Department of Mechanical Engineering, Vaal University of Technology, Vanderbijlpark, South Africa; Department of Mechanical Engineering, Tshwane University of Technology, Pretoria, South Africa; Department of Mechanical and Industrial Engineering, University of South Africa, Florida, South Africa
Adeala, A.A., Department of Mechanical Engineering, Vaal University of Technology, Vanderbijlpark, South Africa; Huan, Z., Department of Mechanical Engineering, Tshwane University of Technology, Pretoria, South Africa; Enweremadu, C.C., Department of Mechanical and Industrial Engineering, University of South Africa, Florida, South Africa
Models for estimating monthly average daily global solar radiation were developed for South African provinces. These models, in addition to the traditional sunshine hours used in existing models incorporates ambient temperature, relative humidity, and wind speed as variable parameters for predicting global solar radiation, making it different from most of the existing models that use only sunshine hours as variable. Meteorological data obtained for nine locations in South Africa were employed in the model formulation. The accuracy of the models were verified by comparing estimated values with measured values in terms of the following statistical error tests: mean bias error, mean absolute bias error, mean absolute percentage error, root mean square error, and the regression coefficient.The values of regression coefficient for the formulated models are between the ranges of 90%-99%. It was also observed that for an accurate estimation of global solar radiation in Eastern Cape Province, all weather elements are needed. This implies that the models give an excellent prediction for global solar radiation for their corresponding locations. Also, different errors calculated for the formulated models are close to zero especially mean absolute percentage error. The result shows that the formulated models are good enough to be used to predict monthly average daily radiation for South Africa and also, the inclusion of some other elements in some of the models improved the accuracy of the predictions made by the models.