Novel cross-domain analysis of cyber-physical power systems using enhanced modelling and simulation techniques

Abstract

The modern power grid demands advanced modelling and simulation methods to address its evolving nature and increasing complexity. A critical focus is the appli-cation of cyber-physical systems (CPS) theory, especially with the rise of 5G, smart inverter-based resources (IBRs) and the prevalence of extreme weather events and cyber incidents. Research in cyber-physical power systems (CPPS) is pivotal for en-hancing grid reliability and resiliency by enabling for the design and analysis of future grid-applicable cyber technologies. While power system and communication network simulators suffice in their re-spective domains, integrated analysis is required. Consequently, this thesis pro-poses a whole-system graph-based CPPS model that links reduced power systems with abstracted communication networks, enabling cross-domain analysis. A novel Cyber Node Importance Index (CNII) assesses the criticality of cyber nodes by con-sidering factors such as cascading failures and centrality measures. The CNII is applied on a GB-based CPPS model, demonstrating its applicability in identifying vulnerabilities and informing mitigation strategies for real and large infrastructure. This thesis further details a methodology for the conversion of power system mod-els from offline to real-time to enable real-time testing and dynamic CPPS analysis under communication contingencies. Also, the architecture of a real-time testbed is proposed. This testbed aims to evaluate CPPS resilience to cyberattacks and com-munication disruptions, integrating power system simulation, communication emula-tion, control schemes and industrial control system (ICS) protocols. Key contributions of the thesis focus on improvements in taxonomy, vulnerability analysis, offline to real-time power system model conversion and real-time testbed design, collectively advancing CPPS modelling and simulation approaches.