|On the dynamics of high-concentrated mud suspensions|
Winterwerp, H. (1999). On the dynamics of high-concentrated mud suspensions. Communications on Hydraulic and Geotechnical Engineering, 99.3. PhD Thesis. TU Delft. Faculty of Civil Engineering and Geosciences: Delft. viii, 172 + appendices pp.
Deel van: Communications on Hydraulic and Geotechnical Engineering. Delft University of Technology. Department of Civil Engineering: Delft. ISSN 0169-6548
|Auteur|| || Top |
Many harbour basins and navigation channels suffer from rapid siltation, forcing managing authorities to carry out expensive maintenance programmes to safeguard navigation. This rapid siltation is often attributed to the occurrence of high concentrations of suspended cohesive sediment and/or fluid mud observed in the turbidity maxima of estuaries or during episodic events in coastal areas. The present study is dedicated to the transport and fate of such High-Concentrated Mud Suspensions (HCMS) with emphasis on the processes in the vertical.
HCMS is defined as a suspension with Newtonian fluid properties, that measurably affects the turbulent mixing processes in the water column. From an extensive literature review it is concluded that HCMS is encountered all over the world under a large variety of hydrodynamic and meteorological conditions. Its occurrence may vary largely with time and in space, however. Abundant availability of mobilisable mud seems to be a sufficient condition for such occurrences.
The behaviour of HCMS in the water column is governed by vertical mixing processes, and by (hindered) settling processes. These mixing processes are affected by sediment-induced buoyancy effects and are a function of the mass concentration of the suspended sediment, whereas a proper description of the (hindered) settling processes requires both the mass and volumetric sediment concentrations. The relation between mass and volumetric concentrations is provided by a fractal description of the mud floes, implying a power law behaviour of various mud properties. From this description a new formulation for the settling velocity as a function of floc size is derived, which is consistent with Stokes' law in the case of massive particles with a fractal dimension 3 (e.g. sand), and which agrees well with empirical data from the literature. The hindered settling formula by Richardson and Zaki, derived for fairly large, massive particles, does not account for the floc structure typical for cohesive sediment, and a new formula is proposed. This formula also compares well with empirical data from the literature.
The evolution of floc size, hence settling velocity, in a turbulent environment is described through a new flocculation model in a Eulerean framework, that includes the effects of turbulence-induced aggregation and floc breakup. This model predicts that the growth of floes in open water systems can seriously be limited by a limited residence time of these floes in the water column, as a result of small water depth and/or of long flocculation times. The model also predicts gelling concentrations in estuaries of the right order of magnitude; gelling values in coastal areas under storm conditions are grossly underpredicted at present. Though this flocculation model compares well with the scarcely available empirical data and yields qualitatively sound results, extensive further validation against comprehensive data sets is required before the model can be deployed with confidence for practical applications.
The sediment-induced buoyancy effects, mentioned in the third paragraph, already become manifest at moderate suspended sediment concentrations, for non-cohesive as well as cohesive sediments. An important implication of these buoyancy effects is the existence of a saturation concentration Cs, which is a measure for the maximal sediment load that can be carried in suspension by a turbulent flow. At concentrations beyond C,, the flow becomes super-saturated and both the turbulence field and the concentration profile collapse.
For steady flow conditions, a simple scaling law for C, can be derived on the basis of a critical flux Richardson number, which value is known from stratified flow experiments reported in the literature. For tidal flow conditions, scaling laws for C. are derived from an integral entrainment model. As these are fairly complicated, these scaling laws are not easy to apply in practice.
The saturation concentration for mud can be compared with the equilibrium concentration which is well-known for sand suspensions. A major difference with sandy suspensions, though, is that sand immediately forms a rigid bed upon deposition, whereas cohesive sediment generates a fluid mud layer, the initial concentration of which is determined by the gel point of the mud. At the fluid mud - water interface little turbulence can be generated, which explains the above mentioned collapse. The fluid mud can become rigid by consolidation processes, after which the production of turbulent energy at the water-mud interface becomes possible again, so that mixing is restored.
This consolidation process is described with a new consolidation model in a Eulerean reference frame, that includes both the effects of consolidation and hindered settling, using a fractal description of the material functions. The consolidation model compares excellently with a numerical benchmark experiment described in the literature, and the combined model gives a reasonable description of a hindered settling - consolidation experiment. However, in combination with the simple Bingham model used in this study, this consolidation model largely overpredicts the strength evolution of a consolidating mud layer.
The various process formulations have been implemented in a one-dimensional vertical numerical model, referred to as the lDV POINT MODEL in this thesis, which allows the simulation of the effects of flocculation and gellation, settling, hindered settling and lutocline formation, consolidation, remixing, and sediment-induced buoyancy effects on the turbulence field of HCMS in estuarine and coastal environments. The various formulations also provide the relevant scaling parameters of the processes which govern the dynamics and appearances of HCMS under a wide variety of conditions.
This model concept is validated through application to well-documented laboratory experiments and field measurements, i.e. flume experiments with sand to study the velocity profile in sediment-laden flow, consolidation and entrainment experiments with cohesive sediment in an annular flume, measurements in and upstream of the turbidity maximum in the Ems Estuary, and measurements during storm conditions in the Maasmond area, the entrance to the Port of Rotterdam. In general, the measurements and model simulations agree well.
The suspended sediment concentration profiles in general and the rapid settling around slack water in particular, as observed in the Ems River, can only be simulated properly if the effects of both flocculation and sediment-induced buoyancy are accounted for in the lDV POINT MODEL. These simulations also indicate that the floc size may vary by an order of magnitude over the tidal cycle; computed floc sizes and settling velocities agree reasonably with in situ measurements.
The measured time series of suspended sediment concentration at 0.15 and 0.55 m above the sea bed in the Maasmond area can only be simulated properly if the effects of hindered settling, sediment-induced buoyancy effects, salinity-induced buoyancy effects and augmented mixing by waves are all accounted for in the IDV POINT MODEL. A temporary fluid mud layer of about 1 dm appears to be formed on the North Sea sea bed around slack water, which is mixed rapidly during accelerating tide. The simulations also indicate that the flow in the Maasgeul, the access channel to the Port of Rotterdam, is not able to keep the available sediment in suspension.
From a series of sensitivity analyses it is concluded that under stormy conditions, auto-saturation can occur, i.e. the flow can erode so much sediment that it becomes super-saturated. This is probably the explanation for the occurrence of mud banks observed at several locations in the world.
This study revealed the important role of the aforementioned processes in the rapid siltation phenomena described in the first paragraph. Implementation of appropriate models of these processes in a full three-dimensional model is required to help the managing authorities to control their maintenance problems and to improve the economic exploitation of their port.