CPS: Synergy: Certifiable, Attack-resilient Submodular Control Framework for Smart Grid Stability

ABSTRACT

Large-scale Cyber-Physical Systems (CPS) must be resilient to natural disturbances and malicious attacks in a timely fashion with certifiable guarantees on safety and stability. These requirements must be satisfied even when the system occupies a large geographical area in which multiple disturbances or attacks may occur simultaneously. This effort aims to research and develop scalable and attack resilient control algorithms for CPS, based on the insight that the physical dynamics of CPS possess inherent structures, such as submodularity and bounded curvature, that can be exploited to develop efficient algorithms with provable optimality guarantees. Major activities for this year focused on research and development of control algorithms for small-signal and transient stability. We showed that the problem of selecting a minimum-size set of generators to participate in wide-area control can be viewed as combinatorial optimization of a function with bounded curvature, and demonstrated that our proposed algorithm, MinGen, requires fewer generators to ensure stability compared to current state-of-the-art methods.