Separation of normal and abnormal components from vibration signal of rotating machinery

摘要

In condition monitoring of rotating machinery, the vibration signal is widely used as a state variable of the machines. Since the normal component and the abnormal component are included in this vibration signal, the effective diagnosis method can be established by separating these components. In this paper, the method for separating vibration signal of time domain and spectrum of frequency domain to the normal component and the abnormal component based on statistical features is proposed. The condition diagnosis using these components that separated by this method is carried out in order to verify the effectiveness.