摘要
为提高锂电池SOH的估算精度,搭建了一种基于改进鹈鹕优化算法(POA)结合反向传播(BP)神经网络的估算模型。通过NASA公开数据集,提取了多组与锂电池SOH相关的健康因子,并进行相关性分析,选取相关性较好的健康因子作为模型输入。通过改进后的POA算法对BP神经网络的权值和阈值进行寻优。将所提算法与BP神经网络、粒子群优化算法(PSO)结合BP神经网络、POA算法结合BP神经网络方法进行比较,仿真结果表明:所提方法的均方根误差更小,决定系数更高,具有更好的实际应用价值。
Lithium battery health status(SOH)is the basis for stable battery operation.Improving the accuracy of SOH estimation of lithium batteries can effectively improve their operational reliability.In order to improve the accuracy of SOH estimation of lithium batteries,an estimation model based on improved Pelican optimization algorithm(POA)combined with back propagation(BP)neural network is built.Firstly,several groups of health factors related to lithium battery SOH are extracted through NASA public data set,and a correlation analysis is made,and health factors with good correlation are selected as model inputs.Then the weights and thresholds of BP neural network are optimized by the improved POA algorithm.Compared with BP neural network,particle swarm optimization algorithm(PSO)combined with BP neural network and POA algorithm combined with BP neural network,the proposed method has a lower root-mean-square error and a higher determination coefficient,and thus possesses more practical application values.
作者
赵辉
朱文彬
岳有军
王红君
ZHAO Hui;ZHU Wenbin;YUE Youjun;WANG Hongjun(Tianjin Key Laboratory of Complex System Control Theory and Applications,Tianjin University of Technology,Tianjin 300384,China)
出处
《重庆理工大学学报（自然科学）》
北大核心
2023年第10期255-262,共8页
Journal of Chongqing University of Technology：Natural Science
基金
天津市自然科学基金重点项目(08JCZDJC18600)
天津市教委重点基金项目(2006ZD32)。
关键词
锂离子电池
健康状态
改进鹈鹕优化算法
BP神经网络
lithium-ion battery
state of health
improved pelican optimization algorithm
BP neural network
