Seismic Accelerograms

The paper presents results from a recent study in progress, involving an extensive analysis, based on several deterministic and stochastic indices, of the frequency content of ground motions recorded during strong Vrancea seismic events. The study, continuing those initiated by Lungu et al. in the early nineties, aims to better reveal the characteristics of the analyzed ground motions. Over 300 accelerograms, recorded during the strong Vrancea seismic events mentioned above and recently re-digitized, are used in the study. Various analytical estimators of the frequency content, such as those based on Fourier spectra, power spectral density, response spectra and peak ground motion values are evaluated and compared. The results are correlated and validated by using the information provided by various spectral bandwidth measures, as the Vanmarcke and the Cartwright and Longuet-Higgins indices. The capacity of the analyzed estimators to describe the frequency content of the analyzed ground motions is assessed comparatively and the most relevant parameters are being identified.

A recently proposed method for scaling real accelerograms to obtain sets of code-compliant records is assessed. The method, which uses combined time and amplitude scaling, corroborated with an imposed value of an instrumental, Arias type intensity, allows the generation of sets of accelerograms for which the values of the mean response spectrum for a given period range are not less than 90% of the elastic response spectrum specified by the code. The method, which is compliant with both for the Romanian seismic code, P100-1/2006, and Eurocode 8, was described in previous papers. Based on dynamic analyses of single-degree-of freedom (SDOF) and of multi degree-of-freedom (MDOF) systems, a detailed application and assessment of the method is performed, for the case of the long corner period design spectrum in Bucharest. Conclusions are drawn on the advantages of the method, as well as on its potential improvement in the future.