Page 296 - Numerical Analysis and Modelling in Geomechanics

ENRICO PRIOLO 277

Figure 9.9 Simplified model of a massive structure. The picture shows the model
geometry and the locations of source and receivers (triangles). The structure height H is
held fixed at 10 m in all experiments (from Chiaruttini et al. (1996)).
The size of the rectangular elements in the half-space is equal to 180 m ×180 m,
and corresponds to G=5.2 nodes per minimum wavelength.
The simulations show that some models indeed resonate, storing part of the
incoming energy. With realistic parameters, the lowest resonance frequency is
due to pure shear deformation and is controlled by the shear velocity V s and
height H of the load (f=V /4H), rather than by the inertial properties. Flexural
s
modes are excited only at higher frequencies. The acceleration predicted at the
top of the structure may be 5 to 7 times higher than at the base, depending on the
mass of the structure (Figures 9.11 and 9.12). The gradual release of trapped
energy produces a ground-roll which lasts several seconds after the wave front
has passed. The ground-roll amplitude depends on the structure’s mass and can be
as large as 30% of the peak acceleration. Outside the resonance conditions, the
ground motion is almost unaffected by the presence of the artifact, and the
horizontal motion on top of it is nearly twice the motion at ground level. The
shape of the embankment has only a marginal influence. The perturbation
obviously reaches its highest values close to the structure, but it may still be
relevant at several hundred metres distance (Figure 9.12), especially for the largest
structures. For instance, for an embankment with L=200 m and H=10m, the
level of ringing remains as high as 25%, up to distances of at least 600 m.
Similar results should be expected when the incident field is an upcoming shear
wave. Finally, the presence of an elastic attenuation in the embankment does not
significantly alter the preceding conclusions, unless it has very high values (e.g.
Q<15).
The modelling results indicate that the soil-structure interaction may
substantially alter the free-field ground motion. From a practical point of view,
the main conclusions of this study are: (1) careful analysis is necessary when
interpreting seismic records collected in the vicinity of large artifacts; (2) seismic
hazards at a site may depend on the presence of man-made structures such as
embankments, dams, tall and massive buildings. Finally, this study can easily be
extended to simulate the presence of multiple structures.

Conclusions
In this paper, the 2-D Chebyshev spectral element method (SPEM) to help
solving engineering seismology problems has been reviewed. Its effectiveness in