Estimação de parâmetros do motor de indução utilizando um modelo contínuo no tempo

Abstract

This work describes a method for estimating all electrical parametersof an induction cage motor using a continuous time model and a recursive least squares algorithm. The work presents the linear-in-the-parameters model used by the least squares algorithm. This procedure is based a the motor current-voltage model and assumes that the rotor speed is know. Parameter bias is reduced by including residual terms in the estimation model. Derivatives, which, appear in the regressive model, are estimated by fitting a polynomial to a short window of data and subsequently differentiating such a polynomial. Not all parameters are directly estimated by the algorithm, thus, a procedure has been developed to determine the rotor parameters and also the mutual inductance. The procedures presented in this work were validated using both simulated and real data. The results confirm the good performance of the method.