Journal of Energy Storage 29 (2020) 101377

Contents lists available at ScienceDirect

Journal of Energy Storage

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

Studies on thermal management of Lithium-ion battery pack using water as T

the cooling fluid
S.D.V.S.S. Varma Siruvuri, P.R. Budarapu⁎
School of Mechanical Sciences, Indian Institute of Technology, Bhubaneswar 752050, India

ARTICLE INFO ABSTRACT

Keywords: Study of battery thermal management is critical for safe and better performance of Lithium-ion batteries, con-
Lithium-ion batteries sidering several recent battery failures and explosions. Lithium-ion batteries are generally used in stacks to meet
Battery thermal management system the high energy requirements. Thus, the heat generated in a battery pack must be properly managed for efficient
Liquid coolant operation. A battery thermal management system (BTMS) is an ideal tool to understand the thermal health of a
Flow rate
battery. An attempt is made here to study the influence of water as the coolant on the battery thermal man-
Flow reversal
agement. A variable heat generation that depends on the discharge rate is assumed in the simulations. A 5C
charge/discharge rate (per hour) per battery and two packs of 5 batteries each are considered for the analysis.
Due to symmetry only half pack (2.5 batteries) with symmetry boundary conditions is considered in the
simulations. The influence of volume flow rate, flow direction, and contact area on the battery thermal beha-
viour are analysed. Results indicate that a 5C charge/discharge produces maximum temperatures of 319.31 K
and 316.93 K at flow rates of 0.25 × 10 6 m3/s and 1.6 × 10 6 m3/s, respectively. Whereas, the uncooled
batteries with 5C charge/discharge produces a maximum temperature of 336.623 K. Therefore, the uncooled
peak temperature is estimated to be reduced by 5.85% by circulating the coolant at 1.6 × 10 6 m3/s.
Furthermore, by reversing the direction of water flow in one of the channels, leads to redistribution of heat and
hence the temperature distribution in the battery pack is observed to be more uniform. Flow reversal also helped
to reduce the peak temperatures by 2.77 K in the battery pack. All the calculations are performed using the
commercial software COMSOL.

1. Introduction must be properly managed to avoid thermal runaway, electrolyte fire

and explosion [8]. Therefore, an effective battery thermal management
With rapid developments in Lithium-ion batteries (LiBs) in the re- system (BTMS) is essential for efficient performance of Lithium-ion
cent years, they have been extensively used in several applications batteries. The heat generated by electrical resistance and electro-
ranging from small power requirements in micro devices to battery chemistry reactions in Lithium-ion battery pack of 14 prismatic bat-
operated vehicles [1,2]. Due to the fast depletion of conventional en- teries was investigated in [9]. The lifetime of commercial Lithium
ergy sources, and their quite operation and less environment pollution, battery systems subjected to cyclic loads and correlations of capacity
battery operated vehicles, also known as electric vehicles (EVs) are versus weight and resistance, storage of energy are reported in [10–12].
playing a significant role in transportation [3]. Lithium-ion batteries Degradation of battery performance and failure is a complex phe-
possess high power density, energy efficiency, long life span and cycle nomena associated with the non-linear systems such as Lithium-ion
life, low self-discharge rate, operation safety, non-existence of the batteries. The chemistry of electrode materials in Lithium-ion batteries
memory effect and low operating costs as compared to other batteries, and the heat generation is studied in [13] at various charge and dis-
which made them the first choice for the electric vehicles [4,5]. A Li- charge rates through a multiphysics modeling and computer simula-
thium-ion battery consists of several electrochemical cell units con- tion. Some parameters influencing the battery performance include:
nected in parallel. The performance of LiBs is closely related to thermal design, chemistry, dynamic environment, and the actual operating
and Li+ transport behaviour [6]. Therefore, the performance and life of conditions, such as: temperature, discharge rate [14,15]. The heat
Lithium-ion batteries are sensitive to temperature fluctuations during generation phenomena and critical thermal issues of Lithium-ion bat-
the charge and discharge process [7]. Such temperature fluctuations teries to provide an effective thermal management solution for the high

⁎
Corresponding author.
E-mail address: [email protected] (P.R. Budarapu).

https://doi.org/10.1016/j.est.2020.101377
Received 12 August 2019; Received in revised form 24 February 2020; Accepted 15 March 2020
Available online 23 April 2020
2352-152X/ © 2020 Elsevier Ltd. All rights reserved.
S.D.V.S.S.V. Siruvuri and P.R. Budarapu Journal of Energy Storage 29 (2020) 101377

Fig. 1. (a) Schematic of the battery pack highlighting various dimensions, where the batteries are arranged in two half packs. Each half pack consists of 2.5 batteries,
where the symmetry boundary conditions are applied on the cross section of cell 3. (b) Battery pack with circuitous channel in design 1. Cooling liquid is passed
through the circuitous rectangular cooling channels placed between each layer of batteries. (c) Dimensions of the circuitous channel considered in design 1. (d)
Design 2, where all the arrangements of design 1 remain same, except the cooling circuit with an extra turn is used. This will help to increase surface area of the
cooling fluid. (e) Meshed battery pack including the circuitous channel.

energy density Lithium-ion batteries, considering operating tempera- cooling fluids and reported that perfluorinated polyether provided the
ture, different cooling or heating methods like: thermoelectric coolers optimal temperature range. A mini channel liquid cooled cylinder was
[16], heat pipe [17] and optimisation methods [18] are reviewed in proposed in [4], where the authors reported that the maximum tem-
[19–25]. The effect of temperature on the capacity fade and aging of perature could be controlled under 313 K, by maintaining no less than
Lithium-ion batteries are investigated in [26], considering the electrode four mini-channels and an inlet mass flow rate of 10 3 kg/s. An alu-
structure, including electrode thickness, particle size and porosity. minium based mini-channel tubes is tested on a Lithium-ion cell under
Several analytical [27], numerical [28] and computational techniques different discharge rates for its thermal management, using different
[29,30] to analyse heat generation during discharge of battery opera- flow rates and configurations in [39]. A flow boiling in mini-channel
tion were proposed in the literature. A dimensional analysis is applied utilising dielectric hydrofluoroether liquid based thermal management
in [31] to predict the Lithium ion battery’s energy density in order to system is discussed in [40]. A liquid cooling method with U-shaped
obtain the sets of parameters that influence the performance, where the serpentine-channel configuration, to study the effect of cooling with
authors have proposed a novel empirical model to predict the energy number and layout of channels along with the inlet temperature of
density. coolant, on cooling performance of BTMS are analysed [41]. A chan-
However, for fast charge and discharge processes air cooling is not neled liquid cooling thermal management system of Lithium-ion battery
efficient [32–35]. Alternately, liquid cooling will be more efficient and pack for electric vehicles to study the thermal behaviour, and hence to
suitable for cooling large surface areas [36]. The thermal performance investigate the effects of discharge rates and the heat exchange area
of conventional dry air cooling and mist cooling is studied using ex- between neighbouring batteries is discussed in [42]. At low C-rates (C/
periments and numerical simulations [37] and the mist cooling is re- 10), particularly at temperatures below 0∘ C, Li-ion cells exhibits a re-
ported to offer lower and more uniform temperature distribution versible capacity fade, mainly influenced by anodic and cathodic acti-
compared to dry air cooling. The thermal behaviour of Lithium-ion vation energies [43]. The efficacy of using vortex generators in the
batteries under load and the capability of various cooling fluids in cooling channels of lithium-ion battery BTMS was investigated in
maintaining the working conditions within a safe range was in- [44,45].
vestigated in [38], where the authors tested the capability of different In this study, a new circuitous rectangular channel is designed for

2
S.D.V.S.S.V. Siruvuri and P.R. Budarapu Journal of Energy Storage 29 (2020) 101377

the circulation of liquid coolant at various flow rates. The maximum respectively, called as coarse, regular and fine grid sizes in the con-
operating temperature in the battery pack is controlled by varying the vergence studies. Considering the coarse grid, the peak temperature in
flow rate. The effects of various parameters like volume flow rate, the battery pack is observed to be 319.298 K, whereas, the peak tem-
contact area, and the flow direction of the cooling fluid on the tem- perature is found to be 319.31 K and 319.53 K, respectively, by
perature distribution across the battery pack are also studied. The main adopting regular and fine grid sizes, respectively. Therefore, a regular
novelties of the present study are: (i) design of the circuitous cooling grid size with the maximum and minimum element sizes of 20 mm and
channels such that major portion of the battery will be in contact with 3.78 mm for cells, and 6.39 mm and 1.22 mm for both cooling plates
cooling medium, as compared to simple straight channels. The pro- and cooling channels, respectively, are adopted in all the simulations. A
posed cooling channel is an integral part of the intermediate plate be- discretised battery pack considering regular mesh is shown in Fig. 1(e).
tween two cells. Furthermore, in electric vehicles, the cells are gen- The total number of elements in the present model are found to be
erally stacked in multiple rows and columns. Therefore, the influence of 302,871 with 247,417 elements of 4 node tetrahedron, 524 elements of
a cell on the heat dissipation characteristics of another cell placed next 5 node pyramid and 54,930 elements of 6 node prism, and the total
to it, is required to be addressed. (ii) Studies on the influence of volume number of nodes are equal to 481,480.
flow rate of coolant, contact area, and the flow direction on the tem- A constant C-rate is assumed during the charge/discharge. Coolant
perature distribution across the battery pack. (iii) Flow reversal, which flow is assumed to be fully developed laminar with maximum Reynolds
is found to help in redistributing the heat and hence the temperature number equal to 2300. Batteries are held in contact under atmospheric
distribution in the battery pack, apart from reducing the peak operating pressure (101325 Pa) and contact surface roughness-asperities average
temperatures by 2.77 K in the battery pack. height equal to 2.6 × 10−6 m [46]. The amount of heat transferred to
The arrangement of the article is as follows: Section 2 is devoted for the cooling plate is given by heat participation coefficient based on the
discussing the modelling aspects of Lithium-ion battery packs in Charron’s relation. The pressure at the outlet is also assumed to be at-
COMSOL. Results considering the influence of flow rate of cooling fluid, mospheric. In numerical simulations, the one dimensional Lithium-ion
flow area and reversed flow direction are discussed in Section 3. Key battery module and the three dimensional heat transfer module are
results are summarised in Section 4. coupled through a variable heat generation given by:

2. Modelling aspects (L neg + Lsep + L pos)

Qh = comp1.aveop1(comp1.liion.Qh )*
Lbatt (1)
In order to estimate the heat generated in the battery pack, in-
dividual batteries/cells are arranged in a rectangular (prismatic) where Qh is heat generated in one dimensional lithium-ion model
fashion, as shown in Fig. 1(a). Fig. 1(b) shows the battery pack along (comp1.liion) using an average coupling component for cathode, anode
with the circuitous channel in design 1. Dimensions of the cooling and separator (aveop1) and L neg , Lsep , L pos and Lbatt are the lengths of
channel used in design 1 are shown in Fig. 1(c). The influence of water anode, separator, cathode and battery, respectively. This coupling is
as the cooling fluid is tested by estimating the maximum temperature in achieved in two steps: First, the electric current and the heat generation
the battery pack at various flow rates, see Fig. 1(b). Due to symmetry terms are estimated through the Lithium-ion battery model considering
only half pack of batteries are considered in the simulations. Each half the cathode, anode and separator. In the second step, the generated
pack consists of 2.5 batteries, where the symmetry boundary conditions heat is used to predict the temperature profile of the cells by coupling
are applied on the cross section of cell 3, see Fig. 1(a). Cooling liquid is the heat transfer module. In the heat transfer module, it is assumed that
passed through a 1 mm × 5 mm circuitous rectangular channel, cut out the properties like thermal conductivity, thermal capacity and density
in a 80 mm × 120 mm × 2 mm aluminium sheet. The aluminium sheet are volume averaged. Modules such as heat transfer and laminar flow
is sandwiched between the Lithium-ion batteries of size are coupled by employing the multiphysics non-isothermal flow, and
80 mm × 120 mm × 12 mm. The simulations are performed con- Kays-Crawford heat transport model. Work done due to change in
sidering Lithium-Manganese Oxide prismatic Battery with LiPF6 elec- pressure and viscous dissipation are also considered.
trolyte in ethylene carbonate (EC): ethylmethyl carbonate (EMC) (3:7 Specific flow rate of cooling fluid is prescribed at the inlet, whereas,
by volume) for liquid electrolyte, Graphite electrode LixC6 mesocarbon atmospheric boundary conditions are maintained at the outlet. Free
microbeads (MCMB) for anode and (Lithium Manganese oxide) LMO convection boundary conditions with heat transfer coefficient equal to
electrode LiMn2O4 Spinel for cathode are considered, with a battery 5 W/m2K are considered along the sides of the battery pack. Initially
capacity of 1.5 Amp-hour and a discharge current of 90 Amps. The heat both the Lithium-ion battery packs and the inlet water are maintained
generated is estimated based on the battery charge/discharge rate. A 5C at the ambient temperature, i.e., 310 K. The battery is assumed to
charge/discharge rate (per hour) per battery and two packs of 5 bat- charge/discharge at a rate of 5C lasting for a period of 62.234 s. Cooling
teries each are considered for the analysis. However, a discharge cur- plates are made of aluminium, and water is considered as the cooling
rent of 90 Amps is considered in the analysis, and the capacity of liquid. Thermal properties of cooling plate, water and the cell used in
adopted battery is 1.55 Amp-hr, such that the charge/discharge hap- this study are summarised in Table 1, where the coefficient of thermal
pens in 62.234 s. Heat generated from the batteries during charge/ conductivity (λ) of Lithium-ion battery is listed in Table 2.
discharge is transferred to the cooling plate through the contact sur- Furthermore, in order to study the effect of coolant surface area in
faces, which is further transferred to the cooling liquid flowing through contact with batteries, a circuit having four coils, see Fig. 1(d), is also
the channels. tested in the analysis. We name the cooling circuit with 3 coils as ‘de-
The influence of mesh size is throughly tested before fixing the sign1’ (D1) and the circuit with 4 coils as ‘design2’ (D2).
element size of various components of the battery pack. For instance,
considering the thickness of cooling plate as 2 mm, which is the lowest
dimension in the battery pack, the element sizes are selected accord-
Table 1
ingly. A grid convergence test is performed by considering design 1
Material and geometric properties of the battery pack [47].
with flow rate equal to 0.25 × 10 6 m3/s. This is because, the tem-
perature distribution for this particular case is found to be higher, as Material ρ(kg/m3) Cp(J/kgK) λ(W/mK) μ(Pa-s)
compared to other cases. Three different grid sizes, where the minimum
Water 1000.0 4128.0 0.6 1.003 × 10 3
element size is associated with the cooling plate are selected as:
Aluminium 2700.0 900.0 238.0 -
1.92 mm, 1.22 mm and 0.639 mm, and the element sizes associated Lithium-ion battery 2055.2 1399.1 Diagonal -
with the battery are selected as: 5.88 mm, 3.78 mm and 2.1 mm,

3
S.D.V.S.S.V. Siruvuri and P.R. Budarapu Journal of Energy Storage 29 (2020) 101377

Table 2 were considered, with a constant 5C charge/discharge. To begin with,

Coefficient of thermal conductivity (λ) of Lithium-ion batteries [47]. the maximum temperature observed in the battery pack without any
X Y Z cooling is found to be 336.623 K, see Fig. 2(a). The maximum tem-
perature observed at the lowest flow rate, i.e. 0.25 × 10 6 m3/s, is
X 29.55 0 0 equal to 319.31 K, as shown in Fig. 2(b), which drops down to 316.93 K
Y 0 0.89724 0
as flow rate is increased to 1.6 × 10 6 m3/s, see Fig. 2(c). Continued
Z 0 0 29.55
increase in mass flow rate leads to the decrease of maximum tem-
perature in the battery pack in a linear fashion, see Fig. 2(d). Therefore,
3. Results and discussion introduction of cooling fluid reduces the peak temperature of the bat-
tery pack by a minimum of 17.7 K (5.143 %) at 0.25 × 10 6 m3/s flow
The performance of lithium-ion battery pack is significantly influ- rate and a maximum of 20.07 K (5.85 %) at 1.6 × 10 6 m3/s flow rate,
enced by the surface area of cooling fluid identified by the number of as compared to the case without cooling.
cooling channels, volume flow rate and the direction of flow of the As shown in Fig. 1(a), cell 2 in the battery pack is sandwiched be-
cooling fluid. Thus, the influence of above three variables on the tween cells 1 and 3, and is connected to the cooling channels on both
thermal management of battery pack is investigated here. The cooling sides, resulting in an efficient cooling. Distribution of temperature
performance is indicated by the reduction in maximum temperature of across all the surfaces of cell 2 considering design 1, at flow rates of
battery pack. Maximum temperature is captured at the end of full 0.25 × 10 6 m3/s, 1.0 × 10 6 m3/s, and 1.6 × 10 6 m3/s are shown in
charge. Fig. 3. Based on the temperature distribution in Fig. 3, at lower flow
rates the hot areas are observed to be concentrated in the lower portion
of the cell, see Fig. 3(a) and (b). With increase in flow rate, the hot areas
3.1. Influence of volume flow rate are more distributed, as shown in Fig. 3(c), apart from the drop in peak
temperature. Furthermore, due to the presence of cooling surfaces on
In order to pin-point the effect of volume flow rate, seven different both sides of the cell 2, the temperature distribution along either side of
volume flow rates ranging from 0.25 × 10−6 m3/s to 1.6 × 10 6 m3/s

Fig. 2. Temperature distribution of the battery pack (a) without cooling and with flow rate equal to: (b) 0.25 × 10 6 m3/s and (c) 1.6 × 10 6 m3/s. (d) Variation of
the maximum temperature in the battery pack with flow rate.

4
S.D.V.S.S.V. Siruvuri and P.R. Budarapu Journal of Energy Storage 29 (2020) 101377

Fig. 3. Temperature distribution in cell 2 considering design 1, with different flow rates: (a) 0.25 × 10 6 m3/s (b) 1.0 × 10 6 m3/s (c) 1.6 × 10 6 m3/s.

Fig. 4. Temperature distribution of the battery pack with flow rate equal to 1.0 × 10 6 m3/s, considering (a) design 1 and (b) design 2.

cell 2 is 317.7 K, whereas, the peak temperature in the entire battery

pack is 318.13 K.

3.2. Influence of increase in flow area

In order to study the thermal performance of the battery pack with

extra coolant surface area, an extra coil in the circuitous rectangular
cooling channel is added, see Fig. 1(b)–(d). Thus, the design with extra
coil, known as design 2, leads to an increase of 13.8 % surface area as
compared to design 1. Simulations are performed considering different
flow rates and the obtained peak temperatures are compared to the
peak temperatures obtained considering design 1, see Figs. 4 and 5.
Distribution of the temperature in the battery pack at a flow rate of
1.0 × 10 6 m3/s, without (design 1) and with an extra surface area
(design 2) are shown in Fig. 4(a) and (b), respectively. The peak tem-
peratures in designs 1 and 2 as a function of flow rate of the cooling
fluid is plotted in Fig. 5. According to Fig. 5 a peak temperature dif-
Fig. 5. Comparison of maximum temperature as a function of flow rate in de- ference of 0.934 K at 1.6 × 10 6 m3/s and 2.127 K at 0.25 × 10 6 m3/s
sign 1, design 2 and flow reversal. were observed. Peak temperature variation by reversing the flow
through channels under different flow rates is compared with that of
designs 1 and 2 in Fig. 5. According to Fig. 5, the peak temperatures in
the battery surface is symmetric. To summarise, due to the presence of
design 2 and flow reversal are almost same. However, a significant
cooling surfaces on both sides of cell 2, it is efficiently cooled as com-
difference in peak temperatures with flow rate is evident between de-
pared to cell 1. This is confirmed by comparing the temperature dis-
sign 1 and, design 2 and flow reversal. Variation of temperature across
tribution at flow rate 1.0 × 10 6 m3/s, where the peak temperature in
the thickness of cell 2, at different locations along the x-direction with

5
S.D.V.S.S.V. Siruvuri and P.R. Budarapu Journal of Energy Storage 29 (2020) 101377

Fig. 6. Temperature profile across the thickness of the cell 2, along (a) the x-direction and (b) the z-direction. A flow rate of 1.0 × 10 6 m3/s is considered to generate
the plots.

Fig. 7. Temperature distribution along the thickness direction in cell 2 with flow rate 1.0 × 10 6 m3/s, considering (a) design 1 (b) flow reversal in design 1 (c)
design 2 and (d) flow reversal in design 2.

flow rate equal to 1.0 × 10 6 m3/s is plotted Fig. 6(a). Three different the temperature on the cell surfaces (front and back) is lower than cell
locations along the x direction, i.e. x = 0, 40 mm and 80 mm are temperature in the middle of the cell thickness, see Fig. 6(a). Therefore,
considered and at each x location, temperatures at four different loca- based on Fig. 6(a), as the distance along the thickness (y) direction
tions along the y direction are plotted in Fig. 6(a). Furthermore, the increases the temperature is initially found to be increased, followed by
temperature variation considering designs 1 and 2 and flow reversals at a reduction, where the temperature is initially raised from 315.75 K to
the same locations are also shown in Fig. 6(a). According to Fig. 6(a), 316.5 K, followed by a drop to 315.5 K. This is because, the points on
the temperature variation along the cell thickness considering design 1 cell 2 that are located close to cell 1, are exposed to higher temperatures
is not significant. The main reason for this is that the cooling fluid is as the outer surface of cell 1 is cooled by natural convection, see
heated up to its maximum by the time it reaches the bottom of cell 2, Fig. 7(b)–(d). Furthermore, the heat generated in cell 2 also adds up,
see Figs. 3 and 7(a). This is also confirmed by the fluid temperature in raising the temperature in the middle. On the other hand, the points
cooling channels at various flow rates in design 1, shown in close to cell 3, are exposed to the cooling fluid which is relatively at
Fig. 8(a)–(c). Therefore, the temperature is uniformly distributed (along lower temperature compared to that of with coolant present in the
the cell thickness) at the mid section, as shown in Fig. 6(a). On the other cooling channel sandwiched between cell 1 and cell 2, and thus the heat
hand, as shown in Fig. 7(b)–(d), the temperature distribution along the removal is efficient. The over all drop in temperature, along the
cell thickness is not uniform for design 2 and flow reversal. As a result, thickness direction of cell 2 is observed to be around 0.45 K and the

6
S.D.V.S.S.V. Siruvuri and P.R. Budarapu Journal of Energy Storage 29 (2020) 101377

Fig. 8. Temperature distribution at the center of the width of the cooling channel sandwiched between cell 1 and cell2 (a) Design 1 with flow rate 0.25 × 10 6 m3/s
(b) Design 1 with flow rate 1.0 × 10 6 m3/s (c) Design 1 with flow rate 1.6 × 10 6 m3/s (d) Design 2 with flow rate 1.0 × 10 6 m3/s (e) Flow reversal in design 1
with flow rate 1.0 × 10 6 m3/s (f) Temperature difference considered between the water in the cooling channel and the nearest cell 2 surface across the height of the
cell in design 1 with flow rate 1.0 × 10 6 m3/s .

peak temperatures in design 2 are found to be 1.0 K - 1.5 K lower than centre along thickness direction. The reason for this is that the fluid in
the peak temperatures in design 1. the cooling channels on both sides of the cell 2 is at relatively higher
Furthermore, temperature of cell 2 along the z direction at top and temperatures, see Fig. 8(a)–(c). Furthermore, as the temperature of the
bottom locations, through the thickness, for designs 1 and 2 and flow coolant in the cooling channel increases from top to bottom, see
reversal are plotted in Fig. 6(b). In design 1, variation of cell 2 tem- Fig. 8(a)–(c), the temperature across the cell thickness at the mid height
perature along the thickness at the bottom of the cell is observed to be of the cell remain constant at ≈317 K, as shown in Figs. 6(a) and 7(a).
reducing towards the centre and increasing towards the surfaces. This is According to Fig. 8(a)–(e), the peak temperature in the cooling channel
due to heating of the cooling fluid by the time it reaches cell bottom and is observed to be reducing with increase in flow rate, and the reduction
hence higher temperature on the cell surfaces, see Figs. 6(b), 7(a) and in design 2 is more effective than that of design 1. The temperature
8(a)–(c). Such high temperatures are not uniformly distributed along difference between the water in the cooling channel and the nearest
the cell thickness, leading to slight drop in temperature towards the surface of the cell 2, along the z direction for designs 1 and 2 and flow

7
S.D.V.S.S.V. Siruvuri and P.R. Budarapu Journal of Energy Storage 29 (2020) 101377

Fig. 9. Flow reversal considering design 1: (a) case 1 and (b) case 2.

Fig. 10. Temperature distribution of the battery pack with flow rate equal to 1.0 × 10 6 m3/s in (a) FR1 (b) FR2.

reversal are plotted in Fig. 8(f). The temperature difference in Fig. 8(f), 3.3. Influence of direction of flow
is observed to be continuously increasing with the distance along the z
direction for all cases. Whereas, the temperature difference is highest As observed in Section 3.2, due to the conventional flow of the
for design 1 and lowest for design 2 and flow reversal, for the reasons cooling fluid (entry from top and exit from bottom or entry from bottom
explained above. and exit from top) through all the channels, higher temperatures are
To summarise, distribution of temperature across the thickness of concentrated on one side of the battery pack. One way to reduce such
cell 2, on left and right and the bottom and top locations along the x steep temperature gradients is to reverse the direction of flow in the
and z-directions, considering designs 1 and 2, are plotted in Fig. 6(a) alternate channels, as indicated in Fig. 9. In this study, two different
and (b), respectively. Based on Fig. 6(b), a temperature difference of 3 K cases, as shown in Fig. 9 are identified to reverse the direction of flow of
is noticed between the top and bottom surfaces, where the top surface is the cooling fluid, considering design 1, and named as flow reversal in
cooler than the bottom surface. The reason being, the cold fluid en- design 1 (FR1). The main idea in cases 1 and 2 shown in Fig. 9, is to
tering at the top of the battery pack absorbs maximum heat at the top reverse the flow direction in alternate cooling channels. This means, the
raising its temperature, as it reaches the bottom of the battery pack. inlet of cooling channel 1 (inlet 1 in Fig. 9(a)) will be the outlet of
Therefore, the bottom of the battery pack will always be at higher cooling channel 2 (outlet 2 in Fig. 9(a)) and vice-versa. Similar analogy
temperatures, compared to the top battery pack, see Figs. 2 and 4. is applied to the second pack of batteries in Fig. 9(b). The above two
Furthermore, in design 2, flow reversals 1 and 2 are found to be effi- cases are repeated considering design 2 as well, and those studies are
cient in cooling the battery pack as compared to design 1, see Figs. 6, 7 indicated as flow reversal in design 2 (FR2). The peak temperature of
and 8. the pack with different flow rates considered in FR1 is plotted in Fig. 5
and observed that FR1 is efficient compared to conventional flow in
design 1 and also the peak temperatures are observed to be decreasing
at a slower rate with distance. A volume flow rate of 1.0 × 10 6 m3/s is

8
S.D.V.S.S.V. Siruvuri and P.R. Budarapu Journal of Energy Storage 29 (2020) 101377

Fig. 11. Temperature distribution of cell 2 with flow rate equal to 1.0 × 10 6 m3/s, (a) considering design 2, and with (b) FR1 and (c) FR2.

Fig. 12. Temperature profile of the battery pack along the y direction in design
1 and flow reversal 1.

considered in all further flow reversal simulations. Distribution of

temperature in the battery pack in FR1 and FR2 are shown in Fig. 10(a)
and (b), respectively. Furthermore, the temperature distribution in cell
2 considering design 2 and flow reversals are shown in Fig. 11.
Comparing Figs. 11(b), (c) and 3(b), the maximum temperature in
cell 2 is found to be reduced by 1 K, by reversing the direction of flow of
the cooling fluid. This is due to the fact that the inlet for the cold
coolant is always on one side in the conventional flow. Therefore, due
to the maximum temperature difference between the coolant and the
battery surface at the top, the coolant absorbs the maximum energy
from the top of the battery pack and becomes hot while travelling to the
bottom, as shown in Fig. 8. As a result, the conventional flow always
Fig. 13. (a) Temperature profile in the cell 2 along the z direction, at different x
results in steep temperature gradient along the z direction of the battery
and y locations. (b) Temperature profile in the cell 2 along the z direction, at
pack, see Fig. 4. On the other hand, a uniform temperature at the top
different x locations and at the centre of the cell in the thickness direction.
and bottom surfaces can be achieved by reversing the flow direction of
the cooling fluid as shown in Fig. 9. Accordingly, a change in tem-
perature distribution across the thickness of the battery pack is evident redistributed from the bottom to the top of the battery pack in FR1 and
in Figs. 10. FR2, no significant drop/change in the temperature across the thickness
The influence of reversed flow direction of the cooling fluid on the of cell 2 is observed, see Figs. 11 and 7(b) and (d). In other words, the
temperature distribution considering designs 1 and 2, along the thick- trend is found to be similar for both FR1 and FR2, with FR2 being more
ness (y) direction at various locations along the x and z directions are effective in maintaining lesser temperatures, as shown in Fig. 6(a).
plotted in Figs. 11 and 6(a) and (b), respectively. Since the heat is However, the temperature distribution along the thickness (y) direction
at various locations at the top and bottom z locations for designs 1 and

9
S.D.V.S.S.V. Siruvuri and P.R. Budarapu Journal of Energy Storage 29 (2020) 101377

Fig. 14. Variation of temperature with time along the y and z directions, in (a) design 1 and (b) flow reversal 1.

Fig. 15. Temperature distribution in cell 2 in design 1, considering 0.25 × 10 6 m3/s flow rate at (a) 3C-rate (b) 5C-rate and (c) 10C-rate.

2, are compared to the flow reversals 1 and 2 in Fig. 6(b). According to in temperature of 3 K from the bottom to the top of cell 2 in the con-
Fig. 6(b), both FR1 and FR2 follow a similar trend at z = 0 and ventional flow is also observed. On the other hand, at z = 0, reversing
120 mm, with the heat dissipation being more efficient in FR2. A drop the flow direction of the cooling fluid was instrumental to redistribute

10
S.D.V.S.S.V. Siruvuri and P.R. Budarapu Journal of Energy Storage 29 (2020) 101377

summarise, by reversing the flow direction of the cooling fluid the

operating life of the battery pack can be significantly improved.
Variation of the temperature along the thickness of the battery pack
in design 1 and flow reversal 1 at different heights along the z direction
are plotted Fig. 12. The temperature profiles are captured at various
locations along the z direction, i.e. at z = 14, 60 and 106 mm from the
bottom surface, see Fig. 10. According to Fig. 12, an increase of tem-
perature from 316 K to 318 K, with a slight drop in cell 2, is observed at
the bottom of the battery pack (z = 14 mm) in design 1. Whereas, by
reversing the flow direction the temperature rises from 315.9 K to
316.3 K in cell 1, which further drops to 313.8 K in cell 2, and finally
reached to 316 K in cell 3, see Fig. 12. This is due to the presence of cold
fluid on one side of cell 2 and hot fluid on the other side of cell 2, at the
bottom of the battery pack during the flow reversal. A similar trend is
observed at the middle of the battery pack (z = 60 mm) as well, where
the temperature is slightly increased from 315.9 K to 317 K in design 1.
On the other hand, the temperature fluctuates about 316 K along the
cell thickness in flow reversal 1, since a combination of cold and hot
Fig. 16. Comparison of peak temperature with 0.25 × 10 6 m3/s flow rate and
fluids will be available on either side of cell 2 at the top of the battery
without cooling at different C-rates.
pack, leading to a redistribution of temperature, see Fig. 10. This in-
dicates a more uniform distribution of the heat across the thickness of
the heat at the bottom of the battery pack. This happened because an the battery pack with the flow reversal, as highlighted in Fig. 10, as
alternative cold inlet and hot outlet are available on either surface of opposed to the concentrated hot/cold regions estimated in the con-
cell 2. ventional flow in Fig. 4. The redistribution of heat with flow reversal
Furthermore, the presence of alternative cold inlet and hot outlet leads to a uniform temperature in the middle (z = 60 mm) of the
leads to the distribution of heat from y = 0 and y = 12 mm, as shown in battery pack. However, the temperature drops and rises along the
Fig. 6(b), where the temperature in cell 2 is noticed to be reduced from thickness at the top (z = 106 mm) of the battery pack, in both design 1
316.5 K at y = 0 to 313.5 K at y = 12 mm. Therefore, an over all drop and flow reversal 1 due to the availability of coolant at lower tem-
of 2.77 K is achieved at y = 12 mm by redistributing the heat energy peratures, as shown in Figs. 4 and 10.
through flow reversal. In the similar lines, the temperature in cell 2 is A comparison of temperature profiles in cell 2, considering design 1
increased from 314 K at y = 0 to 316.5 K at y = 12 mm in the flow and FR1, along the z direction at different x and y locations are shown
reversal, see Fig. 6(b). However, the temperature remains almost same in Fig. 13(a). Distribution of temperature along the z direction at y =
at 313.5 K in the conventional flow. Furthermore, the temperatures at y 14, 20 and 26 mm, i.e. on the outer, middle and inner surfaces of cell 2
= 12 mm, z = 0 in the flow reversal, are same (313.5 K) as the tem- considering design 1 and FR1 are plotted in Fig. 13(a). As highlighted in
peratures at z = 120 mm in the conventional flow and vice-versa. This Fig. 13(a), a more uniform temperature on the surface at y = 20 mm,
confirms the redistribution of thermal energy from the bottom to the along the z direction is observed with flow reversal, where the tem-
top of the battery pack and hence, the redistribution of temperatures. perature is observed to change from 316.45 K to 316.3 K at various
As a result of redistribution the entire surface of the cell will be sub- locations along the x direction, i.e., x = 20, 40 and 60 mm from the
jected to uniform temperatures by reversing the coolant flow direction, lower left corner. Whereas, the temperature continue to drop from
as shown in Fig. 11(b) and (c). On the other hand, in design 1, the top bottom to top (with and without flow reversal) on the surface at y =
half of the cell surface is always subjected to low temperatures and the 14 mm. However, the temperatures on the surface at y = 14 mm, on the
bottom half to high temperatures, see Fig. 11(a). Such thermal gra- top of cell 2 are lower with flow reversal. The surface at y = 14 mm on
dients are not good for the long service life of the battery pack. To cell 2 is close to one side of the cell 1, whose the other side of cell 1 is

Fig. 17. Temperature distribution of the battery pack with flow rate equal to 0.25 × 10 6 m3/s, considering (a) design 1 and (b) design 3.

11
S.D.V.S.S.V. Siruvuri and P.R. Budarapu Journal of Energy Storage 29 (2020) 101377

Fig. 18. Temperature distribution of the cells 1 and 2 from designs 1, 2 and 3 with 0.25 × 10 6 m3/s flow rate in cell 1 with (a) design 1 (b) design 2 (c) design 3 and
in cell 2 with (d) design 1 (e) design 2 and (e) design 3.

exposed to the atmosphere, whereas, the surface at y = 20 mm on cell 2 temperatures (0.25 K) are observed in FR2, compared to FR1. More-
is in the middle. Therefore, a flow reversal brings in more uniform heat over, the extra cooling surface area resulted in lower temperatures (by
distribution, which is evident from the more uniform temperature dis- 0.65 K) in design 2, compared to design 1.
tribution in Fig. 13(a). Also, according to Fig. 8(f), the temperature Variation of temperature with time along the y and z directions in
difference is lowest with flow reversal as compared to design 1. Fur- design 1 and flow reversal 1 are plotted in Fig. 14(a) and (b), respec-
thermore, FR1 and design 2 are efficient in reducing the peak tem- tively. Based on Fig. 14 (a), the temperature continuously rises with
peratures in the battery pack compared to design 1. On the other hand, time. Moreover, the temperature is observed to be lower on the back
the surface at y = 26 mm on cell 2 is located close to cell 3, where a surface of the cell, compared to the centre. The temperature difference
flow reversal is almost inverted the temperature profile compared to the is higher at the top and decreases towards the bottom, as shown in
conventional flow, see Fig. 13(a). The influence of flow reversal is Fig. 14(a). This is due to the cold fluid enters at the top at initial
significant on this surface as the temperatures are observed to change temperature and heats up absorbing the heat from the cell during its
from 313 K to 316 K, where the temperature with conventional flow on travel towards the bottom of the cell. Thus, the heat absorbing capacity
the same surface are observed to vary from 318 K to 314.5 K, see reduces as the cooling fluid moves down the cooling channel. Fur-
Fig. 13(a). Furthermore, distribution of temperature on the surface at y thermore, as one side of cell 1 is exposed to ambient air, which has less
= 20 mm, along the z direction considering designs 1 and 2 and flow heat absorbing capacity compared to the side exposed to cooling fluid,
reversals 1 and 2 are plotted in Fig. 13(b). A uniform temperature with more heat will be absorbed by the coolant from the cell 1 than from cell
flow reversal is further confirmed in Fig. 13(b), where slightly lower 2. As a result, the front surface of cell 2 ( y = 14 mm) will be at relatively

12
S.D.V.S.S.V. Siruvuri and P.R. Budarapu Journal of Energy Storage 29 (2020) 101377

higher temperatures compared to back surface ( y = 26 mm). The flow 13.8 % more surface area compared to design 1. Based on the present
reversal in design 1 is more effective in maintaining lower temperatures studies, design with flow reversal is recommended for an efficient,
compared to conventional flow, see Fig. 14(b). Since the front surface uniform heat transfer and temperature distribution, and hence, an ef-
( y = 14 mm) of the cell is exposed to conventional flow, and the back ficient operation of the battery with long life.
surface ( y = 26 mm) of the cell 2, is exposed to reversed flow direction
the temperature at the top and bottom of the cell are reversed, see Declaration of Competing Interest
Fig. 14(b). However, the temperatures on the back surface ( y = 26 mm)
of the cell 2 are at lower temperatures compared to front surface The authors declare that they have no known competing financial
( y = 14 mm). Also, the peak temperatures in FR1 are lower by 1.5 K interests or personal relationships that could have appeared to influ-
than that of conventional flow. ence the work reported in this paper.
The thermal behaviour of battery pack at different C-rates is studied
by considering design 1, with flow rate equal to 0.25 × 10 6 m3/s. A Acknowledgements
comparison of peak temperatures at various C-rates with 0.25 × 10 6
m3/s flow rate and without cooling is given in Fig. 16. Figure indicates PRB is grateful to the School of Mechanical Sciences, Indian
that the peak temperature in battery pack with and without cooling is Institute of Technology Bhubaneswar for providing the computational
increasing with C-rate. Furthermore, the cooling effectiveness is evident resources to carry out this work.
in Fig. 16, where a maximum and minimum peak temperature differ-
ence of 17 K and 0.13 K at 5C and 2C-rates, respectively, are noticed. References
Also, an observation of Fig. 15 indicates that the temperature gradient
along the cell height is increasing with C-rate. Therefore, the considered [1] A. Greco, X. Jiang, D. Cao, An investigation of lithium-ion battery thermal man-
flow rate is capable of maintaining a uniform temperature distribution agement using paraffin/porous-graphite-matrix composite, J. Power Sources 278
throughout the cell only at lower C-rates, i.e., less than 5C. (2015) 50–68.
[2] M. Al-Zareer, I. Dincer, M.A. Rosen, Novel thermal management system using
In the present design of the battery pack, cell 1 is exposed to en- boiling cooling for high-powered lithium-ion battery packs for hybrid electric ve-
vironment and hence always maintains higher temperatures due to hicles, J. Power Sources 363 (2017) 291–303.
inefficient cooling. Therefore, a new design by adding an extra cooling [3] H. Fathabadi, High thermal performance lithium-ion battery pack including hybrid
active–passive thermal management system for using in hybrid/electric vehicles,
plate on to the front face of cell 1 is developed and named as ‘design3’. Energy 70 (2014) 529–538.
Simulations are performed with design 3 by considering 0.25 × 10 6 [4] J. Zhao, Z. Rao, Y. Li, Thermal performance of mini-channel liquid cooled cylinder
m3/s flow rate and thus the obtained temperature profile of the pack is based battery thermal management for cylindrical lithium-ion power battery,
Energy Convers. Manage. 103 (2015) 157–165.
compared with that of design 1 in Fig. 17. Addition of extra cooling [5] S.S. Madani, M.J. Swierczynski, S.K. Kr, A review of thermal management and
panel has definitely improved the cooling performance of battery pack, safety for lithium ion batteries, 2017 Twelfth International Conference on
where a drop of 0.6 K in the peak temperature and better heat dis- Ecological Vehicles and Renewable Energies (EVER), (2017), pp. 1–20, https://doi.
org/10.1109/EVER.2017.7935914.
tribution is noticed in cells 1 and 2, see Figs. 17 and 18. A comparison
[6] P.C. Mahamud R, Reciprocating air flow for li-ion battery thermal management to
of temperature distribution in cells 1 and 2, in designs 1,2 and 3 is improve temperature uniformity, Power Sources 196 (2011) 5685–5696.
shown in Fig. 18, where design 2 is observed to possess the lowest peak [7] A. Greco, D. Cao, X. Jiang, H. Yang, A theoretical and computational study of li-
temperature, followed by designs 3 and 1. thium-ion battery thermal management for electric vehicles using heat pipes, J.
Power Sources 257 (2014) 344–355.
Furthermore, identification of the key parameters studied in [8] L. Lu, X. Han, J. Li, J. Hua, M. Ouyang, A review on the key issues for lithium-ion
Sections 3.1, 3.2 and 3.3, through a sensitivity analysis can help to battery management in electric vehicles, J. Power Sources 226 (2013) 272–288.
optimise the performance of Lithium ion batteries. For instance, a [9] Y. Li, Z. Zhou, W.-T. Wu, Three-dimensional thermal modeling of li-ion battery cell
and 50 v li-ion battery pack cooled by mini-channel cold plate, Appl. Therm. Eng.
sensitivity based on various common methods has been applied in [31] 147 (2019) 829–840.
to identify the input parameters influencing the energy conversion [10] K.A. Severson, P.M. Attia, N. Jin, N. Perkins, B. Jiang, Z. Yang, M.H. Chen,
factor of flexoelectric materials. M. Aykol, P.K. Herring, D. Fraggedakis, et al., Data-driven prediction of battery
cycle life before capacity degradation, Nat. Energy (2019) 1.
[11] A. Greco, X. Jiang, D. Cao, An investigation of lithium-ion battery thermal man-
4. Conclusions agement using paraffin/porous-graphite-matrix composite, J. Power Sources 278
(2015) 50–68.
[12] D. Davies, M. Verde, O. Mnyshenko, Y. Chen, R. Rajeev, Y. Meng, G. Elliott,
Battery thermal management system considering a liquid coolant
Combined economic and technological evaluation of battery energy storage for grid
has been proposed here to investigate the thermal behaviour of battery applications, Nat. Energy 4 (1) (2019) 42.
pack. A battery pack consisting of five rectangular cells of size [13] A. Nazari, S. Farhad, Heat generation in lithium-ion batteries with different nominal
capacities and chemistries, Appl. Therm. Eng. 125 (2017) 1501–1517.
80 mm × 120 mm × 12 mm is considered and the liquid coolant is
[14] M.R. Palacín, A. de Guibert, Why do batteries fail? Science 351 (6273) (2016)
passed through a 1 mm × 5 mm circuitous rectangular channel, cut out 1253292.
in a 80 mm × 120 mm × 2 mm aluminium sheet, sandwiched between [15] K. Liu, Y. Liu, D. Lin, A. Pei, Y. Cui, Materials for lithium-ion battery safety, Sci.
the cells. A 5C charge/discharge rate per battery with a cycle time of Adv. 4 (6) (2018) eaas9820.
[16] Y. Liu, S. Yang, B. Guo, C. Deng, Numerical analysis and design of thermal man-
62.234 s is considered for the analysis. Two such battery packs of five agement system for lithium ion battery pack using thermoelectric coolers, Adv.
cells each are considered while modelling. The influence of volume flow Mech. Eng. 6 (2014) 852712.
rate, flow direction, and contact surface area on the battery thermal [17] Z. Rao, Y. Huo, X. Liu, Experimental study of an OHP-cooled thermal management
system for electric vehicle power battery, Exp. Therm. Fluid Sci. 57 (2014) 20–26.
behaviour are analysed. [18] L. Martin-Martin, J. Gastelurrutia, G.S. Larraona, R. Anton, L. del Portillo-Valdes,
Results indicate that the liquid coolant is efficient in reducing the I. Gil, Optimization of thermal management systems for vertical elevation appli-
peak temperatures from 336.623 K to 319.31 K and 316.93 K at flow cations powered by lithium-ion batteries, Appl. Therm. Eng. 147 (2019) 155–166.
[19] J. Kim, J. Oh, H. Lee, Review on battery thermal management system for electric
rates of 0.25 × 10 6 m3/s and 1.6 × 10 6 m3/s, respectively, which is vehicles, Appl. Therm. Eng. (2018).
5.85 % reduction in peak temperature by circulating the coolant at [20] J. Deng, C. Bae, J. Marcicki, A. Masias, T. Miller, Safety modelling and testing of
1.6 × 10 6 m3/s. Furthermore, by reversing the flow direction of the lithium-ion batteries in electrified vehicles, Nat. Energy 3 (4) (2018) 261.
[21] W. Wu, S. Wang, W. Wu, K. Chen, S. Hong, Y. Lai, A critical review of battery
coolant in one of the channels leads to the redistribution of heat and thermal performance and liquid based battery thermal management, Energy
hence the temperature distribution in the battery pack is observed to be Convers. Manage. 182 (2019) 262–281.
more uniform. Flow reversal also helped to reduce the peak tempera- [22] H. Jouhara, N. Khordehgah, N. Serey, S. Almahmoud, S.P. Lester, D. Machen,
L. Wrobel, Applications and thermal management of rechargeable batteries for in-
tures upto 2.77 K in the battery pack. The developed methodology is
dustrial applications, Energy 170 (2019) 849–861.
tested for various cases of normal and reversed flows of the cooling [23] G. Qian, X. Liao, Y. Zhu, F. Pan, X. Chen, Y. Yang, Designing flexible lithium-ion
fluid in two different designs, where the second design consists of batteries by structural engineering, ACS Energy Lett. (2019).

13
S.D.V.S.S.V. Siruvuri and P.R. Budarapu Journal of Energy Storage 29 (2020) 101377

[24] J. Kim, J. Oh, H. Lee, Review on battery thermal management system for electric [36] D. Chen, J. Jiang, G.-H. Kim, C. Yang, A. Pesaran, Comparison of different cooling
vehicles, Appl. Therm. Eng. 149 (2019) 192–212. methods for lithium ion battery cells, Appl. Therm. Eng. 94 (2016) 846–854.
[25] S. Yang, C. Ling, Y. Fan, Y. Yang, X. Tan, H. Dong, A review of lithium-ion battery [37] L.H. Saw, H.M. Poon, H. San Thiam, Z. Cai, W.T. Chong, N.A. Pambudi, Y.J. King,
thermal management system strategies and the evaluate criteria, Int. J. Novel thermal management system using mist cooling for lithium-ion battery packs,
Electrochem. Sci. 14 (7) (2019) 6077–6107. Appl. Energy 223 (2018) 146–158.
[26] Z. An, L. Jia, Y. Ding, C. Dang, X. Li, A review on lithium-ion power battery thermal [38] C. Menale, F. D’Annibale, B. Mazzarotta, R. Bubbico, Thermal management of li-
management technologies and thermal safety, J. Therm. Sci. 26 (2017) 391–412. thium-ion batteries: an experimental investigation, Energy 182 (2019) 57–71.
[27] M. Parhizi, A. Jain, Analytical modeling and optimization of phase change thermal [39] C. Lan, J. Xu, Y. Qiao, Y. Ma, Thermal management for high power lithium-ion
management of a li-ion battery pack, Appl. Therm. Eng. 148 (2019) 229–237. battery by minichannel aluminum tubes, Appl. Therm. Eng. 101 (2016) 284–292.
[28] F.G. Hadia, Y.T. Zhang, Thermal management simulation of lithium ion batteries for [40] Z. An, L. Jia, X. Li, Y. Ding, Experimental investigation on lithium-ion battery
EV/HEV, Applied Mechanics and Materials, vol. 457, (2014), pp. 350–353. thermal management based on flow boiling in mini-channel, Appl. Therm. Eng. 117
[29] B. Mortazavi, H. Yang, F. Mohebbi, G. Cuniberti, T. Rabczuk, Graphene or h-BN (2017) 534–543.
paraffin composite structures for the thermal management of li-ion batteries: a [41] T. Deng, G. Zhang, Y. Ran, Study on thermal management of rectangular li-ion
multiscale investigation, Appl. Energy 202 (2017) 323–334. battery with serpentine-channel cold plate, Int. J. Heat Mass Transf. 125 (2018)
[30] Q. Cheng, Porous graphene sponge additives for lithium ion batteries with excellent 143–152.
rate capability, Sci. Rep. 7 (1) (2017) 925. [42] C. Zhao, W. Cao, T. Dong, F. Jiang, Thermal behavior study of discharging/charging
[31] C.C. Kwasi-Effah, T. Rabczuk, Dimensional analysis and modelling of energy density cylindrical lithium-ion battery module cooled by channeled liquid flow, Int. J. Heat
of lithium-ion battery, J. Energy Storage 18 (2018) 308–315. Mass Transf. 120 (2018) 751–762.
[32] W. Li, M. Xiao, X. Peng, A. Garg, L. Gao, A surrogate thermal modeling and para- [43] J.P. Singer, K.P. Birke, Kinetic study of low temperature capacity fading in li-ion
metric optimization of battery pack with air cooling for EVs, Appl. Therm. Eng. 147 cells, J. Energy Storage 13 (2017) 129–136.
(2019) 90–100. [44] B. Mondal, C.F. Lopez, A. Verma, P.P. Mukherjee, Vortex generators for active
[33] K. Chen, W. Wu, F. Yuan, L. Chen, S. Wang, Cooling efficiency improvement of air- thermal management in lithium-ion battery systems, Int. J. Heat Mass Transf. 124
cooled battery thermal management system through designing the flow pattern, (2018) 800–815.
Energy 167 (2019) 781–790. [45] J. Xie, Y. Xie, C. Yuan, Numerical study of heat transfer enhancement using vortex
[34] K. Chen, M. Song, W. Wei, S. Wang, Design of the structure of battery pack in generator for thermal management of lithium ion battery, Int. J. Heat Mass Transf.
parallel air-cooled battery thermal management system for cooling efficiency im- 129 (2019) 1184–1193.
provement, Int. J. Heat Mass Transf. 132 (2019) 309–321. [46] P. Zhang, T. Cui, Q. Li, Effect of surface roughness on thermal contact resistance of
[35] H. Zhou, F. Zhou, L. Xu, J. Kong, et al., Thermal performance of cylindrical lithium- aluminium alloy, Appl. Therm. Eng. 121 (2017) 992–998.
ion battery thermal management system based on air distribution pipe, Int. J. Heat [47] Y. Huo, Z. Rao, X. Liu, J. Zhao, Investigation of power battery thermal management
Mass Transf. 131 (2019) 984–998. by using mini-channel cold plate, Energy Convers. Manage. 89 (2015) 387–395.

14