基于级联奇异值分解和Hilbert变换的滚动轴承故障诊断  

Rolling Bearing Fault Diagnosis Based on Cascaded Singular Value Decomposition and Hilbert Transformation

作　　者：宋耀文[1] 王彩[1] 程刚[1] 陈曦晖[1] 

机构地区：[1]中国矿业大学机电工程学院,江苏徐州221116

出　　处：《煤矿机械》2014年第4期255-257,共3页Coal Mine Machinery

基　　金：国家高技术研究发展计划重大项目(2012AA06A406)

摘　　要：针对Hilbert变换包络解调法在强噪声影响下不能有效提取轴承故障信息的问题,研究了级联奇异值分解降噪的方法,并采用简便方法对重构矩阵的时间延迟和奇异值降噪阶次进行选择。应用结合后的方法分析含故障信息的滚动轴承实验信号,结果表明该方法能有效降低噪声并突显故障特征频率。Due to the envelope demodulation based on Hilbert transformation failing to extract the fault information from beatings, the cascaded singular value decomposition method used to reduce noise is studyed and simple methods are adopted to choose the parameters of the time delay in reconstruction matrix and the singular value de-noising order . The combined method is used to analyse the rolling beating experiment signal containing the fault information,and the results show that this method can reduce the noise effectively and highlight the fault characteristic frequency.

关 键 词：HILBERT变换 级联奇异值分解 降噪 时间延迟 

分 类 号：TH133.33[机械工程—机械制造及自动化]