Analysis of liquefaction-induced instability and failure in slopes and embankments is an important problem of geotechnical earthquake engineering. While various empirically-based and/or simplified approaches have been developed and used in practice, the current state of knowledge of stress-strain-strength behavior of liquefiable soils, the wealth of observed response of soil slopes and embankments during recent earthquakes, and the significant progress in numerical modeling of engineering structures in recent years, have prepared the ground for a more rigorous, yet realistic, approach to the analysis of liquefaction-induced instability and failure. This project is to develop a multiscale approach, which allows the soil response to be captured in both pre- and post-failure stages with a unified constitutive model. Moreover, since the traditional finite element (FE) approach for fully-coupled effective stress analysis of soils is severely limited in handling the near- and post-failure response of a geotechnical system, a meshfree method is proposed to overcome the known difficulties of an FE approach, and to produce reliable and robust solutions in all stages of a liquefaction analysis.
The intellectual merit of the project lies in the development and introduction of two very powerful constitutive and numerical methods in the analysis of soil liquefaction. The proposed methods will be implemented in a public domain open source code, which can be used by researchers and practitioners. The developed computational tool will be validated and verified against a series of experimental and field observations on the responses of soil slopes and embankments.
The broader impacts of the project include the potential use of the techniques developed in this research in other fields of engineering such as bio-poromechanics and hot rolling processes in metals and alloys. The graduate students trained in the course of this project will have an excellent opportunity to develop a thorough understanding of the mechanisms involved in liquefaction and liquefaction-induced deformation and failure as well as an expertise in modeling of geotechnical systems using two of the most advanced methods of analysis and modeling in engineering mechanics. The depth and breadth of such research training should be of significant value to geotechnical engineering practice.
The main outcome of this project was the development of a two-scale constitutive model for saturated granular materials that are susceptible to liquefaction. The two-scale model is able to resolve the intricate mechanisms of soil internal resistance to shear loading by integrating the microscale shear-induced mechanisms with the macroscale features of the soil response. Key features such as grain rotations within highly localized shear zones are considered. The model was shown to capture the essential features of the stress-strain-strength response of sands in the pre- and post-failure conditions, particularly when localized shear zones emerge within the soil. The project also produced a multiscale analysis framework that can be used for post-failure stability and deformation analysis of geostructures that contain zones of liquefied material. The multiscale constitutive model and analysis framework are shown to produce realistic patterns of failure within heterogeneous soil specimens that are subjected to biaxial loading. Four graduate students and three undergraduate students were involved in the research activities and received training in the course of the project. One PhD dissertation and one master's thesis were produced. The project also provided significant opportunities for development of educational modules on earthquake engineering that were used for outreach to grade school and undergraduate students.
