Optimization of thermal non-uniformity challenges in liquid-cooled lithium-ion battery packs using NSGA-II

Abstract Heat removal and thermal management are critical for the safe and efficient operation of lithium-ion batteries and packs. Effective removal of dynamically generated heat from cells presents a substantial challenge for thermal management optimization. This study introduces a novel liquid cooling thermal management method aimed at improving temperature uniformity in a battery pack. A complex nonlinear hybrid model is established through traditional full-factor design and back propagation neural network (BPNN) approximation. This model links input parameters such as the number of baffles, baffle angle, and inlet speed to output parameters including maximum temperature, temperature difference, and pressure drop. Global multiobjective optimization is carried out using the Nondominated Sorting Genetic Algorithm II to sidestep locally optimal solutions. Pareto optimal solutions are sorted using multiple criteria decision-making techniques. Through thermal management optimization, the maximum temperature rise of the battery relative to the initial temperature is controlled within 7.68 K, the temperature difference is controlled within 4.22 K (below the commonly required 5 K), and the pressure drop is only 83.92 Pa. Results presented in this work may help enhance the performance and efficiency of battery-based energy conversion and storage. The optimization technique used in this work helps maximize the benefit of an innovative battery thermal management technique.