A geometric approach to nonlinear fault detection and isolation in a hybrid three-cell converter

ABSTRACT

In this paper, the problem of fault detection and isolation in a three-cell converter is investigated using a nonlinear geometric approach. This powerful method based on the unobservability distribution is used to detect and isolate the faulty cell in the three-cell converter. First, a model describing the faults in the cells is presented. The geometric approach is then applied on this faulty model to generate residual signals based on a sliding-mode observer that allows the detection of faults in the three-cell converter. Numerical results show the effectiveness of the proposed sliding-mode residual generators for fault detection and isolation in the three-cell converter.

KEYWORDS:

• Fault detection and isolation
• geometric approach
• switched systems
• three-cell converter
• sliding-mode residual generator

Acknowledgments

Part of this work was conducted when the first author was visiting the Estimation, Modeling and ANalysis (EMAN) group at the Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) devision at King Abdullah University of Science and Technology (KAUST). He would like to thank KAUST for its support and generous hospitality.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

S. Khelouat  http://orcid.org/0000-0003-4400-1286

T. M. Laleg-kirati  http://orcid.org/0000-0001-5944-0121

A. Benalia  http://orcid.org/0000-0003-1260-6723

D. Boukhetala  http://orcid.org/0000-0003-2112-0709

M. Djemai  http://orcid.org/0000-0002-5375-9470

Additional information

Notes on contributors

S. Khelouat

S. Khelouat was born in Algeria, in 1978. He received the B.Sc from UMMTO University of Algeria in 2004 on Automatic Control and M. Sc on Automatic of Industrial Systems in 2008 from Ecole Nationale Polytechnique (ENP) of Algeria. He is currently Phd student at Ecole Nationale Polytechnique (ENP). He is Senior reader at Ecole Nationale Superieure de Technologie of Algeria since 2013. His fields of interest are mainly related to nonlinear control, observation and fault detection theory including hybrid systems, applied to power converter and industrial process.

T. M. Laleg-Kirati

T. M. Laleg Kirati is an associate professor in the division of Computer, Electrical and Mathematical Sciences and Engineering at KAUST and member of the Computational Bioscience Research Center (CBRC). She joined KAUST in 2011 after being a postdoctoral fellow and a researcher at the French Institute for research in Computer Sciences and Control Systems (INRIA) in Bordeaux. She received her PhD in Applied Mathematics in 2008 from INRIA and Versailles University. Professor Laleg’s work is in the general area of mathematical control theory, systems modeling, signal processing and their applications. Her primary research goals are directed towards developing effective estimation methods and algorithms to understand complex systems, extract hidden information and design control and monitoring strategies. Her research projects are motivated by real world problems in engineering and biomedical fields. She was elevated to an IEEE senior member in 2015. She won the best paper awards with her students at the IEEE International Conference of the IEEE Engineering in Medicine and Biology Society and the IEEE International Conference on Control Engineering and Information Technology. She has been recently recognized for being among the three finalists in the academic of distinction category (engineering) of the leadership excellence for women awards. She is also member of the steering committee of KAUST women in data science.

A. Benalia

A. Benalia was born in Algeria, in 1972. He received the B.Sc from the Polytechnic School of Algiers in 1996 and Phd from Paris-Sud university, France in 2004. He is full professor since 2007 and vice rector of the scientific research since 2015 at the university of Laghouat, Algeria. He is with LACoSERE laboratory of the Laghouat university, Algeria. His fields of interest are mainly related to nonlinear control including hybrid system, sliding mode, and flatness theory, with applications to multilevel converters, renewable energy conversion, active power filters and fuel cell system.

D. Boukhetala

D. Boukhetala is a full Professor in Automatic Control at the Ecole Nationale Polytechnique, Algeria. From 1996 to 1999 he was the head of the Department of Automatic Control. From 2005 to 2013 he was the Director of the Process Control Laboratory and from 2010 to date he is the Director of Postgraduate Studies and Scientific Researchat EcoleNationale Polytechnique of Algeria. His research interests are decentralized control, advanced control and fault detection theory applied to robotics, industrial process, power systems and Smart Grids.

M. Djemai

M. Djemai is full professor at the University of Valenciennes, France since 2008. He is with LAMIH–Laboratory of Industrial and Human Automation, Mechanics and Computer Science, CNRS. UMR 8201. He is member of IFAC TC 2.1 control system and IFAC TC. 1.3 on Discrete Event and Hybrid Systems. His research interests are mainly related to nonlinear control, observation, and fault detection theory including hybrid system control, sliding mode, and variable structure systems, and time scale approach, with applications to power systems, robot and vehicles.