Bathymetry inference from free-surface flow features using large-eddy simulation
Abstract/Contents
- Abstract
- Turbulent fluid flow over complex boundaries is of interest in a variety of environmental and engineering applications. For example, dunes are present in nearly all fluvial channels and impact flow resistance, sediment transport, and deposition within many rivers. The goal of this work is to contribute to an improved understanding of bedform-generated turbulence and its interaction with the flow surface. To achieve this goal, we develop a high-performance parallel moving-grid simulation code, test it on several stratified flow scenarios, and use it to study turbulent flow dynamics over wavy boundaries. The object-oriented code solves the time-dependent, three-dimensional, incompressible Navier-Stokes and scalar transport equations in generalized curvilinear coordinates in a large-eddy simulation (LES) framework. Parallelism is enabled via the message-passing interface (MPI) for use in distributed memory parallel computing environments. The aim of the moving grid method was to reduce the numerical diffusion associated with the transport of miscible fluids in environmental flows. This is a difficult problem from a numerical point of view because of the presence of numerical diffusion which tends to overpredict the amount of mixing that takes place. At the heart of the method is a second-order accurate arbitrary Lagrangian Eulerian (ALE) scheme based on r-adaptivity of the underlying numerical discretization to simulate flow with moving generalized curvilinear coordinates. The motion of the mesh is based on the fluid velocity field, however certain adjustments to the Lagrangian velocities are introduced to maintain quality of the mesh. The adjustments are based on the variational approach of energy minimization to redistribute grid nodes closer to the areas of rapid solution variation. Additionally, the moving grid method enforces consistency with continuity (CWC), which ensures local conservation of the advected quantity. CWC is essential for the numerical simulation of stratified flows in which negative values of temperature and/or salinity are unacceptable. Furthermore, the ability to arbitrarily move the simulation mesh at every time step allows: (i) computation of the free surface in cases in which the rigid-lid approximation is invalid and (ii) computation of bed motion (for example, due to ripple dynamics), which affects the flow field in a coupled hydrodynamic moving-bed simulation. Using the Navier-Stokes solver, we studied free-surface turbulence arising from different bedforms in a channel flow to infer the characteristics of the bedforms. As we discovered, coherent structures that produce surface signatures, such as boils of the third kind, are strong even in flat-bed channel flow. Therefore, inferring bedforms from free-surface signatures is difficult because the relative impact of the bedforms is weak in comparison to the background turbulence arising from the flat-bed channel flow. Nevertheless, despite the difficulty in inferring bedform shape, the mean depth can be inferred from the free-surface turbulence by means of an image processing technique to count the frequency of new upwellings. Using this method, it was shown that for a fixed Reynolds number based on the channel depth, the upwelling frequency for flat-bed open channel flow follows an inverse square power-law as a function of channel depth. Therefore, after both the mean surface velocity and the upwelling frequency are measured, the mean depth can be inferred. Additionally, we described how the value of TKE at the free-surface can be used to calculate the effective roughness once the channel mean depth is inferred from the counting technique. The roughness, in turn, provides an idea of the structure of the bedforms. Finally, we investigate the effect of sensor measurement noise on the statistical inverse problem of inferring bedform topography of an open channel from the mean flow features at the free surface. We considered two types of sensor measurement noise: homogeneously distributed, corresponding to instrumental measurement errors, and localized, corresponding to a "faulty" sensor. The arbitrary bathymetry inference problem was solved with an optimization-based approach operating on a database of precomputed simulations of laminar flow over parameterized sinusoidal bedforms.
Description
| Type of resource | text |
|---|---|
| Form | electronic; electronic resource; remote |
| Extent | 1 online resource. |
| Publication date | 2013 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Associated with | Koltakov, Sergey |
|---|---|
| Associated with | Stanford University, Institute for Computational and Mathematical Engineering. |
| Primary advisor | Fringer, Oliver B. (Oliver Bartlett) |
| Thesis advisor | Fringer, Oliver B. (Oliver Bartlett) |
| Thesis advisor | Gerritsen, Margot (Margot G.) |
| Thesis advisor | Iaccarino, Gianluca |
| Advisor | Gerritsen, Margot (Margot G.) |
| Advisor | Iaccarino, Gianluca |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | Sergey Koltakov. |
|---|---|
| Note | Submitted to the Institute for Computational and Mathematical Engineering. |
| Thesis | Thesis (Ph.D.)--Stanford University, 2013. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2013 by Sergey Urievich Koltakov
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...