A database of damage-consistent (intensity-based) natural and synthetic accelerograms for seismic risk assessment

The actual practice of seismic-risk assessment for critical infrastructures is based on Uniform Hazard Spectra (UHS) derived from Probabilistic Seismic Hazard Assessment (PSHA) in the format of (pseudo) spectral accelerations. The spectral (force-based) method is still widely applied in earthquake engineering, especially in the design of systems. It is also applied in fragility analysis. The UHS and subsequently the time-history derived based on it has the problem that in reality it does not represent a uniform hazard but the weighted contribution of earthquakes leading to very different intensities at the plant site. In this work the outline of a methodology proposed to develop a database of damage-consistent accelerograms is presented. Macroseismic intensity represents a measure of the strength of an earthquake record inferred from observed damage. According to the European Macroseismic Scale (EMS), different earthquakes characterized by the same macroseismic intensity should lead to the same mean observed damage on buildings with homogenous characteristics (vulnerability classes). At the same time, macroseismic intensity allows to consider the variability on the ground motion parameters associated with the same level of damage. Therefore, to perform non-linear time-history analyses with intensity-consistent sets of accelerograms, means to subject the structures to the same damaging potential and to catch the variability on ground motion parameters. In order to do this, a database of intensity (EMS) consistent natural accelerograms is developed starting from existing catalogues of records and observed intensity. In order to extend the availability of intensity-consistent accelerograms, a methodology is then proposed to assign the macroseismic intensity to physics-based simulated accelerograms. This methodology is based on correlations between intensity and ground motion parameters and tested using non-linear SDOF systems representative of the non-linear behavior of different buildings vulnerability classes.