基于模式识别的电动汽车电池故障自动诊断方法  

Automatic Diagnosis Method of Electric Vehicle Battery Fault Based on Pattern Recognition

作　　者：张增丽 白新雷 马文建 ZHANG Zengli;BAI Xinlei;MA Wenjian(State Grid Hebei Electric Power Co.,Ltd.Marketing Service Center,Shijiazhuang 050035,China)

机构地区：[1]国网河北省电力有限公司营销服务中心,河北石家庄050035

出　　处：《河北电力技术》2022年第4期10-14,88,共6页Hebei Electric Power

基　　金：国网河北省电力有限公司科技项目(kj2020-085)。

摘　　要：针对电动汽车电池故障诊断精准性差,诊断结果的FDR值较高问题,提出基于模式识别的电动汽车电池故障自动诊断方法,分析根据电动汽车的电池组成结构,明确电池组工作原理,采用LMD方法提取出电池运行信息内特征向量,将采集得到的特征向量汇总建立识别特征样本库,结合模式识别技术中的贝叶斯算法构建故障自动诊断模型,通过阈值对比确定电池故障具体发生位置,实现基于模式识别的电动汽车电池故障自动诊断。Aiming at the problems of poor accuracy of electric vehicle battery fault diagnosis and high FDR value of diagnosis results,an automatic diagnosis method of electric vehicle battery fault based on pattern recognition is proposed.According to the battery structure of electric vehicle,the working principle of battery pack is clarified.LMD method is used to extract the battery run characteristic vector inside information,collected feature vectors arc summarized to establish the recognition feature sample database,and the automatic fault diagnosis model is constructed by combining the Bayesian algorithm in pattern recognition technology.The specific location of battery faults is determined by threshold comparison,and the automatic diagnosis of battery faults in electric vehicles based on pattern recognition is realized.

关 键 词：模式识别 电动汽车 电池结构 故障诊断 特征向量 判别函数 

分 类 号：U472.9[机械工程—车辆工程]