Ensemble of One-Class Classifiers for Detecting Faults in Induction Motors

Document Type

Conference Proceeding

Publication Date

8-27-2018

Publication Title

Canadian Conference on Electrical and Computer Engineering

Volume

2018-May

Abstract

This paper studies the use of an ensemble of one-class classifiers for broken rotor bars detection in an induction motors. To achieve this goal, the current signal of induction motor is considered into account for the sake of detection. The fault detector is a multiple classifiers system (MCS), which combines various one-class classifiers to enhance the accuracy of the monitoring system compared to individual one-class classifiers. One-class classifiers are combined in different manners to form the ensembles. These include random subspace, bagging and boosting strategies. These ensemble-based schemes are constructed in homogeneous and heterogeneous configuration and compared together for the purpose of fault detection in induction motors.

DOI

10.1109/CCECE.2018.8447805

ISSN

08407789

ISBN

9781538624104

Share

COinS