Civil Engineering Association

Seismic risk computation for the base-isolated reactor building of the IRIS NPP M.

This report deals with the fragility assessment of the IRIS reactor building in its base-isolated version, following the previous configuration without any isolation system [1]. The aim of this step consists in the evaluation of the effectiveness of the base-isolation when applied to the IRIS NPP for the reduction of the seismic risk and in the comparison between the performance of the traditional and the isolated reactor building. Since the behavior of the isolators is markedly
hysteretic, the hypothesis of linearity of the building response, typical of the traditional building, has been removed herein, and a suitable force-displacement literature model is adopted to
represent the isolators inelastic response to horizontal loading.
The probabilistic assessment is based on the procedure described in [2], on a nonlinear analytical
model, by performing sequential seismic analyses in the MATLAB framework [3], with the application of an explicit direct integration method. Previous studies [4] on an extensive finite element model of the isolated IRIS reactor building allow to introduce the rigid body condition for the structural equation of motion. The most important requirement of the procedure remains to reduce as much as possible the
uncertainties related to the incomplete knowledge and accuracy in defining models and methods; this reduction is here sought by refining analysis procedures and using consolidated analytical and numerical tools. The characterization of the isolator devices has been preliminarily performed by testing scaled prototypes, in view of further full scale laboratory experimentations; the definition of the limit state domain for the reference device in terms of the total vertical and horizontal forces has been also evaluated.
The definition of the random variables and the generation of the seismic excitations are also key
aspects of the analysis. They represent significant requisites for setting up the probabilistic
assessment of the response of the structural system.
The exceedance probability of the control system limit state domain is here computed via Monte Carlo simulations; to reduce the computational effort, the response surface (RS) method is used to express the seismic response as a function of the variation of the adopted random parameters. The generation of the RSs is performed in terms of mean and standard deviation of the minimum distance from a specific limit domain. In such setting, the RS evaluation must be repeated for every value of peak ground acceleration; on the other hand, to evaluate the isolators’ behavior, the seismic behavior of the isolated building can be captured by means of a very simple mechanical model which can be based on the hypothesis of rigid-body motion of the building. Finally, the results of the fragility analysis are computed, also in view of a refinement of the response surfaces, within a complete risk assessment for a prototype site.