A short-term prediction-based efficient optimization power control strategy for heavy-duty hybrid electric vehicle

This study proposes a short-term prediction-based power control strategy using the modified iteration sequential clustering quadratic programming (MISCQP) algorithm for heavy-duty series hybrid electric vehicles (SHEVs). In this strategy, a power preconditioning method is established on the basis of demand power prediction which guarantees the stable power output under transient high-power condition. Through the prediction sequence, MISCQP algorithm is proposed to solve receding horizon problem and achieve real-time control by improving the iteration efficiency. For this purpose, the clustering algorithm is designed to skip the unnecessary short step in the iteration which is too few to obtain sufficient descent. The iteration points in various iteration domains are clustered and the corresponding cluster centers are obtained. Next, the aforementioned clustering results are introduced to improve the termination criterion. The updated criterion turns to skip the short step when the distance of cluster centers of various iteration domains varies within the set threshold. Finally, the performance of the proposed strategy is validated both in simulation and hardware-in-loop tests. The results reveal that the proposed strategy achieves 5.00 %, 5.86 %, 6.27 % less fuel consumption while maintaining stable power output under all the driving cycles. And the average iteration number of proposed strategy is decreased by 10.36 %, 8.47 %, 9.21 %, respectively.