An unsupervised, on-line system for induction motor fault detectionusing stator current monitoring

摘要

A new method for on-line induction motor fault detection ispresented in this paper. This system utilizes artificial neural networksto learn the spectral characteristics of a good motor operating on-line.This learned spectrum may contain many harmonics due to the load whichcorresponds to normal operating conditions. In order to reduce thenumber of harmonics which are continuously monitored to a manageablenumber, a selective frequency filter is employed. This frequency filteronly passes those harmonics which are known to be of importance in faultdetection, or which are continuously above a set level, to a neural netclustering algorithm. After a sufficient training period, the neuralnetwork signals a potential failure condition when a new cluster isformed, and persists for some time. Since a fault condition is found bya relative comparison to a good condition, on-line failure prediction ispossible with this without requiring information on the motor or loadcharacteristics. The detection algorithm was implemented and itsperformance verified on various fault types