A Novel Marine Gear Fault Diagnosis Method Based on VMD and MDE

Abstract

Aiming at the signal of marine gear is always with large noise and difficult to extract the fault features, a fault diagnosis method based on variational mode decomposition (VMD) and multiscale dispersion entropy (MDE) is proposed. The marine gear vibration signal was transformed by the VMD, the problem of parameter selection which affects the decomposition accuracy of VMD was studied, and the corresponding parameter selection method was proposed. Later, the intrinsic mode functions (IMFs) which can effectively characterize the signal characteristics were selected by frequency domain correlation coefficient and energy fluctuation coefficient, and reconstruction signal was formed. Finally, the MDE value of the reconstruction signal was calculated and used to compose the multiscale feature vector, the vector as input to the least squares support vector machines (LS-SVM) for fault diagnosis. Through the analysis of the test gear vibration signal, and the comparison is made with common signal decomposition method, the results show that the proposed method can effectively realize the fault diagnosis of gear.