Fault diagnosis for disturbed nonlinear systems based on robust discrete uncoupled state multiobservers

Abstract

In this paper, robust Unknown Input MultiOb-servers (UIMOs) are designed, in the presence of disturbances, for nonlinear systems based on a discrete uncoupled state mul-timodel. The diagnosis is conducted in the presence of actuator and sensor faults. Firstly, a robust detection against disturbances is investigated. Then, a new bank of UIMOs is designed to accomplish the isolation task using generated structured residuals. Sufficient conditions are developed and given in terms of linear matrix inequalities with equality constraints to compute the uncoupled state multiobserver gains. The simulation results prove the efficiency of the proposed scheme of fault detection and isolation and show that the proposed fault detection and isolation technique is suitable for the disturbed Multiple-Input Multiple-Output (MIMO) nonlinear systems.