Master’s Thesis/ Internship Project – Energy Forecasting

Locatie: Gouda
Opleidingsniveau: HBO/WO

Context

The Dutch electricity grid is under huge pressure due to rapid electrification and the growth of renewable energy sources, resulting in increasing grid congestion. These challenges demand new, innovative digital solutions that push the limits of what is physically possible.

In projects for Distribution System Operators (DSOs), there is a growing need for accurate energy forecasting within system-operations departments. The goal is to push the limits of the electricity grid by temporarily overloading it beyond static electric power limits while staying within thermal limits, using thermal inertia to absorb high peaks and manage congestion.

Through our subsidiary Phase to Phase, the market leader for grid-calculation software in the Netherlands, we have access to State-of-the-Art (SotA) software and extensive power-grid knowledge.

Research Proposal

We are looking for a talented student in Computer Science, Electrical or Mechanical Engineering, Applied Mathematics, or a comparable field who is driven to contribute to the energy transition through innovative digital solutions. 

There is vibrant research in 48-hour-ahead energy forecasting, with techniques such as Graph Neural Networks (GNNs), Transformers, State-Space Models, or other deep learning methods appearing in recent papers. You will study the current state of the art, translate the most promising techniques into practical solutions, and deliver a working prototype that demonstrates the quality of your chosen algorithm. You will receive 15-minute-interval energy load data for model development.

You will work under a primary supervisor; depending on your background, interests, and needs, additional experts will join the guidance team. In addition, we’re here to offer the support and guidance you need to reach your personal learning goals.

Depending on your interests, we will define a suitable research question. For example:

• Can we improve on existing gradient-boosting methods for energy forecasting using SotA techniques such as GNNs, Transformers, or State-Space Models?

If you are interested in the topic but your background does not exactly match or if you are looking for an applied-sciences (HBO) internship, please contact us and we will try to find a suitable solution.

Keywords

• Energy forecasting
• Congestion management
• Deep learning
• Optimization
• Grid calculation

Recruiter

Dorith Baan
Tel.: 0182594000
Email: jobs@technolution.nl