Flow reattachment on the roof of a 6 m cube

Five ultrasonic anemometers are used to measure flow velocities above the roof of a 6 m cube and at a reference point upstream. Various analysis techniques are applied to the data in order to illustrate the differences between: the mean reattachment position when the mean wind is normal to the windward face (0°), which is at x/h  ≈0.6; the median instantaneous reattachment position at x/h=0.66x/h=0.66 and the reattachment position that would occur if the wind direction was held at a constant 0°, which is at x/h=0.75x/h=0.75 and is also the position of zero conditionally averaged u velocity at instants when the v velocity is zero. It is also shown that the flow is highly unsteady and that the reattachment length varies from negligible separation to no reattachment, which occurs for 20% of the time. Some of these variations are related to fluctuations in the onset wind speed and direction, but they are also influenced by the dynamic response of the separation vortex system. The formation and shedding of vortices means that certain frequencies in the turbulence spectrum, around 1 Hz, are slightly amplified, whereas frequencies above 10 Hz are filtered out as a result of the inertia of the vortex system. The effects of reattachment length on the pressure distribution are briefly considered but it is shown that these do not account for the differences between the Silsoe field data and typical wind-tunnel results. It is suggested that the pressure differences may be related to Reynolds number, but it appears that this is not associated with changes in reattachment length.