Applied Thermal Engineering 101 (2016) 284–292

Contents lists available at ScienceDirect

Applied Thermal Engineering

j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / a p t h e r m e n g

Research Paper

Thermal management for high power lithium-ion battery by

minichannel aluminum tubes
Chuanjin Lan a, Jian Xu a, Yu Qiao b, Yanbao Ma a,*
a School of Engineering, University of California, Merced, CA 95343, USA
b Department of Structural Engineering, University of California San Diego, La Jolla, CA 92093, USA

H I G H L I G H T S

• A new design of minichannel cooling is developed for battery thermal management system.
• Parametric studies of minichannel cooling for a cell are conducted at different discharge rates.
• Minichannel cooling can maintain almost uniform temperature (Tdiff < 1 °C).
• Pumping power assumption is only about 5 milliwatt.

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

Article history: Lithium-ion batteries are widely used for battery electric (all-electric) vehicles (BEV) and hybrid elec-
Received 21 October 2015 tric vehicles (HEV) due to their high energy and power density. An battery thermal management system
Accepted 20 February 2016 (BTMS) is crucial for the performance, lifetime, and safety of lithium-ion batteries. In this paper, a novel
Available online 4 March 2016
design of BTMS based on aluminum minichannel tubes is developed and applied on a single prismatic
Li-ion cell under different discharge rates. Parametric studies are conducted to investigate the perfor-
Keywords:
mance of the BTMS using different flow rates and configurations. With minichannel cooling, the maximum
Electric vehicle
cell temperature at a discharge rate of 1C is less than 27.8 °C, and the temperature difference across the
Lithium-ion battery
Thermal management cell is less than 0.80 °C using flow rate at 0.20 L/min, at the expense of 8.69e-6 W pumping power. At
Minichannel cooling higher discharge rates, e.g., 1.5C and 2C, higher flow rates are required to maintain the same tempera-
ture rise and temperature difference. The flow rate needed is 0.8 L/min for 1.5C and 2.0 L/min for 2C, while
the required pumping power is 4.23e-4 W and 5.27e-3 W, respectively. The uniform temperature distri-
bution (<1 °C) inside the single cell and efficient pumping power demonstrate that the minichannel cooling
system provides a promising solution for the BTMS.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction influence the performance, cost, reliability, and safety of EVs. There-
fore, a battery thermal management system (BTMS) is crucial for
While the traditional transportation vehicle with an internal com- the EVs [4–10].
bustion engine contributes about 13% of annual world greenhouse During thermal management study for lithium-ion batteries, ad-
gas (GHG) emissions [1], battery electric vehicles (BEV) and hybrid equate knowledge of heat generation and thermal behavior inside
electric vehicles (HEV) are emerging replacements for traditional the battery is required to predict battery temperature. Studies have
vehicles to reduce GHG emissions [2]. EVs and HEVs are not only been done on the thermal modeling of batteries at different oper-
cleaner and more environmentally friendly, but are also more eco- ating conditions, i.e., at normal discharge rates and thermal abuses
nomically effective as the operating cost is reduced dramatically [3]. [8,11–20]. For normal operating conditions, Pesaran et al. [11] de-
Due to their high energy density, high power density, long life, and veloped a lumped capacitance battery thermal model to predict the
environmental friendliness, Li-ion batteries are widely used for BEVs thermal performance and impact of the temperature on vehicle level
and HEVs. However, poor performance at low temperature, degra- performance. Based on this lumped model, the thermal behavior
dation of electrodes at high temperature, and safety issues due to of modules and packs were evaluated. In another study by Chen et al.
thermal runaway associated with the Li-ion batteries will directly [12], a detailed three-dimensional thermal model was developed
to examine the thermal behavior of a lithium-ion battery, consid-
ering the layered-structure of the cell stacks, the case of a battery
* Corresponding author. Tel.: +1 (209) 228 4046; fax: 209 228 4046. pack, and the gap between both elements. Using this detailed model,
E-mail address: [email protected] (Y. Ma). the asymmetric temperature profile and the anomaly of temperature

http://dx.doi.org/10.1016/j.applthermaleng.2016.02.070
1359-4311/© 2016 Elsevier Ltd. All rights reserved.
 C. Lan et al. / Applied Thermal Engineering 101 (2016) 284–292 285

distribution on the surface can be simulated precisely. Besides these

modeling techniques for normal conditions, different models were
also developed for oven exposure testing. A one-dimensional pre-
dictive model for a 18,650 lithium-ion cell was developed by
Hatchard et al. using the kinetics of jelly-roll material decomposi-
tion reactions [13]. To consider geometrical features, a three-
dimensional model was developed by Kim et al. to determine the
local hot spot propagation through the cell [14]. Their results showed
that cell size greatly affected the thermal behavior of a cell due to
different heat transfer area per unit volume. Guo et al. proposed
another three-dimensional model to predict the thermal abuse per-
formances of lithium-ion batteries with high capacity, and analyzed
the temperature distribution under the abuse conditions [15]. The
model predictions were compared to experimental test results and
a good agreement was observed. For other abuse situations, thermal
modeling for the battery pack has also been developed in a one-
dimensional lumped model [16] and further in a three-dimensional
model [17]. Thermal runaway caused by nail penetration was ex-
perimentally studied by Doh et al. [18] and Chiu et al. [19], who also
modeled the complex reactions and mechanisms during the thermal
runaway.
With profound understanding of the thermal behavior of battery
cells at different operating conditions, different battery thermal man-
agement systems (BTMS), e.g., air cooling, liquid cooling, and phase
change material (PCM) cooling, have been applied to avoid the safety
issues from thermal aspect and to maintain the optimal operating
temperature. Forced air cooling with different structures has been
applied by manipulating the position of the air-inlet and the air-
outlet along with longitudinal or horizontal battery packs [21–26].
However, compared with the effectiveness of passive cooling by PCM,
the active forced air cooling is not a proper thermal management
system to keep the temperature of the cell in the desirable oper-
ating range without expending significant fan power [27]. Another
advantage of the PCM cooling is that the heat generated during the
discharge can be stored as latent heat in the PCM and transferred (e) Details of the minichannel geometry
back to the cell module during the relaxation. Therefore, the battery
temperature can be kept above the environment temperature, which
can increase the overall energy efficiency of the battery system Fig. 1. Different designs of minichannel cooling system: (a) one strip with four
minichannels; (b) one strip with eight minichannels; (c) two strips with four
[28–31]. Compared with the PCM cooling and air cooling, liquid minichannels each; and (d) four strips with four minichannels each (blue arrows
cooling systems can provide more effective heat transfer with dif- indicate the inflow direction and orange ones represent outflow direction); (e) details
ferent channel designs [32–34]. The cold-plate structure of the S-type of the minichannel geometry.
with guide plates was introduced by Zhang et al. to avoid the heat
concentration and increase the heat transfer area [32]. To enhance
the performance of the conventional channel with minimum pres- 55 Ampere-hours. The heat generation inside battery is assumed
sure penalty, an oblique minichannel liquid cold plate was developed uniform, but the thermal conductivity is anisotropic. Other proper-
by Jin et al. to cool down the EV batteries without over-designing ties of the battery are assumed homogenous. The aluminum
the cooling system [33]. In spite of these studies, a safer and more minichannel tubes wrap around three sides of the battery, as shown
cost-effective thermal management system is still required. in Fig. 1(a)–(d). The geometry of the tubes is shown in Fig. 1(e): the
In the present study, a new battery thermal management system height of channel (h) is 3 mm, and width (w) is 3 mm. The thickness
using aluminum minichannel tubes was designed. Different designs of aluminum between the outer surface and channel (δ) is 1 mm, and
of tube systems were parametrically studied at different discharg- the thickness between two neighbor channels will be twice δ [35].
ing rates. The numerical modeling of the battery and cooling system With this minichannel cooling system, the temperature distribution
design is introduced in the next section. To examine the perfor- across the battery is studied at different discharge rates. The desired
mance of this BTMS at different discharge rates, different designs temperature range for battery performance is between 15 °C and 35 °C
of tube systems are applied and the results are shown in the third [36]. If the battery temperature is below this range, battery performance
section. The conclusion on the applicability of the minichannels will be lowered due to poor ion transport. On the other hand, a tem-
cooling system is given in the last section. perature higher than that range will cause faster side reactions, leading
to higher dissipation rates of cyclable lithium and active materials.
2. Method
2.2. Governing equations
2.1. Physical problem
The energy conservation equation of the battery cell is given as
The computational domain consists of a prismatic geometry as follows:
the representative of one single battery cell, the minichannel cooling
system, and the fluid. The dimensions for the cell are 173 mm (z: height) ∂Tb ∂ ⎛ ∂Tb ⎞ ∂ ⎛ ∂Tb ⎞ ∂ ⎛ ∂Tb ⎞
ρbC b = ⎜ kx ⎟+ ky + ⎜ kz ⎟ + Qb (1)
by 168 mm (x: width) by 39 mm (y: depth), and the capacity is ∂t ∂x ⎝ ∂x ⎠ ∂y ⎜⎝ ∂y ⎟⎠ ∂z ⎝ ∂z ⎠
286 C. Lan et al. / Applied Thermal Engineering 101 (2016) 284–292

Table 1 neglected (water density is assumed constant). After the conver-

Material properties [13,19]. gence of flow fields, in the second step, the heat transfer in both
Parameter Aluminum Battery solid domain and fluid domain are solved together. Direct solver is
k (W m−1 K−1) 238 3.4 (cross-plane)/34.0 (in plane) used for time dependent simulation and the absolute tolerance is
ρ (kg m−3) 2700 1700 set as 0.001. Time step is set as one second since temperature field
Cp (J kg−1 K−1) 900 830 advances smoothly and monotonously. In this way, the simulation
can be much faster than solving coupled flow and heat equations.
Table 2
Heat generation rate of 55 Ah lithium-ion battery monomer at different discharge
rates [22]. 2.5. Model validation

Discharge rate 1C 1.5C 2C

In this study, free tetrahedral mesh is used for the mesh con-
Heat generation rate Qb (W) 7.60 15.60 23.89 struction. A mesh independence check is first performed. Most
elements are assigned to the fluid domain and the meshes at two
corners are specially refined. Sparse mesh is used for the heat trans-
where ρb and Cb are the average density and average specific heat fer in the solid battery domain. Two meshes with different numbers
of the battery, respectively; kx, kz are the in-plane thermal conduc- of elements are used for the mesh independence check. The ge-
tivity along the width and height direction, and ky is the cross- ometry in Fig. 1(a) is used for the validation case. Uniform inlet
plane thermal conductivity in the depth direction. These battery velocity is set as 0.023 m/s and the discharge rate is 1C (Qb = 7.60 W).
material parameters are given in Table 1 [13,19]. Qb is the uniform For the fine mesh, 418,192 elements are used in total, 303,492 of
volumetric heat generation rate across the whole battery. Its value which are used for fluid domain. For the coarse one, 202,778 ele-
at different discharge rates for the lumped model of a single cell is ments are used in total, 135,312 of which are used for fluid domain.
extracted from Xu’s work [22], as listed in Table 2. For the results, the maximum velocity obtained by fine mesh is
Liquid water is used as the cooling fluid inside the aluminum 0.0378 m/s and pressure drop is 34.54 Pa. The maximum temper-
minichannels. The energy conservation equation for water is: ature in the battery is 34.87 °C. Using the coarse mesh, the maximum
velocity obtained is 0.0374 m/s and pressure drop is 35.63 Pa. The
∂Tw  maximum temperature in the battery is 34.57 °C. The difference in
ρw C w + ∇ ⋅ ( ρwC w vTw ) = ∇ ⋅ (kw ∇Tw ) (2)
∂t the maximum velocity and difference in maximum temperature are
both 1%, while the pressure difference is about 3%. These results val-
where ρw, Cw and kw are the average density, specific heat and thermal idate the mesh independence, and the coarse mesh design is used
conductivity of water, respectively. Tw is the water temperature and in this study.

v is the velocity vector of water. The motion of incompressible After validating the mesh design, the numerical result of heat
liquid water is governed by the mass and momentum conserva- transfer and laminar flow also needs validation. As for the heat trans-
tion equations: fer simulation, energy conservation can be checked for the whole
 system. The input heat generation rate is 7.60 W from the battery
∇⋅v = 0 (3)
at a discharge rate of 1C, and the thermal power carried away by
 the water is 7.67 W, which is the only option for heat loss from the
⎡ ∂v  ⎤ 
ρw ⎢ + ( v ⋅∇ ) v ⎥ = −∇P + μ∇ 2v (4) system. The difference is 1%, thus validating the results for energy
⎣ ∂t ⎦ conservation. As for the laminar flows inside minichannels, the fric-
where P and μ are the static pressure and dynamic viscosity of water. tion loss at the end of the long section before reaching the bend
can be compared with analytical solutions. The long straight section
2.3. Initial and boundary conditions of the minichannel is long enough for the flow to be fully devel-
oped before reaching the bend. For fully developed laminar flow
The initial temperatures of the battery, the cooling channels, and inside rectangular channels with geometrical aspect ratio α (w/h),
the water are set at 27 °C. At flow inlets, the velocity and temper- the friction factor (f) can be predicted by [37]:
ature of water are assumed to be uniform and constant. For the flow
( f Re) fd = 24 (1 − 1.3553 α + 1.9467 α 2 − 1.7012 α 3 + 0.9564 α 4 − 0.2537 α 5 )
outlet, a constant zero pressure is specified and an outflow bound-
ary condition is used for energy equations. No slip boundary (5)
condition is used for the inside surfaces of minichannels. Thermal where Re is Reynolds number and f is Darcy friction factor, respec-
contact resistance between aluminum minichannels and the battery tively, given as: Re = ρwU 0 Dh μ , and f = ( ΔP ∗ Dh ) ( 21 ρwU 0 2 ∗ 4 L ). Dh is
is not taken into consideration in this study. Since the battery is the hydraulic diameter of the rectangular minichannel, U0 is the
usually located in a narrow and tight space, the air surrounding the average velocity at the channel inlet, μ is the dynamic viscosity of
battery can be quickly heated up without effective ventilation. There- water, ΔP is the pressure drop, and L is the length. For a square duct,
fore, in this study, convective cooling by surrounding air is assumed (f Re)fd is calculated to be 14.23 as α = 1. The result from simula-
to be negligible in the situation of having no fans installed. In other tions is f Re = 13.36, which differs from predictions by 6%.
words, a thermal insulation boundary condition is applied to all the In the end, the pressure drop of laminar flow through the whole
outer surfaces of the battery. As a result, the minichannel cooling minichannel with two 90° bends is calculated and the result is com-
is the only way for the heat dissipation of the battery. pared with theoretical predictions. To account for the entrance length
effect, the apparent friction factor is calculated for both the developing
2.4. Numerical method and fully developed laminar flow regions in the channel, given as
1
The commercial finite element software package COMSOL is used
to solve the conjugate heat transfer problem. Each simulation is ⎣⎢{
fapp Re = ⎡ 3.2 ( x + ) }
0.57 2
+ ( f Re)
2
fd
⎤2
⎦⎥
(6)
performed in two steps. In the first step, flow fields are solved using
a stationary iterative solver with relative tolerance set as 0.001, since In this equation, entrance length is defined as: x + = L ( Dh ⋅ Re) .
the effect of heat transfer on flow fields is quite small and can be More details for theoretical predictions are given in Liu’s work [38].
 C. Lan et al. / Applied Thermal Engineering 101 (2016) 284–292 287

As for the excess pressure loss due to 90° square bends, for slow of different flow rates, are studied here. The maximum tempera-
flow or negligibly small Re, the pressure loss is linear in the veloc- ture of the battery and the temperature difference between the
ity rather than quadratic, given as [39]: maximum temperature and minimum temperature of the battery,
denoted as Tmax and Tdiff, will be mainly discussed for different cases.
μU 0
ΔPexcess = K L (7) Uniformity index Tuni, which is used to quantify the temperature dif-
Dh ference inside the whole battery is also discussed. It is defined as:
where KL is the coefficient given for different bend geometries, which
is 4.2 for a 3D square corner [39]. To validate the pressure loss based Tuni = Tdiff Tavg (8)
on this equation, a small inlet velocity is applied as U0 = 0.00579 m/s.
where Tavg is the average temperature of the whole battery. Accord-
The total pressure loss calculated from simulation is 6.58 Pa, which
ing to the definitions, the smaller these values are, the better
is 6% different from the predicted value 7.03 Pa based on Eq. (5), (6)
performance it is for the thermal management system design.
and (7). All the above mentioned tests for energy conservation and
pressure drop comparison indicate that the numerical model used
in this study is valid. 3.1. Thermal management at a discharge rate of 1C (Qb = 7.60 W)

3. Results and discussions 3.1.1. Different designs for minichannel system

Four different minichannel cooling systems with a different
After validating the numerical model, the effects of different number of strips and a different number of minichannels are chosen
minichannel designs at different discharge rates, as well as the effects to study the effect of different geometric designs, as shown in

30 3
Tdiff ( C)
Tmax ( C)

29
2

28 1 by 4 1 by 4
1 by 8 1 1 by 8
2 by 4 2 by 4
27 4 by 4 4 by 4

0
0 1200 2400 3600 0 1200 2400 3600
t (s) t (s)

(a) (b)

(c) (d)
Fig. 2. (a) Temporal history of maximum temperature; (b) temporal history of temperature difference; (c) temperature distribution after 1 hour of discharging at 1C (Qb = 7.60 W)
with a flow rate at 0.05 L/min; (d) temperature distribution of the top one-eighth of the battery. Since no heat flux goes through the dashed plane in (c) due to symmetry
and the thermal insulated outer surfaces of the battery, the simulation can be performed on this much smaller geometry than the original whole shape, thus saving com-
putational resources and showing enlarged views.
288 C. Lan et al. / Applied Thermal Engineering 101 (2016) 284–292

Fig. 3. Design with different flow directions. Blue arrows indicate the flow inlet and orange ones represent flow outlet.

Fig. 1(a)–(d). In design (a), one strip with four minichannels is The temperature differences Tdiff across the whole battery are
wrapped in the middle of the battery; in design (b), one strip with shown in Fig. 2(b). It can be seen that the temporal change of Tdiff is
eight minichannels is used; in design (c), two strips with four similar to Tmax for each corresponding case. This is because near the
minichannels each are located in the upper and lower parts of the minichannels inlet, the local battery temperature is always close to
battery, respectively; in design (d), four strips with four minichannels the inlet water temperature since the battery and the inlet water
each are distributed uniformly. All the flow inlets are set on one side are only separated by a thin layer of aluminum (Fig. 2(c)). However,
of the battery and flow outlets on the other side. near the minichannels outlet, as the water is warmed up, the heat
The maximum temperatures of the battery Tmax for the four dif- dissipation efficiency for the battery is reduced and the maximum
ferent designs are shown in Fig. 2(a). The discharge rate is fixed at battery temperature occurs. Since the local battery temperature near
1C (Qb = 7.60 W), and the flow rate is 0.05 L/min. Since the total the minichannels inlet is the minimum battery temperature and
number of minichannels is different for designs (a)–(d), the inlet remains almost unchanged after 1 hour of discharging, the varia-
flow velocity varies accordingly for different designs due to differ- tion of temperature difference depends mainly on the maximum
ent total cross-section areas. Design (a), which uses one strip with temperature variation. Tdiff, Tmax and Tuni after 1 hour of discharging
four minichannels, shows the highest temperature rise from 27.0 °C at 1C are shown in Table 3, as well as pressure drop and pumping
to 30.08 °C, after 1 hour of discharging. Compared to design (a), Tmax power required. The pumping power for design (a) is more than 5
for design (b) (one strip with eight minichannels) changes from times larger than design (b), and more than 23 times larger than
27.0 °C to 29.75 °C. This indicates that a strip with more minichannels design (d). From the comparison in Table 3, design (d) requires
can reduce the temperature rise, though the flow rate is lower. For minimum pumping power and obtains the best thermal manage-
design (c), which has the same total number of minichannels as ment performance (minimum Tdiff, Tmax and Tuni), while the only
design (b), Tmax increases from 27.0 °C to 29.35 °C. This tempera- shortcoming is the cost of the channel materials. Therefore, design
ture increase is smaller than that of design (b). From the comparison (d) will be used for all the remaining subsections.
between design (b) and design (c), it can be seen that a wider dis-
tribution of minichannels has better thermal management 3.1.2. Different flow directions for performance enhancement
performance than concentrating all minichannels at one place. For Based on design (d) in Fig. 1(d), five designs with different flow
design (d), which has four strips evenly distributed, Tmax has the directions are compared to study the effect of flow direction, as
minimum increase from 27.0 °C to 29.20 °C. These results show that shown in Fig. 3. The blue arrows indicate the flow inlets, and the
the design using more minichannels and a wider distribution has orange arrows indicate the flow outlets. The design “Direction 1”
the minimum temperature increase. This is reasonable since the heat has been used in Subsection 3.1.1.
can be more easily dissipated by minichannels when the contact The temporal variations of Tmax and Tdiff are shown in Fig. 4. Similar
areas are larger and more distributed. trends for Tmax and Tdiff are observed for all cases. Among the five

29
2
Tmax ( C)

Tdiff ( C)

28 Direction 1 Direction 1
Direction 2 1 Direction 2
Direction 3 Direction 3
Direction 4 Direction 4
Direction 5 Direction 5
27
0
0 1200 2400 3600 0 1200 2400 3600
t (s) t (s)

(a) (b)
Fig. 4. Temporal history of (a) maximum temperature and (b) temperature difference using different flow directions, at a discharge rate of 1C (Qb = 7.60 W).
 C. Lan et al. / Applied Thermal Engineering 101 (2016) 284–292 289

Table 3
Comparison of pressure drop, pumping power, temperature difference, maximum
temperature and uniformity index for different designs at a discharge rate of 1C
(Qb = 7.60 W) with a flow rate at 0.05 L/min.
29
Case ΔP (Pa) Pumping Power (W) Tdiff (°C) Tmax (°C) Tuni

1 by 4 41.8 8.71e-6 3.04 30.08 0.102

1 by 8 15.8 1.66e-6 2.72 29.75 0.093

Tmax ( C)
2 by 4 15.8 1.66e-6 2.30 29.35 0.079
4 by 4 7.03 3.66e-7 2.18 29.20 0.076
28

designs, case ‘Direction 1’ obtains the minimum Tmax and Tdiff. Con- 0.05 L/min
sidering the complexity and cost of the inlet manifold, the case 0.10 L/min
‘Direction 1’ is also the best choice. The worst thermal perfor- 27 0.15 L/min
mance occurs for case ‘Direction 2’, which has alternative flow 0.20 L/min
direction. For the other three cases, variations of Tmax and Tdiff show
close performance. Based on these results, the design ‘Direction 1’ 0 1200 2400 3600
will be used for all the remaining subsections.
t (s)

3.1.3. Different flow rates for performance enhancement (a)

Since the pressure drop and the required pumping power are
quite small as observed in Table 3, the flow rate can be increased
to further improve the thermal management performance. In this
study, the flow is kept in laminar regime. 2
The maximum temperatures and temperature differences of the
battery at different flow rates are shown in Fig. 5. It can be seen that
Tmax and Tdiff get lower as the flow rate increases. Using flow rates
as 0.15 L/min and 0.20 L/min, Tmax becomes stable in no more than T diff( C)
1200 s. From Fig. 5(c), it can be seen that the high temperature area
of the battery is close to the minichannel outlet, since the water 1
near the outlet is already heated up and the water temperature is
close to the maximum temperature of the battery. When flow rate
is 0.15 L/min, Tmax is kept at 28 °C and Tdiff is less than 1 °C. Once the 0.05 L/min
flow rate is increased to 0.20 L/min, Tmax gets to 27.81 °C and Tdiff 0.10 L/min
becomes 0.80 °C after 1 hour of discharging at 1C (Qb = 7.60 W). As 0.15 L/min
the flow rate increases, the uniformity index Tuni also reduces as Tmax 0.20 L/min
and Tdiff, as shown in Table 4. These results demonstrate that the 0
0 1200 2400 3600
minichannel cooling system can maintain Tdiff to be less than 1 °C, t (s)
and it only requires 4.28e-6 W and 8.69e-6 W pumping power for
0.15 L/min and 0.20 L/min flow rate, respectively. (b)
3.2. Thermal management at a discharge rate of 1.5C (Qb = 15.60 W)

In this subsection, the thermal performance for a higher dis-

charge rate at 1.5C (Qb = 15.60 W) is studied. Based on the results
in Subsection 3.1, flow rates higher than 0.20 L/min are used. The
results are shown in Fig. 6. Both Tmax and Tdiff decrease as the flow
rate increases; however, the effect of flow rate on cooling
performance gets smaller as flow rate increases. From Fig. 6(c), it
can be seen that the high temperature area of the battery is near
the center of the battery in cross-plane (y) direction, slightly closer
to the minichannel outlet. Due to the high flow rates used, the outlet
temperature of the minichannel is much lower than the maximum
temperature of the battery. When the flow rate increases from
0.20 L/min to 0.40 L/min, the Tmax drops from 28.66 °C to 28.16 °C

Table 4
(c)
Comparison of pressure drop, pumping power, temperature difference, maximum
temperature and uniformity index using different flow rates at a discharge rate of 1C. Fig. 5. Temporal history of (a) maximum temperature and (b) temperature differ-
ence using different flow rates, at a discharge rate of 1C (Qb = 7.60 W). (c) Temperature
Flow rate ΔP (Pa) Pumping Tdiff (°C) Tmax (°C) Tuni distribution after 1 hour of discharging at 1C (Qb = 7.60 W), using a flow rate at
(L/min) power (W) 0.20 L/min.
0.05 7.03 3.66e-7 2.18 29.20 0.076
0.10 15.8 1.64e-6 1.31 28.32 0.047
0.15 27.4 4.28e-6 0.97 27.98 0.035
0.20 41.8 8.69e-6 0.80 27.81 0.029
290 C. Lan et al. / Applied Thermal Engineering 101 (2016) 284–292

29

1.5

Tdiff ( C)
Tmax ( C)

28
1

0.20 L/min 0.20 L/min

0.40 L/min 0.5 0.40 L/min
0.60 L/min 0.60 L/min
27 0.80 L/min 0.80 L/min
1.00 L/min 1.00 L/min
0
0 800 1600 2400 0 800 1600 2400
t (s) t (s)

(a) (b)

(c)
Fig. 6. Temporal history of (a) maximum temperature and (b) temperature difference using different flow rates, at a discharge rate of 1.5C (Qb = 15.60 W). (c) Temperature
distribution after 2400s of discharging at 1.5C (Qb = 15.60 W), using a flow rate at 1.00 L/min.

at t = 2400 s. When the flow rate is further increased to 0.6 L/min, 3.3. Thermal management at a discharge rate of 2C (Qb = 23.89 W)
Tmax reduces to 28.01 °C. Finally, when flow rate increases from
0.80 L/min to 1.00 L/min, Tmax only reduces 0.04 °C and Tuni only A harsher situation is studied in this subsection as the discharge
reduces 3% while the required pumping power increases nearly twice rate increases to 2C (Qb = 23.89 W). According to the previous dis-
as much as shown in Table 5. There is no more potential benefit by cussion for a discharge rate of 1.5C (Qb = 15.60 W), flow rates at no
increasing the flow rate further, since the heat conduction rate inside less than 1.00 L/min are used for the cooling system performance anal-
the battery is limited by the material property, and the maximum ysis at a discharge rate of 2C.
temperature cannot be reduced any more by the minichannel cooling Results for Tmax and Tdiff using different flow rates at a discharge
system. This will be explained in more detail in Subsection 3.3. Con- rate of 2C (Qb = 23.89 W) are shown in Fig. 7, respectively. At a flow
cerning the pumping power cost, 0.80 L/min is the best option for rate of 1.00 L/min, Tmax reaches to 28.38 °C. As flow rate increases
cooling performance at a discharge rate of 1.5C (Qb = 15.60 W). to 2.00 L/min, Tmax reduces to 28.27 °C, but the required pumping
power becomes seven times larger (Table 6). When the flow rate
increases further to 4.00 L/min, there is not much improvement for
cooling performance as Tmax, Tdiff and Tuni remain almost unchanged.
Table 5
Comparison of pressure drop, pumping power, temperature difference, maximum
However, the pumping power is doubled to 3.61e-2 W. Concern-
temperature and uniformity index using different flow rates at a discharge rate of ing the pumping power cost, 2.00 L/min is the best option for cooling
1.5C (Qb = 15.60 W). performance at a discharge rate of 2C (Qb = 23.89 W).
Flow rate ΔP (Pa) Pumping Tdiff (°C) Tmax (°C) Tuni To further explore the cooling system improvement using higher
(L/min) power (W) flow rate, the temperature contour of the battery and minichannels
0.20 41.8 8.74e-6 1.65 28.66 0.058
using a flow rate at 4.00 L/min is shown in Fig. 7(c). The tempera-
0.40 160 6.66e-5 1.15 28.16 0.041 ture of the minichannels is quite low at both the inlet and the outlet,
0.60 314 1.96e-4 1.00 28.01 0.036 due to the large flow rate. The maximum battery temperature is at
0.80 508 4.23e-4 0.94 27.94 0.034 the center of the battery in the y (depth) direction, due to the low
1.00 743 7.74e-4 0.90 27.90 0.033
cross-plane thermal conductivity of the battery. Though the heat
 C. Lan et al. / Applied Thermal Engineering 101 (2016) 284–292 291

29 2

Tmax ( C)

Tdiff ( C)
28
1

1.00 L/min 1.00 L/min

2.00 L/min 2.00 L/min
27 3.00 L/min 3.00 L/min
4.00 L/min 4.00 L/min
0
0 400 800 1200 1600 0 400 800 1200 1600
t (s) t (s)
(a) (b)

(c)
Fig. 7. Temporal history of (a) maximum temperature and (b) temperature difference using different higher flow rates, at a discharge rate of 2C (Qb = 23.89 W). (c) Tem-
perature distribution of battery and minichannels after 1800 s of discharging at 2C (Qb = 23.89 W), using a flow rate at 4.00 L/min.

can be quickly taken away by the large flow rate, the heat transfer 4. Conclusion
inside the battery is constraint by the (cross-plane) thermal con-
ductivity of the material and the geometrical thickness in cross- Parametric studies were carried out to demonstrate the feasi-
plane (y) direction. Increasing the flow rates for the minichannels bility of adopting the minichannels cooling system for high-capacity
is not an effective solution for this high discharge rate. However, lithium-ion battery thermal management. Effects of different geo-
reducing the thickness of the cross-plane (y) direction should be metric designs, flow directions, and flow rates, were studied through
an applicable solution. Though there is not much room to improve parametric study by monitoring maximum temperature rise of the
the cooling performance, maintaining Tmax as low as 28.38 °C and battery Tmax, temperature difference across the battery Tdiff and tem-
Tdiff as low as 1.38 °C is still acceptable for thermal management perature uniformity Tuni. Under the same total flow rate inside the
performance. cooling system, the more minichannels are used, the better cooling
performance can be achieved, though the inlet flow speed is reduced
accordingly. Additionally, with the same total number of
Table 6 minichannels, the case using distributed distribution of minichannels
Comparison of pressure drop, pumping power, temperature difference, maximum shows better cooling performance than the case using concen-
temperature and uniformity index using different flow rates at a discharge rate of trated distribution. Moreover, the effects of the flow rate and flow
2C (Qb = 23.89 W).
direction were studied. Results indicate that the best performance
Flow rate (L/min) ΔP (Pa) Power (W) Tdiff (°C) Tmax (°C) Tuni is achieved when all the flow inlets are aligned along one side of
1.00 743 7.74e-4 1.38 28.38 0.049 the battery, instead of alternating inlets and outlets.
2.00 2.53k 5.27e-3 1.26 28.27 0.045 In addition, this study shows that at a discharge rate of 1C, using
3.00 5.19k 1.63e-2 1.23 28.23 0.044 a flow rate of 0.20 L/min, Tmax is well controlled at 27.81 °C and Tdiff
4.00 8.66k 3.61e-2 1.21 28.21 0.044
is 0.80 °C after 1 hour of discharging, with only 8.69e-6 W pumping
292 C. Lan et al. / Applied Thermal Engineering 101 (2016) 284–292

power required. At a discharge rate of 1.5C, as flow rate increases [16] K. Smith, C.Y. Wang, Power and thermal characterization of a lithium-ion battery
pack for hybrid-electric vehicles, J. Power Sources 160 (2006) 662–673.
to 0.80 L/min, Tmax is 27.94 °C. Because only slight improvement on
[17] H. Sun, X. Wang, B. Tossan, R. Dixon, Three-dimensional thermal modeling of
cooling performance can be achieved using flow rates higher than a lithium-ion battery pack, J. Power Sources 206 (2012) 349–356.
0.80 L/min, 0.80 L/min is adopted as the best solution for cooling [18] C.H. Doh, D.H. Kim, H.S. Kim, H.M. Shin, Y.D. Jeong, S.I. Moon, et al., Thermal
performance at a discharge rate of 1.5C, considering the pumping and electrochemical behaviour of C/Li x CoO 2 cell during safety test, J. Power
Sources 175 (2008) 881–885.
power cost for higher flow rates. When the discharge rate in- [19] K.C. Chiu, C.H. Lin, S.F. Yeh, Y.H. Lin, K.C. Chen, An electrochemical modeling
creases to 2C, at a flow rate of 1.00 L/min, Tmax reaches to 28.27 °C of lithium-ion battery nail penetration, J. Power Sources 251 (2014) 254–263.
and Tdiff is 1.26 °C, with 5.27e-3 W pumping power required. Further [20] U.S. Kim, J. Yi, C.B. Shin, T. Han, S. Park, Modelling the thermal behaviour of a
lithium-ion battery during charge, J. Power Sources 196 (2011) 5115–5121.
increase in flow rate cannot reduce the temperature near the battery [21] M.S. Wu, K.H. Liu, Y.Y. Wang, C.C. Wan, Heat dissipation design for lithium-ion
center in cross-plane (y) direction, but can only reduce the water batteries, J. Power Sources 109 (2002) 160–166.
temperature at the outlet. [22] X. Xu, R. He, Review on the heat dissipation performance of battery pack with
different structures and operation conditions, Renew. Sustain. Energy Rev. 29
In conclusion, we have demonstrated that the minichannels (2014) 301–315.
cooling system can be applied for the battery thermal manage- [23] R. Mahamud, C. Park, Reciprocating air flow for Li-ion battery thermal
ment. This system can reduce both the maximum temperature rise management to improve temperature uniformity, J. Power Sources 196 (2011)
5685–5696.
and temperature difference across the whole battery, at little expense [24] T. Wang, K.J. Tseng, J. Zhao, Z. Wei, Thermal investigation of lithium-ion battery
of pressure drop and pumping power. module with different cell arrangement structures and forced air-cooling
strategies, Appl. Energy 134 (2014) 229–238.
[25] S.K. Mohammadian, Y. Zhang, Thermal management optimization of an
Acknowledgement air-cooled Li-ion battery module using pin-fin heat sinks for hybrid electric
vehicles, J. Power Sources 273 (2015) 431–439.
[26] L. Fan, J.M. Khodadadi, A.A. Pesaran, A parametric study on thermal management
This work was funded by the U.S. Department of Energy under of an air-cooled lithium-ion battery module for plug-in hybrid electric vehicles,
the ARPA-E program (Award No. DEAR0000396). J. Power Sources 238 (2013) 301–312.
[27] R. Sabbah, R. Kizilel, J.R. Selman, S. Al-Hallaj, Active (air-cooled) vs. passive (phase
change material) thermal management of high power lithium-ion packs:
References limitation of temperature rise and uniformity of temperature distribution,
J. Power Sources 182 (2008) 630–638.
[28] S. Al Hallaj, J. Selman, A novel thermal management system for electric vehicle
[1] EPA, Global greenhouse gas emissions data. <http://www.epa.gov/ batteries using phase-change material, J. Electrochem. Soc. 147 (2000) 3231–
climatechange/ghgemissions/global.html>, 2015 (accessed 07.09.15). 3236.
[2] J. Tollefson, Charging up the future, Nature 456 (2008) 436–440. [29] S. Al-Hallaj, J. Selman, Thermal modeling of secondary lithium batteries for
[3] S.F. Tie, C.W. Tan, A review of energy sources and energy management system electric vehicle/hybrid electric vehicle applications, J. Power Sources 110 (2002)
in electric vehicles, Renew. Sustain. Energy Rev. 20 (2013) 82–102. 341–348.
[4] T.M. Bandhauer, S. Garimella, T.F. Fuller, A critical review of thermal issues in [30] R. Kizilel, R. Sabbah, J.R. Selman, S. Al-Hallaj, An alternative cooling system to
lithium-ion batteries, J. Electrochem. Soc. 158 (2011) R1–R25. enhance the safety of Li-ion battery packs, J. Power Sources 194 (2009)
[5] L.G. Lu, X. Han, J. Li, J. Hua, M. Ouyang, A review on the key issues for lithium-ion 1105–1112.
battery management in electric vehicles, J. Power Sources 226 (2013) 272– [31] Z. Rao, S. Wang, A review of power battery thermal energy management, Renew.
288. Sustain. Energy Rev. 15 (2011) 4554–4571.
[6] Z.H. Rao, S.F. Wang, A review of power battery thermal energy management, [32] Y. Zhang, X.L. Yu, Q.K. Feng, R.T. Zhang, Thermal performance study of integrated
Renew. Sustain. Energy Rev. 15 (2011) 4554–4571. cold plate with power module, Appl. Therm. Eng. 29 (2009) 3568–3573.
[7] Q. Wang, P. Ping, X. Zhao, G. Chu, J. Sun, C. Chen, Thermal runaway caused fire [33] L. Jin, P.S. Lee, X.X. Kong, Y. Fan, S.K. Chou, Ultra-thin minichannel LCP for EV
and explosion of lithium ion battery, J. Power Sources 208 (2012) 210–224. battery thermal management, Appl. Energy 113 (2014) 1786–1794.
[8] R. Spotnitz, J. Franklin, Abuse behavior of high-power, lithium-ion cells, J. Power [34] Y. Huo, Z. Rao, X. Liu, J. Zhao, Investigation of power battery thermal
Sources 113 (2003) 81–100. management by using mini-channel cold plate, Energy Convers. Manag. 89
[9] P. Balakrishnan, R. Ramesh, T.P. Kumar, Safety mechanisms in lithium-ion (2015) 387–395.
batteries, J. Power Sources 155 (2006) 401–414. [35] X. Xie, Z.J. Liu, Y.L. He, W.Q. Tao, Numerical study of laminar heat transfer and
[10] L. Lu, X. Han, J. Li, J. Hua, M. Ouyang, A review on the key issues for lithium-ion pressure drop characteristics in a water-cooled minichannel heat sink, Appl.
battery management in electric vehicles, J. Power Sources 226 (2013) 272–288. Therm. Eng. 29 (2009) 64–74.
[11] A.A. Pesaran, Battery thermal models for hybrid vehicle simulations, J. Power [36] J.P. Rugh, A. Pesaran, K. Smith, Electric vehicle battery thermal issues and
Sources 110 (2002) 377–382. thermal management techniques, presented at the SAE 2011 Alternative
[12] S. Chen, C.C. Wan, Y.Y. Wang, Thermal analysis of lithium-ion batteries, J. Power Refrigerant and System Efficiency Syposium, 2011.
Sources 140 (2005) 111–124. [37] R.K. Shah, A.L. London, Laminar Flow Forced Convection in Ducts: A Source Book
[13] T. Hatchard, D.D. MacNeil, A. Basu, J.R. Dahn, Thermal model of cylindrical and for Compact Heat Exchanger Analytical Data, vol. 1, Cambridge, MA, Academic
prismatic lithium-ion cells, J. Electrochem. Soc. 148 (2001) A755–A761. Press, 1978.
[14] G.H. Kim, A. Pesaran, R. Spotnitz, A three-dimensional thermal abuse model [38] D. Liu, S.V. Garimella, Investigation of liquid flow in microchannels,
for lithium-ion cells, J. Power Sources 170 (2007) 476–489. J. Thermophys. Heat Transfer 18 (2004) 65–72.
[15] G. Guo, B. Long, B. Cheng, S. Zhou, P. Xu, B. Cao, Three-dimensional thermal [39] M.V. Koch, K.M. VandenBussche, R.W. Chrisman, Micro Instrumentation: For
finite element modeling of lithium-ion battery in thermal abuse application, High Throughput Experimentation and Process Intensification – A Tool for PAT,
J. Power Sources 195 (2010) 2393–2398. Hoboken, NJ, John Wiley & Sons, 2007.