Diagnóstico activo de fallas basado en consistencias para sistemas híbridos usando algoritmos genéticos y procesos de Markov

Abstract

This work is about a design of a fault diagnoser for a kind of hybrid system using a set of residuals obtained from the system structural model by genetic algorithms. These algorithms verify the main diagnosability properties. Result diagnosis presents ambiguity to reduce this an active diagnosis based on a Markov Decision Process was implemented.