Skip to Main content Skip to Navigation
Theses

Diagnostic des systèmes linéaires en boucle fermée

Abstract : In most of the works concerning system supervision, the methods developed for fault detection and isolation are synthesized from an open-loop representation of the system. But considering real industrial applications, it appears that , most of the times, the system is inserted in a control loop with output feedback. In this context, the task of diagnosis is tedious for several reasons. Firstly, the controller is designed in order to attenuate the effect of the faults, consequently their detection becomes challenging. Secondly, due to the output feedback, the system inputs are correlated with the system outputs, which can complicate the fault isolation. The works presented in the present thesis can be divided into two parts. The first one focuses on the systematic analysis of the sensitivity of the several signals of the control loop with respect to the faults (both additive and parametric faults have been considered). The sensitivity analysis is carried out to select the signal encompassing most information on fault in order to be used for fault diagnosis. In the second one, a fault detection and isolation method based on input-output decoupling is presented for closed-loop linear systems with additive faults. The point is to compute an output feedback such that each fault affects only one output, and thus ease the diagnosis
Document type :
Theses
Complete list of metadatas

Cited literature [86 references]  Display  Hide  Download

https://hal.univ-lorraine.fr/tel-01752956
Contributor : Thèses Ul <>
Submitted on : Thursday, March 29, 2018 - 2:02:51 PM
Last modification on : Tuesday, April 10, 2018 - 1:27:58 AM

File

2007_BAIKECHE_H.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01752956, version 1

Collections

Citation

Hamid Baïkeche. Diagnostic des systèmes linéaires en boucle fermée. Autre. Institut National Polytechnique de Lorraine, 2007. Français. ⟨NNT : 2007INPL078N⟩. ⟨tel-01752956⟩

Share

Metrics

Record views

156

Files downloads

94