Title: Corrosion pitting damage detection of rolling bearings using data mining techniques

Authors: Yonglai Zhang; Xiaofeng Zhou; Haibo Shi; Zeyu Zheng; Shuai Li

Addresses: Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, 110016, China; University of Chinese Academy of Sciences, Beijing, 100049, China; Key Laboratory of Networked Control System, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, 110016, China ' Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, 110016, China ' Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, 110016, China ' Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, 110016, China ' Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, 110016, China

Abstract: Detection of rolling bearings is very crucial for the reliable operation in the process of condition monitoring of rotating machinery. In this paper, a novel monitoring method using support vector data description (SVDD) with principal component analysis (PCA) for fault diagnosis of corrosion pitting on the raceways and balls in rolling bearings is proposed to improve diagnostic accuracy based on feature extraction dataset of vibration signals. The feasibility and validity of the proposed monitoring scheme are investigated through case study. Experiment results show that the proposed method can achieve 92.85% accuracy, 93.11% sensitivity, and 90.47% specificity based on an imbalanced dataset.

Keywords: machine learning; rolling bearings; corrosion pitting; support vector data description; SVDD; principal component analysis; PCA; damage detection; bearing damage; data mining; condition monitoring; rotating machinery; feature extraction; vibration signals.

DOI: 10.1504/IJMIC.2015.072614

International Journal of Modelling, Identification and Control, 2015 Vol.24 No.3, pp.235 - 243

Received: 17 Jan 2015
Accepted: 09 Mar 2015
Published online: 22 Oct 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article