Sensor Fault Diagnosis of Singularly Perturbed Systems Applied to F-8 Flight Dynamic

Abstract

The paper outlines sensor fault-detection and isolation of singularly perturbed systems (SPS) using the slow subsystem. In deterministic case, the fast subsystem is considered as a modeling error by the approximation of the SPS with the slow subsystem. A second approach was to design a set of observer-based residuals using the slow subsystem. In stochastic case, the slow subsystem is used to approximate the initial system and to the sensor fault-diagnosis. The goal of the paper is to generate slow subsystem based residuals to fault-detection of SPS. We applied the results on a F-8 aircraft-dynamic to validate the proposed approaches.