Flow control for oblique shock wave reflections

Shock wave-boundary layer interactions are prevalent in many aerospace applications that involve transonic or supersonic flows. Such interactions may lead to boundary layer separation, flow unsteadiness and substantial losses in the total pressure. Flow control techniques can help to mitigate these adverse effects and stabilize the interaction. This thesis focuses on passive flow control techniques for oblique shock wave reflections on flat plates and presents experimental results for both laminar (part I) and turbulent interactions (part II). Particle image velocimetry (PIV) measurements were used as the main flow diagnostics tool throughout this thesis, where especially the laminar case proved to be challenging due to its very small boundary layer thickness of ∼0.2 mm.
Laminar boundary layers are extremely prone to separation and long separation bubbles (∼50δ99) were recorded even for relatively weak shock waves (p3/p1∼1.2). The bubble has a flat / triangular shape and extends mostly upstreamof the incident shockwave. The incoming boundary layer is lifted over the bubble and remains in an apparent quasilaminar state up to the incident shockwave, afterwhich the boundary layer quickly transition into a turbulent state (30-40δ99). Only for very weak shock waves it was found that the boundary layer can remain laminar up to reattachment.
The separation bubble for laminar interactions can be removed by enforcing boundary layer transition a short distance upstream of the interaction. Transition strips that introduce three-dimensional features in the flow were found to be more effective at this task than purely two-dimensional trips (e.g. a step) and could therefore be placed closer to the interaction while still maintaining their effectiveness. Forced boundary layer transition, however, comes at the price of having a substantially thicker (∼50%) turbulent boundary layer downstream of the interaction, which is the result of losses at the trip, a larger portion of turbulent flow and higher shock-induced losses. It therefore appears that there is no added value to tripping the boundary layer for laminar flat plate interactions, especially given the fact that the untripped laminar interaction shows no signs of any large-scale type of unsteadiness.
For the turbulent interactions, micro-ramp vortex generators were studied as flow control devices. Micro-ramps transport high-momentum fluid towards the near-wall region of the flow by the action of streamwise vortices, thus creating a fuller boundary layer profile that is less prone to separation. A net transport of streamwise momentum has been observed up to 5-7δ99 downstream of the micro-ramp, after which a plateau level is reached in which, on average, no momentumis added or removed fromthe nearwall region of the flow. Consequently, a similar distance between the trailing edge of the micro-ramp and the onset of the interaction is required to ensure a maximumreduction in separation bubble size and shock unsteadiness. The application of micro-ramps leads to a spanwise modulation of the separation bubble, with the micro-ramp being most effective along its centreline. The control effectiveness of the micro-ramp is virtually independent of the Reynolds number and is slightly reduced for higherMach numbers.