基于DDPG算法的纯电动汽车热管理控制策略研究
Research on the Thermal Management Control Strategy of Electric Vehicles Based on the Deep Deterministic Policy Gradient Algorithm
孙卓松 1袁晓红1
作者信息
- 1. 武汉理工大学现代汽车零部件技术湖北省重点实验室,武汉 430070
- 折叠
摘要
针对纯电动汽车热管理控制策略问题,设计了 一种基于深度确定性策略梯度(Deep Deterministic Policy Gra-dient,DDPG)算法的控制器.首先对纯电动汽车热管理系统的电池包及乘员舱热负荷进行了分析;随后,根据分析结果设计并训练了基于DDPG算法的控制器,该控制器在确保电池安全及乘员舱舒适性的同时,尽可能达到节能的目的;最后,通过不同工况下的仿真对比,验证了基于DDPG算法的纯电动汽车热管理控制策略的有效性,结果表明,DDPG控制器综合性能优于基于规则的控制器.
Abstract
This paper presents a controller for thermal management control strategies in electric vehicles,designed u-sing the Deep Deterministic Policy Gradient(DDPG)algorithm.The thermal load of the battery pack and passenger cabin of the electric vehicle thermal management system is analyzed,and a controller is subsequently designed and trained using the DDPG algorithm.The aim of this controller is to ensure battery safety and passenger comfort,while also achieving maximum energy savings.The effectiveness of the electric vehicle thermal management control strategy based on the DDPG algorithm is verified through simulation comparisons under different operating conditions.The results indicate that the comprehensive performance of the DDPG controller is superior to that of rule-based controllers.
关键词
纯电动汽车/热管理系统/控制策略/深度强化学习/深度确定性政策梯度Key words
electric vehicle/thermal management/control strategy/deep reinforcement learning/deep deter-ministic policy gradient引用本文复制引用
出版年
2024
NETL
NSTL
万方数据