Fast fault estimation for linear systems with unmatched faults

Abstract

This paper proposes a fast fault estimation (FE) method for a class of linear time-invariant systems, which relaxes the strict observer matching condition. Firstly, the integration unknown input observer is designed for the more widely left invertible systems. Secondly, for strongly observable systems, by digesting the idea of a finite-time observer, the finitetime integration unknown input observer is designed, and the faults are estimated exactly in pre-designed arbitrary finite time. Furthermore, for strongly detectable systems, after the system decomposed into strongly observable part and nonstrongly observable part, faults are able to be reconstructed very rapidly due to the finite-time observation of the strongly observable subsystem. The simulation results illustrate the effectiveness and the advantage of the proposed fast fault estimation schemes.

Finite‐time observer design for singular systems subject to unknown inputs

“…Subsequently, Lee and Park considered the fault estimation problem via finite-time UIO method [39]. More recently, under strongly detectable condition, Li et al proposed a class of FTO-based fault fast reconstruction method by geometric theory [42]. It should be pointed out that all the results regarding FTOs mentioned above are focused on non-singular system.…”

Singular System Full-Order and Reduced-Order Fixed-Time Observer Design

“…The merits of the FTO are that it can reach an accurate estimation of original system states in an arbitrarily pre-defined time. By using this technique, Lee et al considered the fault estimation problem [24], and Li et al studied the fault fast reconstruction problem by geometric theory [26]. It should be noted that all the results regarding FTOs mentioned above are focused on non-singular system.…”