In the context of escalating environmental and climate concerns, the new energy sector has experienced rapid development. In recent years, electric vehicles (EVs) powered by lithium-ion batteries have seen a significant increase in market share. During discharge, lithium-ion batteries generate substantial heat, but their performance is highly sensitive to temperature. The optimal operating temperature range for an EV battery pack is typically 25–40°C, with a maximum temperature difference not exceeding 5°C. For modern EVs, indirect liquid cooling via cold plates is a mainstream thermal management approach. The cooling performance of such systems depends on factors like flow channel geometry, dimensions, distribution, coolant velocity, and pressure drop. In this study, we focus on the structural design and optimization of a liquid cooling system for an EV battery pack, aiming to enhance temperature uniformity and reduce energy consumption.
The thermal behavior of an EV battery pack is critical for safety and efficiency. We begin by analyzing the heat generation of a square lithium iron phosphate (LFP) battery cell, commonly used in EV applications. The Bernardi model, based on the first law of thermodynamics, provides a heat balance equation that accounts for enthalpy changes, phase transitions, irreversible heat, and capacitive effects during charge-discharge cycles. The heat generation rate is expressed as:
$$Q = I \left[ (U_{\text{OCV}} – U) + T_b \frac{\partial U_{\text{OCV}}}{\partial T_b} \right]$$
where \( Q \) is the heat generation rate (W), \( I \) is the current (A), \( U_{\text{OCV}} \) is the open-circuit voltage (V), \( U \) is the working voltage (V), and \( T_b \) is the battery temperature (K). For the specific LFP battery used in our EV battery pack, experimental data fitting yields a simplified expression:
$$Q = 4 \times 10^{-3} I^2 + 0.114I$$
This allows us to calculate the heat generation per unit volume under different discharge rates, as summarized in Table 1. The heat generation increases substantially with higher discharge rates, emphasizing the need for effective cooling in high-power EV battery pack operations.
| Discharge Rate | Heat Generation (W/m³) |
|---|---|
| 0.5 C | 6,593.14 |
| 1.0 C | 15,934.07 |
| 1.5 C | 28,021.98 |
| 2.0 C | 42,857.15 |
| 2.5 C | 60,439.56 |
The EV battery pack in this study consists of 24 battery cells arranged in a 3P8S configuration (three modules in parallel, each with eight cells in series). Each module is separated by aluminum alloy side plates with polycarbonate insulation, and aerogel pads are placed between cells to minimize thermal interaction. The liquid cooling plate is positioned at the bottom of the EV battery pack, with thermal interface materials enhancing heat transfer. Key material properties are listed in Table 2.
| Component | Specific Heat (J/(kg·K)) | Density (kg/m³) | Thermal Conductivity (W/(m·K)) |
|---|---|---|---|
| Battery Cell | 1,633 | 2,136.75 | 29.0 (Y, Z directions); 1.0 (X direction) |
| Thermal Pad | 1,800 | 2,000.00 | 1.8 |
| Aerogel | 1,180 | 230.00 | 0.025 |
| Side Plate | 900 | 2,700.00 | 209.0 |
| Insulation Plate | 1,260 | 1,150.00 | 0.2 |
To address temperature uniformity issues in the EV battery pack, we designed a single-inlet-double-outlet parallel flow channel cold plate. This structure features a central inlet that splits coolant flow into two symmetrical branches exiting at the sides, reducing the flow path length and mitigating coolant temperature rise along the channel. The initial dimensions are 201 mm × 144 mm × 6 mm for the cold plate, with a main channel width of 8 mm, branch channel widths of 5 mm, and a channel thickness of 3 mm. The cooling performance is evaluated using computational fluid dynamics (CFD) simulations with a 50% ethylene glycol solution as coolant. The coolant properties vary with temperature, as shown in Table 3.
| Temperature (°C) | Density (kg/m³) | Specific Heat (J/(kg·K)) | Thermal Conductivity (W/(m·K)) | Dynamic Viscosity (Pa·s) |
|---|---|---|---|---|
| 20 | 1,073.38 | 3,281 | 0.380 | 0.00394 |
| 25 | 1,071.00 | 3,300 | 0.384 | 0.00339 |
| 30 | 1,066.00 | 3,339 | 0.391 | 0.00256 |
| 35 | 1,063.66 | 3,358 | 0.394 | 0.00226 |
Under a 1 C discharge rate, with coolant inlet velocity of 1 m/s and temperature of 25°C, the single-inlet-double-outlet parallel flow channel achieves a maximum temperature of 31.22°C and a maximum temperature difference of 4.81°C for the EV battery pack. The temperature differences among the three modules are 4.78°C, 4.70°C, and 4.75°C, with deviations not exceeding 0.08°C. This demonstrates balanced cooling and high temperature consistency, crucial for the longevity and safety of the EV battery pack. The design effectively counters the reduced cooling near the outlet due to coolant warming.
As discharge rates increase, the thermal load on the EV battery pack rises. We evaluated the cooling performance of a single cold plate at various discharge rates. The results, presented in Table 4, show that at 0.5 C and 1 C, the cooling system maintains temperatures within the optimal range. However, at 1.5 C, the maximum temperature difference exceeds 5°C, and at 2 C, both maximum temperature and temperature difference surpass acceptable limits, indicating the inadequacy of a single cold plate for high-rate scenarios.
| Discharge Rate | Maximum Temperature (°C) | Maximum Temperature Difference (°C) |
|---|---|---|
| 0.5 C | 27.5 | 3.2 |
| 1.0 C | 31.22 | 4.81 |
| 1.5 C | 35.95 | 8.5 |
| 2.0 C | 41.74 | 13.01 |
| 2.5 C | 48.1 | 18.3 |
To enhance cooling for high discharge rates, we proposed a four-cold-plate configuration with two arrangements: Scheme 1 places cold plates on the top, bottom, and sides of the EV battery pack, while Scheme 2 places cold plates on the top, bottom, and between modules. For vertical cold plates, a single-inlet-single-outlet parallel flow channel is used to minimize gravitational effects. The cooling performance is compared in Table 5. Scheme 2 consistently outperforms Scheme 1, especially at higher discharge rates. At 2.5 C, Scheme 2 reduces the maximum temperature by 1.27°C and the maximum temperature difference by 1.1°C compared to Scheme 1. This is attributed to better heat transfer from direct contact with modules and thermal isolation between modules, preventing heat diffusion in the EV battery pack.
| Discharge Rate | Scheme 1 Max Temp (°C) | Scheme 2 Max Temp (°C) | Scheme 1 Max ΔT (°C) | Scheme 2 Max ΔT (°C) |
|---|---|---|---|---|
| 0.5 C | 26.1 | 25.96 | 2.8 | 2.68 |
| 1.0 C | 28.5 | 28.3 | 3.9 | 3.75 |
| 1.5 C | 31.2 | 30.8 | 5.2 | 4.9 |
| 2.0 C | 34.8 | 34.1 | 7.1 | 6.5 |
| 2.5 C | 39.4 | 38.13 | 9.8 | 8.7 |
We optimized the single-inlet-double-outlet parallel flow channel structure to improve flow uniformity and reduce pressure drop, both critical for the efficiency of the EV battery pack cooling system. The optimization targets were minimizing the flow deviation \( f_r \) and pressure drop \( p_0 \), defined as:
$$f_r = \sum (\bar{f} – f_i)^2$$
$$p_0 = p_1 – p_2$$
where \( \bar{f} \) is the average flow rate, \( f_i \) is the flow rate in channel i, \( p_1 \) is inlet pressure, and \( p_2 \) is outlet pressure. Initial design showed a maximum flow deviation of 20% and a pressure drop of 3,653.4 Pa. Using ANSYS DesignXplorer with a Kriging interpolation response surface and multi-objective genetic algorithm (MOGA), we optimized channel thickness H and widths W1 to W7. The parameter ranges were H: 3–5 mm, W1–W7: 4–8 mm. The optimized values are H = 5 mm, W1 = 8 mm, W2 = 4 mm, W3 = 8 mm, W4 = 4 mm, W5 = 8 mm, W6 = 4 mm, W7 = 8 mm. Post-optimization results in Table 6 show a flow deviation reduction to 2.4% and pressure drop to 2,559 Pa, a 30% decrease, meeting design criteria for the EV battery pack cooling system.
| Channel | Initial Flow Rate (kg/s) | Optimized Flow Rate (kg/s) |
|---|---|---|
| f1 | 0.00697 | 0.01060 |
| f2 | 0.00582 | 0.01081 |
| f3 | -0.00586 | -0.01055 |
| f4 | -0.00698 | -0.01080 |
The optimized flow channel was compared with a traditional serpentine flow channel under a 2.5 C discharge rate. The optimized single-inlet-double-outlet channel reduced the maximum temperature from 30.92°C to 28.70°C (a drop of 2.22°C) and the maximum temperature difference from 4.5°C to 3.38°C (a drop of 1.12°C). This highlights the superiority of the parallel design in enhancing temperature uniformity and cooling efficiency for the EV battery pack. The serpentine channel suffers from coolant temperature rise along the path, leading to uneven cooling, whereas the parallel design ensures more balanced heat dissipation.
In conclusion, this study presents a comprehensive approach to designing and optimizing a liquid cooling system for an EV battery pack. The single-inlet-double-outlet parallel flow channel effectively improves temperature consistency by mitigating coolant warming effects. For high discharge rates, a four-cold-plate configuration with cold plates placed between modules offers superior cooling performance and thermal isolation. Optimization of channel dimensions significantly enhances flow uniformity and reduces pressure drop, contributing to energy efficiency. These findings provide valuable insights for developing advanced thermal management systems in EV battery packs, ensuring safety, longevity, and optimal performance under varying operational conditions. Future work could explore optimal channel spacing, coolant velocity effects, and rounded corners to further minimize flow resistance and vortices in the EV battery pack cooling system.