Experimental investigation of jet mixing of a co-flow jet airfoil

5th Flow Control Conference28 June - 1 July 2010, Chicago, IllinoisAIAA 2010-4421Experimental Investigation ofJet Mixing of a Co-Flow Jet AirfoilB. P. E. Dano, D. Kirk and G.-C. ZhaDept. of Mechanical and Aerospace EngineeringUniversity of Miami, Coral Gables, FL 33124AbstractDownloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421The jet mixing of a co-flow jet (CFJ) airfoil is investigated to understandthe mechanism of lift enhancement, drag reduction, and stall margin increase.Digital Particle Image Velocimetry, flow visualization and aerodynamic forcesmeasurements are used to reveal the insight of the CFJ airfoil mixing process.At low AoA and low momentum coefficient, the mixing between the wall jet andmainflow is dominant with large structure coherent structures for the attached flows.When the momentum coefficient is increased, the large vortex structure disappears.At high AoA with flow separation, the CFJ creates a upstream flow strip betweentwo counter rotating vertical shear layer, i.e., the outer shear layer and inner flowinduced by CFJ. The UFS is characterized with large vortex free region. The co-flow walljet is deflected normal to the airfoil surface characterized with a saddle point. Withincreased momentum coefficient of the CFJ, the saddle point moves downstream andeventually disappears when the flow is attached.Turbulence plays a key role in mixing the CFJ with mainflow to transport highkinetic energy from the jet to mainflow so that the mainflow can remain attached at highAoA to generate high lift. When the flow is separated, increased CFJ momentumcoefficient also increases the turbulence intensity at jet injection mixing region.AoA Angle of AttackC Chord lengthCC Circulation controlCFD Computational Fluid DynamicsCFJ Co-Flow JetC L Lift CoefficientC D Drag CoefficientCµ Jet Momentum CoefficientD DragL LiftLE Leading edgeM Mach numberNomenclatureCopyright © 2010 by all the authors of this paper. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.

Downloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421flight. A movable flap at the airfoil TE is suggested to overcome the blunt TE drawback[14] but subsequently imposes a weight penalty. To maintain sufficient stall margin, LEblowing is usually needed. A considerable penalty of the CC airfoil is the dumped blowingjet mass flow, which is imposed on the propulsion system. Furthermore, for a CC airfoil,the drag measured in the wind tunnel is not representative of the actual drag that occursduring flight. This is because, in a wind tunnel test, the penalty to draw the mass flowfrom the free stream as the supply for the jet injection acts as a side force and can not beincluded in the drag measurement. The actual drag, also called the “equivalent” drag, musttherefore include this penalty [15, 16, 17], which is composed of the ram drag andcaptured area drag.Most recently, a new ZNMF jets flow control airfoil based on fluidic actuators(hereafter called co-flow jet airfoil or CFJ airfoil) has been developed by Zha et al. [18-20, 23], to avoid dumping the jet mass flow and achieve performance enhancementwithout relying on the Coanda effect. The co-flow jet airfoil is to open an injection slotnear leading edge and a suction slot near trailing edge on the airfoil suction surface. Ahigh energy jet is injected near leading edge tangentially and the same amount of massflow is sucked in near trailing edge. The turbulent shear layer between the main flow andthe jet causes strong turbulence diffusion and mixing under severe adverse pressuregradient, which enhances lateral transport of energy from the jet to main flow and allowsthe main flow to overcome severe adverse pressure gradient and remain attached at highangle of attack(AoA). The high energy jet induces high circulation and hence generateshigh lift. The energized main flow fills the wake and therefore reduce drag. The zero netjet mass flow of a CFJ airfoil can minimize the penalty to propulsion system.The purpose of this research is to investigate the turbulent jet mixing of a CFJairfoil in order to understand the mechanism responsible for lift enhancement, dragreduction, and stall margin increase. Digital Particle Image Velocimetry, flowvisualization and aerodynamic forces measurements are used to reveal the insight of theCFJ airfoil mixing process.II. Experimental setupA schematic of a CFJ airfoil concept is shown in Fig. 1. A high energy jet isinjected near the LE in the direction tangential to the main flow and the same amount ofmass flow is drawn into the airfoil near the TE. Pressurized air is injected in a spanwiselong cavity near the LE, and then exits through the spanwise long rectangular slot. ADuocel high density aluminum foam was placed between the inlet and the exit slot toequilibrate the pressure and ensure a uniform exit velocity. Similarly, a spanwise longcavity placed near the trailing edge is used to let the air settle down before being suckedthrough the three suction ports. The injection and suction slot height are 0.65% and1.42% of the chord, respectively, the airfoil chord is 12" and the span is 24". Theinjection slot and suction slots are located at 7.5% and 88.5% of the chord, respectively.All airflow and aerodynamic variables were acquired at the University of Miami24”x24”wind tunnel facilities. Fig. 2 and Fig. 3 show an overview of the UM Aero-labfacilities and the CFJ airfoil. The injection and suction flow conditions wereindependently controlled. A compressor supplies the injection flow line and the flow rate

Downloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421controlled using a Koso Hammel Dahl computer controlled valve. A vacuum pumpgenerates the necessary low pressure for suction and is controlled with a manual needlevalve. Both mass flow rates in the injection and suction lines are measured usingOripac orifice mass flowmeters equipped with high precision MKS pressuretransducers. The aerodynamics variables are measured using an AMTI 6 componentstransducer. All wind tunnel freestream, CFJ airflow, and aerodynamic variables wererecorded using a state-of-the-art Labview data acquisition system. LE trip was notimplemented since we found that the results are insensitive to the LE trip. All the data(e.g., wind tunnel speed, aerodynamics forces, blowing and suction mass flow rates) wereacquired at a rate of approximately 50 samples per second, allowing for limited airfoilflutter analysis around stall angle(s). The range of angles of attack (AoA) varied between0 o and 30 o . The nominal free stream velocity was V ∞ =10m/s for all tests and the chordReynolds number was about 1.89x10 5 . Similarly to C L and C D , a jet momentumcoefficient Cµ was defined as follows:mV jC (1)21/ 2 V SWith increasing AoA, the pressure over the airfoil is expected to decrease andfacilitate the CFJ jet penetration. This results in a higher Cµ. Fig. 4 shows the value ofCµ and corresponding jet velocity for all tests and illustrates that compensation wasrelatively well achieved for lower Cµ. For this paper, three nominal values of Cµ atAoA=0 o are presented: 0.06, 0.15 and 0.25, corresponding to mass flow rates of 0.030kg/s, 0.045 kg/s and 0.060 kg/s, respectively. The jet velocity is varied correspondingly atabout 23m/s, 34m/s, and 45m/s.Various laser flow visualization techniques were used to monitor the circulationover the upper surface. A LaVision Digital Particle Image Velocimetry (DPIV) systemwith a Litron Nano Nd:YAG 200 mJ/pulse was used to monitor and acquire the velocityfield surrounding the airfoil. An adaptive 64x64 to 32x32 pixel cross-correlation analysismethod was used with masking over the airfoil, resulting in 75x100 total vectors(including the airfoil). A series of 1,000 velocity fields were acquired for each AoA andCµ. A customized flow seeder using the same particles for PIV tracers was used for flowvisualization. The flow along the span direction was found to be uniform within 1%discrepancies between the root and the tip, while the suction showed a 7% skewnesstoward the root at the suction side. All the results presented were acquired along sectionB, on Fig. 3. III. Large Structure Vortex DetectionFor each series of test, the 1,000 instantaneous DPIV samples were averaged to yieldthe mean velocity and vorticity fields. A large structure vortex detection algorithm(VDA) based on Graftiaux et al.[21] was used to describe the vortical structure [2]. Forall grid points P of the DPIV grid, the vortex detection function, Γ 1 was calculated over asub-area centered on P. The sub-area size, A M , centered about each velocity grid pointwithin the flow, ranged from a 3x3 to 11x11 square grid with a grid node spacing of 2.4mm. In other words, the vortex structure size detected will be greater than 2.4x2.4 mm 2 .

The vortex detection function is defined by:1 ( P)AM( PM Uhp( M )) zˆ1dA M dAPM U M A sin()( )M1 (2)AhpADownloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421where M are the discrete grid points within the sub-area, A M , that correspond to velocitydata points. The larger the value of Γ 1, the greater the swirl component of the flow causedby the fluctuating velocity. The criterion |Γ 1 |>0.90 is used to identify a vortex motionwithin the flow, where Γ 1 varies between -1 and 1. Based on the sign of Γ 1 , the rotation iseither clockwise (CW) or counterclockwise (CCW). Upon detection of a vortex, a circleof the size of the detected vortex is plotted and filled with a color corresponding to thevortex rotation direction: blue=CW and red=CCW. When applied to the 1,000instantaneous flow fields, the VDA results are compiled into a single figure showing thedistribution of the detected vortices.Aerodynamic PerformancesIV. Results and DiscussionWind tunnel results for lift coefficient (C L ) and drag coefficient (C D ) for thebaseline and CFJ airfoils for different angle of attacks and different Cµ are shown inFig.5a and 5b. Compared to baseline, the CFJ airfoil shows a dramatic gain in C L and ismaintained with higher values of Cµ. Lift increase at AoA=0 o varies between 10% and80% for the range of Cµ used. The max C L increase at stall AoA varies between 22% and72% for the range of Cµ used.Fig. 6 regroups the performance increase for all Cµ at zero AoA and stalled AoA.These results can be explained by the combined effect of the co-flow jet and suctionincreasing the circulation and preventing flow separation. Drag coefficient results for theCFJ show a drastic decrease in drag compared to baseline. Moreover, zero to negativedrag (thrust) can be observed for a large range of AoA values. This trend is furtherincreased for higher Cµ. This can be explained by the reduced or reversed wake velocitydeficit resulting in drag reduction and thrust generation [17,18]. Since the aerodynamicefficiency of an aircraft is measured by L/D, the CFJ airfoil hence could be an efficientdevice to simultaneously increase lift and reduce drag with certain expense of pumpingpower consumption. The baseline results show a stall occurring around AoA=22°,whereas the CFJ airfoil results show a systematic increased stall AoA between AoA=25°and AoA=30° for all Cµ. Due to the large amount of data recorded, only the resultspertaining to the mixing and flow mechanism of the CFJ performance increase arepresented here. All the following results are for a freestream velocity of U ∞ =10m/s.Two samples of instantaneous smoke flow visualization for the baseline and CFJairfoil at AoA=25 o are shown in Fig. 7. One can see that the baseline airfoil flow islargely separated. On the top surface, Kelvin Helmholtz (KH) vortices can be seenshedding from the separation point. The flow from the intrado past the TE generates a

large recirculation that feeds the separated wake. In contrast, the CFJ airfoil shows asmooth flow hugging the top surface with a thin downward inclined wake past the TE.At low AoA, a closer look at the jet flow in the vicinity of the jet exit without andwith CFJ, shown in Fig. 8, reveals a street of small KH that dissipates rapidly. Thecoherent vortex structure rotates clockwise with no CFJ, and the rotating direction isreversed with CFJ due to the higher jet velocity than that of the freestream. The vortexstreet modulates in size with increasing Cµ and disappears for Cµ> 0.14.Downloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421General Mixing CharacteristicsA comparison of the time averaged velocity field for baseline with a typical CFJat AoA=25 o is shown in Fig. 9. For clarity, only 1 in 5 vectors are shown in all directions.The baseline airfoil shows a large separation originating from the LE, characteristic of astalled airfoil. The flow is observed passing from the TE to the LE all the way to theseparation point and is then entrained in the outer flow, creating a shear layer.Instantaneous PIV snapshots show that the shear layer region contains vortices, but inaverage, the air flow appears to merge nicely. The average velocity in the wake isobserved to have a much slower magnitude than the freestream.In contrast, the CFJ airfoil shows a smooth flow along the entire surface of theairfoil with no sign of separation. The injection jet velocity is about V j /U ∞ =2.2, but aregion of high velocity originating from the LE due to the strong LE suction effect isobserved, with values up to V max /U ∞ =3.0. Comparisons of the VDA results, shown in Fig.10, provide more information on the vortical structure. For baseline, the shear layer isrevealed with the multiple large CW vortices originating from the flow separation point.In contrast, the VDA results for the CFJ airfoil show that no large vortical structure ispresent.The wakes of airfoils are always of great interest. A comparison of the averagevorticity field for baseline and CFJ at Cµ=0.06 at AoA=25 o is shown in Fig. 11. The flowof the baseline airfoil from the intrado passes the TE and creates a large shear layer thatturns rapidly horizontal. VDA results (not shown) confirm that this shear layercorresponds to numerous CCW vortices spatially averaged. Further downstream, the flowof the baseline airfoil turns back, creating a large vortical structure. The air in theseparated region flows upstream along the extrado as observed in Fig. 9a. When the CFJin turned ON, the wake consists in a thin layer with opposite vorticity that extendsdownward with an angle of about 25 o from the freestream direction. This shear layer ischaracteristic of a jet, which reduces the drag or generates thrust depending on theintensity of the CFJ[17,18]. The VDA results reveal a few large structure vortices thatdisappear completely for higher Cµ.Fig. 12 to 14 show the time averaged PIV CFJ airfoil velocity fields and theVDA results for AoA=30 o for all three Cµ. For the lower Cµ=0.06, one can see that theflow is largely separated since the CFJ is not strong enough to achieve attached flow atsuch high AoA. The outer flow shear layer is seen shedding numerous CW (blue)vortices. Between the outer flow and the jet flow, the flow can be observed travelingupstream to meet the separation point, which is reminiscent of the baseline stalled flow.The upstream flow strip (UFS) is “sandwiched” between the outer shear layergenerating CCW vortices and the flow induced by the CFJ injection that produces CW

Downloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421vortices. The UFS is a region of large vortex free. Below the UFS, the jet appears todecay rapidly and strongly interact with the return flow from the TE.A white triangular marker and a dash line indicate the location of the saddle pointarea. The jet is deflected normally to the airfoil surface feeding the UFS and the outershear layer. The jet-to-return saddle point area is characterized by a rapid change fromthe CCW vortex of outer shear layer to the CW vortices induced by the CFJ. With theincreased Cµ, the saddle point is pushed downstream along the airfoil surface. The UFSbecomes thinner and the large vortices appear to increase in number and size. Finally, theresults for the higher Cµ=0.25 are shown in Fig. 14. The UFS appears to be completelysuppressed, the number of vortices is greatly decreased and the saddle point is, again,pushed further downstream. Inspection of instantaneous samples confirms that the abovefeatures are not artifact from the averaging process and show that the UFS is acombination of pairing of two counter-rotating vortices funneling flow from the wake tothe LE separation point, as shown in Fig. 15a. The jet deflection normal to the surfacealso appears to be eased by outer shear layer CW rollups, as seen in Fig. 15b. The flow inthe vicinity of the TE for Cµ=0.25 is shown in Fig. 16. While the jet action of delayingthe separation benefits to the airfoil circulation, the suction appears to slightly acceleratethe flow from the TE in the upstream direction. The wake is typical of a separated flow.CW vortices are observed on top of the shear layer with the “delayed” outer flow andCCW vortices are shedding off the TE and the shear layer flow from the intrado.Turbulent MixingCombining the time averaged velocity and VDA results provides ampleinformation on the mean and the vortical characteristics of the flow, but mixing shouldalso be considered through non-vortical fluctuations. The intermittent characteristics ofthe flow can be observed using the turbulent kinetic energy field (TKE). CorrespondingTKE fields for Fig. 12-16 are shown in Fig 17 to Fig. 19. For Cµ=0.06, most TKEoriginates in the jet area, indicative of the high velocity gradient caused by the injectionjet penetration. The outer shear layer starting from LE and stretching to downstreamalso shows high value TKE, again mostly due to the high velocity gradient between thehigh speed free stream and the UFS. While the wake above the TE shows very low levelsof fluctuation, the area below the TE shear layer shows very high TKE values, indicativeof a high intensity shear layer between the flow from the upper suction surface and lowerpressure surface, as shown in Fig. 7. The TE shear layer turbulence intensity is reducedwhen Cµ is increased. This is because the increased CFJ intensity decreases the velocitydissimilarity between the suction and pressure surface.For Cµ=0.25 shown in Fig. 19, the LE separation point and the CFJ injection areashows the highest value of TKE. The flow is about to be attached. Observations ofinstantaneous DPIV velocity fields shows that an intermittent flow pattern is establishedwith either a separated flow or a re-attached flow. In contrast, past the TE, the level ofturbulence in the wake is largely decreased due to reduced wake velocity deficit. Theincreased turbulence intensity at the CFJ injection area indicates the increasedturbulence mixing intensity, which transports high kinetic energy from the jet to the mainflow so that the main flow can overcome the severe adverse pressure gradient at highAoA and remain attached.

When the flow is attached, the turbulence kinetic energy is significantly reduceddue to the disappearance of large structure vortices as shown from Fig. 20 to 22.IV. ConclusionDownloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421The turbulence mixing mechanism of a CFJ airfoil is investigated in this paper. Atlow AoA and low momentum coefficient, the mixing between the wall jet and mainflowis dominant with large structure coherent structures for the attached flows. When themomentum coefficient is increased, the large vortex structure disappears.At high AoA with flow separation, the CFJ creates a upstream flow strip betweentwo counter rotating vertical shear layer, i.e., the outer shear layer and inner flowinduced by CFJ. The UFS is characterized with large vortex free region. The co-flow walljet is deflected normal to the airfoil surface characterized with a saddle point. Withincreased momentum coefficient of the CFJ, the saddle point moves downstream andeventually disappears when the flow is attached.Turbulence plays a key role in mixing the CFJ with mainflow to transport highkinetic energy from the jet to mainflow so that the mainflow can remain attached at highAoA to generate high lift. When the flow is separated, increased CFJ momentumcoefficient also increases the turbulence intensity at jet injection mixing region. Whenthe flow is attached, the turbulence kinetic energy is significantly reduced due todisappearance of large structure vortices.Acknowledgement:This research is supported under ARO/AFOSR Grant 50827-RT-1SP. Weappreciate the assistance from P. Ly, M. Castillo, A. Beaudry and K. Chin-Sim foroperating the wind tunnel tests.REFERENCES[1] W. L. I. Sellers, B. A. Singer, and L. D. Leavitt, “Aerodynamics for RevolutionaryAir Vehicles.” AIAA 2004-3785, June 2003.[2] M. Gad-el Hak, “Flow Control: The Future ,” Journal of Aircraft, vol. 38, pp. 402–418, 2001.[3] M. Gad-el Hak, Flow Control, Passive, Active, and Reactive Flow Management.Cambridge University Press, 2000.[4] S. Anders, W. L. Sellers, and A. Washburn, “Active Flow Control Activities at NASALangley.” AIAA 2004-2623, June 2004.[5] C. P. Tilmann, R. L. Kimmel, G. Addington, and J. H. Myatt, “Flow Control Researchand Application at the AFRL’s Air Vehicles Directorate.” AIAA 2004-2622, June 2004.[6] D. Miller, , and G. Addington, “Aerodynamic Flowfield Control Technologies forHighly

Integrated Airframe Propulsion Flowpaths.” AIAA 2004-2625, June 2004.[7] V. Kibens and W. W. Bower, “An Overview of Active Flow Control Applications atThe Boeing Company.” AIAA 2004-2624, June 2004.[8] R. Holman, Y. Utturkar, R. Mittal, and L. Cattafesta, “Formation Criterion forSynthetic Jets,” AIAA Journal, vol. 43, No. 10, pp. 2110–2116, 2005.[9] T. C. Corke and M. L. Post, “ Overview of Plasma Flow Control: Concepts,Optimization, and Applications.” AIAA Paper 2005-0563, Jan. 2005.[10] C. Enloe, T. E. McLaughlin, G. I. Font, and J. W. Baughn, “ Frequency Effects onthe Efficiency of Aerodynamic Plasma Actuator .” AIAA Paper 2006-0166, Jan. 2006.Downloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421[11] Raman, G. , “ Using controlled unsteady fluid mass addition to enhance jet mixing.,” AIAA Journal, vol. 35, No. 4, pp. 647–656, 1997.[12] Freund, J.B. and Moin, P. , “ Jet mixing enhancement by high-amplitude fluidicactuation,” AIAA Journal, vol. 38, No. 10, pp. 1863–1870, 2000.[13] N. Wood and J. Nielsen, “Circulation Control Airfoils-Past, Present, Future.” AIAAPaper 85-0204, 1985.[14] R. J. Englar, “Circulation Control Pneumatic Aerodynamics: Blown Force andMoment Augmentation and Modifications; Past, Present and Future.” AIAA 2000-2541,June 2000.[15] M. Wilson and C. von Kerczek, “An Inventory of Some Force Procedure for Use inMarine Vehicle Control.” DTNSRDC-791097, Nov, 1979.[16] G. S. Jones, “Pneumatic Flap Performance for a 2D Circulation Control Airfoil,Steady & Pulsed.” Applications of Circulation Control Technologies, Chapter 7, p. 191-244, Vol. 214,Progress in Astronautics and Aeronautics, AIAA Book Series, Editors: Joslin, R. D. andJones, G. S., 2006.[17] G.-C. Zha, W. Gao, and C. Paxton, “Jet Effects on Co-Flow Jet AirfoilPerformance,” AIAAJournal, No. 6,, vol. 45, pp. 1222–1231, 2007.[18] G.-C. Zha and D. C. Paxton, “A Novel Flow Control Method for AirfoilPerformance Enhancement Using Co-Flow Jet.” Applications of Circulation ControlTechnologies, Chapter 10, p. 293-314, Vol. 214, Progress in Astronautics andAeronautics, AIAA Book Series, Editors: Joslin, R. D. and Jones, G.S., 2006.[19] G.-C. Zha, C. Paxton, A. Conley, A. Wells, and B. Carroll, “Effect of Injection SlotSize on High Performance Co-Flow Jet Airfoil,” AIAA Journal of Aircraft, vol. 43, 2006.

Downloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421[20] G.-C. Zha, B. Carroll, C. Paxton, A. Conley, and A. Wells, “High PerformanceAirfoil with Co-Flow Jet Flow Control,” AIAA Journal, vol. 45, 2007.[21] Graftieaux, L., Michard, M. and Grosjean, N., 2001. “Combining PIV, POD andvortex identification algorithms for the study of unsteady turbulent swirling flows”,Meas. Sci. Technol., 12, pp. 1422-1429.[22] Dano, P.E., and Liburdy, J.A., “Vortical Structure of a 45 o Inclined Pulsed Jet in aCrossflow”, AIAA-2006-3543,2006.[23] Zha, G.-C., Gao, W. and Paxton, C. D. "Jet Effects on Co-Flow Jet AirfoilPerformance ", AIAA Journal, Vol.45, No.6, 2007, pp1222-1231

Fig. 1. Schematic and concept of CFJ airfoilDownloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421Fig. 2 Wind tunnel facilityCBAFig. 3 CFJ airfoil and PIV plans locationsFig. 4 Variation of Cµ and jet exit velocity for various AoA

(a)(b)Downloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421Lift Improvement (%)90%80%70%60%50%40%30%20%10%0%20%at AoA=0degat Stall AoA35%Fig. 5 Lift and Drag coefficient results comparisonfor baseline and CFJ at various AoA and Cµ46%59%0.06 0.14 0.25Cµ77% 82%Drag Reduction (%)1200%1000%800%600%400%200%0%at AoA=0degat Stall AoA92% 53%445%58%1009%0.06 0.14 0.25CµFig. 6 Performance enhancement at zero AoA and stall AoA.(a)(b)65%Fig. 7 Smoke flow visualization for (a) Baseline and (b) CFJ (Cµ=0.14) at AoA=25 o

(a)Downloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421(b)Fig. 8 Close-up flow visualization for CFJ at(a) Cµ=0 and (b) Cµ=0.02 for AoA=5 o

(a)(b)Downloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421Fig. 9 Time averaged velocity field at leading edge, for (a) baseline and(b) CFJ (Cµ=0.06) for AoA=25 o(a)Fig. 10 VDA results at the leading edge for (a) baseline and(b) CFJ (Cµ=0.06) for AoA=25 o(a)(b)(b)Fig. 11 Time averaged vorticity field at the trailing edge for (a) Cµ=0and (b) Cµ=0.25 for AoA=25 o

(a)(b)Fig. 12 Leading edge results for (a) mean velocity and (b) VDA for Cµ=0.06 andAoA=30 oDownloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421(a)Fig. 13 Leading edge results for (a) average velocity field and (b) VDA for Cµ=0.14 andAoA=30 o(a)(b)(b)Fig. 14 Leading edge results for (a) average velocity field and (b) VDA for Cµ=0.25 andAoA=30 o

(a)Downloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421(b)Fig. 15 Instantaneous DPIV samples Cµ=0.06 and AoA=30 o

(a)(b)Downloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421Fig. 16 Trailing edge results for (a) average velocity field and (b) VDA for Cµ=0.25 andAoA=30 o(a)(b)Fig. 17 TKE fields at (a) Leading edge and (b) trailing edge for Cµ=0.06 and AoA=30 o(a)Fig. 18 TKE fields at (a) Leading edge and (b) trailing edge for Cµ=0.14 and AoA=30 o(b)(a)(b)Fig. 19 TKE fields at (a) Leading edge and (b) trailing edge for Cµ=0.25 and AoA=30 o

(a)(b)Fig. 20 TKE fields at (a) Leading edge and (b) trailing edge for Cµ=0.06 and AoA=25 oDownloaded by Gecheng Zha on January 16, 2013 | http://arc.aiaa.org | DOI: 10.2514/6.2010-4421(a)(b)Fig. 21 TKE fields at (a) Leading edge and (b) trailing edge for Cµ=0.14 and AoA=25 o(a)(b)Fig. 22 TKE fields at (a) Leading edge and (b) trailing edge for Cµ=0.25 and AoA=25 o