Page 294 - Numerical Analysis and Modelling in Geomechanics

ENRICO PRIOLO 275
realistic. This supports the need for methods that accurately model realistic
geologic structures.
In addition to the earthquake data, environment microtremors were recorded at
the same site of the Catania station (Priolo et al., 2001). The aim of this data
acquisition was to improve our prediction of the seismic ground motion locally.
To this end, Nakamura’s approach was followed which, as proven, provides the
main features of the dynamic ground response through the calculation of the
spectral ratio between the horizontal and vertical components (i.e., H/V ratio) of
background microtremors (Nakamura, 1989). In Figure 9.8, the H/V spectral
ratios obtained from (i) the environment seismic noise, (ii) the seismograms
recorded by the accelerometric station during the December 13, 1990 M=5.8
Eastern Sicily earthquake, and (iii) the seismograms computed by the 2-D
spectral element method for the same earthquake, are compared. The ratios
obtained from the seismic noise measurements well detect the fundamental mode
of vibration at about 1.5 Hz, but they miss the peak at the higher frequency of 4–
5 Hz. This fact may either confirm that Nakamura’s method is only reliable for
identifying the fundamental mode of vibration, or that the peak at 4–5 Hz in the
earthquake records is a feature of the earthquake source. The H/V ratios
determined from the synthetic seismograms are generally noisier, but they well
reproduce the behaviour of the spectral ratios determined from the full ENEA-
ENEL recordings for frequencies larger than 1 Hz. The origin of the two
additional peaks at 0.4–0.8 Hz is currently under investigation.

Influence of a massive structure on the free surface ground
motion
In this study, the 2-D SPEM was used to investigate how seismic ground motion
is affected by the presence of a massive structure. The analysis of soil-structure
interaction in seismic hazard studies is usually concerned with its influence on
the response of and damage to the structure itself. One should expect, however,
that large massive loads perturb the free-field ground motion in their
surroundings. Recent numerical investigations indicate, in fact, that long-range
soil-structure interaction for large buildings may have had a role in determining
the abnormal amount of damage on the Mexico City clay basin during the 1985
Michoacan earthquake (Wirgin and Bard, 1996). The aim of this work was to
investigate how much, and over what distance range, surface loads alter the free-
field ground motion.
Two-dimensional models were considered where a linear elastic structure with
a quadrangular cross-section, resting on an elastic homogeneous half-space, is
impinged upon by the surface waves generated by a vertical and impulsive point
source located on the ground surface at some distance (Figure 9.9). The
scattering of Rayleigh waves and the response of the structure are extensively
analysed in a parametric way: by varying the size, mechanical parameters, and