applied

sciences
Article
Effects of Differential Subgrade Settlement on Slab Track
Deformation Based on a DEM-FDM Coupled Approach
Xuhao Cui 1 , Rui Zhou 2, * , Gaoran Guo 3 , Bowen Du 4 and Hanlin Liu 2

1 School of Civil Engineering, Beijing Jiaotong University, Beijing 100044, China; xhcui01@bjtu.edu.cn
2 Institute of Urban Smart Transportation & Safety Maintenance, Shenzhen University, Shenzhen 518060, China;
liuhanlin@szu.edu.cn
3 School of Civil Engineering, Wuhan University, Wuhan 430072, China; 2017102100030@whu.edu.cn
4 School of Computer Science and Engineering, Beihang University, Beijing 100083, China;
dubowen@buaa.edu.cn
* Correspondence: zhourui@szu.edu.cn

Abstract: Slab track structures become deformed under the effects of differential subgrade settlement.
According to the properties of the China Railway Track System (CRTS) II slab track on a subgrade, a
three-dimensional (3D) coupled model based on both the discrete element method (DEM) and finite
difference method (FDM) was developed. The slab track and subgrade were simulated using the
FDM and DEM, respectively. The coupled model was verified. The deformation of the slab track and
contact forces of gravel grains in the surface layer of the subgrade were studied under differential
subgrade settlement. The effects of settlement wavelength, settlement amplitude, and other types
of settlements were also discussed. The results demonstrate that the settlement amplitude and
settlement wavelength of the subgrade have significant effects on track deformation. The deformation
amplitude of the slab track increases nonlinearly with an increasing settlement amplitude of the





 subgrade. Increases in the settlement wavelength and amplitude of the subgrade significantly
Citation: Cui, X.; Zhou, R.; Guo, G.;
increase the maximum value of the contact force of the gravel grains in the subgrade. The maximum
Du, B.; Liu, H. Effects of Differential contact force of gravel grains near the boundaries of the settlement section can reach two to three
Subgrade Settlement on Slab Track times that of the unsettled condition, which makes it easy to accelerate the plastic settlement of
Deformation Based on a DEM-FDM the subgrade.
Coupled Approach. Appl. Sci. 2021,
11, 1384. https://doi.org/10.3390/ Keywords: high-speed railway; slab track; differential subgrade settlement; deformation; DEM-FDM
app11041384 coupled model

Academic Editor: Seung-Yup Jang

Received: 30 December 2020
Accepted: 27 January 2021
1. Introduction
Published: 3 February 2021
Recently, high-speed railways have developed rapidly and the China Railway Track
System (CRTS) II slab track has been widely used [1–3]. With increasing train speed, re-
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
quirements for the smoothness and stability of railways have also increased [4]. However,
published maps and institutional affil-
differential subgrade settlement commonly occurs under the long-term effects of train
iations. load [5]. The differential subgrade settlement has become a key problem in the maintenance
of the CRTS II slab track. Differential subgrade settlement causes track irregularities, inten-
sifies wheel-rail interactions, and changes the mechanical properties of track structures [6,7].
Therefore, estimating the effects of differential subgrade settlement on track deformation is
a crucial step for conducting an in-depth analysis of the relationship between traffic service
Copyright: © 2021 by the authors.
statuses and subgrade settlement [8,9]. A site map and cross-section of the CRTS II slab
Licensee MDPI, Basel, Switzerland.
This article is an open access article
track are presented in Figure 1. The CRTS II slab track system mainly consists of the rail,
distributed under the terms and
fastener, precast slab, cement emulsified asphalt (CA) mortar layer, and concrete base.
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).

Appl. Sci. 2021, 11, 1384. https://doi.org/10.3390/app11041384 https://www.mdpi.com/journal/applsci

Appl. Sci. 2021, 11, 1384 2 of 15

Figure 1. China Railway Track System (CRTS) II slab track.

Some studies on the mapping relationship between subgrade settlement and track
deformation have been conducted. Guo et al. [10] derived an analytical expression for the
mapping relationship between subgrade settlement and track deformation based on the
two-dimensional Winkler foundation beam theory. However, it had a deviation from the
actual three-dimensional situation. Gou et al. [11] presented an analytical model to study
the effects of the lateral and vertical settlements of bridge pier on track geometry. Sun
et al. [12] studied the mapping relationship between ballast settlement and rail deformation
based on a theoretical model using an iterative method. However, the subgrade was not
modeled. Jiang et al. [13] studied the track deformation caused by the settlement of bridge
pier based on the principle of minimum potential energy. Cai et al. [14] developed a finite
element model for a double-block track by using the software ABAQUS (Dassault Systemes,
Paris, France.). However, the soil was considered as a linear elastic body, which cannot
accurately simulate the deformation of the soil subgrade. Xiao et al. [15] analyzed the effects
of differential subgrade settlement on the stress and displacement of the CRTS III slab track
based on the finite element method. In their study, the subgrade was also simplified as an
elastic body. Chen et al. [16] derived an analytical expression for track deformation under
pier settlement conditions for the slab track-bridge system. The dynamic responses of the
vehicle-track system under bridge pier settlement were also analyzed. Zhou et al. [17]
investigated the effects of vertical deformation of the continuous beam structure of the
CRTS I slab track on rail deformation based on the principle of stationary potential energy.
It can be seen from the above that there are more studies on bridge pier settlement, but
relatively few studies on subgrade settlement. The subgrade is often considered as an
elastic body by the previous studies, which has shortcomings in accurately simulating the
cumulative deformation and rheological behavior of the actual subgrade.
Besides, many researchers have investigated the effects of subgrade settlement on
the stress and dynamic responses of the track structure. Xiang et al. [18] established a
numerical model for a slab track based on the finite element method and studied the effects
of differential subgrade settlement on the stress of the slab track. However, the stress of the
subgrade had not been paid attention to. He et al. [19] studied the dynamic responses of
the vehicle-track system at the transition zones of the railway based on the vehicle-track
interaction theory. Xu et al. [20] established a dynamic analysis model for a ballastless track
based on the vehicle-track interaction theory and finite element method. Che et al. [21]
established a numerical model for a subgrade by using the discrete element method (DEM)
and studied the dynamic characteristics of the subgrade under cyclic loading. The results
indicated that the subgrade plays a crucial role in the stability of railway tracks and
affects the settlement of tracks. Li et al. [22] developed a coupled model that incorporates
both the DEM and the finite difference method (FDM) to study the dynamic behaviors
of the railway ballast assembly. The results showed the superiority of the DEM-FDM
coupled model in simulating the interaction between continuum and discrete domains.
Zhu et al. [23] conducted numerical triaxial tests, in which the rubber membrane and soil
Appl. Sci. 2021, 11, 1384 3 of 15

body were simulated by using the FDM and DEM, respectively. The modeling method of
the DEM-FDM model in [22] was adopted in this paper.
In summary, the differential subgrade settlement has an important effect on the stress
and deformation of the track structure. The analysis methods adopted by previous studies
mainly include the analytical method and the finite element method. The nonlinear contact
characteristics between layers of a track-subgrade system are often ignored in analytical
methods and the mechanical properties of track structures cannot be accurately reflected.
A finite element model can accurately simulate the stress and deformation of a track
structure. However, the subgrade was simplified as an elastic body and the stress state of
the subgrade was rarely paid attention to. The track structure undergoes the following
deformation in a subgrade settlement section, which changes the load of the track structure
on the subgrade and affects the continued settlement of the subgrade [24]. Therefore, there
are coupling effects between subgrade and track structure that must be considered. The
subgrade of a high-speed railway is largely composed of graded gravel. The subgrade may
be cracked or otherwise deteriorated when the settlement is large. In this case, it is not
appropriate to simulate the subgrade as a linear elastic body. In DEM, the granular material
is considered as assembled particles interacting by force-displacement law and Newton-
second law [25]. It is suitable for simulating discontinuous and large deformation problems,
which is consistent with the situation when the subgrade produces a large settlement. The
non-uniform contact between the track structure and subgrade in a settlement section
cannot be accurately simulated by using the DEM-FDM coupled approach.
Based on the above analysis, the slab track was considered as a continuous domain
using the FDM. The subgrade was considered as a discrete domain using the DEM. DEM-
FDM coupled simulation was realized by adopting interface elements for data interaction.
This paper is organized as follows. First, the description of the DEM-FDM coupled mecha-
nism was presented. Second, the coupled model was developed and verified. Finally, the
effects of differential subgrade settlement on track deformation and contact forces of gravel
grains were studied. The effects of settlement wavelength, settlement amplitude, and other
types of settlement were also discussed. It is expected that the obtained results could
provide helpful guidance for the maintenance and optimization of high-speed railways.

2. DEM-FDM Coupled Mechanism

The coupling of the DEM and FDM was realized by adopting interface elements
between the discrete and continuous domains. Interface elements transfer the load from
the discrete domain to the continuous domain and transfer the nodal velocity from the
continuous domain at the interface to the discrete domain. The interface between the
discrete and continuous domains is divided into triangles. A contact diagram for the
discrete and continuous domains is presented in Figure 2 [26].
Figure 2 represents a ball particle in contact with a triangular interface element. C
refers to the contact point and CP refers to the position of the point closest to C on the
triangular interface element. The positions of the three vertices of the triangular interface
element are denoted as xi (i = 1, 2, 3). The areas of the three triangles created by connecting
the three vertices of the triangle at CP are denoted as Ai (i = 1, 2, 3). The total area of the
3
three triangles is denoted as A = ∑ Ai .
i =1
The weighting factors ωi for each vertex are determined by dividing the triangular area
opposite from the vertex by the total area of the larger triangle as ωi = Ai /A. Therefore,
the expression ∑ ωi = 1 is satisfied, which ensures that the sum of the values extrapolated
from CP to the vertices is equal to the value at CP.
Appl. Sci. 2021, 11, 1384 4 of 15

Figure 2. Contact diagram for the coupled model.

Here ri (i = 1, 2, 3) refers to the vectors pointing from CP to the triangle vertices (i.e.,
→
ri = xi − CP). Let the forces applied at the contact point C be denoted as F . Because the
contact point C on the ball particle and the point CP on the interface element may not be at
the same location, the total moment of the ball particle acting on the interface element is
calculated as
→ →
M = (C − CP) × F . (1)
Take the forces applied on the triangle vertices to be Fi (i = 1, 2, 3), then the following
equations can be obtained:
→
∑ Fi = F, (2)
→ →
∑ r i × Fi = M. (3)
In the local coordinate system, the x component coincides with the normal direction
of the triangle and the y component coincides with the tangential direction of the triangle.
Therefore, in the local coordinate system, the equivalent force system on the vertices of the
triangle can be expressed as
→ →
∑ F i,x = F x, (4)
→ →
∑ F i,y = F y, (5)
→ →
∑ F i,z = F z = 0, (6)
→ → → → →
∑ ( r i,y × F i,z − r i,z × F i,y ) = Mx , (7)
→ → → → →
∑ ( r i,z × F i,x − r i,x × F i,z ) = My , (8)
→ → → → →
∑ ( r i,x × F i,y − r i,y × F i,x ) = Mz . (9)
Additionally, the following constraints are introduced to derive a specific solution to
the equations above:
→ →
∑ ( r i,z × F i,z ) = 0. (10)
The equations above are the basic governing equations for the adopted DEM-FDM
coupled method. During the simulation process, the node velocity, displacement, and load
information are continuously transmitted between the continuous and discrete domain to
realize coupling calculations.
Appl. Sci. 2021, 11, 1384 5 of 15

3. DEM-FDM Coupled Model

3.1. Modeling
The CRTS II slab track was considered as a continuous domain and the subgrade was
considered as a discrete domain. An FD model of a CRTS II slab track was developed by
using the software of FLAC3D (Itasca, Minneapolis, MN, USA). A DE model of the surface
layer of the subgrade was developed by using the software of PFC3D (Itasca, Minneapolis,
MN, USA). The FLAC3D can be loaded with the program load command in PFC3D 6.0
version [26]. So the PFC3D and FLAC3D can be run at the same time. The DEM-FDM
coupled model is presented in Figure 3. Because the slab track mainly experiences vertical
deformation under differential subgrade settlement, the lateral deformation of the track
structure can be ignored. Therefore, the horizontal size of the track structure was reduced
and only a strip-shaped model was established. The FD model consisted of the precast
slab, CA mortar layer, and concrete base. The rails and fasteners were not taken into
account. The width of the model along the transverse direction was 0.6 m. The length of
the model along the longitudinal direction was 30 m. The thicknesses of the precast slab,
CA mortar layer, and concrete base in the model were 0.2, 0.03, and 0.3 m, respectively. The
FD model of the CTRS II slab track has meshed with solid elements. This model contained
39,600 solid elements and 50,484 nodes. It was assumed that there is no separation or
slipping between the precast slab, CA mortar layer, and concrete base. The symmetry
constraints were applied to the two ends of the slab track.

Figure 3. The coupled model.

The surface layer of the subgrade is in direct contact with the slab track. The dynamic
stress in the surface layer of the subgrade attenuates quickly under train load and the
dynamic stress in the bottom layer of the subgrade is significantly reduced [27]. Therefore,
only the surface layer of the subgrade was considered in this study. Ball units were
used to model the gravel grains in the subgrade. Considering that the irregular shape of
the particles will affect the mechanical properties, the contact behaviors between grains
were simulated using the rolling resistance model to represent the interlocking between
irregularly shaped grains [28]. The thickness of the surface layer of the subgrade was 0.4 m.
The bottom and side boundaries of the surface layer of the subgrade model were simulated
with wall elements. The gradation curve of the gravel used in this study is presented in
Figure 4.
The average particle size of the gravel in the model was less than 40 mm. When the
container size is greater than eight times the average particle size, the boundary effects
of the container can be ignored [29]. Therefore, the boundary effects in this model can be
ignored. The surface of the slab track was defined as a coupled interface for data interaction.
The parameters for the model are listed in Table 1.
Appl. Sci. 2021, 11, 1384 6 of 15

Figure 4. Gradation curve of the gravel on the surface layer of the subgrade.

Table 1. Parameters for the DEM-FDM coupled model.

Components Parameter Value

Elastic modulus (MPa) 3.55 × 104
Precast slab Density (kg/m3 ) 2400
Poisson’s ratio 0.2
Elastic modulus (MPa) 8.00 × 103
CA mortar layer Density (kg/m3 ) 2400
Poisson’s ratio 0.2
Elastic modulus (MPa) 2.20 × 104
Concrete base Density (kg/m3 ) 2400
Poisson’s ratio 0.2
Density of gravel (kg/m3 ) 2300
Normal contact stiffness between gravel grains (N/m) 1.0 × 108
The surface layer of the subgrade Shear contact stiffness between gravel grains (N/m) 1.0 × 108
Sliding Friction coefficient 0.7
Rolling friction coefficient 0.25

3.2. Description of Subgrade Settlement

By fitting observation data from high-speed railway subgrade settlement, it was
determined that the shape of the subgrade in a settlement section is similar to a cosine
curve [30]. Therefore, a cosine curve was adopted to describe the differential settlement of
the bottom of the surface layer of the subgrade in this study, as shown in Figure 5.

Figure 5. Differential subgrade settlement.

In the settlement section, the expression for subgrade settlement along the longitudinal
direction of the railway line can be written as
  
1 1 2πx
z= f − cos , (11)
2 2 l

where z refers to the value of the settlement, f refers to the settlement amplitude, x refers
to the position coordinate, l refers to the settlement wavelength.
Appl. Sci. 2021, 11, 1384 7 of 15

3.3. Model Verification

According to the document “Code for Soil Test of Railway Engineering TB10102-
2010” [31], a foundation coefficient K30 test was performed using the coupled model
for verification. An FD model of a loading plate was established above the subgrade.
According to the requirements in the aforementioned document, a preload of 0.4 MPa
was first applied to the loading plate and then unloaded after the subgrade stabilized.
Reloading was then performed. The load and displacement of the loading plate were
recorded. The schematic diagram of the K30 test and the resulting load-displacement curve
are presented in Figure 6.

Figure 6. Schematic diagram and simulation results of the K30 test.

The load Ps at the reference settlement value Ss (1.25 mm) was obtained from the
load-displacement curve in Figure 6b. The foundation coefficient K30 can be calculated
as follows:
Ps
K30 = , (12)
πr2 Ss
where r represents the radius of the loading plate (0.15 m).
The calculation results indicated that the foundation coefficient K30 of the coupled
model was 200.26 MPa/m, which meets the requirements of the document “Code for
Design of High Speed Railway TB 10621-2014,” [32], which requires that the foundation
coefficient K30 of the surface layer of the subgrade should be no less than 190 MPa/m. This
indicates the practicality of the coupled model in this paper.
When the displacement of the loading plate is 1.25 mm, the resulting stress on the
loading plate and force chains in the surface layer of the subgrade are presented in Figure 7.
In Figure 7, the force chains represent contact between gravel grains. The radius of a force
chain corresponds to the value of the associated contact force.
As shown in Figure 7a, under the action of the loading plate, the strongest force chains
in the surface layer of the subgrade are mainly distributed under the loading plate and have
the shape of a rough truncated cone. The value of the force chain in contact with the loading
plate is the largest and continues to attenuate as the chain spreads downward. It can be seen
from Figure 7b that the coupled interface can transfer the supporting force of the subgrade
onto the loading plate and change the stress state of the loading plate. It can be inferred that
the node velocity, displacement, and load information can be stably transferred between
the continuous domain and discrete domain through the coupled interface.
Appl. Sci. 2021, 11, 1384 8 of 15

Figure 7. Stress on the loading plate and force chains in the surface layer of the subgrade.

4. Results and Discussion

4.1. Effects of Subgrade Settlement
Figure 8 shows the longitudinal stress on the slab track and force chains in the surface
layer of the subgrade under the settlement condition of 10 mm/15 m (settlement ampli-
tude/settlement wavelength). The calculation results of the unsettled condition are also
presented to highlight the effects of differential subgrade settlement.

Figure 8. Longitudinal stress on the slab track and force chain in the subgrade.

As shown in Figure 8a, the longitudinal stress on the slab track and the values of the
force chains in the subgrade are both small when there is no settlement in the subgrade. In
this case, the distribution of force chains is relatively uniform and the limited differences
between the distributions of force chains are caused by the bulk properties of the gravel in
the subgrade.
It can be seen from Figure 8b that the upper side of the track structure is subjected to
compressive stress and the lower side of the track structure is subjected to tensile stress
in the settlement section. Near the boundaries of the settlement section, there is a small
area in which the upper side of the track structure is in a tension state and the lower side is
in a compression state. This is because the slab track exhibits upward arching near both
boundaries of the settlement section. The contact forces between the gravel grains in the
settlement section are smaller than those when there is no settlement in the subgrade. In
Appl. Sci. 2021, 11, 1384 9 of 15

the center of the settlement section, there is no contact between the gravel grains and the
bottom surface of the concrete base. In other words, a gap appears between the concrete
base and the subgrade, meaning the track structure is in a suspended state near the center
of the settlement section. However, the force chains in the surface layer of the subgrade
near the boundaries of the settlement section are densely distributed and the values of the
contact forces are relatively large. This indicates that the gravel contact forces at the center
of the settlement section are reduced, but the contact forces between the gravel grains near
the boundaries of the settlement range are increased.
The average and maximum contact forces between the gravel grains under different
conditions are computed. The average and maximum contact forces between the gravel
grains in the unsettled condition are 2.83 × 102 and 2.62 × 103 N, respectively. Under
the settlement condition of 10 mm/15 m, the average and maximum contact forces are
2.87 × 102 and 7.58 × 103 N, respectively. Compared with the unsettled condition, the
average and maximum contact forces between the gravel grains under the settlement
condition of 10 mm/15 m increased by 1.41% and 189.31%, respectively. This indicates that
the average contact force between the gravel grains in the surface layer of the subgrade
increases slightly, but the maximum value of the contact force increases significantly
under subgrade settlement condition. The large contact force of the gravel grains near
the boundaries of the settlement section easily accelerates the plastic settlement of the
subgrade, causing expansion of the settlement section.
Figure 9 shows the deformation of the slab track under the settlement condition of
10 mm/15 m.

Figure 9. Deformation of the slab track.

As shown in Figure 9, the slab track exhibits deformation under the forces of gravity
in the subgrade settlement section. The deformation curve of the slab track is close to a
cosine shape, but it does not completely coincide with the subgrade settlement curve. The
deformation wavelength of the slab track is greater than the settlement wavelength of
the subgrade, but the deformation amplitude of the slab track is less than the settlement
amplitude of the subgrade. Furthermore, there are significant upward arches in the slab
track structure near the boundaries of the settlement section. It should be noted that
the train load is not considered in the analysis in this paper. Only the self-weight of the
structure is considered.

4.2. Effects of Settlement Wavelength

Variation in settlement wavelength has different effects on the deformation of the slab
track. Figure 10 shows the deformation curve and deformation amplitude of the slab track
when the settlement wavelength increases from 5 to 25 m and the settlement amplitude
remains unchanged at 10 mm.
Appl. Sci. 2021, 11, 1384 10 of 15

Figure 10. Deformation of the slab track at different settlement wavelengths.

It can be seen from Figure 10a that the deformation wavelength of the slab track
gradually increases with an increase in the settlement wavelength of the subgrade. As
shown in Figure 10b, the growth trend of the deformation amplitude of the slab track
first accelerates and then decelerates with an increase in the settlement wavelength of the
subgrade. The growth rate of the deformation amplitude of the slab track is the fastest when
the settlement wavelength of the subgrade increases from 10 to 15 m. When the settlement
wavelength reaches 25 m, the deformation amplitude of the slab track is 8.87 mm, which is
closer to the settlement amplitude of the subgrade.
Figure 11 shows the gravel contact forces in the subgrade for different settlement wave-
lengths, as well as the growth rate of the contact forces compared to the unsettled condition.

Figure 11. Contact force characteristics of the gravel grains in the subgrade at different settlement wavelengths.

As shown in Figure 11a, the average and maximum contact forces of the gravel grains
increase when the settlement wavelength increases from 0 to 20 m. However, when the
settlement wavelength is greater than 20 m, the average and maximum contact forces
decrease with increasing settlement wavelength. This is because the slab track is still in a
suspended state when the settlement wavelength is less than 20 m and a longer section
of the track is suspended with an increase in the settlement wavelength. A longer section
of suspended track causes the gravel grains near the boundaries of the settlement section
to bear an additional load. When the settlement wavelength is greater than 20 m, the
subgrade and slab track are in contact again due to the deformation of the slab track.
Appl. Sci. 2021, 11, 1384 11 of 15

Therefore, the contact forces of the gravel grains near the boundaries of the settlement
section are reduced.
Figure 11b indicates that an increase in the settlement wavelength has little effect
on the average contact force. However, the maximum contact force of the gravel grains
changes significantly with an increase in settlement wavelength. The maximum contact
force can reach approximately 2.5 times that in the unsettled condition.

4.3. Effects of Settlement Amplitude

Figure 12 shows the deformation curve and deformation amplitude of the track when
the settlement amplitude increases from 5 to 30 mm and the settlement wavelength remains
unchanged at 15 m.

Figure 12. Deformation of the track at different settlement amplitudes.

Figure 12a indicates that the deformation amplitude of the slab track gradually in-
creases with an increase in the settlement amplitude of the subgrade. However, the change
in the deformation wavelength of the slab track is not obvious. As shown in Figure 12b,
the deformation amplitude of the slab track increases nonlinearly with an increase in the
settlement amplitude of the subgrade. The growth rate of the deformation amplitude of the
slab track decreases when the settlement amplitude of the subgrade is greater than 10 mm.
In this case, there is a gap between the concrete base and subgrade. The track structure is
in a suspended state.
Figure 13 shows the gravel contact forces in the surface layer of the subgrade under
different settlement amplitudes, as well as the growth rate of the contact forces compared
to the unsettled condition.
Figure 13 indicates that the average contact force of the gravel grains in the subgrade
increases linearly with an increase in the settlement amplitude. However, the growth rate
of the average contact force is small (less than 4%) compared to the unsettled condition. The
maximum contact force increases nonlinearly with an increase in the settlement amplitude.
The growth rate of the maximum contact force decreases and the track is in a suspended
state when the settlement amplitude of the subgrade is greater than 10 mm. When the
settlement amplitude is 5 mm, the maximum contact force is 138.55% greater than that in
the unsettled condition. It can be inferred that the maximum contact force of the gravel
grains in the subgrade increases significantly even when the settlement amplitude of the
subgrade is small.
Appl. Sci. 2021, 11, 1384 12 of 15

Figure 13. Contact force characteristics of gravel grains in the subgrade at different settlement amplitudes.

4.4. Effects of Other Types of Settlement

Different types of subgrade settlement have different effects on the deformation of the
slab track and the contact force of the gravels in the subgrade. Site investigation shows that
the angular type and faulting type subgrade settlement often occur at the transition zones.
Therefore, the situations under differential subgrade settlement with angular type and
faulting type are also studied in this paper. It is assumed that the longitudinal length of the
angular type settlement is 15 m, and the settlement amplitude is 10 mm. The amplitude
of the faulting type settlement is taken as 10 mm. The deformation curves of the track
structure are shown in Figure 14.

Figure 14. Deformation of the slab track under other types of settlement.

It can be seen from Figure 14a that the deformation curve of the slab track is in good
agreement with the shape of the subgrade settlement with angular type. The deformation
curve of the slab track is smoother compared to the settlement curve of the subgrade.
As shown in Figure 14b, the slab track undergoes severe settlement in the faulting type
settlement section. There is a significant difference between the track deformation and
the subgrade settlement. In this case, there is a clear gap between the concrete base
and subgrade.
The calculation results indicate that the average contact force of the gravel grains in
the subgrade under angular type settlement is 2.93 × 102 N. The maximum contact force is
Appl. Sci. 2021, 11, 1384 13 of 15

1.01× 104 N. The average and maximum contact forces of the gravel grains in the subgrade
under faulting type settlement are 2.91 × 102 and 1.30× 104 N, respectively. The growth
rates of the gravel contact forces under angular and faulting types settlement compared to
the unsettled condition are shown in Figure 15. It can be seen that the maximum contact
forces of the gravel grains are significantly increased by the angular and faulting types
settlements, especially the faulting type settlement.

Figure 15. Growth rates of contact force under other types of settlement.

5. Conclusions
An FD model of a CRTS II slab track and a DE model of a subgrade were established.
DEM-FDM coupled simulations were realized by adopting interface elements for data
interaction. The effects of differential subgrade settlement on the deformation of the slab
track and contact forces of gravel grains in the subgrade were studied. The following
conclusions can be drawn.
(1) The slab track has a smaller deformation amplitude and larger deformation wave-
length compared to the subgrade. The contact force of the gravel grains near the
boundaries of the settlement section is relatively large, which easily causes the plastic
settlement of the subgrade and expansion of the settlement section.
(2) The deformation wavelength of the slab track increases with an increase in the settle-
ment wavelength of the subgrade. The slab track structure is in a suspended state
when the settlement wavelength is short. When the settlement wavelength is 20 m and
settlement amplitude is 10 mm, the maximum contact force can reach approximately
2.5 times that in the unsettled condition.
(3) The deformation amplitude of the slab track increases nonlinearly with an increase
in the settlement amplitude of the subgrade. The maximum contact force of the
gravel grains increases significantly, even if the settlement amplitude of the subgrade
is small.
(4) The angular and faulting types of settlements cause severe deformation of the track
structure. There is a clear gap between the concrete base and subgrade under a faulting
type settlement, which is harmful to the service state of the railway line. Therefore,
the faulting type settlement should be strictly controlled during maintenance.
However, the DEM-FDM coupled model in this paper still needs validation by more
tests for practical applications. In the future, the effect of the contact force of the gravel
grains on the service life and maintenance of the subgrade will be further studied. The
effect of external weather on the deformation of the track will also be investigated.
Appl. Sci. 2021, 11, 1384 14 of 15

Author Contributions: Funding acquisition, B.D.; investigation, X.C.; methodology, R.Z. and B.D.;
validation, G.G. and H.L.; visualization, R.Z.; writing—original draft, X.C. All authors have read and
agreed to the published version of the manuscript.
Funding: This research was funded by the Research and Development Program of China Railway
Corporation under Grant P2018G005-2, Research Project of Beijing-Shanghai High Speed Railway
Co., Ltd. under Grant 2020-15, and Shenzhen Science and Technology Innovation Commission under
Grant 827-000468.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Luo, J.; Zhu, S.; Zhai, W. An efficient model for vehicle-slab track coupled dynamic analysis considering multiple slab cracks.
Constr. Build. Mater. 2019, 215, 557–568. [CrossRef]
2. Hu, N.; Dai, G.L.; Yan, B. Recent development of design and construction of medium and long span high-speed railway bridges
in China. Eng. Struct. 2014, 74, 233–241. [CrossRef]
3. Jiang, L.; Zhang, Y.; Feng, Y.; Zhou, W.; Tan, Z. Dynamic response analysis of a simply supported double-beam system under
successive moving loads. Appl. Sci. 2019, 9, 2162. [CrossRef]
4. Gou, H.; Shi, X.; Zhou, W.; Cui, K.; Pu, Q. Dynamic performance of continuous railway bridges: Numerical analyses and field
tests. Proc. Inst. Mech. Eng. Part F J. Rail Rapid Transit. 2018, 232, 936–955. [CrossRef]
5. Kang, G. Influence and control strategy for local settlement for high-speed railway infrastructure. Engineering 2016, 2, 374–379.
[CrossRef]
6. Kouroussis, G.; Connolly, D.P.; Alexandrou, G.; Vogiatzis, K. The effect of railway local irregularities on ground vibration.
Transport. Res. Part D Transp. Environ. 2015, 39, 17–30. [CrossRef]
7. Kouroussis, G.; Connolly, D.; Alexandrou, G.; Vogiatzis, K. Railway ground vibrations induced by wheel and rail singular defects.
Veh. Syst. Dyn. 2015, 53, 1500–1519. [CrossRef]
8. Gou, H.; Zhou, W.; Bao, Y.; Li, X.; Pu, Q. Experimental study on dynamic effects of a long-span railway continuous beam bridge.
Appl. Sci. 2018, 8, 669. [CrossRef]
9. Wang, S.-J.; Xu, Z.-D.; Li, S.; Dyke, S.J. Safety and stability of light-rail train running on multispan bridges with deformation. J.
Bridge Eng. 2016, 21, 06016004. [CrossRef]
10. Guo, Y.; Gao, J.M.; Sun, Y.; Zhai, W. Mapping Relationship between Rail Deflection of Slab Track and Subgrade Settlement. J.
Southwest Jiaotong Univ. 2017, 52, 1139–1147+1215.
11. Gou, H.Y.; Yang, L.C.; Leng, D.; Bao, Y.; Pu, Q.H. Effect of bridge lateral deformation on track geometry of high-speed railway.
Steel. Compos. Struct. 2018, 29, 219–229.
12. Sun, Y.; Guo, Y.; Chen, Z.; Zhai, W. Effect of differential ballast settlement on dynamic response of vehicle–track coupled systems.
Int. J. Struct. Stab. Dy. 2018, 18, 1850091. [CrossRef]
13. Feng, Y.; Jiang, L.; Zhou, W.; Lai, Z.; Chai, X. An analytical solution to the mapping relationship between bridge structures vertical
deformation and rail deformation of high-speed railway. Steel Compos. Struct. 2019, 33, 209–224.
14. Cai, X.; Liu, W.; Wang, P.; Ning, X. Effect of Land Subsidence on Regularity of Double-Block Ballastless Track. Eng. Mech. 2014, 31,
160–165.
15. Xiao, W.; Guo, Y.; Gao, J.; Zhai, W. Effect of uneven subgrade settlement on the CRTS III slab track stress and deformation of
high-speed railway. J. Railw. Sci. Eng. 2015, 12, 724–730.
16. Chen, Z.; Sun, Y.; Zhai, W. Mapping relationship between pier settlement and rail deformation of high-speed railways-part (II):
The longitudinal connected ballastless track system. Sci. Sin. Technol. 2014, 44, 778–785.
17. Zhou, W.; Nie, L.; Jiang, L.-Z.; Feng, Y.; Tan, Z.; Chai, X. Mapping relation between pier settlement and rail deformation of unit
slab track system. Structures 2020, 27, 1066–1074. [CrossRef]
18. Xiang, J.; Lin, S.; Yu, C. Transfer rules and effect of stress and deformation of ballastless track under uneven subgrade settlement.
J. Traff. Trans. Eng. 2019, 19, 69–81.
19. He, C.; Chen, Z.; Zhai, W. Mapping relationship between uneven settlement of subgrade and rail deformation in subgrade-bridge
transition section and its dynamic application. Sci. Sin. Technol. 2018, 48, 881–890. [CrossRef]
20. Xu, Q.; Fan, H.; Li, B.; Zhou, X. Limited value for uneven settlement of subgrade under CRTS−II type slab track. J. Cent. South.
Univ. (Sci. Tech.) 2013, 44, 5038–5044.
21. Che, X.; Zhang, Z.; Li, Z.; Zhang, X. Discrete element modeling of ballasted railway track with silty-clay-subgrade under cyclic
load. Fresen. Environ. Bull. 2018, 27, 380–386.
22. Li, L.; Liu, W.; Ma, M.; Jing, G.; Liu, W. Research on the dynamic behaviour of the railway ballast assembly subject to the low
loading condition based on a tridimensional DEM-FDM coupled approach. Constr. Build. Mater. 2019, 218, 135–149. [CrossRef]
23. Zhu, H.X.; Yin, Z.Y.; Zhang, Q. A novel coupled FDM-DEM modelling method for flexible membrane boundary in laboratory
tests. Int. J. Numer. Anal. Methods Geomech. 2020, 44, 389–404. [CrossRef]
24. Guo, Y.; Zhai, W. Long-term prediction of track geometry degradation in high-speed vehicle–ballastless track system due to
differential subgrade settlement. Soil. Dyn. Earthq. Eng. 2018, 113, 1–11. [CrossRef]
Appl. Sci. 2021, 11, 1384 15 of 15

25. Zhang, Z.; Zhang, X.; Qiu, H.; Daddow, M. Dynamic characteristics of track-ballast-silty clay with irregular vibration levels
generated by high-speed train based on DEM. Constr. Build. Mater. 2016, 125, 564–573. [CrossRef]
26. Itasca Consulting Group, Inc. PFC 6.0 Documentation, User’s Guide; Itasca Consulting Group, Inc.: Minneapolis, MN, USA, 2019.
27. Jiang, G.L.; Kong, X.H.; Meng, L.J.; Wang, Z.M. Dynamic Characteristics of Soil Subgrade Bed for Ballastless Track. J. Southwest
Jiaotong Univ. 2010, 45, 855–862.
28. Lu, M.; McDowell, G.R. The importance of modelling ballast particle shape in the discrete element method. Granul. Matter. 2007,
9, 69. [CrossRef]
29. Li, N.; Zhou, R.; Hu, Q.; Liu, X. Mechanical fault diagnosis based on redundant second generation wavelet packet transform,
neighborhood rough set and support vector machine. Mech. Syst. Signal. Process. 2012, 28, 608–621. [CrossRef]
30. Guo, Y.; Zhai, W.; Sun, Y. A mechanical model of vehicle-slab track coupled system with differential subgrade settlement. Struct.
Eng. Mech. 2018, 66, 15–25.
31. Ministry of Railways of P.R. China. Code for Soil Test of Railway Engineering; TB10102-2010; Ministry of Railways of P.R. China:
Beijing, China, 2011; pp. 273–280.
32. Ministry of Railways of P.R. China. Code for Design of High Speed Railway; TB 10621-2014; Ministry of Railways of P.R. China:
Beijing, China, 2014; pp. 40–42.