Abstract
The solar radiation measured under cloudless sky conditions is a function of the site, elevation, turbidity, solar elevation angle, and many atmospheric constituents for a particular location. This work tests simple clear sky irradiance models in different South African conditions, using irradiance data from five stations, namely University of Fort Hare, Stellenbosch, Vuwani, Pretoria and University of Free State. This study determines clear sky models that best characterise the diurnal curves of solar irradiance across different locations in South Africa using 1-minute resolution data. The study also compares the performance of simple and more complex clear sky models in different South African conditions, using statistical methods. These models are the Haurwitz, Berger–Duffie, Adnot–Bourges– Campana–Gicquel, Kasten–Czeplak, Ineichen–Perez, Bird–Hulstrom, and Simplified Solis models. The models were tested and calibrated by adjusting their input variables. In simple models, the adjustment of scaling parameters resulted in attaining optimal fits of the models to the measured data. The results indicated that the simple and complex models performed excellently by providing better accuracy of modelled clear sky solar radiation as their input parameters were adjusted to fit the data. Further analysis showed that Haurwitz, Adnot–Bourges–Campana–Gicquel, Ineichen–Perez, Bird–Hulstrom, and Simplified Solis were comparable in their estimates. The Berger–Duffie and Kasten–Czeplak models performed worse when averaged in all stations and acceptably good, respectively. The amplitude scaling parameter a and the site’s altitude exhibited a linear relationship for the models Haurwitz, Berger–Duffie, and Adnot–Bourges–Campana–Gicquel.. That suggests that the irradiance scaling increases with the site’s altitude.
M.Sc. (Physics)