Robust speed estimation of an asynchronous machine aided by a dynamic neural network

This article presents the estimation of angular velocity of an asynchronous machine aided by an artificial neural network. The input variables of the network are utilised as outputs for a stationary reference point of the stator. They are compared, according to the topologies considered, of voltages, currents, and instantaneous active and reactive power consumed by the machine, and speed feedback. As a result of modelling the machine and using the neural network, different estimators are proposed, studied and characterised. Analysis of the simulation results obtained with the aid of Simulink Toolbox and Nnet confirm the performance of the estimators studied.