Case study of intelligent thermal performance optimization for hybrid battery thermal management systems through CFD-driven surrogate modeling and multi-criteria decision analysis

Battery thermal management systems (BTMSs) play a critical role in maintaining the safety, thermal uniformity, and energy efficiency of lithium-ion battery packs operating under demanding conditions. However, the coupled thermal interactions among multiple cooling technologies and design variables make the development of efficient BTMS configurations computationally intensive. This study presents a case study of an artificial intelligence-assisted framework for the thermal design optimization of a hybrid BTMS integrating phase change material (PCM), a micro heat pipe array (MHPA), and liquid cooling. Computational fluid dynamics (CFD) simulations were first employed to generate thermal performance data for various design configurations characterized by four geometric parameters. Two metaheuristic-enhanced multilayer perceptron neural network (MLPNN) models, namely GA-MLPNN and PSO-MLPNN, were then developed to predict the maximum battery temperature (T max ), temperature difference (ΔT), and energy density (ED). The PSO-MLPNN model demonstrated superior prediction capability for ΔT (R 2 > 0.997), whereas GA-MLPNN achieved higher accuracy in estimating T max (R 2 ≈ 0.997) and ED (R 2 > 0.999). The trained surrogate models were integrated with the NSGA-III algorithm to identify Pareto-optimal BTMS designs considering competing thermal and energy objectives. Results revealed that thermally favorable configurations achieved ΔT values of approximately 1.82 °C and T max values around 38.04 °C at the expense of reduced ED, while maximizing ED (≈157 Wh kg −1 ) led to higher thermal gradients (ΔT approaching 4.72-4.87 °C). Finally, the MARCOS multi-criteria decision-making method was employed to rank Pareto solutions under different operational scenarios and identify practical compromise designs. The proposed framework provides an efficient and application-oriented strategy for the thermal design of advanced hybrid BTMSs and offers valuable insights for next-generation electric vehicle battery cooling applications.