Comparison of Some Empirical Solar Radiation Estimation Models Based on Air Temperature and Extraterrestrial Radiation in Kahramanmaras Conditions

Abstract

In this study, it is aimed to calibrating, testing and comparing the Bristow – Campbell, Chen, Annandale and Hargreaves – Samani empirical models, which are used in solar radiation (R_s) estimations based on the air temperature (T) and extraterrestrial radiation (R_a), in accordance with Kahramanmaraş conditions. First of all, using the long-term average daily values of the T and R_s data measured by the Regional Directorate of Meteorology (1938 – 2020), the calibration equations of the models were created using the Microsoft Excel program solver add-on. Then, these equations were tested by using the daily average T and R_s data measured in Kahramanmaras Sutcu Imam University (KSU) in the July - October period of 2021. The daily average R_a values were estimated depending on latitude and time. The long-term average measured daily R_s values varied between 4.992 – 32.557 MJ/m²/day. The daily average R_s values estimated by the calibration equations of the Bristow – Campbell, Chen, Hargreaves – Samani and Annandale models using the long-term average daily T, R_s and R_a data varied between 4.591 – 31.832 MJ/m²/day, 5.680 – 32.692 MJ/m²/day, 3.508 – 36.673 MJ/m²/day and 3.508 – 36.673 MJ/m²/day, respectively. The mean absolute percentage error (MAPE) values calculated as an indication of the deviation between the measured and estimated daily average R_s values using the models were determined as 7.921%, 8.016%, 10.178% and 10.178%, respectively. Similarly, the MAPE values for the estimations made with the year 2021 data were obtained as 10.192, 10.424%, 14.192% and 14.291%, respectively. While daily solar radiation values were estimated with an accuracy rate of approximately 90% (MAPE ≅ 10%) with Bristow – Campbell and Chen models, this rate was approximately 85% (MAPE ≅ 15%) for Hargreaves – Samani and Annandale models. It has been concluded that using these four models, which are calibrated in accordance with the local conditions, daily average solar radiation values can be estimated with high accuracy.