Uma proposta para proteção térmica do motor de indução trifásico utilizando redes neurais artificiais e sensor de baixo custo

Abstract

This work proposes the study, development and also implementation of a system for temperature monitoring and projection of this value a few minutes in advance by using a low cost temperature sensor. Specifically, this work uses a neural network to estimate the carcass temperature in a three phase induction motor based on the machine thermal dynamics. The database used for the development of this work was gathered from the experiments performed on a test bench. The prototype was validated with two machines working with a variation of conjugate and different rates of voltage unbalance.