Abstract:

An energy management strategy based on equivalent factor Q-learning algorithm was proposed to improve the fuel economy of fuel cell hybrid electric vehicles (FCHEVs) and maintain the battery energy balance. The objective function of minimizing equivalent hydrogen consumption and maintaining battery state of charge (SOC) was constructed to establish the energy flow conversion balance model of FCHEVs power source, and the equivalent factor of hydrogen consumption was obtained through the energy conversion balance mechanism. The transfer probability matrix of required power was solved under urban dynamometer driving schedule + world light vehicle test cycle (UDDS+WLTC) conditions, and the output power of fuel cells and batteries was optimized offline by Q-learning algorithm. The forward simulation model was established based on MATLAB/Simulink platform, and the vehicle simulation was carried out under different working conditions. The results show that the equivalent of the hydrogen consumption per 100 km is 0.730 kg under WLTC cycle conditions in the proposed strategy, which is close to that based on the dynamic programming (DP) control strategy, and the SOC is kept within a reasonable range, which verifies the effectiveness of the proposed strategy. And the adaptability of the proposed control strategy is also verified in the actual working conditions of Xining City.

Key words: fuel cell hybrid electric vehicle (FCHEV), equivalent factor, Q-learning algorithm, energy management