Estimation strategies for fault isolation of linear systems with disturbances

Abstract

This study presents fault estimation and diagnostics for a linear system with disturbances. Through a series of transformations on state and output equations, the given system is partitioned into three subsystems: a disturbance and fault-free subsystem, a disturbance-dependent one and an actuator fault-dependent one. The transformations enable the state vectors of the second and third subsystems to be described in terms of only output measurements, respectively. An unknown input observer is constructed for the first subsystem and is then used for actuator fault estimation. The proposed diagnostic method is verified through simulation studies performed on the control of a vertical takeoff and landing aircraft in the vertical plane.