Fault Detection of Rolling Bearing Using Performance Index of Statistical Process Control

Rolling bearings are among the main components that support the shaft in rotating machines. Rolling bearings are typically subjected to harsh operating environments, and hence, experience wear and tear significantly. Because the failure of rolling bearings can result in the downtime of rotating machines and lead to significant monetary losses, the health monitoring of such machine components is an ongoing issue in industries. However, the mutual interference of complex mechanical elements with various functions complicates the health monitoring process, increases time consumption, and consequently, increases costs. In this study, the acceleration data of rolling bearings obtained in the time domain was analyzed using the statistical process control (SPC) performance index. Through the analysis of variance via SAS JMP, the analyzed data was found to have statistical significance, and the box plot classified the failure of each parameter based on 0.8 and 1.5 of the SPC performance index. Consequently, when the SPC performance index was considered, the multinominal logistic regression analysis provided a satisfactory evaluation of the probability of failure occurrence.