Computationally Efficient Energy Management for Hybrid Electric Vehicles Using Model Predictive Control and Vehicle-to-Vehicle Communication
Fengqi Zhang, Xiaosong Hu, Teng Liu, Kanghui Xu, Ziwen Duan, Hui Pang
Abstract
Rapidly-evolving connected vehicle technologies offer growing opportunities to improve the performance of energy management for hybrid electric vehicles (HEVs). In this context, a computationally efficient energy management approach based on a model predictive control (MPC) framework is proposed to obtain the optimal torque split and gearshift for a parallel HEV. The velocity is predicted by exploiting the information through V2V (Vehicle to Vehicle) communication. The energy management problem is then reformulated by introducing equivalent consumption minimization strategy (ECMS) into MPC framework. A penalty index is directly incorporated into the objective function to avoid frequent gearshift activities considering both fuel economy and drivability. To derive an analytical optimal solution, the energy management is further simplified by assuming that the gearshift command keeps constant over each prediction horizon. As a result, the torque split and gearshift are jointly optimized in the same framework. A sensitivity study for different parameters is conducted, and an equivalence factor (EF) adaptation law is also devised for the ECMS-based MPC. Lastly, simulations are performed in standard cycles and three predicted cycles over different prediction horizons, verifying that the ECMS-based MPC produces a promising computational efficiency, relative to traditional dynamic programming (DP)-MPC, while fuel economy globally converges to that produced by DP.