SEISMIC FRAGILITY CURVES WITH UNCONVENTIONAL INTENSITY MEASURES FROM PHYSICS-BASED SIMULATIONS

This paper investigates the applicability of 3D physics-based simulations (PBS) of ground shaking scenarios for seismic fragility studies. Specifically, the aim is to validate their use for empirical fragility functions estimation employing unconventional intensity measures (IMs). This approach takes advantage of the PBS ability to obtain peak, spectral, and integral IMs in a wide range of locations, compared to the few physical stations that record time history ground motion during an earthquake. Given the existence of a detailed, disaggregated damage database that includes building typologies representative of the Italian context, as well as a validated PBS ground shaking scenario, the case study chosen for this analysis is the 2009 L’Aquila earthquake. Empirical fragility curves were estimated avoiding aggregation of data for effectively reducing possible uncertainty in resulting estimates. For validation, the resulting empirical fragility functions were compared with those obtained in a previous work using ShakeMap for the ground motion characterization. Finally, two different approaches were employed to compare the fragility curves for the different unconventional IMs and to find the optimal one for fragility function derivation. The results obtained showed that for reinforced concrete (RC) buildings spectral acceleration IMs, consistently demonstrated superior performance in both procedures. For masonry buildings, peak ground displacement (PGD) emerged as the best performing IM, with peak ground velocity (PGV) and cumulative absolute velocity (CAV) ranking as less effective IMs. Independently of the building material, peak ground acceleration (PGA) proved to be a reliable IM in both approaches.