The global transition towards sustainable transportation is unequivocally centered on the proliferation of electric vehicles (EVs). This shift is fundamentally enabled by the remarkable properties of lithium-ion batteries, which offer high energy density, extended cycle life, and commendable stability. However, the very operation of these electrochemical power sources generates significant amounts of heat. While manageable under steady conditions, the dynamic nature of real-world driving—characterized by rapid acceleration, regenerative braking, and varying loads—poses a formidable challenge to the battery management system. In these dynamic scenarios, power fluctuations can lead to sudden and intense heat generation within the battery pack. If not effectively managed, this can result in localized hotspots, significant temperature gradients, accelerated degradation, and in extreme cases, thermal runaway. Consequently, a sophisticated and responsive BMS is not a luxury but a critical necessity for ensuring the safety, longevity, and reliable performance of EVs.
The primary objective of any thermal battery management system is to maintain the battery pack within an optimal temperature window, typically between 20°C and 45°C, while minimizing the maximum temperature difference (ΔT) between cells to under 5°C. Exceeding these limits compromises efficiency and safety. Various cooling strategies have been explored, including air cooling, liquid cooling, and phase change material (PCM) integration. Among these, direct refrigerant cooling (or direct cooling) has emerged as a highly promising solution for a battery management system. This approach utilizes the refrigerant from the vehicle’s air conditioning loop, allowing it to evaporate and absorb heat directly within cooling plates attached to the battery modules. Its advantages include superior heat transfer efficiency due to the latent heat of vaporization, reduced system complexity and weight by eliminating secondary coolant loops, and excellent performance under high ambient temperatures.
Despite its potential, much of the existing research on direct-cooling BMS has been conducted under simplified, static conditions. Studies often evaluate thermal performance at fixed ambient temperatures (e.g., 25°C or 30°C) and constant discharge rates (e.g., 1C, 2C). While these provide valuable foundational data, they fail to capture the transient, variable, and often harsh thermal loads experienced during actual driving cycles. A significant research gap exists in the experimental investigation of direct cooling battery management systems under authentic dynamic operational profiles. Furthermore, traditional evaluation metrics frequently focus only on the pack’s maximum temperature and overall ΔT, providing an incomplete picture of thermal homogeneity. A more comprehensive analysis should consider temperature uniformity both longitudinally (from the cooling plate contact surface to the top of the cell) and transversely (across the pack parallel to the cooling plate), as these gradients critically impact cell-to-cell aging and performance variance.
This study aims to address these gaps by conducting an experimental investigation into the thermal performance of a direct refrigerant cooling battery management system for a lithium-ion battery pack under three distinct and representative dynamic driving conditions. We evaluate the system’s efficacy not only through traditional metrics but also by introducing detailed assessments of longitudinal and transverse temperature uniformity. When the baseline system is challenged by a strenuous progressive acceleration profile, we identify its limitations and propose a simple yet effective structural enhancement—the integration of aluminum fins—to drastically improve thermal homogeneity. Our work provides critical experimental insights and data to guide the development of more robust and effective thermal management strategies for next-generation electric vehicles.
Experimental Methodology and System Description
1. Direct Cooling Battery Thermal Management Experimental Platform
The core of this investigation is a custom-built direct cooling battery management system experimental platform. The system is designed to mimic a vehicle’s thermal management circuit. Its key components include a variable-speed compressor, a condenser, a cabin evaporator, fans, dedicated cooling plates for the battery, electronic expansion valves (EEVs), and solenoid valves. A crucial feature of this platform is the independent control of the cabin air-conditioning loop and the battery cooling loop via solenoid valves. For the purposes of this study, only the battery cooling loop was activated, utilizing a single cooling plate. The entire platform was housed inside a walk-in constant temperature and humidity environmental chamber, which provided a precisely controlled ambient environment for all tests, with temperature accuracy within ±0.5°C.
The refrigerant, R134a, follows a standard vapor-compression cycle. The compressor circulates the high-pressure vapor, which is condensed into liquid in the condenser. The liquid refrigerant then passes through an EEV, where it undergoes an isenthalpic expansion, dropping in pressure and temperature. This low-pressure, two-phase mixture enters the battery cooling plate. In direct contact with the heat generated by the batteries, the refrigerant absorbs this heat primarily through evaporation (latent heat). The resulting superheated vapor is then sucked back into the compressor to complete the cycle. The opening of the EEV dedicated to the battery cooling plate is controlled by a Proportional-Integral-Derivative (PID) algorithm to maintain a constant superheat of 6°C at the plate’s outlet, ensuring efficient evaporator operation and preventing liquid floodback to the compressor.
2. Battery Pack and Cooling Plate Configuration
The test subject was a battery pack configured from twelve commercial 52 Ah prismatic lithium iron phosphate (LiFePO₄) cells connected in series (12S). The nominal voltage of the pack was 38.4 V. Each cell had dimensions of 148 mm (length) × 28.5 mm (width) × 118 mm (height). The cells were arranged in a 2 (parallel) × 6 (series) layout on top of the primary cooling plate.
The cooling plate, central to the direct cooling BMS, was fabricated from aluminum with dimensions of 450 mm × 295 mm × 23 mm. Internally, it featured a serpentine flow channel, as illustrated in the schematic, constructed from embedded copper tubing (10 mm outer diameter). The channel consisted of 9 passes with a pitch of 50 mm, providing a long, effective flow path for refrigerant evaporation and heat absorption. To minimize thermal contact resistance between the battery pack and the cooling plate, a 2 mm layer of thermally conductive but electrically insulating silicone grease was applied at the interface. Furthermore, to isolate the battery pack from ambient air currents inside the environmental chamber that could distort thermal readings, the entire assembly was enclosed in an insulated box.
In the enhancement phase of the experiment, a simple aluminum fin (360 mm × 179 mm × 2 mm) was added on top of the cooling plate, sandwiched between the plate and the battery pack with the same thermal interface material. The fin’s purpose was to enhance longitudinal heat conduction from the upper portions of the cells down to the actively cooled plate, thereby mitigating vertical temperature gradients.
3. Dynamic Operating Conditions and Data Acquisition
To simulate real-world driving, three dynamic discharge profiles were developed and executed using a Neware battery tester (60V/100A):
- Steady Operation: A constant, low-power discharge at 0.5C (26 A) for 2400 seconds, representing highway cruising.
- Alternating Load Operation: A pulsed profile simulating urban driving with stop-and-go traffic. It consisted of a high-load pulse at 1C (52 A) for 60 seconds, followed by a low-load period at 0.5C for 540 seconds. This cycle was repeated 6 times for a total of 3600 seconds.
- Progressive Acceleration Condition: A strenuous profile representing aggressive driving or rapid acceleration. The discharge rate increased stepwise: 0.5C for 300 s, then 1C for 300 s, and finally 1.5C (78 A) for 300 s, totaling 900 seconds.
All tests were conducted at three ambient temperatures: 25°C, 30°C, and 35°C. Before each discharge test, the battery pack was charged to 100% State of Charge (SOC) using a constant-current constant-voltage (CC-CV) protocol.
Thermal performance was monitored using K-type thermocouples. Six thermocouples were strategically attached to the surfaces of selected battery cells to capture spatial temperature distribution: positions at 1 cm and 9 cm from the bottom along the cell’s central axis, as well as at the geometric center of a cell. An additional thermocouple was placed on the cooling plate surface near the refrigerant outlet. Temperature data, along with battery voltage, current, and refrigerant loop parameters (pressures, temperatures), were recorded at a frequency of 1 Hz.
4. Thermal Performance Evaluation Metrics
A comprehensive set of four key performance indicators (KPIs) was defined to evaluate the direct cooling battery management system:
- Maximum Pack Temperature (Tmax): The highest temperature recorded among all battery surface monitoring points during the test. It indicates the peak thermal stress on the cells.
$$ T_{max} = \max(T_1, T_2, …, T_n) $$ - Maximum Pack Temperature Difference (ΔTpack): The difference between the highest and lowest temperatures recorded anywhere on the battery pack surface. This is the traditional measure of overall thermal uniformity.
$$ \Delta T_{pack} = T_{h} – T_{l} $$
where \( T_h \) is the highest pack temperature and \( T_l \) is the lowest. - Cell Longitudinal Temperature Difference (ΔTcell): The temperature difference between the top (9 cm) and bottom (1 cm) monitoring points on a single cell. This metric specifically evaluates the effectiveness of heat removal in the vertical direction, highlighting potential issues with the thermal path from the cell core to the cooling plate.
$$ \Delta T_{cell} = T_{a} – T_{b} $$
where \( T_a \) and \( T_b \) are the highest and lowest temperatures on that specific cell, respectively. - Pack Transverse Temperature Difference (ΔTh): The temperature difference between cells at the same height (e.g., all points at 1 cm height or all points at 9 cm height) across the pack. This metric reveals temperature imbalances along the flow direction of the refrigerant or due to pack geometry, which can lead to uneven aging.
$$ \Delta T_{h} = T_{i-u} – T_{i-d} $$
where \( T_{i-u} \) is the highest temperature at a given height \( i \), and \( T_{i-d} \) is the lowest temperature at that same height.
The target for an effective battery management system is to maintain \( T_{max} \) < 45°C and \( \Delta T_{pack} \) < 5°C. Our introduced metrics \( \Delta T_{cell} \) and \( \Delta T_{h} \) provide deeper insight into how well these targets are met spatially.
Results and Discussion: Performance Under Dynamic Conditions
1. Steady Operation Condition
The steady 0.5C discharge represents a mild thermal load. The direct cooling BMS demonstrated excellent capability in handling this condition across all tested ambient temperatures. As summarized in Table 1, the maximum pack temperature \( T_{max} \) peaked at 35.9°C in the 35°C ambient test, well within the safe limit. The system’s dynamic response is clear: after an initial temperature rise, the cooling power of the evaporating refrigerant balanced the heat generation, causing temperatures to peak and then gradually decline by the end of the discharge cycle.
The overall uniformity was good, with final \( \Delta T_{pack} \) values of 2.8°C, 3.0°C, and 3.5°C for 25°C, 30°C, and 35°C ambients, respectively. The slight increase with ambient temperature is expected due to the higher starting temperature and reduced temperature differential for heat transfer. Analysis of the detailed metrics showed that both longitudinal (\( \Delta T_{cell} \)) and transverse (\( \Delta T_{h} \)) differences were minimal, staying below 3°C and 1°C, respectively. This confirms that under low, constant loads, the single-sided direct cooling plate provides sufficient and relatively uniform cooling for the entire pack. The refrigerant loop data showed stable compressor power and pressure drop after an initial transient period, indicating efficient and steady operation of the thermal battery management system.
| Ambient Temp. (°C) | Peak Tmax (°C) | Final Tmax (°C) | Final ΔTpack (°C) | Final ΔTcell (Max, °C) | Final ΔTh (Max, °C) |
|---|---|---|---|---|---|
| 25 | 26.7 | 25.2 | 2.8 | < 3.0 | < 1.0 |
| 30 | 30.6 | 27.9 | 3.0 | < 3.0 | < 1.0 |
| 35 | 35.9 | 33.4 | 3.5 | < 3.0 | < 1.0 |
2. Alternating Load Operation Condition
The pulsed discharge profile presented a more demanding test for the thermal battery management system. As seen in the temperature profile, \( T_{max} \) exhibited clear cyclic behavior, rising during each 1C pulse and falling during the subsequent 0.5C period. This demonstrates the system’s ability to respond dynamically to changing loads. Even at the end of the most strenuous pulse under 35°C ambient, \( T_{max} \) reached only 35.7°C. The pack cooled effectively during the low-load phases, ending the test at 31.0°C.
The overall temperature uniformity showed greater variation during this dynamic test. The maximum \( \Delta T_{pack} \) observed during the pulses was higher than in steady operation, but the final values at the end of the test were 6.0°C, 6.7°C, and 6.6°C. While slightly above the 5°C ideal, these values are still within a manageable range for a dynamic cycle, and they were decreasing as the system approached a quasi-steady periodic state. The longitudinal and transverse differences remained well-controlled, with \( \Delta T_{cell} \) stabilizing below 2°C and \( \Delta T_{h} \) below 1°C. This indicates that while the overall pack sees a larger spread between its hottest and coldest points during transients, the internal gradients within and between cells are not severe. The direct cooling BMS successfully managed the thermal transients of an urban driving cycle.
| Ambient Temp. (°C) | Peak Tmax during cycle (°C) | Final Tmax (°C) | Final ΔTpack (°C) | Final ΔTcell (Max, °C) | Final ΔTh (Max, °C) |
|---|---|---|---|---|---|
| 25 | 26.6 | 23.8 | 6.0 | < 2.0 | < 1.0 |
| 30 | 31.0 | 27.5 | 6.7 | < 2.0 | < 1.0 |
| 35 | 35.7 | 31.0 | 6.6 | < 2.0 | < 1.0 |
3. Progressive Acceleration Condition: Identifying a Limitation
The progressive acceleration profile, culminating in a 1.5C discharge, pushed the baseline direct cooling battery management system beyond its effective operating limits, particularly under high ambient temperature. The results, summarized in Table 3, reveal critical shortcomings. Under 35°C ambient, the pack’s \( T_{max} \) soared to 49.8°C, dangerously close to the upper safety threshold and clearly indicative of insufficient cooling capacity for such a high, sustained load. More critically, the thermal homogeneity completely deteriorated. The final \( \Delta T_{pack} \) reached an alarming 16.5°C, vastly exceeding the 5°C target.
Our detailed metrics provide clear insight into the root cause. The longitudinal temperature difference \( \Delta T_{cell} \) for central cells exceeded 11°C. This signifies a major failure in vertical heat conduction: the heat generated in the upper portions of the cells could not be efficiently transported down to the single-sided cooling plate. The cell was effectively insulating itself. Simultaneously, the transverse difference \( \Delta T_{h} \) at the 1 cm height reached 5.9°C, indicating significant temperature variation along the length of the cooling plate, likely due to refrigerant heating and potential dry-out along the serpentine channel. In this strenuous scenario, the baseline BMS design was inadequate, risking accelerated aging and safety hazards due to both high absolute temperature and extreme non-uniformity.
| Ambient Temp. (°C) | Final Tmax (°C) | Final ΔTpack (°C) | Final ΔTcell (Cell-10, °C) | Final ΔTh (at 1cm height, °C) |
|---|---|---|---|---|
| 25 | 34.1 | 7.8 | 6.3 | ~3.0 |
| 30 | 42.1 | 13.6 | 10.5 | ~4.5 |
| 35 | 49.8 | 16.5 | 11.2 | 5.9 |
4. Enhancement with Fins: Dramatic Improvement in Homogeneity
To address the critical issue of poor longitudinal heat conduction identified in the aggressive driving test, we modified the battery management system by integrating a simple aluminum fin between the cooling plate and the battery pack. The fin acts as an extended surface, creating a higher-conductivity thermal pathway from the top of the cells to the actively cooled base.
The improvement was transformative, as detailed in Table 4. Under the same punishing 35°C ambient progressive acceleration test, the fin-enhanced system reduced the peak \( T_{max} \) from 49.8°C to 40.9°C—a reduction of nearly 9°C, bringing it safely back within the optimal range. Most strikingly, the overall pack uniformity was restored: \( \Delta T_{pack} \) plummeted from 16.5°C to just 5.0°C, meeting the critical design target.
The mechanism of improvement is perfectly captured by the detailed metrics. The longitudinal gradient \( \Delta T_{cell} \) for the central cell was slashed from 11.2°C to 4.6°C. This confirms that the fin successfully facilitated the vertical transfer of heat, preventing the upper cell region from becoming a thermal bottleneck. Interestingly, the transverse uniformity also improved significantly, with \( \Delta T_{h} \) at the 1 cm height dropping from 5.9°C to 1.2°C. We hypothesize that the fin, by providing lateral thermal connectivity between cells adjacent to the cooling plate, helped equalize temperatures along the pack, mitigating the “stream-wise” heating effect of the refrigerant. This demonstrates that enhancing longitudinal conduction can have a positive secondary effect on transverse homogeneity in a direct cooling BMS.
The enhancement can be conceptualized through a simplified thermal resistance model. Without fins, the thermal path from the top of a cell to the refrigerant has a high resistance \( R_{cell,vert} + R_{contact} \). The fin introduces a parallel, lower-resistance path. The effective thermal resistance \( R_{eff} \) for heat transfer from the cell top to the coolant is given by:
$$ \frac{1}{R_{eff}} = \frac{1}{R_{cell,vert} + R_{contact}} + \frac{1}{R_{fin}} $$
where \( R_{fin} \) is the conductive resistance of the fin. Even a fin with moderate conductivity significantly reduces \( R_{eff} \), leading to lower temperature differences for the same heat flux \( q \), as governed by:
$$ \Delta T = q \times R_{eff} $$
This model explains the drastic reduction in \( \Delta T_{cell} \) observed experimentally.
| System Configuration | Final Tmax (°C) | Final ΔTpack (°C) | Final ΔTcell (Cell-10, °C) | Final ΔTh (at 1cm height, °C) |
|---|---|---|---|---|
| Baseline (No Fin) | 49.8 | 16.5 | 11.2 | 5.9 |
| With Aluminum Fin | 40.9 | 5.0 | 4.6 | 1.2 |
| Improvement (Δ) | -8.9 °C | -11.5 °C | -6.6 °C | -4.7 °C |
Conclusion
This experimental investigation provides a comprehensive evaluation of a direct refrigerant cooling battery management system under realistic, dynamic electric vehicle operating conditions. We moved beyond static testing to analyze steady, pulsed, and progressively accelerating discharge profiles at relevant ambient temperatures. Our findings clearly demonstrate the context-dependent performance of such a system.
The direct cooling BMS proved to be highly effective and sufficient for maintaining safe and uniform temperatures during steady, low-power operation and dynamic urban-style driving cycles with alternating loads. In these scenarios, the latent heat absorption of the evaporating refrigerant provided responsive and adequate cooling, with all key metrics, including our newly defined longitudinal and transverse uniformity measures, remaining within acceptable bounds.
However, the study crucially identified a limitation of a single-sided cooling plate configuration under strenuous, high-rate continuous acceleration, especially in a hot (35°C) environment. The baseline system failed to maintain thermal control, with the maximum pack temperature reaching 49.8°C and, more critically, the overall temperature difference soaring to 16.5°C. Detailed analysis attributed this primarily to poor longitudinal heat conduction within the cells, leading to severe vertical temperature gradients (\( \Delta T_{cell} > 11°C \)).
The proposed and validated solution—the integration of a simple aluminum fin between the cooling plate and the battery pack—was remarkably successful. This low-complexity modification drastically improved the thermal performance of the battery management system. It reduced the peak temperature by nearly 9°C and, most importantly, brought the maximum pack temperature difference down to 5°C, hitting the critical design target. The fin achieved this by effectively reducing the longitudinal thermal resistance, slashing \( \Delta T_{cell} \) by over 6°C. A secondary, beneficial effect was the improvement in transverse uniformity, as the fin helped equalize temperatures across the pack.
In conclusion, while direct refrigerant cooling is a potent technology for EV thermal management, its design must be robust enough to handle extreme dynamic loads. This work underscores that thermal homogeneity, particularly in the longitudinal direction, can be a critical failure mode. It demonstrates that strategic material addition (like fins) to enhance heat spreading can be a simple, cost-effective, and highly impactful method to upgrade the performance and safety of a direct cooling battery management system. The metrics and methodologies introduced here provide a valuable framework for the future design and validation of advanced, robust thermal BMS capable of ensuring electric vehicle safety and longevity under all real-world driving conditions.