基于EMD与自相关的能量算子解调机械故障诊断  

Fault Diagnosis of Machinery Based on the EMD and the Autocorrelation Energy Demodulation

作　　者：王少锋[1] 王戈[1] 王建国[1] 仲济祥 

机构地区：[1]内蒙古科技大学机械工程学院,内蒙古包头014010

出　　处：《机械设计与制造》2016年第6期174-178,共5页Machinery Design & Manufacture

基　　金：内蒙古自治区高等学校科学研究项目(NJZY154);国家自然科学基金地区基金项目(21366017);内蒙古科技大学创新基金项目(2014QDL025)

摘　　要：针对在强噪声背景下早期轴承故障振动信号的非线性,非平稳性以及信号出现的复杂调制现象,提出一种基于EMD与自相关函数相结合的能量算子解调故障诊断方法。该方法首先根据信号的小波包熵值对信号小波包降噪,而后降噪信号进行EMD分解,提取相关度最大的IMF分量进行自相关函数分析的方法进一步抑制噪声对提取特征频率的干扰,然后对降噪处理过的信号进行能量算子解调,从而实现提取轴承的故障信号的幅值和频率信息。对机械故障振动信号进行实验分析表明,相对于单纯的小波包分析预处理存在的降噪效果不理想以及普通Hilbert解调法的运算精度满足不了诊断需求的情况,该方法能够有效解调出故障频率信息,以实现对故障类别的推断。Aiming at the bearing vibration signal under strong noise background of nonlinear and non-stationary and phenomenon of complex modulation signal,puts forward a autocorrelation function based on EMD,the combination of energy operator demodulation method for fanlt diagnosis.The method firstly according to the signal of wavelet packet entropy of wavelet packet denoising and signal denoising are decomposed with EMD.And extract the IMF components which have the largest correlation auto-correlation function analysis method to further suppress noise to extract characteristic frequency interference.Noise treated signal energy operator demodulation,so as to achieve the extraction of information about the amplitude and frequency of the signals of bearing fault.Experiment on mechanical fault vibration signal analysis showed that compared with the pure wavelet packet analysis pretreatment of the denoise reduction effect is not ideal,and the precision of an ordinary Hilbert demodulation method cannot satisfy the diagnosis requirement.This method can effectively demodulation failure frequency information,in order to realize the inference of the failure.

关 键 词：小波包熵值 EMD 自相关函数 能量算子 调制解调 故障诊断 

分 类 号：TH16[机械工程—机械制造及自动化] TH17