In this project, the student will investigate different algorithms to create forecast models for renewable electricity generation. More concretely, the student will incorporate two important factors in the analysis:
(1) how can uncertainties in the forecast be effectively quantified, and
(2) how can spatial hierarchies be preserved while making these forecasts.
Forecasting renewable electricity generation accurately is an increasingly important challenge. Currently forecast errors are scaling linearly with installed capacity - a situation which is clearly not tenable in the long run. This means that uncertainties in forecasts must be clearly quantified and communicated to all important stakeholders. Likewise, incorporating hierarchical aspects is an important topic as uncertainties vary considerably across production source and location (i.e. simply adding uncertainty estimates in region 1 and region 2 will not produce correct overall uncertainty estimates).
The project starts per 1 September 2025.
Number of placements available: 1 per semester.
Prerequisites
- Familiarity with Python programming langauge.
- Some knowledge of forecasting and/or machine learning is highly encouraged.
Faculty Department
The research domain of the ELECTA division covers the broad spectrum of electrical energy systems and robust control of industrial systems. As the largest research group on energy systems and fault tolerance in the Benelux, ELECTA's vision is to be widely recognized as centre-of-excellence on these topics, where fundamental research is coupled with immediately and prospectively applicable solutions for the industry. Alongside this development of know-how, a pivotal importance is given to sharing gained knowledge with the academic community, students, industry and the society as a whole.