Application of Artificial Neural Network for Estimating Hourly Solar Radiation
Abstract
In this work, we have considered the application of artificial neural network (ANN) for the estimation of hourly solar radiation. The inputs to the network are four instantaneous radiation values (irradiance) taken at 15 minutes interval. The cascade forward propagation network was used. Solar irradiance data taken at 15minutes interval between 09:00 - 12:00 hours universal time in the month of January, 2010 for our location, Akoka, Lagos (lat. 6.510N; lon. 3.400E ) were used to train and simulate the network. The network was simulated with data that were not used for the training process. The MBE and RMSE obtained are 2.59kJ/m2 , 1.88kJ/m2 , 9.35kJ/m2 and 8.37kJ/m2 , 16.71kJ/m2, 19.95kJ/m2 respectively for each hour during the period between 09:00 and 12:00 hours universal time.