Study of Cooling Performance of Liquid-Cooled EV Battery Module According to the TIM Compression Ratio

Worldwide, the issues of energy shortage and environmental pollution have become prominent, leading the automotive industry to focus extensively on the development of environmentally friendly vehicles such as hybrid electric vehicles (HEVs) and electric vehicles (EVs), instead of traditional internal combustion engine (ICE) vehicles. Consequently, the automobile industry is transitioning from using conventional fossil fuels to batteries as the power source. HEVs and EVs require large amounts of power to drive electric motors during vehicle acceleration, thereby necessitating the use of lithium-ion batteries, which are characterized by high energy density and low self-discharge. In this process, current consumption is significant, resulting in the generation of a substantial amount of heat in the battery pack. Prolonged exposure to heat can degrade the battery’s lifespan and cause swelling, that is, expansion of the battery cells owing to gas generation inside them (Oh et al., 2014). This swelling is associated with uncontrollable escalation of abnormal heat, leading to thermal runaway, which can ultimately cause fires in the entire battery pack. Therefore, designing cooling systems for battery packs that are used as power sources in HEVs (Hybrid Electric Vehicles) and EVs (electric vehicles) is a critical challenge.

The cooling methods for the battery packs used in HEVs and EVs broadly include air cooling, phase change material (PCM)-based cooling, and liquid cooling. First, in air cooling, external ambient air is drawn in through a blower motor to cool the battery. This method is predominantly used in hybrid vehicles because the electric power requirement is relatively low owing to combined use of an engine and an electric power source. Although the density and specific heat of air are lower than those of liquids, the output of the electric power source used in HEVs is relatively modest compared to those used in pure EVs, and therefore, system design for cooling performance in HEVs is comparatively simple (Wang et al., 2016). Second, in the PCM-based cooling method, PCMs such as wax or paraffin are commonly used. A PCM transitions between the solid and liquid states by releasing or accumulating the heat generated by the battery. This facilitates maintenance of the battery pack temperature within the operational temperature range through the absorption and storage of excess heat or release of stored heat, as needed (Li et al., 2020). Finally, liquid cooling, which involves the utilization of an insulated liquid coolant, is used to cool batteries in pure EVs, which have high power demands. Depending on the installation location within the vehicle, liquid cooling systems are classified into two main types. Direct liquid cooling involves circulation of a coolant between battery cells to cool them directly (Larrañaga-Ezeiza et al., 2022). By contrast, in indirect liquid cooling, cooling plates installed beneath the battery cells are utilized to create a network of cooling channels that dissipates heat indirectly (Deng et al., 2018). This form of cooling is widely adopted in view of the specific requirements of EVs with high power output (Anisha & Kumar, 2023; Thakur et al., 2020).

In indirect liquid cooling, a gap exists between the cooling plate and the battery module inside the assembly of the battery pack. This gap can lead to noise generation and discontinuity of heat transfer, which would not warrant enhancement of the battery pack’s cooling performance (Viswanath et al., 2000). To address this issue, a thermal interface material (TIM) is applied between the cooling plate and battery module to improve contact thermal conductivity (Gwinn & Webb, 2003; Kumar Swamy & Satyanarayan, 2019).

TIMs are broadly classified into pad- and gel-type materials. While these materials are prone to deformation under pressure, they increase the contact area under compression, thus effectively filling the gap between the cooling plate and battery to improve heat transfer (Sarvar et al., 2006). Furthermore, contact thermal resistance decreases as the pressure applied at the junction between the two surfaces increases, as in Eq. (1) related to bond line thickness (BLT). According to this equation, as the contact thermal resistance decreases, thermal conductivity increases (Sarvar et al., 2006). Here, BLT represents the contact distance between two solid surfaces, as illustrated in Fig. 1.

We utilize compression-induced changes in the thermal conductivity of TIMs as an input parameter in our analysis. Thermal conductivity of TIMs for the computational works in this study is measured by following the ASTM-D5470 standard, wherein the heat generated upon compression of the TIM is measured. Variations in thermal conductivity with changes in the compression ratio of the TIM are observed. Based on these observations, a computational fluid dynamics (CFD) analysis is performed to simulate the changes in the thermal conductivity of the TIM with varying thickness under different compression ratios in a liquid-cooled battery module. The objective is to demonstrate the improvement in cooling performance without increasing the power consumption of the pump to provide additional coolant supply. A detailed explanation of the method used to measure thermal conductivity is provided in Sect. 2.5.4.

The fluid governing equations used for the thermal and flow analysis of the coolant are based on the assumptions of steady state, incompressibility, and a Newtonian fluid. These equations include the continuity (5), momentum (6), and energy conservation equations (White & Xue, 2021). (7) For simulating the behavior of the coolant. Equation (5) is derived from the law of conservation of mass flow rate, which means that mass is conserved when the coolant flows through a closed curved surface inside of cooling plate. Equation (6) is a Navier–Stokes equation derived from Newton’s second law, which is a momentum equation of a coolant in the form of a nonlinear partial differential. Equation (7) is an energy equation used to simulate in consideration of heat transfer and temperature change of the coolant. Here, the \(\mu \Phi\) term means dissipation of frictional heat inside the fluid due to viscosity of the coolant.

Internal spaces of a battery pack has an air portion and a solid parts that constitutes the battery cell itself and module assembly, but convective heat flow in the air area was excluded from the analysis in this study. However, boundary conditions considering the temperature in the air space were reflected on the solid surface with the boundary condition Eq. (9). Equation (9) is convection heat transfer equation. \({T}_{\text{b}}\) and \({T}_{\infty }\) denote battery surface temperature and surrounding temperature.

To implement the thermal behavior of a battery module composed of lithium-ion battery cells, a heat generation model is required. The commonly used formula for calculating the heat generated by a battery is Bernardi’s heat generation model which incorporates Ohm’s law, and it is expressed as Eq. (9) (Bernardi et al., 1985; Zhu et al., 2020).

Here, I denotes the magnitude of battery current, regardless of whether the battery is charging/discharging, and R represents the equivalent internal resistance of the battery. U_OC denotes the open-circuit voltage of the battery circuit, and \(\frac{\text{d}{U}_{\text{OC}}}{\text{d}T}\) represents the entropy coefficient indicating the entropy change of the battery cell (Zhu et al., 2020). Furthermore, the charging and discharging rates of the battery are set to 5C rate, as indicated in Table 1, to observe its behavior under extreme conditions.

In this study, the fluid domain near the cooling plate part, which represents the lowest part of the liquid cooling system in the liquid-cooled battery module, consists of the coolant, while the solid domain consists of the battery cell, compression pad, cooling plate, TIM, heat sink, and busbar. To perform the thermal and fluid analysis, we used the commercial CFD software package Ansys Fluent. The mesh was generated using Fluent Meshing, and it consisted of approximately 5,200,000 elements for one battery module. To ensure stability in the analysis process and reduce computational costs, we adopted a polyhedral mesh configuration (Sosnowski et al., 2018). To enhance the accuracy of heat transfer and fluid flow analysis, particularly in the coolant, three prism layers were stacked, as illustrated in Fig. 11.

The initial battery cell temperature and the coolant inlet temperature were assumed to be room temperature, that is, 25 °C. To closely simulate the actual flow conditions, the SST K-omega turbulence model, a standard turbulence model, was used. Additionally, Table 6 summarizes the properties of the coolant and the flow conditions. Propylene glycol, commonly used for cooling battery packs, was used as the cooling fluid (Kwon & Park, 2016.) Considering the coolant supply rate, viscosity and density of coolant (Table 6), coolant circuit dimension (Fig. 7) and flow velocity inside of coolant circuit (Fig. 2d), \(Re\) is averagely in the range of 1200–2000 under these conditions and in some regions of coolant circuit turbulence might occur.

The walls on the coolant side of coolant passage were set to be adiabatic, while the remaining walls, excluding the coolant section, were set to be convective. This is because the parts other than the coolant were exposed to air, and the coolant section was blocked from air exposure by the upper and lower cooling plates. That is to say, the parts of coolant, busbar, battery cell, compression pad, end plate, heat sink, and TIM contact are in the heat conduction mode. In addition, the wall part in contact with air is in convective heat transfer mode.

We analyzed the cooling performance of a liquid-cooled battery module by considering the change in the thermal conductivity of the TIM owing to changes of its compression ratio. During CFD-based numerical analysis of heat and fluid flows, all boundary conditions were set appropriately as Table 7. The analysis was conducted by considering the compression-ratio-dependent variation in the thermal conductivity of the TIM. After completion of the CFD calculations, the maximum, minimum, and average temperatures in each case were examined, and the differences in the cooling performance of the liquid-cooled battery module were analyzed.

Figures 12 and 13 depict the streamlines and pressures of the coolant flowing in the coolant jacket at the bottom of the battery module No.21, respectively. Figure 14 depicts the overall temperatures in each case of battery module No.21 by considering the TIM-compression-dependent cooling performance. Figure 15 shows the probe plane at the center of the yz plane of the battery module No.21. The yz plane was selected made to observe the temperature distributions of all cells (Fig. 16). Moreover, the central region in the vertical direction of the cells was selected to easily observe the cooling effect resulting from coolant flow because this area experiences active coolant flow.

Figure 16 depicts the temperatures of the cells, compression pads, end plates, and heat sinks based on the probe plane selected in Fig. 15. The cooling effect is the most pronounced at the highest compression ratio of H_TIM = 3.0 mm, where the compression ratio is 25% in relatively thick regions without raggedness and 75% in the thin regions with raggedness. In all cases presented in Figs. 14 and 16, the asymmetric distribution of cell temperatures can be attributed to the streamline distributions of coolant as shown in Fig. 12, where fluid flow is more vigorous near the coolant inlet side and relatively less effective in areas further from the coolant supply.

Figure 17 illustrates the maximum, minimum, and average temperatures of the liquid-cooled battery cells in each case under TIM compression. In Case 1 (H_TIM = 4.0 mm), the maximum battery temperature was 63.35 °C. Comparatively, a cooling effect of approximately 3.20 °C reduction was observed when the TIM was compressed between case 1 (63.35 °C @H_TIM = 4.0 mm) and case 6(60.15 °C @H_TIM = 3.2 mm). The minimum temperature was observed in Cell 01, which was the closest to the coolant, while the average temperatures in Cases 1 and 6 were 54.725 °C and 51.862 °C, respectively, indicating a reduction of approximately 2.863 °C in Case 6. Furthermore, the pressure drop was consistently calculated as 1.53 kPa in all cases, suggesting the existence of a uniform pressure differential because the experimental setup did not alter the fluid flow structure.

As the compression ratio of the TIM material increases, the thermal conductivity \(({k}_{\alpha }\),\({k}_{\beta })\) increases and the rate of change \(\left(\frac{{\Delta k}_{\alpha }}{{\Delta H}_{\alpha }},\frac{\Delta {k}_{\beta }}{\Delta {H}_{\beta }}\right)\) also increases (Table 4). What is more, the heat conduction length decreases as the compression ratio increases, and this \(\left(\frac{\partial }{\partial {x}_{i}}\left({k}_{i}\frac{\partial {T}_{j}}{\partial {x}_{i}}\right)\right)\) accelerates the heat conduction from cell to coolant. The size variations of the cooling plate located under the bottom of each battery cell vary the heat flow rate. Heat is absorbed to the coolant with some distribution and it does not show a proportional decrease depending on the size change in compressibility of TIM. Therefore, as the compression ratio of the TIM increases, the thickness of TIM decreases and the cooling efficiency relatively increases.

In this study, thermal cooling analysis of a liquid-cooled battery module was conducted by considering changes in the thermal conductivity of the TIM depending on its compression ratio due to height variations resulting from assembly of the EV battery module. In addition, we explored the variation in the thermal conductivity of the battery module with the compression ratio of the TIM.

In future studies, we aim to perform numerical analysis of the battery module through structural analysis coupled with pressure profiles for verifying the cooling analysis results more precisely. This approach will enhance the accuracy of the cooling analysis by considering the coupled structural and thermal effects.