Compound faults diagnosis of rotating machinery via enhanced two-layer sliding correlated kurtosis

Abstract

Compound faults identification from vibration signals is still a challenge for rotating machinery because the multiple periodic impulsive fault signals may oscillate with the same frequency. In this research, a method termed enhanced two-layer sliding correlated kurtosis (TLSCK) is presented for isolating and identifying the compound defects from the monitored data containing strong background noise and compound faults. The method contains two main steps: the dual-tree complex wavelet package transform (DTCWPT) based correlated kurtogram is conducted on the raw monitored data as a frequency filtering step; further, the enhanced TLSCK method is conducted to diagnosis the compound defects from above filtered signals. The output signal of the enhanced TLSCK could locate the occurrence of interested fault impulses, while the unwanted vibration components and residual noise are well eliminated. Both numerically simulated and experimentally measured vibration data of damaged rolling bearings are analyzed via the presented method to test its performance, and the analysis results validate that the proposed method is effective in detecting compound faults of rotating machinery.