Energy management strategy for hybrid electric vehicle integrated with waste heat recovery system based on deep reinforcement learning

作     者：WANG Xuan WANG Rui SHU GeQun TIAN Hua ZHANG XuanAng WANG Xuan;WANG Rui;SHU GeQun;TIAN Hua;ZHANG XuanAng

作者机构：State Key Laboratory of EnginesTianjin UniversityTianjin 300072China 

出 版 物：《Science China(Technological Sciences)》 (中国科学（技术科学英文版）)

年 卷 期：2022年第65卷第3期

页      面：713-725页

核心收录：

学科分类：082304[工学-载运工具运用工程] 08[工学] 080204[工学-车辆工程] 0802[工学-机械工程] 0823[工学-交通运输工程] 

基　　金：supported by the National Natural Science Foundation of China(Grant No.51906173) 

主　　题：hybrid electric vehicles organic Rankine cycle waste heat recovery deep reinforcement learning energy management system 

摘      要：Hybrid electric vehicles(HEVs)are acknowledged to be an effective way to improve the efficiency of internal combustion engines(ICEs)and reduce fuel *** the ICE in an HEV can maintain high efficiency during driving,its thermal efficiency is approximately 40%,and the rest of the fuel energy is discharged through different kinds of waste ***,it is important to recover the engine waste *** of the great waste heat recovery performance of the organic Rankine cycle(ORC),an HEV integrated with an ORC(HEV-ORC)has been ***,the addition of ORC creates a stiff and multi-energy problem,greatly increasing the complexity of the energy management system(EMS).Considering the great potential of deep reinforcement learning(DRL)for solving complex control problems,this work proposes a DRL-based EMS for an *** simulation results demonstrate that the DRL-based EMS can save 2%more fuel energy than the rule-based EMS because the former provides higher average efficiencies for both engine and motor,as well as more stable ORC power and battery ***,the battery always has sufficient capacity to store the ORC ***,DRL showed great potential for solving complex energy management problems.