Application of nonlinear site response analysis in Coastal Plain South Carolina

The 1933 Long Beach, 1957 San Francisco, 1967 Caracas, 1985 Mexico City, 1989 Loma Prieta, and 1994 Northridge earthquake events left evidences of how the local site condition can affect the characteristics of propagating earthquake wave from the bedrock through the top soil. The ground motion amplitude, frequency content or the duration can be affected by the local site condition and thus can cause significant amplification or de-amplification to the original bedrock motion which can seriously affect the structures. Quantification of such site effect on ground motions is a challenging task. This dissertation is dedicated to improve the existing ground response quantification techniques and the related knowledge base.;The first major attempt towards ground response quantification was the development of the 1994 NEHRP (BSSC, 1995) seismic site factor provision. Since the development of the NEHRP provisions, several studies have found these factors to produce inadequate predictions for the state of South Carolina. In an attempt to generate seismic site factors based on conditions specific to South Carolina Coastal Plain (SCCP), a series of nonlinear one-dimensional ground response analyses are performed by this author as part of a research team considering appropriate soil profiles and location specific ground excitations. After the generation of this new site factor model, a systematic repercussions study is performed by applying earthquake loads, considering both NEHRP and the new site factors, on typical highway bridge structures.;Being exposed to the realm of nonlinear site response studies, the author feels that this sector lacks sufficient benchmarking studies over the code usage protocols and therefore its use in the practitioners' community has been limited to date (Stewart et al., 2008; Matasovic and Hashash, 2012). This author performs a benchmarking study over several widely used nonlinear one-dimensional site response analysis tools considering the ground motions and soil profiles specific to the Charleston, SC region. A few key issues are addressed: (i) the modeling techniques of several NL site response programs are reviewed for the site; (ii) comparative study over the site factors computed based on several nonlinear and equivalent linear analysis programs produce important insights; and (iii) a guideline stating the conditions required for selecting a nonlinear analysis over an equivalent linear analysis program.;One dimensional site response analysis is limited to horizontally layered ground conditions. Earlier studies showed that topographic variations such as ground slopes can significantly affect the computed surface response. Considering the conditions specific to the Charleston, South Carolina area, two-dimensional finite element models of a range of sloping (mild and infinite) ground cases are analyzed. Based on the outcomes, a slope adjustment factor is proposed which modifies the existing one-dimensional site factors to account for the ground inclination in the design.;Shear wave velocity (VS) is an important input for any seismic site response study. The author observed significant shear strain accumulation at VS contrast locations at the layer interfaces in previous studies. A numerical investigation dedicated to the effect of such stiffness contrast on the seismic surface response is performed. Smoothening these contrast locations reduced the imposed shear strain, thus lesser damping and higher surface spectral accelerations are obtained.