第一作者钟先友男，博士，讲师，1977年生

基于MCKD和重分配小波尺度谱的旋转机械复合故障诊断研究

钟先友，赵春华，陈保家，田红亮

(三峡大学水电机械设备设计与维护湖北省重点实验室，湖北宜昌443002)

摘要：针对强噪声环境下旋转机械复合故障信号难于提取与分离的问题，提出了基于最大相关峭度解卷积(Maximum Correlated Kurtosis Deconvolution，MCKD)和重分配小波尺度谱的旋转机械故障诊断方法。机械信号中存在的噪声会降低重分配小波尺度谱的时频分布可读性，故先要对信号进行MCKD降噪，同时从振动信号中分离出各个故障成分，然后进行Hilbert变换得到包络成分，最后再对包络成分进行重分配小波尺度谱分析，根据尺度图中冲击成分的周期诊断转机械复合故障，算法仿真和应用实例验证了该方法的有效性。

关键词：最大相关峭度解卷积；重分配小波尺度谱；复合故障；最小熵解卷积

基金项目：国家自然科学基金资助项目(51205230, 51275273,51405264)

收稿日期：2014-09-12修改稿收到日期：2014-12-03

中图分类号:TH113.1；TN911.7

文献标志码：A

DOI:10.13465/j.cnki.jvs.2015.07.025

Abstract:Aiming at the problem that the rotating machinery composite faults signal is difficult to be extracted and segmented under strong noise background，a fault diagnosis method for rotating machinery based on maximum correlated kurtosis deconvolution (MCKD) and reassigned wavelet scalogram was proposed. The noise in the rotating machinery vibration signal would reduce the readability of its time-frequency representation，so the noise was reduced by using MCKD，and the fault components were separated from the vibration signal, and then the envelopes were obtained by Hilbert transform and analyzed with the reassigned wavelet scalogram. The composite faults of rotating machinery were diagnosed according to the periods of impulsire components in the scalogram. Some simulation and application examples validate the effectiveness of the method.

Rotating machinery fault diagnosis based on maximum correlation kurtosis deconvolution and reassigned wavelet scalogram

ZHONGXian-you，ZHAOChun-hua，CHENBao-jia，TIANHong-liang(Hubei Key Laboratory of Hydroelectric Machinery Design & Maintenance，China Three Gorges University, Yichang 443002, China)

Key words:maximum correlated kurtosis deconvolution; reassigned wavelet scalogram; composite fault; minimum entropy deconvolution

旋转机械的故障诊断对于减少经济损失和避免人员伤亡具有十分重要的意义。实际生产中旋转机械的故障通常往往表现为复合故障。国内外学者对旋转机械诊断方法进行了大量研究，但是绝大多数是对旋转机械的单一故障进行研究，旋转机械的复合故障诊断仍然是机械故障诊断的难点问题。Purushotham等[1]将小波分析与隐马尔科夫模型相结合，成功地实现了滚动轴承多种故障的分类。Sugumaran等[2]结合决策树和最接近支持矢量机，实现了对多种轴承故障的分类。……