Study on the Formation and Evolution of Vortex Ring Structure behind MVG with Turbulent Inflow

Transcription

1 Study on the Formation and Evolution of Vortex Ring Structure behind MVG with Turbulent Inflow Yonghua Yan Chaoqun Liu Technical Report

2 Study on the Formation and Evolution of Vortex Ring Structure behind MVG with Turbulent Inflow Yonghua Yan 1 and Chaoqun Liu 2 University of Texas at Arlington, Arlington, Texas, Abstract In this paper, high order LES has been conducted for investigation on the flow around micro vortex generator (MVG) at Mach number 2.5 and Reθ=5760. The mechanism of vortex ring generation behind MVG, especially in the initial stage of the ring generation has been studied in detail. The results show that the ring structures come from both the sheer layer on the upper boundary of momentum deficit and also from the sheer layer in the lower boundary layer. The complicated interaction among vortex structures has also been studied. Nomenclature MVG = micro ramp vortex generator M = Mach number Re θ = Reynolds number based on momentum thickness h = micro ramp height δ = incompressible boundary-layer nominal thickness x, y, z = spanwise, normal and streamwise coordinate axes u,v,w = spanwise, normal and streamwise velocity LES = large eddy simulation Subscript 0 = inlet w = wall = free stream I. Introduction In the supersonic ramp jets, shock boundary layer interaction (SBLI) can significantly reduce the quality of the flow field by triggering large-scale separation, causing total pressure loss, making the flow unsteady and distorting. Micro vortex generators (VG) are a kind of low-profile passive control device designed for the boundary layer control. In contrast to the conventional VG (widely used in aviation applications and with height (h) of the order of the boundary-layer (δ)), micro VG has a height approximately 20-40% (more or less) of the boundary layer. Among these micro VGs, mircoramp vortex generators (MVG) are given special interest by engineers because of their structural robustness. MVG generate a pair of streamwise vortex, which remains in the in the boundary layer for relatively long distance, and the corresponding down-wash effect will bring momentum exchange to the boundary layer which makes it less liable to be separated. During such process, a specific phenomenon called momentum deficit, i.e., a cylindrical region consisted of low speed flows, will be formed after the MVG 1. In Lin s review 2 on the low-profile vortex generator, it was mentioned that a device like MVG could alleviate the flow distortion in compact ducts to some extent and control boundary layer separation due to the adverse pressure gradients. Similar comments were made in the review by Ashill et al 3. The formal and systematic studies about the micro VGs including micro ramp VG can be found in the paper of 1 PhD student 2 Professor and AIAA associate fellow. 1

3 Anderson et al 4. Micro-actuators including MVG were found to have comparable effects as the boundary layer bleeding technique, and so MVG was considered to be very practical for the flow control in supersonic inlets. A series of experimental and computational investigations have been carried out since then. It is reported that Babinsky 1,5-7 did the very prominent experimental studies. He made a series of experiments on different kinds of mirco VGs and investigated their control effects in detail. The mechanism of MVG flow control from his work concludes that a pair of counter-rotating primary streamwise vortices is generated by MVG, which are mainly located within the boundary layer and travel downstream for a considerable distance. Secondary vortices are located underneath the primary ones and even more streamwsie vortices could be generated under suitable conditions. Streamwise vortices inside the boundary later bring low momentum fluid up from the bottom and high momentum fluid down to the boundary layer. A striking circular momentum deficit region is observed in the wake behind the MVG. The vortices keep lifting up slowly, which is thought to be the consequence of the upwash effect of the vortices. Numerical simulations have been made on MVG for comparative study and further design purposes. Ghosh, Choi and Edwards 8 made detailed computations under the experimental conditions given by Babinsky. These numerical studies include RANS computations and hybrid RANS/LES computations using immersed boundary (IB) techniques. The fundamental structures, like the streamwise vortices and momentum deficit, were reproduced by the computation. Lee et al 9 also made computations on the micro VGs problems by using Monotone Integrated Large Eddy Simulations (MILES). In their computation, the MVG is placed in a domain with the configuration following the real wind tunnel. The fundamental wave system of the MVG were reproduced in the computation, which consists of the main shock, expansion waves and re-compression shock like that reported by Babinsky 7. The momentum deficit was captured too. They 10 further tested several variations of the standard MVG and micro vane to enhance the control effect. Although there are experiments and computations on MVG problems, only two-dimensional structural information was available and confirmed by experiment. The mechanism of the 3D vortex structure behind MVG is still unclear. In this study, we try to understand the mechanism of the flow structure expecially the vortex structure behind the MVG. Numerical simulations are made on supersonic ramp flow with MVG control at M=2.5 and Reθ=5760. The trailing edge declining angle of the MVG is 70 in computation. In order to make simulations, a kind of large eddy simulation method is used by solving the unfiltered form of the Navier-Stokes equations (NSEs) with the 5th order bandwidth-optimized WENO scheme, which is generally referred to the so-called implicitly implemented LES. Without explicitly using the subgrid scale (SGS) model as the explicit LES, the implicitly implemented LES uses the intrinsic dissipation of the numerical method to dissipate the turbulent energy accumulated at the unresolved scales with high wave numbers. There are two main subfields about this category, i.e., the MILES by Boris, Fureby and Grinstein, et al, and the implicit LES by Visbal, Rizzetta and Gaitonde, et al. The first subfield is based on modified equation analysis, and typically uses the high order monotone scheme like flux-corrected transport (FCT) scheme or piecewise parabolic method (PPM). The ENO algorithm was also reported being used as the limiter in Ref. 40. This kind of method can be used to solve the supersonic problems with shock waves, but the order of the scheme should not be competitive to the modern high order schemes like the compact schemes or WENO schemes with 5th order of accuracy or higher. The second one 41 specifically uses the high order compact scheme by Lele and the high order Pade-type low-pass spatial filter. However, the published applications of the method are only for the low speed flow. When the same numerical algorithms were used on supersonic problems 43-44, the Smagorinsky dynamic SGS model was incorporated in the simulation, which implies the existence of issues related to the numerical stability. A series of shock-capturing schemes were also tried for large eddy simulation 45-46, including the WENO scheme. As mentioned in Ref. 46, at low Mach number the investigated compact differencing and filter scheme formulation may give better results but as the Mach number increases the relative suitability of the ENO method increases. However, the ENO scheme still produces numerical turbulence thus stabilizing filters is need, while the WENO scheme does not need filtering. Recently, an 2

4 evaluating computation was reported on circular cylinder flow using implicitly implemented LES by the 5th WENO scheme 47. Comparisons were made between the computation and the experiment. The results show that the numerical algorithm is feasible and efficient. For the studied supersonic MVG controlled ramp flow problem, there are complex shock wave system, strong shock-vortex interaction and small scale structures. Considering the above status of implicitly implemented LES, the method by solving the Navier-Stokes Eqs with the 5th order bandwidth-optimized WENO scheme is used in the paper and considered as certain implicitly implemented LES. In our previous paper 48, the flow field around the MVG and surrounding areas has been studied in details. Further more, 3-D structure of the shocks is also obtained by our numerical simulation 48. The hairpin vortex was formed and then travelled downstream. According to the analyses, a dynamic vortex model can be given in Fig.1 (half domain). The dominant vortex near the MVG is the primary vortex; underneath there are two first secondary counter-rotating vortices, which later leave the body surface and become fully 3D separations by the way of spiral points in body surface. These vortices will Figure 1. The dynamic vortex model (Li and Liu 2 ) merge into the primary vortex propagating downstream, while new secondary vortex will be generated under the primary vortex. This dynamic vortex model is mostly confirmed by the experiment work of Mohd R. Saad 49 et al in recent (Fig. 2). Figure 2. Surface flow visualisation image and the vortex model given by Mohd R. Saad[] et al 3

5 After the MVG, a strong momentum deficit was found behind MVG which causes a strong circular shear layer 1,48, as shown in Fig. 3. The result is in consistency with the referenced computations and experiments 8,10,48. For clarity, the typical structure of the deficit is provided in Fig. 3 as well with spanwise streamlines. Inside the deficit area, there are two counter-rotating primary vortices which are illustrated in Fig. 3. In adjacent MVG region, the shape of deficit appears to be a circle, and usually has a root connected to the boundary layer. At underneath of the circle, there are two streamwise high velocity regions. Fig. 4 gives a qualitative comparison with experiment (Babinsky et al 1, 2009) in the time and spanwise averaged velocity profile behind MVG. Qualitative agreement is achieved. Figure 3. The momentum deficit (a) Averaged velocity profile behind MVG (b) Averaged velocity profile by Babinsky et al (2009) Figure 4. Qualitative comparison of averaged velocity profile behind MVG with experiment To reveal the coherent structure of the flow, the iso-surface of λ 2 scalar field is given in Fig. 5. It is very clear that there is a chain of vortex rings, starting from behind of the trailing-edge of MVG. The rings are generated almost erectly at first and then be continuously distorted and enlarged while 4

6 propagating downstream. These rings could be a dominant factor of the mechanisms of MVG in control of shock boundary layer interaction. (a) global view (b) close-up view behind the MVG Figure 5. Vortex rings shown by iso-surface of λ 2 Fig. 6 demonstrates the instantaneous numerical schlieren at the central plane. We can see many vortex rings appear in circular shapes. Informed with the prediction of vortex rings, the experimentalists in UT Arlington used some technology to validate the discovery. They used the particle image velocimetry (PIV) and the acetone vapor screen visualization to track the movement of the flow. More specifically, the flash of a laser sheet is used to provide the light exposure at a time interval of micro seconds. Fig. 7 presents a typical image at the center plane using PIV and the acetone vapor technology(lu et al 52 ). It is clearly demonstrated that a chain of vortex rings exists in the flow field after the MVG, same as shown in LES results (see Fig. 6). Figure 6. The numerical shilieren at the center plane 5

7 a) Using PIV b) Using the acetone vapor Figure 7. The laser-sheet flash image at the center plane (Lu et al 2010) Our numerical discoveries of the vortex ring structures are also confirmed by 3-D PIV experiment (Fig. 8) conducted by Sun et al at Delft University 53. Compared the two results, we can find the similar distribution of streamwise ( ω z ) and spanwise vorticity ( ω x ) components, which also proves the existence of ring structures. The Kevin-Helmholtz vortices part in Fig. 9 corresponds with the ring head in Fig. 5. The underneath part which is illustrated as streamwise vortices are two counter rotating primary vortices which are considered to be the main source of the ring structure as explained later. Figure 8. K-H rings behind MVG by (Sun et al 2011) Figure 9. Distribution of Kelvin-Helmholtz vortices and streamwise vortices from LES II. Numerical methods, grid generation and turbulent inflow It is definitely necessary to find physical mechanisms of MVG for design engineers. RANS, DES, RANS/DES, RANS/LES, etc are good engineering tools, but they may not be able to reveal the mechanism and get deep understanding of MVG. We need high order DNS/LES. A powerful tool is the integration of high order LES. In this paper, an approach called monotone integrated LES (MILES) 9,40 6

8 was adopted at Mach number 2.5 and Re θ =5760, in which the numerical dissipation is used as the sub-grid stress model. Flows around MVGs are studied with back edge declining angle 70 o (see Fig. 10).The geometries for the cases are shown in Fig 11. (in which δ 0 represents the incompressible boundary layer nominal thickness). A general grid partition technique is used in this grid generation. According to experiments by Babinsky 1, the ratio h/δ 0 of the models range from 0.3 to 1. The appropriate distance from the trailing-edge to the control area is around 19~56h or 8~19δ 0. In this study, the height of MVG h is assumed to be δ 0 /2 and the horizontal distance from the apex of MVG to the ramp corner is set to be 19.5h or 9.75δ 0. The distance from the end of the ramp to the apex is h. The distance from the starting point of the domain to the apex of MVG is h. The height of the domain is from 10h to 15h and the width of the half domain is 3.75h. As shown in Fig. 6, three regions are divided as: the ramp region, MVG region and fore-region. Between each two regions, there is a grid transition buffer. Because of the symmetry of the grid distribution, only half of the grids need to be generated. The grid number for the whole system is: n spanwise n normal n streamwise = Using the inflow flow profile described in the next section, a data summary is given in table 1 about the geometric parameters of the grid system. Fore-region Grid transition MVG Region Grid transition Ramp Region h Figure 10. The sketch of MVG at β =70 Figure 11. The schematic of the half grid system L x L y L z x + y + z δ δ δ Table 1. The geometric parameters for the computation The details about the geometric objects, grid generation, computational domain, etc, which are introduced in our previous paper, 48, 50 will not be repeated here. The adiabatic, zero-gradient of pressure and non-slipping conditions are adopted at the wall. To avoid possible wave reflection, the non-reflecting boundary conditions are used on the upper boundary. The boundary conditions at the front and back boundary surfaces in the spanwise direction are treated as the periodic condition, which is under the consideration that the problem is about the flow around MVG arrays and only one MVG is simulated. The outflow boundary conditions are specified as a kind of characteristic-based condition, which can handle the outgoing flow without reflection 51. New fully developed turbulent inflow boundary conditions are generated in this paper using following steps: a) A turbulent mean profile is obtained from previous DNS simulation results 8 for the streamwise velocity (w-velocity) and the distribution is scaled using the local displacement thickness and free stream velocity. The basic transfer is based on the assumption that the same distribution exists between the relations of 7

9 * U / U e ~ y / δ. And the averaged streamwise velocity of MVG case can be reached by interpolation (3rd Spline interpolation); b) The pressure is uniform at inlet and has the same value as the free stream value. The temperature profile is obtained using Walz s equation for the adiabatic wall: first, the adiabatic wall temperature is 2 determined using: Tw Te ( 1 r( 1) 2 M e ) = + γ, where the subscript e means the edge of the boundary layer and r is the recovery factor with value 0.9; next, the temperature profile is obtained by Walz s equation: 2 T T 1 2 ( ) 2 e = Tw Te r( γ ) M e U U ; e c) The fluctuation components of the velocity are separated from the total velocity at every instantaneous data file (totally 20,000 files). And such fluctuations are rescaled in the same way. 2 Because T T T T r ( 1) 2 M ( U U ) 2 and e = γ, considering the non-dimensional form and ignore the T e e w e e e 2 U, we get dt = r ( γ ) M UdU, or ( ) ρ T by = ; ρ T 1 e T = r γ M U U. Density fluctuation is determined 2 1 e ρt d) Finally, the transformed parameters are u = U + u, v = V + v, w = w, ρ = ρ + ρ, p =, γ 2 M T = T + T. To check the flow properties before the MVG, we analyzed the relevant flow parameters on a spanwise cross section which is illustrated in Fig 12. The cross section is 11.97h ahead the apex of MVG. As a result, the displacement thickness δ * = 0.371h, the momentum thickness θ = 0.275h, nominal boundary layer thickness δ = 2.36h. Thus, we can obtain a shape factor H as about 1.35, which shows the flow before the MVG is fully developed turbulence flow. Figure 12. The spanwise cross section on which the flow parameters are checked Fig. 13 shows the inflow boundary layer velocity profile in log-coordinates on the same cross section. There is a well-defined log region and the agreement with the analytical profile is well established. These results are typical for a naturally grown turbulent boundary layer in equilibrium. Figure 13. Inflow boundary-layer profile comparison with Guarini et al s 8

10 III. Analysis of the structure of vortex rings The vorticity component which revolves towards the vertical direction( ω in our case) is not shown in Fig. 12 and 13. If this missing part was provided, we can see the vortex ring structure clearly by the combination of all the components of vorticity as shown in Fig. 14. According to further analysis, we find that the ring structure is mainly composed by ω and ω while the streamwise vorticity component ω z is absolutely the major source of two counter-rotating primary vortex inside the ring structure. constitutes the head and part of the bottom of the rings and x y ω y forms the two sides. y ω x Figure 14. Vortex rings shown by the components of vorticity Figure 15. Vortex rings shown by the components of vorticity and λ 2 Compared with the ring structure shown by λ 2 method in Fig 15, we can find that both methods can capture the structure very well and the results are in accordance with each other. However, by using λ 2 method, the streamwise vortices are also illustrated which make the ring structure to be more complicated. In Fig 16, only two components of vorticity, ω and ω, are used to show the real structure of vortex x rings. It can be found clearly that the rings are not perfect and are not closed at the bottom, especially in their initial phase of evolution. In the initial stage, only part of the ring structure can be illustrated, and the y 9

11 vortex structure seems to break down. However, vortex will never break down in the continuous field for viscous flow. By plotting the vortex lines inside one of the ring vortex in Fig 17, it shows the ring part is only the concentration of the vortex lines which finally scatter into the flow field. When the vortex lines become scattered, the corresponding vorticity distribution becomes more even and the vortex structure thus disappears. (a) global view (b) close-up view (only half of the domain is illustrated) Figure 16. Vortex rings shown by ω x and ω y (only positive values are used) 10

12 Figure 17. Vortex lines in one of the vortex rings Fig 17 also shows the spanwise vorticity distribution on the central plane. On the upper bound of the momentum deficit area, the positions of peak areas (red spots) corresponds to the positions of 3D ring head from Fig 16 which shows spanwise vorticity reaches large value on the ring heads. Actually, each vortex ring is a highly rotating vortex tube. However, on the lower bound of the momentum deficit, there are no those peak areas to be found which also confirms that the rings are not closed. To analyze the origin of these vortex rings, the vortex lines in the first two distinguishable rings (which are marked in Fig 16(b)) are given in Fig 18 and 19, while the vortex lines in the third ring is given already in Fig 17. Figure 18. Vortex lines in the first vortex ring 11

13 Figure 19. Vortex lines in the second vortex ring In Fig 18, it shows the origin of the first vortex ring mainly comes from the vortex lines of the two sides nearly in the same spanwise plane of the ring. In Fig 17, the corresponding vortex lines of the third ring mainly come from the upstream boundary layer on both the surface of the ramp and the MVG. However, the vortex lines for the second ring seem to be more complicated, it contains both the situations of the first and third rings. To explain the different modes of vortex line distribution, which are all concentrated at the ring head and form the vortex tube in the initial stage of the whole ring structure behind MVG, we need to find out the origin of these rings. Fig 20 shows the averaged streamwise velocity distribution at different spanwise sections behind the MVG, through which we can find the existence of high sheer layer at the upper bound of the momentum deficit area. By calculating the second derivative of the velocity profile at the first section (as shown in Fig 21), an inflation point is clearly found at the upper bound of the momentum deficit. It is obvious that the existence of the inflection point in shear layer will cause the flow losing the stability and generate vortex rollers by the Kelvin-Helmholtz instability (K-H) theory. So the mechanism for the vortex ring generation may be a result caused by the K-H type instability, and the lost of the stability of the shear layer results in the roll-up of the vortex rings. In our previous paper 54 a stability analysis is made upon this velocity profile including the influence of viscosity of the fluid. The result still shows that the velocity profile is unstable. In Fig. 22, the 3D distribution of three components of vorticity is given right after the MVG. It shows that spanwise vorticity is first generated at the top of momentum deficit area where clusters the streamwise vortices. After that, the ring sides are also generated on the boundary of the momentum deficit. This confirms our assumption above. The top boundary of the momentum deficit has the most instable shear layer, so the spanwise vorticity is supposed to be new generated and forms the ring structure in the end. A ring shows up soon due to the evolution of the unstable shear layer. Also we can find that the whole ring is generated on the boundary of the momentum deficit, not from inside nor outside region. So, the shear layer instability is the main mechanism of the vortex generation for the first ring. As a result of the sheer layer instability, the vortex lines for the first ring must come from the two sides due to the rolling up of the spatial vortex lines made by the sheer layer. 12

14 Figure 20. Averaged streamwise velocity at different sections Figure 21. Inflection points (surface for 3-D) Figure 22. The instantaneous vorticity distribution 13

15 t 1 t 2 t 3 Figure 23. Spanwise vorticity distribution at three continuous times on the central plane (left) and 3D space (right) 14

16 Fig 23 gives the spanwise vorticity distribution at three continuous times on the central plane and in 3D space. It shows that the third vortex ring in Fig 16 is mainly originated at the lower boundary layer and lift up to the upper bound of the momentum deficit, which finally forms a vortex ring and be part of the ring structure. During the process, the vortex lines which are contained in the third ring is stretched and lift up as well gradually, and they finally forms the lines shown in Fig 17. In our previous paper 54, we come to the conclusion that the ring structure is formed by the sheer layer at the upper bound of the momentum deficit, however, it also shows part of the ring structure is generated by the lower boundary layer and be lift up to the upper positions. By tracking the vortex rings carefully, we can also find that there will be no major structure generated by the lower boundary layer in the downstream and be lift up to form the vortex rings. So this situation only happens in the initial stage of the ring generation and evolution nearby the MVG. In Fig 23, we can also find that during the process of forming the ring heads by the upper strong sheer layer at the momentum deficit, spanwise vorticity is continuously be transformed in to ring structure from the lower boundary layer or the shredding of the MVG surface. Thus, the constitution of the second ring in Fig 16 is a little complicated than the other ones since those lines come from both rolling up effect of the sheer layer and the lower boundary layer. The negative spanwise vorticity structures can also be found in both 2D (in blue color) and 3D (in purple color) in Fig 23. These structures are mainly from the lower boundary layer and the shredding on the MVG surface. Although 3D structures can also be generated by the negative vorticity and those structures can be lift up to even the upper bound of the momentum deficit, the vortex ring structure is absolutely constituted by positive spanwisely rotating vortex tubes. However, the existence of the negative vortex did influence the evolution of the ring structure a lot. In Fig 24, we tracked some vortex structures of positive and negative vortices (in the circled area), which shows the process of diminishing between the counter rotating vortex structures. The complicated influence between these vortex structures may continuously made the rings be badly distorted, which is also considered to be the influence to the ring structure induced by the lower boundary layer. t 1 t 2 15

17 t 3 t 4 t 5 Figure 24. Spanwise vorticity distribution at five continuous times on the central plane (left) and 3D space (right) Although we found that part of ring structure is formed by lower boundary layer, the lower boundary layer is still a shear layer. The conclusion all vortex rings are generated by shear layer still stands. The mechanism of the ring generation is a real 3D process. The interaction between the vortices and vortex rings shown by vorticity is a new interesting phenomenon. However, the origin of the vortex ring structure and the interaction inside the structure still need to by study in details. Conclusion Base on numerical results, the vortex ring structure generated by MVG is studied in this paper. The mechanism for the vortex rings which has been confirmed by a series of experimental works was analyzed, and it can be found that the existence of the high shear layer and inflection surface generated by the momentum deficit on the upper bound of the momentum deficit will cause the corresponding Kelvin-Helmholtz instability, which is one of the very important sauce of the vortex rings. However, the 16

18 lower boundary layer does contribute to the formation of ring structure and there exist complicated interactions among the vortex structures generated on both upper and lower boundary layers. References 1 Babinsky, H., Li, Y., and Ford, C.W.P., "Microramp Control of Supersonic Oblique Shock-Wave/Boundary-Layer Interactions" AIAA J., Vol. 47, No. 3, 2009, pp Lin J. C. "Review of Research on Low-Profile Vortex Generators to Control Boundary-Layer Separation." Progress in Aerospace Sciences, Vol. 38, pp (2002). 3 Ashill, P. R., Fulker, J. L., "Hackett, K. C. :A Review of Recent Developments in Flow Control." The Aeronautical Journal, 109, 1095, pp (2005). 4 Anderson, B. H., Tinapple, J. Surber, L. "Optimal Control of Shock Wave Turbulent Boundary Layer Interaction Using Micro-Array Actuation." AIAA paper Holden, H. A., Babinsky, H."Vortex Generators Near Shock/Boundary Layer Interactions" AIAA paper Ford, C. W., P,, Babinsky,H. "Micro-Ramp Control for Oblique Shock Wave/Boundary Layer Interaction." AIAA paper Holden, H.,Babinsky, H. "Effect of Microvortex Generators on Separated Normal Shock/Boundary Layer Interactions." AIAA J.pp (2007). 8 Ghosh, S., Choi J., Edwards, J. R. "RANS and Hybrid LES/RANS Simulations of the Effects of Micro Vortex Generators Using Immersed Boundary Methods." AIAA paper Lee, S. Loth E., Wang, C. "LES of Supersonic Turbulent Boundary Layers with µvg s." AIAA Paper Lee, S., Loth, E. "Supersonic Boundary Layer Interactions with Various Micro-Vortex Generator Geometries. " AIAA Paper Dolling, D. S., Murthy, M. T "Unsteadiness of the Separation Shock Wave Structure in a Supersonic Compression Ramp Flow field." AIAA J., Vol. 12, pp (1983). 12 Dolling, D. S. "High-Speed Turbulent Separated Flows: Consistency of Mathematic Models and Flow Physics." AIAA J., Vol. 36,pp (1998). 13 Dolling, D. S. "Fifty Years of Shock-Wave/Boundary-Layer Interaction Research: What next?." AIAA J., Vol. 39, pp (2001). 14 Settles, G. S. Dodson, L. J. "Supersonic and Hypersonic Shock/Boundary Layer Interaction Database." AIAA J., Vol. 32, (1994). 15 Dussauge, J. P., Dupont, P., Debieve, J. F. "Unsteadiness in Shock Wave Boundary Layer Interaction with Separation." Aerospace Science and Technology, Vol. 10, No. 2, pp.85-91(2006). 16 Andreopoulos, J., Agui, J. H., Briassulis, G. "Shock Wave-Turbulence Interactions," Annu. Rev. Fluid Mech., Vol. 32, pp (2000). 17 Loginov, M. S., Adams, N. A., Zheltovodov, A. A. "Large-Eddy Eimulation of Shock-Wave/Turbulent-Boundary-Layer Interaction."J. Fluid Mech., Vol. 565, pp (2006). 18 Wilcox, D., Turbulence Modeling for CFD, DCW Industries, Inc., Zheltovodov, A. A. "Advances and Problems in Modeling of Shock Wave Turbulent Boundary Layer Interactions." Proceedings of the international Conference on the Methods of Aerophysical research, Institute of Theoretical and Applied Mechanics, Novosibirsk, Russia, pp (2004). 20 Rizzetta, D., Visbal, M. "Large Eddy Simulation of Supersonic Compression-Ramp Flow by High-Order Method. "AIAA J., Vol. 39, No. 12, pp (2001). 21 Kaenel, R.V, Kleiser, L. Adams, N. A. Vos, J. B. "Large-Eddy Simulation of Shock-Turbulence Interaction." AIAA J., Vol. 42, No. 12, pp (2004). 22 Wu M. Martin, M. P. "Direct Numerical Simulation of Supersonic Turbulent Boundary Layer over a Compression Ramp." AIAA J., Vol. 45, No. 4(2007). 23 Adams, N. A."Direct simulation of the turbulent boundary layer along a compression ramp at M = 3 and Re θ = "J. Fluid Mech., Vol. 420, pp (2000). 24 Martín, M. P. "Direct Numerical Simulation of Hypersonic Turbulent Boundary Layers." Part 1. Initialization and Comparison with Experiments. J. Fluid Mech., vol. 570, pp (2007). 17

19 25 Ringuette, M. Wu, M., Martín, M. P. "Low Reynolds Number Effects in a Mach 3 Shock/Turbulent-Boundary-Layer Interaction. "AIAA J., Vol. 46, No. 7, pp (2008). 26 Wu, M., Martín, M. P. "Analysis of Shock Motion in Shockwave and Turbulent Boundary Layer Interaction Using Direct Numerical Simulation Data." J. Fluid Mech., Vol. 594, pp (2008). 27 Ringuette, M. J. Wu, M., Martín, M. P. "Coherent Structures in Direct Numerical Simulation of Turbulent Boundary Layers at Mach 3." J. Fluid Mech., vol. 594, pp (2008). 28 Priebe, S. Wu, M., Martín, M. P. "Direct Numerical Simulation of a Reflected-Shock-Wave/Turbulent-Boundary-Layer Interaction." AIAA J., Vol. 47, No. 5, pp (2009). 29 Bookey, P. Wyckham, C., Smits, A. "Experimental Investigations of Mach 3 Shock-Wave Turbulent Boundary Layer Interactions." AIAA Paper , LCattafesta III, L.N, Sheplak, M., "Actuators for active flow control," Annual Review of Fluid Mechanics 43 (2011) doi: /annurev-fluid Blinde, P.L., Humble, R. A., van Oudheusden, B.W., Scarano, F., "Effects of micro-ramps on a shock wave/turbulent boundary layer interaction," Shock Waves 19 (6) (2009) doi: /s Saad, R., Erdem, E., Yang, L., Kontis, K., "Experimental studies on micro ramps at Mach 5," in: Paper 2816, Proceedings of the 28th International, Symposium on Shock Waves, Manchester, U.K., July, Sun, Z., Schrijer, F.J.J., van Oudheusden, F.S.B.W., "The three dimensional flow organization past a micro-ramp in a supersonic boundary layer," in press, Physics of Fluids (2012). 34 Ko nig, B., Pa tzold, M., Lutz, T., Kra mer, E., Rosemann, H., Richter, K., Uhlemann, H., "Numerical and experimental validation of three dimensional shock control bumps," Journal of Aircraft 46 (2) (2009) doi: / Marxen, O., Rist, U., "Mean Flow Deformation in a Laminar Separation Bubble: Separation and Stability Characteristics", Journal of Fluid Mechanics (2010), vol. 660, pp Smits, A. J., Dussauge, J. P. "Turbulent Shear Layers in Supersonic Flow," 2nd ed., Springer Verlag, New York(2006). 37 Boris, J. P. Grinstein, F. F.,. Oran, E. S, Kolbe, R. J. "New insights into large eddy simulation. "J. Fluid Dyn. Res., Vol. 10, 1992, pp Fureby C., Grinstein, F. F. "Monotonically Integrated Large Eddy Simulation of Free Shear Flows." AIAA J., Vol. 37, No. 5, pp (1999). 39 Fureby C., Grinstein, F. F.. "Large Eddy Simulation of High-Reynolds-Number and wall-bounded Flows." J. Comput. Phys., Vol. 181, pp (2002). 40 Grinstein, F. F., Margolin, L. G., Rider, W. J. "Implicit Large Eddy Simulation," Cambridge university press(2007). 41 Visbal, M. R., Morgan, P. E., Rizzetta, D. P."An Implicit LES Approach Based on High-Order Compact Differencing and Filtering Scheme. " AIAA paper Morgan, P. E. Rizzetta D. P., Visbal, M. R. "Large-Eddy Simulation of Separation Control for Flow over a Wall-Mounted Hump." AIAA J., Vol. 45, No. 11, pp (2007). 43 Rizzetta, D. P. Visbal, M. R. Gaitonde, D. "Large-Eddy Simulation of the Supersonic Compression-Ramp Flow by High-Order Method," AIAA J., Vol. 39, No. 12,(2001). 44 Rizzetta, D. P. Visbal, M. R. "Application of Large-Eddy Simulation to Supersonic Compression Ramps." AIAA J., Vol. 40, No.8, pp (2002). 45 Garnier, E. "On the Use of Shock-Capturing Schemes for Large-Eddy Simulation. " J. Comput. Phys., Vol. 153, pp (1999). 46 Ladeinde, F. Cai, X., Visbal, M. R., Gaitonde, D. V. "Turbulent spectra characteristics of high order schemes for direct and large eddy simulation," Applied Num. Math., Vol. 36, pp (2001). 47 Gao, R. Yu, J, Kong, W., Yan, C. "Evaluation of a WENO Method in Implicit Large Eddy Simulation of Circular Cylinder Flow," The 2nd International Conference on Computer and Automation Engineering (ICCAE), Vol. 5(2010). 48 Li, Q. and Liu, C., LES for Supersonic Ramp Control Flow Using MVG at M=2.5 and Reθ=1440, AIAA , presented in AFOSTR Program Review in Mohd R. Saad, Hossein Zare-Behtash, Azam Che Idris and Konstantinos Kontis. "Micro-Ramps for Hypersonic Flow Control". Micromachines 2012, 3, 1-x manuscripts; doi: /mi30x000x 50 Li, Q., and Liu, C., Numerical Investigations on the Effects of the Declining angle of the trailing-edge of MVG, AIAA paper

20 51 Liu, C. and Chen, L., Study of Mechanism of Ring-Like Vortex Formation in Late Flow Transition, AIAA Paper Lu, F., Pierce, A. and Shih, Y., Experimental study of near wake of micro vortex generators in supersonic flow, AIAA paper Sun, Z., Schrijer, F.F.J., Scarano, F., and Oudheusden, B.W.V., "PIV Investigation of the 3D Instantaneous Flow Organization behind a Micro-Ramp in a Supersonic Boundary Layer, " University of Delft, ISSW28, July 17-22, 2011, Manchester, UK 54 Yonghua Yan, Caixia Chen, Ping Lu, Chaoqun Liu, "LES study on the mechanism of vortex rings behind supersonic MVG with turbulent inflow " AIAA