Estimación de la ocurrencia de una falla en una red eléctrica de media tensión haciendo uso de un algoritmo de Machine Learning

Abstract

The focus of the paper is on the methodology used to estimate the occurrence of faults in medium voltage electrical distribution networks using neural networks, detailing the data collection and processing, as well as the training of such networks. The barriers faced by companies in the electric power distribution and commercialization industry, particularly service interruptions due to network failures, are addressed. It details how corrective maintenance generates high costs and prolonged outage times, while preventive maintenance, although useful, often results in the replacement of components before they are needed. Against this backdrop, the increasing use of technology has driven the implementation of predictive maintenance, which uses constant monitoring and historical data to predict failures.