GNGTS 2022 - Atti del 40° Convegno Nazionale

324 GNGTS 2022 Sessione 2.2 which the median of the SA values be as close as possible to the local design spectrum, the modification of the target spectrum for the simulated accelerogram leads to a compensation for the discordance affecting the subset of the real accelerograms. In the implementation of the proposed procedure some alternative options could be followed. The set of periods for which the misfit is calculated could be spaced by a linearly or logarithmically constant interval. Another possibility of alternative choices concerns the selection of accelerograms that can be representative of scenarios most contributing to hazard for each of the 3 relevant period intervals. It could be desirable that accelerograms representative of all these scenarios be included in the final set of 7 accelerograms. Therefore the programme allows to choose, for the real accelerograms, in addition to the accelerogram best fitting the entire design spectrum, other 3 accelerograms providing the best fitting in each of the mentioned period range. This criterion is then used also for the selection of the 3 simulated accelerograms. Tests of applications of the procedure according to these alternative options are underway for sites located in the different geological domains of the Apulia region, in order to assess whether the different options may be preferable in the presence of different characteristics of the basic seismic hazard. Fig. 2 - Spectral acceleration response (coloured curve) of the real (a) and simulated (b) accelerograms selected, through the proposed procedure, for the site response simulations at Altamura and comparison with the local design spectrum on soils of category A for the Ultimate State limit (black line).