Journal of Energy Storage 40 (2021) 102796

Contents lists available at ScienceDirect

Journal of Energy Storage

journal homepage: www.elsevier.com/locate/est

Experimental and numerical investigation of the LiFePO4 battery cooling by

natural convection
Orhan Kalkan a, b, Ali Celen a, *, Kadir Bakirci b
a
Department of Mechanical Engineering, Faculty of Engineering, Erzincan Binali Yıldırm University, Erzincan 24100, Turkey
b
Department of Mechanical Engineering, Atatürk University, Erzurum, 25240, Turkey

A R T I C L E I N F O A B S T R A C T

Keywords: In this study, the thermal performance of a LiFePO4 (LFP) pouch type battery in the range of 1C-5C discharge rate
LFP battery at 23 ◦ C ambient temperature and natural convection conditions is experimentally and numerically investigated.
Natural convection Time-dependent temperature changes of the battery are imaged with a thermal camera for each discharge, and
Thermal performance
the maximum, minimum and average changes on the battery surface are examined. In addition, the heat transfer
Cooling
rate through the battery is calculated for each discharge rate under natural convection conditions of the battery,
and the temperature and velocity distributions on the battery surface are investigated. For the experiment with
the highest discharge rate (5C), the measured maximum and average battery temperatures are determined as
52.2 ◦ C and 50.1 ◦ C, respectively. The calculated maximum temperature difference on the battery surface is also
6.7 ◦ C. Moreover, the homogeneity index defined for battery surface temperature is 0.147 at 5C discharge rate. It
has been observed that natural convection cooling is not sufficient for the battery at high discharge rates
Furthermore, a new correlation equation has been proposed for the estimation of the heat transfer rate from a 20
Ah LFP pouch type battery under natural convection conditions at 23 ◦ C ambient temperature.

next 15 years [6,7].

1. Introduction Rechargeable Lithium ion (Li-ion) batteries are widely used in ap­
plications such as mobile phones, vehicles, data centers, etc. As an en­
Nowadays, energy demand is increasing with the rapid development ergy storage medium, many types of batteries have been developed both
of industry and technology. As stated in the International Energy in terms of their electrochemical structure and geometric structure. Li-
Outlook 2020 report, energy expenditures in the world will show an ion batteries, which are the most popular today, have prominent ad­
increment swiftly in the coming years, especially in high income regions vantages such as low self-discharge rate, high power density, high cycle
[1]. It is predicted that the global warming linked to the spread of life and high energy capacity. Owing to these advantages, automotive
greenhouse gasses around the world, reduction of fossil fuels and their and electronics industries show great demand for Li-ion batteries. Li-ion
destructive effects will feed the possibility of energy crisis in the future batteries with large capacity are used to meet the required power in the
[2,3,4]. The use of renewable energy sources includes solar, wind, relevant application areas. [8]. On the other hand, Li-ion batteries still
hydro, geothermal, biofuels, etc. has become even more important. have critical problems because they contain high energy material and
However, the main problem with the energy to be obtained from all flammable electrolyte solution, as well as disadvantages such as aging,
these sources is the non-uniform energy output [5]. For this reason, the thermal leakage, high cost, safety, electronic waste. Among the afore­
energy outputs in question need to be stabilized by storing in any type of mentioned drawbacks, safety is a priority for any manufacturer. These
energy. The storage of this energy is possible in today’s technology, safety risks that may cause battery fires are tried to be eliminated using
mostly by using batteries. After the energy is stored as chemical form in positive temperature coefficient element (PTC), non-flammable mate­
batteries, it can be transformed into different energy forms regarding the rials in the electrolyte, overload barrier, add-on antioxidants and similar
field of application. It is known that the most common use and tech­ techniques [9,10]. Maximum battery temperature and temperature
nological development of batteries are generally for electric vehicles. It uniformity between the cells are two significant parameters affecting the
is estimated that the number of electric vehicles will account for safety, health and performance of the battery. Li-ion batteries should
approximately 15% to 30% of the total number of vehicles used in the have an ideal operating temperature range which is between 20 ◦ C and

* Corresponding author.
E-mail address: [email protected] (A. Celen).

https://doi.org/10.1016/j.est.2021.102796
Received 22 March 2021; Received in revised form 10 May 2021; Accepted 30 May 2021
Available online 13 June 2021
2352-152X/© 2021 Elsevier Ltd. All rights reserved.
O. Kalkan et al. Journal of Energy Storage 40 (2021) 102796

Nomenclature β Volumetric thermal expansion coefficient

Θ Homogeneity index
C NTGK model constant σ+ Effective electric conductivities for the positive electrodes
g Gravitational acceleration [m/s2] [S/m]
GrL Grashof number σ− Effective electric conductivities for the negative electrodes
h Average heat convection coefficient [W/m2K] [S/m]
k Heat conduction coefficient of air [W/mK] φ+ Phase potentials for the positive electrodes [V]
L Characteristic length of the battery body [m] φ− Phase potentials for the negative electrodes [V]
NuL Average Nusselt number
Subscripts
Pr Prandtl number
corr Correlation
Q̇surf Heat transfer rate through the battery surface [W]
max Maximum
Q̇gen Heat generation [W] min Minimum
RaL Rayleigh number ref Reference
t Time [s]
T Battery temperature [K] Acronyms
Ts Battery surface temperature [K] CC Constant current
T∞ Ambient temperature [K] CV Constant voltage
u Air velocity in the x-axis direction [m/s] DoD Depth of discharge
v Air velocity in the y-axis direction [m/s] IR Infrared
v Kinematic viscosity [m2/s] LFP Lithium Ferro Phosphate
NCM Lithium Nickel Manganese Cobalt Oxide
Greek symbols PTC Positive temperature coefficient element
α Thermal diffusivity [m2/s]

40 ◦ C, while the temperature change between cells should be a internal temperature of the battery can be visualized via a thermal
maximum of 5 ◦ C [11,12,13,14]. When the operating temperature of camera when the battery is charged or discharged. For the discharge
Li-ion batteries exceeds 50 ◦ C, battery aging process accelerates and with 2C discharge rate, the maximum temperature difference of 2 ◦ C was
capacity fade occurs [15,16]. Therefore, an appropriate thermal man­ observed between the visualized inner and outer surfaces of the battery
agement system needs to be developed for an electric vehicle to operate at the end of the experimental process. It should be noted that C-rate can
safely with high efficiency [17]. When designing thermal management be defined as the ratio of charge/discharge current to battery nominal
systems, the surface temperature distribution of the battery should be capacity. As a result, it has been shown that internal battery temperature
considered as an important criterion. One of the equipment used in the has a significant effect to assess the effects of temperature on battery
experimental determination of the surface temperatures of the batteries safety. Wang et al. [22] investigated the thermal behavior of a battery
is the thermal camera and the studies in the literature regarding the has NCM (LiNi1-x-yCoxMnyO2) cathode material under natural con­
thermal imaging of the batteries are summarized below: vection conditions. Temperature dissipating on the battery surface for
Kanbur et al. [18] propose a novel method to estimate battery different depth of discharge (DoD) and discharge rates via a thermal
thermal performance based on surface temperature differences, losses of camera. Two different indexes are defined to determine the temperature
heat and exergy. In the experimental study, the Li-ion battery of a mobile distribution and fluctuations on the battery surface. It is understood that
phone was imaged via a thermal camera. Firstly, the camera records by these indexes defined can show the temperature homogeneity holisti­
thermal imaging from the battery surface for 1 min so that it can be used cally in spatial distribution graphs depending on different discharge
in data training. The proposed method then estimates the temperature rates and DoD. Zhang et al. [23] have experimentally determined the
change at the surface for the last 5 min of the test period. The error of heat transfer coefficient and specific thermal capacity of a Li-ion battery
estimation of the temperature difference occurred on the battery surface cell via an IR camera. The specific thermal capacity of the battery was
was found to be less than 1%. Moreover, exergy loss and heat dissipation also obtained with the battery surface temperatures measured using the
estimations have approximately 2.8% and 5.3% error rates. Akbarzadeh lumped capacitance method. However, researchers conducted this study
et al. [19] presented a novel method to determine the thermal capacity for a cell phone battery and stated that this method is not suitable for
and anisotropic thermal conductivity of a Li-ion battery with a capacity large size batteries. Kim et al. [24] investigated the degradation due to
of 43 Ah. Moreover, the thermal performance of a battery module with a high discharge rate and temperature in the anode and cathode materials
voltage of 48 V was simulated applying a 3D thermal model at the of a LFP pouch type battery. The thermal performance of the battery was
ambient temperature conditions. According to the thermal character­ evaluated by monitoring the battery temperatures due to internal
ization results, the thermal capacity of the battery increases as the resistance changes with an IR camera. As a result, it has been determined
operating temperature increases. However, the thermal conductivity of that the internal resistance of the cell directly affects the battery life.
the battery is independent of temperature. Mevawalla et al. [20] Panchal et al. [25] examined the thermal performance of a prismatic LFP
mathematically modeled a LFP battery with 20 Ah capacity using battery by imaging the surface temperature and temperature distribu­
COMSOL software. The battery surface and tab temperature results tions with a thermal camera at different discharge rates. Besides, the
under 1C, 2C, 3C and 4C discharge rates are validated with experimental internal resistance, heat generation and voltage variations of the battery
results. Moreover, the temperature distributions on the battery surface were also evaluated. Furthermore, the thermal model of the battery was
were investigated with the numerical solution. It was determined that developed using Matlab Simulink and this model was verified by
the amount of heat generated at the cathode was twice that of the anode experimental measurements. Bazinski et al. [26] experimentally inves­
under 2C and 3C discharge rates. Zhu et al. [21] developed a new tigated the heat conduction of two different sized pouch type LFP bat­
technique to measure the internal temperature of Li-ion batteries. The teries with 14.5 Ah and 10 Ah capacities. The batteries were designed in
battery to be used in experimental study is specially designed by the experimental system to act as fins for a heat source and the heat
covering one surface with an optical glass. In this way, the transient conduction of the battery was solved by the one-dimensional heat

2
O. Kalkan et al. Journal of Energy Storage 40 (2021) 102796

transfer equations used for the fins. As a result, the average heat con­
duction coefficients of the batteries with 14.5 Ah and 10 Ah capacity
were found to be 28.9 W/m.K and 35.1 W/m.K, respectively. Jaguemont
et al. [27] developed a 3D thermal model for a lithium titanium oxide
battery. The developed model contains a heat transfer model that can
offer a faster solution by ignoring electrochemical processes. Simula­
tions of the battery under different ambient temperature conditions
were performed and the temperature results obtained were compared
with the battery surface temperatures measured experimentally by a
thermal camera. Xu et al. [28] experimentally studied the thermal
runaway of Li-ion battery modules with different electrical connection
configurations. The propagation characteristic of thermal runaway was
imaged using an IR camera. With thermocouples placed in the cell, the
temperature of the battery was measured as 160 ◦ C during thermal
runaway. In addition, the electrical model of the battery module was
developed and the amount of electrical charge transferred from other
cells to the cell that was leaking during thermal runaway was also
calculated. Li et al. [29] modeled the temperature distribution on the
surface of a prismatic Li-ion battery using an IR camera and a 3D
scanner. In order to accurately analysis the relationship between the
battery surface topography and the battery surface temperature, a data
fusion model is proposed and accordingly a common calibration method
for parameter identification of the data fusion model is presented. As a
result, the researchers showed that the multiple physical measurement
system can achieve a position matching deviation of 0.19 mm with high
resolution and high data acquisition speed.
In this study, the thermal behavior of a LFP pouch cell with a ca­
pacity of 20 Ah under natural convection conditions and different
discharge rates was experimentally and numerically investigated. The
distribution of surface temperatures of the battery at different discharge
rates depending on time was imaged using a thermal camera, and the
variation of the average battery temperature over time, the average and Fig. 1. Representation of the experimental setup.
maximum battery temperatures reached at the end of each discharge
and the temperature differences on the battery surface were obtained. gradually as there is a transfer of electrons between electrodes during
The experimental results obtained were compared with the numerical discharge and the discharge process is terminated at a voltage value of 2
results, and the heat transfer rate through the battery surface by both V, which is the discharge cut-off voltage recommended by the
methods was determined during the battery discharge. In addition, a manufacturer.
homogeneity index is defined as a measure of the temperature distri­
bution homogeneity on the battery surface and calculated for the certain 2.2. Governing equations and boundary conditions
experimental conditions. In addition, a correlation equation has been
proposed to determine the heat transfer rate required for cooling the LFP The numerical model of cooling the battery under natural conditions,
pouch cell with 20 Ah capacity under natural convection conditions. whose thermal performance was examined experimentally, was gener­
ated. The acceptances regarding this numerical model can be listed as
2. Material and method follows:

2.1. Experimental procedure • Air flow is laminar and two dimensional.

• Air, which is the coolant, is assumed to be an ideal gas.
Transient thermal imaging experiments of a Li-ion battery with a • Heat transfer by radiation on the surface has been neglected.
nominal capacity of 20 Ah at different discharge rates was carried out • Boussinesq approach is valid in the boundary layer on the battery
using the apparatus shown in Fig. 1. The schematic view of the experi­ surface. However, it was assumed that the temperature-dependent
mental setup showing the connection line between the battery and the physical properties of the air for buoyancy formulation did not
experimental tools is also shown in Fig. 2. In the experiments, a thermal change.
infrared (IR) camera with 160 × 120 pixel resolution was used and the • In the initial condition, the battery surface has a homogeneous
maximum, minimum and average temperatures of the battery were temperature distribution.
determined by means of the camera software. It should be noted that
thermal visualizations of battery surface were obtained for the 0%, 20%, With these assumptions, the energy conservation equation, mo­
40%, 60%, 80% and 100% DoDs. The technical characteristics of the Li- mentum conservation equation and continuity equation to be used in
ion battery used in the experimental study are given in Table 1. Exper­ solving the boundary layer on the battery surface are governed as fol­
iments were carried out at 23 ◦ C ambient temperature and five different lows [30].
discharge rates of 1C-5C. The battery was fully charged with constant Energy conservation equation [30]:
current constant voltage method (CC/CV) using a direct current (DC)
power source before each experiment. Discharge of the battery was ∂T ∂T
+u +v
∂T ∂2 T
=α 2 (1)
performed using the constant current (CC) method with an electronic ∂t ∂x ∂y ∂y
load device. During discharge, the voltage and current values of the Momentum conservation equation [30]:
battery were recorded by the electronic load device. While the battery
has a voltage of 3.65 V in a fully charged state, its voltage decreases

3
O. Kalkan et al. Journal of Energy Storage 40 (2021) 102796

Fig. 2. Schematic view of the experimental setup.

Table 1 gβ(Ts − T∞ )L3

Technical specifications of LFP battery.
RaL = GrL Pr = (7)
vα
Parameter Value The following empirical correlation developed by Churchill and Chu
Brand/model A123 Systems/AMP20 [32] to be used in natural convection and heat transfer applications on
Cathode material LiFePO4 the surface of vertical plates has also been used for calculations of nat­
Anode material Graphite
ural convection and heat transfer at the battery surface.
Electrolyte material Carbonate based
Nominal capacity 20 Ah { }2
Nominal voltage 3.3 V 0, 387RaL 1/6
NuL = 0, 825 + [ ]4/9 (8)
Cut-off charge voltage 3.65 V
1 + (0, 492/Pr)9/16
Cut-off discharge voltage 2V
Weight 496 g
Dimensions 7.2 × 160 × 227 mm To calculate the heat transfer rate (Q̇surf ) through the battery surface
experimentally cooled by natural convection at 23 ◦ C ambient temper­
ature, Newton’s cooling law equation given below was used.
∂u ∂u ∂v ∂2 u
+ u + v = gβ(T − T∞ ) + v 2 (2) NuL L
∂t ∂x ∂y ∂x h= (9)
k
Continuity equation [30]:
Q̇surf = hAs (Ts − T∞ ) (10)
∂u ∂v
+ =0 (3)
∂x ∂y Here; h is the average heat transfer coefficient, k is the heat con­
The initial and boundary conditions determined for the solution of duction coefficient of the air, As is the surface area of the battery, Ts is
the conservation Eq. (1)-(3) are given below: the average surface temperature of the battery and T∞ is the ambient air
Initial conditions: temperature.
In order to determine the homogeneity of the battery surface tem­
u(x, y, 0) = 0; T(x, y, 0) = T∞ , {0 ≤ x < ∞, 0 ≤ y < ∞} (4) peratures under different operating conditions, the Θ dimensionless
Boundary conditions: number is defined and can be called the homogeneity index. This
number is calculated as follows:
∂u
u(x, 0, t) = 0; u(x, ∞, t) = 0; (x, ∞, t) = 0; {0 ≤ x〈∞, t〉0} (5) Tmax − Tmin
∂y Θ= (11)
Tmin

T(x, 0, t) = Ts ; T(x, ∞, t) = T∞ ;
∂T
(x, ∞, t) = 0; {0 ≤ x〈∞, t〉0} (6) Here, Tmax and Tmin are the maximum and minimum temperatures on
∂y the battery surface reached at the end of the discharge, respectively.
After the nondimensionalization of the parameters in the equations
given above, Rayleigh number is defined as the product of the Grashof
2.3. Numerical solution and mesh analysis
and Prandtl numbers as follows [31].

The numerical solution of the cooling of the battery under natural

convection conditions according to the initial and boundary conditions

4
O. Kalkan et al. Journal of Energy Storage 40 (2021) 102796

specified in the model generated was carried out using ANSYS-Fluent battery surface according to the grid number. The variation of the
software. Electrochemical-thermal coupled model of the battery was average battery temperatures reached at the end of the battery discharge
generated using the NTGK model which is developed by Kwon et al. [33] with 5C-rate and the heat transfer rate according to the grid number is
as a semi-empirical electrochemical model. Considering the NTGK given in Table 2. It was observed that the model with 802,682 grid
model, the volumetric current transfer density (j) is directly related to number had 0.10% and 0.39% relative error rate in the average battery
the φ+ and φ− phase potentials. The equation of volumetric current temperature and heat transfer rate results, respectively.
transfer density depending on U and Y model parameters, electrode
layer specific area (α) and phase potentials is given as follows. 3. Results and discussions
j = αY[U − (φ+ − φ− )] (12)
The experimental and numerical variation of the voltage drop during
The function of the U and Y model parameters depending on the the discharge of the battery was obtained for 1C-5C discharge rates and
depth of discharge (DoD) is given below. is given in Fig. 4. Discharge of the battery, which has a voltage of
( ) approximately 3.6 V when fully charged, was terminated at a cut-off
∑ 5
U= n (
an (DoD) − C2 T − Tref
)
(13) voltage of 2 V. As it can be understood from Fig. 4, it has been
n=0 observed that as the discharge rate increases, both the discharge time
and the nominal voltage value decrease. In addition, it is seen that the
( ) [ ( )]
∑
5
1 1 experimental and numerical results are compatible with each other.
Y= bn (DoD)n exp − C1 − (14) Experiments using the thermal imaging method were carried out
T Tref
during the LFP battery was discharged at 23 ◦ C ambient temperature at
n=0

Here, C1 and C2 are the battery NTGK model constants, and Tref is the 1C-5C rates. The change of the battery surface temperature at 1C-5C
reference temperature. U and Y functions for LFP battery as follows [34]. discharge rates and different DoD was measured with a thermal cam­
era and is shown in Fig. 5-9. When these figures were examined, it was
U = 3, 6 − 0, 804(DoD) + 1, 075(DoD)2 − 1, 177(DoD)3
( ) seen that the maximum battery temperatures were reached at the end of
+ 0, 00, 095 T − Tref (15) discharge (100% DoD), and these temperatures increased with
increasing discharge rate (1C-5C). Again, as can be understood from
[
Y = 1168, 59 − 8928(DoD) + 52, 504, 6(DoD)2 − 136, 231(DoD)3 Fig. 5-9, in the experiment performed at each discharge rate, it was seen
]
[ (
1 1
)] that there is more heating in the anode (negative) and cathode (positive)
+ 158, 531, 7(DoD)4 − 67, 578, 5(DoD)5 exp − 1800 − regions of the battery compared to the other regions, up to 80% DoD
T Tref
level. It was seen also that there is more warming at the cathode end
(16)
compared to the anode tip. This temperature difference is due to the use
The electrochemical and ohmic heat generation equation in the of aluminum at the cathode terminal and copper material at the anode
battery for NTGK model is defined as follows [34]: terminal, and the difference in the heat conduction coefficient of these
[ ]
dU
Q̇gen = j U − (φ+ − φ− ) − T + σ + ∇φ+ ∇φ+ + σ − ∇φ− ∇φ− (17)
dT Table 2
Here, T represents the battery temperature, σ+ and σ− represent the Mesh independency analysis.
electrical conductivity of the positive and negative electrodes, Grid Avg. battery Relative Heat transfer Relative
respectively. number temp. ( ◦ C) error rate rate (W) error rate

A hybrid mesh structure created by using ANSYS-Fluent mesh 269,460 54.38 %12.32 9.96 %20.48
module was used for the model for the solution of flow and energy 429,104 47.68 %3.44 7.92 %2.90
802,682 49.32 %0.10 7.69 %0.39
equations with the finite volume method, as shown in Fig. 3. Mesh in­
1,431,871 49.27 %0.43 7.66 %0.91
dependence analysis of the model was performed based on the change of 2,108,169 49.48 – 7.73 –
the battery surface temperature and the heat transfer rate through

Fig. 3. Mesh structures of the battery and air domains.

5
O. Kalkan et al. Journal of Energy Storage 40 (2021) 102796

The maximum and minimum temperatures measured on the battery

surface at the end of 1C discharge were measured as 33.7 ◦ C and 30.0 ◦ C,
respectively.
Fig. 6 depicts thermal images of the surface temperatures at certain
DoD levels at the 2C discharge rate of the battery are given. The general
characteristic structure seen in the surface temperature distribution at
1C discharge rate was also preserved at 2C discharge rate. Likewise, it is
seen that the maximum temperature zone on the battery surface is
concentrated in the middle of the battery body towards the end of the
discharge. The measured maximum, minimum and average tempera­
tures on the battery surface at the end of 2C discharge are 38.5 ◦ C,
33.9 ◦ C and 36.9 ◦ C, respectively.
When Fig. 7 is examined, surface temperature distributions at certain
DoD levels are seen at the 3C discharge rate of the battery. As in 1C and
2C discharge rates, in the experiment performed at 3C discharge rate, it
is understood that the maximum temperature zone towards the end of
the discharge is concentrated in the middle regions of the battery.
However, in this experiment, the maximum temperature zone at the end
of discharge does not have a homogeneous distribution as in the previ­
Fig. 4. Time-dependent voltage variation at different discharge rates. ous experiments. The maximum, minimum and average temperatures
measured on the battery surface at the end of the 3C rate discharge are
two materials. Besides, the maximum temperature field is concentrated 42.9 ◦ C, 37.6 ◦ C and 41.2 ◦ C, respectively.
in the middle parts of the battery surface at the range of 80% − 100% Fig. 8 shows the surface temperature distributions for different DoD
DoD. levels during the battery discharge at 4C rate. In the experiments, it is
In the experiments conducted at 1C discharge rate, thermal images of clearly seen that the cathode terminal (positive tab) is warmer than the
the battery surface at different DoD levels are given in Fig. 5. Temper­ anode terminal (negative tab). During the discharge, the maximum
ature contours show that the surface temperature increases in direct temperature zone condenses in the anode and cathode terminals up to
proportion to the discharge time. The surface temperature, which was 80% DoD, while there is a temperature distribution throughout the body
measured as 24.4 ◦ C at the beginning of the discharge, increased to at the end of the discharge (100% DoD). The maximum, minimum and
32.2 ◦ C at the end of the discharge (100% DoD). The maximum tem­ average temperatures measured on the battery surface at the end of the
perature zone on the surface shifts towards the middle parts of the 4C rate discharge are 47.9 ◦ C, 41.6 ◦ C and 46.1 ◦ C, respectively.
battery from near the anode cathode ends during the discharge period. In Fig. 9, surface temperature distributions are given during the

Fig. 5. Temperature distribution on the battery surface during the discharge of 1C.

6
O. Kalkan et al. Journal of Energy Storage 40 (2021) 102796

Fig. 6. Temperature distribution on the battery surface during the discharge of 2C.

Fig. 7. Temperature distribution on the battery surface during the discharge of 3C.

7
O. Kalkan et al. Journal of Energy Storage 40 (2021) 102796

Fig. 8. Temperature distribution on the battery surface during the discharge of 4C.

Fig. 9. Temperature distribution on the battery surface during the discharge of 5C.

8
O. Kalkan et al. Journal of Energy Storage 40 (2021) 102796

discharge of the battery at 5C rate according to certain DoD levels. 5C rate discharge are 52.2 ◦ C, 45.5 ◦ C and 50.1 ◦ C, respectively.
Although we can see the temperature distribution characteristic seen in As a result of the time-dependent temperature data obtained from
the experiment at the 4C discharge rate also in this experiment, it is thermal images, the maximum, minimum and average temperature
understood that the temperature of the cathode terminal at each DoD values on the battery surface were determined experimentally and
level is higher than the anode terminal. The measured maximum, min­ numerically, and the change in time for each discharge rate was given in
imum and average temperatures on the battery surface at the end of the Fig. 10. It is seen that the maximum, minimum and average battery

Fig. 10. Time-dependent a) maximum, b) minimum and c) average battery temperature variation at 1C-5C discharge rates.

9
O. Kalkan et al. Journal of Energy Storage 40 (2021) 102796

temperatures increase in direct proportion to the discharge rate. The

graphs of the temperature over time show that the battery discharge
time decreases as the discharge rate increases. For example, while the
battery was discharged in one hour at a 1C discharge rate, the battery
was discharged in approximately twelve minutes at a 5C discharge rate.
Moreover, the average battery temperature at the end of the discharge of
the battery at 1C and 2C speeds does not exceed the upper operating
temperature limit (40 ◦ C) recommended for Li-ion batteries. However,
this limit is exceeded at the end of the discharge at 3C, 4C and 5C rates.
Furthermore, it is seen that there is an acceleration in the maximum,
minimum and average battery temperature increase for each discharge
rate towards the end of the discharge. In short, it can be stated that
natural convection cooling at high discharge rates (3C-5C) is not suffi­
cient for the battery.
The values of the temperature homogeneity index (ϴ) previously
defined for the battery cooling tests under natural convection conditions
performed between the discharge rates of the battery from 1C to 5C are
given in Fig. 11. It can be concluded that as the discharge rate increases,
the homogeneity index increases, so the temperature homogeneity on Fig. 12. The variation of the time-dependent heat dissipation at 1C-5C
the battery surface gradually decreases. discharge rates.
Eq. (10) was used to experimentally calculate the heat transfer rate
through battery surface, and the heat transfer coefficient by convection
points (x = 0 mm, x = 155 mm and x = 210 mm) along the battery body
in this equation was determined by using Eq. (9). The time-dependent
are 29.0 ◦ C, 28.9 ◦ C and 28.5 ◦ C, respectively.
change of the heat transfer rates for each discharge rate are given in
Fig. 14 represents the air velocity vectors and velocity boundary
Fig. 12. As can be understood from the figure, the heat removed from the
layer changes at different discharge rates in the same plane under nat­
battery increased with the increasing discharge rate, and this can be
ural convection conditions. It is understood that the average air velocity
explained by the increase in the convective heat transfer coefficient and
in the velocity boundary layer increases with the increase of the
the temperature difference between the average battery surface tem­
discharge rate. Average boundary layer velocities of the lowest, middle
perature and the ambient temperature. However, it is obvious that the
and top points (x = 0 mm, x = 155 mm and x = 210 mm) along the
surface and ambient temperature difference will also increase at each
battery body at the maximum discharge rate (5C) are 0.002 m/s, 0.077
discharge rate, as the average battery surface temperature will increase
m/s and 0.106 m/s, respectively. The average boundary layer velocities
during the battery discharge period. Consequently, it is an expected
at the lowest, middle and top points (x = 0 mm, x = 155 mm and x =
result that the heat transfer rate will increase with time. Besides, it can
210 mm) along the battery body at the minimum discharge rate (1C) are
be stated that the time-dependent increase in the heat transfer rate ac­
0.002 m/s, 0.050 m/s and 0.061 m/s, respectively.
celerates towards the end of the discharge.
Using the results of the experimental study, a correlation based on
Battery and air temperatures in the cross-section plane for certain
the discharge rate (DOD) and the current drawn from the battery (I) was
discharge rates is given in Fig. 13. The temperature boundary layer
developed with the regression analysis method for the estimation of the
formed on the battery surface was simulated, and as expected, it was
heat transfer rate through the battery surface under natural convection
determined that the battery temperature and temperature boundary
conditions and is given in Eq. (18). The results of the equation and the
layer thickness increased with the increase of the discharge rate.
experimental results are compared in Fig. 15. It is clear that the equation
Average boundary layer temperatures at the lowest, middle and top
can predict within a maximum ±20% error range. In the regression
points (x = 0 mm, x = 155 mm and x = 210 mm) along the battery body
analysis, a total of 505 data were used and 426 of these data were below
under the maximum discharge rate (5C) are 35.0 ◦ C, 35.5 ◦ C and
the ±10% error band, while only 79 of them were within the ±10–20%
34.2 ◦ C, respectively. Considering the minimum discharge rate (1C), the
error band. Moreover, the R2 value obtained in the regression analysis
average boundary layer temperatures of the lowest, middle and top
performed is approximately 0.99.
[ ]
Q̇surf,corr = 0, 593(DOD)3 − 0, 641(DOD)2 + 0, 856(DOD) + 0, 04 e0,023 I
(18)

5. Conclusions

In this study, the thermal performance of a LFP pouch type battery

with a capacity of 20 Ah under 23 ◦ C ambient temperature and natural
convection conditions was experimentally and numerically investigated.
The temperature change on the battery surface with the discharge of the
battery at 1C-5C rates was visualized via a thermal camera. As a result of
the experiments and calculations, the following conclusions were
reached:

• While the maximum temperature zone on the battery surface is

concentrated in the anode and cathode regions up to about DoD of
80%, it concentrates in the middle of the battery at the end of
discharge.
• There is a 56% increase in the average battery temperatures between
5C and 1C discharge rates. In addition, it was determined that the
Fig. 11. The variation of the homogeneity index at 1C-5C discharge rates.

10
O. Kalkan et al. Journal of Energy Storage 40 (2021) 102796

Fig. 13. The variation of air and battery temperatures for different discharge rates.

Fig. 14. The variation of air velocity vectors for different discharge rates.

11
O. Kalkan et al. Journal of Energy Storage 40 (2021) 102796

[3] M. Lenzen, Life cycle energy and greenhouse gas emissions of nuclear energy: a
review, Energy Conver. Manag. 49 (8) (2008) 2178–2199.
[4] M.K. Tran, M. Akinsanya, S. Panchal, R. Fraser, M. Fowler, Design of a hybrid
electric vehicle powertrain for performance optimization considering various
powertrain components and configurations, Vehicles 3 (1) (2021) 20–32.
[5] O. Ellabban, H. Abu-Rub, F. Blaabjerg, Renewable energy resources: current status,
future prospects and their enabling technology, Renew. Sustain. Energy Rev. 39
(2014) 748–764.
[6] T. Zhang, C. Gao, Q. Gao, G. Wang, M. Liu, Y. Guo, …, Y.Y. Yan, Status and
development of electric vehicle integrated thermal management from BTM to
HVAC, Appl. Therm. Eng. 88 (2015) 398–409.
[7] M. Aneke, M. Wang, Energy storage technologies and real life applications–a state
of the art review, Appl. Energy 179 (2016) 350–377.
[8] C. Peng, M.D. Liao, X.L. Lv, L. Chen, S.P. Hou, D. Min, …, J.H. Lin, Large-scale
synthesis of highly structural-connecting carbon nanospheres as an anodes material
for lithium-ion batteries with high-rate capacity, Chem. Eng. J. Adv. 2 (2020),
100014.
[9] P. Ramadass, W. Fang, Z.J. Zhang, Study of internal short in a Li-ion cell I. Test
method development using infra-red imaging technique, J. Power Sources 248
(2014) 769–776.
[10] J.X. Weinert, A.F. Burke, X. Wei, Lead-acid and lithium-ion batteries for the
Chinese electric bike market and implications on future technology advancement,
J. Power Sources 172 (2) (2007) 938–945.
[11] Celen, A., & Kalkan, O. Numerical Investigation on the usage of finned surface in
lithium nickel manganese cobalt oxides batteries by using air cooling method.
Fig. 15. Regression analysis of experimental and calculated heat transfer rates. Energy Storage, e216.
[12] Q. Wang, B. Jiang, B. Li, Y. Yan, A critical review of thermal management models
and solutions of lithium-ion batteries for the development of pure electric vehicles,
average battery temperatures exceeded 40 ◦ C, known as the upper Renew. Sustain. Energy Rev. 64 (2016) 106–128.
limit temperature, at the end of discharge at 3C, 4C and 5C rates. [13] Z. Zhang, J. Wang, X. Feng, L. Chang, Y. Chen, X. Wang, The solutions to electric
vehicle air conditioning systems: a review, Renew. Sustain. Energy Rev. 91 (2018)
• The highest measured maximum battery temperatures at 5C and 1C 443–463.
discharge rates are 52.2 ◦ C and 33.7 ◦ C, respectively. There was a [14] O. Kalkan, A. Celen, K. Bakirci, A.S. Dalkilic, Experimental investigation of thermal
55% increase between the highest and lowest discharge rates for performance of novel cold plate design used in a Li-ion pouch-type battery, Appl.
Therm. Eng. 191 (2021), 116885.
maximum battery temperature. As a result, it has been determined
[15] Panchal, S. (2014). Impact of vehicle charge and discharge cycles on the thermal
that natural convection cooling of the battery is insufficient for these characteristics of lithium-ion batteries (Master’s thesis, University of Waterloo).
discharge rates since the battery surface temperature difference [16] Panchal, S. (2016). Experimental investigation and modeling of lithium-ion battery
cells and packs for electric vehicles (Doctoral dissertation).
exceeded the 5 ◦ C.
[17] S. Yang, C. Ling, Y. Fan, Y. Yang, X. Tan, H. Dong, A review of lithium-ion battery
• The maximum and minimum heat removed from the battery surface thermal management system strategies and the evaluate criteria, Int. J.
is calculated as 11.5 W and 2.6 W for 5C and 1C discharge rates, Electrochem. Sci. 14 (7) (2019) 6077–6107.
respectively. [18] B.B. Kanbur, V. Kumtepeli, F. Duan, Thermal performance prediction of the battery
surface via dynamic mode decomposition, Energy 201 (2020), 117642.
• For the estimation of heat transfer rate through the battery surface, a [19] M. Akbarzadeh, T. Kalogiannis, J. Jaguemont, J. He, L. Jin, M. Berecibar, J. Van
correlation equation has been developed as a function of discharge Mierlo, Thermal modeling of a high-energy prismatic lithium-ion battery cell and
rate and current and has an error band of ± 20%. module based on a new thermal characterization methodology, J. Energy Storage
32 (2020), 101707.
• Values defined as temperature homogeneity index are calculated as [20] A. Mevawalla, S. Panchal, M.K. Tran, M. Fowler, R. Fraser, Mathematical heat
0.123, 0.135, 0.138, 0.151 and 0.147 for 1C, 2C, 3C, 4C and 5C transfer modeling and experimental validation of lithium-ion battery considering:
discharge rates, respectively. tab and surface temperature, separator, electrolyte resistance, anode-cathode
irreversible and reversible heat, Batteries 6 (4) (2020) 61.
[21] S. Zhu, J. Han, T.S. Pan, Y.M. Wei, W.L. Song, H.S. Chen, D. Fang, A novel designed
CRediT authorship contribution statement visualized Li-ion battery for in-situ measuring the variation of internal
temperature, Extreme Mech. Lett. 37 (2020), 100707.
[22] S. Wang, K. Li, Y. Tian, J. Wang, Y. Wu, S. Ji, Infrared imaging investigation of
Orhan Kalkan: Conceptualization, Methodology, Investigation, temperature fluctuation and spatial distribution for a large laminated lithium–ion
Writing - original draft, Writing - review & editing. Ali Celen: Funding power battery, Appl. Therm. Eng. 152 (2019) 204–214.
acquisition, Project administration, Conceptualization, Methodology, [23] X. Zhang, R. Klein, A. Subbaraman, S. Chumakov, X. Li, J. Christensen, …, S.
U. Kim, Evaluation of convective heat transfer coefficient and specific heat capacity
Writing - original draft, Writing - review & editing. Kadir Bakirci: of a lithium-ion battery using infrared camera and lumped capacitance method,
Methodology, Supervision, Writing - original draft, Writing - review & J. Power Sources 412 (2019) 552–558.
editing. [24] H.J. Kim, J.H. Lee, D.H. Baek, J.K. Lee, A study on thermal performance of
batteries using thermal imaging and infrared radiation, J. Ind. Eng. Chem. 45
(2017) 360–365.
[25] S. Panchal, I. Dincer, M. Agelin-Chaab, R. Fraser, M. Fowler, Experimental and
Declaration of Competing Interest
simulated temperature variations in a LiFePO4-20 Ah battery during discharge
process, Appl. Energy 180 (2016) 504–515.
The authors declare that they have no known competing financial [26] S.J. Bazinski, X. Wang, B.P. Sangeorzan, L. Guessous, Measuring and assessing the
effective in-plane thermal conductivity of lithium iron phosphate pouch cells,
interests or personal relationships that could have appeared to influence
Energy 114 (2016) 1085–1092.
the work reported in this paper. [27] J. Jaguemont, N. Omar, F. Martel, P. Van den Bossche, J. Van Mierlo, Streamline
three-dimensional thermal model of a lithium titanate pouch cell battery in
extreme temperature conditions with module simulation, J. Power Sources 367
Acknowledgement (2017) 24–33.
[28] C. Xu, F. Zhang, X. Feng, F. Jiang, D. Ren, L. Lu, …, M. Ouyang, Experimental study
This work was supported by Research Fund of the Erzincan Binali on thermal runaway propagation of lithium-ion battery modules with different
parallel-series hybrid connections, J. Clean. Prod. 284 (2021), 124749.
Yildirim University. Project Number: FBA-2019–657.
[29] X. Li, Z. Zhang, W. Wang, Y. Tian, D. Li, J. Tian, Multiphysical field measurement
and fusion for battery electric-thermal-contour performance analysis, Appl. Energy
References 262 (2020), 114518.
[30] J.H. Merkin, I. Pop, Conjugate free convection on a vertical surface, Int. J. Heat
Mass Transf. 39 (7) (1996) 1527–1534.
[1] L. Capuano, “International energy outlook 2020 [IEO2020),” [Online]. Available:
[31] T.L. Bergman, F.P. Incropera, D.P. DeWitt, A.S. Lavine, Fundamentals of Heat and
https://www.eia.gov/outlooks/ieo/pdf/ieo2020.pdf. Accessed: Jan. 10, 2021.
Mass Transfer, John Wiley & Sons, 2011.
[2] M. Höök, X. Tang, Depletion of fossil fuels and anthropogenic climate change—A
review, Energy Policy 52 (2013) 797–809.

12
O. Kalkan et al. Journal of Energy Storage 40 (2021) 102796

[32] S.W. Churchill, H.H. Chu, Correlating equations for laminar and turbulent free [34] H. Zhang, C. Li, R. Zhang, Y. Lin, H. Fang, Thermal analysis of a 6s4p Lithium-ion
convection from a vertical plate, Int. J. Heat Mass Transf. 18 (11) (1975) battery pack cooled by cold plates based on a multi-domain modeling framework,
1323–1329. Appl. Therm. Eng. 173 (2020), 115216.
[33] K.H. Kwon, C.B. Shin, T.H. Kang, C.S. Kim, A two-dimensional modeling of a
lithium-polymer battery, J. Power Sources 163 (1) (2006) 151–157.

13