Document
AIAA-92-4089-CP
Measurement of Attachment-Line Location in a W'md-Tunnel and in Supersonic Flight Naval K. Agatwal t, Stan J. Miley 'I, Michael C. Fischer* NASA Langley Research Center Hampton, VA 23665 and Bianca T. Anderson* and RobertJ. Geenen* NASA Dryden Flight Research Facility Edwards, CA 93523 Abstract The phase-reversal phenomenon was As a part of the supersonic laminar flow examined to determine the attachment-line control research program, experiments are location on a highly swept wing. Test results being conducted to measure the indicate that the phase-reversal attachment-line flow characteristics and its phenomenon may not be applicable in location on a highly swept aircraft wing. flight. Additional data analyses is Initially, subsonic wind tunnel tests were continuing to account for phase differences performed on two-dimensional models to introduced by the recording system. A final develop sensors and techniques for the determination on the existence of phase- flight application. The wind tunnel test reversal in flight will require the results of results suggest that, under certain further analysis and possibly additional conditions at the leading edge there are low flight testing. Some representative frequency streamwlse velocity fluctuations attachment-line data are presented and present in the laminar boundary layer which discussed as well as the results from the undergo a phase change at the stagnation wind tunnel investigation.
point Therefore, if present in flight, this distinct change in phase, i.e., time shift of Nomenclature signals from either side of the stagnation point, can be used to determine the DFRF NASA/Dryden Flight Research stagnation-line or attachment-line location. Facility HLFC Hybrid Laminar-Flow Control t Senior Scientist, AS&M, Inc., Associate Fellow AIAA HSCT High Speed Civil Transport <JI Senior Research Associate, NRC, Member AIAA LaRC NASA/Langley Research Center * F- l 6XL SLFC Technical Manager, LFCPO/FAD, Member AIAA Laminar Flow Control LFC *Aerospace Engineer, Member AIM M Mach Number Copyright © 1992 by the American Institute of Aeronautics and Astronautics, Inc. No NASA National Aeronautics and Space copyright is asserted in the United States Administration under Title 17, U.S. Code. The U.S.
Government has a royalty-free licence to Uoo Free-stream tunnel velocity exercise all rights under the copyright claimed herein for Governmental purposes.
Reynolds number Re All other rights are reserved by the copyright owner.
a angle-of-attack line corresponds to the locus of stagnation points along the leading edge. However, for Introduction swept wings, there is no true stagnation point at the leading edge, due to a spanwise component of flow which is inherent with In a Joint Boeing and National Aeronautics and Space Administration (NASA) program, wing sweep. Knowledge of the accurate extensive flight tests on a Boeing- 757 using location of the attachment line is necessary hybrid laminar-flow control (HLFC) have to determine the state of the attachment- shown that a significant amount of laminar line boundary layer, which is of major flow can be achieved at subsonic speeds. importance since the local state of the laminar-flow control (LFC) is also a attachment-line boundary layer is the potential high-payoff technology for the starting point for the streamwise flow over High Speed Civil Transport (HSCT). the wing. If this boundary layer is laminar, Therefore, NASA has initiated a program to then the streamwlse flow boundary layer is initially laminar. However, if the determine the feasibil1ty of achieving attachment-line boundary layer is turbulent, extensive laminar flow on a highly-swept wing at supersonic speeds. The General then the streamwise flow will be turbulent Dynamics F - l 6XL was chosen as the test also. To maintain laminar flow on a swept aircraft due to its possession of a double wing, the attachment-line boundary layer delta-wing planform representative of future must remain laminar above a certain HSCT's proposed by Industry. Preliminary momentum-thickness Reynolds number, feasibility studies have shown that otherwise, once transition occurs, it contaminates the remaining spanwise part maintenance of laminar flow through active boundary-layer control ls viable on the of the wing boundary layer with turbulence.
inboard (subsonic leading edge) section of Transition criteria for the attachment-line boundary layer indicate that for laminar this type of wing planform. There are two F - l 6XL aircraft currently involved in flow, the attachment line should be maintained in the region of the leading edge supersonic laminar-flow research at the NASA Dryden Flight Research Facility with the highest degree of curvature. This (DFRF). The F- l 6XL-l has a single-place again emphasizes the importance of cockpit (see figure 1) while the F- l 6XL-2 has determining its location in flight as accurately as possible.
a dual-place cockpit. Current research activity at DFRF is utilizing the F- l 6XL- l with a laminar-flow suction glove (figure 1) During the initial phases of the suction glove in a co-operative Rockwell/NASA experiment on the F- l 6XL-1, experience experlmen t designed to establish the with pressure distribution measurements in feasibility of obtaining laminar flow on a the leading edge indicated that surface supersonic highly swept wing and to provide static pressure taps could not provide the data for CFD code calibration. The small necessary spatial resolution or accuracy dark area Is the perforated titanium laminar needed to precisely determine the flow suction glove (active). while the white attachment line location. Because of the area Is a non-suction (passive) glove fairing. thin boundary layer at supersonic speeds The F- l 6XL-2 will be used for flight research and the high surface curvature at the leading beginning July 1992. A leading-edge passive edge, pressure orifice interference appeared to have a pronounced effect on the scatter of glove will be tested Initially on the F- l 6XL-2 and will focus on providing both leading- the data to the extent that the interpolating edge physics Information and data for code process necessary to locate the attachment calibration.
line became highly questionable. Therefore, a new technique to measure the attachment- The attachment line Is the spanwlse dividing line location on the F- l 6XL- l glove using an line between upper-surface and lower- array of closely spaced hot-film sensors surface flow on the leading edge of swept flush mounted on the wing leading edge, was wings. For unswept. Le., straight wings, this Investigated.
Wind-tunnel tests were conducted on two- Velocity Calibration Wind Tunnel at the dimensional models, i.e., a cylinder and an Instrument Research Division of the Langley airfoil to develop sensors and measurement Research Center. This wind tunnel can techniques for the flight application. This reach test-section flow speeds up to 85 study looked at the phase reversal across m/ sec and has a turbulence level ranging the leading-edge stagnation point as a between 0. 7% at low speeds to 0.3% at possible means to determine the location of higher speeds. A 50 hp motor drives a the stagnation point. Wind tunnel tests squirrel cage blower at one of two selectable indicated, as discussed in the following speeds. The test section velocity is section, that the existence of a low- controlled by opening and closing 12 radial frequency disturbance is important in the damper vanes located downstream of the use of the phase-reversal technique. diffuser, and changing the motor speed. The Mangalam and Kubendran indicated a wind tunnel is shown schematically in figure direct coupling between the dominant 2. More details of the wind tunnel and the frequencies of flow oscillations at the test section are contained in ref. 8.
1 ea ding edge, laminar-separation unsteadiness, and vortex shedding. The Two models used were a circular cylinder influence of cylindrical-body wake (diameter 101mm), and a NACA 0012 airfoil oscillations on the pressure field and (chord 255mm). Surface-mounted micro- associated feedback mechanism have also thin hot-film sensor arrays were wrapped 2 3 been discussed by Morkovin . Tam around the leading edge of the cylinder and studied the character and origin of the the airfoil model. Surface-mounted multi- different components which make up the element hot-films have been utilized at noise signature of a wing. Paterson et al. NASA Langley Research Center to study the show that the condition necessary for the boundary-layer state, transition, and tonal behavior to exist Is that the wing must separation for both the subsonic and 6 1 be totally laminar. The tone originates from supersonic regimes - O. These sensors are discrete vortex shedding which formed by vapor deposition of nickel and accompanies the breakdown of a laminar copper on a 50 micron thick Kapton Sheet.
wake. Paterson et al. concluded that the The Kapton sheet is glued (about 20 micron discrete tonal behavior disappears, with the thick) to models and provides nearly non- appearance of a turbulent boundary layer. intrusive surface based measurements. A Phase reversal In the signal from sensors 33-elemen t sensor array (sensor size: located across a laminar separation bubble 1.145mm x 0.125mm, and spacing 0. 77mm) was observed by Stack et al. 5 on low was used on the leading edge of both Reynolds number airfoils. Scott et al. models. A 50-element sensor array with observed a phase reversal In hot-film spacing of 1.52 mm was also used near the sensors across the stagnation location on trailing edge of the airfoil model covering an airfoil using the same wind tunnel as X/C of 0.6 to 0.9. Flat ribbon cables were used in the present investigations. soldered to copper leads (laid on the same Mangalam et al. also indicated the Kapton sheet as the sensors) from each presence of phase reversal from sensors sensor and routed through ports on the across the attachment line on a swept wing models. Care was taken to ensure that the in wind-tunnel tests. connecting wires did not interfere with the flow at the test regions. DISA 55Ml0/55Ml2 This paper will desctibe in detail, the results constant temperature anemometer bridges of subsonic wind-tunnel tests on a cylinder were used to operate the sensors. The and an airfoil model, and preliminary results sensors were uncalibrated and were from flight tests at supersonic conditions.
operated at an overheat ratio of 1.3. The models were mounted on a circular plate Test Apparatus attached to the wind-tunnel floor, which could be rotated by an external screw The ground experiments were performed in mechanism to vary the angle-of-attack. The the 12- by 17-inch test section of the cylinder model and the wing instrumented with hot-film sensor arrays, are shown in changed from subcrltical to supercritical, figures 3 and 4, respectively. i.e., laminar separation being replaced by transition to turbulence and subsequent The F- l 6XL test vehicle planform consists of turbulent separation, the tone disappeared.
a inboard section with a sweep of 70 degrees A spectrum from one of the sensors showing and a outboard section with a sweep of 50 this is given in figure 8. A phase reversal in degrees. The NASA supersonic laminar flow this tonal signal was observed at the research program and the aircraft are stagnation line. Figure 9 shows time described In more detail by Fischer and histories of six sensors band-pass filtered Vemuru l 1 _ A 20-element hot-film array for the tonal signal. Phase reversal occurs (sensor size: 1.145mm x 0.125mm) with a between sensors 4 and 5 indicating that the sensor spacing of 0.0286 Inch, was wrapped stagnation line lies between these two around the leading edge just inboard of the sensors. The location of the stagnation line active glove as shown In figures 5 and 6. was confirmed by surface pi::essure Care was taken to maintain wrinkle-free measurements. In addition, by comparing Installation of the sensor array due to the phase relationships between sensors on the sharp leading edge and surface curvature same side of the cylinder, waves propagating near the leading edge. Due to the large in the boundary layer upstream from the variations In the temperatures associated rear of the cylinder were also observed.
with altitudes and compressibility, the hot- Rotating the cylinder moved the phase- film sensors were operated with a reversal effect through the sensor array temperature compensated anemometer accordingly so that the location of the sys tern described In ref. 12. This stagnation line was consistent with angle anemometer system utilizes a second hot- setting.
film array to measure the Instantaneous local temperature. This temperature signal In addition to the tonal signal. phase spectra of the sensor signals showed a is utilized by the anemometer circuit to broadband phase reversal over the low adjust the sensor overheat The temperature sensors are at the same chord location as frequency range. This was particularly the flow sensors, and are located about 1.5 evident for supercritical flow where the tonal Inches In spanwlse direction from the flow signal was absent. Figure 10 shows a phase sensor array: refer to figure 6. Due to the spectrum corresponding to the condition of limited number of anemometers which could figure 9. The tonal signal lies within the be accommodated onboard, only 12 sensors frequency band of phase reversal, but ( 5 through 15: see figure 7) were used. The otherwise can not be distinguished from sensor output was recorded on an onboard other sources. Figures 11 and 12 show 14 track FM tape recorder. The tape amplitude and phase spectra for a recorder was operated at 15 Inch/ sec., which supercritical flow case where the tonal resulted In a frequency response of 10 kHZ. signal ts not present. Phase reversal is from the sensors. Time code was recorded present, but no discrete source is on one channel of the recorder which was identifiable. A consistent time history, as later used In digitization of the data.
shown In figure 8, was not found.
Results and Discussion The broadband phase reversal pointed to other aerodynamic sources besides the Wind Tunnel Investigations: vortex shedding of the cylinder. A hot-wire probe was installed in the test section and The results from the cylinder model were the test condition repeated. Spectra of flow consistent with those of Mangalam and disturbances in the tunnel were obtained: an Kubendran . A tonal signal in the flow at example of which ts given in figure 13. It is the stagnation line was observed as long as seen to compare with the spectra of the the cylinder was subcrltical, i.e., laminar cylinder hot-film sensors of figures 8 and 11.
separation aft of maximum thickness. Using It was evident that the wind tunnel was a tuft it was observed that as the flow supplying an aerodynamic disturbance of sufficient strength to produce a phase- aerodynamic disturbance might exist reversal signal at the stagnation line of the associated with shock formation in model. supersonic flight. The results from the wind tunnel tests showed that a disturbance was The behavior of the airfoil was different from necessary, but gave no assurance that one the cylinder in that no self-generated would be encountered in flight. The flight aerodynamic disturbance, i.e., vortex investigation relied on the possible shedding was present. Detailed existence of a disturbance associated with measurements on the airfoil model the shock structure. As this was the first indicated that in the present set-up it was time that such high-frequency flow sensors possible to obtain laminar flow on one side had been located on the leading edge and of the airfoil model only. Paterson et al. flown at supersonic speeds, there was no showed that for a tonal signal to exist, the data from previous work to fall back on wing must be totally laminar. Achieving regarding the possibility of such a laminar flow on both sides of the airfoil was disturbance.
not possible, even at the low Reynolds number of the test, due to presence of the The aircraft was flown through a series of sensor connecting leads on the wing. Also, test conditions for which laminar flow was such extensive laminar flow is not likely to known to exist. With the present exist in flight experiments. Except for the configuration, laminar flow can only be absence of the vortex shedding signal, the obtained at angles-of-attack below the amplitude spectra were similar to those of corresponding trim angles for the test the cylinder. An example is given in figure matrix of Mach numbers and altitudes. This results in pitch-over maneuvers being 14. Random disturbances in the tunnel flow interacting with the wing leading edge, with required to reach the test matrix points, intensity increasing with the angle-of- leading to transient conditions and small attack, provided a signal necessary for the data windows.
application of the phase-reversal technique.
The stagnation line location was easily In all the test conditions flown, no identifiable by phase reversal between consistent phase-reversal behavior was adjacent sensors. A representative phase found. Analyses of the leading-edge hot-film spectrum for this is shown in figure 15. The sensor signals from the F- l 6XL aircraft location of the stagnation line as indicated show a complicated flow pattern. Unlike the by phase reversal was consistent with that wind-tunnel tests, a true attachment line indicated by pressure distribution exists here. The flow is laminar or turbulent measurements and extended over a wide depending upon local values of attachment- range of angles-of-attack. line momentum-thickness Reynolds numbers. Figure 16 shows a representative The results of the wind tunnel test showed time history of selected leading-edge that the stagnation line could be located by sensors as the aircraft ts maneuvered looking for phase reversal between adjacent through a range of angles-of-attack. Figures hot-film sensors. However, this technique is 17 and 18 show the corresponding angle-of- dependent on the presence of an attack and Mach number time history. The aerodynamic disturbance. For the turbulence level from sensor 10 has been technique to be of practical use, this added to the curve in figure 17 so that disturbance must be known to exist a priori. correlation between leading-edge flow state and angle-of- attack can be seen. At angles- Flight Investigations: of- attack less than 3 degrees, the leading- edge flow is laminar. Between 3 and 5 A preliminary investigation of the existence degrees, the leading-edge flow is turbulent.
of a suitable aerodynamic disturbance and This is due to the attachment line moving possible use of the phase-reversal below the leading edge with increasing technique was carried out on the F-16XL angle-of-attack. The radius of curvature of aircraft. It was assumed that a suitable the surface increases leading to an increase In the momentum-thickness Reynolds References number which causes transition in the attachment line and subsequent turbulent 1 Mangalam, S.M., and Kubendran, L.K., flow. At angles-of-attack above 5 degrees, "Experimental Observations on the the strong favorable pressure gradient at the Relationship Between Stagnation Region leading edge causes the boundary-layer Flow Oscillations and Eddy shedding for flow to relaminarlze as It progress to the Circular Cylinder," ICASE/NASA LaRC upper surface. The lower surface flow Instability Workshop, NASA Langley becomes turbulent above 5 degrees due to Research Center, 1989, pp. 3 72-386.
Increasing momentum-thickness Reynolds number. A demarcation between upper Morkovln, M.V., "Recent Insights Into surface flow and lower surface flow Is Instability and Transition to Turbulence In apparent In figure 16. The turbulence grows Open-Flow Systems," AIAA Paper No. 88- from the attachment line. For the range of 3685, 1988.
angles-of-attack shown In figure 17 the attachment line traverses through the entire Tam, C.K.W., "Discrete Tones of Isolated sensor array on the leading edge. This Airfoils," Journal of the Acoustical Society movement of attachment line Is not obvious of America, Vol. 55, No. 6, June 1974, pp.
from figure 16, but the effect of movement Is 1173-1177.
seen In the laminar and turbulent regions.
Paterson, R.W., Vogt, P.G., and Fink, M.R., As stated before, a consistent behavior of "Vortex Noise of Isolated Airfoils", Journal phase reversal In the signals from the of Aircraft, Vol. 10, No. 5, May 1973, pp. 296- leading-edge hot-film sensors was not seen 30'2.
In these first exploratory flights. As seen from the wind-tunnel results, in the case of 5 Stack, J.P., Mangalam, S.M., Berry, S.A., "A an airfoil, the phase shift is not evident from Unique Measurement Technique to Study the time-histories and the spectral analysis Laminar-Separation Bubble of the signal Is necessary. But, since the Characteristics on an Airfoil", AIAA Paper data were recorded on a tape recorder and No. 88-3675, 1988.
also the data analysis process requires digitizing the signals from a copy of the Scott, M.A., Carraway, D.L., and Lee, C.C., original FM fllgbt data tape, each analog "A Comparison of Techniques for the Determination of Stagnation Location," The tape recorder In this process may introduce phase shifts between the respective data 14th International Congress on channels. Additional analysis of the data, to Instrumentation In Aerospace Simulation Include any such phase shift Introduced, Facilities, Rockville, MD, October 27-31, will be required before a final determination 1991.
on the existence of phase reversal can be made. It is planned to Investigate the flow Mangalam, S.M., Maddalon, D.V., Saric, further outboard on the wing In future flights W.S .. and Agarwal, N.K., "Measurements of before a decision Is made regarding the Crossflow Vortices, Attachment-Line Flow, usefulness of the phase-reversal technique and Transition using Mlcrothln Hot Films," for supersonic flight. AIAA Paper No. 90-1636, 21st Fluid Dynamics, Plasma Dynamics and Lasers Acknowled~ements Conference, Seattle, WA, June 18-20, 1990.
The authors acknowledge the help of Wusk, M.S., Carraway, D.L., and Holmes, microelectronics and technical support B.J., "An Arrayed Hot-Film Sensor for section of electronics technology branch, Detection of Laminar Boundary-Layer Flow La RC In fabricating the sensors, and the F - Disturbance Spatial Characteristics," AIAA l 6XL technical staff at DFRF for performing Paper No. 88-4677-CP, AIAA/NASA/AFWAL the flight tests. Sensors & Measurements Technologies Conference, Atlanta, GA, September 7-9, 11 Fischer, M. C., and Vemuru, C.S., 1988. "Application of Laminar Flow Control to the High Speed Civil Transport - The NASA 9 Holmes, B.J., Carraway, D.L., Manuel, G.S., Supersonic Laminar Flow Control Program," and Croom, C.C., "Advanced Measurement SAE Paper No. 912115, Aerospace Techniques", Proceedings of Research in Technology Conference and Exposition, Natural Laminar Flow and Laminar Flow Long Beach, CA, September 23-26, 1991.
Control. NASA CP-2487, 1987.
12 Chiles, H., R., and Johnson, J. B., 10 Agarwal, N.K., Maddalon, D.V., Mangalam, "Development of a Temprature- S.M., and Collier, F. S., Jr., "Cross-Flow Compensated Hot-Film Anemometer Vortex Structure and Transition System for Boundary-Layer Transition Measurements using Multi-Element Hot Detection on High-Performance Aircraft," Films," AlAA Paper No. 91-0166, 29th NASA TM-86732, 1985.
Aerospace Sciences Meeting, Reno, NV, January 7-10, 1991 Flow control mechanism (12 radial damper vanes) Test section (12 x 17 in.)
drive Diffuser motor
l Exhaust
Figure 2: Schematic of Velocity Calibration Wind-Tunnel of Instrument Research Division at Langley Research Center.
Figure I : F- I 6XL- l aircraft.
HOT-FILM SENSOR ARRAY ACTIVE GLOVE Figure 5: Location of hot-film sensor array on Figure 3: Cylinder model in wind-tunnel.
the F- l 6XL- l aircraft.
PT 1
Figure 6: Close-up view of hot-film sensor array Figure 4 : Airfoil model in wind-tunnel.
on the F- l 6XL- 1 aircraft.
Aircraft Leading Edge ~ 5 VI ,_ 0.0286 in~" 1 VI C Cl) V') Time - mSec Figure 7: Location of hot-film sensors on the Figure 9: Time-histories of hot-film signals leading edge (not to scale).
located across the stagnation-line of cylinder model.
al "O Cl) Cl) VI n:; .c ~ 40 0..
a.
~
0~--------~-------~ 100 200 0 100 200 Frequency - Hz Frequency - Hz Figure 10: Phase spectrum of hot-film sensors 4 Figure 8: Power spectral density of hot-film signal (sensor 4 of figure 9): cylinder model, Re = and E of figure 9.
2.33x 10 .
al co -0
"
Cl.I a,
~
~ 40
~40 a.
a.
~
~
~.
~
100 200 100 200 Frequency - Hz Frequency - Hz Figure I I: Power spectral density of hot-film Figure 13: Power spectral density of hot-wire signal (sensor 4 of figure 9) for supercritical signal: tunnel free-stream, U..., = 42.6 m/sec.
case, Re=3.36x 105.
8Qr------------------ QL_ _______ __i_ _______ ___., 0 200 100 200 Frequency - Hz Frequency - Hz Figure 12: Phase spectrum of hot-film signals Figure 14: Power spectral density of hot-film (sensors 4 and 5 of figure 9) for supercritical signal; airfoil model, Re = 7.31 x 10 • a. = 0 case.
degrees.
Laminar ~ ro
f. 8
ar
'O 6 Laminar 0 100
i
Frequency • Hz 'o ..9,/ OI ~ Figure 15: Phase-spectrum of hot-fllm sl~nal; atrfotl model.
0 10 15 20 40 5 25 30 35 llme • Sec ..., .i.
~ ,.
Figure 17: Angle-of-attack, a variation for the
"
case shown In figure 16.
I.. • y, "'" Ill -, 2.0 .-------------------~ r· ~ E :::i ..
z 10
! 1.8 -
---\
T ~ QJ E Vl ::, Z L6 - .r::.
~ 1.4
1.2 '--
,,...,__ --•-"'• w~•---
I I
I I I I I I I I I I I I
1.0 0 5 10 15 20 25 30 35 40 0 5 15 20 25 30 35 40 Time • Sec nme - Sec Figure 18: Mach Number variation for the case Figure 16: Time-histories of hot-film signals shown in figure 16.
from F- I 6XL- l leading-edge sensors.