Application of an improved minimum entropy deconvolution method for railway rolling element bearing fault diagnosis

• Published in: Journal of sound and vibration Vol. 425; no. C; pp. 53 - 69
• Main Authors: Cheng, Yao, Zhou, Ning, Zhang, Weihua, Wang, Zhiwei
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier Ltd 07.07.2018; Elsevier Science Ltd; Elsevier
• Subjects: Computer simulation; Coordinate transformations; Deconvolution; Entropy; Fault detection; Fault diagnosis; Image processing systems; Mathematical analysis; Minimum entropy deconvolution; Optimization algorithms; Particle swarm optimization; Particle swarm optimization algorithm; Railway; Roller bearings; Rolling element bearing; Signal processing; Spherical coordinates; Subways; Fault diagnosis; Railway; Minimum entropy deconvolution; Rolling element bearing; Particle swarm optimization algorithm
• ISSN: 0022-460X; 1095-8568; 1095-8568
• DOI: 10.1016/j.jsv.2018.01.023

Minimum entropy deconvolution is a widely-used tool in machinery fault diagnosis, because it enhances the impulse component of the signal. The filter coefficients that greatly influence the performance of the minimum entropy deconvolution are calculated by an iterative procedure. This paper proposes an improved deconvolution method for the fault detection of rolling element bearings. The proposed method solves the filter coefficients by the standard particle swarm optimization algorithm, assisted by a generalized spherical coordinate transformation. When optimizing the filters performance for enhancing the impulses in fault diagnosis (namely, faulty rolling element bearings), the proposed method outperformed the classical minimum entropy deconvolution method. The proposed method was validated in simulation and experimental signals from railway bearings. In both simulation and experimental studies, the proposed method delivered better deconvolution performance than the classical minimum entropy deconvolution method, especially in the case of low signal-to-noise ratio. •An improved deconvolution is proposed for fault detection of rotating machinery.•The filter of the new method is solved by particle swarm optimization algorithm.•The proposed method outperforms the classical minimum entropy deconvolution.