THESIS - Forecasting FRR imbalance price using machine learning

Background 

The rapidly developing electrical power system with increased dynamics from more renewable energy has led to a need for increased flexible production. One way the Nordic Transmission System Operators are meeting these challenges by, is to utilize the regulation market by activating the frequency restoration reserve (FRR) products to stabilize the system.  

Historically, hydropower has been the main technology providing manual FRR (mFRR) for the power system. Lately new forms of energy storage and flexibility have entered the market of frequency regulation, like e.g. lithium-ion batteries, EV-charging stations, heat pumps and other aggregated loads. Fortum is a large provider of such services in the power system, through its large number of hydropower generating units, a growing battery energy storage fleet and a virtual battery of aggregated loads. 

Objective 

The objective of the Master Thesis Project is to do a comprehensive literature study of the balancing market, as well as theoretical models that can forecast FRR direction activation (up, down or unregulated).  

The thesis aims to develop and evaluate machine learning models to predict the direction of aFRR and mFRR imbalance prices in the electricity market. These imbalance prices play a critical role in maintaining the balance between supply and demand in real-time, and accurate predictions could help grid operators optimize decisions. This research explores a variety of data-driven approaches, leveraging historical price data, grid frequency, demand-supply forecasts and other market signals. 

Our offer 

At Fortum we see a Thesis Project as a good way to get to know you during your study time and for you to learn more about Fortum and how our business is working. A Supervisor will be appointed to support and guide you through the Thesis Project and you will also create a network of people within the Company as well as with other Students doing their Thesis with us.  

For a well performed and approved work, we will provide a taxable lumpsum award. 

This Thesis Project is preferably executed at either of our offices in Stockholm/Solna. 

Your Qualifications

We believe that you are studying Mechanical Engineering, Electrical Engineering, Engineering Physics or similar, at a Master Programme level. Course in Machine Learning is a pre-requisite. The Project can be performed in either Swedish or English. 

We expect you to be analytical, energetic and able to work independently. You like challenges and are interested in solving complicated technical problems. 

Interested? 

To apply, register your CV and a personal letter of maximum one page or a Power Point presentation of maximum two slides, in our Recruitment Tool via "Apply now", latest by 24th November 2024. Applications will be evaluated continuously and an applicant on-boarded before this date. 

The personal letter shall include an approach on how you would like to investigate the idea. Mark your application with the reference number HYG-176. 

We see this project suitable for two students. When applying, each submit an own application, in which is stated with whom one wants to co-work with.  

If you have General questions about Theses or Internship at Fortum, please address those to [email protected]. 

Please note that we have more Master Thesis Projects open for application at our recruitment site.