Estimation of the dynamic seismic response of buildings through transfer functions: a probabilistic approach

Abstract

The significant number of seismic ground motion records to be considered when designing or assessing civil structures is a common restriction. This is because the large computational time involved in the calculation of the nonlinear dynamic response of complex multi-degree-of-freedom system. There are several strategies to overcome this limitation; however, the reliability estimation of the analyzed systems can be compromised. This research is focused on developing a simplified methodology to obtain a probabilistic and reliable estimate of the seismic response of buildings. To do so, a set of transfer functions associated to the dynamic response of a building model are obtained. Then, an optimal transfer function is identified as the one maximizing the prediction of engineering demand parameters (EDPs) when the structure is subjected to a large set of ground motions records. Results show that using a reduced number of records allows developing an enhanced strategy to obtain reliable results in terms of the main statistical moments of the EDPs. This increased capacity to analyze complex systems in an affordable time has important consequences in the identification of optimal designs in terms of material-performance relationship, as well as in the estimation of the expected risk.