|Numerical modelling of near-bottom sediment transport: turbulence modulation, new process models and application to the Scheldt and the Belgian coast|
Bi, Q. (2015). Numerical modelling of near-bottom sediment transport: turbulence modulation, new process models and application to the Scheldt and the Belgian coast. PhD Thesis. KU Leuven, Arenberg Doctoral School: Leuven. xvi, 228 pp.
Hydraulics Laboratory, Department of Civil Engineering, KU Leuven, more
Sediment transport due to fluid motion is a crucial process in many environmental and engineered systems. Therefore, understanding sediment transport is critical for predicting sediment movements and evaluating the short and/or long-term influence to the surface water systems. Despite the importance of sediment transport, the fundamental aspects involved are far from being completely understood. At the core of the problem is the complex interaction between a turbulent flow field and sediment particles. In this sense, for sediment transport modelling, how to account for as many turbulence generation or destruction mechanisms as possible is one of the keys to improving the accuracy of prediction. The most widely used and validated model for turbulence is two-equation k-e model. However, the traditional k-e model cannot provide accurate predictions when considering the presence of sediment in water due to missing the terms in the equations that can account for energy transfer between sediments and turbulence, as well as the terms for inter-particle interactions. Moreover, the standard k-e model is only valid for fully developed turbulent flow, so it often has difficulties to deal with near-bottom layer (e.g. low-Reynolds effects and high sediment concentrations). Therefore, an alternative approach to overcome these difficulties is two-phase flow theory. In the first part of this study, a modified two-equation k-e model with additional turbulence modulation terms, accounting for the influence of particles on the turbulent flow field, has been proposed. These extra terms are derived using a two-phase flow approach. In the numerical tests, the modified two-phase k-e model reproduces the features of turbulence modulation observed in the experiments on sediment-laden flow.
Another important aspect in sediment transport is the treatment of the near-bottom layer. It has been hypothesized for more than a decade that currently used sediment transport models for morphodynamic studies (e.g., harbour siltation, system response to dredging or structures) can be improved considerably in their predicting capacity when the near-bottom high-concentration effects in the last few centimetres above the bed can be accounted for with a physics-based model. Indeed, experimental observations and theoretical considerations (partially based on two-phase micro-scale models) show that the flow and suspension capacity of the water column are strongly affected by the processes of particle-fluid interactions in the benthic layer. Hence, the second part of this study focusses on the efficient modelling of near-bottom sediment transport in morphodynamic modelling codes through an innovative methodology. In this part, the performance of a few new physics-based process models has been investigated by implementation into a numerical model for the simulation of the flow and morphodynamics in the Western Scheldt estuary. In order to deal with the complexity within the research domain, and improve the prediction accuracy, a two-dimensional (2D) depth-averaged model has been set up as realistic as possible, i.e. including two-way hydrodynamic-sediment transport coupling, mixed sand–mud sediment transport (bedload transport as well as suspended load in the water column) and a dynamic non-uniform bed composition. A newly developed bottom friction law, based on a generalised mixing-length (GML) theory, is implemented, with which the new bed shear stress closure is constructed as the superposition of the turbulent and the laminar contribution. It allows the simulation of all turbulence conditions (fully developed turbulence, from hydraulic rough to hydraulic smooth, transient and laminar), and the drying and wetting of intertidal flats can now be modelled without specifying an inundation threshold. Erosion and deposition in these areas can now be estimated with much higher accuracy, as well as their contribution to the overall net fluxes. Furthermore, Krone’s deposition law has been adapted to sand–mud mixtures, and the critical stresses for deposition are computed from suspension capacity theory, instead of being tuned. The model has been calibrated and the results show considerable differences in sediment fluxes, compared to a traditional approach and the analysis reveals that the concentration effects play a very important role. The new bottom friction law with concentration effects can considerably alter the total sediment flux in the estuary, not only in terms of magnitude, but also in terms of erosion and deposition patterns.