一种平滑固有时间尺度分解法在故障诊断中的应用

The Application of a Smooth Intrinsic Time Decomposition Method for Rolling Bearing Fault Diagnosis

为提高滚动轴承振动信号故障信息提取精度,针对故障诊断过程中存在的噪声干扰问题,文章提出了一种平滑固有时间尺度分解法(Smooth Intrinsic Time Decomposition, SITD)的算法,将小波分析法嵌入到ITD分解过程中,采用了一种自适应阈值函数选取小波系数,使信号重建过程中获得更加精细的有用信号信息。将此方法应用于滚动轴承内圈故障和外圈故障诊断,结果表明与传统ITD方法比较,SITD方法不仅可有效消除背景噪声,同时保留冲击特征,还减少了端点效应,提高了滚动轴承的故障诊断精度。

In order to improve the accuracy of fault information extraction for rolling bearing vibration signal, aiming at the problem of noise interference in the process of fault diagnosis, a Smooth Intrinsic Time-scale Decomposition (SITD) algorithm was proposed. The wavelet analysis was imbedded in iteration procedures of Intrinsic Time-scale Decomposition (ITD). In order to obtain better results, the signal recovery from the noisy version of Proper Rotation Components (PRCs) in wavelet domain using an adaptive threshold function was researched. The proposed method was applied to fault diagnosis with inner race fault and out race fault for rolling bearing, and the results indicate that compared with the traditional ITD method, the SITD method could effectively eliminate the background noise in the vibration signal and reserve the impulse features of the bearing signal, further overcoming the endpoint effect problem. The accuracy of rolling bearing fault diagnosis has been improved.