Predicting output of solar plants using weather information

What is it about?

Solar power varies with cloud coverage and aerosol loading, as well as by time of day. Here we demonstrate a systems approach to short-range forecasting of power from solar farms. We use weather models, artificial intelligence, real-time observations, and historical observations to make these forecasts. We show a high degree of success. The work done in this project brought together collaborators from private industry, academia, and government labs to build and demonstrate this system. We used methods from social science to integrate the team and assessed the economic value of the final system.

Featured Image

Why is it important?

In order to better manage solar plants and integrate them into the grid, it is important to have accurate forecasts of expected production. We analyzed multiple methods and showed the value of such forecasts.