Working title: Advanced modelling techniques for power markets with large share of hydropower
The objective of the PhD work is to identify what uncertainties that need to be considered in future decision support tools and how these uncertainties can be represented in long-term large-scale fundamental power market models with storage representation (hydropower).
The following research questions are defined:
What are the necessary uncertainties to consider in the future decision support tools and how can they be handled in power market models in general?
How can the realization of short- and long-term uncertainty be represented and how does the representation of different uncertainty at different time frames affect the model results?
How to deal with the interplay between variability and uncertainty in future market simulator?
The work of this PhD focusses on power market models used in the hydro-dominated Nordic power market with increasing shares of VRE and connections to Europe. The scope is fundamental power market models used for dispatch and price forecasting. While in most power systems, models used in operation and dispatch are short-term models covering the next few days or weeks, in the Nordic power system dominated by long-term storage of hydropower reservoirs, the scheduling of hydro resources requires a horizon of 3-5 years.
The PhD work will tackle modelling challenges using formal optimization, based on a bottom-up modelling approach. Using artificial intelligence and machine learning are therefore less relevant. However, using machine learning techniques to generate scenarios, make approximations for the system description to reduce the complexity.