Using artificial neural networks in the induction motor DTC scheme

The purpose of this paper is to show the potential use of artificial neural networks (ANNs) in the induction motor direct torque control (DTC) scheme. The speed estimators using ANNs are contrasted with the speed adaptive flux observer (Luenberger observer). The paper proposes a complete DTC scheme using two different ANNs to estimate the rotor speed, one of them is a classical feedforward neural network (FFNN) and the other one is a FFNN chosen by genetic algorithms. The performance of the proposed scheme is carried out by simulation tests using MATLAB/SIMULINK.