Aplicação de variáveis meteorológicas no cálculo de previsão da energia produzida por um sistema fotovoltaico

Abstract

Although Brazil has abundant resources for electrical energy generation through the solar photovoltaic source, it is perceived that the contribution of this source in Brazilian energy matrix is not very expressive, representing only 0.01% of the internal supply in 2017, according to the Ministério de Minas e Energia. One characteristic of this energy source is the uncertainty and variability of production, since the generation of photovoltaic solar power is strongly dependent on the meteorological conditions, making the production control and its insertion in the electric network more difficult. Using as a premise the impact of meteorological conditions on solar photovoltaic production, become necessary to consider these factors on the implementation of methods that estimate the production of electrical energy, making it possible for the user to define how this potential will be used not only by small users of the mini and micro generation, but also by large industrial plants which compound the National Interconnected System. One of the ways to get the future production of energy by a photovoltaic system is to estimate the solar irradiation that will reach the panel and, acquiring its value, using this to calculate the energy that is going to be generated by the incidence of that radiation over the analyzed system. Solar irradiation can be predicted by a statistical artifice called Single Exponential Smoothing (SES) method that uses previously collected values to estimate the next value of a numerical sequence. However, even though SES is effective on typical days, it’s been seen that its result dispel from actual values on days whose climatic behavior and, consequently, the amount of incident irradiation on the surface experiences changes in relation to the same day of previous years. In order to adjust the SES to obtain greater assertiveness on the evaluation of future irradiation, and later, estimate the electric energy generated by a photovoltaic system, the present research will apply the software called Solar Energy Analyzer, developed by the group, to associate the meteorological variables of precipitation, temperature and clouds with SES prediction, obtaining a consistent estimated irradiance in relation with the meteorological forecast data, which makes possible to estimate the electrical energy production considering the particularities of the meteorology of the study area. This work has as study object the photovoltaic system connected to network installed in the Escritório Verde located at Universidade Tecnológica Federal do Paraná, Curitiba Campus. The incident irradiation over the photovoltaic system in study was evaluated on February and March of the year 2019 for validation and calibration. On April of the same year, the already calibrated and validated method was used to obtain the irradiation and the electrical energy produced by the system, resulting as estimated accumulated energy error 9,25%, that express the assertivity of the developed method.