Project: Efficient planning of hydropower systems

In a future scenario, Sweden's production of electricity will consist of a larger share of intermittent types of energy production. Based on this, this project aims at developing new techniques for optimizing the planning and operation of large scale hydropower systems.

Project information

Project manager
Magnus Perninge
Other project members
Alexander Svensson Marcial and Johan Jonsson, Linnaeus University; Robert Eriksson, KTH Royal Institute of Technology
Participating organizations
Linnaeus University, KTH Royal Institute of Technology
Swedish Energy Agency, Swedish National Grid
Mathematics (Department of Mathematics, Faculty of Technology)
Electrical engineering (Department of Physics and Electrical Engineering, Faculty of Technology)

More about the project


In many places, hydropower forms the backbone of the electric energy production system, providing ancillary services to the grid operator. The increased penetration of intermittent renewable sources that we are presently witnessing will result in an increased demand for balancing services and higher volatility in electricity prices, adding complexity to the planning problem for hydropower producers.

The increased uncertainty calls for an adequate risk assessment already in the planning stage. Moreover, the optimization problems that arise when the hydrological couplings in the system are considered is of an in nite dimensional character. Combined, this necessitates the development of new, efficient algorithms. Our aim is to investigate how machine learning can be combined with methods from robust control and risk analysis to find efficient strategies for hydropower producers and analyse the performance of the developed strategies.


The results within the project can be divided into the following categories:

  • Modelling power markets
  • Modelling hydropower systems
  • Showing existence of optimal controls
  • Robust control
  • Risk analysis
  • Numerical computation algorithms

Research theme

The project is part of the research in the theme AI and machine learning for optimization and operations reserach.