基于主成分分析的串联电池组故障诊断实用方法

为了保证电动汽车和储能系统的安全运行，电池组的故障诊断研究备受关注。针对目前面向电池组故障诊断方法相对匮乏且实用性不佳的问题，提出了一种基于主成分分析(PCA)的故障诊断实用方法，以准确地区分组内的电池单体故障和连接故障。首先，提出了非硬件冗余的交叉测量拓扑，分别用不同数量的传感器测量电池和连接板；然后，分析组内测量电压的变化特点，引入PCA对故障特征进行提取，为了保证PCA模型适配，提出了PCA实时建模与故障诊断一体化的思路，并基于此设计了完整的故障诊断方案；最后，利用实验对所提方法进行验证，结果表明所提方法能够可靠区分电池单体故障和连接故障，准确检测阈值法无法检测的电池单体故障，且强鲁棒于荷电状态、健康状态和温度差异的影响。现场运行数据也证实了所提方法能够有效避免发生虚警。

A practical fault diagnosis method for series-connected battery packs based on principle component analysis

In order to ensure the safe operation of electric vehicles and energy storage systems, the research on fault diagnosis for battery packs has attracted much attention. Aiming at the problems that the current diagnosis methods of battery packs are relatively scarce and not practical, a practical fault diagnosis method based on PCA(Principal Component Analysis) is proposed to accurately identify the battery cell fault and connection fault within a pack. Firstly, a non-hardware redundant cross-mearsurement topology is proposed to measure the batteries and the connecting plates with different number of sensors. Then, the variation characteristics of measured voltage in the pack are analyzed and the fault features are extracted by introducing PCA. In order to ensure the adaptation of PCA model, the idea of integrating PCA real-time mode-ling and fault diagnosis is proposed, and based on this, a complete fault diagnosis scheme is designed. Finally, experiments are carried out to verify the proposed method. The results show that the proposed method can reliably distinguish the battery cell fault from the connection fault, accurately detect the battery cell fault that cannot be detected by the threshold method, and is robust to the influence of state of charge, state of health and temperature difference. Field operation data also confirm that the proposed method can effectively avoid false alarm.