A microprocessor-based fault monitor for rotating rectifiers of brushless ac exciters using a pattern-recognition approach

Abstract

A microprocessor-based fault monitor for rotating rectifiers of brushless ac exciters using a pattern-recognition approach is presented in this paper. The features of the frequency spectrum of the voltage waveform induced in a search coil plated in the stator slots of the brushless ac exciter are used in pattern recognition. Based on these features a classifier is designed using the nearest-neighbour rules. A prototype of the microprocessor-based fault monitor has been built. Studies show that the fault monitor can accurately recognize the various classes of faults of the rotating rectifier of the brushless ac exciter.<>