Short Term Solar Radiation Forecast from Meteorological Data using Artificial Neural Network for Yola, Nigeria

Short term forecast for solar radiation using Artificial Neural Networks has been designed for Yola. Twelve months hourly data were used and two distinct ANNs (ANN1 and ANN2) were employed. ANN1 was applied to the time series of solar radiation data alone to forecast future solar radiation; and wind speed, wind direction, ambient temperature, relative humidity and barometric pressure were used as inputs to ANN2 to obtain forecast for the same solar radiation. The results for the ANN1 and ANN2 models forecasts of the 1 to 6 hr ahead solar radiation with respect to their performance evaluation have shown that the maximum values of MAE, RMSE and MRE were 49.93, 63.20 and 0.194 respectively for ANN1; and 159.00, 177.21 and 0.582 respectively for ANN2. The coefficient that determines the strength of their correlation with the observed data (R) were found to be 0.992 for ANN1 and 0.939 for ANN2 hence ANN1 is tentatively a better solar radiation forecasting model for Yola.