Detecção de falhas em sistemas dinâmicos sujeitos a retardo no tempoe incertezas paramétricas

Abstract

This master thesis proposes a strategy for fault detection in continuous time linear systems, subject to time-delay and uncertain parameters, based on robust filter design as state observers. The filter design is based on Lyapunov-Krasovskii theory and on linear matrix inequalities (LMIs). From the residue generated between the outputs of the system and the observer, a normal operating threshold is calculated. Therefore, when the residue is greater than this normal operation threshold the system is considered in fault. Numerical examples illustrate the effectiveness of the proposed method