Sub-Optimal Geometric Distance Algorithm for Active Fault Detection and Diagnosis: Reduced Nominal Performance Degradation Approach

摘要

This work presents a sub-optimal alternative formulation for a concave minimization problem used in the design of auxiliary inputs for active fault detection and diagnosis. The proposed method is developed in the framework of ellipsoidal reach sets and guaranteed fault diagnosis within a given time horizon. Moreover, the method enables the design of auxiliary inputs in the null input space of the nominal system, thereby reducing the nominal performance degradation. Its properties are investigated for a second-order system and an over-actuated aircraft to demonstrate the method’s numerical efficiency and the auxiliary input’s efficacy in separating the reach sets for the nominal-faulty and faulty-faulty cases.