Estimation of Global Solar Radiation using Back Propagation Neural Network: A case study Tripoli, Libya

Abstract

Adequate information on global solar radiation with relevant meteorological parameters at any location is necessary for planning, designing, and prediction of the efficiency and performance of solar energy applications. Measurements of global solar radiation in developing countries are very difficult and not readily available because of the cost of the equipment and their maintenance. Libya as one of the developing countries is facing challenges in global solar radiation measurements and recording. This study employed the application of Back Propagation Artificial Neural Network (BPNN) for the estimation of global solar radiation at Tripoli, Libya. For this reason, meteorological data for the period of January 1995 to December 2010 for important cities of Libya (Tripoli) is collected from Libyan National Meteorological Center Climate and Climate Change. The data consists of the monthly average sunshine hours, rainfall, max. temperature, wind speed, mean evaporation, and relative-humidity. Sensitivity analysis was employed and three different model combinations (M1, M2, and M3) were carried out for the model development. The performance efficiency of the models was evaluated. The obtained results show that BPNN has been proved satisfactory in all the analyses and the analysis of the results indicated that BPNN-M3 has the highest accuracy and reliability.