Loading...

Probabilistic Modeling of Telecommunication Network Failure and Recovery in Community Resilience Analysis

Naderi, Mohammad | 2021

155 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 55396 (09)
  4. University: Sharif University of Technology
  5. Department: Civil Engineering
  6. Advisor(s): Mahsoli, Mojtaba
  7. Abstract:
  8. This thesis extends a probabilistic framework for evaluating the seismic resilience of communities by probabilistic modeling of the telecommunication infrastructure failure and recovery along with its interdependencies with other infrastructures. This framework, Rtx, that is being extended under continuous development at Sharif University of Technology, is based on modeling the community state from hazard incident until complete recovery by a combination of risk and agent-based simulation models. In this thesis the framework is extended with telecommunication infrastructure models. This infrastructure is modeled by developing a library of probabilistic models to predict damage to network components, functionality and serviceability of infrastructure components by considering infrastructure interdependencies, network analysis, and economic and social consequences. In this framework, simulation begins with hazard models that characterize the magnitude, rupture location, ground failure and ground shaking intensity at the location of all community infrastructure components. In the second stage, risk models predict the damage to vulnerable components of the telecommunication infrastructure, including switching centers, aggregation layer nodes, base stations, microwave towers, and telecom cables due to ground shaking and ground failure, and the damage state of other community infrastructure components. Among these models, a group of novel probabilistic models are developed to assess the damage of telecommunication stations that are located inside or on top of buildings, microwave links functionality and the number of cable ruptures due to ground failure. In the next stage, telecom network model determines the serviceability of telecommunication infrastructure components and community population access level to the network by considering the damage state of telecom components and power network serviceability to telecom stations. Next, recovery process of the community to its normal state is modeled using agent-based simulation. Agents, in this simulation, are responsible for the recovery of all or a part of an infrastructure that each agent decides on the recovery process of its infrastructure independently. During agent-based simulation, recovery models predict time and cost of repair and restoration operations of community infrastructures by considering the effect of telecom network unserviceability. Recovery agents of telecommunication infrastructure include repair agent of telecom stations and cables and portable base station operation agent. At the beginning of this simulation, increase in call demand is predicted according to serviceability status of telecommunication components. In this thesis, three main sources for call request are considered that include calls within the affected area, calls from outside the area, and calls to emergency services. When the simulation is concluded, recovery time and total economic, direct social and indirect social losses to the community are evaluated. Reiterating this process through Monte Carlo sampling results in the probability distribution of total community cost that is then used to quantify a community resilience measure. The framework is showcased by a comprehensive application to a virtual community comprising a population of 125,000, 17,000 buildings, telecommunication infrastructure, power network and healthcare system. This analysis represents various insights regarding the community resilience such as the contribution of each loss category to the total community loss, total number of people without access to telecommunication services during the recovery phase, and the recovery time of each infrastructure system. Disaggregation of cost into various categories reveals that social costs have the largest share, followed by indirect economic costs. Moreover, telecom station repair costs account for the largest share of direct economic costs associated with telecommunication infrastructure recovery. The results show that due to rapid recovery strategies in mobile phone network, population access to this network recovers much faster than the landline network. Additionally, the results indicate that the main cause of call request is network congestion, followed by telecom component unserviceability. The proposed framework can be used as a decision support tool for community resilience enhancement policies. For instance, in the case study presented herein, the proposed framework is utilized to assess installation of backup generators in base stations, deploying portable base stations, and base station retrofit actions.
  9. Keywords:
  10. Community Resilience ; Infrastructure ; Telecommunication Network ; Probabilistic Modeling ; Agent Based Simulation ; Probabilistic Seismic Structures Vulnerability ; Call Demand

 Digital Object List

 Bookmark

No TOC