基于小波包的轴承信号降噪和特征提取的研究

Study of Bearing Signal De-noise and Feature Extraction Based on Wavelet Packet

为有效识别轴承故障特征,以轴承内圈故障的信号为例,采用在非平稳信号消噪和以频带能量分布作为故障特征方面有着广泛应用的小波包进行Mat-lab仿真,获得小波包降噪后的信号和作为内圈故障特征的频带能量分布。通过分析频带能量,其结果与实际故障相一致,得出小波包在轴承故障特征提取方面有着一定的优越性。

In order to effectively identify the bearing fault feature, the fault signals of bearing inner ring is adopted. Wavelet packet is adapted to simulate based on Mat-lab because it is widely used in non- stationary signal de-noise and using band energy distribution as fault feature. Then the de-noised signals and the band energy distribution are obtained. Through analyzing the band energy, the result proves consists with the physical fault. So it comes into conclusion that wavelet packet has a certain advantage in the extraction of bearing fault feature.