Integrating Machine Learning for Safe and Sustainable Power System Operation

ABSTRACT

Efforts toward decarbonization and environmental sustainability have driven the unprecedented integration of new energy technologies into power systems. These changes span the electrification of multiple sectors, the widespread adoption of renewable resources, and the rise of digitalization to enhance large-scale grid monitoring and control. These transformations have also accelerated the adoption of machine learning (ML) and artificial intelligence (AI) in decision-making processes, introducing complex modeling and computational challenges, including increased reliance on data-driven tools. In this talk, I will present my work, which (i) develops a methodology for fine-tuning low-fidelity surrogate models for power systems and (ii) establishes a framework for end-to-end learning with hard constraints to enable safe real-time operations.

BIOGRAPHY

Gonzalo Constante is a postdoctoral scholar at Purdue University. He earned his Ph.D. in Electrical and Computer Engineering in December 2022 from The Ohio State University (OSU) and obtained an M.S. degree in ECE in 2018. Before attending OSU, he completed a B.E. in Power Engineering at Escuela Politécnica Nacional, Ecuador, in 2014. His research interests include modeling, optimization, simulation, and the economics of power and energy systems, focusing on developing physics-based and data-driven tools for modern power systems. Gonzalo has received several awards, including a Fulbright Scholarship (U.S. Department of State) and a Presidential Fellowship (OSU).