Microsoft Word - revised_version_of_the_paper

One of the most critical problems in induction motor drive is the knowledge of the current value of some induction motor parameters such as the rotor time constant, which depends strongly on temperature and frequency. This problem won’t be easy to solve because of the high coupling between parameters in the nonlinear induction motor model. In order to determine the rotor time constant, this paper will focus on comparative study between two approaches to estimate this parameter using the indirect vector control of induction motor. The first identification is based on The High Gain Observer which ensures decoupling of the desired electrical variable, and requires less computational time. The second identification is based, on Fuzzy Compensator that uses the fuzzy logic principles. Possible exploitation of the proposed study in the fault detection issue is discussed through simulation of the rotor time constant variation and compensation using MATLAB/SIMULINK ®software, and, experimentally through a real time implementation using the dSPACE 1104 board. Key-Words: Induction Motor, Online Identification, High Gain Observer, Fuzzy Compensator, Rotor Time Constant, fault detection, Real Time Implementation, dSPACE 1104. Nomenclature sd v , sq v d, q Axis Stator voltages rd v , rq v d, q Axis Rotor voltages sd i , sq i d, q Axis Stator currents rd i , rq i d, q Axis Rotor currents s ω , r ω Reference frame and the slip speed ω The Rotor speed sd φ , sq φ d, q Axis Stator fluxes rd φ , rq φ d, q Axis Rotor fluxes s L , r L Stator and Rotor Inductances s R , r R Stator and Rotor Resistances s T , r T Stator and Rotor time constants M Mutual inductance σ The Total leakage factor s θ ,θ d, q Reference frame and rotor position rd ε , rq ε Inputs of the Fuzzy Compensator r̂ T Δ The Output of the fuzzy Compensator r̂ T Estimated value of the rotor time constant e T , l T Electromagnetic and Load torque JΔ The total inertia r f The friction coefficient p The number of pole pairs * Denotes reference value $ Denotes an estimated value