The invention discloses a lithium battery fault analysis method based on an extended multi-dimensional time sequence, and the method comprises the steps: selecting a time sequence window with the width of w in a battery voltage sequence with the length of N, moving the time sequence window for m-1 times in total, carrying out the statistics of a group of battery voltage time sequence data in the time sequence window moved each time, and generating the battery voltage time sequence data of an m * w two-dimensional matrix with m rows and w columns; and calculating the variation coefficient of the release rate vector of the battery voltage time sequence data in the ith moving time sequence window, and finally, averaging the variation coefficients of the release rate vectors of the battery voltage time sequence data in the m moving time sequence windows to obtain the variation coefficient of the battery voltage time sequence data of the two-dimensional matrix. According to the method, the dimension of the lithium battery time sequence can be expanded, the nonlinear relation related to influence factors is processed, the influence of one-dimensional data sparsity is solved, and the potential correlation or dependency relation of the lithium battery time sequence is analyzed.
本发明公开了一种基于扩展多维时间序列的锂电池故障分析方法,在长度为N的电池电压序列中,选择宽度为w的时序窗口,共计移动m‑1次,统计每次移动的时序窗口内一组电池电压时序数据,生成m行w列的m×w的二维矩阵的电池电压时序数据。计算第i次移动时序窗口内电池电压时序数据的发放率向量的变异系数,最后对m次移动时序窗口内电池电压时序数据的发放率向量的变异系数求平均值,得到二维矩阵的电池电压时序数据变异系数。本发明可以将扩展锂电池时间序列维度,处理影响因素相关的非线性关系,解决一维数据稀疏性的影响,分析其潜在的关联或依赖关系。
Lithium battery fault analysis method based on extended multi-dimensional time sequence
一种基于扩展多维时间序列的锂电池故障分析方法
2024-02-20
Patent
Elektronische Ressource
Chinesisch
Model-Based Multi-Fault Diagnosis for Lithium-Ion Battery Systems
SAE Technical Papers | 2022
|Model-Based Multi-Fault Diagnosis for Lithium-Ion Battery Systems
British Library Conference Proceedings | 2022
|