Hourly probabilistic forecasting of solar power

The growing integration of variable renewable energy resources, poses new challenges to the electric grid. Forecast of renewable generations may be deployed to mitigate some of these challenges. There are always errors associated with even the most robust solar power point forecasts; so a great deal of attention has been paid to probabilistic forecasting to quantify the uncertainty and risk associated with point forecasts. The ensemble learning tool, the random forest is used to combine the individual models to obtain hour-ahead combined point forecasts and then to generate the ensemble-based probabilistic solar power forecasts. Comparisons are provided with probabilistic forecasts that are generated by the analog ensemble technique as well as the baseline persistence probabilistic technique. The evaluation is carried out over the entire year, and it is found that the ensemble-based and analog ensemble probabilistic forecasts have almost the same accuracy.