Forecasting and Downscaling Solar Irradiance Using Transformers

As solar energy continues to grow as a key component of the global energy mix, accurate forecasting of solar irradiance becomes more crucial for ensuring reliable electricity supply. Reliable forecasts are necessary for managing the variability of solar power generation and optimizing the integration of renewable energy sources into the grid. However, current forecasting methods often fail to capture the fine temporal variations in solar irradiance, particularly in regions where local weather conditions play a significant role. This research addresses the growing need for accurate solar irradiance forecasting to optimize the integration of solar energy into the grid. By working with raw, unprocessed data, the research aims to preserve important short-term variations that are key for accurate predictions. The focus of the study was on downscaling global solar irradiance data from a 15-minute resolution to a higher 5-minute resolution for Brookings, South Dakota. A transformer-based model was utilized to predict solar power output, exploring different approaches to evaluate the effectiveness of various downscaling methods. The model was trained on historical data to generate 24-hour short-term forecasts, with performance assessed using standard error metrics. The results demonstrate the potential of transformer models to enhance solar irradiance forecasting while also highlighting the challenges and advantages of using raw data for this task.