Computational results are presented which co-validate published experimental observations of steady streaks, deliberately generated by the steady wake of a wire placed upstream of a flat plate with a prescribed leading edge. The largest streak occurs when the wake is generated from a wire placed upstream of the wind tunnel contraction. Normal vorticity passing through the contraction leads to the creation of streamwise vorticity in the test-section via tilting and stretching. The computational results allow the original experiment to be reinterpreted as a receptivity experiment that demonstrates the boundary layer is more receptive to steady streamwise vorticity than normal vorticity. It also suggests an interesting mechanism for the generation of Klebanoff streaks in wind tunnels. The effect of shifting the attachment point at the leading edge on receptivity is also demonstrated. The streak growth is compared to the Optimal streak often used in computational studies. The modal growth of the streak generated by free-stream normal vorticity is found to have a streamwise location of peak energy close to the Optimal streak for wavelengths larger than the leading edge thickness. However, the location of the peak energy for the streamwise vorticity streak varies substantially with wavelength. Differences in wall-normal profiles are also noted.

The design of underwater vehicle propulsors is sensitive to the strength of the adverse pressure gradient (APG) in the stern region. In order to assess this sensitivity, boundary layer properties for different stern shapes derived from the unappended BB2 generic underwater vehicle geometry, an axisymmetric body of revolution, were computed using four different computational fluid dynamics (CFD) Reynolds-Averaged Navier-Stokes (RANS) based turbulence models. It was found that as the stern became steeper (larger APG), there was an increased difference between the boundary layer prediction from the four turbulence models. Initially, boundary layer parameters for the unappended BB2 geometry were validated against experimental measurements from the Defence Science and Technology Group low speed wind tunnel for a Reynolds number of approximately four million with zero angle of incidence. It was found that the Baseline Reynolds stress model (BSL-RSM) closely matched the experimental skin friction and boundary layer profiles at all measurement locations, notably in the stern region, which is a significant improvement from previous studies.

Defining a link between wind-tunnel settling chamber screens, flow quality and test section boundary-layer spanwise variation is necessary for accurate transition prediction. The aim of this work is to begin establishing this link. The computed, steady, laminar wake of a zither (screen model) with imperfect wire spacing is tracked through a contraction and into a model test section. The contraction converts the zither wake into streamwise vorticity which then creates spanwise variation (streaks) in the test-section boundary layer. The magnitude of the spanwise variation is sensitive to the zither open-area ratio and imperfections, but the observed wavelength is relatively insensitive to the zither wire spacing. Increased spanwise variation is attributed to large wavelength variation of drag across the zither, and not the coalescence of jets phenomena. The linear stability of the streaks is predicted using the parabolized stability equations with the e N method. A standard deviation of zither wire position error of 38.1 µm (15 % of wire diameter) for a zither of 50 % open-area ratio is found to suppress Tollmien-Schlichting wave growth significantly.

No abstract

No abstract