Computer simulation of dynamic load transfer and wave propagation in particulate media

A numerical study has been conducted to investigate the local dynamic load transfer processes and the resulting wave motion in granular and particulate materials. Of specific interest were the effects of dynamic intergranular contact constitutive relations, wavelength, particle size and shape, and local microstructure or fabric. This research also has developed a relationship between the microstructure of granular material and its macroscopic wave propagational behaviors. Granular media were simulated by two dimensional assemblies of circular and elliptical particles. Dynamic loading was transient in nature and was modeled by triangular time dependent pulses with durations of 60-900 $\mu$s. The discrete element method was the numerical routine used to simulate the dynamic particulate behavior. Several intergranular contact laws were studied, and comparisons of the numerical predictions with existing experimental data were made to determine a suitable dynamic contact law. It was found that a nonlinear hysteretic law produced the best agreement with the experimental data. By varying the input loading period on a single disk particle chain, the effects of wavelength on wave propagation were investigated, and an observed ringing phenomenon was explained from the viewpoint of soliton dynamics. Elliptical particles with different aspect ratios were used for a preliminary investigation of particle shape effects. Large random granular assembly generators were developed to construct granular media with different anisotropies and microstructures. Using discrete element modeling to simulate the wave motion, the effects of this anisotropy/microstructure were investigated by propagating plane waves in two perpendicular directions in these assemblies. Several granular fabric measures were used to characterize the microstructures, and detailed comparisons were made to determine which fabric was the most successful to correlate with the wave propagational characteristics. The void polygon fabric measure was found to describe these characteristics better than branch and path fabrics