Examining the assumption of homogeneous horizontal layers within seismic site response analysis

Volpini, Carolina and Douglas, John (2017) Examining the assumption of homogeneous horizontal layers within seismic site response analysis. In: 3rd International Conference on Performance-based Design in Earthquake Geotechnical Engineering, 2017-07-16 - 2017-07-19.

[thumbnail of Volpini-Douglas-PBD-III-2017-Homogeneous-horizontal-layers-within-seismic-site-response-analysis]
Preview
Text. Filename: Volpini_Douglas_PBD_III_2017_Homogeneous_horizontal_layers_within_seismic_site_response_analysis.pdf
Accepted Author Manuscript

Download (982kB)| Preview

Abstract

One-dimensional analyses can be conducted to estimate the impact of superficial soil layers on earthquake ground motions. Such analyses are based on the assumption that all boundaries are horizontal and that the response of a soil deposit is predominantly caused by horizontal shear waves propagating vertically from the underlying bedrock. This assumption is made even for sites with a relative large surface area, e.g. the footprint of large infrastructure such as power plants. An important step then is to create a model of the near subsurface that is representative of the overall area under analysis. This means it is essential to evaluate geomechanical characteristics of the soil at certain locations and extend these measurements over the whole site. As a consequence of this, it is assumed that the soil characteristics, which include stratigraphy, geometry and geotechnical properties, are homogeneous. Recent observations, however, have clearly demonstrated that even over a small area (~1km2) ground conditions can vary greatly. The purpose of this study is to examine the impact of relaxing the assumption of infinite horizontal layers by undertaking a parametric study of the variability in amplification across areas with gently dipping subsurface layers. Starting from a 1D approach the influence of dipping layers is evaluated through simplified but geometrically representative models. Randomization of shear-wave velocity profiles using the Toro (1995) method, as implemented in STRATA, is used to compute the variability in site amplification that would be captured by a standard 1D technique. This provides a baseline for comparison with the variability introduced by the dipping layers. Subsequently, two-dimensional simulations are conducted for the same sites with dipping layers to estimate the error made through the assumption of 1D response. The goal of this study is to understand when the 1D assumption can be used in the presence of dipping layers.