Sustainable Engineering Chair
PhD position (II) in data science at the Université libre de Bruxelles (ATM)
“Explainable machine learning applied to the development of smart grids”
In the coming years, the electric grid will need to evolve drastically in order to meet the demands imposed by climate change and the clean energy revolution. The increasing presence of intermittent renewable energy is also affecting the reliability of the grid, making it harder to match supply and demand. The development of smart grids can be a solution to this problem. A smart grid can be conceptualized as a digital twin of the electric grid, able to predict its behavior in real time. Machine learning, together with the ever-growing availability of sensors, offers the tools to build such a digital model of the grid. Machine-learning techniques have already been employed with great success to discover fundamental patterns, forecast the behavior of dynamical systems and model the input-output relationship of complex systems. However, machine-learning models are often black-box models, meaning that their decision process is not understandable and transparent. This is an obstacle that limits heavily the use of machine learning for critical infrastructure, such as the electrical grid where each action taken by the control algorithm should be subjectable to review. On top of that, it is important to avoid that the output of the machine-learning model perpetuates the systemic bias present in the training data.
The objective of this PhD project is to leverage physical information and employ explainable machine-learning tools (such as symbolic regression and physics-based modelling) to provide insights on the inner working mechanism of a smart grid.
These insights will be exploited with two main goals:
- Improve the robustness and interpretability of the machine-learning model.
- Explore the possibility of including socio-economic constraints (affordability, pollution, etc.) in the smart grid.
Offer Description
This PhD project is part of the Sustainable World Initiative & Fellowship For Transformation (SWIFFT). SWIFFT is a joint research project funded by the Helios Foundation that brings together researchers from ULB and VUB to tackle different aspects of the energy transition, ranging from sustainable engineering, transformation and wellbeing to finance and law for a just energy transition (www.swifft.be). This PhD project will be co-supervised by Prof. Alessandro Parente and Dr. Alberto Procacci.
Prof. Alessandro Parente research is in the field of turbulent/chemistry interaction in turbulent combustion and reduced-order models, non-conventional fuels and pollutant formation in combustion systems, novel combustion technologies, numerical simulation of atmospheric boundary layer flows, and validation and uncertainty quantification
Dr. Alberto Procacci works on the development of reduced-order models, with a focus on physically constrained models, forecast of dynamic systems and data assimilation. His overall research goal is to combine physical knowledge with state-of-the-art machine-learning techniques to develop surrogate models that can be employed as digital twins.
Requirements
Skills and Qualifications
We seek a candidate with a strong background in the following fields:
1 – Machine-learning and deep learning models,
2 – Data science,
3 – Numerical modelling.
Offer requirements
- A Master of Science in engineering, data science, physics, or applied mathematics with a focus on either energy engineering, electrical engineering or machine-learning.
- A qualification equivalent to first-class honors degree is preferred.
- Experience in numerical methods, excellent computational skills, expertise in programming (Python, Fortran, C++).
- Interest in social studies, deep learning, data-driven modelling, optimization,
- English language is mandatory.
Selection process
The selection process is based on two steps:
- Evaluation of the documents provided by the applicant
- Interview of each candidate having the eligibility requisites (evaluated through the first step). Interviews will be organized remotely.
The list of documents to be provided:
- Letter of motivation (approx. 1 page),
- Copies of degree and academic transcripts (with grades and rankings),
- Summary of the master’s thesis (approx. 1 page),
- Short CV including a publication list (if any),
- Two reference letters from academics,
- Proof of English language skills.
Where to apply?
Please submit your application through the SWIFFT portal using the button below, or send it to:
Please name your documents following this format: SURNAME-NAME_NameOfTheDocument_position2.pdf. (ex.: SURNAME-Name_CV_position2.pdf; SURNAME-Name_MotivationLetter_position2.pdf)