Title: Fault diagnosis and isolation of a complex system using a neural network observer

Authors: Rania Loukil; Mohamed Chtourou; Tarak Damak

Addresses: Control and Energy Management Laboratory, National Engineering School of Sfax, University of Sfax, Tunisia ' Control and Energy Management Laboratory, National Engineering School of Sfax, University of Sfax, Tunisia ' Sciences and Techniques of Automatic and Computer Engineering Laboratory, National Engineering School of Sfax, University of Sfax, Tunisia

Abstract: In this work, we use the approach based on neural observer in order to introduce the diagnosis of a non-linear system. The synthesis of such a trained specific observer using the back-propagation algorithm leads to an estimation study then a determination of fault diagnosis and isolation of single actuator fault based on residual generation. The robustness of the proposed observer is tested through a physical example. Finally, a comparison of observers' performances will be interesting for judging the effectiveness of this approach. So, the obtained results will be compared to the sliding mode observer and the classic Luenberger observer.

Keywords: neural observers; back-propagation; fault diagnosis; isolation; residual generation; sliding mode observer; Luenberger observer; automatic control; neural networks; complex systems; nonlinear systems; actuator faults.

DOI: 10.1504/IJAAC.2013.057043

International Journal of Automation and Control, 2013 Vol.7 No.3, pp.147 - 165

Received: 28 Dec 2012
Accepted: 02 Jun 2013
Published online: 12 Jul 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article