Harnessing AI for Grid Management

In his seminar, AI4OPT External Advisory Board member Patrick Panciatici explored the intersection of artificial intelligence and power system operations. Drawing on four decades of experience, Panciatici addressed how machine learning, when combined with traditional optimization methods, can significantly improve the management of complex power grids, particularly in the face of uncertainty and operational constraints.

Topology Optimization as a Case Study

A central focus of the seminar was topology optimization—a challenging problem in grid operations that involves configuring the structure of the network to ensure stability and efficiency. Panciatici emphasized that data-driven models, when integrated with domain expertise and proven optimization strategies, can enhance both the efficiency and scalability of these systems. This approach holds promise for enabling smarter and more adaptive decision-making frameworks.

Toward a Foundational Model for Power Systems

Looking ahead, Panciatici proposed the development of a foundational model for power systems. The model would capture generalizable patterns across tasks and support robust, flexible decision making.

A Legacy of Innovation and Collaboration

Panciatici brings a wealth of experience to this topic, having served as Senior Scientific Advisor at RTE, the French transmission system operator, and as a key contributor to numerous international research initiatives. He remains actively involved in academic and industry collaborations through roles at the University of Liège and several European research organizations. His work continues to shape the evolving dialogue between AI research and critical infrastructure.