An Ensemble Application of Conflict-Resolving ART-Based Neural Networks to Fault Detection and Diagnosis

Accurate fault detection and diagnosis is important for secure and profitable operation of modern power systems. In this paper, an ensemble of conflict-resolving Fuzzy ARTMAP classifiers, known as Probabilistic Multiple Fuzzy ARTMAP with Dynamic Decay Adjustment （PMFAMDDA）, for accurate discrimination between normal and faulty operating conditions of a Circulating Water （CW） system in a power generation plant is proposed. The decisions of PMFAMDDA are reached through a probabilistic plurality voting strategy that is in agree- ment with the Bayesian theorem. The results of the proposed PMFAMDDA model are compared with those from an ensemble of Probabilistic Multiple Fuzzy ARTMAP （PMFAM） classi- fiers. The outcomes reveal that PMFAMDDA, in general, out- performs PMFAM in discriminating operating conditions of the CW system,