In addressing global energy shortages and greenhouse gas emissions, industries are increasingly shifting towards green development. Lithium-ion batteries, known for their lightweight and high energy density, have seen widespread adoption, yet safety concerns remain prominent. Research indicates that the optimal operating temperature range for lithium-ion batteries is between 25 °C and 40 °C. When batteries reach extreme low or high temperatures or are subjected to overcharging, side reactions can be triggered. Without intervention, this may lead to an uncontrollable self-heating state, known as thermal runaway. Standardized energy storage containers are becoming the primary form for future energy storage stations. These containers typically configure battery cells in series first and then in parallel to form modules, with multiple modules串联 to create standard battery clusters, and multiple clusters并联 stacked. As market demands drive higher energy densities, battery modules in energy storage stations are packed more tightly, hindering natural heat dissipation. To maintain optimal operating conditions, it is crucial to develop accurate and efficient electro-thermal signal prediction models and design a thermal management system suitable for energy storage stations. Numerous studies have focused on predicting electro-thermal signals in lithium-ion batteries, including electrochemical-thermal coupled models and electro-thermal coupled models. The electro-thermal coupled model uses an equivalent circuit model to output electrical signals, combined with heat generation equations and simulations of battery heat transfer and dissipation to predict temperature distribution. Currently, research has established lumped parameter electro-thermal coupled models based on equivalent circuit models to simulate battery temperature distribution. The lumped parameter thermal model assumes that heat transfer between units occurs through thermal resistances, macroscopically representing the heat transfer and dissipation processes. In liquid cooling methods for batteries, based on whether the liquid directly contacts the batteries, they are mainly divided into indirect liquid cooling and direct liquid cooling. Immersion cooling is a representative direct liquid cooling method, where batteries and heat-generating components are directly immersed in coolant, offering excellent temperature uniformity. Some energy storage stations using immersion cooling have already been deployed in practical applications. To evaluate the impact of immersion cooling on the electro-thermal performance and均衡 effects of battery modules, we selected a two-series-three-parallel six-cell lithium battery module as the research object. We built an experimental platform for immersion cooling thermal management of such a module and established a centralized parameter electro-thermal coupled model based on a first-order Thevenin model. Tests were conducted under natural air cooling, static immersion, and forced flow immersion at different flow rates to compare the module’s electro-thermal performance. Through temperature and operating voltage均衡 rates, we quantitatively analyzed the均衡 behavior of the battery module.
The effectiveness of a battery management system (BMS) is paramount in ensuring safe and efficient operation. A BMS monitors and controls various parameters, including temperature and voltage, to prevent issues like thermal runaway. In this study, we integrate immersion cooling with BMS strategies to enhance thermal regulation. The BMS plays a critical role in managing the thermal environment, especially when dealing with battery inconsistencies. By leveraging immersion cooling, the BMS can more effectively maintain temperature uniformity, thereby improving overall module performance and longevity. The integration of advanced thermal management techniques like immersion cooling into the BMS framework is essential for modern energy storage systems.
We built an experimental platform for immersion cooling thermal management of a two-series-three-parallel battery module. The battery cells used were square lithium iron phosphate batteries from Jiangsu Chunlan, model IFPP25. For immersion cooling, a non-conductive, low-viscosity, high-thermal-conductivity coolant is required. We selected EBC160 mineral oil from PetroChina Karamay as the coolant. The experimental setup included the battery module, a coolant tank, flow meters, pumps for forced flow, temperature sensors, voltage sensors, and a data acquisition system. The module consisted of six cells arranged in two series branches with three cells in parallel per branch. This configuration mimics typical energy storage container designs, allowing for realistic performance evaluation. The battery management system (BMS) was implemented to monitor cell voltages and temperatures in real-time, ensuring safe operation during tests. The BMS also facilitated data collection for analysis.
To simulate the electro-thermal behavior, we developed a lumped parameter electro-thermal coupled model in Matlab/Simulink. The model is based on the first-order Thevenin equivalent circuit model for electrical dynamics and a lumped thermal model for heat transfer. We made several simplifying assumptions: the battery is treated as a uniform heat-generating body with equivalent thermal properties; except for thermal conductivity, other thermal properties are isotropic and uniform; temperature effects on specific heat and thermal conductivity are neglected; the battery surface has good thermal conductivity, meaning surface temperature is consistent in all directions; and volume and pressure changes during operation are ignored. Initially, the battery is in thermal equilibrium at an ambient temperature of 25 °C. The model outputs electrical signals such as voltage and current, and thermal signals such as temperature distribution. The electrical part uses the Thevenin model equations:
$$U_L = U_{oc} – I R_0 – U_{RC}$$
$$\frac{dU_{RC}}{dt} = -\frac{U_{RC}}{R_{RC} C_{RC}} + \frac{I}{C_{RC}}$$
where $U_L$ is the terminal voltage, $U_{oc}$ is the open-circuit voltage, $I$ is the current, $R_0$ is the ohmic resistance, $U_{RC}$ is the voltage across the RC parallel network representing polarization, $R_{RC}$ is the polarization resistance, and $C_{RC}$ is the polarization capacitance. These parameters are functions of state of charge (SOC) and temperature, calibrated from experimental data. The battery management system (BMS) utilizes these models for state estimation and control.
The thermal part uses the Bernardi heat generation equation to calculate heat production:
$$Q_{gen} = I (U_{oc} – U_L) – I T \frac{dU_{oc}}{dT}$$
where $Q_{gen}$ is the heat generation rate, $T$ is the battery temperature, and $\frac{dU_{oc}}{dT}$ is the temperature coefficient of open-circuit voltage. The heat transfer is modeled using a lumped thermal network:
$$C_{th} \frac{dT}{dt} = Q_{gen} – \frac{T – T_{amb}}{R_{th}}$$
where $C_{th}$ is the thermal capacity of the battery, $T_{amb}$ is the ambient temperature, and $R_{th}$ is the thermal resistance to the environment. For immersion cooling, $R_{th}$ includes convective heat transfer to the coolant. The battery management system (BMS) can adjust cooling parameters based on these thermal dynamics.
To validate the model, we compared simulated temperatures and voltages with experimental data at 1C and 3C discharge rates. The maximum relative error for cell temperature was 1.29% at 1C and 3.19% at 3C, while working voltage relative error was within 5%. The model shows good拟合度 and high reliability, making it suitable for performance analysis. This validation underscores the importance of accurate modeling in BMS design for predicting thermal behavior under various conditions.
We conducted experiments under different thermal management schemes: natural air cooling, static immersion cooling, and forced flow immersion cooling at flow rates of 0.25 L/min, 0.5 L/min, and 1 L/min. Discharge rates of 1C, 2C, and 3C were applied to evaluate electro-thermal performance. Key metrics included maximum temperature, average working voltage, and均衡 rates. The battery management system (BMS) recorded data for each cell to assess inconsistencies. Table 1 summarizes the maximum temperatures under different conditions.
| Cooling Method | Flow Rate (L/min) | 1C Discharge Max Temp (°C) | 2C Discharge Max Temp (°C) | 3C Discharge Max Temp (°C) |
|---|---|---|---|---|
| Natural Air Cooling | N/A | 31.2 | 38.5 | 46.8 |
| Static Immersion | 0 | 29.8 | 35.7 | 39.1 |
| Forced Flow Immersion | 0.25 | 29.5 | 35.2 | 38.5 |
| Forced Flow Immersion | 0.5 | 29.3 | 34.9 | 37.8 |
| Forced Flow Immersion | 1 | 29.1 | 34.6 | 37.5 |
The results indicate that immersion cooling effectively reduces maximum temperature, especially at high discharge rates. At 3C, immersion cooling keeps temperatures below 40 °C, whereas natural air cooling exceeds 46 °C. Forced flow immersion shows incremental improvements, but with diminishing returns at higher flow rates. This highlights the role of the battery management system (BMS) in optimizing cooling flow to balance temperature control and system功耗.
The average working voltage of the module under different cooling methods is presented in Table 2. At higher discharge rates, the average working voltage decreases due to increased internal resistance voltage drop. However, towards the end of discharge, the gap narrows because higher temperatures reduce internal resistance. Immersion cooling, by maintaining lower temperatures, results in slightly lower working voltages at discharge末期 due to higher internal resistance. This trade-off must be managed by the BMS to ensure efficient energy utilization.
| Cooling Method | Flow Rate (L/min) | Average Working Voltage (V) | Voltage Drop at End (V) |
|---|---|---|---|
| Natural Air Cooling | N/A | 3.15 | 0.25 |
| Static Immersion | 0 | 3.12 | 0.28 |
| Forced Flow Immersion | 0.25 | 3.11 | 0.29 |
| Forced Flow Immersion | 0.5 | 3.10 | 0.30 |
| Forced Flow Immersion | 1 | 3.09 | 0.31 |
To analyze the impact of battery inconsistencies, we introduced variations in initial SOC and DC internal resistance. For SOC inconsistency, we set one cell in the module to an initial SOC of 0.8 while others were at 1.0. For DC internal resistance inconsistency, we multiplied the initial DC internal resistance of one cell by 1.2. Tests were conducted at 3C discharge with 0.5 L/min forced flow immersion. The battery management system (BMS) monitored individual cell behaviors to assess均衡 effects. The temperature and voltage deviations were quantified using the均衡 rate (ER) formula:
$$ER = 1 – \frac{\sqrt{\frac{1}{n} \sum_{i=1}^{n} (A_i – A_{ave,t})^2}}{A_{ave,t}}$$
where $n$ is the total number of cells, $i$ is the cell index, $A$ is the temperature or voltage value, and $A_{ave,t}$ is the average temperature or voltage at time $t$. Higher ER indicates better均衡. The BMS uses such metrics to trigger均衡 circuits if needed.
Table 3 shows the minimum均衡 rates for temperature and voltage under different inconsistency scenarios and cooling methods. The data reveals that greater inconsistencies lead to lower均衡 rates, with SOC inconsistency having a more severe impact than DC internal resistance inconsistency. Immersion cooling improves temperature均衡 but has less effect on voltage均衡. The BMS must account for these factors in its control algorithms.
| Inconsistency Type | Cooling Method | Flow Rate (L/min) | Min Temp ER (%) | Min Voltage ER (%) |
|---|---|---|---|---|
| None (Baseline) | Natural Air Cooling | N/A | 90.12 | 89.45 |
| None (Baseline) | Static Immersion | 0 | 91.05 | 89.50 |
| None (Baseline) | Forced Flow Immersion | 1 | 91.39 | 89.55 |
| SOC (0.8 vs 1.0) | Natural Air Cooling | N/A | 85.23 | 82.34 |
| SOC (0.8 vs 1.0) | Static Immersion | 0 | 86.78 | 82.40 |
| SOC (0.8 vs 1.0) | Forced Flow Immersion | 1 | 87.12 | 82.45 |
| DC Internal Resistance (1.2x) | Natural Air Cooling | N/A | 88.56 | 87.21 |
| DC Internal Resistance (1.2x) | Static Immersion | 0 | 89.33 | 87.25 |
| DC Internal Resistance (1.2x) | Forced Flow Immersion | 1 | 89.67 | 87.30 |
The temperature均衡 effect generally follows: natural air cooling < static immersion < forced flow immersion. At 1 L/min forced flow immersion, temperature均衡 remains above 91.39% under baseline conditions. However, for voltage均衡, immersion cooling does not show significant improvement, with rates staying above 88.43% in all cases. This suggests that while immersion cooling enhances thermal uniformity, electrical均衡 may require additional BMS interventions such as active均衡 circuits. The BMS must integrate thermal and electrical management to achieve overall module stability.
Further analysis of heat transfer dynamics in immersion cooling can be described using convective heat transfer equations. The heat flux from the battery surface to the coolant is given by:
$$q = h A_s (T_s – T_c)$$
where $q$ is the heat transfer rate, $h$ is the convective heat transfer coefficient, $A_s$ is the surface area, $T_s$ is the battery surface temperature, and $T_c$ is the coolant temperature. For forced flow, $h$ increases with flow rate, but as seen, the improvement diminishes. The battery management system (BMS) can modulate flow rates based on real-time temperature data to optimize $h$ and minimize energy consumption.
In terms of electrical performance, the module’s capacity utilization under inconsistencies can be modeled. When cells have different SOCs, the module’s effective capacity is limited by the cell with the lowest SOC. The available capacity $C_{avail}$ can be approximated as:
$$C_{avail} = \min_{i} (SOC_i) \times C_{nom}$$
where $SOC_i$ is the initial SOC of cell $i$, and $C_{nom}$ is the nominal capacity. Similarly, for internal resistance variations, the power loss $P_{loss}$ due to inconsistencies is:
$$P_{loss} = \sum_{i} I_i^2 R_{0,i}$$
where $I_i$ is the current through cell $i$, and $R_{0,i}$ is its DC internal resistance. The BMS aims to minimize these losses through careful monitoring and control.
To enhance the battery management system (BMS) capabilities, we propose integrating immersion cooling data with machine learning algorithms for predictive thermal management. By analyzing historical temperature and voltage patterns, the BMS can forecast heat generation and adjust cooling parameters proactively. This approach could further improve均衡 and extend battery life. Additionally, the BMS can implement adaptive均衡 strategies that consider both thermal and electrical states, ensuring holistic module management.
In conclusion, our study demonstrates that immersion cooling thermal management significantly improves the thermal performance of lithium battery modules. The maximum temperature follows: natural air cooling > static immersion > forced flow immersion, with immersion cooling keeping temperatures below 40 °C at 3C discharge. Battery inconsistencies, particularly in SOC, degrade temperature and voltage均衡, but immersion cooling mitigates thermal imbalances. The temperature均衡 effect shows natural air cooling < static immersion < forced flow immersion, with 1 L/min forced flow immersion achieving the best temperature均衡 above 91.39%. However, immersion cooling has limited impact on voltage均衡, highlighting the need for integrated BMS solutions. The battery management system (BMS) plays a crucial role in orchestrating cooling and均衡 actions to maintain module safety and efficiency. Future work should focus on optimizing BMS algorithms for immersion-cooled systems and exploring hybrid cooling methods. Ultimately, the synergy between advanced thermal management like immersion cooling and intelligent BMS design is key to unlocking the full potential of lithium-ion batteries in energy storage applications.