Fuzzy Influence Diagrams in Power Systems Diagnostics

Abstract

In this paper, influence diagram with fuzzy probability values, as a graphical tool for the diagnostic reasoning in power system has been proposed. Instead of Bayesian networks that are using conditional probability tables, often difficult or impossible to obtain, a verbal expression of probabilistic uncertainty, represented by fuzzy sets is used in this paper. The proposed methodology enables both type of inference support: bottom-up and top-down inference, including decision nodes in the analysis. This inference engine is illustrated on the case of the detection of the cause of excessive tripping of transformer breaker in the substation.