EMD-DCS based pseudo-fault feature identification method for rolling bearings

Pseudo-fault feature is the fault feature included in the vibration signals of healthy parts, which is caused by the faulty parts in the system. In the paper, a pseudo-fault feature recognition method based on empirical mode decomposition (EMD) and degree of cyclostationary (DCS) was proposed to identify the pseudo-fault feature of a rotor bearing system. The technical difficulties of rolling bearing fault diagnosis based on the single-channel pseudo-fault signal were analyzed by comparing the healthy and pseudo-fault signals of the rolling bearing. A dynamic model of the rotor-bearing system considering the rolling bearing slipping rate was established. The pseudo-fault feature of the rolling bearing was analyzed by the time-frequency method and the cyclic stationary method. The feature identification process of the rolling bearing pseudo-fault based on EMD-DCS was presented. An experiment of feature identification was carried out by using a rolling bearing fault simulator. The experimental results show that the EMD-DCS based method can effectively distinguish pseudo-fault features of rolling bearings from fault features. The research in the paper has theoretical significance and practical application value to ensure the equipment operation safety. © 2020, Editorial Office of Journal of Vibration and Shock. All right reserved.