Congestion and observability across interdependent power and telecommunication networks under seismic hazard. Issue 4 (November 2021)

• Record Type: Journal Article
• Authors: Talebiyan, Hesam; Leelardcharoen, Kanoknart; Dueñas-Osorio, Leonardo; Goodno, Barry J; Craig, James I
• Is Part Of: Earthquake spectra. Volume 37:Issue 4(2021)
• Page Start: 2892
• Page End: 2919
• Publication Date: 2021-11
• Subjects: Lifeline systems -- probability theory -- telecommunications systems -- electric power systems -- resilience -- interdependency; Earthquake engineering -- Periodicals; Earthquake engineering; Periodicals; 624.1762
• DOI: 10.1177/87552930211026690
• Languages: English
• ISSNs: 8755-2930

Abstract

This article quantifies the seismic performance of interdependent electric power and telecommunication systems, while also identifying variables with the highest impact on design. We introduce interdependent power and telecommunication models, which probabilistically simulate the physical dependency of telecommunication systems on power via interdependent adjacency and coupling strength, while a topology observability analysis quantifies the cyber dependency of the power system on telecommunications. We also use new functionality-based performance measures, including data congestion in telecommunications and partial observability in power systems, given communication demands upsurging after earthquakes. As an application, our methodology assesses the performance of stylized power and telecommunication systems in Shelby County, TN. Results show that neglecting retrials, congestion, and power interdependency lead to significant overestimation of the performance of telecommunication systems, particularly at low-to-medium hazard levels. Sensitivity results also reveal that decreasing the strength of coupling across systems is one of the most effective ways to improve the seismic performance of evolving cyber-physical systems, particularly when increasing observability in the power system through telecommunication end offices with richer data flow pathways.