EVALUATING LIFELINE RESPONSE TO EARTHQUAKES: A SIMULATION METHODOLOGY.

DONALD EDWARD NEWSOM, Purdue University

Abstract

Modern society is increasingly dependent on lifelines, land-based network type systems providing utility services, transportation, etc. The functioning of these systems in a post-earthquake environment is critical. However, little research has been done to evaluate the system response of lifelines to earthquakes. Several factors make this evaluation difficult. The system response depends on the physical distribution of lifeline facilities in relation to the local geology and the specific earthquake. The inherent redundancy of lifelines, both between and within nodes of the network, means that the actual configuration of the system must be considered.In the simulation methodology developed here, the response of a lifeline to a series of specific hypothesized earthquakes is modeled. The electric power transmission system in a major metropolitan area is used as a case study. The physical and electrical configurations of the system. are modeled in detail. Post-earthquake reconfiguration of the system in light of specific damages indicates the extent of disruption. A discrete event simulation of post-earthquake recovery indicates the duration of disruption.The simulation begins with the choice of a particular hypothesized earthquake. The severity of ground motion is found at each site where lifeline facilities are located. At each site the dynamic characteristics of equipment and any supporting structure are used together with the local seismic environment to find the severity of shaking applied to each piece of equipment. The local seismic environment is also used to estimate the reduction in power demand due to damage to customers. Frogility curves are used to find probabilities of failure for each piece of equipment in each of several failure modes. In each replication of the simulation a sot of specific equipment failures is chosen consistent with the probabilities of failure. The surviving parts of the system aro reconfigured, using the inherent system redundancy, to deliver power. Post-earthquake recovery operations are simulated day-by-day until full recovery is achieved under emergency operating conditions. Results of the simulation include identification of which equipment had to bo repaired, man-hours and other resources required, and customer service statistics for each day.This research draws upon the thesis of Peter J. Feil, "A Methodology for the Reconfiguration of a Severely Damaged Electrical Power System." His work included the development of structures for entry and retrieval of data describing the system, the reconfiguration methodology, and o methodology for identifying and evaluating the repairs necessary to restore the system.Several experiments assessed the basic seismic response of the system and the effects of random variations in equipment damoges, repair times, and travel times. Sensitivity analyses were performed, examining variations in post-earthquake demand reduction, spare transformer availability, loss of generating capacity, increased seismic resistance of certain equipment, and priorities for assigning repair crows.Recovery from a magnitude 8.3 earthquake, the largest considered, typically took 6 days. Up to 8% of the customers in the hardest-hit area had no power until then. Many others had only partial supply of their demand. Recovery from a magnitude 7.5 earthquake typically took 3 days. Several critical sites and types of equipment were identified as consistently having highly disruptive failures. Hypothetically improving the seismic resistance of some of the critical equipment vastly improved system response. Data deficiencies were encountered in the areas of soll profiles, equipment fragility, and post-earthquake demand reduction. Sensitivity analyses showed that these data are often important.

Degree

Ph.D.

Subject Area

Engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS