Proposta de um sistema de IoT para redes de distribuição: entidades virtuais, visualização de informação e inteligência artificial aplicadas a smart grids

Abstract

The amount of generated and stored data has been growing significantly. Driven by factors related to the internet of things (IoT), this trend is expected to continue. In the electricity sector, the implementation of smart grids (SGs) aims to provide new services and overcome challenges, such as reducing non-technical losses, integrating distributed and intermittent energy sources, as well as meeting the growing demand. For SGs to achieve these goals, the collected data must be transformed into useful information and presented efficiently. In this context, this work presents a proposal for a monitoring and analysis system for distribution networks and SGs, applying concepts related to digital twins, information visualization techniques, and artificial intelligence algorithms. A tree visualization, a map, and the parallel coordinates technique are used. Then, monitoring scenarios with more than a million entities are presented, confirming the efficiency of the proposal. In addition, recurrent neural networks are used to perform load prediction. For the evaluation of the forecast models, several error metrics are used, along with the parallel coordinates visualization technique. Thus, the implemented application materializes several important concepts and functionalities for monitoring and analysis systems, in the context of smart grids.