Loading...

Probabilistic Modeling of Water Distribution Infrastructure and Its Inter- Dependencies in Community Resilience Analysis

Khayambashi, Kamiar | 2021

464 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54155 (09)
  4. University: Sharif University of Technology
  5. Department: Civil Engineering
  6. Advisor(s): Mahsuli, Mojtaba; Safaie, Ammar
  7. Abstract:
  8. This research develops a probabilistic framework for evaluating community resilience subject to earthquake events by modeling the water distribution system and its prevailing inter- and intra-dependencies with other infrastructure systems of the community. This framework is entitled “Rtx,” which is under continuous development at the Center for Infrastructure Sustainability and Resilience Research (INSURER) at Sharif University of Technology, Tehran, Iran. It quantifies the community resilience by integrating hazard models, risk models, and recovery models within an agent-based simulation. Comprehensive modeling of the water distribution system requires a library of probabilistic models, which is developed herein. These models predict the damages incurred by the water network components and functionalities thereof, considering the inter- and intra-dependencies, time-variant water demand, water network serviceability, components recovery, and economic and social consequences. In the aforementioned framework, simulation starts with hazard models, which predict the occurrence, magnitude, rupture location, and consequently, ground shaking and ground failure intensity at the location of infrastructure components. Afterward, risk models predict the damages incurred by water infrastructure components, comprising water tanks, pumping stations, treatment plants, wells, tunnels, pipelines, and also, components of other infrastructure systems due to both ground shaking and ground failure hazards. Among these models, novel probabilistic models are presented for predicting the repair rate of the pipeline using the Bayesian linear regression method. Additionally, a new probabilistic model for predicting the number and combination of pipeline failures is proposed. Upon predicting the damage state of the infrastructure components, a group of models evaluates the serviceability of the water network using the network capacity and the time-variant demand. To this aim, given the damage state of water network components and the serviceability state of the power network, the functionality status of water network components is predicted in order to evaluate the water network serviceability capacity. Moreover, the time-variant demand on the water network is modeled using the damage and habitability of buildings as well as the relocation of the population due to people leaving their houses for shelters. Then, given the post-earthquake modified capacity and modified demand, the water network model evaluates the water accessibility using a network connectivity method. Subsequently, the proposed framework simulates the required operations for the post-earthquake recovery of the community using the recovery models and agent-based simulation. In order to model the recovery process of water distribution infrastructure more accurately, a semi-structured interview is conducted with field experts. According to this interview, In the recovery simulation, the recovery process of water distribution infrastructure starts with damage inspection of components and ends with the completion of their repair. Water network recovery operations comprise the inspection and repair of water stations and the inspection and repair of water lines, including trunk, transmission, and distribution pipes and tunnels. The required time and cost for each of the aforementioned operations are computed by the recovery models. At the end of the simulation, the recovery period and direct and indirect economic and social costs incurred by the community are estimated. In this research, indirect social costs stem from the life quality reduction due to homelessness and the lack of access to water and/or power, which is translated into cost using the notions of the value of statistical life and life expectancy. At last, the probability distribution of the community total loss is estimated using the Monte Carlo sampling method. This distribution is then used as a basis for assessing the community resilience measure. The proposed framework is showcased on a virtual community comprising a population of 125,000, a portfolio of 17,000 residential, commercial, and administrative buildings, and water, power, and treatment infrastructure systems. This framework provides several insights about the testbed. Some of these insights are the contribution of different categories of costs in the total loss, the number of people without water access during the recovery period, and the time required to recover each infrastructure system. Analysis results reveal that the majority of direct costs of water distribution infrastructure recovery are due to water station repair costs. Moreover, the results indicate that the direct and indirect costs induced from the damages occurred to water distribution infrastructure are comparable to direct economic costs of the community. The results also demonstrate the water accessibility of the community during the recovery period after the occurrence of earthquakes with different moment magnitudes. Furthermore, the capabilities of the proposed framework for decision-making on different resilience enhancement policies such as water network distribution pipelines retrofit or the presence of reserve water storage in healthcare facilities have been examined. In the case of the community, after the occurrence of the scenario earthquake in the testbed, the costs incurred by the community are more than the gross domestic product (GDP), and consequently, a catastrophe occurs. The majority of community costs are due to social costs
  9. Keywords:
  10. Community Resilience ; Agent Based Simulation ; Damage ; Interdependency ; Probabilistic Modeling ; Water Distribution Network Resiliance ; Infrastructure ; Probabilistic Seismic Structures Vulnerability ; Seismic Risk Index

 Digital Object List

 Bookmark

No TOC