Sie sind auf Seite 1von 5

Under Sleeper Pads in Turnouts

Modern railway tracks distribute the loads from rail vehicles via the rails, rail
seats and sleepers as evenly as possible. There are limitations when it comes
to turnouts. Points show an irregularity in regard of the vertical elasticity. Mostly
they are too stiff. But their elasticity can be adjusted by under sleeper pads.

1 Challenge Variations in stiffness along the tongue and traditional ballasted track, or for character-
frog area, as well as the constantly chang- izing flat elastic elements such as ballast
Modern railway tracks need to be able to ing load-bearing surface area of the sleep- mats and USP.
bear the loads from rail vehicles via the ers in the ballast superstructure result in
rails, rail seats and sleepers as evenly as discontinuity of the track system. The bedding modulus of different kinds
possible and distribute such to the track of ballasted track ranges from roughly
superstructure and to the subgrade. By dis- The differences in the load-distributing over 0.05 N/mm3 (very soft) to more than
tributing these loads sufficiently, stresses the turnout length is a 3-dimensional prob- 0.40 N/mm3 (very hard) on average, de-
can be kept as low as possible, helping to lem which can be grasped and analyzed pending on the installation conditions [2].
minimize maintenance expenses and thus using Finite Elements Methods (FEM). By For new rail lines, the higher stiffness mainly
increase the operating life of the track sys- installing additional elastic elements with results from the construction of compacted
tem. defined stiffness it is possible to increase substructure and anti-frost layers. Moreo-
the load-distributing effect of the rails. At ver, the use of the Dynamic Track Stabilizer
Although one can draw on tried and trusted the same time the load-bearing character- (DGS) in ballasted track also results in con-
calculation methods, as per Zimmermann istics of the track frame can be optimized. solidation. While measures of this kind in-
[1] for beams on elastic foundation in re- One cost-effective way of approaching this crease the load-bearing ability of the track
spect of the load distribution effect, there target is to use under sleeper pads (USP) bed, at the same time the rails function as
are limitations when it comes to turnouts. with varying degrees of stiffness. a load-distributing element is reduced, with
Due to their geometry (Fig. 1), turnouts negative ramifications for dynamic effects
show an irregularity which can result in In the following a short presentation of the in wheel/rail contact, and this can lead to
fundamental impact of the bedding modu- increased stresses on the ballast.
lus, the possible increase in superstructure
elasticity via USP and a 3-dimensional FEM Higher degrees of bedding stiffness due
turnout model for optimizing the USP ar- to consolidated ballast and subgrade with
rangement is provided. greater load-bearing capacity can be adjust-
ed by installing elastic elements with lower
levels of bedding modulus.

2 Bedding Modulus Regarding the beam on elastic foundation

according to Zimmermann, a reduction
The static and dynamic loads on the track of the bedding modulus from C to C* de-
superstructure stemming from rail traf- creases the ballast pressure by the factor
fic mainly depend on the behaviour of the 4
track bed, as well as the geometry and C * /C
Fig. 1: Turnout with rigid crossing frog and
check rails stiffness characteristics of the track frame. If the effect of a reduced ballast pressure
For traditional ballasted track, the elasticity is quantified with regard to consideration
derives primarily from the flexibility of the of track stability using the 2nd power law,
various degrees of rail deflection within dif- ballast bed and the subgrade. It is gener- a reduction in the ballast pressure of 15%
ferent areas, even if other boundary condi- ally expressed via the bedding modulus C results for example in a lengthening of the
tions remain unchanged. and represents the relationship between intervals for track maintenance by a factor
the surface pressure and the related rail of 1.4. With regard to consideration of track
deflection. stability using the 4th power law, the same
p reduction in the ballast pressure results in
C= [N/mm3] with lengthening the interval by a factor of 1.9,
Harald Loy y i. e. the duration until the next track mainte-
p = pressure between sleeper and ballast nance is almost doubled!
y = rail deflection [mm] The applicability of this assessment of
track stability is backed up adequately by
In simplified terms, the bedding modulus the derived results of the AASHO Road
indicates how much pressure [N/mm] re- Tests [3] and the experience gathered by
System Development sults in a deflection of 1 mm. As surface Deutsche Bahn following introduction of
pressure is already included, the bedding heavy superstructures using UIC 60 rails
Address: Getzner Werkstoffe GmbH
Herrenau 5, A-6706 Buers modulus is used to describe stiffness in [4]. In this regard, reference is made to a
cases where the elasticity primarily results direct hard mounting of the sleepers on
from a surface mounting, for example with the ballast.

RTR 2/2009 35
Under Sleeper Pads in Turnouts

This is also valid for upgrading of tracks, as tive experiences with USP in turnouts have
the costs involved for such work are also led to the development of a standardized
comparatively low. design for the Austrian Federal Railways.

With ballasted track, elastic USP not only

allow for a longer rail bending line, which re- 4 Geometrical Discontinuities at
duces loads on the ballast, they also help to Turnouts
prevent contact abrasion as the top layer of
ballast can become embedded in the pads. Compared to straight track, for which calcu-
Hard contact points between the bottom of lations are easy to conduct due to the rela-
the sleeper and the ballast are alleviated tively homogenous geometry with constant
Fig. 2: Concrete sleepers with Sylomer and the track mounting is more homogene- rail profiles and sleeper mounting surfaces,
under sleeper pads (USP) ous. The pads also help to prevent sudden calculating elastic elements in turnouts is
settling of sleepers due to cavitations [7]. far more complex. The main reasons for
higher effort include the varying profiles of
In regard to the flexibility of the ballasted Due to stabilization of the top ballast layer, the rails, the additional construction ele-
track, the resulting bedding modulus of the migration of ballast rocks due to dynamic ments and the generally strong variations in
standard gauge track should be set in a way forces is shifted to lower layers, which can the sleeper conditions. These parameters
that sufficient rail deflection is ensured, have a benign effect on the long-term qual- result in varying degrees of vertical load
which should not be less than 1.0 1.2 mm ity of the superstructure. deflection.
[6], taking into consideration the load-dis-
tributing effect in the track and the turnouts. Even under the assumption of declining ef- Looking at a standard right-hand turnout,
The limiting factor is generally the maximum fectiveness over the effective lifetime of there is an initial jump in stiffness at the
permissible amount of rail foot tension. the rail, USP still cannot result in any detri- transition point from the stock rail sleep-
mental impact on the track superstructure. ers, which are not supposed to be tamped
In this regard, they can be seen as fail-safe in the middle, to the heavier turnout sleep-
3 Increasing Superstructure elements. The track superstructure will al- ers which are often located in front of the
Elasticity with Under Sleeper ways exhibit more favourable characteris- switch panel (Fig. 3).
Pads (UPS) tics than structure without any USP.
In the switch panel itself, switching ma-
USP are a cost-effective way for subse- There are numerous ways to ensure ade- chines which are integrated into hollow
quently increasing the elasticity of the su- quate adhesion of the pads to the sleep- steel sleepers can result in additional dis-
perstructure and reducing wear and tear ers. One possibility is to glue the pad to the continuities. The switch rails themselves
on the ballast (Fig. 2). USP are elastic ele- cured concrete, but the general trend in the have an increasing moment of inertia from
ments located between the sleeper and the future shows that the pads are integrated the toes to the centre of the turnout. During
ballast and are today available in a continu- directly into the sleepers as a part of the passage of the vehicle, while the stock rail
ous range from approximately 0.02 N/mm3 manufacturing process. A tight connection is loaded with one wheel load, the softer
to harder than 0.30 N/mm3 for standard can be achieved using a plastic mesh, half switch rail results in a lower distribution of
products, Determination of bedding modu- of which is integrated into the sleeper pad load when subject to the load from the op-
lus is given by BN 918 145 01 on a pre- and the other half of which can be vibrated posite wheel load. In the area of the clo-
scribed load plate with ballast profile [5]. into the wet concrete of the sleeper. sure rail, the continuously growing sleeper
areas are also noticeable. The massive frog
The investment cost is far lower than those Tests in Germany have shown that the use together with the wing rails results in the
for the installation of ballast mats. Ballast of sleeper pads allows for significantly im- greatest stiffness. Due to the check rails
mats are mainly used where direct mount- proved track behaviour and dynamic vibra- and the long sleepers, the distribution of
ing to a hard subgrade, e. g. on a bridge, tion behaviour compared to traditional bal- the loads is greatest in this area. Immedi-
should to be avoided. lasted track [8]. In Austria, turnouts with ately after the last long sleeper there are
USP were already installed in 2002 and often eccentrically placed short sleepers
USP does not necessitate the installation measurements have shown a reduction of with shortened sleeper heads on the inte-
of elastic railpads. The track frame con- vibrations in the 40 Hz 50 Hz frequency rior side. This results in a strong load on
sisting of the rails and concrete sleepers, range. Moreover, substantially less subsid- one side, with corresponding torsion of the
which is rigidly connected into a load-bear- ence was found in turnouts with rigid cross- track, depending on the track design.
ing structure via the rail fasteners, can thus ing frogs, even compared to constructions
remained unchanged in its traditional role. with moveable crossing frogs [9]. The posi- The geometrically determined discontinuity

wing rails
switch rails
closure rails crossing nose
switch toes through rails
flange way

switching switch heels

stock rails
check rails

switch panel closure panel crossing panel

Fig. 4: FEM turnout model for the optimization

Fig. 3: Components of a standard right-hand turnout of under sleepers pads

36 RTR 2/2009
Under Sleeper Pads in Turnouts

in the turnout area causes strong localized

variations in the loads on the superstruc-
ture. With a lower bedding modulus, the
bending lines can be lengthened which re-
duces the amount of pressures on the bal-
last. At the same time, by optimizing the
distribution of the bedding modulus with
various USP it is possible to smooth out
the differences in stiffness stemming from
the geometry. Thus, careful design and in-
stallation of various types of sleeper pads
can achieve a double positive effect in the
turnout area: Loads on the superstructure
can be reduced and the turnout can be
smoothed out'.

5 FEM Turnout Model

It is necessary to understand the entire
system in order to be able to transpose the
positive experiences gained with padded
sleepers to applications involving turnouts.
And furthermore to balance out the geo- Fig. 5: Deflection pattern of the rail through the entire turnout by reduced bedding modulus step
metrically determined differences in verti- by step
cal load distribution. The Finite Elements
Method (FEM) can help us to grasp this
system (Fig. 4). the rail fasteners (rail plate/baseplate func- entire structure (subsidence and rises);
tion under pressure loads - fastening clips the strains result from the kinematic con-
Using FEM and high-performance comput- function under tension loads). Moreover, ditions and the parameters of the con-
ers, it is possible to generate a complete the behaviour of the bedding (sleeper pads struction components. A holistic model of
turnout with padded sleepers and to per- and ballast) can be modelled using any de- this nature can be helpful for grasping the
form analyses in respect of its load-bearing sired non-linear deflection curves. system comprising the elastically mounted
functions. In contrast to the manual calcu- turnout in the ballasted track and making
lation method according to Zimmermann, As the individual elastic levels are not su- targeted adjustment for the further devel-
there is no need for idealization by transfor- perposed by the formation of resultant opment of such.
mation of the entire structure (transverse spring stiffness, it is possible to simulate a
sleepers in longitudinal sleeper superstruc- division of the flexibility in levels above and
ture, resultant stiffness for several elas- below the sleepers. As a result, it is also 6 Analysis of Load-Bearing
tic levels). But in order to cut calculation possible to individually observe deforma- Behaviour
times, the individual components are re- tion variables.
duced by base elements to their relevant In order to understand the bedding impact
functionality and the necessary degrees of Parameterization of the model allows for on the vertical rail deflection, the stiffness
freedom. Another advantage of the model is any standard turnout to be generated with of the ballast and the subgrade throughout
that variable parameters can be assigned the defined characteristics and for calcu- the entire turnout area is uniformly and
for all geometries and stiffness in the rail lations to be carried out with a moving stepwise reduced within the framework of
elements (tongue, crossing frog, check load collective, corresponding to the load a simulated calculation. The maximum rail
rails, etc.). The same is valid for sleeper impact of the bogies. As a result, one ob- deflection can therefore be calculated for
parameters and the non-linear behaviour of tains all of the vertical deformation in the each load position, based on the load col-





5FM  'BY 

Under Sleeper Pads in Turnouts

es compared to a turnout without USP. Even

with the relatively stiff pads used in this
case (> 0.2 N/mm3) the forces transmitted
into the superstructure can be reduced on
the order of 10 % to 30 %, depending on the
part of the turnout.

7 Summary
The use of under sleeper pads (USP) can
increase the elasticity of track superstruc-
ture with relatively low investment costs. At
the same time the ballast, which is a latent
source of track instability, is stabilized, as
individual ballast rocks can become embed-
ded in the surface layer of the USP. Loads
on the superstructure are reduced by a
more homogenous mounting of the sleep-
ers and track stability is improved.

Moreover, in turnouts, the geometrically

determined differences in stiffness can be
smoothed out. To achieve this target, USP
Fig. 6: Comparison of rail deflection patterns with optimized track using different USP along the with various degrees of stiffness can be
turnout used, positioned in a way that the entire
construction features improved load-bear-
ing conditions. This allows the track to be
smoother. Track irregularities resulting from
turnouts can thus be mitigated and vibra-
tions can be reduced.

Using a turnout model based on the Finite

Elements Method, it is possible to analyze
the load-bearing behaviour of the construc-
tion. As data can be varied for different pa-
rameters, it is possible to analyze a very
wide range of geometric boundary condi-
tions and stiffness conditions. The chal-
lenge involved is to take into consideration
the non-linear bedding properties.

Using the FEM turnout model, an optimized

arrangement of Sylomer USP can be calcu-
lated for any situation.

Fig. 7: Reduction of rail seat forces along the turnout with optimized solution

lective for a typical bogie (axle load 220 kN, comparison with an improved arrangement [1] Eisenmann, J.: Die Schiene als Trger und Fahr-
bahn. Die Eisenbahnschiene (Herausgeber Fasten-
3 m spacing). The deflection patterns pre- of stiffness using various types of USP. rath). Verlag Wilhelm Ernst & Sohn 1977, S.978.
sented in Fig. 5 represent the envelope for [2] Eisenmann, J. und Rump, R.: Ein Schotteroberbau
the maximum vertical deformations of the The first curve shows as an example the fr hohe Geschwindigkeiten. ETR Eisenbahntech-
nische Rundschau 46, (1997), Heft 3, S.99108.
relevant rail. deflection curve without any additional USP.
[3] AASHO-Road-Test Special Report 61G. Deutsche
The rail deflection of 0.8 mm in the regular bersetzung in Betonstraenjahrbuch 1962/63,
One can see that a lowering of the bedding track results exclusively from the defined Beton-Verlag, Dsseldorf.
modulus results in an increase in rail de- bedding modulus of the ballast and the sub- [4] ORE-Frage D.161, Bericht Nr.4 vom Sept. 1987: Dy-
namische Auswirkungen der Anhebung der Radsat-
flection as desired. But at the same time grade of 0.2 N/mm3. If the vertical deforma- zlast von 20t auf 22,5t und der geschtzte Anstieg
the various differences in the stiffness can tion is increased by the uniform installation der Oberbauelastizitt.
also be seen along the path of the turnout. of additional under sleeper pads, the result- [5] DB Systemtechnik: BN 918145-01: Technische
Lieferbedingungen Spannbetonschwellen mit elas-
The differences in the deformation curves ing bedding modulus of roughly 0.10 N/mm3 tischer Sohle Elastische Schwellensohlen, (Jan.
are magnified. The lower the bedding modu- leads to an increase in deflection to 2004)
lus is, the more pronounced the differenc- 1.3 mm before and after the turnout. The [6] Eisenmann, J., Leykauf, G. und Mattner, L.: Vor-
schlge zur Erhhung der Oberbauelastizitt. ETR-Ei-
es are. In this regard, the amount of deflec- differences in vertical rail deflection can be senbahntechnische Rundschau 43 (1994). H. 7/8,
tion is lowest in the crossing frog due to the smoothed out by the installation of various S.491494
higher level of rail stiffness and the large USP, in the event of identical initial levels. [7] Mller-Boruttau, F., Kleinert, U.: Betonschwellen mit
elastischer Sohle, ETR 50 (2001). H3 S. 90-98.
surface area of the sleepers, whereas im- This allows for a more homogenous pattern
[8] Leykauf, G., Stahl, W.: Untersuchungen und Er-
mediately after the last long sleeper. of deflection to be achieved (using different fahrungen mit besohlten Schwellen. EI Der Eisen-
USP types). The turnout is smoothed in its bahningenieur (55) 6/2004, S. 8-16.
By optimizing the arrangement of USP with function as a load-bearing element. [9] Schilder, R.: USP Under Sleeper Pads
Schwellenbesohlungen Ein Bericht ber bisherige
varying degrees of stiffness, these differ- Erfahrungen, VG Salzburg Congress, Band 65,
ences can be smoothed out. Fig. 6 shows a Fig. 7 shows the reduction of rail seat forc- Sept. 2004

38 RTR 2/2009