Grain-resolving simulations of submerged cohesive granular collapse

We investigate submerged cohesive granular collapses via fully coupled, grain-resolving direct numerical simulations, for weakly polydisperse, loosely packed columns. We quantify the time-dependent spreading velocity of the collapsing front, as well as the final runout length and deposit thickness, as functions of the aspect ratio of the initial particle column and a cohesive number formulated as the ratio of cohesive to gravitational forces. We find that the dependence of these properties on the initial aspect ratio and the cohesive number can be accurately captured by piecewise power-law relationships. By means of computational particle tracking we obtain insight into the Lagrangian dynamics of the granular collapse, and specifically into which regions of the initial particle column travel farthest or end up at the surface of the final deposit. In this regard, we observe fundamentally different behaviors for short and tall columns. The simulations furthermore enable us to identify aggregates of particles held together by cohesive forces that undergo very little deformation during the collapse process. Finally, the simulations provide information on the emergence of cohesive and contact force chains, as well as on their preferred spatial orientation.