International Journal of Economic and Environment Geology (May 2019)
Performance Evaluation of Models Established for the Estimation of Diffused Solar Radiation: Case Study Lahore, Pakistan
Abstract
A suitable design of solar power project requires accurate measurements of solar radiation for the site of investigation. Such measurements play a pivotal role in the installation of PV systems. While conducting such studies, in general, global solar radiation (GSR) is recorded, whereas diffuse component of solar radiation on a horizontal surface is seldom recorded. The objective of the present study is to assess diffuse solar radiation (DSR) on horizontal surfaces by using polynomial models for Lahore, Pakistan (27.89 N, 78.08 E) and by correlating clearness index with diffuse fraction. The established models are compared with some of the existing models from the literature. Performance of models is evaluated by employing five goodness-of-fit (GoF) tests that are, mean bias error (MBE), root mean square (RMSE), Coefficient of Determination (R2), Mean Absolute Percentage Error (MAPE) and Akaike’s Information Criterion (AIC). The comparison of the results of goodness-of-fit tests with those of existing models indicate that the models established in the present study are performed better as compared to the existing models. The values of statistical error analysis further suggested that a cubic model with a good accuracy of 97.5% and AIC of -22.8 is relatively more suitable for this climatic region for estimating diffuse solar radiation. The study shows that the model developed is in good agreement with Elhadidy and Nabi model with an accuracy of 96.1% and AIC of 4.4 and satisfactory results are obtained for Lahore. The findings can help to give a generous understanding of solar radiation in order to optimize the solar energy conversion systems. The results of this study provide a better understanding of the associations between global solar radiation, clearness index and diffused fraction for the region under study
