Leveraging generative models for enhanced solar irradiance ramp detection

Abstract

Short-term variations in PV power are an increasingly important challenge for solar energy integration. By anticipating sudden changes in irradiance caused by passing clouds, all-sky imager-based solar nowcasting can help address this challenge. However, the utility of nowcasting systems is highly dependent on the quality of the forecast. While recent data-driven models have shown great potential in standard forecast metrics such as RMSE and forecast skill, they tend to produce smoothed forecast curves and may not be well suited to detect ramps. An alternative approach is to forecast cloud motion and irradiance separately. Here, we show that by using generative video prediction to create future sky conditions, ramps can be predicted more accurately.