Evaluation of One-Dimensional Seismic Site Response Analyses at Small to Large Strain Levels

Thesis (Master's)--University of Washington, 2016-08

One dimensional (1D) site response analysis using total stress approach is a popular framework for evaluating seismic hazard at a site where no significant excess pore water pressure generation is expected. Several site response analysis codes are available, but their variabilities to predict site response at shear stress levels approaching the shear strength of the soil have not been demonstrated. This study evaluates the performance of different soil models employed in each code to predict the nonlinearity behavior of soil over a wide range of strain levels. This research performed a set of 1D site response analyses utilizing input motions, scaled to various intensity levels, against sites that were underlain by cohesive deposits with determined shear wave velocity profiles. The analyses utilized several available nonlinear soil models while the model-to-model variability was characterized. These codes were then validated against free-field downhole data from a vertical array at a relatively well characterized site. The evaluations of the variabilities of ground motion amplitude, duration, response spectra and cyclic hysteresis loop at various strain levels were performed by comparing all predictions to data from a vertical array. The results showed that all codes give consistent predictions with reasonable accuracy at small to moderate shear strain levels. At larger shear strain levels, only some of current nonlinear soil models were capable of predicting reasonable cyclic behavior in terms of being able to approach the peak shear strength with reasonable damping behavior. The analyses show that the variability of predicted peak shear strain parameters are higher than peak acceleration and shear stress parameters. It also shows that the coefficient of variation of the ground motion parameters predicted by all codes tended to increase at greater shear strain levels.