Utilização das previsões da temperatura ambiente como entrada para a estimativa de radiação solar

Abstract

Due to the great need to use solar radiation for renewable energy generation and the fact that many locations do not have radiation meters, this study emerged as a research project to verify whether it is possible to calculate radiation values using correlation models between radiation and temperature, based on temperature forecast data. Data from one year and one month of forecasts were tested for two locations on the northeastern coast, and the four best correlation models that obtained good results in coastal regions were selected for calculating radiation values. The performance of the models was evaluated using two performance coefficients: the coefficient of determination (R²) and the root mean square error (RMSE). It was observed that the four models obtained similar results, with the Li et al. (LI) which presented the best results in the evaluations with R² values approximately 0.67 in the measured values and 0.33 in the predicted values, and the RMSE obtaining values close to 2.8 MJ.m-²d-¹in the measured values and 4.1 MJ. m-²d-¹ in the predicted values.