Comparison of Regional Empirical Models Based on Sunshine Duration for Determining Solar Irradiance

Abstract

Solar radiation is a crucial parameter required in various fields, particularly for obtaining energy from solar power plants. In this context, the primary objective of this study is to compare commonly used empirical models based on sunshine duration to determine solar radiation for neigh boring provinces centered around Van, a province in eastern Turkey, known for its high solar energy potential. Six empirical models developed based on sunshine duration, utilizing meteorological data obtained from the Turkish State Meteorological Service, were used to estimate solar radiation. Performance evaluation of the models was carried out using several statistical metrics, including Mean Bias Error (MBE), Root Mean Square Error (RMSE), normalized Root Mean Square Error (nRMSE), Mean Absolute Bias Error (MABE), Mean Absolute Percentage Error (MAPE), t -statistic (t-stat), and Coefficient of Determination (R 2 ). Results obtained from the regression analysis revealed that the lowest value of the coefficient of determinatio n was 0.6444 for A & gbreve;r & imath;, while the highest value was 0.8674 for Siirt. In these provinces, exponential and linear models yielded the most successful results, respectively. Additionally, the predictions made using the logarithmic model resulted in significantly poor outcomes in all study regions, with Van having the lowest coefficient of determination at 0.2430. Hakkari demonstrated the best results with a coefficient of determination of 0.7230 using the cubic model, and & Scedil;& imath;rnak yielded the highest resu lt of 0.6795 with the linear model. The results indicate that empirical models based on sunshine duration possess varying prediction capacities depending on the climatic conditions, and therefore, the success of solar radiation estimation relies on the choice of empirical coefficients.