GNGTS 2022 - Atti del 40° Convegno Nazionale

332 GNGTS 2022 Sessione 2.2 damage distributions for masonry, RC and all buildings within the detailed study area mapped in panel (a) of the same figure. Results show that both approaches, PBS and ShakeMap, generally well reproduce the observed seismic damage in the detailed study area. Predictions are aligned with observations for all levels of damage from DS2 to DS5, indicating the general fitness of both procedures to reproduce observed seismic damage. Fig. 3 - (a) Expected exceedance probabilities of preselected damage levels in the detailed study area. (b) Comparison between observed and predicted global damage distributions for masonry, RC and all buildings in the same area. Conclusions. This work validates the use of ground shaking scenarios obtained by 3D PBS as a basis to constrain the intensity measures for empirical fragility analyses from past earthquakes, with application to the 2009 L’Aquila earthquake. Fragility curves are calibrated for several masonry and RC building typologies representative of the Italian building stock by characterizing the ground motion intensity at the buildings located within the L’Aquila municipality, using the broadband shaking scenarios from the PBS. The comparison of the fragility curves derived from PBS with those obtained from the ShakeMap available for the L’Aquila earthquakehighlights that thePBS-based fragility analysis does not showany systematic bias with respect to standard approaches. The validation of the PBS approach is confirmed also by the comparison of damage predictions with the post-earthquake observations. This study points out the advantages of simulation-based seismic shaking scenarios for empirical fragility studies, particularly when strong-motion recordings are insufficient or not available, as for historical earthquakes.