基于盲源分离去噪和HHT的旋转机械故障诊断

摘要

In the characteristics extraction of the rotating machine with Hilbert-Huang transform,the vibration signals from the sensors mounted on the machine are generally suffered by the disturbance from different types of noise.The neglect of the noise generally causes worse effect of analysis.In order to overcome this deficiency,by means of combining with the blind source separation,a new algorithm,which is named Hilbert-Huang transform based on the fast independent component analysis,is proposed to resolve the mode fission.The simulation and case analysis show that the proposed method is very effective to waken the phenomenon,and to extract the characteristic frequency of the signal,further to realize the fault diagnosis of rotating machinery.%在利用Hilbert-Huang变换对旋转机械的故障信号进行特征提取时,传感器所获得的信号往往受到不同类型的噪声干扰,而忽略噪声的影响常常产生很差的分析效果.为克服此不足,结合盲源分离,提出了一种解决HHT分析中模态裂解现象的方法,即基于快速独立分量分析消噪的HHT分析方法.仿真与实例结果表明,该方法能有效抑制HHT过程中的模态裂解现象,有效提取信号的特征频率,进而实现旋转机械故障诊断.