Application of Material Point Method and Mohr-Coulomb Strain Softening Constitutive Model in Simulations of Multiphase Granular Flows

Abstract

Sudden displacement of large volumes of liquid–granular mixtures in nature (dam-breaks, landslides, floods, etc.) are often reported to be deadly and destructive. The flow behavior of those mixtures is complex and depends on a large variety of parameters, e.g., size distribution of the solid phase, viscosity of the fluid, predisplacement packing conditions, ratio of solid to liquid phase, and geometry of the domain. Because of the multitude of parameters and the large displacements involved, the numerical modeling of these phenomena is complex. A two-phase double-point material point method formulation in Anura3D, a particle-based continuum numerical method, was tested against two experimental cases. Model simulations showed that simple constitutive models such as Mohr-Coulomb (MC) with perfect plasticity can be sufficient to accurately model bulk granular flow behavior. However, with slightly different initial conditions, these flows can exhibit more complex features such as progressive block failures, which necessitates a more advanced solid constitutive model such as MC strain softening. Further, other simulation parameters like wall friction boundary conditions and fluidization threshold are also crucial in these types of numerical simulations. The ability of such models to capture complex failure modes is critical to assess dam safety. The sudden release of a mixture of granular materials and water, for example, after failure of a tailings dam, can have devastating consequences on downstream infrastructure and communities. We investigated how such flows can be accurately predicted using the computer model Anura3D, which uses a special technique, where material points separately represent the water and the grains because those points can efficiently represent the large deformations that these types of granular flows exhibit. We found that, at higher water content, a simple model of the mechanical properties of the granular and water mixture suffices. However, at lower water content, the more complex behavior of the granular flow requires a specific model that takes into account the local history of deformations. This study is relevant to the analyses of dam safety, post-wild fire hillslope debris flows, and landslides.