基于最大相关峭度解卷积与形态滤波的齿轮故障特征提取

Gear feature extraction based on maximum correlated kurtosis deconvolution andmathematical morphological filtering approach

摘要:

为了准确地进行齿轮故障特征提取,结合最大相关峭度解卷积和形态滤波,给出了一种新的方法。首先利用最大相关峭度解卷积恢复信号中的周期性故障特征并实现信号的降噪,再运用形态差值滤波器对解卷积后的信号进行滤波以增强信号中的冲击特征并解调出包络,最后求取包络谱以进行故障特征提取；通过齿轮断齿故障振动数据的分析,验证了方法的有效性。

Abstract:

In order to extract gear fault features effectively, a mode based on maximum correlated kurtosis deconvolution (MCKD)and mathematical morphological filtering is proposed. Firstly, the periodic fault features in the signal are recovered by MCKD and the noise of the signal is reduced. Then the morphological difference filter is used to filter the deconvolution signal to enhance the impact characteristics in the signal, and to get the envelope signal. Finally, the envelope spectrum of the filtering results is obtained to extract the fault features. The analysis of broken tooth of gear fault data shows that the method can extraction of gear fault features effectively.