Main content area

Fuel economy optimization of power split hybrid vehicles: A rapid dynamic programming approach

Yang, Yalian, Pei, Huanxin, Hu, Xiaosong, Liu, Yonggang, Hou, Cong, Cao, Dongpu
Energy 2019 v.166 pp. 929-938
algorithms, decision making, dynamic programming, energy, fuels, vehicles (equipment)
Fuel economy of hybrid vehicles is affected by their powertrain configurations, powertrain parameters, and energy management strategies. It is most beneficial to optimizing all the three factors simultaneously. However, when the design search space is large, an exhaustive, optimal control strategy, such as dynamic programming (DP), is too computationally expensive. Hence, a faster optimization method with higher computational efficiency and acceptable accuracy is required. Based on the DP approach, an approximate optimization method, called rapid dynamic programming (Rapid-DP), is developed and discussed in this paper. This method effectively reduces the decision-making time (the time can be reduced by a factor of 700, compared to the DP approach) for optimal control. The optimization processes and results are described and then compared with the original DP and PEARS + methods under two different driving cycles: FTP72 and HWFET. In conjunction with particle swarm optimization (PSO), the rapid-DP is leveraged, for the first time, to optimize key powertrain parameters for power split hybrid electric vehicles. Based on two power-split hybrids: Toyota Prius and Prius++, the joint optimization approach is exploited to examine vehicular fuel savings attributed to synergistic parameters optimization and operating-mode increase. The multi-mode configuration with optimal component parameters is demonstrated to be most fuel-efficient, with 6.56% and 3.15% fuel reductions under FTP72 and HWFET cycles, respectively, with respect to the original Prius 2010.