Experimental application of vibrational resonance on bearing fault diagnosis

Abstract

The vibrational resonance (VR) method is used to analyze the simulated and experimental vibration signals with different bearing faults. At first, we get the curve of the response amplitude at the characteristic frequency versus the amplitude of the auxiliary signal to obtain the VR phenomenon. Based on the curve, we obtain the optimal amplitude of the auxiliary signal which can induce the strongest VR. Then, we process the simulated and experimental bearing fault signals accordingly. In addition, we made a comparison of the proposed methodology with the envelope spectrum of the time domain signal. The results show that the response amplitude at the characteristic frequency is amplified excellently compared with the corresponding characteristic frequency in the raw signal. The VR method is better than the envelope spectrum in characteristic frequency extraction. Furthermore, the amplitude at the characteristic frequency is much larger than those at other frequency components after the VR processing. Therefore, the VR can be used for bearing fault diagnosis. We think it also has a certain values of reference in other weak feature extraction issues.