基于小波去噪和EEMD＿HHT边际谱的滚动轴承故障诊断

Rolling Bearing Fault Diagnosis Based on Wavelet denoising and Ensemble Empirical Mode and Hilbert-Huang Transform Marginal Spectrum

本文将小波及聚合经验模态分解(ensemble empirical mode decomposition, EEMD)和Hilbert-Huang变换(HHT)边际谱的故障分析方法相结合,应用于强噪声背景下轴承信号故障特征提取。首先将轴承信号利用小波变换进行降噪处理,然后采用EEMD方法将轴承振动信号分解成若干个固有模态函数(IMFs);然后对各IMF进行Hilbert变换,求出轴承振动信号的HHT边际谱,最后根据边际谱能够区分不同工况下的正常和故障轴承,正确率为100%,并且通过谱图及局部细化图能够分析其频率特征。结果表明,这种方法能够有效提取轴承故障特征信息,提高轴承故障诊断率。

The paper combines Wavelet with Ensemble Empirical Mode and Hilbert-Huang Transform MarginalSpectrum fault analysis methods to apply to rolling bearing signals＇fault feature extraction under strong noisebackground. Firstly rolling bearing signals were denoised with wavelet transformation, and then were decomposed intoseveral IMFs , then each IMFs carried on Hilbert transformation to get HHT Marginal spectrum of the rolling bearingvibrating signals, finally normal and faulty rolling bearings of different work states can be classified based on thismarginal spectrums with 100% accuracy , and their frequency features can be analyzed with marginal spectrums and theirlocal refinement diagrams. The result shows this method can extract the fault features of rolling bearings and raise therolling bearing＇s fault diagnosis rate.