Unexpected Fluctuation in Seismic Response Accompanied By Model Uncertainty in Numerical Simulation

Abstract eng:
The focus of this paper is the unexpected fluctuation found in the seismic response due to a slight difference in the settings of structural parameters. Over the past few decades, the performance of computers has improved considerably, and simulation-based techniques with a large mass of information on objective structures can easily be applied to damage evaluations, hazard assessments, and so on. For simulation-based techniques, the numerical models must be set appropriately. Thus, design parameters or other equivalent values for the targeted structures are often incorporated into the numerical models, although it is known that there are certain discrepancies between the design parameters and the actual properties of existing structures. Since the numerical results necessarily fluctuate according to these discrepancies, a kind of “uncertainty” should be taken into account for the structural parameters and for the corresponding responses in order to ensure the reliability of the numerical results. Although it is still quite a task to determine the uncertainty of structural parameters quantitatively, examining the statistical properties of the fluctuation in seismic response caused by certain gaps in the structural parameters is of significant importance. To statistically examine the fluctuation properties of the seismic response, a large number of numerical tests were conducted using several numerical models of the common type and various input earthquakes. This study employed models with single- and multi-degrees of freedom that were modeled by shear-spring masses. The skeletons of the shear springs were modeled by normal bilinear or trilinear types, that are generally applied to concrete structures, base-isolated members, and so on. The yield force and the elastic period were taken as the input structural parameters in this study because the performance of structures can be clearly determined with them and they have a direct effect on the seismic response. The maximum story drift angle was taken as the response quantity, as it is an important factor for evaluating the damage level when building structures. To discuss the fluctuation in seismic response accompanied by slight changes in the yield force and the natural period, the maximum gap between the maximum story drift angles for the corresponding yield force and natural period within a certain band, was examined under several combinations of structural properites and various input earthquakes. Introducing the ratio of fluctuation, Rfj, as an evaluation index, the statistical properties of the seismic response were summarized as follows: 1) The mean of Rfj with a slight change in the yield force highly depends on the ductility level. 2) The mean of Rfj with a slight change in the natural period of structures depends more on the natural period than on the ductility level. 3) A slight difference in structural parameters, for instance, 2% in yield force or 0.01 s in natural period, may possibly cause a difference in the maximum story drift angle of around 10%. 4) Rfj agrees well with the beta distribution.

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