Weak Fault Enhancement Method for Bearing Fault Diagnosis by Using MWS Stochastic Resonance

Abstract

Large mechanical equipment is subject to periodic vibrations in harsh operating environments for a long time. Bearing failure produces violent changes in the behaviors of large rotating machinery as the safety and reliability of the scene. Therefore, it is especially significant to effectively identify the early failure of the bearing. Since the signal of bearing fault belongs to a low-frequency weak fault, it is hard to classify the characteristic frequency. To solve this shortcoming, a new stochastic resonance model MWS, which is established on the joint monostable function and Woods-Saxon, is presented and used to strengthen the characteristics frequency, also the bearing fault characteristic frequency is gained. The simulation analysis shows that the proposed method can effectively extract the weak fault characteristics which are submerged in the noisy environment, and the bearing test proves that the MWS effectively extract the weak feature information in the bearing fault signal.