基于多尺度递归定量分析的电池健康状态评估  

Battery health status evaluation based on the quantitative analysis of multiscale recursive

作　　者：周宣昊 张小龙 ZHOU Xuanhao;ZHANG Xiaoong(Haixi Power Supply Company of State Grid Qinghai Electric Power Company,Golmud Qinghai 616000,China)

机构地区：[1]国网青海省电力公司海西供电公司,青海格尔木616000

出　　处：《蓄电池》2024年第5期201-211,共11页Chinese LABAT Man

摘　　要：使用美国宇航局艾姆斯研究中心数据仓库收集的随机电池使用实验数据集,提取了影响电池容量的2个特征参数。通过集合经验模态分解法将2个特征参数(温度和电压)进行分解,然后根据高、中、低频的分量特征进行叠加重构,获得重构的分量。随后通过Visual Recurrence Analysis(RQA)程序对重构分量进行分析,得到非线性特征参数—标准偏差、平均重缩放距离、递归熵。利用得到的非线性特征参数对BP神经网络进行训练,再把测试数据输入训练后的BP神经网络中进行处理,最终得到电池健康状态预测结果。This article uses a random battery usage experimental dataset collected from NASA's Ames Research Center Data Warehouse.Two characteristic parameters affecting battery capacity were extracted from NASA's random battery usage dataset.Then,the two characteristic parameters(temperature and voltage)are decomposed using the ensemble empirical mode decomposition method,and the reconstructed components are obtained by overlaying and reconstructing the high,medium,and low frequency component features.Subsequently,the reconstructed components were analyzed using the Visual Recurrence Analysis(RQA)program to obtain nonlinear feature parameters such as standard deviation,average rescaling distance,and recursive entropy.Train the error back propagation algorithm(BP)neural network using the obtained nonlinear feature parameters,then input the test data into the trained BP neural network for processing,and finally obtain the battery health state prediction result.

关 键 词：电动汽车 动力电池 锂离子电池 NASA 随机电池使用实验数据集 集合经验模态分解法 随机游走 温度 电压 多尺度 递归 定量分析 BP神经网络 健康状态 容量预测 

分 类 号：TM912.2[电气工程—电力电子与电力传动]