A novel digital twin framework of electric power infrastructure systems subjected to hurricanes

Abstract

The electric power network (EPN) is one of the most critical infrastructure systems as most lifeline, economic, and social systems depend heavily on it, and any disruption in the network may affect the well-being of modern societies. Being the most vulnerable to natural hazards, the resilience of the EPN has received plenty of attention in recent years, particularly considering the increasing frequency and severity of natural hazards associated with climate instabilities. The data revolution and the recent advances in the fields of artificial intelligence (AI), machine learning (ML), and the Internet of Things (IoT) have prompted researchers to take the next step and expand the available predictive models toward digital twins (DT). However, there is still a lack of an applicable framework for a DT of infrastructure systems in the face of disasters. In this paper, a novel DT framework of the EPN when subjected to hurricanes is proposed that combines physics-based and data-driven models while also employing a dynamic Bayesian network (DBN). The DBN can be updated in near real-time via data sensing to provide a DT that is simple, computationally feasible, scalable, and capable of modeling and estimating the failure and performance states of the various elements of the EPN. The proposed DT framework is applied to Galveston Island's EPN, and the results are validated using historical data, demonstrating that the DT can produce detailed and highly accurate estimations to be used in decision-making for community resilience planning.