基于DT-CWT和MCKD的滚动轴承早期故障诊断方法  

Bearing Early Fault Diagnosis Based on DT- CWT and MCKD

作　　者：任学平[1] 王朝阁[1] 张玉皓[1] 汪军[1] 

机构地区：[1]内蒙古科技大学机械工程学院,内蒙古包头014010

出　　处：《仪表技术与传感器》2016年第5期106-110,共5页Instrument Technique and Sensor

基　　金：内蒙古自治区自然科学基金资助项目(2012MS0717)

摘　　要：滚动轴承早期故障信号具有非平稳、能量低等特征,为了能够准确、有效地检测出轴承故障,提出了双树复小波(dual-tree complex wavelet transform,DTCWT)和最大相关峭度反褶积(maximum correlated kurtosis deconvolution,MCKD)相结合的诊断方法。首先运用双树复小波对采集到的振动信号进行分解,再重构单支信号,由于噪声的干扰,从重构后分量的频谱中很难对故障做出正确的判断。然后对包含故障特征的分量进行最大相关峭度反褶积处理以消除噪声成分,凸现故障特征信息。最后对降噪信号求取Hilbert包络谱,便能准确获得故障特征频率。通过信号仿真和实验数据分析验证了该方法的有效性。Rolling bearing early fault signal has non-stationary and lower energy characteristics,in order to be able to accurately and effectively detect the bearing fault,dual tree complex wavelet（ DTCWT） and maximum correlated kurtosis deconvolution（MCKD） diagnosis method of combining were put forward. First of all,the dual tree complex wavelet adopted to collect vibration signal was decomposed,and then single signal was reconstructed. Due to the interference of noise,it is difficult to judge from the component of the spectrum reconstruction for failure to make the right judgment. Then,the containing fault feature component was processed by using MCKD to reduce the strong background noise and enhance the fault information. Finally,the fault frequency can be obtained accurately by Hilbert envelope spectrum. Through the signal simulation and experimental data analysis,the effectiveness of the proposed method was verified.

关 键 词：轴承故障 最大相关峭度反褶积 Hilbert包络谱 双树复小波 

分 类 号：TH16[机械工程—机械制造及自动化]