Full text of DOI:10.2514/2.3135

JOURNAL OF AIRCRAFT
Vol. 40, No. 3, May–June 2003
Effect of a Trailing-Edge Jet on Fin Buffeting
¤
†
‡
S. Phillips, C. Lambert, and I. Gursul
University of Bath, Bath, England BA2 7AY, United Kingdom
Effect of a trailing-edge jet on the interaction of a leading-edge vortex with a Ž n and resulting Ž n buffeting was
investigated in water-tunnel as well as wind-tunnel experiments. Flow visualizationshowed that Ž n-induced vortex
breakdown can be delayed into the wake even for the head-on collision of the leading-edge vortex with the Ž n.
Hence it was demonstrated that the adverse pressure gradient caused by the presence of the Ž n could be overcome
with a de ected trailing-edge jet. Delay of vortex breakdown into the wake even for relatively small values of jet
velocity ratio is possible for de ected jets, whereas the effectiveness of the jet with no de ection is very limited.
Buffeting response of a  exible Ž n in wind-tunnel experiments showed that there was considerable delay of the
onset of buffeting to higher angles of attack with increasing jet momentum for ¯ = 30 and 45 deg. Depending on the
Ž
n location with respect to the leading-edge vortex, it was possible to shift the buffeting envelope as much as 12 deg
in incidence of the wing. Experiments with varying wing sweep angle showed that jet blowing was as effective in
attenuating Ž n buffeting for less slender wings. Wind-tunnel experiments also showed that the nozzle geometry is
very important and causes very different buffeting envelopes for the same momentum coefŽ cient.
Introduction
is the dominant factor in determining the location of vortex break-
down and the magnitude of Ž n buffeting. Limited effectiveness of
the  ow-control techniques just mentioned is caused by incapabil-
ity of altering the external pressure gradient while modifying the
structure of the vortices.
XPERIMENTAL evidencesuggeststhat severalunsteady ow
phenomenacan causeŽ n buffeting.Thesephenomenaandtheir
E
1
physical models were recently discussedby Gursul and Xie. How-
ever, vortex breakdown phenomenon is the most important source
of buffeting over delta wings. A wide variety of investigationswas
conductedon both simpliŽ ed Ž n-delta-wingconŽ gurationsand Ž ns
on actual model aircraft in order to understand the mechanisms
of Ž n buffeting and the relation to the vortical  owŽ elds over the
The purpose of this study is to investigate a different aspect of
vortex control technique, which is more likely to alter the exter-
nal pressure gradient, and involves a jet at the trailing edge of the
wing. As the beneŽ cial effects of trailing-edge blowing on vortex
2
0¡23
2
¡13
breakdown over delta wings are well known,
this has good
wing.
given by Wolfe et al.
A good summary of the experimental investigations is
8
potential to overcome the adverse pressure gradient caused by the
trailing edge and the Ž n and to delay vortex breakdown and attenu-
ate Ž n buffeting. Thrust vectoringremains as a preferredmethod of
integrating propulsive and lift systems for modern Ž ghter aircraft,
Both structural-controland  ow-controlmethods have been used
by previousinvestigatorsto attenuateŽ n buffeting.Structuralmeth-
ods include increasing the stiffness and damping. There have also
been attempts to reduce the buffeting with active vibration con-
2
4
which can have thrust/weight ratio larger than unity. This means
thatlargeamountof mass injectionthroughtrailing-edgejets is pos-
sible. This study concentrateson the effectsof a trailing-edgejet on
1
4
15
trol techniques using piezoelectricactuators or an active rudder
without any knowledge of the  owŽ eld. Unfortunately, only mod-
est reductionsin Ž n responseare possiblewith the structuralcontrol
methodsin the absenceofany ow-controlmethod.Regarding ow-
control methods, several techniques are available in order to alter
the position of the vortex with respect to the Ž n or delay vortex
Ž
n buffeting.
Experimental Setup
1
6
17
Water-Tunnel Experiments
breakdown.These includeblowing and suction on the wing sur-
3
18
face, fences, and variable position leading-edge extensions.
A
Initial experiments with trailing-edge blowing were performed
in a water tunnel located in the University of Bath because of its
advantages in  ow visualization. The water tunnel has a horizon-
tal working section, with a cross-sectional area of 15 by 20 in.
(381 by 508 mm). The delta wing and Ž n were designed and
scaled to those used in the wind-tunnel experiments previously re-
general review of vortex breakdown control methods is presented
by Mitchell and Delery.19 Common to these investigationswas the
introduction of modiŽ cations well upstream of the breakdown and
close to the origin of the vortex. However, these methods are not
effective over a wide range of angle of attack encountered dur-
ing a maneuver because of drastic changes in the position of the
vortex. Moreover, the effectiveness of these techniques reaches a
saturation with increasing control parameter (such as the blowing
coefŽcient) as one tries to delay the location of vortex breakdown
because vortex breakdown phenomenon strongly depends on the
external pressure gradient. Trailing edge of the wing and, more im-
portantly,the Ž n itselfproducesan adversepressuregradient,which
2
5
ported elsewhere. The wing and rigid Ž n are shown in Fig. 1.
The chord length is c D 250 mm, giving a Reynolds number around
4
Re D 2:5 £ 10 for U
D 10 cm/s. At the maximum angle of attack
1
® D 40 deg, the blockage ratio was 6%, and no correction on the
data was attempted. The Ž n was attached to the trailing edge of the
wing with a screw, and the location of the Ž n could be varied from
y
=s D 0.0 to 1.0 with increments of 0.1.
f
For the trailing-edge blowing experiments a rectangular nozzle
Received 16 February 2002; revision received 17 September 2002;
withanaspectratioof6 anda widthof30mm wasplacedunderneath
the wing as shown in Fig. 1. The centerlineof the nozzle coincided
accepted for publication 18 September 2002. Copyright °c 2003 by the au-
thors. Publishedby the American Institute of Aeronautics and Astronautics,
Inc., with permission. Copies of thispaper may bemade for personal orinter-
nal use, on conditionthat the copier pay the$10.00per-copy fee to the Copy-
right Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923;
include the code 0021-8669/03 $10.00 in correspondence with the CCC.
with the approximatepositionof the vortexaxis (y
the range of angleof attack tested. The spanwiselocationof the vor-
=s ¼ 0.6) within
v
tex was estimated from the pressure measurements across the span
2
5
at x=c D 0.5 in the absenceof the Ž n. It was found thatthe location
of the suction peak is only slightly affected by the angle of attack,
showing that the spanwise location of the leading-edge vortex is
roughly constant. An important parameter was jet de ection angle
¯, which was generated by using a thin plate with the same width
¤
Student, Department of Mechanical Engineering.
Postgraduate Student, Department of Mechanical Engineering.
Reader in Aerospace Engineering,Department of Mechanical Engineer-
†
‡
ing. Member AIAA.
5
90

PHILLIPS, LAMBERT, AND GURSUL
591
a) Design of  exible Ž n (dimensions in mm)
Fig. 1 DeŽ nition of Ž n location yf and jet de ection angle ¯.
b) Half-delta wing modal, Ž n and splitter plate
Fig. 3 Overview of the experimental setup in wind tunnel.
shaping. The advantages that the wood sections offer are low ma-
terial density and ease of fabrication. These sections were attached
to the spar with small bolts. With this design the contributionof the
wood sectionsto the bendingstiffnessof the spar is minimized. The
dimensionsof the sparwere chosento obtainthe naturalfrequencies
of the Ž rst bending mode for a typical modern combat aircraft. The
thickness of the spar was 2 mm. The leading edge of the Ž n was
double bevelled at an angle of 30 deg. The main dimensions of the
Ž
n are given in Fig. 3a. The spar was attached to the delta wing by
a bracket near the trailing edge of the wing.
The experimentalsetup, which uses a half-model delta wing and
a splitter plate, is shown in Fig. 3b. Three delta-wing models with
sweep angles 3D 65, 70, and 75 deg were used in the experiments,
although the majority of the tests were carried out for 3 D 75 deg
as in our previous investigations and in water-tunnel experiments.
All three wings had a chord length of c D 500 mm and a thickness
of 15 mm. The lee surface was  at, whereas the leading edges were
bevelled at 45 deg on the windward side. The Reynolds number
Fig. 2 Schematic of the experimental setup in water tunnel.
as the nozzle to de ect the jet (see Fig. 1). The volumetric  ow rate
of the jet was measured by a rotameter, which was placed at the
dischargeside of a submersiblepump as shown in Fig. 2. The max-
5
based on the chord length was Re D 3:5 £ 10 . For 3 D 75 deg the
dimensions of the delta wing and Ž n are scaled to those used in
the water-tunnelexperiments.Buffetingresponseof this  exible Ž n
was investigated by measuring the Ž n vibration levels sensed by a
tip accelerometerattachedto the spar. The measurementuncertainty
for the tip accelerationis estimatedas 2%. In additionto calculating
the rms value of the Ž n tip acceleration,the spectra of the tip accel-
eration were examined for each case. The Ž n vibrations occurred
at the natural frequency of the Ž rst bending mode, which is hardly
in uenced by the angle of attack.
imum jet velocity obtained was Ujet=U
1
D 8.9, which corresponds
to a momentum coefŽcient of
¯
2
jet
1
2
2
C
¹
D ½U
A
jet
½U S
w
D 0:708
1
Here Ajet and S
denote cross-sectionalarea of nozzle exit and sur-
w
face area of the wing. As the thrust/weight ratio can be larger than
2
4
unity, it can be shown that the momentum coefŽ cient can take up
valueson the orderof unity.The valuesofthe momentumcoefŽ cient
used in the experiments are therefore realistic and also consistent
An identicalnozzlescaled to one used in the water-tunnelexperi-
ments was connectedto pressuredair supplyto producethe trailing-
edge jet. The mass  ow of the jet was monitoredby a rotameter,and
2
0¡23
with the range used by the other investigators.
In the water-
tunnel experimentsdye  ow visualizationwas used to visualizethe
vortex trajectories and breakdown location. A digital video cam-
era was used to record the  ow visualizationand to further analyze
the results. The measurement uncertaintyfor vortex breakdown lo-
cation was 1% of the chord length. Estimated uncertainty for the
momentum coefŽ cient was 2%.
the maximum momentum coefŽcient obtained was C
¹
D 0.287 for
3 D 75-deg delta wing. Because of the limitations of the air sup-
ply, this maximum momentumcoefŽ cientwas considerablysmaller
than that used in the water-tunnelexperiments.
Results
Wind-Tunnel Experiments
Water-Tunnel Experiments
Previous wind-tunnel investigation25 using a  exible Ž n showed
Additional experimentsin a 2.12 by 1.51 m low-speed wind tun-
nel were carriedout to demonstratethe applicabilityof the proposed
control methods at higher Reynolds numbers. A  exible Ž n shown
inFig. 3a wasdesignedand fabricated.Itconsistsofa thinaluminum
spar surroundedby several wood segments to provide aerodynamic
that the Ž n location y =s is a very important parameter in addition
f
to the angle of attack. Outboard Ž n locations produce very light
buffeting, surprisingly even at high angle of attack. Inboard Ž n lo-
cations can produce the heaviest buffeting, depending on the angle

5
92
PHILLIPS, LAMBERT, AND GURSUL
a)
b)
c)
a)
b)
Fig. 5 Flow visualization for yf /s = 0.2, ® = 31 deg, ¯ = 30 deg: a) Ujet/
U1 = 0 and b) Ujet/U1 = 8.9 (C¹ = 0.708).
Fig. 4 Flow visualization for a) yf /s = 0.2, ® = 25 deg; b) yf /s = 0.2,
= 35 deg; and c) yf /s = 0.6, ® = 23 deg for no jet blowing.
®
a)
of attack. At low to moderate angles of attack, the buffetingis light.
An example is shown in Fig. 4a for y
high angle of attack, in particular when the shear layer impinges
on the Ž n, the heaviest buffeting is observed. Here, the shear layer
refersto the regionof largevelocitygradientformed betweennearly
stagnant ow nearthe centerof vortexand the freestream.An exam-
=s D 0:2 and ® D 25 deg. At
f
ple is shown in Fig. 4b for y
interestingconŽ gurationis for y
=s D 0:2 and ® D 35 deg. A particularly
f
=s D 0:6, which correspondsto the
f
approximate position of the vortex axis within the range of angle
of attack tested. An example is shown in Fig. 4c for ® D 23 deg.
Because of the adverse pressure gradient produced by the presence
of the Ž n, vortex breakdownis always observedupstreamof the Ž n.
This head-on collisionof the vortex with the Ž n producesmoderate
level of buffeting, when compared to heavy buffeting produced for
inboard Ž n locations. The fact that the head-on collision does not
produce large buffeting when the center of the vortex impinges on
b)
Fig. 6 Flow visualization for yf /s = 0.4, ® = 31 deg, ¯ = 30 deg: a
U1 = 0 and b) Ujet/U1 = 8.9 (C¹ = 0.708).
)
Ujet
/
4
;10
the Ž n is consistent with the previous work.
focused on the Ž n locations y
=s D 0.2, 0.4, and 0.6 in this study.
Figures 5–7 show  ow visualizationpictures for a) no  ow con-
trol and b) jet blowing at the maximum  ow rate (Ujet=U
D 8.9,
=s D 0.2, 0.4, and 0.6. The angle of attack is
Consequently, we
f
ratio is possible for de ected jets, whereas the effectiveness of the
jet with no de ection (¯ D 0 deg) is very limited. Positive effect of
jet de ection in delaying vortex breakdown was noted previously
1
2
1
C
®
¹
D 0.708) for y
f
for a delta wing and with no Ž n. Our results for y
f
=s D 0.2 is
D 31 deg, and the jet de ection angle ¯ D 30 deg. It is seen that
consistent with those of Ref. 21. Similar observationscan be made
from Fig. 8b for a higher angle of attack, although  ow control is
less effective overall. No attempts were made to study the effect of
de ection angle in detail, and only four values of ¯ were used in
the currentexperiments.Nevertheless,the results suggestthat there
might be an optimum value of ¯, which may depend on the angle
of attack. Consequently we performed detailed  ow visualization
experiments for ¯ D 30 and 0 deg for comparison.
vortexbreakdown is delayedinto the wake for all three Ž n locations
when the jet is turned on. In fact, even for the head-on collision
(
y
f
=s D 0.6) it is possible to eliminate the breakdown completely.
Effect of the velocity ratio Ujet=U
on the location of vortex break-
down is shown in Fig. 8 for ® D 30 and 40 deg for various jet de-
ection angles ¯. Figure 8a shows that complete elimination of
vortex breakdown even for relatively small values of jet velocity
1


PHILLIPS, LAMBERT, AND GURSUL
593
a)
b)
Fig. 7 Flow visualization for yf /s = 0.6, ® = 31 deg, ¯ = 30 deg: a) Ujet/
U1 = 0 and b) Ujet/U1 = 8.9 (C¹ = 0.708).
Fig. 9 Variation of breakdown location with angle of attack for
¯
= 0 deg.
Figures 9 and 10 show the variation of breakdown location as
a function of angle of attack for three Ž n locations for ¯ D 0 and
3
0 deg, respectively.In each case no  ow control(Ujet=U
1
D 0) and
D 0.708)
a)
jet blowing at the maximum  ow rate (Ujet=U
D 8.9, C
1
¹
are compared. Each data point represents the time-averagedbreak-
down location,and largescatterof data is observed,in particularfor
the jet blowing, which indicatesthe unsteadycharacterof vortex/jet
interaction.It is seen in Fig. 9 that, for ¯ D 0 deg, vortexbreakdown
location is delayed 20 to 30% of the chord length with a trailing-
edge jet. For ¯ D 30 deg shown in Fig. 10, the delay is around 40 to
6
0% of the chord length. For y
f
=s D 0.2 and 0.4 vortex breakdown
can be completely eliminated up to an angle of attack of around
3
5 deg. Even for the head-on collision case y
f
=s D 0.6 for which
vortex breakdown is always upstream of the Ž n, breakdown can be
eliminated up to ® D 30 deg with jet blowing. A summary of the
results presentedin Figs. 9 and 10 is shown in Fig. 11 together with
the resultsfor no Ž n case. It is seen that for Ujet=U
location for all three Ž n locations is further upstream compared to
D 0 breakdown
1
no Ž n case,and the effectof Ž n locationis negligiblefor xbd=c · 0.6.
ForUjet=U
1
D 8.9(C
¹
D 0:708)breakdownlocationfor y =s D 0:2,
f
0
.4, and no Ž n case roughlycollapse,whereasvortexbreakdownfor
the head-on collision case is further downstream.
b)
The results presented so far show that vortex breakdown caused
by the presence of a Ž n can be completely eliminated by a trailing-
edge jet even at high angle of attack. Most likely reason behind this
improvementis thatthe trailing-edgejet createsa favorablepressure
Fig. 8 Variation of vortex breakdownlocation as a function of velocity
ratio for yf /s = 0.2, jet de ection angles ¯ = 0, 15, 30, and 45 deg, and
for a) ® = 30 deg and b) ® = 40 deg.

5
94
PHILLIPS, LAMBERT, AND GURSUL
a)
b)
Fig. 11 Variation of breakdown location with angle of attack for ¯ =
3
0 deg and a) Ujet/U1 = 0 and b) Ujet/U1 = 8.9 (C¹ = 0.708).
Fig. 10 Variation of breakdown location with angle of attack for
= 30 deg.
¯
gradient near the trailing edge21 and reduces the adverse pressure
gradient caused by the Ž n. This is similar to the early experiments
of Lambourneand Bryer,26 who showedthat a downward de ection
of a trailing-edge ap delayedthe vortexbreakdown.Likewise, vor-
tex breakdowncould be completely eliminated on a cambered delta
wing. It can be argued that a de ected jet acts like a de ected  ap
and accelerates the  ow near the trailing edge, and hence reduces
the adversepressure gradient. However, there are other factorssuch
as entrainment effects suggested by Ref. 20 and the interaction be-
a)
2
7
tween the jet and the wing vortices. As the jet exhausts into a
2
7
cross ow, a counter-rotatingvortex pair is generated. Wang et al.
showed that wing vortices can be drawn toward the jet center by
the induced velocity created by the jet vortices. Figures 12 show

ow visualizationpicturesfor a) Ujet=U
a “steady-statecase”), c) right after the jet is turned off, for the fol-
1
D 0, b) Ujet=U D 8:9 (in
1
b
)
lowing parameters: y
=s D 0:2, ® D 30 deg, ¯ D 30 deg. It is seen
f
that the leading-edge vortex is drawn toward the jet and is nearly
parallel to the jet in the steady-state case. When the jet is turned
off, the wing vortex realigns itself to become nearly parallel to the
freestream.Then the vortexbreakdownslowly propagatesupstream
and eventually reaches a steady-statelocation similar to that shown
in Fig. 12a. The role of jet entrainment and jet/vortex interactions
deserves further study.
c)
Wind-Tunnel Experiments
Fig. 12 Flow visualization for a) Ujet/U1 = 0, b) Ujet/U1 = 8.9 (C¹ =
To quantify the effect of trailing-edgejet on the buffeting of the
0.708), and c) right after the jet is turned off, y /s = 0.2, ®= 30 deg, and
f

exible Ž n, detailed experiments were carried out for 3 D 75-deg
¯ = 30 deg.

PHILLIPS, LAMBERT, AND GURSUL
595
Fig. 13 Variation of rms tip acceleration as a function of incidence for
yf /s = 0.2.
Fig. 14 Variation of rms tip acceleration as a function of incidence for
yf /s = 0.4.
delta wing for y
D 0, 15, 30, and 45 deg. Figure 13 shows the variation of the
rms tip accelerationas a function of angle of attack for various mo-
mentum coefŽ cients for the Ž n location y
the effect of the jet momentum is very small for ¯ D 0 deg, but there
is considerable delay of the onset of buffeting with increasing jet
momentumfor ¯ D 30 and 45 deg. It is possibleto shiftthe buffeting
envelopeas much as 8 deg in angle of attack at the largest momen-
tum coefŽ cient. Hence, blowing at the trailing edge appears to be
more effective at the higher jet de ection angles. These results are
consistentwith the  ow visualizationpictures obtained in the water
f
=s D 0.2, 0.4, 0.6, and jet de ection angles
¯
tunnel (see Figs. 8–10). Figure 14 shows the variation of the rms
tip acceleration as a function of angle of attack for various ¯ and
=s D 0.2. It is seen that
y
=s D 0:4. Again there is negligible effect for ¯ D 0 deg, whereas
delays of as much as 12 deg in the buffeting onset incidence are
possibleat high jet de ection angles. The results for y
f
f
=s D 0.2 and
f
0.4 are very similarwith clear delaysin buffetingresponseto higher
angles of attack when the jet is turned on, although the buffeting
levels are somewhat smaller for y
Figure 15 shows the variation of the rms tip acceleration as a
function of angle of attack for various ¯ and y
=s D 0.6. Note that
f
=s D 0.4.
f

5
96
PHILLIPS, LAMBERT, AND GURSUL
Some experimentswere performed in order to study the effect of
nozzle geometry. In particular, the nozzle aspect ratio and the noz-
zle width to wing semispan ratio might be important parameters.
The dimensions, width/semispan ratio, and aspect ratio of several
nozzles tested are given in Table 1. The original nozzle is denoted
as case A in the table. To make a comparison of the effect of noz-
zle geometry, the momentum coefŽ cient was kept the same for all
nozzles tested. The last column in Table 1 shows the velocity ratio
U =U for the maximummomentumcoefŽcientused for the origi-
jet
1
nal nozzle(C
D 0.287).Figure16 showsthe variationof the rms tip
¹
Fig. 15 Variation of rms tip acceleration as a function of incidence for
yf /s = 0.6.
for thisŽ n locationthespanwisepositionofthe vortexaxiscoincides
with the location of the Ž n leading edge. For this head-on collision
case vortex breakdown is generally upstream of the Ž n (except for
large values of jet momentum coefŽ cient), and the overall level of
buffeting is smaller than that of y
f
=s D 0:2 and 0.4. It is seen in
Fig. 15 that the variation of the rms tip acceleration with angle of
attack is also more gradual compared to other cases. The effect of
jet blowing at the trailing edge is very small for this Ž n location for
all jet de ection angles tested.
Fig. 16 Variation of rms tip acceleration as a function of incidence for
yf /s = 0.4 and C¹ = 0.287.

PHILLIPS, LAMBERT, AND GURSUL
597
Table 1 Dimensions, width/semispan, aspect ratio,
and velocity ratio for nozzles
accelerationas a function of angle of attack for four nozzles at var-
ious jet de ection angles for C D 0:287 and y =s D 0.4. Buffeting
¹
f
response differs very little for ¯ D 0 and 15 deg, but there are large
effects of nozzle geometry for ¯ D 30 and 45 deg. At these large
de ection anglesthe originalnozzle appearsto be more effectivein
delaying the onset of buffeting, although there is an accompanying
increase in the maximum levels of buffeting.Also, althoughthe ve-
Dimensions,
Ujet=U1
2
Nozzle
mm
Width/semispan Aspect ratio (C¹ = 0.287)
A
B
C
D
60 £ 10
60 £ 5
0.45
0.45
0.45
0.22
6
12
24
6
4
5.66
8
60 £ 2.5
30 £ 5
locity ratio Ujet=U
see Table1) it providesthe bestperformance.In fact, there is a trend
1
is the smallest for the original nozzle (case A,
8
Fig. 17 Variation of rms tip acceleration as a function of incidence for
yf /s = 0.2 and ¤ = 65 deg.
Fig. 18 Variation of rms tip acceleration as a function of incidence for
yf /s = 0.4 and ¤ = 65 deg.

5
98
PHILLIPS, LAMBERT, AND GURSUL
Fig. 20 Maximum shift in incidence in buffet envelope as a function
of momentum coefŽ cient for yf /s = 0.4 and ¯ = 45 deg.
1
5, 30, 45 deg; and Ž n locations y
f
=s D 0.2, 0.4, and 0.7 (which
correspondstohead-oncollisioncaseforthiswing).Notethatforthe
samejetvelocitythedimensionlessmomentumcoefŽ cientissmaller
for 3 D 65-deg wing as a result of larger wing area. Consequently,
the maximum momentumcoefŽ cient that can be attained is smaller.
Nevertheless, the results are qualitatively similar to those obtained
for 3 D 75 deg. Except the head-on collision case, the jet blowing
appearsto beeffective,in particularathigherjet de ectionangles.A
bettercomparisonfortheeffectofwingsweepisshowninFig. 20for
y =s D 0.4 and ¯ D 45 deg. Maximum shift in incidence in buffet
f
envelope with respect to C
the momentum coefŽ cient C
¹
D 0 case is plotted as a function of
¹
for all wings. Although there is large
scatterin data (presumablybecauseof the effectof nozzlegeometry
with respect to the wing), the trend conŽ rms that jet blowing is as
effective for decreasing wing sweep angle.
Conclusions
Effect of a trailing-edge jet on the interaction of a leading-edge
vortex with a Ž n and resulting Ž n buffeting was investigated in
water tunnelas well as wind-tunnelexperiments.Flow visualization
studies showed that vortex breakdown is induced as a result of the
adverse pressure gradient imposed by the Ž n, but can be delayed
into the wake with increasing jet momentum coefŽ cient. Even for
the head-on collision of the leading-edgevortex with the Ž n, it was
possible to eliminate the breakdown completely. Delay of vortex
breakdown into the wake even for relatively small values of jet
velocityratio is possiblefor de ectedjets, whereasthe effectiveness
of the jet with no de ection is very limited. For ¯ D 30 deg, vortex
breakdowncan be delayedup to 60% of the chord lengthdepending
on the Ž n locationwith respectto the leading-edgevortex.Buffeting
response of a  exible Ž n in wind-tunnel experiments showed that
the effect of jet momentum was very small for ¯ D 0 deg, but there
was considerabledelay of the onset of buffetingto higher angles of
attack with increasingjet momentum for ¯ D 30 and 45 deg. It was
found that it was possible to shift the buffetingenvelopeas much as
1
2 deg in incidence of the wing. The wind-tunneltests showed that
blowing at the trailing edge appears to be more effective at higher
jet de ection angles, which is consistentwith the  ow visualization
results obtained in the water tunnel. It appears that jet blowing is as
effective for less slender delta wings.
Fig. 19 Variation of rms tip acceleration as a function of incidence for
yf /s = 0.7 and ¤ = 65 deg.
that jet blowing becomes less effective with increasingUjet=U
the constant C
. Clearly there remain several aspects of the effect
of nozzle geometry to be studied in future work.
1
for
The results presented in this paper show that vortex breakdown
causedby the presenceof a Ž n can be delayedby a trailing-edgejet,
with resultingattenuationof Ž n buffeting.It is suggestedthat the jet
creates a favorable pressure gradient and reduces the adverse pres-
sure gradient caused by the Ž n. However, there are several factors
that need further studies: effect of jet entrainment, vortex interac-
tions between the jet and wing vortices, and the effect of nozzle
geometry. Our limited wind-tunnel experiments showed that the
nozzle geometry is very importantand causes very differentbuffet-
ing envelopes for the same momentum coefŽ cient.
¹
Finally, some experiments were carried out to study the effect
of wing sweep. One expects that adverse pressure gradient which
exists as a result of the trailing edge depends on the slendernessof
the wing. With this in mind, detailed experiments were performed
for 3 D 65 deg wing, using the originalnozzle.Figures 17–19 show
the variation of the rms tip acceleration as a function of angle of
attack for various jet velocity ratios; jet de ection angles ¯ D 0,

PHILLIPS, LAMBERT, AND GURSUL
July 1997.
599
Acknowledgments
13Gursul, I., and Xie, W., “Interaction of Vortex Breakdown with an
This work was sponsored by the European OfŽ ce of Aerospace
Researchand Development,Air ForceOfŽ ce of ScientiŽc Research,
USAF, under Contract F61775-01-WE006. The authors acknowl-
edge the help of K. Wong in wind-tunnel experiments.
Oscillating Fin,” AIAA Journal, Vol. 39, No. 3, 2001, pp. 438–446.
Hauch, R. M., Jacobs, J. H., Dima, C., and Ravindra, K., “Reduction
of Vertical Tail Buffet Response Using Active Control,” Journal of Aircraft,
1
4
Vol. 33, No. 3, 1996, pp. 617–622.
1
5
Breitsamter, C., and Laschka, B., “Aerodynamic Active Control for
References
EF-2000 Fin Buffet Load Alleviation,” AIAA Paper 2000-0656, Jan. 2000.
1
16
Gursul, I., and Xie, W., “Buffeting Flows over Delta Wings,”AIAA Jour-
Bean, D. E., Greenwell, D. I., and Wood, N. J., “Vortex Control Tech-
nal, Vol. 37, No. 1, 1999, pp. 58–65.
Triplett, W. E., “Pressure Measurements on Twin Vertical Tails in
Buffeting Flow,” Journal of Aircraft, Vol. 20, No. 11, 1983, pp. 920–
nique for the Attenuation of Fin Buffet,” Journal of Aircraft, Vol. 30, No. 6,
1993, pp. 847–853.
2
17McCormick, S., and Gursul, I., “Effect of Shear Layer Control on Lead-
ing Edge Vortices,” Journal of Aircraft, Vol. 33, No. 6, 1996,pp. 1087–1093.
9
25.
3Lee, B. H. K., and Brown, D., “Wind-Tunnel Studies of F/A-18 Tail
18
Gursul, I., Srinivas, S., and Batta, G., “Active Control of Vortex Break-
Buffet,” Journal of Aircraft, Vol. 24, No. 1, 1992, pp. 146–152.
down over a Delta Wing,” AIAA Journal, Vol. 33, No. 9, 1995, pp. 1743–
1745.
4
Washburn, A. E., Jenkins, L. N., and Ferman, M. A., “Experimental
Investigation of Vortex-Fin Interaction,” AIAA Paper 93-0050, Jan. 1993.
19Mitchell, A. M., and Delery, J., “Research into Vortex Breakdown Con-
trol,” Progress in Aerospace Sciences, Vol. 37, 2001, pp. 385–418.
5
Meyn, L. A., and James, K. D., “Full Scale Wind Tunnel Studies of
2
0
F/A-18 Tail Buffet,” AIAA Paper 93-3519, 1993.
Helin,H. E.,andWatry,C.W.,“Effects ofTrailing-EdgeJet Entrainment
6
Lee, B. H. K., and Tang, F. C., “Characteristics of the Surface Pressures
on Delta Wing Vortices,” AIAA Journal, Vol. 32, No. 4, 1994, pp. 802–
804.
on a F/A-18 Vertical Fin due to Buffet,” Journal of Aircraft, Vol. 31, No. 1,
1
994, pp. 228–235.
21Shih, C., and Ding, Z., “Trailing-Edge Jet Control of Leading-Edge
Vortices of a Delta Wing,” AIAA Journal, Vol. 34, No. 7, 1996, pp. 1447–
1457.
7Bean, D. E., and Wood, N. J., “Experimental Investigation of Twin-
Fin Buffeting and Suppression,” Journal of Aircraft, Vol. 33, No. 4, 1996,
pp. 761–767.
22Vorobieff, P. V., and Rockwell, D. O., “Vortex Breakdown on Pitching
Delta Wing:Control by Intermittent Trailing-EdgeBlowing,”AIAA Journal,
Vol. 36, No. 4, 1998, pp. 585–589.
8Wolfe, S., Canbazoglu, S., Lin, J. C., and Rockwell, D., “Buffeting of
Fins: An Assessment of Surface Pressure Loading,” AIAA Journal, Vol. 33,
No. 11, 1995, pp. 2232–2234.
2
3
Mitchell, A. M., Molton, P., Barberis, D., and Delery, J., “Control of
9Mayori,A., and Rockwell, D., “Interaction ofa Streamwise Vortexwith a
Thin Plate: A Sourceof TurbulentBuffeting,”AIAA Journal, Vol. 32, No. 10,
Leading-Edge Vortex Breakdown by Trailing Edge Injection,” AIAA Paper
99-3202, 1999.
1
994, pp. 2022–2029.
24Ross, H., “X-31, The First Aircraft Designed for High Angle of Attack
Manoeuvring,” NATO AVT Meeting, May 2001.
10Wolfe, S., Lin, J. C., and Rockwell, D., “Buffeting at the Leading-Edge
2
5
of a Flat Plate due to a Streamwise Vortex: Flow Structure and Surface
Pressure Loading,” Journal of Fluids and Structures, Vol. 9, No. 4, 1995,
pp. 359–370.
Lambert, C., and Gursul, I., “Buffeting of a Flexible Fin over a Delta
Wing,” AIAA Paper 2001-2426, June 2001.
2
6
Lambourne, N. C., and Bryer, D. W., “The Bursting of Leading Edge
11Canbazoglu, S., Lin, J. C., Wolfe, S., and Rockwell, D., “Buffeting of
Fins: Distortion of Incident Vortex,” AIAA Journal, Vol. 33, No. 11, 1995,
pp. 2144–2150.
Vortices—Some Observations and Discussion of the Phenomenon,” Aero-
nautical Research Council, R&M 3282, London, April 1961.
2
7
Wang, F. Y., Proot, M. M. J., Charbonnier, J. M., and Sforza, P. M.,
12Gordnier, R. E., and Visbal, M. R., “Numerical Simulation of the Im-
pingement of a Streamwise Vortex on a Plate,” AIAA Paper 97-1781, June–
“Near-Field Interaction of a Jet with Leading-Edge Vortices,” Journal of
Aircraft, Vol. 37, No. 5, 2000, pp. 779–785.