Derin sinir ağları kullanarak Eskişehir'deki güneş radyasyonu tahmini

According to the WEO, the global demand for energy is presumably going to increase, due to growing up the world's population during the upcoming two decades. As a result of that, apprehensions about environmental effects, which appear as a result of greenhouse gases, are grown and cleaner energy technologies are developed. This clearly shows that extended growth of the worldwide market share of solar energy. For this reason, a predictive model that effectively observes solar energy conversion and its performance is considered as the best choice for design solar energy plants. In this thesis, to find the most accurate model for solar radiation forecasting, many studies have been conducted. In this work, based on three different approaches, the problem was solved by using classic mathematical models, artificial neural network (ANN), and deep neural network (DNN) methods. Besides, four new models have been developed for solar radiation prediction. The data are obtained from the Turkish State Meteorological of Service. The obtained results are compared and evaluated by the root mean square error (RMSE), mean absolute error (MAE), mean bias error (MBE) and test statistics (t-statistic). As a result of the analysis, the best result was given by the deep learning (DL) method. Keywords: Solar radiation forecasting, ANN, DNN, Artificial intelligence, Hybrid models