一种自适应Morlet小波滤波方法及其在滚动轴承早期故障特征提取中的应用

An adaptive Morlet wavelet filter method and its application in detecting early fault feature of ball bearings

针对滚动轴承早期故障信号微弱,故障特征难以提取的问题,提出了一种基于混洗蛙跳算法(shuffled frog leaping algorithm,SFLA)的自适应Morlet小波滤波方法.首先利用自相关分析去除宽频随机噪声,然后通过SFLA优化Morlet小波的滤波参数,获得在最小信息熵下的中心频率和滤波带宽.由自适应Morlet小波滤波器获得的滤波信号,其中的冲击成分可以很好地被表征.最后对滤波后的信号做包络谱分析即可提取滚动轴承的故障频率.实验表明,自适应Morlet小波滤波方法可以成功地从低信噪比信号中提取出周期冲击特征,对于滚动轴承早期故障振动信号,能够有效地提取冲击特征频率实现滚动轴承早期故障诊断.

Considering the early fault of ball bearings being weak and the difficulty of detecting the fault feature,an adaptive Morlet wavelet filter method based on shuffled frog leaping algorithm（ SFLA） is proposed. First,the auto-correlation analysis is utilized to filter the broadband random noise.Then,the optimal center frequency and the filter bandwidth under the minimum information entropy are acquired by optimizing the filtering parameters of Morlet wavelet through SFLA. The filtered signal can be obtained by applying the adaptive Morlet wavelet filter,and the impulse features can be well highlighted. Finally,the filtered signal is analyzed by the envelope spectrum to extract the fault frequencies of the ball bearings. Experimental results indicate that the proposed method can successfully detect the periodic impact features from the lowsignal-to-noise ratio（ SNR） signal. Furthermore,in the processing of the early fault vibration signals of the ball bearings,the proposed method can be adopted to obtain the impulse feature frequencies effectively,which is used to diagnose the early fault of ball bearings.