Regresar

Speed control for dc motors using neural networks

Abstract:

In this work the problem of identifying and controlling a DC motor system using Neural Networks is presented. The system consists of a DC machine with a tachogenerator coupled to its shaft, a pre-amplifier, a servo amplifier unit and the power supply for the system. Two control schemes were implemented, inverse plant identification and copy of an existing controller (PID), both were used in combination with the adaptive pbkp_rediction scheme, with the aim of assuring the learning of trained Neural Networks. All the schemes were trained with signals from a computer (input signals) and the voltage from the tachogenerator (output signals). The results show that Neural Networks are a valid option when DC motor system must be performed.