Comparison of Incremental Dynamic, Cloud, and Stripe Methods for Computing Probabilistic Seismic Demand Models

Abstract

Probabilistic seismic demand models (PSDMs) are mathematical relations between measures of earthquake intensity and measures of structural response. The models include best estimates as well as considerations of uncertainty. Three methods for obtaining the demand models are compared in this paper: cloud, incremental dynamic, and stripe. A PSDM is derived using each method for a sample reinforced concrete bridge using a single IM (first-mode spectral acceleration) and EDP (column drift ratio). Merits and drawbacks of each method are discussed. For the structure considered, cloud and incremental dynamic analyses yield interchangeable PSDMs for the same amount of computational effort. Stripe analysis should be used only for individual intensity levels, not as a means of generating complete PSDMs.