Self‐learning energy management for plug‐in hybrid electric bus considering expert experience and generalization performance. (13th March 2020)

• Record Type: Journal Article
• Authors: Guo, Hongqiang; Zhao, Fengrui; Guo, Hongliang; Cui, Qinghu; Du, Erlei; Zhang, Kun
• Is Part Of: International journal of energy research. Volume 44:Number 7(2020)
• Page Start: 5659
• Page End: 5674
• Publication Date: 2020-03-13
• Subjects: energy management -- expert experience -- generalization performance -- plug‐in hybrid electric bus -- Q‐learning; Power resources -- Periodicals; Power (Mechanics) -- Periodicals; Power resources -- Research -- Periodicals; 621.042
• DOI: 10.1002/er.5318
• Languages: English
• ISSNs: 0363-907X

Abstract

Summary: A self‐learning energy management is proposed for plug‐in hybrid electric bus, by combining Q‐Learning (QL) and Pontryagin's minimum principle algorithms. Different from the existing strategies, the expert experience and generalization performance are focused in the proposed strategy. The expert experience is designed as the approximately optimal reference state‐of‐charge (SOC) trajectories, and the generalization performance is enhanced by a multiply driving cycle training method. In specific, an efficient zone of SOC is firstly designed based on the approximately optimal reference SOC trajectories. Then, the agent of the QL is trained off‐line by taking the expert experience as reference SOC trajectories. Finally, an adaptive strategy is proposed based on the well‐trained agent. Specially, two different reward functions are defined. That is, the reward function in the off‐line training mainly considers the tracking performance between the expert experience and the SOC, while mainly considering the punishment in the adaptive strategy. Simulation results show that the proposed strategy has good generalization performance and can improve the fuel economy by 22.49%, compared to a charge depleting‐charge sustaining (CDCS) strategy. Abstract : The self‐learning energy management includes three steps. Step 1 : An efficient zone of SOC is defined based on the approximately optimal reference SOC trajectories, and the approximately optimal reference SOC trajectories are taken as expert experience in off‐line training. Step 2 : The agent is adequately trained off‐line by a multiply driving cycle training method. Step 3 : Adaptive energy management control is realized by the well trained agent and a defined reward function.