Akademik Veri Yönetim Sistemi
JOURNAL OF ELECTRICAL ENGINEERING AND TECHNOLOGY, cilt.21, sa.2, ss.1, 2026 (SCI-Expanded, Scopus)
Solar irradiation prediction is a very important topic for the regions where solar energy systems will be established and for investment policies to be made in the coming years for these increasingly popular systems. However, the continuous change in solar irradiation due to time and atmospheric conditions necessitates developing and evaluating different prediction models. This study aims to assess and analyze the prediction performance of global solar irradiation on a horizontal plane using different time series algorithms, based on meteorological parameters obtained historically and seasonally in Hakkâri, Türkiye, which has a high solar energy potential, between 2019 and 2022. In this context, Autoregressive Integrated Moving Average (ARIMA), Seasonal Autoregressive Integrated Moving Average (SARIMA), Simple Exponential Smoothing (SES), and Holt-Winters Exponential Smoothing (HWES) models are preferred as time series-based solar irradiation prediction models. The study presents a comparative analysis based on the results obtained among the time series algorithms used as prediction models. It details their applications and determines the potential of solar irradiation prediction models. The performance of each prediction model is compared using performance evaluation criteria commonly used in the literature. To increase the success of the models, Grid Search optimization algorithm for hyperparameter optimization and K-Fold cross-validation method for the acceptability of the results are used. It is observed that the success rates of the optimized models are significantly increased compared to their standard forms. The results indicate that the proposed time series models, in terms of evaluation criteria, show that ARIMA and SARIMA models yield successful results and SES and HWES models yield approximately similar results. Overall, this study presents a comprehensive view of time series models that can be used to improve solar irradiation prediction.