A demand-based framework for resilience assessment of multistate networks under disruptions. (June 2022)

• Record Type: Journal Article
• Authors: Geng, Sunyue; Liu, Sifeng; Fang, Zhigeng
• Is Part Of: Reliability engineering & system safety. Volume 222(2022)
• Publication Date: 2022-06
• Subjects: Multistate networks -- Resilience assessment -- Markov reward process -- Universal generating function; Reliability (Engineering) -- Periodicals; System safety -- Periodicals; Industrial safety -- Periodicals; Fiabilité -- Périodiques; Sécurité des systèmes -- Périodiques; Sécurité du travail -- Périodiques; 620.00452
• DOI: 10.1016/j.ress.2022.108423
• Languages: English
• ISSNs: 0951-8320

Abstract

Abstract: Multistate networks composed of many independent components are susceptible to natural disasters and intentional attacks. Resilience is described as the capability of keeping and enhancing the required functionality under disruptions, which is of great importance to design and analyze multistate networks. While the resilience engineering community makes great contributions to resilience assessment, approaches for measuring network performance are limited and the practical demand of users is neglected in multistate networks. This paper proposes a demand-based framework for resilience assessment under disruptions in order to address these issues. After specifying network structure with the help of reliability block diagram, component behavior is modeled as a Markov reward process. Both the probability of staying in a particular state and the cost of recovery actions can be calculated. Furthermore, network performance with multiple levels is quantified using universal generating function. Finally, network resilience is comprehensively evaluated from the perspectives of absorption, adaptation and recovery. Besides assessing overall resilience with the consideration of network demand, the proposed framework conducts quantitative analysis of multistate networks. A case study of the potable water network is used to demonstrate the applicability of the proposed framework. Highlights: A Markov reward process model is developed for multistate components. Network performance with multiple levels is quantified under disruptions. A new metric is put forward to measure network resilience comprehensively.