|dc.description.abstract||Earthquakes are one of the greatest natural disasters to human life and properties. Following lessons from previous earthquake disasters, the performance-based seismic design is increasingly accepted by engineers to prevent seismic disasters. In performance-based seismic design, realistic and reliable design response spectra are required to reliably and accurately predict responses of designing structures. However, the mostly used ground response spectra, i.e., Newmark design spectrum and Uniform Hazard Spectrum on soil surface, and floor response spectrum constructed by current methods do not properly meet
the requirements of performance-based seismic design:
1. Newmark design spectrum exhibits lower amplitudes at high frequencies and higher
amplitudes at low frequencies. Thus, it cannot realistically and reliably reflect seismic
features of target sites.
2. Variability of soil parameters, nonlinear property of soils, and vector-valued seismic
site response analysis are not integrated into the process of constructing Uniform
Hazard Spectrum on soil surface in modern methodologies. Thus, the desired design
response spectrum is not realistically and reliably represented.
3. An efficient method to generate probabilistic floor response spectrum considering random ground motions has not been addressed. The direct spectra-to-spectra method
to generate floor response spectrum is superior to the time history analysis method in
efficiency. However, this method is not applicable currently to generate probabilistic
floor response spectrum.
The objective of this study is bridge the gap between performance-based seismic design
and realistic design response spectra.
1. Considering the problem of Newmark design spectrum, this study establishes a system
of site design spectrum coefficients to modify the Newmark design spectrum. The
modified Newmark design spectrum could more realistically and reliably represent
seismic features of target sites.
2. To obtain more realistic and reliable Uniform Hazard Spectrum on soil surface, this
study integrates the variability of soil parameters, the nonlinear property of soils,
and the vector-valued seismic site response analysis into the process of constructing
Uniform Hazard Spectrum on soil surface.
3. This study investigates contribution of ground motions (i.e., tuning cases) to the
uncertainty of floor response spectrum, and established the statistical relationship
between t-response spectrum and ground response spectrum. Using this statistical
relationship, probabilistic floor response spectrum by the direct spectra-to-spectra
method considering random ground motions could be constructed.
With results of this study, the most economic solution to the balance between the safety
and economy is expected to reliably obtain for performance-based seismic design for the