Reconfiguração otimizada de redes de distribuição de energia elétrica eom penetração fotovoltaiea, com a utilização de armazenadores de energia e com o auxílio de inteligência artificial

Abstract

This work presents a study about the reconfiguration of electrical distribution systems, with aim to minimize its technical losses, or be it, Joule effect losses. The effect of Photovoltaics penetration and energy storage were considered herein. To achieve this objective, in a way to at the end of this research left available a useful and applicable tool, resources from Artificial Intelligence techniques were used, such as Artificial Neural Networks and Particle Swarm Optimization. The Artificial Neural Networks were used to estimate the generated power by means of Photovoltaics. Weather data were obtained and standardized with the projection for a defined horizon which allowed the estimation of the potential generated power. By means of bibliographic studies e simulation, the NARX architecture was chosen. A battery energy storage long-term scale study was performed to minimize technical losses on distribution grids. Combinatory analysis were performed to reconfigure the distribution grid with factbility tests, according to graph theory. The power flow was implemented by means of the three-phase Newton-Raphson method. With Particle Swarm Optimization, the optimized configurations of the grid were tested for the network losses minimization. The computer language used, for all of the proposed algorithms, was the available on MATLAB® software. The grid topology analyzed is that from IEEE 37 buses (modified).