JOURNAL OF AIRCRAFT

Vol. 48, No. 3, MayJune 2011

Full-Scale Flight Tests of Active Flow Control to Reduce

Tiltrotor Aircraft Download

Downloaded by UNIVERSITY OF ILLINOIS on November 23, 2014 | http://arc.aiaa.org | DOI: 10.2514/1.46956

Michael A. McVeigh
The Boeing Company, Philadelphia, Pennsylvania 19142
Hassan Nagib
Illinois Institute of Technology, Chicago, Illinois 60616
Tom Wood
Bell Helicopter Textron, Inc., Hurst, Texas 76053
and
Israel Wygnanski
University of Arizona, Tucson, Arizona 85721
DOI: 10.2514/1.46956
The vertical force, or download, acting on the airframe of current tiltrotor congurations during hover amounts to
approximately 10% of rotor thrust, or about 6000 lb for the V-22. Various mechanical means have been
experimentally tried to reduce this penalty, but none has been implemented, largely because of mechanical
complexity. This paper describes the research conducted on the application of active ow control to the problem,
since this technique may offer a solution without large weight penalties and unacceptable complexity. The research
was conducted as part of the Defense Advanced Research Projects Agency Micro Adaptive Flow Control program.
The work culminated in June 2003, when the NASA/U.S. Army/Bell XV-15 tiltrotor aircraft was used to successfully
demonstrate the effectiveness of active ow control in reducing airframe download during hover. The wing aps
were tted with zero-mass-ow actuators that periodically injected/removed air in the ap upper surface boundary
layer through slots from the interior of the ap. The active ow control was effective in delaying ow separation from
the ap, which reduced the download on the wings. The ight tests were the culmination of extensive laboratory
experiments on two-dimensional models and on a powered full-span 16%-scale model of the XV-15 aircraft. The XV15 ight tests conrmed the laboratory ndings by successfully reducing the download measured in hover by as much
as 14%, demonstrating that the aerodynamic principles of active ow control can be applied to full-scale air vehicles.

V1
W
x

f
K

Nomenclature
A
c
Cd
CP

=
=
=
=

CT
CW
C
C1
DL
f
F
h
R
T
v
Vj
VT

=
=
=
=
=
=
=
=
=
=
=
=
=

disc area of two rotors, m2

wing chord, aps up, m
2
section drag coefcient, drag=
1=2
V1
c
rotor power coefcient, 550 total rotor
horsepower=

AVT3 
rotor thrust coefcient, total thrust=

AVT3 
aircraft weight coefcient, weight=

AVT2 
momentum coefcient, 2
h=x
Vj =V1 2
2
section lift coefcient, lift=
1=2
V1
c
wing download, N
blowing frequency, Hz
nondimensional frequency, fx=V1
slot width, m
rotor radius, m
thrust of two rotors, N
rotor wake velocity, m=s, positive down toward wing
rms jet exit velocity, m=s
rotor-tip speed, m=s

=
=
=
=
=
=
=

remote windspeed, m=s

aircraft weight, N
distance from actuation slot to trailing edge, m
angle of attack, 
ap deection, 
Krueger ap deection, 
air density, kg=m3

I. Introduction
HIS paper presents a technical history of the rst application of
periodic active ow control (AFC) to a manned operational
aircraft. It describes a number of the important milestones of this
cooperative research program and focuses on some of the problems
encountered and the solutions provided.
Tiltrotor aircraft (Fig. 1) experience a large download force in
hover primarily caused by the rotor downwash impinging on the
wings and that part of the fuselage between the wings. To minimize
the download and increase net vertical lift, current design practice [1]
is to deect the aps/ailerons to an angle of about 65 to reduce the
area exposed to the rotor downwash. The resulting download is
approximately 10% of the rotor thrust. However, when the aperons
are deected to a larger angle, the ow separates in the vicinity of the
ap shoulder and the download increases substantially. Various ways
to decrease tiltrotor download have been explored via twodimensional (2-D) ap tests [2,3]. Steady blowing [4] and wingmounted mechanical devices [5] have been demonstrated to reduce
the download. However, weight and complexity have prevented
these solutions from being reduced to practice.
While steady blowing over the deected ap has been shown to
delay separation and reduce the vertical drag on 2-D apped airfoils
at large negative angles of attack, a more attractive technique is
periodic excitation [6], which provides similar benets but at much
lower energy requirements.

Received 1 September 2009; revision received 6 November 2010; accepted

for publication 6 November 2010. Copyright 2010 by the American
Institute of Aeronautics and Astronautics, Inc. All rights reserved. Copies of
this paper may be made for personal or internal use, on condition that the
copier pay the $10.00 per-copy fee to the Copyright Clearance Center, Inc.,
222 Rosewood Drive, Danvers, MA 01923; include the code 0021-8669/11
and $10.00 in correspondence with the CCC.

Senior Technical Fellow, Mail Stop P23-15, P.O. Box 16858;

michael.a.mcveigh@boeing.com. Senior Member AIAA.

Professor, MMAE Department, 10 West 32nd Street, Room 243;

nagib@iit.edu. Fellow AIAA.

Director of Flight Technology and Simulation, Mail Stop 1343, 600 East
Hurst Boulevard; twood@bellhelicopter.textron.com.

Professor, AME Department; wygy@ame.engr.arizona.edu. Fellow

AIAA.
786

787

MCVEIGH ET AL.

Institute of Technology (IIT), was tasked with the development of

these actuators to be tted inside the aps of the aircraft [7]. Under his
leadership, the IIT team successfully developed palm-sized electromagnetic actuators for the ight demonstration. This effort is detailed
in [7].

II. Active Flow Control Parameters

NASA/U.S. Army/Bell XV-15 tiltrotor in hover.

The application of periodic AFC to the tiltrotor hover download

problem was suggested by J. McCroskey, of NASA Ames Research
Center, during a visit to Tel-Aviv University where he had observed a
demonstration of stall delay on a wing section by a pulsating zeromass-ux jet emanating from a narrow slot. Heeding his suggestion,
the validity of using the same technique to reduce vertical drag was
proven by testing the airfoil at a large negative angle of attack. The
result (Fig. 2) demonstrated that periodic actuation at a reduced
frequency F  1:6 required an order-of-magnitude lower level of
blowing than steady blowing for approximately the same level of
drag reduction, i.e., 30%. These results provided the impetus for the
initiation of an extensive AFC research program using V-22 wing
section models and a powered scale model of the V-22 Osprey
tiltrotor. Most of the aerodynamic tests that followed were aimed at
reducing the download, and they were conducted at the University of
Arizona on the V-22 model in the hover mode. Based on the test
results showing that AFC was effective, a program decision was
made by R. Wlezien of the Defense Advanced Research Projects
Agency to conduct a full-scale demonstration of AFC on the NASA/
U.S. Army/Bell XV-15 tiltrotor airplane. To prepare for this, a 16%scale model of the XV-15 wing airfoil and a full-span 16%-scale
powered model of the aircraft were built to explore AFC parameters
before designing the aircraft hardware and preparing the aircraft for
hover testing. This involved close cooperation with the Bell
Helicopter Textron, Inc., XV-15 ight-test team. Finally, in order to
gain additional insight into the phenomena observed in the tests, a
large-eddy simulation (LES) computational uid dynamics (CFD)
model of the XV-15 airfoil was developed and exercised to visualize
the ow patterns and relate these to the test results.
From the V-22 model download testing, model-scale AFC
actuators were already available for use in the XV-15 model tests.
However, small, light, robust, and reliable actuators suitable for ight
were not available and had to be developed in time for the hover
testing on the XV-15 aircraft. Hassan Nagib, from the Illinois

The reduced frequency F relates the forcing frequency to the

speed of the ow:
F 

fx
V1

This term represents the ratio of the distance over which the
control of separation has to be achieved (i.e., the distance from the
slot to the trailing edge of the controlled surface) to the wavelength of
the dominant oscillation.

III. Model-Scale Actuators

Actuators used during the V-22 model tests and the initial XV-15
model tests were supplied by Domzalski Machine of Mesa, Arizona.
In the later XV-15 model tests, Nagib et al. [7] designed an improved
set of electromagnetic actuators (Fig. 3). This design used voice coils
to deliver high peak jet velocities over a wide frequency range
without the need for a cooling system. The square frames were
designed to t together to form a linear spanwise array beneath slots
cut into the upper airfoil surface cover plates at various chordwise
distances. The top surfaces of the actuator frames were sealed against

= - 115 , R e = 0.15 M
1.6
V

Drag Coefficient, C d

Downloaded by UNIVERSITY OF ILLINOIS on November 23, 2014 | http://arc.aiaa.org | DOI: 10.2514/1.46956

Fig. 1

The main parameters characterizing AFC on a surface are the

amplitude and frequency of forcing, the waveform type (sinusoidal,
burst mode, or amplitude modulated), the ow Reynolds number, the
slot width, and a characteristic surface length downstream of the
actuator. Actuators are usually placed upstream of the separated ow
region.
Assuming that separated ow is independent of Reynolds number,
provided it is sufciently high, two nondimensional parameters are
generally used to characterize the ow: a momentum coefcient and
a reduced frequency. The momentum coefcient C of a jet of width
h is the ratio of the jet momentum to the momentum of the freestream.
In the case of a purely sinusoidal jet ow, the jet momentum is
proportional to the amplitudes of the velocity oscillations. Thus,
p







2h Vjet
RMS 2
2h Vjet peak= 2 2

C 
c
c
V1
V1




h Vjet peak 2

c
V1

1.4
F+ = 0.8

1.2
Steady Blowing
F+ = 1.6

1.0
AFC

0.8

0.1

1.0

10

Momentum Coefficient, C (%)

Fig. 2 Demonstration of effect of unsteady vs steady blowing on

vertical drag reduction of a apped airfoil.

Fig. 3

Schematic of actuators.

788

MCVEIGH ET AL.

the lower surface of the cover plate. However, for the model tests,
there was not enough room in the aps to accommodate the actuators,
so they were mounted in the main element of the airfoil, and ducts
were used to channel the output to slots in the ap upper surface.
Since the mechanism for delaying separation depends on the
magnitude and frequency of the excitation, the model actuators
provided a broad range of these parameters. The general frequency
range of the coils was from 50 to 300 Hz at an input amplitude of 0 to
9 V per actuator. This in turn provided a range of C from 0 to 2.5% at
0:6 < F < 2:1 and much higher values of C at some specic
frequencies. By varying these values together with changes in
chordwise slot location and ap deection, an optimum conguration for download reduction could be determined.

Downloaded by UNIVERSITY OF ILLINOIS on November 23, 2014 | http://arc.aiaa.org | DOI: 10.2514/1.46956

IV. Large-Eddy Simulation Floweld Computations

To get an idea of the ow patterns and pressures to be expected in
the tests, calculations were performed using LES [8] for an airfoil at
an 85 angle of attack with a ap deected at 80 . Some of the
calculated instantaneous pressure contours are shown in Fig. 4.
During hover, the airfoil is a bluffbody for which the drag is
associated with strong vortex shedding. The pressure contours, at the
time when the oscillatory Cd is close to its maximum, show the
presence of a large vortex close to the base of the airfoil. This vortex is
created by the roll up of the mixing layer that separates the dead air in
the lee of the airfoil from the rotor downwash that surrounds the
wing. This mixing layer is susceptible to the KelvinHelmholtz
(KH) instability that creates the smaller array of vortices visible
below the leading and trailing edges of the airfoil. The strength of
the KH vortices increases initially with increasing distance from the
separation location before dissipation sets in and weakens them. The
instability of the vortex array generates a large circulation pattern that
rotates in the counterclockwise direction in the mixing layer created
near the leading edge of the airfoil. Near the trailing edge, the
circulation is in the clockwise direction.
These circulation patterns affect the instantaneous pressure
distribution around the airfoil and cause an oscillation in the
stagnation location on the upper surface that results in periodic vortex
shedding. When the KH eddies are created naturally, due to random
disturbances in the ow, they vary in strength and are more
susceptible to early roll up near the base of the airfoil. This generates
low pressure on the bottom surface of the airfoil and results in large
drag oscillations. The contours shown in Fig. 4b correspond to an
instantaneous minimum in Cd , because the large circulation pattern
(Fig. 4a) has been swept downstream, and only some shear layer
(KH) vortices are present close to the base of the airfoil. In Fig. 4c,
eddies created by controlled periodic excitation (AFC) at a point near
the ap shoulder are more regular and resistant to the secondary
instability and large roll up. The pressure contours of Fig. 4c suggest
that a weak circulation pattern might have been created one chord

a) Baseline at maximum
computed cd

length below the airfoil. Such a pattern has only a minor effect on the
pressure distribution over the airfoil and will substantially reduce the
oscillations in drag. Two additional observations may be made by
comparing ow patterns in Figs. 4a and 4b with Fig. 4c:
1) The periodically excited ow stays attached to the ap instead
of separating from it.
2) Periodic excitation applied to the trailing-edge ap effectively
regulates the shedding of vortices created in the mixing layer formed
near the leading edge of the airfoil. Thus, application of additional
excitation near the wing leading edge must be coordinated with the
phase of the excitation on the ap in order to be effective.

V. Wind-Tunnel Tests
Tests were made on 2-D models of a apped XV-15 airfoil in
which various means to reduce vertical drag were evaluated. The
more promising approaches were then tested in a hover download test
rig using the powered model of the XV-15.
A.

Two-Dimensional Airfoil Vertical Drag Tests: Test Setup

Experiments [9,10] were conducted on a 2-D 10.08 in. chord (16%

scale) apped XV-15 wing airfoil model in University of Arizonas
24 by 41 in. wind tunnel. Part of each sidewall was removed to avoid
large blockage corrections (Fig. 5). The sidewalls terminated
upstream of the airfoil at the scaled height of the rotor above the wing
of the airplane. The sidewall opening allowed the ow to go around
the wing section as if in free air.
The XV-15 wing airfoil (NACA 64A223 mod) spanned the 24 in.
height of the tunnel. It was machined from aluminum and provided
with a chordwise row of pressure taps at the center span. Five 0.04in.-wide inline spanwise slots were machined into the ap upper
surface for uidic actuation. Each slot was spaced apart by a short
distance from its neighbor to preserve the structural integrity of the
surface. The interruption of the slots caused only a small local
reduction in the momentum of the jets.
The 25% chord plain ap was hollow to accommodate AFC
actuation ducting. The interior of the main airfoil, also hollow, served
as a settling chamber for actuation from outside the tunnel or
accommodated a spanwise array of internal actuators.
The ducts from the actuators to the slots and all their connections
were identical. The actuators were all manufactured with identical
impedance and were designed to have the same performance so as to
give as much spanwise uniformity as possible. The slots delivered the
air at an angle of approximately 30 to the ap upper surface, which
was determined to be optimal based on prior testing. The amplitudes
of the jet oscillations from each slot were measured, and the actuators
were matched so that their sensitivity to input voltage did not vary by
more than 5% from slot to slot. The sensitivity to frequency
variation was greater, amounting to 10% for frequencies of 160 Hz
and higher. This variance was halved at forcing frequencies that were

b) Baseline at minimum
computed cd

c) Periodic excitation applied from

a slot at 10% flap chord

Fig. 4 Computed instantaneous pressure contours showing characteristic structures in oweld.

789

MCVEIGH ET AL.

Krueger Flap
41 in.

Tunnel
Wall
10 in. Chord
(Flaps Up)

Fig. 8 XV-15 airfoil with Krueger ap.

No Walls
Tunnel
Wall

Fig. 5

2-D airfoil in partially open wind-tunnel test section.

Test Results Without Krueger Flap

The 2-D tests explored the effect of actuator-slot position,

amplitude of the periodic excitation C , its frequency F, ap
deection angle, and the airfoil angle of attack with respect to the
tunnel velocity, representing the rotor downwash velocity.
Based on data from wake surveys, the vertical drag on the wing
sections was about 95% pressure drag, so the measured pressure drag
was taken as representative of the total vertical drag. The normalized
drag and lift coefcients on the 2-D model (Fig. 6) were obtained at a
representative   85 . The Reynolds number was 250,000, and
the slot was located at 10% ap chord, which was upstream of the
observed ow separation point. The download without actuation (the
slot was taped over) decreased with increasing f , reaching a
minimum value of 0.72 at 55 , and then it increased at higher angles

Normalized Lift and Drag Coefficients

= - 85 , R e = 0.25 x 106, F + = 0.95

Reference Value

1.0

10% Slot Location

0.8

C
0% - Slot Closed

0.6

0% - Slot Open

C = 0.3

0.3
D

CD

C.

0.25%

0.4

0.50%
0.2

CL
C
L

1.45%

CL = 0.24

2.90%
0
0

20

40

-0.2

60

80

100

Flap Angle,

Fig. 6

2-D airfoil section drag and lift, with and without AFC.

Test Results with Krueger Flap

A Krueger ap (Fig. 8) was attached to the leading edge of the 2-D

airfoil and evaluated for its effect on reducing the download [11]. The
length of the Krueger was approximately 25% of the chord and was
shaped to conform to the front portion of the lower surface of the
airfoil rearward of 10% chord. Figure 9 presents the variation of Cd
with ap deection for three Krueger settings of 90, 100, and 110 ,
with and without AFC applied to the ap or to the Krueger.
The passive Krueger (i.e., no AFC applied), when deected at 90
to the chord, showed a small reduction in drag with the minimum
occurring at 63 ap deection. When strong AFC (C  3:4%) was

= - 85/Flap = 80/F+ = 0.59

- 0.05

Slot
Position:

7%

- 0.10

35%
15%

- 0.15
- 0.20

10%

- 0.25
Delta Drag Based on Lower Surface 50% Chord Pressures
- 0.30

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Normalized Vertical Drag Coefficient, Cd

B.

Normalized Delta Drag Coefficient, Cd

Downloaded by UNIVERSITY OF ILLINOIS on November 23, 2014 | http://arc.aiaa.org | DOI: 10.2514/1.46956

below 100 Hz. Following a few tests, the ow through the slots was
checked again to make sure that there was no degradation in actuator
performance.

due to increasing separation to a constant Cd 0:83 at f > 70 .

When the slot was opened (but with no AFC applied), internal cavity
resonance, which included the sealed ducting, further reduced the Cd
to 0.61 by maintaining attached ow on the ap up to f  70 .
When AFC was applied at a reduced frequency of F  0:95, and
the minimum Cd was moved to f 80 , depending on the value of
C . The largest reduction was with C  2:9%, which reduced the
relative drag by
Cd  0:3 (41%) compared with the minimum
without AFC. Lowered section drag was accompanied by an increase
in the orthogonal lift force (lower part of Fig. 6), which is a chordwise
force toward the rear of the aircraft sometimes referred to as
suckback. This increased from about zero at 55 < f < 60 to a
normalized Cl  0:22 at f  80 . On the tiltrotor aircraft, this force
would be canceled by a small amount of forward nacelle tilt.
Determination of the optimal slot location, and the best values of
C and F , would have required many measurements of the
chordwise pressures distributions. To avoid this, the drag was derived
from the Cp measured at a single pressure tap located at midchord on
the lower surface of the airfoil. Since the pressure over the airfoils
lower surface was constant and the maximum Cp over the upper
surface was unity (stagnation), the
Cp was used to determine the
drag accurately and consistently following a short calibration
process.
The effect of C on the section drag is presented in Fig. 7 for
various locations of the slot at a constant value of F  0:59. Note
the large reduction in drag at small values of C that occurred with
the slot positioned at 7 and 10% chord and the subsequent
insensitivity of Cd to further increases in C . For the more rearward
slot positions of 15 and 35%, however, increasing C above 0.5%
continued to reduce the drag signicantly but not below the levels
attained with the slot at 10% chord.

0.9

Passive Krueger

Active Krueger

0.7

Active Flap and

Passive Krueger

0.6
0.5

Both Flap
and Krueger
Active

Slot at 10% Flap Chord

0.4
0.3

Krueger Off

Baseline / Krueger Off

Krueger 90

0.2

Krueger 100

0.1

Flap
AFC On,
C = 3.4%

AFC Off / Krueger 90

Krueger AFC On / C = 3.4%/ Krueger 90
Both AFC On / C = 3.4%/ Krueger 90

Krueger 110

0.0
40

Momentum Coefficient, C (%)

Fig. 7 Effects of slot location and C
on the download of the 2-D section.

Krueger Off

0.8

45

50

55

60

65

70

75

80

85

90

95

100

Flap Deflection,

Fig. 9

Effect of Krueger ap on vertical drag of XV-15 airfoil section.

790

Normalized Drag Coefficient, C d

MCVEIGH ET AL.

Angle of Attack = -80

Krueger Angle = 95
Flap Deflection = 55
C = 0.5%
Slot at 10% Flap Chord

0.8

0.7

0.6

26%
26%

0.5

0.4

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Nondimensional Frequency, F +

Variation of Cd with actuation frequency F
.

Fig. 10

0.2

Normalized Cd , CL

Downloaded by UNIVERSITY OF ILLINOIS on November 23, 2014 | http://arc.aiaa.org | DOI: 10.2514/1.46956

C = 1.5%, R e = 1.33 x 105, F + = 0.7

CL
0.1

Test
-0.1

Cd
-0.2
0

10

15

20

25

30

35

Slot Position, % Chord

Fig. 11 LES-computed variation of
Cl and
Cd with slot location.

applied to only the Krueger near the leading edge, a drag reduction of
almost 30% occurred compared with the baseline with no Krueger
and AFC off. With AFC applied to the ap but not to the Krueger,
minimum drag occurred at a ap setting of 83 , regardless of Krueger
angle. When both the ap and the Krueger had AFC applied to them,
a reduction in minimum drag of about 50% was achieved. Active
control over the Krueger ap and no control over the trailing-edge
ap provided approximately the same download alleviation as active
control over the ap, with or without the Krueger; that is, when AFC
was applied to the ap and the Krueger was passive, there was no
benet to adding the Krueger.
The effect of actuation frequency on the drag reduction with
C  0:5% is shown in Fig. 10. For this Krueger conguration, the
minimum drag occurred in the neighborhood of F  0:5.

The location of the oscillatory jet plays a large role in the efciency
of AFC. In Fig. 11, values for download reduction and suckback
force as a function of slot position are shown. These values were
obtained using the LES analysis, and they compare favorably with
the experimental data. Note that the LES predicted the optimal slot
location to be around 10% ap chord.
The time-averaged pressure distributions over the airfoil and ap
for the unforced and forced cases are shown in Fig. 12, where the
experimental results are indicated by interconnected symbols, and
the results obtained by the LES are represented by smooth curves.
There is good agreement between the LES and the experimental
results, validating the calculation technique. The impingement of the
ow on the upper surface results in a positive Cp over the main
element. Stagnation pressure (Cp 1) occurs near the midchord of
the main element. The ow then accelerates toward the leading edge
and toward the ap. The ow separates at the main airfoil leading
edge and at the upper surface of the ap shoulder due to the large ap
angle.
The negative base pressure on the lower surface of the airfoil is
almost constant (Cp < 0), and it is approximately the same over the
separated region of the ap upper surface. AFC applied to the ap
upper surface increases the base pressure on the airfoil lower surface
and substantially lowers the pressure over the upper surface of the
ap by attaching the ow over the shoulder. Both effects reduce the
download.
The Krueger ap effects predicted by the LES computations
(Fig. 13) showed relatively little effect when the ow separated from
both aps, but they showed a large effect when periodic excitation
reattached the ow. When the wavelength of the excitation was
commensurate with the length of the ap, the calculated Cd was 0.4,
while the measured value was 0.38. The computed pressure contours
again indicate that forcing the ow at a suitable frequency regulates
and equalizes the KH eddies that are formed in the mixing layer,
making them more resistant to roll up into a large recirculating eddy
below the airfoil. Coupled with ow reattachment over both aps,
this results in the observed drag reduction.

VI.
A.

Powered XV-15 Model Tests

Test Rig

The powered XV-15 model download experiments were

conducted in the Aerodynamics Laboratory at the University of
Arizona. The laboratory was climate controlled, with bare concrete
oors and walls and a high ceiling. The XV-15 tiltrotor model was
inverted and suspended from an A-frame structure, directing the rotor
downwash up toward the ceiling, which was located more than ve
rotor diameters away (Figs. 14 and 15). This arrangement minimized
the effects of the ground in the hovering condition, and the A-frame
structure provided minimal interference to the ow in the vicinity of
the wing. The motor and drive system was placed on the oor, with

Fig. 12 Calculated and measured Cp distribution.

Downloaded by UNIVERSITY OF ILLINOIS on November 23, 2014 | http://arc.aiaa.org | DOI: 10.2514/1.46956

MCVEIGH ET AL.

a) No excitation

791

b) Excitation from the shoulders of both flaps

Fig. 13 Computed pressure contours for XV-15 airfoil with Krueger ap.

gearing providing the rotational motion up through two stacks to

each rotor. The rotors were not connected to the nacelles to allow for a
direct measurement of the aerodynamic forces on the model. The gap
between the rotors and the nacelles was less than 2 in., or 4.2% of the
4-ft-diam rotors.

The bottom of the fuselage was 18 in. below the crossbar. The
model consisted of a fuselage, tails, wings, adjustable ailerons and
aps, nacelles, and rotors. The bottom surface of the fuselage had
hard points to attach a 0.5 in. aluminum plate. A tripod xture
suspended from the A-frame was attached to the aluminum plate by

Fig. 14 XV-15 model download rig.

Fig. 15 XV-15 model and thrust balance schematic.

MCVEIGH ET AL.

B.

Test Results

to which the ow over the wing sections was 2-D, surface ow

visualization was employed using tufts and surface oil ow (china
clay). The wing leading-edge separation patterns created by the rotor
wake are shown in Fig. 16a, which indicated that the separation from
the wing leading edge was uniform along the span.
Separation from the deected trailing-edge ap (it had a seal
between the main element and the ap surface) appeared to be 2-D,
except just inboard of the engine nacelles (Fig. 16b) where the ow
separates at much smaller ap deections. It is not clear if separation
propagated evenly from the trailing edge when the ap angle was
increased or if it did so along the span from the tip to the root of the
wing, because a very small increase in ap deection resulted in
totally separated ow over the ap. The relative two-dimensionality
of the separation line over the wing suggests that the 2-D airfoil
results were representative.
To judge the effectiveness of AFC, a set of measurements without
AFC were taken with the rotor wake impinging on the wing at an
angle of attack of 88 for which aircraft ight-test data were
available. The wing airfoil was clean and had actuation slots sealed
by a thin tape pulled tightly to maintain the surface geometry.
Moving the slot to different positions did not affect the measurements
when the tape was in place. With the tape removed, it was noticed that
the slot would introduce the effect seen in the 2-D tests; that is, an
open slot leading to a closed cavity could cause cavity resonance and
passively control the separation. However, comparing the taped and
open slot cases showed only slight differences that did not exceed the
scatter in the data.
In the 2-D tests, the addition of the Krueger ap at the leading edge
was shown to reduce the download without AFC. This also occurred
in the 3-D tests (Fig. 17), where the ow on the Krueger stayed fully
attached, as observed with tufts, providing download reduction up to
K  110 . Beyond 110 , the Krueger had increasing amounts of
separation that reduced its effectiveness. At a ap setting of 60 ,
separation was most sensitive to the Krueger deection angle,

0.120
Flap and Aileron Slots Not Taped

2200 rpm

0.115

Download / Thrust

three load cells that directly measured the total aerodynamic

download force. The load cells were arranged in a T-shaped conguration to provide the vertical load, rolling moment, and pitching
moment.
The rotors were 4-ft-diam wooden replicas of the 25-ft-diam XV15 rotor blades incorporating the same twist, planform, and airfoils.
Each set of three blades was set in a clamshell hub that connected to a
vertical driveshaft. Blade collective pitch was set by hand. A balance
mounted on each rotor stack measured rotor thrust. Transducers
installed at the base of each stack, inline with the horizontal drive
shaft, measured rotor torque. This allowed the port transducer to
measure the port rotor torque and the starboard transducer to measure
the torque of both rotors. Subtraction yielded the torque of the
starboard rotor. A 40 hp variable-frequency electric motor drove both
rotors through a linking horizontal cross shaft. The maximum
allowable continuous operating speed of the wooden blades was
2400 rpm, giving a rotor-tip speed of 500 fps.
AFC was provided by a set of ve voice-coil actuators mounted
inside each wing. The exible diaphragms and ducts mated each
actuator to its designated slot. The actuators generated an oscillating
jet at each slot on the ap surface via exible ducts from the wing to
the inside of the ap and then to the wing slots. Three slots were
located on each aileron, and two were along each inboard ap. The
jets from the slots were angled at approximately 30 to the airfoil
surface, a value determined to be effective from past experiments.
Note that the aps and the ailerons were deected at the same angle in
all these tests. A slot was also provided in the wing leading edge, and
the output from the rear faces of the actuators was ducted to this slot
in the same way, as the ducts from the front faces led to the aps. This
enabled periodic blowing and suction over a Krueger ap that could
be attached to the leading edge below the slot. A Krueger ap is not a
part of the XV-15 conguration, but it was planned to be evaluated in
ight if the powered-model experiments indicated merit, although
the 2-D test with a passive Krueger with an active ap did not yield
positive results (Fig. 9).
The test procedure accounted for balance drift caused by
temperature changes during a data run. The model was rst warmed
up at the desired rpm for about 60 s to allow the ow and the balance
readings to stabilize. Data was acquired for 40 s, and then the motors
were shut down. When the rotors came to a stop, a windoff data point
was recorded to see if the load cells had drifted during the run. The
variance in the download force measurements for 20 repeated data
points for each of the three ap deections selected was approximately 1.5%.

The applicability of the 2-D results to the three-dimensional (3-D)

powered model was of concern at the outset. To determine the extent

No Krueger

0.110

Krueger 105

0.105

Krueger 110

0.100

Krueger 120

0.095
0.090
0.085
0.080
55

60

65

Wing leading-edge

70

75

80

85

90

95

Aileron and Flap Angle,

Fig. 17 Effect of Krueger on download of powered XV-15 model with

no AFC.

0.160
XV-15 Aircraft

0.150

a) Separation from the leading edge

Download/Thrust

Downloaded by UNIVERSITY OF ILLINOIS on November 23, 2014 | http://arc.aiaa.org | DOI: 10.2514/1.46956

792

16% XV-15 Model

(2200 rpm, 13.6 )

0.140
0.130
0.120
0.110
0.100
0.090

b) Separation from the flap; the nacelle is on the

right- hand side of each picture
Fig. 16 Flow visualization on right-hand wing of powered model.

10

20

30

40

50

60

70

80

Aileron and Flap Angle,

Fig. 18 Comparison of baseline download between 16% poweredmodel and full-scale aircraft.

793

C.

Comparison of Model Results with Full-Scale Aircraft Data

Download data obtained on the 16% XV-15 model and ight-test

data from the XV-15 airplane are compared in Fig. 18. The ight-test
value of download was deduced from measurements of rotor torque
and aircraft weight using an estimate of installed rotor gure of merit.
With CT matched, the model data agreed with the ight results. This
indicates that a scale model can realistically simulate ight
conditions in hover. Therefore, it is assumed that AFC results at
model scale should give reasonably accurate predictions of what
AFC would give in full scale also.
D.

Results with Active Flow Control Applied

The rst AFC tests that were conducted used AFC on the ap only
(Fig. 19). The intention was to conrm that AFC would work in 3-D,
as it was shown to work in 2-D and to map out its effectiveness. The
angle between the rotor thrust axis and the wing chord was set to
88 . At 2200 rpm and CT  0:016, baseline DL=T was 10.6% at
60 ap deection [9]. With AFC turned on at C  3:2%, the
download was reduced by 13%. Further reducing C to 1.25% only
reduced the download alleviation to 12%.
E.

0.125
90 Hz
90 Hz
90 Hz
90 Hz
180 Hz
180 Hz
Krueger 105 180 Hz
Krueger 105 180 Hz
Krueger 105 180 Hz

Flap Deflection = 74o, 2200 rpm

0.120

Download / Thrust

because this is the ap deection that pulls the stagnation point

farthest from the leading edge. The Krueger was also sensitive to the
smoothness of the transition region between the airfoil and the
Krueger leading edge. Without AFC, the Krueger alone provided a
download reduction of about 10% for the 3-D case, although it was
less effective for the 2-D case (compare the minimum download
shown in Figs. 6 and 9. However, the 2-D experiments showed that
the Krueger maintained attachment up to about k  63 (Fig. 9),
while in the 3-D case, it had no effect on the ap deection at which
the ow separated. The discrepancy in separation angles is attributed
to the swirl angle and unsteady ow caused by the rotor in contrast
with the uniform 2-D ow from the wind tunnel.
Also interesting to note in Fig. 17 is the difference in slope of
DL=T vs ap deection, with and without the Krueger. Without the
Krueger, the slope is much higher from f  65 to 70 , after which it
becomes atter as the ap fully separates. The presence of the
Krueger may have had the effect of shifting the stagnation point more
toward the leading edge and changing the pressure distribution over
the ap. This effect was best seen as the ap entered incipient
separation. The attached Krueger developed a forward force that
opposed the attached ap suckback force, effectively canceling it.
This may be important when AFC is applied to the ap, since it
increases negative pressure on the ap surface, resulting in a
substantial rearward force.

0.115
No AFC Applied

0.110
0.105
0.100
0.095
0.090
0.085
0.080
0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Momentum Coefficient, C (%)

Fig. 20 Effect of C
on XV-15 model download/thrust with passive

Krueger.

stagnation location on the airfoil upper surface due to the high

negative pressure on the activated ap. This, in turn, would affect the
ow over the leading edge, perhaps separating the ow over the
Krueger. Therefore, the rst tests conducted on the powered model
with the Krueger had AFC applied to the ap only, with the Krueger
set to an angle K where the ow was attached as seen from the 2-D
baseline data.
The results were not the same as in the 2-D case. The passive
Krueger with AFC on the ap performed better than with no Krueger
for the same level of AFC, provided that the Krueger angle K was
properly set. In this case, the most benecial setting was 105 . The
unexpected passive Krueger benet with an activated ap indicates
that a Krueger ap could be benecial in reducing download on a
tiltrotor aircraft if employed either without the complexity of
leading-edge actuation or used with leading-edge actuation but no
ap AFC. To examine this further, a C sweep at various K was
made. The results indicated that the value of a passive Krueger ap
was not reduced by the application of AFC over the trailing-edge
ap, even as the C from the ap slot was increased. Figure 20 shows
data with k  105 and f  74 . Note that a threshold value of
C  0:5% was necessary to achieve most of the benets from apmounted AFC. Additional increases in C gave only marginal
benets.
When comparing these data to the minimum DL=T of 0.107,
obtained on the basic model without a Krueger ap or AFC, AFC
alone was responsible for a 16% reduction in download. In
conjunction with a passive Krueger ap, the benet increased to
23%. Figure 21 shows that, even with the passive Krueger set to
K  120 , where the ow over the Krueger was separated, some
improvement beyond an activated ap alone was evident.

Effect of Passive Krueger Flap

The results in Fig. 17 showed that the Krueger with naturally

attached ow would provide a benet without AFC. However, 2-D
studies indicated that with AFC on the ap, there was no observable
benet from the passive Krueger, despite its considerable value for
the baseline (Fig. 9). This might be explained by the shift in the

F.

0.120
No AFC

0.115

Full-Span AFC at 10% Flap Chord

120o Krueger

Slot at 10% Flap Chord

0.115
Baseline Untaped, 2200 rpm

0.110

1.25% C, 2200 rpm

3.2% C, 2200 rpm

0.105

With AFC

0.100
0.095

Download / Thrust

0.120

Effect of Active Krueger Flap

The benets achieved on the airplane model with a passive

Krueger and an active ap, which were not seen in the 2-D tests, were
sufciently encouraging to test with AFC applied to the Krueger ap.
The actuator setup was such that one actuator powered both the ap

0.125

Download / Thrust

Downloaded by UNIVERSITY OF ILLINOIS on November 23, 2014 | http://arc.aiaa.org | DOI: 10.2514/1.46956

MCVEIGH ET AL.

0.110

No AFC, No Krueger

0.105
0.100
0.095

No Krueger,
3.4% C

0.090

Passive Krueger,
3.4% C

0.085

Active Krueger,
3.0% C

0.080
0.090
0.085
50

0.075
50
55

60

65

70

75

80

85

90

Aileron and Flap Angle,

Fig. 19 Download on XV-15 16% powered model with AFC.

55

60

65

70

75

80

85

90

Aileron and Flap Angle,

Fig. 21 Effect of simultaneous actuation on trailing-edge ap and

Krueger.

794

MCVEIGH ET AL.

Cove Cover

slot and the leading-edge Krueger slot on the model. This was done
by exposing the bottom sides of the actuator diaphragms to the
Krueger slot while their top sides were connected to the trailing-edge
ap. Consequently, on an actuator upstroke, compressed air would be
ejected from the ap slot and sucked in through the Krueger slot. The
Krueger was then inherently out of phase with the ap slot for all the
tests, which were also conned to a single frequency of excitation.
Since the ow over the Krueger ap was naturally attached up to
K 110 , the Krueger ap was deected to a passive separated
condition with K  120 . The resulting full-span periodic actuation
over both Krueger and ap yielded a download reduction of 27%
(Fig. 21).

Downloaded by UNIVERSITY OF ILLINOIS on November 23, 2014 | http://arc.aiaa.org | DOI: 10.2514/1.46956

XV-15 Flight Tests

Full-Scale Actuator Development and Integration

Nagib and his team at IIT led the full-scale actuator development
and this effort resulted in a compact and reliable actuator that met the
ightworthiness requirements [7]. The actuators developed for the
XV-15 ight tests were larger and more powerful versions of
the same basic design as the model-scale actuators. Four goals drove
the design. First, the actuators were to be placed as close to the slot
exit as possible to minimize losses. Second, the actuator width was
sized so that the actuator arrays would cover as much of the ap span
as possible. Third, the actuator depth was made as large as possible
without interfering with the structural or the functional components
inside the ap leading edge. Fourth, the actuators were designed and
mounted so that they could be easily and quickly replaced in the event
of a failure.
For the ight tests, each actuator was tted into the ap and placed
against a nose block that contained a specially shaped conduit from
the actuator to the slot on the ap surface. The conduit accelerated the
ow to the slot in the blowing phase, and during the suction phase, the
shape of the conduit was such that ow separation within it was
avoided.
With the dimensions of the actuator dened, the internal actuator
components, the coil, magnet, armature, and piston were optimized
to give the required forcing amplitude of 60 to 100 m=s peak velocity
at the slot exit over the desired frequency range 50100 Hz. A
number of factors were considered in the material selection and
component design. These factors include the signicantly higher
forcing amplitude that was required compared with the model, the
lower frequency range, and the more hostile operating environment.
A larger coil was used, along with a larger, stronger magnet and
armature. The piston was designed to allow longer coil travel. New
high-temperature adhesives were used. As with the actuators in the
model, extensive nite element modelling was used to optimize the
design of the magnet, armature, and actuator frame.
A risk-reduction task was designed to ensure smooth integration of
the actuators into the aircraft and to verify actuator performance and
durability under simulated ight conditions. The task involved
construction of a full-scale model of a 2-D section of the XV-15 ap
with actuators and AFC slot, and testing of the model in the National
Diagnostic Facility (NDF) wind tunnel at IIT at the expected ight
conditions.
The full-scale ap section model (Fig. 22) was designed to hold up
to ve actuators. The model was installed in the NDF test section with
a truncated main wing element that was designed to give separated
ow conditions beneath the ap and main wing similar to those seen
in the powered-model tests. This would produce representative
pressure differences across the actuators diaphragm. The ap and
main element were instrumented with pressure taps. The AFC slot
location and design used on the ap model were identical to the
location and design to be used on the aircraft that was based on the
16% model test results. The interior of the ap model was designed to
closely simulate the interior of the aircraft aps. The wiring and
connectors used in the ap model were identical to those used on the
aircraft. This arrangement allowed troubleshooting of any problems
(such as overheating) under simulated ight conditions.
The pseudoap test was also used to optimize the performance of
the actuator arrays and to estimate their power requirements. Pressure

Actuator
AFC Slotat10% Chord

Cove
Drive
shaft
Flap Nose

Fig. 22

n.
4i

n.
5i
1.7

VII.
A.

AFC Slot

Full-scale pseudoap and the design of the slot on the aircraft.

measurements were made on the surface of the model, allowing

determination of the frequency ranges and forcing amplitudes that
were expected to have the greatest effect on the separation. The
operating environment inside the actual XV-15 aps was simulated
and the thermal operating limits of the actuators were determined.
The pressure difference across the actuator pistons was measured
under these conditions and the actuator operating limits (maximum
amplitude for each frequency) established for different pressure
differentials.
Two commercially available audio ampliers were used to supply
power to the actuators on the aircraft. One amplier was used to drive
the set of 26 actuators on each wing. The ampliers used a highefciency power supply and were internally modied to give the
required output power (approximately 100 W per actuator, i.e.,
2600 W per side) without overloading the power inverters on the
aircraft. A wiring study was performed to determine the optimal
wiring arrangement for the actuators. Detailed impedance measurements were made on individual actuators as well as groups of
different numbers of actuators. Impedance was measured over a
range of frequencies and operating amplitudes. The chosen wiring
conguration gave the optimum distribution of available power
across the span of the wing.
The actuators were installed in the aps of the XV-15 before
installing the aps on the aircraft, Fig. 23.
A detailed series of benchtop calibrations were performed to
conrm the performance of the actuators, to document the total
impedance, and to test the wiring conguration. The actuators were
installed in bays, 2 in.each ap and 4 in.each aileron. Following
installation their output was calibrated in situ using a hot wire,
Fig. 23. A special AFC controller was developed for the aircraft and
was mounted in the instrument panel. Several calibrations
documented the performance of the actuators for different settings
of the AFC controller. They also documented the spanwise variation
in forcing amplitude, the effects of leakage around the actuator
frames, and the effects of varying the venting of the ap interior.
These tests also allowed the operating parameters of the actuators for
the ight tests to be nalized and the actuator health monitoring

Flap Leading Edge

Actuators
Traversing Probe Holder

Jet Calibration Jig

Flap
Leading
Edge

Hot-Wire Probe

Fig. 23 Actuator arrangement and slot calibration jig for the XV-15.

795

MCVEIGH ET AL.

Bell XV-15 during AFC test ights and the arrangement of actuators.

B.

XV-15 Flight-Test Results

The testing took place at Bells Arlington, Texas, ight-test

facility. The XV-15 aircraft was tted with special-purpose aps and
ailerons, built by Bell for the test, Fig. 24. The aps contained the
actuators built by Nagib et al. [7], and the airplane was qualied for
hover and very low-speed forward ight only, because the normally
independent ailerons and aps were connected to simulate the
arrangement on the V-22. Each wing contained 26 actuators: 10 on
the ap and 16 on the aileron. These were supplied with power via
wires through the aps connected to power-conditioning ampliers
driven by the aircraft electrical generators in the fuselage. The
actuators fed air to the slots positioned at 10% chord. The width of the
slots was nominally 0.045 in. A control panel in the cockpit enabled
the pilot to select the frequency and overall blowing output strength
of all of the actuators.
The performance goals, which were established before the ight
tests, are illustrated in Fig. 25, using data from the powered-model
testing. They were 1) to demonstrate that AFC could be used to
reattach fully separated ow on the ap and 2) to produce a useful
reduction in the minimum download by maintaining attached ow on
the ap to higher ap angles that was hitherto possible. A reduction
of 14% or 220 lb was targeted for goal 1, and 9% or 150 lb was
targeted for goal 2. The aircraft was weighed with full fuel, and the
weight of the pilots and their baggage was also determined before
each ight. A fuel ow totalizer gave the amount of fuel consumed,
and hence the weight at any time. Testing was conducted early in the
morning when the winds were less than 4 kt. The test procedure was
to hover at a xed altitude out-of-ground effect with a selected ap
setting, with and without AFC. Data were taken rst with AFC off,
and then it was switched on and nally off again before changing the
amplitude. This procedure was repeated for a range of ap settings,
rotor rpm, C, and F . Video records of tufts on the ap trailing edge

(Fig. 24) showed attachment when the AFC was turned on and
separation when it was turned off.
The essential results are summarized in Figs. 26 and 27. The
curves of the variation of the rotor power coefcient with the weight
coefcient were constructed from data taken at different rpm settings
for a given ap deection. Figure 26 shows the results applicable to
success criterion 2. The AFC-off case is for a ap deection of 70
corresponding to the minimum download condition without AFC.
The AFC-on case is shown for a ap deection of 75 , corresponding
to the minimum download with AFC engaged. The AFC-on curve is
below and to the right of the baseline curve, indicating that, with
AFC, more weight can be carried for a given power. The resulting
download reduction is approximately 150 lb, or 9%, which met the
goal for criterion 2.
Figure 27 shows the results applicable to criterion 1. Here, both
curves are for a ap setting of 75 . At this angle, the ow over the
aps was fully separated without AFC. With AFC applied, the ow
on the ap was reattached, as shown by the reduced power required to
0.00150
0.00145
Success Criterion 2: ~ 150 lb

AFC Off, f = 70

0.00140
0.00135

AFC On

f = 75

0.00130
0.00125
0.00120
0.00115
0.00110
0.0104

0.0108

0.0112

0.0116

0.0120

Fig. 26 XV-15 ight-test results; AFC performance for success

criterion 2.

0.00150
f = 75

AFC On (Flap Only)

0.00145

Power Coefficient, C P

Success
Criterion 1

Success
Criterion 2

0.00140

AFC Off
Success Criterion 1: ~ 220 lb

0.00135
AFC On

0.00130
0.00125
0.00120
Effects of AFC in attaching
the flow persisted for a short
duration during On/Off operation

0.00115
0.00110
0.0104
20

30

40

50

60

70

80

Flap Deflection,

Fig. 25 Denition of ight-test success criteria.

90

0.0124

Weight Coefficient, C w

AFC Off (Flap Only)

Download on XV-15 Wing

Downloaded by UNIVERSITY OF ILLINOIS on November 23, 2014 | http://arc.aiaa.org | DOI: 10.2514/1.46956

systems to be tested. Health monitoring included thermocouple

measurements of actuator temperature and dynamic pressure
measurements inside the actuator bays in the aps.

Power Coefficient, C P

Fig. 24

0.0108

0.0112

0.0116

0.0120

0.0124

Weight Coefficient, C w

Fig. 27 XV-15 ight-test results; AFC performance for success

criterion 1.

Downloaded by UNIVERSITY OF ILLINOIS on November 23, 2014 | http://arc.aiaa.org | DOI: 10.2514/1.46956

796

MCVEIGH ET AL.

hover at the same weight and as conrmed by tuft observation. At

high (constant) values of power coefcient, the reduction exceeded
the 220 lb or 14% criterion, marked by the arrow in Fig. 27, which
again met the goal.
In both the ight and model tests, the procedure was to measure the
download with AFC off, then with it on, and again with it off. If the
process was reversible, the off data would match. However, it was
often noted during the ight tests that the effects of AFC lingered on
after AFC was switched off, sometimes long enough to force the
pilots to repeat the data or to descend slightly before taking the
assigned third reading. It is known [12] that the C required to force
reattachment of separated ow may be as much as an order of
magnitude higher than the C needed to prevent separation.
Consequently, environmental perturbations resulting from rotor-tip
vortices striking the ap at blade passage frequencies may lock onto
the AFC and provide sufcient amplitudes to keep the ow attached.
Such coupling was observed during early V-22 powered-model tests,
and it suggests a different procedure for the application of AFC. In
this case, dynamic coupling between ap deection and AFC
amplitude during the initiation or transition to hover may prove more
efcient. Such an arrangement requires sensors for closed-loop
control and the use of algorithms that take advantage of this
hysteresis.

VIII.

Conclusions

AFC using zero-mass-ux periodic excitation successfully

reduced the download on the wings of hovering tiltrotor aircraft. The
effectiveness of AFC was demonstrated in a research program using
scale models and a full-scale XV-15 tiltrotor aircraft. The following
are the main results of this research program:
1) In 2-D wing/ap tests, the most effective location of the blowing
slot was found to be near 10% ap chord.
2) In the 2-D tests with the 10% slot location, it was found that the
greatest reduction in vertical drag (download) was achieved at C

0:5% and that larger values did not signicantly reduce the download
further.
3) In the 2-D tests, the best nondimensional AFC frequency was
F < 1.
4) On the powered XV-15 tiltrotor aircraft model, a passive
Krueger ap was effective when AFC was applied to the trailingedge ap. This was not found to be the case in the 2-D tests.
5) On the powered XV-15 model, with AFC applied to the trailingedge ap and to the Krueger ap with the two 180 out of phase with
each other, a reduction in download of 30% was achieved compared
with the lowest value obtained without AFC.
6) Hover ight tests of the full-scale XV-15 aircraft showed that the
download could be reduced by 9% (150 lb) by increasing the ap
deection from 70 to 75 and applying AFC to maintain attached
ow on the ap.
7) Hover ight tests of the full-scale XV-15 aircraft showed that the
ow on the ap was completely separated at a ap deection of 75 .
Activation of AFC reattached the ow, and it reduced the download
at the same ap deection by more than 14% at the higher power
settings.
8) Tests on powered full-span scaled tiltrotor models can correctly
reproduce full-scale download aerodynamics.
9) LES of the 2-D ow about the XV-15 wing section, with and
without a Krueger ap, were useful in visualizing and interpreting the
experimental results.

Acknowledgments
Three universities (University of Arizona, Tel Aviv University,
and Illinois Institute of Technology) and two major aerospace
companies (The Boeing Company and Bell Helicopter Textron, Inc.)
were involved in this project, which was administered by the U.S.
Army Research Ofce under contract number DAAD19-99-C-0023.
Many people contributed to this program, which lasted over
4.5 years. The authors wish to acknowledge the assistance of J.
McMichael, R. Wlezien, and S. Walker of the Defense Advanced
Research Projects Agency, who steered this project to its successful
conclusion. There were many technical contributors to this program
whose names do not appear in the list of authors. Aerodynamic
testing and computation were made by N. Anderberg, D. Cerchie, L.
Cullen, A. Darabi, D. Greenblatt, R. Grife, A. Hassan, J. Kiedaisch,
P. Kjellgren, M. Schmalzel, and A. Stalker. Special thanks are due to
D. Hodder for his design of the hover rig and the rigorous calibration
effort. The support of the Bell Helicopter Flight Test Team is greatly
appreciated.

References
[1] McVeigh, M. A., The V-22 Tilt Rotor Large-Scale Rotor Performance/
Wing Download Test and Comparison with Theory, 11th European
Rotorcraft Forum, London, Sept. 1985.
[2] McCroskey, W., Spalart, P., Laub, G., and Maisel, M., Airloads on
Bluff Bodies, with Application to the Rotor Induced Downloads on TiltRotor Aircraft, Vertica, Vol. 9, 1985, pp. 111.
[3] Maisel, M., Laub, G., and McCroskey, W. J., Aerodynamic
Characteristics of Two-Dimensional Wings at Angle of Attack Near
90 Degrees, NASA TM 88373, 1986.
[4] Felker, F., and Light, J., Rotor/Wing Aerodynamic Interactions in
Hover, 42nd American Helicopter Society Annual Forum,
Washington, D.C., American Helicopter Soc., Alexandria, VA,
June 1986.
[5] Wood, T. L., and Peryea, M. A., Reduction of Tiltrotor Download,
49th American Helicopter Society Annual Forum, St. Louis, MO,
American Helicopter Soc., Alexandria, VA, May 1993.
[6] Seifert, A., Bachar, T., Koss, D., Shepshelovich, M., and Wygnanski, I.,
Oscillatory Blowing, a Tool to Delay Boundary Layer Separation,
AIAA Journal, Vol. 31, No. 11, 1993, p. 2052.
doi:10.2514/3.49121
[7] Nagib, H., Kiedaisch, J., Wygnanski, I., Stalker, A. D., Wood, T., and
McVeigh, M. A., First In-Flight Full-Scale Application of Active Flow
Control: The XV-15 Tiltrotor Download Reduction, NATO Specialists Meeting, Prague, NATO, 15 Sept. 2004.
[8] Kjellgren, P., Anderberg, N., and Wygnanski, Download Alleviation
by Periodic Excitation on a Typical Tiltrotor Conguration:
Computation and Experiment, AIAA Flow 2000, Denver, CO, AIAA
Paper 2000-2697, June 2000.
[9] Kjellgren, P., Cerchie, D., Cullen, L., and Wygnanski, I., Active Flow
Control on Bluff Bodies with Distinct Separation Locations, 1st AIAA
Flow Control Conference, St. Louis, MO, AIAA Paper 2002-3069,
June 2002.
[10] Kjellgren, P., Hassan, A., Sivasubramanian, J., Cullen, L., Cerchie, D.,
and Wygnanski, I., Download Alleviation for the XV-15:
Computations and Experiments for the Flow Around the Wing, AIAA
Paper 2002-6007, 2002.
[11] Stalker, A., The Effects of Krueger Flaps and Periodic Perturbations on
the Download Alleviation of a Typical Tiltrotor Aircraft, M.S. Thesis,
AME Department, Univ. of Arizona, Tucson, AZ, 2004.
[12] Nishri, B., and Wygnanski, I., Effects of Periodic Excitation on
Turbulent Flow Separation from a Flap, AIAA Journal, Vol. 36, No. 4,
1998, pp. 547556.
doi:10.2514/2.428