Search

It is well recognized that the macro-scale response of soils is anisotropic in terms of strength, stiffness, permeability, etc. The source of this anisotropy is thought to be an anisotropy of the material itself. This anisotropy can be quantified using statistical methods if DEM numerical simulations or advanced experimental techniques are used. The anisotropic response of soil has been analyzed by many researchers in terms of the fabric tensor, which provides a measure of the orientation of the contacts between particles. Although many approaches for the quantification of the evolution of soil fabric have been used, they have not been previously compared to assess their effectiveness to describe fabric changes. A direct comparison of different methods of fabric quantification is presented in this paper based on the results from DEM simulations under different stress paths and the suitability of each of these methods is discussed. The results highlight the need for more accurate methods and/or approaches to accurately describe the evolution of fabric anisotropy in granular materials. [ABSTRACT FROM AUTHOR]

While there have been many discrete element method (DEM) publications considering the micromechanics of granular materials subject to monotonic loading, studies of the particle-scale material response to cyclic or repeated loading have been comparatively rare. From a geotechnical perspective soil is subjected to repeated loading in a variety of situations. Examples include foundations to railways and roads, foundations to wind turbines, soil adjacent to integral bridges, etc. The work described in this paper extends an earlier study by O’Sullivan et al.. [1]. In this earlier study, DEM simulations of strain controlled cyclic triaxial tests were coupled with laboratory experiments to validate a DEM model. The simulations were performed using the axi-symmetric DEM formulation proposed by [2] and a stress controlled membrane algorithm was used to apply forces to balls along the outer vertical boundaries to model the latex membrane used in the laboratory tests. Specimens of uniform spheres and mixtures of sphere sizes were considered in the validation stage of this research. The earlier study considered strain amplitudes of 1%, 0.5% and 0.1%. In the current study the response is extended to consider the smaller strain amplitude of 0.01%. All of the simulations were carried out in a quasi-static mode and in all cases the maximum stress level mobilized was significantly lower than the peak stress measured in equivalent monotonic physical tests and DEM simulations [2]. In examining the response of the material to the smaller strain amplitude, the macro scale analyses considered the stress strain response and specimen stiffness. At the particle scale, the variation in coordination number and deviator fabric are considered as well as the distribution of the contact forces orientations. The findings may provide insight to the development of continuum constitutive models for cyclic soil response that include fabric parameters [3]. [ABSTRACT FROM AUTHOR]