Probabilistic and Deterministic Wind Speed Prediction: Ensemble Statistical Deep Regression Network

Wind energy as one of the most promising alternatives brings a set serious challenges in operation power systems because uncertain nature wind speed. To address this problem, it is essential to establish framework forecast comprehensive form information about end, an ensemble residual regression deep network designed understand fully time-variant and spatial features from historical data including speed corresponding meteorological data. Then, enhance accuracy, modified error-based loss function proposed. Consequently, provide information, kernel density estimator proposed extract probability functions (PDFs) with high level accuracy reliability. The simulation results comparative analysis on actual dataset London, U.K. demonstrate capability probabilistic approach.