A Data-driven Energy Management Strategy for Series Hybrid Electric Tracked Vehicle based on Power Coordinated Control

The power distribution and fuel economy of hybrid electric vehicles are greatly influenced by energy management strategies (EMSs). In this paper, a series hybrid electric tracked vehicle (SHETV) with an independent dual-side drive and a full-wave rectification is served as the research object. The EMS based on dynamic programming (DP) is first constructed to be used as the global optimal benchmark. Then, a data-driven EMS based on power coordinated control and extracting the global optimal control rules of the DP-based EMS is developed by designing the multi-layer perceptron (MLP) neural network, a filter, and a state of charge (SOC) stabilizer to predict the appropriate power distribution, ensuring that the SOC can be maintained around the initial value and the engine power output can be smoothed. Simulation results indicate that the fuel economy of the proposed EMS is improved by 10.14% compared with the strategy without extracting the global optimal control rules and reaches 89.20% of the DP-based EMS.