Modelling, simulation and identification of induction machines in continuous and discrete time.

摘要

Field orientation control, sometimes called vector control, is applied in the control of induction machines to obtain high performance dynamic responses. The key to field orientation control is knowing the instantaneous magnitude and direction of the rotor flux. The magnitude and direction of the rotor flux can either be measured directly with sensors in the direct field orientation control method or be estimated in the indirect field orientation control method. However the performance of the indirect field orientation control method is sensitive to variations in motor parameters such as the rotor resistance and the magnetizing inductance. Unfortunately motor parameters vary greatly with temperature, frequency and current amplitude.;In this thesis a novel method of using a subspace identification method to estimate the rotor flux directly is presented. Simulations of field orientation control with both open loop and closed loop control schemes are first carried out and the sensitivity of field orientation control to the variations in motor parameters is studied. Furthermore the saturation effect in induction machines is taken into consideration. All of the models of induction machines are discretized in order that the identification algorithms can be implemented with a digital signal processor. Finally a subspace identification method for linear parameter-varying (LPV) systems is used to identify the rotor flux of the induction machine. Positive results have been obtained, indicating the potential of the subspace identification method in field orientation control of induction machines.