Professional Documents
Culture Documents
Gunter Ramberger
Structural Bearings
and Expansion Joints I
for Bridges
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
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.
2. Expansion Joints
2.1 Introduction 51
2.2 The role of expansion joints 51
2.3 Calculation of movements of expansion joints 51
2.4 Construction of expansion joints 58
2.5 Materials for expansion joints 70
2.6 Analysis and design of expansion joints 72
2.7 Installation of expansion joints 84
2.8 Inspection and maintenance 86
2.9 Replacement of expansion joints 87
2.10 References 88
7
1 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 of-
ten 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 develop-
ment of ageing-, ozone- and UV-radiation-resistant elastomers and plastics, new ma-
terials 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 compres-
sion 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.
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
+
about one axis compression)
Horizontal Roller bearing; Link
movement in bearing (tension and
one direction compression);
Rotation Constant line rocker
about one axis sliding bearing
-77
Fig. I .4. I - I : Bearings at an abutment
,
I - ~- ~
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 fol-
lowing: 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 re-
strained. More than three vertical supports of the superstructure lead to statically in-
determinate 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 sec-
tions) 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 distri-
bute 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
I .4 The layout of bearings 11
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).
1 ---_---,--a
Fig. 1.4.1 -7: Horizontal arrangement of the bearings (case a)
Case a) has the advantage that only vertical bearing reactions and no permanent hori-
zontal 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 super-
structure. 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
Formerly, the “classical” arrangement of the bearings for a bridge with two main gird-
ers 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 dif-
ferences cause horizontal restraint forces, provided that bearings have no clearance on
the abutments.
The author prefers the statically determinate system with only one lengthwise re-
strained bearing at the abutment concerned because the actual clearance of a bearing
is not determinable in reality (fig. 1.‘4.2-2).
++-
11, %I,
LA-
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 deter-
minate 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 reac-
tion force. For all bearing constructions it is easier to transfer horizontal forces in com-
bination 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 sub-
and superstructure (fig. 1.4.2-4).
! I
I
HI
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).
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 stiff-
ness 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 bear-
ing. 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. 1.4.2-8: Layout for curved continuous girders (no constraint under overall tem-
pe ra ture)
The elongation is
A,, = k . r, << r,
A- = k . r << r
k proportional elongation
The rotation is
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.
I
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.
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.
- 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.
nosing forces
centrifugal forces
braking forces
traction forces
- wind loads
wind on construction
wind on traffic loads
- settlements of abutments and piers
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
- vehicle impact
- derailment
others
The simplest way to obtain these combinations is to calculate the variable actions, es-
pecially 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 rail-
ways, 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 sec-
tions. The following examples demonstrate the difference:
Fig.1.5.2-1: Influence area for the verticul bearing reaction A, box section.
These restraining forces are possible if the friction between bearing and sub- and su-
perstructure 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 ap-
propriate shapes (fig. 1.6.1-2).
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).
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.
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.
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 unrein-
forced 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 struc-
tural steel and lubrication grease. PTFE is a plastic with high mechanical and chemi-
cal 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 incompres-
sible 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
Anchoring plate
Sliding plate
Polished stainless steel
PTFE (Polytetrafluorethylen)
Lid
Pot -wall
Sealing
Elastomere disc
Pot - bottom
Fig.1.6.3-4: Members of a pot bearing
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 displace-
ments 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.
hb PTFE
Part of sphere
Chrome plated
polished surface
Fig.1.6.4-1: Members of a spherical bearing
I
Fig. 1.6.4-2: Fix spherical bearing
I
I
I
I 1
I
I
I
Fig. 1.6.4-4: Free spherical bearing
1.6 Construction of bearings 27
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 al-
low 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.
!
Fig. I . 6.6-I :Fixed bearing
-top plate
bearing
assembly
base plate
Eurocode 3 may be used for the design of all bearing components made from struc-
tural 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 bear-
ings. For design one should use EC 3 , part 1-4. Concerning stainless steel for sliding
plates see 1.7.3.
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 approxi-
mation for E, co,npr is given by
The maximum stresses under compression between two rigid bodies are
F
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
1.7 Materials for bearings 31
the maximum is not in the middle of one half but nearer the outer side; thus we
(3
a
finally obtain: a + < , El -
- -1 E l compr. This is described very well by the
2
following approximate formula:
for b 2 a
a= Mi? a'b
Under the rotation a we obtain a curvature p = with I = _____
12
~
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 dy-
namic coefficient. After movement has taken place the dynamic friction coefficient re-
mains 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 orientat-
ed 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 val-
ues of the friction coefficient pLdbetween PTFE and stainless steel, depending on the
compression force (EN 1337-2).
I
I I I I
I I I I
0.00 I I I I I I -
0.0 0.5 1.o 1.5 2.0 2.5 3.0 p [kNicm']
Fig. 1.7.3-I : Friction coejficient depending on the compressionforce
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 move-
ments but they occur at high velocity. Wearing is mostly caused by the second case.
For the stainless steel plate, austenitic steel X6CrNiMo17122 according to EU-
RONORM 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.
b=
7
1
1
3F(I-v2) .
2E 1
-f-
1
1
= 1,109
73 E -*-
Cylindrical body on cylindrical body
34 1. Bearings
with
+ 1
~-~
1
Fig. I .8. I - I b: Arrangement of the radii
rl r2
F bearing reaction
1 length of the cylinder
r,, r2 radii of the bodies in contact
E Young's modulus Fig. 1.8.1-2: Stress distribution
V Poisson's ratio (v = 0.3 for steel)
max (3 maximum normal stress at the contact area
b half the width of the contact zone
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 2
+ 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
f 2
(cylindrical body on flat surface): pd 5 18. R . with
E d
maxo, =0.418.
R
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
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.
a c h c , a
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.
36 1. Bearings
cw
Fig. 1.8.2-2: Shearforce
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):
1.9 Installation of bearings 37
e >--FSd Y M p d d
' +
e, 2 FSd ' YMp +
= e 2 +-
-7 - 2t ' f y 3 2 t . f, 3 3
b) Depending on the geometry of the pin plate:
Generally, bearings should not be built directly on the construction beneath. To guar-
antee that the area below a bearing is fully sealed a layer of mortar or of a similar prod-
uct 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 clear-
ance 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 bear-
ings 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 im-
portant. 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 pro-
tected 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 dif-
ferent 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 hard-
ens 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.
- -
I
-
\
i
_m_
/
reinforcement against splitting tension
To replace a bearing, the bridge has to be lifted by one or more hydraulic jacks. For hy-
draulic jacks, adequately stiffened areas to transmit the hydraulic jack forces to the
sub- and superstructure are required. Concrete parts must be reinforced against split-
ting 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
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 ap-
plied 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 se-
cured 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 some-
times can occur (fig.l.11-3 and tig.l.11-2).
I !! I
pipe or
t--------
I I
L ----____ c=
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 suffi-
cient 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.
1.13 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 Deut-
sche 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 (Vor-
tragsveroffentlichung).
Beyer, E. und Wintergerst, L.: Neue Briickenlager, neue Pfeilerform. Der Bau-
ingenieur 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
Nordlin, E., Stoker, S. and Trinble, R.: Laboratory and Field Performance of
Elastomeric Bridge Bearing Pads, Highway Research Board (1968).
[351 Pare u. Keiner: Elastomeric Bridge Bearings. Highway Research Board Bull
242, 1960.
[361 Payne u. Scott: Engineering Design with Rubber
[371 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 Entwick-
lung. Plastik-Konstruktion 1971, Heft 5 , Seite 209 bis 227.
Schonhofer: Neugestaltungen auf dem Gebiet des Auflagerbaues und auf ver-
wandten Gebieten. Werner-Verlag, Dusseldorf 1952.
Sedyter: Uber die Wirkungsweise von Bleigelenken. Beton und Eisen 1926,
Seite 29.
[431 Shen, M. K.: Uber die Losung des Balkens mit unverschieblichen Auflagern.
Der Bauingenieur 39 (1964), Seite 100.
[441 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.
[491 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 Bau-
technischen Eisenbahndienst, 1967, Seite 23 1 bis 277, Abschnitt E Brucken-
und Ingenieurhochbau.
[531 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, Wilming-
ton ( 1959).
CNR-UNI 10018-68 (Italian Standards for rubber bearings).
1.13 References 45
Pfohl, H.: Reaktionskraft am Festpunkt von Briicken aus Bremslast und Bewe-
gungswiderstanden der Lager. Bauingenieur 58 (1983), Seite 453 bis 457.
Eggert, H. und Hakenjos, V.: Die Wirkungsweise von Kalottenlagern. Der Bau-
ingenieur 49 (1974), Heft 3 , Seite 93/94.
Lehmann, Dieter: Beitrage zur Berechnung der Elastomerlager. Die Bautech-
nik 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 Auf-
lagerflachen fur Elastomerlager bei Stahlbetonfertigteilen. Fertigteilbau-
forum 13/83, Seite 1 bis 5 (Betonwerk + Fertigteil-Technik).
Kessler, E.: Die Anwendung unbewehrter Elastomerlager. Betonwerk + Fertig-
teil-Technik, Heft 6 (1987), Seite 419 bis 429.
Bundesminister fur Verkehr: Schlden an Brucken und anderen Ingenieurbau-
werken. Dokumentation 1982. Verkehrsblatt-Verlag, Dortmund.
Bundesminister fur Verkehr: Bericht uber Schaden an Bauwerken der Bundes-
verkehrswege. 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. Disser-
tation 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. Auf-
lage 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 Me-
chanics. 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 multi-
span 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
[ 1 151 Hakenjos, V.: Lager im Bauwesen mit Komponenten aus Kunststoff verdran-
gen hochbeanspruchbare stahlerne Rollenlager. 13th H.F. Mark-Symposium
on 19- 10-94 in Vienna.
[ 1 161 Marioni, A.: Apparecchi di appoggio per ponti e strutture. ITEC, Milano 1983
[ 1 171 Campbell, T. I. and Kong, W. L.: TFE Sliding Surfaces In Bridge Bearings. Re-
port ME-87-06, Ontario Ministry of Transportation and Communications,
Downsview, Ontario, 1987.
[ I 181 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.
[ 1 191 Gent, A. N.: Elastic Stability of Rubber Compression Springs. ASME, Journal
of Mech. Engr. Science, Vol. 6, No. 4, 1964.
[I201 Jacobsen, F. K. and Taylor R. K.: TFE Expansion Bearings for Highway
Bridges. Report No. RDR-3 1, Illinois DOT, June 1971.
[ 1211 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 .
[ 1221 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.
11231 Roark. R. J. and Young, W. C.: Formulas for Stress and Strain. 5th Ed.,
McGraw Hill, New York, 1976.
11241 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.
[ 1251 Roeder, C. W. and Stanton, J. F.: State of the Art Elastomeric Bridge Bearing
Design. ACI Journal, 1991.
[ 1261 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.
[ 1271 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.
11281 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.
[ 1291 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.
[ 1301 Saxena, A. and McEwen, E. E.: Behaviour of Masonry Bearing Plates in High-
way Bridges. Proceedings of 2nd World Congress on Bearings and Sealants,
ACI Publication SP-94, San Antonio, 1986.
[ 1311 Stanton, J. F. and Roeder, C. W.: Elastomeric Bearings Design, Construction,
and Materials. NCHRP Report 248, TRB, National Research Council, Wash-
ington, 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 Bear-
ings (HLMRB), October 1991.
51
2 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 abut-
ments 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 traf-
fic 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 expan-
sion 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 move-
ments used in the construction of bridges.
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 move-
ments (displacements and rotations) from the structural analysis of the system. Move-
ments 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 con-
sidered movement.
.-
I"
n
Temperature: UXt.”= UT ’ C l i ’ ATi
i=l
A possible problem is the change of the location of the fixing point or the unknown lo-
cation 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 abut-
ment 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).
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 cross-
section 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 con-
crete. N\,, is a compression force acting on the composite cross-section.
A, Area of concrete
yr+ -+
positive definition: cp u x
2.3 Calculation of movements of expansion joints 55
'xd 1
(bn Tn
e _ r C 1 ~
F Y I
Fig.2.3.1-5: Displacement of bearings
If a fixing point is located on a high pier the additional movements due to pier defor-
mation must be considered in the structural analysis. The movements can result from
acceleration, braking forces, uniform and non-uniform temperature actions.
u, = sincp. ueff
uy = C0Scp~Ueff
u, = u x .tan€)
h
Fig.2.3.3-2: Bridge with short cantilever on the abutmen2
2.3 Calculation of movements of expansion joints 57
SN+ I /
...............
I - .............
7 Hydraulic jack
Fig.2.3.3-3: Vertical displacement of bearings (due to bearing replacement)
Tr-
2.3.4-1). However, it is possible to avoid this movement by uniform lifting over the
cross-section.
........
- ........
Hydraulic jack
Fig.2.3.4-1: Lijting on one side
'PZ
....~~..........~....
To fulfil the last two requirements a limitation of gap widths is essential. Additional-
ly, it is recommended to avoid slopes exceeding about 3 % and vertical steps between
joined surfaces exceeding 8 mm (fig. 2.4.1- 1).
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.
2.4 Construction of expansion joints 59
In particular, all elastomeric components must be readily accessible and easily re-
placeable.
Flexible filler .
There are covering elements fulfilling the requirements of support, strain distribution
and watertightness without additional sealing, e.g. the following kind of joint con-
struction (fig. 2.4.2-2 and fig. 2.4.2-3).
Flexible filler
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 influ-
ence 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
Sealing element
Cushion element
Fixing element /Edge element ( steel laws
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 cor-
rosion 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
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).
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
Especially when using elastomeric cushions and neoprene extrusion seals, the
restraining actions can exceed 20 kN/m which in some cases is not negligible.
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.
A typical construction is the coupling of V-shaped and hollow section sealing ele-
ments. It is called multiple seal expansion joint. These expansion joints can be classi-
fied by the kind of supporting and linkage. The folding trellis linkages (fig. 2.4.4-3)
satisfy a l l supporting and linkage purposes.
2.4 Construction of expansion joints 65
-7
I
1
I
I
Fig.2.4.4-3: Rails supported by folding trellis linkage
- - I
An additional linkage is needed if the rails are supported by parallel beams. One pos-
sibility 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 num-
ber 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 2. Expansion Joints
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 non-
watertight 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 con-
struction but with several disadvantages. The deformation capacity in the crosswise
direction is severely limited and vertical deformations of the joint can prejudice traf-
fic 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),
- rubber seals between the plates (it makes the joint watertight to a great extent).
Fixing’
Fig.2.4.4-9: Finger joint with supportedfingers
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.
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 ma-
chined 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:
2
AS = A1 . ~
For bridges without a ballast bed the gap between the bridge and the abutment nor-
mally 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.
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 lon-
gitudinal action per track FT, = f 8 [kN/m] (LTI-LT*), where LTi are the expansion
lengths from the fixed bearing.
Characteristic
I Resistance to tearing
Tearing strain min. 350 % min. 400 %
Resistance to crack min. 1 0 N / m m min. 15 N / m m
propagation
Behaviour after a
temperature stress
(1 4 d; 70 "C)
Change of hardness max. +7 Shore A max. +5 Shore A
Change of resistance to max. -20 % max. -15 %
tearing
Change of tearing strain max. -20 % max. -20 %
Resistance against
potassium chloride
(solution: 4 %; 14 d; 23 " C )
Change of volume max. +10 % max. +10 %
Change of hardness max. -5 Shore A max. -5 Shore A
Resistance against hot
asphaltic bitumen
(30 minutes; 220 "C)
Change of resistance to
tearing max. -20 % max. -20 %
Change of tearing strain max. -20 % max. -20 %
Bond with steel - Failure within the
elastomeric material
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 suf-
ficient 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.
- The asphaltic layer over the supporting construction must have the same thickness
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).
q * Rv k stat
R Static load
cp Dynamic factor
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 a 50 40 ?n a [ D E G ]
n
L, = L , - ~ S ,
i=l
Gap width
Rail width
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 consi-
dered 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.
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.
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:
2.6 Analysis and design of expansion joints 75
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):
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 de-
celeration must be considered. However, they are smaller than the horizontal force due
to acceleration and braking. The factor 5 consists of three parts:
tS,kRand cE
accelerat,O"
. The horizontal
The vertical load acting on an intermediate or edge profile is Fv.k.star
loads are determined as follows:
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.
2.6 Analysis and design of expansion joints 77
The static bending moments in the vertical direction can be determined on the sup-
ported 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).
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.
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 contin-
uous girder. The following figures show some calculated results.
2.6 Analysis and design of expansion joints 79
fh F fh
450 450
L
-8 4
lh=lO@10 m L
400 400
350 350
300 300
250 250
1,80 rn
200 200
2.00 rn
150 2,20rn 150
100 100
1 2 3 4 5 6 7 8 9 1011 1 2 3 4 5 6 7 8 9 1 0 7 1
c 1 , . ~ ~ - 1[N/ml
0~ Ch,dyn l o6 “/mI
L L
4501 450
1,40 m 10 m
Ih= 300.108 m>1m45m 15 m
400 ,1.60 rn 4oo Ih=40@10 m Om
1,80 rn 350 I0 rn
I0 m
2.00 rn 300
2.20 rn
200
150
100
1 2 3 4 5 6 7 8 9 1011 1 2 3 4 5 6 7 8 9 1011
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 fol-
lowing 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
d
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 de-
sign velocity. The values Acp of the resonance frequency are comparatively high. Nat-
ural 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 de-
termined. Longer spans cause higher values Acp, leading to a higher stress range. An-
other disadvantage is an increasing number of stress cycles exceeding the cut-off lim-
it, 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.
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 fre-
quency 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.
2.6 Analysis and design of expansion joints 81
In
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 4 U R log Ao
4 4 R
, , D
+=2 MIO No= 5 Mi0 NL= i W MID log N N=
, 2 MIO No- 5 Mi0 NL=100 MIO
Ndarn*
log N
, = a,,,,,.A
A o ~ 1110 o ~ Constant
. ~ ~amplitude
~ ~ stress
~ ~ range for 100 million cycles
82 2. Expansion Joints
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 nor-
mally indicated by the standards.
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).
I
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 fol-
lowing testing arrangement is recommended.
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 thick-
ness which is situated in the middle of the cushion element. LR and BR are the dimen-
sions 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 consi-
dered 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
locomotive’s 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
The behaviour of a rolling leaf joint should be checked in the same way. In most cas-
es 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.
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 im-
mediately. After carrying out the final fixing, the protection against corrosion is com-
pleted.
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) oc-
cur. 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 prestress-
ing work. The joint construction has to accommodate the movements which occur af-
ter the installation. Therefore, the dimension and, by this, the costs of a joint con-
struction 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
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.
result.
- Obstruction or damage of the drainage system.
The adjacent carriageway pavement should also be checked. A jutting joint construc-
tion 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.
Grooved
dowel
Clamping
strip ,-/'
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 ac-
cessible 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 expan-
sion 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. How-
ever, 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 present-
ed 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, Vol-
ume 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 Associa-
tion 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 Ex-
pansion 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 wasserundurchlassi-
ge Fahrbahnubergange von Strassen- und Wegbriicken. Bonn: Bundesministe-
rium fur Verkehr, Ausg. 1992
2.10 References 89