基于二进小波能量自功率谱的转子碰磨故障特征提取

Feature Extraction of Rotor Rub-impact Fault Based on Energy Auto-power Spectrum of Dyadic Wavelet

针对转子动静碰磨故障,通过分析转子运行时产生振动信号的时频特性,提出了一种基于二进小波能量自功率谱的故障特征提取方法。该方法利用二进小波变换得到小波能量系数谱,在小波能量系数谱中找到系数之和变化最大的能量尺度层,对该层进行自功率谱变换后可将碰磨故障的特征量提取出来。

For rub-impact fault between rotor and stator, the paper presents fault feature extraction method based on energy auto-power spectrum of dyadic wavelet by analyzing time-frequency characteristics of vibration signal generated by operating rotor. By method, dyadic wavelet transformation is made to get coefficient spectrum of wavelet energy. Energy level with largest energy sum in coefficient spectrum of wavelet energy is found, and the characteristic quantity of rub-impact fault is extracted after transforming auto-power spectrum of level.