Effect of Number of Records and Their Properties on the Estimation of Seismic Demands By Nonlinear Analysis.

Abstract eng:
Design codes tend to recommend a minimum number of scaled natural earthquake accelerograms as an option to model the seismic input for time-history analysis. However there is still debate as to what that minimum number of records must be to obtain meaningful results when dealing with non-linear inelastic analysis. This article addresses the issues related to the stability of assessed seismic demands as affected by the number of accelerograms used (and their properties) for time-history analysis. To that effect, a large number of inelastic analyses are conducted using a representative inelastic MDOF system. The seismic input is modelled using a database of accelerograms recorded on rock which have been previously used for the calibration of ground motion prediction equations in Europe and the Middle East. The seismic input is scaled to match the intensity of the standard EC8 design spectrum. Scaling criteria used in the study include amplitude and dual scaling (amplitude + time scaling) to comply with EC8 recommendations related to spectrum matching, as well as, by using an optimum spectrum intensity scale. Results are assessed in terms of the stability of the predicted seismic demands as affected by the size of the family of accelerograms used for the analyses (i.e. families made of 3,4,5,…,11 accelerograms), and by the degree of fitting between the design spectrum and the mean response spectrum of the scaled records of the family and by the scaling criterion.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket