基于SVD和SET的滚动轴承故障诊断研究

Fault diagnosis of rolling bearings based on SVD and SET

在工程现场采集到的轴承振动信号中常包含有大量的背景噪声,导致同步提取变换(SET)无法在时频谱图中清晰、准确地表达出滚动轴承的振动信号特征,为此,基于奇异值分解(SVD)在降噪方面的优势,提出了一种基于SVD和同步提取变换的滚动轴承故障诊断方法。首先,通过SET将滚动轴承的一维振动信号转换到二维时频平面内,得到其时频谱图;然后,采用SVD对时频谱图进行了分解,根据奇异值的大小重新构造了时频谱图,此时的频谱图能清晰、准确地表达出滚动轴承振动信号的时频特征;最后,通过逆变换回到时域,对时域信号进行了解调和包络,所得到的包络谱能清晰地表达出滚动轴承的故障特征频率及其倍频。研究结果表明:相比于只进行SET的滚动轴承信号分析方法,该方法能有效地去除时频谱图中的背景噪声;并且,其包络谱能突出故障的特征频率,从而对滚动轴承进行有效的故障诊断。

In engineering,the collected bearing vibration signal often contained a lot of background noise.Aiming at the problem that the synchro extractingtransform(SET)could not clearly and accurately express the characteristics of the rolling bearing vibration signal in the time-frequency spectrum,combining with the advantages of singular value decomposition(SVD)in noise reduction,a fault diagnosis method based on SVD and SET was proposed.Firstly,the one-dimensional vibration signal was converted into the two-dimensional time-frequency plane through SET,and the time-frequency spectrum was obtained.Then SVD was used to decompose the time-frequency spectrum,and reconstruct the time-frequency spectrum according to the size of the singular value.At this time,the spectrogram could clearly and accurately indicate the time-frequency characteristics of the signal.Finally,returning to the time domain through the inverse transformation,the time domain signal was demodulated and enveloped,and the fault characteristic frequency of the rolling bearing and its frequency multiplication could be clearly expressed by the obtained envelope spectrum.The experimental results show that,comparing with only SET for identifying the rolling bearing vibration signal,the background noise can be effectively removed by this method in the time-frequency spectrum,and its envelope spectrum can highlight the characteristic frequency of the fault,thereby enabling effective fault diagnosis.