When Does Advection-Aware Graph Nowcasting Help? A Controlled Study of Distributed Solar Ramp Forecasting with a Self-Supervised Cloud-Motion Estimator

This paper investigates short-term forecasting of cloud-induced power ramps across a network of distributed photovoltaic plants or irradiance sensors. Its advection-aware graph neural network connects each site to upwind sites and sets edge time lags using the cloud-motion vector, allowing ramp information to propagate before physical arrival. The approach is evaluated in a controlled synthetic testbed.