Fault Diagnosis of Anti-Friction Bearings Based on Bi-Dimensional Extensive Empirical Wavelet Decomposition and Optimized Incremental RVM

Abstract

The prompt and efficient diagnosis of anti-friction bearing faults is crucial for the proper functioning of rotating machinery. In this paper, a novel approach to anti-friction bearing fault diagnosis utilizing Bi-dimensional extensive empirical wavelet decomposition combined with optimized incremental RVM is presented. As an enhanced version of ensemble empirical wavelet decomposition, Bi-dimensional extensive empirical wavelet decomposition extends its application from one-dimensional to two-dimensional signal processing, enabling the extraction of more comprehensive anti-friction bearing features. Additionally, the optimized incremental RVM is employed for fault diagnosis of anti-friction bearings. The experimental findings indicate that Bi-dimensional extensive empirical wavelet decomposition surpasses Bi-dimensional empirical wavelet decomposition, and optimized incremental RVM outperforms traditional RVM. The proposed method incorporating Bi-dimensional extensive empirical wavelet decomposition and optimized incremental RVM proves to be highly effective for anti-friction bearing fault diagnosis.