Particle-Resolving Simulations of Dune Migration: Novel Algorithms and Physical Insights

Abstract

Sediment transport is ubiquitous in aquatic environments, and the study of sediment transport is important for both engineering and environmental reasons. However, the understanding and prediction of sediment transport are hindered by its complex dynamics and regimes. In this dissertation, the open-source solver SediFoam is developed for high-fidelity particle-resolving simulations of various sediment transport problems based on open-source solvers OpenFOAM and LAMMPS. OpenFOAM is a CFD toolbox that can perform three-dimensional flow simulations on unstructured mesh; LAMMPS is a massively parallel DEM solver for molecular dynamics. To enable the particle-resolving simulation of sediment transport on an arbitrary mesh, a diffusion-based algorithm is used in SediFoam to obtain the averaged Eulerian fields from discrete particle data. The parallel interface is also implemented for the communication of the two open-source solvers. Extensive numerical simulations are performed to validate the capability of SediFoam in the modeling of sediment transport problems. The predictions of various sediment transport regimes, including `flat bed in motion', `small dune' and `vortex dune', are in good agreement of with the experimental results and those obtained by using interface resolved simulations. The capability of the solver in the simulation of sediment transport in the oscillatory boundary layer is also demonstrated. Moreover, this well-validated high-fidelity simulation tool has been used to probe the physics of particle dynamics in self-generated bedforms in various hydraulic conditions. The results obtained by using SediFoam not only bridge the gaps in the experimental results but also help improve the engineering practice in the understanding of sediment transport. By using the particle-resolving simulation results and the insights generated therein, the closure terms in the two-fluid models or hydro-morphodynamic models can be improved, which can contribute to the numerical modeling of sediment transport in engineering scales.