Estimation d'accelerogrammes spatialement variables sur profils de sols stochastiques : application au seisme d'el asnam de 1980

Abstract

Spatial variability of ground motions has significant influence on dynamic response of extended structures such as pipelines and bridges. When carrying out seismic design of structures, reasonable definition of earthquake ground motions is crucial to structural response analysis. Although a large number of earthquakes have been recorded in the past decades, the site conditions where these earthquakes occured may be much different from those of the structures. In this study, the widely used finite-source empirical Green’s function (EGF) method is extended to synthesize seismic motions across an array of stations located at bedrock in the epicentral region of the 1980 El-Asnam earthquake (North-west Algeria). A parametric model of coherency is presented for base rock based on the simulation results. A Parametric study investigating the sensitivity of coherency to seismic source, propagation path and site conditions, is conducted by using the extended hybrid Green’s function method. The corresponding spatial coherency functions of surface seismic motions at the ElAsnam Sogedia site are calculated using the analytic site specific model developed in the Zerva and Harada approach for a layered, random soil site. A stochastic generation of time histories at various locations on the ground surface of the site, compatible with our estimations of the target power spectral densities and the spatial coherency functions, is presented. The obtained results show that finite-source effects can cause significant loss of coherency at bedrock in the near-field. The layer stochasticity can significantly reduce, or greatly reduce for medium to large separation distances, the coherency near the sites resonant frequency, even for firm soil conditions when its damping ratio is low. For these soil conditions, the layer stochasticity controls the seismic ground strains.