Page 243 - Numerical Analysis and Modelling in Geomechanics

224 SEISMIC MICROZONING USING NUMERICAL MODELLING
scale, to all the villages and gave to each local situation a value of the
amplification factor on the basis of the table defined at the end of the first phase.
It is worth noting that the feeling of mutual support in an emergency was the
main reason for the positive results of the process.
Considerable work was done on site effect analysis: some researchers used
experimental approaches, others developed numerical codes for the evaluation of
the amplification. In the field of experimental approaches both strong events, and
microtremors and small events were used and numerical methods are available
for 1D, 2D and 3D analysis. In this particular case, experimental approaches
were discarded for reasons of both time and cost, as explained above. The choice
of 1D and 2D techniques derives from the quantity and quality of input data
available and the need for fast computation in order to fulfil the time requirement
previously mentioned.
The Umbria Region is now applying a similar procedure in the areas that have
not been affected by the earthquake, as a support for urban planning in the frame
of prevention policies.

The seismic input

The area affected by the earthquake sequence of September 1997 is located in a
region with several dissected seismic structures, still not very well known and
identified as defined seismic sources. Therefore, it being impossible to separate
the seismic hazard contribution coming from all the possible sources to each
village, the cumulative contribution, on a probabilistic basis, was derived from
all relevant neighbouring seismogenetic areas. The probabilistic approach also
fits with the aim of the project, that is, the evaluation of a set of parameters to be
entered in code for building restoration and new building design.
Calculations were made treating seismic source zones as areas of uniformly
distributed seismicity following a Poisson process. According to Cornell’s
(1968) methodology the expected ground motion values, for a 10% probability
of exceedence within a reference time period of 50 years, were computed.
The hazard analysis results are summarized through uniform hazard response
spectra and an equivalent set of time-histories suitable for microzoning study.

Hazard analysis
The seismic hazard calculation is an application of the total-probability theorem
(Harr, 1987) and it is usually presented in the following form:

(8.1)

where P is the probability that a ground-motion amplitude A is greater than a in
time t; f denotes the probability density function of the magnitude distribution;
M