基于短时傅里叶变换和独立分量分析的滚动轴承包络分析

滚动轴承的早期故障信号能量小,频带分布广泛；而传统包络谱分析技术直接在强干扰影响下对滚动轴承的故障特征提取经常失效.提出一种基于短时傅里叶变换(short time Fourier transform,STFT)的能量谱和独立分量分析( independent component analysis,ICA)的抗干扰滚动轴承包络分析新方法.该方法首先对获取的滚动轴承振动信号进行STFT能量谱分析,获取信号采样频带下的能量分布,采用带通滤波器获得高频带能量信号,并提取该包络波形,再通过ICA实现包络波形按源分离去噪,最后通过比较各独立分量的包络频谱与滚动轴承理论计算故障特征频率的匹配性,实现滚动轴承故障的精确诊断.仿真数据和试验验证该方法的可行性.

APPLICATION OF ROLLING ELEMENT BEARING ENVELOPE ANALYSIS BASED ON SHORT TIME FOURIER TRANSITION AND INDEPENDENT COMPONENTS ANALYSIS

The initial fault characteristics of the rolling elements bearing usually shows low energy and dispersed frequency distribution.And the fault feature extraction often fails under influence of strong interference while using traditional envelop analysis.A new envelop analysis based STFT(short time Fourier transform) energy spectrum and independent component analysis(ICA)was proposed.The method firstly analyze the vibration signal with STFT energy spectrum,getting signal frequency distribution of energy.Using filter-pass band obtain high frequency signal and extract the envelop energy section,then through the ICA realize envelope waveform denoising according to source separation,finally calculate the envelope spectrum of the independent components and compare it with the fault characteristics which calculated with the theory of the rolling elements bearing,realize accurate diagnosis of rolling bearings’ faults.Simulation data and experiment verify the feasibility of the method.