Fault diagnosis for HEX/reactor system via invertibility - Université Toulouse III - Paul Sabatier - Toulouse INP Accéder directement au contenu
Communication Dans Un Congrès Année : 2016

Fault diagnosis for HEX/reactor system via invertibility

Mei Zhang
  • Fonction : Auteur
Ze-Tao Li
  • Fonction : Auteur
Boutaib Dahhou
Michel Cabassud

Résumé

A fault diagnosis approach is developed for pneumatic valve used in intensified HEX/reactor system. The pneumatic valve is viewed as an actuator subsystem interconnected with process subsystem in series. A condition of invertibility of the cascade system has been derived in [18], characterizing that effects of faults occurred in actuator subsystem are distinguishable by the global output uniquely. Using this idea, the fault diagnosis scheme proposed in this paper is based on local fault filtering, each subsystem is assigned to monitor one subsystem and provided a decision regarding its health. In particular, the connection point between the two subsystems is not accessible to measurements. An input estimator is then developed to estimate this unknown connection point only rely on the global system output. Different from [18], any information of the derivatives of the output vectors is not involved in the input estimation stage. For that, a high-gain second-order sliding mode observer is considered to exactly estimate the derivatives of the output vectors in a finite time. Numerical simulation examples are given to illustrate the effectiveness of the proposed methods.
Fichier principal
Vignette du fichier
Zhang_20521.pdf (1.43 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01976894 , version 1 (10-01-2019)

Identifiants

Citer

Mei Zhang, Ze-Tao Li, Boutaib Dahhou, Michel Cabassud. Fault diagnosis for HEX/reactor system via invertibility. 2016 24th Mediterranean Conference on Control and Automation (MED), Jun 2016, Athens, Greece. pp.1361-1366, ⟨10.1109/MED.2016.7535988⟩. ⟨hal-01976894⟩
62 Consultations
51 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More