Comparison of Two Solar Probabilistic Forecasting Methodologies for Microgrids Energy Efficiency

In this paper, the performances of two approaches for solar probabilistic are evaluated using a set metrics previously tested by meteorology verification community. A particular focus is put on several scores and decomposition specific metric: continuous rank probability score (CRPS) as they give extensive information to compare forecasting performance both methodologies. The methodologies used produce intra-day forecasts with time horizons ranging from 1 h 6 h. first methodology based steps. step, we generated point forecast each horizon in second use quantile regression methods estimate prediction intervals. directly estimates intervals forecasted clear sky index distribution past data inputs. With also propose add geometric angles Overall, nine compared at six measurements stations different climatic conditions. This paper shows detailed picture overall models consequently may help selecting best methodology.