Solar forecasting for smart grid energy management (Speaker: Prof.dr. Wilfried van Sark, Utrecht University, Copernicus Institute of Sustainable Development)

Speaker: Prof.dr. Wilfried van Sark, Utrecht University, Copernicus Institute of Sustainable Development

With the fast growing amount of photovoltaics (PV) solar panels and electrical vehicles the traditionally organized electricity grid needs to change: electrical power will be flowing in two directions. For proper management of these power flows better supply and demand forecasting is required, focusing on high spatial and time resolutions. Knowing when the sun is shining and when you need your electrical vehicle are essential inputs for local energy management systems.

We have developed a method for solar forecasting based on the use of PV systems as irradiance sensors; we explicitly did not use any weather information. The systems were located mainly in the city of Utrecht. From the cross-correlation of power between the systems, we were able to deduce the direction of clouds passing over the city. The method outperforms forecasting methods based on persistence for cloudy weather, and provides accurate forecasts for periods between 1 and 15 minutes.

In the presentation, I will also present ideas for future work that will include connecting with actual weather forecasts.