Optimization using kriging metamodel and CMA-ES to improve temperature uniformity of electric vehicle liquid-cooled cylindrical Li-ion battery BTMS

Abstract Temperature uniformity in Li-ion battery thermal management systems (BTMS) is crucial for achieving thermal safety and optimal performance in electric vehicles. This study proposes a method to enhance the temperature uniformity of a liquid-cooled BTMS by using a kriging metamodel and the covariance matrix adaptation evolution strategy (CMA-ES). The heat generation rate of the battery was determined based on voltage and current data, followed by thermal fluid analysis to evaluate the cooling performance. Key design variables included the outer and inner dimensions of the cooling plate and the coolant flow rate, constrained by the BTMS mass. The kriging metamodel was used to reduce computational costs significantly, while CMA-ES provided a robust optimization method capable of efficiently handling complex, multimodal design spaces. The combined approach resulted in a 21.15% reduction in temperature difference without increasing the BTMS mass, demonstrating the effectiveness of this optimization strategy in improving the thermal management of Li-ion battery modules.