Interaction of coherent eddies with headland wake flows

Melanie J. O'Byrne 1 , Ross W. Griffiths 1 , Jason H. Middleton 2

1 Research School of Earth Sciences, Australian National University, Canberra, ACT 0200, Australia
2 School of Mathematics, University of New South Wales, Sydney, NSW 2052, Australia

Flow around islands and headlands can trap suspended particles, bring nutrients up from deeper waters, influence the distribution of sediments and provide favourable environments for marine biota.

As few islands or headlands are isolated bodies, it is important to understand how unsteady incident flow affects wake structure and behaviour.

In 2006 we continued laboratory investigations into the effects that perturbations in the oncoming flow have on the wake of an idealised headland, using ultrasonic Doppler techniques to measure the velocity field.

We have completed a series of experiments in a one-metre diameter rotating cylindrical tank. The headland, seafloor (base) and coast (sidewall) were impulsively set in motion relative to the still water. Thus, in a headland reference frame, there is oncoming flow.

We find that for Reynolds numbers above approximately 1000, coherent incident eddies ‘lock-on' to shedding in the headland wake, resulting in a train of vortex dipoles downstream of the obstacle. This finding is potentially relevant to engineering applications such as design of offshore platforms, power cables or bridge pylons.

Rotating experiments also reveal an interesting phenomenon analogous to classic amplitude modulation. For some Reynolds numbers between 800 and 1500, the wake from an upstream disturbance acts as a high-frequency carrier signal, modulating the amplitude of the signal from the headland wake.

We aim to identify flow regimes and geometries where these behaviours may be observed, how they are triggered and thus whether they may be predicted.

The limited flow time in a single revolution of the rotating table results in a transient headland wake, as shown in Figure 10. These still images, from two experiments with a Reynolds number of 1000, compare the unperturbed or natural flow (left column) with the perturbed flow (right column) around an idealised headland. The recirculating headland wakes look similar in both cases nine seconds after impulsively starting rotation. Twenty-four seconds later, the perturbed wake is considerably mixed and coherent eddies from upstream are clearly visible, while the natural headland wake is clear. Thereafter, the wake structure is lost regardless of upstream conditions, in part due to the presence of the curved wall and its thickening boundary layer.

In order to ensure an established wake flow, we have begun studies using a new 3-metre flume in which the working fluid continually recycles through a series of pumps. This permits a typical flow time of 20 minutes. Preliminary results at low Reynolds numbers reveal that incident disturbances can significantly extend the length of the wake bubble. In future experiments, we will study the residence time and amount of leakage of fluid trapped in the headland wake.

Figure 10. Still images of two experiments at a headland Reynolds number of 1000, without (left) and with (right) oncoming flow perturbations, reveal the transient wake structure in our experiments with an impulsively started flow. The idealised headland and coast (a vertical wall) are at the top of each image.