Pattern recognition-a technique for induction machines rotor fault detection "eccentricity and broken bar fault"

Abstract

A pattern recognition technique based on Bayes minimum error classifier is developed to detect broken rotor bar faults and static eccentricity in induction motors at the steady state. The proposed algorithm uses stator currents as input without any other sensors. First, rotor speed is estimated from stator currents, then appropriate features are extracted. The produced feature vector is normalized and fed to the trained Bayes minimum error classifier to determine if motor is healthy or has incipient faults (broken bar fault, static eccentricity or both). Only number of poles and rotor slots are needed as pre-knowledge information. Theoretical approach together with experimental results derived from a 3 hp AC induction motor show the strength of this method. In order to cover many different motor load conditions data are derived from 10% to 130% of the rated load for both a healthy induction motor and an induction motor with a rotor having 4 broken bars and/or static eccentricity.