Distributed Resilience Assessment of Critical Infrastructures with Digital Twins Considering Uncertainty

Abstract

The increasing complexity and interconnectivity of critical infrastructures such as power systems - driven for example by the integration of Internet of Things (IoT) technologies - require understanding and evaluation of their resilience to cyber-physical threats. Digital Twins are a promising tool for analyzing system behavior under stress, e.g. failures and overloads, in a controlled environment. This paper presents a probabilistic simulation-based resilience assessment framework as a Digital Twin approach. Focusing on power systems, we consider cyber-physical threat scenarios, including IoT-based load altering attacks, where compromised smart devices manipulate demand patterns to destabilize the grid. By considering the inherent uncertainty of crises in impact modeling, a distributed computing approach enables simulating a diverse, randomized set of scenarios with high sampling size, allowing for a comprehensive estimation of resilience properties. Results demonstrate how different impact intensities and system configurations affect system performance, characterized by different resilience metrics. Our holistic approach improves risk and resilience analysis of critical infrastructures by incorporating uncertainty into quantitative assessments, supporting crisis management and decision-making.