基于改进EMD与形态滤波的齿轮故障特征提取

Gear fault feature extraction based on improved EMD and morphological filter

针对齿轮故障特征往往被强背景噪声淹没的问题,提出一种改进EMD与形态滤波相结合的齿轮故障特征提取新方法。首先采用开-闭、闭-开级联而成的组合形态滤波器对原始故障信号进行消噪处理,然后通过EMD方法将包含在齿轮故障信号中的各个频率族信号分离,再采用互信息方法消除传统EMD分解结果中包含的虚假分量,最后利用分解得到的各阶固有模态函数为单一分量调制信号的特点,通过差值形态滤波的方式对分量信号进行解调以提取故障特征。齿轮故障实验信号的研究结果表明:该方法可有效的提取齿轮故障特征信息并抑制噪声,而且能够取得比传统包络解调分析更好的效果。

Based on improved EMD and morphological filter,a novel method was proposed to extract gear fault feature.Noises were reduced by morphological filters installed cascadely,performing the operation of power off-on and on-off sequentially.Mutual information was used to remove pseudo-components in EMD after the frequency families were separated by EMD.A difference morphological filter was used for demodulating component signal to extract fault feature based on the fact that the eigen modal function of each order is a monocomponent demodulated signal.Experiment results show that the method is more effective than the envelope demodulation in extracting fault feature and reducing noise.