改进形态滤波与LMD结合的滚动轴承故障诊断方法研究

Study on Improved Morphological Filter and LMD Fault Diagnosis of Roller Bearing

为了消除噪声对滚动轴承故障诊断结果的影响,提出了一种改进形态滤波与局域均值分解(Local mean decomposition,LMD)结合的滚动轴承故障诊断方法,该方法首先利用LMD对滚动轴承的故障信号进行分解,采用峭度和相关系数准则剔除多余的低频分量,再用改进的形态滤波对选出来的PF分量进行滤波解调。最后,对滤波后的信号进行Hilbert包络谱分析,并且与LMD-Hilbert包络谱和直接Hilbert包络谱的结果进行对比分析。实验结果表明:该方法能够有效地提取滚动轴承故障的特征,诊断轴承故障位置。

In order to eliminate the noise influence on roller bearing fault diagnosis,a method of roller bearing fault detection based on improved morphological filter and LMD is proposed.Firstly,the roller bearing fault signals are decomposed by LMD,and the low frequency PF components are eliminated by the spectral kurtosis and correlation coefficient.Then the improved morphological filter is used to filter and demodulate the PF components.Finally,the results are analyzed by the Hilbert envelope spectrum,compared with LMD-Hilbert and Hilbert directly.Experimental result shows that the proposed method is effective and can diagnose the location of bearing fault accurately.