Abstract:
Ground improvement techniques are widely employed in seismically active regions to prevent
ground failure or mitigate the amount of settlements transmitted to buildings. In many cases,
these solutions substantially modify the subsoil conditions and the surface ground motions
intensities accordingly. This thesis assesses the capability of ground response analysis (GRA)
models to predict site amplification effects and quantify the influence of the degree of ground
improvement and geometry of the improved zone on the seismic demand of buildings.
In the first part of this research, a new frequency-dependent equivalent linear (FDEL) algorithm
is proposed to predict ground motions propagating through soil profiles for total stress GRA.
When simulating strong ground motions that exhibit higher levels of soil deformation, the
frequency content of ground motions processed within this FDEL procedure is substantially
improved as compared to the traditional equivalent linear method, with results more consistent
with those from nonlinear time-domain analyses. The new FDEL method is less sensitive to the
parametrization of nonlinear material curves compared to other methods. However, all GRA
methods tend to under-predict the spectral accelerations at high frequencies when compared to
historical downhole array records, with substantial discrepancies across the model predictions.
In the second part, the effects of ground densification and stiffening on seismic site amplification
factors are investigated through a series of parametric GRAs for total and effective stress
conditions. When a densified crust is present with unimproved soft soil layers underneath, the
ground surface accelerations are generally reduced. However, the levels of de-amplification
decrease as the depth of ground densification increases. It was found that the reduction of
foundation displacements across all improved sites is accompanied with an amplification of the
seismic base shear developed in structures. The increase in the shear-wave velocity of the
densified layers provides a poor correlation when the densified crust is underlain by liquefiable
layers. However, the ratio of improved zone thickness to the bottom depth of soft/liquefiable
layers correlates well with the increase in the seismic demand. Examples of possible correlations
that could be developed to better anticipate the effects of ground densification on the seismic
demand of buildings are presented.