Physical and numerical dynamic response modeling of slopes and embankments

This research involves a physical model-based experimental study, along with numerical simulations, to investigate (i) the applicability and validity of established 1-g similitude laws (ii) the repeatability and reliability of 1-g physical modeling technique, and (iii) the effects of topographic amplification on the overall dynamic response of slopes and embankments. The "modeling-of-models" method was employed to investigate the validity and applicability of established similitude laws for 1-g physical modeling technique. In the modeling-ofmodels experimental study, model tests of a prototype were conducted at different scaling factors. The modeling-of-models experimental study including three small-scale laboratory built slopes comprised of model clay conducted under static loading. Additionally, three shake table tests of laboratory-built model clay embankments were conducted under a suite of ground motions including both synthetic and recorded earthquake motions that varied in frequency, duration and amplitude. The largest respective models from these two sets of experiments were considered as the "prototype" and the other two smaller models' geometry, strength, low-strain dynamic properties and frequency content of the input motions were adjusted by applying 1-g similitude laws to reflect the reduced scale of "prototype". The repeatability of 1-g physical modeling technique was investigated for dynamic and static loading conditions by performing two small-scale model tests under identical conditions. Possible sources of uncertainty arising from the model construction process, soil preparation, variation in soil properties as well as boundary conditions were investigated. The physical modeling technique was found to be well controlled, and soil preparation and model construction has a very limited impact on the results. The inherent variation of soil properties (water content, undrained shear strength and shear wave velocity) was negligibly small. The main source of uncertainty in 1-g physical modeling tests pertained to the boundary conditions. The effects of surface topography of slopes and embankments on overall dynamic response was investigated by conducting shaking table tests on cohesive embankment models withdifferent slope inclinations coupled with numerical simulations using two-dimensional finite difference program FLAC. A previously performed centrifuge test of a clean sand model embankment was studied to examine differences in the patterns of topographic amplification in cohesive and granular materials.