2D/1D AGGRAVATION FACTORS: FROM A COMPREHENSIVE STUDY TO ESTIMATION WITH A NEURAL NETWORK MODEL - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2018

2D/1D AGGRAVATION FACTORS: FROM A COMPREHENSIVE STUDY TO ESTIMATION WITH A NEURAL NETWORK MODEL

Résumé

The aim of this paper is to derive "aggravation factors " (AGFs) quantifying the difference between 2D site response and the corresponding 1D estimate for numerous quantities of engineering interest such as response spectra and ten scalar ground motion intensity parameters (GMIPs). The raw results are a huge collection of AGFs derived by numerical simulation for a total 154692 receivers located at the surface of 894 valleys exhibiting a trapezoidal (131) or triangular (18) shape, with six different velocity contrasts. A simple statistical analysis first allows to identify the respective order of magnitude of these AGFs for each GMIP. In a second step, a neural network approach is used to provide a set of analytical relationships quantifying the coupled effects of a limited number of geometrical and mechanical parameters for all the considered GMIPs, and different positions within the valley. The AGFs are found to be component dependent, with larger values for out-of-plane component (SH waves) compared to the in-plane one (SV waves). They are GMIPs dependent, being the largest for Arias Intensity (I A) and peak spectral amplification factor (SAF), intermediate for the Cumulative Absolute Velocity (CAV), and the smallest for all the other indicators. The geometry has a significant control on the AGF. For embanked valleys, the highest AGF occur in the center because of constructive interferences, steep edge slopes have significant but very localized effects on the edges, while gentle edge slopes have significant, long distance effects because of their energetic diffraction power. Diffraction away from the lateral slopes also implies a 10-20% variability of the motion on the rocky edges. Finally, results show also that mechanical parameters within the valley do affect the AGF.
Fichier principal
Vignette du fichier
16ECEE_Stambouli_11624.pdf (2.17 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01826955 , version 1 (30-06-2018)

Identifiants

  • HAL Id : hal-01826955 , version 1

Citer

Ahmed Boudghene Stambouli, Pierre-Yves Bard, Emmanuel Chaljub, Peter Moczo, Jozef Kristek, et al.. 2D/1D AGGRAVATION FACTORS: FROM A COMPREHENSIVE STUDY TO ESTIMATION WITH A NEURAL NETWORK MODEL. 16Ith European Conference of Earthquake Engineering, EAEE + AUTH, Jun 2018, Thessaloniki, Greece. ⟨hal-01826955⟩
288 Consultations
131 Téléchargements

Partager

Gmail Facebook X LinkedIn More