Structural Engineering Documents

Gunter Ramberger

Structural Bearings
and Expansion Joints
for Bridges
I

International Association for Bridge and Structural Engineering

Association lnternationale des Ponts et Charpentes
lnternationale Vereinigung fur Bruckenbau und Hochbau

IABSE
AIPC
IVBH

Copyright 02002 by
International Association for Bridge and Structural Engineering
All rights reserved. No part of this book may be reproduced in any form or by any
means, electronic or mechanical, including photocopying, recording, or by any
information storage and retrieval system, without permission in writing from the
publisher.
ISBN 3-85748-105-6
Printed in Switzerland
Publisher:
IABSE-AIPC-IVBH
ETH Honggerberg
CH-8093 Zurich, Switzerland
Phone: Int. + 41-1-633 2647
Fax:
Int. + 41-1-633 1241
E-mail: secretariat@iabse.ethz.ch
Web: http://www.iabse.ethz.ch

Dedicated to the commemoration of the late Prof. Dr. techn. Ferdinand Tschemmernegg,
University of Innsbruck.

Preface
It is my hope that this treatise will serve as a textbook for students and as information
for civil engineers involved in bridge construction. My intent was to give a short
guideline on bearings and expansion joints for bridge designers and not to mention all
the requirements for the manufacturers of such products. These requirements are
usually covered by product guidelines, which vary between different countries.
Not all the references are related to the content of this document. They are more or less
a collection of relevant papers sometimes dealing with special problems.
I express many thanks to Prof. Dr.-Ing. Ulrike Kuhlmann, University of Stuttgart,
chairperson of Working Commission 2 of IABSE, who gave the impetus for this work;
to her predecessor of the IABSE Commission, Prof. Dr. David A. Nethercot, Imperial
College of Science, Technology and Medicine, London, for reviewing the manuscript,
and Prof. Dr. Manfred Hirt, Swiss Federal Institute of Technology, Lausanne, for his
contributions and comments.

I wish to thank J. S. Leendertz, Rijkswaterstaat, Zoetermeer; Eugen Briihwiler, Swiss

Federal Institute of Technology, Lausanne; Prof. R. J. Dexter, University of Minnesota; G. Wolff, Reissner & Wolff, Wels; 0. Schimetta t, Amt der 00 Landesregierung,
Linz; Prof. B. Johannsson, LuleA Tekniska Universitet, for amendments, corrections,
remarks and comments. I thank also my assistant Dip1.-Ing. Jorgen Robra for his
valuable contributions to the paper, especially for the sketches and drawings, and my
secretaries Ulla Samm and Barbara Bastian for their expert typing of the manuscript.
Finally, I would like to thank the IABSE for the publication of this Structural Engineering Document.

Vienna, April 2002

Gunter Ramberger

Table of Contents
1.

Bearings
1.1 Introduction
1.2 The role of bearings
1.3 General types of bearings and their movements
1.4 The layout of bearings
1.5 Calculation of bearing reactions and bearing movements
1.6 Construction of bearings
1.7 Materials for bearings
1.8 Analysis and design of bearings
1 .9 Installation of bearings
1.10 Inspection and maintenance
1. I 1 Replacement of bearings
1. I 2 Codes and standards
1.13 References

2.

7
7
7
9
16
19
29
33
37
38
39
41
42

Expansion Joints
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
2.10

Introduction
The role of expansion joints
Calculation of movements of expansion joints
Construction of expansion joints
Materials for expansion joints
Analysis and design of expansion joints
Installation of expansion joints
Inspection and maintenance
Replacement of expansion joints
References

51
51
51
58
70
72
84
86
87
88

Bearings

1.1

Introduction

All bridges are subjected to movements due to temperature expansion and elastic
strains induced by various forces, especially due to traffic loads. In former times our
bridges were built of stones, bricks or timber. Obviously, elongation and shortening
occurred in those bridges, but the temperature gradients were small due to the high
mass of the stone bridges. Timber bridges were small or had natural joints, so that the
full elongation values were subdivided into the elongation of each part. On the other
hand, the elongation and shortening of timber bridges due to change of moisture is often higher than that due to thermal actions. With the use of constructional steel and,
later on, of reinforced and prestressed concrete, bridge bearings had to be used. The
first bearings were rocker and roller bearings made of steel. Numerous rocker and
roller bearings have operated effectively for more than a century. With the development of ageing-, ozone- and UV-radiation-resistant elastomers and plastics, new materials for bearings became available. Various types of bearings were developed with
the advantage of an area load transmission in contrast to steel bearings with linear or
point load transmission, where elastic analysis leads theoretically to infinite compression stresses. For the bearings the problems of motion in every direction and of load
transmission were solved, but the problem of insufficient durability still exists. Whilst
it is reasonable to assume the life of steel bearings to be the same as that of the bridge,
the life of a bearing with elastomer or plastic parts can be shorter.

1.2

The role of bearings

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.

1.3

General types of bearings and their movements

Normally, reaction forces and the corresponding movements follow a dual principle a non zero bearing force corresponds to a zero movement and vice versa. An exception
is given only by friction forces which are nearly constant during the movement, and by
elastic restraint forces which are generally proportional to the displacement.
Usually, the bearing forces are divided into vertical and horizontal components.
Bearings for vertical forces normally allow rotations in one direction, some types in
all directions. If they also transmit horizontal forces, usually vertical forces are combined.

1. Bearings

A special type of bearing transmits only horizontal forces, while allowing vertical
displacements.
The following table (Table 1.3-1) shows the common types of bearings, including the
possible bearing forces and displacements. Friction and elastic restraint forces are not
considered.

Symbol Function

Construction

All translation
fixed
Rotation all
round

Point rocker bearing

Pot bearing; Fixed
elastomeric bearing;
Spherical bearing

Horizontal
movement in
one direction
Rotation all
around

Constr. point rocker

sliding bearing;
Constr. pot sliding
bearing; Const.
elastomeric bearing;
Constr. spherical
sliding bearing

Horizontal
movement in
all directions
Rotation all
round

Free point rocker

bearing; Free pot
sliding bearing; Free
elastomeric bearing;
Free spherical sliding
bearing; Link bearing
with universal joints
(tension and
compression)

All translation Line rocker bearing

Leaf bearing
fixed
(tension and
Rotation
about one axis compression)
Horizontal
movement in
one direction
Rotation
about one axis

Roller bearing; Link

bearing (tension and
compression);
Constant line rocker
sliding bearing

Horizontal
movement in
all direction
Rotation
about one axis

Free rocker sliding

bearing; Free roller
bearing; Free link
bearing
~

All horizontal
tranal. fixed
Rotation all
round

HoriLontal force
bearing

Horizontal
movement in
one direction
Rotation all
round

Guide bearing

1.4 The layout of bearings

Tuble 1.3-1

8.2

1.4 The layout of bearings

1.4.1 General
Bearings can be arranged at abutments and piers (fig. 1.4.1-1 ; fig. 1.4.1-2) under the
webs of the main girders, under diaphragms (fig. 1.4.1-3), and under the nodes of
truss bracings. The webs and the diaphragms of concrete bridges have to be properly
reinforced against tensile splitting; steel bridges need stiffeners in the direction of the
bearing reactions to transfer the concentrated bearing loads to the superstructure and
the substructure. Abutments and piers also have to be properly reinforced under the
bearings against tensile splitting.

-77
Fig. I .4. I - I : Bearings at an abutment

I -

~-

Fig. 1.4.1-2: Bearings at u pier

I7

Fig. 1.4.1-3: Bearing at a single pier

10

1. Bearings

The layout of the bearings should correspond to the structural analysis of the whole
structure (super- and substructure together!). If the settlement and the deflection of the
substructure can be neglected the structural analysis of the superstructure, including
the bearings, can be separated from that of the substructure. Sometimes the model for
the analysis, especially of the superstructure, will be simplified by assuming the following: bearings are situated directly on the neutral axis of the girder (fig. 1.4.1-6),the
motion of the bearings occurs without restraint, bearings have no clearance, etc. In this
case we must consider the correct system (fig. 1.4.1-5) at least for the design of the
bearings and take into account the influence of the simplifications on the structure.

&
Fig. I .4.1-4: Reality

A
Fig. I .4.1-5: Correct system

On the abutments or separating piers it is normal to use at least two vertical bearings
to avoid torsional rotations. At intermediate piers one or more vertical bearings may
be used. If more than one bearing is used the rotational displacement at the pier is restrained. More than three vertical supports of the superstructure lead to statically indeterminate bearing conditions, but even the simplest bridge has at least four vertical
bearings. If the torsional stiffness of the superstructure is low (e.g. open cross sections) it may be neglected and the layout with four bearings becomes isostatic. If the
torsional stiffness is not negligible (e.g. box girders) we have to take it into account for
the structural analysis, especially for skewed and curved bridges. On a bridge with n
> 3 vertical supports, n - 3 bearing reactions can be chosen freely within a reasonable
bandwidth. This possibility can be used to prestress the superstructure and to distribute the bearing reactions as desired.
If the bearings are situated (nearly) in a plane we need at least one horizontally fixed
and one horizontally moveable bearing. The moving direction must not be orthogonal

11

I .4 The layout of bearings

to the polar line from the fixed to the moveable bearing. If more than two bearings in
the horizontal direction are necessary, the basic principle should be that an overall
uniform extension, caused by temperature or shrinkage, shall be possible without
restraint.

In general, there are two possibilities for the arrangement of the bearings:
a) arrangement in a horizontal position (fig. 1.4.1-7)
b) arrangement in a position parallel to the road or rail surface (fig. 1.4.1-8).

---_---,--a

Fig. 1.4.1 -7: Horizontal arrangement of the bearings (case a)

-(I

f=--

,,displaced

bridge

Fig. 1.4.1-8: Inclined arrangement ofthe bearings (case b)

Case a) has the advantage that only vertical bearing reactions and no permanent horizontal reactions result from vertical loads, but it has the disadvantage that bridges with
inclined gradients require a step at the expansion joint due to movements in the superstructure. The greater the elongation or shortening, the greater the step required.
Case b) has the advantage that the slope of the expansion joint is independent of the
movement of the bridge. The inclination of the surface of support gives the direction
of the normal force. Besides vertical reaction forces, also horizontal reaction forces
result from vertical loads. Permanent horizontal actions can lead to a displacement
by creep of the concrete and the soil and, thus, to crooked piers.

12

1. Bearings

1.4.2 The layout for different types of bridges

For single span girders the layout of the bearings is straightforward. One fixed and one
moveable bearing is provided on each abutment, all other bearings are just vertical
supports, moveable in any horizontal direction. For wide bridges the horizontally
fixed bearings are located in or near the bridge axis.
Formerly, the classical arrangement of the bearings for a bridge with two main girders consisted of one fixed and one lengthwise moveable bearing at one abutment and
one lengthwise moveable and one free bearing at the other abutment (fig. 1.4.2-1). This
layout has the advantage that longitudinal horizontal forces (braking and traction
forces) can be distributed into the two bearings at the abutment, but it has the
disadvantage that horizontal forces in the cross direction (wind) and temperature differences cause horizontal restraint forces, provided that bearings have no clearance on
the abutments.
The author prefers the statically determinate system with only one lengthwise restrained bearing at the abutment concerned because the actual clearance of a bearing
is not determinable in reality (fig. 1.4.2-2).

LA-:.

++11,

%I,

_-

- -- -

;c

Fig. 1.4.2-1: Classical layout

Fig. 1.4.2-2: Horizontally statically determinate system (better than classical layout)

- _ - _ _ - - - -------- --- .

Fig. 1.4.2-3: System with separated vertical and horizontal bearings (statically determinate system)

1.4 The layout of bearings

13

For skewed or horizontally curved single span bridges we have to decide whether the
horizontal force should be combined with the higher or with the lower vertical reaction force. For all bearing constructions it is easier to transfer horizontal forces in combination with a high vertical force. In this case the resultant force stays nearer to the
centre, its angle to the vertical is smaller and leads to smaller bending moments in suband superstructure (fig. 1.4.2-4).

!
I
I
HI

Fig. 1.4.2-4: Inclination of the resultantforce

Thus, the horizontally constrained bearings for skewed bridges should be placed at the
obtuse corners of the bridge, for curved bridges at the outer side (fig. 1.4.2-5).

Fig. 1.4.2-5: Skewed bridge

Fig.1.4.2-6: Layoutfor continuous girders

14

1. Bearings

For straight continuous girders normally two bearings are used at every abutment and
pier. If the torsional stiffness is high (box girder) the intermediate piers can be reduced
to a round column with one bearing on the axis under the diaphragm. Constrained
bearings in the cross direction are the rule at all piers. If the horizontal bending stiffness is very high we can transfer the horizontal forces only at the abutments. The same
considerations are suitable also for skewed and curved bridges (fig. 1.4.2-6).
Bearings for horizontal forces and guide bearings which transfer only horizontal
forces may be used in combination with leaf or link bearings which cannot transmit
horizontal forces.
The movement of an expansion joint must be linked by a guide like a constraint bearing. The main movement of an expansion joint should be in the axis of the traffic way.
Generally, this direction does not coincide with the direction of the polar line from the
fixed bearing to the moveable bearing at the abutment (fig.1.4.2-7). If all other
bearings have the same angle between the polar line and the moving direction there
results a layout of the bearings with no restraints on uniform elongation or shortening
(e.g. caused by thermal actions or shrinkage), as shown below (fig.1.4.2-8).

Fig.1.4.2-7: Layout for curved bridges

Fig. 1.4.2-8: Layout for curved continuous girders (no constraint under overall tempe ra ture)

Fig. 1.4.2-9: Geometrical situation

15

I .4 The layout of bearings

The elongation is
A,, = k . r, << r,

A- = k . r << r

proportional elongation

The rotation is

A;tana
r+A,

A;tana
r

k.r.tana
=k.tana
r

For Cp, = Cp, the bridge simply rotates as a rigid body without constraint.
One special case of this general rule is well known: the bearings are moveable in the
direction of the polar lines with a = 0 (fig.1.4.2-10). However, this layout has the
disadvantage that generally the main movement of the joint does not coincide with the
movement of the bearing.

Fig.1.4.2-10: Special case with a = 0

1.4.3 Special bearing conditions, advice etc.

It is important to note that the layout of the bearings has a great influence on the structural system. The above mentioned arrangements of bearings are typical for average
bridges. The following examples show some special effects which have to be considered for the design of bridges and bearings. These examples do not lay claim to completeness.
a) The already mentioned bearing layout, consisting of one bearing fixed in all sliding
directions and one fixed lengthwise at one abutment, leads to high constraint forces
not only under horizontal but also under eccentric vertical loading (fig.] .4.3-1). It
is interesting that this eccentric loading has no prying effect if the bearings are
situated directly on the neutral axis of the girder. This effect results only from the
(small) eccentricity of the bearing under the lower flange.

16

I . Bearings

A+
Fig. 1.4.3-1: Prying effect due to a eccentric loading
b) A similar situation occurs for a continuous girder with chequer pattern loading.

~
~

Fig. 1.4.3-2: Prying effect due to chequer pattern loading

c) It is not generally known that a skewed bridge with horizontally fixed bearings only in
one line exhibits the same effect under vertical loading, as the following figure shows:

Fig. 1.4.3-3: Prying forces for a skewed bridge with vertical loading
Similar effects can occur for curved bridges. For the correct analysis of the bearing
reactions it is always necessary to model the bearings at the very point where they
are actually situated, and in combination with the substructure. The deflection of the
substructure can influence the constraint bearing reactions significantly.

1.5

Calculation of bearing reactions and bearing movements

1.5.1 Actions
According to Eurocode 1 (ENV 1991) the actions can be subdivided into:
- permanent actions,
- variable actions,
- extraordinary actions.

1.5 Calculation of bearing reactions and bearing movements

17

The bridge should take up the desired shape under all permanent loads, at the average
temperature (+lo" C in most of the European countries) and, if time-dependant
displacements occur, at the time t = 00, at which time all moveable bearings should be
in the zero adjustment (null position). Variable actions and extraordinary actions lead
to deviation from this form.
Variable actions to consider are:
- traffic loads, considering the applicable dynamic coefficients
- loads due to traffic loads, i.e.
nosing forces
centrifugal forces
braking forces
traction forces
- wind loads
wind on construction
wind on traffic loads
- settlements of abutments and piers
- thermal actions
uniform temperature
vertical temperature gradient
horizontal temperature gradient
temperature differences between individual parts of the bridge (e.g. stay
cables, pylon and stiffening girder)
- creep and shrinkage of concrete
Extraordinary actions to consider are:
- earthquake actions
- vehicle impact
- derailment
- rupture of the conductor line
others
'

1.5.2 Bearing reactions

For permanent actions such as self-weight of the construction, dead load and prestressing, the bearing reactions can be calculated as one load case.
For the analysis of the bearings it is necessary to consider different combinations of
the bearing reactions:
- maximum vertical force and the adjacent horizontal force,
- minimum vertical force and the adjacent maximum horizontal force,
- maximum horizontal force and the adjacent maximum vertical force,
- maximum horizontal force and the adjacent minimum vertical force.
The simplest way to obtain these combinations is to calculate the variable actions, especially the traffic load, according to the influence line. One should bear in mind that
horizontal actions such as centrifugal forces or braking forces are proportional to the
vertical traffic load, but other loads, such as wind or traffic or traction forces for railways, are not.

18

1. Bearings

To obtain the extreme bearing reaction it is necessary to consider that all bridges are
three-dimensional and not merely plane systems.
The influence lines (influence surfaces) of the bearing reactions can be found as the
displacement curves (displacement surfaces) of the system, due to unit displacements
F = 1 or cp = 1, acting at the position and in the direction of the required force. If these
analyses are performed on a three dimensional model, the definitive influence area
will result directly (fig.1 S.2-1; fig.1 S.2-2). If plane models are used for the analyses,
special care is necessary, particularly with continuous girders with open or box sections. The following examples demonstrate the difference:

Fig.1.5.2-1: Influence area for the verticul bearing reaction A, box section.

Fig.1.5.2-2: ZnJuence areafor the vertical bearing reaction A, open section.

1.5.3 Bearing displacements

As already mentioned, the zero adjustment (null position) of every bearing has to be
defined. The displacements are measured from that position. Thus, for concrete and
composite bridges it is usual to consider displacements under time-dependent actions
such as creep and shrinkage from the time of installation of the bearing to the time defined for the null position (normally t = w), from which position the displacements due
to variable actions are measured.
To obtain the maximum displacements and rotations, again we can use influence lines.
The influence line of a displacement can be calculated as the displacement curve due
to the corresponding unit force P = I.
To take into account the imperfections due to installation, the temperature difference
for the calculation of bearing displacements should be assumed higher than for the
structural analysis of the bridge, or some additional displacement should be considered.

I .6 Construction of bearings

1.6

Construction of bearings

Fig. 1.6-1 gives un overview for the most common bearings.

19

20

1. Bearings

1.6.1 Elastomeric bearings

Elastomeric bearings are the simplest types of bearings. In the basic mode they consist merely of an elastomeric block (usually rectangular or round). The elastomeric
works as a soft part between sub- and superstructure and allows movements in all directions by elastic displacements or rotations. Under vertical loads the elastic block
bulges, leading to vertical displacements. A solution to this problem was found by reinforcing the elastic block by thin horizontal steel plates, vulcanized to the elastomer
(fig. 1.6.1- 1). The reinforcing plates prevent the block from bulging, thus leading to
very small vertical displacements, but they do not hinder horizontal displacements in
every direction and also allow small rotations in all directions. Every displacement
and rotation leads to restraining forces and moments which have to be taken into
account on the whole structure.
These restraining forces are possible if the friction between bearing and sub- and superstructure is sufficient. The friction forces F depend on the compressive force C and
the friction coefficient p, with F = C . p. If displacements take place under a small
compressive force, sliding between bearing and sub- or superstructure can occur. To
avoid this it is necessary to use elastomeric bearings with resistance to sliding. This
can be achieved by applying vulcanized plates on the bottom and on the top of the
bearing, which can be connected to the sub- and superstructure by bolts, pins or appropriate shapes (fig. 1.6.1-2).

Fig.1.6.1-1: Elastomeric hearing (unanchored)

Smaller, short time, horizontal forces can be transmitted by the restraining forces. If
these forces are higher or if they are permanent loads a restraining steel construction
is required. In these case the elastomeric bearing transmits the vertical force and
allows rotations, while horizontal forces in one or two directions are transmitted by
the steel construction (fig. 1.6.1-3 ; fig. 1.6.1-4).

Fig. 1.6.1-2: Elastomeric bearing (anchored)

1.6 Construction of bearings

21

Fig. 1.6.1-3: Elastomeric bearing constraint

Combination: elastomeric bearing and steel construction fixed in one direction.

Fig. 1.6.1-4: Fixed elastomeric bearing

Combination: elastomeric bearing and steel construction fixed in two directions.

1.6.2 Steel bearings

Steel bearings are the oldest type of bearings. They have been used for more than 100
years. The principle is simple: a flat plate rolls on another steel plate with a curved surface. If this surface is part of a sphere, theoretically we obtain a point tangency. If this
surface is part of a cylinder, theoretically we obtain a linear tangency. In the first case
we speak of point rocker bearings, in the second case of line rocker bearings. These
bearings allow rotations in all or in one direction, but they do not allow displacements
(fig. 1.6.2-1 ; fig.1.6.2-4).
Under minimal vertical reactions in combination with horizontal loads point rocker
bearings and line rocker bearings can exhibit damage of their connections, because of
tension. In combination with sliding elements these bearings are very sensitive to this
phenomenon, and it causes partial uplift and excessive wear as a result.
Linear tangencies can be found also in roller bearings consisting of a roll and a lower
and an upper plate (fig. 1.6.2-5). These bearings allow rotations in one direction and
displacements in one direction.
The problem with these bearings is a point or linear concentration of the bearing
force, which theoretically leads to infinite stresses. In 1881, the physicist Heinrich
Hertz found the solution of this problem: caused by the elastic deformation the theoretical point of tangency yields to a circle, the theoretical line of tangency yields to a
rectangle. The infinite stresses decrease to high but finite stresses, the so called Hertz
compression stresses over a very small contact zone. If the radius of the sphere or of
the cylinder decreases the Hertz stresses increase. From the local stress concentration
the stresses have to be distributed to the contact zones between bearing and sub- and
superstructure. Therefore, steel bearings normally need thicker plates for the stress
distribution than other types of bearings which transfer the bearing reactions over an
area.

22

1. Bearings

Point rocker bearings are used for bearing reactions in the range 500 and 2500 kN, line
rocker bearings and roller bearings for loads in the range 200 and 20 000 kN.

Fig.1.6.2-I: Fixed point rocker bearing

Fig. 1.6.2-2: Point rocker bearing constraint in one direction

Fig. 1.6.2-3: Free point rocker bearing

Fig. 1.6.2-4: Line rocker bearing

I .6 Construction of bearings

23

i
Fig. 1.6.2-5: Roller bearing (left side without guide rail; right side with guide rail)
The contact zones of steel bearings cannot be protected against corrosion. Therefore
corrosion-resistant layers of high alloyed steel should be used for the contact areas.
This can be done by building up a surface by forging or by welding. Between the mild
steel and the hardened high alloyed steel of the surface there should be a welded or
forged tough buffer zone. The thickness (in mm) of the hardened layer both on the
roller (radius R in mm) and of the plate should be t 2 0,14 . R - 2.

1.6.3 Pot bearings

These bearings were invented in the 1950s. They combine the two desirable properties: rotation capacity with a very small resistance and transmission of the bearing
reaction over a defined area.
The pot bearing consists of a steel pot, filled with an elastomeric disc and a lid or a
piston to the top (fig. 1.6.3-1). When subjected to high compression forces, the unreinforced elastomeric disc behaves similarly to a liquid. Rotations can occur due to the
nearly constant volume of the elastomer (v = 0,5). Of great importance is the sealing
between the elastomeric pad and the lid: if this sealing has a defect the elastomeric pad
escapes like a viscous liquid.
The standard type of pot bearing allows only rotation (fig. 1.6.3-2). Vertical forces are
transmitted to the pad, horizontal forces from the lid to the pot. To release one sliding
direction, an additional construction becomes necessary (fig. 1.6.3-3 and fig. 1.6.3-5).
This sliding construction consists of three components: a polytetrafluorethylene
(PTFE) disc, a surface of polished stainless steel connected to a sliding plate of structural steel and lubrication grease. PTFE is a plastic with high mechanical and chemical resistance, great toughness and very small friction when combined with polished
stainless steel. The PTFE disc is 5 to 6 mm thick, where half a thickness is enclosed by
the lid. This disc has small round pockets on the surface for the lubrication grease
(normally silicon grease) to reduce friction and wearing.
To constrain the movement in one direction an additional guide is used for the lid. This
guiding device allows movements in only one direction (fig. 1.6.3-3).
Pot bearings are used for vertical bearing forces from 1000 kN up to 100000 kN.
Depending on the standard applied the allowable compression between lid and elas-

24

1. Bearings

tomeric pad should not exceed 4.0 kN/cm2. The allowable compression for the PTFE
is 3 kN/cm2for permanent loads and 4.5 kN/cm2for short term loads (traffic, wind etc.).
Pot bearings have the advantage of a very high vertical stiffness (nearly incompressible elastomeric part). It is comparatively independent of the size of bearing and the
applied load. This characteristic is important for the bearing of high velocity railway
bridges. Bearings with low vertical stiffness can lead to damage of the rails.

astomere disc
Lid
Sealing
Pot - wall
Pot - bottom

Fig.1.6.3-1: Function of a pot bearing

Fig. 1.6.3-2: Fixed pot bearing

Fig. 1.6.3-3: Pot bearing constraint in one direction

25

1.6 Construction of bearings

Fig.1.6.3-4: Members of a pot bearing

Anchoring plate
Sliding plate
Polished stainless steel
PTFE (Polytetrafluorethylen)
Lid
Pot -wall
Sealing
Elastomere disc
Pot - bottom

Fig.1.6.3-5: Free pot bearing

1.6.4 Spherical bearings

The basic type of spherical bearing consists of three main parts: the pan, the part of a
sphere and the upper plate made of constructional steel (fig.1.6.4-1). To allow displacements between the parts, sliding surfaces are necessary. The pan has a PTFE
plate on the upper surface, the part of the sphere has a chrome-plated polished surface
on the underface and a PTFE plate also on the upper surface, and the upper plate has a
polished stainless steel plate on the underface. The PTFE plates are chambered over
half the thickness and have lubrication pockets with silicon grease, like the sliding
plates for pot bearings.
The friction resistance of the sliding parts causes reaction moments due to rotations.
They must be taken into account to consider additional design stresses of the bearing
material.
The vertical bearing reaction is transferred over the compressed areas of the PTFE.
The basic model is a moveable bearing (fig. 1.6.4-4). To constrain horizontal displacements an additional construction to connect the upper plate with the pan becomes
necessary (fig.1.6.4-2; fig.1.6.4-3).
British and Italian bearings have one sliding plane only and a deeper concave part to
take over horizontal forces (fig. 1.6.4-5). The construction must be checked for uplift
and exceeding the stresses in the contact area. In the bearings with two sliding planes
the centre of rotation is between the contact areas of the sliding surfaces, whereas in
Italian and British bearings it is somewhere in the bridge structure or in the pier or the
abutment.

26

1. Bearings

Like pot bearings, spherical bearings are used for vertical forces in the range of 1000
to 100 000 kN.

Polished

Sliding plate

hb
Part of sphere

PTFE

Chrome plated
polished surface

Fig.1.6.4-1: Members of a spherical bearing

Fig. 1.6.4-2: Fix spherical bearing

I
I

Fig. 1.6.4-3: Spherical hearing constraint in one direction

I
Fig. 1.6.4-4: Free spherical bearing

27

1.6 Construction of bearings

Fig. 1.6.4-5: Italian and British spherical bearing (one sliding s u f a c e )

1.6.5 Leaf and link bearings

All the above mentioned bearings are able to transfer compression forces. If tensile
forces as well as compressive forces must be transferred, leaf and link bearings are
used. These bearings can only transmit forces in the direction of the leaf. To transfer
forces in the crosswise direction, separate bearings must be used.
A leaf bearing consists of a foot plate, one or two lower leafs with pin holes and two
or one upper leaf with foot plate and pin holes, connected by a pin. Leaf bearings allow free rotation in one direction. Pin and pin holes must have a fit less than 0.3 mm,
as in cases of greater slackness and changing forces the pin will punch the hole. Pin
plate and pin should be of different types of steel to avoid seizure. Pin plates are made
of structural steel, pins often of tempered steel.
For link bearings a pendulum is linked to the foot leaf and to the upper leaf by pins.
Link bearings allow rotation and displacement in one direction. For pin holes and pins
the same rules apply as given for leaf bearings.
Link bearings with universal (Cardan) joints are used only in special cases. They
allow rotation and displacement in all directions.
Displacements 6 of link bearings are always combined with a small displacement 6,

62

in the perpendicular direction. 6, = __ , with R equal to the distance between the

2R
axes of the pins. Therefore this distance should not be too small.

1.6.6 Disc bearings

Disc bearings were introduced in the late 1960s. The vertical loads are transferred by
an elastomeric disc made of polyether-urethane polymer. In contrast to a pot bearing a
transverse extension of the elastomeric disc is possible. Bearing capacity and functioning is comparable with an elastomeric bearing. Rotations around the horizontal
axis are transferred by differential deflection of the disc. The rotations cause a shift of
the axis of the load from the centre of bearing, which must be considered in the design.
Horizontal forces are transferred by a shear-restriction device which allows vertical
deformation and rotation. The basic type is a fixed bearing. Free bearings are constructed by additional sliding elements and (if necessary) guiding systems.

Next Page

28

1. Bearings

Fig. I . 6.6-I :Fixed bearing

Fig. 1.6.6-2: Uni-directional guided

-top plate

bearing
assembly

base plate

Fig. I . 6.6-3: Multi-directional non-guided

Previous Page
1.7 Materials for bearings

1.7

29

Materials for bearings

1.7.1 Steel
Structural steel
Structural steel is used for all parts of bearings which are not under extraordinary
local stress or do not require special properties against corrosion. Structural steel for
bearings can be:
- Non-alloy structural steels according to EN 10025
- Fine-grained structural steels according to EN 10113
- Quenched and tempered steels according to EN I0082
Eurocode 3 may be used for the design of all bearing components made from structural steel according to EN 10025 and EN 10113 and for all connections (bolts, welds
etc.). Quenched and tempered steels are used mostly for non-welded parts under high
pressure (parts with Hertz compression, bolts of leaf and link bearings). In contact
areas with Hertz compression layers of corrosion-resistant hard steel can be applied
by forging or by welding. In the case of hard-surface welding a tough intermediate
(puffer) layer must be welded between the steel and the hard-surface.

Stainless steel
Stainless steel according to EURONORM 88-2 or I S 0 683 can also be used for bearings. For design one should use EC 3 , part 1-4. Concerning stainless steel for sliding
plates see 1.7.3.

1.7.2 Elastomeric parts

Elastomeric parts of bearings consist normally of natural or artificial (chloropren) rubber (NR or CR, respectively). Artificial rubber has the same good properties as natural rubber, and in addition it has a higher resistance against ozone, ultraviolet radiation
and ageing and is more rigid. The characteristic mechanical property is the shear modulus G between 0.7 and 1.15 N/mm2 at room temperature, decreasing with increasing
temperature. When undergoing stress changes the volume of rubber is nearly constant.
So we have a Poissons ratio v = 0.5 and a Youngs modulus of elasticity E =
2 . ( 1 +v) . G -- 3 . G. The fracture strain of rubber lies between 250 % and 500 %. Rubber creeps under stress by up to 50 % of the elastic strain, but creeping ends within
some days or weeks. Rubber does not break under compression, it can only break
under tensile or shear stresses. Compressing a rubber pad changes its shape. The
changing of the shape depends on the possibility of displacement at the compressed
areas. If the compressed areas are fixed to a rigid surface, the displacement remains
small. Thus we obtain the inequality A, > A , > A3 (fig.1.7.2-1).

Fig.1.7.2-I : Vertical displacements depending on the lateral expansion

30

1. Bearings

Fig. 1.7.2-2: Stress distribution

If the surface of the rubber is fixed to a rigid body shear stresses develop between the
two surfaces under compression (fig. 1.7.2-2). Under compression we obtain a virtual
modulus of elasticity E, Lllmpr which depends not only on the shear modulus G but also
on the thickness of the part between two plates. For rectangular parts a good approximation for E, co,npr is given by

1'

conipr

=G

(: )

. (1

0,6

g)

for b 2 a

The maximum stresses under compression between two rigid bodies are

with o = -, F: compression force.

ab
For bending, the effective modulus of elasticity E, bcndlng is lower than E, i<,,,,pr because
we obtain a compression in two half waves under a constant rotation angle a.If both
halves develop a constant displacement, the virtual modulus of elasticity would be the
a
1
same as under compression, but with a + we obtain E, hendlng = el Lo,npr. Actually,
2
4
~

31

1.7 Materials for bearings

is not in the middle of one half but nearer the outer side; thus we
a
- -1 E l compr. This is described very well by the
finally obtain: a + < , El
2
following approximate formula:
the maximum

(3

for b 2 a
Under the rotation a we obtain a curvature p =
and a restraining moment

a=

Mi?
bending

with I =
'

a'b
12

_____

Fig. 1.7.2-3: Rotution - restraining moment

Fig. 1.7.2-4: Displacement - restruining~forces

1.7.3 Sliding elements

For sliding elements in constructional bearings it is normal to use PTFE, also known
by the registered trade names Teflon and Hostaflon. PTFE is a so called thermoplast.
For bearings it is used in the original (virgin) condition, i. e. not sintered and without
fillers. As a counterpart to this rather soft material polished stainless steel plates are
normally used, and sometimes acetal resin plates or hardened chromium-plated steel
plates. Chromium-plated steel plates are not resistant to fluorine ions and are rather
prone to corrosion than stainless steel plates. They are allowed for convex elements
only.
The combination of a soft and a hard part has the advantage that there is no danger of
cold welding which can occur on polished metal or plastic surfaces under high pressure. To minimise the friction silicon grease should be used to provide lubrication. To
keep this grease between the two surfaces the PTFE has lubricant pockets on its surface, so that a permanent lubrication takes place over several years. The PTFE plates
for bearings are normally 5 to 6 mm thick, the depth of lubricant pockets is 2 mm. Un-

32

1. Bearings

der pressure the PTFE yields. To keep the PTFE in the desired shape it is necessary to
keep about half the thickness in a <<chamber>>
with sharp edges. Over the sharp edges
we obtain a small bulge. It is also possible to glue PTFE to a steel surface. In this case
the PTFE is about 2.5 mm thick.
The friction coefficient increases with decreasing temperature and with decreasing
compression. The static friction coefficient (first movement) is higher than the dynamic coefficient. After movement has taken place the dynamic friction coefficient remains at this value and returns to the static value after a few hours. This might depend
on the orientation of the large polymer molecules; during movement they are orientated into the direction of motion within a very thin surface layer. When the motion is
stopped, the orientation is lost within a few hours. Fig. 1.7.3-1 shows the design values of the friction coefficient pLdbetween PTFE and stainless steel, depending on the
compression force (EN 1337-2).

I
I

0.00

0.0

I
I

0.5

1.o

1.5

I
I

I
I
I

2.0

2.5

3.0

p [kNicm']

Fig. 1.7.3-I : Friction coejficient depending on the compressionforce

The design value of the ultimate compression load is

f , = 6 , 5 (1 - 0,02. [6 - 30'C)) kN/cm2 for 6 2 30'C ,

6 : maximum temperature of the bearing.
The wearing of the PTFE depends on
a) the product of compression and velocity of the displacement
b) the total amount of sliding during the life-time
c) the lubrication of the surface (a loss of lubrication leads to extremely high wearing)
d) the roughness and the hardness of the stainless steel surface
e) the contact pressure near the edge of PTFE (ironing effect)

33

I .8 Analysis and design of bearings

For slow movements caused by thermal actions we obtain long sliding movements but
at a low velocity. Quick movements caused by traffic loads have short sliding movements but they occur at high velocity. Wearing is mostly caused by the second case.
For the stainless steel plate, austenitic steel X6CrNiMo17122 according to EURONORM 88-2, surface n (IIIc), should be used. The stainless steel plate must cover
the PTFE plate completely in all situations. The thickness of the plate should be at
least of 1 .5 mm. The connection to the carrying plate of mild steel can be welded or
glued. For 2.5 mm thick plates the connection can be riveted or bolted.

1.8

Analysis and design of bearings

1.8.1 Hertz compression

For the design of bearings the following problems should be addressed: compression
between two spherical bodies, compression between a spherical and a flat body, compression between two cylindrical bodies, compression between a cylindrical and a flat
body along a generator line. As already mentioned, Heinrich Hertz obtained the solution under the following assumptions (1881):
1. The two bodies consist of isotropic, homogeneous and infinitely elastic materials.
2. Only normal stresses (no shear stresses) occur at the contact areas.
3. The radius (width) of the contact areas is small compared with the radii of the
involved bodies.
Hertz found the following maximum compression stresses max (T and widths b on the
contact areas:
Spherical body on spherical body

3F(I-v2) .

b=
7

2E

1
1

-f-

= 1,109
73

Cylindrical body on cylindrical body

-*-

34

1. Bearings

with

F
1

rl

r2

~-~

Fig. I .8. I - I b: Arrangement of the radii

bearing reaction
length of the cylinder
radii of the bodies in contact
Young's modulus
Fig. 1.8.1-2: Stress distribution
Poisson's ratio (v = 0.3 for steel)
maximum normal stress at the contact area
half the width of the contact zone

r,, r2
E
V

max (3
b

For the usual rocker or roller bearings the max (3 beneath the vertical bearing reaction
greatly exceeds the material yield strength (fig. 1.8.1-2). However, at the contact zone
we have not only vertical but also horizontal compression stresses. According to the
von Mises criterion the comparison stress
I

Ov =

d0i2 + O2

+ Oj3 - (3~(32 -

- O3Oi and yielding begins when

reaches
the material yield strength f,. In the present three-dimensional compression regime,
(3" will be less than (3, and yielding will not begin until o1= f,. On the other hand, the
maximum strain does not occur at the surface in the middle of the compression zone,
so that the hardness of the surface is not the only criterion for the assessment of Hertz
compression.

EN 1337-4 - roller bearings

gives for the design line load pd of a roller bearing

(cylindrical body on flat surface): pd 5 18. R .

f 2
E d

tensile strength of the material

R radius of the cylinder
Eddesign value of the modulus of elasticity
f,

with

35

1.8 Analysis and design of bearings

Compared to Hertz's formula with

maxo, =0.418.

we find
maxo, 1 0 . 4 1 8 . f i . f " = 1,77.fu=oRd
.

EN 1337-6 - rocker bearings - gives for the design load Fz,dof a point rocker bearing
(sphere against plane surface) Fz,d5 170. R 2 .

f"

Ed

Compared to Hertz's formula with

we find
m a x o , 10.388..1/170.f, = 2,15f, =oRd.
For cylindrical rocker bearings the same formulae as for roller bearings are used.

1.8.2 Pin and pin plate for leaf and link bearings
A special problem of all leaf and link bearings concerns the design of the pin and the
pin plate. Eurocode 3, part 1- 1, gives simple but satisfactory design rules. The design
values of the shear force and the bending moment for the pin can be found using the
simple model of distributing the force of each pin plate uniformly over the pin.

c ,

In the case of fig. 1.8.2-1 we obtain the shear force and the bending moment according
to fig. 1 3.2-2 and fig. 1.8.2-3.

1. Bearings

36

c
w
Fig. 1.8.2-2: Shearforce

Fig. 1.8.2-3: Bending moment

For normal bridge bearings we have: c = 0, a =

b
.
2

The design values for the resistances are

Shear: F,,,

= 0.6. A . f u p/YMp = 0.6.

d2n
~

.fup/YMp = 0.47 1. df,,

/YMp

The combination of shear and bending has to fulfil the inequality

In this inequality, the central pin plate is controlling.

The bearing resistance of plate (thickness t and yield strength f,) and pin is:

F,,,, = 1.5.t . d . f y/YM,

f,, field strength of the pin
fUptensile strength of the pin
yMp= 1.25 according to EC 3- 1- 1
The bearing capacity of the pin plate at the hole is achieved under one of the following
conditions (EC 3- 1 - 1 gives two possibilities):

37

1.9 Installation of bearings

a) Depending on the pin plate thickness t:

t = min (2a, b),
e >--FSd Y M p
-7 - 2t ' f y
'

e, 2

FSd

'

YMp

2 t . f,

d
= e 2 +3
3

b) Depending on the geometry of the pin plate:

1.9

Installation of bearings

Concerning the installation of bearings, the need for a later simple replacement must
be taken into account. So it should be common practice to put every bearing between
a lower and an upper steel cover plate. These cover plates are anchored or connected
both with the substructure and the superstructure. These cover plates are connected to
the bearings during the installation but remain fixed to the structure while the bearings
are replaced (fig. 1.9-1). Thus, the connection between bearing and cover plates should
be constructed in order to allow a simple release. Bolted connections are often used
but after many years often the bolts can hardly be unscrewed. According to the
author's experience, fastening the bearings with small fillet welds that can be ground
off and remade during the replacement process is simpler.

Fig. 1.9-1: Fixing of a bearing

Generally, bearings should not be built directly on the construction beneath. To guarantee that the area below a bearing is fully sealed a layer of mortar or of a similar product is used. So the height of the bridge at the abutments or piers can be adapted easily
and very exactly. It is useful to fix the bearing to the bridge so that there is no clearance at the upper plate and to adjust the bridge by hydraulic jacks. In this situation the

38

1. Bearings

bearings should be adjusted exactly. Thus, the lower plate will get exactly the desired
inclination (horizontal or parallel to the gradient, see fig. 1.9-1) and all moveable bearings will have the desired pre-adjustment, which depends on the temperature of the
bridge and the expected shrinkage and creep. The installation of the bearings should
be done early in the morning when the bridge has a (nearly) constant temperature. The
designer has to provide a table with the pre-adjustment of every bearing depending on
the measured bridge temperature.
For good functioning, careful handling of the bearings during installation is very important. The bearings must be kept free of dirt, mortar, water and dust, especially from
all moving parts. Many bearings, such as pot bearings and spherical bearings, are protected against dust by rubber bulges, but others are not protected at all. These have to
be cleaned to remove mortar and sand after the installation.
The gap between the lower plate of the bearing and the substructure is normally 3 to 5
cm thick and must be completely filled with a mortar bedding. This can be done in different ways:
- by a fresh mortar bedding, chambered in the centre where the bearing is set. The
excess of mortar will come out on all sides and must be removed.
- by a special joint filling mortar which must be mixed in a pan type concrete mixer
with a precise quantity of water. This mortar is liquid at first and should be poured
in a formwork around the bearing only from one side, so that the air can escape on
the other side. The special mortar fills the gap without air bubbles, it sets and hardens very quickly so that after one day the mortar bedding can be fully loaded and
the formwork removed. If the gap is less than 1 cm a two-component epoxy resin
should be used instead of mortar. Initially this resin is a lighter fluid than mortar,
thus completely filling even very small gaps.
- by boxing up earth-damp mortar in the gap with a wooden stick also from one side
to avoid air bubbles. This method will be difficult for the lower plates with a short
side larger than half a metre.
All mortars should be non-shrinking.

1.10 Inspection and maintenance

Visual tests of all bearings should be done by qualified personnel at regular intervals.
The following properties of the bearings have to be checked:
a) sufficient ability to allow movement, taking into account the temperature of the superstructure
b) correct positioning of the bearings themselves and of parts of the bearing relative to
each other
c) uncontrolled movement of the bearing
d) fracture, cracks and deformations of parts of the bearings
e) cracks in the bedding or in adjacent parts of sub- and superstructure
f) condition of the anchorage
g) condition of sliding or rolling surfaces
h) condition of the anticorrosive protection, against dust, and of the sealings.
For the different types of bearings the following checks are of importance:

39

1.1 1 Replacement of bearings

Elastomeric bearings: Displacements and rotations, cracks in the elastomer.

Roller and rocker bearings: Displacements and rotations, adjustment of the
guiding device, no gap in the contact line.
Pot bearings: Sufficient mesh of the lid in the pot, tight sealing of the elastomer
in the pot (if the sealing has a defect, the elastomer comes out like a pancake!)
Sliding devices - PTFE and stainless steel: Thickness of the PTFE, clean surface of
the stainless steel.
The result of an inspection should be recorded in a report. EN 1337-10 gives an example for such a report.
For maintenance the bearings should be cleaned, lubricated (if necessary and possible) and coated with paint. Small defects should be repaired as far as possible.

1.11 Replacement of bearings

The replacement of bearings is a normal maintenance operation for bridges. Thus, a
bridge designer has to provide measures so that a replacement can be carried out
easily. The owner of a bridge has to define in the tender if the replacement of the bearings must be carried out under full traffic, restricted traffic or without traffic, depending on the importance of the bridge and the possibility of a traffic ban or a traflk
diversion.
In case of a replacement under traffic the jacking equipment should allow the same
movements as the bearing. To allow rotations the jacks around one bearing should be
connected to a single hydraulic circle. That means that the security devices must have
a sufficient clearance. Translations are possible by means of additional sliding constructions.

i
_m_

reinforcement against splitting tension

Fig. 1.I I - I : Stiffened areasf o r hydraulic jacks

To replace a bearing, the bridge has to be lifted by one or more hydraulic jacks. For hydraulic jacks, adequately stiffened areas to transmit the hydraulic jack forces to the
sub- and superstructure are required. Concrete parts must be reinforced against splitting tension, steel parts need stiffeners (fig. 1.11-2). Thus, the construction drawings
must show in which areas or at which points hydraulic jacks can be set, what are the
maximum lifting forces and up to which level the bridge may safely be lifted. This

40

1. Bearings

data is of particular importance if the bridge is supported in a statically indeterminate

way at one abutment or pier, in which case the lifting force depends on the height of
lift. High stresses can be induced in the cross girder or diaphragm by the lifting device.
In such cases it may be necessary to lift the whole cross section uniformly with two or
more hydraulic jacks even for exchanging only one bearing. If more than one jack is
used the forces can be controlled by hydraulic connection of some or of all jacks: all
connected jacks have the same pressure. Hydraulic jacks need some clearance for the
installation. For lifting by a few millimetres up to two centimetres flat piston jacks can
be used. The following table gives a guide for the required clearances:

Force
kN
500
1000
2000
SO00

Required clearance
Normal hydraulicjack
mm
300
360
450
600

Required clearance
Flat piston jack
mm
150
180
200
250

Table 1.11-1: Required clearance for hydraulic jacks

There are flat jacks with a height of 80 mm and a lifting force up to SO00 kN. But their
stroke is only 20 mm and there is no security device. This kind of jack should be applied for special cases only. New bridges should be constructed for normal hydraulic
jacks.
In all situations, during the replacement of a bearing the hydraulic jack should be secured by a mechanical device such as an adjusting nut for the piston or lining plates to
avoid dropping in case of pipe rupture or rupture of the piston sealing which sometimes can occur (fig.l.11-3 and tig.l.11-2).

!!

pipe or

t-------I
L

----____

c=

Fig. 1.1 1-2: Hydraulic jack with lining plates

1.12 Codes and standards

41

Fig.1.I 1-3: Hydraulic jack with thread and nu1

If the replacement of a bearing takes a long time so that displacements of moveable
bearings will occur, the hydraulic jacks have to be equipped with a sliding device,
normally PTFE plus a sliding plate of stainless steel.
Particular care is required when replacing bearings which transmit horizontal forces:
if the friction between the jack and the surface of sub- and superstructure is not sufficient it is necessary to restrain the movement of the bridge by appropriate devices. If
the replacement is done under traffic, in most cases, and especially for railway
bridges, these devices have to transmit all horizontal forces due to a possible loss of
friction.

1.12 Codes and standards

The first attempts to standardize bearings in national codes were made decades ago. In
Europe several codes and national standards are available. The best known national
standards in Europe on this topic are
Germany:
DIN 4141 Lager im Bauwesen (structural bearings),
Teil 1 bis 14.
United Kingdom: BS 5400 Steel, Concrete and Composite Bridges.
Section 9.1 Code of Practice for design of bridge bearings
Section 9.2 Specification of materials, manufacturing and installation of bridge bearings
New European Standards about bearings are the following
EN 1337
Structural bearings with the parts
EN 1337- 1 General design rules
EN 1337-2 Sliding elements
EN 1337-3 Elastomeric bearings
EN 1337-4 Roller bearings

42
EN
EN
EN
EN
EN
EN
EN

1. Bearings

1337-5
1337-6
1337-7
1337-8
1337-9
1337-10
1337-1 1

Pot bearings
Rocker bearings
Spherical and cylindrical PTFE bearings
Guided bearings and Restrained bearings
Protection
Inspection and maintenance
Transport, storage and installation

A recommendable American Standards about bearings is the following:

AASHO-LRFD: American Association of State Highway Officials ( I 994).

1.13 References
Books and special chapters about bearings for bridges:
Eggert H., J. Grote, W. Kauschke: Lager im Bauwesen. Verlag von Wilhelm
Ernst & Sohn, Berlin, Munchen, Dusseldorf 1974.
Lee D.J.: Bridge Bearings and Expansion Joints. Second edition by E & FN Spon,
London, Glasgow, New York, Tokyo, Melbourne, Madras 1994.
Eggert H., W. Kauschke: Lager im Bauwesen. 2. Auflage, Ernst & Sohn, Berlin 1995.
Rahlwes K., R. Maurer: Lagerung und Lager von Bauwerken in: Beton-Kalender
1995, Teil2, Ernst & Sohn, Berlin.

Papers:
Albrecht, R.: Zur Anwendung und Berechnung von Gummilagern. Der Deutsche Baumeister 1969, Heft 4, Seite 326, und Heft 6, Seite 563.
Andra, Beyer, Wintergerst: Versuche und Erfahrungen mit neuen Kipp- und
Gleitlagern. Der Bauingenieur 5 (1962).
Andra, W. und Leonhardt, F.: Neue Entwicklungen fur Lager von Bauwerken,
Gummi- und Gummitopflager. Die Bautechnik 39 (1969), Heft 2, Seite 37 bis
50.
Bayer, K.: Auflager und Fahrbahnubergange fur Hoch- und Bruckenbauten aus
Kunststoff. Verein Deutscher Ingenieure VDI im Bildungswerk BV 1956 (Vortragsveroffentlichung).
Beyer, E. und Wintergerst, L.: Neue Briickenlager, neue Pfeilerform. Der Bauingenieur 35 (1960), Heft 6, Seite 227 bis 230.
Eggert, H.: Briickenlager. Die Bautechnik 50 (1973), S. 143/144.
Bub, H.: Das neue Institut fur Bautechnik. Strasse und Autobahn, Band 20
(1 969), Seite 189.
Burkhardt, E.: Gepanzerte Betonwalzgelenke, Pendel- und Rollenlager. Die
Bautechnik 17 (1939), Seite 230.
Cardillo, R. und Kruse, D.: Paper (61/WA-335) ASME (1961).
Cichocki, F.: Bremsableitung bei Briicken. Der Bauingenieur 36 (1961), Seite
304 bis 305.

1.13 References

43

Clark, E. und Moutrop, K.: Load Deformation Characteristics of Elastomer

Bridge Bearing Pads. University of Rhode Island, May 1962.
Desmonsablon, Philippe: Le calcul des piles ddformables avec appuis en
caoutchouc. Annales des Ponts et Chaussdes, Paris 4/1960.
Eggert, H.: Bauwerksicherheit bei Verwendung von Rollen- und Gleitlagern.
Strasse Brucke Tunnel 1971, Heft 3, Seite 71.
Eggert, H.: Die baurechtliche Situation bei Lagern fur Briicken und Hochbauten. Der Stahlbau 39 (1970), Heft 6, Seite 189.
Einsfeld, U.: Erlauterungen zu den Richtlinien von unbewehrten Elastomerlagern. Mitteilungen Institut fur Bautechnik 6/1972.
Franz: Gummilager fur Brucken. VDI-Zeitschrift, Bd. 101/1959, Nr. 12, Seite
47 1 bis 478.
Gent, A.: Rubber Bearings for Bridges. Rubber Journal and International Plastics 1959.
Grote, J.: Neoprenelager - einige grundsatzliche Erwagungen. Kunststoffe im
Bau 7/1968.
Grote, J.: Unbewehrte Elastomerlager. Der Bauingenieur 44 (l969), Seite 121.
Grote, J.: Vermeidung von Rissen und Dehnungsschaden durch gummielastische Lagerungen. Kunststoffe im Bau 11/1968.
Hakenjos, V.: Untersuchungen uber die Rollreibung bei Stahl im elastisch-plastischen Zustand. Technisch-wissenschaftlicheBerichte der Staatlichen Materialpriifungsanstalt an der Technischen Hochschule Stuttgart 1967, Heft 67/05.
Heesen: Gepanzerte Betonwalzgelenke, Pendel- und Rollenlager. Die Bautechnik, Jahrgang 25 (1 948), Seite 26 1.
Hutten, P.: Beitrag zur Berechnung der Lagerverschiebungen gekrummter,
durchlaufender Spannbeton-Balkenbriicken. Dissertation TH Aachen 1970.
Jorn, R.: Gummi im Bauwesen. Elastische Lagerung einer Pumpenstation. Der
Bauingenieur 36 (1961), Heft 4, Seite 1371138.
Keen: Creep of Neoprene in Shear Under Static Conditions, Ten Years, Transactions of the ASME, Juli 1953.
Leonhardt und Andra: Stutzungsprobleme der Hochstrassenbriicken. Betonund Stahlbetonbau 55 (1960), Heft 6.
Leonhardt, F. und Reimann, H.: Betongelenke, Versuchsbericht, Vorschlage
zur Bemessung und konstruktiven Ausbildung. DAfStb, Heft 175. Berlin:
Verlag Ernst & Sohn 1966, und Leonhardt, F. und Reimann, H.: Betongelenke.
Der Bauingenieur 41 (1966), Seite 49.
Leonhardt, F. und Wintergerst, L.: Uber die Brauchbarkeit von Bleigelenken.
Beton- und Stahlbetonbau 1961, Heft 5, Seite 123 bis 131.
Maguire, C. und Assoc.: Elastomeric Bridge Bearings Pads 1959.
Massonnet: Zuschrift zu B. Topaloff, Gummilager fur Briicken. Der Bauingenieur 39 (1964), Seite 428.
Monnig, E. und Netzel, D.: Zur Bemessung von Betongelenken. Der Bauingenieur 44 (1969), Seite 433 bis 439.
Morton, M.: Rubber Technology. Reinhold Publishing Co. 1959.
Mullins, L.: Softening of Rubber by Deformation. Rubber Chemistry and
Technology (Feb. 1969).

44

[351
[361
[371

[431
[441

[491

[531

1. Bearings

Nordlin, E., Stoker, S. and Trinble, R.: Laboratory and Field Performance of
Elastomeric Bridge Bearing Pads, Highway Research Board (1968).
Pare u. Keiner: Elastomeric Bridge Bearings. Highway Research Board Bull
242, 1960.
Payne u. Scott: Engineering Design with Rubber
Rejcha, C.: Design of Elastomer Bearings. Journal of Prestressed Concrete
Institute Oct. 1964, Vol. 9, Nr. 5.
Resinger, F.: Langszwangungen - eine Ursache von Bruckenlagerschaden. Der
Bauingenieur 46 (1971), Seite 334.
Rieckmann, H.-P.: Einfluss der Lagerkonstruktion auf die Knicklange von
Pfeilern. Strasse Briicke Tunnel 1970, Seite 36 bis 42 und Seite 270 bis 272.
Sasse, H.-R. und Schorn, H.: Bewehrte Elastomerlager - Stand der Entwicklung. Plastik-Konstruktion 1971, Heft 5 , Seite 209 bis 227.
Schonhofer: Neugestaltungen auf dem Gebiet des Auflagerbaues und auf verwandten Gebieten. Werner-Verlag, Dusseldorf 1952.
Sedyter: Uber die Wirkungsweise von Bleigelenken. Beton und Eisen 1926,
Seite 29.
Shen, M. K.: Uber die Losung des Balkens mit unverschieblichen Auflagern.
Der Bauingenieur 39 (1964), Seite 100.
Suess, K. und Grote, J.: Einige Versuche an Neoprenelagern. Der Bauingenieur
38 (1963), Heft 4, Seite 152 bis 157.
Thielker, E.: Elastomeric Bearing Pads and Their Application in Structures,
Paper 207 of Leap Conference (1964).
Thul, H.: Bruckenlager. Der Stahlbau 38 (1969), Seite 353.
Topaloff, B.: Gummilager fur Briicken - Berechnung und Anwendung. Der
Bauingenieur 39 (1964), Seite 50 bis 64.
Topaloff, B.: Gummilager fur Brucken. Beton- und Stahlbetonbau 54 (1959),
Heft 9.
Uetz, H. und Breckel, H.: Reibungs- und Verschleissversuche mit Teflon.
Sonderheft der Staatl. Materialprufungsanstalt an der TH Stuttgart, 7.12.1964,
Seite 61/76.
Uetz, H. und Hakenjos, V.: Reibungsuntersuchungen mit Polytetrafluorathylen
bei hin- und hergehender Bewegung. Die Bautechnik 44 (1967), Heft 5, Seite
159 bis 166.
Uetz, H. und Hakenjos, V.: Gleitreibungs- und Gleitverschleissversuche an
Kunststoffen. Kunststoffe, 59. Jahrgang 1969, Heft 3, Seite 161 bis 168.
Weiprecht, M.: Auflagerung von Briicken. Elsners Taschenbuch fur den Bautechnischen Eisenbahndienst, 1967, Seite 23 1 bis 277, Abschnitt E Bruckenund Ingenieurhochbau.
Zies, K.-W.: Stabilitat von Stutzen mit Rollenlagern. Beton- und Stahlbetonbau
65 ( 1 970), Seite 297.
AASHO-LRFD: American Association of State Highway Officials (1994).
Dupont de Nemours Co.: Design of Neoprene Bridge Bearing Pads, Wilmington ( 1959).
CNR-UNI 10018-68 (Italian Standards for rubber bearings).

1.13 References

45

Ministry of Transport: Provisional Rules for the Use of Rubber Bearings in

Highway Bridges, Memo. 802, London (1962).
Mitteilungen, Institut fur Bautechnik, 1970, Heft 2 und 4, und 1971, Heft 4
und 6.
Ohne Verfasser. Auflager aus Teflon. Ausziige aus dem Journal of Teflon 1964,
1965 und 1966, Druckschrift der Du Pont de Nemours International S.A.
Geneva, Switzerland.
Ohne Verfasser. Bruckenlager. Beratungsstelle fur Stahlverwendung, Dusseldorf, Merkblatt 339,2. Auflage 1968.
ORE Office de Recherches et dEssais: Verwendung von Gummi fur Bruckenlager, Frage D 60, Utrecht (1962, 1964, 1965).
Wiedemann, L.: Zusatzliche Richtlinien fur Lager im Brucken- und Hochbau.
Mitteilungen Institut fur Bautechnik 3/1973, S. 73. Verlag Ernst & Sohn.
Eggert: Vorlesungen uber Lager im Bauwesen. Wilhelm Ernst & Sohn
1980/1981.
Kauschke, W.: Entwicklungsstand der Gleitlagertechnik fur Briickenbauwerke
in der Bundesrepublik Deutschland. Bauingenieur 64 (1989), Seite 109
bis 120.
BattermandKohler: Elastomere Federung, Elastische Lagerungen. W. Ernst &
Sohn, Berlin, Munchen 1982.
Gerb: Schwingungsisolierungen. Berlin, 9. Auflage 1992, Eigenverlag (gegen
Schutzgebuhr erhaltlich).
Grote, J. und Kreuzinger, H.: Pendelstutzen mit Elastomerlagern. Der Bauingenieur 53 (1978), Seite 63/64.
Kanning, W.: Elastomer-Lager fur Pendelstutzen - Einfluss der Lager auf die
Beanspruchung der Stutzen. Der Bauingenieur 55 ( 1 980), Seite 455.
MauredRahlwes: Lagerung und Lager von Bauwerken. Betonkalender 1995,
Ernst & Sohn, Teil 11.
Weihermuller, H. und Knoppler, K.: Lagerreibung beim Stabilitatsnachweis
von Bruckenpfeilern. Bauingenieur 55 (1980), Seite 285 bis 288.
Andra, W.: Der heutige Entwicklungsstand des Topflagers und seine Weiterentwicklung zum Hublager. Bautechnik (1984), Seite 222 bis 230.
Eggert, H.: 7 Grundsatze bei der Lagerung von Brucken. 9. IVBH-Kongress
Amsterdam 1972, Schlussbericht. Internationale Vereinigung fur Briickenbau
und Hochbau, Zurich, Schweiz.
Deinhard, J.M., Kordina, K., Mozahn, R., Storkebaum, K.-H.: Der Schadensfall an der Mainbrucke bei Hochheim. Beton - Stahlbetonbau, 72 (1977), Seite
1 bis 7.
Eggert, H. und Wiedemann, L.: Nutzungsgerechte Lagerung von Stahl- und
Verbundbrucken und unterhaltungsgerechte Konstruktion von Bruckenlagern.
IVBH Symposium Dresden 1975. Vorbericht.
Eggert, H.: Lager fur Brucken und Hochbauten. Bauingenieur 53 (1978), Seite
161 bis 168, und Zuschrift 54 (1979), Seite 200.
Konig, G. et. al.: Spannbeton: Bewahrung im Bruckenbau. Analyse von Bauwerksdaten, Schaden und Erhaltungskosten. Springer-Verlag Berlin, Heidelberg, New York, London, Paris, Tokio 1986.

46

1. Bearings

Pfohl, H.: Reaktionskraft am Festpunkt von Briicken aus Bremslast und Bewegungswiderstanden der Lager. Bauingenieur 58 (1983), Seite 453 bis 457.
Eggert, H. und Hakenjos, V.: Die Wirkungsweise von Kalottenlagern. Der Bauingenieur 49 (1974), Heft 3 , Seite 93/94.
Lehmann, Dieter: Beitrage zur Berechnung der Elastomerlager. Die Bautechnik I (1978), Seite 19 bis 22, I1 (1978), Seite 99 bis 102, I11 (1978), Seite 190
bis 198, IV (l979), Seite 163 bis 169.
Kordina, K. und Nolting, D.: Zur Auflagerung von Stahlbetonteilen mittels
unbewehrter Elastomerlager. Der Bauingenieur 56 (1981), Seite 41 bis 44.
Kordina, K. und Osterath, H.-H.: Zur Auflagerung von Stahlbetonteilen mittels
unbewehrter und bewehrter Elastomerlager. Der Bauingenieur 59 (1 984),
Seite 461 bis 466.
Kessler, E. und Schwerm, D.: Unebenheiten und Schiefwinkligkeiten der Auflagerflachen fur Elastomerlager bei Stahlbetonfertigteilen. Fertigteilbauforum 13/83, Seite 1 bis 5 (Betonwerk + Fertigteil-Technik).
Kessler, E.: Die Anwendung unbewehrter Elastomerlager. Betonwerk + Fertigteil-Technik, Heft 6 (1987), Seite 419 bis 429.
Bundesminister fur Verkehr: Schlden an Brucken und anderen Ingenieurbauwerken. Dokumentation 1982. Verkehrsblatt-Verlag, Dortmund.
Bundesminister fur Verkehr: Bericht uber Schaden an Bauwerken der Bundesverkehrswege. Januar 1984. Eigenverlag BMV.
Beyer, E. und Eisermann, G.: Nachstellbare Bruckenlager. Erfahrungen beim
Bauvorhaben Dusseldorf-Hauptbahnhof. Beton 5/1983.
Dickerhoff, K.J.: Bemessung von Bruckenlagern unter Gebrauchslast. Dissertation Universitat Karlsruhe 1985.
Petersen, Chr.: Zur Beanspruchung moderner Briickenlager. Festschrift J.
Scheer, Marz 1987.
Hehn, K.-H.: Priifeinrichtung zur Untersuchung von Lagern. VDI-Z 118
(1976), Seite 1 14 bis 118.
N.N., Sanierung der Kolnbreinsperre, Projektierung und Ausfuhrung. 1. Auflage Mai 1991. Herausgeber: Osterreichische Donaukraftwerke AG.
Hakenjos, V. und Richter, K.: Dauergleitreibungsverhalten der Gleitpaarung
PTFE weiss/Austenitischer Stahl fur Lager im Briickenbau. Strasse, Briicke,
Tunnel 1 1 (1979, Seite 294 bis 297.
Imbimbo M. und Kelly J.M.: Influence of Material Stiffening on Stability of
Elastomeric Bearings at Large Displacements. Journal of Engineering Mechanics. Sept. 1998.
Zederbaum, J. (1966): The frame action of a bridge deck supported on elastic
bearings. Civil Engineering and Public Works Review 61(7 14), 67-72.
Leonhardt, F. und Andra, W. (1 960): Stutzprobleme der Hochstrassenbrucken.
Beton- und Stahlbetonbau, 55(6), 121-32.
Tanaka, R., Natsukawa, K. and Ohira, T. (1984): Thermal behaviour of multispan viaduct in frame. In International Association of Bridge and Structural
Engineering, 12th Congress, Vancouver, Canada, 3-7 September.
Building Research Establishment (1979) Estimation of thermal and moisture
movements and stresses; Part 2, Digest 228, Watford.

1.13 References

47

[96] Emerson M. (1977): Temperature differences in bridges: basis of design requirements. TRRL Laboratory Report 765. Transport and Road Research Laboratory, Crowthorne.
[97] Emerson M. (1968): Bridge temperatures and movements in the British Isles.
RRL Report LR 228, pp.38. Road Research Laboratory, Crowthorne.
[98] Emerson M. (1973): The calculation of the distribution of temperature in
bridges. TRRL Report LR 561. Transport and Road Research Laboratory,
Crowthorne.
[99] Emerson M. (1976): Bridge temperatures estimated from the shade temperature. TRRL Report LR 696. Transport and Road Research Laboratory, Crowthorne.
[ 1001 Stephenson, D.A. (1961): Effects of differential temperature on tall slender columns. Concrete and Constructional Engineering, 56(5), 175-8: 56( 1l), 401-3.
[ 1011 Garrett, R.J. (1985): The distribution of temperature in bridges. The Journal of
the Hong Kong Institution of Engineers, May, 35-8.
[ 1021 ComitC Euro-International du BCton (1984). Design manual on structural
effects of time-dependent behaviour of concrete (Bulletin No. 142). George
Publishing Company.
[ 1031 ComitC Euro-International du BCton (1985). Manual of Cracking and Deformations. Bulletin 158E, Lausanne.
[ 1041 Neville, A.M., Dilger, W.H. and Brooks, J.J. (1983): Creep of Plain and Structural Concrete. Construction Press, London and New York.
[ 1051 Mattock A.H. (1961): Precast-prestressed concrete bridge 5.Creep and shrinkage studies. Journal of the Portland Cement Association Research and Development Laboratories, May.
[ I061 Institution of Geological Sciences: National Environmental Research Council
(1 976), Atlas of Seismic Activity 1909-1968. Seismological Bulletin No.5.
[ 1071 Dollar, A.T.J., Abedi, S.M.H., Lilwall, R.C. und Willmore, R.L. (1975): Earthquake risk in the UK. Proceedings of the Institution of Civil Engineers, 58,
123-4.
[ 1081 ICE and SECED (1 985): Earthquake engineering in Britain. Proceedings of
Conference of the Institution of Civil Engineers and the Society of Earthquake
and Civil Engineering Dynamics, University of East Anglia, April.
[ 1091 Lee, D.J. (197 1): The Theory and Practice of Bearings and Expanison Joints for
Bridges, Cement and Concrete Association.
[ 1 101 Buchler, W. (1987): Design of Pot Bearings, American Concrete Institute
Publication, SP-94, V01.2, pp. 882-915.
[ 1 1 11 Black, W. (1971): Notes on bridge bearings, RRL Report LR 382, Transport
and Road Research Laboratory, Crowthorne.
[ I 121 Kauschke, W. and Baignet, M. (1987) Improvements in the Long Term Durability of Bearings in Bridges, American Concrete Institute Publication SP-94,
V01.2,577-612.
[ 1 131 Taylor, M.E. (1970): PTFE in highway bridges. TRRL Report LR 491, Transport and Road Research Laboratory, Crowthorne.
[ 1141 Eggert, H., Kauschke, W.: Lager im Bauwesen, Ernst & Sohn, Berlin 1996.

48

1. Bearings

[ 1 151 Hakenjos, V.: Lager im Bauwesen mit Komponenten aus Kunststoff verdran-

[ 1 161
[ 1 171

[ I 181

[ 1 191

[I201
[ 1211

[ 1221

11231
11241

[ 1251
[ 1261

[ 1271

11281

[ 1291

[ 1301

[ 1311

gen hochbeanspruchbare stahlerne Rollenlager. 13th H.F. Mark-Symposium

on 19- 10-94 in Vienna.
Marioni, A.: Apparecchi di appoggio per ponti e strutture. ITEC, Milano 1983
Campbell, T. I. and Kong, W. L.: TFE Sliding Surfaces In Bridge Bearings. Report ME-87-06, Ontario Ministry of Transportation and Communications,
Downsview, Ontario, 1987.
Crozier, W. F., Stoker, J. R., Martin, V. C. and Nordlin, E. F.: A Laboratory
Evaluation of Full-Size Elastomeric Bridge Bearing Pads. Research Report CA
DOT, TL-6574- 1-74-26, Highway Research Report, June 1979.
Gent, A. N.: Elastic Stability of Rubber Compression Springs. ASME, Journal
of Mech. Engr. Science, Vol. 6, No. 4, 1964.
Jacobsen, F. K. and Taylor R. K.: TFE Expansion Bearings for Highway
Bridges. Report No. RDR-3 1, Illinois DOT, June 1971.
McEwen, E. E. and Spencer, G. D.: Finite Element Analysis and Experimental
Results Concerning Distribution of Stress Under Pot Bearings. Proceedings of
1 st World Congress on Bearings and Sealants, ACI Publication SP-70, Niagara
Falls, 198 1 .
Nordlin, E. F., Boss, J. F. and Trimble, R. R.: Tetrafluorethylene (TFE) as a
Bridge Bearing Material. Research Report, M & R 64642-2, California DOT,
Sacramento, CA, June 1970.
Roark. R. J. and Young, W. C.: Formulas for Stress and Strain. 5th Ed.,
McGraw Hill, New York, 1976.
Roeder, C. W., Stanton, J. F. and Taylor, A. W.: Performance of Elastomeric
Bearings. NCHRP Report 298, TRB, National Research Council, Washington,
D. C., October 1987.
Roeder, C. W. and Stanton, J. F.: State of the Art Elastomeric Bridge Bearing
Design. ACI Journal, 1991.
Roeder, C. W., Stanton, J. F. and Feller, T.: Low Temperature Performance of
Elastomers. ASCE, Journal of Cold Regions, Vol. 4, No. 3, September 1990,
pp 113-132.
Roeder, C. W. and Stanton, J. F.: Failure Modes of Elastomeric Bearings and
lnfluence of Manufacturing Methods. Proceedings of 2nd World Congress on
Bearings and Sealants, ACl Publication SP-94, Vol. 1, San Antonio, Texas,
1986.
Roeder, C. W., Stanton, J. F. and Taylor, A. W.: Fatigue of Steel-Reinforced
Elastomeric Bearings. ASCE, Journal of Structural Division, Vol. 116, No. 2,
February 1990.
Roeder, C. W., and Stanton, J. F.: Elastomeric Bearings: A State of the Art.
ASCE, Journal of the Structural Division, No. 12, Vol. 109, December 1983.
Saxena, A. and McEwen, E. E.: Behaviour of Masonry Bearing Plates in Highway Bridges. Proceedings of 2nd World Congress on Bearings and Sealants,
ACI Publication SP-94, San Antonio, 1986.
Stanton, J. F. and Roeder, C. W.: Elastomeric Bearings Design, Construction,
and Materials. NCHRP Report 248, TRB, National Research Council, Washington, D. C., August 1982.

1.13 References

49

[132] Stanton, J. F., Scroggins, G., Taylor, A. W. and Roeder, C. W.: Stability of
Laminated Elastomeric Bearings. ASCE, Journal of Engineering Mechanics,
Vol. 116, No. 6, June 1990, pp 1351-1371.
[ 1331 Structural Bearing Specification. FHWA Region 3 Structural Committee for
Economical Fabrication, Subcommittee for High Load Multi-Rotational Bearings (HLMRB), October 1991.

51

Expansion Joints

2.1

Introduction

As mentioned in chapter 1.1, movements in old stone and timber bridges were small
and no additional devices were necessary to close the gaps between bridges and abutments due to bridge movements. The first expansion joints were built for steel railway
bridges because their movements were not negligible. With the increase of road traffic and of its speed, closing the gaps became necessary for safety reasons, especially
at the moveable bearings. Initially, cover plates were used for expansion joints. For
longer bridges these cover plates were not sufficient, so that finger joints and sliding
plate joints were used. All these types of expansion joints were not watertight and so
the water ran down to the bearings and to the abutments. The first watertight expansion joints were built using steel rails between rubber tubes to absorb the movements.
This principle led to a lot of different multisealed expansion joints which differed in
the means of supporting the steel rails, in the rubber profiles and in controlling the gap
widths. Another type of watertight expansion joint is the cushion joint, consisting of a
rubber cushion with vulcanised steel plates which transfer the traffic loads. In spite of
continuous amendments of all constructions for expansion joints, these still remain
wearing parts, especially in bridges with high traffic density and high traffic loads.
The following chapters give a short survey of expansion joints for different movements used in the construction of bridges.

2.2

The role of expansion joints

The role of expansion joints is to carry loads and to provide safety to the traffic over the
gap between bridge and abutment or between two bridges in a way that all bridge
displacements can take place with very low resistance or with no resistance at all.
A further requirement is a low noise level especially in an urban environment. The
expansion joints should provide a smooth transition from the bridge to the adjacent
areas. The replacement of an expansion joint is always combined with a traffic interruption - at least of the affected lane. Therefore expansion joints should be robust and
suitable for all loads and local actions under all weather conditions, moisture and deicing agents. The replacement of all wearing parts should be possible in a simple way.

2.3

Calculation of movements of expansion joints

Movements of expansion joints depend on the size of the bridge and the arrangement
of the bearings. Normally the form of construction depends on the horizontal translation orthogonal to the joint. But it is necessary to consider all translations and rotations
to ensure that the displacements will not reach the limits of the joint construction.
To describe the movements of an expansion joint in detail we have to consider three
translations and three rotations (fig. 2.3- 1).

52

2. Expansion Joints
/

Fig.2.3-1:Possible movements
These movements result from temperature, displacements due to external loads, and
creep and shrinkage in concrete and composite bridges. We may obtain the movements (displacements and rotations) from the structural analysis of the system. Movements due to loads depend on the location of the loads. The controlling deformations
can be determined with influence lines (fig. 2.3-2 and fig. 2.3-3). The influence line of
a deflection is the bending line due to a unit load acting in the direction of the considered movement.
1

.-

Fig.2.3-2:Influence line for a translation

I"
Fig.2.3-3:Influence line for a rotation
To obtain the displacement caused by a rotation it is also possible to calculate the
rotations; the displacements can be determined from the known rotations.

2.3.1 Horizontal translation in the direction of the bridge axis u,

A change of the environment temperature, creep under normal force and shrinkage
lead to a uniform extension or shortening of the bridge (fig. 2.3.1 -1).
The thermal expansion coefficients of steel and concrete have approximately the same
/ K ). A uniform change of temperature about the cross
value ( a , = 1,0...1,2.
section causes only a horizontal translation of the joint. This applies to composite
bridges, too.

53

2.3 Calculation of movements of expansion joints

Fig.2.3.1- I : Uniformly extension or shortening

n

Temperature:

UXt.= UT C

l i ATi

i=l

Creep and shrinkage of concrete bridges

N,,

Permanent normal force

(compression > 0)

E,,

Shrinkage coefficient

Shrinkage: u,,., = -EcbW

li

i=l

A possible problem is the change of the location of the fixing point or the unknown location of the fixing point. On arch bridges the superstructure is usually fixed at the
crown of the arch. The fixing point is moved by the deformation of the arch due to the
asymmetrical load.
Buried expansion joints are often used for short bridges (Chapter 2.4). If the fixing
point is situated on longer piers, it acts as a horizontal spring bearing. Due to a movement
in the joint a plastic deformation of the asphalt layer occurs and the construction has a
certain rigidity. A different rigidity of the expansion joints on the right and left abutment and a possible longitudinal deformation can lead to the cracking of the asphalt
layer at one abutment. As the rigidity of this joint is higher than the rigidity of the piers
the new fixing point is situated near the undamaged expansion joint (fig. 2.3.1-2).
Cracking of the asphalt layer of
the buried expansion joint

Fixing point after cracking

Fig.2.3.1-2: Change of the fixing point

54

2. Expansion Joints

In the case of an elastic fixing point there are additional movements at expansion
joints due to acceleration and braking forces.
The actual rigidity of piers can differ from the planned rigidity. Moreover, if the bridge
is fixed on more than one pier, the position of the fixing point can differ from the
planned position.
Creep and shrinkage in composite bridges (acting in the concrete parts of crosssection only) mainly lead to deflections which result in rotations above the y-axis (fig.
2.3.1-4). Creep can be considered using a reduced section area and a reduced moment
of inertia, shrinkage by a substitute tensile force Nshacting on the free shrinking concrete. N\,, is a compression force acting on the composite cross-section.

-1

-I-

E,,

Shrinkage coefficient

A, Area of concrete

E, Reduced modulus of elasticity of

concrete to consider creep
Fig.2.3.1-3: Equivalent shrinking force

Fig.2.3.1-4: Deflection under load

Horizontal movements of expansion joints can also be caused by vertical movements

of the abutments. They are caused by foundation settlements or by replacement of
bearings (fig. 2.3.1-5). Statically indeterminate steel and composite bridges can be
prestressed by intentional lifting and/or lowering at the bearings.
yr+

positive definition:

cp

-+
u x

55

2.3 Calculation of movements of expansion joints

'xd 1
(bn

Tn

F Y I
Fig.2.3.1-5: Displacement of bearings
UXd1

= 44 ' (e" +e,>

Uxd n

= $1

'

e, + @" .e ,

If a fixing point is located on a high pier the additional movements due to pier deformation must be considered in the structural analysis. The movements can result from
acceleration, braking forces, uniform and non-uniform temperature actions.

2.3.2 Horizontal translation in direction of the cross-section u,

A horizontal translation in the crosswise direction results if the angle between the
joint and the moving direction of the bearing is not 90 O (e. g. in skew bridges). The
magnitude of the movement depends on the magnitude of the movement in the direction of the bridge axis and on this angle (fig. 2.3.2-1 and fig. 2.3.2-2).

u, = sincp. ueff
uy = C0Scp~Ueff

Fig.2.3.2-1: Skewed bridge

56

2. Expansion Joints

Fig.2.3.2-2: Skewed bearing conditions

2.3.3 Vertical translation u,

Vertical translations u, can be caused by the replacement of bearings (fig. 2.3.3-3) and
the geometrical conditions on the abutment (fig. 2.3.3-1 and fig. 2.3.3-2).

u, = u x .tan)

Fig.2.3.3-1: Sloping bridge with horizontal bearings

Fig.2.3.3-2: Bridge with short cantilever on the abutmen2

2.3 Calculation of movements of expansion joints

SN+

...............

57

.............

7 Hydraulic jack

Fig.2.3.3-3: Vertical displacement of bearings (due to bearing replacement)

2.3.4 Rotation around the bridge axis cpx
In the case of a replacement of one single bearing at one side a rotation cpx occurs (fig.
2.3.4-1). However, it is possible to avoid this movement by uniform lifting over the
cross-section.

Tr-

........

- ........

Hydraulic jack
Fig.2.3.4-1: Lijting on one side

Rotation around the y-axis cpr

This deformation is caused by vertical loading and non-uniform temperature. The
controlling load positions of the traffic loads can be determined with influence lines.

2.3.5

Fig.2.3.5-1: Rotation due to deflections

2.3.6 Rotation around the z-axis cpz
The deformation cpz is caused by non-uniform temperature action in the horizontal
direction, and by wind loads (fig. 2.3.6-1).

58

2. Expansion Joints

'PZ
....~~..........~....

Fig.2.3.6-I: Non-uniform temperature action

2.4

Construction of expansion joints

2.4.1 General
The construction of expansion joints has to fulfil the following requirements:
- movement capacity
- bearing capacity for static and dynamic loading,
- watertightness to save bearings, substructure and possible linkage of expansion
joints from deterioration,
- low noise emission,
- traffic safety.
To fulfil the last two requirements a limitation of gap widths is essential. Additionally, it is recommended to avoid slopes exceeding about 3 % and vertical steps between
joined surfaces exceeding 8 mm (fig. 2.4.1- 1).

Fig.2.4.I-I: Recommended safety requirements

Expansion joints are exposed to pollution. The sealing should not be damaged by
inclusions of bigger external bodies. If the gap width is reduced due to a movement
of the superstructure the joint must be able to expel grit and silt to the carriageway
surface.

59

2.4 Construction of expansion joints

In particular, all elastomeric components must be readily accessible and easily replaceable.
2.4.2 Small movements (up to 25 mm)
For movements up to 15 mm it is possible to construct a continuous asphaltic carriageway pavement with a supporting element covering the gap of the superstructure.
This kind of joint is also called a buried expansion joint (fig. 2.4.2-1). Up to 10 mm a
flat metal plate is sufficient; for movements above 10 mm an elastomeric pad is necessary to avoid pavement cracks at the edges of the supporting plate. An additional reinforcement of the pavement is advisable to provide a uniform strain distribution. The
thickness of the pavement should be at least 80 mm and should be equal to the thickness of the corresponding parts of the superstructure and the abutment. To fulfil this
requirement the cover of the gap is usually extended into a niche.
The asphaltic pavement does not provide sufficient watertightness. An additional sealing is recommended to protect bearings and substructure from deterioration.
Flexible filler

Fig.2.4.2-I: Buried expansion joint

There are covering elements fulfilling the requirements of support, strain distribution
and watertightness without additional sealing, e.g. the following kind of joint construction (fig. 2.4.2-2 and fig. 2.4.2-3).
Flexible filler

Fig.2.4.2-2: Buried expansion joint sealed by a rubber profile

60

2. Expansion Joints
Flexible filler
Reinforcement

//

Fig.2.4.2-3: Buried expansion joint with continuous sealing and additional rubber
projile
For movements between 15 and 25 mm the asphaltic material above the joint can be
replaced by a specially modified asphaltic material. Constructions of this kind are
called asphaltic plug joints (fig. 2.4.2-4 and fig. 2.4.2-5).The thickness should be at
least 80 mm, while the length should not exceed 700 mm.
Though movements exceeding 25 mm could be managed in laboratory tests the influence of temperature and of deformation velocity is not known adequately. Incorrect
placement of material results in tearing of the adjacent carriageway pavement. Further
problems are yielding of asphaltic material under the wheels of standing vehicles,
brake and acceleration forces combined with high environment temperatures, and the
development of rutting.
Because of their low lifetime (though combined with low relative costs) asphaltic plug
joints are recommended for temporary purposes.
,Flexible matenal

Fig.2.4.2-4: Asphaltic plug joint

61

2.4 Construction of expansion joints

Fig.2.4.2-5: Asphaltic plug joint additional sealed by a rubber profile

2.4.3 Medium movements (over 25 mm, up to 80 mm)

The absorption of medium movements requires an elastic expansion element or an expansion gap across the carriageway surface. For traffic safety, gaps below 5 mm or
over 65 mm are not recommended. Thus, the expansion movement of a simple gap
construction is limited to 60 mm.
Expansion joints for medium movements consist of a sealing element, edge elements,
and fixing elements. The sealing element can be replaced by a cushion element that
absorbs movements caused by shear deformation (fig. 2.4.3-1).

Fixing element

Sealing element
/Edge element ( steel laws

Cushion element

Fig.2.4.3-I: Construction methods of expansionjoints for medium movements

Seals of expansion gaps can be constructed as V-shaped sealing strips (fig. 2.4.3-2)
or hollow sections (fig. 2.4.3-4). Movements are absorbed by the folding of these
elements.
There are special seals for pavements and cyclist areas to decrease the width of the gap
to avoid accidents (fig. 2.4.3-3). Traditional cover-plates are prone to rattling and corrosion and hinder the accessibility of possible seals, but they provide the best comfort
for pedestrians with high heel shoes (fig. 2.4.3-6).

62

2. Expansion Joints

Fig.2.4.3-2: V-shaped sealing

Fig.2.4.3-3: Special sealing for sidewalks

El

Fig.2.4.3-4: Hollow section

Fig.2.4.3-5: Expansion joint with V-shaped sealing

Fig.2.4.3-6: Expansion joint with cover plate

Fig.2.4.3-7: Expansion joint for sidewalks

63

2.4 Construction of expansion joints

The use of seals made from cellular neoprene extrusion has the advantage of a closed
carriageway surface. In addition to the function as sealing, they are able to transfer
traffic loads. Movements up to 80 mm can be accommodated (fig. 2.4.3-8).

Fig.2.4.3-8:Seals made from cellular neoprene extrusion

Elastomeric cushion joints (fig. 2.4.3-9) are made from neoprene reinforced with steel
plates. Thus, traffic loads can be transferred without significant deflections. The
movements are absorbed by increasing and decreasing of the widths of the two gaps
on the upper side. The maximum movement is limited by the gap width.
The rubber cover of the bearing plate can wear away under traffic or can be damaged
(e.g. by snow ploughs) which lowers the skid resistance.

I
I

Fig.2.4.3-9:Elastomeric cushion joint

Especially when using elastomeric cushions and neoprene extrusion seals, the
restraining actions can exceed 20 kN/m which in some cases is not negligible.
2.4.4 Large movements (over 80 mm)
For large movements, sealing elements and rail elements are coupled. Additionally to
the components of a single gap construction, intermediate elements (also called rails),
supporting elements and linkage elements are needed (fig. 2.4.4- 1). Linkage elements
cause equal gap widths saving the seals from overextending. They must be able to
sustain acceleration and braking forces.

64

2. Expansion Joints
Intermediate e l e y t S,,eanilg
Fixing element

Linkage element

element

Edge,element ( steel jam

Support element

Fig.2.4.4-1: Construction method of expansion joints for large movements

The following figure shows the coupling of cushion elements. In this case a special
linkage mechanism is not necessary due to the high deformation resistance of the
single elements which actually act as a spring linkage.

Fig.2.4.4-2: Coupled elastomeric cushion joint

A typical construction is the coupling of V-shaped and hollow section sealing elements. It is called multiple seal expansion joint. These expansion joints can be classified by the kind of supporting and linkage. The folding trellis linkages (fig. 2.4.4-3)
satisfy a l l supporting and linkage purposes.

65

2.4 Construction of expansion joints

-7
I

I
I

- -

Fig.2.4.4-3: Rails supported by folding trellis linkage

An additional linkage is needed if the rails are supported by parallel beams. One possibility is the spring linkage (fig. 2.4.4-4). Springs are made of an elastic material. The
portion of the resisting force resulting from friction depends on the number of rails
and supporting beams whereas the portion of spring force is independent at the number of springs because of the series connection. A disadvantage of this kind of linkage
is that acceleration and braking forces cause non-uniform spring deformations. If the
gaps are opened near to the maximum value the seals can be overextended.
Another possibility of linkage of parallel supporting beams is the use of horizontal
parallel linkages (fig. 2.4.4-5).

66

Fig.2.4.4-4: Rails supported by

beams, spacing controlled by
springs

Fig.2.4.4-6: Rails supported by hinged

arranged beams (Swivel System)

2. Expansion Joints

Fig.2.4.4-5: Rails supported by beams.

Spacing controlled by horizontal parallel
linkages

Fig.2.4.4- 7: Linkage

2.4 Construction of expansion joints

67

If the supporting beams are skew (Swivel System, fig. 2.4.4-6) they control the gap
width by means of the kinematic characteristic of the mechanism (fig. 2.4.4-7). The
number of supporting beams does not depend on the number of rails. The higher the
number of rails the more economical becomes the application of hinged supporting
beams.
As an alternative to the application of multiple seal expansion joints, special nonwatertight constructions like cantilever-toothed joints or rolling leaf joints (also called
roller shutter plate expansion joint) are used. Both the cantilever-toothed joint and the
rolling leaf joint are as a rule not watertight, so that an additional drainage system is
necessary.
The cantilever-toothed joint (fig. 2.4.4-8), also called finger joint, is a very robust construction but with several disadvantages. The deformation capacity in the crosswise
direction is severely limited and vertical deformations of the joint can prejudice traffic safety. To accommodate small vertical deformations without hazard the free finger
ends should be rounded. Finger joints with supported fingers (fig. 2.4.4-9) have
proved to be not as good as with cantilever fingers.
The rolling leaf joint (fig. 2.4.4-10) consists of a tongue plate, a rocker plate, and
sliding plates. The acceptable movement depends on the size and number of sliding
plates.
Rolling leaf joints can exhibit the following disadvantages:
- broken hinges (falling shutter plates cause gaps in the motorway),
- wear of the bearing surface,
- breaking of the restraining spring elements.
Some manufactures have carried out important improvements by:
- stronger hinges,
- use of specially designed bearings for the shutter plates,
- stronger restraining elements with elastomeric springs,
- rubber seals between the plates (it makes the joint watertight to a great extent).

Fig.2.4.4-8:Cantilever-toothedjoint or finger joint

68

2. Expansion Joints

Steel fingers with an

elastomeric encasing
\

Fixing
Fig.2.4.4-9: Finger joint with supportedfingers

Fig.2.4.4-10: Rolling leaf joint

2.4.5 Expansion joints for railway bridges
For the expansion joints for railway bridges it is necessary to consider two elements:
- the rails themselves,
- the bridge.
Nowadays, a continuous track without expansion joints is preferred, due to the comfort of the passengers. Therefore, many modern railway bridges have no expansion devices for the track. Eurocode 1, part 3 (ENV 1991-3), gives rules for the maximum expansion lengths for continuous tracks. The expansion length (i.e. the distance between
the thermal centre and the opposite end of the deck) should not exceed 60 m for steel
structures with a ballast bed and 90 m for concrete and composite structures, again
with a ballast bed. If the expansion length exceeds these values expansion devices
should be used.

69

2.4 Construction of expansion joints

Two different types of railway expansion joints are in use by the railway authorities.
The first type consists of a parallel joint in the rail and works according to fig. 2.4.5-1.

Fig.2.4.5-I: Paralleljoint in the rail

The second type is normally used for high speed railways (fig. 2.4.5-2). It consists of
an ending rail with a slope of 1 : r, with r = 70 to 100, and a tapered rail which is machined in the same slope. This expansion joint has the advantage that there is no gap
between the rails and that the wheel load is carried by a full profile section, but the
disadvantage that the rail gauge will be widened by the expansion, according to the
slope:

AS = A1 .

2
r

E. g., for an expansion of A1 = 200 mm and with r = 70 m the gauge is widened by

As = 6 mm. The expansion joints of the rails should not be located directly over the gap
because of the rotation angle (py of the bridge. It is better to adjust the expansion joint
on the embankment, at a short distance from the bridge.

Fig.2.4.5-2: Feathered joint

70

2. Expansion Joints

For bridges without a ballast bed the gap between the bridge and the abutment normally stays uncovered. For bridges with a ballast bed there are two possibilities:
a) to enclose the ballast within the bridge and on the embankment, bridging the gap
only by the rails. This construction causes problems to the automatic track ballast
tamping machine.
b) to build a continuous ballast bed by means of elastic rubber (neoprene) joints or by
sliding cover plates. In this case there is no interruption of the ballast bed and no
problem for the tamping machine. On the other hand the compactness of the ballast
increases and decreases with the expansion in the area of the expansion joint of the
rails.

A completely different design philosophy is to install no special expansion devices on

bridges with a continuous ballast bed as described above (and, sometimes, with normal bolted rails joints). If high forces caused by temperature-induced expansion or
shortening of the bridge occur in the rail, the track will move in the longitudinal direction and will become settled by itself, especially under the vibrations of the passing
trains.
The forces between the track and the bridge have to be considered especially for the
design of the longitudinally fixed bearing(s) of the bridge. ENV 1991-3 gives the longitudinal action per track FT, = f 8 [kN/m] (LTI-LT*), where LTi are the expansion
lengths from the fixed bearing.

2.5

Materials for expansion joints

2.5.1 Steel parts

Normally, the supporting members such as edge elements, rails and cross beams are
made of mild steel protected by coating or of corrosion-resistant steel. The stirrups of
the fixing are curved reinforcing bars. Stainless steel is used for moveable parts like
the bolts of a folding trellis linkage and sliding plates connected with PTFE. Members
that are difficult to access, e.g. niches for linkage elements, also are made of stainless
steel or corrosion-resistant steel.
Steel parts embedded in concrete outside of the zone of carbonation, corrosion-resistant steel and stainless steel do not need any protection against corrosion. Parts made
of mild steel must be protected. Coatings must have a sufficiently high resistance
against mechanical stress, temperature actions, oils, and de-icing salt. The coating
should be chosen in accordance with the appropriate national standards. However, a
coating consisting of a two-component epoxy priming coat with zinc dust and a twocomponent epoxy final coat with micaceous iron ore is recommended. Steel parts embedded within the zone of carbonation need only a priming coat. A protection against
corrosion by means of an elastomeric sheathing is possible if the elastomeric material
satisfies the requirements of resistance and durability. In the case of protection by galvanising, hot-dip galvanising is the normal case. Spray galvanising is expensive but
also possible.

71

2.5 Materials for expansion joints

2.5.2 Elastomeric parts

Elastomeric parts must be resistant to environmental influences, de-icing salt, alkaline
and acidic water. They are classified in two categories (load transferring and non-load
transferring).
Load transferring elements (e.g. cushion elements or elastic bearings of the rails) are
made from polychloroprene or from natural caoutchouc. The material must be ageresisting, despite the presence of de-icing salt.
Non-load transferring elements (e.g. sealings) are made from polychloroprene or from
ethylene-propylene-caoutchoucwith high resistance to tearing and to crack propagation. The thickness should not be below 4 mm.
The following table gives the recommended characteristics of applied elastomers.
Characteristic

I Resistance to tearing
Tearing strain
Resistance to crack
propagation
Behaviour after a
temperature stress
(1 4 d; 70 "C)
Change of hardness
Change of resistance to
tearing
Change of tearing strain
Resistance against
potassium chloride
(solution: 4 %; 14 d; 23 " C )
Change of volume
Change of hardness
Resistance against hot
asphaltic bitumen
(30 minutes; 220 "C)
Change of resistance to
tearing
Change of tearing strain
Bond with steel

min. 350 %
min. 1 0 N / m m

min. 400 %
min. 15 N / m m

max. +7 Shore A
max. -20 %

max. +5 Shore A
max. -15 %

max. -20 %

max. -20 %

max. +10 %
max. -5 Shore A

max. +10 %
max. -5 Shore A

max. -20 %
max. -20 %

max. -20 %
max. -20 %
Failure within the
elastomeric material

Tah.2.5.2-I : Recommended characteristics qf the elastomeric parts

The springs of spring-linked multiple seal expansion joints are made of polyurethane
with a high resistance to crack propagation. The material is able to withstand high
strains. It can be compressed down to 20 % of the original length. A further advantage
is the good damping characteristics.

72

2. Expansion Joints

Asphaltic plug joints are made of a special modified asphaltic material. This must
have a sufficient flexibility to absorb the movements of the gap, combined with a sufficient load bearing capacity. The exact composition of the material depends on the
producer.
However, the binder material usually consists of bitumens modified with plasticizers
and polymers. The aggregates, usually, belong to the basalt group.

2.6

Analysis and design of expansion joints

2.6.1 Buried expansion joints and asphaltic plug joints

Expansion joints have to satisfy the requirements of ultimate limit state and fatigue
strength design. A buried expansion joint or an asphaltic plug joint must only fulfil the
construction requirements given in chapter 2.4. The most important rules are:
- The thickness of the asphaltic layer should be at least 80 mm.
- The asphaltic layer over the supporting construction must have the same thickness
as over the superstructure and over the abutment.
- The length of asphaltic plug joints shall not exceed 700 mm.
Thin cover plates should be verified by a calculation. The spread of the load can be
considered by an angle of 45 O (fig. 2.6.1-1).

Fig.2.6.1-I: Loud spread under a wheel

2.6.2 Single seal and multiple seal expansion joints
In most cases the ultimate limit state of a single seal and of a multiple seal expansion
joint is analysed correctly, while the fatigue was only considered empirically. However, damage is usually caused by fatigue. Therefore a correct analysis is essential
[ 18; 191. The loading acts for a very short time. The probability that the axles of two
vehicles are at the expansion joint at the same time is relatively small and only one
axle need be considered.
As a rule, standards contain a design load of the following type to analyse single
members of a bridge.

73

2.6 Analysis and design of expansion joints

q * Rv k

stat

LR Contact length wheel - carriageway surface

R Static load
cp Dynamic factor

Fig.2.6.2-I: Design wheel loud

One rail of an expansion joint carries only the portion Fv.k.,tat
of the load, depending on
the rail width b, the gap width s and the contact length LR(fig. 2.6.2-4).
a,[-I

I no
0 Y5

n yo
0 85
0 80
0 75
n 70
0 65
0 60
n 55
n 50

Yn

80

70

50

40

?n a [ D E G ]

Fig.2.6.2-2: Factor a,

Fig.2.6.2-3: Arrangement of the wheel loads

L,

= L , - ~ S ,
i=l

Fig.2.6.2-4: Loadper rail

2. Expansion Joints

Effective contact length

Gap width
Number of gaps within the contact length
Rail width
Portion of wheel load
Factor of the influence of the angle between expansion joint and driving
direction (fig. 2.6.2-2)
If a f 90 O the two wheels of the axle do not cause the maximum loading on the rail at
the same time. This fact can be considered by reducing the influence of both wheel
loads by the factor a,.
Horizontal wheel loads result from rolling friction, acceleration and braking forces,
and from the slope of the bridge. Accelerating and braking of a lorry at the expansion
joint cause maximum loads but this is a comparatively rare case and, thus, is considered only for the ultimate limit state analysis. Horizontal forces due to rolling friction
act at each overrunning and exert an influence on the fatigue of the material.

Ultimate limit state

The ultimate limit state is analysed with the single wheel loads of an axle and considering the dynamic factors given in the relevant standards.

The acceleration and braking force are determined from the vertical loading. Edge
profiles and their fixing are designed for a horizontal force due to the full wheel load.
Intermediate profile:

Fh,k,dyn
= 5 . (P,, . Fv.k,stat

Coefficient of static friction of the standard

cpV, ( P ~ Vertical and horizontal dynamic factor

Contrary to the fatigue analysis, for ULS verifications a horizontally and vertically
fixed continuous girder is a suitable model of the rails. Rails and support beams can
be calculated with the E-P or P-P method because actually no yielding occurs due to
the high applicable design loads.
The ultimate limit state is analysed using the semiprobabilistic safety concept as
follows:

75

2.6 Analysis and design of expansion joints

Fatigue design
Failure due to fatigue is the main reason for the observed damage. Three types of
fatigue fractures have been observed (fig. 2.6.2-5):
1) Failure of the welded joint between rail and support beam
2) Failure of the support beam
3 ) Failure of the rail

Fig.2.6.2-5: Possible cracks due to fatigue

For the fatigue design, the stress range is of interest. At first it is determined by using
the loads given in the standards.
The horizontal forces due to rolling friction, slope of bridge and acceleration or deceleration must be considered. However, they are smaller than the horizontal force due
to acceleration and braking. The factor 5 consists of three parts:

5 = 5s + 5 R -k 6 E

against driving dmcilon

gs
gR
gE

Factor due to slope

Factor due to rolling friction
Factor due to locomotive acceleration/deceleration

again~tdriving direction

Fig.2.6.2-6: Determination of the factors

- 5 E [%I

tS,kRand cE

accelerat,O"

76

2. Expansion Joints

The vertical load acting on an intermediate or edge profile is Fv.k.star

. The horizontal
loads are determined as follows:
Intermediate profile:

Fh,k.stat
= 5. Fv,k,stat

Fig.2.6.2-7:Dynamic loading of a rail

The contact time t, of the wheel depends on the contact length LR,the velocity v and
the width of the profile b.
b+L,
t, =V

T half period). The circular frequency is:

The impact load is sine-shaped (t, = -;
2

The impact causes a damped sinusoidal vibration (fig. 2.6.2-8). For the ultimate limit
state analysis the response in the fundamental mode of the system is of interest. It is
considered by the dynamic value given in the applicable standards. Fatigue of material
is caused by the stress range. Normally, only the first and second amplitude of Fv,k,dyn
exceed the constant amplitude fatigue limit.

77

2.6 Analysis and design of expansion joints

Fig.2.6.2-8: Dynamic loading and

response of system

Fig.2.6.2-9: Dynamic model

The static bending moments in the vertical direction can be determined on the supported continuous beam. It depends on the stiffness of the springs if it has to be taken
into account or if the springs can be assumed to be rigid. In the horizontal direction the
consideration of the elastic fixing is essential (fig. 2.6.2- 10).

Fig.2.6.2-10: Vertical and horizontal static system

78

2. Expansion Joints

It is important to use the dynamic stiffness of the springs because it differs from the
static value. Both the spring stiffness and the damping coefficient are determined by
overrun-tests. The frequency fh and the damping coefficient can be determined from
the recorded time-deformation curve. The spring stiffness Ch,dyn in the model is varied
until the lowest natural frequency according to the experiments is observed.
The logarithmic decrement D of the damping coefficient of a spring-linked expansion
joint amounts to approximately 10 %.
Further possibilities to determine the lowest natural frequency are an analysis by FEM
or approximate methods. The following method leads to satisfactory solutions.
The fundamental vibration mode shape of the vertical direction can be described by
the static bending line of a continuous girder.

A sinusoidal loading causes the following bending deflection curve:

The following formula leads to the stiffness of the spring:

The application of the formulae of the frequency and the rotational frequency leads to
the natural frequency of the vertical system:
2= -c
(Jj
m

With known chdyn and equal span widths the frequency fh of the horizontal direction
can be determined in the same way. But the system is an elastically-supported continuous girder. The following figures show some calculated results.

79

2.6 Analysis and design of expansion joints

fh

fh

450

-8

450

lh=lO@10 m
400

400

350

350

300

300

250

250

1,80 rn
200

2.00 rn

2,20rn

150
100

200
150
100

1 2 3 4 5 6 7 8 9 1 0 7 1

1011

c 1 , . ~ ~ - 1[N/ml
0~

Ch,dyn

4501

1,40 m

Ih= 300.108 m>1m45m

450

rn

4oo

1,80 rn

350

2.00 rn

300

,1.60

400

l o6 /mI

10 m

15 m

Ih=40@10 m

Om
I0 rn
I0 m

2.20 rn
200
150
100

C h.dyn

1011

/mI

o6

C h.dyn 1

1011

[N/m]

Fig.2.6.2- I I : Lowest natural frequencies of an elastically supported continuous girder

m Massofrail [kg/m]

Single span [m]

Ih Moment of inertia [m4]

Ch.dyn

Dynamic stiffness of spring [N/m]

fh Lowest natural frequency [Hz]

The dynamic values cp, and cp2 of the first and second modes of the system are added
to the value Acp. With an assumed logarithmic damping coefficient of 10 %, the following diagrams give directly the impact factors Acp (fig. 2.6.2-12). Either the first or
second figure can be used. They are suitable for the vertical and horizontal direction.

80

2. Expansion Joints

Distance of
frequency

The horizontal axis of the diagram (b) contains the natural frequency of the system.
This version shows the frequency of resonance as the maximum of the graph of the design velocity. The values Acp of the resonance frequency are comparatively high. Natural system frequencies near the resonance must be avoided at least for the vertical
bending. The recommended distance from the resonance frequency is also indicated
in the diagram. With a known design velocity a maximum span of the rails can be determined. Longer spans cause higher values Acp, leading to a higher stress range. Another disadvantage is an increasing number of stress cycles exceeding the cut-off limit, which means that more than two modes of the system must be considered.
With the values Acpv and AT,, the dynamic difference moments can be calculated.
Mv.k.dyn

= 'Vv

. Mv.k.stat

Mh.k.dyn

= '(Ph

'

Mh.k.stat

The stress range is determined as follows:

'Ok.rnan.dyn

k .dyn

AMh.k.dyn

WV
wh
The design load of an axle is higher than the actual load. The nominal stresses should
be reduced by the factor fredto get the actual design loads.
The value of the factor depends on the ratio between design load and loading due to
the real traffic situation. The determination of the actual traffic situation requires
extensive data for the real loads and their frequency (fig. 2.6.2-13). Infrequent high
loads exert an advantageous influence on the fatigue behaviour (overloading effect).
The maximum load for fatigue design must be determined considering the real frequency of the actual traffic loads (e.g. there may be load components occurring only
in one of a thousand cases).
Instead of the nominal stress also the design load could be reduced.

81

2.6 Analysis and design of expansion joints

In

0.2

0.4

0.6

0.8

7,O

A 5 i 1 Amax

Fig.2.6.2-13: Example of a typical loading sequence

The stress ranges up to the chosen limit are used to determine a constant amplitude
stress range that causes the same damage (fig. 2.6.2-15).
log Ao

log 4 U R

+=2 MIO

,
No= 5 Mi0

NL= i W MID

log N

Fig. 2.6.2-14: Fatigue strength curve

D
N=
, 2 MIO

No-

5 Mi0

NL=100 MIO
Ndarn*

log N

Fig. 2.6.2-15: Constant amplitude

stress range

This value when compared with the stress range A q m a dyn

x provides the factor that allows the fatigue analysis with design loads given in the standards to be used. For instance, [20] recommends the factor fred= 0.75 for the conditions of traffic in Germany,
to be applied to the loads of German Standard DIN 1072. A maximum stress determined in this way is exceeded in only one of a thousand cases.
The fatigue design has to fulfil the following equation:

YFt

Partial safety factor of the fatigue loading (yFf= I .O)

?/Mf

Partial safety factor of fatigue strength (yMf= 1.15)

A o ~ 1110
, = a,,,,,.A

~ Constant
. ~ ~amplitude
~ ~ stress
~ ~ range for 100 million cycles

82

2. Expansion Joints

Can be ascertained by the analyses of the real sequence using the Palmgren-Miner summation (aloe = 0.4).
Fatigue strength for 100 million cycles

AOL

The construction members of the expansion joint are three-dimensional and compact.
The fatigue strength A q can be taken from the standard used if it contains a suitable
detail category, otherwise tests become necessary. The following testing arrangements
were recently used with success (fig. 2.6.2-16). The required number of tests is normally indicated by the standards.

Fig.2.6.2-16: Recommended arrangement of the tests

The lifetime of a construction can be calculated as a statistical value. It is only applicable for the evaluation of that type of construction.

Ndarnage

365. DTLV . DAAL . p

Yd
Ndamage
DTLV
DAAL = 4
P

Design life - time in years

The number of cycles exceeding the cut-off limit
The average of daily lorry traffic in one direction
The average number of axles of each lorry
The distribution of the DTLV on several lanes
p = 1.O in case of one lane
p = 0.85 in case of two lanes
p = 0.80 in case of three or more lanes

2.6.3 Elastomeric cushionjoint

The loads for the ultimate limit state analysis and the reduced loads for the fatigue
analysis are determined in the same way as for the seal expansions joints. In the vertical direction the analysed element transfers a portion of the wheel load, depending on
the zone of influence. Horizontal loads are determined from the vertical loads using
the factor 5.
Intermediate profile:
Edge profile:

Fh.k.stat

Fh.k.stat

5.

Rv.k.stat

Fv.k.~tat

2.6 Analysis and design of expansion joints

83

The horizontal loading of edge profiles and their fixings are analysed considering the
complete wheel load. Edge profiles and fixings can be analysed in the same way as for
multiple seal joints. A possible intermediate profile can be treated as a single span
beam (fig. 2.6.3-1).

,213
1

,113

,I13
4

213
4

Fig.2.6.3-1: Calculation of the intermediate profile

The elastomeric parts of elastomeric cushion joints have to withstand stresses and
stress ranges due to traffic loads. Their strength can be ascertained by tests. The following testing arrangement is recommended.

Fig.2.6.3-2: Recommended arrangement of the test

The specimen is of the same character as the planned construction and has a length of
at least 1200 mm. The loads are applied through an elastomeric disk of 50 mm thickness which is situated in the middle of the cushion element. LR and BR are the dimensions of the load area according to the applicable standard. If the width of sample is
smaller than LR, only a reduced load acts on the joint construction. It can be considered by a smaller disk and a force than P. The inclination of P depends on the factor
5. It considers the sliding friction or the roller friction, the slope of the bridge and the
locomotives acceleration and is different for the ultimate limit and fatigue tests.
The applied force P has the following value for the ultimate limit test:

84

2. Expansion Joints

Fvk

5tdf

Wheel load of the standard

For the fatigue test the loads are reduced by the factor fred.
'red

= fie, P

The construction is applicable if experiments prove that the full load P can be supported as a static load, the reduced load Predfor 2 millions of cycles.
2.6.4 Cantilever-toothedjoint and rolling leaf joint
The Bernoulli-Euler theory of bending gives correct results provided that the height to
length ratio of a beam is at least 1/5. Fingers of cantilever-toothed joints are often not
within this range. If this requirement is satisfied the ultimate load can be calculated
easily. Otherwise tests become essential.
The fatigue behaviour must be determined by tests anyway because of the three
dimensional character of the connection cantilever / edge element. The testing
arrangement and the applied loads are the same as for cushion joints (fig. 2.6.4-1).
Maximum stresses are caused when the joint expansion is maximum.

Fig.2.6.41 : Recommended arrangement of the test

The behaviour of a rolling leaf joint should be checked in the same way. In most cases neither the application of the Bernoulli-Euler theory of bending is possible nor do
the standards contain suitable detail categories for the fatigue design. The loads must
be placed in the most disadvantageous position.

2.7

Installation of expansion joints

The design of an expansion joint is performed by determination of the extreme values

of the expected movements and the position of installation. The installation data
depends on the planned construction sequence.
The expansion joint is adjusted by means of an auxiliary construction. For a spring
linkage prestressing is necessary (fig. 2.7-1). It is recommended to instal the expansion joint in the early morning when the temperature is distributed almost uniformly
over the whole bridge.

85

2.7 Installation of expansion joints

Immediately before the installation the actual temperature of the bridge is measured.
If it is not within the considered tolerance the adjustment must be corrected. After that
the expansion joint is flushed and fixed temporarily. In the case of a steel bridge it is
provisionally bolted or tack-welded. The auxiliary construction must be removed immediately. After carrying out the final fixing, the protection against corrosion is completed.
In concrete bridges the expansion joints are provisionally fixed by welding together
reinforcement and anchoring. The concrete pour should be at least of the same
strength as the adjacent material of the superstructure. While pouring the concrete the
joint construction should be protected by a cover.

......
Adjustable auxiliary
construction

;a/

Reinforcement

Reinfdcement

Fig.2.7-1: Possible auxiliary constructionfor the installation

In the case of a steel bridge the date of installing the expansion joints has no influence
on the expected range of movement. In the case of a concrete bridge or a composite
bridge, single unidirectional movements (shortening due to creep and shrinkage) occur. These movements begin with the erecting of the construction and stop within
some weeks / months / years. Creep is caused by compressive stresses, especially due
to prestressing. The movement due to prestessing forces occurs during the prestressing work. The joint construction has to accommodate the movements which occur after the installation. Therefore, the dimension and, by this, the costs of a joint construction can be reduced by a late installation.
The variation of creep and shrinkage is shown in the following figures by means of the
coefficient of creep cp(-,t,) and the shrinkage value E,, . In various standards,
t = 5 years (= 1800 days) to t = 20 years is set equal t o t = m.

86

2. Expansion Joints

500

1000

1500

2000

500

1000

1500

Fig.2.7-3: Variation of creep

2000

Time Id1

Time [dl

Fig.2.7-4: Variation of shrinkage

The maximum increments of shrinkage and creep occur immediately after completion
or after prestressing. For example after 100 days (about 3 months), about 50 % of the
expected creep deformations and 25 % of the shrinkage deformation have taken place.

2.8

Inspection and maintenance

Expansion joints should be checked regularly by means of visual inspection. The frequency depends on the sensitivity of the construction. Before the inspection the joint
is cleaned, and cover-plates may need to be removed. The check should involve the
following items:
- Damage of the anticorrosive protection. This should be repaired before advanced
rust formations appear. The new coating must be compatible with the existing one.
- Visible cracks due to fatigue in the steel members.
- Damages to the seals. The soiled water of the carriageway can lead to the deterioration and corrosion of the bearings, the substructure and possible the linkages.
- Workability of the linkage. If it does not fulfil its function, damage of the seals may
result.
- Obstruction or damage of the drainage system.
The adjacent carriageway pavement should also be checked. A jutting joint construction due to wheelers enhances the impact loading. If it is not possible to repair the
entire pavement, asphalt ramps should be erected to protect the joints.
Service-free expansion joints are often demanded by the manufacturers. Nevertheless,
it is recommended to clean the gaps from grit and silt to protect seals and linkage. The
drainage should also be cleaned regularly.

87

2.9 Replacement of expansion joints

2.9

Replacement of expansion joints

The lifetime of an expansion joint should be the same as the lifetime of the carriageway pavement. A complete replacement becomes necessary if the steel parts exhibit
advanced fatigue damage. On steel bridges only the bolted or welded connections are
removed. A replacement on concrete bridges is more expensive.
More frequent is the replacement of single members, especially of the elastomer components. Seals should be replaceable from the carriageway site. Manufacturers offer
different systems for easy replacement (fig. 2.9-1).
Edge or intermediate beam

Edge or intermediate beam

Grooved
dowel

Clamping
strip

ic

,-/'

Polychloroprene sealing element

Polychloroprene sealing element

Fig.2.9-I: Possiblefixings to the seal

The gap width must be opened to at least 25 mm. In the case of an elastic linkage,
smaller widths are possible because the rails can be displaced. On the other hand the
seals must not be stretched fully. Expansion joints for large movements should be accessible from the underside to change members of the linkage like elastomeric
springs.
In the case of a road with several lanes it is desirable to change the seals of the expansion joint in sections. It is possible to join the seals by vulcanization on site.
If a replacement of the rails becomes necessary they can also be joined on site. However, the joints should be situated in zones with minimal stress range and must be
welded very carefully because of the high fatigue loads.

88

2. Expansion Joints

2.10 References
Books about expansion joints for bridges:
Lee D.J.: Bridge Bearings and Expansion Joints. Second edition by E & FN Spon,
London, Glasgow, New York, Tokyo, Melbourne, Madras 1994.

Papers:
Price, A.R. (1982): The service performance of fifty buried type expansionjoints.
TRRL Report SR 740, Transport and Road Research Laboratory, Crowthorne.
Price, A.R. ( I 983): The performance of nosing type bridge deck expansion joints.
TRRL Report LR 1071, Transport and Road Research Laboratory Crowthorne.
Price, A.R. (1984): The performance in service of bridge expansionjoints. TRRL
Report LR 1 104, Transport and Road Research Laboratory, Crowthorne.
Department of Transport (1989): Expansion joints for use in highway bridge
decks. Departmental Standard BD 33/88.
Department of Transport (1989): Expansion joints for use in highway bridge
decks. Departmental Advice Note BA 26/88.
Koster W. (1969): Expansion Joints in Bridges and Concrete Roads. Maclaren
and Sons.
Busch, G.A. (1986): A review of design practice and performance of fingerjoints.
Paper presented to the 2nd World Congress on Joint Sealing and Bearing
Systems for Concrete Structures, San Antonio, Texas, September.
Watson, S.C. (1972):A review of past performance and some new considerations
in the bridge expansion joint scene. Paper presented to regional meetings of the
AASHO Committee on Bridges and Structures, Spring.
Koster W. (1986): The principle of elasticity for expansion joints. Paper presented to 2nd World Congress on Joint Sealing and Bearing Systems for Concrete
Structures, San Antonio, Texas, September.
[ 101 Lee, D.J. (1971): The Theory and Practice of Bearings and Expansion Joints for
Bridges, Cement and Concrete Association.
[ I I ] Demers, C.E. and Fisher, J.W., Fatigue Cracking of Steel Bridge Structures, Volume I : A Survey of Localized Cracking in Steel Bridges - 1981 to 1988, FHWA
Publication No. FHWA-RD-89- 166, McLean, VA, 1990
[ 121 Standard Specifications For Highway Bridges. 15th edition, American Association of State Highway and Transportation Officials, Washington, D.C., 1992
[ 131 Tschemmernegg, F., The Design of Modular Expansion Joints, Proceedings of
the 3rd World Congress on Joint Sealing and Bearing Systems for Concrete
Structures, Toronto, 1991.
[14] Dexter, R.J., Kaczinski, M.R., and Fisher, J.W.; Fatigue Testing of Modular Expansion Joints for Bridges, Proceeding of the 1995 IABSE Symposium, Volume
7312, San Francisco, CA, 1995.
[ 151 TL/TP-FU 92, Technische Liefer- und Priifvorschriften fur wasserundurchlassige Fahrbahnubergange von Strassen- und Wegbriicken. Bonn: Bundesministerium fur Verkehr, Ausg. 1992

2.10 References

89

- RVS 15.45, Briickenausriistung - Ubergangskonstruktion. Wien:

Forschungsgesellschaft fur das Verkehrs- und Strassenwesen, Arbeitsgruppe
c<Briickenbau>>,
Arbeitsausschuss <cBruckenausriistung>>,
Ausg. Januar 1995.
[ 171 Braun, Chr.: Verkehrslastbeanspruchungvon Ubergangskonstruktionen in Strassenbriicken. Bauingenieur 67 (l992), P. 229-237.
[ 181 Tschemmernegg, F. (a.0.): Ermudungsnachweis von Fahrbahnubergangen nach
ENV-1993-1. Stahlbau (1995), P. 202-210.
[ 191 Pattis, A.: Dynamische Bemessung von wasserdichten FahrbahnubergangenModulsysteme (Dynamic Design of Waterproof Modular Expansion Joints).
Ph.D. dissertation. Department of Civil Engineering and Architecture, University of Innsbruck, Austria (Dec. 1993).
[20] Herleitung eines Lastmodells fur den Betriebsfestigkeitsnachweis von StraBenbrucken. Forschung Strassenbau und Strassenverkehrstechnik Heft 430, 1984.
[2 11 Ramberger, G.: Bearings, expansion joints and hydraulic equipment for bridges,
IABSE, IS. Kongress-Bericht Copenhagen, 1996.
[22] Fisher, J.W., Kaczinski, M.R. and Dexter, R.J.. Field and Laboratory Experience
with Expansion Joints. IABSE, 15. Kongress-Bericht Copenhagen, 1996.
[23] Braun, C.: The Design of Modular Joints for Movements up to 2000 mm.
IABSE, 15. Kongress-Bericht Copenhagen, 1996.
[24] Nielsen, H.B.: The Storebaelt West Bridge. Railway Expansion Joints. IABSE,
15. Kongress-Bericht Copenhagen, 1996.
[25] Crocetti, Roberto: Modular Bridge Expansion Joints - Loads, Dynamic Behaviour and Fatigue Performance. Thesis for the degree of Licentiate of Engineering. Department of Structural Engineering, Division of Steel and Timber Structures. Chalmers University of Technology, 1998.
[26] Barnard, C.P., Cuninghame, J.R.: Practical guide to the use of bridge expansion
joints. Application guide 29, Transport research laboratory, UK 1997.
[ 161 Richtlinie