Fault diagnosis for rotating machinery based on multi-differential empirical mode decomposition

The fault diagnosis of rotating machinery has crucial significance for the safety of modern industry, and the fault feature extraction is the key link of the diagnosis process. As an effective time-frequency method, Empirical Mode Decomposition (EMD) has been widely used in signal processing and feature extraction. However, the mode mixing phenomenon may lead to confusion in the identification of multi frequency signals and restricts the applications of EMD. In this paper, a novel method based on Multi-Differential Empirical Mode Decomposition (MDEMD) was proposed to extract the energy distribution characteristics of fault signals. Firstly, multi-order differential signals were deduced and decomposed by EMD. Then, their energy distribution characteristics were extracted and utilized to construct the feature matrix. Finally, taking the feature matrix as input, the classifiers were applied to diagnosis the existence and severity of rotating machinery faults. Simulative and practical experiments were implemented respectively, and the results demonstrated that the proposed method, i.e. MDEMD, is able to eliminate the mode mixing effectively, and the feature matrix extracted by MDEMD has high separability and universality, furthermore, the fault diagnosis based on MDEMD can be accomplished more effectively and efficiently with satisfactory accuracy.