基于EMD和自适应形态滤波的解调方法及其应用研究

A Demodulation Method Based on Empirical Mode Decomposition and Adaptive Morphological Filtering

针对传统包络解调分析中窄带滤波参数难以确定和Hilbert解调的差频现象,提出了一种基于经验模态分解(EMD)和自适应形态滤波的解调方法,进行滚动轴承故障信息的分离和故障特征频率提取。该方法首先应用EMD的自适应滤波特性分离出故障产生的高频调制信号;然后利用基于峭度的自适应形态滤波方法对其进行解调分析,提取轴承故障特征。仿真及实验分析结果表明:该方法自适应较好,能量损失小,能有效地进行滚动轴承故障特征提取,利于轴承的早期故障诊断。

It is hard to determine the parameters of narrowband filter and the difference frequency phenomenon of Hilbert demodulation using traditional envelope demodulation method. A demodulation method which was based on empirical mode decomposition （EMD） and adaptive morphological filtering was proposed to separate fault features and extract fault characteristic frequency of a rolling bearing. First, the high frequency modulated signal was ex- tracted using the adaptive filtering properties of EMD. Then, an adaptive morphological filtering method based on kurtosis was used to demodulate the high frequency modulated signal and extract the fault features of the rolling bearing. The results of simulation and experiment analysis indicate that the fault features of the wiling bearing can be effectively extracted using this method which is well adaptive and suffer little energy loss. It is good for diagnosing the early fault of rolling bearings.