Peridynamic modeling of the micromechanical response of snow under high strain rates

Modeling the mechanical response of snow is a challenge due to the wide range of properties and conditions that impact the mechanics, such as temperature, loading conditions, and the geometry of the microstructure. Current modeling approaches including the finite element and discrete element methods have made great strides in modeling the micromechanical behavior of snow, but they often make simplifications to the microstructure geometry and failure process. However, it has been shown that the geometry of individual snow grains and the sintered bonds that form between them, as well as the failure of these bonds, are important factors in the micromechanical behavior of snow. We investigate a modeling framework called peridynamics that has not been applied to snow problems previously, but has been successfully applied to material fracture problems. This approach can use segmented voxels from mu CT scans of snow as the model input geometry, which allows for highly accurate representations of realistic snow microstructures within the model. We provide the first analyses to determine appropriate simulation domain sizes and the domain resolutions required to capture realistic snow behavior with peridynamics. We show that the approach is able to simulate realistic brittle behavior for snow at very high strain rates (>= 0.1s-1). The model naturally produces highly damaged regions within the sintered bond regions of the microstructure without any preconditioning. We then discuss next steps for further validating peridynamics for snow applications.