Classificação inteligente de falhas de rolamento em motores elétricos de indução trifásicos utilizando a transformada rápida de Fourier e a análise de componentes principais

Abstract

Three-phase induction motors play an important role in the industry, so the correct diagnosis and classification of faults in these machines is essential, as this reduces unscheduled downtimes for corrective maintenance and consequently increases productivity. Thus, this work presents a study related to the diagnosis of bearing faults in three - phase induction motors driven by frequency inverters. For the application of the method, a pre-processing of the current signals used in the Fast Fourier Transform was performed and the Principal Component Analysis was performed. In order to evaluate the classification accuracy in relation to the different levels of fault severity, the classification accuracy of two networks were compared: (i) Multi-layered Perceptron Network and (ii) Fuzzy ARTmap network.