Available via license: CC BY 3.0
Content may be subject to copyright.
IOP Conference Series: Earth and Environmental Science
PAPER • OPEN ACCESS
Research on Fault Diagnosis of External Short Circuit of Lithium Battery
for Electric Vehicle
To cite this article: Changchun Liu et al 2020 IOP Conf. Ser.: Earth Environ. Sci. 440 032106
View the article online for updates and enhancements.
This content was downloaded from IP address 212.80.200.214 on 19/03/2020 at 17:16
Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd
ESMA 2019
IOP Conf. Series: Earth and Environmental Science 440 (2020) 032106
IOP Publishing
doi:10.1088/1755-1315/440/3/032106
1
Research on Fault Diagnosis of External Short Circuit of
Lithium Battery for Electric Vehicle
Changchun Liu1, a, Tao Wu1, b, Cheng He2, c
1School of Environmental and Material Engineering, Shanghai Polytechnic University,
Shanghai 201209, China
2School of Intelligent Manufacturing and Control Engineering, Shanghai Polytechnic
University, Shanghai 201209, China
a651979759@qq.com, b97314950@qq.com, checheng@sspu.edu.cn
Abstract. This study conducted an experimental study on the external short circuit (ESC)
fault characteristics of lithium-ion batteries for electric vehicles. An experimental
platform was established to simulate the electrical behavior of lithium batteries during
ESC failure using a modified first-order RC model. The model parameters are re-
identified by the dynamic neighborhood particle swarm optimization algorithm. An
ESC fault diagnosis algorithm based on two-layer model is proposed. The first layer
performs initial fault detection and the second layer performs accurate model-based
diagnostics. The four new units are shorted to evaluate the proposed algorithm. The
results show that the ESC fault can be diagnosed within 5 s, and the error between the
model and the measured data is less than 0.36 V. The proposed algorithm can make a
correct diagnosis.
1. Introduction
In recent years, the rapid development of advanced lithium-ion battery technology has greatly promoted
the development of electric vehicles worldwide, providing a promising solution to solve the global
energy crisis and environmental pollution [1-2]. During the use of the vehicle, battery failures such as
overcharging, over discharging, and short circuit may occur. These faults may generate heat and gas
inside the lithium battery, which may cause serious consequences such as thermal runaway, fire, or even
explosion. In order to reduce the loss caused by battery failure, fault diagnosis is very necessary and has
great practical value.
2. External short circuit fault test platform
Figure 1 shows the experimental platform for external short circuit (ESC) External Short Circuit fault
test, namely, host computer, electronic load, charge and discharge control / signal detection circuit,
circuit power supply, charging power supply, CAN bus, temperature control box, security Protection
box, sensor, etc. The upper computer is set outside the temperature control box and connected to the
short circuit trigger controller through the CAN bus to ensure safe and reliable operation. Lithium
batteries, relays and sensors are placed in a safety box temperature control box when the safety box is
placed. During the experiment, the voltage, current and temperature changes were recorded by the data
ESMA 2019
IOP Conf. Series: Earth and Environmental Science 440 (2020) 032106
IOP Publishing
doi:10.1088/1755-1315/440/3/032106
2
acquisition unit. Two thermocouples were attached to the anode and cathode of the battery, respectively,
for measuring the temperature of the battery.
Figure 1. External short circuit fault test platform
ESMA 2019
IOP Conf. Series: Earth and Environmental Science 440 (2020) 032106
IOP Publishing
doi:10.1088/1755-1315/440/3/032106
3
3. Model establishment and optimization
R
p
C
p
R
0
V
o
(γ )
d_ESC
+
U
p
-
I
batt
+
-
V
t
Figure 2. First-order RC circuit model of lithium battery
A first-order RC model is used to describe the ESC process lithium battery. The voltage calculation
formula for the first-order RC model is:
,1 , ,
exp( ). [1 exp( )]
pk pk p battk p p p
pp
tt
VVR IRC
(1)
The goal of optimization is to find the best parameters to maximize the model to meet the test data.
The cost function is the root mean square error (RMSE) between the test data and the model prediction
and is described as follows:
2
_, _mod,
1
()
()
n
t testj t elj
j
VV
Jn
(2)
_and
_ are the terminal voltages from the test data and the model simulation, respectively,
j is the sampling point and n is the total number of samples. The particle updates its speed to three items,
its current speed, its individual best historical position (self-optimal position), and its best position within
the neighbourhood (group best position), described as:
11 2 2
( 1) () ( () ()) ( () ())
iiii ii
vk wvkcrpkxk crgkxk
(3)
Where w is the inertia weight,and is the weight coefficient,
and
is a random value. p and g
represent the optimal position and the optimal position of the group, respectively, and are calculated as
follows:
**
( ) { ( ) | ( ( )) ( ( )), [1, ]}
ii i i
p
kx Jx Jx k
(4)
ESMA 2019
IOP Conf. Series: Earth and Environmental Science 440 (2020) 032106
IOP Publishing
doi:10.1088/1755-1315/440/3/032106
4
( ) { ( ) | ( ( )) ( ( ))), [ , ]}
is s
gk xk Jxk Jxk i i
(5)
Where
is the neighbourhood boundary. The neighbourhood starts from k=2 and grows larger until
the neighbourhood of each particle covers the entire group. The neighbourhood k is initially set to a very
small value to make the search results versatile and grows to a larger value during the second half of the
search process to increase the convergence speed of the algorithm.
4. Experimental results and analysis
In order to observe the effectiveness of the DPSO algorithm, the terminal voltage RMSE (root mean
square error) between the test data and the model prediction is used to quantitatively describe the model
accuracy. During the ESC test, the peak currents of the four cells reached 60.1 A, 66.7 A, 69.8 A and
70.6 A, while the voltage dropped to 0.93 V, 1.01 V, 0.79 V and 1.12 V in 0.2 s. Current and voltage
Changes can implement the rules in the top-level algorithm to make the fault model effective. Figure 3
shows a comparison between model predictions and test data measured over 10 seconds. It can be seen
that the measured voltage signal is very close to the model prediction. Obviously, after the first 3 s, two
batteries (Battery 2 and No. 3) were diagnosed as ESC faults because the RMSE during this time period
was below the threshold
; after 5 seconds, all four cells were diagnosed as ESC Failure because the
RMSE during this time period is below the threshold
.
Figure 3. Test results of four units (A) Unit number 11 (B) Unit number 12
(C) Unit number 13 (D) Unit number 14
5. Conclusion
This research studied the ESC fault characteristics of lithium batteries through experiments. The current,
voltage and temperature changes of the lithium battery during the ESC fault are compared and analyzed.
It has proven to be feasible to simulate a fault process using a modified equivalent circuit model. A first-
order RC model is used and the DPSO algorithm is used to optimize the parameters of the model. A
ESMA 2019
IOP Conf. Series: Earth and Environmental Science 440 (2020) 032106
IOP Publishing
doi:10.1088/1755-1315/440/3/032106
5
comparison of model predictions with test data shows that the model has an error of less than 0.241 V.
Based on the above experimental results, a fault diagnosis algorithm based on two-layer model is
proposed. Four new lithium batteries were tested to evaluate the proposed fault diagnosis algorithm. The
results show that the ESC faults of the four batteries can be diagnosed after 5 s, and the error between
the model prediction and the measured data is less than 0.36 V. The algorithm can get the correct
diagnosis.
Acknowledgments
This work was financially supported by Shanghai Polytechnic University Graduate Program Fund
(EGD18YJ0003).
References
[1] Manzetti S, Mariasiu F. Electric vehicle battery technologies: from present state to future systems.
Renew Sust Energy Rev 2015; 51: 1004 – 12.
[2] Zhao X, Doering OC, Tyner WE. The economic competitiveness and emissions of battery electric
vehicles in China. Appl Energy 2015; 156: 666 – 75.
