基于QPSO优化LSTM的锂离子电池荷电状态估计  

State of charge estimation of lithium-ion batteries based on QPSO optimized LSTM

作　　者：刘锐 朱培逸 Liu Rui;Zhu Peiyi(School of Electrical Engineering Yancheng Institute of Technology,Yancheng 224300,China;School of Electrical Engineering Changshu Institute of Technology,Changshu 215500,China)

机构地区：[1]盐城工学院电气工程学院,盐城224300 [2]常熟理工学院电气工程学院,常熟215500

出　　处：《国外电子测量技术》2024年第10期9-16,共8页Foreign Electronic Measurement Technology

基　　金：国家自然科学基金(62106025);江苏省自然科学基金(BK20241968);江苏省产学研合作项目(BY20231393)资助。

摘　　要：锂离子电池荷电状态(SoC)是电池管理系统的关键参数之一,针对单一长短期记忆(LSTM)网络估计精度不高的问题,提出量子粒子群(QPSO)优化的长短期神经网络,引入量子粒子群算法对LSTM神经网络模型关键参数进行优化,进而提高网络对SoC的估计性能。此外,采用INR-18650电池数据集对所提出的模型进行测试,包含3种不同温度(0℃、25℃、45℃)和4种工况包括动态压力测试DST、联邦城市驾驶时间表FUDS,US06高速公路驾驶时间表和北京动态压力测试BJDST。最后,在各工况下分别验证模型性能,并与其他优化算法进行比较,验证结果表明,所提方法在各温度下均能提高模型的SoC估计结果,且不同温度4种工况下的均值绝对误差(MAE)均小于1%和均方根误差(RMSE)均小于1.1%,最大误差均在5%以内。The state of charge of lithium-ion battery is one of the key parameters of the battery management system,and in order to solve the problem of low estimation accuracy of a single LSTM network,this paper proposes a long-term and short-term neural network optimized by quantum particle swarm optimization,and introduces the quantum particle swarm optimization algorithm to optimize the key parameters of the LSTM neural network model,so as to improve the estimation performance of the network on SoC.In addition,the INR-18650 battery dataset was used to test the proposed model,which included three different temperatures(0℃,25℃,45℃)and four working conditions,including dynamic stress test DST,federal city driving timetable FUDS,US06 highway driving timetable and Beijing dynamic stress test BJDST.Finally,the performance of the model is verified under each working condition and compared with other optimization algorithms,and the verification results show that the proposed method can improve the SoC estimation results of the model at all temperatures,and the mean absolute error(MAE)and root mean square error(RMSE)are less than 1%and the root mean square error(RMSE)are all less than 1.1%under the four working conditions at different temperatures,and the maximum error is within 5%.

关 键 词：荷电状态估 长短期记忆 量子粒子群优化算法 电池管理系统 

分 类 号：TM912[电气工程—电力电子与电力传动] TN98[电子电信—信息与通信工程]