Hurst based vibro-acoustic feature extraction of bearing using EMD and VMD

Fault feature extractions of the bearings using the vibration signals are an age old method to anticipate faults in machines. However, the recent research shows that the acoustic sensing using pressure based microphones have significant scope in the area of fault diagnosis. In this paper, initially, the vibro-acoustic features of the bearing at variable speeds are analyzed using variational mode decomposition (VMD) and empirical mode decompositions (EMD). The authors have proposed a novel fault identification method using correlation coefficient (CC) and Hurst exponent to depict the actual fault mode from the decomposed signals. Finally, the vibration and acoustic signals at variable speeds are compared to analyze the effectiveness of the sensing techniques in anticipating faults. These analyses show that most of the times acoustic signals reciprocate the fault mode better than of vibration signals, when extracted using VMD as compared to EMD.