Renewable generation forecast studies – Review and good practice guidance.

Abstract Propelled by the actual demand from the renewable energy industry, the progress of literature on quantitative forecasting models during the past years is extensive. Research provides a vast output of papers on wind speed, wind power, solar irradiance and solar power forecasting models, accompanied by models for energy load and price forecasting for short-term (e.g. for the intraday trading schemes available at many market places) to medium-term (e.g. for day-ahead trading) usage. While the models themselves are, mostly, rather sophisticated, the statistical evaluation of the results sometimes leaves headroom for improvement. Unfortunately, the latter may occasionally result in the rejection of papers. This review aims at giving support at this point: It provides a guide on how to avoid typical mistakes of presenting and evaluating the results of forecasting models. The best practice of forecasting accuracy evaluation, benchmarking, and graphically/tabularly presenting forecasting results is shown. We discuss techniques, examples, guide to a set of paragon papers, and clarify on a state-of-the-art minimum standard of proceeding with the submission of renewable energy forecasting research papers. Highlights • We provide a baseline for the empirical assessment of renewable energy forecasting. • Prices, wind speed/power solar irradiance or price and load forecasts are covered. • We show how to present and evaluate forecasting results. • Performing and evaluating OOS studies and probabilistic forecasts is shown. [ABSTRACT FROM AUTHOR]