Skip to content

PhD Position Energy Flexibility Aggregation and Coordination with Learning-based Predictive Control

  • Genk, Vlaams Gewest, Belgium

Job description

In modern power systems, Distributed Energy Resources (DERs) such as solar panels and micro-generators are being integrated into electricity systems through hierarchical market structures, where groups of DERs interact with the system as a single entity to provide flexibility and increase the penetration of renewable resources.

At the aggregator level, where there is a multitude and variety of devices and appliances across the aggregator’s portfolio (building HVAC, EVs, water tanks etc), the process of control and scheduling is rendered infeasible without automating a big part, or the whole process.

In this settings, the process of scheduling and control flexible devices involves the energy management of numerous small entities, and also needs to be solved online and under uncertainty. This is because, the increasing integration of renewable resources in the electric power grid has increased the need for aggregators to offer/trade (or correct their offer) and dispatch their flexibility close to real-time (e.g. intraday/continuous intraday market) in order to better deal with the high uncertainty due to the availability of renewables.

Regarding these challenges, a certain part of the literature has focused on designing schemes for making efficient dispatch decisions under uncertainty. However, current methods cannot guarantee a good ‘anytime’ solution for real-time control and scheduling mainly due to computational scalability issues.

Recently progress in the AI field has proven the possibility to merge learning based algorithms (to cope with complex stochastic models) and model-based MPC-like approaches (for robustness in real-world safety-critical applications). The idea behind those approaches is that employing structural knowledge about the optimization model is paramount to achieving both feasibility and optimality.

The main goal of this PhD project is to bring energy flexibility to the real world of aggregation and coordination by designing novel approaches that combines classical model predictive controllers and data-driven learning to overcome challenges such as dimensionality, operational and economical constraints under uncertainty.

As a PhD candidate, your role will be engaging the above challenges, with a particular focus on control and optimization methodologies for real-time energy flexibility quantification and activation.

The approach will be based on historic data and demonstrated on real-world use cases.

PhD supervision:

The PhD fellowship is granted within the collaboration between the University of Ghent (UGent) and VITO. The successful candidate will be supervised by Prof. Guillaume Crevecoeur (UGent) and co-promoted by Dr. Carlo Manna (VITO).

Registration and Budget:
Fulltime 4 year position

: The PhD student will be registered at the University of Ghent.
Funding mode:
Grant student at university.
Work Location
: The PhD student will split their time between UGent and the VITO EnergyVille campus, according to a schedule agreed upon by both institutions.

How to apply?

Applications should be submitted online and include a copy of your CV, diploma transcripts and a cover letter.

More information about the application procedure is available on the VITO website.

You can register until September 11, 2024 for the jury of November 8, 2024.

Job requirements

  • You hold M.Sc. degree relevant to the position (e.g., electrical engineering, computer science, etc.);
  • You are fluent in English, both oral and written;
  • You have strong analytical skills, and a strong sense of responsibility, you are an excellent communicator and you are also able to autonomously plan and perform research tasks;
  • Experience with software development using Python, Java, Matlab etc., is a big plus.