基于自相关分析和LMD的滚动轴承振动信号故障特征提取

滚动轴承的故障信号是非平稳的、多分量的调制信号,特别是故障早期,由于调制源弱，早期故障信号微弱且受周围设备的噪声干扰，导致故障特征难以识别。采用自相关分析和局域均值分解（LMD）方法提取故障特征。首先采用自相关分析提取信号中的周期成分,消除噪声的干扰,然后利用局域均值分解方法将多分量的调制信号分解为若干个PF分量之和,再结合共振解调技术对PF分量进行包络分析以提取故障特征频率。实验证明了方法的有效性。

Rolling Bearing Vibration Signal Fault Feature Extraction Based on Autocorrelation Analysis and LMD

Rolling bearing fault vibration signals were  nonstationary and multi-component modulated signals,Especially in the early stages of the fault,fault features were difficult to identify because early fault signals were weak due to the weak modulation source and contained strong noise.The
autocorrelation analysis and LMD method was proposed herein,it can extract fault features effectively.First,the noises were eliminated by using autocorrelation analysis,the periodic components in the signals were extracted.Second,multi-component modulation signals were decomposed into a number of product functions(PFs).Last,applying demodulated resonance technique to each  PF component,the fault feature frequency was extracted through envelopment analysis.The experimental results demonstrate the effectiveness of the method.