A novel phase change material-based thermal management system with fin-enhanced compartments for lithium-ion batteries optimization using artificial neural network and genetic algorithm

The rapid growth of high-power lithium-ion technologies has intensified the demand for efficient Battery Thermal Management Systems (BTMSs). Excessive heat generation during continuous cycling not only accelerates performance degradation but also triggers safety risks such as thermal runaway. Phase change materials (PCMs) offer a promising solution due to their latent heat storage capability. Yet, their naturally low thermal conductivity, risk of leakage, and limited structural adaptability remain key challenges. In this study, a novel PCM-based BTMS is proposed for LiCoO₂-26,650 cells, introducing a cylindrical battery pack that accommodates ten batteries. A unique innovation of this system is the dedicated PCM enclosure surrounding each cell, entirely separated from the battery surface to prevent leakage, contamination, and potential short-circuit hazards commonly associated with direct-contact PCM configurations. To overcome the PCM's low thermal conductivity, three aluminum fins (one central and two lateral) are embedded in every enclosure. Their lengths, along with the battery-to-wall distance, serve as variable geometric parameters that dictate the system's overall heat dissipation capacity. High-fidelity numerical simulations are performed across geometric configurations to evaluate the thermal behavior under continuous operating conditions. These results are then used to develop a robust Artificial Neural Network (ANN) capable of accurately predicting the average battery surface temperature. Finally, a Genetic Algorithm (GA) is incorporated to conduct single-objective optimization, identifying the optimal fin and enclosure configuration that minimizes the battery temperature. The ANN prediction model achieved an R 2 of 0.977, indicating that the ANN captures 97.7% of the variance in battery temperature. The peak battery temperature drops from 323.137 K (in the No-PCM case) to approximately 310.15 K (in No-fin cases), corresponding to a reduction of about 13 K. In the No-PCM system, the temperature rise reached 25.987 K. When PCM is introduced without fins (No-fin system), the temperature rise is reduced to 13.35 K, representing an improvement of approximately 48.6% in cooling performance compared to the No-PCM configuration. In the final step, when fins were added alongside PCM in the optimized design, the battery temperature was further reduced from about 310.15 K to 307.201 K, which corresponds to an additional reduction of approximately 3 K. In the optimized design, the temperature rise was further reduced to 10.051 K, indicating a 61.3% improvement in cooling performance relative to the No-PCM system and a 24.7% improvement over the No-fin designs.