Hour-Ahead Solar Forecasting Program Using Back Propagation Artificial Neural Network

Abstract

Solar photovoltaic power generation highly relies on solar irradiance, cloud cover variability, temperature, atmospheric aerosol levels, and other atmosphere parameters. Accurate forecasting of solar power is crucial to very short-term generation scheduling and on-line secure economic operation. In this paper, hour-ahead forecasting using BP-ANN is proposed. The inputs of BP-ANN include previous intervals of solar irradiation, moving average temperature, moving average relative humidity, time of the day and day of the year index. The supervised learning ANN render a higher accuracy with the good convergence mapping between input to target output data. The simulation of hour-ahead solar irradiation forecasting results from ANN render a better performance compared with autoregressive moving average model in terms of mean absolute Error (MAE), root mean square error (RMSE), mean absolute percentage error (MAPE), mean bias error (MBE) and correlation coefficient (Corr).