基于深度强化学习卡尔曼滤波锂离子电池SOC估计  

Estimating the State of Charge of Li-ion Batteries Based on Reinforcement Learning of Kalman Filters

作　　者：高洪森[1] 游国栋[2] 王雪 房诚信 张尚 GAO Hongsen;YOU Guodong;WANG Xue;FANG Chengxin;ZHANG Shang(Tianjin Lishen Battery Joint-stock Co.,Ltd.,Tianjin 300384,China;College of Electronic Information and Automation,Tianjin University of Science&Technology,Tianjin 300222,China)

机构地区：[1]天津力神电池股份有限公司,天津300384 [2]天津科技大学电子信息与自动化学院,天津300222

出　　处：《天津科技大学学报》2020年第4期65-69,共5页Journal of Tianjin University of Science & Technology

基　　金：天津市科技支撑重点项目(17YFZCNC00230,13JCZDJC29100);天津科技大学实验室创新基金资助项目(1902A030)。

摘　　要：针对锂离子电池的荷电状态(state of charge,SOC)估算精度,设计了一种基于深度强化学习卡尔曼滤波锂离子电池SOC估计算法.首先以锂离子电池二阶RC等效电路为研究对象,应用卡尔曼滤波算法,构建了锂离子电池的离散系统数学模型;然后结合深度强化学习思想,构造了一种深度强化学习卡尔曼滤波算法,该算法利用贝叶斯规则确保最佳协方差.仿真结果表明,深度强化学习卡尔曼滤波算法对锂离子电池荷电状态的精度有较好的估计.In order to improve the estimation accuracy of the state of charge(SOC)of the lithium-ion battery,an algorithm to estimate the state of charge of lithium-ion batteries based on reinforcement learning of Kalman filters was proposed.Firstly,a second-order RC equivalent circuit model was introduced,and a state-space model of coefficient of variation of the lithium-ion battery was structured by using Kalman filter.Then,a reinforcement learning algorithm of Kalman filters was structured,while Bayes rule was used to ensure optimal covariance.Simulation shows that the algorithm can more accurately estimate the state of charge of the batteries.

关 键 词：锂离子电池 荷电状态 卡尔曼滤波 深度学习 

分 类 号：TM912[电气工程—电力电子与电力传动]