Un filtro para detección robusta de fallas en sistemas lineales a tiempo discreto con incertidumbres estructuradas

Abstract

This contribution studies the fault robust detection problem for discrete time linear systems with structured uncertainties. The robustness is analyzed starting from the characterization of two signal types: energy signals and power signals. The power signals are used to characterize the structured uncertainties as perturbations. On base of these two signal types, mixed performance criteria in H2 / H8 are used in order to design a filter for fault robust detection. The filter synthesis is obtained by solution of a convex optimization problem, where the machinery of linear matrix inequalities (LMI) can be used. The filter performance is guaranteed enhanced the sensibility of the faults with respect to the perturbations and the structured uncertainties