|Ph.D Student||Mehr Arilon|
|Subject||Pore Water Pressure Generation in Saturated Sand during|
|Department||Department of Civil and Environmental Engineering||Supervisor||Professor Mark Lloyd Talesnick|
The response of saturated sand to undrained cyclic loading is usually characterized by pore water pressure generation. In order to examine the phenomenon from its roots, it is necessary to perform laboratory tests which simulate what actually happens in the field. One of the conclusions from analyzing triaxial and simple shear test results, is that they do not reliably simulate what occurs on site. Therefore, within the framework of the study new equipment was developed in order to perform torsional shear tests on hollow cylinder specimens of sand.
The study presents results of experiments based on four stress paths relevant to earthquake loading. The test results have been used to analyze existing and new models, in the framework of pore water pressure generation during cyclic loading.
Many researchers use a bounding surface plasticity approach in order to simulate sand behavior under cyclic loading. The present work accentuates the inapplicability of elasto-plastic models based on boundary surface plasticity.
A series of pure deviatoric drained shear tests on hollow cylinder specimens of saturated sand was carried out in which loading, unloading and reloading was performed. The results indicate that there was stress memory during the unloading stage of the shear stress versus shear strain relationship. Nevertheless, there was no evidence of the stress memory in the shear stress versus volumetric strain relationship; the volumetric strains developed from the beginning of loading. This finding defies the existence of a purely elastic zone in sand. The finding casts doubt in the reliability of the division of strain into elastic and plastic portions when simulating earthquake loading of saturated sand.
The fact that there is a memory stress stage in the shear stress-shear strain space in both drained and undrained stress paths (with a slight modification due to the pore pressure build up) leads to the adoption of the Masing's rules for cyclic shear stress versus shear strain response. Modification of the coefficient in the original rules leads to better a fit to the experimental data. Analysis of the experimental results has led to the adoption of the modified hyperbolic model (MKZ) for the backbone shear stress versus shear strain behavior; pore water pressure generation is updated during loading.
The study includes development of a new hyperbolic model which correlates pore water pressure generation to modified strain energy. Combination of this model with the modified hyperbolic model for the backbone shear stress-shear strain relationship, and with the slightly modified Masing rules, provides an acceptable fit to the laboratory test results.
The study presents the application of the chosen model to a seismic, site response problem. A computer program was written for the site response based on a nonlinear solution, which considers pore water pressure generation and dissipation during earthquake loading.