Collaborative Research: Learning for Safe and Secure Operation of Grid-Edge Resources

This NSF project aims to address the challenges and opportunities presented by the rapid proliferation of Grid Edge Resources (GERs) in modern power systems. Examples include distributed generators and smart inverters, smart thermostatically controlled loads, electric vehicles, and battery energy storage systems. Since GERs operate beyond traditional utility network boundaries and are controlled by customers, they introduce variable levels of controllability, observability, and vulnerability to cyber-attacks. The project will bring transformative change to the field of power system management through the development of a new analytical foundation and data-driven control methodologies to ensure the safe and secure operation of GERs. The intellectual merits of the project include the development of novel algorithmically robust data-driven control strategies that can withstand the unavoidable cyber vulnerabilities of GERs, and the advancement of our understanding of GER behavior and its impact on power system dynamics. The broader impacts of the project include enhancing the safety and security of the nation's critical energy infrastructure, improving the reliability of artificial intelligence and data-driven control methods across various safety-critical engineering systems, and promoting diversity and inclusion in two minority-serving institutions. The technical objectives of this project will be achieved by introducing a novel combination of model-based and data-driven control methods to guarantee that GERs are operated without violating power distribution systems’ constraints, despite the lack of direct control and validation capabilities in managing GERs in real-world power systems. Our approach ensures network-safe exploration and data-driven control at any stage of operation, despite model uncertainty. To address the challenge of unavoidable corrupt inputs from GERs, such as corruption in sensed load, we will develop grid edge control algorithms that are algorithmically robust to vulnerabilities in GERs. The proposed methods and results will be tested under realistic scenarios, considering diverse characteristics of various GREs, and under different network operating conditions and constraints, using real-world GER data and industry-standard computer simulations. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.