Grillage Analogy C.S.surana R.agrawal - For Word

Grillage Analogy in
Bridge Deck Analysis
C.S. Surana
IL Agrawal
-
Narosa
Z-
1-1-Toic
-L)%%

,
New Delhi Madras Bombay Calcutta London

1
C.S. Surana
Department of Civil Engineering
Indian Institute of Technology,
Delhi Nevi Delhi-110 016, India
R. Agrawal
Department of Civil Engineering
Institute of Technology, Banaras Hindu University
Varanasi, India
.r
Copyright C 1998 Narosa Publishing House
NAROSA PUBLISHING HOUSE
6 Community Centre, Panchsheel Park, New Delhi 110 017
35-36 Greams Road, Thousand Lights, Madras 600 006 PIA
306 Shiv Centre, D.B.C. Sector 17, K.U. Bazar P.O., New Bombay 400 705 252 2
2F-2G Shivam Chambers, 53 Syed Amir Ali Avenue, Calcutta 700 019
3 Henrietta Street, Covent Garden, London WC2E 8LU, UK
All rights reserved. No part of this publication may be

reproduced, stored in a retrieval system, or transmitted
in any form or by any means, electronic, mechanical,
photocopying, recording or otherwise, without the prior
permission of the publisher.
All export rights for this book vest exclusively with

the publisher. Unauthorised export is a violation of
Copyright Law and is subject to legal action.
ISBN 81-7319-153-0
Published by N.K. Mehra for Narosa Publishing House, 6

Community Centre, Panchsheel Park, New Delhi 110 017,
typeset at Innovative Processors, New Delhi-110 002 and
printed at Rajkamal Electric Press, Delhi 110 033 (India)

_ To
Our Parents
for having shown us the way
Our Families
for their constant support
A
4
V
Preface
Bridge design and construction all over the 'world has undergone remarkable
changes in the past two decades. The increased demand for complex
roadway alignments, advances in construction technology and availability of
computing power for bridge design, are some of the factors for these
developments.
Over the years, a number of methods of analysis of highway bridge
decks have been- evolved. The methods range from the simplified hand
method like Courbon's or graphical methods like Hendry-Jaeger and
Morice & Little etc. to highly sophisticated methods like finite element,
finite strip etc. The former are conservative and the latter which require
fairly complex computer programs and larger computational facilities, are
prone to errors of idealization and interpretation of results.
Grillage analogy method, which is well-established and computer-
oriented, bridges the gap between the two. The present book describes
bridge deck analysis by grillage analogy. The method is versatile in nature
and can be applied to a variety of bridge decks having both simple as well
as complex configurations with ease and confidence. Analysis of bridge
decks employing grillage analogy is possible on commonly available PCs
while retaining the accuracy and versatility of other refined methods that
l y usually require larger computational facilities. A considerable saving in
time in the analysis of the bridge is achieved and the method also provides
a 'feel' of the bridge, behaviour to the designer. Although, the accuracy of
any method of analysis for a particular structure is difficult to predict, the
method of grillage analogy is found to be fairly accurate when compared
to methods like FEM.
The book is mainly intended for professionals and students. Consultants
and researchers who are confronted with the problem of analysis for bridge
design and those who wish to specialize in the subject, will also find it
viii Preface
The designer and consultant handling a variety of structural problems

faces difficulties when confronted with the task of analysis of a bridge deck,
whether simple or complex. Ordinarily, one would like to seek solutions in
the least possible time and, as far as possible, would like to avoid the
cumbersome mathematics involved, without compromising on accuracy. This
objective is not easily realised. One may either have to develop his own
program or to modify available programs to suit the specific requirements.
For this, considerable time, thorough understanding and a lot of confidence is
essential. A ready to use computer program of analysis based on grillage
analogy has been made available in the book. A discussion on the evolving of
the program including the basic assumptions and applications of the concepts
used in grillage analogy has also beep provided for the discerning user. This
endeavour will go a long way towards computerizing the analysis of bridges,
in this country and elsewhere.
The book begins with an introduction of the recent developments in.
the area of bridge analysis, design and construction. Specifications for
bridge loadings recommended by Indian Roads Congress (IRC) and
adopted in India, are described. The loading standards of some other
countries are compared with Indian. Standard Loadings.
A brief review of the important methods of analysis of bridges including
grillage analogy is undertaken and the applicability of each method to
various types of bridges having different plan geometry and. support condi-
tions are discussed in Chapter 2. The merits and shortcomings of each
method is also dealt with.
The procedure for formulation and assembly of matrices using direct
stiffness method which is more suitable for mathematical modeling of plane
grillage, is illustrated, followed by a simple but generalized computer pro-
gram in Chapter 3. The listing of this program is given in Appendix I.
Chapter 4 idealizes the actual bridge deck into a suitable mathematical
model of a grillage. The equivalent elastic properties are evaluated and
assigned to the members of the grillage. Analysis of the idealized grillage
for loading is described in Chapter 5. The interpretation of results obtained
and the local effects to be included in the final design are also outlined.
Chapter 6 discusses a more elaborate computer program written in
FORTRAN based on grillage analogy applied to bridge decks. The Program
Manual and Users' Manual are provided. These are explained so that any
designer, not willing to go through the previous sections of the book and with
little exposure to the structural behaviour of bridges, can still prepare the
input data and analyze a bridge for different IRC loadings. The program
44

A I L
Preface ix
is so designed that loads other than Indian standard loadings, can also be
easily incorporated. The listing of the program is given in Appendix II.
A number of worked out examples of different types of slab, T-beam and
box-girder bridges are given in Chapter 7 to explain the use of the program.
The example are chosen from actual life-size bridges and the solutions are
obtained for IRC loadings and also for user's specified loadings. Input and
selected output modules are given for the convenience of the user.
Exhaustive and relevant references are included after each chapter for
the benefit of the readers.
A diskette containing two programs of the Appendices (in a ready to
use form) can be ordered from the authors through the publisher. The
diskette also contains the input data and exhaustive force responses/output
of all the worked out examples given in the book.
Although the general methods and concepts postulated by Lightfoot, West,
Hambly, Jaeger, Bakht and others are further developed and subsequently
expanded, we are deeply indebted to them and to many others whose
writings, teachings or personal help have shaped our thinking and approach to
the subject matter.
The financial assistance and other help rendered by the Curriculum
Development Cell of I.I.T. Delhi towards writing the book is gratefully
acknowledged. The sabbatical leaves granted to the authors by I.I.T. Delhi
and I.T. BHU, Varanasi, respectively for this joint venture are also thank-
fully acknowledged. We are indebted to our research scholars who helped
in developing and checking the computer programs and to our numerous
students for assistance in worked examples and proof checking. The works
of preparing drawings and typing of the manuscript was done by different
persons at different times. We acknowledge their help.
Any suggestions for further improvement from the readers would be
very much appreciated.
C.S.SURANA
RAMJIAGRAWAL
Contents
Preface
1. Introduction
L1 General 1
1.2 Recent Trends in Analysis and Design of Bridges 2
1.2.1 Structural Systems 2
1.2.2 Computer-Aided Methods of Analysis 3
1.23 Design Methodology 4
1.2.4 Modern Construction Techniques 4
1.3 Structural Forms of Bridge Decks 5
1.4 Form of Construction 6
1.4.1 Slab Bridge 6
1.4.2 Slab-on-Girders Bridge 9
1.4.3 Box-Girder Bridge 12
1.5 Plan Geometry or Planforms 15
1.6 Support Configurations 16
1.7 Bridge Loadings 16
1.7.1 Loading Requirements 18
1.7.2 Dead Loads 18
1.7.3 Live Loads 18
1.7.4 Impact Loads 25
1.7.5 Footway, Kerb, Railing and Parapet Live Loads 26
1.8 Comments on Loading Standards 28
1.9 Organisation of the Text 30
References 33
2. Methods of Bridge Deck Analysis

- 2.1 I ntrod uctio n 35
2.2 Methods of Analysis and their Applicability 35 2.3
Courbon's Method 36
xii Contents
2.4 Orthotropic Plate Theory 38

2.5 Finite Difference Method 40
2.6 Method of Harmonic Analysis 42
2.7 Grillage and Space Frame Analogy 45
2.8 Folded Plate Analysis 47
2.9 Finite Element Method 48
2.10 Finite Strip Method 50
References 52
3. Stiffness Method Applied to Grillage Analysis 55

3.1 Introduction 55
3.2 Matrix Method of Structural Analysis 55
3.3 Degrees of Freedom and Sign Convention 56
3.4 Member Stiffness Matrix 58
3.5 Assembly of Structure Stiffness Matrix 63
3.6 Solution of Simultaneous Equations 65
3.7 Computer Program 66
3.8 Example 69
References 73
4. Transformation of Bridge Deck into Equivalent Grillage 74

4.1 Introduction 74
4.2 Idealization of Physical Deck into Equivalent Grillage 75
4.2.1 IdeAlinition of Deck Structure 75
4.2.2 General Guidelines for Grillage Layout 78
4.2.3 Grillage Idealization of Slab Bridge 79
4.2.4 Grillage Idealization of Slab-on-Girders Bridge 90
4.2.5 Grillage Idealization of Box-Girder Bridge 92
4.3 Evaluation of Equivalent Elastic Properties 94
4.3.1 Flexural Moment of Inertia, I 96
4.3.2 Torsional Inertia, .1 96
4.3.3 Flexural and Torsional Inertias of Grillage
Members: Slab Deck 99
4.3.4 Flexural and Torsional Inertias of Grillage Members:
Slab-on-Girders Deck 101
4.3.5 Flexural and Torsional Inertias of Grillage Members:
Box-Girder and Cellular Deck 103
References 109
Contents xiii
5. Application of Loads, Analysis, Force Responses

and their Interpretations 111
5.1 Introduction 111
5.2 Evaluation and Application of Loads 111
5.3 Identification of Panels in the Grillage 115
5.4 Transfer of Loads to the Nodes 117
5.4.1 Transfer i of Dead Loads 118
5.4.2 Transfer of Live Loads 118
5.5 Grillage Analysis and Force Responses 131
5.5.1 Analysis of Grillage 132
5.5.2 Force Responses 134
5.5.3 Design Envelopes 138
5.6 Interpretation of Results 139
5.6.1 Slab Bridges 139
5.6.2 Slab-on-Girders Bridges 143
5.6.3 Box-Girder Bridges 143
References 145
6. Computer Program 146

6.1 1ntroductiOn 146
6.2 Important Features of the Program 'GABS' 146
6.3 Program Manual for 'GABS' 147
6.3.1 Variables 147
6.3.2 Sign Conventions 150
6.3.3 Main Program and Flow Charts 150
6.3.4 Description of Subroutines 151
6.4 User Manual for 'GABS' 165
6.4.1 Data Input Module 165
6.4.2 Result Output Module 167
6.5 Limitations and. Scope 169
References 169
7. Illustrative Examples 171

7.1 Introduction 171
7.2 Illustrative Examples 172
7.2.1 Example 1: Right Slab Bridge 172 7.2.2
Example 2: Skew Slab Bridge 179 7.2.3
Example 3: Voided Slab Bridge 183 7.2.4
Example 4: Right T-Beam Bridge 187
xiv Contents
7.2.5 Example 5: Skew T-Beam Bridge 193
7.2.6 Example 6: Box-Girder Bridge 199
References 203
Appendix I: Listing of Program GRID 205

Appendix II: Listing of Program 215
GABS Index 251
Chapter 1
Introduction
1.1 , GENERAL
Bridge construction has been one of the important engagements of mankind
from the earliest days. Bridges are one of the most challenging of all civil
engineering works. It has always fired the imagination of people as they
seem to lead to hitherto uncharted territory.
Bridge construction tqday has achieved a world-wide level of importance.
The nuiribers and sizes of bridges have continuously increased in the last
fifty years. Man's increasing mobility through railway and motorised
transport has caused such complex forms of bridges to be built, which had
seemed unrealistic earlier. To cope-up with this demand, tremendous efforts
all over the world in the form of active research in analysis, design and
construction of bridges is continuing.
Over the years, a number of methods of analysis of bridge superstruc-
tures have been evolved and are being used. Courbon's method, Hendry-
Jaeger method and Morice and Little method are some of the methods
which have been in use since long, and, are still popular, as they are found
to be easy, amenable to design graphs and also reasonably accurate for
bridge decks of simple configurations. But these methods are being gradu-
ally replaced where computer facilities are available or more accurate
analysis is desired or the cross-section and/or layouts of the bridge decks
are complex.
Following the advent of digital computers, computer-aided methods like
Finite Element, Finite Difference, Finite Strip have been developed and are
in use to analyse intricate forms of skew, curved, bifurcated and arbitrary
shapes of bridges having usual support conditions and cross-sections. But
these methods are highly numerical and always carry a heavy cost-penalty_
Grillage Analogy is probably one of the most popular computer-aided
methods for analysing bridge decks. The method consists of representing the
actual decking system of the bridge by an equivalent grillage of beams. The
dispersed bending and torsional stiffnesses of the decking system are
assumed, for the purpose of analysis, to be concentrated in these beams. The
2 Grillage Analogy in Bridge Deck Analysis
stiffnesses of the beams are chosen so that the prototype bridge deck and
the equivalent grillage of beams, are subjected to identical deformations
under loading. The actual deck loading is replaced by an equivalent nodal
loading. The method is applicable to bridge decks with simple as well as
complex configurations with almost the same ease and confidence. The
method is easy to comprehend and use. The analysis is relatively inexpen-
sive and has been proved to be reliably accurate for a wide variety of
bridges. The grillage representation helps in giving the designer a feel of
the structural behaviour of the bridge and the manner in which the loading
is distributed and ,eventually taken to the supports.
The book essentially deals with the Grillage Analogy method and its
applications to variety of bridge decksboth simple and complex. But it is
also thought relevant to introduce in brief other existing methods of analysis
to the readers (Chapter 2).
In order to apprise the readers with the developments in bridge engineer-
ing, the present chapter discusses The recent trends in analysis and design of
bridge decks. The structural forms of decks based on types of construction,
planforms and support conditions have been outlined. The loading standards
used in India for concrete highway bridges, are given in some detail. Indian
Roads Congress (IRC) loadings are compared with the standard loadings
adopted by some other developed countries.
1.2 RECENT TRENDS IN ANALYSIS AND DESIGN OF BRIDGES

Bridges form vital links in the communication system and the need to build
bridges across wide rivers with alluvial and scourable beds, deep gorges, open
seas and grade separators on urban highways, calls for the solution of a
multitude of engineering problems. The ultimate aim is to evolve the most
'efficient design with the available resources and a technical know-how
matching the site conditions. It should emphasise on quality and -the life
cycle-cost rather than the minimum material used or minimum cost at the time
of construction.
Structural engineering and construction technology have undergone a sea
change in the last three decades and these have had a great impact on modem
bridge engineering. Significant new developments in types of structural
system, computer-aided analysis, design methods and modern construction
techniques have taken place in recent times. These trends and their scope are
briefly outlined in the following sections.
1.2.1 Structural Systems

In all recent developments of medium- and long-span concrete bridge
construction, prestressing has played a Pentral role and roes* '2rid-cs hui:t today
are made of structural concrete using prestressing,. This technology is
Introduction 3
now over fifty years old and has proven its superiority, reliability and
economy. The technology of prestressing has undergone a lot of change and
the technique of external prestressing was `reinvented'.after more than forty
years. External prestressing has a number of advantages like ease of con-
struction and better quality and in the case of corroded or broken tendons,
these can be exchanged. Cable stayed bridges present a special case of
external prestressing as the stay cables also introduce'prestressing force in
the girders of the bridge_
In addition to external prestressing, there has been a trend to reduce the
dead load of the superstructure leading to new types of bridge systems
known as Alternate Web Systems. Alternate Web System consists of web
trusses made of concrete with folded plate webs of corrugated steel
sheets. An example of such an innovative design is the bridge at Maupre
in France [73
The introduction of partial prestressing, which covers the range from
zero prestressing to full prestressing, allows a much wider application of
prestressing. In.bridge design, partial prestressing leads to a simpler distri-
bution of prestressing cables in the longitudinal direction. Local highly
stressed zones are covered by ordinary reinforcement. In the transverse
direction, partial prestressing is the only -alternative to ordinary reinforce-
ment because the stresses due to traffic can easily exceed those due to dead
load. Furthermore, imposed deformations resulting from temperature,
creep, settlement etc. can cause tensile stresses in fully prestressed bridges
resulting in cracks of large widths. It can be expected that partial prestress-
ing eventually will gain universal acceptance with full prestressing and
reinforced concrete, as the two limiting cases.
It appears that prestressing will remain the driving force behind new
developments in bridge engineering in the near future.
1.2.2 Computer-Aided Methods of Analysis

The introduction and application of computers in planning, design, analysis,
construction management and safety control of bridges led to a revolution in
bridge construction. In the sixties, the problem of indeterminacy posed serious
difficulties to bridge engineers. Today, most engineers and design_offices
have easy access to personal computers and also special purpose structural
software packages have been developed mainly based on computer-aided
methods like finite element, finite difference, finite strip, folded plate and
grillage analogy. These packages are selectively used depending-upon the
complexity of the problem, both for linear and non-linear analyses. However,
the user friendliness of most application
softwares need substantial improvements. The major impact, we can
experience today, however, comes from the use of software with enhanced
graphics capabilities and CAD/CAM systems.
With improved analysis capabilities, structural optimization becomes a
reality which in the past had just been a subject of interest to academicians.
Expert. Systems -will be the next step towards efficient structural optimization.
1.2.3 Design Methodology

In the field of reinforced concrete design, the Limit State Concept is
gaining increasing acceptance and this has improved the performance
characteristics of structures to be built. Two types of limit states, namely
collapse limit state corresponding to maximum load bearing capacity (e.g.
bending, shear, torsion, buckling, fatigue etc.) and the serviceability limit state
which is related to the criteria governing normal use of the structure (e.g crack
width, deflection, vibration, durability etc.) have been incorporated into codes
of practice and specifications for bridges of many countries. In India, while
concept of limit states have already been introduced for building works, the
relevant specifications for bridges are not yet available.
Different design formats have been proposed for structural codes in order
to account for uncertainties both in the relevant load combinations at the
ultimate state and in the strength of structural members or systems. Both
simple as well as more refined methods like partial safety factor method,
companion action factor method, load-reduction factor method etc. are being
used. Given the wide range of_ design situations, it is evident that great
precision of the load combinations can not be expected in the near future.
The design of bridges for earthquake and accidental actions, which are
capable of producing forces exceeding elastic limit of structural members,
is gaining importance. Therefore, in order to protect bridges from these
actions, it is necessary to absorb the energy .at suitable locations by
providing plastic hinges. The inelastic deformations of such hinges are
characterised by the ductility ratio. Accordingly, the design for accidental
actions calls for both strength of members as well as ductility of joints and
members and needs careful consideration.
1.2.4 Modern Construction Techniques

In the last thirty years, construction methods have experienced a rapid
development. The method of construction influences the cost of the bridge
to a significant extent. It is prudent to conceive the method of construction
at the time of selecting the type of bridge for the site and keep this in mind
while designing the bridge.
Introduction 5
Concrete bridges are usually constructed on stationary falsework only in

the case of a small number of spans and also when the superstructure is
located. not at a large height above the ground. For other situations, highly
mechanised construction methods are used involving repetitive
construction operations.
Long span bridges are usually constructed using either launching girders
(mechanical formwork) or: segmental cantilever construction. Medium-span
bridges are constructed adopting the incremental push launching method,
classical balanced cantilever method or cantilever method with launching
gantry i.e. segmental construction.
The importance of the concept of prefabricating entire cross-sectional
blocks and combining these with a span-by-span gantry system is increas-
ing. Prefabrication is very economical for short-span bridges. However,
large number of bearings and joints have a negative effect on durability and
user comfort. Improvements are possible by connecting the girder elements
at the support to form a continuous beam using post-tensioning. .
The objectives of a good construction method should be, to build a
bridge at a minimum cost and with a maximum margin of safety during the
construction phase. The pre-requisite for efficient construction, especially
in the case of long span bridges, is therefore an optimum combination of
design and construction methods. How some of the major developments
affecting analysis, design and construction of concrete bridges, discussed
briefly above, will influence the bridge engineering in future, depends upon
the designer, contractor, client, authorities and last, but not the least, on the
freedom given by Bridge Codes.
1.3 STRUCTURAL FORMS OF BRIDGE DECKS

A bridge may be classified in many ways depending upon its function,.
material of construction, form or type of superstructure, plan geometry,
support conditions or span. It is neither intended to discuss here the choice of
a particular type of bridge for a specific situation nor to present a detailed
routine bridge classification, which may be available in any textboot on
bridge engineering. The present book is concerned essentially with the
analysis of highway bridge decks and hence the main factors which govern
and influence the choice of analytical techniques, to be discussed, are only
identified. They are:
I. Form of construction or type of deck
2. Plan-geometry or planform and
3. Support conditions
6. Grillage Analogy in Bridge Deck Analysis
Each of the above parameters is discussed and illustrated. The description
will be limited to only those types of bridges which can be gainfully
handled by employing the method of grillage analogy. The decks of rein-
forced concrete, prestressed concrete and composite construction shall form
the basis of our discussion in the following sections.
1.4 FORM OF CONSTRUCTION

The principal forms of bridge deck construction have been reviewed and
categorised in this section. Broadly, the forms of construction can be divided
into slabs (solid, voided and pseudo-slabs), slab-on-girders (T-beam and I-
beam) and box-girders (single, multi-cell and multi-spine). Arch, rigid frame,
truss, suspension and cable-stayed bridges are not included in our
discussions as these are not amenable to Grillage Analogy.
1.4.1 Slab Bridge

Slab bridges are easiest to construct and are frequently used for compara-
tively .smaller spans. The form is very efficient at distributing point loads
because of its two-way spanning ability and high torsional strength. It is
relatively easy to construct and this is reflected in its construction cost. The
principal disadvantage is its high self-weight which can be counteracted to
some extent, by providing suitable. variation in thickness or by providing
voids. It may be of reinforced concrete or of prestressed concrete.
Solid reinforced concrete slab of constant depth is normally used for
spans upto 10 m (Fig. 1.1a). For larger spans, say upto 15 m, haunching or
variable depth is adopted to reduce dead load (Fig. 1.1b). A solid slab of
uniform depth is. preferred in highly skewed crossings, particularly if sig-
nificant curvature and variation in width of the deck is involved.
Continuous construction can be used with advantage if the possibility of
uplift at abutments is expected.
Voided slab bridges (Fig. 1.1c) are adopted to reduce the self weight of the
bridge. The voids are usually circular or rectangular. The depth of voids is
generally restricted ,to sixty per cent of the depth of the slab so that the slab
continues to behave like a single plate. If this limit of void-depth is
exceeded, the slab may behave more like a cellular deck (discussed later).
The voids may either run for the full span length or, alternatively, these may
be provided in the central span length only so that solid section is available
near the supports where shear is large.
Voided R.C. slabs with depth upto 100 cm may be adopted for span
range of 8 to 15 m. However, for spans between 15 and 30 m, voided pre-
Introduction 7
___________________________ E
/
a) Solid Slab
______________ E
a
b) Solid Stab
____________________________ fl
1 7:7000007 - -
c) Voided Stab
Figure LI Slab Bridges
stressed concrete slabs of depth upto 1.2 m are cheaper. For moderate skew
crossings having spans of 15 to 25 m, this type of deck with longitudinal
prestressing is useful but for highly skewed crossings, reinforced concrete -
decks are preferred for ease of construction. If the voided section is found
inadequate in shear, it should be kept solid near supports (refer voided slab
example of Chapter 7).
In R.C. slab bridges, span-depth ratio of 15 for simple spans and 20 to 25
for continuous spans are usually adopted for both solid and voided slabs.
For cast-in-situ, prestressed concrete voided slab bridges, this ratio is nearly
30. In precast prestressed voided slabs, the ratio ranges between 25 and 30.
The deck slab overhang, designated as 'a' in Figs. 1.1b and 1.1c may be
provided to produce the desirable aesthetic effect and also to reduce the
dead load and the width of sub-structure.
In many countries, standard precast prestressed beams are employed for
short- and medium-span bridges. These standard beams are closely
positioned across the width of the bridge and in-situ concrete is poured to
give transverse connection in order to create a slab-type deck. This form of
deck is described as Pseudo-slabs. Such type of two-stage casting is also
referred to as Contiguous Construction and the standard beams are termed as
Contiguous Beams.
Many forms of prestressed, precast beams are used in pseudo-slab decks.

The Prestressed Concrete Development Group (PCDG) in U.K. produced a
series of inverted T-beams, 1-beams and box-beams for various spans (Fig.
1.2). The beams are placed side by side and the gaps are filled
In-situ concrete
Shear key 7
a) inverted T-Beams Bond rod
. i . . J. ..
. . . :
r .
. . --. . .c
.
r. 1 .
..1
\
, _,., .
1... ...- '..:.
. . ; :
.....
:. Z.:. . % .
Il
..
Void
b) Box -Beams
Figure 1.2 PCDG Standard Precast Pretensioned Bridge Beams
with in-situ concrete to form an integral slab. Transverse bond rods are
provided just above the bottom flange passing through pre-formed holes in
the precast beams. This type of construction can be used upto 20 m span.
Another standard group of beams, known as M-Beams have been developed
by Cement and Concrete Association (C&CA), U.K., in collaboration with
the Ministry of Transport, U.K. (Fig. 1.3) for short- and medium-span
bridges.
Pre-tensioned cast-in-situ Pseudo-slabs are usually adopted when the
erection of formwork presents no difficulty. The advantage of this form of
construction is that the structure is monolithic and the stress-distribution in
the slab can easily be evaluated. On the other hand, the precast zire.crs v:ith
in-situ concrete filled up, are preferred when there is difficulty in supporting
Introduction 9
Permanent Pocking a) Pseudo- Box

Beams
b) T-Beams
Figure 1.3 C & CA, U.K., Standard M-Beams
the formwork. However, in general, the prestressing of slab, is

uneconomical and the formwork is often heavy.
The concept of Pseudo-slab bridge construction using pre-tensioned
precast beam elements is relatively new to India.
1.4.2 Slab-on-Girders Bridge

Slab-on-girders bridges are by far the most commonly adopted type in
the span range of 10 to 50 m. The majority of beam and slab decks have
number of beams spanning longitudinally between abutments with a thin
slab spanning transversely across the top. T-beam bridges are one of the
most common examples under this category and are very popular because
of their simple geometry, low-fabrication cost, easy erection or casting
and smaller dead loads (Fig. 1.4).
Usually I-section or T-section is used for the beam but T-section is found
to be more efficient. T-beams are economical where depth of section is not a
controlling factor from headroom considerations. The T-beam bridge su-
perstructure may consist of either girders and slab or girders, slab and
diaphragms at the supports or girders, slab, intermediate cross-beams and
10 Grillage Analog' in Bridge Deck Analysis
end diaphragms. However, T-beam bridge with cross-beams extending into

and cast monolithically with the deck slab is found to be more efficient and is
recommended for adoption. Sinip1y supported R.C. T-beam is normally
adopted for spans upto 25 m. Span-depth ratio is generally kept as 10 for.
Services
Precast detachable footpath slab
1
1 1
I
I b
1 I
Cross ben
h
Figure 1.4 T-Beam Bridge
simple spans and 12 to 15 for continuous spans. Higher ratios are possible,
but riding qualities are affected by creep characteristics of concrete. The
girder spacing 'h' (Fig. 1.4) may vary as justified by comparing the cost of
corresponding slab thickness. The usual range of spacing 'h' is between 2 to 3
-
m for these bridges. The stem-width 'b' is about 300 mm. This stem width is
increased to 'B' at the bottom, forming a bulb to accommodate a large
number of reinforcement bars there. This '13' may be kept between 500 to
600 mm. The stem width '6' is increased to '13' in the end-region, to take care
of large shears occuring there.
Slab-on-girders bridge also includes prestressed concrete bridges. Majority
of prestressed concrete bridges, constructed in India are of post-tensioned
type. The bridge decks, with post-tensioned girders suitable for simply
supported construction, may ..have either fully cast-in-situ slab and girders or
deck slab with precast prestressed girders alongwith cross-beams, assembled
together and transversely prestressed (Fig. 1.5a,b). Such types of construction
is convenient for the span range of 20 to 30 m. The span-depth ratio is
usually kept as 20 for simple spans and 25 for continuous spans for
prestressed concrete girder bridges. The girder spacing 'h' (Fig. 1.5a) is
normally kept between 2.0 and 4.5 m. The stem width 'b' should preferably
be a minimum of 300 mm to facilitate prestressing of tendons. The deck slab
overhang 'a' should be provided as required to produce the desirable aesthetic
effect and to reduce transverse moments.
Decks with composite construction are also popular for short- to medium-
span bridges. Composite construction refers to the use of structural elements
made of two materials in combination in such a way that they act
Introduction
11.
Cross ,ocb
beam
a 'I
h _I, h
a) Cast- in- Situ Prestressed Concrete Girder

Deck
25mm gap between precast girders
11111111111111Nal111111enIF
Precast girder
Precast cross- beam
b) Tronsversety Precast Prestressed Girder Deck
Figure 1.5 Bridge Decks with Post-Tensioned Girders
together. Though, normally, this is understood to refer to the use of rolled

steel sections with in-situ concrete slab but it also covers precast prestressed
concrete girders with cast-in-situ R.C. slabs and made to act as T-beams. The
girders and the floor slabs are connected using shear-connectors by which
they work together.
The precast prestressed concrete girders and cast-in-place R.C. deck slab
(Fig. 1.6a) is suited for medium spans i.e. spans between 30 and 60 m. For
such bridges the span-depth ratio of about 18 in case of simple spans and say
20 for continuous spans, are recommended.
The composite construction with steel girder (Fig. I.6b) is economical in
the short span range of 8 to 15 m. The construction *has the advantage of
speedy erection and reduced cost of formwork. The flexural stiffness of a
composite beam is about 2 to 4 times that of the corresponding steel beam
and this property results in reduced deflections and vibrations.
The transition from pseudo-slab to slab-on-girders deck is difficult to
define with complete precision. Pseudo-slab is associated with multiple
longitudinal beams (five or more for a two-lane bridge) at close spacings
whereat, the beams aria sian oriages tall in the category of decks with
smaller number of longitudinal beams under similar situation.
girder
Precast
Cast-in-
situ slab
a) Precast Prestressed Concrete Girders

with' in- situ Stab
n-situ R.C. slab
Steel girder
b) Rolled Steel Girder Sections with Cast-

in- situ R.C. Stab
Figure 1.6 Decks with Composite Construction
1.4.3 Box-Girder Bridge
In recent years, single or multi-ell reinforced and prestressed concrete

box-girder bridges have been widely used as economic and aesthetic solt-
tions for overcrossing, undercrossing, separation structures and viaducts,
found in today's modem highway systems. The main advantage of these
type of bridges lies in the high torsional rigidity available because of the
closed box-section and convenience in varying the depth along the span.
High torsional stiffness gives them better stability and load distribution
characteristics and also makes this form particularly suited for Grade Sepa-
rations, where the alignment of bridges are normally curved in plan. Also,
the hollow section may be used to accommodate services such as water
mains, telephones, electric cables, sewage pipes etc. and the section has an
a d d e d a d va n t 2 o - e of h pi n a B a h r
Introduction 13
The cross-section of a concrete box-girder bridge consists of top and bottom

slabs connected monolithically with vertical webs to form a cellular structure.
The box-girder may be composed of single cell or multiple cells, with or
without side cantilevers. The cells can be either rectangular or trapezoidal,
latter being used increasingly in prestressed concrete elevated roadways. Some
of the commonly adopted cross sections are shown in Fig. 1.7. A multi-cell
deck is distinguished from a voided slab in analysis. In multi-cell decks, the
transverse bending of top slab between webs is also considered. Also, the void
depth in a multi-cell deck may be larger than the limiting values given earlier
for the voided slab, leading to cell-distortions.
p
t w
S
(a) Single -Cell Box-Girder
V./
4 h ___ I
(b) Multi - Celt Box-Girder
T
5
I.-- a
(c) Single-Celt Trapezoidal Box-Girder
Figure 1.7 Box-Girder Bridges
Multi-cell box-girder bridges are constructed with or without transverse

diaphragms. If diaphragms are provided only at the supports, shear defor-
mations in the transverse direction leading to shear-lag and cell-distortions
have to be considered in the analysis. But if additional intermediate dia-
phragms are also provided between the supports, the shear deformations and
cell-distortions are usually small and mav be ored.
In the span range of 20 to 30 m, cast-in-situ multi-cell reinforced concrete

box-girder bridges are widely used. For longer spans, say upto 60 m, post-
tensioned prestressed, cast-in-situ multi-cell box-girders 'have been
employed. Two-cell box-girders have been used for span ranging from 30 to
40 m, while single cell trapezoidal box-girders are built with 30 to 50 m
span. Post-tensioned precast box-girders prove to be economical for larger
spans say upto 100 m.
For spans above 60 m and upto 200 m, segmentally erected prestressed
concrete box-girder bridges with one, two, or more cells, spaced apart, may be
adopted. This form of construction is commonly used in viaducts and is
sometimes known as Spine, Spread or Spaced Box-girders (Fig. 1.8). Spine
beam bridges, as built in practice, may be defined as structural members whose
breadth and depth are small in relation to their length and are therefore
subjected mainly to longitudinal bending, transverse shear and torsion. Trans-
verse diaphragms are normally provided only over the supports. The bridges
are generally prestressed longitudinally and reinforced transversely unless they
are exceptionally wide and in such a situation they may be prestressed
transversely also. These may be continuous or simply supported.
SI b w o -
a 4. b h
+- b
Figure 1.8 Multi-Cell Spine Box-Girders
The span-depth ratio for R.C. box-girder bridges are generally adopted as
16 for simple spans and 18 for continuous spans. For prestressed cast-in-situ
concrete box girders, this ratio ranges from 20 to 25 depending upon support
conditions. In case of precast prestressed box-girder bridges, the span-depth
ratio is taken between 18 and 20 whereas for spine or spaced box-girders, the
ratio lies usually between 18 and 22. In precast, prestressed multi-cell box-
beams, the ratio can be as high as 25 to 30 [9].
The spacing 'h' of box-girders usually lies between 2.0 and 3.5 m. The web
of box-girder superstructure should have a minimum thickness `ki: of 200
mm. It is often useful to increase the thickness of webs near supports to
provide adequate concrete section for shear resistance. Precast box-beams
ordinarily have a width '17' of 1.0 to 1.2 m and height 's' in the range of 0.6 to
1.2 m (Figs. 1.7 and 1.8). The bottom slab thickness 'd,' is kept
Introduction IS
approximately 1/20 of clear span between webs but it should not be less
than 150 mm and may be increased near continuous supports.
1.5 PLAN GEOMETRY OR PLANFORMS
The horizontal and vertical alignments of a bridge are governed by the
geometries of the highway, roadway or channel it crosses. A bridge may
either be right or skew, straight or curved or any combination thereof.
Typical geometrical planforms of bridge decks are illustrated in Fig. 1.9.
(a) Right Deck (b) Skew Deck
(c) Symmetrical Curved (d) Unsymmetrical Curved

Deck Deck
(e) Arbitrary Shaped Decks
(f) Bifurcated Deck

7,777nSupported edge
Free edge
Figure 1.9 Plan Shapes of Bridge Decks

The simplest form is the right deck but the demand for skew bridges is
increasing because of non-availability of space for traffic schemes. The skew
effect becomes more important in design when the skew angle exceeds 15.
The construction of horizontally curved bridges has increased considerably in
recent years for highway bridges. The need for smoother dissemination of
congested traffic and the limitation of right -of-way alongwith economic and
environmental considerations dictate that the bridge alignments meet the
overall requirements of the highway system. Also, the current emphasis on
aesthetic considerations has motivated increased development in designs
which utilizes curved configurations, either symmetrical or asymmetrical.
Due to the geometric complexities, curved girders are subjected to not only
flexural stresses but also to very significant torsional stresses. The deflec -
tions and stresses in such decks are markedly different from those of right
bridge decks.
There is a large range of arbitrary planforms where the free.edges of the
bridge are non-parallel, support lines are non-parallel or supports are randomly
distributed. Bifurcated decks are needed at motor-way exit or entry and may
have an arbitrary geometry.
1.6 SUPPORT CONFIGURATIONS

Figure 1.10 shows some of the support configurations normally used in
highway bridges.
The simple supports are common with slab bridges or with slab-on-girders
bridges of smaller spans. Cantilever and Balanced cantilever -bridges are
constructed for span range of 35 to 60 m having T-beam or box-girder as
their cross-section. Fully continuous bridges are advantageous for spans over
35 m and are suitable with prestressed concrete girders.
Further, the bridge may be placed on rigid supports or flexible (yielding)
supports. The conventional plate, rocker or rocker-cum-roller bearings pro-
vide rigid supports. However, the recent trend is to favour elastomeric
bearings. This provides yielding supports. These are preferred because of
their low height and low cost and require practically no maintenance. Also,
they are easy to replace. These bearings can cope up with complex defor-
mations of skew and curved geometry.
1.7 BRIDGE LOADINGS

The loading has profound effect upon the design, construction and eventu-
ally upon the cost of any bridge of a given span.
Besides carrying their own weight, the bridge decks are designed for
certain loadings imposed partly by the vehicles and the users and partly by
Introduction 17
(a) Simply Supported Arrangement
(b) Cantilever Arrangement

Anchor Canti lever
Suspended span
(c) Balance
d Cantilever
Arrangemen
t
fi
ft 11
-
(d) Continuous Arrangement
Figure 1.10 Support Configurations of Bridge Decks
nature. In order to maintain uniformity in design, loading standards have

been laid down for the guidance of engineers. Different countries, including
India, have their own loading standards.
In India, these standards for Railway bridges are formulated by the
Research Design and Standards Organisation (RDSO) of the Indian Railways
[17]. For highway bridges, Indian Roads Congress (IRC), a statutory body
formed by the Government of India under the Ministry of Surface Transport,
prepares the Codes of Practices. These codes are complied faithfully in the
design of bridges [18, 19, 20]. The Bureau of Indian Standards (BIS), a body
responsible for the "Standardization" in the country, also brings out
specifications ror images [2 i But the specifications laid down by IRC
supercede those of the BIS, wherever at variance.
r
18 Grillage-Analogy in Bridge Deck Analysis

1.7.1 Loading Requirements
The deck of the highway bridge has to support moving loads in the form of
vehicles, men and materials and transmit their effects to the foundation. It has
also to support and carry the self weight of its various components. The
structure is also subjected to vibrations under moving loads giving rise to
what is known as impact loading.
The details of some other loads and forces such as earthquake,'iwind etc.
which also become important in some cases could be referred :from the Codes
of Practice [18, 19]. Only the important loads to be used in- the analysis of
decks are briefly described here.
1.7.2 Dead Loads

The bridge superstructure is to be analysed for its self weight and dead
loads imposed on it as well. The dead loads imposed on the bridge consist of
permanent stationary load such as that of wearing coat, kerb, parapets etc. In
estimating the dead loads, the unit weights of materials specified in reference
[19] may be adopted.
Dead, loads invariably form a relatively large loading component and
result in significant design forces and deformations. It is, however, never a
problem to either estimate these loadings accurately or compute their effects
on the structure.
1.7.3 Live Loads

The main loading on highway bridges is due to the vehicles moving on it,
which are transient and hence difficult to estimate accurately. In order to
analyse the bridge for these moving loads, IRC Code [19] recommends
certain standard hypothetical loading systems. The bridge is then designed for
the maximum response values under these standard loads.
The live loads usually consist of a set of wheel loads which are patch loads
due to tyre contact area. These patch loads may be treated as point loads
acting at the centre of the contact area. This simplification is found to be
acceptable in the analysis.
According to Indian Roads Congress classification, the main live loads for
road bridges can be put into the following four types [19]:
i) !RC Class A LoadingSingle Lane and Two Lanes

Single lane Class A loading is a train load of eight axles of two wheels
each thus having sixteen wheels in total. The total load of the train is 55.4
Ti :Ai Lai: it..ngth of the train is 2U.i in and the distance
between the first and the last axle is 18.8 in. The minimum clear longitu -
.61
Introduction 19
dinal distance between two successive trains is 18.5 m. The minimum centre
line distance of the wheel-line from the edge of the Kerb works out to 400
mm. The configuration of the load as well as the position of each wheel is
given in Figs. 1.11(a) and 1.11(b).
1200 1200
1-- 8300 f--4800 H-4800 --1
1-/-1.- 3200-F-1- 4300 +3000 +3000+ 3000-1

co.
6001100 1200 to
1 Axle load
Co co. co
1 1 I cD (tonnes)
Class A
c`:
fa irt
co co CO 4:0
. (6 -4 . Class B
(a) Class A and Class B Train of Vehicles
X
4- + -F. + + + +
Direction of Motion
1800
+ +
(Left Most
Front Wheel
H81-320044-1.300 +3000+3000+3000-
1 (b) Plan
1800
(c) Section
on x x
(411. Dimension are in mm)
Figure 1.11 (a) 1RC Class A and Class B Loadings
Class A-two laneS loading consists of two class A -single lane trains placed
side by side at specified minimum clearance. Class A loading is
Clear Roadway
.0- "/".0,0.0.0.0.0 7e..0,/,/ fm.r1_________ Iii),

01W
(d) Cross-Section
Max. w =500 for Class A

Max. w :380 for Class B
Carriage Minimum clearances
way.
widih eg, ft 1
5500 Uniformly increasing 150

lo 7500 from 400 to 1200
>7500 1200 150
(All Dimensions are in mm)

Figure 1.11 (b) IRC Class A and Class B Loadings
adopted on all permanent bridges and culverts to be constructed on State and

National Highways.
ii) IRC Class B Loading

The Class B loading is identical to Class A loading as far as positions of
axles are concerned but the magnitude of axle loads is 60% of the -corre-
sponding loads in Class A vehicle (Figs. 1.11a, b). This loading is intended
for temporary structures, timber bridges and bridges in specified areas.
iii) IRC Class AA Loading

This loading is an alternate loading and one train of Class AA vehicle is to
be considered for every two lanes of Class A loading. It consists of either a
tracked vehicle of 70 tonnes or a two axle wheeled vehicle of 40 tonnes.
Detailed dimensions, kerb distances etc. are given Figs. 1.12 (a) and
1.12 (b). Bridges designed for Class AA loading should also be checked for
equivalent lanes of Class A loading since under certain conditions. heavier
stresses are obtained under such equivalent Class A loading. The nose to tail
spacing between two successive vehicles is specified as 90 m.
Introduction 21
3600
_____ 7200 _____
0) Elevation
4- 1-1-++ ++4- + + -r
Suggested Equivalent
Concentrated Loadings o
un
O
Left Most
Front WheeNt- 1-4--++++++1
[4-9@360z3240 ______
(20 toads 3.5t =70t on Two Tracks)
(ii) Nan
35t 35t
2900
(iii) Cross-Section
Figure 1.12 (a) IRC Class AA Tracked Vehicle
iv) IRC Class 70R Loading
This is the revised version of Class AA loading and consists of tracked
.and wheel loadings. The loading is detailed in Appendix I of IRC Code
[191 The minimum clearance between the road face of the kerb and the
outer edge of the track or wheel is same as for Class AA loading. The
spacing between successive vehicles is 30 m. This spacing is measured
from the rear most point of ground contact of the leading vehicle to the
forward most point of ground contact of the following vehicle.
-43001 -...1 300H-700 .I 300k- -.I 300 1.-.
1 1 1 ___ I
375t 6.25 t 6.25 t 3.75 t
(1) Cross Section
-I-
+Mos +
t---
Lett -
10,2
Front Wheel
Ill\
+
-I- + 4. + I-
k 600 4*
__ 1000 4-
600 --1
2200 ___________
) Plan
Carriage Way Minimum
Width Value of C
Single Lane Bridges
3800 and 300
above
Multi L ane B ridg es
- 5500 600
5500 1200
All Dimensions are in mm
Figure 1.12 (b) IRC Class AA Wheeled Vehicle
70R track loading, as before, weighs 70 tonnes (Fig. 1.13). The track
dimensions are slightly different than those of Class AA track loading. For
design purposes, wherever required, each strip loading could be idealised into
a suitable number of point loads say 8 or 10.
70R wheel loading is of two types: (1) 70R Bogie loading weighing 40
tonnes through two axles each weighing 20 tonnes (Fig. 1.14a) and (2) 70R
train loading weighing 100 tonnes through seven axles, one axle of 8 tonnes,
two axles of 12 tonnes each and four axles of 17 tonnes each (Fig. 1.14b).
Introduction 23
4570
____ 7920 _____
(a) Elevation
t ,+ + + + + + + + -i-
vSuggested Equivalent I
Concentrated Loadings c,
4 + 4 +
Left Most / 1-P-96)457= 4113.1
Front. Wheel (20 toads 03.5t =70t on Two

Tracks) (b) Plan
35T. 35T
14 2900
(c) Cross- Section
Clearance 4C' same as for Class AA Loading

Figure 1.13 IRC Class 70R Track Loading
An axle may have four or eight wheels on it. There are two, four wheel
arrangements and one, eight wheel arrangement leading to three alternate
wheel arrangements termed as Col. 'I', Col. 'm' and Col. 'n' arrangements
[19]. AU axles will have the same arrangement of wheels at a time and all
wheels on an axle will have equal loads. The two alternate four wheel
arrangements namely Col. '1' and Col. 'm' are given in Figs. 1.14(a, b). The
eight wheel arrangement namely Col. 'n' is not found critical and is not given
nere. However, if required, details of Col. 'n' arrangement could be obtained
from IRC Code [19].
I- 4880
20t 20t
(i) Elevation
" --"F
4
0
+
I
0 JCI
CO
ci
-I-
Left Mos +t
Front Wheel (ii) Plan
Tr," -1`4 W /-"F-

/
1-e--
n nb
NI
(iii) Cross-section
Note: (i) Min. ro' same as for Class A A Loading
(ii) Max. 'w' =410
(iii) Either a=450 & b =1480 (CoI.1')
or a=795 & b = 790 (CoL`m1);
so that (2a4b) = 2380
Figure 1.14 (a) IRC Class 70R Bogie Loading
IRC Code [19] also gives in Appendix I certain other types of track and
wheeled loadings. These are lighter than Class 70R tracked and wheeled
loadings discussed above. These are to be adopted if a user specifies these for
the bridge.
For detailed loading Standards and their specifications, the reader is ad-
vised to consult the relevant IRC Codes of Practice mentioned under refer-
ences [18, 19].
The above IRC loads are nlared on the. hriricre clerk anr1 mnveri tong;til_ dinally as
well as transversely in small increments to cover the entire bridge
Introduction 25
1
a -
O T
3960
213 137073050 7'713 0 -
1f
1 t 12t 1 t 1 7 1 17t
t P
(i) Elevation
+.
44 +
++++ +4-
00.4co
4 4
(ii) Plan
(Left Most
Front Wheel
Note : Cross-Section and distances are as for
Bogie Loading shown in Fig. 1.14(a)
(All dimension are in mm)
Figure 1.14 (b) IRC Class 70R Wheel Loading
deck. One Class A or Class B loading can be put on every lane of the
roadway of a bridge. For multi-lane bridges, one lane of Class AA or Class
70 R per two lanes of the carriageway is allowed as an alternate to Class A
loading.
1.7:4' Impact Loads

Another major loading on the bridge superstructure is due to the vibra-
tions caused when the vehicle is moving over the bridge. The theoretical
estimation of this load is quite complex as it depends upon a variety of
factors such as roughness of the surface, spring system of the vehicle, con-
dition of expansion joints at the entry to bridge etc.
The IRC Code [l9], however, recommends definite values of impact
factors for the vehicles for simplifying the analysis. The value of impact load
is expressed as a percentage of the live load, depending upon the material
used in the construction of deck of the bridge, ly pc of Luiu!
the bridge span. This percentage can be calculated using suitable formulae
[19] or could be directly read from Figs. 1.15 (a,b) for both steel and concrete
bridges for different types of IRC loadings.
55 54.5
50
74 . 45
L= Span in m
Lo 40
35
900
30 44 13.5+L
0
25 Steel
Q. 20 450
E 6 +L Concrete
15.4
15
10 8.8
5 I 1 I
0 5 10 15 20 25 30 35 40 45 50
Span (m)
Class A and Class B Loadings
Figure 1.15 (a) Impact Percentage Curves
1...7.5 Footway, Kerb, Railing and Parapet Live Loads

The following provisions have been made for footpath, kerb, railing and
parapet live loadings in IRC:6-1987 [I9].
(i) For all parts of bridge floors accessible only to pedestrians and animals
and for all footways, the loading shall be taken as 400 kg/m2. Where
crowd loads are likely to occur, such as on bridges located near towns,
which are either centres of pilgrimage or where large congregational
fairs are held seasonally, the intensity of footway loading be increased
from 400 kg/m' to 500 kg/m2.
(ii) Kerbs, 0.6 m or more in width, shall be designed for the above loads
and for a local lateral force of 750 kg per metre, applied It -,rizontalt: at
the top of the kerb. If the kerb width is less than 0.6 m, no live load
Introduction 27
430 U
25
O .
4.
20 0 15
u10
"" 5 0 5 10 15 20
30 35 40 45 50
(a)
25
Span (m)
Class AA and Class 70R Wheeled Loading
.7.30
25
025C
2:20
10 STEEL
40 45 50
.7.30
L.= Span in m
025C
Concrete & Steel
4
(
2:20
215 J.
4-
10
U CON C. 8.6
o. 5
F 0 5 10 15 20 25 .30 35
Span (m)
(b) Class AA and Class 70R Tracked Loading
Figure 1.15 (b) Impact Percentage Curves
may be necessary in addition to the lateral load specified above. The

horizontal force need not be considered in the design of the main
structural members of the bridge.
(iii) In bridges designed for IRC vehicular loadings, the members supporting
the footways shall be designed for the following live load per square
metre of footway area, the loaded length of footway taken in each case
being such as to produce the worst effects on. the member under
consideration:
a) For effective span of 7.5 m or less, 400 kg/m2 or 500 kg/m2 as the
case may be as per (i) above.
b) For effective spans of over 7.5 in but not exceeding 30 m, the
intensity of load shall be determined according to the equation:
P =Pi (40L 9 300)
c) For effective spans of over 30 m, the intensity of load shall be

determined according to the equation:
P
= ( p 2 6 0 + 4800)(16.5
L A 15 )
where P` = 400 kg/m2 or 500 kg/m2 as the case may be based on
(i) above.
P = the live load in kg/m2-
L= the effective span of the main girder in m.
W = width of the footway in m.
(iv) . Each part of the footway shall be capable of carrying a wheel load
of 4 tonnes, which shall be deemed to include impact, distributed
over a contact area. 300 mm in diameter; the permissible working
stresses shall be increased by 25 per cent to meet this provision.
(v) The railings or parapets shall be designed to resist a lateral horizontal
force and a vertical force each of 150 kg/m applied simultaneously at
the top of the railing or parapet. These forces need not be considered in
the design of the main structural members if footpaths are provided. In
cases where footpaths are not provided, the effect of these forces shall
be considered in the design of the structural system supporting the
railings and the footpath upto the face of the footpath kerb only.
1.8 COMMENTS ON LOADING STANDARDS

Bridge Design Codes of most of the countries prescribe some or the other
form of standard live loading to be used in the design. It is observed that
there is large variation amongst these live loadings and it is difficult to
imagine that the traffic pattern would be differing so widely especially
amongst the developed countries. A comparative study of highway bridge
loadings of different countries over a span of 100 m for two lane bridges was
made by ThomaS [12]. His graphs for maximum bending moments and shear
forces are reproduced in Fig. 1.16. It will be seen that the American loading
(AASHO loading) gives lowest design values and German loading gives the
highest design values over almost the entire span range considered.
The existing IRC loadings are complicated in their application, especially
if various types of live loadings are to be considered alternately in the
Introduction 29
64
56
E 24
-43 16
c
gel 8
0 20 40 60 80 100
span(m)
(a) Maximum bending moment for two lanes
280 Legend :
240 New
zealand Japan
200
Shear force 00
---France
160 --.---West Germany
--IRC (India)
120
---A ASHO group -
80 -)BS (HA group)
----Sweden
40
0 20 40 60 80 100
Span (m)
(b) Maximum shear force for two lanes
Figure 1.16 Comparison of Highway Bridge Loadings
design to determine the severest effects. Class 70R loading is a newly

introduced live loading and can be taken as a replacement of IRC Class AA
loading. But this loading is also not. simple.
The basis for IRC provisions regarding impact is not clear. No systematic
study has been made to derive realistic impact factor for road bridges in our
country. The impact effect need not be considered for the full length of the
load but needs only be applied to the heaviest axle or the pair of adjacent
wheels causing the maximum bending moment or shear.
The task of the designers will be simplified if some concerted efforts are
made to introduce some degree of uniformity into national loading specifi-
cations with respect to international loading specifications. Newly formed
International Organisation for Standardisation (ISO) has taken up the
responsibility of producing a Standard Loading Code and considerable
progress has been made in this direction. It is hoped that common interna-
tionally acceptable Bridge Specifications including Loading Standards will
emerge in the near future.
1.9 ORGANISATION OF THE TEXT

Over the years, a number of methods of analysis of highway bridges have
been evolved. The method of grillage analogy which is a well-established
and computer oriented method, bridges the gap between overly simplified
hand computation methods and sophisticated finite element and finite strip
methods.
To lay the foundation of the book, the present Introductory Chapter is
devoted to discuss some of the recent developments in the area of bridge
analysis, design and construction that have taken place. Various forms of
highway bridge construction, their planforms and support configurations are
outlined. Loading standards recommended by Indian Roads Congress (IRC)
and adopted in India for highway bridges, are described. The details of some
of the important loadings for which the analysis of bridge decks are carried
out, are also given. The Loading Standards of some of the advanced coun-
tries are compared with Indian Standard Loadings. Critical comments on
Indian loadings are given and the scope and need for a unified loading is
emphasised.
In order to make the reader appreciate the relative usefulness and poten-
tialities of grillage analogy, it is considered prudent to discuss various other
methods of analysis of bridge decks also. A brief review of the important
methods of analysis of highway bridges including grillage analogy is under-
taken in Chanter 2. These methods are described in relation to their historical
Introduction 31
background, methodology, applicability, merits and limitations. Further, the

applicability of these methods to various types of bridges having different
plan geometry and support conditions is also presented in a tabular form.
It is assumed that the readers have sufficient exposure to matrix methods
of structural analysis on which the method of grillage analogy is based. But
to maintain a continuity of discussion when the reader is to deal with a
computer program in the succeeding chapter of the book using direct stiff-
ness method, a brief outline of the stiffness method and the formulation and
assembly of stiffness matrix kw skeletal structures are presented in Chapter
3. Gauss-Elimination Procedure and Cholesky's Factorisation method to
solve the large number of resulting simultaneous equations are also dis-
cussed. A simple but a general computer program in FORTRAN illustrating
different steps involved in matrix formulation of skeletal grid having any
planform under externally applied nodal loads, is included and is given in
Appendix I. This will provide a background to understand the comprehen-
sive software developed and later presented in Chapter 6. Also, the designer
will be able to make changes easily in the program, if needed, as per his
requirements. A skew grid is analysed using the program for hypothetical
concentrated load at nodes, as an illustration.
Chapter 4 idealises the actual bridge structure into a suitable equivalent
mathematical model of a grillage consisting of longitudinal and transverse
grid lines. The general guidelines for choosing suitable grid layout are given
and also illustrated through typical examples of slab, slab-on-girders and box-
girder bridges. The procedure for the evaluation of equivalent elastic
properties i.e. flexural moment of inertia and torsional inertia T are also
discussed for'the above types of bridge decks.
The longitudinal and transverse members of the idealised grillage form a
mesh having a number of nodes. Grillage analysis requires that the dead,
live and impact loads actually acting on bridge decks, are transformed into
equivalent loads acting on nodes of the mesh. Chapter 5 discusses different
types of loads and identifies the panels in which the wheel loads of a
vehicular live loading fall. The transfer of loads to nodes of grillage in the
form of equivalent nodal loads is also given. The analysis of grillage is then
carried out and resulting force responses and design envelopes are dis-
cussed. The interpretation of results so obtained and the local effects to be
included in the design, are also outlined. Slab, slab-on-girders and box-
girder decks, both of right and skew configuration, are dealt-with.
Chapter 6 presents a computer program based on grillage analogy devel-
oped in FORTRAN for the elastic analysis of bridge decks, covering right
and skew layouts. The important features of the program are highlighted.
The Program Manual consisting of listing of variables, sign conventions,
descriptions of subroutines, flow charts and the Users Manual comprising
data input and output formats for the force responses, are thoroughly .dis-
cussed and explained with illustrations. The designer, not willling to go
through the previous sections of the book and with little exposure to the
structural behaviour of the bridge, can prepare the input data and can still
analyse the bridge for different IRC loadings. The program is designed in
such a way that any loading standard other than Indian loadings can be
easily incorporated without involving much changes in the program. The
program developed can deal with solid, voided and pseudo-slabs, T-beam
and I-beam bridges with or without cross-beams, single and multi-cell
box-girder bridges and composite bridges. The bfidges supported on neo-
prene bearings (yielding supports) can also be handled. The complete
listing of the- program is given in Appendix II. The limitations and scope
of the program are discussed.
Theoretical discussion .bnly does not convince a designer of the versatility
of a technique unless it is supplemented by its application on a variety of
actual problems. To achieve this objective, number of worked out examples
of different types of slab, T-beam, and box-girder bridges are included in
Chapter 7. The examples chosen are the life size bridges and some of them
have already been constructed in India. The bridges are analysed for Indian
Standard Loadings and for users' specified loadings also in some cases. To
enhance the usefulness of the book, some of the important parameters
encountered in practice e.g. yielding of supports, effects of variation in
numbers .and spacings of longitudinal and transverse grid lines etc. are
highlighted through the above worked out examples.
Relevant references are added after the end of each chapter to enable
the reader to consult further for more details.
The book also contains two appendices. Appendix I pertains to the listing
of the general program 'GRID' as .discussed in Chapter 3 and Appendix II
consists of listing of computer program 'GABS'. Diskette with executable
files of these programs, compatible to IBM PC, alongwith input/output
modules for the worked examples, illustrated in Chapter 7, has been pre-
pared and can be ordered from the authors through the publishers.
Introduction 33
REFERENCES
1. BAKHT, B. and JAEGER, L.G., "Bridge Analysis Simplified", McGraw Hill Book
Co., New York, 1985.
2. CUSENS, A.R. and PAMA, R.P., "Bridge Deck Analysis", John Wiley and Sons,
1975.
3. HAMBLY, E.C., "Bridge Deck Behaviour", Chapman and Hall, London, 1976.
4. HARVEY, DAVID [., "Recent Trends in Short and Medium Span Highway
Bridges in the United Kingdom", Proc. Intl. Conf. on Short and Medium Span
Bridges, Vol. I, 1982.
5. KULKA, E and LIN, T.Y., "Comparative Studies of Medium Span Box-Girder
Bridges with Other Precast Systems", Proc. Intl. Conf. on Short and Medium
Span Bridges, Vol. I, 1982.
6. LEE, D.J., "Progress in Bridge Engineering", Developments in Structural
Engineering, Proc. Fourth Rail Bridge Centenary Conference, Vol. I, 1990.
7. MATHIVAT, IE., "Recent Developments in Prestressed Concrete Bridges", FLP
Notes, 1988.
8. Pormuswamv, S., "Bridge Engineering", Tata McGraw-Hill Publishing Co.
Ltd., New Delhi, 1986.
9. RAINA, V.K., "Concrete Bridge Practice-Construction, Maintenance and
Rehabilitation", Tata McGraw-Hill Publishing. Co. Ltd., New Delhi, 1988.
10. ROWE, R.E. and Somenviu.e, G., "Research on Slab Type and Spine Beam
Bridge", Proc. Intl. Conf. on Developments in. Bridge Design and
Cnnstruction. Crosby Lockwood & Son Ltd., 1971.
11. STEINMAN, D.B. and WATSON, S.R., "Bridges and their Builders", Dover
Publications, New York, 1957.
12. THOMAS, P.K., "A Comparative Study of Highway Bridge Loadings in Different
Countries", U.K. Transport and Road Research Laboratory-Supplementary Re-
port 135 UC, 1975.
13. VicroR, D.J., "Essentials of Bridge Engineering", Oxford & IBH Publishing
Co., New Delhi, 1980.
14. WIELAND, M., "Modern Bridge Engineering in Structural Concrete", Proc. of
Asia-Pacific Conference on Road, Highway and Bridge Maintenance and
Rehabilitation, Bangkok, 1987.
15. WrreFoirr, H., "Building Bridges", Bentan Verlog GmbH, Germany, 1984.
16. "Specifications for Highway Bridges", Pt. I and Pt. III, Japan Road
Association, 1984.
17. "Bridge Rules Specifying the Loads for Designing the Super and Sub-
structures of Bridges and for Assessing the Strength of Existing Bridges",
Govt. of. India Publication, Ministry of Railways, New Delhi, 1977.
18. IRC:5-1985, "Standard Specifications and Code of Practice for Road Bridges,
Section I-General Features of Design", Indian Roads Congress, New Delhi,
I99A.
19. IRC:6-I987, "Standard Specifications and Code of Practice for Bridges, Section
HLoads and Stresses", Indian Roads Congress, New Delhi, 1990.
20. IRC:18-1985. "Design Criteria for Prestressed Concrete Road Bridges (Post-
tensicined Concrete)", Indian Roads Congress. New Delhi, 1990.
21. IS:1343-1981, "Indian Standard Code of Practice for Prestressed Concrete",
Bureau of Indian Standards, New. Delhi, 1981.
Chapter 2
Methods of Bridge
Deck Analysis
2.1 INTRODUCTION
Extensive research into the behaviour of bridge decks under loading had
been carried out pver the past five decades and a number of methods of
bridge deck analysis were evolved from time to time. Prior to the general
use of the computer-aided analysis, hand computation methods and charts
based on some approximations and idealizations, provided convenient
methods of load distribution. These were reasonably accurate for design
purposes. However, with the advent of digital computers, many computer -
aided methods have been developed and are in use although some of these
methods are highly numerical and expensive.
The present chapter aims at, giving a broad idea about the various impor-
tant methods of bridge deck analysis. The applicability of the different
methods in relation to bridge type, plan geometry and support conditions is
also presented in a tabular form.
2.2 METHODS OF ANALYSIS AND THEIR APPLICABILITY

Different techniques commonly in use for the analysis of bridge decks of
various types and configurations are:
1. Courbon's Method
2. Orthotropic Plate Theory
3. Finite Difference Method
4. Method of Harmonic Analysis
5. Grillage and Space Frame Analogy
6. Folded Plate Analysis
7. Finite Element Method and
8. Finite Strip Method.
Each of the above techniques is more suited to a particular type or types

of bridge decks depending upon the closeness of the actual structure with
the assumptions of the method. The applicability of the methods to various
types of bridges is illustrated in Table 2.1. It may be evident that one
particular type of bridge deck can be analysed by more than one method
and in such cases, the choice rests with the designer depending upon the
facility, time available, economy and of course his familiarity with the
method.
The above methods will be discussed briefly in relation to their historical
background, applicability, merits and limitations in succeeding sections. For
more details, the references given at the end of the chapter may be useful.
2.3 COURBON'S METHOD

Courbon presented the method [10] at the time when other sophisticated
and more accurate techniques for bridge deck .analysis were not commonly
available. The method makes simplifying assumptions, restricting its appli-
cability to a pertain extent but the method has been very popular because of
its simplicity. The method is applicable to inter-connected T-beam-bridges
and is still in vogue in India and is recommended-by Indian Roads Congress
for live load distribution strictly within its limitations.
The method is recommended to be used when the following conditions
are satisfied :
i) The ratio of span to width is greater than 2,
ii) The longitudinal girders are inter-connected by abOut five symmetri-
cally spaced cross-beams, and
iii) The cross-beams extend to a depth of at least three-fourth of the
depth of the longitudinal girders.
These conditions are not difficult to satisfy in majority of actual T-beam
bridges. According to Courbon's method, the load R. on any girder i of a
bridge consisting of multiple parallel girders (Fig. 2.1) is computed assum-
ing a linear variation of deflection in the transverse direction. The deflection
of Inertia
I .0-1 1
;
Figure 2.1 Deflection Profile in Courbon's Method

TABLE 2.1 Applicability of Analytical Techniques
Type of Deck Plan Geometry Support Condition
Method of
Analysis Slab Pseudo T- Box- Right Skew 'Curved Arbitrary Simply Continuous Arbitrary
Slab Beam Girder and > 15" Supported
Skew
< 51)
Courbon's Method * *
Orthotropic Plate * * * * *
Theory
Finite Difference * * *
Method
Methods of Bridge Deck Analysis 37

Method of Harmonic * * 4.
Analysis
Grillage Analogy * * * * * * * * * * *
Folded Plate * * * *
Method
Finite Element * * * * * * * * *
Method
Finite Strip Method * * * * * * * *
will be maximum in the exterior girder on the side of the eccentric load (or
c.g. of loads, if there is a system of concentrated loads) and minimum on
the other exterior girder.
The load R1 is given by
PI (PLedEL)
R. =L "
EI S E1 , z 4
or,
(2.1)
R.=--P4(1--L1 edij
Eli cf
where I. = moment of inertia of ith longitudinal girder P
= total live load
e = eccentricity of live load (or- in case of multiple loads,
distance of c.g. of loads from centroid of moment of inertias) d.
= distance of ith girder from centroid of moment of inertia
For a N-girder bridge with. all girders
having same moment of inertia, the above
formula reduces to (2.2)
R. = P (i+N edi) 1TT

The above load on the girder is used to determine the bending moment in the
girders. For evaluating shear force in the girders, the same load distribution is
valid for loads lying beyond 5.5 m from either supports. For loads lying at or
within 5.5 m of either supports, the reactions on the longitudinal girders shall
be greater of the results obtained by (i) assuming the deck slab simply
supported or continuous, as the case may be, with the supports being taken as
unyielding and (ii) following Courbon's method [51]. The Courbon's method
under-estimates the load on interior girders and generally over-estimates the
load on the exterior girders. However, these inaccuracies shall be
significantly reduced if the cross-beams are deeper and more in numbers.
With the availability of more accurate methods, this procedure may be used
for preliminary design of girder-sections only.
2.4 ORTHOTROPIC PLATE THEORY

The use of orthotropic plate theory for the approximate analysis of simply
supported right concrete bridge decks was pioneered by Guyon [16] and
Methods of Bridge Deck
Analysis 39 Massonnet [30]. Design curves were prepared by Morice,
Little and Rowe [32] and a summary of the design technique using these
curves had been presented by Rowe [39]. Later Bares and Massonnet [2]
and Cusens & Pama [11] further developed the method.
The design curves are based on the distribution of deflection due to the
first harmonic load. The maximum 'calculated longitudinal moments and -
stresses are increased by 10 per cent to account for dropping the higher
harmonic loading terms in the analysis. This approximation is reasonable
for longitudinal moments but the transverse moments are highly dependent
on the local distribution of harmonic loads and hence superimposition of
higher harmonic components becomes essential.
The bridge deck is replaced by an equivalent plate with bending and
torsional rigidities in two orthogonal directions and the following well-
known partial differential equation, governing the behaviour of the
equivalent system, is obtained,
4
D2'd + 2 H 4 4
+D, d W
P(x,y) (2.3)
d xw 4 d xd 2 y2 1../ Y d y4
where 21-/ = (Dxy, DD + DI + D2)
Dx and D - are the equivalent flexural rigidities and Dxy and D are the
equivalent torsional rigidities per unit width in longitudinal and transverse
directions respectively. Di and D2 are the coupling rigidities per unit width
arising due to Poisson's ratio effects.
The flexural and torsional rigidities have significant influence on the
load distribution. Their effect is considered through two dimensionless
characterising parameters namely flexural parameter 9 and torsional para-
meter a as given below : b Ds 2a
(2.4)
Dy. 30.25
+ D3w+ Di+ D2
and a = _________________________________ (2.5)
2 11 (px
where, 2a and 2b represent span and width of the equivalent plate respec-
tively. The coupling rigidities D and D2 are small and can be neglected
without introducing any significant error.
The dimensionless distribution coefficients, which are given for nine stan-
dard reference points and load positions across the bridge width, are plotted
or tabulated against values of O. The charts or tables are given for two values
of a, namely a = 0 and a= 1. Values of the coefficient Ka for any interme-
diate value of a are obtained by the following interpolation function,
Ka= Ko + (K1 191Tx (2.6)

-where Ke and K1 are the corresponding coefficients for a = O. and a =
1.0 respectively.
For analysis by this method, the applied loads are converted into equivalent
concentrated loads -at the standard locations. The responses corresponding to
each load at these locations are then added to account for the total effect.
Cusens and Pama [11] have improved the distribution coefficient
approach by taking seven terms of the harmonic series and by extending
the range of a upto 2.0.
Another set of curves are available for transverse moment' coefficients `p'
for different values of B and for standard load eccentricities at various pre-
fixed standard stations with reference to the centre line of the bridge deck.
:The Poisson's zatio is found to effect the transverse moment coefficients and
is assumed as 0.15.
The flexural rigidities DF and D of a given slab, T, box or composite
y
section are computed as usual and pose no problem. The torsional rigidity
of the section is evaluated based on St. Venant's method by taking the
summation of the torsional rigidity of the components forming the section
and from this torsional rigidities D and D of the section are computed.
yx
However, the orthotropic plate idealization does not always represent the
physical behaviour of the bridge decks. In bridge with few girders, say 3,
the bending moments obtained are subject to errors, especially if the
bridge is wide and load occupies only a fraction of the width. Also, the
transverse moments which are complex combination of bending between
girders and bending due to non-uniform girder deflections, can not be
accurately obtained.
Also, the method suffers from the drawback of having to assume the deck
to be uniform throughout and also the design curves involve a certain degree
of approximation in use. As only the first harmonic component is used in
load distribution, the method is not recommended to be used- to estimate
shear. Further, the method can not handle skew bridges.
2.5 FINITE DIFFERENCE METHOD

When more complex boundary conditions are encountered in practice, the
methnri of iprtheltrnpic plate, discussed earlier, becomes cumbersome and
'-aw
difficult to apply. The finite difference method is the answer for such complex
boundary conditions. The method is versatile in nature and has wide
applications:-
The finite difference technique had been used to advantage, first by
Neilsen [34], later by Westergaard [49] and applied to bridge decks by
Naruoka and Ohmura [33], Heins and Hails [17], Robinson [38] and many
others.
In this method of analysis, the deck is notionally divided into grids of
arbitrary mesh size and the deflection values at the grid points are treated as
unknown quantities. The usual governing differential equation of an
orthotropic plate is considered in the finite difference method. The differen-
tial equation and accompanying boundary conditions are expressed in terms
of these unknown deflections. The resulting sets of linear simultaneous
equations are then solved for these unknown deflections. Finally, moments
and shear forces are determined from the known deflection pattern. The
curved deflection profile of the deck is approximated by a series of straight
lines and, naturally, accurate results can be expected only if fine grids are
used.
Finite difference equations for various boundary conditiOns like
simple, fixed, free or a combination of free and simple supports can be
written down for each case. However, a fixed edge condition is not treated
very accurately by this method except when very fine grids are used. The
simultaneous equations formulated from these grid points are solved on
digital computers which have matrix packages specially designed for these
type of problems.
The presence of interconnecting beams below the slab, present a special
boundary situation. A better representation of slab and beam interaction can
be found by treating the two as separate structural elements which are made
compatible by satisfying a set of boundary conditions. The interaction of
beams and the top deck is based on the assumptions that beam and deck slab
have identical deflections, the beams have no torsional rigidity and there is no
horizontal force between the beams and the slab.
The simplicity of the trigonometric solution for simply supported right
bridge decks tends to be lost when extended to the problem of skew decks.
However, the finite difference method has been used extensively for skew
bridges also where the grid may be taken along the orthogonal coordinates as
shown by Jensen [22], Robinson [38] or along skew coordinates as suggested
by Naruoka and Ohmura [33].
Most of the solutions and published values of deflections and moments for
skew necks roicr oni:yr to isotropic plates; the mesh is court4.... and acc
racy is doubtful. Naruoka and Ohmura [33] neglected Poisson's ratio. They
dealt with torsionless plate (H = 0) as well as isotropic plate (H = Dx= Dy).
They had difficulty in setting up satisfactory interpolation equations between
these two limiting cases.
.favour [21] produced influence surfaces for deflections and moments for a
wide range of skew orthotropic slabs for both uniform and concentrated loads
making various over-simplifying assumptions. Schleicher- and Wegener [41]
published tables of deflections and stresses for continuous isotropic skew
slabs under uniform loading. Ghali [13] used finite difference equations at
different segments of a_ two-girder skew bridge to determine influence
coefficients for deflection; shear and moment assuming that the applied loads
act directly on girders leading to erroneous conclusions. Robinson [38]
assumed that a concentrated load might reasonably be distributed over one
mesh quadrilateral. Paterson [35] adopted this assumption in developing an
ALGOL computer program based on standard skew mesh over the skew
orthotropic plates.
Although, finite difference method, applied to skew bridge decks, has been
able to solve a large number of bridge structures but still it has certain
inherent difficulties in its use like adoption of fine mesh work, proper as-
sessment of singularities around the obtuse corner, deterioration in conver-
gence with increasing skew 'angles and in some cases, non-compatibility in
boundary conditions. The number of parameters involved in skew bridges are
such that the preparation of design curves to cover a realistic range of
loadings, skew angles and degree of orthotropy, does not appear to be a
practical proposition.
The finite difference method has also been extended to bridge decks curved
in plan by Heins and Hails [17] based on governing equation of orthotropic
plate in polar coordinates, neglecting Poisson's ratio. However, the efforts
needed to analyse a curved deck by this method is cumbersome
and is generally not recommended.
.-
2.6 METHOD OF HARMONIC ANALYSIS
In harmonic analysis, the applied load is broken into a number of harmonic ,
components, each consisting of a distributed load parallel to the longitudinal
axis of the structure and with intensity varying as pure sine-wave as shown in
Fig. 2.2.
11 It X
Mathematically, P(x) a sin ___ (2.7)
,t=3 L
n=1 n=2 n=3 n=4

-+ + +
Figure 2.2 Representation of Load by Harmonic Series
Under the action of each sine-wave load component, every longitudinal.

strip of the structure deflects and twists in a pure sine-wave form. Since
differential of sine function is a cosine function and vice-versa, the equilib-
rium equations, which can be thought of as differentials of deflections, can
also be expressed as a number of sine or cosine functionS. These resulting
equilibrium equations can be solved as conventional simultaneous equa-
tions. The general theory of Harmonic Analysis is described in detail in
many books of mathematics including Kreyszig [24J.
The concept Of harmonics method, applied to bridges, was established by
Hendry and Jaeger [18) and their method is referred as Hendry-Jaeger
method. In this approach, the actual transverse medium is replaced by a
uniform continuous medium of the same total transverse moment of inertia.
For torsionally weak bridges, the load distribution defined through
distribution coefficient `p' is shown to depend upon dimensionless flexural
parameter a, which is defined as
12 (LI3 Eir a
(2.8)
= 7r4 71 El
where L = span of the bridge, h = distance between adjacent longitudinals,

El = flexural rigidity of a longitudinal girder and Elr= total flexural rigidity
of transverse medium.
Hendry-Jaeger produced design graphs between a and p for T-beam right
bridges having 2, 3, 4, 5 and 6 girders [18]. To cater for a bridge having more
than six girders, the bridge is suggested to be converted into an equivalent 6-
girder bridge. Also, for slab bridges, it is recommended that the deck should
be converted into an equivalent 6-girder bridge. To take into account the
transverse positions of loads, a versus p plotting is done for loads placed on
various girders. In order to account for higher harmonics, the method
suggests modifications in the value of a itself using the same
a p plot. As the method can consider higher harmonic components, it can
estimate shear also fairly, accurately.
Hendry-Jaeger had also given analysis for torsionally stiff girder bridges.
The torsional rigidity, CJ of the girder is considered through a
dimensionless parameter /3 which is defined as,
'= 2 h CJ (2.9)
The extreme value of /3 for torsionally stiff bridges has been assumed as
infinity. Design graphs between a and p for 13 = cc have also been
plotted similar to graphs for torsionally weak bridges namely, 13 = 0.
Coefficients for any intermediate value of /3 may be obtained by using
following inter' polation function;
11(3+16 ____________________________________________________
ffrc7) (2.10)
A
where po and p are the distribution coefficients corresponding to 13 = 0

and p= cc respectively and pp is the distribution coefficient for the desired
value of 11
Hendry-Jaeger suggested method to analyse torsionally weak (13 = 0)
three and four girder skew bridges. Only the .first harmonic components
of the loading and deflection were considered. For more accurate analysis
of 3-girder torsionally weak skew bridges, Jaeger, Bakht and Surana [19]
extended the method by incorporating second harmonic term also in the
analysis. Design graphs were also prepared.
Surana [45] developed a general mathematical model for the analysis of
skew girder bridges of finite torsional rigidity incorporating first three har-
monic components for displacement functions and first two harmonic com-
ponents for rotation function. To take into account the effect of skew angle
A, an additional dimensionless skew parameter K = (h/L) tan A., was
introduced. The analysis of right bridge could be obtained by putting A = 0.
Agrawal [1] and Prasad [37] extended the above method further and
gave design tables and charts for use in design offices for three and four
girder bridges.
Harmonic method has advantage in its inherent capability to completely
identify a girder. bridge by a certain combination of non-dimensional
structural parameters. These parameters help in visualising the structural
behaviour due to change in bridge dimensions which in turn help in

decision-making about the trial sections. The method is best suited -for the
analysis of beams and slab right and skew decks with 3 and 4 girders. But
for a bridge with more girders, the number of coefficients to be handled
for a reasonably accurate result, become rather large.
2.7 GRILLAGE AND SPACE FRAME ANALOGY

For any given deck, there will invariably be a choice amongst a number of
methods of analysis which will give acceptable results. When the complete
field of slab, pseudo-slab and slab-on-girders decks are considered,
grillage analogy seems to be completely universal with the exception of
Finite Element and Finite Strip methods which will always cany a heavy
cost penalty for a structure as simple as a slab bridge. Further, the rigorous
methods of analysis like Finite Element Method, even today, are
considered too complex by some bridge designers.
The grillage analogy method can be applied to the bridge decks exhib-
iting complicated features such as heavy skew, edge stiffening, deep
haunches over supports, continuous and isolated supports etc., with ease.
The method is versatile, in that, the contributions of kerb beams and foot-
paths and the effect of differential sinking of girder ends over yielding
supports (such as neoprene bearings) can be taken into account. Further, it
is easy for an engineer to visualise and prepare the data for a grillage.
Also, the grillage analysis programs are more generally available and can
be run on personal computers. The method has proved to be reliably
accurate for a wide variety of bridge decks.
This method of analysis, based on stiffness matrix approach, was made
amenable to computer programming by Lightfoot and Sawko [25]. West
[47, 48] made recommendations backed by carefully conducted
experiments on the use of grillage analogy. He made suggestions towards
geometrical layout of grillage beams to simulate a variety of concrete slab
and pseudo-slab bridge decks, with illustrations. Gibb [14] developed a
general computer program for grillage analysis of bridge decks using direct
stiffness approach that takes into account the shear deformation also.
Martin [28], then followed by Sawko [40] derived stiffness matrix for
curved beams and proclaimed a computer program for a grillage for the
analysis of decks, curved in plan. The grillage analogy has also been used
by Jaeger and Bakht [20] for a variety of bridges.
The method consists of converting the structure into a network of skeletal
members rigidly connected to each other at nodes. The load-deformatiOn
relationship at the two ends of a skeletal element with reference to the
1

member axis is expressed in terms of its stiffness property. This relationship
which is expressed with reference to the member coordinate axis, is then
transferred to the structure or global axis using transformation matrix, so
that the equilibrium condition that exists at each node in the structure can
be satisfied.
The bridge structure is very stiff in the horizontal plane due to the
presence of decking slab. The translational displacements along the two
horizontal axes and rotation about the vertical axis will be negligible and
may be ignored in the analysis. Thus a skeletal structure will have three
degrees of freedom at each node i.e. freedom of vertical displacement and
'freedom of rotations about two mutually perpendicular axes in the
horizontal plane. In general, a grillage with `71'. nodes will have 3n degrees
of freedom.
All span loadings are converted into equivalent nodal loads by computing
the fixed end forces and transferring them to global axes. A set of simulta-
neous equations are obtained in the process and their solutions result in the
evaluation of the nodal displacements in the structure. The member forces
can then be determined by back substitution.
Bridges are frequently designed with their decks skew to the supports,
tapered or curved in plan. The behaviour and rigorous analysis are signifi-
cantly complicated by the shapes and support conditions but their effects on
grillage analysis are of inconvenience rather than theoretical complexity.
Space Frame idealisation of bridge decks has also found favour with
bridge designers. This idealisation is particularly useful for a box-girder
structure with variable width or depth where the finite strip and folded
plate techniques are inappropriate. However, Scordelis [42] concluded
certain disadvantages of.space frame analysis to the extent that the
computer-time involved is excessive while the solution is still approximate.
Most-road bridges of beams and slab construction can be analysed as
three dimensional structure by a space frame analysis which is an extension
of grillage analogy. The mesh of the space frame in plan is identical to the
grillage, but various transverse and longitudinal members are placed
coincident with the line of the centroids of the downstand or upstand
members they represent. For this reason, the space frame is sometimes
referred to as "Downstand Grillage" [11]. The longitudinal and transverse
members are joined by vertical members, which, being short, are very stiff in
bending. The downstand grillage behaves in a similar fashion as the plane
grillage under actions of transverse and longitudinal torsion and bending in a
vertical plane and consequently, sectional properties of these are calculated in the
'""--"= '
.hirhods of Bridge Deck Analysis 47
Space frame programs have been used in the analysis of box-girder bridge
decks. However, the simulation of boxes by space frames is not capable of
predicting local effects and the method has proved expensive in use.
2.8 FOLDED PLATE ANALYSIS

A folded plate is a prismatic shell formed by a series of adjoining thin plane
slabs rigidly connected along their common edges. A box-girder bridge deck
may be regarded as a.special type of folded plate structure. in that the plates
are arranged so as to form a closed section. Method of analysis originally
developed by Goldberg and Leve [15] for folded plate had been adopted for
the analysis of box-girders. Scordelis [42] initially applied the method of
folded plate to simply supported box-girder bridges and later on extended it
to continuous decks [43] also. An approximate method; known as-Finite
Segment method, was also used by Sc.onielis [42] and Johnson and Lee [23]
in analysing box-girder bridges.
The structure consists of a number of rectangular plates connected at
longitudinal joints. Each plate is initially assumed to be fixed at the longi-
tudinal joints. Edge forces due to surface loads are determined by plate
theory and, for loads in the plane of the plate, by plane stress theory. The
stiffness matrix for each plate is then expressed in terms of the harmonics
of a half-range Fourier series. Each joint has four degrees of freedom i.e.
displacement longitudidally tangential to the joint, rotation about an axis
tangential to the joint and vertical and horizontal displacements. The direct
stiffness method is used to analyse the complete structure. The method is
applied to simply supported structures with support diaphragms. The dia-
phragms are assumed to be infinitely rigid in their own plane but perfectly
flexible in a direction normal to their own plane. The analysis can be ex-
tended to include intermediate diaphragms. Harmonic analysis is used to
represent applied loads.
Folded Plate method is quite suitable for analysis of box-girder bridge
having a few number of cells. The method offers a logical approach in the
sense that it analyses the structure in its correct form without replacing it by
an equivalent structural system. Thus, the field of application of the method
is restricted to the right cellular bridge decks of uniform cross-sections
having intermediate diaphragms but which must be simply supported at the
extreme ends with rigid diaphragms positioned over the supports. However,
within its field of application, the method is efficient in terms of computer
time, is accurate and yields complete information about the elastic stresses in
the structure.
2.9 FINITE ELEMENT METHOD
During the past two decades, the Finite Element Method (FEM) of
analysis has rapidly become a'Arery popular-technique for the computer
solution of complex problems in engineering and the method is now well-
known and established. Its early application to problems of plate flexure
led to its adoption as a convenient tool in the solution of many bridge
deckS where its generality gave it a considerable edge over many other
specialised techniques. The method is able to tackle complex planforms,
irregularly positioned supports, holes in the deck and other anisotropic
features. Thus, the Finite Element Method may seem to be very general in
application and indeed, for difficult bridge deck problems, it is sometimes
the only valid form of analysis.
The technique was pioneered for two dimensional elastic structures by
Turner et al [46] and Clough [9] in the latter. half of fifty's. Since then,
considerable developments have been made in theory and execution of FEM
and it got further impetus due to availability of faster and larger digital
computers. The method was fUrther developed by Zienkiewicz [50], Martin
and Carey [29], Desai and Abel [121 etc. and was applied to bridge decks by
Scordelis [44], Meyer [31], and many others. The current state-of-the-art on
the theory of FEM and its application to bridge structure, is available in
reference 27.
The FEM consists of solving the mathematical model which is obtained
by idealising a structure as an assembly of various discrete two or three
dimensional elements connected to each other at their nodal points, possess-
ing an appropriate number of degrees of freedom. The solution by FEM
essentially involires four basic steps:
1. Discretisation of the structure into finite elements,
2. Evaluation of element properties,
3. Matrix formulation for element assemblage and its solution, and
4. Interpretation of results.
The most important step in the finite element method of analysis is the
formulation of a mathematical model of the actual structure which is
represented as an assemblage of discrete parts, known as elements. Each
element of the model has finite dimensions and properties and in order to
perform subsequent analysis, it is necessary to establish the force-
displacement relationships of each element.
The bridge deck is represented as an elastic continuum and the division
of the structure in 'elements' can be carried out in many different ways. For
example, in a slab deck; the elements may be taken as triangular or quadrilat-
eral plate elements. The representation may be coarse with a small number of
elements or fine using a relatively large number of elements. The actual
choice will depend upon the geometry of the structure, on the importance of
local features and also on the convergence properties of the element.
The usual direct stiffness or diSplacement method of analysis is applied while
assembling the elements and solving the bridge deck for stress responses.
Equilibrium of the internally and externally applied forces at each node of an
element and the compatibility of element deformations are both satisfied. Also,
the internal force-displacement relationships must be established with each
element as governed by the existing geometry and material property
characteristics. The interpretation of stress resultants for design is a cumbersome
process and the job becomes still difficult if large number of elements are
involved and the support conditions and loading systems are complex.
The main advantage of the method over other analytical techniques, is its
generality. Normally, as was pointed out, it seems possible, by using many
elements, to virtually approximate any continuum with complex boundary and
loading conditions to such a degree that an accurate analysis can be expected.
In practice, however, engineering limitations arise, the most important being
the cost of the analysis. As the number of elements increases, the manpower
required to prepare the data and interpret the results increases and also a larger
amount of computer time is needed for the analysis. Furthermore, the non-
availability of the softwares and large computers may prevent the use of a large
number of finite elements to idealise the deck. The method invariably requires
a high speed computer and back up storage for solving any real bridge
problem. Also, the round-off and truncation errors occuring in the analysis
because of finite precision arithmatic are further impediments. It is, therefore,
desirable to use only efficient finite element programs which in turn essentially
depend upon the use of efficient finite elements, programming methods and
the use of appropriate numerical techniques.
The application of FEM to bridge problems will need a thorough under-
standing and knowledge of almost all the facets of advanced structural
mechanics and numerical techniques which many a times a design engineer
may not have. Also, the softwares based on FEM are not so easily available
and also the understanding of its use for the bridge deck analysis is difficult.
Versatile as it may be in application, the method can be shown to be
uneconomical in analysing bridge decks of regular shape specially right
structures with simple end suppbrts.
Thus, the method should be reserved for bridge decks which are inca-
pable of solution by any of the simpler and more economical methods.
, 50 Grillage Analogy in Bridge Deck Analysis
2.10 FINITE STRIP METHOD

The finite strip approach when first published by Cheung [5] in 1968-69, was
recognised as having excellent prospects as a method of analysis for simply
supported bridge deck structures M terms of accuracy and efficiency.
Basically, the method is a hybrid procedure which retains advantages of both,
the orthotropic plate method and finite element concept. The procedure is
applicable to both slab and box-girder bridge decks.
The method may be regarded as a special form of the displacement
formulation of the finite element procedure. The basic difference between the
two methods stems from the assumed displacement patterns. The assumed
displacement functions for a finite element in slab or box-girder analysis
normally takes the form of a two-way polynomial functions. The displace-
ment functions assumed for the corresponding finite strip are combination of
a one-way polynomial function (in the transverse direction) and a harmonic
function (in the span-wise direction). The harmonic functions are chosen to
satisfy the end support conditions. In effect, the finite strip spans -, between
two opposite end supports.
In this method, the structure is assumed to be discretised into a number
of strips. Each strip has constant thickness; however, the thickness can
vary from strip to strip. The strip is treated as a beam shown in Fig. 2.3.
The stiffness matrix of a strip with pre-set end conditions is formulated.
The loading can be point load, patch load or line load.
Simply Supported edge
Auxiliary
Nodal Line Strips
Nodal Lines
Figure 2.3 Nodal and Auxiliary Nodal Lines in Finite Strips
The continuum is divided into strips by fictitious lines called Nodal Lines
and these strips are assumed to be connected to one another along discrete
number of nodal lines which coincide with longitudinal boundaries of the
strips. A displacement function in terms of the nodal displacement para-

meter is chosen and then strain and stress fields within each element are
obtained. Based on the chosen displacement function, it is possible to obtain
stiffness and load matrices which are then assembled to form a set of overall
stiffness equations. Since the band-width and the size of the matrix is usu-
ally small, the equations can be solved easily by any standard band matrix
solution technique to yield the nodal displacement parameters.
For more refined solutions of general applicability to both for slab and
cellular decks. of uniform cross-sections, the concept of Auxiliary Nodal
Lines (ANL) between the normalstrip boundaries was introduced by Loo
and Cusens [26]. In the solution bf the basic force-displacement equation, the
unknown amplitudes at auxiliary nodal lines may be; written in terms of the
unknowns at the boundaries Before the matrix equations are assembled from
individual strips. This is a useful property of the ANL technique as it permits
higher order functions to be incorporated into the. finite strip formulation
without affecting the band-width of the overall matrix equation. The resulting
computer time far an individual problem is only marginally higher than for
the. analysis based on the conventional technique using a third order
polynomial function and nodal lines at the boundaries of each strip.
The accuracy of a finite strip analysis depends mainly upon the number
of finite strips used in representing the actual structure and upon the
number . of terms retained for the Fourier series functions.
The finite strip procedure for rectangular slab-type bridge decks had been
first suggested by Powell and Ogden [36] and later on, the method was
adopted by Scordelis [43] and his collaborators in the University of California
at Berkeley, to deal with right and curved box-structures. Research work by
Cheung [6] and his co-workers in applying the strip method for the analysis of
slab type bridge decks with intermediate column supports [4] and simple
curved slab and box-bridges [7] gave further impetus to the methodology.
Qther studis by Cheung on the analysis of rectangular slabs with end boundary
conditions other than ;simple supports and with variable cross-sections in the
span-wise direction [8], however, have not been convincingly efficient when
compared with methods such as finite element analysis. The same criterion
applies to the finite strip analysis of skew slab bridges -suggested by Brown
and Ghali [3]. Loo and Cusens [26] had also been constantly working on the
method and its applications to bridge problems.
The method of finite strip, within its field of applications, is well suited
for computer use. Any form of loading, including prestressing forces,
may be conveniently handled.
However, the method suffers from the drawback that it is ideally
suitable to only prismatic structures with simply supported ends.
Apparently, it seems that the method has not been applied to skew girder
bridges with diaphragms and only right or circularly curved bridges can be
analysed. Further, each finite strip is assumed to have constant geometry
and material properties in longitudinal direction.
As can be seen froth the discussions of various important methods of
bridge deck analysis, practically every method has its merits and limitations
and some of them even have limited applicability. However, grillage analogy
method seems to be a general, simple, sufficiently accurate, easy to compre-
hend and convenient to work even on easily available Personal Computers.
Therefore, the remaining part of the book deals with different aspects of
Grillage-Analogy method in detail.
REFERENCES
L AGRAWAL, R., "Analysis and Design of Interconnected Skew Girder Bridges",
Ph.D. Thesis, Indian Institute of Technology, New Delhi, 1975.
2. BARES, R. and MAssozorET, C., -"Analysis of Beams and Grids and OrthOtropic
Plates by the Guyon-Massonnet-Bares Method", Crossby Lockwood, London, 1968.
3. BROWN, T.G. and GuAu, A., "Finite Strip Analysis of Skew Slabs", Proc. McGill-
EIC Conference in Finite Element Method in Civil Enginering, 1972.
4. CHEUNG, M.S., CHEUNG, Y.K. and CHAU, A., "Analysis of Slab and Girder
Bridges by the Finite Strip Method", Building Science, Vol. 5, 1970.-
5. CHEUNG,IK., "The Finite Strip Method in the Analysis of Elastic Plates with Two
Simply Supported Ends", Proc. ICE, 40, 1968.
6. CHEUNG, Y.K., "Analysis of Box-Girder Bridges by Finite Strip Method", Proc.
2nd Intl. Symposium on Concrete Bridge Design Chicago, ACI Publications, SP 26,
1969.
7. CHEUNG, Y.K., and CHEUNG, M.S.,,"Analysis of Curved Box-Girder Bridge by
Finite Strip Method," Publication, IABSE Vol. 31/1, 1971.
8. CHEUNG, Y.K., "Finite Strip Method in Structural Analysis", Pergamon Press,
Oxford, England, 1976.
9. CLOUGH, R.W., "The Finite Element in Plane Stress Analysis", Proc. 2nd ASCE
Conf. on Electronic Computation, Pittsburg, Pa., 1960.
10. COURBON, J., "Application de la Resistance des Materiaux au Calculdes Pants",
Dunod, Paris, 1950.
11. CUSENS, A.R. and PAMA, R.P., "Bridge Deck Analysis", John Wiley, London, 1975.
12. DESAI, C.S. and ABEL, J.F., "Introduction to Finite Element Method", Von
Nostrand Reinhold, New York, 1972. 1
13. GHAI I, A., "Designs of Simply Supported Skew Concrete Girder Bridges", Proc.
Intl. Symposium on Concrete Bridge Design, SP 26 Toronto, 1969.
14. -Gm A., "Grillage Analysis. Notes for Course on Bridge Deck Analysis", Civil
Engg. Dept., University of Dundee, 1972.
IS. GOLDBERG, J.E. and LEVE, ELL., "Theory of Prismatic Folded Plate Structures",
Publication IABSE, VoI. 17, 1957.
16. GUYON, Y., :`Calcul des Pants Larges a Poutres Multiples Solidarisees Par des
Entretoises", Annales des Ponts et Chausees, No. 24, 1946.
17. HEINS, C.P. and HAns, R.L., "Behaviour of Stiffened Curve Plate Model", J.
Struct. Div., ASCE, Vol. 95, ST11, 1969.
18. HENDRY, A.W. and JAEGER, L.G., "The Analysis of Grid Frameworks and
Related Structures", Prentice-Hall, Englewood Chatto & Windus, London, 1958.
19. JAEGER, L.G., BAiurr, B., and SURANA, C.S., "Application and Analysis of
Three-Girder Skew Bridges", Proc. Second Intl. Colloquium on Concrete in
Developing Countries, Bombay, 1988.
20. JAEGER, L.G., and &waif, B., "The Grillage Analogy in Bridge Analysis", Cana-
dian Journal of Civil Engineering, 9(2), 1982.
21. JAVOUR, J., Ikew Slabs and Gridwork Bridges", Bratislava Czechoslovakia:
Slovenske Uydavataletsvo Technickij Litratury, 1967.
22. JENSEN, V.P., "Analyses of Skew Slabs", Bulletin Series No. 332 University of
Illinois, Illinois, 1941.
23. JOHNSON, C.D. and LEE, T., "Long Non-prismatic Folded Plate Structures", J.
Struct. Div., ASCE, Vol. 94, 1968.
24. KREYSZIG, E., "Advanced Engineering Mathematics", John Wiley New York, 1962.
25. LIGEar-oar, E. and SAWKO, F., "Structural Frame Analysis by Electronic
Computer: Grid Frameworks Resolved by Generalised Slope Deflection,"
Engineering, 187, 1959.
26. Loo, Y.C. and CUSENS, A.R., "The Finite Strip Method in Bridge Engineering",
Viewpoint Publication, Cement and Concrete Association, London, 1978.
27. MAISEL, B.I., "Review of Literature Related to the Analysis and Design of Thin
Walled Beams", Technical Report 42:440, Cement and Concrete Association,
London, 1970.
28. MARTIN, H.C., "Introduction to Matrix Method of Structural Analysis," McGraw-
Hall Book Co., New York, 1966.
29. MARTIN, H.C. and CAREY, G.F., "Introduction to Finite Element Analysis",
McGraw-Hall, New York, 1973.
30. MASSONNET, C., "Methode de Calcul des Ponts a Poutres Multiples Tenant
Compte de Leur Resistance a la Torsion", Publication, IABSE, No. 10, 1950.
31. MEYER, C., "Analysis and Design of Curved Box-Girder Bridges", Report SESM
70-22, University of California at Berkeley. 1970.
32. MORICE, P.B., LITTLE, G. and RowE, R.E., "Design Curves for the Effects of
Con-Pui-micatIon uri I a, Lcmctit and Concrete Associaton,
London, 1956.
33. NARUOKA, M. and OHMURA, I-I,. "On the Analysis of a Skew Girder Bridge by
the Theory of Orthotropic Parallelogram Plates", Proc. 1SBSE, No. 19, 1959.
34. NEILSEN, N.J., "Bestemmelese of Spaedinger i Piader Ved Anvendelse of
Differensligninger", Doctoral Dissertation, College of Engineering,
Copenhagen, 1918.
35. PATERSON, D.K,W., "Load Distribution in Skew Orthotropic Plates", Ph.D.
Thesis, University of DUndee, 1970.
36. PoWELL, G.I-1 and OGDEN, D.W., "Analysis of Orthotropic Steel Plate Bridge
Decks", Proc. Struct. Div., ASCE, ST5, V.95, 1969.
37. PaAsAn, J., "Modified Hormonics Method for Analysis and Design of Skew
Girder Bridges", Ph.D. Thesis, Indian Institute of Technology, New Delhi, 1982.
38. ROBINSON, K.E., "Behaviour of Simply Supported Skew Slabs Under Concentrated
Loads", Research Report No.8, Cement and Concrete Association, London, 1959.
39. ROWE, R.E., "Concrete Bridge Design", C.R. Books Ltd., London, 1962.
40. Sawx.o, F., "Computer Analysis of Grillages, Curved in Plan", Publication,
IABSE, 1967.
41. SCHLEICHER, C. and WEGENER, B., "Continous Skew Slabs: Tables for Statical
Analysis", Verlog fur Bauwesen, Berlin, 1968.
42. SCORDELIS, A.C., "Analysis of Simply Supported Box-Girder Bridges", Report
SESM 66-17, Dept. of Civil Engg., UniVersity of California at Berkeley, 1966.
43. SCORDELIS, A.C., "Analysis of Continous Box-Girder Bridges", Report SESM
6725, University of California at Berkeley, 1967.
44. SCORDELIS, A.C., "Analytical Solutions for Box-Girder Bridges", Proc. Intl. Conf.
on Developments in Bridge Design and Construction, Crosby Lockwood, London,
1971.
45. SURANA, C.S., "Interconnected Skew Bridge Girders", Ph.D.Thesis, University
of Edinburgh, U.K., 1968.
46. TURNER, M.J., CLOUGH, R.W., MARTIN, H.C. and.Torp, L.J., "Stiffness and
Deflection Analysis of Complex Structures", J, Aero Sci., 23, 1956.
47. WEST, R., "The Use of a Grillage Analogy for the Analysis of Slab and Pseudo-
Slab Bridge Decks", Research Report 21, Cement and Concrete Association,
London, 1973.
48. WEST, R., "Recommendations on the Use of Grillage Analysis for Slab and
Pseudo-Slab Bridge Decks", C&CA/CIRIA. Cement and Concrete
Association, London, 1973.
49. WESTERGAARD, H.M., "Formulas for the Design of Rectangular Floor Slabs and
Supporting Girders", Proc. ACI, 22, 1926.
50. ZIENKIEWICZ, 0.C., "The Finite Element Method in Engineering Science",
McGraw-Hill, London, 1971.
51. IRC:21-1987, "Standard Specifications and Code of Practice for Road Bridge:
Section III-Cement Concrete (Plain and Reinforced)", Indian Roads Congress,
New Delhi, 1991.
Chapter 3
Stiffness Method Applied to

Grillage Analysis
3.1 INTRODUCTION
The bridge deck structure may be considered as an assembly of structural
members connected together at discrete nodes forming a grid. The
deformations and forces at nodes are inter-related by corresponding
stiffnesses. In order to satisfy 'the equilibrium and compatibility conditions
at each node, a large number- of simultaneous equations will result and the
manual solutions of these may be prohibitive. But using the matrix method
of structural analysis as a primary approach, it becomes possible to obtain
the computer-oriented solution.
It is assumed that readers have sufficient exposure to matrix methods of
structural analysis. But to maintain the continuity of discussion in succeeding
chapters, a brief presentation of stiffness method and its formulation for
bridge deck analysis by Grillage Analogy is included. A simple but general
computer program, illustrating different steps involved in matrix formulation
needed, is also presented. A sample grid is analysed as an example.
3.2 MATRIX METHOD OF STRUCTURAL ANALYSIS

For an elastic structure, the actions P (forces and moments) and displace-
ments D (translations and rotations) are directly related as
[F] (P} (3.1)
where [F] is known as Flexibility Matrix of the structure and io defined
as the displacement produced by a unit value of action P. Similarly,
another way of relating P and D is
{P} = [K] {D} (3.2)
where [K] is the Stiffness Matrix and is defined as the action required to
produce a unit displacement.
The above two matrix approaches, namely Flexibility Method and Stiff-
ness Method, are commonly employed in the analysis of skeletal structures.
In the flexibility method, also known as Force Method, the redundant
structure is converted into a 'released' or statically determinate structure by
the removal of sufficient internal or external actions. The solution of the
problem consists in finding the values of those actions which will restore
compatibility of the displacements at the joints and supports of the structure.
The unknowns are, therefore, static unknowns.
In the stiffness method, also known as Displacement Method, the redun-
dant structure is converted into a 'locked up' or fully restrained structure by
locking of every joint and support. The solution of the problem then consists
in finding the values of the displacements which must be applied to all joints
and supports to restore equilibrium conditions at the joints. The unknowns
are, therefore, kinematic unknowns.
Thus it can be seen that the above two methods approach the problem
from two different angle and this represents the essential difference
between them.
It is generally agreed that the stiffness method is more suitable for the
analysis of structures than the flexibility method. One of its advantages over
the flexibility method is that it is conducive to computer programming Once
the analytical model of a structure has been defined, no further engineering
decisions are required in the stiffness method in order to carry out the
analysis. In flexibility approach, the reduced structure is- to be solved a
number of times as the choice of redundants is not unique.
The stiffness method will be used in the analysis of bridge decks by
grillage analogy.
The method is based on certain assumptions. .The important among
these are; (a) Hooke's law appliesleading to the principle of superimpo-
sition, (b) small deformation theory holds true and (c) shear deformations
can be ignored.
The assembly process in stiffness matrix in its original form is
computationally inefficient needing more computer time and storage. This
inefficiency has been overcome in its improved version called Direct Stiff-
ness Method which is readily programmable on digital computer.
Various steps in direct stiffness method are discussed in the following
sections.
3.3 DEGREES OF FREEDOM AND SIGN CONVENTION

Before we discuss the formulation of stiffness matrix for a member, the
concept of Degrees of Freedom (D.O.F.) for a structure is essential. The
degrees of freedom for a structure are the independent deformations which
Stiffness Method Applied to Grillage Analysis 57
define the deformed shape of the structure completely. In general, any

structure has six degrees of freedom at a node e.g. three components of
translation along three orthogonal axes and three components of rotation
about them. These degrees of freedom can be written in the form of a
vector, known as displacement vector. This vector is generally partitioned
into two sub-vectors, namely active displacement vector containing the
degrees of freedom along which free displacement is possible and passive
or support displacemOt vector which corresponds to the restrained degrees
of freedom at support's.
Depending upon the significance of a deformation, in a particular type-
of structure form, some of the above degrees of freedom can be ignored,
being insignificant. Then the total number of degrees of freedom get re-
duced. For example, in a grid structure, the two translations in the plane of
grid and a rotation about the axis perpendicular to the plane of grid are
insignificant and may be neglected. Thus, the degrees of freedom in the
case of grid are reduced to only three at each node. These are translation
perpendicular to the plane of grid and rotations about two orthogonal axes
lying in the plane of the grid, as shown in Fig. 3.1(a) for a member of the
grid in X-Y plane. These are called Global or Structural Degrees of Free-
dom. For a member, the degrees of freedom are six in number, three at
each end of the member (Fig. 3.1b).
83
0
2
(a)
'31
L (b)
Figure 3.1 Global Axes and Degrees of Freedom
Figure 3.2 depicts a member T in conjunction with a set of member

axes Xr., Yr and Z.. Member axis (X.) makes an angle y with the global
X-axis and the member axis (Z.) is parallel to global Z-axis. The possible
displacements of the ends of the member '1' in the directions of member
axes are known as Local or Member Degrees of Freedom.
58 Grillage- Analogy in Bridge Deck Analysis
Zm
Figure 3.2 Local Axes and Degrees of Freedom
In the above figure, the single headed arrows denote translations and the
double headed arrows represent rotations and are taken as positive in the
directions shown following right hand screw rules.
3.4 MEMBER STIFFNESS MATRIX

Consider a grid member (Fig. 3.3) that is fully restrained at ends j and k.
Member stiffnesses consist of reactions exerted at the member ends by the
restraints when unit deformations (one translation and two rotations) are
imposed at each end of the member in turn. It is assumed that the shear
centre and the centroid of the member coincide so that twisting and bending
of the member occur independent of one another.
Ym
3 6
E1, CJ
Xm k 5
Zm (a) L (b)
Figure 3.3 Grid Element with End Displacements
The unit displacements are considered to be induced one at a time while

all other end displacements are zero.
The member stiffnesses corresponding to six possible types of end dis-
placements as shown in Fig. 3.3, are summarized pictorially in Figs_ 3.4 (a)
and 3.4 (b). Forces are represented by single headed arrows and moments
by rinnhle hpnr.id right hand screw rule as mentioned
r -r---
Yrn 1E1
_6E1 L2
pp-XM
-_ 2E1
Zm 12E1 L3
L3
(I) Unit 'Translation
Along Zm at End
Ym
G7
k 'Xm
1.0
Zm
(ii) Unit Rotation About Xm at End 3
2, L
E
1 r. __ Xm
4E1Z m ' 1 . 0
6E1 6E1
L2
(iii) Unit Rotation About Ym at End
Figure 3.4 (a) Determination of Member Stiffnesses Corresponding to Degrees
of Freedom at End j
earlier. All vectors are drawn in positive senses, but in cases where the
restraint actions are actually negative, a minus sign precedes the expression
for the stiffness coefficient. In general, it is possible for the member to
undergo any one or more of the six displacements shown in Fig. 3.4.
All the member stiffnesses shown in the figures are derived by determin -
ing the values of restraint actions required to hold the distorted member in
equilibrium. The stiffness matrix [K] for a grid member as above, therefore,
is of the order 6 x 6 and each column in the matrix represents the actions
caused by corresponding unit displacement. The member stiffness 'matrix
thus obtained is symmetric aria is snown in equation 3.3.
Ym
6E1 6E1
_ Xm
J 0
k
12(1
12E1
L3
(iv) Unit Translation Along Zm at End k
vm
G GI
GAKt X m
1.0
Zm
(v) Unit Rotation About Xm at End k

Ym
4E1
2E1
17
6E1
L 2 6 E 1
L2
(vi) Unit Rotation About Ym at End k
Figure 3.4 (b) Determination of Member Stiffnesses Corresponding to Degrees of
Freedom at End k
The stiffness matrix developed above for a grid member is in terms of

local degrees of freedom, which is different for different members meeting at
a joint. Since, the equilibrium at a joint is to be satisfied, taking into account
the end actions of all members meeting at that joint together with the external
forces if any, a common reference is essential. This is provided by global
degrees of freedom and a relationship between local and global degrees of
freedoms is needed. This relationship is obtained in terms of a matrix known
as Transformation or Rotation Matrix.
Consider a grid member 1-2 in Fig. 3.5 with member axes Xm and Yrn. The
relation between the deformations at node I in the original direction and in
the direction of global axes X and Y is shown. These deformations are related
(Eq. 3.4) with the help of a 3 x 3 rotation matrix [A] in terms of direction
cosines.
"V1'
Sti f fness Method Applied to Grillage Analysis 61
12 El 6E! 12 El 6 El
0 0 2
L2 L3 L
GJ _GJ
0 0 0 0
6E1 L 4 EI 6 El L 2 El
[K.] L 2 L L (3.3)
12 El 0 6 El 11E1 0 6 El.
L3 L3 L2
L2 GJ
0 GJ 0 0
6 EI L 2 EI 6 El L 4E1
L2 L L2 L
Ym Y
Y Xm
m5
ems
Figure 3.5 Transformation of Stiffness Matrix
l 0 0
= 0 cos y sin y (3.4)
_0 sin y cosy_
Now, local member deformations can be related to global member deforma-

---411......+nommo11111111111
62. Grillage Analogy in Bridge Deck Analysis
tions as given in equation 3.5.
cosy siny
0 sin y cos y I
I
tyesm3
Wf
tn4
0 cosy siny ,,
W m]
10 0I
I 0 siny cosy_ O',6
or {c1} = [T] (3.5)
where {dm} and {d.'} are local and global member deformation vectors
respectively and [T] is a 6 x 6 transformation matrix derived from the rota-
tion matrix fill such that
(3.6)
{Pm} = [T] {P:,} (3.7)

[P.} and {P} are the member end action vectors in local and global
degrees of freedom respectively.
Since .transformation matrix [T] is an orthogonal matrix, its inverse
and transpose should be same. Therefore, we can write,
{P',n} = [T]T fP,1 (3.8)
We may write for actions and displacements referred to local axes X. and
Y. as
{P.} = [Kin] {din} (3.9)
and referred to the global axes X and Y as
{Pm} = .1 {d',,} (3.10)
Substituting for {P.} in equation 3.8 above from equation 3.9, we get,
(Pim? = [TY- {dj
or {P,} = [TI [lc [7] {d:}
T
(from equation 3.5)
Thus from equation 3.10 above,
[rir,] = [Tr. [Km] [7] (3.11)
Stiffizess Method Applied to Grillage Analysis 63
Thus, the member stiffness matrix in terms of the global axes is found by
operating on the local member stiffness matrix [IC an] given by equation 3.3
earlier, using the transformation matrix. This takes into account the
material constitutive laws and orientation of the member. For a grid mem-
ber, the final global member stiffness matrix is shown in equation 3.12.
12E1
R
;n:
L3.
6E1 GJ C2 4E1
34
".n2 s CS
Y?
mi3
6E1 GJ4E1"L LGJ52+4E1C2
12E1 6E7 6E1 12E1

L' L2 L2 c L3
M",5 6E1 GJ2 L 2E1 2 GJ 2E1 6E1 GI 4E1 2
Les L _ +-S L L2 L 2 L
S-
M 4 +1 76 L
6E1_ a...ILI 2 6E1 GJ 4E1 GJ 2 4E1
2
L2 L L L L L2 TCS---ECS TS
+r-C
(3.12)
where s = sin y, c = cos y
3.5 ASSEMBLY OF STRUCTURE STIFFNESS MATRIX

The member stiffness, developed in previous sections, gives the relation
between actions and deformations of a single member and satisfies member
constitutive laws. But to satisfy equilibrium condition at any joint, we have to
consider assemblage of all the members say 'n', meeting at that joint. The
structure stiffness matrix element satisfying the joint equilibrium can be
obtained by an assembly of 'n' member stiffness matrix elements. Symboli-
cally, this assembly process can be represented as
[KJ] = I icno. (3.13)

i=i
where [K1] is the assembled structure stiffness matrix corresponding to jth

degree of freedom and [Kr is the corresponding term of ith member
stiffness matrix. This assembled matrix is nothing but the addition of all
internal forces which will be subsequently equated to the externally applied
forces along the same degree of freedom.
Generally this assembled structure stiffness matrix [K) is quite large in size
and hence some technique is to be adopted so that this matrix occupies
the minimum possible storage space in the computer. One of the basic
properties of this matrix is its symmetry and banded nature, advantage of
which is taken to reduce the storage memory (RAM) required. The matrix is
having non-zero terms only in a restricted region which can be bounded by
two lines parallel to the leading diagonal as shown in Fig. 3.6. This is called
band of stiffness matrix and maximum number of non-zero elements in a row
is called band-width of the matrix. The number of elements to be
HalfBand Sky line
K1 0 K13 0 Kik 0 0 0 0 0
K22 0 K25 K.7) 0 0 0 0
K23
K33 K34 K35 K36 )0 0 0 0

Kh4 /(45 Kt.6 K47 KI.8 0 0

K 5 5 K 5 6 K 57 K 5$ 0 0

K66
1(67
K68
1(77 1(78K69
K79

Symmetric K88 K89 1(81

Figure 3.6 Half-Band and Skyline of a Matrix
stored are further reduced by the fact that the lower triangle part of the
matrix can be obtained from the upper triangle part and vice-versa. So only
half band of the matrix need be stored and rest of the elements can be
deduced from this, whenever required. The half band-width of stiffness
matrix of a structure can be obtained as
Half Band-Width (Degrees of freedom per node) x (maximum difference
of numbers of connected nodes f 1)
Storage requirements can be further reduced by using Skyline Technique for
assembly. As shown in Fig. 3.6, Skyline is an envelope drawn in matrix nose
zero Ci erfierit S which do not have any non -zero elements
o

above them. So the Skyline always remains below the half band boundary line
and thus contains smaller number of elements to be stored.
The stability of the structure is now considered by introducing the boundary
conditions. There are two procedures available to take into account the
boundary conditions. In one, the supports are idealised by a lumped stiffness
which can be thought of by proiiding stiff springs along the supports. This is
achieved by adding a high stiffness term corresponding to the passive degree
of freedom. The reactions in that case can be obtained by multiplying the
spring stiffness by the corresponding displacement.
Another approach is based upon the condition that the displacement along
the restricted degrees of freedom i.e. the passive displacement vector, is a null
vector. Considering this, structure stiffness matrix is partitioned into four
submatrices as given in equation 3.14.
[K pp 7, KpRilDp} If} (3.14)

KTpR K RR DR R
where [Km] etc. are partitioned submatrices and {pp} and {DR} are free and
[IC p
restrained displacement vectors respectively. (PI is the external load vector

and {R} is the reaction vector.
From above, we can write
[Kpp] {Dp} [KpR] {DR} = {P} and - (3.15)
[KPR]T {Dp} [KRR] {DR} = {R} (3.16)
In case of rigid supports,{D R} = 0 and the equation 3.15 reduces to
[Kpp] {Dp} = {P} (3.17)
The solution of the above equation will determine the deformation vector
{Dd. This value of {Dp} is substituted in the equation 3.16 to get the
reaction, i.e.
[IV {Dp} = {R} (3.18)
The first method can handle flexible as well as rigid supports, while the
second method is applicable to rigid supports only.
3.6 SOLUTION OF SIMULTANEOUS EQUATIONS

A large number of simultaneous equations will result depending upon the size
of the grid chosen. They are to be solved in an efficient manner utilising
the minim'', computer t;.,... itcratvc., .7
and Gauss-Seidel Method was one of the most popular techniques to

solve the simultaneotis equations. But the method suffers from the disadvantage that the time taken
by computer to solve the equations achieving a particular degree of accuracy can be predicted only
approximately.
Now-a-days, Gauss-Elimination Procedure and Cholesky's Factorization Method [5]
are commonly used to solve the simultaneous equations. The Gauss-Elimination
Procedure consists of making all the terms of stiffness matrix, K for j < i, equal to zero
i.e., the stiffness matrix is reduced to upper triangle matrix by eliminating all the terms
below the leading diagonal. Then deformations are calculated by back substitution.
The Cholesky method consists of_factorization_olstiffness-matrix [K]-into
[KL] and [Ku] such that
[K] = [KL] [Ku] (3.19)
where KL = lower triangle matrix
Ku = upper triangle matrix
Then, from equation 3.2, we can write
{P} = [KL] [Ku] {130}
Or {P} = [KL] {Q} (3.20)
where {Q} = [Ku] {D} (3.21)
{Q} can be obtained by Forward substitution from equation 3.20 as [KL] is a lower
triangle matrix and {D} can be obtained by Backward substitution from equation
3.21 as [KU] is the upper triangle matrix. The vector {D} thus obtained, gives the
nodal deformation of the structure. The member deformations in global as well as
local coordinates can be evaluated from the nodal deformations {D} obtained
above. From member deformations, member forces can be calculated by
multiplying member stiffness matrix with the member deformation vector.
An exhaustive treatment of above sections on stiffness matrix and solution of
simultaneous equations are available in references given at the end of the chapter
and the reader may refer the same for more details.
3.7 COMPUTER PROGRAM

For illustrating the basic steps of computer oriented direct stiffness method
applied to grids, as discussed above, a simple program 'GRID' is presented in
this section. The program will help the reader to understand the approach
more clearly. This will also lay a foundation for a more comprehensive
computer program to be encountered later in Chapter 6. Figure 3.7 gives the
flow chart of the program.
Stffness Method Applied to Grillage Analysis 'V
START
/ I N P U T : G R I D D
n o d e s , N o . o f e
C o o r d i n a t e s o f
M e m b e r p r o p e r t
c o n n e c t i v i t i e s ,
E T A I L S N o . o f
l e m e n t s ,
n o d e s ,
i e s a n d
H a l f b a n d - w i d
Generate stiffness matrix and

assemble global stiffness matrix
for all members
i
TypeINPUT: SUPPORT
of support, No. of DETAILS
supports,
Supported
in case of flexible supportconstant
nodes, Spring
Modify diagonal terms

of stiffness matrix accordingly
Decompose stiffness matrix Subroutine "DECOMP"
INPUT: LOADING DETAILS

No. of loads, Loaded nodes,
Value of load
Generate load
vector
4r
Solve generated Subroutine

load vector "SOLV"
4.
Print
nodal deformations
4.
DO 300
Loop for solving for
all members

Compute
member end forces
Print
member end forces
Figure 3.7 Flow Chart of Program 'GRID'

The computer program described here can be used to determine the de-
formations and member forces in a skeletal grid having any planform under
externally applied nodal loads. The prograril is written in FORTRAN lan-
guage. In this program, Cholesky method is used for the solution of simul-
taneous equations. The program can handle both flexible and rigid type of
supports. In order to reduce the volume of input data, the procedure for
automatic coordinate generation has been adopted. The subroutine
`DECOMP' has been usOd to decompose the stiffness matrix by Cholesky's
method and the solution is obtained by using subroutine `SOLV%
For simplicity, the transfer of loads from panels of the grid to its nodes
has not been dealt with in the program and the forces are assumed to act
only at nodes. Many other complicated, but important features, pertaining
to bridge deck analysis, have:also not been incorporated in the program
and these will be discussed in. Chapter 6.
A list of principal variables used in the program and their description is
given in Table 3.1 alongwith the symbols. The program listing is given in
Appendix I at the end of the book. A diskette containing the program (in a
ready to use form) along with the Input and Output files of the example
can be ordered from the authors through the publishers.
TABLE 3.1 : Variable Description
Variable Symbol in Text Description
X(1),Y(I) Coordinate of ith node
NCN1(I),NCN2(1) j,k Node numbers corresponding
to the ends of ith member
X1( 1) I Moment of inertia of ith member
XJ(I) J Torsional inertia of ith member
NNODES n Total number of nodes
NELEMS m Total number of members
NDOF 3*n Total number of degrees of freedom
NHBAND Half band width of stiffness matrix
Young's modulus of elasticity
G G Shear modulus of elasticity
SM(I,J) Member stiffness matrix
c,s cos y, sin y Cosine and sine of the angle y, the
member is making with global X-axis
P(1) (P) Load vector
D(I) {DJ Deformation vector
GSM(1,1) [K] Global (structure) stiffness matrix
TM(I,J) [TI Transformation matrix
3.8 EXAMPLE
111- ar.d nut of inaluies of the above program have been
illustrated with the help of a simple example of a skeletal grid shown in Fig. 3.8.
Simply
supported
Notes:
(i) Load of 10,000 kg at node11
(ii)Dimensions in cm and
E,Gin kg/cm 2
Figure 3.8 Skew Skeletal GridExample
It consists of three longitudinals and seven transverse members. The longi-

tudinal members are simply supported at the ends on rigid supports. The
transverse members are in skew alignments. The grid is to be analysed for a
vertical load of 10,000 kg. applied at node 11 located at the center of the
grid. The numbers given in boxes in the figure represent the member num-
bers. The Young's modulus of elasticity E and shear modulus of elasticity G
are taken as 142860 kg/cm2 and 62110 kg/cm2 respectively.
The input data with description for this example is given in Table 3.2 and
the result output of the program is presented in Table 3.3. The moment of
inertia I and torsional inertia J for the elements of each group are given as .
input.
The results are self-explanatory. The deformations at all the nodes and
shear force, bending moment and torsional moment for all the members of the
grid are only given for brevity. However, the program appended also has
thc = all thc, in= data, nodal coordinates, moment
of inertia I and torsional inertia J of all the members. The reactions at
support nodes 1, 2, 3, 19, 20 and 21 are to be obtained by algebraically

summing up the shear force in the members meeting at the respective nodes.
TABLE 3.2: Input Data
Description Input Data
Number of Nodes, Number of Elements Number 21,32
of groups describing nodal coordinates From 3
node no. to node no., Node number increment,
X-coord., Y-coord., X-increment,
Y-increment 1,19,3,0.,0.,200.,0.
Same for remaining groups 2,20,3,250.,250.,200.,0.
3,21,3,500.,500.,200.,0.
Number of groups describing member sequences 5
From member no. to member no., Member no.
increment, Node 1, Node 2, First node increment,
Second node increment 1,6,1,1,4,3,3
Same for remaining groups 7,12,1,2,5,3,3
13,18,1,3,6,3,3
19,25,1,1,2,3,3
26,32,1,2,3,3,3
E, G 142.86E3, 62.11E3
No. of groups having different member properties 3
I, J, total no. of members in group, member 370.75E5, 16.5E5,18,1,2,3,4,5,6,
numbers of the group 7,8,9,10,11,12,13,14,15,16,17,18
Same for remaining groups 5.36E6, 5.987E6,4,19,25,26,32
6.67E6, 6.72E5,10,20,21,22,23,
24,27,28,29,30,31
Half band width of matrix 12
Type of supports (1 for rigid and 0 for flexible) 1
No. of supports 6
Node numbers at which supports are provided 1,2,3,19,20,21
No. of loads 1
Node no. at which load is acting 11
Type of load, magnitude of load 1, 10000.
Note: Units kg.cm.
72 Grillage Analogy in Bridge Dec* Analysis
TABLE 33 : Output Data

Total no. of Nodes -7- 21
Total no. of Elements = 32
Half Band-Width = 12
Total no. bf D.O.F. = 63
E = 142860.00 G = 62110.00
Deformations
Nodb No. Vertical Deflection X-Rotation Y-Rotation
I -0.166858E-09 0.502802E-04 0.465766E-04
2 -0.250818E-09 0.641898E-04 0.715232E-04
3 -0.189220E-09 0574660E704 0.484959E-04
4 - 0.887135E-02 0.107143E-04 0.403060E-04
5 - 0.140837E-01 0.641827E-04 0.668466E-04
6 -0.915430E-02 0.717644E-04 0.410519E-04
7 -0.154419E-01 -0343813E-04 0.239530E-04
8 0.256665E-01 0.431990E-04 0.456668E-04
9 -0.157153E-01 0.819048E-04 0.230262E-04
10 -0.179865E-01 -0.721953E-04 0.901100E-06
11 -0.307446E-01 - 0.472005E-18 - 0.329927E-18
12 -0.179865E-01 0.721953E-04 -0.9,01100E-06
13 -0.157153E-01 -0.819048E-04 -0.230262E-04
14 -0.256665E-01 -0.431990E-04 -0.456668E-04
15 -0.154419E-01 0.343813E-04 - 0.239530E-04
16 -0.915430E-02 -0.717644E-04 -0.410519E-04
17 -0.140837E-01 -0.641827E-04 -0.668466E-04
18 - 0.887135E-02 - 0.107143E-04 - 0.403060E-04
19 -0.189220E-09 -0.574660E-04 -0.484959E-04
20 - 0.250818E-09 - 0.641898E-04 - 0.715232E-04
21 -0.166858E-09 -0.502802E-04 -0.465766E-04
Member Forces
Member Shear Force Bending Moment Torsion
No. End 1 End 2
1 0.145468E+04 0.205935E+05 - 311530E+06 0/02739E+05
2 0.114928E+04 0.318142E+06 -.547997E+06 0.231074E+05
3 0.470264E+03 0.563450E+06 -.657503E+06 0.193761E+05
4 - .466629E+03 0.680321E+06 -.586995E+06 0.497521E+04
5 -.121693E+04 0.599063E+06 -.355676E+06 - .519601E+04
6 -.158519E+04 0.355656E+06 -.386176E+05 -.732661E+04
7 0.196012E+04 - .721619E+05 -.319862E-F06 0361352E-1-01
8 0.263379E+04 0.297518E+06 - .824277E+06 0.107522E+05
9 0.406311E+04 0.80306R.E+k"..,:.; - .1615.i.;21-07 .71.7 i :13-;
Contd.
10 - .406311E+04 0.161569E+07 -.803068E+06 0.221355E+05
11 -.263379E+04 0.824277E+06 - .297518E+06 0.107522E45
12 - .196012E+04 0.319862E+06 0.721619E+05 0.361352E+01
13 0.158519E+04 0.386176E+05 -.355656E+06 -.732661E+04
14 0.121693E+04 0.355676E+06 - .599063E46 -.519601E-F04
15 0.466629E+03 0.586995E+06 - .680321E+06 0.497521E+04
16 -.470264E+03 0.657503E+06 -.563450E+06 0.193761E+05
17 - .114928E+04 0.547997E+06 - .318142E+06 0231074E+05
18 - .145468E+04 0.311530E+06 - .205935E+05 0.202739E+05
19 - .943386E+02 -.226007E+03 0.335797E+05 -.288976E+05
20 0.305410E+03 -.267163E+04 -.105307E+06 -.667882E+04
21 0.679011E+03 -.135654E-1-05 - .226501E+06 -.828866E+04
22 u.y.sotwit+Os -.263173E+05. - .304925E+06 - .595134E+04
23 0.750304E+03 -.157253E+05 -.249547E-1-06 - .134105E44
24 0.368260E+03 .149277E+04 - .128707E+06 0.152035E4.04
25 -.425352E+02 0.324875E+05 - .174490E+05 0.221261E+05
26 0.425352E+02 0:174490E+05 -.324875E+05 0.221261E+05
27 - .368260E+03 0.128707E+06 0.149277E+04 0.152035E+04
28 -.750304E+03 0.249547E+06 0.157253E+05 -.134105E+04
29 - .936893E+03 0.304925E+06 0.263173E+05 - .595134E+04
30 - .679011E+03 0.226501E+06 0.135654E+05 - .828866E+04
31 - .305410E+03 0.105307E+06 0.267163E+04 - .667882E+04
32 0.943386E+02 -.335797E+05 0.226007E+03 - .288976E+05
REFERENCES
1. BATHE, K.J., "Finite Element Procedures in Engineering Analysis", Prentice-
Hall, USA, 1982.
2. BEAUFAIT, F.W., RowAN, W.H., HOODLEY, P.C. and HACKETT, R.M.,
"Computer Methods of Structural Analysis", Prentice-Hall, USA, 1970.
3. CHAPRA, S.C. and CANALE, R.P., "Numerical Methods for Engineers" McGraw-
Hill, New York, 1989.
4. CUSENS, A.R. and PAMA, R.P., "Bridge Deck Analysis", John Wiley, 1975.
5. GERE, J.M. and WEAVER, W., "Analysis of Framed Structures", Van Nostrand, 1965.
6. JENKINS, W.M., "Matrix and Digital Computer Methods in Structural
Analysis", McGraw-Hill, New York, 1969.
7. RUBINSTEIN, M.F., "Matrix Computer Analysis of Structures", Prentice-Hall,
USA, 1966.
8. SURANA, C.S., "Grillage Analogy for Analysis of Super-Structures of Bridges",
Proceedings of the Seminar on Modern Trends in Analysis and Design, B.H.U.,
Varanasi, February 1984.
9. SURANA, C.S. and HUMAR, J.L., "Beam and Slab Bridges with Small Skews",
Canadian Journal of Civil Engineering, Vol. 11. No. I, 1984.
Chapter 4
Transformation of Bridge
Deck into Equivalent Grillage
4J INTRODUCTION
In recent years, the Grillage Analogy Method, which is a computer-oriented
technique, is increasingly being used in the analysis and design of bridges.
The method is also suitable in cases where bridge exhibits complicating
features such as heavy skew, edge stiffening and. isolated supports. The use
of computer facilitates the investigation of several load cases in shortest
possible time. The method is versatile in nature and the contribution of kerb
beams and the effect of differential sinking of girder ends over yielding
bearings (such as neoprene bearing) can also be taken into account and large
variety of bridge decks can be analysed with sufficient practical accuracy.
Furthermore, the grillage representation is conducive to give the designer a
`feel' for the structural behaviour of the bridge and the manner in which the
bridge loadings are distributed and eventually taken to the supports.
The method consists of 'converting' the bridge deck structure into a
network of rigidly connected beams at discrete nodes i.e. idealizing the
bridge by an equivalent grillage. The deformations at the two ends of a
beam element are related to the bending and torsional moments through
their bending and torsional stiffnesses.
These moments are written in terms of the end-deformations employing
slope-deflection and torsional rotation-moment equations. The shear force
in the beam is also related to the bending moment at the two ends of the
beam and can again be written in terms of the end-deformations of the
beam. The shear and moment in all the beam elements meeting at a node
and fixed end reactions, if any, at the node, are summed-up and three basic
statical equilibrium equations at each node namely EF = 0, EMx = 0 and
EMy = 0 are satisfied.
Transformation of Bridge Deck into Equivalent Grillage 75 ,
In general, a grid having 'n' nodes will have '3n' nodal deformations
and `3n' equilibrium equations relating to these. Back substitution in the
slope-deflection and torsional rotation-moment equations will give the
bending and torsional moments at the two ends of each beam element.
Shear forces are computed from bending moments and external loads.
When a bridge deck is analysed by the method of Grillage Analogy, there
are essentially five steps to be followed for obtaining design responses:
(i) Idealization Of physical deck into equivalent grillage
(ii) Evaluation of equivalent elastic inertias of members of grillage
(iii) Application and transfer of loads to various nodes of grillage
(iv) Determination of force responses and design envelopes and
(v) Interpretation of results
'The first two steps of grillage analogy are discussed in this Chapter. The
other remaining steps are dealt with in the subsequent chapters.
4.2 IDEALIZATION OF PHYSICAL DECK INTO EQUIVALENT

GRILLA(E
The method of grillage analysis involves the idealization of the bridge deck
as a plane grillage of discrete inter-connected beams. This is the first im- ,
portant step to be taken by the designer and needs utmost care and under-
standing of the structural behaviour of the bridge decks. It is difficult to
make precise general rules for choosing a grillage mesh and much depends
upon the nature of the deck to be analysed, its support conditions, accuracy
required, quantum of computing facility available etc. and only a set of
guidelines can be suggested for setting grid lines. It may be noted that such
idealization of the deck is not without pitfalls and the grid lines adopted in
one case may not be efficient in another similar case and the experience and
judgement of the designer will always play a major role.
4.2.1 Idealization of Deck Structure

A rectangular slab-element subjected to loads normal to its plane is
equivalent in deformability to an assembly of six beams [81 as shown in Fig.
4.1. The properties of grillage members. are given as follows:
yL2) t3
x
Ly 24 (1 y2)
2 3
r
ly
2
-j 24 (1 7 )
.(E Ly (1-3y)) t3
G 24(1y2 )
.(E L, (1.-3y) t3
2
G 24(1y )
= 7(121 +14)13 _________________________________________________ t3-
/ (4.1)
LyJ24(1_72)
where y= Poisson's ratio and E, G Elastic moduli.
Ly
(a) A Slab Element
Y
Torsioniess Diagonal
Members
(b) Plan of Equivalent Grillage
Figure 4.1 Grillage Idealization
For practical purposes, a slab bridge can be regarded as an isotropic plate

and can be conceptually divided into a number of rectangles. Each rect -
angle can then be idealized by the above _mentioned assembly of six
beams (Fig. 4.2a). However, the resulting grillage is a complex one due to
Transformation of Bridge Deck into Equivalent Grillage 77
the presence of diagonal members. This renders the analysis unsuitable, be -

cause of time consuming data input requirements and difficulties in th e
interpretation of results.
The idealization can be made manageable by assuming Poisson's ratio y
to be zero, thereby eliminating the need for the torsionless diagonal mem -
bers. Member prOperties of the resulting assembly of orthogonal beams
(Fig. 4.2b) are given by
3
t Lx 3
/
=
Ix = 24 Y
24
E t3 ) E (L, t3) (4.2)

(L24
), Y
G 24
Jx
WAMOM
MOO_Off
MMAM rAM
(a) Nan of Idealized
Grillage Accounting
for Poisson's Ratio
(b) Nan of Idealized

Grillage ignoring
Poisson's Ratio
Figure 4.2 Plans of Idealized Grillage
In an assembly of orthornal beams, tbR ninment in a beam denends only

upon the curvature of the assembly in the direction of beam. The corre -
sponding moment in a slab depends not only upon the curvature in the
direction of the moment but also on the curvature in the perpendicular

direction. This "slab-action" is represented by the following equations for the
moments in x and y directions, M and My respectively,
= D(d2 c
(4.3)
y ____2 ) M
y
D(d2 2) d
d x2 dy 2
dy d
E t3
where D
12 (1 y 2 )
Second term in the above equations accounts for the effect of
curvature in the direction perpendicular to that of the moment. It can be seen
that the effect of curvature in the perpendicular direction, which is directly
related to y, would not effect the idealization which neglects y.
4.2.2 General Guidelines for Grillage Lay-out

Because of the enormous variety of deck shapes and support conditions,
it is difficult to, adopt hard and fast rules for choosing a grillage lay-out of
the actual structure. However, some basic guidelines regarding the location,
direction, number, spacing etc. of the longitudinal and transverse grid lines
forming the idealized grillage mesh, are described here. But each type of
deck has its own special features and may need some particular arrange-
ments for setting idealized grid lines and therefore has been discussed
separately also.
a) Location and Direction of Grid Lines
Grid lines are to be adopted along 'Lines of Strength'. In the longitudinal
direction, these should be along the centre line of girders, longitudinal webs.
or edge beams, wherever these are present. Where isolated bearings are
adopted, the grid lines are also to be chosen along the lines joining the centres
of bearings. In the transverse direction, the grid lines are to be adopted, one at
each end connecting the centres of bearings and along the centre lines of
transverse beams, wherever these exist.
Ordinarily, the grid lines should coincide with the centre of gravity of the
sections but some shift is permissible, if this simplifies the grid lay-out or if
it assigns more clearly and easily the sectional properties of the grid
members in the other direction.
b) Number and Spacing of Grid Lines
Wherever possible, an odd number of longitudinal and transverse grid lines
are to be adopted. The minimum number of longitudinal grid lines may be
three and the minimum number of transverse grid lines per span may be live.
The ratio of spacing of transverse grid lines to those of longitudinal grid

lines may be chosen between 1.0 and 2.0. This ratio should also, ordinarily,
reflect the span-width ratio of the bridge. Thus, for a short span. and wide
bridge, it should be close to 1.0 and for long span and narrow bridge, this
ratio may be closer to 2.0.
Grid lines are usually uniformly placed, but their spacings can be varied, if
the situation so demands. For example, closer transverse grid lines should be
adopted near a continuous support as the longitudinal moment gradient is
steep at such locations.
It may be noted that with an increase in number of grid lines, the accuracy
of computation increases, but the effort involved is also more and soon it
becomes a case of diminishing return. In a contiguous girder bridge, more
than one longitudinal physical beam can be represented by one grid line. For
slab bridges, the grid lines need not be closer than two to thrrte times the
depth of slab.
4.2.3 Grillage Idealization of Slab Bridge

Following recommendations are made for setting out grid lines in slab
decks with line supports at .either end:
(i) The direction of longitudinal grid line is ordinarily parallel to the free
edge of the deck.
(ii) For bridges supported on discrete bearings, longitudinal grid lines are
placed along the centre of each bearing. One longitudinal grid line
along centre line of each edge beam, if they exist, is also provided.
(iii) Total number of longitudinal members generally be kept between five
and seven (preferably odd number) for two-lane slab decks without
footpath. One additional grid line is provided along the centres of each
footpath, if exist. In case of contiguous beam-slab construction
(pseudo-slabs), one longitudinal grid line may be provided for two or
more physical beams.
(iv) The minimum distance between longitudinal grid lines is limited to 2 to
3 times the slab depth and the maximum separation of .lorigitudinal
members should not be more than one-fourth of the effective span.
(v) In general, transverse grillage members should be at right angles to
longitudinal members. But for bridges with skew angle less than or
upto 15 or where the transverse directions of strength, such as
reinforcement or prestressing, are skew, the transverse grid lines are
oriented parallel to the supports.
(vi) The end transverse grid lines are placed along the centre line of bear-
ings on each side.
(vii) The spacing of transverse grid lines should be small. Their number
depends upon the span of the bridge. Five to seven transverse mem-
bers may be adopted for spans upto 10 m and seven to nine for
bridges with span above 10 m. In regions of sudden change such as
over intermediate supports, a closer spacing is necessary.
Ordinarily, one grid line along the centre of bearings at each end
and brie at the centre of span are provided initially and then other
grid lines are placed in between them.
(viii) As far as possible, the spacings of each of longitudinal as well as
transverse grid lines, are kept uniform_
(ix) It is important that the idealized grillage is supported at the same
positions as the actual deck.
Some of the grillage arrangements for right and skew solid slab, voided
slab and pseudo slab bridges are discussed with examples.
1. Solid Slab Bridge

a) Right bridge with and without footpaths
Figures 4.3 and 4.4 show the plans and sections of slab bridges without
footpath and with footpath respectively. The longitudinal and transverse
grid lines are also shown therein.
b) Skew bridge
Skewness has considerable effect on the behaviour of the deck and criti-
cal design stresses. Skew decks are associated with special characteristics
like hogging moment and large reactions near the obtuse corner and small
reactions and possible uplift at the acute corner. Moreover there is consid-
erable torsion of deck. Thus special attention is required while laying out
the grid lines for a skew bridge.
Decks with skew angles less than 15 can usually be handled as right
decks. Same guidelines as discussed above, for right bridges, are applicable
for such decks with or without footpaths and do not need any further elabo-
ration for setting out the grid lines. However, bridges having skew angles
more than 15 pose problems in regards to the positioning and orientation of
longitudinal and transverse grid lines and are discussed here.
A skew deck can be analysed with grillage having either a parallelogram
mesh as shown in Fig. 4.5(a) or orthogonal meshes as in Figs. 4.5(b) and 4
.5(c).
Transformation of Bridge Deck into Equivalent Grillage .81
-
r -
r-r-T-7 - - 1 XmL
I 1 I I I I I
I 1 1 I I I
I L --1 A_ i I 1
r T T ---
I I I I1 1 11
I _1_11 ! I
I 1 I II I
_1. _____ -- Grid-Lines
I I 1- I I
i I 1 I I 1
L - i-- - -I- --1- - 4-- -4- -
_I
I I ' 1 1 I
I- T - TI - i-I- 1 - - - 4 - - 1 I
I 11_1 I I
I 1
I Ii , ! I
I 1 1 I
,L__4_,_1-1_I_J__ -I X= 0
I I I I I I- I
______
I (a) 11Plan
i of Solid
11 1 1 1
Slab
1
1
LL __i __..,_ _ _1_ L _ _I
-.4 b
1D
I_________________________________________________________________________________1-
(b) Section on A-A
Figure 4.3 Solid Slab Grillage Lay-out without Footpath
While the parallelogram mesh (Fig. 4.5a) is convenient for low skew
angles, it is not appropriate for angles of skew greater than 15 because it has
-.no members close to the direction of dominating structural action. For
bridges with larger skew, say. greater than 15, a parallelogram mesh as Fig.
4.5(a) will result in an over-estimated maximum deflections and moments,
the amount increasing with angle of skew. The quantity of reinforcement in
such grids is likely to be excessive and uneconomical. An orthogonal grid
lay-out as shown in Figs. 4.5(b) or 4.5(c) will be more realistic.
In skew bridges, the direction of principal bending moment across the
Farancl to skcw span at edge to near normal to
.$
411
rI - - r- - r - - -T - - - 7 - I-- -1 X1.
1 1 I I tI I t 1
I 1 I
1I 11 tI I
I -- i --- 4. - - - 4 - - - -I- - - --I
I
I I 1
I
I 1
I I I
1 I
i
r- --II- -- -1-- --1- -- --11----F - - H A
I I .1
I I I 1 I . 1 I I I
1
x
1 1 1
1 I I I
r-- - i I l i t i- t 1 i
1 ii I 1 I
t 1 1 I
11 I. I ___I
t 1 1 i1 t
i 1 I t I I
I l 1
1
t1
r-- - -I1--.-r - -- -t- --- ir - -- -I- - 1
I I I I I I
I 1 Ii I I
1 t
i
t__ __ i__ __i_ ___L _ _ _ I _ _ _ _ L _ _ _I
(a) Plan of Solid Slab
Footpath
u ' ( 1 4 . 1 4 1
i s I , T
(b) Section on A. A
Figure 4.4 Solid Slab Grillage Lay-out with

Footpath
abutment in the centre. This is illustrated in Fig_ 4.6. While adopting the
orthogonal arrangement of grillage, two different cases arise, depending
upon the skew angle and span-width ratio as shown in Fig. 4.7.
Case (a) refers to the situation where skew region is small and right
Nyc m re-,,r,-.s.rt. this rasp when length AE < EG. The
recommended limiting condition is L sin 2, .S B/2 (Fig. 4.7a).
E. 1
Diaphragm Beam
PP-
INN AI
I
1IP"
XIII1111LS
I "
SiIN111111
(a) Skew or Paretlelogram (b) Mesh Orthogonal to
Mesh Span
II II III II II 6..
1IIIIIIIIIIII Edge Beam
IIII II I II 111 k
(c) Mesh Orthogonal to Support ,
Figure 4.5 Grillages for Skew Decks
Figure 4.6 Principal Moment Directions in Skew Slab Decks
Case (b) represents the situation where skew region is large and right
region is small i.e. AE > EG. The recommended limiting condition here may
be L sin A > B/2 (Fig. 4.7b).
The above two cases will decide the orientation of longitudinals with
respect to supports. In case (a), the main spanning will follow the direction
normal to support and the longitudinal grid lines will be taken normal to
support lines and transverse grid lines will be perpendicular to them as shown
in Fig. 4.8 (a).
As against this, in case (b). the main spanning will fnllnw ctepw rur,:ction
and hence longitudinal grid lines will be taken parallel to bridge axis and
Right Region
Skew span
Skew 7: Skew Angie

Region A G
B
(a) Small Skew Region (L

sin NzS. Bi2)
Right Region
Skew L=Skew span

Region
B
(b) La rge Skew Region
( L sin h > 6/2 )
Figure 4.7 Cases of Small and Large Skew Regions
transverse grid lines will be perpendicular to these as shown in Fig. 4.8(b).

The spacing and number of grid lines in both longitudinal and transverse
directions in case (a) and case (b) are discussed below:
Case (a)
The spacing of longitudinal grid lines may be different in skew region and
right region. The spacing of these longitudinal grid lines is to be neither
greater than the spacing of transverse grid lines nor greater than three times
the thickness of slab. In the right region, .one 11/4-)nezlit4difia1 caLli jakcn
along ED and GF and one or more between ED and GF. In skew region,
F C
X-L
-Longitudinds
_ Transversals
(a) Longitudinal Grid Lines

Norm al t o Support s
14 . / * 4 .
/ Transversals
/
/ / Longitudinals
A /
/ 1'4 / /
A
(b) Longitudinal Grid Lines
Parallel to Bridge Axis
Figure 4.8 Grid Lines in Skew Slab Decks
the longitudinal grid lines may originate from the nodes where the transverse
lines meet the longitudinal grid lines. However, if the number of longitudinals
in skew region become too many, alternate or even lesser longitudinal lines
may suffice. As far as possible, the longitudinal grid lines should be
equidistant in skew region and right region each.
Seven transverse grid lines for skew decks of span upto 10 m and nine
grid lines for spans above 10 m may be used. Out of these transverse grid
lines, one line each be located at ends of the span joining the centre line of
bearings on each abutment and one at the centre of span. The remaining
lines can be set in between these. As far as possible, the grid lines be kept
equidistant.
Case (b)
The lay-out of lbrigitudinal grid lines will be similar to as in the case of
right bridge discussed earlier. The actual numbers of transverse grid lines
will be decided so that the spacing of these lines is not greater than 1.5 to 2
times the spacing of longitudinal grid lines. Referring to Fig. 4.8(b), one
grid line each is set along GQ and DR and one at midway between GQ and
DR. Others may be set between these if necessary. In triangular portions
AGQ and DCR, transverse grid lines may be taken from the end of each
longitudinal. However, if the number of transversals in this region
becomes too large then alternate or even lesser transverse lines may also
suffice. Again, in this case also, the transverse grid lines be placed
equidistant, as far as possible.
In skew slab deck with foot path, one extra longitudinal grid line is
taken for the footpath at its centre at each end of the deck width as in the
case of right deck.
Sometimes a thick slab deck with thin cantilever and connecting slabs, as
shown in Mg./4.9 is also encountered in practice. The thin cantilever slab at
the edge, supports the kerb, parapet etc. and the middle slab is meant for
road divider or verge. The longitudinal grid lines can be placed as in Case
(a) or as in Case (b). Location of the longitudinal grid lines as shown, assign
the sectional properties of the transverse grid members better between 3 and
4 in Fig. 4.9(a) or between 4 and 6 in Fig. 4.9(b). It may be also noted that in
Fig. 4.9(b) longitudinal grid lines through 1, 5 and 9 will not have zero end
deflections.
Thin Cantilever
(b)
Figure 4.9 Thick Slab Deck with Thin Cantilever and Connecting Slab
AM .
2. Voided Slab Bridge

For a voided slab deck as shown in Fig. 4.10, the longitudinal and trans-
verse grid lines are set in a similar fashion as in solid slab except that one
longitudinal grid line is adopted covering one or more voids, depending upon
their closeness. The grid lines may preferably be taken at the centre of solid
portion between voids as shown in Fig. 4.10(b). The edge grid lines may be
taken suitably to represent the edge strips of slab. The number of longitudinal
lines may be odd or even depending upon the void position and the width of
the bridge.
Partipet
\ m
(a) Section of a Voided Slab
1 1

I I j I I I - 0

Figure 4.10 Longitudinal Grid Lines for Voided Slab
The number of transverse lines may be located as in solid slab, depending

upon span-width ratio and the spacing of longitudinal grid lines.
Sometimes, the voids do not run throughout the length of the span but are
provided only for about two-third middle segment of the span. In such cases,
the longitudinal grid lines run through the entire span as usual but with
different inertias for voided and solid end portions. In transverse direction,
grid lines are also located at the sections where voided section changes to
solid section.
3. Pseudo-Slib Bridge
As discussed in Chapter 1, contiguous beams spaced closely with in-situ
concrete comes under the category of pseudo-slabs. Usually standard pre-
cast beam sections are used in the construction. In such a situation, the
number of closely spaced beams are large. A grillage arrangement with
longitudinal grid lines coincident with every physical beam will lead to
many grid linpc which ,Pc....*.1 arid unmanageable. Therefore, it
is proper to represent more than one physical beam by a longitudinal grid line.
Sufficient transverse grid lines are to be provided for detailed analysis.

Their precise positions are chosen so that they intersect support beams at the
same points as longitudinal grid lines. For other details regarding the layout
of longitudinal and transverse grid lines, recommendations given in Section
4.2.3 can be followed.
Figure 4.11 shows the details of a bridge deck having inverted T-beams
and in-situ concrete slab forming a Pseudo-slab system. The span and width
8 @ 1= 8 M
V
(a) Cross Section O
I
3 I f
- . (b) Longitudi nal Gri d D
S
Lines'
.1F 5
I
. 1 I
.
I-- - I-- -- - , --4- 1 ----I
I 1
1 1 i
I I I I
I i I
I I I II
I I I I
I, 1 1
I I I
I
I 1
l- - +---- i-- - - + - --- -i Iii;
I I i 1 c.4
1 1 1 1e
1 1 1 1-
1 1 I
----- 1 -.- --1---- - 1 ----1
1 1 1 1
I I I 1
I I I I
11-- -I-- -- ---
1 I -1 i
I i I- 1 - --
1 I
1 1
I 1 I I
I I 1 I I
1--- -i- - --
i E. i
k--
r to n
--Jk-- - -1
Figure 4.11 Grillage Geometry of Inverted T-Beam with in-situ Concrete Slab-Pseudo Slab
Bridge
of the deck are taken as 15 m and 9 m respectively, giving span-width ratio

as 1.67. Nine precast beams at 1 m spacing are assumed.
Five longitudinal grid lines, along the centres of alternate webs of physi-
cal beams are chosen. The edge longitudinal grid line is taken along the
centre of edge physical beam as shown. Thus each grid line represents
more than one physical beam. Seven transverse grid lines are chosen at the
spacing of 2.5 m to represent the transverse medium.
If the number of physical beams are in even number, say 10 in the above
example, the longitudinal grid lines are taken at the centre of in-situ
concrete between the beams as shown in Fig. 4.12.
Voids
i
1-
Figure 4.12 Longitudinal Grid Lines at the Centre of in-situ Concrete

Slab-Pseudo Slab Bridge
Another form of pseudo-slab construction consists of standard precast

box-beams as shown in Fig. 4.13. In such cases, the longitudinal grid lines
are taken along the centre lines of voids. Again, one grid line represents
more than one box-beam. The edge grid line may be taken along the centre
line of first box-beam. The number of transverse beams are chosen
depending upon the span-width ratio and the mesh size. If diaphragms are
provided, the transverse lines may be located at each diaphragm first and if
necessary, more grid lines may be chosen in between these.
EiftiOrD14ffiritrIfffe
(a) Box-Beam Deck

I I -0 1
(b) Longitudinat Grid Line
Figure 4.13 Longitudinal Grid Lines in Pseudo -Slab (Box-Type)

4.2.4 Grillage Idealization of Slab-on-Girders Bridge
The idealization of beam and slab bridge by an assembly of intercon-
nected beams-seems to confirm more readily to engineering judgement than
for slab bridges. The T- and I-beams are by far the most commonly adopted
type of bridge decks consisting of longitudinal girders at definite spacing;
connected by top slab, with or without transverse cross-beams. Usually, the
diaphragms connecting the longitudinal girders, are provided at the sup-
ports.:
The logical choiCe of longitudinal grid lines for T-beam or I-beam decks
is to make them coincident with the centre lines of physical girders and
these longitudinal members are given the properties of the girders plus
associated portions of the slab, which they represent. Additional grid lines
between physical girders may also be set in order to improve the accuracy
of the result. Edge grid lines may be provided at the edges of the deck or at
suitable distance from the edge. For bridge with footpaths, one extra
longitudinal grid line along the centre-line of each footpath slab is also
provided. The above procedure for choosing longitudinal grid lines is appli-
cable to both right and skew decks.
When intermediate cross-girders exist in the actual deck, the transverse
grid lines represent the properties of cross girders and associated deck slabs.
The grid lines are set-in along the centre-Iines of cross-girders. Grid lines
are also placed in between these transverse physical cross-girders, if after
considering the effective flange widths of these girders, portions of the slab
are left out. If after inserting grid lines due to theSe left-over slabs., the
spacing of transverse grid lines is still greater than two times the spacing of
longitudinal grid lines, the left-over slabs are to be replaced by not one but.
two or more grid lines so that the above recommendation for spacing is
satisfied.
When there is a diaphragm over the support in the actual deck, the
grid lines coinciding with these diaphragms should also be placed. A
typical T-beam bridge with grillage lay-out is shown in Fig. 4.14.
When no intermediate diaphragms are provided, the transverse medium
i.e. deck slab is conceptually broken into a number of transverse strips and
each strip is replaced by a grid line. The spacing of transverse grid lines is
somewhat arbitrary but about 1/8 of effective span is generally convenient.
As a guideline, it is recommended that the ratio of spacing of transverse and
longitudinal grid lines be kept between 1 and 2 and the total number of lines
be odd. This spacing ratio may also reflect the span-width ratio of the deck.
Therefore, for square and wider decks: the ratio can he kept as I and for long
and narrow decks it can approach to 2.
(c) Cross-Section
(a) Plan
(b) Longitudinal Section
iiiiiiiii
1101111111
11
( d) G ril l a ge Lay - out
Figure 4.14 T-Beam Bridge and Grillage Lay-out
The transverse grid lines are also placed at abutments joining the centres of
bearings. A minimum of seven transverse grid lines are recommended,
including end grid lines. It is advisable to align the transverse grid lines
normal to the longitudinal lines wherever cross-girders do not exit. It should
also be noted that the transverse grid lines are extended upto the extreme
longitudinal grid lines.
In skew bridges with small skew angle say less than 15 and with no
intermediate diaphragms, the transverse grid lines are kept parallel to the
support lines as shown in Fig. 4.15(a). Additional transverse grid lines are
provided in between these support lines in such a way that their spacing does
not exceed twice the spacing of longitudinal lines, as in the case of right
bridges, discussed above.
IS
In skew bridges with higher skew angle, the transverse grid lines are
set along abutments PQ and WV and also along PR, ST and UV initially
as shown in Fig. 4.15(b). Then extra grid lines are inserted in between PR
and UV. When the span-width ratio and the skew angle are such that the
skew region QR is larger than the spacing of transverse grid lines in right
region VR, additional transverse members parallel to interior transverse
grid lines starting from the interior nodes are to be placed.
FQ
(
Z.
y,y ,R
Lb ;is./
>1 7T ,
(a) Deck with Small Skew (b) Deck with Large Skew
A>15
Figure 4.15 Grillage Arrangement in Skew T-Beam Bridge without Cross Girders
Sometimes, precast I-sections with in-situ reinforced concrete slab

form the deck. In such cases, the grillage is laid out in the same way as
the T-beam monolithic deck. The ratio of moduli of elasticity of slab and
precast beam materials are to be properly accounted for in the analysis,
and will be discussed later in the chapter. This is referred to as a two-
stage construction.
4.2.5 Grillage Idealization of Box-Girder Bridge

Idealization of box-girder bridge is in many ways similar to that of slab-
on-girders construction but there is a behavioural difference betweeen them.
The analysis of box-girder bridge is associated with special problems of
shear deformations (shear lag) due to usually wide flanges of the deck and
distortions of the cells, if sufficient number of intermediate transverse
diaphragms are not provided. Although, other rigorous analytical techniques
are available for the analysis of box-girder bridges, the grillage analogy has
also proved to be sufficiently accurate and can be recommended in many
cases. The grillage analogy in this case has the added advantage of being
relatively inexpensive in computer time and simple to comprehend.
The method is to be adopted where the effects of shear deformations and

cell-distortions are negligible and could be ignored. The method is most
appropriate for multi-cell rectangular box-girder decks (Fig. 4.16). How-
ever, it can also be used for decks with one or two cells only. The outer
webs may be vertical or inclined, as shown in Fig. 4.17.
I-
(a) Deck Section
Figure 4.16 Grid Lines for Multi-Cell Box-Girder Deck

(a) Deck Section
.0
- -- ___________ elipt

Figure 4.17 Grid Lines for Box-Girder Decks with Inclined Webs
Longitudinal grid lines are usually placed coincident with webs of the
actual structure (Fig. 4.16). If the deck has sloping end webs (Fig. 4.17), the
grillage simulation is not so precise and engineering judgement must be used
to position longitudinal members. However, one grid line may be taken at the
junction of the inclined web with slab as shown. Normally, additional
longitudinal grid lines are located along the edges of the side cantilevers with
nominal stiffnesses for the convenience of analysis. Additional longitudinal
grid lines are to be adopted for bridges with footpaths at their centres as done
in the case of slab and T-beam bridges, discussed earlier.
The transverse medium consisting of top and bottom slabs only (with no
diaphragms), is represented by equally spaced transverse grid lines along
the span. The spacing and number of grid lines are similar to as adopted for
slab-on-girders bridge. If the deck is having diaphragms, the transverse grid
lines are placed along each diaphragm including at supports. Additional grid
lines representing the top and bottom slabs are placed in between the dia-
phragms, if needed, to meet the minimum:requirements of transverse grid
lines. A closer spacing of transverse grid lines will result in more continuous
structural behaviour and will provide greater details of forces etc.
For skew box-girder bridges, the procedure to.be followed regarding the
setting of longitudinal and transverse grid lines will be the same as in the case
of slab-on-girders bridge, discussed earlier.
Spaced box-girders, also referred to as Spine box-girders, connected by
top slab only, form a distinct class of decks requiring a special approach.
Figure 4.18 represents a four-cell spaced box-girder bridge.
In deciding the grillage layout for this type of deck, the main problem lies
in defining the stiffness in the transverse direction. In the transverse direction,
the deck is alternately very stiff over the box-beams and very flexible between
the box-beams. It is not possible to replace the transverse media by grillage
beams of uniform stiffness, as is usually possible with other types of decks. In
this type of deck, each physical beam is replaced by grid lines positioned at
the webs as shown in Fig. 4.18(b). This takes into account the effect of abrupt
change of transverse section more correctly. The other details of layout of
transverse grid lines are similar to those of slab-on-girders bridge. The plan of
the arrangement of grid lines is shown in Fig. 4.18(c).
Sometimes, the width of the cell is small in comparison with their spac-
ing and walls are relatively thick preventing distortion of the cell (Fig. 4.19).
For such sections of bridges, the longitudinal grid lines are better placed
coincident with centre lines of boxes with additional nominal grid lines
running in between the boxes as shown in Fig. 4.19 (b). In the transverse
direction, the grid lines will have uniform stiffness corresponding to the
deck slab and other details of layout will be similar to those of slab-on-
girders bridge, discussed earlier.
4.3 EVALUATION OF EQUIVALENT ELASTIC PROPERTIES

After the actual bridge structure is simulated into equivalent grillage, con-
sisting of longitudinal and transverse grid lines meeting at discrete nodes, the
second important step in grillage analogy method is to assign appropriate
elastic properties i.e. flexural and torsional stiffnesses to each member of the
ariliaap cn idpaliceri Thic merle ihP rnrrInlitt;nr of
,
flexural
moment of inertia I and torsional inertia J for the members of the grillage
FF
(a) Section of Spaced Box-Girder Deck
I 1- -v-- i-T ---Y-T---Y- T

L _L _ _ 1_ I _ _1_4. _ __I I
T
-
1 --4----- t i - -- t - i - -- 4--
i
I I 1 I I 1 ,I 1
_ 4 . . . .
t-
I 1 T1
1_1_ I , I I 1 1
I-
1
1
T -7-
--1
r 1--7-
I -- - - I j I I I
II t - t--- - - F 4
--
I I I 1,
4
1 I I
1-- 1- - --- I - I - -- i
-1-
- - I- -1 A--k--

A k A i 7 1c --- i--- 1--
(c) Plan of Grillage Arrangement
Figure 4.18 Grillage Arrangement in Spaced Box-Girder Deck
mesh. This is accomplished by considering isolated sections of the deck as

if they are individual beams and the inertias are calculated for each section
and allotted to the corresponding grillage beams representing that section.
The principles involved and the methodology adopted for evaluating the
various flexural and torsional inertias, are discussed first. Later on specific
bridge decks are considered.
bi
I
i i
er,
(bi- b2) -
\,.../
1-..-b4,-.-1 1--b.-,-
(a) Cross- Section of Deck with Narrow .-1Box-Beams
1 2

to 3
( b4/2

(d) Nominal Stab
Grid Line
(a) Beam Grid Line

Figure 4.19 Grillage Arrangement in Decks with Narrow Box -Beams
4.3.1 Flexural Moment of Inertia, I

The computation of flexural moment of inertia I of different individual
geometrical shapes like slab, T or I beams, box-girders etc. is straight for-
ward and needs no elaboration. However, in beams having Tee, Ell or box-
sections where slab is cast monolithically with the web of the beam, effective
flange-width of the associated slab is to be considered. The Indian Roads
Congress (IRC) recommendations [16] for choosing suitable effective flange
width of beams are being followed in India for road bridges and will be
further discussed elsewhere in this section.
4.3.2 Torsional Inertia, J

The torsional inertia J, often referred to as the Saint-Venant torsion con-
stant also, is generally not a simple geometrical property of the cross-section
as the case with flexuial moment of inertia I and needs careful consider-
ation. There is no accurate analytical procedure for the derivation of J.
However, the approximate method for the evaluatidn of J for different
cross-sections is based on the elastic theory of torsion of prismatic beams
[11, 12] and is discussed here.
Saint-Venant [12] derived an approxiMate expression for computing
the torsional inertia J, of open sections which is applicable to all cross-
sectional shapes without having reentrant corners. The expression
A4
J= (4.4)
40 I
P
where A is the area of cross-section and I is the polar moment of inertia. .

For a rectangle of sides b and d, above expression reduces to,
3 b3 d3
= (4.5)
10 (b2 + d2 )
In the case of a thin rectangle where b >
5d, the J value is more accurately given by
bd3
J= (4.6)
3
If the cross-section has reentrant
corners, J is very much less than that given by equation 4.4 above [13]. In
such cases, the value of J is obtained by notionally sub-dividing the section
into rectangular shapes without having reentrant corners and summing the
values of J of these elements. The value of J of a sub-divided portion with
notional cuts on two opposite faces' is to be computed as if sub-division is a
part of wide thin strip for which J = bd3I 3. Figure 4.20 shows a T-section
with reentrant corners and its sub-division. Thus, if J values of the portions
1, 2, 3 and 4 are designated as J , J , J3 and J4 respectively, then,
1 2
(1 3)
J 2 =- 1 (b 2
Jr = - " di
' i3 ) 3
3 b33 3 b: d34
J
4 2 2
J 3 10 (1 4 10 (b 4 d4)
and for the beam section as a whole,

4-1 Effective Flange Wiciffi---

-MAE=
11111111a
4113
i
Yr
d41:
1I d i
1. ______
b1
ro2
Figure 4.20 Sub-Division of T-Section with Reentrant Corners -rJ2+J3+J4
(4.7)
It may be noted that the value of J of the portion of deck slab forming the
flange is' to be halVed to take into account its continuity in the other direction
[12]. Widths b3 and b4 of segments 3 and 4 are so adjusted that areas b3xd3
and b4xd4 are same as original areas of the respective segments.
While notionally sub-dividing the section, it is worth remembering
Prandtl's membrane analogy. It is shown that the torsional stiffness of a
cross-sectional shape is proportional to the volume under an inflated bubble
stretched across a hole of the same shape. Thus, care is to be taken so as to
obtain the largest numerical value of J of the section as a whole. This is
achieved by choosing the elements so that they maximize the volume under
their bubbles [11, 12]. Sub-dividing the cross-section arbitrarily into rect-
angles and not trying to maximize the. volume under the bubble will result in
lower value of J.
As an alternative to Equations 4.5 and 4.6 above, expression given by
Timoshenko and Goodier [12] can be used for rectangular sections viz. J =
yrbd3 where ty depends on the ratio of sides b and d (b > d) of the rectangle.
The values of tir for different ratios of b and d are given in Table 4.1_
TABLE 4.1
bld 1 1.2 1.5 2 2.5 3 4 5 10 00
.0 .141 0 .166 0 .19 6 0 .229 0 .249 0 .26 3 0 .281 0 .291 0 .312 0333
The coefficient yr for any intermediate value of bid, may be obtained by

linear interpolation. If the depth or width of an element changes along its

length, then the inertias both I and J of the element are to be based upon
the mean dimensions.
It may also be mentioned here that the load distribution is likely to be
more sensitive to the value of flexiaral moment of inertia I rather than to
that of torsional inertia J, and as such a small error in the computation of J
is not likely to much affect the final results. An incorrect sub-division of a
section will invariably lead to an underestimation of J as discussed earlier
and this will only lead to a conservative design to a small extent.
The computations of I and J for different types of decks are now illus-
trated in the following sections.
4.3.3 Flexural and Torsional Inertias of Grillage Members: Slab Deck

For solid slab bridges, the computation of flexural inertia is straight for-
ward. The moment of inertia is calculated about the neutral axis of the
deck. Thus for an isotropic solid slab of depth d, the flexural inertia (i' per
unit width is given by
i 3d =
12 t
If the deck has thin cantilever or intermediate slab strips as in Fig. 4.9, the
longitudinal grid lines can be placed as in Fig. 4.9(a) or as in Fig. 4.9(b). In
Fig. 4.9(a) the flexural inertias of all members are calculated about the deck
neutral axis. However if the grid lines are placed as in Fig. 4.9(b), the thin
slabs above members 1,5 and 9 act primarily as flanges to members 2, 4, 6
and 8 respectively. Consequently the inertias of members 1, 5 and 9 are
calculated about centroid of the thin slab, while for members 2, 4, 6 and 8
inertias are calculated with the flanges as in Fig. 4.9(a) but with small
inertias of members 1, 5 and 9 deducted. In the transverse direction, the thin
slab would bend about its own centroid only so that the thin slab depth is
used for members 1-2, 4-5, 5-6 and 8-9 while the thick slab depth is used for
members 2-3, 34, 6-7 and 7-8.
For a voided slab deck such as in Fig. 4.21, the inertias of longitudinal
grid lines are calculated for shaded section about neutral axis. Transversely,
fa
Neutral
xis
Figure 4.21 Voided Slab Section

the inertia is generally calculated about the centre line of void. However
for void depths less than sixty per cent of the overall depth, the transverse
inertia can usually be assumed to be equal to the longitudinal inertia per
unit width.
Reinforced and prestressed concrete slab bridges usually have similar
stiffnesses in longitudinal and transverse directions with the result that suf-
ficient accuracy is obtained by assuming that the full uncracked concrete
section is effective, with reinforcing steel ignored. If the percentage of
transverse reinforcement is relatively small while longitudinally the bridge
is prestressed or heavily reinforced, one should take into account flexural
cracking, In such cases, the flexural inertias in two directions are calculated
separately for different transformed sections.
When the torsional inertia j per unit width of the slab is computed, two
points are tote noted, namely (a) the deck contributes to torsional stiffness
in both the directions and hence the value of f in each direction is to be
taken as one-half of the computed value and (b) the lateral dimension is
much laiger than the depth. Hence, for the slab,
j = 22: 13 d3 I = d3 per unit width (4.8)
Comparing the values of i and j per unit width for solid slab, it will be
seen that
j = 2i (4.9)
There is no simple and rigorous rule for calculating j for voided slabs
and the above relationship between i and j can be used with advantage to
compute j for voided slabs with sufficient accuracy.
Further, the torques in two orthogonal directions of an orthotropio slab are
of equal magnitudes. Consequently, the transverse and. longitudinal grillage
members should have identical torsional inertia per unit width of the deck. It
is suggested that, for such slabs, the torsional inertia of the transverse and
longitudinal grillage members, per unit width, be computed as,
j = 2 ix i (4.10)
where i and i are the longitudinal and transverse flexural inertias per unit
width of slab respectively.
Alcn if hac hppn nhcerveri that the. simulation of grillage and physical deck is
.
improved if the width of the edge member is reduced to (b 0.3 D) for

calculations of J where D is the thickness of slab and b is the width
of. edge member [6]. However, for simplicity, full width of edge member may be used.
In pseudo-slab construction, usually precast prestressed concrete beams are closely
placed with in-situ reinforced concrete forming the deck. In such cases, the in-situ
concrete slab has lower strength and stiffness than. the prestressed concrete beams so that
it has a modular ratio m = 0.8, compared
to prestressed concrete. Different transformed
cracked section inertias are used in the
two directions because the transverse
reinforcement is usually lighter.
It is to be noted that in a skew gril-
lage, the slab width which is to be
taken for computing the inertias, is the
one normal to the grid line. Referring
to Fig. 4.22, the widths of the slab to be
considered with the grid lines A and B
B of the skew system will be U cos a. and
Figure 4.22 Slab widths for V cos A. respectively. )
inertia in a skew grillage layout
43.4 Flexural
and Torsional Inertias of
Grillage Members: Slab-on-Girders Deck
Slab-on-girders bridge decks consist of a number of beams spanning
longitudinally between abutments with a thin slab spanning transversely
across the top. T-beam bridges are the common examples 'tinder this cat-
egory. The be___ ms may be cast monolithically with the slab or the precast
beams with in-situ slab may be used. The decks may be with or without
intermediate and/or end diaphragms.
The thin slabs can be thought of as flanges of I or T-beams. When such I or
T-beams bend, the flanges are subjected to flexural stresses. An element of
the flange away from the rib or stem of the beam has less stress than the one
directly over the rib due to shearing deformations of the flange. Shear
deformation relieves some amount of compressive stress in more distant
elements. This phenomenon is known as shear lag. The variation of corn:
pressive stress across the width of flange is accounted for by considering the
effective width of flange. The effective width may be smaller than the actual
width and is considered to be uniformly stressed. The effective width of flange
is determined based on the concept of constant compressive stress over the
entire effective width such that the total compressive force carried by the
effective flange is the same as that by the actual flange width with variable
compressive stress (Fig. 4.23). Tnis effective width has been found to depend,
besides other things, on the span and the relative thickness of slab.
HEffective widthH
Variation of-' -1_..-- Assumed Uniform
' Ftexural. ' >--'r----'-'------- Stress over the
----
S t r e s s
Effective width
Figure 4.23 Variation of Flexural Stress in the Flange of T-Beam
For the purpose of calculation of flexural and torsional inertias, the

effective width of slab, to function as the compression flange of T-beam or
L-beam, is needed. A rigorous analysis for its determination is extremely
complex and in absence of more accurate procedure for its evaluation, IRC
recommendations are followed. IRC : 21-1987 [16] recommends that the
effective width of the slab should be the least of the following:
1. In case of T-beams
(i) One-fourth the effective span of the beam
(ii) The distance between the centres of the ribs of the beams
(iii) The breadth of the rib plus twelve times the thickness of the slab
2. In case of L-beams
(i). One-tenth the effective span of the beam
(ii) The breadth of the rib plus one-half the clear distance between the
ribs
(iii) The breadth of the rib plus six times the thickness of slab
The flexural inertia of each grillage member is calculated about its cen-
troid. Often the centroids of interior and edge member sections are located at
different levels. The effect of this is ignored as the error involved is
insignificant.
Once the effective width of slab acting with the beam is decided, the
deck is conceptually divided into number of T or L-beams as the case may
be. Some portion of the slab may be left over between the flanges of
adjacent beams in either directions. In the longitudinal direction, it is suf-
ficient to consider the effective flange width of T, L or composite sections,
in order to account for the effects of shear l ag and irrn^r- th 1-ft
over slab. However, in the transverse direction, the left over slab should
be considered by introducing additional grid lines at the centre of each
left over slab portion.
Sometimes, for the purpose of improving the simulation, it is desirable
to place additional grillage members of nominal stiffnesses between grid
lines representing the beams. The sectional properties of these additional
members are calculated in a similar manner as outlined in Section 4.3.3
for the deck of Fig. 4.9.
The sectional properties of grid lines representing the slab only, are
calculated in the usual way i.e. I = bd3/12 and J = bd316.
If the construction materials have different properties in the longitudinal
and transverse directions, care must be taken to apply correction for this.
For example, in a reinforced concrete slab on precast prestressed concrete
beams or on steel beams (Fig. 4.24), the inertia of the beam element (1 or
J) is multiplied by the ratio of modulus of elasticities of beam Eb and slab
E: materials to convert it into the inertia of slab material. For example the
total torsional' inertia in terms of the slab material will be given by
Effective Flange Width
1 --*-1 J = Jl + (J2 + J3 + J4) Eb
_________ b E s
-d1
d2 J1, J2, J3 and J4 are calculated as below
=12-(13-121 4)
3b 3 d 3
j 22
2
10 (i4 -
13
- Figure 4.24 Torsional Inertia in
Precast Beam with Cast-in-Situ Slab
J3=3b3d3
3b3 d3
J 44
4
10 (hi + 41)
4.3.5 Flexural and Torsional Inertias of Grillage Members : Box-

Girder and Cellular Deck
In the box-girder deck having longitudinal section as shown in Fig: 4.25,
the top and bottom slab flex in unison about their common centre of gravity.
Following the notations given in Fig. 4.25, the moment of inertia of transverse grillage
member per unit width, it, is calculated about the common centroid and is given by
= 1 + t2 d2(4.11) since ti d1 = t2 d2 and (di +
d2) = d, we get
2
i = d t t 2 per unit width

t1+ t2 (4.12)
If the transverse grillage members also
include a diaphragm, as at 'A', (Fig.
4.25) the inertias should be calculated including the diaphragm.
Figure 4.25 Longitudinal Section of Box-Girder Deck
The torsional inertia J of a thick-walled box-section is obtained simply by

deducting the value of J of.the inner boundary from that of the whole section.
Thus in Fig..4.26, where bi dl < 0.6 bd,
3 b3d3 b.?d?
J= (4.13)
10 (b2 + d2) (0 +4)
Where continuity in the transverse direction is present, the torsional inertia in

each direction is to be taken as half of above value.
In thin-walled cellular decks, unlike the flexural inertias, the torsional
inertia of grillage members can not be calculated by isolating certain portion
of the bridge and then assigning their properties to grillage beams. The total
torsional inertia of a section is distributed amongst the various grillage
members in proportion to the width that each member represents. For a
Figure 4.26 Thick-Walled Box-Section

longitudinal grillage member representing a thin-walled cellular section,
the torsional inertia is given by
J =4 A2 (4.14)
f ds
t
where 'A' is the area bounded by the centre line of the closed cross-section
and 's' and 't' refer to the length and thickness of different units respec-
tively.
Referring to the Fig. 4.27, the equation 4.14 can be written as
4A2 4A2
J= (4.15)
(E (Si S2 2S 3
t t t
l 2 3
______________________________________ 4
-7
F - - - 2
Figure 4.27 Thin-Walled Box-Section
The above expression is applicable to closed cross-sections with one or two

symmetric cells. For multi-cellular sections, as shown in Fig. 4.28, the value
of J is obtained on the assumption that-the contribution of the interior webs is
negligible. The J value of the section is given by .
I-
(a) Multi -Celt Cellular Deck
t2
(a) Equivalent One -Cell Cellular Deck
Figure 4.28 Computation of J in Multi-cell Cellular Deck
J 4 b2d2 (4.16)
(b b 2d)
t2 t3
The term 2d/t3 is small compared to other two terms in the denominator of
equation 4.16 and can be ignored. The equation 4.16 can then be re-written as
J 4 bd2ti (4.17)
t 2)
If j denotes torsional inertia per unit width in each direction, then,
L
(b)
= 2d2 r,t2 per unit width of cell (ti + (4.18)

12)
Comparing equations 4.12 and 4.18, it is seen that] =. 2i. This is the same
relationship as was obtained for solid slab bridges earlier and can also be conveniently
used for computing the values of J of the transverse members of multi-cell boxes of
closed section without diaphragms.
For obtaining J values of longitudinal members, cell-widths in cross-sections are
considered and for J values of transverse members, spacings of transverse members are
to be taken.
If j is the torsional inertia per unit_width, referring to Fig. 4.29, the values of J for
longitudinal members 1, 2 and 3, will be given by
(a) Cross-Section of a MuLtf-Cetl. Box-Girder
T
a
(b) Plan
Figure 4.29 Evaluation of J for Longitudinal and Transverse Members of Multi-
Cell Box-Girder Bridge
= 1/2j h and J, = j h
and for transverse members 4 and 5. .1 vaiiips will
J4= = ja
= O M
4 1 6 1 1 . 1 1 1 1 . -
For transverse. grid members of multi-cell box-sections incorporating

diaphragms, the value of J is obtained as the larger of the values
computed from the following alternatives :
(i) treating the transverse members (with the diaphragms) as T, L or
I-sections and then applying the open section formula given by
equation 4.7 and
(ii) considering the' top and bottom slabs only and applying the close
section formula as in equation 4.12.
For precast spaced box-beam bridges (spine bridges) having narrow
cells and relatively thick walls preventing cell-distortions (Fig. 4.19), the sec-
. tional properties of nominal, additional grillage. member 2 is calculated
for width of the slab 'as shown. The properties of the grillage members 1
and 3, are then evaluated' for the section with flanges including the area in
nominal members but with previously calculated properties of the nominal
members deducted from these values.
If the boxes are much wider in comparison to the beam spacing as in Fig.
4.30, the values of I and .1 of the longitudinals can each be taken as one-half
1
(a) Cross- Section of Spaced Box-Beam
'1 2 3 4
1
(b) Arrangement of Grid Lines
(c) Flanged Box-Section for Computing Land

Figure 4.30 Spaced Box-Bridge Having Wide Beams
the value of the flanged box section. For the transverse grillage beam be-
tween members 1-2 (Fig. 4.30b), the value of I will be based on top and
bottom slabs computed about their combined centroid and the value of I for
the beam between 2-3 will be based on top slab about its own centre of
gravity. The values of J for beams between 1-2 and 2-3 can be determined
from the usual relation i.e. j = 2i.
Thus, it is seen that the actual deck of the bridge can be idealised into
suitable grillage with appropriate equivalent inertias allotted 'to its each
member. Other steps in grillage analogy i.e. the application of loads and its
transfer to various nodes of grillage, analysis, stress response,
interpretation of results etc. are dealt with in the next chapter.
REFERENCES
1. BAKHT, B. and JAEGER, LG., "Simplified Analysis for Slab-on-Girder
Bridges", Bridge and Structural Engineer 12(4), 1982, INDIA.
2. BAKHT, B., JAEGER, L.G. and CHEUNG, M.S., "Cellular and Voided Slab
Bridges", Journal of Structural Division, ASCE (ST9), 1981.
3. BAxnr, B. and JAEGER, L.G., "Bridge Analysis Simplified", McGraw Hill,
New York, 1985.
4. BAKHT, B.s JAEGER, L.G., CHEUNG, M.S. and MuFrr, A.A., 'Me State-of-the
Art in Analysis of Cellular and Voided Slab Bridges", Canadian Journal of
Civil Engineering 8(3), 1981.
5. CUSENS, A.R. and PAMA, R.P., "Bridge Deck Analysis", Wiley, London, 1975.
6. HAMBLY, E.C., "Bridge Deck Behaviour", Chapman and HaIl Ltd., London, 1976.
7. JAEGER, L.G. and BAKHT, B., "Bridge Analysis by Micro-computer", McGraw-
Hill, New York, 1990.
8. JAEGER, L.G. and BAxtrr, B., 'The Grillage Analogy in Bridge Analysis",
Canadian Journal of Civil Engineering 9(2), 1982.
9. LtarrFocrr, E. and SAWKO, F., "Structural Frame Analysis by Electronic Com-
puter: Grid Frameworks Resolved by Generalised Slope Deflection",
Engineering, 187, 1959.
Id. MAISEL, B.I., "Review of Literature Related to the Analysis and Design of Thin
Walled Beams", -Technical Report 42.440, Cement and Concrete Association,
London, July 1970.
1 I. OMEN, J.T., "Mechanics of Elastic Structures", McGraw Hill, New York, 1967.
12. TIMOSHENKO, S. and GOODIER, J.N., "Theory of Elasticity", McGraw-Hill, New
York, 1951.
13. WEST, R., "Recommendations and the Use of Grillage Analysis for Slab and
Pseudo-Slab Bridge Decks", Cement and Concrete Association, CIRIA,
London; 1973.
14. WEST, R., "The Use of Grillage Analogy for the Analysis of Slab and Pseudo-
Slab Bridge Decks", Research Report 21, Cement and Concrete Association,
London, 1973.
110 Grillage Analogy in Bridge De& Analysis
15. YET-IRANI, Al. and HusATN, M.H., "A Gridwork Frame Analogy for Plates in
Flexure", Journal of Engineering Mechanics Division, ASCE, June 1965.
16. IRC: 21-1987, "Standard Specifications & Code of Practice for Road Bridges,
Cement Concrete (Plain and Reinforced), Section III", The Indian Roads Congress,
1988.
-
Chapter 5
Application
Application of Loads,
Analysis, Force Responses
and Their Interpretations
5.1' INTRODUCTION

The bridge deck has been transformed into an equivalent grillage consisting
of longitudinal and transverse grid members such that the idealized grillage is
very close to the physical deck. Each member of the grillage is allotted
flexural and torsional inertias which are equivalent to the corresponding
physical properties of the bridge deck. The longitudinal and transverse grid
lines form a mesh having number of nodes. The bridge is mainly subjected to
vertical loads comprising dead, live and impact loads. Grillage analysis
requires that these applied loads be transformed into equivalent loads at
nodes. This is done by finding equivalent nodal loads in terms of either
vertical load only or alternately vertical load, bending moment and torsional
moment.
This Chapter discusses different types of loads, identification of panels in
which the wheel loads of a vehicular loading system fall and transfer of loads
to nodes of grillage in the form of equivalent nodal loads. The analysis of
grillage is then carried out and response envelopes and the interpretation of
results are discussed. Both right and skew decks have been dealt with. Local
effects which are to be accounted for in arriving at the design responses, are
also discussed.
5.2 EVALUATION AND APPLICATION OF LOADS

The loads, consisting of dead, live and impact loads acting on the bridge
superstructure, are to be evaluated and nppropriately distributed to the nodes
of the grillage. Evaluation of each type of the above loads and their place-
ments on the deck are discussed below.
a) Dead Load
The deck of a bridge is subjected to dead loads comprising of its self
weight and weights due to wearing coat, parapet, kerb etc. which are of
permanent stationary nature. The dead loads act on the deck in the form of
distributed load. These dead loads are customarily considered to be borne by
the longitudinal grid members only giving rise to distributed loads on them.
This distributed load on a longitudinal grid member is idealised into equiva-
lent nodal loads. This is specially required to be done *hen the distributed
load is non-uniform. On the other hand, if the self load is uniform all along
the length of the longitudinal grid line then it is not necessary to find the -
equivalent nodal load and instead it can be handled as a:uniformly distrib-
iited load (u.d.l.) itself. Further, if the dead load is u.d.l. but its centre is non
concident with the longitudinal grid line Then it is substituted by a
vertical u.d.l. together with a torsional u.d.l.
Figure 5.1 shows a voided slab with voids running only over part length.
Longitudinal grid lines will have non-uniform loading, higher intensity of
loading in support region and lower intensity of loading in central region.
In such cases equivalent vertical nodal load is computed from the load on
the tributary area. Figure 5.2 shows a solid slab with non-uniform 'spacing
I I ' I I
I I I
I I
1I
I 1
Ii
Longitudinal Grid Line
Transverse Grid Line

II I
2c
-L- Tributary Area
2c
Figure 5.1 Voided Slab Bridge

Application of Loads, Analysis, Force Responses and Interpretations 113
A
Tributary Area
b'p+p 2a pqr,p, 2a pi
Grid Line A: Verticat and Torsions[ u.d.t. Grid

Line B: Verticat u.d.[.
Figure 5.2 Longitudinal Grid Lines with u.d.l. (b* a)
of longitudinal grid lines. The equivalent vertical load along each grid line is
computed in the form of vertical u.d.l. based on its tributary area. For grid .
line 'A' the loading is non-central and hence the equivalent load will consists
of a vertical -and torsional u.d.l., whereas for grid line `B' the loading is
central and hence the equivalent load will be a vertical load only.
The self weight of cross-beams and diaphragms needs further consider-
ations. These beams, located at specific intervals, are actually small discrete
loads on the longitudinal girders. However, for simplicity of computation,
the total weight of all the cross-beams per span should be calculated and
equally divided in the form of distributed loads to various longitudinal
members of the grillage. The dead weight of railings, kerbs, footpaths etc.
is lumped on the edge longitudinal grid lines.
b) Live Load
The main live loading on highway bridges is of the vehicles moving on
it. Indian Roads Congress (IRC) recommends different types of standard
hypothetical 'vehicular loading systems, for which a bridge is to be designed.

The details of these loadings have been discussed in Section 1.7 of Chapter 1
and are also available in IRC Code of Practice [10].
The vehicular' live loads consist of a set of wheel loads. These are dis-
tributed over small areas of contacts of wheels and form patch loads. These
patch loads are treated as concentrated loads acting at the centres of contact
areas. This is a. conservative assumption and is made to facilitate the analysis.
The effect of this assumption on the result is very small and does not make
any appreciable change in the design.
IRC Class A two lane, Class AA Tracked and Wheeled, Class 70R Tracked
and Wheeled loads are shown in Figs. 1.11 to 1.14. Three different wheel
arrangements for Class 70R Wheeled loads are in existence and are shown in
Fig. 1.14. Class 70R Tracked load may be idealised into 20 point loads of 3.5
tonnes each, 10 point loads on each track. The total load of the
vehicle, this case is 70 tonnes.
. One Class A or Class B loading can be adopted for every lane of the
carriageway of the bridge. Thus, for a two-lane bridge, we can have two lanes
of Class'A or Class B loading. However, for all other vehicles, only one
vehicle loading per two lanes of the carriageway is assumed. Thus, for a two-
lane bridge, the alternate live loading will be one lane of Class 70R Train or
Bogie or Track loading. For a three-lane bridge, three lanes of Class A
loading or alternately one lane of Class 70R loading alongwith one lane of
Class A loading is usually assumed.
It is assumed in the design that the vehicles can not go closer to the kerb by
certain . recommended clear distances. These distances are different for
different types flooding and have been given in Section 1.7.3.
The wheel loads of the vehicle will be either in the panels formed by the
longitudinal and transverse grid lines, or on the nodes. The wheel loads
falling in the panels are to be transferred to the surrounding nodes of the
panels to facilitate the analysis. The distribution procedure of point wheel
loads to nodes of the panels will be discussed in the next Section.
In order to obtain the maximum response resultants for the design, different
positions of each type of loading system are to be tried on the bridge deck.
For this purpose, the wheel loads of a vehicular loading system are placed on
the bridge and moved longitudinally and transversely in small steps to occupy
a large number of different positions on the deck. The largest force response
is obtained at each node.
c) Impact Load
Another major loading on the bridge superstructure is due to the vibrations
. Application of Loads, Analysis, Force Responses and Interpretations 115
caused when the vehicle is moving over the bridge. This is considered
through impact loading. IRC gives impact load as a perCentage of live
load. As per IRC Code, impact load varies with type of live loading, span
length of the bridge and whether it is a steel or a concrete bridge. The
impact load can be calculated using formulae or could be directly read
from ready-to-use graphical plot (Fig. 1.15). The impact load, so
evaluated, is - directly added to the corresponding live load.
5.3 IDENTIFICATION OF PANELS IN THE GRILLAGE

When longitudinal and transverse members of the grillage cross each other,
they form panels and the grillage is therefore divided into number of such
panels. All the wheels of the vehicular loading system may not come di-
rectly on the nodes of the grid but usually majority of the wheels fall inside
the panels. These wheel loads acting on the panels are to be transferred to
the contiguous nodes forming the panel, before the grid is analysed by the
grillage analogy. Therefore, it is essential to identify the panels of the
idealised grillage deck in which a particular wheel load is lying.
The identification of panels involves their types and location in the
grillage. The panel may be rectangular, triangular or in the shape of a
parallelogram. Right bridge will have rectangular panels only (Fig. 5.3a)
while skew bridge with orthogonal arrangements of transversals will have
triangular panels near the ends and rectangular panels in the central region
as shown in Fig. 5.3(b). In case of skew deck with small skew angle where
the transversals are taken parallel to the supports, parallelogram panels as
shown in Fig. 5.3(c), will form. In case of skew decks with orthogonal
arrangements of transversals, the upper and lower triangular segments are
formed with polar symmetry, as shown.
In skew bridges, it is likely that only one wheel of an axle may lie in the
triangular panel or in the parallelogram panel, while other wheels are out-
side the deck, as loads P1 and P5 shown in Fig. 5.3 (b) and Fig. 5.3(c). This is
not the case for the right bridge where all the wheels of an axle lie either on
the rectangular panel or outside it (Fig. 5.3a). It is also possible that some of
the axles of the train of wheels be out of the span. This happens when the
axle load in question has either not entered the span or has already crossed
it. In all cases, bridge is analysed only for the wheel loads that lie on or
between the support lines of the bridge deck. Transversely, the wheels are so
positioned that the wheels closest to the kerbs maintain a minimum
specified distance from either of the kerbs.
When wheels of a vehicular loading system enter the deck, the position
of the wheels are considered one by one and the panels in which they fall
are identified. As the coordinates of the nodes and grillage geometry are
+P +P77
+ P6 + P6 Upper
Triangular
Panel
+P5 +P5
t
Ly
Lower
+P4 -1-P4
Triangular
+P3 +P3 Panels
+P
. 2 +P2
+121 -1-P1 Width --1
(a) Rectangular Panels (b) Triangular and Rectangular
in Right Bridge Panels in Skew Bridge
Parattelogrum
panels
&Y
(c) Parattelogrum Panels in Skew Bridge
Figure 5.3 Panels of Bridge

known, the type of panetand its location in the grillage are determined.
The load in.the panel is now to be transferred to the contiguous nodes of
the panel in the form of equivalent loads.
A panel may have more than one wheel loads. In such a situation,. each
load is transferred to the nodes independently and the effects are summed
up algebraically at each node due to all wheel loads falling inside the
panel. Similarly, the nodes will receive loads froti other panels around it
and also any load which may come on it directly.
5.4 TRANSFER OF LOADS TO THE NODES

The grillage analysis requires that loads be transferred to the corresponding
nodes in the form of equivalent-loads. These equivalent nodal loads can be
computed using any one Of the following two approaches :
i) Simple statical approach where the load is apportioned in the form of
-equivalent vertical shear assuming that the panel between contiguous
grillage elements is simply supported along its boundary.
ii) Another approach is .where the equivalent load .consists of vertical
shear and moments assuming that the panel between the contiguous
grillage elements is clamped at its edges.
Although the first approach is simpler, the neglect Of fixed end
moments will lead to .some error. The neglect of flied end moments in the
longitudinal direction does not usually give rise to any significant error but
their neglect in transverse direction can result in some inaccuracy in
transverse moments. This is more so when the actual loads are acting on
the cantilever portion of the slab or when the spacing between girders is
large or when only a few concentrated loads act on the deck. If statical
division of load is to be adopted due to its simplicity, then it has to be
ensured that the transverse spacing of grillage members is small and the
loads are well dispersed longitudinally and transversely. Often, this is
achieved by placing additional 'nominal' longitudinal grid lines in between
or outside the main longitudinal grid lines. These nominal grid lines may be
assigned zero inertia values.
The second approach, where the loads are distributed in the- form of
vertical shear and moments, is more tedious but theoretically superior. As
the computer is invariably used in the analysis of grillage, this tediousness
may not be considered an impediment to its use. However, both the ap-
proaches are in practice and if the grillage mesh size. is small, the results
given by both will be close. But if the mesh size is coarse, only the second
approach is rp.cnrymnprot.d.
Both the approaches for the distribution of dead loads and live loads are
discussed.
5.4.1 Transfer of Dead Loads

The dead load in the form of u.d.l. on the longitudinals, as evaluated
earlier, are distributed to the nodes of the grillage. This is done in two ways:
i) Assuming the longitudinal grid lines to be simply supported at nodes,
the vertical load at each node of the longitudinal grid line due to u.d.l.
are obtained statically. The total equivalent vertical load on each node
is obtained by summing up the loads coming from the1adjacent mem-
bers.
ii) Assuming that the longitudinal grid lines are fixed at nodes, the
equivalent nodal loads will consist of a vertical load and a moment.
Thus, if W per unit length of u.d.l. due to dead load is acting on any
longitudinal grid line as shown in. Fig. 5.4, the vertical load V and
moment M at node 2 1is given by
V
w L 2 ) m (WL WL
2 12
W/unit Length
o. (Z)
nrYerrel 3
1-*-- Li ___________ L
}M2 V
Figure 5.4 Transfer of Dead Load on the Nodes
Similarly, the torsion T per unit length due to the transverse eccentricity of
loads can also be distributed on the nodes treating the concerned longitudinal
grid lines as fixed.
As a sign convention, downward vertical force has been assumed as
positive. For moments, right hand screw rule has been followed.
5.4.2 Transfer of Live Loads-

The live load may either be in the form of u.d.l. on footpaths, if provided,
or vehicular loading system moving on the deck. These live loads are to be
transferred to the nodes of the grillage suitably for the analysis.
a) Live Load on Footpath

- . .
, .; !IL ithe fOCA-
path is to be considered. As per IRC:6-1987 [10] the footpath live load
Application of Loads, Analysis, Force. Responses and Interpretations 119
consists of a uniformly distributed load on its area over part or full length.
Its magnitude etc. are discussed in Section 1.7 of Chapter 1. The footpath
live load may be distributed to the longitudinal grid lines situated in its
Vicinity according to the tributary area of each grid line, like dead load,
discussed above. The transfer of this live load to the nodes of the grillage
will also be done in the same manner as dead load. The bridge structure is
analysed for footpath live load on one side of the deck or ontoth sides and
over part length or full length and the resulting force responses are com-
bined with force responses due to other live loadings, if the combination
results in increase of responses.
b) Vehicular Live Load

The vertical wheel loads acting inside the panels of the grillage are
distributed in the form of equivalent loads to the !lodes of the concerned
panel. This is done by adopting any one of the approaches discussed above.
Both approaches are dealt with for rectangular, triangular and parallelogram
panels.
The transfer of concentrated wheel load lying in the panel of the grillage is
carried out in two steps. First the load is distributed along the direction
parallel to transverse grid lines and then these forces are transferred to the
adjacent nodes of the longitudinal grid lines.
Case I : Equivalent Vertical Nodal Load

The panel is assumed to be simply supported along the boundary.
In a rectangular panel (Fig. 5.5), the equivalent vertical nodal loads are
obtained by simple statical division and are given by
ad ac
P P i = Lx L,, P
L., L,
P= P P= be bd P
i (5.2)
- 1,, L, 4L, Ly
For a lower triangular panel with load P as shown in Fig: 5.6, the equiva-
lent nodal vertical loads are obtained by first distributing thz, load on lon-
gitudinal 1-2 and end transversal 1-3 at edges E and F respectively in a
direction parallel to the transversal 2-3. The equivalent vertical loads at E and
F and on nodes 1, 2 and 3 are given by
(s (I)
Pr= P PF =s P
Ly
X
2 3
y
P F
1 i 4,
Hb a
Figure 5.5 Statical Distribution of Loads in Rectangular Panel
Ly
d
b
3
TM) (
C
)
L.
1
(A)
Figure 5.6 Statical Distribution of Load in Triangular Panel
Hence
C pE C p L:113
F =
Lx L, L.
P (4a L td)
=- 6 - 1 L PE = - p
L.,. 4.
P3 =-- La PF =---d , F
(5.3)
4.
Application of Lo'ads, Analysis, Force Responses and Interpretations 121
The above expressions for nodal forces can also be alternately obtained by
dividing the panel load P in ratios of the equivalent corresponding areas of
triangles formed and the total area of the triangular element, i.e.
area of A BPC
P 1=
area ofd ABC
area oft APC
P2
area of A ABC
area ofd APB

P3 = (5.4)
area of A ABC
By substituting in terms of the dimensions shown in the figure, the nodal loads
P1, P2 and P3 obtained from equation 5.4 will be same as given by equation 5.3
above.
The equivalent nodal loads in upper triangular elements (Fig. 5.3b) will
,be same in magnitude and direction as in lower triangular element given
above. In panels having parallelogram shape and containing the wheel load
P (Fig. 5.7), the equivalent nodal vertical loads at E and F and on nodes I,
2, 3 and 4 are similarly evaluated and given in equation 5.5.
P
E =. 2- P P F P
Y
= ad ac
_P
p
P
P
P
1
2 LXLy
= be p
P4 bd p
(5.5)
3 4
L , L y
I
.
Case If : Equivalent Nodal Vertical Load and Moments

The concentrated wheel load falling in the panel is distributed to the
corresponding nodes of the panel in the form of vertical shear and moments,
treating the panel between the contiguous grillage elements as fixc:d along
the edges. Again, the transfer of wheel load is done in two stages as in Case I,
i.e. first the load is distributed on the longitudinals at its edge parallel to the
transversals in the form of vertical loads and moments and then these are
again transferred to the nodes of the longitudinal as in a fixed beam sub jected
to concentrated load and moments on its span. The distribution of loads to the
nodes of the rectangular, triangular and parallelogram panels is
b a
Ly
Figure 5.7 Statical Distribution of Load in Parallelogram Panel
illustrated and the resulting equations for vertical shear and two moments at
the nodes of the panels in each case are derived.
Vector forces on the diagrams are shown in the positive direction of axes
but proper signs are incorporated in the expressions given: Negative sign
preceding the equation will indicate that the force is in opposite direction to
the vector drawn on the diagrams. Also, the equations pertaining to fixed end
distribution of a point load, a bending moment and a torsional moment for a
beam are given below which will be useful in subsequent discussions.
a) Fixed End Distribution of a Point Load P for a beam AB (Fig. 5.8a)
b 2 (3a + b) D D-
a2 (3b + a) P
I 5
I
A =
ab2 a2 b
M= M= P (5.6)
A L2 L2
b) Fixed End Distribution of a Bending Moment M (Fig. 5.8b) 6

ab
P
A=-P M
M 1,3
A b (2a b) M MB a (2b a) M (5.7)

L2 L2
c) Fixed End Distribution of a Torsional. Moment T (Fig. 5.8c)
TA =1Lb Ts= 7:La (5.8)

P
M
A 61 I . OmB
1____________
4 _______ a w1-4
L
b .).
,.
PA i
1 P
B
(a) Point Load Distribution
MA rm MB,Ai
t's 11/4 B
a _________________________________________________ I - b ______.1
a
_________________________L .i
Pt: PB
(b) Bending Moment Distribution
(c) Torsional Moment Distribution
Sign Convention : Moment Force Torsion
Positive Positive Positive
TA a /6- lAl b
I
Figure 5.8 Fired Ended Beam: Distribution of Forces at the
Ends (a) Rectangular Panel

Consider a rectangular panel having nodes 1, 2, 3 and 4 with a wheel load P
acting as shown in Fig. 5.9. The load P is first transferred to points E and F
lying on the longitudinal members 1-2 and 4-3. Treating the edges E and F
fixed, the vertical forces and moments about X-axis at E and F are given by,
2
P E a 3( 3 b + a ) b 2
( P3F a 3+ Pb )
=
L Dy Ly
M y2 M Y7
MX2 11 ;P 11 MX3
A
t - MXF
MX E
P
F
1 4
M
e )E------ba
l
___1fhP4
_____________ LY
Myi
Figure 5.9 Load Distribution in Rectangular Panel
a2 bn ab2M
p
(5.9)
=
2r
y
'14 L AC
L2 4
Again treating the member 1-2 as fixed at its ends, PE will induce at node
1, vertical force P1 and moment and /1//..x will cause moment Mxi in the
directions shown and their values can be written as
d2 (3 c + d)
P
= Lg PE
x
ar Cd2
Similarly for other nodes of the panel, the vertical force and the two
moments can be evaluated as follows:
At node 2
C2 (3d +c)
P2 ' 3 PE
Lx
2 d
M =MXE M =-- r (5.11)
E
Lx L2
At node 3
g (3d+ c)
3 = L 3 F
c2d
M r ri (5.12)
13 L 2F
At node 4
d2 (3c +d)
P4 L3 F
c d2
Mx4 = MxF MY4 = (5.13)

LI. L2x. "
(b) Triangular Panel
Figure 5.10 shows a lower triangular panel of the grid acted upon by a load P. The load P is
first distributed to members 1-2 and 1-3, parallel to the transversal 2-3 at E and F respectively.
The vertical forces PE and PF and associated moments MXE and MXF at E and F are given by
d
PE=(s d) 2 (2d s) P P F 2 (3 s 2 d ) P
s3 S3
(S d) 2 d p d2 (s d) p
M XE 2 M XF 4
(5.14)
Now, PE will cause vertical forces P. and P2 and moments
M yi and Mn and the moment MXE will induce moments Alm
and Mr at nodes 1 and 2
6\14'\
- e:
\
M
Y2
.4___ Nwr
Mx2 LY _______________________________________________________________________________________
1.R2
C
X
Lx
a
1(A)
.4,45\ Myl
Figure 5.10 Load Distribution in Lower Triangular Panel
respectively as as shown. Similarly PF, on member 1-3 will lead to vertical

forces PHI) and P3(I) and moments M
i(1) and M3(1) at nodes 1 and 3 respec-
lively. The vertical forces and moments due to PEXE and PF. at nodes 1
and 3 are as below
At node I due to PE
P =C2 (3a + c)
L3 E

c
MX1= MXE AlYL = c2 a P (5 .15)
x
At node 2 due to 3
a2 (3 c + a)
P 2 or P E t = LS
=7J
a
M o r M X S =LM X E
s
2
c
Mr2 or MYB L 2 = a PE (5.16)
At node 1 due to PF
C2 (3a+ c) c2 a
Pim= L3 PF
M =
"I) sin a PF (5.17)
At node 3 due to :PF
P3(1) = a2 (3c + a) a2 c
PF
M3(1) LPF (5.18)
I sin a
Now MXF acting in the direction of X-axis is resolved into components MXF
sin -a and MXF cos a in the directions parallel and perpendicular to member 1-
3 respectively. MXF sin a will cause torsional moments M1(3) and M3(3) at
nodes 1 and 3 and MXF cos a will cause bending moments M1(.2) and vertical
force Pim at node 1 and bending moment M3(2) and vertical force P3(2) at node 3.
The expressions for these force resultants are given below
a
M1(3) = L MXF sing M3(3) L = MXF sin a
P1(2)L 3 = 6 ac MXF sina COS CC
c (2a c)
M1(2) = L MXF cosa
!
P3(2) = 6 acL MXF cosa sina
a (2c 1 a)
(5.19)
M MXF cosa
3(2) =
:"1 .
Now at node 1, the total vertical force PA, moment about x-axis MxA and
moment about y-axis M, will be,
128 Grillage Analagy.in Bridge Deck Analysis
PA = PI + P1(1) + PI(2)
c2 (3a+ c) c2 (3a+ c) 6ac
PF M xF sina cos a
L3 P+x L x
3
L x
MgA = Mio) cos a + M x1 + M1(2) cos a + M1(;) Sin a

c2 a..,,
= .- MXE cIvi cot c (2a c) ,, cos2a + c
Lx sin2
2 .c a
. Lx r Lx L!
Al m = M n M1 ( 1 ) sin a M1(2)
sin a + M 1 ( 3 ) cos a c2 a c
= - - rPE - - r +
E
MXF sin a cosa + MXF sin o: cos a

C2 a n c (2a c)
L 2
L 2F
Ls2 Lx
(5 . 2 0 )
Similarly at node 3, the total forces Pc, Mxc and M 1 will be,
P = P + PC 3(1) 30.1
a 2 (3c+ a) , bac
rF + MxF sin a cos a L3x
Mxc = M3(1)cos a + M3(2) cos a +.M3(3)sin a
a2 c a (2c a)
= r2 FcoLa costa MxF + a MxF sine a
"L L! Lx
M = M3(1)sin a M3(2) sin a + M3(3) COS a

L2
cr La(2ca)
= +
2
M xF sin a cos a + a MxF sin a cos a (5.21)
x
Similar expressions can be derived for upper triangular panel also (Fig.
5.11). The distribution of vertical forces, bending and torsional moments on
nodes of the panel are shown.
The resultant vertical forces, at joints A, B and C will remain same as in the
case of lower triangular panel. Also, the bending and torsional moments at the
three nodes will be the same in magnitude but opposite in direction. (c)
Parallelogram Panel
The load in a parallelogram panel can be distributed to its nodes in the
similar manner as in the case of rectangular or triangular panels discussed
- - .
' Application of Loads, Analysis, Force Responses and Interpretations 129
limxi
My1
__ y
41-
;'"s
&-
ty
e 4.ar Mx
MXF
4}h>si>
CV') 17,1
/4.
1
P
3 `ry 4N"' .\ 3(C)
4'62 b d 2 (B)
2
_
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
_
__
_
__
_
Ly_________________________________________________________________"4 M
Y
11M x2
Figure 5.11 Load Distribution in Upper Triangular Panel
earlier. Figure 5.12 shows a

parallelogram panel and the distribution of load to its
nodes in the form of equivalent vertical loads and
moments. The vertical loads and moments at edges E
and F are given as
a2 (3 b + a) b 2 (3a+ b)
E L3 PF Ls
P
2
M =a b P sec a b2a
MF 2 See a
E L2.1. L
=
a2 b
P b2 a
tilXE 2 MXF =
L L2
P3 mx3
0
Y4
tlx2
P2
my 2. MF MXF
d 0 YF_
PF
Lx
PO
ticlx4
P.4
1
01
p H-- b .i* a 1-1 x
Ly
Figure 5.12 Load Distribution in Parallelogram Panel
2b ab2
11.11.E=-- P tan a m= P tan a . (5.22)
L2 YF L
The forces at E and F are distributed at nodes 1, 2, 3 and 4 and the

resultant vertical loads, bending moments and torsional moments at each node
are directly written as
d2 (3c +d)D 6cd

P = 3 FE 3 IVIYE
Lx
d
xk1L E
c d2 d (2 c - d)
PE YE (5.23)
Lt L;
P2 c2 (3d + c) 6cd
P E 3+ Y M
L 3 E 1, E
x
L a .
C 119
M XE
x ' '
e2 d c (2d
2 c) mYE
Y 2
=L 2x E L
(5.24)
(3d+c)
C2 6cd
P3 = L 3 I F + L3 MYF
Lx
Mx3 L i v i x"F
x
,Y3 c2 d c (2d c) MYF

Lx (5.25)
= r F 2
L2 Lx
P4 d2 (3c +d) 6cd

---
3 F 3 MYF
Lx Lx
d
M m ,
xa XF
(5.26)
c d2 d (2c d)
frf r- ________________ MYF
Y4 L2 F 2
Lx
Thus, it is possible to distribute wheel loads falling in panels of different

shapes to the contiguous nodes either in the form of vertical forces only by
statical distribution or by a combination of vertical loads and moments. The
choice rests with the designer. If the computer is used for the transfer of
loads, it is prudent to replace the vertical load lying in the panel by a
combination of vertical loads and moments.
5.5 GRILLAGE ANALYSIS AND FORCE RESPONSES

After the loads are transferred to the nodes of the grillage in the form of
equivalent forces, the grillage may now be analysed to determine nodal
deformations and member forces.
132 Grillage Analogy in Bridge Dick Analysis
5.5.1 Analysis of Grillage

Direct stiffness method is an effective tool in analysing the grillage on a
computer. As mentioned earlier, there are three possible displacements at
each joint of the grillage. These, for a grid in X-Y plane, are joint rotations
about X and Y axes and joint translation in Z-direction, normal to X-Y plane.
The displacements in its own: plane and rotation about Z-axis are small and
are ignored. The analysis of grillage by the stiffness method involves the
following steps
1. Formulation of Stiffness Matrix
The first step in the analysis of grillage involves the generation of the
stiffness matrix of the structure corresponding to the appropriate degrees of
freedom. A 6 x 6 member stiffness matrix [Km] is generated for each mem-
ber of the grillage in terms of its geometric and elastic properties. But the
matrix [K.] developed is in terms of local degrees of freedom which are
different for different members meeting at a joint. Therefore, the matrix
[K.3 for a glid member is transformed into global degrees of freedom by
using a 6 x6 transformation matrix N. This takes into account the orienta-
tion of the member. Thus for a grid member, the global member stiffness
matrix [K,:] is obtained.
Now, to satisfy the equilibrium conditions at a joint, the assemblage of all
the members meeting at that joint is considered. The structure stiffness
matrix [K] is obtained by assembly of elements of global member stiffness
matrix [I C ]. This assembled matrix [K] is nothing but the addition of all
internal forces which will be subsequently equated to the externally applied
loads along the same degree of freedom.
One of the basic properties of the matrix [K] developed is its symmetry
and banded nature and the advantage of this fact is taken in storing only the
banded upper triangular portion of the matrix. This enables us to analyse
bridge with a large number of nodes.
2. Formulation of Load Vectors
External equivalent loads applied at the joints (nodes) of the grillage
constitute the load vector {B}, having moments about X and Y axes and
vertical force along Z axis. The load vectors may either be formed separately
for dead load, live load and impact load or their effects are added together
and a single vector is formulated.
3. Identification of Support Conditions
The stability of the structure is considered by introducing the boundary
'7417.-.t-??2;4.7.'
conditions. The supports of the bridge may be either on compressible

(yielding) neoprene type bearings or rigid (non-yielding) steel or concrete
bearings. Depending on the type of bearings used, stiffness matrix devel-
oped above is to be modified suitably.
The neoprene bearings can be assumed to have negligible rotational
stiffnesses: The axial stiffness of the neoprene beating is taken and added to
the terms in the stiffness matrix at the position corresponding to the vertical
dgfiection of the supported node. Since bearing dimensions are not known
in the beginning, a suitable size of the bearing is assumed initially and its
axial stiffness is computed and used and finally modified, if reactions
obtained so demand, or alternately rigid bearing is assumed initially and
based on the reaction and rotations, suitably dimensioned neoprene bearing
may be adopted and a revised analysis is carried out. For skew bridges, the
stiffness of the bearings has marked effects on. the structural behaviour and
it is very important to account for their effects properly. Yielding of bearing
tends to even out the differences between various reactions. -
When simple rigid type supports are used, the rows and columns corre-
sponding to the vertical deflections are removed from the stiffness matrix.
The load vector has also to be modified by removing the rows corresponding
to the vertical deflection at the supported nodes. The deformation vector that
is obtained by solving the modified stiffness matrix using the modified load
vector does not contain the vertical deflections at the supported nodes. The
deformation vector is modified to include these. The support locations in the
grillage is kept at the same place where it actually exist.
For obtaining the reactions at the supports, the structure stiffness matrix
[K] is partitioned into four sub-matrices pertaining to the free and restrained
deformation vectors {DP} and {DR} and correspondingly load vectors into
external load vector {-P} and reaction vector {R} as discussed in Chapter 3.
The support reactions can be obtained by using Equation 3.18.
4. Solution of Simultaneous Equations
A large number of simultaneous equations will result from the assembly
of stiffness matrix considering equilibrium at each joint. The number of
these equations depends upon the size of the grid. Efficient techniques like
Gauss-Elimination and Cholesky Factorisation are available to solve these
equa-tions. The resulting deformation vector can be used to compute
member forces using basic member properties. The procedure is already
dealt in Chapter 3.
5. Determination of Nodal and M h
Pin Pr nOri:ar!r
The solution of simultaneous equations will yield nodal deformations of

the structure. The member deformations in global as well as local coordi-

nates can be evaluated by multiplying member stiffness matrix with the
member deformation vector. The output or the result consists of vertical
deflection and rotations about X and Y axes at each node, shear in each
member, bending moments at the two ends of the member and torsional
moment in each member. The output also gives reactions at the supports.
The formulation of member stiffness matrix, its transformation, assembly,
support conditions, reactions etc. have been discussed in general in Chapter 3
and may be referred there for detail.
Normally the above processes for the analysis of grillage are handled by
computer using a suitable program meant for the same. Many compact
packages are available for the grid analysis for a set of given loads on the
nodes of the grid and may be used. One such computer program in FOR-
TRAN is given at the end of Chapter 3.
However, a more general computer program for the analysis of grids
subjected to different types of IRC loadings or any other special live loading
termed as User Specified live loading has been developed and discussed in
the next Chapter. The program transfers the load from panels to the nodes
automatically and analyses the bridge for various positions of loads.
5.5.2 Force Responses

As discussed above, the solution of equations yield nodal deformations i.e.
deflection, slope and rotation at each end of the member. The shear force for
a member, the bending moments at the two ends of the member, the torsional
moment in a member and reactions at the supported nodes are the usual
output. However, these output can be modified and more details are possible.
Ordinarily the output is obtained for various longitudinal and transverse
positions of different types of live loading. Invariably the output obtained is
very large. Scanning this output, for a grillage of even moderate size, is a
problem. Therefore, to reduce the output data, only the critical values of the
force responses need to be retained.
For the design of any bridge structure we need the envelope diagrams of
various responses on it. The envelope diagrams are the response diagrams
drawn along the longitudinal grid lines with the largest values of responses
picked up under live load. This may be achieved for a particular live load by
moving, it over the deck in small increments both longitudinally and
transversely and for each of the load positions, the deck is analysed. When
the load moves from one position to the next position, the responses are
again obtained for this new position of load and these values are comoared
with the previous values. The larger values of each force responses like shear
force, bending moment and torsional moment for each grid member
are retained alongwith the corresponding load position, deleting the smaller
values. The process is repeated till the whole length and breadth of the bridge
is covered by the live load.
The load position for each critical value is given through the coordinates of
the left most wheel of the leading axle. This information of load position
could be used for positioning the live load on the deck and carrying out a
manual check if so desired.
The number of movements of loads in longitudinal and transverse direc-
tions will depend upon the factors like span, carriageway width, type of live
loading, extent of accuracy desired, available computer time, etc. However,
as a preliminary guidance, the movements of loads in increments equal to
about 1/15th of span length or half the size of the mesh in longitudinal
direction is chosen. The movement of loads in transverse direction is very
much limited due to the restrictions on wheels from coming closer to the
kerb by a specified distance. The loads may be moved transversely in five to
seven, equal intervals in a two lane bridge and in steps of about 750 mm in
wider bridges.
The initial and final positions of the live loading on the deck should be so
chosen that no critical response is missed out. The initial and final positions
of wheels in longitudinal direction (X min and Xn) and in transverse direction
(Ymin and Yn.) are shown in Figs. 5.13a, b. The coordinates of these initial
and final positions can be calculated with the help of Table 5.1.
TABLE 5.1: Computation of Initial and Final Coordinates of Reference Wheel
Type of Loading Loading Reference

Number
Class A Two lane 2 5300 400 18800

Class 70R Train
Col. 'I' and Col. 'm' 3, 4 2380 1405 13400
Class 70R Bogie
Col.'/' and Col. 'in' 5, 6 2380 1405 1220
Class 70R Track 7 2060 1620 4113
Class A Single lane 8 1800 400 18800
Xthin = 0 Xmax =L+U+Btan
Ynth, = A + S ) Y =BAC-5
where A = Transverse distance of Kerb from exterior grid line (AKERB1*

B = Right distance between exterior longitudinal grid lines
C = Transverse distance between centres of exterior wheels
L = Right span for right bridge and skew span for skew bridge
S = Mini mu m tr an s ver se dis t an ce o f c ent r e o f l e ft mos t wh e el fro m
Kerb (SCLMIN)*
U = Longitudinal length of loading )1. = Skew angle
Notations used in the program. Distances are in mm
_
Loading Axle
p
1"1"-- YMAX
c
_FH44,R- rncn-751---1

Left Most Right Most
Grid Line B _ Grid Line
(i)
++++Cross-section
X=L ++++ 1
XM 1 1 t+
0 CI)
B 4T*-
0 0
;41
(0,0)
+ 4I+
+ + + +
0 0
()Lett Most Wheel of Leading Axle 0 0
Right Most Wheel of Rear Most Axle 0 CD
(ii) Right Bri dge Plan
Figure 5.13 (a) Initial and Final Positions of Loading to Cover Entire
Right Deck
The total number of positions which a live load may occupy on the deck and for
which the bridge will be analysed. are:
F"' 1MAX
Loading Axle
c -0 1 C
4141-p-n'
Left Most ___________________________ Right Most

Grid Line 14 _____________ B _____________ Grid Line
( 1 ) Crosssection
Y=0
OF 1+ - 02)
+ + X M A X
X=L
XMAX
(0,0)
u
I
Left Most Wheel of Leading AX1e 0-0

(I) Right Most Wheel of Rear Most Axle (D
(ii) Skew Bridge Plan
Figure 5.13 (b) Initial and Final Positions of Loading to Cover Entire Skew
Deck
(X . X min +1 ) Y max Y min +1

I n c r e m e n t )
in X direction Increment in Y direction
It is possible to eliminate some of the positions out of the above, based on

experience, judgement and a little manual calculations. Alternately, one
may take the view that as computer is involved in the analysis, more posi-
tions of loads do not matter and carry out the analysis for all the above load
positions.
5.5.3 Design Envelopes .

In order to design a bridge for IRC loading, it is not sufficient to analyse the grillage for any one type of
live load only and obtain response envelope diagrams for it. The maximum responses due to one particular
type of live load may not be critical at all the points on the deck and it has to be scanned for other types of
live loads also to obtain the largest design responses.
To achieve this, each live load system is moved longitudinally and transversely in small increments to
cover the entire deck. The grillage is analysed for each of these positions and the maximum values of
responses are retained alongwith the corresponding load positions; the maximum response results of various
types of live loads are compared with each other and the highest values along with their load positions and
type of loading are retained giving an overall envelope diagram for each response separately.
Normally a one lane bridge is designed for Class A single lane loading. A two-lane bridge is designed for the
following alternate live loadings :
i) Two lanes of Class A loading
ii) One lane of Class 70R Train loading,
iii) One lane of Class 70R Bogie loading, and
iv) One Iane of Class 70R Track loading.
Class 70R Train loading and Class 70R Bogie loading each have been
considered with two alternate types of wheel arrangements on the axle, Col.
`1' and Col. `tn'. Refer section 1.7.3 of Chapter 1.
In addition, the program also caters for any other specified live loading. The
user in such case is required to give full details of the specified live loading.
To summarise, following are the loadings together with. their loading
..:.:_:. program.T ne program can move
various live loads on the bridge deck through prescribed steps along the span
and across the carriageway width.
Loading (1) Dead load
Loading (2) Class A Two lane
Loading (3) - Class 70R Train : Col 1
Loading (4) Class 70R Train : Col m
Loading (5) Class 70R Bogie : Col 1
Loading (6) Class 70R Bogie : Col m
Loading (7) Class 70R Track
Loading (8) Class A Single lane
Loading (9) Specified by user
Loading (1) represents dead load. Loading (2) to Loading (9) represent
various types of live load.
It is built-in the program that live load results are given inclusive of dead
load results. If only live Ibad results are desired, then in the Input data, dead
load can be given as zero.
5.6 INTERPRETATION OF RESULTS

The Output or the result obtained from the analysis of grillage consists of
vertical deflections and X and Y rotations of each node, shear force and
torsional moment of each beam element, bending moments at the two ends of
each beam element and reactions at each support.
The above results are to be judiciously used while designing a bridge deck.
Since the deck has been initially idealised as a grillage and the analysis has
been performed on the idealised grid, the results may sometimes need
modifications and proper interpretations before they are finally used in
design. Some of the important interpretations of the Output and its modi-
fications due to the local effects for slab bridges, slab-beam bridges and
cellular bridges are discussed below.
5.6.1 Slab Bridges

The computer output for deformations like deflection and rotation and force
responses like bending moment, shear force and torsional moment are to be
thoroughly examined and judiciously interpreted in slab bridges.
Modifications in the output results are made, if necessary, due to local effects
which are not considered earlier in the grillage analysis and the modified
responses are to be used in the design for better accuracy. Some
of the significant obseivations pertaining to force responses, for slab bridges

are discussed here.
The slabs are designed on the basis of per unit force response. The
computer output gives response for the width which is represented by a
particular grillage member. Hence, these responses should be converted into
per unit width befOre these values are considered for design.
The. analysis will give positive or negative values for various force
responses with respect to each beam element and due care should be taken to
adhere to the sign conventions adopted in the computer program. For
example, a negative bending moment value at end I of a grillage element will
denote a sagging moment while it will indicate a hogging moment at end 2 of
the same element.
When a grillage member continues across a node, the values of moment at
end 2 of one member and end 1 of the adjacent member in continuation will
be usually different. This is due to the torsional moments in the members
framing in other direction. To deal, with such "Stepping" or "Saw -Tooth" in
the moment value's along a grid line at nodes, these two moments may be
averaged out (Fig. 5.14).
Support Grid Nodes Support
.... ..,
_ l
mertt
\Griltage Output
Figure 5.14 Grillage Output Saw -Tooth Moment Diagram
Only one value of the shear force for a member of the grillage is obtained
from the output and the same may be used in design as such. Similarly,
maximum reactions printed, are taken as design values for reactions at
supported nodes.
The torque per unit width in a true orthotropic slab is same in orthogonal
directions, however, it is often different when read from 2TilIage output.
The torque per unit width at any point should be taken as the average of
these two outputs.
Application of Laads, Analysis, Force Responses and Interpretations 141
The design of the section in bending should be based upon the principal
moments mi and in,. Referring to Fig. 5.15, the principal moments mi, m2 and
their deviation at the nodal point of interest is obtained from the values of
bending moments per unit width mx and my, in X and Y directions and the
corresponding averaged torsional moment per unit width mx,employing the
formulae
2
nix+ my {(mx my )
m1.2 2
- + 2 2 +m 2
(5.27)
ttan2 0 = m xy
mx my
m1
I
2
My
Figure 5.15 Computation of Principal Moments and Their Deviation from
X-Y Axes
In reinforced concrete bridges, the direction of reinforcement may not

always coincide with the direction of principal moment_ This is more so with
skew slab bridges. In such a case, it should be ensured that reinforcement
component in the direction of each principal moment is adequate.
For a critical load combination, if at a bearing point there is a net negative
or downward reaction force as opposed to the usual positive or upward
reaction force then it indicates that under that combination of loading there is
uplift at the bearing point. This may happen for bridges with large skew angle
and there may be one or more such bearing points in the deck.
It may be recalled that while obtaining the above results, conditions of

zero deflections have been imposed at all the bearing points. Thus the
behaviour of the actual structure is not consistent with the assumed condi-
tion of the grid-model. In such cases a repeat-analysis is necessary. In the
repeat-analysis, the condition of zero deflection is not imposed at the sup-
port node/nodes expected to be lifting when analysing under loading form-
ing the above critical load combination. Lifting at the bearing is not desir-
able and should be prevented as far as possible. This could be done by
providing a tension bearing at-such bearing point as opposed to the usual
compression bearing. In case where a downward reaction force is small, a
lateral shift in the position of the bearings can also change it to a net
upward reaction force and condition of no lift could be obtained.
As pointed out in Chapter 4, a slab bridge, strictly speaking, can be
idealised as an assembly of orthogonal beams only if Poisson's ratio v of the
slab material is zero. Ignoring Poisson's ratio which is about 0.15 for
concrete leads to an under-estimation of moments. This under-estimation is
usually negligible for longitudinal bending moments but is not so for trans-
verse bending moments. The error in the prediction of transverse bending
moments can be large in somes cases. This is because the curvature along the
span is considerably greater than that in the transverse direction. The
following relationships can be used to account, approximately, for errors
resulting from the neglect of Poisson's ratio
(M) 70/,)4
(My) = (M)0 + 7(Mx),3 (5.28)
where (M )0and (M)0 are the responses obtained by grillage analysis for
which Poisson's ratio = 0 and (Mx) and (My) are the relevant corrected
moments for a finite value of Poisson's ratio.
Due to dispersion of load, the actual area of application of load is larger
than the area of contact of the wheel with the slab. If this application area is
larger than the grillage mesh, the load can be assumed to be sufficiently
dispersed for the grillage to reproduce the distribution of moments through-
out the slab. No further modification of moments is necessary. On the other
hand, if the area of application of the load is small compared to the grillage
mesh, no worthwhile information will be obtained about the local high values
under the load, though the grillage distributed moment field will simulate
that in the deck. The additional moments due to high local curvatures can be
obtained for the area of slab within the grillage mesh from influence charts
due to Pucher [6].
4
r

5.6.2 Slab-on-Girders Bridges
In beam and slab decks also, the stepping of moments in members on
either side of a. node occur. The difference in bending moments in two
adjacent members meeting at a node will generally be large in exterior
girders. Where all the members meeting at the node are physical beams, the
actual values of bending moment output from the program should be used.
If at a node there are no physical beams in the other direction and the grid
beain elements represent a slab, thd bending moments on either side of the
node should be averaged as there is no real beam of any significant
torsional strength.
Design shear forces and torsions can be read directly from grillage out-
put without any modifications.
In composite construction, where the grillage member stiffnesses are
calculated from properties of two dissimilar materials of slab and beam
elements, the output force response (torque, bending moment or shear) is
attributed to each in proportion to its contribution to the particular stiffness.
The local two-dimensional system of moments and torques in the thin
slab under a concentrated load are not given by the grillage analysis and
can be obtained from Pucher's charts [6]. The local moments obtained from
the charts must be added to those in the slab resulting from twisting and
relative deflection of the supporting beams. If there are no "nominal" gril-
lage members between two physical beams and the transverse members
have not been loaded, these moments can be read directly from grillage
output for the local transverse members. However, if there is a "nominal"
longitudinal member under the load or if transverse members have been
loaded, the slab moments due to twisting of beams are best calculated from
grillage output displacements and rotations of adjacent beams using slope
deflection equations [3].
In case of longitudinal grid lines not physically supported at ends, the
load carried by these lines will flow towards nearby supports via end trans-
versals. If not taken note of this phenomenon this will give lower value of
shear in supported longitudinal grid lines. To account for this under-estima-
tion, shear of these beams be added to the shear of adjacent beams which
are physically supported. Similarly, in order to avoid under-estimation of
bending moment in supported longitudinal beams, the bending moments of
unsupported grid lines also have to be considered for design of supported
longitudinal beams.
5.6.3 Box-Girder Bridges

drawn from the results of the output for
pox-girder or cellular deck again indicate "Saw Tooth" (Fig. 5.14) with
large jumps in moments at the nodes. This is because of the transfer of the
torsional moment in the transverse members at each joint to bending mo-
ments and shear forces in the longitudinal member [3]. The true design
bending moment is obtained by averaging the bending moments on the two
sides of each joint. The top and bottom slab stresses are calculated from these
average moments.
The transverse bending moment in the grillage member is equivalent to
the opposed transverse compression of the top slab and tension of the bottom
slab (or vice-versa) due to transverse flexure without Tlistortion. In
narrow decks, it is usually very small compared to the longitudinal bending
moment (except in the diaphragms) but in wide decks it can be large spe-
cially near skew supports. A grillage output transverse moment diagram has
a saw-tooth shape like the longitudinal moment diagram and the top and
bottom slab stresses are calculated from the average moments.
The slab bending moments are derived from the shear force in the trans-
verse grillage members. The fraction of this shear force carried by each of
the top and bottom slabs of the cell are assumed to be proportional to the
flexural stiffnesses of the slabs. Assuming also that points of contraflexure
lie midway between the webs, the moment at each end of a slab is simply
the shear force it carries multiplied by half the distance between webs. The
transverse slab moment in the cantilever can be taken directly from the
grillage output since this member is not representing the cell. Local moments
are again obtained using influence charts of Pucher t6]. The total design
moment in cellular deck is obtained by adding the slab moments of the cell,
cantilever moment and the local moments.
The torsion-shear flow in the slabs must be calculated from the average
torque per unit width of transverse and longitudinal grillage members.
The grillage output shear force represents the total shear force in each web
of the deck.
Grillage analysis ignores the effects of Poisson's ratio on the interaction of
longitudinal and transverse moments. This neglect of Poisson's ratio
introduces little error in narrow decks while in wide decks with little stiffness
against cell-distortion, the calculated transverse moments can be con-
siderably in error, if moments are small and the Poisson's ratio is significant.
However, since the concrete has relatively low Poisson's ratio of ap-
proximately 0.15, its neglect may not result in any appreciable error.
Reviewing the discussions contained in Chapters 4 and 5, taken together,
on the grillage analogy method, it can be seen that the procedure is straight
forward and is also amenable to an enaine.er who may not he well conversant
with higher mathematics and numerical techniques. The setting out of
the grid lines and the evaluation of elastic properties of the members can be
handled manually following the guidelines provided above. The determi-
nation of the dead loads etc. is simple. The remaining steps in grillage
analogy i.e. transfer of loads to nodes, formulation and inversion of matrix,
solution of equations, evaluation of design responses etc. are done on com-
puter using: a suitable software package which are not difficult to find
now-a-days. Preparation of input data is also simple (as will be shown later)
and can be done mechanically even without understanding the computer
program fully However, the interpretations of results and its modifications
due to local effects pertinent to a particular type of deck may present some
difficulty and will need some understanding of the structural behaviour of
the deck 'in question.
A general grillage analysis computer program developed for dead load
and various types of live loads is discussed in the next chapter.
REFERENCES
1. BAKHT, B. and JAEGER, L.G., "Bridge Analysis Simplified", McGraw Hill, New
York, 1985.
2. GERE, M.J. and WEAVER, W., "Matrix Analysis of Framed Struture", D. .Van
Nostrand Co., USA, 1965.
3. HAMBLY, E.C., "Bridge Deck Behaviour", Chapman and Hall, London, 1976.
4. JAEGER, L.G. and Bmarr, B., "The Grillage Analogy in Bridge Analysis", Canadian
Journal of Civil Engineering, 9(2), 1982.
5. JAEGER, L.G. and BAKHT B., "On the Analysis of Slab Bridges by Grillage Anal-
ogy", Proc. Annual Meeting of the Canadian Society for Civil Engineering,
Montreal, May 1987.
6. PUCHER. A., "Influence Surfaces of Elastic Plates", Springer Verlag, Wien and
New York, 1964.
7. RUSCH. E.H.H. and HERGENROEDER, A., "Influence Surfaces for Moments in Skew
Slabs", Werner-Verlag, Dusseldorf, 1969.
8. WEST. R., "Recommendations on the Use of Grillage Analysis for Slab and
Pseudo-Slab Bridge Decks". Cement and Concrete Association, London, 1973.
9. WEST. R., "The Use of Grillage Analogy for the Analysis of Slab and
Pseudo-Slab Bridge Decks", Research Report 21, Cement and Concrete
Association, 1973.
10. IRC 6-1987, "Specifications and Code of Practice for Road Bridges-Loads and
Stresses", Section II, Cement Concrete (Plain and Reinforced), Indian Roads
Congress. New Delhi - 1990.
11. IRC 21-1987, "Specifications and Code of Practice for Road Bridges". Section
III, Cement Concrete (Plain and Reinforced), Indian Roads Congress. New
Delhi, 1988.
-- -
Chapter 6
Computer Program
6.1 INTRODUCTION
A Computer Program based on grillage analogy is developed in FORTRAN
for the elastic analysis of bridge decks, covering right and skew layouts. The
program is compatible with MS-DOS environment in its present form but it
can also be compiled to run on mini or main-frame computer. The program
is modular and includes different subroutines to execute specific operations.
Additional features and modifications can be easily incorporated for a future
modular expansion. Sufficient number of explanatory comment statements
have been introduced for better and quicker understanding of the program.
However, the program is written with the premise that the reader has some
prior exposure to computer programming. But if one does not want to go
through the drill of understanding the program modules completely, he can
still prepare the input data and can analyze the bridge.
This Chapter deals with the important features of the program. The
Program Manual consisting of listing of variables, sign conventions, main
and subroutine segments and flow charts is given. The Users Manual
comprises data input and output formats. Limitations and scope of the
program are also discussed.
The listing of the program is given in Appendix II at the end of the book.
A diskette with executable file of the program compatible to IBM PC along
with input data and result files for all the worked examples illustrated in
Chapter 7, is available and can be ordered separately.
6.2 IMPORTANT FEATURES OF THE PROGRAM 'GABS'

The program developed can deal with solid, voided and pseudo-slab
bridges, T-beam bridges with or without cross-beams, composite bridges,
single and multi-cell box-girder bridges and box-beam spine bridges. The
bridge may be supported on rigid or flexible bearings. It can handle right
bridges and also skew bridges where transversals are either orthogonal or

Computer Program 147
skew to the longitudinal girders. Analysis is possible for decks with or

without footpaths. The bridges can be analyzed for different classes of Stan-
dard IRC live loadings [6] simultaneously. The program can also analyze the
bridge for any moving or stationary User Specified live loading such as
AASHTO [5], OHBD Truck loading [4] etc. The wheel loads moving on the
deck are automatically transferred from the panels of the grillage to the nodes
in the form of equivalent vertical forces and moments. Similarly, the dead
load which is given in the form of uniformly distributed load and torsional
moment per unit length acting on the longitudinal grid lines, are also
transferred to the nodes of the grillage in the form of equivalent vertical
forces and moments. The dead and live loads can be dealt with separately
and also responses can be obtained when they are combined. The design
force responses stored are the maximum values of responses obtained from
the analysis of each type of live load after combining it with dead load.
Efforts have been made to make the program as general and versatile as
possible without making it unduly lengthy and complicated.
The program 'GABS' consists of a main program and a number of sub-
routines. The main program reads the problem details, structural and geo-
metrical parameters, member designations, support conditions and loading
details. It organizes data and calls different subroutines. The subroutines
generate the global axes oriented structure stiffness matrix, set the support
conditions, decompose the matrix, transfer the loads to the nodes and solve
the matrix and give nodal deformations for different load cases after com-
bining with dead load. Member forces and reactions are calculated in the
main program itself.
TWo versatile subroutines COMPARE 1 and COMPARE 2 are used to
produce the output for the maximum live load design responses and these
responses are printed through subroutines WRITE 1 and WRITE 2. The
main program and subroutines are discussed further in the Program
Manual and User Manual.
6.3 PROGRAM MANUAL FOR 'GABS'

The program manual gives variables used in the program, sign convention
adopted, details of main program and flow charts and description of subrou-
tines used.
6.3.1 Variables
Some of the variables used in the program for input, output and also
those used in more than one subroutines are listed below:
AIMP Impact factor

AKERB Right distance of kerb face from the left most longitudinal
grid line in transverse direction
ANG Skew angle in degrees
B Array of live load vector (3 x N)
BL Live load vector due to single wheel
CH1 CH3 A character for Yes (Y) or No (N)
CHOICE A character for STOP (5) or START (0)
CORRT A numeral for reading dead load
N Overall maximum reaction
CORTN Maximum reaction under a live load
CORRT Overall minimum reaction
CORT Minimum reaction under a live load
DD Number of types of groups of elements
DLNG Spacing of longitudinal grid lines
DLB E Dead load vector of-size (3 x N)
and G Young's modulus of elasticity and shear modulus of elas-
ticity of the bridge material
El, E2 X-projections of bridge beyond centers of supports at the
beginning and at the end
.FZ Shear at end I of the element PT (i,3)
GRIDTYPE A. character representing 'skew' or 'orthogonal' layout of
the bridge
IK2 A character for dead load case
IX Number of supported nodes
1XG, IXL Counters for wheels out of span in longitudinal direction
JDD Array of numbers of supported degrees of freedom
JRR Type of restraint at support
JTT Array containing node numbers of supported nodes
K(I)- Vertical and torsional u.d.l. on longitudinal girders due to
dead load
L Element length
LCASE Load case type
LDING A character array containing the names of all types of load
cases
M Array containing end node numbers of elements
M(I,1), M(I,2) Node numbers of element I
MM Total number of elements
MP I Torsion at end 1 of the element PT (1,4)
MN 1 , MN2,
MN 1 1 , MN22 Node numbers of the contiguous nodes of the panel
MAXB Maximum values of member end deformations

MAXPT Maximum values of member end forces
MAXRT Minimum values of support reactions
MAXRTN Maximum values of support reaction
MOVE A counter for load placing
MQ 1 Moment at end 1 of the element PT (I,5)
MQ2 Moment at end 2 of the element PT (1,6)
N Total number of nodes
NLG Number of longitudinal grid lines
NOLCASE Total number of live load cases desired to be analyzed
NTG Number of transverse grid lines
NW Number of wheels, in. the loading
NWPA Number of wheels per axle
PATYPE Panel identification character (Rectangular '12.' or
Triangular 'T')
P(I, 1) Length of element in X-direction
P(I,2) Length of element in Y-direction
P(I,3) Flexural moment of inertia of element
P(I,4) Torsional inertia of element
POS Wheel position indicative character (YL, YG, XL, XG, etc.)
PW Array containing load on each wheel
RT Array of support reactions
S Stiffness matrix
SANG Skew angle in radians
SPAN Span in X-direction
SCLMIN Minimum distance of the center of left most front wheel
from kerb
SUPK Beating stiffness
T Number of identical elements in a group
WW Band width
XC, YC Arrays containing X and Y coordinates of all wheels
XL, YL X and Y coordinates of center of left most front wheel of
loading at start
XINC, YINC X and Y increments of left most front wheel
XSTOP, YSTOP X and Y coordinates of center of left most front wheel of
loading at end
XLLG Array of cumulative transversal grid line spacings
XLL, YLL X and Y dimensions of panel
XPL, YPL X and Y distances of wheel position with respect to node
MNI
YLLG Array of cumulative longitudinal grid line spacings

(DLNG)
6.3.2 Sign Conventions

The coordinate system and the sign conventions adopted in the program are
as follows:
i) Origin is taken at node on the left most longitudinal grid line as
shown in Fig. 6.1(a) and 211 the coordinates are measured from this
point.
ii) X-axis is along the span length of the bridge and Y-axis is perpendicu-
lar to it; the positive directions being as shown in Fig. 6.1.
iii) Vertical axis is downward positive and vertical forces acting downward
are also taken positive.
iv) Rotations about X and Y axes are depicted by the right hand screw rule,
the positive directions being as shown in Fig. 6.1(b).
v.) The sagging bending moment will be negative at end 1 and positive at
end 2 of the element 1-2 according to the above mentioned vector
notations.
vi) One value of the torsional moment given in the output for a beam
element pertains to the value at end 1 of the element. The positive
direction is along the member axis as per right hand screw rule.
vii) One value of the shear force given by the output corresponds to the
shear at end 1 of the beam element. Shear at end 2 will be equal and
opposite to shear at end 1. The positive value of shear indicates end 2
up with respect to end 1 and vice-versa.
6.3.3 Main Program and Flow Charts

The main program reads all the necessary data required as Input for the
analysis and calls various subroutines during the process. The flow charts
depict the algorithm followed in the program. The main program does the
necessary calculations according to algorithm and prints the input and output
values as per given formats. The details of Data Input and Result Output will
be dealt with under the Section User Manual for 'GABS'. The detailed Micro
Flowchart (Fig. 6.3) illustrates the steps involved in the program. Various
subroutines of 'GABS' are arranged according to a Macro Flow Chart shown
in Fig. 6.2.
Left-most Longitudinal
Grid line
(
Mx 'ex
r __ My ,6y
Z (downward) Z (downward)
(a) Coordinate System (b) Sign Convention Figure
6.1 Coordinate System and Sign Convention
6.3.4 Descriptions of Subroutines

Various subroutines are arranged according to the Macro Flowchart given
in Fig. 6.2. The functions of each subroutines are briefly described here.
1. Subroutine INODE
This subroutine identifies the members of the grillage corresponding to
end nodes of each member. It stipulates that the node numbers, given as
input data, are chosen in such a way that the node number of end 1 of
the member is always lesser than the node number of end 2 of the same
member.
Read and Print Input Data.
Call "INODE" to Identify Member No. Corresponding to End nodes
Call "DEAD LOAD" to Calculate Dead Load Vector
Call "STIFF' to Calculate Global Stiffness Matrix [K]
If Rigid Supports If Flexible Supports
Call "SORT' and "MODSTIF' ___ Clan "DECOMP" to Decompose [K]

to Modify [K]
1 '
Call "IMPACT" to Calculate Impact Factor
For Each
Live Load
Call "LLOAD" and "WLOAD" to Calculate the Configuration
and Placing of Vehicular Live Loads
Call "PAID!" OR "PAIDRB" to Identify Panel and Wheel

Positions in it
For Each
Wheel
Call "LDISTT" OR "LDISTR" to get Load Vectors from
the Wheel Loads
Call "SOLVE" to Analyse for Deformations

,
4
Calculate Member End Forces and Reactions
Call "COMPARE!" to Obtain Design Responses for Current Load Type
Call "WRITE!" to Print Deformations and Force Responses for the

Current Load Type
Call "COMPARE2" to Obtain Final Design Responses by Comparing all Load Types
Call "WRITE2" to Print Final Design Deformations & Force Responses

with Load Types
STOP
X.4
.L. 1:1317
(START)
INPUT
Grid type, Skew angle, No. of members,
No. of nodes, No. of types of groups of
members, E, G, No. of longitudinal grid lines,
Spacing of longitudinal grid lines, X-
Projections of bridge beyond support lines,
Node nos. of one representative element of
the group, Total no. of elements in the group,
X and Y lengths of the element, I and J of the element,
Nodes nos. of remaining elements of the group.
Identifies member Call Subroutine

no. corresponding "INODE'f
to end nodes.
Is IuN PUT. No. of

pported nodes, Type
of restraint, Node nos.
of supported nodes.
Bearing
Stiffness
INPUT
Y-distance of inner edge
/ I N P U T
of kerb from left most
longitudinal grid line.
Call Subroutine
"DEAD LOAD"
YES
INPUT INPUT
Vertical and Torsional udl Vertical load
loadings on longitudinal grid at each node due
lines due to dead loads to dead load
Assemble global
stiffness matrix
in half band-
width form
Sorts out and modifies

stiffness matrix by
Call Subroutines eliminating rows and
"SORT" and columns corresponding
"MODSTIFFI to restrained degrees of
freedom.
Decomposes the
Stiffness matrix Call Subroutine
into upper & lower 4r-- "DECOMP"
triangular matrices
/ INPUT
"CH3"
YES
STOP)
NO
INPUT
TotalTotal no. of live load
cases to be analysed
DO 1000
Loop for solving all
live load cases
I N P U T
Load case type
"LCASE"
Dead load YES

Vector is
formed
NO
INPUT
I n i t i a l & f i n a l X a n d Y
c o o r d i n a t e s o f l e f t m o s t f r o n t
w h e e l o f l o a d i n g , I n c r e m e n t s i n X
a n d Y c o o r d i n a t e s .
Calculates impact
Call Stibroutine factor according
"IMPACT" to type of loading
and span.
Place the loading with

increments in X and Y coordinates
of left front wheel and check
for final values.
Gives no. of wheels; No.

of wheels per axle; X and Y
coordinates of Wheels
for all standard live load 1-Call Subroutine
cases. For User's specified "LLOAD"
loading, coordinates of wheels
and no. of wheels are read in.
Gives the load on each

wheel for standard loadings,
For User's specified loading
min. kerb distance of centre of
left most front wheel and load
on each wheel are read in.
1Call Subroutine
"PAIDRB"
A
For skew bridge, checks if wheel For right bridge, checks if

is outside span or nearer wheel is out of span or out of
to kerb than allowable limits; allowable transverse eccentricity;
Identifies the containing Identifies the panel containing
panel type, its contiguous nodes, wheel load by its contiguous nodes,
panel dimensions and load calculates the wheel position on the
positions of the wheel. panel and the panel dimensions.
r YES
:han _____________________________________________
NO
Triangular Rectangular
Using fixed beam analogy,

these subroutines calculate the
nodal load vector from the
wheel load placement on
the panel.
Are
all wheel
Next loads analysed
0 4
w h e e l t ___ NO and load vectors
added
Enhances the final live

load vector by impact factor
and add dead load vector
Modifies load vector

by eliminating values
corresponding to
restrained degrees of
freedom
Solves for deformations
Call Subroutine - using front and back
)10-
SOLVE' substitution method on the
decomposed stiffness matrix
and load vector
Spi
Inflates the deformation
vector for rigid bearing
case with zeros at restrained
degrees of freedom.
4,
Calculates member end
forces using member
stiffness matrices and
deformation vectors.
4
Calculates support reactions
by using equilibrium
conditions at the supported
nodes.
Compares and retains

Call Subroutine maximum values of member
"COMPARE]." E' end forces, nodal deformations
41
and support reactions for

the current load type with
load position for each maximum value.
Are
all load
NO positions for
urrent load type
exhausted
YES
INPUT
IK2
NO Is
IK2=I
YES
1' Prints the maximum
Call Subroutine values of member end
"WRITEl" forces; deformations,
reactions etc. for the
current load type.
Are all
NO load
Next
load case exhaus
Compares and retains

maximum responses after
comparison amongst Call subroutine
maximum response "COMPARE2"
values for different
load types. Prints the maximum design
values of member forces,
Call Sithroutine deformations, reactions due
to all types of loads with
corresponding. load type
and load positions.
Figure 6.3 Micro Flow Chart
L
17.. 7, ,
2. Subroutine DEAD LOAD

This subroutine reads vertical load per unit length and torsional moment
per unit length on the longitudinal grid lines due to dead load and
transfers this loading to the nodes treating the- grid lines as fixed;
alternately,. the subroutine can also take dead load in the form of nodal
vertical loads: Loading in-either form is to be calculated manually and
- 6 then given as input data. Dead load is designated as load type 1, i.e
LDING( 1 ).
3. Subroutine STIFF
The subroutine assembles the stiffness matrix of the grillage in half
band-ividth form using the element stiffness matrix terms previously
calculated.
4. Subroutines SORT and MODSTIF
These subroutines sort and modify the stiffness matrix by eliminating
rows and columns corresponding to restrained degrees of freedom. If
the support is flexible, this subroutine adds the bearing stiffness to the
LI
appropriate stiffness, matrix terms.
-.
5. : Subroutine DECOMP
This subroutine decomposes the stiffness matrix into upper and lower
triangular matrices, using Choleskey's [I] Factorisation method.
6. Subroutine IMPACT
The subroutine calculates impact factor according to the type of live
loading and spans as per provisions of relevant IRC Code of Practices [6].
7. Subroutine LLOAD
For the type of live load specified, this subroutine identifies the total
number of wheels and number of wheels per axle. Using the coordi-
nates of left front wheel (given as input), it also calculates the coor-
dinates of all other wheels. Following eight types of live load cases
(LDING) are built-in into this subroutine:
LDING(2) = Class A Two Lane
LDING(3) = Class 701i. Train: Column '1' Wheel Configuration
LDING(4) = Class 7CR Train: Column `rd Wheel Configuration
LDING(5) = Class 70R Bogie: Column T Wheel Configuration
LDING(6) = Class 70R Bogie: Column 'm' Wheel Configuration
LDING(7) = Class 70R Track
LDING(8) = Class A Single Lane
LDING(9) = Specified by User
. 1 1 1 1 1 1 . m w . - - s - - - - . : . . . _ , -
162 . Grillage Analogy in Bridge Deck Analysis
In load case 9, the *user specifies the details of loading, namely, the
number of wheels and their coordinates. The wheels are numbered
beginning from left most wheel of leading axle, successive wheels
on the same axle in the positive direction of Y, then the wheels on
next following axle and so on.
8. Subroutine WLOAD
This subroUtine describes magnitudes of load on all wheels for each
type of live loading. The minimum alliArable distance of center of
left most wheel from Kerb is set in. All the wheels are taken as point
loads acting at the centers of wheel contact- area. Class 70R Track
loading is described as ten point loads per track, the load on each
point being 3.5 tonnes at spacing of 457 mm as shown in Fig. 1.13.
For User's Specified loading, the' oad on each wheel inclusive of
impact and minimum kerb distance from center of left most front
wheel are given as input.
9 . Subroutines PAIDRB and PAID1
When grid type is skew and the transversals are parallel to the supports
forming parallelogram panels or when the grid is right i.e. skew angle is
zero, the subroutine PAIDRB is called. On the other hand, when the
skew angle is not zero but the transversals are orthogonal to the
longitudinals, the program subroutine PAID1 is called. These subrou-
tines use the spacing between longitudinal grid lines and the lengths of
elements taken on longitudinals, to calculate the boundaries of the
panels with respect to node 1 (specified earlier) and then compare the
coordinates of wheel load with these panel boundaries to identify the
panels containing the wheels. These subroutines are further discussed
below.
a) Subroutine PAIDRB
For each wheel, the Y-coordinate, measured from the origin i.e. from
the left most longitudinal grid line, is checked. If the Y-coordinate is
less than minimum permissible distance, defined by the distance
(Akerb+SCLMIN), the subroutine terminates with position identifica-
tion character as YL. Also, the subroutine terminates if the Y-coordi-
nate exceeds Y mar (Table 5.1) with position identification character
YG. Similarly, if the X-coordinate is such that the wheel is not on the
span, the subroutine terminates with position identification character
as XL or XG as the case be. The panel is now identified as 'P.' i.e.
rectangular or parallelogram. By using the coordinates of wheels, the
panel containing the wheel is identified by its contiguous nodes MNI,
MN2, MN1 1, MN22 as"shown in Fig. 6.4(a, b). The dimensions of

the panel and the position of the wheel on the panel are calculated and
stored. .
b) Subroutine PAID I
The checks for X and Y coordinates of wheel are as in the case of the
subroutine PAIDRB. If the wheel is on the span with its Y coordinate
within allowable limits, then using the coordinates, the panel contain-
ing the wheel is identified as rectangular or parallelogram (`R') or
Triangular (T), the triangle being always a right angled one. The
contiguous nodes are stored as MN1, MN2, MN11 and MN22 for
rectangular and parallelogram panels and as MN1, MN2, MN22 for
triangulir panel (Fig. 6.4).
MN22
y MN11
MN1 MN11 MN1
(a) Rectangular Ranel (b)Parattetogram

Panel
MN2 MN22 MN1.
MN1 MN 22 MN2
(c) Lower Triangular (d) Upper Triangular

Panel Panel
Figure 6.4 Panel Identification in Subroutines `PAIDRB' and `PAIDIP
The dimensions of the panel and the position of the wheel on the panel
with respect to MN1 are calculated and stored. Distinction between
lower and upper triangular panels as shown in Fig. 6.4(c, d) are also
made using position identification characters as 'XG' or 'XL'. If the load
is out of the permissible range in Y-direction (nearer to Kerbs than
permissible), the program stops. If a particular wheel - load is
longitudinally out of span then it is ignored.
10.: Subroutines LDISTR and LDISTT
These subroutines distribute the load from the panel onto the nodes in
the form of vertical loads and moments in X and Y directions. The
subroutine LDISTR is used for distributing the load to the nodes of
rectangular as well as parallelogram panels while the. subroutine
LDISTT distributes the load to the nodes of triangular panels. The
final expressions for nodal load vectors for rectangular and parallelo-
gram panels are given by equations 5.10 to 5.13 and 5.23 to 5.26
respectively and for triangular panels are given by equations 5.16 to
5.21. These have been used in the program with appropriate signs.
11. Subroutine SOLVE
This subroutine uses the previously decomposed stiffness matrix and
the load vector to solve the set of simultaneous equations by backward
and forward substitution procedure and gives the deformations corre-
sponding to three degrees of freedom. If the support type is rigid, the
deformation vector is 're-inflated' by zeroes at the support nodes. This
subroutine is called directly if the supports are flexible. In case of rigid
supports, the load vector is first reduced in size by eliminating elements
corresponding to degrees of restraints at supported nodes and then the
subroutine SOLVE is called.
12. Subroutines COMPARE1 and COMPARE2
These subroutines are used to produce the output for the maximum
live load design responses.
a) Subroutine COMPAREI
The deflections, slopes, moments, shears and torsions at the nodes and
the reactions at the support nodes are compared for all placings of the
same live loading and the critical values of these force responses are
retained along with the corresponding load positions. Along with a
critical response, wherever relevant, the corresponding values of other
related responses, are also retained. For example, along with maximum
shear force corresponding torsional moment, along with maximum
torsional moment corresponding shear forcc,
minimum reaction corresponding X and Y rotations are also stored.

All the live loading types are exhausted in a similar manner one by
one.
b) Subroutine COMPARE2
The critical responses under each type of live loading are compared
amongst themselves and the most critical amongst these response
values are also stored along with type of loading and load position.
13. Subroutine WRITE1
This subroutine is called to print the maximum deformations and force
responses for each load case at various nodes or for various members
identified by COMPARE1.
14. Subroutine WRITE2
This subroutine is called to pint the final envelope values of deforma-
tions and force responses at the nodes or for the members as the case
be, along with type of live loading and its position on the deck. This
subroutine is also called to print the dead load force responses.
6.4 USER MANUAL FOR 'GABS'

The user manual of the program comprises of the details of Data Input and
Result Output. The final output obtained helps in preparing Envelope
diagrams for different deformations and force responses required in the
design of the bridge.
6.4.1 Data Input Module

The input data -is in tonne-mm unit in the program but. can be in any set
of compatible units. The input details are in Free Format. The input data
consists of the details of the structure e.g. skew angle, grillage layout,
member designations, properties of grillage members, support conditions
etc.; dead load values on each longitudinal grid lines; live load cases and
controls and details regarding the movement of vehicular live load on the
deck, etc. The description of READ STATEMENTS (data input) is given
below in the sequence in which they are needed in the program.
1. Reads a character CH1 which initiates the program. Input 'Y' starts
the program while 'N' stops it.
2. Reads type of grid chosen (GRIDTYPE). Grid type can either be
`SKEW' (any value other than 90) or 'ORTH' (90). The 'SKEW' grid
is designated as where the transversal are not normal to the
longitudinal but are parallel to the supports. An 'ORTH' i.e. or-
.. thogonal (90) grid has its transversal normal to the longitudinal.
a check. This is followed by the analysis results. The analysis results consists
of force responses and deformation responses.
The following is the sequence in which the output result is obtained:
I. Grid details as per input data after suitably titling these.
2. Dead load details as per input data.
3. A list of different types of live loads for which the analysis can be.
done and reference numbers of these live loads.
4. Responses under dead load as follows:
(a) Shear forces and torsional moments in each beam element.
(b) Sagging bending moment at each node on its right and on its left.
(c) Hogging bending moment at each node on its right and on its left.
(d) Rotations about X and Y axes and vertical deflections at each node.
(e) Reactions, rotations about X and Y axes and vertical deflections at
each support. Deflections at supports will occur when grid is
supported on yielding supports.
5. Responses under individual live load together with impact load and
dead load as follows:
(a) Span for impact factor calculation and impact factor taken for the live
load.
(b) Type of live load through its reference number, initial and final posi-
tions of reference wheel through its X and Y coordinates and step
lengths in X direction and Y direction respectively. Reference wheel is
the left most wheel on leading axle of the loading system.
(c) Shear force and torsional moment for each beam element
(i) maximum shear force, corresponding torsional moment and posi-
tion of reference wheel;
(ii) maximum torsional moment, corresponding shear force and
position of reference wheel;
(d) Maximum sagging bending moment at each node:
(i) at right of the node and corresponding position of reference
wheel;
(ii) at left of the node and corresponding position of reference wheel;
(e) Same as at (d) above but for maximum hogging bending moment.
(f) Maximum rotations about X and Y axes and maximum deflection at
each node.
rom.ar.tinri :et earth cttrwsnrt with Correspondingdeflection,
rotations and position of reference wheel.

Computer Program - 169
(h) Same as at (g) above but for minimum reaction.

(1) Maximum X-rotation at each support with corresponding
reaction, deflection, Y-rotation and position of reference wheel. (j)
Same as at (i) above but for maximum Y-rotation.
6. Envelope values based on the above live load cases together with
dead load and corresponding impact loading. Overall envelope valueS
of each of the above responses together with the corresponding live
load and location of reference wheel.
Thus complete information, regarding the type of load system and the
location of wheels on the deck giving the critical or design values.are
known to the designer.
6.5 LIMITATIONS AND SCOPE

Efforts have been made to keep the program as simple and general as
possible without Inaldnpit unduly lengthy and complicated but there is a
wide scope for the improvement, additions and modifications.
Changes are possible in regard to loading standards, movement of vehicu-
lar loads on the deck, output and data generation, etc. The automatic gen-
eration of grillage mesh. and the evaluation of elastic properties of grid
members could also have been in-built in the program as a subroutine instead
of doing the same manually as at present. But this has been done with the
motive of educating the designer and allow him more flexibility in choosing
grid lines depending upon the situations.
The program has been primarily developed for Indian Standard Loadings
specified by IRC. However, any type of loading can easily be incorporated in
the program by modifying the relevant subroutines IMPACT, WLOAD and
LLOAD. A minimal change in the main segment of the program will be
needed in this process.
The program can handle bridges having rigid and yielding simple.sup-
ports. This can be extended to continuous supports also. Curved bridges
have not been included in the program but the method of grillage analogy
is capable of handling such bridges with a reasonable accuracy and a
separate computer program can be written for the same. The analysis of
curved bridges, if incorporated in the program 'GABS' will make it too
lengthy and cumbersome.
REFERENCES
1. GERE. M.J. and WEAveR, W., "Matrix Analysis of Framed Structure', D. Van
Nostrand Co., USA, 1965.
2. NAYAR, K.K., RAGHUPAIHI, M., SEETHARAMULU, K. and &RANA, C.S.,
"Computer Aided Bridge AnalysisA Software Based on Grillage Analogy",
International Conference on Bridges and Flyovers, Hyderabad, February 1991.
3. SHARMA, K.G., &R ANA, C.S. and AGRAWAL, A.K., "Automated Analysis of
T-Beam Bridges Using Grid Method", Proceedings of Second International Con-
ference on Computer Aided Analysis and Design in Civil Engineering, U.O.R.,
Roorkee, January 1985.
4. "Ontario Highway Bridge Design Code", Ministry of Transportation and Com-
munications, Highway Engineering Design, Downsview, Ontario, Canada, 1983.
5. "Standard Specifications for Highway Bridges", American Association of State
Highway and Transportatiion Officials (AASHTO), Washington, D.C., 1977.
6. IRC 6-1987, "Standard Specifications and Code of Practice for Road Bridges -
Loads and Stresses", Section II, The Indian Roads Congress, New Delhi, 1990.
Chapter 7
Illustrative Examples
7.1 INTRODUCTION
This Chapter aims at illustrating the use of the grillage analogy method
discussed in earlier chapters through worked examples. A set of six examples
have been chosen covering different types oDbridge superstructures.
Converting an actual bridge deck to an equivalent grid is not without
pitfalls and needs special attention. This conversion involves setting of longi-
tudinal and transverse grid lines and assigning flexural and torsional inertia
values to various grid members. This has been done for all the bridge
examples presented in this Chapter along with brief discussions. It is ex-
pected that these examples and the detailed recommendations given in
Chapter 4 together, will enable a prospective user of the grillage method to
convert a bridge superstructure to an equivalent grid with requisite accuracy
and confidence.
The computer program 'GABS' discussed in Chapter 6 is a versatile
program. These examples are used to explain the use of the aboVe
program when the loading is IRC loading or a user-specified loading or a
stationary loading.
A user of the program at some stage would like to know the sensitivity
of a solution to the size of the grid mesh, i.e. how will the response result
alter, if instead of a normal sized mesh, as recommended in Chapter 4, a
finer or coarser sized mesh is adopted. Example 1 of a right slab bridge is
devoted to investigate this.
End-reactions are non-uniform for a skew bridge even under a
uniformly distributed loading on the deck, such as 'self load'. Example 2
is of a skew slab bridge and is chosen to illustrate this variation, when the
slab is supported on bearings either rigid or flexible.
Computation of Inertia for a voided slab is not straight forward.
Example 3 of a voided slab bridge illustrates this computatiori, as well as
the handling of 'non-uniform dead loading' along the longitudinal grid
In Example 4, a T-beam bridge has been analysed for dead load as .well as
live load specified by the Ontario Highway Bridge Design (OHBD) Code of
Canada.
In example 5, a skew T-beam bridge is analysed for different IRC
loadings and the envelope diagrams. for B.M and S.F are presented.
A box-girder bridge develops torsional and distorsional Warping under
loading. The bending stresses are also affected by shear lag. All these effects
are ignored in the program "GAB" presented in this book.
In example 6, a box-girder bridge has been analysed by 'GABS' as
well as by Finite Element Method and the results are compared.
72 ILLUSTRATIVE' EXAMPLES
Examples 1, 2 and 3 pertain to a right slab, skew slab and a voided slab
bridge deck respectively. Analysis,of slab-on-girders bridges both right and
skew have been illustrated in examples 4 and 5 whereas example 6 dis-
cusses a box-girder bridge.
All the bridges are simply supported and have 7500 mm carriageway
width. The bridges are of reinforced concrete. The Young's modulus of
elasticity and Shear modulus of elasticity are taken as 2.0 and 0.87 tlmm2,
respectively. Wherever dead loads are considered, density of reinforced
concrete including density of wearing coat is taken as 2.40 dm'. In the
input data, the loading type is designated by a number which refers to the
type as given. in Section 5.5.3 of Chapter 5.
COmplete Input and Output files of all the six examples of.this Chapter
and one example of Chapter 3, with the two programs 'GABS' and
'GRID!, (in a ready to use form) is available on a diskette. The diskette
can be ordered from the authors through the publisher.
7.2.1 Example 1: Right Slab Bridge
A two-lane right slab bridge (Fig. 7.1) is chosen for this example. The
equivalant grid is shown in Fig. 7.2 and is referred to as normal mesh in
further discussions. It consists of seven longitudinal and seven transverse
grid lines. Recommendations given in Chapter 4 are followed in arriving at
the grid pattern. The bridge is analysed for four different types of IRC live
loadings alongwith corresponding impact factors. The IRC live loadings
chosen are (i) Class A-2 lane loading, (ii) Class 70R Train Col. loading,
(iii) Class 70R Bogie Col. '1' loading and (iv) Class 70R Track loading.
These loadings are moved on the bridge one at a time in suitably chosen
intervaIsboth longitudinally as well as transversely so that the load
traverses the entire length and width of the deck. For this example, an
o b .
Illustrative Examples 173
P
Wearing Coat .4600
470lieroo
0
7500
L
300
9000
300
Figure 7.1 Right Slab Bridge
interval of 500 mm has been chosen for longitudinal movement of all types
of loadings. In transverse direction, the intervals are so chosen that the load
traverses the full deck width in 5 or 6 steps.
The input data is given in Table 7.1 and selective output is given in Table
7.2. It may be noted that maximum longitudinal bending moment is obtained
under Class 70R Track loading. The maximum bearing reaction'per unit
width for exterior grid lines is under ClassA.-2 Jane loading and for
utidr.-.1 701-: Train Col. T loading. This shows that for
te r
700
_0
7 000
7500
H 60 1350 =8100
300
42 49 35N,
7
fi 13 20 27 34 41 - he
5 12 19 26 33 40 47 0 o
co
%0
IL
11 18 25 32 39 46 ti+
3 10 17 24 31 38 45
2 9 16 23 30 37 44
1 8 15 22 29 36 43
a a
Inner Transversal Inner Longitudinal
1-.--1500 -1 T I = 4.29 E 10 1- E101-

3.86
700
_t_ .1: 8.58E10 ___________________ j 700
J=7.72 E10
I 1350 '-I
Outer Outer
Transversal
N-10504
T Longitud
T 1000 inal '1

700 I = 3.00 E10 I 600 F.
.1: 6.00 E10 T 6.48E10
750'1
300 700 3,12.97E10
's Indicate Locations of L-4- 6751
arid lines 300
Figure 7.2 Normal Grid Mesh and Sections of Various Grid Members
f a particular bridge any one particular type of IRC live loading does not give
all the critical responses.
TABLE 7.1: Input DataRight Slab BridgeNormal Mesh Size
300 300
1 2 12 1500 0 6.48E+10 -12.97E+10
2 3 3 4 4 5 5 6 6 7
43 44 44 45 45 46 46 47 47 48 48 49
8 9 3 0
1500 0 3.86E+10 7.72E+10
9 10 10,
11 II 12 12 13 13 14
15 16 16
17 17 18 .18 19 19 20 20 21
22 23 23
24 24 25 25 26 26 27 27 28
29 30 30
31 31 32 32 33 33 34 34 35
36 37 37
38 38 39 39 40 40 41 41 42
1 8 1 2
0 1350 3_0E+10 6.0E+10
8 15 15
22 22 29 29 36 36 43
7 14 14
2 9 30 21 21 28 28 35 35 42 42 49
9 16 16 0 1350 4.29E+10 8.58E+10
3 10 10 23 23 30 30 37 37 44
4 11 11 17 17 24 24 31 31 38 38 45
5 12 12 18 18 25 25 32 32 39 39 46
6 13 13 19 19 26 26 33 33 40 40 47
14 3 1 20 20 27 27 34 34 41 41 48
300.
1 7 8 14 15 21 22 28 29 35 36 42 43 49
0 0
0
1 0 0 0 0 0 0 0 0 0 0 0
1
0
0
4
2
Y Y
.
90
0
84 49 4; 2.0 0.87 7 1350 1350 1350 1350 1350 1350
0 700. 500. 280. 27800. 2100.

1
0
0 .1705. 500. 385. 22400. 4015.
Contd.
. .`
0
5
0 1705. 500. 385. 10220. 4015.
1
0
7
0 1920. 500. 440. 13113. 4120.
1
0
1
0
S
This example is further used to study the effect of size of the mesh
formed by grid lines on various force responses. Figure 7.3 shows a coarser
mesh for the same bridge where the number of longitudinal grid lines have
been reduced from 7 to 5 but the number of transverse grid lines have been
kept the same. Figure 7.4 shows a finer mesh for the same bridge where the
number of transverse grid lines have been increased from 7 to 11 but the
number of longitudinal grid lines are kept same as in normal sized mesh of
Fig. 7.2. Both these grids are analysed for the same four types of IRC live
loadings as above, keeping the longitudinal and transverse intervals for
various IRC loadings same as in the analysis of grid of Fig. 7.2. The input
TABLE 7.2: Effect of Mesh Size on Force Responses

(a) Maximum longitudinal bending moments and
corresponding bending compressive stresses
Reference Exterior Grid line Central Grid Line
Grid Load B.M. Comp Mean a Load B.M. Comp. Mean a
Type (t m) Sress a (N/mm2) Type (t m) Stress a (N/mm2)
(N/mm2) (N/mm2)
Normal 7 29.3 2.47 7 19.3 1.75
Coarse 7 34.1 233 2.48 7 28.9 1.75 1.73
Fine 7 29.0 2.44 7 18.8 1.70
(b) Maximum bearing reactions and corresponding bearing reactions per unit width
Reference Exterior Grid Line Central Grid Line
Grid Load Reaction R per unit Mean Load Reaction R per unit Mean
Type R width (r) r Type R width (r) r
(t) (t/m) (t/m) (t) (t/m) (t/m)
Normal 2 7.23 7.42 3 17.50 13.01
Coarse 2 9.59 7.30 7.41 3 26.02 12.tb i2.91
Fine 2 7.32 7.51 3 17.38 12.87
4.6001.r-
700 1000
$
4 @ 2025 8100
1 .,,,7 s' v Lc, V J aS
-r-
13 20. 27 . 34
6
12 : 19 26 33
5
0
0
0
4 11 18 " 25 32 171
10 17 "24 31 eU
0
3
9 16 23 30
2
1 8 15 22 29
A A A A P
Inner Longitudinal Outer Longitudina L
1=5.79 E10
2025 --a`Lt. :FL ___ 1700

700
00
i o _t 1'
-i-
1 =11.56 E10
-7- 1=7.65 E10
( Nate: Inner '4113115 FT J :15.30 E 10
Same as in 04
k- 300
Figure 7.3 CoarseTransversal
Grid Mesh and
andSections
Outer of Various Gridare
Transversal Members
Fig.
7.2 )
^-*1 600h
........ .
700

6e 1 35 0 = 81 0 0
11 v 22 33 44 a 55 v 66a 77
10 21 32 43 54 65 76
.
9 20 31 42 53 64 75
19 30 41 52 63 74
70 X 900..9000
7 18 29 40 51 62 73
6 17 28 39 50 61 72
5 16 27 38 49 60 71
15 26 37 48 59 70
3 14 25 36 47 58 69
2. 13 24 35 46 57 68
1 12 23 34 45 56 67
a
Inner Transversal Outer Transversal
-
1 = 2. 5 7 E 10 . 4 7 50 1. - - T 1=214E10
700 Jf. 4 29E10
70 0 J = 5 1 5 E 1 0
900 k --pl.
3001.r--
( Note: Inner longitudinals and Outer Longitudinals are

Same as in Fig 7.2 /
Figure 7.4 Fine Grid Mesh and Sections of Various Grid Members
_______ 4
1151.111.11."m--
l - ..yilt7,-

data have not been given here for brevity. The governing loadings for
maximum longitudinal bending stresses and bearing reactions were found
to be the same as for normal sized mesh as shown in Table 7.2. The
maximum bending compressive stress remained within 2% of its mean
value and the maximum bearing reactions per unit width remained within
1.5% of its mean value. This' shows that some variation in fineness or
coarseness in mesh pattern can be adopted, if desired,, without affecting the
accuracy in any significant manner.
7.2.2 Example 2: Skew Slab Bridge

A skew slab bridge having 300 skew angle supported on five isolated
bearings at each end is analysed in this example. The example refers to an
actual bridge constructed on NH 8 at KM 247/40 near Ajmer, Rajasthan.
The bridge was analysed and designed by ,the first author.
The cross-section of the bridge, the. positions of the bearings and grid
chosen for the analysis are shown in -Fig. 7.5. Preliminary analysis lead to
the adoption of neoprene bearings of spring stiffness of 40 t/mm at each
support point. The flexibility of the support is considered in the analysis.
245
10490
7
0 9 24 5
(a) PLan at mid-thickness of Slab and EquIvaLent grid
1775 'a 7500 *linqi

275
., :,, ,rw.c. 75(k.a)
\I
1?50
610 --4.1 r 1-1, rw- t 1 610
k--
535 4x2395= 9580 535
/mot
.! :.1.1!t octton
Figure 7.5 Skew Slab Bridge Supported on Isolated Bearings

Table 7.3 gives sectional'widths adopted in computing I and J values for
grid members of various groups. Dead load is assumed to be acting along
longitudinal grid lines. The dead load along the two exterior longitudinal
grid lines inclusive of weight of railing @ 0.15 Um is found to be 4.0 t/rn
each. The dead load along interior longitudinal grid lines is taken as 4.74
Tim each.
The types of liVe load considered on the carriageway are the sarn4 as in
example I, namely, Class A-2 lane loading, Class 70R Train Col. '1' load-
ing, Class 70R Bogie Col. '1' loading and Class 70R Track loading.
Table 7.4 presents grid input data assuming flexible bearings at
supports. The bridge is also alternately analysed for rigid bearings for the
sake of comparison.
For brevity, only bearing reactions are presented and discussed. Bearing
reactions under dead load are given in Fig. 7.6. Also the reactions are
tabulated under dead load and under dead and live load combine in Table
7.5. The curve of Fig..-7.6 of bearing reactions under dead load for rigid
bearings gives the shape of a saddle. The reaction is very large at the obtuse
angled corner. This is followed by a smaller reaction at the second bearing.
The curve then rises again and finally diminishes at the acute angled corner.
The curve corresponding to flexible bearings has more even distribution of
reactions. The same trend is followed by bearing reactions under dead and
live load combine. The minimum bearing reactions under dead and live load
combine, at times, is smaller than its value under dead load alone. This
indicates that for certain positions of live load on the deck, the bearing
50
Q..
. 40
Rigid Bearings
oc 30
Flexible
20
cc 10
0 9 18 27 36 45
Bearing Node Number
Figure 7.6 Bearing Reactions Under Dead Load

MOM
reaction due to live load alone is negative. However, special attention is

required to be paid if the minimum bearing reaction under dead and live
load combine becomes negative. This will indicate an uplift at the bearing.
TABLE 7.3: Widths Adopted in Computing I and J Values for. Grid

Members of Various Groups
Group Typical Width '6' Remarks
Members (mm)
1&2 1-2, 5-6 1807 Along exterior longitudinal grid lines
3&4 I0-H, 13-14 2395 Along interior longitudinal grid lines
5 1-10 245 Along skew transverse lines
6 . 6-14 1645 Along interior transverse lines
7 2-10 1383 Along exterior transverse lines, not
covering full grid width
8 5-13 1514 Along exterior transverse lines covering
full grid width
Notes: 1. Constant thickness of 750 mm is assumed for all members. T hus

flexural mo me nt of inertial I for a me mber of width ` If mm is (750)3 b/12 nun'. 2.
Torsional Inertia J of the me mber is taken as t wice its flexural mo ment of
Inertia, I.
TABLE 7.4: Input Data-Skew Slab Bridge on Flexible Bearings

Y
90
30
80 45 8 2:0 .87 5 2395. 2395. 2395. 2395.
285 285
1 2 8 4382.5 0. 6.35E+10 12.71E+10
2 3 3 4 4 5 41 42 42 43 43 44 44 45
5 6 8 1645. 0. 6.35E+10 12.71E+10
6 7 7 8 8 9 37 38 38 39 39 40 40 41
10 11 12 1382.5 0. 8.42E+10 16.84E+10
II 12 12 13 17 18 19 20 20 21 25 26 26 27 28 29
33 34 34 35 35 36
13 14 12 1645. 0. 8.42E+10 16.84E+10
14 15 15 16 16 17 21 22 22 23 23 24 24 25
29 30 30 31 31 32 32 33
1 10 8 1382.5 2395. 0.86E+10 1.72E+10
9 18 10 19 18 27 19 28 27 36 28 37 36 45
6 14 12 0. 2395. 5.78E+10 11.57E+10
7 15 8 16 14 22 15 23 16 24 22 30 23 31 24 1/ 30
4 31 39 32 40 JV
Contd.
2 10 12 0. 2395. 4.86E+10 9.72E+10
3 11 4 12 11 19 12 20 18 26 20 28 26 34 27 35 34 42
35 43 36 44
5 13 8 0. 2395. 5.32E+10 10.64E+10
9 17 13 21 17 25 21 29 25 33 29 37 33 41
10 1 1 9 10 18 19 - 27 28 36 37 45
40
1040.
1
4.00E-3 .967 4.74E-3 0 4.74E-3 0 4.74E-3 0 4.00E-3 .967
0
1
1
1
0
4 L
2
0 1440 500 280 36500 2840
1
0
3
0 2445 500 385 31000 4755
1
0
5
0 2445 500 385 19000 4755
1.
0
7
0 2660 500 440. 22000 4860
1
0
S
TABLE 7.5: Bearing Reactions Under Dead Load and Dead & Live Loads
Combine
Bearing Dead Load Alone Dead and Live Loads Combine
Node Reaction (t) Min. Reaction (t) Max. Reaction (t)
Flexible Rigid Flexible Rigid Flexible Rigid
Bearing Bearing Bearing Bearing Bearing Bearing
9 or 37 44 51 44 49 68 74
18 or 28 31 23 31 22 55 55
27 or 19 28 29 28 29 52 61
36 or 10 24 27 24 27 42 47
45 or 1 13 11 11 9 20 16
7.2.3 Example 3: Voided Slab Bridge
A voided slab bridge is chosen for this example. The voids are running
through the central part of the span. In support regions where shear is
large, solid section has been provided. Figure 7.7 shows the bridge details
and the proposed grid arrangement is shown in Fig. 7.8.
The example is chosen to explain (a) how inertia values are to be com-
puted for such voided sections and (b) how non-uniform dead load along
the longitudinal grid line is to be dealt with.
The inertia values are computed on the principle that slab has same
Liu8 , 8
supportest 4
1 Mis.mmiM
14441 .
I l
r - , ._,
in Fin r : !---. 7 '
ri
I 1 1 I 1 iI I1 11
il I
II 1 1 I

I I 1 I II
I I I I II : I I
III I
1ii I I I N.
1I I II :I
C
I Ig 0 -
iI1
I , , I I 1 ec. Q
1
_11.11i r 1 8
iilriii Hi!!I I N,
it
1 11 1;111 i I
I II Ii ' 1 I 1 1 II I
I i. I II I Ill!! 1 I 1 1 I
LJLJLJULJLJUII I ..
co o '
0
90
0
7500 9C11,_ISuf:port
_
F()11(3 01 0 010 0 0
H- 7 x 960 u 6720 .4 1-.
480 480

90 1830 1920 1920 1830 90
Figure 7.7 Voided Slab Bridge
7
V
v5
14 21 28 6
13 20 27 3
0 0
0 N
- 2 11
5 0
12 19 26
0
N
(0
3
4
11 18 25 .
30
1
10 17 24
9 16 23
8 15 22
1 A 2 9
i+-1830-441 2 x 19 2 0 1830-44
= 3840
. M . 1 1 1 1 . Figure 7.8 Grid Relating to Voided Slab Bridge
inertia per unit width in the two mutually perpendicular directions. This is
true only if void depth is upto 60 per cent of the overall depth. For voided
section, the moment of inertia per unit width of the slab is computed from
the section showing the void/voids and the same is assumed per unit width
in the orthogonal direction. Torsional inertia .1 of the section is computed
from moment of inertia 1 adopting the formula .1 = 21 for all slab sections.
Table 7.6 gives the sections and their I values. Moment of Inertia values of
members belonging to groups 3 to 10 have been calculated on the basis of I
per unit width (i) of the members of groups I and 2.
The dead load alone the longitudinal grid line is non-uniform as in the end
regions the section consists of solid slab and in the central region the section
consists of voided slab. Such non-uniform loading is dealt in by
1..-:-,,''''''"Ff7e-.72.14.-,1=.0M-;.1';'.:'-,
i,---7411111111.11c12

TABLE 7.6: Sections and Moment of Inertias
Section
x 10 - 4
(D = IOW)
--I 160
800 '(i = 0.8333)
--04 960
p 736 (i 17...7671)
.8333 x 1920 = 1600
1 9 2 0
.7671 x 1920 = 1473
1600 kg 6001
1400
.8333 x 1400 = 1167
I.- 2000
(.8333 + .7671) 1000
66001 0 = 1600
_L
1.10001.-10001
2000 -.4
6110 .7671 x 2000 = 1534
Group Member Typical

Location member

1 Exterior long. 1-2

- grid line
2 do-- 2-3
3 Interior long. 8-9
grid line
4 9-10
5,6 Support 1-8

transverse 8-15
grid line
7,8 First Interior 2-9

transverse 9-16
grid line
9,10 Interior 3-10

transverse 10-17
grid lines
co mp uting gravity lo ad s at the vario us nodes based o n the load o f the

tributary area or contributary area of each node. Table 7.7 gives details for
r-trin +1.Itn e. T L - - ,y;ven in
the same sequence in which node numbering is done.
AIM

TABLE 7.7: Computation of Nodal Loads
Typical Section* Area Length Load
Node (m2) (m)
1 As for member 0.960 1.40 3.23
1-2
2 Mean.of members 0.824 2.0 3.93
1-2 and 2-3
3 As for member 0.677 2.0 3.25 1
2-3
8 As for member 1.92 1.4 6.45
8-9
9 Mean of members 1.648 2.0 7.86
8-9- and 9-10
10 As for member 1.354 2.0 6.50
9-10
* Refer Table 7.6 for Sections.
Table 7.8 gives the input data required for dead load analysis. Data
relating to live load analysis is not given for brevity though it can be given in
the same manner as for earlier examples.
TABLE. 7.8: Input Data-Voided Slab Bridge

Y
90
0
58 35 10 2.0 .87 5 1830. 1920. 1920. 1830.
400. 400.
1 2 4 2000. 0. 8.0+E10 16.0+E10
6 7 29 30 34 35
2 3 8 2000. 0. 7.36E+10 14.72E+10
3 4 4 5 5 6 30 31 31 32 32 33 33 34
8 9 6 2000. 0. 16.E+10 32.E+10
13 14 15 16 20 21 22 23 27 28
9 10 12 2000. 0. 14.73E+10 29.46E+10
10 11 11 12 12 13 16 17 17 18
IS 19 19 20 23 24 24 25 25 26 26 27
1 8 4 0. 1830. 11.67E+10 23.33E+10
7 14 22 29 28 35
8 15 4 0. 1920. 11.67E+10 23.33E+10
14 21 15 22 21 28
9 4 0. 1830. 16.E+10 32.E+10
Contd.
6 13 23 30 27 34
9 16 4 0. 1920. 16.E+10 32.E+10
13 20 16 23 20 27
3 10 6 0. 1830. 15.34E+10 30.68E+10
4 11 5 12 24 31 25 32 26 33
10 17 6 0. 1920. 15.34E+10 30.68E+10
11 18 12 19 17 24 18 25 19 26
10 3 1 7 8 14 15 21 22 28 29 35
0
100
3.23 3.93. 3.25 3.25 3.25 3.93 3.23
6.45 7.86 6.50 6.50 6.50 - 7.86 6.45
6.45 7.86 6.50 6.50 6.50 7.86 6.45
6.45 7.86 6.50 6.50 6.50 7.86 6.45
3.23 3.93 3.25 3.25 3.25 3.93 3.23
1
1
7.2.4 Example 4: Right T-Beam Bridge

A right T-beam bridge having three longitudinals and five transverse beams
as shown in Fig. 7.9, has been analysed. in this example.
The beams will have T-sections because of monolithic construction. The
flange widths is computed following IRC 21-1987 Clause 305.12.2[3]. The
three longitudinal beams between them will cover the entire deck width as
the flange width with the middle beam covering central 3 m width as its
flange. To apply the above clause for assigning the flange widths of
transverse beams, their effective spans are required to be estimated. The
effective span length may be defined as the distance between points of zero
bending moment and its evaluation initially is not straight forward. It is easy
to visualize that under 'loading, an interior transverse beam can develop sag-
ging bending moment at a section where it crosses the central longitudinal
beam. The longitudinal beams usually have small torsional inertia and hence
transverse beams at sections where they end into exterior longitudinal beams,
will develop small bending moments only. It is, therefore, suggested that the
span length of the interior transverse beams for applying the above IRC
Clause be taken as centre to centre distance between exterior longitudinal
beams, which in this case is 6 m. The case of end transverse beam is however
different. The end transverse beam under loading can develop
lIm.M10.*
7500
210 70mm Wearing coat
-
,1
150
C1-1-1-*--3000

--
1 1"-
1- 600
r100.3 3000
51
A B C D Ci e t A'
0.11950 -1..s.-- 1. @1500: 6000 ----1.-11050 I-
-(a) Cross- section at Mid-span
5 _______ 5250 1.1

3601 1 <150 x 300 210
I
10500

E G
(b) Longitudinal section
Figure 7.9 Right T-Beam Bridge
hogging bending moment over interior support. The span length of exterior
transverse beam will be less than centre to centre distance between adjacent
bearings. It may be assumed as 0.8 times the distance between centres of
adjacent bearings, that is 2.40 m. The flange widths of individual transverse
beams applying the relevant Clause will be then as shown in Fig. 7.10. We
now proceed to idealize the bridge deck into grid structure.
Let us take transverse grid lines first. Referring to Fig. 7.9 (b), adopt grid
lines along centres of transverse beams. These are then at E, G and J. The
effective 'slab widths covered by these transverse T-beams still leave out
large widths of slabs between them. To take into account these remaining

Illustrative Examples 189 -i-

--0
dm.
O
O
cV
0
1 MO ,
o
O I
+m, OFR
-
0
on
.0-o

(12
Figure 7.10 Plan of Bridge Showing Widths of Deck Slabs

Assigned to Various Transverse Members
slab strips adopt transverse grid lines between them at F and H giving five
transverse grid lines in half length of the bridge or nine transverse grid lines
in full length of the bridge. The widths of the slab strips represented by these
are shown in Fig. 7.10. The assumed locations of all the nine grid lines
are shown in Fig. 7.11. Now, I and T values to these grid members corre-
sponding to their sections are to be assigned.
In longitudinal direction, referring to Fig. 7.9 (a), grid lines along the
centres of three longitudinal beams are adopted. These are at B, D and B'. I
and J values based on corresponding T-sections of these beams are as-
signed to these grid lines. The spacing of longitudinal beams is 3000 mm
N/RiNsam.,_ MUM
18 36
27 45 63
62
61
a
6 0 0
V
II
in
N
59 _ii
50
eV
co
58
57
56
11
10 55
401500 = WOO I
Figure 7.11 Equivalent Grid
which is more than spacing of transverse grid lines which is 2625 MILL But
the spacing of longitudinal grid lines should be less than spacing of trans-
verse grid lines for the width to span plan dimension of this bridge. Hence
two additional grid lines at C and C' between the longitudinals are intro-
duced. The live load wheel while traversing over the carriageway may come
on the cantilever slab i.e. on the left of B or right of B' and hence additional
longitudinal grid lines, one each, say at centres of each of the kerb slab at A
and A' be adopted. The inertia value of entire transverse section is covered by
die loligatudinal beams at B, D and B' and hence the grid lines at locations A,
C, C' and A' may be assigned zero inertia values or alternately
these grid lines could be assigned small nominal inertia values and corre-
spondingly inertia values of beams at B, D and B' be reduced. The former
option is adopted herein. The location of all seven longitudinal grid lines
is .shown in Fig. 7.11.
The bridge is analysed for dead load and live load. The dead load is due
to self load and is assumed to be acting along the seven longitudinal grid
lines as per corresponding tributary widths. The live load is .due; to OHBD
Truck load and is chosen to demonstrate the use of the program! for a user
specified live loading. The OHBD truck loading is used for the design of
highway bridges in Ontario State of Canada and has been taken from
Ontario Highway Bridge Design Code [2]. The impact factor for the loading
is taken as 0.3. The minimum kerb distance for the loading is taken as 600
mm. The truck loading is given in Fig 7.12. The truck is made to traverse
the length and breadth of the entire carriageway. The lengths of the steps in
longitudinal and transverse directions are kept as 1500 mm and 750 mm.
respectively.
The input data is given in Table 7.9. The longitudinal design bending
moment along exterior and interior beams at one-eighth span point along
6 14 14 20 16 Axle load
3 7 710 8 Wheel load
I
(a) Elevation Showing Magnitude of Axle and

Wheel Loads
250 250
-
.
4 .0
I
. 0 0
- 0 0
GO 0 st
2SCRE3 U
7
L
00
Wheel 1200
Reference T
)7/2fC2 ____ M
1- *Ti 81 44 -1
3600 6000 7200 11
i
1- 180Q0 .. I
(b) Plan Showing Dimensions of Truck Loading
Figure 7.12 OHBD Truck Load

with corresponding positions of the truck is given in Table 7.10. The truck
position is indicated through X-Y coordinates of reference wheel, which is
the left most wheel of the leading axle. The X -Y coordinates are given with
respect to node 1 (Fig. 7.11).
TABLE 7.9: Input Data-Right T-Beam Bridge

Y
90
0
110 63 112.0 .87 7 1050. 1500 1500. 1500. 1050.
1500. 500. 500.
1 2 32 2625. 0. 0. 0..
2 3 3 445 5 667 7 8 8 9 19 20 20 21 21 22 22 23 23 24
24 25 25 26 26 27 37 38 38 39 39 40 40 41 41 42 42 43 43 44 44 45 55 56
56 57 57 58 58 59 59 60 60 61 61 62 62 63
10 11 16 2625. 0. 53.93E+10 2.78E+10
11 12 12 13 13 14 14 15 15 16 16 17 17 18 46 47 47 48 48 49
49 50 50 51 51 52 52 53 53 54
28 29 8 2625. O. 45.65E+10 2.78E+10
29 30 30 31 31 32 32 33 33 34 34 35 35 36
10 19 8 0. 1500. 9.94E+10 .91E+10
19 28 28 37 37 46 18 27 27 36 36 45 45 54
11 20 8 0. 1500..32E+10 .65E+10
20 29 29 38 38 47 17 26 26 35 35 44 44 53
12 21 12 0. 1500. 13.77E+10 .97E+10
21 30 30 39 39 48 14 23 23 32 32 41 41 50 16 25 25 34 34 43
43 52
13 22 8 0. 1500. .29E+10 .58E+10
22 31 31 40 40 49 15 24 24 33 33 42 42 51
1 10 4 0. 1050. 1.27E+10 2.25E+10
46 55 9 18 54 63
2 11 4 0. 1050. .89E+10 1.78E+10
47 56 8 17 53 62
3 12 6 0. 1050..32E+10 .64E-F10
48 57 5 14 50 59 7 16 52 61
4 13 4 0. 1050. .80E+10 1.59E+10
49 58 6 15 51 60
6 3 10 18 28 36 46 54
300.
1
5.2E-04 8.79E-02 2.38E-03 9.64E-02 1.01E-03 0. 2.53E-03 0.
1.01E-03 0. 2.38E-03 -9.64E-02 5.2E-04 -8.79E-02
0
1
1
1
0
1
9
0 900 1500 750 39000 5400
10
0 1800 -3600 0 -3600 1800
-4800 '0 -4800 1800 -10800
-10800 1800 -18000 0 -18000 1800
600
3.9 3.9 9.1 9.1 9.1 9.1 13. 13.10.4 10.4
1
0
S
TABLE 7.10: Design Bending Moment Along Longitudinal Reams

Location Exterior Beam Interior Beam
along Bending Coordinates of Bending Coordinates of
span Moment Reference wheel Moment Reference wheel
(t m) (m) m) (m)
0.125L 177 22.5, 0.9 136 22.5, 3.15
0.25L 297 19.5, 0.9 202 19.5, 3.15
0.375L 361 18.0, 0.9 255 18.0, 3.15
0.500L 379 15.0, 0.9 256 15.0, 3.15
7.2.5 Example 5: Skew T-Beam Bridge -

A skew T-beam reinforced concrete two-lane simply supported bridge with
40 skew angle is chosen for this example. The example refers to an actual
bridge constructed on NH-46 in Tamil Nadu at KM 40/2. The bridge was
analysed and designed by the authors for the Ministry of Surface Transport.
The bridge is shown in Fig. 7.13. The longitudinal and transverse beams are
cast monolithically with deck slab.
The longitudinal and transverse beams are taken as T-beams as in example
4. For end diagonal beams, the effective span has been taken as 0.8 times the
skew distance between centres of bearings for assigning the flange
.01MEMil
KERB
3@ 5500- 16500
Skew span =2070
(a) Plan
111 1
11
I-
450
li w
e
k
rs
a
cross beam
60

A0C0
H 1475 hi-441250r 475
Hnoo I, 2500 .1. 2500 1:1700.1
(b) Right cross - section at mid span
Figure 7.13 Skew T-Beam Bridge
Flange width of beam

Figure 7.14
Beams exist along lines and 0 Additional
transverse grid tines are put along tines 0 0 g0
Width of deck stab contributing to I , J Values
to grid tine , say 3,is (b + c )
Transverse Grid Lines in End Region

width to these beams. Seven longitudinal grid lines as in Example 4 are

adopted. The mathematical model of this bridge requires comparatively a
larger number of transverse grid lines. Figure 7.14 shows the transverse grid
lines in the end region and Fig. 7.15 shows the equivalent grillage for the
bridge. A closely spaced grid pattern in the end region will provide more
accurate force responses. This adoption also ensures equal number of nodes
along each longitudinal grid line, which is a pre-requisite of .the program
'GABS'.
The analysis for only IRC live loads has been carried out. The live loads
considered are same as for Example 1. Table 7.11 gives the input data. The
91
1238
_ 30
78
65 77 in00
52 64 76 89 rc
*4'
z
9 51 63 75 88
26 38 50 62 74 87
13 2750
25 37 61 86
85
. 12 _ 84
24 36 48 6072
11 83 0
2 3 3 5 47 59 71 0
0
10 82
22 34 46 58 7 0 II
0
U)
(--
9 81 csa
21 33 45 576 9 x -
4
8 80
X 0 3 2 44 566 8 1512
7
19 31 43 55 1238
6 6 .79
_____y1
18 30 42
17 29
4
3
2
1 0*
(0,0) 1 .
1--,-4x1250= -`.1 1-.-
1475 5000 1475

Figure 7.15 Equivalent Grillage

200
'150
100
.50
co
0
Bending Moment Envelope for Exterior Beam
7 no
100
oi 50
0
Bending Moment Envelope for Central Beam
30
10
0
Shear Force Envelope for Exterior Beam

50
30
In 10
Shear
Shear Force Envelope for Interior Beam
0246 8 10 12 14 16 18 20'
Span (m)
Figure 7.16 Bending Moment and Shear Force Envelopes for Beams
overall envelope values for bending moments and shear forces under live
loadings for exterior and interior beams are presented in Fig. 7.16. In pre-
paring these envelope diagrams the response values at polar symmetric
points have been compared and wherever these were different, the larger
one has been adopted. The shear force envelope diagrams show sudden
change in values at locations of cross beams. The envelope diagrams for
shear forces also show maximum reaction values at the supports. The reac-
tions at the two ends of the exterior beam differ considerably. The reaction
is minimum at the acute angled corner and is maximum at the obtuse
angled corner. This is expected in a skew bridge.


TABLE 7.11: Input Data-Skew T-Beam Bridge
Y
90
40
168 91 26 2.0 .87 7 1475. 1250. 1250. 1250. 1250. 1475.
450. 450.
2 14 2 0. 1475. 0.12E+10 0.24E+10
78 90
3 15.4 0. 1475. 0.11E+10 0.22E+10
4 16 76 88 77 89
5 17 2 0. 1475. 0.10E+10 0.2E+10
75 87
6 18 6 0. 1475. 0.13E+10 0.26E+10
9 21 11 23 69 81 71 83 74 86
(
7 19 4 0. 1475. 0.144E+10 0.288E+10
13 25 73 85 67 79
8 20 4 0. 1475. 0.304E+10 0.607E+10
12 24 68 80 72 84
10 22 2 0..1475. 0.45E+10 0.9E+10
70 82
1 2 8 1238. 0. 0. 0.
6 7 30 31 37 38 54 55 61 62 85 86 90 91
2 3 16 1050. 0. 0. 0.
3 44 5 5 6 86 87 87 88 88 89 89 90 27 28 28 29 29 30
38 39 53 54 62 63 63 64 64 65
7 8 8 1512. Q. 0. 0.
12 13 31 32 36 37 55 56 60 61 79 80 84 85
8 9 16 2750. 0. 0. 0.
9 10 10 11 11 12 32 33 33 34 34 35 35 36 56 57 57 58 58 59
59 60 80 81 81 82 82 83 83 84
14 15 8 1050. 0. 44.90E+10 2.23E+10
15 16 16 17 17 18 74 75 75 76 76 77 77 78
18 19 4 1238. 0. 44.90E+10 2.23E+10
25 26 66 67 73 74
19 20 4 1512. 0. 44.90E+10 2.23E+10
24 25 67 68 72 73
20 21 8 2750. 0. 44.90E+10 2.23E+10
21 22 22 23 23 24 68 69 69 70 70 71 71 72
40 41 4 1050. 0. 38.0E-F10 1.66E+10
41 42 50 51 51 52
42 43 2 1238. 0. 38.0E+10 1.66E+10
49 50
43 44 2 1512. 0. 38.0E+10 1.66E+10
48 49
44 45 4 2750. 0. 38.0E+10 1.66E+10
45 46 46 47 47 48
1 1412 1238. 1475. 3368E+10 7.98E+10
14 27 27 40 40 53 53 66 66 79 13 26 26 39 39 52
52 65 65 78 78 91
15 27 60. 1250A.07E+10 0.14E+10
16 28 28 40 52 64 64 76 65 77
17 29 6 0. 1250. 0.063E+10 0.12E+10
29 41 41 53 39 51 51 63 63 75
18 30 16 0. 1250. 4.64E+10 0.64E+10
30 42 42 54 54 66 26 38 38 50 50 62 62 74 21 33 33 45
45 57 57 69 23 35 35 47 47 59 59 71
19 31 8 0. 1250. 0.091E+10 0.18E+10
31 43 43 55 55 67 25 37 37 49 49 61 61 73
20 32 8 0. 1250. 0.192E+10 0.384E+10
32 44 44 56 56 68 24 36 36 48 48 60 60 72
22 34 4 0. 1250. 0.283E+10 0.567E+10
34 46 46 58 58 70
6 3 14 26 40 52 66 78
225.
1
0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
0
1
1
0
4
2
0 625 1000 280 47000 2025
1
0
3
0 1630 1000 385 42000 3940
1
0
5
0 1630 1000 385 29000 3940
1
0
7
0 1845 1000 440 32000 4045
1
0
1
7.2.6 Example 6: Box-Girder Bridge

A single cell box of trapezoidal cross-section with 30 m span is considered
for this example. The bridge is shown in Fig. 7.17. Converting a bridge of
this type to an equivalent grid and assigning inertia values to various
members needs special care. Referring to Fig. 7.17(b) six longitudinal grid
lines have been assumed; namely two at C and C' in alignment with locations
of centre of bearings, two at B and B' at locations where inclined webs
intersect the deck slab and two at A and A' where the decking slab ends. The
end ones at A and A' are znecessiated because the cantilever projections are
large and wheels of live load could go on the cantilever
3 5 0 0
1. 4176 --oi
(a) Cross Section
2725 1075 3500 . 1075 * 2725

0,
A BC C B'
i
A'
(b) Location of Longitudinot Grid Lines
7
30 )35
0 4- 225
est
30000
50
450 (c) Longitudinal Section
- 5
41 ____________ 8x 3750 c 30000 ________
(d) Location of Transverse Grid Lines
Figure 7.17 Box-Girder Bridge

slabs beyond B and B'. The moment of inertia of the cross section of the
bridge about a common axis is computed and this is divided equally among
the two longitudinals at B and B'. Similarly torsional inertia of the closed
trapezoidal section is computed using equation 4.15 and one half of this is
assigned to each of the longitudinals at B and B'. Thus it is assumed that in
longitudinal direction, the entire inertia is Concentrated along grid lines at B
and B'. The remaining four longitudinals which are located at A, A',
C and C' each are assigned zero inertia values. Referring to Fig. 7.17 (c,
d), nine transverse grid lines as shown have been assumed. The flexural and
torsional inertia values of these transverse members are computed employing
equation 4.12. The end transverse members also have diaphragms and hence
moment of inertia of diaphragms are added while computing moment of
inertia of end transverse members.
The bridge is analysed for a stationery position of Class 70R Col. 'm'
train loading. The equivalent grid and position of live load on it is shown
in Fig. 7.18. Table 7.12 gives the grid input data. The maximum
longitudinal bending moment at mid-span was found to be 350.0 t.m
giving a bending compressive stress of 1.76 MPa and bending tensile
stress of 3.36 MPa.
To compare the results of aboye grillage analysis, the bridge is alter-
nately analysed by Finite Element Method (FEM). Standard package SAP
IV is used for the analysis. Rectangular plate element with six degrees of
freedom at each node is chosen. Boundary elements are introduced at sup-
port locations for support reactions. Figure 7.19 gives the_ structural
idealisation chosen for the FEM analysis. The maximum bending
compressive and tensile stresses at mid span were found to be 1.78 MPa
and 3.20 MPa respectively. The comparison is quite close and acceptable.
For more detailed information on the comparison, readers may refer to
reference [1].
In the above examples, wherever necessary, the sections adopted for
the grillage .elements are given. The inertia values of these sections can be
taken from the respective grid Input data given for each example.
TABLE 7.12: Input DataBox-Girder Bridge
Y
90
0
93 54 8 2.0 0.87 6 2725.0 737.0 4176.0 737.0 2725.0 450 450
1 2 32 3750. 0. 0. 0.
2 3 3 4 4 5 56 67 7 8 8 9
Contd.
' Illustrative Examples 201
19 20 20 21 21 22 22 23 23 24 24 25 25 26 26 27
28 29 29 30 30 31 31 32 32 33 33 34 34 .35 35 36
46 47 47 48 48 49 49 50 50 51 51 52 52 53 53 54
10 11 16 3750.0 0.0 136.76E+10 203.4E+10
11 12 12 13 13 14 14 15 15 16 16 17 17 18
3.7 38 38 39 39 40 40 41 41 42 42 43 43 44 44 45
t 10 4 0.0 2725.0 0.38E+10 0 .76E+10
37 46 9 18 45 54
10 19 0.0 737.0
4 92.86E+10 185.72E+10
28 37 18 27 36 45
19 28 2 0.0 4176.0 92.86E-F10 185.72E+10
27 36
2 11 14 0.0 2725.0 0.62E+10 1.23E+10
38 47 3 12-39 48 4 13 40 49 5 14 41 50
6 15 42 51 7 16 43 52 8 ,17 4453
11 20 14 0.0 737.0 135.04E+10 270.08E+10
29 38 12 21 30 39 13 22 31 40 14 23 32 41
15 24 33 42 16 25 34 43 17 26 35 44
20 29 7 0.0 4176.0 135.04E+10 270.08E+10
21 30 22 31 23 32 24 33 25 34 26 35
4 3 19 27 28 36
1800.
1
0. 0. 0. 0.0. 0. 0. 0. 0. 0. 0. 0.
0
I
1
0
0
1
3
23980. 3205. 1. 1. 23980. 3205.
1
0
S
795 790
f3205 795 e
-0
positionsof
7
2
9 3
6 w
h
els
1 I I - 53
11101111 43 2
St
12t
51
8@3750m.30000
12t
F
42
MINI
14
cc;
171
17t -
Yir
4 17
ao
40 17t
N
3
30
121
12 39
17
212
0
1075
9
11725.4 I.-3500 --I 27251
1075
Figure 7.18 Equivalent Grid with Position of Live Load

Illustrative Eramples 203
. c
H
0 In on 0 71:7 0 IA 0
01 0 to
.111 -gr-
ati
te r co 0 -t
O
U
1 In
5
O
g
2
0
0
0
3
0
0
0
0
I0 O
w %.t w tp O c-I
-1
1 2 725 1180188411E183 2725 I 1917 11917) 184.1. ["1--
1144 1146
Figure 7.19 Box-Girder Bridge-Structural Idealisation for F.E.M. Analysis
REFERENCES
1. ALOK 13uowt.ucK, "To study the Behaviour of Box-Girder Bridges under Live
Load by Analysing the Bridge Using Several Methods and Comparing the Re-
sults", An M.Tech. Project, Civil Engineering Department, I.I.T. Delhi. 1990.
2. IRC 21-1987, "Standard Specifications and Code of Practice for Road Bridges",
Section III, The Indian Roads Congress, New Delhi. 1991.
3. "Ontario Highway Bridge Design Code", Ministry of Transportation and Com-
munications, Highway Engineering Design. Downsview. Ontario, Canada, 1983.
9
r t
Appendix I
Listing of Programs
Grid'*
* This program can analyse a grid under nodal loading.

C PROGRAM 'GRID'
C THIS PROGRAM CAN ANALYZE A GRID UNDER APPLIED NODAL
C LOADS. THE CONFIGURATION OF THE GRID COULD BE RIGHT,
C SKEW OR QUADILATERAL.
IMPLICIT DOUBLEPRECISION (A-H2O-Z)
COM MON/BX I /X(100), Y(I 00), XI(I00), XJ(100), SM (6,6),
NCD (6), NN(50) COMMON/BX2/GSM(200,50)
COMMON/BX3/NCN1(200),NCN2(200),NSN(20),SUPSTF(20)
COMMON/BX4/P(300),D(300)
COMMON/BX5/GDM(6),DM(6),PM(6),TM(6,6)
COMMON/BX6/LNODE(200),NLT(200),XLOAD(200)
OPEN(UN1T=3,FILE='GRID.IN')
OPEN(UNIT=4,FILE='GRID.OUT)
C READING INPUT DATA

C INITIALISATION OF MATRICES
DO 10 1=1,6
DO 5 J=1,6
SM(I,J)=0.0
TM(I,J) =0.0
5 CONTINUE
10 CONTINUE
C GEOMETRICAL PROPERTIES OF THE GRID

READ (3,*) NNODES, NELEMS
C COORDINATES OF NODES
READ (3,*)NGROUP
DO 20 I=I, NGROUP
READ (3 ,*)N1 ,N2,NINC, AX,AY,XINC,YINC
DO 15 11=N1,N2,NINC
X(I1)7AX
Y(I1) =AY
AX=AX+XINC
AY =AY+Y1NC
15 CONTINUE
20 CONTINUE
C MEMBER CONNECTIVITIES
READ (3,*)NGROUP
DO 30 1=1, NGROUP
READ(3,*)N1,N2,N1NC,NC I ,NC2,NC11NC, NC21NC
DO 25 11=NI,N2,NINC
NCN1(I1)=NCI
NCN2(I1)=NC2
NC 1 =NCI +NC IINC
NC2=NC2+NC2INC
25 CONTINUE
30 CONTINUE
C MEMBER PROPERTIES
READ (3,*)E,G
READ (3,*)NGROUP
Appendix I 207
DO 40 I=1,NGROUP
READ(_S,*)ALAJ,N,(NN(I1),I1 =1,N)
DO 35 I1=1,N
12=NN(11)
XI(12)=AI
XJ(12)=AJ
35 CO N TI NU E
4 0 C O N TI N U E
C HALF BANDWIDTH OF STIFFNESS MATRIX .

READ (3,*)NIIBAND
NDOF=3*NNODES
C INITIALIZE THE STIFFNESS MATRIX

DO 45 IA=I,NDOF DO
45 IB=1,NHBAND
45 GSM(IA,IB)=0.0
C PRINTING INPUT DATA

_WRITE(4,*)'TOTAL NO. OF NODES=',NNODES
WRITE(4,*).TOTAL NO. OF ELEMENTS=',NELEMS
WRITE(4,*)'HALF BAND WIDTH=',NHBAND
WRITE(4,*)'TOTAL NO. OF DOF.=',NDOF
WRITE(4,50)
50 FORMAT (4X,'NODE NO.',9X,'X-COORD.',7X,'Y-COORD.')
DO 60 I=1,NNODES
WRITE (4,55)I,X(I),Y(I)
55 FORMAT(6X,I3,9X,F10.4,6X, F10.4)
60 CO N TI NU E
WRITE(4,*)
WRITE(4,*)'E=',E,' G=',G
WRITE(4,*)
WRITE(4,65)
65 F013.MAT(4X,'MEMBERNO.',8X,'M.I.',14X,T,8X,'NODE1',4X,'NODE2')
DO 75 I=1, NELEMS
WRITE (4,70)LXI(1),XJ(I),NCN1(1),NCN2(1)
70 FORMAT (8X,I3,4X,E14.6,4X,E14.6,3X,13,6X,13)
75 CONTINUE
C MEMBER STIFFNESS MATRIX
DO 90 I=1,NELEMS
C DETERMINATION OF ANGLE OF TRANSFORMATION
II =NCN1(I)
12=NCN2(1)
X L = DS QRT((X(12)-X(I1))**2 + (Y(I2)-Y(I1))**2)
C =(X(12)-X(I1))/XL
S=(Y(I2)-Y(I1))/XL
X1 =12.*E*XI(1)/(XL**3)
X2 =6.*E*XI(I)/(XL**2)
X3 =4.*E*XI(I)/XL
X4 = G*XJ(I)/XL
SM(1,1)=XI
SM(2,1)=X2*S
SM(2,2)=X3*S*S+X4*C*
C
SM(3,1)= -X2*C
SM(3,2)=-X3*C*S+X4*C*S
SM(3,3)=X3*C*C-I-X4*S*S
SM(4,1)=-X1
SM(4,2)=-X2*S
SM(4,3)=X2*C
SM(4,4)=X1
SM(5,1)=X2*S
SM(5,2)=X3/2.*S*S-X4*C*C
SM (5 ,3) =-X3/2.*C*S-X4*C*S
SM(5,4)=-X2*S
SM(5,5)=X3*S*S-i-X4*C*C
SM(6,1)=:-X2*C
SM(6,2)=-X3/2.*C*S-X4*C*S
SM(6,3)=X.3/2.*C*C-X4*S*S
SM(6,4) =X2*C
SM(6,5)=-X3*C*S+X4*C*S
SM(6,6)---X3*C*C+X4*S*S
DO 80 J=1,5
DO 80 K=(.I+1),6
80 SM(J,K)=SM(K,J)
C. " ASSEMBLY OF STIFFNESS MATRIX

NCD(1)=1.1*3-2
NCD(2)=NCD(1)+1
NCD(3)=NCD(2)+1
NCD(4)=12*3-2
NCD(5)=NCD(4)+1
NCD(6)=NCD(5)+1
DO 85 IA=1,6
NI =NCD(IA)
DO 85 IB=1,6
N2=NCD(IB)-NI+1
IF(N2.LT.1) GO TO 85
GSM (N1,N2)=GSM(N1,N2)+ SM(IA,113)
85 CO N TI NU E
90 CONTINUE
C 'SUPPORT CONDITION
C READ SUPPORT TYPES- 0 FOR FLEXIBLE, 1 FOR RIGID
READ (3,*)NST
IF(NST.EQ.0) GC) TO 100
READ (3,*)NSUP
READ(3,*)(NSN(I),I = 1 ,NSUP)
DO 95 I=1,NSUP
IDOF = 3*(NSN(I)-1)+ 1
GSM(IDOF,1)=GSM(IDOF, 1)*1.0E + 6
95 CONTINUE
GO TO 110
100 READ (3,*)NSUP
READ(3,*) (NSN(1),SUPSTF(0,1=1,NSUP)
DO 105 I=1,NSUP
IDOF=3*(I-1)+1
ncivr(1nrw i)=GSM(1DOF,I)+SUPSTF(I)
105 CONTINUE
Appendix 1 209
C DECOMPOSITION OF STIFFNESS MATRIX

110 CALL DECOMP (NDOF, NHBAND, EXIT)
C INIALIZATION OF LOAD VECTOR

DO 115 1=1,NDOF
P(I)=0.0
115 CONTINUE
C LOAD VECTOR
. READ (3,1)NLOADS
DO 120 I=1,NLOADS
READ(3,*)LNODE(I),NLT(I),XLOAD(I)
. 120 CONTINUE
DO 125 I=1,NLOADS
IDOF=3*(LNODE(1)-1)+NLT(1)
P(IDOF) =XLOAD(I)
125 CONTINUE
C SOLUTION OF STIFFNESS MATRIX

CALL SOLV (NDOF, NHBAND)
WRITE(4,*)
WRITE(4,*)'DEFORMATIONS'
WRITE(4,130)
130 FORMAT(4X,'NODE NO.',4X,'VERT.DEP.',6X,'X-ROT.',10X,'Y-ROT.')
DO 140 IA=1, NNODES
11=3*IA-2
I2=I1 +I
13=12+1
WRITE(4,135)IA,D(I1),D(I2),D(13)
135 FORMAT(4X,I3,4X,3(E14.6,2X))
140 CONTINUE
C MEMBER FORCES
WRITE(4,*)
WRITE(4,I'MEMBER FORCES'
WRITE(4;145)
145 FORMAT(1X,'MEM.NO.',IX,'S.FORCE',7X,'B.MOMENT',12X,'TORSION')
WRITE(4,150)
150 FORMAT(26X,'END 1',9X,'END 2')
DO 195 I= LNELEMS
C DETERMINATION OF ANGLE OF ROTATION

I1=NCNI(I)
12=NCN2(I)
XLz---DSQRTaX(I2)-X(11))**2+(Y(12)-Y(I1)j**2)
C=(X(12)-X(II))/XL
S=(Y(12)-Y(I1))/XL
DO 155 IA=1.6
DO 155 113=1,6
155 TM(IA,IB)=0.0
TM(1,1)=1.0
TM(2,2)=C
TM(2,3)=S
TM(3,2)=-S
1:
TM(3,3)=C DO
160 IA=4,6 II 1.
=1A-3
DO 160 IB=-4,6
1
11=1B-3
160 TM(IA,113)=TM(11,1.1) 1
C GLOBAL MEMBER END DEFORMATIONS

GDM(I)=1)(11*3-2)
C
GDM(2) = D (1 1 *3-1)
C
GDM(3)=D(11*3)
GDM(4)=D(12*3-2)
C
GDM(5)=D(1249-1)
GDM(6)=D(12*3)
DO 170 I=1,6
SUM =0.0
DO 165 K=1,6
SUM=SUMA-TM(1,10*GDM(K)
165 CONTINUE
DM(J)=SUM
170 CONTINUE
X I =12.*E*X1(1)/(XL**3)
X2=6.*E*XI(I)/(XL**2)
X3=4.*E*X1(I)/XL
X4 =G*XJ(I)/XL
SM(1,1)=XI
SM(2,1)=0
SM(2,2) =X4
SM(3,1)=-X2
SM(3,2)=0
SM(3,3)=X3
SM(4,1)=-X1
SM(4,2)=0
SM(4,3)=X2
SM(4,4)= X1
SM(5,1)=0
SM(5,2) =-X4
SM(5,3)=0
SM(5,4)=0
SM(5,5)=X4
SM(6,1)=X2
SM(6,2) =0
SM(6,3)= X3/2.
SM(6,4) =X2
SM(6,5)=0
SM(6,6)=X3 DO
175 J=1,5 DO
175 K= (J+1),6
175 SM(J,K)=SM(K,J)
DO 185 2=1,6
SUM =0.0
DO 180 K =1,6
SUM =SUM
Appendix I 211
180 CONTINUE
PM(J) =SUM
185 CONTINUE
WRITE(4,190)I,PM(1),PM(3),PM(6),PM(2)
190 FORMAT(4X,I3,2X,4(E12.6,2X))
195. CONTINUE
STOP
END
C SUBROUTINE DECOMP:
C DECOMPOSITION.OF STIFFNESS MATRIX SUING
C CHOLESICEY'S METHOD.
C N ---->TOTAL NO. OF EQUATIONS
C IBW = >HALF BAND wrai
OF THE MATRIX
C EXIT = >INDICATING VARIABLE
C S = <STIFFNESS MATRIX TO BE DECOMPOSED
SUBROUTINE DECOMP(N,IBW,EXIT)
IMPLICIT DOUBLEPRECISION (A-H2O-Z)
COMMON/BX2/S(200,50)
WRITE(*,*) DECOMPOSITION STARTS'
DO 260 I=1,N.
WR1TE(s,*)I
IP=N-I+ 1
IF (IBW.LE.IP) GOTO 200
GOTO 205
200 IP=IBW
205 DO 260 LI P
SUM =S(I,J)
IQ=IBW-J
IF ((I-1).LE.IQ)GOTO 210
GOTO 215
210 IQ =I-1
215 IF (IQ.EQ.0) GOTO 230
GOTO 220
220 DO 225 K =1,IQ
225 SUM =SUM-S(I-K,1+K)*S(I-K,J+K)
230 IF (I.NE.1) GOTO 235
GOTO 240
235 S(I,J) =SUM/TEMP
GOTO 260 -
C37 IF (SUM.LE.0.0) GOTO 245
240 GOTO 250
245 EXIT=0.0
RETURN
250 EXIT =1.0
255 TEMP=DSQRT(SUM)
S(I,J) =TEMP
260 CONTINUE
RETURN
END
C SUBROUTINE SOLD':
C SOLUTION OF STIFFNESS MATRIX DECOMPOSED BY
212 Grillage Analogy 0: Bridge Deck Analysis
C CHOLESKEY 'S METHOD FOR GIVEN LOAD 5
D7 = -Pl*Pl*Pa- Fr2*P2*P3
D8 =-P1*P2*(PS -P4) D9=
P I *P 1 *P3 -In *P2*F4
M1=3*M(I,1)
M2 =3*M(I,2)
M3 =M1-1
M4 =M1-2
M5=M2-1
M6 =M2-2
MXI = D l*B(Mrh) +D2*B(M3) + D3*(B(M1)-B(M2)) + D7*B(M6) + D8*B(M5)
MY1=D2*B(146.4)+D4*B(M3)+D5*(B(M1)-B(M2))+D8*B(N16)+D9*B(M5)
MX2 =D7*B(M4) + D8*B(M3) + D3*(B(M1)-B(M2)) +D1*B(1V16) +D2*B(M5)
MY2 =D8*B(145.4) +D9*B(M3)+D5*(B(M1)-B(M2))+D2*B(M6)+D4*B(M5)
FZ =D3*(B(M--4)+B(M6))+D5*(B(M3)+B(M5))+D6*(B(M1)-B(M2))
FZ1 =F5
J=I-IBW+ I
IF ((I +1).LE.JBW) GOTO 265
GOTO 265
265 J=1
DO 270 K=1,1-1
270 SUM = SUM-S(K, (I-K+ 1))*D(K)
275 D (I) = SU/s4/S(I ,I)
280 CONTINUE DO
305 I =- I ,N
J=N-I+IBW IF (J.
GT. GOTO 285
GOTO 290
285 J=N
290 SUM =D(N-1+ I)
IF (I, EQ.1) GOTO 300
DO 295 K=(N-I+2),1
'295 SUM =SUM-S(N-I+1,K-N +D*D(K)
300 D(N-I+1)=SUM/S(N-I+1,1)
305 CONTINUE
RETURN
END
Appendix II
Listing of Program
Gabs*
* This program can analyse a grid under generalised deck loading.

PROGRAM 'GABS'
C THIS PROGRAM CAN ANALYSE A GRID UNDER GENERALISED DECK
C LOADING.THE CONFIGURATION OF THE GRID COULD BE RIGHT ,
C SKEW OR TRIANGULAR.
IMPLICIT real*8"(A-H2O-Z)
INTEGER GG,DD,WW,R,T,EE,EE1,EE2,T1,R3,PATYPE,GRIDTYPE
real*8 L,MP1:,MP2,MQ1,MQ2,MX1,MX2,MY1,MY2,E1,E2,
+MAXPT(6,4;500),MAXB(4,500),MAXRT(4, 140),MAXPT2(6,2:4,500),
+MAXB2(2:4;500),MAXRT2(2:4,100),MAXRTN(4,100),CORSPT(2,500),
+MAXR2N(2:4,100),CORSPTI(2,500),CORSPT2(2,500),CORK(100),
+CORX1(100),CORY1(100),CORIX(100),COR1Y(160),COR2X(100),
+COR1X1(100),COR1Y1(100),COR2X1(100),COR2Y1(100),
+ coRRx(2,500), CORRY(2 ,500),CORRX1 (2,500), CORRX2(2,500),
+ CORY(100);COR2Y(100), CORRY1(2,500),CORRY2(2,500),
.+CORD(100),CORD1(100),CORD2(100)
CHARACTER*1 CH1,CH3
CHARACTER*2 POS, INPFILE*12, OUTFILE*12
COMMON/BLOCK2/S(1,50000),U(50000),P(1000,7),PT(1000,7),W(100),
+D(100),B(1000),BD(1000),RT(100),FEL(500,6),RTN(100)
COMMON/BLO CK3/XC (100), YC(100),P W(100),PI ,SANG,PT1 (1000 ,2),
+DLB(1000),XLLG(0:100),YLLG(0:100) -
DIMENSION M(1000,2), ITT(100),JDD(100),JDD1(100),
+PM(1000,1)
DIMENSION DLNG(100),BL(1000)
COMMON/BLOCK6/ SCLMIN,XLL,YLL,AKERB
COMMON/BLOCK8/0MAXPT(6,4,1000), oMAX13(4,1000), OMAXRT(4,100),
+0MAXRN(4,100)
COMMON/XX1/AIMP
CHARACTER*20 FILE1, FILE2
WRITE(*,*) 'Enter the file name for Input '
READ(*,5) FILEI
5 FORMAT(A20)
WRITE(*,*) 'Enter the file name for Output
READ(*,5) FILE2
OPEN(UNIT=3,FILE:=F1LE1)
OPEN(UNIT =4 ,FILE =FILE2)
C OPEN(UNIT=3,FILE='*SRC.EX2')
C OPEN(UNITL-4,FILE='*SRC.GRID3OU')
MOVE=O
XL1 =0.0
YLI =0.0
XINT=0.0
YINT=0.0
DO 10 i=1.6
DO 10 J=1,4
Appendix II 215
DO 10 K=1,500
10 MAXPT(I,J,K)=0.0
DO 15 1=1,4
DO 15 3=1,500
15 MAXB(I,J)=0.0
DO 20 I=1,4
DO 20 3=1,100
MAXRTN(I,J)=0
20 MAXRT(I,J)=0
P1=4*ATAN(1.0)
25 READ(3;230) CHI
IF (CHI.EQ.'N') GO TO 315
READ(3,*) GRIDTYPE
C GRIDTYPE CAN EITHER BE SKEW(ANY VALUE OTHER. THAN 90) OR
C ORT H.(90)
READ (3,*) ANG
READ (3,*) MM,N,DD,E,G,NLG,(DLNG(I),I=1,NLG-1)
READ (3,*) E1,E2
SANG = ANG*PI/180
WRITE (4,235)
WRITE (4,240) ANG
WRITE (4,245) MM,N,DD,E,G
WRITE(4,290)NLG,(DLNG(I),I=1,NLG-1)
WRITE(4,295)
WW =0
1=0
DO 45 GG=1, DD
), I= I+ 1
READ (3,*) M(I,1),M (1,2),T,P(I,1),P(1,2),P(1,3),P(I,4)
WRITE (4,250) GG
WRITE (4,255)T,(P(1,1C2),K2=1,4)
PM(I, 1) = P(I , 1).
L=SQRT(P(I,1)**2+P(I,2)**2)
P(I,1) =P(I,
p(1,2)=P(1,2)/L.
P(I,3)=2*E*P(1,3)/L
H =P(1,3)
P(I,4)=G*P(I,4)/L
P(I,5) = 6*H/(L**2)
P(1,6)=2*H
P(I,7) =3 *HiL
EE=I
EE1=EE+1
EE2=EE+T-1
DO 30 II I =EE1,EE2
30 PM(II I, I) = PM(EE,1)
READ (3,*) (M(IAX, 1), M(IAX ,2),IAX =EE1,EE2) Nil
1+T 1
WRITE (4,260) (M(IAX1,1), M(IAX1,2),IAX1=1,N11)

T1=T-1
DO 40 R=1,T1
I=R+EE
DO 35 .1=1,7
35 P(I,J)=P(I-1,J)
40 CONTINUE
45 CONTINUE
NTG=N/NLG
WW =3*NTG +3
YLLG(0)
DO 50 I=1,NLG-1
YLLG(1) = YLLG(1- 1.)+ DLNG(1)
50 CONTINUE.
YSPAN =YLLG(NLG-1)
XLLG(0) =0
DO 55 I=1,NTG-1
CALL INODE(M,I,I+1,L1)
XLLG(I) = XLLG(I-1) + PM(L1, 1)
55 CONTINUE
SPAN=XLLG(NTG-1)
N3=N*3
NROWS=N3
NROWS1=NROWS
N8 = N3*WW
DO 601=1, 15
60 JTT (I)=0
READ (3,*) IX,JRR, (JTT(1), I=1,IX)
IF (JRR.EQ.1)THEN
READ(3,*) SUPK
ENDIF
WRITE (4,265)
WRITE (4,270) IX,(JTT a1m1=1,1X)
IF (JRR.NE.3) WRITE(4,280)
(JRR.EQ.3) WRITE(4,275)
IF (JRR.EQ.1)THEN
WRITE (4,285)SUPK
ENDIF
READ (3,*) AKERB
AKERB = AKERS - 0.5
NNODES =N
CALL DEAD LOAD (DLB,N,NLG,M,PM,E1,E2)
C THE FOLLOWING DO LOOP STORES IN JDD THE VERT DEFN. NOS
C CORRESPONDING TO THE SUPPORTED NODES.
DO 65 I=1,IX
65 JDD(I)=3*JTT(I)
CALL STIFF(N,MM,WW,P,M,S)
IF (JRR.EQ.3) THEN
--
Appendix II 217
CALL SORT(IX,JDD,JDD1)
CALL MODSTIF(WW,IX,JDD1,S,NROWS)
ELSE
DO 70 1=1,1X
NE=WW*(JDD(I)-1)+ I
S(1,NE)=S(1,NE)-1-SUFK
70 CONTINUE
ENDIF
CALL DECOMP(NROWS,WW,S,U)
C COMPUTATIONS OF LOAD MATRIX STARTS
75 READ (3,230) CH3
IF (CH3.EQ.'S') GO TO 315
NL=N
N9=3*N+WW
READ (3,*) 'NOLCASE
WRITE(4,80)
80 FORMAWING(1) = DEAD LOAD ONLY'/
+ ' LDING(2) = CLASS A - TWO LANE'/
+ ' LDING(3) = CLASS 70-R TRAIN:COL -L'/
+ ' LDING(4) = CLASS 70-R TRAIN:COL -M'/
LDING(5) = CLASS 70-R BOGIE:COL -1.1/
LDING(6) = CLASS 70-R-BOGIE:COL -M'/
LDING(7) = CLASS 70-R TRACK'/
+ ' LDING(8) = CLASS A - SINGLE LANE'/
+ ' LDING(9) = SPECIFIED BY USER '/)
DO 225 IMM1=1,NOLCASE
READ(3,*)LCASE
IF(LCASE.EQ.I)THEN
WRITE(*,*)'DEAD LOAD ONLY'
WRITE(4,*)'DEAD LOAD RESULTS'
DO 85 JA3=1,3*NL
85 B(JA3)=DLB(JA3)
GOTO 135
ENDIF
READ(3,*)XL,YL,XINC,YINC,XSTOP,YSTOP
MODIN1=0
MODIN2=0
XFIN = XL + REAL(INT((XSTOP-XL)/XINC))*XINC
YFIN = YL REAL(INT((YSTOP-YL)/YINC))*YINC
WRITE(4:*)'
WRITE(4,310)LCASE
WRITE(4,*)' ------------------------
CALL IMPACT (LCASE, SPAN,SANG)
WRITE(4,'(10X,A)TTHE RESPONSE IS INCLUSIVE OF DEAD LOAD RESULTS'
WRITE (4,300)XL, YL, XINC, YINC,XFIN, YFIN,AIMP
MOVE =0
IM1 =0
YL1 =YL-YINC
218 :Grillage Analogy in Bridge Deck Analysis
90 XL1 =XL-XINC
YL1 =YL1+YINC
95 XL1 =XLI + XINC
IF (YLLGT.YSTOP)THEN
GO TO 200
ENDIF
IF (XLLGT.XSTOP)THEN
GO TO 90
ENDIF
IF(MODIN1. EQ.1) THEN
XL1 =XL1-1.0
MODIN1 =0
ENDIF
IF(XLI.EQ.0) THEN
XL1=1.0
MODINI =1
ENDIF
IF(MODIN2.EQ.1) THEN
XL1 =XL1 +1.0
MODIN2=0
ENDIF
IF(XLLEQ.SPAN) THEN
XL1=XL1-1.0
MODIN2=.1
ENDIF
C WRITE(*,*)'LOAD CASE ',LCASE,' LEFT FRONT WHEEL AT ',XL1,
C +' AND ',YL1
IM1=IMI+1
CALL LLOAD (LCASE, XLI, YL1, XC, YC, NW; NWPA,IM1)
CALL WLOAD (LCASE, PW,SCLMIN, NW, IM1)
IXL=0
IXG=0
DO 100 JA2=1,3*NL
100 B(JA2)=0.0
DO 110 JA2=1,MM
DO 105 JA3=1,6
105 FEL(IA2,1A3)=0.0
110 CONTINUE
DO 120 I5P=1,NW
IF ((GRIDTYPE.NE.90).bR.(SANG.EQ.0.0)) THEN
CALL PAIDRB(XLLG,YLLG,XC(15P),YC(I5P),SANG,IXL,IXG,A10ERB
, SCLMIN, SPAN, POS, PATYPE, MN1, MN 2, MN11 , MN22, X LL, YLL, XPL, YPL
+ ,NTG,NLG)
ELSE
CALL PAID1(M,PM,N,NLG,DLNG,MN1,MN2,MN11.MN22,PATYPE,
+ XC(I5P),YC(15P),XLL,YLL,XPL,YPL,SANG,IXG,IXL,POS,
+ AKERB,SCLMIN)
ENDIF
Appendix II 219
IF(POS.EQ.'YOGO TO 90
IF(POS.EQ.'YG')GO TO 90
IF(PATYPE.EQ.0)G0 TO 120
IF(PATYPE.EQ.3)THEN
CALL LDISTT (MN1,MN2,MN22,PW(15P),XPL,YPL,XLL,YLL,BL,POS,NL)
ELSEIF(PATYPE.EQ.4)THEN
IF (GRIDTYPE.EQ.90)THEN
SANG! =0.0
CALL LDISTR (MN1,MN2,MN1I,MN22,PW(I5P),SANG1,13L,XPL,YPL,XLL,
+ Yi,L,POS,NL)
ELSE
SANGI=SANG
CALL LDISTR (MN1,MN2,MN11,
+ MN22,PW(15P),SANG1,BL,XPL,YPL,XLL,YLL,POS,NL)
ENDIF
ENDIF
DO 115 JOG=I,3*NL
115 B(JOG)=B(JOG)+BL(JOG)
120 CONTINUE
IF(IXG.EQ.NW)GO TO 90
IF(LXL.EQ.NW)G0 TO 95
MOVE=MOVE+1
DO 125 IFI=1,3*N
125 B(IFI)=B(IFI)*AIMP
DO 130 JA3=1,3*NL
130 B(JA3)=B(JA3)+DLB(IA3)
C COMPUTATION OF LOAD MATRIX ENDS
135 DO 140 I=1,NROWS1
BD(I)=-B(1)
140 CONTINUE
IF (JRR.EQ.3) THEN
DO 150 1=1X,1,-1
J=JDDI(I)
DO 145 IP= J,NROWS1
B(IP)=B(IP+1)
145 CONTINUE
150 CONTINUE
CALL SOLVE (NROWS,WW,U,B,B)
DO 155 I=NROWS+1,NROWSI
155 B(I)=0
DO 165 I=1,IX
J=.113D1 (I)
DO 160 IP =NROWSI,J,-1
B(IP + I) =B(IP)
160 CONTINUE
B(IP +1)=0
165 CONTINUE
LT C."0
CALL SOLVE(NROW S1 , WW , U, B , B)
ENDIF
C COMPUTATIONS OF ELEMENT FORCES STARTS
DO 170 I =1,MM
P1 =P(I,1)
P2 =P(I,2)
P3 =P(I ,3)
P4 =P(1,4)
P5 =P(1,5)
P6 = P(I, 6)
P7 =P(I,7)
D1 =-P1*P1*P4 +P2*P2*P6
D2 =P1*P2*(P4-P6)
D3 = P2*P7
D4 =P2*P2*P4+P 1 *Pl*P6
D5 = -P 1 *P7
D6 = P5
D7 = -Pl*P 1*P4 +P2*P2*P3
D8 =-Pl*P2*(P3 +P4)
D9 =P1*P1*133-P2*P2*P4
M =3*M(I, I)
M2 =3*M(I,2)
M3 =M1-1
M4 =M1-2
M5 =M2-1
M6 = M2-2
MX1 =D1*B(M4)+D2*B(M3)+D3*(B(M1)-B(M2))+D7*B(M6)+D8*B(M5)
MY1 = D2*B(M4) + D4*B(M3) + D5*(B (M1)-B (M2)) + D8*B(M6) + D9*B (M5)
MX2 = D7493(M4) + D 8*B (M3) + D3*(13(M1)-B(M2)) + D 1*B (M6) + D2*B(M5)
MY2 = D8*B (M4) + D9*13(M3) + D5*(B (M1)-B(M2)) + D2*B(M6) + D4*B (M5)
FZ = D3*(B (M4) + B(M6)) + D5*(B (M3) + B (M5 ))+136*(B(M1)-B (M2))
FZ1= FZ
MQ 1= (-MX1)*P2 +(MY1)*P1
MQ2 = (-MX2)*P2 + (MY2)*P1
MP1 = (MX1)*P1 +(MY1)*P2
MP2 = (MX2)*P1 +(MY2)*P2
PT(I,3) =FZ1
PT(I,4) = MP 1
PT(L5) = MQ I
PT(I, 6) =MQ2
170 CONTINUE
C COMPUTATIONS OF ELEMENT FORCES ENDS
DO 175 II =1,1X
RT(11 ) =0
175 CONTINUE
DO 185 11=1,1X
DO 180 12=1,MM
IFWTT(11 ) . EQ. Ivi(12, 1)).OR.(11-f(11).EQ.M(12,2))) THEN
Appendix 11 221
IF (JTT(II).EQ.M(I2,I)) THEN
RT(I1) =RT(I1) +PT(12,3)
ELSE
RT(I1) =RT(I1)-PT(I2,3)
ENDIF
ENDIF
180 CONTINUE
185 CONTINUE
DO 190 II =1,IX
NP1=3*JTT(II)
RT(I1)=RT(11)-BD(NP1)
190 CONTINUE
DO 195 JP8=1,2
DO 195 IP8=1,MM
KP8 =JP8 +4
195 PTI(IP8,11)8)=PT(IP8,KP8)
CALL COMPARE1 (PT,B,RT,MAXPT,MAXB,MAXRT,MAXRTN,MM,N,IX,
+MOVE,PT1,XL1,YLI,LCASE,CORSPT,CORX,CORY,CORXI,CORY1,
+ CORRX, CORRY, CORD I ,JTT)
IF(LCASE.EQ.1) GOTO 200
GO TO 95
200 READ (3,*)IIC2
IF (IK2.EQ.I)CALL WRITE1(B,MAX7F,MAXB,MAXRT,MAXRTN,MM,
+N,IX,MOVE,M,JTT,LCASE,CORSPT,CORX,CORY,CORX1,CORY1,
+CORRX,CORRY,CORDI)
IF(NOLCASE.EQ.1) GO TO 75
DO 205 IL1=1,4
IL2 =11,1+2
DO 205 IL3 =1,MM
PT (IL3,112)=MAXPT(IL1,1,IL3)
DO 205 1=2,4
MAXPT2(IL I ,I ,IL3) =MAXPT(IL1,I,IL3)
CORSPT2(1,IL3)= CORSPT(1,IL3)
CORSPT2(2,IL3) =CORSPT(2,IL3)
205 CONTINUE
DO 210 1=1,N
DO 210 .1=3*1-2,3*1
B(J)=MAX13(1,J)
DO 210 K=2,4
MAXBAK,J)=MAXB(K,J)
CORRX2(1,I)=CORRX(1,I)
CORRY2(1,I)=CORRY(1,I)
CORRX2(2,I)=CORRX(2,I)
CORRY2(2,I) =CORRY(2,I)
210 CONTINUE
DO 215 1=1,1X
RT(I) =MAXRT(I ,I)
RTN(I)=MAXICIN ,
DO 215 J=2,4
MAXRT2(J,I) =MAXRT(J,1) 2'
MAXR2N(.1,1)=MAXRTN(J,I) 2
COR2X(D=CORX(I) 2
COR2Y(I)=CORY(I) 2
COR2X1(I)= CORX I (I)
COR2Y1 (I) = CORY1 (I)
CORD2(I)=CORD1(I)
215 CONTINUE
DO 220 JP8=1,2
KP8=JP8+4
DO 220 108=1,MM
PTI(IP8,JP8) = MAXPT(ICP8, 1, IP8)
DO 220 1=2,4
MAXPT2(KP8,I,IP8)=MAXPT(ICP8,LIP8)
220 CONTINUE
CALL COMPARE2(PT,B,RT,RTN,OMAXPT,OMAXB,OMAXRT,OMAXRN,MM,
+N,IX,IMM1,PT1,MAXPT2,
+MAXB2,MAXRT2,MAXR2N,JTT,COR2X,COR2Y,COR2X1,COR2Y1,
+COR1X,CORIY,COR1X1,CORIYI,CORSPT1,CORSPT2,
+CORRX1,CORRX2,CORRY1,CORRY2,CORD,CORD2)
225 CONTINUE
WRITE(4,80)
LCASE =0
READ(3,*) 11C3
IF (IK3.NE.1) GO TO 75
CALL WRITE2(0MAXPT,OMAX13,0MAXRT,OMAXRN,MM,N,IX,MOVE,M,JTT,
+ LCASE,CORIX,CORlY,CORIX1 , CORlY I ,CORSPT1,CORRX1,CORRYI,CORD)
GO TO 75
230 FORMAT (1A1)
235 FORMAT(2(/),30X,'BRIDGE DETAILS'/29X,16('*')//)
240 FORMAT(10X,'SKEW ANGLE = ',F9.3, ' DEGREES')
245 FORMAT(10X,'NO. OF ELEMENTS= ',13/,10X, 'NO. OF NODES=',
+ 13/,10X,'NO. OF ELEMENT GR.OUPS-=',12/,
+ 10X, 'YOUNGS MODULUS OF ELASTICITY=',
+ E14.7, ' T/SQ.MMVIOX,'SHEAR MODULUS OF ELASTICITY = ',E14.7,
+ ' T/SQ.MMV)
250 FORMAT(//15X,'GROUP',I2/15X,7('-'))
255 FORMAT(20X,'NO. OF ELEMENTS IN THIS GROUP = ',12/20X,
+ 'LENGTH OF ELEMENT IN X-DIRECTION =',F7.0,' MM'/20X,
+ 'LENGTH OF ELEMENT IN Y-DIRECTION =',F7.0,' MM'/20X,
+ 'MOMENT OF INERTIA OF ELEMENT =',E14.7,' MM4'/20X,
+ 'TORSIONAL INERTIA OF ELEMENT =', E14.7,' MM4') 260
FORMAT(/20X,'ELEMENTS IN THIS GROUP:'//7(20X,5(12,'-',
+ I2,3X)/))
265 FORMAT (/15X,'BOUNDARY CONDITIONS-SUPPORT NODES'
+ /15X, 33 ('-'))
270 FORMAT (20X,'NO. OF SUPPORT NODES 12120X,
Appendix Il 223
+ 'NODE NO. OF SUPPORTS:', 10(12,2X))

275 FORMAT (20X,'NON YIELDING SUPPORTS')
280 FORMAT (20X,'YIELDING SUPPORTS')
285 FORMAT(20X,'STIFFNESS OF BEARINGS',2X,E9.3,'N/MM')
290 FORMAT(/5X,'NO. OF LONGITUDNAL GRID LINES = ',14//5X,
+ 'SPACINGS (IN MM) = ',(F8.0,5X))
295 FORMAT(/3X,'DETAILS OF ELEMENTS IN EACH GROUP')
300 FORMAT(/15X,'INITIAL COORDINATE OF '
+,'LEADING LEFTMOST WHEEL IS','(',F8.1,',', F8.1,')',/15X,
+ 'INCREMENT ALONG-X-AXIS=', F7.0,14M',/15X,'INCREMENT ALONG '
+,'Y-AXIS=', F7.0,'MM'
+,/15X,'FINAL CO-ORDINATE OF LEADING LEFTMOST WHEEL IS',
+,/15X,'IMPACT FACTOR= ',F8.4//)

310 FORMAT(' ENVELOPE VALUES UNDER LIVE LOAD ',I3)
315 STOP
END
c***********************************************************************
SUBROUTINE STIFF(N,MM,WW,P,M,S)
IMPLICIT real*8(A-H2O-Z)
a DIMENSION S(1,50000),P(1000,7),M(1000,2)
INTEGER U,WW,R1,R2,R3,R
N3 =N*3
N8 =N3*WW
DO 320 II=1,N8
320 S(1,II)=0.0
DO 325 I=1,MM
P1=P(I,1)
P2=P(I,2)
P3 =P(I,3)
P4 =P(I,4)
P 5 = P ( I , 5 ) .
P 6 = P ( I , 6 ) P 7
= P ( I , 7 )
M 1 = 3 * M ( I , 1 )
M1 = (MI-1)*WW
M2 =(MI-2)*WW
M3 =(MI-3)*WV/
S(1,M3 +1) =S(1,M3 + 1) + P 1 *P I *P4
+P2*P2*P6 S(1,M3+2)=S(1,M3+2)+P1*P2*(P4-
P6) S( I ,M3 +3) =S(1,M3 +3) + P2*P7
S(I,M2+1)=S(I,M2+1)+P2*P2*P4+P1iP1*P6
S(1 , M2 +2) =S(1 ,M2 +2)-P I *P7
S(1,M1+1)=S(1,M1+1)+P5
MI=3*M(I,2)
MI = (MI-I )*WW
M2 =(MI-2)*WW
M3 =(MI-3)*WW
S(1,M3+1)=S(1,M3+1)+Pl*Pl*P4+P2*P2*P6
S(1 ,M3 +2) =S(1,M3 + 2) +P I *P2*(P4-P6)
S(1,M3 +3) =S( I ,M3 +3)-P2*P7
S(1,M2+1)=S(I,M2+1)+P2*P2*P4+Pl*P1*P6
S(1,M2+2)=S(1,M2+2)+P1*P7
S(I,M1+1)=S(1,M1+1)+P5
C ABOVE STATMENTS SET UP DIAGONAL ELEMENTS
U=3*IABS(M(I,2)-M(I,1))
R=3*M(I,1)
RI =(R-1)*'WW+U
R2 =(R-2)*W1V+U
R3 =(R-3)*WW +U
S(1,R3+1)=S(1,R3+1)-Pl*P1*P4+P2*P2*P3
S(1,R3+2)=-S(1,R3+2)-Pl*P2*(P4+P3) S(1
,R3 +3) = S(1,R3 +3)-P2*P7
S(1,R2)=S(1,R2)-P1*P2*(P4+P3)
S(1,R2 + I) = S(1,R2+ 1)-P2*P2*P4 +Pl*Pl*P3
S(1,R2+2)=S(1,R2+2)+P1*P7
S(1,R1-1)=S(I,R1-1)+P2*P7
S(1,R1)=S(1,R1)-PI*P7
S(1,R1+ I)=S(1,R1+1)-P5
325 CONTINUE
C ABOVE STATMENT SET UP OFF DIAGONAL ELEMENTS
RETURN
END
c***********************************************************************
SUBROUTINE INODE (M,I1,I2,16)
DIMENSION M(1000,2)
IF (12.LT.I1) THEN
WRITE(*,*)*ERROR'
READ(*,*)0
ELSE
K1=0
330 K1 =K1+1
IF (M(K1,1).NE.II.OR.M(K1,2).NE.I2) GO TO 330
I6=K1
ENDIF
RETURN
END
c***********************************************************************
SUBROUTINE PAID1(M,PM,N,NLG,DLNG,MN1,Mi',12,MN11,MN22,
+ PT,XL,YL,XLL,YLL,XPL,YPL,SANG,IXG,IXL,POS,AKERB,SCLMIN)
INTEGER PT
CHARACTER*2 POS
DIMENSION M(1000,2),PM(1000,1),DLNG(100),XLLG(100),SLNG(I00)
POS='NN'
Appendix-11 225
X=0
DO 335 I=1,NLG-1
SLNG(I)=0
X=X+DLNG(I)
335 SLNG(1)=X -
SPAN =0
DO 340 !IFF =1 ,(N/NLG-1)
XLLG(IFF) =0
CALL INODE(M,IFF,IFF+1,IN1)
SPAN=SPAN+PM(IN1,1)
XLLG(IFF)=SPAN
340 CONTINUE
IF (YL.LT.(AKERB+SCLMIN))GO TO 470
IF(YL.GT.(SLNG(NLG-I)-(AKERE+SCLMIN)))GO TO 475
IF((XL.LT.(SLNG(NLG-I)*TAN(SANG))).AND.(YL.GT.XL/
+ TAN(SANG))) GO TO 480
IF ((XL.GT.SPAN).AND.(YL.LT.((XL-SPAN)/TAN(SANG))))
+ GO TO 485
XLEN=0.0
DO 345 II =1,NLG-1
CALL INODE(M,I1,II + I ,LI)
345 XLEN=XLEN+PM(L1,1)
IF ((XL.GT.XLEN).AND.(XL.LE.SPAN))GO TO 350
IF (XL.LT.XLEN) GO TO 380
IF (XL.GT.SPAN) GO TO 425
3 5 0 PT = 4
DO 355 II =NLG,N/NLG-I
CALL INODE (M,I1J1+1,L1)
XLEN=XLEN+PM(L1,1)
IF (XL.LE.XLEN) GO TO 360
355 CONTINUE
360 IF(YL.GT.0)GOTO 365
MN 1=11
MN 2=11+1
MN 11 = II +(N/NLG-1)
MN 22=11+N/NLG
XLL=XL-(XLEN-PM(LI,1))
YLL=0.0
XPL=PM(L1,1)
YPL=DLNG(1)
RETURN
365 DO 370 12=1,NLG-1
IF ((SLNG(12)-YL).GE.0.0)G0 TO 375
370 CONTINUE
375 MN 1=11+ (I2-1)*(N/NLG- )
MN 2=iviNi+1
MN 1I=MN1+(N/NLG-1)
MN 22=MN11+1
XLL=XL-(XLEN-PM(L1,1))
IF(12.NE.1)YLL=YL-SLNG(I2-1)
IF(12.EQ.1)YLL=YL
XPL=PM(L1,1)
YPL=DLNG(12)
RETURN
380 POS = 'XL'
DO 390 13= 1,NLG-1
IF(13.NE.1)GOTO 385
IFaL.LE.XLLG(1)).AND.(YL.LE.XL/TAN(SANG))) GOTO 395
GOTO 390
385 IF ((XL.LE.XLLG(I3)).AND.(YL.LE.XL/TAN(SANG)).AND.(YL.GE.SLNG
+ (13-1))) GO TO 395
390 CONTINUE
GO TO 400
395 PT=3
MN 1 =1 ,+(I3-1)*N/NLG
MN 2=MN1+I
MN 22=MN1+N/NLG
MN 11=0
CALL INODE(M,I3,13 +1,131)
XPL=PM(I3I,1)
YPL=DLNG(13)
IF(13.NE.1)XLL=XL-XLLG(I3-1)
IF(I3.NE. DYLL = YL,SLNG(13-1)
IF(I3.EQ.1)XLL=XL
IF(I3.EQ.1)YLL=YL
RETURN
400 PT =4
DO 405 14 =2,NLG-1
IF (XL.LE.XLLG(I4)) GO TO 410
405 CONTINUE
410 DO 415 15=1,NLG-1
IF ((YL-SLNG(15)).LE.0) GO TO 420
415 CONTINUE
420 MN I =14+(15-1)*(N/NLG-1)
MN 2=MN1+1
MN 11 =MN1 +N/NLG- I
MN 22 = MNI 1 + 1
CALL INODE(M,I4,14+1,I3I)
XPL=PM(I3I,I)
YPL=DLNG(I5)
XLL = X L-XL.TL-G(I4-1)
IF(15.NE.1)YLL =YL-SLNG(15-1)
IF(I5.EQ.1)YLL=YL
RETURN
425 POS = 'XG'
XXL= (SPAN +XLLG(NLG-1))-XL
Appendix II 227
YYL=SLNG(NLG-1)-YL
DO 435 16=1,NLG-1
IF(I6.NE.1)GOTO 430
IF((XXL.LEALLG(1)).AND.(YYLLE.XXL/TAN(SANG)))GOTO 440
GOTO 435
430 IF ((XXL.LE.XLLG(16)).AND.(YYLLEJO(L/TAN(SANG)).
+ AND.(YYLGE.SLNG(16-1))) GO TO 440
435 CONTINUE
GO TO 445
440 PT =3
MN I =(146)*N/NLG+N
MN 2=MN1-1
MN 22=MN1-NINLG
MN 11=0
CALL INODE(M,I6,16+1,131)
XPL = PM(I3I, 1)
YPL=DLNG(NLG-16)
IF(I6.NE.1)XL,L=XXL-XLLG(16-1)
IF(I6 .NE. DYLL =YYL-SLNG(I6-1)
IE(16.EQ.1)XLL=XXL
IF(I6. EQ. 1)YLL=YYL
RETURN
445 PT=4
DO 450 17=-2,NLG-1
IF (XXL.LEILLG(17)) GO TO 455
450 CONTINUE
455 DO 460 I8=1,NLG-1
IF ((YYL-SLNG(18)).LE.0) GO TO 465
460 CONTINUE
465 MN22= N-(18-1)*N/NLG-(I7-18)
MN11=MN22-1
MN2=MN22-(N/NLG-1)
MN1=MN2-1
CALL INODE(M,I7,17+1,131)
XPL=PM(I3I,1)
YPL=DLNG(I8)
XLL=XLLG(I7)-XXL
IF(18.NE.1)YLL =SLNG(18)-YYL
RETURN
470 POS='YL'
RETURN
475 POS='YG'
RETURN
480 IXL =Del,4-1
GO TO 490
485 IXG =IXG+ I
490 PT =0
RETURN
END
c**********************************************************************
SUBROUTINE PAIDRB(XLLG,YLLG,XL,YL,SANG,IXL,IXG,AKERB,SCLMIN,
SPAN,POS,PT,MN1,MN2,MN11,MN22,XLL,YLL,XPL,YPL,NTG,NLG)
DIMENSION YLLG(0:100),XLLG(0:100)
INTEGER PT
CHARACTER*2 POS
POS=
PT=0
IF (YL.GT.(AKERB+SCLMIN)).THEN
IF (YL.LT.(YLLG(NLG-1)-(AKERB+SCLMIN))) THEN
IF (XL.GT.(YL*TAN(SANG))) THEN
IF X(XL-SPAN).LT.(YL*TAN(SANG))) THEN
PT =4
IF (YL.LT.0.0) THEN
11=1
495 IF (XL.LT.XLLG(I1)) THEN
MN1=I1
MN2=I1 +1
MN11=MN1+NTG
MN22=MN11+1
XLL=XL-XLLG(I1-1)
YLL = 0
XPL =XLLG(I1)-XLLG(11-1)
YPL=YLLG(1)
ELSE
II =I1 +1
GO TO 495
ENDIF
ELSE IF (YL.GT.YLLG(NLG-1)) THEN
11=1
G=XL-YLLG(NLG-1)*TAN(SANG)
500 IF (G.LT.XLLG(I1)) THEN
MN II = II +NTG*(NLG-1)
MN22 =MN11 +1
MN1=MN11-NTG
MN2 =MN1 +1
XLL=G-XLLG(I1-1)
YLL=YLLG(NLG-1)-YLLG(NLG-2)
XPL=XLLG(I1)-XLLG(I1-1)
YPL=YLL
ELSE
I1 =II +1
GO TO 500
ENDIF
ELSE
11=1
Appendix 11 229
505 IF (YL.LT.YLLG(I1)) THEN

N2=11
ELSE
I1=I1+1
GO TO 505
ENDIF
II =1
G=(XL-YL*(TAN(SANG)))
510 IF (G.LT.XLLG(I1))THEN
N1 =I1
ELSE
11=I1+1
GO TO 510
ENDIF
MNI =N1 +(N2-1)*NTG
MN2=MN1+1
MN11=MN1+NTG
MN22=MN11 +1
XLL=XL-XLLG(N11)-YL*TAN(SANG)
YLL=YL-YLLG(N2-1)
XPL=XLLG(N1)-XLLG(N1-1)
YPL=YLLG(N2)-YLLG(N2-1)
ENDIF
ELSE
IXG
PT =0
ENDIF
ELSE
IXL=IXL +1
PT=0
ENDIF
ELSE
POS = 'YG'
ENDIF
ELSE
POS ='YL'
ENDIF
RETURN
END
C ****************************************************
SUBROUTINE WLOAD (LCASE,PW.SCLMIN,NW,MOVE)
IMPLICIT real*8 (A-H2O-Z)
DIMENSION PW(100),PWI(100)
DO 515 11=1,28
515 PW (11)=0.0
GO TO(10,1,3,3,5,5,7,1,9),LCASE
IF AIL= 1
RETURN
1 SCLMIN=400 PW(I)=1.35
PW(2)=1.35 PW(3)=135
PW(4)=1.35 PW(5)=5.70
PW(8)=5.70 PW(6)=5.70
PW(7)=5.70 PW(9)=3.40
PW(10)=3.40
PW(11)=3.40
PW(12)=3.40
PW(13)=3.40
PW(14)=3.40
PW(15)=3.40
PW(16)=3.40 IF
(LCASE.EQ.2)GO TO 2
RETURN
2 DO 520 1=17,32
520 PW (I)=PW(I-16)
RETURN .
3 SCLMIN=1405
DO 525 1=1,4
PW (1)=2.0
PW (I+4)=3.0
525 PW (1+8)=3.0
DO 530 1=13,28
530 PW (I)=4.25
RETURN
5 SCLMIN=1405
DO 535 1=1,8
535 PW (I)=5.0
RETURN
7 SCLMIN=1620
DO 540 1=1,20
540 PW (1)=3.50
RETURN
9 IF (MOVE.NE.1) GO TO 550
READ (3,*) SCLMIN1
READ (3,*)(PW1(K),K=1,NW)
WRITE(4,545)(PW1(K),K=1,NW)
545 FORMAT('LOAD ON WHEEL' ,10E10.2)
550 SCLMIN =SCLMINI
DO 555 K=1,NW
PW (K) =PWI(K)
555 CONTINUE
10 RETURN
Appendix 11 231
END
c********************************************************i**************
SUBROUTINE LLOAD (LCASE,XT,YT,XW,YW,NW,NWPA,MOVE)
IMPLICIT reals8(A-H2O-Z)
r. DIMENSION XW(100),YIN(100),XW1(100),YW1(100)
-DO 565 11=1,28
XW (II)=0.0
565 YW (11)=0.0
X=XT
Y=YT
GO TO (10,1,3,4,5,6,7,1,9),LCASE
IF AIL=1
RETURN
1 NW =16
NWPA =2
DO 570 11 =1,8'
YW (I1*2-1)=Y
570 YW (11 41)=Y+1800
XW (1)=X
XW (2)=XW(1)
XW (3) =XW(2)-1100
XW (4)=XW(3)
XW (5) =XW(4)-3200
XW (6)=XW(5)
XW (7)=XW(6)-1200.
XW (8)=XW(7)
XW (9)=XW(8)-4300
XW (10)=XW(9)
XW (11)=XW(10)-3000
XW (12)=XW(11)
XW (13)=XW(12)-3000
XW (14)=XW(13)
XV(I5)=XW(14)-3000
XV, (16)=XW(15)
IF (LCASE.EQ.2) GO TO 2
RETURN
2 NW =32
NWPA =4
DO 575 1=17,32
575 XW (1)=XW(I-16)
DO 580 1=9,16
YW (1*2-1)=YW((I-8)*2)+1700
580 YW (I*2)=YW(I*2-1)+1800
RETURN
3 NW=28
NWPA
=4 K-7
585 DO 590 1=1,K
YW(I*4-3)=Y
YW(I*4-2) =Y +450
YW(I*4-1)--= Y+ 1930
590 YW(I*4) =Y +2380
IF (LCASE.EQ.5)GO TO
615 595 DO 600 I=1,4
XW(I) =X
XW(I+4) =XW(I)-3960
XW(I + 8) = XW(I +4)-1520
VW (I+ 12) =XW(I +8)-2130
XW(I+16)=XW(I+12)-1370
XW(I+20)=XW(I+16)-3050
600 XW(I +24) = XW(I +20)-1370
RETURN
4 NW=28
NWPA=4
L=7
605 DO 610 I=1,1,
YW(I*4-3)=Y
YW(I*4-2)=Y+795
YW(I*4-1) = Y + 1585
YW(I*4)=Y +2380
610 CONTINUE
IF (LCASE.EQ.6)GO TO 615
GO TO 595
5 NW =8
NWPA=4
K=2
GO TO 585
615 DO 620 I=1,4
XW(I) =X
620 XW(I+4)=X-1220
RETURN
6 NW=8
NWPA=4
L =.2
GO TO 605
7 NW=20
NWPA=2
DO 625 1=1,10
YW(I*2-1)=Y
625 YVV(I *2) =Y +2060
DO 630 1=1,10
XW(I*2-1)=X-((I-1)*457)
630 XW(I*2)=XW(I*2-1)
RETURN
!F. N!'r 1 )r:Ct TO Fall
READ (3,*)NW I
Appendix 1.1 233
IF(NW1.EQ.1)G0 TO 640
DO 635 12P=2,NW1+1
IF (12P.EQ.NW1+1)GOTO 635
READ (3,*)XW1(I2P),YW1(I2P)
635 CONTINUE
640 NW=NW1
XW(I)=XT
YW(1)=YT
DO 645 I1P=2,NW
XW(I1P) =X1V(1)+ XW 1(I1P)
YW(IIP)=YW(1)+YWI(I1P)
645 CONTINUE
I0 RETURN
END
SUBROUTINE DECOMP(NROWS,IWW,A,U)
IMPLICIT real*8(A-H,G-Z)
DIMENSION A(1,50000), U(50000)
DO 650 I=1,NROWS*IWW
650 U(I)=0.0
DO 670 I=I,NROWS
IP = (I-1)*IWW+ 1
SUM=0.0
DO 655 K=1, I-1
IF ((I+1).GT.(K+IWW)) GO TO 655
SUM= SUM + (U((K-1)*IWW+ I-K+1))**2-
655 CONTINUE
U(IP)=SQRT(A(1,IP)- SUM)
DO 665 J= 2,IWW
SUM=0.0
DO 660 K= 1,1-1
IF ((I+J).GT.(K+IWW)) GO TO 660
SUM= SUM + (U((K-1)*IWW +/-K +I))*(U((K-1)*IWW+I-K +1))
660 CONTINUE
U(IP+.1-1)=(A(1,IP+J-1)-SUM)/U(IP)
665 CONTINUE
670 CONTINUE
RETURN
END
c***********************************************************************
SUBROUTINE SOLVE(NROWS,IWW,U,F,D)
DIMENSION U(50000),F(1000),D(I000),X(1000)
X(1)=F(1)/U(1)
DO 680 I =2,NROWS
SUM=0
DO-675 J=1, I-I
IF ((I-J).GE.IWW) GO "ID t,-75
SUM= SUM + U((J-1)*IWW + I + 1-1)*X(J)

675 CONTINUE
X(I)=(F(I)-SUM)/U(1+(I-1)*IWW)
680 CONTINUE
D(NROWS)=X(NROWS)/U((NROWS-1)*IWW+1)
DO 690 I=NROWS-1,1,-1 72
SUM =0
DO 685 J=2 ,IWW
SUM=SUM + U(IWW*(I-1)+J)*D(I+J-1)
685 CONTINUE 7:
D(1)=(X(1)-SUM)/U((I-1)*IWW +1)
690 CONTINUE
RETURN
END
c***********************************************************************
SUBROUTINE MODSTIF(IWW,IX,JDD,S,NROWS)
DIMENSION S(1,12000),JDD(100)
DO 705 I=IX,1,-1
J=JDD(I)
DO 700 IP = 1,J-1
IF((J-IP +1).GT.IWW) GO TO 700
IA=(IP-1)*IWW+(J-Ip+1)
I B =a p - i r r ww+ r ww
DO 695 1Q-=-IAJB-1
S(1,IQ)=S(1,IQ+1)
695 CONTINUE
S(1,IQ)=0
700 CONTINUE
705 CONTINUE
DO 715 I=IX,1,-1
J =JDD(I)
IA= (J-1)*IWW +1
IB=(NROWS-1 )*IWW
DO 710 IP =IA,IB
S(1,IP)=S(1,1P+IWW)
710 CONTINUE
NROWS=NROWS-1
715 CONTINUE
RETURN
END
SUBROUTINE SORT(IX,JDD,JDD1)
DIMENSION JDD(100),JDD1(100)
DO 720 I=1 ,IX
720 JDD1(I) =JDD(I)
IP=IX+1
Appendix II 235
DO 730 1=1,a-1
IP=IP-1
IK = 1
DO 725 J=1,IP
IF (JDD1(IK).GT.JDD1(J)) GO TO 725
IK
725 CONTINUE
EMP =.113D1(IP)
IDD1(IP)=./DD1(1K)
IDD1(1K) =EMP
730 CONTINUE
RETURN
END
c***************m*******************************************************
SUBROUTINE LDISTR (MN1,M N2,MN11, M N22, PLOAD , SANG,
+ BL,XPL,DLNG,XLL,YLL,POS,NL)
IMPLICIT real*8(A-H, 0-Z)
DIMENSION BL(I000)-
CHARACTER*2 POS
DO 735 I=1 ,3* NL
735 BL(I)= 0
C =XLL
B =YLL/COS(SANG)
D =XPL-XLL
A =(DLNG-YLL)/COS(SANG)
X =XPL
Y =DLNG/COS(SANG)
PE =PLOAD*A*A*(3*B +A)/Y**3
PF =PLOAD*B*B*(3*A+B)/Y**3
BMF =-PLOAD*A*B*B/Y/Y*COS(SANG)
BME=PLOAD*A*A*B/Y/Y*COS(SANG)
BMFI =-PLOAD*A*B*B/Y/Y*SIN(SANG)
BME1 = PLOAD *A*A*B/Y/Y*SIN(SANG)
BL(3*MN I )=PE*D*D*(3*C + D)/X**3-6*BMEI*D*C/X**3
BL(3*M N2) = PE*C*C*(3*D + C)/X**3 +6*BMEI*D*ca**3
BL(3*MN11)= PF*D*D*(3*C + D)/X**3-6*BMF1*D*C/X**3
BL(3*MN22)=PF*C*C*(3*D +C)/X**3 +6*BMF1*D*C/X**3
BL(3*MN1-1)=-(PE*C*D*D/X/X-BME1*D*(2*C-D)/X**2)
BL(3*MN2-1)= -(-PE*C*C*D/X/X-BMEl*C*(2*D-C)/X**2)
BL(3*MN11-1) --(PF*C*D*D/X/X-BMFI*D*(2*C-D)/X**2)
BL(3*MN22-1) =-(-PF*C*C*D/X/X-BMFI*C*(2*D-C)/X**2)
BL(3*MNI-2) =BME*D/X
BL(3*MN2-2) =BME*CIX
BL(3*MN I1 -2) =BMF*D/X
BL(3*M N22-2) = Bne*ca
RETURN
END
c***********************************************************************
236 Grillage Analogy in Bridge Deck Anidysis
SUBROUTINE LDISTT (MN1,MN2,MN22,PLOAD,XPL,DLNG,XLL,YLL,BL,POS,NL)

DIMENSION BL(1000)
CHARACTER*2 POS
DO 740 I=1,3*NL
740 BL(I)=0
A =XLL
D =YLL
C =XPL-XLL
B=DLNG-YLL
X=XPL
Y=DLNG
SINT=X/SQRT(X*X+Y*Y)
TANT=X/Y
COST=SINT/TANT
EF=A/TANT
PE=PLOAD*(EF-D)**2*(2*D+EF)/EF**3
PF=PLOAD*(3*EF-2*D)*D*D/EF**3
BMF=PLOAD*D*D*(EF-D)/EF/EF
BME=PLOAD*D*(EF-D)**2/EF/EF
BL(3*MN1)=(PE+PF)*C*C*(3*A+C)/X**3+6*A*C*BMF/X**3*SINT*COST
BL(3*MN2)=PE'iA*A*(3*C+A)/X**3
BL(3*MN22)=PF*A*A*(3*C+A)/X**3-6*A*C*BMF*SINT*COST/X**3
BL(3*MNI-1)=-(BMF*C/X*SINT*COST*(1+(2*A-C)/X)+PF*A*C*C/X/X
+ +PE*C**2*A/X**2)
BL(3*MN2-1)=PE*A*A*C/X/X
BL(3*MN22-1)=-BMF*A/X*SINT*COST*(1+(2*C-A)/X)+PF*C*A*AfX./X
BL(3*MN1-2)=C/X*(PF*A*C/XJTANT+BME+BMF*((2*A-C)/X*COST*COST
+ -SINT*SINT))
BL(3*MN2-2)=BME*A/X
13L(3*MN22-2) =-PF*C*A*A/X/X/TANT-BMF*A/X*(SINT*SINT-(2*C-A)/X
+ *COST*COST)
IF(POS.NE.'XG')GOTO 750
DO 745 1=1,2
BI.,(3*M N1-1) = -BL(3*MN1-1)
BL(3*MN24)=-BL(3*MN2-I)
745 BL(3*MN22-1)=-BL(3*MN22-I)
750 RETURN
END
C ***************************************,ie*************
SUBROUTINE IMPACT(LCASE,SPAN,SANG)
IMPLICIT REAL*8(A-H2O-
Z)' COMMON/XXUAIMP
real*8 L
WRITE(*, *)'SPAN = ',SPAN
WRITE(*,*)'ANGLE=',SANG
L=SPAN*COS(SANG)/1000 WRITE(*,*)'SPAN
FOR IMPACT FACTOR CALC. =' ,L
Appendix II 237
'IL) WRITE(4,755)L
755 FORMAT(10X,'SPAN FOR IMPACT FACTOR CALCULATION=
',F6.2) GO TO (1,1,3,3,3,3,1,1,9),LCASE
1 IF (L.LE.3.0)AIMP=1.5
IF (L.GE.45.0)AIMP=1.0875
IF (L.GT.3.0.AND.L.LT.45.0)AIMP 1.0+(4.5/(6.0+L))
RETURN
3 IF (LLE.12.0)AIMP=1.25
IF (L.GE.45.0)AIMP=1.0875
IF (L.LT.45.0.AND.L.GT.12.0)AIMP=1.0 +(4.5/(6.0+L))
RETURN
7 IF (L.LE.5.0)AIMP=1.25
IF (L.GT.39.0)GO TO 1
IF (L.GE.9.0)AIMP=1.10
IF (L.LT.9.0.AND.L.GT.5.0)AIMP=1.25-0.15*(L-5.0)/4.0
RETURN
9 A IM P = 1 . 0
RETURN
END
c***************************************************#*******************
SUBROUTINE DEAD LOAD (B,N,NLG,M,PM,E1,E2)
IMPLICIT REAL*8(A-H2O-Z)
real*8 K(100),B(1000),E1,E2
O
DIMENSION M(1000,2),PM(1000,1)
DO 760 I=1,3*N
760 B(I)=0.0
READ(3,*)CHOICE
IF(CHOICE.EQ.1)THEN
READ (3,*)(1C(1),I=1,2*NLG)
WRITE(4,795)
WRITE(4,800)(I,K(I*2-1),K(2*1),I=1,NLG)
WRITE(4,805)E1,E2
J=0
DO 765 I=1,N,NINLG
J=1+2
CALL INODE(M,I,I+1,I1)
AL=PM(I1,1)/2.
B(3*I)=B(3*I)+K(J-1)*AL
B(3*I-1)=B(3*I-1)-K(J-1)*AL*AL/3.
B(3*I-2)=B(3*12)+K(J)*AL
JI=I+N/NLG-1
CALL INODE (M,II-1,II,I1)
AL=PM(I1,1)/2.
B(3*II)=B(3*II)+K(J-1)*AL
B(3*II-1)=B(3*II-1)+K(J-1)*AL*AL/3.
B(3*II-2)=B(3*II-2)+K(.1)*AL
765 CONTINUE
DO 785 i=2.INI
DO 770 J=1,NLG
IF (I.EQ.N/NLG*J)G0 TO 785
IF (I.EQ.N/NLG*J+1)G0 TO 785
770 CONTINUE
DO 775 J=1,NLG
IF (I.LT.N/NLG*J)G0 TO 780
775 CONTINUE
780 CALL INODE(M,I-1,I,I1)
CALL INODE(M,I,I+1,12)
AL= (PM(I1, 1) +PM(I2, 1))/2.
ALl = PM(I1,1)/2.0
AL2 = PM(I2,1)/2.0
B(3*1)=B(3*I) +K(J*2-1)*AL
B(3*I-1)=B(3*I-1)-K(J*2-1)*(-PM(I1,1)**2+PM(12,I)**2)/12
B(3*I-2)=B(3*I-2)+K(J*2)*AL
785 CONTINUE
DO 790 I=1,NLG
B(3*((1-1)*N/NLG+I))=El*K(2*I-1)+B(3*((I-1)*N/NLG+1))
790 B(3*N*IINLG)=E2*K(2*I-I)+B(3*N*I/NLG)
ELSE
READ(3,*)(B(3*I),I=1,N)
WRITE(4,8I0)
WRITE(4,815)(I,B(3*1),I=1,N)
ENDIF
795 FORMAT(/5X,'DEAD LOAD ALONG LONGITUDINAL GRID LINESV5X,39c*'),
+ I/5X,'LONG. GRID',5X,'VERTICAL LOAD',5X,'TORS1ONAL MOMENT'/5X,
+ 'LINE NO.',13X,'TIMM',14X,'T MMTMW/60('*')/)
800 FORIvIAT(8X,I3,9X,E11A,9X,E11.4)
805 FORMAT(!/'THE END PROJECTIONS OVER SUPPORT LINES ARE ',E1I.4,
+' & A,' MM'/' RESPECTIVELY.'//)
810 FORMAT(/5X,'DEAD LOAD AS SUPPLIED,DIRECTLY ON NODES,VERTICAL'
+ ,'. LOAD ONLY',/5X,30('*')//5X,'NODE NO.' ,10X,'LOAD(T)'
+ /30('*'),/)
815 FORMAT((6X,I3,I3X,E11.4))
RETURN
END
c***********************************************************************
SUBROUTINE COMPARE1(PT,B,RT,MAXPT,MAX13,MAX.RT,MAX RTN,MM,N
+ ,IX,IMI,PT1,X,Y
+ ,LCASE,CORSPT,CORX,CORY,CORX1,CORY1,CORRX,CORRY,CORD1
+ ,JTT)
IMPLICIT REAL*8(A-F1,0-Z)
DIMENSION PT(1000,7),B(1000),RT(100),PT1(1000,2),JTT(100)
REAL *8 MAXPT(6,4,500),MAXB(4,500),MAXRT(4,100),MAXRTN(4,100)
+ ,CORSPT(2,500),CORX(100),CORY(100),CORX1(100),CORY1(100)
+ C ORRX(2,500). CORRY(2 .500) CORD I
(100) IF (1M1.GT.1)G0 TO 850
Appendix II 239
DO 820 I=1,4
1 =I +2
DO 820 K=I,MM
MAXPT(J, 1,K) = PT(K, J)
MAXPT(I,2,K)=LCASE
MAXPT(L3,K) = X
MAXPT(I,4,K)=Y
CORSPT(I,K) =PT(K,4)
CORSPT(2,K)=PT(K,3)
820 CONTINUE
KL
DO 825 I=1,N
N11 =3*I-2
N13 =3*I .
CORRX(1,I) -=13(N11 +1)
CORRX(2,I)=B(N13)
CORRY(1,I) =B(N 11)
CORRY(2,I)=B(N13)
DO 825 J=N11,N13
MAXB(1,J)=B(J)
MAXB(2,J)=LCASE
MAXB(3 ,J) =X
MAX13(4,J)=Y
825 CONTINUE
), DO 830 I=1,IX
MAXRT(I ,I) =RT(I)
MAXRT(2,I) =LCASE
MAXRT(3,I) =X
MAXRT(4,I) =Y
CORXI (I) = B(JTT(I)*3 -2)
CORYI (I) =B(JTT(I)*3-1)
CORD1(1) = B(ITT(I)*3)
MAXRTN(1,I) =RT(I)
MAXRTN(2,I)=LCASE
MAXRTN(3,I) =X
MAXRTN(4,I) =Y
WRITE(*,*)LJTT(I),B(JTT(I)*3-2),B(JTT(1)*3-1)
CORX(I)=B(JTT(1)*3-2)
CORY(I)=B(JTT(I)*3-1)
830 CONTINUE
DO 835 1=5,6
DO 835 K=1,MM
J=1-4
MAXPT(I, 1,K) = PT1(K,J)
MAXPT(1, 2,K) = LCASE
MAXPT(1,3,K)= X
MAXPT(I,4,K)=Y
835 CONTINUE
DO 845 1=3,6
DO 845 K=1,MM
IF (1.GT.4)G0 TO 840
IF (MAXPT(I,1,K).LT.0.0)THEN
MAXPT(I,1,K)=0
MAXPT(I,2,K)=0
MAXPT(I,3,K)=0
MAXPT(I,4,K) =0
ENDIF
GO TO 845
840 IF (MAXPT(I,1,K).GT.0.0)THEN
MAXPT(L1,K)=0
MAXPT(I,2,K)=0
MAXPT(I,3,K)=0
MAXPT(I,4,K) =0
ENDIF
845 CONTINUE
RETURN
850 DO 855 1=1,2
J=I+2
DO 855 K=1,MM
IF (ABS(MAXPT(I,1,K)).GE.ABS(PT(K,J)))GO TO 855
MAXPT(I, 1,K) = PT(K,J)
MAXPT(I,2,K) =LCASE
MAXPT(1,3,K)=X
MAXPT(I,4,K) =Y
CORSPT(I,K)=PT(K,54)
855 CONTINUE
DO 860 1=3,4
J=1+2
DO 860 K=1,MM
IF (MAXPT(I,1,K).GE.PT(K,J))GO TO 860
MAXPT(1,1,K)=PT(K,J)
MAXPT(I,2,K) =LCASE
MAXPT(1,3,K)=X
MAXPT(1,4,K)=Y
860 CONTINUE
DO 865 1=5,6
DO 865 K=1,MM
J=I-4
IF (MAXPT(1,1,K).LT.PT1(K,J))G0 TO 865
M A'XPT(I , 1 ,K) = PT1(K,J)
MAXPT(1,2,K)=LCASE
MAXPT(1,3,K)=X
MAXPT(1,4,K)=Y
865 CONTINUE
KL1 = I
DO 870 I=1,N
Appendix II 241
N 1 =3*1-2
IF (ABS(MAX13(1,N11)).GE.ABS(B(N11)))G0 TO 870
MAXI3(1 ,N 1) =B(NI1)
MAXI3(2,N11)=LCASE
MAX13(3,N11)=X
MAX13(4,N11)=Y
CORRX(1 ,I) = B(N1 I +1)
CORRX(2,I)=B(N11+2)
870 CONTINUE
KL1 =1
DO 875 I=1,N
N12=3*1-1
IF (ABS(MAXB(1,N12)).GE.ABS(B(N12)))G0 TO 875
MAXB(1,N12) =-13(N12)
MAX13(2,N12)=LCASE
MAXB(3,N12)=X
MAX13(4,N12)=Y
CORRY(1,I)=B(N12-1.)
CORRY(2,1)=B(N12+1)
875 CONTINUE
DO 880 I=1,N
N13 =3*I
IF (ABS(MAX13(1,N13)).GE.ABS(B(N12)))G0 TO 880
MAXB(1,N13)=B(N13)
MAXB(2,N13)=LCASE
MAXB(3,N13)=X
MAXB(4,N13)=Y
880 CONTINUE
DO 885 I=1,IX
IF(MAXRT(1,1).GT.RT(I))G0 TO 885
MAXRT(1,I)=RT(I)
MAXRT(2,I) =LCASE
MAXRT(3 ,I) =X
MAXRT(4,I)=Y
CORX1(I) = B (JTT(I)*3 -2)
CORY1(I) = B (JTT(I) *3-1)
CORD1(I)=B(ITT(I)*3)
885 CONTINUE
DO 890 I=1,IX
IF (MAXRTN(1,I).LT.RT(I)) GO TO 890
MAXRTN(1,I)=RT(I)
MAXRTN(2,I)=LCASE
MAXRTN(3,I)=X
MAXRTN(4,I)=Y
WRIT E(*,*)i,yrixo, B(JTT(1)*3-2),B(JTT(1)*3-1)
CORX(I)=B(ITT(I)*3-2)
CORY(I)=B(JTT(1)*3-1)
890 CONTINUE
RETURN
END
c***********************************************************************
SUBROUTINE COMPARE2(PT,B,RT,RTN,MAXPT,MAXB,MAXRT,MAXRTN,
+ MM,N,IX,IM1,PT1,
+ MAX PT2 ,MAXB2,MAXRT2;MAXR2N,JTT ,COR2X,COR2Y,C OR2XI ,COR2Y1 ,
+ COR1X COR I Y, COR1XI ,CORlYI,CORSPT1,CORSPT2,
+ CORRRX I , CORRX2 ,CORRY I ,CORRY2 , CORD,CORD2)
IMPLICIT REAL*8(A-H2O-Z) 90
DIMENSION PT(1000,7),B(1 000), RT(100),RTN(100),PT I (1000, 2),
JTT(100)
REAL*8 MAXPT(6,4,500),MAXB(4,500),MAXRT(4,100)
+ ,MAXPT2(6,2:4,500)
+ ,MAXB2(2:4,500),MAXRT2(2:4,100),MAXRTN(4,100),MAXR2N(2:4,100)
+ ,COR2X(100), COR2Y(100),COR2X1(100) ,COR2Y1( I00) ,CORSPT1 (2 ,500)
+ ,CORSPT2 (2,500),COR1X( I00),CORI Y(100),COR1X1(100),COR1Y1 (100)
+, CORRX1(2,500), CORRX2(2,500) ,CORRY1(2,500) ,CORRY2 (2,500) 9
+, CORD( I00),CORD2(100)
IF (Im1.NE.1)G0 TO 925
DO 895 I =1,4
J=I+2
DO 895 K = 1,MM
MAXPT(I, I ,K) = PT(K,J)
MAXPT(I,2,K)=MAXPT2(1,2,K)
MAXPT(I,3,K) = M AXPT2 (I ,3,K)
MAXPT(I,4,K) =MAXPT2(I,4,K)
CORSPT1 (1 ,K) = CORSPT2 (1,K)
CORSPT I (2 ,K) = CORSPT2(2,K)
895 CONTINUE
DO 900 I=1,N
NI I =3*I-2
N13 =3*1
DO 900 J=N11,N13
MAXB(1,J) = B(J)
MAXB(2,J) = M AXB2(2 ,J)
MAXB(3,J) MAXB2(3 ,J)
MAX B(4, J) = M AXB2 (4, J)
CORRX1(1 ,I) =CORRX2(1,I)
CORRY I (1 ,I) =CORRY2(I,I)
C ORRX I (2 , I) = CORRX2 (2,I)
C ORR Y I (2,D= CORRY2(2,I)
900 CONTINUE
DO 905 I = 1,IX
MAXRT(1,1) = RT (I)
M AXRT(2 , I) = M AXRT2 (2,1)
M AXRT (3 , I) = MAXRT2 (3,1)
M AXRT (4, I) = MAXRT2(4,I)
COR1XI(I)=COR2X1(I)
Appendix II 243
CORI Y f(I) =COR2Y1(1)
CORD(I)=CORD2(I)
MAXRTN(1,I) =RTN(I)
MAXRTN(2,I) =MAXR2N(2, I)
MAXRTN(3,I)=MAXR2N(3,I)
MAXRTN(4,I)=MAXR2N(4,1)
COR1X(I)=COR2X(I)
CORIY(I)=COR2Y(I)
905 CONTINUE
DO 910 1=5,6
DO 910 K=1,MM
J=I-4
MAXPT(L1,K)=PT1(K,J)
MAXPT(I,2,K) =MAXPT2(I,2,K)
MAXPT(1,3,K)=MAXPT2(1,3,K)
MAXPT(L4,K)=MAXPT2(I,4,K)
910 CONTINUE
DO 920 I=3,6
DO 920 K=1,MM
IF (I.GT.4)G0 TO 915
IF (MAXPT(I,1,K).LT.0.0)THEN
MAXPT(I,1,K)=0
MAXPT(I,2,K)=0
MAXPT(I,3,K)=0
MAXPT(1,4,K)=0
ENDIF
GO TO 920
915 IF (MAXPT(1,1,K).GT.0.0)THEN
MAXPT(I,1,K)=0
MAXPT(I,2,K) =0
MAXPT(1,3,K)=0
MAXPT(I,4,K)=0
ENDIF
920 CONTINUE
RETURN
925 DO 930 I=1,2
.1= I + 2
DO 930 K=I,MM
IF (ABS(MAXPT(I,I,K)).GT.ABS(PT(K,J)))GO TO 930
MAXPT(I,1,K)=PT(K,J)
MAXPT(1,2,10=MAXPT2(1,2,K)
MAXPT(I,3,K)=MAXPT2(I,3,K)
CORSPT1(I,K)= CORSPT2(I,K)
930 CONTINUE
DO 935 1=3,4
J=1+2
DO 935 K=i,iviNi
244 Grillage Analogy in Bridge Deck Analysis4
IF (MAXPT(I,1,K).GT.PT(K,D)G0 TO 935
MAXPT(I,1,K)=PT(K,J)
MAXPT(I,2,IC) =MAXPT2 (I, 2, K)
MAXPT(1,3,K)=MAXPT2(I,3,K)
935 CONTINUE
DO 940 1=5,6
DO 940 K=1,MM
J=I4
IF (MAXPT(I,1,K).LT.PT1(K,i))GO TO 940
MAXPT(I, I ,K) =PT1(K,J)
MAXPT(1,2,K)=MAXPT2(1,2,K)
MAXPT(1,3,K) =MAXPT2(1,3 ,K)
MAXPT(1,4,K)=MAXPT2(I,4,K)
940 CONTINUE
KL1=1
DO 945 I=1,N
N11=3*1-2
IF (ABS(MAXB(1,N11)).GT.ABS(B(N11)))G0 TO 945
MAX'S(1 ,N11)= B(N I 1)
MAXB(2, N11) = MAXB2(2, N11)
MAXB(3,N11)=MAXB2(3,N11)
MAXB(4,N1I)=MAX132(4,N11)
CORRX1(1,1)=CORRX2(1,I)
CORRX1 (2 ,I) CORRX2(2 ,I)
945 CONTINUE
DO 950 1=1,N
N12=3*1-1
IF (ABS(MAXB(1,NI2)).GT.ABS(B(N12)))GOTO 950
MAXB(1,N12)=B(N12)
MAXB (2, N12) = MAXB2(2 , N 12)
MAXB(3,N12)=MAX132(3,N12)
MAXB(4,N12) = MAXB2(4,N 12)
CORRY I (1 ,1) = CORRY2(1, I)
CORRY I (2,I) = CORRY2(2, I)
950 CONTINUE
DO 955 I=1,N
N13 =3*I
IF (ABS(MAXB(1,N13)).GT.ABS(B(N13)))GOTO 955
MAXB(1,N13)=B(N13)
MAXB(2,N13)=MAXB2(2,N13)
IVIAXB(3 , N13) =MAXB2(3 , N13)
MAXB (4 , N13) = MAXB2(4, N13)
955 CONTINUE
DO 960 I=1,IX
IF(MAXRT(1,I).GT.RT(I))GO TO 960
MAXIIT(11)=RT(T 1
MAXRT(2,I)=MAXRT2.(2,I)
Appendix11 245
MAXRT(3,I)=MAXRT2(3,I)
MAXRT(4,I)=MAXRT2(4,I)
COR1XI(I)=COR2X1(1)
COR1Y1(I)=COR2Y1(1)
CORD(I) =CORD2(I)
960 CONTINUE
DO 965 I=1,1X
IF(MAXRTN(1,11.1-T.RTN(1))00, TO 965
MAXRTN(1,I)=RTN(I)
MAXRTN(2,I)=MAXR2N(2,I)
MAXRTN(3,I) =MAXR2N(3,I)
MAXRTN(4,I)=MAXR2N(4,i)
COR1X(I)=COR2X(I)
COR1Y(1)=COR2Y(1)
965 CONTINUE
' RETURN
END
c***************************************************************i*******
SUBROUTINE WRITE2 (MAXPT,MAXB,MAXRT,MAXRTN,MM,N,IX,MOVE,M,
+ JTT,LCASE,COR1X,CORlY,CORIX1,COR1Y1,CORSPT1,CORRX1,CORRY1
+ ,CORD)
IMPLICIT REAL*8(A-H2O-Z)
DIMENSION M(1000,2);JTT(100)
real*8 MAXPT(6,4,500),MAXB(4,500),MAXRT(4,100),MAXRTN(4,100)
+ ,CORIX(100),COR1Y(100),COR1X1(100),CORIY1(100),CORSPTI(2,500)
+,CORRXI(2,500),CORRY1(2,500),CORD(100)
WRITE (4,1025)
DO 970 I=1,MM
WRITE (4,1030)I,M(L1),M(L2),(MAXPT(K,1,0,CORSPT1(K,1),
+(MAXPT(K,J,1),J= 2,4),K = 1 ,2)
970 CONTINUE
WRITE (4,1015)
WRITE (4,*)'MAX SAGGING MOMENT ON THE ELEMENTS'
WRITE (4,1040)
DO 975 I=1,MM
975 WRITE (4,1035)I,M(I,I),M(I,2),(MAXPT(5,7,1),J= I,4),(MAXPT(4,5,1)
+ ,J=1,4)
WRITE (4,1015)
WRITE (4,*)'MAX HOGGING MOMENT ON THE ELEMENTS'
WRITE (4,1040)
DO 980 I= I,MM
980 WRITE .(4,1035)I,M(I,1),M(1,2),(MAXPT(3,J,I),J = 1,4), (MAXPT(6,1,1)
+ ,J = 1,4)
READ (3,*) IK4
IF (IK4.NE.1) GO TO 990
WRITE (4,11345)LNAME,LNAME,LNAME
DO 985 I=1,N
N11=3*1-2
N12 =31-1
N13 = 3*I
WRITE (4,1050)1,1, (MAX.13(J,N11),J =1,2),(MAX13(1,NI2),J= 1,2),
+ (MAXB(J,N13),J = 1,2)
985 CONTINUE
990 CONTINUE
WRITE (4,1060)
DO 995 I=1,LX
WRITE (4,1065)LITT(1),-MAX.RTN(1,I),COR1X(I),CORlY(I)
+ ,MAXB(I,3*JTT(I)),(MAXRTN(J,I),J =2;4)
995 CONTINUE
WRITE (4,1020)
WRITE (4,1055)
DO 1000 I=1,IX
WRITE (4,1065) LITT(1),-MAXRT(1,1),COR1X1(1),CORIY1(1),CORD(1)
+ ,(MAXRT(I,I),J=2,4)
1000 CONTINUE
WRITE (4,1075)
DO 1005 I=1,IX
N1I=3*.ITT(1)-2
WRITE (4,1085)LITT(1),MAXB(1,N11),CORRX1(1,JTT(1)),
+ C ORRX1(2, ITT(I)),(MAXB(J, N11), I =2,4)
1005 CONTINUE
WRITE (4,1080)
DO 1010 I=1,IX
N12=34TIT(1)-1
WRITE (4,1085)I,JTT(I),MAXB(1,N12),CORRY1(1,ITT(I)),
+CORRY1(2,Ms(I)),(MAXB(J,N12),J=2,4)
1010 CONTINUE
WRITE (4,1020)
1015 FORMAT(/80('*'))
1020 FORMAT(/80('*'))
1025 FORMAT(//,'MAX. SHEAR AND TORSION ON THE ELEMENTS'/
+110('*')/'SNO ELEMENT MAX SHEAR TORSION'
+,' LCASE X Y '
+ ,' MAX TORSION SHEAR',
+ ' LCASE X Y'/19X,'T TMM',12X,'MM MM',5X,
+ TMM T',14X,'MM MM'/110('*')///)
1030 FORMAT(I3,3X,I2, I2,2(2X, E11.4, IX, E11.4,
1X +,F3.0,2X,F6.0,2X,F6.0))
1035 FORMAT(13,3X,I2,'-',12,2(2X,E11.4,1X,F3.0,2X,F6.0,2X,F6.0))
1040 FORMAT(80('*')PSNO ELEMENT AT END1 LCASE X Y '
+ ,' AT END2 LCASE X Y'/18X,'TMM',13X,'MM MM',5X,
+ 'TMM',14X,'MM MM'/80('*')/) 1045
FORMAT(99('*')/'SNO',2X,'JOINT',3X,
+ 'MAX X-ROTATION',1X,1A8,2X,'MAX Y-ROTATION',1X,1A8,2X,
+ 'MAX Z-DEFLECTION',1X,1A8,/22X,'RAD'.24X,'RAD',27X,
'MM',/99('*')/)
Appendix II 247
1050 FORMAT(I3.4X,13,2(4X,E13.6,5X,F6.0),5X,E13.6,6X.
+F6.0)
1055 FORMAT(/,'MINIMUM REACTION ON SUPERSTRUCTURE/K('*')
+PSNO JOINT REACTN COR ROT-X COR ROT-Y
+ CO R D EFL LC A S E X r/
+ 80('*')/)
1060 FORMAT(//,`MAXIMUM REACTION ON SUPERSTRUCTURE/80(n
+PSNO JOINT REACTN COR ROT-X COR ROT-Y COR DEFL
, ' LC A S E X Y '/
+ 80('*')/)
1065 FORMAT(I3,2X,I3,2X,4(E11A,1X),F3.0,2X,F6.0,2X,F6.0)
1070 FORMAT(13,2X,13,2X,3(E11.4,1X),F3.0,2X,F6,0,2X.F6.0)
1075 FORMAT(80(w),/,' SNO JOINT MAX ROT X COR ROT Y '
+ COR DEFL LCASE X Y 'PUNITS IN T AND mm'/
+ 80(P')/)
1080 FORMAT(80C*),/,' SNO JOINT MAX ROT Y COR ROT X '
+,' COR DEFL LCASE X Y '/
+ 80('*')/)
1085 FORIvIAT(I3,2X,I3,3(3X,E11.4),2X,F3.0,2X,F6.0,2X,F6.0)
1090 FORMAT(I3,2X,I3,2X,E11.4,6X,F3.0,2X,F6.0,2X,F6.0)
RETURN
END
c**********************************************************************
SUBROUTINE WRITE1(13,MAXPT,MAXB,IvIAXRT,MAXRTN,MM,N,IX,MOVE,M,
TIT,LCASE,CORSPT,CORX,CORY,CORX1,CORYLCORRX,CORRY,CORD1)
IMPLICIT REAL *8(A-H2O-2;)

DIMENSION M(1000,2),JTT(100),B(1000)
real*8 MAXPT(6,4,500),MAXB(4,500),MAXRT(4,100),MAXRTN(4,100)
+,CORSPT(2,500),CORX(100),CORY(100),CORX1(100),CORY1(100)
+,CORRX(2,500),CORRY(2,500),CORDI(100)
IF (LCASE.NE.1) GO TO 1115
WRITE(4,1245)
DO 1095 I=1,MM
WRITE(4,1260)1,M(I, 1),M(1,2),(MAXPT(K,1,I),K= I ,2),LCASE
1095 CONTINUE
WRTTE(4,1165)
WRITE(4.1250)
DO 1100 I=1,MM
WRITE(4,1265)I,M(1,1),M(1,2),MAXPT(5, 1 , I),MAXPT(4, 1,D,LCASE
1100 CONTINUE
WRITE(4,1165)
WRITE(4,1255)
DO 1105 I= LIN.EvI
WRITE(4,1265)I,M(1,1),M(1,2),MAXPT(3,1,D,MAXY19(6,1,I),LCASE
1105 CONTINUE
WRITE(4,1165)
WRITE(4,1240)
DO 1110 1=1,1X
WRITE(4,1270)LJTT(I),-MAXRTN(I,I),(MAXB(I.J).
+ J=3*-TIT(I)-2,3*JTT(0),LCASE
1110 CONTINUE
WRITE(4,1165)
GO TO 1275
Ills WRITE (4,1175)
DO 1120 I=1,MM
WRITE (4,1180)I,M(I,1),M(I,2),(MAXPT(K,1,1),CORSPT(K,I),
+(MAXPT(K,J,1),J=2,4),K=1,2)
1120 CONTINUE
WRITE (4,1165)
WRITE (4,*)IMAX SAGGING MOMENT ON THE Fr PMENTS'
WRITE (4,1190)
DO 1125 I=1,MM
1125 WRITE (4,1185)1,M(I,1),M(1,2),(MAXPT(5,J,I),J=1,4),(MAXPT(4,J,I)
+,J=1,4)
WRITE (4,1165)
WRITE (4,*)'MAX HOGGING MOMENT ON THE ELEMENTS'
WRITE (4,1190)
DO 1130 I=1,MM
1130 WRITE (4,1185)1,M(1,1),M(1,2),(MAXPT(3,J,I),J=1,4),(MAXPT(6,J,1)
+,J=1,4)
READ(3,*) IKI
IF (1K1.NE.1) GO TO 1140
WRITE (4,1195)LNAME,LNAME,LNAME
DO 1135 I=1,N
N11=3*1-2
N12=3*1-1
N13=3*I
WRITE (4,1200)I,I, (MAX13(J,N I 1),J=1,2),(MAXE (J,N12),J=1,2),
+(MAX13(1,N13),J=1,2)
1135 CONTINUE
1140 CONTINUE
WRITE (4,1210)
DO 1145 I=1,IX
WR1TE(4,12 I5)I,M(1),-MAXEITN(1,1),CORX(1),CORY(1)
,MMCB(1,3*JTT(1)),(MAXRTN(J,1),J=2,4)
1145 CONTINUE
WRITE (4,1170)
WRITE(4,1205)
DO 1150 I=1,1X
WRITE (4,1215) 1,JTT(I),-MAXRT(1,1),CORX1(1),CORY1(1).CORD1(I)
+ ,(MAXRT(J,I),J=2,4)
1150 CONTINUE
WRITE (4,1220)
DO 1155 I=I,DC
JI =3*.ITT(1)-2
WRITE (4,1230)I,ITT(1),MAXB(1,11),CORRX(I,ITT(I)),CORRX(2,TIT(1))
t(MAXE3(J,J1),J=2,4)
1155 CONTINUE
WRITE (4,1225)
DO 11601=1,1X
J13491T(1)-1
WRITE (4,1230)1,JTT(I),MAXB(I,J1),CORRY(1,JTr(1)),CORRY(2,TrraD
+,(MAXB(J,J1),J=2,4)
11Ni CONTINUE
WRITE (4.1170)
1165 FORMAT(/100(*))
1170 FORMAT(/80('4'))
Appendix II 249
1175 FORMAT(//,'MAX. SHEAR AND MAX. TORSION ON THE ELEMENTS'/

+ 1 IO('*')PSNO ELEMENT MAX SHEAR TORSION'
+,' LCASE X Y '
+ ,' MAX TORSION SHEAR'.
+' LCASE X Y719X,'T 'FM1v1',12X,'W MM',5X,
+' TMM T,14X,'MM MKT/110( 1 *V)
7 1180 FOFtMAT(L3,3X,I2,'-',I2,2(2X,E11.4,1X,E11.4,1X,
+ F3.0.2X,F6.0,2X,F6.0))
1185 FORMAT(13,3X,I2,'-',12,2(2X,E11.4,1X,F3.0,2X,F6.0,2X,F6.0))
1190 FORMAT(SOMPSNO ELEMENT AT ENDI LCASE X Y '
+,' AT END2 LCASE X Y'/18X.,'IMM 1 ,13X,MM MM',5X,
+ 'TMM,14X,' MM MM1/80('*')/)
1195 FORMAT(99(n/2X,'SNO',2X,IJOINT,3X,
+ 'MAX X-ROTAIION;IX,1A8,2X,'MAX Y-ROTATION',1X,1A8,2X,
+ 'MAX Z-DEFLECTION',1X,1A8,/22X;RAD',24X,RAD',27X,
+ TM3/1.199C*)/)
1200 FORMAT(13,4X,13,2(4X,E13.6,5X,F6.0),5X,E13.6,6X,
'+ F6.0)
1205 FORMAT(/,'MINIMUM REACTION ON
SUPERSTRUCTURE' 4/80('*')PSNO JOINT REACTN COR ROT-X
COR ROT-Y' +,' COR DEFL LCASE X
+,11X,RAD',9X,RAD',9X.,'nme,11X,Irtute,6X,'mne/ 80(1*')//)
1210 FORMAT(//,'MA)IMUM REACTION ON SUPERSTRUCTURE'
+/80(41PSNO JOINT REACTN COR ROT-X COR ROT-Y '
+ COR DEFL LCASE X Y'/' T'
+111X,'RAD',9X,TAD',9X;rarre,11X,'Enmi,6X,'mm? 80('*')/)
1215 FORMAT(13,2X,13,2X,4(E11.4,1X),F3.0,2X,F6.0,2X,F6.0)
1220 FORMAT(80C*),/,' SNO JOINT MAX ROT X COR ROT Y '
+ COR DEFL LCASE X Y '/
+16X,'RAD`,12X,RAD',11X,'nuns,13X,smm',5X,'mm./ 80(41//)
1225 FORMAT(80(n,/,' SNO JOINT- MAX ROT Y COR ROT X '
+ COR DEFL LCASE X Y '/
+16X;RAD',12X,RAD',11X.Anm1,13X,'ma,5X,'nuni 80 C nil)
1230 FORMAT(13,2X,I3,3(3X,E11.4),2X,F3.0,2X,F6.0,2X.F6.0)
1235 FORNLAT(13,2X,13,2X,E11.4,6X,F3.0,2X,F6.0,2X,F6.0)
1240 FORMAT(//,'REACTION ON SUPERSTRUCTURE AT SUPPORTS'
+/80('*')PSNO JOINT REACTN ROT - X ROT - Y '
+ ,' DEFLECTION LCASE/17X.7,8X,'RAD',10X,'RAD'
+,9X'nulf/80('*')/)
1245 FORMAT(1/,'SHEAR AND TORSION ON ELEMENTS'/
+ 80(*VSNO ELEMENT SHEAR TORSION LCASE'/
+ 19XT,10X,I'mm1/80(1*')/)
1250 FORMAT(//,'SAGGING MOMENT ON ELEMENTS'
+/80('*')/SNO ELEMENT AT ENDI AT END2 LCASE'I
+ 18X,Tmm.',8X,'Tnine/80('*')/)
1255 FORMAT(//,'HOGGING MOMENT ON ELEMENTS'
+/80('*')PSNO ELEMENT AT END1 AT END2 LCASE'/
+ 18X,Trran',8X,Tmrn'/80('*')/)
1260 FORMAT(13,3X,I2,'-',I2,2X,E I X.I3 )
1265 FORMAT(13,3X,I2,'-',12,2X,E1 L 4.1X,E11.4.2X.13 ) 1270
FORMA.T(3,2X,13,3X4(Eii.4, X), iX,I3) 1275 CONTINUE
RETURN
END
F.
Subject Index
Analysis, 35, 111, 131- Clearance, 20, 22-24
139 Arbitrary planforms, Codes of practices (see bridge
16 Assumptions, 36, 56 codes) Composite construction, 10
Axes, 42, 46 Coarse mesh, 41, 171
Axle loading, 19 Contributory area (see tributary area)
Computer aided methods, 1, 3
Band-width, 51, 64 Computer programs
Beam GRID, 32, 66-69, 205-212
contiguous beam, 7, 87 GABS, 32, 146-1,69, 171-
effective flange width, 96, 101 203, 213-250
equilibrium equations, 74 Contiguous beams (see beams)
0
spaced beam, 14 Contiguous nodes, 115, 117
spine beam, 14 Courbon's method, 1, 36-38
Bearing stiffness, 133 Curved deck, 42
Bearings Curved geometry, 16
neoprene, 16, 32, 133
rocker, 16 Data
roller, 16 input, 150, 165
Bending Moment diagram, 140, 143 output, 150, 165
Box-girder bridges, 6, 12-15, 92-94, Dead load, 18, 112, 118
103-109, 143-146 Degrees of freedom, 46, 56-58
Bridge classification, 5 Design curves, 1, 39, 43, 44
Bridge codes, 5, 17, 18 Design envelopes, 31, 138
Bridge deck analysis (see analysis) Diaphragms, 10. 13
Bureau of Indian Standards (BIS), 17 Direct stiffness method, 31, 55-73, 132
Displacement method, 56
Cable stayed bridges, 3, 6 Distortion, 13, 93, 108. 144
CAD/CAM systems, 4 Ductility-ratio, 4
Cantilever bridges, 16
Cantilever construction, 5 Edge grid lines, 87, 89
Cell distortion, 13, 93, 108, 144 Effective width of flange, 96, 101
Cellular deck, 6, 51, 103 Elastic properties, 94-109
Cement and concrete association (C & Envelope diagrams, 134, 138, 196
CA), S Equiiihrium equations. 74
Cholesky's factorisation Equivalent loads (see transfer of
method. 31, 69, 133 loads)
252 Subject Index
Examples, 69-73, 171-203 Idealization

Expert system, 4 bridge deck, 74-94
External prestressing, 3 loading, 117-131
Impact load, 25-27, 114
Fine mesh, 42, 171 Inclined webs, 93
Finite difference method, 1, 40-42 Indian Railways, 17
Finite element Method (FEM), Indian Roads Congress (IRC), 2, 17,
1, 48, 49, 200 . 30, 36, 96, 113
Finite strip method, 1, 50-52 Influence surfaces, 42
Flexibility matrix,55 Interpretation of results, 139-145
Flexible support, 16 Isotropic decks, 41-76
Flexural moment of inertia, 31, 38,
94-109 Kerb, 18, 19, 26, 28, 113, 162
Flexural parameter, 39, 43
Flowcharts, 66-68, 150-160 Launching, 5
Folded plate analysis, 47 Limit state concept, 4
Footpaths, 28, 113, 118 Lines of strength, 78
Force method,. 56 Live load
Force responses, 31, 32, 111, 134 AASHO, 28, 147
FORTRAN, 31, 32., 69, 146 IRC, 18-25, 28-30
Fourier analysis (see harmonic OHBD, 147, 191
analysis) Loading (see dead load, live load,
impact load)
Gauss-Elimination procedure, Load distribution, 35, 43, 99
31, 66, 133 Loading standards, 17, 24, 28, 30
Gauss-Seidel method, 65 Local effects, 111
Grade separation, 12 Longitudinal grid lines, 31, 78
Grillage examples
box-girder deck, 143, 199-203 Matrix method of analysis, 31, 55 M-
grid deck, 69-73. beam, 8
skew deck, 69-73, 179-182 Method of harmonic analysis (see
slab deck, 172-187 Hendry-Jaeger method)
slab-on-girders deck, 187-198 Ministry of surface transport, 17,
voided slab deck, 183-187 193 Ministry of transport, UK, 8
Grillage section properties Moment of inertia (see flexural
box-girder-deck, 103-109 moment .of inertia)
cellular deck (see box-girder deck) MS-DOS, 146
skew deck, 101 Multi-cellular decks, 13, 106
slab deck, 99-101
slab-on-girders deck, 101-103 Neutral axis, 99
Nodal deformations, 66, 131, 133
Harmonic components, 39, 42, 44 Nodal loads, 31, 69, 117
Hendry-Jaegar method, 42-45
design graphs, 43-44 Ontario highway bridge design code,
interpolation function, 44 147, 172. 191
Highway bridge decks, 5, 18 Orthotropic plate theory, .3`6-4.,)
Hypothetical loading system, 18, 144
Subject Index 253
coupling rigidities, 39 Skew grid, 31, 42, 69-73, 80
equations, 39 Skyline technique, 64
flexural parameter, 39 Slab bridges, 6-9, 79-89, 99-101,
torsional parameter, 39 139-142
Orthotropic slab, 38-40, 100,140 Slab-on-girders bridges, 6, 9-12, 90-
92, 101-103, 143
Panels Slope deflection, 74, '75
parallelogram, 115, 121, 128-131 Solid slab, 79, 80, 90, 99, 146
rectangular, 115, 119, 123-125 Spaced box-girders, 94 Space
triangular, 115, 119, .121, 125-129 frame analysis, 46 Spine
Parapets, 18, 26 beam bridges, 14 Spine box-
Partial prestressing, 3 girders, 14, 94, 146
Patch loads, 18, 50, 114 Stiffness method (see direct stiffness
Physical deck, 74, 75, 99, 111 method)
Plan geometry, 5, 15, 31, 35 Subroutines (see computer programs)
Planforms, 2, 5, 15, 69 Support conditions, 2, 5, 31, 35, 132
Plastic hinges, 4 Supports, 6, 13, 15, 69
Poisson's ratio, 40, 42, 77, 144
Post tension, 10, 14, 142 T-beam bridges (see slab-on-girders
Prandtl's membrane analogy, 98 bridges)
Prerast prestressed concrete girders, 11 Torsional inertia, 31, 94-109
Prestressed concrete bridges, 11, 12, 16 Torsional moment, 111
Principal moments, 141 Torsional parameter, 39, 44
Program manual, 32, 147 Torsional rigidity, 39, 44
Pseudo-slab, 7, 8, 11, 87, 101, 146 Transfer of loads, 117-131
Pucher's chart. 142 Transformation matrix, 46, 60
Push launching, 5 Transverse grid lines, 31, 78
Trapezoidal cell, 14
Re-entrant corners, 97 Tributary area, 112, 113
Responses, 31, 32, 111, 119, 134, 147 Trigonometric solution, 41
Result output, 150, 165 Tyre contact area, 18
Rigid frame bridges, 6
Rigid supports, 16, 65, 133 Users manual (see computer programs)
Users specified loading, 32, 134, 147,
Saint-Venant, 40, 96 162
Segmental construction, 5
Shear deformation, 13, 92, 101 Vehicular live loading (see also live
Shear lag, 13, 92, 101, 172 load), 27, 31, 111, 115, 119
Shear modulus, 70, 76, 166 Voided slab (see also slab bridges),
Sign convention, 56, 150 6, 13, 87, 99, 146, 182
Skeletal structures, 46, 56
Skew angle, 80 Warping, 172
Skew decks
finite difference method, 41, 42 Young's modulus. 70, 76, 166
finite element method, 45
orthotropic plate theory, 40
c),
ti
Subject Index
Analysis, 35, 111, 131-139 Clearance, 20, 22-24
Arbitrary planforms, 16 Codes of practices (see bridge codes)
Assumptions, 36, 56 Composite construction, 10
Axes, 42, 46 Coarse mesh, 41, 171
Axle loading, 19 Contributory area (see tributory area)
Computer aided methods, 1, 3
Band-width, 51, 64 Computer programs
C Beam GRID, 32, 66-69, 205-212
contiguous beam, 7, 87 GABS, 32, 146-169, 171-203,
effective flange width, 96, 101 213-250
V equilibrium equations, 74 Contiguous beams (see beams)
spaced beam, 14 Contiguous nodes, 115, 117
spine beam, 14 Courbon's method, 1, 36-38
Bearing stiffness, 133 Curved deck, 42
Bearings Curved geometry, 16
neoprene, 16, 32, 133
rocker, 16 Data
roller, 16 input, 150, 165
Bending moment diagram, 140, 143 output, 150, 165
Box-girder bridges, 6, 12-15, 92-94, Dead Ioad, 18, 112, 118
103-109, 143-146 Degrees of freedom, 46, 56-58
Bridge classification, 5 Design curves, 1, 39, 43, 44
Bridge codes, 5, 17, 18 Design envelopes, 31, 138
Bridge deck analysis (see analysis) Diaphragms, 10, 13
Bureau of Indian Standards (BIS), 17 Direct stiffness method, 31, 55-73, 132._.-
Displacement method, 56
Cable stayed bridges. 3, 6 Distortion, 13, 93, 108, 144
CAD/CAM systems, 4 Ductility-ratio, 4
Cantilever bridges, 16
Cantilever construction, 5 Cell Edge grid lines, 87, 89
distortion, 13, 93, 108, 144 Effective width of flange, 96, 101
Cellular deck, 6, 51, 103 Elastic properties, 94-109
(`am not , Anri nr.rvisrdaCci no,
(.1 Envelope diagrams, 134, 138, 196
CA), 8 Equilibrium equations, 74
Cho lesky' s factorisation Equivalent loads (see transfer of
method, 31, 69, 133 loads)
252 Subject Index
Examples, 69-73, 171-203
Expert system, 4
External prestressing, 3
Fine mesh, 42, 171

Finite difference method, 1, 40-42
Finite element method (FEM),
1, 48, 49, 200
Finite strip method, 1, 50-52
Flexibility matrix, 55
Flexible support, 16
Flexural moment of inertia, 31, 38,
94-109
Flexural parameter, 39, 43
Flowcharts, 66-68, 150-160
Folded plate analysis, 47
Footpaths, 28, 113, 118
Force method,.56
Force responses, 31, 32, 111, 134
FORTRAN, 31, 32, 69, 146
Fourier analysis (see harmonic
analysis)
Gauss-Elimination procedure,
31, 66, 133
Gauss-Seidel method, 65
Grade separation, 12
Grillage examples
box-girder deck, 143, 199-203
grid deck, 69-73
skew deck, 69-73, 179-182
slab deck, 172-187
slab-on-girders deck, 187-198
voided slab deck, 183-187
Grillage section properties
box-girder-deck, 103-109
cellular deck (see box-girder deck)
skew deck, 101
slab deck, 99-101
slab-on-girders deck, 101-103
Harmonic components, 39, 42, 44

Hendrv-Jaear method, 42-45 design
graphs, 43-44
interpolation function, 44 Highway
bridge decks, 5, 18 Hypothetical
loading system, 18, 144
Idealization
bridge deck, 74-94
loading, 117-131
Impact load, 25-27, 114
Inclined webs, 93
Indian Railways, 17
Indian Roads Congress (IRC), 2, 17,
30, 36, 96, 113
'Influence surfaces, 42
Interpretation of results, 139-145
Isotropic decks, 41-76
Kerb, 18, 19, 26, 28, 113, 162
Launching, 5
Limit state concept, 4
Lines of strength, 78
Live load
AASHO, 28, 147
IRC, 18-25, 28-30
OHBD, 147, 191
Loading (see dead load, live load,
impact load)
Load distribution, 35, 43, 99
Loading standards, 17, 24, 28, 30
Local effects, 111
Longitudinal grid lines, 31, 78
Matrix method of analysis, 31, 55 M-

beam, 8
Method of harmonic analysis (see
Hendry-Jaeger method)
Ministry of surface transport, 17,
193 Ministry of transport, UK, 8
Moment of inertia (see flexural
moment of inertia)
MS-DOS, 146
Multi-cellular decks, 13, 106
Neutral axis. 99
Nodal deformations, 66, 131, 133
Nodal loads, 31, 69, 117
Ontario highway bridge design code,

147, 172, 191
Orthotropic plate theory, 38-40
coupling rigidities, 39
equations, 39
flexural parameter, 39
torsional parameter, 39
Orthotropic slab, 38-40, 100,140
is
Panels
parallelogram, 115, 121, 128-131
rectangular, 115, 119, 123-125
triangular, 115, 119,.121, 125-129
Parapets, 18, 26
Partial prestressing, 3
Patch loads, 18, 50, 114
Physical deck, 74, 75, 99, 111
Plan geometry, 5, 15, 31, 35
Planforms, 2, 5, 15, 69
Plastic binges, 4
Poisson's ratio, 40, 42, 77, 144
Post tension, 10, 14, 142
Prandtl's membrane analogy, 98
Precast prestressed concrete girders, 11
Prestressed concrete bridges, 11, 12, 16
Principal moments, 141
Program manual, 32, 147
Pseudo-slab, 7, 8, 11, 87, 101, 146
Pucher's chart, 142
Push launching, 5
Re-entrant corners, 97
Responses, 31, 32, 111, 119, 134, 147
Result output, 150, 165
Rigid frame bridges, 6
Rigid supports, 16, 65, 133
Saint-Venant, 40, 96
Segmental construction, 5
Shear deformation, 13, 92, 101 Vehicular live loading (see also live
Shear lag, 13, 92, 101, 172 load), 27, 31, 111, 115, 119
Shear modulus, 70, 76, 166 Voided slab (see also slab bridges),
Sign convention, 56, 150 6, 13, 87, 99, 146, 182
Skeletal structures, 46, 56
Skew angle, 80 Warping, 172
Skew decks
finite difference method, 41, 42 Young's modulus, 70, 76, 166
finite element method, 45
thccry, '!"
Subject Index 253
Skew grid, 31, 42, 69-73, 80
Skyline technique, 64
Slab bridges, 6-9; 79-89, 99-101, 139-
142
Slab-on-girders bridges, 6, 9-12, 90-
92, 101-103, 143
Slope deflection, 74, 75
Solid slab, 79, 80, 90, 99, 146
Spaced box-girders, 94 Space
frame analysis, 46 Spine
beam bridges, 14 Spine box-
girders, 14, 94, 146
Stiffness method (see direct stiffness
method)
Subroutines (see computer programs)
Support conditions, 2, 5, 31, 35, 132
Supports, 6, 13, 15, 69
T-beam bridges (see slab-on-girders

bridges)
Torsional inertia, 31, 94-109
Torsional moment, 111
Torsional parameter, 39, 44
Torsional rigidity, 39, 44
Transfer of loads, 117-131
Transformation matrix, 46, 60
Transverse grid lines, 31, 78
Trapezoidal cell, 14
Tributary area, 112, 113
Trigonometric solution, 41
Tyre contact area, 18
Users manual (see computer programs)

Users specified loading, 32, 134, 147,
162
G

Grillage Analogy C.S.surana R.agrawal - For Word

Uploaded by

Copyright:

Available Formats

Grillage Analogy C.S.surana R.agrawal - For Word

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Grillage Analogy C.S.surana R.agrawal - For Word

Uploaded by

Copyright:

Available Formats

What is the book mainly intended for?

What is the book mainly intended for?

What methods of bridge deck analysis are described in the book?

What methods of bridge deck analysis are described in the book?

Grillage Analogy in

Bridge Deck Analysis

New Delhi Madras Bombay Calcutta London

All rights reserved. No part of this publication may be

All export rights for this book vest exclusively with

Published by N.K. Mehra for Narosa Publishing House, 6

The designer and consultant handling a variety of structural problems

2. Methods of Bridge Deck Analysis

2.4 Orthotropic Plate Theory 38

3. Stiffness Method Applied to Grillage Analysis 55

4. Transformation of Bridge Deck into Equivalent Grillage 74

5. Application of Loads, Analysis, Force Responses

6. Computer Program 146

7. Illustrative Examples 171

Appendix I: Listing of Program GRID 205

1.2 RECENT TRENDS IN ANALYSIS AND DESIGN OF BRIDGES

1.2.1 Structural Systems

1.2.2 Computer-Aided Methods of Analysis

1.2.3 Design Methodology

1.2.4 Modern Construction Techniques

Concrete bridges are usually constructed on stationary falsework only in

1.3 STRUCTURAL FORMS OF BRIDGE DECKS

1.4 FORM OF CONSTRUCTION

1.4.1 Slab Bridge

Many forms of prestressed, precast beams are used in pseudo-slab decks.

Figure 1.2 PCDG Standard Precast Pretensioned Bridge Beams

Permanent Pocking a) Pseudo- Box

the formwork. However, in general, the prestressing of slab, is

1.4.2 Slab-on-Girders Bridge

end diaphragms. However, T-beam bridge with cross-beams extending into

a) Cast- in- Situ Prestressed Concrete Girder

b) Tronsversety Precast Prestressed Girder Deck

Figure 1.5 Bridge Decks with Post-Tensioned Girders

together. Though, normally, this is understood to refer to the use of rolled

a) Precast Prestressed Concrete Girders

n-situ R.C. slab

b) Rolled Steel Girder Sections with Cast-

1.4.3 Box-Girder Bridge

In recent years, single or multi-ell reinforced and prestressed concrete

The cross-section of a concrete box-girder bridge consists of top and bottom

(a) Single -Cell Box-Girder

Figure 1.7 Box-Girder Bridges

Multi-cell box-girder bridges are constructed with or without transverse

In the span range of 20 to 30 m, cast-in-situ multi-cell reinforced concrete

(a) Right Deck (b) Skew Deck

(c) Symmetrical Curved (d) Unsymmetrical Curved

(e) Arbitrary Shaped Decks

(f) Bifurcated Deck

Figure 1.9 Plan Shapes of Bridge Decks

1.6 SUPPORT CONFIGURATIONS

1.7 BRIDGE LOADINGS

(a) Simply Supported Arrangement

(b) Cantilever Arrangement

nature. In order to maintain uniformity in design, loading standards have

18 Grillage-Analogy in Bridge Deck Analysis