CAMBERING AND PRESTRESSING OF STANDARD

STEEL GIRDERS ON RAILWAY BRIDGES

Ravindra Kumar Goel1 & H.O. Narayan2
(Published in Journal of IPWE, Jan., 2009)

1.0 INTRODUCTION
Indian Railways are having about 1.27 lac bridges out of which about 16000 are steel girder bridges.
Mostly all steel bridges of span 30.5 meters and above are of open web type. Open web girders are
used as through standard spans of 30.5, 45.7, 61.0 and 76.2 meters. Warren truss (triangulated truss)
with vertical members at every panel point is used as standard truss for these girders. As a standard
practice camber is provided in steel girder bridges to offset the effect of deflection under moving load.
In addition to camber, prestressing of truss members is done to counter the stresses likely to develop
under actual loads. This paper briefly summarizes the concepts and the existing design provisions for
cambering & prestressing.
2.0 CONCEPT OF CAMBER & PRESTRESSING
2.1 Concept of camber is often not well understood by bridge engineers and it is wrongly considered to give
prestressing forces in the members. Actually both the terms are having different meaning and different
purposes. To provide camber is a functional requirement to avoid effects of vertical acceleration under
moving trains. Camber is given in such a way that sag of floor system is restricted when Live Load
passes over the bridge. Maximum deflections under specified live loads are found and are given in the
opposite direction during design to work out the cambered lengths of the members. Finally when girder
is erected floor system will have upward sag. When train passes over the girder the floor system will
become horizontal. This is the ideal condition for which the truss has been analysed.
2.2 In case camber is not provided, the deflected shape of the truss under live loads will create additional
stresses in the members. These stresses are called secondary stresses.
2.3 Prestressing is done intentionally in addition to camber to develop reverse kind of stresses in various
members of the truss so that these members remain less stressed under moving loads. No savings of
material are intended as the sectional area provided is sufficient to take the design stresses. In fact we
can design the girders as cambered with or without prestressing. In case, the girders are designed as
cambered with prestressing, the secondary stresses are ignored, thus giving an economical design.

Director, Bridges & Structures Directorate, Research Designs & Standard Organization, Indian Railways, Manak Nagar, Lucknow226011

Asst. Design Engineer, Bridges & Structures Directorate, Research Designs & Standard Organization, Indian Railways, Manak Nagar,
Lucknow-226011

3.0 RELEVANT DESIGN PROVISIONS OF IRS STEEL BRIDGE CODE

3.1 The primary stresses in the design of triangulated structures are defined as axial stresses in members
calculated on the assumption that a)

All members are straight and free to rotate at the joints.

b) All joints lie at the intersection of centroidal axes of the members.

c)

All loads including the weight of the members are applied at the joints.

3.2 In practice the assumptions made above are not realised and consequently members are subjected not
only to axial stresses but also to bending and shear stresses. These stresses are defined as secondary
stresses, and fall into two groups:
a)

Stresses which are the result of eccentricity of connections and of off-joint loading generally e.g.
loads rolling direct on chords, self weight of members and wind loads on members.

b) Stresses which are the result of elastic deformation of the structure and the rigidity of the joints.
These are known as deformation stresses.
3.3 Structures are designed, fabricated and erected in such a manner as to minimise as far as possible
secondary stresses. In the case of truss spans, ratios of width of the members (in the plane of distortion)
to their lengths between centres of inter-sections may preferably be not greater than 1/12 for chord
members and 1/24 for web members, in order to minimize the deformation stresses.
3.4 Secondary stresses which are the result of eccentricity of connections and off-joint loading generally are
computed and combined with the co-existent axial stresses, but secondary stresses due to the selfweight and wind on the member are ignored.
3.5 In all cases of truss members deformation stresses described above are either computed or assumed as
specified and added to the co-existing axial stresses.
3.6 In non-pre-stressed girders, deformation stresses are assumed to be not less than 16 2/3 per cent of the
dead load and live load stress including impact.
3.7 In the case of prestressed girders, deformation stresses may be ignored. However, girders are not to be
designed for prestressing unless it is assured that the standard of workmanship in the fabrication and
erection of girders will be such that correct prestressing can be relied upon.
3.8 The effectiveness of pre-stressing in the web members of spans below 60m (200ft) and in all members
of spans below 45m (150ft) is to be ignored. In actual practice, in standard open web girders, the
effectiveness of prestressing in all the members, is ignored.
3.9 All open web girders for railway bridges of spans 30.5 m (100ft) and above shall be prestressed.
3.10 Rules for prestressing of open web girders are given in Appendix A of Steel Bridge Code.
3.11 Beams and plate girder spans up to and including 35.0m need not be cambered.

3.12 In non-prestressed open web spans, the camber of the main girders and the corresponding variations in
length of members shall be such that when the girders are loaded with full dead load plus 75 per cent of
the live load without impact producing maximum bending moment, they shall take up the true geometrical
shape assumed in their design.
3.13 Where girders are prestressed the stress camber change should be based on full dead load and live load
including impact.
4.0

CAMBER REQUIREMENTS OF STEEL GIRDERS

4.1 All standard plate girders are of less than 35.0 metre span, hence camber is not provided.
4.2 All open web girders of railway bridges are required to be designed as pre-stressed.
4.3 Dead load camber recorded after assembly and erection should be retained during the service life of
girder if there is no distress.
4.4 During technical inspection, camber is checked at every panel point of bottom chords of both truss with
the help of dumpy level or precision level, which will facilitate the inspection officials to understand the
structural condition.
4.5 As far as possible camber observations are required to be taken at the ambient temperature mentioned
in the stress sheet. Details of dead load camber is given in Table 1.0 for open web through girders for
guidance.
Table 1: Design and Dead Load Camber of Through Type
Open Web Girders
(a)

MBG Loading

SPAN
30.5 m
45.7 m
61.0 m
76.2m

(b)

Condition
A
B
A
B
A
B
A
B

L0
0
0
0
0
0
0
0
0

L1
15.5
13.5
23.0
18.0
32.5
24.5
35.0
24.0

L2
23
20
40
32
57
44
62.5
50.5

L3
27
24
55
44
76.5
58.5
88
58

L4
57
46
81
62
101
65

L5
-

Drg. No.
BA-11341 to 57
BA-11361 to 77
BA-11321 to 38

108
70

BA-11151 to 68

HM Loading

SPAN
30.5 m
45.7 m
61.0 m
76.2m

Condition
A
B
A
B
A
B
A
B

L0
0
0
0
0
0
0
0
0

L1
13.88
12.16
23
19.78
32.75
25.93
37.6
29.07

L2
22.87
20.04
34.8
29.92
55.6
44.03
67.35
52.06

L3
28.49
24.96
48.6
41.79
74.63
59.08
93.29
72.12

L4
53.6
46.09
78.12
61.85
105.68
81.7

L5
114.17
88.26

A-Design camber value as per camber sheet of fabrication drawing

B-Dead load camber value after erection & supporting span on bearings after removing intermediate supports.
Note : All dimensions are in mm

Drg. No.
BA-11521 to 38
BA-11501 to 18
BA-11551 to 68
BA-11621 to 39

4.6 It is a good practice to draw dead load camber, last inspection camber readings and present recorded
readings of each panel points on graph paper with different colour for each girder of bridge to ascertain
any loss of camber which will be a permanent record.
4.7 If there is no loss of camber when compared to dead load, camber recorded in bridge register or the last
inspection report, it will assure inspecting official that no internal structural change is taking place and
structure is not distressed.
4.8 Loss of camber in a girder can be due to:
i)

Heavy overstressing of girder or component due to loss of cross section on account of corrosion
or increased load than the designed load or due to fatigue of the structure.

ii)

Adding on extra dead load on girder such as ballast, extra thickness of road material, service
pipelines, etc.

iii) Overstressing of joint rivets i.e. in open web girder at any panel point.
iv) Play between holes and rivet shank on account of elongation of holes or crushing of rivet shanks.
4.9 If any loss of camber is noticed during inspection, following investigation is to be carried out:
1.

Camber readings should be verified again at the ambient temperature at which bearings have been
centralized during erection. At high temperatures lesser camber values are expected due to
longitudinal thermal expansion of girders.

2.

Girder should be thoroughly inspected and checked for loose rivets at panel joints or at splices.
Also all members should be thoroughly inspected for distortion or deformation.

3.

Check whether bearings are frozen.

4.

Stress reading of critical members under maximum load should also be taken by strain gauge
methods to check if any member is over-stressed.

4.10 Loss of Camber may not actually affect the functionality of bridges, specially if, standard steel girders
have been provided. Normally loss of camber should not occur, if the girder has been fabricated &
erected as per laid down quality procedures. In order to ensure this, fabrication of all the open web
girders is inspected by B&S Directorate of RDSO.
5.0 CONCLUSIONS
5.1 It can be seen that providing camber in open web girder bridges is an important functional requirement.
The constant availability of dead load camber during the service life of the bridge is an important parameter
to monitor the health of the bridge.
5.2 Standard open web girders are designed as cambered with pre-stressing, thus ignoring the secondary
stresses and giving an economical design.
5.3 Quality control during fabrication and erection is an important codal requirement to ensure longer service
life of the girders. Therefore, all works of fabrication of open web girders are to be inspected by RDSO.
4

5.4 Loss of camber in standard open web girder, during service life needs to be analyzed critically before
recommending re-girdering or imposing severe speed restrictions.
5.5 The availability of the specified dead load camber after erection is a sign of good workmanship during
fabrication and erection.

x x x

DYNAMIC EFFECTS ON RAILWAY

BRIDGES - A REVIEW
R.K. Goel*
(Published in Indian Railway Technical Bulletin, R.D.S.O. Lucknow, May, 2009)

bl rduhdh ys[k esa] Hkkjrh; jsy ds iqyksa ds vfHkdYi gsrq iz;qDr gksus okys xfr izHkkoksa dk iquoZyksdu
fd;k x;k gSA ;g ik;k x;k gS fd mPpxfr ds iqyksa ds vfHkdYi esa ^xfr fo'ys"k.k^ djus gsrq izko/kkuksa dks
fodflr djus dh rqjUr vko';drk gSA
This technical paper reviews the existing provisions for considering dynamic effects for design of bridges
on Indian Railways. It is observed that there is a need to develop provisions for conducting Dynamic Analysis
of bridges for high speeds.
1.0 INTRODUCTION
The dynamics of railway bridges involves the response of bridges to the movement of vehicles and to the
influence of a number of parameters which increase dynamic strains or stresses
The most important parameters influencing the dynamic effect in railway bridges are the characteristics
of bridge structures (i.e., the length, mass, and rigidity of individual members), the characteristics of vehicles
(i.e., the sprung and unsprung masses, the stiffness of springs), the damping in bridges and in vehicles, the
velocity of vehicle movement, the track irregularities, and so on. The vehicles affect the bridges not only by
vertical forces, but also by movements which generate longitudinal and transverse horizontal forces. The
railway bridges are designed considering the dynamic effect of the moving load by increasing the live load by
an impact factor, which is a function of a single variable i.e. its length. However recent codes of practice
address the problem by giving the full recognition to the fact that resonance in bridges may occur depending
upon the characteristics of the train, parameters defining its passage, the bridge characteristics and track
irregularities.
In the wake of current International practice the provisions in Bridge Rules need to be re-examined and
amended appropriately. The paper describes in detail provisions of different codes in this context and justifies
the need of further studies on existing bridges to suggest new provisions for inclusion in IRS Bridge Rules.
2.0 DYNAMIC EFFECTS OF HIGH SPEED ON BRIDGES
The train running with high speed induces dynamic impact on the bridge structures, influencing their
working state and service life.

*Director/Steel Bridges-I, Bridges & Structures Directorate, Ministry of Railways, RDSO, Lucknow (UP)-226011.

The vibrations of the bridge in turn affect the running stability and safety of the train, and thus becomes
an important factor for evaluating the dynamic parameters in bridge design. Therefore, in many countries, the
dynamic behaviors of bridges have been systematically studied in the development of high speed railway. The
high speed trains provide intensive vibrations similar to the resonance phenomenon.

Fig - 1 Dynamic Effects on Railway Bridges

The resonance occurs if the frequency of an input force coincides with one of the natural frequencies of
the system. The resonant vibration of railway bridges results in the deterioration of passenger comfort, reduction
of traffic safety (a possibility of derailment of vehicles), the destabilization of ballast (higher maintenance
costs) and increased damage in the bridge system from fatigue considerations.
3.0 THE EXISTING PROVISIONS IN BRIDGE RULES
3.1 Coefficient of Dynamic Augment
In design practice, dynamic effect of the moving load is taken care of by increasing the live load by
impact factor or dynamic augmentation factor or dynamic coefficient. This factor depends on many parameters
like the type of loading, speed, type of structure, material of structure, loaded length etc. But for simplicity on
Indian on Railways, impact factor is specified by the Bridge Rules, involving only one parameter, i.e., the
loaded length. All the other parameters are considered to have constant value (the most adverse value). For

Broad Gauge and Meter Gauge steel railway bridges carrying a single track, the impact factor is given by the
following expression:
Impact factor = 0.15

8
6 L

, subject to a maximum of 1.0

Where L is defined as given below

a)

L is loaded length of the span in meters for the position of the train giving the maximum stress in the
member under consideration. For the design of chord members, it will be the whole span of the
truss and for the web members only part of the span is to be loaded.

b) L is taken as 1.5 times the cross-girder spacing for finding stresses in the stringers (rail-bearers).
c)

L is taken as 2.5 times the cross-girder spacing for finding moments in the cross-girders (floorbeams).

3.2 Critical Speed

3.2.1Critical speed is defined as the speed at which the external forcing frequency will be equal to one of the
natural frequencies of the track-bridge vehicles- system, contributing to vertical response of the bridge.
3.2.2Critical speed in the case of steam locomotives and for open web girders only may be calculated by any
of the following methods:
(i) by running trains at varying speed across the bridge and determining the speed giving the maximum
deflection.
(ii) by ascertaining the maximum static deflection under live-load and applying the following formula,
2C

V
d(

WhereV

= critical speed in km/h

= circumference of driving wheels in m.

W = dead load of the span in kN (t) per m

P

= equivalent live load in kN (t) per m run of the train on the span, at the position giving maximum
Bending Moment, and

= maximum static deflection in m caused by the live load; and

(iii)

by the following approximate formula: V

266
L

Where,
V

= critical speed km/h and

= effective length of span in m.

3.2.3Speed restrictions for open web girders for steam traction in the range of critical speed 10 km/h and
determined as above.
4.0 THE PROVISIONS IN THE OTHER CODES OF PRACTICE
4.1 BS 5400 Part - 2
4.1.1. In this standard, dynamic effects are considered in clause 8.2.3. Here equivalent static loadings (RU
and RL loading) are multiplied by appropriate dynamic factors to allow for impact, oscillation and other
dynamic effects including those caused by track and wheel irregularities. The dynamic factor for RU
loading applies to all types of track and is given in Table 1.
Table - 1 Dynamic factor for type RU loading
Dimensional
(m)
Upto 3.6
From 3.6 to 67
Over 67

Dynamic factor for evaluating

Bending Moment
Shear
2.00
0.73 +

1.67

2.16
(L 0.2)
1.00

0.82 +

1.44
(L 0.2)

1.00

4.1.2In deriving the dynamic factor, L is taken as the length (in m) of the influence line for deflection of the
element under consideration. For unsymmetrical influence lines, L is twice the distance between the
point at which the greatest ordinate occurs and the nearest end point of the influence line. In the case of
floor members, 3 m should be added to the length of the influence line as an allowance for load distribution
through track.
4.1.3The dynamic factor for RL loading, when evaluating moments and shears, shall be taken as 1.20, except
for un-ballasted tracks where, for rail bearers and single track cross girders, the dynamic factor shall be
increased to 1.40.
4.2 EN 1991-2
4.2.1In European standard dynamic effects (including resonance) are considered in clause 6.4. In this code
dynamic effects are taken care of in a better way.
4.2.2 A static analysis shall be carried out with the load models (Load Model 71 and where required Load
Models SW/0 and SW/2). The results shall be multiplied by the dynamic factor, f to consider the

dynamic effects. Generally the dynamic factor f is taken as either f2 or f3 according to the quality of
track maintenance as follows:
(a) For carefully maintained track:
22

1.44
0.82, with: 1.00 < f2 < 1.67
L
0.2

(b) For track with standard maintenance:

33

2.16
0.73, with: 1.00 < f < 2.00
3
L
0.2

Where Lf is ''Determinant'' length (in m.).

The dynamic factor f3 shall not be used with:

the loading due to Real Trains.

the loading due to Fatigue Trains.
the load model HSLM.
the load model ''unloaded train''.

For determining whether a dynamic analysis is required a flow chart is used which is given in Fig. 2

Fig - 2 Flow chart for determining whether a dynamic analysis is

required ( EN 1991-2:2003(E), page no 77)

10

Where:
V

is the Maximum Line Speed at the site (km/h)

is the span length

n0

is the first natural bending frequency of the bridge loaded by permanent actions (Hz)

nT

is the first natural torsional frequency of the bridge loaded by permanent actions (Hz)

is the Maximum Nominal Speed (m/s)

(v/g0)lim is given in annex F (EN 1991-2: 2003(E))

5.0 Comparison of Dynamic Factors as Suggested by Different Codes
Quasi static methods which use static load effects multiplied by the dynamic factor f are unable to
predict resonance effects from high speed trains. Dynamic analysis techniques, which take into account the
time dependant nature of the loading from the High Speed Load Model (HSLM) and Real Trains (e.g., by
solving equations of motion) are required for predicting dynamic effects at resonance. A comparison of
Dynamic Factor as per important international codes w.r.to IRS Bridge Rules is shown below.
Table - 2 Comparison of Dynamic Factors according to various codes
Description

IRS

BS 5400 Part-2

UIC 776-1 R

EN 1991-2

Gauge

Broad Gauge and

Metre Gauge (one
formula)
Static

1.4 m gauge and

above

Standard Gauge

Static

Static

Four
250
kN
concentrated loads
preceded
and
followed
by
a
uniformly
distributed load of
80 kN/m
L, length of the
influence line for
deflection
of
element
under
consideration

As per UIC
Leaflet 702

Standard track
Gauge and Wide
track gauges
Static. Dynamic for
special cases of speed,
span, natural frequency,
etc.
Four
250
kN
concentrated
loads
preceded and followed
by
a
uniformly
distributed load of 80
kN/m

Method
analysis

Loading

Basis

of

Axle loads of
245.2 kN (25 t) for
the
locomotives
and a train load of
80.9 kN/m (8.25t)
on both sides of
the locomotives
L, loaded length
of span for the
position of the
train giving the
maximum stress in
the member

11

L,
Characteristic
length,
the
length of the
influence line
for
the
deflection of the
member
to
which
the
calculations
refer.

L, Determinant length,
the length of the
influence
line
for
deflection of the element
being considered

Track
maintenance
standards

Not considered

Not
considered
(Maintenance
of
track and rolling
stock to be of
reasonable standard)

Considered
i) track
maintained to
exacting
standard
ii) other lines

Considered
i) carefully maintained
track
ii) track with standard
maintenance

Fill below
bottom of
sleeper

Reduction in the
coefficient for arch
bridges
and
Concrete
slabs/girders
of
span < 25 m
Reduction
in
Dynamic
coefficient
by
multiplying factor

Not considered

Reduction in the
coefficient, for
depth of fill more
than 1 m, for
arch and concrete
bridges
_

Reduction
in
the
coefficient, for depth of
fill more than 1 m, for
arch and concrete ridges

Multiple
spans

i)

No reduction in
Dynamic
coefficient
Permitted

0.60 for
intermediate
spans of steel
girders

ii) 2/3 for arch

bridges more
than 15m span
Formulae for
coefficient in a
general
case (for BM
and track
with
standard
maintenance)
Upper and
lower
limiting
values

Natural
frequency of
structure

0 .15

8
6 L

i) 1.15 - 2.00 for

BM and SF both
1.10 minimum for
analyzing existing
bridges for speed
clearances of new
stock
Does not consider
specifically

2.16
L
0.2

0.73

2.16
L
0 .2

0.73

2.16
0.2

0.73

BM: 1.00 - 2.00

SF : 1.00 - 1.67

1.00

i) Carefully maintained
track : 1.00 - 1.67
ii) Track with standard
maintenance : 1.00 2.00

Does not consider

specifically

Vibration effects
are in-built in the
coefficient. Also,
stipulates
the
natural frequency
to be within
prescribed values

Vibration effects are inbuilt in the

coefficient. When natural
frequency falls beyond
the limiting
values, and also for closer
values of natural torsional
and bending frequencies,
Dynamic analysis is
prescribed

12

Resonance

Not considered but

critical
speed
defined for steam
locomotives and
open web girders.

Not considered

Not considered

Limitation
for speed

160 & 125 kmph

for passenger and
goods trains

No

No

Limitation
for span

No limitation in
IRS Bridge rules,
however, IRS steel
bridge Code is
applicable upto
100 m span for
simply
supported
spans
only.

No

No

Not considered in Static

analysis.
However,
when falls in the
category
requiring
dynamic analysis, gets
considered.
200 kmph with static
Analysis
350 kmph with
Dynamic analysis
No

6.0 OBSERVATIONS
Following observations are made :
(i) There are no rational provisions in the present IRS Bridge Rule to relate the dynamic effects with
speed of train and standard of track maintenance.
(ii) Method needs to be developed to determine the requirement of dynamic analysis of bridges for
speeds more than 160 kmph for passenger trains &125 kmph for goods trains.
(iii) The procedure for carry out dynamic analysis needs to be developed in context of Indian Railways,
on the pattern of practices followed in other international codes, keeping in view the differences in
gauge, loadings, types of rolling stocks and locomotives etc.
(iv) Effect of track maintenance standard on dynamic augment needs to be studied and quantified. Use
of track recording data can also be made in this regard.
7.0 CONCLUSION
It is evident that the present provisions of IRS Bridge Rules are not in line with international practices
being followed in European Countries. The concept of dynamic analysis for higher speeds needs to be
introduced for which detailed studies are required to be undertaken.

x x x

13

FATIGUE DESIGN OF WELDED CONNECTIONS

FOR RAILWAY BRIDGES
Goel Ravindra Kumar
(Presented in International Seminar organized by
Indian Institute of Welding, Mumbai, Feb. 2005)
ABSTRACT
The design of the connections forms an important part of the overall design of a bridge structure. The bridge
designer while designing the connections has to consider the factors such as optimum location of the joint, the type of
connections bolted or welded, shop connection or field connection, sizes of fabricated members and their transportation
requirement to site besides the strength considerations. Careful attention is required to be given to the strength and fatigue
behaviour of the chosen connection. Different types of welding techniques are used in fabrication of structural members
depending upon their strength requirements and criticality from fatigue consideration. The philosophy of fatigue design
has been briefly discussed in this paper and the Indian Railway practice for design and manufacture of welded components
for bridges has been presented. Shop welding has been adopted by Indian Railways using submerged arc welding technique
to reduce the number of rivets drastically. However, the field connections have still been kept riveted. The performance of
the welded connections provided has been reported satisfactory. The different types of welded and non welded connections
adopted have been listed and the need of research in assessing fatigue strength of welded connections in Indian conditions
has been identified.
Key Words : Welded connections, fatigue behaviour, bridge structure, welding techniques, stress concentration,
welded components, fatigue strength, submerged arc welding technique.

AUTHOR DETAILS
Ravindra Kumar Goel is Director/Steel Bridge in Bridges & Structures Directorate at Research Designs
& Standards Organisation, Ministry of Railways, Lucknow-226011 (U.P.), India. He has earlier worked as
Dy. Chief Engineer/Bridges and In-charge of Bridge Workshop, Northern Railway, Jalandhar Cantt. He has
good experience of fabrication of steel structures and bridges using riveted as well as welded connections.
Fabrication of welded bridge girders was started at Bridge Workshop, Northern Railway, Jalandhar Cantt.
under his guidance. He has also developed and implemented quality system ISO-9002 for fabrication of
steel structures and bridges at the Fabrication Workshop.
1.0 INTRODUCTION
On Indian Railways, majority of bridge superstructures are of steel. Till 1985, only riveted connections
were being used, as proven welding technology and requisite infrastructural facilities for fabrication of bridge
girders was not available. The riveted connections tend to make the structure heavy and thus uneconomical
besides posing maintenance problems. During the last two decades proven welding technology has been
established and use of welded connections has been started on Indian Railways. Railway bridges are subjected
to heavy dynamic loads and fluctuations of stresses may cause fatigue failure of members or connections at
lower stresses than those at which they would fail under static load. Such failures are primarily due to stress
14

concentrations introduced by constructional details. All details are therefore to be designed to avoid as far as
possible stress concentration likely to result in excessive reductions of the fatigue strength of connections. In
view of above the welded connections are designed for non-critical locations only.
2.0 PHILOSOPHY OF FATIGUE DESIGN
2.1 Design Approaches & Design Input Requirement
Two major approaches are extensively followed in engineering applications, they are safe-life design
and fail-safe design approach. The safe-life design approach aims at determination of life of the structure,
before the end of which the structure can be repaired, replaced, or retired. At the same time, it has also been
recognized that some structural damage are inevitable and failure would occur and that the catastrophic
failure is rarely tolerable. Fail-safe design recognizes that fatigue crack may occur and arranges the structure
so that cracks will not lead to failure of the structure before they are detected and repaired. Multiple load
paths and crack stopper built at intervals into the structure are some of the means to achieve fail-safe design.
There are two primary groups of information that are necessary as an input for a comprehensive fatigue
analysis. One group of information is the data related to the material behavior when subjected to cyclic
loading, such as laboratory tests for constructing S-N curve, and other factors that would help to evaluate life
of the structure. The laboratory tests must simulate the stress environment that the structural component will
experience. The second group of information is the determination of the total number of cycles that the
structure will undergo throughout its life (Load Spectrum). With these two groups of information available,
complete fatigue analysis can be done.
Having S-N curve and load spectrum available, cumulative damage theory is introduced that relates
cycles of magnitudes to the S-N curve in order to predict the total life of the structural component.

In evaluating the number of cycles

to failure for given structure subjected to
fluctuating loads, fatigue test data
representing the load environment must be
available. Fatigue test data is represented
in the form of S-N diagrams. The S-N
diagram is a plot of stress amplitude,
stress range or the maximum cyclic stress,
S (selected as the controlled or
independent variable), versus the number
of cycles of failure, N (the dependent
variable).

Log (Stress.S)

2.2 S-N Diagrams

Endurance limit
Log (number of cycles)
(Figure 1) Idealized S-N curve (logarithmic plot)

For most materials, the logarithmic plot of S-N curve is approximated by a straight line. For some

15

alloys, including ferrous alloy, the logarithmic plotting will generate additional straight (horizontal) to account
for the endurance limit also called the fatigue limit, (Figure 1). The endurance limit is an important parameter
while designing an element to have an infinite life.
Factors influencing the endurance limit include, the stress ratio, degree of surface finish, head treatment,
stress concentration and corrosive environment. Therefore, it is expected that the fatigue limit will have a
wide range of values depending on conditions described above. Studies have been conducted abroad on
experimental data and it has been shown that certain alloys, such as ferrous material, exhibit a clear fatigue
limit. For the maximum applied stress below this limit, failure will not occur, and therefore, the material has
infinite life. However, sufficient studies are not available in Indian conditions.
The number of cycles to failure, N, obtained by S-N curve (under load or stress condition) is related to
total life of the member up to failure. In reality fatigue cycles throughout the life of the structural part consists
of two phases crack initiation and propagation. Propagation means stable crack growth up to crack instability.
The S-N curve approach of assessment of fatigue damage does not separate the crack initiation phase from
the propagation phase. Thus it is assumed that the crack is already initiated in the member and the total
number of cycle associated with crack propagation to failure is determined.
3.0 IRS METHODOLOGY OF DESIGN
3.1 Methodology
IRS methodology of fatigue design is based on stress ratio of minimum and maximum principal stresses
to be transmitted by the connection. The allowable stress P depends on the ratio of minimum stress f min to
maximum stress fmax, number of repetitions of stress cycle N, the method of fabrication and the type of
connection. In determining the ratio fmin/ fmax gross area is used. To allow for the effect of fatigue the allowable
working stresses are determined from Appendix G of IRS Steel Bridge Code. This appendix covers mild
and high tensile steel fabricated for connected by welding, riveting or bolting. The allowable stresses given in
the Appendix are the principal stresses at the point under consideration depending upon the weld details.
Thus in the design of girder web, the combined effect of bending and co-existent shear stresses is taken.
There are seven classes of details from A to G where A is the most superior details and G is the most
inferior detail (Figure-2)

(Inferior most)

(Inferior most)
(Figure 2) Detail class of connection

16

The permissible fatigue stress, P is a function of s tress ratio (fmin/fmax on gross area, No. of cycle N
method of fabrication and type of connections where,
fmin

= minimum force/stress primarily function of dead load.

fmax

= maximum force/stress primarily function of DL+LL+ impact.

= No. of cycles for the specified route/section. Permissible fatiguestress values are available
for 0.6, 2, 4 & 10 million of cycles for mild and high tensile steel.

A-G = Details of joints depends upon method of fabrication, type of

force on the weld, location of weld in the member etc.

connections, direction of

Note : For intermediate values of N, log interpolation is permissible.

3.2 Design Steps
(1) For sections, depending upon the number of locomotives run per day the number of cycles of
loading N is decided.
(2) The class of welding and method of fabrication to be complied with is already decided.
(3) The minimum stress or loads for different members for dead load is calculated by any acceptable
method of structural analysis.
(4) The maximum stress or loads for different members for DL+LL+Impact is calculated. The ratio of
fmin/fmax is obtained. This may be (+) ve or (-) ve depending upon the nature of stresses/force.
(5) The permissible stress in fatigue is picked-up from relevant class of details depending upon N and
ratio fmin/fmax.
The actual stress in the members must be less than the permissible stress in the members.
4.0 PERMISSIBLE STRESSES IN WELDS
Since fatigue strength of welded structures depends upon the constructional details, this is to be decided
in consultant with the fabrication agency. It is apparent that any improvement in welding technology adopted
in the fabrication would be taken in to account at the stage of deciding the construction detail. Accordingly,
the permissible stresses and the size of members and weld sizes are determined.
Stresses due to dead load, live load and impact, stresses resulting from curvature and eccentricity of
track, and secondary stresses as defined in clause 3.3.2 (a) of IRS Steel Bridge Code, are considered for
effects due to fatigue. All other items mentioned in clause 3.1 of IRS Steel Bridge Code and secondary
stresses as defined in clause 3.3.2 (b) thereof, are ignored when considering fatigue.
Permissible Stress in butt welds are not to exceed the permissible stresses of the parent metal as specified
in IRS Steel Bridge Code. However, all the butt welds are to be examined radiographically or any equally
effective non-destructive test method

17

The basic permissible stress in fillet welds based on a thickness equal to the throat thickness is limited to
100 N/mm2 (10.2 kg/mm2) where a fillet weld is subjected to shear stress in two directions, the actual stress
shall be taken as the vector sum of the separate shear stresses and not to exceed 100 N/mm2 (10.2 kg/mm2).
Load carrying fillet welds are designed such that the stress on the total effective area of fillet welds does
not exceed the relevant values specified in Table for Class G Constructional details, Appendix G to IRS
Steel Bridge Code, subject to a maximum of 100 N/mm2 (10.2 kg/mm2). These welds are also designed so
that secondary bending stresses are not developed (e.g. single lap joints shall not be used).
5.0 REDUCTION IN PERMISSIBLE STRESSES
5.1 The permissible stresses for field welds of structural members are reduced to 80%. As per existing
policy field welds are not permitted for bridges carrying road/railway loading. Thus all the welding in
railway bridges is limited to shop connections and all the field connections are still riveted.
5.2 If over-head welds are unavoidable, the stresses permitted are reduced to 80% and further reduced to
another 80% if field welding is involved.
5.3 In structures subjected to dynamic loading, tensile or shear stresses in butt welds is not permitted to
exceed 66.67 % of the permissible stresses unless the welds are examined radiographically, ultrasonically
or other non-destructive testing methods which are equally effective and present satisfactory evidence
to the Engineer that welds are meeting the quality requirement. These permissible stresses are reduced
to 80% for over-head welding and further reduced to 80% for field welding.
6.0 WELDED & NON-WELDED CONNECTIONS ADOPTED
From different considerations all the connections in fabrication of steel bridges cannot be welded. Direction
of welding with respect to the direction of principal stresses plays an important role in determining the class of
connection. The basic permissible stresses are determined accordingly and the decision to adopt the connection
as welded is taken on the basis of relative advantages and economy. Sometimes, the classification of connection
so determined, forces the designer to increase the complete cross-sectional area and the advantages sought
by adopting welded connections are nullified. As all field connections are to be riveted one, the choice also
depends on the transportation facilities likely to be available from the shop to the site. The different types of
welded and non-welded connections adopted on IRS bridges are listed as under
Table 1 IRS Type of Connections
Welded
1. Web to flange connection of stringer/plate girder.

1.

2. Connection of end stiffener to web in plate girders

3. Connection of horizontal stiffener to web not used.
4. Shear connector to top flange

2.
3.
4.

5. Longitudinal connections in fabricating built up

members of open web girder bridges
6. Butt welds in web and flange

18

Non-welded
Intermediate stiffener to web in plate
girders.
Lateral bracings (top & bottom)
All diaphragm connections to girders
Lateral connections such as batten and
lacings in built up members.

Out of these welded connections web to flange connection is the main connection involving major
quantum of welding work. This connection is designed to transmit the horizontal shear force combined with
any vertical loads which are directly applied to the flange. Where a load is directly applied to a flange, it is
considered as dispersed uniformly through the flange to the web at a slope of two horizontal to one vertical.
Butt welding has also been successfully adopted in place of spliced joints in plate girder bridges. The typical
cross sections of butt welded joints and the members built up by longitudinal fillet welding using submerged
arc welding are shown in Figure 3.

Figure 3 (a) Typical cross section of butt weld

Figure 3 (b) Typical cross section of a welded plate girder

19

Figure 3 (c) Typical cross sections of the built up members of a truss girder

20

Figure 3 (d) Typical cross sections of other members of a truss girder

7.0 MATERIAL & CONSUMABLES USED

7.1 Raw Steel
Mild steel conforming to IS:2062 Grade B fully killed and fully normalized/controlled cooled is permitted
for use in dynamically loaded structures. However, plates less than 12mm thick need not be normalized/
controlled cooled. Wherever, the service temperatures are likely to go below zero degree census steel
conforming to IS:2062 grade C is recommended ensuring impact properties at (-)200C to (-)400C. There
is a general difficulty in getting rolled sections like angles, channels, I-sections etc conforming to IS:2062 in
Grade B or C apparently due to a small requirement of construction industry. Therefore, rolled section in
Grade A are permitted till such time they are readily available in grade B/C.
High tensile steel complying the requirement of IS:8500 grade 540 (copper bearing quality) is prescribed
for welded work. All finished steel is to be well and cleanly rolled to the dimensions, sections and masses
specified. The finished material is ensured to be reasonably free from surface flaws; laminations; rough/
jagged and imperfect edges; and all other harmful defects.
7.2 Welding Electrodes
All welding electrodes used for fabrication of welded components are to be strictly as per IRS M-28 &
IRS M-39 for metal arc welding and submerged arc welding wire flux combination respectively. Welded
construction work is to be carried out generally in accordance with the provisions of Indian Railway Standard
21

Welded Bridge Code. The electrodes have been classified into 20 different classes. The purpose for which,
each class of electrode is to be used, together with the range of codings is given in of IRS M:28-1976. Filler
wire for CO2 welding should be as per RDSO specification for CO2 welding filler wire (Tentative). RDSO
issues periodical list of approved suppliers of electrodes for metal arc welding.
8.0 WELDING TECHNIQUES
8.1 All welds are done by submerged-arc welding process either fully automatic or semi-automatic. Carbon
dioxide welding or manual metal-arc welding may be done only for welds of very short runs or of minor
importance or where access of the locations of weld do not permit automatic or semi-automatic welding.
8.2 Except for special types of edge preparation, such as single and double U the plates which are to be
joined by welding may be prepared by using mechanically controlled automatic flame cutting equipment
and then ground to a smooth finish. Special edge preparation is made by machining or gouging.
9.0 WELDING PROCEDURES
9.1 The welding procedure is to be such as to avoid distortion and minimize residual shrinkage stresses.
Properly designed jigs are used for assembly. The welding techniques and sequence, quality, size of
electrodes, voltage and current required are monitored as prescribed by manufacturers of the material
and welding equipment.
9.2 Site welding is not to be undertaken except in special circumstances with the approval of the Engineer.
Site welding should be confined to connections having low stresses, secondary members, bracings etc.
9.3 Manual metal arc welding is permitted with adequate precautions as per IS:9595 and under strict
supervision of competent supervisor.
10.0 SEQUENCE OF WELDING AND WELD PASS
Distortions may occur due to heat emission during welding process. To avoid such distortions proper
sequence of welding is followed. The correct sequence is quite often developed with experience for which
frequent interaction of designer with the fabrication is necessary. Some of the cases are illustrated as under:
10.1 For fabrication of welded composite girders, channel shear connectors shall be welded on top flange
plate prior to assembly of I-section. This facilities correction of any distortion of flange plate developed
during the welding of channel shear connectors.
10.2 In making of a typical I-section four fillet welds are to be made. The welding sequence to be followed
is indicated by number 1 to 4 as shown in the Figure 4.
10.3 Whenever a square butt weld in a 10 or 12mm thick plate is required to be made, the sequence to be
adopted is shown in Figure 5.

22

1
10mm
or 12mm
2
(Figure 4) Sketch showing sequence of square butt welding

(Figure 5) Sketch showing the sequence of fillet welding for fabricating the I-section

11.0 QUALITY CONTROL

To ensure required quality of welded connections, a proper quality assurance programme is to be
decided before starting the fabrication. To ensure that quality of fabrication does not suffer in any way,
inspection of bridge girders, has been entrusted to Research, Design and Standards Organisation, Ministry of
Railways, Lucknow. RDSO looks after in detail the various aspects involved in the fabrication and inspection
of these girders (such as specification of steel, welding process, fabrication techniques, stage inspection etc.)
to maintain the quality of not only the end product but quality of the process as a whole.
12.0 CONCLUSION:
Indian Railways is in the process of adopting more and more welded connections for design of railway
bridges. So far the welding has been used to make shop connections in fabricating individual bridge members.
The connections have been made using submerged arc welding technique and besides reducing the dead
weight, these are found quite convenient from the maintenance point of view. However, the use has been
restricted to shop welding only, that too for non critical locations because of proneness of welded connections
to fatigue failure. So far, the welded connections have been used extensively at the location of web to flange
connections of stringers and plate girder bridges. Longitudinal fillet welding has also been used extensively to
fabricate built up members of open web girders.

23

Further research is required to correctly assess the fatigue strength of different types of welded connections
under different type of loading conditions. There is also a need to improve the reliability of welded connections
for increasing its use in Railway bridges which are subjected to dynamic loading. Any improvement in the
welding technology adopted and its reliability from fatigue consideration will greatly help the designers in
adopting welded connections for other critical locations also.
13.0 ACKNOWLEDGEMENT
The author gratefully acknowledges the encouragement and support provided by Executive Director
(B&S), RDSO in preparation of this paper. The assistance provided by Shri A.K. Pandey, Section Engineer
and Smt. Suman Verma, Steno Grade-1 of B&S Dte.,RDSO/ Lucknow is also thankfully acknowledged.

x x x

24

DESIGN CONSIDERATIONS FOR THE DESIGN OF

POT-PTFE BEARINGS FOR RAILWAY BRIDGES
R.K. Goel* & R.N. Shukla**
(Presented in National Seminar of IIBE Delhi State chapter on
Load Transfer Devices and foundations
at Vigyan Bhawan New Delhi, Jan., 2006 )
SYNOPSIS
The important design consideration involved in the design of Pot-PTFE bearing have been discussed in this paper
and the limitations in design of these bearings for Railway Bridges have been illustrated. The design data of various
standard spans has been shown and the need of making in-depth study of the implications of relaxing codal provisions has
been stressed.

1.0 Introduction:
The pot bearing consists of circular, non-reinforced natural rubber or elastomer pad, totally enclosed in
a steel pot with the load applied to the elastomer via a piston attached to the upper bearing plate. A seal is
used to prevent rubber extruding between piston & pot. As the elastomer is fully confined within a metal
cylinder, it provides a load carrying medium whilst at the same time providing the bearing with a multidirectional
rotational capacity. By themselves, pot bearings do not permit translation. In order to permit translational
movement in addition to rotation, plain sliding arrangement is provided over the top plate of pot bearing.
PTFE (Poly tetrafluroethylene) which is having very low coefficient of friction is generally used with stainless
steel to design this sliding arrangement. The weight of Pot-PTFE bearing is about 50% of the weight of
conventional rocker & roller bearing for the same span. Due to its less weight and due to almost no maintenance,
it is desired to provide Pot-PTFE bearings instead of conventional rocker-roller bearings. However, due to
certain design constraints these bearings cannot be designed for all type of spans. This paper describes in
brief the design consideration involved and the limitations in the design of such bearings for railways bridges.
2.0 Material Specifications:
(i)

Mild Steel

IS: 2062 grade-B

(ii) Stainless Steel

IS: 6911

(iii) Cast steel

IS:1030 grade 280-520W.

(iv) Elastomer pad

a) IRC:83 (Part-II) Standard specifications and code of practice

for Road Bridges- Elastomeric Bearings.

* Director/Steel Bridges-I, B&S Dte., RDSO, Ministry of Railways, Lucknow

** Sr. Section Engineer (B&S) RDSO, Ministry of Railways, Lucknow

25

b) IRC-83 (Part-III) Properties of confined elastomer.

(v) PTFE

a) BS:3784 grade A. Specification for poly tetrafluroethylene

(PTFE) sheet.
b) IRC:83 (Part-III) for permissible pressure on confined PTFE.
c) BS:5350: Standard method of test for adhesives, Part-C9,
Floating Roller Peel Test.

(vi) Wiper seal & dust seal : RDSO/M&C/RP-194/94.

3.0 Codes of Practice:
In absence of standard code of practice for Railway Bridges, the design of bearings to Railway Bridges
is done in accordance with following documents:
i)

IRS Bridge Rules

ii)

IRS Steel Bridge Code

iii) IRC-83 Standard specifications and code of practice for Road Bridges, Section:IX Bearings,
Part-III: Pot, Pot-cum-PTFE, Pin and metallic guide bearings.
The relevant clauses for design are discussed as under:
3.1 Design of Elastomeric Pad Diameter and Thickness:
3.1.1Average stress in confined elastomeric pressure pad of Pot bearing shall not exceed 35 Mpa and
extreme fibre pressure shall not exceed 40 MPa.

Fig. 1: Design principle of cylinder wall

26

3.1.2The minimum thickness of the confined elastomeric pressure pad shall not be less than 1/15th of its
diameter or 16mm, whichever is higher and the diameter shall not be less than 180mm.
3.1.3Minimum average stress in confined elastomeric pressure pad of Pot bearing, under any critical
combination of loads and forces that can coexist, shall in no case be less than 5 MPa.
3.2 Design of Pot Wall:
3.2.1The design of Pot wall is done with respect to the hoop tensile stresses in the cross section of the
cylinder wall. Hoop tensile stress in the cross section of cylinder wall due to :
i.

Fluid pressure, sat1 = (di x he x sce) / (2 x bp x hc)

ii.

Horizontal force, sat2 = H / (2 x bp x hc),

Where,
di

= diameter of confined elastomeric pressure pad in mm

he

= thickness of confined elastomeric pressure pad in mm,

sce =Fluid pressure in confined elastomeric pressure pad due to vertical load in MPa.
bp = thickness of cylinder wall in mm
hc

= height of cylinder wall in mm

Total hoop tensile stress (sats,cal) due to fluid pressure and horizontal force i.e. (sat1+ sat2), shall not
exceed the value of permissible stress in axial tension as specified.
3.3 Design of Sealing Ring:
3.3.1For brass sealing ring type internal seal, 2mm thick and 20mm wide split rings made of metallic brass
shall be provided in layers with staggered split positions. Minimum two layers of rings shall be provided

Fig. 2: Allowable strain in elastromeric pressure pad

27

for elastomeric pressure pad of diameter upto 480mm and minimum three layers of rings shall be provided
for elastomeric pressure pad of diameters more than 480mm.
3.3.2The dimension of the confined elastomeric pressure pad shall be such that at design rotation the deflection
at the perimeter shall not exceed 15 percent of the pad thickness below the internal seal.
3.4 Design of PTFE Guided Sliding Assembly
3.4.1PTFE shall be located into recess of a sufficiently rigid metal backing plate by confinement and shall
either be dimpled large sheet(s) or an array of solid (i.e., without dimples) rectangular modules of size
80mm x 50mm. The dimpled large sheets shall be circular or rectangular in shape and may be subdivided

Fig. 3. Typical arrangement of modular pife

Fig. 4. Patern of dimples in recessed pite sheet

28

into a maximum of four parts. For dimpled sheets with smallest dimension (diameter or smaller side)
exceeding 100mm, contact area shall be taken as the gross area without deduction for the area of the
dimples. In arrayed PTFE layout the distance between the individual modules shall not be more than
10mm. The shoulders of the recess should be sharp and square to restrict the flow of PTFE. The
thickness of the PTFE and its protrusion from the recess should be related to its maximum plan dimension
in accordance with Table 1.
Table 1. Dimension of Confined PTFE
Maximum dimension of PTFE
(diameter or diagonal)
(mm)

Minimum thickness
(mm)

Maximum protrusion
above recess
(mm)

< 600

4.5

2.0

> 600, < 1200

5.0

2.5

> 1200, < 1500

6.0

3.0

3.4.2Average pressure on confined PTFE shall not exceed 40 MPa and extreme fibre pressure shall not
exceed 45 Mpa.
3.4.3The characteristic maximum coefficient of friction for steel sliding on uniformly lubricated PTFE shall be
as per Table 2. Liner interpolation may be used for intermediate values. In absence of test data the
coefficient of friction of unlubricated PTFE on stainless steel should be taken as twice the value as given
in Table 2. For design purposes, induced horizontal force caused by the resistance to translational
movement due to friction at the PTFE-stainless steel interface shall be determined considering the
PTFE as unlubricated.
Table 2- Coefficient of Friction for Stainless Steel Sliding on Properly Lubricated PTFE
Average pressure on confined
PTFE (MPa)

Maximum coefficient
of friction

0.08

10

0.06

20

0.04

> 30

0.03

3.5 Design Horizontal Force:

3.5.1For design of Pot bearings or part thereof the design horizontal force to be considered shall be the
resultant of the coexisting active horizontal force, determined from global analysis, and included horizontal
forces, generated due to friction at sliding interface (if any), but shall in no case be less than 10 percent
and greater than 25 percent of the design vertical load.
29

3.6 Permissible stresses in steel

3.6.1Permissible stress in axial tension sat shall not exceed 0.6 fy, where fy = minimum yield stress of steel in
MPa.
3.6.2Maximum bending stress in tension (sbt) or in compression (sbc) in extreme fibre shall not exceed 0.66
fy.
3.6.3Maximum shear stress (tvm) shall not exceed 0.45 fy.
3.6.4Maximum bearing stress (sp) shall not exceed 0.75 fy.
3.6.5Irrespective of any increase in the permissible stress specified the equivalent stress (se) due to coexisting
bending (tension or compression) and shear stress obtained from the following formula and shall not
exceed 0.9fy.
se, cal =

(3 x tvm,cal2 + sbt,cal2)

se, cal =

(3 x tvm,cal2 + sbc,cal2)

or

Where,
tvm, cal =

Calculated value of shear stress in MPa,

sbt, cal =

Calculated value of bending stress in tension in MPa

sbc, cal =

Calculated value of bending stress in compression in MPa

se, cal =

Calculated value of combined stress in MPa

3.6.6When the effect of wind or earthquake is taken into account the above permissible stresses shall not be
increased.
3.6.6The above permissible values are also applicable for stresses on steel structure, adjacent to the bearing.
3.7 Distribution of Vertical Load
3.7.1In absence of 3D FEM analysis, load distribution through the bearing component(s) and to the adjacent
structure shall be calculated considering effective contact area after one vertical to two horizontal (IV:2H)
distribution of confined elastomer stress as shown in Fig. 5. Flexural stress due to active and induced
moments shall be calculated considering the section modulus of the effective contact area as shown in
Fig. 5. Average contact stress, flexural stress and the combined effect shall not exceed the limiting values
as specified. For dispersion through sliding components it should be ensured that the dispersed area is
contained within the particular component, even when maximum design displacement occurs.

30

Fig. 5: Load dispersion through bearing components

3.8 Shear Stress & Bending Stresses

3.8.1Shear stress at cylinder wall and base interface calculated considering 1mm radial slice of the cylinder
due to:
(i) Fluid pressure, tvm1 = (he x sce)/bp
(ii) Horizontal force, tvm2 = 1.5 x H/(di x bp)
Where, Parabolic distribution factor is considered as 1.5.
3.8.2Total shear stress (tvm, cal) due to fluid pressure and horizontal force, i.e. tvm1 + tvm2, shall not exceed
the value of permissible stress in shear as specified earlier in para 3.5.
3.8.3Bending stress at cylinder and base interface calculated considering 1mm radial slice of the
cylinder due to:
(i) Fluid pressure, sbt1 = (6 x sce x he2)/(2 x bp2)
(ii) Horizontal force, sbt2 = 1.5 x 6 x H x ha/(di x bp2)
Where,
ha

height of line of application of design horizontal force from cylinder wall above base interface
in mm,

design horizontal force in N.

31

he

thickness of confined elastomeric pressure pad in mm

sce =

fluid pressure in confined elastomeric pressure pad due to vertical load in MPa

bp

thickness of cylinder wall in mm

di

diameter of confined elastomeric pressure pad in mm.

Parabolic distribution factor is considered as 1.5.

3.8.4Bending stress (sbt,cal) due to fluid pressure and horizontal force, i.e. sbt + sbt2, shall not exceed the
value of permissible bending stress as specified in Para 3.5.
3.8.5Equivalent stress (se, cal) due to combined bending and shear shall be checked in accordance with Para
3.5.
Cantilever projection of guide shall be checked for shear, bending and combined stresses
against permissible values specified in Para 3.5.
3.9 Average Permissible Direct Bearing Pressure on Pier/abutment
3.9.1Average permissible direct bearing pressure scc on the adjacent concrete structure shall be calculated
using the following equation:
scc

sco

(A1/A2)

sco

permissible direct compressive stress in concrete = 0.25 fck, where fck is the characteristic
compressive strength of concrete.

A1

dispersed concentric area, which is geometrically similar to the loaded area A2 and also the
largest area that can be contained in a plane of A1 (maximum width of dispersion beyond
the loaded area face shall be limited to twice the height).

A2

loaded area and,

(A1/A2) < 2

The projection of the adjacent structure beyond the loaded area shall not be less than 150mm. Adequate
reinforcement for spalling and bursting tension shall be provided.
3.9.2In case of coexisting direct and flexural compressive stresses on the adjacent concrete structure, the
following criteria should be satisfied:
scc, cal/scc + sc, cal/sc < 1
Where,
scc, cal

calculated direct compressive stress,

sc, cal

calculated compressive flexural stress.

32

3.9.3When the effect of wind or earthquake is taken into account, the above permissible stresses may be
increased by 25 per cent.
4.0 Design of Pot-PTFE Bearings For Railway Bridges
4.1 The design of Pot-PTFE bearings has been done for large railway bridges based on above codal
provisions and provisional drawings have been issued. The typical design data for 76.2m span is given
in Annexure-I. The sectional elevation and plan of typical sliding bearing for 76.2 m span railway bridge
girder is shown in Fig. 6. The values of critical design parameters in respect of minimum average stress,
average stress and extreme fibre stress are tabulated in Table-3 for different spans.
Table-3 Values of critical design parameters for different Spans for railway bridge girders

Span
(m)

Drg.No.

Loading

Total Perm. Str.

weight of > 5N/mm2
girder
including
Min.
track wt. average
etc. (t)
stress
(N/mm2)

Perm. Stress
2

Perm. Stress

35 N/mm

40 N/mm2

Average Stress

In Extreme fibre stress

Service
condition

Seismic
condition

Service
condition

Seismic
condition

(N/mm2)

(N/mm2)

(N/mm2)

(N/mm2)

76.2

B-11578

HM
(OWG)

405

5.367

21.56

24.11

33.86

37.86

76.2

B-11577

MBG
(OWG)

322

5.677

21.19

24.33

34.74

39.86

61.0

B-11576

HM
(OWG)

286

5.038

22.75

25.15

35.58

39.32

61.0

BA-11575

MBG
(OWG)

221

5.13

22.21

25.21

35.64

40.20

45.7

MBG (under
slung)

140

5.084

26.32

30.68

41.96

48.87

33

Fig. 6 Typical sectional elevation and plans of sliding bearing for 76.2 m span railway
open web girder for H.M. Loading

4.2 It is observed that whereas it is possible to simultaneously meet the design criteria in respect of minimum
average stress and maximum extreme fibre stress for larger spans, the same is not easy for lower spans
e.g. for 45.7m span open web girder (under-slung type), the total weight of girder is so less (only 140t)
that the diameter of elastomer pad has to be reduced substantially to achieve minimum average stress of
5N/mm2. With this reduced diameter of elastomer pad, the extreme fibre stresses under service conditions
as well with seismic effect are exceeding the permissible values (40 N/mm2).
4.3 The codal provision as given in BS-5400: Section 9.1 for design of Pot-PTFE bearing have been
checked up and it is noted that no such limitation in respect of minimum average stress has been specified.
However, keeping in view the stability of girders during seismic conditions the design of Pot-PTFE
bearings for railway bridges has been done for ensuring minimum average stress 5N/mm2.

34

5.0 Testing of Bearing

The actual stress in the elastomer in Pot bearing due to design load effects is limited by the effectiveness
of the seal preventing it from extruding between the piston and the pot wall. Since the details of pot bearings
vary considering and stress analysis is complex, the design is invariably verified by testing. A sample test
proforma is attached as Annexure-II.
6.0 Conclusions
Design of Pot-PTFE bearing is governed by the minimum average stress on the confined elastomer in
the pot. For smaller spans it is not possible to simultaneously meet out the codal provisions for minimum
average stress as well as for maximum extreme fibre stress in the elastomer pad. The design of Pot-PTFE
bearing is therefore restricted due to the stipulated minimum average stress and can be feasible for railway
bridge girders of spans 61.0 m and above.
7.0 Suggestion For Future Study
7.1 The implications of relaxing the existing codal provision for minimum average stress need to be studied
further to take full advantage of Pot-PTFE bearings for all types of Railway Bridges.
8.0 References
[1]

IRC-83 (Part III), Standard Specifications and Code of Practice for Road Bridges, Section:IX, Bearings, Part-III:Pot,
Pot cum PTFE, Pin and Metallic Guide Bearings.

[2]

BS-3784 Grade A, Specifications for Poly tetrafluroethylene (PTFE) Sheet

[3]

BS-5400:Section 9.1, Design of Pot-PTFE Bearings.

[4]

LEE, DAVID J., Bridge Bearings and Expansion Joints, E & FN Spon London, UK.

[5]

IRS Steel Bridge Code, Research Design and Standard Organization, Ministry of Railways, Govt. of India, Lucknow
(U.P.)

[6]

RDSO Design Documents for design of Pot-PTFE bearings for different spans and loadings.

35

ANNEXURE-I
Data for design of sliding and fixed pot PTFE bearing for 76.2m HM loading
1

Clear span

76.2 m

Effective span

78.8 m

D.L of girder including Track, gangway etc

405 t

L.L shear

980.024t

CDA

240.106 t

Vertical load due to wind/Bearing

16.84 t

Lateral load due to wind/Bearing

24.2 t

Lateral load due to seismic effect for zone-iv

25.18 t

Longitudinal force/Bearing

67.5 t

10

Max. horizontal movement -

63.5 mm

11

Total vertical load/bearing with wind = {(3)+(4)+(5)}/4+16.84

{(405+980.024+240.106)}/4+16.84

12

= 423.123t

Total vertical load/bearing with seismic effect for zone - iv

{(405+980.024+240.106)}/4+35.09+13.24

= 454.613t

13

Deflection at centre = 116mm

14

THEETA = ATAN(116/39400) 0.0029 radians < 0.025 or 0.16869 Degree

36

ANNEXURE-II
CHECK LIST OF INSPECTION FOR FABRICATION OF POT PTFE BEARING
1.0 Raw Material
1.
2.

M.S. Steel
Cast Steel

3.
4.
5.
6.
7.

PTFE
Elastomer
Stainless Steel Plate
Internal seal (brass ring)
Wiper seal and dust seal

As per IS 2062 Gr. B or C as mentioned in Drawing

As per IS 1030 Gr. 280-520 W or As per IRC 83 (iii) as mentioned in
drawing or contract agreement
As per drawing or IRC 83 (iii) BS 3483 Gr. A
As per IRC 83 (Part II)
IS-6911
As per IS 410
As per drawing and RDSOs Specification M&C/RP-194/94

2.0 Dimensional Test

2.1 Fixed Bearing (Pot Bearing)
Sl.
Dimension (mm)
Item
No.
Nominal
Measured
1. Pot cylinder and elastomer
a) Inner dia of the Pot
b) Outer dia/size of the pot
c) Center to center distance of holes in the lugs of pot
d) Width of lugs (measure and record width of all lugs)
e) Thickness of the pot wall
f) Thickness of the base of pot
g) Thickness of lugs
h) Thickness of the pot wall
i) Inner height of the pot
j) Thickness of elastomer
2.
Piston/top plate
a) Thickness of the top plate in piston
b) Thickness of the top plate above pot wall
c) Dia of the top plates
d) Center to center distance of holes in the lugs of top
plate
e) Thickness of lugs
3.
Sealing ring
a) Thickness
b) Width
c) Inner diameter
4.
High tensile bolts and washers
a) Dia of bolt sleeve with HT anchor
b) Dia of bolt hole in bolt sleeve
c) Dia of HT anchor
d) Length of bolt sleeve
e) Length of bolt sleeve with anchor
5.
Overall height of bearing in assembled condition

37

Whether within
tolerance

2.2 Free Bearings (Pot cum PTFE bearings)

Sl.
Item
No.
1. Pot cylinder and elastomer
a) Inner dia of the Pot
b) Outer dia/size of the pot
c) Center to center distance of holes in the lugs of
pot

2.

3.

4.

5.

6.

Dimension (mm)
Nominal
Measured

d) Width of lugs (measure and record width of all

lugs)
e) Thickness of the pot wall
f) Thickness of the base of pot
g) Thickness of lugs
h) Thickness of the pot wall
i) Inner height of the pot
j) Thickness of elastomer
Saddle Plate & PTFE
a) Size of the saddle plate
b) Size of PTFE sheet in plan
c) Thickness of PTFE sheet
d) Depth of dimples
e) Dia dimples
f) Spacing of dimples
g) Depth of groove in saddle plate for housing
h) Thickness of saddle plate
i) Thickness of the collar of saddle plate
Top Plate, Stainless Steel & Ear Plate
a) Plan size of underside of top plate
b) Thickness of top plate
c) Plan size of steel plate
d) Size of the lugs of top plate
Sealing ring
a) Thickness and no.
b) Width
c) Inner diameter
High tensile bolts and washers
a) Dia of bolt sleeve with HT anchor
b) Dia of bolt hole in bolt sleeve
c) Dia of HT anchor
d) Length of bolt sleeve
e) Length of bolt sleeve with anchor
Overall height of bearing in assembled condition

x x x

38

Whether within
tolerance

STUDY OF BEHAVIOUR OF STRINGER TO FLOOR

BEAM CONNECTION IN RIVETED
RAILWAY OPEN WEB GIRDER BRIDGES
Ravindra Kumar Goel*1
(Published in Journal of IPWE, April, 2006)
ABSTRACT
Stringer-to-floor-beam connections in riveted railway bridges have in many cases shown to be critical details with
respect to fatigue. These connections, while generally designed with respect to shear forces alone, are often subjected to
repeat secondary bending as a result of their rotational stiffness. The behaviour of double-angle stringer-to-floor-beam
connections in riveted railway bridges has been analyzed based on performance of full scale bridge parts. The response of
these shear connections has been studied under the action of bending moment. It is found that these connections are
capable of developing appreciable moments due to restraint they exert on the rotation of stringer of stringer ends associated
with bending. The resulting bending and axial stresses in the angles and the rivets of the connection might consequently
be considerable. High stress concentrations are also present in these components, which further increase their fatiguedamage susceptibility. The paper discusses the important findings of the study.

1.0 INTRODUCTION
The floor-system in old riveted railway truss bridges is typically designed as a grid structure consisting of
longitudinal and transverse members (stringers and floor-beams) connected through their web plates by
means of riveted double-angles. The main function of these connections is to transfer the end reactions of
stringer to the floor-beam through shear action.
One general assumption that is made in the design of these double-angle connections is that they have
sufficient rotational flexibility to allow for the stringer-end rotation associated with bending without developing
appreciable moment. This assumption is also often adopted today. Analyses of the load-carrying capacity of
double-angle stringer-to-floor-beam connections in existing riveted bridges, and the assessment of their
fatigue strength, are generally made with reference to the shear forces acting on the connections only.
The assumption of null degree-of-fixity might be justifiable when it comes to the ultimate load-carrying
capacity of the connections and the connected members. However, overlooking the effect of the rotational
stiffness of these connections might result in an inadequate estimation of their fatigue strength. A relatively
large number of fatigue-damage cases in double-angle stringer-to-floor-beam connections have been reported,
the majority of which were attributed to the moment acting on these connections.

*Director/Steel Bridges-I, Bridges & Structures Directorate, Ministry of Railways, RDSO, Lucknow (UP)-226011.

39

2.0 BEHAVIOUR OF RIVETED DOUBLE ANGLE CONNECTION

2.1 As per prevailing IRS practice, the connection between stringer and cross beam is designed as a shear
connection to transmit the vertical shear at the joint location considering the stringer as simply supported.
Practically the stringer is not simply supported as the riveted double angle connection is not free to
rotate in the plane of stringer axis. The rotational stiffness of such a joint is also not sufficient enough to
design the stringer as completely fixed at supports. Fig. 1 illustrates the assumed, theoretical and practical
conditions at joint for understanding the proper support conditions of stringer
Unit Load

0
= Deflection
0 = End Rotation

RA

B
RB

Case - 1 Simply Supported Beam

'

0=O

'<

RA

M1< M

M1< M

1
1< '

RA

RB

Case - 2 Beam, Completely Fixed at ends

B
01 > O
<0

Case - 3 Beam, Partially Restrained at ends

RB

Fig 1 : Deflection, Rotation and Bending Moments Generated under Differend Support Conditions

2.2 The rotational stiffness of double-angle stringer-to-floor connections is primarily a function of the flexural
stiffness of the outstanding legs of the connection angles. The effect of the deformation applied to the
connections (by the stringer-end rotation associated with bending) is to subject the outstanding legs in
the upper portion of the connection to an out-of-plane flexure, causing bending and axial stresses in the

40

angles and the rivets, respectively. The magnitude of stringer-end moment and the resulting forces in the
different components of the connection will depend on the amount of applied deformation that can be
accommodated by the flexural flexibility of the outstanding legs.
2.3 A large part of the deformation applied to the connection by the rotation of the stringer end is locked
by the stiffness of the connection. Consequently, the angle is subjected along most of its depth to tensile
forces which are counter balanced by local contact pressure at the bottom of the connection (Fig.2).
The upper part of the connection is subjected to relatively high tensile forces and the magnitude of
stringer end moment is considerable. The studies have shown that the double-angle shear connections

Fig 2 The forces acting on the connection due to the action of moment :
(a) contact pressure on the backside of the outstanding leg
(b) axial forces and stresses in the angle leg connected to the stringer
web (The values shown are typical values observed in studies)

are capable of developing more than 60% of the corresponding moment for a total rigid connection. The
moment-load behaviour can be considered linear and the position of the rotational center is only slightly
affected by the magnitude of the applied load.
3.0 DEFORMATION CHARACTERISTICS OF CONNECTION ANGLES
3.1 As has already been observed, the flexibility of the outstanding legs of double-angle connections subjected
to moment has a major impact on the response of these connections. In particular, the gauge distance
between the rivets and the fillet of the angle plays a dominant role in the behavior of these connections.
41

On the one hand, the flexibility of the outstanding leg at these locations has a decisive effect on the
amount of applied deformation that can be accommodated by the connection and consequently, on the
magnitude of the stringer-end moment. Furthermore, due to its locally higher stiffness, the outstanding

Fig. 3 The distribution of bending stresses in the connection angle

(a) near the fillet along the depth of the connection of the connection
(b) along the gauge distance and through the thickness of the angle
(The values shown are typical values observed in studies elsewhere)

leg along the gauge distance attracts a larger portion of the tensile forces acting on the connection,
resulting in high stress concentration at these locations. The bending stresses in the outstanding leg of the
angle are mostly magnified along the gauge length and reach a maximum near the fillet of the angle as
shown in Fig. 3.
3.2 When the distribution of stresses in the angle shown in Fig. 3 is analyzed, a number of observations can
be made. Although the behavior of the outstanding leg of the angle is bending dominated, some membrane
action appears to exist. Based on the distribution of stresses through the thickness of the angle near the
fillet, the contribution from the membrane action in this location is less than 10% of the bending stresses.
3.3 The rivet-clamping force had a negligible effect on the magnitude of maximum bending stresses near the
fillet of the angle. On the other hand, higher bending stresses were produced in the outstanding leg near
the rivet when the latter had a higher clamping force. The outstanding leg of the connection angle
behaves as though it was partially fixed under the rivet head. The same behavior explains the slightly
stiffer response obtained for the connections with a higher rivet-clamping force. The deformation of the
outstanding leg near the rivet is restricted by the axial and bending stiffness of the latter and is consequently
affected by its clamping force. The axial stiffness of the rivets is typically several times greater than the
flexural stiffness of the outstanding leg of the angle and the contribution form rivet elongation to the
flexibility of the connection is relatively small.
42

4.0 AXIAL AND BEARING STRESSES IN THE CONNECTING RIVETS

4.1 It is known that the clamping force in the fasteners of connections subjected to fluctuating moment and/
or tensile forces has a significant influence on the fatigue strength of these fasteners. For a given applied
load, a higher clamping force will simply result in a smaller increase in the axial force developed in these
fasteners. Several previous investigations relating to riveted connections have concluded that hot-driven
rivets could develop considerable clamping force. The average clamping ratio (i.e. the ratio between
the clamping stress in the rivet and the rivet yield stress) was found to be about 70%. The clamping
force in the rivets of flange-to-web connections was investigated for stringer and the average value of
the clamping ratio found in these rivets was 42%.
4.2 Generally, the clamping force in hot-driven rivets is highly undetermined variable. In addition to its
dependence on the grip length (i.e. shank) of the rivet and the stiffness of the connected plates, it is also
affected by the driving temperature and the driving method (e.g. hand-hammering or using hydraulicpressure riveters). The magnitude of the clamping force is also expected to reduce with time through
relaxation or due to the fretting of the connected components.
4.3 In particular, the rivets in stringer-to-floor-beam connections are expected to have substantially lower
clamping stress. These connections are generally assembled on site, which often involves uncontrolled
riveting conditions and installation problems that could result in faulty, misshapen rivets.
4.4 The effect of the clamping force in the rivets
on their fatigue strength is even more
pronounced when taking account of the
variation in local stresses in the rivet shank,
where the effect of rivet bending is also. In
fact, the bending of the rivets caused by the
flexure of the outstanding legs of the
connection angle, together with the stress
concentration present at the junction
between the rivet shank and its head, are
the major mechanisms behind crack initiation
and fracture in these rivets (rather than the
variation in the nominal axial tensile stress in
these rivets). The out of plane distortion of
the connection angles has been measured
and verified by FE modeling and relationship
Fig. 4 Out-of-plane distortion of the connection angles (FE
has been developed with gauge distance (g) model), and the corresponding displacement of the equivalent Lsegments with an effective width, Ceff
an angle thickness (t) etc. The displacement
of the connection angle with depth of connection Fig.4
4.5 The rivets in the second row of the connection (looking from the top) also experiences relatively high

43

axial and bending stresses. The segmental tensile forces carried by this rivet were also observed high
and were further magnified by the effect of prying. Fatigue failure in the upper two rivets took place in
many of the connections during the fatigue tests of the bridge parts.
5.0 CONCLUSIONS
5.1 Double-angle stringer-to-floor-beam connections have been a common source of fatigue damage in
riveted railway bridges. In many cases, this damage is generated by the secondary stringer-end moment
that develops at the connections as a result of their rotational stiffness. The behavior of double-angle
stringer-to-floor-beam connections has been studied. Static and fatigue tests have also been performed
on three full-scale bridge parts taken from an old riveted railway bridge. The results of the analysis
show that double-angle stringer-to-floor beam connections, although generally assumed to act as simple
shear joints, exert appreciable restraint on the stringer-end rotation associated with bending and, as a
result, high secondary bending moment can be developed at these connections.
5.2 The flexibility of the outstanding legs of the connection angles has a major influence on the response of
the double-angle connections. In particular the gauge distance between the rivets and the fillet of the
angle is found to play a dominant role in the behaviour of these connections. The outstanding legs of the
connection angles are subjected by the action of moment to out-of-plane distortion, which in turn generate
high flexural stresses in the angles. The distribution of these flexural stresses along the depth of the
connection is, however, not uniform. High concentrations of flexural stresses are present near the fillet
of the angle at the rivet gauge distances.
5.3 The magnitude of the clamping force in the rivets has been found to have only a marginal effect on the
rotational stiffness of the connections and the rivet-clamping force virtually does not affect the flexural
stresses in the angles. On the other hand, the resulting stress ranges in the rivets of the connection are
greatly influenced by the magnitude of clamping force in these rivets. The variation in both axial and
bending stresses in the rivets is substantially reduced when a higher clamping force was present.
5.4 Rivet bending due to prying action, together with the stress concentration present at the junction between
the rivet shank and its head, is the major mechanism behind fatigue cracking in the rivets rather than the
variation in the nominal axial tensile stress in these rivets.
6.0 REFERENCES
6.1

Bjorn Akesson, 1994. Fatigue Life of Riveted Railway Bridges, Deptt. Of Structural Engineer, Chalmers University
of Technology, Goteborg, Sweden.

6.2

Design Documents for Design of Stringer-Cross Girder Connection, Research Design & Standard Organisation,
Ministry of Railway, Lucknow (U.P.)-226011

x x x

44

SYNTHESISING S-CURVE AND IUH WITH CATCHMENT

CHARACTERISTICS
Ravindra Kumar Goel*
(Published in Indian Railway Technical Bulletin, R.D.S.O. Lucknow, 2005)

bl ys[k esa gkbMksxzkQ dh vkd`fr dks iznf'kZr djus ds fy, ,l&gkbMksxzkQ rduhd dk mi;ksx fd;k
x;k gSA izLrqr xf.krh; ekMy] ftlesa rhu iSjkehVj iz;qDr fd;s x;s gSa] ,l gkbMksxzkQ dks Hkyh izdkj izLrqr
djus esa l{ke gSA bldk mi;ksx djrs gq;s bUlVSUVfu;l ;wfuV gkbMksxzkQ dk xf.krh; ,Dlizs'ku vodyu
fof/k ls fudkyk x;k gSA ;g rduhd dEI;wVjksa ij ljyrkiwoZd iz;ksx dh tk ldrh gSA miyC/k
gkbMksykftdy vakdM+ksa dk fo'ys"k.k djds la'kysf"kr ,l&gkbMksxzkQ o vkbZ-;w-,p- rS;kj djus esa bl fof/k dk
ljyrkiwoZd mi;ksx fd;k tk ldrk gSA
ABSTRACT
Expressing the shape of a hydrograph has been a matter of interest for hydrologists. Two parameter mathematical
models have been suggested in past to express the shape of a hydrograph. In the present study S-hydrograph or S-curve
approach has been used to derive the expression for the Instantaneous Unit Hydrograph (IUH). The three parametric model
suggested has been found to satisfactorily express the shape of S-hydrograph or S-curve. The expression for Instantaneous
Unit Hydrograph (IUH) can be derived by differentiating the expression of S-hydrograph in terms of the three parameters.
The suggested approach can be conveniently used on computers to describe the shape of hydrographs and S-curves and
offer simplicity in analysis of hydrograph data.

INTRODUCTION
The concept of S-curve is very important in the unit hydrograph theory as unit hydrograph of any
duration can be derived from it. In the present study a three parameter mathematical expression has been
used to describe the shape of the S-curve. Since IUH and S-curve are related with each other, therefore, the
expression for IUH has also been derived in terms of these three parameters.
DATA USED
Available data of gauging site of Bridge No.184 on Garwa Road-Chopan section of Eastern Railway
has been used for the study. The data of 11 no. of single peaked floods were supplied to Hydrological
Directorate (Regional Studies) of CWC by RDSO and the ordinates of Unit Hydrograph were received from
CWC for each of the floods. These unit Hydrograph ordinates have been used to generate S-curve. The
ordinates of the S-curve so generated are expressed as discharge per unit drainage area per unit rate of
rainfall excess to give the saturation value equal to unity.

*Director/Steel Bridges-I, B&S Dte., Research Designs & Standards Organisation, Lucknow

45

EXPRESSION FOR THE S-CURVE SHAPE

The double log plot of the S-curves indicate that the curve follows a straight line for small values of time
t, and deviates gradually from the straight line for larger values. Such type of variation can be adequately
represented by the following form of equation -

1 (ts/t)m/n

Eq. no.1

Where, S is the ordinate of the S-curve (dimensionless), m is slope of the initial straight line and ts & n
are parameters affecting the subsequent deviation from the straight line in the plot.
The above form of equation was fitted to each of the S-curves generated above and the parameters
m, ts and n were determined. The values of these parameters for 11 no. of analysed storms are given in
Table -1.
Table I SAMPLE DATA USED IN STUDY Sub-zone 1(d)
Span :2x18.29(G)
CA:248.60 Km2 Km:73/8
4x30.48(G)
Railway: Eastern No. of R.G : 3
Section : Garwa Road-Chopan Flood on:9/10-7-1988
Date & Time Wtd.Avg.
Losses Net Rainfall
Stage
Obs. disch. Base flow Net flow Ord.Of UH
rainfall
(cm)
(cms)
(cms)
(cms)
(cms)
(cm)
(cm)
(cm)
9-7-88
8.00
0.03
0.03
219.02
0.60
9.00
1.02
0.82
0.20
219.02
0.60
10.00
1.70
0.81
0.89
219.02
0.60
11.00
219.02
0.60
0
12.00
219.02
0.60
0.02
13.00
219.02
0.60
0.60
0
0.08
14.00
219.10
2.00
1.35
0.65
0.20
15.00
0.10
0.10
220.16
122.00
2.10
119.90
0.60
16.00
Total 2.85
1.76
1.09
220.10
112.00
2.85
109.15
110.70
17.00
22.00
98.00
3.60
94.40
100.80
18.00
219.88
77.00
4.35
72.65
87.16
19.00
219.80
68.00
5.10
62.90
67.08
20.00
219.70
56.00
5.85
50.15
56.23
21.00
219.64
50.00
6.60
43.40
46.30
22.00
219.60
44.00
7.35
36.65
40.07
23.00
219.56
41.00
8.10
32.90
33.84
24.00
219.52
36.00
8.85
27.15
30.38
10-7-88
1.00
219.50
33.00
9.60
23.40
25.07

Br. No.184

2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00

219.48
219.46
219.42
219.40
219.38
219.36
219.32
219.30

46

31.00
29.00
24.00
23.00
22.00
21.00
16.00
15.50

10.35
11.10
11.85
12.60
12.35
14.10
14.85
15.50

20.65
17.90
12.15
10.40
8.65
6.90
1.15
0

21.60
19.07
16.53
11.22
9.60
7.99
6.37
0

Runoff =

751.10x60x 60x100
248.60x100 0x1000

1.09 cm % Runoff =

1.09
x100
2.85

38.16%

DEVELOPMENT OF EXPRESSION FOR IUH

IUH is related with the S-curve of unit excess rainfall intensity as under -

U (0, t)

ds
dt

Eq. no.2

Differentiating the expression of S-curve with respect to time,'t' following expression for IUH is obtained:

U (0, t )

(m / ts)(t / ts) m 1 1 (t/ts)m/n

(n 1)

Eq. no.3

Expressions For IUH Peak And Time To Peak

Putting the first differential of U(0,t) with respect to time equal to zero, the following expressions are
obtained for time to peak (tpk) and peak (Umax) of IUH :
tpk

U max

n (m - 1)
m n

n/m

ts

1 (m n) (m
ts
mn

Eq. no.4

n)/m

n(m - 1) n(m-1)/m
(n 1) (n 1)

Eq. no.5

REPRESENTATIVE S-CURVE & IUH

The above expression of S-curve has been fitted to the shape of the S-curve generated for each of the
11 Unit Hydrograph data under study. The determined values of the parameters m, n & ts are shown in
Table-2. The average values of m, n & ts have been used to derive a representative S-curve and IUH for the
Table -2 S-CURVE PARAMETERS AS DETERMINED
Bridge No. 184 Eastern Railway Section:Garwa Road-Chopan Catchment Area=248.60 km2
Parameters of s-curve
No.
Date of storm
n
ts
m
1
19/20-7-92
0.56
8.9
2.45
2
9/10-7-88
0.64
9.4
2.6
3
15/16-8-89
0.65
9.6
2.8
4
14-7-90
0.71
9
5.5
5
12/8/1991
0.55
8.5
5
6
1/2-9-91
0.36
5.75
1.8
7
15/16-7-93
0.66
5.5
2.8
8
4/5-8-94
0.81
5.1
2.83
9
10/11-8-94
0.59
9.5
3.3
10
18-8-94
0.645
8.15
1.95
11
17/18-9-94
0.82
6.25
3
Average
0.64
7.79
3.09

47

1.2000
0 .1 80 0

1.0000

discharge coeff.

0 .1 60 0

0 .1 40 0

d isch arge coeff.

0 .1 20 0

0 .1 00 0

0.8000

0.6000

0.4000

0.2000

0.0000

0 .0 80 0

10

15

20

25

30

tim e,in hrs

Fig. 1(b) Plot of Representative S-curve

0 .0 60 0

0 .0 40 0

0 .0 20 0

0 .0 00 0
0

10

15

20

25

30

tim e in h rs

Fig. 1(a) P lot of R epresentative

Instan teneou s Unit H ydrograph

catchment area of gauging site at Br. No.184. The plots of the representative IUH and S-curve are shown
in Fig. 1(a) & Fig.1(b) respectively.
CONCLUSION
It is evident that the three parametric mathematical expression given above is capable to describing the
shape of representative S-curve and IUH for the catchment under study. The approach can be extended
further to available hydrological data of other catchments and the representative parameters of S-curve can
be determined. These S-curve parameters can be co-related with the catchment characteristics by regression
analysis and synthetic S-curves and IUHs may be obtained directly in the form of a mathematical expression.
REFERENCES
Goel, Ravindra Kumar (1989), "S-curve synthesis using catchment characteristics", M.E. dissertation submitted to Civil
Engg. Department, University of Roorkee, Roorkee.
McCuen, R.H. (1989), "Hydrologic analysis and design".
Report No.RBF-34, (2003), "Validation of flood estimation report no.S/15/1987 subzone-1(d) (Sone Basin)", Research Designs
& Standards Orgnaisation, Ministry of Railways, Lucknow.

48

APPENDIX - I
DERIVATION OF EXPRESSION FOR IUH SHAPE
The first differential of the proposed S-Curve equation with respect to time t, yields the expression for
Instantaneous Unit Hydrograph as follows:

U(0,t)

ds
dt
ds
1 (ts/t)m/n
dt
- n. 1 (ts/t)m/n

m
(ts/t)(m
ts
m
(ts/t){(m
ts
U(0,t)

n)/n

n 1

( m/n).tsm/n .t -(m/n)-1

1 (ts/t)m/n

n)- m(n 1)}/n

m
(t/ts)m-1 1 (t/ts)m/n
ts

(n 1)

(t/ts)m/n 1

(n 1)

( n 1)

This is the expression for the shape of an Instantaneous Unit Hydrograph. Differentiating this expression
w.r.t. time t, once again we get :

m(m - 1)
d
(1/ts)m-1 t m- 2 1 (t/ts)m/n
U(0,t)
ts
dt
m
(t/ts)m-1 { (n 1)}1 (t/ts)m/n
ts
(1/ts) m/n (m/n) t

( n 1)

( n 2)

(m/n) -1

m(m - 1)
(t/ts)m- 2 1 (t/ts)m/n
ts.ts
1 m.2 (n 1)
(t/ts) m
ts 2
n

m/n - 2

( n 1)

1 (t/ts) m/n

49

( n 2)

m
(t/ts)m- 2 1 (t/ts)m/n
2
ts

( n 2)

(m - 1){1 (t/ts) m/n } (m/n)(n 1) (t/ts) m/n

m
(t/ts)m- 2 1 (t/ts)m/n
2
ts

(n 2)

(m - 1) (t/ts) m/n { m - 1 - m - m/n}

m(m - 1)
(t/ts)m- 2 1 (t/ts)m/n
2
ts

(n 2)

1 - (m n)/{n(m - 1)} (t/ts)m/n

Now,

dU
0, whendt
(i)

(t/ts) m

= > t = 0 and U = 0 (minimum)

(ii)

1 (t/ts)m/n

( n 2)

OR 1 (t/s)m/n

(n 2)

=> t = and U = 0 (minimum)

(iii)

1-

(m n)
(t/ts) (n/m)
n(m - 1)

n(m - 1)
m n

and U (m/ts)

(n / m)

n(m - 1)
m n

ts

( n/m)(m -1)

n(m - 1)
1
m n

50

( n 1)

n(m - 1)
m n

(m/ts)

( n/m)(m -1)

( n 1)

m(n 1)
m n

(m/ts)

{n(m - 1)}(n/m)(m-1) {m(n 1)} -(n

(m n){n (m-1)/m - (n 1)}

1/ts

(m n)(m n)/n {n (m - 1)}(n/m)(m-1)

m n (n 1)(n 1)

1)

i.e. IUH Peak,

Umax

(m n) (m n)/n {n (m - 1)}(n/m)(m-1)
1/ts
m n (n 1)(n 1)

and time to peak,

tpk

n(m - 1)
m n

(n/m)

ts

x x x

51

Composite sleepers : An Environment Friendly Alternate

Sleeper for Track and Bridges
Rama Kant Gupta*
(Published in Indian Railway Technical Bulletin, R.D.S.O. Lucknow, No., 2003)

daiksftV Lyhij u;h izdkj dh i;kZoj.k vuqdy

w Lyhij gS tks osLV&IykfLVd ls cuh gksrh gSA buesa ydM+h ds Lyhijksa dk
yxHkx lHkh izdkj dk yphykiu gSA dbZ n`f"Vdks.kksa ls ;g ydM+h ds Lyhijksa ls lqihfj;j gSA la;Dq r jkT; vesfjdk esa jsyos Lyhijksa
ds vfrfjDr ikWyhoqM esd dh lkexzh ls cgqr ls iqyksa dk fuekZ.k fd;k x;k gS blds iz;ksx ls LVs&osLV&IykfLVd ls cus gksus ds
dkj.k dsoy Hkwfe&izn"w k.k de ugha gksrk gS vfirq blls ou&laink dks igqp
a us okyh {kfr esa Hkh deh vkrh gSA bl izdkj bls nks vFkks
esa gjk mRikn dgk tk ldrk gSA bl ys[k esa bldh jpuk rFkk ykHk vkSj iz;ksx'kkyk ,oa {ks= ijh{k.kksa esa izkIr ifj.kkeksa ij izdk'k
Mkyk x;k gSA
la;qDr jkT; vesfjdk esa Lyhijksa dks VkbZ dgk tkrk gSA bl fy, tgka Hkh fons'kh dksMks esa VkbZ dk lanHkZ gks] mls Lyhij gh
ekuk tk;sA
Composite sleepers are new type of environment friendly sleepers made of waste plastics. It has almost all the
flexibilities of wooden sleepers. In many ways, it is even superior to wooden sleepers. In USA, in addition to railway
sleepers, with the material of Polywood make, even many bridges have been constructed. Its use not only reduces the land
pollution from stray waste plastics but also ensure less destruction to the forests. As such, it can be treated as a green
product in two ways. In this paper, its composition including advantages as well as the results achieved in lab test as well
as field trial has been discussed.
(In USA, sleepers are known as ties. Hence, wherever reference of foreign codes in the name of ties is there, the same
should be considered as sleepers).

1.0 Introduction
Sleeper is one of the most important components of the railway track. In olden days, predominantly
wooden sleepers were used. It was having some inherent problems. As such, other type of sleepers like PRC
has been developed for normal track. However, for special locations like bridges and non-standard turnout,
wooden sleeper is still one of the best alternative.
Due to scarcity of wooden sleeper, alternative to wooden sleeper was required to be developed. As
such, RDSO has taken one project regarding development of FRP sleepers and the same was successfully
completed. Initial cost wise, FRP sleepers are comparatively costlier. However, life cycle cost wise, position
of various sleepers for bridges are as below :
Sl. No.
1.
2.
3.

Sleeper
Bridge timber
Steel channel
FRP

Life cycle cost*

Without block cost
With traffic block cost
10725/20376/19386/24744/20733/20733/-

*in Ruppes

*Exectivie Director (Bridges and Structures)/RDSO/Lucknow

52

Initial cost* of the

sleeper
466/10333/20176/-

Keeping in view the cost component, it was tried to search another alternative. In this process, an MOU
was signed with the following two firms, who have shown their interest in development of alternate
composite sleepers in associate on with their foreign collaborators :
i.

M/s Patil Group of Industries, Hyderabad, having association with M/s Tietek/USA.

ii.

M/s Micron, Delhi, having association with M/s Polywood/LSA.

2.0 Composite sleepers as an eco-friendly product : Thermo-plastic based composite sleepers are
eco-friendly and can be categorized as the green product. It is eco-friendly in the sense that it ensures
reuse of the waste plastic in a better way by ensuring its usage in such a fashion, otherwise trees are
required to be cut to have required number of wooden sleepers.turnout sleepers or bridge timbers. The
main constitutents of composite sleepers are : used HDPE, plastics, polystyrene, polyethylene. PVC
and so on, All are falling in the general category of plastic, which has become part and parcel of life in
every field. Furthermore, its application is increasing day by day. Main problem with this material is that
these are non-biodegradable. As such, these used plastic waste are posing environmental hazard to the
society. Environmentalists are having serious concern about its usage as well as disposal of such products.
Non-use of plastic is not possible due to its deep rooted link in the human life. As such, alternative about
its disposal/reuse are required and now a days, it is the serious matter of concern world over.
In any city, plastic is one of the major solid wastes as a non bio-degradable component. For its proper
disposal, scientists had taken the project for its effective reuse. Fortunately, one of its usages in the form
of railway sleepers has come up. By reuse of the plastic, the land pollution, which otherwise was increasing
by non-proper disposal of the same will certainly be reduced. Its use as railway sleepers will ensure
comparatively less cutting of trees for the world over railways. This will further result less destruction to
the forests, which ultimately will ensure more consumption of carbon dioxide and simultaneously more
release of oxygen by the forests to the atmosphere. As such, it is a boon to the society.
Indian Railways have mostly switched over from wooden sleepers to PRC sleepers but in some other
countries, wooden sleepers are still being predominantly used. As per report of Rutgers University/
USA, US rail road on annual basis replace 10 to 15 million wooden sleepers. Even if on the conservative
side, we are taking 15 million wooden sleepers to be replaced on annual basis by two neighbouring
American countries i.e. Canada and USA, its length would be about 38600 kms, which is approximately
equal to circumference of the Earth. If its cubical content is calculated, that is coming to 16,00,000
Cubic meters. Considering total requirement of world over railways, one can get some idea about
quantity of timbers required for use of the same as railway sleepers, which will only be achieved by
deforestation of huge area on the Earth.
3.0 Composition of the composite sleers : Composition of such sleepers are mainly high-density
polyethylene (HDPE) based. Different manufacturers have patented their products based on different
compositions. As such, each material combination followed by different manufacturers will have different
properties from one another just like wood taken from different species of trees have different properties.
Brief ideas about combination of some of the known manufactures are given below :

53

3.1 Tietek Sleepers : Composition of this sleeper is given below Recycled HDPE

55%

Crumbed rubber

12.5%

Glass reinforcfement

12.5%

Fillers

20% (calcium carbonate, Mica etc.)

In addition to that, some other patented items are also there, which are the trade secret of the manufacturer.
3.2 Polywood Sleepers : Its composition is of mainly two items i.e. Polystyrene (PS) and High Density
Polyethylene (HDPE). Its percentage is in the ratio of approximately 35% PS to 65% HDPE. In addition
to that, some other patented items are also there, which are the trade secret of the manufacturer.
4.0 Advantages of composite sleepers
There are so many advantages of composite sleepers, Some of them are listed below :
i.

Composite sleepers being made of non bio-degradable material, its life is expected to be more
than that of wooden sleepers. Life of wooden sleepers is normally 10 to 25 years depending upon
its variety, quality of wood, traffic density and its location of use. As compared to this, life of
composite sleepers is assessed to be around 40 to 50 years.

ii.

All types of flexibilities, which is available with wcoden sleepers like notching, grooving, repairing
of the spike killed area, adging, drilling holes etc are all available with composite sleepers. As such,
it is one of the best replacements of wooden sleeper.

iii. Property of the wooden sleeper is not uniform due to presence of knot and other defects available
in the timbers. Cost of the wooden sleaper also increases which is disproportional with respect to
increase in size. As compared to this, unit cost as well as property of composite sleepers is uniform
in respect to shape and size (including length).
iv. No toxic treatment is involved in composite sleepers like use of creosote in case of wooden sleepers.
It is worthwhile to point out that in USA where environmentalists are very much conscious about
proper disposal of toxic materials, are facing problems regarding disposal of creosote treated
wooden sleepers.
v.

The constitutent material of the composite sleepers are thermo-plastic based. As such, its recycling
is possible.

vi. Composite sleepers consume the waste plastic which otherwise is posing its disposal problem, As
such, it is fantastic case of reuse of waste resources.
vii. More use of composite sleepers will ensure less destruction to the forests.
5.0 Test criteria for passing of the thermo-plastic based composite sleepers : Indian Railways have
developed certain test criteria to ascertain the desired mechanical properties and accordingly passing of
such sleepers. The test criteria are 54

Static Load Test

Impact Load Test

Dynamic fatigue test

Furthermore, some other tests like resistance to ultra violet rays, fire resistance and other tests which
has not been standardized till on date for Indian Railways, was left oto the suppliers of such materials to
submit such reports. Details of the procedure about the various tests are given as below :
5.1 Static Load Test : This test is carried out to assess the adequacy of deisng as per details given below:
l

Load

50t on each rail seat

Loading area

As per size of the bearing plate to be provided.

Rate of loading

5t/min.

Number of samples

5.1.1 Acceptance criteria : No visible cracks should be developed on the outer surface of the sleepers on
holding the 50t load for 5 minutes. Sleeper deflection at rail seat for FRP sleeper was kept as 3mm. In
case of composite sleeper other than FRP, deflection is likely to be more. Minimum to the extent possible
deflection is preferable. Deflection measurement and observation of the cracks under rail seat shall be
done at the interval of 5t load e.g. at 5t, 10t, 15t....
5.2 Impact Load Test : Impact load test is required to assess the shock absorption capacity of the sleeper,
which the sleeper is about to bear during the train running as well as during derailment. The test scheme
envisages dropping of wheel on sleeper placed at 30o slope to horizontal plane at following two locations:
i.

294mm away from centre line of rail toward centre.

ii.

200mm away from sleeper end.

Wheel drop details are :

Weight of wheel - 500 kg.
Height of drop - 75 cm.
No. of drops - 2 at the same location.
5.2.1Acceptance criteria : Only recess should form. No crack should appear on the surface of the sleeper.
5.3 Dynamic Fatigue Test : This test is carred out to assess the structural integrity of the product and to
ensure absence of any void and other inherent manufacturing defects (particularly for FRP sleepers) like
dry patches, resin rich areas, delamination inside the composite sleeper etc. Details about the test are
given as below :
Vertical

4 to 20t on each rail seat.

Horizontal

40% of the vertical load

No. of cycles

2 million
55

Frequency

5 Hz

No. of samples

5.3.1Acceptance criteria : The sleeper should not developed any cracks on the surface of the sleeper,
which is of shear, flexural or torsional rupture (due to local buckling) in nature. Hairline localised cracks
may be permissible provided there is no reduction in the load carrying capacity of the sleeper. Sleeper
deflection at rail seat in case of FRP sleeper was prescribed as 3mm. In case of composite sleeper, the
same should be nearer to this.
6.0 Performance of the composite sleepers : Performance of sleepers in lab test as well as field test,
which is known to RDSO till now is discussed as below :
6.1. Lab Test : Both the products have been lab tested in India as well as abroad. Basic properties requried
as per AREMA as well as Chicago Transit Authority is given in Anenxure -1, alongwith the respective
values achieved against those parameters by the manufactuers of Tietek and Polywood sleepers. AREMA
had introduced Part V in their Code in the name of Engineered Compsite tie in 2003. Similarly, Chicago
Transit Authority specified for composite plastic railroad tie bearing no. CTA 1117-020 of 2002. As
such, these specifications are also of the recent origin. Without any refeence, RDSO had prepared test
criteria about testing of such sleepers. The test criteria has already been discussed in Para 5 above.
Result achieved about such sleepers tested in India will be discussed first, followed by the test results
conducted abroad.
6.1.1Performance under static load test : Under this test, load of 50t is required to be applied at each of
the rail seat and observations to be recorded. Sleeper should not fail in this regard.
Performance of tietek as well as Polywood sleepers is given as below :
S. No.

Applied load in kN

Average deflection of Tietek sleeper

(in mm)

Average deflection of Polywood

sleeper (in mm)

1.

50

1.62

1.91

100

2.45

3.03

150

3.28

3.89

200

4.15

4.57

250

4.98

5.43

300

5.80

6.19

350

6.82

7.04

400

8.20

8.06

10

450

9.73

9.17

11

500

11.83

10.28

12

0
(Load released)

4.29.
(Residual deflection)

1.62
(Residual deflection)

56

Performance of both the sleepers are almost identical as far as deflection/compression is concerned.
After removal of the load, recoupment of the deflection/compression was better in case of Polywood sleepers
than that of Tietek sleepers.
6.1.2Fatigue test : Under this, sleepers are subjected to 2 milion cyucles of dynamic loading, which is
varying from 4 to 20t vertical and 1.6 to 8t horizontal i.e. 40% of the vertical load. Both of these
sleepers were tested at IIT/Madras. Both of the sleepers passed the test criteria and results are almost
identicial.
6.1.3Impact/Derailment load test : Performance of the sleepers in impact/derailment loading is given as
below :
S. Type of sleeper Groove Size after wheel drop in mm
No.
After Ist wheel drop
a
b
c
d
1.
2.
3.
S.
No.

1.
2.
3.

Tietek sleeper
Wooden sleeper
Polywood sleeper

59
61
40

35
41
26

35
50
24

17
21
12

Type of sleeper Groove Size after wheel drop in mm

Tietek sleeper
Wooden sleeper
Polywood sleeper

After 2nd wheel drop

b
c
41
53
57
193
41
32

a
67
81
54

d
26
28
16

Note: Wheel drop hits the corner of the sleeper and

a. Damaged length on the width portion
b. Damaged length on the depth portion.
c. Width of damage in wheel drop.
d. Depth of damage in wheel drop.

Fig. 1: Location of Wheel Drop for Checking the Performance

of Composite Sleepers in Derailment Loading

57

Wheels are dropped from the height of 75 cms twice at the same location and at total 4 locations as as
shown in the Fig.1. Average values of the results in respect of a, b, c & d has also been given in the above
table.
On perusal of the observations, it is clear that as compared to wooden sleepers, Tietek sleepers as well
as Polywood sleepers both performed better. Observations further reveal that with respect to derailment
loading, Polywood sleeper stands first, followed by Tietek sleepers and then after the bridge timber i.e.
wooden sleeper.
6.2 Ageing Test : Ageing is the phenomena, which determines the life of the sleepers. Being non biobiodegradable in nature, composite sleepers are not affected like wooden sleepers. However, it degrades
in other fashion i.e. degradation due to ultraviolet rays. AREMA as well as Chicago Transit Authority
specifies that deterioration of material due to ultra-violet rays should not be more than 0.003 inch/year
(0.076 mm/year). Acceleratec ageing test was got done from reputed testing labs by the respective
manufacturers. Their results are given as below :
6.2.1Tietek sleepers : As per the details submitted by M/s Tietek Inc, through M/s Patil Group, accelerate
ageing test equivalent to 15 years of exposure was got done on Tietek composite sleeper. Details are
given below :
S.
Particulars
No.
1.
2.
3.

Tietek composite sleeper

Soft wood sleeper (Oak)
Hard-wood sleeper (Red
Oak)

Percentage
retention of the
strength
75%
50 to 60%
50 to 55%

The above figures show that deterioration rate of hardwood, as well as of the soft wood is almost the
same with a little bit better performance of softwood. It seems to be somewhat controversial but the factual
position as received, has just been put up for sharing the result. Actually, performance of hardwood is better
than the softwood. However, with respect to composite sleeper, retention of 75% strength equivalent to 15
years exposure seems to be logical since composite materials are non-biodegradable. It only degrades by
ultra-violet rays.
6.2.2Polywood sleepers : Accelerated weather test of the Polywood sleepers was done at University of
Illinois/USA. Result as quoted by M/s Polywood which is equivalent to 20 years of the exposure, is
given below :

58

S. No.
1.

Test Particulars
Tie plate compressive modulus

2.

Face hardness

3.

Spike insertion force

4.

Spike lateral resistance

5.

Spike withdrawal force

6.

% Surface area loss from checks/splits.

Exposure condition
Before exposure
After exposure
Before exposure
After exposure
Before exposure
After exposure
Before exposure
After exposure
Before exposure
After exposure
Before exposure
After exposure

OAK Sleeper
3,200 psi
1,000 psi
3.600 lb.
1,000 lb.
11,000 lb.
3,900 lb
3,500 lb.
2,500 lb.
8,500 lb.
1,900 lb.
0.9%
4.5%

Polywood Sleeper
2600 psi
3100 psi
6,300 lb.
7,800 lb.
7,200 lb.
6,900 lb.
1,700 lb.
6,900 lb.
2600 lb.
3,400 lb.
0.0%
0.0%

Accelerated weather test was got done as per provisions of AREMA contained in Document No. TD
96-010 dates April 1996.
On perusal of the above table, it reveals that except in spike insertion force criteria, in all other respects,
performance of Polywood sleeper improved over age. This is a very unusual phenomenon. further detailed
explanation is not available in the report prepared by Rutgers University of USA, (Annexure - II) which
included the test report of Illinois University/USA. It is worthwhile to point out that Polywood sleeper was
developed at Centre for Advance Materials of Rutgers University, New Jersey, USA.
Based on the test report, it can be concluded that Polywood sleeper may be one of the excellent
composite sleepers of the future. However, details about the mechanisms of improving the performance of
Polywood sleeper with respect to ageing are under correspondence with the manufacturers to ascertain how
the properties are getting improved in due course of time.
6.3 Other Related Test Results : Other related test results as per AREMA standard as well as Chicago
Transit Authority has been shown in Annexure I. Perusal of the Annexure reveals that :
i.

Polywood sleepers are lighter than the Tietek sleepers.

ii.

Coefficient of thermal expansion of Polywood sleepers is more favourable than that of Tietek
sleepers.

iii. Screw spike withdrawal result of Polywood sleeper is also better than that of Tietek sleepers.
iv. Modulus of elasticity in compresison of Tietek sleeper is better than that of Polywood sleepers.
6.3.1About rubber content : Tietek sleeper is containing rubber also, while the Polywood sleeper doesn't.
AREMA standard is silent about specifying any such content. It shows that rubber content is acceptable
to AREMA standard. But the Chicago Transit Authority standard limits rubber contents to maximum
3%. Tietek sleeper as per the composition received through Patil Group is havng 12.5% of rubber
content. Why Chicago Transit Authority is limiting the rubber content of 3% needs to be ascertained for
getting the desired quality of composite sleeper for Indian Railways.
6.4 Field Performance : M/s Patil Group with the proposal of Tietek sleeper approached first. Hence,

59

after successful Lab tests, Tietek Sleepers were provided on the two bridges as per the details given
below :
S. No. Bridge No.
Division/Railway
Date of laying
1.
894 A Moradabad/Northern Railway
26.12.02
2.
42
HWH/Eastern Railway
29.04.03

Fig. 2: Tietek sleeper provided on Bridge No. 42 of Howrah Division

M/s Micron with the proposal of Polywood sleepers approached later on. Their sleepers have also
been successfully lab tested and in many of the parameters, it was found better than that of Tietek sleepers.
These sleepers are yet to be provided on some of the bridges for field tests.
Report reveals that both types of composite sleepers have already been provided in large numbers on
US rail roads and performing well.
Report further reveals that Polywood sleepers have also been provided for points and crossing. Even
with the composition of Polywood inc., 3 bridges as poer details given below have been constructed i.

One bridge of 25ft span was constructed in 1999 for Jeep and Pedestrain at Fort Leonard Wood
in Missouri/USA.

ii.

Another bridge on 30ft span was constructed in 2000 on nature trail along Hudson river in New
Baltimore/USA..

iii. Third bridge having span of 40ft was consutrcted in 2002 on Mullica River in USA.

60

Fig. 3: Polywood Sleeper Provided on One of the Turnout of US Rail Road

6.4.1Performance about gauge retention : On Moradabad Division, the sleepers were provided during
the extreme cold. It was inspected by the author of this paper during the extreme hot period i.e. in June
2003. While laying the sleepers. - 3mm gauge was maintained. At the time of inspection, gauge was
found to be +4mm. As such, there is variation of gauge of 7mm from extreme cold to extreme hot
weather conditions. Considering extreme cold to extreme hot temperature variation of about 40oC and
taking coefficient of thermal expansion as 7.5x105 as conveyed by the manufacturers, variation in length
is coming to 5mm on gauge length i.e. 1673 mm.
Composite sleeper by nature is having more coefficient of thermal expansion as compared to other
sleepers like concrete, steel and wooden. Variation of 7mm in gauge from extreme cold to extreme hot was
taken seriously by Railway Board. However, this is not the proble matic area, if initial adjustment is being
done at the time of laying the sleepers. Regarding this, some literature survey was done. One of them is
AREMA Part V of 2003. It very beautifully explains the thermal behaviour of composite sleepers. Relevant
provisions of Pt. V of AREMA of 2003 are reproduce below :
"Characteristics properly of engineered polymer composites (EPC) ties is higher thermal expansion
coefficient than wood, concrete, steel or engineered wood product (EWP). Theoretically, this means that
EPC ties could grow longer or shorter with changes in temperature as compared to the other tie materials for
the same changes in the temperatures. However, field experience in a variety of locations and climates as
shown that gauge is not affected to the degree predicted by direct calculations. The specification requirement
for thermal expansion ensures that rail gauge will be maintained over a wide range of operating temperature.
While engineered polymer composite ties will undergo, at least, some dimensional changes due to change
in temperature, it should be noted that these changes do not occur instantaneously with the change in ambient
air temperatures. The polymer matrix materials used in these ties are inherently poor heat conductors. While
the surface of the tie may exhibit, a repeated change in temperature (i.g., when exposed to direct sunlight) the
61

bulk of tie will not. Change in gauge will only not be seens as a result of changes in temperature over a single
day-to-night cycle. However, over more long term seasonal changes in temperature, gauge dimensions will
be affected - increasing as temperature gets higher and decreasing as the temperature gets lower as the bulk
of the tie material has time to reach a thermal equilibrium at the new seasonal temperature.
If EPC ties are installed at ambient temperatures below 4oC or above 38oC, gauge should be adjusted
by 0.125" (3.2 mm) (tighter at cold temperatures installations, greater at high temperatures) or otherwise
recommended by the manufacturer. Since gauge adjustment exercise while laying the sleepers at bridge No.
894A was not done, hence variation of gauge from -3 mm to + mm in view of the explanation given in the
AREMA Code is not a matter of concern.
6.4.2Performance about-spike killing : While inspecting the sleepers provided at the above-mentioned
two bridges, one more problem i.e. spike killing was reported on Bridge No. 42 on Howrah division of
Eastern Railway but the same was not reported on bridge No. 894A on Moradabad Division of Northern
Railway. Spike killing tendency reported on HWH division might be due to change in material quantity.
It is worthwhile to mention that the composite sleepers provided on different bridges were of different
lot received at different time. To ensure that no spike killing tendency particuarly at early stage, sleeper
quality needs to be maintained by the manufacturer. In due course of time, spike-killing problem may
arise as we are facing with wooden sleepers. Regarding this, just as we have its solution in wooden
sleeper by insertion of wooden gulli, somewhat similar provision had been made in AREMA Code.
Here Para 5.4.2.8 of the AREMA Code is worth mentioning which is reporudced below :
"Tie plugs may be used in EPC ties, basically in the same fashion as they are used in sawn ties. The
specification for wooden tie plugs are located in Article 3.1.5. Polymer based plugging compounds (e.g.
polyurethane) may also be used".
M/s Patil Group, representative of tietek sleeper came with the idea that just like plugging the wooden
sleeper, Tietek Sleepers can also be plugged with the help of gulli made of the same material of which tietek
sleeper is composed. It is expected that, if the provisions are working satisfactorily in foreign countries, there
is no doubt that the same will not give similar result here also. However, all such provisions needs to be tried
on Indian Condition also for conformation.
6.4.3Riding quality and compression : Till now, riding quality has been reported to be good on bridges
provided with such sleepers Adverse compression has also not been reported.
7.0 Conclusion : Literature survey, test results and performance of the sleeper abroad as well as in this
country, reveals that composite sleeper may be one of the alternative for bridge timber, it is also having
potential in use of the same for non-standard turnout where it is not possible to standardize the PRC
sleepers. Composite sleeper will ensure use of plastic waste, which is creating otherwise environmental
hazard in the form of land pollution as well as choking of the sewerage system. furthermore, adoption of
composite sleeper will ensure less destruction to the forests. In case of import of such sleepers to be
used as bridge timber, cost of the same is likely to be between Rs. 7000 to Rs. 8000 per sleeper. In
case, the same is manufactured in this country itself, cost of the same may come to around Rs.5000/
sleeper. As such, it would be better that such sleepers to the requirement of Indian Railways may be got
manufactured here itself.
62

Annexure - I
Comparison of Various Standards and its Respective Values for Tietek and Polywood Sleepers
S. Mechanical properties/ test method
No.

Size of test
specimen

1. Specific gravity ASTM D 6111-97

2. Density ASTM D6111-97

7" X 9"
7" X 9"

3. Coefficient of Thermal Expansion

ASTM D696-98

7" X 9"

4. Compressive Strength ASTM D6108-97

(Compression Parallel to Grain)
5 Compression Strength ASTM D6108-97
Rail Seat Compression (Compression
Perpendicular to Grain)
6 Permanent Deformation Under Load
(Compression Perpendicular to Grain)
ASTM D6108-97
7. Modules of Elasticity (Compression)
ASTM D6108-97
8. Flexural Strength ASTM D6109-97

4" X 6"
7" X 9"

Chicago
Result of sleeper supplied
AREMA
transit
by
Standards
authority
M/s Tietek M/s Polywood
standards
0.90
Not specified
0.85 to .90
45.0 lbs per cu Not specified 57-66 lbs/cft
53-56
ft.
lbs/cft.
0.000075
75. x 10-5
0.00005
0.00007
Inch/Inch/oF inch/inch per inch/inch per oF
in/in. per oF
o
F
(max)
3,000 psi
4300 psi
minimum
1,000 psi
1200 psi
minimum

7" X 9"

0.015 inch
(max)

0.0135 inch

4" X 6"

170,000 psi
minimum
2,500 psi
minimum
200,500 psi
minimum
1,000 psi
minimum
2,500 lbs.
minimum

1700000 psi

170000 psi

175000250000 psi
2000 - 2500
psi
-

-220000 psi

1500 psi-

Spike 1900 lbs

Screw Screw
5000 lbs.

5000 lbs

15000 lbs.

4 " X 6"

9. Modulus of Elasticity (Flexural)

ASTM D6109-97
10. Shear Strength ASTM D6109-97

4" X 6"

11. Mechanical Fasteners ASTM D6117-97

Screw Spike Withdrawal

7" X 9"

4" X 6"

2000 psi

Note: "-"shows that the respective values are not known

x x x

63

3000 psi

FABRICATION

INTERNAL DEFECTS IN STRUCTURAL STEEL

FOR BRIDGES
R.K. Goel*
(Published in Journal of IPWE, Sept., 2009)

1.0 INTRODUCTION
Steel plate is produced by a process of rolling thick slabs of steel into approximately rectangular plates.
During the process the material is passed through rollers many times, each pass resulting in a decrease in
thickness and an increase in area. Inevitably, the plate material produced by this process is anisotropic. Any
discontinuities that may be present (for example resulting from bubbles or slag trapped in the cooling slab) are
elongated in the direction of rolling and laminar in form.
2.0 ROLLING DEFECTS
2.1 After rolling, plates are usually sheared or flame-cut to a rectangular size at the mills. This may mask
laminar imperfections at the edges/ends, but any such imperfections would be revealed when the fabricator
trims the plate. Laminar imperfections can occur within the body of a plate or at the edges. However,
even mid-plate laminations are likely to appear as edge laminations when the fabricator strips out a
flange or web from the rolled product. Laminations of significant size (area) would impair the structural
performance of welded attachments to the plate surface, and might even give rise to local buckling
failures in an element under in-plane compression. Edge laminations would reduce the integrity of
connections made at or near the edge of a plate, for example in box girders or fabricated channel
sections. Edge defects would also create a potential weakness against corrosion, either as a result of
entrapment of dirt or moisture behind a coating, or by making a discontinuity in the coating; any consequent
local corrosion would tend to open the lamination and aggravate the problem.
2.2 In early steels, and particularly in wrought iron, discontinuities were a significant proportion of the plate
area and the through-thickness performance of the plate could not be relied upon. Modern steels are
generally very clean and laminar imperfections produced by modern steelmaking processes are usually
modest in size and extent, and can be detected by ultrasonic testing (and visually on clean edges). The
quality control procedures in modern steel mills can generally be guaranteed to detect any gross defects
and prevent them from getting into the supply material. Problems usually only arise in material from older
(and often uncertified) mills, which may be supplied through some stockholders.

*Director/Steel Bridges-I, B&S Dte., Research Designs & Standards Organisation, Ministry of Railways, Lucknow

67

3.0 ACCEPTANCE LEVELS FOR INTERNAL IMPERFECTIONS

BS 5400-6 refers to BS 5996 for acceptance levels for internal imperfections. Document BS 5996:
1993 defines six basic acceptance levels (designated B1 to B6) for imperfections within the material body,
and three basic levels for imperfections at the material edge.
4.0 REMEDIAL/REPAIR PROCEDURES
Often the situation arises where unacceptable defects are discovered at a stage in the work when
replacement is not practicable without jeopardizing the completion programme. There are procedures that
have been used successfully in the past to overcome this type of difficulty.
It is important that any repair conforms to the same required standards of workmanship as any other
welding on the fabrication. Special attention needs to be given to the finishing and, where necessary, surface
dressing, to ensure that the fatigue class of the repair is no worse than that required by the designer at that
particular location. It is for this reason that no such repairs should be carried out until the formal approval of
the Design Engineer is obtained.
4.1 Repairs to edges
Shallow edge imperfections, particularly those caused by the overlapping in rolling, can be dealt with by
grinding in from the edge to make a groove suitable for welding, and employing an appropriate procedure to
weld the edges of the plate together. Where the edge imperfection is wider, usually part of an area of delamination in the middle of a larger plate which has been exposed by stripping into narrower pieces, the edges
can be bound together as above and the wider area treated as below.
4.2 Repairs in the middle of plates
In the very unlikely situation necessitating to carry out a repair to the middle of a plate, the affected area
can be clamped together by use of HSFG bolts or by plug welding. If such a repair is contemplated, the size
and spacing of any bolts or welds must be considered carefully. Plug welding might be suggested for better
appearance, although it is very difficult to completely disguise a field of plug welds. A better solution is to cut
out the affected area and let in a new piece of plate, if this can be done.
4.3 Testing
Any weld repairs should be thoroughly tested, by Magnetic Particle Influx (MPI) for single run repairs
and by MPI and Ultrasonic Testing (UT) for deeper repairs.
5.0 MATERIAL WITH THROUGH THICKNESS PROPERTIES
A requirement for material with through thickness properties is a different requirement from that limiting
the extent of internal imperfections. However, specifying through thickness properties in accordance with BS
EN 10164 invokes a requirement for inspection for laminar imperfections of a high sensitivity. Such inspections
should identify any significant imperfections.

68

6.0 INTERNAL DEFECTS IN ROLLED SECTIONS

Rolled sections are susceptible to similar problems but, once again, the quality controls in modern
manufacturing facilities should prevent unacceptable material reaching the market.
When dealing with older rolling plants, or with products from less well controlled mills, sometimes there
are inclusions in the corner of an angle or channel or at the web-to-flange junction of a beam or Tee. Except
in very thin sections this would seldom be a structural or durability problem, but might need to be dealt with
where a gusset or cleat is attached near the corner (angle/channel) or in the middle of the flange (beam/Tee).
7.0 CONCLUSION
It is evident that inspection of raw structural steel for fabrication of bridges is quite important to ensure
quality of fabricated steel girders. Adequate attention need to be given to identify the common defects and
take remedial measures. There is also a need to increase awareness and skills of bridge engineers in this
regard.
8.0 REFERENCES
1.

BS 5400-6: 1999, Steel concrete and composite bridges. Specification for materials and workmanship, steel.

2.

BS 5996: 1993, Acceptance levels for internal imperfections in steel plate, strip and wide flats, based on ultrasonic
testing. (Withdrawn).

3.

BS EN 10160: 1999, Ultrasonic testing of 3. steel flat plate product of thickness equal to or greater than 6 mm
(reflection method).
BS EN 10164: 2004, Technical delivery conditions, Steel products with improved deformation properties perpendicular
to the surface of the product.

x x x

69

FATIGUE TOLERANCE OF WELDED CONNECTIONS

IN RAILWAY BRIDGES
R.K. Goel*
(Published in Indian Railway Technical Bulletin, R.D.S.O. Lucknow, Feb., 2008)

SYNOPSIS

bl rduhdh ys[k esa oSYMsM tksM+ks ls lEcfU/kr QfVx fo"k; ij tkudkjh nh x;h gSA ;g mu oSfYMax
o bathfu;fjax ds tkudkjksa ds fy, gS tks fd LVhy ds LVDpjksa ds fuekZ.k ;k QfVx ds dkj.k =qfVxzLr LVDpjksa
dh ejEer esa yxs gSaA bl rduhdh i= esa oSYMsM duD'ku~l dh QfVx dSVsxjh dk fu:i.k djrs gq,s ;g
n'kkZ;k x;k gS fd fdl izdkj oSfYaM+x ds nkSjku fd;s tkus okys lq/kkj ls tksM+ dh {kerk c<+ tkrh gSA
The paper introduces the fatigue phenomenon as it applies to welded connections. It is intended for
welding and engineering personnel engaged in the fabrication and repair of structures subject to fatigue loading
and potential fatigue failure. It briefly describes the fatigue-cracking phenomenon and summarises the concept
of "detail category" for welded structural components.
1.

INTRODUCTION

Fatigue is defined as cumulative, localised and permanent damage caused by repeated fluctuations of
stress sometimes below the static design stress of the structure. This cyclic loading can lead to gradual
cracking or even catastrophic failure of a structural element. Fatigue accounts for more service failures than
any other failure mechanism in metal structures subjected to dynamic loading.
2.

BASIC CONCEPTS OF FATIGUE

2.1 The severity of the repeated fluctuations

of stress can be characterised by
maximum stress, minimum stress, stress
range and the number of cycles, as
shown in Figure 1. To determine a
materials resistance to fatigue failure,
samples are prepared and subjected to
a fluctuating stress until failure occurs.
Repeated testing at different stress
ranges results in a relationship between
applied stress and cycles to failure as
Figure 1 - Idealised Fluctuating Load Cycle
*Director/Steel Bridges & Structures Directorate, Ministry of Railways, RDSO, Lucknow (UP) - 226 001

70

represented in Figure 2. This figure introduces the concept of a fatigue limit or endurance limit implying
that below a given stress failure by fatigue will not occur. However, all materials subjected to cyclic
stresses will eventually fail by fatigue, so the line never actually becomes parallel to the x-axis.

Figure 2 - Fatigue Endurance Curve

2.2 The fatigue tolerance of metals can be described by the number of cycles it can tolerate before failure
occurs at a given stress range. As can be expected the smaller the stress range, as shown in Figure 1, the
more cycles the material can tolerate. The number of cycles the material can tolerate increases in a
logarithmic fashion as the stress range is reduced. Thus a log-log stress-cycle diagram produces a
straight line as indicated in Figure 3. This method of depicting fatigue performance is known as an SN
curve where 'S' is Stress (range) and 'N' is Number of cycles to failure. To produce the S-N curve
shown in Figure 3 it is necessary to carry out many tests. There will invariably be scatter in the results,
thus it is necessary to establish a relationship that represents a conservative tolerance level.

Figure 3 S-N Curve for a Structural Steel presented on a log-log scale

71

2.3 The predictable fatigue tolerance curve depicted in Figure 3 is only applicable to identical smooth and
uniform fatigue specimens subjected to idealised fluctuating loads. In the real world, the fluctuating loads
to which structures are subjected vary both in intensity and duration. Also fabricated structures contain
stress concentrations with the result that welded components have a poorer tolerance to fluctuating
loads than their non-welded counterparts. Nevertheless there are ways to mathematically deal with
fluctuating stresses of differing magnitudes and the S-N curve provides a well-established, and now
universal method of displaying fatigue tolerance for welded joints provided a suitable lower bound
failure probability is applied.
3.

FATIGUE INITIATION AND PROPAGATION

Fatigue has two parts, initiation and propagation. On a sub-microscopic scale the imperfections in the
metal's internal structure, known as dislocations, play a major role in the fatigue crack initiation phase. After
a large number of loading cycles dislocations pile up and form structures called persistent slip bands. These
leave tiny steps in the surface that serve as stress risers where fatigue cracks can initiate. Once a fatigue crack
has initiated, it then propagates under the action of the applied fluctuating load. Fatigue crack propagation
continues until the component has insufficient cross section to carry the load. The controlling event that
determines the life of the component can either be the initiation event or the propagation event. In a welded
structure there is essentially no initiation event since there are generally enough pre-existing flaws in a welded
structure that the component is immediately in the propagation phase.
Stress raisers and the operating environment also greatly influence the fatigue performance of steel
structures. The fatigue curves presented in Figure 4 show the deterioration of fatigue performance from a
mirror polished specimen tested in air to that exhibiting surface roughening due to underwater salts.

Figure 4 Effect of surface finish and wet environments on fatigue performance

72

4.

DETERMINING FATIGUE PERFORMANCE OF WELDED STRUCTURES

Welded components are less tolerant to fluctuating loads than their non-welded counter-parts for three
reasons:
a)

Welds contain internal flaws which act as the initiation site for crack propagation;

b) Welds create external stress raisers which act as the initiation site for crack propagation;
c)

The process of welding introduces residual stresses in the region of the weld exacerbating the
applied fluctuating stress.

The fatigue tolerance of welded connections can be classified into "detail categories" according to the
type of weld and its orientation with respect to the applied fluctuating loads. The detail categories for steel
structures are used by structural steel designers when fluctuating loads occur during service. The detail category
for any given weld configuration is a number between 36 and 180 that represents the stress range in MPa that
can be tolerated for two million (2x106) fluctuating load cycles.
5.

CLASSIFYING THE FATIGUE TOLERANCE OF WELDED STRUCTURES USING THE

DETAIL CATEGORY

The following detail categories are extracted from the Australian Design Standards for Steel, AS4100
and AS 5100. The significant aspects of the detail categories are:
a)

Welds lying parallel or transverse to the applied cyclic loads;

b) Welds with or without external reinforcement

c)

Welds with or without external discontinuities such as stop-starts and undercut;

d) Fillet welds, partial penetration welds or full penetration butt welds;

e)

Structural members with or without welded attachments.

5.1 Detail Category 180

Plain steel component with all surfaces machined and polished. All changes in cross section are smooth
and gradual. A category 180 structural element is analogous to the uniform polished specimens used to create
the S-N curves depicted in Figure 3. Such an element can tolerate a stress range of 180MPa for 2x106 cycles
and serves as a standard reference by which all weld details can be classified.
5.2 Detail Category 140
As-rolled steel component or hollow section with no flame cut edges joined with a full penetration butt
weld lying parallel to the direction of the fluctuating stress with all reinforcement machined off. The quality of
the weld would meet the requirements of specific welding specification.
5.3 Detail Category 125
Steel component made with butt or fillet welds lying parallel to the fluctuating stresses made by automatic
welding processes without stop/starts.
73

5.4 Detail Category 112

A) Steel component made with butt or fillet welds lying parallel to the fluctuating stresses made by
manual or semi-automatic welding processes containing stop/starts.
B) As-rolled steel component with no flame cut edges joined with a full penetration butt weld lying
transverse to the direction of the fluctuating stress with the weld reinforcement ground flush with the
parent metal. The quality of the weld would meet the requirements of specific welding specification.
5.5 Detail Category 80
As-rolled steel component joined with a full penetration butt weld lying transverse to the direction of the
fluctuating stress with the weld reinforcement in the as-welded condition. The quality of the weld would meet
the requirements of specific welding specification and it may have been made by manual or semi-automatic
welding processes in any welding position.
5.6 Detail Category 71
A) As-rolled steel component joined with a full penetration butt weld lying transverse to the direction
of the fluctuating stress made using a backing strip.
B) A steel component carrying fillet welded non-loaded attachment(s)
C) A full penetration (fillet reinforced butt weld) cruciform joint with the welds transverse to the applied
stress
5.7

Detail Category 45

Steel components under the action of fluctuating loads with attachments welded at the edge or close to
the edge of the component.
5.8 Detail Category 36
Components made with fillet welds or partial penetration butt welds lying transverse to the fluctuating
loads
5.9 Designing Welded Structures Using Detail Categories
Figure 5 shows a compendium of S-N curves from for all the detail categories including the base-line
Category 180 for a non-welded component. With knowledge of the Detail Category the designer of a welded
structure can either design for a fixed life expressed by a number of cycles or an indefinite life based on the
structure being able to tolerate up to one hundred million (108) cycles. If the design called for a fixed life of 2
million cycles the Detail Category "f rn" could be used. For a fixed life of 5 million cycles the fatigue limit "f3"
is appropriate. If an indefinite life of up to 100 million cycles is required then the cut off fatigue limit "f5" is the
stress range to be applied. Table 1 summarises how the tolerable stress range in MPa changes depending on
the required life of the component. The percentages given in brackets refer to the relative fatigue performance
of the welded connection compared to the base line Detail Category 180 for a fixed 2 million-cycle life.
74

Number of stress cycles (nsc)

Figure 5 S-N curve of steel under normal stress

Table 1 can be used as a quick guide to determine the fatigue performance of a given weld configuration.
For example a fillet weld transverse to the fluctuating load is a detail category 36. This means it can tolerate
a fluctuating stress of 36 MPa (based on throat thickness) for a 2 million-cycle life. This is 20% of the fatigue
strength provided by a non-welded category 180 components.
If an indefinite life of up to 100 million cycles is required for the same fillet weld then the allowable
fluctuating stress must be reduced to 15 MPa. This is only 8% of the fatigue strength of the non-welded
component with a life of 2 million cycles.

75

Table 1. Relationship Between Tolerable Stress Range and Design Life for
Weld Detail Categories
Detail Category and stress range Tolerable stress range (MPa) for Tolerable stress range (MPa) for
up to 5 x 106 Cycles
up to 108 Cycles
(MPa) for up to 2x106 Cycles

6.

180 (100%)

133 (74%)

73 (41%)

160 (89%)

118 (65%)

65 (36%)

140 (78%)

103 (57%)

57 (32%)

125 (69%)

92 (51%)

51 (28%)

112 (62%)

83 (46%)

45 (25%)

100 (55%)

74 (41%)

40 (22%)

90 (50%)

66 (37%)

36 (20%)

80 (44%)

59 (33%)

32 (18%)

71 (39%)

52 (29%)

29 (16%)

63 (35%)

46 (26%)

25 (14%)

56 (31%)

41 (23%)

23 (13%)

50 (28%)

37 (21%)

20 (11%)

45 (25%)

33 (18%)

18 (10%)

40 (22%)

29 (16%)

16 (9%)

36 (20%)

27 (15%)

15 (8%)

ATTACHMENTS

Special attention must be drawn to attachments welded onto structural members subject to cyclic loads.
As soon as an attachment is welded onto a structural member the fatigue performance drops significantly to
a design category of 45. Catastrophic structural failures have occurred due to the addition of seemingly
innocuous attachments that have been welded to structural members.
Attachments shall not be welded onto structural members without express instructions from the design
engineer attachments shall be welded in such a manner that the start and stop of the weld does not provide
crack initiation sites.
7.

IMPROVING THE PERFORMANCE OF WELDED CONNECTIONS

7.1 Weld Configuration

Figure 5 and Table 1 demonstrate that the weld configuration has a significant impact on the resultant
fatigue performance. Table 2 provides some readily available methods to improve or optimise the fatigue
performance of a welded joint by changing the joint configuration. An improved weld configuration should
always be the first choice when striving to optimise fatigue performance.
76

Table 2. Fatigue Improvement Methods for Welded Connections

Original Weld Detail

Category Improved Weld

Detail
Category

Partial Penetration Butt Weld

36

Complete Penetration Butt Weld on a

backing strip

71

Fillet Weld transverse to loading direction

71

Fillet reinforced full penetration Butt

Weld transverse to loading direction

80

Fillet reinforced full penetration Butt Weld

transverse to loading direction

71

All undercut, discontinuitys

convexity removed by grinding

and

80

Complete Penetration Butt Weld on a

backing strip

71

Single Vee Butt weld (transverse to

loading direction)

90

Single Vee Butt weld (transverse to

loading direction)

80

Double Vee Butt weld

112

Single Vee or Double Vee Butt Weld

80 or 90 Butt weld with all reinforcement

ground off flush with member

112

Intermittent Fillet welds parallel loading

direction made from one side

80

Continuous Fillet Welds parallel to

loading direction without stop starts
made from one side

90

Double sided fillet or butt welds parallel to

loading direction

112

Double sided fillet or butt welds made

with automatic methods with no stopstarts

125

Double sided fillet or butt welds parallel to

loading direction

112

Stop-starts and stress raisers removed

by grinding

125

7.2 Post Weld Fatigue Improvement Techniques

For welded structures it is possible to apply post weld treatments to improve the fatigue performance.
Remedial work can be applied to welds in the as-welded condition to increase the cyclic stress range that the
weld detail can tolerate for the given design life. The remedial work consists of either removing crack initiators
or reducing residual tensile stresses.
There are a number of post weld improvement methods available. These are classified in Figure 6.

77

P o s t W e ld
Im p r ov em e n t
M e th o d s

W e ld
g eo m e te ry
im p r ov e m e nt
M e th o ds

M ac h inin g
m e th o ds

R e m e lting
m e th o ds

R e s idu a l
s tr e ss
m e th o ds

S pe cia l
w e ldin g
te c h n iqu e s

M e c ha n ic a l
m e th o d s

T he r m al
m e th o ds

M iling
a nd
la th e
tu rnin g

T IG
d r es sin g

W e ld
p rofile

S h o rt
P e e ning

L o ca l
c o m pr e ss
ion

B urr
g rid ing

P las m a
d r es sin g

S p e cia l
e lec tro de

H a m m er
P e e ning

T h e rm a l
s tr e ss
r e lief

U ltra s on ic
im pa ct

s p ot
h e e lin g

Initia l
o ver
lo ad ing

G u nn e rt
m eth od

D is c
g rid ing

Figure 6 - Post Weld Fatigue Performance Improvement Methods

8.

FATIGUE IMPROVEMENT METHODS AVAILABLE FOR FABRICATION SHOPS

8.1 Post Weld Improvement Techniques

Of the fifteen post weld improvement methods mentioned in Figure 6 there are some that can be readily
applied with the minimum of cost using an angle grinder, burr grinder and rotary flap wheel. Ultrasonic impact
testing (UIT) is also emerging as a useful mechanical post weld improvement method.
8.2 Welding
The welding consumables and welding process shall be selected to produce welds that result in weld
beads that blend into the parent metal without excessive reinforcement and without a tendency to produce
undercut. Automatic welding methods are preferable to manual or semi-automatic methods. Unduly course
weld ripples and stop-starts are detrimental to fatigue performance and should be removed by grinding.
Where possible, run on and run off tabs shall be used for butt welds. The weld shall extend into the run
on run off tabs by a distance at least equal to that of the thickness of the parent material and joint shall be
completely filled. The run on and run off tabs shall be removed and the ends of the weld ground flush with the
grinding marks parallel to the direction of loading.

78

8.3 Grinding
Welds, or any sections of weld, which are considered to have excessive reinforcement or unacceptable
profile shall be ground in order to improve fatigue performance. Weld toes may also be ground, with the burr
grinder preferred for these situations.
Rough grinding may be carried out using a grinding stone or angle grinder but care shall be taken to
ensure that deep score marks are avoided. Final grinding shall be carried out using barrel grinders or flap
wheels and should be selected to ensure that the diameters are appropriate for the size of weld in question.
Local burr grinding of the weld toes may be carried out but thinning of the parent material in any instance must
not exceed 5% of the wall thickness.
Any grinding marks visible after completion of the work shall be parallel to the loading direction.
8.4 Burr Grinding
Rotary burr grinding generally provides the minimum amount of dressing necessary to obtain an acceptable
weld. Burr grinders are generally tungsten carbide tools which come in a variety of shapes and sizes. The type
of burr selected should be appropriate for the particular application. Burr grinding marks shall be parallel with
the main direction of loading i.e. transverse to the weld.
8.5 Non-Destructive Examination
Non-destructive examination as specified should be carried out after all grinding and de-burring
workmanship has been completed. Equally important is performing 100% visual inspection to ensure that no
undercutting or gouge marks are evident after the grinding operations.
8.6 Training
Examples of various types of welds are required to show acceptable surface finish. Personnel should be
trained and samples used as benchmarks of acceptable workmanship. The samples should represent the
types of welds being produced.
9.

CONCLUSION

9.1 It is evident that the fatigue tolerance of welded connection is highly sensitive to the quality of welding,
technique of welding and methodology adopted during welding.
9.2 Fatigue performance of welded connections can be significantly improved by changing the configuration
of the welded joint. Fatigue performance can also be improved by adopting post weld treatment
techniques.
9.3 Welded attachments significantly reduce the fatigue performance of structural members and therefore to
be avoided.
9.4 Welded connections when adopted require a good understanding of the fatigue behaviour on the part of
the designer and the engineer incharge of fabrication, as quality control is very important in giving a good
fatigue tolerant connection.
x x x

79

CONSTRUCTION OF STEEL GIRDERS - PRACTICAL

CONSIDERATIONS IN DESIGN OF WEB - FLANGE
CONNECTION
Ravindra Kumar Goel*
(Published in Journal of IPWE, July 2005)
ABSTRACT
Indian Railways is in the process of adopting welded connections for the design of Railway Bridges. The design of
connection is to be done with respect of the fatigue strength for a specified number of cycles of maximum and minimum
stress, to which the bridge component is subjected. The economy of the design depends not only upon the type of weld
and welding methodology adopted but also the ease with which the fabrication can be done. The practical difficulties
associated with the fabrication of girders with welded connections are to be visualized by the designer so as give an
economical design. The paper briefly describes the considerations involved in the design of web & flange connection for
a railway plate girder bridge.

1.0 INTRODUCTION
Double sided filet welds have been conventionally used in web to flange connection of standard plate
girders designed by RDSO. The fillet weld is designed to carry the design horizontal shear at the top edge of
the web plate. Various design considerations and the codal provisions are to be kept into consideration for
the design of this connection. This is the most important connection to be made using advanced welding
technology such as submerged arc welding. The design parameters and constraints involved in the design of
12.2m span welded plate girder to MBG loading for 10 million cycles of fatigue are described in the following
paragraphs.
2.0 DESIGN PROVISIONS
IRS Welded Bridge Code lays down the stipulations regarding maximum permissible stresses to be
taken in different type of welds for design of dynamically loaded structures from fatigue considerations. Since
fatigue strength of welded structures depends upon the constructional details, the type of weld is to be
decided before the permissible stresses and consequently the size of members and the welds are determined.
Stresses due to dead load, live load and impact, stresses resulting from curvature and eccentricity of track,
and secondary stresses as defined in clause 3.3.2 (a) of IRS Steel Bridge code, only are to be considered for
effects due to fatigue. All other items mentioned in clause 3.1 of IRS Steel Bridge code and secondary
stresses as defined in clause 3.3.2 (b) thereof are to be ignored when considering fatigue.

*Director/Steel Bridges-I, Bridges & Structures Directorate, Ministry of Railways, RDSO, Lucknow (UP)-226011.

80

3.0 PERMISSIBLE STRESSES IN WELDS

(a)

Fillet welds
i)

The basic permissible stress in fillet welds based on thickness equal to the throat thickness shall be
100 N/mm2 (10.2 kg/mm2) where a fillet weld is subjected to shear stress in two directions, the
actual stress shall be taken as the vector sum of the separate shear stresses and shall not exceed
100 N/mm2 (10.2 kg/mm2).

ii)

Load carrying fillet welds in dynamically loaded structures shall be designed such that the stress on
the total effective area of fillet welds does not exceed the relevant values specified in Table for class
'G' Constructional details of Appendix 'G' to IRS Steel Bridge code, subject to a maximum of 100
N/mm2 (10.2 kg/mm2).

iii) Load carrying fillet welds in dynamically loaded structures shall be designed so that secondary
bending stresses are not developed (e.g. single lap joints shall not be used).
(b)

Butt Welds

Stresses in butt welds shall not exceed the permissible stresses of the parent metal as specified in IRS
Steel Bridge Code. However, such welds are to be tested for their strength using non-destructive testing
techniques such as radiography, ultrasonic testing or any other equally effective technique.
4.0 DESIGN OF WELD BETWEEN FLANGE PLATE & WEB
Based on the existing design stipulation attempt was made to design a load carrying double sided fillet
weld for the above connection with following parameters:
i)

Clear span

= 12.2m

ii)

Type of loading

= MBG

iii)

No. of fatigue cycles

= 10 million

iv)

Dead load

= 10.24t

v)

Super Imposed Dead Load

vi)

Dead load + Super imposed dead Load +

= 283.86t

5.32t

Impact load + Wind load

4.1 Calculation of Permissible Stress in Fatigue

Stress ratio

f m in
f m ax

10.24 5.32
283.86

0.05482

Accordingly, for class 'G' for 1 x 107 cycles permissible stress in fatigue (tensile) has been worked out
from Appendix 'A' of SBC as under:
81

3.3

Permissible Stress

3.8 - 3.3
x 0.05482
0.1

= 3.574 kg/mm2
= 0.3574 t/cm2
4.2 Calculation of Minimum Size of Fillet Weld
From maximum bending moment consideration, minimum cross section required is as shown in figure 1.
For this cross-section the critical design parameters are as under:
i)

Horizontal shear

0.48 t/cm

ii)

Vertical shear

0.706 t/cm

0.854 t/cm

Therefore, resultant shear (R)

Hence, minimum size "S" of double sided fillet weld is computed as below:
2 x 0.7 S x0.3574 = 0.854
or

S = 1.706 cm = 17.06mm, Say 18mm.

4.2 Limitations on the Fillet Size

As calculated above a minimum size of fillet weld of 18 mm is required. However, the same is considered
un-acceptable as making of 18 mm fillet weld in workshop would require at least two passages of welding
using submerged arc welding technique. This will involve additional work of welding at the fabrication stage
and will adversely affect the productivity besides being uneconomical. Further from metallurgical point of
view it is found not desirable to provide a fillet weld of size greater than the minimum thickness of the connecting
plates. Therefore the option available was either to increase the thickness of web plate or to go for better type
82

of welding. The first option was considered uneconomical and impracticable from fabrication point of view.
On the other hand it was noted that IRS Welding Code vide Clause 13.3 permits higher fatigue strength of
butt welds i.e. equal to that of the parent metal. Therefore, it was found practicable to design the connection
using double bevel butt weld.
5.0 Considerations in Double Bevel Butt weld
5.1 Special Consideration
For butt welding, special considerations are to be made before welding for the preparation of fusion
faces, angle of bevel, root radius and root face so that the limits of accuracy required in welding can be
achieved. The detailed provisions in this regarding are specified in IS-9595 and other related application
standard. However, when no appropriate application standard exists it is recommended that, for manual
welding, the tolerance on limits of gap and root face should be + 0.5 mm on the specified dimensions for
material upto and including 12mm thick. The tolerance on the included angle between the fusion faces of a V
preparation is recommended to be + 5o. However, practically such tolerances are for guidance only. Closer
tolerances are necessary for automatic process and particular requirements depend on the characteristics of
the process. It should be noted that the quality of root face preparation and the appropriate gap has a direct
bearing on the strength of the weld. In order to get a good quality weld following checks should be exercised:
(i)

It shall be ensured, if necessary, by suitable non-destructive tests that the fusion faces and adjacent
surfaces shall be free from cracks, notches or other irregularities which might be the cause of defects or
would interfere with the deposition of the weld material.

(ii)

Fusion faces and the surrounding surface shall be free from heavy scale, moisture, oil, paint or any other
substance which might affect the quality of the weld or impede the progress of welding. Certain proprietary
protective coatings are specially formulated with the intention that they should not interfere with welding.
The effectiveness of such coatings is however need to be demonstrated and established by means of
specimen welding.

5.2 Assembly for Welding

(i) Parts to be welded shall be assembled such that the joints to be welded are easily accessible and
visible to the operator.
(ii) Jigs and manipulators should be used, where practicable, so that the welding can be carried out in
the most suitable position. Jigs shall maintain the alignment with the minimum restraints so as to
reduce the possibility of locked in-stress.
5.3 Alignment of Butt Joints
The root edges or root face of butt joints shall not be out of alignment by more than 25 percent of the
thickness of the thinner material for material upto and including 12mm thick, or by more than 3mm for thicker
material. For certain applications and welding processes, closer tolerances may be necessary.

83

Connection between Web & Flange Plate

5.4 QUALITY CONTROL

To ensure proper quality control on butt weld a proper quality assurance programme is to be decided
before starting the fabrication in accordance with instructions laid down in IRS:B1-2001. The butt weld so
prepared is examined for its integrity and soundness by ultrasonic or other equally effective non-destructive
testing technique.
6.0 SAMPLE PREPARATION AND TESTING
In order to ascertain the further practical difficulties in making a good quality Double Bevel butt joint a
simple joint was got prepared and tested. The welding procedure specification sheet outlining the welding
procedure and the welding parameters is appended as Annexure-I. The sample joint was tested in the
Metallurgical & Chemical Directorate of RDSO. The dye penetration & micro etching tests performed on
the test piece yielded satisfactory results. The test report is appended on Annexure-II. On the basis of
above, the design of flange butt connection using double bevel butt weld for use in 12.2m span welded plate
girder was approved and provisional drawings No RDSO/B-16003, B-16003/1, B-16003/2, B-16003/3
were issued. Feedback from railways on the use of this connection is awaited.
7.0 CONCLUSION
An economical Welded connection has been designed by taking adequate considerations of the fabrication
difficulties, welding technology available and the existing codal provisions. The double bevel butt weld has
been suggested instead of conventional fillet weld. The design is expected to meet the aspirations of field
engineers and result in economy of cost, material & fabrication time.
8.0 REFERENCES
1.

Indian Railway Standard:1962. Code of Practice for the Design of Steel/Wrought Iron Bridges (Steel Bridge Code),
Research Designs & Standards Organisation, Ministry of Railways, Lucknow, India.

2.

Indian Railway Standard:1986. Bridge Rules, Research Designs & Standards Organisation, Ministry of Railways,
Lucknow, India.

3.

IRS:B1:2001. Indian Railway Standard Specification for Fabrication and Erection of Steel Girder Bridges and Locomotive
Turn Table, Research Designs & Standards Organisation, Ministry of Railways, Lucknow, India.

84

ANNEXURE-I
WELDING PROCEDURE DATA SHEET FOR DOUBLE BEVEL BUTT WELD
NAME AND ADDRESS OF FABRICATOR

Bridge Workshop/JRC

WORK ORDER NO.

A-2210

AUTHORITY

CONSIGNEE

DESCRIPTION OF JOB AND

Fabrication of I-Sec. Welded girder Test piece as

per RDSO Drg. No. for HM loading.

MATERIAL SPECIFICATION

IS : 2062-99 Gr. B. Fully killed fully normalized

DESCRIPTION OF COMPONENT

I-section fillet butt weld

WELDING PROCESS

SAW

WELDING POSITION

Vertical

WPSS NUMBER FOLLOWED

JOINT DESIGN DETAILS AND SKETCH

S.N.
1
2

DESCRIPTION
Electrode for (track weld)

SIZE
4mm

TYPE
Coated

Electrode Wire

4mm

Un- coated

WELDING CURRENT

CLASS
BRAND DRYING METHOD
IRSM-28 A-3 TROARO Heated in electric oven
for 2 hrs. at 250oc
IRSM-39
ESSAB -

TYPE
:
POLARITY :
WELDER QUALIFICATION
:

DC
Reverse
B17, 18 : 7310 Pt. 1

NAME OF WELDER

TICKET NO.

1262

DATE OF WELDING

22-9-04

WELDING PARAMETERS
WELD PASS ELECTRODE CURRENT
NO.
OR/WIRE DIA
MM
Amp.
1
4
575+10
2
4
600+10

PROVISION OF RUN IN AND RUN OFF

ARC/
VOLTAGE
Volt
32
36

CLEARING OF WELD BEAD BEFORE

85

WIRE FEED
SPEED
M/min
2.90
3.10

Yes

TROLLY
SPEED
M/min
0.3
0.25

GAS FLOW
RATE
Ltr./min
-

LAYING NEXT WELD BEAD

Yes

RECTIFICATION OF WELD DEFECTS

(IF ANY) :

Yes

INSPECTION OF WELD
PREPARED BY NAME

DESIGNATION
JE/WELDING
SSE/W
ABE/W

SIGNATURE

MICRO ETCHING TEST : FILLET-BUTT JOINT

Description
Weld size
Throat
Root Penetration

Weld size
18

I1-1
H
17

All dimensions in mm

I2-2
V
17

H
18

12
5

1
13
6

86

I1-10
H
17

I2-20
V
18

12
5

H
18

V
17
5

ANNEXURE-II
M&C DIRECTORATE
Test Report N0. 157/04
In reference to above, two samples of bevel weld were received from B&S Directorate to carry out
metallurgical testing.
1.

Marking and M&C Lab Identification No.

S. No. Description of component
Paint marking
1 Fillet weld 18mm size done by SAW technique FILLET WELD
1
in two passes
2 Bevel butt weld with edge preparation as per FILLET BUTT
WPSS
WELD 2

2.

157/04/2

Sample particulars (as furnished)

Particulars
Description of component
Welding process
Welding position
Date of welding

3.

Lab Identification No.

157/04/1

Sample No. 157/04/01

Test piece I-sec. Fillet weld
SAW
Vertical
22-9-04

Sample No.157/04/2
I-section Fillet but weld
SAW
Vertical
22-9-04

Dye penetrating test

Dye penetrating test was conducted on both sample nos. 157/04/1 and 157/04/2. No appreciable
defects were observed.
4.

Macro etching test

One transverse slice of each sample was cut polished & etched. Following observations are observed:

Sample No. 157/04/1

Description of component
Weld size
Throat
Root penetration

Size of weld

I1-1
H
17.0

18
-

I1-2
V
18.8

H
20.4

12.0
7.8

V
19.0
14.0
7.4

Sample No. 157/04/2

Description of component
Weld size
Throat
Root penetration

Size of weld

I1-1
H
15.7

18
-

12.0
8.2

87

I1-2
V
19.2

H
17.5

V
15.0
12.0
7.0

5.

Remark
i)

Dye penetrating test found satisfactory in both the samples.

ii)

Weld joints found satisfactory in both the samples.

x x x

88

INSPECTION
&
REHABILIATION

TRACK STANDARDS ON GIRDER

BRIDGES A REVIEW
R.K. Goel*
(Presented in IPWE Seminar at Secunderabad, Dec., 2001)

1.0 INTRODUCTION
Track standard for steel girder bridges have been laid down by Railway Board through various Manuals
such as Indian Railways P. Way Manual, Indian Railways LWR Manual and special instructions etc. The
instructions contained in manuals are further supplemented through various circulars issued by zonal railways.
RDSO has also issued drawings for providing Steel Channel Sleepers on girder bridges. In this paper an
attempted has been made to review and compile the existing instructions on the subject for better understanding
and easy implementation.
2.0 TRACK STANDARD IN RESPECT OF RAILS AND SLEEPERS OF GIRDER BRIDGES
2.1 For various reasons it is desirables to avoid fish plated joint in the track. On girder bridges, necessary
instructions for continuing LWR are laid down in LWR manual. Accordingly LWR/CWR is not to be
continued on M.G. on bridges having overall length more than 20m.
2.2 On Broad Gauge the Maximum length of girder bridge over which LWR/CWR track can be permitted,
is as under:
Maximum overall length at girder bridges (in meters) on which LWR/CWR can be continued
Rail free fastening on bridges Rail free fastenings on bridges
para 4.5.7 (i) of LWR manual partly box anchored para 4.5.7 (ii)
of LWR manuals
Temperature Zones Rail section used
Type of sleeper used in
Type of sleeper used in
approaches PRC/ST
approaches PRC/ST
I
60 Kg
30
77
52 kg/90R
45
90
II
60 Kg
11
42
52 kg/90R
27
58
III
60 Kg
11
23
52 kg/90R
27
43
IV
60 Kg
11
43
52 kg/90R
27
43

*Director/Steel Bridges-I, Bridges & Structures Directorate, Ministry of Railways, RDSO, Lucknow (UP)-226011

91

2.2.1The table given above clearly indicates that with the change of rail section from 90R/52 kg to 60 kg, the
maximum permissible overall length of girder bridges over which LWR/CWR can be continued, gets
reduced. This aspect therefore, needs to be kept in view while planning Track Renewal works on
Girder Bridges.
2.2.2The following stipulations are required to be followed while laying LWR/CWR on girder bridges as per
above table:
a)

Rail free fastenings shall be provided throughout the length of the bridge between abutments.

b) The approach track upto 50m on both sides shall be anchored by providing any one of the following:

c)

i)

ST sleepers with elastic fastening.

ii)

PRC sleepers with elastic rail clips with fair T or similar type creep anchors.

The ballast section of approach track upto 50 metre shall be heaped upto the foot of the rail on the
shoulders and kept in well compacted & consolidated condition during the of extreme summer and
winter.

d) Single span of the bridge shall not exceed 30.5 metres and shall have sliding bearings on both ends.
e)

Wherever the girder bridges are provided with rail free fastenings and partly box anchored on each
span, 4 central sleepers shall be box-anchored with fair V or similar type creep anchors and the
bridge timbers laid on girders shall be notched to accommodate individual rivets heads instead of
through notch.

f)

The girders shall be centralized with reference to the location strips on the bearing, before laying
LWR/CWR.

g)

The sliding bearings shall be inspected during the months of March and October each year and
cleared of all foreign material. Lubrication of the bearings shall be done once in two years.

2.2.3Welded rails may be provided from pier to pier with rail free fastenings and with SEJ on each pier. The
rail shall be box anchored on four sleepers at the fixed end of the girder if the girder is supported on
rollers on one side and rockers on other side. In case of girders supported on sliding bearings on both
sides, the central portion of welded rails over each span shall box anchored on four sleepers. On both
side of approaches fully creep anchored fastening shall be used. The length of single span bridge permitted
temperature zone-wise shall be as under:
Temperature Zone

IV
III
II
I

Maximum length of single span girder bridge with SEJ (190mm

gap) at 30m away from both abutments with full creep resistant
fastenings at approaches (td = tm)
75m
87m
110m
146m

92

2.2.4LWR/CWR may also be continued over a girder bridge with the provisions of SEJ at the far end
approach of the bridge using rail-free fastenings over the girder bridge. The length of the girder bridge in
this case, however, will be restricted by the capacity of the SEJ to absorb expansion, contraction and
creep, if any, of the rails. The length of the girder bridges with the above arrangement that can be
permitted in various rail temperature zones for LWR/CWR with SEJs having maximum movement of
120 mm and 190 mm are as follows:
Rail temperature
zone

Max. movement of Max. length of bridge with SEJ

SEJ used (mm)
With ST/PRC
With CST-9
approach sleepers
approach
sleepers

Initial SEJ gap to be

provided at td
With ST/PRC
approach
sleepers

With CST-9
approach
sleepers

IV

190

55 m

45 m

7.0 cm

6.5 cm

III

190

70 m

70 m

7.0 cm

6.5 cm

II

190

110 m

100 m

6.5 cm

6.5 cm

190

160 m

150 m

6.5 cm

6.0 cm

II

120

20 m

15 m

4.0 cm

4.0 cm

120

50 m

50 m

4.0 cm

4.0 cm

Note : SEJ is to be installed 10 m away from the abutments

2.2.5It is evident from the above table that the provisions are same for 90R/52 kg or 60 kg rail section.
However, the LWR/CWR can be permitted on bridges of longer length if the approach sleepers are
changed to ST/PRC instead of CST-9.
2.2.6Bridge on which LWR/CWR is not permitted/provided shall be isolated by a minimum length of 36
meters well anchored track on either side.
2.2.7The instructions for permitting short welded rail panels on girder bridges are given in Indian Railways P.
Way Manual as under:
SWR may be continued over girder bridges with un-ballasted decks up to 13.3m opening if the length
of SWR is symmetrical to the centre ling of bridge and up to 6.1m opening if the length of SWR is unsymmetrical
to the centre line of the bridge. No fish-plated joint should be located on the girder or within six meters from
either abutment. In all such cases rail free fastenings, such as rail screws, dog spikes or rail free clips shall be
used, so that relative movement between rail and sleepers may take place.
2.3 SLEEPERS SPACING OF STEEL GIRDER BRIDGEIn case of bridge timbers the sleepers spacing and depth to be followed is again given in IRPW manual
as under:

93

Depth of sleepers
(exclusive of notching)
not less than

Gauge

Clear distance between

consecutive sleepers not to
exceed

B.G.

510 mm

150 mm

M.G

305 mm

125 mm

N.G.

125 mm

125 mm

Length of sleepers

Outside to outside of girder flanges plus

305 mm. but not less than 2440 mm.
Outside to outside of girder flanges plus
305 mm. but not
less than 1675 mm.
Outside to outside of girder flanges plus
305 mm. but not less than 1525 mm

2.3.1The size of bridge timber also depends on centre to centre distance of girders. For sal wood timbers or
other similar timbers the depth of timber is to be kept as under:
B.G.
Centre to Centre
of Girder
Up to 1,850
Up to 2,000
Up to 2,300

M.G.
Depth of sleeper
excluding notching

Centre to Centre of
girder

150 mm.
180 mm.
240 mm.

Up to 1,250
Up to 1,650
Up to 2,000

Depth of sleeper
excluding notching
150 mm.
200 mm.
240 mm.

2.3.2The minimum width of timber is to be kept as 250mm in BG and 200mm in MG.

2.3.3For standard bridge spans, the size of bridge timbers is to be kept as give below:
Spans in Meters
Up to 12.2
Up to 18.3
Up to 24.4
Up to 30.5

B.G.
Size of sleeper
bxd
250 x 150mm.
250 x 150 mm.
250 x 180 mm.
250 x 240 mm.

M.G.
Size of sleeper
bxd
200 x 150 mm.
200 x 200 mm.
200 x 200 mm.
200 x 240 mm.

2.3.4Clear distance between joint sleepers is to be kept less than 200mm for BG as well MG.
3.0 TRACK STANDARD FOR BRIDGE APPROACHES
3.1 With a view to eliminate possibility of derailment on the approaches of major/important bridges, it is
necessary to strengthen the track in the approaches of major/important bridges. According, to Northern
Railway CE circular No.221 on P.Way, the following measures are to be taken for the purpose:
i)

PRC sleepers to Drg. No. T-4089 to Drg. No. T-4097 should be provided under guard rails of all
bridges.

ii)

In case of CST 9, ST and wooden road on the approaches of the girder bridges, PRC sleepers
should be provided under the guard rails and up to a length of 100 m on the approach with full

94

ballast cushion as laid down. However, in case the approach track is LWR/CWR, the PRC
sleepers should be provided only up to the SEJ and the entre breathing length of LWR/CWR
should be provided with box anchoring.
iii) 100% fittings must be ensured at all time on girder bridges both on running rail and the guard rail
and in approach track up to 100 m /breathing length. Any deficiency noticed must be made good
immediately.
iv) Buffer rails, wherever, existing in approaches of bridges having LWR/CWR track, should be replaced
with switch expansion joints.
v)

Full ballast section of LWR track should be provided up to 100m from the abutment.

vi) For 1 rail length from abutment, ballast retainers should be provided to check slipping of ballast
and to ensure full ballast section.
vii) The sleeper spacing on the approaches of bridges up to 100m length should made 60 cm. However,
where it is not immediately possible, the existing sleeper spacing may continue but 60 cm sleeper
spacing must be provided at the time of renewals/deep screening works.
viii) Rail level of track in approaches of bridges should be maintained strictly as per section and dips in
rail level immediately after abutments must be avoided.
ix) In order to avoid bunching of sleepers over girder (a) guard rails must have notching at spike
locations and (b) tie bars should be fixed on outside of sleepers with hook bolts passing through the
tie bars.
x)

It should be ensured that for a length of about 100m on either side of the girders bridges, the width
of cess available on BG is 90 cm clear of full ballast section.

xi) Wherever curve sharper than 3 degree exist on the approaches of major girder bridges, these
should be provided with check rails.
The above stipulations should be followed in case of all bridges over 6.1m span. In case of arch bridges
or other decked bridges with parapets where the height of the bank is considerable and any derailment can
cause disaster, the aforesaid stipulations must be followed. Such bridges are to be personally identified by the
DSE/Co-ordination of the division.
4.0 CONCLUSION
4.1 It is seen that for continuation of LWR in all zones, the maximum overall length of girder bridge is smaller
for 60 kg rail as compared to track with 52 kg/90R rail. It implies that the continuation of LWR on some
bridges may not be possible if rail renewal is done by 60 kg rail, although the LWR was permitted on the
same bridge with 52 kg rail.
4.2 It can be concluded that the track standards as laid down for girder bridges are very sensitive to track
structure on approaches.
4.3 There is a need to carefully plan the Track Renewals on Girder bridges and review the existing track
structure as per the laid down track standards.
x x x
95

REHABILITATION OF DISTRESSED ROAD OVER

BRIDGE AT FIROZPUR CANTT A CASE STUDY
R.K. Goel* & A.K. Sanyal **
(Presented in IIBE Seminar at Vigyan Bhawan, New Delhi, 1999)
ABSTRACT
Most of the Road Over Bridges over Railways in Northern region are quite old and have outlived service life. Some
of them had been constructed with jack arch type deck in lime mortar and brick masonry. The cost of complete rebuilding
of such bridges is normally high. ROB at Firozpur Cantt is one of such Road Over Bridges wherein cost effective rehabilitation
has been done by making optimum use of residual strength of its various components. The weak components have been
strengthened and complete replacement of material has been done where required. This paper deals with the various
aspects of planning and execution of the work.

1.0 INTRODUCTION
The ROB at FZR is an important link between Firozpur City and cantonment. The ROB was constructed
by Municipal Corporation and Railways some time in 1907 for passing local traffic. During the past 90 years
the traffic had grown considerably and the ROB had become the main link to cross the Railway yard.
The need for strengthening of ROB was felt in early nineties as the bridge had already out lived its useful
life and started showing signs of distress. A series of meetings were held by Railways with civil authorities and
it was decided to rebuild the ROB on cost sharing basis. The bridge portion in between the abutments was
to be rebuilt by railways and the approaches by civil authorities. Estimate were prepared but due to resources
constraints, sanction for complete rebuilding could not be obtained. The general details of ROB and its cross
section before strengthening have been shown in Figure 1 & 2 respectively.
In 1994 a high level delegation of Railway and civil officers jointly inspected the bridge and observed
that the existing bridge girders can be made use of after strengthening but the road surface needed to be dug
along the side of railing to find out depth of road surface above the top of girders and to correctly assess the
condition of girders in order to suggest strengthening measures.
As a result of above, the complete rebuilding of ROB was considered avoidable and plans were drawn
by Railways to strengthening the existing ROB over the track portion.
2.0 STRENGTHENING ARRANGEMENTS
The total depth of over burden over the brick masonry arches was found to be 0.47m which was
considered excessive and increasing the dead load over the girder, without serving any purpose. This was in
addition to the dead load due to 0.36m thick brick masonry in jack arches. Measures as detailed below were
taken to strengthen the bridge.
*Dy. Chief Engineer Bridges, N. Rly,

** Asst. Bridge Engineer, N. Rly.

96

2.1 Reduce the dead weight over the girders by replacing jack arches with RCC deck slabs designed for
double lane class A loading.
2.2 Strengthening of existing girders wherever required and to cross frame them for better stability.
2.3 To replace the existing girders where required and to cross frame them for better stability.
2.4 Jacketing of abutments is lower portion.
Figure 3 shows important details of strengthening measures adopted for ROB.
3.0 SALIENT FEATURES OF THE WORK
3.1 The work of execution involved complete block of road traffic. The various activities involved in the
execution of work have been shown in Figure 4. Accordingly road traffic block of 80 days over the
ROB was arranged with the cooperation of civil authorities. The heavy traffic was diverted through the
level crossings. Situated at the periphery of city and the light traffic was permitted through the only level
crossing existing in city area. Figure 5 shows the arrangements made for the diversion of traffic to
various level crossings.
3.2 Arrangement of shuttering for casting of RCC deck slabs was to be done with uninterrupted flow of rail
traffic under the ROB. Figure 6 shows the details of shuttering arrangements for the work.
3.3 In span 1, signs of corrosion and pitting were observed on girder adjoining to outer most girders.
Additional steel plates of 12 mm thickness were welded on such corroded/damaged areas. Figure 7
shows the strengthening arrangements at such isolated locations.
3.4 Complete steel work was given two coats of epoxy (anti corrosive) paint after strengthening and cross
framing.
4.0 Replacement of longitudinal jack arches by RCC slabs
This work was done by providing the special type of shuttering arrangements in between the longitudinal
beams. The wooden logs of about 100mm dia were used as props to support the shuttering plates on the
bottom flanges of the longitudinal RS joints. This has allowed the uninterrupted flow of rail traffic under the
bridge. The details of shuttering arrangements are shown in Figure 6.
5.0 CONCLUSION
Cost effective rehabilitation of ROB at Firozpur was achieved by making full use of the residual strength
of its girders and other components. The approach can be further utilized in rehabilitation of similar other
ROBs where complete rebuilding is not immediately possible due to constraints on resources.

97

Figure 1: General details of R.O.B. at FZR Cant.

Figure 2: Cross section shwoing old arrangement of jack arch roof of R.O.B. at Fzr.

98

Figure 3: Cross section showing arrangement of deck slabs and details of cross - framing of
R.S. Girder after strengthening of R.O.B. at Firozpur

Figure 4: Chart of scheduled activites - rehabilitation of distressed ROB at FZR

99

Figure 5: Traffic diyersion arrangements : for heavy vehicles & light vehicles due to closure of road over bridge

Figure 6 X-Section showing details of shuttering arrangements for R.C.C. Deck Slab

100

Figure 7: Top plan of R.O.B. - FZR showing strengentning arrangements in span No. 1.

x x x

101

INSPECTION & MAINTENANCE OF BEARINGS

FOR STEEL GIRDER BRIDGES
Ravindra Kumar Goel
(Published in Journal of IPWE, Jan., 2007)

1.

INTRODUCTION

The role of bearings is to transfer the bearing reaction from the superstructure to the substructure,
fulfilling the design requirements concerning forces, displacements and rotations. The bearings should allow
the displacements and rotations as required by the structural analysis with very low resistance during the
whole lifetime. Thus, the bearings should withstand all external forces, thermal actions, air moisture changes
and weather conditions of the region.
It is generally assumed that the life of steel bearing is the same in the life of the bridge. The fact remains
that the life of any structural component is dependent on the type of maintenance inputs given. The life of a
bearing with elastomer or plastic parts is generally expected to be smaller. This technical paper discuses the
various types of bearings generally used in Railway Steel Girders and the maintenance inputs required to be
given for their longer life.
2.0 TYPES OF BEARINGS
2.1 The bearing transfers the forces coming from the superstructure to the substructure. It also allows for
necessary movements in the super- structure which are caused by temperature variations. The following
types of bearings are generally used on Indian Railways
a)

Sliding bearing

b) Rocker and roller bearing

c)

Oil Bath bearing

d) Elastomeric bearing
e)

Phosphor Bronze Bearing

f)

Teflon or P.T.F.E. Bearing

g)

Elastomeric Pot Bearing

3.0 FREQUENCY OF INSPECTION

3.1 All bearings should be generally cleaned and greased once in three years to ensure that all translation
and rotational movements are permitted. Each bridge bearing assembly and the adjacent members in
*Director/Steel Bridges-I, Bridges & Structures Directorate, Ministry of Railways, RDSO, Lucknow (UP)-226011.

102

contact should be inspected at least once a year to ascertain their actual condition. Suitable remedial
measures must be taken immediately if defects are noticed. In case of irreparable damages, even the
replacement of the bearing should be planned. The bearings should also be examined carefully after
unusual occurrences like heavy traffic damage, earthquakes and batterings from debris in high floods.
4.

MAINTENANCE OF BEARINGS

4.1 Sliding Bearings

These are of two types :
(i) Flat Bearing
(ii) Centralised Bearing
In the case of sliding bearings, the girder is lifted a little over 6 mm and the bearing surfaces are thoroughly
cleaned with kerosene oil and a mixture of black oil. Grease and graphite in a working proportion may be
applied on bearing surfaces and the girder lowered. Jacking beams may be inserted wherever necessary. The
position of sliding bearings with respect to location strips should be maintained in such a way that the free
sliding movement in longitudinal direction is freely permitted.
4.2 Roller & Rocker bearings
4.2.1The roller and rockers are lifted from their position (by adequate slinging). The bearings are scraped,
polished with zero grade sand paper and grease graphite in sufficient quantity applied evenly over the
bearings, rockers and rollers before the bearings are lowered. The knuckle pins of both the free and
fixed ends should also be greased. While lifting fixed ends, the space between girders (in case of piers),
or between the girder and the ballast wall (in case of abutment) at the free ends should be jammed with
wedges to prevent longitudinal movement of the girders.
4.2.2The tooth bar of the roller assembly should be placed vertically at mean temperature. It will be better to
indicate in the completion drawings of bridge stress sheet, the maximum expansion, and range of
temperature for which the bearing is designed, so that the slant at the time of greasing can be decided
depending on the temperature obtaining at the time of greasing.
4.3 Oil Bath Bearings
In oil bath bearings, the rollers are submerged in oil. These bearings do not need oiling and greasing as
movement of girder causes automatic lubrication. For such bearings, dust covers should invariably be provided
to keep the oil free from dirt. Level of oil in the oil bath should be checked once in a year and the oil to be recouped, if required. Complete change of oil should be done once in 5 years. Normally, girders are not
required to be lifted where oil bath bearings are provided.
4.4 Phosphor Bronze Bearings
Phosphor bronze bearings need not be greased as they are corrosion resistant and retain their smooth
103

surface. However, they should be thoroughly cleaned of dirt and dust etc. to permit smooth translation and
rotation.
4.5 Elastomeric Bearings
a)

Elastomeric bearings are made of natural or synthetic rubber. They are very stiff in resisting volume
change but are very flexible when subjected to shear or pure uni-axial tension. They are generally
reinforced with steel plates in alternate layers to reduce bulging. When used with a steel or concrete
girder these permit moderate longitudinal movements and small rotations at the ends. Excessive
and uneven bulging should be checked during inspection of these bearings.

b) These bearings require periodic cleaning. They may require replacement in service depending on
the condition and usage.
4.6 Teflon or P.T.F.E. Bearings
The coefficient of friction between steel and PTFE is quite low. The mating surface, which forms the
upper component of the bearing, is stainless steel with good surface finish. These bearings also require periodic
cleaning of the bearing surfaces as in case of sliding bearings. The interface should be protected from dust.
Lubricating the mating surface by silicon grease reduces the coefficient of friction and is desirable.
4.7 Elastomeric Pot Bearing
Such bearings are considered to be maintenance free. However, they should be visually examined
during the inspection for any breakage in dust seal and general cleanliness of the bearing assembly should be
ensured.
5.0 CONCLUSION
Bearings are an important component of the bridge superstructure. Proper maintenance of bearings
ensures that the rotational and translational restraints are not developed in the girders. This prevents
development of avoidable secondary stresses in superstructure and substructure. Therefore, it is necessary
to inspect the bearings at the specified interval and do the required maintenance.

x x x

104

REPAIR AND RECLAMATION OF CAST IRON

PENDULUM BEARING OF JUBILEE BRIDGE BY
EUTECTIC WELDING TECHNIQUE
Ravindra Kumar Goela B. Garaib

R.N. Shuklac

(Published in Journal of IPWE, April, 2009)

ABSTRACT
Eutectic welding has been used successfully to repair the cracked Pendulum Bearing of a Railway Bridge. The
Pendulum Bearing is made up of Cast Iron with high percentage of phosphorus content & about 125 years old. The Cast
Iron Pendulum Bearing was of unique design carrying the total vertical load of 602.2t under loaded condition of the bridge.
Since the Railway Bridges are subjected to heavy dynamic loads, the cracks developed & were propagating and this was
considered a safety hazard. In situ welding was carried out to repair cracks on the pedestal of pendulum bearing of the
Bridge in 1979. The repaired bearing had given a satisfactory service life of more than 20 years. Development of fresh
cracks of similar type at some other locations of the pedestal was reported in 2003 and have again been repaired using the
same technique which was established in the field. The paper describes the methodology, chemical composition of the
material, the electrodes used and the procedure followed for the repair undertaken.
Key words: Eutectic Welding, Pendulum Bearing, Cast Iron A-frame, Eutec-Chamfertrode, Grey Cast Iron, V-Groove,
Crack Arrester hole

INTRODUCTION
Bearing is an important structural component in a bridge transferring the load of the superstructure to the
pier or abutment without restricting the rotational and translational movements. Jubilee Bridge over river
Hooghly was constructed in 1885 on Bandel-Naihati section of Eastern Railway. The bridge is constructed
over river Hooghly with one central span of 360 feet and two side spans of 420 feet. The central span is
having an overhang of 120 feet on each pier having fixed type support conditions. The side spans are simply
supported on the edge of over hanged central span on one side and resting on abutment on the other side.
The general view of the bridge is shown in Figure-1.

Figure 1 General view of Jubilee Bridge

a
c

Director/B&S, RDSO, Ministry of Railways, Manak Nagar, Lucknow, bAsst. Research Officer/ M&C, Manak Nagar, Lucknow,
Sr. Section Engineer/Design/B&S, Manak Nagar, Lucknow

105

The abutments are provided with pendulum bearing of a unique design which permit easy translational
and rotational movements. The details of pendulum bearing are shown in Figure-2. Fifty percent of total
weight of the side spans including the traffic loads (approximately 1204t) is transferred to abutment through
two such bearings, provided on each abutments. The translational and rotational movements are taken care
of by the two suspended link members made of mild steel which are free to swing around a horizontal pin
acting as the pivot. The vertical loads are transferred through the cast-iron A-frame to a cast iron pedestal
resting on the abutment. Cast iron A-frame and pedestal below are designed to take care of the design
compressive stresses coming from the superstructure to the bearing.

(a) Cast Iron Base

(b) Another view of Cast Iron A-Frame

(upper past of the bearing)

(c) Cast Iron A-frame with mild

steel link members, suspended
from top in pin
Figure 2 Details of Pendulum Bearing

106

One of the cast iron pendulum bearings on Naihati end of this bridge had developed cracks on its base
during 1979. Repair welding was carried out by M/s L&T using their eutectic castaguna method with their
own electrodes. Specific procedure of surface preparation and welding was developed for undertaking the
repair welding. Similar type of fresh cracks were again observed during 2003 on the bearing on Naihati end.
The cracks were found extended on both sides of the base of the casting. There was no problem on the
earlier repaired cracks except some non-structural hairline surface cracks of minor nature. Thus the welding
done during 1979 was considered satisfactory. Preservation of old records helped RDSO to find out the
material composition and the technique used. The fresh cracks could be repaired using the same procedure.
The details of the problem and the repair procedure adopted are described in this paper.
FIRST INCIDENCE OF CRACKS (1979)
As already mentioned, Cracks were observed at the base of the bearing at Naihati end upstream side
pedestal in August 1979 during periodical inspection. The cracks had initiated from an oblong hole on the
pedestal and extended both side. Similar crack had developed on the other side as shown in Figure 3.

Figure - 3 Details of cracks initiated from oblong hole in Cast Iron Pedestal of Pendulum Bearing

The A-frame consists of grey cast iron box section casting with about 1.5 inch metal thickness. The grey
cast iron was having high percentage of phosphorous. Actual samples of material were taken from the bearing
and tested at RDSO. The phosphorous content was found to be 1.17%.

107

CRACK REPAIR PROCEDURE ADOPTED IN 1979

The cracks were repaired by M/s L&T using Eutectic Welding with Castaguna Method. The complete
process was developed by RDSO in consultation with M/s. L&T, which supplied special low heat-input type
electrodes. The process and technique developed is termed as CASTAGUNA EUTECTIC process because
of very low heat input. The process is described as below:
(i) Half-inch diameter crack arrester hole were made at a distance of half inch from the crack tip to
avoid further propagation of cracks during welding.
(ii) Cracked area was gouged out upto the root of the crack using Eutec-chamfertrode gouging
electrode to a single V preparation with 60-70 angle and 6-8 mm root gap followed by crack
test using chalk-kerosene method.
(iii) A butter layer was deposited on the walls of the groove by Eutectrode-27 electrode using staggered
welding technique to avoid heat concentration. This was checked by touching the area adjacent to
the weld bead with hands. The temperature should be hand touchable.
(iv) Next layer was deposited using Xuper Xyron-242, 3.15 mm dia electrode. Xuper Xyron-242 is
a low heat input Ni-Fe type electrode manufactured by M/s L&T.
The chemical composition of the weld metal is as under
(i)

- 2% (max)

(ii)

Mn - 1.5%(max)

(iii) Si

- 4%(max)

(iv)

Ni

(v) Cu

- 10%(max)

- 45-60%

(v) Peening was done in the hot condition. Subsequent layers were deposited by same electrode of
4.0 mm dia.
(vi) In order to further strengthen the joints stitch welding was carried out. In this process, grooves
were made on weld metal & parent metal on both sides of the cracks in a direction perpendicular
to the crack to house the strengthening bars. In this case, 20 mm dia, 10" long, round EN-8 steel
bars were used. The bars are called CASTAGUNA bars as shown in Figure 4.
(vii) The bars were placed over the grooves & welded. First layer was welded by using Eutectrode-27
and then filled up by Xuper Xyron-242 taking same precaution as above. Inter bar space was
maintained at 2-3 inch approx. & bars were placed in staggered position (brick laying technique)
to avoid stress concentration in a straight line.
(viii) Two bars had to bend in hot condition to match the profile of the section where they were welded.
(ix) Dye penetrant test was carried out to ensure that the welding done is free from crack. Minor
surface cracks were observed which were considered not harmful.

108

Figure 4 Details of Castaguna bars & repair welding

SECOND INCIDENCE OF CRACK (2003)

Similar type of fresh cracks was again observed during 2003 on the bearing on Naihati end. The cracks,
two in numbers, were found extended on both sides of the base of the casting. The length of both the cracks
were 320 mm & located at a distance of 685 mm from end of the bearing on outer side & 435 mm on track
side.
A team of RDSO officers visited the site for detailed inspection. It was observed that there was no
problem with the cracks repaired earlier in 1979 except some non-structural hairline surface cracks of minor
nature, warranting no attention.
There was a crack on the top casting also. It was intimated by Eastern Railway that the crack on the top
casting was observed in May 1999. Nothing was done on the crack since then, except for regular monitoring
& imposing a speed restriction of 10 Km/h. The crack was in the transverse direction extending over the
entire wall thickness & the two parts almost got separated by about 23 mm. In view of the large gap and the
possibility of its further increasing, it was felt necessary to repair the crack on top casting also.
DEVELOPMENT OF CRACK REPAIR METHODOLOGY
Keeping in view of the urgency & requirement of Eastern Railway, RDSO issued a procedure for repair
welding of the bearing. The procedure adopted in 1979 and recent developments in the field of welding were
kept in view while framing the procedure. Welding process and filler metal was selected on the basis of
material composition of the casting, technical suitability and availability of filler metal as discussed below.
Material composition of the casting
The pendulum bearing was made out of grey cast iron weighing approx. 5 tonnes and phosphorous
content as analyzed in the past was noted to be 1.17%. Therefore, welding procedure with least heat input
was to be adopted for repair.

109

Welding process selection

Out of the various welding processes available, gas welding and manual metal arc welding processes
could be employed for in-situ repair welding. In this case, gas-welding process was not suitable because the
process required massive heat input during preheating and welding. Arc welding with grey iron or mild steel
filler metal was not advisable since preheating to 500C approx. is needed which is difficult to achieve at site.
Cold welding using Nickel base filler metal, which generates least heat, was therefore considered the most
appropriate alternative for repair of the cracks in these castings.
Filler metal selection
Various types of Nickel base filler metal (e.g. pure nickel, Nickel-Iron, Monel etc.) were available.
Keeping in view of the parent metal composition, strength etc. Nickel Iron type filler metal was selected.
The following advantages of Ni-Fe filler metal were kept in view:
a)

It has fairly high ultimate tensile strength (320-430 N/mm2 as against 200N/mm2 for cast iron)

b) Welding can be done with minimum preheat or without preheat

c)

Welding current required is fairly low resulting in minimum heat input

d) Weld metal has reduced co-efficient of expansion.

e)

Suitable for cast iron having higher phosphorus content

f)

Cast iron can also be welded with mild steel using this filler metal

The above Ni-Fe type electrode are classified as class G in IRS M-28 which is the governing specification
for MMAW electrodes in Indian Railways. A no of brands of various manufacturers are approved for this
class. Therefore, no difficulty was expected in getting the electrodes.
CRACK REPAIR ADOPTED FOR PEDESTAL IN 2003

Figure 5 Details of cracks in pedestal

110

Surface and edge preparation

a)

Paint, grease and other surface contaminants from the area adjacent to crack was removed by using
solvent, paint remover, cleaning agents etc.

b)

Surface skin was removed by grinding to ensure freedom from scales etc.

c)

A l0 mm dia. crack arrester hole was drilled 25mm ahead of the crack tip.

d)

In the portion of casting having crack extending throughout the cross section, a V groove was made by
removing metal from either side of the crack. The V groove was having 2-3mm root face.

e)

During the process of preparing V groove, oxy-cutting was avoided. Arc gouging using graphite electrode
( Eutectic-chamfertrode) was done initially. Final finishing of the V groove was done by chipping and
grinding only.

f)

The portion where the crack did not extend throughout the section thickness, metal was removed from
either side of the crack up to a depth adequate enough to ensure removal of crack.

g)

V groove was formed thereafter and dye penetrant test was carried out to ensure freedom from any
cracks.

h)

The V groove so formed was heated to a temperature of 200C for 15 minutes using Oxy-Acetylene
flame to ensure removal of impregnated oil, grease followed by wire brush cleaning.

Welding operation
a)

Buttering layer was given by using 3.20 mm dia Eutectrode-27 and ground.

b)

Then CPCI-026 electrode was used for filling the V groove.

c)

Root run was given by 4 mm dia. electrode and subsequent runs was given by higher dia. electrode

d)

Welding was done in small increments and back step sequence was followed to minimize heat input

e)

Welding was done using stringer bead with lowest possible current.

f)

DC reverse polarity was employed

g)

Proper inter-pass cleaning was made and moderate peening with rounded ball head hammer was done
to relieve stresses. Interpass temperature did not exceed 100C.

h)

After welding each run, the weld was inspected visually for presence of crack, if any. After every 3-4
runs, dye penetrant test was carried out to ensure freedom from cracks.

i)

After the groove was filled, the weld was flush ground and dye penetrant test carried out.

j)

Further strengthening of the base was done by Stitch Welding i.e. welding round shaped En 8 steel
bars over the V weld. The details of the same were as below.
i)

Round EN-8 steel bars, dia 20mm & length 250mm was used.

111

ii)

Grooves were made on the weld metal & parent metal on both side by gouging and grinding to a
depth of 10-12 mm in a direction perpendicular to the crack to make seating for the bars.

iii)

The bars were placed in a brick laying fashion to avoid stress concentration in a straight time. The
distance between the successive bars was 50-75mm

iv) Welding was done by using CPCI-026 electrodes taking precautions to avoid over heating.
v)

After completion of the welding, the weldment was tested by dye penetrant test for presence of
any crack. Superficial cracks appeared and considered not harmful.

CAST IRON-STEEL DISSIMILAR METAL WELDING FOR TOP CASTING

The crack in top portion of the pendulum bearing was in the transverse direction extending over the
entire wall thickness & the two parts almost got separated by about 23 mm (Figure 6).
During discussions at E.Rly. Hqrs. it was jointly agreed that reclamation of the portion of the casting by
filling weld metal would not be appropriate since it would involve huge quantity of weld metal and heat input
which was not desirable. It was therefore, considered appropriate to strengthen the top portion by two steel

Figure 6

112

plates joined together on the top by bolts on the flanges and on the casting by lap welding. The steel plates
should be 32 mm thick and be of weldable grade MS conforming to IS 2062 Gr C. The steel plates therefore
should be having identical contour as the top of the castings. The details of the welding and weld dimensions
are shown in Figure 6. It was case of cast iron-steel dissimilar metal welding which also required special
type of electrodes and procedure. The details of the repair technique adopted are given below:
Welding process & filler electrode
MMAW process was recommended using IRS class-G electrodes conforming to IRS M-28:2002.
Eutectrode-27 electrode as butter layer followed by 22*33N electrode. However, M/s. L&T were permitted
to use their own product which was examined and approved by RDSO. The details of welding electrode
used are given in Table 1.
Table 1 Details of welding electrodes used for welding mild steel plate to Cast Iron A-frame
Application

Electrode

C%

Mn%

Si%

S&P%

Ni%

Al%

45-

Cu% Fe%

(Max.)
Recommended by RDSO
Repair
of
cracks,
Joining of MS to Cast
Iron

Class G to

0-

0-

0-

IRS M 28 02

2.0

1.0

4.0

0.030

60

0-

Bal

2.5

Used by firm during actual welding

Repair
of
cracks,
Buttering layer

Eutectrode-27

0
0.5

0.5
2.5

0
1.5

0.040

01.0

01.0

Bal

Repair
of
cracks,
Subsequent runs

CPCI-026

0-

0.030

40-

0-

Bal

2.0

1.5

4.0

60

1.0

2.5

45

0-

0-

2.0

2.5

0.40

60

1.0

2.5

Joining of MS to Cast
Iron for repair of Top
casting

22*33N

0.030

Bal

Surface preparation
1.

Paint, grease and other contaminants from the surface of the casting and the steel plate where lap
welding is to be accomplished was removed. The adjacent areas where welding was to be carried out
was also thoroughly cleaned

2.

The areas of the casting as well as steel plate where welding was carried out was ground after cleaning

3.

The edges of the MS plate and adjacent area was also ground.

Welding operation
1.

One of the steel plates which were welded on the top portion of the casting was tack welded to
facilitate subsequent lap welding.

113

2.

It was ensured that the steel plate was placed squarely on the casting surface and no gap existed
between the two surfaces. The Steel plate was tack welded to facilitate subsequent welding.

3.

After tack welding, lap welding was carried out on location marked A in the sketch. Welding was done
in small increments using back step sequence. Root run was done by using 4 mm dia. electrode.

4.

Welding was thereafter carried out on the B and C locations vertically. Initially welding commenced
from A side upwards to the top flanges. Welding was carried out in small increments alternatively on
both the sides to avoid any undue stresses and distortion.

5.

For welding of the B and C locations, back step sequence was followed and stringer bead technique
was used. DC reverse polarity was employed during the entire welding operation.

6.

After completion of one run of weld on A, B and C sides, second run deposit was carried out.

7.

Dye penetrant test was carried out after 3 runs and finally after completion of the welding.

8.

The sequence of alternate weld metal deposition was carried out till required weld size of 25mm was
achieved (Sketch enclosed).

9.

After completion of the welding of steel plate on one side, welding of steel plate was carried out on the
other side similarly as explained above.

10. During the course of welding, interpass temperature of 100oC maximum was maintained.
11. Proper interpass cleaning was carried out to avoid entrapment of slag, foreign matter etc.
12. After each run of welding, peening with rounded ball head hammer was carried out to relieve stresses.
SELECTION OF WELDING AGENCY
The work was highly technical in nature and high degree of quality control was to be exercised during
repair welding. M/s. L&T were selected based on their credentials experience of past working on the same
bridge in 1979.
CONCLUSION
Structural members of many old bridges are made of Grey Cast Iron. Eutectic welding with Castaguna
Method of repairing cracks is found quite successful for repairing cracks developed in grey cast iron. The
repairs are found intact even after 25 years of service. It is therefore concluded that satisfactory repairs by
welding may be carried out in grey cast iron structures by using low heat input electrodes. The Eutectic
Welding with Castaguna Method can be satisfactorily adopted for repairing cracks in similar load carrying
members of other bridges.
REFERENCES

B&S Directorates File No.CBS/MJB Research Designs Standards Organisation, Ministry of Railways, Lucknow
(UP)-226011.

M&C Directorates File No.M&C/MIN/III/153, Research Designs Standards Organisation, Ministry of Railways,
Lucknow (UP)-226011.

x x x

114

FATIGUE
ASSESSMENT

REVIEW OF FATIGUE PROVISIONS OF BS:5400 FOR

DESIGN OF RAILWAY BRIDGES
R.K. Gupta* & R.K. Goel**
(Published in Indian Railway Technical Bulletin, R.D.S.O. Lucknow, May, 2005)
ABSTRACT

bl ys[k esa Hkkjrh; jsy ds ekud bLikr iqy lafgrk rFkk ch-,l-&5400] Hkkx&10 esa] jsyos iqyksa ds JkafUr
vfHkdYi QSfVx fMtk+bZu gsrq] fn;s x;s izko/kkuksa ds v/;;u dks izLrqr fd;k x;k gSA Hkkjrh; jsy ds ekud
bLikr iqy lafgrk dh dfe;ksa dk mYys[k djrs gq,s ch-,l-&5400] Hkkx&10 ds izko/kkuksa dks le>us dk iz;kl
fd;k x;k gSA ;s izko/kku ,l-,u- doZ (S-N curve) ij vk/kkfjr gSa rFkk yksfMaxl ds izdkj fLFkfr;ksa] :V th,e-Vh- (GMT), tksM+kas (Connections) ds izdkj vkfn dks /;ku esa j[krs gq, cuk,s x;s gSaA jsyos iqyksa ls lEcfUkr
izko/kkuksa dks bl ys[k esa dzeokj rjhds ls izLrqr fd;k x;k gSA bu izko/kkuksa dks Hkkjrh; jsyos ds iqyksa ds
fM+t+kbZu esa ykxw djus esa vkus okyh dfBukb;ksa dh ppkZ djrs gq,s] vuqlU/kku dh vko';drkvksa dk fu/kkZj.k
fd;k x;k gSA
The paper presents a thorough study of existing provisions of IRS Steel Bridge Code and the BS-5400,
Part-10 in respect of fatigue design of Railway Bridges. The short comings of Indian Railway Standard Steel
Bridge Code have been highlighted and the concepts of fatigue design have been discussed. Provisions of
BS-5400, Part-10 which are based on S-N curve approach are found quite elaborate, covering different
aspects like loadings, loading situations, route GMT, class of connection etc. The design provisions for Railway
Bridges have been discussed in this paper in a systematic manner. The limitations of the study are discussed
and the areas of future research have been identified.
1.0 INTRODUCTION
1.1 The Code of Indian Railway Standards for Steel Bridge recommends method to allow for the effect of
fatigue in design of parts of steel bridges, which are subjected to repeated fluctuations of stress. These
fluctuations may cause fatigue failure of members or connections at lower stresses than those at which
they would fail under static load. Such failures are primarily due to stress concentrations introduced by
the constructional details. Thus, all the details are designed to avoid as far as possible the stress
concentrations likely to result in excessive reductions of the fatigue strength of members or connections.
Care is also taken to avoid a sudden reduction of the section of a member or part of a member,
especially where bending occurs.

Executive Director/Bridges & Structures Directorate/RDSO/Lucknow,

Director/Bridges & Structures Directorate/RDSO/Lucknow.

**

117

1.2 Concept of EUDL (Equivalent Uniformly Distributed Load) is used to determine the maximum bending
moment and maximum shear force for the type of IRS loading. The EUDL for maximum bending
moment and maximum shear force depends upon the span and the dynamic augment increases with
speed. The EUDL for maximum bending moment for different spans, including the dynamic augment for
a maximum speed of 160Kmph are given in tabular form for IRS-MBG loading.
1.3 To allow for the effect of fatigue the allowable working stress is determined from the Appendix 'G' of
IRS Steel Bridge Code for wide range of constructional details. The code covers mild and high tensile
steel fabricated or connected by welding, riveting or bolting. The allowable stress 'P' depends on the
ratio of minimum stress fmin to maximum stress fmax, number of repetitions of stress cycle 'N', the method
of fabrication and the type of connection. In determining the ratio fmin/ fmax gross area is considered. The
code classifies the constructional details into seven classes i.e. Class A to Class G according to type of
steel, type of fabrication and connection. All the details are designed such that the stress induced under
design loads are within the allowable limits.
1.4 The allowable stresses are the principal stress at the point under consideration. Thus in the design of
girder the combined effect of both bending and co-existent shear stress is considered and the bridge
members are generally designed for 10 million cycles of stresses produced under the design load.
2.0 SHORT COMINGS OF IRS APPROACH
2.1 There is no rational basis for adopting counts of 10 million number of cycle to determine the allowable
stress levels.
2.2 Fatigue is a cumulative phenomenon; this is not reflected in the above procedure.
2.3 Stress-ratio procedure does not take into account the effect of all stress ranges experienced by a
member.
2.4 Material S-N curve forms the basis of all fatigue analysis and design which is not the case with the
present procedure.
3.0 PROVISIONS OF BS-5400
In order to overcome the above shortcomings of IRS approach, the provisions of BS-5400 were
scrutinized. Its provisions are based on the concept of cumulative fatigue damage. The code concerns with
the fatigue design methodology for highway and railway bridges and takes into consideration the various
drawbacks of IRS approach. The methods of fatigue assessment provided in the code are based on PalmgrenMiner's damage summation model. Fatigue life assessment is based on the S-N curve approach wherein the
number of cycles to failure is dependent only on stress range and not on maximum stress values. For fatigue
assessment of Railway bridges the provisions and design methodology have been described in a systematic
manner. The important provisions concerned with design of railway bridges are discussed below :

118

3.1 STRESS RANGE CALCULATION, sRmax

3.1.1 Cl. 6.1.1 For Welded Details
sRmax is taken as the greatest algebraic difference between principal stresses not more than 45 degree
apart in any one stress cycle. i.e. If smax is 9.4 kg/mm2 & smin is (-) 2.0 kg/mm2, then sRmax would be 11. 4
kg/mm2.
3.1.2 Cl. 6.1.3 For Non-Welded Details
(i) In case, stress range is entirely on compression side i.e. when there is no stress reversal taking
place, the effect of fatigue loading may be ignored.
(ii) In case of stress reversal
The effective stress range to be used in the fatigue assessment
sRmax = 60% of range from zero stress to maximum compressive stress + 100 % of range from zero
stress to maximum tensile stress
i.e. when smax is (+) 9.4 kg/mm2 & smin is (-) 2 kg/mm2 then sRmax would be equal to 7.64 kg/mm2
(= 0.6 x 9.4 + 2).
It is evident that for the same values of maximum and minimum principal stresses, the stress range taken
for fatigue analysis in welded details is much higher as compared to corresponding values for non-welded
details. It therefore, implies that under the same loading conditions the welded details are more susceptible to
fatigue failure then the non-welded.
3.1.3 Clause 7.3.1 For Non-Consideration of Dead Load Stress in Welded Members
In welded members the dead load stress need not be considered whereas for non-welded members,
the dead load stresses will have to be considered in determining the effective stress range when compression
stresses occur.
3.2 Standard Live Loads Considered
3.2.1Type RU Loading : This loading allows for all combination of vehicles currently running or projected
to run on railways in Europe including United Kingdom and is to be adopted for the design of bridges

80 KN/m

250

0.8m

250

1.6m

250

1.6m

250

1.6m

No limitation

80 KN/m
0.8m
No limitation

Fig.1 Type RU Loading

119

carrying main line railways of 1.4m gauge and above. RU loading consists of four 250 kN concentrated
loads preceded and followed by a uniformly distributed load of 80 kN/m. The arrangement of this
loading is as shown in Fig.1
3.2.2Type RL Loading : Nominal type RL loading consists of a single 200 kN concentrated load coupled
with a uniformly distributed load of 50 kN/m for loaded length up to 100m. For loaded lengths in
excess of 100m the distributed a nominal load shall be 50 kN/m for the first 100m and shall be reduced
to 25 kN/m for lengths in excess of 100m, as shown in Fig.2. Alternatively two concentrated nominal
loads, one of 300 kN and the other of 150 kN, spaced at 2.4m intervals along the track, shall be used
on deck elements where this gives a more severe condition. These two concentrated loads shall be
deemed to include dynamic effects. RL loading is a reduced loading for use only on passenger rapid
transit for use only on passenger rapid transit railway systems on lines where main line locomotives and
rolling stock do not operate.

200 kN
50 kN/m

25 kN/m

25 kN/m

100m

No limitation

No limitation

Fig. 2: Type RL Loading

4.0 CALCULATION OF VARIOUS FACTORS

(i) Design Life Factor, k1 = 1.0 for standard design life of 120 years for standard load spectra for a
permanent railway bridge standard loading as given in Table 2 for RU loading and Table 3 for RL loading. (It
is for assumed traffic volume of 18 - 27 GMT)
Table 2* : Standard load spectra for RU loading
Heavy Traffic
Group
number

Medium Traffic
4

Light Traffic
4

Load 0.75 0.65 0.55 0.45 0.35 0.25 0.15 0.05 0.75 0.65 0.55 0.45 0.35 0.25 0.15 0.05 0.75 0.65 0.55 0.45 0.35 0.25 0.15 0.05
proportion,
kw
Range

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
to to to to to to to to t0 to to to to to to to t0 to to to to to to to
0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

120

Length, Total number of live load cycles (nR x 106) for various loading groups and types
L(m)
2

58

65

71

41

18

14

130 16

17

45 217

23

23

26

40

82

40

65

87

0.2

95 132 45

13

12

46

1.5 71

28

18

27

33

32

81

27

53

89 111

18

20

18

34

43

22

47 116

0.1

42

30 202

3.5 11

19

34

62

0.3 11

18

35

52

0.1

44

30

60

10

10

22

40

0.3

43

17

40

0.1 0.5 0.3 10

69

60

15

1.5 74

0.3
0.6

0.3 40

13

0.1 0.5 0.3

24

51

20

99

0.3

30

54

0.1 0.5 0.3

46

27

30

79

0.5

138

0.2 0.5 0.3

10

174

50

60

1.7

134

0.3

136

* Table No. corresponds to the one given in code.

NOTE 1- L is the base length of the point load influence line (see figure 12). For intermediate values of L, permissible stress ranges may be
derived from the spectra for the two adjacent lengths shown in the table and the value interpolated. nR values apply to one track.
NOTE 2 - The values are based on a traffic volume of 27 x 106 tonnes per annum.

Table 3* : Standard load spectra for RL loading

Group number

Load proportion, Kw

0.55

0.45

0.35

0.25

0.15

0.05

Range

0.5
to
0.6

0.4
to
0.5

0.3
to
0.4

0.2
to
0.3

0.1
to
0.2

0
to
0.1

Length, L(m)
2
3
4
5
7
10
15
20
30
50

Total number of live load cycles (nR x 106) for various loading groups and types
9
1
0
0
0
1
1
1
0
1

120
112
29
6
38
10
13
13
8
13

189
68
75
110
65
56
0
0
6
0

42
10
3
0
0
37
49
0
0
0

0
170
74
2
0
0
30
50
0
0

0
0
180
75
77
77
15
80
265
80

* Table No. corresponds to the one given in code.

Note 1 - L is the base length of the point load influence line. For intermediate value of L, permissible stress ranges may be derived from
the spectra for the two adjacent lengths. Shown in the table and the values interpolated. nR values apply to one track.
Note 2 - The values are based on a traffic volume of 27 x 106 tonnes per annum.

121

If design life is other than 120 years, then the value of k1 is to be taken as lesser of the following -

(a)

k1 =

120
Design life in yrs.

120
Design life in yrs

1
m

1
m 2

When m = inverse slope of sr - N curve appropriate to detailed class to be obtained from Table-8.
Table 8* : sr - N relationships and constant amplitude non-propagation stress range values
0 (N/mm2)

Detail class

K2

3.0

0.16 X 1012

25

3.0

0.25 X 10

12

29

F2

3.0

0.43 X 1012

35

3.0

0.63 X 10

12

40

3.0

1.04 X 1012

47

3.0

1.51 X 10

12

53

3.5

4.23 X 1013

78

4.0

1.01 X 10

15

8.0

2.08 X 1022

100
82

* Table No. corresponds to the one given in code.

Note - Values applicable to non-standard criteria may be obtained from Appendix A.

Multiple cycle factor, k2 is to be applied when there are more than one cycle of stress induced by the
loading event.

= 1

R2
R1

R3
R1

R4
R1

1
m
m

m = inverse slope of sr - N curve (Table-8)

sR1, sR2, sR3 are stress ranges in decreasing order of magnitude.
(ii) RU loading factor, k3 depends upon the detailed class of connection & Base length of the point load
influence line. The relevant value are to be taken from table-4

122

Table 4* : Value of k3 for RU loading of Railway bridges

Heavy traffic
Detail stress

D
E
F
F2
G
W

Medium traffic

Light traffic

D
E
F
F2
G
W

1.01
1.13
1.30
1.46
1.62
1.62
2.03
2.03

1.14
1.28
1.46
1.65
1.65
165
1.83
1.83

1.09
1.23
1.37
1.53
1.71
1.92
2.19
2.46

1.09 1.13
1.22 1.30
1.36 1.46
1.561.71.62
5
1.81
1.952.12.03
8
2.03
2.18
2.03

D
E
F
F2
G
W

1.28
1.46
1.46
1.65
1.83
1.83
1.83
1.83

1.37
1.53
1.71
1.92
2.19
2.46
2.74
3.06

1.60
1.79
1.79
2.05
2.31
2.31
2.56
2.87

1.60
1.80
1.80
2.00
2.34
2.50
2.50
2.50

1.71
1.71
1.95
2.20
2.20
2.20
2.20

Length, L(m) Values of k3

< 3.4
3.4 to 4.0
4.0 to 4.6
4.6 to 7.0
7.0 to 10.0
10.0 to 14.0
14.0 to 28.0
> 28.0

1.00
1.09
1.23
1.37
1.53
1.71
1.92
2.19

1.00
1.09
1.22
1.36
1.56
1.75
1.95
1.95

Note - L is the base length of the point load influence line (see Figure 12)
* Table No. corresponds to the one given in code.

(iii) GMT factor, k4 = 1 for GMT of 18 to 27 which is the GMT assumed in standard RU/RL load spectra.
For GMT other than 18 to 27 the values to be taken from Table - 5. It is evident that the factor reduces
with increase in GMT.
Table 5* : Values of k4 for railway bridges
Annual traffic tonnage on one track (million of tones)
42 to 27
27 to 18
18 to 12
12 to 7
0.89
1.0
1.13
1.27

7 to 5
1.42

<5
1.6

* Table No. corresponds to the one given in code.

(iv) Lane factor, k5 takes into consideration the stress induced at a detail due to two tracks. The relevant
values are to be taken from Table - 6.
Table 6* : Values of k5 for railway bridges
P1

R a tio
k5

P2

0 .5 to
0 .6
1 .42

0 .6 to
75
1 .27

R atio

0 .75
to 0.9
1 .12 5

P1
P1 + P 2

0 .9 to
1 .0
1 .0

* Table No. corresponds to the one given in code.

123

P1

-P
2

R a tio

0 .0 to 0 .7

0 .7 to 1 .0

1 .0

0 .89

P2
P1

Table 7* : Values of k6 for RL loading of railway bridges

Detail Class

D
E
F
F2
G
W

Length, L(m)

Values of k6

<3.0
3.0 to 3.4
3.4 to 4.0
4.0 to 10.0
10.0 to 15.0
15.0 to 20.0
> 20.0

1.23
1.34
1.43
1.57
1.77
1.98
2.08

1.28
1.37
1.49
1.62
1.79
1.99
2.05

1.35
1.45
1.55
1.68
1.90
2.00
2.09

1.65
1.71
1.80
1.91
2.10
2.10
2.10

Table No. corresponds to the one given in code.

(v) RL loading factor, k6 is just as the RU loading factor k3 and the relevant value are to be taken from
Table -7. This loading factor is to be considered in place of RU loading factor. As already discussed,
this type of loading is not matching with any of IRS type loading.
5.0 Discussion on k3 & k6
It is apparent that the values of k3 are ranging from 1.0 to 2.19 and similarly values of k6 are ranging
from 1.23 to 2.08 for detailed class of connection 'D' which is the appropriate classification for the type of
connections adopted on Indian Railway bridges. Both these factors increases with L which is the base length
of the point load Influence line. The standard refers to Fig. 12 to elaborate it and makes a further reference to
Fig. 9. As understood, for the road bridges the length is to be taken from lane stress history for the road
bridges as the base length of the loop containing the largest ordinate (+ve or -ve). For an element of the
highway bridge loaded by more than one lane, L is to be determined from the Influence line for the lane
producing the largest value of stress ordinate.
As the sT, the limiting stress range is obtained by multiplication of these factors k1, k2, k3,.. with s0, the
fatigue strength of connection gets improved with any increase in the values of these multiplying factors.
Now, the problem is how to determine 'L' values for railway bridges. For simply supported spans we may
assume 'L' as equal to the effective span of the bridge as the maximum stress ordinate would be developed on
full base length. This may be workable for design of stringers and simply supported plate girder bridges,
however, for truss girder bridges where axial stress are the prime stresses, choosing appropriate value of 'L'
would be difficult. For IRS loading, these factors might be quite different. However, this much is easily
understood that the limiting stress range or the limiting fatigue strength increases if the base length 'L' is more.

124

For simplicity, 'L' may be taken as the appropriate base length of the influence line diagram for the member
under consideration.
6.0 DESIGN sr- N RELATIONSHIP
6.1 Clause 11.2 of BS : 5400 The number of repetitions to failure 'N' of any one stress range sr is obtained
form the equation
m

Nx

k2

or Log 10 N

m log 10 r

log 10 k 2

Clause k2 & m are as given in Table - 8.

The values given in Table-8 are based on two standard deviations below the mean line with a probability
of failure of 2.3%. The probabilities of failure associated with various numbers of standard deviations below
the mean line are given in Clause A.1 of Appendix-A of BS-5400 part-10.
6.2 Clause 11.2-Computation of no. of cycles to failures The number of repetitions to failure 'N' of
any one stress range sr is obtained from the equation
m

K2

or Log10 N = Log10 k2 - m log10 sr

Values of k2 & m are as given in Table - 8.
6.3 Clause 11.3- Treatment of low stress cycles Number of repetitions of each stress range sr less than
s0 should be reduced in the proportion (sr/s0)2.
Where s0 = Stress range given by equation in Clause 11.2
N = 107
i.e. 10 million cycles
6.3.1It is further given for easy calculations that
n
N

n
N

n r
k2

n r
k 2 02

m 2

n r
10 7 0

n
10 7

when r

m 2

when r

125

7.0 CALCULATION OF LIMITING STRESS RANGE, sT

7.1 First find s0 constant amplitude non-propagating stress range for the constructional detail, as chosen
appropriately on the basis of Table - 17, this value is to be taken from Table -8.
e.g. for Detail Class 'D' for riveted type of connections,
m = 3.0
k2 = 1.52 x 1012
s0 = 53 N/mm2
7.2 The limiting stress range sT can now be calculated in accordance with Clause 9.2.2.1 (d)
sT = k1, k2, k3, k4, k5, s0 for RU loading
= k1, k2, k4, k5, k6, s0 for RL loading
8.0 CHECK FOR DESIGN ADEQUACY:
The design adequacy of the given detail may now be checked as per Clause 9.2.2.2 and Clause 9.2.2.3
8.1 Where sRmax (Maximum Stress Range) does not exceed sT, i.e sRmax < sT, the detail may be considered
to have a fatigue life in excess of the specified design life.
8.2 Where sRmax is more than sT, we have following two options:
(i) The detail may be assessed by more precise procedure given in Clause 9.3.
(iii) The detail may be strengthened so as to reduce sRmax or it should be designed to a higher class.
9.0 MORE PRECISE PROCEDURE OF DAMAGE CALCULATION
9.1 General
This method involves a calculation of Miner's summation and may be used for any details for which
the sr-N relationship is known and for any known load for stress spectra.
9.2 Design Spectrum for Standard Loading
9.2.1 The design spectrum can be determined by the use of either table 2 for RU loading or table 3 of RL
loading (amended where appropriate in accordance with 7.3.3). These tables indicate, for simply
supported members, the equivalent frequency of occurrence of stress ranges of varying magnitudes
resulting from the passage of the individual trains forming various standard traffic types, where the stress
ranges are expressed as proportions of the maximum stress range and the load proportion, kw is the
ratio of actual to standard gross weights of vehicles, bogies or axles in a load spectrum.
Now in order to draw the standard load spectra for a particular type of loading one has to determine
the maximum stress range due to that loading and then the histogram of number of cycles verses stress
ranges of varying magnitudes can be plotted with the help of Table 2 or Table 3 of the code for the
126

appropriate base length (L). This method of drawing standard load spectra is based on the understanding
of the provision given in Clause 7.3. and Appendix-E of the code. The annual traffic tonnage for standard
traffic types and the composition of standard traffic mix are given in Table 15 and Table 16 of the code.
Table 15*
Traffic
type

Train type

Train weight,
tones

No. of trains per

annum

Total annual tonnage,

tonnes x 106

Heavy

1120

4821

5.40

1120

7232

8.10

852

15845

13.50
Total

Medium

600

22500

13.50

1120

2411

2.70

1120

6027

6.75

1794

2257

4.05
Total

Light

27.00

27.00

1794

752

1.35

372

14516

5.40

344

23546

8.10

172

47093

8.10

600

4500

2.70

572

2360

1.35
Total

27.00

Table 16*
Train
Type
1
2
3
4
5
6

Train weight, tonnes

Number of trains
per annum

246
253
380
203
209
231

11545
54032
9786
6453
26986
3463
Total

*Table No. corresponds to the one given in code

127

Total annual tonnage,

tonnes x 106
2.84
13.67
2.74
1.31
5.64
0.80
27.00

9.2.2In the case of loading from more than one track, account should be taken of the possibility of stress
fluctuations arising from the passage of trains on not more than two track, both separately and in
combination. As an approximation, the effects of two track loading may be obtained by dividing sRmax
(see 9.3.2.1) by the coefficient k5 which can be obtained from table 6.
9.2.3Where the approach, passage and departure of a unit uniformly distributed load procedures more than
one cycle of stress, as for instance in multi-span longitudinal or cross members or in continuous deck
slabs, all the cycles should be taken into account. The appropriate standard trains composing the load
spectra should be traversed across the relevant point load influence lines and the resulting stress histories
should be analyzed by the reservoir method, given in appendix B, to derive the respective stress spectra.
These should then be combined with the appropriate annual occurrences obtained from table 15 to 16,
proportioned for the required traffic volume and multiplied by the specified design life to produce the
overall design spectrum. As an approximation, the effect of the additional cycles may be obtained by
dividing either sR max (see 9.3.2.1) or sR max/k5 (see 9.3.2.2) by the coefficient k2 which should be
obtained from 9.2.4.
9.3 Design Spectrum for Non-Standard Loading
9.3.1Where the loading does not comply with 7.3.1 the appropriate train should be traversed across the
relevant point load influence lines and the resulting stress histories should be analyzed by the rainflow
method to derive the respective stress spectra. These should then be combined with the appropriate
total occurrences in the design life of the bridge to compile the overall design spectrum. For nonwelded details the stress range should be modified as given in clause 6.1.3.
9.3.2In assessing an existing structure, a design spectrum may be compiled from strain readings or traffic
records obtained from continuous monitoring.
9.4 Simplification of Spectrum
Where a non-standard loading is used in accordance with 7.1 or the stress ranges are obtained from
strain gauge readings, the design spectrum should be divided into at least 10 equal intervals of stress. All the
stress ranges in any one interval should be treated as the mean range in that interval and low stress ranges
should be treated in accordance with clause 11.3.
9.5 Calculation of Damage
Using the design spectrum the value of Miner's summation n/N should be calculated in accordance
with clause 11 and it should not exceed 1.0 for the fatigue life of the detail to be acceptable.
10.0 CONCLUSION:
The British Code (BS-5400, Part 10) addresses the problem of fatigue design exhaustively and if
provisions are based on rational. It takes into consideration the design parameters such as traffic density,
number of lanes, loading type, design life, multiple cycles of stress produced by loading event and suggest a

128

simplified approach for assessing the fatigue life of the component being designed. Fatigue life of the component
depends on the type of connection adopted which is also taken into consideration by choosing different
values of parameters describing the shape of the characteristic S-N curve of the fabrication material.
11.0 LIMITATIONS AND AREA OF FURTHER RESEARCH
11.1 The simplified procedure given in the code is applicable for standard type of loadings applicable to
European conditions. For IRS type of loading the values of loading parameters and various coefficient
needs to be developed for which further study and research is required.
11.2 For using detailed assessment procedure a standard load spectra applicable to types of trains and traffic
density found on Indian Railways needs to be developed.
11.3 The characteristic S-N curves for the steel and the types of connections adopted for design of bridges
in Indian Railways needs to be developed for proper assessment of the fatigue strength of bridge members.
12.0 ACKNOWLEDGEMENT
The author gratefully acknowledges the support provided by Executive Director (B&S), RDSO for
undertaking this study. The assistance provided by Shri R.K. Sharma, Section Engineer,. and Smt. Suman
Verma, P.A. of B&S Dte is also thankfully acknowledged.
13.0 REFERENCES
13.1 British Standards BS 5400 : Steel, Concrete and Composite Bridges, Part 10. Code of Practice for Fatigue, British
Standards Institution, 1980.
13.2 British Standards BS 5400 : Steel, Concrete and Composition Bridges, Part 2. Specification for Loads, British Standards
Institution, 1978.
13.3 IRS Code of Practice for the Design of Steel/Wrought Iron Bridges (Steel Bridge code) Revised 1962, (with all the
amendments).
13.4 IRS Bridge Rules - Revised 1964 Incorporating Correction Slip 1 to 15, Reprinted 1986.
13.5 Loads to be considered in Railway Bridge Design, UIC Code, Leaflet No. 776-1 R 4th Edition, 1-7-94.
13.6 Statistical distribution of axle loads and stresses in railway bridges Report No. ORE D 128/RP5, RP6 & RP7.

x x x

129

FATIGUE LIFE ASSESSMENT OF STEEL BRIDGES

Ravindra Kumar Goel1 & Dinesh Kumar2
(Presented in IPWE Seminar on Bridge Design, Construction,
Rehabilitation & Maintenance at Chennai, Jan, 2006)
ABSTRACT
The paper presents a study of existing provisions of Indian Railway Standard Steel Bridge Code and the BS-5400,
Part-10 in respect of fatigue design of Railway Bridges. The fatigue design concepts of the two codes have been discussed
and the observed short comings of IRS Steel Bridge Code have been listed. Provisions of, BS-5400 Part-10 which are based
on S-N curve approach are found quite elaborate, covering different loadings, loading situations, route GMT, class of
connection etc. The design provisions for Railway Bridges given in BS-5400 have been briefly discussed in this paper and
the assessment of fatigue life of IRS bridges has been done using the simplified method given in the code. Standard type
RU Loading adopted in BS-5400 is found heavier in comparison of IRS Loading, therefore the fatigue life assessment made
is expected to be on conservative side. The limitations and the assumptions of the study are discussed and the areas of
future research have been identified.

INTRODUCTION
The Code of Indian Railway Standards for Steel Bridge recommends method to allow for the effect of
fatigue in design of parts of steel bridges, which are subjected to repeated fluctuations of stress. These
fluctuations may cause fatigue failure of members or connections at lower stresses than those at which they
would fail under static load. Such failures are primarily due to stress concentrations introduced by the
constructional details. Thus all the details are designed to avoid as far as possible the stress concentrations
likely to result in excessive reductions of the fatigue strength of members or connections. Care is also taken
to avoid a sudden reduction of the section of a member or part of a member, especially where bending
occurs.
To allow for the effect of fatigue the allowable working stress is determined from the Appendix G of
IRS Steel Bridge Code for wide range of constructional details. The code covers mild and high tensile steel
fabricated or connected by welding, riveting or bolting. The allowable stress P depends on the ratio of
minimum stress smin to maximum stress smax, number of repetitions of stress cycle N, the method of fabrication
and the type of connection. In determining the ratio smin/smax gross area is considered. The code classifies the
constructional details into seven classes i.e. Class 'A' to Class 'G' according to type of steel, type of fabrication
and connection. All the details are designed such that the stress induced under design loads are within the
allowable limits.

Director, Bridges & Structures Directorate, Research Designs & Standards Organization, Indian Railways, Manak Nagar, Lucknow226001

Section Engineer/Design, Research Designs & Standards Organization, Indian Railways, Manak Nagar, Lucknow-226001

130

The allowable stresses are the principal stress at the point under consideration. Thus in the design of
girder the combined effect of both bending and co-existent shear stress is considered and the bridge members
are generally designed for 10 million cycles of stresses produced under the design load. Concept of EUDL
(Equivalent Uniformly Distributed Load) is used to determine the maximum bending moment and maximum
shear force for the type of IRS loading. The EUDL for maximum bending moment and maximum shear force
depends upon the span and the dynamic augment increases with speed. IRS Bridge Rules gives tables for
determination of EUDL for maximum bending moments and shear forces alongwith the co-efficient of dynamic
augment for maximum speed of 160 kmph.
SHORT COMINGS OF IRS APPROACH
1.

There is no rational basis for adopting counts of 10 million number of cycle to determine the allowable
stress levels.

2.

Fatigue is a cumulative phenomenon; this is not reflected in the above procedure.

3.

Stress-ratio procedure does not take into account the effect of all stress ranges experienced by a
member.

4.

Material S-N curve forms the basis of all fatigue analysis and design which is not the case with the
present procedure.

5.

Standard train load is transformed into equivalent uniform load. Thus, actual variation of stresses in a
member due to passage of train is not accounted.

DESIGN PROVISIONS OF BS-5400

BS-5400 part 10 is a comprehensive code which is based on the concept of cumulative fatigue damage.
The code concerns with the fatigue design methodology for highway and railway bridges and take into
consideration the various drawbacks of IRS approach. The methods of fatigue assessment provided in the
code are based on Palmgren-Miners damage summation model. Fatigue life assessment is based on the SN curve approach wherein the number of cycles to failure is dependent only on stress range and not on
maximum stress values. For fatigue assessment of Railway bridges the methodology has been given in the
code. The methodology for determination of stress range has been described for welded and non-welded
details and a simplified procedure has been given for determining the limiting value of the maximum range of
stress for the specified design life for two different types of standard loadings. The code specifies different
factor k1, k2, k3, k4 & k5 for design parameters such as design life, multiple cycle of stress loading, type of
standard loading, annual GMT and multiple lane loading respectively. The code gives specific methodology
and tables to calculate the factors for different design parameters.
Calculation of Limiting Stress Range, sT
The constant amplitude non-propagating stress range, s0 for the constructional detail is chosen
appropriately on the basis of Table 17 & Table 8 of the code. The limiting stress range sT can now be
calculated for RU loading as under:
131

sT = k1. k2. k3. k4. k5. s0 ..(1)

Check for Design Adequacy
The design adequacy of the given detail is now checked as per Clause 9.2.2.2 and Clause 9.2.2.3 of
the code. Where sRmax (Maximum Stress Range) does not exceed sT, i.e sRmax < sT, the detail may be
considered to have a fatigue life in excess of the specified design life. Where sRmax is more than sT, we have
following two options:
1.

The detail may be assessed by more precise procedure given in Clause 9.3 of the code.

2.

The detail may be strengthened so as to reduce sRmax or it should be designed to a higher class.

STANDARD TYPE LOADINGS

RU Loading
This loading allows for all combination of vehicles currently running or projected to run on railways in
Europe including United Kingdom and is to be adopted for the design of bridges carrying main line railways
of 1.4m gauge and above. RU loading consists of four 250 kN concentrated loads preceded and followed by
a uniformly distributed load of 80 kN/m. The arrangement of this loading is as shown in Fig.1
RL Loading
Nominal type RL loading consists of a single 200 kN concentrated load coupled with a uniformly
distributed load of 50 kN/m for loaded length up to 100m. For loaded lengths in excess of 100m the
distributed a nominal load shall be 50 kN/m for the first 100m and shall be reduced to 25 kN/m for lengths in
excess of 100m, as shown in Fig.2. Alternatively two concentrated nominal loads, one of 300 kN and the
other of 150 kN, spaced at 2.4m intervals along the track, shall be used on deck elements where this gives a
more severe condition. These two concentrated loads shall be deemed to include dynamic effects. RL
loading is a reduced loading for use only on passenger rapid transit for use only on passenger rapid transit
railway systems on lines where main line locomotives and rolling stock do not operate. This loading is
apparently of an exclusive type and not matching with any of IRS type of loading.

80 KN/m
0.8

250

250

1.6

250

1.6

250

1.6

No limitation

80 KN/m
0.8
No limitation

Fig.1 RU Loading

132

200 kN
50 kN/m

25 kN/m

25 kN/m

100m
No limitation

No limitation

Fig.2

RL Loading

COMPARISON OF RU LOADING WITH IRS-MBG LOADING

A comparison of EUDL values of Bending Moment as per RU loading was made with corresponding
values for IRS-MBG loading (Table-1). It is found that the EUDL values as per RU loading are on higher
side as compared to IRS-MBG loading.
Table 1. EUDL for RU Loading and IRS MBG Loading
SNo.

Span
in m

(1)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
22
24
26
28
30

EUDL for maximum B.M. in kN (including dynamic

augment)
RU Loading
IRS MBG loading
(3)
(4)
1000
980
10.07
980
1089
980
1351
1162
1541
1391
1652
1523
1817
1608
1951
1688
2066
1750
2171
1817
2268
2078
2359
2195
2447
2317
2531
2415
2613
2497
2694
2566
2773
2623
2851
2699
2927
2772
3003
2864
3153
3049
3301
3231
3447
3403
3592
3573
3736
3741

133

%age difference
(5)
2.04
2.76
11.12
16.25
10.79
8.50
13.00
15.56
18.04
19.51
9.14
7.47
5.60
4.81
4.64
4.98
5.72
5.63
5.58
4.87
3.42
2.17
1.28
0.52
-0.15

FATIGUE LIFE ASSESSMENT OF STANDARD GIRDERS

The S-N curve based approach of BS-5400 has been tried on exiting design of standard girders. The
stresses calculated during the analysis and the cross sections provided in the existing design have been used
to workout the fatigue life of the members provided. Following assumptions have been made during this
study 1.

Maximum stress range calculated as the difference of dead load stress and the maximum stress likely to
come in the member with DL &live load with impact and occasional load.

2.

The stresses due to load combination with occasional load have been taken into consideration to find
out the maximum stress range. This combination rarely occurs in practice, therefore the analysis can be
considered on conservative side.

3.

Material properties are assumed to be as per Table-8 of BS-5400 and s0 value has been taken to be
corresponding to detailed classification D of this table.

4.

The fatigue life of standard spans has been assessed by calculating the design life factor k1. This factor
has been worked out as sRmax/(s0 x k3x k4) and fatigue life calculations have been done by inversion,
using the equations given in Clause 9.2.3 of BS-5400 Pt.10 by taking fatigue life as minimum of following:
Fatigue Life

120
m
k1 ..(2)

or
Fatigue Life

120
m 2 (3)
k1

Where m = 3.0 taken from Table-8 for detailed Class D.

5.

It is pre-assumed that the loading event produces only one cycle of stress and hence k2=1.0

6.

Value of RU loading factor k3 has been taken from Table-4 of the code considering the case of medium
traffic loading, corresponding to the base length (L) of the influence line diagram for detailed class D.

7.

Value of GMT factor, k4 is assumed as 1.0 for GMT of 18 to 27 million tonnes. For GMT other than
this, the value of k4 have been taken from Table-5 of the code.

8.

Single lane loading has been assumed, hence value of lane factor, k5 is taken as 1.0.

Based on above assumptions, the design calculations for assessment of fatigue life of standard plate
girder spans are given in Table-2 for standard annual GMT of 27 to 18. Table-3 gives the fatigue life
assessment for different annual GMTs on the route.
The fatigue life assessment has also been done for open web girder bridges and the fatigue life of
different members is found different based on the base length (L) of the influence line diagram for that member.

134

Table 2. Fatigue Life Assessment of Standard Plate Girders for standard GMT
Std. Span
(1)
12.2 MBG
18.3 MBG
24.4 MBG
12.2 HM
24.4 HM

Dead load Dead load +I.L.

Drawing No.
(kN)
+ L.L (kN)
(2)
B-1528
B-1529
B-1555
B-16002
B-16001

(3)
158.2
243.6
504.2
158.2
57.62

(4)
2531.18
3108.6
3939.5
0.2955.4
5076.5

A
(mm2)

Stress (N/mm2)
pmin

pmax

(6)

(7)
98.0
131
98.5
106.5
102.6

(5)
43100
46900
77000
43800
94100

6.1
10.2
12.6
5.7
11.6

Stress
Range Rmax
(N/mm2)
(8)
91.9
120.8
85.9
100.7
90.9

L'
(m)

135

RU loading
factor, k3

k1

(10)
1.92
2.19
2.19
1.92
2.19

(11)
0.903
1.040
0.740
0.990
0.783

12.2 HM
Drg. No.B-16002
Fatigue
k1
life, years
(9)
(10)
1.112
70
0.990
123
0.876
178
0.779
253
0.697
354
0.619
506

24.4
Drg.No

(9)
13.1
19.4
25.6
13.1
25.6

Table 3. Fatigue Life Assessment for Different GMTs

Route GMT
(1)
42-27
27-18
18-12
12-7
7-5
<5

GMT
factor,
k4
(2)
0.89
1.0
1.13
1.27
1.42
1.6

12.2 MBG
Drg. No.B-1528
Fatigue
k1
life, years
(3)
(4)
1.014
111
0.903
163
0.799
235
0.711
334
0.636
466
0.564
667

18.3 MBG
Drg. No.B-1529
Fatigue
k1
life, years
(5)
(6)
1.168
55
1.040
99
0.920
154
0.819
218
0.732
305
0.650
437

24.4 MBG
Drg. No.B-1555
Fatigue
k1
life, years
(7)
(8)
0.831
209
0.740
296
0.655
427
0.582
606
0.521
847
0.462
1212

k1
(11)
0.880
0.783
0.693
0.617
0.551
0.489

Table 4(a) Assessment of Fatigue life of various members of standard OWG 30.5m span for standard GMT
(Welded Through Type ,MBG loading, BA-11462)
0(N/mm2) =53

Member

Dead load
(t)

m=3
Dead load
+I.L. + L.L +
Occ. Load
(t)

A
(cm2)

Stress ( t/cm )

DL

DL+IL+LL
+Occ load

Stress Range
Rmax (t/cm2)

L'
(m)

RU
loading
factor, k3

Rmax
N/mm2

k
R

136

L0-L1

10.35

123.74

114.54

0.090

1.080

0.990

31.926

2.46

98.996

L1-L2

10.44

110.63

114.54

0.091

0.966

0.875

31.926

2.46

87.472

L2-L3

18.65

166.28

132.95

0.140

1.251

1.110

31.926

2.46

111.042

L0-U1

-19.67

-133.74

146.48

-0.134

-0.913

-0.779

31.926

2.46

-77.874

U1-L2

11.38

87.66

87.71

0.130

0.999

0.870

25.54

2.19

86.968

L2-U3

-3.9
-3.9

22.77
-50.4

70.67
77.34

0.050
-0.050

0.322
-0.652

0.713

19.16

2.19

60.460

U1-U2

-16.57

-106.19

124.52

-0.133

-0.853

-0.720

31.926

2.46

-71.972

U2-U3
U1-L1& U3L3
Stringer
X-girder

-16.77

-108.67

124.52

-0.135

-0.873

-0.738

31.926

2.46

-73.803

4.22

63.32

80.6

0.052

0.786

0.733

10.642

1.92

73.325

4.13
2.86

140.29
50.59

174.2
234

0.030
0.064

1.036
1.161

1.006
1.097

5.321
5.28

1.53
1.53

100.600
109.740

1
1

Table 4(b) Assessment of Fatigue life of various members of standard OWG 45.7m span for standard GMT
(Welded Through Type ,MBG loading, BA-11482)

0 (N/mm2) =53

m=3
Dead load
Dead load +I.L. + L.L +
A
Member
Occ. Load
(t)
(cm2)
(t)

Stress ( t/cm2)
DL

Stress
Range
DL+IL+LL+ Rmax (t/cm2)
Occ load
1.335
1.171

L'
(m)

RU loading
factor, k3

Rmax
N/mm2

k
Rmax/

47.24

2.46

117.062

0.

137

L0-L1

21.7

176.55

132.28

0.164

L1-L2
L2-L3

21.9
48.57

159.16
304.52

132.28
238.9

0.166
0.203

1.203
1.275

1.038
1.071

47.24
47.24

2.46
2.46

103.765
107.137

0.
0.

L3-L4
L0-U1
U1-L2

48.71
-38.5
27.02

312.76
-205.84
150.57

238.9
223.2
117.78

0.204
-0.172
0.229

1.309
-0.922
1.278

1.105
-0.750
1.049

47.24
47.24
38.38

2.46
2.46
2.46

110.527
-74.973
104.899

0.
0.
0.

L2-U3

-16.6
5.8
5.8

-104.99
-32.6
67.09

91.28
91.28
84.13

-0.182
0.064
0.069

-1.150
-0.357
0.797

-0.968

32.48

2.46

-96.834

0.

1.012

26.67

2.19

98.160

0.

U1-U2

-36.91

-184.17

149.17

-0.247

-1.235

-0.987

47.24

2.46

-98.720

0.

U2-U3

-37.46

-190.53

149.17

-0.251

-1.277

-1.026

47.24

2.46

-102.614

0.

U3-U4
U1-L1&
U3-L3
Stringer
X-girder

-50.81

-257.79

201.06

-0.253

-1.282

-1.029

47.24

2.46

-102.944

0.

5.09

67.61

78.6

0.065

0.860

0.795

11.81

1.92

79.542

0.

3.00
2.68

148.31
55.86

183.1
253.5

0.040
0.016

1.111
1.112

1.071
1.096

5.905
5.28

1.53
1.53

107.090
109.640

1.
1.

U3-L4

Table 4(c) Assessment of Fatigue life of various members of standard OWG 61.0 m span for standard GMT
(Welded Through Type, MBG loading, BA-11582)
0 (N/mm2) = 53
Dead
Member
load (t)

m=3
Dead load +I.L.
+ L.L + Occ.
Load (t)

A
(cm2)

Stress
Stress ( t/cm2)
Range
DL+IL+LL
2
DL
+Occ load Rmax (t/cm )

L'
(m)

RU loading
factor, k3

Rmax
N/mm2

138

L0-L1

42.2

271.5

229.92

0.184

1.181

0.997

63

2.46

99.730

L1-L2

42.2

302.08

229.92

0.184

1.314

1.130

63

2.46

113.031

L2-L3
L3-L4

90.5
90.5

530.88
542.65

402.08
402.08

0.225
0.225

1.320
1.350

1.095
1.125

63
63

2.46
2.46

109.525
112.453

L0-U1

-63.8

-311.57

299.6

-0.213

-1.040

-0.827

63

2.46

-82.700

U1-L2
L2-U3
U3-L4

46
-27.4
9
9

223.45
-153.21
91.74
-42.7

151.09
127.6
115.19
127.6

0.304
-0.215
0.078
0.071

1.479
-1.201
0.796
-0.335

1.174
-0.986

54
45

2.46
2.46

117.447
-98.597

0.997

36

2.46

147.680

U1-U2

-72.5

-318.83

243.6

-0.298

-1.309

-1.011

63

2.46

-101.121

U2-U3

-72.5

-316.25

243.6

-0.298

-1.298

-1.001

63

2.46

-100.062

U3-U4
U1-L1 &
U3-L3
Stringer
X-girder

-96.5

-415.54

324

-0.298

-1.283

-0.985

63

2.46

-98.469

-13.8

90.18

98.69

-0.140

0.914

1.054

15.75

2.19

105.360

8.15
6.08

166.74
68.63

180
288

0.057
0.104

1.183
1.218

1.126
1.114

7.875
5.5

1.71
1.53

112.565
111.400

Table 4 (a), (b) & (c) gives the typical calculations for assessment of fatigue life of various members of 30.5m,
45.7m & 61.0m span standard open web girders conforming to MBG loading.
CONCLUSION
1.0 From maximum bending moment consideration the equivalent uniformly distributed loads as per RU
loading for medium traffic are found heavier as compared to the corresponding EUDL for IRS MBG
loading. The provision of BS-5400 Part 10 for RU loading can be conveniently adopted for assessment
of fatigue life of plate girders designed for IRS loading. The assessed fatigue life of existing IRS plate
girders is found between 99 to 296 years for standard GMT of 18 27 million tonnes. Fatigue life of
girders is affected by the average annual GMT of the route quite significantly. For very low GMT routes
(less than 5 GMT) the fatigue life is found increased by about four times the fatigue life on standard
GMT routes. Therefore, for economical design of bridges, the route GMT should be considered an
important design parameter. The renewal of existing plate girders, designed as per IRS Steel Bridge
Code, can also be planned economically keeping in view the assessed fatigue life with respect to the
GMT of the route.
2.0 The fatigue life analysis of different members of open web girder bridges reveals that the stringers and
cross girders are having fatigue life much shorter to other members. Thus there is a need to rationalize
the design of these members.
LIMITATIONS AND AREAS OF FURTHER RESEARCH
1.

Above study has been conducted for assessing the fatigue life of plate girder bridges based on stress
range concept of BS-5400 with D class of connections. Actual fatigue life of connection is not assessed
which may be somewhat on lower side. Stress concentration factors for various type of connections
need to be established for better assessment.

2.

The material S-N curve given in BS-5400 has been used in the study assuming that the material
characteristics of the Indian Steel (IS-2062) are the same. The results need to be verified by testing on
steel actually used in fabrication of girders.

ACKNOWLEDGEMENT
The authors gratefully acknowledges the support provided by Executive Director (B&S), RDSO for
undertaking this study. The assistance provided by Shri R. K. Sharma, Section Engineer and Smt. Suman
Verma, P.A. of B&S Dte., RDSO/ Lucknow is also thankfully acknowledged.
REFERENCES
1.

British Standard. (1980). BS:5400 Steel, Concrete and Composite Bridges, Part 10, Code of Practice for Fatigue.
British Standards Institution, London.

2.

British Standard (1980). BS-5400 : Steel, Concrete and Composition Bridges, Part 2, Specification for Loads. British
Standards Institution, London.

139

3.

Indian Railway Standard. (1962). Code of Practice for the Design of Steel/Wrought Iron Bridges (Steel Bridge
Code). Research Designs & Standards Organisation, Ministry of Railways, Lucknow (U.P.).

4.

Indian Railway Standard. (1986). Bridge Rules. Research Designs & Standards Organisation, Ministry of Railways,
Lucknow (U.P.).

5.

Office of Research & Experiment. (1976). Statistical distribution of axle loads and stresses in railway bridges.
Report No. ORE D-128/RP5, International Union of Railways, Paris.

6.

Ravi, G. and Ranganathan, R. (1991). Critical study of fatigue design of bridges as per BS:5400-Part 10. International
symposium on Fatigue and Fracture in Steel and Concrete Structures, Structural Engineering Research Centre,
Madras, India.

7.

Gupta, R.K. and Goel, R.K. (2005), Review of Fatigue Provision of BS:5400 for Design of Railway Bridges No. 312313, Vol Lx1, Feb-May, 2005, Indian Railway Technical Bulletin, Research Design Standard Organisation, Lucknow

8.

UIC Code. (1994). Loads to be considered in Railway Bridge Design. Leaflet No. 776-1, International Union of
Railways, Paris.

APPENDIX II. NOTATION

The following symbols are used in this paper:
A

area of cross section;

k1-k5

coefficients in simplified assessment procedure as adopted in BS-5400 Part-10;

no. of repetitions to failure of stress sr;

so

constant amplitude non-propagating stress range (sr at N=107);

spmax

maximum value of principal stress;

spmin

minimum value of principal stress;

sr

range of stress (stress range);

sRmax

maximum stress range (spmax - spmin);

sT

limiting stress range under standard loading.

x x x

140

ADVANCEMENTS IN CODAL PROVISIONS FOR

FATIGUE DESIGN OF RAILWAY STEEL BRIDGES
Ravindra Kr. Goel*
(Presented in National Conference on Advances in Bridge Engineering organized by
IIT/Roorkee at Roorkee, March, 2005)
ABSTRACT
Fatigue is considered an important aspect while designing the dynamically loaded structures. This is particularly
important in steel railway bridges where dynamic loads are quite high in comparison to static loads. Normally fatigue is
critical at the joints and more so with welded connections. Residual stresses and stress concentrations at the joints control
the stress levels to be adopted in the design. Therefore, the connection details are the most important part of any fatigue
assessment for railway steel bridges. This aspect has drawn the attention of structural engineers and there has been
increased awareness to design the steel bridges with respect to fatigue effects at connections. A lot of work has been done
in this regard by European Community under the Euro Code Programme. The behavior of connections under dynamic
loading has also been studied in details and the codal provisions have been drafted in a scientific and rational manner. The
paper attempts to highlight the important advancement made in this area for design of railway steel bridge.

1.0 INTRODUCTION
Railway steel bridges are subjected to high dynamic loads and the check of members for failure against
the fatigue is critical and it affects the overall cost of the bridge. Fatigue check is carried out for loads that are
normally less than the maximum design loads. Normally fatigue is critical at the joints and more so with
welded connections. Residual stresses and stress concentrations at the joints control the stress levels to be
adopted in the design. Therefore, the connection details are the most important part of any fatigue assessment
for railway steel bridges. This aspect has drawn the attention of structural engineers and there has been
increased awareness to design the steel bridges with respect to fatigue effects at connections. A lot of work
has been done in this regard by European Community under the Euro Code Programme. The behavior of
connection under fluctuating stresses has been studied in detail along with other aspects and the provisions
have been drafted for design of steel structures. The paper attempts to discuss the important provisions
related with fatigue design of steel railway bridges.
2.0 IRS PROVISIONS VIS A VIS EURO CODE
2.1 As per IRS code, as on date, fatigue strength is based on the ratio of minimum stress and the maximum
stress the member undergoes. When welded connections are involved, residual stresses also play an
important role. The actual analysis of stresses becomes very difficult as location and type of weld also
become important. It is for this reason that all international codes have changed over their design approach
to what is known as stress range concept from the stress ratio concept.
*

Director/Steel Bridges-I, Bridges & Structures Directorate, Ministry of Railways, RDSO, Lucknow (UP)-226011.

141

2.2 Presently, IRS Code stipulates that all members of the truss and floor system should be designed for 10
million cycles of fatigue loading. However, for a given traffic volume it is well known that the stringers
are subjected to greater number of cycles of loading than the chord members. Hence IRS provisions
lacking in rationality are conservative for the design of chord members. However, as per Euro code,
this difference in the cycles of loading is accounted for by using different influence lengths for different
members. The Euro Code is written based on limit sate concept as against the working stress concept
of IRS Steel Bridge.
2.3 The design approach in Euro Code is based on S-N curve. The quantitative relationship relating fatigue
failure to stress range and no. of cycles is used for the fatigue assessment of a category of the structural
detail. The fatigue resistances given in the code apply only to structures with normal atmospheric action
and sufficient corrosion protection.
3.0 EURO CODE PROVISIONS FOR DETERMINING FATIGUE STRENGTH
3.1 The provisions governing the fatigue strength of steel structures are given in Euro Code EN 1993-1-9:
2002. The material partial safety factors are chosen depending upon the consequences of failure of
structure.
3.2 For the choice of partial factors in the fatigue assessment two reliability concepts are considered:
(i) Damage tolerant concept and
(ii) Safe life concept
3.2.1The damage tolerant concept may be applied when load redistribution between components of structural
elements can occur in case of fatigue damage.
3.2.2The safe life concept should be applied in case where local formation of cracks in one component could
rapidly lead to failure of the structural elements or structure.
3.2.3Structures that are designed to EN 1993-1-10 and which are subjected to regular maintenance are
deemed to be damage tolerant.
3.2.4A pre-requisite of the damage tolerant concept is the existence of a mandatory inspection program that
provides that if a crack is detected, the part will be repaired or replaced. The provisions for inspection
program are specific to member nations and are to be specified by the concerned nation as National
Annexure. The National Annexure may also give the choice of the reliability concept and numerical
values for partial safety factors for fatigue strength gmf
Table 3.1 Partial safety factors for fatigue strength gmf
Safety concept
Damage tolerant concept
Safe life concept
*

Consequence of failure
Low consequence
High consequence
1.00
1.15
1.15
1.35

Table No. corresponds to EN 1993-1-9 : 2002

142

4.0 GENERAL DESIGN METHOD

4.1 The fatigue assessment, in general stress range verification, shall be carried out according to EN 1992,
EN 1993, and EN 1994.
4.2 The safety verification is done by ensuring that the following condition is satisfied :
Design stress range < Allowable stress range (Both for same no. of cycles)
i.e. Ff 2 E

c / Mf

Where:
Parameters for Design stress range
gFf

is the partial safety factor for fatigue loading

is the damage equivalent factor for fatigue which is a function of following:

a)

Span/loaded lengths/Type of loading (l1)

b)

Traffic volume (l2)

c)

Design life (l3)

d)

No. of tracks supported by the member (l4)

(i.e.) l = (l1) x (l2) x (l3) x (l4) < 1.4

j2

is the damage equivalent impact factor

DsE is the stress due to design train loads placed in the most un-favourable position for the element
under consideration.
Parameters for allowable stress Range
Dsc is the reference value of the fatigue strength to be taken from S-N curve corresponding to the
detail category of the joint/connection as specified in Table 8 of EN 1993-1-9:2002.
gMf is the partial safety factor for fatigue strength
5.0

FATIGUE STRESS SPECTRA

5.1 Simplified procedure of fatigue stress range assessment has been described below (1) For the simplified fatigue loading as specified, the following procedure may be used to determine
the design stress range spectrum.
(2) The maximum stress sP, max and the minimum stress sP, min should be determined for a detail by
evaluating influence areas.
(3) The reference stress range Dsp for determining the damage effects of the stress range spectrum
should be obtained from:
Dsp = | sp, max sp, min |

143

(4) The damage effects of the stress range spectrum may be represented by the damage equivalent
stress range related to 2 106 cycles:
E2 = 2p
where, l is the damage equivalence factor
j2 is the damage equivalent impact factor.
(5) For railway bridges the value of f2 should be obtained from the code, EN 1991-2.
5.2 As an alternative to the procedure given above, fatigue stress spectra may be obtained from the evaluation
of stress histories from fatigue load vehicles as specified in EN 1991-2 and EN 1993-1-9.
f2, f3)
6.0 DYNAMIC FACTOR f (f
6.1 The dynamic factor takes into account, the dynamic magnification of stresses and vibration effects in the
structure. However, it does not take account of resonance effects for which dynamic analysis needs to
be carried out to calculate impact values.
6.2 For load model 71 given in figure-1, the dynamic factor f is taken as f2 or f3 depending upon the
quality of track maintenance.

Figure 1 Load Model 71 and characteristic values for vertical loads

(a)

For carefully maintained track:

1.44
2

0 .2

0.82

with: 1.00 < f2 < 1.67

(b)

For track with standard maintenance:

2.16
3

0 .2

0.73

144

with 1.00 < f3 < 2.0

Where, Lf is determinant length, i.e. the length associated with f and to be appropriately taken as per
the Euro Code.
7.0 FATIGUE ASSESSMENT
7.1 The fatigue assessment shall be made as follows:
Ff E2

c
Mf

7.2 The damage equivalence factor l for railway bridges up to 100m span should be determined as follows:
l = l1 l2 l3 l4
where l1 factor for different types of girder that takes into account the damage effect of traffic and
depends on the length of the influence line or area, loading type factor
l2 factor that takes into account the traffic volume, GMT factor
l3

factor that takes into account the design life of the bridge, design life factor

l4 factor to be applied when the structural element is loaded by more than one track, lane factor
7.2.1

l1 may be obtained from Table 9.3 and Table 9.4 of EN 1993-2:2004. These tables are reproduced
below for illustration.
Table 9.3* : l1 for standard rail traffic

L
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
6.0
7.0
8.0
9.0
10.0
12.5
15.0
17.5
20.0
25.0

Type1
1.38
1.38
1.38
1.37
1.17
1.05
0.94
0.81
0.77
0.86
0.97
0.98
0.92
0.88
0.85
0.79
0.75
0.74
0.74
0.76

Type2
1.27
1.27
1.27
1.26
1.23
1.19
1.02
0.82
0.73
0.69
0.63
0.57
0.55
0.56
0.56
0.55
0.56
0.56
0.55
0.59

Type3
1.31
1.31
1.31
1.31
1.28
1.25
1.12
0.96
0.88
0.80
0.79
0.79
0.77
0.74
0.72
0.73
0.73
0.73
0.68
0.56

Type4
1.50
1.50
1.50
1.49
1.46
1.42
1.16
1.00
0.91
0.86
0.79
0.82
0.83
0.83
0.83
0.78
0.77
0.68
0.66
0.58

Type5
1.62
1.62
1.62
1.35
1.29
1.25
1.12
1.15
1.14
1.16
1.12
0.96
0.85
0.77
0.66
0.52
0.51
0.53
0.55
0.59

145

Type6
1.65
1.65
1.65
1.46
1.39
1.35
1.18
1.08
1.07
1.07
1.07
1.04
1.01
0.96
0.91
0.89
0.81
0.72
0.72
0.69

Type
1.69
1.69
1.69
1.53
1.44
1.4
1.17
1.05
1.04
1.05
1.07
1.07
1.06
1.05
1.04
1.00
0.91
0.80
0.70
0.68

Type8
1.65
1.65
1.65
1.64
1.60
1.56
1.40
1.20
0.97
0.93
0.78
0.79
0.73
0.68
0.65
0.60
0.59
0.58
0.58
0.60

EC Mix
1.60
1.60
1.60
1.46
1.38
1.35
1.17
1.07
1.02
1.03
1.03
0.97
0.92
0.88
0.85
0.82
0.76
0.70
0.67
0.66

Table 9.4*: l1 for express multiple units and underground and for rail traffic with 25 t axles
Express multiple units and
underground

Rail traffic with 25 t axles

Type9

Type10

Type5

Type6

Type11

Type12

25t Mix

0.5

0.97

1.00

1.62

1.65

1.83

1.79

1.65

1.0

0.97

1.00

1.62

1.65

1.83

1.79

1.65

1.5

0.97

1.00

1.62

1.65

1.83

1.79

1.65

2.0

0.97

0.99

1.35

1.46

1.81

1.78

1.64

2.5

0.95

0.97

1.29

1.39

1.56

1.74

1.55

3.0

0.85

0.94

1.25

1.35

1.51

1.69

1.51

3.5

0.76

0.85

1.12

1.18

1.21

1.57

1.31

4.0

0.65

0.71

1.15

1.08

1.04

1.30

1.16

4.5

0.59

0.65

1.14

1.07

1.05

1.05

1.08

5.0

0.55

0.62

1.16

1.07

1.07

1.00

1.07

6.0

0.58

0.63

1.12

1.07

1.10

0.87

1.04

7.0

0.58

0.60

0.96

1.04

1.15

0.77

1.02

8.0

0.56

0.60

0.85

1.01

1.14

0.71

0.99

9.0

0.56

0.55

0.77

0.96

1.13

0.67

0.96

10.0

0.56

0.51

0.66

0.91

1.12

0.64

0.93

* Table No. corresponds to the one given in code.

NOTE 1 The underlined values in Table 9.3 and Table 9.4 give the envelope of all the train types shown
in Annex-F of EN 1991-2 (such as freight trains, passenger trains and high-speed trains in any combination)
and cover the worst effect for a given span. The values given for mixed traffic correspond to the combination
of train types given in Annex F of EN 1991-2.
NOTE 2 l1 for express multiple unit, underground and rail traffic with 25 t axles are given in Table 9.4.
NOTE 3 For lines with train type combinations other than those taken into consideration (specialized
lines for example), the National Annex may directly specify values of l1 as demonstrated in Table 9.3 and
Table 9.4.
7.2.2In determining l1 the critical length of the influence line or area should be taken as follows, unless a more
accurate determination is made:
a)

for moments:

for a simply supported span, the span length, Li;

for continuous spans, in midspan sections, see Figure-2, the span length Li of the span under
consideration;

146

b)

c)

for continuous spans in support sections, see Figure-2, the mean of the two spans Li and Lj adjacent
to that support;

for cross-girders supporting rail bearers (or stringers), the sum of the two adjacent spans of the rail
bearers (or stringers) immediately adjacent to the cross-girder;

for a deck plate supported by only cross-girders or cross-ribs (no longitudinal members) and for
those supporting cross-members, the length of the influence line for deflection (ignoring any part
indicating upward deflection), taking due account of the stiffness of the rails in load distribution. For
cross members spaced not more than 750 mm apart, this may be taken as 2 cross-memberspacing + 3 m.

for shear for a simply-supported span (and, as an approximation, for a continuous span):
l

for the support section (Fig.2) the span under consideration Li ;

for the midspan section (Fig.2) 0.4 the span under consideration Li ;

in other cases:

Fig. 2 : Location of midspan or support section

l

the same as for moments.

7.2.3l2, l3 & l4 should be obtained from Table 9.5, 9.6 & Table 9.7 of EN 1993-2:2004 respectively.
These tables are reproduced below for illustration.
Table* 9.5: l2
Traffic per year [106 t/track]
2

5
0.72

10
0.83

15
0.90

20
0.96

25
1.00

30
1.04

35
1.07

40
1.10

50
1.15

Table* 9.6: l3
Design life [years]
1

50
0.87

60
0.90

70
0.93

147

80
0.96

90
0.98

100
1.00

120
1.04

Table* 9.7: l4
1.00
0.90
0.80
0.70
0.60
0.50
1/1+2
1.00
0.91
0.84
0.77
0.72
0.71
4
1 is the stress range at the section to be checked due to load model 71on one track;
1+2 is the stress range at the same section due to load model 71 according to EN 1991-2 on any two tracks.
* Table No. corresponds to the one given in code.

7.2.4. The value of l should not exceed lmax which is equal to 1.4.
7.3

Combination of damage from local and global stress ranges

Where the stress verification in a member is due to the combined effects of flexure of the bridge (global
effects) and flexure of the internal elements (local effects), the combined effects DsE2 should be as follows:
E2 = loc loc loc + glo glo glo

in which the suffix "loc" refers to local effects and "glo" refers to global effects.
8.0 FATIGUE STRENGTH
8.1 For the fatigue strength assessment of bridges, the appropriate fatigue detail categories are to be chosen
and accordingly the values of Dsc and DsR are to be read from the fatigue strength curves shown as
Fig. 7.1 & 7.2 in EN 1993-1-9. These curves have been reproduced below as Fig. 3 & Fig.4.

Fig. 3 - Fatigue strength curves for direct stress ranges

148

Fig. 4 Fatigue strength curves for shear stress ranges

8.2 Fatigue strength categories Dsc for nominal stresses are given in different tables (Table 8.1 to 8.10) for
different types of connection details. Table 8.1 of the code for un-welded details has been shown below
for illustration.
8.3 The detail categories are designated by number which represents in N/mm2, the reference value Dsc or
Dtc for the fatigue strength at 2 million cycles.
8.3 The reference value Dsc or Dtc for the fatigue strength at 2 million cycles have been calculated with
a 75% confidence level of 95% probability of survival for log N, taking into account the standard
deviation and the sample size and residual stress effect.
9.0 FATIGUE STERNGTH MODIFICATION
9.1 For non welded details or stress relieved welded details, the mean stress influence on the fatigue strength
may be determined by adding the tensile portion of the stress range and 60% of the modulus compressive
portion of the stress range.
9.2 The size effect has to be taken into account as given in table 8.1 to 8.10. The fatigue strength then
calculated as Dsc,red = k5 Dsc

149

Table 8.1* : Non-welded details

Table No. corresponds to the one given in code.

150

10.0 CONCLUSION
10.1 It can be seen that the provisions of Euro code are based on a comprehensive and detailed study which
takes into account the type of connection, location of joint, type of weld, type of loading, dynamic
effects, no. of lanes, traffic density, design life, differential loaded lengths etc.
10.2 The effect of various factors has been studied in detail and tables have been derived to easily work out
the effect of different factors. However, these tables are specific to the type of loadings and the standard
train composition being used in Euro Nations.
10.3 The member nations have been delegated the right to determine the values of critical parameters within
the broad framework of the Euro Codes. This has enabled the member nations to use the codes with
variations as per the practical conditions prevailing in their country.
10.4 The provisions of Euro Code are based on Limit state of Design and there are frequent cross references
with other codes of practices therefore, it may not be practicable to adopt the provisions selectively.
Detailed study is required to understand the philosophy behind the design practices suggested. Study is
also required to prepare a national annexure for Indian conditions for the fatigue parameters prescribed
in Euro Codes.
11. REFERENCES
11.1 EN 1993-1-8 : 2002, Euro Code - 3, "Design of Steel Structures, Part 1.8 - Design of joints"
11.2 EN 1993-1-9 : 2002, Euro Code - 3, "Design of Steel Structures, Part 1.9 : Fatigue strength of steel structures".
11.3 EN 1993-2 : 2004, Euro Code - 3, "Design of Steel Structures, Part 2 : Steel Bridges".

x x x

151

REVIEW OF FATIGUE PROVISIONS OF EURO CODES

FOR DESIGN OF RAILWAY STEEL BRIDGES
Piyush Agrawal* & Ravindra Kr. Goel**
(Published in Golden Jubilee Special issue of Indian Railway Technical Bulletin,
R.D.S.O. Lucknow, March 2007)

jsyos iqyksa ij Hkkjh VSfQd gksus ds dkj.k fofHkUu vo;oksa dh QSfVd ds dkj.k Qsy gksus dh lEHkkouk
vfkd jgrh gSA QSfVd ds dkj.k lcls vf/kd uqdlku fofHkUu tksM+ksa ij gksrk gSA vr% QSfVd gsrq fd;s tkus
okyh fMt+kbZu tkp esa tksM+ksa ds izdkj dk fo'ks"k egRo gSA bl fn'kk esa ;wjksfi;u la?k }kjk vk/kqfudre fjlpZ
dk lekos'k djrs gq,s ;wjks dksM fodflr fd;s x;s gSaA bl 'kks/k i= esa bu dksM+ksa dk foLr`r v/;;u djds
jsyos LVhy xMZjksa ds fMt+kbZu ls lEcfU/kr izkfo/kkuksa dks le>kus dk iz;kl fd;k x;k gSA
Railway steel bridges are subjected to high dynamic loads and the check of members for failure against
the fatigue is critical and it affects the overall cost of the bridge. Fatigue check is carried out for loads that are
normally less than the maximum design loads. Normally fatigue is critical at the joints and more so with
welded connections. Residual stresses and stress concentrations at the joints control the stress levels to be
adopted in the design. Therefore, the connection details are the most important part of any fatigue assessment
for railway steel bridges. This aspect has drawn the attention of structural engineers and there has been
increased awareness to design the steel bridges with respect to fatigue effects at connections. A lot of work
has been done in this regard by European Community under the Euro Code Programme. The behaviour of
connection under fluctuating stresses has been studied in detail alongwith other aspects and the provisions
have been drafted for design of steel structures. The paper attempts to discuss the important provisions
related with fatigue design of steel railway bridges.
1.0 INTRODUCTION
The existing fatigue provisions of IRS Steel Bridge code are based on stress-ratio concept which is
obsolete. Internationally the fatigue is considered as a cumulative phenomenon in which stress-range is taken
at the primary design parameter alongwith others such as type of connection, number of cycles of fluctuating
stress, GMT and type of traffic etc. Presently, IRS Code stipulates that all members of the truss and floor
system should be designed for 10 million cycles of fatigue loading. However, for a given traffic volume it is
well known that the stringers are subjected to greater number of cycles of loading than the chord members.
Hence IRS provisions lacking in rationality are conservative for the design of chord members. However, as
per Euro code, this difference in the cycles of loading is accounted for by using different influence lengths for

*Executive Director/ Bridges & Structures Directorate, Ministry of Railways, RDSO, Lucknow (UP)-226011.
**Director/Steel Bridges-I, Bridges & Structures Directorate, Ministry of Railways, RDSO, Lucknow (UP)-226011.

152

different members. The Euro Code follows limit sate concept as against the working stress concept of IRS
Code and it's design provisions are based on S-N curve approach. The quantitative relationship, relating
fatigue failure to stress range is used for the fatigue assessment of a category of the structural detail.
2.0 RELIABILITY CONCEPTS IN EURO CODE
2.1 The provisions governing the fatigue strength of steel structures are given in Euro Code EN 1993-1-9:
2002. The material partial safety factors are chosen depending upon the consequences of failure of
structure. For the choice of partial factors in the fatigue assessment two reliability concepts are considered:
(i) Damage tolerant concept and
(ii) Safe life concept
2.2 The damage tolerant concept may be applied when load redistribution between components of structural
elements can occur in case of fatigue damage.
2.3 The safe life concept should be applied in case where local formation of cracks in one component could
rapidly lead to failure of the structural elements or structure.
Table -1 partial safety factors for fatigue strength gMf
Consequence of failure
Low consequence
High consequence
1.00
1.15
1.15
1.35

Safety concept
Damage tolerant concept
Safe life concept

3.0 FATIGUE DESIGN CRITERIA

3.1 The safety verification is done by ensuring that the following condition is satisfied :
Design stress range < Allowable stress range
(Both for same no. of cycles)
i.e. Ff 2 E

c / Mf

Where:
Parameters for Design stress range
gFf is the partial safety factor for fatigue loading
l

is the damage equivalent factor for fatigue which is a function of following:

a) Span/loaded lengths/Type of loading (l1)
b) Traffic volume (l2)
c) Design life (l3)

153

d) No. of tracks supported by the member (l4)

(i.e.) l = (l1) x (l2) x (l3) x (l4) < lmax
f2 is the damage equivalent impact factor
DsE is the stress due to design train loads placed in the most un-favourable position for the element
under consideration.
Parameters for allowable stress Range
Dsc is the reference value of the fatigue strength to be taken from S-N curve corresponding to
the detail category of the joint/connection as specified in Table 8 of EN 1993-19:2002.
gMf is the partial safety factor for fatigue strength
4.0

FATIGUE STRESS SPECTRA

4.1 Simplified procedure of fatigue stress range assessment has been described below (i) For the simplified fatigue loading as specified, the following procedure may be used to determine
the design stress range spectrum.
(ii) The maximum stress sP,max and the minimum stress sP,min should be determined for a detail by
evaluating influence areas.
(iii) The reference stress range Dsp for determining the damage effects of the stress range spectrum
should be obtained from:
p = | p,max - p,min |
(iv) The damage effects of the stress range spectrum may be represented by the damage equivalent
stress range related to 2 106 cycles:
E2 = ..p
where, l is the damage equivalence factor
F is the damage equivalent impact factor.
f2, f3)
5.0 DYNAMIC FACTOR f (f
5.1 The dynamic factor takes into account, the dynamic magnification of stresses and vibration effects in the
structure. However, it does not take account of resonance effects for which dynamic analysis needs to
be carried out to calculate impact values. The values of this factor would depend upon the traffic load
model. The most commonly used load model as per Euro Codes is the load model 71 which has been
defined in Figure.1. The Euro Code provides for different dynamic impact factors for carefully maintained
track and the track with standard.

154

250

80 KN/m
0.8

250

1.6

250

1.6

250

1.6

80 KN/m
0.8

No limitation

No limitation
Figure 1 Load Model 71 and characteristic values for vertical loads

The expressions for these impact factors are as given below:

(a)

For carefully maintained track:

1 .4 4
2

0 .2

0 .8 2

with: 1.00 < f2 < 1.67

(b) For track with standard maintenance:

2.16
0.73
L
0.2

with 1.00 < f3 < 2.0

Where, Lf is determinant length, i.e. the length associated with f and to be appropriately taken as per
Annexure-I.
6.0 ASSESSMENT OF FATIGUE STRENGTH
6.1 The fatigue assessment shall be made as follows:
Ff E2

c
Mf

6.2 The damage equivalence factor l for railway bridges up to 100m span should be determined as follows:
l = l1 l2 l3 l4
where l1 is the factor for different types of girder that takes into account the damage effect of traffic and
depends on the length of the influence line or area and known as loading type factor. It's value is taken from
Table - 2 corresponding to loaded span length, L.

155

Table-2 l1 values for different type of rail traffic

L

Standard Rail Traffic

Rail Traffic with 25 t

axles

1.0

1.69

1.79

2.0

1.64

1.78

3.0

1.56

1.69

4.0

1.20

1.30

5.0

1.16

1.16

6.0

1.12

1.12

7.0

1.07

1.15

8.0

1.06

1.14

9.0

1.05

1.13

10.0

1.04

1.12

12.5

1.00

1.07

15.0

0.91

0.99

17.5

0.80

0.85

20.0

0.74

0.76

25.0

0.76

0.69

30.0

0.77

0.68

35.0

0.76

0.68

40.0

0.73

0.68

45.0

0.70

0.69

50.0

0.70

0.70

6.3 The above values give the envelope of all the train types as taken in Euro Code EN-1991-2 such as
freight trains, passenger trains and high speed trains in any combination and cover the worst effect for a
given span. The values are specific, therefore to the type of traffic running over the member nations and
need to be separately worked out for any other type of traffic.
6.4 The other parameters l2, l3 & l4 may be taken from Table 3, 4 & 5 respectively. The standard design
life of 100 years has been considered in Euro Code and accordingly the design life factor l2 is taken as
1.0 in Table-3 for a design life of 100 years. Similarly standard GMT of 25 has been taken for which
GMT factor l3 is taken as 1.0 in Table 4. l4 is applicable when the bridge is to be designed for more
than one lane of traffic. This factor takes into account the stress reversal effect under a lane when the
other lanes are loaded. This factor as given in Table-5 is applicable for traffic load model-71 shown in
Figure-1.
156

Table - 3: Values of GMT Factor, l2

Traffic per year [106 t/track]

10

15

20

25

30

35

40

50

0.72

0.83

0.90

0.96

1.00

1.04

1.07

1.10

1.15

Table - 4 : Values of Design Life Factors, l3

Design life [years]

50

60

70

80

90

100

120

0.87

0.90

0.93

0.96

0.98

1.00

1.04

Table - 5: Values of Lane Effect Factor, l4

1/1+2

1.00

0.90

0.80

0.70

0.60

0.50

1.00

0.91

0.84

0.77

0.72

0.71

1
1+2

is the stress range at the section to be checked due to

load model -71on one track;
is the stress range at the same section due to load model 71 according to EN 1991-2 on any two
tracks.

6.5 The values of ' l ' should not exceed lmax which is taken as 1.4 for European conditions. This value
again needs to be re-established for actual conditions prevailing in other nations.
7.0 COMBINATION OF DAMAGES FROM LOCAL AND GLOBAL EFFECTS
Where the stress verification in a member is due to the combined effects of flexure of the bridge (global
effects) and flexure of the internal elements (local effects), the combined effects Ds E2 should be as
follows:
E2 = loc loc loc + glo glo glo
in which the suffix "loc" refers to local effects and "glo" refers to global effects.
8.0 FATIGUE STRENGTH
8.1 For the fatigue strength assessment of bridges, the appropriate fatigue detail categories are to be chosen
and accordingly the values of Dsc and Dtc are to be read from the fatigue strength curves shown in
Fig. 2 & Fig.3.

157

Figure - 2: Fatigue strength curves for direct stress ranges

Figure - 3: Fatigue strength curves for shear stress ranges

8.2 Fatigue strength categories Dsc for nominal stresses are given in tabular form for different types of
connection detail. A typical table for un-welded details has been shown in Table-6 for illustration.
158

Table 6 : Detailed Category of Connections - Non-welded details

159

8.3 The detail categories are designated by number which represents in N/mm2, the reference value Dsc or
Dtc for the fatigue strength at 2 million cycles.
8.4 The reference value Dsc or Dtc for the fatigue strength at 2 million cycles have been calculated with a
75% confidence level of 95% probability of survival for log N, taking into account the standard deviation
and the sample size and residual stress effect.
9.0 FATIGUE STERNGTH MODIFICATION
9.1 For non welded details or stress relieved welded details, the mean stress influence on the fatigue strength
may be determined by adding the tensile portion of the stress range and 60% of the modulus compressive
portion of the stress range.
10.0 SIZE EFFECT
10.1 The size effect is a new concept introduced in Euro Code. It suggests that the fatigue strength of a
particular type of connection reduces as the thickness of the connecting members increases beyond
some specific value. The size effect has to be taken into account as per the modification factor given in
tables classifying detail category of connection and the fatigue strength is then modified as underDsc, red = k5 Dsc
11.0 CONCLUSION
11.1 It is observed that the provisions of Euro code are based on a comprehensive and detailed study which
takes into account the type of connection, location of joint, type of weld, type of loading, dynamic
effects, no. of lanes, traffic density, design life, differential loaded lengths etc.
11.2 The effect of various factors has been studied in detail and tables have been derived to easily work out
the effect of different factors. However, these tables are specific to the type of loadings and the standard
train composition being used in Euro Nations.
11.3 The member nations have been delegated the right to determine the values of critical parameters within
the broad framework of the Euro Codes. This has enabled the member nations to use the codes with
variations as per the practical conditions prevailing in their country.
11.4 The provisions of Euro Code are based on Limit state method of Design and there are frequent crossreferences with other codes of practices developed in Euro Programme. Therefore, it may not be
practicable to adopt the provisions selectively. Detailed study is required to understand the philosophy
behind the design practices suggested.
11.5 As the design parameters given in Euro codes are applicable to the conditions prevailing in European
Nations, it is essential to develop a national annexure for Indian conditions for all such fatigue parameters,
prescribed in Euro Codes.
12.0 REFERENCES
12.1 EN 1993-1-8 : 2002, Euro Code - 3, "Design of Steel Structures, Part 1.8 - Design of joints"
12.2 EN 1993-1-9 : 2002, Euro Code - 3, "Design of Steel Structures, Part 1.9 : Fatigue strength of steel structures".
12.3 EN 1993-2 : 2004, Euro Code - 3, "Design of Steel Structures, Part 2 : Steel Bridges".

160

Annexure 1
DETERMINANT LENGTH, L
(1) The determinant lengths Lf to be used are given in the Table below. These values are recommended
values and are to be specified by the member Nations based upon the conditions prevailing in their
country.
(2) Where no value of Lf is specified the determinant length should be taken as the length of the influence
line for deflection of the element being considered or alternative values should be specified.
(3) If the resultant stress in a structural member depends on several effects, each of which relates to a
separate structural behaviour, then each effect should be calculated using the appropriate determinant
length.
Table 1: Determinant Lengths, Lf
Case

Structural element

Determinant length L

Steel deck plate: closed deck with ballast bed (orthotropic deck plate) (for local and transverse stresses)
Deck with cross girders and continuous longitudinal ribs:
1.1

Deck Plate (for both directions)

3 times cross girder spacing

1.2

Continuous longitudinal ribs (including small cantilevers 3 times cross girder spacing
up to 0.50 m)3

1.3

Cross girders

Twice the length of the cross

girder

1.4

End cross girders

3.6mb

Deck plate with cross girders only:

2.1

Deck plate (for both directions)

Twice cross girder spacing +3m

2.2

Cross girders

Twice cross girder spacing + 3m

2.3

End cross girders

3.6mb

Steel grillage: open deck without ballast bed b (for local and transverse stress)
3.1

Rail bearers:
- simply support

3 times cross girder spacing

Cross girder spacing + 3m

3.2

Cantilever of rail bearer a

3.6m

3.3

Cross girders (as part of cross girder/continuous rail bearer Twice the length of the cross
grillage)
girder

3.4

End cross girders

- as an element of a continuous grillage

3.6m b

cantilevers greater than 0.50m need a special study

It is recommended to apply 3

x x x
161

EMERGING LOAD MODEL (MBG) FOR FATIGUE

ASSESSMENT OF RAILWAY STEEL BRIDGES
R.K. GOEL*
(Published in Golden Jubilee Special issue of Indian Railway Technical Bulletin,
R.D.S.O. Lucknow, March 2007)
SYNOPSIS

bl ys[k esa iqyksa ds tksM+ksa dh QSfVax LVsSaFk dk vkdyu djus ds fy;s vko';d yksM+ ekWMy ,e ch th
fodflr fd;k x;k gSA bl yksM ekWMy ds iz;ksx }kjk ikexszu ekbuj fu;e dk O;ogkfjd o rdZ laxr mi;ksx
djds iqy ds tksM+ksa dh QSfVax LVSaFk dk lgh fu/kkZj.k fd;k tk ldrk gSA ,p ,e yksfMax o vU; yksfMaXl ds
fy;s] blh vk/kkj ij QSfVax yksM ekMy dk fodkl djus dh vko';drk gSA
This paper presents the considerations involved in development of a traffic load model for fatigue
assessment of a connection in railway steel bridges. The relevance of load model in assessing fatigue life has
been illustrated and a load model has been developed for IRS MBG loading with traffic classification as
suburban, light, medium and heavy type. This load model can be adopted for assessment of fatigue strength
of bridge details in a rational manner.
1.0 INTRODUCTION
Railway bridges are subjected to heavy fluctuating dynamic loads. These fluctuations cause fatigue
failure of members or connections at lower stresses than those at which would otherwise fail under static load.
IRS provisions, which are based on stress ratio concept does not taken into account the phenomenon of
fatigue adequately. World over such effects are taken into account following the Palmgren Miner principle of
cumulative damage based on stress range concept. Indian Railways are presently having MBG standards of
loading for design of new bridges. This loading specifies the axle loads and spacings in respect of the loco
only and therefore the effects of axle loads and spacings in respect of wagons are not specifically considered.
For conducting detailed analysis of fatigue strength of bridge components, the effects of individual axles are
required to be accounted for. In this paper an attempt has been made to develop load model based on the
type of trains, axle loads, type of locos and the train lengths etc. The proposed load Model also allows the
accountal of effects of low, medium and high traffic density in assessing the fatigue damage of bridge members.
With this load model it is possible to assess the bridge details for fatigue in a rational manner using the
Palmgren Minor rule based on stress range concept.

*Director/B&S/SB-I, Research Designs & Standards Organisation, Lucknow

162

2.0 PALMGREN-MINER LINEAR DAMAGE RULE

Palmgren has proposed a damage model on the basis of constant energy absorption per cycle. The
energy absorption per cycle leads to linear summation of damage. Miner has subsequently, represented this
concept in mathematical form. Palmgren-Miner rule states that the fatigue damage contribution by each individual
load spectrum at a given stress level is proportional to the number of cycles applied at a stress interval, ni,
divided by the total number of cycles to failure at the same stress level, Ni. It is obvious that each ratio can be
equal to unity if the fatigue cycles at the same stress level would continue until failure occurs. The total
damage, in terms of partial cycle ratios or damage, can be written as -

Total damage =

n1
N1

n2
N2

......

ni
Ni

....

nn
Nn

The Palmgren-Miner rule, described above is considered a simplified and versatile tool for determining
the total life of the structure under study. It is apparent that the detail under consideration is said to have failed
if the Total Damage becomes unity (1.0). The rule does not account for the effect of load sequence and load
interaction on damage accrued, and have an over simplified assumption of linear summation. However,
Palmgren-Miner rule is still widely used to estimate life of a structure, on account of its ease of application.
3.0 TRAIN LOADS
As of today in India the vast
railway network is the principal
mode of transport having
approximately 11000 trains
running daily carrying passengers
and freight. The trains are
broadly classified on the basis of
their usage as either a passenger
or goods train. Further, they are
classified on the basis of their
composition i.e., the type and
number of locomotives and
wagons. Ten standard broad
gauge train types are defined to
represent the existing trainloads.
The trains are modeled as series
of concentrated loads. The train
diagrams for the standard trains
are shown in Fig. 1.

Fig. 1 : IRS Standard Broad Gauge Trains Considered in Fatigue

Load Model (MBG)

163

4.0 RELEVANCE OF FATIGUE STRENGTH TO LOAD MODEL

Following factors primarily affect the fatigue strength of a typical connection
i)

Type of joint details

ii)

Stress range at the location under consideration.

iii) No. of cycles of stress range

The type of joint detailing is decided keeping in view the methodology of fabrication to be adopted.
Once the type of joint detail is finalised, the allowable stress-range, to which the connection can be subjected
to, can be obtained from relevant S-N curve for 'N' number of cycles. 'N' is usually taken as 2 million. S-N
curves developed for the purpose are shown in Fig. 2.

Figure-2 : Fatigue strength curves for director stress ranges

164

This implies that the connection of a particular category is able to safely withstand 2 million cycles of the
allowable stress-range. As practically observed, the different components of the structure undergoes different
no. of cycles of different stress-ranges. Therefore, every connection detail, over a period of time, is subjected
to a stress-range histogram consisting of number of stress-ranges and corresponding number of cycles. A
typical stress-range histogram is shown in Fig. 3.

Figure-3 : Typical stress range histogram

The actual damage to the connection detail is the cumulative effect of all such stress-ranges that are
included in the stress-range histogram. The concept of design life comes into picture at this stage, as the
cumulative damage should be equal to unity, at the end of design life. The stress-range histogram, to which
the detail is subject to, is a function of type of trains, frequency of trains, speed and the GMT etc. In practical
scenario, it is a complex phenomenon of cumulative fatigue damage, which will be very difficult to model
unless some kind of standard of load-frequency distribution is assumed. Therefore, In order to standardize
the stress-range histogram, it is necessary to standardize the load models so that the cumulative fatigue
damage can be adequately assessed over a period of time. The load model will have to specify the distribution
of train types and their frequencies with respect to their cumulative GMT and traffic volume. These parameters
in turn will have to be taken as design input for assessing the fatigue strength of connections.
5.0 TRAFFIC LOAD MODEL FOR MBG LOADING
To develop the load model, details of the existing traffic on selected BG routes of S.E., W, N & Central
Railways have been taken. The present level of traffic, the maximum traffic projections and the future loading
standards have been considered in the above load model. The projected traffic and its classification have
been proposed in such a way that the number of train compositions are as few as possible and all present and
future traffic projection are represented adequately.

165

5.1 Train Composition

5.1.1 Locomotives
The locomotives to be considered for various train compositions are proposed as specified in MBG
loading-1987. The appropriate type of locomotives, which will be more critical from the fatigue considerations
depending upon the loaded length, have been considered.
5.1.2 Passenger train composition
Passenger train having 15 and 22 coaches are running at present in many sections. In addition, it is
expected that in future 26 coaches trains will also be running. Hence trains with 15, 22 and 26 coaches have
been included. The trains having 26 coaches may have two AC coaches whose axle loads are 16.25t, while
all other coaches have 13t of axle loads. It is also expected that EMU rakes, which are presently of 9
coaches will get enhanced to 12 coaches due to increase in passenger traffic.
5.1.3 Goods train composition
Trains with 75 wagons of 4-wheeler and 40 BOX wagons are included, since they shall continue to
remain in service for a long time. Major traffic is expected to be carried by BOXN rakes with 22 t axle loads.
Two compositions have been indicated - one with 5,000 t train and the other with 10,000 t trains - similar to
the existing 4,500t train and 9000t trains. These groups shall represent rakes, with all other types of wagons.
So long as the maximum GMT remains the same, a train with heavier trailing load than 10,000 t can also be
permitted on the same bridge without producing any more adverse effect from fatigue considerations. Only
10,000 t trains have been proposed, though technology mission envisages upto 18,000 t trains.
5.2 Speed and dynamic augment
The speed of passenger train will be limited to 160 kmph and that of goods train to 100 kmph. The
dynamic augment will be taken as pr IRS Bridge Rules.
5.3 Class of Traffic
A distinction has been made between lines with heavy freight, mixed, suburban and light traffic. The
lines are proposed to be classified as follows:
a)

Heavy freight lines - 100 GMT

This line is expected to carry mostly 10,000 t trains made up of BOXN wagons having 22 t axle loads.
Some trains will also be of 5000 t capacity. Only nominal passenger traffic is expected to move on these
lines. This classification will cater for those lines which have large volume of traffic of iron ore or coal.
b)

Mixed traffic lines with heavy traffic - 70 GMT

At present the mixed traffic line Allahabad - Mughalsarai carries 44 GMT and has 48 trains running
daily. Most of the coal and other traffic will be moving in 10,000 t rakes. Similarly, 26 coach passenger rakes
166

are likely to cater to the increase passenger traffic. Therefore, maximum number of trains with 22 and 26
coaches is proposed. Thus there will be 48 trains giving 70 GMT.
c)

Suburban lines - upto 60 GMT

The most intensively used suburban line is on Western Railway, Bombay-Virar, which has an existing
traffic of 200 suburban trains per day and 10 other passenger, trains. Jessop's/ICF BG de EMU stock are
assumed to continue. However, it is expected that these EMU rakes, which are presently of 9 coaches, will
get enhanced to 12 coaches due to increased passenger traffic. At present there are 10 other passenger
trains on the same line. The same is proposed to continue with 22 and 26 coach rakes. Thus there will be
210 trains giving 57.7 GMT.
d)

Mixed traffic line with light traffic - 40 GMT

Five passenger trains each of 15 coaches and 22 coaches are proposed. In addition, there will be 19
freight trains, which include two 10,000t trains. The section will have to total 40 GMT with 33 trains.
It is expected that by judicious distribution of rakes, empty movement of heavy haul rakes will be
negligible and as movement of empty rakes is not likely to govern fatigue design, it has not been considered in
other classes of traffic. The Distribution of various Train Types in above traffic classes is shown in Table-1.

Type
Train

of

Train Type No.

Table 1: Traffic load Model of MBG Loading for Fatigue Assessment

Class of Traffic
Weight GMT
per
per Heavy Freight Mixed Traffic Suburban
Train
Traffic (60
Composition Train Train Traffic (100 Lines with
GMT)
Heavy Traffic
GMT)
(t)
(70 GMT)

Mixed Traffic
Lines with
Light Traffic
(40 GMT)

No. of GMT No. of GMT No. of GMT No. of GMT

Trains
Trains
Trains
Trains

Passenger
Trains

1
2
3
4

1+15
2+22
2+26 AC
EMU 12

900
1400
1700
700

0.33
0.51
0.62
0.26

3
2
-

1.0
1.0
-

6
10
14
-

2.0
5.1
8.7
-

5
5
200

2.6
3.1
52.0

5
5
-

1.7
2.6
-

Freight
Trains
loaded

5
6
7
8

1+75-4
2+40 BOX
2+55 BOXN
2(2+55
BOXN)

3200
3600
5100
10300

1.17
1.31
1.86
3.76

2
2
10
20

2.3
2.6
18.6
75.2

2
4
12

2.3
7.4
45.1

2
5
10
2

2.3
6.5
18.6
7.5

Freight
Trains
empty

1+75-4
Wheeler
2+40 BOX

1100

0.40

0.8

1300

0.47

0.9

39

100.7

48

70.6

210

57.7

33

40.9

Total

10

167

6.0 CONCLUSION
A standard load model for MBG loading has been developed for fatigue assessment of bridge details.
The load model developed can be used for generating the route specific stress range & frequency histograms
and the cumulative fatigue damage can be worked out by using Palmgren-Miner rule. Fatigue assessment of
bridge components with specific detail of connection and a given design life, can be rationally carried out by
using the above load model. The load model for HM Loading can also be developed on similar lines.
7.0 REFERENCES
7.1

Revision of Fatigue Provisions in IRS Steel Bridge Code (2004), First Interim Project Report submitted by Department
of Earthquake Engineering, Indian Institute of Technology, Roorkee to Research Designs Standards Organisation,
Ministry of Railways, Lucknow (UP)-226011.

7.2

Stress Spectra for Fatigue Design of Railway Bridges (1991), Project report submitted by Department of Civil
Engineering, Indian Institute of Technology, Kanpur to Research Designs Standards Organisation, Ministry of
Railways, Lucknow (UP)-226011.

7.3

Technical Documents on Traffic Details for Revision of Fatigue Provisions of IRS Steel Bridge Code (1989), file
No.CBS/PSB, Research Designs Standards Organisation, Ministry of Railways, Lucknow (UP)-226011.

7.4

IRS Bridge Rules (1986), Research Designs & Standards Organisation, Ministry of Railways, Lucknow (U.P.).

x x x

168

ANALYTICAL APPROACH FOR ASSESSMENT OF

FATIGUE LIFE OF STEEL GIRDERS OF BRIDGE NO.110
(Up Line) ACROSS RIVER GANGA ON
LUCKNOW-KANPUR SECTION OF INDIAN RAILWAYS
RAVINDRA KUMAR GOEL*
(Published in Journal of IPWE, July, 2007)
ABSTRACT
Assessment of residual fatigue life of Steel bridges is done by RDSO in case of major bridges referred by Zonal
Railways. Conventionally, such assessment requires gauging & instrumentation on the important and critical members of
the bridge and fatigue testing of steel samples drawn from the bridge. A review of fatigue provisions of British Standard has
shown that BS-5400 takes into account a factor for design life. This factor is dependent on type of loading and the traffic
volume being carried on the bridge. The MBG loading of Indian Railways is found matching with RU loading given in British
Standard. Therefore, the parameters given in BS-5400 for RU loading were considered applicable to analyze the fatigue life
of steel girders on Indian Railways.
The paper describes the simplified approach of British Standard and analyses the life of the existing girders. The
results are compared with those obtained after instrumentation and testing of the bridge. The fatigue life estimated by
using the simplified approach of BS: 5400 is found on higher side as compared to the estimation based on actual testing
where concentration factors at joints have been assumed on conservative side in absence of actual measurements. Therefore,
the simplified approach of BS-5400 Part-10 can be satisfactorily adopted for quick assessment of fatigue life of steel girders
on Indian Railways.

1.0 INTRODUCTION
Indian Railways in its vast network is having 1,19,984 bridges, out of which, large numbers are of steel.
The bridges constructed upto 1905 have been declared as early steel bridges, which are to be replaced as a
matter of policy. There are still a large number of steel bridges in service which are not of early steel and field
engineers, quite often find themselves in dilemma, to decide about their replacement. It is difficult to ascertain,
from fatigue consideration, whether sufficient residual life of the bridge exists or not. To assist the Zonal
Railways, Research Design and Standard Organization take up such works of major bridges on specific
request of the railways. In this context, the work of fatigue life estimation of Ganga Bridge No.110 Up line of
Lucknow-Kanpur section was taken in hand. Based on detailed study it was concluded that the bridge is still
having life of about 40 years and hence the sanctioned work of regirdering was pended, thus saving the public
money of approximately 15 crores.

*Director, Bridges & Structures Directorate, Research Designs & Standards Organization, Indian Railways, Manak Nagar, Lucknow226001

169

There are different approaches to determine residual life of steel girder brides. Palmgren Miner's theory
is most common and easy to understand. The work involves strain gauging of the critical members and getting
actual time histories for a period for 24 hrs and testing the samples of the steel for fatigue characteristics.
Thereafter, the recorded data is analyzed to work out the residual fatigue life of the bridge components. The
method is time consuming and a period of more than six months is required to complete the fatigue tests. A
new approach based on British Standard BS:5400 part 10 was applied to assess the fatigue life of various
members of the steel girders of the bridge to get a quick result, in order to decide not to replace the girders of
the bridge. The analysed values of the residual fatigue life were subsequently confirmed by the conventional
approach.
2.0 OBJECTIVE OF THE STUDY
To analytically assess the residual life of the girders of Br. No.110 Up between Lucknow-Kanpur
section, which is not a early steel girder bridge and already given a service life of about 79 years.
3.0 DETAILS OF BRIDGE
The existing bridge having double BG track consisting of 25x30.4m+1x12.19m span The Up line girders
are of 1926. The general arrangement and the configuration of the girder are shown in fig.1.

Fig. 1 - General configuration of Ganga Bridge, Br. No. 110 (UP Line),
Span - 25 X 30.48 M + 1 X 12.19 M

Fig. 2 - View of Up/Dn line bridge girders from the river bed

170

Fig. 3 - View of the bridge girder from the river bed

4.0 IRS PRACTICE OF FATIGUE DESIGN

The Code of Indian Railway Standards for Steel Bridge recommends method to allow for the effect of
fatigue in design of parts of steel bridges, which are subjected to repeated fluctuations of stress. These
fluctuations may cause fatigue failure of members or connections at lower stresses than those at which they
would fail under static load.
Such failures are primarily due to stress concentrations introduced by the constructional details. Thus
all the details are designed to avoid as far as possible the stress concentrations likely to result in excessive
reductions of the fatigue strength of members or connections. Care is also taken to avoid a sudden reduction
of the section of a member or part of a member, especially where bending occurs.
To allow for the effect of fatigue the allowable working stress is determined from the of IRS Steel Bridge
Code for wide range of constructional details. The code covers mild and high tensile steel fabricated or
connected by welding, riveting or bolting. The allowable stress 'P' depends on the ratio of minimum stress
smin to maximum stress smax, number of repetitions of stress cycle 'N', the method of fabrication and the type
of connection. In determining the ratio smin/smax gross area is considered and the bridge members are generally
designed for 10 million cycles of stresses produced under minimum and maximum of the design load.
Concept of EUDL (Equivalent Uniformly Distributed Load) is used to determine the maximum bending
moment and maximum shear force for the type of IRS loading. The EUDL for maximum bending moment
and maximum shear force depends upon the span and the dynamic augment increases with speed. IRS
Bridge Rules gives tables for determination of EUDL for maximum bending moments and shear forces alongwith
the co-efficient of dynamic augment for maximum speed of 160 kmph.
There is no rational basis for adopting counts of 10 million number of cycle to determine the allowable
stress levels. As per latest research available, Fatigue is a cumulative phenomenon; which is not reflected in
the above procedure. Stress-ratio procedure does not take into account the effect of all stress ranges
experienced by a member. World over, the stress range concept is followed and material S-N curve forms
171

the basis of all fatigue analysis and design. Due to these short comings, it is not considered proper to analyze
the fatigue life of existing bridge by using the provisions of IRS Steel.
5.0 DESIGN PROVISIONS OF BS-5400
BS-5400 part 10 is a comprehensive code which is based on the concept of cumulative fatigue damage.
The code concerns with the fatigue design methodology for highway and railway bridges and take into
consideration the various drawbacks of IRS approach. Fatigue life assessment is based on the S-N curve
approach using Palmgren-Miner's damage summation model wherein the number of cycles to failure is dependent
only on stress range and not on maximum stress values. The methodology for determination of stress range
has been described for welded and non-welded details and a simplified method has been given for determining
the limiting value of the maximum range of stress for the specified design life for two different types of standard
loadings. The code specifies different factor k1, k2, k3, k4 & k5 for design parameters such as design life,
multiple cycle of stress loading, type of standard loading, annual GMT and multiple lane loading respectively.
The code gives specific methodology and tables to calculate the factors for different design parameters.
5.1 Calculation of Limiting Stress Range, sT
The constant amplitude non-propagating stress range, s0 for the constructional detail is chosen
appropriately on the basis of Table - 17 & Table 8 of the code. The limiting stress range sT can now be
calculated for RU loading as under:
sT = k1. k2. k3. k4. k5. s0 ..(1)
5.2 Fatigue Criteria for Design Adequacy
The design adequacy of the given detail is now checked as per Clause 9.2.2.2 and Clause 9.2.2.3 of
the code. If maximum stress range (sRmax) does not exceed limiting stress range (sT), i.e sRmax sT, the detail
may be considered to have a fatigue life in excess of the specified design life.
6.0 STANDARD TYPE LOADINGS
Two type of standard loadings, namely RU loading & RL loading have been considered in BS:5400,
which are shown in Fig.1 & Fig. 2 respectively.

200 kN
50 kN/m

25 kN/m

25 kN/m

100m
No limitation

No limitation
Fig.4 RU Loading

172

80 KN/m
0.8

250

250

1.6

250

1.6

250

1.6

80 KN/m
0.8

No limitation

No limitation
Fig.5 RL Loading

7.0 COMPARISON OF RU LOADING WITH IRS-MBG LOADING

A comparison of EUDL values of Bending Moment as per RU loading was made with corresponding
values for IRS-MBG loading (Table-1). It is found that the EUDL values as per RU loading are on higher side
Table 1. EUDL for RU Loading and IRS MBG Loading
Sl.
No.
(1)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

Span in m
(2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
22
24
26
28
30

EUDL for maximum B.M. in (including dynamic augment)

RU Loading
IRS MBG loading
(3)
(4)
1000
980
10.07
980
1089
980
1351
1162
1541
1391
1652
1523
1817
1608
1951
1688
2066
1750
2171
1817
2268
2078
2359
2195
2447
2317
2531
2415
2613
2497
2694
2566
2773
2623
2851
2699
2927
2772
3003
2864
3153
3049
3301
3231
3447
3403
3592
3573
3736
3741

173

%age difference
(5)
2.04
2.76
11.12
16.25
10.79
8.50
13.00
15.56
18.04
19.51
9.14
7.47
5.60
4.81
4.64
4.98
5.72
5.63
5.58
4.87
3.42
2.17
1.28
0.52
-0.15

as compared to IRS-MBG loading. Therefore, the various factors developed for RU loading as given in BS:
5400 Part 10 have been used for fatigue life analysis of the members of the bridge subjected to MBG loading
which is expected to give a fatigue life on a conservative side.
8.0 FATIGUE LIFE ANALYSIS
The general arrangement drawing and the stress sheet of the bridge have been studied. The stresses
calculated and the cross sections provided in the existing design have been used to workout the stress ranges
to which the members are subjected to and the fatigue life has been estimated as per simplified approach of
BS: 5400 Part -X.
Following assumptions have been made during this study a)

The maximum axial stresses due to EUDL for IRS loadings have been worked out and the maximum
stress range has been calculated as the difference of dead load stress and the maximum stress likely to
come in the member with DL, live load with impact and occasional load.

b)

The axial stresses due to load combination with occasional load have been taken into consideration to
find out the maximum stress range. This combination with extreme loads rarely occurs in practice,
therefore, the analysis is on conservative side.

c)

Material properties are assumed to be as per Table-8 of BS-5400 and s0 value has been taken as 47 N/
mm2 corresponding to detail class 'E' of this table.
Table 8* : sr - N relationships and constant amplitude non-propagation stress range values
0 (N/mm2)

Detail class

K2

3.0

0.16 X 1012

25

3.0

0.25 X 10

12

29

F2

3.0

0.43 X 1012

35

3.0

0.63 X 1012

40

3.0

1.04 X 1012

47

3.0

1.51 X 1012

53

3.5

4.23 X 1013

78

4.0

1.01 X 1015

100

8.0

2.08 X 1022

82

* Table No. corresponds to the one given in code.

d)

The fatigue life has been assessed by calculating the design life factor k1. This factor has been worked
out as sRmax/(s0 x k3) and fatigue life calculations have been done by inversion, using the equations given
in Clause 9.2.3 of BS-5400 Pt.10 by taking fatigue life as minimum of following:

174

120
m
k1 ..(2)

Fatigue Life
or

120
m 2 (3)
k1

Fatigue Life

Where m = 3.0 taken from Table-8 for detailed Class 'E'.

e)

Multiple cycles factor, k2 has been assumed as 1.0 as the loading event stress range histogram is assumed
to have a single stress range for simplicity

f)

Value of RU loading factor k3 has been taken from Table-4 of the code considering the case of heavy
traffic loading, corresponding to the base length (L) of the influence line diagram for the member.

g)

Value of GMT factor, k4 is assumed as 1.0 for GMT of 18 to 27 million tonnes. The GMT carried by the
bridge since 1981 is given in Table -2. Actual GMT data suggests that this assumption is very much on
conservative side.
Table 4* : Value of k3 for RU loading of Railway bridges
Heavy traffic
Detail stress

D
E
F
F2
G
W

Length,L(m)

Values of k3

< 3.4
3.4 to 4.0
4.0 to 4.6
4.6 to 7.0
7.0 to 10.0
10.0 to 14.0
14.0 to 28.0
> 28.0

1.00
1.09
1.23
1.37
1.53
1.71
1.92
2.19

1.00
1.09
1.22
1.36
1.56
1.75
1.95
1.95

Medium traffic

1.01
1.13
1.30
1.46
1.62
1.62
2.03
2.03

1.14
1.28
1.46
1.65
1.65
165
1.83
1.83

D
E
F
F2
G
W
1.09
1.23
1.37
1.53
1.71
1.92
2.19
2.46

Light traffic

1.09 1.13
1.22 1.30
1.36 1.46
1.561.7 1.62
5
1.81
1.952.1 2.03
8
2.03
2.18 2.03

1.28
1.46
1.46
1.65
1.83
1.83
1.83
1.83

* Table No. corresponds to the one given in code.

175

D
E
F
F2
G
W
1.37
1.53
1.71
1.92
2.19
2.46
2.74
3.06

1.60
1.79
1.79
2.05
2.31
2.31
2.56
2.87

1.60
1.80
1.80
2.00
2.34
2.50
2.50
2.50

1.71
1.71
1.95
2.20
2.20
2.20
2.20

TABLE - 2 GMT of Bridge No. 110, (up line, since 1981)

Year

Annual Traffic density, GMT

Annual GMT

Cumulative GMT

1981-82

3.64

3.64

1982-83

3.90

7.54

1983-84

3.22

10.76

1984-85

3.67

14.43

1985-86

3.59

18.02

1986-87

3.65

21.67

1987-88

5.36

27.03

1988-89

5.93

32.96

1989-90

5.88

38.84

1990-91

6.17

45.01

1991-92

5.29

50.30

1992-93

5.67

56.97

1993-94

6.78

62.75

1994-95

6.58

69.33

1995-96

6.58

75.91

1996-97

6.99

82.90

1997-98

6.56

89.46

1998-99

7.58

97.04

1999-00

8.45

105.49

2000-01

16.40

121.89

2001-02

14.60

136.49

2002-03

10.30

146.79

Table - 5* Value of K4 for railway bridges

Annual traffic tonnage on one track
(million of tones)
42 to 27 27 to 18 18 to 12 12 to 7 7 to 5
0.89
1.0
1.13
1.27
1.42
*Table No. corresponds to the one given in code.

176

<5
1.6

h)

For single lane loading value of lane factor, k5 is taken as 1.0.

i)

Value of design life factor k1 has been worked out now by equating the limiting stress (sT) to maximum
stress (sRmax) obtained from analysis as per stress sheet.

The design calculations for assessment of fatigue life of members of truss girders has been done as given
in Table - 3 for standard annual GMT of 27 to 18. The verticals U1-L1 and U11-L11 are found to have
minimum fatigue life of 141 years. As per record the bridge was constructed in 1926, therefore the remaining
fatigue life is estimated as 61 years or say 60 years.
Table - 3 Calculation Sheet For Fatigue Life Assessment
dead load
S.
Member
No.
(t)
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35

2
U 0-U 1
U 1-U 2
U 2-U 3
U 3-U 4
U 4-U 5
U 5-U 6
U 6-U 7
U 7-U 8
U 8-U 9
U 9-U 10
U 10-U 11
U 11-U 12
U 0-L 1
U 1-L 1
L 1-U 2
U 2-L 3
U 3-L 3
L 3-U 4
U 4-L 5
U 5-L 5
L 5-U 6
U 6-L 7
U 7-L 7
L 7-U 8
U 8-L 9
U 9-L 9
L 9-U 10
U 10-L 11
U 11-L 11
L 11-U 12
L 1-L 3
L 3-L 5
L 5-L 7
L 7-L 9
L 9-L 11

3
-47.45
-48.11
-83.19
-83.52
-107.06
-107.21
-107.21
-107.06
-83.52
-83.19
-48.11
-47.45
65.3
-5.99
-33.18
19.36
16.1
-25.24
13.93
16.1
-5.78
-5.78
16.1
13.93
-25.24
16.1
19.36
-33.18
-5.99
65.3
72.95
98.58
110.76
98.58
72.95

A
(cm 2)
4
393.31
393.31
393.31
393.31
393.31
393.31
393.31
393.31
393.31
393.31
393.31
393.31
222.30
64.26
116.13
99.49
64.26
116.13
99.49
64.26
116.13
116.13
64.26
99.49
116.13
64.26
99.49
116.13
64.26
222.3
241.97
267.45
267.45
267.45
241.97

Miin.
Stress L.L with
CDA
min
(t)
( t/cm 2)
5
-0.12
-0.12
-0.21
-0.21
-0.27
-0.27
-0.27
-0.27
-0.21
-0.21
-0.12
-0.12
0.29
-0.09
-0.29
0.19
0.25
-0.22
0.14
0.25
-0.05
-0.05
0.25
0.14
-0.22
0.25
0.19
-0.29
-0.09
0.29
0.3
0.37
0.41
0.37
0.3

6
-93.36
-94.55
-162.81
-163.44
-209.23
-209.54
-209.54
-209.23
-163.44
-162.81
-94.55
-93.36
142.29
-37.52
-77.88
53.93
34.37
-71.22
53.51
34.37
-46.36
-46.78
34.37
53.51
-71.22
34.37
53.93
-77.88
-37.52
142.29
144.35
193.74
217.44
193.74
144.35

Occ.
Load
(t)
7
-58.29
-63.16
-63.65
-66.64
-61.12
-57.7
-57.7
-61.12
-66.64
-63.65
-63.16
-58.29
12.13
-1.53
-9.86
6.98
0
-5.51
4.22
0
-3.11
-3.11
0
4.22
-5.51
0
6.98
-9.86
-1.53
12.13
26.38
27.93
31.92
27.93
26.38

Max.
Max.
Load
Stress
Dl+LL+
max
OL
(
t/cm 2)
(t)
8
9
-199.1
-0.51
-205.82 -0.52
-309.65 -0.79
-313.6
-0.8
-377.41 -0.96
-374.45 -0.95
-374.45 -0.95
-377.41 -0.96
-313.6
-0.8
-309.65 -0.79
-205.82 -0.52
-199.1
-0.51
219.72
0.99
-45.04
-0.7
-120.92 -1.04
80.27
0.81
50.47
0.79
-101.97 -0.88
71.66
0.72
50.47
0.79
-55.25
-0.48
-55.67
-0.48
50.47
0.79
71.66
0.72
-101.97 -0.88
50.47
0.79
80.27
0.81
-120.92 -1.04
-45.04
-0.7
219.72
0.99
243.68
1.01
320.25
1.2
360.12
1.35
320.25
1.2
243.68
1.01

177

Stress
Range
Rmax
(t/cm 2)
10
-0.39
-0.4
-0.58
-0.59
-0.69
-0.68
-0.68
-0.69
-0.59
-0.58
-0.4
-0.39
0.7
-0.61
-0.75
0.62
0.54
-0.66
0.58
0.54
-0.43
-0.43
0.54
0.58
-0.66
0.54
0.62
-0.75
-0.61
0.7
0.71
0.83
0.94
0.83
0.71

RU
Rmax k1 = Design
loadin
L'
Rmax
Life
g
N
(m) factor,
/mm 2 0 *k3 years
k3
11
32.16
32.16
32.16
32.16
32.16
32.16
32.16
32.16
32.16
32.16
32.16
32.16
32.16
5.41
29.17
26.26
5.334
23.35
20.44
5.334
17.53
17.53
5.334
20.44
23.35
5.334
26.26
29.17
5.41
32.16
32.16
32.16
32.16
32.16
32.16

12
2.19
2.19
2.19
2.19
2.19
2.19
2.19
2.19
2.19
2.19
2.19
2.19
2.19
1.37
2.19
1.92
1.37
1.92
1.92
1.37
1.92
1.92
1.37
1.92
1.92
1.37
1.92
2.19
1.37
2.19
2.19
2.19
2.19
2.19
2.19

13
-39
-40
-58
-59
-69
-68
-68
-69
-59
-58
-40
-39
70
-61
-75
62
54
-66
58
54
-43
-43
54
58
-66
54
62
-75
-61
70
71
83
94
83
71

13
0.379
0.389
0.563
0.573
0.670
0.661
0.661
0.670
0.573
0.563
0.389
0.379
0.680
0.947
0.729
0.687
0.839
0.731
0.643
0.839
0.477
0.477
0.839
0.643
0.731
0.839
0.687
0.729
0.947
0.680
0.690
0.806
0.913
0.806
0.690

15
2206
2045
671
637
398
416
416
398
637
671
2045
2206
382
141
310
370
203
307
452
203
1109
1109
203
452
307
203
370
310
141
382
366
229
158
229
366

9.0 FATIGUE LIFE ASSESSED USING CONVENTIONAL APPROACH

The fatigue life assessment of various members was also done based on strain gauging. The observed
values of the stress range and cycles during the 24 hours of period are given in Table - 4. Considering the
different safety factors and S-N curve developed for limited number of fatigue testing of samples, residual life
of different members has been worked out using conventional Palmgrain - miner technique. The estimated
residual life has been shown in Table -5.
Table -4 Observed Stress Ranges and Cycles During 24 Hrs
S. No.

Member

Observed Stress Range

(N/mm2)

Modified stress range

(N/mm2) Observed X 2.5

No. of Cycle

1.

Bottom Chord

0 - 10

0-25

761

10 - 20

25-50

14

20 - 30

50-75

23

30 - 40

75-100

12

End Raker

0 - 10

0-25

845

10 - 20

25-50

17

20 - 30

50-75

32

Vertical

0 - 10

0-25

1448

10 - 20

25-50

173

Diagonal

0 - 10

0-25

845

10 - 20

25-50

20

20 - 30

50-75

21

Top Chord

0 - 10

0-25

725

10 - 20

25-50

34

20 - 30

50-75

2.

3.
4.

5.

Table - 5 Estimated Residual Fatigue Life ( As Per Conventional Approach)

Member
Bottom Chord
Vertical
End Raker
Diagonal
Top Chord

Residual Life in years

40
50
61
85
133

178

10.0 COMPARISON
It is apparent that the fatigue life assessment based on actual testing is giving a residual life of 50 years for
vertical as against the estimate of 60 years given by analytical approach based on BS:5400. The actual life
assessment, however, is found governed by bottom chord as the residual fatigue life for this member is found
40 years only. This can be attributed to conservative factors adopted in the damage calculations as the actual
stress concentrations at joints are not measured.
11.0 CONCLUSIONS
1.

The RU loading of BS:5400 is found matching with the IRS MBG loading. The design factors given in
BS:5400 are therefore applicable for IRS bridges.

2.

The analytical approach based on simplified method of BS:5400 can be satisfactorily adopted to make
a quick assessment of residual fatigue life of the steel bridges on Indian Railways.

12.0 AREAS OF FURTHER RESEARCH

1.

Above study has been conducted for assessing the fatigue life of plate girder bridges based on stress
range concept of BS-5400 with 'E' class of connections. Actual fatigue life of connection is not assessed
which may actually be governing the fatigue strength of members. Stress concentration factors for
various type of connections need to be established for better assessment.

2.

The material S-N curve given in BS-5400 has been used in the study assuming that the material
characteristics of the Indian Steel (IS-2062) are the same. Detailed study on fatigue characteristics of
Indian steel is, therefore required and the results to be validated.

13.0 REFERENCES
1.

British Standard, British Standards Instit, 1980 "BS:5400 Steel, Concrete and Composite Bridges, Part 10, Code of
Practice for Fatigueution, London.

2.

British Standard, British Standards Institution, 1980, "BS-5400 : Steel, Concrete and Composition Bridges, Part 2,
Specification for Loads, London.

3.

Indian Railway Standard. (1962). "Code of Practice for the Design of Steel/Wrought Iron Bridges (Steel Bridge
Code)". Research Designs & Standards Organisation, Ministry of Railways, Lucknow (U.P.).

4.

Indian Railway Standard. (1986). "Bridge Rules". Research Designs & Standards Organisation, Ministry of Railways,
Lucknow (U.P.).

5.

Office of Research & Experiment. (1976). "Statistical distribution of axle loads and stresses in railway bridges".
Report No. ORE D-128/RP5, International Union of Railways, Paris.

6.

Ravi, G. and Ranganathan, R. (1991). "Critical study of fatigue design of bridges as per BS:5400-Part 10". International
symposium on Fatigue and Fracture in Steel and Concrete Structures, Structural Engineering Research Centre,
Madras, India.

7.

Gupta, R.K. and Goel, R.K. (2005), "Review of Fatigue Provision of BS:5400 for Design of Railway Bridges" No. 312313, Vol Lx1, Feb-May, 2005, Indian Railway Technical Bulletin, Research Design Standard Organisation, Lucknow

179

8.

UIC Code. (1994). "Loads to be considered in Railway Bridge Design". Leaflet No. 776-1, International Union of
Railways, Paris.

9.

Estimation of the residual life of floor system of Ganga Bridge near Balawali ( Civil Engineering Report No. 175 )

10.

Investigations on the assessment of residual life of early steel/wrought iron girder bridges ( Civil Engineering Report
No. 245 )

11.

Assessment of fatigue life of early steel/wrought iron girder bridges ( Report No. BS-5 )

12.

Guidelines for assessment of residual life of early steel/wrought iron girder bridges (Report No. BS-39 )

13.

Statistical distribution of axle loads and stresses in railway bridges ( ORE Report No.1 Question D-128 )

14.

6. Fatigue life of riveted railway bridges by Bjorn Akesson, Chalmers university of technology, Sweden .

15.

Condition monitoring and life assessment of railway bridge 449/3/34 over river Tapti near Bhusaval by Department of
civil engineering IIT, Mumbai

16.

Assessment of residual fatigue life of Ganga Bridge near Kanpur, N. Rly. (Report No. BS- 70 )

17.

Code of practice for fatigue ( BS-5400: Part 10 )

x x x

180

FATIGUE ASSESSMENT CRITERIA FOR DESIGN AND

ANALYSIS OF STEEL GIRDER BRIDGES FOR HEAVY
AXLE LOAD OPERATIONS
Piyush Agarwal* & R.K. Goel**
(Presented in IPWE Seminar held at Jaipur, Jan., 2007)

1.0 INTRODUCTION
1.1 Railway bridges are subjected to heavy fluctuating dynamic loads. These fluctuations cause fatigue
failure of members or connections at lower stresses than those at which would otherwise fail under static
load. IRS provisions, which are based on stress ratio concept does not taken into account the
phenomenon of fatigue adequately. World over, such effects are taken into account following the Palmgren
Miner cumulative damage rule based on stress range concept. British Code BS-5400 part 10 and more
recently Euro Code EN-1993-1-9: 2002 address the fatigue effects in a rational manner by taking
design parameters such as route GMT, type of traffic, design life and detailed category of connection.
However, these provisions are specific to loading conditions prevailing in their countries and the design
parameters as given in these codes can not be straightaway applied for design of bridges in traffic
conditions prevailing in India.
1.2 Recently, Indian Railways has taken a leap forward to cope up with the increased demand of freight
transportation by deciding to have a dedicated freight corridor (DFC) for which a new set of standard
designs would have to be prepared. This would be a big challenge for Bridge Designers to develop the
new designs, based on rational criteria for fatigue in accordance with established international practices.
Development of a fatigue load model for such a dedicated freight route is the first design input required
by the designers. In this paper an attempt has been made to develop such a load model based on the
anticipated axle load, type of locos and the train lengths. This model with modifications can be
subsequently used for analyzing the existing bridges on feeder routes.
2.0 PLAMGREN-MINER LINEAR DAMAGE RULE
2.1 Palmgren has proposed a damage model on the basis of constant energy absorption per cycle. The
energy absorption per cycle leads to linear summation of damage. Miner has subsequently, represented
this concept in mathematical form. Palmgren-Miner rule states that the fatigue damage contribution by
each individual load spectrum at a given stress level is proportional to the number of cycles applied at a
stress interval, ni, divided by the total number of cycles to failure at the same stress level, Ni. It is

Executive Director/Bridges & Structures /RDSO/Lucknow,

Director/Bridges & Structures /RDSO/Lucknow.

**

181

obvious that each ratio can be equal to unity if the fatigue cycles at the same stress level would continue
until failure occurs. The total damage, in terms of partial cycle ratios or damage, can be written as -

Total damage, Dd =

n1
N1

n2
N2

......

ni
Ni

....

nn
Nn

2.2 The Palmgren-Miner rule, described above is considered a simplified and versatile tool for determining
the total life of the structure under study. It is apparent that the detail under consideration is said to have
failed if the Total Damage becomes unity (1.0). The rule does not account for the effect of load sequence
and load interaction on damage accrued, and have an over simplified assumption of linear summation.
However, Palmgren-Miner rule is still widely used to estimate life of a structure, on account of its ease
of application.
3.0 TRAIN LOADS AND TRAFFIC MODEL
3.1 Railway Board vide its letter No.2006/CE-II/TS/2 dated 26-10-06 has advised the axle load and
Track Loading Density (TLD) for design of foundation and bridges. As per this letter axle load of 32.5t
and TLD of 12 t/m has to be adopted for design of bridges. Based on this information, the freight train
compositions including their length, weight and GMT etc. have been developed for light, medium and
heavy traffic classifications.
3.2 The wagon details for DFC are yet to be finalised, therefore the wagon details (axle spacings) of IRS
HM loading have been taken in the above analysis. No passenger trains have been considered and the
existing combinations of HM loading have been modified to develop the above load model for fatigue
assessment. The traffic classification and load model developed have been shown in Table-1 & Table2 respectively. The salient features of the load model are described as under:
3.2.1 Loading :
For 32.5 t, none of the existing loadings given in IRS Bridge Rules is suitable. Therefore, the train
formations have been tentatively developed on the basis of the details of locomotives and wagons for HM
loading as available in Bridge Rules with following modificationsi)

The number of train formations, have been reduced to 12 from 17 by removing the similar type of
formations.

ii)

Trailing Load density: The Gondola wagon as taken in HM loading has been adopted with same
axle spacing and increased axle load of 32.5 t. It gives a trailing load density of 12.79 t/m. Railway
board vide its letter no.2006/CE-II/TS/2 dated 26-10-06 has instructed to design the bridges for
TLD of 12 t/m. During discussion with Wagon Directorate of RDSO, it has been learnt that
wagons of BOXN type are being modified for heavy axle load and these would be giving a TLD of
12.33 t/m. The proposed TLD is therefore, slightly on higher side and the designs of standard
spans shall be safer. The locomotives of HM loading have been adopted with their axle spacings
unchanged. However, the axle loads of all the locomotives have been proposed as 32.5 t for stress
analysis. This will take care of any possible increase in axle loads of the locomotives.
182

iii) Tractive Effort: In the HM loading the tractive effort of locomotives has been observed as 60t, 45t
and 30.5t depending upon the number of locomotives coupled together. The maximum tractive
effort of HM loading is 135 t with three locomotives of WAG6C & WAG6B. It has been learnt that
increasing the axle load increases the tractive effort of locomotive. As discussed with concerned
directorates there is no possibilities of increasing the tractive effort of an electric loco beyond 75t.
In such a case only 2 locomotives would be normally sufficient for heavy haul. For heavier traction,
three or four locomotives of lesser tractive effort can be coupled. In the proposed loading the
maximum tractive effort of 180 t has been proposed with a combination of 3 locomotives. The
comparison of locomotives and their tractive efforts for double, triple or quadruple traction has
been shown in table given below:
Tractive Effort

S.No.

No. of Locos

1.

2.

135 t@ 60 t each

180t@ 60 t each

3.

121.8 t@30.45t each

180t@ 45 t each

HM loading

Heavy Loading

120 t @ 60 t each

150t@ 75 t each

iv) Braking Forces: The braking force of the locomotive has been indicated against the train formations
in Table-2. The braking force of the train load has been taken as 13.4% of train load as has been
done in the case of HM loading. Therefore it is in accordance with prevailing practice.
3.2.2 Speed :
It has been advised by Railway Board that the maximum permissible speed of freight trains of heavy axle
load would be 100 kmph. It was also learnt that speed trials are required to be done with 10% extra speed.
Further keeping in view the possibilities of 10% increase in future, design speed of 125 kmph has been
proposed. It is also in conformity with the Coefficient of Dynamic Augment (CDA), given in IRS Bridge Rules
which has been developed for a speed of 125 kmph.
Table-1 Traffic Classification for Fatigue Assessment

Type of
Train

Train
Length of
Weight per
formatio
train (First
Train Composition
Train
n
axle to
(t)
No.
Last axle)
1
2

Freight trains
loaded
(Gondola
Type
Wagon)

Freight trains
empty

3
4
5
6
7
8
9
10
11
12

2WAG6C Type+40
2(2WDG2
Type
+40)
3WDG2+55
3WDG2 Type+55
3WAG6C Type+55
3WAG6A+75
3WAG6B+75
3WAG6C+75
4WAG6C Type+75
4WDG2 Type+75
2WAG
6C
Type+40
3WDG2+55

GMT
per
train

Heavy
(150 GMT)
No. of
Trains
per day

Class of Traffic
Medium
(100 GMT)

No. of
GMT Trains
per day

GMT

Light
(50 GMT)
No. of
Trains
GMT
per
day
5
10.145
2
8.116

445.100
886.184

5560
11120

2.029
4.058

4.058

5
2

10.145
8.116

612.952
612.952
618.482
820.312
821.662
821.762
842.684
835.364
445.100

7519
7690
7690
10119
10119
10119
10470
10470
1392

2.744
2.807
2.807
3.693
3.693
3.693
3.822
3.822
0.51

1
4
4
5
5
5
8
9
-

2.744
11.228
11.228
18.465
18.465
18.465
30.576
34.398
-

5
4
4
4
3
3
1
1
1

13.72
11.228
11.228
14.772
11.079
11.079
3.822
3.822
0.51

3
2
2
1
1
1
1

8.232
5.614
5.614
3.693
3.693
3.693
0.51

612.952

1788

0.65

0.65

0.65

183

Table - 2 Train Formations Considered in Load Model

184

4.0 RELEVANCE OF LOAD MODEL WITH DESIGN LIFE OF BRIDGE

4.1 Following factors primarily affect the fatigue strength of a typical connectioni)

Type of joint details

ii)

Stress range at the location under

consideration.

iv) No. of cycles of stress range

4.2 The type of joint detailing is decided keeping in
view the methodology of fabrication to be adopted.
Once the type of joint detail is finalised, the
allowable stress-range, to which the connection can
be subjected to, can be obtained from relevant SN curve for 'N' number of cycles. 'N' is usually
taken as 2 million. S-N curves developed for the
purpose are shown in Fig. 1.
4.3 It implies that the connection of a particular category
is able to safely withstand 2 million cycles of the
allowable stress-range. As practically observed,
185

Figure-1 : Fatigue strength curves for director

stress ranges

the different components of the structure undergoes different no. of cycles of different stress-ranges.
Therefore, every connection detail, over a period of time, is subjected to a stress-range histogram
consisting of number of stress-ranges and corresponding number of cycles. A typical stress-range
histogram is shown in Fig. 2.

Figure-2 : Typical stress range histogram

4.4 The actual damage to the connection detail is the cumulative effect of all such stress-ranges that are
included in the stress-range histogram. The concept of design life comes into picture at this stage, as the
cumulative damage should be equal to unity, at the end of design life. The stress-range histogram, to
which the detail is subject to, is a function of type of trains, frequency of trains, speed and the GMT etc.
In practical scenario, it is a complex phenomenon of cumulative fatigue damage, which will be very
difficult to model unless some kind of standard of load-frequency distribution is assumed. Therefore, In
order to standardize the stress-range histogram, it is necessary to standardize the load models so that
the cumulative fatigue damage can be adequately assessed over a period of time. The load model will
have to specify the distribution of train types and their frequencies with respect to their cumulative GMT
and traffic volume. These parameters in turn will have to be taken as design input for assessing the
fatigue strength of connections.
5.0 FATIGUE ASSESSMENT CRITERIA
5.1 It is well known that the fatigue provisions of IRS Steel Bridge Code are based on stress ratio concept,
which is quite obsolete. Revision of fatigue provisions is already being done with active support of IIT/
Roorkee. The new provisions are to be based on stress range concept using the Palmgren- Miner
cumulative damage rule. It has been observed that the provisions of British Code BS-5400 are based
on stress range concept and are well understood by practicing design consultants. Therefore, the fatigue

186

assessment criterion has been framed based on BS-5400. Accordingly, the fatigue assessment to be
done in accordance with Palmgren-Miner summation rule by assessing accumulated damage as described
below.
5.2 From the train configurations given in fatigue load model (Table-2) stress histories shall be determined at
the structural detail (including such secondary effects as would be relevant) and stress-ranges and their
numbers shall be evaluated by rain flow method or reservoir method. Fatigue damage analysis shall be
done by applying appropriate partial safety factors to stress ranges and characteristic S-N curves in
accordance with relevant provisions of BS:5400.
5.3 The damage summation shall be performed as per Clause 8.4, 9.2 and 11.1 of BS: 5400 Part-10 as
under:

Dd

n
N

Where,
n

is the number of cycles associated with stress range modified with appropriate partial safety factor
for load

is the number of cycles corresponding to the design S-N curve modified with appropriate partial
safety factor for material.

5.4 Failure shall assumed when Dd > 1.0

5.5 Partial safety factor for load, gfl & partial safety factor for material strength, gm shall be taken as per
BS:5400 Part-1 & Part-3.
5.6 Following parameters shall be considered in assessment of fatigue damage:
5.6.1 Fatigue load Model and traffic classification as per Table-2. Bridges shall be designed for Medium
traffic for annual GMT of 100.
5.6.2 S-N curves for Direct stress range as per BS: 5400 Part -10.
5.6.3 Joint detail classification as per Table-17 of BS: 5400 Part -10. The design shall be in conformity with
the description and requirements of the connection detail chosen.
5.6.4 Appropriate Correction factors for stress concentration as given in Appendix-H of BS: 5400 Part-10
for the detailed classification shall be applied.
5.6.5 Design life of 120 years shall be considered.
5.6.6 Maximum design speed 125 kmph shall be considered.
6.0 CONCLUSIONS
6.1 A rational approach, in accordance with latest international practice has been suggested for design of
new steel girder bridges for heavy axle load operations. The approach follows stress-range concept
187

and Palmgren-Miner cumulative damage rule, which forms the basis of fatigue provisions of British
Standards and Euro Codes. The approach can also be applied for fatigue assessment of existing
bridges for heavy axle load operations.
6.2 A fatigue load model has been developed keeping in view the requirements of heavy haul on Dedicated
Freight Corridor, which is to come in near future. The load model takes into account the parameters of
locomotives and wagons that are necessary inputs required for design of steel girder bridges.
6.3 Fatigue assessment criteria has been proposed in accordance with British Standard BS: 5400, keeping
in view the fact that the provisions are well understood and practiced by leading design consultants in
India. It is expected that a rational fatigue assessment procedure would be followed in developing the
future designs of steel bridges on Indian Railways.
7.0 REFERENCES
7.1

Revision of Fatigue Provisions in IRS Steel Bridge Code (2004), First Interim Project Report submitted by Department
of Earthquake Engineering, Indian Institute of Technology, Roorkee to Research Designs Standards Organisation,
Ministry of Railways, Lucknow (UP)-226011.

7.2

Stress Spectra for Fatigue Design of Railway Bridges (1991), Project report submitted by Department of Civil
Engineering, Indian Institute of Technology, Kanpur to Research Designs Standards Organisation, Ministry of
Railways, Lucknow (UP)-226011.

7.3

Technical Documents on Traffic Details for Revision of Fatigue Provisions of IRS Steel Bridge Code (1989), file
No.CBS/PSB, Research Designs Standards Organisation, Ministry of Railways, Lucknow (UP)-226011.

7.4

IRS Bridge Rules (1986), Research Designs & Standards Organisation, Ministry of Railways, Lucknow (U.P.).

7.5

British Standards BS 5400 : Steel, Concrete and Composite Bridges, Part-1 General Statement, British Standards
Institution, 1980.

7.6

British Standards BS 5400: Steel, Concrete and Composite Bridges, Part-3 Code of Practice for Steel Bridges, British
Standards Institution, 1980.

7.7

British Standards BS 5400 : Steel, Concrete and Composite Bridges, Part-10 Code of Practice for Fatigue, British
Standards Institution, 1980.

7.8

EN 1990:2002 (Eurocode - Basis of Structural Design) - (For safety, comfort, deformation including twist and deflection)

7.9

EN 1991-2:2003 (Eurocode I - Action on Structures, Part 2 - Traffic Loads on Bridges) - (Natural frequency range and
Loading for fatigue estimation)

7.10 EN 1992-1:2004 (Eurocode 2 - Design of Concrete Structures, Part - I - General Rules and Rules for Buildings)
7.11 EN 1992-1-1:2004 (Eurocode 3 - Design of Steel Structures, Part I - 1 - General Rules) - (Classification of cross
sections)
7.12 EN 1993-1-8:2002 (Eurocode 3 - Design of Steel Structures, Part 1-8 - Design of Joints) - (Classification of HSFG Bolts)
7.13 EN 1993-1-9:2002 (Eurocode 3 - Design of Steel Structures, Part 1-9 - Fatigue Strength of Steel Structures)
7.14 EN 1993-2:2004 (Eurocode 3 - Design of Steel Structures, Part 2 - Steel Bridges) - (Requirements for fatigue assessment,
Road and Rail Bridges)
7.15 EN 1994-2:2003 (Eurocode 4 - Design of Composite Steel and Concrete Structures, Part 2 - Rules for Bridges) - (Width
of effective flange, shear connectors)

x x x

188

FATIGUE ASSESSMENT OF CONCRETE BRIDGES EURO NORMS

Ravindra Kumar Goel*
(Published in IRICEN Journal of Civil Engineering, Dec., 2008)

1.0 INTRODUCTION
Throughout the life of a bridge, constant road or rail traffic loading will produce large numbers of
repetitive loading cycles in bridge elements. Both steel (reinforcing and pre-stressing) and concrete components
which are subjected to large numbers of repetitive loading cycles can become susceptible to fatigue damage.
As a consequence, fatigue assessment is required to be undertaken for structures and structural components
which are subjected to regular load cycles. The design provisions are explicitly given in Euro Code (EC 2-2).
The relevant features of this code are described in this paper for the general awareness of bridge designers of
Indian Railways.
2.0 COMPONENTS NOT REQUIRING FATIGUE ASSESSMENT
Some exceptions, where fatigue verification is generally not required are as under:
l

Footbridges, except those components very sensitive to wind action. The most common cause of
wind-induced fatigue is vortex shedding. EN 1991-1-4 covers wind-induced fatigue.

Buried arch and frame structures with a minimum earth cover of 1.0 m (road bridges) or 1.5 m
(railway bridges). This assumes a certain amount of arching of the soil, which suggests that span
should also be relevant.

Foundations.

Piers and columns not rigidly connected to bridge superstructures. "Rigid" in this context is intended
to refer to moment connection as pinned connections will not usually lead to cycles of significant
live load stress range.

Retaining walls of embankments for roads and railways.

Abutments which are not rigidly connected to bridge superstructures (with the exception of the
slabs of hollow abutments)

Prestressing and reinforcing steel in regions where, under the frequent combination of actions only
compressive stresses occur at the extreme concrete fibres. This is because the strain and hence
stress range in the steel is typically small while the concrete remains in compression.

* Director/Bridges & Structures Directorate/RDSO/Lucknow.

189

3.0 INTERNAL FORCES AND STRESSES FOR FATIGUE VERIFICATION

3.1 Euro Norm 2-1-1/6.8.2(1)P requires stresses to be calculated assuming cracked concrete sections,
neglecting the tensile strength of the concrete. Shear lag should be taken into account where relevant
(2-1-1/5.3.2.1 refers). 2-1-1/6.8.2(2)P additionally requires the effect of different bond behaviour of
prestressing and reinforcing steel to be taken into account in the calculation of reinforcement stress. This
results in an increase in stress in the reinforcing steel from that calculated using a cracked elastic crosssection analysis by a factor, ?, given in 2-1-1/(6.64).
3.2 Euro Norm 2-1-1/6.8.2(3) requires fatigue verification to be undertaken for the design of shear
reinforcement, which is a new check for UK practice. Steel forces are calculated from the truss analogy
using a compressive strut angle of qfat. For fatigue calculation, it is important to use a realistic estimate of
the stress range. It is therefore appropriate that this angle is taken greater than that assumed for the
ultimate limit state design (within the angular limits of 2-1-1/6.2.3(2)), since the latter is the angle at the
ultimate limit state after a certain amount of plastic redistribution has taken place to reduce the stress in
the links and to use them optimally. As a result, qfat may be taken as:

tan

fat

tan

1.0

2-1-1/(6.65)

where q is the angle of concrete compression struts to the beam axis assumed in the ultimate limit state
shear design. For shear reinforcement inclined at an angle ? to the horizontal, the steel force can be determined
by re-arranging 2-1-1/(6.13) thus:

V s
Asw z cot fat cot sin

where Dn is the shear force range.

4.0 COMBINATION OF ACTIONS
The calculation of the stress ranges for fatigue verification to EC2 requires the applied load to be divided
into non-cyclic and fatigue-inducing cyclic action effects. The basic combination of the non-cyclic load is
defined by expressions (6.66) and (6.67) of EN 1992-1-1 and is equivalent to the definition of the frequent
combination for the serviceability limit state. The cyclic action is then combined with the unfavourable noncyclic action to determine the stress ranges - 2-1-1/(6.68) and (6.69) refer.
The non-cyclic action gives a mean stress level upon which the cyclic part of the action effect is
superimposed as illustrated in Figure-1 for reinforcement. Mean stress is important as it determines whether
the sign of the stress in an element reverses in the course of a cycle of loading. In Figure-1, the reinforcement
stress range for a given cyclic action is less for the smaller tensile mean stress as part of the cyclic loading then
causes compression in the concrete, which reduces the stress in the reinforcement for that part of the cycle.

190

Bar
stress

Stress range from cyclic loads

(e.g. passage of fatigue load model)

Tension

Reduced stress range

for same vehicle

Mean stress level

from non-cyclic loads

Compression

Time

Figure 1 - Stress ranges for reinforcement fatigue verification caused by same

cyclic action at different mean stress levels

5.0 VERIFICATION PROCEDURE FOR REINFORCING AND PRESTRESSING STEEL

5.1 The number of cycles to fatigue failure of a steel component is a function of the stress that each loading
cycle induces in the component and the type of component. Since the relationship of stress range (DsR)
to the number of cycles to failure (N) is exponential, the relationship is normally plotted graphically in the
form of a log DsR - log N curve. These types of curve are commonly referred to as S-N curves.
5.2 Euro Norm 2-1-1/6.8.4(1) allows the damage produced by cycles of a single stress range of amplitude
DsR to be determined by using the corresponding S-N curves for reinforcing and pre-stressing steel.
The form of these curves is illustrated in Figure-2 for reinforcement; the diagram for prestressing steel is
similar, using 0.1% proof stress in place of yield stress. Recommended values defining the appropriate
S-N curve geometry for the steel component under consideration are given in 2-1-1/Tables 6.3N and
6.4N for reinforcement and pre-stressing steel respectively. The recommended parameters therein may
be modified in the National Annex.

log
Rsk
b = k1
log fyk

b = k2
1

Figure -2 - Characteristic fatigue strength curve (S-N curve) for reinforcing steel

5.3 Euro Norm 2-1-1/6.8.4(1) and also 2-1-1/2.4.2.3(1) require a partial factor, gF,fat, to be applied to all
fatigue loads when calculating the stress range. The value for gF,fat is defined in the National Annex and
is recommended by EC2 to be taken as 1.0. The resisting stress range at N* cycles, DsRsk, given in 21-1/Tables 6.3N and 6.4N, also has to be divided by the material partial safety factor gS,fat. The
recommended value for gS,fat from 2-1-1/2.4.2.4(1) is 1.15.
191

5.4 In real fatigue assessment situations for concrete bridge design, there will be more than one stress range
acting on the steel element throughout its design life. 2 1 1/6.8.4(2) allows multiple amplitudes to be
treated by using a linear cumulative damage calculation, known as the Palmgren-Miner summation:
D Ed
i

n
N

1.0

2-1-1/(6.70)

where:
n(Dsi)

is the applied number of cycles for a stress range of Dsi

N(Dsi)

is the resisting number of cycles for a stress range of Dsi, i.e. the number of loading cycles to
fatigue failure

5.5 For most bridges, the above is a complex calculation because the stress in each component usually
varies due to the random passage of vehicles from a spectrum. Details on a road or rail bridge could be
assessed using the above procedure if the loading regime is known at design. This includes the weight
and number of every type of vehicle that will use each lane or track of the bridge throughout its design
life, and the correlation between loading in each lane or track. In the majority of cases this would
require lengthy calculations.
5.6 As an alternative to the use of 2-1-1/(6.70), 2-1-1/6.8.5 allows the use of simplified fatigue Load
Models 3 and 71, from EN 1991-2, for road and rail bridges respectively, in order to reduce the
complexity of the fatigue assessment calculation. It is assumed that the fictitious vehicle/train alone
causes the fatigue damage. The calculated stress from the vehicle is then adjusted by factors to give a
single stress range which, for N* cycles, causes the same damage as the actual traffic during the bridges
lifetime. This is called the "damage equivalent stress" and is discussed in section below.
5.7 Euro Norm 2-1-1/6.8.4(3) requires that, where prestressing or reinforcing steel is exposed to fatigue
loads, the calculated stresses shall not exceed the design yield strength of the steel as EC2 does not
cover cyclic plasticity.
5.8 Euro Norm 2-1-1/6.8.4(5) relates to assessment of existing structures, which is strictly outside the
scope of EC2, so its inclusion is curious. Its reference to corrosion is not explicit about either the degree
of corrosion or its nature (e.g. general or pitting) so a single value of stress exponent to cover all
situations is dubious. Nevertheless, it was not intended that any such allowance for corrosion be made
in new design.
2-2/6.8.4(107) permits no fatigue check to be conducted for external and unbonded tendons lying
within the depth of the concrete section. This is because the strain, and hence stress, variation under
service loads is small in such tendons. Consideration should be given to fatigue in external tendons
which are outside the depth of the structure (such as in extradosed bridges) as the fluctuation in stress
might be more significant here. This situation is covered by EN 1993-1-11.

192

6.0 VERIFICATION USING DAMAGE EQUIVALENT STRESS RANGE

In the damage equivalent stress range method described by 2-1-1/6.8.5(1) and 2 1 1/6.8.5(2), the real
operational loading is represented by N* cycles of an equivalent single amplitude stress range, Dss,equ(N*)
which causes the same damage as the actual traffic during the bridge's lifetime. This stress range may be
calculated for reinforcing or prestressing steel using 2-2/Annex NN. 2-1-1/6.8.5(3) contains a verification
formula for reinforcing steel, prestressing steel and splicing devices:

F , fat

*
s ,equ ( N )

Rsk

( N* )

2-1-1/(6.71)

s , fat

where:
Dss,equ(N*) is the appropriate damage equivalent stress range (converted to N* cycles) from 2-2/
Annex NN.
DsRsk(N*) is the resisting stress range limit at N* cycles from the appropriate S-N curves given in 2-11/Tables 6.3N or 6.4N
6.2 Euro Norm 2-1-1/(6.71) does not cover concrete fatigue verification. 2-2/Annex NN3.2 provides a
damage equivalent verification for concrete in railway bridges but there is no similar verification for
highway bridges. For highway bridges, concrete can be verified using the methods in 2-2/6.8.7.
7.0 OTHER VERIFICATION METHODS
7.1 Euro Norm 2-1-1/6.8.6(1) and (2) give alternative rules for fatigue verifications of reinforcing and
prestressing steel components. These methods are intended as an alternative to checking fatigue resistance
using 2-1-1/6.8.4 or 6.8.5.
7.2 Euro Norm 2-1-1/6.8.6(1) allows the fatigue performance of reinforcement or prestressing steel to be
deemed satisfactory if the stress range under the frequent cyclic load combined with the basic combination
is less than k1 for unwelded reinforcement or k2 for welded reinforcement. The values of k1 and k2 may
be given in the National Annex and EC2 recommends taking values of 70 MPa and 35 MPa respectively.
The meaning of "frequent cyclic loading" is not explained but it implies a calculation based on the fatigue
load models in EN 1991-2. Assuming this to be the case, it will usually be preferable to perform a
damage equivalent stress calculation using 2-2/Annex NN as this also uses the fatigue load models of
EN 1991 2 and will lead to a more economic answer.
7.3 Euro Norm 2-1-1/6.8.6(2) allows the stress range alternatively to be calculated directly from the frequent
load combination to avoid the need to calculate stress ranges from fatigue load models or directly from
traffic data. However, the recommended allowable stress ranges above would mean that elements
would rarely pass such a check.
7.4 Where welded joints or splicing devices are used in prestressed concrete construction, 2-1-1/6.8.6(3)
requires that no tension exists in the concrete section within 200 mm of the prestressing tendons or
193

reinforcing steel under the frequent load combination when a reduction factor of k3 is applied to the
mean value of the prestressing force. The value of k3 is defined in the National Annex. EC2 recommends
taking a value of 0.9. This value is increased to 1.0 in the UK's National Annex (in the same manner as
rsup and rinf in 2-1-1/5.10.9) to limit the number of load cases to be considered for SLS and fatigue
design. The criterion ensures that the stress range for such details is kept small for the majority of cycles
since the concrete will generally remain in compression.
7.5 VERIFICATION OF CONCRETE UNDER COMPRESSION OR SHEAR
The general fatigue verification procedure for concrete given in 2-2/6.8.7(101) requires a cumulative
damage summation, like that in 2-1-1/6.8.4, to be carried out using traffic data. For rail bridges, this lengthy
calculation can be avoided by using the simplified damage equivalent stress verification of 2-2/Annex NN.3.2.
Neither the Annex nor 2-2/6.8.7(101) itself, however, gives appropriate data for road bridges.
7.6 As a simpler alternative, 2-1-1/6.8.7(2) gives a conservative verification based on the non-cyclic loading
used for the static design:
c , max

f cd , fat

0.5 0.45

c , min

2-1-1/(6.77)

f cd , fat

but limited to 0.9 for fck < 50 MPa or 0.8 for fck > 50 MPa, where:
sc,max is the maximum compressive stress at a fibre under the frequent load combination (compression
measured as positive);
sc,min is the minimum compressive stress under the frequent load combination at the same fibre where
sc,max occurs. sc,min should be taken as 0 if negative (in tension);
fcd,fat is the concrete design fatigue compressive strength defined in the code as:

f cd , fat

k1

cc

t 0 f cd 1

f ck
250

2-2/(6.76)

Where
k1

is a coefficient defined in the National Annex and is recommended by EC2 to be taken as 0.85;

bcc(t0) is the coefficient for concrete strength at first cyclic loading from 2 1 1/3.1.2(6);
t0

is the age of the concrete in days upon first cyclic loading i.e. the age at which live load is first
applied;

fcd

is the design compressive strength of concrete. A value for acc of 1.0 is intended to be used here
in conjunction with k1 = 0.85, as k1 performs a similar function of accounting for sustained
loading;
For concrete road bridges, this alternative concrete fatigue verification is unlikely to govern design,
194

other than possibly for very short spans where the majority of the concrete stress is produced by live load. It
will therefore generally be appropriate to use this simplified check. No guidance is given on the calculation of
the concrete stresses; ignoring concrete in tension will be a conservative assumption.
7.7 Euro Norm 2-1-1/6.8.7(3) permits the above simplified verification of concrete to be applied to the
compression struts of members subjected to shear and requiring shear reinforcement. Since the
compression struts have transverse tension passing through them (see discussions in section 6.5 of this
guide), fcd,fat has to be reduced by the factor v, defined in 2-1-1/6.2.2(6) and the verification becomes:
c ,max

f cd , fat

c ,min

0.5 0.45

7.8 The stresses

f cd , fat

c ,max

and

c ,min

can be calculated for reinforced concrete beams, with shear

reinforcement inclined at an angle ? to the horizontal, from the following expression obtained by rearranging 2-1-1/(6.14):

VEd 1 cot 2
bw z cot
cot

VEd is the relevant shear force under the frequent load combination and the other symbols are defined in
2-1-1/6.2.3. The concrete stress increases with reducing strut angle q so in this case it is conservative
to base q on its ULS value in the above calculation rather than the larger angle qfat from 2-1-1/(6.65).
7.9 For members subjected to shear but not requiring shear reinforcement, 2-1-1/6.8.7(4) provides the
following expressions for assuming satisfactory fatigue resistance in shear:
V Ed ,min
For V
Ed , max

0 :

VEd ,max
VRd ,c

0.5 0.45

VEd ,min
VRd ,c

2-1-1/( 6.78)

but limited to 0.9 for fck < 50 MPa or 0.8 for fck > 50 MPa,
V Ed ,min
or, for V
Ed , max

0 :

VEd ,max
VRd ,c

0.5

VEd ,min
2-1-1/( 6.79)

VRd ,c

where:
VEd,max

is the design value of the maximum applied shear force under the frequent load combination;

VEd,min

is the design value of the minimum applied shear force under the frequent load combination in
the cross-section where VEd,max occurs;

195

VRd,c

is the design shear resistance from 2-2/(6.2.a).

8.0 CONCLUSION
It is seen that Euro norms contain explicit provisions for accounting fatigue effects in concrete structures.
The fatigue assumes special significance in the design of railway bridges, where dynamic impacts are
considerable. Therefore, it is necessary to design the new concrete bridges on Indian Railways taking into
due consideration, the fatigue effects. Since existing IRS Concrete Bridge Code does not contain such
provisions, use of Euro norms may be considered.
9.0 REFERNCES
Proceedings of Workshop on Euro Codes, 11th to 13th Sept., 2008, organised by Indian Concrete Institute (New Delhi
Centre).\
Euro Code EN 1992-2-2

x x x

196

DETERMINATION OF FATIGUE LOAD PARAMETERS

USING ARTIFICIAL NEURAL NETWORKS
Ravindra Kumar Goel*
(Published in Indian Railway Technical Bulletin, R.D.S.O. Lucknow, Feb., 2009)

1.0 INTRODUCTION
An artificial neural network (ANN or just NN) is a computational structure modeled loosely on biological
processes. Neural Network (NN) explore many competing hypotheses simultaneously using a massively
parallel network. These networks are composed of non-linear relatively computational elements interconnected
by links with variable weights. It is this interconnected set of weights that contains the knowledge generated
by the NN. Neural networks are considered as universal function approximators which can map any nonlinear
function. Because of being flexible function approximators, they are powerful methods for pattern recognition,
classification, and forecasting. They are also being explored for decision support and knowledge induction.
This potential of the ANN's has been exploited for the determination of Fatigue Load Parameters (l1) for the
MBG and HM loadings on the basis of parameters values and load model data available in Euro Codes.
2.0 BACKGROUND OF ARTIFICIAL NEURAL NETWORKS
In general, NN models are specified by network topology, node characteristics, and training or learning
rules. NNs are composed of a large number of simple processing units, each interacting with others via
excitatory or inhibitory connections. Learning is achieved through a rule that adapts connection weights in
response to input patterns. Alterations in the weights associated with the connections permits adaptability to
new situations. A feed-forward neural network consists of:
l

Layers

Network input layer

Hidden layers (A, B)

Output layer (C)

No. of units per layer (neuron) which has:

Connections
Weights
Bias
Activation function

* Director/Bridges & Structures Directorate/RDSO/Lucknow.

197

Fig. 1 - A feed forward neural network

The input p is transmitted through a connection that multiplies its strength by the weight w, to form the
product wp. The neuron has a scalar bias, b which is added to the product wp. Here the input, n (wp+b) is
the argument of the transfer function (activation function) f, which produces the output a. In general sigmoid
function for hidden layer and linear function for output layer are used as activation function.

Fig. 2 - Structure of a unit in neural network

3.0 FUNCTIONING OF NEURAL NETWORKS

Fig. 3 - Functioning of neural networks

198

4.0 ADOPTED NEURAL NETS AND OUTPUTS

4.1 Software package
NeuroIntelligence and NeuroSolutions software packages were used for this study. According to
NeuroIntelligence manual, "NeuroIntelligence is neural network software designed to assist experts in solving
real-world problems. NeuroIntelligence supports all stages of neural network design and application. It can
be used to:
1) Analyze and preprocess datasets,
2) Find the best neural network architecture,
3) Train, test and optimize neural networks,
4) Apply the network to solve the problem.
NeuroIntelligence is optimized to solve forecasting, classification and function approximation problems."
However, both packages are widely used depending on the computational environment and the finances that
can associated with analytical efforts.
4.2 Steps in designing the Neural Networks model
4.2.1 Variable selection
Success in designing a neural network depends
on a clear understanding of the problem. It is
important to know which input variables are important
that can affect the output (l1). It is necessary because
of the powerful ability of neural networks to detect
complex nonlinear relationships among a number of
different variables. The decomposition of the problem
into a form amenable to solution is a formidable task
requiring an understanding of the intricate relationships
between the variables involved and the variables which
contribute effectively towards the solution sought. The
Fig. 4 - ANN Predicted partial safety factors for Train 1
determination of l1 for the Indian trains consisting of
MBG and HM was initiated from the data available for the EU trains as embodied in the Eurocodes. In order
to facilitate computation the data as contained in the Eurocodes as well as the data available for the trains of
the Indian Railways had to be converted into a coherent database with the same denominators. This was
achieved by reducing the train into its decomposed form of axle weights and spacings between them and also
included the train velocity and loaded length forming the inputs. l1 was the only output variable. Initial tests on
Eurocode data (fig.4) showed immense promise for the application of neural networks for the above purpose.
However, it was also noted that there was a major change of behaviour of l1 when the loaded length was less
than 20 m. necessitating the bifurcation of the problem into two parts.
199

4.2.2 Training, testing, and validation sets

The data sets are divided into three distinct sets called the training, testing, and validation sets. The
training set is the largest set and is used by the neural network to learn the patterns present in the data. The
testing set is used to evaluate the generalization ability of a supposedly trained network. A final check on the
performance of the trained network is made using the validation set.
A number of random rows were selected for training, validation and test sets. This was done so that the
network structure could train/generalize itself to all type of data of different trains. For the current problem,
division was done as follows:
Training set 70%
Validation set 15%
Test set 15%
4.2.3 Data preprocessing
Data preprocessing refers to analyzing and transforming the input and output variables to minimize
noise, highlight important relationships, detect trends, and flatten the distribution of the variable to assist the
neural network in learning the relevant patterns. Data were preprocessed by scaling down them between the
upper and lower bound of the transfer function. In the present case all preprocessing have been done by
taking global maxima & minima (combining training and query data) in the scaling factor.
Numeric values are scaled using the following formula:
l

SF = (SRmax-SRmin)/(Xmax-Xmin)

Xp = SRmin + (X-Xmin)* SF

Where:
X - actual value of a numeric column
Xmin - minimum actual value of the column
Xmax - maximum actual value of the column
SRmin - lower scaling range limit
SRmax - upper scaling range limit
SF - scaling factor
Xp - preprocessed value
Input columns scaling range: [-1, 1] (Linear transfer function)
Output column scaling range: [0, 1] (Logistic or Sigmoid transfer function)

200

4.2.4 Neural network structure

Number of Input neurons

= Number of input variables after preprocessing

= 29 (MBG )
= 37(HM )

Number of Output neurons

= Number of output variables

= 1

Number of hidden layers

= 2

Number of neurons in hidden

layers

= Depends on network architecture

Search parameters:
Search Method: Exhaustive
Fitness criteria: Inverse Training error
Error function: Sum-of-squares
For example: Network structure of MBG train type 2 is [29-14-5-1] i.e. 29 inputs, 2 hidden layers with
14 and 5 neurons respectively, and one output.
4.2.5 Neural network training
Many training algorithm were tried and on the basis of capability of a algorithm to correctly depict the
target trends (values), the best algorithm was selected.
Training algorithm - Online Back Propagation
Learning Rate - 0.6
Learning momentum - 0.7
Number of iterations - 100000
4.2.6 Evaluation criteria
The error function minimized in the neural networks training is the sum of squared errors.
Error = (Target values - Output values)2 /(No. of values)
4.3 Outputs
The outputs from two independent studies were collated to arrive at the partial factors Recommended.
The studies were such that in each case the optimal network architecture was adopted based on algorithms
native to the software itself. A consequence of the above strategy was that the outputs were arrived at by
adopting different network architecture in each case. Further, the order of preference adopted was such that
the MBG trains were considered prior to the HM trains. The core strategy consisted of initiating the analysis

201

with the Euro trains in the training and testing dataset and then querying on the MBG train type 1. The outputs
for train 1 were then folded back into the data set and the network was re-trained with the additional data
available, this procedure had been employed successfully in prediction earlier with remarkable success. In
this manner the outputs for each train was used to augment the training dataset. The HM trains were queried
using the Euro data as well as the complete set of data derived for the MBG trains, and the data was folded
back into the training dataset as mentioned before. The outputs from the two studies conducted are as stated
in Tables 1 to 4. The values recommended in Table 5 & 6 are the enveloping values emerging from the two
independent studies. Graphical representation of typical curves for MBG loading is as shown in Fig..5.

Fig. 5 - Graphical representation of partial factors for MBG loading

5.0 CONCLUSION
Use of Artificial Neural Network Techniques has been utilized successfully in determining the partial
safety factor (loading factor, 'l1') for fatigue assessment of steel bridges subjected to Indian Railway Standard
loadings. The study suggests new areas in the field of analysis and design of bridge engineering where availability
of reliable data in sufficient quantity can be used to established relationships with the design input and output
parameters.

202

Table 1 Neurosolutions: Evaluation of Partial Safety Factors for MBG Loading

Span
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
6.00
7.00
8.00
9.00
10.00
12.50
15.00
17.50
20.00
25.00
30.00
35.00
40.00
45.00
50.00
60.00
70.00
80.00
90.00
100.00

Train-1 Train-2 Train-3 Train-4 Train-5 Train-6 Train-7 Train-8 Train-9

1.3501 1.3408 1.4576 1.4306 1.7804 1.3795
1.09
1.0863 1.3533
1.3611 1.2903 1.4257 1.3975 1.7483 1.2607 1.035 1.0626 1.3142
1.3621 1.176 1.3189 1.4058 1.6915 1.1259 0.9749 1.0407 1.2884
1.2758 1.0296 1.0149 1.4069 1.6324 0.9836 0.9012 1.0197 1.2695
1.1021 0.915 0.8023 1.3881 1.5773 0.8389 0.8134 0.9967 1.2551
1.0572 0.8668 0.7038 1.345 1.5229 0.7014 0.7169 0.9583 1.2433
1.064 0.8285 0.6317 1.2688 1.4679 0.5822 0.6187 0.8731 1.232
1.005 0.7614 0.5558 1.1444 1.4133 0.5145 0.541 0.7463 1.2168
0.9186 0.7427 0.5138 0.9912 1.3448 0.4923 0.5046 0.6802 1.1885
0.9076 0.7975 0.5172 0.8771 1.2018 0.4893 0.4938 0.6849 1.1246
0.9725 0.8568 0.9217 0.8895 0.7969 0.6255 0.6409 0.9526 0.8695
0.9684 0.8535 0.9124 0.8901 0.7938 0.6214 0.6371 0.9496 0.8681
0.9634 0.8496 0.902 0.8906 0.7906 0.6172 0.6334 0.9471 0.8667
0.9575 0.845 0.8904 0.8909 0.7874 0.6127 0.6297 0.9448 0.8654
0.9503 0.8397 0.8772 0.891 0.7841 0.608 0.6262 0.9428 0.8641
0.9257 0.822 0.8368 0.8894 0.7758 0.5954 0.6177 0.9381 0.8613
0.8886 0.7968 0.7839 0.8839 0.7672 0.5815 0.6097 0.9328 0.8589
0.8349 0.7624 0.7185 0.8724 0.7585 0.5666 0.6023 0.9263 0.8567
0.7636 0.7183 0.6451 0.8522 0.7499 0.5509 0.5953 0.918
0.855
0.6001 0.6068 0.5084 0.7751 0.7332 0.5181 0.5822 0.8967 0.8527
0.4935 0.4943 0.4243 0.6496 0.7173 0.4853 0.5699 0.871 0.8509
0.4562 0.4234 0.3874 0.5242 0.6998 0.4541 0.5574 0.8431 0.8473
0.4483 0.3991 0.3754 0.4459 0.6739 0.4255 0.5433 0.8145 0.8374
0.4466 0.397 0.3739 0.409 0.6301 0.4004 0.5262 0.7857 0.8146
0.4442 0.3987 0.3758 0.3935 0.5659 0.3791 0.5049 0.7564 0.7712
0.4355 0.3947 0.3805 0.3852 0.4342 0.3483 0.4513 0.6936 0.6236
0.4264 0.3848 0.3834 0.3832 0.3641 0.3316 0.3994 0.619 0.4836
0.4193 0.3758 0.384 0.3804 0.3323 0.3243 0.365 0.5281 0.4075
0.4145 0.3696 0.3829 0.3765 0.3138 0.3217 0.3454 0.431 0.3726
0.411 0.3657 0.381 0.3722 0.3015 0.3211 0.3334 0.3551 0.3613

203

Train-10
0.8146
0.7738
0.7461
0.7224
0.6922
0.6519
0.6028
0.5532
0.5152
0.4908
0.6086
0.6049
0.6009
0.5965
0.5918
0.5784
0.5625
0.5442
0.5236
0.4788
0.4352
0.3993
0.3739
0.3575
0.3477
0.3386
0.3352
0.3334
0.3321
0.3311

Table 2 Neurosolutions: Evaluation of Partial Safety Factors for HM Loading

Span
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
6.00
7.00
8.00
9.00
10.00
12.50
15.00
17.50
20.00
25.00
30.00
35.00
40.00
45.00
50.00
60.00
70.00
80.00
90.00
100.00

Train-1
1.7915
1.6939
1.6022
1.5160
1.4192
1.2956
1.1516
1.0210
0.9291
0.8768
0.8060
0.7571
0.7049
0.6512
0.5980
0.4801
0.3981
0.3509
0.3276
0.3181
0.3317
0.3515
0.3681
0.3795
0.3866
0.3914
0.3875
0.3789
0.3679
0.3567

Train-2
1.6321
1.4326
1.2695
1.1489
1.0311
0.9063
0.7970
0.7248
0.6910
0.6874
0.9428
0.9117
0.8752
0.8328
0.7846
0.6439
0.5013
0.3942
0.3327
0.2902
0.2851
0.2916
0.3022
0.3114
0.3179
0.3256
0.3283
0.3277
0.3254
0.3228

Train-3
1.5883
1.3656
1.1263
0.9544
0.8252
0.7187
0.6391
0.5896
0.5658
0.5617
0.9429
0.9053
0.8596
0.8051
0.7425
0.5736
0.4416
0.3696
0.3370
0.3201
0.3249
0.3353
0.3456
0.3537
0.3590
0.3634
0.3631
0.3608
0.3578
0.3547

Train-4
1.8991
1.8853
1.8615
1.8299
1.7867
1.7387
1.6969
1.6454
1.5507
1.4251
1.0412
1.0283
1.0127
0.9936
0.9699
0.8874
0.7791
0.6769
0.6077
0.5579
0.5575
0.5627
0.5626
0.5569
0.5476
0.5241
0.4995
0.4792
0.4677
0.4677

Train-5
1.8449
1.8224
1.7911
1.7534
1.7124
1.6727
1.6423
1.6265
1.5406
1.4174
0.9994
0.9866
0.9705
0.9504
0.9260
0.8438
0.7393
0.6374
0.5612
0.4923
0.4845
0.4972
0.5129
0.5245
0.5289
0.5158
0.4899
0.4739
0.4704
0.4663

Train-6
1.9569
1.9457
1.9275
1.9047
1.8802
1.8569
1.8347
1.7962
1.7072
1.5693
1.1716
1.1676
1.1629
1.1575
1.1515
1.1347
1.1170
1.1000
1.0840
1.0509
0.9972
0.8629
0.5897
0.3801
0.3075
0.2760
0.2726
0.2773
0.2897
0.3136

Train-7
1.8654
1.8587
1.8347
1.7878
1.7281
1.6637
1.6033
1.5482
1.4969
1.4489
0.9122
0.8838
0.8544
0.8247
0.7956
0.7302
0.6811
0.6492
0.6303
0.6087
0.5833
0.5454
0.4958
0.4300
0.3487
0.2501
0.2362
0.2365
0.2418
0.2531

Train-8
1.9703
1.9689
1.9657
1.9598
1.9492
1.9319
1.9039
1.8584
1.7980
1.7396
1.1553
1.1507
1.1452
1.1387
1.1310
1.1064
1.0742
1.0363
0.9953
0.9122
0.8296
0.7317
0.5723
0.3711
0.2733
0.2392
0.2369
0.2407
0.2496
0.2659

Train-9
1.9710
1.9698
1.9671
1.9620
1.9529
1.9378
1.9120
1.8675
1.8066
1.7521
1.1669
1.1638
1.1600
1.1558
1.1505
1.1346
1.1148
1.0926
1.0695
1.0232
0.9725
0.8978
0.7369
0.4678
0.3155
0.2610
0.2584
0.2668
0.2849
0.3170

Train-10
1.9719
1.9709
1.9688
1.9650
1.9588
1.9494
1.9342
1.9071
1.8560
1.7771
1.2046
1.2039
1.2032
1.2024
1.2017
1.1999
1.1981
1.1963
1.1943
1.1893
1.1829
1.1592
1.1092
0.9456
0.7756
0.5977
0.4813
0.3630
0.2994
0.2822

Train-11
1.9034
1.9045
1.9005
1.8973
1.8941
1.8731
1.7976
1.6941
1.6289
1.5830
1.0110
1.0040
0.9961
0.9884
0.9804
0.9588
0.9346
0.9072
0.8769
0.8081
0.7308
0.6498
0.5714
0.5021
0.4464
0.3785
0.3547
0.3546
0.3648
0.3755

Train-12
1.8979
1.8959
1.8911
1.8771
1.8239
1.6865
1.5663
1.5020
1.4566
1.4193
1.0680
1.0650
1.0625
1.0593
1.0557
1.0441
1.0265
0.9982
0.9507
0.7458
0.4502
0.3171
0.2808
0.2676
0.2615
0.2588
0.2629
0.2719
0.2849
0.3019

Table 3 NeuroIntelligence: Evaluation of Partial Safety factors for MBG Loading

Span
20.0000
25.0000
30.0000
35.0000
40.0000
45.0000
50.0000
60.0000
70.0000
80.0000
90.0000
100.0000

Train-1
0.6957
0.6782
0.6551
0.6281
0.6002
0.5733
0.5482
0.5040
0.4711
0.4547
0.4509
0.4506

Train-2
0.7630
0.7558
0.7391
0.7095
0.6716
0.6338
0.6012
0.5541
0.5252
0.5038
0.4790
0.4459

Train-3
0.7545
0.7550
0.7552
0.7548
0.7555
0.7450
0.7050
0.5419
0.4799
0.4470
0.4287
0.4227

Train-4
0.4269
0.4113
0.3917
0.4076
0.4085
0.4114
0.4161
0.4284
0.4365
0.4338
0.4233
0.4113

Train-5
0.7078
0.7101
0.7118
0.7130
0.7136
0.7139
0.7137
0.7121
0.7077
0.6974
0.6737
0.6242

Train-6
0.6830
0.6623
0.6431
0.6267
0.6133
0.6179
0.6236
0.6310
0.6284
0.6195
0.6116
0.6023

Train-7
0.7383
0.7382
0.7382
0.7383
0.7388
0.7395
0.7404
0.7425
0.7436
0.7424
0.7375
0.7267

Train-8
0.7539
0.7567
0.7586
0.7598
0.7604
0.7602
0.7592
0.7528
0.7389
0.7226
0.7089
0.6943

Train-9
0.7556
0.7553
0.7546
0.7529
0.7497
0.7436
0.7321
0.6829
0.6133
0.5575
0.5211
0.5064

Train-10
0.6660
0.6511
0.6294
0.6318
0.6316
0.6268
0.6162
0.5758
0.5185
0.4657
0.4298
0.4081

Table 4 NeuroIntelligence: Evaluation of Partial Safety factors for HM Loading

Span
20.00
25.00
30.00
35.00
40.00
45.00
50.00
60.00
70.00
80.00
90.00
100.00

Train-1
0.3954
0.4008
0.4067
0.4125
0.4175
0.4209
0.4222
0.4179
0.4083
0.3999
0.3948
0.3920

Train-2
0.5649
0.5681
0.5686
0.5659
0.5595
0.5494
0.5359
0.5021
0.4667
0.4371
0.4156
0.4007

Train-3
0.4744
0.4735
0.4723
0.4702
0.4669
0.4615
0.4530
0.4235
0.3897
0.3719
0.3650
0.3622

Train-4
0.5769
0.5760
0.5752
0.5747
0.5745
0.5747
0.5752
0.5777
0.5821
0.5882
0.5950
0.6006

Train-5
0.5812
0.5839
0.5870
0.5904
0.5943
0.5988
0.6038
0.6157
0.6298
0.6453
0.6603
0.6724

Train-6
0.4647
0.4612
0.4601
0.4617
0.4662
0.4736
0.4826
0.4946
0.4851
0.4647
0.4471
0.4319

Train-7
0.4189
0.4149
0.4099
0.4038
0.3963
0.3875
0.3780
0.3623
0.3552
0.3531
0.3525
0.3523

Train-8
0.5060
0.5059
0.5059
0.5058
0.5056
0.5052
0.5045
0.5009
0.4919
0.4746
0.4505
0.4271

Train-9
0.6057
0.6039
0.6016
0.5988
0.5951
0.5904
0.5845
0.5683
0.5452
0.5150
0.4810
0.4487

x x x
204

Train-10
0.5082
0.5126
0.5185
0.5262
0.5361
0.5484
0.5632
0.5991
0.6367
0.6665
0.6839
0.6906

Train-11
0.4940
0.4863
0.4796
0.4735
0.4683
0.4639
0.4604
0.4563
0.4557
0.4577
0.4609
0.4626

Train-12
0.4556
0.4532
0.4513
0.4497
0.4484
0.4474
0.4466
0.4455
0.4448
0.4443
0.4441
0.4438

IMPACT OF INCREASING AXLE LOAD ON FATIGUE

LIFE OF STANDARD STEEL GIRDER BRIDGES - A
STUDY BASED ON REVISED FATIGUE PROVISIONS
R.K. Goel* & H.O. Narayan**
(Presented in IPWE Seminar at IRICEN, Jan, 2010)
SYNOPSIS
Increasing the axle load of railway wagons has been a matter of concern for railway engineers as it's effect on residual
fatigue life of steel girder bridges was not clearly understood. There had been no accepted provisions or guidelines to
assess the fatigue life of steel girder bridges till RDSO discussed the "Draft Provisions for Fatigue Assessment" in 78th
BSC. The simplified approach of these provisions has been used to assess the design fatigue life of standard plate girder
bridges for MBG loading and new 25 t Loading -2008 for different average annual GMTs of the routes. It has been found that
the reduction in design fatigue life depends on the bridge span. The design fatigue life gets improved substantially, if the
coefficient of dynamic impact (CDA) is controlled either by imposing speed restriction or by improving track maintenance
on girder bridges.

1.0 INTRODUCTION
1.1 Indian Railway has recently permitted over utilization of carrying capacity of it's wagons for generating
extra revenue. This has increased axle loads of wagons. The magnitude of trailing load (TLD) has been
increased from Carrying Capacity (CC) to CC+6+2 and subsequently to CC+8+2. A new loading
standard 25 t loading -2008 has also been included in existing IRS Bridge Rules and this loading has
been permitted on existing bridges. RDSO has also issued guidelines for running 25 t loading - 2008 on
existing bridges. It is to be noted that most of the railway bridges have been designed for Indian Railways
standard loadings such as BGML, RBG & MBG which are having varying equivalent uniformly distributed
loads. The net effect of increased trailing load is to increase the number of cycles of maximum stress
range to which the members are subjected and therefore greater fatigue damage is caused therein. The
ultimate effect of running heavy axle load is thus, to reduce the residual fatigue life of the existing bridges.
1.2 The issue of revision of fatigue provisions in IRS steel bridge code had been under discussion for quite
some time. Detailed studies have been carried out by RDSO on the issue and draft provisions have
been issued for adoption. These provisions have also been discussed in 78th Bridge Standard Committee
and as per Railway Board Orders the workability and suitability of these provisions is being ascertained
before final adoption. In this connection a study on assessment of fatigue life of existing standard steel
girders has been done by RDSO based on the simplified approach of the draft revised provisions. This
* Director/Steel Bridges - I, Bridges & Structures Directorate, Research Designs & Standards Organization, Manak Nagar, Lucknow
(U.P.)-226001
** Asstt. Design Engineer, Bridges & Structures Directorate, Research Designs & Standards Organization, Manak Nagar, Lucknow
(U.P.)-226001

205

paper presents the approach used and the results of study, showing the extent to which the fatigue life of
standard steel girders has been affected by running 25 t axle load vis--vis MBG loading.
2.0 SIMPLIFIED APPROACH OF REVISED IRS PROVISIONS
2.1 Fatigue stress spectra
2.1.1For the simplified fatigue loading the following procedure is adopted to determine the design stress
spectrum.
2.1.2The recommended equivalents for train loads shall be adopted in accordance with existing provisions of
IRS Bridge Rules, including the dynamic impact factor f, which is calculated as (1.0 + CDA), where
CDA is the coefficient of Dynamic Augment as specified in IRS Bridge Rules.
2.1.3The maximum stress sp,max and the minimum stress sp, min should be determined for a detail or structural
connection. The sp, max is for Dead Load + Full Live Load with dynamic impact factor ' ' whereas the
P,min is for dead load effects only.
2.1.4The reference stress range DsP for determining the damage due to the stress spectrum should be obtained
from:
P , max

P ,min

2.1.5The damage effects of the stress range spectrum may be represented by the damage equivalent stress
range related to 2 million cycles as
*

E ,2

Where,
l

is the damage equivalence factor, and

is the dynamic impact factor (1.0 + CDA),

CDA

is the coefficient of Dynamic Augment as specified in Bridge Rules

2.1.6The value of damage equivalence factor, 'l ' are specific to the type of loading and the values are
dependent on loaded length, 'L' which is defined in para 2.3.3.
2.2 Fatigue assessment
The fatigue assessment shall be carried out by ensuring the satisfaction of the following criteria:
Ff *

E,2

Mf

Where,
DsC is the reference value of the fatigue strength at NC = 2 million cycles
gMf is the partial safety factor for material
206

gFf is the partial safety factor for loads

2.3 Damage equivalence factors
2.3.1The damage equivalent factor for railway bridges should be determined from:
l = l1 *l2 *l3 *l4 subject to the condition that l < max
where
l1 is a factor that takes into account the damaging effect of traffic and depends on the base length of
the longest loop of the influence line diagram
l2 is a factor that takes into account the annual traffic volume in million tonnes
l3 is a factor that takes into account the design life of the bridge in years
l4 is a factor to be taken into account when the bridge structure is loaded on more than one track
lmax is the maximum value taking into account the fatigue limit and is equal to 1.4
2.3.2

The loaded length for the determination of the appropriate l1 should be taken as follows:
(a)

(b)

(c)

for moments:
-

For a simply supported span, the span length, L

For cross girders supporting rail bearers (or stiffeners), the sum of the spans of the
rail bearers (or stiffeners) carried by the cross girder.

for shear for a simply supported span

-

For the support section, the span length

For the mid-span section, 0.4 * the span under consideration

In other cases
-

(d)

the same as for moments

for truss members

-

base length of the largest loop of Influence line diagram

2.3.3The value of l2 , in terms of the annual volume of traffic may be obtained from the following expression:
2

0.5193 * Ta

0.2036

Where Ta is the annual volume of traffic expressed in million tonnes.

2.3.4Unless otherwise specified by the competent authority the value of l3 will be taken as 1.04 for a design
life of 120 years. For other values of design life the corresponding value may be calculated from the
following expression where LD is the design life in years.

207

0.3899 * LD

0.2048

2.3.5The value of l4, assuming 15% of the total traffic on both tracks crosses whilst on the bridge, unless
specified otherwise by the competent authority, shall be obtained from
4

0.7926 * a 2

0.7280 * a 0.9371

where
a=

1+2

Ds1 = Stress range at the section being checked due to train on one track.
Ds1+2 = Stress range at the same section due to train load on two tracks.
The values of l4 may be calculated for other proportions of crossing traffic from
5
4

1 n a5

1 a) 5

where
n is the proportion of traffic that crosses whilst on the bridge.
2.3.6 The value of l should not exceed lmax which is specified as 1.4.
3.0 Determination of l1 Parameters
3.1 The fatigue life assessment has been done for MBG loading and 25t loading as given in IRS Bridge Rule.
The fatigue load model for these loading have been developed and l1 parameters worked out in
association with IIT/Roorkee using Artificial Neural Network Technique.
3.2 The value of l1 is to be obtained from tables 1 & 2 for MBG loading and 25t loading - 2008 respectively
as a function of the loaded length. These values have been worked out as per the train types included in
the respective standard fatigue load models for MBG loading and 25t loading. The loaded length shall
depend upon the influence line diagram of the structural detail/connection under consideration.
4.0 ASSESSMENT METHOD AND ASSUMPTIONS
(i) Maximum stress range taken as the difference of dead load stress and the maximum stress likely to
come in the member with DL, Impact load and live load. The maximum bending stresses due to
equivalent uniformly distributed load for IRS loadings given in Bridge Rules have been worked out
and the maximum stress range calculated.
(ii) In the analysis of plate girder only the bending stresses due to maximum bending moment have
been taken into consideration to find out the maximum stress range and the design fatigue life of
spans has been assessed.
(iii) Occasional loads have not been considered.

208

Table 1: l1 for MBG Loading

Span (m) Train-1 Train-2 Train-3 Train-4 Train-5 Train-6 Train-7 Train-8 Train-9 Train-10
0.50

1.1996

1.3377

1.3342

1.1477

1.3353

1.5986

1.5681

1.3336

1.2672

1.2849

1.00

1.1775

1.3266

1.3111

1.1360

1.3320

1.5893

1.5698

1.3318

1.2590

1.2748

1.50

1.1615

1.3106

1.2820

1.1218

1.3314

1.5756

1.5730

1.3293

1.2552

1.2634

2.00

1.1470

1.2903

1.2674

1.1039

1.3318

1.5650

1.5769

1.3267

1.2537

1.2473

2.50

1.1288

1.2651

1.2559

1.0820

1.3317

1.5617

1.5809

1.3252

1.2523

1.2290

3.00

1.1050

1.2339

1.2460

1.0502

1.3303

1.5618

1.5839

1.3254

1.2502

1.2174

3.50

1.0741

1.1963

1.2367

1.0063

1.3281

1.5614

1.5856

1.3270

1.2474

1.2121

4.00

1.0386

1.1537

1.2270

0.9639

1.3254

1.5620

1.5859

1.3291

1.2442

1.2077

4.50

0.9986

1.1100

1.2157

0.9294

1.3222

1.5627

1.5856

1.3312

1.2406

1.2024

5.00

0.9503

1.0694

1.2014

0.8953

1.3188

1.5636

1.5860

1.3327

1.2367

1.1961

6.00

0.8699

0.9994

1.1606

0.8247

1.3112

1.5663

1.5876

1.3362

1.2286

1.1843

7.00

0.8451

0.9574

1.1341

0.7469

1.3031

1.5590

1.5823

1.3184

1.2216

1.1666

8.00

0.8300

0.9389

1.1075

0.6797

1.2957

1.5299

1.5736

1.3142

1.2173

1.1446

9.00

0.8175

0.9282

1.0628

0.6437

1.2884

1.4952

1.5621

1.3117

1.2146

1.1102

10.00

0.8676

0.9588

1.0191

0.5600

1.1653

1.2961

1.3417

1.1167

1.1238

1.0127

12.50

0.8462

0.9549

0.9896

0.5421

1.1416

1.2469

1.2912

1.1127

1.1205

0.9585

15.00

0.8263

0.9376

0.9505

0.5439

1.1252

1.1911

1.2392

1.1087

1.1114

0.9302

17.50

0.8162

0.9168

0.9056

0.5563

1.0956

1.1457

1.1787

1.1047

1.0814

0.9170

20.00

0.8558

0.9087

0.8477

0.5905

1.1604

1.1808

1.1717

0.9881

1.1445

0.8985

25.00

0.7783

0.7977

0.7779

0.5846

1.1146

1.0770

1.1297

0.9740

1.0955

0.8607

30.00

0.7553

0.7700

0.7163

0.5551

1.0888

1.0312

1.0675

0.9646

1.0616

0.8149

35.00

0.7307

0.7225

0.6843

0.5103

1.0649

0.9918

1.0005

0.9729

1.0455

0.7745

40.00

0.7162

0.7040

0.6632

0.5391

1.0403

0.9096

0.9651

0.9787

1.0377

0.7437

45.00

0.6884

0.6742

0.6497

0.5367

1.0243

0.8560

0.9229

0.9781

0.9996

0.7234

50.00

0.6494

0.6555

0.6411

0.5154

1.0048

0.7993

0.8834

0.9739

0.9546

0.7101

60.00

0.5494

0.6384

0.6346

0.4315

0.9696

0.7129

0.8208

0.9619

0.8765

0.6847

70.00

0.5282

0.5849

0.6367

0.4211

0.8868

0.6692

0.7579

0.9474

0.8288

0.6658

80.00

0.5020

0.5398

0.6156

0.4219

0.8213

0.6506

0.7126

0.9286

0.8037

0.6516

90.00

0.4739

0.5128

0.5324

0.4127

0.7781

0.6456

0.6807

0.9048

0.7785

0.6367

100.00

0.4521

0.4893

0.5039

0.4047

0.7205

0.6416

0.6593

0.8836

0.6921

0.6272

209

Table 2: l1 for 25 T Loading

Span
(m)

TrainTrainTrain- TrainTrain-2 Train-3 Train-4 Train-5

Train-7 Train-8 Train-9
1
6
10
11

0.50

1.2278 1.3574 1.5043 0.6903 1.4369 1.6074 1.3131

1.3236

1.3923 1.3121 1.2930

1.00

1.2042 1.3435 1.4886 0.6762 1.4284 1.6003 1.3160

1.3199

1.3883 1.3031 1.2889

1.50

1.1862 1.3283 1.4708 0.6625 1.4198 1.5970 1.3230

1.3156

1.3842 1.2998 1.2847

2.00

1.1707 1.3124 1.4509 0.6500 1.4114 1.5948 1.3297

1.3106

1.3804 1.2988 1.2801

2.50

1.1556 1.2951 1.4279 0.6383 1.4038 1.5928 1.3301

1.3049

1.3766 1.2982 1.2746

3.00

1.1410 1.2750 1.4001 0.6246 1.3971 1.5910 1.3218

1.2987

1.3712 1.2965 1.2681

3.50

1.1227 1.2503 1.3660 0.6073 1.3915 1.5895 1.3132

1.2922

1.3639 1.2932 1.2605

4.00

1.0910 1.2195 1.3274 0.5859 1.3865 1.5889 1.3132

1.2858

1.3617 1.2884 1.2515

4.50

1.0393 1.1835 1.2924 0.5631 1.3811 1.5906 1.3154

1.2802

1.3644 1.2829 1.2412

5.00

0.9765 1.1453 1.2656 0.5443 1.3724 1.5930 1.3182

1.2761

1.3674 1.2778 1.2296

6.00

0.8968 1.0749 1.2128 0.5239 1.3494 1.5961 1.3216

1.2728

1.3610 1.2736 1.2039

7.00

0.8796 1.0144 1.1493 0.5148 1.3371 1.5981 1.3215

1.2705

1.3307 1.2669 1.1742

8.00

0.8692 0.9717 1.0911 0.5069 1.3073 1.5971 1.3204

1.2680

1.2870 1.2586 1.1416

9.00

0.8626 0.9473 1.0537 0.4997 1.2605 1.5930 1.3186

1.2665

1.2515 1.2494 1.1163

10.00 0.9047 0.9682 1.0109 0.5456 1.0825 1.3503 1.1921

1.1708

1.0788 1.1092 0.9719

12.50 0.8925 0.9620 0.9572 0.5315 1.0711 1.3002 1.2022

1.1743

1.0851 1.0845 0.9463

15.00 0.8806 0.9198 0.9468 0.5220 1.0534 1.2432 1.1971

1.1877

1.0566 1.0317 0.9249

17.50 0.8274 0.8937 0.9165 0.5157 1.0076 1.2011 1.1833

1.1917

0.9964 0.9946 0.8925

20.00 0.8262 0.8995 0.8962 0.5253 0.9897 1.2124 1.1808

1.1590

0.9845 0.9422 0.8102

25.00 0.8065 0.8712 0.8906 0.5104 0.9690 1.1650 1.1806

1.1591

0.9911 0.9302 0.7540

30.00 0.7899 0.7774 0.8768 0.5091 0.9401 1.1004 1.1797

1.1588

0.9804 0.9241 0.7012

35.00 0.7706 0.7643 0.8657 0.5055 0.9058 1.0415 1.1781

1.1577

0.9472 0.8938 0.6568

40.00 0.7554 0.7290 0.7835 0.5019 0.8702 0.9786 1.1759

1.1546

0.9056 0.8550 0.6224

45.00 0.7270 0.7037 0.7470 0.4955 0.8395 0.9766 1.1733

1.1497

0.8677 0.8156 0.6101

50.00 0.6863 0.6814 0.7148 0.4877 0.8171 0.9733 1.1706

1.1444

0.8364 0.7789 0.6778

60.00 0.5832 0.6552 0.6815 0.4813 0.8197 0.9241 1.1637

1.1318

0.7877 0.7392 0.6610

70.00 0.5486 0.6046 0.6601 0.4823 0.8529 0.8500 1.1552

1.1164

0.8249 0.7086 0.6352

80.00 0.5090 0.5570 0.6458 0.4861 0.7976 0.7703 1.1484

1.0992

0.7609 0.6729 0.6087

90.00 0.4738 0.5367 0.5849 0.4999 0.7276 0.7245 1.1400

1.0743

0.6793 0.6579 0.5898

100.00 0.4532 0.5173 0.5554 0.5113 0.6520 0.6921 1.1286

1.0313

0.6131 0.6460 0.5724

210

(iv) For plate girders, the loaded length for considering l1 has been taken, as effective span length.
(v) GMT factor is taken as per proposed fatigue criteria in 78th BSC.
(vi) Fatigue categories are chosen as per the Table - 3 which is based on the tables given in draft
provisions, discussed in 78th BSC and the engineering judgment applied. The analyzed fatigue life
may vary on this account.
Table - 3 Fatigue categories adopted for member detail/connections
S. Member detail or connection to be
No. assessed

Stringer, X-girder & Plate girders

(welded type)
Stringer, X-girder & Plate girders
(rivetted type)
Fillet weld of web-flange connection of
stringer/x-girder
Gusset connections

Stringer and x-girder connection

1
2

Fatigue category
Category
Reference
Details 5 & 6 of
100
Table 9.2
Detail 8 of Table
80
9.1
Detail 8 of Table
80
9.5
Detail 8 of Table
80
9.1
Detail 11 of Table
100
9.1

Remark

Bending stresses at mid

of span
Shear stress at throat
area of weld.
Axial stresses on net
area.
Shear stress on shank
area of rivet/bolt

5.0 FATIGUE LIFE ASSESSED

5.1 Based on above assumptions, the design calculations for assessment of fatigue life of plate girder bridges
are given in Table - 4 for MBG loading and in Table - 5 for 25t loading.
5.2 It is to be noted that the annual GMT for a particular route may not be comprised of the trains causing
maximum stress range. Situation may vary route wise and partial GMT may be due to trains causing
much lower stress range. In such cases a more detailed fatigue assessment would be required taking
into consideration the representative load model of actual set of trains running on the route.
5.3 It is further to be observed that due to change of loading, the 1values and the maximum design stressranges have increased marginally. However, the effect on design fatigue life has been considerable on
higher spans as compared to smaller spans.
Table - 4 Assessed fatigue life of standard plate girder bridges (MBG Loading)
Std. Span
12.2m MBG
12.2m MBG
18.3m MBG
18.3m MBG
24.4m MBG
24.4m MBG
12.2m BGML
18.3m BGML
24.4m BGML

RDSO Drg. No.

B-16009 (4 million)
B-16012 (10 million)
B-16010 (4 million)
B-16013 (10 million)
B-16011 (4 million)
B-16005 (10 million)
B-11003 (2million)
B-11004 (2 million)
B-11005 (2 million)

Stress Loaded Loadin Design Life (years) for GMT & corresponding average route
GMT factor (Lamda2)
Range Length
g
Rmax
L'
factor,
5
10
20
30
40
50
(N/mm2) (m) Lamda1 0.721 0.830
0.956
1.038
1.101
1.152
118.33
13.1
1.28 64.802 32.533 16.333 10.914
8.200
6.568
95.75
13.1
1.28 182.219 91.480 45.926 30.690 23.057
18.469
109.35
19.4
1.18 141.727 71.152 35.721 23.870 17.933
14.365
84.22
19.4
1.18 507.201 254.633 127.834 85.426 64.177
51.409
93.02
25.6
1.125 394.140 197.872 99.339 66.383 49.871
39.949
84.15
25.6
1.125 642.927 322.772 162.043 108.285 81.351
65.166
112.88
13.1
1.28 27.440 13.776 6.916
4.622
3.472
2.781
111.86
19.4
1.18 42.671 21.422 10.755
7.187
5.399
4.325
107.2
25.6
1.125 66.311 33.290 16.713 11.168
8.390
6.721

211

Table - 5 Assessed fatigue life of standard plate girder bridges (25 t Loading)
Std. Span
RDSO Drg. No.

Loaded
Design Life (years) for GMT & corresponding
Stress Length
average route GMT factor (Lamda2)
Loading
Range
L'
factor,
Rmax
(m) Lamda1
5
10
20
30
40
(N/mm2)
0.721
0.830
0.956 1.038
1.101

50
1.152

12.2m MBG

B-16009 (4 million)

118.31

13.1

1.29

62.437

31.346

15.737 10.516

7.900

6.329

12.2m MBG

B-16012 (10 million)

95.73

13.1

1.29

175.603

88.159

44.259 29.576

22.220

17.799

18.3m MBG

B-16010 (4 million)

114.75

19.4

1.21

101.106

50.759

25.483 17.029

12.793

10.248

18.3m MBG

B-16013 (10 million)

88.38

19.4

1.21

361.811 181.642

91.191 60.938

45.781

36.673

24.4m MBG

B-16011 (4 million)

98.3

25.6

1.18

238.428 119.699

60.093 40.157

30.169

24.167

24.4m MBG

B-16005 10 million)

88.92

25.6

1.18

389.066 195.325

98.060 65.529

49.229

39.435

12.2m BGML B-11003 (2million)

112.86

13.1

1.29

26.440

13.274

6.664

4.453

3.345

2.680

18.3m BGML B-11004 (2 million)

117.38

19.4

1.21

30.446

15.285

7.674

5.128

3.852

3.086

24.4m BGML B-11005 (2 million)

113.28

25.6

1.18

40.122

20.143

10.112

6.758

5.077

4.067

6.0 EFFECT OF SPEED ON THE ASSESSED DESIGN FATIGUE LIFE

6.1 The effect of speed on coefficient of dynamic augment (CDA) has been given in 3.3.1 (a) of Bridge
Rules. The relevant clause is reproduced as under:
"CDA laid down in Clauses 2.4.1 and 2.4.2 (diminished) according to Clauses 3.3 where applicable
may be multiplied by the factor (Vr/V) where Vr is the permissible speed and V is(i) 125 km/h for trains hauled by diesel and electrical locomotives and 80 km/h for steam locomotives
on BG.
(ii) 100 km/h for trains hauled by diesel and electric locomotives and 60 km/h for steam locomotives
on MG."
6.2 It is evident from the perusal of Table 4 & 5 that the assessed fatigue life is quite low for high GMT
routes and the spans designed for 2 million cycles as per old fatigue provisions based on stress ratio
concept. This is apparently due to the fact that the stress ranges considered in analysis are based on full
CDA (Coefficient of Dynamic Augment) which is applicable for a speed 125 kmph for goods trains.
Practically, goods trains do not run with a speed more than 100 kmph. Therefore, the design fatigue life
has been re-assessed with reduced stress ranges corresponding to a sectional speed of 100 kmph by
proportionately reducing the CDA as per clause 3.3.1 (a) of Bridge Rules. The results for re-assessed
design life for MBG loading and 25t loading are shown in table 6 & 7 respectively.
6.3 It is further noted that the new 25t loading has been permitted with a sectional speed of 75 kmph due to
strength considerations. Accordingly, the fatigue life of bridges with new 25t loading at 75 kmph has
been again worked out with reduced CDA values corresponding to 75 kmph and the results are given
in Table - 8. The assessed fatigue life of different standard spans for 50 GMT at varying speeds for 25t
loading are given in Table - 9 for comparison.
212

7.0 LIMITATION OF THE RESULTS OBTAINED

The results of analysis may be interpreted keeping in view the following:
i)

Assessment is based on average GMT of the route over the life span of the bridge. The GMT being
carried in present may be actually different.

ii)

The GMT may be comprised of no. of trains which may not give the maximum stress range.

iii)

Partial factor of safety is assumed as 1.0 for analysis of existing bridges.

v)

It is assumed that the physical condition of the bridge is otherwise sound from other considerations and
effect of corrosion, pitting and other defects developed during service have not been accounted for.
Table - 6 Assessed fatigue life for different GMTs (MBG Loading) with sectional
speed of 100 kmph
Std. Span

12.2m MBG
12.2m MBG
18.3m MBG
18.3m MBG
24.4m MBG
24.4m MBG
12.2m BGML
18.3m BGML
24.4m BGML

RDSO Drg. No.

B-16009 (4 million)
B-16012 (10 million)
B-16010 (4 million)
B-16013 (10 million)
B-16011 (4 million)
B-16005 (10 million)
B-11003 (2million)
B-11004 (2 million)
B-11005 (2 million)

Design Life (years) for GMT & corresponding average route

Stress
Loaded
Loading
GMT factor (Lamda2)
Range
Length
factor,
Rmax
L'
5
10
20
30
40
50
Lamda1
(N/
(m)
0.721
0.830
0.956
1.038
1.101
1.152
2
mm )
109.8
13.1
1.28
93.378 46.879 23.535 15.727
11.815 9.465
88.8
13.1
1.28
263.261 132.166 66.352 44.340
33.311 26.684
102.5
19.4
1.18
194.372 97.582 48.989 32.737
24.594 19.701
78.9
19.4
1.18
697.510 350.174 175.800 117.479 88.258 70.698
87.7
25.6
1.125 525.457 263.798 132.436 88.500
66.487 53.259
79.31
25.6
1.125 858.574 431.034 216.394 144.606 108.637 87.024
104.7
13.1
1.28
39.619 19.890
9.986
6.673
5.013
4.016
104.8
19.4
1.18
58.666 29.452 14.786
9.881
7.423
5.946
101.1
25.6
1.125
88.272 44.316 22.248 14.867
11.169 8.947

Table - 7 Assessed fatigue life for different GMTs (25 t Loading) with sectional
speed of 100 kmph
Std. Span

12.2m MBG
12.2m MBG
18.3m MBG
18.3m MBG
24.4m MBG
24.4m MBG
12.2m BGML
18.3m BGML
24.4m BGML

RDSO Drg. No.

B-16009 (4 million)
B-16012 (10 million)
B-16010 (4 million)
B-16013 (10 million)
B-16011 (4 million)
B-16005 (10 million)
B-11003 (2million)
B-11004 (2 million)
B-11005 (2 million)

Stress Loaded
Loading
Range Length
factor,
Rmax
L'
Lamda1
2
(N/mm ) (m)
109.7
88.8
107.5
82.8
92.7
83.8
104.7
109.9
106.8

13.1
13.1
19.4
19.4
25.6
25.6
13.1
19.4
25.6

1.28
1.28
1.18
1.18
1.125
1.125
1.28
1.18
1.125

213

Design Life (years) for GMT & corresponding

average route GMT factor (Lamda2)
5
0.721
90.297
253.445
139.052
497.486
317.497
519.735
38.142
41.991
53.493

10
0.830
45.332
127.238
69.809
249.755
159.395
260.925
19.149
21.081
26.855

20
0.956
22.758
63.878
35.046
125.386
80.022
130.993
9.613
10.583
13.482

30
1.038
15.208
42.687
23.420
83.789
53.475
87.537
6.424
7.072
9.010

40
1.101
11.425
32.069
17.595
62.948
40.174
65.763
4.826
5.313
6.769

50
1.152
9.152
25.689
14.094
50.424
32.181
52.679
3.866
4.256
5.422

Table - 8 Assessed fatigue life for diff erent GMTs

CDA with 75 Kmph

GMAmf=

GMAfF=

Fatcat=
Std. Span

RDSO Drg. No.

Stress Range
Loaded
Rmax (N/mm2) Length L'
(m)

Loading
factor,
Lamda1

100

100

Design Life (years) for GMT & corresponding average

(Lamda2)
5

10

20

30

0.721

0.830

0.956

1.038

1.

214

12.2m MBG

B-16009 (4 million)

101.15

13.1

1.28

134.198

67.372

33.823

22.602

16

12.2m MBG

B-16012 (10 million)

81.84

13.1

1.28

377.543

189.540

95.156

63.588

47

18.3m MBG

B-16010 (4 million)

100.2

19.4

1.18

196.021

98.409

49.405

33.015

24

18.3m MBG

B-16013 (10 million)

77.2

19.4

1.18

700.297

351.573

176.502

117.948

88

24.4m MBG

B-16011 (4 million)

87

25.6

1.125

432.826

217.294

109.089

72.899

54

24.4m MBG

B-16005 (10 million)

78.7

25.6

1.125

706.208

354.541

177.992

118.944

89

12.2m BGML B-11003 (2million)

96.5

13.1

1.28

56.800

28.516

14.316

9.567

7.

18.3m BGML B-11004 (2 million)

102.5

19.4

1.18

59.018

29.629

14.875

9.940

7.

24.4m BGML B-11005 (2 million)

100.3

25.6

1.125

72.686

36.491

18.320

12.242

9.

Table - 9 Assessed fatigue life of std. plate girder bridges for 50 GMT at different speeds (25 t Loading)
Std. Span

RDSO Drg. No.

12.2m MBG
12.2m MBG
18.3m MBG
18.3m MBG
24.4m MBG
24.4m MBG
12.2m BGML
18.3m BGML
24.4m BGML

B-16009 (4 million)
B-16012 (10 million)
B-16010 (4 million)
B-16013 (10 million)
B-16011 (4 million)
B-16005 (10 million)
B-11003 (2million)
B-11004 (2 million)
B-11005 (2 million)

Stress Range Rmax (N/mm2) for

CDA at different speeds
CDA100
CDA75
CDA125
118.31
109.7
101.15
95.73
88.8
81.84
114.75
107.5
100.20
88.38
82.8
77.20
98.3
92.7
87.00
88.92
83.8
78.70
112.86
104.7
96.50
117.38
109.9
102.50
113.28
106.8
100.30

Design Life (years) for 50 GMT at d

speeds
125 kmph
100 kmph
6.3
9.2
17.8
25.7
10.2
14.1
36.7
50.4
24.2
32.2
39.4
52.7
2.7
3.9
3.1
4.3
4.1
5.4

8.0 CONCLUSIONS
8.1 The design fatigue life of standard plate girders is found reduced as a result of running of 25t loading 2008 vis--vis existing MBG loading.
8.2 Reduction in design fatigue life is nominal (3-4%) for smaller span (12.2m) and significantly high (40%)
for higher span (24.4m).
8.3 Route GMT and sectional speed are important parameters which affect the design fatigue life considerably.
The Effect of reducing CDA is to improve the design fatigue life by reducing the maximum design stress
range and the reduction is substantial.
8.4 Plate girders designed for 10 million cycles as per existing fatigue criteria in IRS Steel Bridge Code give
better estimate of fatigue life as compared to those designed for 4 million cycles.
8.5 Plate girders designed for 2 million cycles as per existing fatigue criteria in IRS Steel Bridge Code gives
very low estimate of fatigue life and the same need to be verified again using detailed fatigue life analysis
with respect to train loads, sectional speeds and actual GMTs on the routes where these girders are
provided.
Note : Similar Analysis of open web girder of 30.5m span designed for BGML loading (Drg. No. BA-11122) has also been done
subsequently for MBG loading & 25t loading for varying speeds. The analysis has also been done for open web girder for 30.5m designed
for 25t loading (Drg. No. B-11679). The results are given in Annexure-I (a to g). The analysis needs to be revised based on revised l1
parameters or any change in fatigue category proposed to be adopted.

215

Annexure-I (a)
Analysis of Open Web Girders at different Speeds with reduced CDAs
OWG assessed With revised Lambda

With full CD
GMAmf=

gussets

30.5m span, Design for BGML loading, BA-11122

80

Analysed for 25 t Loading

80

Member

Stress Range
Rmax (N/mm2)

Loaded Length L'

(m)

Loading factor,
Lamda1

Design Life (years) for GMT & correspondin

GMT factor (Lamda2)
5

10

20

30

0.721

0.830

0.956

1.038

216

L0-L1

79.84

31.926

1.03

430.078

215.914

108.396

72.436

L1-L2

79.84

31.926

1.03

430.078

215.914

108.396

72.436

L2-L3

93.89

31.926

1.03

194.896

97.845

49.121

32.825

L0-U1

62.06

31.926

1.03

1471.530

738.760

370.883

247.843

U1-L2

121.8

25.54

1.07

45.406

22.795

11.444

7.648

L2-U3

110.7

19.16

1.07

72.402

36.348

18.248

12.194

U1-U2

82.54

31.926

1.03

365.611

183.550

92.148

61.578

U2-U3

82.54

31.926

1.03

365.611

183.550

92.148

61.578

U1-L1& U3-L3

85.1

10.642

1.33

90.404

45.386

22.785

15.226

Stringer

94.11

5.321

1.37

47.854

24.024

12.061

8.060

X-girder

97.22

10.642

1.33

47.183

23.687

11.892

7.947

Annexure-I (b)

OWG assessed With revised Lambda

With CDA at
GMAmf=

gussets

30.5m span, Design for BGML loading, BA-11122

80

Analysed for 25 t Loading

80

Member

Stress Range Rmax

(N/mm2)

Loaded Length L'

(m)

Loading factor,
Lamda1

Design Life (years) for GMT & corresponding a

factor (Lamda2)
5

10

20

30

0.721

0.830

0.956

1.038

217

L0-L1

75.61

31.926

1.03

561.026

281.654

141.400

94.491

70

L1-L2

75.61

31.926

1.03

561.026

281.654

141.400

94.491

70

L2-L3

88.91

31.926

1.03

254.315

127.675

64.097

42.833

32

L0-U1

58.77

31.926

1.03

1919.895

963.854

483.888

323.360

24

U1-L2

110.37

25.54

1.07

73.465

36.882

18.516

12.373

9.2

L2-U3

103.46

19.16

1.07

100.736

50.573

25.389

16.966

12

U1-U2

78.16

31.926

1.03

477.138

239.540

120.257

80.362

60

U2-U3

78.16

31.926

1.03

477.138

239.540

120.257

80.362

60

U1-L1& U3-L3

70.53

10.642

1.33

226.156

113.538

57.000

38.090

28

Stringer

86.22

5.321

1.37

73.388

36.843

18.497

12.360

9.2

X-girder

90.21

10.642

1.33

68.003

34.140

17.139

11.453

8.6

Annexure-I (c)
OWG assessed With revised Lambda

With CDA a
GMAmf=

gussets

30.5m span, Design for BGML loading, BA-11122

80

Analysed for 25 t Loading

80

Member

Stress Range Rmax

(N/mm2)

Loaded Length L'

(m)

Design Life (years) for GMT & corresponding a

factor (Lamda2)
Loading factor,
Lamda1
5
10
20
30
0.721

0.830

0.956

1.038

218

L0-L1

71.37

31.926

1.03

743.641

373.334

187.427

125.248

94

L1-L2

71.37

31.926

1.03

743.641

373.334

187.427

125.248

94

L2-L3

83.93

31.926

1.03

336.981

169.176

84.932

56.756

42

L0-U1

55.48

31.926

1.03

2201.999

1105.480

554.990

370.873

27

U1-L2

103.64

25.54

1.07

99.884

50.145

25.175

16.823

12

L2-U3

96.92

19.16

1.07

138.565

69.565

34.924

23.338

17

U1-U2

73.78

31.926

1.03

632.323

317.448

159.370

106.499

80

U2-U3

73.78

31.926

1.03

632.323

317.448

159.370

106.499

80

U1-L1& U3-L3

64.61

10.642

1.33

346.989

174.201

87.455

58.442

43

Stringer

78.33

5.321

1.37

117.257

58.867

29.553

19.749

14

X-girder

83.19

10.642

1.33

101.000

50.706

25.456

17.011

12

Annexure-I (d)
OWG assessed With revised Lambda

With full CDA

GMAmf=

gussets

30.5m span, Design for BGML loading, BA-11122

80

Analysed for MBG Loading

80

Member

Stress Range Rmax

(N/mm2)

Loaded Length L'

(m)

Loading factor,
Lamda1

Design Life (years) for GMT & corresponding a

factor (Lamda2)
5

10

20

30

0.721

0.830

0.956

1.038

219

L0-L1

75.03

31.926

0.96

821.411

412.377

207.028

138.347

10

L1-L2

75.03

31.926

0.96

821.411

412.377

207.028

138.347

10

L2-L3

88.31

31.926

0.96

370.663

186.086

93.422

62.429

46

L0-U1

58.38

31.926

0.96

2796.692

1404.037

704.875

471.035

35

U1-L2

116.3

25.54

1.08

54.372

27.297

13.704

9.158

6.8

L2-U3

107.7

19.16

1.13

63.433

31.846

15.988

10.684

8.0

U1-U2

77.59

31.926

0.96

697.291

350.064

175.744

117.442

88

U2-U3

77.59

31.926

0.96

697.291

350.064

175.744

117.442

88

U1-L1 & U3-L3

85.1

10.642

1.27

113.260

56.860

28.546

19.076

14

Stringer

83.15

5.321

1.34

97.589

48.993

24.596

16.437

12

X-girder

94.89

10.642

1.27

66.556

33.414

16.775

11.210

8.4

Annexure-I (e)
OWG assessed With revised Lambda

With full CDA

GMAmf=

gussets

30.5m span, Design for BGML loading, BA-11122

80

Analysed for MBG Loading

80

Member

Stress Range Rmax Loaded Length L'

(N/mm2)
(m)

Loading factor,
Lamda1

Design Life (years) for GMT & corresponding a

factor (Lamda2)
5

10

20

30

0.721

0.830

0.956

1.038

220

L0-L1

71.05

31.926

0.96

1071.850

538.106

270.148

180.527

13

L1-L2

71.05

31.926

0.96

1071.850

538.106

270.148

180.527

13

L2-L3

83.63

31.926

0.96

483.675

242.821

121.905

81.463

61

L0-U1

55.28

31.926

0.96

3105.026

1558.831

782.587

522.966

39

U1-L2

109.61

25.54

1.08

72.608

36.452

18.300

12.229

9.

L2-U3

101.02

19.16

1.13

86.713

43.533

21.855

14.605

10

U1-U2

73.47

31.926

0.96

909.887

456.795

229.327

153.248

11

U2-U3

73.47

31.926

0.96

909.887

456.795

229.327

153.248

11

U1-L1& U3-L3

78.53

10.642

1.27

167.708

84.195

42.269

28.246

21

Stringer

76.18

5.321

1.34

149.704

75.157

37.731

25.214

18

X-girder

88.32

10.642

1.27

94.452

47.418

23.806

15.908

11

Annexure-I (f)
OWG assessed With revised Lambda

With full CD
GMAmf=

30.5m span, Design for BGML loading, BA-11122

Analysed for MBG Loading

Stress Range Rmax
(N/mm2)

Loaded Length L'

(m)

Loading factor,
Lamda1

L0-L1

67.07

31.926

0.96

L1-L2

67.07

31.926

L2-L3

78.94

L0-U1

Member

gussets
80
80

Design Life (years) for GMT & corresponding av

factor (Lamda2)
5
0.721

221

10
0.830

20
0.956

30
1.038

1420.272

713.026

357.964

239.210

0.96

1420.272

713.026

357.964

239.210

31.926

0.96

640.901

321.754

161.532

107.944

52.19

31.926

0.96

3105.026

1558.831

782.587

522.966

U1-L2
L2-U3

102.92

25.54

1.08

98.743

49.573

24.887

16.631

94.14

19.16

1.13

122.362

61.430

30.840

20.609

U1-U2

69.36

31.926

0.96

1205.660

605.283

303.873

203.064

U2-U3

31.926

0.96

1205.660

605.283

303.873

203.064

U1-L1& U3-L3

69.36
71.94

10.642

1.27

Stringer

69.20

5.321

1.34

257.182
239.291

129.114
120.133

64.820
60.311

43.316
40.303

3
3

X-girder

81.46

10.642

1.27

140.238

70.404

35.345

23.620

Annexure-I (g)

with lambda values for 25 t loading

GMAmf=
Fatcat=
Fatcat/GMAmf=

30.5 m span, Design for 25 t loading, B-11679

Analysed for 25 t Loading
Member

Stress Range
2
Rmax (N/mm )

Loaded Length L'

(m)

Loading factor,
Lamda1

1
80 gussets
80

Design Life (years) for GMT & corresponding

GMT factor (Lamda2)
5
0.721

10
0.830

20
0.956

30
1.038

222

L0-L1

55.99

31.926

1.03

2201.999

1105.480

554.990

370.873

27

L1-L2

55.99

31.926

1.03

2201.999

1105.480

554.990

370.873

27

L2-L3

73.77

31.926

1.03

632.741

317.658

159.475

106.570

80

L0-U1

65.05

31.926

1.03

1169.463

587.111

294.750

196.968

14

U1-L2
L2-U3

72.86
81.252

25.54
19.16

1.07

558.154

280.213

140.677

94.007

70

1.07

327.776

164.555

82.612

55.206

41

U1-U2

82.41

31.926

1.03

368.436

184.968

92.860

62.054

46

U2-U3
U1-L1& U3-L3

82.41
74.82

31.926
10.642

1.03
1.33

368.436

184.968

92.860

62.054

46

Stringer

79.49

5.321

1.37

169.509
324.480

85.100
162.900

42.723
81.781

28.550
54.651

21
41

X-girder

85.52

10.642

1.33

262.435

131.752

66.144

44.201

33

x x x

LOAD RATING

INCREASE IN DESIGN LONGITUDINAL FORCES

OVER A CENTURY DUE TO HEAVY HAUL
AND IMPLICATIONS ON EXISTING BRIDGES
Ravindra Kumar Goel*
(To be published in Journal of IPWE, April 2010)

1.0 INTRODUCTION
The construction of railways in India started more than one hundred and fifty years ago when less
powerful and lighter locomotives were used and the bridges carried lighter loads. The earlier Bridge Rules
did not make any provisions for tractive and braking forces and these forces were generally not considered
while designing the bridges. A reference to the tractive and braking forces was first made in the Bridge Rules
of 1923 by which time many of the main lines had already been constructed. The provision for longitudinal
forces was revised first in 1926 and again in 1933. The rules were further modified in 1941 which are
continuing till date. This paper gives systematic development of provision for longitudinal forces since
beginning.
2.0 THE EARLIEST PROVISIONS
Prior to the formulation of the Bridge Rules, under statutory obligations, in 1892, bridges were built to
comply with the British Board of Trade Rules. These rules specified the permissible stresses for wrought iron
and steel but not the longitudinal loads. After 1892, bridges were built according to the standards laid down
in the Bridge Rules (which were revised first in 1903) and again in 1923, 1926, 1933, 1941 and 1964).
Reference to the tractive and braking force first appeared in the 1923 Rules. Accordingly to these rules the
force specified under the relevant clauses of the British Standard specifications for girder bridges were required
to be used.
3.0 BRIDGE RULES, 1926
In the 1926 Rules, the provision for longitudinal forces was made in clauses 30 & 31 which are reproduced
below:
Clause 30 and 31 of 1926 Bridge Rules for the longitudinal forces
"30 (a) Longitudinal forces :Where a structure carried a railway provision shall be made for the stresses
due to the tractive effort of the live load and the braking effect resulting from the application of the brakes to
such load while passing thereover, these forces being considered as acting at rail level.

Director/Steel Bridges - I, Bridges & Structures Directorate, Research Designs & Standards Organization, Manak Nagar, Lucknow
(U.P.)-226001

225

31 (a)

For railway worked by steam or electric locomotives, the amount of the tractive effort on one track
shall be ascertained by multiplying one and three-quarter times the maximum end shear due to the
live load on that track by a factor equal to

20
L 75
Also limited to a maximum of 0.15
Where L = the span in feet. The factor shall not exceed 0.15 as a maximum.
The braking effect shall be similarly determined using a factor equal to

12
0.75
L 90
(b) In the case of line worked solely on the electrical multiple unit system, the amount of the tractive effort on
one track shall be ascertained by multiplying the sum of the actual wheel loads on the span by a factor
equal to

3
0.10
L 10
Where L = the span in feet. The factor shall not exceed 0.20 as a maximum.
The braking effect shall be similarly determined by using a factor of 0.20 for all spans."
4.0 MODIFICATION TO ALLOW DISPERSION, 1933
In the 1933 Rules the provisions for longitudinal forces were further modified as a special clause was
introduced to allow for the dispersion of the horizontal forces through the track at the approaches of the
bridge on the basis of the experiments carried with steam locomotives on bridges. The relevant clause is
reproduced below:
"22. Longitudinal forces : Where a structure carried a railway, provision as under shall be made for the
stresses due to the tractive effort of the live load and to the braking effect resulting from the application of the
brakes to such load while passing thereover.
These forces shall be considered as acting horizontally through the girder seat where girders have sliding
bearings or through the knuckle pin.
For spans supported on sliding bearings, the horizontal forces shall be considered as being divided
equally between the two ends; for spans which have roller bearings at one end, the whole of the horizontal
forces shall be considered to act through the fixed end.
(a) Tractive effort : For railways worked by steam or electric locomotive the amount of tractive force on

226

one track shall be ascertained by multiplying the EUDL on one track taken from the Table of loads for
calculating bending moments by one of the following factors :
For broad gauge spans ,

32
L 90

Where L = the length of the bridge in feet subject to a maximum of 120 ft. for HM loading and 100 ft.
for ML and BL loading.
For bridges or spans exceeding 120 ft. in length for HM loading and 100 ft. for ML and BL loadings the
tractive effort shall be assumed to be constant.
For metre gauge spans,

20
L 55

Where L = the length of the bridge in feet subject to maximum of 80 feet for all loadings.
For bridges or spans exceeding 80 ft. in length the tractive effort shall be assumed to be constant.
(b) Braking effect : For railways worked by steam or electric locomotives the braking effect on one track
shall be ascertained by multiplying the EUDL on the track taken from the Table of loads for calculating
bending moments, by one of the following factors:
For broad gauge spans,

32
145 L

0.03

For metre gauge spans,

35
165 L

0.01

Where L = the length in feet of the bridge or the length of the train whichever is less.
(c) In the case of lines worked solely on the electrical multiple unit system, the following allowance shall be
made:
(i) Tractive effort : 25 percent of the sum of the driving axles on the bridge.
(ii) Braking effort : 20 percent of the sum of all axles on the bridge.
In all cases the amounts calculated under heads (a), (b) & (c) shall be reduced to allow for dispersion of
the horizontal forces through the track at the ends of the bridge by amounts as given in the following table:
Standard

Allowance for dispersion of horizontal forces through the track

Broad gauge
BL

7.5 tons

ML

9.0 tons

HM

11.5 tons
227

Metre gauge
BL

5.0 tons

ML

6.0 tons

HM

7.5 tons

In the case of bridges up to 80 feet long, the above allowances may be increased by 50 percent.
In all cases the net horizontal forces after deducting dispersion shall be assumed to be distributed equally
amongst the spans in the length L as defined."
5.0 RELAXATIONS FOR EXISTING BRIDGES, 1941
The provisions for longitudinal forces in Bridge Rules of 1933 were considered to be too severe for
existing bridges and a special clause was added relaxing the forces for examining them in Bridge Rules of
1941. The relevant clause is reproduced below "3.4 For the purpose of calculating, the longitudinal loads on existing bridges, clause 2.8 shall be held to
apply generally excepting that the maximum tractive load may be assumed to be 25 percent of the axle loads
of the coupled wheels of actual engines under considerations, and the braking load to be 20% of the actual
braked engine axle loads plus 10 percent of the other braked axles running or proposed. The amount of the
longitudinal load so calculated may be reduced, where no rail expansion joints exist, by an allowance for
dispersion in tonnes per track, equal to the following percentage of the weight of the rails in kg per m.
(a) For loaded length, L, equal to or exceeding 30m (100 ft.) - 20 percent.
(b) For loaded length, L, less than 30m (100 ft.) - 20% increasing uniformly to 30% for L=0
Note : In order to obtain the allowance for dispersions in tons, when the weight of the rails is
expressed in lb per yard, the percentages given above should be halved."
In the Bridge Rules of 1941, the longitudinal forces were calculated for different spans using the formulae
given in 1993 Rules and were given in an appendix and the formulae themselves were not given. Further, the
provisions for the additional dispersion of force to the approaches for spans upto 80 ft. length was dispensed
with vide correction slips.
6.0 CORRECTION SLIPS, 1968
The above provisions were modified in January, 1968 Correction Slip No.4 and tractive effort of different
type of locomotives such as steam locomotive, DC locomotive, Diesel or AC locomotives were mentioned.
The modified clause read as under:
"3.4 Delete the existing clause 3.4 as amended by A&C Slip No.2 dated 09-06-67 and substitute as
under:
"3.4 For the purpose of calculating the longitudinal loads on existing bridges, clause 2.8 shall be held to

228

apply generally. For trains hauled by steam or DC locomotives, the maximum tractive force may be assumed
to be 25% of the axle load of the coupled wheels on actual engines under consideration and the braking load
to be 20% of the actual braked engine axle load plus 10% of the other braked axle loads. For trains hauled
by Diesel or AC Electric locomotives, the tractive force shall be as specified for the locomotive distributed
equally amongst the driving axles. Braking force for such locomotives shall be as specified for them distributed
equally amongst the braked axles together with 10% of the weight of the braked trailing axles covering the
loaded length. However, the maximum horizontal force on any pier or abutment would be as specified in
Appendices VII & VIII so long as not more than two locomotives are coupled together".
7.0 REVISION OF LONGITUDINAL FORCES, 1975
In 1975, the Table of longitudinal force (inclusive of dispersion) as appearing in Bridge Rules 1964,
were revised while introducing the Revised BG standard of loading. These values of longitudinal forces are
based on the enveloping values of live loads, tractive effort and braking force to cover the following conditions
of loading:
(i)

Double headed operation exerting 75t tractive effort using two locomotives - one of 45t tractive effort
and the other of 30t tractive effort.

(ii)

Increased likely tractive effort of 53t to 54t for the BOBO + BOBO type locomotive.

Due to introduction of air-brake system, clause 3.4 of 1964 Bridge Rules incorporating provisions
regarding longitudinal loads on existing bridges was also revised as under:
8.0 CORRECTION SLIPS 1975 & 1993
Clause 3.4 (As amended vide A&C Slip No. 4 dated 30-1-1968) was further amended in 1975 vide
Correction Slip No.10 as under:
"Delete the last two sentences "Braking force for such locomotives not more than two locomotives are
coupled together" and substitute the following:
The braking force for such locomotives shall be as specified for them distributed equally amongst the
braked axles, together with 10% of the weight of the braked trailing axles covering loaded length, if fitted with
vaccum brakes, or 20% of the weight of the braked trailing axles covering the loaded length, if fitted with air
brakes."
However, above provisions were again modified in 1993 vide Correction Slip No. 21 and the braking
force for locomotives fitted with air brakes was limited to 13.4% of the weight of the braked trailing axles
covering the loaded length.
The provision regarding dispersion of longitudinal forces as appearing in 1964 Bridge Rules, was retained.
This provision reads as under:
"3.4.1 The amount of longitudinal load so calculated be reduced where no rail expansion joint exists by
an allowance for dispersion in tones per track equal to the following percentage of the weight of the rails in kg.

229

per metre.
(a) For loaded length, L, equal or exceeding 30m (100 feet) - 20%
(b) For loaded length, L, less than 30m (100 feet) - 20% increasing uniformally to 30% for L = 0.
Note -In order to obtain the allowance for dispersion in ton, when the weight of the rail is expressed in
lb per yard the percentage given above should be halved.
9.0 Correction Slip No. 12, 1980
Clause 2.7.1 - Add the following at the end of the clause :
"For sliding bearing of PTFE/elastomeric type .0.10."
Clause 2.8.1.3 - Delete the existing Cl. 2.8.1.3 and substitute as under:
"2.8.1.3 For spans supported on sliding bearings, the horizontal loads shall be considered as being
distributed between different supports as below:
a)

2 supports directly under the loaded span: each 40% of the horizontal load due to tractive/braking
efforts after deducting force dispersed as stipulated in clauses 2.8.2.4.1 and 2.8.2.4.3.

b) Other two adjacent supports each 20%.

For spans which have roller bearings at one end, the whole of the horizontal load shall be considered to
act through the fixed end."
Clause 2.8.2.4 - Replace existing clause 2.8.2.4 by the following:
"2.8.2.4.1. In case of bridges, provided with through welded rails, rail free fastenings and adequate
anchorage of welded rails on approaches (by providing adequate density of sleepers, ballast cushion and its
consolidation etc., but without any switch expansion joints) dispersion of longitudinal load P through the rails
away from the loaded length may be allowed to the extent of 25% of the value of longitudinal load P as
obtained through Appendices VII, VIII and VIII(a) subject to a minimum of 16t for BG, 12t for MGML and
10t for MGBL. This will also apply to bridges with jointed track with rail free fastenings but without any
switch expansion or mitered joints. Where suitably designed elastomeric bearings are provided the aforesaid
relief may be further increased by 40% thereof."
Introduce a clause 2.8.2.4.2 as below:
"2.8.2.4.2 The total dispersion under clause 2.8.2.4.1 shall not exceed the capacity of the rails for
transferring the longitudinal load to the approaches nor should it exceed the capacity of the anchored length of
the track on the approaches in resisting the longitudinal load."
10. CORRECTION SLIP NO. 23, 1999 AND CORRECTION SLIP NO. 32 OF 2005
In 1999 vide Correction Slip No. 23, Clause 2.8 & 3.4 were revised and further modified vide Correction
Slip No. 32 of 2005. The modified provisions are attached as Annexure-I. These provisions are being
followed as on date for working out design longitudinal forces on railway bridges.
230

11. IMPLICATIONS OF INCREASED LONGITUDINAL FORCES

11.2 In the light of above history, it is evident that no consideration of longitudinal forces was made in the
design of bridges before 1923. It does not necessarily mean that the bridge engineers were not aware
about the implications of tractive effort and braking force on the design of bearings and substructures.
Actually the loads carried and the speeds were so small that the longitudinal forces whatsoever generated,
were sufficiently taken care of by the gravity type of masonry structures. Problems started only with
increase in axle loads, train lengths and their speed, warranting powerful locomotives. The increase in
axle load and longitudinal forces over the last 100 years is shown in Table-1.
Table - 1
S.
No.

Loading Std.

Year

Max. Axle
load

TLD

Max. Tractive
Effort

Approx.
population

Standard B

1903

18.0

1.2 t/ft.

Not specified

1%

Standard B 25%

1908

22.5

1.5 t/ft.

Not specified

1%

BGML

1926

22.9t

7.67 t/m

47.6t

90%

RBG

1975

22.5t

7.67 t/m

75t

1%

MBG

1987

25t

8.25 t/m

100t

7%

25t loading

2008

25t

9.33 t/m

126t

DFC

2008

32.5t

12.13 t/m

126t

11.2 It is worth mentioning that the girder bridges older than 1905 have either been rebuilt or strengthened
under the policy of replacing early steel girders. The increased longitudinal forces are most critical to
bridges in permitting heavier axle load. Tractive effort and speeds are required to be limited even after
considering dispersion of longitudinal forces by maintaining tract structure on bridge approaches in
sound condition. In many cases, the bearings of the bridges have been found failing in longitudinal forces
and replacement has been recommended. The position of such bridges is shown in Table - 2
Table - 2 Details of spans where strengthening of bearings is required to run 25t loading-2008
S.N.

Loading

Type of girder

Standar Min. strengthening required at fixed end/sliding bearings

d Span

BGML

Plate Girder

24.4m

Existing bearings as per RDSO Drg.No. BA-11005/1 Alt. 2

to be strengthened by providing 2 Nos. additional anchor
bolts.

-do-

Open
girder

web 30.5m

Strengthening to be done as per RDSO Drg. No. RDSO/B1521.

-do-

Underslung
girder

30.5m

Strengthening to be done as per RDSO Drg.No.CBS-2534

and rivets of locations strips are to be replaced by property
class 8.8 as per I.S.Code: 1367 part 3.

-do-

Open
girder

web 45.7m

Strengthening to be done as per RDSO Drg. No. RDSO/B1522.

231

-do-

Open
girder

web 61.0m

Strengthening to be done as per RDSO Drg. No. RDSO/B1523.

-do-

Open
girder

web 76.2m

Strengthening to be done as per RDSO Drg. No. RDSO/B1524.

RBG

Plate Girder

24.4m

Existing bearings as per RDSO Drg.No.BA-11075/1 to be

strengthened by providing 2 Nos. additional anchor bolts.

-do-

Open
girder

web 30.5m

Strengthening to be done as per RDSO Drg.No.BA-11344

Alt. 4.

-do-

Underslung

30.5m

Strengthening to be done as per RDSO Drg.No.BA-11404

Alt.2 and rivets of locations strips are to be replaced by
property class 8.8 as per I.S.Code: 1367 part 3. 2 Nos.
additional anchor bolts also be provided. .

10

-do-

Open
girder

web 45.7m

Strengthening to be done as per RDSO Drg.No.BA-11364

Alt.3.

11

-do-

Open
girder

web 61.0m

Strengthening to be done as per RDSO Drg.No.BA-11324

Alt. 3.

12

-do-

Open
girder

web 76.2m

Existing bearings to be replaced with Drg. No RDSO/B1524.

12.0 CONCLUSION
12.1 It is observed that the design longitudinal forces have increased considerably over the last 100 years due
to increase in axle loads and speeds. The old bridges, built prior to 1926 are particularly susceptible to
carry present day longitudinal loads.
12.2 The increase in longitudinal forces over the last 100 years on Indian Railways has made the bearings and
substructures susceptible to overstressing and increased the possibility of failures.
12.3 An holistic review of the effect of heavy axle load on the bridges needs to be made and more resources
need to be allotted to increase awareness and skills of engineering officials in various aspects of bridge
engineering.
12.4 Specialised institutions for bridge health monitoring need to be developed.
13.0 REFERENCES
13.1 RDSO, Ministry of Railways, Civil Engineering Report No.C-64, Progress Report No. 1, "Test on Sone Bridge Longitudinal Force on Bridges", April, 1968.
13.2 RDSO, Ministry of Railways, Civil Engineering Report No.C-64, Progress Report No. 14, "Summarised Integrated
Report - Longitudinal Forces on Bridges".
13.3 RDSO, Ministry of Railways, Civil Engineering Report No.C-119 "Bearing Stool for Measurement of Longitudinal
Force Dispersion to Bridge Sub-structure - Longitudinal Force on Bridge", April 1972.
13.4 RDSO File No.CBS/25t Axle load, "Guidelines for Checking the Suitability of Existing Bridges for 25t Loading", letter
to all zonal railways No.CBS/25t Axle Load dated 20-04-2009.

232

Annexure-I
Longitudinal Forces on Bridges
(As per Correction Slip No. 23 & 32 modified up-to-date)
2.8 LONGITUDINAL FORCES
2.8.1Where a structure carries railway track, provision as under shall be made for the longitudinal loads
arising from any one or more of the following causes:
(a) the tractive effort of the driving wheels of locomotives;
(b) the braking force resulting from the application of the brakes to all braked wheels;
(c) resistance to the movement of the bearings due to change of temperature and deformation of the
bridge girder. Roller, PTFE or elastomeric bearings may preferably be provided to minimize the
longitudinal force arising on this account.
(d) Forces due to continuation of LWR/CWR over the bridges (Abeyance till further orders).
2.8.1.1 Total longitudinal force transferred to sub-structure through any bearing due to causes mentioned in
Clause 2.8.1 shall not be more than the limiting resistance at the bearing for the transfer of longitudinal
force.
2.8.2For Railway Bridges, the value of longitudinal force due to either tractive effort or the braking force for
a given loaded length shall be obtained from the Appendices VIII, VIII (a), XXIV and XXVII.
2.8.2.1
(a)

For bridges having simply supported spans, the loaded length shall be taken equal to

The length of one span when considering the effect of longitudinal forces on
(i) the girders
(ii) the stability of abutments
(iii) the stability of piers carrying sliding or elastomeric bearings under one span loaded condition, or
(iv) the stability of piers carrying one fixed and one free (roller or PTFE) bearings.

(b) The length of two spans when considering stability of piers carrying fixed or sliding or elastomeric
bearings, under the two span loaded conditions. The total longitudinal force shall be considered divided
between the two spans in proportion to their lengths.
2.8.2.1.1 In case of continuous span bridges, appropriate loaded length shall be considered which will give
the worst effect.
2.8.2.2

No increase shall be made in the longitudinal force for the dynamic effect.

2.8.2.3

The longitudinal forces shall be considered as acting horizontally through the knuckle pins in case
of bearings having rocking arrangement or through girder seats in case of sliding, elastomeric or
PTFE bearings for the design of bearings and sub-structure.

233

2.8.2.4.1 For sub-structure having sliding or elastomeric bearings, following percentage of net longitudinal
force from the loaded spans after allowing for dispersion as per Clauses 2.8.3.1, 2.8.3.2 and 2.8.3.3 shall be
considered for the design:
Abutment 50%
Pier

40%

In case of multi-span bridges, the design of sub-structure shall also be checked for 20% of net longitudinal
force transferred from the span adjoining to the spans directly supported by the sub-structure under consideration
and considering the directly supported spans as unloaded. However, this force shall not be more than the
limiting resistance of the bearings on the sub-structure for the transfer of longitudinal force under unloaded
condition.
2.8.2.4.2 For spans having roller or PTFE bearings at one end, the whole of the net longitudinal force
after allowing for dispersion as per Clauses 2.8.3.1, 2.8.3.2 and 2.8.3.3 shall be considered to act through
the fixed end.
2.8.2.4.3 Forces due to continuation of LWR/CWR Till such time the forces due to continuation
of LWR/CWR on bridges on Indian conditions are finalized, provisions of UIC 774-3R October 2001
edition with modifications should be used with the provisions of values for the track resistance as specified
therein subject to the provision that a track resistance of 60 kN/m in loaded condition shall be taken
provisionally till field observations in Indian condition are finalized by RDSO. Accordingly, forces are to be
worked out for design of the bridges (abeyance till further order).
2.8.3 Dispersion and distribution of longitudinal forces.
2.8.3.1 In case of bridges having open deck provided with through welded rails, rail-free fastenings
and adequate anchorage of welded rails on approaches (by providing adequate density of sleepers, ballast
cushion and its consolidation etc., but without any switch expansion joints) the dispersion of longitudinal force
through track, away from the loaded length, may be allowed to the extent of 25% of the magnitude of
longitudinal force and subject to a minimum of 16t for BG and 12t for MMG or MGML and 10t for MGBL.
This shall also apply to bridges having open deck with jointed track with rail-free fastenings or ballasted deck,
however without any switch expansion or mitred joints in either case. Where suitably designed elastomeric
bearings are provided the aforesaid dispersion may be increased to 35% of the magnitude of longitudinal
force.
NOTE: Length of approach for the above purpose shall be taken as minimum 30m.
2.8.3.2 The dispersion of longitudinal force indicated in Clause 2.8.3.1 shall not exceed the capacity
of track for dispersing the longitudinal force to the approaches nor shall it exceed the capacity of anchored
length of the track on the approaches to resist dispersed longitudinal force. This aspect may be given special
attention for the stability of track in case of multi-span bridges provided with elastomeric bearings on all
spans.

234

2.8.3.3 In case of multi-span bridges having continuous spans, or flexible supports such as tall or
hollow RCC piers or steel trestles, or flexible bearings (elastomeric bearings) on all supports, or any other
special features, which are likely, to affect the distribution of longitudinal forces significantly, the dispersion
and distribution of longitudinal forces shall be determined by suitable analysis. The analysis shall take into
account stiffness and frictional characteristics of various resisting elements viz., supports, bridge girders,
bearings, rail-girder fixtures, track on bridge and approaches etc.
2.8.3.4 For the design of new bridges and in case of rebuilding of existing bridges, dispersion of longitudinal
force shall not be allowed.
2.8.4 Where the bridge carries more than one track, longitudinal forces shall be considered to act
simultaneously on all tracks. The maximum effect on any girder with two tracks so occupied shall be allowed
for, but where there are more than two tracks, a suitable reduction may be made in these forces for the
additional tracks.
2.8.5. When considering seismic forces, only 50% of gross tractive effort/braking force, to be reduced
by taking dispersion and distribution of longitudinal forces, shall be considered along with horizontal seismic
forces along/across the direction of the traffic.

x x x

235

LOAD RATING PROCEDURE FOR CAPACITY

ASSESSMENT OF BRIDGES FOR RUNNING
HEAVY AXLE LOAD TRAFFIC
R.K. Goel*
(Published in Indian Railway Technical Bulletin, R.D.S.O. Lucknow, Sept. 2009)

1.0 INTRODUCTION
A load-rating factor is a numeric value indicating a structure's ability to carry a specific load. Load rating
factors are computed by applying standard design loads along with the structure's self-weight and super
imposed dead loads. Rating factors are computed for various structural components and are equal to the
ratio of the component's live load capacity and the live load applied to that component; including all appropriate
load factors. A load-rating factor greater than 1.0 indicates that a member's capacity exceeds the applied
loads with the desired factors of safety. A rating factor less than 1.0 indicates a member is deficient such that
a specific vehicle cannot cross the bridge with the desired factor of safety. A number near to zero indicates the
structure cannot carry its own dead weight and maintain the desired safety factor. The lowest load rating
factor of various components generally controls the load rating of the entire structure. Additional factors may
be applied to account for variability in material, load application, and dynamic effects.
2.0 TYPE OF LOAD RATINGS
Two levels of load rating are performed for the bridge. An Service Load Rating corresponds to the
design stress levels and/or factors of safety and represents the loads that can be applied on a regular basis.
The Ultimate Load Rating correspond to the maximum load limits above which the structure may experience
damage or failure. It may be noted that occasional loads (i.e. extreme wind or seismic loads) are generally not
likely to come when the live loads are there on the bridge and hence to be considered accordingly.
3.0 ADVANTAGE OF LOAD RATING
The bridges for which simple design calculations indicate a load rating less than 1.0, more detailed
analysis can be made to take advantage of the conservations adopted in over simplified design procedures
and design load factors applied. Due to lack of sufficient information the material and cross-section properties
are generally under-estimated and girder end supports are assumed to be simply supported. In fact, even
relatively simple bearings offer resistance to bending and thus have a substantial effect on the midspan
bending moments. Inaccuracies associated with conservative assumptions are compounded with complex
framing geometries. With modern computational facilities, it is possible to generate a model of the structure
that is capable of reproducing the measured strains. Decisions concerning load rating are then based on the
*

Director/B&S, Research Designs & Standards Organization, Manak Nagar, Lucknow-226001

236

performance of the model once it is calibrated against measured values.

4.0 LOAD RATING EQUATION
4.1 The main purpose for obtaining an accurate model is to evaluate how the bridge will respond when
standard design loads or permitted loads are applied to the structure. Since load testing is generally not
performed with all type of load models, an analysis must be performed to determine load-rating factors
for most critical type of load model. Load rating is accomplished by applying the desired rating loads to
the model and computing the stresses on the primary members. Rating factors are computed using the
load rating equation.
. .C A1. D
RF = c s
A 2 .L(1 + I)
where:
RF =

Rating Factor for individual member.

Member Capacity.

Dead-Load effect.

Live-Load effect.

A1 =

Factor applied to dead-load.

A2 =

Factor applied to live-load.

Coefficient of Dynamic Augment, either from IRS Bridge Rules or measured.

fc =

Condition factor

fs =

System factor

4.2 It is important to understand that load testing and the integrated approach are most applicable to obtaining
Service Load Rating values. This is because it is assumed that all of the measured and computed responses
are linear with respect to load. It is an excellent method for estimating service load stress values but it
generally provides little additional information regarding the ultimate strength of particular structural
members. Therefore, load factor rating values also need to be computed using conventional assumptions
to assess member capacity. It is to be noted that load responses should never be permitted to reach the
inelastic range.
4.3 Load Factor Rating values needs to be computed to ensure a factor of safety between the ultimate
strength and the maximum allowed service loads. The safety to the public is of vital importance but as
long as load limits are imposed such that the structure is not damaged, then the safety is no longer an
issue.

237

5.0

LOAD RATING PROCEDURE

5.1 Field Data Generation: field data is required to be generated for use in developing a load rating for
the superstructure. These procedures will only complement the rating process, and must be used with
due consideration to the substructure and inspection reports indicating physical strength parameters.
5.2 Preliminary Investigation: It is intended to verify the linear and elastic behavior through continuity of
strain histories, locate neutral axis of flexural members, detect moment resistance at girder supports, and
qualitatively evaluate behavior.
5.3 Development of representative model: Use graphic pre-processors to represent the actual geometry
of the structure, including span lengths, girder spacing, skew, transverse members, and deck. Identify
gage locations on model identical to those applied in the field.
5.4 Simulate load test on computer model: Generate two dimensional model of test vehicle and apply to
structural model at discrete positions along the same paths as defined during field tests. Perform analysis
and compute strains at gage location for different loading conditions.
5.5 Compare measured and initially computed strain values: Various global and local error values at
each gage location are computed and visual comparisons are made with the help of a graphical processor.
5.6 Evaluate modeling parameters: Improve model based on data comparisons. Engineering judgment
and experience is required to determine which variables are to be modified. A combination of direct
evaluation techniques and parameter optimization are used to obtain a realistic model.
5.7 Model Evaluation: In some cases it may not be desirable to rely on secondary stiffening effects as
their effectiveness may be questionable at higher load levels. It is beneficial, though, to quantify their
effects on the structural response so that a representative computer model can be obtained. The stiffening
effects that are deemed unreliable can be eliminated from the model prior to the computation of rating
factors. For instance, if a non-composite bridge is exhibiting composite behavior, then it can conservatively
be ignored for rating purposes. However, if it has been in service for 50 years and it is still behaving
compositely, chances are that very heavy loads have crossed over it and bond-breaking if any, would
have already been occurred. Therefore, probably some level of composite behavior can be relied upon.
5.8 Perform load rating: Apply load model as per IRS loading given in Bridge Rules to the calibrated
model and perform load rating. For MBG and 25t loading the load model are given in Appendix I & II
respectively.
5.9 Consider other factors: Other factors such as the condition of the bridge and/or substructure, traffic
volume, and other information available in the inspection report should be taken into consideration and
the rating factors adjusted accordingly. The load Resistance Factors for the purposes are given in
Table-1.

238

Table 1 Load Resistance Factors for Bridge Rating

Dead Load

DC (Dead-load effects due to structural components and

attachments)

1.25

DW (Dead-load effect due to wearing surface and utilities i.e.

super imposed dead load)

1.50

Inventory or service

1.75

Operating or ultimate

1.35

Condition Factor, c to be Good or Satisfactory

applied on member capacity
Fair

1.00

Live Load

Poor

0.95
0.85

System Factor, s to be applied Welded Members in Two-Girder/Truss/Arch Bridges

on member capacity.
Riveted Members in Two-Girder/Truss/Arch Bridges

0.85
0.90

Three-Girder Bridges with Girder Spacing ? 1.8m

0.85

Four-Girder Bridges with Girder Spacing ? 1.2m

0.95

All other Girder Bridges and Slab Bridges

1.00

Floorbeams with Spacing > 3.6m and Non-Continuous Stringers

0.85

Redundant Stringers and Subsystems Between Floorbeams

1.00

6.0 CONCLUSIONS
6.1 It can be seen that load rating of bridges can help us to ascertain the actual carrying capacity with
respect to the design loads in a rational manner and decisions can be taken with confidence to permit
heavy axle load operation on the basis of Load Rating Values.
6.2 The procedure takes into account the redundancies, the reserve strength, the additional capacities of the
bridge components while ensuring safety with respect to specified factors of safety and actual condition
of the bridge.
6.3 The purpose of instrumentation of the bridges can be effectively served by suggesting on overall Load
Rating Factor for the bridge covering all aspects of different bridge components.

239

Appendix - I

240

Appendix - II

241

Appendix - II

242

Appendix - II

243

244
x x x

EXPERIENCE OF PILOT PROJECT ON LOAD

RATING OF BRIDGE NO. 502 ON WESTERN RAILWAY
BY LIVE LOAD TESTING
Mahesh K.1 Gupta & R.K. Goel2
(Published in Journal of IPWE, Jan 2010)

1.0 INTRODUCTION
1.1 Rating bridges against the specified design loading provides a simple numerical method of comparing
the assessed load capacity of existing bridge against the specified design loading. The concept utilizes all
aspects of design loading including the speed and thus, the same basic live load travelling at a lower
speed is considered a lower loading. Rating the bridge by live load testing is considered advantageous
as it takes into account the structural redundancies in the system and thus enables bridge owners to take
full advantage of its reserved capacities.
1.2 A pilot study was conducted by Western Railways to perform live load testing on Bridge No. 502 on
river Narmada near Surat. This multi-truss bridge was selected for testing due to the complexity of
analyzing such a large structure. The goal of the testing was to obtain and then utilize field measurements
to verify an analytical model from which accurate load ratings could be obtained.
1.3 All field testing operations were conducted on April 28th, 2006. The Structural Testing System (STS)
was used for measuring strains at 36 locations on the superstructure while it was subjecting to a slow
moving live load. The response data was then used to "calibrate" an analytical finite element model of the
structure which was in turn used to develop load ratings using the American Railway Engineering and
Maintenance of Way Association (AREMA) design guidelines.
1.4 The field testing and load rating report was sent to B&S Directorate of RDSO for critical examination.
It was observed that through, the testing of the bridge was done as per actual trainloads run over the
bridge and the analytical model was calibrated on the basis of field measurements, the load rating initially
given was for Cooper-E 80 loading and not as per the standard IRS loading. M/s. Bridge Diagnostic
Inc. were subsequently supplied with the IRS MBG loading and new 25t loading and modified load
ratings were obtained. This paper describes the load rating procedure adopted and the results of the
pilot study.

1. Executive Director/B&S, Research Designs & Standards Organization, Manak Nagar, Lucknow-226011
2. Director/B&S, Research Designs & Standards Organization, Manak Nagar, Lucknow-226011

245

2.0 DETAIL OF BRIDGE NO. 502

2.1 The bridge is located at km 323/33 to 325/10 in Ankaleswar-Bharuch (AKV-BH) section of New
Delhi-Mumbai main line. The GMT of the section on both lines Up & Dn is 53. The details of the
bridge are given in Table - 1
Table - 1 Salient Features of Bridge No. 502 on River Narmada
Location/Route

323/33 325/10, New Delhi-Mumbai

Section

Ankaleswar Bharuch (AKV-BH)

Year of Construction

1933 to 1935 (Opened for traffic on 20th Dec 1935)

Weight

2x24 MT + 1x825 MT + 15x905 MT

Loading

BGML of 1926 (As per old bridge inspection registers)

Structure Type

282 feet span Steel Truss

Total Number of Spans

15, plus 1 249 feet Truss and 1 60 feet approach span.

Number of tracks

Span Lengths

282.0 (85.95m)

Structure width

34.67 (10.57m)

Deck Type

Wooden Sleepers / no ballast i.e. open deck

Spans Tested

Test Reference Location (X=0,Y=0)

Northeast bearing point.

Test Vehicle Direction

South i.e. New Delhi to Mumbai side

Test Beginning Point

Front axle at X = -258 feet (-78.64 meter)

Lateral Load Position(s)

Y1 = West Track, Y2 = East Track,

Number/Type of Sensors

36 strain transducers (BDI Make)

Data Acquisition Sample Rate

40 Hz

Number of Test Vehicles

3 2 WCAM1 Locomotives, Engine No. 21844 & No. 21835

followed by 1 loaded BRNAHS SECR, loaded with sleepers.

Structure Access Type

Scaffolding / Ladders

2.2 Details of Major Activities on the Bridge

1.

First time DN line Stringer angles replaced in 1968-69 (span no 5 to 11) and G.I. sheet cover was
provided at stringer level for protection against train droppings.

2.

Oil bath provided in 1979-80.

3.

Top Stringer angle of UP line replaced completely in 2001-02.

4.

Perforated bottom cover plate of Bottom Chord replaced at many locations in various spans time
to time as per requirement.

246

5.

6.

Through stringer angle replacement work is being done for DN line (2009) due to
i)

Corrosion/Pitting of angles.

ii)

Cracks observed at the root of the angle.

Corrosion protection of top stringer angles is also in progress by providing a cover plate for both
lines.

3.0 CONCEPT OF LOAD RATING

A load-rating factor is a numeric value indicating a structure's ability to carry a specific load. Load rating
factors are computed by applying standard design loads along with the structure's self-weight and super
imposed dead loads. Rating factors are computed for various structural components and are equal to the
ratio of the component's live load capacity and the live load applied to that component; including all appropriate
load factors. A load-rating factor greater than 1.0 indicates that a member's capacity exceeds the applied
loads with the desired factors of safety. A rating factor less than 1.0 indicates a member is deficient such that
a specific load cannot cross the bridge with the desired factor of safety. A number near to zero indicates the
structure cannot carry its own dead weight with the desired safety factor. The lowest load rating factor of
various components generally controls the load rating of the entire structure. Additional factors may be applied
to account for variability in material, load application, and dynamic effects.
4.0 EVALUATION OF SUPPORT STIFFNESS AND STRUCTURAL REDUNDANCY
4.1 Advantage of Load Rating
The bridges for which simple design calculations indicate a load rating less than 1.0, more detailed
analysis can be made to take advantage of the conservations adopted in over simplified design procedures
and design load factors applied. Due to lack of sufficient information the material and cross-section
properties are generally under-estimated and girder end supports are assumed to be simply supported.
In fact, even relatively simple bearings offer resistance to bending and thus have a substantial effect on
the midspan bending moments. Inaccuracies associated with conservative assumptions are compounded
with complex framing geometries. With modern computational facilities, it is possible to generate a
model of the structure that is capable of reproducing the measured strains. Decisions concerning load
rating are then based on the performance of the model once it is calibrated against measured values.
4.2 A common requirement in identification of structural redundancies is to determine effective spring stiffness
that best represents in-situ support conditions. It is generally simple to evaluate a spring constant in
terms of moment per rotation however this value generally has little meaning to the engineer. A more
conceptual approach is to evaluate the spring stiffness as a percentage of a fully restrained condition.
e.g. zero% being a pinned condition and 100% being fixed. This is best accomplished by examining the
ratio of the beam or slab stiffness to the rotational stiffness of the support.
4.3 As an illustration, a point load is applied to a simple beam with elastic supports (Figure-1). By examining
the moment diagram, it is apparent that the ratio of the end moment to the midspan moment (Me/Mm)
equals zero if the rotational stiffness (Kr) of the springs is equal to zero. Conversely, if the value of Kr is
247

set to infinity (rigid) the moment ratio will equal 1.0. If a fixity term is defined as the ratio (Me/Mm), which
ranges from 0 to 100 percent, a more conceptual measure of end restraint can be obtained.

Figure - 1 Moment Diagram of Beam with Rotational End Restraint

5.0 MODEL CORRELATION AND PARAMETER MODIFICATION

5.1 The accuracy of the model is determined numerically by the analysis using several statistical relationships
and through visual comparison of the strain histories. The numeric accuracy values are useful in evaluating
the effect of any changes to the model, where as the graphical representations provide the engineer with
the best perception to understand clearly as why the model is responding differently than that indicated
by measurements. Member properties that cannot be accurately defined by conventional methods or
directly from the field data are evaluated by comparing the computed strains with the measured strains.
These properties are defined as variable and are evaluated such that the best correlation between the
two sets of data is obtained. It is the engineer's responsibility to determine which parameters need to be
refined and to assign realistic upper and lower limits to each parameter. The evaluation of the member
property is accomplished with the aid of a parameter identification process (optimizer) built into the
analysis.
5.2 In short, the process consists of an iterative procedure of analysis, data comparison, and parameter
modification. It is important to note that the optimization process is merely a tool to help evaluate various
modeling parameters. The process works best when the number of parameters is minimized and reasonable
initial values are used. The typical plot of measured values and the computed values by the calibrated
model are shown in Fig - 2.

248

Figure - 2 Typical plot of measured and calculated values after calibration

6.0 LOAD RATING EQUATION

6.1 The main purpose for obtaining an accurate model is to evaluate how the bridge will respond when
standard design loads or permitted loads are applied to the structure. Since load testing is generally not
performed with all type of load models, an analysis must be performed to determine load-rating factors
for most critical type of load model. Load rating is accomplished by applying the desired rating loads to
the model and computing the stresses on the primary members. Rating factors are computed using the
load rating equation.

RF

. s .C A 1 .D
A .L( 1 I )
2
c

where:
RF

= Rating Factor for individual member.

= Member Capacity.

= Dead-Load effect.

= Live-Load effect.

249

A1, A2

= 1.0 for Allowable Stress & 1.30 for Maximum limit state.

fc & fs

= Condition and System factors

= Coefficient of Dynamic Augment, from IRS Bridge Rules

The various factors proposed to be applied in AREMA are given in Table - 2

Table - 2 Load Resistance Factors for Bridge Rating
DC (Dead-load effects due to structural components and 1.25
attachments)
DW (Dead-load effect due to wearing surface and utilities i.e. super 1.50
imposed dead load)
Inventory or service
1.75

Dead Load

Live Load
Condition Factor, c to be
applied on member capacity

Operating or ultimate

1.35

Good or Satisfactory

1.00

Fair

0.95

Poor

0.85

System Factor, s to be applied Welded Members in Two-Girder/Truss/Arch Bridges

on member capacity.
Riveted Members in Two-Girder/Truss/Arch Bridges

0.85
0.90

Three-Girder Bridges with Girder Spacing ? 1.8m

0.85

Four-Girder Bridges with Girder Spacing ? 1.2m

0.95

All other Girder Bridges and Slab Bridges

1.00

Floor beams with Spacing > 3.6m and Non-Continuous Stringers

0.85

Redundant Stringers and Subsystems Between Floor beams

1.00

7.0 LOAD COMBINATIONS ADOPTED FOR MODIFIED LOAD RATING

7.1 The details of "IRS MBG Loading" and "IRS 25t Loading - 2008" were supplied to M/s. Bridge
Diagnostics, USA and load rating has been given for one of the load combinations given. Figure-3 and
Figure-4 show the details of the load combinations used in the load rating. In MBG loading lighter loco

Figure - 3 Modified B.G. Loading-1987

Load Combination-2 ( Appendix-I of Bridge Rules)
250

Figure - 4 : "25t Loading - 2008"

Double Headed Diesel Locomotive Load Combination-1 (Appendix-XXII of Bridge Rules)
(22.5t axle load) combination was chosen whereas Double Headed Diesel Locomotive of the 25t
loading-2008 was chosen as the representative rating locomotives for their respective categories. It
may also be noted that the ratings were calculated based on the assumption that the applicable rating
locomotive was occupying both tracks simultaneously.
7.2 The load rating has been done by considering full coefficient of dynamic augment as per IRS Bridge
Rules. Figure 5 gives the schematic diagram showing the location of strain transducers.

Figure - 5 Locations of strain transducers in truss

8.0 CAPACITY ASSESSMENT AND LOAD RATING

8.1 The member capacities were analyzed based on yield strength of 240 MPa, The analysed capacities for
each member are provided below in Table 3. The two types of load rating are provided (a) Normal allowable rating which is with respect to permissible stresses.
(b) Maximum allowable rating which is with respect to maximum allowable stress. The load rating
values for various members of the bride are given in table 4 & 5 for MBG loading & 25t loading
- 2008 respectively.
251

Table-3 Member Capacities

Normal Allowable Capacity
S.No.

Member
Des.

Maximum Allowable Capacity

TENSIO COMP. MOMEN SHEAR TENSIO COMP. MOMEN SHEAR

N (kN)
(kN) T (kN-m) (kN)
N (kN)
(kN) T (kN-m) (kN)

Top-Lat

1038

829

1510

1222

U6-U8

10654

10182

15496

14888

U4-U6

10654

10182

15496

14888

U2-U4

8926

8587

12984

12552

L0-U2

8926

8815

12984

12872

Top-EndLat

1495

1374

2174

2012

Port-Lat

1166

993

1695

1459

FB-1CP

6534

6534

3396

1598

9504

9504

4940.23

2740

FB-2CP

7641

7641

4199

1598

11114

11114

6107.96

2740

10

FB-Ends

5427

5427

2595

1598

7894

7894

3775.10

2740

11

Stringers

2490

2490

736

714

3622

3622

1070.49

1223

12

U2-L2

2632

1929

3829

2857

13

U2-L4

5926

5699

8620

8331

14

U4-L4

2927

2674

4257

3917

15

U4-L6

4035

3561

5870

5226

16

U6-L6

2927

2674

4257

3917

17

U6-L8

3090

2739

4494

4018

18

U8-L8

3090

2350

4494

3473

19

L7-M7

1431

1431

2082

2082

20

L6-M7

1495

1377

2174

2017

21

L5-M5

1431

1431

2082

2082

22

L4-M5

1495

1377

2174

2017

23

L3-M3

1431

1431

2082

2082

24

L2-M3

1495

1377

2174

2017

25

M1-L2

1495

1377

2174

2017

26

L1-M1

1431

1431

2082

2082

27

L0-L4

5886

5886

8562

8562

28

L4-L6

7265

7265

10567

10567

29

L6-L8

9048

9048

13161

13161

30

Port-Int-Bot

1166

1166

1695

1695

252

Table - 4 Rating Factors (MBG Loading)

Moment

Shear

Rating Factor
Floor System

Impact
Factor

Normal
Allow.
Stress

Max.
Allow.
Stress

Rating Factor
Dead Load
Response
(kN-m)

Live Load
Response
(kN-m)

Normal
Allow.
Stress

Dead Load
Response
Max. Allow.
(kN)
Stress

Live Load
Response
(kN)

Stringers

43.10%

1.69

2.5

37

289

2.2 6

3.9

-14

-217

Floor Beam (No Cover

Plates)

40.40%

2.7

3.94

24

679

1.68

2.94

72

647

40.40%

2.01

2.94

80

1176

1.68

2.94

70

647

40.40%

1.94

2.85
123
Compression Force

1495

3.5 4

6.14
37
Tension Force

Floor Beam (1 Cover

Plate)
Floor Beam (2 Cover
Plates)

Truss End Post

L0-U2

Impact
Factor

23.70%

Rating Factor

Dead Load
Response
(kN)

Live Load
Response
(kN)

Rating Factor

Dead Load
Response
Max. Allow.
(kN)
Stress

314

Live Load
Response
(kN)

Normal
Allow.
Stress

Max.
Allow.
Stress

2.24

3.54

-1777

-2536

na

na

na

na

Normal
Allow.
Stress

Top Chords
U2-U4

23.70%

1.89

3.03

-2002

-2816

na

na

na

na

U4-U6

23.70%

1.75

2.82

-2505

-3544

na

na

na

na

U6-U8

23.70%

1.82

2.93

-2423

-3444

na

na

na

na

L0-L4

23.70%

na

na

na

na

2.3 7

3.70

1144

1620

L4-L6

23.70%

na

na

na

na

1.5 3

2.49

1966

2794

L6-L8

23.70%

na

na

na

na

1.5 4

2.49

2447

3473

L1-M1

23.70%

11.05

16.08

-3

-104

2.9 7

4.31

-3

391

L2-U2

23.70%

52.95

75.92

210

-33

2.3 3

3.48

210

841

L3-M3

23.70%

59.60

85.35

74

-20

2.7 2

4.02

74

404

L4-U4

23.70%

1.43

2.31

-646

-1144

10 .42

14.30

-646

277

L5-M5

23.70%

18.25

26.18

66

-66

3.1 6

4.67

66

349

L6-U6

23.70%

2.26

3.44

-296

-852

4.9 1

6.94

-296

530

L7-M7

23.70%

13.50

19.33

75

-90

3.0 2

4.46

75

363

L8-U8

23.70%

76.01

99.99

-87

-24

66 .02

95.20

-87

39

M1-L2

23.70%

5.82

8.48

24

-195

8.4 8

12.40

24

140

L2-M3

23.70%

4.78

7.12

-66

-222

38 .96

55.91

-66

32

U2-L4

23.70%

49.46

68.40

1176

-112

1.8 8

2.96

1251

2010

L4-M5

23.70%

4.27

6.37

-78

-246

24 .98

35.77

-78

51

U4-L6

23.70%

9.79

13.64

680

-350

1.7 1

2.65

740

1562

L6-M7

23.70%

4.32

6.43

-70

-245

17 .22

24.69

-70

73

U6-L8

23.70%

3.12

4.49

170

-754

2.99

236

1153

Bottom Chords

Verticals

Diagonals

253

Table -5 Rating Factors (25t Loading)

Moment
Floor System

Stringers
Floor Beam (No
Cover Plates)
Floor Beam (1 Cover
Plate)
Floor Beam (2 Cover
Plates)

Impact
Factor

Shear

Rating Factor

43.10%

Normal
Allow.
Stress
1.58

40.40%

Rating Factor
Dead Load
Live Load
Dead Load Live Load
Resp.
Max. Allow. Response Response (kN- Normal Max. Allow. Response
(kN-m)
m)
(MPa)
(MPa)
Stress
Allow. Stress
Stress
2.34

37

309

2.1 4

3.7

14

228

3.09

4.5

24

593

1.9 2

3.36

72

566

40.40%

2.28

3.34

80

1037

1.9 2

3.36

70

566

40.40%

2.21

3.25
123
Compression Force

1312

4.1 8

7.23
37
Tension Force

266

Rating Factor
Truss End Post

L0-U2

Impact
Factor

Normal
Allow.
Stress

Rating Factor

Dead Load
Response
Max. Allow.
(MPa)
Stress

Live Load
Response
(MPa)

Dead Load Live Load

Response
Resp.
Normal Max. Allow.
(MPa)
(MPa)
Allow. Stress
Stress

23.70%

2.19

3.46

-1777

na

na

na

na

U2-U4

23.70%

1.88

3.02

-2002

-2827

na

na

na

na

U4-U6

23.70%

1.73

2.79

-2505

-3583

na

na

na

na

U6-U8

23.70%

1.80

2.89

-2423

-3490

na

na

na

na

L0-L4

23.70%

na

na

na

na

2.3 5

3.68

1144

1631

L4-L6

23.70%

na

na

na

na

1.5 3

2.48

1966

2808

L6-L8

23.70%

na

na

na

na

1.5 3

2.48

2447

3498

L1-M1

23.70%

10.93

15.92

-3

-106

3.1 2

4.54

-3

371

L2-U2

23.70%

51.95

74.48

210

-33

2.66

3.98

210

735

L3-M3

23.70%

55.17

79.00

74

-22

2.9 3

4.34

74

374

L4-U4

23.70%

1.48

2.39

-646

-1105

12 .54

17.22

-619

228

L5-M5

23.70%

20.17

28.93

66

-60

3.0 9

4.56

66

357

Top Chords

Bottom Chords

Verticals

L6-U6

23.70%

2.49

3.79

-296

-773

5.7 5

8.12

-296

453

L7-M7

23.70%

14.91

21.35

75

-82

3.1 3

4.63

75

351

L8-U8

23.70%

74.46

99.99

-87

-25

67 .06

96.71

-87

38

M1-L2

23.70%

5.37

7.81

24

-211

8.2 3

12.03

24

145

L2-M3

23.70%

4.62

6.87

-66

-229

35 .69

51.23

-66

35

U2-L4

23.70%

54.54

75.42

1176

-102

1.8 8

2.97

1251

2006

L4-M5

23.70%

4.02

6.01

-78

-261

27 .8

39.80

-78

46

U4-L6

23.70%

11.59

16.13

680

-296

1.7 6

2.75

740

1509

L6-M7

23.70%

4.29

6.39

-70

-246

19 .18

27.51

-70

66

U6-L8

23.70%

3.66

5.27

170

-643

2.2 1

3.29

236

1046

Diagonals

254

9.0 DISCUSSION OF LOAD RATING VALUES

9.1 It can be observed from the table 3 & 4 that as a whole the load ratings are very good, with all ratings
well above 1.0. The critical rating factor for Normal Allowable Stress was 1.43 for the MBG Loading
and 1.48 for the 25t Loading. In both cases, the load rating was controlled by the compressive strength
capacity of the vertical member L4-U4.
9.2 It is interesting to note that the critical loading factor is 1.48 for 25 t loading-2008 which is more than the
corresponding value for MBG loading. In view of 25t laoding-2008 being heavier in terms of EUDLs
given in the Bridge Rules, there appears a contradiction. However, a critical analysis explains the
contradiction. The load ratings have been done with respect to only one load combination of each
loading i.e. Combination-II of MBG loading and Combination-I of 25 t - loading. A careful observation
of Bridge Rule shows that there are only two load combinations in MBG loading and Combination-II is
more severe due to reduced axle spacing of the loco. Whereas, there are 5 load combinations in 25t
loading and the load Combination-I is not the severe most.
9.3 Therefore, the rating factors for 25 t - loading needs to be worked out for all the other four load
combinations and the most critical rating factor needs to be considered. However, the rating factor for
MBG loading is acceptable and implies that the bridge, although designed for BGML loading is sufficiently
strong to take the loads of MBG loading. This, however further needs to be examined from fatigue
considerations also.
10.0 CONCLUSIONS
10.1 The load rating by live load testing is an specialized job which requires high degree of technical understanding
of the bridge design, careful planning for instrumentation, special purpose structural analysis software
suiting to the rating methodology, accurate strain transducers and deep sense of commitment. Such
works need to be handled by specialized agencies under the supervision of competent Bridge Designers.
10.2 It is observed that the existing bride no. 502 of 1935, designed for BGML loading gives a load rating of
1.43 for MBG loading which is an indication of its good condition to safely carry the present day
loadings.
10.3 The load rating for load combination-1 of 25t loading-2008 is 1.48 which needs to be worked out for
other remaining combinations. However, apparently the bridge is sound enough to carry the 25t loading
with specified factors of safety.
10.4 Separate fatigue assessment is required to be done to get an idea of the fatigue strength of various
components of the bridge.
11.0 OBSERVATIONS
11.1 Although the load rating of the bridge is good for MBG loading& 25t loading, still the replacement of
top angles of stringers has been done due to following reasons:

255

i)

Fatigue cracks in top angles have been observed and these angles have been replaced on Up line
in 2001-02 and being replaced on Dn. Line in 2009.

ii)

The replacement work of top angle of stringers is being done on condition basis due to corrosion/
pitting. The protection measures against corrosions are also being taken.

12.0 LIMITATIONS
12.1 Occasional loads (e.g. seismic, wind & longitudinal loads) have not been considered in determining the
above load ratings.
12.2 Fatigue assessment of connections and members has not been done which is necessary to know its
overall effect on the fatigue life of the bridge components.
12.3 Coefficient of Dynamic Augment (CDA) has been considered as per IRS Bridge Rules. Dynamic analysis
of the bridge is further required to assess the actual effects of the bridge dynamics, in case the rating is
required for higher speeds.
13.0 ACKNOWLEDGEMENTS
The authors sincerely acknowledge the assistance rendered by M/s. Bridge Diagnostics, USA to supply
the updated member capacities and modified load rating as per standard IRS loading even after completion
of pilot project.
14.0 REFERENCES
14.1 Bridge Diagnostics, Inc. USA, "Field testing and load rating on bridge no. 502 - 282 foot steel truss", report submitted
to Western Railway, Oct' 06 & updated July, 2009.

x x x

256

CHECKING THE SUITABILITY OF EXISTING

ROLLER-ROCKER BEARINGS FOR HIGHER
AXLE LOAD
Ravindra Kr. Goel1, Hari Om Narayana1 & Sujeet Nath Gupta3
(Presented in IPWE Seminar held at Jaipur, Jan., 2007)

1.0 INTRODUCTION
Conventional roller and rocker bearings are provided in open web girder bridges since long. These
bearings have been designed as per the loading standard (mostly BGML or RBG) prevailing at the time of
construction of the bridges. As the traffic volume is increasing and extra revenue is proposed to be generated,
a policy decision has already been taken to make some of the routes fit for running 25t axle load. The existing
bridges are required to be checked for carrying this increased axle load, on restricted speed, if required. The
implications of higher axle load are also to be examined on existing bearings with respect to original design
parameters of BGML/RBG loading. It has been seen that zonal railways are not having complete awareness
about the various design parameters involved in design of rocker-roller bearings and replacement of bearings
are proposed without exercising adequate design check. The paper describes in detail the relevant provisions
for analysis of such bearings and comments on the adequacy of existing bearings of standard spans.
2.

LOADS AND LOAD COMBINATIONS

2.1 Loads
The loads to be considered for design/analysis of bearings are to be taken from Clause 2.1 of Bridge
Rules. The following loads are to be invariably considered. It is seen that the lateral horizontal forces due to
wind (not due to Earthquake) are governing in steel girders. Therefore, the effect of wind load is to be
considered unless some special conditions are found prevailing at the site of bridges.
a)

Dead Load

b) Live load including raking force

c)

Impact load

d) Longitudinal load
e)

Wind load

f)

Forces due to curvature and eccentricity

Director, Bridges & Structures Directorate, Research Designs & Standards Organization, Indian Railways, Manak Nagar, Lucknow226001
2
Sr. Section Engineer/Design, Research Designs & Standards Organization, Indian Railways, Manak Nagar, Lucknow-226001
3
Section Engineer/Design, Research Designs & Standards Organization, Indian Railways, Manak Nagar, Lucknow-226001

257

a)

Dead Load

Dead load can be taken from the design sheets or drawings of the existing bridge to be checked. In the
absence of reliable documents the dead load has to be estimated on the basis of detailed survey of the
different bridge components and connections.
b)

Live Load

At present it is customary to use the concept of EUDL for working out the live load effect on bearing.
EDUL for shear for the particular span (effective) is to be considered for calculating live load on each bearing.
Annexure 2 of RDSO letter No.CBS/Golden/Q/Strength dated 8-8-06 can be referred for the purpose.
This EUDL table has been prepared for 2WAG-9/2WDG4 with BOXN wagons of 25t axle load.
c)

Impact Load

For working out the impact load impact factor CDA can be derived from Clause 2.4.1 of Bridge Rules.
Impact factor can be reduced for reduced speed. Proportionately if the bearings are to be checked for
operation of traffic at restricted speed as per the following expression.
Im p act L o ad : C D A x L .L . x
d)

S p eed R estricte d in k m p h
125

Longitudinal Force

It is an important design parameter, considerably affecting the design of bearings. The longitudinal force
actually applied at the bearing depends upon the maximum tractive effort of loco running over the route,
maximum braking force applied by the train, dispersion of longitudinal force through the girder and track.
Clause 2.8.3.2 of Bridge Rules (as per A&C Slip No.22) has to be referred in this regard which is reproduced
below for ready reference.
"In case of bridges having open deck provided with through welded rails, rail-free fastenings and
adequate anchorage of welded rails on approaches (by providing adequate density of sleepers, ballast
cushion and its consolidation etc., but without any switch expansion joints) the dispersion of longitudinal
force through track, away from the loaded length, may be allowed to the extent of 25% of the magnitude
of longitudinal force and subject to a minimum of 16t for BG and 12t for MMG or MGML and 10t for
MGBL. This shall also apply to bridges having open deck with jointed track with rail-free fastenings
or ballasted deck, however without any switch expansion or mitred joints in either case. Where suitably
designed elastomeric bearings are provided the aforesaid dispersion may be increased to 35% of the
magnitude of longitudinal force."
Note: Length of approach for the above purpose shall be taken as minimum 30m
As per this clause the dispersion in longitudinal force (tractive effort or braking force) may be permitted
upto 25% by ensuring adequate anchorages of welded rails on approaches.

258

Longitudinal force can be taken from Annexure 5 of RDSO letter No.CBS/Golden/Q/Strength dated
08-08-06 as the maximum value of tractive effort or braking force for effective span under consideration.
These values are again for 2WAG-9/2 WDG4 loco with BOXN wagons of 25t axle load.
e)

Wind Load

Wind load is to be taken from the existing design sheets which is generally for a wind load intensity of
150 kg/m2. As per A&C Slip No.34 of Bridge Rule, the wind load shall not be considered acting alongwith
live load if the wind load intensity is more than 100 kg/m2. Therefore, the wind load for analysis of existing
bridges for permitting higher axle load shall have to be modified accordingly. Wind load has following two
effects:

f)

i)

To apply overturning effect on bearings causing increased vertical load on leeward side bearings.

ii)

To apply a lateral force on bridge simultaneously with longitudinal force. Resultant of the two
forces is to be resisted in shear by saddle bolts and anchor bolts.

Forces due to curvature or eccentricity of track

Usually, the girder bridges are on straight. However, whenever they are on curve the extra load on one
girder should be calculated as per Clause 2.5.3 (a) of Bridge Rules and the extra horizontal load due to
centrifugal force should be taken as per Clause 2.5.3 (b) of Bridge Rules. Eccentricity effects are usually not
considered in girder bridges unless they are large enough.
2.2 Load Combinations
Following load combinations are to be considered:
a)

For checking bed plates and bearing pressure on bed blocks

i)

DL + LL + Impact

ii)

DL + LL + Impact + Overturning effect due to wind (Permissible stresses are increased by 16.67%)

iii) DL + LL + Impact + Overturning effect due to wind & longitudinal force.

(Permissible stresses are increased by 33.3%)
b)

For checking of anchor bolts and saddle bolts in horizontal shear

i)

Resultant horizontal force comprising longitudinal force and wind load.

(Permissible stresses are not increased in this combination)

3.0 ASSESSMENT OF EXISTING BEARING
3.1 Check for adequacy of maximum bed plate pressure
This check is conducted to ensure that the concrete/bed stone is safe against crushing under the vertical
loads. The maximum pressure calculated should be with in the allowable pressure which can be taken from
259

Clause 3.16 of Steel Bridge Code. This clause is reproduced below for ready reference
"Allowable Working Pressure under Bearings or Bed Plates
The area of bearings or bed plates shall be so proportioned that when the eccentricity of loads
due to combination mentioned in Clause 3.2.1 the maximum pressure on material forming the bed
shall not exceed the following limits: Granite

36 kg/cm2

Sand Stone

(33 tons/ft2)

29.5 kg/cm2 (27 tons/ft2)

Cement Concrete:
As laid down for permissible bearing pressure in Plain concrete in Table III and III(a) of the IRS
Concrete Bridge Code-1962.
Reinforced Concrete:
As laid down for permissible stress in direct compression for the specified crushing strength at 28 days
for ordinary Portland cement (or the equivalent period of time for other cement) given in Table III and III(a)
of IRS Concrete Bridge Code-1962
The above-mentioned limits may be exceeded by 331/3 per cent for combinations mentioned in
clauses 3.2.2 and 3.2.3
The centre of pressure under flat bearing plates attached to the girders shall be assumed to be at
one-third of the length from the front edge."
3.2 Maximum pressure on concrete
Maximum pressure on concrete can be calculated from the formula:
P

V
L B

LF L a 6
L B2

WL 6 L a
L2 B

Where
V

Vertical load with overturning effect of wind

La

Lever arm, in the height from the center of knuckle to the bottom of base plate or top of
pier/abutment

Length of base plate

Width of base plate

LF

Maximum longitudinal force

WL

Wind load

260

3.3 Check for anchor bolts

Anchor bolts are provided at the base plate to restrain the bed plate against any horizontal movement in
X & Y direction. These bolts are checked basically to resist resultant force comprising of longitudinal force
and lateral force due to wind effect. The shear stress on each bolt can be calculated by the following formula:
Resultant Force

Shear stress per bolt =

d2

Where
n =

No. of anchor bolts provided

d =

Diameter of anchor bolts

3.4 Check for saddle bolts

These bolts are provided to connect the bottom chord of open web girders with the saddle plate. These
bolts provide an interface for the transfer of horizontal force in X, Y direction. These bolts are checked for
shear stress which is equal to
Shear stress per bolt

Resultant force
n

d2

Where

3.5

No. of bolts

Dia of saddle bolt

Relaxation of Permissible stresses

Permissible stresses for retaining existing bridges can be relaxed as per clause 3.20.2 of Steel Bridge
Code which is reproduced below:
"Mild Steel, Wrought Iron and Early Steel Girders
Bridge spans other than open web girder spans may, if they are kept under regular observation by
the Bridge Engineer and his staff, be retained in use, provided that if the impact effect-specified in
clause 3 of the Bridge Rules (Revised 1964) for the maximum permissible speed over the bridges is
allowed for the calculated stresses for various combinations of loads as laid down in relevant clauses
do not exceed the working stresses specified for those combinations by more than 11 percent. Under
the same conditions, permissible shear and bearing stresses on rivets may be increased by 25 per cent.
This increase in rivet stresses shall not be allowed if the stresses are calculated by the method given in

261

APPENDIX- E.
Under the conditions specified above, open web girder spans may be retained in use, provided
that the calculated tensile and compressive stresses do not exceed the specified working stresses by
more than 5 per cent. The permissible shear and bearing stresses on rivets may be increased by 10 per
cent."
4.0 ADEQUACY OF BEARINGS OF STANDARD SPANS
The adequacy of existing roller-rocker bearings of standard open web girder span of BGML & RBG
loadings have been checked by following above provisions and the results are shown in Table-1 (a) & (b).
The analysis has been done by taking into consideration wind load of 100 kg/m2 alongwith live load as per
A&C Slip No.34 of Bridge Rules with and without 25% dispersion of longitudinal forces as per clause
Table-1 Adequacy of Bearings of Existing OWG Spans for 25t Axle Load
S.N.

Span

STRENGTHENING REQUIRED AT FIXED END

RBG STANDARD
BGML STANDARD

(a) With 25% Dispersion of Longitudinal Forces

1

30.5
(through type)

45.7
(through type)

61.0
(through type)

76.2
(through type)

Existing bearing safe.

2 Number additional anchor bolts to be

provided and saddle bolts be replaced by
higher strength bolts of property class 6.6.

2 Number additional anchor bolts to

be provided and saddle bolts be
replaced by higher strength bolts of
property class 6.6.
2 Number additional anchor bolts be
provided

- do -

- do -

2 Number additional anchor bolts be 2 Number additional anchor bolts be

provided
provided.

(b) Without Dispersion of Longitudinal Forces

30.5
(through type)

45.7
(through type)

61.0
(through type)

76.2
(through type)

2 Number additional anchor bolts to

be provided and saddle bolts be
replaced by higher strength bolts of
property class 6.6.

4 Number additional anchor bolts to be

provided and saddle bolts be replaced by
higher strength bolts of property class 8.8.
Overstressing on Bed Block is less than
10% and may be permitted on condition
basis.
4 Number additional anchor bolts to 4 Number additional anchor bolts to be
be provided and saddle bolts be provided and saddle bolts be replaced by
replaced by higher strength bolts of higher strength bolts of property class 6.6.
property class 6.6.
2 Number additional anchor bolts to 2 Number additional anchor bolts to be
be provided and saddle bolts be provided and saddle bolts be replaced by
replaced by higher strength bolts of higher strength bolts of property class 6.6.
property class 6.6.
2 Number additional anchor bolts be
-doprovided

262

2.8.3.2 amended vide A&C Slip No.22 of Bridge Rules. The adequate anchorage of welded rails on
approaches has to be ensured accordingly. The checking of these bearings has been done keeping in view,
the speed for which the standard span has been cleared by RDSO vide letter No.CBS/Golden/Q/Strength
dated 08-08-06.
5.0 CONCLUSION
5.1 It is seen that the design of existing roller-rocker bearings is governed significantly by longitudinal forces.
Forces on bearing are considerably reduced by allowing dispersion of longitudinal forces. Railways may
thus ensure proper anchorage of track on approaches to take advantage of clause 2.8.3.2 of Bridge
rules in this respect.
5.2 The existing bearings can be safely retained by changing saddle bolts with higher class (Property Class
6.6 or 8.8) and providing extra anchor bolts as found necessary.
5.3 As the longitudinal forces are to be resisted at the fixed end only, no action is required to be taken at
roller end, provided the physical condition is other satisfactory.
6.0 REFERENCES
6.1

IRS Steel Bridge Code (1962), Research Designs & Standards Organisation, Ministry of Railways, Lucknow (U.P.).

6.2

IRS Bridge Rules (1964), Research Designs & Standards Organisation, Ministry of Railways, Lucknow (U.P.).

6.3

IRS Drawings and Design Documents of relevant Standard Spans, Research Designs & Standards Organisation,
Ministry of Railways, Lucknow (U.P.).

6.4

RDSO letter CBS/Golden/Q/Strength dated 08-08-2006 with Annexures.

x x x

263

EFFECT OF INCREASING SPEED ON DYNAMIC

IMPACT - AN ANALYTICAL STUDY ON STANDARD
STEEL GIRDER BRIDGES
Ravindra Kumar Goel*
(Published in IRICEN Journal of Civil Engineering, Sept., 2009)

1.0 INTRODUCTION
Dynamics of railway bridges involves the response of bridges to the movement of vehicles and to the
influence of a number of parameters which increase dynamic strains or stresses. The most important parameters
influencing the dynamic stresses in railway bridges are the frequency characteristics of bridge structures (i.e.,
the length, mass, and rigidity of individual members), the frequency characteristics of vehicles (i.e., the sprung
and unsprung masses, the stiffness of springs), the damping in bridges and in vehicles, the velocity of vehicle
movement, the track irregularities, and so on. The vehicles affect the bridges not only by vertical forces, but
also by movements which generate longitudinal and transverse horizontal forces. The coefficient of dynamic
augment which is used in IRS Bridge Rules is a function of loaded length only and does not represent the
actual dynamic impact of moving train loads at high speeds. The paper presents the results of an analytical
study undertaken on standard railway bridges which suggested that dynamic analysis of bridges is required to
be done for design of bridges for high speeds.
2.0 IMPORTANCE OF DYNAMIC EFFECTS
2.1 With more and more high-speed railways being built in the world, more and more emphasis has been
given on the subject of dynamic interaction of vehicles and bridges. On the one hand, the train running
with high speed induces dynamic impact on the bridge structures, influencing their working state and
service life. On the other hand, the vibration of the bridge in turn affects the running stability and safety
of the train, and thus becomes an important factor for evaluating the dynamic parameters in bridge
design. Therefore, in many countries, the dynamic behaviors of bridges have been systematically studied
in the development of high speed railway.
2.2 The railway bridges subjected to high speed trains provide intensive vibration similar to the resonance
phenomenon. The resonance occurs if the frequency of an input force coincides with one of the natural
frequencies of the system. The resonant vibration of railway bridges results in the deterioration of passenger
comfort, reduction of traffic safety (a possibility of derailment of vehicles), the destabilization of ballast
(higher maintenance costs) and increased damage in the bridge system from fatigue considerations.

*Director/Steel Bridges-I, Bridges & Structures Directorate, Ministry of Railways, RDSO, Lucknow (UP)-226011.

264

2.3 The dynamic effects of vehicles on railway bridges as shown in Fig. 1, [2] highlight the cause and effect
relationships that exist. It may be appreciated that the problems to be addressed are multi-disciplinary in
nature and require a reasonable degree of interaction between the executing agencies and the research
people. The codal provisions of other countries can not be directly adopted without quantifying the
parameters in indian context and verifying the relationship for Indian condition.
Dynamic effects of railway vehicles
on bridges

Influence of
moving forces
and masses

Of the
track

Deterministic

Stochastic
also due to traffic flow

Vertical effects

Horizontal effects

rregularities

Of
vehicle
wheels

Starting
forces

Longitudinal

Braking
forces

Transverse

Lateral
impacts

Centrifugal
forces

Fig. 1 Dynamic effects of railway vehicles on bridges

2.4 In practice the design of railway bridges includes the dynamic effect of the moving load by increasing the
live load by an impact factor. However recent codes of practice address the problem by giving the full
recognition to the fact that resonance in bridges may occur depending upon the characteristics of the
train, its passage and the bridge characteristics. Considerable research has been done elsewhere on the
subject of dynamic interaction of vehicles and bridges. However, significant work has not been done in
India on existing bridges to confirm their adequacy for high speed trains.
2.5 Under current design practice, the impact factor is considered a function of length only. However,
dynamic response depends on a number of factors, including vehicle properties, bridge characteristics,
and pavement roughness [9, 10]. Although designs that comply with current codes may satisfy safety
and strength requirements, they may also greatly underestimate actual bridge response in many cases
265

[11, 12, 13, 14]. Consequently, some bridge structures may suffer distress as a result of unexpected
dynamic response. For example, stresses generated by heavy vehicles moving at high speeds over a
rough bridge deck may greatly exceed those predicted by incrementing static live loads by a dynamic
impact factor prescribed in bridge codes. Existing analysis and design procedures do not always predict
these unexpected and undesirable results.
3.0 CODAL PROVISIONS FOR DYNAMIC EFFECTS OF TRAIN LOADS
3.1 Steel Bridge Code and Bridge Rules in India
3.1.1The present design of railway bridges is done in accordance with guidelines and provisions existing in
Steel Bridge Code [22] and Bridge Rules [23]. In design practice, dynamic effect of the moving load is
taken care of by increasing the live load by impact factor or dynamic augmentation factor or dynamic
coefficient. This factor depends on many parameters like the type of loading, speed, type of structure,
material of structure, loaded length etc. But for simplicity an impact factor is specified by the Bridge
Rules in India involving only one parameter, i.e., the loaded length. All the other parameters are considered
as constants in the expression for impact factor. For Broad Gauge and Meter Gauge steel railway
bridges carrying a single track, the Coefficient of Dynamic Augment (CDA) is given by the following
expressionCDA = 0.15 +

8
6 L

, subject to a maximum of 1.0

Where L is defined as given below:

a)

L is loaded length of the span in meters for the position of the train giving the maximum stress in the
member under consideration. For the design of chord members, it will be the whole span of the
truss and for the web members only part of the span is to be loaded.

b) L is taken as 1.5 times the cross-girder spacing for finding stresses in the stringers (rail-bearers).
c)

L is taken as 2.5 times the cross-girder spacing for finding moments in the cross-girders (floorbeams).

3.1.2It is important to note that the expression for Co-efficient of Dynamic Augment (CDA) was proposed in
1981 based on actual field observations made on existing bridges. The formula is applicable for speeds
upto 160 kmph on Broad Gauge and 100 kmph on Meter Gauge for passenger trains. It is apparent that
the effect of higher speeds is not reflected in this expression.
3.2 BS 5400-2: 1978
In this standard [24], dynamic effects are considered in clause 8.2.3. Here equivalent static loadings
(RU and RL loading) are multiplied by appropriate dynamic factors to allow for impact, oscillation and other
dynamic effects including those caused by track and wheel irregularities.

266

Table 2.1 Dynamic factor for type RU loading

Dimension L
(m)

Dynamic factor for evaluating

Bending Moment

Shear

2.00

1.67

upto 3.6
from 3.6 to 67

2 .16
0 .82 +
( L 0.2)

0.73 +

over 67

1.00

1 .44
( L 0 .2)
1.00

The dynamic factor for RU loading applies to all types of track and is given in Table 2.1 (Table no. is
with reference to BS:5400 - 2)
In deriving the dynamic factor, L is taken as the length (in m) of the influence line for deflection of the
element under consideration. For unsymmetrical influence lines, L is twice the distance between the point at
which the greatest ordinate occurs and the nearest end point of the influence line. In the case of floor members,
3 m should be added to the length of the influence line as an allowance for load distribution through track.
The dynamic factor for RL loading, for evaluation of moments and shears, shall be taken as 1.20,
except for unballasted tracks where, for rail bearers and single track cross girders, the dynamic factor shall be
increased to 1.40.
3.3 EN 1991-2: 2003(E)
3.3.1In European standard [25] dynamic effects (including resonance) are considered in clause 6.4. In this
code dynamic effects are taken care of in a better way. A static analysis shall be carried out with the load
models (Load Model 71 and where required Load Models SW/0 and SW/2). The results shall be
multiplied by the dynamic factor, f to consider the dynamic effects. Generally the dynamic factor f is
taken as either f2 or f3 according to the quality of track maintenance as follows:
(a)

For carefully maintained track:

(b)

1.44
L
0. 2

0.82 , with
1.00

1.67:

2.00

For track with standard maintenance:

2.16
L 0 .2

0.73 , with: 1.00

Where Lf is ''Determinant'' length (in m.).

The dynamic factor f shall not be used with:
- the loading due to Real Trains.
- the loading due to Fatigue Trains.
267

- the load model HSLM.

- the load model ''unloaded train''.
3.3.2The requirements for determining whether a dynamic analysis is required are shown in Fig. 6.10
(Figure no. is with reference to Euro Code)
Where:
V is the Maximum Line Speed at the site (km/h)
L is the span length
h0 is the first natural bending frequency of the bridge loaded by permanent actions (Hz)
hT is the first natural torsional frequency of the bridge loaded by permanent actions (Hz)
v is the Maximum Nominal Speed (m/s)
(v/h0)lim is given in annex F (EN 1991-2: 2003(E))

Fig. 6.10 - Limits of bridge natural frequency n0 [Hz) as a function of L(m)

3.3.3Quasi static methods which use static load effects multiplied by the dynamic factor f are unable to
predict resonance effects from high speed trains. Dynamic analysis techniques, which take into account
the time dependant nature of the loading from the High Speed Load Model (HSLM) and Real Trains
(e.g., by solving equations of motion) are required for predicting dynamic effects at resonance.
268

Fig. 2 Flow chart for determining whether a dynamic analysis is required

( EN 1991-2: 2003(E), page no 77)

269

3.3.4The standard [25] gives a flow chart (Figure 2) for determining whether dynamic analysis is required.
This chart cannot be used directly for bridges on Indian Railways as the procedure given involves
parameters which are meant for train given in European Load Models. As per this flow chart for
continuous girder dynamic analysis is not considered necessary if design speed is less than 200 kmph.
4.0 MATHEMATIC MODELING OF MOVING LOAD
4.1 Mondal [28] has obtained the response of a load train by assembling the responses of the different point
loads Pk (Fig. 3). The differential equation corresponding to mode 'n' is given as under:
k

M n qn ( t ) 2

n M n qn (t )

2
n

M n q n (t )

Pk

(vt d k )

k 0

Fig. 3 Train load travels across the beam

4.2 Software package SAP2000 (Ver. 10.0.5) has been used to validate the results of mid span dynamic
displacement with respect to time for a simply supported beam which is subjected to a single point
wheel load traversing from one end to other end. The results are shown in Fig. 4.

Fig. 4 Comparison between moving load analysis and SAP2000 results

270

4.3 Moving load analysis of standard spans

Two standard plate girder bridges and two standard welded truss bridges have been considered as per
details given below:
a)

12.2 m Span Plate Girder Bridge:

It consists of two girders each having 2 Flange plates 550 x 25, 1 Web plate 1250 x 10
Clear Span = 12.2 m
Effective Span = 13.1 m

b)

24.4 m Span Plate Girder Bridge:

It consists of two girders each having 2 Flange plates 620 x 45, 1 Web plate 1980 x 14
Clear Span = 24.4 m
Effective Span = 25.6 m

c)

30.5 m Span Welded Truss Bridge:

Clear Span = 30.5 m
Effective Span = 31.926 m

d)

61 m Span Welded Truss Bridge:

Clear Span = 61 m
Effective Span = 63 m

4.4 Train definitions

For analyzing all the above mentioned bridges the moving load analysis has been carried out by the train
descriptions. Trains are classified on the basis of their usage like passenger and goods train. For the present
study 10 Modified Broad Gauge trains (which are defined as MBGT1 to MBGT10) and 12 Modified Broad
Gauge trains for HM routes (which are defined as HMT1 to HMT12) have been considered as per Bridge
Rules. Trains are modeled as a series of concentrated axle loads moving across the bridge.
4.5 Moving load analysis and dynamic analysis
After completing the modeling of all type of bridges, for all type of trains, MBGT1 to 10 and HMT1 to
12 moving load analysis and dynamic analysis are performed in SAP2000 v 10.0.5.
4.5.1In moving load analysis multi-step static analysis is performed to get the bending stress histories at
midspan of girder bridges and combined stress histories (axial and bending) of different members of
truss bridges considering coefficient of dynamic augment as per Bridge Rules [23].

271

4.5.2The parameters which are involved in moving load analysis mainly are (a) duration of loading i.e. the
time required of a train to pass over the bridge (b) time step for discretization of the load. For multi-step
static analysis speed of the train (v) and time step (dt) are considered as 10 m/sec and 0.05 sec respectively
for finding the stress histories. The minimum spacing between two wheels (lw), considering all the trains
is 1.7 m. For getting all the cycles in the stress history time step should be less than (lw/v).
4.5.3Dynamic analysis is performed by linear direct-integration time-history analysis for different velocities of
train considering zero percentage (0 %) and two percentage (2 %) damping ratio. From modal analysis
time periods of two vertical modes are taken for defining mass and stiffness proportional damping.
4.5.4The main important parameters for dynamic analysis are (a) velocity (b) time step
4.6 Results and discussions
4.6.1The midspan displacement responses of plate girder bridges for 0.0 % damping and 2.0 % damping due
to passage of MBGT5 with respect to the change of velocities are shown in Figs. 5 and 6, considering
time step as lw/v.
4.6.2The midspan displacement responses of welded truss bridges for 0.0 % damping and 2.0 % damping
due to passage of MBGT5 and HMT8 with respect to the change of velocities are shown in Figs. 7, 8,
9 and 10 considering time step as lw/v.
4.6.3Also the dynamic response is compared with the response of moving load analysis i.e. multi-step static
with coefficient of dynamic augment. When the response of moving load analysis crosses the response
of dynamic analysis then the corresponding velocity is named as cut-off velocity. Table 1 shows the cutoff velocities of all the bridges due to passage of MBGT5 and HMT8
Table 1 Cut-off velocity of Girder and Truss Bridges due to passage of
MBGT5 and MBGT8
Type of
Bridge

Type of Train
MBGT5

Truss
HMT8
Girder

MBGT5

Cut-off Velocity (m/sec)

Effective Span
(m)

Damping ratio = 0.0%

Damping ratio = 2.0%

31.926

312.67

320

63

233.65

240.76

31.926

85.89

557.22

63

286.26

296.49

13.1

298.25

428.89

25.6

269.87

276.47

272

Fig. 5 Variation of midspan displacement due to passage of MBGT5 on 12.2 m

Plate Girder Bridge with respect to velocity

Fig. 6 Variation of midspan displacement due to passage of MBGT5 on 24.4 m

Plate Girder Bridge with respect to velocity

273

Fig. 7 Variation of midspan displacement due to passage of MBGT5 on 30.5 m Welded Truss
Bridge with respect to velocity

Fig. 8 Variation of midspan displacement due to passage of MBGT5 on 61.0 m

Welded Truss Bridge with respect to velocity

274

Fig. 9 Variation of midspan displacement due to passage of HMT8 on 30.5 m Welded Truss
Bridge with respect to velocity

Fig. 10 Variation of midspan displacement due to passage of HMT8 on 61.0 m

Welded Truss Bridge with respect to velocity

275

5.0 CONCLUSION
1.

Mid span displacement on standard spans as per dynamic analysis are varying with speed.

2.

The static analysis is adequate upto a cut off velocity which is generally more than 160 kmph. Dynamic
analysis is required for different types of spans for permitting higher speeds.

3.

The cut off velocity is different for different type of bridges and length of spans for a standard type train.

4.

The cut off velocity is different for different types of trains in a load model.

5.

Cut off velocity can be increased by providing external damping devices on the bridge.

6.

For present speeds (i.e. upto 160 kmph) CDA based static analysis is safe.

6.0 REFERENCES
1.

Paul D.K., Pandey A.D., Bhargava Ashish and Tadigotla Sashidhar, ''Rationalisation of Fatigue Criteria for Railway
Bridges in India'', Bridge Engineering - Some Issues of Research Interest, Edited by Krishna P.-Professor of Bridge
Engineering, Railway Chair, Department of Civil Engineering, IIT Roorkee, Roorkee- 247667, India, June- 2002.

2.

Fryba L. (1996), "Dynamics of Railway Bridges", Thomas Telford Services Ltd, London, UK, 1st Ed.

3.

Xia H. and Zhang N. (2005), "Dynamic analysis of railway bridge under high-speed trains", Computers and Structures,
Vol. 83, pp 1891-1901.

4.

Xia H., Zhang N., De Roeck G. and Maeck J. (2003), "Experimental analysis of high-speed railway bridge under Thalys
trains", International Journal of Sound and Vibration, Vol. 268, pp 103-113.

5.

Xia H., Zhang N. and De Roeck G. (2003), "Dynamic analysis of high speed railway bridge under articulated trains",
Computer and Structures, Vol. 81, pp 2467-2478.

6.

Xia H., Zhang N. and Gao R. (2005), "Experimental analysis of railway bridge under high-speed trains", International
Journal of Sound and Vibration, Vol. 282, pp 517-528.

7.

Fryba L (2001), "A rough assessment of railway bridges for high speed trains", Engineering Structures, Vol. 23, pp
548-556.

8.

Wang T. L. (1993), "Impact in a railway truss bridge", Computer and Structures, Vol. 49, pp 1045-1054.

9.

Chan H.T. and O'Connor C. (1990a), "Vehicle Model for Highway Bridge Impact", Journal of Structural Engineering,
Vol. 116, No. 7, pp 1772-1793.

10.

Kou, J.K. and DeWolf J.T. (1997), "Vibrational Behavior of Continuous Span Highway Bridge-Influencing Variables",
Journal of Structural Engineering, Vol. 123, No. 3, pp 333-344.

11.

Chang D. and Lee H. (1994), "Impact Factors for Simple-Span Highway Girder Bridges", Journal of Structural
Engineering, Vol. 120, No. 3, pp 704-715.

12.

Huang D., Wang T.L. and Shahawy, M. (1992), "Impact Analysis of Continuous Multigirder Bridges Due to Moving
Vehicles", Journal of Structural Engineering, Vol. 118, No. 12, pp 3427-3443.

13.

Inbanathan M.J. and Wieland M. (1987), "Bridge Vibrations Due to Vehicle Moving Over Rough Surface", Journal of
Structural Engineering, Vol. 113, No. 9, pp 1994-2008.

14.

Yang Y.B., Liao S.S. and Lin B.H. (1995), "Impact Formulas for Vehicles Moving Over Simple and Continuous Beams",
Journal of Structural Engineering, Vol.121, No. 11, pp 1644-1650.

276

15.

Chan H.T. and O'Connor C. (1990b), "Wheel Loads From Highway Bridge Strains: Field Studies", Journal of Structural
Engineering, Vol. 116, No. 7, pp 1772-1793.

16.

Fafard M., Laflamme M., Savard M. and Bennur M. (1998), "Dynamic Analysis of Existing Continuous Bridge",
Journal of Bridge Engineering, Vol. 3, No. 1, pp 28-37.

17.

Wang T.L., Huang D. and Shahawy M. (1992), "Dynamic Response of Multigirder Bridges", Journal of Structural
Engineering, Vol. 118, No. 8, pp 2222-2238.

18.

Yang Y.B. and Lin B.H. (1995), "Vehicle-Bridge Interaction Analysis by Dynamic Condensation Method", Journal of
Structural Engineering, Vol. 121, No. 11, pp 1636-1643.

19.

Cao L.J., Allen J.H., Shing P.B. and Woodham D. (1996), "Behavior of RC Bridge Decks With Flexible Girders", Journal
of Structural Engineering, Vol. 122, No. 1, pp 11-19.

20.

Huang D., Wang T.L. and Shahawy M. (1995), "Vibration of Thin-Walled Box-Girder Bridges Excited by Vehicles",
Journal of Structural Engineering, Vol. 121, No. 9, pp 1330-1337.

21.

Humar J.L. and Kashif A.H. (1995), "Dynamic Response Analysis of Slab-Type Bridges", Journal of Structural
Engineering, Vol. 121, No. 1, pp 48-61.

22.

Government of India Ministry of Railways (Railway Board). ''Indian Railway Standard Code of Practice for the Design
Steel or Wrought Iron Bridges Carrying Rail Road or Pedestrian Traffic (Steel Bridge Code)'', Adopted - 1941, Revised
- 1967.

23.

Government of India Ministry of Railways (Railway Board). ''Bridge Rules - Rules specifying the loads for design of
super-structure and sub-structure of bridges and for assessment of strength of existing bridges'', Adopted - 1941,
Revised - 1964.

24.

BS 5400-2: 1978, Steel, concrete and composite bridges-Part 2: Specification for loads.

25.

EN 1991-2 : 2003 (E), Euro code 1: Actions of structures- Part 2: Traffic loads on bridges.

26.

Chopra A.K. (2005), "Dynamics of Structures - Theory and Applications to Earthquake Engineering", Prentice-Hall of
India Private Limited, New Delhi, India, 2nd Ed.

27.

Goicolea J.M., Dominguez J., Navarro J.A. and Gabaldon F. (2002), "New Dynamic Analysis Methods for Railway
Bridges in Codes IAPF and Eurocode- 1", Computational Mechanics Group, Universidad Politecnica de Madrid,
Spain.

28.

Mondal S.B. (2008), 'Influence of Train Speed on Fatigue Performance of Railway Steel Bridges", Dept. of Earthquake
Engineering, IIT/Roorkee.

x x x

277

TECHNICAL PAPERS ON BRIDGE PERFORMANCE,

IXTH INTERNATIONAL HEAVY HAUL CONFERENCE
HELD AT SHANGHAI, CHINA - A REVIEW
R.K. Goel*
(Published in IRICEN Journal of Civil Engineering, Dec., 2009)

1.0 GENERAL
Heavy haul technology is regarded as an important concept for the freight transportation development
of the world and it is considered as one of the important signs of railway modernization. Railway heavy haul
started in the 1950s, and become an essential transport method for bulk freight in many countries, because of
its comparative advantages such as large capacity and high efficiency. The breakthroughs, particularly in
aspects of traction power, high capacity and light weight evolution of wagons, synchronous control technology,
braking technology, etc., have boosted development of heavy haul transportation since 1980s. IXth International
Heavy Haul Conference, IHHC-2009 was successfully held in Shanghai through the joint efforts of Ministry
of Railways, People's Republic of China and International Heavy Haul Association. This technical paper
summarizes the technical papers presented on the performance of Railway bridges due to heavy haul.
2.0 TECHNICAL PAPERS ON BRIDGE PERFORMANCE
The conference was attended by more than 300 eminent persons, including representatives of railway
researchers, technical and management personnel from the USA, South Africa, Australia, Sweden, India,
Russia, Brazil, Germany, China etc., as well as representative of suppliers and manufacturers of railway
products and railway operators. A number of technical papers were presented on bridge performance and
one of the technical session was exclusively for Bridges. The Summary of technical papers presented in the
conference on bridge performance along with salient aspects covered and important conclusions drawn are
given herewith.
2.1 Analysis on Dynamic Characteristics of Bridges with 32-m Pre-stressed Concrete Simple Supported
Beam under Wagon Impact by Jianzhen XIONG, Manmang GAO & Jingchuan YAO
2.1.1 Summary
A super-long bridge on Datong-Qinhuangdao Line has single-track, pre-stressed concrete simple
supported beams and 38 arches of 32m span. The pier is circular with variable cross section and 15.6 20.6m high. Testing data shows: the weak lateral stiffness of pier causes lateral vibration responses of the
deck and pier exceeding the safe threshold in the Bridge Examination Code of China Railways when freight

*Director/SB-I, B&S Dte., Research Designs & Standards Organization, Lucknow

278

trains pass the bridge. In view of this, this paper establishes a train-bridge simulating calculation model and
applies coupling vibration analysis method to study train and bridge's vibration responses in case of two
reinforcement solution when a heavy-haul train(composed of C80) passes at 60 ~ 120 km/h with regard to
the two typical piers of 32m simple supported beam, the paper makes further analysis concerning effect of
pier's lateral stiffness on the train (composed of C80) running. In this way, the paper provides technical
support for the operational safety and pier reinforcement to operate 20000t-level super-long train on DatongQinhuangdao Line.
2.1.2 Salient aspects:
1.

It was found that the vibrations in lateral direction exceed the threshold values specified in Bridge
Examination Code of China Railways.

2.

Train-bridge simulating calculation model was developed using Coupling Vibration Analysis Method.

3.

Two pier reinforcement solutions were studied with a heavy haul train (composed of C-80) passing
at 60 - 120 kmph w.r.to piers of 32m simple supported beam.

4.

Train vibration responses were also studied to consider the effect of piers lateral stiffness on the
train.

2.3 Important Conclusions:

1.

Before the pier reinforcement, the lateral vibration of bridge and pier are comparatively larger. The
running speed of C80 heavy-haul train should not be exceed 60 km/h, if the safe limit in the Bridge
Examination Code of China Railway is satisfied.

2.

The train-bridge coupling calculation before pier reinforcement indicates that the weak pier lateral
stiffness is the main cause of the large lateral vibration of bridge.

3.

The reinforcing method of pier top lateral connection can reduce the pier lateral vibration effectively,
and enhance the maximum running speed of C80 heavy-haul train to 75 km/h after reinforcement
from the previous under 60 km/h before the reinforcement.

4.

When the pier height is less than 16m, the arrangement of double column piers can reduce the
bridge laters vibration effectively.

2.2 Discussion on Bridge Reinforcement Measures under Heavy Haul Conditions on DatongQinhuangdao Line by Xurong WANG & Junjian ZHANG
2.2.1 Summary
Since 2005, Datong-Qinhuangdao Line operated 20,000 t-level heavy haul trains, with axle load raised
from 21t up to 25t and annual traffic volume reaching 200 million tons. At present, the line faces the challenges
of increasing running speed of empty trains to 100 km/h, super-long and high-speed load creates great impact
and destruction on bridges. Considering the problem of 32m girder on Datong-Qinhuangdao Line, such as
weak lateral connection, lateral vibration exceeding limitation, crack of diaphragm, the paper comes up with
279

measures for lateral prestressed reinforcement of girders after theoretically and calculation on the basis of
scientific detection of the bridges. In addition, the paper explains the quality assurance measures for strength
and elasticity modulus of C50 high-performance concrete that needs control, as well as the uncurling technology
of steel strands with small shrinkage. In the end, the paper provides feasible suggestion on bridge reinforcement
solution under heavy-haul transport condition.
2.2.2 Salient aspects:
1.

Axle load has been raised from 21 to 25 t since 2005.

2.

Challenge of running empty train to 10 kmph.

3.

Super long and high speed load creates heavy impact and damage to bridges.

4.

Problem of weak lateral connection of girder with pier, lateral vibration exceeding limitation, crack
of diaphragm.

5.

Bridges were studied scientifically and reinforcement in the manner of lateral prestressing of girders
suggested.

2.2.3 Important Conclusions:

1.

Lateral prestressing of PSC girders proposed.

2.

Quality assurance measures explained for strength and elasticity modulus of C-50 high performance
concrete.

2.3 Experimental Study on Lateral Dynamic Behaviors of Bridges with 32m Prestressed Concrete Girder
on Datong-Qinhuangdao Line by Yiqian YANG, Penghui LIU, Jingchuan YAO, Jing YIN, Zaitian KE &
Zhengshen DONG
2.3.1 Summary
This paper makes a comprehensive analysis on lateral dynamic behaviors of bridges with 32m prestressed concrete girders on Datong-Qinhuangdao Line based on measured data about lateral vibrations of
29 arches of 32m girders and 36 pier on 11 bridges. The results show that the lateral vibration of 32m is
mainly caused by the pier's vibration. Snake movement of trains and lateral swing vibration of bogies affect
lateral vibration depending on running speed, pier height and train load. For rounded shape and doublecolumn piers with height at 10 - 20m, the lateral amplitudes exceed ordinal values of "the Bridge Examination
Code of China Railways". It is suggested to upgrade or abandon C62, C62A, C62B, C62N, C63, C63A,
C64 wagons, reinforce rounded shape and double-column piers with height at 10 - 20m. The lateral dynamic
behaviors of piers improved remarkably after reinforcement.
2.3.2 Salient aspects:
1.

Analysis of lateral dynamic behavior of bridges with 32m PSC girders.

2.

Pier vibration is primarily responsible for lateral vibration of PSC girders.

280

3.

Lateral swing vibration of bogies and snake like movement of train.

4.

Lateral vibrations are dependent on running speed, pier height and train loads.

5.

Lateral amplitude found to exceed ordinary values of the "Bridge Examination Code".

2.3.3 Important Conclusions:

1.

It has been suggested to either upgrade or abandon C62, C62A, C62B, C62M, C63, C63A,
C64 wagons.

2.

Reinforce rounded shaped and double column piers with height 10m to 20m to control pier
vibrations lateral dynamic behavior improved remarkable after reinforcement.

2.4 Fatigue Life Prediction of Prestressed and Reinforced Concrete Bridges by R.M. DE Souza,
R. L. Queiroz, S.L.A. Lobato, R.A.C. Sampaio & A.A. de Azevedo
2.4.1 Summary
Evaluation of fatigue damage of the elements of a structure is essential for determining its life expectancy,
and is of great interest to both practitioners and researchers. This is especially true for structures subjected to
varying loads with great intensity and with a high number of cycles, such as Railway bridges. This paper
presents a methodology, and corresponding computer implementation, for estimating fatigue life of reinforced
and/or prestressed concrete girders and columns of railway bridges, following the recommendations of CEBFIP model code 90. Based on the concepts of influence line according to Muller-Breslau Principle, the
development computer program first determines the variation of the internal forces in a given section of a
girder or column, due to the passage of different train types, and considering an appropriate impact coefficient.
Starting from the temporal series of the internal forces, another module of the program performs a nonlinear
biaxial analysis of the section of interest, using a fiber-section model, and then computers the strain and
corresponding stress variation in several components of the section (concrete, reinforcing/bars, prestressing
cables, etc.). As the amplitudes of the stress cycles are not constant, Rain-flow algorithm is used to count
stress cycles, and appropriate S-N curves for concrete and steel are used to obtain the damage caused by
each cycles. Miner's rule is then employed to determine the complete damage caused by the entire load
block, which corresponds to the passage of a pier of trains (in the loaded and un loaded condition). Depending
on the damage caused by each load block, and the number of daily passage, the program is able to compute
life expectancy in years. as an application example, the program has been used to evaluate, under current and
future traffic conditions, the life expectancy of some structural members of a bridge of Carajas Railway, a
heavy haul line located in Brazil. The bridge under study is formed by simply supported prestressed concrete
beams with as span length of about 30m, supported on reinforced concrete columns. The obtained results
revealed that the fatigue life expectancy for the analyzed columns of this bridge is satisfactory, and that the
proposed methodology is appropriate for this type of study.

281

2.4.2 Salient aspects:

1.

Paper presents a methodology and corresponding computer implementation for estimating fatigue
life of RCC/PSC girders and columns of railway bridges.

2.

Methodology is based on concepts of influence line according to Muller -Breslau Principle.

3.

In a given section of bridge the program determines the variation of the internal forces due to
passage of different train types, considering an appropriate impact coefficient.

4.

Another module of the programme performs non-linear bi-axial analysis of the section of interest
and computes the stress variations in several components of the section i.e. concrete, reinforcing
bar, prestressing cables, etc.

5.

As the amplitudes of the stress cycles are not constant, rainflow algorithm is used to count cycles.

6.

Appropriate S-N curves are used for concrete and steel to obtain damage caused by each cycle.

7.

Miner's Rule is then applied to determine the complete damage caused by the entire load block
(load model).

8.

Depending upon the damage caused by each block, the number of daily passages, the programme
is able to compute life expectancy in year.

2.4.3 Important Conclusions:

1.

The obtained results of fatigue life expectancy for the analyzed columns of the bridge were found
satisfactory.

2.

Influence of sequential arrangement (disposition of locomotives & freight cars) has a very small
effect on fatigue life expectancy.

3.

For PSC/RCC bridges since service stresses due to life loads are very low, the damage due to
fatigue is small (practically negligible).

2.5 Research on Adoptability of Existing Railway Bridge to Heavy Haul Transport in Respect of
Strength and Fatigue by Yuling ZHANG
2.5.1 Summary
This paper introduces the basic conditions of design of existing domestic railway bridges. In view of the
three possible solutions after separation of freight and passenger traffics, i.e. (1) 25t axle load; (2) 26.5t axle
load (80t-level load), 7.58t/m uniform load; (3) 300 kN axle (95t-level load), 10 t/m uniform load, the paper
analyses exiting railway bridges on common lines and sidings through assessing the adaptability of design
carrying capacity to strength; considering possible annual traffic volume, make analysis of influence on bridge
fatigue. The research results show that bridges shall be reinforced according to infinity fatigue life principle to
operate heavy-haul trains on existing lines.

282

2.5.2 Salient aspects:

1.

Foresee separation of existing mixed traffic lines into separate lines for traffic and passengers.

2.

In case the existing lines are converted into traffic lines with heavier axle loads, the adaptability of
existing track and bridges is to be seen. Bridges have been focused.

3.

Bridge are segregated w.r.to their design standards. Most are of Zhong live load standard. Such
as ours e.g. BGML, RBG, MBG.

4.

Potential capacity analysis of existing design state was done in view of the fact that larger impact
coefficient was used in old designs for steam engines.

5.

Carrying capacity analysis were done separately for steel bridges and concrete bridges.

2.5.3 Important Conclusions:

1.

The results shows that the bridges shall be reinforced according to infinite fatigue life principle to
operate heavy haul trains on existing lines.

2.

The effective method to enhance applicability of bridge to heavy haul transportation is to determinate
fatigue limit of different fatigue detail by testing research.

3.

Research the fatigue design method and reinforced design based on infinite life theory.

4.

The research above was stated to be based on design state of bridge. Final precision conclusions
should be determined by analyzing the practical condition of real bridge and railway track.

2.6 Static and Dynamic Analysis of a Reinforced Concrete Bridge along the Carajas Rail Road by
Ritermayer M Teixeira & Denio & R.C. de Oliveira
2.6.1 Summary
This work presents results of static and dynamic analysis carried out on 503m length bridge along
Carajas railroad crossing Cajuapara river in Brazil. The bridge's behavior became important due to some
changes in the static and dynamic acting loads once the future train will be heavier, becoming necessary a
more accurate analysis to able a efficient strengthening design. Thus, a mathematic model was developed
using the finite element method basis to evaluate its behavior and an "in loco" structural monitoring to get
experimental data as strains and vibrations under train passage. Several accelerometers and strains gauges
were placed along the slab, beams and columns. Numerical and experimental results are shown and the main
conclusions validating the applied methodology are presented as well.
2.6.2 Salient aspects:
1.

Bridge consisting of 22 statically determinate spans of 25m each with overall length of 593, was
analyzed statically and dynamically.

2.

Bridge was modeled using SAP 2000 standard software module, and finite elements.

3.

Strains and vibrations were measured to monitor slabs under rolling loads.
283

4.

The measured/observed dynamic effects were 41% of the calculated values.

2.6.3 Important Conclusions:

1.

Experimental dynamic and static effects were very close for the monitored longitudinal beam.

2.

Interference of train's mass was clear and recommended to be considered for future computational
analysis.

3.0 CONCLUDING REMARKS:

The review of technical papers on bridge performance has been done to enhance the awareness of
Indian Railways personnel about the actions being taken in the world to analyze/strengthen the existing bridges
for heavy haul and to make use of their findings for possible application on Indian Railway Bridges.

x x x

284

Seismic Isolation of Bridges

R.K. Goel*
(To be published in Indian Railway Technical Bulletin, RDSO, Lucknow)
SYNOPSIS

rhoz HkwdEih; lEHkkouk okys {ks=ksa esa] visf{kr HkwdEifujks/kh iqyksa ds vfHkdYi ds fy,s HkwdEi foyxu ;qfDr;ksa
dk iz;ksx iqyksa dh ykxr dks de dj ldrk gSA HkwdEih; tkZ dk rsth ls {kj.k djus ds fy,s Hkh dqN
;qfDr;k iz;ksx dh tkrh gSa] ftlls iqyksa esa gksus okyh HkwdEih; {kfr dks de fd;k tk ldrk gSA igys ls cus
gq,s iqyksa esa bu ;qfDr;ksa dk iz;ksx djus ls iwoZ] mu iqyksa dh fLFkfr o fuekZ.k lEcU/kh tkudkjh dk leqfpr
v;;u djuk vko';d gSA bl rduhdh ys[k esa blh lEcU/k esa vko';d tkudkjh nh x;h gSA
Use of seismic isolation devices, for design of railway bridges in seismically critical zones can be made
to reduce the cost of construction. Some energy dissipation devices are also used to dissipate the seismic
energy efficiently, in order to reduce the structural damages due to earthquake. Proper studies are necessary
about the locations and constructional details of existing bridges before these devices are selected for use on
existing bridges. This technical paper gives necessary information related with the subject.
1.0 Introduction
1.1 The traditional method of seismic design of bridges involves estimating seismic demands for design
earthquake motion and accordingly the different components of the bridge are designed to provide the
requisite strength. The ductile detailing is done in piers or abutments so that they can withstand inelastic
deformations during severe earthquakes. The structures are conventionally fixed to the substructure,
therefore earthquake motion is transmitted as such.
1.2 There is no attention paid towards decreasing the seismic demand in the traditional approach. As
inelastic deformations are permitted, the bridges designed to permit some damage and cracking to nonstructural elements such as ballast walls, locations strips of bearings etc. Minor dislocation to
superstructure and some misalignment can also be permitted. The conventional construction can cause
very high accelerations in stiff piers and large displacements in flexible piers. These two factors cause
difficulty in ensuring safety of components.
1.3 A base isolation design is based on innovative concept of decreasing the seismic demand of the bridge
by replacing conventional bearings with some kind of isolation bearing, that lengthens its time period
and thus reduces transmission of acceleration to the superstructure. Some damping element is also
introduced to restrict the amplitude of motion caused by the earthquake. There is a significant reduction

Director, Bridges & Structures Directorate, Research Designs & Standards Organization, Indian Railways,
Lucknow-226001

285

Manak Nagar,

in acceleration and displacement at the bearing level, thereby providing protection to superstructure.
1.4 The base isolated bridges are most suitable for important bridges that require higher level of protection.
The advantages of seismic isolation include the ability to eliminate or significantly reduce structural and
non-structural damage to the bridges and enhance their overall serviceability.
1.5 Some tectonic conditions such as near field earthquakes or soil-foundation condition, particularly soft
soil may, however preclude the use of seismic isolation.
2.0 Developments in Seismic Isolation
2.1 The concept of base isolation is not new. The first patent for seismic isolation was taken out in 1909 and
since then several proposals have been made to isolate the structure from ground vibrations. Significant
development has occurred in rubber technology and sliding systems that have led to the current status of
seismic base isolation systems as one of the reliable technique of passive structural control.
2.2 The following developments are responsible for evolution seismic base isolation technique to practical
reality:
(i) Development of high quality elastomeric natural rubber bearings, which can be designed to produce
low horizontal stiffness and high vertical stiffness.
(ii) The design and manufacture of mechanical energy dissipators and high damping elastomers that
are used to control displacements to acceptable levels and also to resist wind loads.
(iii) The development of computer softwares for analysis of seismically isolated bridges. These softwares
can model non-linear behaviour of isolator units.
(iv) The development of shake table testing techniques to evaluate performance and validate computer
modeling.
(v) The development of techniques of evaluating site-specific parameters of Design Basis Earthquakes.
3.0 Basic elements of seismic isolation
There are three basic elements in seismic isolation system:
(i) A flexible mounting between the superstructure and substructure to lengthen the time period.
(ii) A damper or energy dissipater so that the displacement between can be controlled.
(iii) A mechanism to provide rigidity under low lateral loads such as wind or minor earthquakes.
4.0 Types of base-isolation systems
The most practical devices developed for base isolation of bridges are following:
(i) Elastomeric bearings
(ii) Lead rubber bearings (LRB)

286

(iii) Friction pendulum system (FPS)

4.1 Elastomeric bearings
These bearings consist of thin layers of natural rubber vulcanized and bonded to steel plates. This type
of bearing is the most practical form of introducing flexibility into an isolated structure. The vertical rigidity is
maintained by steel shims, which are bonded to each layer of rubber. These steel shims constrain lateral
deformation of the rubber under vertical load. This results vertical stiffness in bearing of a similar order of
magnitude to conventional bearing. A typical elastomeric bearing is shown in Fig. 1a. The natural rubber
exhibits a mechanical behaviour which in the simplest description can be represented as a combination of
viscoelastic and hysteretic behaviour. Low damping natural rubber bearing exhibits linear elastic and linear
viscous behaviour to large shear strains. Figure 2 shows hysteretic behaviour of elastomeric bearing. The
equivalent damping ratio is typically less then 0.05 for shear strains in the range of 0 to 200%. High damping

(a)

(b)
Figure 1: Construction of elastomeric bearing

287

Figure 2: Hysteric force-displacement relation of elastomeric bearing

Figure 3: Force-displacement loops of high damping rubber bearing

288

rubber (HDR) bearings are made of specially compounded rubber which exhibits equivalent damping ration
of 0.10 to 0.20 Figure 3 shows hysteresis behaviour of high damping rubber bearing.
4.2 Lead rubber bearing
Lead rubber bearings are constructed of low damping natural rubber with a preformed central hole. A
lead core is press-fitted in the hole (Fig. 1b). The lead core deforms in almost pure shear, yields at low level
of shear stress and produces hysterisis behaviour, which is stable for a number of cycles. The lead recrystallizes

Figure 4: Construction of friction pendulum (or FPS) bearing

at normal temperature (about 20o C) so that repeated yielding does not cause fatigue. Lead bearings exhibit
characteristic strength, which ensures rigidity under service loads.
4.3 Friction Pendulum System (FPS)
Sliding bearings limit the transmission of force to the isolated structure to a limit of mR. While this is
desirable, the lack of restoring force results in significant dispersion in peak displacement response and leads
to development of permanent displacements. To avoid these undesirable features sliding bearings should be
used in combinations with a restoring force mechanism. The simplest way to introduce restoring force is to

(Figure 5a) Idealized force response spectrum

289

(Fig. 5b) Idealized displacement response spectrum

Figure 6: Response spectra for increasing damping

provide a spherical sliding surface as in FPS bearings, Fig. 4. The coefficient of friction of sliding bearings
depend on a number of parameters of which the composition of sliding surface, bearing pressure and velocity
of sliding are most important.
5.0 Force reduction due to period lengthening
5.1 The basic concept in passive base isolation systems is the period lengthening as a result of introducing
isolation bearing between super structure and the substructure. The reduction in the force response is
illustrated in Fig. 5a. It is dependent on the nature of earthquake motion and period of fixed base

290

structure. Further with the period shift, the displacement across the flexible mount is increased, Fig 5b.

Figure 7: Hysteretic force-deflection curve

However as seen in Fig. 6, the displacement can be controlled by increasing the damping. It is seen that
increasing the damping also reduces the forces at given period.
5.2 Care must be taken not to introduce excessive damping into the system because that could produce high
accelerations due to higher mode participation in the response.
6.0 Energy Dissipation
6.1 The most effective way of introducing substantial damping is through hysteretic energy dissipation. The
term hysteric refers to the offset in the loading and unloading curves under cyclic loading. Work done
during loading is not completely recovered during unloading and the difference is dissipated as heat. Fig
7 shows an idealized force-displacement loop where the enclosed area is a measure of energy dissipated
during one cycle of motion.
6.2 Several mechanical devices which use friction or the plastic deformation of either mild steel or lead to
achieve this feature has been developed in New Zealand are shown in Fig. 8. By addition of special
purpose fillers to elastomers it is possible to increase their hysterisis damping without unduly affecting
their mechanical properties. So far, it has not been possible to achieve the same level of energy dissipation
as is possible with a lead-rubber elastomeric bearing or viscous dampers.
6.3 Friction is another source of energy dissipation, which is used to limit deflections. However, with the
exception of friction pendulum system, it is a difficult source to quantify. A further disadvantage is that

291

Figure 8: Various mechanical energy dissipators

most frictional devices are not self-centering; a permanent offset between the sliding parts may result
after an earthquake.
6.4 Hydraulic damping has been used successfully in some bridges. Potentially high damping are possible
from viscous fluid flow but the maintenance requirements and high initial cost have restricted the use of
such devices. Viscoelastic, friction and shape memory alloys dampers are some more examples of
energy dissipation.
7.0 Feasibility of Seismic isolations for new construction
The bridges are generally suitable for seismic isolations under following conditions:
(i) The subsoil conditions do not produce a long period ground motion such as occurred in 1985 in
Mexico City. The base isolation will not be workable for long period ground motions.
(ii) The most significant benefits obtained from isolation are in bridges for which the fixed base fundamental

292

period of structure is less than 1.0 sec. bridges for which fundamental period is closer to 2.0 sec,
base isolation may not be that effective.
(iii) The site permits horizontal displacement at the base of the order of 200mm or more.
(iv) The other consideration when assessing the suitability of a bridge for seismic isolation is soil condition
and the geology of site. Generally, stiffer the soil, the more effective the isolation is for soft soil
condition, base isolation will not be effective. Another geologic consideration is the distance from
a major fault. For near-fault situations, generally the design forces and displacements are amplified
so much that base isolation will not be effective.
8.0 Feasibility of base isolation in retrofitting
Retrofit of existing bridges to improve their seismic performance involves additional factors compared
to new constructions because of existing constraints. Some bridges are more suitable for retrofit by base
isolation than by other methods. For example bridge structures can be retrofitted by replacement of traditional
steel bearing with elastomeric bearings. This replacement will cause not only reduction of earthquake-induced
forces but allow the redistribution of seismic forces to other regions of structures.
9.0 CONCLUSION
9.1 Design of bridges, based on seismic isolation is an effective and economic solution to reduce the seismic
acceleration and displacements in high seismic zones where soil conditions are stiff.
9.2 Use of energy dissipation devices can also be made to control the structural damages due to seismic
vibration.
9.3 Retrofitting requirements of existing bridges needs to be studied in detail before using seismic isolation
devices to improve their seismic
10. ACKNOWLEDGEMENT
The author sincerely acknowledge the contribution of Dr. S.K. Thakkar (Ex. Professor Bridge Chair,
IIT, Roorkee), in providing valuable input for this paper which is intended to improve general awareness of
bridge engineers of Indian Railways.

x x x

293

EMERGING PROVISIONS FOR SEISMIC DESIGN

OF RAILWAY BRIDGES
Ravindra Kumar Goel
Director/B&S, Research Design & Standards Organisation, Lucknow
(Paper to be presented in International Seminar of IIBE at Hyderabad, Apirl, 2010)
ABSTRACT
Assessment of seismic loads for design of railway bridges is done as per seismic coefficient method given in Indian
Railway Standard Bridge Rules. Bridge Rules has derived its seismic loading provisions from IS 1893 (1984). In these
provisions, design seismic coefficient does not depend on the flexibility of the bridge. Moreover, the ductility of bridge
components is not considered while calculating the design seismic loads. Similarly, there are no details about response
spectrum and time history analysis. IS 1893 itself has under gone a complete change after the Bhuj earthquake in 2001.
Research Design & Standards Organisation has studied the latest developments in the field of earthquake engineering in
association with IIT, Kanpur and developed the guidelines for seismic design of Railway Bridges. The international
practice of assessment of seismic forces on realistic basis has been suggested and seismic load combinations have been
rationalised accordingly. For the first time, the seismic isolation devices have been suggested to counter seismic ground
motions. This paper gives the salient features of the emerging provisions for seismic design of railway bridges.

1.0 INTRODUCTION
1.1 Provisions for assessment of seismic loads for design of Railway bridges are given in Indian Railways
Standard Bridge Rules. Bridge Rules of the Indian Railways has derived its seismic loading provisions
from IS 1893 (1984). In these provisions, seismic coefficient method is used for bridges, wherein
design seismic coefficient does not depend on the flexibility of the bridge. Moreover, the ductility of
bridge components is not considered while calculating the design seismic loads. Similarly, there are no
details about response spectrum and time history analysis.
1.2 IS:1893 was initially published in 1962 as "Recommendations for Earthquake Resistant Design of
Structures" and then revised in 1966. As a result of additional seismic data collected in India and further
knowledge and experience gained the standard was revised in 1970, 1975 and then in 1984.
1.3 This standard has been further reviewed on account of earthquakes in various parts of the country
including that in Uttar-Kashi, Latur and Bhuj and technological advancement in the field and the Sectional
Committee CED:39 of BIS, decided to split the standard into five parts to deal with different types of
structures.
2.0 REVIEW OF LATEST PROVISIONS OF IS:1893
2.1 IS 1893(Part 1):2002 specifies the Criteria for Earthquake Resistant Design of Structures and deals
with general provisions and Buildings. This standard was finalized in 2002 and it covers general principles
and design criteria, combinations, design spectrum, main attributes of buildings, dynamic analysis, apart
from seismic zoning map and seismic coefficients of important towns, map showing epicenters, map
294

showing tectonic features and lithological map of India.

2.2 IS 1893 (part-3) is on "Bridges and Retaining Walls," and deals with bridges. This part is still in draft
stage. Final version of the draft is in circulation for comments and likely to be finalized shortly. The
proposed draft has been formulated to ensure that bridges possess at least a minimum strength to
withstand earthquakes. The intention is not to prevent damage to them due to the most severe shaking
that they may be subjected to during their lifetime. Actual forces that appear on different portions of
bridge during earthquakes may be greater than the design seismic forces specified in this standard.
However, ductility arising from material behavior, detailing and over strength arising from the additional
reserve strength, over and above the design force, are relied upon to take care of this difference in actual
and design lateral loads.
2.3 The reinforced and pre-stressed concrete components are proposed to be designed under reinforced
so as to cause a tensile failure. Further, they should be suitably designed to ensure that premature failure
due to shear or bond does not occur. Ductility demand under seismic shaking is usually not a major
concern in bridge superstructures. However, the seismic response of bridges is critically dependent on
the ductile characteristics of the substructures, foundations and connections.
3.0 IMPORTANT MODIFICATIONS IN DRAFT IS:1893 (PART-3)
3.1 Following major improvements have been suggested in draft IS:1893 (Part - 3) as compared to IS:
1893: 1984.
a)

Relative values of seismic zone factors are the same as included in IS 1893 (Part 1): 2002. Three
methods, namely seismic coefficient method, response spectrum method and time history method
are given for estimating design seismic forces which recognizes the flexibility of bridges.

b) The concept of ductility and over-strength is brought in the draft explicitly, by introducing the
response reduction factors.
c)

Different response reduction factors have been proposed for the different components of the bridge,
depending on the redundancy, expected ductility and over-strength in them.

d) The design force level for bridge has been raised from the existing level and brought in line with
IS1893(Part1):2002.
e)

The concept of capacity design is introduced in the design of connections, substructures and
foundations.

f)

The soil-foundation factor is dropped. The effect of soil on response is represented in the response
spectrum.

g)

Provision for dislodging of girders in the bearings is introduced.

h)

Use of vertical hold-down devices, stoppers, restrainers and horizontal linkage elements to account
for the large displacements generated during seismic shaking is recommended for preventing falling

295

of spans.
j)

A minimum width of seating of superstructure over substructures to avoid dislodging of spans from
atop the substructure is required for all bridges.

k) The method of computing earth pressures for C- ? soil is included in the section on Retaining Walls.
4.0 BASIC DIFFERENCES IN SEISMIC DESIGN OF RAILWAY AND HIGHWAY BRIDGES
4.1 The design considerations involved in case of seismic design of railway bridges are somewhat different
as compared to highway bridges. The important differences are as given under:
(i) Simple span structures are preferred over continuous structures for railway bridges. Many of the
factors that make continuous spans attractive for highway bridges are not as advantageous for
railway use. Continuous spans are also more difficult to replace in emergencies than simple spans.
(ii) The ratio of live to dead load is much higher for a railway bridge than for a similarly sized highway
bridge. This can lead to serviceability issues such as fatigue and central deflection governing the
designs rather than strength.
(iii) Design impact load on railway bridges is higher as compared to highway bridges.
(iv) Interruptions in service are typically much more critical for railway than for highway agencies.
Therefore constructability and maintainability without disruption to traffic are crucial for railway
bridges.
(v) Since the bridge supports the track structure, the combination of track and bridge movement
cannot exceed the tolerances in track standards. Interaction between the track and bridge should
be considered in design and detailing.
(vi) Seismic performance of highway and railway bridges can vary significantly. Railroad bridges have
performed well during seismic events.
(vii) Track structure (along with guard rail) serves as an effective restraint (and damping agent) against
bridge displacements in case of railway bridges.
(viii) Railway bridge owners typically expect a longer service life from their structures than highway
bridge owner expect from theirs.
(ix) Trains operate in a controlled environment, which makes type of damage permissible for railway
bridges that might not be acceptable generally for highway users.
4.2 Keeping in view above a need was felt to develop new provisions for seismic design of railway bridges
in conformity with latest developments in the country and elsewhere.
5.0 FUNDAMENTAL PRINCIPLES OF SEISMIC DESIGN
5.1 Major development have taken place internationally in the field of earthquake engineering and the
earthquakes are now known to take place along defined tectonic plate boundaries and near to identifiable

296

geological features. The seismically critical areas can be more accurately identified and the magnitudes
of earthquakes in such areas can be reasonably predicated.
5.2 Following principles of design are followed internationally(i) Small to moderate earthquakes should be resisted within elastic range of the structural components
without significant damage.
(ii) Realistic ground motion intensities and forces are used in the design procedures.
(iii) Exposure to shaking from large earthquake should not cause collapse of all or part of the bridge.
Where possible damage that does occur should be readily detectable and accessible for inspection
and repair.
5.3 The peak ground acceleration, generated during an earthquake causes inertial forces in various bridge
components and the design forces for the component is somewhat different depending upon it's ductility.
The design seismic force is, therefore modified by a factor called Response Reduction Factor. This
factor depends on energy dissipation or ductile characteristics of the bridge component.
5.4 The seismic zone map as given in IRS : 1893 (Part -1) : 2002 classifies the country in four seismic zones
based on the maximum considered earthquake. The seismic zone factors (Z) given, are nothing but the
peak ground accelerations (PGAs) expected in these zones. The structures are proposed to be designed
for a of peak ground acceleration of Z/2 only. Thus, the design basis earthquake is one which gives a
peak ground acceleration (PGA) of Z/2. An earthquake of higher PGA, occurring in the service life of
the bridge will, therefore, be resisted by mobilization of the over strength or reserve strength of the
structure.
5.5 Keeping the above principles in view, the guidelines have been developed for seismic design of railway
bridges. The salient features of the guidelines are briefly discussed hereafter.
6.0 CONCEPTUAL CONSIDERATION
6.1 Conceptual considerations are aimed at providing simplicity, symmetry, and displacement capacity in
the bridge so as to improve its seismic resistance. This is similar to the role of architectural planning and
detailing in the seismic performance of buildings. In the past earthquakes, it is seen that bridges with
preferred configurations, superstructure, substructure and ground conditions have performed better
than non preferred type. Bridges of non preferred types require special considerations in modeling,
analysis, design, and construction. The selection of an appropriate structure type and configuration
should take the seismic hazard at the site, the soil conditions and the bridge performance requirements
into account. In general, sites near active faults, with potentially liquefiable or unstable soil conditions
and with unstable sloping ground conditions should be avoided. The measures to improve the soil
conditions should also be considered as an alternative.
6.2 Configuration
Criteria for determining an adequate structural configuration and layout include simplicity, symmetry and
297

regularity, integrity, redundancy, ductility and ease of inspection and repair. Bridge should be simple in
geometry and structural behavior. Simple structural configuration provides a direct and clear load path
in transmitting the inertial forces from superstructure to ground. The bridge behavior under seismic loads
can be predicted with more certainty and accuracy with fewer dominant modes of vibration. Bridges
with features such as extreme curvature or skew, varying stiffness and mass and abrupt changes in
geometry require special attention in analysis and detailing to avoid permanent damages and failure.
6.3 Superstructure
6.3.1

Simple spans of standard configuration are preferred by railways since they have performed
well during past earthquakes and can be returned to survive or replaced. In simple spans,
lateral load on piers depend on the weight of adjacent spans. If spans are of equal length,
then, all the piers are subjected to almost same lateral seismic force.

6.3.2

In continuous bridges, since all the piers are connected through deck, the lateral seismic force
on a pier also depends on the stiffness of pier. In such cases, large lateral forces may get
transmitted to one single pier of large stiffness. Continuous spans may reduce the likelihood of
unseating at the pier and thus prevent release of seismic energy. Moreover, all the longitudinal
forces are resisted at the fixed end at any one of the piers or abutments, therefore that pier or
abutment is likely to suffer more damages during seismic ground motion.

6.3.3

Long spans produce higher load demands on fewer foundations which will increase foundation
vulnerability and reduce redundancy.

6.2.4

Excessive ballast and other non structural weight should be avoided as much as possible
while designing the superstructure.

6.4 Substructure
Wide seat width at the abutment and the pier allows for large displacements without unseating the bridge
spans. Multiple columns provide redundancy in the substructure which is needed to survive the higher level
ground motions.
6.5 Ground Conditions
6.5.1

The foundation soil should be investigated for susceptibility to liquefaction and slope failure
during the seismic ground motion. To the extent possible, bridges in the region of high seismicity
should be founded on stiff, stable soil layers.

6.5.2

Large diameter pile foundations may be used to withstand the slope failure or to carry the
bridge loads through liquefiable soil layer to competent material.

6.6 Conceptual design suggestions in terms of configuration, superstructure, substructure and ground
conditions are given in Table-1, along with the non preferred types, for which special designs and details
are required. These considerations shall be followed as much as practical and a balance shall be maintained

298

Table 1 - Seismically Preferred and Not Preferred Aspects of Bridges

1.0
1.1
1.2
1.3
1.4
1.5
2.0
2.1
2.2
2.3
2.4
3.0
3.1
3.2
4.0
4.1

Seismically preferred
Configuration
Straight bridge alignment
Normal piers
Uniform pier stiffness
Uniform span stiffness
Uniform span mass
Superstructure
Simply supported spans
Short spans
Light spans
No intermediate hinges within span
Substructure
Wide seats
Multiple column
Ground conditions
Stiff, Stable soil

Seismically not preferred

Curved bridge alignment
Skewed piers
Varying pier stiffness
Varying span stiffness
Varying span mass
Continuous spans
Long spans
Heavy spans
Intermediate hinges
Narrow seats
Single column
Unstable soil

between functional requirements, cost and seismic resistance features.

7.0 REGUALR AND IRREGULAR BRIDGES
7.1 Regular Bridge - A regular bridge has no abrupt or unusual changes in mass, stiffness or geometry
along its span and has no large differences in these parameters between adjacent supports (abutments
excluded). A bridge shall be considered regular for the purposes of this guidelines, if
(a) It is straight or it describes a sector of an arc which subtends an angle less than 90 at the center of
the arc.
(b) The adjacent piers do not differ in stiffness by more than 25%. (Percentage difference shall be
calculated based on the lesser of the two stiffnesses as reference).
(c) If multi-column piers are used then the stiffness of the most stiff column within a pier shall not be
25% more than the stiffness of the most flexible column in that pier.
7.2 Irregular Bridge - All bridges not conforming to above, shall be considered irregular. Further, arch
bridges of span exceeding 30m, cable stayed bridges, suspension bridges, and other innovative bridges
shall also be treated as irregular bridges. Response Spectrum (Multi Mode) Method of analysis is
mandatory for irregular bridges.
8.0 SEISMIC ZONE MAP AND ZONE FACTORS
8.1 For the purpose of determining design seismic forces, the seismic zone map of India as given in IS:1893
(Part-1) shall be used.

299

Table 2 - Zone Factor Z for Horizontal Motion

Seismic
Zone

II

III

IV

0.10

0.16

0.24

0.36

8.2 The PGA (Peak ground acceleration also called zero period acceleration, ZPA), associated with each
zone, is called zone factor, Z. The zone factors are given in Table-2.
9.0 IMPORTANCE FACTOR

Table 3 - Importance Factor for bridges

Bridges included
i) All important bridges irrespective of route.
ii) Major bridges on group A, B and C routes.
(Route classification as per IR P.Way Manual)
i) Major bridges on all other routes.
ii) All other bridges on group A, B and C routes.

Importance Factor
1.5

1.25

All other bridges

1.0

9.1 The values of importance factor I, proposed for different bridges are given in Table-3. The importance
factor reflects strategic importance of the route and functionality of the bridge in the post earthquake
period.
10.0 METHOD OF SEISMIC ANALYSIS
10.1 The seismic forces for bridges shall be generally estimated by Seismic Coefficient Method (Single Mode
Method) described as under Elastic Seismic Acceleration Coefficient, 'Ah'
Ah =

ZIS a
2g

Where,
Z = Zone Factor, given in Table 2 for horizontal motion.
I = Importance Factor, given in Table 3,
Sa
g

Spectrum Acceleration Coefficient along the considered direction given as follows:

300

For rocky, or hard soil sites (Type I)

Sa
g

2.50
1.00/T1

T1 0.40
0.40 T1 3.00

0.33

T1

3.00

For medium soil sites (Type II)

Sa
g

2.50
1.36/T1

T1 0.55
0.55 T1 3.00

0.45

T1

3.00

T1
T1

0.67
3.00

T1

3.00

For soft soil sites (Type III)

2.50
1.67/T1
0.56

0.67

Sa/g

Sa
g

T1
Fig. 1 Response Spectrum for 5% damping

301

Table 4 - Multiplying Factors for Other Damping percentages

Damping %
0
2
5
7
10
15
20
25
30

Factors
3.02
1.40
1.0
0.90
0.80
0.70
0.60
0.55
0.50

T1 = Fundamental natural period of the bridge (along the considered direction).

10.2 A plot of

Sa
versus T1 is given in Fig.1 for 5% damping. For other damping values, the multiplying
g

factors are given in Table - 4.

10.3 Design Seismic Force Resultants for Bridge Components - The design seismic force resultant V
at a cross section of a bridge component due to earthquake shaking along a considered direction shall
be given
Table 5: VALUES OF RESPONSE REDUCTION FACTOR (R)
Bridge component and Type

Superstructure
-RCC
1.0
-PSC
1.0
-Steel
1.0
Substructure
RCC Piers with ductile detailing
- Single Column, Wall Type
2.5
- Frame Type
3.25
RCC Piers without ductile detailing
- Single Column, Wall Type
2.0
- Frame Type
2.5
Steel Framed Construction
2.5
RCC Abutment
2.0
PCC Abutment
1.5
Connections (including bearings)
Superstructure to abutment
0 .8
Superstructure to column
1
Columns or piers to foundations
1
Expansion joints within a span of the superstructure
0.8
Note: Response reduction factor is not to be applied for the calculation of displacements.

302

by V

Ve
R

Where,
Ve = Maximum elastic force resultant at the chosen cross-section of that bridge component
R = Response Reduction Factor for the component as given in Table 5.
10.4 Other Methods of Analysis - Under specific circumstances, the guideline suggests the use of other
methods of estimation of seismic forces such as a)

Response Spectrum (Multi Mode) Method

Table - 6 : Calculations of seismic coefficients
(Implications of proposed provisions v/s existing provisions)

S.N.

Existing IRS Bridge Rules

h = 0 I

For Zone V
0= 0.08

I = 1.5

4
= 1.0 for hard soil

Proposed Provision

1 Sa
. I.
2 R g

Ah=

0.36
=
= 0.18
2
2
I = 1.5 Important Bridge + major Bridges on A,B & C
routes.
= 1.25 Major Bridges + Other Bridges on ABC routes.
= 1.0 All other bridges.
(a) For rigid structures i.e. T<0.4 sec

Sa
= 2.5
g

(for hard soil)

R = 2.5 for RCC pier with ductile detailing.

1 Sa 1
=
(2.5) = 1.0
R g 2.5
h = 0.08 x 1.5
= 0.12

(i)

Ah = 0.18 x 1.5 x 1
= 0.27 i.e. increased by 125%

(ii) Generally small bridges are rigid and having short

piers. So important factor may be taken as 1.0.
Ah = 0.18x1.0x1.0
= 0.18 i.e. increased by 50% only

303

(a)
= 1.0 for hard soil

For structures with some flexibility i.e. T = 1.0

Sec.

Sa
= 1.0
g
R = 2.5 for RCC pier with ductile detailing
Therefore,

1 Sa 1
=
R g 2.5
h = 0.08 x 1.5 x 1.0
= 0.12
6.
= 1.0 for hard soil
h = 0.08 x 1.5 x 1.0
= 0.12

Ah = 0.18 1.5

1
2. 5

= 0.108
i.e. reduced by 10%
(c) For flexible structures i.e.
T = 3.0 Sec.

Sa
= 0.33
g
R = 2.5
Therefore,

1 Sa 0.33
=
= 0.133
R g 2.5

Ah = 0.18 x 1.5 x 0.133 = 0.036

~ 0.04 (say)
i.e. reduced by 66.7%

b) Time History Method

These methods of analysis require selection of appropriate time histories from actual earthquakes and
use of modeling of bridge. Therefore, consultation from experts in the field is advised.
11.0 IMPLICATION OF THE PROPOSED PROVISIONS
11.1 The design values of horizontal acceleration as per proposed guidelines have been worked out and
compared with those obtained as per IS Bridge Rules. The comparison is shown in Table 6.
11.2 It is seen that the horizontal design seismic acceleration is less for tall piers having more flexibility.
12.0 MODIFICATION IN SEISMIC LOAD COMBINATION
12.1One of the most significant aspects of emerging provisions is reduction of partial load factors for limit
state design of Reinforced Cement Concrete (RCC)/Prestressed Cement Concrete (PSC) structures.
Table 12 of IRS Concrete Bridge Code is proposed to be modified and the load combination for
seismic case shall be as shown in Table 7.
12.2 It can be seen that the overall seismic design loads are now more rationally accounted for an economical
design.

304

Table 7- Comparison of Proposed Partial Load Safety Factors with those existing in IRS Concrete Bridge Code
S.N.

ULS/
SLS

ULS

SLS

305
3

During
construction

Existing partial safety factors as Partial Safety Factor for Load (fL)
Remarks
per Concrete Bridge Code
as proposed
1. The partial load factor is 1.6 in existing load
a)
1.4 DL + 2.0 SIDL +
a)
1.25 DL + 1.5 SIDL + 1.5 factor is reduced to 1.5 in new guidelines. This is
1.6 EQ + 1.70 EP
EQ +1.5 EP
IS:1893 part-3.
b)
1.4 DL + 2.0 SIDL +
2.
It is proposed to consider =0.3 only fo
1.75 (LL + IL) +1.25 LL(F) b)
1.25 DL + 1.5 SIDL + with EQ load. This is as per proposed draft IS:189
+1.25 EQ
(LL + IL*+ LL(F) ) +1.20 EQ +1.5 of occurring seismic load with live load is very l
(Erection loads are also to be EP + 1.4 PS + 1.4 HY + 1.4 BO
partial factor is reduced to 0.30. This factor may
considered along with)
bridge is located in yard or near a stopping s
combination we are considering partial safety fact
a)
1.0 DL + 1.20 SIDL+ a)
1.0 DL + 1.20 SIDL+ 1.0 load i.e. the earthquake effects are not augmente
1.0 EQ + 1.0 EP
EQ + 1.0 EP
primary load combination without live load.
increase this factor in specific cases of seismicall
b)
1.0 DL + 1.20 SIDL + b)
1.0 DL + 1.20 SIDL + rational manner.
1.0 (LL + IL) + 1.0 LL (F) + (LL + IL* + LL (F) ) + 1.0 EQ
3.
Impact load is to be considered with liv
1.0 EQ
superstructure only.
(Erection loads are also to be
4.
fL=1.4 for DL and 2.0 for SIDL are red
considered along with)
respectively for a realistic assessment. Earth wo
bridges shall not be considered as SIDL.
5.
fL=1.5 is proposed for EP for realistic as
No. separate combination

1.0 DL + 1.2 SIDL + 1.0 EQ + 1.1. Live Loads are not considered with EQ loads durin
ER
1.3 EP + 1.0 PS + 1.0 HY + 1.0 BO

DL = dead load,
LL = live load,
IL* = Impact load (To be considered for design of superstructure only)
EP = earth pressure,
HY = hydrodynamic load,
SH = shrinkage load,

DL(S)
LL (F)
EQ
PS
BO
ER

= superimposed dead load (SIDL)

= live load on footpath,
= earthquake load,
= prestressing load,
= buoyancy load,
= Erection load,

Note : The partial load factors given are for general guidance only and can be changed depending upon the location of the bridge and intensity of traffic.

13.0 ADDITIONAL STUDIES FOR SPECIAL CASES

13.1 For specific cases of bridges, some additional studies/analysis are proposed, which are described in
Table 8.
Table 8 - Cases Requiring Special Studies/Analysis
Cases in which additional special studies/
Special studies/analysis
analysis is required
In zone IV and V, bridges with individual span Modeling of the bridge including geometrical
length more than 120 m and/or pier height is nonlinearity, P-delta effect and soil-structure interaction
more than 30 m
is needed.
Pushover analysis may be done to ascertain the energy
dissipation characteristics of ductile members.
Continuous deck bridge of length larger than Spatial variation of ground motion shall be considered.
600 m, or Geological discontinuity exists at the
site
Bridge site close to a fault (< 10 Km) which 1. Site specific spectrum shall be obtained. Else, nearmay be active.
source modifications shall be done. Specialist
literature shall be required to obtain site specific
spectrum.
2. If bridge is crossing the fault, detailed geological
studies shall be performed to estimate past
movements across the fault. Bridge to be designed
so as to withstand the expected fault displacements.
Help from geological/ seismological persons with
enough experience will be required to calculate
fault movement.
In zone IV and V, the soil condition is poor, Site specific spectrum shall be obtained.
consisting of marine clay or loose sand (e.g.,
where the soil up to 30m depth has average SPT
N value equal to or less than 20)
Site with loose sand or poorly graded sands Liquefaction analysis is required.
with little or no fines. Liquefiable soil.

14.0 POST EARTHQUAKE OPERATIONS AND INSPECTIONS

The response of railway tracks and bridges to an earthquake would depend on distance from epicenter
and nature of attenuation. The post earthquake train operations in the region shall be cautiously started. The
guidelines provide specific instructions to be followed for safe operation of trains once an earthquake has
been reported to traffic controller.
15.0 CONCLUSIONS
15.1 Seismic design of bridges has under gone major philosophical changes all over the world.
15.2 The selection of type of bridge, arrangement of spans and configuration are most important factors to be
kept into consideration for seismic resistant design of bridges.

306

15.3 Railway bridges are having different considerations for seismic design as compared to highway bridges.
Post earthquake safety precautions are equally important to prevent further derailments and to ensure
that the relief operations are carried out safely.
15.4 The emerging provisions for seismic design of railway bridges are in line with the recent philosophical
changes taking place in national and international circles. However, there is a need of increased awareness
about the correct use of provisions for design of new bridges.
15.5 There is a need of undertaking seismic audit of existing railway bridges and to develop a methodology
for retrofitting of bridges found seismically deficient. It is a huge work and additional resources in terms
of trained technical manpower are required to be generated.

x x x

307

For Official Use Only

Technical Paper No.355

GOVERNMENT OF INDIA
MINISTRY OF RAILWAYS

COMPENDIUM
OF
TECHNICAL PAPERS
ON
DESIGN, FABRICATION, INSPECTION,
REHABILITATION, LOAD RATING &
FATIGUE ASSESSMENT
OF
STEEL GIRDER BRIDGES

Compiled
By
R.K. Goel
Director/B&S

January-2010

Research Designs & Standards Organisation

Lucknow-226011

Dedicated
To
Them
Who Believe
In
Selfless Service

FOREWORD
Indian Railways is having a large population of steel girder bridges to maintain. Design & fabrication of
steel bridges require good understanding of the codal provisions and the design principles involved. The
concepts are often complicated and need to be explained properly to the beginners. It is important to
maintain proper documentation of such concepts. B&S Directorate of RDSO has been in the field of designing
steel bridges and inspecting the work of fabrication to ensure that the fabrication is done as per standard
drawings with proper implementation of the designers concepts. The maintenance problems arising in the
field have also been addressed time to time. In the process, the design concepts have been well understood
by the RDSO officials and clarifications are given to the design staff of zonal railways whenever needed.
Several technical papers have been written time to time by various authors to document the important
concepts such as cambering in open web girders, assessment of residual fatigue life, design of welded
connections, repairs/rehabilitation of bridges and bearings, load rating of bridges etc. Technical papers have
also been written on other aspects of bridges such as estimation of design discharge, track standards on
girder bridges, composite bridge sleepers, bridge inspection and maintenance of bridge bearings.
A good effort has been made to compile all such papers and issue the compilation in the form of a
compendium by Shri R. K. Goel, Director/SB-I. Significant contribution has been made by Shri R.K.
Sharma, JE/Library and Smt. Suman Verma, Steno Grade-1, B&S Directorate in searching and arranging the
papers in presentable form. I believe that the compendium will serve as a good reference to the designers
and field bridge engineers of zonal railways.
Comments, if any shall be highly appreciated.

(Mahesh K. Gupta)
Executive Director

INDEX
SL.

TITLE

PAGE NO.

NO.
1.

DESIGN

1.1

Cambering and Pre-stressing of Standard Steel Girders on Railway Bridges

1.2

Dynamic Effects on Railway Bridges A Review

1.3

Fatigue Design of Welded Connections For Railway Bridges,

14

1.4

Design Considerations in Design of Pot-PTFE Bearings for Railway Steel Bridges,

25

1.5

Study of Behaviour of Stringer to Floor Beam Connection in Rivetted Open Web

Girder Bridges

39

1.6

Synthesizing S-Curve and IUH with Catchment Characteristics

45

1.7

Composite Sleepers : An Environment Friendly Alternate Sleeper for Track and Bridges

52

2.

FABRICATION

2.1

Internal Defects in Structural Steel for Bridges

67

2.2

Fatigue Tolerance of Welded Connections in Railway Bridges

70

2.3

Construction of Steel Girders- Practical Considerations in Design of

Web-Flange Connection.

80

3.

INSPECTION & REHABILITATION

3.1

Track standards on girder bridges A Review

91

3.2

Rehabilitation of Distressed Road Over Bridge at Firozpur Cantt A Case Study

96

3.3

Inspection & Maintenance of bearings for Steel Girder Bridges

3.4

Repair and Reclamation of Cast Iron Pendulum Bearing of Jubilee Bridge by

Eutectic Welding Technique

102
105

4.

FATIGUE ASSESSMENT

4.1

Review of Fatigue Provisions of BS:5400 for Design of Railway Bridges

117

4.2

Fatigue Life Assessment of Railway Steel Bridges

130

4.3

Advancements in Codal Provisions for Fatigue Design of Railway Steel Bridges

141

4.4

Review of Fatigue Provisions of Euro Codes for Design of Railway Steel Bridges

152

4.5

Emerging Load Model (MBG) for Fatigue Assessment of Railway Steel Bridges

162

4.6

Analytical approach for assessment of fatigue life of steel girders of bridge no. 110 (UP line)
across river Ganga on Lucknow-Kanpur section of Indian railways

4.7

169

Fatigue Assessment Criteria for Design and Analysis of Steel Girder Bridges for Heavy
Axle Load Operations

181

4.8

Fatigue Assessment of Concrete Bridges Euro Norms

189

4.9

Determination of Fatigue Load Parameters using Artificial Neural Networks

197

4.10

Impact of Increasing Axle Load on Fatigue Life of Standard Steel Girder Bridges
A Study Based on Revised Fatigue Provisions

5.

LOAD RATING

5.1

Increase in Design Longitudinal Forces over a Century Due to Heavy Haul and
Implications on Existing Bridges

5.2

225

Load Rating Procedure for Capacity Assessment of Bridges for Running Heavy Axle
Load Traffic

5.3

205

236

Experience of Pilot Project on Load Rating of Bridge No. 502 on Western Railway by
Live Load Testing

245

5.4

Checking the Suitability of Existing Roller-Rocker Bearings for Higher Axle Load

257

5.5

Effect of Increasing Speed on Dynamic Impact An Analytical Study on Standard

Steel Girder Bridges

5.6

264

Technical Papers on Bridge Performance, IXth International Heavy Haul Conference

Held at Shanghai, China A Review

278

5.7

Scismic Isolation of Bridges

285

5.8

Emerging Provisions for Seismic Design of Railway Bridges

294

DESIGN