Title:

Perspectives on using Machine Learning to operate large power grids

 

Abstract:

Managing complex power systems requires solving large-scale decision-making problems under uncertainty and operational constraints. This presentation investigates the integration of machine learning with standard optimization techniques to enhance system operation, with a focus on topology optimization as a representative use case. By combining data-driven models with domain-specific knowledge and optimization methods, we aim to improve both efficiency and scalability. The discussion concludes with a forward-looking perspective on developing a foundational model for power systems—one that captures generalizable patterns across tasks and supports robust, flexible decision-making.

 

Bio:

Patrick Panciatici - Former Senior Scientific Advisor at RTE (French Transmission System Operator).

He joined EDF R&D in 1985 and then RTE in 2003 when he participated in the creation of an internal R&D department at RTE.

He has 40 years of experience in power systems: modeling, simulation, control and optimization.

He interacts with a large network of international experts and academic teams worldwide on these topics.

He is a member of CIGRE and a Fellow of IEEE.

He has been the RTE representative in two US initiatives: PSERC and Bits&Watts.

He is currently: Member of the Advisory Board of the NSF AI Research Institute for Advances in Optimization (AI4OPT),

Scientific Collaborator of the University of Liège (Belgium),

Co-PI of the ANITI Industrial Chair "POPML4PS: Combining Polynomial Optimization and Machine Learning : Application to Power System Decision Support Tools" (Toulouse), Associate Researcher at the Laboratory of Signals and Systems, a French laboratory jointly run by the CNRS, CentraleSupélec and the University of Paris-Saclay,

and Independent adviser of CRESYM "Collaborative Research for Energy SYstem Modelling".