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

A new method for on-line induction motor fault detection is presented in this paper. This system utilizes artificial neural networks to learn the spectral characteristics of a good motor operating on-line. This learned spectrum may contain many harmonics due to the load which corresponds to normal operating conditions. In order to reduce the number of harmonics which are continuously monitored to a manageable number, a selective frequency filter is employed. This frequency filter only passes those harmonics which are known to be of importance in fault detection, or which are continuously above a set level, to a neural net clustering algorithm. After a sufficient training period, the neural network signals a potential failure condition when a new cluster is formed, and persists for some time. Since a fault condition is found by a relative comparison to a good condition, on-line failure prediction is possible with this without requiring information on the motor or load characteristics. The detection algorithm was implemented and its performance verified on various fault types.<
>