Summary of Final Contract Report ObjectiveThe objective of this investigation is to develop and evaluate a new analytical approach for estimating site-dependent earthquake ground motion for engineering design purposes.The analytical approach is based on assuming physical models for the earthquake source, path wave-propagation effects, and nonlinear soil response at the site, then calculating ground motion using a combination of seismic (layer matrix) and finite element methods. The approach is structured to potentially provide a three-dimensional ground motion capability, although the nonlinear soil response calculations are here limited to one dimension. The predictive capabilities of this approach were examined by comparing ground motion calculations to empirically derived ground motion trends and to ground motion simulations performed using the equivalent linear method.
Principal ResultsThe most significant results of this work are the following: (1) A superior soil model, the endochronic model, has been introduced to site evaluation studies.(2) Comparison of the endochronic model with the conventional equivalent linear model has shown the latter to yield excessively conservative ground motion predictions for loose soil deposits in the near field.
SummaryFor distances less than about 20 kilometers, there is minimal earthquake ground motion data available, and substantial disagreement exists among empirical relations for predicting design motion characteristics in this distance range. This disagreement reflects the uncertainties associated with extrapolation of the empirical relations to short distances, where data is sparse. To the extent one can construct adequate analytical models for the seismic source, path wave-propagation effects, and site nonlinear response, it becomes possible to simulate ground motion for geologic environments, distance ranges, and earthquake magnitudes for which data is scarce or nonexistent.The SH component of rock-site ground motion at distances 5, 10, and 30 kilometers was synthesized for an earthquake model which consisted of an assemblage of discrete, circular cracks, representing a composite, extended source. It was found that strong motion observations are, in general, inconsistent with a rupture mechanism in which crack growth stops instantaneously. Once more realistic rupture deceleration was incorporated into the model, the SH component of motion alone was apparently adequate for simulating ground motion consistent with observed peak acceleration, peak velocity, and strong motion duration.A new nonlinear constitutive model for soils, the endochronic model, was fit to cyclic shear data for dry sand, and an equivalent linear constitutive model was fit to the same sand data. Using the endochronic model, close agreement was achieved between calculated (solid curves) and observed (dashed curves) stress-strain behavior of dry sand over 300 cycles of deformation, as illustrated below.
Crete 2Only-five parameters were required to accomplish the fit to the sand data.The nonlinear e...