Ex-ante and ex-post decomposition strategy for ultra-short-term wind power prediction

Highly reliable wind power prediction is feasible and promising for smart grids integrated with large amounts of wind power. However, the strong fluctuation features of wind power make the wind power less predictable. This paper proposes a novel wind power prediction approach, incorporating wind power ex-ante and ex-post decomposition and correction. Firstly, the initial wind power during the wind power decomposition stage is decomposed into a trend, fluctuation, and residual data, respectively, and the corresponding preliminary prediction models are developed, respectively. Secondly, in the error correction stage, the errors produced by the preliminary prediction model are corrected by persistence methods to compensate for final prediction errors. Moreover, the proposed model's comprehensive deterministic and probabilistic analysis is investigated in depth. Finally, the outcomes of numerical simulations demonstrate that the proposed approach can achieve good performance since it can reduce wind power forecast errors compared to other established deterministic models and uncertainty models.