The optimized real-time energy management strategy for fuel-cell hybrid trucks through dynamic programming

Fuel-cell hybrid trucks (FCHTs), as primary force in fuel-cell vehicles, currently lack comprehensive development of rule-based energy management strategies (EMSs). The current optimization rule of the dynamic programming (DP) algorithm exhibits poor adaptability to changes in operating conditions and battery state of charge (SOC). Therefore, the work proposed real-time rule EMS of DP-optimized FCHTs. First, DP was taken offline to determine the optimal relationship between fuel cell power and demand power under driving conditions and SOCs. Secondly, real-time online recognition was achieved through driving pattern recognition (DPR) controller. A backpropagation neural network, optimized by the northern goshawk algorithm, was used for constructing DPR. Finally, the MATLAB/Simulink simulation showed that the proposed strategy exhibited superior DPR accuracy. Compared to rule-based strategies, significantly lower hydrogen consumption reduced vehicle-operating costs.