Loss Modeling for Downtime, Deaths and Decision-Making - 1362005

Project Title—ID Number Loss Modeling for Downtime, Deaths and Decision-Making - 1362005
Start/End Dates 10/1/05 – 9/30/06
Funding Source PEER-NSF
Project Leader (boldface) and Other Team Members James L. Beck (Caltech/F), Judith Mitrani-Reiser (Caltech/GS), Keith Porter (Caltech/F), Gee Hecksher (Architectural Resources Group/I)
F=faculty; GS=graduate student; US=undergraduate student; PD=post-doc; I=industrial collaborator; O=other

Project goals and objectives

The main goal of this project is to continue to work towards closing the gaps that exist in the loss estimation module of the PEER performance-based earthquake engineering methodology. We aim to close these gaps by Year 10 in order to complete, document and package the PEER methodology. Specific objectives for Year 9 are:

  1. Joint methodology document. Prepare a joint document of PEER's loss estimation methodology with PI Miranda at Stanford which presents PEER's damage- and loss- analysis methodologies for the repair–cost decision variable.
  2. Indirect loss associated with downtime. In coordination with PI Comerio at UC Berkeley and BIP Hecksher, complete the PEER damage- and loss-analysis methodology for the decision variable of indirect losses arising from downtime, including the probability of building closure (thereby creating a virtual building inspector).
  3. Decision-analysis performance metric. In coordination with PIs May and Ince, work on the decision-analysis framework that uses the "3 Ds" (deaths, dollars, and downtime) as decision variables but also allows decision makers to account for their risk attitude.

Role of this project in supporting PEER's mission (vision)

PEER promotes and integrates the research of diverse scientists and engineers (earth sciences, engineering seismology, engineering, architecture, economics, and public policy) to create a practical methodology capable of estimating building and bridge performance in terms of interest to owners and society, namely, dollars, deaths, and downtime. The present research advances this mission in several ways.

  1. Finalize building damage and loss analyses. The joint methodology document will finalize the EDP-to-DV analysis and resolve any lingering questions of competing approaches within PEER.
  2. Advance downtime modeling. The indirect-loss research will advance, or finalize with consensus, probabilistic modeling of building closure and downtime and how to quantify the latter in financial terms.
  3. Advance decision-making. The decision-analysis framework will allow for the risk attitudes of decision-makers and will inform the decision-making stage of the PEER methodology, including addressing the cost of uncertainty.
  4. Speed methodology dissemination. By collaborating with ATC-58 researchers (e.g. Hecksher, Miranda and Porter), this research will be informed by practitioners' needs, will facilitate technology transfer from PEER to ATC-58 and then to practice.

Methodology employed

Caltech is approaching the issue of building downtime by considering both building closure (red- tagging) and building inoperability, since a building that is green-tagged may not be operational. The Applied Technology Council's Procedures for Post-earthquake Safety Evaluation of Buildings document (ATC-20, 1989 and 1995) offers guidelines for rapid and detailed building evaluations after seismic events. We use the available structural performance data from simulations (EDP's from nonlinear structural analysis and DMs from fragility functions) in accordance with these guidelines to create an event tree, shown in Figure 1 below, whose branches provide the probabilities of a building being green-, yellow- or red-tagged. Based on ATC guidelines, a building is unsafe to inhabit if it receives a red-tag (which may have been initially a yellow-tag). Conditional on the closure of a building, we can establish a probability distribution for building downtime based on mobilization time (time to inspect, acquire building permits, acquire financial means, etc.) and the repair durations of the damaged building assemblies. The repair strategy is owner-dependent, but we will provide an envelope of repair durations based on slow- and fast-track strategies for the repair work.

Caltech model for building closure

A building can receive a green tag but still be inoperable because of severe damage of its nonstructural components and equipment. The Caltech downtime model currently includes all nonstructural components, equipment and occupancy-specific systems for which there are existing fragility functions (i.e., wallboard partitions, raised access flooring, suspended ceilings, glazing, elevators, domestic water, fire response, etc.) and that are considered critical for building operability by ATC (ATC-20 1989, ATC-20 1995, ATC-58 2005) and other researchers (e.g. Porter et al. 1993). As in the case of building closure, conditional on building inoperability, we can establish a probability distribution for building downtime based on mobilization time and the repair durations of the damaged building assemblies.

Brief Description of previous year's achievements, with emphasis on accomplishments during last year (Year 8)

In the previous year, Caltech has completed the damage and loss analysis of the benchmark building. The Caltech, Stanford, and UCLA research teams submitted two papers to 8NCEE and will give a joint oral presentation at the conference. Also, the MATLAB loss estimation toolbox is complete and now available (contact Judy Mitrani-Reiser: judith@caltech.edu). Various PEER researchers and industrial collaborators interested in loss modeling have met several times to discuss the similarities and differences in their approaches for the PEER methodology. The outcome of this meeting was a consensus that the PEER approaches can be adopted as tools for assisting engineers, architects, building owners, and city officials in design-related decisions. Caltech also developed a preliminary model for building closure and building inoperability, and is finalizing it in Year 9.

Other similar work being conducted within and outside PEER and how this project differs

ATC-58 workers this year are advancing from a 25% draft of their analysis methodology. They appear not yet to have begun to address fatalities or downtime. There is no indication that they will address decision-making through formal decision analysis.

Describe any instances where you are aware that your results have been used in industry

  1. The Kajima Corporation of Japan has developed a PBEE-like methodology, partly in collaboration with Caltech.
  2. John A. Martin Associates has developed loss-estimation software for CSMIP that relies in part on fragility functions developed by Caltech researchers.
  3. The California Earthquake Authority is reviewing its seismic retrofit incentive program using products of Caltech's PBEE research.
  4. Commercial Internet users have accessed the Caltech-produced PEER Testbeds webpage more 4,000 times per month as late as November 2005. Two of the most-popular documents downloaded were Porter's 2003 PEER Methodology Overview and the 2002 Caltech documents of the methodology for performing and depicting sensitivity studies.

Expected milestones & deliverables

  • 31 Mar 06: Completion of methodology to establish the probability that certain category of facilities will not be operational.
  • 30 June 06: Completion of methodology to establish probability distribution for downtime, given closure and the recovery strategy.
  • 31 July 06: Completion with Miranda of documentation of PEER's methodology for damage and loss estimation: EDP to DV=repair costs.
  • 31 Aug 06: Report documenting EDP-to-DV=indirect losses from downtime.
  • 30 Sep 06: Report documenting initial decision-analysis framework that uses the "3 Ds" (deaths, dollars, and downtime) as DVs but also allows the decision maker to account for his or her risk attitude.

Member company benefits

Architectural Resources Group (Hecksher's company) will gain early access to and influence over the downtime methodology.

Back to Funded Project Archive main page