UniWind: Toward Unified Day-Ahead Wind Power Forecasting via Physics-Informed State Routing

2026-07-01

AI Summary

This study addresses the prediction bias in wind power forecasting caused by the entanglement of meteorological factors and latent turbine operational states—such as curtailment or shutdown—by proposing UniWind, a novel framework that integrates physics-informed state routing. UniWind synergistically combines site-calibrated physical priors with operational-state-aware expert corrections through several key innovations: knowledge-guided supervised state routing, bounded expert adjustment, shared physical power curves, physics-based upper-bound constraints, and zero-shot cross-site transfer capability. Comprehensive evaluations across more than twenty real-world wind farms, including both full-sample and zero-shot cross-site experiments, demonstrate that UniWind substantially improves prediction accuracy and robustness compared to existing approaches.

Abstract

Day-ahead wind power forecasting is essential for cost-effective power-system operation. It is primarily driven by future meteorological conditions while retaining temporal dependencies in power generation. In practice, observed wind-farm power often entangles physically available power with local environmental effects and latent operational states, such as shutdowns and curtailment. Existing physical models provide useful constraints but adapt poorly across wind farms, whereas data-driven models can capture rich correlations but often conflate meteorological effects with state-induced deviations. In this study, we propose UniWind, a wind power forecasting model based on physics-informed state routing. UniWind first employs a Physical Prior Estimator to construct a site-calibrated physical prior by combining site-conditioned monotonic warping with a shared physical power curve. It further applies a physical upper-bound constraint to shape this prior as a soft envelope of available wind power generation. UniWind then proposes a Latent State Encoder to model operating-state embeddings and transforms the physical prior into final power forecasts through a State-aware Power Corrector, which uses knowledge-guided supervised state routing and bounded, state-specific expert correction. Full-shot and cross-farm zero-shot experiments on more than 20 real-world datasets demonstrate the accuracy and robustness of UniWind.

Problem

• wind power forecasting
• latent operational states
• physical constraints
• meteorological effects
• state-induced deviations

Innovation

• physics-informed
• state routing
• wind power forecasting
• latent state modeling
• zero-shot generalization