International Journal of Control, Automation, and Systems (2015) 13(6):1-12 DOI 10.1007/s12555-014-0140-2 ISSN:1598-6446 eISSN:2005-4092 http://www.springer.com/12555 High Gain Adaptive Observer Design for Sensorless State and Parameter Estimation of Induction Motors Abdelaziz Maouche*, Mohammed M’Saad, Bachir Bensaker, and Mondher Farza Abstract: This paper addresses the problem of accurately estimating the mechanical and magnetic state variables as well as the stator and rotor resistances of induction motors using only the stator cur- rent measurements and the supplied stator voltages from an appropriate nonlinear parametrization. The involved estimation is carried out by a high gain adaptive observer designed bearing in mind the avail- able fundamental results together with the useful implementation features, namely conception simplici- ty and computational efficiency. An exponential convergence of the state and parameter estimation er- rors is established under admissible assumptions, namely the persistent excitation requirement has been particularly reduced thanks to the introduction of unknown parameter characteristic indices. The effectiveness of the adaptive observer is highlighted throughout simulation results involving a typical induction motor. Keywords: Characteristic indices, high gain adaptive observer, nonlinear parametrization, sensorless induction motor. 1. INTRODUCTION An important research activity has been devoted to the conception of high performance sensorless induction motors over the last years. Many challenging observation and control problems have been investigated to this end thanks to high gain, sliding mode, backstepping and adaptive control principles. This leads to various funda- mental contributions with successful experimental evaluations on the observability and observer design, the observer based fault tolerant detection and the high per- formance control system design with and without pa- rameter adaptation [1-10]. A remarkable effort has been devoted to the adaptive observer design providing an admissible estimation of both the state variables and the model parameters. The resulting adaptive observers have been used to design suitable diagnosis procedures as well as adaptive control- lers. Three approaches have been used for adaptive ob- server design: The first approach is based on the ex- tended Kalman filter which is mainly motivated by its simplicity and wide use in engineering systems [7,11,12]. Indeed, the extended Kalman filter suffers from the lack of fundamental stability and convergence results prevent- ing thereby from having those necessary insights to properly specify the design parameters, namely the co- variances matrices of the state disturbances and noise measurements which are crucial from both stability and performance point of view. The second approach is based on high gain design which is particularly appealing by its exponential convergence under admissible as- sumptions [13-16]. The intrinsic sensitivity to noise measurement of the high gain has been recently investi- gated using adequate time varying observer gains as well as suitable filtering actions [17-19]. The third approach is based on the sliding mode principle which is mainly mo- tivated by a finite time convergence and robustness con- siderations [20-23]. It is however worth noticing that the sliding mode systems are commonly implemented using adequate approximations of the sign function altering thereby the finite time convergence. Moreover, there is no great awareness about the high gain nature of sliding mode systems and their robustness nature. This paper provides a high gain adaptive observer for sensorless induction motors which perform an accurate joint estimation of all the state variables, namely, the electrical, mechanical and magnetic ones, together with the stator and rotor resistances using only the stator cur- rents measurements and the stator voltages. Two under- lying issues are worth to be highlighted to better under- stand the nature of the considered estimation problem. Firstly, the observability of IM’s may be lost for specific operating conditions when their rotor speed is not meas- ured. This requires a suitable procedure to properly proc- ess the singularities occurring whenever the observability condition is lost. Secondly, the induction motor design observer model is nonlinear with respect to the rotor and stator resistances and hence the adaptive observer design © ICROS, KIEE and Springer 2015 __________ Manuscript received April 2, 2014; revised November 10, 2014; accepted December 18, 2014. Recommended by Associate Editor Choon Ki Ahn under the direction of Editor Fuchun Sun. This work was supported by the French and Algerian PROFAS program, it was done with the GREYC Laboratory, UMR 6072 CNRS, ENSICAEN, Caen, France. Abdelaziz Maouche and Bachir Bensaker are with the Depart- ment of Electronics, Badji Mokhtar University, BP 12 23000, Annaba, Algeria (e-mails: {maouche.abdelaziz, bensaker_bachir} @yahoo.fr). Abdelaziz Maouche, Mohammed M’Saad, and Mondher Farza are with the GREYC UMR 6072 CNRS, ENSICAEN, 6 Bd Mare- chal Juin, Caen, France (e-mails: {abdelaziz.maouche, mohammed. msaad, mondher.farza}@ensicaen.fr). * Corresponding author.