It is expected that the Nordic power system of the future will increasingly consist of solar, wind and water energy. The industry will increasingly need tools that can express uncertainty in the forecasts in a satisfactory manner.
The idea of this project is to investigate whether it is possible to create hybrid hydrological models that use Bayesian networks in combination with physically based models. We want to develop models that can deal with uncertainties and also be quick to run in an operational setting.
This will require knowledge of hydrology and hydrological models, as well as knowledge of Bayesian networks and how these can be used purely technical modeling to represent hydrological processes.
To investigate this type of hybrid models can be very transferable to other disciplines relevant to the University of Agder in particular areas that CIEM is working on, as emergency response and crisis management in case of fire. The algorithms and methods to be examined in this project will have great relevance.