Professional Documents
Culture Documents
Abstract: The performance of lithium-ion batteries used in electric vehicles (EVs) is greatly affected by temperature. Hence, an efficient
Downloaded from ascelibrary.org by 103.141.55.162 on 05/11/23. Copyright ASCE. For personal use only; all rights reserved.
battery thermal management system (BTMS) is needed to ensure the safety of batteries and prolong the cycle life. In order to find a more
efficient type of cooling plate for the rectangular batteries, the three-dimensional models of four common cooling plates with different internal
structures are established. After a series of computational fluid dynamic simulations and comparisons, the most optimum structure of the
cooling plate is obtained. Subsequently, the effect of different mass flow rates is investigated among the different cooling plates. It indicates
that the cooling plate with convex structure has a better cooling performance than the other three, and the heat transfer performance of various
cooling plates changes a lot with the increasing of mass flow rate. The convex structured cooling plate could be applied for optimizing
the performance for electric vehicles. DOI: 10.1061/(ASCE)EY.1943-7897.0000648. This work is made available under the terms of
the Creative Commons Attribution 4.0 International license, https://creativecommons.org/licenses/by/4.0/.
Author keywords: Electric vehicle; Battery thermal management system; Cooling plate; Simulation.
Fig. 2. (Color) Different structures of cooling plate: (a) convex structure; (b) honeycomb structure; (c) airfoil structure; and (d) U-shaped structure.
simulation, the flow rate was set to 0.0223 kg s−1 first, which was a of the battery box walls, and the heat convection resistance with
real value of the pump. The diameter of the inlet was set to 20 mm, ambient air. Because of the complex heat transfer condition, it
and the dynamic viscosity of the medium was 0.00339 Pa · s. can be assumed that the heat could only be taken away by the cool-
Therefore, the Reynolds number can be calculated by Eq. (1), ant. The boundary condition can be seen in Fig. 4.
which is 432. Thus, the inlet condition can be seen as the laminar As shown in Fig. 4, the model was the conjugate heat transfer
flow. and the inlet temperature of the glycol solution was set to be model: only one direction of the plate can absorb the heat from
298.15 K. The capacity of the rectangular Li-ion battery was 100 the battery, and the other outside surfaces of cooling plate are adia-
Ah and the mean voltage was 3.66 V. The battery is shown in batic of surroundings. Then the heat is removed by the coolant.
Fig. 3(a). The heat generation inside the battery is a complex pro- Flow-related parameters (such as velocity) change with time,
cess, so it is good to directly measure the heat flux of a battery in a and the flow of the fluid at this time is called unsteady flow.
certain charge or discharge rate. When the battery pack discharged The convergence variables are X-velocity, Y-velocity, Z-velocity,
at 1C, the heat flux on the bottom of the battery was tested by two continuity equation, and energy. The convergent criteria were
heat flow densitometers, and the location can be seen in Fig. 3(b). defined that all these residuals were less than 0.00001.
Then the data were recorded every 10 min, and the average results In this simulation, three factors were considered to evaluate the
of five trials show the mean heat flux was 1,257.86 W m−2 . performance of cooling plate: pressure drop (ΔP), maximum
Because the pack was sealed by the plastic, it can be assumed that temperature of cooling plate (T max ), and standard deviation of tem-
the heat could only be taken away by the cooling plate. The basic perature (T σ ). All three factors can be monitored by STAR-CCM+.
information of the battery is given in Table 3 The pressure drop between the inlet and outlet reflects the
pump power consumption. The battery power will reduce with the
ρud
Re ¼ ð1Þ decreasing of the flow resistance of the coolant. This paper did not
μ just simulate a single U-shaped channel, it also simulated many
The cooling plates were coupled with the bottom of battery different vortex generators inside the cooling plate. Fig. 5 shows
modules. The other sides of the plates were exposed to air inside the drag force and resistance of fluid around a three-dimensional
the battery box. Therefore, the heat resistance may include the ra- object. Therefore, the formula for calculating the pressure drop
diant heat resistance inside the box, the heat conduction resistance of an interior channel as follows:
F D ¼ Ff þ F p ð2Þ
Table 2. Freezing point and boiling point at different concentrations
Concentration Freezing point (K) Boiling point (K) where FD = total resistance; Ff = frictional resistance, which
0.0 273.2 373.2 is caused by the shearing stress from the vortex generator; and
10 270 374.3 Fp = pressure resistance, which is caused by the normal stress from
20 265.4 375.4 the vortex generator
30 259.1 377.6
40 250.9 378.8 ρu20 ρu20 ρu20
FD ¼ CD A ; Ff ¼ Cf A ; Fp ¼ Cz A
50 239.4 380.4 2 D 2 f 2 D
60 224.9 383.2
ð3Þ
Fig. 3. (Color) Battery experiment: (a) battery setup; and (b) sensor locations.
where ρ = density of the coolant (kg=m3 ); and u, v, and w = where u~ = velocity vector of the coolant (m=s); and fx , f y , and
velocity vectors of the refrigerant in three coordinates (m=s). f z = body force of the refrigerant in three coordinates.
(a) (b)
Fig. 6. (Color) Effects of different mesh sizes on simulation results: (a) pressure loss; and (b) outlet temperature.
Fig. 7. (Color) Mesh of cooling plate: (a) full domain; and (b) section.
Fig. 8. (Color) Effect of different structures: (a) pressure loss; (b) maximum temperature; and (c) standard deviation of temperature.
Fig. 9. (Color) Temperature nephograms of cooling plates: (a) convex structure; (b) honeycomb structure; (c) airfoil structure; and (d) U-shaped
structure.
plate (T max ), and standard deviation of the temperature (T σ ) change Fig. 8(b) shows the maximum temperature of the cooling
with the discharge of battery under the mass flow rate of 0.0223 kg=s. plate. It can be seen that all these structures displayed a similar
It can be seen in Fig. 8(a) that the cooling plate with the convex maximum temperature. The largest temperature was given by
structure had the least resistance because of its more reasonable the convex structure, which reached 305.1 K. The smallest temper-
structure, followed by the U-shaped structure, the airfoil structure, ature was 304.5 K of the U-shaped structure. Therefore, the main
and the honeycomb structure, which has the greatest resistance. influence factors of the maximum temperature are not the structure
At the highest flow rate, the pressure loss of the honeycomb arrange- of the internal fluid. As shown in Fig. 9, we can clearly see that the
ment exceeded 72.1% of the convex structure and 51.5% of the area on the right of the outlet is hotter. This is because as the fluid
airfoil structure under the same flow rate, which undoubtedly greatly flows inside the channel, it absorbs the heat from outside con-
increases the power consumption, and it is unreasonable to adopt the stantly, so the temperature will increase along with the flow direc-
cooling plate with honeycomb structure and airfoil structure. tion. Additionally, the edge of the cooling plate has the no fluid
Fig. 10. (Color) Comparisons of the different structures with different mass flow rates: (a) pressure loss; (b) maximum temperature; and (c) standard
deviation of temperature.
through it. The area, as a result, would be the most dangerous part range is wide enough to compare the four structures. The pressure
of the battery. loss with different flow rates are shown in Fig. 10(a). The increase
The internal channel has little effect on the maximum temper- of ΔP corresponds to Eq. (3), which is the quadratic function of the
ature of the cooling plate, but changes the standard deviation of the mass flow rate. In the figure, we can see that the four kinds of cool-
temperature greatly. We can find that honeycomb structure not only ing plates have similar ΔP at low mass flow rate. Additionally, the
has the maximal resistance; it also has the worst temperature airfoil structure and honeycomb structure both increased more
uniformity. From Fig. 8(c) we can find the standard deviation is quickly than the other two. In other words, in the case of power
0.98% smaller than the honeycomb structure. Except the honey- consumption, neither the airfoil structure nor the honeycomb struc-
comb structure, the other three structures behave with similar tem- ture is the best choice.
perature uniformity. In Fig. 10(b), the maximum temperature of the cooling plate
According to the preceding analysis, the cooling plate with the always appeared at the edge of the outlet. And the maximum
convex structure can meet the requirements of minimum resistance temperature of structures was less than 1 K at different mass flow
as well as the good heat transfer effect. rates, so it can be inferred that most of those types of cooling
plates would have the same problem at a certain temperature.
At the same time, the different between the minimum and maxi-
Effect of Different Flow Rates mum temperatures should not surpass 10 K; otherwise, the bat-
As shown in Fig. 10, in order to explore the performance of cooling teries’ efficiency would be affected and human lives could be
plates with different mass flow rates, the inlet mass flow rates were threatened. Therefore the mass flow rate of this situation should
set as 0.00558, 0.1115, 0.0223, and 0.0446 kg=s; the simulation be no less than 0.0223 kg=s.