Constitutive modeling of post-liquefaction cyclic response of sands

点击翻译按钮获取中文摘要

Large post-liquefaction deformation is a major cause for seismic induced hazards in soil foundation and geotechnical structures, and has been a subject of extensive research since its observations in several well documented earthquakes. While there has been advances in constitutive modeling of granular materials, particularly within the platform of critical state soil mechanics and with the use of bounding surface plasticity, significant challenge is still remaining in modeling the shear strains in the post-liquefaction regime. The objective of this project is to extend the formulation of the SANISAND family of models. Despite of the elegant mathematical formulation of the SANISAND up to the point of initiation of liquefaction (nearly zero mean effective stress), the predicted shear strain accumulation at low mean confinement pressure is not in agreement with laboratory experiments, and the model predictions in the post-liquefaction regime significantly differ from those observed in the experiments. The complexity of the mechanisms involved, and experimental limitations on measurement of the involving components such as micro-mechanical fabric and stiffness, justify the use of mathematical and computational tools in accordance with physical postulates, such as discrete element modeling (DEM), for understanding the involved mechanisms. The research proposal has 4 well-defined parts. The first part is to study the nature of the problem from a particle-to-particle perspective using DEM simulations. The second part deals with the extension of the formulation of SANISAND model, where the model extension will be inspired from the DEM simulation results, and incorporated to an existing model implementation in 2- and 3-dimensional continuum based finite difference platform. In the third phase the resulting developed and implemented constitutive model will be verified and validated, to ensure about its performance and robustness, so the compiled model can be made accessible as a Dynamic Linked Library (DLL) to other modelers in the field. Finally, the developed tool will be used in modeling the liquefaction response in the foundation of a tailings dam during a seismic event.****