PhysDiffWind: A physics-constrained retrieval-augmented diffusion framework for offshore wind speed forecasting

doi:10.1016/j.asoc.2026.114647

PhysDiffWind: A physics-constrained retrieval-augmented diffusion framework for offshore wind speed forecasting

2026-01-13

0

PRE

AI

DOI：10.1016/j.asoc.2026.114647

Original

Original request for help

Save

Abstract

En 中文

• A physics-constrained retrieval-augmented diffusion framework is proposed for offshore wind forecasting. • Retrieve multivariate sequences using temporal convolutional encoding and embedding similarity. • Incorporate physical constraints through pressure gradient and air density embeddings. • Extensive experiments demonstrate superior accuracy and uncertainty quantification over state-of-the-art baselines.

PhysDiffWind: A physics-constrained retrieval-augmented diffusion framework for offshore wind speed forecasting

Abstract

Journal

Organization