Shedding light on the bluff body wake instability

Abstract: The flow around a vehicle, for example, results in a large resistance and can generate a complex swirl pattern behind the body of the vehicle, which can lead to strong alternating lateral forces. There is therefore a clear need for more fundamental understanding of the instability of this so-called wake region, which arises behind bluff bodies. In this thesis, the focus lies on the relation between the wake inlet conditions and the wake characteristics of a bluff body. This relation is studied experimentally, in the Reynolds number range 2.9E3 - 5.5E4.  The experiments are carried out in a specially designed test-section, which comprises a rectangular forebody with a smooth leading edge and a blunt trailing edge. The perforated top and bottom surface of this body allow for boundary layer modification by means of wall suction. Hot-wire anemometry and pressure measurements show that wall suction, in the order of 1% of the free-stream velocity, leads to a significant change of the boundary-layer properties; the boundary-layer thickness decreases and the wall-shear stress increases.  Laminar boundary layers take the form of the asymptotic suction boundary layer and turbulent boundary layers are shown to relaminarize, when subject to wall suction above a critical value. A modification of the boundary layer leads to an increase of the vortex shedding frequency and a decrease of the base pressure. Empirical relations for the Strouhal number and the base pressure coefficient are derived. The boundary layer conditions are included in these relations and as result, they hold for both laminar and turbulent boundary layers.Two-dimensional velocity fields, obtained by means of particle image velocimetry, reveal the effect of boundary-layer modification on the near-wake topology. By dividing the velocity component into (i) mean, (ii) periodic and (iii) random contributions, it can be shown that the confluence point moves upstream, the curvature of streamlines enclosing the trailing edge increases, and the Reynolds normal and shear stresses increase. The experimental study is continued by introducing interference elements in the wake. The qualitative effect of a splitter plate on the wake characteristics is not altered by boundary-layer modification, but the critical splitter plate length and gap width for which the vortex shedding frequency is suppressed do change. A sensitivity map is obtained experimentally by placing a control rod at various locations in the wake. Considering all limitations of such an experiment, it was found to be in good agreement with the result of a linear stability analysis carried out on the measured mean velocity field.