An expert system for diagnosing the state of marine electric machine insulation

Abstract

The paper presents an expert system for assessing the state of degradation of electric machine insulation. The system identifies the state of electric machine insulation by making use of the Bayesian decision rule and information from two different sources: the object itself and its environment, i.e. operating conditions. The hypothesis on the uniform probability distribution of reliability states is verified on the basis of the degradation hazards which exist in an assumed location and operation time of the machine; then classifiers of the reliability states are generated. In the paper the causes and rules assumed in the program as well as simulation algorithm are described.