Preliminary analysis of mechanical bearing faults for predictive maintenance of electrical machines

Abstract

Nowadays, electrical machines are used in numerous applications, where unexpected faults are to be prevented. Sophisticated technologies are demanded to be able to manage big data of machines conditions and store these datasets remotely using cloud computation. This data is necessary for algorithms to be trained and predict further failures. This paper presents a study of bearing faults for predictive maintenance. The data collection in lab environment and its preliminary analysis is introduced. The impact of different control modes and loads on global parameters of rotating machines is discussed. The fault classification and prediction techniques are presented.