Journal of Energy Storage 28 (2020) 101235

Contents lists available at ScienceDirect

Journal of Energy Storage

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

Battery thermal management with thermal energy storage composites of T

PCM, metal foam, fin and nanoparticle
Mohammad Mahdi Heyhata, Sepehr Mousavib, Majid Siavashib,
⁎

a
Faculty of Mechanical Engineering, Tarbiat Modares University, Tehran, Iran
b
Applied Multi-Phase Fluid Dynamics Lab., School of Mechanical Engineering, Iran University of Science and Technology, Heydarkhani St., Narmak, Tehran, Iran

A RT ICLE INFO ABSTRACT

Keywords: This study aims to analyze the thermal performance of the passive thermal management system (TMS) of the
Phase change material (PCM) 18,650 lithium-ion battery with application of phase change materials (PCM). To improve performance of TMS,
Battery thermal management system nanoparticles, fins and porous metal foam are used beside the PCM, and their effects on the system performance
Metal foam are compared. The local thermal non-equilibrium (LTNE) model and non-Darcy law are considered to simulate
Nanoparticle
the nano-PCM melting inside the porous media. Numerical results are validated through previously published
Fin
Heat transfer enhancement
experimental data and results are presented for two, 4.6 W and 9.2 W, heat generation rates. Sole effects of
adding nanoparticles to the PCM, utilizing different numbers of fins, and application of the metal foam on the
system performance are scrutinized. Results indicated that the porous-PCM composition performs more effi-
ciently than the nano-PCM and the fin-PCM ones. In addition, ΔTavg, battery parameter is introduced and its
variations are analyzed to judge about the effect of each technique to reduce the battery mean temperature.
Using the porous-PCM led to 4–6 K reduction in the battery mean temperature with respect to the pure PCM.
Moreover, for the porous-PCM composition a delay is observed in the PCM melting initiation time that can
adversely affect the performance of battery TMS.

1. Introduction efficiency, capacity, power, safety and life cycle of the lithium-ion
batteries [1]. Also, reaching to temperatures lower than the battery
Increased environmental pollution and global warming issue have operational range will cause to lose capacity and autonomy of the
forced researchers to find a clean and suitable alternative for fossil battery as well as some other problems during its setting up [2].
fuels. Vehicles are one of the main consumers of fossil fuels. The pure Battery TMSs keep the battery temperature uniform and in a safe
electric vehicles (EVs) and hybrid electric vehicles (HEVs) with good range. Various methods are used in the battery TMSs, including the air
features such as fast drive starts, zero CO2 emissions, and low noise cooling [3,4], liquid cooling [5,6,10,11], and PCM cooling [7–9,12–14]
pollution are most suitable alternatives for the conventional vehicles. systems. Using PCM materials in the battery TMS was first proposed by
However, EVs and HEVs still have a long way to be competed with the Al-Hallaj and Selman [7]. They used the paraffin wax with 56 °C
internal combustion engine cars, in terms of their efficiency and cost. melting temperature to control the temperature of a single battery pack
Rechargeable electrochemical batteries are also used as energy sources including 8 cylindrical 18,650 Sony batteries. Their outcomes showed
in EVs and HEVs. Different types of electro-chemical batteries are that using PCM led to a much more uniform temperature profile com-
available, although lithium-ion batteries with high capacity, power, pared to the methods such as the air TMS. Another advantage of that
energy density, and lifecycle as well as low self-discharge rates are method was the low manufacturing and maintenance costs, simplicity,
widely used. The overall performance of EVs and HEVs mostly depends and the ability to control the temperature fluctuations in the battery.
on the performance of their batteries so that by increasing their effi- Sabbah et al. [10] compared the performance of both air and PCM
ciency, the overall efficiency increases significantly. in TMSs of lithium-ion batteries. The results demonstrated that the
This type of batteries generates a large amount of heat, especially performance of the air-cooling management system highly depends on
during the fast discharge process. Therefore, the absence of an appro- the ambient temperature and the discharge rate of the battery, which
priate thermal management system (TMS) will increase the battery could not have the ability to keep the battery temperature in the safe
temperature with subsequent devastating effects on its performance, range at a moderate discharge rate and high ambient temperature.

⁎
Corresponding author.
E-mail addresses: [email protected] (M.M. Heyhat), [email protected] (S. Mousavi), [email protected] (M. Siavashi).

https://doi.org/10.1016/j.est.2020.101235
Received 12 September 2019; Received in revised form 30 December 2019; Accepted 22 January 2020
Available online 14 February 2020
2352-152X/ © 2020 Elsevier Ltd. All rights reserved.
M.M. Heyhat, et al. Journal of Energy Storage 28 (2020) 101235

Nomenclature Tfr freezing temperature, K

ur r-direction velocity, ms 1

Amush mushy zone constant ut Total velocity, ms 1

asf specific surface area, m 1 uz z-direction velocity, ms 1

Cf inertia coefficient of fluid flow in metal foams

Cp specific heat at constant pressure, J kg 1K 1 Greek letters
df equivalent diameter of metal fiber, m
dp equivalent pore size, m αf thermal diffusivity, m2s 1
dnp nanoparticles diameter, m β thermal expansion coefficient, K 1
dpcm equivalent diameter of a PCM molecule, m ΔH latent enthalpy, kJ kg 1
fl liquid fraction Δt time-step, s
g gravitational acceleration, ms 2 ΔTavg, battery average PCM temperature difference, K
H total enthalpy, kJ kg 1 Δr mesh size in r-direction, mm
h sensible enthalpy, kJ kg 1 Δz mesh size in z-direction, mm
hsl latent heat of fusion, kJ kg 1 ɛ porosity
hsf heat transfer coefficient between metal foam & PCM, μ Dynamic viscosity, Pa.s
W m 2K 1 ν Kinematic viscosity, m2 s−1
Keff Effective thermal conductivity, W m 1K 1 φ Nanoparticles volume fraction
Kb Boltzmann constant ψ Stream function, kg s−1
K Permeability, m2 ψ* , m
M Molecular weight, g. mol 1
N Avogadro number Subscript
nr mesh numbers in r-direction
nz mesh numbers in z-direction eff Effective value
Pr Prandtl number f Fluid (PCM)
Ra Rayleigh number fe Effective value for PCM
Re Reynolds number s Solid(metal foam)
T temperature, K se Effective value for metal foam

Moreover, the system required a very high airflow in turbulence range concluded that fins have the higher impact. Also, simultaneous appli-
of fan, while both the discharge rate and the ambient temperature were cation of fins and nanoparticles could lead to the greatest performance
high. The fact that is not practically possible in electric vehicles. On the with the lowest melting time. Sushobhan and Kar [17] observed that
other hand, PCM thermal management system could manage to keep applying CuO nanoparticles with 5% and 2.5% volume fractions could
the battery temperature even below the safe range in high discharge improve the PCM performance as well as its melting rate.
rates, while the ambient temperature was 52 °C and also managed to Wang et al. [21] scrutinized the effect of using fins on improving the
keep the battery temperature more uniformly. battery TMS performance, experimentally. Paraffin with 44 °C melting
The use of PCM in battery TMSs has some disadvantages, including temperature was used in their system. Adding 8 fins with less than 3%
the limited heat absorption capacity, highly increase in the battery volume fraction of nanoparticles in the PCM enclosure leads to 8 °C
temperature after the fully melting of PCM, and requiring a re-pre- reduction in the average battery temperature. The surface temperature
paration process (PCM solidification). Also, the use of PCM in cold profile became more uniform and battery thermal resistance reduced by
weather lead to reduction of the rate of battery warming up in the in- 55%.
itial driving time that can damage the battery [11]. In order to elim- Yang et al. [22–24] in different numerical and experimental works
inate these unfavorable facts, researchers have suggested the use of the proved that the open-cell metal foams can significantly improve the
combination of passive and active methods such as using PCM and heat transfer rate and temperature uniformity during the PCM melting.
airflow or PCM and liquids, simultaneously [12]. In hybrid TMSs, the Li et al. [26] reported that using metal foams in the passive TMS of a
use of PCMs significantly reduces the need for an active TMS. In most lithium-ion battery reduces the battery temperature by 3 to 10 °C
cases, at low discharge rates or moderate ambient temperatures, compared to the pure PCM, depending on the discharge rate and por-
probably there is no need for active cooling systems and when the osity of the foam.
ambient temperature or the discharge rate is high, use of PCM will The effect of each above-mentioned methods, i.e. utilizing metal
delay the increase in battery temperature and activation of the active foam, fins, and nanoparticles, on improving PCM performance depends
cooling system, which will result in a significant cost saving [13]. on the geometry, flow characteristics and boundary conditions of the
PCM materials have an inherent defect which is their low thermal studied problems. Also, these methods are different from each other
conductivity coefficient. For most PCM materials this coefficient is based on their costs and capacities for PCM latent heat storage. To the
around the small amount of 0.2, which leads to some major problems authors’ knowledge, the performance of these three methods have not
arisen by utilizing PCMs to absorb the heat generated by the battery. On been compared for thermal management of lithium-ion batteries.
the other hand, after the complete melting of PCM, the low PCM Accordingly, this study aims to investigate the effect of these methods
thermal conductivity makes it an insulator which inhibits the heat on the performance of the passive TMS of the 18,650 lithium-ion bat-
transfer and leads to a rapid increase in the temperature. For this tery in the high heat generation rate condition. Furthermore, in the
reason, the PCM conductivity coefficient must be increased so that the most of the previous studies on PCM melting in porous media, it is
PCM can properly transfer the generated heat. Various methods are assumed that the porous material and PCM are in local thermal equi-
available to enhance the heat transfer rate in PCMs such as using na- librium (LTE), while this assumption could lead to high deviations from
noparticles [14–17], fins [18–21], and metal foams [22–26]. Mousavi the real results. The local thermal non-equilibrium model (LTNE) could
et al. [20] compared the effect of nanoparticles and fins on the melting provide more accurate results with respect to the LTE model, hence, in
time of a vertical cylindrical thermal energy storage system and this study LTNE is implemented for numerical simulation of the PCM

2
M.M. Heyhat, et al. Journal of Energy Storage 28 (2020) 101235

melting inside the porous medium. Table 1

Thermophysical properties of n-eicosane PCM (Tref = 323.15 K ) [28].
2. Problem description and boundary conditions Solid (at Liquid (at 323.15 K)
298.15 K)
The performance of PCM in the battery TMS was studied.
Density, ρ (kg/m3) 910 769
Furthermore, the effects of using nanoparticles, fins and metal foams on
Latent heat of fusion, hsl (kJ⁄kg) 248 –
PCM performance were investigated and compared with each other. Melting point, Tmelt (K) 309.55 –
The pure physical geometry of the problem, i.e. without foam and fins, Molecular weight, M (g⁄mol) 282.54 282.54
is shown in Fig. 1. The 18,650 lithium-ion battery is located in the Specific heat capacity, cp (J⁄(kg.K)) 1926 2400
central axis of the aluminum enclosure. The inner diameter, the wall Thermal conductivity, k (W/(m.K)) 0.423 0.146
Thermal expansion coefficient, β – 8.161×10-4
thickness and the height of the enclosure are 31 mm, 5 mm and 70 mm, (1⁄K)
respectively. The diameter and the height of the battery are 18 mm and
65 mm, respectively. The battery is considered to be made of solid
aluminum with constant properties. Table 2
The PCM used is n-eicosan with a melting temperature of Thermophysical properties of copper (foam) and aluminum (wall materials).
309.55–309.65 K. The thermophysical properties of PCM and alu-
Density, Specific heat capacity, cp Thermal conductivity, k
minum are presented in Table 1 and Table 2, respectively. The ambient ρ (kg⁄m3) (J⁄(kg.K)) (W⁄(m.K))
and initial temperatures of TMS and the battery are assumed to be
298.15 K. So the PCM is somewhat sub-cooled. The enclosure is in- Cu 8900 386 387.6
sulated from the lower part, while the upper part of the PCM and the Al 2700 900 238

outer wall of the enclosure are considered to have constant heat transfer
coefficient, h, equal to 10 W/m2K. The geometry is axisymmetric.
µ = (9 × 10 4T 2 0.6529 T + 119.94) × 10 3
(1)
Therefore, a 2D model is used in the simulation. The heat generation
rate for the 18,650 Lithium-Ion batteries, considering 5C discharge rate There is a discontinuity for thermo-physical properties in the mushy
and 10C discharge rate, are 4.6 W and 9.2 W, respectively [27]. zone. Therefore, a linear change of properties according to temperature
The material of the fins, nanoparticles and metal foam are assumed is considered in this zone to avoid any trouble caused by this dis-
to be copper. The thermophysical properties of copper are provided in continuity in numerical solutions. Boussinesq's approximation is used to
Table 2. Effect of using 3%, 5% and 7% volume fraction nanoparticles consider buoyancy forces with a reference temperature of 323.15 K.
on PCM performance was studied. In order to study the effect of using
fins, three models with 1, 3 and 5 fins were considered. The physical
geometry of these cases are shown in Fig. 2. Moreover, the effect of 3. Governing equations
using metal foam was studied in 3 models with 85%, 90% and 95%
porosity. The schematics of battery TMS units with nanoparticles, fins To model the solid-liquid phase changing the enthalpy-porosity
and metal foam are shown in Fig. 3. The melting process of PCM is method was applied. Therefore, the entire computational region is
simulated by the enthalpy-porosity method [29,30]. The flow is con- considered to be porous so that for the liquid and solid phases the
sidered to be laminar and incompressible for the melted PCM. The porosity is 1 and 0, respectively. The porosity within the mushy zone
movement of the solid parts of PCM is disregarded because of negligible equals the liquid fraction. The continuity, momentum, and energy
volume change during the melting process. Moreover, the impacts of equations in 2D axisymmetric cylindrical coordinate for region filled by
viscous losses and contact resistance were ignored [28]. PCM and metal foam are as follows. It should be noted that for simu-
All PCM properties are considered to be invariant except for the lating nano-PCM, simply the PCM properties are replaced with nano-
dynamic viscosity which is a function of temperature and obtained from PCM's.
Eq. (1) [28]: The continuity equation is given by [31]:

Fig. 1. Schematic of battery TMS unit, (a) battery and its enclosure, and (b) geometrical dimensions and boundary conditions.

3
M.M. Heyhat, et al. Journal of Energy Storage 28 (2020) 101235

Fig. 2. Schematics of battery TMS units with fins (a) 1 fin, (b) 2 fins, and (c) 3 fins.

1 (rur ) uz where K is metal foam permeability, µ is the viscosity of PCM and CF is

+ =0
r r z (2) drag or inertial factor. Sr and Sz in the above equations are functions of
porosity which are defined by:
Where velocity components in r and z directions are ur and uz, re-
spectively. The momentum equations are expressed based on the (1 fl ) 2
Sr = Amush (ur up )
Brinkman-Forchheimer extended Darcy model in the r and z directions fl 3 + (5)
as follows [31]:
(1 fl )2
Sz = Amush (uz wp)
( ) ( ( ) )
ur ur ur P 1 ur ur 2u
t
+ ur r
+ uz z
= r
+µ r r
r r r2
+
z2
r
fl 3 + (6)

µ CF The following equation is applied to compute the liquid fraction:

u
K r K
ur ur + Sr (3)
0 if T Ts

( (r ) + ) +
1 if Tl T
( )= fl =
uz uz uz P 1 uz 2u
z
+ ur + uz +µ g (T Tm)
t r z z r r r z2 T Tl
if Tl T Ts
T l Ts (7)
µ CF
u
K z
uz uz + Sz
K In Eqs. (5) and (6), up and wp are pull velocities where for reducing
(4) the computational costs they were ignored. To avoid zero in the

Fig. 3. Schematics of battery TMS units with (a) nanoparticle, (b) fins and (c) metal foam.

4
M.M. Heyhat, et al. Journal of Energy Storage 28 (2020) 101235

denominator, ε as a very small parameter (equals to 0.001) is used. The (e 2 )2

RB =
mushy zone constant, Amush, determines the damping properties. The (e 2 ) e 2ks + (2e 4 (e 2 ) e 2) kf (15)
liquid stays motionless in a smaller time for the greater value of Amush in
solidification process. Thus, the mushy zone constant, has an important ( 2 2e )2
RC =
role on the melting process, takes different values for various materials 2 2 (1 2e 2 ) ks + 2( 2 2e 2 (1 2e 2 )) kf (16)
in different conditions. Voller and Prakash [30] examined the role of
Amush values of 1.6 × 103 and 1.6 × 105 on the PCM performance. The RD =
2e
flow in the two-phase region was lessened by increasing of Amush. e 2ks + (4 e 2) k f (17)
Shmueli et al. [32] studied the influences of Amush amounts (105, 107,
108, 109 and 1010) on melting process of a PCM in a vertical cylinder. 2 (2 (5/8) e 3 2 2 )
=
They found that the values of 105 and 108 for Amush led to the smallest (3 4e 2 e) (18)
and longest melting time, respectively. Therefore, in order to determine
a proper value for Amush, the experimental and numerical results should e = 0.339 is the experimental constant in the Boomsma and Poulikakos
be compared. In the current study, several numerical simulations were model.
carried out for various values of Amush and obtained results were com- The fluid mixing in porous media at pore scale can result in heat
pared with experimental data of [33]. Comparison of the outcomes transfer enhancement. In order to take the effects of this phenomenon,
revealed that the amount of 50,000 for Amush could lead to acceptable thermal dispersion model developed by Georgadis and Catton [35] is
results. employed, which can be expressed by the following equation:
There are two common viewpoints of thermal equilibrium and 0.36 f Cpf df v
thermal non-equilibrium in the energy analysis in the combination of Ktd =
(1 ) (19)
metal foam and PCM. In this study, two separate non-equilibrium
equations were used for metal foam and PCM. The energy equation in Then the effective thermal conductivities of the PCM and the metal
the cylindrical coordinate system for metal foam is [31]: foam can be described as
Tm 1 T
r m +
2T
m ks = 0, kf = kpcm kpcm, eff = ke + ktd (20)
m Cp, m (1 ) = Km, eff hsf asf [Tm Tpcm]
t r r r z2
kf = 0, ks = km km, eff = ke (21)
(8)
Tm is the metal foam temperature and Km, eff is the effective thermal The empirical models provided by Calmadi and Mahajan were used
conductivity of metal foam. hsf and asf are the heat transfer coefficient for computing the permeability, the inertial factor and specific surface
and the specific surface area of the metal foam, respectively. area as the following equations [31]:
The energy equation for PCM is [31]: K = 0.00073(1 ) 0.224d
f
1.11d 3.11
p (22)

(u )=k (r ) +
Hpcm Tpcm Tpcm Tpcm 2T
1 pcm
+ pcm Cp, pcm r + uz pcm, eff df
t r z r r r z2 cf = 0.0012(1 ) 0.132 ( ) 1.63
dp (23)
+ hsf asf [Tm Tpcm]
(9) 3 df
asf =
Tpcm is the PCM temperature and kpcm, eff is the effective thermal con-
dp 2 (24)
ductivity. In Eq. (9), the first term in the left side Hpcmis the PCM total The correlations which proposed by Churchill and Chu [36] are
enthalpy which is calculated as follows. employed to calculate the heat transfer coefficient between the metal
foam and the PCM.
Hpcm = h + H (10)

where h is the sensible enthalpy and is given by:

1
T
kf 0.518Rad 4
hsf = 0.36 + 4
h = href + cp dT df 9

Tref (11) 1+ ( ) 0.559 16 9

Pr
(25)
Moreover, ΔH in Eq. (10) is the latent enthalpy which changes from
g Tdf 3
zero (in the solid form) to hsl (in the liquid form). It can be computed as: Rad =
f f (26)
H = fl hsl (12)

In order to solve the above equations, the parameters related to

4. Nano-PCM properties
metal foam must be determined such as permeability, pore size, effec-
tive thermal conductivity, inertial factor, and specific surface area.
For all cases of this study, nano-PCM is assumed as a Newtonian
Calculating these parameters is highly complicated and is dependent to
fluid [25]. Table 2 lists the thermophysical properties of Cu nano-
inner micro structures of metal foam .Various experimental and semi-
particles. The density and specific heat capacity of nano-PCM were
experimental relationships have been introduced. In the present work,
calculated by the subsequent relations [37]:
3D model of Boomsma and Poulikakos [34] was used for computing the
effective thermal conductivity. This model is validated by experimental nPCM = np + (1 ) PCM (27)
results with porosity between 0.88 to 98%.
np (Cp )np + (1 ) PCM (Cp) PCM
2 (Cp)nPCM =
Ke = (28)
2(RA + RB + RC + RD ) (13) nPCM

where subscripts nPCM and np stands for nano-PCM and nanoparticle

4
RA = properties, respectively, while ϕ is the volume fraction of nanoparticles.
(2e 2 + (1 e )) ks + (4 + 2e 2 + (1 e )) kf (14) Latent heat of nano-PCM (hsl(nPCM)) is evaluated by [37]:

5
M.M. Heyhat, et al. Journal of Energy Storage 28 (2020) 101235

(1 )( hsl ) PCM experiment conducted by Jones et al. [33] the melting process of pure
hsl (nPCM ) =
nPCM (29) n-eicosan in a vertical cylinder was investigated. Fig. 5 illustrates the
obtained liquid fraction of the present work and numerical results of
The melting process of PCM is greatly affected by the viscosity and
Muhammad et al. [28] and experimental data of Jones et al. [33]. In
thermal conductivity. Therefore, one must be careful about assigning a
another experiment conducted by Tian and Zhao [31], the effect of
model for these two properties. One of the earliest model for thermal
using metal foam on the melting process of paraffin wax was studied. A
conductivity is the Maxwell model [38] which has been used in various
composite of copper metal foam with 95% porosity and
applications [39]. Thus, to model the thermal conductivity of solid
200 × 120 × 25 mm extent and RT58 paraffin with 48–52 °C melting
nano-PCM the Maxwell model is applied:
point were put under 1.6 kW/m2 constant heat flux. Temperature
knPCM knp + 2kPCM 2 (kPCM knp) variations were measured using thermocouples in different points.
=
kPCM knp + 2kPCM + (kPCM knp) (30) Fig. 6 shows a comparison of temperature variations in two points
y = 0 and y = 8 mm in both numerical-experimental study of Tian and
There are some models which include the nanoparticle shape and Zhao [31] and current study. It is clear that there is a good agreement
size. One of them is Corcione's model [40] which is used to calculate the between the previously reported data and outcomes of the present
thermal conductivity of liquid nano-PCM: work.
10 0.03
knPCM T knp
= 1 + 4.4 Re 0.4Pr 0.66 0.66
kPCM Tfr kPCM (31) 6. Results and discussion

where Pr, T, Tfr, and knp are Prandtl number, the temperature of PCM, 6.1. The effect of nanoparticles on PCM performance improvement
the freezing temperature of PCM, and nanoparticle thermal con-
ductivity. Also, Re is nanoparticles Reynolds number, which is defined Using nanoparticles is one of the ways to improve heat transfer. In
as: current study, Cu nanoparticles were used. The presented relationships
2 PCM kb T
in Section 4 were used in order to calculate the nano-PCM properties.
Re = The changes in liquid mass fraction of PCM with respect to time for
µPCM 2dnp (32)
4.6 W and 9.2 W heat generation rates are presented in Fig. 7. It is
µPCM is the viscosity of PCM, dnp is diameter of nanoparticles equals obvious that adding Cu nanoparticles to PCM and the increase of its
20 nm and kb is Boltzmann constant, kb = 1.38 × 10 23 . volume fraction lead to slight improvement in heat transfer as well as
Finally, Corcione's model [40] is used to assess the viscosity of nano- melting rate. Adding Cu nanoparticles by 7% volume fraction results in
PCM as follows: decreasing the melting time by 5% at the best case.
µnPCM 1 It must be considered that the purpose of improving PCM perfor-
= mance in TMS is to reduce the battery mean temperature. The changes
µPCM 1 34.87(dnp/ dPCM ) 0.3 1.03
(33)
in battery mean temperature with respect to time for pure PCM and
where: nano-PCM are shown in Fig. 8. The trend of battery mean temperature
1 (the volume-weighted average of temperature in battery domain)
6M 3
change shows that the addition of nanoparticles does have a positive
dPCM = 0.1
N PCM0 (34) impact, though small, on heat transfer improvement and battery mean
temperature reduction. For example, in case with 9.2 W heat generation
In above equation N is Avogadro number, M is the molecular weight
rate and at 600 s, by adding 7% nanoparticles, the battery mean tem-
of pure PCM, and ρPCM0 is the density of pure PCM at T0 = 293 K.
perature was decreased just 0.3 K. But this trend is just up to complete
melting so that the pure PCM shows better performance in the liquid
5. Numerical calculation and validation
phase. This may be due to the negative impact of adding nanoparticles
on free convection heat transfer within fully liquid PCM. Also, it was
The ANSYS-Fluent commercial software is used for solving the
assumed the nanoparticles were homogeneously dispersed in the PCM,
governing equations. In order to resolve the problem a uniform struc-
but in a practical work because of agglomeration of nanoparticles, the
tured grid is created. SIMPLE algorithm is applied to link the pressure
use of nano-PCM is very challenging.
and velocity equations. The second order upwind scheme is employed
for discretizing the quantities. Moreover, PRESTO is utilized for the
pressure correction. the under-relaxation factors were used to enhance 6.2. The effect of fins on PCM performance improvement
the convergence rate. These values were set to 0.5, 0.3 and 0.9 for
momentum, pressure, and liquid fraction, respectively. The set point The changes in PCM liquid mass fraction are shown in Fig. 9 for
values for residuals of the continuity, momentum and energy equations 4.6 W and 9.2 W heat generation rates for the enclosures without fin
were 10−3, 10−5 and 10−8, respectively, in order to meet the con- and with 1, 3, and 5 fins. The results show that adding fins lead to heat
vergence condition. In every time step, the convergence criteria were transfer improvement and reduced melting time in both heat generation
met after about 100–200 iterations. Mesh and time step independencies rates. However, it seems that it has more significant effect in higher
are inspected in 3 mesh sizes and 3 time steps. The complete melting heat generation rate. The addition of 5 fins reduces the melting time by
time for different mesh sizes and time steps for the pure PCM with 18% at a heat generation rate of 9.2 W at best.
9.2 W heat generation rate are provided in Table 3. The results show
that 213 × 60(Δz≅Δr≅0.32 mm) mesh numbers and 0.05 s time step Table 3
are suitable for the current study. The numerical simulations were Melting time of pure PCM at 9.2 W heat generation rate, for different mesh sizes
implemented on a workstation cluster with 16 GB RAM and Intel Xeon and time steps.
E5-2620 v4 (20 M Cache, 2.10 GHz) CPU. The simulations took about 4 Mesh (nz × nr) Time step (s)
days time to resolve completely the melting course while using 100% of 0.1 s 0.05 s 0.01 s
the workstation power.
163 × 46 (Δz≅Δr≅0.42mm) 1450 s 1430 s 1430 s
The employed correlations and numerical method are validated
213 × 60(Δz≅Δr≅0.32mm) 1418 s 1420 s 1421 s
with previous experimental [31,33] and numerical [28] studies. Phy- 238 × 68(Δz≅Δr≅0.29mm) 1420 s 1419 s 1420 s
sical geometries of two experiments are shown in Fig. 4. In the

6
M.M. Heyhat, et al. Journal of Energy Storage 28 (2020) 101235

Fig. 4. Geometry of two experiments, (a) the experiment of Jones et al. [33], and (b) the experiment of Tian and Zhao [31].

Fig. 5. The variations of liquid fraction in the present study, experimental data
of Jones et al. [33] and numerical results of Muhammad et al. [28]. Fig. 6. Temperature variation at two points y = 0 mm and y = 8 mm in nu-
merical-experimental study of [31] and the current study.
The battery mean temperature is shown with respect to time in
Fig. 10 for unfinned and finned configurations. It is obvious that adding notably in comparison to the pure PCM. The reason of this phenomenon
fins results in considerable reduction in battery mean temperature. But is that by use of metal foam the temperature of the PCM increases more
the point is the comparison between the enclosures with 5 and 1 fin. uniformly, and as it can be inferred from Fig. 12, the equilibrium model
Although an improved heat transfer and reduced melting time is ex- was unable to predict this behavior.
pected from enclosure with 5 fins but after complete melting, better The pore density is one of the main characteristics of a porous
performance is observed for 1 fin compared to 5 fins. The contours of medium and the concentration of pores for a metal foam can be ex-
temperature for cases with fins are shown in Fig. 11. These results also plained by the number of pores per inch (PPI). 1 PPI means 1 pore exists
show that after complete melting of PCM, the case with 1 fin has the per one-inch length of the metal foam. In the same porosity, higher PPI
better performance compared to the case with 5 fins. This phenomenon shows a higher number of pores meaning that the average diameter of
is probably due to the significant reduction in free convection heat the pores is smaller. The simulations have been carried out for the
transfer in the 5 finned configuration compared to the 1 finned case. melting process of PCM in three types of metal foams with 90% porosity
and 10, 20 and 30 PPIs at 9.2 W heat generation rate. The variation of
the battery means temperature for the samples are shown in Fig. 13.
6.3. The effect of metal foam on PCM performance improvement
The results show that under the considered conditions in this study, a
larger PPI performs better to reduce the mean battery temperature, so
In the current study, the local thermal non-equilibrium (LTNE)
the metal foam with a 30 PPI is used to perform further studies.
model is used for validations. The changes in the liquid fraction of PCM
Porosity is the main parameter in the porous medium. It is the ratio
and battery mean temperature over time for local thermal equilibrium
of the volume of free space to the total volume of the porous media. By
(LTE) model and LTNE model are shown in Fig. 12. As it is demon-
reducing the porosity, the effective conduction ratio is increased
strated, significant errors can be observed for the LTE model in both
(Eqs. (13)–(21)) and the free convective heat transfer is decreased
figures. For example, using metal foam, delays the melting start time

7
M.M. Heyhat, et al. Journal of Energy Storage 28 (2020) 101235

Fig. 7. Effect of the nanoparticle concentration on the liquid fraction variations Fig. 9. The effect of fin number on the liquid fraction variations at two heat
at two heat generation rates. generation rates.

(Eqs. (22) and (23)). The PCM liquid mass fraction with respect to time delay is caused by the increase in the effective conduction factor of the
in 4.6 W and 9.2 W heat generation rates are shown in Fig. 14. The solid PCM which leads to more uniform temperature distribution
results indicate that the effect of metal foam is more significant in high through the PCM. As the result, for the battery surface temperature
heat generation rates. Also the porosity reduction in both heat gen- (which is the point where the PCM melting starts) more time is required
eration rates leads to heat transfer improvement in melting process. The to reach the PCM melting temperature. By studying the battery mean
battery mean temperature is shown with respect to time in Fig. 15 for temperature diagrams in Fig. 15, it can be deduced that this phenom-
4.6 W and 9.2 W heat generation rates. These two figures display that enon adversely affects the battery TMS performance. Therefore, trying
using metal foams leads to considerable improvement in heat transfer to reduce this delay time, can hope to improve the performance of
and battery mean temperature. On the other hand, the porosity seems porous-PCM composition of TMS.
to have minor impact on battery temperature reduction. Even after The contours of temperature and velocity for the pure PCM and the
complete melting, it turns to have negative effect on battery tempera- porous-PCM cases are shown in Figs. 16 and 17, respectively. The re-
ture at some point. The best outcomes were observed for 90% porosity sults illustrate that the use of metal foam reduces the overall system
compared to 85% and 95%. It should be noted that finding optimal temperature and makes the system temperature contour more homo-
porosity for metal foam in every problem is of the most importance geneous. For example, at 1500s, the difference between the highest and
because porosity is directly related to important parameters such as the lowest temperatures for the pure PCM case is about 4 K; while, it is
cost, density, thermal capacity and unit weight. Another notable phe- about 2 K for the porous-PCM case. The velocity contours show that the
nomenon, which can be observed in Fig. 14 (A and B regions), is the utilization of metal foam reduces the velocity and free convection heat
considerable delay in PCM melting starting time for the porous-PCM transfer.
composition in comparison with the pure PCM. The reason for this

Fig. 8. Effect of nanoparticle concentration on the battery mean temperature at two heat generation rates, (a) 4.6 W, and (b) 9.2 W.

8
M.M. Heyhat, et al. Journal of Energy Storage 28 (2020) 101235

Fig. 10. The effect of fin number on the battery mean temperature at two heat generation rates, (a) 4.6 W, and (b) 9.2 W.

Fig. 11. Temperature contours for cases with fins at 4.6 W heat generation rate at 3000 s.

6.4. Comparing nano-PCM, fins and metal foam effects on PCM better behavior in liquid phase. On the other hand, using metal foam
performance and fins lead to heat transfer improvement as well as melting time re-
duction. The results show that the effects of using metal foam and fins
In this section the effects of three methods on PCM function on the are more considerable on higher heat generation rates. Metal foam is
TMS of battery are compared. observed to have the best function among the three methods in both
Fig. 18 represents the liquid fraction contours for different cases at heat generation rates. It is notable that the volume fraction of used
4.6 W heat generation rate at 2400 s. Results indicate that use of the copper in each of these methods is different but it is approximately in
mentioned three methods has significant effects on the PCM melting the same range. For example the volume of Cu for 5 fins is nearly 7.6%
and flow characteristics. Application of the metal foam leads to a uni- of total volume of PCM enclosure which can be compared with 7%
form heat transfer inside the PCM and provides a better melting per- volume fraction nanoparticles and metal foam with 90% porosity. This
formance. For a better comparison, a new parameter called improved comparison reveals that with nearly the same volume fraction of
battery mean temperature ΔTavg, battery is introduced. This parameter copper, use of metal foam can be the most effective way to improve the
represents the temperature difference between the case of using pure performance of the TMS of battery. Another salient point in Figs. 19 and
PCM and each of other methods. This parameter for three methods with 20 is the behavior of ΔTavg, battery for the samples with metal foam (as
respect to time are shown in Figs. 19 and 20 for 4.6 W and 9.2 W heat indicated with C and D dashed lines in these figures). A significant
generation rate, respectively. The results show that using nanoparticles reduction in ΔTavg, battery is observed in these areas that are caused by
does not lead to positive effect on battery temperature reduction the delay in PCM melting initiation time (as was previously discussed in
especially after complete melting which the pure PCM does show a Section 6.3.). The peak of ΔTavg, battery is also observed in the other two

9
M.M. Heyhat, et al. Journal of Energy Storage 28 (2020) 101235

Fig. 12. Comparison of the two models (LTE vs. LTNE) for simulation of porous PCM with 9.2 W heat generation rate; (a) liquid fraction, and (b) The battery mean
temperature.

Fig. 13. The effect of pore density on the battery mean temperature at 9.2 W
heat generation rate. Fig. 14. The effect of metal foam porosity on the liquid fraction variations at
two heat generation rates, (a) 4.6 W, and (b) 9.2 W.

methods. Due to the minor melting delay for them, their peaks are not
comparable with the peak of porous-PCM cases. There is also another by 10%. However, after complete melting of PCM, no significant dif-
peak created for ΔTavg, battery after the time that PCM is fully melted. ference is observed between the pure and modified PCM with fins or
Results of Figs. 19 and 20 clearly indicate that the performance of nanoparticles. In the case of porous-PCM, the results show that because
porous-PCM composition can be considerably improved by decreasing of the delay in the start time of melting, the stored energy is about half
this delay time. of the energy absorbed by the pure PCM. In return, after starting the
In order to study the rate of energy absorption in an accurate and melting process, the stored energy increases dramatically and at the end
detailed way, the total thermal energy of PCM section including PCM of the melting process, it reaches the same range as that of the pure
sensible and latent thermal energy and sensible energy of copper used PCM. Eventually, after a complete meltdown, the total stored thermal
in nanoparticles, fins and metal foam in each case of study are calcu- energy in the PCM containing metal foam is about 85–95% of the total
lated and are shown in Fig. 21. The results show that in the current energy of the pure PCM. Even though using metal foam has an adverse
study, generally pure PCM has better performance in heat storage. In effect on the total amount of energy, it has shown an ideal performance
order to reach a better understanding, the dimensionless number of as a thermal management system. In fact, using metal foam in the
heat storage ratio is introduced which is defined as the ratio of absorbed specified conditions and geometry, such as the current study, results in
energy in the modified PCM to pure PCM. The change in this parameter more heat disposal from the system to the environment.
over time is presented in Fig. 22. Results show that until the complete
melting of PCM, the heat storage rate is slightly improved by fins and
nanoparticles, and by using 5 fins the total absorbed energy is increased

10
M.M. Heyhat, et al. Journal of Energy Storage 28 (2020) 101235

Fig. 15. The effect of metal foam porosity on the battery mean temperature at two heat generation rates, (a) 4.6 W, and (b) 9.2 W.

Fig. 16. Temperature contours for the pure PCM and porous-PCM cases at 4.6 W heat generation rate.

11
M.M. Heyhat, et al. Journal of Energy Storage 28 (2020) 101235

Fig. 17. Velocity and stream function contours for the pure PCM and porous-PCM cases at 4.6 W heat generation rate at 3000 s.

Fig. 18. Liquid fraction contours for different cases at 4.6 W heat generation rate at 2400 s.

12
M.M. Heyhat, et al. Journal of Energy Storage 28 (2020) 101235

Fig. 21. Variations of heat storage for three methods at 9.2 W heat generation
rate.

Fig. 19. Variations of ΔTavg, batteryfor three methods at 4.6 W heat generation
rate.

Fig. 22. Variations of heat storage Ratio for three methods at 9.2 W heat
generation rate.

Fig. 20. Variations ofΔTavg, for three methods at 9.2 W heat generation management system performance.
• Using fins in 4.6 W and 9.2 W heat generation rates, led to battery
battery
rate.
mean temperature reduction of 2 and 4 K, respectively. The results
7. Conclusions indicated that the addition of fins did not always lead to system
performance improvement. In the studied configurations, it revealed
In this research the melting process of PCM in passive thermal that the thermal management system had better performance with 3
management system of 18,650 lithium ion battery was simulated in fins than 5 fins.
4.6 W and 9.2 W heat generation rates. Moreover, the effects of three • Using metal foam in 4.6 W and 9.2 W heat generation rates led to
major PCM heat transfer improvement methods such as using nano- battery mean temperature reduction of 4 and 6 K, respectively. The
particles, fins and metal foam were scrutinized and compared with each change in porosity from 95% to 85% did not lead to significant
other. The results showed that: change on thermal management system performance.
• In approximately the same volume of copper, using metal foam was

• The addition of nanoparticles into PCM in both 4.6 W and 9.2 W more efficient for thermal management system improvement.
heat generation rates had insignificant effect on thermal
Finally, investigation of the performance of a passive thermal

13
M.M. Heyhat, et al. Journal of Energy Storage 28 (2020) 101235

management system in real driving cycles. Also, design and study of a [15] F. Selimefendigil, H.F. Oztop, A.J. Chamkha, Natural convection in a CuO–water
passive thermal management system for a battery pack instead of a nanofluid filled cavity under the effect of an inclined magnetic field and phase
change material (PCM) attached to its vertical wall, J. Therm. Anal. Calorim. 135
single cell, in addition to, design and evaluation of a hybrid tempera- (2019) 1577–1594, https://doi.org/10.1007/s10973-018-7714-9.
ture control system (consisting of active and passive temperature con- [16] F. Selimefendigil, H.F. Öztop, Corrugated conductive partition effects on MHD free
convection of CNT-water nanofluid in a cavity, Int. J. Heat Mass Transf. 129 (2019)
trol systems) can be suggested as some future research works. 265–277, https://doi.org/10.1016/j.ijheatmasstransfer.2018.09.101.
[17] B.R. Sushobhan, S.P. Kar, Thermal modeling of melting of nano based phase change
CRediT authorship contribution statement material for improvement of thermal energy storage, Energy Procedia 109 (2017)
385–392, https://doi.org/10.1016/j.egypro.2017.03.035.
[18] F. Bayrak, H.F. Oztop, F. Selimefendigil, Effects of different fin parameters on
Mohammad Mahdi Heyhat: Formal analysis, Investigation, temperature and efficiency for cooling of photovoltaic panels under natural con-
vection, Sol. Energy. 188 (2019) 484–494, https://doi.org/10.1016/j.solener.2019.
Writing - review & editing, Supervision. Sepehr Mousavi:
06.036.
Methodology, Software, Validation, Formal analysis, Data curation, [19] F. Selimefendigil, H.F. Öztop, Forced convection in a branching channel with partly
Writing - original draft, Visualization. Majid Siavashi: elastic walls and inner L-shaped conductive obstacle under the influence of mag-
netic field, Int. J. Heat Mass Transf. 144 (2019) 118598, , https://doi.org/10.1016/
Conceptualization, Methodology, Formal analysis, Investigation, j.ijheatmasstransfer.2019.118598.
Resources, Writing - review & editing, Supervision, Project adminis- [20] S. Mousavi, M. Siavashi, M.M. Heyhat, Numerical melting performance analysis of a
tration, Funding acquisition. cylindrical thermal energy storage unit using nano-enhanced PCM and multiple
horizontal fins, Numer. Heat Transf. Part Appl. 75 (2019) 560–577, https://doi.org/
10.1080/10407782.2019.1606634.
Declaration of Competing Interest [21] Z. Wang, H. Zhang, X. Xia, Experimental investigation on the thermal behavior of
cylindrical battery with composite paraffin and fin structure, Int. J. Heat Mass
Transf. 109 (2017) 958–970, https://doi.org/10.1016/j.ijheatmasstransfer.2017.
The authors whose names are listed immediately below certify that 02.057.
they have NO affiliations with or involvement in any organization or [22] X. Yang, P. Wei, X. Cui, L. Jin, Y.-L. He, Thermal response of annuli filled with metal
foam for thermal energy storage: an experimental study, Appl. Energy. 250 (2019)
entity with any financial interest (such as honoraria; educational 1457–1467, https://doi.org/10.1016/j.apenergy.2019.05.096.
grants; participation in speakers’ bureaus; membership, employment, [23] X. Yang, Q. Bai, Z. Guo, Z. Niu, C. Yang, L. Jin, T.J. Lu, J. Yan, Comparison of direct
consultancies, stock ownership, or other equity interest; and expert numerical simulation with volume-averaged method on composite phase change
materials for thermal energy storage, Appl. Energy. 229 (2018) 700–714, https://
testimony or patent-licensing arrangements), or non-financial interest doi.org/10.1016/j.apenergy.2018.08.012.
(such as personal or professional relationships, affiliations, knowledge [24] X. Yang, J. Yu, Z. Guo, L. Jin, Y.-.L. He, Role of porous metal foam on the heat
transfer enhancement for a thermal energy storage tube, Appl. Energy 239 (2019)
or beliefs) in the subject matter or materials discussed in this manu- 142–156, https://doi.org/10.1016/j.apenergy.2019.01.075.
script. [25] K. Kant, A. Shukla, A. Sharma, P. Henry Biwole, Heat transfer study of phase change
materials with graphene nano particle for thermal energy storage, Sol. Energy 146
(2017) 453–463, https://doi.org/10.1016/j.solener.2017.03.013.
References [26] W.Q. Li, Z.G. Qu, Y.L. He, Y.B. Tao, Experimental study of a passive thermal
management system for high-powered lithium ion batteries using porous metal
[1] S. Landini, J. Leworthy, T.S. O'Donovan, A review of phase change materials for the foam saturated with phase change materials, J. Power Sources 255 (2014) 9–15,
thermal management and isothermalisation of lithium-ion cells, J. Energy Storage https://doi.org/10.1016/j.jpowsour.2014.01.006.
25 (2019) 100887, , https://doi.org/10.1016/j.est.2019.100887. [27] H. Yang, H. Zhang, Y. Sui, C. Yang, Numerical analysis and experimental visuali-
[2] A.N. Jansen, D.W. Dees, D.P. Abraham, K. Amine, G.L. Henriksen, Low-temperature zation of phase change material melting process for thermal management of cy-
study of lithium-ion cells using a LiySn micro-reference electrode, J. Power Sources lindrical power battery, Appl. Therm. Eng. 128 (2018) 489–499, https://doi.org/
174 (2007) 373–379, https://doi.org/10.1016/j.jpowsour.2007.06.235. 10.1016/j.applthermaleng.2017.09.022.
[3] P. Nelson, D. Dees, K. Amine, G. Henriksen, Modeling thermal management of li- [28] M.D. Muhammad, O. Badr, H. Yeung, Validation of a CFD melting and solidification
thium-ion PNGV batteries, J. Power Sources 110 (2002) 349–356, https://doi.org/ model for phase change in vertical cylinders, Numer. Heat Transf. Part Appl. 68
10.1016/S0378-7753(02)00197-0. (2015) 501–511, https://doi.org/10.1080/10407782.2014.994432.
[4] Y. Wei, M. Agelin-Chaab, Development and experimental analysis of a hybrid [29] F. Iachachene, Z. Haddad, H.F. Oztop, E. Abu-Nada, Melting of phase change ma-
cooling concept for electric vehicle battery packs, J. Energy Storage 25 (2019) terials in a trapezoidal cavity: orientation and nanoparticles effects, J. Mol. Liq. 292
100906, , https://doi.org/10.1016/j.est.2019.100906. (2019) 110592, , https://doi.org/10.1016/j.molliq.2019.03.051.
[5] Z. Qian, Y. Li, Z. Rao, Thermal performance of lithium-ion battery thermal man- [30] V.R. Voller, C. Prakash, A fixed grid numerical modelling methodology for con-
agement system by using mini-channel cooling, Energy Convers. Manag. 126 (2016) vection-diffusion mushy region phase-change problems, Int. J. Heat Mass Transf. 30
622–631, https://doi.org/10.1016/j.enconman.2016.08.063. (1987) 1709–1719, https://doi.org/10.1016/0017-9310(87)90317-6.
[6] L. Feng, S. Zhou, Y. Li, Y. Wang, Q. Zhao, C. Luo, G. Wang, K. Yan, Experimental [31] Y. Tian, C.Y. Zhao, A numerical investigation of heat transfer in phase change
investigation of thermal and strain management for lithium-ion battery pack in heat materials (PCMs) embedded in porous metals, Energy 36 (2011) 5539–5546,
pipe cooling, J. Energy Storage 16 (2018) 84–92, https://doi.org/10.1016/j.est. https://doi.org/10.1016/j.energy.2011.07.019.
2018.01.001. [32] H. Shmueli, G. Ziskind, R. Letan, Melting in a vertical cylindrical tube: numerical
[7] S.A. Hallaj, J.R. Selman, A novel thermal management system for electric vehicle investigation and comparison with experiments, Int. J. Heat Mass Transf. 53 (2010)
batteries using phase‐change material, J. Electrochem. Soc. 147 (2000) 3231–3236, 4082–4091, https://doi.org/10.1016/j.ijheatmasstransfer.2010.05.028.
https://doi.org/10.1149/1.1393888. [33] B.J. Jones, D. Sun, S. Krishnan, S.V. Garimella, Experimental and numerical study of
[8] Y.-.H. Huang, W.-.L. Cheng, R. Zhao, Thermal management of Li-ion battery pack melting in a cylinder, Int. J. Heat Mass Transf. 49 (2006) 2724–2738, https://doi.
with the application of flexible form-stable composite phase change materials, org/10.1016/j.ijheatmasstransfer.2006.01.006.
Energy Convers. Manag. 182 (2019) 9–20, https://doi.org/10.1016/j.enconman. [34] K. Boomsma, D. Poulikakos, On the effective thermal conductivity of a three-di-
2018.12.064. mensionally structured fluid-saturated metal foam, Int. J. Heat Mass Transf. 44
[9] G. Karimi, M. Azizi, A. Babapoor, Experimental study of a cylindrical lithium ion (2001) 827–836, https://doi.org/10.1016/S0017-9310(00)00123-X.
battery thermal management using phase change material composites, J. Energy [35] Z. Liu, Y. Yao, H. Wu, Numerical modeling for solid–liquid phase change phe-
Storage 8 (2016) 168–174, https://doi.org/10.1016/j.est.2016.08.005. nomena in porous media: shell-and-tube type latent heat thermal energy storage,
[10] R. Sabbah, R. Kizilel, J.R. Selman, S. Al-Hallaj, Active (air-cooled) vs. passive (phase Appl. Energy 112 (2013) 1222–1232, https://doi.org/10.1016/j.apenergy.2013.02.
change material) thermal management of high power lithium-ion packs: limitation 022.
of temperature rise and uniformity of temperature distribution, J. Power Sources. [36] Y. Zhao, C.Y. Zhao, Z.G. Xu, H.J. Xu, Modeling metal foam enhanced phase change
182 (2008) 630–638, https://doi.org/10.1016/j.jpowsour.2008.03.082. heat transfer in thermal energy storage by using phase field method, Int. J. Heat
[11] L. Ghadbeigi, B. Day, K. Lundgren, T.D. Sparks, Cold temperature performance of Mass Transf. 99 (2016) 170–181, https://doi.org/10.1016/j.ijheatmasstransfer.
phase change material based battery thermal management systems, Energy Rep 4 2016.03.076.
(2018) 303–307, https://doi.org/10.1016/j.egyr.2018.04.001. [37] F. Iachachene, Z. Haddad, H.F. Oztop, E. Abu-Nada, Melting of phase change ma-
[12] R.D. Jilte, R. Kumar, M.H. Ahmadi, L. Chen, Battery thermal management system terials in a trapezoidal cavity: orientation and nanoparticles effects, J. Mol. Liq.
employing phase change material with cell-to-cell air cooling, Appl. Therm. Eng. (2019), https://doi.org/10.1016/j.molliq.2019.03.051.
161 (2019) 114199, , https://doi.org/10.1016/j.applthermaleng.2019.114199. [38] J.C. Maxwell, A Treatise on Electricity and Magnetism, Oxford University Press,
[13] Q. Wang, B. Jiang, B. Li, Y. Yan, A critical review of thermal management models Cambridge, 1904.
and solutions of lithium-ion batteries for the development of pure electric vehicles, [39] M. Al-Jethelah, S.H. Tasnim, S. Mahmud, A. Dutta, Nano-PCM filled energy storage
Renew. Sustain. Energy Rev. 64 (2016) 106–128, https://doi.org/10.1016/j.rser. system for solar-thermal applications, Renew. Energy 126 (2018) 137–155, https://
2016.05.033. doi.org/10.1016/j.renene.2018.02.119.
[14] F. Selimefendigil, H.F. Öztop, Effects of an inner stationary cylinder having an [40] M. Corcione, Empirical correlating equations for predicting the effective thermal
elastic rod-like extension on the mixed convection of CNT-water nanofluid in a conductivity and dynamic viscosity of nanofluids, Energy Convers. Manag. 52
three dimensional vented cavity, Int. J. Heat Mass Transf. 137 (2019) 650–668, (2011) 789–793, https://doi.org/10.1016/j.enconman.2010.06.072.
https://doi.org/10.1016/j.ijheatmasstransfer.2019.03.093.

14