基于SVD——形态降噪的TKEO故障诊断方法研究

Research on TKEO fault diagnosis method based on SVD-morphological noise reduction

针对强噪声干扰背景下微弱故障特征信息难以提取的问题,提出了一种基于奇异值分解(SVD)-形态降噪的Teager能量算子(TKEO)故障诊断方法。首先对轴承振动信号进行SVD,对得到的分量信号进行形态滤波,以滤除噪声干扰;然后利用峭度准则对分量信号进行筛选,并对其进行重构;最后利用TKEO计算重构信号的瞬时能量,得到信号的能量谱,提取振动信号的特征。将提出的方法应用于滚动轴承故障分析,结果表明该方法能清晰地提取故障特征信息。

Aiming at problem of extracting weak characteristics from the fault signals containing strong background noise, a method based on singular value decomposition （SVD）-morphological noise reduction for Teager-Kaiser energy operator（TKEO） fault diagnosis is proposed. Firstly, bearing vibration sigual is decomposed by SVD, the component signals are filtered by morphological filter to remove the noise; Secondly, the component signals are screened and reconstructed using kurtosis criterion ; Finally, the instantaneous energy of reconstructed signals are calculated by using the TKEO, the energy spectrum of the signals are then obtained, from which the characteristics of the vibration signals are extracted. The experimental results show that the proposed method is capable of extracting fauh features and has offered an approving performance on fault diagnosis of rolling bearing.