Estimation en ligne de l'état et des paramètres du moteur asynchrone triphasé

Abstract

Obtaining high performances with an asynchronous machine requires complex commands which require in particular the knowledge of the rotor parameters and states. However, the rotor states and parameters of the asynchronous cage machine are by nature inaccessible. Consequently, their estimation is essential. In this context, our work is pressed on the two following shutters:  Development of a sensorless direct vector control by the use of the Kalman filter extended to the mechanical speed of rotation. The results of simulation show that this filter presents an interesting robustness with respect to the external disturbances and parametric variations of the machine. These performances are translated directly on the total system drive, in which one observes remarkable improvements as well in statics and in dynamics.  Improvement of the static and dynamic performances of an indirect vector control by the use of the extended Kalman filter to rotor resistance in order to estimate on line its variation with the temperature and the frequency. According to the results of simulation of the total drive system, the on line adaptation of this rotor resistance in the block of the " autopilotage" permits to guarantee artificial decoupling between the flux and the torque. This constitutes the well-known performances of the vector control.