Air and PCM cooling for

battery thermal management

considering battery cycle life
Chen, F., Huang, R., Wang, C., Yu, X., Liu, H., Wu, Q., Qian, K. &
Bhagat, R.
Author post-print (accepted) deposited by Coventry University’s Repository

Original citation & hyperlink:

Chen, F, Huang, R, Wang, C, Yu, X, Liu, H, Wu, Q, Qian, K & Bhagat, R 2020, 'Air and
PCM cooling for battery thermal management considering battery cycle life', Applied
Thermal Engineering, vol. 173, 115154.
https://dx.doi.org/10.1016/j.applthermaleng.2020.115154

DOI 10.1016/j.applthermaleng.2020.115154
ISSN 1359-4311

Publisher: Elsevier

NOTICE: this is the author’s version of a work that was accepted for publication in
Applied Thermal Engineering. Changes resulting from the publishing process, such
as peer review, editing, corrections, structural formatting, and other quality
control mechanisms may not be reflected in this document. Changes may have
been made to this work since it was submitted for publication. A definitive version
was subsequently published in Applied Thermal Engineering, 173, (2020)
DOI: 10.1016/j.applthermaleng.2020.115154

© 2020, Elsevier. Licensed under the Creative Commons Attribution-

NonCommercial-NoDerivatives 4.0 International
http://creativecommons.org/licenses/by-nc-nd/4.0/

Copyright © and Moral Rights are retained by the author(s) and/ or other copyright
owners. A copy can be downloaded for personal non-commercial research or study,
without prior permission or charge. This item cannot be reproduced or quoted extensively
from without first obtaining permission in writing from the copyright holder(s). The
content must not be changed in any way or sold commercially in any format or medium
without the formal permission of the copyright holders.

This document is the author’s post-print version, incorporating any revisions agreed during
the peer-review process. Some differences between the published version and this version
may remain and you are advised to consult the published version if you wish to cite from
it.
 Air and PCM Cooling for Battery Thermal Management Considering Battery

Cycle Life

Fenfang Chena, Rui Huanga, Chongming Wangc, Xiaoli Yua, Huijun Liua, b,
Qichao Wua, Keyu Qiana, Rohit Bhagatc

a
College of Energy Engineering, Zhejiang University, Hangzhou, China, 310027
b
Zotye Automotive Engineering Research Institute, Hangzhou, China, 310018
c
Institutes of Future Transport and Cities, Coventry University, Coventry, United Kingdom, CV1 5FB

 Corresponding author e-mail address: [email protected] (Dr Chongmign Wang); [email protected] (Prof Xiaoli Yu)
Highlights

⚫ Performance of air and PCM battery cooling is studied.

⚫ Realistic charge/discharge current profile is applied in simulations.
⚫ Battery non-uniform degradation and battery cycle life are analysed.
⚫ A performance index for assessment of battery thermal management system is proposed.
Abstract

Cycle life of lithium-ion batteries (LIBs) is essential for the application of hybrid electric vehicles
(HEV) and electric vehicles (EV). Since temperature greatly affects degradation rate and safety of LIBs,
battery thermal management system (BTMS) is required. In this paper, the performance of active air
cooling and passive phase change material (PCM) cooling for BTMS are assessed in terms of battery
thermal states and cycle life. A coupled one-dimensional electrochemical and two-dimensional thermal
models are developed to simulate the temperature of a battery module with 16 cylindrical (26650)
graphite-LiFePO4 lithium-ion battery cells. The model is validated with the experimental data taken
from literature. By applying a realistic current profile of a HEV to the battery model, simulations are
performed at various ambient temperatures, inlet velocities of air cooling and PCM phase change
temperatures. The battery cycle life and its non-uniformity across the module are estimated with a
battery degradation model with inputs of battery temperature results. The study shows that active air
cooling has a better cooling effect than PCM cooling, especially at high ambient temperatures. But the
active air cooling leads to a large temperature non-uniformity at low inlet air velocities. The cycle life
of the battery module under air cooling is longer than that of PCM cooling, although a larger life non-
uniformity is observed. Furthermore, two methods are compared by a newly proposed evaluation index
called cyclical cost. This index considers both the battery cycle life and the parasitic power
consumption of the BTMS. The result demonstrates that air cooling has a lower cyclical cost than PCM
cooling. When the inlet velocity of the air cooling system increases, the cyclical cost has a trend of
decreasing first and then increasing. This paper provides a guide for the development of BTMS to
further prolong the cycle life and reduce total operating cost of LIBs.

Keywords: Lithium-ion battery; thermal management; air cooling; PCM cooling; cycle life
Greek symbols
α Transfer coefficient
γ Bruggeman tortuosity exponent
δ Cyclical cost (€ cycle-1)
ε Volume fraction
ζN Maximum cycle life difference rate
η Local surface over potential (V)
θ Capacity loss (%)
λ Thermal conductivity (W m-1 K-1)
ρ Density (kg m-3)
σ Electronic/ionic conductivity (S m-1)
τ Price (€)
Φ Potential (V)
∆ Change
∇ Divergence

Symbols
Acell Area of the positive electrode (both sides) (m 2)
c Lithium-ion concentration (mol m-3)
Cp,i Heat capacity (J kg-1 K-1)
Di Diffusion coefficient (m2 s−1)
F Faraday’s constant (C mol−1)
h Heat transfer coefficient (W m-2 K-1)
H Enthalpy (J kg-1)
Hfuel Lower heating value of the fuel (J L-1)
I Current (A)
i Current density (A m−2)
jn Local charge transfer current density (A m−2)
j0 Exchange current density (A m−2)
k0 Reaction rate constant (m2.5 mol-0.5 s-1)
Li Thickness of component (m)
N Number of cycle before end of life
̅
N Mean cycle life of the module
ΔN Maximum cycle life difference across the module
P Power (W)
p Pressure (Pa)
q Total heat generation (J m-3)
qact Active heat generation (J m-3)
qohm Ohmic heat generation (J m-3)
qrea Reaction heat generation (J m-3)
r Radius distance variable of particle (m)
R Characteristic radius of electrode particles (m)
Rgas Gas constant (J mol−2 K−1)
Sa Specific surface area (m-1)
t Time (s)
t+ Li+ transference number
T Temperature (℃)
ΔT Maximum temperature difference (℃)
Ueq Open circuit potential of the electrode (V)
Ui Open circuit voltage (V)
v Thermodynamic factor
Vin Inlet velocity (m s-1)

Subscripts
0 Initial or equilibrated state
1 Solid phase
2 Liquid/electrolyte phase
n Negative electrode
s Separator
p Positive electrode
air Air
amb Ambient
AC Air conditioner
batt Battery
eff Effective value
fan Fan
fuel Fuel
max Maximum
pcm Phase change material
ref Reference value
surf Surface of active material

Abbreviations
BTMS Battery thermal management system
EV Electric vehicle
HEV Hybrid electric vehicle
LIB Lithium ion battery
PCM Phase change material
PCT Phase change temperature
SOC State of charge
1. Introduction
Air pollution, global climate change and fossil energy shortage are the major driven forces behind
the development of the energy and transportation sectors. In the automobile industry, traditional engine-
powered vehicles are gradually replaced by hybrid electric vehicles (HEVs) and electric vehicles (EVs)
[1]. For the energy storage system in HEVs/EVs, lithium-ion batteries (LIBs) are widely used for its
high voltage, high energy density, and low self-discharge rate.
Even though LIBs are the best energy storage system so far, they still have some drawbacks, such
as degradation [2], low energy density compared with internal combustion engines [3] and high cost
[4]. Low energy density limits the driving mileage of HEV/EVs. Degradation further reduces driving
mileage and causes range anxiety, and it also increases the operating cost of HEV/EVs if it needs
replacement. Approximately 50% of the vehicle cost is from the battery pack. In addition, recycling of
LIBs is another challenging issue that has not been solved [5]. Therefore, reducing battery degradation
rate and improving the cycle life of LIBs are of great relevance for future clean transportation. There
are two research directions in this area. One is technology advancement in battery cells, such as new
material [6], material modification [7] and the improvement of manufacture technology. The other is
battery management, including but not limited to, online fault diagnosis [8,9], equalization control
[10,11] and thermal management. Battery thermal management has been received increasing attention
since the temperature is one of the most sensitive factors to battery cycle life. More importantly, high
temperature may lead to the risk of battery thermal runaway.
Studies of battery thermal management have been carried out to control the temperature of LIBs,
from basic air cooling systems to compound systems like PCM and heat pipe compound battery thermal
management system (BTMS) [12–16]. In addition, novel thermal management methods are reported
[17,18]. The main goal of these studies is to control the temperature of LIBs within the safety range
for avoiding battery thermal runaway. However, even the temperature of the LIBs has been controlled
under the threshold of the thermal runaway, the temperature still has a great impact on battery cycle
life. Experimental tests [19] show that the cycle life of the battery cell reduced by 50% if the
temperature rises from 25 C to 45 C [19]. Neubauer et al. [20] of National Renewable Energy
Laboratory examined the sensitivity of electric vehicle battery pack degradation to battery thermal
management. The battery life was estimated when it was cooled by a passive system and a powerful
liquid cooling system with three different operation modes (key-on, standby and combined). The results
show that applying liquid cooling systems can significantly reduce capacity loss and resistance growth
rate, and combined key-on and standby system offered best reduction in degradation. Yuksel et al. [21]
assessed the influence of thermal management on battery degradation. The battery life was simulated
under the conditions of with and without air-cooling system. Applying an air-cooling system doubled
life compared with no BTMS.
However, the existing studies looked mainly at the impacts and sensitivities of various factors on
battery life, and thermal management is one of the many factors. Heat generation and heat transfer
processes were modelled by using lumped parameter method, which only solved the average
temperature of the whole battery pack rather than temperature distributions. Therefore, indexes of
temperature non-uniformity and degradation non-uniformity were not obtained. Meanwhile, BTMS
investigated in these studies were limited to several specific scenarios, and the influences of BTMS
parameters on battery cycle life has not been studied.
The novelty of this paper is that it quantitatively evaluates the degradation characteristics,
including cycle life and its non-uniformity across the module, of the battery module under different
cooling configurations. There is limited literature covering this area, however, it is of great significance
for BTMS. Since extending battery cycle life is one of the purposes of battery thermal management,
the authors think that battery life should must be considered as one of the indexes of BTMS. However,
the previous studies mainly take battery maximum temperature, temperature non-uniformity and power
consumption as evaluation indexes for BTMS. Hence, this study proposed a battery health-conscious
evaluation index for BTMS, which considers both power consumption of BTMS and cycle life of
batteries. The simulation results and newly proposed evaluation index can provide new ideas for battery
health-conscious thermal management.
In this study, a one-dimensional electrochemical model is developed for the lithium-ion battery
cell in COMSOL, a multi-physics modelling software, where the heat generation can be solved
accurately. Then, a two-dimensional thermal model is developed in COMSOL to simulate the heat
transfer processes in the battery pack. The electrochemical and thermal models are coupled together
via heat generation and temperature. The model is validated using experimental data from the literature.
The simulation is carried out by using realistic current load and thermal management methods of active
air cooling and passive phase change material (PCM) cooling are involved. On this basis, the time-
varying temperature field of the module can be obtained. The degradation process of the battery cell is
assessed by a battery capacity loss model. The effects of ambient temperatures, inlet velocities of air
cooling and PCM phase change temperatures are studied.
2. Model development
2.1 Electrochemical model
In this study, a one-dimensional electrochemical model developed by Doyle and Newman [22,23]
is used to predict heat generation of LIBs. This model is widely published in literature [24–26] and will
therefore not be repeated here. The governing equations are provided in Table 1 for completeness. A
cylindrical type of the graphite-LiFePO4 battery manufactured by A123 system (ANR26650M1A) is
chosen in this study. Fig. 1(a) and Fig. 1(b) show the internal structure of the battery cell and the
schematic of the computational domain of the one-dimensional electrochemical model, respectively.
The positive and negative current collectors are neglected in our study. Geometrical and design
parameters, kinetic and transport properties, and thermal properties of the battery used for modelling
are listed in Table 2.

Fig. 1 Internal structure of the battery cell (a) [27] and the schematic of the computational domain (b)

Table 1: Governing equations for the 1-D electrochemical model

Description Governing equations
Electrochemical kinetics    F    F  
jn = j0 exp  a   − exp  − c   
 RgasT   RgasT 
     

j0 = Fk0c2a ( c1,max − c1, surf )

a
c1,surf
c

 =1 -2 -Ueq

U eq ,i
U eq ,i =U ref ,i +
T
(T − T )ref

Si ( SOC )
= U ref ,i ( SOC , Tref ) + (T − Tref )
F
 ci
SOC =
ci ,max

Charge conservation
  i1 +   i2 = 0

  i1 = − Sa jn

i2 = Sa jn

i1 = − 1eff 1

2 RgasT 2eff   ln f  
i2 = − 2 + 1 +  (1 − t+ )  ( ln c2 )
eff

  ln c2 
2
F

Mass conservation
c1 1   2 c1 
= −r D1
t r 2 r  r 

c2 S j
2 +   J2 = a n
t F
c2 S j
2 +   J2 = a n
t F
Energy conservation U i
qrea = S a jnT
T
qact = Sa jn

qohm = −i1 1 − i2 2

q = qrea + qact + qohm

Table 2: Main parameter of ANR26650M1A type LiFePO4 cylindrical battery

Parameter Positive Separator Negative
nominal capacity (Ah) 2.3
Cell radius (m) 0.013
Cell height (m) 0.065
2
Acell (m ) 0.18a
Li (m) 8.0×10-5a 2.5×10-5a 3.4×10-5a
Ri (m) 5×10-8a - 5×10-6a
ε1,i 0.374 a
- 0.58a
ε2,i 0.444 a
0.54b 0.357a
c1,i (mol m-3) 22806×0.03b 30555×0.8b
c2,0 (mol m-3) 1200b 1200b 1200b
c1,max (mol m-3) 22806a - 30555a
αa, αc 0.5 a
- 0.5a
γi 1.5 a
1.5a 1.5a
 Uref,i Fig. 1A(a) c Fig. 1A(a) c
ΔSi Fig. 1A(b) d Fig. 1A(b) e
σ1 (S m-1) 0.5b - 100b
σ2 (S m-1) Fig. 2A a
 39000  1 1 
D1,p (m2 s-1) D1, p (T ) = 1.18  10−18 exp   −  a,b

 R  Tref T  

 35000  1 1 
D1,n (m2 s-1) D1,n (T ) = 3.9  10−14 exp   −  a,b

 R  Tref T  
D2 (m2 s-1) 2.0×10-10a
 30000  1 1 
k0,p (m2.5 mol-0.5 s-1) k0, p (T ) = 3  10−11  exp   −  b

 R  Tref T  

 20000  1 1 
k0,n (m2.5 mol-0.5 s-1) k0,n (T ) = 1.4  10−12  exp   −  b

 R  Tref T  

v (
 =0.601 − 0.24 10−3 c + 0.982 1 − 0.0052 (T − 294.0) 10−9 c3 )
t+ 0.363b
Thermal conductivity (W m-1 K-1) 0.2b
Heat Capacity (J kg-1 K-1) 1100a
Density (kg m-3) 2047a
Tref (K) 298.15
Rgas (J mol−2 K−1) 8.314
F (C mol-1) 96,487
a
Ref.[25]; b Ref.[26]; c Ref.[28]; d
Ref.[29]; e
Ref.[30]

2.2 Thermal model

For the simplicity of numerical simulation, thermal radiation in the whole model is neglected.
Besides, because there is no liquid motion existing in paraffin-graphite composite (thus, its viscosity
is regarded as 105 kg m-1 s-1) [31], the momentum equation is not considered for the melted PCM
[31,32]. Therefore, the only one governing equation of both battery cell and PCM domains is the energy
conservation equation:
T
 i C p ,i =  ( i T ) + qi (1)
t
where the subscript i refers to battery cell or PCM; qi is the heat source term. In the domain of battery
cells, qi is the total heat generation coming from the 1-D electrochemical model. The average
temperature of the battery cell calculated by Eq.(1) is used for 1-D electrochemical model. In the
domain of PCM, the value of qi is 0. Equivalent heat capacity method is employed to take the melting
latent heat of PCM into account, with which the apparent heat capacity of the PCM during phase
transition is defined as:

C p, pcm = Ceq + CL (T ) (2)

where Ceq is the equivalent heat capacity of the solid-liquid mixture; CL(T) is the latent heat distribution
within phase transition interval.
In the domain of cooling air, fluid flow and heat transfer are governed by the continuity,
momentum conservation and energy conservation equations:
 air
+   (  air v ) = 0 (2)
t

 ( air v )
+   ( air vv ) = −p (3)
t

 (  air C p ,air T )
+   (  air C p ,air vT ) =   ( air T ) (4)
t

Temperature and heat flux are continuous at the interface of PCM/cell and air/cell, where the
energy conservation equation can be expressed as:

T T
batt = pcm (5)
n n

T
batt =h(Tbatt − Tair ) (6)
n

In PCM cooling condition, PCM is cooled by ambient air in the way of natural convection, the
governing equation is:

T
pcm =hnatural (Tpcm − Tamb ) (7)
n

2.3 Battery Degradation model

The battery degradation is reflected in capacity loss and impedance rise. Here, we only consider
the capacity loss of LIBs. Wang et al. [33] developed a capacity loss model for ANR26650M1A (the
same with the one studied in this paper) by carrying out a serious of experiments. In their model, the
relationship between capacity loss and the temperature is formulated with an Arrhenius type equation.
The generic capacity degradation model given by Wang et al. is expressed as follows:
  −31700 + 370.3  ( I CELL / Cnom
CELL
)  Ah0.55
 CYC =A  exp  (8)
 R T CELL

 
whereθCYC is the capacity loss; A is a constant for given current and its values are listed in Table 3;

ICELL is the charge/discharge current; CCELL

nom is the nominal capacity of the cell; I
CELL CELL
/Cnom is the
charge/discharge rate (Crate); TCELL is the cell temperature and Ah is the total Ah-throughput.
In the dynamic cycling assessment of this study, it is assumed that during a very short time step
from t to t+Δt, the charge/discharge current and cell temperature remain unchanged. The accumulated
capacity loss at a time t+Δt can be expressed as follows:
0.55
 CYC 0.55 1

  t 
t +t =t +t  
CYC
+ Aht +t  (9)
 t +t  
 

 −31700 + 370.3  CrateCELL

,t +t

t +t =At +t  exp   (10)
 R  T CELL
 t +t 

where ΔAht+Δt is the ampere-hour processed during time step Δt, which can be calculated as:

( +t | )  t
t +t 1 CELL
Aht +t =  I CELL ( t )  dt  | It | + | I tCELL (11)
t 2
At+Δt is considered as a constant in each time step Δt, the value of which is estimated by linear
interpolation between the tabulated values in Table 3.
Table 3: Values of coefficient A in Eq.(8)

Crate 1/2 2 6 10
𝐴 31,630 21,681 12,934 15,512

In addition, during the capacity loss assessment, we also have following simplifications: (1) The
ageing mechanisms are the same during charge and discharge; (2) The capacity loss during storage
(calendar life) is neglected.
In the process of simulation, the temperature profile of each cell in dynamic cycling can be
calculated by the thermal model, and current profile is the input parameter of the electrochemical model.
Based on temperature and current profiles, the accumulated capacity loss of cells can be solved by
using the abovementioned capacity degradation model.
2.4 Model Validation

Fig. 2 Comparison of simulated results with the experimental data from literature: (a) Cell voltage
and (b) surface temperature.
Fig. 2 shows the simulated results of cell voltage and surface mean temperature, along with
experimental results published in Ref. [25]. The boundary conditions used in the simulation are the
same with experimental conditions. Natural heat transfer coefficient (hnatural) of the cell surface is set
as 10 W m-2 K-1 [25,34]. Fig. 2 shows that experimental data and simulated results match well,
indicating the electro-thermal model is able to simulate battery thermal performance at different
discharge rate.

3. Cooling configurations and simulation setup

Fig. 3 shows a battery module with 16 cells is selected for this study. In the air cooling BTMS,
the cooling air is sent from the left side of the module (inlet) to the right side (outlet) through the gaps
between battery cells and an optimized gap size of 7 mm is adopted according to Ref. [35]. In the PCM
based BTMS, the gaps are filled with PCM. In general, the mass of the PCM can be estimated by the
equation as follows [36]:

Qbatt = mpcm  C p , pcm  ( PCT − Tinitial ) + mpcm  L (13)

where Qbatt is the total heat generation of the battery, mpcm, PCT, Tinitial, L is the mass, phase change
temperature, initial temperature and latent heat of PCM, respectively. Because the value of Cp,pcm is
much lower than L and Tinitial (Tamb) is one of the variables, the first item on the right side of Eq. (14)
is not considered. Then a minimum mpcm of 24.02 g cell-1 (corresponding gap size between cells is
about 5 mm) can be obtained according to the battery heat generation and thermo-physical properties
of PCM (see below). On this basis, the PCM based BTMS is assumed to have a same gaps size (7 mm)
as the air cooling BTMS, which makes it work better.
A grid independence study is carried out to decide a proper element size. The temperature of a
probe point on battery surface and total heat rate at the outlet are used to test the grid independence.
Fig. 4 shows that the values of the two selected parameters. It is found that the results stabilize when
the grid number reaches 27776, corresponding element size of 1 mm. Thus, element size of 1 mm is
used in later study. Fig. 5(a) and Fig. 5(b) show the overall grid structure of the computational domain
and the refined mesh at the interface of battery cells and cooling air.

Fig. 3 Geometry of the battery pack

Fig. 4 Grid independent study

Fig. 5 2D Mesh of the battery module: (a) overall grid structure and (b) refined mesh at the interface
of air and the battery cell.
Fig. 6 shows the current profile of ANR26650M1A selected in this study. This profile is the
simulation result based on a hybrid electric bus and the total driving time and distance of the cycle are
approximately 3000 s and 16.55 km, respectively [37]. It’s defined that the discharge current is positive.

The average heat generation rate under this load at 30 ℃ is 1.43 W cell-1, which is equal to that under

4.25 C-constant current discharge.

Fig. 6 Current profile of battery cells obtained from literature (Ref. [37])
In the study of air cooling, the battery pack is cooled by either 20 C air-conditioned air when the
ambient temperature (Tamb) is higher than 20 C, or directly cooled by ambient air when Tamb is equal
to or below 20 C. The inlet velocity (Vin) is swept from 0 to 8 m s-1, and Tamb swept from 0 and 50 C.
Table 4 lists the simulation boundary conditions/settings for air cooling cases.
Table 4: Simulation setup for air cooling cases
Type Setting
Fluid dynamic conditions Turbulence
Turbulence model k-epsilon [35,38]
 Inlet velocity [m s-1] 0, 0.5, 1, 2, 3, 4, 6, 8
Tamb > 20 C; Tin = 20 C
Inlet temperature [C]
Tamb ≤ 20 C; Tin = Tamb
Outlet gauge pressure [Pa] 0
Wall of the module Symmetry
Wall of the cell Fluid-solid coupling
Ambient temperature [C] 0, 10, 20, 30, 40, 50
Instead of taking heat away from the module by cooling medium in active BTMS, passive PCM
based BTMS controls the battery temperature by accumulating or discharging heat of PCM and only
part of the heat is dissipated to the ambient through sides of the module. Therefore, thermo-physical
properties of the PCM and heat transfer boundary conditions of the module sides are important. In this
study, a paraffin-graphite composite PCM [39] with relatively large thermal conductivity is chosen
[40,41], which helps to dissipate the heat from the surface of battery cells to the side of the module and
achieve an uniform temperature field. Two sides of the module are modelled as natural convection air
cooling with a hnatural of 10 W m-2 K-1 [25,34] and the others are symmetry. Then, the effect of PCT to
battery thermal states and life is investigated, whilst the other thermo-physical properties of PCM, such
as latent heat, specific heat capacity and transition interval, are kept constant. The Tamb is swept from
0 C to 50 C. Table 5 and Table 6 list the boundary conditions of the model and thermo-physical
properties of PCM, respectively.
Table 5: Boundary conditions for PCM cooling cases
Type Boundary condition
Sides of the module Symmetry; natural heat transfer
Heat transfer coefficient [W m-2 K-1] 10 [25,34]
Wall of the cell Conduction
Ambient temperature [C] 0, 10, 20, 30, 40, 50

Table 6: Thermo-physical properties of PCM [39]

parameter Value/range
Latent heat [J kg-1] 181000
Specific heat capacity [J kg-1 K-1] 1980
Thermal conductivity [W m-1 K-1] 16.6
Density [kg m-3] 866
Transition interval [C] 3
Phase change temperature [C] 35, 40, 45, 50, 55
4. Results and discussion
The results and discussion are divided to three sub-sections, including the results of air cooling
and PCM cooling, and the comparison of air cooling and PCM cooling.
4.1 Air cooling

Fig. 7 Temperature fields of battery module at the end of the driving cycle
Fig. 8 Module mean temperature curves through the whole cycle: (a) Tamb=30 C and (b) Tamb=50 C
Fig. 7 and Fig. 8 and presents the temperature fields of battery module at the end of the driving
cycle and the mean temperature of all cells in the module throughout the whole cycle at different Tamb
and Vin. Due to the dynamic changing charge/discharge rates, the module mean temperature fluctuates
significantly, but it is always kept between 20 C and 40 C, even when Vin reduces to 0.5 m s-1, which
is attributed to the using of 20 C air-conditioned air. The red dash line represents the case where Vin =
0 m s-1 and only natural convection occurs. It indicates that the module will be overheated, leading to
the risk of thermal runaway.
Fig. 9 Mean temperature curves of selected cells in the module: (a) Tamb = 30 C, Vin = 0.5 m s-1; (b)
Tamb = 30 C, Vin = 8 m s-1; (c) Tamb = 50 C, Vin = 0.5 m s-1; and (d) Tamb = 50 C, Vin = 8 m s-1
Fig. 9 plots the mean temperature curves of different cells within the module in different
conditions. The code of cells in the module is shown in Fig. 3. The maximum temperature difference
between cells at Vin = 0.5 m s-1 is more than 2 C in most of time, whereas the difference is minor at
Vin = 8 m s-1.
 Fig. 10 Module temperature at different Tamb and Vin: (a) Mean temperature and (b) maximum
temperature difference
Fig. 10(a) shows the time-averaged temperature of the module through the whole cycle decreases
with the decrease of Tamb and the increase of Vin. There is little difference in temperature when Vin
exceeds 0.5 m s-1, indicating that the cooling efficiency is reduced.
Fig. 10(b) shows the maximum temperature difference (ΔT) between cells in the module. The
maximum temperature difference (ΔT) is defined by:

T = Tmax − Tmin (12)

where T̅ max and T̅ min are the maximum and minimum space and time-averaged cell temperature
among the module, respectively. It can be seen that ΔT increases first and then decreases when Vin
increases from 0 m s-1 to 8 m s-1. The maximum ΔT is observed at Tamb = 50 C and Vin = 0.5 m s-1.
Fig. 11 shows the trajectories of capacity loss predicted by the degradation model [33]. The battery
life is defined as the number of cycles that lead to a 20% capacity loss.
̅ ), maximum cycle life difference (ΔN) and maximum cycle life
We use the mean cycle life (N
difference rate (ζN) to evaluate the degradation of the module, which are given by the following
equations:
16

N i
N= i =1
(13)
16

N = Nmax − Nmin (14)

 N
N= (15)
N
where Ni is the cycle life of each cell in the module, Nmax and Nmin are the maximum and minimum
cycle life of the cell in the module.

Fig. 11 Capacity loss trajectories: (a) Tamb = 30 C and (b) Tamb =50C.

Fig. 12 Cycle life of the module at different Tamb and Vin: (a) Mean cycle life, (b) maximum cycle life
difference and (c) maximum cycle life difference rate.
̅ at different Tamb and Vin. Combining the results of Fig. 12(a) and Fig. 10,
Fig. 12(a) illustrates N
we can see that a high temperature leads to an enlarged capacity loss rate. Fig. 12(b) shows that ∆N is
the highest at Tamb = 0 C and Vin = 0.5 m s-1, and it decreases with the increase of Tamb and Vin. Since
the temperature distribution at Vin = 0 m s-1 is relatively uniform, ΔN is minor. However, ΔT is lowest
at Tamb = 0 C and it reaches the highest at Tamb = 50 C, which shows the opposite trend with ΔN. This
is because battery capacity loss and temperature follows an Arrhenius type relation, meaning that the
degradation rate is more sensitive to temperature and little temperature difference will lead to large
degradation non-uniformity in a cold environment. Fig. 13(a) demonstrates abovementioned rule by
plotting simulation results of ΔT and ΔN of this study. Results obtained at same Tamb, which also reflects
different temperature levels of battery module, are plotted together.
Fig. 12(c) shows ζN varies little at different Tamb if Vin remains unchanged. As expected, with
increased Vin, ζN shows a trend of increasing first and decreasing afterwards. The maximum ζN is
observed at Vin = 0.5 m s-1 and it is higher than 20%. It should be noted that ζN at a low Tamb is not
significantly larger than the other conditions due to prolonged cycle life, though its ΔT is the largest.
Fig. 13(b) plots the relationships between ΔT and ζN at different Tamb. It shows that a small ΔT will lead
to a large ζN at the low-temperature condition.

Fig. 13 Relationship between temperature difference and degradation non-uniformity at different

temperatures: (a) Effects of ΔT to ΔN and (b) effects of ΔT to ζN.
4.2 PCM cooling

Fig. 14 Temperature fields of the module at the end of the driving cycle

Fig. 15 Mean temperature curve of the module through the whole cycle at different PCT: (a) Tamb =
30 C and (b) Tamb = 50 C.
Fig. 14 and Fig. 15 show the temperature fields of the module at the end of the cycle and the mean
temperatures of all cells in the module throughout the whole cycle at different Tamb and PCT,
respectively. Different from cases of air cooling, the module temperature keeps rising rather than
fluctuates in a range. Mean temperature curves of selected cells in the module are shown in Appendix
Fig. 11A.
Fig. 16(a) shows the time-averaged temperature of the module throughout the whole cycle. We
can see the module mean temperature rises with the increase of Tamb. The PCM with a lower PCT has
a better cooling effect at low Tamb. However, when Tamb reaches a certain value, the module mean
temperature rises significantly. This is because PCM is completely melted before the end of cycle in
these cases, which can be seen from the liquid phase fraction results shown in Fig. 17. Fig. 16(b) shows
the ΔT is lower than 0.4 C in all conditions, indicating PCM cooling can achieve much better
uniformity than its air cooling counterpart. Meanwhile, a strong correlation between ΔT and liquid
phase fraction can be found in Fig. 16(b) and Fig. 17. A higher liquid phase fraction leads to smaller
ΔT before complete phase transition. But once the PCM has completely melted, ΔT will rise rapidly.

Fig. 16 Module temperature at different Tamb and PCTs: (a) Mean temperature and (b) max.
temperature difference.
 Fig. 17 Liquid phase fraction of PCM at the end of the cycle

Fig. 18 Capacity loss trajectories: (a) Tamb =30 C and (b) Tamb = 50 C.
Fig. 19 Cycle life of the module at different Tamb and PCT: (a) Mean cycle life, (b) maximum cycle
life difference and (c) maximum cycle life difference rate.
Fig. 18 and Fig. 19 show the capacity degradation trajectories of representative cases and cycle
life of the module, respectively. Fig. 19(a) and Fig. 19(b) illustrate that when Tamb increases from 0 C
to 50 C, the N
̅ drops from 900-1000 to less than 200, and ΔN reduces from 30-40 to less than 10. Fig.

19(c) shows the ζN is lower than 5% in all cases. It is the highest at Tamb = 0 C and then decreases with
the increase of Tamb. However, there are sharp rises at high Tamb in some cases. Fig. 19 also shows that
the PCM with lower PCT can achieve improved cycle life. However, it is more likely to be completely
melted in continuous charging/discharging cycle with high rates and at high Tamb conditions, which
will put the battery module at the risk of accelerating the degradation process. In consequence, it’s a
conflict while choosing the type of PCM especially when the battery pack is to be used in a wide Tamb
range.

4.3 Comparison of air and PCM cooling strategies

Fig. 20 Pressure drop (a) and temperature increase (b) between inlet and outlet of cooling air

Fig. 21 Total power consumption of air cooling

In this section, the performance of air cooling and PCM cooling are compared with respect to their
power consumption and battery cycle life.
Because PCM cooling is a passive BTMS and does not consume energy, we emphasize on the
energy consumption analysis of air cooling. Power consumption considered in this study includes the
pumping power of the fan and refrigerating power of the air conditioning system. The ideal pumping
power consumption is calculated using Eq.(16):
 Pfan = p V (16)

where Δp is the pressure drop from the inlet to the outlet of duct; V̇ is the volume flow rate of cooling
air.
The refrigerating power is evaluated by assuming that the refrigeration cycle of the air
conditioning system has a constant Coefficient of Performance (COP) of 2 in all conditions [42]. Thus,
the ideal refrigerating power consumption can be calculated by Eq.(17) :

PAC = mair C p,air Tair / COP (17)

where ṁ air is the mass flow rate; ΔTair is the temperature increase of the cooling air from the inlet to
the outlet.
Fig. 20 presents the pressure drop and temperature increase between the inlet and the outlet of
cooling air. Fig. 21 shows the ideal power consumptions of fan and air conditioning system. A sharp
rise of power consumption while increasing Vin can be observed, which is attributed to the dramatically
rising fan power. Total power consumption also increases when Tamb rises to 30 C because the air-
conditioning system is turned on to produce 20 C cooling air.
To evaluate the cost performance of air cooling and PCM cooling systems, this study take battery
cycle life as an additional performance indicator of BTMS on top of battery maximum temperature,
temperature non-uniformity and power consumption which are taken as the main performance
indicators in most previous studies and an evaluation index of cyclical cost (δ) is proposed to link
indicators of power consumption and battery cycle life together. δ is defined as the ratio of the total
cost, including purchasing cost of the battery module and parasitic energy cost of BTMS to the module
cycle life (capacity loss reaches 20%). δ is expressed as the following equation:

 batt  Cnom
PACK
+   Neof  Qp
= (18)
Neof

f
= (19)
H fuelPT

where τbatt is the battery; CPACK

nom is the nominal capacity of the battery module; Neof is the number of

cycles of the battery pack before end of life; Qp is the parasitic energy consumption per cycle; τf is the
fuel price; Hfuel is the lower heating value of the fuel; ηPT is efficiency of the hybrid powertrain and is
assumed to be 30.1% [43]. Table 7 lists the values of Qp, τf and Hfuel obtained from Ref. [44]. Such
economic index is also a common practice in the researches of power management for HEV [37,44–
46].
Table 7: Main parameters for cost evaluation[44]

parameter Value

battery price [€ kWh-1] 900

Fuel price [€ L-1] 1.34
Diesel lower heating value [MJ L-1] 38.6

Cases of air cooling at Vin = 0.5 m s-1, PCM cooling at PCT = 45 C and no BTMS are chosen for
̅ , Nmax and Nmin of three conditions. It shows air cooling can
comparison. Fig. 22(a) presents the N
achieve the highest cycle life and the largest non-uniformity. Fig. 22(b) shows δ of the battery module
can be reduced by using active air and passive PCM cooling. Besides, air cooling has at least 50%
lower δ than that of PCM cooling though it has parasitic power consumption. δ of the PCM cooling is
also more sensitive to the Tamb, especially when Tamb exceeds the PCT of the PCM. Fig. 23 shows δ in
air cooling condition at different Tamb and Vin. δ of battery module first decreases and then increases
with the increase of Vin, and there are optimal Vin for the lowest δ at different Tamb.

Fig. 22 Cycle life (a) and cyclical cost (b) comparison of air cooling (Vin = 0.5 m s-1), PCM cooling
(PCT = 45 C) and no BTMS
 Fig. 23 Cyclical costs in air cooling condition at different Tamb and Vin
5. Conclusions
This study analyses the thermal states and cycle life of a battery module under a dynamic current
load obtained from a hybrid electric bus (16.55 km in approximately 3000 s) when active air cooling
using air-conditioned air and passive PCM cooling are independently applied for thermal management.
The effects of inlet velocity (Vin), PCM phase change temperature (PCT) and ambient temperature
(Tamb) are studied. The following are conclusions drawn from results and discussion.
1. Active air cooling using air-conditioned air is able to keep the module temperature under safety
range even at Tamb as high as 50 C, whereas passive PCM based BTMS in which PCM is cooled by
natural convection with air at ambient temperature is difficult to achieve this. However, a more uniform
temperature field can be obtained with PCM cooling.
2. Due to the lower temperature field in air cooling condition, module cycle life is extended by up to
600% compared with that of the PCM cooling. In the air cooling condition, the maximum cycle life
difference rate (ζN) is over 20% at a low Vin, e.g. 0.5 m s-1. With the increase of Vin, ζN is reduced. In
PCM cooling condition, ζN is lower than 5% and the largest value is 4.1%, which has a strong link to
the liquid phase fraction of PCM.
3. The cyclical cost (δ) of the battery module using air-conditioned air cooling is much lower than
that using PCM cooling. In PCM cooling, δ can be double due to limited module cycle life. For the air
cooling, the δ first decreases and then increases while increasing Vin. There exists an optimal Vin for
the lowest δ at different Tamb.
4. To extend the life and reduce operating cost of LIBs by means of thermal management, active air
cooling is the better choice than passive PCM only cooling. For the purpose of both extending the life
and improving the uniformity of the LIB pack, PCM and active compound system may be the best
choice.

Acknowledgements
The authors highly acknowledge the finical support from Science and Technology Planned Project of
Zhejiang Province (No. 2018C01057).
Reference
[1] M.Z. Jacobson, Review of solutions to global warming, air pollution, and energy security,
Energy Environ. Sci. 2 (2009) 148–173.
[2] A. Barré, B. Deguilhem, S. Grolleau, M. Gérard, F. Suard, D. Riu, A review on lithium-ion
battery ageing mechanisms and estimations for automotive applications, J. Power Sources. 241
(2013) 680–689.
[3] M.M. Thackeray, C. Wolverton, E.D. Isaacs, Electrical energy storage for transportation -
Approaching the limits of, and going beyond, lithium-ion batteries, Energy Environ. Sci. 5
(2012) 7854–7863.
[4] G. Zubi, R. Dufo-López, M. Carvalho, G. Pasaoglu, The lithium-ion battery: State of the art
and future perspectives, Renew. Sustain. Energy Rev. 89 (2018) 292–308.
[5] W. Lv, Z. Wang, H. Cao, Y. Sun, Y. Zhang, Z. Sun, A Critical Review and Analysis on the
Recycling of Spent Lithium-Ion Batteries, ACS Sustain. Chem. Eng. 6 (2018) 1504–1521.
[6] K. Zhao, L. Zhang, R. Xia, Y. Dong, W. Xu, C. Niu, L. He, M. Yan, L. Qu, L. Mai, SnO2
Quantum Dots@Graphene Oxide as a High-Rate and Long-Life Anode Material for Lithium-
Ion Batteries, Small. 12 (2016) 588–594.
[7] W. Liu, X. Li, D. Xiong, Y. Hao, J. Li, H. Kou, B. Yan, D. Li, S. Lu, A. Koo, K. Adair, X.
Sun, Significantly improving cycling performance of cathodes in lithium ion batteries: The
effect of Al 2 O 3 and LiAlO 2 coatings on LiNi 0.6 Co 0.2 Mn 0.2 O 2, Nano Energy. 44
(2018) 111–120.
[8] W. Gao, Y. Zheng, M. Ouyang, J. Li, X. Lai, X. Hu, Micro-short-circuit diagnosis for series-
connected lithium-ion battery packs using mean-difference model, IEEE Trans. Ind. Electron.
66 (2019) 2132–2142.
[9] B. Xia, Y. Shang, T. Nguyen, C. Mi, A correlation based fault detection method for short
circuits in battery packs, J. Power Sources. 337 (2017) 1–10.
[10] J. Wei, G. Dong, Z. Chen, Y. Kang, System state estimation and optimal energy control
framework for multicell lithium-ion battery system, Appl. Energy. 187 (2017) 37–49.
[11] M.M. Hoque, M.A. Hannan, A. Mohamed, A. Ayob, Battery charge equalization controller in
electric vehicle applications: A review, Renew. Sustain. Energy Rev. 75 (2017) 1363–1385.
[12] L.H. Saw, Y. Ye, A.A.O. Tay, W.T. Chong, S.H. Kuan, M.C. Yew, Computational fluid
dynamic and thermal analysis of Lithium-ion battery pack with air cooling, Appl. Energy. 177
(2016) 783–792.
[13] E. Jiaqiang, D. Han, A. Qiu, H. Zhu, Y. Deng, J. Chen, X. Zhao, W. Zuo, H. Wang, J. Chen,
Q. Peng, Orthogonal experimental design of liquid-cooling structure on the cooling effect of a
liquid-cooled battery thermal management system, Appl. Therm. Eng. 132 (2018) 508–520.
[14] D. Zou, X. Ma, X. Liu, P. Zheng, Y. Hu, Thermal performance enhancement of composite
phase change materials (PCM) using graphene and carbon nanotubes as additives for the
potential application in lithium-ion power battery, Int. J. Heat Mass Transf. 120 (2018) 33–41.
[15] Y. Ye, Y. Shi, L.H. Saw, A.A.O. Tay, Performance assessment and optimization of a heat pipe
thermal management system for fast charging lithium ion battery packs, Int. J. Heat Mass
Transf. 92 (2016) 893–903.
[16] J. Zhao, P. Lv, Z. Rao, Experimental study on the thermal management performance of phase
change material coupled with heat pipe for cylindrical power battery pack, Exp. Therm. Fluid
Sci. 82 (2017) 182–188.
[17] M. Al-Zareer, I. Dincer, M.A. Rosen, Novel thermal management system using boiling
cooling for high-powered lithium-ion battery packs for hybrid electric vehicles, J. Power
 Sources. 363 (2017) 291–303.
[18] A.M. Sefidan, A. Sojoudi, S.C. Saha, Nanofluid-based cooling of cylindrical lithium-ion
battery packs employing forced air flow, Int. J. Therm. Sci. 117 (2017) 44–58.
[19] J. Wang, J. Purewal, P. Liu, J. Hicks-Garner, S. Soukazian, E. Sherman, A. Sorenson, L. Vu,
H. Tataria, M.W. Verbrugge, Degradation of lithium ion batteries employing graphite
negatives and nickel-cobalt-manganese oxide + spinel manganese oxide positives: Part 1,
aging mechanisms and life estimation, J. Power Sources. 269 (2014) 937–948.
[20] J. Neubauer, E. Wood, Thru-life impacts of driver aggression, climate, cabin thermal
management, and battery thermal management on battery electric vehicle utility, J. Power
Sources. 259 (2014) 262–275.
[21] T. Yuksel, S. Litster, V. Viswanathan, J.J. Michalek, Plug-in hybrid electric vehicle LiFePO 4
battery life implications of thermal management, driving conditions, and regional climate, J.
Power Sources. 338 (2017) 49–64.
[22] M. Doyle, Modeling of Galvanostatic Charge and Discharge of the Lithium/Polymer/Insertion
Cell, J. Electrochem. Soc. 140 (1993) 1526.
[23] T.F. Fuller, Simulation and Optimization of the Dual Lithium Ion Insertion Cell, J.
Electrochem. Soc. 141 (1994) 1.
[24] P. Amiribavandpour, W. Shen, D. Mu, A. Kapoor, An improved theoretical electrochemical-
thermal modelling of lithium-ion battery packs in electric vehicles, J. Power Sources. 284
(2015) 328–338.
[25] E. Prada, D. Di Domenico, Y. Creff, J. Bernard, V. Sauvant-Moynot, F. Huet, Simplified
Electrochemical and Thermal Model of LiFePO 4 -Graphite Li-Ion Batteries for Fast Charge
Applications, J. Electrochem. Soc. 159 (2012) A1508–A1519.
[26] Y. Ye, Y. Shi, A.A.O. Tay, Electro-thermal cycle life model for lithium iron phosphate
battery, J. Power Sources. 217 (2012) 509–518.
[27] J. Voelcker, Lithium batteries take to the road, IEEE Spectr. 44 (2007) 26–31.
[28] M. Safari, C. Delacourt, Modeling of a commercial graphite/LiFePO4 cell, J. Electrochem.
Soc. 158 (2011) 562–571.
[29] J.L. Dodd, Phase composition and dynamical studies of lithium iron phosphate, 2007.
http://thesis.library.caltech.edu/1662/.
[30] Y. Reynier, R. Yazami, B. Fultz, The entropy and enthalpy of lithium intercalation into
graphite, J. Power Sources. 119–121 (2003) 850–855.
[31] Y. Li, Y. Du, T. Xu, H. Wu, X. Zhou, Z. Ling, Z. Zhang, Optimization of thermal management
system for Li-ion batteries using phase change material, Appl. Therm. Eng. 131 (2018) 766–
778.
[32] N. Javani, I. Dincer, G.F. Naterer, B.S. Yilbas, Heat transfer and thermal management with
PCMs in a Li-ion battery cell for electric vehicles, Int. J. Heat Mass Transf. 72 (2014) 690–
703.
[33] J. Wang, P. Liu, J. Hicks-Garner, E. Sherman, S. Soukiazian, M. Verbrugge, H. Tataria, J.
Musser, P. Finamore, Cycle-life model for graphite-LiFePO4 cells, J. Power Sources. 196
(2011) 3942–3948.
[34] K. Shah, A. Jain, Prediction of thermal runaway and thermal management requirements in
cylindrical Li-ion cells in realistic scenarios, Int. J. Energy Res. 43 (2019) 1827–1838.
[35] J. Zhao, Z. Rao, Y. Huo, X. Liu, Y. Li, Thermal management of cylindrical power battery
module for extending the life of new energy electric vehicles, Appl. Therm. Eng. 85 (2015)
33–43.
[36] Z. Rao, Q. Wang, C. Huang, Investigation of the thermal performance of phase change
material/mini-channel coupled battery thermal management system, Appl. Energy. 164 (2016)
659–669.
[37] X. Hu, L. Johannesson, N. Murgovski, B. Egardt, Longevity-conscious dimensioning and
power management of the hybrid energy storage system in a fuel cell hybrid electric bus, Appl.
Energy. 137 (2015) 913–924.
[38] F. He, H. Wang, L. Ma, Experimental demonstration of active thermal control of a battery
module consisting of multiple Li-ion cells, Int. J. Heat Mass Transf. 91 (2015) 630–639.
[39] R. Sabbah, R. Kizilel, J.R. Selman, S. Al-Hallaj, Active (air-cooled) vs. passive (phase change
material) thermal management of high power lithium-ion packs: Limitation of temperature rise
and uniformity of temperature distribution, J. Power Sources. 182 (2008) 630–638.
[40] J. Jaguemont, N. Omar, P. Van den Bossche, J. Mierlo, Phase-change materials (PCM) for
automotive applications: A review, Appl. Therm. Eng. 132 (2018) 308–320.
[41] T. ur Rehman, H.M. Ali, M.M. Janjua, U. Sajjad, W.M. Yan, A critical review on heat transfer
augmentation of phase change materials embedded with porous materials/foams, Int. J. Heat
Mass Transf. 135 (2019) 649–673.
[42] M. Shams-Zahraei, A.Z. Kouzani, S. Kutter, B. Bäker, Integrated thermal and energy
management of plug-in hybrid electric vehicles, J. Power Sources. 216 (2012) 237–248.
[43] J. Thomas, Drive Cycle Powertrain Efficiencies and Trends Derived from EPA Vehicle
Dynamometer Results, SAE Int. J. Passeng. Cars - Mech. Syst. 7 (2014) 1374–1384.
[44] X. Hu, N. Murgovski, L.M. Johannesson, B. Egardt, Comparison of three electrochemical
energy buffers applied to a hybrid bus powertrain with simultaneous optimal sizing and energy
management, IEEE Trans. Intell. Transp. Syst. 15 (2014) 1193–1205.
[45] S.J. Moura, H.K. Fathy, D.S. Callaway, J.L. Stein, A stochastic optimal control approach for
power management in plug-in hybrid electric vehicles, IEEE Trans. Control Syst. Technol. 19
(2011) 545–555.
[46] S. Zhang, R. Xiong, F. Sun, Model predictive control for power management in a plug-in
hybrid electric vehicle with a hybrid energy storage system, Appl. Energy. 185 (2017) 1654–
1662.
 Appendix

Fig. 1A Reference open circuit potential (a) and entropy change (b)

Fig. 2A Electrolyte conductivity as a function of Li concentration

Fig. 3A Temperature fields of the module at the end of the driving cycle of air cooling at Tamb = 0 C
 Fig. 4A Module mean temperature curves through the whole cycle of air cooling at Tamb=0 C

Fig. 5A Mean temperature curves of selected cells in the module of air cooling at Tamb=0 C: (a) Vin =
0.5 m s-1 and (b) Vin = 8 m s-1

Fig. 6A Capacity loss trajectories of air cooling at Tamb=0 C

 Fig. 7A Velocity and pressure fields of the fluid domain of air cooling at Tamb = 30 C

Fig. 8A Temperature fields of the module at the end of the driving cycle of PCM cooling at Tamb = 0
C
Fig. 9A Liquid phase fraction of PCM in the module at the end of driving cycle
Fig. 10A Mean temperature curves of the module of PCM cooling at Tamb=0 C
Fig. 11A Mean temperature curves of selected cells in the module of PCM cooling: (a) Tamb = 0 C,
PCT = 35 C, (b) Tamb = 0 C, PCT = 55 C, (c) Tamb = 30 C, PCT = 35 C, (d) Tamb = 30 C, PCT =
55 C, (e) Tamb = 50 C, PCT = 35 C and (f) Tamb = 50 C, PCT = 55 C
Fig. 12A Capacity loss trajectories of CPM cooling at Tamb=0 C