University of Birmingham

An Investigation of Subgrade Differential Settlement

on the Dynamic Response of the Vehicle Slab-Track
System
Zhang, Xiaohui; Burrow, Michael; Shunhua , Zhou

DOI:
10.1177/0954409715613538
License:
None: All rights reserved

Document Version
Peer reviewed version

Citation for published version (Harvard):

Zhang, X, Burrow, M & Shunhua , Z 2015, 'An Investigation of Subgrade Differential Settlement on the Dynamic
Response of the Vehicle Slab-Track System', Proceedings of the Institution of Mechanical Engineers, Part F:
Journal of Rail and Rapid Transit. https://doi.org/10.1177/0954409715613538

Link to publication on Research at Birmingham portal

Publisher Rights Statement:

Version of Record available at:
http://dx.doi.org/10.1177/0954409715613538

Checked Jan 2016

General rights
Unless a licence is specified above, all rights (including copyright and moral rights) in this document are retained by the authors and/or the
copyright holders. The express permission of the copyright holder must be obtained for any use of this material other than for purposes
permitted by law.

•Users may freely distribute the URL that is used to identify this publication.
•Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private
study or non-commercial research.
•User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and Patents Act 1988 (?)
•Users may not further distribute the material nor use it for the purposes of commercial gain.

Where a licence is displayed above, please note the terms and conditions of the licence govern your use of this document.

When citing, please reference the published version.

Take down policy
While the University of Birmingham exercises care and attention in making items available there are rare occasions when an item has been
uploaded in error or has been deemed to be commercially or otherwise sensitive.

If you believe that this is the case for this document, please contact [email protected] providing details and we will remove access to
the work immediately and investigate.

Download date: 17. 7. 2020

An Investigation of Subgrade Differential Settlement on the Dynamic

Response of the Vehicle Slab-Track System

Xiaohui Zhang1, Michael Burrow2, and Shunhua Zhou1
1
Key Laboratory of Road and Traffic Engineering of the Ministry of Education, Tongji University, People’s

Republic of China
2
School of Civil Engineering, University of Birmingham, UK

Corresponding author: Xiaohui Zhang

Email: [email protected] [email protected]

Tel.: +86 13651976796 +44 07871621415

Mailing address:

Room 438

School of Transportation Engineering, Tongji University

No. 4800, Cao’an Road

Jiading District

Shanghai

China.

1 / 25
Abstract: The construction of non-ballasted slab railway track on existing subgrade soils, or on

embankments, is at an early stage of development in Chinese railways. Developing appropriate

standards for the allowable amount of subgrade differential settlement which takes into account the

dynamic response of the train-track system is one of a number of issues that need to be addressed. To

inform the development of such standards, a model based on the theory of vehicle-track coupling

dynamics, which considers the self-weight of the track structure, was built to investigate how

differential settlement, in terms of the amplitude, wavelength and position of the settlement along the

track can affect various railway performance related criteria including ride quality, stability, vehicle

safety and potential damage to the train wheel and the rail (i.e. forces at the wheel/rail contact and in

the fasteners). The performance of the model was compared favorably with other widely used

models described in the literature. The analysis of the study to inform design standards using the

developed model demonstrated that the magnitude of the differential settlement influences passenger

comfort the most compared to the other performance criteria. There exists for the CRTS I

track-form considered a particular wavelength (8m for the specific conditions considered) which

results in all measures of performance being at their maximum values. Further, the longitudinal

position of the settlement waveform in relation to the joints between two concrete slabs, a factor

which is not considered in design standards, was shown to influence component deterioration,

passenger comfort and safety. The greatest propensity to cause component damage occurs when

the beginning or end of the differential settlement waveform corresponds with the inter-slab joint of

a concrete base. Accordingly, it is recommended that current design standards should be modified

to specify appropriate combinations of amplitude, wavelength and position of the differential

settlement which give acceptable measures of performance.

2 / 25
Key words: CRTS I slab-track; subgrade differential settlement; vehicle-track coupling dynamics;

structure self-weight; dynamic response

1 INTRODUCTION

China has embarked on an extensive programme of building high speed railway lines and the current

length of its high speed network of approximately 11,000 km (with a similar length planned),

accounts for about half of the world’s total. The majority of these lines have been built using

non-ballasted track-forms on bridges, through tunnels or on pile reinforced embankments. Indeed

slab-track because of its superior stability and low maintenance requirements also is increasingly

becoming the track-form of choice in China for many other lower speed lines. However because of

the expense required to build slab-track on engineered structures, less expensive solutions are being

investigated including constructing the slab-track directly on the existing subgrade soils or on

embankments where necessary. However, a major issue for such slab-track systems, no matter the

line speed, concerns the allowable amount of differential settlement as this directly affects safety,

passenger comfort and damage to the wheel-rail. Differential settlement can occur in slab-track

particularly on soft subgrade soils due to changes in track support condition. Japan has developed

the differential settlement control standards for non-ballasted railway, of 12.5mm for a 20 m length
[1]
of track (i.e. a chord length of 20m) .In China, the Suining-Chongqing railway sets 20mm for a

chord length of 20m as the control standard [2].

A number of authors have undertaken studies using numerical models of the train-track system to

determine the limits of differential settlement (as a function of the amplitude and wavelength of the

settlement) of slab track systems according to a number different limiting performance criterion (e.g.

safety). For example, from the point of view of the concrete base, Zhou [3] and Chen P [4] utilizing a

3 / 25
three-dimensional finite element model (FEM) suggested that for an assumed settlement waveform

of wavelength of 20m, the differential settlement should be limited to less than 15mm, depending on
[5]
the tensile strength of the concrete slab. On this basis, Chen RP considered the effect of

temperature change on the deformation and tensile stress of the track plate and concrete base, and
[3]
suggested a stricter limiting value for subgrade differential settlement. However Zhou , Chen P [4]

and Chen RP [5] did not consider functional aspects of the track such as safety and passenger comfort.
[6]
Liu developed an FEM using beam, shell and spring elements to simulate a train traversing a

foundation subject to differential settlement, but their model is unable to consider track geometry
[7]
irregularities. In terms of passenger comfort, Cai CB for an assumed settlement waveform of

wavelength of 20m, utilizing a conventional vehicle and slab track coupled model and found that

amplitudes not greater than 20mm would keep the vertical acceleration of the vehicle body to within

acceptable limits. By means of a similar method, Cai XP [8] analyzed the dynamic characteristics of
[7]
the train-track system due to subgrade differential settlement. However, both t Cai CB and Cai

XP [8] did not consider the effect of the train speed on the differential settlement limits. Utilizing a

numerical dynamic model of the train slab-track system which assumed that the subgrade differential
[9]
settlement was entirely translated into the rail surface differential settlement, Han related

differential settlement limits to train speed. However, the assumption made by Han regarding

subgrade settlement and rail surface settlement results in track support conditions which differ
[10-11]
significantly from those found in practice. Xu developed a train-track model in which the

vehicle was represented as a numerical multi-rigid vehicle system, and the track support system was

represented as a three-dimensional FEM. Using this model Xu calculated the dynamic response of

different kinds of non-ballasted track systems and suggested standards for the subgrade differential

settlement of waveforms of a length of 20m.

4 / 25
 A limitation of the studies described above is that they focus on limits of differential settlement

solely as a function of the amplitude of the differential settlement and do not consider other

potentially important parameters associated with settlement such as the wavelength and position of

the differential settlement in relation to the slabs which make up many non-ballasted track systems.

In addition, conventional train slab-track coupled dynamic analysis methods assume that the

subgrade stiffness and damping is zero in the area where the differential settlement takes place. In

such cases where the self-weight of the track structure is not considered, the track structure settles

only when the train arrives (see Figure 1(a)). In reality the differential settlement already exists

before the train arrives (see Figure 1(b)). Consequently, modeling the track in this way results in an

incorrect dynamic response of the system and can therefore lead to incorrect estimates of allowable

differential settlement.

(a) [Insert Figure 1(a).]

(b) [Insert Figure 1(b).]

Fig.1 Track structure diagram before the train arrives: (a) model that doesn’t consider self-weight, (b) reality

To address the issues described above apparent in the identified existing studies, a model based on
[12]
the theory of vehicle-track coupling dynamics , which considers the self-weight of the track

structure, was developed to investigate allowable subgrade differential settlement as a function of the

amplitude and wavelength and position of the settlement waveform. To determine the allowable

settlement limiting performance criteria which consider stability, safety and potential damage to the

train wheel and the rail were considered.

5 / 25
2 THE VEHICLE SLAB-TRACK COUPLED MODEL

2.1 Calculation model

The vehicle dynamics model is based on multi-rigid system dynamics theory described in the

literature [12]. The rail is modeled as a simply supported Euler beam with self-weight; the track plate

and the concrete base are modeled as the free-free Euler beam with self-weight; the emulsified

cement asphalt mortar (CA mortar) and the subgrade are regarded as discrete spring-damping

systems. In real environments under cyclic and environmental (temperature) loading, the CA mortar

may tear from the track plate and the concrete base. Therefore to account for this, the stiffness and

damper values in the proposed model were set to zero when subjected to tension. Through

mechanical analysis, the Euler beam oscillation differential equation in the vertical direction

considering the self-weight can be written as follows:

EI
∂ 4 y ( x, t )
∂x 4
+ mr
∂ 2 y ( x, t )
∂t 2
= F ( x, t ) + mr g （1）
where y( x, t ) is the vertical displacement of the Euler beam; mr is the mass of per unit length of the

Euler beam; EI is the flexural rigidity of the Euler beam cross section and F ( x, t ) is the external

force.

The partial differential equations of the vertical vibration of rail, track plate and concrete base can

be obtained by determining the external forces of the structures. Solving the fourth order partial
[13]
differential equations requires the Ritz method , and the basic form of second order ordinary

differential equations of the modal coordinates of rail, track plate and concrete base can be obtained.

These are as follows:

Rail:

6 / 25
 Er I r k π 4
（）
N 4
g
qɺɺk (t ) + ( ) qk (t ) = -∑ Frsi (t )Yk ( xi )+∑ Pj (t )Yk ( xwj ) + 2mrr lr (1 − cos kπ ) k = (1 ~ NM )2
mrr lr i =1 j =1 kπ

Track plate:

（3）
n0 n0
Es I s ls β sk4 Tk (t ) + ms lsTɺɺk (t ) = ∑ Frsi (t ) X k ( xsi ) − ∑ Fcai (t ) X k ( xsi ) + ms gGS (k ) k = (1 ~ NS )
i =1 i =1

Concrete base:

（4）
m0 m0
ɺɺ (t ) = ∑ F (t ) D ( x ) − ∑ F (t ) D ( x ) + m gGB(k ) k = (1 ~ NB)
Eb I b lb β bk4 Bk (t ) + mb lb B k cai k bi fdi k bi b
i =1 i =1

where N , n0 and m0 are respectively the number of fasteners, the number of coordinate nodes of one

track plate, and the number of coordinate nodes of one concrete base; Frsi (t ) , Fcai (t ) and Ffdi (t ) are

respectively the fastener force, the CA mortar reaction force and the subgrade reaction force; Pj (t )

is the wheel-rail contact force of the wheel j ; qk (t ) , Tk (t ) and Bk (t ) are respectively the regular modal

coordinates of the rail, the track plate and the concrete base; Yk , X k and Dk are respectively the

orthogonal function department of simply supported Euler beam of the rail, free-free Euler beam of

the track plate and free-free Euler beam of the concrete base; β sk and β bk are respectively the

constants of the track plate and the concrete base; NM , NS and NB are respectively the modal

orders of the rail, the track plate and the concrete base; GS (k ) and GB(k ) are respectively the

additional functions of self-weight of track plate and concrete base, and their values are:

ls (k = 1)

0 (k = 2)
GS (k ) = 

 ls β sk
（5）
− e − ls β sk sin ls β sk  els β sk + e − ls β sk cos ls β sk 
e + − Ck  −  (k > 2)
 2β sk β sk  2β sk β sk 

7 / 25
 lb (k = 1)

0 (k = 2)

GB(k ) = 
 lb βbk
（6）
− e − lb βbk sin lb βbk  elb βbk + e − lb βbk cos lb β bk 
e + − Ck  −  (k > 2)
 2 βbk β bk  2 βbk βbk 

where Ck is the coefficient of the free beam.

The wheel-rail contact force is solved using the nonlinear elastic contact theory developed by

Hertz [12]. The fastener force and the CA mortar reaction force can be obtained as follows:

 Frsi (t ) = C pi [ Z r ( xi , t ) − Z s ( xi , t )] + K pi [ Z r ( xi , t ) − Z s ( xi , t )]
（7）
ɺ ɺ

 Fcai (t ) = Cai [ Zɺ s ( xi , t ) − Zɺb ( xi , t )] + K ai [ Z s ( xi , t ) − Zb ( xi , t )]

where Zɺ r ( xi , t ) and Z r ( xi , t ) are respectively the vertical velocity and displacement of the rail;

Zɺ s ( xi , t ) and Z s ( xi , t ) are respectively the vertical velocity and displacement of the track plate;

Zɺ b ( xi , t ) and Z b ( xi , t ) are respectively the vertical velocity and displacement of the concrete base;

C pi and Cai are respectively the damping of the fastener system and the damping of the CA mortar;

K pi and K ai are respectively the stiffness of the fastener system and the stiffness of the CA mortar.

In the model suggested herein, the differential settlement is simulated through the subgrade

reaction force:

C Zɺ ( x , t ) + K fi [ Z b ( xi , t ) − z ( xi )] else
Ffdi (t ) =  fi b i
0 when Zb ( xi , t ) ≤ z ( xi )
（8）
where Zɺb ( xi , t ) and Z b ( xi , t ) are respectively the vertical velocity and displacement of the concrete

base; z ( xi ) is the subgrade differential settlement value; C fi is the damping of the subgrade; K fi is

the stiffness of the subgrade.

8 / 25
 Solutions of the vehicle model and the proposed slab track structure dynamics model considering

self-weight all adopt the explicit integration method suggested in the literature [14].

2.2 Model verification

The model suggested herein was verified by a comparative analysis of the outputs calculated by the

model with that: (1) computed by the railway vehicle and slab track vertical coupled software VICT

for the vibration responses produced by a dipped welded joint excitation[12]; (2) computed by an

FEM model for the static responses due to subgrade differential settlement regardless of wheel-rail

contact force; (3) recorded in the literature[11] for the vibration responses due to subgrade differential

settlement under a moving train load.

(1) VICT

Figure 2(a) shows the comparison of the wheel-rail contact force of the proposed model (with

self-weight) and the VICT model (without self-weight) due to a dipped welded joint (5º) for a train

travelling at a speed of 100km/h (i.e. without differential settlement). Note the stiffness of the

fastening system is the dynamic stiffness from field test on the existing high speed railway lines. The

difference in the wheel-rail contact force between the two models illustrates that the wheel-rail

contact force is relatively unaffected by the structure’s self-weight when there is no subgrade

differential settlement. Figure 2(b) shows the comparison of the fastener force at the time when the

wheel-rail contact force reaches a peak value. The lower part of Figure 2(b) shows that the difference

in the resulting forces between the two models is approximately equal to 370 N per fastener, which

is equal to the weight of the rail per fastener. Accordingly, this illustrates that the model proposed is

9 / 25
as accurate as the widely used VICT model, but at the same time can take into account of

self-weight.

(a) [Insert Figure 2(a).]

(b) [Insert Figure 2(b).]

Fig.2 Comparison of the vibration responses of the case with no differential settlement: (a) wheel-rail contact force,

(b) fastener force

(2) FEM

The static responses due to subgrade differential settlement regardless of wheel-rail contact force can

be calculated by assuming the wheel-rail contact force to be zero because the self-weight of the track

structures and the concrete base is taken into consideration in the proposed model. Figure 3(a) shows

the vertical displacement of the rail computed using the proposed model and an FEM suggested in
[3]
the literature due to differential settlement (assumed to be a cosine curve with a wavelength of

30m and an amplitude of 45mm) without train loading. By inspection of Figure 3(a), the vertical

displacements in the two models show good agreement with a maximum difference of less than 5%.

Figure 3(b) compares the fastener force of the two models in the area where the differential

settlement takes place, in which the value that is greater than zero means the compressive force and

the value that is less than zero means the tensile force. By inspection of Figure 3(b) the fastener

forces computed by the two models show good agreement in magnitude and frequency. For example,

the fastener forces in both models increase in magnitude rapidly at the inter-slab joints between

adjacent concrete track plates and concrete base slabs.

(a) [Insert Figure 3(a).]

(b) [Insert Figure 3(b).]

10 / 25
 Fig.3 Comparison of the static responses: (a) rail vertical displacement, (b) fastener force

(3) Existing literature

A comparison of the maximum forces and accelerations experienced by the vehicle body and track

structural components due to subgrade differential settlement under a moving train load was made
[11]
between the proposed model and that calculated by a FEM described in the literature . The

comparison is shown in Table 1 for a train travelling at a speed of 350km/h, with a differential

settlement wavelength of 20m and amplitude of 20mm. In both models track irregularities were

incorporated using the German railway spectra of low irregularity [12].

Table 1 Comparison of dynamic responses

Item Proposed model literature [11] difference

The maximum acceleration of vehicle body (m/s2) 1.323 1.217 8.7%

The maximum wheel-rail contact force (kN) 90.213 85.628 5.3%

The maximum fastener compressive force (kN) 47.934 46.352 3.4%

The maximum fastener tensile force (kN) 15.271 13.265 15.0%

The maximum CA mortar pressure (MPa) 0.138 0.128 7.8%

The differences in the dynamic responses are within 15%. The difference may be partly

explained by the fact that the models are unlikely to have exactly the same track irregularity function

and this would have resulted in different dynamic responses. The possible differences in track

irregularity functions between the two models results from the randomized process associated with

converting a railway track irregularity from the frequency to the time domain for use within the

model.

11 / 25
 It is evident from the above comparisons with the other models available in the literature that the

proposed model suggested herein can calculate with a sufficient degree of accuracy the dynamic

response of the railway vehicle and track structure components due to subgrade differential

settlement. Compared to the VICT, the model takes the effect of track structure self-weight into

consideration, which is more representative of the actual situation when the differential settlement

occurs, and compared to the 3D FEM model, the model is much more efficient computationally but

is still capable of computing the dynamic response of the railway vehicle and the track structures

components to a sufficient degree of accuracy.

3 Subgrade differential settlement

In order to inform construction and maintenance standards and to suggest limits of differential

settlement., a study was undertaken to determine functional performance criteria associated with

train stability, safety and wheel-rail damage as a function of (i) the amplitude of differential

settlement, (ii) the wavelength of differential settlement, and (iii) the position of differential

settlement along the length direction of the track. The track was modeled as a straight section

without any radius of curvature and the differential settlement was represented as a cosine function

with a variable amplitude, wavelength and position along the track.

The measures of functional performance criteria chosen are:

(a). Stability: The maximum vertical acceleration of the train in keeping with a number of railway

organizations was used as a measure of stability. In Chinese design standards, an upper limit of the

vertical acceleration 0.13g is specified [15].

(b). Safety: The likelihood of the derailment of a train is commonly measured by the axle load

decrement ratio (PD) and an upper limit of 0.6 for the PD is specified in Chinese standards [15].

12 / 25
 (c). Wheel and rail damage: The likelihood of excessive wheel or rail deterioration can be

measured by the wheel-rail contact force and the maximum allowable value is suggested to be

250kN according to the British Rail research [16].

3.1 Amplitude of differential settlement

In order to study the influence of differential settlement on the three measures of track performance

four values of the amplitude of the differential settlement, 10mm, 20mm, 30mm and 40mm were

used together with three train speeds of 100km/h, 200km/h and 300km/h. The wavelength of the

settlement profile was maintained at 20m. Note a wavelength of 20m is specified in a number of

design standards [17].

Figure 4 shows the computed vertical acceleration for the four amplitudes chosen. It can be seen

from the figure that the differential settlement causes the vehicle body to accelerate through two and

half cycles. For all amplitudes the wavelengths of the first two complete cycles of acceleration are

nearly constant with the speed of the vehicle, whilst the wavelength of the last half cycle (i.e. △)

increases with speed, i.e. the distance that the vehicle body requires its acceleration to return to its

initial state increases with the driving speed.

[Insert Figure 4.]

Fig.4 Vertical acceleration of vehicle body

Figure 5 shows the maximum vertical acceleration of the vehicle body as a function of the

amplitude of the modeled differential settlement from which it can be seen that the maximum of the

vertical acceleration increases with amplitude as expected, with a corresponding decrease in stability.

The maximum acceleration for amplitudes of 10mm and 20mm for all three train speeds considered

13 / 25
is well below the maximum allowable limit stipulated in the Chinese standards (i.e.0.13g). However,

for differential settlement amplitudes of 30mm and 40mm the stipulated limit is exceeded for speeds

of 200km/h and 300 km/h.

[Insert Figure 5.]

Fig.5 Maximum vertical acceleration of vehicle body

Figure 6 shows the maximum computed PD value as a function of settlement amplitude. As is

shown in Figure 6, the maximum PD value increases with the amplitude of the differential settlement,

although the value in all cases is less than that stipulated in the Chinese standards.

[Insert Figure 6.]

Fig.6 Maximum axle load decrement ratio

Figure 7 shows the wheel-rail contact force for vehicle speeds of 100km/h and 300km/h. From

Figure 7, the variation of the wheel-rail contact force, in the 20m section where the differential

settlement has been modeled, increases with the amplitude of differential settlement and the

fluctuation of the wheel-rail contact force is generally greater at a speed of 300km/h compared to a

speed of 100km/h, although the maximum values in all cases however is much less than the 250kN

suggested by British Rail.

[Insert Figure 7.]

Fig.7 Wheel-rail contact forces

3.2 Wavelength of differential settlement

As the concrete base in general has a higher stiffness than the soil subgrade, it resists to an extent

subgrade settlement, and therefore prevents the rail from settling as much as the subgrade. The

14 / 25
wavelength of the differential settlement can affect this difference to a large extent and, ultimately,

therefore the dynamic response of the system. Therefore to better understand the effect of the

wavelength of the differential settlement, the functional performance of the system was investigated

for 8 different wavelengths (5m, 6m, 7m, 8m, 10m, 15m, 20m and 30m) with an amplitude of 20mm

at train speeds of 200km/h and 300km/h.

Figure 8 shows the maximum vertical acceleration as a function of the wavelength of subgrade

differential settlement for train speeds of 200km/h and 300km /h. For a train speed of 200km/h it can

be seen that the maximum vertical acceleration exceeds that stipulated in Chinese standards for

wavelengths between 8m to 15m. When the train speed is 300km/h, the maximum vertical

acceleration is equal to or exceeds the upper limit of 0.13g for the wavelengths investigated between

8m to 20m.

[Insert Figure 8.]

Fig.8 Maximum vertical acceleration

In terms of the overall trend, the maximum vertical acceleration at the two train speeds considered

increases rapidly with wavelength up to a maximum value of 8m. Thereafter at both speeds the

acceleration decreases. The reason for the increase and then decrease of acceleration with

wavelength of the differential settlement seen in Figure 8 may be explained as follows. When the

wavelength of the differential settlement is small, the deflection of the concrete base is less than the

magnitude of the differential settlement because of the flexure rigidity of the concrete base, resulting

in a gap between the concrete base and the subgrade. As the wavelength of the differential settlement

increases, the deflection of the concrete base increases correspondingly, and with it the vertical

acceleration, until the base contacts the surface of the subgrade. However, the wavelength of the rail

15 / 25
deflection increases with the wavelength of the subgrade differential settlement and results in the

dynamic response reducing correspondingly. The increase in vertical acceleration due to the former

effect predominates at wavelengths of differential settlement up to 8m. At higher wavelengths the

effect on the increase in the wavelength of the rail deflection on reducing the vertical acceleration

predominates.

From the analysis above, at a wavelength of 8m is the wavelength the base contacts the surface of

the subgrade (i.e. the gap between the concrete base and the subgrade is zero). Because the height of

the gap depends on the flexure rigidity of the concrete base cross section, the length of the concrete

base and the track structure design, the particular wavelength (8m for the specific conditions

considered) which results in the acceleration being at its maximum can be considered to depend in

part at least on the design of the track structure and the concrete base.

Figure 9 shows the axle load decrement ratio (PD) as a function of the wavelength of the

differential settlement for train speeds of 200km/h and 300km/h. As is shown in the Figure 9, the

change of the axle load decrement ratio with wavelength is similar to that of the vertical acceleration,

and both reach a peak value at wavelengths of 8m. For a train speed of 200km/h the axle load

decrement ratio exceeds the limit of 0.6 (i.e. the Chinese standards) for wavelengths of between 8m

and 10m. When the vehicle speed is 300km/h, the axle load decrement ratio exceeds the limit of 0.6

for wavelengths of between 7m and 15m.

[Insert Figure 9.]

Fig.9 Axle load decrement ratio

16 / 25
 Figure 10 shows the wheel-rail contact force for a train speed of 300km/h for all wavelengths

considered. It can be seen that the impact of subgrade differential settlement on wheel-rail contact

force has a process of weakness after a first enhancement. The maximum of the wheel-rail contact

force of 221kN occurs when the wavelength is 8m which is less than that of 250kN suggested by

British Rail.

[Insert Figure 10.]

Fig.10 Wheel-rail contact force

3.3 Position of differential settlement waveform

CRTS I slab track consists of discrete concrete sections (see Figure 11) and therefore the

longitudinal position of the differential settlement waveform with respect to these sections may

influence the dynamic response of the system.

In order to study the impact of the longitudinal position of the differential settlement waveform, 8

different positions of the simulated differential settlement waveform were considered within the

model. The first position is such that the start of the settlement waveform corresponds to the

beginning of a concrete base section and the last position of the waveform corresponds to the end of

the section. It should be noted that there is a horizontal gap between any two adjacent concrete bases,

so the first position of the waveform and the last position do not correspond. The other 6 positions of

the waveform were arranged at equidistant intervals as shown in Figure 11. For each case the

maximum amplitude of the settlement was 20mm and the wavelength of the waveform was 20m (to

match the existing Chinese railway standards for differential settlement). Excitation was provided by

17 / 25
a train travelling at a speed of 300km/h and the track was modeled as being smooth to simply the

analysis (i.e. the track irregularity spectrum was not taken into consideration).

[Insert Figure 11.]

Fig.11 Position of subgrade differential settlement

Figure 12 shows the vertical acceleration of the vehicle body and the axle load decrement ratio for

the eight different cases as the train passes the position where the subgrade differential settlement

occurs. For all eight cases it can be seen that the waveform and the magnitude of the acceleration

curves do not change appreciably, and that the computed axle load decrement ratios in all cases are

within a range of between 0.23 and 0.29 (i.e. there is approximately a 20% difference in the

decrement ratio depending on the location of the settlement waveform). Whilst values of the

decrement ratios are less than that stipulated in the Chinese standard (i.e. 0.6) it should be noted that

the Chinese standard includes the effects of track irregularities and therefore any corresponding

dynamic effects.

(a) [Insert Figure 12(a).]

(b) [Insert Figure 12(b).]

Fig.12 Dynamic responses: (a) vertical acceleration of vehicle body, (b) axle load decrement ratio

Figure 13(a) shows the wheel-rail contact force for each of the eight positions considered. As

shown in the Figure 13(a), the minimum and maximum values of the wheel-rail contact force are

relatively unchanged for the eight cases, but the waveforms of the contact force in each case are

significantly different. Further it may be seen that the wheel-rail contact forces change markedly at

the inter-slab joint between two adjacent concrete bases because the rail support condition changes at

these locations. This may be further understood with reference to Figure 13(b) which shows the

18 / 25
forces within the rail fasteners due to differential settlement alone (i.e. without train loading) and

Table 2 which shows the maximum compressive and tensile forces in the fasteners with and without

train loading.

(a) [Insert Figure 13(a).]

(b) [Insert Figure 13(b).]

Fig.13 Wheel-rail contact force and fastener force: (a) wheel-rail contact force (b) fasteners force without train

loading

Table 2 Maximum fastener compressive and tensile forces

Position of Without train loading As the train passes through

settlement
Compressive force /kN Tensile force /kN Compressive force /kN Tensile force /kN
waveform

1 24.75 12.58 47.94 16.58

2 17.41 13.13 39.79 17.41

3 18.95 12.81 39.13 15.8

4 21.65 10.90 42.11 14.29

5 14.48 11.72 41.55 15.76

6 22.11 10.66 40.20 14.94

7 20.19 13.03 40.81 15.82

8 18.72 13.92 42.29 23.56

For all positions of the settlement waveform the maximum compressive force in a fastener is

greater than the tensile force without train loading. The magnitudes of the compressive and tensile

forces experienced by the fasteners are respectively greatest in case 1 and case 8 where the start of

the settlement waveform corresponds to the inter-slab joint as described above. As the train passes

through, the maximum compressive force in case 1 experienced by the fastener is between 13% and

19 / 25
22% greater than that in the other cases and the maximum tensile force in case 8 is 35% to 65%

greater than that in the other cases. Further analysis shows that some of the fasteners within the

area of differential settlement are always in compression with and without train loading (see Figure

14(a)), whilst some fasteners experience both compressive and tensile forces as shown in Figure

14(b). Those that experience both compressive and tensile cyclic loading forces are likely to have

lower service lives than those which are subject to compressive forces alone [18].

(a) [Insert Figure 14(a).]

(b) [Insert Figure 14(b).]

Fig.14 Fastener forces of (a) fasteners subject to compressive forces alone (b) fasteners subject to both compressive

and tensile cyclic forces

4 Concluding discussion

A railway vehicle and slab track vertical coupled dynamics numerical model was described in this

paper and was shown to give similar results to three different existing railway vehicle/ track dynamic

models in terms of the forces and accelerations experienced by the vehicle body and track structural

components. The proposed model has an advantage over these and other existing widely used

dynamic models of the train track system in that it is more computationally efficient and therefore

can be used in environments which do not have access to the computing facilities which are required

to run similar FEMs (i.e. outside of research establishments).

Nevertheless, it is recognized that further refinement of the developed model is both desirable and

necessary. The construction of non-ballasted slab railway track on existing subgrade soils, or on

20 / 25
embankments, is at an early stage of development in Chinese railways, so it was not possible to

compare the outputs of the model directly with field measurements. However, when such data

becomes available verification of the model via the comparison of predicted settlements with those

measured in the field should help to refine the model and the parameters adopted. Additional work

is also required to refine the way in which the rail-sleeper fastening system has been modeled, since

although the fastener force model has been developed according to the literature, whereby the

fastener force is the same in tension and compression, the fastener force in reality exhibits some

nonlinearity especially with loading frequency and temperature change.

Subgrade dynamics, which may impact the dynamic response of the slab track system, have not been

accounted for in the proposed model since the emphases of the paper is on the ride quality, stability,

vehicle safety and potential damage to the train wheel and the rail (rather than the subgrade). For

this purpose, the literature suggests that a model, such as that proposed, can still give sufficient

accuracy for the task at hand, albeit not including subgrade dynamics. However, further

investigation is recommended to ascertain both how the dynamic response due to differential

settlement can be incorporated and how the impact of including subgrade dynamics may affect the

accuracy of the model.

The developed model was used to carry out a number of studies to inform railway design

standards with respect to allowable subgrade differential settlement under CRTS I type slab track

systems. These studies investigated the influence of the settlement waveform, in terms of its

amplitude, wavelength and position on measures of track performance associated with passenger

21 / 25
ride quality, railway vehicle safety and track component damage. The following findings may be

drawn from the analysis.

(1) The dynamic response of the train-track system increases with the amplitude of the

differential settlement and the train speed. Whilst the amplitude of the differential settlement

affects all of the criteria investigated, its greatest influence is on passenger comfort (vertical

acceleration of the vehicle body). When the differential settlement amplitude is greater than 20mm

for a wavelength of 20m the limit stipulated in Chinese design standards is exceeded for speeds of

200km/h and 300 km/h.

(2) The dynamic response of the system, in terms of stability, safety and damage to the wheel-rail,

was shown to be a function of the wavelength of the subgrade settlement and that there exists for the

slab-form considered a particular wavelength at which the measures of response are at their

maximum values. Therefore, for any particular slab-track form, the combinations of wavelength

and settlement amplitude which cause the stability, safety and damage criteria to be at their

maximum values should be used as design criteria, rather than the current situation where standards

suggest a maximum allowable settlement amplitude for one wavelength only (e.g. 20 mm for a 20 m

wavelength).

(3) CRTS I slab track consists of a number of discrete lengths of concrete slab and it was shown

that the position of the settlement waveform in relation to the joints between two slabs influences

greatly the wheel-rail and the fastener forces. The position can also effect the magnitude of the

measures of safety (PD) and passenger comfort by up to 20%. The greatest potential to cause

fatigue damage (and therefore possible early failure of track components) occurs when the beginning,

or end, of the differential settlement waveform corresponds with the inter-slab joint of the concrete

base. Since such areas may be subject to water ingress and thereby softening of the underlying

22 / 25
subgrade promoting settlement, it is suggested that particular attention should be given to these areas

in terms of monitoring condition and associated maintenance.

It may therefore be seen that when developing design standards for slab-track it is necessary to

stipulate maximum values of the allowable settlement, in terms of its amplitude, wavelength and

position (in relation to jointed slab-track) and that the allowable values should also be a function of

train speed.

Acknowledgements

The work described in this paper was supported by the National Natural Science Foundation of

China (51478353) and the Research Fund for the Doctoral Program of Higher Education of China

(20110072110026). Both organizations are thanked with gratitude.

The authors are also most grateful for the support of Tongji University and the University of

Birmingham is also thanked for facilitating this research through an exchange programme.

References:

[1] Zhou M. The Influence of Slab Track and Train Operation Safety on Subgrade's Uneven Settlement. Master

Thesis, Tongji University, 2011.

[2] Ministry of Railways of the People’s Republic of China. The design and technical condition of the

non-ballasted track of the integrated test section of Suining-Chongqing railway. 2004.

23 / 25
[3] Zhou M, Gong QM, Wang BL, and Zhou SH. Effect of differential settlement of subgrade on dynamic

response of ballastless track. Railway Standard Design, 2010, 10, 1-4.

[4] Chen P, Gao L and Ma MN. Limited value of subgrade settlement and its influence on mechanical

characteristics of ballastless track in high-speed railway. Construction & Design for Project, 2008, 05, 63-66.

[5] Chen RP, Wang ZZ, Jiang HG and Bian XC. Control standard of differential settlement in high-speed

railway slab based on bending fatigue strength. Journal of Zhejiang University (Engineering Science), 2013, 05,

796-802.

[6] Liu RB. Analysis of Differential Settlement of Subgrade Effecting on Slab Track Dynamics. Master Thesis,

Southwest Jiaotong University, 2008.

[7] Cai CB, Zhai WM, and Wang KY. Calculation and Assessment Analysis of the Dynamic Performance for

Slab Track on Sui-Yu Railway. China Railway Science, 2006, 27, 17-21.

[8] Cai XP. Dynamic characteristics of train-ballastless track under the conditions of differential settlement of

subgrade. Advances in Environmental Vibration - Proceedings of the 5th International Symposium on

Environmental Vibration, ISEV 2011, 2011, 647-652.

[9] Han YT and Yao L. Analysis of the dynamic performance for slab track settlement on embankment. Journal

of Railway Engineering Society, 2006, 27, 17-21.

[10] Xu QY, Li B, and Zhou ZH. Study on the limited value for the differential settlement of subgrade under

CRTS-I type slab track. China Railway Science, 2012, 33, 1-6.

[11] Xu QY, Li B, and Fan H. Influence of differential settlement of subgrade on dynamic characteristic of

train-ballastless track on subgrade coupling system. Journal of Railway Science and Engineering, 2012, 9,

13-19.

[12] Zhai WM. Vehicle-Track Coupling Dynamics (Third Edition), 2007 (The Science Press, China)

[13] Li DX. Advanced Structure Dynamics (Second Edition), 2010 (The Science Press)

[14] Zhai WM. Two simple fast integration methods for large-scale dynamic problems in engineering.

International Journal for Numerical Methods in Engineering, 1996, 39, 4199-4214.

[15] GB5599-85:1985. Evaluation standard of the dynamic characteristic of railway vehicle.

24 / 25
[16] Clark RA, Dean PA, Elkins JA and Newtons SG. An investigation into the dynamic effects of railway

vehicle running on corrugated rails. J Mech Engng Sci, 1982, 24, 65-76.

[17] Ministry of Railways of People’s Republic of China. Design specification of the integrated test section of

Suining-Chongqing non-ballasted railway, 2004.

[18] Shi J, Burrow M P N, Chan A H, et al. Measurements and simulation of the dynamic responses of a

bridge–embankment transition zone below a heavy haul railway line [J]. Proceedings of the Institution of

Mechanical Engineers, Part F: Journal of Rail and Rapid Transit, 2013, 227, 254-268.

25 / 25