APRENDIZADO DE MÁQUINAS APLICADO NA PREVISÃO DE FALHAS EM PENEIRAS VIBRATÓRIAS

Abstract

This work presents a study on the application of machine learning to the predictive maintenance technique, more specifically, vibration analysis of vibrating screens, with the objective of reducing the response time for diagnosis of defects present in the equipment via vibration analysis. Such approach is justified by the lack of academic research on the sieving of grains and the existing vibrational failures in this process, thus contributing to the reduction of downtime and cost reduction in the activity of sieving, also contributing in other studies the areas of engineering. This purpose will be achieved through bibliographic research and preliminary analysis on the subject, followed by a qualitative research, of exploratory character, which will make possible the analyses that will be carried out throughout the work. The study proved that the use of machine learning algorithms helps notably in the visualization and analysis of vibrational data, resulting in analyses not only faster, but also with greater reliability and assertiveness.