Open Source Tool for Probabilistic Short-Term PV and Wind Power Forecasting

The effective integration of renewable energy sources (RES) into the power grid is a challenging problem faced by power system operators and various stakeholders. In this context, a large number of forecasting models has been developed aiming at alleviating this problem. However, most of those approaches are deterministic models that generate single point forecasts without providing any information about the uncertainty associated with the prediction. Probabilistic forecasting has recently emerged in order to capture the uncertainty induced by the stochastic nature of RES. Nevertheless, the limited number of probabilistic approaches reported in the literature deploy complex black box models that may require long training times, extensive hyperparameter tuning, feature engineering, and expert domain knowledge. Under those conditions, their practical implementation at scale and their integration into system operation may be questioned. To this end, we introduce an open source tool for probabilistic short-term PV and wind power forecasting that can be readily deployed by anyone with basic programming knowledge. Finally, the performance of the implemented algorithms is evaluated using real data.