Seismic demand analysis using environmental contours of vector-valued ground motion intensity measures

Abstract

A study is presented on seismic demand analysis using environmental contours of vector-valued ground motion intensity measures. The environmental contours specify combinations of intensity measures for estimating an engineering demand parameter with a given exceedance probability. Copulas are employed to model the joint distribution of vector-valued ground motion intensity measures for developing the environmental contours. A framework is presented for copula estimation of spectral accelerations using ground motion records from stations on stratified sand and silt deposits in Mexico City. The statistical dependence between pairs of spectral accelerations at a range of periods is analyzed in terms of Kendall’s τ and used for estimating parametric copula models. Goodness of fit of the estimated copulas is examined and the Akaike information criteria is applied for selection of copula models. Environmental contours of spectral accelerations are developed adopting the inverse FORM approach and employed to estimate mean exceedance rates of maximum interstory drift of a building model. Both, the cases of deterministic and stochastic response models of interstory drift are examined. The accuracy of the environmental contour estimates is assessed by comparison with Monte Carlo simulations and estimates from long-term analyses. It is found that the approximation of the failure surface in the inverse FORM approach is quite appropriate for a range of return periods for which very precise or reasonably accurate response estimates were obtained from the environmental contours.