coatings
Article
Study on the Evolution of Wheel Wear and Its Impact on
Vehicle Dynamics of High-Speed Trains
Maorui Hou 1,2, * , Bingzhi Chen 2 and Di Cheng 1
Citation: Hou, M.; Chen, B.; Cheng,
D. Study on the Evolution of Wheel
Wear and Its Impact on Vehicle
Dynamics of High-Speed Trains.
Coatings 2022, 12, 1333. https://
doi.org/10.3390/coatings12091333
Academic Editor: Antolin Hernández
Battez
Received: 26 June 2022
Accepted: 8 September 2022
Published: 14 September 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Coatings 2022, 12, 1333. https://doi.org/10.3390/coatings12091333
1 Railway Science and Technology Research and Development Center, China Academy of Railway Sciences
Corporation Limited, Beijing 100081, China
2 College of Locomotive and Vehicle Engineering, Dalian Jiaotong University, Dalian 116028, China
* Correspondence: houmaorui@126.com
Abstract: Wheel wear is one of the most critical factors affecting the vehicle performances and
maintenance costs of railway vehicles. However, previous research has to ignore the initial wheel-rail
profiles for the evolution of wheel wear. Therefore, this work investigates the relationship between the
evolution of wheel wear corresponding to different initial wheel-rail profiles and vehicle dynamics,
wheel-rail deterioration. Firstly, the evolution of wheel wear during a long service period is measured
from two high-speed railway trains running on two different lines. Contact geometry, e.g., equivalent
conicity and contact pair distribution, are extracted. After that, the influence of wheel wear on the
vehicle dynamic performance is studied using a multi-body dynamic software. The calculated contact
parameters, e.g., pressure, shear traction, and creepage, are used to analyze the distribution of rolling
contact fatigue. Based on the experimental and simulation results, the initial wheel and rail profiles
significantly affects the wheel wear pattern, the thin rim wheel has uniform wear, and other wheels
occurs hollow wear. The hollow wear can lead to gradual deterioration of vehicle dynamics, which
conversely aggravates the wheel reprofiling.
Keywords: wheel wear; wheel-rail interface; vehicle system dynamics; multi-body simulation
1. Introduction
The wheel-rail profile is crucial in defining how wheels and rails interact dynam-
ically [1–3]. In practice, minor tread alterations may significantly alter the equivalent
conicity and wheel-rail contact locations, altering the vehicle’s stability [4].
Wheel wear of the Chinese high-speed train is relatively complicated. The main
characteristics of the wheel-rail inter-face can be summarized as follows: high operational
speed (300–350 km/h), long operational distance (2300 km/day), ice, sand or another
environment. In railway applications, wheel wear due to rail-wheel interactions is a critical
issue. Increasing vehicle speeds and traffic volumes increase wheel wear, which reduces
wheel re-profiling intervals and decreases vehicle performance dynamically [5,6].
Many scholars have studied wheel wear through field measurements, laboratory
tests, and theoretical prediction. Numerous research on the influence of wheel-rail profiles
on dynamic performance [7,8] and deterioration in wheel-rails [9,10] have been done.
In laboratory tests, wear coefficients in various wheel-rail contact conditions were also
estimated [11]. Ostash et al. and Andreiko et al. researched the new high-strength wheel
steels with the best combination of high wear and fracture resistance [12,13]. Qi et al. [14]
analyzed the abnormal vibration problem of carbody caused by the wheel rail profile
through field tests. In many cases, the simulated wear distributions are successfully
developed to forecast the development of the wheel profile [15–18]. Cui et al. [19] proposed
a re-profiling threshold according to the wear forms for high-speed train’s wheels. Wang
et al. [20] studied the wheel re-profiling interval for the CRH380B EMU according to the
dynamic performance based on extensive field data set. An analysis of wheel-rail contact
https://www.mdpi.com/journal/coatings
Coatings 2022, 12, 1333 2 of 14

bandwidth and its impact on wheel tread wear was carried out by Gan et al. [21]. It can
be found from these studies that most of the existing studies focus on the new wheel-rail
profile in the literature, but the research on wheel wear pattern and wheel-rail interaction
for different service lines is few. However, the difference in the shape of the worn wheels
has a great impacts on vehicle dynamics of high-speed trains.
In this work, to analyze how wheel wear evolves and affects vehicle dynamics, wheel
wear evolution, wheel-rail contact pairs distribution, and vehicle dynamics behavior are
experimentally and numerically studied for high-speed trains. Section 2 demonstrates the
wheel wear evolution (e.g., the changes in wheel profiles and depth of the worn shape) in
one reprofiling cycles. In Section 3, wheel-rail interactions, such as the wheel-rail contact
pairs distribution and equivalent conicity, are comparatively analyzed for the worn wheels.
In Section 4, the effects of wheel wear on the dynamics behavior and wheel deterioration
are investigated through a multi-body analysis of the vehicle. Finally, some conclusions
are drawn.

2. The Evolution of Wheel Wear

From 2016, two high-speed lines in China, A and B, were tested for a long period of
time to record wheel profiles. The curve distribution characteristics of these two lines are
similar. The curve radius of the main line is basically greater than 55,000 m. Curves with a
radius less than 2000 m are mostly located around stations or railroad hubs, accounting for
less than 1%.
Table 1 illustrates the basic information of these two lines, including their length,
climates, operation speed, track types, and operation modes. The following table summa-
rizes the essential characteristics of these two railway lines, including climates, operating
speeds et al.
Table 1. Information about wheel-rail systems.

Annual
Rail Wheel
Axle Passing Rail Wheel
Name Length/km Speed/(km/h) Climates Track Type Hard- Hard-
Load/t Ton- Material Material
ness/HB ness/HB
nage/Mt
CRTS I
A 1776 200–250 Sandy (unballasted 14 4.2 U71MnK ER8 260–300 ≥245
track)
CRTS I
Dampness-
B 857 200–250 (unballasted 14 6.3 U71MnK ER8 260–300 ≥245
heat
track)

Wheel profiles are measured by the MiniProf instrument every 20,000 km or 50,000 km.
The MiniProf systems are high-precision (0.01 mm), full-contact profile measuring tools for
monitoring and analyzing the cross-sectional profiles of wheels. During a wheel re-profiling
interval, the wheel is tested 10 times for Line A, and the wheel is tested 7 times for Line B.
Figure 1 depicts the connection between wheel wear and operating distance for
all 64 wheels in 8 cars, including the maximum, minimum, median, abnormal, and quan-
tiles of wear depth. According to the statistical results of two high-speed lines, the wheel
wear gradually increases and presents an approximately linear relationship. Significantly,
the error of wheel wear gradually increases with operation distance. In the Line A, the
average wear rate is 0.05 mm/10,000 km, the maximum wear and the average wear are
about 1.5 mm and 1.1 mm at 218,000 km, respectively.
Similarly, in the Line B, the wheel wear rate is segmentally linear: the wheel wear
rate is about 0.022 mm/10,000 km before 113,000 km, and it further increases to about
0.025 mm/10,000 km after 113,000 km. At 218,000 km, the maximum wear on a wheel is
about 0.75 mm, and the average wear is about 0.55 mm.
The worn profiles of a wheel with Line A are shown in Figure 2a. As shown in
Figure 2b, the firstly-measured wheel profile be the reference, the profiles tested after
different operation distance is subtracted from the reference profile in the normal direction
 1.6 0.8
1.4

1.2
0.6
Coatings 2022, 12, 1333 1.0 3 of 14

Wear [mm]

Wear [mm]
0.8
0.4
0.6

0.4
to obtain the wear distribution. The calculations demonstrate 0.2 that when hollow wear begins
0.2
in the wheel, the dominant wear concentrated on the lateral position between −20~30 mm.
0.0
At lateral2.6 positions 13.2 15.2 −17.7
between 19.3 −
35 and 21.845 mm, flange wear occurs with a11.3
coordinate system
Coatings 2022, 12, x FOR PEER REVIEW  0.0 3  of  15 
  5.4 7.7 10.1 3.2 5.7 8.5 15.2 21.8
defined at theOperation
nominal
distancerolling
[10,000 km] radius. Noticeably, the flange wear
 
Operationhas not
distance been
[10,000 km] observed
 
for Line A.
(a)  (b) 
1.6 0.8
Figure 1. Wheel wear in one re‐profiling interval: (a) Line A; (b) Line B. 
1.4

1.2 Similarly, in the Line B, the wheel wear rate is segmentally linear: the wheel wear 
0.6
1.0 rate is about 0.022 mm/10,000 km before 113,000 km, and it further increases to about 0.025 
Wear [mm]

Wear [mm]
0.8 mm/10,000 km after 113,000 km. At 218,000 km, the maximum wear on a wheel is about 
0.4
0.6 0.75 mm, and the average wear is about 0.55 mm. 
0.4
The worn profiles of a wheel with Line A are shown in Figure 2a. As shown in Figure 
0.2
2b, the firstly‐measured wheel profile be the reference, the profiles tested after different 
0.2
operation distance is subtracted from the reference profile in the normal direction to ob‐
0.0
tain the wear distribution. The calculations demonstrate that when hollow wear begins in 
0.0
2.6 5.4 7.7 10.1 13.2 15.2 17.7 19.3 21.8 3.2 5.7 8.5 11.3 15.2 21.8
the wheel, the dominant wear concentrated on the lateral position between −20~30 mm. 
Operation distance [10,000 km] Operation distance [10,000 km]
 
At lateral positions between −35 and −45 mm, flange wear occurs with a coordinate system   
defined at the nominal rolling radius. Noticeably, the flange wear has not been observed 
(a) 
for Line A. 
(b) 
Figure 1. Wheel wear in one re‐profiling interval: (a) Line A; (b) Line B. 
Figure 1. Wheel wear in one re-profiling interval: (a) Line A; (b) Line B.
Similarly, in the Line B, the wheel wear rate is segmentally linear: the wheel wear 
rate is about 0.022 mm/10,000 km before 113,000 km, and it further increases to about 0.025 
mm/10,000 km after 113,000 km. At 218,000 km, the maximum wear on a wheel is about 
0.75 mm, and the average wear is about 0.55 mm. 
The worn profiles of a wheel with Line A are shown in Figure 2a. As shown in Figure 
2b, the firstly‐measured wheel profile be the reference, the profiles tested after different 
operation distance is subtracted from the reference profile in the normal direction to ob‐
tain the wear distribution. The calculations demonstrate that when hollow wear begins in 
the wheel, the dominant wear concentrated on the lateral position between −20~30 mm. 
At lateral positions between −35 and −45 mm, flange wear occurs with a coordinate system 
defined at the nominal rolling radius. Noticeably, the flange wear has not been observed 
   
for Line A. 
(a)  (b) 

Figure 2. MeasuredFigure 2. Measured wheel profiles and wheel wear distribution of the Line A: (a) wheel profiles; (b) 
wheel profiles and wheel wear distribution of the Line A: (a) wheel profiles;
wheel wear distribution. 
(b) wheel wear distribution.
As  shown  in  Figure  3,  there is  a  slight change  in  the  wheel rim. In  addition,  a  lateral  wear 
As shown in Figure 3, there is a slight change in the wheel rim. In addition, a lateral
position of −35 to 50 mm with a broad wear range was stated for the ‘uniform wear’. As a result of 
Coatings 2022, 12, x FOR PEER REVIEW  4  of  15 
  wear position of −35 to 50 mm with a broad wear range was stated for the ‘uniform wear’.
the economic reprofiling strategy, initial flange wear is clearly visible. 
As a result of the economic reprofiling strategy, initial flange wear is clearly visible.

     

(a)  (b) 
Figure 2. Measured wheel profiles and wheel wear distribution of the Line A: (a) wheel profiles; (b) 
wheel wear distribution. 

As  shown  in  Figure  3,  there is  a  slight change  in  the  wheel rim. In  addition,  a  lateral  wear 
position of −35 to 50 mm with a broad wear range was stated for the ‘uniform wear’. As a result of 
   
the economic reprofiling strategy, initial flange wear is clearly visible. 
(a)  (b) 
Figure 3. Wheel profiles and wear distribution of the Line B: (a) wheel tread profiles; (b) wheel wear 
Figure 3. Wheel profiles and wear distribution of the Line B: (a) wheel tread profiles; (b) wheel wear
distribution. 
distribution.
3. Wheel‐Rail Contact Geometry 
  The  contact  geometry  is  defined  by  gauge,  rail  cant,  and  profile  of  the  wheel  that 
contact with railhead. In China, it uses a gauge of 1435 mm, a cant of 1:40, a wheelback 
distance of 1353 mm, a wheel diameter of 860 mm. In addition, the evolution of equivalent 
conicity, wheel‐rail contact pairs distribution for the two high‐speed lines are compared. 

3.1. Equivalent Conicity 
Coatings 2022, 12, 1333 4 of 14

3. Wheel-Rail Contact Geometry

The contact geometry is defined by gauge, rail cant, and profile of the wheel that
contact with railhead. In China, it uses a gauge of 1435 mm, a cant of 1:40, a wheelback
distance of 1353 mm, a wheel diameter of 860 mm. In addition, the evolution of equivalent
conicity, wheel-rail contact pairs distribution for the two high-speed lines are compared.

3.1. Equivalent Conicity

The equivalent conicity is a key parameter and it relates to the dynamic performance
and running stability of the vehicle. Even in the steep curve, a high level of comparable
conicity permits good curving behavior without flange contact. However, this also reduces
the essential speed, perhaps resulting in severe hunting instability. Low equivalent conicity,
on the other hand, provides for steady vehicle operation at high speeds but leads to
flange contact.
When the operation distance reaches the limit value, the wheel should be re-profiled
to restore worn profiles to the original shapes. In order to prolong the wheelset life, one
frequent method is economical re-profiling, which uses the smallest amount of material cut
at each re-profiling operation. The vehicle management department has proposed different
re-profiling strategies.
Figure 4 plots the initial wheel and rail profiles. Hence, in Figure 4a, the flange profile
has 2 mm gaps for the wheel re-profiled. However, the two wheel profiles is coincide
completely for the wheel tread section between the −20 and 40 mm. Exploring reasons for
Coatings 2022, 12, x FOR PEER REVIEW  5  of  15 
  thin flange due to remove little material on turning lathes. To address wheels with severe
flange wear, a variety of narrow flange profiles (changing flange thickness) can greatly
extend the service lifetimes of high-speed wheels.

30
0
Vertical coordinate [mm]
Vertical coordinate [mm]

25 line A
60D
line B
20 -10 0 line A
-2 line B
15
-4
-20
10 -6
-8
5 -30
-10
0 10 20 30
0
-40
-5 -40 -20 0 20 40
-60 -40 -20 0 20 40 60
Lateral coordinate [mm] Lateral coordinate [mm]  
 

(a)  (b) 
Figure 4. Profiles of two railway lines’ wheels and rails: (a) re‐profiling shapes and (b) rail profile. 
Figure 4. Profiles of two railway lines’ wheels and rails: (a) re-profiling shapes and (b) rail profile.

The measured rail profiles of the two lines are identical to the 60D profile, as shown 
The measured rail profiles of the two lines are identical to the 60D profile, as shown in
in Figure 4b. The rail profile of the Line B with a certain positive deviation slightly bugles 
Figure 4b. The rail profile of the Line B with a certain positive deviation slightly bugles
at the
at the working
working sideside gauge.  The rail
gauge. The rail grinding
grinding profile
profile by  the grindstone,
by the grindstone, so
so the
the working
working 
accuracy is far less than turning lathes to restore the original tread profile. The rail grind‐
accuracy is far less than turning lathes to restore the original tread profile. The rail grinding
ing strategies adopted by different engineering departments is slightly different, resulting 
strategies adopted by different engineering departments is slightly different, resulting in
in slight manufacturing error between the measured profiles and the 60D. 
slight manufacturing error between the measured profiles and the 60D.
Based on the UIC 519 standard, the equivalent conicity is determined here using the
Based on the UIC 519 standard, the equivalent conicity is determined here using the 
measured profiles [22]. The equivalent conicity of two lines after reprofiling the wheel are
measured profiles [22]. The equivalent conicity of two lines after reprofiling the wheel are 
compared, as shown in Figure 5. Due to the thin flange of the Line B, the wheel-rail gap
compared, as shown in Figure 5. Due to the thin flange of the Line B, the wheel‐rail gap 
increases, the wheel enters the flange root, and the equivalent conicity begins to increase
increases, the wheel enters the flange root, and the equivalent conicity begins to increase 
gradually when the wheelset lateral displacement reaches 10 mm.
gradually when the wheelset lateral displacement reaches 10 mm. 

0.8

line A
0.6 line B
conicity [-]

0.4
 ing strategies adopted by different engineering departments is slightly different, resulting 
in slight manufacturing error between the measured profiles and the 60D. 
Based on the UIC 519 standard, the equivalent conicity is determined here using the 
measured profiles [22]. The equivalent conicity of two lines after reprofiling the wheel are 
compared, as shown in Figure 5. Due to the thin flange of the Line B, the wheel‐rail gap 
Coatings 2022, 12, 1333 5 of 14
increases, the wheel enters the flange root, and the equivalent conicity begins to increase 
gradually when the wheelset lateral displacement reaches 10 mm. 

0.8

line A
0.6 line B

Equivalent conicity [-]

0.4

0.2

0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Wheelset lateral displacement [mm]
 
Figure 5. Equivalent conicity curves.
Figure 5. Equivalent conicity curves. 

Figure 6 depicts the connection between equivalent conicity and operation distance
Figure 6 depicts the connection between equivalent conicity and operation distance 
for all 64 wheels in eight cars of a single train. Figure 6a shows that the equivalent conicity
for all 64 wheels in eight cars of a single train. Figure 6a shows that the equivalent conicity 
of the Line A gradually rises monotonically, and the increasing rate of equivalent conicity
of the Line A gradually rises monotonically, and the increasing rate of equivalent conicity 
is about 0.006/10,000 km. The average statistic conicity starts from 0.07 and ends at 0.19 in
Coatings 2022, 12, x FOR PEER REVIEW  is about 0.006/10,000 km. The average statistic conicity starts from 0.07 and ends at 0.19 in 
6  of  15 
  one re-profiling cycle. However, the equivalent conicity is relatively stable for the Line B at
one re‐profiling cycle. However, the equivalent conicity is relatively stable for the Line B 
about 0.03 in a whole re-profiling cycle, as shown in Figure 6b. As the wide range of wheel
at about 0.03 in a whole re‐profiling cycle, as shown in Figure 6b. As the wide range of 
tread wear, the equivalent conicity does not increase.
wheel tread wear, the equivalent conicity does not increase. 
0.3
0.08

0.06
Equivalent conicity [-]

Equivalent conicity [-]

0.2

0.04

  0.1
0.02

0.0 0.00
2.6 5.4 7.7 10.1 13.2 15.2 17.7 19.3 21.8 3.2 5.7 8.5 11.3 15.2 21.8
Operation distance [10,000 km] Operation distance [10,000 km]
   

(a)  (b) 
Figure 6. Equivalent conicity revolution of different railway lines: (a) Line A; (b) Line B. 
Figure 6. Equivalent conicity revolution of different railway lines: (a) Line A; (b) Line B.

3.2. Wheel-Rail Contact Pairs Distribution

3.2. Wheel‐Rail Contact Pairs Distribution 
The wheel rail 
The  wheel and  and profiles 
rail profiles obtained
obtained  onfield 
on  the  the field test every
test  every  50,000are 
50,000 km  kmused 
are used
to  to
compute the development of the wheel-rail contact pairings distribution. When the
compute the development of the wheel‐rail contact pairings distribution. When the wheel‐
wheelset’s lateral displacement changes from −10 to 10 mm, the contact point is com-
set’s lateral displacement changes from −10 to 10 mm, the contact point is computed every 
puted every 0.5 mm.
0.5 mm. 
The evolution of contact point is illustrated in Figure 7. The wheel-rail contact point
The evolution of contact point is illustrated in Figure 7. The wheel‐rail contact point 
for a lateral displacement is shown by the thin line connecting the wheel to the rail, and
for a lateral displacement is shown by the thin line connecting the wheel to the rail, and 
the number above the wheel denotes the wheelset lateral displacement. Figure 7a depicts
the number above the wheel denotes the wheelset lateral displacement. Figure 7a depicts 
how the wheel-rail contact points are dispersed in the center tread after a wheel repair, but
how the wheel‐rail contact points are dispersed in the center tread after a wheel repair, 
the flange contact will occur when the wheelset travels more than 8 mm from its original
but the flange contact will occur when the wheelset travels more than 8 mm from its orig‐
inal  position.  With
position. With the
the increasing
increasing ofof 
operating distance,
operating  distance, wheel-rail
wheel‐rail contact points
contact  areare 
points  mainly
distributed in three areas: − 10~0 mm, 2~8 mm, and 9~10 mm. The
mainly distributed in three areas: −10~0 mm, 2~8 mm, and 9~10 mm. The wheel‐rail con‐ wheel-rail contact
bandwidth gradually increases, and the contact point gradually draws closer to the rail
tact bandwidth gradually increases, and the contact point gradually draws closer to the 
shoulder. During this period, the equivalent conicity steadily rises, resulting in a reduction
rail shoulder. During this period, the equivalent conicity steadily rises, resulting in a re‐
in the vehicle’s running stability and riding quality.
duction in the vehicle’s running stability and riding quality. 
Figure 7b illustrates the distribution of wheel‐rail contact pairings always keeps the 
center wheel tread and rail head and the bandwidth of the wheel‐rail contact has been 
somewhat improved. 
Coatings 2022, 12, x FOR PEER REVIEW  7  of  15 
 
Coatings 2022, 12, 1333 6 of 14

 
 

(a)  (b) 
Figure 7. Wheel‐rail contact pairs distribution: (a) Line A (b) Line B. 
Figure 7. Wheel-rail contact pairs distribution: (a) Line A (b) Line B.

4. Dynamics Analysis of Wheel Wear 
Figure 7b illustrates the distribution of wheel-rail contact pairings always keeps the
center wheel tread and rail head and the bandwidth of the wheel-rail contact has been
With the goal of better understanding the impact of wheel wear, the numerical sim‐
somewhat improved.
ulation was used to analyze the dynamics performance and wheel degradation of a high‐
speed vehicle. 
4. Dynamics Analysis of Wheel Wear
With the goal of better understanding the impact of wheel wear, the numerical
4.1. Numerical Modeling 
simulation was used to analyze the dynamics performance and wheel degradation of
Following the analysis of wheel wear in various lines, the impact of wheel wear on 
a high-speed vehicle.
vehicle dynamics is investigated further in this section of the paper. The commercial pro‐
gram Simpack creates a vehicle model based on characteristics taken from a high‐speed 
4.1. Numerical Modeling
train,  as seen  in Figure  8 and  Table  2. Multi‐body  dynamics is used  to  create a  vehicle 
Following the analysis of wheel wear in various lines, the impact of wheel wear on
model  with  two  bogie  subsystems  that  completely  encompass  nonlinear  wheel‐rail 
vehicle dynamics is investigated further in this section of the paper. The commercial pro-

 
 Coatings 2022, 12, 1333
Coatings 2022, 12, x FOR PEER REVIEW  8  of 715 
of 14
  Coatings 2022, 12, x FOR PEER REVIEW  8  of  15 
 

gram Simpack creates a vehicle model based on characteristics taken from a high-speed
contact, and suspension parts. The carbody and frame, as well as the wheelsets, have six 
contact, and suspension parts. The carbody and frame, as well as the wheelsets, have six 
train, as seen in Figure 8 and Table 2. Multi-body dynamics is used to create a vehicle
degrees of freedom, but the axle‐box has just one degree of freedom, so the model has 46 
degrees of freedom, but the axle‐box has just one degree of freedom, so the model has 46 
model with two bogie subsystems that completely encompass nonlinear wheel-rail contact,
degrees of freedom. Using the Kalker nonlinear creep and the Hertz contact theory, the 
degrees of freedom. Using the Kalker nonlinear creep and the Hertz contact theory, the 
and suspension parts. The carbody and frame, as well as the wheelsets, have six degrees of
creep force between the wheel and the rail is calculated. Hertz theory is used to solve the 
creep force between the wheel and the rail is calculated. Hertz theory is used to solve the 
freedom, but the axle-box has just one degree of freedom, so the model has 46 degrees of
normal contact problem, and the creep force is estimated using the FASTSIM technique. 
normal contact problem, and the creep force is estimated using the FASTSIM technique. 
freedom. Using the Kalker nonlinear creep and the Hertz contact theory, the creep force
The suggested friction coefficient between wheel and rail for running dynamics analysis 
The suggested friction coefficient between wheel and rail for running dynamics analysis 
between the wheel and the rail is calculated. Hertz theory is used to solve the normal
is 0.4 (dry condition). 
is 0.4 (dry condition). 
contact problem, and the creep force is estimated using the FASTSIM technique. The
The track irregularity is also considered here, and it comes from the measured track 
The track irregularity is also considered here, and it comes from the measured track 
suggested friction coefficient between wheel and rail for running dynamics analysis is 0.4
irregularity data with a length of 5000 m, as shown in Figure 9. 
irregularity data with a length of 5000 m, as shown in Figure 9. 
(dry condition).

 
 
Figure 8. The numerical model. 
Figure 8. The numerical model. 
Figure 8. The numerical model.
Table 2. Critical parameters of high‐speed train. 
Table 2. Critical parameters of high‐speed train. 
Table 2. Critical parameters of high-speed train.
Specification  Value  Specification  Value 
Specification 
Specification Value 
Value Specification 
Specification Value 
Value
Carbody weight (t)  33.79  Wheelset distance (m)  2.50 
Carbody weight (t) 
Carbody weight (t)
33.79  Wheelset distance (m) 
33.79 Vertical stiffness of primary springs (MN/m) 
Wheelset distance (m)
2.50 
2.50
Bogie frame weight (t)  2.70  1.18 
Bogie frame weight (t) 
Bogie frame weight (t) 1.87  2.70  Vertical stiffness of primary springs (MN/m) 
2.70 Lateral stiffness of primary springs (MN/m) 
Vertical stiffness of primary springs (MN/m) 1.18 
1.18
Wheelset weight (t)  0.98 
Wheelset weight (t) 
Wheelset weight (t) 1.87  Lateral stiffness of primary springs (MN/m)  0.98 
1.87 Lateral stiffness of primary springs (MN/m) 0.98
Bogie distance (m) 17.50 
Longitudinal stiffness of axle‐box joint 
17.50 Longitudinal stiffness of axle-box joint (MN/m) 13.70
Bogie distance (m)  Longitudinal stiffness of axle‐box joint  13.70 
Bogie distance (m)  17.50  (MN/m)  13.70 
(MN/m) 
  The track irregularity is also considered here, and it comes from the measured track
 
irregularity data with a length of 5000 m, as shown in Figure 9.

   
   
(a)  (b) 
(a)  (b) 
Figure 9. Track irregularity: (a) Lateral, (b) Vertical. 
Figure 9. Track irregularity: (a) Lateral, (b) Vertical.
Figure 9. Track irregularity: (a) Lateral, (b) Vertical. 
4.2. Running Stability and Ride Quality
4.2. Running Stability and Ride Quality 
4.2. Running Stability and Ride Quality 
The vehicle stability and ride quality under wheel wear are initially assessed based on
The vehicle stability and ride quality under wheel wear are initially assessed based 
the The vehicle stability and ride quality under wheel wear are initially assessed based 
measured wheel and rail profiles. Accelerations gathered from the carbody floor and
on the measured wheel and rail profiles. Accelerations gathered from the carbody floor 
on the measured wheel and rail profiles. Accelerations gathered from the carbody floor 
the bogie frame above thethe 
wheelset areare 
used to perform a dynamic assessment of lateral
and  the  bogie  frame  above  wheelset  used  to  perform  a  dynamic  assessment  of 
and  the  bogie  frame  above  the  wheelset  are  used  to  perform  a  dynamic  assessment  of 
lateral stability and ride comfort. This is known as the Sperling index is a commonly used 
lateral stability and ride comfort. This is known as the Sperling index is a commonly used 
for assessing the vehicle’s comfort, and is described as: 
for assessing the vehicle’s comfort, and is described as: 

 
 
 𝐴
𝑊 7.08 𝐹 𝑓   (1)
𝑓
Coatings 2022, 12, 1333 8 of 14
where A is the vibration acceleration’s amplitude in the frequency domain, f is the fre‐
quency, and F(f) is the frequency‐weighted function. According to GB 5599‐2019 [23], the 
limit value of ride comfort is set at 2.5 for the excellent level, 2.75 for the medium level, 
stability and ride comfort. This is known as the Sperling index is a commonly used for
and 3.0 for the qualifying level. 
assessing the vehicle’s comfort, and is described as:
The bogie lateral acceleration is filtered by bandpass filter 0.5–10 Hz, then a limit of 
8 m/s2 defined in GB5599 was utilized to analyze bogie hunting. Filter parameters are set 
s
A3
as follows: filter type is Butterworth, and window length is 500 m. 
W = 7.08 10
F( f ) (1)
f
Figure 10 depicts the impact of wheel wear on bogie frame’s nonlinear critical speed, 
lateral stability, and lateral acceleration. To maintain the high‐speed train’s stability, the 
where A is the vibration acceleration’s amplitude in the frequency domain, f is the frequency,
maximum speed should be more than 350 km/h for the 250 km/h class high‐speed railway, 
and F(f)taking into account the 40% safety margin. Figure 10a shows that the maxi‐mum speed of 
is the frequency-weighted function. According to GB 5599-2019 [23], the limit
value oftwo lines is 450 km/h after wheels are re‐profiled. The maximum speed for Line A steadily 
ride comfort is set at 2.5 for the excellent level, 2.75 for the medium level, and 3.0
for the qualifying level.
falls from 450 km/h to 370 km/h as the service distance increases, but it is still more than 
Thethe running speed of 250 km/h. The non‐linear critical speed fluctuates within 450 km/h 
bogie lateral acceleration is filtered by bandpass filter 0.5–10 Hz, then a limit of
8 m/s2 defined in GB5599 was utilized to analyze bogie hunting. Filter parameters are set
with little change corresponding to the B line. 
as follows: filter type is Butterworth, and window length is 500 m.
Figure 10b demonstrates that the lateral ride index stables at 1.5 at the beginning of 
Figure 10 depicts the impact of wheel wear on bogie frame’s nonlinear critical speed,
the first 50,000 km, then it is more significant than 2.75 for 152,000 km, so the ride quality 
lateral stability, and lateral acceleration. To maintain the high-speed train’s stability, the
class decrease from “excellent” to “good”. However, the lateral ride index is retained at 
maximum speed should be more than 350 km/h for the 250 km/h class high-speed railway,
1.5, keeping as “excellent” in the Line B. 
taking into account the 40% safety margin. Figure 10a shows that the maxi-mum speed of
By comparing the findings from the two groups in Figure 10c, the lateral acceleration 
two lines(Peak‐to‐Peak) increases monotonically with the operating distance for line A, the accel‐
is 450 km/h after wheels are re-profiled. The maximum speed for Line A steadily
falls from 450 km/h to 370 km/h 2as
eration goes from 1.38 m/s the service
 to 3.49 m/s distance increases, but it is still more than
2 in 218,000 km. Contrary, the lateral acceleration of 

the running speed of 250 km/h. The non-linear

bogie frame basically fluctuates between 1.2 m/scritical2 and 1.4 m/s
speed fluctuates
2.  within 450 km/h
with little change corresponding to the B line.

500 3.0 line A

line A
line B line B
475
2.5
450
Critical speed [km/h]

2.0
Lateral ride index

425

400 1.5

375
1.0
350
0.5
325

300 0.0
-5 0 5 10 15 20 25 -5 0 5 10 15 20 25
Operation distance [10,000 km] Operation distance [10,000 km]
   

(a)  (b) 

6  
line A
5 line B
Acceleration(m/s2)

0
-5 0 5 10 15 20 25
Operation distance [10,000 km]
 

(c)   

Figure 10. Vehicle dynamics index. (a) critical speed; (b) lateral ride index and (c) carbody
  acceleration.

Figure 10b demonstrates that the lateral ride index stables at 1.5 at the beginning of
the first 50,000 km, then it is more significant than 2.75 for 152,000 km, so the ride quality
class decrease from “excellent” to “good”. However, the lateral ride index is retained at 1.5,
keeping as “excellent” in the Line B.
 Coatings 2022, 12, x FOR PEER REVIEW  10  of  15 
 

Coatings 2022, 12, 1333 9 of 14

Figure 10. Vehicle dynamics index. (a) critical speed; (b) lateral ride index and (c) carbody acceler‐
ation. 

By comparing theTime histories of the lateral acceleration are shown in Figures 11 and 12 for Line A. 
findings from the two groups in Figure 10c, the lateral acceleration
In Figure 11a,b, the lateral acceleration increases gradually. When the operating distance 
(Peak-to-Peak) increases monotonically with the operating distance for line A, the accelera-
is under 50,000 km, the maximum acceleration is 0.75 m/s2. However, when the operating 
tion goes from 1.38 m/s2 to 3.49 m/s2 in 218,000 km. Contrary, the lateral acceleration of
distance reaches 100,000 km, the acceleration exceeds 1.0 m/s
2 and 1.4 m/s2 .
2, with a maximum value of 
bogie frame basically fluctuates
1.2 m/s2. 
between 1.2 m/s
Time histories ofThe frequency spectra of carbody acceleration (FFT) are further shown in Figure 11c, 
the lateral acceleration are shown in Figures 11 and 12 for Line A.
In Figure 11a,b, the lateral acceleration increases gradually. When the operating distance
it can be seen that the central frequency of carbody lateral acceleration is less than 2 Hz 
is under 50,000 km, the maximum acceleration is 0.75 m/s2 . However, when the oper-
after wheel re‐profiling. The core frequency rises to 3.5 Hz at 54,000 km and it rises to 5–6 
ating distance reaches 100,000 km, the acceleration exceeds 1.0 m/s2 , with a maximum
Hz at 100,000 km or more. 
2
value of 1.2 m/s .

     

(a)  (b) 

0.10  
Power spectral density /[(m/s2)2Hz1]

0 km 54,000 km 101,000 km
152,000 km 218,000 km
0.08

0.06

0.04

0.02

0
0 1 2 3 4 5 6 7 8 9 10
Frequency /Hz
 

(c)   

Figure 11. VibrationFigure 11. Vibration on carbody ends for the Line A: (a) time‐domain of the service distance less 

on carbody ends for the Line A: (a) time-domain of the service distance less than
than 54,000 km; (b) time‐domain of the service distance more than 101,000 km;(c) frequency domain. 
54,000 km; (b) time-domain of the service distance more than 101,000 km;(c) frequency domain.
The frequency spectra of carbody acceleration (FFT) are further shown in Figure 11c, it
Figures 12 illustrate the bogie frame’s lateral acceleration. This graph shows how the 
bogie frame’s lateral acceleration amplitude steadily grows as the operating distance in‐
can be seen that the central frequency of carbody lateral acceleration is less than 2 Hz after
wheel re-profiling.creases. When the operation distance of the vehicle is less than 50,000 km, the maximum 
The core frequency rises to 3.5 Hz at 54,000 km and it rises to 5–6 Hz at
100,000 km or more.acceleration becomes 2 m/s2; when the operation distance is more than 100,000 km, the 
maximum acceleration reaches 4 m/s2. The frequency spectra of bogie frame vibration is 
Figure 12 illustrate the bogie frame’s lateral acceleration. This graph shows how
shown in Figure 12c. After the maintenance of wheel, the central frequency of the carbody 
the bogie frame’sis more diminutive than 2 Hz, and it will increase to 3.5 Hz with an operation distance of 
lateral acceleration amplitude steadily grows as the operating distance
increases. When the operation distance of the vehicle is less than 50,000 km, the maximum
54,000 km. Similar to the carbody, the central frequency of the bogie frame also rises to 5–
acceleration becomes 2 m/s2 ; when the operation distance is more than 100,000 km, the
maximum acceleration reaches 4 m/s2 . The frequency spectra of bogie frame vibration is
  shown in Figure 12c. After the maintenance of wheel, the central frequency of the carbody
is more diminutive than 2 Hz, and it will increase to 3.5 Hz with an operation distance of
54,000 km. Similar to the carbody, the central frequency of the bogie frame also rises to
5–7 Hz with an operation distance of 100,000 km. Lesser low frequencies in the same plot
relate to stiff motion of the carbody and hunting motion of the wheelset.
 Coatings 2022, 12, x FOR PEER REVIEW  11  of  15 
 

7 Hz with an operation distance of 100,000 km. Lesser low frequencies in the same plot 
Coatings 2022, 12, 1333 10 of 14
relate to stiff motion of the carbody and hunting motion of the wheelset. 

   

(a)  (b) 

0.6 0 km 54,000 km 101,000 km  

Power spectral density/[(m/s2)2/Hz1]

152,000 km 218,000 km
0.5

0.4

0.3

0.2

0.1

0
0 1 2 3 4 5 6 7 8 9 10
Frequency/Hz
 

(c)   
Figure 12. Vibration on bogie frame ends for Line A: (a) service distance less than 54,000 km; (b) the 
Figure 12. Vibrationservice distance more than 101,000 km;(c) frequency domain. 
on bogie frame ends for Line A: (a) service distance less than 54,000 km; (b) the
service distance more than 101,000 km;(c) frequency domain.
Line B, in contrast to Line A, exhibits a modest variation in the lateral acceleration as 
Line B, in contrast to Line A, exhibits a modest variation in the lateral acceleration as a
a result of the change in service distance. Figures 13 and 14 illustrate the time histories 
result of the change in service distance. Figures 13 and 14 illustrate the time histories and
and frequency spectra, respectively. The lateral vibration acceleration of the carbody has 
frequency spectra,no noticeable change at different operation distance, where its maximum value is less than 
respectively. The lateral vibration acceleration of the carbody has no
noticeable change 0.5 m/s
at
2. Moreover, the frequency domain characteristics of vibration are consistent, its 
different operation distance, where its maximum value is less12 than
Coatings 2022, 12, x FOR PEER REVIEW  of  15 
central frequency keeps 1.7 Hz. The lateral acceleration also does not change significantly 
  2
0.5 m/s . Moreover, the frequency domain characteristics of vibration are consistent, its
for the bogie frame, and the maximum value is less than 1 m/s2. 
central frequency keeps 1.7 Hz. The lateral acceleration also does not change significantly
 
for the bogie frame, and
 
the maximum value is less than 1 m/s2 .

   

(a)  (b) 
Figure 13. Vibration on carbody ends for Line B: (a) time‐domain and (b) frequency‐domain. 
Figure 13. Vibration on carbody ends for Line B: (a) time-domain and (b) frequency-domain.
    

(a)  (b) 
Coatings 2022, 12, 1333 11 of 14
Figure 13. Vibration on carbody ends for Line B: (a) time‐domain and (b) frequency‐domain. 

 
 

(a)  (b) 
Figure 14. Vibration on bogie frame ends for Line B: (a) time‐domain and (b) frequency‐domain. 
Figure 14. Vibration on bogie frame ends for Line B: (a) time-domain and (b) frequency-domain.

4.3. Rolling Contack Fatigue

4.3. Rolling Contack Fatigue 
The fatigue-related
The  damage
fatigue‐related  caused
damage  by the
caused  by recurrent actionaction 
the  recurrent  of a consequent creep force
of  a  consequent  creep 
inforce in addition to the ordinary contact forces occurring through the same region is re‐
addition to the ordinary contact forces occurring through the same region is referred to
asferred to as rolling contact fatigue (RCF). 
rolling contact fatigue (RCF).
Based on the “Code for Design of High-Speed Railway” [24] and considering the
Based on the “Code for Design of High‐Speed Railway” [24] and considering the ac‐
actual curve parameters, two curves with different radii are chosen here to investigate the
tual curve parameters, two curves with different radii are chosen here to investigate the 
rolling contact fatigue under wheel wear in the curved track. The curve parameters are
rolling contact fatigue under wheel wear in the curved track. The curve parameters are 
shown in Table 3.
shown in Table 3. 
Table 3. Parameters of the curves.
Table 3. Parameters of the curves. 
Radius/m Cant/mm Speed/(km/h) Rail Cant Deficiency/mm
Radius/m  Cant/mm  Speed/(km/h)  Rail Cant Deficiency/mm 
1000 110 110 32.8
1000  110  110  32.8 
600 125 90 34.3
600  125  90  34.3 

It is proposed in this paper that an engineering model established by Ekberg et al. [25]
It is proposed in this paper that an engineering model established by Ekberg et al. 
be used to evaluate the potential RCF problems. The surface fatigue index FIsurf is used to
[25] be used to evaluate the potential RCF problems. The surface fatigue index FIsurf is used 
anticipate the occurrence of surface fatigue flaws on a surface.
to anticipate the occurrence of surface fatigue flaws on a surface. 
2πabk
FIsur f = µ − (2)
3Fz
  where µ is the traction coefficient, a, and b are the semiaxes of the contact patch, Fz is the
normal force magnitude, and k is the yield stress in pure shear, which is 584 MPa in the
high-speed railway. The contact patch traction coefficient, is defined as the quotient of
the tangential and normal forces in the contact patchTraction coefficient is defined as the
symbol of tangential and normal forces:
q
Tx 2 + Ty 2
µ= (3)
Fz

where Fx , Fy is the longitudinal, lateral creep forces, respectively. If the inequality FIsur f > 0,
fatigue damage will occur.
The influence of RCF on vehicle dynamic performance is investigated here, where two
different curve radius 1000 m and 600 m are chosen here, and the track irregularity is not
considered in the calculation. The rolling contact fatigue index of the outer and inner rail for
the guide wheel is shown in Figure 15. As shown in Figure 15a, the FIsurf is positive in the
inner rail with a curve radius 1000 m, this may lead to the rolling contact fatigue cracks on
 two different curve radius 1000 m and 600 m are chosen here, and the track irregularity 
not considered in the calculation. The rolling contact fatigue index of the outer and inn
rail for the guide wheel is shown in Figure 15. As shown in Figure 15a, the FIsurf is positiv
in the inner rail with a curve radius 1000 m, this may lead to the rolling contact fatigu
cracks on wheel surface. This is caused by the fact that the lateral wheelset displaceme
Coatings 2022, 12, 1333 on the curves is about 9 mm, and two contact points occur on the outer rail, where one 
12 of 14
which occurs at the flange. Consequently, the FIsurf of the outer rail is larger. The norm
force and contact area diminish to a considerable amount as wheel wear progresses, ther
fore the surface fatigue index rises. In addition, Figure 15b illustrates that the FIsurf of th
wheel surface. This is caused by the fact that the lateral wheelset displacement on the curves
radius 1000 m is negative, while the FIsurf of radius 600 m is positive. This indicates th
is about 9 mm, and two contact points occur on the outer rail, where one of which occurs at
the possibility of the wheel forming rolling contact fatigue crack on the curve with radiu
the flange. Consequently, the FIsurf of the outer rail is larger. The normal force and contact
600 m is slightly greater than that of the radius 1000 m. Furthermore, the FIsurf progre
area diminish to a considerable amount as wheel wear progresses, therefore the surface
sively rises with wheel wear, especially the running distance is greater than 152,000 km
fatigue index rises. In surf
the FI addition, Figure 15b illustrates that the FIsurf of the radius 1000 m is
 changes from negative to positive, and the possibility of rolling contact fatigu
negative, while the FI surf of radius 600 m is positive. This indicates that the possibility of
crack increases under long‐term action. 
the wheel forming rolling contact fatigue crack on the curve with radius 600 m is slightly
It could speed up the growth and occurrence of RCF on both the up and down slop
of the radius 1000 m. Furthermore, the FIsurf progressively rises with
greater than thatbecause the lateral slip. RCF may also be influenced by the worse geometry match at th
wheel wear, especially the running distance is greater than 152,000 km, the FIsurf changes surf of th
hollow’s outer edges, which has a sharper contact geometry. Additionally, the FI
from negative to Line B is small, so the possibility of a fatigue crack is less. 
positive, and the possibility of rolling contact fatigue crack   increases under
long-term action.
0.3 0.2 0km
0km
57,000km
54,000km
113,000km
0.2 101,000km 0.1 152,000km
152,000km 218,000km
0.1 218,000km 0.0

FIsurf
FIsurf

0.0 -0.1

-0.1 -0.2

-0.2 -0.3
R1000 R600 R1000 R600 R1000 R600 R1000 R600
Inner rail Outer rail
Inner rail Outer rail
Radius [m]  
Radius[m]  

(a)  (b) 

Figure 15. Rolling Figure 15. Rolling contact fatigue index: (a) Line A; (b) Line B. 

contact fatigue index: (a) Line A; (b) Line B.

It could speed up In summary, it can be concluded that wheel‐rail profiles have great influence on v

the growth and occurrence of RCF on both the up and down slopes
hicle dynamic performance and rolling contact fatigue damage. Therefore, in order to o
because the lateral slip. RCF may also be influenced by the worse geometry match at the
tain ideal vehicle dynamics performance, the wheel‐rail profile can be optimized. 
hollow’s outer edges, which has a sharper contact geometry. Additionally, the FIsurf of the
Line B is small, so the possibility of a fatigue crack is less.
In summary, it can be concluded that wheel-rail profiles have great influence on vehicle
dynamic performance and rolling contact fatigue damage. Therefore, in order to obtain
 
ideal vehicle dynamics performance, the wheel-rail profile can be optimized.

5. Conclusions
This article illustrates the progression of wheel wear on high-speed trains at 250 km/h
on two railway lines. Experimental and numerical analysis were used to investigate the
impacts of wheel wear and its influence on vehicle dynamics.
(1) The initial wheel-rail profiles significantly influence the wheel wear pattern, the
thin rim wheel has uniform wear, and other wheels occurs hollow wear. With increasing
running distance, the equivalent conicity of the hollow wear wheel increases, while the
uniform wear wheel is stable in one re-profiling cycle.
(2) The hollow wear can lead to gradual deterioration of vehicle dynamics, which
conversely aggravates the wheel reprofiling. However, for the uniform wear of the wheel,
the vehicle dynamic behavior did not deteriorate significantly, particularly, the possibility
of a rolling contact fatigue crack is also tiny.
(3) The nonlinear critical speed for Line A presents a trend of continuous decrease
as the operation distance increases, while the Line B basically remains stable. The lateral
ride index increased to a certain extent, and the ride quality decreased from “excellent” to
“good” for the Line A, while the ride index has a slight change for the Line B.
(4) In the spectra of lateral bogie frame accelerations, at frequencies between 5 and
7 Hz, hollow wear increased the amplitudes significantly. Furthermore, it established that
this frequency range was consistent with bogie hunting, hence verifying the reason of the
anomaly warnings.
(5) At present, Chinese railway officials have selected a predetermined re-profiling pe-
riod of 220,000 km. Therefore, for the uniform wear of the wheel, under certain monitoring
Coatings 2022, 12, 1333 13 of 14

conditions to prevent the occurrence of low equivalent conicity, the re-profiling cycle of
high-speed wheels can be appropriately extended to prolong the fixed re-profiling interval
and reduce the cost of operation and maintenance for the high-speed train.

Author Contributions: M.H.: methodology, funding acquisition, writing—original draft. B.C.:

investigation, writing—review and editing. D.C.: methodology, software. All authors have read and
agreed to the published version of the manuscript.
Funding: This research was funded by the China Railway (Grant nos. J2021J007) and China Academy
of Railway Sciences (Grant nos. 2019YJ162).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Seo, J.-W.; Hur, H.-M.; Kwon, S.-J. Effect of Mechanical Properties of Rail and Wheel on Wear and Rolling Contact Fatigue. Metals
2022, 12, 630. [CrossRef]
2. Yang, X.; Tao, G.; Li, W.; Wen, Z. On the formation mechanism of high-order polygonal wear of metro train wheels: Experiment
and simulation. Eng. Fail. Anal. 2021, 127, 105512. [CrossRef]
3. Polach, O.; Nicklisch, D. Wheel/rail contact geometry parameters in regard to vehicle behaviour and their alteration with wear.
Wear 2016, 366–367, 200–208. [CrossRef]
4. Enblom, R. Deterioration mechanisms in the wheel–rail interface with focus on wear prediction: Aliterature review. Veh. Syst.
Dyn. 2009, 47, 661–700. [CrossRef]
5. Zhai, W.; Jin, X.; Wen, Z.; Zhao, X. Wear Problems of High-Speed Wheel Rail Systems Observations, Causes, and Counter-measures
in China. Appl. Mech. Rev. 2020, 72, 1–23. [CrossRef]
6. Shi, H.; Wang, J.; Wu, P.; Song, C.; Teng, W. Field measurements of the evolution of wheel wear and vehicle dynamics for
high-speed trains. Veh. Syst. Dyn. 2017, 56, 1187–1206. [CrossRef]
7. Zhu, A.; Yang, S.; Li, Q.; Yang, J.; Li, X.; Xie, Y. Simulation and measurement study of metro wheel wear based on the Archard
model. Ind. Lubr. Tribol. 2019, 71, 284–294. [CrossRef]
8. Kou, J.; Zhang, J.; Zhou, H.; Wang, C.; Sun, L. Effect of the worn status of wheel/rail profiles on wheel wear over curved tracks.
J. Mech. Sci. Technol. 2021, 35, 945–954. [CrossRef]
9. Wang, W.J.; Zhong, W.; Guo, J.; Liu, Q.Y.; Zhu, M.H.; Zhou, Z.R. Investigation on rolling contact fatigue and wear properties of
railway rails. Proc. Inst. Mech. Eng. Part J J. Eng. Tribol. 2009, 223, 1033–1039. [CrossRef]
10. Karttunen, K.; Kabo, E.; Ekberg, A. Numerical assessment of the influence of worn wheel tread geometry on rail and wheel
deterioration. Wear 2014, 317, 77–91. [CrossRef]
11. Ding, H.; He, C.; Ma, L.; Guo, J.; Liu, Q.; Wang, W. Wear mapping and transitions in wheel and rail materials under different
contact pressure and sliding velocity conditions. Wear 2016, 352–353, 1–8. [CrossRef]
12. Ostash, O.P.; Anofriev, V.H.; Andreiko, I.; Muradyan, L.A.; Kulyk, V. On the concept of selection of steels for high-strength
railroad wheels. Mater. Sci. 2013, 48, 697–703. [CrossRef]
13. Andreiko, I.; Kulyk, V.; Ostash, O.P. Resistance of steels of railroad wheels to corrosion-fatigue fracture. Mater. Sci. 2012, 47,
608–612. [CrossRef]
14. Qi, Y.; Dai, H.; Song, C.; Qu, S. Shaking analysis of high-speed train’s carbody when cross lines. J. Mech. Sci. Technol. 2019, 33,
1055–1064. [CrossRef]
15. Braghin, F.; Lewis, R.; Dwyer-Joyce, R.; Bruni, S. A mathematical model to predict railway wheel profile evolution due to wear.
Wear 2006, 261, 1253–1264. [CrossRef]
16. Dirks, B.; Enblom, R.; Berg, M. Prediction of wheel profile wear and crack growth—comparisons with measurements. Wear 2016,
366–367, 84–94. [CrossRef]
17. Tao, G.-Q.; Du, X.; Zhang, H.-J.; Wen, Z.-F.; Jin, X.-S.; Cui, D.-B. Development and validation of a model for predicting wheel wear
in high-speed trains. J. Zhejiang Univ.-Sci. A 2017, 18, 603–616. [CrossRef]
18. Wang, Z.; Wang, R.; Crosbee, D.; Allen, P.; Ye, Y.; Zhang, W. Wheel wear analysis of motor and unpowered car of a high-speed
train. Wear 2019, 444–445, 203136. [CrossRef]
19. Cui, D.; Li, L.; Wang, H.; Wen, Z.; Xiong, J. High-speed EMU wheel re-profiling threshold for complex wear forms from dynamics
viewpoint. Wear 2015, 338–339, 307–315. [CrossRef]
20. Wang, J.; Song, C.; Wu, P.; Dai, H. Wheel re-profiling interval optimization based on dynamic behavior evolution for high speed
trains. Wear 2016, 366–367, 316–324. [CrossRef]
Coatings 2022, 12, 1333 14 of 14

21. Gan, F.; Dai, H.; Gao, H.; Chi, M. Wheel–rail wear progression of high speed train with type S1002CN wheel treads. Wear 2015,
328–329, 569–581. [CrossRef]
22. UIC Code519; Method for Determining the Equivalent Conicity, 1st ed. International Union of Railways: France, Paris, December 2004.
23. GB/T 5599; Specification for Dynamic Performance Assessment and Testing Verification of Rolling Stock. Ministry of Railways of
the People’s Republic of China: Beijing, China, 2019.
24. TB 10621-2014; Code for Design of High Speed Railway. National Railway Administration of the People’s Republic of China:
Beijing, China, 2015.
25. Ekberg, A.; Kabo, E.; Andersson, H. An engineering model for prediction of rolling contact fatigue of railway wheels. Fatigue
Fract. Eng. Mater. Struct. 2002, 25, 899–909. [CrossRef]