In this study, I provide a technical and economic analysis of liquid immersion cooling for lithium-ion battery energy storage systems (BESS). I begin with a technical comparison of four system configurations: pack-immersed cabinets, cluster-immersed cabinets, pack-immersed containers, and cluster-immersed containers. I then develop a comprehensive economic model incorporating initial investment, operational costs, revenue streams, and battery degradation. The results indicate that immersion-cooled BESS can achieve favorable economic metrics, with a static payback period of 4.65 years, a dynamic payback period of 5.81 years, a net present value of 4.3409 million CNY, and an internal rate of return of 18.14% for a pack-immersed container system. Sensitivity analysis reveals that peak electricity pricing has the most significant impact on economic performance. The choice of immersion liquid significantly affects viability, with transformer oil, silicone oil, and hydrocarbon fluids showing promising results, while fluorinated liquids prove economically unattractive due to high costs. This work demonstrates that immersion-cooled BESS can be economically competitive when system design and operational parameters are optimized.
Electrochemical energy storage has emerged as a leading technology among various storage methods due to its minimal geographic constraints and mature industrial development. Among cooling approaches, immersion liquid cooling offers distinct advantages over air cooling and phase-change material cooling. By allowing direct contact between batteries and the cooling fluid, thermal resistance is reduced, heat dissipation area is increased, and both the maximum battery temperature and temperature uniformity are effectively improved.
Previous research has demonstrated the thermal management capabilities of immersion cooling. Studies have shown that immersion cooling outperforms forced air cooling for lithium-ion battery packs. Experimental investigations have confirmed that immersion cooling can limit the maximum temperature and temperature gradient of battery packs to 40℃ and 3℃ respectively. Temperature is a critical factor affecting battery lifespan, with higher temperatures accelerating capacity fade. The potential of immersion cooling to extend battery life makes a full lifecycle economic analysis essential.
While much existing literature focuses on the thermal performance of immersion cooling or the economic analysis of conventional BESS, there is a notable gap in comprehensive economic assessments of immersion-cooled systems. This work aims to address this gap by developing an economic analysis methodology specifically for immersion-cooled BESS, including sensitivity analysis of key uncertainty factors.
Technical Configuration Analysis of Immersion-Cooled BESS
System Architecture and Components
The forced-convection immersion cooling system is considered for two application scenarios: BESS cabinets (typically used for commercial and industrial storage) and BESS containers (used for large-scale storage). Two technical approaches are evaluated: pack immersion and cluster immersion. The system components include battery modules, immersion fluid, heat exchange equipment, and piping systems.
Pack Immersion vs. Cluster Immersion
Pack immersion involves each battery pack being individually cooled by immersion, with packs isolated from each other. Cluster immersion involves immersing an entire battery cluster in the cooling fluid. Table 1 summarizes the qualitative differences between these approaches.
| Indicator | Pack Immersion | Cluster Immersion |
|---|---|---|
| Thermal management unit | Battery pack | Battery cluster |
| Initial investment cost | Higher | Lower |
| Operation and maintenance cost | Lower | Higher |
| Sealing complexity | More complex | Simpler |
| Design complexity | Simpler | More complex |
| Immersion fluid volume | Less | More |
Heat Generation Model for BESS
Using Wuhan, China as the geographic reference, I calculate the total heat load of the battery system by considering both internal heat generation from the batteries and heat exchange with the external environment. The key physical quantities and their calculation formulas are presented in Table 2.
| Quantity | Symbol | Formula | Value |
|---|---|---|---|
| Battery container heat generation power | P1 | $$P_1 = I^2R$$ | 43200 W |
| Temperature rise heat of battery container | Qx | $$Q_x = cm\Delta t$$ | 91584 kJ |
| Cooling power of battery container | P2x | $$P_{2x} = k\left(P_1 – \frac{Q_x}{t_2}\right)$$ | 39624 W |
| Equivalent sky temperature | ts | $$t_s = 0.0552(t_e + 273)^{1.5} – 273$$ | 24℃ |
| Solar radiation equivalent temperature | tsol,eq | $$t_{sol,eq} = \frac{\rho I_H}{\alpha_e}$$ | 48.35℃ |
| Radiative heat transfer coefficient | αer | $$\alpha_{er} = \varepsilon \times 5.67 \times \frac{\left(\frac{t_e + 273}{100}\right)^4 – \left(\frac{t_s + 273}{100}\right)^4}{t_e – t_s}$$ | 5.0 |
| Sky radiation equivalent temperature | tsky,eq | $$t_{sky,eq} = \frac{\alpha_{er}(t_e – t_s)(1 – C_H H – C_M M)}{\alpha_e}$$ | 7.85℃ |
| Sky radiation correction factor | ε | $$\varepsilon = 1 + \frac{t_{sky,eq} – t_{sol,eq}}{t_i – t_e}$$ | 7.75 |
| Net radiative heat input | Pnet | $$P_{net} = \varepsilon K_T S \Delta t_2$$ | 5900 W |
| Total heat generation of battery container | P | $$P = P_{2x} + P_{net}$$ | 48826 W |
Based on the total heat calculation, I determine that the BESS cabinet system requires a 5.7 kW air-cooled chiller, while the BESS container system requires a 56 kW air-cooled chiller. The immersion fluid volumes for the four configurations are set as: pack-immersed cabinet: 350 L; cluster-immersed cabinet: 500 L; pack-immersed container: 4200 L; cluster-immersed container: 8000 L. These volumes are used to calculate the total immersion fluid cost.
Cost and Economic Analysis Model for Immersion-Cooled BESS
Cost Analysis of Immersion-Cooled BESS
The cost per watt-hour (CNY/Wh) is the ratio of total system cost to total battery capacity. Component prices are sourced from industry data and procurement information. Table 3 shows the cost breakdown for four immersion-cooled BESS configurations.
| Component | Pack-immersed cabinet (233 kWh) | Cluster-immersed cabinet (233 kWh) | Pack-immersed container (5 MWh) | Cluster-immersed container (5 MWh) |
|---|---|---|---|---|
| Battery module | 12.49 | 11.99 | 260.20 | 241.50 |
| PCS | 1.86 | 1.79 | 38.84 | 36.04 |
| EMS | 0.37 | 0.36 | 7.77 | 7.21 |
| Transformer | 1.12 | 1.07 | 23.30 | 21.63 |
| Assembly | 0.56 | 0.54 | 11.65 | 10.81 |
| Cables | 0.56 | 0.54 | 11.65 | 10.81 |
| BMS | 1.68 | 1.61 | 34.95 | 32.44 |
| Immersion fluid (hydrocarbon) | 1.75 | 2.50 | 21.00 | 40.00 |
| Thermal management circuit | 1.13 | 1.03 | 11.06 | 9.20 |
| Total price | 21.51 | 21.42 | 420.42 | 409.64 |
Battery modules account for approximately 60% of total system cost, the largest single component. PCS contributes 8-10%, and BMS contributes 7-9%. The combined cost of immersion fluid and thermal management circuit ranges from 7.6% to 16.5% of total system cost across the four configurations, with immersion fluid consistently costing more than the thermal management circuit.
In terms of unit cost (CNY/Wh), pack immersion is generally more expensive than cluster immersion due to the additional cost of individual battery pack enclosures. Due to economies of scale, BESS containers achieve lower unit costs than BESS cabinets.
Different immersion fluid types significantly affect system cost. Four common fluids are considered: fluorinated liquid (570 CNY/L), hydrocarbon fluid (50 CNY/L), transformer oil (15 CNY/L), and silicone oil (18 CNY/L). Cluster immersion systems require more fluid than pack immersion, resulting in higher total fluid costs. Transformer oil and silicone oil show comparable costs, while hydrocarbon fluid is moderately higher. Fluorinated liquid costs are substantially higher than all other options.
Economic Analysis Model for Immersion-Cooled BESS
The core of the economic analysis is to calculate revenue and costs over the operational lifecycle, determining net profit and net cash flow. Key economic metrics are:
- Payback time (PBT): The time required for net revenue to recover the initial investment. Both static PBT (SPBT) and dynamic PBT (DPBT) are calculated.
- Net present value (NPV): The sum of discounted future cash flows minus the initial investment.
- Internal rate of return (IRR): The discount rate at which NPV equals zero.
The SPBT is calculated as:
$$SPBT = Z – 1 + \frac{|F_{Z-}|}{F_Z}$$
where Z is the year when cumulative net cash flow becomes positive, FZ− is the cumulative net cash flow at year Z-1, and FZ is the net cash flow in year Z.
The DPBT is calculated as:
$$DPBT = Z_H – 1 + \frac{|F_{H-}|}{F_H}$$
where ZH is the year when cumulative discounted net cash flow becomes positive, FH− is the cumulative discounted net cash flow at year ZH-1, and FH is the discounted net cash flow in year ZH.
The NPV is calculated as:
$$NPV = \sum_{t=0}^{n} F_{nt}(1 + i)^{-t}$$
where n is the operation period in years, t is the year index, Fnt is the net cash flow in year t, and i is the discount rate.
The IRR satisfies:
$$\sum_{t=0}^{n} F_{nt}(1 + IRR)^{-t} = 0$$
Battery capacity degradation is considered in the economic analysis. The actual discharge capacity in year n is given by:
$$Q_n = \left[1 – \frac{0.2(n – 1)}{10}\right] Q_0$$
where Q0 is the initial capacity and Qn is the capacity in year n.
Revenue Model
Revenue from the battery energy storage systems includes peak-valley arbitrage (approximately 60% of total revenue), as well as peak shaving, frequency regulation, and capacity compensation (approximately 40%). The annual operating revenue In is:
$$I_n = \eta Q_n P_f N$$
where η is the charge/discharge efficiency (90%), Pf is the discharge electricity price (1.16 CNY/kWh), and N is the annual number of charge/discharge cycles (365 for daily cycling).
Peak shaving revenue Ctf is:
$$C_{tf} = \eta W N \Delta P$$
where ΔP is the peak-valley price difference and W is the system power.
Frequency regulation revenue Ctp is:
$$C_{tp} = \eta Q W N$$
where Q is the demand response electricity price.
At the end of operation, a residual value of 5% of system cost is recovered.
Expenditure Model
Annual operating expenditure includes charging costs, depreciation, operation and maintenance (O&M) costs, and financial costs. The depreciation cost Czj assuming a 25-year lifespan and 5% residual value is:
$$C_{zj} = C_{js} \times 0.95 / 25$$
where Cjs is the construction cost.
O&M cost Cyw is:
$$C_{yw} = k C_{js}$$
where k is the O&M cost coefficient (4% in this study).
Assuming 80% of the project cost is financed through a bank loan at interest rate R (3.95%) over P years (25 years) with equal installments, the financial cost Ccw is:
$$C_{cw} = \frac{80\% C_{js} R (1 + R/12)^{12P}}{(1 + R/12)^{12P} – 1}$$
Taxes Cs (including income tax and value-added tax) are also included. The annual operating expenditure Cn is:
$$C_n = \eta Q_n P_c N + C_{yw} + C_{zj} + C_{cw} + C_s$$
where Pc is the charging electricity price (0.55 CNY/kWh).
Profit and Net Cash Flow
Annual profit Jn is:
$$J_n = I_n – C_n$$
Annual net cash flow Fn is:
$$F_n = J_n + C_{zj}$$
Economic Analysis and Discussion
Baseline Economic Performance
Table 4 summarizes the economic indicators for the four immersion-cooled BESS configurations, calculated using a discount rate of 6.5%.
| Indicator | Pack-immersed cabinet | Cluster-immersed cabinet | Pack-immersed container | Cluster-immersed container |
|---|---|---|---|---|
| SPBT (years) | 6.14 | 6.09 | 4.65 | 4.47 |
| DPBT (years) | 8.37 | 8.28 | 5.81 | 5.53 |
| NPV (million CNY) | 10.02 | 10.20 | 434.09 | 457.19 |
| IRR (%) | 12.13 | 12.26 | 18.14 | 19.04 |
For the pack-immersed container system, the SPBT is 4.65 years, DPBT is 5.81 years, NPV is 4.3409 million CNY, and IRR is 18.14%. All metrics fall within economically reasonable ranges (SPBT and DPBT under 10 years, NPV > 0, IRR > discount rate). Container systems outperform cabinet systems due to lower unit costs, and cluster immersion outperforms pack immersion due to lower initial investment.
Impact of Battery Life Extension on Economics
Immersion cooling improves battery lifespan by maintaining lower and more uniform temperatures. Studies indicate that immersion cooling can reduce battery capacity fade by up to 20% compared to air cooling, which reduces the frequency of battery replacement. Table 5 shows the economic impact of considering this lifespan extension for the pack-immersed container system.
| Indicator | Air-cooled battery lifespan | Immersion-cooled battery lifespan |
|---|---|---|
| SPBT (years) | 4.65 | 4.61 |
| DPBT (years) | 5.81 | 5.72 |
| NPV (million CNY) | 434.09 | 484.50 |
| IRR (%) | 18.14 | 18.54 |
When battery lifespan extension is considered, NPV increases by 11.6%, and IRR improves. Compared to indirect liquid cooling and air cooling, immersion cooling shows IRR improvements of 8.8% and 28.0% respectively, demonstrating that its profitability can offset higher initial costs.
Sensitivity Analysis
Peak and Valley Electricity Prices
Tables 6 and 7 show the sensitivity of SPBT and IRR to peak and valley electricity prices.
| Peak price (CNY/kWh) | Valley price (CNY/kWh) | ||
|---|---|---|---|
| 0.52 | 0.55 | 0.58 | |
| 1.10 | 5.04 | 5.37 | 5.76 |
| 1.16 | 4.40 | 4.65 | 4.93 |
| 1.22 | 3.90 | 4.09 | 4.32 |
| Peak price (CNY/kWh) | Valley price (CNY/kWh) | ||
|---|---|---|---|
| 0.52 | 0.55 | 0.58 | |
| 1.10 | 16.27 | 14.87 | 13.43 |
| 1.16 | 19.49 | 18.14 | 16.77 |
| 1.22 | 22.61 | 21.30 | 19.97 |
A wider peak-valley price difference significantly improves economic performance. When the peak price increases to 1.22 CNY/kWh and the valley price decreases to 0.52 CNY/kWh, the SPBT drops from 4.65 to 3.90 years and the IRR increases from 18.14% to 22.61%.
Sensitivity Coefficients
Tables 8 and 9 show the sensitivity analysis factor (SAF) values for the uncertainty factors.
| Uncertainty factor | SAF value | Absolute value ranking |
|---|---|---|
| Peak electricity price | -2.39 | 1 |
| Valley electricity price | 1.10 | 3 |
| System structure cost | 1.23 | 2 |
| Immersion fluid cost | 0.02 | 4 |
| Uncertainty factor | SAF value | Absolute value ranking |
|---|---|---|
| Peak electricity price | 3.54 | 1 |
| Valley electricity price | -1.39 | 3 |
| System structure cost | -1.59 | 2 |
| Immersion fluid cost | -0.07 | 4 |
Peak electricity price has the highest SAF absolute value for both SPBT and IRR, making it the most influential uncertainty factor. System structure cost is the second most influential factor, while immersion fluid cost has a relatively minor impact in this analysis.
Impact of Immersion Fluid Type and System Structure
Table 10 and 11 show the impact of immersion fluid type and system structure on SPBT and IRR.
| Immersion fluid type | Pack-immersed cabinet | Cluster-immersed cabinet | Pack-immersed container | Cluster-immersed container |
|---|---|---|---|---|
| Fluorinated liquid | >25 | >25 | 13.31 | >25 |
| Hydrocarbon fluid | 6.14 | 6.09 | 4.65 | 4.47 |
| Transformer oil | 5.57 | 5.30 | 4.41 | 4.02 |
| Silicone oil | 5.63 | 5.38 | 4.43 | 4.06 |
| Immersion fluid type | Pack-immersed cabinet | Cluster-immersed cabinet | Pack-immersed container | Cluster-immersed container |
|---|---|---|---|---|
| Fluorinated liquid | -4.59 | -12.75 | 5.81 | -1.23 |
| Hydrocarbon fluid | 12.13 | 12.26 | 18.14 | 19.04 |
| Transformer oil | 13.86 | 14.81 | 19.37 | 21.59 |
| Silicone oil | 13.68 | 14.54 | 19.25 | 21.32 |
Transformer oil, silicone oil, and hydrocarbon fluids all show good economic performance, with IRR values of 21.59%, 21.32%, and 19.04% respectively for the cluster-immersed container configuration. Fluorinated liquid performs poorly with negative IRR values, making it economically unattractive.
Figure 12 summarizes the general relationship between immersion fluid price and IRR across a broader price range of 10-1200 CNY/L.
Table 12 provides a price overview of different immersion fluid categories.
| Category | Name | Price range (CNY/L) |
|---|---|---|
| Mineral oil | Transformer oil | 10-30 |
| Mineral oil | Other mineral oils | 20-50 |
| Silicone oil | 5 cSt silicone oil | 20-50 |
| Silicone oil | 10 cSt silicone oil | 20-40 |
| Silicone oil | 20 cSt silicone oil | 15-30 |
| Refrigerant | R134a | 60-100 |
| Ester | Synthetic ester oil | 30-80 |
| Ester | Pentaerythritol ester | 40-100 |
| Synthetic oil | PAO | 50-200 |
| Synthetic oil | GTL | 100-250 |
| Fluorinated liquid | NOVEC 7100 | 400-600 |
| Fluorinated liquid | NOVEC 7200 | 580-700 |
| Fluorinated liquid | NOVEC 7500 | 800-1200 |
As immersion fluid price increases, IRR decreases. If the fluid price does not exceed 150 CNY/L, the immersion-cooled BESS retains good economic viability when considering battery life extension. Above this price threshold, IRR may fall below the discount rate, making cost recovery challenging. System structure also matters: cluster-immersed containers have 10.0% lower unit cost than cluster-immersed cabinets, resulting in a 55.3% improvement in IRR.
Conclusions
This study provides a comprehensive technical and economic analysis of immersion cooling for lithium-ion battery energy storage systems. The main findings are:
- Cost structure: Battery modules account for approximately 60% of total system cost for immersion-cooled BESS, making this the most significant cost component.
- Economic viability: Under the assumptions of this study, immersion-cooled BESS can achieve economically reasonable performance. For the pack-immersed container system: SPBT = 4.65 years, DPBT = 5.81 years, NPV = 4.3409 million CNY, and IRR = 18.14%.
- Most sensitive factor: Peak electricity price has the highest sensitivity coefficient for both SPBT (SAF = -2.39) and IRR (SAF = 3.54), making it the most influential uncertainty factor for economic performance.
- Immersion fluid selection: Transformer oil (IRR = 21.59%), silicone oil (IRR = 21.32%), and hydrocarbon fluids (IRR = 19.04%) show good economic performance. Fluorinated liquids are economically unattractive with negative IRR values.
- System structure: Lower unit cost configurations achieve better economic performance. The cluster-immersed container system has 10.0% lower unit cost than the cluster-immersed cabinet system, resulting in a 55.3% improvement in IRR.
- Battery life extension: When the beneficial effect of immersion cooling on battery lifespan is considered, NPV improves by 11.6% compared to the case without lifespan consideration.
This work demonstrates that immersion-cooled BESS can be economically competitive when appropriate system design and operational parameters are selected. Future research should investigate economic performance under different operational profiles, evaluate additional immersion fluid types, and consider long-term maintenance costs including fluid replacement and disposal.