Sistema de Inferencia Difuso para Identificar Eventos de Falla en Tiempo Real del STE usando Registros SOE

The aim of this paper is to present the modeling and validation of a fuzzy inference system to identify electric fault events in real time from the STE (Energy Transportation System) using SOE registries. This system will help us to solve the problem of detecting the beginning of the fault event and extracting the useful information of it. This research is part of the Colciencias – ISA UNAL project whose purpose is to design and implement intelligent computational tools to perform fault automatic diagnosis (abnormal events) in real time. The fuzzy inference model chosen is of Mamdani’s type, in which the set of rules is built from expert knowledge, who defines the relationship that should be expected from the combination of the possible states of the input linguistic variables, such as: “amount of SOE signals”, “average time between signals”, “kind of SOE signal”, etc. The fuzzy inference system was successfully validated in real registries of SOE using thousands of signals, collected by the Colombian STE, and handled by the ISA Interconexión Eléctrica S.A. E.S.P. company. Keywords—Artificial Intelligence, Fuzzy Logic, Diagnosis of Fault Events, Real Time Analysis, Energy Transportation Systems.