Headland wake flows and the effects of an upstream disturbance
Date
2010
Authors
O'Byrne, Melanie Jane
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This thesis presents laboratory experiments examining the dynamics of headland wakes with and without an incident disturbance in the oncoming flow. In particular, we explore the effects on the headland wake structure of the wake of another obstacle located upstream of the headland, and the extent to which incident variability influences wake shape and structure. Flow around islands, headlands and reefs is known to bring up nutrients from deeper waters, influence the distribution of sediments and pollutants, and provide favourable environments for marine life. This thesis was motivated in part by the lack of laboratory studies on headland wake flows. For both island and headland wake flows, previous experimental studies have neglected the high levels of turbulence inherent in environmental flows. As complex flows in nature usually include eddies and turbulence generated upstream of a region of interest such as coastal topography, the influence of incident disturbances on topographic wakes in shallow flow must be understood. The focus is on an idealised headland as a vertical half-cylinder against a vertical wall. The incident disturbance has been produced by a small vertical circular cylinder. Using flow visualisation and ultrasonic Doppler velocity measurements, the three-dimensional flow field has been examined in a circular towing tank, impulsively started flow in a circular rotating tank, and continuous flow in a channel over a range of Reynolds numbers and two aspect ratios. The primary set of results were obtained in a laboratory flume in which flow passed a stationary headland, with or without an incident disturbance, thus including the effects of bottom and side-wall boundary layers. In the other two classes of experiment, the headland was towed through stationary water so that there were initially no bottom or side-wall boundary layers; and a circular tank with the headland fixed to the wall was impulsively started from rest and rotated at a fixed angular velocity through a single revolution. The latter gave a completely uniform oncoming flow outside temporally developing boundary layers. Experiments were carried out at Reynolds numbers (based on the headland radius) ranging from 300 to 6000, with the majority of experiments having Reynolds numbers between 350 and 1600. This range typically characterises coastal headland flows, where the Reynolds number is based on an eddy viscosity rather than molecular viscosity in laboratory experiments. The very small aspect ratio of water depth to headland length scale characteristic of coastal oceanic flows ({u039B} {u226A} 1) was not accessible in the laboratory; the headland radius in this thesis was of the order of the water depth. In order to characterise the incident upstream disturbance, the wake of a vertical circular cylinder was first studied. This body was chosen for its simple geometry and the large volume of previous published work. At Reynolds numbers below 150, the cylinder produced a single dominant frequency in its wake. At Reynolds numbers over 300, velocity fluctuation frequencies in the cylinder wake varied as a function of cross-stream and downstream position. A suite of experiments exploring the relationship between the dimensionless Strouhal frequency and Reynolds number found the Strouhal number to be consistently 20-25% smaller, for a given Reynolds number, than the empirical curves in Roshko (1954a). The differences have been attributed to the much smaller aspect ratio (i.e. shallow flow) in this study and the level of freestream fluctuations in the flume. For the isolated headland with no intended incident disturbance, key results are that the Strouhal number is not a useful single parameter to characterise velocity fluctuations in the wake, as the dimensionless frequency varies with cross-stream and downstream measurement position. When shear-layer width becomes comparable to the headland radius, at a location three to four radii downstream of the headland, the wake is fully developed and the Strouhal number remains unchanged with distance further downstream. This study therefore suggests that headland wakes are best characterised by the largest, lowest-frequency structure that persists beyond three headland radii. A further important result is that the modified wake parameter P, introduced as an alternative to the Reynolds number to reflect the relative importance of bottom friction over lateral stresses, is a better parameter for characterising the vertical structure (in particular) of the flow than the Reynolds number typically cited for wake studies. This is due to the fact that P includes the influence of shallowness and aspect ratio on the wake. The experiments cover a range of P similar to those in coastal flows. Finally, the structure and behaviour of the headland wake when an incident disturbance is present has been investigated. The introduction of an upstream disturbance plays a crucial role in the downstream evolution of the headland wake. The magnitude and intensity of fluctuations is greater with an incident disturbance than without. The peak fluctuation amplitude occurs closer to the headland when there is an incident disturbance than for the isolated headland under the same flow conditions. Further downstream, velocity fluctuations in the wake decay in magnitude more rapidly as a result of the incident disturbance. Flow visualisation and fluctuations of both horizontal and vertical velocities all indicate that coherence in the far wake is destroyed by a coherent frequency advected from upstream of the headland. Thus we find substantial differences between the essentially three-dimensional mean flow and its fluctuations when a wake flow with coherent structures is introduced upstream of the headland. An overarching observation is that in the flows studied here, which would commonly be treated as 'shallow-water' flows, the depth-dependence of the mean flow and the three-dimensionality of the flow in the wake are dominant characteristics that cannot be neglected. Although we have concentrated on the case with a vertical coastal boundary and vertical headland topography, and the study is restricted to aspect ratios of order one, our findings have implications in the ocean, atmosphere and built environments.
Description
Keywords
Citation
Collections
Source
Type
Thesis (PhD)
Book Title
Entity type
Access Statement
Open Access
License Rights
Restricted until
Downloads
File
Description