EP008-05
A close look at the effect of grain shape on bedload transport

Monday, 7 December 2020: 16:16
Virtual
Eric Deal1,2, Qiong Zhang3, J Taylor Perron4, Santiago Benavides5, Ken Kamrin3 and Jeremy G Venditti6, (1)Massachusetts Institute of Technology, Earth, Atmospheric, and Planetary Sciences, Cambridge, MA, United States, (2)ETH Zurich, Zurich, Switzerland, (3)Massachusetts Institute of Technology, Mechanical Engineering, Cambridge, MA, United States, (4)MIT---EAPS, Cambridge, MA, United States, (5)Massachusetts Institute of Technology, Earth, Atmospheric and Planetary Sciences, Cambridge, MA, United States, (6)Simon Fraser University, Department of Geography, Burnaby, BC, Canada
Recorded Presentation
Abstract:
It is well established that grain shape influences granular dynamics, as demonstrated by the dependence of the angle of repose on grain shape. In addition, grain shape influences fluid flow around grains, as demonstrated by the dependence of settling velocity on grain shape. Both of these effects should impact bedload transport, yet grain shape is typically ignored in bedload transport laws and models because there is no widely accepted framework that accounts for grain shape.

To address this knowledge gap, we observed bedload transport of both glass spheres and natural grains (river gravel) in a laboratory flume over a wide range of flow conditions. Using computer-vision-based particle tracking, including a novel machine-learning algorithm for tracking natural grains, we determined trajectories for the majority of grains in the experiments (>95% for spheres, >70% for gravel). Together with flow velocity fields from particle image velocimetry, the grain trajectories offer detailed quantification of bedload transport. We measure a factor of 2-3 higher mass flux of spheres versus natural gravel for the same Shields number. We probe this difference through comparisons of mean fields of several parameters, including grain and fluid velocity profiles, packing fraction, and granular temperature profiles.

We then compared the laboratory results for glass spheres with numerical simulations designed to mimic the lab conditions as closely as possible. The numerical simulations are based on a Lattice-Boltzmann fluid method coupled to a discrete particle model (LBM-DEM), allowing us to capture the dynamics of fluid-grain and grain-grain interactions at the grain scale. Comparing the mean-field measurements from the flume experiments with the simulations, we find that the simulations match the observations closely across the full range of boundary shear stresses tested, allowing us to use the simulations to explore grain-scale solid- and fluid-phase stresses that are impossible to measure in the flume.

<< Previous Abstract | Next Abstract >>