“Our goal is to create reliable models for intelligent forecasts of both electricity usage and solar power production. This may become the foundation for creating smart power grids”, says Joakim Munkhammar, Project Manager for the Built Environment Energy Systems Group at Uppsala University.
Integrating solar energy in the power system on a large scale can be problematic, since solar electricity production varies rapidly when clouds pass over the solar panels. Changes between sunny and cloudy conditions could present a challenge to local grid owners with a large share of solar power in their grids, not just in Sweden but in large parts of the world.
Uppsala University has launched the project together with the Foundation for Collaboration between the Universities in Uppsala, Business, and the Public Sector, and Region Uppsala. The idea is to train computers to predict solar power production and electricity usage by using statistical models and artificial intelligence. The project is funded by the Swedish Energy Agency’s research and innovation programme ‘Samspel (in Swedish)’.
“The number of local, decentralised energy systems such as solar power plants is increasing a lot right now in many parts of the world. Since renewable energy often presents great challenges to the energy system, the results of our project could become very important for the power grids of the futures,” says Fredrik Björklund, Project Manager at the Foundation for Collaboration between the Universities in Uppsala, Business, and the Public Sector.
Variations in solar irradiance have been studied for over 100 years, and in recent years this research has become all the more relevant in the design and application of solar power facilities. The project, which is named ‘Development and evaluation of forecast models for solar electricity and electricity usage over time and space’, builds on previous projects involving statistical modelling of electricity use and solar irradiance, developed by Uppsala University.