Optimization of fuel cell thermal management system based on back propagation neural network

Abstract

Two thermal management control strategies, namely flow following current and power mode and back propagation neural network auto-disturbance rejection method, were proposed to solve significant temperature fluctuation problems, long regulation time, and slow response speed in fuel cell thermal management system variable load. The results show that the flow following current and power control strategy can effectively weaken the coupling effect between pump and radiator fan and significantly reduce the overshoot and adjustment time of inlet and outlet cooling water temperature and temperature difference reactor. Although the control effect of the neural network and strategy is insufficient under maximum power, the overall control effect is better than that of the flow following the current control strategy.