The electrical power grid is one of the most essential infrastructures in modern society. The next-generation electrical power infrastructure, known as the “smart grid,” is a comprehensive upgrade of power and energy systems with emerging information and communication technologies. While these fundamental changes are reinforcing the functionalities of power grid, they also raised serious security concerns.

Dr. Sun and Dr. He tackles the power grid security challenges from the information, signal processing, and networking perspectives.  When the team received the National Science Foundation grant for addressing the smart grid security issues in 2011, this topic was traditionally studied by power engineers who rarely integrated information technology into this area. The team has overcome the barriers between the power engineering society and the information technology society, and advanced the understanding of power grid vulnerability through novel attack strategies.

The research team discovered two cyber-physical attack models, joint attacks and sequential attacks, which can cause power grid cascading failures (i.e. large-scale blackout). In the joint attack model, multiple malignancies (e.g. failures) occurring simultaneously, at transmission lines, substations, or both. In the sequential attack, a series of malignancies orchestrated at different operation points of a power system, resulting from multiple attacks and system responses. They developed vulnerability analysis tools that identified specific attack strategies. The results surprised the power industrial and research community that smart attackers can cause significant damage to the power grid if they use the joint/sequential attacks targeting some hidden critical components. These hidden critical components were previously overlooked by mainstream vulnerability analysis approaches which focus primarily on the loads and connectivity.  Furthermore, the nominee developed visualization tools for decision-making assistance.

The team is currently working on defending cyber-physical attacks against power system, using machine learning tools and adaptive control.