[SY-F2] The Role of Grain Shape in Discrete Element Modeling of Snow Mechanics
The macro-mechanical properties of the heterogeneous material snow are determined by its microstructure. Single ice crystals are connected by sintering and form a complex ice skeleton. In this microstructure fracture occurs under high strain rate deformations, which are relevant for applications such as avalanche risk forecasting or snow mobility. During such a deformation snow transforms from a porous sintered into a granular material. Both states can be represented by the discrete element method, which is well suited for the mechanical description of snow. However, a convenient but strongly simplified assumption is a spherical shape of the snow particles. The grain shape is important in both structural states: it affects the granular dynamics and controls the number and size of inter-particle contacts, which are crucial for the strength and fracturing behavior of the structure. Yet, a detailed reconstruction of the snow grains with discrete elements is neither preferable for the computational efforts nor necessary for most applications. In laboratory experiments, we investigate the mechanical properties of ice beads and two natural snow types, differing in crystal shape and size. The ice beads allow one-to-one comparison with the DEM simulations and are used as a well-defined intermediate step to later adapt the model to the natural snow types. Compression tests were performed under variation of the sintering time, which is related to the bond size and therefore critical for the strength of the snow microstructure. By means of 3D computed tomography images, the number and size of bonds between the ice beads are quantified and simulations can be performed with the same bead configuration as in the experiments. The relation of the sintering state and the macroscopic strength of the sample is examined in experiments and simulations. Once the simulations perform well in the bead case, the model is modified in terms of bond number and size, to match the mechanical behavior of the natural snow types. Accounting for the grain shape in our DEM snow model remarkably improves the simulations of snow micro- as well as macro-mechanical processes.