Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 189 (2017) 186 – 192

Transportation Geotechnics and Geoecology, TGG 2017, 17-19 May 2017, Saint Petersburg,
Russia

High Speed Trains Geotechnics: What Is a Tolerable Bump?

Jean-Louis Briauda*, Somayeh R. Taftib
a
Distinguished Professor, Texas A&M University, College Station, Texas,USA, 77843-3136; e-mail: briaud@tamu.edu
b
PhD Candidate, Texas A&M University, College Station,Texas,USA, 77843-3136;e-mail:srezaeitafti@tamu.edu

Abstract

The problem of the bump at the transition between the embankment and the bridge is an important concern for railways and
highways. These bumps can lead to a rough riding surface which creates discomfort at high speed and high maintenance costs.
The current study addresses the problem for High Speed Trains (HST). One reason for the development of a bump is the
difference in stiffness between the compacted soil embankment and the bridge typically resting on deep foundations. A 4-D finite
element model, developed in LS-DYNA, was used to simulate the effect of this difference in stiffness at the transition between
the embankment and the bridge. The modulus of the embankment is varied from 5 MPa for a very soft embankment to 120 MPa
for a very stiff embankment. The bridge is considered to be rigid by comparison. This study will show that the dynamic
amplification factor, DAF, defined as the ratio of the maximum dynamic rail/wheel impact force to the static load on the wheel,
changes dramatically with the embankment modulus and train speeds.
© 2017 The Authors. Published by Elsevier Ltd.
© 2017 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
Peer-review under responsibility of the scientific committee of the International conference on Transportation Geotechnics and
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Geoecology.
Peer-review under responsibility of the scientific committee of the International conference on Transportation Geotechnics and Geoecology
Keywords: bump; transition; HST; 4-D finite element model; DAF

Nomenclature

E Elasticity modulus
ν Poisson's Ratio
ρ Unit mass
Keq Equivalent spring stiffness of train suspension system
Ceq Equivalent damper coefficient of train suspension system
K1,K2 Primary spring stiffness of train suspension system
C1,C2 Primary damper coefficient of train suspension system
K3 Secondary spring stiffness of train suspension system
C3 Secondary damper coefficient of train suspension system

1877-7058 © 2017 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the scientific committee of the International conference on Transportation Geotechnics and Geoecology
doi:10.1016/j.proeng.2017.05.030
 Jean-Louis Briaud and Somayeh R. Tafti / Procedia Engineering 189 (2017) 186 – 192 187

Mcar Car body mass

Mbogie Bogie mass
VR Rayleigh wave speed
VS Shear wave speed
DAF Dynamic amplification factor

1. Introduction

The problem of bumps at the transition between the main track and the bridge is a main concern of the railway
industry (Davis et al., 2003 [1]; Davis and Li, 2006 [2]; Li et al., 2003 [3]; Li and Davis, 2005 [4]; Plotkin et al.,
2006 [5]). It is also a concern for highways at transitions between the road on top of the embankment and the bridge
(Wahls, 1990 [6]; Stark et al., 1995 [7]; Briaud et al., 1997 [8]; Long et al., 1998 [9]; Seo et al., 2002 [10]; Dupont
and Allen, 2002 [11]; Seo, 2005 [12]). A major source of track bumps is the transition zone between compacted soil
embankments and bridge abutments resting on deep foundations. This irregularity is due to the difference in stiffness
between the two rolling surfaces (Davis and Plotkin, 2009 [13]) that leads to a dynamic oscillation of the train
wheels and to a cyclic variation of the contact force between the wheels and the rail. This dynamic effect results in
an impact force near the bump or dip and associated deterioration of the track near bridges. This additional force acts
at the train-track interface and may result in the formation of a bump or dip in the track. This dynamic effect
becomes more intense as these irregularities, bump or dip, increase in size (Plotkin et al., 2006 [5]; Banimahd, 2008
[14]; Nicks, 2009 [15]; Davis and Plotkin, 2009 [13]). The dynamic loads caused by the difference in stiffness or by
a bump or by a dip can vary in the range of 1.5 to 3 times the static load (Davis et al., 2003 [1]). In addition, an HST
can intensify the impact loads due to the higher train speeds (Banimahd, 2008 [14], Nicks, 2009 [15]). Because the
soil and embankment modulus is an important parameter in this case (Farritor, 2006 [16]), the current study will
include the effect of changes in this modulus on the track response in the case of no bump or rip. This study also
numerically models the influence of high train speeds beyond the range studied by Banimahd (2008) [14] and Nicks
(2009) [15].

2. The four dimensional finite element model: numerical simulations

A 4-D model of a railroad track on soil subgrade was simulated in LS-DYNA to evaluate the response of the
coupled train-track-soil system at speeds up to 200 m/s (720 km/h) without any bump or dip. The Finite Element
Model (FEM) components included the train and the track/natural soil.

2.1. Track/soil model

The track was the steel rail and was modeled as a solid element with elastic material properties (Table 1). The
steel rail was attached to the model of the railroad ties. The railroad ties were modeled as solid elements with elastic
concrete material properties. They were spaced at 0.7 m from center to center and had dimensions of 0.3 m x 0.2 m
x 2.4 m. The track and subgrade mesh is shown in Fig. 1. To perform the parametric studies, different soil moduli
(Es) were considered for the subgrade (Table 1). To study the effects of the change in track stiffness when passing
from the embankment to the bridge, the assumption is made that the track is placed on soil with a specified soil
modulus from 5 MPa to 120 MPa (Table 1) and that it is placed on a rigid base representing the bridge.
188 Jean-Louis Briaud and Somayeh R. Tafti / Procedia Engineering 189 (2017) 186 – 192

Table 1. Elastic material properties of Track.

Section E (MPa) ν ρ(kg/m3)
Rail 210e3 0.25 7897
Ties 20e3 0.30 1000
Subgrade 5 to 120 0.35 1260
Ballast 120 0.35 1260

(b)
(a)

Fig. 1. (a) Cross section; (b) side view of the finite element model.

Fig. 1 shows a cross section and side view of the model used in the parametric studies. The ballast thickness was
300 mm. The total length of the track and soil model was 180 m long to ensure that one-fourth of a train car can run
until a steady state situation is reached and that the effects of the boundary conditions are avoided. In this research,
only one-half of the full model was simulated because the mesh was symmetrical with respect to the track
centerline. The ballast and the natural soil were assumed to be elastic because the strains are very small. The
boundary conditions imposed on the model (Fig. 2) are roller supports on the sides of the embankment model to
allow for vertical motion and pin supports at the bottom of the embankment model to restrict both horizontal and
vertical movements on that boundary. It should be noted that at centerline, only the horizontal motion in y-direction
is restricted.
(a)

(b)

Fig. 2. Boundary conditions (a) Cross section; (b) side view.

 Jean-Louis Briaud and Somayeh R. Tafti / Procedia Engineering 189 (2017) 186 – 192 189

2.2. Train model

In this study, one-fourth of a train car, meaning one bogie, including the suspension system represented by
springs and dampers, was modeled (Fig. 3). To simplify the train model, an equivalent spring stiffness (Keq) and
damper coefficient (Ceq) for the primary and secondary suspension systems were considered to model the train
suspension system. The HST suspension system specifications are presented in Table 2. A bogie with the same
dimensions as the one Nicks (2009) used in her simulations was used in this train model (Fig. 3). All material and
section properties for the parts of the bogie were assumed to be rigid or solid (Table 3), except for the springs,
dampers and contact surface of the wheels. A “surface-to-surface” contact was applied to define the contact between
the rails and the wheels in LS-DYNA. The outer, elastic elements of the wheel were defined as the slave surface; the
top, outer elements of the rails comprise the master surface. If the slave nodes penetrate the master nodes, a penalty
algorithm will place normal interface springs between the contact surface and the penetrating nodes (Hallquist 2006
[16]). In LS-DYNA, based on Coulomb formulation, values of 0.4 and 0.35 were chosen for the static and dynamic
friction coefficients, respectively (Hallquist 2006 [17]).

Table 2.HST specifications (X2000).

Value
Parameter Value Value
(metric Parameter Parameter
(kN/m) (kN.s/m)
tonnes)
Mcar 55.08 k1,k2 3.28e3 c1,c2 90

Mbogie 3.26 k3 1.31e3 c3 30

Axle Load 17.40 Keq 1.09e3 Ceq 26

190 Jean-Louis Briaud and Somayeh R. Tafti / Procedia Engineering 189 (2017) 186 – 192

Fig. 3.Train dimension (Nicks, 2009 [15])

Table 3.Material properties of different parts of bogie.

Parts Material Property E (MPa) ν ρ (kg/m3)

Wheel Rigid 2e5 0.28 7850
Axle Rigid 2e5 0.28 7850
Side Frame Rigid 2e5 0.28 3816
Bolster Rigid 2e5 0.28 2897
Car Mass Rigid 2e5 0.28 2.2e7
 Jean-Louis Briaud and Somayeh R. Tafti / Procedia Engineering 189 (2017) 186 – 192 191

3. Dynamic response of high speed train in transition zones: parametric study

To evaluate the effect of the modulus of the approach embankment on the response of the HST in the
embankment to bridge transition zone, a no-bump flat track was numerically modeled. The modulus of the approach
embankment was varied from 5 MPa for a very soft soil to 120 MPa for a very stiff one. In this simulation, a train
with an axle load of 174 kN is moving from an embankment onto a bridge at speeds varying from 10 m/s (36 km/h)
to 200 m/s (720 km/h). The Dynamic Amplification Factor is defined as the ratio of the highest vertical impact force
between the train and the track near the transition divided by the static value. The high impact force is created
because the embankment which has a lower modulus settles more than the bridge thus forming a natural bump at the
transition. A consistently high DAF can lead to deterioration of the track near the transition from then embankment
to the bridge. The results show that for the transition zone without a bump, the DAF increases as the train speed
increases up to a critical value and then the DAF decreases for higher speeds (Fig. 4). The critical speed increases
and the DAF decreases as the subgrade modulus increases (Fig. 4). A relationship between this critical speed and the
soil properties seems to exist.

Dynamic Amplification Factor

3.6

3.2 Es=5 MPa

Es=10 MPa
2.8
Es=25 MPa

2.4 Es=40 MPa

Es=70 MPa
2
DAF

Es=100MPa

1.6 Es=120MPa

1.2

0.8

0.4

0
0 50 100 150 200 250
Train Speed, m/s

Fig. 4.Dynamic amplification factor vs. train speed for different subsoil modulus.

4. Conclusion

The problem of the development of a bump due to the difference in stiffness between the compacted soil
embankment and the bridge typically resting on deep foundations for a HST was evaluated by using a 4-D finite
element numerical modeling of the track/soil/train assembly using LS-DYNA. The train travelled at speeds ranging
from 5 m/s to 200 m/s and a critical speed was identified for the dynamic amplification factor (DAF). From the
numerical simulations, the following results have been obtained:
x The maximum DAF decreases as the subsoil modulus increases.
x The maximum DAF occurs at the Critical Speed and can be as high as 3.2 for a very soft soil and as low
as 1.2 for a very stiff soil. The higher values of the DAF can lead to high maintenance costs,
uncomfortable ride, and possibly derailment.
x The Critical Speed depends on the soil modulus and increases with the soil modulus.
192 Jean-Louis Briaud and Somayeh R. Tafti / Procedia Engineering 189 (2017) 186 – 192

Acknowledgements

We would like to acknowledge the University Transportation Center for Railway Safety (UTCRS) and the Texas
A&M Transportation Institute (TTI) for their financial support of this project.

References

[1] D. Davis, D. Otter, D. Li, S. Singh, Bridge approach performance in revenue service. Railway Track and Structures, 99(12), Issue 12, New
York, 2003, pp. 18.
[2] D. Davis, D. Li, Design of track transitions. Research Results Digest 79. Transportation Technology Center, Inc., Pueblo, Colorado, 2006, 4-
15.
[3] D. Li, L. Smith, B. Doe, D. Otter, S. Uppal, (2003). Study of bridge approach and track transition degradation—Factors and mitigation,
Federal Railroad Administration Contract DTFR53-01-H-00305, Washington, D.C., 2003.
[4] D. Li, D. Davis, Transition of railroad bridge approaches. Journal of Geotechnical and Geoenvironmental Engineering,2005, 131(11), 1392-
1398.
[5] D. Plotkin, D.D. Davis, S. Gurule, S.M. Chrismer, Track Transitions and the Effects of Track Stiffness. Proceedings of the AREMA 2006
Annual Conference, Louisville, KY, USA, Sept. 18-2, 2006.
[6] H.E. Wahls, Design and construction of bridge approaches. NCHRP, Synthesis of Highway Practice 159, Transportation Research Board,
D.C., 1990.
[7] T. Stark, S. Olson, J. Long, Differential movement at the embankment/structure interface-mitigation and rehabilitation. Final Report, Report
no.IAB-H1 FY93,Illinois Transportation Research Center, Illinois DOT, Springfield, Illinois, 1995.
[8] J. Briaud, R. W. James, S. B. Hoffman, Settlement of bridge approaches:(the bump at the end of the bridge),NCHRP, Synthesis of Highway
Practice 234, Transportation Research Board, Washington, D.C., (1997).
[9] J. Long, S. Olson, T. Stark, E. Samara, Differential movement at embankment-bridge structure interface in illinois. Transportation Research
Record: Journal of the Transportation Research Board,1998, (1633), 53-60.
[10] J. Seo, H. Ha, J. Briaud, Investigation of Settlement at Bridge Approach Slab Expansion Joint: Numerical Simulations and Model
Tests,Report 0-4142-2, Texas Transportation Institute, College Station, 2002.
[11] B. Dupont, D. L. Allen, Movements and settlements of highway bridge approaches, Research Report, Kentucky Transportation Center, 2002.
[12] J.B. Seo, The Bump at the End of the Bridge: An Investigation, Thesis for the Degree of Doctor of Philosophy, Texas A&M University,
USA, 2005.
[13] D. Davis, D. Plotkin, Track settlement at bridge approaches. Railway Track and Structures,2009, 105(2)
[14] M. Banimahd, Advanced Finite Element Modelling of Coupled Train-Track Systems: A Geotechnical Perspective,Thesis for the Degree of
Doctor of Philosophy, Heriot-Watt University, UK, 2008
[15] J.E. Nicks, The Bump at the End of the Railway Bridge, Thesis for the Degree of Doctor of Philosophy,Texas A&M University, USA, 2009.
[16] S. Farritor, Real-Time Measurement of Track Modulus from a Moving Car,Final Report, U.S. Department of Transportation, FRA,
Washington, DC, January 2006.
[17] J.O. Hallquist, LS-DYNA Theory Manual. Livermore Software Technology Corporation,2006. http://www.lstc.com/pdf/ls-
dyna_theory_manual_2006.pdf.
[18] MANUAL, K. U. LS-DYNA keyword User’s manual, 2007.
[19]Manual, L. K. U.Version 971, Vol.1-2