Implementation of a neural network to adaptively identify and control VSI-Fed induction motor stator currents

Abstract

This paper presents a prototype hardware implementation of a continually online trained artificial neural network (ANN) to adaptively identify the electrical dynamics of an induction machine and control its stator currents from a pulsewidth modulated voltage-source inverter. A single-transputer-based hardware platform is described, and the effects of computational speed limitations on the controller bandwidth are discussed. Captured results are compared with simulation results to practically verify the success of the adaptive neural network identification and control scheme.