Predicting global solar radiation in nigeria using adaptive neuro-fuzzy approach

Abstract

Accurate Solar radiation prediction is essential for efficient and reliable solar power project and design. In this study, the potential of Adaptive Neuro-Fuzzy Inference System (ANFIS) is evaluated to predict the global solar radiation on horizontal surface in Nigeria from measured meteorological data. 10 years data ranging from (2002–2012) was analyzed based on the series of measured meteorological data, 70% of the data were used for training the model and 30% were used for testing. The prediction accuracy of the developed model was compared with existing models. Statistical parameters Root Mean Square Error (RMSE) and coefficient of determination (R2) were used to assess the model performance. The result obtained prove that the developed ANFIS model provides more accurate prediction of RSME = 0.8093 MJ/m2 and R2 = 0.86877 MJ/m2 at the training phase and RMSE = 1.6954 MJ/m2, R2 = 0.73632 MJ/m2 at the testing phase. The obtained results prove the developed ANFIS model is an efficient tool for solar radiation prediction.