Addressing ground motion simulation validation for earthquake engineering applicationsC. Galasso and F. Zareian Design of new structures and assessment of existing ones require ground motion signals (e.g., accelerograms) as input of engineering structural analysis. Real (e.g., recorded) accelerograms, coming from real events, are the best representation of seismic loading for earthquake engineering applications. However, the inherent scarcity or total absence of suitable real accelerograms for some specific scenarios (e.g., large magnitude events on nearby faults) makes utilization of alternative options unavoidable. Simulated (or synthetic) accelerograms, generated by seismologists by modeling the seismological source and accounting for the wave propagation from source to site, are an attractive alternative with respect to real accelerograms as input to seismic analyses of both existing and new structures. There are many simulation methods available but, to date, there are concerns among engineers regarding the fact that simulated records may not be equivalent to real records (considered as a benchmark by many) in estimating the seismic demand and then the induced damage potential on structures. This study addresses the issue of engineering validation, on the basis of a large statistical analysis, of ground motion simulation in terms of elastic and post-elastic structural response to four historical Californian earthquakes; i.e., 1979 M 6.5 Imperial Valley earthquake, 1989 M 6.8 Loma Prieta earthquake, 1992 M 7.2 Landers earthquake and 1994 M 6.7 Northridge earthquake. For each earthquake, a hybrid broadband ground motions simulation methodology was used, which combines a deterministic and rigorous approach at low frequencies with a semistochastic approach at high frequencies (> 1Hz). The results of this study are directly relevant to the engineering community although they may also provide feedback for seismologists who generate simulated accelerograms for engineering applications. |
