Volume 45: Energy Transitions toward Carbon Neutrality: Part VIII

Reinforcement Learning-Based Energy Management for Fuel Cell Vehicles Incorporating Temperature Dynamics Qilin Shuai, Yiheng Wang, Zhengxiong Jiang, Qingsong Hua

Abstract

The energy management system and thermal control of fuel cell in fuel cell vehicles plays a crucial role in ensuring their stable and efficient operation. This study presents a novel fuel cell powertrain energy management system control strategy considered the temperature fluctuation based on deep reinforcement learning. A comprehensive SIMULINK model, encompassing fuel cell cooling system and stack models, was constructed for the fuel cell, followed by simulation testing under various temperature scenarios. To validate the robustness and stability of the control system, the standard operating conditions – US06 were employed for experimental verification. The experimental results highlight the effectiveness of the designed fuel cell energy management system in achieving transient temperature stabilization. Additionally, the results revealed that stable operation temperatures correlate with reduced hydrogen consumption. Furthermore, it’s noted that fuel cell hydrogen consumption displays substantial variation under uniform operating conditions at varying temperatures. This highlights the key role of temperature in fuel cell performance. These findings serve as valuable reference points for the refinement of energy management system designs with thermal control of fuel cell, contributing to the advancement of fuel cell vehicle technology

Keywords fuel cell vehicle, energy management system, temperature, deep reinforcement learning, PEMFC model, Hydrogen consumption