The increasing integration of intermittent renewable energy sources has created a critical demand for efficient and reliable energy storage solutions. Among these, electrochemical energy storage, particularly lithium-ion battery energy storage systems (BESS), has gained significant prominence due to its geographical flexibility and technological maturity. The performance, safety, and lifespan of a lithium-ion battery are intrinsically linked to its operating temperature. Effective thermal management is therefore paramount. While air cooling and phase change material (PCM) cooling are common, immersion liquid cooling offers distinct advantages. By submerging the battery cells or modules directly in a dielectric fluid, the thermal interface resistance is drastically reduced, and the effective heat transfer area is maximized. This results in superior temperature uniformity and control over the maximum temperature of the lithium-ion battery pack, which directly translates to enhanced safety, performance, and longevity.
Although the technical merits of immersion cooling for lithium-ion batteries, such as high heat transfer coefficients and excellent temperature uniformity, are well-documented in research, a comprehensive assessment of its economic viability for large-scale energy storage applications remains relatively scarce. This analysis aims to bridge that gap by conducting a detailed techno-economic evaluation of immersion-cooled lithium-ion battery energy storage systems. The focus will be on comparing different architectural implementations, modeling their lifecycle costs and revenues, and identifying the key factors that influence their financial attractiveness.
1. Technical Architecture and System Configurations
The design of an immersion-cooled lithium-ion battery energy storage system can be categorized based on the scale of the immersion unit and the overall system container. The primary configurations are pack-level immersion and cluster-level immersion, deployed within either cabinet-style units (typical for commercial & industrial applications) or containerized systems (for utility-scale storage).
The core components of a forced-convection immersion cooling system remain consistent across configurations and include the lithium-ion battery packs or clusters, the dielectric immersion fluid, a liquid-to-liquid or liquid-to-air heat exchanger (often part of a chiller unit), circulation pumps, and the associated piping network. A schematic representation of such a system highlights the flow of immersion fluid from the battery tank, through the heat exchanger for heat rejection, and back to the battery tank via pumps.
1.1 Pack-Level vs. Cluster-Level Immersion
The fundamental architectural decision lies in choosing the smallest thermal management unit. In pack-level immersion, each individual battery pack (e.g., a 1P52S configuration) is housed in its own sealed, fluid-filled enclosure within the larger cabinet or container. In cluster-level immersion, multiple packs are combined into a larger battery cluster (e.g., 1P260S), and this entire cluster is submerged in a single, larger tank of immersion fluid.
The structural differences lead to significant variations in initial investment, operational costs, and design complexity, as summarized in the table below.
| Metric | Pack-Level Immersion | Cluster-Level Immersion |
|---|---|---|
| Thermal Management Unit | Individual Battery Pack | Entire Battery Cluster |
| Initial Capital Cost | Higher (more enclosures, seals) | Lower (fewer, larger enclosures) |
| Operating & Maintenance Cost | Potentially Lower (smaller fluid volume, simpler leak isolation) | Potentially Higher (larger fluid volume, complex maintenance) |
| Sealing Complexity | More Complex (many small seals) | Simpler (fewer, larger seals) |
| System Design Complexity | Simpler (modular, standard packs) | More Complex (custom large tank, internal structure) |
| Immersion Fluid Volume | Lower | Higher |
1.2 System Scaling: Cabinet vs. Container
The choice between a cabinet and a container system is primarily driven by the required energy capacity. A typical commercial cabinet might house ~233 kWh of lithium-ion battery capacity, while a standard 20-foot or 40-foot container can be configured for multi-MWh scale, such as a 5 MWh system used in this analysis.
The component specifications scale accordingly. For instance, a 233 kWh cabinet system with pack-level immersion might use five individual pack tanks containing a total of approximately 350 liters of fluid, cooled by a small-capacity chiller (e.g., 5.7 kW). In contrast, a 5 MWh container with cluster-level immersion could use eleven large cluster tanks holding around 8000 liters of fluid, requiring a much larger chiller unit (e.g., 56 kW). The piping network also becomes more complex, evolving from a simple two-stage distribution in a cabinet to a three-stage distribution in a container. The table below outlines the component differences for the four primary system configurations analyzed.
| System Configuration | Battery & Tank Layout | Estimated Immersion Fluid Volume | Chiller Power Requirement |
|---|---|---|---|
| Cabinet (Pack Immersion) | 5 x 1P52S Pack Tanks | 350 L | 5.7 kW |
| Cabinet (Cluster Immersion) | 1 x 1P260S Cluster Tank | 500 L | 5.7 kW |
| Container (Pack Immersion) | 88 x 1P52S Pack Tanks | 4200 L | 56 kW |
| Container (Cluster Immersion) | 11 x 1P416S Cluster Tanks | 8000 L | 56 kW |
1.3 Thermal Load Calculation for System Sizing
Accurate sizing of the cooling system (chiller) is critical for both performance and cost. The total heat load ($P_{total}$) on the thermal management system originates from the lithium-ion battery’s internal heat generation during operation and the environmental heat gain/loss through the container walls. A simplified model for a containerized system is presented below.
The heat generation from the lithium-ion battery pack during discharge can be estimated as:
$$P_{bat} = I^2 R$$
where $I$ is the current and $R$ is the internal resistance.
The sensible heat required to be removed to maintain temperature is:
$$Q = m c_p \Delta T$$
where $m$ is the battery mass, $c_p$ is the specific heat capacity, and $\Delta T$ is the allowable temperature rise.
The cooling power needed from the chiller, considering a safety factor ($k$), is:
$$P_{chiller} = k \left( \frac{P_{bat} – Q/t_{cycle}}{} \right)$$
where $t_{cycle}$ is the charge/discharge duration.
Environmental heat gain, particularly solar radiation, must also be accounted for. The net radiative heat input ($P_{rad}$) depends on the equivalent sky temperature, solar irradiance, and the container’s surface properties. The total cooling load is then:
$$P_{total} = P_{chiller} + P_{rad}$$
For a sample 5 MWh lithium-ion battery container in a warm climate (e.g., Wuhan, China), this calculation yielded a total cooling load of approximately 48.8 kW, leading to the selection of a 56 kW chiller unit. This process ensures the immersion cooling system is adequately sized to handle the worst-case thermal load from the lithium-ion battery system.
2. Economic Modeling and Cost Framework
The economic viability of immersion cooling for a lithium-ion battery energy storage system is assessed by comparing the lifecycle costs against the projected revenue streams. This requires building a detailed cost model and a discounted cash flow (DCF) analysis.
2.1 Cost Breakdown for Immersion-Cooled BESS
The capital expenditure (CAPEX) for an immersion-cooled lithium-ion battery system includes the cost of the battery cells/packs, the power conversion system (PCS), the immersion fluid, the thermal management loop (pumps, chillers, piping), enclosures, and balance of plant (BOS) components like transformers, cabling, and management systems (BMS, EMS).
Cost data is synthesized from industry reports, procurement tenders, and manufacturer quotations. A key metric is the cost per watt-hour (CNY/Wh). The analysis reveals that the lithium-ion battery pack itself constitutes the largest portion of the total system cost, typically around 60%. The immersion fluid and the dedicated thermal management loop together represent a significant added cost compared to traditional air-cooled systems, ranging from approximately 7.6% to 16.5% of the total CAPEX depending on the configuration and fluid type.
The table below provides a comparative cost breakdown for the four main system configurations, assuming a hydrocarbon-based immersion fluid.
| Cost Component | Pack-Imm. Cabinet (233 kWh) | Cluster-Imm. Cabinet (233 kWh) | Pack-Imm. Container (5 MWh) | Cluster-Imm. Container (5 MWh) |
|---|---|---|---|---|
| Lithium-Ion Battery Pack | 12.49 | 11.99 | 260.20 | 241.50 |
| PCS | 1.86 | 1.79 | 38.84 | 36.04 |
| Immersion Fluid | 1.75 | 2.50 | 21.00 | 40.00 |
| Thermal Management Loop | 1.13 | 1.03 | 11.06 | 9.20 |
| BMS, EMS, Transformer, etc. | 4.28 | 4.11 | 89.32 | 82.90 |
| Total Cost (Million CNY) | 21.51 | 21.42 | 420.42 | 409.64 |
| Cost per Watt-hour (CNY/Wh) | 0.923 | 0.919 | 0.841 | 0.819 |
The cost per watt-hour clearly shows the economies of scale (containers are cheaper per Wh than cabinets) and the cost advantage of the cluster-level immersion architecture due to reduced enclosure complexity, despite its higher fluid volume.
The choice of immersion fluid has a dramatic impact on CAPEX. Fluids like engineered fluorocarbons are excellent thermal performers but can cost over CNY 500/L. Hydrocarbon oils, transformer oil, and silicone oils offer a lower-cost alternative, typically ranging from CNY 15-50/L. The total fluid cost for different system configurations and fluid types is a critical variable in the economic model.
2.2 Revenue and Operational Expenditure Model
The primary revenue stream for a grid-connected lithium-ion battery energy storage system is energy arbitrage (peak shaving/valley filling), often constituting about 60% of total revenue. Additional revenues come from ancillary services like frequency regulation and capacity payments.
Annual Revenue ($I_n$): The arbitrage revenue in a given year $n$ depends on the system’s usable energy capacity, which degrades over time.
$$I_n = \eta \cdot Q_n \cdot (P_{discharge} – P_{charge}) \cdot N + I_{ancillary}$$
where:
- $\eta$ is the round-trip efficiency (assumed 90%).
- $Q_n$ is the actual dischargeable capacity in year $n$ (kWh). Battery capacity fade is modeled as: $$Q_n = \left[1 – \frac{0.2(n-1)}{L_{80}}\right] Q_0$$ where $Q_0$ is initial capacity and $L_{80}$ is the cycle life to 80% capacity.
- $P_{discharge}$ and $P_{charge}$ are the discharge (peak) and charge (off-peak) electricity tariffs (e.g., CNY 1.16/kWh and CNY 0.55/kWh).
- $N$ is the number of full cycles per year (e.g., 365).
- $I_{ancillary}$ represents revenue from other grid services.
Annual Operational Expenditure ($C_n$):
$$C_n = C_{electricity} + C_{O&M} + C_{depreciation} + C_{finance} + C_{tax}$$
- $C_{electricity} = \eta \cdot Q_n \cdot P_{charge} \cdot N$
- $C_{O&M}$: Annual operation and maintenance cost, often a percentage of CAPEX (e.g., 4%).
- $C_{depreciation}$: Straight-line depreciation over project life (e.g., 25 years) with a residual value (e.g., 5% of CAPEX).
- $C_{finance}$: Interest payments on debt (e.g., 80% of CAPEX at 3.95% annual rate, amortized over 25 years).
- $C_{tax}$: Corporate income tax and value-added tax.
Annual Net Cash Flow ($F_n$):
$$F_n = (I_n – C_n) + C_{depreciation}$$
This represents the cash generated by the project each year after accounting for all costs and non-cash expenses.
2.3 Financial Performance Indicators
The economic viability is judged using standard project finance metrics:
Net Present Value (NPV): The sum of the discounted annual net cash flows over the project life ($T$), minus the initial CAPEX.
$$NPV = -CAPEX + \sum_{n=1}^{T} \frac{F_n}{(1 + r)^n}$$
where $r$ is the discount rate (e.g., 6.5%). A positive NPV indicates a profitable project.
Internal Rate of Return (IRR): The discount rate ($r^*$) that makes the NPV equal to zero.
$$-CAPEX + \sum_{n=1}^{T} \frac{F_n}{(1 + r^*)^n} = 0$$
A higher IRR denotes a more attractive return on investment.
Payback Period (PBP): The time required for the cumulative cash flows to repay the initial investment.
- Static Payback Period (SPBP): Does not consider the time value of money.
- Dynamic Payback Period (DPBP): Uses discounted cash flows.
2.4 Impact of Immersion Cooling on Lithium-Ion Battery Lifespan
A crucial, often undervalued, economic benefit of immersion cooling for lithium-ion battery systems is lifespan extension. The cycle life ($L_{80}$) of a lithium-ion battery is highly sensitive to its maximum operating temperature ($T_{max}$) and temperature gradient ($\Delta T$). Empirical data shows that operating at 50°C can reduce cycle life by over 70% compared to operating at 25°C. Immersion cooling’s superior thermal control can maintain the lithium-ion battery at a lower, more uniform temperature. Studies suggest immersion cooling can reduce $\Delta T$ by 3-5°C compared to indirect liquid cooling or air cooling, potentially extending battery life by 20-50%. This directly increases $L_{80}$ in the revenue model, delaying the costly replacement of the lithium-ion battery packs and increasing total energy throughput over the project’s life, thereby significantly improving NPV and IRR.
3. Economic Analysis Results and Sensitivity
Applying the economic model to the four configurations with a 25-year project life, a discount rate of 6.5%, and assuming hydrocarbon fluid yields the following results. The analysis includes one major lithium-ion battery pack replacement during the project life for standard cooling, but this can be delayed or avoided with immersion cooling.
| System Configuration | Static PBP (Years) | Dynamic PBP (Years) | NPV (Million CNY) | IRR (%) |
|---|---|---|---|---|
| Pack-Imm. Cabinet | 6.14 | 8.37 | 1.00 | 12.13 |
| Cluster-Imm. Cabinet | 6.09 | 8.28 | 1.02 | 12.26 |
| Pack-Imm. Container | 4.65 | 5.81 | 43.41 | 18.14 |
| Cluster-Imm. Container | 4.47 | 5.53 | 45.72 | 19.04 |
The results clearly demonstrate the economic superiority of larger-scale containerized systems due to lower per-unit costs. Furthermore, within each scale, the cluster-immersion architecture shows marginally better financial metrics than pack-immersion due to its lower CAPEX, despite higher fluid volume. For the pack-immersed container system, the NPV is strongly positive, the IRR (18.14%) significantly exceeds the discount rate (6.5%), and the payback periods are within acceptable ranges, indicating a robust economic case.
When the lifespan-extending effect of immersion cooling on the lithium-ion battery is incorporated (e.g., increasing $L_{80}$ by 20% and reducing replacement frequency), the economic metrics improve further. For the pack-immersed container, NPV can increase by over 10%, and IRR rises accordingly.
3.1 Sensitivity and Break-Even Analysis
The project’s economics are sensitive to several key parameters. A sensitivity analysis measures how much the IRR or NPV changes in response to a change in an input variable.
Sensitivity Coefficient ($S_{AF}$):
$$S_{AF} = \frac{\Delta A / A}{\Delta F / F}$$
where $\Delta A/A$ is the relative change in the output metric (e.g., IRR) and $\Delta F/F$ is the relative change in the input factor (e.g., electricity price difference).
The analysis for the pack-immersed container system reveals the following order of sensitivity (absolute value of $S_{AF}$ for IRR):
- Peak Electricity Price ($S_{AF} \approx +3.5$): The most sensitive factor. A 1% increase in peak price improves IRR by about 3.5%.
- System Structure/CAPEX ($S_{AF} \approx -1.6$): A 1% increase in CAPEX reduces IRR by about 1.6%.
- Valley Electricity Price ($S_{AF} \approx -1.4$): A 1% increase in charging cost reduces IRR by about 1.4%.
- Immersion Fluid Cost ($S_{AF} \approx -0.07$): The least sensitive among these, as it constitutes a smaller portion of total CAPEX.
The high sensitivity to the peak-valley price spread underscores the importance of favorable electricity market conditions for any lithium-ion battery energy storage project’s profitability.
Break-Even on Immersion Fluid Cost: A critical analysis is to determine the maximum allowable immersion fluid cost that still yields an acceptable return. Holding all other factors constant and targeting an IRR equal to the discount rate (6.5%, the break-even point), the model can be solved for fluid cost. For the cluster-immersed container, using a low-cost fluid like transformer oil (~CNY 15/L) yields an IRR of ~21.6%. The IRR remains above the hurdle rate even with fluid costs up to approximately CNY 150/L for this configuration. However, using expensive fluorinated fluids (e.g., >CNY 500/L) can drive the IRR negative, rendering the immersion-cooled lithium-ion battery system economically unviable unless other benefits (like unparalleled safety or ultra-long life) justify the premium.
4. Conclusion
This technical and economic analysis demonstrates that immersion cooling is not only a technically superior thermal management solution for lithium-ion battery energy storage systems but can also be economically viable under the right conditions. The choice between pack-level and cluster-level immersion involves a trade-off between higher initial enclosure costs and lower fluid/system costs, with cluster-level generally offering a slight economic edge. The significant economies of scale make containerized systems far more financially attractive than cabinet-sized units.
The financial success of an immersion-cooled lithium-ion battery project hinges on several factors: achieving a low enough system CAPEX, selecting a cost-effective yet performant immersion fluid, and operating in a market with a sufficient electricity price spread. The single most influential factor for profitability is the peak discharge price. While the upfront cost of the immersion cooling subsystem is higher than that of air cooling, its ability to extend the cycle life of the expensive lithium-ion battery pack by maintaining optimal temperatures creates substantial long-term value, improving key metrics like NPV and IRR.
In summary, for utility-scale lithium-ion battery energy storage, immersion cooling presents a compelling value proposition that balances enhanced safety and performance with a positive economic return, especially when implemented in an optimized cluster-immersed, containerized architecture using mid-range dielectric fluids. As the costs of immersion components and fluids decrease with market adoption and scale, the economic case for this technology is expected to strengthen further.