Driving information process system-based real-time energy management for the fuel cell bus to minimize fuel cell engine aging and energy consumption

The traditional velocity prediction method has an unsatisfactory prediction effect on the starts and stops phase of the bus urban driving cycle, which caused the fuel cell engine large fluctuation of output power, life aging, and consequently low energy efficiency of the fuel cell bus (FCB). In addition, the load weight of the FCB fluctuates more than that of traditional passenger cars or logistics vehicles. From the viewpoint of frequent starts, stop and random change of load weight during the driving process, this paper proposes a driving information process system (DIPS) based real-time energy management, which adds pattern recognition to the long short-term memory (LSTM) neural network-based velocity prediction to improve starts and stops phase prediction precious. The prediction results show that this method can significantly improve the prediction effect of the low-velocity urban bus driving cycle. Moreover, the change of load weight caused by the change of passengers is considered to the energy management. The simulation results show that the DIPS-based real-time energy management equivalent hydrogen consumption is 2.5% higher than that of the traditional MPC-based energy management. Finally, the effectiveness and real-time performance of the energy management are verified by hardware in the loop test.