Bedrock erosion by sliding wear in channelized granular flow

Friday, 19 December 2014
Chi-Yao Hung1, Colin Peter Stark2, Hervé Capart1, Breannan Smith2, Henrique Teles Maia2, Liming Li2 and Meredith D Reitz3, (1)National Taiwan University, Taipei, Taiwan, (2)Columbia University in the City of New York, Palisades, NY, United States, (3)Lamont-Doherty Earth Obs, Palisades, NY, United States
Boundary forces generated by debris flows can be powerful enough to erode bedrock and cause considerable damage to infrastructure during runout. Bedrock wear can be separated into impact and sliding wear processes. Here we focus on sliding wear. We have conducted experiments with a 40-cm-diameter grainflow-generating rotating drum designed to simulate dry channelized debris flows. To generate sliding erosion, we placed a 20-cm-diameter bedrock plate axially on the back wall of the drum. The rotating drum was half filled with 2.3-mm-diameter grains, which formed a thin grain-avalanching layer with peak flow speed and depth close to the drum axis. The whole experimental apparatus was placed on a 100g-ton geotechnical centrifuge and, in order to scale up the stress level, spun to a range of effective gravity levels. Rates and patterns of erosion of the bedrock plate were mapped after each experiment using 3d micro-photogrammetry. High-speed video and particle tracking were employed to measure granular flow dynamics. The resulting data for granular velocities and flow geometry were used to estimate impulse exchanges and forces on the bedrock plate. To address some of the complexities of granular flow under variable gravity levels, we developed a continuum model framed around a GDR MiDi rheology. This model allowed us to scale up boundary forcing while maintaining the same granular flow regime, and helped us to understand important aspects of the flow dynamics including e.g. fluxes of momentum and kinetic energy. In order to understand the detailed processes of boundary forcing, we performed numerical simulations with a new contact dynamics model. This model confirmed key aspects of our continuum model and provided information on second-order behavior such as fluctuations in the forces acting on the wall. By combining these measurements and theoretical analyses, we have developed and calibrated a constitutive model for sliding wear that is a threshold function of granular velocity and stress.