Battery energy storage systems (BESS) are critical for stabilizing power grids with high renewable penetration, but their thermal management remains a key challenge. Inefficient cooling accelerates cell degradation and risks thermal runaway. Conventional indirect liquid‑cooling systems rely on chiller units that consume significant electricity, while direct immersion systems are costly and complex. To address these issues, we propose a novel cooling architecture that combines falling‑film evaporation with mechanical vapor recompression (MV R). This design leverages the high heat transfer coefficient of phase‑change cooling and the thermodynamic advantage of vapor recompression to dramatically reduce energy consumption. Our system uses water as the working fluid, operates under vacuum, and eliminates the need for a separate refrigerant loop. In this paper, we present the process design, simulate the system using Aspen Plus, compare its performance with a conventional air‑conditioning‑based cooling system under a 210 kW cooling load, and analyze the influence of key parameters. We also evaluate the economic feasibility through annualized cost and payback period.
The core idea is to let a thin water film evaporate on the surface of a plate that is in thermal contact with the battery modules. The generated vapor is then compressed by an MV R compressor, raising its saturation temperature, and finally condensed in an external cooler, releasing the heat to the ambient. The condensed water is returned to the evaporator after pressure reduction. This closed‑loop cycle consumes only the compressor work, which is much smaller than the work required by a conventional vapor‑compression chiller. The following sections detail the system layout, simulation model, results, and economic assessment.
System Description
Figure 1 (inserted below) shows the principle of the indirect liquid‑cooling module. Each battery cell is attached to a metal falling‑film plate (stainless steel, aluminum, or alloy) with thermal grease. The plate forms a sealed vacuum cavity on the back side. A liquid distributor at the top uniformly wets the outer surface of the plate. Water flows down as a thin film, absorbs the heat conducted through the plate, and evaporates under the low pressure inside the cavity. The vapor is drawn to a vapor‑liquid separator, where any entrained liquid is returned to the distributor. The dry vapor then enters the MV R compressor.
The overall system flow is depicted in the schematic. After evacuation and water charging, the cycle operates as follows:
- Water from the separator bottom is pumped to the distributor and forms a falling film on the battery‑facing plates.
- Heat from the battery evaporates the film; the vapor flows to the separator.
- Dry vapor is compressed by the MV R compressor, increasing its temperature and pressure.
- Hot compressed vapor enters a condenser (cooler) where it is completely liquefied by external cooling water or air.
- Liquid water passes through a pressure‑reducing valve, flashes to the low pressure of the vacuum cavity, and returns to the distributor for reuse.
This design keeps the battery modules at atmospheric pressure and avoids direct contact with the coolant, ensuring safety and reducing coolant inventory.
Simulation Model and Assumptions
We built a steady‑state model in Aspen Plus using the IAPWS‑95 property method for water/steam. The falling‑film evaporator and vapor‑liquid separator are combined into a single EVAP block, assuming complete evaporation of the water feed. The compressor is modeled with an isentropic efficiency η_c. The condenser (COND) is assumed to fully condense the steam, and the throttle valve (VALVE) is isenthalpic. The main equations are:
Compressor power:
$$ W = m_v \Delta H_{id,v} / \eta_c $$
Evaporator cooling load:
$$ Q = m_w \Delta H_{w,e} $$
where \(m_v = m_w\) under the assumption of full evaporation and negligible losses.
Table 1 lists the baseline design parameters used in the simulation.
| Parameter | Value |
|---|---|
| Cooling load | 210 kW |
| Cooling temperature (evaporation) | 303 K (30 °C) |
| Operating pressure (saturation at 303 K) | 4 247 Pa |
| Falling‑film water flow rate | 311 kg h⁻¹ |
| Compressor temperature rise | 15 K |
| Compressor isentropic efficiency | 0.80 |
The simulation results at the baseline point (303 K evaporation, 15 K compression) give a net cooling capacity of 204.492 kW (accounting for slight heat losses) and a compressor power consumption of only 14.227 kW.
Comparison with Conventional Refrigeration
We compared our MV R system against a commercial room‑level chilled‑water air‑conditioning unit that provides 210 kW of sensible cooling (Table 2). The conventional system has a COP of 3.3 (first‑level energy efficiency in China). Table 3 summarizes the performance comparison.
| Parameter | Value |
|---|---|
| Total cooling capacity | 210 kW |
| Sensible cooling capacity | 210 kW |
| Air flow rate | 53 000 m³ h⁻¹ |
| Number of fans | 4 |
| Dimensions (W×D×H) | 2 350 mm×1 100 mm×2 450 mm |
| Weight | 1 000 kg |
| System | Cooling load (kW) | Compressor/Chiller power (kW) | Energy saving vs. conventional |
|---|---|---|---|
| MV R falling‑film system | 204.49 | 14.227 | – |
| Conventional chiller (COP = 3.3) | 210 | 63.636 (theoretical) / 61.967 (actual at 204.49 kW) | 77.04 % |
At the same effective cooling load, our MV R system consumes only 14.227 kW, which is 77 % less than the conventional chiller’s 61.967 kW. The large saving stems from the fact that the MV R cycle avoids a separate refrigerant loop; it compresses only the water vapor generated by the load, rather than driving a complete vapor‑compression cycle through a condenser and expansion device.
Influence of Key Thermodynamic Parameters
We investigated the sensitivity of the MV R system performance to evaporation temperature (cooling temperature) and compressor temperature rise. Figure 5 (original) shows the compressor power as a function of cooling temperature for different temperature rises. To avoid referencing figure numbers, we present the data in Table 4.
| Cooling temp. (K) | ΔT = 10 K | ΔT = 12 K | ΔT = 15 K |
|---|---|---|---|
| 283 | 8.34 | 10.02 | 12.54 |
| 293 | 7.89 | 9.48 | 11.87 |
| 303 | 7.49 | 9.00 | 11.28 |
| 313 | 7.11 | 8.55 | 10.72 |
It is clear that increasing the cooling temperature reduces compressor work because the vapor density rises, leading to a smaller volumetric flow rate, and the pressure ratio decreases. Similarly, a lower compression temperature rise (i.e., smaller pressure ratio) directly reduces work. For example, at a cooling temperature of 303 K, raising the compression rise from 10 K to 15 K increases power by 52.3 % (from 7.49 to 11.28 kW).
The energy‑saving ratio relative to the conventional chiller (COP = 3.3) also varies. Table 5 lists the saving percentage for different combinations.
| Cooling temp. (K) | ΔT = 10 K | ΔT = 12 K | ΔT = 15 K |
|---|---|---|---|
| 283 | 82.8 | 80.4 | 76.2 |
| 293 | 83.4 | 81.1 | 77.0 |
| 303 | 83.9 | 81.7 | 77.7 |
| 313 | 84.4 | 82.2 | 78.3 |
At a fixed compression rise, higher cooling temperature yields slightly better savings. Conversely, increasing the compression rise reduces the saving margin. In practice, the cooling temperature should be chosen as high as permissible by the battery’s temperature limit, and the compressor should be selected for the lowest possible discharge temperature that still provides sufficient driving force for heat rejection.
Economic Analysis
We performed an economic comparison using annualized cost (AC) and payback period (PBP). The initial investment for the MV R system includes the compressor, condenser, vapor‑liquid separator, valves, and pump (Table 6). The reference conventional system includes a chiller, cooling tower, and associated piping; its capital cost is estimated at ¥24 200 based on market data for a 210 kW unit.
| Component | Quantity | Unit price (CNY) |
|---|---|---|
| Steam compressor | 1 | 12 000 |
| Condenser (cooler) | 1 | 3 000 |
| Vapor‑liquid separator | 1 | 1 000 |
| Pressure‑reducing valve | 1 | 400 |
| Circulation pump | 1 | 500 |
| Valves and fittings | 1 set | 300 |
| Total | 17 200 |
The annualized cost model uses an interest rate i = 10 %, a system lifetime n = 15 years, a salvage value S = 20 % of initial cost, maintenance cost Cmc = 10 % of annualized capital, and electricity price Cee = 0.8 CNY/kWh. The annual operating hours are 3 000 h (typical for stationary battery energy storage system applications). The relevant formulas are:
$$ C_{acc} = C_{ccm} \cdot \frac{i(1+i)^n}{(1+i)^n-1} $$
$$ S_a = S \cdot \frac{i}{(1+i)^n-1} $$
$$ C_{rec} = C_{ee} \cdot W \cdot t $$
$$ AC = C_{acc} + C_{mc} – S_a + C_{rec} $$
For the payback period (PBP), we use the discounted cash flow formula:
$$ PBP = \frac{\ln\left(1 – \frac{C_{ccm}}{B_j}(i-r)\right)}{\ln\left(\frac{1+r}{1+i}\right)} $$
where Bj is the annual savings in the first year, and r = 6 % is the discount rate.
Table 7 presents the economic comparison. The MV R system achieves an annual electricity cost of only ¥34 145, compared to ¥148 721 for the conventional system, saving ¥114 576 per year. The annualized cost (AC) is ¥36 525 vs. ¥152 069. The payback period is only 0.163 year (about 2 months), making the investment extremely attractive.
| Item | MV R system | Conventional system |
|---|---|---|
| Initial capital cost Cccm (CNY) | 17 200 | 24 200 |
| Annualized capital Cacc (CNY) | 2 262 | 3 182 |
| Annual maintenance Cmc (CNY) | 227 | 319 |
| Annual electricity cost Crec (CNY) | 34 145 | 148 721 |
| Annualized cost AC (CNY) | 36 525 | 152 069 |
| Annual savings (CNY) | 115 544 | – |
| Payback period PBP (years) | 0.163 | – |
Conclusion
We have proposed and simulated an MV R falling‑film indirect liquid‑cooling system tailored for battery energy storage system thermal management. The system uses water as the coolant, operates under vacuum, and consumes only a fraction of the electricity required by conventional chiller‑based cooling. At a 210 kW cooling load and 303 K evaporation temperature, the compressor power is merely 14.2 kW, representing a 77 % energy saving compared to a first‑level efficiency chiller. Sensitivity analysis shows that higher cooling temperature and lower compression temperature rise further improve efficiency. Economically, the system costs only ¥17 200 in initial investment and yields a payback period of about 2 months, providing a clear economic incentive. This work demonstrates that integrating MV R with falling‑film evaporation is a promising pathway toward low‑energy, cost‑effective cooling for large‑scale battery energy storage system installations.