摘 要：针对动力电池组中单体故障难以及时识别的问题，本文提出一种基于信息熵和Z分数的故障识别预警方法。该方法通过滑动窗口计算单体电压时间序列的信息熵，以量化其波动紊乱程度；引入Z分数衡量各单体熵值相对于群体均值的偏离程度，从而实现异常程度的标准化度量。在此基础上，结合三级预警机制，对异常单体进行分级预警。实际车辆运行数据的试验验证表明，该方法在充电过程中能有效识别电压异常单体，具有较好的预警能力和可解释性。
关键词：动力电池；信息熵；Z分数；异常检测；故障预警
中图分类号：U469.72；TM912 文献标志码：A DOI：10.15917/j.cnki.1006-3331.2026.03.002
Research on Entropy Based Risk Identification and Early Warning Technology for Power Battery Single Unit Faults
LEI Jianxin, ZHOU Hang, ZHONG Xinhao, XIAO Ye, WANG Fan
Abstract: To address the problem that single-cell faults in a battery pack are difficult to identify promptly,this paper proposes a fault identification and early warning method based on information entropy and Z-score.This method calculates the information entropy of the single-cell voltage time series using a sliding window to quantify the degree of fluctuation disorder.It introduces the Z-score to measure the deviation of each cell's entropy value from the group mean,thereby achieving a standardized measurement of the abnormality degree.On this basis,combining with a three-level warning mechanism,abnormal cells receive graded warnings.The experimental verification using actual vehicle operation data shows that this method can effectively identify voltage-abnormal cells during charging,demonstrating good warming capability and interpretability.
Key words: power battery; information entropy; Z score; anomaly detection; fault warning
