Energy Management Strategy for Plug-in Hybrid Vehicles Based on Model Predictive Control and Local Encryption Dynamic Programming

A model predictive control (MPC) energy management strategy (EMS) coupled with offline dynamic programming (ODP) based on historical average vehicle speed, ODP-MPC, is proposed in this paper. The effectiveness of ODP-MPC is verified using historical traffic flow datasets from the open literature. The simulation results show that ODP-MPC can reduce fuel consumption by 1.1% to 7.3% compared to MPC. Moreover, at the prediction area H_p=3(3s), the fuel consumption of ODP-MPC is only 2.1% higher than that of the DP algorithm. This indicates that ODP-MPC can approximate the theoretical fuel economy. As for the computational effort, the online computation time of ODP-MPC is improved by 6.3%~22.9% compared to MPC, but still less than the 1s time step. Reducing the number of grid cells (m) or increasing the distance step (dis_tf) in offline DP reduces the offline computational cost and the fuel economy of ODP-MPC. The coupled locally encrypted meshing strategy (LEMS) in ODP-MPC resolves the trade-off between computational cost and fuel-saving performance. Compared to ODP-MPC with m=1001 and dis_tf=10, when m=21 and dis_tf=30, the fuel consumption of ODP-MPC with LEMS remains nearly unchanged, while the computational cost is reduced by 99.1%.