改进PCA和JADE盲源算法应用于电机系统振动信号的分离

Improved PCA and Blind JADE Algorithm in Separating Vibration Signals of Motor System

以电机系统振动信号为采集对象,对采集来的复杂、混叠的、带噪声的振动信号进行盲源分离。引入改进的PCA去噪、JADE算法进行分离。结果表明:给出的分离算法能够在复杂振动信号中较为清晰地分离、提取出主要源信号,可以有效地提取各个振动源信号的特征,为电机系统噪声抑制算法的设计和电机故障诊断提供有效信息。

Taking vibration signals as the object of acquisition, then the blind source separation was carried out to these complex and aliasing signals with noise, and the improved PCA and JADE algorithm were intro- duced to separate the sampling data. Separation results show that the given separation algorithm can have major source signals successfully and clearly separated from the complex vibration signals, including characteristics of the vibration signals with various sources. This provides useful information for designing noise suppression al- gorithm of the motor system and for diagnosing motor fault.