Multi-level optimization of thermal management systems with compact large-bend configurations in hybrid electric vehicles

The thermal management system design for specialized hybrid vehicles presents unique challenges, particularly due to the presence of a 90° bend in the airflow path. This study proposes a multi-level modelling and optimization methodology aimed at meeting cooling requirements while minimizing airflow pressure drop across all system components, despite spatial constraints. The integrated modelling framework incorporates a zero-dimensional heat transfer model, a one-dimensional model for non-uniform airflow, and three-dimensional computational fluid dynamics simulations. Coupled with the NSGA-II algorithm, the optimization addresses multiscale structures, including fin design, radiator arrangement, radiator dimensions, and air duct size. The 90° bend induces substantial non-uniform flux across the radiator's airflow face, reducing heat transfer performance by up to 10 % compared to uniform airflow. However, adjusting the radiator thickness and expanding the air duct can mitigate this non-uniformity, reducing it by up to 40 %. Results indicate that the total pressure drop initially decreases and then increases as the air duct expands and the radiator thins, due to the interplay of pressure drop across components, allowing for the identification of an optimal radiator height. This optimization framework and the insights gained provide valuable guidance for developing integrated thermal management systems for large-bend configurations in hybrid electric vehicles.