American Society of Civil Engineers


A Definition Undone: Explicit Estimation of PMLs in the Age of Reliance on Design Ground Motion Records


by Mahmoud M. Hachem, (Senior Structural Engineer, Skidmore, Owings & Merrill LLP, 1 Front Street, San Fancisco, CA 94111 E-mail: mahmoud.hachem@som.com), Terrence F. Paret, (Senior Principal, Wiss, Janney, Elstner Associates, Inc., 2200 Powell Street, Suite 925, Emeryville, CA 94608; E-mail: tparet@wje.com), and Gary R. Searer, (Associate, Principal, Wiss, Janney, Elstner Associates, Inc., 2550 N. Hollywood Way, Suite 502, Burbank, CA 91505; E-mail: gsearer@wje.com)
Section: Suggested Improvements to Guidelines, Standards, and Analysis Procedures 2, pp. 584-592, (doi:  http://dx.doi.org/10.1061/41084(364)54)

     Access full text
     Purchase Subscription
     Permissions for Reuse  

Document type: Conference Proceeding Paper
Part of: Improving the Seismic Performance of Existing Buildings and Other Structures
Abstract: The term PML, or Probable Maximum Loss, has been employed for many years in the insurance, lending and engineering industries as a means to define the monetary risk — expressed as a percentage of replacement value — associated with a particular property experiencing an earthquake. PML studies have mostly been used to facilitate real estate transactions. Over time, a number of definitions of the PML have appeared. Most often, PML estimates are generated by engineering firms utilizing shorthand "black box" methods implemented by commercially available computer software. Generic characteristics of the property such as structural type, number of stories, age and more are input and a PML is output. Notwithstanding the authors’ suspicions about the validity of such shorthand methods — particularly in light of the technical definition of the PML, more rigorous engineering and cost estimating methods can also be used to explicitly estimate building response and damage distribution, develop repair scopes and methodologies, and derive costs. However, such explicit methods — while capable of generating substantially more rational and defensible estimates than shorthand black box software — also fall short because they are incapable of yielding answers consistent with the technical definition of PML. The discussion herein explores the vagueness of the definition of PML and the obvious inconsistencies between the explicit statistical significance embedded in the definition of PML and what the engineering methods traditionally used to conduct seismic response assessment can deliver, with foci on the ground motion side, on the capacity side of engineered materials and systems, and on the fragility of nonstructural systems. Many of these inconsistencies are applicable to both "shorthand" and explicit PML estimation methods. Included, for example, are discussions of selection and scaling of ground motion records — a first step in any explicit assessment — as this pertains to definition of a design event upon which the PML estimate is based.


ASCE Subject Headings:
Ground motion
Risk management