Probabilistic seismic assessment is of interest both in the design of new structures, and in the assessment of existing structures. An accurate prediction of the structural response distribution requires a large number of dynamic analyses, using a large number of ground motions. Such a task, however, requires substantial computational work, and, furthermore, is impeded by the scarcity of high intensity records. In this research a methodology is proposed for sampling optimized suites of ground motions, which methodology is comprised of a vector-valued intensity measure (IM), and a ground motion selection and modification (GMSM) method. Using the optimized suites, an optimized response prediction is obtained, which presents a compromise between the reduction in the number ground motions used, and the loss of accuracy. The proposed vector-valued IM is comprised of the spectral acceleration at the fundamental period, the Normalized Spectral Area (NSA) parameter, and, optionally, a measure of the nonlinearity level. NSA is defined as the area of the displacement response spectrum between the fundamental period and the ultimate elongated period, normalized to the spectral displacement at the fundamental period. In this way, it captures the effect of the excitation spectral characteristics (i.e. frequency composition) on the response. NSA is intended to have high correlation to various relative response parameters. With the proposed GMSM method, optimized suites are formed through stratified sampling on the NSA parameter, using datasets of ground motions normalized to the spectral acceleration at the fundamental mode, in order to replicate the IM true central tendency and true dispersion. Consequently, the response parameter true central tendency and true dispersion are also replicated. Stratified sampling results in a reduced standard error of the mean IM, contrasted to random sampling. The advantage of the proposed GMSM method is that when there is sufficiently high correlation between NSA and the relative response parameters, the standard error of the mean is also reduced.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:654757 |
| Date | January 2013 |
| Creators | Theophilou, Artemis I. |
| Publisher | University of Surrey |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://epubs.surrey.ac.uk/811162/ |
Page generated in 0.0021 seconds