Received November 28, 2019, accepted December 20, 2019, date of publication December 31, 2019,

date of current version January 9, 2020.

Digital Object Identifier 10.1109/ACCESS.2019.2963197

A Detailed Reliability Study of the Motor

System in Pure Electric Vans by the
Approach of Fault Tree Analysis
XIONG SHU 1,3 , YINGFU GUO 1 , WENXIAN YANG 2, KEXIANG WEI 3,

YUN ZHU 3 , AND HONGXIANG ZOU 3

1 School of Mechanical Engineering, Hunan University of Science and Technology, Xiangtan 411201, China
2 School of Engineering, Newcastle University, Newcastle upon Tyne NE1 7RU, U.K.
3 Hunan Provincial Key Laboratory of Vehicle Power and Transmission System, Hunan Institute of Engineering, Xiangtan 411104, China
Corresponding author: Kexiang Wei (kxwei@hnie.edu.cn)
This work was supported in part by the Hunan Province Science and Technology Innovation Program under Grant 2017XK2303,
Grant 2019RS2044, and Grant 2019JJ50099, and in part by the Research Foundation of Education Department of Hunan
Province, China, under Grant 18B389.

ABSTRACT Attributing to the unique feature of zero carbon emission, electric vehicles (EVs) are attracting
increasing interest in recent years, but their reliability, particularly the reliability of their critical components,
is still a matter of concern today. In order to address this issue, much effort has been made before to
assess the reliability of drive motor in the EVs. However, drive motor and motor controller are logically
integrated and requested to work as one system in the EVs. In contrast to the individual reliability analysis
of them, the combined assessment of the two parts can provide a more reliable prediction to the reliability
of the entire motor system. Moreover, both drive motor and motor controller are composed of multiple
components. The structure, type, and characteristics of these components may affect the reliability of the
motor system as well. But these issues have not been considered in the previous research. In order to fill
this gap of knowledge, the reliability of the entire motor system of pure electric vans that includes both
drive motor and motor controller is investigated in this paper. In the research, the theoretical failure rates of
subassemblies and components in drive motor and motor controller are predicted first. Then based on the
failure rate prediction results, the reliability of the entire motor system (comprising both drive motor and
motor controller) is assessed. Based on the assessment results, some interesting conclusions with respect to
the most vulnerable subassemblies and components in the entire motor system and the potential disadvantage
of existing reliability research are finally obtained. It is deemed that these new findings will be of great
significance to the future reliability design and maintenance of pure electric vans.

INDEX TERMS Reliability, electric vehicles, motor system, motor controller, fault tree analysis.
I. INTRODUCTION diesel and petrol vehicles at some point in the near future. For
Encouraged by financial subsidies and policy dividends, example, it would be 2040 in the UK. This means that there
the market of electric vehicles is booming across the world will be many pure electric vehicles running on the road in the
today. The recent survey has shown that China, the UK, following years. That is why their reliability in the application
the United States, and Germany currently top the world is receiving more concern than ever before.
ranking of the sales of pure electric vehicles [1]. Take the Broadly speaking, pure electric vehicles can be classified
Chinese market as an example, total 984,000 pure electric into two categories, i.e. passenger cars and commercial vehi-
vehicles were sold in China in 2018, which was an increase cles. Their market share splits are about 80% versus 20%. For
of 50.8% over the same period of the previous year [1]. It is example, 788,000 passenger cars and 196,000 commercial
deemed that pure electric vehicles will replace traditional vehicles were sold in China in 2018 [1]. In terms of relia-
bility, the practice has shown that the latter is more prone to
The associate editor coordinating the review of this manuscript and develop failure in the application, although such a comment
approving it for publication was Lorenzo Ciani . has not yet been officially reported before in open literature.

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see http://creativecommons.org/licenses/by/4.0/
VOLUME 8, 2020 5295
 X. Shu et al.: Detailed Reliability Study of the Motor System in Pure Electric Vans by the Approach of Fault Tree Analysis

For this reason, the research in this paper will be focused on to fill these gaps of knowledge by looking into the reliability
investigating the reliability of commercial vehicles. Herein, issues of all subassemblies and components in both the drive
considering more than 90% of sold commercial vehicles are motor and motor controller. The outcome of this research
pure electric vans [1] and the motor system is the most key should be a good complement to the current knowledge of
part in these vehicles, the reliability of the motor system in electric vehicles and it will also bring benefits to the emerging
pure electric vans will be studied in the following. electric vehicle industry.
The motor system in pure electric vans is used to con- Fault tree analysis, an effective methodology for conduct-
vert the electric energy into mechanical energy to drive the ing reliability and safety assessment, has been extensively
vehicle. Any reliability issue existing in the motor system used to investigate the reliability issues in various systems.
may lead to dangerous accidents on the road. Therefore, For example, the potential failures of wind turbine were qual-
it is essential to investigate its reliability issues, identify itatively and quantitatively evaluated by using the fault tree
the vulnerable parts in the system, and then improve its analysis method in [14], and it was found that the majority
design based on the investigation results [2], [3]. In order of floating turbine failures are due to marine conditions, salt-
to meet this purpose, much effort has been made before. spray, and high wind speed; the fault tree analysis method
A detailed literature review has shown that so far, the majority was further improved in [15] in order to adapt to performing
of published works in the relevant area are focused on how the qualitative analysis of complex systems containing multi-
to improve the control and fault tolerance performance of component systems; the dynamic fault tree analysis model
motor systems. For example, the fault tolerance performance was developed in [16] to determine the average maintenance
of multi-phase permanent magnet synchronous motor was period of a floating wind turbine; the dynamic fault tree
investigated in [4] with the aid of a multi-level Markov model; models were developed in [6] and [17] for assessing the relia-
the similar research on motor drives was also conducted bility of a fault-tolerant control system and a vehicle guidance
in [5] by using the Monte Carlo method in order to get bet- system in unmanned aerial vehicles; and so on. In view of
ter fault-tolerance mechanisms. The reliability of automated these previously successful applications of fault tree analysis,
guided vehicles’ (AVGs) control system was studied in [6] it will be taken in this paper to investigate the reliability of the
by using a combined fault tree and Petri net approach; the motor system of pure electric vans. The remaining parts of the
research was further improved by the authors in [7] through paper are organized as below:
introducing genetic algorithm method into the maintenance In Section 2, the motor system of pure electric vans will
strategy optimization. Fault classification method and fault- be briefly introduced so as the reader can have a basic under-
tolerant control strategies were adopted in [8] in order to standing of it. In Section 3, the fault tree of the drive motor
improve the dynamic performance and control stability of and motor controller will be developed and following which
electric vehicles. The reliability of drive motor and electronic the mathematical methods for predicting the failure rates of
converter for power supply was also studied by scholars. For basic events will be developed. To ease understanding, a case
example, the reliability of the drive motor of EVs was studied study is conducted in Section 4, in which the failure rate,
in [9], [10] by using a combined fault tree and Petri net reliability of the subassemblies in the motor system of the
approach, and the fault logical causes by the winding insu- pure electric vans will be calculated and analyzed. The paper
lation and bearing with higher fault rate were investigated will be finally concluded in Section 5.
in [11] by the approach of fault tree analysis. The reliability of
bidirectional DC/DC converters of EVs was studied in [12]; II. THE MOTOR SYSTEM IN PURE ELECTRIC VANS
the age and life of power converter components were esti- As shown in Fig.1, the motor system of pure electric vans is
mated in [13] based on a survey, and so on. However, it is a gearless system. It is simply composed of a drive motor and
noticed that in previous research, the reliability study was an associated motor controller. The drive motor will be used
mainly focused on the analysis of drive motor, and the reli- to drive the wheels directly via rear axle. The motor controller
ability issues in the motor controller were rarely discussed. is used to control the torque, speed and rotation direction of
This may lead to unreliable research conclusions due to the drive motor according to the vehicle driving demand. Fig.1,
fact that the drive motor and motor controller are logically the black and red dot lines are the Control Area Network
integrated. They are always requested to work as one system (CAN) lines, it connecting the motor controller, Vehicle Con-
and cannot be dealt with individually. A combined reliability trol Unit (VCU) and the units that are designed to implement
assessment of the two parts will provide a more reliable different control functions.
prediction of the reliability of the entire motor system. How- From Fig.1, the drive motor is the core component that is
ever, such detailed reliability research of the entire motor used to convert electrical energy into mechanical energy [18].
system has not been conducted before. This motivates the In order to ease understanding, the structure of the drive motor
research of this paper. Moreover, both drive motor and motor is shown in Fig.2.
control are composed of multiple components. The structure, As shown in Fig.2, the drive motor comprises a shell, stator,
type, and characteristics of components may also affect the rotor, bearing, and position sensor. As soon as electric power
reliability of the entire motor system. However, this has not is applied to the stator, a rotating magnetic field will be gen-
been considered before either. The purpose of this research is erated. Then, the rotor will be driven to rotate in the magnetic

5296 VOLUME 8, 2020

X. Shu et al.: Detailed Reliability Study of the Motor System in Pure Electric Vans by the Approach of Fault Tree Analysis

FIGURE 1. Schematic diagram of the motor system in pure electric vans.

FIGURE 3. Schematic diagram of the motor controller. 1 - phase lines,

2 - power line, 3 – wiring slot, 4 - copper row, 5 - discharging module,
6 – housing shell, 7 - communication module, 8 - master chip,
9 - thin-film capacitor, 10 - control module, 11 - current transducer,
12 - driver module, 13 - IGBT, 14 - seal ring of shell.

FIGURE 2. The structure of the drive motor in pure electric vans. 1 - rotor,
2 - rotor end cover, 3 - bearing, 4 - spline, 5 - fastening screw of rotor
balance end cap, 6 - inlet of cooling water, 7 - jacket of cooling water,
8 - stator winding, 9 - core of stator, 10 - sealing rings of bearing, 11 - end
cap fastening screw, 12 - junction box, 13 - seal ring, 14 - cable fixed
head, 15 - fastening screw of rear end cap, 16 - position sensor,
17 - grease seal, 18 - three-phase lines, 19 - position sensor.

field due to the electromagnetic induction effect. Simulta-

FIGURE 4. Fault tree and reliability block diagram of the motor system.
neously, the relative position of the stator and rotor will be
measured by the position sensor. Then the measured position
will be input into the motor controller to ensure desired three- chip based on the position signal measured by the position
phase voltages are output from the motor controller. During sensor.
the operation of the motor system, the temperature of the
drive motor is detected by the temperature sensor (such as III. RELIABILITY STUDY OF MOTOR SYSTEM
PT1000 shown in Table 6) in real-time. When the sensor fails, Since fault tree analysis has been successfully applied to
its resistance value will far exceed normal value and even will investigate the reliability of numerous systems [20], it will
be recorded as infinity. In this case, the automatic detection be taken to investigate the reliability of the motor system.
program or system protection program of the motor controller Since the motor system is composed of a drive motor and
will assume that the drive motor has been overheated, thereby motor controller, its fault tree model and reliability block
shutting down the system automatically. diagram is shown in Fig.4. Accordingly, the reliability study
From Fig.3, it is seen that the motor controller is mainly in this section will comprise two parts. The first part is for
composed of outer casing, thin-film capacitor, copper row, investigating the reliability of the drive motor, and the second
Insulated Gate Bipolar Translator (IGBT), driver module, is for investigating the reliability of the motor controller.
control module, active discharging module, communication
module, sealing ring, and cooling water jacket. Among them, A. RELIABILITY STUDY OF DRIVE MOTOR
the module refers to the equipped printed circuit board, which As mentioned earlier, the drive motor is composed of bear-
consists of connections, components, and board. For exam- ings, rotor, stator, sensors, and other associated components.
ple, the control module consists of a control chip, resistances, Its safety and efficiency can be affected once a fault occurs
capacitances, connections, and wire board. In the operation, in any of these components. The survey has shown that the
an IGBT gate control signal will be output from the control failures of oil seal, bearing, stators, and rotor winding account

VOLUME 8, 2020 5297

 X. Shu et al.: Detailed Reliability Study of the Motor System in Pure Electric Vans by the Approach of Fault Tree Analysis

FIGURE 5. Reliability block diagram of the drive motor.

TABLE 1. Failure events of drive motor components.

FIGURE 6. Fault tree of the drive motor.

handbook of the US Navy’s mechanical equipment reliability

prediction program [22], the failure rate of the basic events
in Fig.6 can be estimated by



 λ1,3 = λWI ,B CV 1 CT 1 Calt
λ2 = λSH ,B Cf 2 CT 2 CDY CSC




λ 4 = λM ,B CM





λ5 = λTD,B + λS5 + λT 5 + λp5 + λX 6





λ6 = λTD,B + λS6 + λT 6 + λp6 + λX 6


(2)


 λ7 = λG,B CGS CGP CGA CGL CGT CGV
λ8 = λSE,B CQ CH CF8 CV 8 CT 8 CN CPV




λ 9 = λF,B CSZ CL CT 9 CI CSC9 CK

for more than 80% of the total number of failures of the drive




motor [21], herein the reliability study of the drive motor λ10 = λBE,B CV 10 CCW CT 10 CSF




will be focused on the study of these vulnerable components

λBE,B = 2 × 105 / (Ls /LA )y


which show high failure rate. Based on this understanding,
the reliability block diagram of the drive motor is developed. All parameters in these equations are explained in Table 2 and
It is shown in Fig.5. the failure rate unit of each component is FPMH (failures per
For a reliability model shown in Fig.5, its reliability can be million hours).
evaluated using the following equation. From (2), it is seen that the failure rate of the components
Y4 is dependent on their working environments, loads, and many
Rm (S1 ) = P (mx1 ) P (mx2 ) . . . P (mx4 ) = P(mxi ) (1) other factors. Take the failure of the bearing as an example,
i=1
where P(mxi ) refers to the reliability probability function of its basic failure rate is
the i-th component in Fig.5. λBE,B = 2 × 105 /L10 (3)
In order to investigate the reliability of the components in
the drive motor, the fault tree of the drive motor is developed, where, L10 = (Ls /LA )y , Ls is basic dynamic load, LA is
see Fig.6. In the figure, the top event S1 is ‘drive motor equivalent radial load, and Y is a bearing life constant
failure’, and the basic events from em1 to em10 are the failure (Y = 3.0 for ball bearings, and 3.3 for roller bearings), and
events of drive motor components. In this model, gm1 to gm4 L10 can be also converted to hours by using the equation
are intermediate events (or logic gate events), which are the
logical combination of the relevant basic events. All these 106
L10h = L10 (4)
events are explained in Table 1. 60n
From Fig.6, it is found that the reliability of the drive motor where, L10h is the bearing life with operating hours, n is the
relies on the reliability of bearing, motor rotor, motor stator, rotational speed in the unit of revolution per minute.
transducers, and other associated components. In order to From the Fig.6, the reliability of the rotor relies on the
understand the reliability of the drive motor, it is essential reliability of its winding and shaft. Therefore, the failure rate
to evaluate the failure rates of these components. As long as of the rotor is the sum of the failure rates of its winding and
the failure rates of these components and their corresponding shaft, i.e.
reliability probability functions are obtained, the reliability
of the drive motor can be readily assessed. According to the λgm1 = λm1 + λm2 (5)

5298 VOLUME 8, 2020

X. Shu et al.: Detailed Reliability Study of the Motor System in Pure Electric Vans by the Approach of Fault Tree Analysis

TABLE 2. Parameters in equation (2).

Likewise, the failure rate of the stator, transducers and motor controller can be described by using the diagram in
the other components in the drive motor can be respectively Fig.7.
expressed as Based on Fig.7, the reliability of the motor controller can
 be predicted by [27]
 λgm2 = λm3 + λm4
 Y5
λgm3 = λm5 + λm6 (6) Rc (S2 ) = P (cx1 ) P (cx2 ) . . . P (cx5 ) = P(cxi ) (7)
i=1
λgm4 = λm7 + λm8 + λm9 + λm10


where P(cxi ) indicates the reliability probability function of
the i-th component in Fig.7.
The physical means of all variables in (5) and (6) can be Define the failure of the motor controller as the top event
found in Table 1. (S2) and the failures of these components as the basic events
ec1 to ec14, the fault tree of the motor controller can be
B. RELIABILITY STUDY OF MOTOR CONTROLLER constructed. It is shown in Fig.8. Where, the intermediate
As shown in Fig.3, the motor controller is composed of events are gc1 to gc4, which are the logical combination
power electronic components and a housing shell. Due to the of the relevant basic events. All these events are explained
housing shell is reliable and has little chance to affect the in Table 3.
reliability of the motor controller, it will not be considered Herein, it is worth noting that the capacitor mentioned
in the following study. The motor controller consists of bus- in Table 3 is used to balance the power during the change
bar capacitors, control module, driver module, discharging of power frequency and the power output from the rectifier.
module, communication module, and IGBT. Since a fault On the one hand, it will absorb the energy produced by the
occurring in any of these components can lead to the failure drive motor when power switching devices are switched off
of the motor controller, the reliability block diagram of the in the scenario of an urgent stop. It will provide instantaneous

VOLUME 8, 2020 5299

 X. Shu et al.: Detailed Reliability Study of the Motor System in Pure Electric Vans by the Approach of Fault Tree Analysis

TABLE 3. Failure events of motor controller components. equation [24]

" Py
πtc13 i τi

#
i=1
λc13 = 0.1 + 1.4 × 10−3
τon + τoff
X 
j
× (πn )i (1Ti )0.68
× 10−9 (8)
i=1

where (πt )i is the i-th temperature factor related to the i-

th junction temperature
 of the capacitor mission profile;
1 1
2900 303 − 273+t
πtc13 = e

Ac13
is a factor reflecting the influ-
ence of ambient temperature tAc13 ; τon is the total working
time ratio of the capacitor; τoff is time ration for the capacitor
being in storage (or dormant) mode; (πn )i is the i-th influence
factor related to the annual cycles number of thermal varia-
tion with the amplitude 1Ti ; 1Ti refers to the i-th thermal
amplitude variation of the mission profile.
The control module in Table 3 comprises a master chip,
a printed circuit board (PCB), and some surface-mounted
components (SMCs). The SMCs include inductors, resistors,
capacitors, transformers, integrated chips, diodes, etc. So,
the reliability of the control module relies on the reliability
of the master chip, PCB and SMCs. The failure rate of these
components can be evaluated by [23], [24], [26]
 " r √ #
Nt 1 + 0.1 S
 λc1 = 5 × 10 πtc1 πc Nt 1 +
−3
πL

 + Np
S 3





   Xj 
−3
(πn ) (1Ti )0.68 × 10−9


 × 1+3 × 10
i=1



 X
λc2 = λs × 10−9





 X  Xj 
λf + 1+3×10 −3 0.68

+ (πn )i(1Ti )


i=1
X



 ×10−9 × λd
λ (C π π
i2 iE ) πiQ

= + C

FIGURE 7. Reliability block diagram of the motor controller. 
 ci i1 it

λcc = λcb πcE πcQ πcT πcS πcch





λcg = λgb πgE πgQ πgT





λ = λrb πrE πrQ πrR


 cr


λcm = λmb πmE πmQ πmr πmA πmC πmK

(9)
 
1 1
1740 303 − 273+t
where πtc1 = e Ac1
is a factor indicating the
influence of ambient temperature tAc1 ; πC indicates the influ-
ence of the number of layers; Nt is the total number of holes;
S is the surface area of the board; NPP is P the number of
Ns + Nf
tracks. Its default value is NP = 2 , in which
Ns is the number of lines that connect to every com-
ponent, Nf is the number of components that connect
FIGURE 8. Fault tree of motor controller. to every hole; πL is the coefficient of line track width,
1Ti is the variation of temperature; λd is the interconnection
failure rate. Other parameters in (8) are further explained
in Table 4.
peak power as required so as the inverter can be well protected
Then, the failure rate of the control module in the motor
during the sudden change of power. At present, the capac-
controller can be evaluated by [27].
itors used in the motor controllers of pure electric vans are
film capacitors. Its failure rate can be evaluated by using λgc1 = λc1 + λc2 + λc3 (10)

5300 VOLUME 8, 2020

X. Shu et al.: Detailed Reliability Study of the Motor System in Pure Electric Vans by the Approach of Fault Tree Analysis

TABLE 4. Other parameters in equation (8). TABLE 6. Parts of the drive motor.

TABLE 5. Performance parameters of the motor system being

investigated.

FIGURE 9. The drive motor being investigated.

other components in the motor controller, respectively (show

in Table.3).

IV. CASE STUDY

As the hardware components of the driver module, dis-
A case study is conducted in this section in order to quan-
charging module, and communication module of the motor
titatively assess the reliability of the motor system in pure
controller are similar to the control module, their failure rate
electric vans by using the mathematical methods described in
can be evaluated in the same way, i.e.
section 3. The performance parameters of the motor system
are listed in Table 5.
λgc2 = λc4 + λc5 + λc6



λ

= λc7 + λc8 + λc9
gc3 A. EVALUATION OF FAILURE RATES OF DRIVE MOTOR
(11)


 λgc4 = λc10 + λc11 + λc12 PARTS
λgc5 = λc13 + λc14 To ease understanding, the drive motor of interest is shown in

Fig.9 and its parts are listed in Table 6.
where λgc2 − λgc5 refer to the failure rate of the driver As mentioned earlier, the failure of the drive motor is
module, discharging module, communication module and the mainly caused by the faults developed in bearing, rotor,

VOLUME 8, 2020 5301

 X. Shu et al.: Detailed Reliability Study of the Motor System in Pure Electric Vans by the Approach of Fault Tree Analysis

TABLE 7. Parameters for calculating the failure rates of drive motor components.

stator, and other associated components. Therefore, accord- motor components can be readily obtained. They are listed
ing to the Handbook of Reliability Prediction Procedures for in Table 7.
Mechanical Equipment NSWC-09 [22] and Chinese elec- Then, calculate the failure rate of every component in the
tric vehicle technical requirements standards [25], [26], all fault tree model of the drive motor by substituting these
parameters used for calculating the failure rate of these drive parameters into (2)-(6). The obtained calculation results

5302 VOLUME 8, 2020

X. Shu et al.: Detailed Reliability Study of the Motor System in Pure Electric Vans by the Approach of Fault Tree Analysis

FIGURE 10. Calculation results of the failure rate of drive motor components.

TABLE 8. Calculation results of the failure rates of drive motor

components.

TABLE 9. Physical means of the parameters.

FIGURE 11. Parts of the motor controller.

driver module, discharging module, communication module,

and IGBT. All these parts are illustrated in Fig.11.
From the design documents and Bill of Material (BOM)
provided by the motor systems manufacturer, it is known that
the model of busbar capacitor is C362H557K19802 and the
model of IGBT is FS400R07A3E3. As there are many SMCs
are listed in Table 8. In order to facilitate interpretation,
on the PCB and they have a variety of specifications, it is
the results are further illustrated in Fig.10.
not realistic to evaluate the failure rate of every component
From the calculation results shown in Table 8 and Fig.10, it
one by one. To deal with this issue, the SMCs in BOM list
is found that the oil seal of the bearing is the most vulnerable
are classified first according to their types, such as resistance,
part in the drive motor, followed by the position sensor and
inductance, capacitance, etc. Then, the failure rate of the same
rotor/stator windings. The temperature sensor is also prone to
type of SMCs is calculated by using the following counting
develop failure in operation. By contrast, the motor bearing,
method.
the spline, and the shaft are relatively more reliable. The fas- 
tening screw and stator core are the most reliable components 

 λtci = ni λci
in the drive motor, they are nearly free of fault.  λtcc = nc λcc



λtcg = ng λcg (12)
B. EVALUATION OF FAILURE RATES OF MOTOR 

 λtcr = nr λcr

CONTROLLER PARTS 
λ = n λ

tcm m cm
The failure of the motor controller is primarily caused by
the faults occurring in the busbar capacitor, control module, All variables in the above equations are explained in Table 9.

VOLUME 8, 2020 5303

 X. Shu et al.: Detailed Reliability Study of the Motor System in Pure Electric Vans by the Approach of Fault Tree Analysis

TABLE 10. Parameters used for calculating the failure rates of SMCs.

FIGURE 12. Calculation results of the failure rates of different motor controller parts.

According to IEC TR62308-2004, MIL-HDBH-217E and in Table 12. In parallel, the calculation results are also illus-
the technical standards for the Chinese electric vehicle indus- trated in Fig.12 so as the reader can have an intuitive under-
try [23], [24], [26], the failure rates of all SMCs are calcu- standing of the reliability of the different parts of the motor
lated. The parameters used for calculating the failure rate are controller.
listed in Table 10, and failure rate calculation results are listed From Fig.12, it is found that the PCB of the control module
in Table 11. is the most vulnerable part in the motor controller (i.e. its
Based on the data listed in Table 11, the failure rate of failure rate is as high as 3.058), followed by the PCB of
motor controller parts is calculated by substituting these the drive board module (i.e. its failure rate is 1.116). The
parameters into (8-12). The calculation results are listed master chip and communication chip, as well as optocoupler,

5304 VOLUME 8, 2020

X. Shu et al.: Detailed Reliability Study of the Motor System in Pure Electric Vans by the Approach of Fault Tree Analysis

TABLE 11. SMCs on boards and their failure rate calculation results. TABLE 12. Failure rates of motor controller parts.

The comparison of Table 8 and Table 12 has shown that the

overall reliability of the motor controller is lower than that
of the drive motor. In the global view of the motor system,
its control module is most vulnerable, the failure rate of
which is as high as 3.44. The second is the driver module.
Its failure rate is 1.402. The motor rotor, stator, transducers,
and communication module have similar failure rates. They
are 0.2998, 0.2523, 0.2580, and 0.2830, respectively. The
understanding of such a figure is of great significance to the
reliability design and maintenance of pure electric vans.

C. UNRELIABILITY INDEX
Although the calculation results presented above have
indicated the reliability of individual components, the infor-
mation provided by them remains the ‘snapshots’ of the reli-
ability of the motor system. For this reason, the unreliability
indices of the drive motor and motor controller as well as
the entire motor system are evaluated based on the above
calculation results in order to obtain a more comprehensive
understanding of the reliability of the motor system. Herein,
the unreliability indices of the drive motor and its controller
are estimated with the aid of Eqns. (1) and (7). The equations
for calculating the unreliability indices are

 Fm (t) = 1 − Rm (S1 )

Fc (t) = 1 − RC (S2 ) (13)
Fmc (t) = 1 − (1 − F m (t))(1 − F c (t))



where Fm , Fc and Fmc refer to the unreliability indices of the

drive motor, motor controller, and the whole motor system,
respectively; Rm , Rc refer to the reliability indices of the drive
motor, motor controller, respectively.
The corresponding estimation results when the service life
are also prone to develop a fault in operation. But they are of the motor system reaches respectively 10,000, 30,000,
relatively more reliable than the PCB of the control and driver 50,000 and 70,000 hours are shown in Fig.13.
module. Their failure rate varies in the range of 0.2-0.35. From Fig.13, it is clearly seen that
In comparison, the IGBT and SMCs, as well as the PCB of the (1) the motor controller is much less reliable than the drive
communication module and discharging module, have little motor. This agrees very well with the research conclu-
chance to develop a fault in operation as their failure rate is sion obtained from the failure rate calculation results
below 0.1. in Section 4.2;

VOLUME 8, 2020 5305

 X. Shu et al.: Detailed Reliability Study of the Motor System in Pure Electric Vans by the Approach of Fault Tree Analysis

because the drive motor and motor controller are logi-

cally integrated and are always requested to work as one
system in the EVs. The reliability of the entire motor
system will be less estimated if only the reliability issues
in the drive motor are considered in the assessment.

ACKNOWLEDGMENT
The authors gratefully acknowledge Prof. B. Pailthorpe and
Dr. N. Bordes for useful discussions and consultations.

REFERENCES
[1] China Passenger Car Association. Electric Vehicle Sales Analysis for the
Global Market from 2018 to 2019. Accessed: Feb. 30, 2019. [Online].
FIGURE 13. Calculation results of the unreliability index of the drive
Available: http://www.cpcaauto.com/newslist.asp?types=csjd&id=9435
motor and motor controller.
[2] F. P. Santos, P. Teixeira, and C. G. Soares, ‘‘Operation and maintenance of
floating offshore wind turbines,’’ in Floating Offshore Wind Farms. Cham,
(2) both drive motor and motor controller will become more Switzerland: Springer, Mar. 2016, pp. 181–193, doi: 10.1007/978-3-319-
27972-5_10.
and more unreliable with the increase of their service [3] F. H. Gandoman, A. Ahmadi, P. V. D. Bossche, J. Van Mierlo, N. Omar,
time, i.e. the longer the service time, the higher their A. E. Nezhad, H. Mavalizadeh, and C. Mayet, ‘‘Status and future perspec-
unreliability indices will tend to be; tives of reliability assessment for electric vehicles,’’ Rel. Eng. Syst. Saf.,
vol. 183, pp. 1–16, Mar. 2019, doi: 10.1016/j.ress.2018.11.013.
(3) the most important is that despite the service time, the cal- [4] I. Bolvashenkov, J. Kammermann, and H. G. Herzog, ‘‘Research on relia-
culation result of the unreliability index of the entire bility and fault tolerance of multi-phase traction electric motors based on
motor system is much higher than the corresponding Markov models for multi-state systems,’’ in Proc. SPEEDAM, Aug. 2016,
pp. 1166–1171, doi: 10.1109/SPEEDAM.2016.7525928.
value of the unreliability index of the drive motor. This [5] M. S. Blanco, ‘‘The economics of wind energy,’’ Renew. Sustain. Energy,
suggests that the entire motor system, in fact, has less vol. 13, pp. 387–401, Aug./Sep. 2009, doi: 10.1016/j.rser.2008.09.004.
reliability than drive motor. [6] R. Yan, L. M. Jackson, and S. J. Dunnett, ‘‘Automated guided vehicle
mission reliability modelling using a combined fault tree and Petri net
approach,’’ Int. J. Adv. Manuf. Technol., vol. 92, no. 5, pp. 1825–1837,
V. CONCLUSION Mar. 2017, doi: 10.1007/s00170-017-0175-7.
In order to provide a more reliable prediction to the entire [7] R. D. Yan, S. J. Dunnett, and L. M. Jackson, ‘‘Novel methodol-
ogy for optimizing the design, operation and maintenance of a multi-
motor system in pure electric vans, a detailed study of the AGV system,’’ Rel. Eng. Syst. Saf., vol. 178, pp. 130–139, Oct. 2018,
reliability issues in both drive motor and motor controller doi: 10.1016/j.ress.2018.06.003.
are investigated in this paper by the approach of fault tree [8] D. Wanner, ‘‘Faults and their influence on the dynamic behavior of elec-
tric vehicles,’’ Ph.D. dissertation, Dept. Vehicle Eng., Roy. Inst. Technol.
analysis. From the research reported above, the following Vehicle Dyn., Teknikringen, Stockholm, Sweden, 2013.
conclusions can be drawn [9] B. Wang, G. Tian, Y. Liang, and T. Qiang, ‘‘Reliability modeling and
evaluation of electric vehicle motor by using fault tree and extended
• Due to aging, the reliability of both drive motor stochastic Petri nets,’’ J. Appl. Math., vol. 2014, pp. 1–9, Apr. 2014, doi:
and motor controller will decrease gradually with the 10.1155/2014/638013.
increase of their service time; [10] B. Wang, Y. Liang, C. Yang, and Z. Sang, ‘‘Reliability modeling and
assessment of electric vehicle motor using fault tree and fuzzy Petri nets,’’
• Despite the service time, in the entire motor system, Int. J. Grid Distrib. Comput., vol. 9, no. 8, pp. 121–136, Aug. 2016,
the motor controller is more unreliable than drive motor. doi: 10.14257/ijgdc.2016.9.8.12.
For example, when the motor system continuously [11] X.-H. Zhu, S.-M. Gui, and N. Shi, ‘‘Grey prediction model of motor relia-
bility of electric vehicle,’’ Electr. Mach. Control, vol. 16, no. 8, pp. 41–42,
runs for 30,000 hours, the unreliability of drive motor Aug. 2012.
is 0.033, but the unreliability of motor controller has [12] M. Khalilzadeh and A. Fereidunian, ‘‘A Markovian approach applied to
reached as high as 0.147, which is about 4.5 times of reliability modeling of bidirectional DC-DC converters used in PHEVs and
smart grids,’’ Iranian J. Elect. Electron. Eng., vol. 4, no. 12, pp. 13–301,
the unreliability of the former; Dec. 2016, doi: 10.22068/IJEEE.12.4.301.
• The control module is not only the most vulnerable part [13] Y. Song and B. Wang, ‘‘Survey on reliability of power electronic systems,’’
of the motor controller but also the most vulnerable part IEEE Trans. Power Electron., vol. 28, no. 1, pp. 591–604, Jan. 2013,
doi: 10.1109/tpel.2012.2192503.
of the entire motor system. This is due to the application [14] J. Kang, L. Sun, and C. G. Soares, ‘‘Fault tree analysis of floating offshore
of a large number of electronic components (the total wind turbines,’’ Renew. Energy, vol. 133, pp. 1455–1467, Apr. 2019, doi:
number of electronic devices in the control module is 10.1016/j.renene.2018.08.097.
[15] T. P. KhanhNguyen, J. Beugin, and J. Marais, ‘‘Method for evaluating
up to 610 PCS), and more electronic devices will lead an extended fault tree to analyses the dependability of complex systems:
to more connecting lines, thinner line widths and more Application to a satellite-based railway system,’’ Rel. Eng. Syst. Saf.,
through holes on the PCB, which will reduce the relia- vo133, pp. 300–313, Jan. 2014, doi: 10.1016/j.ress.2014.09.019.
[16] X. Zhang, L. Sun, H. Sun, Q. Guo, and X. Bai, ‘‘Floating offshore
bility of the control module; wind turbine reliability analysis based on system grading and dynamic
• The reliability issues in the drive motor and motor con- FTA,’’ J. Wind Eng. Ind. Aerodyn., vol. 154, pp. 21–33, Jul. 2016,
troller should be investigated together when assessing doi: 10.1016/j.jweia.2016.04.005.
[17] B. Hu and P. Seiler, ‘‘Pivotal decomposition for reliability analysis of fault
the reliability of the motor system. Otherwise, an unre- tolerant control systems on unmanned aerial vehicles,’’ Rel. Eng. Syst. Saf.,
liable reliability prediction may be obtained. This is vol. 140, pp. 130–141, Aug. 2015, doi: 10.1016/j.ress.2015.04.005.

5306 VOLUME 8, 2020

X. Shu et al.: Detailed Reliability Study of the Motor System in Pure Electric Vans by the Approach of Fault Tree Analysis

[18] M. Hernandez, M. Messagie, M. De Gennaro, and J. Van Mierlo, WENXIAN YANG received the Ph.D. degree in
‘‘Resource depletion in an electric vehicle powertrain using different LCA mechanical engineering from Xi’an Jiaotong Uni-
impact methods,’’ Resour., Conservation Recycling, vol. 120, pp. 119–130, versity, in 1999. Then, he did a research with
May 2017, doi: 10.1016/j.resconrec.2016.11.005. the City University of Hong Kong, from 2001 to
[19] E. Dabla, C. Martin, F. Pérès, F. Andrianoelison, and M. Piton, ‘‘Contribu- 2002, Nottingham Trent University, from 2002 to
tion of Bayesian network to the prediction of reliability of IGBT module 2004, Cranfield University, from 2004 to 2007,
used in railway applications,’’ in Proc. 16th IFAC Symp. Inf. Control Prob- the University of Durham, from 2007 to 2009, and
lems Manuf., Bergamo, Italy, Jan. 2018, doi: 10.1016/j.ifacol.2018.08.265.
the U.K. Offshore Renewable Energy CATAPULT
[20] S. Kabir, ‘‘An overview of fault tree analysis and its application in model-
Centre, from 2009 to 2013. In May 2013, he joined
based dependability analysis,’’ Expert Syst. Appl., vol. 77, pp. 114–135,
Jul. 2017, doi: 10.1016/j.eswa.2017.01.058. Newcastle University as a Lecturer, where he was
[21] L. Wang, A. Liao, and Y. Ding, ‘‘Reliability evaluation of rolling bearings promoted to Senior Lecturer, in 2018. With the expertise in machine con-
for traction motors of bogies based on proportional hazards model,’’ Meas. dition monitoring, fault diagnosis, signal processing, pattern recognition,
Control Technol., vol. 31, no. 1, pp. 14–19, Jan. 2018. artificial intelligence, system reliability analysis, and reliability design, he
[22] Handbook of Reliability Prediction Procedures for Mechanical Equipment, has been consistently striving to use advanced condition monitoring and
Nav. Surf. Warfare Center Carderock Division, Potomac, MD, USA, 2009, reliability analysis techniques to help the enterprises to increase economic
pp. 14–29. benefits by lowering the downtime and operation and maintenance costs of
[23] Military Handbook Reliability Prediction of Electronic Equipment MIL- their production equipment.
HDBH-217E, Dept. Defense Assistance Mil., Federal Agencies, Ind.,
Washington, DC, USA, 1990.
[24] Reliability Date Handbook-Universal Model for Reliability Prediction of
Electronic-Components, PCBs and Equipment, Int. Electro Tech. Commis-
sion, Geneva, Switzerland, 2004, pp. 5–10.
[25] Safety Requirements for Electric Vehicles—Part 3: Protection Against
Electric Shock, document GB/T 18384.3, National Standardization Man-
agement Committee, China, 2015, pp. 2–9. KEXIANG WEI received the Ph.D. degree in
[26] Drive Motor System for Electric Vehicles—Part1: Specification, docu- mechanical engineering from Shanghai Jiao-
ment GB/T18488.2, National Standardization Management Committee, tong University, in 2006. Then, he held a
China, 2015, pp. 4–15. postdoctoral position at Shanghai Jiaotong Uni-
[27] E. A. Elsayed, Reliability Engineering, 2nd ed. Beijing, China: Publishing versity for two years. He was invited to the Uni-
Housing of Electronics Industry, 2012, pp. 99–104 and 45–120. versity of Sydney as a Visiting Scholar, from
2010 to 2011. He is currently a Professor of
mechanical engineering with the Hunan Insti-
tute of Engineering. His research interests include
machinery dynamics, vibration control, condition
monitoring, and fault diagnosis.

XIONG SHU received the B.S. degree in mechan-

ical engineering from the Hunan Institute of Sci-
ence and Technology, Hunan, China, in 2011, and YUN ZHU received the B.S. and M.S. degrees
the M.S. degree in mechanical engineering from from the Changsha University of Science and
the Changsha University of Science and Technol- Technology, Changsha, China, in 2009 and 2013,
ogy, Hunan, in 2014. He is currently pursuing respectively, all in mechanical engineering. He is
the Ph.D. degree in mechanical engineering with currently a Teacher majoring in automobile engi-
the Hunan University of Science and Technology. neering with the Hunan Institute of Engineering.
He has been engaged in research on new-energy His current research interests include electric vehi-
vehicles, since 2012. His research interests include cle energy management, reliability analysis tech-
signal processing, fault diagnosis, and reliability analysis techniques of niques, and energy harvesting.
power train system of electric vehicle.

HONGXIANG ZOU received the B.S. degree from

the Huazhong University of Science and Tech-
nology, Wuhan, China, in 2001, the M.S. degree
YINGFU GUO received the B.S. and M.S. degrees from the Changsha University of Science and
from Xi’an University of Science and Technology, Technology, Changsha, China, in 2014, and the
1983 and 1989, respectively, and the Ph.D. degree Ph.D. degree from Shanghai Jiao Tong Univer-
from the China University of Mining and Tech- sity, Shanghai, China, in 2017, all in mechani-
nology, in 2002. He is currently a Professor of cal engineering. He is currently a Distinguished
mechanical engineering with the Hunan University Researcher of Hunan Provincial Key Laboratory of
of Science and Technology. His research interests Vehicle Power and Transmission System. His cur-
include machine condition monitoring, fault diag- rent research interests include smart structure design and dynamic analysis,
nosis, reliability analysis, and signal processing. reliability analysis techniques, and vibration control.

VOLUME 8, 2020 5307