基于主元分析与动态时间弯曲的故障诊断方法及应用研究

Fault Diagnosis Method and Applied Research Based on Principal Component Analysis and Dynamic Time Warping

摘要:

轴承是工程实际中常用而又极易损坏的部件,特别是对其早期微弱响应的辨识,具有重要的社会价值和意义；为提高运转轴承的安全可靠性和可维护性,提出了基于主元分析与动态时间弯曲距离的故障诊断方法,它可以准确对早期微弱动态响应辨识、诊断;该方法首先将典型故障样本信号与待测信号小波去噪并EMD分解,并对若干固有模态分量主元分析求取主元,然后对主元分量进行分析,获得相关特征值组成特征向量,计算待测信号与已知故障样本信号特征向量的弯曲距离,弯曲距离越小表明两信号越相似,从而辨识故障;此外,还可将其应用于转子、碰磨、齿轮故障诊断中,工程应用实例表明该方法可以准确故障分类,高效故障诊断。

Abstract:

Bearings are often used in engineering practice but it is easily damaged. Especially identification of their weak response in the early, which has important value and significance. To improve safety reliability and maintainability of the operation bearing, proposed fault diagnosis method and applied research based on principal component analysis and dynamic time warping. It can accurate identify the dynamic response of the early weak and diagnosis.First, fault samples and measured signals were de-noised by wavelet then,they were EMD decomposed, Several IMF components were striked to PCA,And all PCA components were analyzed to obtain the relevant eigenvectors composed by eigenvalues,Calculated the Dynamic Time Warping of eigenvectors that signals being measured between the known sample signals of fault feature vectors,In terms of the two signals,the smaller the distance, the more similar,Moreover, The method can also be applied to the fault diagnosis of rotor, rubbing and gear.Application examples of engineering show that this method can accurate classify faults, efficient troubleshooting.