MINISTRY OF AVIATION
AERONAUTICAL RESEARCH COUNClL
CURRENT PAPERS
Measurements of Buffeting on Slender Wing Models
bY
D. G. Mabey, M.Sc.(Eng.)
LONDON. HER MAJESTY’S STATIONERY OFFICE
1967
TEN SHILLINGS NET
U.D.C. bow 533.6.013.43 : 533.693.3 : 532.527 : 533.6.011.32/34
C.P. No.9174
March IV66
hiEASUREZ?JXTS OF BIJFFETING ON SLENDER 'iKlXG MODELS
by
D. G. Mnbey, M.Sc.(Eng.)
Tests on two solid slender mng models show that there are two stsges in
the mng buff'eting at subsonm speeds;
(1) a mild buffeting due to leading-edge vortices, end
(4 a severe buffeting at vortex breskdom.
At supersonic speeds the mild buffeting is weaker, probably because of
the reduced size and intensity of the leading-edge vortices.
Tests on two aeroelastic mcdels at subsonic speeds show that the mild
buffeting is principally at the wing jrd or higher distortion modes. The order
of magnitude of this buffeting suggests that on a supersonio transport aircraft
this might be pcrceptlblc during subsonic climb-out at high I&$.
*Replaces R.A.d. Tech. Report No.66086 - A.R.C.27923.
2
CONTENTS
1 lKI'RU!XlCTION
2 EliPEWl~ISNl'AL DSTAIIS
2.1 Alcdels
2.2 Measurements
2.3 Test conditions
3 RESULTS
3.1 MC&l 1
3.2 1;l0aei 2
3.j Model 3
3.4 Node1 4
4 DISCUSSION
5 CONCLUSIONS
Acknowledgements
Appendix A Tip buffetmg stresses and amplitudes
Table 1 Test conditions
S.ynbols
References
Illustrations
Detachable abstract cards
&
3
3
3
5
5
5
5
8
a
9
10
12
12
13
15
16
17
Figures l-22
3
1 INTRODUCTI'3J -
The flow at mnci&nce over slondcr v+g aircraft with sharp leading-edges
is dominated by a paw of vortices which strcngly influence the aircraft hadingS.
The effect of the vortices on the steady loads is well known'. Flowever the effect
on buffet* characteristics is uncertain, although significant pressure flUCha-
tions have been measured under the leading-edge vortices*. No wing buf'feting has
been reported on the low speed HP.115 but a very light buffeting has been
reported on the high speed Bristol 221 at subsonic speeds.
The presont wind tunnel tests, on two solid models (Models 1 and 2) were
made to lnvcstigate the basic nature of slender vring buffeting. The measurements
consisted of the variation of unstoaZy wing-root stress ltith mcidcncc. Using
thx method two separate stages of "ing buffeting were identified,
(1) a mild buffeting associated with the cxistencc of Isating-edge
vortices, and
(2) 3 a strong buffeting foll&ing vortex brcskdo~vn .
Accurate mcas+rcmcnts of buf'foting on solid wind tunnel models are diffi-
cult bccauso the models normally have unrepresentative aerodynamic and structural
damping coefficients and unrcprosontatlvc distortion mcdcs. Hence tho two aorc-
clastlc models (Ncdels* 3 and 4) wore also tested.
This report proscnts buffeting measurements on all f'cur wings and briefly
discusses their possiblo implicatiord.
2 CXWRINEWM.L DETAILS
2.1 I.Iodels
Buffeting measurcmonts wro .mdc on fdur slcndor wing models ~onstructod
for other static and dynamic tests during tho evolution of a supersonic trans-
port dosign. The cxtcrnd shapos of those models differ significantly from
the Concord prototype and production aircraft.
Flg.1 shows the general arrangomont of Model 1. This is a l/75 scale
solid bronze model of an early supersonic transport dosign. Four semi-conductor
strain gauges at the wing-root wore wired so that the signals from the port and
starboard xvlngs wcrc adtitlvc for symmetric doformations of the model, while
signals resulting from antisymnotric doformations nor‘c virtually eliminated.
Before the buffeting tests the model was ground rescnanco tostcd on its support-
ing sting. Idcntiflcd symwtric distortion modes arc sho:m in Fig.2. The
'Throughout this report buffet will bo dcfincd ns the applied acro&ynamic excitation and buffeting as the structural response. The sonsitivo semi- conductor strain gauges used on those models are probably able to detect buffoting which would bo impcrccptiblc to passongors.
4
wind-off structural &mp.ing is lcw in both of these modes because of the solid
model construction.
Pig.3 show the generzil arrangement of Model. 2. This is a l/30 scale solid
steelmcdel of a supersonic transpart design. Model 2 was fitted with four.
strain gauges on tine port wing so that symmetric and sntisymmetric defcrm?ticns
mould produce a signal. Distortion mcdes identified in a ground resonance test
are shovm in Flg.4.
F1g.5 shcws a photograph of Model 3, which is generally stilar in external
shape to Models 1 and 2. This is a l/55 scale flutter model of-an early super- sonic transport design. The aircraft skin, spars and ribs are represented by
etching from light alloy; the volume within the model is stabilised with
?xpancied plastic foam. The model could be mounted on Its flutter excitation
support so that total damping measurements (wind on) could be made after the
buffeting measurements.
Wire strain gauges were provided on this model for flutter tests with cdy
two active gauges in each bridge at the port tip and at tke pert wing-rmt SC
thd both sg?rmetrlc arCi antisyrmnetrio deformation 3 would prcduce a signal as on MC&A 2. (Semi-ccncluctcr strain gauges ~culd have been preferred but could not
be applied without altering the model inert&s.) Some syz!metric and snti-
symmetric distortion m&es are shown u1 F'lg.6. (The frequencies measure&in the
initial ground resonance .test are slightly different from those subsequently
measured in the tunnel because of small Liffcrences in the support and excita-
tion systems.) The wind-off structural damping is higher than.onMcdels 1 snd 2
and more typical of probable full scale values.
Fig.7 s@s the general arrsngcment of Hodel 4. This is a defknsticn model o?' an early superscnic trsnspc& design. The mcdel'ccnstruction was
- . ~i.mi1e.r to that of Ikdd J and the model v,as mcunted on a solid sting. Semi- conckctor strain gauges were applied directly to the outside skin of this
deformation model (where the inaktia distributions here not critical) and then covered with a thin srsldite fairing. .Twc bridges, caoh with four active gauges were provided near the wing-root sncl at 5C$$ semi-span as shotm in Fig.7. The
gauges were wired to add the port snd stsrbosrd signals as on Model 1. The
epproximate distortion modes which appeared in the buffet signals eze shown in Fig.8. The total damping (wind on) codd not be measured because no provision
was made for exciting this m&.el.
5
It should be noted that the "full scale" frequencies of Figs.2, 4 and 8
are simply obtained by multiplymg the model frequencies by the model scale,
and are not &redly related to a specific aircraft. The values for Model 3
me, of course, related to a specifio aircraft.
2.2 Xeasurements
The rms wing buffeting signals on all four wings were measured on a
spectrum analyser (the tuned signals were measured with 6$ bandwidth). To
avoid overstressing models 3 snd 4 (9 and 30 lb msxknum lift respectively)
the steady lift and pitching mament at every kinetic pessure were monitared
before the buffeting tests.
The total damping measurements on&de1 3 were derived from veotar plots of model response to meohanical excitation at particular frequenoies.
2.3 Test ocditions
The models were tested in tither the 13 x 9 ft (low speed) or the
3 x 3 ft (subsonic/trans~ic/supersonic) wind tunnel. Table j gives the test
conaitions. The low Reynolds number of the Ncdel 3 and 4 tests will be noted.
Ballotini (small glass spheres) were applied in narrow bands to fix transition
on Models 1, 2 and 4 but not on Model 3 (to avoid altering the model inertia
distributions).
3 RESULTS
3.1 Model 1
Model 1 was first briefly tdsted in the 3 ft tunnel snd a mild buffeting
was detected at about 6' incidence (a) at allMach numbers (M) from 0.40 to 0.80. Above lkO.80 this mild buffeting was not detected over the inoreased
signal arising from tunnel unsteadiness. Oil flow photographs (Fig.9) show
that this buffeting is associated with the existence of the leading-edge
primary vortex. Owing to some asynrmetry this vortex appears first on the port
wing at a = 6' (Fig. y(a)). As incidence increases to (1. = 7' there is an
intermediate flow with several spanwise vortices near both tips (Fig. T(b)).
At a = 8' these vortloes hnve combined to form one single vortex on the port
wing; there are still two vortices on the starboard wing (Fig. T(o)).
Model1 could not be tested at incidences much higher than a = 14' in
the 3 ft tie1 because of excessive blockage, even 111 the 36 x 26 in slotted
working seotion. Hence the tests were extended to the 13 x 9 ft low speed wind
tunnel, where blockage effects would be insignificant even at a = 30'.
6
Fig,10 shows the variation of total wing-root strain signtil from a = 1 O to
a = 3o” in t1,e 13 h 9 ft talel*. The 0-e comprises several different sections
which are related with the flow develo,pment. From a = 1 ' to a = 6' there is a
constant signal owing to the flow unsteadiness induced by the support and the
support vibration**. Then from a = 6’ to a = IO0 the szgnal builds up SteadilY
.a8 the leading-edge vortices develop. era a = 10' t0 D = 22.5’ there 1s a-
constant signal implying "a plateau" in the buffeting. From a = 22.5' to
a = 26’ the signal builds up rapidly as the vortex breakdmn point, clearly
visible by condensation in Lthe vortex core, moves upstream from the trailang-edge
on to the port wing. From a = 26' to a = 29' thore is ti further sudden build up
of signal Thich mJst have been caused by vortex breakdavn on the starboard wing.
This vortex breskdu~.was not shown by condensation in the core, probably because
the core was larger than for the port wing. From.a e 29' to 30' the buffeting
falls rapidly.
The wing buffeting was next er&ed in more detazl by measuring its
response %lhen tuned to the frequency of the distortion modes. Fig.11 shows
the variation of tuned wing-root strain signal with incidence. The major
portion of the buffeting at vortex break&mm is clearly associated with the wing
response at its fundsmental frequency of 300 c/s (Fig. II(a)). (This response may come from a larger low frequency componen t in the excitation or to the larger
scale of the disturbance.) The fundamental mode is not significantly excited by
the formation of the leading-edge vortex; most of the response from a = 6' to
a = IO0 comes from the thrrd deformation mode at 700 c/s (Flg.ll(b)). This mode
is also excited by vortex breakdown, but not as strongly as the fundamental made. . A mode at 1330 c/s was also weakly excited by both vortex formation and vortex
breakdown (Fig.ll(c)). t
Sideslip significantly lowers the incidence far vortex breakdovm; the
vortex breaks down first in the.letig wing on which the sweepbaok is reduced.
Thus at F.10' sideslip (possible limiting values for oross wind landings), vortex
breakdown occurs at about 15' inoidence (Fig.12).
Model 1 was then tested again in the 3 ft tunnel to investigate buffeting
over a wide range of density at subsonic and supersonic speeds,
b This test was made in two stages, from(t=1°t060andthenfroma~60 because of the support geometry. ** Subsequent experience with Model 4 showed that this relatiwly high signal was associated ~lth the unsteadiness produced by the bluff mcdel support, rather than the tunnel unsteadiness, which was very low.
7
The tests atM = 0.50 clearly showed that buffeting could be detected
(F1g.13) at stream densities as low as p = 0.0091 lb/IX3 (a tunnel total
pressure of 4 in Hg absolute). Hence buffeting tests at this density on
Model 3 (a fragile aeroelastic model) would be possible if the dampings On both
models were comparable. The sensitivity of the semi-conductor wmg-root strain
gauge bridge on Model I was only twice that of the wire strain gauge bridge Of
Model 3 because Model 1 was much stiffer than Model 3. An interestirg feature
of these results not noticed III the previous tests was the small peak in the
signal at a = 7.5'. A corresponding peak was subsequently found on Model 3 at
'7.0' incidence (3.3).
Another interesting feature of these results is that at the higher kinetic
pressures the wing fundsmental mode at 300 c/s was excxted by the mild buffeting
on vortex formation (Fig.14). The other modes at 700 and 1300 c/s were also
exalted together with sn additional mode at 1920 C/S. Although the particular
modes detected depend critically on the somewhatarbitraryposition of the strain
gauges, excitation of modes over this frequency range suggests that the excita-
tion spectrum associated with the f'ormaticn of the leading-edge vortex is
relatively flat for values of *E/v from 2.0 to 13.0.
The first supersonic test was made atM = 1.4, the lowest Mach number at
which reflected shock waves were clear of the model. The buffeting onset was not
sharply defined, except at the highest density (Fig.15). However there was a
signal peak at a = 8' which recalls that at a = 7', M = 0.50, (Fig.13). A vapour 5 screen m the working section revealed an upper surface vortex at incidences
higher than 7'. No vortex instability was apparent in the vapour screen at 0 iXZ-2.
'I Fig.15 also shows that the "plateau level of buffeting is not muoh higher
than the signal at zero incidence. The buffeting at supersonic speeds 1s prob-
ably less than at subsonic speeds because the leading-edge vortices are weaker.
(The leading-edge vortices develop much less non-linear lift at supersonic
speeds; see e.g. Ref.6.) The leading-edge vortices are also much smaller at
supersonic speeds as Rig.33 of Ref.7 clearly demonstrates.
Tests atM = 1.6 were inconclusive. The signal at zero incidence was much
hxgher (probably because of quadrant vibration excited by the tunnel ah&k) and
no signal peak was detected.
a
3.2 Model 2
In broad outline the buffeting oharacteristic of Iviodel 2 (Fig.16) resemble
those of Model 1 (Pi&IO). On Model 2 the existenoe of a vortex is apparent at
a = GO, there is a small signal peak at a = 8' and a "plateau" in the signal
from a = 12O to a = 21°. Vortex breakdavn reaohes the trailing edge of the
starboard wing at a i: 23' and the port wing at ~1 = 260r.
The wing fundamental mode is strongly excited (Fig.lT(a)) snd the second
distortion mode is also excited (Fig.lT(b)). The third distortion mode is
weakly excited and its signal actually falls from a = C ' to a = 6' (Fig.l7(c)).
The total signal shows a similar fall (Fig.19 and this oould be associated with
the suppression of the lower surface vortex as the inoidenoe is increased from
o" to 6O.
The effects of sideslip on the incidence for vortex breakdown were similar
to those inMode 1 (Fig.12) and are not presented here. A full acaount of the
tests inMode 2, including a comparison of wing-root buffeting signals and
static pressure fluctuations measured under the vortex, will be presented in a
separate report8.
3.3 Model 3
Sinoe Model 3 had a representative structure the safe limits for lift and
pitching mcunent were far less than for the solid models. Safe limits of lift
and pitching moment (based on a factor of safety of 3) were estimated and
rigorously applied. The earlier results fromNcde1 1 (3.1) suggested that the
interesting incidence range would be from a = 6' (buffeting onset) to CI = 12'
(the buffeting plateau). To attain this incidenoe range the kinetic pressure was severely restricted. In the low apeed 13 x 9 ft tunnel operating at atiospherio density the maximum velocity was only 150 ft/s. In the 3 x 3 ft tunnel the f'ree- Stream density o0uld be reduced and hence the velocity inoreased for the same
kinetic pressures.
In the 13 x 9 ft tunnel no buffeting could be detected above the relatively
high signal at zero incidence. In the 3 ft tunnel the buffeting measurements were made atM = 0.50 and free stream densities of 0.004.8 and O.OOg6 lb/ft3
(equivalent to a0 at-d 115 knots IUs).
* Model shaking prevented tests at higher incidence at the test speed of 252 ft/s. The step in the buffeting signal between vortex breakdown on the port and star- board wings onLIode1 1 (Fig.10) was not observed on Model 2.
?
Pig.l8(a) shows typical plots of the variation of total tip strain signal
with incidence atM = 0.50. Buffeting onset is at a = 5' and 4" at p = 0.0048
and 0.0096 lb/f't3 respeotively. At both densities there is a peak in the signal
near a = 7 ', similar to that at a = 8' onMcde1 1. Even at the lowest density
the maximum lncidenoe ws.s restricted to a = 9' and so the existence of the
buffeting plateau was not established. Fig.l8(b) shows the data corrected for
electrical pick up and tunnel unsteadmess. It is assumed that there is no
correlation between the buffeting signal and the signal at zero incidence, i.e.
/m.
+ Flg.19 shows that the tip buffeting is proportional to p rather than
*, which tiplies the predominan
~pingT'lo. The pred oe of structural damping over aeroaynamic
onnnance of st:uc&urd damping on solid wind tunnel models
has been notxed in previous buffet tests but the predominance of structural
damping on thus aercelastic model was not expeated. However, an examination of
the speotra of the-tip strain signal Fig.20 shows the interesting faot that most
of the buffet- is at 3rd symmetria distortion mode at 273 c/s and not at the
?st symmetric distortion mode as on unswept wings. (Fzg.20 shows that the 3rd
symmetric distortion mode predominates even 111 the smaller wing-root strain
signal.) Direct measurements show little increase in total clamping with stream
density for the 3rd mode (Fig.Zl(c)). Hence structural dsmping must predominate
for this mode. (The mode at 155 c/s could have been symmetric or antisymmetric
(see Fig.6).)
At the wing fundamental frequency f = j27.5 c/s, there is a large
variation in total d-pm& with stream density (Fig.21(a)) but this m&e was
not sufficiently exoited by the buffet to &&de if either
TS = p
or TS = p$
is appropriate.
The model and aircraft tip stresses will be identical because of the
dimensional similarity relationships II . Tip stresses and amplitudes during
subsonic climb-out are estimated in Appendix 1 and discussed in Section 4.
3.4 Model4
(1)
(2)
Model 4 was tested to determine how buffeting varied with velocity and
density for the different vibration modes. In the low speed 13 x 9 ft tunnel
tests at atmospherlo density were frustrated by Reynolds number effects, e.g.
IO
as the D-81 velocity increased from 175 zo 250 ft/s the incidence for vortex formation fell from a = 8' to 6'. In the 3 ft tumel, tests at a constant 1~
density (p = 0.005 lb/f+?) over a speed range from&l = 0.30 to 0.70 were also
unsuccessful. If the tunnel kinetic pressure-is increased by increasing the
speed, rather than the free stream density, the frequency parameter WC/V iS
reduced. Any change in frequency parameter 'is really undesirable, even if the
excitation spectrum is relatively fl& as It is m these tests. However tests
at a constant speedM = 0.40, over a 4/l range of density, yieided some interest-
ing data.
Fig.22 attempts to reduoe the buffeting measurements by both equations (1)
and (2) above. The arbitrary scales used have been adJusted so that points for
the highest density are common to both equations.
Data are provided for the three modes excited. The fundamental mode at
f = 263 c/s is not very strongly excited (of the VRS spectra for Male1 3 shown
in Pig.20)). However careful examination suggests that the buffeting data for
this m&e do appear to fit equation (2), rather than equation (I), implying that
aerodynamic dsmping predominates over structural-damping (of the contribution of
aerodynamic damping to the total dampme, for the fLndsmenta1 mode onMode 3,
Fig.2l(a)). The second mcde at f = 380 c/s 1s strongly excited, but the data are
inoonolusive (Fig.22(b)). The response of this mode is probably -ted-by a
mixture of aercdynamic and struotural dsmping. The %hird mode at. f = 1470 c/s
is fairly strongly excited but structural dampini obviously predominates
(Fi&22(C)) .
It is worth noting that the model response over this frequency range again implies that the spectrum of excitation from the leading-edge vortex is relatively
flat in the range of wG/v from 2.5 to 13.5 as onMode 1.
1 4 DISCEXXON I
Tests on the solid models 1 and 2 indicate-that slender wings can r,espond
to the buffet excitations provided by leading-edge vortices ena vortex breakdown
at subsonio speeds up to M = 0.00. The more unportant problem is to determine
if the aircraft response would be signfioant. To answer this question, even on an unswept wing4, aeroelastio models must be tested because the mcde shapes,
frequencies and totai dsmpLngs are then appropriate to the aircraft. This is
partiCL&Wly important for buffeting tests on slerder wings, where-higher modes
than the fun&mental arp most strongly exoited (note the large differences
between the 3rd mde shapes snd frequencies for the solid models and the aero- elastic models which have simiiar external geometry).
11
IiIeasurements on the aoroclastic Model 3 (Appendix A) show that tho full
scale tip amplitude at 7’ incidonco and 270 knots SAS would be about kO.3 inch
at 5 c/s. (Thoso figures arc cstimatcd after applying Wry large scaling factors
and ignore any change in nodo shapo bctuoen 0 and 270 knots IUS.) Because the
model was supportod at tho centre of tho fuselage, amplitudes and accolcrations
of the aircraft fuselage oannot bo directly estlqatod. Howovor, any buffeting 12 a700vc 0.02g might be considorod uncomfortoblo . Honco, during the subsonic
climb after take-off to about 5000 ft altitude, there will be a short poriod
when the combination of incidence and EAS will be suoh that buffeting induced
by tho leading odgo vortices might bo dotectcd. Tho tats on Modal 3 can only
give the probable order of magnitude of buffeting on a typical slender wing
aircraft.
The normel flight cnvclope of a slondcr wing aircraft is unlikoly to
cxtond into the vortex breakdown rogion so that the buffeting induced by vortex
broakdoun is rathor aoadcmic. Honcvcr, it is intoresting to recall that vortex
breakdown did cxcito the wing fundamental mode on Models 1 and 2. Models 3 and
4 could not be tostod at tho incidoncos roquircd for vortex breakdown. Tests
of three idonticol 65' dolta wing made respootively of stool, light alloy and
mngnosium should provide data to substantiate tho buffoting scale relationships
appropriate for vortex breakdown 13 .
Theso present tests give no information about tho prossure fluctuations
oxciting tho buffeting apart from showing that the buffet oxoltation spootrum
is relatively flat over the froquoncy rang0 from
2<G/V<12 .
Typical prossuro fluctuation moasuromontse on kIodo1 2 on the YRS platoau at
a = 12O show
A;//9 = 0.02
and at vortex breakdown a = 2l+'
AG/‘/s = 0.07 .
A moro dotnilod study of prossuro fluctuations on slondor wings to extend 14 tho maasurcmcnts of I'ryott and Onon soems desirable. In particular the planned
(1966) flight comparison of prossuro fluctuations measured on tho FD2 and the
Bristol 221 might rovoal any basic difforonces in vortex structure owing to
planform changes or to tho change from a round to a sharp leading-edge. (The
occasional mild buffeting on the Bristol 221 rcportod by pilots dcos not by
itself prove that the prossuro fluctuations unclor tho vortox arc smaller than
on tho FD2. Tho structure of tho 221 is difforcnt from that of the FD2 and the
pilot sits 6 ft furthor forward.)
12
A fiti question concerna possible Reynolds nunber effects on the buffeting
charaoteristxs of aler,der wings. Wtn sharp leading-edge3 to fix separation
lines Reynolds nuzber effects on the developed VOrkX flow shotid be nuoh smaller
than on wings with round leading-edges (tnere is certainly a large Reynolds
nunber effect on tie vortex on the FD2 aircraft i?lth round leadmg-dgea 15
)a Homover Reynolds number effects might be significant when the snzi~l streamwise
vort~cea ore rolling up to form the main mrtex (Figq9), although no Reynolds
number effects are apparent in this region on Model1 111 the 3 ft tunnel aa
Reynolds number inorcaaed from 0.23.106 to 2.66.10~ (Fig.13): Agam, removal
of the roughness bitnds on the ning anh nose of Model 1.5.n the 13 x 9 ft tunnel
did not alter the incidence for vortex formation or vortex breakdozn, or the
buffeting. Hence the present mensuxmenta should give a fair u-&cation of
both the oh.araoter and order of magrutude of the buffeting on full scs.le slender
rin.g aircraft.
5 co~!cLusIo~
Test‘s on two solid slender wing models show that there are two stages in
the wing butfeting at subsonic speeds (Figs.10 and 16):-
(1) D. mild buffeting when the leading-edge vortioca sr& fully established, _.
Q.nd (2) a severe bufYet at vortex breakdown.
At supersonic speeds the mild~buffeting is weaker, probably because of the
reduced size and intensity of the leading-edge vortices.
Teats on tVro aeroelnstio mgdels show that the mild buffeting is principally
at the wing 3rd distortion mode (Fig.20). The order of magnitude of this buffeting suggests that on.a auperaonlo transport aircraft buffeting or rough
vibration night be noticed-during a amall port of the aubaonio climb-out titer
klke-off.
The ~dmr is grateful. to Hr. L. Martin of the Dyntio Teat Seotlon B.A.c.
Fiston, for all theSmenauremcnts of node shapc- u and damping feotora and also for the ground resonanoe teat of yodel 3.
13
A Apperldic
TIPBUPFET STRESSESAWdTLiDES
The stress induced by the wing buffet on Moclcl 3 is
5 = E(l/h) (AR/k) 2
where d = stress lb/m*
E : Youngs mcdulus lb/in2 (l.107 for light alloy)
h = gauge factor (2 for wire gauges)
AR = resxtmce change R = resistance
(the faotor 2 is required because there are only tvro active gauges in the
bridge onMode 3).
Hence u = E(AV/V)
(3)
(4)
where AV = 131s unsteady voltage
V = steady voltage (6.3V)
From Fig.lS(b) and equation (4) the total rms stress at a = 7’ and 115 kt
EAS is
u = 65 lb/m* .
Similarly from Fig.*O(b) and equation (4) we have a tip stress at the
third deformation mode at 273 o/s (5 c/s full scale)
u = 40 lb/in2 .
In an attempt to determine the order of magnitude of the tip smpLtu&s
associated with, but not directly related to these buffet stresses the model
was subJected to a ground resonanoe test. The third symmetric distortion mode
was excited by the normal push-rod sting used for oxoiting symmetrio mdes* and the excitation ucreasd until the unsteady signal from the tip gauges was 26 PV
* This sinusoidal excitation attempts to represent the partion of the distributed "white noise" loading exciting the model. This loading takes no aoooud of the degree of correlation of the pressure fluctuations at different points under the vortex, or of possible interactions between the response of the fairly closely spaced deformation modes.
14 A~pendixA
rms (Fig.20(b)). The tip amplitude was then approximately fO.001 inch (peak to
peak) oorresponding to +o.060 inch full scale.
These stresses and defleotions will increase at the higher EAS(typically
about 270 knots) required during the rapid subsonio &Limb-out of the aircraft
needed to conserve fiel. Only a small increase in total damping for the 3rd mcde is suggested by extrapolation of Fig.21(c). Hence the buffet- stress end defleo-
tions may be approximately inoreased by the ratio of the kinetic pressures,
i.e. (270/115)2 = 5.5.
Hence
u = 220 lb/in2 ) at 5.0 o/a full scale
and the tip deflection = +o.jin
(peak to peak)
Table 1
Test condltlons
MaxiiTUSl working Reyholds FuFllamenta1
!hael Tunnel section number(s) VRS bridge (", c)n((o, G)U sting support Date of tests
R wing f$equsncy
S
1 3 f t Tabs 2.66.40~ 4 active semi- 285 1.65 M=OAo to
Solid sting October 1964 con6uctor gauges
M=O.80
l3x9ft low speed o.92.106 November 1964 M=O .23
3 ft Tabs 2.66.10~ 1 Lko.40 to k0.80
1
supersonx 2.20.106
Id=1.4 and
&I .6
I Janusry 1965
1
2 ipgft low speed 2.56.10~ 4 active semi- 130 1.87 6 component July 1965 k0.23 conductor gauges balance
3 3 ft subsonic 0.34.106 2 active wire gauges 127 1 .oo Flutter Janusry 1965 M=0.50 + 2 external excitation
resistors sting
4 3 ft subsonic 0.5~0~ 4 active semi- 263 1.52 Solid sting day 1965 M=O.40 conductor gauges
I , , / , , , , , , , / ~ , , , I ,
Tabs = Top and bottom slotted section
0
0 crit
c
EAS
k2
11
G 9 R
V
Yii?S
a
?
Y
P w = 2llf
aping critical damping
average chord
equivalent airspeed knots
frequency c/s
structural dsmping coeffuient % crit
Mach number
rms pressure fluctuation lb/in2
kmetio pressure &pV* lb/in2
Reynolds number based on c
free stream veloczty ft/s
wing-root strain
incidence
sideslip
aerodynamic damping coefficient% crit
free stream density
circular frequency (radians/s)
17
R.lPEZNCES
s Author
1 D. Kiichemarn
2 K. G. Turner
3 N. Lambourne
4 D. G. Mabey
5 I. McGregor
6 L. C. Squire
7 L. C. Squire
J. G. Jones
A. Stanbrook
0 R. F. Keating
G. Moss
Further oxperimentcl investigation of the character-
istics of cambered gothic wings at Mach numbers from
0'4 to 2'0.
A.R.C. R & M 3310, December 1961
An experimental investigation of the characteristics
of some plane and cambered 65' delta wings ct Mach
numbers from 0.7 to 2'0.
A.R.C. R & M 3305, July ?961
Unpublished M.O.A. Report
9 D. D. Davis Jnr The use of wind tunnels to predict flight buffet
W. B. Huston loads.
10 D. D. Davis Jnr
D. E. Uornom
11 W. H. Melbourne
12 H. F. Huddleston
Title. etc.
A non-linear lifting surface theory for wings of
smell aspect ratio with edge separations.
A.R.C.17769 April 1955
Unpublished M.O.A. Report
The breakdown of certain types of vortex.
A.R.C. C.P.915 September 1965
Unpublished M.O.A. Report
The vapour screen method cf flow visualisation.
Journ. Fluid Mech. Vol Ii, p.481-511, 1961
NACA RX L57D25, June 1957
Buffet tests on an attack airplane model with
emphasis on analysis of data from wind tunnel test%
NACA RM L57H13, February 1958
Aerodynamic investigation of tho wind loads on a
cylindrical lighthouse.
Unpublished 1t.O.D. Report
Verbal estimates of the severity of vertical
sinusoidal vibration,
Unpublished work
18
P.lzmimTcES (conta)
NO. Autnm - Title, etc.
13 D. G. Mabey Unpublinhcd iI.0.h. Report
14 L A. Fyatt Preliminary low speed measurements of akin friction and
T. B. Owen surface pressure fluctuations in a slender wing at
mcidence.
A.R.C. 25436, S.zptcmbor I%3
15 D. G. Mabey Comparison of seven wing buffet boundsries measured in
wind tunnels and in flight.
A.R.C. C.P.840, Soptenbcr 1964
24”
MODEL SCALE ‘/75
T- 6.2”
FIG. I G.A. OF MODEL I
, ,*1 / ,,, ,,, ,,, ,/,,, ,,
(FULL SCALE FREQUENCIES IN BRACKETS)
12 DISTORTION
MODE
f-2846 C/i (3 8 c/s)
912 - -0.00077
t!??
DISTORTION
MODE
f = 312 2 C/S (4 =/s>
35?
DISTORTION
MODE
T” 4- f = 769 5 c/5
DISTORTiON (IO. 3 c/s>
MODE q2 = 0 0013
FG.2 MODEL I SYMMETRIC DISTORTION MODES
J
FIG.3 G.A OF MODEL 2
(FULL SCALE FREQUENCIES IN BRACKETS)
MODE
% = O-0042
2 nd DISTORTION 53 (5.0c,s)
MODE. O-00082
\ \ .- \ I
e
/ /-
- 280 c/s (9.3 c/s)
3” DtSTORflON MODE $2 =
0~00048
4!+ DISTORTlON MODE
FIG.4 MODEL 2 DISTORTIQN MODES
.
tt DISTORTION
MODE f = 127-5 C/S (2.3 C/S)
2 nd DISTORTION MODE
f = 15.5 c/s (2-s c/s)
’ 3 rd 01 STORTI ON
f = 273 C/5
MODE , 0 0055 f .- -
’ 4th OlSTORTlOFl f = 300 c/s
MObE
7th DlSTORTlON MODE
2 o*oos \ ,
\ .- - --.’ : -. - .
w
/ - -_ :-- /
/ :
‘(5-5 c/5)
(9.7 c/s) f = 535 cJ.5
@= 0.013
lFIG.6 (a) SYMMETRIC MODES
FIG.6 MODEL 3 DISTORTION MODES
( FLILl- SCALE FREQIJE.NCIES IN 0RAWElS)
f- 153 c/s (2 8 C/S)
?/zoo 006
2 nd
DISTORTION MODE
f = 178 C/S (3 -2 c/5)
Mozmz+
3d DISTORTION
f = 254 c/S (4.6 C/S)
f = 342 C/S (6-Z c/s)
f/2 = O-006
FIG. 6 (b) ANTISYMMETRIC MODES
FIG.6 (CONCLD) MODEL 3 DISTORTION MODES
(FULL SCALE FREQUENCIES IN BRACKETS)
2 nd
DISTORTIOhJ MODE
3 t-c+ DISTORTION
MOk + = 434 G/S (5.8 G/S)
I
4th DISTORTION MODE
(IO * 3 G/S)
FIG. 8 MODEL 4 DISTORTION MODES (APPROX~
.
.
c
400
WRS SIGNAL
P” 300
200
100
0
13x9 ft 13x9 ft TUNNEL TUNNEL
V= 252 V= 252
BREAKDOWN BREAKDOWN
5 IO 15 20 25 ao 30
FIG.10 MODEL I VARIATION OF TOTAL WING-ROOT STRAIN SIGNAL WITH INCIDENCE
WRS 13 X 9 Ft TUNNEL
200
WRS
SIGNAL
FlG.Il(a) f= 3OOc/s
100 I 4 VORTEX *..’
FORMAT IO d - =--a @ A ++c
VO’ftTEX BREAKDOWN
0 5 IO 15 20 25 30 a0
FIG.11 (b) f= 700 c/s
100 H/R5
SIqNAL
/uV
0 5 IO I5 20 25 cc0
30
FIG.ll(c) f = 1330 c/s
* I
FIG. II MODEL I VARIATION OF TUNED WING-ROOT STRAIN SIGNAL WITH INCIDENCE
I3 %S f t TUNNEL
v = 252 ft/s
--
PORT WING
-1
IO
-zoo -loo O-
IO p” io
FIG.12 MODEL I EFFECT OF SIDESLIP ON INCIDENCE
FOR VORTEX BREAKDOWN
WRS SIGNAL
800
P”
700
600
500
400
300
200
I00
I 0
A h ,“ = 00366 lb/f t3
(‘~0 0205 lb/f?
c rk h
(’ = 0.0091 lb/f
50
FlG13MODEL I VARIATION OF TOTAL WING-ROOT STRAIN SIGNAL WITH INCIDENCE AND DENSITY M= 0.50
FIG. 14 MODEL I VARIATION OF TUNED
WING-ROOT STRAIN SIGNAL WITH INCIDENCE M= 0.50
1000
WRS
SIqNAL
900
PV
800
700
P
6OC
sot
400
30G
zoo
100
0
3Ft TUNNEL
D I f -0 - 0278 lb/f
I I
I h
h Y p=O 0045 lb/k3
h pt = 16 in Hj
pt=9 In Hj
Y D pb=4 In Hj
IY!.IIl 0 cc 15 O0
FIG.15 MODEL I VARIATION OF TOTAL WING- ROOT STRAIN SIGNAL WITH INCIDENCE AND DENSITY M=Iv4
1400
WV SQNAL
1200
1000
800
600
400
200
0
--- --_--
13x 9Ft TUNNEL.
v = 252 W/S
PORT WING
5 IO 15 20 25 oc”
30
FIG ~ 16’MODEL 2 VARIATION OF TOTAL WING-ROOT STRAIN SIGNAL WITH INCIDENCE ,..
1000 WRS
SIGNAL
P”
6OC
60C
4oc
20(
0
13 r9 f t TUNNEL
v= 252 f t/5
I 0 0
/ . .
z
0
0
5
t
-
--I-- ) 15
FIG 17 (a) f = 128 C/S 4Ob
WRS SIGNAL
P”
200 WRS
SIGNAL
P”
FIG.17tb) f=lSoC/S
0 5 IO 15 20 25 do 30
FIG.17 (Cl f = 277 C/S
FIG.17 MODEL 2 VARIATION OF TUNED WING-ROOT STRAIN SIGNAL WITH INCIDENCE
40
TS
SIGNAL
I1” 30
20
IO
0
3 f t TUNNEL
2.0 40 6.0 8.0 ct IC
FIG.18 (a) UNCORRECTED FOR “PICK UP” AND UNSTEADINESS
40
TS SIGNAL
P” 30
O0
FIG.18 (b) CORRECTED FOR “PICK UP” AND UNSTEADINESS
FIG.18 MODEL 3 VARIATION OF TOTAL TIP-STRAIN SIGNAL WITH INCIDENCE AND DENSITY M=O.50
TS/e (AR~ITARY
UNIT5)
( ARBllARY UN ITS)
IO
3Ft TUNNEL
I + 0’.00;6 Ib/ft3 ,* &’ 81, 0 O-0048 lb/d
1’ 5 I I
0 c8-@---&-4 ‘s
I I I 40 6-O 8 0 at IO.
FIG.(a) T S SIGNAL/(DENSITY) v INCIDENCE
IO, I I I I I
+ t 0*00;6 lb/&’
Q 0*0048 Ib/ft’
,0°
I I I +@
..i. B 2.0 70 - 6.0 8-O cc
,o.o” I
FIG.19 (b) T S SIGNAL / (DENSITY)’ V INCIDENCE
FIG. 19 MODEL 3 COMPARISON OF BUFFETING SCALING LAWS M=O. 50
20 T-5
PV
I * FIRST THREE DEFORMATlON
3’ MODES
0 SOFT SUPPORT
IO -\‘PICK UP)‘--2
P I
7th DEFOR’MATION _
0 Yp
MOOE
v ?
R
50 125 162 274 535 IO80 f c/s
FIG.20 (a) TIP-STRAIN p = 0. 00478 lb/ft3 cc=7.10°
30
TS
rv 20
IO
0
I, -1
-L I 2
JICK Up” ,
ti
RST THREE DE FORMATION MODES
Ll SOFT SUPPORT
50 125 162 274 535 1080 f c/s
FIG. 20 (b) TIP-STRAIN f=O -00956 lb/ft3 cc m7.0’
20
WRS
Y
IO
0
-. -PICK
UP” 1 r
* FIRST THREE ’
’ DE FORMATIObl I
’ 2
iii
MODES
YARD SUPPORT
50 125 162 274 535 1080 C/S
FIG. 20(c) WING-ROOT STRAIN p=O~OO956 lb/ft’ cc=9.0°
FIG.20 MODEL 3 SPECTRA OF STRAIN SIGNALS M-0,50
)I+
2.0 4-O 6.0 . 8.0 cc FIG.21 (a) f=l27.5 c/s
20 4.0 60 FIG.21 (b) f=l55 c/s
0 01 , +
iT + g2 2 t
! 0 II\ .-
WIND OFF
0 2-o 40 6-O 8-O ,& IO-O”
FIG. 21 (c) f=273 c/s
FIG. 21 MODEL 3 VARIATION OF TOTAL DAMPING COEFFIENTS FOR SYMMETRIC MODES WITH INCIDENCE
AND DENSITY, M=O-50
3Ft TUNNEL
5-o IO-0 15 0 s 20, ,o”
I 7 c-e. I I J n 5.0 IO.0 150 cc to-o0
FIG. 22 (a) FUNDAMENTAL f= 263 c/s
FIG 22 MODEL 4 COMPARISON OF BUFFETING
SCALING RELATIONSHIPS FOR DIFFERENT MODES M=0*40
I20
100
80
60
40
20
0
0 0~00.5
X 0~010 I I -0 0.020
I EQN I WRS cc
I
-I----
FIG22 (bl) f = 380 C/S
FIG.22(CONT’D) MODEL 4 COMPARISON OF BUFFETING SCALING RELATIONSHIPS FOR
DIFFERENT MODES M=0.40
WRS
00
60
40
20
0
FIG.22 (b2) f = 380 C/S
FIG.22 (CONT’D) MODEL 4 COMPARISON OF BUFFETING SCALING RELATIONSHIPS FOR
DIFFERENT MODES M= O-40
60
40
20
0
80, I I I
WR#
6C
40
2c
I EQN 2
WRS Oc ,=‘+
FIG 22(C) f= 1470 C /S
FIG.~~(coNc~D) MODEL 4 COMPARISON OF BUFFETING SCALING RELATIONSHIPS FOR
DIFFERENT MODES M = 0.40 Prmted in &gland for Eer Nojetty’s Stationery Ojfrce by the Royal Aimsaft BstablMment, Pamborough. Dd.125875. 1.U
A.R.C. C.P. No.917 piarch lw6
533.6.013.43 : HabCY, D.G. 533.6gJ.3 :
NCAsuRSSN=TS OF BUFFEWX ON SLEti WING HOCEls 532.527 t 533.6.011.32/3&
Tests on Tao solid slender wing mdels show that them am tvm staws in Chs ting bulfetlng at absoalc spads;
(1) a mild bufletlw due to leadlw-adge vortlOes, and
(2) a evers b”ffeClW at mrtcx breakdorm.
P.T.O.
, . . , , , , , , , * , , , , ,,v, , , , , , * , , , , Iv, , “ , , , , , , .#, . , , , , I , , . , , . . , . , , , , , , , . ,:,.<p>, ,
A.R.C. C.P. No.917 Hamh1566 ’
tlabey. D.O.
PEAS-S OF BUFFETING DN SEWER WIND “OOELS
533.6.013.43 : 533.693.3 : 532.527 1 ~3.6.011.32/34
A.R.C. C.P. No.9l7 March 1966
t’abcy, D.C.
lEh%RE-ImTs OF BUFFETING ON SIXElm “,ND HODELS
533.6.013.W : 533.673.3 : 532.527 : 533.6.011.32134
Tests on Tao solid slender wing models show ttat there are tvm stages In the wllg bulfetlng at subsonic sped.%
(1) a mild buffeting due to leading-edge vortices, and
(2) a s.e~eR buffeting at WI-teX breakdown.
At mpr~~nlc speeds the mlld bulretlng is weaker. pmbably because or the reduced size and Intensity of the leadlng+cl@ vortices.
P.T.O.
,
, , I . , , , > , , ,
Tests on two semelastic mdsls St subsonlc speeds show tlmt the mild buffeting is prinoi~lly St the wing 3rd OP higher diStm-t.lo" IPdnS. Ths order 01 magnitude of this bUffetAM meats ttat 0" B S"PrSo"ia transport aimraft this &!ht be penxptlble durin(( slbsonlc cllmbaut at hleh a..¶.
Tests on tm acmelast maels at mbsollic speeds show tkit this 0114 lbsts on tm asmSl~st.1~ models at sbsonle speeds show that tha m‘ld buffaCIng is prlnclgally at tla wing 3rd or bighsr dIstortlo" mdes. The buffeting Is prlnai~lly at the wing 3rd or hi&w dlsiortion mxlas. T~s oI-dSr Of lsagnltude Oi this buffetlW SupbeSts tkzt on a S"prso"lC order or menltti 0r this burreting mggests that on a suprsonic transport almr-aft this tight be perceptible dwlng subsonic clinb-out st tIX"spOrt SlrCIQft Chls might be perceptible during srbsonic cll,ri,-out St hiah Ms. hm EAS.
C.P. No. 917
Pubhshed by HER MAJESTY’S STATIONERY OrFJCE
To be purchased from 49 High Holborn. London w c 1 423 Oxford Street. London w 1 13~ Castle Street. Edmhurgh 2
109 St Mary Street. CardIt? Brazennose Street, Manchester 2
50 Fan&x Street. Bristol 1 35 Smallbrook. Rmgway, Bmmngham 5
80 ChIchester Street. Belfast 1 or through any bookseller
C.P. No. 917
S 0 CODE No. U-9017-17
Top Related