In the context of global carbon neutrality goals, battery energy storage systems have become pivotal for integrating intermittent renewable energy sources such as wind and solar power. These systems are increasingly deployed for grid frequency regulation and peak shaving, demanding high power density and fast response times. However, the inherent challenge in battery energy storage systems lies in thermal management. During high-rate charging and discharging, particularly at 3C rates, lithium-ion cells generate significant heat due to electrochemical reactions and ohmic resistance. This heat, if not effectively dissipated, leads to accelerated aging, capacity fade, and potential safety hazards such as thermal runaway. Conventional thermal management strategies, primarily bottom liquid cooling, face limitations under these high heat flux conditions due to the restricted heat transfer area and the temperature gradient that develops along the height of the cell. Our work addresses this critical bottleneck by proposing a novel gas-liquid dual-phase thermal management design for battery energy storage systems.
Our design integrates two complementary cooling mechanisms. The bottom of the cell pack retains a conventional indirect liquid cooling plate. Critically, we introduce a forced air convection loop at the top of the cells. The gaseous working medium (air) passes over the top surfaces of the cells, absorbing heat. This heated air is then channeled through a dedicated path to the bottom of the pack, where it passes through the fin structure of the liquid cooling plate. Here, the air transfers its absorbed heat to the coolant via the cold plate, cooling down before being recirculated to the top. This creates a dual-phase, dual-working-medium heat exchange loop. A typical configuration of the proposed system is shown below.
We built a detailed computational fluid dynamics (CFD) model using Flotherm-XT software to simulate the thermal performance of a 1P48S battery module (48 cells in series) under a full 3C charge-discharge cycle. The simulation parameters are summarized in the table below.
| Component | Density (kg/m³) | Specific Heat (J/kg·K) | Thermal Conductivity (W/m·K) | Dynamic Viscosity (Pa·s) |
|---|---|---|---|---|
| Cell (LFP) | 2205.5 | 1008.7 | X: 17.8, Y: 4.2, Z: 2.3 | – |
| Coolant (Water-Glycol) | 1077 | 3473 | 0.4 | 0.0046 |
| Cold Plate (Aluminum) | 2700 | 900 | 177 | – |
| Thermal Interface Material | 2000 | 1200 | 1.5 | – |
| Busbar (Aluminum) | 2702 | 903 | 237 | – |
| Housing (Steel) | 7850 | 470 | 52 | – |
| End Plate | 2760 | 963 | 96.2 | – |
| Aerogel Insulation | 320 | 1000 | 0.025 | – |
The baseline performance of the traditional bottom liquid-cooled battery energy storage system served as our reference. In this configuration, only the bottom surface of the cells is in direct thermal contact with the cold plate via a thermal interface material. Under the 3C cycle, the maximum cell temperature reached 54°C, with a significant temperature difference of 15°C across the module. This high temperature and large gradient are detrimental to the cycle life and performance consistency of the system. In contrast, our initial gas-liquid dual-phase design, without optimization, significantly improved the thermal profile. The maximum cell temperature was reduced to 46°C, and the module-level temperature difference was lowered to 10.6°C. The primary reason for this improvement is the doubling of the effective heat transfer area. The top surface of the cells, which previously only had contact with the ambient air or busbars, now actively participates in heat exchange with a forced air stream.
A key challenge we identified in the initial gas-liquid design was the temperature rise of the air as it flowed from the inlet to the outlet of the pack. The heated air in the downstream sections had a significantly reduced capacity to cool the cells. To solve this, we developed an “alternating gas channel” design. This structure splits the top air channel into layers. The cold air first enters the top layer. At a specific point along the flow path, a portion of the air is directed down to the first half of the cells. After absorbing heat, this air is channeled back to the top layer to exit. Simultaneously, the cold air that remained in the top layer is now directed to the second half of the cells. This ensures that the cells at the rear of the pack are also cooled by fresh, low-temperature air, rather than pre-heated air. This design further improved temperature uniformity.
The optimization of the system involved a detailed parametric study of the fin structures in both the top and bottom sections, grounded in heat transfer theory using the Number of Transfer Units (NTU) method. The total heat transfer rate Q can be expressed as:
$$ Q = \varepsilon m C_p (T_f – T_w) $$
Where \( \varepsilon \) is the effectiveness, \( m \) is the mass flow rate, \( C_p \) is the specific heat of air, \( T_f \) is the fluid temperature, and \( T_w \) is the wall temperature. The effectiveness is given by:
$$ \varepsilon = 1 – \exp(-\text{NTU}) $$
$$ \text{NTU} = \frac{h A}{m C_p} $$
The convective heat transfer coefficient h is a function of the Reynolds number (Re) and the fin geometry. In the laminar regime, the Nusselt number \( Nu \propto Re^{0.5} \), leading to \( h \propto u^{0.5} d^{-0.5} \), where u is the velocity and d is the fin gap. Under a constant fan power constraint (\( P = \Delta P \times v \)), the pressure drop \( \Delta P \propto u/d^2 \) leads to the relationship \( u \propto d \). Substituting these relationships into the NTU expression shows that the overall heat transfer is maximized at a specific fin gap. Our simulations for the top fins confirmed this, and we identified an optimal gap of 3 mm.
| Fin Gap (mm) | Max Cell Temperature (°C) | Max Module Temperature Difference (°C) |
|---|---|---|
| 3 | 47.5 | 8.0 |
| 6 | 48.2 | 8.8 |
| 11 | 49.1 | 9.5 |
| 18 | 50.5 | 10.2 |
We then turned our attention to the bottom liquid cooling plate, which serves as the heat sink for the entire system. To enhance the heat exchange between the heated air returning from the top and the liquid cold plate, we replaced the standard straight fins with corrugated fins. The corrugated geometry induces periodic flow separation and secondary flows, which break the thermal boundary layer and significantly increase the local convective heat transfer coefficient. We optimized both the gap and the wavelength of these corrugated fins. The optimal gap was found to be 6 mm, which balances the increase in heat transfer area and the induced flow resistance. The optimal corrugation wavelength was found to be 90 mm.
| Parameter | Max Cell Temperature (°C) | Max Module Temperature Difference (°C) |
|---|---|---|
| Straight Fins (Gap 3mm) | 49.0 | 9.5 |
| Corrugated (Gap 3mm) | 47.5 | 8.8 |
| Corrugated (Gap 6mm, Wavelength 90mm) | 46.0 | 8.5 |
| Corrugated (Gap 10mm) | 48.0 | 9.0 |
The combined effect of the alternating gas channel, optimized top fins, and optimized bottom corrugated fins resulted in a highly efficient thermal management system. A key metric we used to evaluate the overall performance is the heat exchange efficiency, \( \eta \). The total heat generated by the 48 cells during the 3C cycle, \( Q_{gen} \), is calculated by integrating the power curve over time:
$$ Q_{gen} = 48 \times \int_{0}^{2400} P(t) dt = 3,755,904 \, \text{J} $$
The heat absorbed by the cooling system \( Q_{abs} \) is calculated from the temperature change of the cells:
$$ Q_{abs} = C M \Delta T $$
For the baseline bottom liquid-cooled battery energy storage system, \( Q_{abs,base} \) was 1,464,802 J, yielding an efficiency \( \eta_{base} = 61\% \). For our optimized gas-liquid dual-phase system, \( Q_{abs,new} \) was 938,976 J, yielding a significantly higher efficiency \( \eta_{new} = 75\% \). This 14% absolute improvement in efficiency is a direct result of the increased heat transfer area and the enhanced heat transfer coefficients provided by the air cooling loop and optimized fin structures. The efficiency is defined as:
$$ \eta = \frac{Q_{abs}}{Q_{gen}} $$
The implications for the longevity of the battery energy storage systems are profound. The cycle life of lithium iron phosphate (LFP) cells is highly sensitive to operating temperature. In our simulations, we correlated the reduced cell temperatures with cycle life. For LFP cells operating above 30°C, cycle life degrades by approximately 1000 cycles for every 5°C reduction in continuous operating temperature. By lowering the maximum temperature by 8°C and significantly reducing temperature gradients, our design suppresses the parasitic side reactions that degrade the SEI layer and cause lithium inventory loss. The resulting average temperature reduction translates directly to a projected increase in pack cycle life of over 1400 cycles, enhancing the economic viability of the battery energy storage system.
| Thermal Scenario | Average Pack Temperature (°C) | Projected Cycle Life (Cycles until 80% SOH) |
|---|---|---|
| Bottom Liquid Cooling Only | ~50 | ~3800 |
| Optimized Gas-Liquid Dual-Phase | ~42 | ~5200 |
In conclusion, we have successfully demonstrated a novel gas-liquid dual-phase thermal management strategy for high-power battery energy storage systems. By integrating a forced air cooling loop on top of the cells with the existing bottom liquid cooling plate, we effectively doubled the heat transfer area. Through systematic optimization of the air channels (alternating flow) and fin geometries (top fins and bottom corrugated fins), we achieved a maximum cell temperature reduction of 8°C and a module-level temperature gradient reduction of 6.5°C under a demanding 3C charge/discharge cycle. The heat exchange efficiency of the battery energy storage system was increased from 61% to 75%. This thermal performance improvement is projected to extend the operational lifetime of the pack by more than 1400 cycles, significantly improving the total cost of ownership. The core technical pathways presented—enhancing heat transfer area and suppressing condensation via gaseous media coupling—provide critical guidance for the thermal management of next-generation high-power battery energy storage systems and other high-power density electronics. The findings lay a strong theoretical and practical foundation for designing more efficient, durable, and safer energy storage solutions. Our work confirms that dual-phase coupling is a viable and highly effective strategy for managing the extreme heat fluxes expected in future grid-scale battery energy storage systems.