This thesis focuses on the control of shock wave / boundary layer interactions (SWBLIs) in supersonic inlets. The overall aim of this study is to determine to what extent vortex generators (VGs) can mitigate flow separations within supersonic inlets. To achieve this, an experimental investigation was undertaken in a small-scale wind tunnel, because small-scale wind tunnels are much more amenable to numerous measurement techniques than real inlets.

A new geometry was designed and developed as part of this study to make the configuration more representative of typical inlet conditions than previous studies. The flow-field that was determined to be relevant, yet still simple, comprised of a Mach 1.4 normal shock followed by a region of subsonic diffusion created by a 6° straight-angled diffuser. A shock holder was used to improve shock stability and the Reynolds number was set to 25 million per metre throughout.

This flow was examined in three different shock positions. The flow was found to be highly sensitive to the relative position of the shock and diffuser: when the shock was positioned somewhat upstream of the diffuser the flow was relatively benign and, apart from small corner separations, the majority of the boundary layers remained attached; but as the shock was moved close to the diffuser, separation was introduced on the channel floor; and once inside the diffuser the entire diffuser boundary layer was separated.

In the first instance, VGs were employed on the channel floor. While the VGs were able to produce a thin attached region on the channel floor and therefore improve the centre-span region somewhat, they were detrimental to the corner separations. As a result, corner suction was employed to reduce the prominence of the corner flows. Corner suction dramatically reduced the corner separations, however, flow separation still dominated the diffuser because separation was now introduced at centre-span.

These results demonstrate that a strong coupling between the centre-span and corner flow regions exists in this configuration. When flow control was applied to one region, although the flow is improved locally, increased losses in other regions tended to offset this gain. This is because an overall improvement in the pressure within the diffuser can only be maintained if all areas can sustain the pressure-rise. As a result, only when all the problem regions were appropriately controlled—the corners using suction and the centre-span using VGs—could a notable improvement in the flow be obtained. In this combined configuration, more than 50% of the diffuser-span remained attached throughout the diffuser and there was a 15% drop in stagnation pressure losses and 6% increase in the wall-pressure recovery.

Although further VG studies are required, this investigation does suggest that VGs do have the potential to alleviate the current dependency on boundary-layer bleed for flow control in supersonic inlets.