U-042

A Neuro-Fuzzy Based Parameter Identification of an Indirect Vector-Controlled Induction Motor Drive

Authors: Lino Rosell Valdenebro, Jaime Reyes Hernandez, Edson Bim
Affiliation: Universidad Estadual de Campinas, Brazil

Abstract
In this paper an adaptive field oriented control of the induction motor drive is proposed. The adaptive scheme use a neuro-fuzzy approach for the identification of the rotor time constant, which is used to adjust the estimate of the slip angular speed. First a fuzzy logic estimator was developed and tuned, then the fuzzy estimator was implemented by a dynamic backpropagation neural network-based controller. The fast convergence of the fuzzy control is maintained by using adaptive step size of the control variable. The neural network implementation adds the advantage of fast computation, either by a dedicated hardware chip or by digital signal processor (DSP)-based software. Computer simulations results are obtained and the performance of the resulting field oriented controller is analyzed.

Lino Rosell Valdenebro
lino@dsce.fee.unicamp.br