Diagnóstico de Falhas em Máquinas de Indução Via Imagens Térmicas Utilizando Redes Neurais Convolucionais

Abstract

In this article, an algorithm combining the Region of Interest identification technique and a Convolutional Neural Network model to diagnose faults in an induction motor is introduced. The algorithm developed in PYTHON on the Google Colab platform was applied to classify 10 different failure modes and the healthy operation of an induction motor, presenting a classification accuracy superior than the Decision Tree, Random Forest, Support Vector Machine and K - Nearest Neighbors.