Proceedings of Indian Geotechnical Conference December 22-24,2013, Roorkee NUMERICAL MODELING OF GRANULAR SOILS UNDER CYCLIC TRIAXIAL TESTING Akhila Manne, Research scholar, Geotechnical Engineering Laboratory, Earthquake Engineering Research Centre, IIIT Hyderabad, akhila.manne@research.iiit.ac.in Dr. Neelima Satyam. D, Assistant Professor, Geotechnical Engineering Laboratory, Earthquake Engineering Research Centre, IIIT Hyderabad, neelima.satyam@iiit.ac.in ABSTRACT: The complexity involved in understanding physical phenomena is obvious and can be attempted by simulating them. Simulation of complex phenomena such as liquefaction of soil requires the understanding of mechanism rather than the basis for the cause. Numerical simulations are capable of providing a comprehensible basis for such complex phenomena. In this study, it has been attempted to understand the behavior of granular soils under cyclic loading by 3D modeling of cyclic triaxial test. Nonlinear behavior of soil can be best embodied by Discrete Element Modeling (DEM) as it considers soil as discrete assembly of particles. So, using DEM cyclic triaxial testing has been modeled in undrained condition for an assemblage of random spherical particles of equal and different sizes. From the analysis it has been identified that at a constant confining pressure and loading, uniformly graded sample has a greater resistance to failure during cyclic loading than non-uniform distribution. The outcome of the study is useful to understand liquefaction phenomena. Keywords: DEM, Dynamic soil properties, Liquefaction, cyclic triaxial test INTRODUCTION The nonlinearity of soil is prevalent when exposed to high strain amplitudes (> 10 -3 ). High strains or dynamic forces acting on soil result in stiffness degradation decrease of effective stress and large deformations. Nonlinear behavior of soils has been investigated by many researchers and various methods have been postulated to demarcate and quantify such behavior. Most of the researchers successfully employed laboratory testing and theoretical analysis to arrive at crucial conclusions. Inspite of the analysis, the mechanism involved in the resulting behavior of soil is yet to be understood. Numerical simulation studies can be employed for providing insight into the complex behavior of soil. Soil is an aggregation of discrete particles and it requires to be modeled as discrete elements. Discrete element methods have the capacity to capture the mechanical interaction of discrete particles which cannot be solved by continuum based methods. From the discrete particle model of an idealized granular assembly, the motion of individual particles can be estimated. Using this method, quantification of required properties is from microscopic to macroscopic behavior of the considered assembly. Particle interactions and overall behavior of the system apart from internal and external physical conditions are predominantly influenced by material parameters such as the stiffness of the particles, nature of fluid filling the space between them, grain geometry (size, shape, and surface roughness). The state of the granular particles ascertains the forces acting on them. A constitutive model that considers such factors is to be postulated and used for the study. To obtain a profound understanding of behavior of granular soils, the material/fabric can be studied by subjecting it to different stress paths, stress level, stress history and confining pressures. Most of the research has been done by considering particles as spheres [1, 2] or ellipse [3, 4, 5, 6]. In this study cyclic triaxial test has been modeled using spherical particle shape for simplicity and robustness. For the simulation, PFC 3D (Particle Flow Code) [7] has been applied.