How AI can help operate tomorrow’s power grid and support voltage stability in the face of challenges posed by renewables and new loads
A path, a research project
Adrien Bolland is a postdoctoral researcher at the F.R.S.-FNRS, working at the Montefiore Institute of the University of Liège. His research combines deep reinforcement learning and the operation of electrical power systems.
He trained as an electrical engineer at the University of Liège and obtained a master's degree in signal processing and intelligent robotics in 2020. He then started a Ph.D. under the supervision of Prof. Damien Ernst, which he defended in April 2025. His thesis focuses primarily on understanding the influence of exploration on policy gradient algorithms and on proposing alternative approaches to perform maximum-entropy exploration.
During his Ph.D., he had the opportunity to work on several applications in the energy sector in collaboration with various industrial and academic partners. He studied the joint design and control of microgrids, the strategic bidding of batteries on the continuous intraday electricity market, the operation of renewable energy communities, and the estimation of costs in unit commitment problems with simulation-based inference.
Since October 2025, he is an F.R.S.-FNRS postdoctoral researcher in the laboratory of Prof. Louis Wehenkel. In short, he is studying how AI can assist transmission system operators. These operators are responsible for keeping voltages stable across the electrical network. This task is becoming increasingly difficult, as the network integrates more distributed and renewable energy sources, along with new types of loads with complex and unknown behaviors. He is convinced that reinforcement learning can help with decision-making by accounting for the resulting fast, complex and uncertain dynamics, and he will spend three years studying how this conviction can become reality.
