Detecção de falhas em motores elétricos utilizando aprendizado de máquina

Abstract

This work proposes a method for detecting failures in electric motors, using machine learning techniques. For training and testing of the model, a data base containing 1951 time series obtained from 4 different sensors that simulate 6 states of operation: normal, unbalanced, horizontal misalignment, vertical misalignment, internal bearing failure and external bearing failure.