Improved multi-dimensional dynamic programming energy management strategy for a vehicle power-split hybrid powertrain

Dynamic programming is a widely used algorithm to solve the optimal control problem for hybrid electric vehicle in the whole driving scenario, but the problems of interpolation error and dimensional disaster have not been completely solved. This study proposes an adaptive adjustment method to solve the interpolation error problem, in which the solution without theoretical error can be obtained at the cost of a tiny driving cycle accuracy. Furthermore, a dynamic equivalent consumption factor calculation method is proposed to analyze the change of energy storage quality in the battery pack and the instantaneous equivalent consumption of the motors. The effectiveness of the improved algorithm is verified on a passenger car with hybrid powertrain using planetary gear mechanism. The results show that the established multi-dimensional dynamic programming algorithm has a relatively stable energy -saving ability. The maximum gap of fuel consumption rate is 3.2% in the four test driving cycles. Besides, there is only a maximum difference of 0.29W between the original driving cycles and the new cycles generated by the adaptive adjustment method. And obvious dynamic equivalent consumption factor changes of 15.7% in WLTC-class3, 10.3% in CLTC-P, 11.8% in FTP75 and 11.0% in NEDC are observed through the proposed calculation method. (c) 2022 Elsevier Ltd. All rights reserved.

Automated summary

This study proposes an adaptive adjustment method and a dynamic equivalent consumption factor calculation method to solve the problems of interpolation error and dimensional disaster in the optimal control problem for hybrid electric vehicles. The effectiveness of the improved algorithm is verified through practical testing.