Safety research on the power Li-ion battery cells for electric vehicles · 2020. 9. 25. · I . Safety research on the power Li-ion battery cells for electric vehicles . Sheng Yang

I

Safety research on the power Li-ion battery cells for electric vehicles

Sheng Yang

June 2020

Submitted in fulfilment of the requirements of the degree of

Doctor of Philosophy

Faculty of Science, Engineering and Technology

Swinburne University of Technology

Melbourne, Australia

II

III

Abstract

The introduction of electric vehicles (EVs) is an effective measure to significantly

reduce greenhouse gas emissions associated with the usage of fossil fuels.

Lithium-ion (Li-ion) batteries have been considered as dominant energy storage

systems for EVs due to their advantages over other energy storage systems. However,

the poor thermal stability of Li-ion batteries has attracted more and more attention on

their safety and reliability as the development and expansion of EVs market.

The derivative of force with respect to displacement can be used to characterize the

stiffness of Li-ion batteries. In chapter 2, the stiffness of Li-ion batteries in the

quasi-static compression tests is analysed. It is found that the stiffness curve distinctly

shows three stages corresponding to densification stage, microscopic damage stage

and macroscopic failure stage. The Li-ion battery’s stiffness increases in the

densification stage while decreasing in the microscopic damage stage and

macroscopic failure stage. Hence, the constitutive model of the jellyroll of Li-ion

batteries is improved with considering microscopic damage, which is then validated

by the established explicit finite element model (FEM) of the Li-ion batteries. The

voltages and temperatures of Li-ion batteries are also measured to further compare

their responses at different stages. It is found that the internal short circuit of Li-ion

batteries at the fully charged state occurs in the microscopic damage stage while that

of Li-ion batteries at low and medium SOCs occurs in the beginning of macroscopic

failure stage.

Understanding the mechanical responses of Li-ion batteries during the EVs

collision is the key to solve subsequent thermal runaway problems, which involves

complicated and coupled behaviors. In chapter 3, several dynamic compression tests

are performed on two types of 18650 Li-ion batteries, namely LiNiCoAlO2 (NCA)

and LiNiCoMnO2 (NCM). Experimental results indicate that their abilities to resist

deformation both have positive relationship with the loading rate, namely, the strain

rate hardening behaviors. The strain rate hardening behaviors of NCM Li-ion batteries

IV

can be observed at low loading rate while that of NCA Li-ion batteries can only be

observed when the loading rate rises to a certain value. The constitutive model of the

jellyroll of Li-ion batteries is proposed with considering the strain rate hardening

behaviors, which is then validated by the established explicit FEM of the Li-ion

batteries. The proposed model can be used to evaluate the safety performance of

Li-ion batteries under crash accidents and provide useful information for the structure

design of battery packs in EVs.

The thermal runaway induced by the internal short circuit (ISC) of Li-ion batteries

under mechanical abusive conditions is another focus of EV industries. In chapter 4,

the coupled electrochemical-electric-thermal model is improved with considering the

material properties and the damaged area of the short circuit object, which can predict

the occurrence of thermal runaway of Li-ion batteries under various ISC conditions.

Simulation results indicate that the safety performance of Li-ion batteries under

mechanical abusive conditions can be improved by appropriately increasing the

adhesion strength between the aluminum current collector and the positive electrode.

All above-mentioned outcomes can provide valuable guidance for the safety design

of battery packs in EVs.

Key words: Li-ion batteries, Safety performance, Mechanical abusive conditions,

Constitutive model, Finite element model, Electrochemical-electric-thermal coupled

model

V

Acknowledgement

This research work is completed under the careful guidance and patient help of my

principal supervisor Professor Cheng Lin and second supervisor Associate Professor

Wenwei Wang at Beijing Institute of Technology, China. It includes the selection of

the topic, the formulation of research content and program, the performance of related

work and the writing of this thesis. These two supervisors both give me meticulous

guidance, care and help not only in the side of my study and research but also in the

side of my life and emotion. In the side of my study and research, their profound

knowledge and excelsior attitude make me effectively carry out the related research of

safety performance of Li-ion batteries. In the side of my life and emotion, their

elegant style of conversation and common touch benefit me a lot and they also teach

me how to deal with people and things. Besides, they also try to create various

opportunities to send me to foreign countries to have academic communications with

world famous scholars, learn the latest research trends in the international frontier and

expand my academic horizons and enrich my research theories.

On the occasion of completing this thesis, I also would like to specially express my

sincere gratitude to my principal supervisor Associate Professor Weixiang Shen and

second supervisor Professor Guoxing Lu at Swinburne University of Technology for

their meticulous care and careful help. Their rigorous and realistic scientific attitude,

excelsior scientific style, sharp and meticulous scientific thinking and innovative

research awareness are beneficial to my whole life. Moreover, A/Prof. Shen strictly

requires every cooperative paper with excelsior attitudes and paid a lot of efforts to

review each paper and improve their qualities. Prof. Lu provided me a strong support

on my experiments.

Finally, I would like to express my sincere gratitude to Professor Rui Xiong. He

gives me a strong help and patient guidance on my research work. Besides, he offers

me a help, points out the aims for me and promotes me to move on without hesitation

when I lost the direction of both my study and life. Besides, I want to thank for the

VI

financial assistance of China Scholarship Council and Swinburne University of

Technology. Thanks also go to all experts and professors who reviewed this thesis,

your valuable comments and suggestions significantly improve this thesis.

I am very appreciated for my parents for their care, help, support, understanding

and tolerance, who bring me up and teach me to use ethical rules to guide my

behaviors. I will use the rest of my life to repay them.

VII

Declaration

This is to certify that the thesis entitled "Safety research on the power Li-ion

battery cells for electric vehicles" submitted in fulfillment of the requirements for the

Degree of Doctor of Philosophy in the Faculty of Science, Engineering and

Technology of Swinburne University of Technology, is my own original work and

that it contains no material which has been accepted for the award of any other degree

or diploma, except where due reference is made in the text of the thesis. To the best of

my knowledge, it contains no material previously published or written by another

person except where due reference is made in the text of the thesis.

Sheng Yang

June 2020

VIII

List of Publications

1. SCI journal papers

[1] WW Wang*, S Yang, C Lin. Clay-like mechanical properties for the jellyroll of

cylindrical Lithium-ion cells. Applied Energy, 2017, 196: 249-258.

[2] WW Wang, S Yang*, C Lin, WX Shen, GX Lu, YD Li, JJ Zhang. Investigation

of mechanical property of cylindrical lithium-ion batteries under dynamic

loadings. Journal of Power Sources, 2020, 451:227749.

[3] S Yang, WW Wang*, C Lin, WX Shen, YD Li. Improved constitutive model of

the jellyroll for cylindrical lithium ion batteries considering microscopic

damage. Energy, 2019, 185: 202-212.

[4] WW Wang, S Yang*, C Lin, YD Li. State of charge dependent constitutive

model of the jellyroll of cylindrical Lithium-ion cells. IEEE ACCESS, 2018, 6:

26358-26366.

[5] S Yang, WW Wang*, C Lin, WX Shen, YD Li. Investigation of internal short

circuits of lithium-ion batteries under mechanical abusive conditions. Energies,

2019, 12: 1885.

2. EI and conference papers

[1] WW Wang*, S Yang, C Lin, YD Li. Measuring the internal short circuit

resistance under mechanical abusive conditions. International Conference on

Electric and Intelligent Vehicles (ICEIV2018), Melbourne, Australia, Nov,

2018.

[2] WW Wang*, S Yang, C Lin, YD Li. Mechanical and electrical response of

cylindrical Lithium-ion cells at various State of Charge. Energy Procedia 2018;

145; 128–132.

[3] WW Wang*, S Yang, C Lin, FC Sun. Mechanical behaviors of cylindrical

Lithium-ion cells at various State of Charge. International Symposium on

Electric Vehicles (ISEV 2017), Stockholm, Sweden, July 2017.

[4] WW Wang*, S Yang, C Lin. Clay-like mechanical properties of components for

IX

the jellyroll of cylindrical Lithium-ion cells. Energy Procedia 2016; 104;

56–61.

[5] WW Wang*, S Yang, FC Sun, QQ Yu. The Clay-like mechanics model of

cylindrical Lithium-ion battery cells under radial compression. Energy Procedia

2016; 88; 652–655.

X

Contents

Chapter 1 Introduction ............................................................................. 1 1.1 Background and motivation .............................................................................. 1

1.2 Researches on mechanical properties of Li-ion batteries ................................. 1

1.2.1 Researches on mechanical properties of Li-ion batteries under

quasi-static loadings ......................................................................................... 2

1.2.2 Researches on mechanical properties of Li-ion batteries under dynamic

loadings ............................................................................................................ 4

1.3 Research contents of this thesis ........................................................................ 5

Chapter 2 Investigation of safety performance of cylindrical Li-ion batteries under quasi-static loadings ...................................................... 6

2.1. Analysis of mechanical, electric and thermal responses of Li-ion batteries

under quasi-static loadings ...................................................................................... 6

2.2. Constitutive model for jellyroll of Li-ion batteries under quasi-static loadings

............................................................................................................................... 12

2.3. Validation and discussion ............................................................................... 16

2.4. Conclusions .................................................................................................... 19

Chapter 3 Investigation of safety performance of cylindrical Li-ion batteries under dynamic loadings ......................................................... 21

3.1. Analysis of mechanical, electric and thermal responses of Li-ion batteries

under dynamic loadings ........................................................................................ 21

3.2. Constitutive model for jellyroll under dynamic conditions ........................... 27

3.3. Validation and discussion ............................................................................... 28

3.4. Conclusions .................................................................................................... 31

Chapter 4 Investigation of Internal Short Circuits of Li-ion Batteries under Mechanical Abusive Conditions ................................................. 32

4.1 Investigating thermal responses of a Li-ion battery under various ISC

conditions .............................................................................................................. 32

4.1.1 Improved electrochemical-electric-thermal model ............................... 32

4.1.2 Effect of various ISC types on thermal responses ................................ 36

XI

4.1.3 Effect of the area of the ISC object on thermal responses .................... 39

4.1.4 Effect of the SOC of a Li-ion battery on thermal responses ................. 41

4.2 Validation and discussion ................................................................................ 44

4.3 Conclusions ..................................................................................................... 46

Chapter 5 Conclusions and Expectations ............................................. 47 5.1 Conclusions of this research ........................................................................... 47

5.2 Innovations of this research

............................................................................................................................... 47

5.3 Expectations in the future ............................................................................... 48

XII

List of Figures

2.1 (a) Measured force-time curve for cell 1 at the SOC of 100%; (b) Derivative of the

force with respect to time-stiffness; (c) Measured voltage and temperature

responses for cell 1 ............................................................................................... 7


force with respect to time-stiffness; (c) Measured voltage response for cell 2 at

the SOC of 20% ..................................................................................................... 8


force with respect to time-stiffness; (c) Measured voltage response for cell 3 at

the SOC of 60% ..................................................................................................... 9


force with respect to time-stiffness; (c) Measured voltage and temperature

responses for cell 4 at the SOC of 100% ............................................................. 10

2.5 Measured force-displacement relations at various SOCs when the Li-ion battery is

compressed between two rigid flat plates and their analytical fittings: (a) SOC=0;

(b) SOC=0.2; (c) SOC=0.4; (d) SOC=0.5; (e) SOC=0.6; (f) SOC=0.8 ............... 13

2.6 Schematic description of Li-ion battery deformation compressed between two

plates .................................................................................................................. 14

2.7 Calculated and analytical stress-strain curves at various SOCs: (a) SOC=0; (b)

SOC=0.2; (c) SOC=0.4; (d) SOC=0.5; (e) SOC=0.6; (f) SOC=0.8 .................... 15

2.8 Simulated and measured force-displacement curves on Li-ion cells between two

rigid planes: (a) SOC=0; (b) SOC=0.2; (c) SOC=0.4; (d) SOC=0.5; (e) SOC=0.6;

(f) SOC=0.8 ......................................................................................................... 18

2.9 Simulated and measured force-displacement curves on Li-ion cells (a) under rigid

rod indentation, SOC=0.2; (b) under rigid rod, SOC=0.5; (c) under a

hemispherical punch, SOC=0.2; (d) under a hemispherical punch, SOC=0.5 .... 19

3.1 Compression tests at different loading rates: (a) INSTRON VHS 8800; (b) MTS

EXCEED E45; (c) Measured mechanical responses for NCM Li-ion batteries at

different loading rates; (d) Measured mechanical responses for NCA Li-ion

batteries at different loading rates ........................................................................ 22

XIII

3.2 (a) INSTRON 5985; (b) FLUKE TI 400; (c) HIOKI 8880; (d) A NCA Li-ion

battery at the SOC of 100% ................................................................................. 23

3.3 (a) Measured load-time curve for NCM Li-ion battery at the SOC of 0.4; (b)

Measured voltage response; (c) Measured temperature response ...................... 24

3.4 (a) Measured Load-time curve for NCA Li-ion battery at the SOC of 0.4; (b)

Measured voltage response; (c) Measured temperature response ........................ 25

3.5 Calculated stress-strain curves for NCM Li-ion batteries at different loading rates

.............................................................................................................................. 28

3.6 Calculated stress-strain curves for NCA Li-ion batteries at different loading rates

.............................................................................................................................. 28

3.7 Simulated and measured force-displacement curves for NCM Li-ion batteries at

different loading rates: (a) v=0.5 mm/min; (b) v=5 mm/min; (c) v=25 mm/min;

(d) v=3 m/s ......................................................................................................... 29

3.8 Simulated and measured force-displacement curves for NCA Li-ion batteries at

different loading rates: (a) v=5 mm/min; (b) v=25 mm/min; (c) v=60 mm/min; (d)

v=3 m/s .............................................................................................................. 30

4.1 Thermal responses of various ISC conditions for a Li-ion battery at SOC = 0.2 and

0.5 mminr :(a) Positive–negative, (b) copper-positive, and (c)

aluminum-negative ............................................................................................ 38

4.2 Thermal response of various ISC areas, SOC = 0.4, positive–negative (a)

0.5 mminr , and (b) 1.5 mminr = .................................................................... 40

4.3 Thermal response of various ISC areas, SOC = 0.4, aluminum-negative (a)

=0.5 mminr , and (b) =1.5 mminr ........................................................................ 41

4.4 Thermal response of various capacities, =1.5 mminr , positive–negative (a) SOC =

0.2, and (b) SOC = 0.6 ....................................................................................... 42

4.5 Thermal response of various capacities, =1.5 mminr , aluminum-negative (a) SOC

= 0.2, and (b) SOC = 0.6 ...................................................................................... 43

4.6 Thermal responses under various loadings (unit: oC) at SOC = 0.2: (a) two rigid

plates, (b) a rigid rod, and (c) a hemispherical punch ........................................ 44

4.7 Thermal response of the battery compressed by a rigid rod (unit: oC) at SOC = 0.6

.............................................................................................................................. 45

XIV

List of Tables

2.1 Specifications of 18650 Li-ion batteries ................................................................ 6

2.2 Fitting parameters ki, 0ia , 0ib and 0ic at various SOCs .................................. 12

2.3 Simulation parameters ......................................................................................... 17

3.1 Specifications of NCA 18650 Li-ion battery ......................................................... 21

3.2 Fitting parameters 1a , 1b and 1c for NCM and NCA L-ion batteries at various

loading rates ......................................................................................................... 27

4.1 Governing equations of the axisymmetric model ................................................ 33

4.2 Physiochemical parameters used in the axisymmetric model .............................. 34

4.3 The definition of those parameters used in improved coupled model ................. 35

4.4 Parameters of different internal short circuit (ISC) types .................................... 35

Chapter 1 Introduction

1

Chapter 1

Introduction

1.1 Background and motivation

Electric vehicles (EVs) have been considered as an effective measure to alleviate

the emission of greenhouse gas associated with the usage of fossil fuels due to their

unique characteristics [1-7]. Li-ion batteries have been used as the dominant energy

storage for EVs due to their advantages over other types of energy storage in terms of

high energy/power density and long cycle life [8-15].

However, the poor thermal stability of Li-ion batteries has affected the safety and

reliability of EVs, especially when the energy/power density and physical size of

Li-ion batteries are increased to make EVs to meet the demand for long driving

mileage [16-18]. For an example, when Li-ion batteries operate above the normal

temperature range caused by mechanical abuse (e.g. EV crash) a series of exothermic

reactions may be triggered inside Li-ion batteries, leading to thermal runaway

[19-24].

Furthermore, highly energetic electrode materials, flammable electrolyte and the

oxygen produced by side reactions make thermal runaway of Li-ion batteries

extremely dangerous [25-30]. The catastrophic accidents associated with thermal

runaway are induced inside Li-ion batteries, which have been an obstacle for further

development and popularization of EVs. Therefore, there is an urgent need to

investigate the safety performance of Li-ion batteries and explore the mechanism of

thermal runaway under abusive conditions.

1.2 Researches on mechanical properties of Li-ion batteries

Li-ion batteries are mainly composed of two parts: a shell casing and the jellyroll.

The jellyroll consists of several identical single cells, which is composed of five

layers: positive/negative electrode (or cathode/anode), separator and copper/aluminum

foil. The electrolyte and separator are designed as a good ionic conductor but also act

as an electronic insulator, which allow Li-ions to pass while blocking electrons.


2

However, ISCs may be triggered inside a Li-ion battery under mechanical abusive

conditions (e.g. EV crashes) due to the failure of separator, or physical contact

between two electrodes or an electrode and a current collector (an aluminum or

copper foil) or two current collectors. This kind of ISC is the major cause for thermal

runaway. Hence, the understanding of mechanical properties of Li-ion batteries is the

key to reveal the complex mechanism of thermal runaway.

1.2.1 Researches on mechanical properties of Li-ion batteries under quasi-static

loadings

Some researchers focused on the mechanical property of separator and shell casing

of Li-ion batteries. Sheidaei et al. investigated the effect of electrolyte on the

mechanical property of a single layer of separator at both machine and transverse

direction [31]. They found that the mechanical properties of separators are anisotropic,

namely, the separators have different mechanical properties at different directions.

Zhang et al. investigated the mechanical properties of separators after several

charge/discharge cycles [32]. They found that the cycle number of separators had an

effect on their failure. Wierzbicki et al. calculated the compression resistance of the

metal casing and the positive end-cap of 18650 Li-ion batteries [33]. Although the

hardness of metal casing and the positive end-cap is high, their contributions to the

peak load of 18650 Li-ion batteries are small due to their small thickness.

Some researchers focused on the mechanical property of anode and cathode active

materials of Li-ion batteries. Zhang et al. utilized simulation to study the stress inside

electrode particles induced by the insertion of Li-ions [34]. They found that

decreasing the size while increasing the aspect ratios of electrode particles could

reduce the stress inside the electrode induced by the insertion/removal of Li-ions. Zhu

et al. studied the deformation mechanisms of electrodes of Li-ion batteries through

scanning electron microscope (SEM) [35]. They captured the derivation and

development of a number of cracks during the deformation process of electrodes.

Zhang et al. proposed the constitutive model of electrodes of Li-ion batteries, which

was used to predict their failure behaviors [36]. They also found that electrode failure


3

was the combination of tensile and compression failure.

Some researchers focused on the mechanical property of electrolyte of Li-ion

batteries. Dixon et al. compared the deformation behaviors of Li-ion batteries with

and without electrolyte under mechanical tests [37]. It was found that the electrolyte

would decrease the force and displacement level of Li-ion batteries, namely, the

measured force and displacement of Li-ion batteries with electrolyte were smaller

than that of Li-ion batteries without electrolyte at the same loading scenarios.

Some researchers focused on the mechanical characteristics of whole Li-ion

batteries. Sahraei et al. utilized the experimental data measured from compression

tests to obtain the stress and strain relationship of pouch Li-ion batteries [38]. They

also found that the aluminum shell of pouch Li-ion batteries had obvious constraints

on its geometry structure, which affect its deformation and failure behaviors in some

cases. However, it was different from 18650 Li-ion batteries, whose steel shell had

little effect on its deformation and failure behaviors. Luo et al. investigated the

development of damage inside a pouch Li-ion battery under an indentation test. They

observed an important inflection point on the measured load-displacement curve of

pouch Li-ion batteries, which could be considered as the initiation of damage inside

Li-ion batteries [39].

An individual Li-ion battery is composed of several components with each

component made of different material. The material properties of some components

exhibited anisotropic. Xu et al. took the material anisotropy into consideration and

then improved the constitutive model of jellyroll, which could better predict the

mechanical responses of 18650 Li-ion batteries under bending cases [40]. The

mechanical properties and failure mechanisms of 18650 Li-ion batteries under axial

direction were investigated in [41] which found three causes as follows: 1) the

physical contact between the steel casing and the jellyroll due to the fracture of

separator. 2) the physical contact among different components of the jellyroll. 3) the

physical contact between the safety value and the jellyroll. They may contribute to the

trigger of ISC of Li-ion batteries under axial direction.


4

Xu et al. [42] investigated the deformation and failure behaviors of 18650 Li-ion

batteries at various SOCs under both compression and bending tests. They found that

the structure stiffness of Li-ion batteries increased as the increase of SOCs while the

failure strain of Li-ion batteries had an inverse linear relationship with their SOCs,

namely, the failure strain of Li-ion batteries decreased as the increase of SOCs.

However, the SOC-dependent behaviors were found insignificant for pouch Li-ion

batteries under spherical punch head indentation tests [43]. They also found that the

SOC-dependent mechanical behaviors of Li-ion batteries are due to the internal stress

triggered by the constraint of volume expansion.

The above-mentioned researches focused on the mechanical properties of Li-ion

batteries at the fully discharged state or the effect of the SOC of Li-ion batteries on

their mechanical behaviors using experimentation. In this study, the constitutive

model of the jellyroll of Li-ion batteries is established under quasi-static loadings,

considering the effect of SOC.

1.2.2 Researches on mechanical properties of Li-ion batteries under dynamic

loadings

For EV industry and customers, the safety performance of EVs especially at the

time of vehicle collision is one of their focuses, which are concerned with mechanical

responses of Li-ion battery under dynamic loadings.

Kisters et al. investigated the effect of electrolyte on the dynamic responses of

Li-ion batteries [44]. They found the peak force of the elliptic Li-ion battery with

electrolyte increased by 25% under dynamic loadings while that without electrolyte

increased by 65%. Chen et al. performed drop-weight tests on prismatic Li-ion

batteries [45]. They observed three failure modes of Li-ion batteries under dynamic

loadings, namely there was a total loss of battery capacity, partial loss of battery

capacity and no loss of battery capacity. The above studies focused on

experimentation and result analysis.

Xu et al. utilized simulation to study the dynamic responses of cylindrical Li-ion

batteries [46]. Kermani et al. utilized a Johnson-Cook model to predict the mechanical


5

responses of Li-ion batteries under dynamic tests [47].

Currently, all corresponding researches under dynamic loadings are still in the

preliminary stage and there is no systematic research method, so it is necessary to

investigate mechanical properties of Li-ion batteries under dynamic loadings to reveal

complex mechanism of thermal runaway, which has recently become a hotspot of the

research on battery safety.

1.3 Research contents of this thesis

Thermal runaway process of a Li-ion battery induced under mechanical abusive

conditions can be simply divided into three stages. In stage 1, the ISC is triggered

inside a Li-ion battery and battery temperature starts to increase due to the large Joule

heat generated by large ISC current. It should be noted that Joule heat induced by ISC

current is much larger than the heat generated by decomposition of solid electrolyte

interphase (SEI). In stage 2, thermal contraction of separator inside the battery will be

induced once temperature exceeds critical value of o120 140 C , which will cause

more serious ISC and stimulate battery temperature rise again. In stage 3, a chain of

side reactions inside a Li-ion battery is induced and battery temperature will increase

rapidly, leading to thermal runaway. The process of thermal runaway can’t be stopped

once it is triggered inside a Li-ion battery.

This research is hence divided into two parts to investigate the safety performance

of Li-ion batteries under various mechanical abusive conditions. The first part is

aimed to predict the trigger of the ISC inside a Li-ion battery. It analyzes the

mechanical, electric and thermal responses of Li-ion batteries and investigates the

failure and out-of-control behaviors of Li-ion batteries after ISC. Then, the

constitutive model of Li-ion batteries based on experimental data is proposed to

predict mechanical responses at the trigger of ISC inside Li-ion batteries. The second

part is aimed to predict the thermal runaway triggered inside a Li-ion battery after ISC.

It investigates the development process of ISC of Li-ion batteries and proposes the

electrochemical-electric-thermal coupled model of Li-ion batteries to predict the

thermal responses and the trigger of thermal runaway after various ISCs.

Chapter 2 Investigation of safety performance of cylindrical Li-ion batteries under quasi-static loadings

6

Chapter 2

Investigation of safety performance of cylindrical Li-ion

batteries under quasi-static loadings In this chapter, quasi-static compression tests are performed on 18650 Li-ion

batteries at various SOCs. The stiffness, electric and thermal characteristics of 18650

Li-ion batteries under quasi-static loadings are analyzed to investigate its safety

performance. The constitutive model of the jellyroll of a 18650 Li-ion battery is

proposed by considering the microscopic damage, which is then validated by the

established explicit FEM of a Li-ion battery.

2.1. Analysis of mechanical, electric and thermal responses of Li-ion

batteries under quasi-static loadings

Quasi-static compression tests are performed on 18650 Li-ion batteries at the SOCs

of 20%, 60% and 100% and the loading rate is 0.5 mm/min. The specifications of

these 18650 Li-ion batteries are shown in Table 2.1. The measured load-displacement

relation is converted into the load-time relation to easily compare with the

corresponding relation of voltage and temperature versus time. Besides, the stiffness

of a Li-ion battery is obtained through the derivative of force with respect to

displacement to further investigate its mechanical responses under compression tests.

Table 2.1. Specifications of 18650 Li-ion batteries

Items Specifications

Normal capacity 2100 mAh

Normal voltage 3.6 V

Charge voltage 4.2 0.05 V

Cut off voltage 2.5 V

Cathode Material LiNiCoMnO2

Fig. 2.1 (a) describes the force-time relation and Fig. 2.1 (b) describes the

stiffness-time relation. It indicates that the deformation process of a Li-ion battery can


7

be divided into three distinct stages corresponding to densification stage, microscopic

damage stage and macroscopic failure stage.

Figure 2.1. (a) Measured force-time curve for cell 1 at the SOC of 100%; (b) Derivative of the

force with respect to time-stiffness; (c) Measured voltage and temperature responses for cell 1.

In stage 1, the stiffness of a Li-ion battery keeps rising as the compression continues

until it reaches its maximum value, as shown in Fig. 2.1 (b). In stage 2, the decrease in

stiffness of a Li-ion battery indicates that microscopic damage appears inside the

battery, including micro cracks or micro holes. The number and size of those micro


8

cracks or holes increase as the compression continues until the measured force

reaches its maximum value, as depicted in Fig. 2.1 (a) and (b). Finally, the

deformation process enters stage 3 as the occurrence of macroscopic fracture

indicated by the decrease in the measured force.


force with respect to time-stiffness; (c) Measured voltage response for cell 2 at the SOC of 20%.

Fig. 2.1 (c) shows the measured voltage and temperature with respect to time. It is

noteworthy that the maximum force occurs at 1068 s (see Fig. 2.1 (a)) while the


9

voltage starts to decrease and temperature starts to increase at 972 s, which indicates

that the ISC has been triggered before the measured peak force, namely, at stage 2.

After the ISC is triggered, the battery voltage slowly decreases from 4.11 V at 972 s to

4.03 V at 1072 s and the battery temperature slowly increases from 29.2 °C at 972 s to

35.18 C at 1072 s.


force with respect to time-stiffness; (c) Measured voltage response for cell 3 at the SOC of 60%.

Once the measured force reaches its maximum value, the voltage rapidly decreases


10

from 4.03 V at 1072 s to 1 V at 1102 s and the temperature rapidly increases from

48.1 C to 745.52 C within 4 s, which is contributed from the internal

macroscopic fracture.


force with respect to time-stiffness; (c) Measured voltage and temperature responses for cell 4.

The Li-ion batteries at the SOCs of 20% and 60% are chosen to perform

indentation tests to investigate the trigger of ISC at low and medium SOCs. Fig. 2.2 (c)

shows the voltage response of cell 2 at the SOC of 20% and Fig. 2.3 (c) describes the


11

voltage response of cell 3 at the SOC of 60%. The stiffness of a Li-ion battery is also

analyzed during the indentation tests, as shown in Fig. 2.2 (b) and Fig. 2.3 (b),

respectively. Several conclusions can be obtained from the experimental results as

observed in Figs. 2.2 and 2.3. Firstly, the deformation process of Li-ion batteries at

low and medium SOCs also shows three distinct stages. Secondly, for Li-ion batteries

at low and medium SOCs, the battery voltage is stable before the maximum force and

it then rapidly decreases after the maximum force. Thirdly, the ISC of Li-ion batteries

at the SOCs of 20% and 60% is triggered at the stage 3 (i.e. macroscopic failure

stage).

The other kind of indentation test is performed on cell 4 at the fully charged state to

further investigate its electrical and thermal response after the trigger of ISC. The

difference of the indentation test for cell 4 from cell 1 is that the indenter will be

stopped but its position is remained when the decrease is observed in the battery

voltage. It should be mentioned that the mechanical testing platform will maintain the

load until the indenter is removed. Fig. 2.4 describes the measured mechanical,

electrical and thermal responses of cell 4. The battery voltage slowly decreases from

4.11 V at 882 s to 3.87 V at 1162 s and then rapidly decreases to 1.07 V at 1182 s, as

shown in Fig. 2.4 (c). Whereas the battery temperature slowly increases from the

initial value of 28.4 C at 852 s to 49.19 C at 1292 s. The indenter is removed

when the battery temperature continuously decreases. Fig. 2.4 (c) shows that the

battery temperature increases again after the indenter is removed.

Again, several conclusions can be obtained from the above-mentioned experimental

results. Firstly, the ISC isn’t triggered when a Li-ion battery is compressed to the

initiation point of stage 2, no matter at the low, medium SOCs and even at the SOC of

100%. Secondly, the thermal runaway won’t be induced by the ISC triggered at stage

2 but can be induced at stage 3, especially for those Li-ion batteries at the SOC of

100%, where the battery voltage and temperature responses are rapid and dramatic.


12

2.2. Constitutive model for jellyroll of Li-ion batteries under

quasi-static loadings

In this section, several 18650 Li-ion batteries at various SOCs are compressed

between two rigid plates and Fig. 2.5 describes the measured results. In all the

measured curves, it can be seen that there is one part from 0 mm to 3 mm representing

linear relationship and there is the other part from 3 mm onwards representing

exponential relationship. The Gaussian function is chosen to fit the exponential part

and a linear function simply passes the point (0, 0) and the inflection point in 3 mm.

Table 2.2 Fitting parameters ki, 0ia , 0ib and 0ic at various SOCs

SOC 0 0.2 0.4 0.5 0.6 0.8

Coefficient k 360.1 459.6 562 514 693.9 715.48

Coefficient 0a 73630 76360 78250 72030 74340 80190

Coefficient 0b 9.793 9.772 9.729 9.119 8.909 8.813

Coefficient 0c 3.307 3.38 3.435 3.121 3.125 3.055

Thus, the fitting function is described as follows

20 0 0

( 3)( ) exp( (( ) / ) ) ( 3)i

i i i

k w wH w a w b c w

(2.1)

where ki, 0ia , 0ib and 0ic are the fitting parameters. Table 2.2 lists the values of

those fitting parameters corresponding to each SOC and w is the displacement of top

plate in x direction.

The deformation process of a 18650 Li-ion battery compressed between two rigid

plates can be described in Fig. 2.6. H is the applied load corresponding to the

measured load and w is the displacement of top rigid plate in the x direction

corresponding to the measured displacement. Moreover, some following

simplifications are made based on the experimental observation:

(1) Length of a battery in the z direction remains unchanged during the whole

deformation process.


13

Figure 2.5. Measured force-displacement relations at various SOCs when the Li-ion battery is

compressed between two rigid flat plates and their analytical fittings: (a) SOC=0; (b) SOC=0.2; (c)

SOC=0.4; (d) SOC=0.5; (e) SOC=0.6; (f) SOC=0.8.

(2) The perimeter of outside metal casing of the battery remains unchanged during

the whole deformation process.

(3) Shape of a battery in the x-y plane is comprised of three parts corresponding to

one rectangular part and two semicircular parts.


14

Figure 2.6. Schematic description of Li-ion battery deformation compressed between two plates.

The strain in x direction at time t (i.e. x ) is calculated as follows.

( ) 2 ( ) 2x t r t R w (2.2)

( ) 2 ( ) 2x t t r t t R w dw (2.3)

( ) ( )( ) 2x

x t t x t dwdx t R w

(2.4)

0

2ln( )2 2

w

xdw RR w R w

(2.5)

The total volume of the 18650 Li-ion battery at time t is calculated as:

( ) (2 ( ) ( ) ( ) ( )) ( )( )2 2w wV t b t r t r t r t L L R R (2.6)

The volumetric strain of the 18650 Li-ion battery at time t (i.e. v ) is calculated as

follows:

( ) ( )( )2 2 2 2w dw w dwV t t L R R (2.7)

2( ) ( ) ( )( ) ( ) 2 2 2 2 2( ) ( )( )

2 2

V

dw w dw w dwR RV t t V td w wV t R R

(2.8)

Solving Eqs. (2.4) and (2.8) yields:

2ydwdR w

(2.9)

0

2ln( )2 2

w

ydw R wR w R

(2.10)

The equivalent stress and strain can be expressed as:


15

Figure 2.7. Calculated and analytical stress-strain curves at various SOCs: (a) SOC=0; (b)

SOC=0.2; (c) SOC=0.4; (d) SOC=0.5; (e) SOC=0.6; (f) SOC=0.8.

2 2 2 2 2 22 ( ) ( ) ( ) 6( ) 2i x y x z y z xy yz zx (2.11)

2 2 2 2 2 22 3( ) ( ) ( ) ( ) 3 2i x y x y x y xy yz zx (2.12)

The shear forces inside the Li-ion battery and its strain in z direction are ignored,


16

namely, 0 xy , 0 yz , 0 zx , and 0 z . The internal stress in x direction (i.e.

x ) is compressive stress while that in y direction is tensile stress (i.e. y ). The

following equations can be obtained based on the plastic flow rule:

2 2 2

222 2( )

x yix

x x y x y

dwd d dR w

(2.13)

2 2 2

222 2( )

y xiy

y x y x y

dwd d dR w

(2.14)

where is plastic ratio. Solving Eqs. (2.13-2.14) yields

2 22 2

x y

y x

R wR w

(2.15)

x can be calculated as:

xHbL

(2.16)

y can be obtained after solving Eqs. (2.15) and (2.16):

66y x

R wR w

(2.17)

Solving Eqs. (2.5), (2.10-2.12) and (2.16-2.17) can calculate the equivalent

stress-strain relation of a 18650 Li-ion battery. Fig. 2.7 describes the calculated

stress-strain results of Li-ion batteries corresponding to various SOCs.

2.3. Validation and discussion

The mechanical structure of an individual 18650 Li-ion battery mainly consists of

four parts corresponding to a steel casing, a positive end-cap, a jellyroll and a metal

core. In this study, only the jellyroll is considered in establishing the FEM of a Li-ion

battery since the effect of the other three parts on the battery's failure stress and strain

are small, ignoring their effect is beneficial to the mesh generation and the

improvement of computational efficiency and convergence. Moreover, the five layers

of jellyroll are integrated into a homogenized representative volume element (RVE)

during the establishment of its FEM. Hyperworks/Ls-dyna software is utilized to carry

out the simulation analysis. The mechanical property of the jellyroll is modeled by


17

crushable foam (Material type 63) and its mechanical behaviors are considered as

isotropic. The material property of two plates is set as rigid, which is consistent with

the experimental setup. Fig. 2.7 are used as the input to crushable foam to model the

mechanical responses of 18650 Li-ion batteries at various SOCs. Table 2.3 lists other

simulation parameters. The displacement of the top plate is considered as the

simulated displacement and the contact force between the top plate and jellyroll is

considered as the simulated load.

Table 2.3. Simulation parameters.

Component Young’s modulus Poisson ratio Density Tensile cut-off value

Jellyroll jellyroll 1.5 GpaE jellyroll 0.15 -32.8 g cm 10 Mpaf

Three loading cases are chosen to validate the constitutive model, including the

compression test between two rigid plates, the indentation test with a rigid rod of 12

mm in radius and the compression test with a hemispherical punch of 7 mm in radius.

Fig. 2.8 shows the simulated force-displacement curves for the compression between

two rigid plates. In all curves, the simulation results in the part from 0 mm to 5 mm

have a good agreement with those corresponding experimental results. In the part

after 5 mm, the simulation results for Li-ion battery at the SOC below 0.6 (i.e.

SOC<0.6) also have a good agreement with those corresponding experimental results

(see Fig. 2.8 (a)-(d)) while those simulation results for Li-ion battery at the SOC over

0.6 (i.e. SOC>=0.6) do not agree well with those corresponding experimental results

(see Fig. 2.8 (e)-(f)). However, the simulated forces at 8 mm are 74146 N for a Li-ion

battery at the SOC of 0.6 and that for a Li-ion battery at the SOC of 0.8 is 80498 N.

The measured force at 8 mm is 69421 N and 75673 N, respectively.

Fig. 2.9 (a)-(b) show the simulated force-displacement curves for the indentation

test with a rigid rod of 12 mm in radius and Fig. 2.9 (c)-(d) show the simulated

force-displacement curves for the compression test with a hemispherical punch of 7

mm in radius. In all curves, the simulation results also have a good agreement with

those corresponding experimental results, especially for the peak force.


18

Figure 2.8. Simulated and measured force-displacement curves on Li-ion cells between two rigid

planes: (a) SOC=0; (b) SOC=0.2; (c) SOC=0.4; (d) SOC=0.5; (e) SOC=0.6; (f) SOC=0.8.

The reasons for the errors between experimental and simulation results are as

follows.

(1) Ignoring the steel casing, the metal core and the positive-end cap in the

establishment of the FEM is responsible for the errors in the simulations especially in

the part from 0 mm to 4 mm.

(2) There are some gaps inside a real 18650 Li-ion battery. However, solid elements


19

are utilized to establish the FEM of the jellyroll. This simplification is contributed to

the errors in simulation results especially in the part from 0 mm to 4 mm.

(3) The electrolyte isn’t taken into consideration in the establishment of the FEM of

the jellyroll and its material property is set as isotropic. This simplification may also

lead to the errors in simulation results for the part after 4 mm.

Even though many factors contribute to the errors, overall simulation results for

three loading cases agree well with those corresponding experimental results.

Figure 2.9. Simulated and measured force-displacement curves on Li-ion cells (a) under rigid rod

indentation, SOC=0.2; (b) under rigid rod, SOC=0.5; (c) under a hemispherical punch, SOC=0.2;

(d) under a hemispherical punch, SOC=0.5.

2.4. Conclusions

In this chapter, the stiffness, electric and thermal characteristics of a 18650 Li-ion

battery under quasi-static loadings are analyzed and it is proposed that the


20

deformation process of a 18650 Li-ion battery can be divided into three stages,

corresponding to densification stage, microscopic damage stage and macroscopic

failure stage. The effect of the SOC of a Li-ion battery on its failure and

out-of-control behaviors after ISC under quasi-static loadings are investigated. The

constitutive model of the jellyroll of a 18650 Li-ion battery is proposed by

considering the microscopic damage. The proposed model can be used to predict the

mechanical responses on the trigger of ISC under various quasi-static loadings, which

can benefit to the safety of Li-ion batteries in EVs.

Chapter 3 Investigation of safety performance of cylindrical lithium-ion batteries under dynamic loadings

21

Chapter 3

Investigation of safety performance of cylindrical Li-ion

batteries under dynamic loadings In this chapter, compression tests at various loading rates are performed on two

types of 18650 Li-ion batteries. The effect of the loading rate on the mechanical

property of a 18650 Li-ion battery is investigated. The mechanical, electric and

thermal responses of a 18650 Li-ion battery under dynamic loadings are analyzed.

The failure and out-of-control behaviors of a 18650 Li-ion battery after ISC under

dynamic loadings are investigated. The constitutive model of the jellyroll of a Li-ion

battery is proposed to describe its dynamic behavior, which is then validated by the

explicit FEM of a Li-ion battery.

3.1. Analysis of mechanical, electric and thermal responses of Li-ion

batteries under dynamic loadings

Compression tests at various loading rates are carried out on NCM and NCA 18650

Li-ion batteries. The specifications of the NCM Li-ion battery are the same as those in

the chapter 2.1 and those of the NCA Li-ion battery are listed in Table 3.1.

Table 3.1. Specifications of NCA 18650 Li-ion battery

Items Specifications

Normal capacity 2500 mAh

Normal voltage 3.7 V

Charge voltage 4.2 0.05 V

Cut off voltage 2.5 V

Cathode Material LiNiCoAlO2

Firstly, two types of Li-ion batteries are compressed at various loading rates to

investigate their dynamic responses and all batteries are discharged to their cut-off

voltage to avoid the occurrence of fire or even explosion during tests. Compression

tests at high speed are performed by INSTRON VHS 8800 as shown in Fig. 3.1(a)


22

and those at low speed are performed by MTS EXCEED E45 100 KN as shown in Fig.

3.1(b). Moreover, the compression tests at high speed are recorded by a high-speed

camera (i.e. Phantom V2512), as shown in Fig. 3.1(a). The measured load-time

relations are all converted into load-displacement relations to easily compare the

mechanical responses at different loading rates.

Figure 3.1. Compression tests at different loading rates: (a) INSTRON VHS 8800; (b) MTS

EXCEED E45; (c) Measured mechanical responses for NCM Li-ion batteries at different loading

rates; (d) Measured mechanical responses for NCA Li-ion batteries at different loading rates.


23

The loading rate for NCM Li-ion battery includes 0.5 mm/min, 5 mm/min, 25

mm/min and 3 m/s, respectively, while that for NCA Li-ion battery includes 0.5

mm/min, 5 mm/min, 25 mm/min, 60 mm/min and 3 m/s, respectively. The reliability

of experimental data of compression tests between two plates is ensured by repeating

each test at least three times. The measured load-displacement curves for NCM Li-ion

battery compressed at different loading rates are shown in Fig. 3.1(c) and those for

NCA Li-ion battery are described in Fig. 3.1(d). There are several similarities between

the mechanical response of NCM and NCA Li-ion battery. Firstly, the general trend of

mechanical response of NCM Li-ion battery compressed at different loading rate is

almost the same as that of the NCA Li-ion battery. Secondly, the measured load for

NCM and NCA Li-ion battery both increases as the increase in the loading rate,

namely, the strain rate hardening behaviors. There are also some differences between

the mechanical response of NCM and NCA Li-ion battery. Firstly, the strain rate

hardening behavior of a NCM Li-ion battery can be obviously observed from

measured load-displacement curves at low loading rates corresponding to 0.5 mm/min,

5 mm/min and 25 mm/min. However, the measured load-displacement curve for NCA

Li-ion battery is almost the same at the loading rate of 0.5 mm/min, 5 mm/min and 25

mm/min. Their strain rate hardening behaviors can also be obviously observed when

the loading rate reaches 60 mm/min.

Figure 3.2. (a) INSTRON 5985; (b) FLUKE TI 400; (c) HIOKI 8880; (d) A NCA Li-ion battery at

the SOC of 100%.


24

Figure 3.3. (a) Measured load-time curve for NCM Li-ion battery at the SOC of 0.4; (b) Measured

voltage response; (c) Measured temperature response.

NCM and NCA Li-ion batteries at the SOC of 0.4 are compressed between two

plates to further investigate the effect of the differences in the cathode materials on

their safety performance. It should be noted that these compression tests are

performed by INSTRON 5985 and the loading rate is set as 60 mm/min, as shown in

Fig. 3.2 (a). During the compression tests, the battery temperature is recorded in situ

by FLUKE TI 400 as shown in Fig. 3.2 (b) while the battery voltage is measured in


25

situ by HIOKI MR 8880 as shown in Fig. 3.2 (c). The measured relations of load,

voltage and temperature versus time of NCM Li-ion batteries are shown in Fig. 3.3

and those measured relations of NCA Li-ion batteries are shown in Fig. 3.4.

Figure 3.4. (a) Measured Load-time curve for NCA Li-ion battery at the SOC of 0.4; (b)

Measured voltage response; (c) Measured temperature response.

For the NCM Li-ion battery compressed between two plates, the maximum force

occurs at 7.83 s and its measured value is 92.47 kN, as depicted in Fig. 3.3 (a). The

battery voltage firstly remains stable before the measured force reaching its maximum


26

value and rapidly decreases from the initial voltage of 3.64 V at 8 s to 2.12 V at 8.5 s

and then continues to decrease to 0.01 V at 10.5 s, as depicted in Fig. 3.3 (b). The

battery temperature also firstly maintains stable before the maximum force and slowly

increases from the initial temperature of 17.03 C at 8.88 s to its maximum

temperature of 102.78 C at 86.58 s and then starts to slowly decrease, as depicted

in Fig. 3.3 (c).

For the NCA Li-ion battery compressed between two plates, the maximum force

also occurs at 7.83 s while its measured value is 90.817 kN, as depicted in Fig. 3.4 (a).

The battery voltage also firstly maintains unchanged before the maximum force and

rapidly decreases from the initial voltage of 3.64 V at 8.3 s to 1.65 V at 8.85 s and

then continues to decrease to 0.07 V at 8.9 s, as depicted in Fig. 3.4 (b). The battery

temperature also firstly maintains unchanged before the maximum force and slowly

increases from the initial temperature of 20.38 C at 9.99 s to its maximum

temperature of 112.51 C at 74.37 s and then starts to slowly decrease, as depicted

in Fig. 3.4 (c).

Several conclusions can be obtained from the experimental results. Firstly, the ISC

is triggered at the peak force for NCM and NCA Li-ion battery at the SOC of 0.4

under dynamic loadings, but the thermal runaway isn’t triggered inside both batteries.

Secondly, for NCM and NCA Li-ion battery at the SOC of 0.4, the battery voltage

both rapidly decreases to almost zero while the battery temperature both slowly

increases to the maximum temperature. Thirdly, the general trends of electrical and

thermal responses of the NCM Li-ion battery are almost the same as those of the NCA

Li-ion battery during the compression tests at the loading rate of 60 mm/min except

that the maximum temperature for the NCA Li-ion battery after the occurrence of ISC

is slightly higher than that of the NCM Li-ion battery. This is due to that the normal

capacity of NCA Li-ion battery is 2500 mAh while that of NCM Li-ion battery is

2100 mAh.

The compression tests are also performed on NCM and NCA Li-ion batteries at the

fully charged state (i.e. the 100% SOC), which causes strong fire during the


27

compression tests as shown in Fig. 3.2 (d). Thus, the battery voltage and temperature

can’t be measured due to the strong fire. This indicates that catastrophic accidents

may be caused at the collision of EVs especially when the ISC is triggered at the fully

charged state. Hence, more researches should be carried out to investigate the safety

performance of Li-ion batteries at the fully charged state under dynamic loading.

3.2. Constitutive model for jellyroll under dynamic conditions

For NCM and NCA Li-ion batteries, their deformation processes under

compression test at high loading rate is similar to those at low loading rate, which can

be seen in Fig. 2.6. Moreover, Fig. 3.1 shows that the general trend of mechanical

behaviors of NCM and NCA Li-ion battery at the high and low loading rate is similar.

Hence, those measured load-displacement curves at different loading rates are fitted

by Gaussian functions.

21 1 1( ) exp( (( ) / ) )H w a w b c (3.1)

where 1a , 1b and 1c are the fitting parameters. Table 3.2 lists the value of those

fitting parameters corresponding to each loading rate and w is the compression

displacement in x direction.

Table 3.2. Fitting parameters 1a , 1b and 1c for NCM and NCA Li-ion batteries at various

loading rates.

NCM Li-ion batteries NCA Li-ion batteries

Loading rete mm/min m/s mm/min m/s

0.5 5 25 3 5 25 60 3

Coefficient 1a 74.99 75.96 87.11 61.67 102.8 95.43 90.95 72.18

Coefficient 1b 10.26 9.927 10.19 6.835 9.981 9.766 9.566 7.014

Coefficient 1c 3.771 3.56 3.836 2.377 3.028 3.019 3.105 2.087

The calculation process from the measured load-displacement relationship to the

stress-strain relationship is the same as those explained in chapter 2.2. The calculated

equivalent stress and strain for NCM and NCA Li-ion battery compressed at different


28

loading rates is shown in Figs. 3.5 and 3.6, respectively.

Figure 3.5. Calculated stress-strain curves for NCM Li-ion batteries at different loading rates.

Figure 3.6. Calculated stress-strain curves for NCA Li-ion batteries at different loading rates.

3.3. Validation and discussion

The established FEM in chapter 2.3 is also utilized to validate the constitutive

model of a Li-ion battery under dynamic loadings, which is proposed in chapter 3.2.

All settings used in this section are consistent with those described in chapter 2.3

except for the following aspects. Firstly, the dynamic loading rate in the experiment is

used in the simulation. Secondly, the calculated stress-strain relations shown in Fig.

3.5 and 3.6 are used as the input to crushable foam to model the mechanical responses

of NCM and NCA Li-ion batteries at different loading rates, respectively.

The comparisons between the simulation and experimental results for NCM and

NCA Li-ion batteries compressed at different loading rates are shown in Fig. 3.7 and


29

Fig. 3.8, respectively.

Figure 3.7. Simulated and measured force-displacement curves for NCM Li-ion batteries at

different loading rates: (a) v=0.5 mm/min; (b) v=5 mm/min; (c) v=25 mm/min; (d) v=3 m/s.

Overall, the simulation results agree well with the corresponding experimental results

for both NCM and NCA Li-ion batteries. For the NCM Li-ion battery compressed at

the loading rate of 0.5 mm/min, the measured and simulation force at 8 mm is 52.08

kN and 55.05 kN respectively, as depicted in Fig. 3.7 (a). For that at the loading rate

of 5 mm/min, the measured force at 8 mm is 56.928 kN and the corresponding

simulation force is 59.57 kN, as shown in Fig. 3.7 (b). For that at the loading rate of

25 mm/min, the measured force at 8 mm is 63.557 kN whereas the corresponding

simulation force is 67.69 kN, see Fig. 3.7 (c). For that at the loading rate of 3 m/s, the

measured force at 6.08 mm is 57.919 kN and that at 6.318 mm is 58.2 kN while the

simulation force at 6.0866 mm and 6.286 mm is 59.893 kN and 67.193 kN


30

respectively, as shown in Fig. 3.7 (d).

Figure 3.8. Simulated and measured force-displacement curves for NCA Li-ion batteries at

different loading rates: (a) v=5 mm/min; (b) v=25 mm/min; (c) v=60 mm/min; (d) v=3 m/s.

For the NCA Li-ion battery compressed at the loading rate of 5 mm/min, the

measured force at 8 mm is 66.489 kN while the simulation force is 8.324 kN, as

depicted in Fig. 3.8 (a). For that at the loading rate of 25 mm/min, the measured and

simulation force at 8 mm is 67.105 kN and 69.944 kN respectively, as shown in Fig.

3.8 (b). For that at the loading rate of 60 mm/min, the measured force at 8 mm is

70.819 kN and the corresponding simulation force is 74.541 kN, see Fig. 3.8 (c). For

that at the loading rate of 3 m/s, the measured force at 6.7 mm is 71.398 kN while the

simulation force at 6.714 mm is 74.904 kN, as shown in Fig. 3.8 (d).

The reasons listed in chapter 2.3 are also responsible for the errors between the

experimental and simulation results in this section, which are induced by the


31

simplification involved in the establishment of FEM. Moreover, the prediction

accuracy of simulation force at 6.3 mm for the NCM Li-ion battery compressed at the

loading rate of 3 m/s is only approximately 84%. For the NCA Li-ion battery

compressed at the loading rate of 3 m/s, the relative error between the measured force

from 4.5 mm to 6.5 mm and the corresponding simulation force is also large. The

reasons are as follows. Firstly, under compression tests at the loading rate of 3 m/s,

aluminum foil is covered on the top and the bottom plate respectively to prevent the

corrosion of electrolyte. It may cause a slight slip during the compression process.

Secondly, according to the recorded videos for the compression tests at the loading

rate of 3 m/s, the movement process of the Li-ion battery can be divided into three

stages. The Li-ion battery firstly moves with the bottom plate (i.e. the indenter) and it

then shoots from the bottom plate to the top plate due to its inertia. Finally, the bottom

plate starts to compress the Li-ion battery. Hence, some measurement errors may be

caused during the compression test. Nevertheless, the established FEM has

sufficiently good accuracy to simulate the mechanical responses of 18650 Li-ion

batteries under dynamic loadings.

3.4. Conclusions

In this chapter, the mechanical, electric and thermal responses of a 18650 Li-ion

battery under dynamic loadings are analyzed. The effect of the loading rate on the

mechanical property of a 18650 Li-ion battery is investigated. The effect of the

cathode materials on the mechanical, electric and thermal responses of a 18650 Li-ion

battery is also investigated. These investigations have helped to understand the

mechanism of the failure and out-of-control behaviors of a 18650 Li-ion battery after

ISC under dynamic loadings. The constitutive model of the jellyroll of a Li-ion

battery is proposed to describe its dynamic behavior and can be used to predict the

mechanical responses at the trigger of ISC under dynamic loadings, which can

provide valuable guidance for the structure design of battery pack for EVs.

Chapter 4 Investigation of Internal Short Circuits of Lithium-Ion Batteries under Mechanical Abusive Conditions

32

Chapter 4

Investigation of Internal Short Circuits of Li-ion Batteries

under Mechanical Abusive Conditions In this chapter, the material property and the damaged area of the short-circuit

object is taken into consideration to improve the electrochemical-electric-thermal

coupled model, which is then used to investigate the thermal responses of a Li-ion

battery under various ISC conditions and predict the triggering of thermal runway.

4.1 Investigating thermal responses of a Li-ion battery under various

ISC conditions

4.1.1 Improved electrochemical-electric-thermal model

COMSOL Multiphysics (version 5.2a) is utilized to establish an axisymmetric

electrochemical-electric-thermal coupled model to investigate the thermal response of

Li-ion battery under various ISC conditions. In this model, the electrolyte is 3:7 EC:

EMC LiPF6 and the material of positive and negative electrode is

LiyCo1/3Ni1/3Mn1/3O2 and LixC6, respectively. Table 4.1 and 4.2 list the governing

equations and parameters used in the model. In Table 4.1, qr is defined as reaction

heat produced by the reaction overpotential, qj represents ohmic heat generated by

charge transport in the electrodes and electrolyte and qe is defined as reversible heat

(i.e. entropy heat). Some parameters listed in Table 4.2 are taken from [48-50] and the

others are estimated by a trial-and-error approach through simulation.

The small ISC resistance will induce large short-circuit current when the ISC is

triggered inside a Li-ion battery. Hence, the ohmic heat generated by large

short-circuit current should be taken into consideration when establishing the coupled

model, which is calculated by the following equations.

2S S

SISC

I RqV

(4.12)


33

Table 4.1. Governing equations of the axisymmetric model.

Physical and chemical mechanisms Equations

Solid phase: Charge conservation ( )eff Lis j (4.1)

Electrolyte phase: Charge conservation ( ) ( ln )eff eff Lie D ek k c j (4.2)

Electrolyte phase: Conservation of Li+ species 0( ) 1( )eff Lie e

ec tD c jt F

(4.3)

Solid phase: Conservation of Li+ species 22 ( )s s sc D cr

t r r r

(4.4)

Energy conservation ( )

( )pe r j

c Tk T q q q

t

(4.5)

Electrochemical kinetics 0 exp[ ] exp[ ]Li a cs

F Fj a iRT RT

(4.6)

Overpotential s e U (4.7)

Exchange current density 0 ,max , ,( ) ( ) ( )a a ce s s e s ei k c c c c

(4.8)

Ohmic heat lneff eff effj s s e e e eq k k c (4.9)

Reaction heat ( )Lir s e jq j U (4.10)

Entropy heat ( )jLie

Uq j T

T

(4.11)

,s SS

s

IR

(4.13)

where ,s S represents the solid potential drop along the ISC object in the axial

direction, SR is defined as ISC resistance, and ISCV represents the volume of the ISC

object.

For a Li-ion battery under mechanical abusive conditions, the ISC resistance has a

strong correlation with the internal damaged area, which determines the short-circuit

current. The material property of the ISC object has some effects on the heat

dissipation. Hence, the above electrochemical-electric-thermal coupled model is

improved by considering the damaged area and the material property of the

short-circuit object. As discussed above, four types of ISC may be triggered inside the


34

Li-ion battery under mechanical abusive conditions corresponding to

negative–positive ISC, aluminum-negative ISC, copper-positive ISC, and

aluminum–copper ISC.

Table 4.2. Physiochemical parameters used in the axisymmetric model.

Parameter Unit Cu Foil Negative

Electrode

Separator Positive

Electrode

Al Foil

Density 3kg/cm 38.9 10

31.2 10 45.3 10 32.9 10 32.7 10

Specific heat J/(Kg K) 385 1150 2050 1150 900

Thermal conductivity W/(cm K) 4.0 0.004 0.005 0.004 2.38

Electron conductivity S/cm 55.8 10 1.0 0.1 53.54 10

Thickness cm 410 10 498 10 417 10 492 10 415 10

Particle radius cm 410 10 48 10

Initial electrolyte

concentration

3mol/cm 0.001

Porosity 0.4

These parameters used in the improved coupled model at the negative–positive ISC

are calculated as follows.

2inS r (4.14)

1 1pos negS

pos neg

L LR

S S (4.15)

( )pos pos neg neg

Spos neg

L S L SL L S

(4.16)

pos pos neg negS

pos neg

C m C mC

m m

(4.17)

pos nega

pos neg

pos neg

L Lk L L

k k

(4.18)


35

pos pos neg negr

pos neg

L k L kk

L L

(4.19)

The definitions of these parameters are listed in Table 4.3. It should be mentioned that

the ISC resistance mainly consists of two parts corresponding to the intrinsic and

contact resistance. It is difficult to determine the precise value of the contact

resistance especially for that involves in the ISC of Li-ion battery triggered under

mechanical abusive conditions due to the limitation of measuring techniques [48].

Hence, the ISC resistance only considers the intrinsic resistance in this chapter. Those

parameters under the other three ISC types are also calculated based on Equations

(4.14-4.19) and the values of those parameters are listed in Table 4.4.

Table 4.3. The definition of those parameters used in improved coupled model.

Parameter Definition Parameter Definition

L Thickness ak Thermal conductivity in the axial direction

m Mass of the ISC object rk Thermal conductivity in the radial direction

Density S Area of the ISC object

C Specific heat inr Radius of the area of the ISC object

k Thermal conductivity

Table 4.4. Parameters of different internal short circuit (ISC) types.

Parameter Unit Negative-Positive Al-Negative Cu-Positive Cu-Al

SR 0.1S

-39.8 10

S

-29.2 10S

-95.96 10S

S 3kg/cm 62003.8 10 61399 10 63458 10 -

SC J/(Kg K) 1150 1086 955.7 -

ak W/(cm K) 0.004 0.0046 0.0044 -

rk W/(cm K) 0.004 0.32 0.396 -

,s S V , ,s pos s neg , ,s al s neg , ,s pos s cu -

When ISC is triggered inside a battery, the un-shorted electrode layers and the

shorted electrode layers formed a closed loop current path, where the un-shorted

electrode layers were responsible to provide the energy to the shorted electrode layers


36

[48]. Hence, the temperature in shorted electrode layers is higher than that in the

un-shorted electrode layers inside the Li-ion battery after ISC and the established

axisymmetric electrochemical-electric-thermal coupled model only takes the shorted

electrode layers into consideration.

Only one shorted electrode layer is taken into consideration in the establishment of

the axisymmetric electrochemical-electric-thermal coupled model and its radius is 3

cm. This simplification is explained as follows. Firstly, it is considerably difficult to

solve a detailed electrochemical-electric-thermal model and even a single layer has a

high computation burden. Secondly, all electrode layers inside the Li-ion battery are

fairly equivalent in the aspects of heat and electricity [51]. Thirdly, one established

shorted electrode layer has a similar working principle a whole Li-ion battery since it

consists of one layer of aluminum and copper current collector, one layer of separator

and one layer of positive and negative electrode. The effect of the electron transport

may be underestimated by this simplification to a certain extent, but it is still within a

reasonable range. The surface heat transfer coefficient is set to 5 W/(m2k) since this

chapter also investigates the effect of the material property of the ISC object on the

thermal response. This boundary condition may have some slight differences with the

real condition since the heat dissipation is also affected by those neglected layers, but

the effect brought by this setting should not be exaggerated and it is still within a

reasonable range. Besides, it is considered that thermal runaway is triggered inside a

Li-ion battery when its internal temperature exceeds o120 C .

4.1.2 Effect of various ISC types on thermal responses

The established axisymmetric electrochemical-electric-thermal coupled model is

utilized to investigate thermal responses of a Li-ion battery under various ISC types,

where the damaged area of the ISC object (i.e. the ISC resistance) and the SOC of the

battery is the same. Table 4.4 shows that the resistance of the copper-aluminum ISC

corresponding to 95.96 10 S is much less than that value of the


37

aluminum-negative ISC corresponding to 39.8 10 S , which indicates that the

temperature rise of battery under copper-aluminum ISC is much more serious than

that under aluminum-negative ISC. Besides, recent research found that the ISC

process may be interrupted inside a Li-ion battery under copper-aluminum ISC due to

the melting of ISC current path induced by large Joule heat [50], which isn’t hence

investigated in this study.

Figure 4.1 shows the simulation thermal responses of a Li-ion battery under

different ISC types, where the SOC is 0.2 and the radius of the area of the ISC object

is 0.5 mm (i.e. rin=0.5 mm). The simulation thermal response under positive-negative

ISC is shown in Fig. 4.1 (a). It indicates that the temperature rise in the positive

electrode is similar to that in the negative electrode, which is higher than that in the

separator. For the simulation response under copper-positive ISC, the temperature rise

in the negative electrode is similar to that in the separator, which is slightly less than

that in the positive electrode, as shown in Fig. 4.1 (b). The simulation thermal

response under aluminum-negative ISC is described in Fig. 4.1 (c). The temperature

rise in the positive electrode is similar to that in the separator, which is less than that

in the negative electrode. The comparison between Fig. 4.1 (a) and (b) shows that the

temperature rise in the positive and negative electrode of a Li-ion battery under

positive–negative ISC is slightly higher than that under copper–positive ISC this is

due to the fact that the higher thermal conductivity of copper and higher resistance

under the positive-negative ISC than that under copper–positive ISC. Hence, the

influence of material property of the ISC object on the thermal response of a Li-ion

battery is important, which should be considered in the

electrochemical-electric-thermal coupled model.


38

Figure 4.1. Thermal responses of various ISC conditions for a Li-ion battery at SOC = 0.2 and

0 5 mm .inr . (a) Positive–negative, (b) copper-positive, and (c) aluminum-negative.

Several conclusions can be obtained from the simulation results of a Li-ion battery at

the SOC of 0.2 and the radius of the area of the ISC object is 0.5 mm, as shown in Fig.

4.1. Firstly, the thermal response of a Li-ion battery under the aluminum-negative ISC

is stronger and quicker than that under other two types of ISC. Secondly, the

maximum temperature occurs inside the negative electrode for a Li-ion battery under

aluminum-negative ISC while that occurs inside the positive electrode under other


39

two types of ISC. Hence, the hazard level of the aluminum-negative ISC is higher

than that of other two types of ISC since the initial temperature of these side reactions

occurring inside the negative electrode is lower than these inside the positive

electrode. The adhesion strength between the positive active materials and the

aluminum foil can be appropriately increased to improve the safety performance of

Li-ion batteries under mechanical abusive conditions. However, it should be noted

that plenty of glue for strengthening adhesion will degrade electrochemical

performance of Li-ion batteries.

4.1.3 Effect of the area of the ISC object on thermal responses

In this study, the established axisymmetric electrochemical-electric-thermal

coupled model is utilized to investigate thermal responses of the Li-ion battery

induced by different damaged areas of ISC object, where the SOC of the battery and

the type of ISC is the same as before. In general, the common ISC of a Li-ion battery

triggered under mechanical abusive conditions is the positive–negative ISC due to the

failure of separator. Moreover, the simulation results shown in Fig. 4.1 indicate that

the temperature rise in a Li-ion battery under the copper-positive ISC is slightly lower

than that under the positive–negative ISC. Hence, this section only investigates the

effects of different damaged areas of ISC object on the temperature rise in a Li-ion

battery under the positive–negative and aluminum-negative ISC.

Fig. 4.2 describes the simulation thermal response of a Li-ion battery at different

radii of the ISC object, where the SOC is 0.4 and the ISC type is the positive–negative

ISC. Fig. 4.2 (a) shows the simulation result for a Li-ion battery with the ISC object at

the radius of 0.5 mm, the temperature in the negative and positive electrode increases

from the initial value of o20 C to the maximum value of approximately o75 C

within 0.96 s. For a Li-ion battery with the ISC object at the radius of 1.5 mm, the

temperature in the negative and positive electrode increases from the initial value of o20 C to the maximum value of approximately o93 C within 0.17 s, as shown in

Fig. 4.2 (b).


40

Figure 4.2 Thermal response of various ISC areas, SOC = 0.4, positive–negative (a)

0.5 mminr , and (b) 1.5 mminr = .

The simulation results of a Li-ion battery at different radii of the ISC object is

depicted in Fig. 4.3, where the SOC is 0.4 and the ISC type is aluminum-negative ISC.

For a Li-ion battery with the ISC object at the radius of 0.5 mm, the temperature in

the negative and positive electrode increases from the initial value of o20 C to the

maximum value of approximately o153 C and o138 C within 0.47 s, respectively,

as shown in Fig. 4.3 (a). For a Li-ion battery with the ISC object at the radius of 1.5

mm, the temperature in the negative and positive electrode increases from the initial

value of o20 C to the maximum value of approximately o196 C within 0.1 s and o140 C within 0.13 s, respectively, as shown in Fig. 4.3 (b).

Several conclusions can be obtained from the simulation results shown in Fig. 4.2

and 4.3. Firstly, for a Li-ion battery under both two types of ISC, the temperature rise

inside a Li-ion battery with a large damaged area is higher and quicker than that with

a small damaged area. Secondly, for a Li-ion battery under the positive–negative ISC,


41

the thermal runaway may not be triggered inside the battery even with a large

damaged area, as shown in Fig. 4.2 (b), while the thermal runaway may be triggered

inside a Li-ion battery under the aluminum–negative ISC even with a small damaged

area, as shown in Fig. 4.3 (a). It accounts for the phenomenon observed during the

compression tests of 18650 Li-ion batteries that the thermal runaway may be triggered

inside a Li-ion battery compressed at a small displacement, while it may be not

triggered inside a Li-ion battery compressed at a large displacement.

Figure 4.3. Thermal response of various ISC areas, SOC = 0.4, aluminum-negative (a)

=0.5 mminr , and (b) =1.5 mminr .

4.1.4 Effect of the SOC of a Li-ion battery on thermal responses

The SOC of a Li-ion battery has positive correlation with its voltage and capacity.

Li-ion batteries in EVs may be at various SOCs at the time of vehicle collision. In this

study, the established axisymmetric electrochemical-electric-thermal coupled model is

utilized to investigate the thermal response of a Li-ion battery at various SOCs, where

the damaged area of the ISC object and the type of ISC is the same.


42

Fig. 4.4 describes the thermal response of Li-ion battery at different SOCs, where the

ISC type is positive–negative ISC and the radius of the area of the ISC object is 1.5 mm

(i.e. 1 5 mm .inr ). For a Li-ion battery at the SOC of 0.2, the temperature in the

negative and positive electrode increases from the initial value of o20 C to the

maximum value of approximately o71 C within 0.12 s, as shown in Fig. 4.4 (a). For a

Li-ion battery at the SOC of 0.6, the temperature in the negative and positive electrode

increases from the initial value of o20 C to the maximum value of approximately o181 C within 0.63 s, as shown in Fig. 4.4 (b).

Figure 4.4. Thermal responses of various capacities, =1.5 mminr , positive–negative (a) SOC = 0.2,

and (b) SOC = 0.6.

The comparison between the simulation results shown in Fig. 4.4, Fig. 4.2 and Fig.

4.1 (a) shows that the effect of the increase in the SOC of a Li-ion battery on its

thermal response is more important than the effect of the decrease in the ISC

resistance when the ISC resistance decreases to a pretty low value. It explains for the

phenomenon observed under the indentation tests that the thermal runaway isn’t


43

triggered inside the 18650 Li-ion battery at the SOC less than 0.4. The serious damage

inside the Li-ion battery is corresponding to a low ISC resistance when the internal

macroscopic fracture occurs under indentation tests.

Figure 4.5. Thermal response of various capacities, =1.5 mminr , aluminum-negative (a) SOC = 0.2,

and (b) SOC = 0.6.

Fig. 4.5 describes the thermal response of a Li-ion battery at different SOCs, where

the ISC type is aluminum–negative ISC and the radius of the area of the ISC object is

1.5 mm. For the Li-ion battery at the SOC of 0.2, the temperature in the aluminum

current collector and negative electrode increases from the initial value of o20 C to

the maximum value of approximately o153 C and o144 C within 0.07 s, as shown

in Fig. 4.5 (a). For the Li-ion battery at the SOC of 0.6, the temperature in the

aluminum current collector and negative electrode increases from the initial value of o20 C to the maximum value of approximately o424 C and o408 C within 0.36 s,

as shown in Fig. 4.5 (b).


44

4.2 Validation and discussion As discussed in chapter 4.1.1, some simplifications are made in the establishment of

the axisymmetric electrochemical-electric-thermal coupled model in the COMSOL

Multiphysics, which only considers one shorted layer of electrode. For the validation,

obviously the predicted temperature from this established model cannot compare with

the measured temperature from a whole Li-ion battery, however it can be used to

predict the thermal runaway and then compare with the measured temperature

triggering thermal runaway of a whole Li-ion battery. The prediction of triggering

thermal runaway inside the Li-ion battery under mechanical abusive conditions is the

focus for the safety of EV industries. In this study, three loading cases corresponding

to the compression test between two plates, the compression test under a rigid rod,

and compression test under a hemispherical punch are performed on 18650 Li-ion

batteries at the SOC of 0.2 and 0.6. The specifications of the Li-ion batteries and the

mechanical testing platform are the same as those in the chapter 2.1. An infrared

camera (FLUKE TI 400) is used to record the surface temperature of a 18650 Li-ion

battery to determine if the thermal runaway is triggered.

(a)

(b)

(c) Figure 4.6. Thermal responses under various loadings (unit: oC) at SOC = 0.2: (a) two rigid plates,

(b) a rigid rod, and (c) a hemispherical punch.

All three loading cases are performed on 18650 Li-ion batteries at the SOC of 0.2,

where the radius of the rigid rod is 12 mm and that of the hemispherical punch is 7

mm. Fig. 4.6 shows that the recorded thermal responses of the 18650 Li-ion batteries.


45

It should be mentioned that the left infrared image in Fig. 4.6 shows the maximum,

minimum, and average temperature of the Li-ion battery at the initial time of

macroscopic failure (i.e. peak force) under compression test while the right infrared

image in Fig. 4.6 shows the maximum, minimum, and average temperature of the

Li-ion battery at the time of reaching its maximum temperature. Moreover, the

moment of the trigger of a macroscopic failure inside the Li-ion battery is

corresponding to the initial time of recording temperature.

For a 18650 Li-ion battery compressed between two plates, its surface temperature

increases to the maximum value of o58.6 C after 300 s, as depicted in the right

infrared image in Fig. 4.6 (a). For a 18650 Li-ion battery compressed by a rigid rod,

its surface temperature increases to the maximum value of o92.5 C after 100 s, as

depicted in the right infrared image in Fig. 4.6 (b). For a 18650 Li-ion battery

compressed by a hemispherical punch, its surface temperature increases to the

maximum value of o81.5 C after 200 s, as depicted in the right infrared image in

Fig. 4.6 (c). These experimental results indicate that the thermal runaway isn’t

triggered inside the 18650 Li-ion battery at the SOC of 0.2 under all three loading

cases, they have a good agreement with the simulation results, where the thermal

runaway isn’t triggered inside the Li-ion battery no matter the sizes of the damaged

area and ISC types, as shown in Fig. 4.1, 4.4 (a) and 4.5 (a).

Figure 4.7. Thermal response of the battery compressed by a rigid rod (unit: oC) at SOC = 0.6.

For 18650 Li-ion batteries at the SOC of 0.6, the compression test under a rigid rod

is only carried out since the thermal responses of 18650 Li-ion batteries under three

loading cases are similar. Fig. 4.7 shows that the recorded thermal response of a

18650 Li-ion battery at the SOC of 0.6. The thermal runaway is triggered inside the

Li-ion battery only after 10 s, which simultaneously produces plenty of smoke, as

shown in the right infrared image of Fig. 4.7. It should be noted that the surface

temperature of 18650 Li-ion battery is obviously far higher than o90 C when the

thermal runaway is triggered. Moreover, the temperature measurement is also


46

influenced by the generated smoke. Hence, the recorded temperature (i.e. o90 C ) in

the right infrared image of Fig. 4.7 is the temperature of the generated hot smoke

instead of the surface temperature of 18650 Li-ion battery. This experimental result

indicates that the thermal runaway is triggered inside the 18650 Li-ion battery at the

SOC of 0.6 under compression tests. For simulation results, the thermal runaway is

also triggered inside the Li-ion battery at the SOC of 0.6, no matter the types of ISC,

as shown in Fig. 4.4 (b) and 4.5 (b). The good agreement between the simulation and

experimental results validates the established model.

4.3 Conclusions

In this chapter, an improved axisymmetric electrochemical-electric-thermal coupled

model is established by considering the material property and the damaged area of the

short-circuit object, which can be utilized to predict the thermal responses of a Li-ion

battery under various ISC conditions and then determine if the thermal runaway is

triggered.

Chapter 5 Conclusions and Expectations

47

Chapter 5

Conclusions and Expectations

5.1 Conclusions of this research

This research investigates the mechanical, electric and thermal responses of Li-ion

batteries under both quasi-static and dynamic loadings in chapters 2-3 and establishes

the corresponding constitutive model of jellyroll of Li-ion batteries. It also establishes

an improved electrochemical-electric-thermal coupled model to investigate the

thermal responses of Li-ion batteries under various ISC conditions in chapter 4.

In summary, this research work provides the prediction of the mechanical responses

and the mechanical safety tolerance corresponding to the trigger of ISC of a Li-ion

battery under various mechanical abusive conditions and the determination of the

trigger of thermal runaway induced inside a Li-ion battery after various ISCs. This

research forms a primary theoretical system for the analysis of mechanical property of

Li-ion batteries and provides the theoretical foundation for the safety design of Li-ion

battery systems in EVs. This will benefit to the promotion of the healthy and rapid

development of EV industries.

5.2 Innovations of this research

(1) It is found that the deformation process of a 18650 Li-ion battery can be divided

into three stages corresponding to densification stage, microscopic damage stage and

macroscopic failure stage. The constitutive model of the jellyroll is proposed by

considering microscopic damage, which can be utilized to evaluate the safety

performance of 18650 Li-ion batteries under quasi-static loadings.

(2) The constitutive model of the jellyroll suitable for dynamic loadings is proposed

based on the experimental results, which can be utilized to evaluate the safety

performance of 18650 Li-ion batteries under dynamic loadings

(3) The electrochemical-electric-thermal coupled model is improved by considering

the material property and the damaged area of the short-circuit object, which can be

utilized to determine if the thermal runaway is triggered inside a Li-ion battery under

Chapter 5 Conclusions and Expectations

48

various ISC conditions.

The innovations of this thesis provide valuable guidance for the structure design of

battery packs for EVs and significantly improve safety of Li-ion batteries in EVs.

5.3 Expectations in the future

This research has made some achievements on the investigation of the safety

performance of Li-ion batteries under mechanical abusive conditions. However, there

is still a lot of work to do in solving the safety problems of Li-ion batteries. The

followings are the list of some problems for future research:

(1) The constitutive model of the jellyroll suitable for dynamic loadings needs to be

improved based on the investigation of the mechanical behaviors of the

components of the jellyroll under dynamic loadings.

(2) The evolution and development of microscopic damage inside the Li-ion batteries

under various mechanical abusive conditions need to be further investigated.

(3) The failure criterion of Li-ion batteries needs to be proposed, which can be

utilized to determine the type and position of ISC inside the Li-ion batteries under

various mechanical abusive conditions.

(4) The electrochemical-electric-thermal coupled model of Li-ion batteries needs to

be improved with adding a mechanical model.

49

Bibliography

[1] John B, Margaret OM. Development of a driving cycle to evaluate the energy economy of

electric vehicles in urban areas. Applied Energy, 2016, 177: 165-178.

[2] Li Y, Davis C, Lukszo Z, Weijnen M. Electric vehicle charging in China's power system:

Energy, economic and environmental trade-offs and policy implications. Applied Energy,

2016, 173: 535-554.

[3] Zhao Y, Noori M, Tatari O. Vehicle to Grid regulation services of electric delivery trucks:

Economic and environmental benefit analysis. Applied Energy, 2016, 170: 161-175.

[4] Sun FC, Xiong R. A novel dual-scale cell state-of-charge estimation approach for

series-connected battery pack used in electric vehicles. Journal of Power Sources, 2015,

274: 582-594.

[5] Zhang S, Xiong R, Cao JY. Battery durability and longevity based power management for

plug-in hybrid electric vehicle with hybrid energy storage system. Applied energy, 2016,

179: 316-328.

[6] Sun FC, Xiong R, He HW. A systematic state-of-charge estimation framework for

multi-cell battery pack in electric vehicles using bias correction technique. Applied energy,

2016, 162: 1399-1409.

[7] Chen Z, Xiong R, Tian J. Model-based fault diagnosis approach on external short circuit

of lithium-ion battery used in electric vehicles. Applied energy, 2016, 184: 365-374.

[8] M. Li, Jun Lu, Zhongwei Chen, et al. 30 Years of Lithium-Ion Batteries. Advanced

materials, 2019, 30 (31): 1800561.

[9] Aiping Wang, Sanket Kadam, Hong Li, et al. Review on modeling of the anode solid

electrolyte interphase (SEI) for lithium-ion batteries. NPJ Computational materials, 2018,

4: UNSP 15.

[10] Winter M, Barnett B, Xu kang. Before Li-ion batteries. Chemical reviews, 2018, 118 (23):

11433-11456.

[11] Idrees M, Batool S, Kong J, et al. Polyborosilazane derived ceramics-Nitrogen sulfur dual

doped graphene nanocomposite anode for enhanced lithium ion batteries. Electrochemica

acta, 2019, 296: 925-937.

[12] Zeng ZQ, Murugesan V, Han KS, et al. Non-flammable electrolytes with high

50

salt-to-solvent ratios for Li-ion and Li-metal batteries. 2018, 3 (8): 674-681.

[13] Yao ZJ, Xia XH, Xie D, et al. Enhancing ultrafast Lithium ion storage of Li4Ti5O12 by

tailored TiC/C core/shell skeleton plus nitrogen doping. Advanced functional materials,

2018, 28 (31): 1802756.

[14] Griffith KJ, Wiaderek KM, Cibin G, et al. Niobium tungsten oxides for high-rate

lithium-ion energy storage. Nature, 2018, 559 (7715): 556.

[15] Liu J, Bao ZN, Cui Y, et al. Pathways for practical high-energy long-cycling lithium metal

batteries. Nature energy, 2019, 4 (3): 180-186.

[16] Jansen AN, Kahaian AJ, Kepler KD, et al. Development of a high-power lithium-ion

battery. J. Power Sources, 1999, 81: 902–905.

[17] Lee KJ, Smith K, Pesaran A, et al. Three dimensional thermal-, electrical-, and

electrochemical-coupled model for cylindrical wound large format lithium-ion batteries. J.

Power Source, 2013, 241: 20–32.

[18] Feng XN, Lu LG, Ouyang, MG, et al. A 3D thermal runaway propagation model for a

large format lithium-ion battery module. Energy, 2016, 115: 194–208.

[19] Maleki H, Deng GP, Anani A, et al. Thermal stability studies of Li-ion cells and

components. J Electrochem Soc, 1999, 146: 3224-3229.

[20] Maleki H, Deng GP, Kerzhner-Haller I, et al. Thermal stability studies of binder materials

in anodes for lithium-ion batteries. J Electrochem Soc, 2000, 147: 4470-4475.

[21] Biensan P, Simon B, Peres JP, et al. On safety of lithium-ion cells. J Power Sources, 1999,

81: 906-912.

[22] Gnanaraj JS, Zinigrad E, Asraf L, et al. The use of accelerating rate calorimetry (ARC) for

the study of the thermal reactions of Li-ion battery electrolyte solutions. J Power Sources,

2003, 119: 794-798.

[23] MacNeil DD, Lu ZH, Chen ZH, et al. A comparison of the electrode/electrolyte reaction at

elevated temperatures for various Li-ion battery cathodes. J Power Sources, 2002, 108:

8-14.

[24] Q. Wang, P. Ping, X. Zhao, et al. Thermal runaway caused fire and explosion of lithium

ion battery. J Power Sources, 2012, 208: 210–224.

[25] XN Feng, XM He, MG Ouyang, et al. Thermal runaway propagation model for designing

a safer battery pack with 25 Ah LiNixCoyMnzO2 large format lithium ion battery. Applied

51

energy, 2015, 154: 74-91.

[26] Andrey WG, David F, Julian W, et al. Thermal-runaway experiments on consumer Li-ion

batteries with metal-oxide and olivine-type cathodes. RSC advances, 2014, 4: 3633.

[27] PF Huang, P Ping, K Li, et al. Experimental and modeling analysis of thermal runaway

propagation over the large format energy storage battery module with Li4Ti5O12 anode.

Applied energy, 2016, 183: 659-673.

[28] Krishna S, Divya C, Ankur J. Experimental and theoretical analysis of a method to predict

thermal runaway in Li-ion cells. J. Power Source, 2016, 330: 167–174.

[29] GF Guo, B Long, B Cheng, et al. Three-dimensional thermal finite element modeling of

lithium-ion battery in thermal abuse application. J. Power Source, 2010, 195: 2393–2398.

[30] XN Feng, MG Ouyang, X Liu, et al. Thermal runaway mechanism of lithium ion battery

for electric vehicles: A review. Energy storage materials, 2018, 10: 246-267.

[31] A Sheidaei, X Xiao, X Huang, et al. Mechanical behavior of a battery separator in

electrolyte solutions. J Power Sources, 2011, 196: 8728-8734.

[32] X. Zhang, J.E. Zhu, E. Sahraei. Degradation of battery separators under charge-discharge

cycles. RSC Advances, 2017, 7: 56099-56107.

[33] Wierzbicki T, Sahraei E. Homogenized mechanical properties for the jellyroll of

cylindrical Lithium-ion cells. J Power Sources, 2013, 241: 467-476.

[34] X Zhang, W Shyy, AM Sastry. Numerical Simulation of Intercalation-Induced Stress in

Li-Ion Battery Electrode Particles. J Electrochem Soc, 2007, 154: A910-A916.

[35] Zhu JE, Zhang XW, Luo HL, et al. Investigation of the deformation mechanisms of the

lithium-ion battery components using in-situ micro tests. Appl Energy, 2018, 224:

251–266.

[36] C. Zhang, J. Xu, L. Cao, et al. Constitutive behavior and progressive mechanical failure of

electrodes in lithium-ion batteries. J Power Sources, 2017, 357: 126–137.

[37] B. Dixon, A. Mason, E. Sahraei, Effects of electrolyte, loading rate and location of

indentation on mechanical integrity of li-ion pouch cells. J Power Sources, 2018, 396:

412–420.

[38] Sahraei E, Hill R, Wierzbicki T. Calibration and finite element simulation of pouch

lithium-ion batteries for mechanical integrity. J Power Sources, 2012, 201: 307-321.

[39] HL. Luo, Y. Xia, Q. Zhou. Mechanical damage in a lithium-ion pouch cell under

52

indentation loads. J. Power Sources, 2017, 357: 61-70.

[40] J. Xu, B.H. Liu, X.Y. Wang, et al. Computational model of 18650 lithium-ion battery with

coupled strain rate and SOC dependencies. Applied Energy, 2016, 172: 180-189.

[41] J.E. Zhu, X.W. Zhang, S. Elham, et al. Deformation and failure mechanisms of 18650

battery cells under axial compression. J Power Sources, 2016, 336: 332-340.

[42] J. Xu, BH. Liu, DY. Hu. State of charge dependent mechanical integrity behavior of 18650

Lithium-ion batteries. Scientific reports, 2016, 6: 21829.

[43] W Li, Y Xia, JE Zhu, et al. State-of-charge dependence of mechanical response of

lithium-ion batteries: A result of internal stress. Journal of the electrochemical society,

2018, 165 (7): A1537-A1546.

[44] T. Kisters, E. Sahraei, T. Wierzbicki. Dynamic impact tests on lithium-ion cells.

International journal of impact engineering, 2017, 108: 205-216.

[45] XP Chen, T Wang, Y Zhang, et al. Dynamic mechanical behavior of prismatic lithium-ion

battery upon impact. International journal of energy research, 2019, DOI:

10.1002/er.4774.

[46] J. Xu, B.H. Liu, L.B. Wang, et al. Dynamic mechanical integrity of cylindrical lithium-ion

battery cell upon crushing. Engineering Failure Analysis, 2015, 53: 97-110.

[47] G. Kermani, E. Sahraei. Dynamic impact response of lithium-ion batteries, constitutive

properties and failure model. RSC Advances, 2019, 9: 2464-2473.

[48] Zhao W, Luo G, Wang CY. Modeling Internal Shorting Process in Large-Format Li-Ion

Cells. Journal of the electrochemical society, 2015, 162 (7): A1352–A1364.

[49] Zhang MX, Liu LS, Stefanopoulou A, et al. Fusing Phenomenon of Lithium-Ion Battery

Internal Short Circuit. Journal of the electrochemical society, 2017, 164 (12):

A2738–A2745.

[50] Zhao W, Luo G, Wang, CY. Modeling Nail Penetration Process in Large-Format Li-Ion

Cells. Journal of the electrochemical society, 2015, 162: A207–A217.

[51] Tommy GZ, Marten B, Goran L. Investigation of Short-Circuit Scenarios in a Lithium-Ion

Battery Cell. Journal of the electrochemical society, 2012, 159: A848–A859.

Safety research on the power Li-ion battery cells for electric vehicles · 2020. 9. 25. · I . Safety research on the power Li-ion battery cells for electric vehicles . Sheng Yang

Documents