Adaptive Hierarchical Energy Management Design for a Plug-In Hybrid Electric Vehicle

To promote the real-time application of the advanced energy management system in hybrid electric vehicles (HEVs), this paper proposes an adaptive hierarchical energy management strategy for a plug-in HEV. In this paper, deep learning (DL) and genetic algorithm (GA) are synthesized to derive the power split controls between the battery and internal combustion engine. First, the architecture of the multimode powertrain is founded, wherein the particular control actions, state variables, and optimization objective are explained. Then, the hierarchical framework for control actions generation is introduced. GA is utilized to search the global optimal controls based on the powertrain model provided in MATLAB/Simulink. DL is applied to train the neural network model that is connecting the inputs and control actions. Finally, the effectiveness of the presented integrated energy management strategy is validated via comparing with the original charge depleting/charge sustaining policy. Simulation results indicate that the proposed technique can highly improve the fuel economy. Furthermore, a hardware-in-the-loop is conducted to evaluate the adaptive and real-time characteristics of the designed energy management system.