Probabilistic implementation of a mechanics-based procedure for seismic risk assessment of classes of RC buildings

Student: Rosamaria Iaccino
Supervisor: Dr. G. Magenes

Abstract

The present work has the final objectives of implementing in probabilistic terms a mechanics-based procedure for seismic risk assessment of existing R.C. buildings, which were designed and built before or after the application of modern seismic codes, and of comparing different seismic vulnerability/risk methodologies. A very unfavourable situation is to find, in seismic zone, structures designed without following seismic criteria. It occurred because the territory, at the time of design, was not classified as seismic zone. Another reason of having not ductile structures, characterized by a collapse mechanism of soft storey, is the fact that old seismic codes are inadequate to seismicity of the territory where these buildings have been built.

In this work, a simplified deformation-based method for seismic vulnerability assessment is applied to Imperia Province, in Western Liguria, and the results are given for each census area.
This seismic evaluation procedure was originally proposed by Glaister and Pinho [2002] and was successively modified by Crowley, Pinho and Bommer [2004]; it outlines a deterministic approach for the estimation of loss due to a given earthquake within a particular building class over a particular region.

To take into account the values' variability of some parameters, as geometric dimensions of columns and beams or concrete and steel strains, the method is developed in probabilistic terms. In fact, characterizing these input parameters (including building height) as random variables, defined by a log-normal statistical distribution, the equations of the procedure are used to describe a correlated Joint Probability Density Function (JPDF) of displacement d and period T. The JPDF takes into account the fact that d and T are not two independent variables and it is integrated over the failure domain, in order to compute the probability of failure. In particular, this is calculated by estimating the Cumulative Density Function (CDF) of ? and T using first order reliability methods FORM. Capacity is then compared with demand, considering the displacement response spectrum given by application of Sabetta-Pugliese attenuation relationship.

The results are finally compared with those ones of a previous work, in which the displacement-based method proposed by Calvi [1999] was applied.