针对不同碰摩程度的转子故障,利用小波基函数“db12”,进行6层小波分解,对故障特征比较明显的小波频带,利用Morlet小波尺度图进行时频特征提取,给出了各频带内分解信号的特点以及其频带能量比例。结果表明,采用两种小波基函数和频带能量图可以很好地提取不同碰摩情况下的故障特征。轻微碰摩时,碰摩引起的高低频分量很弱;中度碰摩时,引起比较稳定的X/2谐波系列;严重碰摩时,高于转频的谐波成分,能量均达到最大。
Mother wavelet function "db12" was adopted for rotor fault with different rub-impact degrees. The displacement signals were decomposed into 6 levels. Reassigned Morlet wavelet scalogram has been used to extract fault features for wavelet frequency bands with obvious time-frequency features. The decomposed signals in different decomposed levels together with their frequency-band energy ratios were discussed in details. The results showed that fault features could be extracted well by combining two mother wavelet functions and frequency-band energy plot in different rub-impact conditions. In the stage of slight rub-impact, high and low frequency components are all small; in the stage of moderate rub-impact, stable X/2 based harmonic components occurred and in the stage of serious rub-impact, the energy of harmonic components are greater than rotational frequency and reached the maximum.