A short-term wind power probability prediction method based on soft clustering and similarity measurement

Abstract

With the rapid development of wind energy, probabilistic forecasting of wind power becomes increasingly crucial for reliable operations of power grids. This paper proposes a wind power interval prediction method based on temporal data soft clustering and similarity measurement (SCSM). First, a soft clustering module is used to cluster wind power data with probabilities. Next, a similarity measurement module assesses the similarity between wind power data based on soft clustering results and generates probability interval predictions by referring to historical prediction errors. Finally, the effectiveness of the proposed method is validated using real wind power data.