Integrated Thermal Management Methodology Based on Proximal Policy Optimization With Memory Function for Battery Electric Vehicles

Thermal management strategy (TMS) plays a pivotal role in ensuring operational safety, energy efficiency in battery electric vehicles (BEVs) by maintaining critical components within optimal thermal operating ranges. This study proposes the integrated TMS for BEVs by integrating proximal policy optimization (PPO) with a gated recurrent unit (GRU)-based memory mechanism. By leveraging GRU-enhanced PPO (HPPO), proposed TMS captures long-term temporal dependencies in thermal dynamics, enabling adaptive coordination of cooling components while maintaining battery, motor, and cabin temperatures within optimal ranges. Simulation results demonstrate that the HPPO-based TMS reduces energy consumption by 25.76% compared to conventional rule-based TMS, achieving a balance between thermal stability and energy efficiency. Further analysis reveals a significant improvement in root mean square error of controlled components’ temperature. Moreover, in the three cases where the sensitivity of the proposed TMS is verified, the proposed TMS achieved energy improvement of 23.61%, 25.19% and 21.54% compared to traditional rule-based TMS, respectively. These advancements highlight the potential of memory-augmented reinforcement learning for multi-objective thermal optimization in next-generation BEVs, offering a scalable solution for extreme climate adaptability.