基于Wigner-Ville分布的角域故障特征提取方法的研究

将采集的时域非平稳信号经过角域重采样转换成角域准平稳信号,再应用时频分析的方法进行分析,可提高分析精度.实际的工程信号不可避免的都存在噪声干扰,必须对信号作预处理,将信号中的噪声部分消除而又不至于将有用信息滤除,从而能够准确地提取出故障特征,第二代小波去噪方法可以较好地满足此类要求.角域信号带有相位信息,对角域信号应用基于Ga-bor重构的时变滤波技术遮掩掉信号中的各阶比分量,提取出仅包含有故障特征的信号,再计算其Wigner-Ville分布,可将故障定位在某一具体角度上,这对于叶轮机、齿轮等的故障检测具有重要的现实意义.

Research on Method of Angle Domain Fault Feature Extraction on Wigner-Ville Distribution

The acquired time domain non-stationary signals are transferred into angle domain quasi-stationary signals by the method of angle domain resampling,then analyzed by time-frequency analysis method,which can improve the accuracy of analysis.Since the actual engineering signals are inevitably disturbed by noise,it is necessary to pretreat the signals,which will keep the useful information but eliminate noise signals so that accurate fault feature can be extracted.It is argued that the second generation wavelet de-noising can better meet these demands. Angle domain signal is endowed with phase information.The application of angle domain signals to time-varying filter based on Gabor reconstruction masks other order components and extracts only signal of fault feature.Its Wigner-Ville distribution is then calculated to locate fault on a specific angle,which has a great realistic significance in turbines and gear fault detection.