Seismic risk is usually estimated under the hypothesis of stationarity. It is known that in some cases this is not sufficiently accurate. One example is the aftershock seismic risk evaluation, used to classify buildings, in terms of their usability. The proposed approach considers the occurrence of events as a Poisson process with variable rate. The number of events in the unit time is given by the Omori modified law and the magnitude distribution by the Gutenberg-Richter relationship. Since in the early days after the event, field data are not accurate enough to evaluate the model parameters, a Bayesian approach is used. The prior data, coming from the recent Italian sequences, and the posterior data, coming from the field, are merged, considering their uncertainties. An oriented NS fault (normal, inverse or strike slip) and a main shock of given magnitude is considered. Elastic spectral ordinates are attenuated from source to site, considering a given soil type. Buildings are considered to be uniformly distributed at every distance from the fault. Each building has similar seismic response, represented by a bilinear relationship. The performance point is obtained through the N2 method. The response of the damaged structure to aftershocks is still assumed bilinear, passing thorough the origin and the performance point in the main shock. For a given main shock, soil type and building type, uncertainties on the attenuation relationships, on the aftershock magnitude and epicentre and on the Life Safety limit state are considered. The probability of unusability in the main shock and in the aftershocks at each location from the epicenter, given the main shock, soil type and building type, has been evaluated. Results are sensitive to the exposure interval T1-T2, where T1 is the starting and T2 the ending time, measured from the main shock. The results are then compared against the minimum requirement concept, reported in ATC20 and in the Italian usability methodology. In the paper the full methodology is described and several important items are discussed, such as the effects of the spatial localization of strong aftershocks, the aftershock activity decay and the building resilience.
Goretti, A., Giannini, R. (2010). A BAYESIAN MODEL FOR THE SHORT-TERM USE OF BUILDINGS AFFECTED BY AFTERSHOCKS. In Proceeding of 14 European Conference on Earthquake Engineering.
A BAYESIAN MODEL FOR THE SHORT-TERM USE OF BUILDINGS AFFECTED BY AFTERSHOCKS
GIANNINI, Renato
2010-01-01
Abstract
Seismic risk is usually estimated under the hypothesis of stationarity. It is known that in some cases this is not sufficiently accurate. One example is the aftershock seismic risk evaluation, used to classify buildings, in terms of their usability. The proposed approach considers the occurrence of events as a Poisson process with variable rate. The number of events in the unit time is given by the Omori modified law and the magnitude distribution by the Gutenberg-Richter relationship. Since in the early days after the event, field data are not accurate enough to evaluate the model parameters, a Bayesian approach is used. The prior data, coming from the recent Italian sequences, and the posterior data, coming from the field, are merged, considering their uncertainties. An oriented NS fault (normal, inverse or strike slip) and a main shock of given magnitude is considered. Elastic spectral ordinates are attenuated from source to site, considering a given soil type. Buildings are considered to be uniformly distributed at every distance from the fault. Each building has similar seismic response, represented by a bilinear relationship. The performance point is obtained through the N2 method. The response of the damaged structure to aftershocks is still assumed bilinear, passing thorough the origin and the performance point in the main shock. For a given main shock, soil type and building type, uncertainties on the attenuation relationships, on the aftershock magnitude and epicentre and on the Life Safety limit state are considered. The probability of unusability in the main shock and in the aftershocks at each location from the epicenter, given the main shock, soil type and building type, has been evaluated. Results are sensitive to the exposure interval T1-T2, where T1 is the starting and T2 the ending time, measured from the main shock. The results are then compared against the minimum requirement concept, reported in ATC20 and in the Italian usability methodology. In the paper the full methodology is described and several important items are discussed, such as the effects of the spatial localization of strong aftershocks, the aftershock activity decay and the building resilience.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
