Research on energy management strategy for PHEVs based on vehicle speed prediction and DP-MPC

Abstract
Addressing the issue that the Charge-Depleting Charge-Sustaining (CD-CS) energy management strategy for Plug-in Hybrid Electric Vehicles (PHEVs) cannot achieve optimal economy throughout the entire driving cycle, and overcoming the challenges of the Dynamic Programming (DP) algorithm requiring prior knowledge of the entire driving cycle and its high computational burden, this paper proposes an energy management strategy based on vehicle speed prediction, Dynamic Programming and Model Predictive Control (DP-MPC). This strategy, under the framework of Model Predictive Control (MPC), improves the DP algorithm utilizing the rolling optimization of MPC, transforming the global optimization of the driving cycle in DP into a local optimization within the prediction horizon of the driving cycle. Firstly, the studied PHEV powertrain configuration and key components’ parameters were determined, and a simulation model was established using AMESim. Secondly, the PHEV energy management strategy based on DP was analyzed. A deep learning-based vehicle speed prediction model (VSPNet) was designed. Within the MPC framework, the vehicle speed prediction model was integrated with the DP model to propose the DP-MPC energy management strategy, converting the global optimization problem into a local one. Lastly simulation analysis was conducted under the China Light-Duty Vehicle Test Cycle (CLTC) conditions. The impact of three different prediction horizons (5s, 8s, and 10s) on energy management was compared. The comparative results indicate that a prediction horizon of 8s yields lower fuel consumption for the PHEV. Under 10, 15, and 20 repeated CLTC cycles, compared to the DP-based energy management strategy, the equivalent fuel consumption per 100 km is only 18.9%, 18.5%, and 16.1% higher, respectively. Compared to the ECMS-based energy management strategy, the equivalent fuel consumption per 100 km is reduced by 4.8%, 6.8%, 7.0%. Compared to the rule-based energy management strategy, the equivalent fuel consumption per 100 km is reduced by 16.3%, 28.1%, 30.0%.