In the context of the global energy transition, battery energy storage systems have become essential for balancing renewable energy fluctuations and ensuring grid stability. Among various thermal management strategies, immersion liquid cooling offers superior heat dissipation for large-capacity battery packs due to direct contact with dielectric fluids, eliminating thermal contact resistance and maximizing heat transfer area. This study focuses on a 280 Ah large-capacity battery energy storage system, investigating the effects of battery spacing, coolant inlet/outlet configurations, flow velocity, and coolant type on cooling performance. Numerical simulations are conducted using a validated electrochemical-thermal model, and sensitivity analysis of coolant thermophysical properties is performed. The findings provide practical guidance for designing efficient immersion cooling systems for battery energy storage systems.
1. Numerical Model and Methodology
We employed a three-dimensional computational fluid dynamics (CFD) model to simulate the thermal behavior of a battery pack consisting of 52 cells (4×13 arrangement) with a nominal capacity of 280 Ah each. The battery dimensions are 204 mm × 174 mm × 72 mm. The housing height is 230 mm, with a wall clearance of 25 mm. The vertical spacing between cells is fixed at 10 mm, while the horizontal spacing di is varied from 0 to 10 mm. Nine different inlet/outlet configurations are considered, with positions relative to the bottom of the housing as summarized below.
| Configuration | Inlet Height (mm) | Outlet Height (mm) |
|---|---|---|
| Case1 | 182 | 182 |
| Case2 | 182 | 102 |
| Case3 | 182 | 20 |
| Case4 | 102 | 182 |
| Case5 | 102 | 102 |
| Case6 | 102 | 20 |
| Case7 | 20 | 182 |
| Case8 | 20 | 102 |
| Case9 | 20 | 20 |
The heat generation within each cell is calculated using the Bernardi equation, assuming uniform internal heat generation:
$$ q_v = \frac{1}{V_b} \left( I^2 R_t – I T \frac{\partial U_0}{\partial T} \right) $$
where \(V_b\) is the cell volume, \(I\) is the current, \(R_t\) is the internal resistance, and \(T\frac{\partial U_0}{\partial T}\) is taken as 11.16 mV. For 1 C discharge, the volumetric heat generation rate is 14386 W/m³. The battery is modeled as an orthotropic material with thermal conductivities: \(\lambda_x = \lambda_z = 21.6\) W/(m·K) and \(\lambda_y = 2.1\) W/(m·K). Density is 2118 kg/m³ and specific heat is 1029.4 J/(kg·K).
The fluid domain is governed by the continuity, momentum, and energy equations:
$$ \frac{\partial \rho}{\partial t} + \nabla \cdot (\rho \mathbf{v}) = 0 $$
$$ \frac{\partial (\rho \mathbf{v})}{\partial t} + \nabla \cdot (\rho \mathbf{v} \mathbf{v}) = -\nabla p + \mu \nabla^2 \mathbf{v} + \rho \mathbf{g} $$
$$ \rho c_p \frac{\partial T}{\partial t} + \nabla \cdot (\rho c_p \mathbf{v} T) = \nabla \cdot (\lambda \nabla T) $$
Initial temperature is 25 °C, coolant inlet temperature 25 °C, and outlet is pressure outlet. The housing exchanges heat with ambient air at 25 °C through natural convection with a coefficient of 5 W/(m²·K). A polyhedral mesh with approximately 6.05 million cells is used after grid independence verification, yielding a relative error below 1% for both maximum temperature \(T_{\text{max}}\) and maximum temperature difference \(\Delta T_{\text{max}}\).
2. Effect of Horizontal Battery Spacing
We first investigated the influence of horizontal spacing di (0, 2.5, 5, 7.5, 10 mm) using Case5 configuration (inlet/outlet at 102 mm) with an inlet velocity of 0.4 m/s. The results show that increasing the spacing initially improves cooling performance, but beyond an optimum value, the effect reverses.
| Spacing (mm) | \(T_{\text{max}}\) (°C) | \(\Delta T_{\text{max}}\) (°C) |
|---|---|---|
| 0 | 37.19 | 10.94 |
| 2.5 | 36.61 | 10.64 |
| 5 | 35.35 | 9.37 |
| 7.5 | 35.56 | 9.62 |
| 10 | 35.75 | 9.86 |
The best performance is achieved at 5 mm spacing, where both \(T_{\text{max}}\) and \(\Delta T_{\text{max}}\) are reduced by 1.84 °C and 1.57 °C respectively compared to zero spacing. This occurs because increasing the spacing reduces flow resistance and improves coolant circulation, but further increases lower the local flow velocity, weakening convective heat transfer. The velocity contours at mid-height confirm that a 5 mm gap provides a favorable balance.
3. Effect of Inlet/Outlet Configuration
Using the optimal spacing of 5 mm and inlet velocity 0.4 m/s, we evaluated nine configurations. The maximum temperature and temperature difference are shown below.
| Case | \(T_{\text{max}}\) (°C) | \(\Delta T_{\text{max}}\) (°C) |
|---|---|---|
| Case1 | 35.48 | 9.86 |
| Case2 | 35.50 | 9.88 |
| Case3 | 35.51 | 9.91 |
| Case4 | 35.29 | 9.32 |
| Case5 | 35.35 | 9.38 |
| Case6 | 35.39 | 9.41 |
| Case7 | 35.46 | 9.73 |
| Case8 | 35.48 | 9.76 |
| Case9 | 35.50 | 9.80 |
The inlet position has a much stronger impact than the outlet position. For example, when the outlet is fixed at 182 mm, lowering the inlet from 182 mm to 102 mm reduces \(\Delta T_{\text{max}}\) by 5.6%, while further lowering to 20 mm increases it. This is because the inlet location determines the vortex formation and flow distribution within the battery energy storage system housing. The velocity and vorticity fields at the vertical mid-plane reveal that Case4 (inlet at 102 mm, outlet at 182 mm) creates stronger vortices that enhance mixing and heat transfer. Changing the outlet position alone alters \(\Delta T_{\text{max}}\) by less than 1%. Therefore, we recommend positioning the inlet at mid-height for optimal cooling in battery energy storage systems.
4. Effect of Coolant Inlet Velocity
We varied the inlet velocity from 0.2 m/s to 1.6 m/s using Case4 configuration. Both \(T_{\text{max}}\) and \(\Delta T_{\text{max}}\) decrease with increasing velocity, but the reduction diminishes at higher velocities.
| Velocity (m/s) | \(T_{\text{max}}\) (°C) | \(\Delta T_{\text{max}}\) (°C) |
|---|---|---|
| 0.2 | 38.37 | 11.83 |
| 0.4 | 35.29 | 9.32 |
| 0.6 | 33.66 | 7.93 |
| 0.8 | 32.49 | 7.00 |
| 1.0 | 31.78 | 6.40 |
| 1.2 | 31.30 | 6.00 |
| 1.4 | 31.15 | 5.82 |
| 1.6 | 31.00 | 5.70 |
From 0.2 to 0.4 m/s, \(\Delta T_{\text{max}}\) drops by 21.2% (from 11.83 to 9.32 °C), while from 1.4 to 1.6 m/s, the reduction is only 2.0%. This trend is typical for forced convection: higher velocities enhance the heat transfer coefficient, but the marginal benefit decreases. Additionally, at velocities ≥0.8 m/s, the battery pack temperature reaches a steady state earlier in the discharge process, indicating faster thermal equilibrium. For practical battery energy storage system design, an inlet velocity around 0.4–0.8 m/s offers a good trade-off between cooling performance and pumping power.
5. Effect of Coolant Type
We compared five dielectric fluids: synthetic oil, MIVOLT-DF7, FC-72, deionized water, and silicone oil. Their thermophysical properties are listed below.
| Coolant | Density (kg/m³) | Specific Heat (J/(kg·K)) | Thermal Conductivity (W/(m·K)) | Dynamic Viscosity (Pa·s) |
|---|---|---|---|---|
| Synthetic oil | 807 | 2523 | 0.159 | 0.0070 |
| MIVOLT-DF7 | 916 | 1907 | 0.129 | 0.0150 |
| FC-72 | 1680 | 1100 | 0.057 | 0.0006 |
| Deionized water | 998 | 4182 | 0.598 | 0.0010 |
| Silicone oil | 965 | 1460 | 0.160 | 0.0965 |
Simulations were performed at 1.6 m/s using Case4 configuration. The results are summarized below, along with the pressure drop and average heat transfer coefficient on the battery surface.
| Coolant | \(T_{\text{max}}\) (°C) | \(\Delta T_{\text{max}}\) (°C) | \(\Delta P\) (Pa) | \(h\) (W/(m²·K)) |
|---|---|---|---|---|
| Deionized water | 28.64 | 3.61 | 3560 | 22.69 |
| Synthetic oil | 30.93 | 5.72 | 2543 | 20.15 |
| FC-72 | 31.11 | 6.02 | 4865 | 19.88 |
| MIVOLT-DF7 | 31.74 | 6.22 | 3120 | 19.44 |
| Silicone oil | 34.63 | 8.78 | 4230 | 18.81 |
Deionized water exhibits the best cooling performance, with \(T_{\text{max}}\) of 28.64 °C and \(\Delta T_{\text{max}}\) of 3.61 °C, well within the recommended range for battery energy storage systems. Its high specific heat and thermal conductivity, combined with moderate viscosity, lead to the highest heat transfer coefficient. Silicone oil performs worst due to its extremely high viscosity, which inhibits flow and reduces convection. Although FC-72 has low viscosity, its poor thermal properties result in inferior performance. The pressure drop varies with density and viscosity; synthetic oil offers the lowest pumping loss, while FC-72 requires the highest.
To further understand the influence of individual thermophysical parameters, we performed a sensitivity analysis by varying each property ±40% relative to synthetic oil (Group A) and deionized water (Group B). The results reveal the order of importance: density > specific heat > thermal conductivity > dynamic viscosity.
| Parameter | Variation Range | \(\Delta T_{\text{max}}\) range (Group A, °C) | \(\Delta T_{\text{max}}\) range (Group B, °C) |
|---|---|---|---|
| Density | 60%–140% | 7.65 → 5.20 | 4.65 → 3.06 |
| Specific heat | 60%–140% | 7.42 → 5.26 | 4.29 → 3.24 |
| Thermal conductivity | 60%–140% | 6.89 → 5.48 | 4.02 → 3.35 |
| Dynamic viscosity | 60%–140% | 5.70 → 6.20 | 3.22 → 3.75 |
The density affects the kinetic energy of the coolant and the thermal capacity per unit volume; higher density enhances both mixing and heat storage. Specific heat directly influences the heat absorbed per unit temperature rise. Thermal conductivity reduces thermal boundary layer resistance, and viscosity increases flow resistance, worsening heat transfer. This analysis provides a rational basis for selecting or developing coolants for battery energy storage systems. For instance, increasing density (e.g., using nanoparticles) could significantly improve cooling, but care must be taken to avoid excessive pumping power.
6. Conclusions
This numerical study systematically evaluated the cooling performance of an immersion liquid cooling system for a 280 Ah large-capacity battery energy storage system. The key findings are:
– An optimal horizontal battery spacing of 5 mm reduces both maximum temperature and temperature difference by more than 1.5 °C compared to zero spacing.
– The coolant inlet position has a dominant effect on the flow field and thermal uniformity; the best performance is achieved with a mid-height inlet (102 mm) and a top outlet (182 mm).
– Increasing inlet velocity improves cooling, but the benefit diminishes beyond 0.8 m/s. A velocity of 0.4–0.8 m/s is recommended for practical designs.
– Deionized water outperforms other coolants, yielding a maximum temperature difference of only 3.61 °C at 1 C discharge. Silicone oil is the least effective.
– Sensitivity analysis shows that density has the greatest impact on cooling performance, followed by specific heat, thermal conductivity, and dynamic viscosity.
These insights serve as a valuable reference for engineers designing efficient and reliable thermal management systems for battery energy storage systems, particularly for large-format cells with high heat generation rates.