Fusion fault diagnosis approach to rolling bearing with vibrational and acoustic emission signals

As the key component in aeroengine rotor systems, the health status of rolling bearings directly influences the reliability and safety of aeroengine rotor systems. In order to monitor rolling bearing conditions, a fusion fault diagnosis method, namely empirical mode decomposition (EMD)-Mahalanobis distance (E2MD) and improved wavelet threshold (IWT) (E2MD-IWT) for vibrational signals and acoustic emission (AE) signals is developed to improve the diagnostic accuracy of rolling bearings. The IWT method is proposed with a hard wavelet threshold and a soft wavelet threshold. Moreover, it is shown to be effective through numerical simulation. EMD is utilized to process the original AE signals for rolling bearings so as to generate a set of components called intrinsic modes functions (IMFs). The Mahalanobis distance (MD) approach is introduced in order to determine the smallest MD between the original AE signal and IMF components. Then, the IWT approach is employed to select the IMF components with the largest MD. It is demonstrated that the proposed E2MD-IWT method for vibrational and AE signals can improve rolling bearing fault diagnosis, beyond its ability to effectively eliminate noise signals. This study offers a promising approach to fault diagnosis for rolling bearings in aeroengines with regard to vibration signals and AE signals.