An ANN optimal preview controller technique for induction motor

An artificial neural networks (ANN) technique for on-line speed control of a three-phase induction motor (IM) is presented in this paper. This novel technique is based on the optimal preview controller. The proposed technique comprises a new error system and vector control of the IM. Preview feedforward steps are introduced into the control law to enhance the transient response and to improve the robustness of the controlled system. A feedforward neural network trained with the backpropagation algorithm has been developed to embody the characteristic of the above optimal preview controller within a small, accurate and global system. The training was successful over a large range of training data. Test results conducted over several data ranges have shown accurate and fast performance in predicting the controller output variables.