Wenbo Chen, an AI4OPT graduate research assistant and Machine Learning Ph.D. student at Georgia Tech, presented his Ph.D. defense seminar titled "Synergizing Machine Learning and Optimization: Scalable Real-time Risk Assessment in Power Systems" on April 5th. His focus was on integrating renewable energy into power systems and the need for real-time risk assessment, crucial for US Independent System Operators (ISOs). Traditional optimization-based methods were impractical due to time and complexity.

To tackle these challenges, Wenbo developed optimization proxies using machine learning, aiming to integrate Machine Learning (ML) with optimization for real-time risk assessment. He created a just-in-time ML pipeline for Security-Constrained Economic Dispatch (SCED) problems in the US energy market, handling load variability and complex decision-making.

Wenbo also introduced an End-to-End Learning and Repair (E2ELR) architecture, enhancing feasibility and scalability in real-time risk assessment methodologies. Wenbo's research provides insights into integrating AI and optimization in power systems, benefiting industry stakeholders and policymakers working on sustainable energy solutions. His innovative approaches offer potential solutions to real-world challenges, advancing real-time risk assessment in large-scale power systems.

To read more details, click here.