Diagnostics of gear deterioration using EEMD approach and PCA process. (February 2015)

• Record Type: Journal Article
• Authors: Yang, C.Y.; Wu, T.Y.
• Is Part Of: Measurement. Volume 61(2015:Feb.)
• Page Start: 75
• Page End: 87
• Publication Date: 2015-02
• Subjects: Ensemble empirical mode decomposition (EEMD) -- Post-processing EEMD -- Principal component analysis (PCA) -- Gear deterioration -- Artificial neural network (ANN) -- Marginal envelope spectrum -- Gear diagnosis; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2014.10.026
• Languages: English
• ISSNs: 0263-2241

Abstract

Highlights: Extract gear fault features through Hilbert–Huang transform approach. Obtain gear meshing frequency-related signal component using ensemble EMD. Utilize time–frequency and marginal envelope spectrum for feature analysis. Principal component analysis is to enhance computational performance of classifier. Achieve intelligent diagnosis of gear fault via HHT, PCA and ANN. Abstract: The research objective is to evaluate the diagnostic effectiveness of the gear deterioration through the feature extraction process as well as the feature sifting approach. The gear fault-related features in the vibration signals are extracted by using the ensemble empirical mode decomposition method (EEMD) and the marginal Hilbert spectrum analysis. The extracted features in different scales of time domain and frequency domain are sifted through the principal component analysis (PCA). The features of high priority in the principal space represent the majority of the dynamical characteristics of the faulted gears. The artificial neural network (ANN) is employed to classify the selected dominated principal components for the purpose of diagnosing the gear deterioration. The diagnostic results obtained using the ANN classifier show that the different types and levels of gear malfunctions can be identified effectively by the proposed approach. It is also noted that the PCA process can simultaneously enhance the accuracy of gear fault diagnosis and reduce the feature dimensionality for the purpose of increasing the computational efficiency.