A detection method for bearing faults using null space pursuit and S transform

Abstract

In this paper, a novel measurement method based on null space pursuit and S transform is presented to detect the faults of the bearing vibration signal. Since the features of bearing faults in the vibration signal may be submerged by machine-borne vibration and the high intensity of noise, it is very difficult to detect the bearing faults directly. The proposed method is a hybrid of null space pursuit and S transform. As a signal separation approach, the null space pursuit based on operator can extract the fault components from the vibration signal efficiently. Then, S transform is introduced to represent a “Magnitude-Frequency” and a “Time-Frequency” contour of the extracted fault components. Finally, several real vibration signals containing bearing fault signature are investigated to verify the effectiveness of the proposed method.