Strain localization is a ubiquitous feature of granular materials undergoing nonhomogeneous deformation. Localized deformation typically is followed by a reduction in the overall strength, and thus can have a significant impact on material and structural behavior. Because shear bands are quite often observed in soils, it is of considerable interest and importance to the geotechnical community to be able to capture the full effects of strain localization in predictive models for analysis and design. Of key relevance are the ability to predict when a shear band forms, how this narrow zone of discontinuity is oriented within the material, and how the propagation of the shear band influences the post-localization constitutive response. Currently, even the most advanced and well-calibrated numerical models cannot predict the onset of localization, as the mechanisms governing localized deformation still are not properly understood.
The development of more accurate mathematical models of soil behavior thus requires a more fundamental understanding of the localization phenomena; in particular, the important factors responsible for the inception and development of localized deformation. The objective of this research is to combine state-of-the-art geotechnical experimental techniques with advanced finite element modeling to obtain a more thorough understanding of the strain localization process in sands. A meso-scale modeling approach will be used, which will treat specimen response as a structural response and will incorporate the measured spatial density variation and other imperfections (natural and imposed) to analyze the specimen response as a boundary-value problem. Experimentally, the technique of X-Ray Computed Tomography (CT), widely used in medical applications, will be used to capture meso-scale density variations in plane strain specimens of sand. Digital Image processing techniques will aid in transferring of the CT results as input into the finite element models. Finally, Digital Image Correlation (DIC) will be used to track local, in-plane displacements throughout deformation. The modeling will consider effects of both strong and weak imperfections, both imposed and naturally occurring. By capturing all of these imperfections, the potential of existing strain localization models for application to practical boundary-value problems can, finally, truly be assessed.
This NSF award will enable identification of important factors that contribute to the initiation of strain localization in sands, yielding tremendous insight as to why persistent shear bands form where they form in granular materials in general. That the Geomechanics and Geotechnical Systems Division of NSF contributed to sponsoring the 6th International Workshop on Bifurcations and Instabilities in Geomechanics (IWBI) in 2002 highlights the need for engineering input in this active research area. Recognition of the fundamental deficiencies of the standard FE method and development of techniques to circumvent these difficulties have immense implications to how geotechnical engineers analyze boundary-value problems in practice, particularly in the regime of instability and softening. Furthermore, the use of advanced scanning and data imaging techniques available in other fields, such as those used in medical and materials sciences, will put the field of geotechnical engineering at parity with current technology. The proposed partnership between numerical and experimental research will ignite a more thorough approach to investigating the localization phenomenon.
The second PI is a recent member of the faculty at JHU, and the proposed research will help her to develop a strong research group in advanced geotechnical experimentation that can provide mentoring to women and minorities. Currently the first PI supports two underrepresented graduate students (Black and Hispanic) while the second PI supports two undergraduates, one of whom is a woman, in her research group. Both schools have been very conducive to departmental support of undergraduate involvement in research, and to support of underrepresented students. Through research exchange programs with local high schools, the laboratory and simulation components of the proposed research will serve as ideal avenues to engage high school students in geotechnical engineering and the research process in general. The union of numerical and experimental research will offer Stanford University and Johns Hopkins University graduate students a more multifaceted approach to graduate education.
