Next-Month Prediction of Hourly Solar Irradiance based on Long Short-Term Memory Network

Authors : İnayet Özge AKSU

Pages : 225-232

Doi:10.21605/cukurovaumfd.1273795

View : 9 | Download : 8

Publication Date : 2023-03-30

Article Type : Research Paper

Abstract :Today, in parallel with the population growth and the advancement of technology, development concerns have started to arise in terms of country administrators. Therefore, alternative solutions to classical energy sources are sought. Renewable energy sources are one of the preferred energy sources today. The popularity of renewable energy sources, including solar energy, is increasing day by day. Solar energy has the potential and accessibility to spread faster than other renewable energy sources. Since Türkiye is located in a region with a high potential in terms of solar energy, which is generally called the sun belt, it is a right decision to prefer solar energy as an energy source in our region. In this study, time series prediction using Long Short-Term Memory insert ignore into journalissuearticles values(LSTM); Network method is used for short-term solar irradiance estimation. In order to demonstrate the success of the results, a comparison was made with the Artificial Neural Network insert ignore into journalissuearticles values(ANN); method. Finally, prediction results of solar irradiance were compared with statistical tests and error analyzes were given in numerically.
Keywords : Uzun kısa süreli hafıza, Güneş ışınımı, Zaman serisi tahmini, Güneş enerjisi