【摘要】 列车轮对轴承故障振动(特别是轮对轴承存在复合故障时)一般是由多个相互独立的振动信号源和噪声混叠而成,常见方法在诊断轮对轴承复合故障时易出现误诊。独立分量分析(ICA)方法能对各个独立源进行估计,可实现列车轴承复合故障的精确诊断;但很多ICA算法是在未考虑噪声的模型下推导出来的,且列车轮对状态监测信号受诸多干扰因素的影响;为此,本文提出将时延自相关降噪与ICA相结合提取并分离列车轮对轴承复合故障特征信息的方法。仿真与实际应用结果表明,该方法能有效分离出轮对轴承复合故障信号中的典型故障,可进一步降低列车轮对轴承故障诊断的误诊率。更多还原

【Abstract】 fault vibration of wheelset bearing(especially the compound fault)is generally independent vibration sources and noise,misdiagnosis fault diagnosis of train wheelset bearing.The method,Independent Component Analysis(ICA),can be used for estimating the independent sources from mixed signals,accurate diagnosis of wheelset bearing.onitoring signal of train wheelset state usually affected by many factorsany Algorithms of ICA are deduced with the model noise a method combining time delay autocorrelation and ICA to extract and separate the fault feature.imulation and practical application show that this method effective separat typical fault from compound fault,can further reduce misdiagnosis rate of train wheelset bearing fault diagnosis.更多还原