Energy Conversion and Management 204 (2020) 112280

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: [Link]/locate/enconman

A novel battery thermal management system coupling with PCM and T

optimized controllable liquid cooling for different ambient temperatures
⁎ ⁎
Depeng Konga, , Rongqi Penga, Ping Pingb, Jin Dua, Guoming Chenb, Jennifer Wenc,
a
Center for Offshore Engineering and Safety Technology, China University of Petroleum (East China), Qingdao 266580, PR China
b
College of Chemical Engineering, China University of Petroleum (East China), Qingdao 266580, PR China
c
Warwick FIRE, School of Engineering, University of Warwick, Library Road, Coventry CV4 7AL, UK

A R T I C LE I N FO A B S T R A C T

Keywords: In order to improve the working performance of the lithium-ion battery pack in continuous operation under
Phase change material different ambient temperatures, a coupled composite phase change material and liquid cooling thermal man-
Liquid cooling agement system is proposed. The simulation for this system under a cycle that a 3C rate discharging and then a
Lithium-ion battery 0.5C charging was conducted, as well as comparison tests concerning factors such as cell-to-cell spacing, cell-to-
Coupled thermal management
tube distance, channel number and coolant velocity. Simulation results showed that the coupled system with
Different ambient temperatures
Controlling strategy
suitable design exhibited good thermal performance at an ambient temperature of 30 °C, which kept the max-
imum surface temperature and the temperature difference of the battery pack at 41.1 °C and 4 °C at the end of 3C
discharge. Then, the latent heat of phase change material was also recovered by the liquid cooling during the
0.5C charge. Specially designed experiments have also been conducted to verify the effectiveness and practic-
ability of the proposed coupled system. Based on this system, a liquid cooling strategy was proposed for con-
trolling the velocity and inlet temperature of coolant by monitoring the temperature of the phase change ma-
terial and environment. This further improved the thermal performance of the battery pack during cycling at
different ambient temperatures and significantly reduced the unnecessary power consumption of liquid cooling
during this process.

1. Introduction than air cooling, simpler structure than liquid cooling and wider ap-
plication for the batteries of different shapes than heat pipes which only
From numerous secondary power batteries, lithium-ion battery fit for rectangular surface [14].
(LIB) with high energy density and long cycle is extensively applied as Nevertheless, two important barriers limit the applications of PCM-
the heart and the energy supporting equipment of electric vehicles based BTMS [2]. One is that the pure PCM, such as the paraffin, is not
(EVs) [1]. One of the most significant factors affecting the performance, capable of meeting the requirements of the fast response to thermal
lifespan and safety of LIB is the operating temperature [2]. Therefore, surge owing to its relatively low thermal conductivity [15]. Therefore, a
an effective battery thermal management system (BTMS) is imperative great deal of effort has been made to improve its thermal conductivity
to guarantee that the battery pack can be operated in the desired adding materials with high thermal conductivity, such as expanded
temperature range of 20–50 °C [3] and keep the temperature distribu- graphite (EG) [16,17], metal foam [18,19], and carbon fiber [20,21].
tion from cell to cell within 5 °C [4]. Consequent to these requirements, The other is that heat accumulation may be likely to happen in the case
numerous investigations on the BTMS have been performed based on of running out of the available latent heat due to the low surface heat
the employment of different heat transfer medium such as air [5,6] or/ transfer coefficient of PCM [22]. This limitation leads to a rapid decline
and liquid cooling [7,8], heat pipe cooling [9,10] and phase change in heat dissipation effectiveness during continuous charge-discharge
materials (PCMs) cooling [11,12]. As a novel solution for thermal cycles [2,22]. Furthermore, the complete melting of PCM would cause
management applications, PCM can absorb a large amount of heat thermal insulation from batteries to heat sink, resulting in heat accu-
during its phase change process, while the battery temperature is mulation [23] and increases the risk of thermal runaway.
maintained around the phase change temperature for a period of time To overcome these shortages, PCM coupled with other heat dis-
[13]. Besides, the PCM-based BTMS provides higher cooling capacity sipation methods were proposed. Both passive (fin structure [24,25]

⁎
Corresponding authors.
E-mail addresses: kongdepeng@[Link] (D. Kong), [Link]@[Link] (J. Wen).

[Link]
Received 9 July 2019; Received in revised form 27 September 2019; Accepted 9 November 2019
Available online 20 November 2019
0196-8904/ © 2019 Elsevier Ltd. All rights reserved.
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

Nomenclature Ttube surface temperature of internal channel, K

△T temperature difference, K
Ac cross sectional area of channel, m3 v velocity, m s−1
Cp heat capacity, J kg−1 K−1
d cell-to-tube distance of the PCM, mm Greek symbols
dc hydraulic diameter of channel, m
H enthalpy, J K−1 β liquid fraction, %
h convection heat transfer coefficient, W m−2 K−1 μ dynamic viscosity, Pa s
k thermal conductivity, W m−1 K−1 ρ density, kg m−3
L latent heat, J
P pressure, Pa Subscripts
Pc wetting perimeter of channel, m2
Qgen the heat generation of the battery, W amb, 0 ambient temperature
Qb,j joule heat, W b battery
Qb,r reversible entropic heat, W c coolant
Qb,p irreversible electrochemical reaction heat, W l liquid phase
Rc thermal contact resistance, K W−1 max maximum
Re Reynolds number min minimum
rc radius of channel s solid phase
S thermal contact surface t tubes
s cell-to-cell spacing, mm 5 Point 5
T temperature, K

and heat pipe [26,27]) and active (thermoelectric coolers (TECs) [28], while the effect was not obvious at 37 °C. Qin et al. [32] further de-
air cooling [29–33] and liquid cooling [14,34–37]) coupling strategies signed a coupled BTMS that each battery was surrounded symmetrically
have been proposed. Passive coupling has the distinct advantages of with four PCM filled cavities and vents. Comparing with the natural
lightweight and no additional energy consumption. Due to the ex- convection, the maximum temperature difference and the maximum
tensive application in the field of latent heat thermal energy storage temperature in the active strategy with the forced air cooling are re-
(LHTES) [38,39], fin structures with PCM have been gradually con- duced 1.2 °C and 16 °C under 3C discharge rate, respectively. Coupled
sidered as the thermal management of LIB. Our previous study [25] PCM and air cooling BTMS has advantages in many aspects such as the
showed that PCM-fin structure could enhance the heat transfer effi- cost, weight and maintainability, but its thermal performance need
ciency between PCM and ambient due to improved heat conduction and more improvement under abuse conditions with elevated temperatures.
large heat transfer surface, prolonging the duration of thermal control Another commonly used active cooling method is the liquid cooling
of the battery module under continuous cycles. Heat pipe, as a highly system. Generally, it is easier to implement the liquid cooling in pris-
efficient heat transfer device, is also widely used in PCM based BTMS. matic batteries than cylindrical ones due to the poor contact between
Huang et al. [26] designed such a system and found that the heat ac- the cylindrical surface and the flat tube/plate walls [1,14]. In contrast,
cumulation of PCM can be effectively reduced by timely transporting PCM is easily shaped into various shapes, which could be integrated
heat to the module’s side direction. into BTMS for batteries with various geometries. For the cylindrical
In practical applications, passive coupled PCM BTMS could be in- battery pack, coupled PCM-liquid cooling system seems to be an ef-
sufficient. To enhance the heat dissipation efficiency, auxiliary active fective solution to not only reduce the potential failure probability of
cooling by air or liquid was introduced. A thermoelectric cooler (TEC) is PCM, but also achieve rapid heat dissipation for PCM-based BTMS.
a solid-state device based on the working principle of the thermo- Song et al. [36] investigated a coupled BTMS that the heat generated by
electric effects in semiconductors [40], which would cease temperature batteries was conducted to a liquid cooled mini-channel cold plate
gradients when an electrical current is applied. For standby battery of through the PCM and heat spreading column. Their results demon-
outdoor base station, Song et al. [28] proposed a BTMS based on the strated that both the battery temperature ramp-up rate and the battery
TEC and PCMs, and reported the TEC could cool or heat PCMs and keep steady-state temperature were significantly reduced by the coupled
the battery module in an appropriate temperature range during dif- cooling, in comparison with single PCM or liquid cooling condition.
ferent ambient temperatures and cycling process. However, the low Recently, Kshetrimayum et al. [37] developed a thermal management
efficiency and lifetime of TECs [41,42] limited their application in the method involving an integrated design of PCM and cooling plate and
battery pack for EVs. the thermal performance of this coupled system was predicted nu-
Considering that natural convection is not sufficient to cool batteries merically under normal operation and nail penetration event, respec-
and PCMs down during continuous cycling, introducing forced air tively. Simulation result of thermal runaway indicated that the max-
convection to enhance the heat transfer coefficient seems to a viable imum temperature of the cells adjacent to thermal abused cell was
solution. Ling et al. [29] coupled the PCM with forced air for the maintained below 90 °C, and the heat propagation to the adjacent cells
thermal management of cylindrical batteries and investigated the per- was prevented.
formance of the as-proposed system experimentally and numerically. A Currently, most studies about BTMS based on PCMs coupled with
maximum temperature of 46 °C was recorded for the system in cycling liquid cooling could suppress the maximum temperature of battery pack
with discharge rate of 2C under the ambient temperature of 34 °C. A within 50 °C, even in high rate discharge. Nevertheless, it is a difficult
tube-shell battery pack with a PCM coupled with forced air cooling problem to balance the cooling efficiency between the PCM and liquid
BTMS was designed by Jiang et al. [31]. It was found that this coupled cooling in coupled PCM and liquid cooling system. In the designed
system could control the cell temperature below 52 °C and kept the system by the Zheng et al. [34], dense aluminum cooling tubes install
maximum temperature difference within a low value of 1–2 °C. Besides, around the prismatic batteries and composite PCM fills up the clearance
the introduction of the baffles remarkably improved the heat transfer of of batteries and cooling tubes. Simulation results showed that limited
air cooling in the coupled system at the ambient temperature of 28 °C, PCM was hard to take the advantage of the latent heat, and the

2
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

comparison tests indicated that the change of PCM phase change in- scanning calorimetry (DSC) measurement (METTLER DSC-3), as shown
terval had little impact on the cooling performance of this coupled in Table 1. Five S-shaped aluminum tubes were installed around the
system. Similarly, Ling et al. [14] pointed out that when the forced batteries as the liquid tube from top to bottom, and the liquid tubes
liquid convection increased up to 1000 W m−2 °C−1, the active cooling were supported by brackets at both ends. The coolant in the liquid tubes
was so intense that PCM temperature even fell below its melting point, was (1:1) mixture water-ethylene glycol, which could provide a greater
which implied that all the heat was stored in the PCM in the form of resistance against biological fouling and cold weather [42,44], and its
sensible heat and the latent heat was not activated. These results in- temperature dependent physical and thermal properties are shown in
dicate that the coupling system with highly efficient liquid cooling Table 2. The flow of coolant in the topmost liquid tube was shown in
system and insufficient PCM may make the PCM latent heat insignif- Fig. 1(a). It is noteworthy that all the coolant flows in the left side of the
icant, even some other materials with higher thermal conductivity re- battery pack and flows out the right side, which could improve the
placing the PCM could be better. However, it is the latent heat of the cooling efficiency of the liquid cooling. The entire battery pack was
PCM that enables more uniform temperature control and lower power packaged with insulating material to ensure electrical safety. Besides,
consumption in the coupled system. For the pure liquid cooling system, the flow inlet of the top tube is on the left side in the Y direction, and
it may require more sophisticated design and higher cost to achieve the the lower tube was opposite the top tube and so on.
similar performance. Therefore, it is worth investigating the design of a
coupled PCM-liquid cooling system that PCM is sufficient to take full 2.2. Model analysis
use of the latent heat during the discharge process and the liquid
cooling is light and simple enough to recover the latent heat during the Simulation can be very helpful to implement a battery model and
charge process. Moreover, the performance of coupled PCM-liquid achieve best design solutions in terms of cooling and thermal man-
cooling system during continuous charge-discharge cycle has rarely agement [45]. In this work, numerical simulations were conducted
been systematically studied for battery packs, especially at different using COMSOL Multiphysics to test the thermal performance of the
ambient temperatures. system shown in Fig. 1. The 3D-models contain four subdomains of
To address the aggressive thermal management challenges under LIBs, PCM, liquid tubes and coolant.
continuous cycling and abuse ambient temperature, a novel Coupled For a cylindrical battery, it consists of connector, mandrel, canister
PCM-Liquid cooling thermal management System (CPLS) is proposed in and active battery materials. A uniform battery heat generation model
the present study, in which the PCM is used as the main heat dissipation based on the 1D-electrochemical and 3D-thermal coupling models
material to keep the battery at the optimum operating temperature (Pseudo Two-dimensional model) in our previous work was used here
during discharging, and the liquid cooling acts as the role of recovering [25]. The thermos-physical properties of the battery materials/com-
the latent heat of the PCM timely during charging. In order to increase ponents and the geometric parameters of the battery are listed in
the utilization of PCM at different ambient temperatures, a controlling Table 3. For the composite PCM, the properties were assumed to be
strategy was proposed to control the velocity of the liquid cooling by constant and identical for both liquid and solid phase [46,47]. The
the feedback of temperature changes at designated location of the motion of solid PCM, the volume variation and the convective heat
battery pack in this study. Besides, the coolant also could cool/heat the transfer between PCM and EG were neglected during the period of
PCM to prevent the batteries from suffering in a hot/cold environment. phase change [48]. For the coolant, the Reynolds number Re here de-
The velocity and temperature of the coolant could be adjusted based on scribes the liquid coolant through the flow channel and is defined
the strategy to reduce needless power dissipation and achieve reliable below:
temperature performance.
ρc vc dc
A combined experimental and numerical investigation on the Re =
uc (1)
thermal characteristics of the CPLS was carried out. The thermal
characterization of the CPLS was firstly investigated during one cycle. 4Ac 4πrc2
Then, the effects of various factors, such as PCM geometric parameters, dc =
Pc 2πrc (2)
channel number and coolant velocity on the battery temperature were
analyzed to optimize the design. The comparison was subsequently where ρc is density of the coolant, vc is the velocity, dc is the hydraulic
conducted between the numerical predictions and experimental mea- diameter of the channel, Ac is the cross sectional area of the channel, Pc
surements to verify the thermal performance of CPLS in practical ap- is the wetting perimeter of the channel, rc is the radius of the channel.
plications. Finally, the controlling strategy for the liquid cooling system The inlet velocity for the channel is from 0.1 m/s to 1 m/s, and the
in CPLS was introduced and charging-discharging cycle tests were initial temperature of inlet coolant ranges from 10 °C to 40 °C. The
conducted to rank the effectiveness of the controlling strategy at dif- values of Reynolds number for the coolant flow are among
ferent ambient temperatures.
Table 1
2. Simulation solution Specifications and parameters of NMC battery and PCM.
Specifications Value
2.1. Concept description
Diameter of cell 21.7 ± 0.2 mm
Height of cell 70.9 ± 0.2 mm
Commercial 21,700 Li-ion batteries of LiNiMnCoAlO2 (NMC)
Mass of cell 69 ± 2 g
cathode were considered. The basic parameters of NMC battery are Nominal voltage 3.65 V
tabulated in Table 1. The energy content per cell for 21,700 battery is Nominal capacity 4 Ah
higher by ~50% compared to the 18,650 battery [43], which involves Recommended temperature of cell 0–45 °C (charge)
−20–60 °C (discharge)
more heat generation and stricter requirement for BTMS.
Thermal conductivity of PCM 5.023 W m−1 K−1
Fig. 1 shows the schematic diagram of the battery pack with CPLS. Density of PCM 870 kg m−3
Six 21,700 Li-ion batteries were connected in series firstly and then four Specific heat capacity of PCM 2412 J kg−1 K−1
series were connected in parallel (6S4P). The batteries were inserted Phase-transition temperature range 31–36 °C
into PCM matrix which consists of pure PCM, lipid and expanded gra- Latent heat absorption 119.24 kJ kg−1
Density of aluminum tube 2700 g cm−3
phite (EG). The thermal properties of the PCM matrix such as thermal
Heat Capacity of aluminum tube 900 J kg−1 K−1
conductivity, phase-transition temperature and latent heat capacity had Thermal conductivity of aluminum tube 238 W m−1 K−1
been measured by laser flash apparatus (TA FL4010) and differential

3
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

Fig. 1. Schematic diagram of the battery pack with the CPLS: (a) the overall system, (b) cross-section.

Table 2 ∂Tb
ρb cp, b = ∇ ·(kb ∇Tb) + Qgen
Thermo-physical properties of water-ethylene glycol (1:1) mixture [34]. ∂t (3)
Property Algebraic equations where ρb, kb and Tb are the density, thermal conductivity and tem-
−3 perature of battery, respectively. Qgen is the heat generation of the
Density (kg m ) ρc = -0.387Tc + 1081
Dynamic viscosity μc = 5.86 × 10−6Tc2 − 3.489 × 10−4Tc + 8.36 × 10−3 battery:
(Pa s)
Qgen = Qb, j + Qb, p + Qb, r (4)
Heat capacity Cp,c = 3.85 × 10−3Tc + 3.203
(kJ kg−1 K−1)
Heat conductivity kc = 8.3 × 10−4tl + 0.364
where Qb,j, Qb,p and Qb,r are the joule heat, irreversible electrochemical
(W m−1 K−1) reaction heat and the reversible entropic heat, respectively.
The heat transfer inside the PCM is governed by the following
equations:
118.4–1692.1 calculated with Eq. (1), which falling in the laminar re-
∂HPCM
gime. Besides, the uniform inflow velocity was set for the coolant inlet ρPCM = ∇ ·(kPCM ∇TPCM )
∂t (5)
and the inner wall of the wavy channel is assumed to be no-slip, whilst
0-pressure condition and the suppressed back flow were set for the TPCM
coolant outlet. HPCM = ∫T0
cPCM dT + βL
(6)

where ρPCM, Hpcm and kPCM are the density, enthalpy and effective
2.3. Governing equations thermal conductivity of PCM, TPCM and T0 represent the temperature of
PCM and ambient environment, respectively. L is the latent heat of the
As shown in Fig. 1(a), the generated heat from the battery charge PCM. β is the liquid fraction of the PCM, which can be expressed as:
and/or discharge was absorbed by the PCM firstly, and stored as the
latent heat. The coolant within the tubes removed the heat away and ⎧0 TPCM < Ts (solid phase )
recovered the latent heat of the PCM. The governing equations for four ⎪ TPCM − TS
β= Ts ≤ TPCM < Tl (Solid/ liquid phase )
subdomains are described as follows. ⎨ TL − TS
⎪1 TPCM ≥ Tl (liquid phase ) (7)
For the LIBs, the energy conservation equation can be written as: ⎩

Table 3
Thermo-physical properties of battery materials/components and geometric parameters of the battery [49–51].
Parameters Cathode current collector, aluminum Anode current collector, copper Anode Separator Cathode Mandrel Connector /Canister
foil foil

Thickness (um) 10 10 92 22 87
Thermal conductivity (W m−1 K−1) 238 398 5 1 5 0.26 44.5
Density (kg m−3) 2702 8933 1500 1200 2660 1150 7850
Specific heat capacity (J kg−1 K−1) 903 385 700 700 1437.4 1700 475
Particle radius (um) 5.3 13.1
Solid phase volume fraction 0.43 0.58
Electrolyte phase volume 0.32 0.54 0.32
Bruggeman coefficient for tortuosity 1.5 1.5 1.5

4
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

where Ts and Tl are the solid and liquid phase temperatures of the PCM, ambient is defined as follows:
respectively.
∂T
For the coolant, the conservation equations for the continuity, mo- − kPCM = h (TPCM − Tamb)
∂n (16)
mentum and energy terms are:
∂ρc ⇀
where kPCM represent the thermal conductivity of PCM, and h is the
+ ∇ ·(ρc vc ) = 0 natural convection heat transfer coefficient, whose value is
∂t (8)
5 W m−2 K−1 in this simulation. In addition, the radiation heat transfer
⇀
∂vc ⇀ ⇀ ⇀ was not taken into account.
ρc + ρc (vc ·∇) vc = −∇P + ∇ ·(μc ∇vc )
∂t (9) Sensitivity analysis of with three girds involving 2137975,
5,183,760 and 12,384,819 cells were conducted. Fig. 2(a) shows the
∂Tc
ρc Cp, c + ∇ ·(ρc Cp, c →
vc Tc ) = ∇ ·(kc ∇Tc ) predicted maximum temperature responses of the battery pack with
∂t (10)
CPLS during one cycle. It can be seen there is almost no difference
For the tube, the energy conservation equation can be expressed as between the predictions with different grid resolutions. Hence, the grid
follows: structure used in this study as shown in Fig. 2(b) includes 2,133,962
domain units, 28,836 boundary units and 904 vertices.
∂Tt
ρt Cp, t = ∇ ·(kt ∇Tt )
∂t (11)
2.4. Model verification
where (ρc, ρt), (Cp,c, Cp,t) (kc, kt) and (Tc, Tt) are density, specific heat
capacity, thermal conductivity and temperature of coolant and tubes,
For the purpose of verifying the heat generation rate of the battery,
respectively; →vc , P, and μc denote the coolant velocity vector, static
the experiment and simulation on the thermal behavior of 21,700
pressure and dynamic viscosity, respectively.
battery were firstly carried out at 3C discharge rate under natural
During the process of simulation, the initial state is specified as
convection conditions. The battery was placed in the thermostat in-
follows:
cubator, for which the Tamb fluctuation was controlled within ± 1 °C.
t = 0; T (x , y, z ) = Tamb; (12) Before the tests, the battery was cycled under 0.2C by the battery cycler
to ensure its actual capacity and reliability. Then, three K-type ther-
where Tamb is the ambient temperature.
mocouples with diameter of 0.5 mm were arranged on different loca-
Boundary condition at the interface between battery and the PCM is
tions of battery (near the top, on the middle, near the bottom of battery)
defined as follow:
to record the temperature, as shown in Fig. 3(a). The battery was dis-
∂T ∂T charged at 3C with a constant current. During the test, the thermostat
− kb = −kPCM
∂n ∂n (13) incubator temperature was kept at 30 °C. The comparison of battery
where kb and kPCM are the thermal conductivity of battery and PCM, temperature between experimental data and simulated results is pre-
and ∂T represents temperature gradient. sented in Fig. 3(a). It is clear to find that the simulation results agree
∂n
The contact surface between PCM and tubes shares similar well with the experimental results, and the maximum discrepancy is
boundary condition: less than 1.3 °C. The experimental results indicate the calculation
method of the heat generation rate of battery is accurate. In addition,
∂T ∂T the maximum temperature difference among the three locations is
− kPCM = −ktube
∂n ∂n (14) slight, thus it could assume that the temperature of the battery in the
Boundary condition at the interface between tube and the liquid middle can represent the entire battery temperature in following ex-
coolant is defined as follow: periments.
To further verify the accuracy of the prediction, cycling tests were
∂T ∂T
− ktube = −kc + hc (Tc − Ttube ) conducted on batteries wrapped with sufficient PCM at 30 °C. To obtain
∂n ∂n (15)
representative data, 3 continuous charge-discharge tests were carried
where kc is the thermal conductivity of coolant, hc is the heat transfer out. For each charge-discharge process, the battery was firstly charged
coefficient of the liquid coolant, Ttube is the surface temperature of in- from 2.75 V to 4.2 V at 0.5C, then a constant voltage charge process was
ternal channel. conducted with a cutoff current of 10 mA. Resting the 10 min, the
The boundary condition of the interface between the PCM and the battery was discharged to 2.75 V at 1.5C or 3C. After the same laying-

Fig. 2. (a) Maximum temperature responses of the battery pack with CPLS predicted with different gird resolutions during one cycle. (b) The finite element model for
the system with nonsymmetrical geometry.

5
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

T1

T2

T3

Fig. 3. (a) Comparison of measured and predicted battery temperatures during 3C discharge. (b) Comparison of the measured and predicted temperatures at 1.5C or
3C cycling.

aside time, the charge–discharge process was repeated twice. The pre- temperature uniformity than the case without liquid cooling, which
dicted and measured temperatures of the middle location on the battery indicated that the introduction of liquid cooling in CPLS alleviated the
surface were selected for comparison. Fig. 3(b) depicts the predicted heat accumulation of batteries more efficiently.
and measured data at 1.5C or 3C cycling. It shows that the predicted To further facilitate the comparison, the evolution of battery max-
battery temperature compared well with the measurements for both imum temperature (Tmax) and PCM liquid fraction for battery pack
1.5C and 3C cyclings. This indicates the robustness and accuracy of the under PCM without liquid cooling and CPLS during one cycle were
thermal model used in simulation. Besides, these results demonstrated extracted as shown in Fig. 5. Compared with the CPLS, it can be seen
that sufficient PCM could provide effective cooling to control the that the temperature rise of the PCM case was faster as the discharge
temperature rise of LIB during the continuous cycle. depth increasing, and the Tmax was significantly higher at the 1200 s.
Besides, the slope of liquid fraction in the PCM case was greater than
that in the CPLS case during the discharging. These results indicated
3. Results and discussions
that the latent heat in the PCM case was quickly consumed, and heat
accumulation occurred in the middle of the pack at the end of 3C dis-
3.1. Thermal characterization of CPLS
charge as shown in Fig. 4(b). The liquid cooling in CPLS case dissipated
some part of the heat transferring from PCM to the environment timely
In order to verify the influence on the heat dissipation performance
and promptly, which decreased the heat accumulation in the PCM. The
of CPLS, simulations were carried out for battery pack under three
thermal performance difference between the two cases was more ob-
thermal management modes, natural convection, PCM without liquid
vious during the charge process. For the PCM case, the PCM continued
cooling and CPLS. The battery pack was predicted under one cycle,
to absorb the heat released by the battery charging, and the heat
which was 3C discharge (from 0 s to 1200 s) then 0.5C charge (from
transferring speed to the air through PCM was very slow, so the liquid
1200 s to 8400 s) and no rest time. The Tamb was set as 30 °C, and the
fraction reached 100%. Conversely, the CPLS fully recovered the PCM
coolant flow rate was fixed at 0.2 m/s in CPLS.
latent heat at the end of charging.
Fig. 4 shows the temperature contours of the battery pack with three
It reveals that the CPLS not only could fully take advantage of the
thermal management modes at the end of 3C discharge (1200 s). The
PCM latent heat to control the temperature rise of the battery pack, but
temperature of the case under natural convection was above 80 °C,
also prolong the PCM liquefaction time by the liquid cooling during
much higher than the other cases. Its poor temperature performance
discharging. Besides, the CPLS also could recover the latent heat of the
demonstrated that the natural convection did not apply to the battery
PCM timely before the next discharging.
pack under abuse conditions. Both cases with PCM demonstrated an
obvious reduction of heat accumulation by taking advantage of the high
latent heat. Among them, the battery pack with CPLS exhibited better

Fig. 4. Temperature contours of the battery pack with three thermal management modes at 1200 s.

6
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

3C discharge (1200 s). Apparently, with the augment of the cell-to-cell

spacing from 2 mm to 5 mm, the temperature distribution of the battery
pack and PCM was remarkably improved, especially the temperature of
the center batteries gradually decreased and finally approached the
temperature of marginal batteries in the X direction. The reason is that
the smaller cell-to-cell spacing leads the PCM between adjacent cells
completely melt under the same heat dissipation of liquid cooling
during the middle stage of discharging. When further increasing the
thickness from s = 5 mm to s = 9 mm, the difference of the temperature
contour was no longer apparent. However, the thicker the cell-to-cell
spacing is, the heavier and more complex the structure is, as well as
lower energy density.
To further confirm the effect of cell-to-cell spacing on the thermal
performance, the Tmax, temperature difference (△T) of the battery pack
and PCM liquid fraction under different cell-to-cell spacing at 1200 s
Fig. 5. Evolution of Tmax and PCM liquid fraction for battery pack under PCM are displayed in Fig. 7(a). From s = 2 mm to 5 mm, the cooling effect
without liquid cooling and CPLS during one cycle. was remarkably improved that the Tmax and △T were controlled from
49.5 °C and 11.5 °C to 41.1 °C and 4 °C, respectively. It reveals that the
3.2. Effect of PCM geometric parameters small spacing induces thermal aggregation, which could lead to a high
local temperature rise, so unexpectedly high Tmax and △T are prone to
In this section, the effect of PCM geometric parameters was in- appear. When further increasing the spacing, the Tmax and △T continued
vestigated on the thermal performance of the battery pack, including to reduce at a lower rate. The liquid fraction of the PCM is a significant
cell-to-cell spacing (s) and cell-to-tube distance (d) as shown in index denoting the phase change progress and the effectiveness of the
Fig. 1(b). When a parameter varies, the other parameters are main- PCM itself. It can be seen that the PCM liquid fraction decreased as the
tained as their default values. The test conditions such as ambient PCM volume increasing as shown in Fig. 7(a). These results illustrate
temperature, coolant velocity and cycle rate were the same as Section that although thicker cell-to-cell spacing can improve thermal perfor-
3.1. mance, the increasing PCM volume inevitably leads to a decrease of the
PCM utilization and battery pack energy density.
Fig. 7(b) demonstrates the Tmax of battery pack and PCM liquid
3.2.1. Impact of cell-to-cell spacing fraction under different cell-to-cell spacing at the end of 0.5C charge
A wider cell-to-cell spacing could help in reducing the maximum (8400 s). The case with larger cell-to-cell spacing has a lower final Tmax,
temperature and temperature difference (△T) of the battery pack [14], which represents a lower starting temperature in the next cycle. The
but wider cell-to-cell spacing also results in a larger volume of PCM, ability of the liquid cooling system to recover the latent heat of PCM is
which would pose a greater challenge to the heat dissipation capability also a key indicator of the thermal performance of the coupling system.
of the liquid cooling system. To investigate the influence of cell-to-cell At 8400 s, the PCM liquid fraction was changed from 2.89% of
spacing on the temperature distribution of the battery pack, eight types s = 2 mm to 0.13% of s = 5 mm, and after s = 6 mm, the PCM liquid
of cell-to-cell spacing, which changed from 2 mm to 9 mm, was chosen fraction reduced to zero before the end of charging. Owing to the same
to study the thermal influence. Fig. 6 illustrates the temperature con- heat released by the discharging, the thicker cell-to-cell spacing is, the
tours of the battery pack with different cell-to-cell spacing at the end of

s s s s

s s s s
Fig. 6. Temperature contours of the battery pack with different cell-to-cell spacing at 1200 s.

7
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

Fig. 7. (a) Tmax, △T of the battery pack and PCM liquid fraction under different cell-to-cell spacing at 1200 s, (b) Tmax of battery pack and PCM liquid fraction at
8400 s.

longer liquid cooling could take away the heat absorbed by the PCM. absorbed by the PCM in the early stage of 0.5C charge. After that, the
PCM liquid fraction was reduced because the heat taken away by the
continuous liquid cooling is larger than the generated heat by 0.5C
3.2.2. Impact of cell-to-tube distance charge. From the cases of d = 2 mm to d = 5 mm, it reveals that al-
In the CPLS, the cell-to-tube distance (d) is another important factor. though the increased cell-to-tube distance leads to the increase of the
The shorter cell-to-tube distance brings the coolant tubes closer to the PCM volume, it also causes the excessive heat transfer distance from
batteries, and the volume of PCM is smaller in Y direction, which may batteries to tubes so that the negative influence of decreasing the
improve the cooling efficiency of the liquid cooling system. cooling effect of liquid cooling is more significant than the positive
In this section, the cell-to-tube distance was set as 1, 2, 3, 4 and effect of increasing the PCM volume on thermal performance of overall
5 mm, respectively, and the cell-to-cell was kept constant at 5 mm. system. Consequently, these results imply that a cell-to-tube distance of
Fig. 8(a) shows the evolution of Tmax in the battery pack under different 2 mm is enough in this study.
cell-to-tube distances during one cycle and Fig. 8(b) illustrates the △T of
battery pack at 1200 s. During the discharging, the Tmax and △T were
significantly suppressed from d = 1 mm to d = 2 mm, especially the △T 3.3. Thermal behavior of liquid cooling
for the case of d = 2 mm had been reduced to the desired range
(< 5 °C). Whereas the improvement of the battery thermal performance 3.3.1. Impact of channel number
was no longer obvious from d = 3 mm to 5 mm. During the 0.5C charge, Suitable channel number is essential for improving the efficiency of
it can be seen that the decreasing rate of Tmax in each case was inversely liquid cooling system during charge and discharge process. The cooling
proportional to the increase of cell-to-tube distance. It could be ex- performance of the CPLS is expected to increase as the channel number
plained that shorter cell-to-tube distance would result in a closer increases. In this section, the channel numbers from 3 to 6 were com-
cooling distance, which causes a better heat dissipation of liquid pared in terms of their performance in controlling the battery tem-
cooling under the same coolant flow velocity. perature and recovering the latent heat of PCM, and the coolant flows
Fig. 9(c) describes the liquid fraction evolution of the PCM from from the left side of the battery pack to the right side in all case as
1200 s to 8400 s. Apparently, the closer cell-to-tube distance led to the shown in Fig. 1(a). Besides, the flow velocity of the coolant was set as
higher PCM liquid fraction at the end of discharge, but for the charging, 0.2 m/s, and the initial temperature of the battery pack was 30 °C.
the liquid fraction of PCM in the case of d = 1 mm had dropped to 0% The evolution of Tmax of the battery pack during one cycle is shown
at the time of 8000 s, and other cases could not fully recover the PCM in Fig. 9, as well as the PCM liquid fraction. In all cases, there was no
latent heat. It is worth noting that the PCM liquid fraction continued to difference in the curves of the Tmax and liquid fraction before the last
increase from 1200 s to nearly 1350 s, which is mainly since the heat stage of the discharging. The more channel number was, the more Tmax
released by the batteries at the end of 3C discharge was gradually and liquid fraction decreased at the end of discharging. However, the

Fig. 8. (a) Evolution of Tmax in the battery pack under different cell-to-tube distances during one cycle, (b) △T of battery pack at 1200 s, (c) Evolution of liquid
fraction of the PCM during 0.5C charge process.

8
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

Fig. 9. Evolution of Tmax of the battery pack (a) and PCM liquid fraction (b) under different channel number during one cycle.

difference of Tmax and liquid fraction from 3 channels to 6 channels on the PCM solidification process during the charging. It can be ex-
were 1.5 °C and 6.2%, respectively. It reveals that most of the heat plained that conduction is the dominant mode of heat transfer during
generated by the battery discharging is absorbed by the PCM, and the the solidification process of PCM, and the presence of liquid cooling
remaining part is taken away by the liquid cooling, thereby no heat enhances significantly the heat transfer efficiency between PCM and
accumulation occurs in the battery pack even at 3 channels. Compared ambient environment due to liquid’s high specific heat capacity and
to the effect of PCM parameters, the channel number has little influence fluidity. However, the more the channels number is, the heavier and
on the thermal performance of the battery pack during 3C discharge. It more complex the structure of coupled system is, as well as higher
also indicates that if the latent heat could be taken full use, the main difficulty during the manufacturing process. Consequently, these results
heat material of CPLS is PCM, and the liquid cooling is prone to prolong imply that 5 channels are enough in this study.
the PCM liquefaction time during discharging.
For the charging, the more the channel number was, the more Tmax
and PCM liquid fraction decreased. What’s more, the downtrend of 3.3.2. Impact of coolant velocity
maximum temperature and PCM liquid fraction increased with the It was assumed that altering the coolant velocity would have in-
growth of channel number. To further facilitate the comparison, the fluences both on the temperature performance of the battery pack and
temperature contours of the battery pack with different channel num- the recovery of PCM latent heat. In this section, eight inlet flow velo-
bers at the end of charging are displayed in Fig. 10. For the case of 3 cities, including 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.8 and 1 m/s, were in-
channel, the temperature of the overall system was above 31 °C, which vestigated at the ambient temperature of 30 °C. The Tmax, Tmin and △T
is higher than the initial phase change temperature of PCM, so the PCM under different flow velocities at 1200 s are shown in Fig. 11(a). From
liquid fraction was maintained at 19% and there was still a serious heat 0.1 to 0.5 m/s, the Tmax and △T decreased 1.27 °C and 1.17 °C, respec-
accumulation on both sides of the battery pack. It implies that the heat tively. but the increase of inlet velocity has little effect on the Tmin. As
stored in the system cannot be fully dissipated in one cycle and will the coolant velocity further increasing, the thermal performance of
accumulate as the cycles continue, leading to a rapid increase in the battery pack was improved insignificantly. Similar to the study of the
battery temperature, which could spiral out of control. From 4 channels channels number, the effect of the flow velocity in the coupled system
to 6 channels, the problem of heat accumulation was gradually im- on the thermal performance of the battery is far less than the influence
proved in the coupled system. In the case of 5 channels, the temperature of the PCM parameters, it further demonstrates that if the sufficient
of the whole system was reduced to below 31 °C, and the PCM liquid PCM could absorb the heat generated by battery pack, liquid cooling
fraction was 0.35% as shown in Fig. 9(b). For the case of 6 channels, the may have a limited impact on the thermal performance of the battery
temperature contour has been maintained at 30.5 °C, and the latent heat pack.
of the PCM was already recovered at 8000 s as shown in Fig. 9(b). Fig. 11(b) presents the curve of PCM liquid fraction during charging
These results demonstrate the liquid cooling has a significant effect at different coolant velocities. The PCM liquid fractions were close in all
velocities at the end of discharging. It demonstrates that the PCM had

Fig. 10. Temperature contours of the battery pack with different channel numbers at 8400 s.

9
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

Fig. 11. (a) Tmax, Tmin and △T under different flow velocities at 1200 s. (b) Curve of PCM liquid fraction during charge progress.

absorbed the majority of heat generated by the battery, so that the 3.4. Experimental verification
change in flow velocity has little impact on the liquid fraction of the
PCM. For the charging, the PCM liquid fraction in the case of 0.1 m/s 3.4.1. Experimental setup
was the highest at 1200 s and 8400 s compared to other cases. Its liquid Based on the above numerical study, a prototype BTMS coupled
fraction of 2.3% at the end of charge means the latent heat of PCM had PCM and liquid cooling was designed and fabricated. The assembly
not been fully recovered. In the case of 0.2 m/s, the PCM liquid fraction process of the CPLS for battery pack is shown in Fig. 12. During this
was 0.35% at 8400 s. For the cases with other coolant velocities, the process, twenty-four 21,700 batteries (purchased from Lishen Battery
PCM liquid fraction all had been dropped to zero before 8400 s. Com- Joint-Stock Co., Ltd, Tianjin, China) were connected in the way of 6S4P
bined with the results in Fig. 12(a) and Fig. 12(b), it can be concluded in the insulating package, which was fabricated by 3D printer, and the
that as the coolant flow velocity increased, the thermal performance of material is polylactic acid (PLA, 0.025 W m−1 K−1). In this battery
the battery pack and the recovery of PCM latent heat increased. pack, the cell-to-cell spacing was 5 mm and the cell-to-tube distance
Nevertheless, faster coolant flow velocity (faster than 0.5 m/s in this was 2 mm. The aluminum tubes realized the S-shape by the bender, and
study) contributes little to the improvement in the heat exchange effi- the ends of tubes were connected with the adiabatic hose to realize the
ciency between the PCM and the coolant. Instead, it would incur higher function of transferring coolant. Then, PCM (purchased from Saimo
pump power consumption. For this reason, the coolant velocity of New Energy Technology Co., Ltd, Hubei, China) enwrapped the bat-
0.2 m/s was employed in the liquid cooling system. teries and aluminum tubes to form a thermal enhancement structure.
After using nickel belts to connect each battery, the baseplates were
wrapped by the fire retardant foam (20 mm Rubber Foam Flame Re-
tardant Sheet, 0.034 W m−1 K−1).
The schematic diagram of the experimental setup is shown in

Fig. 12. Photos of production process of the battery pack with the CPLS.

10
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

Fig. 13. Sketch of the experimental system and thermocouples locations in the battery pack.

pack, 3 continuous charge-discharge cycles by the battery cycler were

set as follows. In each cycle, the charge and discharge rates were 0.5C
and 1.5C, respectively, and the rest time was 10 min between each
process.

3.4.2. Experimental comparison under cycle test

Fig. 15 depicts the comparison of the Tmax and Tmin under the CPLS
between experiment data and predicted results. It can be seen that the
Tmax and Tmin of the experimental battery pack were maintained in the
range of 20 °C–30 °C, and the △T at the end of each discharge was 4.2 °C
on average. Because the Tmax did not reach the lowest phase transition
temperature of the PCM during the three cycles, the PCM acted as a
heat-conducting medium between the battery pack and tubes. There-
fore, the excellent temperature performance of battery pack was mainly
dependent on the liquid cooling system. It is worth noting that the
measured Tmax and Tmin of the battery pack during the experiment in-
Fig. 14. Comparison between the predicted and measured Tmax and Tmin under creased with the number of cycles compared with the previous cycle
the CPLS. while the predicted Tmax and Tmin of the battery pack were the same at
the end of each discharge. The main reason is that due to the limited
Fig. 13. A battery cycler (CT-4004-30V30A-NFA, Shenzhen Neware size of the water tank in the experimental system, the coolant with high
Electronics Co. Ltd, China) with an accuracy of ± 0.01% was used to temperature circulated into the water tank after heat exchange with the
charge and discharge the battery pack. The K-type thermocouples with battery pack, which caused the temperature of the coolant in the water
the diameter of 1 mm and 1 s response time were attached on thirteen tank to gradually increase from 20 °C to 22 °C, as shown in Fig. 14.
different locations as exhibited in Fig. 13. Among them, eight were used Hence, the measured temperature of the battery pack was also con-
to record the temperature for designated batteries, three are for the tinuously increased. For the simulation, the coolant inlet temperature
temperature at different locations of PCM, one recorded the tempera- was set to 20 °C, the temperature profiles of Tmax and Tmin during the
ture of water tank, and one recorded the Tamb. All thermocouples lead three cycles were consistent. These results show that the predicted re-
wires were connected to a National Instruments CompactDAQ-9213 sults agree well with the experimental data, which proves that the CPLS
data acquisition system whose data was stored in the PC. The coolant provided effective cooling to control the temperature rise of the battery
was water-ethylene glycol (1:1) mixture (glycol purchased from pack during cycles.
Dongsheng Chemical Co., Ltd, Jiangxi, China) stored in a water tank.
The flow rate of coolant was controlled by the hydraulic pump (pur- 3.5. Optimization of the controlling strategy for liquid cooling in CPLS
chased from Richuang rubber & plastic electric appliance Co., Ltd,
Guangzhou, China), and the turbine flowmeter (purchased from Previous experimental studies were carried out at low discharge rate
Shengjie Instrumentation Technology Co., Ltd, Hefei, China) was used and relatively a low ambient temperatures, and the coolant velocity in
to measure the inlet flow of the main tube. During the experimental CPLS was maintained at 0.4 m/s during all cycle. These not only led to
test, the Tamb was maintained at 20–22 °C. The temperature of the the PCM latent heat being completely unutilized but also brought un-
coolant was measured by a thermocouple inserted into the water tank. necessary energy consumption. For practical applications, the CPLS
The flow velocity of coolant for each tube was set to 0.4 m/s controlled needs to cope with severe conditions such as high discharge rate and
by a hydraulic pump and measured by a flowmeter. For the battery high/low temperature environment. In order to reduce unnecessary

11
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

energy consumption, it is also necessary to optimize the control of li- liquid cooling at the coolant velocity of 0.2 m/s throughout the cycle
quid cooling to ensure that the CPLS could adapt to different ambient made the PCM latent heat utilization low, and unnecessary energy
temperatures while maintaining good thermal performance. consumption occurred in the middle and late charge process.
At Tamb of 10 °C, the coolant inlet temperature was set to 20 °C to
3.5.1. Role of liquid cooling for CPLS under different ambient temperatures improve the temperature uniformity of the battery pack. As shown in
The role of liquid cooling for CPLS is different under different am- Fig. 15(c), the CPLS with the velocity of 0.2 m/s obtains a better
bient temperatures. Therefore, it is necessary to analyze the effect of the thermal performance of battery pack than that in the 0.001 m/s case
liquid cooling on CPLS at different ambient temperatures. To achieve a during the overall cycle, especially the temperature difference. Similar
better thermal management performance, the inlet coolant was actively to the situation in Tamb of 20 °C, the same problems that the utilization
heated/cooled for ambient temperatures of 10 °C and 40 °C. Fig. 15 of PCM latent heat is low and unnecessary energy consumption still
presents the Tmax, Tmin and PCM liquid fraction profiles of the battery exists + in the liquid cooling in the middle of the charge process.
pack under the velocity of 0.2 m/s and 0.001 m/s at four Tamb of 10 °C, At Tamb of 40 °C, PCM has completed phase change from solid phase
20 °C, 30 °C and 40 °C during 3C discharge and 0.5C charge process. The to liquid phase. The coolant inlet temperature was set to 25 °C. As
velocity of 0.001 m/s approximately simulates the case that hydraulic shown in Fig. 15(d), the Tmax in the case of 0.2 m/s was suppressed as
pump lay idle. 52.1 °C at 1200 s, which is below the upper limit of safe temperature
Similar to Section 3.1, better thermal performance and temperature 55 °C, while Tmax in the 0.001 m/s case reached 73.6 °C. However, the
uniformity were obtained as shown in Fig. 15(a) in the case with the △T in the case of 0.2 m/s reached 11.1 °C at 1200 s, higher than the
velocity of 0.2 m/s than 0.001 m/s during the overall cycle at 30 °C. suitable temperature difference. It can be explained that the batteries at
Besides, the PCM liquid fraction with the case of 0.2 m/s was reduced to the edge of battery pack are more prone to emerge the heat accumu-
zero at 8400 s, while the PCM liquid fraction in the case of 0.001 m/s lation due to the low heat transfer efficiency between PCM and the
was still more than 99% during the charging. The results show that environment, thereby it causes the large temperature difference com-
keeping the liquid cooling at normal coolant velocity during the cycle is pared with the batteries surrounded by more liquid tubes. During the
necessary for the heat dissipation performance of battery pack and the charging, the recovery of PCM latent heat in the 0.2 m/s was sig-
latent heat recovery capability of PCM at the ambient temperature of nificantly superior to the other one owing to the continuous thermal
30 °C. exchange between PCM and the coolant. It indicates that the liquid
At Tamb of 20 °C, the liquid cooling inlet temperature was set to cooling with the inlet coolant velocity of 0.2 m/s and inlet temperature
20 °C. Compared with the case of 0.001 m/s, better thermal perfor- of 25 °C greatly improves the heat dissipation of the CPLS in high
mance of battery pack was also found in the case of 2 m/s during the temperature environment. However, there still exited an unnecessary
overall cycle as shown in Fig. 15(b), but the temperature difference in energy consumption of the liquid cooling during the middle and late
the 0.2 m/s case was larger at 1200 s due to the lower Tmin. It should be stages of charge process.
noted that only 13% of PCM latent heat was utilized in the case of It can be concluded that the role of liquid cooling in the CPLS is
0.2 m/s during the cycle, while it was 63% in the case of 0.001 m/s. different under different ambient temperatures. When Tamb is lower
These results reveal that although the CPLS could maintain the tem- than the PCM phase change interval, the PCM is used as the main heat
perature performance of the battery pack at an appropriate range, the dissipating material during discharging. The liquid cooling system

Fig. 15. Tmax and Tmin profiles of the battery pack under different velocity at four Tamb of 30 °C (a), 20 °C (b), 10 °C (c) and 40 °C (d) during one cycle.

12
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

realizes the function of recovering the PCM latent heat and reducing the accumulation after absorbing the heat generated by the batteries due to
△T of battery pack during charging. When Tamb is equal to or higher the poor heat transfer efficiency between PCM and the environment.
than the PCM phase change interval, the heat dissipation of the battery Therefore, its ability to recover latent heat is also worse than other
pack depends on the liquid cooling, and the PCM only acts as the regions, and Point 5 could represent the region that has the worst heat
thermally conductive medium from batteries to coolant until the latent dissipation capability in the CPLS.
heat could be utilized. Besides, it is not necessary to maintain a coolant Based on the role of liquid cooling on the CPLS, it needs to adjust the
velocity of 0.2 m/s throughout the cycle considering the energy con- coolant velocity timely in the CPLS according to the temperature of
sumption, and heating or cooling the inlet coolant at abuse environ- PCM. Therefore, a segment function was used to control the coolant
ments (10 °C/40 °C) can obtain a better thermal performance of the velocity of the liquid cooling, which was triggered by Point 5. The
battery pack. Consequently, the liquid cooling could be optimized by an triggering mechanism is governed by the following equations:
effective controlling strategy in the CPLS.
⎧ 0.001 m / s, T5 ≤ 31 °C
vc = 0.2 m / s , 31 C< T5 ≤ 36 °C
°
⎨
3.5.2. Optimization of the controlling strategy in liquid cooling ⎩ 0.5 m / s, T5 > 36 °C (17)
In the CPLS, the controllable variables of the liquid cooling system
are velocity and inlet temperature of the coolant. The main role of li- According to the study in Section 4.1, the ambient temperature af-
quid cooling is to recover the PCM latent heat. When the PCM tem- fects significantly the thermal performance of the battery pack. At
perature is lower than its phase change interval, the liquid cooling normal ambient temperatures such as 20 °C and 30 °C, there is no need
system could keep the coolant velocity low. When the PCM temperature to adjust the coolant inlet temperature. Nevertheless, in the cold en-
is in its phase change interval, the coolant velocity should be appro- vironment such as 10 °C, it is necessary to appropriately increase the
priately increased to prolong the duration of PCM liquefaction process. inlet temperature to maintain the temperature uniformity of the battery
When the PCM temperature is higher than the phase change interval, it pack. In the high temperature environment such as 40 °C, it is also
is necessary to continue to increase the coolant velocity. While taking necessary to lower the coolant inlet temperature. It could effectively
away the heat of battery pack through the heat conduction of PCM, the suppress the Tmax of the battery pack and improve the ability of the
liquid cooling system also could accelerate the solidification process of liquid cooling system to recover the PCM latent heat. Therefore, it is
PCM to utilize the PCM latent heat as soon as possible. Therefore, the necessary to adjust the inlet temperature of the coolant timely in the
temperature of PCM could be used as the control indicator for the CPLS according to the ambient temperature. Another segment function
coolant velocity change of the liquid cooling system. It is necessary to is used to control the inlet temperature of the liquid cooling system and
determine an appropriate monitoring point representing the PCM the Tamb is used as the trigger point. The trigger mechanism is governed
temperature. The coolant velocity of the liquid cooling system is con- by the following equations:
trolled by the temperature of the monitoring point. Five monitoring °
points in the middle cross section of the battery pack were selected ⎧ 20 C, Tamb < 20 °C
based on the above criteria shown in Fig. 16. Tc = Tamb, 20 ° C≤ Tamb ≤ 31 °C
⎨ °
Fig. 17 presents the temperature rise curves of the monitoring points ⎩ 25 C, Tamb ≥ 31 °C (18)
for the cases of 0.2 m/s at different ambient temperatures. At Tamb of
30 °C, as shown in Fig. 17(a), the time for the temperature of five
monitoring points to the lowest phase change temperature of PCM 3.5.3. Continuous charge and discharge at different ambient temperatures
(31 °C) were almost the same. Point 4 was the first point whose tem- To investigate the effectiveness of the as-designed liquid cooling
perature approached to the highest phase change temperature of PCM control strategy at different ambient temperature, the performance of
(36 °C). However, the temperature of Point 5 (T5) took the highest the CPLS for battery pack was predicted under 3 continuous cycles at
temperature at 1200 s, and then it was also kept the highest during the Tamb of 10 °C, 20 °C, 30 °C and 40 °C. During the three cycles, the dis-
charge process. At Tamb of 20 °C and 10 °C, T5 was still higher than other charge and charge rate were 3C and 0.5C, respectively, and there is no
monitoring points during the charging. At Tamb of 40 °C, T5 was higher rest time.
than the other monitoring points throughout the cycle. The main reason At Tamb of 30 °C, the temperature evolution of Point 5 and the cor-
is that the Point 5 is in the centered-edge of the battery pack in Y di- responding change of coolant velocity are shown in Fig. 18(a), and the
rection. As mentioned in Section 3.2.1, this PCM region is prone to heat evolution of Tmax, Tmin and PCM liquid fraction are depicted in

Fig. 16. Schematic of middle cross section (a) of battery pack, and the selected monitoring points (b) in this middle cross section.

13
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

Fig. 17. Temperature rise curves of the each monitoring points based the cases of 0.2 m/s at different ambient temperatures.

Fig. 18(b). It can be seen that the temperature change of Point 5 trig- at the ambient temperature of 30 °C.
gered a change of coolant flow velocity in the CPLS timely. Therefore, At Tamb of 20 °C, the temperature evolution of Point 5 and the cor-
the battery pack achieved desirable temperature rising with the assis- responding velocity change of coolant are shown in Fig. 19(a), and the
tance of PCM latent heat and liquid cooling as shown in Fig. 18(b). As a evolution of Tmax, Tmin and PCM liquid fraction are depicted in
result, the Tmax, Tmin and PCM liquid fraction were 40.7 °C, 37 °C and Fig. 19(b). For each cycle, the temperature of Point 5 rose above 31 °C
80% at the end of 3C discharge during each cycle, which indicates that between the middle stage of 3C discharge and the early stage of 0.5C
the controlling strategy is suitable in the CPLS during cycling conditions charge. Hence the coolant velocity kept at 0.001 m/s except for this

Fig. 18. (a) Temperature evolution of Point 5 and the corresponding velocity change of coolant, and (b) evolution of Tmax, Tmin and PCM liquid fraction at Tamb of
30 °C.

14
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

Fig. 19. (a) Temperature evolution of Point 5 and the corresponding velocity change of coolant, and (b) evolution of Tmax, Tmin and PCM liquid fraction at Tamb of
20 °C.

period so that the power consumption of the liquid cooling was largely of coolant are shown in Fig. 21(a), and the evolution of Tmax, Tmin and
reduced. As shown in Fig. 19(b), the temperature performance of bat- PCM liquid fraction are depicted in Fig. 21(b). The coolant velocity of
tery pack was kept within the appropriate range. The PCM liquid liquid cooling was reduced from 0.2 m/s to 0.001 m/s according to the
fraction reached 19.2% at the end of 3C discharge in the first cycle, and controlling strategy in the middle stage of the first cycle. After that, the
then run up to 23.8% in the next two cycles. As a result, the controlling liquid cooling was maintained at a low power consumption until the
strategy in liquid cooling not only maintained the thermal performance next cycle. In the following two cycles, this phenomenon also occurred
of the battery pack at a good level, but also greatly reduce the power in the charging. It demonstrates that the CPLS with the controlling
consumption of the CPLS at Tamb of 20 °C. strategy still achieved the energy savings of the liquid cooling at high
At Tamb of 10 °C, the temperature evolution of Point 5 and the cor- temperature environment. As shown in Fig. 21(b), the Tmax and △T were
responding velocity change of coolant are shown in Fig. 20(a), and the reduced to 50.1 °C and 10.1 °C compared to the case with Tamb of 40 °C
evolution of Tmax, Tmin and PCM liquid fraction are depicted in in Section 4.1. In the last two cycles, the Tmax and Tmin were suppressed
Fig. 20(b). The coolant inlet temperature was heated to 20 °C due to to 39.4 °C and 35.2 °C due to the PCM latent heat participating in the
Tamb of 10 °C. While the temperature curve of Point 5 was similar to the heat dissipation of the battery pack, and the PCM liquid fraction was
case with Tamb of 20 °C, its duration time of keeping the coolant velocity maintained at 49.6% at each end of 3C discharge. The results show that
of 0.2 m/s was shorter than that in the case of 20 °C. The latent heat the CPLS with the controlling strategy in liquid cooling exhibits the
utilization rates of PCM were 10.2%, 17.5% and 18.1%, respectively, at ability of the rapid response in suppressing the temperature rise of the
each end of 3C discharge. It could be explained that a large amount of battery pack and recovering the PCM latent heat, therefore the thermal
heat generated by the battery pack was absorbed by the sensible heat of performance of battery pack was greatly improved at 40 °C, and the
PCM and the external ambient due to the obvious temperature differ- power consumption of the liquid cooling is decreasing.
ence, so the PCM liquid fraction was low at the end of first 3C dis- These results confirm that CPLS with effective controlling strategy
charge. It was found that the controlling strategy in liquid cooling also not only further improved the thermal performance of battery pack in
exhibited an acceptable thermal performance at Tamb of 10 °C. dealing with different ambient temperatures, but also significantly re-
The case with Tamb of 40 °C represents an extreme environment. duced the unnecessary power consumption of liquid cooling during the
According to the triggering mechanism described in Section 3.4.2, the charging and discharging.
inlet coolant would be cooled to 25 °C when Tamb is beyond 31 °C. The
temperature evolution of Point 5 and the corresponding velocity change

Fig. 20. (a) Temperature evolution of Point 5 and the corresponding velocity change of coolant, and (b) evolution of Tmax, Tmin and PCM liquid fraction at Tamb of
10 °C.

15
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

Fig. 21. (a) Temperature evolution of Point 5 and the corresponding velocity change of coolant, and (b) evolution of Tmax, Tmin and PCM liquid fraction at Tamb of
40 °C.

4. Conclusion the battery pack depends on the liquid cooling. The heat released by
the battery charging and discharging or stored in PCM itself is
A novel BTMS integrating PCM and liquid cooling was proposed for conducted to the liquid cooling system and taken away by the
the cylindrical NMC battery pack to address the aggressive thermal coolant flow to control the temperature performance of battery
challenge during continuous operation under different ambient tem- pack and recover the PCM latent heat timely.
peratures. Three different cooling methods including air, PCM without (4) For the controlling strategy, the heat dissipation of liquid cooling
liquid cooling and the coupled system for battery pack were in- should be strengthened as the PCM liquid fraction increasing.
vestigated by the simulation to verify the effectiveness of the CPLS. Therefore, the temperature of PCM region where the heat is easily
Then, factors such as cell-to-cell spacing and cell-to-tube distance in the accumulated and the heat dissipation is the worst could be used as
PCM, channel number and coolant velocity in the liquid cooling were the control indicator for the coolant velocity of the liquid cooling
discussed to investigate their influences on the design of CPLS. Based on system. Besides, the ambient temperature affects significantly the
the numerical study, the prototype CPLS were measured under con- thermal performance of the battery pack and the recovery of PCM
tinuous cycles to compare with the numerical predictions. Further, the latent heat, it is necessary to adjust the inlet temperature of the
controlling strategy for liquid cooling was designed and discussed in coolant timely in the CPLS according to the change of the ambient
detail, and 3 continuous charge-discharge cycles were used to study the temperature.
effect of CPLS with the controlling strategy under different ambient (5) The CPLS with the controlling strategy could further limit the
temperature. The following conclusions could be drawn: maximum temperature and improve the temperature uniformity
during the 3 continuous cycles at different ambient temperature,
(1) Increasing appropriately the cell-to-cell spacing could improve the while increasing the utilization of PCM latent heat and reducing the
thermal performance, after the cell-to-cell spacing exceeds 5 mm, power consumption of the liquid cooling.
the enhancement of the heat dissipation efficiency is insignificant in
this coupled system. Although the longer cell-to-tube distance be- Declaration of Competing Interest
yond 2 mm brings the better temperature of battery pack during
discharging, the excessive heat transfer distance from batteries to The authors declare that they have no known competing financial
tubes would contribute to the negative influence of decreasing the interests or personal relationships that could have appeared to influ-
cooling effect of liquid cooling especially in charging, which is more ence the work reported in this paper.
significant than the positive effect of increasing PCM volume.
(2) When the PCM could absorb sufficiently the heat released by the Acknowledgements
battery discharging, the liquid cooling has a limited effect on the
thermal performance of the battery pack, and its function is prone This work is financially supported by the National Natural Science
to prolong the PCM liquefaction time and further reduce the battery Foundation of China (No. 51604297), the Key Research and
temperature during the discharge process. However, liquid cooling Development Program of Shandong Province, China (No.
could accelerate the PCM solidification process significantly during 2018GSF120011), the Fundamental Research Funds for the Central
battery charging or rest. The growth of the channel number or the Universities (Nos. 18CX07005A and 19CX07006A).
coolant velocity brings the lower battery temperature and PCM li-
quid fraction at the end of charging. The design of the liquid cooling References
system should be carefully considered to satisfy the balance be-
tween the thermal performance and structural complexity in the [1] Liu H, Wei Z, He W, Zhao J. Thermal issues about Li-ion batteries and recent pro-
coupled system. gress in battery thermal management systems: a review. Energy Convers Manage
2017;150:304–30.
(3) The role of the liquid cooling in the CPLS is different under different [2] Wu W, Wang S, Wu W, Chen K, Hong S, Lai Y. A critical review of battery thermal
ambient temperatures. When the Tamb is lower than the PCM phase performance and liquid based battery thermal management. Energy Convers
change interval, the PCM is used as the main heat dissipating ma- Manage 2019;182:262–81.
[3] Saw LH, Ye Y, Tay AAO. Electro-thermal analysis and integration issues of lithium
terial during discharging. The liquid cooling system mostly realizes ion battery for electric vehicles. Appl Energy 2014;131:97–107.
the function of recovering the PCM latent heat and reducing △T of [4] Pesaran AA. Battery thermal models for hybrid vehicle simulations. J Power Sources
the battery pack during charging. When the Tamb is equal to or 2002;110:377–82.
[5] Seham S, Martin A-C. Analysis of cooling effectiveness and temperature uniformity
higher than the PCM phase change interval, the heat dissipation of

16
D. Kong, et al. Energy Conversion and Management 204 (2020) 112280

in a battery pack for cylindrical batteries. Energies 2017;10:1–17. lithium ion batteries combining phase change materials with forced-air cooling.
[6] Liu Y, Zhang J. Design a J-type air-based battery thermal management system Appl Energy 2015;148:403–9.
through surrogate-based optimization. Appl Energy 2019;252:113426. [30] Xie Y, Tang J, Shi S, Xing Y, Wu H, Hu Z, et al. Experimental and numerical in-
[7] Zhao J, Rao Z, Li Y. Thermal performance of mini-channel liquid cooled cylinder vestigation on integrated thermal management for lithium-ion battery pack with
based battery thermal management for cylindrical lithium-ion power battery. composite phase change materials. Energy Convers Manage 2017;154:562–75.
Energy Convers Manage 2015;103:157–65. [31] Jiang G, Huang J, Liu M, Cao M. Experiment and simulation of thermal manage-
[8] Qian Z, Li Y, Rao Z. Thermal performance of lithium-ion battery thermal manage- ment for a tube-shell Li-ion battery pack with composite phase change material.
ment system by using mini-channel cooling. Energy Convers Manage Appl Therm Eng 2017;120:1–9.
2016;126:622–31. [32] Qin P, Liao M, Zhang D, Liu Y, Sun J, Wang Q. Experimental and numerical study on
[9] Liang J, Gan Y, Li Y. Investigation on the thermal performance of a battery thermal a novel hybrid battery thermal management system integrated forced-air convec-
management system using heat pipe under different ambient temperatures. Energy tion and phase change material. Energy Convers Manage 2019;195:1371–81.
Convers Manage 2018;155:1–9. [33] Wu W, Yang X, Zhang G, Chen K, Wang S. Experimental investigation on the
[10] Wei A, Qu J, Qiu H, Wang C, Cao G. Heat transfer characteristics of plug-in oscil- thermal performance of heat pipe-assisted phase change material based battery
lating heat pipe with binary-fluid mixtures for electric vehicle battery thermal thermal management system. Energy Convers Manage 2017;138:486–92.
management. Int J Heat Mass Transf 2019;135:746–60. [34] Zheng Y, Shi Y, Huang Y. Optimisation with adiabatic interlayers for liquid-domi-
[11] Zou D, Liu X, He R, Zhu S, Bao J, Guo J, et al. Preparation of a novel composite nated cooling system on fast charging battery packs. Appl Therm Eng
phase change material (PCM) and its locally enhanced heat transfer for power 2019;147:636–46.
battery module. Energy Convers Manage 2019;180:1196–202. [35] Lopez CF, Jeevarajan JA, Mukherjee PP. Evaluation of combined active and passive
[12] Huang Y-H, Cheng W-L, Zhao R. Thermal management of Li-ion battery pack with thermal management strategies for lithium-ion batteries. J Electrochem Energy
the application of flexible form-stable composite phase change materials. Energy Convers Storage 2016;13:031007–31010.
Convers Manage 2019;182:9–20. [36] Song L, Zhang H, Yang C. Thermal analysis of conjugated cooling configurations
[13] Lv Y, Situ W, Yang X, Zhang G, Wang Z. A novel nanosilica-enhanced phase change using phase change material and liquid cooling techniques for a battery module. Int
material with anti-leakage and anti-volume-changes properties for battery thermal J Heat Mass Transf 2019;133:827–41.
management. Energy Convers Manage 2018;163:250–9. [37] Kshetrimayum KS, Yoon Y-G, Gye H-R, Lee C-J. Preventing heat propagation and
[14] Ling Z, Cao J, Zhang W, Zhang Z, Fang X, Gao X. Compact liquid cooling strategy thermal runaway in electric vehicle battery modules using integrated PCM and
with phase change materials for Li-ion batteries optimized using response surface micro-channel plate cooling system. Appl Therm Eng 2019;159:113797.
methodology. Appl Energy 2018;228:777–88. [38] Mat S, Al-Abidi AA, Sopian K, Sulaiman MY, Mohammad AT. Enhance heat transfer
[15] Azizi Y, Sadrameli SM. Thermal management of a LiFePO4 battery pack at high for PCM melting in triplex tube with internal–external fins. Energy Convers Manage
temperature environment using a composite of phase change materials and alu- 2013;74:223–36.
minum wire mesh plates. Energy Convers Manage 2016;128:294–302. [39] Mahmoud S, Tang A, Toh C, Al-Dadah R, Soo SL. Experimental investigation of
[16] Ling Z, Chen J, Fang X, Zhang Z, Xu T, Gao X, et al. Experimental and numerical inserts configurations and PCM type on the thermal performance of PCM based heat
investigation of the application of phase change materials in a simulative power sinks. Appl Energy 2013;112:1349–56.
batteries thermal management system. Appl Energy 2014;121:104–13. [40] Shi S, Xie Y, Li M, Yuan Y, Yu J, Wu H, et al. Non-steady experimental investigation
[17] Yan J, Li K, Chen H, Wang Q, Sun J. Experimental study on the application of phase on an integrated thermal management system for power battery with phase change
change material in the dynamic cycling of battery pack system. Energy Convers materials. Energy Convers Manage 2017;138:84–96.
Manage 2016;128:12–9. [41] Seo Young K, Keunhee L, Sunghoon P, Jaehoon K. Thermal design analysis and
[18] Li WQ, Qu ZG, He YL, Tao YB. Experimental study of a passive thermal management performance test of a 1kW thermoelectric battery cooler. Fourteenth intersociety
system for high-powered lithium ion batteries using porous metal foam saturated conference on thermal and thermomechanical phenomena in electronic systems.
with phase change materials. J Power Sources 2014;255:9–15. ITherm; 2014. p. 1417–21.
[19] Qu ZG, Li WQ, Tao WQ. Numerical model of the passive thermal management [42] Arora S. Selection of thermal management system for modular battery packs of
system for high-power lithium ion battery by using porous metal foam saturated electric vehicles: a review of existing and emerging technologies. J Power Sources
with phase change material. Int J Hydrogen Energy 2014;39:3904–13. 2018;400:621–40.
[20] Samimi F, Babapoor A, Azizi M, Karimi G. Thermal management analysis of a Li-ion [43] Quinn JB, Waldmann T, Richter K, Kasper M, Wohlfahrt-Mehrens M. Energy density
battery cell using phase change material loaded with carbon fibers. Energy of cylindrical li-ion cells: a comparison of commercial 18650 to the 21700 cells. J
2016;96:355–71. Electrochem Soc 2018;165:A3284–91.
[21] Babapoor A, Azizi M, Karimi G. Thermal management of a Li-ion battery using [44] Chen K, Li X. Accurate determination of battery discharge characteristics – a
carbon fiber-PCM composites. Appl Therm Eng 2015;82:281–90. comparison between two battery temperature control methods. J Power Sources
[22] Wu W, Zhang G, Ke X, Yang X, Wang Z, Liu C. Preparation and thermal conductivity 2014;247:961–6.
enhancement of composite phase change materials for electronic thermal man- [45] Cicconi P, Germani M, Landi D, Mengarelli M. Virtual Prototyping Approach to
agement. Energy Convers Manage 2015;101:278–84. Evaluate the Thermal Management of Li-Ion Batteries; 2014.
[23] Wu W, Wu W, Wang S. Thermal management optimization of a prismatic battery [46] Rao Z, Wang Q, Huang C. Investigation of the thermal performance of phase change
with shape-stabilized phase change material. Int J Heat Mass Transf material/mini-channel coupled battery thermal management system. Appl Energy
2018;121:967–77. 2016;164:659–69.
[24] Wang Z, Zhang H, Xia X. Experimental investigation on the thermal behavior of [47] Tiari S, Qiu S, Mahdavi M. Discharging process of a finned heat pipe–assisted
cylindrical battery with composite paraffin and fin structure. Int J Heat Mass Transf thermal energy storage system with high temperature phase change material.
2017;109:958–70. Energy Convers Manage 2016;118:426–37.
[25] Ping P, Peng R, Kong D, Chen G, Wen J. Investigation on thermal management [48] Tan FL, Hosseinizadeh SF, Khodadadi JM, Fan L. Experimental and computational
performance of PCM-fin structure for Li-ion battery module in high-temperature study of constrained melting of phase change materials (PCM) inside a spherical
environment. Energy Convers Manage 2018;176:131–46. capsule. Int J Heat Mass Transf 2009;52:3464–72.
[26] Huang Q, Li X, Zhang G, Zhang J, He F, Li Y. Experimental investigation of the [49] Zhao C, Cao W, Dong T, Jiang F. Thermal behavior study of discharging/charging
thermal performance of heat pipe assisted phase change material for battery cylindrical lithium-ion battery module cooled by channeled liquid flow. Int J Heat
thermal management system. Appl Therm Eng 2018;141:1092–100. Mass Transf 2018;120:751–62.
[27] Jiang ZY, Qu ZG. Lithium–ion battery thermal management using heat pipe and [50] Hosseinzadeh E, Genieser R, Worwood D, Barai A, Marco J, Jennings P. A sys-
phase change material during discharge–charge cycle: a comprehensive numerical tematic approach for electrochemical-thermal modelling of a large format lithium-
study. Appl Energy 2019;242:378–92. ion battery for electric vehicle application. J Power Sources 2018;382:77–94.
[28] Song W, Bai F, Chen M, Lin S, Feng Z, Li Y. Thermal management of standby battery [51] Dong T, Peng P, Jiang F. Numerical modeling and analysis of the thermal behavior
for outdoor base station based on the semiconductor thermoelectric device and of NCM lithium-ion batteries subjected to very high C-rate discharge/charge op-
phase change materials. Appl Therm Eng 2018;137:203–17. erations. Int J Heat Mass Transf 2018;117:261–72.
[29] Ling Z, Wang F, Fang X, Gao X, Zhang Z. A hybrid thermal management system for

17