Assessment of seismic collapse uncertainties of steel moment-resisting frames by means of single-degree-of-freedom systems

Abstract

The collapse evaluations presented to date are of various approximations, such as significant degrees of uncertainty in ground motion characterization and structural collapse modeling. Accordingly, determining the variance of collapse capacity has an important role in evaluation of the seismic risk of structures. However, its process includes time-consuming steps, for example, incremental dynamic analysis and nonlinear dynamic collapse analyses of the archetype models. In order to save time and preserve accuracy, this study attempts to assess the seismic collapse uncertainty of single-degree-of-freedom systems for evaluating the same uncertainties in steel moment frames. A combination of extended incremental dynamic analysis and Latin hypercube sampling method was utilized to study the epistemic and aleatory effects on multi-story steel moment frames (3-, 5-, 8- and 15-story frames) as well as single-degree-of-freedom models. The results indicated that due to aleatory and epistemic uncertainties, the variance of collapse capacity in three-, five- and eight-story frames was close to the results of single-degree-of-freedom systems in which P-Δ effects were assumed as small, whereas the variance of collapse capacity of a 15-story frame (T₁>2 s) was compatible with the results of single-degree-of-freedom systems when P-Δ effects were considered large.