As the global energy transition accelerates toward decarbonization, the integration of wind and solar power into electrical grids has introduced significant challenges related to frequency stability. The inherent intermittency and volatility of renewable energy sources require fast-responding and high-capacity energy storage systems to balance supply and demand. Among various storage technologies, the battery energy storage system based on lithium-ion chemistry has emerged as a promising solution due to its high energy density, rapid response, and long cycle life. However, when operating under high charge/discharge rates—typical for frequency regulation and power smoothing—thermal management becomes critical. Excessive heat generation can degrade battery performance, reduce lifespan, and even lead to thermal runaway. In this work, we propose a liquid-cooled battery energy storage system tailored for high-rate frequency regulation applications. Through experimental characterization of a power-type 100 Ah lithium iron phosphate cell and numerical simulations of a 4 MW/1 MWh system, we demonstrate that liquid cooling effectively maintains battery temperature within safe limits under demanding operation profiles, including 4C continuous discharge and automatic generation control (AGC) frequency regulation scenarios. Our findings confirm that the proposed battery energy storage system design, combined with liquid thermal management, enables stable and reliable long-term operation for grid frequency support.
Experimental Characterization of Power-Type Lithium-Ion Cells
To understand the electro-thermal behavior of the battery energy storage system, we first conducted systematic experiments on a 100 Ah power-type lithium iron phosphate (LFP) cell. The cell design employs smaller particle sizes and thinner electrodes to reduce ionic resistance and polarization, enabling high-rate capability. We used a first-order equivalent circuit model (ECM) to describe the electrical behavior. The ECM parameters—open-circuit voltage (OCV), ohmic resistance, polarization resistance, and polarization capacitance—were identified through hybrid pulse power characterization (HPPC) tests. The test rig included a battery cycler, thermal chamber, and data acquisition system, as summarized in Table 1.
| Component | Specification |
|---|---|
| Battery cycler | Current range: 0–300 A, voltage accuracy: ±0.1% |
| Thermal chamber | Temperature range: –20 to 80 °C, stability: ±0.5 °C |
| Data acquisition | Sampling rate: 1 Hz, 16-bit resolution |
| Thermocouples | Type K, accuracy: ±0.5 °C |
The OCV as a function of state of charge (SOC) was measured at 25 °C. The ohmic resistance $R_\Omega$ and polarization parameters $R_p$ and $C_p$ were extracted from the voltage response during pulse tests. Figure 2 in the original study (not reproduced here) showed that $R_\Omega$ remained between 0.29 and 0.35 mΩ over most of the SOC range, while $R_p$ increased near the end of discharge. The polarization capacitance $C_p$ exhibited an inverse relationship with $R_p$. These low internal resistances confirm the suitability of the power-type cell for high-rate applications, as less heat is generated per unit current. Using the ECM, the cell voltage during operation is given by:
$$U = U_{\text{OCV}} + I R_\Omega + U_p(0) e^{-\frac{t}{R_p C_p}} + I R_p \left(1 – e^{-\frac{t}{R_p C_p}}\right)$$
The heat generation rate $Q_{\text{gen}}$ comprises irreversible and reversible (entropic) components:
$$Q_{\text{gen}} = Q_{\text{ir}} + Q_{\text{re}} = I(U_{\text{OCV}} – U) – I T \frac{dU_{\text{OCV}}}{dT}$$
The entropic coefficient $dU_{\text{OCV}}/dT$ was measured experimentally and is shown in Figure 5 of the original work. Thermal tests at 1C and 4C discharge rates under natural convection revealed that at 4C, the cell temperature rose by approximately 20 °C, reaching 45 °C at end of discharge—exceeding the recommended operating range for long-term stability. This motivated the adoption of active liquid cooling for the battery energy storage system.
System Integration and Simulation Model
We designed a 4 MW / 1 MWh liquid-cooled battery energy storage system based on the power-type cells. Each cell has dimensions of 174 mm × 72 mm × 204 mm (length × width × height). The module configuration is 1P12S (one parallel, twelve series), arranged as 3 rows and 4 columns. Liquid cooling is achieved using 65 mm wide, 6 mm thick micro-channel aluminum flat tubes, which are bent and welded to form cooling plates. Twelve modules are connected in series to form a battery string, placed on two trays for modular maintenance. The coolant pipes are connected in series within each module and in parallel between layers. The overall system houses eight battery racks, each equipped with an independent pump and heat exchanger.
A coupled electro-thermal-fluid simulation was developed using the finite volume method in Fluent. The electrical model (ECM) and heat generation model were implemented via user-defined functions (UDFs). The governing equations include continuity, momentum, energy for fluid, and energy for solid domains:
$$\frac{\partial \rho}{\partial t} + \nabla \cdot (\rho \mathbf{u}) = 0$$
$$\frac{\partial}{\partial t}(\rho \mathbf{u}) + \nabla \cdot (\rho \mathbf{u}\mathbf{u}) = -\nabla P + \nabla \cdot (\mu \nabla \mathbf{u})$$
$$\frac{\partial}{\partial t}(\rho c_p T) + \nabla \cdot (\rho c_p \mathbf{u} T) = \nabla \cdot (k \nabla T)$$
$$\frac{\partial}{\partial t}(\rho c_p T) = \nabla \cdot (k \nabla T) + Q_{\text{gen}}$$
The thermophysical properties of coolant, aluminum tube, and battery are given in Table 2.
| Material | Density $\rho$ (kg/m³) | Thermal conductivity $k$ (W/(m·K)) | Specific heat $c_p$ (J/(kg·K)) | Viscosity $\mu$ (Pa·s) |
|---|---|---|---|---|
| Coolant (water-glycol) | 1075 | 0.387 | 3300 | 0.0015 |
| Aluminum tube | 2719 | 202.4 | 871 | — |
| Battery (orthotropic) | 1855.9 | 1.2 (through-plane) / 20.0 (in-plane) | 1343 | — |
The simulation was validated by comparing with single-cell experiments under 4C discharge with liquid cooling (3 L/min, inlet temperature 25 °C). The temperature at the positive tab was monitored. Figure 5 of the original work (not shown) confirmed that the model predicts temperature rise within 3% error, proving its accuracy.
Thermal Management Performance Under High-Rate Conditions
We conducted three simulation cases to evaluate the battery energy storage system’s thermal behavior:
| Case | Condition | Coolant flow rate | Inlet temperature |
|---|---|---|---|
| 1 (Baseline) | 4C constant current discharge | 3 L/min per module | 25 °C |
| 2 | 4C constant current discharge | 1.5 L/min per module | 25 °C |
| 3 | AGC frequency regulation (4C peak) | 3 L/min per module | 25 °C |
Case 1: 4C Discharge with Nominal Flow
Figure 6 (original) showed the temperature evolution within the battery module during a 4C discharge. The maximum temperature reached 32.8 °C at the end of discharge, well within the optimal range of 15–35 °C. The temperature difference between cells was about 3 °C. The temperature rise of the battery and coolant outlet is shown in Figure 7. The battery temperature rise was 7.8 °C, and the coolant outlet temperature increased by 4.1 °C. The parabolic shape of the temperature curve indicates that the heat transfer gradually balances with heat generation, preventing runaway temperature increase. This confirms that the liquid-cooled battery energy storage system can maintain safe temperatures under sustained high-rate operation.
Case 2: Effect of Reduced Flow Rate
When the coolant flow rate was halved to 1.5 L/min per module, the maximum battery temperature at the end of 4C discharge reached 36 °C, and the inter-cell temperature difference increased to 7 °C. The lower flow rate reduces the convective heat transfer coefficient and causes greater temperature rise along the flow path, leading to thermal non-uniformity. Such gradients accelerate differential aging and compromise the overall lifetime of the battery energy storage system. Therefore, adequate flow rate is essential for long-term reliability.
| Parameter | Case 1 (3 L/min) | Case 2 (1.5 L/min) |
|---|---|---|
| Max battery temperature (°C) | 32.8 | 36.0 |
| Temperature rise of battery (°C) | 7.8 | 11.0 |
| Coolant outlet temperature rise (°C) | 4.1 | 9.0 |
| Max temperature difference between cells (°C) | 3.0 | 7.0 |
Case 3: Actual AGC Frequency Regulation Profile
We simulated the battery energy storage system under a realistic AGC signal lasting 15000 seconds. The current profile and corresponding heat generation rate are shown in Figure 10 (original). Despite rapid current variations, the thermal inertia of the system smooths temperature fluctuations. The maximum battery temperature remained below 33.5 °C while the heat generation rate averaged around 65,000 W/m³ during high-power periods. The temperature difference between cells stabilized at less than 2 °C (Figure 12). This demonstrates that the liquid-cooled battery energy storage system can effectively manage thermal loads under the highly dynamic conditions of frequency regulation, ensuring both safety and longevity.
Conclusion
In this work, we systematically investigated the applicability of a power-type lithium-ion battery energy storage system with liquid cooling for high-rate frequency regulation. Experimental results on 100 Ah LFP cells revealed that natural convection is insufficient for 4C operation, as temperatures approached 50 °C. Through a comprehensive numerical model validated by experiments, we analyzed the thermal performance of a 4 MW/1 MWh battery energy storage system under various scenarios. The liquid cooling scheme with 3 L/min per module maintained battery temperatures below 33 °C and inter-cell温差 within 2 °C under both 4C continuous discharge and real AGC profiles. These findings confirm that the proposed battery energy storage system integrated with liquid thermal management is a viable solution for grid-scale frequency regulation, meeting the demands of high power density and long operational life. Future work will explore optimization of cooling channels and control strategies to further enhance efficiency and cost-effectiveness of the battery energy storage system.