then jacked down onto the permanent
bearings.
The location of the roll/slide path is usually inboard from the bearing positions
and clear from the face of the abutment, to avoid congestion in the bearing shelf
area. Consideration also needs to be given to how the structure is supported
during the transferring operations, as some structures are more tolerant of
differential settlement of supporting temporary works and varying support
reactions than others. It may be necessary to interlink the hydraulic circuits to
supporting jacks to ensure that the supporting reactions are evenly distributed
and to avoid over loading individual sledges, rollers and jacks.
Depending on the weight of the bridge and friction resisting movement, the
plant used for moving the bridge usually consists of a system of winches for the
lighter structures or hydraulic jacks for the heavier ones. The system will also
need to be designed to withstand a force normal to the main pulling force to
ensure that it is suitably robust; this value is normally taken as 30% of the
maximum pulling force. Provision should be made for reversing the direction
of sliding or rolling, should there be an overshoot or if crabbing occurs.
Closed circuit television cameras are often used to provide information on
progress so that action can be taken if required. This is particularly relevant if
the chosen method is susceptible to seizing if one end of the structure moves
differentially in relation to the other.
Due to the importance of completing the slide/roll within the specified
possession, a trial slide/roll, typically of 2 m, is usually undertaken in advance
of the possession to prove the system. Wherever possible, the complete slide
track or roller path should be installed in advance.
In exceptional
circumstances, where this is not possible, it is essential that the construction of
the slide track or roller path be simulated as closely as possible during the trial
slide/roll, to validate this part of the possession programme.
From experience, a key consideration in determining the feasibility of a sliding
or rolling method of installation is the positional tolerances that need to be met
and whether they can be attained within the limitations of time and space during
a possession.
It is possible to use sliding/rolling techniques also to remove the existing bridge
so that it can be safely demolished without affecting the critical items of work.
28
In this case, there must be enough space to accommodate a similar slide track or
roller path on the opposite side of the railway.
3.5.3 Transporting

This is a relatively recent technique for moving railway bridges, first used to
erect a new bridge in the early 1990s. The technique consists of using multiaxle highly manoeuvrable vehicles sometimes described as self propelled lifting
vehicles (SPLVs), to lift the ready assembled bridge from temporary works at a
nearby site and transport it to its final position. This method of installation is
particularly suited to bridges over highways or where the presence of overhead
electrification or very restricted possession duration precludes crane erection.
The method is generally more expensive than erection by crane and is
comparable with sliding and or rolling.
SPLV units are typically 2.4 m wide and 8.4 m long and can lift 96 tonnes.
The units can be interconnected to increase the lifting capacity. The SPLVs can
move in any direction in the horizontal plane, and rotate 360 degrees about a
vertical axis; they also have a limited capability to lift and lower the structure.
The vertical movement range is generally limited by the stroke of the jacks on
each wheel (typically 600 mm) and some of the available range will be utilised
in compensating for unevenness of the terrain over which the trailers have to
pass. On flat level ground most of the range may be available but on rough or
uneven ground little may be available for raising/lowering the structure.
Additional vertical movement can be provided using timber block towers on top
of the vehicles, but removing the timbers is a time consuming operation and
therefore this should be kept to the minimum practical. In detailing the
structure, consideration should be given to the accuracy to which the structure
can be positioned: a tolerance of 25 mm should normally be allowed for plan
position of components.
The main considerations, when this type of method is being adopted, are:
The surface on which the vehicles are to traverse. An existing highway is

generally considered to be suitable, although the effect of the trailers on
particularly weak or vulnerable surfaces, services and basements etc.
should be checked. Unmade ground, especially newly exposed
embankments when installing a new structure where no previous structure
existed, or winter conditions, are less suitable for SPLVs because of the
additional risks.
The effects on the bridge while it is being transported need to be clearly

understood and designed for. This is not only a capacity issue but also one
of deformation (including twist), as the vertical reactions from the
temporary support positions can cause horizontal displacements that cause
difficulties with positioning the structure, particularly for high skew
structures.
Access is required to a suitable site where the bridge can be temporarily

erected. This site needs to be near the bridge site, usually not more than
500 to 700 m away, to keep traffic disruption and any relocation of
roadside furniture to a minimum.
Specialist advice should be sought from the contractor offering the SPLV
system to make sure the vehicles can traverse the route between the remote
site and the bridge site.
29
The bridge should ideally be fully waterproofed with track and ballast
already laid before transporting it in, order to realise the maximum benefit
from the system.
Consideration should be given to using the same system for removal of the
existing bridge.
Urban areas generally are less suitable for this kind of treatment because of
the lack of suitable vacant site within easy reach of the bridge site and the
presence of significant roadside furniture that will require temporary
relocation.
3.6
Replacement of the permanent way

The permanent way will require removal and reinstatement by a specialist

contractor at the beginning and end of the installation possession. There are
various methods and types of specialist plant available for this work, which will
depend on the length of track to be removed, its condition and the site
constraints. In some cases, the bridge may be erected complete with track,
which must then be joined to the track either end of the bridge.
Where there is third and fourth rail power system, this is usually removed and
reinstated with the permanent way. Where there is an OLE system, it is
common practice either to design the structure so that it can be slid/transported
into position (so that the OLE is unaffected) or to design the structure in
discrete elements that can be installed in ways that require the OLE wires only
to be slewed a small amount. Removal and reinstatement of OLE is very time
consuming and should be avoided wherever possible.
Other railway infrastructure will also require reinstatement during the main
bridge erection possession or in subsequent possessions. As a minimum, time
will generally be required at the end of the main possession to restore and test
signalling track circuits.
3.7
Buildability
When designing and detailing a steel railway structure, proper consideration

needs to be given to the practicality of fitting the structure together, assuming
that the work may be done in inclement weather and at night. This leads to the
development of details that can accommodate realistic site erection tolerances.
To ensure fit of the structure in the limited possession time available, all
components should be fully trial erected at the fabricators works. This is
usually done before the application of protective treatment and waterproofing
but after all welding is complete (assuming all site joints are bolted, see
discussion below). The trial erection should include placement of any precast
concrete elements such as sill units and ballast walls and ancillary items such as
ballast plates, as any lack of fit with these items is far easier and more
economical to address in the fabricators workshop rather than on site during the
installation possession. For general advice on trial erection, see GN 7.04[47].
Whichever method of erection is used, it is best to minimise the site activities as
far as possible. Activities such as waterproofing should ideally be undertaken
before delivery to site. This usually means that they are carried out under
controlled conditions. Obviously, when steelwork connections still have to be
30
made at site there will be areas of steel that will have been masked and
therefore require to be coated or waterproofed on site. These will be relatively
small areas and can therefore be accomplished quite quickly. Pre-waterproofing
the deck also removes a major weather-dependant activity from the critical path
and this in itself is a major benefit.
The design should generally use shop welded and site bolted connections for
steelwork, although there is a place for site welding in some circumstances (for
example where pre-assembly is possible outside of possessions). Structural
connections that require site welding during a possession should not be
specified, as it is not easy to carry out welded connections to the specified
quality in the limited periods of time available during a possession. However,
bearing location plates are a particular exception, where the welded detail is not
subject to fatigue loading and welding is specified to improve site tolerance.

Site bolting generally consists of using field bolts at key locations in the entire
structure in order to pull the components fully together at each joint location,
before replacing them progressively with the specified permanent fasteners.
This ensures that the permanent fasteners will function satisfactorily as designed
in the completed structure.
In situ site splicing of the main girders can be time consuming because
temporary works in the form of trestles with working space for personnel at just
the correct level are required near the splice positions. Such splices in box
girders can be particularly difficult and, being on the critical path, should be
considered carefully if possession time is limited. Where possible, splices
should be made in advance of the possession, preferably at ground level.
Transportation of the fabricated elements to site also needs to be considered by
the designer. For transport by road, structural components should where
possible be detailed so that the loads comply with ordinary size and weight
limits. This conflicts, to some extent, with the desire, for reasons of durability
and maintenance, to minimise the number of site connections and thus the
number of pieces that are transported. Also, the limited times available for
erection often dictate the use of the largest possible component sizes and thus
the transport of larger loads. Larger loads are termed abnormal indivisible
loads and movement requires notification to the police. Very large loads
require special Movement Orders. See GN 7.06 for more detailed advice.
Transport by rail should generally be avoided, if at all possible, because of the
risk it imports into the construction activities. If rail transport is considered,
specialist advice should be sought to ensure that members fit within the relevant
loading gauge, which will be dependent on, among other things, the selected
transporting vehicle type.
Generally, the following checklist is suggested:
Check requirements for transportation of large and heavy loads
Ensure that the size and weights of prefabricated units are appropriate to
the capacity, reach and jib height of readily available cranes that will suit
the site.
Make site setting out simple and provide generous tolerances for location of
the structure. (Typically, when erecting piecemeal by crane all bearing
fixing details should accommodate tolerances of at least 25 mm in plan
31
position, 10 mm in vertical level and be able to accommodate lack of fit

in bearing surfaces.)
Provide prefabricated components with purpose-made lifting points

symmetrically disposed about the centre of gravity and at same level,
allowing units to be lifted and balanced easily.
Carry out a trial erection of the bridge components at the fabricators

works.
Do not make assumptions about critical parts of existing structures that are
to be reused (for example, the thickness of the existing bridge abutments
should be checked by coring or trial pits). Investigate thoroughly.
Consider carefully the method of installing the bearings. RT/CE/C/008[12]

provides detailed guidance notes.
As far as possible, avoid the use of methods and materials requiring precise
quality control or which are susceptible to bad weather.
Avoid minor operations on the critical path for erection.

Further guidance on buildability is given in CIRIA Report 155[53]. The model

specification document RT/CE/C/008[11] also includes guidance on a variety of
railway bridge installation methods.
For a bridge replacement, seizure of the bearings, as a result of corrosion etc.
could result in the existing deck structure acting as a prop between abutments.
Stability checks on the abutments should be considered for each phase of
construction.
Where cranes are positioned on ground immediately behind existing abutments,
to lift in the replacement bridge girders and deck, the stability of the abutment
under the erection loading must be considered.
Dealing with services
Where an existing bridge is to be reconstructed, the method and sequence of
dealing with the existing cables during the works should be considered at an
early stage. The expense and possession time required to disconnect and
reconnect cables can be considerable, particularly where fibre optic cables are
present. Temporary slewing of the cables onto temporary supports is usually
preferable to disconnection, but the practicality of this depends on the amount of
slack available, the condition of the cables, and whether they are on the outside
or the inside of a curve. The presence of a temporary cable bridge may
obstruct crane working, and affect the size of crane required.
London Underground lineside services (high and low voltage cables, signal and
communication cables, compressed air mains plus third party cables) are
generally supported across bridges off the parapet steelwork. Diversion and
reinstatement for a bridge replacement is such a significant cost of a project that
alternatives to roll-in and cranage require consideration. The use of SPLVs (see
Section 3.5.3) to transport a complete bridge deck under cable runs, supported
by a temporary transfer bridge, has been successfully employed.
32
FORMS OF CONSTRUCTION
4.1
Influences on form of construction
In addition to the influences on bridge design discussed in Sections 2 and 3,

there are three basic sets of parameters that determine the form of construction
of a railway bridge:
Available construction depth
Span and geometric configuration
Limitations imposed by the substructure
Construction depth

For some new bridges on new railway alignments, construction depth is not
particularly constrained; the track level and the road level beneath can be fixed
at levels that suit the structure. In such cases, there are more options to choose
the form of supporting girders and deck for maximum efficiency and economy,
and for aesthetic considerations.
Replacement bridges are much more likely to be constrained to a shallow
construction depth, because of the need to maintain a clearance below and to
avoid lifting the track. For very short spans, deck-type structures can be
entirely arranged within a shallow construction depth but in many cases the only
way to support the track is to arrange a shallow deck spanning transversely to
longitudinal main girders either side of the track. This form is known as half
through construction or, for top-braced deep trusses over longer spans,
through construction.
Span and geometric configuration
Clearly, the span has a direct influence on the depth of the main girders of a
bridge and thus on whether the girders can be arranged within the available
construction depth.
Single span bridges have historically been very common and even multiple span
bridges have tended to be built as a series of individual spans. There are many
good reasons for this but the discontinuity at intermediate supports has often led
to corrosion and maintenance problems. Today continuous construction is
preferred, where possible, but this then requires consideration of the interaction
of bridge and track (in response to longitudinal loading and thermal effects) and
the consequences on the reactions at supports; see discussion in Section 6.4.4.
Many bridges span skew to the abutments that support them. This can give rise
to track twist problems and difficulties in detailing the end cross girders and in
arranging bridge articulation. On new bridges, there may be scope for building
abutments square to the span, even when the railway crosses the road or river at
a skew, but replacement bridges usually have to be built to suit the existing
arrangement of the substructures.
Limitations imposed by the substructure
In addition to the limitations on replacement of existing skew bridges,
replacement on an existing substructure often constrains the width of the bridge.
The strength and form of construction of the abutments and intermediate
33
supports are likely to have a strong influence on the detailing of the bearings
and sill beams.
4.2
Shallow deck-type bridges
There are three forms of shallow slab or plank construction (i.e. where the
deck acts mainly as a beam spanning between abutments):
Solid steel slabs.
Orthotropic deck.
Steel beam sections encased in concrete (filler beam construction).
4.2.1 Solid steel slabs

The simplest bridges comprise simply supported slabs spanning longitudinally

between abutments. Solid steel slabs can be used to form such decks for
railway bridges up to approximately 3 m. Their advantage is a very low
structure depth, much less than alternatives such as prestressed concrete. The
inefficient use of material is offset by the low cost per tonne, due to the
minimal fabrication required.
The slabs simply sit on the abutments on elastomeric strip bearings. Transverse
continuity is not normally provided. To provide parapet supports and robust
kerbs, independent reinforced concrete units spanning between abutments are
often used, as shown in Figure 4.1. These provide the robust kerb and should
be designed to carry the loads from a derailed train.
Traditionally, quite wide slabs have been used, with typically two 1.7 m wide
slabs per track (this width corresponds approximately to half the normal centreto-centre spacing between tracks). The slabs are 200 to 250 mm thick and no
fabrication is involved, other than cutting to size. Waterproofing is provided on
the top surface of the slabs; at the joints the waterproofing usually cracks in
service but some leakage is normally accepted.
Usually, serviceability limits on deflection and rotation (see Section 9) govern
the design and steel stresses are relatively low. In the absence of welding,
fatigue is not normally a problem.
Steel slab
Concrete parapet unit
Figure 4.1
Simplified part cross-section of a solid steel slab deck

superstructure
4.2.2 Orthotropic deck

For spans of up to 9 m, a very shallow structure depth of approximately 300 to
400 mm can be achieved using all-steel units spanning longitudinally between
abutments. (For spans beyond the capability of solid slabs, orthotropic decks
offer the shallowest structure depth of any form of deck type construction.)
34
The deck units comprise a steel deck plate (typically 20 to 25 mm thick) with
structural Tee sections welded to its lower face, usually at about 600 mm
spacing. The Tees usually sit on elastomeric pad or strip bearings and end
diaphragms are created by casting concrete around the Tees at the ends.
The deck units are relatively flexible transversely.
Parapet/robust kerb
containment can be achieved using independent parapet walkway units located
clear of the tracks, in the same way as for solid steel slabs. See Figure 4.2
Figure 4.2
Diagrammatic cross-section of an orthotropic deck

Alternatively, parapets and robust kerbs can be provided by bolt-down steel

units at either side of the deck. To resist the effects of the horizontal loads on
robust kerbs, transverse bracing must then be provided between the Tee sections
(see Figure 4.3).
Figure 4.3
Orthotropic deck with robust kerb
Orthotropic deck construction is also used spanning transversely as the deck of

some half through bridges (both with plate girders and with the BR standard box
girders). See discussion about standard box girder bridges in Section 4.4.
4.2.3 Filler beam construction

Filler beam construction uses rolled steel sections embedded within and
generally acting compositely with a reinforced concrete slab. See Figure 4.4.
The concrete is cast to the level of the top of the bottom flanges and covers the
tops of the sections.
To achieve transverse continuity, the bottom layer of reinforcement is threaded
through holes drilled through the webs of the rolled sections and the top
reinforcement is above the steel sections. Structure depths vary from a
minimum of about 400 mm. In the UK, spans up to about to 13 m have been
constructed and for 13 m span, UC sections can be used (the largest UC
sections are about 400 mm deep). In Europe, spans up to about 30 m have
been constructed; for these, European wide flange beams (HE and HD series,
many of which are similar to UB sections) are used.
35
Figure 4.4
Filler beam construction
The deck can be poured in situ, if construction is outside a possession, but

usually the deck is constructed in a number of units, which are precast and
connected together using shear-key joints. Non-keyed joints may be acceptable
(either leakage of water through the joints would be accepted or there would be
a suitable sealing joint bridging adjacent units). Joints between units are
normally arranged between tracks (i.e. the unit would be the width of track plus
half the track spacing either side). Complete decks can be precast for
installation as single units, where appropriate access and cranage are available.

Filler beam construction is also used transversely as the deck of some half
through bridges. See further discussion below.
4.3
Half through plate girder bridges
Half through bridges are able to achieve shallow construction depths over
relatively large spans and various forms have been widely used for railway
underbridge construction. Essentially, such bridges comprise a pair of plate
girders that span between abutments and a deck, spanning transversely between
the plate girders. The deck may be of composite construction or an orthotropic
deck. The lateral stability of the top flange (which is in compression) is
achieved through U-frame action (see Section 7.1.3). The bridges are usually
single spans, although continuous construction has been employed. To carry
two tracks, a variation using three plate girders has been developed see
further discussion below.
British Rail developed a series of standard half through bridge types, some of
them now outmoded but many of which have led to modern derivatives. This
series of bridge types is described below, although not all details would now be
acceptable for replacement bridges.
4.3.1 Z Type
The Z type evolved from the A type bridge, which was used between about
1950 and 1970. The A type comprises two I-section plate girders spanning
longitudinally and simply supported on the abutments with a deck spanning
transversely between the main girders just above bottom flange level. Cross
girders are UC sections 152 mm or 203 mm deep at 600 mm spacing and with a
concrete infill that is flush with the flanges; the ends of the cross girders are
connected through a 4-bolt shear plate detail. Each deck supports one track;
where there are two tracks, the arrangement is as shown in Figure 4.5. The
span range for this type was 6 to 15 m.
36
Reinforced concrete deck slab

and steel girders @ 600 centres
Figure 4.5
Brick/masonry
protection
Simplified cross-section of the former A Type decks

The main problem with the A Type was that when two decks are provided to
carry adjacent tracks, the gap between the flanges of the girders in the six foot
(the space between adjacent tracks) is too small to permit inspection and
maintenance. For this reason, the webs were often protected by brickwork for
their full height between the flanges of the main girders (although brickwork has
not proved successful in preventing corrosion). Brickwork was also usually
provided to the trackside faces of the webs to provide additional protection to
the waterproofing and steelwork.
The Z Type deck which developed from the A Type solved the issue of
inspection/maintenance access to the six foot girders by offsetting their flanges
(hence Z) to create space between the bottom flanges and webs for access,
but maintained a narrow gap between the top flanges. A filler is usually
provided at the top flanges to close the gap.
The deck is a modified filler beam type construction (but see comment in
Section 7.3 about the design of the deck slab) and the cross girders are bolted to
the web through an end plate connection detail. The slab is reinforced
transversely and longitudinal reinforcement is provided to control cracking.
Structure depth is the depth of the slab (typically 325 mm in the standard
details, which use HEM 160 sections), plus the thickness of the bottom flanges
(which can add up to about 120 mm, including doubler plates).
The web stiffeners of the main girders are on the outside and connected to the
top flange but not to the bottom flange. Walkways are usually provided as
shown in the Figure, supported by cantilever brackets bolted to the main
girders.
The standard design developed by British Rail and Railtrack for the Z Type is
available as a set of standard drawings and a design manual[56]. The latest
design was issued in 1996 (i.e. prior to the latest revision of BS 5400-3). The
deck was designed fully but the drawings only provide details for the main
girders: the actual girder sections have to be designed to suit the actual span
geometrical constraints.
In the 1996 form of the Z type, the top flanges of both main girders project
above ballast level, but do not exceed 110 mm above rail level, thus minimising
the structure width (see Figure 4.6). However, current requirements for the
provision of a robust kerb result in the outer girders being made deeper, to
extend at least 300 mm above rail level. They are thus deeper than the inner
girders and this in turn requires the deck to be widened on the outside.
37
Every third cross girder is connected through a 6-bolt shear plate detail to form
U-frames with the stiffened web; the other cross girders have a 4-bolt
connection. The span range for the Z type is 6 to 17 m.
Protection
to web (paving slab)
Cross girders @ 600 centres
Note that the outer girders are now required to be deeper, extending to 300 mm above
rail level
Figure 4.6
Simplified cross-section of (1996) Z Type deck units

Track with a curved horizontal alignment is usually canted. For low values of
cant the variation in ballast depth is not a problem and Z Type decks are
installed with the soffit horizontal. As the cant increases, the height of ballast
to be retained may approach the top of the girder on the outside of the curve.
To avoid this effect, the decks are often canted to match the inclination of the
track.
The decks are normally simply supported on fabricated steel line pedestal
bearings, as detailed on the standard drawings. If the decks are canted, the
bearings are either increased in height under the outer girder, or located on
bearing plinths with levels adjusted to suit the deck inclination and geometry.
The standard details also include restraints designed to resist soffit collision
loads from highway traffic where headroom is less than 5.7 m.
The deck are normally assembled and concreted prior to installation as single
units. Smaller span decks have been completed at the fabricators premises and
delivered to site; others have been constructed in an adjacent site compound.
Crane installation is the most common method, but transporters have been used
on a number of projects, particularly where overhead electrification systems are
present.
4.3.2 London Undergrounds modified A type

On London Underground, the practice for half through bridges with minimum
construction depth is to utilise a reinforced concrete deck the same depth as the
cross girders, very much like the A type. However, intermediate web stiffeners
are provided on the inner faces of the webs. At U-frame positions, the cross
girders are bolted to the stiffener endplates as well as to the main girder webs.
Haunches are provided against each web to contain track ballast and protect
internal stiffeners; the haunches are cast integral with the deck.
4.3.3 U Type decks

The U type is a development by Cass Hayward and Partners of the Z Type.
The concept replaces the main girders and cross girders with a single piece
U-shaped fabrication, usually with a concrete slab cast on top of the steel floor
to create a composite structure.
The steel floor plate of the composite U type is stiffened transversely with
upstanding ribs and shear connectors are welded to the plate. The floor plate
acts as the tension flange of the main girder. See Figure 4.7. Because there
are few external stiffeners and the soffit is completely free of any stiffening and
attachments, maintenance requirements are significantly reduced. The outer
38
girder is deep enough to act as a robust kerb and, where there are two tracks,
the inner girder is kept below the lower structure gauge.
Steel floor plate

(stiffened on top surface)
Figure 4.7
Cross-section of a composite U Type deck
The structure depth is about 220 - 230 mm and the span range is similar to that
for Z types, i.e. 6 to 17 m.
4.3.4 B and C Type decks

For spans beyond the Z/U Type range, the depth of the main girders needs to
be increased and this can be done by increasing the girder spacing, whilst
keeping the structure outside the lower sector structure gauge (see Section 2.7.3
and Figure 2.4). The top flange level is usually more than 300 mm above rail
level and can be considered to act as a robust kerb to restrain derailed vehicles.
This form of bridge is known as the B type (with two main girders, for single
tracks) and the C type (with three-main girders carrying two tracks). The top
flange width of the central girder of the C type is restricted by the need to
maintain clearance to both tracks. Where U-frames are arranged under both
tracks, the design of the cross girder connections at the centre girder needs to
provide for significant end fixity and the main girder web and stiffeners need to
cater for stress reversals induced by the U-frames as the tracks are alternately
loaded. Alternatives to the double U-frames include an arrangement of
L-frames under each track (with no moment connections to the centre girder
and that girder designed without intermediate restraints) and a combination of
U-frames under one track and L-frames under the other.
Both types use a similar floor construction to the Z type (although with deeper
cross girders for the wider span). The span range for these types is 12 to 23 m
for the B type and 9 to 20 m for the C type. Complete bridges can be installed
as single units by sliding or transporter methods.
The C Type has some disadvantages and two-girder forms are preferred where
possible. However, for replacement bridges the constraints of, for example,
avoiding major change to the loading on the abutments may necessitate use of
the C type.
39
600
203 x 203 UC

Cross section
Figure 4.8
Simplified cross section of a B Type deck
Figure 4.9
Simplified cross section of a C Type deck
4.3.5 D and E Type decks

The D and E type decks were developed by British Rail for larger spans. The
height of the main girders exceeds the platform gauge and the width between is
further increased to provide horizontal clearance to accommodate the tracks and
a safe cess walkway at track level (see discussion in Section 2.7.3(i) and
Figure 2.2). With these arrangements, spans up to 50 m (D type) or 45 m
(E type) can be achieved.
The form of the floor depends on whether structure depth is critical. A
shallower depth is obtained using cross girders at 1.5 m centres and with the top
flanges encased within the slab; the slab is haunched down to bottom flange
level. Haunching helps to reduce the deck self weight, but also pushes the
bottom concrete soffit and longitudinal reinforcement higher, which limits the
effects of global longitudinal bending on crack widths. Structure depth is about
500 mm (plus the bottom flange of the main girder) for the D type and about
600 mm (plus bottom flanges) for the E type.
1500 max.
Figure 4.10 Simplified cross section of a D Type Deck
40
Figure 4.11 Simplified cross section of an E Type Deck

The bottom reinforcement in the slab is threaded through holes in the cross
girder webs. See Figure 4.12.
When the D and E types were developed, the cross girders were connected with
shear plates and a 4-bolt endplate, which minimises erection time during
possessions but is comparatively inefficient for U-frame stiffness. A deeper
connection with stiffened end plates provides a more rigid U-frame.

Reinforcement passing through

cross girder webs
Concrete soffit raised

above bottom flanges
Figure 4.12 Simplified part longitudinal section of D/E Type deck floor
There have been numerous developments of the E type, using different deck
slab constructions and also using an all-steel orthotropic deck (see
Section 4.3.6).
One alternative that has been used where construction depth is not so critical
uses UC section cross girders at up to 3 m spacing with a slab on top, acting
compositely. This gives maximum economy for the slab construction, although
the slab is quite heavily reinforced longitudinally, to span between the beams.
Two-track half through bridges with the main girders reduced in height to fit
below the platform gauge limits and external walkways similar to the B Type
single-track decks at top flange level are often also referred to as narrow
E Type decks. Their advantage is that the cross girder span and construction
depth are reduced compared to the full height E Type and spans of up to 26 m
can be achieved with a structure depth of about 500 mm.
The logistics of fabrication and transport to site often dictate the physical size of
components and spliced connections may be required for construction of very
long span main girders. These bridges are usually constructed alongside/near to
the existing structure and installed by sliding or transporter methods.
4.3.6 Use of orthotropic deck floors

The use of an orthotropic deck, rather than a composite deck, is an alternative
for most of the types discussed above. The floors are usually based on the
fairly heavy orthotropic deck used with the standard box girders (Section 4.4)
For this form of floor, the Tee stiffeners are usually arranged at 600 mm
41
centres and are connected to the main girders with 4 or 6-bolt connections
(U-frames are usually arranged at every third cross girder).
The deck plate is usually kept clear of the web between cross girder positions
and is not assumed to contribute to the bending strength of the main girders.
See details in Section 10. The design of steel deck floors was developed for the
standard box girder bridges (see below) and steel decks used with half through
plate girders are often based on the standard details for the box girder bridges.
Lightweight orthotropic floors (with longitudinal and transverse stiffening
elements) are likely to be susceptible to fatigue damage arising from stress
reversal in the continuous components of the floor and are generally not
preferred. Network Rail also requires such decks to be subject to a full
dynamic analysis for rail speeds in excess of 100 mph.
4.4
Standard box girder bridges

The standard box girder bridge design was developed by British Rail and
Railtrack as an alternative to the E Type deck. The concept aimed to reduce the
span of the deck, and therefore its depth, and to simplify erection by crane
during possessions. Single track and double track versions are detailed on
standard Network Rail drawings[57].
The bridges are half through types, employing trapezoidal steel box girders with
inclined inner webs and orthotropic deck units (sometimes referred to as
battledeck units). Two versions exist for double track (Figure 4.13). The
three girder version uses the same girders as the single track design on the
outside of the two tracks and an additional six foot girder with both webs
sloping located in the space between the tracks.
The cess walkway and continuous position of safety is provided at the top flange
level, rather than at the level of the track. The arrangement of footway is
shown in Figure 4.14. It is convenient to accommodate the services beneath the
walkway, rather than at track level.
A key feature of the behaviour of the box girder type is that the floor of the
bridge is simply supported between the box girders to avoid end fixity effects
being generated and transferred to the box girder web. The virtual pinned
connection is achieved through a special type of shear plate detail that allows
relative rotation between the ends of the cross girders and the box girder. As
well as the very great benefit to fatigue performance (see comment in
Section 8.6.3), the pinned connection offers the benefit that highly skew
arrangements can be achieved without the complications in trimmer details
inherent in other forms of construction. The box girders have the torsional and
distortional stiffness necessary to carry the eccentric loading applied through the
shear plate and the box girders are supported on linear rocker bearings beneath
the inner sloping web.
42
Ballasted
track
Cantilevered
walkway
Single track
design
Main box girder
Tee rib deck units
Three girder
design

Two girder
double track
design
Figure 4.13 Simplified cross sections of the three design ranges

covered by the standard box girder bridge design
Structure depths vary from 260 mm for the single-track and three-girder
designs, to over 350 mm for the wide two-girder design.
Up to 27 m span, the main girders of all three types can be fitted to platform
gauge clearances and walkways are provided at top flange level by cantilevering
from the main girders (Figure 4.14).
Cantilevered walkway
with services below
Main girder below

platform gauge
Figure 4.14 Part section showing box girder with cantilever walkway
The two-girder double track design can be used above
maximum span of 39 m. Within this range, the height
exceeds platform gauge and the deck width is increased
clearance between the girders for a walkway/continuous
ballast level, in a similar manner to the D/E Type decks.
27 m span, up to a
of the main girders
to provide sufficient
position of safety at
The standard drawings (1996 issue) include full details of the components within
the standard decks and substructure loadings such that numerical analysis is not
43
required. The drawings cater for the span ranges mentioned above, various
deck widths and skews of up to 55 degrees.

The box girders form confined spaces and, for smaller spans (which need
smaller boxes), other structural configurations should be considered wherever
possible. Entry to the box girders is obtained at the girder ends and the
standard drawings contain designs for inspection units, which either bolt onto
the ends of the girders, or are integral with them, cantilevering back from the
girder ends. Two versions are included for side entry from the external
elevation and top entry from track level. It is also acceptable to design an
independent access chamber/pit butting up to the girder ends, as part of the
substructure design. The six foot girders of the standard three girder design
have particularly narrow inspection units at the lower end of the span range and
consideration should be given to increasing the six foot dimension (track
spacing), or changing to an alternative form of construction. Weathering steel
should be considered so that the operations of applying and maintaining a
protective coating within a confined space can be eliminated (although a coating
is applied to the lower half of the inside of the box before it is closed during
fabrication).
Box girder bridges of larger spans, increased widths and continuous multi-span
girders have also been designed to suit specific sites. Such designs are not
covered by the range of the standard drawings and must be fully analysed.
Consideration should be given to the provision of additional access hatches
within the span of larger span and continuous girders for inspection
requirements.
4.5
Slab-on-beam composite girder bridges
Where construction depth is not critical, steel plate girders or rolled sections
spanning longitudinally, with a reinforced concrete deck acting compositely with
the steel sections, may be employed (i.e. a deck similar to that employed for
highway bridges).
This form of construction can be used for spans from 8 m to over 50 m, in
single span or continuous span construction.
Structure depth to span
proportions usually lie in the range 12 to 15 for simple spans.
The girders are supported on structural bearings and the deck slab can be
designed to support parapets and to provide a robust kerb.
44
Figure 4.15 Cross-section of a composite superstructure

For highway bridges, deck slabs are often cast in situ but this is impractical for
reconstruction during typical railway possessions.
To overcome this,
superstructures have been constructed in advance and installed by sliding, or
transporter methods (see Section 3.5). Precast decks have also been used, with
the shear connection to the main girder completed in situ.

Composite construction with multiple I-section main girders is also used for
railway bridges. See Figure 4.16.
Figure 4.16 Composite deck-type bridge with multiple girders
4.6
Truss girder bridges
The design of truss railway bridges is beyond the scope of this publication, but
the following comments are given, for general information.
Half through and through superstructures with truss girders have been employed
for large span structures and a few new and replacement bridges of this type are
still being built. In most cases, a truss configuration is the only viable option
for a large span. Historically, a wide range of truss forms have been used but
today only Warren Truss, Modified Warren Truss or Bowstring Truss forms are
normally considered.
For spans under about 80 m, half through construction is likely to be used and
U-frame action will need to be developed to restrain the top chords. The details
of the connection of cross girders to bottom chord nodes are inevitably complex
and very careful consideration needs to be given to design for fatigue
endurance. For larger spans, through construction will be used and the top
chords are stabilised laterally by bracing in the plane of the chords.
45
Usually, cross girders are provided at positions of the nodes on the bottom
chords and the floor spans longitudinally between these cross girders.
Historically, stringers or waybeams were provided, but this open form of
construction has disadvantages in having many intricate details, which unless
very carefully designed give rise to fatigue problems, difficulties in maintenance
and openness to noise transmission. Currently there is a strong preference for a
reinforced concrete slab extending between the trusses whilst spanning
longitudinally between cross girders. A robust kerb is provided inside the line
of the trusses to contain derailed vehicles and provide protection to the truss
girders. In some cases, where weight or structural depth is at a premium,
orthotropic decks may be used.
Where through construction with overhead bracing is used, consideration of
vertical clearances must be made to allow clearances to all vehicles identified
for use on the route. In most cases, clearance to overhead electrification
equipment is also required, even if the line is not currently electrified.
Signal sighting may be an issue with deep trusses and may result in wider
spacing of the trusses than is needed simply for clearances.

Because of their physical size, truss girder bridges are normally assembled on
site from a series of individual members/subassemblies and installed by sliding
during possessions. Crane and transporter methods have been used, but are rare
because of the weight involved and time constraints.
4.7
Direct fastening construction
The forms of construction described above all cater for conventional ballasted
track. Where the depth available for construction is severely limited by
headroom and track level constraints, dispensations to use lower depths of
ballast are often necessary and, in some cases, the rails have to be directly
fastened to the structure. Direct fastening versions of all of the above solutions,
with the exception of solid steel slabs are available.
Direct fastening is normally achieved by either mechanical fixings, or
embedding rails within an elastomer in a trough attached to, or cast into the
deck.
Direct fastening is generally avoided, if at all possible, because of the onerous
maintenance requirements.
4.7.1 Direct fastening systems

There are two basic systems:
Mechanical fixings between the deck and rails
Embedding the rails within an elastomeric compound in grooves formed

within the deck.
Mechanical fixings can comprise baseplates, or clip housings that are cast/fixed
using bolts/studs into the deck. Special adjustable sprung baseplates with
elastomeric pads beneath and spring-loaded fixings are often used. They have
the advantage of reducing the stiffness of the track and can also reduce noise.
By eliminating the ballast and the sleepers, construction depth is reduced by
approximately 450 mm compared to conventional track (with 300 mm of ballast
beneath sleepers).
46
A further reduction in construction depth can be achieved where embedded rails

are used. For this track form, the rails are contained within grooves, which are
usually created using steel trough section rail bearers, or cast into concrete
members. They are commonly used for paved tramways. The rails are aligned
within the grooves using wedges of hardened elastomer and then embedded by
pouring liquid elastomer into the trough. The resultant track is more flexible
than mechanically fixed track and is generally less prone to component failure.
The construction depth savings at least equal those with mechanical fixings and
if the grooves can be incorporated into a longitudinal spanning deck, a further
saving of approximately 150 mm is achievable.
4.7.2 Transition arrangements

Normal ballasted track is quite flexible and deforms within acceptable limits
under traffic. Consequently, there is a severe change in flexibility at the start
and end of a length of directly fastened track. The horizontal and vertical
alignment of the track at the approach to such a structure tends to become
uneven under traffic and difficult to maintain. The mechanical/other fixings at
the start and end of the directly fastened track are also subject to alternating
loadings, as the adjacent ballasted track deflects under axle loads.
One solution to the transition problem between ballasted and directly fastened
track is to provide running on slabs. A running on slab, sometimes referred to
as a transition slab, is usually a reinforced concrete slab, butting up to the end
of the bridge deck, supported at one end on the bridge abutment. Guidance
should be sought from the railway authority about transition arrangements.
On London Underground, bridge running-on slabs are deemed unnecessary,
because axle weights and train speeds are significantly lower.
Ballast gluing has also been used to create a transition within ballasted track on
the approaches to existing bridges, by injecting layers of ballast with a resin
compound.
4.7.3 Waybeams
Waybeams are longitudinal members to which rails are directly fastened. They
might be used for some light rail, short span, low speed situations or for
temporary works but are no longer used for mainline railways. For small
spans, waybeams can be used to span longitudinally between abutments and they
can also be used spanning between cross girders in trusses, where open
construction is acceptable.
Steel plate decking is often used in conjunction with waybeams (either side of
the track); the plate must then be designed to cater for derailment.
4.8
Integral construction
Historically, there have been different ways in which bridge movements were
catered for (if at all). For shorter spans, the girders were usually simply seated
on the abutments, with no specific provision for either movement or restraint.
Medium to long spans often had bearings that were intended to allow thermal
movements to occur but in practice they often seized up over time and
unintended integral behaviour occurred. Currently, there is a significant
interest in integral bridge construction for highway bridges, with the intent of
eliminating all movement joints and, where possible, all bearings as well. The
47
superstructure and substructure act together in response to loading and thermal

expansion/contraction.
Fully integral construction, where there is rotational continuity between
substructure and superstructure, is unlikely to be appropriate for replacing
existing railway bridges, because of the difficulties in making the necessary
connections within possessions, the likelihood of critical details being hidden
beneath the track, the maintenance liabilities and the onerous fatigue effects that
such connections would attract.
However, in some circumstances a form of pinned integral construction may
be appropriate, particularly where it is beneficial for the deck structure to act as
a prop between existing abutments that would otherwise have inadequate
stability. This may be the case for some replacement bridges.

Some general guidance on integral construction may be found in an SCI

publication[43], but if integral construction is to be considered for a railway
bridge, careful consideration must be given to the particular features and
requirements for railways and early consultation with the railway authority is
recommended.
48
DESIGN STANDARDS
5.1
HMRI principles and guidance
Railway Safety Principles and Guidance[49] published by HM Railway

Inspectorate (HMRI) gives general recommendations (Part 1) and detailed
guidance relevant to bridges (Part 2 Section A, Guidance on the Infrastructure).
Strictly, this guidance is not mandatory. However, HMRI is required by law
(under the Railways and Other Transport Systems (Approval of Works, Plant
and Equipment) Regulations[32]) to approve most new railway works, including
bridges, and approval is unlikely to be given to departures from the guidance
without good reason. Many of the HMRI Guidance provisions, moreover, are
repeated in Network Rail standards or incorporated by reference into them.
For bridgeworks within stations, reference to the Fire Precautions (Sub-surface
Stations) Regulations[33], Section 12, is required, to determine whether the
structure requires an enhanced fire rating. London Underground Engineering
Standard E1045[21] lists the stations that must comply with these regulations.

5.2
Railway standards
For UK mainline railways, there are two principal classes of document related
to design and construction: Railway Group Standards and Network Rail
Company Standards. London Underground has its own set of Standards.
5.2.1 Railway Group Standards

Railway Group Standards are published by the Railway Safety and Standards
Board (formerly Railway Safety). The organization is a stand-alone company
owned by Network Rail, train operating companies and other interested parties.
The Standards are intended to set out minimum mandatory provisions for
ensuring safety on Network Rail infrastructure and they impose requirements on
train operators and station operators as well as on Network Rail.
In addition to mandatory standards, the Railway Safety and Standards Board
also publishes certain Railway Approved Codes of Practice. These contain
recommendations which are non-mandatory but which are deemed to constitute
suitable and sufficient means of meeting the mandatory requirements of a
Railway Group Standard.
5.2.2 Network Rail Company Standards

Network Rail publishes its own Company Standards (currently many of these
are still in the name of the former Railtrack), which respond to Railway Group
Standards by adding to and/or setting out in greater detail how the requirements
are to be met. The Company Standards are divided into Procedures (generally
management processes) and Specifications (generally more detailed technical
requirements).
Network Rail also publishes a number of non-mandatory Company Codes of
Practice (again, many of these are currently still in the name of Railtrack).
49
5.2.3 London Underground Standards

London Undergrounds Chief Engineers Directorate is responsible for the
preparation and publishing of Engineering Standards.
If standard clearances or any other Engineering Standard cannot be complied
with, a concession to Standards must be obtained using the process described in
Engineering Procedure P2006[28].
5.3
Technical approval
5.3.1 Network Rail

Network Rail operates a technical approval process for bridges and other civil
engineering work, as set out in Company Specification RT/CE/S/003 Technical
approval of design, construction and maintenance of civil engineering
infrastructure[14]. As far as bridge design is concerned, Network Rails process
is similar to (though not identical with) the Highways Agencys technical
approval procedures, with an Approval in Principle (AIP) submission - Form A
- and a Certificate of Design and Checking - Form B.
The AIP submission is a powerful control on design. The designer is required
to list on Form A the standards proposed to be used and any proposed
departures. A description of the proposed work is also required, normally
accompanied by appropriate drawings. The amount of detail called for at this
stage varies from job to job; as a general rule, the more unusual the design the
more information will need to be provided. Form A then has to be signed in
acceptance on behalf of Network Rail, with or without comments on the
proposals prior to detailed design.
5.3.2 London Underground

London Underground operate a similar process for technical approval, as set out
in Engineering Standard E100820].
5.4
Particular railway standards
5.4.1 Railway Group Standards GC/RT5110 and GC/RT5112

GC/RT5110[6] Design requirements for structures sets out general requirements
relating to design of all structures. Apart from some provisions concerning
bridge parapets over electrified railways, the requirements are all fairly high
level, i.e. they specify principles to be followed rather than giving detailed or
quantified rules.
GC/RT5112[7] Loading requirements for the design of bridges is more
prescriptive. It specifies that railway bridges shall generally be designed for
full RU type loading as specified in BD 37/88 Loads for Highway Bridges
although in exceptional cases a lighter loading may be permitted with suitable
justification. Continuous bridges must also be checked for the effects of load
model SW/0 (defined in a diagram) applied to any one or two tracks. An
appendix lists the loads and load effects to be considered (including primary and
secondary traffic loads, superimposed loads, loads due to the environment and
accidental loads) but does not give values.
50
Notwithstanding the documents title, GC/RT5112 also gives requirements for

the deformation of railway bridges under traffic, by reference to a leaflet
776-3R[35], published by the UIC (Union International des Chemins de Fer).
5.4.2 Railway Approved Code of Practice GC/RC5510

GC/RC5510[5] Recommendations for the design of bridges gives a set of
recommendations for the design of bridges and supports the requirements of
GC/RT5110 and GC/RT5112. Although strictly the recommendations are not
mandatory, designers would be wise to follow them unless they have good
reason not to or unless they are specifically instructed not to by Network Rail.
Any departures from the recommendations would need to be identified and
justified in the relevant AIP submission.

Some of the recommendations are general or high level but many are specific
and quantitative. They cover loading, structural design and non-structural
matters (for example clearances, and provision of lineside walkways). In many
cases, they refer to other documents such as Railway Group Standards, British
Standards, DMRB Standards (see Section 5.6) and UIC Leaflets. They cover
train speeds up to 125 mph (200 km/h) but above this speed only in very
general terms, such as the need to seek specialist advice.
Structural design recommendations are generally based on BS 5400. There are
a few modifications recommended to Part 4 and to Part 5 (in the form of the
Yellow Document, see Section 5.6), and to BD 37/88. The current issue of
GC/RC5510 was published before the latest revision of Part 3 but it is implied
that the revised Standard should now be used. Use of the latest revision of
Part 3 is particularly important for designing girders restrained by U-frames and
for design of bearing stiffeners, where the new rules are significantly different.
5.4.3 Network Rail Company Standard for bridge design

Network Rail is currently preparing a Company Standard covering the design of
bridges. The current documents most relevant to designers are those published
by Railway Safety and Standards Board, as given above in 5.4.1 and 5.4.2.
5.4.4 Network Rail Model Clauses

It is arguable whether preparation of contract specifications forms part of design
proper, but there is certainly a connection; it is important that the quality of
materials and workmanship in a finished bridge structure be consistent with the
design assumptions. Network Rail document RT/CE/C/008[12] Model clauses for
specifying civil engineering work is a comprehensive model framework from
which it is recommended that contract specifications be prepared. It is
published as a number of separate material-related sections. Section 90 covers
steelwork for bridges and is intended to be consistent with design in accordance
with GC/RC5510. Section 100 covers bearings (other than steel), Sections 110
to 115 cover waterproofing and Section 80 covers structural concrete
5.5
BS 5400
The design and construction of bridges in the UK is covered by British Standard

BS 5400: Steel, concrete and composite bridges[1]. The document combines
codes of practice to cover the design and construction of bridges and
specifications for the loads, materials and workmanship. It is based on the
51
principles of limit state design. It comprises the following Parts:

Part
Part
Part
Part
Part
Part
Part
1
2
3
4
5
6
7
Part 8
Part 9
Part 10
General statement
Specification for loads
Code of practice for design of steel bridges
Code of practice for design of concrete bridges
Code of practice for design of composite bridges
Specification for materials and workmanship, steel
Specification for materials and workmanship, concrete,
reinforcement and prestressing tendons
Recommendations for materials and workmanship, concrete,
Bridge bearings
Code of practice for fatigue

The general principles of the limit state design approach are given in Part 1; it
states that two limit states are adopted in BS 5400, the ultimate limit state (ULS)
and the serviceability limit state (SLS).
Part 2 specifies loads that are to be taken into account in the design. Parts 3, 4,
5 and 10 are Codes, which are manuals of good practice for the design of
bridges. Implicit in the Codes is the assumption that workmanship and
materials will be in accordance with the Specifications of Parts 6 and 7. These
two Parts are written in a form suitable for incorporation in contract documents.
In particular, Part 6 provides a comprehensive specification for the various
forms of steel (plates, sections, bolts, welds, etc.) and the quality of
workmanship employed in fabrication and erection.
In Part 3 (as amended in 2000), reference is made to a number of product
standards for the steel material, the most commonly recognised of which is
BS EN 10025[2]. Part 6 refers to these and other supporting standards for
materials, workmanship, inspection and testing, etc. See further comment in
Section 7.5.
5.5.1 Interrelation of Parts 3, 4 and 5 of BS 5400

Part 5 of BS 5400 deals with the design of composite bridges, but provides
detailed requirements only for the interaction between steel and composite
elements. Design of the separate elements is referred to Parts 3 and 4 of the
Standard.
5.5.2 Revised versions of Parts 2 and 5 of BS 5400

Since Parts 2 and 5 of BS 5400 were published by BSI, there have been
technical changes to those documents that have been implemented not through
the revision of the documents but by issue of new documents by the Highways
Agency (those documents are now part of the Design Manual for Roads and
Bridges (DMRB), see below). In both cases, the revised documents are
applicable to railway bridges as well as to highway bridges.
At the time of writing, it is expected that BSI will be reissuing Parts 2 and 5,
updated with the relevant technical revisions in the DMRB documents.
52
5.6
Design Manual for Roads and Bridges
The Design Manual for Roads and Bridges (DMRB) comprises a collection of
Standards and Advice Notes issued by the Overseeing Organisations
responsible for highways in the UK. The documents give guidance to the
designer and provide interpretation and application of BS 5400. The documents
also correct typographical errors in the Standard and amend it where considered
appropriate. Most notable in the latter category are documents BD 37 and
BD 16.
BD 37/88 was issued in 1988 and made significant revisions to the highway
bridge loading; the document included a complete re-presentation of BS 5400-2,
with amendments incorporated. The document is referred to in current railway
standards (notably GC/RT5112 and GC/RC5510) rather than the BSI document.
BD 37 was reissued in 2001 (BD 37/01), including updates to the wind loading
and some small revisions to railway bridge loading. It is expected that the
railway standards will be revised to refer to the latest version of BD 37.

BD 16/82 implements and makes extensive changes to BS 5400-5; the changes

were so extensive that an unofficial compilation of BS 5400-5 and BD 16/82
(sometimes referred to as the Yellow Document) was produced by the
Highways Agency. GC/RC5510 recommends the use of this Yellow Document,
rather than the BSI document. At the time of writing (March 2004), BD 16/82
and the Yellow Document are still current.
There is no general requirement for designers of railway bridges to use or
comply with other documents in the DMRB, although a few of those documents
are specifically referred to. It is recognised, however, that DMRB standards
often represent industry best practice and they may generally be used for
guidance provided that they do not conflict with any particular requirements or
recommendations applicable to railways.
5.7
Matters not covered by standards
Ideally, from the point of view of designers, everything required to design a

bridge would be covered by published standards. Regrettably, however, life is
not as neat as this. For various reasons (e.g. changes in commercial policy,
technological development, increases in speed of trains, recent problems or
incidents or accidents), the railway authority may impose requirements that are
not covered in any published standard.
Designers would therefore be well advised to maintain close liaison with
Network Rail or the appropriate London Underground infrastructure company
during all stages of the design process, especially the early stages leading up to
formal AIP submission. Particular matters in respect of which Network Rail is
known to have current requirements not properly covered in standards include
the following:
Dynamic effects due to passenger trains at speeds above 125 mph

(200 km/h) or heavy freight trains at speeds above 60 mph (100 km/h).
Effects on local deck elements of direct-fastened bridges due to heavy

instantaneous wheel loading caused by wheel flats.
Deck-end uplift of bridges under traffic where the deck overhangs the
bearings appreciably.
53
Design of bridge strike impact protection beams for heavily-skewed

bridges.
Design requirements for bridges on high speed lines must also comply with the
Railways (Interoperability) (High Speed) Regulations[34] and its associated
technical specification for interoperability. Such bridges are outside the scope
of this publication. Advice should be sought from the railway authority.
5.8
Development of Standards

Network Rail and London Underground Standards are subject to a continual

process of review and updating to take account of changing circumstances. The
onus is on the designer to ensure that reference is made to the latest issues of
the relevant documents.
54
DESIGN PROCEDURES
In this Section, some references are made to specific clauses in BS 5400-3 and
in the Yellow Document version of BS 5400-5. For brevity, the clause
references are given in the form 3/1.2.3 and 5*/1.2.3, respectively.
6.1
General

The design process encompasses the whole range of activities from initial
selection of the basic form of the bridge, through the sizing of individual
elements to detailed numerical checking against recognised criteria. Earlier
Sections of this publication have set out functional requirements, practical
considerations for construction and the forms of construction commonly used for
railway bridges. This Section discusses the influence of these requirements and
considerations on the design process and indicates how the key aspects of the
various design standards referred to in Section 5 are taken into account during
design. Further aspects of design, relating to specific details, are discussed in
Section 10.
Adequacy of structural performance is measured against three basic criteria:
strength, fatigue endurance and deformation performance. For railway bridges,
all three criteria can have a strong influence on design, even for very modest
spans. It is therefore important for railway bridge design that all three are
considered from the earliest stages of design.
Generally, railway bridge design in the UK will be in accordance with BS 5400.
The various documents referred to in Section 5 are essentially either
implementation directives or supplementary documentation. The design rules
for structural elements are independent of the bridge type, so general guidance
on application of code rules for steel and composite highway bridges is
applicable to railway bridges. However, the forms of construction and design
loadings are different, so different emphasis is needed. Some of the guidance
below has been extracted from SCI publication P289[50], but the extracts are
limited to those aspects most likely to be relevant to railway bridges; if more
detailed guidance is needed (such as on the design of beams with longitudinal
stiffeners) reference should be made to that publication.
The first step in the design process is to identify and understand the client and
project requirements (this includes obtaining the relevant standards, etc.). A
key aspect of those requirements is the bridge loading; guidance on loading is
given in Section 6.3. Based on the requirements, it will usually be possible to
select an initial form of construction and make an approximate sizing of the
principal elements; guidance is given in Section 6.4.
Before setting out on verification of the adequacy of the initial selection, global
analysis is needed to determine the load effects (forces, moment and
displacements); see Section 6.5. Guidance on verifying adequacy is given in
Sections 7 to 9. Bearing selection/design and interface with the supporting
substructures is also very important; see Sections 7.8 and 7.7 respectively.
55
6.2
Design basis
BS 5400 adopts a limit state approach, in which nominal loads are multiplied
by load factors (fL) to derive design loads. These design loads are then
applied to the structure in isolation or in combination to determine the most
adverse internal forces (bending moments, shears etc.), which are in turn
utilised to calculate resultant design load effects. The design resistance is in
turn based on characteristic material properties, reduced slightly by the
application of another partial factor on material strength (m). The criterion for
structural adequacy is expressed as:
R* S*

i.e. the design resistance R* shall be at least equal to the design load effects S*
see Clauses 2.3 and 5 of Part 1.
In this verification process, in addition to the partial factors m and fL, there is
a third factor f3 to be applied. Unfortunately, there is an inconsistency between
Parts 3, 4 and 5 of BS 5400 in applying that factor. In Part 3 (steel bridges),
the calculation of design resistance involves division by f3 whereas in Part 4
(concrete bridges) the calculation of design load effects involves multiplication
by f3. It is not important whether f3 is applied as a divisor on the strength side
or as a multiplier on the loading side but, clearly, care must be taken when
considering a composite structure that f3 is neither omitted nor applied on both
sides. It is recommended here that f3 always be applied as a divisor on the
strength side.
Attention is also drawn to the different treatments of the partial factor m.
Part 3 gives values for m that are to be applied in various circumstances to
expressions for design strengths (resistances); the factor is explicitly applied in
the calculation. In Part 4, m is often implicitly included in expressions for
design strength (such as ULS moment resistance of a slab).
The following notes offer a fuller explanation of the key terms used in the
above discussion:
Nominal loads (Qk)
are unfactored loads, they are the loads that would be

applied to a structure in a perfect situation, i.e. in the
case where the dimensions, densities and/or applied
loads are known precisely.
Load factors (fL)
are the partial factors applied to the nominal loads to

cater for uncertainty in the value of the applied load
(i.e. a vehicle/train may be overloaded etc.), and for
the probability of coincident loading events occurring
simultaneously (load combination).
Load combination
A combination of types of load that might reasonably

coexist. It is reasonable to design for the worst
credible combinations of loads that might occur, rather
than assuming that the most adverse of each type of
load occurs simultaneously.
(For example, it is
statistically unlikely that an overloaded train runs at full
speed on the coldest day of the year in an extreme
wind.)
56
Design loads (Q*)
are the products of the nominal loads and their load

factors.
Design load effects (S*) are the moments, forces, stresses and displacements
resulting from the response of the structure/ element to
the design loads
Characteristic strength
A value of material strength that can reasonably be

relied upon to be achieved when the structure is
constructed.
Usually the guaranteed or specified
minimum value is used.
A partial factor to allow for uncertainty in achieving

the characteristic strength
f3
A partial factor to allow for other uncertainties for

example in the accuracy of model analysis.

6.3
Loading
The loading to be used for railway bridge design is essentially that specified in
BS 5400-2. However, as mentioned in Sections 5.5 and 5.6, that document was
modified by BD 37/88 and more recently by the updated BD 37/01[40]. For
mainline railways, document GC/RT5112 specifies that BD 37/88 shall be used
and document GC/RC5510 gives recommendations that further modify
BD 37/88. References that are made to Part 2* in this Section refer equally to
BS 5400-2 and to both issues of BD 37 unless otherwise noted.
The loading specification in these documents gives values for all live loads and
defines the basis of deriving dead loads. The specification includes a summary
table of the various load factors that are to be applied in the various load
combinations for each of the load types. An extract from that table, giving only
the factors relevant to railway bridges and modified according to GC/RC5510,
is included here as Table 6.1.
The principal types of loading to be considered are:
Dead load.
Superimposed dead loads.
Railway Live Loading.
Accidental loading.
Each of these types is discussed below.

Further specification clauses relate to loads due to wind and temperature effects.
In any design, all loadings should be considered, although for many railway
bridges, the loading combinations with temperature and wind loads do not
govern.
57
Table 6.1
Clause
number
5.1
Summary of load factors for railway loading
Load
Dead:
Steel
Concrete
5.2
Superimposed
dead:
Ballast
other loads*
5.1.2.2
&
5.2.2.2
5.3
Reduced load factor for dead and

superimposed dead load where this has a
more severe total effect
Wind:
during erection
with dead plus superimposed dead load only,
and for members primarily
resisting wind loads
with dead plus superimposed dead plus other
appropriate combination 2
loads
relieving effect of wind

5.4
Temperature:
restraint to movement,
except frictional
frictional bearing restraint
5.6
effect of temperature
difference
Differential settlement
5.7
Exceptional loads
5.8
Earth pressure:
retained fill and/
or live load
5.9
6.8
7
8
vertical loads
non-vertical loads
relieving effect
Erection: temporary loads
Vehicle collision
loads on bridge
supports and
superstructures:
Foot/cycle track
bridges:
Railway bridges:
ULS: ultimate limit state
Effects on all elements

excepting elastomeric
bearings
Effects on elastomeric
bearings
live load & effects due to
parapet load
type RU and RL, and
SW/0 primary and
secondary live loading
Limit
state
ULS
SLS
ULS
SLS
ULS
SLS
ULS
SLS
ULS
fL to be considered in
1
2
3
1.10 1.10 1.10
1.00 1.00 1.00
1.20 1.20 1.20
1.00 1.00 1.00
1.75 1.75 1.75
1.20 1.20 1.20
1.20 1.20 1.20
1.00 1.00 1.00
1.00 1.00 1.00
ULS
SLS
ULS
SLS
1.10
1.00
1.40
1.00
ULS
SLS
1.10
1.00
ULS
SLS
ULS
SLS
ULS
SLS
ULS
SLS
ULS
SLS
1.00
1.00
ULS
SLS
ULS
SLS
SLS
ULS
SLS
ULS
1.30
1.00
1.30
1.00
1.00
0.80
1.20 1.20 1.20 1.20 1.20
1.00 1.00 1.00 1.00 1.00
to be assessed and agreed between
the engineer and the appropriate
authority
1.20 1.20 1.20 1.20 1.20
1.00 1.00 1.00 1.00 1.00
1.50 1.50 1.50 1.50 1.50
1.00 1.00 1.00 1.00 1.00
1.00 1.00 1.00 1.00 1.00
1.15 1.15
1.00 1.00
1.50
SLS
ULS
SLS
ULS
SLS
combination
4
5
1.10 1.10
1.00 1.00
1.20 1.20
1.00 1.00
1.75 1.75
1.20 1.20
1.20 1.20
1.00 1.00
1.00 1.00
1.00
1.50
1.00
1.40
1.10
1.25
1.00
1.20
1.00
1.25
1.00
1.20
1.00
SLS: serviceability limit state
Shaded values are those given by GC/RC5510

* Other loads includes track and ballast in excess of that between top of sleeper and 300 mm below sleeper
(see GC/RC5510)
NOTE. For loads arising from creep and shrinkage, or from welding and lack of fit, see Parts 3, 4 and 5 of
BS 5400, as appropriate.
6.3.2 Dead load

The dead load is the weight of the structural elements. It can be calculated
from the specified geometry and the densities of the materials. Note that the
Part 2* partial factors on dead load for steel and concrete are varied by
GC/RC5510.
58
6.3.3 Superimposed dead load

Superimposed dead load covers the weight of the waterproofing layer, the
ballast, the track and any walkways, parapets, services etc. Again, this can be
calculated from the specified geometry and densities of the materials. However,
because there may be less control over the extent of the items making up
superimposed dead load during the life of the structure, the partial factors are
usually larger than those applied to dead load.
For main line railways, the density to be used for ballast is given in
GC/RC/5510, which also clarifies the appropriate partial load factors. Ballast
depth is usually taken to the top of the sleepers; allowance should be made for
increased ballast depth where the deck is sloped for drainage or precambered.
Weights of track
Weights of track components are given in RT/CE/C/025 Assessment of
underbridges[13]. For weights of London Underground track components, refer
to Engineering Standard E3314[23].
To avoid double-counting, track weights are often quoted as extra over
values (i.e. the weight of the sleepers, rails, etc.) minus the weight of the
displaced ballast).

6.3.4 Railway live loading

There are two classes of live load primary live load and secondary live load.
The primary live loads are the weight of the traffic (i.e. vertical loads) and the
secondary live loads are horizontal loads that may or may not coexist with the
primary loads. However, because railway secondary live loads frequently do
coexist, both primary and secondary loads must be considered together in
combinations 1, 2 and 3.
For bridges carrying mainline traffic, the applicable nominal loading in known
as Type RU; in some cases an alternative loading, known as SW/0 must also be
considered. For passenger rapid transit railway systems, the applicable nominal
loading in known as Type RL.
For mainline railways, document GC/RT5112 specifies that, in normal
circumstances, bridges be designed for RU loading (currently it refers to RU
loading as specified in BD 37/88). Additionally, for continuous bridges, load
model SW/0 must also be considered.
For London Underground bridges, document E3304[22] specifies that bridges
carrying mainline traffic be designed for RU loading and designates to which
sections of the system the requirement applies. Elsewhere, the design loading is
RL loading, as specified in BD 37/88.
RU and SW/0 load models
Type RU loading is not the weight of a real train but is a simplified model
developed by the International Union of Railways (UIC) that, when applied to a
simply supported beam, would produce load effects approximately equal to or
slightly in excess of those that would be derived if the (static) weight of real
trains were applied. The derivation of the RU load model is given in UIC
document 776-1R[36].
59
Type SW/0 loading is a special loading that only needs to be applied to

continuous bridges, where the RU model does not give the worst loading effects
at the intermediate supports. The derivation of the SW/0 loading is also given
in UIC document 776-1R. The load model should be curtailed if it produces a
more onerous effect.
Details of RU and SW/0 loading are given graphically in Figure 6.1 and
Figure 6.2
250 kN
250 kN
250 kN
250 kN
80 kN/m
80 kN/m
0.8 m
Figure 6.1
1.6 m
1.6 m
15 m
0.8 m
Type RU loading (UIC loading 71)
133 kN/m
Figure 6.2
1.6 m
133 kN/m
5.3 m
15 m
Type SW/0 loading (for continuous spans only)
RL Loading
Type RL loading is also a simplified model that is intended to produce load
effects equivalent to those of real mass-transit and light railway trains. The
loading model is shown in Figure 6.3. For deck elements, Part 2* states that an
alternative of two concentrated loads also needs to be considered.
200 kN
25 kN/m
25 kN/m
100 m
Figure 6.3
Type RL loading
Dynamic effects of vertical loading

RU, RL and SW/0 loads represent the static weight of the traffic. To take
account of the magnification in load effects that may develop under the
approach, passage and departure of a train travelling at speeds up to 200 km/h,
a dynamic factor is applied. (See comment in Section 2.3 on the aspects of
dynamic behaviour that are considered aspects a, b, c and f are taken into
account by the dynamic factor.)
60
For RU loading, the value of the dynamic factor is given in an expression that
includes the length of the influence line for the element being considered,
which is taken to represent the natural frequency of the element. Different
factors are given for moment effects and shear effects, because the
magnification of bending moments is greater than that of shear forces.
For RL loading, two simple factors are given, one for ballasted track, one for
unballasted track, both irrespective of span, but the factor does not allow for
lurching (aspect f in Section 2.3), which is separately allowed for by applying
56% of the load to one rail and 44% to the other rail.
For train operating speeds in excess of 200 km/h (125 mph) the dynamic factors
are not applicable and special consideration is needed. Consult the railway
authority for appropriate requirements.
Secondary live loads
Secondary live loads are those caused by the change in speed/direction of the
train that causes primary loading. The three types of secondary load are:
centrifugal loading, nosing and longitudinal loading (braking/traction).

Centrifugal Loading
The centrifugal load due to a mass travelling around a curve at speed is easily
calculated. In Part 2*, the speed applicable is 10 km/h higher than the highest
speed expected. This gives the formula:
P(vt + 10) f
127r
2
Fc
Where P is the static load, vt is the train speed (km/h) and r is the radius of
curvature (m). The parameter f is a reduction factor that recognizes that trains
travelling in excess of 120 km/h will be lighter (lower mass) and thus the
centrifugal force is less.
Nosing
Lateral oscillation of the train on the track also gives rise to lateral forces. The
value of 100 kN given in Part 2* was deduced for this dynamic wheel/rail
interaction force from measurements of forces on rails. It should be applied on
both straight and curved track.
Longitudinal Loads
Forces due to traction and braking both act along the rails. With continuous
welded track, some of the force is transmitted beyond the bridge (see
Section 6.5.8). Traction and braking differ in that there are usually only a
small number of driving axles, but wheels are braked all along the train. For
long loaded lengths, the braking loads are therefore significantly higher than the
traction loads.
The length to be used in deriving the dynamic factor is given in Table 17 of Part 2.
An addendum to the Table is given in Appendix I of GC/RC5510 but that Addendum
erroneously lists concrete slab decks as elements of battle deck floor; that line of the
table should be considered as separate from the definitions for battle deck floors.
61
Simplified RU and RL loading

As a further simplification of the RU and RL load models, Appendix D of
Part 2* provides tables of equivalent static and dynamic uniformly distributed
loads for bending moments and end shear forces for simply supported beams.
The tabulated loads are the total loads that, when applied uniformly to a simple
span, give exactly the same bending moment and shear force as would RU/RL
loading, disposed along the span so as to achieve maximum effect.
Note that the tables apply only to simply supported beams spanning parallel to
the direction of the track. For cross girders, such as in a typical half through
bridge deck, the RU/RL load model should be applied directly.
Distribution of axle and wheel loads
For the determination of bending moments in transversely spanning elements,
the railway loading may be applied along the lines of the rails (usually taken as
at 1.5 m centres, see page 1). RU loading should be shared equally between
the two rails; RL loading should be shared 56%:44%, as noted above.
Guidance on the local effects due to wheel loading is given in Appendix I of
GC/RC5510.
6.3.5 Accidental loading

Accidental loading can arise from the derailment of a train or from the impact
of a vehicle collision.
Derailed trains
Document GS/RC5510 recommends that robust kerbs be provided to contain
the wheels of derailed trains and gives values of the design load for such kerbs.
Main girders are deemed to act as robust kerbs provided that they are of
sufficient height and are sufficiently restrained laterally.
Clause 8.5 of Part 2* specifies the vertical loading to represent the effects of a
derailed train. This clause applies a static load anywhere within 2 m either
side of the track (in practice, up to the face of the robust kerb, where this is
closer). Note that Clause 7.3.3 of GC/RC5510 says that 19.1 of that document
gives both the horizontal loading on the kerb and the vertical loading for
derailed trains, but 19.1 only gives the value of the horizontal loading; use the
Part 2* value for the vertical loading.
In addition to the loading representing the effects of a derailed train, Clause
4.6.1 of Part 2* also requires a check on the overall stability of the structure in
the event of derailment (as a ULS condition). Damage to local elements such as
walkways and the parapet is acceptable but the whole structure should not
overturn or collapse. (This requirement can dictate the position of main girders
in deck type construction or preclude the use of large side cantilevers.)
Impact from road traffic
Bridges liable to accidental impact from road vehicles should be designed to
resist the appropriate loads specified in GC/RC5510 Appendix J. That
document refers in turn to BD 60/94[42].
For London Underground bridges, refer to BD 37/01 Section 6.8 for impact
loads.
62
6.4
Initial design
6.4.1 Choice of construction type

The choice of the form of construction will usually be constrained by the
available construction depth and the span. The following Tables give some
indicative guidance on the construction depths for various forms of construction
Table 6.2
Construction depths, for spans up to 20 m

Construction depth (mm) for simple span (m) (See Note 1)
Bridge type
Solid steel slabs
Orthotropic deck

Filler beam
12
16
20
850
n/a
n/a
n/a
n/a
950
1100
n/a
n/a
n/a
1000
1200
1400
n/a
(2)
(2)
n/a
(2)
Z types
n/a
1000
1000
1000
n/a
U type
n/a
950
950
950
n/a
B and C types
n/a
1020(2)
1020(2)
1020(2)
n/a
(2)
(2)
1225
1225(2)
D and E types
n/a
n/a
1225
Standard box girder

(single track)
n/a
n/a
960
960
960
Standard box girder

(two track)
n/a
n/a
1100
1100
1100
Composite deck-type
n/a
1500
1800
2100
2500
n/a = not applicable

(1) Construction depths include a depth of 700 mm for track, ballast and waterproofing, assumed
to be 375 mm, 300 mm and 25 mm respectively.
(2) Plus the thickness of the bottom flanges of the main girder and, where applicable, any
clearance between the cross girder soffit and the bottom flange of the main girder.
Table 6.3
Construction depths, for spans exceeding 20 m
Bridge type
Construction depth (mm) for simple span (m) (See Note 1)

20
D and E types
1225
30
(2)
40
(2)
50
(2)
1300
1300
over 50
(2)
1300
1300(2)
Truss girders
n/a
n/a
n/a
1300
1300
Standard box girder

(single track)
960
n/a
n/a
n/a
n/a
Standard box girder

(two track)
1100
1170
1170(3)
1170(3)
1170(3)
Composite deck-type
2500
3000
4000
5000
5000+
n/a = not applicable

(1) Construction depths include a depth of 700 mm for track, ballast and waterproofing, assumed
to be 375 mm, 300 mm and 25 mm respectively.
(2) Plus the thickness of the bottom flanges of the main girder and, where applicable, any
clearance between the cross girder soffit and the bottom flange of the main girder.
(3) Standard design covers spans up to 39 m only
Where the ballast depth can be reduced, the construction depths in the above
tables can be reduced.
See discussion under ballast in Section 2.2.1.
Providing a clear gap between the cross girders and the bottom flanges of B, C,
D, E and Z types will increase the depth slightly.
63
6.4.2 Size of main girders

Up to about 25 m span, fatigue considerations are likely to determine the size of
the main girder flanges and thus the sizes will depend on the fatigue class and
railway traffic. The appropriate limiting stress range can be easily determined
from Part 10 (using the simplified procedure, see Section 8.5) and the bending
moment is easily calculated from the equivalent UDL for RU loading including
dynamic effects, given in Appendix D of Part 2*.
Above 25 m, strength at ULS is likely to govern. The design bending moment
is easily calculated from the equivalent UDL for RU loading, multiplied by the
load factor, and flange sizes may be determined on the basis of yield strength in
tension and a lower strength in compression (typically between about 70% and
80% of yield for half through bridges, depending on the lateral torsional
buckling slenderness that might be expected).
A further consideration with longer relatively shallow main girders on half
through type bridges with composite concrete floors is compliance with the
crack limits at the slab soffit, under combined global and local effects. This
may require greater main girder stiffness to limit global strains. Guidance for
standard Z type girders is given in the related Railtrack User Manual. For
other half through bridge types an early check on this limit should be made.

6.4.3 Skew superstructures

When the railway crosses the road, river or other railway below at a skew,
decisions have to be made at the outset how the support of the superstructure is
to be arranged.
In some cases, the bridge supports can be arranged square to the main girders,
both at abutments and at intermediate supports, and thus a skew structural
configuration is avoided. More commonly, the lines of support are skew to the
main girders. There are then decisions to be made about the arrangement of
cross girders and the twist effects under live load have to be considered.
The arrangement of cross girders in heavily skewed half through bridges needs
particular consideration because some U-frames become L-frames and the end
trimmers may be better arranged either separately supported or pinned (rather
than moment-connected) to the main girders.
In skewed bridges, the twist experienced by rail vehicles, as the bridge deflects
under live load, can be significant and the limitations on twist of the track may
even determine the size of the main girders, in order to provide sufficient
stiffness. Trimmer beams often require intermediate support bearings to reduce
deflection. In composite deck type bridges, the use of plan bracing to the
bottom flange can create an effective box section, which will increase torsional
stiffness significantly and thus reduce twist.
These aspects should be considered at the initial design stage, since they affect
the whole performance of the bridge and poor choices at this stage cannot be
corrected by later detailing refinements.
Some general guidance of the layout of skew railway bridges is given in
GN 1.02[47]. Bridge articulation is discussed in Section 7.8.2.
64
6.4.4 Multiple spans

Where there are multiple spans, a key issue is whether spans should be
independent or continuous over intermediate supports. Factors affecting this
decision may include:
Structural efficiency (in terms of strength), which is improved when spans

are continuous).
Articulation arrangements (see also section 7.8.2), particularly on longer

span bridges, where the cumulative movement on a continuous bridge can
be difficult to accommodate and which might also require special
arrangements in the track such as breather switches.
Transmission of horizontal loads. On flexible intermediate piers it may be

difficult to design to carry the large horizontal loads from braking and
traction on separate spans; on long continuous spans the total load at the
fixed abutment may be very large.
Deflections at intermediate supports (see further comment in Section 9).
Track/bridge interaction (see Section 6.5.8).
The possibility of uplift at the end supports, with multiple spans.

(Depending on the structural configuration, the high live load to dead load
ratio can result in uplift at the end bearings of continuous spans.) Tiedown arrangements to resist uplift under live load are not generally
acceptable.
The ability of the structure and substructures to accommodate differential

settlement, especially in areas of mining subsidence. This may be a
particular problem when reusing existing foundations.
6.4.5 Fatigue and deformation

Either fatigue or deformation performance may well be the governing criterion
for a railway bridge and it is wise to carry out a preliminary evaluation of these
effects before commencing on detailed design, even if it is thought that strength
is likely to be the governing criteria.
6.4.6 Use of weathering steel

As mentioned in Section 2.6, the use of weathering steel may offer benefits in
reducing maintenance requirements. The selection of weathering steel should be
considered at the initial design stage, since the choice will influence costs,
future maintenance regimes and material thickness requirements.
Weathering steel is particularly worthy of consideration for deck type railway
bridges, but less suited to half through bridges, where all surfaces on the track
side of the girder and areas where rain water can lie such as the flanges will
need to be protected against corrosion.
Although the corrosion rate of weathering steel is much lower than of carbon
steel, it cannot be discounted and an allowance for some loss of section over the
life of the bridge must be made. See Section 7.1.6.
The need to detail the superstructure and substructure to deal with run-off from
weathering steel surfaces, in ways that will avoid unsightly rust staining, should
be borne in mind when choosing the structural configuration.
65
6.5
Global analysis
A global analysis is required to establish the maximum forces and moments at

the critical parts of the bridge, under the variety of possible loading conditions.
Local analysis of the deck slab is usually treated separately from the global
analysis.
Even for a simply supported span, it is now common practice to use a computer
analysis to derive load effects and displacements. Software packages are
available over a wide range of sophistication and capability, and the selection of
program will usually depend on the designers in-house computing facilities.
6.5.1 Types of computer model

The basis of most commonly used computer models is the grillage analogy, as
described by West [54] and Hambly[48]. In this model, the structure is idealised as
a number of longitudinal and transverse beam elements in a single plane, rigidly
interconnected at nodes. Transverse beams may be orthogonal or skewed with
respect to the longitudinal beams.

For a model of a composite deck-type bridge, each beam element represents

either a composite section (e.g. main girder with associated slab) or a width of
slab (e.g. a transverse element may represent a width of slab equal to the
spacing of the transverse elements).
Grillage models cannot determine U-frame effects in half through bridges.
However, I-section main beams (which are at the edges of the deck) have a very
low torsional stiffness; this effectively makes the deck simply supported along
its edges. The end rotations of the successive cross girders given by the model
can then be used to determine load effects in the U-frames (see Section 7.4.1).
A typical grillage model for a skew half through with composite deck is shown
in Figure 6.4. The main girders are each represented by longitudinal elements,
as mentioned above, and the longitudinal bending of the slab is represented by
intermediate longitudinal beam elements; locating these under the lines of the
rails simplifies application of loading. For skew bridges, the end trimmer is
often simply supported independently of the main beams and this is also shown
in the model.
Main girder
Outer rail, track 2
Inner rail, track 2
Inner rail, track 2
Trimmer beam
Outer rail, track 1

Main girder
Figure 6.4
Grillage model for skew half through bridge
To determine non-buckling U-frame effects directly, linear 3D modelling

software can be used as an alternative to the expressions in BS 5400-3. This
alternative will be appropriate for heavily skewed bridges in particular. In such
an analysis it will be necessary to allow for the flexibility of the connection
between cross girders (or deck slab) and the girders; the stiffness of the
effective web stiffeners that form part of the U-frame must also be modelled.
66
For evaluation of the buckling effects in U-frames, the 3D software must have
non-linear or large-displacement analysis capability or the ability to determine
eigenvalue solutions for buckling. (Such analysis could be used as an
alternative to determining slenderness through the simplified rules for uniform
situations that are in Part 3, where agreed by the railway authority.)
The increasing availability of analytical software may lead to wider use of the
more sophisticated models, though at present the use of simple grillages is
strongly predominant, with the U-frame forces (due to both buckling and nonbuckling effects) being calculated in accordance with Part 3 (see Section 7.4.1).
6.5.2 Parameters for global analysis

Half through bridges

The main longitudinal beam elements should coincide with the main girder axes.
It is usual to neglect beneficial contribution of the deck so, for determining
forces in the main girders, non-composite section properties should be used.
However, when considering the SLS crack width checks, the longitudinal shear
connection and fatigue checks on the deck and the longitudinal shear connection,
an effective width of deck (taking account of shear lag) should be included in
determining the beam properties for grillage analysis. (Rules for determining
shear lag are given in Part 3 and Part 5.) Because the whole deck is in tension
longitudinally, tension stiffening in cracked concrete (for a description of this
effect, see Johnson[51]) may be assumed in deriving the effective area of the
reinforcement in the composite section . If a 3D finite element model is used, it
can take account of the interaction between deck elements and the main beams,
provided that suitable element properties are selected.
For the transverse elements, which would normally coincide with the cross
girders, the stiffness is the combined bending stiffness of the beams and slab
(calculated separately), or of the steel deck plate and cross girder.
The stiffness of the intermediate longitudinal beam elements is that for the slab
or for the deck plate, as appropriate.
Torsional stiffness of the slab should be shared between longitudinal and
transverse elements, as for deck type bridges (see below). The main girders
should be given the St Venant torsional stiffness of the I-section, which will be
small and should not generate significant torsional effects.
Grillage models can also be used for half through box girder bridges; the
longitudinal elements should be given the torsional properties of the box and
dummy stubs may be used to cantilever to the ends of the deck elements, which
should be connected with a pin element (or a very short vertical element of very
low bending stiffness), if the deck is pin-jointed to the cross girders.
Deck type bridges
It is usual to choose the longitudinal beam elements so that they coincide with
the main girders. In a deck type model, the transverse beams do not represent
discrete elements and the spacing can be chosen by the designer to facilitate the
analysis. Generally, the spacing should not exceed about 1/8 of the span.
Uniform node spacing should be chosen in each direction where possible.
Gross section properties should be used in the global analysis (Clauses 3/7.2,
5*/5.1.1 and 5*/6.1.4.1). Section properties for longitudinal beams should
67
assume that the fully effective slab is shared between the beams. Long- and
short-term load effects should be determined separately, because they should be
applied separately to long- and short-term section properties in the stress
analysis of sections.
There is no need for the section properties of transverse beam elements to
represent transverse bracing or cross girders when the bracing is not continuous
(i.e. the bracing is only between pairs of girders).
Section properties for transverse beam elements representing the slab alone
should use a width equal to the element spacing. Torsional stiffness of the slab
should be divided equally between the transverse and longitudinal beams; use
bt3/6 in each direction, where b is the width of slab appropriate to the element
concerned.
6.5.3 Choice of cracked or uncracked analysis


Longitudinal shear connection between the slab and the main girders is usually
provided by the cross girder connections. Therefore, the slab will be in tension
in simply supported bridges and in midspan regions in continuous bridges.
Wherever the concrete tensile stress exceeds 0.1 fcu cracked section properties
should be used. See discussion in Section 6.5.2.
Deck type bridges
For the loads applied to a continuous composite structure at ULS, global
analysis may be carried out assuming initially that the concrete slab is uncracked
over internal supports; up to 10% of the support moments may then be
distributed to the span. Alternatively, and more usually, the concrete may be
assumed to be cracked for a length of 15% of the span on each side of an
internal support (Clause 5*/6.1.4). Similarly, the slab may be assumed initially
to be uncracked for SLS and fatigue analysis, but if the concrete stress exceeds
0.1 fcu, either a new analysis assuming cracked concrete at the supports is
required or the midspan moments must be increased without corresponding
reduction in hogging moment (Clause 5*/5.1.1).
Generally, it is advisable for the global analysis for both SLS and ULS to
assume from the outset that the concrete is cracked adjacent to internal supports
for about 15% of the span. No further redistribution should then be made.
Cracked section properties should include the effective area of the reinforcement
over the full width of slab acting with the steel girder. This effective area
should take account of the slightly lower modulus of elasticity of the
reinforcement.
6.5.4 Analysis of staged construction

For half through railway bridges, the beneficial effect of structural interaction
between the deck slab and the main girders is generally neglected in the design
of the main girders, and thus consideration of staged construction is not
necessary for their design. However, separate analysis of composite and noncomposite models will be needed to determine longitudinal shear at SLS and
longitudinal fatigue effects in the slab. With composite cross girders, long- and
68
short-term effects will act on different effective sections and staged construction
should be considered.
Deck type bridges
It is usual for the deck slab of composite deck type bridges to be concreted in
stages, and for the steel girders to be unpropped between supports during this
process, although opportunity for propping may arise where a bridge is
assembled and concreted off-line. Part of the load is thus carried on the steel
beam sections alone, part by the composite sections. A number of separate
analyses may be required, one representing each different stage that occurs.
This series of analyses will follow the concreting sequence and will take account
of the distribution of the weight of wet concrete (Clauses 5*/5.1 and 5*/12.1).
These models will be a series of partially composite structures. Where a
number of installation methodologies are being considered, each will require a
separate series of stage construction models.

6.5.5 Local analysis of the deck slab

Load is transferred from the rails, through ballast, to rectangular patches on the
bridge floor, one under each rail/sleeper connection. The size of the loaded
area is given by dispersal rules in GC/RT5510, Appendix I. The local effects
of this patch loading need to be considered. Local analysis is needed to obtain
influence lines for the bending of the slab or deck plate and for deriving the
loading history for fatigue calculations. For the local analysis, it may be
sufficient to model only part of the bridge deck, using a refined grillage
analysis, finite element model or possibly Pucher charts[52]; alternatively, part of
the global model can be refined.
For a half through deck with cross girders, all cross girders need to be modelled
and a finer mesh of longitudinal elements is needed for evaluating the total
bending locally between cross girders for a steel deck bridge, three lines of
beam elements per rail should be sufficient.
For a deck type bridge, with longitudinal main beams, a finer mesh of
transverse elements should be used a longitudinal spacing equal to the slab
thickness should be sufficient.
6.5.6 Deformations
The total deflections under unfactored dead and superimposed loads should be
calculated to enable the beams to be pre-cambered during fabrication, where this
is required (Network Rail do not require spans less than 12 m to be precambered, see GC/RC5510, clause 19.8.4). This information should be
produced by the designer and included on the drawings. See GN 4.03[47].
Deformations under live load are needed, to check the deformation performance
(see Section 9).
6.5.7 Dynamic response

On high-speed routes (line speed over 200 km/h), there may be interaction
between the vibrational response of the bridge and loading frequency generated
by the passing axles of the train, above that allowed for in the derivation of the
dynamic factors applied to the live loading. Specialist advice should be sought
from the railway authority.
69
6.5.8 Track/bridge interaction for longitudinal loads

As well as the dynamic interaction between the train and the bridge, there may
be track/bridge interaction in resisting longitudinal loads (braking and traction)
and expansion/contraction. See guidance in GC/RC5510, clause 12.1, on when
consideration of this interaction is needed. For bridges not covered, the
document refers to UIC leaflet 774-3R[37] for guidance. Where consideration is
needed, advice should be sought from the railway authority. See further
comment in Section 9.4.2.
Track/bridge interaction should be considered at the initial design stage, as well
as at the detailed design stage.

6.6
Detailed design
The detailed design stage confirms or refines the outline design produced in the
initial design stage. It is essentially a checking process, taking the results of
global analysis from the application of a complete range of loading conditions to
a mathematical model to generate calculated forces, stresses and deflections at
critical locations in the structure. These forces, stresses and deflections should
be checked to ensure that they comply with the various rules and the adequacy
of local detailing should be checked. The detail of the checking process should
be sufficiently thorough to enable working drawings to be prepared, in
conjunction with a specification for workmanship and materials, and the bridge
to be constructed.
Design for strength, fatigue endurance and deformation performance are
discussed in Sections 7, 8 and 9 respectively.
6.7
Design of half through bridges
In half through bridges, the degree of structural interaction between the deck
and the main girders is a key consideration for designers. Generally, practice
has been to ignore the beneficial effects of interaction and to take account of it
where it produces adverse effects on the element being considered. Table 6.4
summarises the advice given earlier in Section 6 and also in later Sections.
70
Table 6.4
Allowing for deck/main girder interaction in half through bridges
ULS effects
Strength of main
girders (longitudinal
bending)
For global bending analysis, use bare steel section properties.

For stress analysis of beam sections, use bare steel section properties.
For global bending analysis of simple spans, use bare steel section
properties. For global bending analysis of continuous spans, use bare
steel section properties in sagging regions and composite properties in
hogging regions. May take account of shear lag.
Axial and bending
For local analysis of slab:
strength of slab
Where the slab is in global compression:
(longitudinally
Use ordinary cracked section properties for bending strength
between cross girders)
Where the slab is in global tension:
Use only reinforcement areas (no concrete) for strength.
May enhance reinforcement area to allow for tension stiffening effect.
Longitudinal shear
between deck and
girders*
SLS effects
Longitudinal shear
between deck and
girders

Crack widths in slab
Deflection of main
girder
For global bending analysis of simple spans, use bare steel section
properties. For global bending analysis of continuous spans, use
composite properties
For properties in global bending analysis:
Take account of shear lag
In global sagging regions use cracked section properties but
reinforcement area may only allow for tension stiffening effect
May consider slab to be uncracked where tensile stress < 0.1 fcu
For stress analysis to derive longitudinal shear, use properties as for
global bending analysis.
For local analysis of slab:
Use ordinary cracked section properties for crack widths.
Use only reinforcement areas (no concrete) for crack widths.
Check first with bare steel properties if the limits are exceeded slightly,
check again with composite section properties (reinforcement in cracked
regions may be enhanced by tension stiffening), taking account of shear
lag.
Fatigue effects
In main girder
For global bending analysis, use bare steel section properties.

For stress analysis use bare steel section properties.
In longitudinal shear
connection*
For global bending analysis, use bare steel section properties

For section properties in stress analysis of cross sections:
Use uncracked properties for stresses due to global effects.
Use ordinary cracked section properties for stresses due to local
bending.
Use only reinforcement areas (no concrete) for stresses in
reinforcement.
May take account of shear lag.
In slab reinforcement
* Logically, the analysis for longitudinal shear should be based on the normal composite section, taking
account of concrete, to produce a safe result. However, existing bridges dating back to the 1950s, which
traditionally only used a standard four-bolt connection to the cross girders, and which were never designed
for longitudinal shear, have not exhibited problems in those connections in practice. This is considered to
stem from the conservatism of the other design assumptions and the unaccounted benefit of the shear
transmission between the concrete and the main girders. Therefore, using the same analysis principles as
for the slab design is regarded as acceptable practice, and simplifies design
71
DESIGN FOR STRENGTH
In this Section, some references are made to specific clauses in BS 5400-3 and
to the Yellow Document version of BS 5400-5. For brevity, the clause
references are given in the form 3/1.2.3 and 5*/1.2.3, respectively.
While bridges must obviously be designed to be strong enough to carry the
maximum loading that is anticipated during the life of the bridge, it must be
remembered that a railway bridge is subjected, throughout its life, to very many
repeated loadings close to that maximum loading. It is often the case that
design for fatigue endurance governs, rather than design for strength.
Deformation of the bridge under load is also a primary concern for the safety of
the railway traffic and the comfort of passengers and deformation considerations
may govern in some situations. See sections 8 and 9 for guidance on design for
fatigue endurance and for deformation performance.

7.1
General
The main longitudinal beams must be designed to provide adequate strength in

bending and shear to resist the combined effects of global bending, local effects
such as compression over bearings, and structural participation with bracing
systems. Transverse beams and deck slabs must also be designed to provide
adequate strength in bending and shear, although their design is likely to be
governed by fatigue.
BS 5400-3 provides two sets of Clauses for evaluating the strength of beams,
one for beams without longitudinal stiffeners (3/9.9) and one for beams with
longitudinal stiffeners (3/9.10 and 3/9.11). The first evaluates the strength of
the beam cross section as a whole, and the second evaluates the strength in
terms of its elements (flange panels and web panels, between stiffeners). In the
present publication, guidance is restricted to parallel flange beams without
longitudinal stiffeners (beams of variable depth are treated as longitudinally
stiffened beams and are therefore also not covered here).
Additionally, for beams without longitudinal stiffeners, Part 3 (and also Part 5)
recognises a difference between sections that it terms compact, where the
plastic moment resistance of the section can be developed, and those that it
terms non-compact, where local buckling of the web or flange limits the
moment resistance to that at first yield in an extreme fibre.
7.1.1 Requirements at ultimate and serviceability limit states

Part 3 states that beams should be designed to satisfy the requirements at ULS
and that, in determining bending resistance at ULS, the effects of shear lag may
be ignored.
Clause 3/9.2.3 states that the requirements at SLS need only be checked for the
structural steel elements when one (or more) of the following conditions arise:
(i)
Shear lag is significant; this arises when the effective breadth ratio is
less than a restricting value given by Clause 3/9.2.3.1(a).
72
(ii)
Tension flange stresses have been redistributed at ULS in accordance

with Clause 3/9.5.5. This redistribution is not usually employed in
railway bridges.
(iii)
An unsymmetric beam is designed as a compact section at ULS (this

may be the case for composite deck-type bridges and for cross girders
in half through bridges).
For composite beams, Part 5 requires that the stresses in reinforcement and
concrete should be checked at SLS, taking account of any co-existent stresses
due to local bending of the slab (Clauses 5*/5.2.4 and 5*/5.2.6). Calculation of
stresses at SLS should allow for the effects of any shear lag (Clauses 3/8.2 and
5*/5.2.3). Crack widths in tensile regions of the slab should be checked at
SLS.
7.1.2 Section classification

Definition of compact and non-compact sections

Compact sections are those where the full plastic moment can be developed
before, and maintained after, the onset of buckling (Clause 3/9.3.7.1). Limits
are given in Part 3 for breadth/thickness ratios for which this requirement is
deemed to be true (Clause 3/9.3.7).
Part 3 does not explicitly define non-compact sections, it merely refers to
them as the alternative to compact sections. There is, nevertheless, a limit to
the outstands of compression flanges that applies to non-compact sections
(Clause 3/9.3.2.1); this limit allows a flange to develop its full yield strength,
although it cannot sustain plastic strain like the flange of a compact section.
Design implications for compact sections
The following remarks apply chiefly to deck-type bridges, although they would
also apply where cross girders in a half through bridge were designed to act
with a slab as a compact section (although this is not usual). In half through
construction, there is usually little or no advantage in utilising compact capacity
for the main girders.
When a plastic moment is developed in a beam, the stress distribution through
the section takes the form of rectangular blocks rather than the usual triangular
distribution (Figure 7.1). The amount by which the plastic moment exceeds the
elastic moment when yield is just reached depends on the relative proportions of
web and flange; the ratio between the two moments is commonly known as the
shape factor.
elastic
plastic
Sagging bending
Figure 7.1
elastic
plastic
Hogging bending
Stress distributions in composite beams
Because the bending resistance of compact sections at ULS is determined by

redistribution of the stresses at any cross section, the design requirement for a
compact section is simply that the plastic moment capacity is adequate to carry
the total moment acting at that cross section at ULS (Clause 5*/6.2.2). It is not
73
necessary to check the summation of the separate ULS stress distributions for
each stage of construction (see Section 7.1.4).
Although the strength of compact sections is determined from a plastic stress
distribution, the distribution of bending moments should still be determined by
elastic analysis. Further, because in continuous bridges the design loading for
the two regions (hogging and sagging) differs, each region may be designed on
its own merits; compact moment resistance may be used in midspan regions of
deck type bridges when non-compact resistance is used at intermediate supports.
If the beam is compact throughout the span, and the member is not too slender,
the effects of differential temperature, creep and settlement may be neglected at
ULS [see Clause 3/9.2.1.3(iv)]. This allowance appears to be based on the
presumption that the secondary moments (due to temperature gradient and
shrinkage) can redistribute in such cases.

Unsymmetric beams designed to plastic moment capacity will reach yield in one
flange before the other and before the full moment resistance is developed.
Consequently, the stress in the more highly stressed flange must be checked
elastically at SLS (Clause 3/9.9.8) to ensure that yield is not exceeded at that
state and taking into account construction in stages for composite beams (see
Section 7.1.4). The effects of temperature gradient and shrinkage must then be
included.
To achieve plastic moment in a composite section, the shear connectors must be
capable of transferring an appropriate shear between the beam and the slab.
It should be noted that steel beams that are compact when acting compositely
with the slab might not be compact when acting alone during construction. In
such a case, the checks for the construction stage must be made on the basis of
non-compact sections.
Design implications for non-compact sections
The design strength of non-compact beam sections is determined by essentially
elastic stress distributions and limiting stresses. In most cases, therefore,
checks need only be made at ULS. Checks at SLS are required in only a few
circumstances (Clause 3/9.2.3.1).
A non-compact composite section must provide sufficient moment resistance for
the total moment acting at that section (determined as a summation, as explained
in section 7.1.4). However, because a non-compact section is not able to
redistribute stresses, a check must also be made on the summation of the
elastically determined stresses.
7.1.3 Resistance of beams to buckling

Lateral torsional buckling (LTB) is a well-known phenomenon in slender beams
that are free to buckle between positions of restraint. True lateral torsional
buckling only occurs where the beam is entirely free between discrete restraint
positions; where there is a continuous deck at tension flange level, or some
intermediate level between flanges, the mode is actually lateral distortional
buckling, but the code treats this as lateral torsional buckling, for simplicity.
In design, lateral torsional buckling behaviour is taken into account by use of a
buckling curve such as that illustrated diagrammatically in Figure 7.2.
74
1.0
MR /Mult
0.0
Figure 7.2
Slenderness
Typical LTB buckling curve

The proportion of the bending resistance of the cross section that can be
developed depends on the beams LTB slenderness, which can be expressed in
terms of an effective length and various section properties. Part 3 gives rules
for determining the effective length of a beam in bending, depending on the
restraint to the compression flange. It provides for four different ways in which
a compression flange can be stabilised against buckling laterally out of its plane.
Continuous direct restraint
The first and most direct form of restraint is when the flange is connected
directly to a deck (e.g. top flange of a single span composite beam and slab
bridge and the midspan portion of a continuous composite bridge). Then the
effective length is zero (Clause 3/9.6.4.2.1) and MR simply equals Mult. (See
Section 7.2.1 for definitions.)
Discrete intermediate lateral and torsional restraints
The second form of restraint is mainly only seen during construction of decktype bridges, when the compression flanges of the bare steel I beams are
typically restrained at discrete positions by means of lateral or torsional
restraints.
Where there is plan bracing to the top flange of a simply supported span, it
provides discrete lateral restraint that are so stiff that they are fully effective,
and then the buckling effective length is equal to the spacing of the restraints
(Clause 3/9.6.4.1.1.1). Similarly, a direct connection of the flange to another
structure (say a parallel length of completed deck) would be fully effective.
However, plan bracing to the compression flange of deck-type spans during
construction is not essential if stiff torsional restraints are provided between two
beams at discrete intervals. This effectively creates a torsional spring restraint
that modifies the buckling mode between support positions. The effective length
with this form of restraint is given by Clause 3/9.6.4.1.2. If the torsional
stiffness provided by the action of the pair of beams is sufficiently high, the
effective length can be restricted to about one-third of the span, but any further
restriction cannot practically be achieved (see the distributed restraints curve in
Figure 8 of Part 3).
75
The bottom flanges adjacent to intermediate supports of deck-type spans are

often restrained at discrete positions by cross bracing acting in conjunction with
the deck slab. This form of restraint is usually so stiff that it provides fully
effective discrete lateral restraints.
Discrete U-frame restraints
In half through construction (other than the Standard Box Girder), the
compression flanges are restrained at discrete positions not by cross bracing but
by the stiff frames formed by the cross girders (acting compositely with the
deck slab) and the web stiffeners. These are referred to as discrete U-frames.
The stiffnesses of such restraints are evaluated according to Clause 3/9.6.4.1.3
and the effective length is determined according to Clause 3/9.6.4.1.1.2.
In determining the stiffness of a U-frame, the factor f for the flexibility of the
connection is usually the dominant parameter. Unfortunately, the example
connections shown in BS 5400-3 for the three classes of flexibility do not bear
much relation to typical U-frame connections. Guidance on the appropriate f
factor for endplate connections is given in Section 10.

Continuous U-frame restraint

In some small bridges, the compression flange may receive restraint from the
U-frame action of a slab (at the level of the tension flange) with an unstiffened
web plate. This is termed continuous U-frame restraint and the effective length
is determined (Clause 3/9.6.4.2.2) by considering the restraint as a series of Uframes, each of unit length.
Beams without intermediate restraint in half through bridges
Because the use of U-frame action in half through bridges leads to additional
design requirements and has fatigue and other drawbacks, for shorter spans and
where space permits, it may sometimes be preferable to widen and thicken the
main girder compression flanges and to rely only on torsional restraint at the
ends of the girders. No intermediate U-frames would be created. If this choice
is made, the consequences on installation arrangements (notably the stability of
girders during erection) must be considered.
Additionally, in the absence of intermediate restraint, alternative measures
should be provided to ensure that the main girder has reasonable robustness in
the event of accidental impact from derailed trains. It may be necessary to
consider protective elements or to enhance the top flange to resist lateral impact
loading.
7.1.4 Composite beams construction in stages

Composite deck-type bridges are usually built in stages: the steel girders are
erected first; the concrete slab is the cast over one part of the bridge; the slab is
then cast successively over the other parts until complete; finally, the
waterproofing, ballast, services and parapets are added. At each stage, the
structure is different, and the distribution of bending moments (in a continuous
bridge) and the interaction between steel and concrete elements are different.
Additionally, the response of the completed structure to live load is different
again, because the concrete is stiffer in response to short-term loading.
Consequently, for deck type bridges, it is necessary to determine moments and
shears at the various different cross sections as summations of a number of
values. This summation is necessary even for beams that are deemed to be
76
compact, although in those cases some of the load effects can be neglected,
because it is assumed that they can redistribute within the structure at ULS.
The total stresses and strains in the fibres of a composite beam where the main
girders are unpropped and the deck slab is added in stages are determined as the
summations of the distributions for each stage (and similarly for short- and
long-term loads). This summation is shown diagrammatically in Figure 7.3.
The position of zero stress will therefore not necessarily correspond with any
particular neutral axis level.
+
Steel
section
Figure 7.3
Long-term
composite
section
+
Short-term
composite
section
=
Total
stress
distribution
Summation of stresses for unpropped staged construction

Where SLS must be checked as well as ULS, the stress distributions for SLS
and ULS must be calculated separately, each using its appropriate set of partial
factors for the various loads.
As mentioned in Section6.7, in half through bridges it is usual to neglect the
beneficial action of the deck slab in tension at ULS but it is necessary to
consider the effect of composite action on the slab and its connection to the
cross girders at SLS and for evaluation of fatigue effects . This will involve
some consideration of staged construction and summation of stresses, in a
similar way to that for deck type bridges although, where the slab is cracked in
tension, there is no distinction between long- and short-term effects.
7.1.5 Restraint of beams at supports

Bracing or some other restraint system is required to restrain the primary
members (the longitudinal beams) and to provide a load path for transferring
horizontal forces (transverse to the main beams), particularly at support
positions.
Generally, beams are required (Clause 3/9.12.5.1) to be provided with torsional
restraints at supports, because they will have been designed to resist lateral
torsional buckling.
These torsional restraints must also be capable of
transferring lateral loads such as wind load, nosing and centrifugal loads down
to the level of the bearings.
In half through bridges, torsional restraint at the support, although usual, is not
always essential, if other means of restraint is provided within the span. See
further discussion in Section 7.4.1. Lateral forces are usually transferred
directly from the bottom flange to the bearings.
7.1.6 Thickness of weathering steel

Where unpainted weathering steel is used, an additional thickness of steel (not
included in the calculation of section resistance) should be provided to cater for
the loss due to slow rusting (the so-called rust patina). The additional
77
thickness on each exposed surface depends on the atmospheric environment;

recommendations are given in clause 4.5.6 of BS 5400-3.
7.2
Design resistance of beams without

longitudinal stiffeners
Bending and shear resistances of a beam without longitudinal stiffeners are

assessed separately and the effects of interaction between the two are taken into
account by expressions that define a limiting envelope for the combined load
effects. Naturally, the combination of bending and shear is greatest at internal
supports of continuous spans, though it is still necessary to check other regions
for combined effects, particularly where there is a change of section.
7.2.1 Bending resistance

In BS 5400-3, the bending resistance MD of a beam without longitudinal

stiffeners is expressed in terms of the limiting moment of resistance MR divided
by partial factors m and f3 (Clause 3/9.9.1.2). The value of MR is determined
by the strength of the cross section Mult multiplied by a reduction factor (where
appropriate) to allow for lateral torsional buckling (Clause 3/9.8 and Figure 11).
For non-compact sections, the value of Mult is the strength at first yield in an
extreme fibre of the cross section (determined from the effective section
properties, with no allowance for shear lag but reduced for holes and slender
webs, Clause 3/9.4.2). For compact sections, the value of Mult is the plastic
moment capacity of the effective section, Mpe.
The use of this approach for compact sections, using a plastic section property
and a curve that is essentially dependent on elastic buckling, is difficult to
envisage in physical terms, but calibration against experimental tests confirms
that it is a reasonable evaluation.
For composite sections in deck-type bridges, the elastic bending resistance is not
normally governed by the strength of either the concrete or the reinforcement,
but the plastic moment capacity does depend on the strength of those elements.
(And, as noted in Section 7.2.2, the value of Mpe is needed even for noncompact sections.) See Section 7.5 for comment on values of design strength
for concrete and reinforcement.
7.2.2 Determining the limiting moment of resistance

Figure 11 gives the value of the ratio MR/Mult for the value of a modified
slenderness parameter. The evaluation of the modified slenderness parameter
requires the designer to calculate first the effective length for lateral torsional
buckling (le), second the slenderness parameter LT and third a modification to
take account of the yield strength and the relationship of Mult to Mpe. The
parameters involved in this process depend on bracing arrangements and
moment variation along the beam (Clauses 3/9.6 and 3/9.7).
Effective length
The effective length for lateral torsional buckling is given by Clause 3/9.6,
depending on the restraint to the compression flange, as discussed in
Section 7.1.3.
78
Slenderness LT
Once the effective length has been determined, the slenderness is calculated
according to Clause 3/9.7. The expression for LT includes a parameter that
allows for any moment variation over the half-wavelength of buckling. Where
there is little variation, the parameter is close to 1.0; at intermediate supports of
continuous deck type bridges, the variation over the half wavelength adjacent to
the support can make a significant reduction in the value of LT (although can
always be taken conservatively as 1.0).
Modified slenderness
As mentioned above, the determination of MR involves the use of the modified
slenderness parameter and a buckling curve. The modification to the value of
the slenderness parameter LT is the multiplication by two ratios. The first ratio
is the value y / 355 . This allows for the actual strength of the material
(Figure 11 of Part 3 is drawn for the specific value of 355 N/mm2 for the yield
strength).

The second ratio is the value
M ult / M pe . For compact sections, this value is
unity. For non-compact sections, the value is less than unity: although it may
seem surprising that this requires the calculation of Mpe for a non-compact
section, the modification is needed because the buckling curve in Figure 11 has
been calibrated, for both compact and non-compact sections, against the value
of Mpe.
The buckling curve in Part 3
Figure 11 in Part 3 is in two parts, one for welded members and one for nonwelded or stress-relieved members (the allowance for residual stress is different
between the two parts). Also, in each Figure, three curves are given, for
different values of the ratio effective length/half-wavelength of buckling (Re/Rw);
the need to consider this ratio arises from the theoretical basis for the buckling
curve.
The derivation of a buckling curve assumes an initial imperfection and this is
usually expressed as a function of the effective buckling length (the PerryRobertson approach). But a strut or a beam can buckle in only a finite number
of half waves and the half-wavelength of buckling may be somewhat greater
than the effective length. This is taken into account in Part 3 by the use of a
series of buckling curves, for different ratios of half-wavelength to effective
length. Where the load effect varies over the span length, it is also important to
take account of the actual half waves, so that the most severely loaded half
wave is checked.
For an ordinary case, with fully effective restraints, the ratio Re/Rw is 1.0.
For a more detailed explanation of the determination of slenderness and limiting
moment of resistance, see Commentary on BS 5400-3:2000, Code of practice for
the design of steel bridges [46].
7.2.3 Treatment of staged construction

As mentioned in Section 7.1.4, the stresses through a non-compact section that
is constructed in stages have to be determined by summation of several stress
distributions. (This applies also to compact beams treated as non-compact at
79
SLS.) The resulting stress variation is therefore non-linear and discontinuous,

and could not be derived from any single application of loads to a single
section.
For these cases, verification of adequate bending resistance at ULS depends on
two requirements: first, that the total moment (the sum of all the stage
moments) should not exceed MD (see Section 7.2.1 for comment on bending
resistance); and second, that the total accumulated stress at any fibre should not
exceed the factored yield strength (Clause 3/9.9.5.4). This second requirement
also applies to the SLS check on compact sections.
7.2.4 Shear resistance

Economic design of longer spans, usually minimising the cross-sectional area of

the steel girder, results in relatively thin webs. Unfortunately, an unstiffened
slender web is unable to develop full shear yield resistance, because its capacity
is limited by buckling. However, when the web is provided with transverse
stiffeners, the buckling resistance is increased. The increase arises firstly from
the constraint of the rectangular panel and secondly because some of the shear is
carried by tension field action. The mechanism for the latter is discussed in
Reference 46.
The design shear resistance VD (Clause 3/9.9.2.2) is the shear capacity of the
web panel, including the contribution to the tension field action from the
flanges.
Intermediate web stiffeners are an effective means of increasing the limiting
shear stress when the web depth/thickness is in excess of about 75. Typically,
intermediate web stiffeners are provided at a spacing of between 1.0d and 2.0d.
The same rules for calculating shear resistance apply to webs in both compact
and non-compact sections. However, the webs in compact sections are less
slender, by definition.
Form of intermediate transverse web stiffeners
The most common form of transverse web stiffening is a flat plate welded on
one face of the web. Such stiffeners are typically used in deck type
construction when the web depth/thickness (d/t) exceeds about 50. A web d/t
up to as much as 200 may be used in a web stiffened only by transverse
stiffeners. These stiffeners are usually attached to the top flange but not to the
bottom flange.
In half through construction, U-frame action normally depends in part on the
stiffness and strength of the web stiffeners that align with the cross girders.
They must therefore be designed for their combined actions as web stiffeners
and U-frame members and must be spaced to suit the cross girders. Generally,
cross girders are likely to be more closely spaced than is necessary either for
stiffening of the slender web or for U-frame action; it is common to position
stiffeners at every second or third cross girder, with the other cross girders
simply bolted to the plain web. In deep girders, Tees rather than flats have
been used on older bridges, but this is not common now.
7.2.5 Bendingshear interaction

Under combined bending and shear, a bending resistance Mf equal to that
provided by the flanges alone, ignoring any contribution from the web, can be
80
Shear force
sustained at the same time as the shear resistance VR if contribution from the
flanges is ignored. Further, it has also been shown that the full bending
resistance MD can be developed if the shear is not more than half of VR, and that
full shear resistance VD can be developed if the bending is not more than half of
Mf. These limits to the interaction are expressed in Clause 3/9.9.3.1 and are
shown graphically in Figure 7.4.
VD
VR
VR
2

Mf
2
Figure 7.4
Mf
MD
Bending moment
Interaction between shear and bending resistance
Bending-shear interaction is checked for the worst moment and worst shear
anywhere within a panel length (i.e. between intermediate web stiffeners),
rather than at a single section. This is only slightly conservative for panels
adjacent to internal supports of continuous beams but seems rather onerous for
midspan regions of beams of compact section in deck-type bridges.
For beams constructed in stages, the moment acting on the section should be
taken as the total moment for sections designed as compact, but for sections
designed as non-compact an effective bending moment must be derived for use
in the interaction formulae. This is obtained by multiplying the extreme fibre
total stress by the modulus for that fibre in the section that is appropriate to the
stage of construction being checked (see Clause 3/9.9.5.5). The designer
should take care to ensure that the fibre for which total design stress is used to
determine the equivalent bending moment is the same fibre that determines the
bending resistance.
7.2.6 Longitudinal shear connection

In the following discussion, references are to clauses in the Yellow Document,
rather than in BS 5400-5 (see Section 5.5.2).
In composite sections, to provide the necessary shear transfer between the steel
girder and the concrete slab, shear connectors are required. The shear flow
varies along the length of the beam, being highest near the supports, and it is
customary to vary the number and spacing of connectors commensurately, for
economy. The most commonly used form of connector is the headed stud,
though channel connectors are sometimes used.
81
Shear connectors must be designed to provide static strength and for fatigue
loading. With non-compact sections, the required resistance at SLS generally
governs the design for static strength. Shear flows should be calculated at
supports, at midspan and at least one position in between, i.e. quarter points.
The ULS need only be considered for non-compact sections when there is uplift
applied to the connector or redistribution of tension flange stresses (Clause
6.3.4). Fatigue may well govern the spacing of connectors in midspan regions.

For compact sections, the shear resistance is checked at ULS by calculating the
shear flow on the basis of elastic section properties, and assuming the concrete
to be uncracked and unreinforced (Clauses 6.3.1 and 5.3.1). This requirement
ensures in practice that there is sufficient connection to develop the plastic
moment resistance.
The nominal static strength of shear connectors is given in Table 7 and the
design static strength in Clauses 5.3.3.6 and 6.3.4. The design of the
connectors must provide a resistance per unit length of at least the maximum
design load shear flow over 10% of the length of the span each side of a
support. In other parts of the span, a series of groups of connectors at constant
spacing may be used to provide a stepped' resistance, subject to the provision
of sufficient total resistance over each length. The maximum calculated shear
flow within the length of any such group must not be more than 10% in excess
of its design resistance per unit length. The design static strengths are given in
Clauses 5.3.2.5 and 6.3.4, and the design procedures in Clauses 5.3.3.5 and
6.3.4. Note that document GC/RC5510 Appendix B modifies the partial factors
on strength for shear connections and allows connectors to be grouped with
wider spacings between groups.
It should be noted that although the code requires consideration of uplift when
there is tension field action (Yellow Document Clause 5.3.3.6), there are no
provisions for the determination of its value, its point of application or the
length over which it may be resisted. Such uplift is usually ignored by
designers.
Transverse reinforcement is required in the slab to provide shear resistance at
ULS in a similar manner to the requirements for shear stud spacing (Clause
6.3.3). The required area of reinforcement is usually provided in multiple beam
construction by continuity from midspan of the slab of the bottom layer of
transverse reinforcement.
In half through construction, as mentioned in Section 6.5.2, it is usual to neglect
composite action at ULS but the adequacy of the longitudinal shear connection
at SLS should be checked. The shear connection may be provided only by the
cross girder connections, in which case they must be designed for the combined
effects, or shear connectors may be provided along the web to help transfer the
longitudinal shear. The fatigue effects of longitudinal shear forces should also
be checked. For cross girders that project beyond the edge of the deck (for
example as shown in Figure 10.17), the design should take into account bending
effects generated in the cross girders by longitudinal shear effects.
7.3
Filler beams
The design of filler beam construction should generally be in accordance with

Clause 8 of the Yellow Document. However, the document does not give any
specific recommendations for determination of transverse moments for railway
82
bridges (clause 5*/8.3 applies only to highway bridges). In the absence of such
recommendations, a grillage model may be used; the application of railway
loading along the lines of the rails is a more direct way of determining the
transverse load effects than the means of allowing for the transverse distribution
of highway loading that is used in clause 5*/8.3.
The bond strength between the steel beams and the concrete is also only given
in Part 5 for highway and footway bridges. For railway bridges, it would seem
appropriate, for regular filler beam construction to use the same value. By
regular is meant the use of longitudinally spanning filler beams, with no
tensile effects transverse to the beams. However, filler beam construction is
used as the deck slab of half through bridges and in such circumstances there is
tensile stress that is transverse to the cross girders. It is considered unreliable
to assume any bond stress in such circumstances (in early Z type bridge decks
significant cracks were found to occur parallel to and at every cross girder).

For regular filler beam bridge design, the rules in Part 5* can generally be
used for railway bridges. It is advisable to discuss the design basis for a filler
beam deck with the railway authority at an early stage; reference to more recent
documents than BS 5400 may be agreed (for example, guidance can be obtained
from a UIC publication[55] and detailed rules for filler beam decks are given in
the draft Eurocode 4[30]). Use of rules other than BS 5400 will require the
acceptance as a Departure from Standard by the railway authority.
7.4
Restraint systems
7.4.1 Bracing in half through bridges

Design forces for U-frames
Intermediate U-frames are to be designed to resist horizontal forces FR acting
normal to the compression flange at its centroid (Clause 3/9.12.3.2). Equal and
opposite forces are assumed to act on the two flanges. In addition, horizontal
forces FC are to be applied, in the same manner as FR, to U-frames that are
subject to vertical loading.
To determine the magnitude of FC, it is necessary to know the rotation at the
ends of the cross girders of the U-frame, assuming them to be simply
supported. Further, the value of the rotation is not the absolute value but the
value relative to the U-frames either side. The relative value is needed to
represent the fact that if all cross girders rotated the same amount, there would
simply be a bodily horizontal displacement of the flanges and FC would be zero.
The relative rotation is then used to determine the value of FC, depending on the
stiffnesses of the U-frame and of the flange in horizontal bending (Clause
3/12.3.3).
A similar requirement applies to end U-frames, except that the forces are
defined in clause 3/9.12.5.2 as FS and FL.
Note that the force FR may be in either direction (inward or outward) and that
whilst the greatest value of FC is likely to be outward (restraint to the U-frame
under greatest load), adjacent frames would experience inward forces and thus it
is easier to assume that design value of FC may be in either direction.
Additional loads on U-frames will occur where walkways are attached to the
main girder web stiffeners that form part of the U-frame. U-frames must also
83
be checked for horizontal loading arising from the action of the main girders as
robust kerbs.
Where Z-section girders are used, the asymmetry of the main girders leads to
additional effects (twist and lateral displacements). These effects are usually
resisted by the U-frames (in addition to the resistance they provide against the
effects of the FR and Fc forces).
Non-uniform beams and restraint systems

Code rules are generally expressed for relatively simple uniform situations,
because the formulation of rules to express complex behaviour can only cover
standardised situations. In half through construction there are two aspects of
non-uniformity that often give rise to questions.
The first is when the beam cross section changes significantly along the span.
This is quite likely when doubler plates are used to enhance the capacity in
midspan and are then curtailed somewhere about a quarter or a fifth of the span
from the support. If the effective U-frames remain the same throughout the
span and there are three or more half waves, then an analysis on the basis of
midspan properties will still give a reasonable value for effective length of
midspan regions. But this value will be an overestimate of the effective length
for the end regions (the main beam being weaker relative to the U-frames, a
greater effective length would be given if its properties were used to determine
effective length). Use of the effective length derived for midspan in conjunction
with beam properties for the end region should be conservative for the end part
of the span. (It is not the case that the weaker compression flange in the end
region offers less lateral restraint to the midspan region, because, in the model,
there is zero lateral moment and zero displacement at the nodes of the half-wave
deflected profile.)
The second is the restraint provided at the ends of skew bridges. The endmost
intermediate frames are likely to be L-frames rather than U-frames and, for high
skews, there may be no end restraint at all. End restraint is not always needed
for obtaining an adequate value for the effective length (this is a change from
the rule in the previous issue of Part 3), but the question arises of how to meet
the design rules for restraints. It is suggested that restraint forces FS are
determined in accordance with 3/9.12.5.2 for the end of the beam and that, if
there is inadequate restraint (i.e. the bearing cannot provide the restraint and
there is no restraint from a trimmer), the beam is then checked for its adequacy
to carry the force (in plan bending of the flange) back to the nearest L-frame.
The L-frame is then checked for its adequacy to carry the force. For the
determination of effective length, L-frames will be stiffer than the equivalent Uframe, so the assumption of uniform stiffness of intermediate U-frames should
be conservative for that purpose.
Additionally, the determination of appropriate Fc forces and other load effects
generated by the flexing of the whole structure is difficult for skew bridges,
particularly near the supports, because the vertical deflections and rotations are
different at the two ends of the cross girder. Account needs to be taken of
differential deflection and of restraint provided by the torsional and warping
stiffness of the main girder.
In determining the forces on L- and U-frames towards the end of main girders,
it is necessary to consider the level of torsional restraint provided by the main
girder to the cross girder that results in partial fixity to the cross girder. The
84
resultant load effects are increased in skew bridges where one end of the L- or
U-frame is near a fixed support and the other end is subjected to main girder
vertical deflection. The connection, intermediate stiffener and end of the cross
girder are designed to resist these effects. It is normal practice to neglect the
beneficial effects of fixity to the end of the cross girder when designing the
central portion of the cross girder.
7.4.2 Bracing in deck type bridges

For simply supported spans, intermediate bracing is not needed to restrain the
main beams, except possibly for the construction condition. For continuous
spans, beams are often braced adjacent to intermediate supports to provide
restraint to the compression flange. To assist in resisting accidental soffit
impact loads, it may be necessary to brace pairs of girders, typically at 3 to 4 m
centres.

Bracing across the full width of a multiple girder bridge will tend to act as a
stiff transverse element when loads are other than uniform across the deck
width. The members of such bracing will be subject to load reversal when
different parts of the deck width are loaded (i.e. loading from one track, then
from the other); the bracing connections may be subject to significant fatigue
loading. A different arrangement with multiple beam bridges is to brace the
beams in pairs, with no bracing in the bays between the pairs. This makes the
transverse bracing non-participating in transverse load distribution.
Stability of the compression flange must be achieved during construction as well
as in service. In sagging regions, the resistance of a composite beam at ULS is
usually governed by the tensile yield of the bottom flange. During concreting,
with the weight of wet concrete carried only by the steel beams, lateral torsional
buckling and the stability of the top (compression) flange may well govern the
design. Adequate bracing to the top flange must be provided for this condition,
although it may be temporary and can be removed after concreting. Transverse
bracing (between a pair of beams) is often sufficient for this purpose, although
plan bracing to the flanges may be required for carrying lateral wind loads in
longer spans (but it adds to the complexity of fabrication and erection, and
should be avoided where possible).
7.5
Material strength and notch toughness
7.5.1 Steelwork
Strength
The most cost-effective grade of structural steel (in terms of strength per unit
cost of material) is grade S355 to BS EN 10025[2]. That grade is almost
exclusively used in highway bridges and for a large number of railway bridges.
However, where the design of a railway bridge is governed by very onerous
fatigue conditions or serviceability limits (deformation, dynamic response or
control of cracking), grade S275 may be sufficient.
Notch toughness
All parts of structural steelwork are required to have adequate notch toughness,
to avoid the possibility of brittle fracture (Clause 3/6.5). Brittle fracture can
initiate from a stress concentration when loading is applied suddenly, if the
material is not sufficiently tough'. The degree of toughness required is
expressed as a Charpy impact value (determined from a test carried out on a
85
sample of material) and the requirement depends on the thickness of the

material, its minimum temperature in service, the stress level and rate of
loading.
The design requirements are expressed in Part 3 in terms of a limiting thickness
for the given conditions and specified material. If the thickness of the part does
not exceed the limit, there is sufficient toughness.
The most obvious condition that needs to be considered is the lowest
temperature that the steel will experience. Minimum bridge temperatures are
specified in Part 2 (or BD 37), Clause 5.4.3. For steel railway bridges (i.e.
Group 2 in clause 5.4.3) minimum bridge temperatures are typically 19 C in
England and 23 C in Scotland.
Other conditions relate to the stresses locally at the detail. Special k-factors' to
account for stress level (k), rate of loading (ks), stress concentration (kg) and
potential fracture initiation site (kd), are given in Clause 3/6.5.3. Multiplying
all four factors together gives a single overall factor k.

Steel material, to one of the Standards listed in Clause 3/6.1.2, will have a
specified minimum Charpy impact value at a given test temperature.
Commonly, steel grade S355J2 to BS EN 10025[2] is chosen for bridge
steelwork, and this has an impact energy of 27 J at 20 C. (The impact
energy is indicated by the J2 part of the designation.)
The maximum permitted thickness of the steel part is given in Clause 3/6.5.4
in terms of an equation relating the overall factor k and design minimum
temperature, the chosen material yield strength and the Charpy test temperature
for the chosen material grade. The maximum thickness is directly proportional
to the value of k.
For the specific case of k = 1.0, a simple tabular presentation of limiting
thickness for the grades of material covered by Clause 3/6.1.2 is given in
Table 3c of Part 3. For grade S355 steel, the table gives thickness limits at
20C of 50 mm for grade J2, 60 mm for grade K2 and 86 mm for grade NL
(to BS EN 10113-2). In the absence of gross stress concentrations and sudden
loading, the actual limit for a component depends on the values of factors k
and kd.
Factor k varies from unity for parts under significant tensile stress to a value of
2 for parts always in compression (see Table 3b of Part 3). Clearly, this means
that parts in compression can be thicker than parts in tension, for a given
toughness grade.
The value of factor kd depends on the likely flaw size at the potential fracture
initiation site (i.e. where a brittle fracture would start). Although brittle
fracture is a different phenomenon from fatigue, details susceptible to fatigue
are also more susceptible to brittle fracture. For that reason, Table 3a of Part 3
refers to the BS 5400-10 fatigue class details in determining the kd factor. Some
class F details and all class F2 and G details lead to a kd value less than unity
The combination of the two factors can lead to a k factor as low as 0.5 for
tension flanges but up to 2 for compression flanges. Consequently, the
thickness limits in Table 3c may be halved or doubled for a particular detail; it
86
may be difficult to obtain a suitable grade of thick S355 material for the former
case.
When the required flange thickness would exceed the toughness limit, doubler
plates can be used, each plate being about half the thickness of a single plate.
This option is also sometimes chosen to avoid the use of the higher quality NL
grade in thick plate. Additionally, Network Rails current practice is to limit
plate thickness to 75 mm, even if thicker flanges could be chosen of a grade to
give sufficient toughness. (It is understood that thicker material may be
acceptable to Network Rail in specific cases).
Note however that the
termination of doubler plates is a class G detail and the toughness requirement is
likely to govern the positions where doubler plates can be curtailed in tension
flanges.
It is not necessary to use the same grade throughout the whole structure thick
tension flanges could be K2 grade while the remainder is J2, for example but
the designer should be aware of the possibilities for confusion if more than one
grade is used.
Note also that, for k = 1, Table 3c is slightly conservative for higher values of
the thickness limit. See GN 3.08[47] for a modified version of Table 3c for use
with thicker material.

7.5.2 Reinforced concrete

Concrete grade 40 is most commonly used for bridge decks and reinforcement
is grade 460 to BS 4449.
7.6
Slab design
Slabs should be checked for adequacy at both ULS and SLS for bending and
shear, including the combination of local and global effects where the slab is
designed to participate compositely with main members. Derailment effects
must also be checked at ULS.
Limiting stresses and crack widths are given in the Yellow Document.
In half through bridges, local (longitudinal) bending arises from spanning
transversely between cross girders and from relative deflection of adjacent cross
girders. The slab will also be subject to longitudinal tensile strains and these
must be considered when checking the crack widths at SLS and the resistance to
local bending at ULS. Tension stiffening may be taken into account , as it has
a beneficial effect on the bending resistance of slabs under tensile loading. See
comment on global analysis in Section 6.5.2.
Slabs spanning transversely in half through decks will only act compositely with
cross girders if there is adequate shear connection. Bond alone cannot be relied
upon to generate composite action; either sufficient shear connectors must be
provided or sufficient reinforcement passed through holes in the webs of the
cross girders. Where there is no composite action, the transverse bending
strength at ULS may be taken as the sum of the separate resistances of the slab
and cross girder; at SLS loads are shared on the basis of equal deflection of the
slab and the cross girder.
87
7.7
Interfaces with substructures
The design of new substructures or the checking of existing substructures must

be compatible with the design assumptions used for the superstructure.
(Particular attention may be needed to ensure compatibility when different
design organisations are involved.) Where an existing substructure is to be
reused, it is important to ensure (by adequate survey and study of existing
documentation) that the substructure is suitable for supporting bearing
configurations, positions and loadings that may be significantly different from
those which existed previously. Particular aspects to check are, for example,
hidden vaults or thinner abutment walls that may exist beneath new bearing
positions. Generally, the objective is usually to ensure that the maximum load
effects from the new structure are no more than from the existing structure.

Often, new sill beams are used to replace non-existent or localised bedstones (or
even simply to speed the installation process. New sill beams may be used to
span over or beyond weaknesses in existing substructures. Allowance should be
made in their design for tolerance in the relative position of new bearings and
existing substructure, to ensure that proper fit is achieved at site. The design of
sill beams must be coordinated with that of any ballast walls, other ballast
retention features, arrangements for bearing inspection/replacement and
walkway approach details.
A key consideration is ensuring satisfactory
waterproofing and drainage details that will keep water away from the bearings.
Document GC/RC5510, clause 7.5.7 and Appendix C, gives detailed advice on
the re-use of existing substructures.
7.8
Bearings
7.8.1 Bearing types

General guidance on the types of bridge bearings that can be used with steel
bridges is given in RT/CE/C/008 Section 100GN[12] and GN 3.03[47].
Traditionally, line rocker bearings and roller bearings have been used with
longer (i.e. span over 15 m) half through plate girders. Elastomeric pot
bearings are now used in many instances, particularly with composite beam and
slab bridges.
Fabricated line rocker bearings
Line rocker bearings allow rotation to occur about an axis along the line of
contact whilst providing rotational restraint about the horizontal axis square to
the line of contact. They have the advantage of being highly durable, not being
reliant on the use of moving parts or on low friction or compressible materials
with limited life. The facility to restrain rotation about one axis can be
advantageous in achieving overturning resistance or stability against buckling at
supports.
Line rocker bearings are usually aligned square to the girders although for small
skews, alignment with the skew may be more suitable (see further comment in
GN 3.03[47]). They are less suitable on highly skewed bridges, except for the
standard box girder bridges (where they provide the only torsional restraint to
the boxes and are aligned square to the girders).
88
Roller bearings
Roller bearings may be needed where torsional restraint to the beam is required
(in similar circumstances to those where line rockers are needed), but
proprietary roller bearings are expensive. Purpose-mode roller bearings can be
designed and fabricated but specialist input should be obtained.
Pot bearings
Elastomeric pot bearings are now commonly used for railway bridges. They
provide rotational freedom transversely and longitudinally and are much cheaper
than an alternative spherical bearing. They are particularly useful in highly
skew half through bridges.
7.8.2 Articulation
The articulation of a bridge the arrangement of restraints provided at the
various bearings is a subject that requires careful consideration. Horizontal
forces need to be properly restrained but expansion and contraction need to be
allowed for. In railway bridges, the longitudinal track bridge interaction needs
special consideration, see clause 12.1 of GC/RC5510. General guidance on
bridge articulation is given in GN 1.04[47].

7.8.3 Holding down details

In some cases, restraint must be provided against uplift at bearings against the
effects of vehicle collision with the bridge soffit. The choice of suitable detail
that does not compromise bearing function, inspection and maintenance requires
careful consideration.
Structural configurations that result in uplift at bearings due to live load should
be avoided.
89
DESIGN FOR FATIGUE ENDURANCE
In this Section, any references to specific clauses are to clauses in BS 5400-10,

unless noted otherwise.
8.1
General
Fatigue is frequently a more significant factor in the design of steel railway

bridges than it is in highway structures. This is because live loads regularly
approach the maximum design loading, because many steel railway bridges have
short spans with a high live/dead load ratio and because floor members carry
predominantly live load, subjected to a large number of axle loads per train. In
such cases, the section sizes of members may be governed by fatigue stress
limitations rather than by ultimate loads. Fatigue should be considered at an
early stage during preliminary design.

Because of fatigue considerations (and also deflection limitations see

Section 2.2.4), it is rarely worth using a steel grade higher than S355 for
railway bridges and in many cases grade S275 is appropriate. Generally,
fatigue considerations tend to determine member sizes for spans under about
25 m.
Fatigue design is usually carried out to BS 5400-10, for which it is necessary to
establish a standard traffic mix (heavy, medium or light) and annual tonnage to
represent the traffic that it is anticipated the bridge will carry during its design
life. This information should be provided by the railway authority. Guidance is
given in GC/RC5510 Appendix D.
Fatigue failure of steel depends on the propagation of cracks in regions that are
subject to fluctuating stress. The fatigue life depends on the size of the initial
imperfection or stress concentration and on the range of the stress variation.
Part 10 allows for the size of imperfection or stress concentration by means of a
comprehensive classification of welded and non-welded details (Table 17). The
designer must identify the appropriate classification for the detail being
considered.
Good detail design, avoiding features with poor fatigue classes and
unpredictable secondary stresses, is more important than refinement of
calculation. The actual fatigue life of the structure is much more sensitive to
stresses at the critical details than to the number of cycles. Careful specification
and supervision of fabrication is essential to ensure that the design assumptions
are realised in practice.
8.2
Methods of assessment
The Code provides two methods for the assessment of fatigue life of railway
bridges. The methods involve different determinations of the effective range of
stress variation. In order of increasing complexity, they are:
Without damage calculation limiting stress range (Clause 9.2).
With damage calculation application of a loading spectrum (Clause 9.3).
90
The first method is most commonly used and is much the quickest, though
somewhat conservative. In some cases, where this method indicates failure by a
small margin, designers may opt to recheck by the second method.
The procedure for the damage-calculation method is complex, although it is
fully described in Part 10. That method would normally be used only on large
and complex structures, or where a non-standard fatigue loading spectrum must
be considered.
8.3
Fatigue loading
Fatigue loading is specified by Part 10 but essentially consists of the repeated

application of proportions of the standard loading, RU or RL loading as
appropriate. The standard loading, including the relevant dynamic factor and
lateral loads, is applied to such parts of the bridge as necessary to determine the
maximum and minimum stresses at the detail being considered. Where the
passage of a train results in more than one stress cycle, the loading is also
progressively applied across the bridge, to determine the magnitude of the
individual stress ranges.

In the simple method (without damage calculation), only the full RU or RL

loading is applied in this way.
In the damage calculation method, numerous proportions of standard loading are
applied as a progressive load across the bridge, on each track. The spectra
given in Part 10 have up to 7 components, so for a twin-track bridge, 14
loading transits must be considered; the procedure is therefore complex,
although designers should be able to make use of customised spreadsheets to
eliminate much of the tedium.
For light railways, a relevant loading spectrum must be determined, for use
with the damage calculation method. Loading events on light railway structures
can be very frequent and it is important not to underestimate the potential
growth in traffic over the life of the bridge.
8.4
Classification of details
Fatigue detail classifications relate to the potential imperfections at welds, holes

or other discontinuities, and their relationship to the direction of the stress
variation. The greater the imperfection, the lower the stress range that can be
tolerated for a given fatigue life. The complete range of classifications is shown
in Table 17 of Part 10. This Table shows a range of typical details in classes
A to G, plus W (for welds) and S (for shear connectors). For a given stress
range, class A has the highest fatigue life, class W the lowest (class S is a
special case).
The attachment of web stiffeners or other short elements not carrying load in
the stressed direction produces a Class F fatigue detail. Doubler plates and
other long welded attachments, such as bearing plates and patch plates
introduce a Class G detail , as do other attachments that are close to the edge of
the flange. Shear connectors produce a Class F detail in the stressed direction
of the flange plate (as well as the Class S detail that requires special attention).
Grip bolted splices introduce Class D details. Cross bracing, which introduces
transverse restraint forces, involves a variety of details, including Class W.
91
Because railway fatigue loading is onerous, particularly on members that

experience a stress cycle for every axle, there is a strong incentive to use the
highest categories of detail that are practicable. For example, for cross girders
it is best from a fatigue point of view to use rolled sections without any
fabricated attachments. However, there is a warning note to Table 17 that
classifications better than Class D require a higher level of workmanship and
inspection than can normally be achieved (i.e. when fabrication is to
BS 5400-6).
To ensure that the workmanship will be appropriate to the fatigue life and detail
class assumed by the designer, Part 10 recommends that all areas where class F
or better is necessary should be marked on the drawings. The marking should
show the fatigue class and the direction of stress to which it relates (see Clause
5.3.1). Workmanship levels, which are specified in Part 6, relate the level of
inspection and the acceptance criteria for imperfections to the minimum fatigue
class required by the designer.
Although deterioration due to fatigue, to a stage where failure occurs, is an
ultimate limit state, as noted in clause 3.2 of Part 1, the partial factors fL and
f3 are both taken as 1.0, as stated in clause 7.1 of Part 10.

8.5
The simplified procedure for fatigue assessment involves the application of

standard railway loading (on up to two tracks) to appropriate lengths of the
point load influence line (for the position being considered) and the
determination of the algebraic maximum and minimum values of stress. The
stress range is then compared with a limiting stress range that depends on the
non-propagating stress range for the class of detail and five modifying factors.
The modifying factors allow for:
Design life (where it is other than 120 years)
The number of loading cycles from the passage of one train
The type of loading (heavy, medium light traffic) and the base length of the
load influence line
The annual tonnage of traffic
The ratio of the stress due to one track loaded to the stress with two tracks
loaded
Tables and expressions are given in Part 10 for each of these modifying factors
and for the value of the non-propagating stress range for the class of detail.
Note that the limiting stress range for unwelded reinforcing bars is given in
Appendix B of GS/RC/5510 (this varies the recommendations in clause 4 of
BS 5400-4).
8.6
Considerations for particular details
8.6.1 Doubler plates

Because railway bridge loading is relatively onerous and constraints on
construction depth result in shallow steel girders, the use of doubler plates to
achieve the necessary flange areas is common. However, the transfer of load
92
into and out of the doubler plates at their ends produces a Class G fatigue detail
for the main flange plate and a class W detail for the weld; the location of this
termination and the detailing of the end of the doubler plate are primary
considerations for the designer.
In a simple span, a doubler plate on the top flange may contribute nearly as
much cross sectional area as the main flange plate; it is essential for strength
over more than half of the span. But its point of termination is likely to depend
more on where the Class G detail is acceptable for fatigue reasons than where
its contribution to bending resistance is needed.
Guidance on detailing of doubler plates is given in Section 10.2.2, page 1.
8.6.2 Deck type bridges

As mentioned in Section 7.4.1, in multiple track deck-type bridges there is

usually a pair of girders under each track, without any bracing between adjacent
pairs of girders. If there were continuous transverse bracing, it would be
subject to loading from every passage of a train. That load would have to be
transferred through the webs of the main girders and would inevitably result in
Class W fatigue details. It is better to avoid this problem by omitting bracing
between adjacent pairs of girders.
With this structural arrangement, the transverse bending effects in the slab are
usually fairly modest and the design of the slab is controlled mainly by practical
considerations of slab depth and strength considerations at ULS and SLS.
At intermediate supports, stress ranges in the bottom flanges can be quite
significant. The welding of a bearing plate to the underside of the flange
introduces a class G detail and this imposes onerous limitations the flange may
need to be sized to keep the stress range within the limits.
8.6.3 Half through bridges

In half through bridges, the cross girders are subject to onerous fatigue loading
(they effectively see every axle load) and, because of constraints on depth,
are likely to be significantly stressed by that loading. Consequently, the beams
must have the highest classification that is practicable; if rolled sections are used
and welding is avoided (other than at the ends), Class C can be achieved in the
tensile regions. On the top flange, Class F is inevitable where there are shear
connectors, but the stress range should be lower. Where there is a steel deck,
fatigue endurance will govern the design of both the Tee ribs and the deck
plate.
At the ends of the cross girders, the necessity for U-frame action to stabilise the
top flange results in stiff connections and poor fatigue class details. Careful
analysis is necessary to determine the magnitude of the moments that are
transferred and the connection can then be detailed accordingly. (Moments are
developed due to Fc and FR forces and to the differential deflection of the main
girders, especially in skew bridges.) Note that welding patch plates to the web
introduces a long attachment (class G) and this may control the sizing of the
flanges (in order to keep stress variations in the web within fatigue limits). If
these long attachments are welded to the top of the tension flange, this will
also affect consideration of toughness (see Section 7.5.1). A continuous
attachment to the web, such as the shear plates on the standard box girders,
only introduces a class D detail.
93
The standard box girder configuration was developed with a virtually

moment free connection between the cross girders and the boxes (the Tee rib
deck units in Figure 4.13 are simply supported at the face of the web). The
pinned connection is needed because the torsional stiffness of the boxes is so
great that, with a more rigid connection, there would be such end restraint to
the cross girders that the moments developed would give major problems for the
fatigue design of the connection and of the web of the box section.
It should be noted that the use of steel deck on closely spaced cross girders
results in more than a single load cycle on the cross girders (and the deck over
them), due to the passage of a single RU train. Consequently, the value of the
factor k2 (one of the factors referred to in Section 8.5 above) is more accurately
determined by use of the reservoir method (that procedure is normally only
used in the with damage calculation method). The passage of the RU train
results in four small stress ranges and one large stress range but the factor k2 is
typically only slightly less than unity when calculated by this method.
8.6.4 Filler beam deck bridges

Rolled sections are normally used in filler beam construction. The drilled holes
through the webs of the beams introduce a Class D detail but this does not
normally govern design.
94
DESIGN FOR DEFORMATION

PERFORMANCE
9.1
General
In comparison with road bridges, bridges supporting rail traffic are subject to
more onerous deformation limits, to ensure the safety and comfort of rail traffic
and passengers. Excessive bridge deformations can endanger traffic by creating
unacceptable changes in vertical and horizontal track geometry, creating
excessive stresses in rails and creating unacceptable vibrations in the bridge
structure. Excessive deformations can also affect the loads imposed on the
track/bridge system, and create conditions which cause passenger discomfort.

The relevant existing standards for deformation limits are GC/RT5112[7],

GC/RC5510[5] and BD37[40], together with the referenced UIC Leaflet 776-3R[35].
It is likely that deformation requirements will be rationalised in the future, as
Network Rail intends to adopt EN1990 Annex A2[31] (at the time of writing this
Annex and the supporting National Annex are still under development). Also,
the UIC has a current project to update the deformation requirements to reflect
the studies carried out by the UIC in support of the development of Annex A2.
Generally, deformations are calculated using nominal loads and, if relevant,
taking into account the deformation of substructures (most substructures are
massive and have negligible influence on the total deformation at track level).
The live load to be taken into account includes vertical loading enhanced by
dynamic factors, centrifugal loads, nosing and longitudinal loads due to traction
and braking. For ballasted decks, effects such as creep and settlement of
foundations may be assumed to be addressed by track maintenance.
9.2
Deflection under permanent loads
The midspan vertical deflection of the bridge under dead and superimposed dead
loads is checked to ensure that the natural frequency of the structure is within
the known limits of validity of the allowances for dynamic load effects in the
dynamic factors used in design. When determining the deflection, nominal
loads should be used (i.e. take FL = 1.0), except that an additional 100 mm of
ballast depth should be taken into account (see GC/RC5510 Appendix E).
The limits on deflection under permanent loads are given in UIC 776-3R, but
note that Appendix I GC/RC5510 gives a correction to the formula for the
upper limit in Figure 1.
The vertical deflection under dead loads is also used to establish the required
camber. See sections 2.4.2 and 6.5.6.
9.3
Deformation due to rail loading
When determining the live load deformation, all tracks should be considered
loaded (if that is more onerous).
95
9.3.1 Vertical deflection at midspan

Vertical deflection at midspan is checked to ensure acceptable vertical track
radii, that the structure is not significantly different in performance to existing
rail bridges and to ensure acceptable levels of vertical acceleration inside
coaches corresponding to satisfactory passenger comfort.
Permitted span/deflection ratios as a function of structural configuration, span
length, speed and passenger comfort are given in UIC Leaflet 776-3R.
GC/RC5510 clause 19.8.2 gives advice on the selection of the appropriate level
of passenger comfort and corresponding table from UIC 776-3R. Although the
Tables in UIC 776-3R allow greater deflection for ordinary lines with lower
speed trains, an upper limit of span/600 should be observed for all Network
Rail bridges (this requirement is already in the draft Annex A2[31]).
9.3.2
Track twist
Track twist under load occurs when, for a cross section normal to the track at
any given position, the deflection of the structure under one rail is different
from that of the other rail. Excessive track twist will cause a train to derail.

Track twist should always be checked on skew bridges. It can also occur, to a
lesser extent, on bridges subject to eccentric live load, e.g. on a twin-track
bridge with one track loaded.
Track twist should be considered for all locations, from off the bridge (no twist)
through the transition region onto the bridge, across the bridge, and through the
transition region off the bridge. Twist is checked along the centreline of each
track over a gauge length of 3 m parallel to the tracks, taking into account the
worst possible combination of tracks loaded and position of rail loading. For
bridges carrying more than one track, identification of the critical loading
pattern requires careful consideration.
The limiting values of track twist are given in UIC Leaflet 776-3R. These
limits apply to the total twist at rail level. Where the track is on a transition
between level and canted track, there is an intended rate of change of cant,
which means that there is an intended twist in the track. The total track twist is
the sum of twist due to structural deformation and any intended track twist.
Twist effects are likely to be particularly severe for highly skewed bridges,
where track twist limitations may govern design, and in regions of skewed
intermediate supports in a series of simply supported spans.
9.3.3 Uplift at bearings

Unrestrained uplift at any bearing is not permissible, to avoid the resultant
vertical displacement of the track and to avoid premature failure of the bearing.
In particular, uplift should be checked for bearings at the acute corners of
highly skewed bridges and at the ends of continuous structures.
See
Sections 4.1 and 6.4.4.
9.3.4 Uplift at the ends of a deck

Uplift at the end of a deck, commonly known as kick-up, occurs where the
deck beneath the track extends (away from the span) beyond the bearings. The
vertical deflection in the span causes a rotation about a transverse axis at the
end of the deck and the small change of gradient at the bearing causes the end
of the deck to deflect upward.
96
Limits are given in UIC776-3R for uplift to avoid destabilising the ballast and
limit uplift forces on track components and ensure acceptable additional stresses
in the rails. Current thinking is that the limits in UIC 776-3R are too lax. To
maintain acceptable track quality, recent European research indicates that the
uplift should not exceed about 2 mm. This limit is likely to be included in a
future Network Rail standard on the design of bridges.
9.3.5 Vertical acceleration of the deck

Special considerations apply to the design of rail bridges for speeds over
200 km/h, to guard against excessive vibrations. Excessive vibrations can lead
to ballast instability and unacceptable reduction in wheel rail contact forces.
The design of these bridges is outside the scope of this guidance and specialist
advice should be sought from the railway authority.
9.4
Deformation due to rail loading and

temperature
9.4.1
Rotation of the ends of the deck about a transverse axis
The rotation at the end of a deck, or the relative rotation between adjacent deck
ends (where there is a series of separate spans) must be limited, to limit the
uplift forces on track components, to ensure acceptable additional stresses in the
rails and to limit angular discontinuities in rail expansion devices and at
switches.
The limits are given in UIC Leaflet 776-3R. Where track/bridge interaction
effects are required to be taken into account (see below), the associated checks
on limiting additional rail stresses may be critical.
For decks with non-ballasted track, the effect of rotation of the end of the deck
and any uplift at the end of the deck need to be taken into account when
checking the load effects on the rail fastenings and compared with the relevant
limit state (including fatigue) performance characteristics of the rail supports and
fastening system.
9.4.2 Track/bridge interaction

Vertical deflection in the span causes a rotation about a transverse axis at the
end of the deck and, depending upon the height of the upper surface of the deck
above the bearing, a corresponding longitudinal displacement. Together with
longitudinal displacement of the substructure, due to traction and braking loads
and temperature contraction and expansion, these actions develop additional
stresses in the rails and additional forces on the bearings. This combined
response of the bridge to variable actions is called track/bridge interaction and
can dictate the design of long span bridges or continuous bridges.
Where track/bridge interaction checks are required to be carried out by
GC/RC5510 (clause 12.1), clause 6.5.4 of BS EN1991-2[4] should be assumed to
supersede the advice in UIC Leaflet 774-3R[37].
9.4.3 Longitudinal displacement of the end of the deck

Unless track/bridge interaction effects are required to be taken into account by
GC/RC5510 clause 12.1, there are no specific requirements to be met.
97
9.5
Lateral deformation
The total lateral deformation of the deck under variable loads (centrifugal load,
nosing, wind load, lateral temperature gradient) should be limited, to ensure
acceptable track geometry and passenger comfort.
Limits are given in UIC Leaflet 776-3R. The limits are defined in terms of the
maximum permitted change in track radius and the maximum change of angle at
the end of a deck. The maximum change of angle is about a vertical axis and
should be assumed to apply to both ends of a deck and to the maximum total
change of angle between adjacent decks.
To avoid the occurrence of resonance between the lateral motion of vehicles on
their suspension and the bridge, the following limits must be observed:
The lateral flexibility of the bridge should not exceed the limit in
GC/RC5510 clause 19.8.3.
The lateral frequency under permanent loads should not be less than the
limit in GC/RC5510 clause 19.8.3.

The above limits are not likely to be critical for short to medium span bridges
with solid decks and high in-plane shear stiffness.
9.6
Effect of deformation on clearances
When checking clearances, allowance should be made for deflection of the

structure. In some situations the requirement to maintain clearances within a
limited range can dictate the design. For example, variations in vertical
clearance and stepping distances between a train and a platform can dictate the
design of an underline rail bridge at a station where the platform is supported on
a separate structure and does not move with the bridge under rail loading.
Specialist advice should be sought in such cases.
98
10
BRIDGE DETAILS
This Section presents a series of details for the various forms of construction
presented in Section 4 and for general aspects common to several forms of
construction. Comments are given in relation to the following considerations:
Fatigue endurance
Fabrication
Installation
Durability
For any detail, consideration should also be given to ensuring ease of inspection
and maintenance for all elements and connections.
All details should be checked for adequacy of access for welding, bolting,
inspection and application of protective treatments

Details that are optimum in relation to one consideration may well not be
optimum in relation to other considerations. Alternative details are presented
where appropriate.
10.1 Shallow deck-type bridges

10.1.1 Filler beam deck
Filler beam decks can be constructed with either UC or UB sections (or the
corresponding HD and HE European sections). The use of UC or HD sections
results in a shallower depth and the spacing between beams is usually less than
the flange width. A typical section is shown in Figure 10.1. Holes drilled in
the beam webs should be at lest 15 mm greater than the size of the rebars.
Transverse reinforcement
Permanent formwork
Holes drilled
through web
Figure 10.1 Cross section of filler beam deck

Since there is no welding to the rolled section, only the drilling of the webs, the
fatigue classification is very good (class C generally, class D at the holes).
Fabrication work is minimal, although for longer spans the rolled sections will need
to be precambered (Network Rail requires precambering for spans over 12 m).
99
Where the deck cannot be precast to full width (permitting erection by lifting in
the complete deck), longitudinal joints will be needed between the partial width
units. This may have to be an in situ infill between the sections either side of
the joint and with wider gaps between UB or HE sections the reinforcement
may need to be arranged with overlapping links to provide a shear key.
10.2 Half through plate girder bridges

10.2.1 General
The key structural components of a half through plate girder bridge are shown
in Figure 10.2. This shows a skew configuration, which is common. The
configuration is applicable to bridges with composite decks, filler beam decks
and all-steel decks. For illustrative purposes, support arrangements for the
trimmers are different at each end and rocker bearings are shown at one end
only (it is common practice for short spans to use line rockers at both ends).
Main girder
Doubler plate

Cross girder forming

part of U-frame
Trimmer girder
Intermediate bearing
(if required for
long trimmer)
Independent
cross-girder
Guided
bearing
Web stifener forming

part of U-frame
Trimmer girder
on independent
bearings at its ends
Rocker
bearing
Web stiffener forming

part of L-frame
(associated cross-girder
trimmed into trimmer girder)
T section
bearing stiffener
Figure 10.2 Components of half through plate girder bridge

(bridge floor not shown)
10.2.2 Features common to most half through plate girder

bridges
Web stiffeners
The intermediate stiffeners are on the outside of the web (to avoid vulnerability
in the event of a derailment and to avoid awkward waterproofing details) and
are usually simple flats, welded to the web and the top flange. The stiffeners
are not attached to the bottom flange. See Figure 10.3. If the top flange
outstand is the same width or narrower than the stiffener, the weld is kept clear
of the edge of the flange by using a quadrant cut-out (minimum radius 40 mm).
100
A cope hole is not necessary at the top flange/web junction; fabricators can
grind the flat locally to miss the continuous longitudinal weld, before the
stiffener is welded to the main girder.
5t w max.

Figure 10.3 Typical web stiffener for half through plate girder
This provides a simple and economical fabrication detail. All welds are fillet
welds and there is no fitting to the web or flange. Using the minimum size of
weld for strength minimises the distortion of the web during welding. The
fabricator should be allowed the option of laying the weld to the lower portion
of the stiffener after the end plate of the cross girder is bolted to the web, to
facilitate fit-up to the end plate.
The detail introduces a class F fatigue detail in the web and top flange.
Avoiding the attachment to the bottom flange allows the bottom flange to be
class D, if there are no discontinuities in the web/flange weld. Application and
maintenance of protective coatings is straightforward.
Stiffeners will probably not be required at every cross girder position but should
be aligned with the cross girders where they do coexist (and thus form Uframes).
Where stiffeners need to be welded to the bottom flange as well, for example to
help transmit large moments from cross girders or to transmit restraint forces
from lateral bracing, fillet welds should be sufficient.
Bearing stiffeners
With I-section girders, Tee-section stiffeners, one either side of the web, are
usually sufficient to transmit the bearing reactions, except where the design
thermal movement (from mean position) exceeds about 30 mm, equivalent to a
length from the fixed bearing of about 50 m. See Figure 10.4(a). Boxed
stiffeners have been used but these can give rise to fabrication/assembly
problems and do not provide significant structural benefit, except where the
torsional stiffness can be used to limit the effective length of the main girder in
a tight clearance situation, such as when located in the six-foot space. Open
stiffeners are therefore preferred. On the inner face of the web, Tees are less
vulnerable than flats in the event of a derailment and are substantially protected
by an RC haunch (such as that shown in Figure 10.20), where one is provided.
101
Open section bearing stiffeners should be fitted to the bottom flange and then
fillet welded all round. This allows much of the ULS reaction to be transmitted
in direct bearing. However, direct bearing cannot be relied upon to transmit
fatigue load effects (see general advice in GN 2.04[47]) and fillet welds will be
sized to suit fatigue limitations. Alternatively, full penetration butt welds can be
specified at the bottom flange although this can lead to distortion of the flange,
particularly with one-sided stiffeners. With box section bearing stiffeners,
fitting is much more difficult to achieve and should not be assumed; full
penetration welds at the bottom are also very difficult to achieve.

With Z-section main girders, an asymmetric arrangement must be adopted. The

detailing of the stiffeners depends greatly on whether the trimmer girder is
connected to the stiffener or whether it is on independent bearings. Examples
where trimmers are simply supported on the bottom flange are shown in
Figure 10.4(b). The alternative shown in that figure is a common detail but it is
difficult to control the squareness of the flange to the web and its flatness,
because of the asymmetric welding along the flange.
CL Bearing
Fitted and
fillet welded
Bearing plate
welded to flange
Bearing
End plate
(where required)
Alternative with
boxed stiffener
a) Symmetrical arrangement with I-section main girder
Fitted and
fillet welded
Pack plate
welded to flange
Alternative with
cranked stiffener
b) Arrangement with Z-section main girder

Figure 10.4 Bearing stiffeners in half through plate girder bridge
102
Doubler plates
Because of limitations on thickness of flange plates to meet toughness
requirements (see Section 7.5.1), the flanges are sometimes fabricated by
welding doubler plates to the primary flange plates. These doubler plates are
often shorter than the primary flange plates, in recognition of the reducing
moments away from midspan. Where the doublers are terminated in the span, a
detail such as that shown in Figure 10.5 should be used.
Minimum edge distance
25 mm (or weld leg length
+ 15 mm if greater)
R = 75
Class G at end
Doubler plate
Plan

2 mm greater than
weld leg length
Detail at end
Figure 10.5 Doubler plate detail at termination
The taper in width and thickness are both needed to ensure a smooth flow of
load between doubler and primary flange and the radius helps to minimise the
potential for fatigue cracking. Nevertheless, the termination introduces a class
G detail into the primary flange plate and the shear through the weld is a class
W detail.
The termination detail is also a potential site for the initiation of brittle fracture
and the termination of doubler plates in tensile regions must be considered very
carefully (see discussion in Section 7.5.1).
If thicker flange plates are acceptable, rather than doublers, the fatigue
classifications would be better, even where the plate is spliced with a full
strength butt weld.
Trimmer girders
Trimmer girders support the end of the deck and may act as part of end Uframes or as part of a bracing system at an end support. On a skew half
through bridge the trimmers also provide end support to trimmed cross girders
that are curtailed because of the skew. Typical details of trimmer to a steel
deck and the connection of a trimmed cross girder are shown in Figure 10.6.
The Figure also shows the provision of a concrete end face to the trimmer - the
use of concrete (with drip feature) to protect the external face of the trimmer
beam and provide the transition interface between the deck and the ballast wall
(see also Section 10.6) is desirable because of the difficult access for
maintenance.
103
Trimmers for two-track half through bridges are usually deeper than the
intermediate cross girders - the extra depth extends below the bottom of the
main girder and the ends may be tapered to the same depth as the intermediate
cross girders. Where there is also a significant skew, an intermediate bearing
(at the midspan of the trimmer) may be provided to limit deflection.
For details of the connection of trimmers to the main girders in half through
bridges, see Figure 10.16. Where no moment connection is provided between
the trimmer and the main girder, the trimmer may sit simply on the bottom
flange (as shown in Figure 10.4) or be supported completely independently of
the main girders.
Deck plate
Tee section
UC section trimmer

Location key
Sectional elevation
Figure 10.6 Connection of trimmer and trimmed cross girders on a

skew bridge with steel deck
Where composite decks are used, connection of cross girders to trimmers of
skew bridges may be made using an end plate detail, as shown in Figure 10.7.
A notched end to the cross girder is simpler than a tapered detail but fatigue at
the re-entrant corner needs to be checked. A minimum radius to the corner
must be specified and the cut should be machined or ground, if necessary, to
ensure a smooth profile.
Outline of trimmer girder
Full penetration
butt welds
Outline of
trimmer girder
Alternative notched detail

Plan
View in direction of arrow
Figure 10.7 Bolted connection of trimmed cross girder to skew trimmer

Fanned cross girders
Cross girders are normally arranged square to the main girders. Where there is
significant skew, they are usually trimmed at the ends of the span, as shown in
Figure 10.6. With a small skew (up to about 25), the endmost cross girders
are often detailed in a fanned arrangement, between the skew trimmer and the
square arrangement in the middle of the span, to avoid trimmed cross girders
meeting trimmers at a very acute angle (see Figure 10.8).
104
Main girder
Trimmer
Fanned cross
girders
Figure 10.8 Arrangement of fanned cross girders
10.2.3 Half through bridges with composite deck

The following discussion and details relate to decks where either:
the cross girders are part of a filler beam deck (encased, but with flanges
exposed, similar to Figure 10.1 but with more widely spaced beams), or
the slab is haunched around the cross girders (such as in Figure 4.12), or
the slab is on top of the cross girders.

Cross girder connections

Bolted stiffened end plate connections to the face of the web are usually
sufficient to transfer the end moments from the cross girders (due to Fc and FR
forces).
There are a number of details that are commonly used for end plate
connections, some of which are illustrated below.
Figure 10.9 Six-bolt stiffened connection to shallow cross girder in a

single-track bridge
The arrangement shown in Figure 10.9 is suitable at web stiffener locations
(where U-frames are created) with shallow cross girders (160 mm to 225 mm
deep) at close centres (600 mm) carrying a single track. The detail shows a
clear gap, ideally of 50 mm, between the cross girder soffit and the top of the
bottom flange; this avoids corrosion traps and allows access for maintenance of
protective coatings etc. although this does add to the overall structure depth.
Where the deck is filler beam construction, the gap below the cross girder is
usually filled and the slab extended down locally to the flange. Bolt heads
105
should normally be on the outer face, especially when the endplate is encased in
the concrete of the deck slab (so that any future failure can be detected).

Where the web is thin (less than 20 mm), a patch plate may need to be welded
to the web to cater for local bending due to the tensile forces transferred by the
bolts; the plate also helps to ensure flatness for fit-up. See Figure 10.10. This
patch plate introduces a class G detail on the web and is thus a lower class than
due to the attachment of the stiffeners. (Note that attachment to the flange
should be always be avoided, since the class G detail would introduce onerous
requirements for toughness of thick flanges in tension see discussion in
Section 7.5.1.) Alternatively, to avoid the use of a patch plate, a thicker web,
than required for shear, may be used; a 20 mm web was sufficient for the 1996
Z type single-track bridge.
Figure 10.10 Six-bolt connection to a thin web

Connections of cross girders for twin tracks and for wider cross girder spacings
may be made using a similar detail to that in Figure 10.10 but the cross girders
will be deeper (up to about 400 mm). More bolts will be required and a larger
stiffener will be needed.
With both these details, shear is transferred in friction at SLS (and for fatigue
design) but the bolts can be assumed to slip into bearing/shear at ULS. The
welds around the end of the cross girder must be designed to transmit both
shear and tension/compression and must be checked as a class W fatigue detail.
The details in Figure 10.9 and Figure 10.10 can usually be assumed to achieve
a flexibility parameter f (clause 3/9.6.4.1.3) of 0.2 1010 rad/N mm.
For the same bridge configuration, cross girders between web stiffener locations
would be connected as shown in Figure 10.11. A four-bolt connection is
sufficient in most cases, for closely spaced single-track cross girders.
In each case, welding the end plate to the cross girder potentially distorts the
end plate (so that its surface is no longer plane) and the fabricator may choose
to machine the plate after welding; a slightly thicker plate (than required by the
designer for strength) will be provided to allow for this machining. End plates
thinner than 25 mm should not be specified.
106
Figure 10.11 Four-bolt connection of single-track cross girder that is

not part of a U-frame

In twin-track bridges, the combination of vertical and longitudinal shear is likely

to be too great for the small number of bolts within the depth of the cross
girder. In such cases a shear plate detail, such as shown in Figure 10.12 may
be used; vertical shear is carried in bearing, longitudinal shear is carried in
friction. This also facilitates erection of cross girders and provides redundancy
against a cross girder dropping in the unlikely event of premature bolt failure.
Figure 10.12 Four-bolt connection using a shear plate

Where the structure depth is tightly constrained, the cross girders may sit on
and be connected to the bottom flange, as shown in Figure 10.13. Bolt spacing
and edge distances should comply with the usual rules in Part 3. To minimise
construction depth , the heads of the bolts should be countersunk in the bottom
flange. The bolts must then be tightened from above with shallow cross
girders, it may be necessary to reduce the width of the top flange locally, to
allow access for the tightening wrench from above the deck. With this detail
the flexibility parameter f (clause 3/9.6.4.1.3) may be taken as
0.1 1010 rad/N mm.
In addition to the cross girder end plate to main girder web bolts being designed
to resist vertical shear, it is good practice to design the main girder web to
bottom flange weld also to resist the local effects of the cross girder reaction.
This provides redundancy to the connection and avoids any risk of overloading
the main girder bottom flange to web weld (one side of which is permanently
hidden from inspection).
107
Figure 10.13 End plate connection bolted to main girder

Instead of flat endplate details, such as illustrated above, bolted shear plate
details, similar to that in Figure 10.12, have been used at U-frame connections,
to speed erection and to reduce the number of bolts, but it is a more costly
detail and can give fit-up problems.

Swan neck cross girders

Where construction depth is very tightly constrained, it may be necessary to use
swan neck cross girders, similar to that shown in Figure 10.14. The soffit of
the cross girder can then be at the same level as the soffit of the main girder
and the depth of the cross girder is minimised by using a heavy bottom flange
and by designing for composite action with the deck slab (which can be
achieved through studs welded on the web of the cross girder or by passing
longitudinal bars through the web).
Figure 10.14 Swan-neck cross girder
108
Typical connection of cross girders in London Underground half through

bridges supporting single tracks
For London Underground half through bridges, web stiffeners are provided on
the inside faces of the webs, protected by reinforced concrete haunches. Flats
are used for the web stiffeners and these stiffeners are terminated at horizontal
endplates, as shown in Figure 10.15. This detail facilitates both vertical and
horizontal bolting of the cross girders to the main girders. After all bolting is
completed, the cross girders are encased in concrete and the haunches cast
integrally with the deck slab.
This connection achieves a flexibility parameter f of 0.2 10-10 rad/Nmm.

For double track bridges on London Underground, limitations on construction

depth usually require three main girders; similar connection details are used.
HSFG
bolts
Figure 10.15 U-frame connection on London Underground bridges

Connection of end trimmers to main girders
If trimmers are expected to provide the torsional restraint to the ends of the
girders, they will require moment connection to the main girders via the bearing
stiffener.
If torsional restraint is provided to the main girders by a line rocker bearing,
either the trimmer should not have a moment connection to the main girder (use
a pinned support on the bottom flange or support the trimmer independently of
the main girders) or, if a rigid connection is used, the rocker bearing must be
designed for the end moment from the trimmer girder.
Examples of trimmer connections are shown diagrammatically in plan in
Figure 10.16.
With a Z-section main girder, a pinned support for the end trimmer can be
provided on the inner ends of line rocker bearings, as shown in outline in
Figure 10.4(b).
109
Trimmer square to main girder

Trimmer at up to 30 skew to main girder
Figure 10.16 Connection of trimmers to main girder (plan views) where

bearings allow lateral rotation
10.2.4 Half through plate girder bridge with battledeck floor

Figure 10.17 presents an isometric detail of the connection of a steel battledeck
floor to an I-section main girder in a half through bridge. This form of
construction is discussed in Section 4.3.6.
110
Ballast plate omitted

Bolted splice in
deck plate
Figure 10.17 Arrangement of battledeck floor and main girder

Connection details at the end of the cross girders are shown in Figure 10.18.
The triangular stiffener is sometimes elongated to provide room for a lifting eye
for the deck panel.
Comments on the various components are as follows:
End plate to cross girder connection
Shear and moment from the cross girder are transmitted through an end plate
that is friction grip bolted to the web.
The thickness of the end plate to the cross girder should not be less than
25 mm, to minimise the distortion during welding and bending under tensile
loading on the upper bolts. Patch plates (welded on the face of the web) will
only be required if the web is thin, 16 mm thick or less. Weld sizes should be
kept to the minimum compatible with strength and fatigue endurance, to
minimise distortion to the end plate. As for cross girders for composite decks,
the fabricator will probably machine the end plates and this would normally be
done after the Tees are welded to the deck plate, to ensure that the cross girders
are to a common length and the faces of the end plates are in a common plane.
111
30 deck plate
End plate
8 No. holes
for M24 HSFG bolts
Tee section
(cut from UB)

Figure 10.18 Connection of battledeck cross girders to main girder

As explained above, allowing a gap between the bottom flange of the cross
girder and the top surface of bottom flange of the plate girder simplifies the
connection detail. The gap should be sufficient to allow all surfaces to be
maintained (at least 50 mm is desirable).
The HSFG bolts are designed against slip at SLS and under fatigue loading, and
the welds between the end plate and the cross girder will probably be designed
by fatigue considerations (Class W detail).
Deck plate
Between cross girders, the deck plate is kept well clear of the main girder web,
for access during installation and maintenance. The internal corners of the plate
should be rounded, minimum radius 75 mm or three times the plate thickness.
Splices in the deck plate are usually bolted, if the panels have to be placed
during a possession. Welded splices can be used if the bridge is preassembled
at site (before lifting or rolling/sliding into position). Welded splices should be
at least 100 mm from the nearest Tee, to keep the weld away from the greatest
fatigue stress range; the site-bolted splices should be located midway between
the Tees, to maximise access for tightening. Splices must be able to transmit
local shears due to derailment loading, so bolted splices have two rows of bolts
on either side of the joint.
112
Holes for M24 countersunk

HSFG bolts
30 deck plate
50 thick
splice plate
Figure 10.19 Bolted splice in deck plate

Where high strength clamp-type fasteners are used, because they are more
quickly installed in the limited possession period, the splice must be offset from
the midway position, with the fasteners on one side located centrally between
the Tees. The fasteners on the side closer to the Tee are installed before the
Tee is welded to the plate; the fasteners on the other side are installed on site.
This arrangement allows space for the larger tightening tools for clamp-type
fasteners.
The fabricator should be allowed the opportunity to vary the length of the deck
panels to suit erection requirements and fabrication facilities (for example for
machining end plates to a common plane)
Alternative detail
Where construction depth is at a premium, the cross girder is placed
immediately on top of the bottom flange, as shown in Figure 10.13. When this
arrangement is chosen, the cross girder bottom flange is preferably bolted to the
bottom flange of the main girder. The edge spacing of the bolts should comply
with the Part 3 rules to prevent interface corrosion.
10.2.5 Waterproofing and ballast retention

In half through bridges, the ballast must effectively be retained in a trough and
the waterproofing must extend up the sides of the trough. It is best to avoid
using the structural elements as the sides of the trough: with a concrete floor,
the slab can be returned up the face of the web (as haunches); with a steel
battledeck, a steel ballast plate is used.
RC haunch
Where the slab is returned up the face of the web, it should be keyed to the web
with shear connectors. The keying is needed to avoid separation, rather than to
generate composite action; 19 mm diameter studs should be adequate for this
purpose. At the top of the wall, a weather flat should be provided, to avoid
interface corrosion of the structural web. See Figure 10.20. The waterproofing
to the deck extends up the face of the haunch as far as the weather flat. The
protective coating system should extend over the full face of the web, to reduce
any risk of corrosion. A sealant compatible with the waterproofing should be
applied between the top of the concrete and the weather strip, and to any
construction joint forming a kicker.
113
At least top of
rail level
Figure 10.20 Arrangement of RC haunch

Note that this detail is not suitable for Z-section main girders, because that
configuration will have been chosen to minimise the width between the webs of
the main girders. On Z type decks the web face is usually protected by
concrete slabs, up to the underside of the top flange.
Ballast plate
With steel decks, a non-structural ballast plate is provided to retain the ballast
and keep both the ballast and any drainage away from the structural
connections.
A common and simple arrangement for a ballast plate on a battledeck, is shown
in Figure 10.21. The ballast plate is fabricated in short sections that can each
be manhandled. The sections are bolted together through flanges at their ends
and these also provide stiffening against the pressure from the ballast. The
plates may simply lie against the web or may be fastened near the top edge.
However, this arrangement has the disadvantage that access to the structural
connections is only possible if the ballast is removed
114
Handle
Protection plates fitted to

ballast plates
Removable ballast plate

with spray applied waterproof membrane
Channel filled with sealant
Spray applied waterproof

membrane to deck

Figure 10.21 Ballast plate (sloping face)

An arrangement that allows ready access for inspection of the connections
(albeit during possessions) would be as shown in Figure 10.22. If the portion
of ballast plate above the top of the ballast is within the overhang of the flange
from the web, it does not need to be designed as a robust kerb.
12 thick removable plate
12.5
12 thick shaped
stiffener
12 thick ballast plate
Deck plate
Tee section
(cut from
UB or UC)
Figure 10.22 Ballast plate (vertical face)

Waterproofing
It is common practice and generally more economical to specify spray applied
waterproofing systems, which, for steel decks, are usually applied in the
fabrication works (except locally at site joints). However, where all the
115
waterproofing system has to be applied on site, during a possession (and thus

time is likely to be at a premium), sheet waterproofing membrane systems may
be an alternative.
Compatible waterproofing membrane and mastic sealants should be specified,
preferably from the same manufacturer, to avoid guarantee / warranty issues.
Corrosion protection coatings to steel need to be compatible with the
waterproofing system.
Additional protective boarding is usually fitted to give protection to the
waterproofing membrane on the surface of the ballast plates and concrete
haunches. This boarding affords additional protection to the membrane during
routine track maintenance and hand or mechanical packing of the permanent
way. The concrete is usually rebated to accommodate the boarding and to
define its extent.
10.3 Standard box girder bridges

All the principal details for the standard box girder bridges are given in the
Network Rail drawings and these should be used where standard box girders are
required.
10.4 Slab-on-beam composite girder bridges

Details for slab on plate girder railway bridges are generally similar to those
used in highway bridges, except that section sizes are likely to be greater and
that more attention needs to be paid to the avoidance of lower class fatigue
details. General guidance on a range of details is available in the Guidance
Notes[47].
Plan bracing to the bottom flanges is not usually required, unless there is a need
to improve torsional stiffness, for example when the span is long and a single
track is eccentric to the pair of girders supporting it or if there is a risk of
vehicle impact from below. Plan bracing would create a pseudo-box that would
limit any twist under traffic that would otherwise be unacceptable. If plan
bracing is needed, any connections to the flange should be bolted, to avoid the
lower class fatigue details associated with the type of welded connections that
would be needed.
10.5 Bearing details

10.5.1 Fabricated line rocker bearings
Line rocker bearings can be fabricated at modest cost and provide a long-life
durable bearing. Typical arrangements are shown in Figure 10.23. The Figure
shows a detail with positional freedom of the lower bearing plate it permits
rapid installation and allows adjustment in plan position before making the
perimeter welds (note that the welds should be laid before the bridge is open to
traffic, or there may be insufficient resistance to longitudinal or lateral loads).
The bearings are fastened to the main girder by means of bolts installed in
threaded holes in the plate that is welded to the underside of the bottom flange.
A similar detail can be used to fix the bearing to a bottom plate that is in turn
116
site welded to a location plate, as illustrated in Figure 10.23. The use of a

site weld allows freedom in positioning the bearing during installation.
For short spans (up to about 18 m) line rocker bearings can be used at both
ends of the span, provided that there is sufficient clearance to the keep strips to
permit thermal expansion/contraction.
CL girder
CL bearing/stiffener
Threaded holes
Top pack plate
Shear key
Top bearing plate

Bottom bearing
plate with 6 No
27 holes
Radiused bearing
all faces machined
Keep strip
Site weld

Bottom pack plate

with 6 No threaded
holes for M24 bolts
Elevation
Figure 10.23 Fabricated line rocker bearing
10.5.2 Roller bearings

Where roller bearings are required, proprietary bearings should normally be
specified.
10.5.3 Pot bearings

Pot bearings provide support without rotational restraint. They are commonly
used on deck-type bridges and on longer half through bridges. Proprietary
bearings should normally be specified and these will be fixed to top plates on
the main girders and to bottom plates on the bearing shelf. For installation
during a possession, the same freedom in plan position can be achieved by site
welding the bottom plate, as for rocker bearings.
10.6 Deck end waterproofing details

The completion of the waterproofing system from the bridge to the abutment is
usually a critical path activity.
Installation of the components and
accommodation of deviations from nominal geometry are key issues.
The use of pre-cast concrete ballast walls (on the abutment) are preferred by
many railway contractors, particularly if the deck is being lifted into position as
a single fully assembled element. The ballast walls can be fitted following the
installation of the deck, allowing any tolerance issues between the deck and the
sill beams to be taken up by the ballast walls. The ballast walls can be adjusted
for line and level on mortar beds and fastened using locating bars. An
important detail is then the continuity of the waterproofing membrane from the
deck to the ballast wall. Various details have been developed for this and one
117
example is shown simply in Figure 10.24. This shows a small bridging plate
across the movement joint and an arrangement to continue the membrane over it
and behind the ballast wall.
In choosing any arrangement at the abutment, it should be assumed that there is
a strong probability that failure of the waterproofing across the movement joint
will occur in service; leakage through any failure should be able to drain away
from structural components, particularly any parts that are difficult to inspect
and maintain. Bearings should be on plinths; drainage paths should be clear and
not easily blocked by debris,
Whatever configuration is adopted, the details should be agreed at an early stage
so that the steelwork can be suitably detailed and the erection scheme can be
developed.

Waterproofing membrane
Figure 10.24 Waterproofing at end of steel deck

Water management, including waterproofing and drainage arrangements, is a
key consideration at the ends of the deck. Account needs to be taken of likely
flow rates and water disposal arrangements.
10.7 Footways
Where a cess walkway cannot be accommodated between the main girders of a
half through bridge, it is likely that a footway will have to be provided on
cantilever brackets outside the main girder. In many cases, services will also be
required to be supported by the cantilever.
A typical arrangement is shown in Figure 10.25. Good line and level of the
parapet rails is needed and this is best achieved by alignment during trial
erection. To replicate the alignment during final erection, the detail shows a
locating plate that is positioned in contact with the web stiffener after the
handrails have been aligned and then welded to the face of the bracket.
Restoring this contact during final erection ensures good alignment.
HSFG bolts should be used for the brackets and for fastening the handrails, to
ensure that slip does not take place in service; any slip would probably result in
out-of-line handrails.
118
The top surface of the main girders may be used as part of the walkway and,
for safety, a non-slip surface should be provided. Details should be compatible
with the protective treatment to the girders.
More substantial details are needed for supporting London Underground
services.
Non-slip surface

Services
HSFG bolts
Locating plate
welded during
trial erection
Figure 10.25 Footway detail
119
11
REFERENCES

British Standards
1.
BS 5400 Steel concrete and composite bridges

Part 1:1988 General statement
Part 2:1978 Specification for loads
Part 3:2000 Code of practice for design of steel bridges
Part 4:1990 Code of practice for design of concrete bridges
Part 5:1979 Code of practice for design of composite bridges
Part 6:1999 Specification for materials and workmanship, steel
Part 7:1978 Specification for materials and workmanship, concrete,
Part 8:1978 Recommendations for materials and workmanship,
concrete, reinforcement and prestressing tendons
Part 9:1983 Bridge bearings
Part 10:1980 Code of practice for fatigue
2.
BS EN 10025:1993 Hot rolled products of non-alloy structural steels.

Technical delivery conditions
BSI, 1993
3.
BS EN 10133-2 Hot rolled products in weldable fine grain structural steels.

Part 2: Delivery conditions for normalized/normalized rolled steels
BSI, 1993
4.
BS EN 1991-2:2003 Eurocode 1. Actions on structures. Traffic loads on

bridges
BSI, 2003
Railway Safety and Standards Board documents

Railway Group Standards and Railway Approved Codes of Practice were
formerly issued by Railway Safety.
5.
GC/RC5510, Recommendations for the design of bridges

Railway Safety, Aug 2000.
6.
GC/RT5110, Design requirements for structures

Railway Safety, Aug 2000.
7.
GC/RT5112, Loading requirements for the design of bridges

Railway Safety, May 1997.
8.
GC/RT5203 Infrastructure requirements for personal safety in respect of

clearances and access
Railway Safety, February 2001
9.
GC/RT5212 Requirements for defining and maintaining clearances

10. GM/RT2149 Requirements for defining and maintaining the size of railway
vehicles
120
Network Rail documents

11. RT/CE/C/002 Application and reapplication of protective treatments for
Railtrack Infrastructure
Railtrack, February 2002
12. RT/CE/C/008, Model clauses for specifying civil engineering work
(Various Sections, on different topics)
Railtrack, (Various dates)
13. RT/CE/C/025 Structural assessment of underbridges,
Network Rail, February 2004
14. RT/CE/S/003, Technical approval of design, construction and maintenance
of civil engineering infrastructure
Railtrack, October 2001 (due to be revised, mid 2004)
15. RT/CE/S/039 Specification RT98 Protective treatments for Railtrack
infrastructure, Railtrack, February 2002
16. RT/CE/S/049 Track design handbook,
17. RT/CE/S/101, Track design requirements
Railtrack, October 1997

18. RT/CE/S/102, Track construction standards

London Underground standards
19. CED-ST-3108 Bridges and covered ways
Published by Chief Engineers Directorate, 1995
20. E1008 New or altered assets approvals prior to bringing into use
21. E1045 Fire safety classifications of stations
22. E3304 Structural loading
23. E3314 Assessment of railway underbridges,
24. E3504 Structural steelwork,
25. E8001 Track component configuration plain line,
26. E8013 Clearance requirements
27. E8052 Track asset walkways,
28. P2006, The Management of concessions,
29. TE MTS 0314
121
CEN Standards
30. prEN 1994-2 Design of composite steel and concrete structures. Rules for
bridges.
CEN TC/250/SC4 committee document, due to be published as EN 1994-2
in 2005
31. EN prAnnex A2 Basis of structural design Application for bridges
CEN TC/250 committee document, due to be published in 2004
Regulations
32. The Railways and Other Transport Systems (Approval of Works, Plant and
Equipment) Regulations 1994 (SI 1994 No. 157)
HMSO, 1994
33. The Fire Precautions (Sub-surface Railway Stations) Regulations 1989
(SI 1989 No. 1401)
34. HMSO, 1989The Railways (Interoperability) (High-Speed) Regulations
2002 (SI 2002 No. 1166
The Stationery Office, 2002

Other documents
35. Deformations of Bridges, UIC Leaflet 776-3R
International Union of Railways, 1989 (amended 1990)
36. Loads to be considered in the design of railway bridges, UIC Code 776-1R,
International Union of Railways, 1977 (amended 1987)
37. Track-bridge interaction. Recommendations for calculations (2nd edition),
UIC Leaflet 774-3R
International Union of Railways, October 2001
38. The use of weathering steel in bridges, Publication No. 81
ECCS, 2001
39. Weathering steel bridges (brochure)
Corus, 2002
40. BD 37/01 Loads for highway bridges,
Design Manual for Roads and Bridges, The Stationery Office, 2001.
41. BD 65/97 Design criteria for collision protection beams,
Design Manual for Roads and Bridges, The Stationery Office, 1997
42. BD 60/94 Design of highway bridges for vehicle collision loads
Design Manual for Roads and Bridges, The Stationery Office, 1994
43. BIDDLE, A.R., ILES, D.C. and YANDZIO, E.
Integral Steel Bridges Design Guidance
The Steel Construction Institute, 1997
44. BROWN, C.W.
Reducing Noise emission from steel railway bridges (P173)
45. BROWN, C.W. and ILES, D.C.
Design of steel bridges for durability (P241),
122
46. BROWN, C.W. and ILES, D.C

Commentary on BS 5400-3:2000 Code of practice for the design of steel
bridges
47. EVANS, J.E. and ILES, D.C.
Steel Bridge Group: Guidance notes on best practice in steel bridge
construction (P185)
(See Appendix A for a full list of the individual Notes)
48. HAMBLY, E.C.
Bridge deck behaviour
E & F N Spon, 1991
49. HM RAILWAY INSPECTORATE
Railway safety principles and guidance
Part 1
Part 2 Section A Guidance on the infrastructure
HSE Books, 1996

50. ILES, D.C.

Design guide for composite highway bridges (P289)
51. JOHNSON, R.P.
Analyses of a composite bowstring truss with tension stiffening
Bridge Engineering (Volume 156 of the Proceedings of the Institution of
Civil Engineers), June 2003
52. PUCHER, A.
Influence surfaces of elastic plates
Springer-Verlag, Vienna, 1964
53. RAY, S.S., BARR, J. and CLARK, L.
Bridges design for improved buildability, CIRIA Report 155, 1996
54. WEST, R.
Recommendations on the use of grillage analysis for slab and pseudo-slab
bridge decks
Cement and Concrete Association and Construction Industry Research and
Information Association, 1973
55. ProfilARBED (Editors)
Design tables for filler beam railway bridges
UIC, International Union of Railways, 1999
56. Standard Z Type Superstructure Design manual and drawings
Railtrack, 1996
57. Standard box girder bridge design Users manual and drawings
Railtrack, 1996
123
APPENDIX A
STEEL BRIDGE GROUP

GUIDANCE NOTES
The following is a list of all the Guidance Notes in the SCI publication Steel
Bridge Group: Guidance notes on best practice in steel bridge construction.

GN
Title
Section 1 Design general
1.01 Glossary
1.02 Skew bridges
1.03 Bracing systems
1.04 Bridge articulation
1.06 Permanent formwork
1.07 Use of weather resistant steel
1.08 Box girder bridges
1.09 Comparison of bolted and welded splices
1.10 Half through bridges
Section 2 Design detailing
2.01 Main girder make-up
2.02 Main girder connections
2.03 Bracing and cross-beam connections
2.04 Bearing stiffeners
2.05 Intermediate transverse web stiffeners
2.06 Connections made with HSFG bolts
2.07 Welds how to specify
2.08 Attachment of bearings
Section 3 Materials and products
3.01 Structural steels
3.02 Through thickness properties
3.03 Bridge bearings
3.04 Welding processes and consumables
3.05 Surface defects on steel materials
3.06 Internal defects in steel materials
3.07 Specifying steel material
Section 4 Contract documentation
4.01 Drawings
4.02 Weld procedure trials
4.03 Allowing for permanent deformations
4.04 Alternative construction sequences
Section 5 Fabrication
5.01 Weld preparation
5.02 Post-weld dressing
5.03 Fabrication tolerances
5.04 Plate bending
5.05 Marking of steelwork
5.06 Flame cutting of structural steel
5.07 Straightening and flattening
5.08 Hole sizes and positions for HSFG bolts
5.09 The prefabrication meeting
125
GN
Title
Section 6 Inspection and testing
6.01 Weld quality and inspection
6.02 Surface inspection of welds
6.03 Sub-surface inspection of welds
6.04 Hydrogen/HAZ cracking and segregation cracking in welds
6.05 Weld defect acceptance levels
Section 7 Erection and in-situ construction work
7.01 Setting bearings
7.02 Temperature effects during construction
7.03 Verticality of webs at supports
7.04 Trial erection and temporary erection
7.05 Installation of HSFG bolts
7.06 Transport of steelwork by road
7.07 Site welding
7.08 Method statements
Section 8 Protective treatment
8.01 Preparing for effective corrosion protection
8.02 Protective treatment of bolts

Section 9 Other topics

9.01 Construction (Design and Management) Regulations
126
APPENDIX B
WORKED EXAMPLE
The worked example is of a twin-track bridge spanning 36 m. The bridge is

square at its ends and the chosen form of construction is similar to an E type,
except that the slab is wholly on top of the cross girders.
References in the margins are generally to clauses in BS 5400-3 although some
references to Part 10 clauses are given in the format 10/9.1.2, as used in the
main text of this publication.
Index to worked example

Page
General arrangement
129
Design standards
130
Design parameters
130
Loading
131
Main girder - section properties
132
Intermediate U-frames properties
133
End U-frames properties
135
Slenderness of main girders
136
Moment of resistance of main girder
136
Check on dynamic performance
137
Vertical deflection
137
Curtailment of doubler plate
138
Length of curtailment
138
Check on curtailment of bottom flange
139
Shear resistance
140
Fatigue main girder
142
Welds between web and flange at the support
143
Cross girder - loading
144
Section properties of cross girder
145
Fatigue cross girder
146
Connection of cross girders to main girders
147
Bolted connection
149
Cross girder connection not part of U-frame
152
Lateral bending of main girder top flange
153
Intermediate Stiffeners
154
127
Silwood Park, Ascot, Berks SL5 7QN

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Design guide for steel railway bridges
Subject
Example: Design of girders for a half through plate

girder bridge
Client
CALCULATION SHEET
Sheet
of
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Cross section
1000
3400
1950
1435
1435
1650
200
2800
Rail level
1488
900
50 clear gap
25 thick waterproofing
9640

Longitudinal section on deck

Rail &
368 sleeper
490
Ballast
Ballast
300 under sleeper
1318
25 Water
proofing
250
356 x 368 x 202 UC

Cross girders
1500 centres
375
1500 centres
Adjustment to the average ballast depth to allow for camber and drainage fall on deck
is not shown in this example, for brevity.
Arrangement of doublers and web stiffeners
36 m span
Doubler flange
Top flange
Rail level
D = 2.80 m
Free
Fixed
Elevation
P:\PUB\PUB800\SIGN_OFF\P318\P318Example1D08.doc
129
Rev
Made by
General arrangement
c/c tracks
1965
29

Telephone: (01344) 623345
Fax: (01344) 622944
CALCULATION SHEET
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
Sheet
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Design standards
GC/RT5110
GC/RT5112
GC/RT 5203
GC/RC5510 (Note Appendix B, which modifies Parts 3, 4 and 5 of BS 5400)
BS 5400-3:2000
BS 5400-4:1990
BS 5400-5:1979 (including AMD 3998, May 1982)
BS 5400-10:1980 (including AMD 9352, March 1999) use simplified procedure (see
Clause 9.1.2).

BD 37/01 (Includes modified BS 5400-2)

DTP Yellow document (BS 5400-3 plus BD 16/82)
Design parameters
2 standard gauge tracks on straight alignment, track speed up to 100 mph.
A cess walkway is provided on one side of the track and a Continuous Position of
Safety on the other side. The track is located with approximately 100 mm clearance
between the outer edges of the walkway/CPOS and the inner edges of the top flange,
to allow for possible future realignment of the track.
Heavy traffic, 27 106 tonnes/annum to be considered
Use grade S355 steel and grade C40 reinforced concrete
Combined response of bridge and track to variable loads (rail traffic loads and
temperature) regarding track/bridge interaction does not need a specific check because
rail expansion devices are to be provided at each end of the bridge. Thus longitudinal
displacement of upper surface at end of deck due to temperature effects of the deck
does not need to be checked.
130
Clause
references
are to
BS 5400-3
unless noted
otherwise

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Sheet
Client
CALCULATION SHEET
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Loading
SLS
Nominal
ULS
kN/m Factor
Load Factor
Load
DEAD LOAD
Main girder
Cross girders
Slab
Haunch
Ballast
Track
Waterproofing
Services
0.267 m2 77 kN/m3 plus 10%

2.04 kN/m 4.80 m / 3.0 m crs
4.8 m 0.25 m 25 kN/m3
0.96 m 0.25 m 25 kN/m3
4.525 m 0.49 m 21 kN/m3
2No 2.0 kN/m (5.07/9.64)
(4.8+0.89 m) 0.025 m 24 kN/m3
Say
22.61
6.53
30.00
6.00
46.56
3.89
3.41
0.80
1.10
1.10
1.20
1.20
1.75
1.20
1.75
1.20

Track position, relative to girders
24.87
7.18
36.00
7.20
81.48
2.52
5.97
0.96
166.18
1.00
1.00
1.00
1.00
1.20
1.00
1.00
1.00
22.61
6.53
30.00
6.00
55.87
2.10
3.41
0.80
127.32
factors:
BD 37/01 and
GC/RC/5510
Ballast density
GC/RC/5510
Track weight
taken as extra
over ballast
Allow 2kN/m
for track with
3rd rail
Centre
of tracks
5.07 m
9.64 m
SLS
Nominal
ULS
kN/m Factor
Load Factor
Load
LIVE LOAD
EUDL
2 tracks 4162 kN/36 m (5.07/9.64)
Nosing (EUDL) 100 kN 2/36m 1.338/9.64
Walkway (far) 5 kN/m2 0.7 m 1.05/9.64
121.68
0.77
0.38
1.40
1.40
1.50
170.35
1.08
0.57
172.00
FATIGUE LOAD
1 track
4162 kN (6.77/9.64) / 36 m
2 tracks
2 4162 kN (5.07/9.64) /36 m
Nominal
kN/m
81.23
121.68
ULS
Factor
Load
1.00
81.23
1.00 121.68
Ratio 1 track / 2 tracks
81.23/121.68
0.67
BENDING MOMENTS
ULS Dead load
ULS Live load
168.03 362/8
172.00 362/8
26920 kNm
27860kNm
54770 kNm
SLS Dead load

SLS live load
127.32 362/8
135.08 362/8
20630 kNm
21880 kNm
42500 kNm
Fatigue load
121.68 362/8
19710 kNm
131
1.1
1.1
1.0
133.85
0.85
0.38
135.08
EUDL:
BD37
Table 27

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
CALCULATION SHEET
Sheet
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Main girder - section properties

900 x 60
1000 x 60
120
Compression
bf o = 482
8 FW
t w = 20
989
y c = 1340
Plastic
neutral axis
Elastic
neutral axis
D = 2800
1571
y t = 1460
Tension
120
Plastic
Elastic

Girder make-up
Doubler
Top flange
Web
Bottom flange
Bottom doubler
yield strength
335 N/mm2
335 N/mm2
345 N/mm2
335 N/mm2
335 N/mm2
900 60 mm
1000 60 mm
2560 20 mm
900 60 mm
800 60 mm
Elastic properties for vertical bending

IXX
= 0.4141 m4
A
= 0.3091 m3
ybfl
Ztfd
= 0.3236 m3
ytfl
Ztf
Zbf
= 0.2957 m3
= 0.2836 m3
Zbfd
= 0.2672 m2
= 1.460 m
= 1.340 m
Plastic properties
Mpe
107600 kNm (based on yield strengths appropriate to element thicknesses)
Zpe = Mpe/335 = 0.3212 m3 ybfl 1.691 m
ytfl 1.109 m
Shape limitations
bf0 = 482 mm > 7t fo 355 / y
= 420 1.029 = 432 mm Non-compact
Clear web depth = 2544 mm m = (1109 128)/2544 = 0.386

34t w 355
Limit for compact web:
m
yw
9.3.7.3.1
9.3.7.2
= 680 1.014 / 0.386 = 1787 mm Non-compact

Effective sections
Web:
yc
tw
yw
355
(1460 120 )
20
345
= 60 < 68 Web fully effective
355
132
9.4.2.5.1

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Sheet
Client
CALCULATION SHEET
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Intermediate U-frames properties

Stability of the compression flange is achieved by U-frame action. The configuration
of the intermediate U-frames is:
Centroid
D = 2800
400 x 40
stiffener
I1
= 1943
= 2137
Neutral
axis of
cross girder
I2
B = 9640

Stiffness of crossbeams
Effective width of slab acting with crossbeam on each side is the lesser of:
lR/2 = 1.5/2 = 0.75 m and
B/8 = 9.64/8 = 1.205 m
Effective width = 2 0.75 = 1.5 m
Gross area of concrete within the effective width may be used
9.6.4.1.3
1.5 m effective width
250
Elastic
neutral axis
194
431
375
356 x 368 x 202 UC
Concrete grade is C40, hence short term modulus Ec = 31 kN/mm2

Modular ratio = 205/31 = 6.61
Second moment of area of composite section I2 = 0.003424 m4
ysteel = 431 mm
yslab = 194 mm (NA in slab)
The neutral axis is just in the slab. Gross section properties can be used
The top of the slab is 250 + 375 + 120 + 50 = 795 mm above the main girder
soffit
Take the centroid of the main girder top flange as 62 mm below the top of the doubler
Hence distance of centroid of top flange to top of slab:
d1 = 2800 62 795 = 1943 mm
Distance from centroid of top flange to NA of cross girder and slab
d2 = 2800 62 795 + 194 = 2137 mm
133
4/4.3.2.1

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
CALCULATION SHEET
Sheet
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Stiffness of effective web stiffeners

I1
A
yw
ys
=
=
=
=
0.0005277 m4
0.0288 m2
127 mm
293 mm
32t w = 640
20
400 x 40
stiffener
293
127

Flexibility of cross girder/main girder connection
Consider the connection as the middle of the categories in clause 9.6.4.1.3 (i.e.
comparable to those shown in Figure 42, Type (b).
BS 5400-3
Hence f = 0.2 1010 rad/Nmm = 0.2 104 rad/kNm

Lateral deflection of U-frame under unit loads is given by:
R =
R =
9.6.4.1.3
d13
uBd 22
+
+ fd 22
3 EI1
EI 2
1.9433
3 205 10 6 0.005277
0.5 9.64 2.137 2

205 10 6 0.004102
R = 0.0000226 + 0.0000314 + 0.0000913
+ 0.2 10 4 2.137 2
0.0001453 m/kN
134

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
CALCULATION SHEET
Sheet
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
End U-frames properties

Centroid
400
250
750
d 1 = 1.915 m
2.74 m
400 x 40
700
d 2 = 2.327 m
250
Neutral
axis
700
20
l2
400 x 50
9.64 m
Trimmer
Disc type bearings

(Free rotation
both axes)
Effective width of slab acting with trimmer = 0.75 +0.400 = 1.15 m

Second moment of area of composite section I2 = 0.01022 m4
Height of NA = 538 mm above bottom flange, 412 mm below top of slab

Top of slab is 795 mm above the soffit, as for the intermediate cross girders
Centroid of top flange is 90 mm below line of top of doubler (but no doubler at this
section)
Hence d1 = 2800 90 795 = 1915 mm
Hence d2 = 2800 90 795 + 412 = 2327 mm
Stiffness of effective bearing stiffener
400
16 t w= 320
290 x 30
16 t w= 320
500 x 50
t w = 20
I1 = 0.005833 m4
Consider the connection of the trimmer to the main girder as the stiffest of the
categories in clause 9.6.4.1.3 (i.e. comparable to those shown in Figure 42, Type (c).
Hence f = 0.1 1010 rad/Nmm = 0.1 104 rad/kNm
e,max = R for the end U-frame

R =
d13
uBd 22
+
+ fd 22
3 EI1
EI 2
e,max =
1.9153
0.5 9.64 2.327 2
+ 0.1 10 4 2.327 2
+
6
6
3 205 10 0.005833 205 10 0.01022
= 0.0000020 + 0.0000125 + 0.0000541 = 0.0000686 m/kN

135
BS 5400-3

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
CALCULATION SHEET
Sheet
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Slenderness of main girders

l e = k 2 k3 k5l 1
9.6.4.1.3
9.6.4.1.1.2
l 1 = (EI c l R R )0.25
Ic for a 1000 60 flange and 900 60 doubler = 0.008645 m4
l1 = 205 10 6 0.008645 3.0 0.0001453
0.25
= 5.272 m
l 13
X =
2 EI c e,max
5.272 3
= 0.852
2 205 10 6 0.00865 0.0000686
X =

k5 = 2.22 +
0.69
= 2.730
X + 0 .5
k2 = 1.0 (load on girder at bottom flange level)

k3 = 1.0 (no rotation restraint in plan at supports)
l e = 1.0 1.0 2.730 5.272 = 14.39 m

For slenderness of U-frames, LT is given by:
LT =
9.7.2
le
ry
The value of ryc is that for the top flange plus one third of the depth of the web.
For that section:
Ic = 0.008645 m4
A = 0.131
ryc = 0.257 m
Hence
LT =
14.39
0.257
= 56.0
Moment of resistance of main girder

Consider buckling mode:
L/le = 36/14.39 = 2.50 Hence there will be two half waves and lw = 18.0 m
For input to Figure 11:

yc M ult
M
355
pe
LT
= 56.0
BS 5400-3
335 95010

355 107600
= 52.6
136

Telephone: (01344) 623345
Fax: (01344) 622944
CALCULATION SHEET
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
Sheet
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
le/lw = 14.39/18.00 = 0.80

From Figure 11b) (or Annex G) MR/Mult = 0.698
Hence MR = 0.698 95010 = 66320 kNm
MD = MR/mf3 = 66320/(1.05 1.1) = 57420 kNm

This is slightly more than the applied moment and will probably be satisfactory but the
combined effect with lateral bending still has to be checked. See sheet 25.
Check on dynamic performance

(This check would normally follow immediately after calculation of loading and section
properties, i.e. after Sheet 4 in this example, but is presented here for convenience)

Dead load 127.5 kN/m (nominal DL + 100 mm extra ballast)

Deflection =
5wL4
5 127.5 36 4
=
1000 = 32.8 mm
384 EI 384 205 10 6 0.414
GC/RC5510
Appendix E
This value is below the upper limit for a span of 36 m in Figure 1 of UIC 776-3R
(Corrected formula in GC/RC5510 gives a limit of 39 mm for a 36 m span)
End rotation is given approximately by: 4 32.8 /36000 = 0.0036 rad
This value is below the limit in Table 1 of UIC 776-3R - OK
Live load (Nominal value, 2 tracks) 121.6 kN/m
Deflection =
5wL4
5 121.6 36 4
=
1000 = 31.3 mm
384 EI 384 205 10 6 0.414
Span/deflection = 36000/31.3 = 11503

This ratio is better than the limit of 800 in Table 4 of UIC 776-3R for a single deck
and for Speed Range 2.
Vertical deflection
There is sufficient clearance that the vertical deflections do not infringe clearance
requirements above the road below.
Details of deflection calculations are not presented here, for brevity, but the following
were confirmed:
The track twist with one (250 kN) axle 3 m onto the bridge on one track easily UIC 776-3R
complies with twist limitations
The limitation on uplift at the ends of the deck and rotation about a transverse
axis at the end are also easily satisfied
Lateral deflection is satisfactory by inspection
137

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
CALCULATION SHEET
Sheet
10
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Curtailment of doubler plate

As well as the curtailment of the doubler plate within the span, the bottom flange is
reduced in thickness at the ends (the flange will be fabricated by butt welding several
lengths of plate and advantage can be taken to reduce the bottom flange size near the
ends of the span).
Section properties for reduced girder section
=
=
=
=
=
0.2980 m4
0.1849 m3
0.2789 m3
0.2641 m3
61940 kNm
= 0.2132 m2
= 1128 mm
= 1612 mm
A
ybfl
ytfl
1000 x 60
2.74
Using the slenderness derived for the full section, MD will be

approximately: MD = MR/mf3 = 0.698 61940/(1.05 1.1)
= 37430 kNm
900 x 60
800 x 60
Allowing 5% to cover interaction with lateral bending, the section can be curtailed
when M = 0.95 37430 = 35560 kNm
Bending moment varies approximately as:
2 x 2
M x = M centre 1
L
M
M (1 - (x/18.0))
Where x is the distance from midspan

Hence the distance at which the doubler can be curtailed
35560
x = 18.0 1
= 10.66 m from midspan - Say 7.0 from each end
54770
Length of curtailment
From above, the moment 7 m from the support = 34510 kNm
Average stress in the doubler plate:
= 34320 kNm /0.3164 m3 = 108 N/mm2
Load in doubler = 5832 kN
Assume 8 mm fillet welds along both sides. Around the nose of the
doubler the weld size may be greater, depending on the taper in
thickness of the doubler. That weld must be checked for fatigue (as
Class W) checks not presented in this example.
Design strength of welds =
0.5 y + 455
3 m f3
Capacity of two 8 mm FW = 2
8 173
2
0.5 (335 + 455) )

= 173 N/mm2
1.732 1.05 1.2
= 1957 kN/m
Length required = 5832 / 1957 = 2.98 m Detail as 3.0 m long

Curtailment length

Ixx
Ztf
Zbf
Zbfd
Mult
900 x 60
138
Tapers and
radius as in
Figure 10.5
14.6.3.11

Telephone: (01344) 623345
Fax: (01344) 622944
CALCULATION SHEET
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
Sheet
11
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Check on curtailment of bottom flange

If the bottom flange doubler plate is curtailed within the span, the class G detail at the
end of the doubler plate must be checked for adequacy in fatigue and the toughness of
the bottom flange must also be adequate for the level of tensile stress at that location
For fatigue endurance, a similar location and curtailment detail to that on the top
flange would be needed.
Properties for section without any doublers
Ixx
Ztf
Zbf
Mult
=
=
=
=
0.2232 m4
0.1727 m3
0.1609 m3
53900 kNm
A
ybfl
ytfl
= 0.1652 m2
= 1388 mm
= 1292 mm

Assume minimum effective bridge temperature Ue = 20C

From Table 3c, maximum permitted thickness at 20C is 60 mm for grades S355K2,
S355N, S355M and 86 mm for grades S355NL and S355ML, for k = 1
BS 5400-3
For the Class G detail, kd = 0.5. Assume kg = 1.0 and ks = 1.0

BS 5400-3
The value of k is given by Table 3b:

For 0.25y < max 0.5y k = 1.25
For max 0.25y k = 1.5
(max tensile)
Hence the limiting thicknesses are:
For 0.25y < max 0.5y
0.5 1.25 60 = 37.5 mm for grades K2, N, M
0.5 1.25 86 = 53.8 mm for grades NL, ML
For max 0.25y
0.5 1.5 60 = 45 mm for grades K2, N, M
0.5 1.5 86 = 64.5 mm for grades NL, ML
To ensure that max 0.25y the total ULS moment must be limited to:
335 0.25 0.1609 = 13480 kNm
x = 18.0 1
13480
= 15.7 m from midspan Say 2.0 m from end of girder
54770
Such a curtailment does not offer much saving in material and forces the use of the
higher NL or ML grades. It will be more economic to carry the bottom doublers over
the full span and to use grades K2, N or M.
For a bottom flange with a continuous doubler, the worst likely detail would be
Class F, for which k = 1.0 and at higher tensile stresses k = 1. The limiting
thickness for toughness would then be 60 mm for grades S355K2, S355N, S355M (or
slightly greater than 60 mm, following guidance in GN 3.08).
139
P185

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Sheet
Client
CALCULATION SHEET
12
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Shear resistance
Loading at ULS
Shear (kN)
2991
3084
40
6115
166.18 36/2
2 (2094 1.4) 5.07/9.64
100 1.4 1.338/9.64
Dead load
Live load (RU EUL)
Nosing
RU EUDL
- BD37
3000
Clear Depth of web dwe = 2560 mm

Spacing of stiffeners a = 3000 mm
9.9.2.1

Flanges are parallel and straight, so use clause

9.9.2.2
3000
2560
2560
1.172
d we
yw
tw
355
2560
345
20
355
126
For mfw = 0, Figure 12 gives l/u = 0.609

Effective width of flange for mfw (bottom flange is smaller if no doubler)
b fe = is the lesser of 10 t f
355
= 10 60
yf
355
335
9.9.2.2
= 618 mm
and bf/2 = 450 mm

Hence bfe = 450 mm
mfw =
yf bfe tf2
2
tw
2 yw d we
335 450 60 2
= 0.00600
2 345 2560 2 20
For mfw = 0.005, Figure 13 gives l/u = 0.695 and for mfw = 0.0105, Figure 14
gives l/u = 0.790
So for mfw = 0.00600 l/u = 0.714
= 345 / 3
= 199 N/mm2
l = 0.714 199 = 142 N/mm2

VD
tw (d w hh )
l
m f3
20 (2560 0 )
3
= 6294 kN
1.05 1.1 142 10
140
9.9.2.2

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
CALCULATION SHEET
Sheet
13
29
of
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
VD > 6115 kN so this is just satisfactory. However, the value of shear stress at which
tension field action starts, o is only 78 N/mm2 and the difference between that and the
stress at ULS (120 N/mm2) gives rise to large tension field forces and to large forces
on the intermediate stiffener and on the endpost. To reduce those forces, it is better to
increase the web thickness to 22 mm (additional intermediate stiffeners could be added
in the two panels at each end but this creates extra U-frames and a slightly thicker web
locally is more economic).
For a 22 mm web
d we
tw
yw
355
2560
22
345
355
115

mfw = 0.00546
For mfw = 0, Figure 12 gives l/u = 0.680
For mfw = 0.005, Figure 13 gives l/u = 0.795 and
For mfw = 0.010, Figure 14 gives l/u = 0.831
So for mfw = 0.00546 l/u = 0.798 and l = 0.798 199 = 159 N/mm2
VD
22 (2560 0 )
3
= 7753 kN
1.05 1.1 159 10
At the support,
2560 2 22 2
b 2 t 2
2
o = 3.6 E 1 + w = 3.6 205000 1 +
= 94 N/mm

a b
3000 2650
= 6115/(2560 22) = 109 N/mm2

The end support must be designed to resist an additional moment My given by:
My =
8( o ) t w l 2s
d = arctan(2560/3000) = 40.5 degrees

My =
8 (109 94 ) 22 2560 2
40.5
10 6 = 427 kNm
This moment is easily resisted by using an endplate to the girder, of a similar size to
the bearing stiffeners and at a distance of 400 mm from the bearing stiffener.
141
9.14.3.4

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
CALCULATION SHEET
Sheet
14
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Fatigue main girder

At midspan
The top flange is Class F due to the transverse stiffener to

flange weld on the underside and due to the doubler to
flange fillet weld on the upper side. The bottom flange is
Class D, due to the web/flange weld, but the web is class G
at the patch plates for the cross girders and is slightly
further from the NA than the top flange.
10507
Class F (weld
clear of flange
edge)
Moment due to unfactored live load at mid span

M = 121.7 362/8 = 19720 kNm
Sheet 3
Elastic modulus at top of bottom flange = 4.141 1011 / 1340 = 3.090 108 mm3
Sheet 4

Stress range R
19720 106
3.090 108
= 64 N/mm2 = Rmax
Limiting stress range T = k1 k2 k3 k4 k5 o`
10/9.2.2.1
k1 = 1.0 (design life 120 years)

k2 = 1.0 (single cycle of loading)
k3 = 2.19 (length of influence line > 28 m, heavy traffic)
k4 = 1.0 (27 million tonnes/year)
k5 = 1.27 (Since P1/(P1 +P2) = 6.773/(6.773 + 3.372) = 0.668)
10/Table 6
For Class G, o = 29 N/mm
10/Table 8
T = 1.0 1.0 2.19 1.0 1.27 29 = 81 N/mm2 > 64 N/mm2 OK

At the end of the doubler
The top flange plate is Class G
Doubler
The end of the doubler is 4.0 m from the end of the span,
i.e. 14.0 m from midspan
The moment due to unfactored live load can be taken as:
Moment at midspan [1 (14.0/18.0)2]
= 19720 0.395 = 7791 kNm
Elastic modulus = 1.849 1011 mm3
Stress range R
7791 10 6
1.849 10 8
R = 75
Class G at end
Sheet 10
= 42 N/mm2 = Rmax`
For Class G, o = 29 N/mm2
T = 1.0 1.0 2.19 1.0 1.27 29 = 80 N/mm2 > 42 N/mm2 OK

Note that, for the class G detail in compression, kd = 0.5 and k = 2, which gives a
limiting thickness of 60 mm for grades S355 K2, N and M, which is adequate.
142
6.5.3
6.5.4

Telephone: (01344) 623345
Fax: (01344) 622944
CALCULATION SHEET
Job No.
BCR 876
Job Title
Subject

girder bridge
Sheet
Client
15
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Welds between web and flange at the support

ULS loading
Shear at support = 6115 kN (Sheet 12)

Shear flow in web is given by:
q =
VAy
I
Where A is the area and y is the distance to its centroid, of the part of the section
further from the neutral axis than the position being considered.

Elastic properties of reduced main girder section (with bottom doubler & 20 mm web) Sheet 10
IXX
= 0.2980 m4
A
= 0.1654 m2
= 0.1849 m3
ybfl
= 1.128 m
Ztf
= 0.2789 m3
ytfl
= 1.612 m
Zbf
= 0.2641 m3
Zbfd
The bottom flange is larger than the top flange, so consider the shear flow at the
web/flange junction of the reduced girder section (i.e. without doubler).
Area of bottom flange
Area of bottom flange doubler
Distance to centroid of bottom flange
Distance to centroid of bottom flange doubler
q =
=
=
=
=
0.054
0.048
1.038
1.098
m2
m2
m
m
6115 (0.054 1.038 + 0.048 1.098)

= 2316 kN/m
0.298
Shear capacity of two 8 mm fillet welds

plate on Sheet 10).
= 1957 kN/m (see calculations for doubler
Shear capacity of two 10 mm fillet welds = 2447 kN/m 10 mm welds satisfactory

Note that in earlier calculations, the dimensions of the welds were assumed to be 8 mm
leg length. The larger size has negligible effect on those calculations and in practice,
smaller deeper penetration welds might be used, in which case the 8 mm dimension
would still be appropriate.
Fatigue loading
Live load shear = 2 2094 5.07/9.64
= 2204 kN
Range of shear flow: 2329 (2204/6149)
= 835 kN/m
Stress range = 835 / (2 10 0.7071)
= 59 N/mm2
For Class W, o = 25 N/mm2

k factors are as for other main girder checks (Sheet 14)
T = 1.0 1.0 2.19 1.0 1.27 25 = 70 N/mm2 > 59 N/mm2 OK
143

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Sheet
Client
CALCULATION SHEET
16
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Cross girder - loading

Live loading
Consider the proportion of live loading carried by a single cross girder. Assume worst
effect is with one of the concentrated loads over a cross girder
0.8
1.6 m
250kN
1.6 m
250kN
1.6 m
250kN
0.8
250kN
80kN/m

1.5 m
1.5 m
80kN/m
1.5 m
1.5 m
1.5 m
1.5 m
1.5 m
1.5 m
Loading on the central cross girder may be estimated initially by simple rule of
thumb distribution from consideration of statics reactions but will need to be verified
by suitable analysis. Here, assume that the loading is simply the load due to one of the
concentrated edge loads (i.e. 250 kN/track)
(A grillage model showed this assumption to be a slight overestimate in this case)
For dynamic factor, take length of influence line for deflection as twice the main girder BS 5400-2
spacing
Table 17
Dynamic factor = 0.73 +
2.16
19.28 0.2
1.245
Loading on a single cross girder, from one track: 250 1.245 = 311 kN
For worst case loading on a cross girder, assume
tracks centrally located and assume that the cross girder
is simply supported. Bending moment in cross girder,
due to loads applied along the lines of the rails (1.5 m
spacing of wheel loads, inner rails 1.9 m spacing) is
given by:
2.37
1.5
1.9
1.5
2.37
Bending moment diagram
M = 311 (2.37 + 3.87) fL = 970 1.4 = 1358 kNm

Dead loads
Total loads
Cross girders
Slab
Haunch
Ballast
Track
Waterproofing
9.6 m 2.04 kN/m

1.5 9.6 m 0.25 m 25 kN/m3
2No 1.5 0.96 m 0.25 m 25 kN/m3
1.5 9.05 0.49 m 21 kN/m3
2No 3 m 2.0 kN/m
1.5 (9.6+0.89) 0.025 24 kN/m3
SLS
Nominal
ULS
kN Factor Load (kN) Factor Load (kN)
19.58 1.10
21.50
1.0
19.58
90.00
1.2
108.00
1.0
90.00
18.00 1.20
20.70
1.0
18.00
139.69 1.75
244.50
1.2
167.63
6.00
1.2
7.20
1.0
6.000
9.75 1.75
17.10
1.2
11.70
419.90
312.91
Calculate ULS moment due to this total load as a UDL (conservative)

M = 419.9 9.64 / 8 = 506 kNm
Total ULS moment = 1358 + 506 = 1864 kNm
144
e
3
Job No.
BCR 876
Job Title
Subject

girder bridge
17
of
29
Rev

Telephone: (01344) 623345
Fax: (01344) 622944
CALCULATION SHEET
Client
Beam section is compact, by inspection
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
For concrete strength 40 N/mm2 and steel yield strength 345 N/mm2, the equivalent
steel area of the slab in a plastic section is:
o
e
From a balance of tensile and compressive areas, the plastic neutral axis is 10 mm into
the top flange of the UC. Mpe is then given by:
Mpe = 345 (18260 135+ 25700 177 2 7440 5) 106 = 2407 kNm
MD = 2407/1.1 1.05 = 2084 kNm > 1870 kNm OK
Elastic section properties for SLS and fatigue checks
b/L = 0.75 / 9.64 = 0.077
For fatigue loading, consider the loading under the inner rail of one track (position of
maximum moment from one track). Determine effective breadth for quarter span
regions (slightly conservative).
Effective width of slab = 2 750 0.95 = 1425 mm
Modular ratio = 205/31 = 6.61

Section properties of composite section
I2
= 0.002634 m4
ysteel = 399 mm
yslab = 226 mm (NA just in slab)
Zbf = 6.602 106 mm3
Ztf
= 108.8 106 mm3
Effective breadth 1.425 m

226
NA
399
For SLS checks (required because the section is considered compact at ULS) a slightly
greater effective width can be used for midspan regions, where the applied moment is
greatest, but there is little difference in section properties
Check at SLS
C r e a
T h i s
e
m
d
a
o
e
n
r i
2 7
a l
M a
s
r c h
c o p
2 0 0 6
r i g h t
Concrete grade is C40, hence short term modulus Ec = 31 kN/mm2
i
r
From Table 4, with b/L = 0.077, = 0.95
Effective breadth ratio depends on b/L
Area of UC section = 25700 mm2
1500 250 (0.4 40 m / 345) = 18260 mm2 (m for steel = 1.05)
Section properties of cross girder
Sheet
Average load factor is at least 1.27 times greater at ULS than at SLS, so an upper
bound for SLS design moment would be 1864/(1.27 1.1 1.05) = 1268 kNm
Stress in bottom flange = 1268 / 6.602 = 192 N/mm2 Satisfactory
145
5*/6.2.2

Telephone: (01344) 623345
Fax: (01344) 622944
CALCULATION SHEET
Job No.
BCR 876
Job Title
Subject

girder bridge
Sheet
Client
18
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Fatigue cross girder

Maximum moment due to loading from one track (take track as centrally located)
Loading from one track, including dynamic factor = 311 kN
Reaction at RH main girder, due to load on LH
track = 311 3.12 / 9.64 = 100.7 kN
Sheet 16
P1
P2
Moment at inner rail = 100.7 5.77 = 581 kNm
Sheet 16
155.5
Total unfactored moment due to two tracks

M = 311 3.12 = 970 kNm
Hence P1 / (P1 + P2) = 581 /970 = 0.60
155.5
210.3
100.7
BMD for one track loaded

Stress range at top flange level
= 970 106 / 108.8 106 = 9 N/mm2

Stress range at bottom flange level
= 970 106 / 6.602 106 = 147 N/mm2

T = k1 k2 k3 k4 k5 o`
10/9.2.2.1d
k1 = 1.0 (design life 120 years)

k2 = 1.0 (single cycle of loading)
k3 = 1.92 (length of influence line = 19.28 m, heavy traffic)
10/Table 4
k4 = 1.0 (27 million tonnes/year)
10/Table 5
k5 = 1.27 (Since P1/(P1 +P2) = 0.60)
10/Table 6
Top flange is Class F (shear studs are welded to flange), o = 40 N/mm2
10/Table 8
T = 1.0 1.0 1.92 1.0 1.27 40 = 98 N/mm2 > 9 N/mm2 OK

Bottom flange is Class C (rolled section, as rolled condition, no welds)
o = 40 N/mm2
T = 1.0 1.0 1.92 1.0 1.27 78 = 190 N/mm2 > 147 N/mm2 OK
146

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
CALCULATION SHEET
Sheet
19
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Connection of cross girders to main girders

The magnitude of the design moment at the connection depends on the values of the
restraint forces FR and Fc
The force FR is given by:
fc l w
fc EI c
but not greater than

FR =
2
ci fc 667 R
ci fc 16.7l R
Consider the stress in the flange as that due to ULS loading (Sheet 3)
fc = 54770 106 / 3.091 106 =177 N/mm2
ci

2 ES
2LT
= 0.3212/0.3091 = 1.039
LT = 56 (Sheet 8)
ci
2 205000 1.039
56 2
670 N/mm2
R = 0.0001453 mm/N (Sheet 6)

lw
= 18.0 m (Sheet 8)
Ic
= 0.008645 m4 (Sheet 8)
18.0
177
FR =
670
177
667
0
.0001453
= 66.7 kN
But not more than

9
177 205000 8.645 10
FR =
16.7 3000 2
670 177
= 4233 kN
Hence FR = 66.7 kN
147
9.12.2
9.12.3.3

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Sheet
Client
CALCULATION SHEET
20
of
29
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
The force Fc depends on the rotation of the end of the cross girder, relative to those
either side of it, when the cross girders are all simply supported at their ends.
Fc
d2
1.5 R +
l 2R
12 EI c
Consider the cross girders at midspan, under a loading arrangement as below:

250kN
250kN
250kN
250kN
80kN/m

U-1
U0
U+1
U+2
The UDL is omitted on one side as it gives a worse differential loading between
adjacent U-frames. Simple static distribution with 30% redistribution to adjacent
cross girders gives loading at U-1, U0, U+1 and U+2 to be 0, 209 kN, 234 kN and
150 kN. (A grillage analysis would given similar values but with a little more
distribution to frame U-1).
The average of the values at U-1 and U+1 is 117 kN, so the differential loading on
U0 is 209- 117 = 92 kN.
Using the above estimate of loading that causes relative displacements, the end
rotations may be calculated simply by equating them to the area of the M/EI diagram
between the middle and the end of the cross girder. Use the properties of the effective
section for U-frame action (Sheet 5).
In this case, the diagram is polygonally shaped and the area can be taken approximately .
as 2/3 height times length.
The live load midspan bending moment on the most heavily loaded cross girder is
derived on Sheet 16 as 1358 kNm so the applicable moment for relative loading is:
92/250 1358 = 500 kNm
= 2/3 500 4.82 / 205 106 0.002634 = 0.00297 radians
d2
= 2137 mm (Sheet 5) and depth to mid-depth of UC = 2381 mm
Fc =
0.00297 2137
1.5 0.0001453 +
3000 3
10 3
29.0 kN (ULS value)
12 205000 8.645 10 9
The moment at the endplate connection is the product of the force (FR + Fc) and the
lever arm from the centroid of the top flange to the mid-depth of the UC.
Hence moment on connection:
M = (66.7 + 29.0) 2.381 = 228 kNm at ULS
And M = (66.7 42000/55080 + 29.0 1.1/1.4) 2.381 = 178 kNm at SLS
148
Rev

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Sheet
Client
CALCULATION SHEET
21
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Bolted connection
280
500
40
90
300
90
180
375
180

15
200
440
Consider the hogging moment at the connection of frame U0

Check tension in top row of bolts (SLS)
Bolt tensions
Moment
from
Sheet 20
Inertia of bolts about centre of bottom flange (for hogging bending)

Ibolts
= 2 (0.0812 + 0.2612 + 0.4412 + 0.5312 +0.6212)

= 1.874 m2
SLS load at top row of bolts = 178 0.621/1.874 = 59 kN

Add 10% for prying Load = 65 kN
Preload 207 kN (M24 general grade) OK
14.3.6
Bottom
flange
Check shear at interface (SLS)
Dead load shear at end of cross girder = 156 kN (Sheet 16)

Dynamic factor for shear = 0.82 +
1.44
19.28 0.2
1.164
Live load shear 2 209 1.164 5.07 / 9.64 1.1 = 282 kN

Total shear 156 + 282 = 438 kN
N = 10 bolts Shear / bolt = 43.8 kN
PD
k h Fv N
14.5.4.2
m f3
kh = 1.0 and = 0.5 For resistance of a single bolt, N = 1
Prestress load Fv = 207 65 = 142 kN
149

Telephone: (01344) 623345
Fax: (01344) 622944
CALCULATION SHEET
PD
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
Sheet
22
of
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
1.0 142 0.5 1

= 59.2 kN > 43.8 kN OK
1 .2 1 .0
Consider the sagging moment at the connection of frame U-1

Check tension in bottom row of bolts (SLS)
40
Inertia of bolts about top row of bolts (assuming that this is the
effective line of action of the compressive component of the
moment)
Ibolts = 2 (0.0902 + 0.1802 + 0.3602 + 0.5402) = 0.924 m2
Fc and FR forces are as for frame U0 but in outward direction
SLS load at bottom row of bolts = 178 0.540/0.924 = 104 kN

Add 10% for prying Load = 114 kN
Preload 207 kN (M24 general grade HSFG) OK
Bolt tensions
Check shear at interface (SLS)
Dead load shear at end of cross girder = 156 kN (Sheet 16)

Live load shear = 0 (no load on this crossframe in the sagging moment condition)
Shear / bolt = 156/10 = 15.6 kN
Prestress load Fv = 207 114 = 93 kN
=
1.0 93 0.5 10
= 38.8 kN > 15.6 kN OK
1.2 1.0
Check bending of endplate at ULS
Moment on endplate M = 228 kNm (Sheet 20)
200
ULS load at top row of bolts = 228 0.621/1.874 = 76 kN

Moment at face of gusset (assuming cantilever bending of rigid
endplate and allow 10% for prying action)
M = 76 0.085 1.10 = 7.11 kNm
Effective width of plate = 40 + 90/2 = 85 mm
(top of endplate to midway between bolts)
30 gusset
Assuming a 35 mm thick endplate

Plastic moment capacity of 85 mm width Zpe = 352 85/4 = 26000 mm3
Bending resistance MD = 26000 345 / 1.05 1.1 = 7.78 kNm > 7.11 kNm OK
150
Rev
Made by
Shear at ULS is OK by inspection, since the prestress load is greater and the applied
shear is less than (1.4/1.1) times the SLS shear but combined vertical and longitudinal
shear needs to be checked.
PD
29

Telephone: (01344) 623345
Fax: (01344) 622944
CALCULATION SHEET
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
Sheet
23
of
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
There will be a similar bending in the main girder web plate, so a similar total
thickness will be required. Web thickness = 20 mm, so weld on a 20 mm patch plate
Check welds to gusset plate, under tensile forces (ULS)
Bolt forces = 76 kN, 65 kN and 54 kN in the top

three rows (no extra 10%)
Total force = 195 2 = 390 kN
Load on weld to end plate = 76 kN / 85 mm
= 894 kN/m per side
Capacity of 8 mm FW = 979 kN/m (Sheet 10) OK

Shear on attachment to UC = 390 kN / (2 sides 500 mm) = 390 kN/m

Bending on horizontal interface
M = 2 (76 0.260 + 65 0.170 +54 0.080) = 70.4 kNm
Z of 500 mm long welds = 2 0.502 / 6 = 0.0833 m2
Bending load on weld = 70.4 / 0.0833 = 845 kN/m

390 2 + 845 2
= 931 kN/m
< Capacity of 8 mm FW (979 kN/m, Sheet 10) OK

By inspection, a 8 mm FW between the web stiffener and the web is likely to be
adequate to transmit the loading from the bolts, via the web, into the stiffener.
Detailed checks should be made on that connection (calculations not included here).
Fatigue loading on welds
The fatigue loading on the welds can be taken very conservatively as 1/1.4 times the
above ULS loading (the loads that give rise the calculated forces in the welds include a
significant contribution from the dead loading on the main girder), that is 931/1.4 =
665 kN/mm
Stress in two 8 mm fillet welds = 665/(2 5.66) = 59 N/mm2
The limiting stress range is derived as on Sheet 18, using the same k factors and a
value of 0 = 25 N/mm2 (for a class W detail)
T = 1.0 1.0 1.92 1.0 1.27 25 = 61 N/mm2 > 59 N/mm2 OK
151
Rev
Made by
Requirement for patch plate on main girder
Resultant load =
29

Telephone: (01344) 623345
Fax: (01344) 622944
CALCULATION SHEET
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
Sheet
24
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Cross girder connection not part of U-frame

Vertical shear
Dead load shear (SLS) = 156 kN (as for U-frame cross girders)
Live load shear (for worst loading) = 311 1.1 = 342 kN (Sheet 16)
Total load = 156 + 342 = 498 kN
Slip resistance per bolt (M24 HSFG)
PD
1.0 207 0.5 10

= 86.3 kN
1.2 1.0
Capacity of 6 bolts = 518 kN OK

A 25 mm (minimum) endplate will be used

The weld between UC web and endplate will need to be checked for adequacy in
fatigue. A 10 mm FW will probably be OK (calculations not presented here).
Longitudinal shear
Shear lag factor at SLS is given by b/L ratio
b = 4.8 m hence b/L = 4.82/36.0 = 0.15
= 0.46 (Table 4)
Table 4 of
Part 3
For calculation of shear flow at support, assume slab is uncracked at the ends
Section properties for long-term and short-term composite sections give the following:
Ay
= 5.23 10 5 mm 1 (long term) and
I
Ay
= 8.63 10 5 mm 1 (short term)
I
(Section properties not presented here)

Dead load shear = 127.32 36/2 = 2292 kN
Shear flow = 2292 5.23 105 = 120 kN/m
Live load shear = 2094 2 5.07/9.64 = 1101 kN
Shear flow = 1101 8.63 105 = 95 kN/m
Total shear flow = 120 + 95 = 215 kN/m
Shear on first non U-frame cross girder will be the most onerous
Longitudinal shear = 215 1.5 = 323 kN
This load is too great to be carried in friction at the same time as the vertical shear.
A shear plate detail will be required (see Figure 10.12). The shear plate will transfer
vertical shear in bearing and longitudinal shear in friction 6 bolts should be provided.
152
RU EUDL
BD37

Telephone: (01344) 623345
Fax: (01344) 622944
CALCULATION SHEET
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
Sheet
25
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Lateral bending of main girder top flange

Now that the relative rotation of the most heavily loaded cross girder has been obtained 9.12.3.3(b)
( = 0.00297 radians), the lateral bending of the top flange can be evaluated, using
Appendix E of Part 3.
My
(L l e ) 1.25
5EI cd 2
1 +
) 2.8 + 3.5( fc ci
Ll e (1 fc ci
)2
From Sheet 8:
le = 14.39 m
Ic = 0.008645 m4

From Sheet 19:

fc = 177 N/mm2
ci = 670 N/mm2
From Sheet 8, lw = 18.0 m, which is 6 times lR and lR has been calculated in
accordance with 9.6.4.1.1.2, so:
Take ci = 1.25ci = 838 N/mm2
From Sheet 5:
d2 = 2137 mm
My
(36 14.39) 1.25

5 205 106 0.008645 0.00297 2.137
1 +
36 14.39(1 177 838)

2.8 + 3.5(177 838)2
My = 196.0 kN
No directly applied lateral load, so My = My

MDy = (335 8.645 109 / 500) / (1.05 1.1) = 6272 kNm
M y
M
+
M D M Dy
54770 196.0
+
57420 6272
= 0.954 + 0.031 = 0.985 <1 OK
153

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Sheet
Client
CALCULATION SHEET
26
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Intermediate Stiffeners
Loading
9.13.3
Axial loading F = Ftw + Fwi + the load transferred from interaction with the
crossbeam)
Bending M = (FR + Fc)d + bending due to eccentricity of F
3.00
Consider the stiffener adjacent to the end

frame, where Ftw will be greatest, and assume
that the FR and Fc forces are those determined
as the greatest effects in a U-frame
88 N/m
Elastic
neutral axis
Stiffener
57 N/m

6149 kN
M = 16920 kNm
To determine Ftw first evaluate o
b 2 t 2
b
= 3.6 E 1 + w 1 1
2.9 E t w
a b
9.13.3.2
To evaluate 1 calculate the stresses in the main girder at the position of the first Uframe (assume bottom doubler extends over full span, use moduli for 20 mm web).
End reaction in main girder = 6115 kN (Sheet 12)
Moment = 6115 3.00 (166.18 + 171.90) 3.002 /2 = 16820 kNm
Stress at top of web = 16820 106 / 1.921 108 = 88 N/mm2
Stress at bottom of web = 16820 106 / 2.955 108 = 57 N/mm2 (tension)
Hence b = 73 N/mm2 and 1 = 16 N/mm2
Take = 6115 / (2560 22) = 109 N/mm2 (using a 22 mm web)
2
2
2
16
2560 22
2560
2

1
= 75 N/mm
2.9 205000 22
3000 2560
o = 3.6 205000 1 +
Panel length a is greater than length of stiffener ls therefore:
Ftw = ( o) twa = (109 75) 22 3000 = 2244 kN

To determine Fwi (for buckling check) consider the slenderness of the effective stiffener 9.13.3.3
section (see Sheet 6 for details of effective section).
ls
rse
ys
355
(Stiffener section properties with 22 mm web calculated

as for section with 20 mm web, see Sheet 6)
ys = 345 N/mm2 and rse =
5.639 10 8 31480
154
134 mm

Telephone: (01344) 623345
Fax: (01344) 622944
CALCULATION SHEET
2560
134
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
Sheet
27
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
345
= 18.8
355
From Figure 24, with =18.8, ks = 0.036

Fwi =
l 2s
tw ks R
a
Since there are no longitudinal stiffeners
R = R + 1 +
where R is the lesser of and o

6
R = 75 + 16 + 73 / 6 = 103 N/mm2

Fwi =
2560 2
22 0.036 103 10 3
3000
178 kN
Force transferred from the cross girder is the dead load of 1.5 m length of deck and
the live load (taken to be the maximum due to RU load on most heavily loaded cross
girder)
Dead load component = 140 1.5 = 210 kN
Live load component
= 2 250 DF 5.07 /9.64
Where DF is the dynamic factor for shear, given by:

1.44
0.82 +
= 1.164
19.28 0.2
Live load component
= 500 1.164 0.526 = 306 kN
Total load from cross girder = 210 + 306 = 516 kN

Horizontal forces causing bending of the effective stiffener.
FR = 66.7 kN (acting in either direction)
Fc = 29.0 kN (acting outward on the most heavily loaded U-frame)
Strength of effective stiffener section
Stresses at the top of the web

Axial stress due to Ftw = 2244 103 / 31480 = 71 N/mm2
Bending due to eccentricity of Ftw from centroid of stiffener
M = 2244 118 103 = 265 kNm
Stress in web = M / Z = 265 106 / 4.78 106 = 55 N/mm2

Stress in stiffener outstand = M/Z = 265 106 / 1.855 106 = 143 N/mm2 (tension)
Total vertical stress in web = 71 + 55 = 126 N/mm2
155
(Sheet 19)
(Sheet 20)

Telephone: (01344) 623345
Fax: (01344) 622944
Job No.
BCR 876
Job Title
Subject

girder bridge
Sheet
Client
CALCULATION SHEET
28
of
29
Rev
Made by
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Equivalent stress in web is given by:
e =
9.13.5.1
( 1 + k b )2 + es2 2 es2 ( 1 + k b ) + 3 R2
k is the lesser of 0.77 and 2y/b. At the edge of the panel 2y/b = 1 so k = 0.77
e =
(16 + 0.77 73)2 + 1262 126(16 + 0.77 73) + 3 1092
Strength =
y
m f3
345
= 299 N/mm2
1.05 1.1
= 218 N/mm2
OK
Total stress in stiffener = 71 143 = 72 N/mm2 (i.e. tension) OK

Stresses at level of the NA of the cross girder
Ignore the tensile load due to the shear transmitted from the cross girder

Vertical stresses in web and stiffener due to Ftw are unchanged

Bending due to horizontal restraint forces at top flange level
Ignore Fc, because it acts outwards and reduces the compressive stress in the web.
Take FR as acting inward. Hence M = 66.7 2137 /1000 = 143 kNm
Stress in web = 143 106 / 4.78 106 = 30 N/mm2
Stress in stiffener = 143 106 / 1.855 106 = 77 N/mm2 (tension)
Total stress in web = 71 +55 +30 = 156 N/mm2
k is the lesser of 0.77 and 2 799/2560 = 0.62 (NB b is now tensile)
e =
(16 + 0.62 73)2 + 156 2 156(16 + 0.62 73) + 3 1092
= 254 N/mm2 OK
Total stress in stiffener = 71 143 77 = 149 N/mm2 (i.e. tension)
OK
Check buckling of stiffener within the middle third

Assume 2/3 of shear from cross girder is in the effective stiffener, as tensile load
Net axial load = 2244 + 178 2/3 516 = 2078 kN
Bending due to axial components = (2244 2/3 516) 118 103 = 224 kNm
Bending due to FR (taken as acting inward): M = 66.7 1737 /1000 = 116 kNm
Verify that:
P
Ase ls
M xs
Z x y
9.13.5.3
m f3
156

Telephone: (01344) 623345
Fax: (01344) 622944
CALCULATION SHEET
Job No.
BCR 876
Job Title
Subject

girder bridge
Client
Sheet
29
of
DCI
Date
Dec 2003
Checked by
CHP
Date
Feb 2004
Since ys and y are equal, the value of is the same as on Sheet 27, i.e. = 18.8 and
thus ls = c = 0.965 ys = 333 N/mm2
2244 103
340 106
1
+
6
31480 333 4.78 10 345 1.2 1.1

0.198 + 0.206 = 0.404 0.758 OK

Other important items not covered, for reasons of brevity, include:
Design of trimmer girders (Note that the dynamic factor on trimmers is greater
than for the ordinary cross girders)
Design of shear connectors on cross girders
Adequacy in carrying derailed trains
Water management
Selection of bearings
157
Rev
Made by
For single-sided stiffener, ls is given by Figure 37, Curve D, according to the value
of .
In practice, all these items would be checked or designed.
29

Railway Bridges 2

Uploaded by

Copyright:

Available Formats

Railway Bridges 2

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Railway Bridges 2

Uploaded by

Copyright:

Available Formats

What is the purpose and scope of the document?

What is the purpose and scope of the document?

What design considerations are discussed in the document?

What design considerations are discussed in the document?

P318: Design Guide for Steel Railway Bridges

Created on 27 March 2006

SCI PUBLICATION P318

Design Guide for Steel

D C ILES MSc ACGI DIC CEng MICE

Created on 27 March 2006

P318: Design Guide for Steel Railway Bridges

2004 The Steel Construction Institute

P318: Design Guide for Steel Railway Bridges

Created on 27 March 2006

Metronet Rail SSL Ltd

P318: Design Guide for Steel Railway Bridges

Created on 27 March 2006

P318: Design Guide for Steel Railway Bridges

Created on 27 March 2006

DESIGN FOR STRENGTH

DESIGN FOR FATIGUE ENDURANCE

DESIGN FOR DEFORMATION PERFORMANCE

STEEL BRIDGE GROUP GUIDANCE NOTES

P318: Design Guide for Steel Railway Bridges

Created on 27 March 2006

P318: Design Guide for Steel Railway Bridges

Created on 27 March 2006

P318: Design Guide for Steel Railway Bridges

Created on 27 March 2006

P318: Design Guide for Steel Railway Bridges

Key functional requirements

Provision of satisfactory support to the railway traffic and infrastructure

Provision of adequate clearances between the structure and the traffic on

The first requirement can be expressed in terms of requirements for:

Strength and fatigue endurance

Limiting the bridge deformation

Created on 27 March 2006

The second requirement is expressed in terms of various clearance gauges that

The railway infrastructure

2.2.1 Permanent way (track)

P318: Design Guide for Steel Railway Bridges

Created on 27 March 2006

Continuous welded rail

P318: Design Guide for Steel Railway Bridges

Created on 27 March 2006

P318: Design Guide for Steel Railway Bridges

Types of track commonly found on the UK rail network

Track for Network Rail

CEN 60 FB rail on type 600 steel sleeper

BS113A FB rail on F27 or F40 concrete sleeper

BS113A on EF29 shallow concrete sleeper

BS113A on timber sleeper with baseplate

BS113A on type 500 or 560D steel sleeper

Track for London Underground

Bull Head rail on timber sleeper

* Track depth is measured from top of rail to underside of sleeper

Created on 27 March 2006

It is difficult in practice to arrange and maintain a satisfactory transition

P318: Design Guide for Steel Railway Bridges

Because track maintenance is highly mechanised, the introduction of short