Thermal management is a cornerstone for the safety, performance, and longevity of traction batteries in new energy vehicles (NEVs). An effective Battery Management System (BMS) must maintain the battery pack within its optimal operating temperature range, typically between 15°C and 35°C, while ensuring minimal temperature variation between cells. Traditional solutions, primarily air-cooling and liquid-cooling, present significant trade-offs. While liquid-cooling is prevalent for its superior performance over air-cooling, its architecture is inherently complex, involving secondary coolant loops, heat exchangers (chillers), and pumps, which increase cost, weight, and potential failure points like coolant leakage.
To address these challenges, this study proposes, designs, and validates an integrated thermal management architecture centered on Refrigerant Direct Cooling (RDC). This innovative approach leverages the vehicle’s existing air-conditioning (AC) refrigerant circuit to directly cool the battery pack, eliminating the intermediate liquid loop and its associated components. The core of this system is the refrigerant cold plate, which replaces the conventional liquid cold plate. This article details the system design principles, presents a comparative thermodynamic analysis against liquid-cooling, and provides extensive experimental validation under various thermal loads. The findings demonstrate that the RDC-based Battery Management System offers a simpler, more efficient, and highly effective solution for battery thermal management.
The proposed Battery Management System integrates the battery thermal management function directly into the vehicle’s heat pump air-conditioning system. The schematic architecture of this integrated system is shown above and consists of the following key components arranged in a vapor-compression cycle: a compressor, a condenser, a liquid receiver (or accumulator), two parallel evaporator branches, and corresponding electronic expansion valves (EXVs).
One branch is dedicated to cabin cooling via the traditional evaporator. The novel second branch is dedicated to battery cooling. High-pressure liquid refrigerant from the condenser is metered through EXV2, transforming into a low-pressure, low-temperature two-phase mixture. This refrigerant is then directed into the refrigerant direct cooling plate attached to the battery pack. Here, the refrigerant absorbs the battery’s waste heat through latent heat of vaporization (boiling), providing highly efficient cooling. The vaporized refrigerant from both branches is then suctioned back into the compressor, completing the cycle. This integrated BMS architecture fundamentally simplifies the system by removing the chiller, coolant pump, coolant reservoir, and associated piping, leading to potential benefits in cost, weight, and reliability.
1. Comparative Analysis: Refrigerant Direct Cooling Plate vs. Liquid Cooling Plate
The heart of the thermal management shift lies in the design of the cooling plate itself. A detailed comparison between the traditional liquid cooling plate (LCP) and the proposed refrigerant direct cooling plate (RDCP) is crucial for understanding the performance and design implications. The structural differences are significant, primarily driven by the thermodynamic properties of the working fluids (liquid coolant vs. boiling refrigerant).
| Parameter | Refrigerant Direct Cooling Plate (RDCP) | Liquid Cooling Plate (LCP) | Comparative Advantage |
|---|---|---|---|
| Cooling Principle | Direct cooling via refrigerant phase change (latent heat absorption). | Indirect cooling via sensible heat transfer to a single-phase coolant. | RDCP offers higher heat flux capability due to latent heat, enabling lower achievable temperatures. |
| System Architecture | Integrated into AC loop. No secondary loop components. | Requires separate coolant loop with chiller, pump, and piping. | RDCP architecture is significantly simpler, with fewer parts, lower mass, and reduced leak potential. |
| Plate Material & Thickness | Aluminum, 1.2 mm typical wall thickness. | Aluminum, 1.5 mm typical wall thickness. | RDCP enables ~20% weight reduction in the plate itself due to thinner walls. |
| Flow Channel Geometry | Narrower, multi-branch, optimized for two-phase flow distribution and boiling. | Wider, simpler channels optimized for low-pressure-drop liquid flow. | RDCP design is more complex but allows for compact packaging. Channel height can be reduced by ~33%. |
| Heating Capability | Requires separate heating method (e.g., PTC film, pulse self-heating). | Can provide heating via warm coolant from a positive temperature coefficient (PTC) heater or recovered waste heat. | LCP has an inherent advantage in heating functionality. RDCP heating requires add-ons. |
| Control Parameter | Superheat at plate outlet (typically 3-7°C). | Coolant inlet temperature and flow rate. | RDCP control is tied to the AC cycle, requiring precise EXV control for optimal performance. |
The design of the RDCP’s internal microchannel or multi-branch pathway is critical for ensuring even refrigerant distribution and effective heat transfer across the entire battery surface area. In contrast to the LCP, where temperature change is primarily linear (sensible), the RDCP maintains a nearly isothermal surface during the boiling process, which is fundamental for superior temperature uniformity in the battery pack. The governing heat transfer equation for the battery-to-plate interface can be expressed as:
$$
\dot{Q}_{batt} = \frac{T_{batt,avg} – T_{sat}}{R_{th,contact} + R_{th,plate}}
$$
Where:
– $\dot{Q}_{batt}$ is the total heat generation from the battery pack (in W).
– $T_{batt,avg}$ is the average battery surface temperature.
– $T_{sat}$ is the saturation temperature of the refrigerant inside the cooling plate.
– $R_{th,contact}$ is the thermal resistance of the interface material (e.g., thermal pad).
– $R_{th,plate}$ is the conductive resistance of the cooling plate material.
For an RDC system, $T_{sat}$ is a function of the system’s low-side pressure, controlled by the compressor and EXV. A key performance metric for the Battery Management System is the maximum temperature difference ($\Delta T_{max}$) across the battery surface, which should be minimized (< 5°C) to prevent cell imbalance and degradation.
2. System Modeling and Thermodynamic Fundamentals
The performance of the integrated RDC-based BMS can be modeled by analyzing the vapor-compression cycle. The cooling capacity provided to the battery ($\dot{Q}_{evap,batt}$) is a critical output. It can be derived from the refrigerant mass flow rate ($\dot{m}_{ref}$) and the enthalpy change across the cooling plate.
$$
\dot{Q}_{evap,batt} = \dot{m}_{ref, batt} \cdot (h_{out, batt} – h_{in, batt})
$$
Where $h_{in, batt}$ and $h_{out, batt}$ are the specific enthalpies of the refrigerant at the inlet and outlet of the RDCP, respectively. The total system capacity is shared between the cabin evaporator and the battery cooling plate:
$$
\dot{Q}_{evap,total} = \dot{Q}_{evap,cabin} + \dot{Q}_{evap,batt} = \dot{m}_{ref, total} \cdot \Delta h_{evap}
$$
The compressor work input, $\dot{W}_{comp}$, determines the system’s Coefficient of Performance (COP). For the battery cooling function specifically, we can define a subsystem COPbatt:
$$
COP_{batt} = \frac{\dot{Q}_{evap,batt}}{\dot{W}_{comp, batt}}
$$
Where $\dot{W}_{comp, batt}$ is the portion of compressor work attributable to the battery cooling load. In an integrated system, this is not trivial to isolate, but system-level optimization aims to maximize overall COP. The EXV controlling the battery branch (EXV2) regulates the degree of superheat ($SH_{batt}$) at the RDCP outlet:
$$
SH_{batt} = T_{out, batt} – T_{sat}(P_{suction})
$$
Maintaining a stable, positive superheat (e.g., 5°C) is essential to ensure full evaporation without liquid refrigerant entering the compressor, while also maximizing the effective heat transfer area of the plate.
3. Experimental Test Bench and Methodology
A full-scale test bench was constructed to validate the thermal performance of the RDC-based Battery Management System. The core setup involved a simulated battery module (pseudo-battery) equipped with electrical heating pads to accurately emulate heat generation during driving or fast charging. A custom-designed aluminum RDCP was clamped onto the pseudo-battery surface with a 3mm-thick thermal interface pad in between. Multiple T-type thermocouples were embedded at strategic locations on both the battery surface and the RDCP to map temperature distribution. The system was charged with R-134a refrigerant, and a programmable EXV controller was used to maintain the target superheat.
To evaluate performance under diverse and demanding conditions, four distinct steady-state heating power levels were applied to the pseudo-battery, simulating different discharge rates or fast-charging scenarios. The test matrix is summarized below:
| Test Case | Battery Heat Load, $\dot{Q}_{batt}$ (kW) | Ambient Temperature, $T_{amb}$ (°C) | Target RDCP Outlet Superheat, $SH_{batt}$ (°C) | Primary Evaluation Objective |
|---|---|---|---|---|
| Case 1 | 1.4 | 25 | 5 | Low-load efficiency and temperature uniformity. |
| Case 2 | 2.4 | 25 | 5 | Moderate-load performance. |
| Case 3 | 3.0 | 25 | 5 | High-load capability and stability. |
| Case 4 | 3.5 | 25 | 5 | Peak-load thermal management limit. |
| Transient Test | 3.5 (from 45°C initial temp) | 25 | 5 | Cool-down rate comparison vs. Liquid Cooling Plate. |
The pseudo-battery’s thermal properties were characterized to ensure realistic simulation. Key parameters used in the analysis are listed below. These parameters are essential for any modeling or simulation of the Battery Management System thermal behavior.
| Thermophysical Property | Symbol | Value | Unit |
|---|---|---|---|
| Battery Specific Heat Capacity | $c_{p,batt}$ | 2000 | J/(kg·K) |
| Battery Thermal Conductivity | $k_{batt}$ | 21 | W/(m·K) |
| RDCP-Battery Contact Area | $A_{contact}$ | 551,324 | mm² |
| Interface Thermal Resistance | $R_{th,contact}$ | 0.0002 | m²·K/W |
| Total Refrigerant Flow Path Length | $L_{path}$ | 1820 | mm |
4. Results and Discussion: Performance of the RDC-Based BMS
The experimental results conclusively demonstrate the advantages of the refrigerant direct cooling approach for the Battery Management System. The steady-state performance across the four heat load cases is summarized in the following table. Key parameters such as system pressures, refrigerant temperatures, and resulting battery temperatures were recorded.
| Performance Metric | Case 1 (1.4 kW) | Case 2 (2.4 kW) | Case 3 (3.0 kW) | Case 4 (3.5 kW) |
|---|---|---|---|---|
| Liquid Line Pressure (Pre-EXV) | 16.1 bar | 16.4 bar | 16.5 bar | 16.6 bar |
| RDCP Inlet Pressure | 5.13 bar | 5.45 bar | 5.83 bar | 6.18 bar |
| RDCP Outlet Pressure / Suction Pressure | 4.65 bar | 4.86 bar | 5.14 bar | 5.56 bar |
| Battery Surface Max Temp ($T_{max}$) | 21.5 °C | 22.8 °C | 25.4 °C | 26.6 °C |
| Battery Surface Min Temp ($T_{min}$) | 16.8 °C | 17.9 °C | 20.8 °C | 21.8 °C |
| Battery Surface Temp Difference ($\Delta T$) | 4.7 °C | 4.9 °C | 4.6 °C | 4.8 °C |
The most critical finding is the exceptional temperature uniformity achieved. Across all tested power levels—from a modest 1.4 kW to a high 3.5 kW heat load—the maximum temperature difference across the battery surface remained consistently below 5°C. This meets and exceeds the stringent uniformity requirement for modern lithium-ion battery packs, a direct result of the near-isothermal cooling provided by the boiling refrigerant within the RDCP. The Battery Management System successfully maintained the average battery temperature well within the optimal 15-35°C window under all conditions, showcasing its robust thermal regulation capability.
The second major test involved a direct head-to-head transient cool-down comparison between the RDC system and a benchmark liquid cooling system (using a standard LCP and a 50/50 glycol-water coolant at 15°C inlet temperature). Both systems started with an initial battery core temperature of 45°C and were subjected to a constant 3.5 kW heat load. The results, plotted as temperature vs. time, were decisive.
The RDC-based system demonstrated a dramatically faster cool-down rate. The temperature profile can be modeled as an exponential decay towards a steady-state value:
$$
T_{batt}(t) = T_{ss} + (T_{initial} – T_{ss}) \cdot e^{-t / \tau}
$$
Where $T_{ss}$ is the steady-state temperature and $\tau$ is the system time constant. The RDC system exhibited a significantly smaller $\tau$, indicating faster thermal response. At the 1200-second (20-minute) mark, the core temperature of the battery cooled by the RDCP had plummeted to 27.3°C. In stark contrast, the battery cooled by the liquid cooling plate was still at 38.2°C. This performance gap of nearly 11°C highlights the superior heat extraction capacity of the phase-change cooling process. The refrigerant’s latent heat absorption allows it to remove large amounts of heat with minimal temperature glide, whereas the liquid coolant relies solely on sensible heat gain, leading to a higher steady-state temperature for the same flow rate and inlet temperature.
5. Conclusion and Outlook for BMS Integration
This study comprehensively validates the refrigerant direct cooling strategy as a highly effective and efficient solution for the thermal management core of a modern Battery Management System (BMS). By integrating battery cooling directly into the vehicle’s heat pump AC circuit, the proposed architecture achieves significant simplification, removing the secondary liquid loop, chiller, and coolant pump. This translates to tangible benefits in weight, cost, component count, and system reliability.
The experimental results confirm two paramount advantages: first, outstanding temperature uniformity ($\Delta T < 5°C$) across a wide range of operational heat loads, which is critical for cell balancing and long-term pack health; and second, a markedly superior cooling power and transient response compared to conventional liquid cooling, enabling better thermal management during high-stress events like fast charging or aggressive driving.
The design and control of the refrigerant direct cooling plate are key to this success. Future work will focus on optimizing the plate’s microchannel geometry for even lower pressure drop and more uniform flow distribution, as well as developing advanced control algorithms for the dual-EXV system to dynamically and efficiently allocate cooling capacity between the cabin and the battery under all climatic and vehicle operating conditions. Furthermore, integrating a heating function, potentially through a reversible heat pump cycle or strategic placement of auxiliary heaters, will be essential for all-climate vehicle operation.
In conclusion, the integration of the cabin air-conditioning and battery thermal management systems through refrigerant direct cooling represents one of the most promising and logical pathways for the evolution of thermal management in new energy vehicles. This approach aligns perfectly with the industry’s goals of simplification, performance enhancement, and cost reduction, establishing a new benchmark for next-generation Battery Management Systems.