Cyber-Physical Networked Microgrids

Power grid stability and security are challenging problems with significant economic and social impacts that have been exacerbated in recent years by the increase in extreme weather events and cyberattack concerns. Recently, networked microgrids (NMs) have become an emerging paradigm that demonstrates resiliency benefit to their local customers. U.S. DOE has envisioned that the R\&D of NMs will be the next wave of smart grid research to achieve the nation\u27s grid modernization vision towards climate adaptation and resiliency. However, lack of awareness of stability margin, inadequate capability to respond to grid disturbances, and vulnerabilities to communication failure, delay, and cyberattacks all contribute to undermining the capability of NMs to improve distribution grid resiliency. To tackle these issues, a set of novel methods are devised in the cyber and physical layers of NMs. First of all, Formal Analysis (FA) via reachable set computation is established in the physical layer of NMs to efficiently assess their stability in the presence of heterogeneous uncertainties induced by high penetration of distributed energy resources (DERs). FA is further combined with quasi-diagonalized Gersgorin theory to identify systems\u27 stability margins. Second, based on FA, Distributed Formal Analysis (DFA) is developed to efficiently investigate the stability of interconnected power grids by decoupling large-scale NMs via N+M decomposition approach. A programmable data exchange mechanism is also developed in DFA to enable a privacy-preserving approach which helps guarantee the privacy and security of information. Third, as an essential function for NM operation, Compositional Power Flow (ComPF) is devised for assessing steady states of NMs, which for the first time takes into account power sharing and voltage regulation between microgrids. Fourth, to quantify unintentional islanding hazards, this dissertation contributes a DER-Driven Non-Detection Zone (D2NDZ) method, which is a data-driven, learning-based approach. Beyond the NM physical layer, a Software-Defined Active Synchronous Detection (SDASD) is designed and implemented in the cyber layer of NMs to protect them from cyberattacks on Software-Defined Networking (SDN) and power bot attacks on DERs. Case studies have demonstrated and validated the effectiveness of FA/DFA, ComPF, D2NDZ, and SDASD in analyzing NM stability under disturbances, quantifying steady state power flows in islanded NMs, determining unintentional islanding frequencies, and protecting NMs from attacks. The new technologies collectively lead to a set of powerful tools for planning, operating, and protecting future NMs