基于小波与小波包的变速器故障特征提取

Fault Feature Extraction of Gearbox Based on Wavelet and Wavelet Package

本文结合自适应小波变换滤波去噪方法与小波阈值去噪方法,提出了一种可用于变速器故障振动信号去噪的双层滤波去噪算法。该算法的滤波过程分为两层,第一层滤波采用自适应小波变换滤波算法;第二层滤波采用经典的小波阈值去噪算法对信号进行二次去噪。最后,将去噪后的故障信号采用小波包进行了分解,并提取了小波包频带能量作为故障特征向量。

In this paper, a double filtering algorithm is proposed for de-noising of malfunction gearbox vibration signal, which is produced by combining adaptive wavelet transform de-noising and wavelet De-noising by threshold. The de-noising process of this algorithm consist of two phases, at first phase, de-noising is implemented by adaptive wavelet transform. At the second phase, the classical wavelet de-noising algorithm by threshold is used for the signal de-noising. At last, the de-noised signal is decomposed by wavelet package. The power of frequency bands is extracted as fault feature vector.