Estudo do sombreamento parcial em módulos fotovoltaicos através da resistência série e das redes neurais artificiais

Abstract

Given the scenarios of increasing world population, the concentration of CO2, fuel costs, global consumption of energy and climate change, there was a need to search for alternative energy sources. In this context, solar photovoltaic, the result of research and investments over the past five decades, had a great impact in the last decade, recording a significant increase in the production of photovoltaic cells and modules and installations of photovoltaic systems worldwide. One of the goals of this work was to study the behavior of PV modules by partial shading situations in different conditions of temperature and solar radiation. Within this partial shading scenario that impacts virtually all electrical parameters of a module, the study and calculation of the series resistance of the modules was done in conditions of partial shading proposals to verify the relationship between the value of the series resistance and the state of shading partial, allowing identification of a possible state of shading from monitoring the resistance series. Another objective of this work was to apply Artificial Intelligence (AI) resources in the form of Artificial Neural Networks (ANN) to, after the proper training and learning of ANN from the database collected in the field under conditions of partial shade, they can identify the parameters of the PV modules within the various conditions of partial shading proposals. To pursue the objectives of this work four photovoltaic modules were used, two with 40 W nominal power with over fifteen years of manufacturing, assigned by the Instituto Federal de Goiás (IFGoiano) Urutaí Campus, at the city of Urutaí-GO and two new without any use of 75 W nominal power assigned by the Center for Research on Alternative Energy Sources, School of Electrical Engineering, Universidade Federal de Uberlândia (UFU).