Using wavelet based neural networks for feedback signals estimation of a vector controlled induction motor drive

Abstract

In this investigation, a vector controlled induction motor drive is simulated and the feedback signals of this vector controlled drive are estimated using neural networks. The neural networks receive the machine terminal signals as inputs and estimate the rotor flux and unit vectors cosθ e and sinθ e as outputs. These outputs are used in the vector controlled drive system. The calculated feedback signals by the neural networks are not sensitive to the motor parameter variations. In this paper, three types of neural networks (i.e. multilayer perceptron (MLP), radial basis function (RBF) and wavenet) are used and the obtained results are compared. Finally, on the basis of the advantages of wavenets, the results prove the accuracy and effectiveness of the wavenet based estimator. © 2009 Springer Berlin Heidelberg.