The relentless evolution and adoption of electric vehicles (EVs) have brought the performance and safety of the EV battery pack into sharp focus. As the core energy storage component, the lithium-ion battery’s operational behavior is critically dependent on its temperature. Effective thermal management is not a luxury but a necessity to ensure safety, maximize lifespan, and maintain performance. Among various cooling strategies, liquid cooling using cold plates has emerged as a dominant solution due to its superior heat transfer efficiency and reliability compared to air cooling, and its lower cost and complexity relative to phase-change materials. This study delves into the cooling performance of an extruded aluminum alloy liquid cooling plate, employing computational fluid dynamics (CFD) to analyze and subsequently optimize its design. The primary objective is to minimize temperature non-uniformity within the EV battery pack module, thereby enhancing its overall thermal management system.
The foundation of this analysis is a detailed three-dimensional model. The EV battery pack assembly comprises multiple modules positioned atop the cooling plate. For simulation efficiency, a representative section was modeled, consisting of the cooling plate, a thermally conductive interface pad, and a simplified battery module acting as a uniform heat source. The cooling plate itself, fabricated via extrusion and friction stir welding, features a single-inlet, single-outlet design with parallel flow channels. Its primary dimensions form the basis for the computational domain.
The cooling performance is governed by the fluid dynamics and heat transfer within the plate. To accurately simulate this, appropriate physical models and boundary conditions must be established. The cooling fluid is a 50/50 mixture by volume of water and ethylene glycol. Its thermophysical properties, essential for the simulation, vary with temperature. Key properties at a relevant temperature of 25°C are summarized in the table below.
| Property | Value at 25°C | Unit |
|---|---|---|
| Density ($\rho$) | 1073.35 | kg/m³ |
| Specific Heat Capacity ($c_p$) | 3281 | J/(kg·K) |
| Thermal Conductivity ($k$) | 0.380 | W/(m·K) |
| Dynamic Viscosity ($\mu$) | 0.00394 | Pa·s |
The flow regime is determined by calculating the Reynolds number ($Re$), which is the ratio of inertial forces to viscous forces:
$$Re = \frac{\rho v L}{\mu}$$
where $v$ is the characteristic velocity and $L$ is the characteristic length (hydraulic diameter). For the given inlet flow rate of 15 L/min and the channel geometry, the calculated Reynolds number significantly exceeds 4000, confirming the flow is fully turbulent. Therefore, the standard k-epsilon ($k$-$\varepsilon$) turbulence model was selected for its robustness and accuracy in simulating such flows. The transport equations for turbulent kinetic energy $k$ and its dissipation rate $\varepsilon$ are:
$$\frac{\partial (\rho k)}{\partial t} + \frac{\partial (\rho k u_i)}{\partial x_i} = \frac{\partial}{\partial x_j}\left[\left(\mu + \frac{\mu_t}{\sigma_k}\right) \frac{\partial k}{\partial x_j}\right] + G_k – \rho \varepsilon + S_k$$
$$\frac{\partial (\rho \varepsilon)}{\partial t} + \frac{\partial (\rho \varepsilon u_i)}{\partial x_i} = \frac{\partial}{\partial x_j}\left[\left(\mu + \frac{\mu_t}{\sigma_\varepsilon}\right) \frac{\partial \varepsilon}{\partial x_j}\right] + C_{1\varepsilon} \frac{\varepsilon}{k} G_k – C_{2\varepsilon} \rho \frac{\varepsilon^2}{k} + S_\varepsilon$$
Here, $G_k$ represents the generation of turbulent kinetic energy due to mean velocity gradients, $\mu_t$ is the turbulent viscosity, and $\sigma_k$, $\sigma_\varepsilon$, $C_{1\varepsilon}$, $C_{2\varepsilon}$ are model constants.
The boundary conditions for the driving durability simulation scenario are critical inputs. The battery module generates a constant heat load of 500 W. The cooling fluid enters the plate at a mass flow rate of 0.2678 kg/s (equivalent to 15 L/min) and a fixed temperature of 25°C. The outlet is set as a pressure outlet. The initial temperature for both the battery and the environment is set at 40°C, simulating a challenging thermal condition for the EV battery pack. The material properties for all solid components are listed below.
| Component | Density (kg/m³) | Specific Heat (J/(kg·K)) | Thermal Conductivity (W/(m·K)) |
|---|---|---|---|
| Aluminum Plate (6xxx series) | 2700 | 900 | 138 (Isotropic) |
| Thermal Interface Pad | 1830 | 1000 | 3.0 (Isotropic) |
| Battery Module (Simplified) | 2218 | 1008 | Anisotropic: 5.3 (X), 23.4 (Y), 17.2 (Z) |
The computational mesh was generated within STAR CCM+, employing a polyhedral core mesh for all domains. To accurately resolve the high-temperature gradients near the walls, three prism layers were applied on the fluid-solid interfaces with a stretching factor of 1.2. The mesh was refined to ensure solution independence, particularly in regions of complex flow transition.
Baseline Design Analysis and Initial Results
The initial cooling plate design, termed the baseline, featured wide parallel flow channels. The simulation results revealed significant thermal gradients. The temperature distribution on the battery module showed a clear pattern: the lowest temperature of 26.3°C occurred at the bottom surface in contact with the cooling plate, while the highest temperature of 34.1°C was found at the top of a module farthest from the inlet. This resulted in a substantial module temperature difference ($\Delta T_{module}$) of 7.8°C. Furthermore, even on the module’s top surface, a temperature spread of 1.1°C was observed.
The root cause was identified by examining the fluid flow field. The velocity contour plot showed a highly non-uniform distribution. In the inlet header region, the flow was poorly distributed among the parallel channels. A significant portion of the coolant rushed through the channels closest to the inlet port’s direct path, while the channels farther away, particularly those on the opposite side of the header, experienced much lower flow rates. This maldistribution led to uneven cooling, where areas above low-flow channels became hotspots, critically undermining the thermal homogeneity of the EV battery pack.
First Optimization: Flow Channel Refinement
To address the flow maldistribution issue, the first optimization (Opt1) focused on the channel geometry. The original wide channels were subdivided into a larger number of significantly narrower channels. The goal was to increase the flow resistance of the individual channels relative to the inlet/outlet headers, thereby promoting a more balanced flow distribution according to principles of parallel flow networks. The total cross-sectional area for flow was maintained to keep the system pressure drop within a reasonable range.
The simulation of Opt1 showed a marked improvement. The module’s maximum temperature decreased from 34.1°C to 33.6°C, and the overall $\Delta T_{module}$ was reduced to 7.1°C. The top surface temperature difference also narrowed slightly to 1.0°C. Analysis of the new flow field confirmed that the flow distribution in the inlet header region became more uniform. However, a new problem emerged. The flow distribution in the outlet header region became highly uneven. The fluid exiting the numerous narrow channels converged in the outlet chamber. Due to the chamber’s geometry and the outlet port location, the coolant found a path of least resistance, creating a high-velocity “short-circuit” flow directly from the channels nearest the outlet port. Consequently, channels on the far side of the outlet header experienced very low flow, leading to inadequate cooling in the corresponding regions of the EV battery pack module. While the inlet issue was mitigated, the problem had shifted to the outlet.
Second Optimization: Outlet Header Streamlining
The second optimization (Opt2) specifically targeted the flow maldistribution in the outlet region identified in Opt1. The geometry of the outlet header chamber was modified. The large, open cavity was streamlined by lowering its roof in a smooth, curved profile, effectively guiding the flow from all channels more uniformly toward the outlet port. This redesign aimed to reduce recirculation zones and eliminate the preferential short-circuit path, forcing a more balanced collection of coolant from every channel.
The results from the Opt2 simulation demonstrated the success of this approach. The flow field in the outlet region showed dramatically improved uniformity. The high-velocity shortcut was eliminated, and all channels contributed more equally to the outflow. This translated into better thermal performance. The module’s maximum temperature dropped further to 33.1°C, and the minimum temperature also decreased significantly to 25.9°C. While the overall $\Delta T_{module}$ was 7.2°C (slightly higher than Opt1 due to the lower minimum temperature), the most critical metric for cell longevity and performance—the temperature spread across the module plane, especially at the top—was minimized. The top surface temperature difference was reduced to 0.7°C, indicating excellent thermal homogeneity. The progression of results across all three designs is summarized comprehensively below.
| Performance Metric | Baseline Design | Optimization 1 (Narrow Channels) | Optimization 2 (Narrow Channels + Streamlined Header) |
|---|---|---|---|
| Module Maximum Temperature ($T_{max}$) | 34.1 °C | 33.6 °C | 33.1 °C |
| Module Minimum Temperature ($T_{min}$) | 26.3 °C | 26.5 °C | 25.9 °C |
| Module Temperature Difference ($\Delta T = T_{max} – T_{min}$) | 7.8 °C | 7.1 °C | 7.2 °C |
| Module Top Surface Temperature Difference | 1.1 °C | 1.0 °C | 0.7 °C |
| Cooling Fluid Outlet Temperature (Approx.) | ~26.9 °C | ~26.9 °C | ~25.8 °C |
| Flow Distribution | Poor at Inlet | Good at Inlet, Poor at Outlet | Good at Inlet and Outlet |
Discussion and Implications for EV Battery Pack Design
The iterative CFD-driven optimization process yielded significant insights. The baseline design suffered from fundamental flow distribution problems, leading to substantial thermal gradients undesirable for any EV battery pack. The first optimization, channel refinement, successfully addressed the inlet-side maldistribution by increasing channel resistance. This highlights a key design principle: achieving balanced flow in parallel-channel cold plates often requires careful sizing of channel hydraulic diameter relative to header dimensions to ensure equitable pressure drop distribution.
However, Opt1 also revealed that the outlet header design is equally critical. An ill-designed outlet chamber can undo the benefits of a good inlet design by creating preferential flow paths. The second optimization demonstrated that streamlining the outlet header to guide flow smoothly toward the exit port is an effective solution. The combined strategy of channel refinement and header streamlining proved most effective in minimizing temperature spreads on the battery module surface.
The improvements, though seemingly modest in absolute temperature reduction (a 1°C drop in max temperature from baseline to Opt2), have profound implications. For lithium-ion batteries, operating at a lower and more uniform temperature directly reduces the rate of degradation mechanisms like solid electrolyte interphase (SEI) growth and lithium plating. This translates directly into extended cycle life and enhanced safety for the EV battery pack. Moreover, a lower temperature difference across the module ensures more balanced cell aging and state-of-charge (SOC) characteristics during operation and charging, which is crucial for the effectiveness of the battery management system (BMS).
Conclusion
This study successfully employed high-fidelity CFD simulation to analyze and optimize the cooling performance of an extruded aluminum liquid cooling plate for an EV battery pack. The analysis underscored that the cooling efficiency and thermal homogeneity are not merely functions of material or flow rate but are deeply sensitive to the geometric design of the flow channels and their distribution headers. Through a two-stage optimization process—first refining the channel width to improve inlet flow distribution, and subsequently streamlining the outlet header to correct outlet flow maldistribution—the thermal management performance was significantly enhanced. The optimized design achieved a lower maximum module temperature and, most importantly, a substantially reduced temperature variation across the module plane. These improvements contribute directly to increased safety, longevity, and reliable performance of the EV battery pack. The methodology and findings provide a valuable reference for the structural optimization of extruded aluminum liquid cooling plates, emphasizing the necessity of a holistic design approach that considers the entire fluid path to ensure effective and uniform thermal management.