Document
NASA Technical Memorandum 83210 NASA-TM-83210 19820004168
AEROELASTICITY MATTERS: SOME REFLECTIONS
ON TWO DECADES OF TESTING IN THE NASA
LANGLEY TRANSONIC DYNAMICS TUNNEL
..•. .1
Wilmer H. Reed III
~ ... _,. ;·-l r:: .. -... ~ :_, '.~ I,
September 1981
r~o\l 9 1981
L/'.NGLEY RCSE,I\RCH CENTER L![:m.'\RY, l~ilSi\ l-ll,!\'::10N, VIRGINIA
NI\SI\
National Aeronautics and Space Administration Langley Research Center Hampton, Virginia 23665
3 1176 00191 1438
NASA Technical Memorandum 83210
AEROELASTICITY MATTERS: SOME REFLECTIONS
ON TWO DECADES OF TESTING IN THE NASA
LANGLEY TRANSONIC DYNAMICS TUNNEL
Wilmer H, Reed III
September 1981
Nl\51\
National Aeronautics and Space Administration Langley Research Center Hampton, Virginia 23665 AEROELASTICITY MATTERS: SOME REFLECTIONS ON TWO DECADES OF TESTING IN THE NASA LANGLEY TRANSONIC DYNAMICS TUNNEL Wilmer H. Reed 111* ABSTRACT For more than two decades the unique capabili- ties of the TDT have been applied in a great Testing of wind-tunnel aeroelastic models has variety of aeroelastic investigations. The scope become a well established, widely used means of of this work ranges from flutter-proof tests of studying flutter trends, validating theory and high-performance aircraft to fundamental research investigating flutter margins of safety of new on aeroelastic phenomena. Figure 1 depicts some vehicle designs. The Langley Transonic Dynamics of the major technical areas currently under study Tunnel was designed specifically for work on in the TDT. For example, the facility is used to dynamics and aeroelastic problems of aircraft and verify, by means of dynamic models, the flutter space vehicles. This paper presents a cross sec- safety and aeroelastic characteristics of most U.S.
tion of aeroelastic research and testing in the high-speed military aircraft and commercial trans- facility since it became operational more than two port designs; to explore flutter trends and aero- decades ago. The paper illustrates, by means of elastic characteristics of new configurations; for examples selected from a large store of experience, active control of aeroelastic response of airplanes the nature and purpose of some major areas of work and rotorcraft; for ground wind loads, flutter and performed in the Transonic Dynamics Tunnel. These buffet testing of space launch vehicles; and for areas include: specialized experimental techniques; unsteady aerodynamic load measurements on oscil- development testing of new aircraft and launch lating wings and control surfaces.** These and vehicle designs; evaluation of proposed "fixes" to other specific areas of work performed in the TDT solve aeroelastic problems uncovered during devel- have been reviewed in various prior publica- opment testing; study of unexpected aeroelastic tions.(2-ll) A survey by Reed(8) indicates that phenomena (i.e., "surprises"); control of aero- the predictions from a variety of aeroelastic model elastic effects by active and passive means; and, studies in the TDT have, in general, been substan- finally, fundamental research involving measurement tiated in full-scale flight tests. The intent of of unsteady pressures on oscillating wings and this paper is to portray the broad picture of aero- control surfaces.
elastic testing and research performed in the TDT since it became fully operational in 1960.
I. INTRODUCTION The paper is organized along the following Aeroelastic problems encountered by high-speed lines. First, to set the stage, some salient fea- aircraft and launch vehicles most often arise in tures of the facility are described followed by a the transonic speed range, the very range where, discussion of specialized testing techniques devel- regrettably, aerodynamic theory is least developed.
oped for the study of aeroelastic problems. Then, Designers and researchers, therefore, must place drawing from a considerable store of research and heavy reliance on wind-tunnel models to aid in testing experience acquired over the years, exam- clearing new designs for safety from flutter and ples are selected and used to illustrate the nature buffet, evaluating solutions to aeroelastic prob- and purpose of major activities performed in the lems, and studying aeroelastic phenomena at tran- TDT. These activities include: development test- sonic speeds. The lack of suitable wind-tunnel ing of new aircraft and launch vehicle designs; facilities prompted A. A. Regier in 1951 to propose evaluation of proposed "fixes" to solve aeroelastic that the NACA construct a large transonic wind problems uncovered during development testing; tunnel to be dedicated specifically to work on study of unexpected aeroelastic phenomena (i.e., dynamics and aeroelasticity problems associated "surprises"); control of aeroelastic effects by with development of high-speed aircraft. (1) He active and passive means; and, finally, fundamental recommended that such a tunnel should: (1) be as research involving measurement of unst~ady aero- large as possible; (2) be capable of operating over dynamic pressures due to dynamic motions of lifting a wide density range; (3) use air or Freon as the surfaces.
test medium; and (4) operate through the critical transonic speed range, up to Mach number 1.2.
It should be understood that the paper, being in Regier's proposal began to become reality in 1955 the nature of a general survey, presents only a when work started on the conversion of a large sub- glimpse of these many-faceted aeroelastic scudies.
sonic tunnel, the Langley 19-foot pressure tunnel, In most cases, however, more detailed information to a l6-foot (4.87-m) transonic tunnel with may be found in the references cited in the paper.
Freon-12 or air as the test medium. This new facility, designated the Transonic Dynamics Tunnel II. TRANSONIC DYNAMICS TUNNEL FACILITIES (TDT), became fully operational in 1960 and has served ever since as a National facility dedicated The operating characteristics of the TDT and almost exclusively to work on dynamics and aero- some major features which make this facility par- elasticity problems of flight vehicles, particu- ticularly suited for experimental work on dynamics larly in the critical transonic speed range. and aeroelasticity problems are shown in Figure 2.
*
Chief Scientist, Loads and Aeroelasticity DiVision, NASA Langley Research Center.
** The oral version of the paper included a short kaleidoscopic-type motion picture showing a "diary" of flutter-model testing in TDT from 1960 to present.
Figure 1. Some aeroelastic technology areas supported by the Langley Transonic Dynamics Tunnel.
These features include a slotted 4.9 M x 4.9 M Further advantages associated with Freon as com- (16' x 16') test section, variable Mach number pared with air are a nearly three-fold increase in from a to 1.2, variable total pressure from 0.01 Reynolds number for comparable dynamic pressures to 1.0 atmosphere using either air or Freon-12 as and much reduced tunnel drive power requirements.
the test medium. The large test section is impor- tant because large models allow more accurate simulation of pertinent structural details, such as control surfaces and greater Reynolds numbers. A TUNNEL CHARACTER I STI CS - wide range of density variation is required in order to simulate altitude changes which often • TEST SECTION ------------- 4. 9m x 4. 9m (16' x 16') affect flutter. The use of Freon-12 gas • MACH RANGE --------------0 TO 1.2 (dichlorodifluoromethane) as a wind-tunnel test • TEST MED I UM------------- AI R OR FREON-12 medium for dynamically-scaled aeroelastic model • TOTAL PRESSURE ----------0.01 TO 1. 0 ATMOS.
testing has several extremely desirable features, • REYNOLDS NO. (MAX)- 3 x 106/m (l07/ftl the most important of which is its high density (the molecular weight of Freon is about four times SPEC I AL TEST I NG FEATURE S - greater than that of air). A denser test medium • COMPUTERIZED DATA ACQUISITION SYSTEM makes it easier to satisfy the density ratio scal- ing parameter which requires that the density of • "Q-STOPPER" FOR FLUTTER TESTING the model relative to the density of its surround- • TUNNEL -FAN SAFETY SCREEN ing fluid be the same as for the full-scale air- • SUSPENSION SYSTEMS FOR "FREE-FLYING" MODELS plane in the atmosphere. Thus, the use of Freon • GUST GENERATOR permits heavier and consequently more rugged, less expensive dynamically-scaled models. Also, because Figure 2. Transonic Dynamics Tunnel features.
Freon has a low speed of sound (about one-half that of air), the scaled vibration frequencies of models in Freon are half what they would be in air. This In addition to the Freon test medium, other has the advantage of easing data acquisition fre- features which make the facility uniquely suited quency requirements and of reducing the scaled for work on dynamics and aeroelastic problems are rotation speeds of model propellers and rotors.
given in Figure 2. These include a computerized feature of this mount system is a water-cooled data acquisition system especially designed to electromagnetic shaker installed inside the model rapidly process large quantities of dynamic data with field coils attached to the sting and moving for use in guiding the progress of tests. (10) coils attached to the model. This shaker is used Another special feature is a capability to rapidly to excite model vibration modes for the purpose of decrease the tunnel Mach number and dynamic pres- determining the ~erodynamic damping in each mode.
sure by means of quick opening valves which bypass a portion of the flow around the test section.
This capability reduces the risk of damage or loss of expensive models when flutter is encountered.
However, despite all reasonable precautions, models do, from time to time, experience catastrophic flutter, and therefore safety screens are provided to protect the tunnel fan from model debris. Also, to enable simulation of airplane free-flight dynamic motions in the wind tunnel, special model mount systems have been developed. And, to study airplane gust response problems, a system of oscillating vanes at the entrance of the test sec- tion is used to generate a sinusoidal variation in tunnel flow angle.
III. EXPERIMENTAL TECHNIQUES Wind-tunnel tests of aeroelastic models require specialized experimental techniques seldom found in other types of wind-tunnel studies. Over the past years, a variety of new or improved experi- Figure 3. Launch vehicle buffet models.
mental techniques has been developed by the staff of TDT and others to broaden capabilities for study of dynamics and aeroelasticity problems of Shown on the right side of the figure is a aircraft and space launch vehicles. Although most O.OSS-scale model of the space shuttle used in of the techniques currently in use have been flutter and buffet studies. This model was iso- described in earlier publications, it is felt that lated from the sting by means of a pair of pneu- a brief review of a few of these would provide use- matic springs which allowed freedom in pitch and ful background for the present paper.
plunge modes. A locking system was provided to restrain the model in the event of structural Model Mount Systems failure or dynamic instability of the model on its In wind-tunnel-model studies of dynamics and soft suspension system. During the tests the aeroelastic problems careful attention must be model was "flown" at low lift coefficients by given to the manner in which models are suspended adjusting the sting angle and positioning the in the tunnel. A model rigidly mounted in the orbiter elevons to unload the wing.
tunnel does not, in general, properly simulate the dynamic characteristics of its free-flying counter- Airplane Models. Large complete airplane part. On the other hand, a "soft" suspension sys- models are routinely used in the TDT to study a tem that provides a model the freedom of motion variety of dynamic aeroelastic problems including needed to simulate free-flying conditions may at flutter, buffet, gust response, and stability the same time introduce instabilities of its own.
derivative measurements. In most instances the Considerable effort has therefore gone into the free-flight dynamic characteristics of the air- development of suitable "free-flight" mount systems craft play an important role in these studies mak- for use with aeroelastic models of both launch ing it necessary that such characteristics be vehicles and airplanes.
simulated in the model tests. To satisfy this need the so-called two-cable mount system was Launch Vehicle Models. An effective technique developed and has been used extensively in the TDT for predicting the buffet response of launch and in other tunnels as well. This basically vehicles by use of a suitably scaled aeroelastic simple model-mount system was first described and model is described by Hanson and Jones. (3) This (12) technique involves a complete-vehicle model sus- analyzed by Reed and Abbott. Although ini- pended in the wind tunnel on a sting mount which tially developed primarily for use in flutter provides the model freedom to respond essentially work, the two-cable mount, with variations, has in its "free-free" bending modes. Two such sting- since proven its versatility and usefulness in mounted models are shown in Figure 3. On the left other areas as will be discussed later.
side of the figure is shown a 0.08-scale Saturn 1- Apollo used in buffet response and aerodynamic A schematic diagram of the basic two cable damping studies in the transonic speed range. The mount system is shown in Figure 4. Two loops of mount system is designed to restrain the model in small-diameter cable extend in mutually perpen- the longitudinal and lateral directions but soft dicular planes from the model to the tunnel walls, springs are provided to give the model a rigid- one loop upstream and the other downstream.
body-pitch degree of freedom which simulates that Cables pass through pulleys located within the of the full-scale vehicle with its control system.
model contour and tension is applied by stretching The difference in the weight and lift force acting a soft spring in the rear cable. In addition, a on the model is compensated for by means of a snubber cable system (not shown in Fig. 4) is cable system through the sting mount and remotely provided for emergency restraint. These small located adjustable springs. Another important diameter cables cause little aerodynamic inter- need to simultaneously satisfy both Mach number ference and have negligible mass, compared with and Froude number (gravity scaling parameter). In that of the model.
Freon this is possible but only when the model length scale factor is about 1/4. Except for small fighters most models tested in the TDT have length scale factors less than 1/4 and consequently fly at smaller angles of attack (lift coefficient) than does the airplane. To permit simulation of the proper load factor on cable-mounted models a lift-balancing device (see Fig. 4) has been devel- oped to counteract the lift in excess of the model weight. This device, described by Hanson,(15) con- sists of a soft pneumatic spring which, by means of a cable attached near the model center of gravity, applies a relatively constant force with minimum restraint to model motion. This point- force simulation of gravity is only approximate, of course. Although the total lift is correct, the total lift distribution may not be, due to inertia and pitch rate effects. Comparative wind- tunnel/flight buffet studies by Hanson(ll) of the Figure 4. Two-cable mount system.
F-lll variable sweep fighter indicate, however, that the buffet response predicted by the model correlate well with that measured in flight as Remotely operable trim controls on the model indicated by Figure 5. This figure shows various are provided to keep the model centered in the buffet response quantities obtained in high-g tunnel throughout the test range. Pitch and roll flight maneuvers and in the wind tunnel on a are usually sufficient and a single operator or liB-scale flutter model of the F-lll "flown" to "pilot" can fly the model using a miniature air- simulate these high-g conditions.
plane-type control stick which positions the model • AIRPLANE control surfaces.
1.25
o
1.0
o MODEL
Although in principle the mount system is .75 C.G. ACCELERATION, simple, in practice a detailed stability analysis INDICAlED AIRPLANE .5 PERCENT DESIGN of the model/mount system is needed to guide the BUFFET ONSET .25 choice of design parameters (e.g., cable geometry, 01.-_1--_1--_1...-;Y--1--_1-----l cable tensions, and model e.g.) for each new model.
In fact, a stability analysis for the cable mounted HORIZONTAL TAIL model is usually more complex than that of the air-
"r
BEND I NG MOMENT.
plane because of the added degrees of freedom of PERCENT DES I GN the mount which may become unstable. Stability
:~: t_--l-_....I--~L ~
analysis procedure for cable mounted models may be 1.00 found in references such as 12, 13, and 14.
•
o •
. 75 WING ROOT In addition to the requirement for a mount sys-
.~
BEND I NG MOMENT, .50 tem stability analysis, it is also considered good PERCENT DESIGN .25 practice for the pilot and test crew to gain
2ff1
"flying" experience on a "dummy" model prior to risking the more expensive aeroelastically scaled
1.2 1.0 o
model. The dummy model is built much stiffer than the aeroelastically scaled model but has the same geometry, and total mass and inertia propertjes.
Figure 5. Buffet response measurements (Ref. 15).
As mentioned, the two-cable mount system has been adapted for studying a variety of aeroelastic Gust Response problems in addition to flutter. Some of these The need for a wind-tunnel technique for study non-flutter aeroelastic applications relating to of the dynamic response of airplanes to atmo- aircraft stability, control, and loads have been spheric turbulence, particularly in the transonic discussed in review papers by Rainey and Abel, (7) speed range, stimulated the development of a unique airstream oscillator system for use in the Reed, (8) Abbott, (6) and Hanson. (15) The following TDT. This technique involves measuring the fre- section describes an adaptation of the mount sys- quency response function of a "free-flying" aero- tem to enable study of high angle-of-attack phe- elastic model using as input a sinusoidal vertical nomena, in particular, aircraft buffet response.
gust field generated by oscillating vanes located upstream of the test section as indicated in Aircraft Buffet Figure 4. The airstream oscillator system con- For some aeroelastic phenomena such as buffet, sists of two sets of biplane vanes on each side the model lift coefficient becomes an important of the test-section entrance. The vanes are scaling parameter. Models flown in the two-cable oscillated sinusoidally in pitch at frequencies mount are usually trimmed to be in equilibrium at up to 20 Hertz. Trailing vortices from the vane the center line of the test section where aerody- tips pass downstream near the side walls of the namic lift on the model exactly balances the model test section and induce a reasonably uniform dis- weight. This "l-g" condition on the model does tribution of vertical velocity components across not in general represent a l-g condition for the the model span. The model, suspended on the two- airplane. The reason for this has to do with the cable mount system, is free to respond to these MODEL 1 0 TEST
3 "gust" inputs in approximately the same manner as
16 X 10-
--CALC.
would the full-scale flexible airplane in free flight. Guidelines for designing a model mount FLI GHT TE ST.
system that is not only stable but also produces negligible distortion of the airplane short period free-flight mode have been developed and demon- strated in preliminary experimental gust response studies by Gilman and Bennett. (13) To further verify the validity of wind-tunnel data obtained by the airstream oscillator tech- nique a comparative wind-tunnel/flight-analysis study was undertaken jointly by NASA Langley, the Air Force Flight Dynamics Laboratory, and the Boeing Company (Wichita Division). Because con- siderable flight data were available for the B-52E airplane in the form of gust frequency response functions, this airplane was chosen as the test
o
vehicle. These flight-measured frequency response functions were determined from simultaneous mea- surements of the random atmospheric turbulence input and the associated airplane response using
- 200 ~-'--------=-=---'---:~--'------!--::--'-_-!--:--'----.J
power spectral analysis techniques such as
o .04 .08 . 12 . 16 .20
described in reference 16.
REDUCED FREQUENCY. k Shown in Figure 6 are the 1/30-size model of Figure 7. B-52 model and airplane gust the B-52E, mounted on cables in the TOT together frequency response measurements.
with the gust vanes upstream of the test section and a gust measurement probe in the vicinity of Stability Derivative Measurements the model. Some results from the study reported Aeroelasticity has naturally a major influence by Redd, Hanson, and wynne(l7) are presented in on the stability and control characteristics of Figure 7, which shows frequency response for a flexible aircraft at high speeds. The loss of nondimensional wing bending-moment coefficient at aileron control effectiveness or the change in the midwing span per degree of vertical gust angle lift-curve slope with increasing speed are promi- as a function of reduced frequency k (based on nent examples. Paralleling the methods developed mean aerodynamic semichord). Shown for comparison for flight testing, wind-tunnel testing techniques with the measured flight data are the measured and have been developed in the TOT for extracting air- calculated frequency response functions for the plane stability derivatives from aeroelastically model.
scaled models. These techniques again employ "free-flying" cable-mounted models. In the ,~ind The major response in this case comes from the tunnel, as in flight, the model response to known short period mode at reduced frequency k = 0.08.
inputs, such as control surface deflections or At very low frequencies the model response is external forces applied through the suspension affected by the mount system vertical plunge mode cables, is measured and used in the equation of and the airplane response by spurious pilot-induced motion of the model and suspension system to solve motions; at higher reduced frequencies (k > 0.14) for the unknown aerodynamic coefficients. Two such the low level of gust input in the wind tunnel techniques for identifying aerodynamic parameters leads to measurement inaccuracies and scatter in from tests of cable mounted models are described the model test data. The overall satisfactory below.
correlations between wind tunnel, flight, and analytical predictions indicate this to be a use- The first, and simpler, of these techniques ful valid wind-tunnel technique for airplane gust enables one to determine the aileron effectiveness loads research.
and roll damping from aeroelastic model tests.
This approach developed by Abel(7,18) is based on the assumption that the dynamic roll response of the model to sinusoidal aileron oscillations can be represented by a single-degree-of-freedom system in roll. Using a flutter model of a large cargo transport aircraft, the C-14l, Abel obtained aile- ron effectiveness and damping in roll data in the wind tunnel which agreed well with flight measured data. The aileron effectiveness of this model was also determined successfully by an alternate method described by Grosser(19) in which the model was supported by a sting-mounted pylon with springs to provide the model freedom of motion.
The second method is patterned after modern sys- tem identification techniques in current use for extracting airplane stability derivatives from Figure 6. B-52 model used in wind-tunnel flight test data. This technique has been applied gust response studies (Ref. 17).
yaw rates (determined from rate gyros in the model) by Bennett, Farmer, Hohr, and Hall(20) to deter- are suitably mixed and fed back as input commands mine both longitudinal and lateral stability to the torque motors. These stability augmenta- derivatives using dynamic models of the F-14 and tion features of the active mount system have been space shuttle orbiter. In these studies the two- analyzed and modeled by Chin and Barbero(2l) and cable mount system was modified by the addition of demonstrated in the wind-tunnel studies of Bennett servo torque motors and load measuring cells in each of the cable loops as indicated in Figure 8. et al. (20) using the shuttle orbiter model. In this case the model with its c.g. moved aft to produce static longitudinal instability was effec- tively stabilized by means of feedback propor- tional to the output of a model pitch rate gyro.
Subcritical Testing Techniques , I To reduce risk of damage or destruction of I expensive aeroelastically scaled models due to the I I sudden encounter of aeroelastic instabilities,
~:::: - _~ ~:: :>~:ULL£YS
various subcritical testing techniques have been developed to aid in predicting instabilities from [ VERTICAL ", CABLE ' response measurements obtained under stable safe test conditions. With regard to flutter testing, a comprehensive symposium(22) on this subject was held in 1975 which covered many aspects of wind- tunnel flight flutter testing. Host of the sub- critical flutter testing techniques in current use in the TDT were presented by various authors at this symposium: notably, by Foughner; Hammond and Figure 8. Active two-cable mount system.
Doggett; Bennett and Desmarais; and by Houbolt.
Hany of these procedures have been implemented on By means of this "active" mount system, dynamic TDT data acquisition systems and others are being excitation forces can be applied to the model developed to aid the test engineer in guiding the through the cables to provide known inputs suitable conduct of flutter test on a near-real-time basis.
for the parameter identification analyses. The longitudinal and lateral derivative sets determined Recently, an upsurge of renewed interest in for both modes were in reasonable agreement with forward swept-wing concepts, which are usually derivatives determined from other sources. Table I more prone to static divergence than to flutter, shows, for example, estimates of the stability sparked the development of subcritical static derivatives for the shuttle orbiter determined by divergence testing techniques. Using a series of the test technique of Reference 20 together with simple flat aluminum-plate models in the TDT initial estimates by the manufacturer based on Ricketts and Doggett(23) evaluated six different other wind-tunnel tests and theoretical analysis.
subcritical divergence testing techniques. Four of these were based on measurements of static data such as strain-gage measured mean bending moments TABLE I. SHUTTLE ORBITER STABILITY DERIVATIVES at the wing root; the other two methods were based on dynamic measurements of such quantities as Stability Initial Ref. 20 modal frequencies and peak response amplitudes.
derivative estimate estimate Two of these static methods are illustrated in Longitudinal: Figure 9. Both use as input data the load/angle- of-attack gradient, Ai' measured at dynamic pres- -2.51 -3.04 CZ ex sures that are well below the divergence point C C -2.40 -2.68 + while holding Hach number constant (Fig. 9a). In m ma q one method a so-called "divergence index" param- C 0.164 0.164
rna eter, 6, is calculated using measured values of
the loadiangle-of-attack gradient and dynamic Lateral: pressure as indicated by the equation in Figure 9b.
C -0.920 -0.935 When plotted against dynamic pressure, q, the yS divergence index is a straight line which passes -0.046 -0.032 CZ S through unity at q = 0 and crosses the q-axis at -0.288 -0.306 CZ the predicted divergence dynamic pressure. Experi- p ence with the method has shown that the predicted CZ 0.127 0.103 r divergence condition is accurate even when extrapo- C 0.044 -0.022 lated from subcritical data acquired at dynamic ns pressures well below the divergence point.
C 0.194 0.011 n p C -0.231 -0.167 The other static method is illustrated in Fig- nr ure 9c. It is based on the observation by Flax(24) that experimental procedures developed in the 1930's by R. V. Southwell for the prediction of The active mount can be used also to augment the column buckling loads from measurements of the rate stability of an otherwise unstable model configura- of change of column lateral deflection with load tion. In this mode electrical signals proportional were equally applicable to aeroelastic studies such to such variables as the torque motors' angular as aileron reversal and static divergence. By the position and rotational rate and/or model pitch and 1960 62 64 66 68 70 72 74 76 78 80 I I I I I I I I I 1 I I I I I I I t I I I I LOCKHEED ELECT. ~ H C·141------'O'"T r-- ......- ...
F·IlI --------'I'.'- ...,.~~ " ............. ~ ........ ~---__'.~"' ............
LOCKHEED HELICOPTER--- ...
BELL HELICOPTER------- ...... ~ ...
WING C-5A----------~,,:.: :~; "" • .._-----'1' ...
BEND ING BOEING SST--------~ • .._-_'I'.'--'.~._-'I' ...
MOMENT LOCKHEED SST--------'/' ...
m ....
HOIl • W ANGlE OF ATTACK H4 _ ...
OC~ ...... SOUTHWELL METHOD DI VERGENCE I NDEX METHOD S·JA """ XV·I5 TILT ROTOR • ....
H5 _ B·! "'" •••• DIVERGENCE LIGHT WT. FIGHTER (GO I •• ...
SLOPE.
B·52CCV • .... INDEX, , C ·5A ACTI VE LOAD CONTROL ...
qDIV t; H6 --------------------~ ... " ...... J'Y".,.or_ ... .,.~__'I' ...
qDIV SPACE SHUTILE ::============='.~=='.~==~.~""'""' GULFSTREAM III .,.
YH7 STORES FLUTTER SUPPR. --------------"' ...... - ...........
YH6 STORES FLUTTER SUPPR. ---------------'lj.'-~.or__ ...
I ci L1::.bl'--L_.l--L_.l--L----""'1--J B·767-------------------- __ -... ...
HELICOPTER ROTORS----------------------T ... ,iq DYNAMIC PRESSURE, q Figure 10. Aircraft flutter investigations.
Figure 9. Subcritical divergence prediction methods.
C-14l Jet Cargo Transport Aeroelastic model investigations in TDT of the C-14l involved three models in nine separate wind- Southwell method the gradient A is plotted against A/q (see Fig. 9c). The projected diver- tunnel entries over a period extending from 1962 gence dynamic pressure is simply the slope of a through 1968. During early phases of the program least squares straight-line fit of the measured the major focus of these studies was on T-tail data points.
flutter. To permit better simulation of struc- tural details believed to be important in T-tail These subcritical divergence prediction methods, flutter a relatively large model of the aft fuse- developed using inexpensive flat-plate models, were lage and T-tail ,,,as used. A view of this rather then employed with a high degree of success in unusual model suspended in the TDT test section is transonic tests of aeroelastically tailored sho,Yn in Figure 11.
forward-swept wing models as will be discussed later in the paper.
IV. VEHICLE DEVELOPMENT TESTING On the average about one,,·c'li.rd of all tests scheduled in the TDT ar~ t~ s~~~ort aerospace vehicle develop:'fien'c ',i.·O :ra::s. '.:Lese sturli.es involve dynam~.c, aero8~.as Licall~T sCf11e:-l ~'lojels of vehicle prototypes a:."H.i arr.:! COi.1(.~i1cted alm.l~~ ~,!it:l analyses to assur,= t:1al: n·=H \:l.~si(""',:"'.s {·.rill ' 1 8."JP.: ale·· quate margins of safety from flutcer and other aeroelastic problems, particularly in the impor- tant transonic speed range. As indicated in Fig- ure 10 such studies have supported most of the m~ajor high-performance aircraft developed in the U.S. since 1960. The TDT is utilized also in launch vehicle aeroelastic investigations such as buffet at transonic Mach numbers and ground wind load effects on the vehicle while erected on the Figure 11. Models used in C-14l flutter pad prior to launch. (3,4) clearance program.
To illustrate how aeroelastic model studies in Because the lower order vibration modes of the the TDT have been used to support development of wing and fuselage interact with the T-tail flutter flight vehicles, two examples selected from those mode, the structural dynamics of the forepart of shown in Figure 10 will be discussed briefly in the fuselage and wing were scaled using beams this section. These vehicles are the C-14l, a which simulate the mass and stiffness but not the military jet transport. developed in the early aerodynamic surfaces. Also, free rigid-body modes sixties, and the space shuttle, a hybrid airplane/ were simulated by supporting the model at its e.g. space launch vehicle whose recent highly successful by means of flexible rods and cables as shown in first flight represented the culmination of over 10 years of extensive research and development.
the figure.
The other two C-14l models were a complete cable-mounted flutter model shown in Figure 11 and a so-called "dummy" flutter model \Vhich \Vas simi- lar in size and \Veight to the flutter model but \Vhich \Vas much more rigid to avoid the risk of flutter. The use of such dummy models is highly were studies by Reed(25) and Hess(26) to determine desirable in flutter studies involving model sus- whether the shuttle, having a noncircular cross pension systems designed to simulate free-flight section and large-area surfaces normal to wind conditions.
flow while on the launch pad, might encounter cer- tain wind induced aeroelastic instabilities such Construction of the C-14l flutter model was the as "galloping"--a lateral instability of the "hole pod and spar method typically used in high-aspect- vehicle akin to the galloping phenomena of iced ratio flutter models. The stiffness is provided transmission lines--or "stop sign" flutter--a by the spar and removable pods provide the aero- longitudinal torsional instability of the whole dynamic surfaces. Removable weights within the vehicle associated with separated flow and akin to wing allowed different fuel loadings to be stall flutter of wings. Other early studies, sum- simulated.
marized by Runyan and Reed, (27) investigated transonic wing buffet characteristics associated In addition to satisfying the primary objective with reentry at angles of attack approaching of the test which was to demonstrate the required 90 degrees as well as the transonic flutter of flutter margin (132% of maximum dynamic pressure) wings in close proximity.
at Mach numbers up to the airplane maximum dive value, the flutter model also provided useful Once the shuttle configuration had essentially byproduct information such as, mentioned earlier, gelled, a series of aeroelastic model investiga- the development of new testing techniques and data tions in the TDT were undertaken. Starting in 1972, on aileron effectiveness and roll rate.
with ground wind load tests of an early configu- ration and flutter tests of semispan orbiter wing Space Shuttle model, the series was successfully concluded in Space shuttle, the product of a marriage of 1979 with flutter/buffet-model tests of the com- airplane and launch vehicle mated in a configura- pIe te space shu t tIe vehicle \,hich demons t ra ted the tion never before tried, stimulated intensive required flutter margin over the Mach number range interest and study in many areas of aeroelasticity.
0.6 to 1.15. Shown in Figure 12 is the family of Early exploratory wind-tunnel experiments using aeroelastic models used in these studies which simple models were made to assess the likelihood cover ground wind loads, flutter, buffet, and as a of the shuttle's encountering heretofore unimpor- spinoff mentioned earlier, stability derivative tant aeroelastic phenomena. Included, for example, measurements on the orbiter.
Figure 12. Space shuttle aeroelastic model studies in TDT.
V. AERO ELASTIC "FIXES" Flutter The effects of high angle of attack on flutter and buffet loads of the F-14 were investigated If aeroelastic-type problems are uncovered dur- using the cable-mounted flutter model. During ing the course of a vehicle development program, these tests, which preceded the first flight of model studies in the TDT often play a key role in finding and evaluating effective "fixes." In this the prototype, it was discovered that the flow section some examples are given in which such over the "over-wing" fairings caused the fairings, which were essentially cantilevered from a point fixes suggested by model tests were subsequently near the swing-wing hinge line, to deform and implemented on the full-scale counterparts.
oscillate in a "hand-clapping" manner. Several Ground Wind-Induced Oscillations potential fixes were evaluated. Of these a stiff- Two of these examples relate to the hazards of ening "strake" was proven to be effective on the model and was later incorporated in the airplane wind-induced oscillations of erected structures at the launch pad. In ground-wind-load studies using design. Figure 14 shOloJs the F-14 flutter model (without stiffeners) and the airplane with the a 0.03-scale model of the Saturn V-Apollo and stiffeners installed. umbilical tower, it was found that certain combi- nations of wind velocity (about 57 knots) and direction caused near sinusoidal oscillations that were severe enough to exceed the structural design limits of the vehicle. Various possible solutions to the problem were investigated such as the addi- tion of aerodynamic spoilers or damping devices to the structure (Farmer and Jones(4». Of these damping was selected as being the most feasible solution. Figure 13 shows the effectiveness of damping on reducing dynamic wind-induced loads to acceptably safe levels. This solution was then implemented on the vehicle in the form of a viscous damper strut connecting the Apollo capsule atop the vehicle to the adjacent umbilical tower structure. The damper remained attached to the vehicle until just before lift-off.
DESIGN BENDING MOMENT Figure 14. F-14 aeroelastic model tests RESULTANT LOAD ING J revealed flutter problem on over- wing fairing.
-<>- clcc = 0.0191 DYNAMIC VEHICLE --0- clc = 0028 LOADS BASE c .
BENDING --- STEADY LOADS Another example in which problems were first MOMENT uncovered and later solved by means of flutter model tests relates to flutter of wings with externally mounted stores. An extensive series of wind-tunnel tests of a one-quarter-scale F-16
w ~ m ~ ~ ~
flutter model was performed to define the flutter SIMULATED WIND VELOCITY, knots characteristics associated with various external store configurations the airplane must carry.
Figure 13. Wind-tunnel predicted response of From the large number of store configurations Saturn V to ground-wind loads.
tested a few were found to be flutter critical within the design operating envelope. As reported by Foughner and Bensinger~28) satisfactory solu- The other example of a wind-induced oscillation tions were developed, evaluated in the wind tunnel, problem concerns the Titan III umbilical mast, a and eventually implemented on the airplane. One ISO-foot "A" frame structure used in transporting of these flutter-critical store configurations the erected vehicle to the launch pad. Hurricane included air-to-air missiles mounted on launchers winds at Cape Kennedy induced violent oscillations at each wing tip and on pylons at the outboard of the mast to the extent that they rocked the under-wing station. Flutter occurred at a Mach base, damaged the supporting foundation piers, and number of about 1.1 near the required flutter mar- threatened the main structure. The wind-tunnel gin of safety boundary. On the basis of analytical investigation, using a 7-1/2 percent dynamically predictions three possible "fixes" were evaluated scaled model of the mast and transporter, revealed on the model: moving the under-wing missiles for- that by altering the upper quarter of the model ward, stiffening the missile launchers, and adding mast from a closed to an open-grid type structure ballast weight to the missile launchers. Of these the oscillations were greatly reduced. On the the latter solution was chosen for implementation basis of these findings, the Air Force implemented on the F-16. The addition of 4.7 pounds of bal- the solution on the actual mast at Cape Kennedy last weight in the wing tip launchers completely which significantly extended the range of wind con- eliminated the flutter tendencies of this store ditions under which Titan III ground-handling configuration.
operations can be conducted safely.
of safety from flutter. The engine mount systems Another wing-store flutter problem identified were redesigned to provide "fail-safe" redundan- in the F-16 model studies could be corrected cies such that the failure of anyone component in simply by revising the fuel usage sequence of the mount system would not cause flutter. Whirl externally mounted fuel tanks. These tanks are flutter avoidance has now also become a design compartmented into separate forward, mid, and aft consideration for prop-rotor V/STOL aircraft (see sections from which fuel can be drawn. The origi- nal fuel sequence, which was found to be flutter Kvaternik and Kohn(33» and modern wind turbine critical, called for first emptying the forward generators.
and aft compartments and then the center compart-
we "
ment. As shown in Figure 15, by merely reversing ~, , the order of fuel usage from the tank, i.e., by
,,,'
emptying the center compartment first, the flutter problem was eliminated.
~~ c::::::z.;1 )~
~
i'~Q "',, 1'" C"' ROm
l.D
FLUTTER BOUNDARY ~
<:---')""'.r,m·7:"7,7:,,!r---~ EMPTY FULL EMPTY 12 .8 .4 figure 16. Lockheed Electra propeller MACH whirl flutter model.
Figure 15. External tank fuel usage sequence change increases flutter speed of F-16 Nonlinear Aerodynamic Effects (Ref. 28). Although conventional linear aerodynamic theory predicts flutter to be independent of the steady- state aerodynamic loads, wind-tunnel and flight VI. AEROELASTIC "SURPRISES" tests have sometimes shown otherwise. During development testing of two different prototype During the development or early service life of aircraft, flutter-type instabilities were observed new vehicle configurations aeroelastic "surprises" in severe maneuvers whereas in earlier flight- are sometimes encountered which, to understand and flutter clearance tests at similar Mach numbers account for in future designs, warrant extensive and altitudes but in l-g flight, the flutter modes wind-tunnel testing and analysis. Three such sur- were well damped. In both instances the insta- prises encountered on new aircraft designs will be bility phenomena encountered in flight were later discussed in this section.
essentially duplicated in the TDT using flutter models of the prototype aircraft.
Propeller Whirl Flutter Soon after coming into service two Lockheed T-Tail Deflected-Elevator Flutter. The first Electra turbopropeller-driven transports were lost incident involved T-tail empennage flutter of a in mysterious accidents. The suspected cause of large cargo transport airplane, the C-141. The these accidents was a new form of aeroelastic instability occurred during high-altitude tests instability that had been discovered analytically at a Mach number near 0.8 but only in maneuvers in the late thirties by Taylor and Browne(29) in a when the elevator was deflected more than 8 degrees in either direction. The instability was charac- study of vibration isolation of aircraft engines terized by limited amplitude oscillations involv- but was found to be of no significance for air- ing coupling between elevator rotation and stabi- planes of that time. Later to become known as lizer torsion. Subsequent flight investigations propeller whirl flutter, the instability involves of various proposed solutions, including vortex a coupling of the aerodynamic and gyroscopic forces generators, dampers, and elevator mass balance, of the propeller with the stiffness and inertia led to the selection of mass balance. forces of the mount resulting in a precession-type motion of the propeller. A wind-tunnel model investigation was urgently carried out in the newly An experimental study using the existing C-14l commissioned TDT using a complete flutter model of high-speed flutter model was undertaken in the the airplane. Figure 16 shows the model in the TDT TDT. Results from this study by Sandford and test section mounted on a rod suspension system to Ruhlin(34) are summarized in Figure 17. It was simulate free flight. Results from the model tests found that the basic instability phenomenon by Abbott, Kelly, and Hampton(30) and analyses by encountered on the airplane in flight was repro- (31) (32) duced in the wind tunnel, although at higher Reed and Bland and by Houbolt and Reed speeds, and the elevator mass-balance solution for indicated that propeller whirl flutter could occur the airplane also eliminated flutter on the model.
at high forward speeds but only if the power plant support stiffness is severely reduced due to some form of damage to its mount structure. In an undamaged condition the airplane had ample margin AIRPLANE FLUTTER 400
Icel >8
FLI GHT ENVELOPE V , KEAS knots FLIGHT WIND TUNNEL WING SWEEP 55 DEG, W1Nii SWEEP 67 DEG, M= o.~] a::;3 DrG.
M~ 0.94 0= it) DlG.
W'NG 11P ,8 ,4 ,6 1.0 0 Ar.CEL.
MACH NUMBER Figure 17. Comparison of flight measurements and TIME model-predicted flutter of T-tail with SCALED fH,Q,: 2,2 Hl fUll SCALt ft1CQ .. : 7..1 tlZ deflected elevator (Ref. 34).
Figure 18. B-1 shock-induced instability studied in TDT.
Shock-Induced Flutter. The other incident involving self-excited oscillations due to maneu- vers occurred during high-altitude flight-load Wing/Store Flutter demonstration tests of the B-1 bomber. In this High-speed strike aircraft are required to instance the instability appeared as a limited carry an ever-increasing number of wing-mounted amplitude bending oscillation of the outer-wing external stores such as fuel tanks and armament.
panels. The oscillations occurred near critical Out of the many combinations of store loadings Mach number conditions for the airfoil and only at possible some inevitably cause reduction in the high positive angles of attack. A qualitative airplane's flutter speed and as a consequence explanation for these so-called "shock-induced restrict its operating envelope. Two promising self-excited bending oscillations" is given by concepts for alleviating wing/store flutter prob- Stevenson. (35) Some simplified calculations by lems have been demonstrated in recent investiga- Ashley(36) relate the instability to chordwise tions in the TDT using the models shown in Fig- shock movement with angle of attack and important ure 19. The first involves active controls tech- phase lags known to be present in the shock oscil- nology wherein the onset of flutter is sensed by lation. This instability does not represent a accelerometers on the wing which are used, by limitation to the B-1 as it occurs outside the means of computer-implemented feedback control range of normal flight operations.
laws, to activate control surfaces that produce aerodynamic forces to oppose flutter (see To study the phenomena further the B-1 flutter Refs. 37 and 38, for example). The second flutter model which was shown to be free of flutter prob- suppression method is a basically passive concept lems in earlier tests under simulated l-g condi- called the decoupler pylon which dynamically tions, was tested again in the TDT. As with the decouples the store from the wing by means of a C-14l model/flight correlation studies described self-aligning store suspension system that is soft previously, attempts were made to simulate the in pitch. (39) load factor and flight conditions for which the B-1 encountered shock-induced oscillations.
Results from both the flight and wind-tunnel tests are shown in Figure 18. Again, the instability phenomenon encountered in flight was demonstrated in the tunnel, although at slightly different con- ditions than in flight.
VII. CONTROLLING AEROELASTIC EFFECTS Aircraft design options for controlling aero- elastic effects can be broadened significantly through application of advanced concepts such as active control technology and aeroelastically tailored composite structure. Potential aero- elastic benefits made possible by these advanced ----'- technologies are being evaluated experimentally in the TDT using dynamic-elastic models. Three such research investigations have been selected for dis-
~~BfC~NGUi.ARJ'1i&i '
cussion in this section of the paper. These studies concern flutter suppression of wings with Figure 19. Wing/store flutter suppression external stores, helicopter vibration control, and studies in the Langley TDT.
divergence of forward swept wings.
repair of helicopter components. Oscillatory These active and passive flutter suppression loads transferred from the rotor to the airframe concepts have both been investigated throughout the are the primary contributors to airframe vibra- transonic speed range using aeroelastic models of tions. By means of active feedback control, the the YF-17 and the F-16. Some typical flutter-mode vibratory loads are reduced at their source--the damping trends measured on the F-16 model with rotor--in contrast to some conventional passive stores are shown in Figure 20. Damping measured means of vibration control which are designed to for an active flutter suppression system, the isolate the airframe from the vibration source. decoupler pylon, and, for comparison purposes, the In this concept, known as higher harmonic control, nominal design are plotted as a function of dynamic the rotor blade pitch angle is commanded to oscil- pressure and Mach number, holding the tunnel pres- late at the blade passage frequency (e.g., 4 per sure altitude constant. For the active system the revolution for a four-bladed rotor) with appro- flaperon control surfaces on each wing were actu- priate amplitude and phase so that the transmitted ated by signals which had been suitably filtered, loads are reduced to a minimum. By use of an mixed and fed back from accelerometers mounted on adaptive automatic control system and optimum con- each wing panel at the location indicated by the trol theory the blade pitch control inputs are sketch in Figure 20. For the decoupler-pylon case continuously updated to provide "optimum" vibra- only one of the three stores on each wing was so tion reduction under changing flight conditions. mounted, that being a GBU-8 guided bomb which was located between the tip-mounted missile and the Hind-tunnel model investigations of this con- inboard-mounted fuel tank. The figure indicates that for both systems the flutter mode was well cept have been reported by Hammond(41) and Molusis, damped and there were no signs of impending flutter Hammond, and Cline. (42) The model, shown in Fig- up to dynamic pressure nearly 100% greater than ure 21, is known as the Aeroelastic Rotor Experi- that for which the nominal design encountered mental System (ARES) and is an outgrowth of a flutter. Also, studies by Reed, Foughner, and model used in earlier rotor aeroelastic investiga- Runyan(40) have shown that in addition to increas- tions in the TDT. (9) The rotor control system ing the flutter speed, the decoupler pylon makes consists of a conventional swash plate that is flutter relatively insensitive to store center of remotely driven by three electro-hydraulic servo gravity and inertia changes.
actuators to provide the necessary high-frequency response characteristics. The closed-loop higher harmonic control system makes use of digital optimal control theory, the controlled quantities being either the vibratory forces and moments as measured by a strain-gage balance on which the rotor was mounted(41) or vibratory accelera- tions.(42) DAMP ING
~ ~~i~L:I~~ 5l
.4 LONG ITUD I NAL VERTI CAL 3
.... 4231t~
ACCELERATION.
ACCELERATION.' J
9 .2 .1 o .20 .30 .40 O~ AOVANCE RATIO .7 .6 .5 .9 .6 .7 .8 .5 LATERAL ~\ACII NU~\B[R ACCELERAT ION •. 4 9 .3 Figure 20. Damping trends of F-16 .2 model (Ref. 40).
.1 .2 .3 .4 The feasibility of using the decoupler pylon on ADVANCE RATIO the F-16 airplane has been studied under a NASA contract by General Dynamics, Fort Worth, in which Figure 21. Test results of closed-loop factors other than flutter, such as maneuver loads, adaptive controller for helicopter store ejection, gust response, etc., were con- vibration reduction (Ref. 42).
sidered. The study indicated the concept to be feasible for flight applications and a program has been initiated to evaluate the decoupler pylon on Also shown in Figure 21 are data indicating the an F-16.
effectiveness of the higher harmonic control in reducing the vibration response for various Helicopter Vibration Control advance ratios. It can be seen from these data that the system is highly effective in reducing Another application of active control technology the vibration levels in the vertical and hori- under study in the TDT concerns the reduction of zontal directions over the entire range of advance helicopter vibration through rotor loads control.
ratios tested. The lateral vibration results are Most present-day helicopters experience excessive mixed, the controller being effective at the vibrations in some regions of operation. These higher advance ratios but causing increased vibrations lead to pilot fatigue, passenger annoy- response at lower advance ratios. These reduc- ance, and the need for frequent maintenance and tions in vibration level were accompanied by increased blade and pitch link loads but they were small relative to design limits for the model.
These tests are believed to be the first time use of an adaptive control system employing optimal control theory for such purpose. Preparations are underway to demonstrate the system in flight on an OH-6A helicopter.
Divergence of Fonmrd-S,'/ept Wings "0 ''0 As with active controls, the use of aeroelasti-
~.l
.
cally tailored composite structure opens up new '00 '00 >--.
I ANA1Y~S / D'N
design options for controlling aeroelastic response. / .. -
1M 160
.. '"
EXmlMlNl ANALYSIS / >sf In this approach the orientation of the fibers in SU8ClUTl<Al 120 12.
no<NlQ\J' fibrous composite materials in wing skins becomes EXPEIUMENJ
•• SU8C1UTlCAt
a design variable by means of which aeroelasticity HCHNIQUE may be put to beneficial uses. A noteworthy , , .~ 0 0 example is the recent rekindling of interest in ., 0 .. .0 .. 1.0 . , ..
0 .. -8 to 1.'2
I.'
MAOf NVMNR .~Q!~~~ forward-swept wings made possible by the applica- tion of aeroelastically tailored composites. By Figure 22. Divergence tests in TDT of such means Krone(43) showed in analytical studies aeroelastically tailored forward- that the low static divergence speed of forward- swept wings (Re f. 44).
swept wings, a problem which had originally led to rejection of the concept, could now be avoided with relatively small structural weight penalties. there is much need for experimental data on surface pressures and flow field measurements on bodies and This rebirth of interest in forward-swept wings lifting surfaces. Such measurements are needed to brought about by aeroelastic tailoring technology gain better understanding of the physics of aero- has led to considerable research effort by the Air dynamic flow fields about realistic configurations Force Flight Dynamics Laboratory, the Defense at transonic speeds, and to obtain a data base for Advanced Research Projects Office, and NASA to use in the validation of aerodynamic theories and better understand the aeroelastic characteristics in the design of active controls.
of forward-swept wings. Aeroelastically tailored models of forward-swept wing designs by Grumman Until recently most of the unsteady pressure Aerospace Corp. and Rockwell International were measurement work had been performed in European tested in the TDT to provide correlation with wind tunnels (notably by the NLR in The analysis and to gain confidence that divergence can Netherlands, ON ERA in France, RAE in England, and be efficiently avoided within the design flight DFVLR in Germany), whereas in the United States, envelope. Results from these tests together with the emphasis was on large aeroelastic models. At a discussion of the evolution and background of present, there is a sizable program underway and aeroelastic tailoring as applied to forward-swept planned for measuring steady and oscillating pres- wings have been gresented by Hertz, Shirk, Ricketts, sure distributions on a variety of lifting sur- faces at Langley in the TDT. In a recent survey and Weisshaar. (4) Photographs of the O.S-scale Grumman model and the O.6-scale Rockwell model paper on unsteady pressure measurement programs mounted in the wind tunnel are shown in Figure 22.
conducted in European wind tunnels, 0Isen(4S) Also shown in the figure are the experimental observes that this new interest in unsteady pres- divergence speed boundaries (projected from test sure measurements in the U.S. appears to be matched data at speeds below the divergence speed as by a growth of interest in flutter model testing described in Reference 23) and calculated diver- in Europe.
gence boundaries. Thus far divergence has been the major focus of aeroelastic studies of forward- The planforms presently under investigation in swept wings. Other aeroelastic characteristics the TDT include a clipped delta-wing model of such as buffet, flutter, and gust response, will interest for supersonic transport and fighter no doubt become the subject of future studies for applications and a high-aspect-ratio (10.7) swept this class of aircraft.
wing of interest for energy efficient transports with active controls. Both models are equipped VIII. UNSTEADY PRESSURE MEASUREMENTS with active leading- and trailing-edge active con- trol surfaces and a large number of static orifices Up to this point, the paper has dealt with the and dynamic pressure transducers on their upper use of scaled flexible models to study various and lower surfaces.
aeroelastic stability and response problems. These models play the role of mechanical analogs of the The test program involving the high-aspect-ratio full-scale structure. When mounted in a wind tun- model is described, and some selected results from nel that properly simulates the flow field, such initial wind-tunnel tests ara presented by models perform the difficult time and space inte- Sandford, Ricketts, Cazier, and Cunningham. (46) grations of the aerodynamic loads to produce This model has five leading-edge and five trailing- response characteristics that would be expected on edge control surfaces which can be oscillated the full-scale article. Whereas tests of this kind individually and in various combinations. It is show the net effects of the aerodynamic flow field instrumented with 2S2 static orifices and 164 in around the model, they give little insight into the situ dynamic pressure transducers. Figure 23 shows physics of the flow itself. Because unsteady aero- a photograph of the model mounted in the TDT and dynamic forces, particularly in the transonic range, some unsteady pressure measurements associated with are probably the weakest link in the chain of tech- the oscillation of an outboard leading-edge and a nologies used in aeroelastic design and analysis, trailing-edge control surface. The data are for operational more than two decades ago. The many- faceted aspects of experimental aeroelasticity were illustrated by way of examples selected from a broad base of testing experience in the TDT.
The coming decade should see continued production use of this unique National aerospace facility.
X. ACKNOWLEDGE~illNTS The author is indebted to Mr. Andrew D. Carey of the Photographic Branch and Mr. John H. Cline of the Configuration Aeroelasticity Branch for their untiring efforts devoted to the preparation 100r.
of film clips used in the oral version of this paper.
:~~~[O~~£:: -100-·· ....
XI. REFERENCES ±4" 1. Garrick, I. E., and Reed, W. H. III: Histori- .81k .15c .80c .15c cal Development of Aircraft Flutter. AIAA Jour. of Aircraft, Vol. 18, No. 11, November 1981. Also available as AIAA 6CP~~ .
6CPb~~~~
paper no. 8l-059l-CP.
o •.. "-->-40 o fRACTION OF CHORD 1.0 fRACTION Of CHORO 2. Regier, A. A.: The Use of Scaled Dynamic Models in Several Aerospace Vehicle Studies.
Figure 23. Unsteady pressure measurements on Proceedings of AS ME Winter Annual Meeting, model with oscillating control surfaces.
S~nposium on Use of Models and Scaling in Shock and Vibration. Philadelphia, Pa.
Edited by Welfred and Baker, published by Mach number 0.6 and a reduced frequency, based on ASME, N.Y., 1963.
root semichord, 6f k = 0.21. The calculated results,which show good correlation with experi- 3. Hanson, P. W., and Jones, G. W., Jr.: On the ment everywhere except in the regions of the con- Use df Dynamic Models for Studying Launch trol surface hinge, were obtained using a subsonic Vehicle Buffet and Ground Wind Loads. Pro- kernel function method (RHO IV) for wings with ceedings of Symposium on Aeroelastic and oscillating controls. (47) Dynamic Modeling Technology, U.S. Air Force Systems Command, RTD-TDR-63-4l97, Part 1, Other experimental studies underway at Langley 1964.
include unsteady pressure measurements on a two- dimensional dynamic-stall model that can be oscil- 4. Farmer, M. G., and Jones, G. 1-1., Jr.: Summary lated at large pitch amplitudes in order to simu- of Langley Wind Tunnel Studies of Ground- late the dynamic-stall response of helicopter Wind Loads on Launch Vehicles. Compilation blades. Also, steady and unsteady pressures are to of Papers Presented at the NASA LRC Meeting be measured for a pitching rectangular-wing model.
on Ground Wind Load Problems in Relation to This rectangular wing is especially suited for use Launch Vehicles, TMX-57779, June 7-8, 1966.
in validating 3-D transonic unsteady aerodynamic theories which are currently limited to simple wing 5. Guyett, P. R.: The·Use of Flexible Models in planforms.
Aerospace Engineering. Royal Aircraft Establishment Technical Report No. 66335, Finally, mention should be made here of a coop- Nov. 1966.
erative unsteady pressure measurement project between Lockheed, the U.S. Air Force, NLR (The 6. Abbott, F. T.: Some Current Techniques in Netherlands), and NASA referred to as the "LANN" Experimental Aeroelasticity. Proceedings of wing. This wing has a supercritical airfoil sec- ASME 1967 Winter Annual Meeting Symposium on tion and is highly instrumented for measuring sur- Solid-Fluid Interaction Problems in face pressures. The wing is scheduled for testing Mechanics, Nov. 12-16, 1967.
at NLR in late 1981 and in the Langley National Transonic Facility, at flight Reynolds numbers, in 7. Rainey, A. G., and Abel, I.: Wind Tunnel 1983. The data will be useful for examining Techniques for the Study of Aeroelastic Reynolds number effects on steady and unsteady Effects on Aircraft Stability, Control, and aerodynamics of supercritical wings. These tests Loads. AGARD Proceedings No. 46 - Aero- will require significant advances in the state of elastic Effects from a Flight Mechanics the art of unsteady pressure measurements due to Standpoint, Marseilles, France, April 21-24, the high dynamic pressures and the extremely low 1969.
temperatures required to achieve high Reynolds numbers.
8. Reed, W. H. III: Comparisons of Flight Measurements with Predictions from Aero- IX. CONCLUDING REMARKS elastic Models in the NASA Langley Transonic Dynamics Tunnel. AGARD Conference Proceed- This paper has documented that aeroelasticity ing No. 187 on Flight/Ground Testing Facili- matters by presenting a cross section of aero- ties Correlation, Valloire Savoie, France, elastic research and testing performed in the NASA 1975.
Langley Transonic Dynamics Tunnel since it became 22. Many Authors, Flutter Testing Techniques, 9. Hammond, C. E., and Weller, W. H.: Wind-Tunnel Testing of Aeroelastically Scaled Helicopter NASA SP-4l5. 1976.
Rotor Models. Presented at the 1976 Army 23. Ricketts, R. H., and Doggett, R. V., Jr.: Science Conference, West Point, New York, June 22-25, 1976. Wind Tunnel Experiments on Divergence of Forward-Swept Wings. NASA TP-1685, Aug.
10. Cole, P. H.: Wind Tunnel Real-Time Data 1980.
Acquisition System. NASA TM-8008l, April 1979. 24. Flax, A. H.: The Influence of Structural Deformation on Airplane Characteristics.
11. Hanson, P. W.: Evaluation of an Aeroelastic Jour. of the Aero. Sci., Vol. 12, Jan 1945, Model Technique for Predicting Airplane pp. 94-102.
Buffet Loads. NASA TN D-7066, 1973.
25. Reed, W. H. III: Wind Effects on Space 12. Reed, W. H. III, and Abbott, F. T., Jr.: A Shuttle Vehicles Erected on Launch Pad.
New Free-Flight Mount System for High-Speed Proceedings of the Third International Con- Wind-Tunnel Flutter Models. Proceedings of ference on Wind Effects on Buildings and Symposium on Aeroelastic and Dynamic Modeling Structures, Japanese Organizing Comm., cds., Technology. RTD-TDR-63-4l97, Pt. I, U.S. Air Saiko Co., Ltd. (Tokyo), 1971., Force, Mar. 1964, pp. 169-206. pp. 1127-1140.
26. Hess, R. W.: Wind-Tunnel Roll-Damping Measure- 13. Gilman, J., Jr., and Bennett, R. M.: A Wind Tunnel Technique for Measuring Frequency- ments of a Winged Space Shuttle Configuration Response Functions for Gust Load Analyses. in Launch Attitude. NASA TN D-7394, Dec.
J. Aircraft, Vol. 3, No.6, Nov.-Dec. 1966, 1973.
pp. 535-540.
27. Runyan, H. L., and Reed, W. H. III: Shuttle 14. Barbero, P., and Chin, J.: Users Guide for a Dynamics and Aeroelasticity--An Appraisal.
Computer Program to Analyze the LRC 16' AlAA Astronautics and Aeronautics, Vol. 9, Transonic Dynamics Tunnel Cable-Mount System.
No.2, Feb. 1971, pp. 48-57.
NASA CR-1323l3, 1973.
28. Foughner, J. T., Jr., and Bensinger, C. T.: 15. Hanson, P. W.: The Prediction of Structural F-16 Flutter Model Studies with External IHng Response to Buffet Flow: A State-of-the-Art Store. NASA TM-74078, Oct. 1977.
Review. NASA TM X-72627, 1974. Also see Proceedings of the 39th AGARD Structural and 29. Taylor, E. S., and Browne, K. A.: Vibration Materials Meeting, Munich, Germany, Isolation of Aircraft Power Plants. Jour.
Oct. 7-12, 1974.
Aero. Sci., Vol. 6, Dec. 1938, pp. 43-49.
16. Houbolt, J. C., Steiner, R., and Pratt, K. G.: 30. Abbott, F. T., Jr., Kelly, H. N., and Hampton, Dynamic Response of Airplanes to Atmospheric K. D.: Investigation of Propeller-Power- Turbulence Including Flight Data on Input and Plant Auto-Precession Boundaries for a Response. NASA TR R-199, 1964. Dynamic-Aeroelastic Model of a Four-Engine Turboprop Transport Airplane. NASA TN 17. Redd, L. T., Hanson, P. W., and Wynne, E. C.: D-1806, June 1963.
Evaluation of a Wind-Tunnel Gust Response Technique Including Correlations with 31. Reed, W. H. III, and Bland, S. R.: An Analyti- Analytical and Flight Tests. NASA TP-150l, cal Treatment of Aircraft Propeller Preces- Nov. 1979. sion Instability. NASA TN D-659, Jan. 1961.
18. Abel, I.: Evaluation of a Technique for Deter- 32. Houbolt, J. C., and Reed, W. H. III: mining Airplane Aileron Effectiveness and Propeller-Nacelle Whirl Flutter. Jour.
Roll Rate by Using an Aeroelastically Scaled Aerospace Sci., Vol. 29, March 1962, Model. NASA TN D-5538, Nov. 1969.
pp. 333-347.
19. Grosser, W. F.: A Transonic Speed Wind Tunnel Kvaternik, R. G., and Kohn, J. S.: An Experi- 33.
Investigation of the Rolling Effectiveness of mental and Analytical Investigation of a Large Swept-Wing Transport Aircraft with Proprotor Whirl Flutter. NASA TP-l047, Conventional-Type Ailerons and Various Dec. 1977.
Spoiler Configurations, AlAA paper No. 65-789, Nov. 1965.
34. Sandford, M. C., and Ruhlin, C. L.: Wind- Tunnel Study of Deflected-Elevator Flutter 20. Bennett, R. M., Farmer, M. G., Mohr, R. L., and Encountered on a T-Tail Airplane. NASA Hall, W. E., Jr .• :. Wind Tunnel Technique for TN D-5024, Feb. 1969.
Determining Stability Derivatives from Cable Mounted Models. J. Aircraft, Vol. 15, No.5, Stevenson, J. R~: Shock-Induced Self-Excited 35.
May 1978, pp. 304-310.
Airfoil Bending Oscillations. Los Angeles ,Div., Rockwell International, paper pre- 21. Chin, J., and Barbero, P.: User's Guide for a sented to Aerospace Flutter and Dynamics Revised Computer Program to Analyze the LRC Council, Las Vegas, Nev., Oct. 19-20, 1978.
Transonic Dynamics Tunnel Active Cable-Mount System. NASA CR-132692, 1975.
36. Ashley, H.: Role of Shocks in the "Sub- 47. Rowe, W. S., Sebastian, J. D., and Petrarca, Transonic" Flutter Phenomenon. Jour. of J. R.: Reduction of Computer Usage Costs in Aircraft, Vol. 17, No. 3, Mar~h 1980, Predicting Unsteady Aerodynamic Loading pp. 187-197.
Caused by Control Surface Hotions - Analysis and Results. NASA CR-30009, Jan. 1979.
37. Noll, T. E., Huttsell, L. J., and Cooley, D. E.: Investigation of International Con- trol Laws for Wing/Store Flutter Suppression.
A Collection of Technical P~pers - AlAA/ASME/ ASCE/AHS 21st Structures, Structural Dynamics and Materials Conference, Part 2, May 1980, pp. 585-594. (Available as AIAA-80-0764.)
38. Peloubet, R. P., Jr., Haller, R. L., and Bolding, R. M.: F-16 Flutter Suppression System Investigation. A Collection of Technical Papers - AlAA/ASME/ASCE/AHS 21st Structures, Structural Dynamics and Materials Conference, Part 2, May 1980, pp. 620-634.
(Available as AIAA-80-0768.)
39. Reed, W. n., III, Foughner, Jerome T., Jr., and Runyan, Harry L., Jr.: Decoupler Pylon: A Simple, Effective Wing/Store Flutter Sup- pressor. J. Aircraft, Vol. 17, No.3, March 1980, pp. 206-211.
40. Reed, W. H. III, Cazier, F. W., Jr., and Foughner, J. T., Jr.: Passive Control of Wing/Store Flutter. NASA TM-8l865, Dec.
1980.
41. Hammond, C. E.: Wind Tunnel Results Showing Rotor Vibratory Loads Reduction Using Higher Harmonic Blade Pitch. A Collection of Technical Papers - AlAA/ASME/ASCE/AHS 21st Structures, Structural Dynamics, and Materials Conference, Part 2, May 1980.
(Available as AlAA 80-0667.)
42. Molusis, J. A., Hammond, C. E., and Cline, J. H.: A Unified Approach to the Optimal Design of Adaptive and Gain Scheduled Con- trollers to Achieve Minimum Helicopter Rotor Vibration. Presented at the 37th Annual AHS Forum, New Orleans, LA, May 1981. (Available as AHS preprint 81-20.)
43. Krone, N. J., Jr.: Divergence Elimination with Advanced Composites. AlAA paper No. 75-1009, Presented at AlAA 1975 Aircraft Systems and Technology Meeting, Los Angeles, August 1975.
44. Hertz, T. J., Shirk, M. H., Ricketts, R. H., and Weisshaar, T. A.: Aeroelastic Tailoring with Composites Applied to Forward Swept Wings. Presented at Fifth DOD/NASA Confer- ence on Fibrous Composites in Structural Design, New Orleans, LA, Jan. 27-29, 1981.
45. Olsen, J. J.: Unsteady Pressure Measurements on Oscillating Models in European Wind Tunnels. Tech. Memo. AFDV AL-TM-80-l-FIBR, Harch 1980.
46. Sandford, H. C., Ricketts, R. H., Cazier, F. W' Jr., and Cunningham, H. J.: Tran- o sonic Unsteady Airloads on an Energy Effi- cient Transport Wing with Oscillating Control Surfaces. Journal of Aircraft, Vol. 18, July 1981, pp. 557-561.
1. Report No. 3. Recipient's Catalog No.
I 2. Government Accession No.
NASA TM-83210
4. Title and Subtitle 5. Report Da te
Aeroe1asticity Matters: Some Reflections on Two Decades September 1981
of Testing in the NASA Langley Transonic Dynamics Tunnel 6. Performing Organization Code
505-33-53-01
7. Author(s) 8. Performing Organization Report No.
Wilmer H. Reed III
10. Work Unit No.
9. Performing Organization Name and Address
NASA Langley Research Center
11. Contract or Grant No.
Hampton, Virginia 23665
13. Type of Report and Period Covered 12. Sponsoring Agency Name and Address
Technical Memorandum
National Aeronautics and Space Administration
14. Sponsoring Agency Code
Washington, DC 20546
15. Su pplementary Notes
Presented at International Symposium on Aeroe1asticity
Nuremberg, Germany, October 5-7, 1981
16. Abstract
Testing of wind-tunnel aeroe1astic models has become a well established, widely used
means of studying flutter trends, validating theory and investigating flutter margins
of safety of new vehicle designs. The Langley Transonic Dynamics Tunnel was designed
specifically for work on dynamics and aeroe1astic problems of aircraft and space
vehicles. This paper presents a cross section of aeroe1astic research and testing
in the facility since it became operational more than two decades ago. The paper
illustrates, by means of examples selected from a large store of experience, the
nature and purpose of some major areas of work performed in the Transonic Dynamics
Tunnel. These areas include: specialized experimental techniques; development
testing of new aircraft and launch vehicle designs; evaluation of proposed "fixes" to
solve aeroe1astic problems uncovered during development testing; study of unexpected
aeroe1astic phenomena (Le., "surprises"); control of aeroelastic effects by active
and passive means; and, finally, fundamental research involving measurement of
unsteady pressures on oscillating wings and control surfaces.
'7. Key Words (Suggested by Author(s) I 18. Distribution Statement
Flutter
Unclassified-Unlimited
Subject Category 02
Aeroe1asticity
19. Security Classif. (of this report) 20. Security Classif. (of this page) 22. Price· 21. No. of Pages
Unclassified
Unclassified
17 A02
For sale by the National Technical Information Service, Springfield, Virginia 22161 NASA-Langley, 1981