基于小波相关滤波-包络分析的早期故障特征提取方法  

Approach to extraction of incipient fault features based on wavelet correlation filter and envelope analysis

作　　者：曾庆虎[1] 邱静[1] 刘冠军[1] 张勇[1] 

机构地区：[1]国防科学技术大学机电工程与自动化学院,长沙410073

出　　处：《仪器仪表学报》2008年第4期729-733,共5页Chinese Journal of Scientific Instrument

基　　金：国家"十一.五"部委预研项目(51317050301)资助

摘　　要：噪声是影响齿轮、滚动轴承等机械设备早期故障诊断正确性的主要因素,利用小波相关滤波法的降噪特性,将小波相关滤波降噪方法和Hilbert包络谱分析相结合,提出了小波相关滤波-包络分析的早期故障特征提取新方法,即首先利用小波相关滤波方法作为包络分析的前置处理手段提取振动信号的微弱故障信息特征,以求得信噪比较高的小波系数;然后对高频段尺度域的小波系数进行Hilbert包络细化谱分析,得到早期故障的特征频率。仿真信号和诊断实例分析结果表明,该方法比直接小波系数包络分析法更能有效抑制噪声,凸现早期故障频率。Noise is the biggest obstacle that makes the incipient fault diagnosis results of gear and rolling element bearing uncorrected, an approach to the extraction of weak fault features from vibration noise based on wavelet correlation filter and envelope analysis was proposed, firstly, the weak fault information features were picked up from the vibration noise using the de-noising characteristic of wavelet correlation filter as the preprocessing of the envelope analysis. Then, in order to get fault feature frequency, the de-noised wavelet coefficients in high scales that represent high frequency signal were analyzed by Hilbert envelope spectrum and the characteristic frequency of incipient fault was obtained. The simulation signal and diagnosing example analysis results reveal that the proposed method is more effective than the method of direct wavelet coefficient envelope spectrum analysis in de-noising and clarifying incipient fault.

关 键 词：小波相关滤波 包络分析 HILBERT 特征提取 故障诊断 

分 类 号：TH133[机械工程—机械制造及自动化] TP391[自动化与计算机技术—计算机应用技术]