Some Features of The Unsteady Pressure Field in Transonic Airfoil Buffeting
Some Features of The Unsteady Pressure Field in Transonic Airfoil Buffeting
Some Features of The Unsteady Pressure Field in Transonic Airfoil Buffeting
AIRCRAFT 781
ARTICLE NO. 79-0351R
Comparisons are made between the unsteady transonic flowfields of two airfoils: a Whitcomb supercritical
airfoil, and a conventional NACA 0012 section. Wind tunnel experiments on these airfoils included penetration
into buffeting as a result of high cl and/or high M^. Fluctuating surface pressure, lift, and shock location were
measured on both airfoils. Two-point pressure cross-correlations were used to determine coherence and
propagation direction of pressure fluctuation patterns on the upper surface of each airfoil. Between the upper-
surface shock and the trailing edge (a region with intense pressure fluctuations), pressure disturbances
Downloaded by UNIV OF ADELAIDE (INTERNET) on February 1, 2015 | http://arc.aiaa.org | DOI: 10.2514/3.57965
propagated upstream in attached flow, but traveled downstream when extensive separation existed. In the latter
case, convection velocities were found to be frequency dependent. Another cross-correlation, relating surface-
pressure fluctuations to unsteady lift, was employed to establish which regions of the pressure fields were of
primary importance in producing buffeting forces. While many of the principal features of the pressure/lift
cross-correlations were common to both airfoils, some specific differences were found. For example, the
supercritical airfoil exhibited less periodicity in its cross-correlation. This result was attributed to the flat-
topped, aft-cambered shape of the supercritical airfoil section, which reduced the coupling between shock
oscillations and lift fluctuations.
Nomenclature Introduction
= pressure coefficient = (p —px) Iq^
= pressure coefficient for locally sonic flow
= pressure coefficient at trailing edge
T RANSONIC buffeting limits the maneuverability of
combat aircraft, and consequently remains the subject of
continuing study. 1>2 Because of the highly complex nature of
= air foil chord, m this challenging problem, involving the interaction between an
= section lift coefficient =L/q00c unsteady, separated, three-dimensional, transonic flow and a
= unsteady bending moment coefficient = m'b/Mb multi-component, three-dimensional structure that is
D = drag force per unit span, N/m vibrating in several modes, theoretical prediction of buffeting
f = frequency, Hz intensity will remain an impossibility for a long time to come.
f* = dimensionless frequency -fcl U^ Although rapid progress is currently being made in com-
L = lift force per unit span, N/m putational fluid dynamics, the calculation of even the two-
M = Mach number dimensional flowfield of a rigid airfoil in transonic buffeting
= mean spanwise bending moment at midspan, has yet to be accomplished. On the experimental side, at-
m-N tempts to use wind-tunnel measurements of surface-pressure
= fluctuating spanwise bending moment at mid- fluctuations to predict flight buffet loads have proven to be
span, m - N very difficult. 3 ' 4 Sophisticated models of transonic buffeting
P = static pressure, Pa flows must be developed before buffeting of aircraft, or even
q = dynamic pressure = p U2 /2 two-dimensional airfoils, can be predicted correctly by either
R(X,T) = autocorrelation coefficient of unsteady variable numerical or semiempirical methods. Definition of the
*(0,Eq.(2) necessary flow models depends upon the experimental
R(x,y,r) = cross-correlation coefficient relating unsteady determination of the significant details of transonic buffeting.
variablesx(t) andy(t), Eq. (3) McDonnell Douglas Research Laboratories (MDRL) has
Rec = Reynolds number = U^c/v^ been conducting an investigation of transonic airfoil
U = streamwise flow speed, m/s aerodynamics in cooperation with the Douglas Aircraft
= pressure fluctuation convection speed, m/s Company (DAC) and the NASA Ames Research Center
= unsteady downwash velocity at upper edge of (ARC). One element of this program has been research into
wake near trailing edge, m/s the unsteady aerodynamics responsible for buffeting. An
= chordwise distance, m initial series of experiments dealt exclusively with a Whit-
= chordwise displacement of shock, m comb-type supercritical airfoil.5'6 Within the ranges of cl and
= kinematic viscosity, m 2 /s MO,, studied, regions of intense pressure fluctuations were
= density, kg/m 3 identified, variations of fluctuating pressure power spectra
= cross-correlation time delay, s with local flow conditions were determined, and spanwise and
= root-mean-square value of unsteady component chordwise coherences of pressure fluctuations and wake
of variable downwash velocity oscillations were established.
( )c = freestream value Further transonic airfoil buffeting experiments were
performed recently. These new experiments covered greater cf
Presented as Paper 79-0351 at the AIAA 17th Aerospace Sciences and MOO ranges for the Whitcomb-type supercritical airfoil,
Meeting, New Orleans, La., Jan. 15-17, 1979; submitted Feb. 16, including penetration into heavy buffeting. In addition, a
1979; revision received March 5, 1980. Copyright © American In- comparable set of tests was performed on the conventional
stitute of Aeronautics and Astronautics, Inc., 1979. All rights NACA 0012 airfoil section to provide a hitherto unavailable
reserved.
Index categories: Nonsteady Aerodynamics; Transonic Flow.
set of unsteady-flow data for comparison with the super-
* Scientist, McDonnell Douglas Research Laboratories. Member critical airfoil results. New instrumentation provided lift
AIAA. fluctuation and shock oscillation measurements.
782 F. W. ROOS J. AIRCRAFT
The objective of these experiments was to clarify and define standard statistical procedures.9 Intensities, or root-mean-
the unsteady pressure field acting on the airfoils as each square levels, of fluctuating variables were evaluated, and
airfoil developed buffeting through increasing cl and/or M^. auto- and cross-correlations were computed. Power spectra
Cross correlation of unsteady variables was used to determine were generated by Fourier transformation of autocorrelation
the coherence, spatial and temporal scales, and propagation functions. In terms of arbitrary fluctuating variables x ( t ) and
directions and speeds of unsteady pressure disturbances in the y ( t ) , the various statistical quantities were defined as follows:
flow. Intensity
-1.4 -1.6
c/ = 0.42 ±0.02
-1.2 -1.4 Rec = 2 x 1
(boundary layer
tripped @ 18%c)
-1.0
-0.6
-0.4
-0.2
Downloaded by UNIV OF ADELAIDE (INTERNET) on February 1, 2015 | http://arc.aiaa.org | DOI: 10.2514/3.57965
0.82
0.4 I— R e = 2 x 106
(boundary layer tripped @ 35%c)
Fig. 2 Evolution with Increasing lift of pressure distribution on Fig. 4 Pressure distribution on NACA 0012 airfoil as a function of
supercritical airfoil. increasing Mach number.
M = 0.82
Rer = 2 x 10°
(boundary layer
0.20 tripped @ 35%c)
-0.8 -
-0.6 -
-0.4 «-
-0.2
c / = 0.55 0.04
Rec = 2 x 106 (x' s /c) Supercritical airfoil
0.20
(boundary layer c, c/ = 0.42 ± 0.02
-0.20
tripped @ 35%c) 0.03
0.01
0.10
0 0.20
Moo
0.75 0.78 0.80 0.82 0.84 0.86 0.88 0.90 0.92
MOO
0.70 ' 0
Fig. 6 Chordwise fluctuating pressure distribution for supercritical Fig. 9 Variation of supercritical airfoil flowfield unsteadiness with
airfoil: evolution with Mach number variation (shaded area indicates increasing Mach number.
location of shock region).
0.04 -0.10
c/ - 0.42 0
0.03
Downloaded by UNIV OF ADELAIDE (INTERNET) on February 1, 2015 | http://arc.aiaa.org | DOI: 10.2514/3.57965
Rec = 2 x 106
(boundary layer
0.10
tripped @18%c)
0.02 i
0.20
0.01
0.30
1.0
!
Airfoil
0.8 Conventional 0.80
Supercritical 0.82
R
(P90'P90- r ) 0.6 Supercritical 0.84
R
(P80-P90' r )
00
0.4
0.2
Shock
100 i—-
contribution to the high-frequency pressure fluctuations at the aerodynamics and airfoil motion, these pitching moment
wall beneath the separation remains to be defined. fluctuations differ from those associated with the torsional
Pressure disturbance propagation appears to be more buffet response sometimes exhibited by finite wings.4'17
complex when partial separation exists (e.g., shock-induced Changes in the location and/or strength of the shock in turn
separation followed by reattachment). In such cases, the affect the flow separation, and the convection of such changes
cross-correlations clearly show both up- and downstream is indicated by the sweep of the correlation contours toward
propagation of pressure fluctuations, indicating that both the the trailing edge with increasing T. Other noteworthy features
acoustic and convective modes were active. Frequency- of the correlation map in Fig. 15 are the low correlation level
resolved cross-correlations show that the higher frequencies on the forward part of the airfoil and the slight periodicity
(/* > ~ 7) travel downstream with the frequency-dependent (one cycle identifiable on either side of r = 0).
speed characteristic of the convective mode, while the lower The next three pressure/lift cross-correlation maps were
frequencies show a mixed behavior. This kind of mixed-mode obtained with the conventional NACA 0012 airfoil. The first
disturbance propagation appears on the 0012 section of these, presented in Fig. 16, represents the airfoil in a
whenever the flow reattaches, whether or not significant buffet-free condition (M00=0.68, c, = 0.42). As with the
buffeting exists, and also appears on the supercritical airfoil supercritical airfoil, the unsteady lift that does exist is
when trailing-edge separation occurs downstream of (rather
than at) the shock (e.g., M^ = 0.87 and 0.90 in Fig. 3).
Pressure/Lift Cross Correlations
Downloaded by UNIV OF ADELAIDE (INTERNET) on February 1, 2015 | http://arc.aiaa.org | DOI: 10.2514/3.57965
produced mainly by pressure fluctuations near the shock. reaches the shock so that buffeting exists, pressure distur-
However, in this case the pressure fluctuations clearly bances that arise in the shock-induced separation region and
originate at (or aft of) the trailing edge and propagate forward are subsequently convected downstream lose coherence before
until they affect the shock. reaching the trailing edge (Fig. 15); consequently these
At the highest lift measured (c, = 0.74 at M^ =0.68), the pressure disturbances do not have an organized effect on
0012 airfoil presents an interesting case of incipient buffeting circulation adjustment (which corresponds to lift variation) at
(Fig. 17). The flow remains attached at the trailing edge for the trailing edge. On the conventional 0012 airfoil, flow
this condition, and there is a large shock-induced separation separation begins as a shock-induced separation bubble that
bubble on the airfoil. The strongest pressure/lift correlation grows until it reaches the trailing edge. In this case, convected
appears to be more closely associated with the separation pressure fluctuations remain coherent all the way to the
bubble than with the shock. Because the flow reattaches, the trailing edge, and therefore have a more profound effect on
contours of strong correlation do not stretch toward the circulation changes that occur there.
trailing edge as prominently as they do when the flow is fully Shock-displacement/lift-fluctuation cross correlations
separated. The pronounced positive correlation at the trailing revealed that coupling between shock motion and unsteady
edge again appears to result from the prior half-cycle of lift lift during buffeting is also greater for the conventional
and pressure fluctuation. The greatest positive correlation for airfoil. The magnitude of the peak correlation is higher and
r>0 occurs not near the trailing edge, but under the the periodicity is more pronounced for the 0012 airfoil,
separation bubble. As is true for the supercritical airfoil, the consequences of the differences in chordwise pressure
time-dependent boundary between the positive and negative gradients on the upper surface of the two airfoils. The steep
correlation zones may be an indication that pitching moment chordwise gradients on the 0012 section mean that small
fluctuations exist. One result peculiar to this map is the displacements of the shock correspond to appreciable lift
substantial negative pressure/lift correlation near the leading changes and substantial variations in shock strength. Because
edge of the airfoil (where fluctuation intensities were low). of the flat upper surface, the supercritical airfoil has a
The pronounced periodicity of the pressure/lift cross- correspondingly flat upper-surface pressure distribution that
correlation map in Fig. 17 was characteristic of the 0012 allows the shock to move chordwise with comparatively little
airfoil. Heavy buffeting of this airfoil, which occurred at change in lift, and almost no change in strength.
M^ =0.80, c/ = 0.44, resulted in similarly strong periodicity,
as Fig. 18 shows. Comparison of Figs. 17 and 18 reveals that Conclusions
the frequency indicated by the cross-correlation map is not the Experimental study of unsteady pressure fields that are
same in the two cases. There is no connection between this responsible for transonic airfoil buffeting has provided new
periodicity and spanwise bending vibrations of the airfoil insight into the structure of the associated unsteady flow
model. The frequency appearing in these cross-correlation mechanisms.
maps varies according to flow conditions and is always For both a conventional NACA 0012 airfoil and a Whit-
greater than the spanwise bending mode frequency. comb-type supercritical section, the surface-pressure and lift-
Many of the features of the pressure/lift cross-correlation force fluctuations are associated primarily with the upper-
for the conventional 0012 airfoil in heavy buffeting are surface shock and the region of separated flow between the
similar to those of the supercritical section, suggesting that shock and the trailing edge. On the aft-cambered supercritical
these features are common to transonic airfoil flows in airfoil, separation and the corresponding increase of C'p
general. As with the supercritical section, the strongest always begin at the trailing edge; for the conventional 0012
correlation is associated with movement of the shock wave airfoil, the C'p rise first appears beneath a shock-induced
and the resulting shock-induced separation, and a significant separation bubble that grows toward the trailing edge with
correlation of opposite sign again appears near the trailing increasing c, or M^.
edge. The manner in which the correlation contours sweep Between the upper-surface shock and the trailing edge (a
toward the trailing edge with increasing T again identifies this region of intense pressure fluctuations), pressure disturbances
phenomenon with a shear-layer convection process. The fore on both airfoils are propagated upstream acoustically when
and aft phase reversal of the correlation in Fig. 18 is similar to the flow is attached, but are convected downstream in fully
that shown by the supercritical airfoil, although the negative separated flow. In the latter situation only, the pressure
peak is stronger and the positive peak is weaker for the fluctuation convection speed is strongly frequency-dependent,
conventional section. This difference in relative strengths of with the higher frequency (smaller scale) disturbances moving
positive and negative correlation peaks implies that pitching more rapidly.
moment fluctuations may be less important, relative to lift Cross correlations relating local surface pressures to lift
fluctuations, for the conventional airfoil than they are for the fluctuations were helpful in defining the spatial and temporal
supercritical section. Although slight correlation does exist structure of the buffet-producing unsteady pressure field. For
788 F. W. ROOS J. AIRCRAFT
5
both airfoils, pressure fluctuations at the trailing edge are out Roos, F.W., "Surface Pressure and Wake Flow Fluctuations in a
of phase with those near the shock, in terms of contribution to Supercritical .Airfoil Flowfield," AIAA Paper 75-66, Pasadena,
unsteady lift. This suggests that pitching moment fluctuations Calif., Jan. 1975.
6
might also be present in airfoil buffeting. Hurley, F.X., Spaid, F.W., Roos, F.W., Stivers, L.S. Jr., and
The supercritical airfoil differs from the conventional Bandettini, A., "Supercritical Airfoil Flowfield Measurements,"
section in two aspects: 1) a reduced tendency to develop Journal of Aircraft, Vol. 12, Sept. 1975, pp. 737-744.
7
Whitcomb, R.T., "Review of NASA Supercritical Airfoils,"
buffeting fluctuations as flow separation effects grow with 1C AS Paper No. 74-10, presented at the 9th Congress of the In-
increasing M^ and 2) less-periodic lift and pressure fluc- ternational Council of Aeronautical Sciences, Haifa, Israel, Aug.
tuations once significant buffeting does develop. These 1974.
8
differences are attributed to reduced coupling of lift, shock Roos, F.W., "Hot-Film Probe Technique for Monitoring Shock-
location, and flow separation fluctuations for the super- Wave Oscillations," Journal of Aircraft, Vol. 16, Dec. 1979, pp. 871-
critical airfoil, which results from the characteristic flat- 875.
9
topped, aft-cambered shape of the supercritical section. Bendat, J.S. and Pier sol, A.G., Random Data: Analysis and
Measurement
10
Procedures, Wiley, N.Y., 1971.
Roos, F.W. and Riddle, D.W., "Measurements of Surface-
Pressure and Wake-Flow Fluctuations in the Flow Field of a Whit-
Acknowledgment comb
H
Supercritical Airfoil," NASA TN D-8443, Aug. 1977.
Huston, W.B., "A Study of the Correlation Between Flight and
This research was conducted under the McDonnell Douglas Wind-Tunnel Buffet Loads," AGARD Rept. I l l , 1957.
12
Independent Research and Development Program in co- Monaghan, R.C., "Flight-Measured Buffet Characteristics of a
Downloaded by UNIV OF ADELAIDE (INTERNET) on February 1, 2015 | http://arc.aiaa.org | DOI: 10.2514/3.57965
operation with the NASA Ames Research Center. Supercritical Wing and a Conventional Wing on a Variable-Sweep
Airplane," NASATP-1244, May 1978.
13
DeAngelis, V.M. and Monaghan, R.C., "Buffet Characteristics
of the F-8 Supercritical Wing Airplane," NASA TM 56049, Sept.
1977.
References 14
Coe, C.F., Chyu, W.J., and Dods, J.B. Jr., "Pressure Fluc-
1
Unsteady Airloads in Separated and Transonic Flow, Proceedings tuations Underlying Attached and Separated Supersonic Turbulent
of 44th Meeting of AGARD Structures and Materials Panel, Boundary Layers and Shock Waves," AIAA Paper 73-996, Seattle,
AGARD-CP-226, Lisbon, Portugal, April 1977. Wash., Oct. 1973.
2 15
Aircraft Stalling and Buffeting, AGARD-LS-74, presented at the Willmarth, W.W., "Pressure Fluctuations Beneath Turbulent
von Karman Institute, Brussels, and NASA Ames Research Center, Boundary Layers," Annual Review of Fluid Mechanics, Vol. 7, 1975,
Calif., March 1975. pp. 13-38.
3 16
Mullans, R.E. and Lemley, C.E., "Buffet Dynamic Loads During Cliff, W.C. and Sandborn, V.A., "Measurements and a Model
Transonic Maneuvers," McDonnell Aircraft Co., AFFDL-TR-72-46, for Convective Velocities in the Turbulent Boundary Layer," NASA
1972. TND-7416, Oct. 1973.
4 17
Cunningham, A.M. Jr. and Benepe, D. B. Sr., "Prediction of Moss, G. F. and Pierce, D., "The Dynamic Response of Wings in
Transonic Aircraft Buffet Response," Proceedings of the 44th Torsion at High Subsonic Speeds," Proceedings of 44th Meeting of
Meeting of AGARD Structures and Materials Panel, AGARD-CP- AGARD Structures and Materials Panel, AGARD-CP-226, Lisbon,
226, Lisbon, Portugal, April 1977. Portugal, April 1977.
The current generation of internal combustion engines is the result of an extended period of simultaneous evolution of
engines and fuels. During this period, the engine designer was relatively free to specify fuel properties to meet engine per-
formance requirements, and the petroleum industry responded by producing fuels with the desired specifications. However,
today's rising cost of petroleum, coupled with the realization that petroleum supplies will not be able to meet the long-term
demand, has stimulated an interest in alternative liquid fuels, particularly those that can be derived from coal. A wide
variety of liquid fuels can be produced from coal, and from other hydrocarbon and carbohydrate sources as well, ranging
from methanol to high molecular weight, low volatility oils. This volume is based on a set of original papers delivered at a
special workshop called by the Department of Energy and the Department of Defense for the purpose of discussing the
problems of switching to fuels producible from such nonpetroleum sources for use in automotive engines, aircraft gas
turbines, and stationary power plants. The authors were asked also to indicate how research in the areas of combustion, fuel
chemistry, and chemical kinetics can be directed toward achieving a timely transition to such fuels, should it become
necessary. Research scientists in those fields, as well as development engineers concerned with engines and power plants, will
find this volume a useful up-to-date analysis of the changing fuels picture.
TO ORDER WRITE: Publications Dept., AIAA, 1290 Avenue of the Americas, New York, N. Y. 10019