Micro Ramps in Supersonic Turbulent Boundary Layers: An experimental and numerical study

Micro Ramps in Supersonic Turbulent Boundary Layers: An experimental and numerical study

Sun, Z.

Scarano, F. (promotor)

Aerospace Engineering

Aerodynamics Wind Energy and Propulsion

2014-01-22

The micro vortex generator (MVG) is used extensively in low speed aerodynamic problems and is now extended into the supersonic flow regime to solve undesired flow features that are associated with shock wave boundary layer interactions (SWBLI) such as flow separation and associated unsteadiness of the interaction system. Numerous experimental and numerical studies have shown that despite their small size, such devices can alter the boundary layer properties very efficiently, when compared to the conventional vortex generators. In order to assist a more efficient design of MVGs, fundamental studies have been carried out to understand the associated wake properties such as the increased boundary layer mixing and the structure and stability of the induced vortex system. The present work is conducted in the framework of such fundamental studies. The micro ramp is among the most commonly used MVG devices and has been selected for the present investigations. The research is based both on wind tunnel experiments and numerical simulations in order to build a more comprehensive and detailed understanding of the flow behind a micro ramp immersed in a supersonic turbulent boundary layer. The choice of the experimental approach is justified by the fact that the incoming turbulent boundary layer exhibits a high Reynolds number (Re?=13,600), which makes it too challenging for extensive CFD investigation by using LES or DNS approaches. Variants of the micro ramp configuration as well as the attendant SWBLI can be studied efficiently by wind tunnel experiments adopting PIV as velocity field diagnostics. The use of numerical simulations by the implicit large eddy simulation (ILES) technique for one specific case enables the detailed inspection of the flow field that adds to the understanding of the flow development in regions or aspects where the experimental method provides limited access. Finally, there is general interest to know that till what extent numerical simulations can correctly identify the governing mechanisms of the boundary layer flow manipulation by micro ramps. Tomographic PIV is used as three-dimensional flow diagnostic technique in the investigation of flow organization in the micro ramp near wake (x/h?9~15). From the experimental data it is observed that the mean flow features a conical wake containing a pair of steady vortices aligned in streamwise direction. This is considered to be the basic mechanism of the boundary layer flow manipulation, whereas the wall-normal velocity component features a central focussed upwash with downwash motions at the sides. Simultaneously, a deficit region of streamwise velocity is produced in the center of the wake. The shear layer surrounding the wake is subject to Kelvin-Helmholtz (K-H) type instability and the instantaneous flow organization exhibits the formation of coherent K-H vortices that are arc shaped and dominate the velocity field fluctuations across the shear layer. Conditional averaging of the 3D velocity field yields the salient features of the interaction between the streamwise vortices and the K-H vortices whereas the former are found to be weakened at the generated of K-H vortices. The downstream decay of the flow features that are introduced by the micro ramp is relevant to its positioning with respect to the point of interaction between shock wave and boundary layer, indicating the relevance of investigating the further downstream development. Therefore experiments are conducted with large format PIV camera to study the decay in the center plane of the micro ramp far wake (x/h?12~32). In order to find a proper scaling parameter of the micro ramp wake, two geometrically similar micro ramps with different sizes are employed. Both streamwise and wall-normal velocity components exhibit a power-law decay in agreement with theories for the fully developed turbulent flow regime. The wall-normal velocity decays faster, approximately at a rate 2.5 times of the momentum deficit. The self-similarity of the velocity profiles is also examined. The streamwise velocity exhibits a good degree of self-similarity in the upper and lower shear layer, while the wall-normal component has overlapped upwash profiles. Concerning the turbulent properties, a strong anisotropy of velocity fluctuations is observed at upstream locations (x/h<20), nonetheless both fluctuation components decay to a similar magnitude when approaching the downstream end of the measurement domain (x/h>20). The organization of instantaneous vortical field is also investigated in the attempt to better understand their effect on the wake decay. Spatial auto-correlation of the instantaneous velocity fields yields the streamwise evolution of the average distance between vortices. Vortex pairing is identified in the range x/h=18~22 through an increase of such distance. The detection of counter-rotating vortices in the lower part of the wake suggests that the K-H vortices produced in the upper region of the shear layer propagate into the region close to the wall after vortex pairing, which eventually gives rise to ring-vortex formation in the later stage of the wake. A numerical study using ILES with high order scheme is carried out in collaboration with the University of Texas at Arlington. In order to establish a fair comparison with the experimental data, the flow conditions are made as similar as possible, matching the free stream Mach number and the ratio between micro ramp height and boundary layer thickness. The attendant limitations on computational resources limit the Reynolds number based on boundary layer momentum thickness to about one-third of that in the experiments. The comparison covers the most relevant quantities, such as the streamwise and wall-normal velocity and the peak vorticity. An overall good agreement is observed. A noticeable discrepancy involves underestimation of upwash motion: the wall-normal velocity amounts to 70% of the measured data. In the observation of instantaneous flow, vortex pairing is also identified and the spatial-temporal evolution of the K-H vortex is studied by tracking, which confirms the flow model conjectured from the planar PIV study in the center plane.

turbulence
flow control
Particle Image Velocimetry
Large Eddy Simulation
supersonic flow

https://doi.org/10.4233/uuid:6e1a39dd-1581-4d87-b623-97d1dc39fb78

9789461919243

Institutional Repository

doctoral thesis

(c) 2014 Sun, Z.