Multi-objective design of split-return serpentine channels for liquid-cooled battery thermal management

Effective thermal management is essential for high-capacity lithium-ion batteries used in high-rate energy-storage applications. This study optimizes a liquid-cooled battery thermal management system (BTMS) by combining Response Surface Methodology (RSM) with the Non-dominated Sorting Genetic Algorithm II (NSGA-II). A body-fitted flow channel is designed, and a Box–Behnken scheme defines the simulation matrix. RSM provides surrogate regression models for the objectives—heat removal and hydraulic loss—while NSGA-II identifies the Pareto front to achieve multi-objective optimization. Compared with a linear optimization baseline, the evolutionary search yields an optimal configuration of (channel height h f , inlet angle α f , annular-channel thickness δ f ) = (5.48 mm, 50°, 3.00 mm). At this setting, the BTMS achieves a maximum module temperature of 301.03 K and a maximum intra-module temperature difference of 2.04 K. The system also maintains the maximum module temperature below 302.5 K and the temperature difference under 2 K when the number of cells increases to 21, demonstrating robustness and scalability. These results show that the RSM–NSGA-II framework enhances cooling performance and predictive fidelity relative to linear methods, and it provides a practical route to balancing temperature uniformity and pumping losses in liquid-cooled battery packs.