Affiliations: [a] Laboratory of Sciences and Techniques of Automatic Control and Computer Engineering (Lab-STA), National School of Engineering of Sfax, University of Sfax, Sfax, Tunisia | [b] Department of Systems Engineering, University of Valladolid, Valladolid, Spain | [c] Laboratory of Computer and Automatic for System (LIAS), Higher School of
Engineering of Poitiers (ESIP), University of Poitiers, Poitiers, France
Correspondence:
[*]
Correspondence to: Fernando Tadeo, Department of Systems Engineering, University of Valladolid, Valladolid, Spain. Tel.: +34 983 423 566; Fax: +34 983 423 161; E-mail: [email protected].
Correspondence:
[*]
Correspondence to: Driss Mehdi, Laboratory of
Computer and Automatic for System (LIAS), Higher School of Engineering of Poitiers (ESIP), University of Poitiers, Poitiers, France. Tel.: +33 549 453 678; Fax +33 549 454 034; E-mail: Driss.Mehdi@univ-poitiers.
Abstract: The sensor faults in the induction motor can cause the degradation of the performances of the system and even lead sometimes to the instability. This paper focuses on the design of a state space fuzzy observer that simultaneously estimate descriptor system states and sensor faults. By using the estimates of descriptor states and faults, and the linear matrix inequality (LMI) technique, a fault-tolerant control scheme is worked out. The developed approach takes into account the stability of the closed-loop system and the design of non-linear fuzzy inference systems based on TakagiSugeno (TS) fuzzy models. The TS fuzzy model is employed to approximate the nonlinear induction motor in the synchronous d-q frame rotating with field-oriented technique. The output fuzzy controller is able to maintain the damaged system at some acceptable level of performance even in the intermittent loss of sensor measurement signal. The gains of the observer and the controller are obtained by solving a set of Linear Matrix Inequalities (LMIs). Finally, simulation and experimental results are given to illustrate the effectiveness of the proposed approach.
