applications of circulation control technologies
TRANSCRIPT
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 1/637
Applications
of
Circulation
Control Technologies
Edited by
Ronald D. Joslin
Office
of
Naval Research
Arlington Virginia
Gregory S. Jones
NASA Langley Research Center
Hampton Virginia
olume
2 4
PROGRESS IN
ASTRONAUTICS AN D AERONAUTICS
Frank K. Lu Editor-in-Chief
University
of
Texas at Arlington
Arlington Texas
Published by the
American Institute
of
Aeronautics and Astronautics Inc.
1801 Alexander Bell Drive Reston Virginia 20191-4344
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 2/637
American Institute of Aeronautics and Astronautics, Inc., Reston, Virginia
1 2 3 4 5
Copyright 006 by the American Institute of Aeronautics and Astronautics, Inc. Printed in the United
States of America. All rights reserved. Reproduction or translation o f any part of this work beyo nd that per-
mitted by Sections 107 and 1 08 of the U.S. C opyright Law w ithout the permission
of
the cop yright owner is
unlawful.
The
code following this statement indicates the copy right owner’s consent that copies o f articles in
this volume may be mad e for personal or internaluse, on condition that the copier pay the per-copy fee ( 2.50)
plus the per-page fee ( 0.50) through the C opyright Clearance Center. Inc., 222 Rosew ood Drive, Danvers,
Massachusetts 01923. This consent does n ot extend to other kinds of copying, for which permission requests
should be addressed to the publisher.
Users
should employ the following code when reporting copying from
the volume to the Copy right Clearance Center:
1-56347-789-0/06 2.50
S
Data and information appearing in this book are for informational purposes only. AIAA is not responsible for
any
injury
or damage resulting from use or reliance, nor does AIA A warrant that use or reliance will be free
from privately owned rights.
ISBN
1-56347-789-0
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 3/637
Progress in Astronautics and Aeronautics
Editor in Chief
Frank
K
Lu
University
o
Texas at Arlington
Editorial Board
David A Bearden
The Aerospace Corporation
John D Binder
viaSolutions
Steven A Brandt
U S Air Force Academy
Abdollah Khodadoust
The Boeing Company
Richard C Lind
University
o
Florida
Richard
M
Lloyd
Raytheon Electronics Company
Fred R DeJamette Frank Pai
North Carolina State University
University
of
Missouri Columbia
Gail Klein
Jet Propulsion Laboratory
Ning Qin
University
of
Shefield
George Eitalbery Oleg Yakimenko
German Dutch Wind Tunnels
US
aval Postgraduate School
Sanjay Garg Ben T Zinn
NASA Glenn Research Center
Georgia Institute o Technology
Eswar Josyula Peter H Zipfel
US ir Force Research Laboratory
U.S. Air Force Research Laboratory
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 4/637
Foreword
HIS collection of papers represents a compilation of the state of the art in
irculation control technologies by two of the foremost experts in the field.
The volume is conveniently organized to enable experts and beginners alike to
quickly obtain a thorough historical overview and then be brought up to speed
on the latest research. The final chap ter delves into new areas and draw s attention
to exciting new ideas in circulation control. wide range of advanced exper
imental and numerical methods are discussed by a panel of international
experts. The text will prove to be of great value to workers in this field.
Frank K.
Lu
Editor in Chief
Progress in
Astronautics and Aeronautics
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 5/637
Preface
HE G ENESIS of this volume originated during the planning of the
NASA
T NR Circulation Control Workshop, which was held March 2004 in
Hampton, Virginia. Over two full days, 3 papers and 4 posters were presented,
with 110 scientists, engineers, and program managers in attendance. This book
was conceived to distribute this rich body of technical information on circulation
control to a broader audience and to provide historical documentation to support
future circulation control applications. Since that workshop, the papers have
been updated and peer-reviewed to arrive at a compilation of the state of the
art in circulation-control technologies.
The goals of this book are 1) to summarize the history and the state of the art
in circulation control technology, 2) to provide a single up-to-date knowledge-
base for circulation control design, analysis, and experimental testing, and
3
to highlight prediction tools for circulation control. Goals 1 and are clearly
achieved in the chapters by the diverse applications and significant breadth of
insights offered by the experts in this field. Goal
3
is most notably achieved by
the use and discussion of the diverse range of computational fluid dynamics
CFD) tools for circulation control. Results showing the successful prediction
of performance and inadequacies of some predictions are presented for
completeness.
The book is divided into four sections. The first major section presents a
historical overview of circulation control. Because the overview papers are
very thorough, many of the remaining chapters present brief introductions. The
second major section covers experiments and applications. Section I1 is
divided into A. fundamental flow physics, B. aerospace applications, and C.
nonaerospace applications. The third major section covers CFD-based prediction
tools and som e validation with experiments most of which are detailed in
Section 11). Section I11 is subdivided by the different predictive applications.
Finally, the last section consists of a single chapter, which introduces a vision
for the use of circulation control in a broad spectrum of nonvehicle applications.
Although less rigorous than most chapters, this final chap ter exposes the reader to
some new insights into applications of circulation-control technologies.
Ronald
D
Josl in
Gregory
S
Jones
December
2 5
xix
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 6/637
Table
of
Contents
Preface xix
I Overview
Chapter 1 Advantages of Combining BLC Suction with Circulation
John L Loth. West Virginia University. Morgantown. West Virginia
Control High-Lift Generation 3
Nomenclature 3
Introduction 4
Designing a CC Technology Demonstrator STOL Aircraft 5
1974 Flight Testing
of
the WV U CC Technology Demonstrator 12
1979 CC Flight Tests with a G rumman Aerospace A-6A 16
Conclusions 18
References
2
Chapter 2 Overview of Circulation Control Pneumatic
Aerodynamics: Blown Force and Moment Augmentation and
Modification as Applied Primarily to Fixed-Wing Aircraft
23
Robert J Englar.
Georgia Institute of Technology. Atlanta. Georgia
Nomenclature 23
Introduction 24
Coanda E ffect 25
Applications of Circulation Control. Past and Present 28
Powered Lift and Engine Thrust Deflection 48
Other Aircraft Applications 53
Nonflying Applications of Circulation Control
57
Conclusions 63
References 64
Chapter 3 Exploratory Investigations of Circulation Control
Robin Imber.
Naval Air Systems Comm and. Patuxent River. Maryland;
Technology: Overview for Period 1987-2003 at NSW CCD
Ernest Rogers and Jane Abramson.
Naval Surface Warfare
Center-Carderock Division. West Bethesda. Maryland
69
Nomenclature 69
Introduction
7
ix
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 7/637
Dual-Slotted Cambered Airfoil LSB) 70
Self-Driven Rotary Thruster TIPJET) 73
Annular Wing CC-Duct) 79
Circular Wing CC-Disc) 85
Miniature Oscillatory Valve CC-Valve) for Unsteady W ing Load Reduction 91
Dual-Slotted Low Aspect Ratio Wing CC Hydrofoil) 93
Status of Design Capability 99
Conclusions 100
References 101
1I.A. Experiments and Applications: Fundamental Flow Physics
Chapter
4
Measurement and Analysis of Circulation
F. Kevin Owen.
Complere Inc. Paczjic Grove. California;
Andrew K Owen.
Control A irfoils 105
University o Oxford. Oxford. England. United Kingdom
Nomenclature 105
Introduction 106
Experimental Details 107
Sample Results 107
Conclusions 112
References 112
Chapter 5 Some Circulation and Separation Control Experiments 113
Dino Cerchie. Eran Halfon. Andreas Hammerich. Gengxin Han. Lutz Taubert.
Lucie.Trouve. Priyank Varghese. and Israel Wygnanski. University of Arizona.
Tucson Arizona
Nomenclature 113
Introduction 114
Discussion of Results 118
Conclusions 162
Acknowledgments 164
References 164
Chapter 6 Noise Reduction Through Circulation Control 167
Scott E Munro. Krishan K Ahuja. and Robert
J
Englar.
Georgia Institute of Technology. Atlanta. Georgia
Nomenclature 167
Introduction 168
Background 169
Facilities and Instrumentation 171
Technical Approach 173
Results and Discussion 174
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 8/637
xi
Conclusions
184
Acknowledgments 186
References 186
1I.B. Experiments and Applications: Aerospace
Chapter 7 Pneumatic Flap Performance for a Two-Dimensional
Circulation Control Airfoil 191
Gregory S Jones.
NASA Langley Research Center Hampton Virginia
Nomenclature
Introduction
NASA CC Requirements
Theoretical Considerations
GACC Airfoil Design
Experimental Setup
Airfoil Performance
Conclusions
Appendix
References
191
192
193
195
202
207
216
236
237
241
Chapter 8 Trailing Edge Circulation Control of an Airfoil at
Michael G Alexander. Scott G Anders. and Stuart
K
Johnson.
NASA Lungley
Transonic Mach Numbers 245
Research Center Hampton Virginia
Nomenclature 245
Introduction
246
Instrumentation 251
Facli y 252
Test Procedures and Conditions
253
Test Conditions 254
Discussion of Results 254
Conclusions 263
Acknowledgments
275
References
275
Model Description 247
Chapter 9 Experimental and Computational Investigation into the
Gerald Angle 11. Brian O ’Hara. Wade Huebsch. and James Smith.
West Virginia
Use of the Coanda Effect on the Bell A821201 Airfoil 277
University Morgantown West Virginia
Nomenclature 277
Introduction 278
Experimental Apparatus and Procedure
279
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 9/637
xii
Computational Model and Procedure 282
Experimental Results 285
Computational Results 286
Conclusions 290
References 291
Chapter 10 Novel Flow Control Method for Airfoil
Ge-Cheng Zha and Craig D Paxton. University of Miami. Coral Gables. Florida
Performance Enhancement Using Co-Flow Jet 293
Nomenclature 293
Introduction 294
Results and Discussion 296
Conclusions 311
Acknowledgments 312
References 312
Chapter 11 Experimental Development and Evaluation of
Robert J Englar.
Georgia Institute of Technology. Atlanta. Georgia;
Pneumatic Powered-Lift Super-STOL Aircraft 315
Bryan A Campbell. NASA
Langley Research Center. Hampton. Virginia
Nomenclature 315
Introduction 316
Experimental Apparatus and Test Techniques 320
Wind-Tunnel Evaluations and Results 321
Comparison of Measurements and Predictions 331
Potential Applications 333
Conclusions
333
Acknowledgments 335
References 335
Chapter 12 Use of Circulation Control for Flight Control 337
Steven l Frith and Norman
J
Wood, University of Manchester. M anchester.
England. United Kingdom
Nomenclature 337
Introduction 338
Half-Span Cropped-Delta Model 339
Full-Span UAV Configuration 345
Conclusions 352
Acknowledgments 353
References 353
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 10/637
xiii
1I.C. Experiments and Applications: Nonaerospace
Chapter 13 Pneumatic Aerodynamic Technology to Improve
Robert J Englar. Georgia Institute
of
Technology. Atlanta. Georgia
Performance and Control of Autom otive Vehicles 357
Nomenclature 357
Introduction 357
Basics of Pneumatic Circulation Control Aerodynamics 358
DOE Pneumatic Heavy Vehicle Model Test Results 360
Pneumatic HV Fuel Economy Testing 367
Updated Wind Tunnel Evaluations 371
Pneum atic Sport Utility Vehicles PSUVs) 374
Conclusions 379
Recommendations 380
Acknowledgments 381
References 381
Chapter 14 Aerodynamic Heat Exchanger: A Novel Approach
Richard J Gaeta. Robert J Englar. and Graham Blaylock.
to Radiator Design Using Circulation Control 383
Georgia Institute of Technology. Atlanta. Georgia
Nomenclature 383
Introduction 383
Technical Approach 386
Results 389
Conclusions
395
Acknowledgments
397
References 397
1II.A. Tools for Predicting Circulation Control Performance:
NCCR 1510 Airfoil Test Case
Chapter 15 Investigation of Turbulent Coanda Wall Jets Using
D N S a n d R A N S 401
Hermann
F
Fasel. Andreas Gross. and Stefan W e n . University
of
Arizona.
Tucson. Arizona
Nomenclature 401
Introduction
4 2
Investigated Configurations 403
Numerical Approach 404
Turbulent Wall Jet on a Circular Cylinder
4 5
Circulation Control Airfoil 415
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 11/637
xiv
Conclusions 418
Acknowledgments 419
References 419
Chapter 16 RANS and Detached-Eddy Simulation of the
Eric G Paterson and Warren J Baker. Pennsylvania State University.
NCCR Airfoil 421
University Park. Pennsylvania
Nomenclature 421
Introduction 422
Geometry. Conditions. and Data 424
Computational Methods 425
Grid Generation 427
Initial and Boundary Conditions 429
Results 430
Conclusions 441
Acknowledgments 442
References 442
Chapter 17 Full Reynolds-Stress Modeling of Circulation
Control A irfoils 445
Peter A Chang 111. Joseph Slom ski. Thomas Marino. Michael P Ebert.
and Jane Abramson. Naval Surface Warfare Center-Carderock
Division. West Bethesda. Maryland
Nomenclature 445
Introduction 446
Mathematical Development 448
Results 453
Conclusions 465
Acknowledgments 465
References 465
1II.B. Tools for Predicting Circulation Control Performance:
NCCR
103RE
Airfoil Test Case
Chapter 18 Aspects of Num erical Simulation of Circulation
R Charles Swanson. Christopher L Rumsey. and Scott G Anders.
Control A irfoils 469
NASA Langley Research Center. Hampton. Virginia
Nomenclature 469
Introduction 470
GeometryandGrid 472
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 12/637
xv
Numerical Method 475
Boundary and Initial Conditions 476
Turbulence Modeling 476
Jet Momentum Coefficient 478
Numerical Results 478
Conclusions 495
Acknowledgments 497
Appendix: Coordinates of 103RE Airfoil 497
References 497
Chapter
19
Gregory McGowan. Ashok Gopalarathnam. Xudong Xiao. and Hassan Hassan.
Role of Turbulence Modeling in
Flow
Prediction of
Circulation Control Airfoils 499
North Carolina State University. Raleigh. North Carolina
Nomenclature 499
Introduction
5
Formulation
of
the Problem 501
Results and Discussion 502
Conclusions
51
Acknowledgments 510
References
51
1II.C. Tools for Predicting Circulation Control Performance:
GACC Airfoil Test Case
Chapter 20
Warren J Baker and Eric G Paterson. Pennsylvania State University.
Simulation of Steady Circulation Control for the
General Aviation Circulation Control GACC) Wing 513
University Park. Pennsylvania
Nomenclature 513
Introduction
514
Geometry. Conditions. and Data 515
Computational Methods 516
Grid Generation 518
Initial and Boundary Conditions 521
Computational Resources 523
Results 523
Conclusions 536
Acknowledgments 537
References 537
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 13/637
xvi
Chapter
21
Gregory McGowan and Ashok Gopalarathnam.
North Carolina State University.
Com putational Study of a Circulation Control
Airfoil Using FLUENT 539
Raleigh. North Carolina
Nomenclature 539
Introduction 540
Configurations and Experiments 541
Numerical Approach 542
Results 545
Conclusions 552
Acknowledgments 553
References 553
1II.D. Tools for Predicting Circulation Control Performance:
Additional CFD Applications
Chapter
22
Yi Liu. Lakshmi N Sankar. Robert J Englar. Krishan K Ahuja. and
Richard Gaeta. Georgia Institute
of
Technology. Atlanta. Georgia
Computational Evaluation of Steady and Pulsed
Jet Effects on a C irculation Control Airfoil 557
Nomenclature 557
Introduction 558
Mathematical and Numerical Formulation 559
Results and Discussion 561
Conclusions 575
Acknowledgment 575
References 575
Chapter 23
Jubaraj Sahu. U S Army Research Laboratory. Aberdeen Proving Ground.
Time-Accurate Simulations of Synthetic
Jet-Based Flow Control for a Spinning Projectile
579
Maryland
Nomenclature 579
Introduction 580
Computational Methodology 581
Projectile Geometry and Computational Grid 584
Results 586
Conclusions 594
References 595
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 14/637
xvii
IV Exploring a Visionary Use of Circulation Control
Chapter 24 Coanda Effect and Circulation Control for
Terence R Day.
Vortex Dynamics Pty Ltd Mount Tamborine
Nonaeronautical Applications 599
Queensland Australia
Introduction 99
Applications 600
Conclusions 612
Acknowledgments 612
References 612
Index 615
Author Index 623
Supporting M aterials 625
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 15/637
I. Overview
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 16/637
Chapter 1
Advantages of Combining BLC Suction with
Circulation Control High-Lift Generation
John L. Loth*
W est Virginia University, Morgantown, West Virginia
Nomenclature
CB
=
circulation control blowing efficiency factor
CL
=
lift coefficient
C = blowing coefficient
Di
=
induced drag
D,,
=
parasite drag
CLopt
= optimum lift coefficient where aircraft
L / D
is maximum
mcc
= circulation con trol blowing m ass flow rate
p t
=
total pressure in the compressor bleed air supply duct
q ,
=
dynamic pressure
S
=
wing area
t,, = non-dimensional circulation control blowing slot height
tn
=
non-dimensional ejector nozzle slot height
ts
=
non-dimensional suction slot height
V,,
=
circulation control blowing velocity
V ,
=
equivalent airspeed, corrected for position error
Vi
=
indicated airspeed
V , = free stream velocity
p =
air density
r
=
circulation con trol rounded trailing edge radius
Subscripts
a =
angle of attack
c
= chord
co = free stream conditions
*F’rofessor, Mechanical and Aerospace Engineering. Associate Fellow AIAA.
Copyright 005 by the American Institute of Aeronautics and Astronautics, Inc. All rights
reserved.
3
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 17/637
4
J.
L.
LOTH
I Introduction
HE PU RPOSE of this paper is to present the advantages of combining bound-
T
ry layer control (BLC) by suction with circulation control (CC) by blow ing
for aircraft high lift generation. In the short take-off and landing (STOL) mode,
the sharp trailing edge of the wing must be converted into a rounded Coanda
surface for CC blowing. Jet engine hot, high-pressure compressor bleed air is
the most commonly used source for the blowing air. Ducting this hot, high-
pressure air to the CC blowing slot involves problems arising from factors
such a s duct size, weight, pressure loss, required insulation and thermal expan-
sion joints, and jet engine take-off thrust loss. It is shown here how adding an
ejector for BL C suction just upstream of the CC blow ing slot can diminish the
impact of the aforementioned problems. It can reduce the amount of compressor
bleed air required, and thus duct size, by more than 50%, provide structural
cooling, and improve the CC blowing to free stream velocity ratio, bringing it
closer to four, where the theoretical lift augm entation ratio reaches a maximum.
Flight test results using such a configuration are provided, together with sol-
utions for in-flight transition from the C C rounded trailing edge to a sharp trailing
edge for low-drag cruise. Data were collected in 1974 during flight testing of the
first CC Technology D emonstrator Aircraft, at West Virginia University.
The use of blowing air to augm ent airfoil lift had already been proposed',2 in
the
1920s.
D a ~ i d s o n , ~n his
1960
British patent application, referred to the
concept of blowing over a circular cylinder as circulation control (CC). To
improve the lift-to-drag ratio, Kind and Maull? at Cambridge University, experi-
mented with C C on elliptical airfoils. Kind is also credited with developing the
first boundary layer theory for CC blowing to correlate his experimental
results. At zero angle of attack, elliptic airfoils produce two nearly identical
suction peaks at their leading and trailing edges; this results in an aft shift of
the center of pressure and thus nose-down pitching moment. Typical streamlines
for such an air foil, computed by S h r e ~ s b u r y , ~re shown in Fig.
1.
A schematic of a CC blowing slot is shown in Fig.
2.
Blowing air must be
supplied uniformly to the blowing slot. By Coanda turning, the jet generates a
high suction force on the rounded surface. The angular position of the lower
surface stagnation point, where the Coanda jet separates when meeting the
flow from below the airfoil, determines the circulation and lift produced. Even
today, most disagreements between computational and experimental results are
Fig.
1
Com puted stream lines for an e lliptic airfoil?
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 18/637
COMBINING BLC SUCTION AND CC HIGH-LIFT GENERATION 5
High -pressure fluid flowing from
Fig. 2 Schematic of a CC trailing edge.
a result of the sensitive relationship between circulation and location of this lower
surface stagnation point.
In the late 1960s, Robert W illiams,6 then at NSRD C (Naval Ship Research and
Development Center), started to expe riment with CC airfoils developed by Kind.
Williams697 nvestigated the feasibility of a heavy-lift helicopter with dual
plenum elliptic rotor blades and valves to control CC blowing rate to allow
high forward speed. In 1968, the Office of Naval Research (ONR) contracted,
with West Virginia University (WVU), research on CC airfoils, including
testing at high Reynolds number and away from wind tunnel wall interference.
Loth and Fanucci considered the possibility of protruding a CC blown airfoil
from the roof of one of the WVU flight test aircraft. However, this would not
be safe, because the roll moment produced by such a CC airfoil would exceed
the available aircraft aileron control. To satisfy the contract requirement of
flight testing CC technology, they decided it would be safer to fly a fixed-wing
aircraft with CC blown wings. In the case of blower failure, an elliptic airfoil
would not be flyable; therefore, new CC wings were designed at WVU, which
were in-flight convertible from a high-speed, low-drag, conventional sharp trail-
ing edge to a rounded trailing edge with CC blowing during slow-speed flight
testing. In the following five years, several such CC airfoils were tested at
WVU in the wind tunnel there. A
comparison of blowing air requirements and
lift capability for various high-lift systems was com pleted in 1973, as shown in
Fig.
3.
This indicates that CC high-lift generation is more conservative in
blowing air requirement than other methods.
11. Designing a CC Technology Demonstrator STOL Aircraft
A Bede-4 homebuilt kit was found to provide an economical and suitable
frame for test flying a CC wing. The simplest arrangement for in-flight conver-
sion from a sharp trailing edge to a rounded CC blown trailing edge was first
investigated. This is a forward folding flap, which exposes its semicircular
hinge to provide a rounded trailing edge for CC blowing, as shown in Fig. 4.
Dr. Norio Inumaru, visiting WVU from NRL in Tokyo, Japan, designed its
drooped leading edge to prevent leading edge stall at high lift. The test model
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 19/637
6
J.
L.
LOTH
Fig.
3
Performance comparison between powered h igh-lift systems.
was fabricated by riveting a sheet metal cuff to the wing leading edge and filling
its cavity with foam. In 1970, Model A wing was tested in CC mode in the two-
dimensional 8 x 10ft NSRDC wind tunnel, both in the sharp and round trailing
edge configurations (Fig. 5 ) . The test data for
CL,
a, and
C ,
are shown in F ig.
6 .
Below stall, in the angle of attack range - 2
<
a
<
8 deg, they could be
L.E.
DROOP DESIGN
Fig.
4
WVU Model
A CC
wing, wind tunnel tested at NSRDC in
1970.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 20/637
COMBINING BLC SUCTION AND CC HIGH-LIFT GENERATION
7
Fig.
5
WVU M odel
A
wing: left, in cruise; right, in
CC
mode.
curve-fitted by a linear equation:
In CC mode, the Model A chord length was reduced to 88% relative to cruise
mode. All test data show n are referenced to cruise chord length, which effectively
lowers CLm,, for the Model A in CC mode. The sharp trailing edge wing had a
two-dimensional value of C
= 0.09.
Thus, for curve fitting test data in CC
mode, C was reduced to 0.09 x
88% =
0.08. Excellent curve fitting was
obtained by replacing SCL/SC, with cB/ , where CB is constant and
named the blowing efficiency factor. Model A wing test data with C B=
4.3
pro-
vided the best curve fit, as shown in Fig.
7
using:
The disappointing performance of the Model A wing prompted a new design called
the Model B wing. Instead of folding the flap inward for
CC
high lift generation, its
flap was folded out. The 20% longer chord length, as shown in Figs.
8
and 9, was
expected to increase the
CC
blowing efficiency factor CB from
4.3
to
4.3
x
(120 /
88 )
=
5.9.
In the Model
B
wing, great care was taken to achieve blowing
0.4 0.6 0.8 1.0
0.2
I
CP
Fig.
6
WVU Model
A CC
wing,
1970
wind tunnel test results.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 21/637
a
J.
L.
LOTH
5
4
1
0
0 0.2
0.4
0.6 0.8
CU
Fig. 7 WVU Model
A
CC wing em pirical curve fit with CB = 4.3.
slot uniformity. This was accomplished by machining and bolting segmented
aluminum nozzle blocks to an aluminum 3-in.-diam air supply duct, which
also served as the rounded CC trailing edge. This provided a uniform 0.012
in.-wide primary blowing slot (Fig.
9).
The WVU wind tunnel model tests on a
two-dimensional version of the Model B wing are described in Refs. 9 and 10.
When applied to the CC Technology Demonstrator aircraft, the source of
blowing air had to be selected. Boasson, in his dissertation, proved theoretically
that the lift augmentation ratio CB eaches a maximum when Vcc/Vm = 4.'' Such
low CC blowing velocity requires a high mass flow rate. Then Ap, the duct
friction loss inside the air supplying 3-in.-diameter CC rounded trailing edge,
L.E.
DROOP
DESIGN
Fig.
8
WVU Model B CC wing, wind tunnel tested at WVU and flown on the CC
Technology D emonstrator.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 22/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 23/637
10
J.
L.
LOTH
Inserting values for slot height gives
3 1
-vs
+-v,
=
1
4 4 or
v,
= 4 - 3 v s
Momentum equation:
tsv,2 t ,v,2 T=
s 1
PVCC
(4)
Inserting values for slot height gives
3 1
4 4
V,2+-V,2-0.5V,2=
1
Substituting
V,
from above gives the quadratic equation:
5V;
-
12Vs
+ 6 = 0
Solving for
Vs < V,
gives
Vs =
0.7 when inserted in the preceding relation,
which gives
V,
=
1.9. This means that
t,V,
=
a
x
1.9
=
47.5%; in other
words, the nozzle needs to supply only 47.5% of the CC blowing air. The
balance of the blowing air tsVs = x 1.9
=
52.5% is supplied by the BLC
suction slot and need not be supplied through the CC rounded trailing edge
duct. The C C jet exits at near ambient pressure with thrust
Tcc=
hccVCc.The
required ejector nozzle thrust T, is only 0.83Tcc,
This demonstrates that incorporating an ejector can 1) provide cooling
by boundary layer suction, 2) increase CC blowing momentum by (1-0.83)
or 17%, 3) lower the velocity ratio
Vcc/VW
to increase blowing efficiency
factor
CB.
Furthermore,
it
reduces compressor bleed air mass flow rate required
by 52.5% which lowered duct pressure loss with associated duct size and w eight
savings. In the WVU wind tunnel model tests, the availability of flap hinge
suction also allowed the CC flap to be deflected up to 15 deg without stall for
additional lift augmentation. Arrows in Fig. 11 highlight the special features of
this aircraft.
Arrows have been used to show the CC blowing slot on the top of the 3-in.-
diameter rounded trailing edge. Suction boundary layer control (BLC) is
shown at the flap hinge. The pilot can dump the blowing air overboard by
actuating an air dump valve to achieve Direct Lift Control, called (DLC) as
indicated.
A layout of the WVU CC Technology Demonstrator aircraft is shown in Fig.
10 with a GTC 85-72 gas turbine mounted in the rear passenger seat area. Note
that the jet engine exhaust discharges upwards, to prevent igniting the blacktop
on the parking area.
In Fig. 12 are shown details of the flap retraction and extension mechanism
by a two horsepower electric motor. It turns the CC air supply duct inside the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 24/637
COMBINING BLC SUCTION AND CC HIGH-LIFT GENERATION
11
EMPTY WT- 1720 I@
W
BROSS
2400 b
Fig. 10 WVU CC T echnology Demonstrator dimensions and layout.
cabin, which is welded by bell cranks to the two 3-in.-diameter CC rounded
trailing edges. For increased roll control at low speed, the ailerons are drooped
and blown with compressor bleed air supplied via small ejectors inside the
cabin for coo ling purposes. T o increase aileron effectiveness, they are connected
to a flow diverter valve, which alters the left and right wing blowing rate. The
bolt shown in the air splitter tee serves as a hinge for the splitter valve inside
this tee.
Fences and structure used to strengthen the
cavity at the
bottom of
the wing , into which the
CC rounded trailing edge retracts.
BLC at flap
hinge line
Fig.
11
WVU CC Technology Demonstrator location of CC, BLC, DLC, and space
for flap folding.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 25/637
12
J.
L.
LOTH
Fig. 12 Compressor bleed air enters into a worm -gear driven pipe, connected by
bell cranks to the left and right CC rounded trailing edges.
111.
Prior to
25
h of flight testing, which started on
10
April 1974, the
CC
slot was
tested for blowing uniformity and its ejector for providing adequa te cooling to the
fiberglass wings. The flight tests, performed by test pilot Shawn Roberts, started
with calibrating airspeed and position error, with the use
of
a Pitot tube mounted
with a boom to the left wing tip (Fig.
13).
This boom also contained angle of
1974 Flight Testing of the WVU CC Technology D emonstrator
Fig. 13 Large position error in cockpit speed indicator against equivalent airspeed
based on boom-mounted pitot tube readings.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 26/637
COMBINING BLC SUCTION AND CC HIGH-LIFT GENERATION
13
attack and yaw measuring instrumentation. The aircraft in flight is shown in
Fig.
14,
with the CC blowing flap deployed. A summary of the flight test data
is shown in Fig.
15.
Shown are three scales for the lift coefficient, all based on
dynamic pressure q , calculated using equivalent airspeed and reduced to sea-
level density. The left column indicates the trimmed-out aircraft CL. The
middle column has the tail download added to the lift and is termed CL
wing
On the far right column is the maximum CL value, which occurred at the
flap centerline. For example, the average wing lift coefficient increased from
2.0
without blowing, at
C =
0, to
4.3
with blowing at
C
=
0.12.
Near stall
the difference in angle of attack was negligible, thus the blowing efficiency
factor
C
from Eq.
(2)
can be solved from:
C =
ACLI
C
=
(4.3
-
2 ) / m
=
6.6
G
5 )
This is more than
10
better than could be expected by extrapolating the Model
A
test results for the increased chord length, or C =
4.3
x
(1.2/0.88)
=
5.86.
This improvement can be credited to the utilization of an ejector in the Model
B wing. It is interesting to calculate the CC blowing air horsepower required if
the blowing air were supplied at standard sea-level conditions. The blowing
slot height of
0.050
in. along the two 100-in.-long CC flaps resulted in blowing
slot area
A,, =
10
in.’. Consider flight with the propeller at idle, with
C
=
0.12
and CL
=
3.8.
From the definition of C = T,,/ q,S,) and CL=
L/ q,S,), calculate blowing momentum
Tcc=
(0.1213.8)
x W =
2400
lb) =
76
lbf
(6)
Fig. 14 WVU CC Technology Demonstrator during 25 h of flight testing, starting
10 April
1974.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 27/637
14
J. L.
LOTH
r
0
f
U
t
5.6 ’
t.3
2 1
-
20 .
w
3
t
0
1.7
12
L O
.I
OM
-
I I
-
i NOTS
5 3 Sp 473
Fig.
15 WVU CC
Technology Demonstrator flight performance map with CC
blowing efficiency factor
C,
=
6.6.
At sea level density,
CC
velocity would be
0.002377
76
10/144)
)’”=
78
f t /s
7)
and mass flow rate w ould be
rit
= pAccVcc= 0.002377 x 10/144) x 678 = 0.1 12slug/s
8)
Then the blowing power kinetic energy required is equal to
(9)
V2
2
i z s
= 0.112
x
0.5
x
6782= 25742
ft.lbf/s
= 46.8
hp
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 28/637
COMBINING BLC SUCTION AND CC HIGH-LIFT GENERATION 15
To minimize blowing power required," the CC blowing velocity
V,,
should
equal 4 times the flight velocity
V
At a flight speed of 39
kn, V,,
should then
be: 4
x
39
=
156
kn
=
260 ft/s. For
T,, =
76 lbf, the CC blowing mass flow
rate should be 76/260
=
0.288 slug/s or kinetic power required would be as
low as
0.5
x
0.288
x
2602
=
9734 ft.lbf/s
=
17.7 hp. This reduction in C C
blowing power required demonstrates the advantage of optimizing the ratio
The pilot was quite satisfied with the handling qualities, and surprised how
well the direct lift control DLC valve worked to make correction on the glide
angle on approach to landing without inducing significant attitude changes.
The CC flap deployment and stowing process worked well and required less
than a 17 lb change in stick force, as shown in Fig. 16. To significantly
reduce the stick force during flap deployment or retraction, Loth'* filed
U.S.
Patent 4,600,172, which allows converting a Fowler flap into a CC rounded
trailing edge flap by only folding out a rounded trailing edge, which also
supplies the blowing air (Fig. 17). The BLC suction is sufficiently strong to
hold the Fow ler flap against the CC pipe w ithout the need for mechanical attach-
ments. The ability to stow away the CC rounded trailing edge for high-speed
low-drag cruise is an important aspect for operation with circulation control
high lift systems.
Slow flight was the most challenging aspect of the flight test program. With
the propeller at 135 hp, the aircraft could be slowed to 23.5
kn
indicated,
which corresponds to 33.2
kn
calibrated airspeed. This corresponded to a
trim lift coefficient of 5.1 and wing average lift coefficient of 5.6 while
blowing at 13 psig. Then there is little or no power to spare to prevent the
onset of stall, which always started with a rapid roll and up to 1000 ft of alti-
tude loss. Clearly this represents flying on the backside of the power curve, as
V C C / V ~ .
FLAP FOLDING ANGLE /3* w,T:\yE
LAP
ULL
OUT RETRACTED
Fig. 16
WVU CC
Technology Demonstrator shows acceptable trim force during
flap folding.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 29/637
16
J.
L.
LOTH
Fig. 17
U.S.
Patent 4,600,172 allows flap stowing with greatly reduced actuator
torque. *
shown in Fig. 18. Note it takes only half as much power to cruise twice as fast
at 70 kn.
IV. 1979
CC Flight Tests with a Grumman Aerospace A-6A
The WVU successful demonstration of CC flight on a fixed-wing aircraft
motivated the Navy to contract with Grumman Aerospace to convert an A-6A
bomber to STOL operation with CC blowing. The challenge of converting an
existing large aircraft to operate with CC blowing far exceeded that of building
the small WVU CC Technology Demonstrator from scratch. Bob Englar, at
NSRDC, began, in 1974, a careful CC wind tunnel test program to cover the
Fig. 18 WVU CC Technology Demonstrator performance safety is limited by the
effect
of
high induced drag on pow er required.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 30/637
COMBINING BLC SUCTION AND CC HIGH-LIFT GENERATION
17
entire range of operational aspects for the A-6A. His well-publicized test results
are currently considered to be the most reliable available data and to which com-
putational solutions are being compared. 13- l7 Of particular interest is Englar’s16
“STOL Potential of the Circulation Control Wing for High-Performance Air-
craft.” This contains the performance map of a wind tunnel study of the A-6A
wing without a tail, as shown in Fig. 19. When linearized using Eq. ( l ) , the
best fit constants are CL = 0.09 (per degree) and
CB
= 6.3, as shown in
Fig. 20. These results are similar to those found for the WVU CC Technology
Demonstrator, although the A-6A has a greater percent of wing area equipped
with CC blowing. The drawback of modifying an existing large aircraft is that
the CC blowing system had to be an add-on-feature with no possibility for
rounded trailing edge retraction to maintain its low-drag, high-speed cruise
capability. The magnitude of the CC rounded trailing edge is clearly visible in
a close-up photo; see Fig. 21 and in-flight Fig. 22.
The STOL performance data for the A-6A were close to those predicted
from wind tunnel tests, resulting in 1) 140 increase in usable
CL;
) 30-35%
reduction in take-off and approach speeds; 3) 60-65% reduction in take-off
and landing ground roll; and 4) 75% increase in payload.
1.4L - . .
4 o 4 a I Z 1 6 20 2 4 zs
a n degrees
Fig. 19 Wind tunnel test data for wing of Grumman A-6A.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 31/637
18
4.4
3.6
2.8
>
(CI
w
C
-I
5
2
1.2
0.4
0.4
J. L.
LOTH
0
,, G=,2
C =
a *
I
,
,
,,
f
,
,
,/'. - '
,r
,
,
,
8
,/ ,
'
,
,* /I
,
,/
,
2
,4 8
,
'
/
,
f
,i
IS
,
/
,
I
*'
,
-r
12 20 2E
a
in
degrees
-4 4
Fig. 20 Empirical curve fit to A-6A wind tunnel data shows blowing efficiency factor
Ce
=
6.3.
V. Conclusions
In 1974, the WV U C C Technology Demonstrator STO L was the first aircraft
to demonstrate the high-lift capability of CC . Its wings incorporated an in-flight-
retractable CC rounded trailing edge to enable high-lift generation by CC
blowing on an extended wing, and in-flight conversion to a reduced wing area
with sharp trailing edge for low drag, high speed cruise. The use of a retractable
CC rounded trailing edge required supplying the hot high-pressure C C blowing
air through the rounded trailing edge. To minimize air pressure loss by friction,
the duct cross-sectional area had to be at least twice that of its choked-flow area
A*. To achieve that, an ejector was inserted inside the CC blowing plenum. The
area of its choked flow nozzle was five times smaller than the flow area in its
rounded trailing edge. A dding such an ejector provided several other advantages:
1) Its entrainment provided boundary layer suction just upstream of the CC
blowing slot, which increased the blowing efficiency factor CB.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 32/637
COMBINING BLC SUCTION AND CC HIGH-LIFT GENERATION
19
Fig. 21 Grumman A-6A CC conversion by an add-on fixed
CC
blowing duct
improving
STOL
performance at the expense of cruise speed.
2 The entrained flow rate about equaled the nozzle flow rate, thereby
doubling the CC blowing mass flow rate. This lowered the Vcc/Vm velocity
ratio to increase the blowing efficiency factor CB. It also increased the CC
blowing slot height by a factor of four.
3) Ejector entrainment provided wing structure coo ling and reduced thermal
expansion problems.
Fig. 22 Grumman A-6A during CC flight tests in 1979.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 33/637
20
J.
L.
LOTH
4) Ejector entrainment reduced the amount of compressor bleed air required
with its associated take-off thrust loss.’o,’’9’8
5 During take-off in a jet aircraft, thrust loss associated with compressor
bleed, is at least two and one-half times the bleed air momentum
6) In 1979 the second CC Technology Demonstrator STOL aircraft, a con-
verted Grumman A-6, demonstrated excellent STOL performance in terms of
increased lift-off weight capability and reduction in required runway length.
7) More research is needed to reduce induced drag associated with flying at
high-lift coefficients. Not having to fly on the backside of the power curve
would greatly increase safety in STOL flight.
Since 1974, numerous other applications for
CC blowing over a rounded trail-
ing edge have demonstrated the versatility of this technology.
For example:
1)
Wake drag reduction behind cars, trucks, torpedoes, etc.
2 ) Propeller downwash drag reduction on tilt rotors.
3) Improved performance of low drag horizontally m ounted radiators in cars.
4) Lightweight, hot exhaust gas deflectors on helicopter engines in ground
5 ) Providing an alternative to a helicopter tail rotor, to cancel rotor torque.
6) Noise reduction by wake dissipation on helicopter rotors.
7) Improved effectiveness and control with upper surface blowing (USB).
8)
Providing pneumatic control on fixed flight control surfaces.
These developm ents indicate that the future for new CC applications is bright.
effect.
References
‘“Wings with Nozzle Shaped Slots,” NACA Translation TM 521, July 1929 (from
Berichte D er Aerodynamischen Veruchsenstalt in W ien, Vol.
1, No. 1, 1928).
*“The Use of Slots for Increasing the Lift of Airplane Wings,” NACA Translation PW
635, Aug. 1931 Proc eeding s L’Aeronautique, June 1931).
3Davidson,
I.
M., “Aerofoil Boundary Layer Control System,” British Patent
No. 913,754, 1960.
4Kind, R. J., and Maull, D. J.,
“An
Experimental Investigation of a Low-Speed
Circulation Controlled Airfoil,”
The Aeronautical Quarterly,
Vol.
XIX, May 10, 1968,
’Shrewsbury, G., “Numerical Evaluation of Circulation Control Airfoil Performance
Using Navier Stokes Methods,” AIAA Paper 86-0286, Jan. 1986.
6Williams, R. M., “Some Research on Rotor Circulation Control,” Proceedings of the
Third CALIAVLABS Symposium, Vol. 11, June 1969.
’Williams, R. M., and Howe, H. J., “Two-Dimensional Subsonic Wind Tunnel Tests on
a 20% Thick, 5% Cambered Circulation Control Airfoil,” NSR DC T N AL-176, 1070, AD
877764.
‘Loth, J. L., “Some Aspects of STOL Aircraft Aerodynamics,” Business Aircraft
Meeting, Wichita, KS, 3-6 April 1973, Paper No. 730328.
pp. 170-182.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 34/637
COMBINING BLC SUCTION AND CC HIGH-LIFT GENERATION
21
’Loth, J. L., Fanucci, J. D., and Roberts, S. C., “Flight Performance of a Circulation
Control STOL Aircraft,” AIAA Paper 74-994, 6th Aircraft Design, Flight Test and Oper-
ations Meeting, Los Angeles, CA, Aug. 1974.
“Loth, J. L., Fanucci, J. D., and Roberts,
S.
C., “Flight Performance of a Circulation
Control S TO L Aircraft,” Journal of Aircraft, Vol. 13, No. 3, 1976, pp. 169-173.
“Loth, J. L., and Boasson, M., “Circulation Control STOL Optimization,” Journal of
Aircraft, Vol. 21, No. 2, 1984, pp. 128-134.
‘’Loth, J. L., “Retractable Rounded Trailing Edge for Circulation Control Wing,” U.S.
Patent No. 4,600,172, issued 15 July, 1986.
Englar, R. J., “Investigation into and Application of the High Velocity Circulation
Control Wall Jet for High Lift and Drag Generation on STOL Aircraft,” AIAA Paper
74-502, 17-19 June 1974.
Englar, R. J., “Circulation Control for High Lift and Drag Generation on STOL
Aircraft,”
AIAA Journal
of
Aircraft,
Vol. 12, No.
5 ,
1975, pp. 457-463.
”Englar, R. J., Trobaugh , L. A., and Hemmerly, R. A., “Development of the Circulation
Control Wing to Provide ST OL Potential for High Performance Aircraft,” AIAA Paper 77-
578,6-8 June 1977.
16Englar,R. J., Trobaugh, L. A., and Hemmerly, R. A., “STOL Potential of the Circula-
tion Control Wing for High-Performance Aircraft,”
Journal of Aircraft,
Vol. 15, No. 3,
”Englar, R. J., Hemmerly, R. A., Moore, H., Seredinsky, V., Valckenaere, W. G. nd
Jackson, J. A., “Design of the Circulation Control Wing STOL Demonstrator Aircraft,”
AIAA Paper 79-1842, Aug. 1979; also published in Journal of Aircraft, Jan. 1981.
18Loth, J. L., “Circulation Control STO L Aircraft Design Aspects,” NASA Circulation
Control Workshop, 19-21 Feb. 1986, NASA A mes Research Center, NASA Pub. CP-
”Loth, J. L., Funk, M.
S.,
“Thrust Savings Limitations with B lown High Lift W ings,”
AIAAIAHSIASEE Aircraft Design, Systems and Operations Meeting, St. Louis, MO,
14-16 Sept. 1987.
13
14
1978, pp. 175-181.
2432, pp. 569-588.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 35/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 36/637
24
R. J.
ENGLAR
S =
ground roll
Td=
duct total temperature
V =
freestream velocity
vj
=
blowing jet velocity, isentropic
x/c = nondimensional chord-wise location
xTO = takeoff distance
a =
angle of attack
= ratio of specific heats
Sf flap deflection angle
j =jet deflection angle
p =
freestream density
pj =
blowing jet density
Ic
=
yaw angle (side wind angle)
I. Introduction
HE USE of pneumatic devices in the form of blown jet airfoils has been
T mployed or been under consideration in the field of aerodynamics as far
back as the 1930s, and perhaps even earlier.’,* In most of these devices,
which generally fall into the categories of jet flaps or blown flaps, a jet sheet
exits from the trailing edge of the airfoil at a fixed angle or tangent to a flap
with a sharp trailing edge. This augments aerodynamic forces by entraining
and deflecting the airfoil flowfield pneumatically, rather than solely by deflect-
ing a mechanical surface. These are “pneumatic flap” lift augmentors, which
have been shown to be successful if a sufficient onboard source of compressed
air is available. The aerodynamic concept now known as “circulation control”
(Fig. 1) is a logical follow-on to these devices, with one very important differ-
ence, which has made a significant performance improvement. The tangential
jet sheet exits over the curved trailing edge of the surface replacing the flap,
and this curvature can turn through a full 180deg or more. The jet remains
attached to that curved surface because of a balance between the subambient
pressure in the jet sheet and the centrifugal force in the jet going around the
curvature. Initially, at very low blowing values, the jet entrains the boundary
layer to prevent aft flow separation, and is thus a very effective boundary
layer control (BLC; see Fig. 1 lift plot). Eventually, as the jet continues to
turn, a rise in the static pressure plus viscous shear stress and centrifugal
force combine to separate the jet sheet, and a new stagnation point and stagna-
tion streamline are formed on the lower surface. The large flow entrainment rate
of this jet and the large deflection of the stagnation streamline produce a pneu-
matic camber, and thus pneumatic control of the airfoil’s circulation and lift.
Although it is a very effective BLC device, the interest in this concept comes
from its ability to further augment the circulation and lift, and thus giving
rise to the name “circulation control (CC).” Several additional benefits
became obvious from early experimental investigations of the concept as a
means of lift augmentation:
1) Only very sm all flap size or even nonmoving con trol surfaces were
2)
Lift augmentation could be achieved independent of airfoil angle of attack.
required.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 37/637
OVERVIEW
OF
CC PNEUMATIC AERODYNAMICS 25
TANQENTIAL
BLOWING
OVER
ROUNDED
TRAlLlNQ EDGE
SURFACE
Fig. 1 Basics
of
circulation control aerodynam ics.
3) Jet turning angle was no longer limited by physical je t exit angle or blown
flap deflection angle.
4) Very high force augmentation was generated per unit of input blowing
momentum.
Roughly 70 years have passed since the very earliest revelation of this type
of curved pneumatic device , and a very large variety of pneumatic configurations
have been proposed and evaluated. The author has been actively involved
with many of these since around 1967. To further expose this wide range of
actual and potential applications, this paper will discuss a large number of
these pneumatic devices with which the author is familiar from both past and
current research, as well as provide an indication of w here the use of CC aerody-
namics may be heading. It is by no means a com plete and exhaustive study of all
known efforts, but rather contains representative cases from a wide variety of
pneumatic force/moment augmenting and modifying devices. This paper con-
centrates primarily on fixed-wing aircraft, but CC is certainly not limited to
that application alone. The following examples will confirm the multiple uses
of CC devices as the following: 1) aerodynamic force and moment augmentors
(Fig. 1 shows
ACL/CIL 80,
or
8000
return on the invested momentum);
2
aerodynamic force and moment reduction if/when needed (drag in climb
out and cruise); 3) aerodynamic moment control and stability augmentation;
and 4) aerodynamic device simplifier (moving parts elimination, complexity
and weight reduction).
11. Coanda
Effect
The CC concept is actually based on the now well-known Coanda E f f e ~ t , ~ - ~
named after the Romanian inventor Henri Coanda, who claimed to have
discovered
it
in Paris prior to 1935. There is a Romanian postage stamp (and
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 38/637
26
R. J.
ENGLAR
associated story) showing that Coanda had originally used the Coanda device
for a totally different purpose: as a means to deflect the exhaust of a radial
piston engine away from a wooden aircraft fuselage. During its first flight,
these shielding plates actually entrained the hot exhaust flow inward, igniting
and destroying the aircraft. Figure 2 shows the basic Coanda device as later
formulated by him (after the fire exhaust incident) and its application to a
fixed-wing aircraft (which in this case appears to be a form of BLC device).
Note that in these (and in all other Coanda cases found), Coanda aligns acute-
angle “steps” downstream of one side of a jet nozzle to deflect the jet to that
side and entrain large masses of fluid from the opposite side. The distinctive
steps and angles were intended to generate a separated vortex flow at each
corner, and thus enhance mixing there. The concept was applied by Coanda to
many other devices, including car engine exhaust scavengers, wind-tunnel
turning vanes, thrust augmentors, water propulsion units, injection wind
tunnels, deflection surfaces, and rotary pumps. However, efficiency questions
arose because of added friction along all the steps and separated flow at each
corner. Nevertheless, the concept forms the basis for the present CC aerody-
namics; an infinite number of small-angled steps simply becomes a continuous
curved surface with even greater entrainment capability and less energy loss
due because its lack of discrete corners.
The following discussion will present a number of favorable applications of
CC aerodynamics, where the governing difference between CC and the jet
flap/blown flap will be the continuously curving surface downstream of the tan-
gentially blown jet, with force
augmentation/modification
being m ainly a factor
of je t blowing parameters, not the ang les of the sharp flap trailing edge or the je t
angle relative to the chord line. The main emphasis here will be on fixed-wing
devices and applications. Application of CC to rotary-wing aircraft offers
many additional benefits, as discussed in Refs. 6 to 9. These are based on
Original Coanda
Device, Approx. 1935
Fig. 2 Coanda devices and high-lift, low-drag wing?
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 39/637
OVERVIEW
OF
CC PNEUMATIC AERODYNAMICS 27
high-lift generating capability independent of angle of attack, as Fig. 3 shows,
thus eliminating the previously required cyclic and collective pitch blade
mechanisms.
Before proceeding, it is important to define the blowing m omentum coefficient
C as
where the last definition only holds for two-dimensional incompressible
flow ( p j= pm) . Typically, the jet velocity in ft/s is calculated from isentropic
relationships as:
where the subscript d implies total conditions in the blowing plenum duct,
subscript
3
is freestream, R =
1716
ft2/(s2
OR),
and
y
=1.4 for air. A jet
X
U
T H I C K N E S S I C H OR D , tlc
Fig.
3
Maximum lift of blown circulation control airfoils.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 40/637
28
R. J.
ENGLAR
expansion to the actual static pressure just outside the je t slot would yield higher
calculated values of vj and thus C,, but would vary as the external flow
conditions or shape changed,
so
would be hard to duplicate as a universal
design parameter. Mass flow
m
is almost always measured under test conditions
using appropriate flow meters, but can be calculated isentropically as well using
compressible flow relationships. Before going any further, please note that
there is nothing that prohibits the jet velocity from being supersonic unless the
geometry is such that a shockdown back to subsonic flow causes the jet to
detach from the curved surface. It will soon be seen that it is often advantageous
to have a higher speed je t than a lesser speed one. The momentum term mvj can,
of course, also be thought of as a jet thrust.
111. Applications of Circulation Control, Past and Present
A. Circular Cylinder Stopped-Rotor Aircraft
An early application of CC was developed by the British National Gas Turbine
Establishment (NGTE) in the mid-l960s, when it was desired to produce a stop-
pable-rotor VTOL aircraft. In this concept, a blown two-bladed rotor could
produce very high lift per blade just to get the aircraft to hover, then be
stopped and stowed within the helicopter fuselage for forward fixed-wing
flight.697A circular-cylinder cross-section slotted-pipe rotor appeared to be an
ideal solution , because, as Fig. 3 shows, its thickness/chord ratio of
1.0
presents
the possibility of C,
= 4.rr
if flow can be made to stay attached. As the figure
shows, values even greater than 4.rr were generated by blown CC cylinder
rotor blades when excess thrust in the vertical direction (the jet flap effect) was
included at higher t c values. However, the high drag of a 100% thick circular-
cylinder airfoil proved to be a difficult problem and reduced the aerodynamic
efficiency of these airfoils to unacceptable values.
A similar circular lifting surface' was also pursued by NASA Langley in the
1960s to provide lift on takeoff and landing by blowing on the circular fuselage
cross-section of a hyperson ic aircraft, as well as for return after launch of missile
or rocket boosters having circular cross-sections. Whereas lift coefficient values
over
20
were measured a t very low Reynolds numbers for an end-plated-cylinder
tunnel model with multiple slots, a single-slotted cylinder produced Cl= 18 at
C,
= 6. This lift augmentation of only three times the input
C,
implied the
need for a large air supply. The associated drag coefficient of over 9 gave a
lift/drag ratio of only
2 ,
or even less if the blowing coefficient were added to
the drag to yield an equivalent drag coefficient. Clearly, high lift was available,
but the lift-associated drag and required blowing coefficient posed serious
problems.
B. Elliptic-Airfoil CC Rotor
As interest in circular cylinder CC blades for helicopters was lessening in the
United Kingdom, it was rising dramatically in the United S tates in the late 1960s
as a possible means to increase rotorcraft performance while greatly simplifying
the entire rotor system m echanical hardware. Th e US effort was centered a t the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 41/637
OVERVIEW
OF
CC PNEUMATIC AERODYNAMICS
29
Navy’s David Taylor Naval Ship R&D C enter (DTNSRDC), where the approach
taken was to develop lower-drag, high-lift rotor blade sections by converting the
circular cylinder profile into a much thinner blown elliptic airfoil. These efforts
also became the basis for fixed-wing efforts, and are presented here to clarify
understanding of these CC pneumatic devices. Figure 4 shows several such
single-slotted CC rotor elliptic airfoils, where the obtainable Cl is lower than
for the cylindrical airfoil, but the required
C,
is a factor of 10-20 less. Note
that this performance is all at angle of attack a = 0 deg, providing a nonpitching
alternative to both the mechanical cyclic and collective angle of attack variation
required of conventional rotor blades. Note the very high force augmentation,
ACl/C, of 80, representing an 8000 return on the momentum invested. Also
shown for comparison is a typical 30-deg jet flap applied to a 15% thick
ellipse airfoil; the greatly reduced force augmentation of the jet flap is evident
because the jet exits from the lower surface of the airfoil at a fixed angle. It
should be clear that CC is not a jet flap, but achieves its high-lift capability
because the stagnation stream line movement and resulting circulation can be
controlled and increased well beyond that of a sharp trailing edge. Figure 5
shows the equivalent lift-to-drag ratio of sample elliptic CC airfoils, where the
YOMEMUM COEFFICIENT C
Fig. 4 Typical blown-lift capabilities of two-dimensional CC elliptic airfoils at
a =
0 deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 42/637
30
R. J.
ENGLAR
equivalent drag in the denominator now includes a severe penalty for compressor
power required as well as for intake ram pressure.lo9l1Maximum equivalent L ID
values roughly 6-7% greater than the conventional unblown rotor blade NAC A
0012 a irfoils (varying only
a
re seen for the 20% CC ellipse (at
a
= 0
deg), but
at a lift coefficient 30% higher, at about 1.3. Furtherm ore, the Cl can be increased
up to 6 or 7 if desired, but at a lesser
LID,,.
For additional comparison , if the equivalent drag is defined as merely adding
C, to the measured drag (i.e., CDE
=
Cd C,), then
LID,,
values of over 120 at
Cl= 2.5 are possible, all at a = 0 deg (almost three times the Cl of the 0012
airfoil at stall). The efficiency and simplicity of CC was obvious from these
tow-dimensional airfoil results, and a serious effort to develop these CC airfoils
was undertaken. Reference 12 summarizes much of this Navy effort at
DTNSRDC for the years 1969 through 1983, as well as providing a summary
of CC-related research conducted by other agencies (in the United States and
abroad) outside the Navy from 1956 to 1983.
X
3
Y
d
SECTIONAL
LIFT COEFFICIENT
C/
Fig. 5 Equivalent efficiencies for CC and conventional two-dimensional airfoils.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 43/637
OVERVIEW
OF
CC PNEUMA TIC AERODYNAMICS 31
In 1979, a CC R otor flight demonstrator based on a Kaman H-2 helicopter was
flown with pneumatic aerodynamic and control systems replacing conventional
mechanical cyclic and collective blade pitch. 13-
l
Whereas this flight vehicle
was hindered by control system response phasing problems, which limited its
flight test envelope, it did demonstrate the ability to substitute pneumatics for
mechanical blade lift and control devices for hover and forward flight. It also
led to the possibility of higher harmonic control of helicopters, where cyclic
lift variations at frequencies higher than one per revolution were possible to elim-
inate rotor-induced vibrations. The absence of blade collective and cyclic pitch
links is possible; they can be replaced by internal control cams or valves to
vary blowing pressures.
C. Circulation Control Airfoil Development
Considerable CC airfoil development was ongoing at this time, both exper-
imentally and analytically. A number of CFD techniques using various
Navier-Stokes codes have been developed and used to understand the relevant
viscous flowfields. These will not be discussed here, but can be found summar-
ized in much more detail in Refs. 12 and 16.A typical example of CFD-calcu-
lated streamlines and velocity vectors” is shown in Figs. 6 and 7 for a generic
flat-sided semi-elliptic CC airfoil. Of particular interest here are the computed
velocity vectors and streamlines downstream of the slot on the blown trailing
edge (Fig. 7), where the stagnation point of the jet sheet appears to be turned
nearly 130-140 deg from the jet exit. A considerable number of additional
CFD analyses, both subsonic and transonic, have been conducted by various
investigator^.^^ ^^ ^^
Fig. 6 Computed streamlines for simplified CC airfoil a 2 deg, Cl =
4.6).”
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 44/637
32
R. J.
ENGLAR
Fig. 7 Computed velocity vectors and streamlines," CC airfoil.
A num ber of experimental programs have also been conducted to understand
the CC phenomenon and the details within the blown curved surface region.
Two-dimensional laser-velocimeter measurements at Lockheed for the same
CC airfoil as in Figs. 6 and
7
showed mean velocities that confirmed the CFD
results already presented. Again, je t flow turning to a separation point/stagnation
streamline approximately 130-145 deg from the slot was seen (Fig.
8).
Exper-
imental investigations by the present author of a very similar generic airfoil
(Fig. 9), used surface static pressure, static pressure across the je t, and a rotatable
hot-film shear stress probe to measure the actual separation point location (where
Fig.
8
CC velocity vectors recorded by Lockheed Laser Doppler Velocimeter."
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 45/637
OVERVIEW OF CC PNEUMATIC AERODYNAMICS
33
G
0
t
3
MOMEN TUM C OEFFICIENT, C,
NONDIMENSIONALCHORDMSE STATION x / C
Fig.
9
Two-dimensional semi-ellipse CC model geometry, plus measured lift and
static pressures as functions of C, and slot height.
shear stress = 0) as a function of blowing and slot height. T he resulting C, nd C,
distributions are seen in Fig. 9. As Fig. 10 shows, jet turning as high as 170-
175 deg was measured for this airfoil. At a constant C greater turning occurred
with a smaller slot height because the resultant je t velocity and entra inment are
higher a s je t area reduces. Figure
9
(left plot) shows that this greater velocity
and je t turning clearly results in generation of higher Cl, where values close
to nine are possible at a
=
0 deg (although tunnel flow impingement occurs
here). Figure 9 (right plot) presents associated static pressure distributions on
the airfoil. These analytical and experimental data confirm the effectiveness of
blowing to greatly deflect the entire flowfield and then strongly increase the
circulation and lift on these very generic airfoils, to the point that very high
lift is produced without wing flaps and slats and at Odeg angle of attack.
Some additional information on generic CC airfoil performance is provided in
Ref.
20.
One last note on CC airfoil performance: as previously mentioned, smaller
slot height yields a larger return in Cl at constant C, than does a larger slot
height, primarily because of greater
vj/Vm
and extra flowfield entrainment.
Figures 9 and 10 show this trend. However, if the static pressure coefficient
just outside the slot exit C,,,)s known or can be determined , a new parameter
defined in Fig. 11 can be used (when vj is expanded to this local condition to
yield CBLc) o collapse the different slot height results (left) into a single curve
(right).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 46/637
34
R. J.
ENGLAR
Fig. 10 Blowing jet separation point location measured by hot-film shear stress
probe.
c, LC
CONVENTIONAL MOMENTUM COEFFICIENTS
{EWANDED TO fREESTREAM CONDITIONSI
LOCAL
MOMENTUM
C O E f f l C l E N l S
IEXPANDED
TO
LOCAL WND ITIONS AT SLOT)
Fig.
11
Comparison with momentum coefficients based on local jet exit static
pressure (right) and variation with slot height.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 47/637
OVERVIEW
OF
CC PNEUMATIC AERODYNAMICS
35
D.
X-Wing
Aircraft
An extraordinary use of unusual CC airfoils is the X-Wing VTOL
~ e h i c l e , ~ l - ~ ~combined rotary/fixed-wing aircraft equipped with a four-
bladed rotor, which was designed to take off and hover w ith the same nonmecha-
nical cyclic and collective benefits as those already described. How ever, forward
flight at speeds roughly twice the limit on conventional rotors could be achieved
using a “reverse velocity” blown rotor/wing concept (Fig. 12). Typically, as
vehicle speed increases, the retreating blade of a rotor sees a resulting velocity
that is the difference between the vehicle forward speed and the blade rotational
velocity; this can rapidly become a reverse flow at the blade trailing edge, an
unacceptable region that moves further outboard as speed increases. Lift on
that “stalled” blade segm ent can actually be negative; the rotor might not be trim-
able in roll, and d rag increases dramatically. The X-wing avoids this problem at
high speeds by employing
CC
on each end of the blade (Fig. 12), and a “clever”
control system can blow whichever slot is currently on the airfoil’s trailing edge.
Thus, the airfoil never experiences flow from the “wrong” direction. The entire
system can be simplified even further by use of simultaneous blowing from
both leading and trailing edges of the double-ended airfoils24 (Fig. 12, right).
Note that even if the flow is coming from the wrong direction (dashed curve),
the dual-slotted airfoil still yields
80-90
of the single-slotted airfoil’s lift,
even when the leading edge is counter the conventional direction (compared to
little, zero, or negative lift from a conventional airfoil). This allows rotor-
borne flight at a much higher speed until eventual conversion to a fixed wing
in an X-configuration is achieved, with the representative TE slots on each
blade of the now fixed wing being used for roll and pitch control without
moving surfaces. This concept was actually “flown” full-scale in the NASA
Ames 40 x 80 ft tunnel and successfully completed the transition from hover
to stopped-wing using pneumatics. Two representative configurations from
Refs. 16 and 25 are shown in Fig. 13.
D U A L
PLENUM
AIR FOIL SECTION
HELO DIRECTION
Fig. 12 Dual blow ing on a reverse velocity rotor2’ and blow n lift
of
dual-slotted
CC
airfoils.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 48/637
36
R. J.
ENGLAR
LIGHTLY
LOAOEO
LOW
TIP SPEED
TAIL F A N
Fig.
13
X-wing rotor configuration with rotor stopped, and con trol systems.
E.
Circulation Control Wing (CCW )
The high-lift capability independent of angle of attack, which was demon-
strated by the C C rotor airfoils above , led to the application of C C as a simplified
very-high-lift dev ice for STO L aircraft. The airfoil in Fig. 1 is representative of
this simplified pneumatic concep t, where both the mechanical TE flap and the LE
flap or slat have been replaced with nonmoving pneumatic systems. Primary
development of the concep t took place in conjunction with C C rotor development
efforts at the Navy's DTNSRDC12,16,26-30n the late 1960s to early 1980s.
Initially, the concept was modeled as a small add-on device that would
convert the wing flap's sharp TE into the round CC Wing (Fig. 14, right),
which was tested at DTNSRDC in specialized two-dimensional high-lift test
facilities. Com pared to results from a family of more conservative blown
flaps (Fig. 15, from Ref.
30),
the CCW profiles showed two significant advan-
tages. They could generate greater Cl than the blown flap, because of much
greater streamline displacement, and had no sharp TE to limit streamline
turning, or for the same chord-length device, could generate the same incremental
Cl at much less C, required. An alternative 180 deg rotatable C CW TE is also
shown in Fig. 14 (left), which, although
it
may be mechanically simpler, pays
the penalty of losing wing area in the blown high-lift mode.
Numerous two- and three-dimensional wind-tunnel evaluations and feasibility
studies led up to the flight test of a fixed CCW dev ice on an A-6/CCW STO L
demonstrator a i r ~ r a f t ~ l - ~ ~n 1979. Flow visualizations in Fig. 16 show a full
180 de g of je t turning on a static 1/8-scale model of the test aircraft in the
DTNSRDC tunnel, and Fig. 17 shows the CCW installation on the fixed flap
of the A-6 flight-test aircraft. Because this was a proof-of-concept flight test,
the CCW device was not retractable and the air supply lines were mounted exter-
nally and cross-ducted in the fuselage, where they connected to the high-pressure
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 49/637
OVERVIEW
OF
CC PNEUMATIC AERODYNAMICS 37
CONVENTIONAL
Fig.
14
Retractable/storable CCW trailing edges.
bleed ports of the standard J-52-P8A turbojet engine . Results using only availab le
bleed air from the engines confirmed maximum
C
values 120% greater than the
conventional Fowler flap, or, even more applicable, 140% increase in the usable
lift coefficient at takeoff/approach angles of attack. Also confirmed were 30-
35% reductions in the takeoff and approach speeds resulting in 60-65
reductions in takeoff and landing ground roll distances, and yielding values as
short as 600-700 ft. This full-scale confirmation of CCW also implied that
FLAP
CHORD
LENGTH
MINIMUM
Cp
REQUIRED
Fig. 15 Comparisons between CCW and blown flap airfoils at a = 0 deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 50/637
38
R. J.
ENGLAR
Fig. 16 CCW jet turning on the A d / C C W wind tunnel model at DTNSRDC.
there was sufficient extra
CL
generated to increase the liftable payload by 75 if
the conventional takeoff ground roll distance were used. Also show n was that the
additional lift-induced drag resulted in much steeper glide slopes on approach,
where higher engine power settings (which could also be used for quicker
response during waveoff) were offset by this excess drag.
Fig. 17 A-6/CCW STOL flight demonstrator aircraft.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 51/637
OVERVIEW
OF
CC PNEUMATIC AERODYNAMICS 39
Fig.
18
WVU STOL demonstrator
CC
airfoil?'
A smaller CCW demonstrator based on a prop-driven BD-4 general aviation
aircraft had been flown earlier by West Virginia University (WVU).'6,35,36 n
the flight-tested configuration, the CC blown cylinder was mounted at the TE
of a hinged flap that rotated 180 de g aft to increase the effective high-lift area
by 20%, and included a boundary layer control (BLC) suction slot at the flap
hinge upper surface (Figs. 18 and 19). The blowing air was supplied by an
onboard 200 hp compressor (APU), which provided enough air for blown ailer-
ons in addition to the CCW . The section lift coefficient on the blown CCW wing
section was increased by a factor of nearly 2.5 with blowing. Wing dow nwash on
the tail reduced trimmed
C
increase to a factor of 1.92, but provided three-
dimensional lift augmentations of ACL/C,
=
15.2, a significant increase
should the required airflow be available from a general-aviation aircraft
engine, for example, if using a supercharger or turbocharger.
Fig.
19
WVU BD-4 based STOL demonstrator a ir ~ r a ft .3 ~
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 52/637
40
R. J.
ENGLAR
CORRECTED mnm. F ,
MS
x 1 6
Fig. 20 Thrust performance of J-52-PSA turbojet engine with bleed.
Both of these fixed-wing flight programs demonstrated the feasibility of CCW
as an operational
STOL
system in terms of high lift, short takeoff and landing,
and simplicity, but also identified issues still to be resolved. Among these were
the drag of the device in cruise flight (WVU solved this but at the cost of a mech-
anical 180-deg rotating flap that stowed in the aft wing cavity, and GTR I solved it
with the dual-radius airfoil discussed in Sec.
111.
F.) and, of course, the need for an
onboard air source. Figure
20,
which presents turbojet engine ground test data31
taken during the A-6/CCW program, show that the airflow acquired from high-
pressure compressor bleed ports could be increased up to three to four times that
of the standard engine spec bleed limit without overheating, but obviously at the
cost of takeoff thrust lost. Similar data for turbofan engines show that engine core
bleed is much more costly in thrust loss (although lower-pressure fan bleed is
possible), and thus the idea of an ejector to trade excess pressure for extra
mass flow appears feasible. However, the need to reduce CCW drag in cruise
is a necessity for operational aircraft.
F. Advanced CCW Airfoils
DTNSRDC
and G rum man took tw o approaches to the d rag p r ~ b l e m , ~ ~ ’ ~ ~ -
one a fixed simple radius reduction and the second a very-small-chord deflectable
CCW flap. From the nondeflectable ~ t a n d p o i n t , ~ ~supercritical-type airfoil was
employed as the baseline because it already had a bluff base thickness between
0.005~ nd 0 . 0 1 0 ~ . CW rounded and semiround (96 deg arc instead of 180 deg)
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 53/637
OVERVIEW
OF
CC PNEUMATIC AERODYNAMICS 41
designs were tested, including a series of smaller radii, looking for reduced drag
without loss of lift augmentation. The CCW/Supercritical airfoil was developed
primarily from a low-speed standpoint, where a T E radius of 0.009~ as found to
produce very little drag penalty yet have superb lifting capability. Figure 21
shows its lift curves at constant C compared with a family of mechanical
multi-element flaps. Not only do the no-moving-parts CCW airfoils generate
the same or greater lift as the maximum Cl of a triple-slotted-flap airfoil with
mechanical slat, they also do so at
a
= 0 deg. Note that the large leading edge
of the supercritical airfoil provides a natural nonmoving LE device, which gen-
erated sim ilar stall angles to the mechanical slat. One further benefit is the cruise
drag polar (Fig. 22), which is slightly higher unblow n than the baseline supercriti-
cal airfoil, but with very slight blowing can reduce
d
to less than the baseline
while also increasing lift, both at constant
a.
For clarity here, measured
d
includes C because the wind-tunnel balance cannot easily separate blowing
thrust from drag. That is why Fig. 22 shows negative drag recorded with
blowing. This is accounted for in the equivalent drag term, LID,,, where
C,
is
included (see Ref. 1 0 for a more detailed explanation). Additional benefits of
blown CC airfoils at speeds up to transonic were shown in the compressible
flow tests of Ref.
40,
where blowing was seen to produce a very favorab le bound-
ry
layer/shock interaction, drag reduction, and increased
Cl
(Fig. 23).
The second approach to the drag problem was a simple CCW flap with a curved
upper surface and a sharp trailing edge (Fig. 24, from Refs. 38 and 41). Here, a
short-chord flap (less than 0 . 1 0 ~ ) ivots about a hinge on the lower surface and
exposes a smaller-radius CCW surface downstream of the tangential slot. This
radius is approximately the airfoil thickness at the slot location, less the slot
height. The upper surface of this flap is a second arc of much larger radius, the
radius being chosen to keep the arc close to the airfoil aft contour. As the small
flap is deflected on this dual-radius CC W airfoil, the large radius produces an arced
CC aft surface with a turning arc much larger than the flap deflection angle.
MULREWEHTMECHANICAL
~ W TIRFOILS
N D W V H G - P A R T CCWlsUPERCRlTlcALAIRFOIL
Fig.
21
CCW/Supercritical, dual-radius CCW , and conventional mechanical flap
airfoil comparisons.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 54/637
42
R. J.
ENGLAR
Fig.
22
Low-speed drag polars for CCW /Supercritical airfoil.
Freedream Mach Amber, Iy
Fig. 23 Transonic lift caused by blow ing for three pneumatic ellipse airfoils '
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 55/637
OVERVIEW
OF
CC PNEUMATIC AERODYNAMICS
43
Supercritical Contour
Fig.
24
Dual radius CCW airfoil with LE blowing.
For the 90-deg flap shown here, the jet turning angle is about 135 deg (compare to
Figs.
8
and lo ), limited by the TE com er. With flap retracted to 0 deg, the airfoil is
in a sharp-TE cruise configuration. The slight mechanical addition provides
unblown camber as well. The leading edge employs an inverted tangential slot
to replace any mechanical flap there. Lift data for the 90 deg flap configuration
are also shown in Fig. 21, w here Cl increases of 35% over the CC W/Supercritical
airfoil occur, with considerably greater increases over the conventional flaps.
Figures 25 and 26 also show additional advantages of this configuration: the
ability to dramatically interchange lift and drag as the small-chord CCW flap
2-D CW SUPERCRITICALIRFOIL, UA LRADIUS LAPS,
DRAG POLARS, THE PENALTYOR LIFT
Fig.
25
Drag polars of CCW dual-rad ius airfoil at various flap ang les.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 56/637
44
R. J.
ENGLAR
2-D CCW /SUPERCRITICAL AIRFOIL, DUAL-RADIUS CCW,
u,
degrees
Fig.
26
Dual-radius 90-deg flap CCW airfoil lift as functions of Y and C I ~ ~ .
is deployed (Fig. 25) and the increased lift and stall
a
s LE blowing is activated
(Fig. 26) .The thrust/drag interchange in Fig. 25 implies the potential for high lift
and drag for STOL approach (remember, induced drag due to high lift is not
included in this data for two-dimensional airfoils) or high lift and reduced drag
for takeoff. Figure
26
shows the capability of this nonmoving LE device to
reattach flow, prevent stall, and dramatically increase
C
Figure 27 combines
the above data in terms of Cl vs LID,, where the equivalent drag coefficient is
defined as CDE= Cd C to account for the blowing required to yield these
drag changes. These data include four CCW flap angles and various LE and
TE blowing values. Figure 27 includes a locus of achievable Cl vs the associated
efficiencies in comparison to the clean cruise airfoil (flap = 0 deg, CE = 0,
C
=
0). This plot confirms the ability of CCW airfoils to generate very
high lift and associated drag (reduced LID,) for approach, plus much higher
L /D e
at somewhat lower Cl for takeoff and climbout. Because of the latter, the
30-deg CCW flap at reduced
C,
appears to be an excellent configuration for
takeoff.
One additional benefit results for the 0-deg flap CCW case. When in cruise,
drag is low because of the sharp TE, but should blowing be initiated without
flap deflection (Fig. 28) , significant lift is generated by the flap curvature,
while drag reduction occurs due to thrust recovery. Note the comparison to the
NASA Energy Efficient Transport slotted, flapped airfoil. Not only is lift
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 57/637
OVERVIEW
OF
CC PNEUMATIC AERODYNAMICS 45
CCW Dual-Radius Airfoil DRAG POLARS: the Penalty for Lift?
2-D Lift Coefficient,
C,
Fig. 27 Lift and equivalent efficiencies of dual-radius
CCW
airfoil.
greater for the CCW cruise airfoil, the drag polars move into the thrust recovery
region. From these results, one can also immediately realize the potential of this
high-lift system as a nonmoving roll/yaw device. Blowing only the undeflected
right wing’s flap will produce a lift (roll with right wing up) and favorable yaw
(nose left), thus yielding favorable roll/yaw coupling from a nonmoving surface,
instead of the usual adverse roll/yaw coupling from a conventional aileron.
A study41 was conducted for NASA Langley Research Center to evaluate the
effectiveness of applying this concept to an Advanced Subsonic Transport. Here,
the dual-radius CCW of Fig. 24 was applied to a 737 wing characterized in
Fig. 29. The typical 15 moving elements per wing were replaced with the CCW
single element flaps and LE blowing, yielding perhaps a maximum of three
components per wing (the outboard C CW flap became the aileron, and blowing
differentially on the CCW flap replaced the spoilers for roll). Using only fan
bleed air (and the associated lower thrust lost), replacing the conventional flaps
with CCW was able to triple the usable lift at takeoff and produce the ground
roll reductions shown in Fig. 30. For lighter aircraft weights, blown takeoff
rolls of 400-500 ft are possible with
0
kn headwind, about a third that of the
conventional aircraft; w ith a 20 kn headwind (wind over deck), 200-300 ft
Next Page
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 58/637
46
R.
J.
ENGLAR
Fig.
28
Comparison of cruise dual-radius CCW Odeg flap) wi t h mechanical flap
airfoil.
737
WlNOlFLAPICONTROL SYSTEM
Fig.
29
Pneumatic airfoils simplify wing complexity.
Previous Page
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 59/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 60/637
48
R.
J.
ENGLAR
Angle of Attack,a,deg
Fig. 31 Lift augmentation on the GTRI HSCT/CCW semispan model with blown
IV.
Powered L ift and Engine Thrust Deflection
A.
CCW/USB
Mechanical flaps have been employed to entrain and deflect thrust from
engines mounted on the wing upper surface (upper surface blowing, USB ), and
it was envisioned that the entrainment capabilities of CCW could do the same
without the mechan ical complex ity ; thus the CCW/USB c o n ~ e p t ~ ' , ~ ~ - ~ *as
born (Fig. 32). Subsonic wind-tunnel
investigation
at DTNSRDC showed
no-moving-part pneumatic capability to entrain and turn USB engine thrust
well past the 60 or so degrees of a mechanical USB system, but also continuing
through 90 deg, and then rotating the thrust forward as a thrust reverser through
165
deg (Fig.
33).
Th e possibility then exists fo r high lift and thrust reversing all
in one system jus t by varying the CC W blowing rate, with a possibility of VTOL
in between (depending on installed thrust levels). Wind-on data (Fig.
34
show
very interesting lift-drag polars at a = 0 deg, with the ability to vary lift and
drag by blowing alone, independent of angle of attack. The enhanced lift capa-
bility is far more than mere thrust deflection (i.e., ACL=
CT
[sin a
a] .
t
W R I P B L E M U S T O E F W 3 l D + l
UJETOCC.PLENUHPFESWW
U A R m
Fig. 32 USB and CCW /IJSB pow ered-lift concepts.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 61/637
OVERVIEW
OF
CC PNEUMATIC AERODYNAMICS 9
T = ENQINETHRUST LB)
1 * ANQLE OF AT YXK
q = DYNAMIC PRESSURE FT T IhVi
P, = SLOT PRESSURE
mV,
= SWT MOMENTUM
T = RESULTAKTTHRUST
L
0 CCW ALONE)
25.41LB
486T L5
Fig.
33
CCW/USB model static thrust deflection by blow ing only.
results fro m the increased velocity fro m the engine exhaust being entrained onto
the blown lift surfaces, and the greatly increased circulation lift beyond the
powered wing only.
A full-scale ground test was performed by the present author at NASA A me s
with the CCW/USB mounted behind one engine of the NA SA Q uiet Short-haul
THRUST
DRAG
Fig. 34
CCW/USB
model lift-drag polars,
Y
=
0
deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 62/637
50
R.
J.
ENGLAR
Fig. 35
CCW/USB
test assembly on the QSRA aircraft.
Research Aircraft (QSRA), with the aircraft mounted on a force b a l a n ~ e . ~ ~ , ~ ’ - ~ l
Figure
35
shows the installation behind the left inboard engine of the QSRA.
Thrust deflections as high as over
100
deg were recorded behind this single oper-
ating engine. These data are expected to improve if the two engines per wing are
operated together and the two exhaust sheets converge for even better turning. As
a result of this test, Navy feasibility studies46were conducted for a sea-based tur-
bofan-powered STOL aircraft using both CCW and CCW/USB (Fig. 36 . These
studies, based on the preceding powered model wind-tunnel tests, showed takeoff
ground rolls of
100-200
ft (Fig. 37), varying with weight, blowing, and thrust
levels, and resulting from powered C , values of
8-9.
Fig.
36
Proposed
CCW/USB
Navy
STOL
aircraft.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 63/637
OVERVIEW OF CC PNEUMATIC AERODYNAMICS
51
GROSS
WEIQHT. L B l l O m
Fig. 37 Takeoff ground rolls for proposed
CCW/USB
Navy STOL aircraft.
B.
CC/Jet Deflection
In a related effort,52CC entrainment was also applied to high-performance air-
craft to yield thrust deflection for much higher engine exhaust velocities, where
lesser jet turning could still provide excellent STOL potential due to higher
thrust/weight ratio. An example is shown in Fig.
38.
In-house unpublished exper-
imental work by the present author provided similar studies, where we were able
to deflect supersonic jets from rectangular nozzles by more than 80 deg, using
blowing jet momentum values around 10 of the eng ine thrust.
BLOWING WY) ENT W,
I,
US
Fig.
38
Pneumatic thrust deflection of rectangular jet exhaust left, from Ref. 52)
and unpublished static test results of a similar configuration right).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 64/637
52
R.
J.
ENGLAR
C. Pneumatic Channel Wing
A configuration using similar thrust deflection capability of a CC trailing
edge has recently been under development by G TRI for
NASA
Langley Research
Center. Called the Pneumatic Channel W ing (Fig. 39), it employs blowing at the
TE of a 180 deg channel (similar to the much earlier but unblown C uster Channel
Wing) to entrain the propeller’s thrust, augment the velocity in the channel, and
thus generate high-powered lift. Figure
40
(from Refs. 5 3 and 54) shows typical
GT RI wind-tunnel lift data as a function of both blowing and thrust compared to
the baseline unblown channel wing configuration, where untrimmed CL,, is
increased by a factor of over 7 to a value of 10.5- 11. Reference
54
shows pre-
dicted takeoff ground rolls of less than 100 ft on a hot day at 3000 ft altitude
using wing angle of attack of only 10deg. Further details of this concept’s capa-
bilities and the associated data are found in the paper by E nglar and Cam pbell in
this volume, and al so in NASA C P 2005-213509.
Fig. 39 Conceptual pneumatic channel wing and semispan model in GTRI MTF
tunnel.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 65/637
OVERVIEW OF CC PNEUMATIC AERODYNAMICS
53
a,degrees
Fig. 40 Pneum atic channel wing lift from thrust and blowing?4
V. Other Aircraft Applications
A. CC Propeller
In a manner somewhat similar to the CC rotor above, CC airfoils have
also been incorporated into general aviation propeller designs to replace
complex and expensive mechanical variable-pitch blades with fixed-pitch
pneumatic blades that change aerodynamic and thrust characteristics through
mass flow variation to each blade. Figure
41
shows a proposed application,
where the propeller blade airfoil is the CCW/Supercritical type of
Fig.
21.
References 55, 56, and 57 discuss feasibility studies, and concluded
that such a pneumatic variable-pitch propeller was possible and held interesting
promise depending on the details and costs of an air compressor (such as an
aircraft supercharger or turbocharger) to supply the blowing. The study also
envisioned supersonic jet blowing to be a possible problem, but much
of
the
CCW data already presented above have blowing pressures well above choked
(sonic).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 66/637
54
R.
J.
ENGLAR
F low Control V a l v e
Compressor
Fig.
41
Circulation control propeller system .
B. Moment Control Stability Augmentation and Induced Drag Reduction
The preceding data and applications show the ability to pneumatically
augment or modify lift and drag without use of moving parts (except possibly
very short chord dual-radius CCW devices) and with a high rate of return on
input jet momentum. The application to a pneumatic rudder or even winglets
can provide side force generation as well.58 It should be obvious that augmenting
the aerodynamic force capability of any control surface by blowing can also
either increase the control power or reduce the required area of the device,
with associated benefits including maintaining stability levels but reducing
cruise drag. A few further and less obvious examples of pneumatic control
devic es will now be mentioned; many o thers can be found in Refs. 12, 16, and 58.
The aft suction peak downstream of the CCW slot (usually at 95 chord or
greater) produces very large nosedown pitching moment, which, besides
having to be trimmed, a lso produces greatly enhanced longitudinal pitch stability.
In fact, the A-6/CCW flight demonstrator had such large negative values of
dCM/dC l that the center of gravity of the flight-test aircraft was moved af t by
an additional 10- 15 chord to aid in trim, and the aircraft still had greater longi-
tudinal stability than the conventional A-6, flaps A clever application
of CC W for mom ent control is shown in Fig. 42 on a forward swept wing.59 Pre-
viously, increasing blowing on an aft-swept trailing edge pulled the center of
pressure (CP) outboard and aft, but during this tunnel e~aluation,~’he C P was
made to move outboard and thus forward with blowing. T he amoun t of
xcp
move-
ment and the resulting moment were controlled by which segments of the TE
slots were blown, and by how much.
Pneumatic roll control by differential wing blowing can produce phenomenal
rolling mo ment increments (Fig. 43), where only one wing of a CCW configur-
ation is blown. A second innovation is also show n here: letting the slot continue
around the wing tip. Now, high suction peaks at the wing tip, having a maximum
moment arm, can yield even greater rolling moment. For reference, a conven-
tional 0.20-chord aileron deflected down 30 deg on the outboard
50
span of
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 67/637
OVERVIEW
OF
CC PNEUMATIC AERODYNAMICS
55
WEIGHT FLOW-0.25 Iblsec
Fig.
42 CCW
applied to forward swept w ing for pitching moment reduction.
this wing produced an incremental hCrol1
=
-0.03. Figure 44 suggests one
further advantage of the CC wing tip, where blowing down around the t ip directly
counteracts the tip vortex rollup and relocates it further outboard, creating an
effective aspect ratio increase. Figure
44
shows the effective drag reductions
due to tip blowing*l on an already high-aspect-ratio CC rotor blade. At the
higher C values where induced drag usually dominates, C reductions of 17-
19
are
seen, with greater percentage reductions at lower CL. Lower aspect
ratio aircraft wings using this technique should yield even greater CD eduction.
One can also alter the spanwise lift distribution with spanwise tapered blowing to
approximate an elliptic distribution, and thus minimize induced drag both in
cruise and during climbout.
In 1986, the present author experimentally applied blowing from a tangential
slot along the nose of a generic high-a vortex-lift configuration, thus turning the
Coefficients based on full
span and area
Fig. 43 Roll due to C C wingtip blowing,
Y
=
0
deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 68/637
56
R.
J.
ENGLAR
WNG LIFT COEFFICIENT
Fig. 44 Tip blowing for induced-drag reduction.
fuselage into a side-force and yawing-moment generator. Other investigators
have more recently tried similar schemes, but the results shown in Fig. 45 sum -
marize these effects. At a = 35 deg, the conventional rudder was useless because
of fuselage blockage and separated flow (see
C ,
=
0
curve), but blowing on the
right side of the nose restored d irectional stability when the vehicle was yawed to
the left, and vice versa. Large side force values were also generated by blowing.
C. Microflyer and Pulsed Blowing
A combination of all of the above force and moment control applications has
been pursued recently at GTRI relative to a very small unmanned aerial vehicle
(UAV), the Pneumatic M icroflyer.60 Jet turning on a small-scale, low-
Reynolds-number wing is illustrated in Fig. 46. Pneumatic lift and control sur-
faces will be driven by gas generated by a GTRI proprietary engine powering
the flapping wings of a 6-in.-span flying-insect-like UAV. The opportunity also
exists here to take advantage of pulsed blowing, investigated in Refs.
61
and
6 2 for application to blown flaps and in Ref. 6 3 relative to
CCW.
Here, for
properly shaped blowing wave forms, mass flow required was greatly reduced
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 69/637
OVERVIEW OF CC PNEUMATIC AERODYNAMICS
57
c
I
a
CY
SIDE FORCE
COEFF.
0.40
0 .30
0.20
a . 10
0.00
0.10O . ’ O
0.201
1
CN
YAWING
MOMENT
COEFF.
I
I
a
c
t
y1
cn
2
0.10
I 5I
0.05
0.00
-0.0s
0.10
- loo -5O 50 l o c
SIDESLIP ANGLE NOSE RICHT-
Fig. 45 Tangential forebody blowing for yaw and side force control, a =
35
deg.
experimentally by up to
40
to 50 (Fig. 47 , or conversely, greater lift could be
generated by the same mean mass flow levels. Also, as a simplifying means, all
pneumatic Microflyer control moments would be generated by differential
blowing, rather than by very small moving mechanical parts.
VI.
Nonflying Applications
of
Circulation Control
A number of nonflying applications have been investigated, where the
CC
phenomenon was used to augment or modify flowfields for unique purposes.
In order to provide pitch and/or yaw control for submarines without using
mechanical stem planes, a dual-slotted “pneumatic” stem plane64 was designed
for submerged applications (Fig. 48). Here, up or down pitch of the submarine
(or right or left yawing moment) could all be provided by blowing the appropriate
slot. Towing basin tests of this concept verified that blowing w ater from the slots
when underwater was equally as effective as pneumatic devices (even if the
power required might be higher), and provided the opportunity for smaller
stem-, bow-, or sail-planes, or avoided the possible control plane jam problem
of moving hydrodynamic surfaces.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 70/637
58
R.
J.
ENGLAR
Fig. 46 Low-Reynolds-numbermicroflyer wing with
CCW
turning.
Applications of pneumatics similar to the CC rotor were both conceived as the
CC fan (Fig. 49, from Ref.
65)
and the CC windmill. Here, variation in blowing
parameters through the individual blade slots could vary the output of the fan, or
conversely, for a pneumatic windmill, vary the sensitivity of each individual
blade to the incoming wind angle and strength, as well as the radial load distri-
bution on the blades. For the windmill, blade pitch would not be required to
change mechanically for maximum performance or avoidance of rotor overspeed.
More recently, application of pneumatic concepts to improve the aerodynamic
performance of automotive vehicles has been heavily pursued at GTRI. Tests on
European Formula 1cars (Fig. 50) have verified that proper application of blowing
can dramatically increase the download frequently required for higher-speed cor-
nering of these cars, or reduce the required wing area and its associated drag (note
the absence of the conventional inverted fore and aft multi-element wings). The
high suction (negative static pressure) difference across a blown lifting wing
inverted on a race car can also entrain sufficient flow to provide cooling
through a radiator located therein. Figure 51 shows a Formula SAE car with an
aft blown wing including a pneumatic radiator (unit developed and tested at
GTR166 with assistance from the G T Motorsports team). It i s now possible to
have a multipoint aerodynamic race car design that had previously been prohibited
by the
“nonmoving-aerodynamic-components”
ule. More details of these con-
cepts and of testing of this device are found in another paper by Gaeta and
Englar in this volume and in NASA CP 2005-213509,2005.
A G TRI program originally intended only to reduce aerodynamic drag on pro-
duction cars for increased fuel economy has recently led to additional benefits.
Blowing on the curved aft panels67 of a generic streamlined car (Fig. 52)
showed drag reductions of up to 35 , but also drag increases of over 100 by
blowing different elements, which could be used as a form of aerobraking. Lift
could also be increased by up to 170 over the unblown car, or conversely, a
lower surface slot could also yield downforce if
so
desired. GTRI tests also
showed that yawing and pitching moments could be dramatically changed by
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 71/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 72/637
60
R.
J.
ENGLAR
Optional
End
P l a t e
Fig. 48 Blown m odel stern plane design, two-slotted.
FAN INFLOW
Fig.
49
Circulation control fan concept.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 73/637
OVERVIEW OF CC PNEUMATIC AERODYNAMICS
61
Fig. 50 Pneumatic Formula 1 car model in G TRI tunnel.
blowing, and lateral and directional stability could be restored by blowing only
one side of the slot. Interestingly, the blowing required for all of this could
be provided by turbochargers or superchargers now being installed on high-
performance cars.
These experimental data for automobiles have now been extended to and
adapted in a GTRI program for the Department of Energy68969o improve the
aerodynamics, performance and economics of heavy vehicles
(HV;
i.e., large
tractor/trailer trucks). Figure
53
shows blowing on all four aft comers of
the trailer; this combination is able to reduce drag, turbulent separated flow,
Fig. 51 GT RI-patented pneumatic aerodynam ic heat exchanger installed on
formula SAE race car by GT M otorsports Team.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 74/637
62
R.
J.
ENGLAR
Fig.
52
GTRI P neumatic Futu recar model for drag reduction, showing jet
turning.
spray, and aft suction on the back doors. Blowing the lower slot only can increase
download and aid in braking or provide traction in wet/icy weather, while
blowing the top slot only can generate lift and thus reduce effective weight on
the tires and rolling resistance. Blowing either side slot can offset yaw due to
gusts or sidewinds (which can yield a large component of increased highway
drag), or can help to restore lateral/directional stability. Because the response
of the blowing system can be virtually instantaneous (pressure of only
13-
14 psig can produce sonic jet velocity), safety of operation is very promising,
including the ability to prevent jackknifing by generating opposite yawing
moment for the trailer. Blowing on the trailer top leading edge also appears prom-
ising, because it can provide not only a boundary layer control device, but also
can entrain flow up through the cab/trailer gap and eliminate strong separation
and vorticity there, plus enhance cooling. Wind-tunnel investigations of this
concept on a smaller-scale model of a blown pneumatic heavy vehicle (PHV)68,69
have shown drag reductions of up to 80 relative to a baseline generic HV
Fig.
53
Pneum atic heavy vehicle configuration with potential for
5
blowing slots.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 75/637
OVERVIEW OF CC PNEUMATIC AERODYNAMICS
63
Fig. 54 Pneumatic heavy vehicle test rig undergoing road tests.
model, p lus the ability to increase drag if needed for braking, as well as provide
side forces and lateral/directional control in side winds. They have also con-
firmed that blowing only one vertical side slot at the rear of the trailer can e lim-
inate the destabilizing yawing moments due to sidewinds and generate
counteryaw in the opposite direction if needed. These tunnel tests have led to
development of a full-scale test vehicle and on-road test program of a PHV
test rig (Fig. 54), now ongoing for DOE. More data on this pneumatic ground-
vehicle program are found in another paper by Englar in this volume as well
as in NASA CP 2005-213509,2005.
VII. Conclusions
A. Capabilities
The high flow-entrainment capability of tangential blowing over curved aero-
dynamic surfaces has been show n in the preceding discussions to yield augmen-
tation and control of virtually all aerodynamic/hydrodynamic forces and
moments by simplified means, which frequently require no moving external com-
ponents. The capabilities of the CC devices demonstrated include the following:
1) Two-dimensional lift coefficients as high as 20 without moving parts and
similar high Cl for download as desired in automotive applications. This extra
high lift can also provide aircraft Super-STOL capability or the downsizing of
wing area for more efficient cruise.
2) Lift augmentations
ACl/C,
of
80
and very effective boundary layer
control.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 76/637
6
R.
J.
ENGLAR
3)
Drag reduction due to flow reattachment and thrust recovery, or drag
increase due to flow turning and lift-induced drag, and the ability to pneumati-
cally activate these as needed by the pilot or driver-this is particularly appli-
cable in au tomotive usage.
4) Aerodynam ic mom ent increases from blowing o r differential blow ing to
provide large control increases compared with those of mechanical devices, or
to allow control surface downsizing.
5) Pneumatic engine thrust deflection to 165 deg or more without moving
surfaces.
6) Pneumatic propellers or rotor blades to achieve variable thrust and control
moment without mechanical cyclic pitch.
7) Automotive applications to vary all forces and moments, including racing
vehicle download and drag, without moving parts, using only onboard air sources
such as turbochargers. Also , a low-drag aerodynamic heat exchanger using pneu-
matic-generated pressure difference can cool the vehicle while controlling aero-
dynamic forces and moments.
B. Future of Circulation Control
The preceding capabilities offer the potential
for aerodynamiclhydrodynamic
vehicles simplified by pneumatic multipurpose sui aces synergistically augm ent-
ing l i f , drag, moments, control, stability, and propulsive functions without any
moving mechanical parts.
The force augmentation capability also offers the
potential for reduction in wing and control surface areas for improved cruise per-
formance, or multipoint designs with lift/control surfaces sized for optimal
points of operation. Future investigations could include improved pulsed
blowing to even further reduce the required input mass flows, or to simplify
the operation of complex devices such as higher harmonic rotors. Application
of CC pneumatics to automotive and hydrodynamic vehicles offers the use of
aerodynamic surfaces for functions not currently employed, such as aerodynam ic
drag reduction or increase, download, heat exchange, thrust augmentation, and
stability and control. The opportunity to incorporate all of these devices into a
synergistic blown vehicle from the initiation of the design, rather than as an
add-on, offers the potential for a very effective and efficient multipurpose
vehicle, in which the pneumatic effectiveness, including the propulsion system
air supply source and the control systems, is incorporated from the very begin-
ning. A perfect example of how CC could be applied to a new and unique
Super-STOL vehicle would be its application to the new NASA Extreme-
STOL concept aircraft, where desired goals include C of 10, balanced field
lengths of
2000
ft or less and, of course, the necessity to trim and control this
vehicle at very low speeds, plus the ability to interchange drag increase and
drag elimination between approach and takeoff operations, respectively.
References
‘“The Use of Slots for Increasing the Lift of Airplane Wings,” NACA Translation, PW
’“Wings with Nozzle Shaped Slots,’’ NACA Translation, T M 521, July 1929; Berichte
635, Aug. 1931 (Proceedings L’Aeronautique, June 1931).
Der Aerodynamischen Vereuchsenstalt in Wie n,
Vol. 1, No. 1, 1928.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 77/637
OVERVIEW OF CC PNEUMATIC AERODYNAMICS
65
3Metral, A. R., “On the Phenomenon of Fluid Veins and their Application, the Coanda
Effect,”
A F
Translation, F-TS-786-RE, 1939.
4Sproule, R. S., and Robinson, S. T., “Combined Intelligence Objective Sub-
Committee Report,”
WF
Document Library Item
5
File No. IX-1, X-2, XII-1,
D52.420127, 1944.
’Voedisch, A., Jr., “Analytical Investigation of the Coanda Effect (Project No. FP-
188),” Air Material Command, Wright Field, Dayton, OH, Rept. F TR-2155-ND, April
1947.
6Cheeseman, I. C., and Reed, A. R., “The Application of Circulation Control by
Blowing to Helicopter Rotors,”
Journal of Royal Aeronautical Society, Vol. 71, No.
848, 1966.
’Cheeseman, I. C., “Circulation Control and Its Application to Stopped Rotor Aircraft,”
AIAA Paper 67-747, Oct. 1967.
8Lockwood, V. E., “Lift Generation on a Circular Cylinder by Tangential Blowing from
Surface Slots,” NASA Langley Research Center, Technical Note D-244, May 1960.
Englar, R. J., “Circulation Control Pneumatic Aerodynamics: Blown Force and
Mom ent Augm entation and Modification; Past, Present and Future,” AIAA Paper 2000-
2541, presented at AIAA Fluids 2000 Meeting, June 2000.
“Englar, R. J., and Williams, R. M., “Test Techniques for High Lift Two-D imensional
Airfoils w ith Boundary Layer and Circulation Control for Application to R otary Wing A ir-
craft,” Canadian Aeronautics and Space Journal, Vol. 19, No. 3, 1973 pp. 93-108.
Englar, R. J., “Two-Dimensional Subsonic Wind Tunnel Tests on a Cambered 30-
Percent-Thick Circulation Control Airfoil,” NSRDC, Technical Note AL-201, AD 913-
41 lL , May 1972.
Englar, R. J., and Applegate, C. A., “Circulation Control-A Bibliography of
DTNSRDC Research and Selected Outside References (Jan. 1969 through Dec. 1983),”
DTNSRDC-84/052, Sept. 1984.
13Wilkerson,J. B., Barnes, D. R., and Bill, R. A., “The Circulation Control Rotor Flight
Demonstrator Test Program,” A merican Helicopter Society, Paper AHS 79-5 1, May 1979.
14Mayfield, J., “Aeronautical Engineering-Navy Sponsors Coanda R otor Program,”
Aviation Week and Space Technology,
31 March 1980, pp. 69-74.
”Wilkerson, J. B., Reader, K. R., and Linck, D. W., “The Application of Circulation
Control Aerodynamics to a Helicopter Rotor Model,” American Helicopter Society,
Paper AHS-704, May 1973.
16Nielson,J. N. (ed.), “Proceedings of the Circulation Control Workshop , 1986,” NASA
Ames Research Center, NASA CP-2432, Feb. 1986.
”Shrewsbury
,
G., “Numerical Evaluation of Circulation Control Airfoil Performance
Using Navier-Stokes Methods,” AIAA Paper 86-0286, Jan. 1986.
18Novak, C. J., and Cornelius,
K.
C., “An LDV Investigation of a Circulation Control
Airfoil Flowfield,” AIAA Paper 86-0503, Jan. 1986.
Englar, R. J., “Experimental Investigation of the High Velocity Coanda Wall Jet
Applied to Bluff Trailing Edge Circulation Control Airfoils,” DTNSRDC, Report 4708,
Aero Report 1213, AD-A-019-417, Sept. 1975; also M.S. Thesis, Dept. of Aerospace
Engineering, Univ. of Maryland, College Park, MD, June 1973.
”Wood, N., and Nielson, J., “Circulation Con trol Airfoils Past, Present, and Future,”
AIAA Paper 85-0204, Jan. 1985.
”Williams, R. N., Leitner, R. T., and Rogers, E. O., “X-Wing: A New Concept in Rotary
VTOL,” presented at AHA Symposium on Rotor Technology, Aug. 1976.
9
11
12
19
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 78/637
66
R.
J.
ENGLAR
”Reader,
K.
R., and Wilkerson, J. B., “Circulation Control Applied to a High Speed
Helicopter Rotor,” DTNSRDC, Rept. 77-0024, Feb. 1977.
23Rogers, E. O., Schw artz, A. W., and Abram son, J. S., “Applied Aerodynamics of
Circulation Control Airfoils and Rotors,” presented at 41 st Annual
AHS
Forum, May 1985.
240ttensoser, J., “Two-Dimensional Subsonic Evaluations of a 15-Percent Thick Circu-
lation Control Airfoil with Slots at Both Leading and Trailing Edges,” NSRDC, Rept.
4456, July 1974.
25Williams, R. M ., and Cheeseman,
I.
C., “Potential Acoustic Benefits of Circulation
Control Rotors,” presented at AHS Meeting
on
Rotor Acoustics, NASA Langley Research
Center, May 1978.
26Englar, R. J., “Investigation into and Application of the High Velocity Circulation
Control Wall Jet for High Lift and Drag Generation on STOL Aircraft,” AIAA Paper
74-502, June 1974.
27Englar,R. J., “Subsonic Two-Dim ensional Wind Tunnel Investigations of the High Lift
Capability of C irculation Control Wing Sections,” DTSNRDC, Rept. ASED-274, April 1975.
”Englar, R. J., “Circulation Con trol for High Lift and Drag Generation
on
STOL Air-
craft,”
A I M Journal
of
Aircraft,
Vol. 12, No.
5 ,
1975, pp. 457-463.
29Englar,R. J., Trobaugh, L. A., and Hemmerly, R. A., “Development of the Circulation
Control Wing to Provide S TO L Potential for High Performance Aircraft,” AIAA Paper 77-
578, June 1977.
30Englar, R. J., “Circulation Control Technology for Powered-Lift STOL Aircraft,”
Lockheed Horizons,
No. 24, Sept. 1987.
31Englar, R. J., Hemmerly, R. A., Moore, H., Seredinsky, V., Valckenaere, W. G.,nd
Jackson, J. A., “Design of the Circulation Control Wing STOL Demonstrator Aircraft,”
AIAA Paper 79-1842, Aug. 1979; also published in
Journal of Aircruft,
Vol. 18, No. 1,
32Englar,R. J., “Developm ent of the A-6/Circulation Control Wing Flight Demonstrator
Configuration,” DTNSRDC, Rept. ASED-79/01, Jan. 1979.
33Mayfield, J., “Circulation Control Wing Demonstrates Greater Lift,”
Aviation Week
and Space Technology,
March 19, 1979.
34Pugliese,A. J., and Englar, R. J., “Flight Testing the Circulation Control Wing,” AIAA
Paper 79-1791, Aug. 1979.
35Loth, J. L., Fanucci, J. D., and Roberts, S. C., “Flight Performance of a Circulation
Control STOL Aircraft,” AIAA Paper 74-994, April 1974; also published in
Journal of
Aircruft,
Vol. 13, No. 3, 1976, pp. 169-173.
36Roberts,
S.
C., “West Virginia University Circulation Control STOL Aircraft Flight
Test,” WVU Aerospace, Technical Rept. No. 42, July 1974.
Englar, R. J., “Low-Speed Aerodynamic Characteristics of a Small Fixed-Trailing-
Edge Circulation Control Wing Configuration Fitted to a Supercritical Airfoil,”
DTNSRDC, Rept. ASED-81/08, March 1981.
38Englar,R. J., and Huson,
G. .
Development of Advanced Circulation Control Using
High-Lift Airfoils,” AIAA Paper 83-1847 July 1983 ; also published in
Journal ofAircraft,
39Carr, J. E.,
“An
Aerodynamic Comparison of Blown and Mechanical High Lift Air-
foils,” AIAA Paper 84-2199, Aug. 1984.
40Englar, R. J., “Two-D imensional Transonic Wind Tunnel Tests of Three 15-Percent-
Thick Circulation Control Airfoils,” NSRDC, Technical Note AL-182, AD 882-075,
Dec. 1970.
1981, pp. 51-58.
37
V O ~ .1, NO. 7, 1984, pp. 476-483.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 79/637
OVERVIEW OF CC PNEUMATIC AERODYNAMICS
67
41Englar, R. J., Smith, M. J., Kelley, S. M., and Rover 111, R. C., “Development of
Circulation Control Technology for Application to Advanced Subsonic Transport
Aircraft,” AIAA Paper 93-0644, Jan. 1993; also published in Journal of Aircraft,
Vol.
42Englar, R. J., Niebur, C. S., and Gregory, S. D., “Pneumatic Lift and Control Surface
Technology Applied to High Speed Civil Transport Configurations,” AIAA Paper 97-
0036, Jan. 1997.
43Mavris, D. N., K irby, M. R., Lee, J. M., Q ui,
S.,
Roth, B., Tai, J., and Englar, R. J.,
“Systems Analyses of Pneumatic Technology for High Speed Civil Transport Aircraft,”
GTR I Final Technical Rept. A-5676, Oct. 1999.
44Ni~ho l s ,. H., Jr., Englar, R. J., Ha m s, M. J., and Huson,
G. .
Experimental Devel-
opment of an Advanced Circulation Control Wing System for Navy STOL Aircraft,”
AIAA Paper 81-0151, Jan. 1981.
45Harris, M. H ., Nichols, Jr., J. H., Englar, R. J., and Huson,
G. .
Development of the
Circulation Control WingIUpper Surface Blowing Powered-Lift System for STOL Air-
craft,” Proceedings
of
the ICASIAIAA Aircraft Systems and Technology Conference,
Paper
ICAS-82-6.5.1,
Aug. 1982.
46Yang, H. T., and Nichols, Jr., J. H., “Design Integration of CC W /U SB for a Sea-Based
Aircraft,” Paper ICAS-82-1.6.1, Aug. 1982.
47Englar, R. J., Nichols, Jr., Harris, J. H., Eppel J. C., and Shovlin, M. D. “C irculation
Control Technology Applied to Propulsive High Lift Systems,” Society of Automotive
Engineers, Paper 841497, Oct. 1984.
48Lowndes, J. C., “Aeronautical Engineering: Stud ies Show Lift Coefficient Tripling,”
Aviation Week and Space Technology, Dec. 1, 1980.
49Englar, R. J., Nichols, Jr., J. H., Ham s, M. J., Eppel, J. C., and Shovlin, M. D., “Devel-
opment of Pneumatic Thrust-Deflecting Powered-Lift Systems,” AIAA Paper 86-0476,
Jan. 1986.
”Eppel, J. C. , Shovlin, M. D., Jaynes, D. N., Englar, R. J., and Nichols, Jr., J. H., “Static
Investigation of the CCW/USB Concept Applied to the Quiet Short-Haul Research Air-
craft,” NASA , TM 84232, July 1982.
51Shovlin, M. D., Englar, R . J., Eppel, J. C., and Nichols, Jr., J. H., “Large-Scale-Static
Investigation of Circulation-Control-Wing Concepts Applied to Upper-Surface-Blowing
Aircraft,” NA SA, Technical Paper 2684, Jan. 1987.
52Bevilaqua,P. M., and Lee, J. D., “Design of Supersonic Coanda Jet Nozzles,” in Proceed-
ings of the Circulation Control Workshops 1986, NASA, CP 2432, pp. 289-312, Feb. 1986.
53Englar, R. J., and Campbell, B. A., “Developm ent of Pneum atic Channel W ing
Powered-Lift Advanced Super-STOL Aircraft,” AIAA Paper 2002-2929; presented at
AIAA 20th Applied Aerodynamics Conference, June 25, 2002.
54Englar, R. J. and Campbell, B . A., “Experimental Developm ent and Evaluation of
Pneumatic Powered-Lift Super-STOL Aircraft,” NASAIONR Circulation Control Work-
shop, March 2004; also published in NA SA C P 2005-213509, 2005.
55Braslow, A. L., “Aerodynam ic Evaluation of Circulation Control Propellers,” Bio-
netics Corp., NASA Contractor Report 165748, June 1981.
56Taback, I., Braslow, A. L., and Butterfield, A. J., “Circulation Control Propellers for
General Aviation, Including a BASIC Computer Program,” NASA Contractor Rept.
165968, April 1983.
57Gam er, D., “No Moving Parts, The Circulation Control Airfoil and Fluidic Propeller,”
EAA
Sport Aviation,
Vol.
37, No. 3, 1988, pp. 27-30.
31, NO.
5
1994, pp. 1160-1177.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 80/637
68
R.
J.
ENGLAR
58Wilson, M. B., and von Kerczek, C .,
“An
Inventory of Some Force Producers for Use
in Marine Vehicle Control,” DTNSRDC-79/097, Nov. 1979.
59Wellman, L. K., and Jacobsen, C., “Wind Tunnel Investigation of the Application of
Circulation Control to a Forward Swept Wing,” DTNSRDC /ASED-82/05, June 1982.
Englar, R. J., “Pneumatic High-Lift and Control Surfaces Applied to Micro-Aerial
Vehicles,”
Proceedings
of
GTRI International Conference on Emerging Technologies
f o r
Micro
Air
Vehicles,
Feb. 1997.
610 yle r, T. E., and Palm er, W. E., “Exploratory Investigation of Pulsed Blowing fo r
Boundary Layer Control,” North American Rockwell, Rept. NR72H-12, Jan. 1972.
62Walters, R. E., et al. “Circulation Control by Steady and Pulsed Blowing for a
Cambered Elliptic Airfoil,” West Virginia Univ., Dept. of Aerospace Engineering, Rept.
63Jones,
G.
S.,
and Englar, R. J., “Advances in Pneumatic-Controlled High-Lift Systems
Through Pulsed Blowing,” AIAA Paper 2003-3411; presented at AIAA 21st Applied
Aerodynamics Conference, June 2003.
64Englar, R. J., and W illiams, R. M ., “Design of a Circulation Control Stem Plane for
Submarine Applications,” NSRDC, Technical N ote AL-200, M arch 197 1.
6 5 F ~ ry , . J., and Whitehead, R. E., “Static Evaluation of a Circulation Control Centrifu-
gal Fan,” DTNSRDC, Rept. 77-0051, AD A041-463, June 1977.
aeta, R. J., and Englar, R. J., “Pneumatically Augm ented Aerodynamic Heat Exchan-
ger,” Paper presented at NASA/ONR Circulation Control Workshop, March 2004; also
published in NASA C P 2005-213509, 2005.
67Englar, R. J., Smith, M. J., N iebur, C. S., and Gregory, S. D., “Development of Pneu-
matic Aerodynamic Concepts for Control of Lift, Drag, Moments and Lateral/Directional
Stability of Automotive Vehicles,” Society of Automotive Engineers, Paper 960673, Feb.
1996; also published in S AE SP-1145, “Vehicle A erodynamics,” Feb. 1996.
68EnglarR. J., “Drag Reduction, Safety Enhancement and Performance Improvement for
Heavy Vehicles and SUV s Using Advanced Pneum atic Aerodynamic Technology,” 2003
SA E International Truck and Bus Meeting and Exhibition, Society of Automotive Engin-
eers, Paper 2003-01-3378, Nov. 2003.
Englar, R. J., “The Application of Pneumatic Aerodynamic Technology to Improve
Performance and Control of Advanced Automotive Vehicles,” NASA/ONR Circulation
Control Workshop, March 2004; also published in NASA CP 2005-213509, 2005.
60
TR-32, July 1972.
69
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 81/637
Chapter
3
Exploratory Investigations of Circulation Control
Technology: Overview for Period
1987-2003
at NSWCCD
Robin Imber*
Naval Air Systems Command, Patuxent River, Maryland
and
Ernest Rogerst and Jane Abramsont
Naval Sur ace War are Center-Carderock Division, West Bethesda, Maryland
Nomenclature
A
= area of
CC
slot, or area of foil planform
A R =
aspect ratio
C , = momentum coefficient of slot flow ( r i zv j /qS)
C L
=
lift coefficient ( L / q S )
C , =
drag coefficient
( D / q S )
C , = power coefficient
CT= thrust coefficient
c
=
chord length
D = drag force
d
= diameter, or camber line offset
h =
CC slot exit height (gap)
L = lift force
rit
=
mass flow ( p A V )
PR = pressure ratio
q
=
dynamic pressure
(&pV2)
S =
planform area of lifting surface
t = thickness of airfoil
*Aerospace Engineer.
'Aerospace Engineer, retired.
This material is declared a work
of
the
U.S.
overnment and is not subject to copyright protection
in the United States.
69
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 82/637
7
R.
IMBER, E. ROGERS, AND
J.
ABRAMSON
V = velocity
a =
angle of attack
p
=
density
u
=
rotor solidity, cavitation index
ACL/AC, = lift augmentation ratio (slope of
CL
vs C, curve)
Subscripts
=
induced
= j e t
I. Introduction
EGINNING in 1967, when the Naval Surface Warfare Center, Carderock
B
Division (NSWCCD) was known as the David Taylor Model Basin
(DTMB), researchers there were involved with many circulation control (CC)
exploratory projects, including C C-airfoils, CC-centrifugal fans, dual-directional
CC-airfoils, CC-fixed wings, including the A6 aircraft modification, CC-rotor-
craft, including XH2-CCR and X-wing, CC-hydrodynamic applications, and
valving systems for CC.
The first Circulation Control Workshop (unpublished) was held at DTMB in
1971, and the second was held at the National Aeronautics and Space Adminis-
tration, Ames Research Center, in 1986.' Papers, presentations, and reports of the
research performed from 1967 to 1985 at what is now NSW CCD (there have been
several name changes since 1967) are cited in the proceedings from the 1986 CC
Workshop and more can be found in Ref. 2.
This overview is intended as a brief summary of the highlights of six of the
major CC experimental investigations that have taken place at NSWCCD since
the second CC Workshop, specifically between 1987 and 2003, and was pre-
sented at the 2004 Circulation Control Workshop held in Hampton, V irginia.3
The following investigations are discussed: 1) The Dual-Slotted Cambered
Airfoil, LSB;
2)
The Self-Driven Rotary Thruster, TIPJET; 3) The Annular
Wing, CC-Duct; 4) The Circular Wing, CC-Disc;
5 )
The Miniature Oscillatory
Valve, CC-Valve; and 6) The Dual-Slotted Low Aspect Ratio Wing, CC-
Hydrofoil. For further details regarding these investigations, the reader is encour-
aged to examine the publications that are referenced for each of the projects.
Used throughout this summary is a frequently used measure of CC perform-
ance, the lift augmentation ratio. This ratio is defined as the ratio of the gain in
lift
A C L )
o the change in slot flow momentum (AC,). In this review, the ratio
is determined from experimental data by assessing the slope of the lift response
in the low blowing (low
C
range, where the response is usually linear.
11.
Dual-Slotted Cam bered Airfoil (LSB)
The Dual-Slotted Cam bered Airfoil, also referred to as the LSB (lower surface
blowing), was designed and tested in 1987 by Abramson and colleague^.^ ^ The
inclusion of a lower su rface slot along with the usual upper surface slot provides
the ability to produce lift in both positive and negative d irections. The presence of
camber in this model, along with the objective of preserving the contour of a
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 83/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD 71
Fig. 1 Photograph of the dual-slotted cambered airfoil (LSB).
proven (parent) single-slot CC airfoil, means that the geometric properties of the
lower-slot region are not the sam e as those of the upper-slot region. Additionally,
when operating in the lower surface blowing mode, the
CC
section functions with
negative camber, all with unexplored consequences at that time.
A photograph of the LSB m odel is shown in Fig. 1 and a cross-section draw ing
of the model is shown in Fig.
2.
The LSB has a 17% thickness ratio with 1.1%
camber. It was constructed with a 12-in. chord and a 36-in. span. The upper slot
is located at 96.8% chord and the lower slot is slightly further aft at 97.0% chord.
The airfoil was experimentally evaluated in the
NSWCCD
8 x 10 ft wind
tunnel configured with two-dimensional wall inserts. Testing included three
blowing modes: upper surface only, lower surface only, and dual blowing. The
wind tunnel dynamic pressure ranged from 20 to 60psf, Reynolds number
ranged from
0.8
to 1.4 x lo6, and angle of attack (AOA) ranged from 0 to
+10 deg. Two slot height-to-chord (h/c) ratios were set: 0.0013 and 0.0020.
The maximum momentum coefficient
C)
was
0.22.
One of the main design goals was to have the dual-slotted model perform as
well, when using only the upper slot, as the single slotted “parent” model. The
dual-slotted model had the sam e cross-section as the parent m odel. Lift perform-
ance results from the single and dual slot models are shown in Fig. 3. The com-
parison shows that there was no detrimental effect in adding the second slot.
The second design objective was to increase the control range so that force
control in both d irections was available. F igure 4 displays a plot of lift coefficient
Air Supply Ducts
Fig.
2
Cross-section of LSB airfoil
Upper Surface
Blowing Slot
Coanda
Surface
Lower Surface
Blowing (LSB)
Slot
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 84/637
72
R. IMBER, E. ROGERS, AND J. ABRAMSON
ACr
due to
blowing
Momentum Coefficient, Cp
Fig. 3 Comparison of lift performance for dual-slottedLSB and single-slot parent
airfoil?
against momentum coefficient, and reveals that the goal of doubling the control
range was met.
An unanticipated finding was that the performance of the lower slot, in terms
of measured lift augmentation ratio, was noticeably better than that for the upper
slot (80 comp ared with 60). This empirically unexpected performance enhance-
ment illustrates the need for a well-validated computational code (computational
fluid dynamics; CFD) to help guide future CC designs. Another finding was the
effect of simultaneous blowing5 At the only ratio of dual blowing examined,
Lift
Coefficient
Momentum Coefficient
Fig. 4 Control range increase demonstrated with upper and lower slot ~ a p a b il it y .~
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 85/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD
73
where the low er slot flow momentum level was 25 that of the upper slot, acti-
vation of the lower slot decreased lift. This dual-slotted airfoil helped pave the
way for another dual-slotted CC investigation discussed later, the CC -Hydrofoil.
Summarizing the key findings from the Dual-Slotted Cambered A irfoil inves-
tigation, it was found that 1) incorpora tion of a lower slot did not affect perform-
ance of the upper slot; 2) the available lift control range was doubled, as
expected; 3) a lift augmentation ratio of
80
for the lower slot was obtained;
and 4) simultaneous blowing can be used to decrease (control) the lift increment
produced by single-slot operation.
111. Self-Driven Rotary Thruster (TIPJET)
Experimentally investigated in 1991 by several NSW CCD engineers, the Self-
Driven Rotary Thruster was the first integrated lift/reaction-drive rotor system
combining Cbanda CC aerodynamics with cold cycle reaction drive technol-
ogies.6-8
The rotor was developed as part of the TIPJET unmanned air vehicle, shown
conceptually in Fig. 5 . The design involves a stoppable two-bladed rotor concept
where, after lifting off vertically in rotary mode and accelerating forward, the
rotor transitions to a fixed wing to enable high-speed flight. A “cold cycle” gas
generator, such as the fan stage of a turbofan engine, supplies the compressed
air for both the circulation control and the tip jets that provide rotor drive torque.
Figure 6 shows a sketch of the com pletely pneum atic rotor. Circulation control
slots are located along most of the rotor blade span on both the leading and trail-
ing edges. Reaction drive nozzles are located at the rotor tips. Thus, a single
source of air pressure provides flow for the CC slots to augment rotor lift (vertical
Fig. 5
TIPJET
vertical takeoff and landing unm anned air veh icle with circulation
controlled stoppable rotor.6
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 86/637
74
R. IMBER, E. ROGERS, AND J. ABRAMSON
No drive shaft, unarticulated, flat-pitch blades
NOZZLE
AIR
SUPPLY
DRIVE
FORCE
Fig.
6 Sketch of TIPJET ompletely pneumatic rotor?
thrust), while at the same time providing the rotor torque drive via the tip-jet
nozzles. (The full-scale application concept called for in-flight controllable slot
gap settings.)
A detailed investigation of the TIPJET rotor in hover took p lace in 1991.7 The
primary objective of the hover experiment was to evaluate the interactions
between the lift and drive systems. Drawings of the aluminum rotor model are
shown in Fig.
7
and specifications of the model are listed in Table 1. The 80-
in. rotor blade is tapered, with no twist and zero pitch angle. The thickness and
camber varies linearly with radius from the 25% to 95% span locations.
Figure
8
is a photograph of the blade tip region, showing the CC slot along the
span and the tip-drive nozzle. During the hover test, the rotor could be driven
by either an electric drive motor that enabled the rotor to be operated at selected
rpm settings while investigating specific performance attributes, or by the tip-jet
reaction drive.
To better understand the performance of the integrated lift/drive system, a
detailed investigation was conducted with the tip nozzles closed and the rotor
mechanically driven. Figure 9 shows the experimental data for the measured
rotor thrust coefficient as a function of momentum coefficient for several slot
height settings. The slope of the curve is 29 at the lower values of CJu. This
ratio is higher than that of any previously tested
CC
rotor.
As
shown in Fig. 9,
this measure of efficiency was independent of the four slot heights tested.
To determine if the level of understanding of the performance of the fully
pneumatic rotor system was sufficient for successful incorporation into a flight
vehicle, numerical calculations were developed and compared to the experimen-
tal performance. The results of this comparison, show n in Fig. 10, indicate excel-
lent correlation for both the rotor thrust developed and drive power required.
The ultimate goal of the experimental roto r investigation was to determine the
aeromechanics of the model rotor in self-drive mode. The behavior of the
rotational speed in response to pressure input was unknown at the beginning of
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 87/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD
75
SLOT HEIGHT ADJUSTMENT
SIDE-B Y-SIDE OPPOSING
SCREWS
Fig. 7 Draw ings with d etails of
TIPJET
pneum atic rotor?
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 88/637
76
R.
IMBER, E. ROGERS, AND
J.
ABRAMSON
Table
1
TIPJET rotor specifications’
Blade
Fixed pitch angle
OJ,
deg
Rotor diameter, ft
Number of blades
Chord, in.
25 span
Solidity ratio
Geometric twist, deg
93 span
Airfoils
0
6.67
2
7.95
5.40
0.110
0
25 span 93 span
Thickness ratio
( t / c )
0.213 0.170
Trailing edge radius (rt, /c)
0.05 0.03
Slot height
( h / c )
variable variable
Area/nozzle, in.* 0.764
Camber ratio d / c ) 0.053 0.01 1
Tipjet nozzles (rectangular)
the test. When a slot height and tip nozzle area are set, the blade pressure input is
the only determining factor of the operating condition.
Shown schematically in Fig. 11, as pressurized air is introduced into the
rotor, the rotor rotates in reaction to the flow from the tip nozzles. As the
nozzle flow increases, the rotor lift will increase as a result of using a cambered
airfoil and, most especially, as a result of the increased circulation due to the CC
slots. This additional lift increases the required drive torque, which then limits
the rotational rate for a given air pressure setting. In essence, pressure input
simultaneously influences the lift and produces the torque drive. Identifying
the behavior of a system coupled by these two effects was on e of the ob jectives
of the experiment.
Fig.
8
Tip region of the fully pneumatic rotor model.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 89/637
INVESTIGATIONS
OF
CC TECHNOLOGY AT NSWCCD
77
c
f
SLOT
MOMENTUM
COEFF.
/
SOLIDITY, C ~ / O
Fig.
9
TIPJET rotor thrust performance when mechanically driven at constant
rpm.’
It was discovered that, in full self-drive mode, the rotational speed is stable
and exhibits a self-limiting maximum for a given ratio of slot area to nozzle
area. The data in Fig. 12 reveal the nature of the self-limiting rotational speed.
Rotational tip speed is shown as a function of blade pressure for several slot
height settings. At each of the slot settings, the pressure input response of the
rotor is to increase the rotational rate until a limiting tip speed is reached. It
was demonstrated that the value of the limiting tip speed is a function of the
slot height setting; increasing the slot height results in a lower limiting tip
speed. However, because CC-based lift is not solely dependent on local velocity,
the lift response is not limited to a tip speed maximum. Figure 13 shows that the
rotor lift is essentially always the same linear function of the applied pressure,
independent of actual rotational rate (compare Figs. 12 and 13).
A major finding was that a non-shaft-driven completely pneumatic rotor
inherently seeks a rotational rate that results in lift being a near-linear function
of the blade pressure input, and this linear lift is easily controllable by pressure
throttling. In addition to the specific TIPJET vehicle application, this
capability can be applied to systems that require m echanically simple, easily con-
trolled thrusters.
Summarizing some of the key findings from the Self-Driven Rotary Thruster
investigation, it was found that 1) a lift augmentation ratio of 29 was obtained
when in the mechanically driven mode;
2)
the pneumatic rotor inherently seeks
equilibrium and the self-limiting rotational rate is a function of slot-to-drive-
nozzle area ratio (the resulting rotor lift is a near linear function of the blade
pressure); 3) there is a significant impact on induced power efficiency because
of the non-lifting tip nozzle region; and 4) the presence of the tip nozzle jet
has no discernible impact on the external aerodynamics of the lift system.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 90/637
78
R. IMBER, E. ROGERS, AND J. ABRAMSON
D
c
Fig. 1 TIPJET performance numerical analysis correlation with experimental
results.'
cc
slot Slot
flow
area
t
I
Internal air
r
otor
Drive
rotor
drive
torque
Fig. 11 Conceptual schematic of TIPJET rotational rate equilibrium.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 91/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD
79
a,
a,
v
Blade
Root
Pressure
Ratio, PR oo,
Fig. 12 TIF'JET rotor characteristics when self-driven via tip-jet nozzles?
IV. Annular Wing (CC-Duct)
In the mid-1990s the performance characteristics of an annular wing, having
both inner and outer trailing edge circulation control slots, were explored.
The focus
of
this investigation was to apply full, or partial, perimeter trailing
Blade Root Pressure Ratio, PR,,,,
Fig. 13 Relationship of blade pressure to thrust developed.' (Data and symbols are
the same as in Fig. 12.)
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 92/637
80
R.
IMBER, E. ROGERS, AND
J.
ABRAMSON
Fig.
14
Annular wing m odel installed
in free jet wind tunnel?
edge CC fluid ejection to a propulsor duct to enhance maneuvering control on
watercraft via a thrust vectoring capability. An existing CC-Duct model with
inner and outer trailing edge CC slots had been borrowed from West Virginia
University. The model has been used in the 1970s to investigate the attributes
of variable diffusion for ducted fans on aircraft.’ The CC-Duct model is shown
in Fig. 14 as it was installed in the Atlantic Applied Research Corporation
open-jet acoustic tunnel in 1993. The model originally had a motor housing
and stator that were removed for the CC-Duct investigation discussed here.
The objective for removing the motor housing and stator was to have a simple
configuration in which to establish an understanding of the performance
attributes, and to provide a data set for correlation to basic ring-wing theory.
No propeller was present in any of the test series.
A cross-section of the top of the C C-Duct is shown in Fig. 15 and its geometry
is presented in Tab le
2.
The model is 18411. in d iam eter with a 10-in. chord and a
20 thick uncambered foil section. The inner and outer slots are located around
the full trailing edge circumference at 97% chord.
During the Duct investigation, metal foil tape was applied and burnished well
to the external surface of the model, over the slot region, in order to temporarily
block
off
portions of the slot. This commonly used technique for exploratory CC
research provides the ability to control the distribution of slot flow. By using the
tape, several configurations, focusing on potential attributes of the Duct, were
tested.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 93/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD
81
Dimensions n inches
Fig.
15
Annular wing m odel cross-section (at top).’
The model and experimental arrangement provided the opportunity to
examine many interesting flow control configurations. Icons representing six
basic configurations, or modes of operation, and a brief explanation of each,
are displayed in Table
3.
The dotted line on each of the icons represents the trail-
ing ed ge of the CC-Duct. The solid lines represent the portion of either the inner
or outer slot that is open and where the fluid ejection occurs. The author suggests
reviewing Table 3 before reading further.
As
a simple illustration of force vectoring capability, a long strand of yam was
positioned in the center of the CC-Duct. The photograph on the left of Fig. 16
shows the configuration with no CC blowing. The yam is aligned with the
tunnel free stream velocity, along the longitudinal axis of the Duct. The photo-
graph on the right of Fig. 16was taken fo r a com plimentary halves configuration,
where the ou ter lower slot is active and the upper inner slot is active. The ya m
strand is now at an angle to the free stream, indicating the w ake deflection due
to the force vectoring brought about by the active flow control.
Quantitative data from the configuration on the right in Fig.
16
are shown in
Fig. 17 as a plot of force developed as a function of C The reference area used
Table 2 Annular wing m odel specifications
Model geometry Dimensions, in.
Outside diameter 18.2
Inside diameter 14.2
Chord
10
Slot gap
0.009
h l c
0.0009
Slot position 0.970c
d/c (16.2110) 1.62
AR
(effective) 2.1
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 94/637
82
R.
IMBER, E. ROGERS, AND
J.
ABRAMSON
Table
3
CC annular wing modes of operation as a m arine propulsor duct
Aft view of Slot ejection
ducta configuration Effect Operation benefit
Inner slot only Increased duct Higher prop efficiency
flow-through: accelerating
nozzle
Outer slot only Decreased duct flow: Reduced cavitation
diffusion
Complementary Side-force: yaw Steerage
quadrants
Complementary Side-force: pitch Depth keeping
quadrants
Alternating Vortex generation:
very
high Braking: crash-back
drag,
no
side force
Both slots Drag reduction, auxiliary Cru ise efficiency, dock
thruster side positioning
aDashed line is trailing edge; solid lines represent active slot.
for the force coefficients is duc t length times duct diameter (c x d ) . The lift force
developed for this configuration, even at zero model pitch angle, is more than
twice that available on a passive duct.
When interpreting the CC performance of ring wings, in comparison to that
of flat wings or airfoils, it is important to be observant of how the performance
parameters are nondimensionalized. In Fig. 17, note that the
CL
versus
C
performance curve for the duct is close to that of some CC airfoils (e.g., see
Fig.
4).
This result occurs even though the finite wing effect of shed wake
vorticity causes a downwash that should decrease the net lift response to
roughly half that of a two-dimensional CC foil. The explanation for the
apparent two-dimensional-like performance level has to do with the reference
areas used for the coefficients. Consistent with the practice for CC airfoils and
wings, the duct
C
is based on the full slot length
T d
x c). At the same
time, as is standard practice, CL is defined based on projected area
(c
x d ) ,
without consideration that there are two lifting surface areas involved: the
upper and lower halves of the ring-wing. Therefore, the
C L
that should be
compared to that of an airfoil is half the
CL
shown in Fig. 17, and thus
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 95/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD
83
No Blowing 180 I 180 deg blowing
Fig. 16 Lateral force capability; wake deflection with asymm etric trailing edge
CC
blowing?
matches what would be expected from a wing having an as ect ratio (AR) of 2.1,
which is the equivalent
AR
or the geom etry of the duct. (See the later discus-
sion and performance of the CC-Hydrofoil flat wing, which has about the same
aspect ratio.)
Slot Flow Momentum Level
CF =
mV,/(
pVm2 dc)
Fig. 17 CC annular wing lift and drag performance' (the drag force is in the
direction of a negative propulsive force).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 96/637
84
R.
IMBER, E. ROGERS, AND
J.
ABRAMSON
Test results met expectations including that the drag (plotted as negative
thrust) is a linear function of
C ,
as shown in the experimental data, and
derived as in the following equations.
From the experimental data,
CL =
1 o G
From Hoerner and Borst
AR = 2.1 2)
and for lift-induced drag
then
One of the findings from the CC-Duct investigation was the ability for
braking, or control of induced drag, without a change in net lift. Alternating
the active inside and outside slots every 90deg creates two pairs of counter-
rotating vortices. Figure 1 8 shows the measured performance demonstrating
this capability. The drag is about the same as it was when lift was being devel-
oped. Figure 19 represents a
VSAERO
potential-flow solution of surface pressure
and wake filaments for a CC-Duct braking configuration. (A discussion of
Slot
Flow Momentum Coefficient (Cp)
Fig. 18 Braking configuration measured data; lift-induced drag without net lift?
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 97/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD
85
Wake Filaments looking upstream
Fig.
19
Braking configuration com putational solution."
potential flow solution techniques related to this work, and CC applications in
general, can be found in Ref. 11.)
The key findings from the CC-Duct investigations include 1) lift and side
force can be generated using specific blowing segments; 2) at zero angle of
attack (AOA), forces of almost 2.5 times a conventional ring-wing are
possible;
3)
a braking force via induced drag is available without the devel-
opment of lift; and 4) the performance met expectations and it was found that
the performance can be predicted using a potential flow code, although the
slot flow requirements have to be estimated from the historical CC airfoil
database.
V. Circular Wing (CC-Disc)
In 1995, Rogers and Imber created a circular wing model with a full perimeter
circulation control capability.
l 2
The purpose was to investigate the effectiveness
of CC on very low aspect ratio wings and to explore the attributes of a CC-
enhanced omnidirectional type of control surface or vehicle. The CC-Disc,
also known as the Coanda Disc, was tested in the
8
x loft Subsonic Wind
Tunnel at NSWCCD, employing a six-component external balance for force
and mom ent measurement. Figure 20 is a photograph of the anodized aluminum
2-ft-diam. model and Table 4 lists the model specifications. The Disc has a
19% thick cross-section with 2.4% camber. Figure 21 shows a drawing of the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 98/637
8
R.
IMBER, E. ROGERS, AND
J.
ABRAMSON
Fig. 2 Circular wing model with full perimeter circulation control.
centerline cross-section with surface pressure tap locations. The baseline slot
height was 0.032 in.
The major configurations investigated are shown in Fig. 22. The circular icons
represent a planform view of the Disc with the shaded sections representing the
perimeter sections where fluid ejection occurred. In Fig. 22, the free-stream flow
would be directed from the top to the bottom of the page. The three groups of
configurations were 1) increasing area centered about the trailing edge; 2) con-
stant area at variable azimuth; and 3) lateral and asymmetric variations. The
same slot tape-over technique used in the CC-Duct test was employed for flow
control configuration changes on the CC-Disc.
Resu lts from the first configuration group are show n in Fig. 23 for the model at
zero pitch angle, gradually increasing the flow ejection circumference region.
Lift is show n as a function of the region of blowing, starting with unblown and
then, centered about the trailing edge, increasing the perimeter region blown
until there was full 360-deg fluid ejection. The lines on the plot are for constant
Table 4 Specifications of circular wing
iameter (chord)
2
ft
Thickness ( t l c ) 19
Reference area
(S) 3.14
ft2
Aspect ratio AR) 1.27
Camber
2.4
Coanda radius
r s / c = 0.050
r t e / c= 0.040
Slot position
3.2
from edge
Slot lip thickness
0.026
in.
Slot height ( h ) 0.032 in.
0.0013
0.027
h l c
hlrs
Pressure tap diam eter 0.040 in.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 99/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD
87
Surface Pressure Taps
Centerline Cross-section
Fig. 21 Circular wing model with full perim eter circulation control.'*
blowing coefficient C The optimum, or highest lift, configuration varied
somewhat with the C, level. The overall highest lift was obtained using a
225-deg perimeter of fluid ejection centered about the trailing edge. It is
notable that high lift performance was obtained even with full perimeter
blowing, showing that the omnidirectional configuration is viable.
Figure 24 reveals lift performance results for the 225-deg perimeter configur-
ation at several
AOA
and C, levels. The m aximum C, was more than twice that
available fro m a disc without circulation control. Both w ith and without blowing,
the slope of C,/a closely matches the slope calculated from inviscid theory. The
same configuration and data collection series is shown in Fig. 25 plotted as lift
lncreasina Area Cen tered Abou t the Trailina edae:
Fig. 22 Circular wing test configurations: planform view (sections with blowing
shown in black).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 100/637
aa R.
IMBER, E. ROGERS, AND
J.
ABRAMSON
Configuration
Fig.
23
Lift as a function of azimuthal mass ejection coverage.'*
coefficient squared against drag coefficient. The measured induced drag matches
the prediction of lifting surface wing theory. These matches suggest that, for low
aspect ratio w ings, there are no basic effects unique to lift developm ent by m eans
of the Coanda form of CC.
Angle-of-Attack (deg)
Fig. 24 Influence of angle of attack?
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 101/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD
89
CD
Fig.
25
Aerodynamic efficiency: induced drag with incidence?
As
discussed in more detail in Ref. 12, the circular wing with CC has two
aerodynamic centers:
1)
lift due to angle of attack and
2 )
lift due to circulation
control. The pitching moment map in Fig. 26 demonstrates the results of
these two centers. The pitching moment was established about the center of
moment
Fig. 26 Pitching moment m ap resulting from two aerodynamic centers?
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 102/637
90
R.
IMBER, E. ROGERS, AND
J.
ABRAMSON
CP
Fig.
27
Sensitivity to slot gap setting?
the disc and the plot can be used to determine the combination of pitch angle
and level of circulation control that will provide a specific desired maneuvering
effect.
During part of the CC-Disc investigation, the slot height around the perimeter
was changed to determine the sensitivity of wing performance to slot gap setting.
For the range of
C,
shown in Fig.
27,
there was essentially no change in lift
performance for a 4:l slot gap change. The lack of performance change with
the large change in slot gap is considered a desirable attribute, because it
allows flexibility in the selection of a slot-flow supply system.
Fo r many of the configurations investigated, upper and lower surface pressures
were measured every 10 deg in azimuth. A sample of the upper surface pressure
distributions of four configurations is shown in Fig. 28. Th e area of fluid ejection
is easily seen as the darker shaded regions around the perimeter, which have sig-
nificant low pressure values and steep pressure gradients. The radial lines that
appear in Fig.
28
are a product of the plotting software and not a true character-
istic of the overall pressure distribution.
There were many key findings from the Circular Wing test. The investigation
demonstrated that
1
circulation control is effective on very low aspect ratio
lifting surfaces and, fo r a circular planform, can provide an om nidirectional capa-
bility when full perimeter blowing is applied;
2)
with at least
225
deg of flow
control around the perimeter the lift produced was more than double that of an
unblown circular wing (the limit to augmented lift is believed to be the result
of excessive wall je t turning);
3)
roll con trol was demonstrated using asymmetric
blowing; 4 ift control without change in pitching moment was demonstrated
when blowing only the lateral edges; and 5 ) sensitivity to the 4:l change in
slot gap is minimal.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 103/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD
91
High
pressure
Low
pressure
Full perimeter blowing
Fig. 28 Upper surface pressure resu lts showing four configurations?
VI. Miniature Oscillatory Valve (CC-Valve)
for Unsteady Wing Load Reduction
There were two main objectives for this project. The first objective was to
demonstrate that mass ejection in the trailing edge region of a hydrofoil could
be used to cancel periodic unsteady hydrodynamic loading. The second objective
was to show that a practical closed-loop control system could be devised and that
the required oscillatory valving could be miniaturized and incorporated into the
trailing edge region of a hydrofoil. The focus of the miniature valve design and
control demonstration w as to develop the capability to cancel unsteady foil forces
and be automatically adaptive to upstream disturbances. Fry and Jessup designed
and tested the slot control valve in 1993.13-15
An overall sketch of the 15-in. chord, dual-slotted hydrofoil used in this dem-
onstration is shown in Fig. 29. The flow ejection for this application was normal
to the surface, as in a je t flap. (The concept could be adapted to the production of
tangential mass ejection). Figure 30 shows a trailing edge section-cut of the
dual-slotted, surface-normal
mass ejection into the boundary layer
Fig.
29
Dual-slotted hydrofoil used for miniature valve proof of ~ 0 n c e p t . l ~
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 104/637
92 R. IMBER, E. ROGERS, AND J. ABRAMSON
electromagnet ic
actuator
Fig. 3 Trailing edge of dual-slotted hydrofoil: rotor pivots to open /close upper
lower slots. Constant fluid pressure results in high-response-rate, efficient system.
4
model and Fig. 3
1
depicts the actuator mechanism construction. A small rocker
valve embedded in the trailing edge uses an electromagnetic actuator attached to
a permanent magnet assembly to produce a high-frequency response of the rocker
valve. As shown in Fig. 30, there is an optional nonmovable trailing edge tail
section. The main feature of the actuator was that it controlled the slot exit
area and not the fluid pressure.
A
schematic of the water tunnel installation is shown in Fig. 32. The hydrofoil
was attached to one side of a 24-in. water tunnel test section. A freewheeling pro-
peller provided a periodic upstream flow disturbance. An external pump and fluid
lines delivered fluid to the rocker valve assembly region. A controller was
employed to send signals to the magnet assembly installed in the trailing edge
body of the hydrofoil.
Permanent
Actuator Rotor
Magnet
Coil
Valve
Assembly
Assembly Assembly
Fig. 31 Actuator mechanism constr~ction.'~
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 105/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD
93
24-inch
water tunnel
Fig. 32 Operational schematic of actuator test.
During operation, fluid pressure to the trailing edge of the foil is
ot
throttled;
it is simply redirected as needed. This low-rate perpendicular ejection of mass
into a boundary layer, from a slot at the trailing edge, may mean that the flow
effect responsible for any change in foil lift is the same as the flow effect attrib-
uted to a Gurney flap. (The term “low-rate’’ ejection is used to m ake a distinction
from the high momentum level of a true jet flap.)
As an example of the experimental results, two plots of force against fre-
quency are show n in Fig.
33.
The top plot shows the periodic force spectrum pro-
duced by the hydrofoil due to the upstream flow disturbance. The bottom plot
shows the force variation with the active flow control system operating.
As shown, the targeted hydrofoil load spikes were successfully eliminated by
the system, with three frequencies simultaneously reduced.
The key findings from the Miniature Oscillatory Valve project include the fol-
lowing: 1) functional model-scale actuators can follow steady or time-varying
input signals up to 500Hz; hydrofoil forces were successfully varied up to
110Hz; nd 3) alleviation of a high-frequency periodic hydrofoil loading is
feasible.
VII.
Dual-Slotted Low Aspect Ratio Wing (CC Hydrofoil)
In 2002, Rogers16 was the principal investigator for an in-depth low aspect
ratio hydrofoil investigation that employed dual-slotted trailing edge CC. The
experimental investigation took place primarily in the Navy’s 10-ft-Large
Cavitation Channel (LCC) and is extensively documented in Ref. 16. The ques-
tion as to the effect of cavitation on the performance of a CC-foil had been pon-
dered for many years and was finally answered during this research.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 106/637
94
R. IMBER, E. ROGERS, AND J. ABRAMSON
Frequency (Hz)
Fig. 33 Load cancellation effectiveness. Two plots show three frequencies of
unsteady lift simultaneously reduced to the broadband fl0 0r. l~
There are several compelling reasons to incorporate circulation control fluid
dynamics on underwater vehicles. T he buoyancy of waterborne vehicles means
that they can and d o operate down to extremely low speeds, where conventional
control surfaces have very limited force generation capability. Because the forte
of CC is to leverage the momentum flux from a slot to make a planar surface
produce m uch greater force than otherwise possible, it becomes an attractive con-
sideration for low-speed m aneuvering enhancement. Furthermore,
CC
augmenta-
tion has its best pumping-power efficiency at low speed, in terms of the control
force advantage over a conven tional surface.
Photographs of the model installed in the water tunnel on a reflection plane,
and of the dual-slotted cross-section, are shown in Fig. 34. A more inclusive
view of the test section is provided in Fig. 35. The half-span model of aspect
ratio 2.0 has an uncambered 20 thick elliptical profile essentially identical to
a previously tested CC airfoil, thus allowing a comparison of two- and three-
dimensional (finite wing) performance. The slot-height-to-chord ratio of approxi-
mately 0.0018 was maintained ove r the full 2-ft span of the tapered planform. The
fluid pressure in the dual plenums could be individually regulated and model
loads were measured by a multicomponent balance.
Fundamental lift performance is shown in Fig. 36 with lift coefficient as a
function of
C
for the model at
10
deg AOA. At low
C
levels, the wing per-
formed slightly better than expected, producing a lift augmentation ratio of 36
in the initial linear portion of the curve. Transition from a linear to a square-
root-like response to C occurred, as expected, at the higher blowing levels.
An important discovery was made early in the test. With only the upper slot
blowing, lift roll-off occurred at a much lower
C
than expected, as shown on
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 107/637
INVESTIGATIONS
OF
CC TECHNOLOGY AT NSWCCD
95
Fig. 34 Dual-slotted low aspect ratio circulation control hydrofoil.'6
Fig. 35 Dual-slotted hydrofoil installed in the L arge Cavitation Channel.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 108/637
96
R.
IMBER, E. ROGERS, AND
J.
ABRAMSON
Cp
total
(upper lower)
Fig. 36 Lift performance and benefit of dual-slot activation.16
the curve in Fig.
36
labeled “single slot”. It was concluded that excessive turning
of the wall je t was causing the loss in lift. The lower slot was then em ployed to
produce a very sm all counter flow, no larger than
5
of the upper flow, to see if it
would prevent the excessive turning. This dual-flow configuration produced the
greatly improved perform ance shown in the upper line in Fig. 36. Note that there
was no performance penalty at low C, for the dual-flow (where the benefit was
not needed) and the investigated C range extended to 0.5, a very high value for
CC. (Recall that the prior dual flow experiment on the LSB airfoil had used a
much higher percentage of second slot flow, 25 , with an accompanying
decline in lift.)
The comparison of actual to expected performance for the hydrofoil, shown
in Fig. 37, shows excellent agreement. For the three-dimensional foil the
average value of C vs C is about
50
of that seen in the corresponding two-
dimensional airfoil data. This is the same ratio of three- to two-dimensional per-
formance a s found for the C, vs AOA on conventional wings of the sam e aspect
ratio as compared to an airfoil. Also, similar to the circular wing discussed earlier,
the induced drag performance of the hydrofoil matched predictions based on con-
ventional lifting line theory.
One of the major test objectives was to determine where the minimum
pressure occurs on the model and what the impact of subsequent cavitation
would be on the ability of the jet to induce circulatory lift, or even to remain
attached. Cavitation occurs when the minimum pressure reaches the value corre-
sponding to the vaporization of water, about 0.5 psia depending on temperature.
The cavitation index, sigma
a),
s the term for the absolute value of the pressure
coefficient that will result in vaporization and is a function of the test section
static and dynamic pressures.
0
0.5
1
1.5
2
2.5
3
3.5
0 0.1 0.2 0.3 0.4 0.5 0.6
with 5% lower slot Cµ
no lower slot flow
AOA = 10°
R52
w i t h 2 n d s l o
t f l o w
single slot
CL
(max conventional)
Cµ total (upper + lower)
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 109/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD
97
Slot Momentum Coefficient total), Cy
Fig.
37
Comparison of actual and expected performance.16
The data plotted in Fig. 38 show that after the onset of cavitation , the lift con-
tinued to increase in response to increasing duct pressure. Eventually the lift
began to roll over, but not abruptly. At no time did the Coanda je t detach
prematurely from the trailing edge due to cavitation. For this particular model
CP
Fig. 38 Lift response to Coanda surface cavitation developrnent.I6
0 0.05 0.1 0.15 0.2 0.25 0.3
CDCD
CL
CD
0
0.5
1
1.5
2
2.5
3
3.5
airfoil data
wing data
CD
drag pred.
0= AO A o
tols-dn2/w
prediction
prediction
CD
Slot Momentum Coefficient (total), Cµ
G3
0
5.0
1
5.1
2
0 50.0 1.0 51.0
σ 5.31=
2.01
6.6
nonoitativacf otesno
r of ecaf r usadnaoC
σ 6.6=
°0= AO A
tolselgnis
Cµ
CL
= pC nim ezir opavdluowr etawhcihwta
5.31=amgis
2.01=amgis
6.6=amgis
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 110/637
98
R.
IMBER, E. ROGERS, AND
J.
ABRAMSON
Fig. 39 Cavitation induced by decreasing tunnel static pressure at moderate lift
coefficient.I6
design, cavitation initiated on the nozzle lip face; see the sketch in Fig. 39. The
photograph in Fig. 39 shows some interesting flow visualization on the CC-
Hydrofoil, compliments of the cavitation that caused the white bubbles of vapor-
ized water. Cavitation is not likely to occur operationally, bu t if it does, it is not
catastrophic to the fundamental C C effect.
Another advantage of the dual slots is the ability to vector the je t thrust. In fact,
in static conditions, as representative of extremely low speed operations, the
direction of jet thrust can be vectored essentially through a full 360deg
because the two jets merge to form a free planar jet. The photographs in
Fig. 40 show qualitatively the results obtained when sequencing through a
range of pressure differentials between the upper and lower slots. The tests
were conducted in air, and air pressure was used in the model. The wall jet
vector directions are visualized with yam tufts. In the photograph at the top
left, only the lower slot is active, and the ejected air follows the curved trailing
edge an d vectors out the leading edge of the wing. In the top center photograph,
a small amount of upper surface slot flow is introduced, which results in lifting
the wall jet off the wing surface. In the top right image, the upper flow is
increased to produce a vertical thrust vector effect from the planar jet
formed by the me rger of the two slot flows. The bottom set of images represent
a con tinued increase in the upp er slot pressure until it equa ls the lower pressure
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 111/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD
99
single slot blowing (lower)
180
deg redirectionof the
wall
jet
Fig. 40 Model checkout in air and dual-slot operation with no freestream,0.2 psig.
The tw o wall jets merge to form a steerable planar jet.16
(bottom right photograph) and the jet flow direction is now
180
deg from the
direction shown in the top left photograph. Verification and quantitative data
for this thrust vectoring capability were measured in water using a load cell
and revealed a thrust efficiency of
70-80 .
Among the many findings from the CC-Hydrofoil test, the investigation
demonstrated that 1) cavitation has a benign effect on the Coanda wall jet and
there is no performance detriment with the onset of cavitation; 2) wake velocity
profile filling is viable with dual slots;
3)
a low flow rate from the second slot can
eliminate one form of the CC lift limit; and 4) dual slots permit 0 to 360 deg static
thrust vectoring and this merged-dual-jet m ode may be viable as a je t flap for lift
augmentation at extremely low speeds, where the coefficient of momentum
would be too high for a viable CC mode.
VIII.
Status of Design Capability
To date, design implementation of CC technology has been based on the his-
torical CC airfoil database and potential flow solutions (two-, three-dimensional)
where local increment in lift is specified directly as an empirical relationshi
between slot momentum flux and two-dimensional lift augmentation.
W hereas potential-flow-based techniques can readily address the “what-if” of
a proposed CC application, they cannot guide the “how-to” in terms of trailing
edge design details and the subsequent mass flow requirements, nor can they
identify the performance boundaries.
To support future CC applications, there exists a need fo r viscous-based com-
putational codes to guide the subtle design details of the Coanda trailing edge
region, as well as the upstream contour. Until a CF D code is available that has
been properly and thoroughly validated for sensitivity to CC contour changes
R
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 112/637
100
R.
IMBER, E. ROGERS, AND
J.
ABRAMSON
(airfoil data exists for this), design details will have to continue to be based on the
historical database and intuition. Reliance on past practices as the only guide will
result in a level of conservative design that may fail to uncover the full perform-
ance potential of lift augmentation by active flow control.
Another benefit of future CFD will be the ability to develop additional insight
into the fluid dynamics of circulation augmentation. For that insight to be valid,
the challenge is to ensure that the numerically modeled flow physics is correct
and not just fortuitous in producing correlation with a given set of experimental
data. At NSWCCD, and other organizations, there are ongoing CFD CC vali-
dation efforts.
IX.
Conclusions
The re have been many diverse experimental CC investigations at NSW CCD
since 1986, the time period reviewed in this summary. Each of the projects
built on lessons learned from previous experiments dating back to the
1970s
for fixed and rotary wing, air and water applications. The experimental results
increased insight into the fundamentals of CC and were used to correlate compu-
tational codes for the evaluation of various proposed app lications.
The six exploratory investigations revealed many new findings and, although
similar in many ways, all were unique in basic geometry. Of the four dual-slotted
models, one was a two-dimensional foil, two w ere three-dimensional foils (one
with tangential mass ejection, the other with perpendicular ejection), and the
fourth was an annular wing (duct). The dual slots provided either increased
control range, extension of lift limit, or increased maneuvering steerage com-
pared with a single-slot configuration, or, as in the case of the miniature oscillat-
ing valve, unsteady load reduction. Tw o of the investigations took place in water
and four in air; however, each would be viable in either fluid regime.
Of relevance to hydrodynamic applications, there had been the question of
what would happen if cavitation occurred in the Coanda-slot region, which is
where minimum pressure occurs. In a special series of experiments, it was
revealed that the onset of cavitation did not have a disruptive effect on perform-
ance. Lift continued to be augmented in response to increased slot flow with
an eventual smooth roll-off in lift as the cavitation became more extensive.
Cavitation resulting from incorporating CC is not foreseen as an issue for pre-
sently contemplated hydrodynamic applications.
Four of the investigations were low aspect ratio geom etries and these helped to
extend the understanding of the viability of CC on short-span surfaces. The suc-
cessful omnidirectional demonstrations of the CC-Disc suggest future application
to very maneuverable low
R
vehicles or to appendages. It is helpful to know
that analysis of the low
R
nvestigations determined that there are no basic
effects unique to wing lift developed by means of the Coanda form of CC, as
compared to the classical lift development approaches.
The TIPJET was the only rotary device reviewed and is unique in that it is
driven in the rotary mode by the same air source that provides the blade lift aug-
mentation. The findings from the TIPJET hover test were significant in furthering
rotary-wing CC knowledge and in demonstrating a novel application of CC .
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 113/637
INVESTIGATIONS OF CC TECHNOLOGY AT NSWCCD
101
In conclusion , the six diverse exploratory applications presented all met basic
performance expectations, suggesting a certain degree of maturity of this technol-
ogy. These experimental results, plus those reported by other organizations, now
make for a rather extensive body of know ledge on the subject, ready to encourage
additional creative applications and to support a variety of full-scale applications.
Com plementing this broad range of empirical knowledge, the anticipated emer-
gence of thoroughly validated viscous-based two- and three-dimensional compu-
tational codes for CC w ill contribute to the achievem ent of the full performance
potential of active flow control by efficiently allowing design refinement of how
the circulation control effect is implemented.
References
‘Nielsen, J. N. (ed.),
Proceedings of the Circulation-Control Workshop 1986
NASA
Ames Research Center, NASA/CP-2432, Feb. 1986.
Englar, R. J., and Applegate, C. A., “Circulation Control-A Bibliography of
DTNSRDC R esearch and Selected Outside References (Jan. 1969-Dec. 1983),”
DTNSRDC-84/052, Sept. 1984.
31mber, R. I., “Exploratory Investigations of Circulation Control Technology: Over-
view for Period 1987-2003 at NSWCCD,” NASAICP-2005-213509, Proceedings of
the
2 4 NASAIONR
Circulation Control Workshop, compiled by
G.
S. Jones and
R. D. Joslin, March 2005.
4Abram son, J.
S.,
and Rogers, E.
O.,
“Design of a Circulation Control Airfoil Having
Both Upper and Lower Surface Trailing Edge Slots (Model LSB17),” DTNSRDC/TM-
16-86/03, Sept. 1986.
’Abramson, J., “Characteristics of a Cambered Circulation Control Airfoil Having Both
Upper and Lower Surface Trailing Edge Slots,’’NSW CCD-50-TR-2004/030, April 2004.
6Reader,
K. R., Abramson, J. S., Schwartz, A. W., and Biggers, J. C., “Tipjet VTOL
UAV Summ ary: Volum e 1-1 200-Pound Tipjet VTOL Unmanned Aerial Vehicle,”
DTRC/AD-89/01, Jan. 1989.
’Schwartz, A., and Rogers, E., “Hover Evaluation of an Integrated Pneumatic Lift/
Reaction-Drive Rotor System,” 30th Aerospace Sciences Meeting and Exhibit, AIAA
Paper 92-0630, Jan. 1992.
‘Schwartz, A. W., and Rogers, E. O., “TIPJET VLAR U AV: Technology Development
Status,” Presented at the 20th Annual Symposium and Exhibit of the Association for
Unmanned Vehicle Systems, June 1993.
'Waiters,
R. E., and Ashworth, J. C., “Experimental Investigation of a Circulation Con-
trolled Shrouded Propeller,” West Virginia Univ., Morgantown, WV, TR-39, Feb. 1974.
“Hoerner, S. F., and Borst, H. V. (ed.), Fluid-Dynamic
Lif t ,
Hoerner Fluid Dynamics,
Vancouver, WA, 1985.
“Rogers, E. O., and Abramson , J., “Selected Topics Related to Operational Applications
of Circulation Control,” NASA/CP-2005-213509, Proceedings of the 2004 NASAIONR
Circulation Control Workshop, compiled by
G .
S. Jones and R. D. Joslin, March 2005.
‘’Imber, R., and Rogers, E., “Investigation of a Circular Planform W ing with Tangential
Fluid Ejection,” 34th Aerospace and Sciences Meeting and Exhibit, AIAA 96-0558, Jan.
1996.
13Fry,D. J., and McG uigan, S., “Hydrofoil Circulation Control Via a M iniature Valve for
Alternating Flows Between Tw o Exit Slots,’’ CDNSW CISH D-1401-02, Dec. 1993.
2
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 114/637
102 R.
IMBER, E. ROGERS, AND
J.
ABRAMSON
Fry, D. J., Louie, L. L., and Jessup, S. D., “A Water Tunnel Evaluation of a Novel
Actuator and Active Control System to Cancel Unsteady Foil Forces,” CDNSWC/
SHD-1401-04, Dec. 1993.
”Louie, L., Fry, D.
J .
and Jessup,
S.
D., “An Active Control System to Cancel Unsteady
Foil Forces,” DE-Vol. 7 5 , Active Control of Vibration and Noise, American Society of
Mechanical Engineers, New York, 1994.
16Rogers,
E.
O., and Donnelly, M . J., “Characteristics of a Dual-Slotted Circulation
Control Wing of Low Aspect Ratio Intended for Naval Hydrodynamic Applications,”
42nd Aerospace Sciences Meeting, AIAA 2004-1244, Jan. 2004.
14
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 115/637
1I.A. Experiments and Applications: Fundamental
Flow Physics
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 116/637
Chapter 4
Measurement and Analysis of Circulation
Control Airfoils
F.
Kevin Owen*
Complere Inc. Pacific Grove California
and
Andrew K. Owent
University of Oxford Oxford England United Kingdom
Nomenclature
c
airfoil chord
CL
lift coefficient
C L infinite aspect ratio lift coefficient
C
blowing momentum coefficient
U
mean axial velocity
U
edg e velocity
u rms velocity fluctuations
x
streamwise position
y
crosswise position
airfoil angle of attack
i induced flow angularity
eff airfoil effective angle of attack
*Consultant.
'Research Assistant. Department of Engineering Science.
Copyright 005 by the authors. Published by the American Institute of Aeronautics and
Astronautics, Inc., with permission.
105
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 117/637
106
F. K. OWEN AND A. K. OWEN
I. Introduction
IRC UL AT ION control (CC) airfoil concepts have been studied extensively
C
or more than four decades. These studies have included low-speed airfoil,
helicopter rotor, and flight demonstrator
configuration^.'-^
In these, and other
studies of CC, the sharp trailing edges of otherwise conventional airfoils are
replaced with rounded or bluff surfaces, typically with either circular or elliptic
cross-sections, with thin tangential blowing slots located on the aft upper surface.
The se rounded trailing edges allow the rear stagnation point to move. This move-
ment is controlled by the relative blowing momentum of fluid injected through
the slots, and by the properties of the external flow field. By blowing through
the slot, a jet sheet is issued, which, as a result of the balance of centrifugal
force and subambient static pressure within the jet, r emains attached to the airfoil.
At low blowing rates, this Coanda effect entra ins upper surface boundary layer
flow and prevents trailing edge separation. As the blowing momentum is
increased, the rear stagnation point is moved further around the trailing edge
and the wake deflection angle is increased. An effective camber is introduced,
and the lift is increased. Blowing rates can be adjusted until the airfoil static
pressure distribution is that predicted by inviscid potential flow. With increased
blowing, the je t controls the location of the airfoil stagnation points, and therefore
the circulation and lift. However, eventually there c om es a point where there is no
longer a balance between the static pressure and centrifugal force and je t blow-off
occurs, with a corresponding dramatic decrease in lift.
Lift values greater than those predicted by inviscid potential flow theory
are generated in the CC regime. Pneumatic camber similar to a mechanical
high-lift system can be obtained. However, CC lift augmentation is far more
efficient than conventional high-lift dev ices, because they only have to overcom e
the viscous losses in the flow. By compensating for the viscous losses, the flow
field more closely resembles the ideal inviscid case. Accordingly, lift augmen-
tation several times that attainable with jet flap or blown devices has been
achieved.
Unfortunately, the precise determination of CC airfoil performance for design
and computational fluid dynamics (CFD) assessment purposes is difficult to
achieve. The most serious problem encountered in testing these high-lift
devices is the interference produced by wind tunnel test section wall separation.
Ow ing to the strong adverse pressure gradients o n the airfoil upper surface, strong
secondary flows can be generated in the sidewall boundary layers. Th e problem
is further compounded by significant spanwise circulation gradients, because
circulation must decrease toward the wall. Even at moderate lift, these factors
can generate trailing edge vorticity more characteristic of a three-dimensional
than an infinite span wing. A great deal of research and analysis is still required
in order to properly establish a reliable database for full-scale model development
and CFD code validation.
To address this shortfall, a wind tunnel investigation has been conducted of a
two-dimensional CC airfoil section equipped with trailing edge blowing. The
tests were conducted in the NASA Ames 2 x 2 ft Variable Density Transonic
Wind Tunnel over a range of freestream Mach number and unit Reynolds
numbers. Detailed nonintrusive flowfield measurements of the mean flow
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 118/637
MEASUREMENT AND ANALYSIS OF CC AIRFOILS
107
and turbulent properties were obtained in the airfoil wake for a number of
different blowing coefficients. In this paper, some of these results have been
related to the CC airfoil performance obtained from direct surface pressure
measurements. The analysis shows that wind tunnel wall interference can
have significant influence on high-lift test results. This influence must be
accounted for before wind tunnel test data can be used for design extrapolation
or for turbulence modeling and CFD assessments. Corrections have been made
for finite aspect ratio (AR) wind tunnel wall interference in order to provide
interference-free benchmark data for turbulence modeling and CFD code devel-
opment and validation.
11
Experimental Details
The work described in this report was conducted in the NASA Ames 2 x 2 ft
Variable Density Transonic Wind Tunnel at a freestream Mach number of 0.5
and at a unit Reynolds number of 3.2 x 106/ft. The test model spanned the test
section and was held at zero angle of attack for the present work. The model
was a symm etric 6-in. chord airfoil, 20 thick, 3 cam ber ellipse with a nom-
inally circular arc trailing edge. An adjustable, nominally 0.010-in. tangential
blowing slot was located on the upper surface, 1-2 before the usual upper
surface separation point, at the 96 chordwise location. Transition strips were
attached to the airfoil section at the 17 chord on both the upper and lower
surfaces. The 1.25-mm-wide strips consisted of 0.13-mm nominal diameter
glass beads. Transition effectiveness was verified by the sublimation technique.
A regulated 3000psig air system was utilized to supply the internal plenum
of the model, and a maximum internal pressure of 60 psig was attainable. The
resulting high internal contraction ratio ensured adequate two-dimensionality
of the je t exit flow. The je t exit velocity was calcu lated from isentropic relation-
ships referenced to tunnel static conditions.
There w ere a total of 9 1 pressure taps on the model, 5 9 of which were posi-
tioned along the centerline. Of these taps, 24 were on the upper surface and 35
on the lower surface. The airfoil performance data were obtained by direct
integration of these centerline pressure taps.4 The flow-field measurements
were obtained using a two-component laser velocimeter with conditional sampl-
i n g ~ a p a b i l i t y . ~he effective sensing volume approximated a cylinder with a
200 diameter and 3 mm length, with its axis aligned with the cross-stream
direction. Detailed measurements of the mean axial and vertical velocities, turbu-
lent intensities, and turbulent shear stress distributions were obtained.
111.
Sample Results
Examples of laser velocimeter wake measurements at 5% chord downstream
of the trailing edge for a zero angle of attack airfoil case are shown in Figs. 1and 2.
These results show the effects of jet blowing on the near-wake axial velocity
profiles. In the zero blowing case, there is a small wake displacement due to
the airfoil camber that produces lift at zero angle of attack. There is also a
large region of reversed flow typical of a blunt body recirculation zone. With a
small amount of blowing
(C,
0.024), there is a significant downward wake
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 119/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 120/637
MEASUREMENT AND ANALYSIS OF CC AIRFOILS
109
01
0 0.005 0.01 0.015 0.02 0.025
0 03
0.035
Blowing Mom entum Coefficient,C
Fig. 3 Measured wake angles.
camber predictions. However, these results are significantly higher than the lift
computed from the measured airfoil surface pressure distributions. However,
as expected, we have seen seed particle deposits on the test section windows,
which suggest that strong secondary flows are generated in the wind tunnel
sidewall boundary layers. This shed vorticity will induce unknown flow angular-
ity in the freestream flow ahead of the model, thus ch anging the airfoil’s effective
AOA. However, from the wake measurements, we are able to calculate the
induced flow angularity as a function of jet blowing momentum coefficient.
These results, calculated assuming a semi-elliptic lift distribution, are shown in
Fig. 5 With this information, we are able to compute the finite AR lift coefficients
that are shown in Fig. 6 . These results are in excellent agreement with the
surface pressure, direct lift measurements shown in Fig.
7.
This comparison
shows that sidewall effects are indeed significant, because agreement is not
reached until an induced freestream downwash for a fully three-dimensional
wing is introduced, that is,
CL 2m f
2.5
01
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035
Blowing Mo mentum Coefficient, C
Fig. 4 Infinite AR lift coefficients.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 121/637
110
F. K. OWEN AND A. K. OWEN
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035
Blowing Momentum Coefficient, C,
Fig. 5 Induced flow angularity.
1.2 I
0 1
0
0.005 0.01 0.015 0.02 0.025 0.03 0.035
Blowing Mom entum Coefficient, C,
Fig. 6 Calculated finite
AR
lift coefficients.
0 1
0
0 005 0.01 0.015 0.02 0.025 0.03 0.035
Blowing Momentum Coefficient, C,
Fig.
7
Measured lift coefficients.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 122/637
MEASUREMENT AND ANALYSIS OF CC AIRFOILS 111
0 1
-
40
50
nit I I
2
$
0.05
Q
- 0
~1 0.05
Ll
i -0.1-
- . 18
WakeTurbulence Level, u '/Ue
Fig. 8
Wake turbulence profile C, =
0 x / c =
0.17).
0 10
20
3
where
Wake turbulence measurements indicate that large-scale fluctuations are
introduced by jet blowing and that wake unsteadiness may well be present at
the higher blowing rates just before jet detachment. In the no blowing case
shown in Fig.
8,
small-scale turbulence dominates, and local RMS turbulence
intensities are related to the local mean velocity gradients as in a plane-
mixing layer. Thus, using the measured local turbulence levels and the
measured local mean velocity gradients, we can calculate the effective
mixing length for this flow. There is good agreement between this calculated
mixing length to wake width ratio of 0.2 compared to the nominal value of
0.18
for a plane-mixing layer. However, o nce je t blowing is initiated, as
shown in Fig. 9, a wide highly turbulent core develops that is indicative of
high turbulent kinetic energy production in the blown jet wake. Turbulent
length scales are increased by a factor of three, an indication of large-scale
turbulent mixing andor wake unsteadiness.
0 1
1 t
, I
WakeTurbulence Level, u '/Ue
Fig.
9
Wake turbulence profile C,
=
0.009
x/c
=
0.17).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 123/637
112 F. K. OWEN AND A. K. OWEN
IV. Conclusions
New CC test measurements and analysis have been presented that show the
need for caution when attempting to use wind tunnel test results for CFD code
validation, or for design purposes. In particular, the results have identified
the quantitative extent wall influence can have on CC test results; for example,
lift augmentation reduced from 68 to 42. Th e results also suggest that turbulence
models must be modified to account for the effects of unsteady, large-scale
turbulent mixing. The agreement between the measured and the calculated
finite AR lift coefficients suggests that if we know the effective angle of
attack, then simple inviscid theory may well be adequate for lift coefficient pre-
dictions. In turn, the analysis suggests that two-dimensional CFD computations
could well be meaningless unless the airfoil effective a ngle of attack is known.
Full three-dimensional calculations will probably be required to account for
wall interference; that is, effective angle of attack and effective camber,
especially at high lift.
Estimates of the errors caused by non-uniform flow due primarily to wall
boundary layer separation are essential. These initial investigations suggest
that angle of attack corrections of at least 1 5CL
will be required. Clearly,
this can be a substantial correction factor, because lift coefficients well in
excess of 2.0 can be expected for high-lift systems. Effects on the estimated
drag coefficient are even more acute. Typical drag coefficients show errors of
over 100 at induced ang les of less than 1 deg. Indeed, at lift slopes typical of
those at transonic speeds, angle of attack errors of
0.01
deg can lead to drag
measurement uncertainty of more than one drag count. Clearly, in any high-lift
experiments, accurate estimates or measurements of induced flow angularity
must be made before useful design estimates or meaningful comparisons with
CFD calculations are undertaken. A detailed review and analysis of finite AR
CC exper iments must be conducted to assess wind tunnel wall effects on
experimental data previously reported in the literature. Although induced flow
angularity is a fundamental consequence of the flow around finite
AR
lifting
wings, our experiments and calculations show that these problems could be
ameliorated to some extent by testing higher AR wings, and by measuring the
induced flow angularity upstream.
References
‘Kind, R. J., and Maull, D. J., “An Experimental Investigation of a Low Speed
Circulation Controlled Airfoil,”
The Aeronautical Quarterly
Vol. 19, 1968, pp. 170- 182.
’Cheeseman, I. C., and Seed, A. R., “The Application
of
Circulation Control by Blowing to
Helicopter Rotors,”
Journal
of
the Royal Aeronautical Society
Vol. 71, 1966, pp. 451-464.
3Englar, R. J., “Development
of
the A-6 Circulation Control Wing Flight Demonstrator
Configuration,” DTNSRDC Rept. ASED-79/01, Jan. 1979.
4Wood, N. J., and Conlon, J. A., “The Performance of a Circulation Control Airfoil at
Transonic Speeds,” AIAA Paper 83-0083, Jan. 1983.
50wen,F.
K.,
“Application
of
Laser Velocimetry to Unsteady Flows in Large Scale High
Speed Wind Tunnels,” International Congress on Instrumentation in Aerospace Sim ulation
Facilities, Inst.
of
Electrical and Electronics Engineers Publ. 83CH1954-7, September 1983.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 124/637
Chapter 5
Some Circulation and Separation
Control Experiments
Dino Cerchie,* Eran Halfon,+ Andreas Hammerich,* Gengxin Han,s Lutz
Taubert,* Lucie-Trouve? Priyank Varghese,* and Israel Wygnanski**
University of Arizona,
Tucson,
Arizona
Nomenclature
c = chord length
C
=
drag coefficient
D / q
c )
CDp
=
form drag coefficient
( s ( p
-
, )
dy/q
c )
2 1/2
C
=
integrated force coefficient
(Ct +
CO,)
C
= lift coefficient
L / q
c )
C
=
pressure coefficient
( p -
, ) / q
CQ
=
steady volume flow coefficient (Q/SU, )
C = steady momentum coefficient
[(2
h/c) (Uslot /
U,)’]
CMac=mom ent coefficient about the aerodynamic center
(c,)
=
oscillatory mom entum coefficient [(h/C)(USlotMax/ u,)~I
d
=
reference length, diameter
= frequency of excitation
h = slot height
J
=jet momentum
q
=
dynamic pressure ( J p U k )
F+
=
nondimensional frequency (f d l
U,)
*Research A ssociate.
’Research Assistant. Currently at Tel-Aviv University, Ram at-Aviv, Israel.
*Research Assistant.
gPostdoctoral Fellow.
TResearch Assistant. Currently at L’Ecole Nationale Supdrieure de M dcanique et d’Adrotechni-
**Professor.
Copyright 005 by the American Institute of Aeronautics and Astronautics, Inc. All rights
que, Poitiers, France.
reserved.
113
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 125/637
114 D. CERCHIE ET AL.
Q = volume flow through the slot
Ree
=
Reynolds number
(U ,8 /u )
U , = freestream velocity
UJ
=
slot velocity
UJ
=
maximum slot velocity
a =
ang le of attack or slot location on a circu lar cylind er
Sf
=
flap deflection
8
=
angular distance from the leading ed ge of a cylinder
I. Introduction
TH IN je t being em itted tangentially from a slot milled in a circular cylinder
A r other convex, highly curved surface, alters its direction and wraps itself
around the surface.
A
circular cylinder can turn a je t around and alter its direction
by more than 180deg. Th e centrifugal force ac ting on the deflected je t is balanced
by the pressure difference between the surface of the cylinder and the ambient
fluid. Integrating this pressure results in a force that is approximately equal to
twice the jet momentum emitted at the slot (Fig.
1).
Blunting a trailing edge of
an airfoil and blowing over its upper surface will deflect the fluid downward,
changing the “Kutta condition,” and provide a powerful means of increasing
the usable lift.
This is loosely referred to as supercirculation. One may divert the flow
around a blunt trailing edge by using suction, as it was aptly demonstrated
by Prandtl,’ who removed the boundary layer from on e side of a circu lar cylinder
and attached the flow on the side of the suction slot and generated lift. This idea
m
B
Fig. 1 Stream lines representing a wall jet flowing around a cylinder.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 126/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
115
was applied by Schrenk to thick airfoils that were otherw ise plagued by ea rly sep-
aration.2 Steady suction has been characterized historically by a volume flow
coefficient CQ, because i ts primary aim was to remove low-momentum fluid
from the boundary layer of a given freestream. Excessive suction could also
provide circulation control (CC), which implied an increase of lift above and
beyond the expected value generated by incidence and camber. The use of con-
formal mapping correctly predicted the lift generated by a strong, slot s ~ c t i o n , ~
which was directly proportional to the sink strength associated with the suction
and depended on the location of the slot on the airfoil. Th e suction contribution
to lift is given by ACL = 2 c Q cot(c$/2), where 4, in this case, represents the
location of the slot in the mapped “circle plane”. The drag penalty associated
with suction is very large
(ACD=
2cQ ), and it was theoretically predicted and
experimentally verified by this model. Slot suction for the purpose of lift
enhancement (CC) did not withstand the test of time because of the associated
drag increase and the large ducts that were required to remove the low-pressure,
external fluid.
As
the thickness of airfoils diminished with the quest to increase
speed, they could not accommodate large internal ducting. Nevertheless,
surface suction and multiple slot suction is still considered to be useful for
drag reduction and for delaying transition to turbulence.
The integration of propulsion with lift generation is a long-sought dream
advocated by many
researcher^.^
Th e advent of jet propulsion seemed to offer
such an opportunity, but it quickly became apparent that materials withstanding
the heat were too heavy and too costly for aeronautical applications. In most
instances (the application to MIG-21 is an exception), only the compressed air
generated prior to combustion by turbojet engines was ducted to slots and
blown over flaps to augment their lift.
A
number of production aircraft used
this form of lift augmentation (e.g., Lockheed F104 Starfighter, Blackburn
NA 39 Buccaneer, Dassault Etandard-IVM).
In the application of blowing, a distinction is made between boundary layer
control (BLC) and circulation control (CC). The first function of the jet, as it
blows over the surface, is to increase the mean kinetic energy of the fluid
within the boundary layer
so
that the latter may advance without separation
into a region of rising pressure, for example, over the upper surface of
a highly deflected trailing-edge flap. An adequate jet momentum is expected
to generate a lift coefficient that is approximately predicted by a potential
flow solution. In this regime of boundary layer control, the lift increment is
roughly proportional to the first power of the jet momentum (ACL oc
C,).
An
increase of jet momentum augments the lift further, but this augmentation
is only proportional to the square root of the jet momentum
(ACL oc JC, .
This is the regime of supercirculation, where the jet departs from the
trailing edge with sufficient downward momentum to increase appreciably the
circulation around the wing. Poisson-Quinton4 is credited with establishing
these criteria, as well as the critical value of
C,),,,
that empirically determ-
ined the momentum required to pass from one flow regime to the other
over an airfoil with a deflected flap at arbitrary angle
8
Circulation
control may also be obtained by blowing the jet obliquely from the trailing
edge of the wing, as was done on pure “jet-flap” experiments; however, there
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 127/637
116
D. CERCHIE ET AL.
C, =0.24
Calculated
using
row
of sinks
Fig. 2 Calculated and measured streamlines around a cylinder.
is a substantial gain in lift when the jet is blown over a suitably designed
solid flap.4
A number of theoretical methods have been developed for predicting the ACL
resulting from supercirculation. Stratford attemp ted to calculate the lift by assum-
ing that the “jet-flap” was equivalent to a physical flap.5 More realistic assump-
tions were made by Helmbold,6 S p e n ~ e , ~egendre,* and Woods,’ who replaced
the jet by a vortex sheet originating at the trailing edge. Woods used the hodo-
graph method, whereas Spence7 and Malavard” linearized the problem, assum-
ing small incidence and small jet deflection.
In all the theoretical models, the mixing of the je t with the ambient flow is neg-
lected. In reality, the je t entrains fluid from i ts surroundings and that entrainment
is well represented by placing a suitable distribution of sinks along its path”
(Fig.
2).
When a strong jet flows over a curved flap or the upper surface of an
airfoil, this distribution of sinks contributes to circulation,’2 which is also pro-
portional to
JC,.
When the jet is emitted from the trailing edge of bluff
bodies (e.g. circular or elliptic cylinders), the entrainment that takes place on
both sides of the jet contributes to form drag.”
Some aspects of the ideal flow models are controversial and they have not
been entirely resolved to date, for example, the prediction that the entire jet
momentum should be recovered as thrust regardless of the jet ’s initial inclination
angle relative to the oncoming stream. This result was proven experimentally up
to a flap deflection of 60 deg, at which approximately
90
of the je t m omentum
was recovered as thrust as long as the value of C was quite large. At la rger flap
deflections, the C required to overcome separation and o ther “real flow” effects
(mixing) became excessive, and the thrust recovery almost entirely vanished
when the flap deflection exceeded
90
deg.
The effects of steady blowing, steady suction, or periodic excitation on
circulation and drag are assessed presently. This report represents an ongoing
research with the purpose of improving our understanding of each technique
and to sorting out the leading parameters that affect, control, and manipulate
the flow. We shall start by examining the flow over a flapped, conventional,
symmetrical airfoil, the NACA 0015 (Fig. 3a). The Kutta condition is fixed
and the impact of the increased circulation is easily recognized when compared
to the standard airfoil performance. Thereafter, we have replaced the normal,
26 chord simple flap with a stubby, 8 chord flap consisting of a circular
cylinder that blends into a wedge having an included angle of
40
deg at its
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 128/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS 117
a
b)
Stream
Fig.
3
Airfoil models used.
trailing edge. This configuration was extensively studied at S t a n f ~ r d ' ~n con-
junction with strong steady blowing. The circular trailing edge facilitates the
generation of supercirculation, but the trailing edge wedge predetermines the
Kutta condition provided the flow over the flap is attached (Fig. 3b). The
flow over the small, blunt, and concave trailing edge brought into focus the
need to investigate the controlled flow over a concave surface in the presence
of adverse pressure gradient more extensively. Such flows were investigated
over wall-mounted humps, started by Stratf~rd,'~ho coined the concept of
a boundary layer that is maintained on the verge of separation over an extensive
distance. When periodic excitation was applied to such a boundary layer,15 the
skin friction was increased while the shape factor was reduced, and it thinned
and stabilized the boundary layer and enabled it to better overcome the
imposed pressure gradient. If the pressure recovery region at the rear of the
hump is made steeper, the boundary layer separates, but
it
has to reattach
farther downstream due to the presence of the long flat surface that extends
beyond the trailing edge of the hump. The control of this flow is reduced to
control of a separation b ~ b b 1 e . l ~ ~ ' ~ecause the hump used in Refs. 16
and
17
is based on Glauert's GLAS I1 airfoil (GLAS stands for Glauert's Laminar
Airfoil Section), it was investigated in the present context (Fig. 3c). The flow
around a thick elliptical cylinder was later examined. Its maximum thickness-
to-chord ratio is 30 , and its leading and trailing edges are circular. This
geometry easily lends itself to a change in the actuation location and in the
slot width. The pressure gradient near the leading edge resembles the pressure
gradient experienced by a standard airfoil, whereas the flow near the trailing
edge is complicated by the fact that the Kutta condition is not well defined.
The circular cylinder was the last test article to be examined, because
it
is
the most widely researched flow, but
it
might be the most difficult one to
control. The Kutta condition is not determined and the parameters affecting
flow reattachment interact and affect the circulation in a more complex
manner than on previous configurations due to the strong coupling between
the flows near the leading and trailing edges.
It is believed that by increasing the complexity and the number of degrees
of freedom that are associated with the various configurations selected, the
dominant variables controlling the flow will be identified. Typical questions to
be answered include the following:
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 129/637
118
D. CERCHIE ET AL.
1) What is the best method o r a combination of methods to increase lift?
2)
Is
C, the unique parameter that governs BLC and CC control and are they
occurring sequentially as C, is increased beyond a prescribed threshold level
(
C J c r i t ?
3) Is separation effectively controlled by suction?
4)
Is
C, displaced by CQwhen suction is used for LBC or CC?
5) When and why is periodic excitation (active flow control, AFC) more
effective than blowing or suction?
6) How sensitive is each method to the location of the actuation, and how is it
affected by the configuration on which it is em ployed?
The present chapter focuses on som e of these questions, in an attempt to categor-
ize the effects of the leading parameters in a rational manner.
11.
Discussion of Results
A. Flow Control over an Airfoil with a Conventional Flap
Most aerodynamic control of lift experiments begin with a standard NACA
airfoil and then either progress in the direction of more custom lofting, lift
augmentation devices or flow control to achieve not only the desired loads, but
more favorable distribution of the load along the airfoil surface. We will
discuss the impact of the total load and distribution of the load on a standard
airfoil using both a trailing edge flap and flow control.
Data were collected using a NACA 0015 airfoil with a simple 26% chord
flap at
Re < x lo5.
A schematic drawing is included in Fig. 3a, showing a
cross-section through the airfoil model. Some early observations carried out by
Greenblatt and Wygnanski indicate that the flow over a deflected flap at
Sf= 20 deg separates around a = - 2 deg." Even at a = 0 deg, both steady
blowing and periodic excitation are beneficial. Consider injection of mom entum
at
C,
=
3% (Fig.
4).
For a flap deflection of
Sf
=
20 deg, both steady blowing and
periodic excitation at very low frequency generate a lift increment of ACL
=
0.5
relative to the baseline airfoil performance, whereas periodic excitation at
F+
= 1.1 generated an inferior lift increment of only ACL
=
0.35. Repeating
the same experiment at a lower C, of 1.2% shows slightly lesser periodic exci-
tation performance at
F+
> 1.1, and even poorer steady blowing performance
(see Fig. 5 for Sf= 20 deg). At Sf= 35 deg and at the high C, of 3%, both
steady blowing and low frequency excitation (F+ = 0.3) peaked out by generat-
ing
ACL =
0.4 and 0.52 relative to the baseline flapped airfoil, respectively. An
increase in the flap deflection beyond this angle caused a reduction in the lift
increment generated by steady blowing until, at Sf> 50 deg, the injection of
steady momentum became detrimental to the generation of lift (i.e., the baseline
CL
exceeded the value obtained by using steady blowing). The efficacy of the
low-frequency periodic excitation at C,
=
3% did not deteriorate with increasing
flap deflection beyond Sf= 35 deg, whereas the excitation at the higher fre-
quency of F+ =
1.1
improved with increasing flap deflection until the two
curves crossed over around Sf= 65 deg. At the lower level of C, = 1.2%,
the increase in flap deflection beyond
Sf=
35 deg rendered the steady blowing
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 130/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
119
1.6-
1.4-
1.2-
1
o-
0.8-
R e =
300K
-#-Baseline
AFC <C
>
= 3 , F+= 0.3
I
+AFC<CP>=3 ,F =1.1
- -
Blowing
C
=
3
P
0.6 I I I
20
40
60
Flap
deflection Sf(')
Fig. 4 Effect of blowing and AFC atC, =
3 ,
as a function of flap deflection on a
NACA 0015 at a =
0
deg.
ineffective, if not detrimental, whereas even higher frequency excitation
remained effective (Fig.
5 ) .
The pressure distribution associated with the three modes of flow control at
C = 3 and
Sf=
35 deg is plotted in Fig.
6 .
The constant, low pressure on
the upper surface of the flap ( x / c
> 0.75)
indicates that the baseline flow was
'L 1.81
1.6
1.4 -
1 2 -
1
o
-
0.8 -
Re = 300K
AFC <C > = 1.2 , F+= 1.1
FC <CP>= 1.2 , F = 2.5
P
+
--C Blowing
C = 1.2
. .
I I I
20 40 60
Flap
deflection
4
)
Fig. 5 Effect
of
blowing and AFC a t C, =
1.2 ,
as a function of flap deflection on a
NACA 0015 at a =
0
deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 131/637
120
D. CERCHIE ET AL.
Fig.
6
Pressure distribution over
NACA
0015 with 26 chord deflected flap.
totally separated over the deflected flap. The flow was partially attached by the
periodic excitation at
F+
= 1.1 and completely attached by the low-frequency
excitation at F+ = 0.3 and by the steady blowing. The reattachment of the
flow over the flap changes the circulation around the airfoil and has a far-reaching
effect on the upstream pressure distribution all the way to the leading edge of the
airfoil. The acceleration of the flow upstream of the slot is of particular interest in
this case.
It seems reasonable to examine various flow control mechanisms providing
identical circulation and C This approach is of practical interest, because a
potential designer may be required to generate a prescribed lift by various tech-
niques available and should be familiar with the consequences, such as drag,
pitching moment, momentum input, and so on, associated with generating the
required lift. During the experiments discussed here, the flap was deflected at
two angles;
Sf=
20
and
40
deg, with periodic excitation being applied through
a 0.06 in. slot at the interface of the main element and the flap shoulder.
Figure 7 includes two pairs of angle of attack (AOA) sweeps with and without
AFC (periodic excitation) at the two different flap deflections discussed
previously. Some features are imm ediately apparent in this figure. All of the con-
figurations share the same d C L /d a when a > 0 deg. The deflection of the flap on
the model increases the effective camber of the model, even if the flow over the
flap is separated, causing a shift upward (or to the left) of the C vs a curves.
However, when the flow over the flap separates (see curve corresponding to
Sf=
40
deg that uses AFC in Fig.
7 ,
here is a
shift
to the right with a dCL/
d a 0 in the range -4
<
a
<
-2 deg. This effect is not seen as clearly for
the baseline airfoil sweep with Sf
=
20 deg because of the low
Re
of the exper-
imen t, although the flow separates partially from the flap a x - deg. The other
two curves, plotted in Fig. 7, are not expected to have a discontinuity in dC L/d a.
The baseline flow over the flap that is deflected at Sf=
40
deg is separated over
the entire range of
a,
considered, and the periodically exc ited flow at Sf
=
20 deg
is attached over the flap until a = astall.he excitation level for Sf= 20 deg
F+
= 0.9,
(c,)
=
2.2%) was specially selected in order to overlap the lift
0.5 1.0
1
0
-1
-2
-3C P
X/
C
Re = 300K α = 0°, δ f = 35°
Baseline
AFC <Cµ> = 3%, F
+ = 0.3
AFC <Cµ> = 3%, F
+ = 1.1
Blowing Cµ = 3%
0.0
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 132/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
121
Fig.
7
NACA
0015
airfoil performance with AFC.
curve generated when the flap deflection of the basic airfoil was Sf
=
40 deg. This
enables a detailed comparison to be made between the effect of flap deflection
and periodic excitation. In fact, the curve for
Sf=
20 deg with AFC falls on
top of the curve for
S =
40
deg without AFC until the occurrence of stall.
Although the stall angle is somewhat higher for Sf= 20 deg in conjunction
with AFC, the resulting CLma, s approximately the same for both cases.
When the same AFC is applied while
S = 40
deg, the maximum lift coeffi-
cient,
C = 2.25
at
a =
10 deg. At negative angles of attack
(-8 < a
<
-4
deg), the
ACL
generated by the application of AFC to
S f =
40
deg is commensurate with the ACL observed at 20 deg flap deflection
for a
< astall,
ecause the flow over the flap is attached for both 8 values in
the respective range of a. Inspite of the flow separation from the upper surface
of the deflected flap at
S =
40
deg, the airfoil continues to generate a higher
lift than for Sf= 20 deg, primarily due to the deflection of the flow by the
lower surface.
In the discussion that follows, we exam ine pressure distributions measured on
the surface of the airfoil model, which produced three different lift coefficients
CL= 1.0,
1.35,
1.5). These “sectional” cuts through the CL vs a) curves
show the different approaches that a designer could select to produce a specific
lift and are marked in Fig. 7 to aid the reader. We consider this to be an important
technique to evaluate different flow control strategies, rather than simply look at
the relative benefit in performance that the control can provide at a fixed geo-
metric configuration. Figure 8 shows four pressure distributions that generate
c,
= 1.0.
In the absence of AFC and with the flap deflected to
40
deg, a slight pitchdown
attitude
a= -2
deg) generates a small suction peak near the leading edge
(LE)
and mild adverse pressure gradient along the upper surface of the entire main
element. The flow over the flap is undoubtedly separated and, consequently,
there is a drag penalty associated with this configuration. When
S f
20 deg
for the baseline airfoil, the incidence must increase to a = + 2 deg in order to
-10 0 10 200
1
2
C L
α (°)
Re = 200K F+
= 0.9
δ f = 20° Baseline
δ f = 20° <C
µ> = 2.2%
δ f = 40° Baseline
δ f = 40° <C
µ> = 2.2%
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 133/637
122
D. CERCHIE ET AL.
Fig. 8 Pressure distribution that develops at C,
= 1.0
for the NACA
0015.
maintain the same lift, and a larger suction peak is created at the LE, while
separation on the flap is pushed slightly farther downstream.
When the appropriate level of AFC is applied while maintaining
Sf
=
20
deg,
the AOA can once again be returned to
a
=
-
deg, resulting in a more uniform
pressure distribution over the upper loft and reducing the suction peak near the
LE. Because
a
s the same for the two flap deflections as the total circulation,
the flow near the LE is identical over the upper surface, a s is the pressure distri-
bution in the range
0 < x / c <
0.4 for the two cases considered
Sf=
40 deg and
Sf= 20 deg with AFC). At x/c
>
0.5 and in the absence of AFC, the pressure
remains constant over the upper surface of the airfoil and the deflected flap,
because it is dominated by the “base pressure” (
C p RZ
- .7) of the recirculating
region downstream. On the other hand, when AFC is applied and the flap is only
deflected at 20 deg, the flow accelerates upstream of the slot (i.e., for
0.6
<
x/c
<
0.74). The flow over the flap is fully attached with a pressure coef-
ficient at the trailing edge being positive ( C px 0.25), suggesting that the flow
downstream of the trailing edge continues with its downward mom entum, gener-
ating perhaps a “jet flap” effect. Increasing the flap deflection to 40 deg while
maintaining the AFC results in an attached flow with C p x 0 at the trailing
edge (TE), while heavily loading the aft region of the flapped airfoil. This is
achieved while maintaining a favorable pressure gradient over the entire upper
surface of the main element by placing the airfoil at a
= - 8
deg. In this case
the flow acceleration upstream of the slot is magnified. This behavior would be
especially advantageous for laminar flow applications where delay of transition
is important. It is easy to identify the upstream influence of the AFC along the
upper surface of the main element.
Figure 9 shows the pressure distributions over the model that generated
CL=
1.35.
The two cases that share the same angle of attack and lift
(a
2
deg,
S
- 20 deg, C
= 2.2
and Sf= 40 deg, C = O%), indicate that
the pressure distributions on both the upper and lower surfaces over the upstream
half of the airfoil are almost identical. The case w ith AFC show s the flow over the
f -
1
0
-1
-2C p
X/C
Re = 200K F+ = 0.9
δ f = 20° α = 2° Baseline
δ
f
= 20° α = -2° <C
µ
> = 2.2%
δ f = 40° α = -2° Baseline
δ f = 40° α = -8° <C µ> = 2.2%
0.0 0.5 1.0
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 134/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
123
CD
Fig.
9
Pressure distributions that develop
C =
1.35 for the NACA
0015.
flap is attached with a pressure coefficient close to zero at the TE. The heavily
deflected flap in the absence of AFC has the sam e
C p
-0.7 at the TE as
it
did
at CL= 1.0, indicating that in both cases the recirculating wake region has
approximately the sam e dimension. It implies that the flow over the flap was com -
pletely separated at both angles of incidence and that the additional lift was gen-
erated by the main element. Once again, the flow with AFC accelerates before
reaching the slot. This reduces the adverse pressure gradient on the upper
surface, making the airfoil less susceptible to stall at this value of CL.The basic
airfoil with the flap deflected at 20 deg also generates
CL
= 1.35, but at
a =
8
deg. Under these conditions, the flow is still separated over the flap, but
the wake is narrow as evidenced by
C p
-
0.2
at the TE. One may reattach the
flow to a deflected flap at 40 deg through AFC enabling
CL
=
1.35 at
a
= -4 deg, due to the increased suction on the aft portion of the main elemen t
upstream of the slot (i.e., at 0.4
<
x / c
<
0.74).
In conclusion, the results described in Fig. 9 are similar to those associated
with C L = , but with the effect of AFC being accentuated. Therefore, not
only does the AFC prevent flow separation downstream of the actuation location,
thereby increasing the circula tion, but it also lowers pressure on the upper surface
upstream of it, enhancing the lift. The prime benefit is in the form-drag reduction,
which was reduced by a factor of four in the range
0.5
<
CL
<
1.
The total drag
was reduced only by a factor of two and there is some uncertainty in the drag
estimate. Nevertheless, the effect of AFC is significant and will be discussed in
full in Section D, p. 144.
The pressure distributions at a given CL,with AFC being applied, suggest that
a can be significantly reduced depending on the deflection of the flap needed to
produce the same lift coefficient. For the lower flap deflection case, the suction
peak i s reduced by approximately 40% and the flow over the flap is fully attached
with a TE pressure coefficient close to that of the freestream. At the higher flap
2
1
0
-1
-2
-3
-4C p
X/C
Re = 200K F+ = 0.9
δ
f
= 20° α = 8° Baseline
δ f = 20° α = 2° <C
µ> = 2.2%
δ f = 40° α = 2° Baseline
δ f = 40° α = -4° < C
µ> = 2.2%
0.5 1.00.0
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 135/637
124 D. CERCHIE ET AL.
deflection, the suc tion peak is further attenuated when sufficient lift is generated ,
even at zero A OA. Although the flow over the flap may not be fully attached, the
upstream effect of the A FC is strong enough to load the main elemen t sufficiently
to generate the necessary lift. This behavior is not unique to this airfoil.
One can conclude from the data presented that AFC contributes through three
distinct mechanisms to airfoil perform ance: first by p reventing flow separation on
a deflected flap (this mechanism w as investigated by Nishri and Wygnanski” and
Darabi and Wygnanski2’); secondly, by enhancing circulation through the invis-
cid je t flap effect; and, thirdly, associated with turbulent entrainment of the flow
and the reduction of the static pressure both upstream and downstream of the slot
location. Poisson-Quinton, in Ref. 4, identified the first two mechanisms using
steady blowing. Acceleration due to entrainment, while being present in those
cases, is more prominent when oscillatory excitation is used.
Th e demarcation between (BLC) and (C C) was quite well defined when steady
blowing or suction w ere used to control the flow, because C C implied “an artifi-
cial increase in circulation ove r that which could be expected from incidence and
camber in unseparated flow”.21 Because the lift over streamlined bodies (over
which the flow is totally attached) is predicted by inviscid solution, a comparison
of pressure distribution both m easured and calculated w as essential. The pressure
distributions calculated from viscous and inviscid solutions using the “Xfoil”
program, assuming that the flow is entirely turbulent in the viscous case, are
plotted in Fig.
10,
together with the measured results with and without the use
of AFC. When
Sf= 20
deg, a =
2
deg and, in the absence of actuation, the exper-
imental results agree quite well w ith X foil’s viscous prediction, with the excep-
tion of the base pressure observed over the separated flap. Th e experimental data
for the forced flow suggest that the flow ove r the flap was attached as a result of
the excitation and, as a consequence, the pressure over the entire upper surface
was reduced (Fig.
10).
The measured pressure distribution resulting from excitation at
(
CY)= 2.2
at F+
=
0.9 fell short of the expected inviscid values, sugges ting that this level of
excitation is below the
CJcrit
that separates the BLC and the CC regimes.
Similar conclusions may be drawn for the results obtained for
Sf=
40 deg and
a
=
-4
deg, except that the ideal flow solution overpredicts the forced results
by a larger amount, and the viscous solution does as poorly in predicting the
base-flow pressure distribution. Both examples (Fig.
10)
confirm the suggestion
that periodic excitation, at the level and frequency used, keeps the flow attached
(i.e., controls separation), but does not enhance the circulation above the normal
inviscid limit.
A complimentary example where the enhancement of circulation was
achieved without reattaching the flow over the flap was provided by
H.
Nagib
(personal communication,
2003)
who examined the control of the flow over a
three-element airfoil with a slot located at the shoulder of a highly deflected,
simple flap. In this case, periodic excitation affected the pressure distribution
on the main element and on the leading edge slat without causing reattachment
to the flap itself (Fig. 11). Circulation was increased and the pressure upstream
of the actuation was lowered, in spite of the fact that the pressure over most of
the flap remained unchanged. Because the upstream effect of AFC may be
more significant than the downstream effect, it is possible that the demarcation
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 136/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 137/637
126
D. CERCHIE ET AL.
CP
Fig. 11 Pressure distributions over a three-element airfoil with and without
CC
(H.
Nagib, private communication).
the absence of flap deflection (Fig. 12), but the rounded trailing edge generates a
large form drag (Fig. 13) and a wider wake than is generally expected from the
NAC A 0015 at a given
Re.
In fact, by deflecting the flap to 15 deg, the C,,,,, as
well as the C, attained at small angles of incidence, is slightly reduced, but this
reduction does not affect the form drag or even the total drag. Strong blowing
approaching C, 1 has been used in previous experiments for CC.13 In the
present investigation
C,
0.1, in order to use the upper limit of
C,
=
0.1 for
comparison with data acquired by Hynes.13 For C, = lo%, CLmaxs increased
to 1.5 in the absence of flap deflection, and it attained C = 2.5 for
S = 60
deg (Fig. 12).
In the absence of blowing, the minimum form drag is attained at incidence
4 < a
<
6 deg, regardless of the flap deflection (Figs. 12 and 13). The total
drag was determined from wake surveys and corrected for buoyancy (both
Betz's and Jones's corrections were used; however, there was no difference
between the two methods of correction). It behaves in a similar manner to
CDp
provided
8
>
30 deg (Fig. 14).
It is interesting to note that the total drag is always lower than the CDp nd the
difference between the two increases with increasing
SF
Because the skin friction
drag is generally positive, there has been a search for experimental error and
uncertainty. It is possible that the number of pressure taps near the leading
edge is insufficient, but
it
is equally plausible that vortices shed from a lifting
airfoil over which the flow is partially separated (either due to high incidence
or large
Sf)
increase CDp,making CDp
>
C,. This excess is more noticeable
whenever AFC is used. The constant presence of vortices downstream of
-0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x/c
Cp
Baseline, slat 2
F = 120 Hz, Uj/Uinf = 2.8
Rec = 0.75e6
Flap = 40 deg.
alpha = 13 deg.
ADVINT/ATT 5% Model in NDF at IIT
Nagib & Kiedaisch; 2002
Slot Location for AFC
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 138/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
127
C, s a,Baseline &Steady Blowing
Fig. 12
CL - a
curves on the truncated NACA 0015, for
C,
=
0
and
C,
=
10 .
a bluff body induces low “base pressure” near the base of such a body and
contributes to form drag. This possibility should be examined more closely in
the future , particularly near bluff bodies where the skin friction drag is negligible
relative to
CDp.
We shall now focus on the effect of increasing
C,
on the characteristics of
the airfoil when
Sf= 30
deg, at a constant representative C, = 1. In the
absence of blowing, CL=
1
is attained at
a
x 7 deg, but at
C,
= 0.1 it is
achieved at
a x -0.7
deg (Fig. 12). There is a coupling between
a
and the
C,
necessary to provide the required lift. This relationship is not linear (Fig. 15a),
although it is explored in the region where dCL/da is constant (Fig. 12). The
highest effect on reducing the incidence required to generate the necessary lift
corresponds to 0.025 < C, < 0.075. The moment coefficient about the quarter
chord location (the aerodynamic center) behaves in a similar manner, implying
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 139/637
128
D. CERCHIE ET AL.
C, vs. C,,,
Baseli ne, Uh+=12m/s, Re=2.P105
Fig. 13 Form drag polars for
C,
=
0 and
C,
=
10 .
that a desired pitching moment can be obtained at a prescribed lift by trading
incidence with C, (Fig. 15a). It is interesting to note how the form drag increases
with increasing
C,,
(Fig. 13), whereas the total drag turns to thrust with
increasing
C
(Figs. 14 and 15b).
The increase in C D p is attributed primarily to the low pressure generated on the
convex surface downstream of the flap shoulder due to the Coanda effect. The
concave surface resulting from the presence of the cusp generates positive
pressure, but the surface is too small to affect the
CDp
in a meaningful way
(Fig. 16). The increase in incidence necessary to generate the proper CL at
lower values of
C
also results in an increased suction at the LE and a reduction
in
CD p .
One may now examine the lift increment generated by increasing
C,
while maintaining a and afconstant (Fig. 17). It is interesting to note that at
low levels of C,, ACL cc C;, and only at higher C,, it becomes linearly
dependent on this parameter. This is contrary to the accepted n o t i ~ n ~ ” ’ ~ ’ ’ ~
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 140/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
129
CL vs. C D ,Basel ine, Ui,=12mls, Re=2.7*105
Fig. 14 Drag polars for C, =
0
at four flap deflections and for
8f
= 30 deg, but for
O < C , < l O .
that in the
BLC
regime (i.e., at small
C,)
ACL cc
C,,
and it only becomes
proportional to JC, in the CC regime. The difference may stem from the low
values of C, that are presently considered. The critical C, distinguishing
between the two regimes is 3.5
<
C, < 5.5 in the range of flap deflections
investigated.
Suction at equivalent C, = 10 generates lower lift than blowing, although
the stall angle increases somewhat by using suction. A comparison between
the two methods is shown in Fig.
18
for
S f =
30
deg. Whereas, for blowing,
most of the added momentum coefficient is manifested as thrust (the CD of the
baseline airfoil is 0.04 while with C, =
10
the CDx
-0.06),
for suction
CD x
0.02,
implying that some 80 of the suction momentum generated drag.
It is interesting to note that, based only on CDp,suction generates apparent
thrust, whereas blowing increases the drag. The explanation for this can come
from comparing the pressure distributions corresponding to C, = 1 (Fig.
19).
In one case there is a negative pressure peak over the deflected flap, whereas
in the other it occurs near the leading edge.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 141/637
130
D. CERCHIE ET AL.
a)
a Pi k h i n g M o m e n t a C14 vs. Cp
9
8
7
6
5
a 4
3
2
0
1
Variation
o f
a & Pitching Moment fo r f ixed
C L = l
F ap=3Oo
..............
..............
-2 J . . . 4.5
0
b)
4
c [ ,.I
8
10
CDp & CrJ vs.
C,
@
C p l
for
various
C,,
hap=300
0.15 ...... ...... ................. ......
...... ......
. . . . . . . . . . ........
......
Dp ..;
............ ....
:.
.....I..
..
-
+- CD
_ _ . I
c
0 1
0
1 2
4 5
6 7
8 9 10
c [ I
Fig. 15
Dependence of a)
(Y
and CM,, and
b) CD
and
CDp
on
C,
at
CL= 1.
C. Controlling the GLAS I1 Airfoil
The Glauert Laminar Airfoil Section I1 (GLAS 11)has a maximum thickness-
to-chord ratio of
31.4
and was designed to operate with massive suction
through a slot located at 69 of the chord. The des igner intended to have favor-
able pressure gradients over most of the upper and lower lofts, and maintain
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 142/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
131
C,
profiles for different C,,,CL-l, qlSp=3O0
Fig. 16 Pressure distributions for various values of C, at C, = 1.
laminar flow over a large fraction of the airfoil surface. Suction provided a
pressure discontinuity across the slot that led to a positive pressure along the
entire concave recovery ram terminating with stagnation pressure at the trailing
edge. With adequate suction ’ the measured
L I D
varied between
250
and
550
for
C,
>
1
and
Re
x lo6. In the absence of suction, L I D > 12 for the same
Re,
but
C, was reduced to C,
0.6.
The L I D increased to approximately 30 at
C,Sweep, Steady Blo win g for different
Flap
Angles
Fig. 17 Dependence of CL on C, at
(Y = 3
deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 143/637
132
D. CERCHIE ET AL.
CL vs. a or various C,
CL
vs. Cm and C,,
Fig.
18
Comparison between the performance
of
the airfoil using strong suction or
blowing at C ,
=
10 .
C,
-
Compar i son be tween S teady Suc t i on
&
B l o w i n g
fi x ed C,-I ,C,=lO ,~,,,=3O0
Fig.
19
Representative pressure distributions on the truncated 0015 airfoil using
suction at
C , = 10 .
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 144/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
133
Re 3 x lo6. It appeared that the flow was intermittently reattaching to the ram p
just upstream of the TE, resulting in large drag oscillations. Blowing appeared to
be less effective than suction, requiring larger mass flux to forcibly reattach the
flow. Either way, the
C,
required to keep the flow attached was in excess of
20%.
Many additional research papers followed , culminating in test flights on a glider.
If comparable performance could be achieved using periodic excitation at
considerably smaller
C,
levels, the use of GLAS-type airfoils could be
revived. Furthermore, there has been a considerable interest in a hump that
was placed on the wind-tunnel wall and whose shape represents the upper loft
of a GLAS I1 airfoil for validation of CFD codes. The differences between the
flow over such a hump and over the airfoil should be fully understood and
properly documented.
Th e dependence of
CL
on
a
s plotted in F ig.
20
for the baseline configuration
at
Re
<
0.5 x lo6.
dCL/da is not constant at these low Reynolds numbers,
but the discontinuity in the slope diminishes with increasing Re. At
Re
=
1.17
x lo5 there is a sudden increase in the lift at
a 20
deg. With
increasing
Re,
the kinks in the CL-a curves occur at lower incidence (e.g.,
a 16
deg at
Re 1.7 x lo5)
and they become more moderate. Also, the
maximum lift experienced by the airfoil decreases from being CL,,
=
1.7 at
Re=1.17
x
lo5
to
C
=
1.05
at
Re
=
4.8
x
lo5.
The data acquired at the
Compressed Air Wind Tunnel of the W L Z 3 grees fairly well with the present
results. The
CL a
curve reproduced from Ref.
23
was taken at
Re = 4.06 x
lo5 and could be obtained by interpolating the present data taken
at 3.5 x lo5 < Re < 4.8 x lo5 (Fig. 19).
No
wind tunnel corrections were
applied to either set of data.
Figure
21
shows the typical pressure distributions measured on the surface of
the baseline airfoil at
1.2 x
lo5 <
Re
<
4.8 x lo5,
corresponding to
CL 0.5,
suggest that a laminar boundary layer could be maintained over the lower loft
C,
YS. a
-
Basel ine
Fig. 20 Dependence of C on (Y and on
Re
for the baseline airfoil.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 145/637
134
D. CERCHIE ET AL.
Baseline
C,
pro f i les at
q 4 . 5
or di f ferent Re
Fig.
21
Pressure distribution measured on the baseline airfoil at constantC but
different values of
Reynolds number.
of the airfoil up to
x /c
ranging between
0.5
and 0.7 due to the favorable pressure
gradient existing on that surface. On the upper loft, however, the location where
the C p s minimum depends strongly on Re as it changes from x /c =
0.07
at the
lowest Re used presently to x /c = 0.33at Re = 4.8 x lo5. The
C p
measured near
the TE indicates that the flow is mostly separated in this region, although the base
pressure increased with increasing
Re
suggesting that the mean size of the separ-
ated region was reduced.
Applying the strongest available suction (C, x 19%), blowing (C, x 22%),
or periodic excitation (at
(C,)
x 2.1%) to this airfoil at x / c
=
0.62 and
Re
1.2
x lo5
results in a tremendous increase in lift and the straightening of
the CL a curves (Fig. 22). The actuation location was moved upstream
because the levels of actuation mentioned above were unable to affect the flow
at small angles of incidence at the original slot location suggested in the litera-
ture. The Australian researchers faced similar difficulties and they, too, moved
the suction location. The chosen location of the slot corresponds to the separa tion
line predicted by C FD and it will be discussed separately. Con trary to the obser-
vations of Glauert et a1.,22 blowing was more effective than suction, approxi-
mately doubling the CL attained at a < 10 deg. The maximum CL was
obtained in both instances at
a,,,
=
24 deg; however, in the case of blowing,
CL,,
= 5,whereas by using suction,
CLm,, =
3.5 only. Periodic excitation at
F+
= 0.7 and approximately 1/10 of the steady momentum input for blowing
performed almost as well as the steady blowing up to
a
14 deg. At higher inci-
dence, the difference in
C,
between the steady blowing and the periodic exci-
tation became noticeable, yielding
CLm,, =
3.5 for the periodic excitation.
Because the
C
obtained for the basic a irfoil was only 1.7, suction and peri-
odic excitation represent 100% increase in
CLm,,,
whereas the much stronger
steady blowing represents almost 200% increase in this coefficient.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 146/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
135
Fig. 22 Dependence
of C,
on a and on the control parameters used and the
corresponding drag polars
for
Re
=
1.17 X lo5.
The drag polars for the corresponding cases are also plotted in Fig. 22. They
reveal very interesting features associated with each method of boundary layer
and circulation control. The baseline drag at
CL
< 0 is approximately
C, 0.11. This number agrees very well with the National Physics Laboratory
(NPL) results at comparable
Re
and
C,
values.23 The baseline pressure drag
CDp
=
0.086 was almost constant for all C,
<
0, and it implies that the skin fric-
tion drag is approximately 0.024 for the negative lift coefficients.
The application of suction at a
<
6 deg generated a small dC L/da and it pos-
sessed relatively large C,. For C,
>
0.4, suction reduced the drag to C, 0.02.
This represents a substantial drag reduction relative to the baseline airfoil that
attains C, 0.4 around a 20 deg. The form drag in this case is negative, indi-
cating that there is a low pressure at the leading edge of the airfoil and a high
pressure on the rear ramp. Strong, steady blowing generated thrust in addition
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 147/637
136
D. CERCHIE ET AL.
to lift enhancement (Fig. 22). The maximum thrust measured at CL 1.4 is
C,
=
-0.27, and it exceeds the total momentum input of C, 22.6 . Accord-
ing to these results, blowing at high levels of C, is much superior to suction. It is
well known that there is a drag penalty associated with the removal of boundary
layer flow through the airfoil’s surface. According to Poisson-Quinton4 his drag
is equal to CDsink 2cQ= 0.0312 and, were it not for this penalty suction, would
have generated thrust as well. When massive blowing is used, the jet momentum
is recovered as thrust, but in addition there is a source flow that should contribute
to thrust. In this case the thrust attributed to the source CD,,,,, = - cQ 0.034.
The maximum thrust recovered may be CT= C, +2cQ 0.26. Because the
energy spent to generate suction is similar to that of blowing, the superiority
of steady blowing is clear in this case. The drag associated with periodic exci-
tation is larger than either steady blowing or suction, but because the
(C,)
spent in this case is but a small fraction in comparison to the steady cases, the
comparison is inappropriate.
Repeating the same experiment at
Re
= 2.35
x lo5
reduces the respectiveC
values by a factor of 4. In the case of periodic excitation, it also affected F+Y
lowering it to F+ = 0.35. The baseline CL a curve is now much more normal,
particularly for a > 10 deg (Fig. 23), reaching a CL,, of 1.45. The
CLma,
gen-
erated by steady blowing dropped from CLmax to 3.1, whereas suction
attained a
CLmaX
f 2.5 only. The efficacy of periodic excitation was further
reduced as a result of the concomitant change in
F+
yielding a
CLm,
of 2.25,
in spite of the fact that for
a <
14 deg it generates a higher lift than the steady
suction does at a C, that is an order of magnitude larger.
The drag polar of the basic airfoil suggests that the flow may be attached to the
TE ramp at
a =
18 deg corresponding to CL
=
1.2 (Fig. 23). For this C,, the
L I D
of the basic airfoil is 15. Steady suction at C, = 4.7 generated identical total
drag at CL
=
1.2. There is a major difference in the form drag of these two
cases. Whereas the CDpgenerated by the suction is
so
small and perhaps even
negative, the
CDp
associated with the basic airfoil is larger than the total drag
CDp
=
0.113, whereas
C,
=
0.085). Active flow control (AFC) generates a
higher total drag at this C, C,
=
0.1 l), whereas steady blowing reduces the
C, to 0.015. In fact, at CL= 0.8 the drag associated with the steady blowing
at C,
=
5.5 vanishes, implying that a wake generated by this C, resembles a
wake of a self-propelled, two-dimensional body. The actual C, required to
propel an aircraft at this CL is higher because of the added induced drag. In con-
trast to the results accumulated on the NACA 0015 and its truncated derivative,
CDp s usually smaller than C, except near the stall angle; it is possible that the
concave curvature affects the result as well.
The pressure distributions over the airfoil at
C,
=
1 and
CL
=
2.1 and
Re
values
of 1.17 x
lo5
and 2.35 x
lo5
are plotted in Fig. 24 for the three control mechan-
isms used. Because the same CL was obtained at different incidence angles as a
result of differences in C, and the specific method of control used, a comparison
of pressures near the LE is inappropriate. Nevertheless, the efficacy of the control
scheme reflects on the incidence angle for which a given
CL
is achieved.
Consider the case for C, = 2.1 and
Re
=
1.17 x lo5(Fig. 24b). It appears that
the flow is accelerating toward the slot for all three of the control schemes used.
Suction, however, provides the least maximum acceleration, but its effect is felt
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 148/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
137
Uinf=I4mls,
Re=2.35*1
O5
Fig. 23 The variation
of
C, with a and the corresponding drag polars at
Re = 2.35 X lo5 and the various control parameters used.
farther upstream from
x / c
= 0.5
up to the slot located at
x / c
=
0.62. At the slot
itself, suction brought the flow to stagnation (C, ), whereas on the reminder
of the ramp (0.65 < x / c < l),an almost constant, slightly negative base pressure
is maintained (e.g., C= -0.2). Active flow control accelerates the flow
upstream of the slot, but
i t
also maintains a good pressure recovery downstream,
culminating w ith C = +0.2 near the TE. Strong blowing (at C that is an order
of magnitude larger than the AFC applied) provides a favorable pressure gradient
just upstream of the slot and a very-low-pressure bubble immediately down-
stream. The bubble is generated by the oblique injection angle of the jet,
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 149/637
P
n
n
Fig. 24 Pressure distributions measured at CL= 1 and 2.1
for
two values
of Re
and various control mechanisms.
Next Page
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 150/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
139
because the slot is inclined at 30 deg to the downstream surface. After reattach-
ment the C p is still negative because the curvature of the surface is convex;
however, at x/c 0.68 the curvature of the surface changes sign and also the
sign of the
Cp ,
which becomes positive downstream of this chordwise location.
The large positive pressure on the TE ramp contributes to the thrust generated
by the steady blowing (Fig. 22). The measured
C p
can be larger than unity
because the high-speed jet that emanates from the slot has a larger total pressure
than the freestream
Cp,,, =
5 ) .
To attain CL= 2.1 at Re = 2.35 x lo5 requires a much larger incidence than
for the lower
Re,
partly because of the lower C, available. In this case the same
C,
is obtained for AFC and for steady suction (in spite of the large difference in
the level of actuation) at
a =
18 deg, enabling a direct comparison between the
pressure distributions upstream and downstream of the slot. It is clear (Fig. 24d)
that steady suction is more effective in accelerating the flow upstream of the slot
than AFC is; however, the latter is more effective in the pressure recovery region
on the ramp. The maximum C p due to steady blowing measured around
x/c = 0.7 was reduced from
C p=
+ 5 to
C p
= +1.25, because the total pressure
of the freestream was increased by a factor of 4. It suggests that the regularly
defined pressure coefficient
so
widely used on normal airfoils is not adequate
in the case of blowing.
The L E edge radius of curvature of the GLAS I1 airfoil is very small, pointing to
a potential problem with this design. There i s a large separation bubble on the lower
surface of the airfoil that reduces greatly the favorable pressure gradient on this
surface. This is most obvious when the flow over the ramp is separated (Figs.
24b, 24c, and 24d). Laminar flow on this airfoil is probably achieved on the
upper surface around CL 1 by a proper com bination of incidence, BLC and Re.
The application of AFC at low values of C,
<
2 is considered in Fig. 25
with particular emphasis being placed on C,
< 1
at incidence a
= 0
and
6 deg. At
a = 0
deg there is an increase in lift of
ACL=
0.4 at
C, <
0.4 ,
whereas at a = 6 deg the same increment in lift requires 0.4
<
C?
< 0.9
depending on the reduced frequency
F+
used. The largest increase in
CL
for
the smallest input in C, corresponds to a reduced frequency of 0.7
<
F+
<
1.
For F+
= 1.8
the sudden increase in
CL
requires a lower input of momentum,
but the ACL is somewhat smaller.
The pressure distributions taken in the region of transition from the “low to
high” CL suggest that even a partial attachment of the flow downstream of the
slot changes the circulation affecting the entire pressure distribution on the
upper surface of the airfoil. Active flow control enables the boundary layer to
overcome a very severe adverse pressure gradient existing on the convex part
of the ramp (i.e.,
0.6
<
x/c
<
0.7), even if it does not succeed in attaching the
flow all the way to the TE (Figs. 26 and 27). The data presented in Fig. 26 corre-
spond to C,
=
0.25 and 0.75 for a =
0
and 6 deg, respectively. It is noted that
the LE stagnation point occurs on the upper surface around x/c 0.01 for
a =
0 deg, and almost precisely at the LE for
a =
6 deg. The existence of a
bubble near the LE of the lower surface is observed. The acceleration of the
flow over the leading
50
of the chord on the upper surface is followed by a
smooth deceleration towards the slot and towards the inflection point on the
surface of the TE ramp, provided the
C ,
is close to the threshold value at
Previous Page
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 151/637
140
D.
CERCHIE ET
AL.
Effect of F’ and C,
on
C, at
GO*,
U i , p l 4 m l ~
Effect of F’ and C on C, at
a= ,
i,,=14mls
Fig. 25 Increase of
C,
with
C,
for a variety of
F’
at two values of a.
which CL ncreases (Fig. 26). A further increase in
C
results in a “spike” in the
observed C p us t upstream of the slot and in a higher pressure recovery over the
ram p (Fig. 27).
These results seem to contradict the previous concept that required a threshold
value of CPcrit o overcome separation before circulation can be increased. In this
case the control of separation (downstream effect) and the increase in circulation
(upstream effect) occurred simultaneously.
The effects of AFC on drag are shown in Fig. 28 for
a =
6 deg,
Re
= 235 x lo5 and for three values of F+ ranging from 0.36 to 1.8. For
F+
>
1, the drag is reduced to approximately one-third of its value in the
absence of excitation, whereas for
F+
= 0.36, the reduction is only two-thirds.
Although the high-frequency periodic excitation reduced the total drag at all
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 152/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS 141
C,-0.26 , GOO,
U,,,+4mls
Fig. 26 Effects of F + on the pressure distributionfor
C,
= 0.25 and 0.75 at two
values
of a.
amplitudes
(C,
> 0), the excitation at F+ = 0.36 increases the total drag for
0
<
C,
< 0.6 . The highest frequency of excitation experimented with to
date,
F+
= 1.8, reduced the d rag at the lowest amplitudes imposed. The response
of CDp o the imposed excitation is quite different. First, CDpwas not increased
by excitation at F+ = 0.36; secondly, this frequency was able to lower the
CDpas effectively as
F+
= 1.8; and thirdly, higher CDp esulted from excitation
at F+
1.
The present results are in general agreement with the parametric
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 153/637
142 D.
CERCHIE ET
AL.
F'=0.36,a=0°,
Uim=14mls
F =0.36,a=6 , U i e l m l s
Fig. 27 Effects of C, on the pressure distribution for two values of F + and a.
study of Nishri and Wygnansk ilg on the reattachment of flow to a generic flap by
using AFC. The high CDpassociated with excitation at F+ x 1 results possibly
from the strongest eddies that are consistently present over the ramp and that
are generated by this frequency. Excitation at lower frequencies results in
stronger eddies that are not always present over the surface, because their
wavelength exceeds the length of the surface, whereas excitation at F+
>>
1
generates weaker eddies, so their constant presence has a lesser effect on
the flow.
Nishri and co lleagu es also observed that to maintain the flow attached requires
a much smaller momentum input than to force a separated flow to reattach.
In other words, a hysteresis should be present whenever the level of actuation
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 154/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
143
a=6 ,
P a n d C,
effect
on
CD,
CD,,
at Ui,,,=14mls
Fig. 28 Effects of C, on the drag for three values of F + at a = 6 deg.
is increased until the flow reattaches or when the fluctuation level is
decreased until the flow separates. Such a hysteresis was observed and is
plotted in Fig.
29.
A comparison among all three methods of control discussed is shown in
Fig. 30 for a prescribed incidence of 6 deg. Active flow control results in a
sudden increase in ACL 0.6 when (C,) 0.5 ; suction requires a
C, = 1.5
to obtain the same result, and blowing requires a higher threshold.
Seifert et al. made similar observations on the NACA 0015 airfoil.24 At larger
values of
C,,
blowing may surpass AFC, but the practical implications of this
are questionable in view of the large momentum required.
a=6*,
FreqrSSOHz, F'11.8, U,,,=14mls
Fig.
29
Hysteresis effect caused by an increase or a decrease in C,.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 155/637
144
C L
D.
CERCHIE ET
AL.
a=6',
C , sweep at U,,,=l4m/s
Fig.
30
Effects of suction, blowing, and AFC on lift at small values of m omentum
input.
D.
Flow Around an Elliptical Airfoil
The elliptical cylinder represents an aerodynamic body that has no definite
Kutta condition at its trailing edge, but it generates pressure distributions that
are similar to an airfoil at an angle of attack. The modified 30 ellipse has
both leading and trailing edge cylinders that have adjustable slot widths and
exhaust locations (measured as included angles either from the LE or TE;
(Fig. 3). The focus in the present paper is only on fluidic actuation emanating
from the TE cylinder. Because of the undefined Kutta condition, the elliptical
airfoil has the adjustable circulation characteristics of a cylinder when flow
control is introduced.
Steady suction, steady blowing, and oscillatory excitation have been tested
using the model. The effectiveness of the three flow control categories at
a
=
0
deg has been quantified (Fig. 31). The slot location
c
was varied around
the TE of the model in order to generate the plotted data. The three flow types
were run at a blowing coefficient of 1.9 . The oscillatory flow produced the
best lift results, followed by suction and then blowing. The slot was inclined at
25 deg to the local surface to enhance mixing with the boundary layer. This
may provide the reason for the poor performance of the blowing relative to the
other control methods. If the exhaust angle had been tangential, blowing might
have extended the attached flow region farther around the TE and increased
the lift. The angle where the applied flow control provided the best lift moved
toward the TE until the control was insufficient to counter the adverse pressure
gradient upstream of the slot. As the momentum of the applied flow control
was increased, the location maximizing the lift moved towards the TE, regardless
of the method of control used. Only
C,
values generated by
AFC
are shown in
Fig. 31, indicating that an increase in (C,) from 1.9 to 2.38 increases both
C
and + by at least 5 deg, to 130 deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 156/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
145
0.41
/>
-@-
Suction C@ 1 go/,)
+AFC f = 130Hz,<C,> = 1.9p)
w . w - I
I
I
100 110 120 1 o 140
TE
slot exhaust angle
( )
Fig. 31
ellipse.
Comparison among three different types of
flow
control on a 30 thick
The pressure distributions at C
0.44
give some indication where the lift is
generated for the three different control types (Fig. 32). The pressure distribution
over the elliptical body is nearly constant for the three cases from 15 to 85
chord. The real difference is at the leading and trailing edges. Active flow
Fig.
32
Comparison among different flow control techniques at constant C, 30
thick ellipse.
0.0 0.2 0.4 0.6 0.8 1.0
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
120° Slot Location
110° Slot Location
C P
x / c
AFC C L = 0.448 <C µ> = 1.9% TE = 110°/0.030"
Blowing C L = 0.436 C µ = 1.9% TE = 110°/0.030"
Suction C L = 0.441 C µ = 1.9% TE = 120°/0.030"
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 157/637
146
D.
CERCHIE ET
AL.
control and steady suction produced the same increase in C p at the LE w hen the
control was applied at the TE. These two control approaches also produced
the same pressure distribution along the lower surface near the TE. Active
flow control and blowing looked similar along the upper surface near the TE
where the higher velocity slot flows were entering the flow field. The low-
pressure peak associated with the slot flow was not present at the TE for the
suction case, as was expected.
As a test note, the AFC data plotted represent the time-averaged pressure data.
The data w ere gathered at
600
Hz, and 700 samples were gathered for each static
port. Based on the number of static ports on the model, the typical data runs were
just over a minute in duration.
The effect of (C,) on
C p
is plotted in Fig. 33. The AFC “off’ (baseline)
condition is also plotted for reference. This baseline pressure distribution
agrees well with the CFD results generated using NASA’s CFL3D program.
The three AFC magnitudes increased the lift by increasing the velocity along
the entire upper surface of the ellipse. The lower surface velocity is hardly
affected by AFC. The application of AFC near the TE increased the circulation
along the entire span of the airfoil, not just locally at the TE. A closer look at
the C p at the trailing edge region for various AFC magnitudes is presented in
Fig. 34. The baseline condition shows symmetrically separated flow on the
upper and lower surfaces from around
97
chord to the TE. Application of
AFC resulted in a large, time-averaged increase in the velocity on the upper
surface. The separated region shrank on the upper surface, but it was hardly
affected on the lower one, until, for
(C,) =
2.38 , the flow over the entire
upper surface is attached.
Fig.
33
Supercirculation on a
30
thick ellipse using AFC (CFD courtesy of
A.
Hassan, Boeing).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 158/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
147
130
Slot Location f f 0
+ - - - - -
\
cp
1
.o
-0.6
I \
-0
-0
0
o
02
AFC C;= .619 -dP>2.38%
-m- Baseline
FL3D Re =340K Fully Turbulent
Fig.
34 Zoom
into
TE
region
of
Fig.
33.
The effectiveness of AFC applied to the TE element of the ellipse as a function
of a s plotted in Fig. 35. The data show a nearly constant benefit of AFC until
a
nears stall. The increase in
C
is approximately 40 .
Angle of attack sweeps were performed with steady suction being applied
from the TE cylindrical cross-section (Fig. 36). Rather than increasing the
blowing coefficient, the slot angle was altered for each of these sweeps.
The intent was to assess the influence of
a
on the optimal slot location and
compare these data to the results shown in Fig. 31 at a =
0
deg. The initial
run was at
C =
120 deg, or 5 deg from the peak performance point shown on
Fig. 31. The lift benefit ACL from steady suction was nearly constant for
a
<
4 deg. The slot angle C was then reduced by 10deg, but a (i.e., a corre-
sponding to
CL,,,)
increased only by 3 deg. Another
10
deg reduction in
C
only
increased
a,,,
by 1 deg. From these data it is clear that the zero angle of attack
data provide a good insight into the optimum performance location. They also
shows that the upstream pressure gradient has a very big influence on the per-
formance of control applied at the TE region. In fact, the first decrease in lift
(stall) occurs for the flow separating from the TE cylinder; thereafter the lift
increases again with increasing a until the flow separates from the LE of the
ellipse (Fig. 36).
The same test was repeated for a larger slot angle and larger suction coefficient
(Fig. 37). The trends seen at the smaller slot-width were validated. The larger
suction coefficient produced larger
C
However, the
a
at which
CLm,,
was
realized was not directly proportional to the change in the suction slot location
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 159/637
148 D.
CERCHIE ET
AL.
CL
1.2
0.8
0.4
0.0
1.6
F* =
0.35, E = 130 /0.030 , 1Oms-l
+Baseline
Cp>=l 2%
C#>=l.8%
0
5
10 15 20
( )
Fig. 35 Angle
of
attack sweep using AFC.
or its width. Fig ure
38
show s the effect of steady suction o n the
C
at
a =
0
deg
where the basic ellipse provides n o lift. T he low er surface velocity is decreased
only slightly by the suction and almost at the same increment along the entire
lower surface. T he strongest effect occurs on the upper surface whe re the increase
in velocity is constant from the L E to 90 of the chord. The change in the TE
CL -W-
Baseline
20
-
10
-
00
.2
=
0 ,
=
0.030 ,
p
1.9%
0.8
0.4
0.0
.d/
0 5
10 15
Angle
of
Attack
( )
Fig. 36 Suction on
30
ellipse:
C,
=
1.9
and h =
0.030
in., U = 10 d s .
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 160/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
149
Angle
of
Attack a
)
Fig.
37
Suction on
30
ellipse at
C,
= 3.5 .
static pressure is very evident on this plot and it is similar to the AFC data shown
in Fig.
33.
Comparing the pressure distribution at C = 0.8 with and without the
suction necessitates a comparison between = 4 deg and =
10
deg (Fig. 37),
which are very differently loaded along the chord and generate different moment
around the aerodynamic center. Active flow control tests at the TE were then
conducted for a variety of slot angles, excitation frequencies, and amplitudes.
Fig.
38
Pressure distribution at C, = 3.5 .
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 161/637
150
ACL
D.
CERCHIE ET
AL.
Fig. 39
Scatter
in
data when C, used as parameter.
Data acquired at two slot locations are plotted in Fig. 39 against CJ. The data
did not co llapse onto a sing le curve, indicating that som e other parameter or par-
ameters need to be included for this type of control and geometry. There are two
broad clusters of points depending on the slot location. There is also a depen-
dence on frequency with the lowest frequency of excitation generating the
lowest
ACL.
Empirical correlation trials resulted in the data plotted in Fig.
40.
This set of
data appears to collapse all the results onto a single curve as a function of the
product of
(C,)
the square root of F+, and the angle denoting the distance
from the TE. The length scale used in the definition of F+ depends on the
same angle, as it represents the length from the slot to the theoretical TE of
the ellipse. In this case the data collapse onto a single curve whose generality
is yet to be proven.
E.
Controlled Flow Around a Circular Cylinder and Reexamination
of Some Old Results
The flow around the circular cylinder is discussed last because it represents the
highest degree of complexity. Similarly to the ellipse it has no defined TE separ-
ation location or imposed Kutta condition. Unlike the previous four geometries
discussed, the LE and the T E flows are closely coupled because of their immedi-
ate proximity. The flow is also sensitive to transition location, which is, in turn,
sensitive to a variety of inflow parameters.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 162/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
151
Fig.
40
Collapse of data based on em pirically derived flow control parameter when
AFC is used.
In the absence of an external stream, a wall jet created by steady blowing
wraps itself around a convex surface of the cylinder, following it up to, and some-
times beyond half of its circumference (Fig. 1). In the example shown, a jet of
momentum
J
emanating to the right from a slot located on top of the cylinder
encircles it before separating to the left from its lower surface. The change in
the direction of the flow generates a low-pressure region on the right-hand
surface, which, when integrated, yields a side force whose magnitude is almost
equal to twice the je t mom entum. This force multiplier makes som e applications
of wall jets over curved surfaces very attractive, arousing interest in improving
the understanding of this flow.
One of the unique characteristics of the curved wall jet is its phenomenal
rate of growth from the surface, and its high turbulence level, which is
attributed to the streamwise vortices generated by a centrifugal in~tability.'~
The cylinder over which these measurements were made was carefully
designed with the jet emerging tangentially to the surface after passing
through a smooth contraction. The resulting flowfield is a consequence of
instability.
Therefore, it is important to know how sensitive this flow is to the detailed jet
characteristics leaving the nozzle and to the initial width of the jet relative to
the radius of the cylinder (Fig. 41). There is a large region around the surface
of the cylinder where the pressure is constant in spite of the rapidly thickening
je t flow. The pressure distribution is not seen to be very sensitive to the jet-related
Reynolds number.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 163/637
152 D.
CERCHIE ET
AL.
Fig. 41 Pressure distributions on a cylinder for various Reynolds num bers.
Two additional cylinders were constructed that were more suitable for
wind-tunnel tests in which an external stream of variable velocity could be
applied. One of the cylinders (2 in Fig. 42) was machined from two parts and
has a con tinuous nozzle along its span. The figure shows that the nozzle design
is a critical feature in determining the external flow characteristics. The best
tangential design, nozzle 1, maintained attached flow at least 40-deg further
around the cylinder than the other two slot designs. Nozzle 3, which had a seg-
mented slot cut through the cylinder’s wall, demonstrated the worst performance.
This poor tangential flow characteristic was previously discussed as a possible
reason for the lower blowing control performance near the TE of the ellipse
relative to suction and to AFC control.
Using slot design 2, the pressure distributions around the cylinder, as a func-
tion of slot location relative to the freestream, were generated for a given-jet-to
freestream velocity ratio of 14.5 (Fig. 43). The data possess the sam e trend as for
the ellipse, where the maximum surface velocity continues to increase until the
slot has rotated far enough from the natural baseline separation point (around
60-70 deg for that Reynolds number) that the added momentum can no longer
hold the flow attached. The maximum C p generated is -2 2 for an injection
angle
+=
120 deg measured from the LE of the cylinder. It is interesting to
note that the region of separated flow on both sides of the cylinder is reduced
as the performance of the cylinder is increased. For the smallest angle of
30 deg, the separated region extends from approximately 140 to 290 deg,
whereas of extends only from 240 to 290 deg for the performance slot angle of
130 deg, which already exceeds the most effective location of 120 deg.
A comparison of the pressure distributions for the three flow control
approaches over a single slotted cylinder is plotted in Fig. 44. The slot
location was rotated from the LE toward the TE and the data plotted
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 164/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
153
Comparisono f Pressure
Distribution
o f C o d a Flow
ReM=llOOO
Fig.
42
Pressure distributions on a cylinder for various slot geom etries.
correspond to the best-performance location. The data are presented in a
manner similar to airfoil data in percent of chord, using the cylinder’s diameter
as the chord, rather than degrees
around the cylinder. The two blowing
cases illustrate a standard observation: the stronger the blowing, the farther
Cp dist r ibut ion for d i f ferent s lot locat ions
Re=26000, U,/Um=14.5, b/R=O.Oll
Fig. 43 Pressure distributions on a cylinder for various slot locations.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 165/637
154
D.
CERCHIE ET
AL.
Comparison for three Cases
Fig. 44 Pressure distributions on a cylinder for three different
flow
control
approaches.
downstream into the adverse pressure region the jet can be introduced
for maximum performance. Note also that the suction and oscillatory (AFC)
flow control have nearly the same pressure distribution from the LE to the
slot, while the AFC input magnitude is less than one-third the suction
magnitude. The AFC and steady blowing pressure distributions differ from the
steady suction case aft of the slot location. The steady suction pressure distri-
bution has a sharp increase in pressure whereas the steady blowing and AFC
possess a more gradual change in pressure toward the TE. Judging from the
pressure distribution, the flow is fully attached for the steady blowing at
C
=
39 ,
but it is separated downstream of the slot when comparable suction
is used. Active flow control manages to attach the flow over
95
of the chord
at C = 9 .
The integrated force coefficient CFwas measured and plotted in Fig. 45a as a
function of slot angle relative to the L E for different slot velocities and heights. In
general, a higher
C
generates a higher force. An attempt to normalize this data
with respect to the jet momentum (as was done in the absence of an external
stream) is plotted in Fig. 45b. The data nearly collapse to a single curve for
small jet injection angles relative to the ideal LE, but the scatter is still large
downstream of the slot location corresponding to
C
The pressure distributions along the cy linder for three different slot velocities
are plotted in Fig. 46. In the first case the pressure distribution is normalized by
the freestream dynamic pressure q , whereas in the second it is normalized by the
cylinder diameter and by the je t mom entum
J
Again, the trend is a proportional
relationship between slot velocity and the freestream, but neither
q
nor
J
normal-
ize the pressure distributions correctly. In the first case the higher the ratio
between the jet and the freestream velocity, the higher the negative C while
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 166/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
155
a)
Fig. 45 Total force generated on cylinder for different blowing conditions: a)
C,,
b)
CFIC,.u
in the second the roles are reversed. Some other parameters are required for better
correlation of the data, although se lf-similarity in
C
may not be possible for the
entire range of velocity ratios.
A
possible parameter would include C and a correction to that term
that would be flow configuration dependent. One such approach was
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 167/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 168/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
157
the following form for discussion purposes:
In fact, when a je t of kinematic mom entum
J
and volumetric flow q is injected
parallel to an infinite stream of velocity
U,,
in a way that retains constant
pressure on the surfaces bounding a control volume, the application of the
momentum theorem yields an invariant
which is proportional to Englar’s simplified version of C The two
derivations lead to the same result when is calibrated in the absence of an
external stream, which is normally the case. The last equation is applied to
some early da ta in boundary layer flow and circulation control to determine its
applicability.
The blown flap data from Poisson-Quinton and Lepage2 is adjusted using the
modified flow coefficient and the results are replotted in Fig. 47 together with the
original results that use jus t C Similar reprocessing was carried ou t to the w in
data of Williams21 (Fig. 48) and to the NACA 84-M airfoil data of Attinello
(Fig. 49). In all these cases the data scatter is no greater using the modified par-
ameter, and in some of the more critical, low-momentum blowing tests the
inclusion of
CQ
seems to fit the data better. Large
C,
values in testing are
typically the result of large slot velocities. For these conditions, the impact of
the correction term diminishes, allowing for reasonable performance predictions
to be made for a given configuration. In an effort to minimize the momentum
input required for effective flow control, this modified parameter provides a
better insight into the measured data for smaller flow coefficients. In Fig. 49
the data collapse better using the modified parameter for the small corrected
flow coefficient conditions, even matching the measured data while it shows a
negative lift increment when the flow coefficient is negative. However, it is not
suggested that these data justify a new term for design purposes. What this
does show is that in order to optimize flow control using a minimum input, a
flow-related correction parameter would have to be added to
C,.
It is unknown
whether a correction factor can be introduced that would collapse all flow geome-
tries and control strategies to a unique line without the introduction of terms that
account fo r the pressure gradient and boundary layer characteristics in the region
of the slot.
An interesting characteristic of flow control is shown in Fig. 50. The cylinder
exhibits a lift hysteresis around its maximum lift condition due to changes in slot
9
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 169/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 170/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
159
CL
vs.
Cp or (0.5*CpCQ)
(from
John
Attinello,
1961)
Fig. 49 Blowing over a wing section using data from R ef. 2 (Attinello).
location, under steady blowing. This phenomenon is similar to an airfoil when the
angle of incidence is altered around
astall.
herefore, the initial flow condition is
also important to the performance of BLC and CC, as was discussed by Nishri
and W y g n a n ~ k i . ' ~revious tests on a Wortmann airfoil27 have shown that this
hysteresis can be reduced or eliminated by using AFC. The pressure distribution
around the cylinder w ith a slot located at 120 deg from the L E indicates the flow
can be either attached, yielding a maximum lift, or completely detached, provid-
ing no lift at all (Fig. 50b). The latter pressure distribution is very close to the
baseline data.
The significance of a single slot location to the performance of the prescribed
control mechanism has been amply demonstrated, but very few attempts have
been m ade to add another slot with another array of actuators or jets . A second
slot was thus added to cylinder 3 (Fig. 42) to determine its possible impact.
Figure 51 shows the dual-slot configuration, where both slots are on the same
side of the cylinder and they inject flow in the same direction. The performance
of this configuration relative to the single-slot performance for the same blowing
coefficient is compared. The dual-slot arrangement provides 30 more lift rela-
tive to the single slot by improving the downstream flow conditions. It is interest-
ing to note that the second slot provides only a limited control of separation
downstream, but it alters mostly the pressure distribution upstream, contributing
to supercirculation. This emphasizes the significance of the location of the fluidic
actuator and suggests that a distributed actuation may be more effective, as the
designers of the NOTAR helicopter demonstrated.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 171/637
160
a)
Pressure Com parison,Cp=O,57,or=190°
Fig. 50 Lift and pressure hysteresis for changing blowing location on a cylinder.
Control of lift is not the sole purpose of BLC, because drag reduction for
cruise might be as important. By blowing from two symmetrical slots located
at
a
=
&
110 deg away from the lead stagnation point, the pressure or velocity
over the cylinder can be changed w ithout introducing lift (Fig. 52) .The inviscid
value of the pressure coefficient of is realized at 50 of the chord position.
The cylinder’s drag drops from approximately
C D >
1.0 for the natural case (not
blown) case to a CD
x
0.4 for the symmetrically blown case. Proper accoun ting
should be m ade in order not to waste momentum, because too much flow control
may be applied for given conditions. Figure 53 shows that for symm etric blowing
on the cylinder at C = 0.46, the total drag reduction is 0.61 and all of the
injected flow is recovered as thrust. However, when C, is increased to 1.02,
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 172/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
161
Comparison of Pressure Distribu tion slot located at u=9O0
Cylinder # 3, Re=39000, Cp=1.31
Fig.
51
Pressure d istributions on a single- and dual-slot cylinder.
the total drag reduction is only 0.69 and some of the injected flow is wasted.
The latter case may provide other benefits such as enhanced stability, but the
performance suffers. It is also interesting to note the increase in the wake's
vortex shedding frequency, which is associated with the reduction in drag
(Fig. 53) and is equivalent to the momentum thickness in the wake. This is
Double slo tted Cylinder 1 o o , 0 sweep bac k,
Re135000, Steady Suction, C, per slot
Fig. 52 Pressu re distributions on a du al-slotted cylinder for different suction C,.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 173/637
162
D.
CERCHIE ET
AL.
C
0.00
0.46 1.02
C D ~
.02 0.26
0.18
Co
1.04
0.43
0.35
0.00 -0.14 0.32
Fig. 53 Cp distribution and wake profiles associated with a dual-slotted cylinder.
consistent with observations suggesting that a dom inant S trouhal number associ-
ated with vortex shedding is constant, requiring a frequ ency increase resulting
from a de crease in the mom entum thickness.
111.
Conclusions
Flow control tests ove r five two-dim ensional, aerodynamic, and bluff bod ies
provided s om e new insight into the parameters gov erning active control of sep-
aration and circulation. Suction, blowing, and periodic forcing enable one to
tailor the pressure distribution over the airfoil surface in a similar fashion to
lofting or the introduction of passive devices such as flaps or control surfaces.
Th is chapter questions some of the acce pted concepts associated with separ-
ation and circulation control, although it is not able to provide definitive
answers to these questions. For exam ple, the enhan cemen t of circulation can be
achieved without a ttach ing the flow; how ever, by forcing separa ted flow to reat-
tach to the surface, circulation is generally enhanced. Flow reattachmen t and the
control of circulation d o not have to occur sequentially, as it is widely believed,
leading to so me critical value of an input param eter that separates the two.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 174/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
163
Th e use of steady blow ing is ineffective a t low C,, but its ability to contribute
to thrust while increasing the circulation make it attractive at
C ,
> 5 . Because
0.05
represents a typical ratio of thrust to lift on a civilian airplane in cruise,
steady blow ing might beco me the technique of choice whenever the integration
of propulsion and ae rody nam ics are considered. Suction and periodic excitation
are much more effective in reattaching the flow at low C , values and enhanc ing
the lift, but the input momentum is usually not recovered. Steady suction, in
particular, contributes to drag. The low levels of C , required to attach the flow
by periodic excitation m ake it the most attractive technique for lift enhancem ent.
The momentum coefficient that is widely used to correlate the data is not as
universal a quantity as it is believed to be. It is mostly deficient at low-level
inputs that characterize AFC, requiring consideration of mass addition and
perhaps other variables (e.g., changes in the Kutta condition and pressure
gradient at the slot). A corrected flow coefficient term was discussed and
applied successfully to some well-known flow control cases. The modified
control parameter may be especially useful for correlating data using low-
intensity, pulsed blowing or suction. A better understanding of the performance
of slot design, its width, and velocity for various geometries is required.
The drag of a two-dimensional bluff body is dominated by pressure drag
because the skin friction contribution to drag is small in comparison. However,
in the presence of fluidic control the reliance on C D p
s being the m ajor contri-
butor to drag is wrong. In most cases, as C , increases,
so
does
C D p
while the
actual drag decreases. This is most obvious when the TE is blunt and the flow
is attached due to fluidic control. One wonders how to normalize the pressure
distribution around an airfoil when
C ,
is of order
1
because then
qc J
Th e con-
trolled flows over the ellipse and the circular cylinder have many features in
common. On the ellipse, however, one may change the pressure gradient
upstream of the slot by changing the location of the slot relative to the center
of the ellipse o r by cha nging the thickness ratio of the ellipse. Th e TE perform-
ance becomes less predictable as the pressure gradient along the forward section
of the elliptical airfoil is increased. An empirical parameter was introduced to
scale these results and make a rational comparison with other geometries that
are
creating large circulation at the TE. The circular cylinder data demonstrated
the importance of a good slot design as well as the superior efficiency of
oscillatory flow control relative to steady blowing or suction. It also showed
that there is a limit to blowing that should be observed if the overall system
performance is to be objective.
Th e curvature of the surface downstream of a slot is an important variable, a s
was demonstrated on the truncated NACA and GLAS I1 airfoils. A concave
surface generates a positive pressure in the TE region that reduces the drag;
however, it is susceptible to centrifugal instability and may generate streamwise
vortices whose effect on separation and CC is not well understood. Periodic
excitation em ana ting from a two-dimensional slot utilizes the inherent convective
instability of the flow that is either separated or is on the verge of separation.
Curvature renders another mode of instability that can be used to delay separ-
ation. It may enable the replacement of slot blowing by compact individual
nozzles whose spanwise separation triggers the centrifugal instability associated
with curvature.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 175/637
164
D.
CERCHIE ET
AL.
Acknowledgments
This work was supported in part by a grant from ONR that was monitored by
R. Joslin. The authors are indebted to A. Hassan for his help in providing the CFD
results that guided some of the experiments on the ellipse and the GLAS I1 airfoil.
The authors wish to acknowledge H. Nagib for providing them with an important
figure and for many helpful discussions related to AFC.
References
‘Prandtl, L., “The Generation of Vortices in Fluid of Small Viscosity,” Journal of the
Royal Aeronautical Society, Vol. 31, 1927, p. 735.
2Lachmann,
G.
V. (ed.),
Boundary Layer and Flow Control: Its Principles and Appli-
cation,
Pergamon Press, New Y ork, 1961.
3Goldschm ied, F. R., “Integrated Hull Design, Boundary Layer Control and Propulsion
of Submerged Bodies,” Journal of Hydronautics, Vol. 1, 1967 p. 2.
4Poisson-Quinton, Ph., “Recherches Theoriques et Experimentales sur le Controle de
Couche Limites,” VII International Congress of Applied Mechanics, 1948.
’Stratford, B. S., “Early Thoughts on the Jet Flap,” Aeronautical Quarterly, Vol. VII,
1956, p. 45.
6Helmbold, H. B., “The Lift of a Blowing Wing in a Parallel Stream,” Journal of the
Aeronautical Sciences,
Vol. 22, 1955, p. 341.
7Spence, D. A., “The Lift Coefficient of a Jet-Flapped Wing,” Proceedings of Royal
Society Series A , Vol. 238, 1956, p. 46.
‘Legendre, R., Influence de 1’Emission d’un Jet au bord de Fuite d’un Prof1 sur 1’Ecou-
lement autour de ce Profil, Comptes Rendus, AcadBmie des Sciences, Paris, 1956.
’Woods, L. C., “Some Contributions to Jet-Flap Theory and to the Theory of Source
Flow from A erofoils,” A.R.C. Current Paper 388, 1958.
“Malavard, L.,
Sur une Thkorie Lineaire du Souflage au bord de Fuite d’un Profil
d’Aile, Comptes Rendus, AcadBmie des Sciences, Paris, 1956.
“Wygnanski, I., “The Effect of Jet Entrainment on Loss of Thrust on a Two-
Dimensional Jet-Flap Aerofoil,” Aeronautical Quarterly, Vol. 17, 1966, pp. 31 -51.
‘2Wygnanski, I., and Newman, B.
G.,
“The Effect of Jet Entrainment on Lift and
Moment for a Thin Airfoil with Blowing,” Aeronautical Quarterly, Vol. XV, 1964, p. 122.
13Hynes,C. S., “The Lift, Stalling and Wake Characteristics of a Jet Flapped Airfoil in a
Two Dimensional Channel,” Stanford Univ., Stanford, CA, SUDAAR No. 363, 1968.
14Stratford, B.
S., “An
Experimental Flow with Zero Skin Friction Throughout its
Region of Pressure Rise,”
Journal of Fluid Mechanics,
Vol.
5 ,
1959, pp. 17-35.
‘’Elsberry,
K.,
Loeffler, J., Zhou, M. D ., and Wygnanski, I., “An Experimental Study of a
Boundary Layer that is Maintained on the Verge of Separation,”
Journal
of
Fluid Mech-
anics, Vol. 423, 2000, pp. 227-261.
‘ eifert, A., and Pack, L. G., “Active Separation Control on Wall Mounted Hump at
High Reynolds Numbers,” A I M Journal , Vol. 40, No. 7, 2002, pp. 1363- 1372.
17Greenblatt, D., Paschal,
K.
B., Yao, S. C., Harris, J., Schaeffler, N. W., and
Washburn, A. E., “A Separation Control CFD Validation Test Case,” AIAA Paper
2004-2220, June 2004.
18Greenblatt, D., and Wygnanski, I., “The Control o f Flow Separation by Periodic
Excitation,” Progress in Aerospace Sciences, Vol. 36, 2001, pp. 487-545.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 176/637
CIRCULATION AND SEPARATION CONTROL EXPERIMENTS
165
‘’Nishri, B., and Wygnanski, I., “Effects of Periodic Excitation on Turbulent Flow
Separation from a Flap,”
AIAA Journal,
Vol. 36, No. 4, 1998, pp. 547-556.
”Darabi, A., and Wygnanski, I., “Active Management of Naturally Separated Flow Over
a Solid Surface, Part
11:
The Separation Process,” Journal
of
Fluid M echanics, Vol. 510,
”Williams, J. “British Research on Boundary Layer Control for High Lift,” Boundary
Layer and Flow Control: Its Principles and Applications,
edited by G.V. Lachmann,
Pergamon Press, New York, 1961.
”Glauert, M. B., Walker, W.
S.,
Raymer, W.
G.,
nd Gregory, N., “Wind Tunnel Tests
on Thick Suction Airfoil with a Single Slot,” Aeronautical Research Council R
&
M,
No. 2646, Oct. 1948.
23Salter, C., Miles, C. J. W., and Owen, R., “Tests on GLAS I1 Wing Without Suction in
the Compressed Air Wind Tunnel,” Aeronautical Research Council R
&
M No. 2540,
Feb. 1948.
24Seifert, A., Bachar, T.,
Koss,
D., Shepshelovich, M., and Wygnanski, I., “Oscillatory
Blowing, a Tool to Delay Boundary Layer Separation,”
AZAA Journal,
Vol. 31, 1993,
pp. 2052.
25Neuendorf,R., Lourenco, L., and Wygnanski, I., “On Large Streamw ise Structures in a
Wall Jet Flowing over a Circular Cylinder,”
Physics of Fluids,
Vol. 16, No. 6, 2004,
26Englar, R. J., “Test Techniques for High Lift Two-Dimensional Airfoils with Bound-
ary Layer and Circulation Control for Application to Rotary Wing Aircraft,” Canadian
Aeronautics and Space Inst. Annual General Meeting, “Practical Aspects of V/STOL
Wind Tunnel Testing,” M ay 1972, pp. 1-50.
”Neuburger, D., and Wygnanski, I., “The Use of a Vibrating Ribbon to Delay Separ-
ation on Two Dimensional Airfoils,”
Proceedings
of
Air Force Academy Workshop on
Separated Flow,
F.J. Seiler Research Labs. Rept. TR-88-0004, 1987.
2004, pp. 131 144.
pp. 2158-2169.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 177/637
Chapter 6
Noise Reduction Through Circulation Control
Scott E. Munro,* Krishan K. Ahuja,+ and Robert
J.
Englar'
Georgia Institute of Technology, Atlanta, Georgia
Nomenclature
a
speed of sound
c chord
c1
airfoil lift coefficient
h slot height
riz
mass flow rate
p pressure
q dynamic pressure,
pV2
R radial d istance from je t exit to m easurement location
r radius of CCW surface
T
temperature
V velocity
a angle of attack
Af
frequency bandwidth for narrowband acoustic spectra
polar angle (with respect to the flow axis)
p
density
Re Reynolds number
Subscripts
s associated with slot
T associated with tunnel freestream
*Graduate Student, School of Aerospace Engineering; currently at Naval Air Warfare Center,
'Regents Researcher and Professor, Georgia Tech Research Institute, and School of Aerospace
*Principal Research Engineer, Georgia Tech Research Institute, ATAS Lab. Associate Fellow
Copyright 005 by the authors. Published by the American Institute of Aeronautics and
Wea pons Division, China Lake, California. Student Member A I M .
Engineering. Fellow, AIAA.
AIAA.
Astronautics, Inc., with permission.
167
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 178/637
168
S. E. MUNRO, K. K. AHUJA, AND
R.
J. ENGLAR
associated with je t
o ambient condition
I. Introduction
NE of the major environmental dilemmas facing today's aircraft industry is
oise pollution from aircraft, especially around the airport. There is a large
emphasis on minimizing community noise due to operation of aircraft at and
around airports. Thus, airlines, aircraft manufacturers, The National Aeronautics
and Space Administration (NASA), and the Federal Aviation Administration
(FAA) have made reducing aircraft noise a priority. NASA has proposed a
goal of lowering total aircraft noise emissions by
20
(effective pressure noise
level) EPNdB by
2020.
In order to meet this goal, NASA and other organizations have been
encouraging innovative research to help reduce aircraft noise. Because a
major contributor to aircraft noise on approach is airframe noise (or perhaps
even on takeoff if the engine noise is eliminated), reducing this noise would
be helpful in reaching the industry goals. The major contributors to airframe
noise are the landing gear, the slats, and the flaps. Much work has been
done in these areas in the last five years in an effort to reduce their noise
emissions. Of course, the best solution would be to have an aircraft without
these protrusions into the flowfield. Obviously, an aircraft without landing
gear would have serious drawbacks, but there are alternative high-lift systems
that could replace conventional wing flaps and slats, which have shown great
promise in maintaining and even surpassing the lifting benefits of conven-
tional flaps.
Circulation control wings (CCW ) have been researched and developed exten-
sively, primarily for the purpose of increasing performance and reducing o r repla-
cing the conventional flap system of an aircraft.' Ov er the years, C CW systems
have gone through many configuration designs for many different applications,
including versions for rotorcraft, fighter aircraft, and short haul transports.'
However, there has been limited research investigating the possible acoustic
benefits provided by such a system, other than occasional references to smaller
noise footprints due to shorter takeoff and landing distances. The only known
work on the acoustics of CCW is that of Salikuddin et a1.,2 who evaluated the
noise field of an upper surface blown wing with circulation control (CC). That
study, however, did not provide an indication of the acoustic benefits of a
CCW compared with a conventional wing for the same lift.
Because CCW systems have already been shown to be an adequate replace-
ment for conventional flap systems in the aerodynamic realm,' they are immedi-
ately a candidate for reducing airframe noise as they eliminate much of the
structure of the conventional flap system that protrudes into the flow.
However, there are many issues that need to be resolved before the claims of
lower noise are validated. The CCW system has never been evaluated on an
acoustics basis,
so
it must be optimized for this, while maintaining, at a
minimum, the lift characteristics of a conventional system. The acoustic
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 179/637
NOISE REDUCTION THROUGH CC
69
impact of several parameters must be investigated, such as the blowing slot
height, slot velocity, and CCW geometric configuration (i.e., flap type and deflec-
tion angle). In order to correctly define the best combination, new areas of
research will have to be investigated, including jet noise of extremely high
aspect ratio (AR) nozzles, and the effects of jet turning on its noise propagation.
These many issues are the motivation of the present study. The current work
involves both experimental and computational efforts. Only experimental
results are presented in this chapter. Computational results are presented in
Chapter
22
of this volume and in Ref.
3.
11 Background
The CCW concept has been researched since the
1960s.
The CCW uses a
rounded trailing edge (Fig. l). Air is blown tangentially along the upper
surface from a plenum supply inside the wing through a slot just upstream of
the rounded trailing edge (TE). Blowing moves the upper surface separation
point around the TE, thus changing the TE stagnation point location, and
hence the circulation for the entire wing. The higher-speed air moving
along the surface also causes a suction peak in this region and contributes to
increased lift.
The slot flow remains attached to the surface due to the so-called Coanda
e f f e ~ t . ~t low blowing velocities, the tangential blowing behaves similarly to
a boundary layer control (BLC) device by adding energy to the slow-moving
flow near the surface. At higher blowing rates, the lift is increased by the
change in circulation already described. A CCW can be designed without any
mechanical moving elements if desired. This is achieved using a rounded TE,
where the amount of lift is controlled by the pressure valve to the supply
TANQENTIAL BLOWING OVER ROUNDED TFWILINQ EDGE SURFACE
Fig. 1 Schem atic circulation control wing concept.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 180/637
170
S. E. MUNRO, K. K. AHUJA, AND
R.
J. ENGLAR
plenum. This eliminates the need for flaps with hinges, tracks, screw drives, and
hydraulics.
The increment in lift generated is controlled by the nondimensional parameter
C,,
defined using slot and freestream properties:
With a wing, the nondimensionalizing area is the wing surface
S.
For an airfoil,
C,
is typically given in C,/ft, because the chord is the only availab le reference
length. In general, a given C, will provide a g iven increment in the lift coefficient
over the entire range of angles of attack below stall. The exception to this is
when the slot jet velocities or slot heights are large enough to cause the jet to
separate prematurely. Thus, C, is used extensively in the literature when
discussing CC.
The large, circular trailing edges used in many of the early experiments
evolved into a dual-radius hinged flap, mainly because the nonsharp TE
greatly increased drag.195-8 The hinged flap was a comprom ise of several
desired features. The flap had a curved upper surface, like the cylindrical TE,
but a flat lower surface. This overcame the problem of high drag in cruise associ-
ated with the nonsharp TE of the early designs. Overall, the hinged-flap,
dual-radius design still maintained most of the CC lift advantages, but greatly
reduced the d rag problem associated with the circular TE system.
The flap itself has several mechanical advantages compared to conventional
Fowler flap systems. The flap is about one-fourth to one-third the size of a con-
ventional flap. This means lower flap weight, and so fewer structural components
are required to hold
it
in place.8 The flap is also a simple hinged flap, rather than a
complex Fowler-type flap that requires complex gearing, tracks, and through
gaps, which most likely contribute to airframe noise on their own. The reduced
size and simplicity of the CCW system, even with a small flap, clearly offers
some advan tage over a conventional system.
There are many potential uses for circulation control. How ever, the two appli-
cations that have received the m ost research attention have been C C rotors (C CR )
and CCW applied to an aircraft for short takeoff and landing (STOL) capability.
The reader is referred to Refs. 1 and
5
where further details and citations on
CCW research can be found. Some research pertinent to the present work is
briefly mentioned below.
The Navy sponsored a full-scale flight-test program on an A -6/CCW in the
late 1970s.The design, tests, and results are documented in Refs. 9- 11.Research
has also been carried out to investigate applying the CC system to a B oeing 737
type of aircraft. A summary of the effort is docum ented in Ref.
6.
The only known
acoustic work on CC W configurations was performed by Salikuddin e t a1.2 There
are other otential uses for CC , including autom otive and in heli-
copter~,’,’~where noise reduction may also be appropriate. The acoustic benefits
shown in this paper should be applicable to other areas also.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 181/637
NOISE REDUCTION THROUGH CC 171
111. Facil it ies and Instrumentation
Th e anechoic flight simulation facility (AFSF) was used in the experiments. It
is located at Georgia Tech Research Institute (G TR I) located at its Cobb County
Research Facility in Smyma, Georgia. The AFSF operates in an open-jet wind-
tunnel configuration. It is an anechoic facility that allows acoustic measurements
to be made in the presence of a freestream (Fig. 2). Th e tunnel inlet has a square
inlet that converges down to a 28-in. round duct. The duct terminates in an anec-
hoic room as an open jet. Protruding out from the downstream wall is the collec-
tor, which is 4 ft wide by 5 ft high. Th e collector duct extends outside the building
and ends at a centrifugal fan powered by a diesel engine. The facility is open
circuit, drawing air from outdoors. The details of the facility can be found in
Refs. 14 and 15.
In the current experiments, the wings are mounted via mounting brackets to
the open jet. T his locates the wing across the je t opening immediately d own-
stream of the end of the duct. Figure 3 shows one of the conventional wings
mounted at the exit of the open jet. The ambient pressure in the chamber, the
plenum pressure for the slot, pressures in the air supply line venturi mass flow
meter, and pressure in the inlet (for freestream velocity) were monitored on indi-
vidual pressure transducers and manually recorded for each test point.
Acoustic measurements were made with B&K, 4135, 1 4-in. microphones.
One microphone was mounted on a traverse system that translated the micro-
phone from angles of 3 0 deg to 90 deg (where 0 deg is the freestream direction).
This system was arranged to make all measurements in the flyover plane. The
Collector
Fig. 2 Schematic of anechoic flight simulation facility (AFSF).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 182/637
172
S. E. MUNRO, K. K. AHUJA, AND
R.
J. ENGLAR
Fig.
3
Photo
of
a conventiona l wing mounted in AFSF.
microphone w as connected to a multichannel digital frequency analyzer, which is
run by software on a PC.
Figure 4 shows a schematic of the blowing system for the CCW . It consists of
high-pressure 3/4-in. tubing, a mass flow venturi, pressure gauges, and a muffler.
On the upstream end, the tubing is connected to an existing high-pressure line
with a control valve upstream. The flow passes through a mass flow venturi,
and then goes through more tubing to an in-house built muffler, which absorbs
the upstream valve noise. Downstream of the muffler, the air passes through
more tubing to inlets for the CCW plenum.
The test model wing used in Ref.
6
was used as the test model for this study.
This CC W model, shown in Fig. 5 has a supercritical baseline airfoil shape, but
has many different detachable CCW TE configurations. These included different
sized flaps and cylindrical trailing edges. Based on past aerodynam ic studies, the
best overall aerodynamic characteristics were obtained with the sm all CC W flap
configurations. The sm all deflectable flap allowed for low drag during cruise, but
by blowing over the curved upper surface with the flap deflected, significant flow
turning could still be achieved when desired. The highest lift configuration was
found to be w ith the flap deflected 90 deg. This was used a s the starting configur-
ation for the current acoustic tests.
The conventional wing had the same general shape as the CCW over most of
the chord. However, its trailing edge was altered with a cutout for a stowed flap.
A single-slotted Fowler flap was attached. Two different flaps were tested. The
flap was deflected 30 or 40 deg from the chord line to simulate a landing configur-
ation. Both flaps spanned the entire wing, but one flap had a cutout in at the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 183/637
NOISE REDUCTION THROUGH CC
173
1
High Pressure
Piping
Fig.
4
CCW blowing system configuration.
midspan point. Figure 6 shows the airfoil profile of the model and a drawing
depicting the flap cutout. The photo in Fig. 3 is of the model installed in the
AFSF. The cutout is to simulate the cutouts on a real aircraft. Cutouts are
often present for structural reasons or to prevent engine exhaust from impinging
on a lowered flap.
IV.
Technical Approach
The current work focused on optimizing a CC W system for low noise impact
while maintaining aerodynamic performance sufficient for direct comparison to a
conventional flapped-wing configuration. The first step was to determine if and
how a CCW configuration can have lower noise than a conventional system.
This step involved side-by-side comparison of representative configurations
Supercritlcsl Contour
Fig.
5
Schematic of CCW flap-wing configuration, generic supercritical airfoil
shape.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 184/637
174 S. E. MUNRO, K. K. AHUJA, AND
R.
J. ENGLAR
a
Fig. 6 a) Schematic
of
conventional flap-wing configuration, generic supercritical
airfoil shape;
b)
drawing
of
conventional wing with flap with cutout.
under the same conditions, that is, the same freestream flow and lift conditions.
Because there are several variations of CCW systems that have been researched,
a basic study of different CCW configurations was carried out. The test models
were used in other aerodynamic experiments, so this also allowed the use of
these data when making the acoustic comparisons.
Th e optimized blowing configuration was compared with a conventional wing
system. Basic noise spectra of the CCW and conventional wing configurations
were acquired at several mean flow velocities and angles of attack. Specific
cases where the different configurations had the same lift coefficient were then
com pared directly. Lift da ta from previous studies were used for this comparison.
V.
Results and Discussion
A. Acoustic Optimization of Existing CCW State-of-the-Art
Configurations
The C CW concept has been around fo r nearly 40 years, and there have been
many advances, changes, and modifications to the basic concept to improve its
overall performance. To attempt to test acoustically all the different configur-
ations would be unreasonable, because many of the changes were made
to improve the system. There is little reason to test acoustically a system that
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 185/637
NOISE REDUCTION THROUGH CC
175
is technologically surpassed by a better version. Thus, the goal of the current
study is to investigate two or three of the best performing CCW configurations.
Based on previous aerodynam ic work, the CCW with its flap deflected 90 deg
was chosen as the starting point for the study (a possible high-lift configuration
for landing approach). This had the best overall high-lift aerodynamic perform-
ance of several configurations tested in previous stud ies. The flap was eventually
adjusted to 30-deg deflection to prevent flap-edge vortex shedd ing noise that was
present in the 90-deg arrangement.
Six slot heights were chosen for the optimization study, ranging from 0.003 to
0.020 in. These dimensions were chosen because they w ere typical slot heights
used in earlier aerodynamic studies.6
A
wide range of slot Mach numbers was
evaluated, ranging from
0.3
to 1.2. The acoustically optimized CCW test con-
figuration was com pared with a conventional flap configuration. The conven tional
model had the same generic airfoil shape as the CCW, except near the trailing
edge to accom modate the conventional flap. The flap chord was about 30 of
the wing chord and deflected 40 deg to simulate a landing configuration. Data
were acquired for each test configuration at freestream speeds of 100, 150,
200, and
250
ft /s (nominal) and at geom etric angles of attack of 0,7, and 14 deg.
The m ajority of the data presented in this section was acquired at a geom etric
angle of attack of 0 deg and at the highest freestream velocity of about 240 ft/ s
unless otherwise noted. Figure
7
show s acoustic spectra for several slot velocities
with no freestream flow for the CCW with the 90-deg flap configura tion. It show s
a similar trend to the basic jet velocity scaling property develo ed for round jets.
For the measured velocities, V 8 scaling of jet noise theory' predicts about a
19 dB increase between the two m ost extreme cases, which is similar to that
Frequency,
kHz
(Af
=
32 Hz
Fig. 7 CCW blowing system noise spectra with no freestream flow; VT=
0
ft/s,
h
=
0.006
in.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 186/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 187/637
NOISE REDUCTION THROUGH CC
177
experience/theory on round jets16 this will provide for the jet mixing noise inten-
sity proportional to slot exit area. This translates into a 3 dB increase in noise
after shifting the spectrum for 0.006 in. to the left ove r the spectrum for
h
0.012
in. by a factor of one octave to allow for the shift in the noise frequen-
cies proportional to a characteristic length. This number is somewhat smaller than
the observed difference in the sound pressure levels (SPLs) of the two spectra in
Fig. 8.All of these arguments assume that we can apply the lessons learned from
round jets to very high
AR
jets. Yet, since the noise increase is of the order of
3
dB, it can be said that internal noise is not significant in this case. The fact
that the observed difference in spectral SPLs is more than the expected 3 dB
could also be associated with the scrubbing noise of the CCW slot jet moving
over the rounded edge. If so, it is genuinely produced outside and is not contami-
nated by any internal noise.
W e believe that the data may be contaminated by noise generated internal to
the wing below about 2 kHz. A muffler was built and installed in the supply line
downstream of all valves to eliminate as much upstream noise as possible.
However, due to the small thickness of the wing, inlets into the wing plenum
are smaller than desired. This results in a relatively high velocity flow entering
into the plenum, with no space to absorb the noise generated.
It is believed that these noise sources may be causing a majority of the noise
below 2 kHz where the noise is not following the typical
V8
et no ise scaling. For
the time being, this will be noted and data below
2
kHz will be disregarded as
either somewhat corrupted by internal noise or not understood.
Figure 9 shows the noise spectra for several slot jet velocities at a constant
freestream velocity and constant slot height of 0.003 in. There are several
things to note. First, with no blowing there is a large-amplitude, well-defined
tone. It is also important to note that, in general, the very low frequency
noise
f<
kHz, approx.) is much greater compared to the data in Fig. 7.
Some of this is from the tunnel noise itself (below about 500 Hz), but most
of it is flow noise associated with the freestream flow around the wing.
The tone is believed to be due to the shedding of vortices off the bluff trail-
ing edge of the deflected flap Stshedding 0.2 would produce a shedding fre-
quency of approximately 600Hz). Notice that blowing, even at low slot jet
velocities, significantly reduces the magnitude of the tone. However, in this
case
it
is not completely eliminated; in fact,
it
dominates the spectra at all
blowing velocities.
The aforementioned tone was unexpected. This presented a problem, because
the tone dominated the spectrum at all blowing conditions; thus, any acoustic
benefit derived from using the CCW over a conventional wing would be lost if
the flap were deflected to
90
deg. Because of this, it was decided that reducing
the flap deflection might produce a less dominant tone, but still provide
enough lift with the right amount of blowing to equal that of a conventional wing.
Figure
10
shows two curves with the flap set to 30 deg. In this case note that
the tone is completely eliminated with a small amount of blowing. The compu-
tational study also produced the sam e result, and is presented in Ref. 3 and also in
Chapter 22 of this volume. Not only is this advantageous for the current study, but
this result could be used in other applications where similar shedding produces a
distinct tone.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 188/637
78
S. E. MUNRO, K. K. AHUJA, AND R. J. ENGLAR
2
'fo
X
II
p
n
n
v
Frequency,
kHz (Af
=
32
Hz)
2
'p
z
X
II
b)
Frequency,kHz
(Af = 32 Hz)
Fig. 9 CCW with 90-deg flap and freestream velocity,
=
90 deg, V ,
=
220 ft/s,
a ) f =
0 60 kHz;b ) f =
0-5
kHz.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 189/637
NOISE REDUCTION THROUGH CC
179
Frequency,
kHz
(Af
=
32
Hz
Fig. 10
CCW
with 30-deg flap and freestream velocity,
=
90 deg,
VT= 220
ft/s,
f = -5
kHz.
Data for test conditions similar to those for the 90 deg deflection are shown in
Fig.
11.
Again, w ith no blowing the tone is present. However, with small amounts
of blowing, the separation is eliminated, and hence the tone is completely elimi-
nated. Because this configuration showed more promise, the remaining par-
ameters were optimized using the 30-deg flap configuration. Both slot height
and slot je t velocity were examined.
The effect of slot height was investigated next. Figure 1 2 shows data with
similar freestream conditions but different slot heights. It is important to note
that this figure compares different CCW configurations with the same lift. For
the same C, at different h the slot velocity will be different, because C, is depen-
dent on mass flow from the slot. The goal is to compare the same lift, so it is best
to look at the data where
C,
is constant, because the same
C,
will give the same
lift in most cases. There is som e variation of lift with
h
for high
C,,
but in the
C,
range of interest here, does not have an independent effect on the results. Thus,
the data in Fig.
1 2
show that there is a lower noise from the larger slot heights for
a given lifting condition. This makes sense, because C, is proportional to mass
flow through the slot. By increasing the slot height but maintaining the same
mass flow (and hence same
C,
the jet velocity of the slot is lower. At this
point it appeared that the most appropriate conditions for comparing a CCW
system to a conventional system had been found: maximize the slot height so
that jet velocity is minimized.
Unfortunately it was found that above a slot height of about 0.012 in. the noise
began to increase (for constant
C, .
This was contrary to the logical trend
associated with what should be happening, so some attention was given to
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 190/637
180
S. E. MUNRO, K. K. AHUJA, AND
R.
J. ENGLAR
2
X
Frequency, kHz
Af=
32 Hz
Fig.
11
CCW with 30-deg flap and freestream velocity,
=
90
deg,
V ,
=
220
ft/s,
f
=
0-60 kHz.
Frequency, kHz Af= 32 Hz
Fig. 12 CCW with 30-deg flap at three different h , p
=
0.04, = 90 deg,
VT
=
220 f t / s , f =
0-60
kHz.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 191/637
NOISE REDUCTION THROUGH CC
181
why this was happening. If one looks more closely at
C,,
it contains a mass
flow term. Initial results indicated that reducing the slot velocity reduced the
noise. In the equation this means that
V,
would decrease. If one defines the
mass flow term based on the mass flow “in” rather than “out” the problem
becomes evident:
Density will vary with the pressure in the plenum
( p
P / R T , but it varies pro-
portionally to slot velocity (as V, decreases, P decreases, and hence p decreases).
Area is constant in the plenum regardless of slot height. Thus, in order to offset
the decrease in V, and
p,
Vi, must increase. When this occurs, the internal noise
associated with internal velocities will also increase. Figure 13 shows overall
sound pressure level (OASPL) plotted against h for constant
C,.
If it is
assumed that the highest slot velocity is dominated by external jet noise, the
decrease in noise due to falling V, can also be plotted. In the figure the
highest V, occurs at the smallest
h.
The drop in OASPL should follow the V8
scaling law. However, in this case keep in mind that the slot velocity drops
due to an increase in slot area. Thus, the final estimated curve shows dropping
OASPL due to slot velocity, but a t a lower rate than
V8
because of an increase in
slot area.
Note that the experimental data follow V8 scaling for some time but eventually
increase away from the estimated dropoff. It is believed that this increase is due to
Fig. 13 OASPL for various p
=
constant.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 192/637
82
S. E. MUNRO, K. K. AHUJA, AND
R.
J. ENGLAR
the increasing dominance of internal noise as the slot velocity is reduced while
the internal velocity is increased.
Although this finding was unfortunate, it was not terribly detrimental to the
study as long as one keeps in mind that proper design of the internal system
will decrease the CCW noise further (in essence it should continue to drop
along the estimated slot velocity dotted curve in Fig. 1 3 as the slot velocity is
decreased). Thus, any benefit found will be enhanced with careful design of
the internal system.
B. Determining an “Equal Lift” Condition
The next step was figuring out how to compare the two lift augmentation
systems. Aerodynamic data from previous studies were used for this (specifically
from Ref.
6 ) .
Aerodynamic data were available for both conventional wing con-
figurations and the CCW in the form of lift curves
cl
vs
a
curves). This was
convenient, because for a CCW, a given
C
will generally provide a Acl over
the entire angle of attack range (not including the extreme high jet velocities
and large slots where the jet separates f rom the surface). Thus, once the lift for
the unblown CC W was found, this cou ld be com pare d to the c1 for the conven-
tional airfoil and the needed
Acl
could be calculated by subtracting the
two values. This Acl was then used to de term ine the
C
needed to match lift pro-
vided by the conventional wing flap system. Essentially, each C is analogous to
a flap setting that shifts the baseline lift curve by a given amount. For the
particular CCW configuration (CCW with flap at 30 deg), a C of about 0.04
produced about the same amount of lift as the conventional wings used in the
experiments.
C. CCW
vs
Conventional Wings
Two conventional wing configurations were tested: one configuration with a
30-deg flap spanning the entire span of the wing, and one with a flap deflected
40
deg spanning the entire wing except for a cutout region in the center span
(see Fig.
6
for a drawing; Fig. 3 for a photo of it installed in the AFSF). These
wings are the same basic airfoil shape as the CCW. The wings were tested at
the same flow conditions as the CCW.
Initially, the conventional wing with the 30-deg flap was tested. Figure 14
shows a comparison between the conventional wing with the 30-deg flap and
the CCW configuration with lowest noise for the equivalent lift case. Because
the
h
0.012 in. data were the m inimum CC W noise condition, they are pre-
sented in the figure. In the range between 1 and 10kHz, the CCW has noise
levels similar to those of the conventional system. Unfortunately, this was not
the desired result, although it does provide assurance that using the CCW
system does not increase the noise to the environment in its minimum noise
configuration.
However, many aircraft have a cutout in flaps across the span. Th is difference
contributes a fair share of noise to a conventional wing system, because flap edg e
noise has been identified as a major contributor to airframe noise. Thu s, this wing
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 193/637
NOISE REDUCTION THROUGH CC
183
h
n
v
Frequency, kHz
A f
= 32 Hz)
Fig.
14
CCW
and conventional wing
2-D
flap at similar lift condition,
=
90
deg,
V T
= 220 ft/s.
was missing a noise source that would most likely be greatly reduced in a CC W
system. Be cause the C CW flap is much smaller, there is no need for a gap in the
flap to avoid en gine exhaust. Its small size would a lso in many case s redu ce the
need for gaps due to structural concerns. Th us the CCW system with a full span
flap is not unreasonable.
Acoustic tests were performed on the new configuration, similar to the pre-
vious tests. Figure
15
shows the comparison of the wing with the cutout flap
with the CC W . As expected, the cutout in the flap increased the noise on the con-
ventional system significantly and shows a significant advantage to using a CC W
system in the region below 10kHz and some advantage up to 40 kHz. Beyond
40 kHz, the two systems have similar noise levels. The data in this figure and
following figures have different frequency ranges to emphasize the areas in
the frequency spectrum where there are differences between the two
systems. Similar results can be seen at other freestream velocities and angles
of attack; however, the magnitude of the difference varies some depending on
the conditions.
Up to this point, only d ata from a micropho ne at 90deg have been shown.
This is only part of the noise picture; the changes in directivity of the noise
between the two systems must be compared as well. Data were acquired at
30
60, and
90
deg. It should be noted that there are some differences depending
on the angle. Note that the 60 and
90
deg positions do not actually have a line-
of-sight path to the slot exit, which is located on the top surface of the wing. It
is also worth noting that the jet from the slot leaves the trailing edge of the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 194/637
184
S. E. MUNRO, K. K. AHUJA, AND
R.
J. ENGLAR
Frequency,
kHz
(Af = 32
Hz
Fig.
15 CCW
and conventional wing with cutout at similar lift condition, =
90
deg,
VT
= 220 ft/s.
wing at about 56 deg. Even with freestream velocity, the je t stays relatively
close to that angle for some time beyond the trailing edge of the wing.
Figure
16
compares the data for the two wing systems at 30 deg and
60
deg. At 30 deg the CCW system produces no real advantage over a conven-
tional system. However, there is still some noise reduction in favor of the
CCW system at
60
deg, similar to the
90
deg data shown earlier. These results
indicate that a CCW system certainly has potential for reducing airframe noise.
The results also show some trends of high-AR jets; however, there is still
much left to study and resolve before all the aspects of the CC wing noise
issues are solved and helpful to the design of a practical low-noise CCW system.
T o resolve som e of the questions brought up by the CCW and to eliminate the
possibility of internal noise contamination, a high-AR nozzle has been designed
and fabricated. This nozzle is presently being tested by the authors in an anechoic
facility and the intent is to produce a database of quality high-AR jet noise data
that can be used to verify the speculations about internal noise in the experim ents
presented here. In addition, these data will be used to augment the present results
by demonstrating the even greater benefits possible for a CC W high-lift configur-
ation in reducing airframe noise.
VI.
Conclusions
Following on from the great interest in reducing aircraft noise, an innovative
concept for eliminating a conventional flap system has been tested for its possible
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 195/637
a
NOISE REDUCTION THROUGH CC
Frequency, kHz
(Af = 32
Hz
185
Fig. 16
CCW
and conventional wing with cutout at similar lift condition,
VT
=
220
ft/s: a) =
30
deg,
b) = 60
deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 196/637
186
S. E. MUNRO, K. K. AHUJA, AND
R.
J. ENGLAR
acoustic advantages. Previous studies have show n that the CC wing is an aerody-
namically viable alternative for conventional mechanical flaps. This study show s
that there is also a substantial advantage in the acoustic realm. The results pre-
sented showed a lower noise spectrum for a CC W system compared to a conven-
tional system for the same lifting condition. It should be noted that even if the
CCW produces noise comparable to that of a conventional wing it is an advan-
tage. This is because a CC W is expected to be much lighter than a conventional
wing.
It was also noted that the internal noise of the CCW blowing system of the
model inhibited finding the full possible advantage a C CW system can offer. It
is believed that careful design of a CCW blowing system, including internal
details, could further improve the results shown here.
Acknowledgments
This work was sponsored by NASA Grant NAG1-2146 through NASA
Research Center Langley, under its Breakthrough Innovative Technology
Program. The authors are grateful to L. Sankar of the AE school for many
helpful discussions. Thanks are also due to C. Jameson for designing the
HARN nozzle and to Rick Gaeta for his assistance in the experiments and
many useful discussions.
References
‘Englar, R. J., “Circulation Control Pneumatic Aerodynamics: Blown Force and
Moment Augmentation and Modification; Past, Present & Future,” AIAA Paper 2000-
2541, June 2000.
*Salikuddin, M., Brown, W. H., and Ahuja, K. K., “Noise from a Circulation Control
Wing with Upper Surface Blowing,”
Journal
of
Aircraft, Vol.
24, No. 1, 1987.
3Liu, Y., Sankar, L. N., Englar, R. J., and Ahuja, K. K., “Numerical Sim ulations of the
Steady and Unsteady Aerodynamic Characteristics of a Circulation Control Wing,” AIAA
Paper 2001-0704, Jan. 2001. Also, see Chapter 22 of this volume.
4Dunham, J., “A Theory of Circulation Control by Slot-Blowing Applied to a C ircular
Cylinder,”
Journal
of
Fluid M echanics,
Vol. 33, No. 3, 1968, pp. 495-514.
5Englar, R. J., and Applegate, C. A., “Circulation control-A Bibliography of
DTNSR DC Research and Selected Outside References: January 1969 through December
1983,” David W . Taylor Naval Ship Research and Development Center, DTNSRDC-84/
052, 1984.
6Englar, R. J., Smith, M . J., Kelley, S. M., and Rover, R. C., 111, “Development of Cir-
culation Control Technology for Application to Advanced Subsonic Transport Aircraft,”
AIAA Aerospace Sciences Meeting, AIAA Paper 93-0644, Jan. 1993.
’Englar, R. J., and Huson, G. G., “Development of Advanced Circulation Control
Wing High Lift Airfoils,” AIAA Applied Aerodynamics Conference, AIAA Paper 83-
1847, July 1983.
Englar, R. J., “Low-Speed Aerodynamic Characteristics of a Small, Fixed
Trailing-Edge Circulation Control Wing Configuration Fitted to a Supercritical Airfoil,”
David W. Taylor Naval Ship Research and Development Center, DTNSRDC/ASED-
81/08, March 1981.
8
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 197/637
NOISE REDUCTION THROUGH CC 187
’Nichols, J. H., Jr., Englar, R. J., Harris, M. J., and Huson, G. G., “Experimental Devel-
opment of an Advanced Circulation Control Wing System
for
Navy STOL Aircraft,”
AIAA Aerospace Sciences Meeting, AIAA Paper 81-0151, Jan. 1981.
“Pugliese, A. J., and Englar, R. J., “Flight Testing the Circulation Control Wing,” AIAA
Aircraft Systems and Technology Meeting, AIAA Paper 79-1791, Aug. 1979.
“Nichols, J. H., Jr. et al., “Development of High Lift Devices for Application to
Advanced Navy Aircraft,” DTNSRDC, Rept. DTNSRDC-80/058, AD A084-226, April
1980.
”Lane, P. Jr., “Ground Controls,”
Racecar Engineering,
Vol. 9, No.
8,
1999, pp. 20-23.
13Reader, K. R., “Hover Evaluation of the Circulation Control High Speed Rotor,”
David W. Taylor Naval Ship Research and Development Center, Rept. 77-0034, June
1977.
14Ahuja,
K. K.,
Tanna, H.
K.,
and Tester, B. J., “An Experimental Study of Trans-
mission, Reflection and Scattering of Sound in a Free Jet Flight Simulation Facility and
Comparison with Theory,” Journal of Sound and Vibration, Vol. 75, No. 1, 1981,
15Ahuja,
K. K.,
Teste r, B. J., and Tanna, H.
K.,
“The Free Jet as a Simulator of Forward
Velocity Effects on Jet Noise,” NASA Contractor Rept. No. 3056, 1978.
16Ahuja,
K. K.,
and Bushell,
K.
W.,
“An
Experimental Study of Subsonic Jet Noise and
Comparison with Theory,”
Journal
of
Sound and Vibration,
Vol. 30, No. 3, 1973,
”Tarn, C.
K.
W., and Zaman,
K.
B. M. Q., “Subsonic Jet Noise from N on-Axisymmetric
18Tam, C.
K.
W., and Auriault, L., “Jet Mixing Noise from Fine Scale Turbulence,”
‘’Tarn, C.
K.
W., Golebiowski, M., and Seiner, J. M., “On the Two Components
of
”Tam, C.
K.
W., “Influence of Nozzle Geometry on the Noise of High Speed Jets,”
’lAhuja,
K. K.,
“Correlation and Prediction of Jet Noise,” Journal of Sound and
pp. 51-85.
pp. 317-341.
and Tabbed Nozzles,” AIAA Paper 99-0077, 1999.
AIAA Paper 98-2354, 1998.
Turbulent Mixing Noise from Supersonic Jets,” AIAA Paper 96-1716, 1996.
AIAA Paper 98-2255, 1998.
Vibration,
Vol. 29 No. 2, 1973, pp. 155-168.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 198/637
1I.B.
Experiments and Applications: Aerospace
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 199/637
Chapter 7
Pneumatic Flap Performance for a Two-Dimensional
Circulation Control Airfoil
Gregory
S
Jones*
NASA Langley Research Center, Hampton, Virginia
Nomenclature
A, =
effective cross-sectional area of two-dimensional model
b
=
airfoil two-dimensional span, in.
C = pressure coefficient
c = airfoil chord, in.
d=
section profile-drag coefficient
Cl
= section lift coefficient
C,
cos
a C
sin
a)
C = moment coefficient
C = normal force coefficient
C = fluidic pow er coefficient
CT
=
thrust coefficient
=
C
C =
momentum coefficient = rizuj /q(wc))
D = drag, lbf
h = slot height of Coanda jet , in.
H = tunnel height, in.
L
= lift, lbf
M = mach number
riz = mass flow, lbm /s
Z,J,K = pressure tare coefficients for balance
*Research Scientist, Flow Physics and Control Branch. Senior Member AIM
Copyright
005
by the American Institute of Aeronautics and Astronautics, Inc.
No
copyright is
asserted in the United States under Title 17, U.S. Code. The U.S. Government has a royalty-free
license to exercise all rights under the copyright claimed herein for Governmental purposes. All
other rights a re reserved by the copyright owner.
191
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 200/637
192 GREGORY S. ONES
NPR = nozzle pressure ratio =
P,/P,)
f= fluid power, ft.lb/s
P = pressure, lbf/in.2 or lb f/ft2
p’
=
fluctuating pressure, lb f/in.2 or lbf/ft2
q
= dynamic pressure, lbf/ft2
=
pU 2
r
= trailing edge radius, in.
s
= airfoil reference area, ft2
T = static temperatu re, OR
t
= airfoil thickness, in.
U = velocity, ft/s
u’ =
fluctuating velocity, ft/ s
w = slot width, in.
a
=
angle of attack, deg
Sjet
= reactionary force angle, deg
p = Prandtl-Glauert compressibility
d m )
Ojet = Coanda jet separation angle, deg
E = blockage interference ratio, u / U
p
= density, lbm/ft3
r= circulation
Subscripts
jet, j = conditions at slot exit
rake = conditions at rake location
ram = conditions at engine inlet
AOA = angle of attack, deg
o = stagnation or total conditions
= freestream conditions
I. Introduction
IRCULATION control (CC) technologies have been around since the early
C
930s, and have been successfully demonstrated in laboratories and flight
vehicles alike, yet there are few production aircraft flying today that implement
these advances. These technologies are generally related to pneumatic devices
falling into categories including jet flaps, blown flaps, and Coanda surfaces.
Recent interest in CC aerodynamics has increased for both military and civil
applications, with em phasis on p roviding better vehicle performance and predic-
tion capability.’ The history of Coanda-driven CC has met with varying degrees
of enthusiasm as the requirements for improved high-lift systems continue to
increase. Current lift coefficient goals for extremely short take-off and landin
(ESTOL) vehicles are approaching 10 and lift-to-drag ratios greater than
25.
Personal air vehicles (PAV) have a field length goal of 250 ft.3 To achieve
these goals will require more than what a conventional high-lift system can
provide. In addition to high-lift and cruise drag requirements, the next generation
of aircraft will need to address other issues, including weight4 and noise.5 Con-
ventional high-lift systems that use flaps and leading edge (LE) slats can be
associated with significant weight and volume penalties of a typical wing assem-
bly. These assemblies are also complex (up to 3 and
4
subelements) and very
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 201/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
193
sensitive to location relative to the main element of the wing. The need to sim-
plify and reduce the weight of these systems without sacrificing performance is
the focus of this effort.
Coanda-driven CC techniques generally offer high levels of lift for small
amounts of These systems are perceived to be simpler and less
weighty than conventional high-lift systems. However, advanced system
studies of CC systems being applied to m odem aircraft have been limited o r non-
existent, and so the ability to buy its way onto an aircraft is generally unproven.
Nevertheless, several blocks to real aircraft applications reappear in every discus-
sion of CC. These include, source of air (typically bleed or bypass air from the
engine or added auxiliary power unit), unknown weight penalties related to the
internal air delivery system, engine out conditions, drag penalty associated
with blunt trailing edge (TE), and large pitching mom ents associated with aircraft
trim. Although this is not a comprehensive list, these issues will be used as a
guide in developing a C C wing for general aviation applications.
A primary objective of this effort is to evalua te the benefits of pulsed C C and
to reduce the mass flow requirements for a given lift performance as well as to
reduce the cruise drag penalty associated with a large CC trailing edge. Second-
ary objectives of this study were to evaluate the dual blown pneum atic concept as
a control device and to determine potential benefits of returned thrust (i.e., thrust
is lost at the engine due to bleeding mass from the eng ine, so how much thrust is
returned to the aircraft through the wing).
11.
NASA CC Requirements
Application of CC to different aircraft platforms is driven by requirements that
are dictated by mission.’ The National Aeronautics and Space Administration
(NASA) Vehicle Integration, Strategy and Technology Assessment (VISTA)
office describe many of these missions. Each of the vehicle sectors within the
VISTA program could benefit from CC technologies, but personal air vehicles
(PAV ) and ES TOL vehicles seem to benefit the most. The personal air vehicles
shown in Fig.
1
have characteristics that resemble general aviation vehicles but
meet stiffer requirements for field length (i.e., high lift), noise signatures, and
cruise efficiency
( L / D ) .
With a fresh look at point-to-point travel, NASA’s
PAV program will address airport infrastructure, ease of use, and reductions in
the cost of travel.
Today’s small aircraft utilize significantly oversized wings for cruise and
simple hinged flaps for high lift. These systems are adequate for the current
Fig.
1
Notional concepts
of
NASA
personal air vehicles.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 202/637
194 GREGORY S. ONES
airport infrastructure. However, as these airport requirements become more strin-
gent, high lift and cruise efficiency must be improved. The PAV goals used for
this effort included a
250
ft field length that will require resizing the wing with
a
C
= 4.0,
yielding an
L/D,,,
of
20.
In the near term, reduced approach
speeds enable a 1000 t field length a nd ca n improve safety in addition to redu-
cing community noise signatures. If equivalent control margins and gust sensi-
tivity are achieved, safety (in terms of accident avoidance reaction time and
survivability) is proportional to the approach speed. These reduced speeds
require more efficient high-lift systems. Circulation control technologies have
been identified as a candidate simplified high-lift system. It may be necessary
to integrate this system with other active flow control technologies (combining
higher altitude cruise, gust alleviation, limited powered-lift, and
so
on).
Air sources for C C systems for small aircraft may have a low penalty. Current
high-performance small aircraft are turbocharged for altitude compensation. At
landing and takeoff conditions, compressed air is thrown out the wastegate of
the turbocharger (approx.
2’
lbm /s). Th is is a potential sourc e for air augmenta-
tion to a CC system. Because engine out conditions are an issue for CC appli-
cations, another air source alternative is using the wake vortex energy to
power a wingtip-turbine. Regardless of the air source, it is important to optimize
the efficiency of the CC system for minimizing mass flow at a given lift require-
ment. The NASA ESTOL vehicle sector requirements are
directed to a
100-passenger class vehicle that would include the following elements:
1) 52000 ft balanced field length (related goal of
C
= lo); 2 ) cruise at
M
=
0.8; 3) noise footprint contained within the airport boundary; and
4)
anding speed
-50
kt. The current state-of-the-art aircraft systems can only
achieve two or three of these elements simultaneously. Circulation control has
the potential of enabling the achievement of all the elements of the desired capa-
bility set and could be integrated to the high-lift, flight controls, and propulsion
systems as shown in a notional aircraft in Fig. 2 . It is recognized that the
integration of the propulsion system and the wing is paramount to the success
LEADING EDGE
Active FlowControl
(High
Lf l )
Fig.
2
Notional concept of NASA ESTOL 100-passenger vehicle showing potential
CC
vehicle.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 203/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
195
of either of these vehicle concepts. The focus of this chap ter will be targeted at a
two-dimensional baseline CC airfoil proposal that could be applied to the outer
wing panel of either concept.
111. Theoretical Considerations
The two-dimensional aerodynamic performance is traditionally categorized
into lift, drag, and pitching moment elements. Most fluid mechanics devices
that alter the forces on a body are characterized into two force categories:
1)
induced forces resulting from circulation and 2 ) reaction forces caused by
jet momentum. This section will focus on lift and drag forces associated with
active flow control systems that utilize pneumatic flow control. Pneumatic or
blown active flow control systems can be related to boundary layer control
(BLC) and/or supercirculation modes. These modes are often characterized by
the fluidic power required to achieve the performance augmentation.
To achieve the maximum performance on a body, it is desired to drive the
stagnation streamlines toward the equivalent inviscid ~olution.~ractically, this
is achieved by moving the boundary layer separation to the TE. This is the
performance limit for BLC techniques. To achieve supercirculation
it
is necess-
ary to extend the effective TE beyond the physical TE location with a virtual or
pneumatic flap, as simulated in Fig. 3.
To understand the lim its of airfoil performance, i t in necessary to be aw are of
the inviscid lift characteristics. The influence of the airfoil thickness on the
maximum theoretical inviscid lift coefficient (not including jet thrust or
camber effects) can be described as
c,= =
2 4
+
:)
For a limiting case of t / c of
100%
(i.e., circular cylinder), the maximum lift
coefficient is 4.rr and can be related to classic unblown circulation rc round
the body
lo:
L
=
pure
( 2 )
Fig.
3
CFD simulation
of
pneumatic flap and streamline tuning using a Coanda jet.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 204/637
196 GREGORY S. ONES
The magnitude of the circulation r, s a function of geometry alone and will be
referred to as induced lift and can be related to the modified pressure on the inte-
grated boundary of the body:
257
L
= S p r s i n e d e
3)
Recall that for an inviscid solution (circular cylinder), the normal force is solely
directed in the vertical plane and that drag is zero. As seen in Fig.
4,
the stream-
lines are significantly influenced by the m agnitude of the circulation r,. n prac-
tice, the inviscid limit is never reached because of flow separation. However, for
an airfoil employing a BLC or a CC device, the m aximum inviscid lift is possible.
When a pneumatic system that adds mass is used, an additional circulation
term is added to the induced circulation to account for the reactionary forces
produced by the jet, as described in Eq. (4) :
L
= ~ u ( r cq e t )
(4)
where
r,,,
=
EE a+
8jet)
PUW
and can be related to lift and drag as
CYLINDER MAPPED
INTOAIRFOlL
LE STAGNATION=
TE STAGNATION
LE
STAQNATDN
TE
STAGNATION
Fig. 4 Classic lift resulting from circulation for a circular cylinder and mapped into
airfoil profile.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 205/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
197
This reactionary force term can affect lift or drag depending on the orientation of
the jet exit angle Sj,,) at the boundary of the body. F or pneumatic systems this
reactionary force should not be confused with the thrust vectoring that an articu-
lating nozzle generates on an eng ine nacelle. The reactionary force that is charac-
teristic of a pure jet flap is at a fixed jet angle, as shown in Fig. 5
The efficiency of a pure jet flap (typically vectored normal to surface), com-
pared to typical C C airfoils (vectored tangential to the upper surface), is realized
in the differences in the induced effects that accompany the pressure field. It is
recognized that both of these airfoil techniques benefit from induced forces
and reaction forces that can be correlated to jet position and orientation. Nomin-
ally, je t flap airfoils depend largely on the reaction force of the je t m omentum.
Coanda-type CC systems capture the induced forces more efficiently and
typically deliver larger lift gains than a pure jet flap.
The combined induced circulation and reactionary forces are generally cap-
tured experimentally with a balance, integrated surface pressures, and/or
wind-tunnel wall pressure signatures combined with wake rake pressures. The
force balance is a direct measure of both induced circulation and reaction
forces. Because these forces
are
integrated and summed at the balance, the
ability to decompose the induced and reactionary components are dependent
on knowing the vectored force associated with the jet . Integrated surface press-
ures are representative of induced circulation forces alone. To obtain the total
forces along the boundary of the body, reactionary forces must be added at the
appropriate
S
angle. The integrated wind-tunnel wall signature and wake
rake must also account fo r the reaction forces generated by the jet .
For typical CC systems, the jet exit is nominally directed aft, resulting in a
reactionary thrust force that contributes very little to lift (except when an aft
camber causes a small S as shown in Fig.
6
It should be recognized that the
benefit of turning the flow with the wall bounded jet along the Coanda surface
is reflected in the two-dimensional induced circulation found in the modified
surface pressure field.
The reactionary force of the C C system augments the thrust produced by the
primary propulsion system (Fig. 7). Returning a portion of the thrust that was bled
from the engine to supply the C C subsystem reduces the overall system penalty
associated with CC. The recovery of this thrust will be dependent on the effi-
ciency of the Coanda nozzle and internal losses of the CC air delivery system,
and so on.
PURE JET FLAP
Fig. 5 Thrust vectoring using a classic pure jet flap.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 206/637
198 GREGORY S. ONES
Fig.
6
Schematic
of
flow angles associated with typical Coanda-drivenflow.
It is known that nozzle efficiency is very dependent on nozzle aspect ratio
(AR). Propulsion system studies of rectangular nozzle losses are generally
limited to ARs less than 10. Because there is no database for large-AR nozzles
( h / w
>
1300, similar to those used in CC airfoils), it would not be practical to
extrapolate to obtain thrust recovery. However, for this tw o-dimensional study
(where nozzle AR is meaningless), it is appropriate to neglect the nozzle effi-
ciency and assume no losses. For two-dimensional CC studies the thrust can
be described at the jet exit of the airfoil by the momentum or thrust coefficient:
Fig. 7 Block diagram
of
reactionary forces for an integrated wing and propulsion
system.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 207/637
where
and
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
199
m
=
p et U.,
thw
9)
Th e tradeoffs of engine thrust against reduced engine thrust augmented with
CC thrust will involve detailed specifications of the geometry of the airfoil, the
intake lip, internal diffusers, ducting, compressor, and jet-nozzle designs.
Obviously the results would be applicable for that design only. In the absence
of these details, some general estimates of the benefits or penalties of CC
system s can be formulated by estimating the power requirements of CC.
For a crude estimate of fluid pow er
(Pf),
t is assumed that the je t is taken fro m
a large reservoir. The total power expended will then be at least equal to the
power required to supply the jet velocity head Pjetplus the power lost at the
intake as the fluid is drawn into the large reservoir Pram. his ideal power can
be described as1*
where
1 m
Pjet pu2
2 J P
and
Pram mu: 13)
Hence, the power (ft . lb/s) required supplying a flow with a total momentum
coefficient
C,
is
Pf = c,u”uc a
[
+
2( )2] (qcoucos )
and nondimensionally
(14)
If the je t slot height h is constant and is known for a rectangular wing, the fluid
power can be expressed in terms of just the parameters C, and height-to-chord
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 208/637
GREGORY
S.
ONES
00
ratio
h l c :
Figure
8
shows the nondimensional ideal power for a typical CC jet orifice.
A.
Two-dimensional drag characteristics for blown airfoils are often complicated
by the juncture flow created by the wind tunnel and airfoil model. To avoid these
issues, the most reliable measurement technique for experimentally determining
the drag of a blown airfoil is the momentum-loss method that employs a wake
rake and is described in de tail by Betz13 and Jones.14 The profile drag can be
determined by integrating the wake measured one to three chords
downstream of the TE
Two-Dimensional Drag with Blown Systems
For blown airfoils, it is important to note that the measured profile drag from a
wake rake must be corrected by subtracting the momentum that was added by the
CC
system.” The total horizontal forces on a two-dimensional model do indeed
1.2 -
1.0
-
0.8
w
0.6
0.4
-
0.2
-
Power
0.00 0.05 0.10 0.15 0.20
CP
Fig. 8 Ideal power requirements for typical Coanda jets having different jet exit
heights.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 209/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
201
exceed that indicated by conventional wake rake calculations by the quantity
riz U,. Considering a frictionless hypothetical case w here the jet is exhausted at
a total head equal to freestream total head easily confirms this principle. Here,
the wake will indicate zero drag, but the model will experience a thrust of
mu,.
The way the net forces are book kept results in
This is equivalent to what a force balance would measure, assuming that the air
source is considered to be internal to the model.
B.
Equivalent
Drag
To make direct comparisons of different blown systems such as traditional C C
airfoils, je t flaps, blown flaps, engine augmented pow ered lift systems, and so on,
it is necessary to define an equivalent lift-to-drag ratio. For powered airfoil
systems, the system efficiency should contain the effects of the energy that is
required to obtain the airfoil performance. This also avo ids the infinite efficiency
that would occur when the drag goes to zero as a result of blowing. A correction
can be made through an equivalent “kinetic energy” drag coefficient that is
related to the power described previously. This equivalent drag can be described as
Dequiv
=
Dprofi le -k Dpower + D r a m + Dinduced
where
Dpr o f i l e
is the profile drag,
Dpower
s fluid power, Dram s momentum drag
force required to ingest the blowing flow rate at the engine inlet, and Dinduced
is induced drag (equal to zero for two-dimension). For two dimensional flows,
the equivalent drag becomes
mq?
m
Dequiv drag
+
pU
2uc.a P
and
19)
The practical implementation of the Betz and Jones wake integration techniques
for blown systems is described in Ref.
18.
When the rake drag coefficient is
applied to the equivalent drag, it becomes
It should be noted that the kinetic energy or power that is added to the equivalent
drag, dominates the equation and leads to drag values that are not practical, and
masks the thrust generated by a typical CC airfoil.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 210/637
202 GREGORY S. ONES
C.
Mass Flow
Requirements
To optimize the performance of a CC system at the lowest mass flow, it is
necessary to recognize the relationships between mass flow,
C,,
and slot geo-
metry. Figure 9 highlights this relationship for a given freestream condition
and geometry, which is consistent with experiments described in this report.
Assuming that the performance is dominated by the jet velocity ratio, reducing
the slot height would result in a lower mass flow requirement.
IV.
GACC Airfoil Design
The General Aviation Circulation Control (GACC) wing concept was initially
developed for PAV19 and is now being considered for the ESTOL concept
described previously. To address the requirements of PAV, the airfoil design
and initial performance goals of this wing concept were as follows:
1)
To achieve two-dimensional
Cl
= 3 using a simplified Coanda-driven CC
trailing edge.
2) To provide a pneumatic flap capability that will minimize cruise drag and
provide potential roll and yaw control (dual blowing is defined as upper and lower
Coanda surface blowing). This is based on closing the wake of the bluff TE
associated with typical blunt Coanda surfaces.
3)
To provide the capability to change the C oanda surface shape (e.g., circu-
lar, elliptical, and biconvex).
4) To provide pulsed pneumatic control to minimize the mass flow require-
ments for high lift.
5 ) To provide distributed flow control to customize the spanwise loading on
the airfoil.
To establish a relevant CC airfoil geometry that is readily available to the aero-
dynamic community (not restricte2 due to proprietary issues) and that has the
2bO
200
I 1 5 0
100
50
0
0.0 25.0
50.0
75.0 100 0 125.0
Ujet/uop
Fig.
9 CC
mass flow requirements, chord
=
9.4 in., q = 10 psf,
To =
75°F.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 211/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 212/637
204
GREGORY
S.
JONES
GACC AIRFOIL
PROFILE
BLUNT
LEA DING EDGE
RADIUS-1.93%
Fig.
10
Seventeen-percent thick
GACC
profile with circular trailing edge.
problem occurs beyond the target lift coefficients of 3, so LE control will not be
addressed for this study.
It was decided to modify the GA(W )-1 with Coanda-type TE s by altering only
the aft lower section of the original airfoil. The original GA(W )-1 chord line was
used as the reference for angle of attack (AOA) on the GAC C airfoil design, as
shown in Fig. 10. The tradeoffs of sizing the Coanda surface can be related to
optimizing the lift and drag for high lift or cruise conditions.25926 ominally, a
larger TE Coanda radius of curvature would lead to a higher CC lift coefficient,
as well as a higher cruise drag as a result of an increase in the TE diam eter. The
shaded area shown in Fig. 11 highlights the region of effective Coanda turning
and proven lift performance highlighted by the A-G/CCW flight dem~nstrator.~’
The A-6/C CW airfoil2’ was a 6% thick supercritical wing section that incorpor-
ated a state-of-the-art large circular TE radius of 3.67% chord. This large TE
functioned to guarantee a successful flight demonstration of the high-lift
0.000-
0.000 0 001 0.002
0 003 0.004 0.005
h / C
Fig. 11 Effective Coanda performance for d ifferent radius and jet slot heights.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 213/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL 2 5
system2* only. Any operational use of this design would require a mechanical
retraction of the C C system in to the w ing to avoid a large cruise drag penalty.
To minimize the GA CC airfoil drag performance without the use of a mech-
anical system a dual-blowing pneumatic concept with a small radius TE was
designed. A baseline circular
r / c
of 2 was chosen for the G ACC . Three differ-
ent TE shapes were designed to be interchangeable and integrate with the G ACC
model, as shown in Fig 12.The distance between the slots remained fixed and
used the circular shape as a baseline. Both the elliptic and biconvex shapes
extended the chord by 1 (0.174n.). The 2:l elliptic shape reduced the Y/C to
1 and the biconvex shape had an
Y/C
of 0.
To com pare steady, pulsed, and dual blowing using a common model required
careful design of the internal flow path, as shown in Fig. 13.The ability to inde-
pendently control the upper and lower slot flow enables the investigation of both
positive and negative lift as well as drag and thrust for both high-lift and cruise
conditions. A pulsed ac tuator system was integrated into the upper plenum of the
model for investigation of unsteady circulation control.
To obtain a uniform flow path and create a two-dimensional flow environment
at the Coanda surface it was necessary to carefully design the internal flow path of
all three air sources in the model, a s shown in Fig. 14.Twenty actuators were dis-
tributed in the upper plenum along the span to optim ize the pulsed authority to the
VARIABLE
UPPER SLOT
2 : l
BI CONVEX
Fig.
12
Sketch of three interchangeable
TE
shapes for the GACC airfoil.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 214/637
206
UPPER STEADY
MANIFOLD
GREGORY S. ONES
ACTUATOR UPPER
DIFFUSER SLOT
LOWER STEADY
MANIFOLD
Fig. 13 Sketch of internal
flow
path of the GACC airfoil.
upper Coanda je t for the high-lift mode. A ir for all three sources was fed from
one end of the model and was expanded into large plenums then channeled to
the trailing edge je t exit. Both the upper and low er slots were adjustable
(0.005 < h < 0.025) and were fed from a smooth contraction that had a
minimu m area ratio of
10.
It is difficult to create an infinite or two-dimensional environment with a
fixed-wall wind tunnel for blown airfoil systems. One must consider the rela-
tive size of the model to the size of the test section and the expected trajectory
of the jet created by the blown system. To minimize the impact of the wind-
tunnel interference for CC systems, several experimental design considerations
were considered: Solid Blockage (physical chord and span related to wind-
tunnel cross-section), Wake Blockage (how much streamline turning will be
20ACTUATORS
wlDlFFUSERS
INSTRUMENTED
TRAILING EDGE
COANDA
SURFACE
Fig. 14 Sketch
of
GACC model with upper skin removed to highlight the flow path
and instrumentation of the upper plenum.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 215/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
207
achieved with blown system), and
Juncture
flow regions (aspect ratio of
model).
The GACC model was sized and built for the NA SA Largley Research Center
(LaRC) Basic Aerodynamic Research Tunnel (BART) and had a chord-to-test-
section height ratio of 0.23, an aspect ratio of
3
based on a chord of 9.4 in. and
a two-dimensional wall-to-wall span of 2 8 in. These values are conservative
for the unblown c ~ n f i g ur a t io n ,~ ~owever, once blowing is applied, the influence
of the Coanda jet on streamline turning could be significant. A two-dimensional
RANS code (FUN2D) was used to evaluate the streamline turning related to
Coanda blowing and supercirculation high-lift condition^.^ The free ai r results
of this preliminary C FD evaluation indicated streamline turning and wake deflec-
tion would not impact the tunnel walls for the BA RT test conditions but would be
influenced by the presence of the solid tunnel walls. The study of wall interfer-
ence is ongoing for this experiment.
V. Experimental Setup
Experimental results have been obtained for a GAC C airfoil in the open return
Langley BAR T, as seen in Fig. 15. The tests were conducted over a Mach number
range of 0.082 to 0.1 16 corresponding to dynamic pressures of 10 and 20 psf,
respectively. Lift, drag, pitching moment, yawing moment, and rolling moment
measurements were obtained from a five-component strain gauge balance.
Drag data were also obtained from a wake rake. Airfoil surface pressure measure-
ments (steady and unsteady) w ere used to highlight boundary layer transition and
separation.
A block diagram of the BA RT data acquisition is shown in Fig. 16. T o capture
the transients and time-dependent characteristics of the pulsed flowfield two
approaches were developed: arrayed thin films and miniature pressure trans-
ducers. This report will focus only on the miniature pressure transducers. The
small scale of the model did not lend itself to using off-the-shelf pressure trans-
ducers. Custom differential pressure gauges were designed and fabricated using
GACC
CHORD
9.4
Fig.
15
Sketch of the GACC setup in the Basic Aerodynamic Research Tunnel.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 216/637
208 GREGORY S. ONES
UNSTEADY
P+p’
Fig. 16 Block diagram of BAR T data acq uisition for G ACC setup.
MEMS sensors attached directly to the skins of the model leading and trailing
edges. These transducers were not temperature compensated, making real-time
calibration necessary. To keep the measured errors from exceeding
0.05
of
the full scale (2 psid), a reference pressure was monitored and calibrations
were performed when necessary. This was also the case with the ESP system
for ten independent 32-port modules with ranges of 10 in.
H20
psid, and
2.5 psid.
The five-component strain gauge balance was also custom designed and fab-
ricated for the GACC model. Normal, axial, pitching moment (ref. 50% chord),
rolling moment, and yawing moment lim its are shown in Table
1.
A drawback to
the GACC balance was that the axial resonance of the balance/model system was
too close to the dynamics of the loaded airfoil, resulting in vibration of the model.
Table 1 GAC C balance limits
Normal, Axial, Pitching Rolling Yaw ing
lbf lbf mom ent, in. lbf mom ent, in. lbf mom ent, in. lbf
100
1 1600
400
40
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 217/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
209
This vibration did not always exist, but led to larger than expected errors in the
axial force measurement. Therefore the drag data will be reported only from the
wake rake results. The
GACC
model has three plenums, which are required for
use in different modes of operations (e.g., high-lift, cruise, pulsed, and
so
on).
Each plenum is supplied with air that is independently regulated, as shown in
Fig. 17. To achieve the potential mass flow requirements for the largest slot
area, a 2000 psia high-pressure external air source (3000 psia max) was used.
The air is preheated to com pensate for Joule Thompson effects and temperatures
are maintained to within 1 R.
The mass flow was measured with three independent turbine meters. These
flow meters are precalibrated and compensated for density variation at the
point of measurement (accuracy
= 1%
reading). The high-pressure plenum
that supplies the pulsed actuation system is buffered with a 7.1 ft3 air tank to
eliminate the pulsed backpressure flow at the control and flow measurement
station. The pressure limits of each of these systems were driven by the pressure
ratio at the slot exit.
As
a result of pressure losses in the system the upper and
lower plenums were limited to
50
psid and the actuator pressure was limited to
200
psid. These limits enabled sonic capability at the slot exit.
A
trapeze system was used to couple the air delivery system to the model as
shown in Fig. 18. Special attention was given to the calibration of the balance due
to the number of airlines that cross the balance. U npressurized calibration results
are applied to a
6
x
21 calibration matrix and account for the linear interactions
(first order) and the second-degree nonlinear interactions of the balance.30931ach
pressure line was then independently loaded and characterized with no flow (see
appendix to this chapter). With the model mounted vertically in the tunnel, the
only loads experienced by the model as a result of the air delivery system were
thrust loads along the span of the model. This is the same as the side-force, which
is not gauged or measured. The flexible hoses maintain a vertical orientation to
the model and eliminate horizontal forces being applied to the balance.
VOLUME
BOOSTER
Fig.
17 GACC
air de livery system having three independent air supply lines.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 218/637
210 GREGORY S. ONES
LOWER
JET
AIR
SUPPLY
Fig. 18 GACC balance and model interface with air delivery through trapeze
system.
Measurement of the drag was initially obtained with the balance and reported
in Ref. 19. How ever, upon careful inspection of the issues related to junc ture flow
interference and balance vibration, it was determined that the drag information
from the balance was unreliable. A total head wake rake was designed and fab-
ricated for the BAR T. The streamw ise location of the rake was determined based
on a balance of streamline turning (flow angle at the rake face) and the sensitivity
of the pressure transducers. CFD and wind-tunnel wall-pressure signatures were
used to identify that the jet wake was aligned with the freestream streamlines at
x / c greater than 3.5 from the TE of the model. An example of the wall-pressure
signature is shown in Fig. 19 for typical high-lift conditions.
The magnitude of the wall-pressure signatures shown in Fig. 19 indicates that
a correction may be warranted for the dynamic pressure and angle of attack.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 219/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
21
1
-0.50
-0.25
ACp
O O0
0.25
0.50
-4.0
-3.0
-2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0
XIC LE
REF)
Fig.
19
Wind-tunnel wall-pressure signatures for different lift coefficients solid
symbols for upper wall, open symbols for lower wall), h =
0.020
in., q =
10
psf,
circular trailing edge.
Several wall correction techniques are described in the
1998
AGARD “Wind
Tunnel Wall Corrections” report.32Corrections of two-dimensional experiments
for wall effects are compounded by the two-dimensional AR and the juncture
flow of the model and wind-tunnel wall interface. As a first approximation of
the wall interference characteristics, corrections for two-dimensional lift interfer-
ence are made using a classic approach described in the appendix. It is recognized
that these corrections are inadequate and that the wall signature method may
be more appropriate.
evaluation^^^
of the wall-signature method are ongoing
and are not applied to the data presented in this report.
The wall-signature pressure distribution is also used to locate the streamwise
wake rake position for this experiment. The criteria for the rake measurements
are based on a tradeoff of transducer sensitivity and flow angularity of the flow
at the probe tip. Based on these criteria, the wake rake was located
3.6
chords
downstream of the TE of the model at an AOA of Odeg. The wake profiles
shown in Fig. 20 are representative of the effectiveness of the streamline
turning created by the circular CC airfoil configuration. The errors associated
with the integration of the wake to determine measured drag are related to
the nonzero pressures outside the wake region. Although the rake spans the
entire test section, only
86
is used for the wake integration, thus eliminating
the influence of the floor and ceiling boundary layers. The measured drag was
determined to have a repeatability of d
=
f0.0005. For the momentum
sweep at AOA = 0, the wake moved approximately one chord below the center-
line. An example of an AOA sweep at a fixed blowing rate is shown in Fig. 21.
The wake moved approximately 1.5 chords below the centerline prior to
stalling.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 220/637
21 2 GREGORY S. ONES
0.0125
0.0075
0.0025
CP
-0.0025
-0.0075
NO BL OWING
-0.0125 I
I
-2 -1 0 1 2
U C
(WAKE POSITION)
Fig. 20 Wake profile of GACC with circular trailing edge, AOA = 0,
x/c = 4.64.
The measurement of the nondimensional momentum coefficient can be
obtained from parameters described in
Eq.
(8). Using mass flow and measured
pressure ratios to obtain Ujet, the momentum coefficient can be calculated
without any knowledge of slot height. This is the preferred method because of
the potential errors in measuring the slot height
of
the small-scale model used
CP
AOA
- - --10.0
- -
-6.0
0.0
-0.10
-0.15
0.20
*C,o,,
2 -1 0 1 2
U C
(WAKE POSITION)
Fig. 21
C
= 0.075, x/c = 4.64.
Wake profile of GACC with circular trailing edge, -10
c
AOA c 10,
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 221/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
213
in this test. However, post-test evaluation of the mass flow data revealed
problems with the turbine meters that can be associated with the turbine
meters being located in high-pressure legs of the flow path. This resulted in the
use of slot height to determine the momentum coefficient.
Slot height is a critical parameter for correlation to airfoil performance and
was given careful attention. Nominally, the slot height was set with a digital
height gauge (accuracy
=
0.0001 in.) under no flow conditions. The height was
then readjusted to obtain a uniform velocity along the span of the slot. The slot
height was locked into place with a push-pull set of screws located approxi-
mately 1 in. from the slot exit inside the settling region of the jet plenum. The
0.010 in. TE of the stainless steel skin was observed under load with a microtele-
scope and did not appear to move. However, post-test span wise jet velocities
measured at the slot exit with a hot wire probe, shown in Fig. 22, indicate vari-
ations of 20% relative to the reference jet velocity determined from pressure
ratio. Most of these variations can be identified with the wake of the internal
push-pull screws used for setting slot height. The variations of the low je t vel-
ocities are larger than the higher jet velocities. It was also discovered that the
extreme inboard and outboard slot velocity (not shown) was significantly
lower than the core region of the span. This is attributed to internal flow separ-
ation at the inlet and exit of the flow manifold internal to the model. Although
affecting only the extreme
0.5
in. sections of the span,
it
does effectively
reduce the length of the blowing section of the jet.
The large-scale span-wise variation is thought to be due to internal flow vari-
ations an d/or errors in setting the slot height under loaded conditions. Setting the
final slot height was done on site with the model mounted in the tunnel and mass
flow being added. The confined space of the small wind tunnel made setting the
slot height difficult because of issue of accessibility and noise. Pressurizing the
0
0.2 0.4 0.6 0.8 1
SPANISPAN,,,
Fig.
22
Example
of
spanwise velocity deviation for different jet exit M ach numbers
biconvex TE configuration,h = 0 020 in.).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 222/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 223/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL 215
2.000
1.500
0.000
0 0.2 0.4 0.6 0.8 1 1.2 1.4
UJET-HW/ JET-REF
Fig.
24
Normalized velocity profiles at the upper surface exit plane of the
biconvex
TE.
higher than was thought at the time of setup, as shown in Fig. 25 for the circular
TE. The calculated slot height also varied up to 18% with increasing nozzle
pressure ratio. An average of slot height for the varying mass flow was used for
reporting purposes. Extrapolating the biconvex calculated profile to the
unblown condition results in a 0.021 in. setup. This is consistent with the slot
height measured in the post-test slot profile hot-wire measurements shown
in Fig. 24.
c
1
oo
1.10 1.20 1.30
NPR
Fig. 25 Slot height variations as internal plenum pressure increases for
circular TE.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 224/637
216 GREGORY S. ONES
VI. Airfoil Performance
Airfoil performance will be discussed for two modes of the GACC airfoil: the
high lift mode w ith upper slot blowing and the cruise mode with upper and lower
slot (dual) blowing. The efficiency of pulsed blow ing will be discussed as part of
the high-lift mode.
A. High Lift Mode
1. Baseline (No Blowing)
Lift, drag, and pitching m oment will be used to establish the two-dimensional
baseline performance of the GACC airfoil with different TEs. The original
GA CC airfoil was designed around the circular TE having an
r / c
of 2%. There-
fore, the circular TE will be used as the reference for the elliptic and biconvex
trailing edges. Comparing the lift performance of the three TEs with no
blowing in Fig. 26, the circular TE has a lift enhancement of ACl = 0.16 at a
zero degree AOA relative to the biconvex and elliptic TEs. This is also reflected
in the TE pressures shown in F ig. 27.
Com parisons of the drag performance for the three TE are show n in Fig. 28.
There are few differences in the indicated drag. This can be related to boundary
layer transition fixed at
5
chord and the fixed trailing height established by the
steps created by the upper and lower slots. Minimum drag occurs at zero lift and
AOA = . The airfoil efficiency shown in Fig. 29 indicates that the circular TE
is more efficient than the elliptic or biconvex TEs with no blow ing. The peak effi-
ciency occurs at AOA
=
6 deg and is consistent with the differences in lift. The
2.0
1.5
1 o
C, 0.5
0 0
-0.5
-1 o
20
-1 0 0 10 20
AOA
Fig. 26 Baseline lift coefficient with no blowing balance data).
Next Page
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 225/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
21
7
-5.0
-4.0
3 . 0
-2.0
-1
.o
CP
0.0
1
o
2.0
0.00 0.25
0.50
0.75
1 oo
XIC
Fig. 27 Pressure distribution for GACC airfoil, no b lowing, AOA =
0.
drag polar shown in Fig. 30 illustrates a relatively flat drag characteristic for the
region of lift, which is consistent with cruise conditions (e.g.,
Cl
0.5).
2.
Circular
TE
The circular Coanda TE will be used as a reference for comparisons of
performance throughout the rest of this paper. This section will highlight the
circular TE performance for high-lift conditions. Although somewhat arbitrary,
the initial goal of this effort was to generate a lift coefficient of 3 at an AOA
0.10
0.08
0.06
CD
0.04
0.02
0.00
-20 -10 0 10 20
AOA
Fig. 28 Baseline drag coefficient with no blowing (wake rake).
Previous Page
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 226/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 227/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
219
4.0
3.0
2.0
CL
1 o
0.0
-1
.o
-20 -1
0 0
10 20
AOA
Fig. 31
represent lower b lowing).
Airfoil lift performance with circular
TE
and h l c
=
0.0022 (open symbols
of upper and lower blowing with variations in AOA enables the designer to cus-
tomize lift and drag for either approach or takeoff conditions. The open symbols
shown in Fig. 31 highlight the lower Coanda blowing. The pneumatic flap effect
of lower blowing com pensates for the T E camber as demonstrated by zero lift at
AOA 0 (Cp,o,,
0.024). These effects are more related to cruise drag and will
be discussed later in this chapter. The efficiency of the Coanda blowing can be
related to the slot height and the radius of the Coanda surface. For a fixed
Coanda surface radius of Y/C 2 , an h / C of 1.4 performed better than an
h / C
of 2.2 , as shown in Fig. 32.
3.5
3.0
2.5
2.0
1.5
l
1 o
0.5
0.0
0.00
0.02 0.04
0.06 0.08
0.10
CP
Fig. 32 Lift performance
of
circular
TE,
AOA =
0.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
0.00 0.02 0.04 0.06 0.08 0.10
C
Cl
0.0014
0.0022
h/C
∆CL
∆C
= 60.3
∆CL
∆C
= 45.3
BOUNDARY LAYER
CONTROL
SUPERCIRCULATION
CONTROL
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 228/637
220
GREGORY
S.
ONES
The lift augmentation for the small slot was 60.3 in the separation control
regime compared to the 45.3 augmentation for the larger slot. To extend into
the supercirculation regime it is necessary to push the rear stagnation beyond
the physical TE, forming a pneumatic flap. A shift in the lift augmentation effi-
ciency highlights this effect, as shown in Fig. 32. The limit of the separation
region for this airfoil occurs at a C of approximately 0.03 and a lift coefficient
of 1.8. To predict the mass flow requirements and lift performance in the super-
circulation region, it is possible to extend the supercirculation lift augmentation
line.
The drag characteristics corresponding to Eq. (18)are shown in Fig. 33. Thrust
is generated for the low blowing rates that are characteristic of most CC airfoils
including GA CC. Combinations of Coanda blowing and AOA allow for variable
drag at a fixed lift condition. As an example, the drag can be varied by
ACd 0.060 at a lift coefficient of 2.0. This would include both a thrust and
drag capability. The limitations of this capability are related to the LE stall
characteristics and may be augmented with LE active flow control.
T o gain a greater understanding of drag characteristics for this airfoil, the total
drag measured in the wake can be decomposed into a two-dimensional circula-
tion induced force represented by the pressure distribution on the airfoil
(shown in Fig. 34) and the reactionary force created by the Coanda je t evaluated
at the jet exit. The reactionary force and the induced force can be combined to
create the total force measured. Because the total drag force is known from the
wake rake data and the reactionary force C is equivalent to C then the two-
dimensional circulation induced force will become
0.10
0.05
0.05
-0.1 0
-1 .o 0.0 1 o 2.0 3 0 4.0
Cl
Fig.
33
Airfoil drag polar for circular TE, h l c =
0.0022,
wake rake data (open
symbols represent lower blowing).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 229/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
WC
22
b) COANDA SURFACE
- I
Degrees)
Fig. 34 GACC pressure distribution with circular TE, AOA
=
0
h / c = 0.00106:
a)
airfoil pressure distribution; b) expanded view of circular TE pressure distribution.
An example of the two-dimensional circulation induced drag force is shown in
Fig. 35. These data corresponds to the lift data in Fig. 32. It can be observed
that the slope change related to the supercirculation region in the lift data is
also evident in the drag data, occurring at a momentum coefficient of approxi-
mately 0.03.
Th e efficiency of a blown airfoil has traditionally been related to an equivalent
drag as described earlier in the text. The equivalent drag shown in Fig. 36
highlights the conversion of measured thrust to equivalent drag for two slot
configurations. Although this enables the comparison of one blown system
with another, it is dangerous for the designer to use these values, as seen by
comparing Figs. 35 and 36. The efficiency of the airfoil can be represented by
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 230/637
222
GREGORY
S.
ONES
0.00 0.02 0.04 0.06 0.08
0.10
CP
Fig. 35
Drag performance of circular TE, AOA =
0.
the lift-to-equivalent-drag ratio shown in Fig. 37. Com parison of the two slot con-
figurations indicates a greater efficiency of the larger slot. This is a result of the
drag benefits of the larger slot and is related to the turbulence characteristics of
the Coanda jet. The peak efficiency occurs in the vicinity of the transition from
boundary layer control to supercirculation (refer to Fig. 32).
The two-dimensional L I D equivalent efficiency of the airfoil can also be
related to the fluidic power required of the high-lift system, as shown in
Fig. 38. The corresponding equivalent drag data are shown in Fig. 39. The
fluidic power can be related to the reactionary thrust component described in
Fig. 35. The dashed line in Fig. 35 represents the contribution of the fluidic
'd
EQUlV
CP
Fig. 36 Equivalent drag of circular TE,
AOA
=
0.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 231/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
223
U D
EQUlV
0.00
0.02
0.04
0.06
0.08
0.10
CP
Fig. 37 Efficiency of circular TE, A O A =
0.
pow er to the equivalent drag. Any values that deviate abov e or below this line can
be related to the two-dimensional c irculation induced effects described above and
highlight the magnitude
of
the dominating contribution of the fluidic pow er to the
equivalent drag.
Evaluating the measured drag per fluidic power reveals that the most efficient
use of the fluidic pow er occurs in the boundary control region. This is shown in
Fig.
40,
where
ACd /Cp ,
s a minimum. Th e magnitude
of
the incremental thrust
LID
EQUIV
0.00
0.1
0 0.20 0.30
CPf (FLUIDIC POWER)
Fig.
38
Pumping power required to achieve equivalentG A C C irfoil efficiency for
circular
TE,
AOA =
0.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 232/637
224
‘d
EQUlV
GREGORY
S.
ONES
0.00
0.10 0.20
0.30
0.40
Cpf (FLUIDIC POWER)
Fig. 39 Fluidic power required to achieve equivalent drag for circular TE,
AOA
= 0.
for the larger slot height is 0.9324 at a fluidic power of 0.03873 shown in Fig. 41.
This corresponds to a thrust of 0.0295 see Fig. 35).
This also illustrates a benefit of a blown system compared to other active flow
control techniques such as synthetic ets and suction systems. Without the benefit
of the reactionary force of the jet, the best performance a traditional active
flow control system could achieve would be related to moving or attaching the
0.00
0.05
0.10
0.1
5
0.20
0.25
Cpf (FLUIDIC POWER)
Fig. 40 Drag efficiency per fluidic power for GACC airfoil with circular TE,
AOA
=
0.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 233/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
d
225
Fig. 41 Drag per power ratio for GAC C airfoil with circular TE, AOA =
0.
boundary layer to the most aft portion of the airfoil. This would result in a
minimum drag associated with skin friction alone. For a tangentially blown
system typical of CC airfoils, the reactionary forces enable penetration into the
outer flowfield that is not available to unblown systems. To make a direct com-
parison of these different active flow control systems it would be necessary to
equate the relevant power watts, horsepower, and so on) to achieve a comparable
drag performance.
Another performance parameter of interest is the lift-increment-per-power
ratio, ACl /Cp ,shown in Fig. 42. This parameter is occasionally used for direct
0
1 2 3
ACI
Fig. 42 Lift per power ratio for GACC airfoil with circular
TE,
AOA
= 0.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 234/637
226
GREGORY
S.
ONES
Table 2 Comparison of GACC lift increment-per-power
to similar powered systemsI2
Item
ACL/Cp, (ACi
0.5
ACLICp, (AC1
=1.0)
GACC h/c 0.0014)
Elliptic CC43
TE blown flap3'
Flap knee44 (BLC mode)
44.3
40.4
42.6
26.8
31
28.6
33.2
7.48
comparisons of similar power-augmented devices. The comparisons are made at
ACl
values of
0.5
and
1.0,
which are consistent w ith the boundary control region,
and the initial stage of supercirculation. For the GACC airfoil, the smaller slot
develops more lift for a given power setting than the larger slot in the boundary
layer control region. As the power (or momentum) is increased into the supercir-
culation region, the influence of slot height on lift-to-power augmentation
decreases. Comparisons of the power requirements for the GACC and other
similar airfoils are shown in Table
2 .
The G ACC airfoil performance is compar-
able to that of a similar CC airfoil and blown flaps with active flow control. The
pitching moment characteristics of the GACC airfoil are shown in Fig.
43.
These
values are consistent with other CC airfoils.
3. Per ormance Com parisons
of TE
The following section will focus on comparisons of the different shape
TEs with a fixed slot height of h/c
0.0022.
The shapes include circular, ellip-
tic, and biconvex profiles, having effective TE rad of r / c
2, 1,
and
0 ,
-1
.o
0.0 1
o 2.0
3.0
4.0
Cl
Fig. 43 Twenty-five percent chord pitching moment characteristics of GACC,
h l c =
0.0022.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 235/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
227
0.00 0.02 0.04 0.06 0.08 0.10
CP
Fig. 44 Comparison of lift performance for the GACC airfoil for different TE
shapes,
h l c
= 0.0022.
respectively . The lift performance of the larger radius configuration is higher than
the other configurations, as seen in Fig.
44.
A
comparison of the drag performance, shown in Fig.
45,
highlights the
improvement of the drag as a function of the smaller r / c . The elliptic TE
( r / c
1 )
has less drag than the circular TE ( r / c 2 ) throughout the bound-
ry
layer and supercirculation region. Transitioning from the boundary layer
region to the supercirculation region, the total thrust of the elliptic TE exceeds
0.00 0.02 0.04 0.06 0.08 0.10
CF
Fig. 45 Comparison of the thrust performance of the GAC C having three different
TE shapes.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 236/637
228
GREGORY
S.
ONES
the reactionary thrust, implying a net two-dimensional circulation induced thrust.
The d rag performance of the biconvex shape mimics the circular TE performance
in the BLC region. The thrust for the biconvex configuration extends beyond the
reactionary thrust throughout the supercirculation region. Comparisons of drag
polars for the three different TEs are shown in Fig. 46. The effectiveness of
the sharp TE is reflected in the increased thrust for the biconvex TE.
Com parisons of pitching m oments for the three TEs are shown in Fig. 47. The
biconvex TE has the lowest pitching moment for any given lift. The benefits of
high thrust and low pitching moment com e at the price of mom entum coefficient;
for example, for a lift coefficient of
2 ,
the thrust of the biconvex is 110 counts
larger and the moment is 50 counts smaller than the circular TE performance.
However, the momentum coefficient increased by a factor of
2 .
B. Cruise Configuration
T o address the issue of a blunt TE for typical C C configurations at cruise, the
GA CC w as designed w ith a dual blowing capability, that is, upper and/or lower
blowing on the Coanda This enables the operator to augment the
system thrust while providing roll and/or yaw control. The following section
will address only the dual-blown circular TE performance.
1.
It should be recognized that the cruise condition for this airfoil would be oper-
ated at a substantially higher Mach num ber and higher dynamic pressure, thereby
reducing the momentum coefficient. These low-speed data do not account for the
airfoil compressibility and potential shock manipulation that typical CC configur-
ations may provide. For cruise conditions, the CC performance characteristics are
Dual Blowing for Circular Coanda Sur ace
0.00 0.50 1.00
1.50 2.00 2.50
3.00
Cl
Fig.
46
Comparison of drag polars for three different TE shapes, h / c
= 0.0022.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 237/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
229
0.00 0.50 1.00
1.50
2.00 2.50 3.00
Cl
Fig.
47
Comparison of pitching moments (referenced to 50 chord) for three
different TE shapes,
h / c
=
0.0022.
limited to the boundary layer control region. Nominally, lift coefficients that are
of the order 0.5 are desired during cruise operations.
To
characterize the lift performance of the dual-blown configuration
of
the
GACC airfoil, the upper blowing condition was fixed and the lower blowing
was swept, as shown in Fig. 48. As expected, the upper blowing performance
remains proportional to the lift. Combining this upper blowing with lower
blowing will result in a lift reduction. However, this reduction does not occur
until the initial stages of thrust.
0.001 0.010
0.100
1.000
CI+UPPER LOWER)
Fig. 48 Lift performance for dual blowing, h / c
= 0.0022.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 238/637
230
GREGORY
S.
ONES
0.001
0.010 0.100 1.000
CP(UPPER
LOWER)
Fig.
49
Drag characteristicsof the circular dual blown configuration, h / c
=
0.0022.
The effectiveness of the dual blown configuration is realized in the drag per-
formance. The drag characteristics associated with Fig. 48 are shown in Fig. 49.
The drag performance seems to be independent of upper blowing in the boundary
layer control region. The d rag polar, shown in Fig.
50,
indicates that thrust can
be adjusted for a given lift (e.g., for a fixed Cl 0.5, a ACd
-0.043
can be
adjusted using dual blowing).
-0.5
0.0
0.5 1 o
1.5
2.0
Cl
Fig. 50 Drag polar for the dual blowing cruise configuration
of
the
GACC
airfoil,
circular TE, h / c
=
0.0022 (upper and lower).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 239/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
23
0.20
0.15
0.10
0.05
c p 0.00
0.05
0.10
0.15
0.20
2 1
0 1
2
WAKE R AKE
POSITION
(UC)
Fig.
51
Wake profiles for the dual blowing cruise configuration of the GACC irfoil,
circular TE, reference Cpupper0.003,
h / c
= 0.0022 (upper and lower).
The wake profile shown in Fig.
51
corresponds to the fixed upper blowing of
C
0.003. As the lower blowing rate increases, the profile goes from a single
peak to a double peak, then returns to a single peak. This indicates that the upper
and lower jet s are independent and do not mix efficiently for the blunt circular TE.
The equivalent drag for the circular dual-blown configuration is shown in
Fig.
52 .
The minimum equivalent drag occurs at a combined momentum
CD
EQUlV
0.000
0.025 0.050 0.075 0.100
CI+UPPER
LOWER)
Fig. 52 Equivalent drag for the GACC dual blown circular TE.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 240/637
232
GREGORY
S.
ONES
U D
EQUlV
0.000
0.025
0.050
0.075
0.100
CP(UPPER LOWER)
Fig. 53 Airfoil efficiency for the GACC dual blown circular TE.
coefficient of 0.03 and a fixed upper momentum coefficient of
0.003.
This is con-
sistent with a measured total drag of -0.012 . The peak efficiency, shown in
Fig. 53, occurs at a total momentum coefficient of 0.021. This is consistent
with the measured d rag transitioning from drag to thrust.
2. Pulsed Blowing
As will be shown in this section, pulsed blowing from the upper slot is
intended to reduce the mass flow requirements for a co m arable steady
blowing performance?6937 The G ACC pulsed b lowing systemZBis based on a
high-speed valve that delivers a high volumetric flow to the upper jet ex it. The
actuator is close coupled (internally located x/c 0.90) to the jet exit through
a rapid diffuser to deliver a pulse of air that can be varied in magnitude, fre-
quency, and duty cycle. An exam ple of the pulse train i s shown in Fig. 54. The
quality of the rise time and decay of the pulse train i s related to the overall actua-
tor authority. The rise and decay time of the pulse train is dependent on the
internal volume located internally just upstream of the jet exit. This includes
the 10:
1
contraction and the settling area downstream of the rapid diffuser exits.
The time-dependent pulse train is referenced to the jet exit or 0 deg of the
Coanda surface. The averaged pressure field is compared to a comparable steady
blowing condition, shown in Fig.
55.
The separation associated with this con-
dition was identified to occur in the range 75 < < 90 deg, whereas steady
blowing was in because
60 < <
75 deg. This corresponds to the lift perfor-
mance shown in Fig.
56.
The mass flow reduction of
55
corresponds to the
40 duty cycle shown in Fig. 54. It should be emphasized that this reduction
is limited to the BLC region because of current limits in actuator authority.
The turbulence magnitude and frequency of the steady je t increases just dow n-
stream of the jet exit, then increases along the Coanda surface to peak at
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 241/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
233
Time (sec)
Fig.
54
Time record of circular Coanda surface pressures with pulsed upper
blowing,
35
Hz, 40 duty cycle, circular TE, h / c = 0.00106.
6
30
deg, shown in Fig.
57.
The magnitude and frequency then decays until
the jet separates from the Coanda surface in the range
60
<
6
<
75
deg. For
the pulsed jet, the turbulence magnitude and frequency of the jet-on portion
of the pulse train increases just downstream of the jet exit, then increases
along the Coanda surface to peak at 6
60
deg, as shown in Fig.
58.
The
magnitude and frequency then decay until the jet separates from the Coanda
surface in the case 75 <
4
< 90 deg.
4 P E G )
Fig.
55
Comparison
of
steady and pulsed pressure distribution for the circular TE,
h / c =
0.00106.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 242/637
234
GREGORY
S.
ONES
2.0
DC=80
C=6O
DC=40 \
DC=20
c, 1.0
55
0.5 I
STEADY
0.0
I I I I
0
0.005 0.01 0.015 0.02 0.025
CP
Fig. 56 Comparison of lift performance for steady and pulsed blowing on the
circular TE, h / c
= 0.00106.
Fig. 57 Frequency content of the pressure field on Coanda surface, steady jet,
circular TE,
h / c
= 0.00106: a) nondimensional spectra for steady jet; b) expanded
view of frequency content for the influence of the shear and entrained flow.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 243/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
235
0.01 0.10 1.00 10.00 100.00
8 10 12 14 16 18
b)
FU U (LREF:TE DIAMETER)
Fig.
58
Frequency content of the pressure field on Coanda surface for the pulsed jet,
actuator drive;
35
Hz,
40
duty cycle, circular TE, h / c =
0.00106:
a)
nondimensional spectra for pulsed jet; b) expanded view of frequency content for
the pulse-on portion of pulse train.
1.5
c,
1.0
0.5
DC=40
DC=S?
0.000 0.005 0.010 0.015 0.020 0.025
Ck
Fig. 59 Mass flow reduction for pulsed elliptic TE, h / c =
0.0022,
BLC region.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 244/637
236
GREGORY
S.
ONES
The performance benefit of the pulsed elliptic T E is significantly less than that
of the circular TE, shown in Fig. 59. For a lift coefficient of
1.0
there is a
29
reduction of mass flow for the pulsed ellip tic T E compared to the 55 reduction
of the circular TE. There was no measurable benefit in mass flow reduction for
the pulsed biconvex TE.
The effectiveness of the pulsed blowing can be related to radius of curvature of
the Coanda surface and jet separation. The pulsed effectiveness for larger
Y/C
that
is represented by the
2%
circular TE, moved the time-averaged separation
beyond the maximum TE location of x c 1.0, that is, from the upper Coanda
surface to the lower Coanda surface. Several factors contribute to
the effectiveness of the pulsed jet, including a larger instantaneous velocity,
the increased turbulence (for mixing), pulse frequency, pulse duty cycle,
and the limitation of a steady jet to remain attached to a sm all radius of curvature.
Further research is needed to isolate these parameters.
VII.
Conclusions
The results of this study have addressed two of the major hurdles that limit the
application of CC to aircraft: 1) reducing the C C mass flow bleed requirements
from the engine and 2 ) Conversion of the cruise drag associated with a blunt
TE to thrust through a pneumatic cruise flap. The efficiency of the GACC
airfoil is compared to other CC airfoils in Fig.
60.
The details of the other CC
airfoil data are described in Ref. 17 and shown here to capture the range of
possibilities for the GACC configuration.
U D
(EQUIV)
Fig. 60 Comparison of GACC efficiency with similar CC airfoils, AOA=O unless
otherwise noted (curves do not necessarily represent the envelope of maximum
efficiencycI/cdequiva em h
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 245/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
237
Th e improved efficiency of the cambered rounded ellipse airfoil' is believed
to be a function of the larger radius of the circular TE used on the elliptical
airfoil. The increased efficiency of the camber for the elliptical airfoil is also
shown for the
t / c
0.20
configuration.18 The cam ber effects of the GACC
airfoil are demonstrated in the generation of higher lift for comparable momen-
tum coefficients. Comparing the GACC efficiency to a typical blown flap38
reveals the lift benefit of attaching the je t through C oan da turning. It is specu-
lated that the blown flap prematurely separates, limiting its lift performance to
Cl
<
2. Reshaping the blown flap to the dual-radius CC flap profile enables the
jet to remain attached to the TE of the flap, extending its lift performance to
Cl RZ 5 . It should be noted that L E blowing w as required to extend the lift coef-
ficient beyond
Cl RZ
for the dual radius flap.39 Th e poor efficiency of the jet
flap is generally related to the large blowing requirements associated with the
reactionary force:' and the min imal effect on the two-dimensiona l induced
pressure field.
The efficiency of the GACC's dual blown configuration highlights the low-
speed cruise conditions. Nominally, the lift requirements for cruise are
Cl
0.5. Recall from Fig. 50 that most of the real drag is in the form of
thrust. It is also unclear what jet U to use in the
C,
equation, because the
upper and lower blowing were controlled independently. The general perform-
ance of the GACC airfoil is good, but has not been tested to its limits. It is rec-
ommended that LE active flow control be added to extend the limits of lift. It is
also important to extend the pulsed performance benefits into the supercirculation
region as well as evaluating the Three-dimensional (induced drag) effects.
Selecting the GA CC airfoil section for use on an ESTOL o r PAV vehicle will
require a system study that accounts for integration of the engine and CC system.
A trade study of thrust from the engine alone or a coupled system of engine and
the C C airfoil thrust should highlight the benefit of the CC system. The GACC
airfoil does seem to be an excellent candidate for the outboard portion of the
ESTOL wing, having good lift augmentation capability and good roll and yaw
potential.
Appendix
A. Wall Interference
As a first approximation of the wall interference characteristics, corrections
for two-dimensional lift interference can be made using a classic approach
described by Krynytzky and Hackett41 and Allan and V i n ~ e n t i . ~ ~or a small
model centrally located between two closed parallel walls, corrections for
angle of attack, lift, and pitching moment can be estimated using the
following equations:
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 246/637
238
GREGORY
S.
ONES
where
and
and
2
ACm -*(L)92
PH
CL
4cotT
[
1 + (2 M2)&]quncom
Examples of the wall interference corrections described by Eqs. (A.l-A .4) are
small, as seen in Figs. A.l-A.4.
. a . + *
. . . .
.0000
0.0025
-0.0050
20 -1
0 0 10
AOA
0
Fig. A . l Angle of attack correction from wall interference (circular
TE).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 247/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
239
0.025
M
-0.0501 I
20
-10 0 10
AOA
Fig.
A.2
Lift corrections from wall interference (circular TE).
0.010
0.008
0.006
ACm 0.004
0.002
A0 177
0 134
0 093
o . o o o ~
m e * * *
0.002
20
-10 0 10
20
AOA
Fig. A.3 Moment corrections from wall interference (circular TE).
A0 177
0 134
+
0 093
r e
20
-1
0 0 10
20
AOA
Fig. A.4 Dynamic pressure corrections from wall interference (circular
TE).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 248/637
240
GREGORY
S.
ONES
B. Balance Corrections
are given by
Data reduction equations and tare corrections fo r pressure lines across balance
NF ~ F ( N F S C ) interactions)
AF 6AF(AFSC)-
AFinteractions
+ Pressurecorrection)
PM &M(PMsc) PMinteractionsPressureCorrection)
YM &M(YM sc) YMinteractionsPressureCorrection)
RM ~ M ( R M S C )
-
RMinteractions)
Pressure tare correction for axial, pitching moment, and yawing mom ent forces
are given by
where
where
Pressure Tarecorrection l P a c t +
2 P u p p e r
+ P l o w e r + 4 P ac tP u p p er + sPactPlower
The accuracy of the balance is highlighted in Table A. 1. The rolling moment and
yawing moments are meaningless for two-dimensional testing and will be
Table A .l GACC strain gauge balance accuracy (95 confidence level)
Normal Axial Pitching mom ent Rolling Yawing mom ent
( FS) ( F S )
( FS)
moment ( FS)
(
FS)
0.04
0 39 0.12 0.07
1.64
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 249/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
24
Table A.2 Summary
of
GACC pressure resolutions
Po
A P
AP
AP AP wall
freestream freestream model static rake signature
AP
MEMS
PSIA PSID PSID in. H 2 0 in. H 2 0 PSID
15 1 2.5 10 10
5
0.1080 0.0072 0.0360 0.0052 0.0052 0.0720
A m i n
(PSF)
A m i n
(PSF)
A m i n
(PSF)
A m i n
(PSF)
A m i n
(PSF)
A m i n
(PSF)
ignored except when calculating the interactions to obtain corrected normal,
axial, and pitching moments.
C. Pressure Measurement Limits
The errors associated with the pressure data described above are related to the
resolution of the pressure instrumentation. Nominally the pressure instrumenta-
tion errors are characterized by percent of full scale. See Table A.2 for a
summary of the GACC pressure instrumentation.
References
‘Jones,
G.
S., and Josh , R . D. ,
Proceedings
of
the
2004
NASAIONR Circulation
Control Workshop, NASA /CP-2005-2 13509, June 2005.
’Rich, P., McKinley, R. J., and Jones, G. S., “Circulation Control in NASA’s Vehicle
Systems Program,”
NASA/CP-2005-213509/PTl
une 2005, pp. 1-36.
3Moore, M. D., “Wake Vortex Wingtip-Turbine Powered Circulation Control High-Lift
System,” NASA/CP-2005-213509/PT2 une 2005, pp. 641 -656.
4McLean, J. D., Crouch, J. D., Stoner, R. C., Sakurai, S., Seidel, G. E., Feifel, W. M ., and
Rush, H. M., “Study of the Application of Separation Control by Unsteady Excitation to
Civil Transport Aircraft,” NASA/CR-1999-209338, June 1999.
’Streett, C. L., “Numerical Simulation of a Flap-Edge Flow Field,” Fourth AIA A/C EAS
Aeroacoustics Conference, AIAA Paper 98-2226, June 1998.
6Wood, N., and Nielson, J., “Circulation Control Airfoils Past, Present, and Future,”
AIAA Paper 85-0204, Jan. 1985.
’Englar, R. J., “Circulation Control Pneumatic Aerodynamics: Blown Force and
Moment Augmentation and Modifications: Past, Present, and Future,” AIAA Paper
2000-2541, June 2000.
8Jones,
G. S.,
Bangert, L.
S.,
Garber, D. P., Huebner, L . D., McKinley, R.
E.,
Sutton, K.,
Swanson, R. C., and Weinstein, L., “Research Opportunities in Advanced Aerospace
Concepts,” NA SA, TM -2000-210547, Dec. 2000.
’Davenport, F. J., “A Further Discussion of the Limiting Circulatory Lift of a Finite-Span
Wing,” Journal of the Aerospace Sciences, Vol. 27, Dec. 1960, pp. 959-960.
“Smith, A. M. O., “High-Lift Aerodynam ics,” Journal ofA irc ra f, Vol. 12, No. 6,19 75 ,
‘‘McCormick, B. W., Jr., Aerodynamics of
V / S T O L
Flight, Dover Publications,
”Wilson, M. B., and von Kerczek, C., “An Inventory of Som e Force Producers for Use
pp. 501-530.
Mineola, New York, 1999.
in Marine Vehicle Control,” DTNSRDC-79/097, Nov. 1979.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 250/637
242
GREGORY
S.
ONES
13Betz, A., “Ein Verfahren zur direkten Ermittlug des Profilwiderstandes,” ZFM, Vol.
14Jones, B. M., “The Measurement of Profile Drag by the Pitot T raverse Method,”
”Schlichting, H., Boundary Layer Theory, 6th ed., McGraw-Hill, New York, 1968.
16Rae,W . H., and Pope, A., Low-Speed W ind Tunnel Testing, 2nd ed., Wiley, New Y ork,
1984.
17Kind, R. J., “A P roposed M ethod of C irculation C ontrol,” Ph.D. Thesis, Univ. of
Cambridge, June 1967.
“Englar, R. J., and Williams, R. M., “Test Techniques for High Lift, Two-Dimensional
Airfoils with Boundary Layer and Circulation Control for Application to Rotary Wing Air-
craft,’’ Canadian Aeronautics and Space Journal, Vol. 19, No. 3, 1973, pp. 93-108.
”Jones,
G.
.,
Viken,
S.
A., Washbum, A. E., Jenkins, L. N., and Cagle, C. M., “An
Active Flow Circulation Controlled Flap Concept for General Aviation Applications,”
AIAA Paper 2002-3157, June 2002.
”Lan, C. E., and Roskam, J., Airplane Aerodynamics and Perjormance, Roskam Avia-
tion and Engineering, 1981.
”Homer,
S.
F., and Borst, H. V., Fluid-Dynamic
Lift,
Homer Publishing, 1985.
”Raymer, D. P.,
Aircraft Design: A Conceptual Approach,
AIAA Education Series, 3rd
ed., AIAA, Reston, VA, 1999.
23McGhee, R. H., and Bingham, G.H ., “Low-Speed Aerodynamic Characteristics of a
17-Percent Thick Supercritical Airfoil Section, Including a Comparison Between Wind-
Tunnel and Flight Data,” NASA, TM X -2571, July 1972.
24Cagle, C. M ., and Jones, G. S., “A Wind Tunnel Model to Explore Unsteady
Circulation Control for General Aviation Applications,” AIAA Paper 2002-3240,
June 2002.
25Englar, R. J., and Williams, R. M., “Design of Circulation Controlled S tem Plane for
Submarine Applications,” David Taylor Naval Ship R&D Center, Rep. NSRDC/AL-200
(AD901-198), March 1971.
26Englar,R. J., “Low Speed Aerodynam ic Characteristics of a Small Fixed Trailing Edge
Circulation Control Wing Configuration Fitted to a Supercritical Airfoil,” DTNSRDC,
Rep. DTNSRDC/ASED-81/08, March 1981.
27Englar, R. J., H emm erly, R. A., Tay lor, D. W., M oore, U. H., Seredinsky, V., Valck-
enaere, W.
G.
nd Jackson, J. A., “Design of the Circulation Control Wing S TOL D emon-
strator Aircraft,” AIAA Paper 79- 1842, AIAA Aircraft S ystems and Technology Meeting,
Aug. 1979; also Journal ofAircraft, Vol. 18, No. 1, 1981, pp. 51-58.
”Pugliese, A. J., and Englar, R. J., “Flight Testing the Circulation Control Wing,” AIAA
Paper 79-1791, presented at AIAA Aircraft S ystems and Technology Meeting, Aug. 1979.
29Kuethe, A. M ., and C how, C.-Y., Foundations
of
Aerodynamics, Bases
of
Aerody-
namic Design,
5th ed., Wiley, New York, 1998.
30Preller, R. F., and Rose,
0
J., “Langley Wind Tunnel Force Reduction Program,”
31Smith, D. L., “An Efficient Algorithm using M atrix Methods to Solve W ind Tunnel
32Ewald,B. F. R. (ed.), “Wind Tunnel Wall Correction,” AGARDograph 336, Oct. 1998.
331yer, V., Kuhl, D. D., and Walker, E. L., “Improvem ents to W all Corrections at the
16, 1925 , pp. 42-44 .
Reports and Memorandum, No. 1688, Brit. A.R.C., Jan. 1936.
NASA, CR-165650, NOV.1980.
Force-Balance Equations,” NASA , TN-D-6860, Aug. 1972.
NASA Langley 14x22-FT Subsonic Tunnel,” AIAA Paper 2003-3950, June 2003.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 251/637
PNEUMATIC FLAP PERFORMANCE OF CC AIRFOIL
243
34Rose, R. E., Hammer, J. M ., and K izilos, A. P., “Feasibility Study of a Bi-Directional
Jet Flap Device for Application to Helicopter Rotor Blades,” Honeywell, Doc. No. 12081-
FR1, July 1971.
Rogers, E.
O.,
and Donnelly, M. J., “Characteristics of a Dual-Slotted Circulation
Control Wing of Low Aspect Ratio Intended for Naval Hydrodynamic Applications,”
AIAA 42nd Aerospace Sciences Meeting, AIAA Paper 2004-1244, Jan. 2004.
360y le r, T. E., and Palmer, W. E., “Exploratory Investigation of Pulse Blowing for
Boundary Layer Con trol,” North American Rockwell, Rept. NR72H-12, Jan. 15, 1972.
37Walters, R. E., Myer, D. P., and Holt, D. J., “Circulation Control by S teady and Pulsed
Blowing for a Cambered Elliptical Airfoil,” West Virginia Univ., Aerospace Engineering
TR-32, Morgantown, WV, July 1972.
38Lawford, J. A., and Foster, D. N., “Low-Speed Wind Tunnel Tests on a W ing Section
with Plain Leading- and Trailing-Edge Flaps Having Boundary-Layer Control by
Blowing,” British Aeronautical Research Council R&M 3639, 1970.
39Englar,R. J., and Huson, G.
G.,
“Development of Advanced Circulation Control Using
High-Lift Airfoils,” AIAA Paper 83-1847, July 1983.
40Williams, J., and Alexander, A. J., “Som e Exploratory Three-Dimensional Jet-Flap
Experiments,”
Aeronautical Quarterly,
Vol.
8,
1957, pp 21 -30.
41Krynytzky, A., and Hackett, J. E., “Choice of Correction Method,” AGARDograph
336, Section 1.4, Oct. 1998.
42Allan, H. J., and Vincen ti, W. G., “Wall Interference in a Two-Dimensional-Flow
Wind Tunnel with Consideration of the Effect of Compressibility,” NACA Rept. 782,
1944.
43Englar,R. J., “Two-Dimensional Subsonic Wind Tunnel Test of Two 15-percent Thick
Circulation Control A irfoils,” NSRD C, Technical Note AL-211, Aug. 1971.
44Alvarez-Calderon, A., and Arnold, F. R., “A Study of the Aerodynamic Characteristics
of a High-Lift Device Based
on
a Rotating Cylinder and Flap,” Stanford Univ., Dept.
of
Mechanical Engineering Technical Rept. RCF-1, Stanford, CA, 1961.
35
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 252/637
Chapter
8
Trailing Edge Circulation Control of an Airfoil at
Transonic Mach Numbers
Michael
G.
Alexander,* Scott
G.
A n d e q t and Stuart K. Johnsont
NASA Langley Research Center, Hampton, Virginia
Nomenclature
b =
model span, in.
c = chord, in.
cref
=
reference chord,
30
in.
CD = discharge coefficient
Cl
= sectional lift coefficient
C =
sectional pitching moment coefficient
C p= pressure coefficient
C,
= momentum coefficient
g , = gravitation constant
=
32.174 lbm-ft/lbf-s
h
=
average measured slot height, in.
riz
=
mass
flow,
lbm/s
P, = freestream static pressure, psia
Po = total pressure, psia
q = dynamic pressure, psi
r
=
radius
s
= reference area, ft2
t = airfoil thickness
To= total temperature, O R
V =
velocity, ft/s
h / c =
nondimensional slot height
*Aerospace Engineer. Associate Member AIM
'Aerospace Engineer.
Copyright
005
by the American Institute of Aeronautics and Astronautics, Inc.
No
copyright is
asserted in the United States under Title 17, U.S. Code. The U.S. Government has a royalty-free
license to exercise all rights under the copyright claimed herein for Governmental purposes. All
other rights a re reserved by the copyright owner.
245
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 253/637
246
M
G.
ALEXANDER, S.
G.
ANDERS, AND S. K. JOHNSON
x = chordwise distance, in.
y = span distance, in.
a
=
angle of attack, deg
p = density, lbm/ft3
y
= ratio of specific heat
ACl/C,
= lift augmentation ratio
y / b = nondimensional span location
Subscripts
jet
=
air flow that exits nozzle
1
=
lower and lift
plenum
=
airfoil plenum
s
= slot
u =
upper
co = infinity
TE = trailing edge
0.25 = quarter chord
I. Introduction
IRCULA TION control (CC) is considered one of the most efficient methods
C
or lift augmentation at low Mach numbers.' The device augments an air-
foil's lifting capability by tangentially ejecting a thin jet of high-momentum
air over a rounded trailing edge (TE).* The jet will remain attached to the
surface as along as the low static pressures created by the jet are large enough
to balance the centrifugal forces acting to detach the jet (Fig.
l).3
The jet
moves the separation point around the TE toward the lower surface of the
wing and entrains the external flowfield. This entrainment and separation point
movement produces a net increase in the circulation of the wing, resulting in
lift a ~ gm e n t a t i o n . ~
Tangential Blowing Over
a Rounded Coanda Surface
Fig. 1 Tangential blowing over a Coanda surface.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 254/637
CC OF AIRFOIL AT TRANSONIC MACH NUMBERS 247
Numerous experimental C C tests using the Coanda effect to enhance lift have
been conducted at subsonic velocities on relatively thick (15%) airfoil
section^ ^
The focus of this experiment is to evaluate the effectiveness of T E C C on a thin
airfoil section at transonic Mach num bers. A wind-tunnel test was conducted on a
6
thick slightly cambered elliptical airfoil with both upper- and lower-surface
slot blowing. Parametric evaluations of jet slot heights and Coanda surface
shapes were conducted at momentum coefficients
Ce
from 0.0 to 0.12. The
data were acquired in the NASA Langley Transonic Dynamics Tunnel at
Mach = 0.8 at a =
3
deg and Mach
=
0.3 at a
=
6
deg, at Reynolds numbers
per foot of 1.0 x lo6 and 3.6 x lo5 espectively.
11. Model Description
The configuration tested in this experimental investigation is a semispan
rectangular circulation control airfoil (CCA) with zero leading edge (LE) and
TE sweep, having a circular end plate at the tip. The model, as shown in
Fig. 2a, was mounted in the wind tunnel on a splitter plate located 3 ft
off
the tunnel wall. The model incorporated CC by blowing tangentially from
spanwise rectangular slots located upstream of a trailing edge “Coanda
surface”. The model has two separate and isolated internal plenums that feed
air to either the upper or lower rectangular slot nozzle. The rectangular
slot exits are located at x/cref= 0.9 and extend the full span
(60
in.) of the
Fig. 2 a) CCA m odel view from right rear quarter, looking upstream):
b)
CCA
airfoil section.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 255/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 256/637
CC OF AIRFOIL AT TRANSONIC MACH NUMBERS
249
Table
1
Coanda radius and slot height dimensions
Coanda
1.78: 1 2.38: 1 2.98: 1
Chord, in.
rs, in.
rTE, n.
rsIC
bE/C
Guidelines: r/c
hllrs
W r s
h3vS
hl/rTE
h2/rTE
~ ~ I Y T E
Guidelines: hl r
27.82
1.44
0.25
0.052
0.009
0.024
0.039
0.051
0.14
0.22
0.29
28.09
2.57
0.19
0.091
0.007
0.02 to 0.06
0.014
0.022
0.028
0.18
0.30
0.38
0.01 to 0.08
28.36
4.02
0.15
0.142
0.005
0.009
0.014
0.018
0.23
0.37
0.48
was aligned with the slot exit to ensure the minimum exit area occurred at
x/c, f
=
0.9. The horizontal axis of the ellipse was then mapped to the camber
line of the elliptical airfoil that formed a 5-deg converging nozzle at the slot
exit. The Coanda surface spanned the TE of the model 60 in.).
Reference 6 provided gu idelines for Coanda surface radii of curvature as listed
in Table 1. It is not possible to meet the entire guideline radii of curvature on a 6%
thick airfoil. It was therefore decided that preference would be given to the slot
radius of curvature in an effort to achieve initial jet attachment. As a result, a
family of elliptical Coanda surfaces was chosen that have large slot radii of
curvature and small TE radii of curvature.
C.
Slot
Definitions
Three upper and lower slot heights for each Coanda surface were possible for
this wind-tunnel investigation. The slot heights are given in Table 2. A fourth slot
height (h4) was constructed during the test using the upper surface small slot
h/c
=
0.0012) aft skin by applying four layers of tape at 0.0035 in. per layer
Table
2
Slot and chord measurements
Slot c, in. h, in. hlc
hl 27.82 0.0350
0.0012
h2 28.09 0.0560 0.0020
h3 28.36 0.0730 0.0026
h4 28.36 0.0210 0.0007
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 257/637
250
M G. ALEXANDER, S. G. ANDERS, AND S. K. JOHNSON
for a total thickness of 0.014 in. The resulting “half height” slot was used with the
2.98:
1
Coanda, resulting in an exit
h
= 0.021 in. or
h / c
= 0.0007. The aft upper
and lower removable surfaces were designed to set the slot heights by varying the
internal mold line while not disturbing the outer mold line of the model. Average
measured slot height h and chord lengths were used to determine the height-
to-chord ratio h/c )of each slot. Table 2 lists the measured height and chords
and the resulting h / c . Slot height against Coanda radius information is shown
in Table
1.
D.
Aft Surface
Three sets of aft surfaces were manufactured and attached to the main airfoil
body, which formed the upper and lower external airfoil contour as well as the
internal 5-deg convergent nozzle contour (Fig. 5 ) .
The aft skins also contained chordwise surface static pressure taps at
y / b = 0.5. Any aft surface in combination with any Coanda surface ensured
the minimum nozzle area was located at the nozzle exit. Each aft surface also
established a discrete slot height above the Coanda surface.
E. End Plate
The CCA model used a circular end plate to promote two-dimensional
flow conditions. The end plate was a 30-in.-diam circular plate constructed
from a 0.25-in.-thick aluminum plate with the outside edge beveled. The
design of the end plate was based on sizing criteria found in Ref. 7. A remo-
vable cutout located at its TE allowed for Coanda surface removal and
replacement.
F.
Internal Plenum
As seen in Fig. 2b, the airfoil section is divided in to contiguous, separate, and
isolated upper and lower plenums. The ratio of the slot height to plenum height
ranged from 3.8 to 12.8 depending on the slot height. This ensured low flow
velocities in the plenum that helped maintain uniform plenum flow.
Aft Lower
Surface
Fig.
5
Aft surface identification.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 258/637
CC OF AIRFOIL AT TRANSONIC MACH NUMBERS 251
The m odel has the capability of holding six removab le, 0.050-in.-thick, high-
pressure-loss screens. The screens were fastened to the plenum floor and
extended to the plenum ceiling. Each screen has a porosity of
30%
and is
capable of being placed in both upper and lower plenums at the three locations.
The screen’s porosity was sized using the method described in Ref.
8.
A small
parametric test was performed using the CCA airfoil and the plenum screens
to determine which screen combination created the optimum uniformed flow in
the spanwise direction. From those data, it was determined to use one screen
in each plenum in the aftmost position. The aft screen was located at
x/cref
=
0.72 and ran full spanwise and parallel to the slot nozzle.
G.
Boundary Layer Trip
A boundary layer trip strip’ was located 1.5 in. (measured along the surface)
aft of the LE on the upper and lower surface. The trip strip used epoxy dots
having a diameter of 0.038 in., a thickness of 0.015 in., and an edge-to-edge
spacing distance between the epoxy dots of 0.098 in.
111.
Instrumentation
All pressures were obtained using miniature electronic pressure scanners.
A.
CCA Surface Static Pressures
A total of 83 external static surface pressure taps was loca ted at
y / b
=
0.5
on
the upper and lower airfoil surface (42 upper and 41 low er taps). There are two
spanw ise rows of ten static pressures taps located at
x/cref=
0.5 and 0.8 on each
upper and lower airfoil surface.
B. Coanda Surface Static Pressures
Each C oanda surface had a total of 19 static surface pressure taps located
at y / b = 0.5 every 10 deg radially from 0 deg to 180 deg with 0 deg and 180
deg at the nozzle exit (Fig. 6 ) .
Table 3 Internal plenum pressure tap locations
Taps v l b X1Cre-f
0.2
0.2
0.45
0.5
0.55
0.8
0.3
0.8
0.8
0.8
0.8
0.8
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 259/637
252
M G. ALEXANDER, S. G. ANDERS, AND S. K. JOHNSON
0
180
Fig.
6 Coanda surface tap placement.
C. Total Pressures
Each plenum had six total pressure taps. Their locations are given in Table 3.
Pressure taps at x/c,f
=
0.8 are located aft of the high-loss screen and pressure
taps
x/c,f
= 0.3 are used to determine the total pressure entering the plenum
from the intake nozzle. The total pressure for the plenum was averaged using
taps
2 ,
3, 4 , and
6
to obtain the nozzle exit total pressure.
D.
Thermocouples
The plenum has two iron-constantan, type-J thermocouples located in each
plenum aft of the aft plenum screen that were used to measure plenum total
temperature.
IV.
Facility
A. Model Support
The Transonic Dynamics Tunne l” (TDT) model support systems used for this
test were a sidewall turntable and splitter plate, as depicted in Fig. 7. The splitter
plate was located
3
ft from the tunnel wall using wall standoffs. The rigid support
and model instrumentation was placed inside an aerodynamic shape or “canoe”
located between the splitter plate and the tunnel sidewall.
B.
Air Supply
Air was supplied to the test section via two 1-in. high-pressure flex lines deli-
vering a maximum of
1
lbm/s at 200 psia. Total temperature of the supp ly air was
uncontrolled and ranged from 13°F to 70°F . Each supply line was attached to a
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 260/637
CC OF AIRFOIL AT TRANSONIC MACH NUMBERS
253
Topview
Fig. 7 CCA model installation in the TDT.
control valve that regulated total pressure to the CC A m odel. A m anually oper-
ated crossover line located upstream of the control valve allowed mass flow to be
diverted from one line to another. After the control valve, each line of the supply
air went through its dedicated critical flow venturi and then entered the model
plenum.
V. Test Procedures and Conditions
A. Lift and Pitching Moment
The sectional lift coefficient [Eq . (l)]and quarter chord pitching m oment coef-
ficient [Eq. 2)] were obtained by numerically integrating (with the trapezoidal
method) the local pressure coefficient at each y / b
=
0.5 chordwise orifice from
the upper and lower surface of the m odel:
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 261/637
254
M G. ALEXANDER, S. G. ANDERS, AND S. K. JOHNSON
B.
Mass
Flow
The momentum coefficient is calculated using
The ideal jet velocity (ft/s) was calculated based on the assumption that the
slot jet flow expands isentropically to the freestream static pressure [Eq. (4)]:
Mass flow was determined using Eq.
5 )
The discharge coefficient was ob tained from critical flow venturi calibrations
conducted in the NASA Jet Exit facility. The conditions at the critical flow
venturi were calculated from a static pressure measurement taken at the throat
and a total pressure and temperature near the venturi throat.
VI. Test Conditions
The test conditions and ranges are outlined in Table 4. No corrections were
applied to account for tunnel flow angularity, wall interference effects, or end-
plate effects.
VII. Discussion
of
Results
The higher Reynolds number data will be presented first, because the exper-
imental investigation at transonic conditions was the main testing objective.
A.
Mach
= 0.8, a = 3
deg
1.
Coanda Su8ace Effect
In Figs.
8
and
9,
Coanda surface effects are presented for the upper and
lower slot blowing, respectively. At Mach = 0.8 at
a
=
3
deg, each Coanda
Table 4 CCA test range
of
conditions
Mach Po sia P sia To F Relft
0.3 2.7-4.1 2.6-3.8 67-94 3.6 lo5 o 5.5 lo5
0.8 3.0-4.1
2.0-2.7
95-125 7.8 lo5 o 1.0 lo5
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 262/637
CC OF AIRFOIL AT TRANSONIC MACH NUMBERS
255
Coanda Surface Slot Height
Fig.
8
Coanda surface effect, upper slot blowing, M ach
=
0.8, a
=
3
deg.
surface was capab le of generating incremental lift and pitching moment at each
blowing condition. Upper slot blowing generated positive lift and negative
pitching moment increments, whereas the lower slot blowing generated nega-
tive lift and positive pitching moment increments. Generally, the data in
Fig. 8 display three distinct regions. The first region is characterized by an
increasing lift increment with increasing C, followed by a plateau region in
most cases and then, finally, a region of negative lift increment with further
increasing
C,.
As the Coanda surfaces lengthened, increasing
C,
stretched
the regions further. The Coanda surface effect observed in these data indicates
the longer Coand a surface is more effective over the mid- to high-C, range,
whereas all three Coanda surfaces are equivocal at the low end of
C,.
The
data suggest the jet on the longer Coanda surface remains attached longer
over a larger range of momentum coefficients, but, conversely, the jet separates
much sooner on the smaller Coanda surfaces. This data trend is generally fol-
lowed in Fig.
9
for lower surface blowing. However, the low er surface blowing
is not as effective in producing lift increm ent as the upper surface blowing over
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 263/637
256
M G. ALEXANDER, S. G. ANDERS, AND S. K. JOHNSON
Coanda Surface Slot Height
Fig.
9
Coanda surface effect, lower slot blowing, M ach
=
0.8, a
=
3 deg.
the same range of momentum coefficients. Differences in upper and lower slot
blowing are probably due to angle of attack (AOA), camber, and jet exit
angle. Also, as seen in Fig. 9, none of the Coanda surfaces tested on the
lower surface was capable of generating incremental lift or pitching moment
for
h/c = 0.0026.
The lift augmentation ratio
ACJC,)
for upper and lower slot blowing
is presented in Figs.
10
and
11,
respectively. The upper and lower slot
blowing data indicated that the larger the Coanda surfaces, the greater the
magnitudes of lift augmentation. It was observed that as
C,
increased, lift
augmentation decreased in magnitude with the exception of the data obtained
at
h/c
=
0.0026
(Fig. 11) which, as previously noted, generated insignificant
lift increment. Maximum augmentation was typically achieved on each
Coanda surface at momentum coefficients less than
0.005.
It appeared that
the larger Coanda surface was more effective over a larger range of C, at
any given h / c .
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 264/637
CC OF AIRFOIL AT TRANSONIC MACH NUMBERS
257
Coanda Surface
Slot
Height
Fig. 10 Lift augmentation, Coanda surface effect, upper slot blowing, Mach =
0.8,
a
=
3
deg
2. Slot Height Effect
In Figs.
12
and 13, slot height effects are presented for the upper and lower slot
blowing. The data are the same as previously presented, but replotted to better
evaluate slot height effect. At Mach = 0.8 at a = 3 deg, the smallest slots
were most capable of generating incremental lift and pitching moment at each
blowing condition.
The lift augmentation ratio for the upper surface slot blowing slot height effect
is presented in Figs. 14 and 15. It is observed that the sma ller the slot h / c on any
given Coanda surface, the greater the lift augmentation. As stated earlier, as
C ,
increased, the augmentation diminished.
B. Mach
=
0.3 and
Y = 6
deg
1 Coanda Su ace Effect
In Figs.
16
and 17, Coanda surface effects are presented for the upper and
lower slot blowing, respectively. At Mach = 0.3 at
a
= 6 deg, each Coanda
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 265/637
258
M G. ALEXANDER, S. G. ANDERS, AND S. K. JOHNSON
Coanda Surface Slot
Height
Fig. 11 Lift augmentation, Coanda surface effect, lower slot blowing, Mach =
0.8,
a = 3 deg.
surface was capable of generating incremental lift and pitching moment at each
blowing condition. Increasing incremental lift and moments are observed with
increasing blowing rate with upper slot blowing, creating positive lift
increments and negative pitching moment increments, whereas lower slot
blowing created negative lift and positive pitching moment increments.
Upper and lower slot blowing incremental lift and moment data trends for
each Coanda surface displayed a marked decrease in effectiveness at higher
blowing rates. Also observed is an apparent “pinch down” in the
h / c
=
0.0012 and 0.0020 slot data from C
=
0.06 to
0.08
that diminished
as the Coanda surface increased. This may indicate a reattachment effect
(in the immediate region of the slot) followed by a lull where there is little
flow turning with C increment. The lull is then followed by a period of
flow turning around the Coanda bulb as a result of the increased
C
On the
upper surface blowing (Fig. 16), as the slot size
h / c
was increased, the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 266/637
CC OF AIRFOIL AT TRANSONIC MACH NUMBERS
259
Coanda Surface Slot Height
Fig. 12 Slot height effect, upper slot blowing, Mach =
0.8,
(Y
= 3
deg.
preferred Coanda surface went from 1.78:l at h / c
=
0.0012 to 2.98:l at
h/c = 0.0026. It is observed in Fig. 17 that the lower slot blowing force and
moment increments followed the same trend as the upper slot blowing, but
had reduced absolute values of force and moment increments than that of
the upper surface blowing (Fig. 16). Differences in upper and lower slot
blowing are probably caused by AOA, camber, and jet exit angle. At
Mach = 0.3 at a = 6 deg, the smaller slot h/c
=
0.0012) on the smaller
Coanda surface (1.78: 1) generated the largest increments over the largest
C
range, making it the preferred surface at this test condition.
The lift augmentation ratio for upper and lower slot blowing is presented in
Figs. 18 and 19, respectively. As was seen in the M = 0.8 data, the lift augmenta-
tion decreased with increasing
C
Unlike the M
=
0.8 data, the smallest Coanda
surface generated the largest augmentation ratio from all of the data shown.
How ever, the smallest Coanda surface did not achieve the largest augmentation
ratio for all slot heights. At
h / c =
0.0012 the 1.78:1 Coanda surface achieves the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 267/637
260
M G. ALEXANDER, S. G. ANDERS, AND S. K. JOHNSON
Coanda Surface Slot Height
Fig. 13 Slot height effect, lower slot blowing, M ach =
0.8,
cu =
3
deg.
largest augmentation ratio. At h/c = 0.0026, the 2.98: 1 Coanda surface achieves
the largest augmentation ratio.
2.
Slot Height Effect
In Figs. 20 and 21, slot effects are presented for the upper and lower slot
blowing. The data are the same data as previously presented, but replotted to
better evaluate slot height effect. For each Coanda surface the data suggest that
the smaller the h/c he greater ACl and
ACm
generated for the upper (Fig. 20)
and lower (Fig. 21) slot blowing.
The lift augm entation ratio for the upper and lower slot blowing is presented in
Figs. 22 and 23. In Fig. 22, at each Coanda surface tested, the smaller the slo t, the
greater its augmentation ratio becomes.
C.
Nozzle Pressure
Ratio
In Fig. 24, incremental lift data are presented at Mach num bers of 0.8 and 0.3
as a function of nozzle pressure ratio
NPR).
The surface and slot height noted in
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 268/637
CC OF AIRFOIL AT TRANSONIC MACH NUMBERS
261
Coanda Surface
Slot
Height
Fig. 14 Lift augmentation, slot height effect, upper slot blowing, Mach
=
0.8,
a = 3 deg.
the figure was the best configuration for each Mach number. The NPR data are
presented as an aid in interpreting the data. For NPR values greater than 1.893,
the exit slot is choked and therefore the jet is supersonic.
D. Velocity Ratio
In Fig. 25, incremental lift data are presented at Mach numbers of 0.8 and 0.3
as a function of ve locity ratio for the same configurations used in the NPR figures.
These data are presented for reference purposes similar to the NPR data to orient
the reader to the ranges of velocity ratios tested.
E.
Pressure Distributions
Figure 26a presents data taken at Mach
=
0.8 at a = 3 deg, for the 2.98:l
Coanda surface and h / c = 0.0012 slot configuration. A C , effect was not
observed on the LE of this airfoil. The data suggest a possible weakening of
the upper surface shock with increasing
C,.
In Fig. 26, which shows the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 269/637
262
M G. ALEXANDER, S. G. ANDERS, AND S. K. JOHNSON
Coanda Surface Slot Height
Fig.
15
Lift augmentation, slot height effect, lower slot blowing, Mach = 0.8,
a
=
3 deg.
Coanda surface pressures, the pressure data suggested a shock just aft of the
nozzle exit with flow reattachment and pressure recovery. The surface pressure
data indicated the shock moved aft with increasing
C .
Also, note at
C
= 0.017 and 0.02, the jet completely detaches from the surface.
Figure 27 presents data taken at Mach = 0.3 at a = 6 deg for the 1.78:l
Coanda surface and
h / c
= 0.0012 slot configuration. A
C
effect is observed
on the LE at this test condition. As
C
was increased, the LE suction
peak broadened further downstream up to a
C
=
0.046. The data indicated
at C 0.046 that no further enhancement of the LE suction is observed.
In Fig. 27, which shows the Coanda bulb pressures, the pressure data at
C
2
0.046 suggested a shock just aft of the nozzle exit followed by flow
reattachment. As C is increasing, an increasing negative pressure field
is seen over the remaining length of the Coanda bulb surface. In addition,
the surface pressure data suggest that the shock may be moving aft with
increasing C
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 270/637
CC OF AIRFOIL AT TRANSONIC MACH NUMBERS
263
Coanda Surface
Slot
Height
Fig.
16 Coanda surface effect, upper slot
blowing,
Mach
=
0.3, a
=
6 deg.
VIII.
Conclusions
A wind-tunnel experiment conducted at Mach numbers 0.3 and 0.8 on a
two-dimensional, 6% thick airfoil with a modified TE to enhance the Coanda
effect by tangential jet slot blowing was accomplished. Incremental sectional
lift and quarter-chord pitching moment and lift augmentation ratio data
were presented to support any indications of slot height and Coanda surface
effects.
At the transonic cruise condition, Mach = 0.8 at
a
= 3 deg, it was found that
the effectiveness increased with decreasing slot height and increasing Coanda
surface elliptical ratio. The 2.98:l Coanda surface with the upper slot blowing
position having a slot height of h / c = 0.0012 slightly outperformed the lower
slot position with the upper slot, generating a maximum ACl of 0.25 at a C, of
0.008.
At the lower speed and Reynolds number condition, Mach =
0.3
at
a
=
6
deg,
it
was found that the effectiveness increased with decreasing slot height and
decreasing Coanda surface elliptical ratio. The 1.78:1 Coanda surface with the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 271/637
264
M G. ALEXANDER, S. G. ANDERS, AND S. K. JOHNSON
Coanda Surface Slot Height
Fig. 17 Coanda surface effect, lower slot blowing, Mach = 0.3, a = 6 deg.
upper slot blowing position having a slot height of h / c = 0.0012 gave the
maximum ACl generated at 0.75 at a C, of 0.085.
Increasing incremental lift and moments are observed with increasing
blowing rate with upper slot blowing creating positive lift increments and
negative pitching moment increments, whereas lower slot blowing creates nega-
tive lift and positive pitching moment increments. Lower slot blowing was
not as effective in producing lift and pitching moment increments at transonic
velocities as the upper slot blowing over the same range of momentum
coefficients.
The pressure distribution on all Coanda bulbs at Mach 0.8 suggests the
jet detached from the bulb surface at the higher blowing rates, indicating
a limit to the amount of blowing that can be accomplished without losing
effectiveness. Trailing edge blowing influenced the flowfield upstream of
the slot.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 272/637
NoI
h/c=0.0020)
Fig. 18 Lift augmentation,Coanda surface effect, upper slot blowing, Mach = 0.3, Y = 6 deg.
-I
n
9
z
cn
P
5
5
I
z
5
rn
n
cn
N
Q
cn
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 273/637
Slot h/c=0.0012)
Fig.
19
Lift augmentation, Coanda surface effect, lower slot blowing, Mach
=
0.3,
a = 6
deg.
N
3,
Q
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 274/637
a
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 275/637
c
v
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 276/637
Coanda l.781)
Coanda 2.381)
Fig. 22 Left augmentation, slot height effect, upper slot blowing Mach = 0.3,
a
= 6 deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 277/637
Coanda 2.381)
N
?
rn
X
D
z
n
rn
v,
D
z
n
v
rn
z
v
P
Fig.
23
Left augmentation, slot height effect, lower slot blowing Mach
=
0.3,
a
= 6 deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 278/637
CC OF AIRFOIL AT TRANSONIC MACH NUMBERS 271
_
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
i \
~
Mach AOA ; CoanddS; /c ,
............................ I= 0.3; a = 6f 1.78:1)/0.0012 I
M
=
0.8,
a =
3 f 2.98:1)/0.0012
........................................
_............
i
...................................................................
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
-0.4
-0.5
-0.6
-0.7
-0.8
Fig. 24
Upper surface blowing
L ~ ~ ~ ~ I i
c . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . .. .
.
. . . . . .
......
.... .......,...
c L
0
1
2 3 4 5
6
NPR
Lower surface blowing
0 1 2 3 4
5 6
NPR
Nozzle pressure ratio vs
A C ,
upper and lower slot blowing.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 279/637
272 M G. ALEXANDER, S. G. ANDERS, AND S. K. JOHNSON
Upper surface blowing
0.8
0.1
0.6
0.5
0.4
0.3
0.2
0.1
0
0.5 1 1.5 2
2.5
3
3.5
4 4.5
UjetN inf
Lower surface blowing
cI
0
-0.1
-0.2
-0.3
-0.4
-0.5
-0.6
-0.1
I " " I " " ~ " " " " I " " ~ " " " " _
~ Mach
AoA; CoandaiSlotG / c
0.5 1
1.5 2
2.5
3 3.5 4
4.5
U je t N in f
Fig. 25 Velocity ratio vs AC upper and lower slot blowing.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 280/637
CC OF AIRFOIL AT TRANSONIC MACH NUMBERS
273
Upper surface leading edge pressure distribution
Fig. 26 Pressure distribution, C, effect, upper slot blowing, Coanda 2.98:1), slot
( h / c
= 0.0012), Mach = 0.8, (Y = 3 deg.
-1.5
-1
-0.5
0
0.5
-0.1 0 0.1 0.2 0.3 0.4
Upper surface leading edge pressure distribution
No BlowingCµ = 0.002
Cµ = 0.003
Cµ = 0.004
Cµ = 0.006Cµ = 0.008
Cµ = 0.009
Cµ = 0.011
Cµ = 0.012Cµ = 0.014
Cµ = 0.017
Cµ = 0.02
CP
x/c
Upper Surface
Airfoil Leading Edge
-1.5
-1
-0.5
0
0.5
0.92 0.94 0.96 0.98 1
Upper surface trailing edge pressure distribution
No Blowing
Cµ = 0.002
Cµ = 0.003
Cµ = 0.004
Cµ = 0.006
Cµ = 0.008
Cµ = 0.009
Cµ = 0.011
Cµ = 0.012
Cµ = 0.014
Cµ = 0.017
Cµ = 0.02
CP
x/c
Coanda Bulb
Upper Surface
Aft upper surface
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 281/637
274
M G.
ALEXANDER, S.
G.
ANDERS, AND S.
K.
JOHNSON
Upper Surface Leading Edge Pressure Distribution
Fig. 27 Pressure distribution,
C
effect, upper slot blowing; Coanda 1.78:1), slot
( h / c = 0.0012), Mach = 0.3,
a
=
+ 6
deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 282/637
CC OF AIRFOIL AT TRANSONIC MACH NUMBERS
275
Acknowledgments
The authors would like to acknowledge the assistance of those individuals
whose efforts made this test possible. From the TDT, Chuck McClish, Don
Keller, Jennifer P. Florance, and Wesley Goodman, from Lockheed-Martin,
Jerome Cawthorn, and from the Naval Surface Warfare Center, Carderock
Divison, Ernest R ogers and Jane Abramson.
References
‘Novak, C. J., Cornelius, K. C., and Road, R. K., “Experimental Investigations of
Circular Wall Jet on a Circulation Control Airfoil,” AIAA Paper 87-0155, Jan. 1987.
Englar, R. J., “Investigations into and Application
of
the High Velocity Circulation
Control Wall Jet for High Lift and Drag Generation on STOL Aircraft,” AIAA Paper
74-502, June 1974.
3Ahuja, K. K., Sankar, L. N., Englar, R. J., Munro,
S.
and Liu, Yi., “Application
of
Circulation Control Technology to Airframe N oise Reduction,” G TRI Rept. A5928/ 1,
NASA Grant NAG-1-2146, Feb. 2000.
4Abramson, J., “The Low Speed Characteristics
of
a 15-Percent Quasi-Elliptical
Circulation Control Airfoil with Distributed Camber,” David W. Taylor Naval Ship
R&D Center, Rept. DTNSR DC/ASE D-79/07 (AD-A084-176), May 1979.
’Nielsen, J. N., and Bigger, J. C., “Recent Progress in Circulation Control Aerody-
namics,” AIAA Paper 87-0001, Jan. 1987.
6Rogers, E., and Abramson, J., “Selected Notes on Coanda Circulation Control
Airfoils,” unpublished notes, NSWC, Ap. 2002.
’Hoerner, S. F., and Borst, H. V., Fluid-Dynamic Lift, Hoerner Fluid Dynamics,
Bakersfield, CA; 2nd ed., June 1992.
‘Blevins, R. D., Applied Fluid Dynamics Handbook, Krieger Publishing Company,
Melbourne, Florida. Reprint Edition, June 2002.
’Holmes, J. D., “Transition Tr ip Technique Study in the McAir Advanced Design Wind
Tunnel,” Technical Mem orandum 4395, M ay 1984. (The McAir name is used interchange-
ably with McDonnell Aircraft Company.)
“Staff, Aeroelasticity Branch, “The Langley Transonic Dynam ics Tunnel,” LWP-799
Sept. 1969.
Englar, R. J., “Two-Dimensional Transonic Wind Tunnel Test
of
Three 15-percent
Thick C irculation Control Airfoils,” Technical Note AL-128, Dec. 1970.
”Alexander, M. G., Anders, S. G., Johnson, S. K., Florence, J. P., and Keller, D. F.,
“Trailing Edge Blowing on a Two-Dimensional Six-Percent Thick Elliptical Circulation
Control Airfoil up to Transonic Conditions,” NASA Technical Memorandum 2005-
213545, March 2005.
2
11
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 283/637
Chapter 9
Experimental and Computational Investigation into
the Use of the Coanda Effect on the
Bell A821201 Airfoil
Gerald Angle
II,*
Brian O'Hara,* Wade Huebsch,' and
James
Smith'
W est Virginia University, Morgantown, West Virginia
Nomenclature
A =
area
b
=
span,
f t
C
=
coefficient of
c = chord, ft
D
= reference lengths, f t
F
= download force, lbs
h =
height,
f t
L =
length,
f t
P = pressure, lb/ft2
Re =
Reynolds number
V
=
velocity, ft/s
p
=
density, slug/ft3
Subscripts
=jet
p
=
blowing
= freestream
1, 2, 3, 4 =
reference numbers
*Graduate Research Assistant, Mechanical and Aerospace Engineering. Student Member AIAA.
'Assistant Professor, Mechanical and Aerospace Engineering. Mem ber AIAA.
'Professor, Mechanical and Aerospace Engineering. Member AIAA.
Copyright 005 by the American Institute of Aeronautics and Astronautics, Inc. All rights
reserved.
277
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 284/637
278
G. ANGLE II, B. O HARA, W . HUEBSCH, AND
J.
SMITH
I Introduction
HE COA ND A effect can be described as the balance between the inertial and
T
ormal pressure gradient forces in a near-surface jet of a fluid. A simple case
used to describe this phenomenon is a two-dimensional wall jet , which entrains
the surrounding fluid. As the boundary layer is entrained, the local pressure in the
boundary layer is reduced, creating a pressure gradient that pulls or entrains the
jet towards the surface. From the conservation of momentum, as fluid is
entrained, the jet velocity is reduced. Eventually, the jet velocity is low
enough that the fluid viscosity creates an adverse pressure gradient, again separ-
ating the flow. Expanding this concept to a convexly curved surface, a pressure
gradient is created, forcing the jet to bend around the surface, until the adverse
pressure gradient is reached.
Newman' determined that the flow in a curved wall je t is relatively insensitive
to Reynolds number Re as defined below, provided it is in excess of a threshold
value of
40,000.
Thus
'*
P PcO)V,.VcO
PV2
R e = [
where
P
is the local pressure,
P ,
is atmospheric pressure,
vj
and
Vi,
are the je t
and freestream velocities, and
p
and
v
are the density and viscosity of air. An
approximation of a Coanda jet is a constrained jet, where the streamlines of
the freestream act as a restricting surface. Early experimentation into constrained
jets determined that the inflow velocities of the jet flow do not differ from the
constrained and unconstrained cases, provided that the momentum of the jet is
sufficiently higher than that of the freestream. Looking in more detail at the
boundary layer of the confined jet as the Reynolds number increases, the flow
tends to compress slightly, which inhibits its boundary layer development.
This delay in boundary layer growth hinders the entrainment of the flow, main-
taining the composition of the je t and increasing the bulk jet velocity. The goal of
this work is to use blowing slots to induce the C oanda effect in the leading edges
(LE) and trailing edges (TE) of the airfoil.
Param eters other than the freestream velocity that affect the ability for flow to
remain attached to a curved surface include the four primary variables-radius of
curvature, slot location, slot size (height and span), and blowing pressure-which
are characterized by the coefficient C, as defined in Eq.
2 ) :
pj
vj hb
1
2p ,
V i
b
,
=
where
p
is the density, V is velocity, h is the slot height, b is the span, c is the
airfoil chord, and the subscripts and represent the je t and freestream
values, respectively. General trends exist for these parameters. For instance, as
the slot size is reduced, the separation of the flow is delayed because less mass
flow can be added to the boundary layer, and because of higher jet velocity (at
the same C, . For a given slot location, an increase in the radius of curvature,
or the blown pressure, results in a delay of the onset of flow separation. This
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 285/637
COANDA EFFECT ON THE BELL A821201 AIRFOIL 279
delay in separation, controlled by the interaction of all three of the variables,
experiences a theoretical upper limit of 245 deg, measured from the slot
opening, according to Newman.’ These Coanda jets, placed on the LE and TE
of the main wing of the V-22 “Osprey,” can be used to reduce the downforce
caused by the rotorwash.
This paper expands upon the experimental results shown by Angle et a1.,2 and
compares computational methods to simulate this flow phenomena. Discussions
of the experimental apparatus and computation methods are presented. The
experimental results and computational fluid dynamic (CFD) predictions are
shown, together with their comparison and recommendations for further testing.
11
Experimental Apparatus and Procedure
A model of the Bell A821201 airfoil coordinates provided in Felker’ and
Felker and Light,g with a 19-in. chord length and an 18-in. span (Fig. 1) was con-
structed and tested at the West Virginia University Aerodynamic Wind Tunnel
Facility. The reader is referred to Angle et a1.* for additional information on
the model geometry and wind tunnel facility. This model produced a test
section blockage of 15 , which is relatively high fo r wind tunnel testing.
How ever, this size was needed for the desired instrumentation fo r the two-dimen-
sional preliminary testing of this concept. Force coefficients can be adjusted to
account for solid blockage using the formula presented by Barlow et al.3 and
restated in Eq. 3):
Fig. 1 CAD drawing
of
the experimental model.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 286/637
280
G. ANGLE II, B. O’HARA, W . HUEBSCH, AND
J.
SMITH
where CD s the adjusted download force coefficient, CD is the measured down-
load force coefficient, A is the model frontal area, and is the test section cross-
sectional area. Surface pressure readings w ere taken on this model using m ultiple
static pressure ports, as discussed by Angle et a1.2 The aerodynamic forces were
measured using a three-load cell (0-25 lb each) system, two in the download
direction to provide force and moment, and the third in the normal direction to
measure force, as shown in Fig.
2.
The structure supporting the model in the
test section produced a measurable drag that had to be accounted for when calcu-
lating the drag on the wing. Because the experimental apparatus was the same,
the instrumentation error is the same as that discussed by Angle et a1.,2 which
was found to be 0.1
1
lb for the force measurements and an error of 0.19 in the
pressure coefficient value.
The drag on the support apparatus was determined from the standard drag
coefficient for a cylinder, from Young et al.4 The resulting moments about the
pivot point, above the test section, were removed from the recorded moments,
resulting in
Eq. (4)
for the determination of download force on the model:
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 287/637
COANDA EFFECT ON THE BELL A821201 AIRFOIL 281
where D represents the resulting forces, L denotes the corresponding moment
arms, and the subscripts are as shown in Fig. 2. This figure shows the attachment
points for the two load cells used to determine the drag on the system, as in Angle
et a1.,2 and a third load cell was added to measure the force normal to the drag.
Surface pressure taps were also provided on the model, but not repeated for the
tests associated with this phase of the project.
The large test section, 4 x 6ft, of the Closed Loop Wind Tunnel at West
Virginia University, was used for this testing. The maximum airspeed of this
test section is just above
60
ft/s; however, because of blockage effects, only
59 ft/s could be achieved during testing. The resulting Reynolds number
was 6 x
lo5
based on airfoil chord length. Once the model was installed in
the test section and the load cells calibrated, testing was conducted with the
results shown in Fig.
3.
To perform a test the wind tunnel was brought to the
desired airspeed and data were collected from the load ce lls. Data were collected
for each test point for a four-minute test samp le, with repeats of the baseline after
every five tests. Use of the term baseline refers to testing with zero pressure on
both the LE and TE blowing slots. After collection of the data, the following
procedure was used to reduce the raw voltage data from the load cells. The
voltage values were taken through the calibration curves shown in Fig. 3. A base-
line average was computed from the three baseline runs to be used as the refer-
ence force as well as the zero pressure force value. A simple percent reduction
was calculated between the time average data for each run and the baseline
average.
15
m
-0.01 0.005 0 0.005 0.01 0.01
5 0.02
Signal Vo l tage (V)
Fig. 3 Calibration curves for the three load cells.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 288/637
282
G. ANGLE II, B. O HARA, W . HUEBSCH, AND
J.
SMITH
111. Com putational Model and Procedure
Because Fluent 6.1 was the computational solver used for this study, its grid
generator, Gambit 2.1, was used to create the computational grid and boundaries.
A two-dimensional grid was created based on a cross-section of the WV U wind
tunnel. The overall dimensions of the grid can be seen in Fig. 4. The general setup
used was a two-dimensional cross section of the wind-tunnel test section with a
scale model of a Bell A821201 airfoil equipped with 0.0625-in. blowing slots.
The LE and TE blowing slots are located at 1.61 and 70.55 of the chord
length, respectively. The chord length for this model as in the experimental
setup was 19 in. The displayed measurement of 16 .76 in., in Fig. 4, is the
length from the LE of the airfoil to the end of the 67 deg deflected flap. The
width of the computational test section was 48 in. and the length was set as
84 in. This length was chosen so that most of the wake profile could be captured.
Gambit 2.1 allowed the creation of various types of boundaries. At the top of
the grid a velocity inlet that produced a uniform airflow downward was created.
The bottom of the grid was specified as a pressure outlet. Each of the blowing
slots was created as velocity inlets. The rest of the boundaries were set as no-
slip walls. The mesh was created using unstructured triangular cells. Additional
grid poin ts were clustered around the blowing slots and immediately dow nstream
of the wing, where large gradients and flow separation were expected. A total of
2131 grid points were created on the surface of the airfoil. This resulted in an
60
84
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 289/637
COANDA EFFECT ON THE BELL A821201 AIRFOIL
283
average y + value of approximately 12 for cells next to the wall and very close to
the blowing slot; Fluent, Inc., recommends having a mesh with
y+
values
between
1
and
5
However, use of the enhanced wall treatment model in
Fluent can allow for coarser meshes to be solved. The entire grid comprised
2,184,528 triangular cells and 1,093,464 nodes. Figure
5
shows the overall
grid, and Fig. 6 a close-up of the grid on the LE.
Attempting to match the experiment, the boundary and initial conditions
needed to be correlated to accepted input values. The velocity inlet at the top
of the grid was set to 59 ft /s directly downward, the mean experimental velocity.
Because the experiment used varying plenum pressures, both the experimental
and computational inputs were converted to the blowing slot momentum coeffi-
cient. Using
Eq.
(2), Tables 1 and 2 were created, which show the blowing
slot momentum coefficient for the computational and experimental tests. The
blowing slot velocities were then determined to be 0, 10, 60, 130, and 200 ft/s.
The rest of the initial conditions were set as standard atmospheric conditions.
Fig.
5
Full computational grid.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 290/637
284
G. ANGLE II, B. O’HARA, W . HUEBSCH, AND
J.
SMITH
Table
1
Computational slot velocity and
correspond ing momentum coefficient
Slot
v
t/s
P
0
10
60
130
2
0
0.0002
0.0068
0.0319
0.0756
After the initial conditions w ere all set, the first step was to find an initial sol-
ution. The commercially available codes in Fluent 6.1 were used as the solver.
For each blowing slot velocity, a laminar solution with first-order accuracy
was found. This was done to help the higher order solver converge upon a sol-
ution. Each lam inar solution was than solved again using second-order upwind-
ing accuracy and Fluent 6.1’s two-equation renormalization group kinetic
energy-dissipation (RNG k-e) solver. Other solver settings used in Fluent 6.1
were two-dimensional, double precision, segregated solver, cell-based solution,
with enhanced wall treatment. The
RNG
k-e solv er with enhanced wall treatment
was selected bec ause it has been fo und to yield good results fo r cases dealing with
circulation control (CC ) by Ch ang et al.5 A quick study was also carried out com-
paring the different solver types available in Fluent 6.1. The RNG k-e solver
Fig.
6
Computational grid near the leading edge.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 291/637
COANDA EFFECT ON THE BELL A821201 AIRFOIL 285
Table 2 Experimental slot pressure and
corresponding momentum coefficient
Slot
P
si
c
0
5
10
15
20
25
0
0.0116
0.0232
0.0348
0.0464
0.0580
produced results that appeared to be very realistic, while taking considerably less
time than the five-equation RSM turbulence model.
IV. Experimental
Results
Data from the normal load cell were found to be negligible because they were
of the order of less than
1
lb. This corresponds to a deflection of less than five-
thousandths of an inch, indicating an error of the order of the resolution of the
load cells in the download direction. The baseline test case (nonactive
blowing) experienced a total download force of 18.75 lb, measured from the
two load cells, at the test Reynolds number of 5.94 x
lo5.
As seen in Fig. 7,
which is nondimensionalized by dividing out the no-blowing download force,
for lower blowing coefficients there is an increase in the download force with
1.04
1
1.02
u
g 0.98
y 1
B
n
-
-
m
.96
q
0.94
0.92
-.-
0
0.02 0.04
Blowing Coef fic ient
( )
Fig.
7
Download force variation with blowing coefficient.
0.06
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 292/637
286
G. ANGLE II, B. O’HARA, W . HUEBSCH, AND
J.
SMITH
the LE slot active, and a smaller increase when the TE is activated. As the
blowing coefficient is increased, the LE slot decreases the nondimensional dow n-
load force, while the TE slot produces a fairly constant increase in download
above the baseline value as the blowing coefficient is increased. The curve
showing data for both slots active demonstrates the combined effects of the
individual blowing slots.
The data are summarized in Table 3, where a positive value indicates a
reduction in the download on the A821201 airfoil model. These results show
that, with the current configuration, the LE is more effective at reducing the down-
load force. However, when using both slots there is still an 8 reduction in the
force. It should be noted that no effort has yet been made to optimize slot place-
ment and that the TE flap is deflected according to current V-22 operating
practices. These results do show the overall viability of the blowing slot mechan-
ism as a means of reducing the downwash force. There is also the potential to use a
variant of the technique discussed in this paper to assist in the control of the pitch-
ing moment of the airfoil. By adjusting the blowing pressures separately, the
pitching moment can be altered. With further testing, this potential benefit can
be be tter defined. Additional experimental data can be found in A ngle e t a1.*
V.
Com putational Results
Immediately behind the separation points, both at the LE and TE, turbulent
eddies formed. These turbulent eddies generally caused lower pressures that
increased the download. With the blowing slots in place
it
was found that
these eddies could be reduced in size. This reduction in size is a result of the pos-
ition of the separation point. Although the separation point is a good indicator of
how much the download is being reduced, it is also useful to be able to visualize
the areas of lower pressure, for exam ple, where the flow is circulating, with path-
line, vorticity, and vector plots.
Circulating flow is easily seen by plotting pathlines in the regions behind the
blowing slots. Figures 8 and 9 show particle tracks, which are colored by particle
identification, near the LE and deflected flap of the airfoil. These figures are
helpful from a potential flow point of view and seem to be similar to other
Table 3 Experimental reductions in download force
Percent reduction
Internal
pressure, psig c
LE
only
TE
only Both
0
5
10
15
20
25
0
0 0 0
0.01 2.84
-0.35 -3.12
0.02 0.63 -1.16 .59
0.03 3.88
1.07 2.29
0.05
6.67 -1.13 5.08
0.06 9.23 -0.77 8.68
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 293/637
COANDA EFFECT ON THE BELL A821201 AIRFOIL 287
Fig. 8 Pathlines colored by Particle ID near the leading edge.
active CC studies such as the one carried out by Sw anson et a1.6 Pathlines c an only
tell part of the story for download reduction; they illustrate the path of ai r particles
but do not really show turbulence or velocity gradients. Upstream of the wing,
the flow appears to be mostly uniform and lamin ar, and immediately downstream
of the wing large amounts of turbulence form. This is shown in Fig. 10, a
Fig.
9
Pathlines colored
by
Particle
ID
near the trailing edge.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 294/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 295/637
COANDA EFFECT ON THE BELL A821201 AIRFOIL 289
Fig.
12
Vector plot near the trailing edge.
shows a similar trend. The computational results were all computed a t sea-level
standard atmospheric conditions whereas the experiment was conducted in Mor-
gantown, W V, at an elevation of 1240 ft. Despite this difference, which was
accounted for in the use of force and pressure coefficients, the amount of down-
load when compared to the baseline tests for each approach is very similar, as
shown in Fig. 14. Figure 14 is a more appropriate indicator of how the
....................................................................................
-
A
-
...................................................................................................
-
2
xperimental
5 ..............................................................................
0 1
0.01
0.02
0.03 0.04
0.05
0.08 0.07 0.08
Blow ing Coef f i c ien t Cp)
Fig.
13
Comparison
of
download between experimental and computational
techniques.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 296/637
0 0.01
0.02 0.03 0.04 0.05 0.06 0.07 0.08
Blo win g Coeff ic ient (Ck)
Fig. 14 Comparison
of
the percent download reduction between experimental and
computational techniques.
computational tests compare with the experiments than Fig. 13 because of its
nondimensional nature.
VI. Conclusions
This chapter has presented CC as a method to reduce the force felt by a surface
in the wake of a rotor. Typical applications of
CC
are looking at the airflow over
the surface of the airfoil, where this particular application is looking at the flow
approximately normal,
5
deg angle of attack using the conventional definition.
This difference in flow characteristics seems to have slightly altered the trends
present in the conventional application of active CC methods. At low blowing
coefficients there is a small increase in the force, followed by a decrease in the
force. Some of this decrease in force is a result of the reduction in wake area,
but it is not clear that this is the only aspect capable of reducing the force.
Further investigation will help clarify the true force reducing mechanism(s),
which could include the jet momentum conservation.
The trends in the experimental and computational tests show that active CC,
through the use of blowing slots on the LE and TE of the Bell A821201 airfoil,
can reduce the download force felt from the rotor wash of a tilt-rotor aircraft.
Experimental testing demonstrated a reduction of approximately 10 from the
baseline 18.7 lb download. The baseline download of the computational tunnel
simulation was found to be 241b and had a maximum reduction of around
12 . The percent reduction of the download provided a reasonable match in
both the trend and magnitude.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 297/637
COANDA EFFECT ON THE BELL A821201 AIRFOIL 291
Many aspects of using CC need to be investigated further. Some of these
include looking into optimizing the placement of the LE and TE slots. Current
testing has only studied one location for each of the slots. With decen t agreement
between the computational model and the experimental results, the process of
finding optimum placement will be simplified. Currently, new experimental
and continuing computational models are under development to address
aspects of the current data. The new experimental model will be sized to fit
into the small test section of the WVU Closed Loop Wind Tunnel to allow for
testing at different Reynolds numbers and take test section blockage into
account. A cost/benefits analysis is also being conducted to determine the prac-
tical application of using such a system on a tilt-rotor aircraft to increase the such
aircrafts’ performance.
References
‘Newman, B. G., The Dejexion of Plane Jets By Adjacent Boundaries-Coanda Effect;
Contained in Boundary Layer and Flow Control,
Vol. 1, Pergamon Press, New York,
1961, p. 232.
’Angle,
G.,
Riba, C., Huebsch, W., Thompson, G., and Smith, J., “Download Wake
Reduction Investigation for Application on the V-22 ‘Osprey’,” Society of Automotive
Engineers Technical Paper 2003-01-3020, Sept. 2003.
3Barlow, J. B., Rae, W. H., and Pope, A., Low-Speed Wind Tunnel Testing, 3rd Ed.,
Wiley, New York, 1999.
4Young, D. F., Munson, B. R., and Okiishi, T. H.,
A Brief Introduction to Fluid
Mechanics, Wiley, New York, 1997.
’Chang, P. A. 111,Slomski, J., Marino, T., and Ebert, M. P., “Numerical Simulation of
Two- and Three-Dimensional Circulation Control Problems,” A I M Paper 2005-0080,
Jan. 2004.
wanson, R. C. , Rumsey, C . L., and Anders, S. G., “Progress Towards Computational
Method for Circulation Control Airfoils,” AIAA Paper 2005-0089, Jan. 2005.
’Riba, C. A., “Circulation Control for Download Wake Reduction in the V-22
Aircraft,” Masters Thesis, Department of Mechanical and Aerospace Engineering, West
Virginia Univ., Morgantown, WV, 2003.
‘Felker, F. F., “Wing Download Results from a Test of a 0.658-Scale V-22 Rotor and
Wing,”
Journal of the American Helicopter Society,
1992, pp. 58-63.
’Felker, F. F., and Light, J.
S.,
“Reduction of Tilt Rotor Download Using Circulation
Control,” Proceedings of the Circulation-Control Workshop, 1986, pp. 429-447.
Englar, R. J., “Experimental Investigation of the High Velocity Coanda Wall Jet
Applied to Bluff Trailing Edge Circulation Control Airfoils,” Masters Thesis, Univ. of
Maryland, College Park, MD, 1973.
“Felker, F. F., Shinoda, P. R., Heffernam, R. M., and Sheehy, H. F., “Wing Force and
Surface Pressure Data from a Hover Test of a 0.658-Scale V-22 Rotor and W ing,” NASA
TM-102244, Feb. 1990.
10
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 298/637
Chapter 10
Novel Flow Control Method for Airfoil Performance
Enhancement Using Co Flow Jet
Ge-Cheng Zha* and Craig
D.
Paxtont
University of Miami, Coral Gables, Florida
Nomenclature
C L= lift coefficient
C =
drag coefficient
C,
=
momentum coefficient
c, = specific fuel consumption
D = drag
E
= endurance
F = thrust
m = mass flow rate
k = turbulent kinetic energy
M = Mach number
P ,
=
total pressure
R = range
Re = Reynolds number
S
=
wing span area b
x chord)
U
= velocity
V
= velocity
PR
= total pressure ratio of engine compressor
Wo= takeoff gross weight
W1 = empty weight
y+
=
nondimensional length scale for turbulent boundary layer
*Associate Professor, Department of M echanical and Aerospace Engineering.
+Graduate Student, Department of M echanical and Aerospace Engineering.
Copyright 005 by the American Institute of Aeronautics and Astronautics, Inc. All rights
reserved.
293
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 299/637
294 G.-C.HA AND C. D. PAXTON
a = angle of attack
p = density
7= efficiency
=
ratio of specific heats
E = turbulent dissipation rate
Subscripts
= freestream
=jet injection
L = landing
TO = takeoff
T
=
touch ground
I. Introduction
ACHIEVE high-performance aircraft design, revolutionary technology
T dvancement should be pursued to d ramatically reduce the weight of aircraft
and fuel consumption, and significantly increase aircraft mission payload and
maneuverability. Both military and commercial aircraft will benefit from the
technology.
Flow control (FC) is the most promising route to break through the conven-
tional aerodynamic design limit and bring dramatic performance improvement
to aircraft.'-3 The National Aeronautics and Space Adm inistration (NASA),
U.S. Air Force, and aerospace industry have recently made great efforts to
deve lop f low con t ro l t e~hnology .~ -~o enhance lift and suppress separation,
various flow control techniques have been used, including a rotating cylinder
at the leading ed ge (L E) and trailing ed ge (TE),3,899 irculation control (CC)
using tangential blowing at LE and TE,1°-16 multi-element pulsed
jet separation c o n t r 0 1 , ' ~ -~ ~nd
so
on.
When a flow control technique is developed, there are three issues that may
need to be considered: 1) effectiveness-the FC method should provide substan-
tial improvement in aerodynamic performance, which primarily includes lift
enhancement, drag reduction, and stall margin increase (suppression of separ-
ation); 2 energy efficiency-the FC method should not cause significantly
more energy expenditure, otherwise the penalty may outweigh the benefit for
the whole aircraft as a system, including minimal penalty to the propulsion
system and minimal weight increase resulting from the FC system;
3
easy
implementation-the FC technique should not be too difficult to implem ent.
The rotating cylinder method is generally most effective when the LE or T E
are thick, and hence may be more applicable to a low-speed airfoil. It also
needs a system to drive the rotating system and can increase aircraft weight.
The multi-element airfoil can generate high lift, but generally comes with
large drag and weight penalty due to the moving parts. In addition, the high-
lift flap system increases noise during landing.23
relies on a local favorable pressure gradient on a curved
surface to attach the flow-the Coanda effect. Such a favorable pressure gradient
A
cc
airfoillO,l
1,15,16
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 300/637
CONTROL METHOD USING CO-FLOW JET
295
exists at the airfoil LE as a result of the suction and at the end of the TE because
of the low base pressure when the TE is blunt. T o make the C C airfoil effective,
the blunt TE is therefore needed. However, this will create large drag at cruise.
To overcome the dependence on a large TE for the CC airfoil, a movable flap
at the airfoil T E has been suggested by Englar. The moving parts will increase
the weight penalty to the aircraft. At large angle of attack (AOA), because the
main flow cannot resist the large adverse pressure gradient, the local TE favor-
able pressure gradient cannot be achieved and hence the Coanda effect is difficult
to maintain. If only TE blowing is used, the CC airfoil will usually stall at a
smaller AOA than the regu lar noncontrolled airfoil.24 To increase stall margin,
LE blowing needs to be added.24
A considerably high penalty placed by the CC airfoil on the propulsion system
is the dumped blowing jet mass flow. The blowing air for the wing is usually
sourced from the engine compressor bleed. The mass flow rate of the engine
bleed is directly proportional to the decrease of thrust; that is, an engine will
suffer 1% thrust decrease for 1% bleed flow used for wing flow control, and
suffer 1-3% fuel consumption increase depending on whether the bleed is
from the compressor front stage or back stage.
To avoid the jet mass flow rate penalty caused by blowing, the synthetic je t or
pulsed je t with open or closed loop feedback control are These methods
need a jet generation system, and complicated actuation and sensor systems,
which may increase the degree of difficulty in implementing the FC system
and increase the weight of the aircraft as well. Because the interaction of the syn-
thetic je t with the main flow is generally weak, its effectiveness in enhancing lift
and suppressing separation may not be as dramatic as desired. For example, the
results show n in Ref. 19 using the periodic synthetic jet show about 35% increase
of the
C
and little increase of stall AOA, while the co-flow je t airfoil tested in
Ref. 25 increases the
C
and AOA range by 220 and 153%, respectively, with
C = 0.28. A movable flap is also used with the synthetic je t flow control airfoil
studied in Ref. 19, which will increase the aircraft weight.
The new airfoil flow control technique using the co-flow je t (CFJ)26suggested
in this paper is aimed at considering all the three issues mentioned above, that is,
effectiveness, energy efficiency, and ease of implementation. The co-flow jet
airfoil opens an injection slot near the LE and a suction slot near the TE on
the airfoil suction surface. The slots are opened by translating a great portion
of the suction surface downward. A high-energy jet is injected tangentially
near the LE, and the same amount of mass flow is sucked in near the TE. The
turbulent shear layer between the main flow and the jet causes strong turbulence
diffusion and m ixing, which enhances lateral transport of energy from the je t to
the main flow and allows the main flow to overcome a severe adverse pressure
gradient and remain attached at a high AOA. The strong adverse pressure gradi-
ent enhances je t mixing,27 and the stall margin is significantly increased. The
high jet velocity induces high main flow velocity on the suction surface and
hence creates high circulation and lift. The energized main flow will fill the
wake velocity deficit, which results in a reduced drag or thrust (negative drag).
A CFJ airfoil wing does not need a high-lift flap system and can therefore
reduce noise during landing. A CFJ airfoil does not rely on the Coanda effect
at the LE or TE and thick LE or TE are not required. Hence, the low form
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 301/637
296
G.-C.HA AND C. D. PAXTON
drag of modem airfoils can be maintained. The CFJ technique can be applied to
any type of airfoil, including low-speed thick a irfoils and high-speed thin airfoils.
The level of lift enhancement, drag reduction, and stall mar in increase of the
CFJ
airfoil is very dramatic, as proved by wind-tunnel tests.
Because a
CFJ
airfoil blows and sucks the sam e amount of mass flow, the je t
mass flow can be recirculated through the propulsion system instead of being
dumped away. This can significantly reduce the penalty of energy expenditure
to the overall airframe-propulsion system when compared to the blowing-only
methods. The
CFJ
can always be on during the entire flight mission. The lift
enhancement and drag reduction can be controlled by adjusting the injection
total pressure, and hence the je t mass flow rate, throughout the mission according
to different needs. No moving parts a re required.
The
CFJ
airfoil concept suggested in this paper appears to have the following
advantages:
1 It is very effective in enhancing lift and suppressing separation.
2 It dramatically reduces drag o r creates thrust (like a bird wing generating
both lift and thrust) and hence can achieve very high C,/Co at low
AOA
(cruise),
and very high lift and drag at high AOA (takeoff and landing).
,**
3
It can significantly increase
AOA
operating range and stall margin.
4)
It imparts only a small penalty to the propulsion system.
5 )
It can be applied to any airfoil, thick or thin.
6)
It can be used for the entire flying mission instead of only during takeoff
7
It can be used for low- and high-speed aircraft.
8)
It is easy to implement, with no moving parts.
The preceding advantages of the CFJ airfoil may derive the following superior
aircraft performances: 1 extremely short takeoff and landing distances;
2
super-
sonic aircraft having small wing size matching cruise need, but also having high
subsonic performance (e.g., high lift as low drag at M
< 1); 3
high maneuver-
ability, high safety, and fast acceleration military aircraft;
4)
very economic
fuel consumption; 5 ) small wing span for easy storage, light weight, and
reduced skin friction and form drag; 6 low noise because of no high-lift flap
system at landing (at takeoff, the wing thrust or reduced wing drag will rely
less on the engine thrust and hence will have less nozzle jet velocity, which
will result in lower noise); 7 heavy lift rotorcraft with effectively no dynamic
stall; and 8) stealth aircraft with no moving control surface.
The purpose of this research project is to study the working principle and
demonstrate the superior performance of the
CFJ
airfoil based on
CFD
simulation
and experiment. This paper p resents the
CFD
results and analysis, which are the
basis fo r the wind-tunnel tests of proof of concept conducted in Refs. 25 and 28.
The detailed
CFD
da ta also provide a very useful qualitative physical insight into
the working mechanism of the
CFJ
airfoil.
and landing.
11 Results and Discussion
As a reference, the CFJ airfoil is compared with the baseline airfoil with no
flow control. Figure 1 shows the baseline airfoil,
NACA2415,
and the airfoil
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 302/637
CONTROL METHOD USING CO-FLOW JET
basel ine air fo i l
suct ion slot
nject ion
slot
297
Fig.
1
Baseline airfoil
NACA2415
and the airfoil with
CFJ
slot.
with CFJ slot. The chord length is 0.3 m. The coflow jet airfoil is modified from
the baseline airfoil by translating the suction surface vertically lower by 1.67% of
the chord. The slot surface shape is exactly the same as the original baseline
airfoil suction surface. The slot inlet and exit are located at 6.72 and 88.72%
of the chord from the LE. The slot inlet and exit faces are normal to the slot
surface to ensure that the jet will be tangential to the main flow. The slot inlet
and exit area are 1.56 and 1.63% of the chord.
The Fluent CF D software is used as the tool to simulate the airfoil flows in this
study. The mean flow governing equations are the two-dimensional compressible
Navier-Stokes equations. The
k - E
turbulence model with wall function is used
to save CPU time. The solutions of two typical cases are compared with the
solutions using the
k--E
model integrating to the wall. The results show little
difference. When the wall function is used, y t is of the order of 15-100.
When the turbulent boundary layer is solved by integrating to the wall, the y?
is of the order of 1. The wall function method therefore requires less grid to
resolve the boundary layer and significantly saves CPU time. The reason that
k--E
is used is because of its capability of taking into account the turbulent bound-
ary layer history effect by solving the complete transport equations of
k
and
&,
and
the
k--E
model is more capable than algebraic models to predict the separated
flows, which occur when the airfoil stallsat high AOA.
The full turbulent boundary layer assumption is used because the C FD solver
does not have a transition model. The 0-mesh is generated as shown in the
zoomed region around the airfoil in Fig. 2. The baseline mesh has the dimensions
240
x
100 in the d irection around the airfoil and in the radial direction, respect-
ively. In the CFJ slot, the mesh size is 80 x 12 in the streamwise and spanwise
directions, respectively. A rectangular farfield boundary is used with the down-
stream boundary extended to 30 chord length, upstream, lower and upper bound-
ry
to 20 chord length. The y+ ranges from 15 -30 on the airfoil surface. The
freestream Mach number is 0.3 and the Reynolds number is 1.9 x lo6. For all
the computation, the jet inlet holds a constant total pressure equal to 1.315Pt,.
The static pressure at jet suction is iterated to match the jet injection mass flow
rate.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 303/637
298
G.-C.HA AND C. D. PAXTON
Fig. 2 Zoomed m esh around the airfoil with co-flow slot.
A. CFJ Airfoil Performance
Figure 3 presents the lift coefficient against angle of attack for the baseline
airfoil and the CFJ airfoil. For the baseline airfoil, the lift coefficient predicted
by CFD agrees excellently with the experiment results at Re =
3
x
lo6
before
CFD predicts a little delayed stall and higher lift coefficient in the stall
region. Figure
3
indicates that the lift of the CFJ airfoil is increased significantly.
The zero-lift
AOA
for the baseline airfoil is - 2 deg, and is -6 deg for the CFJ
airfoil. The stall
AOA
is increased by 2 deg. Hence the operating range of
AOA
is
increased totally by
38%.
The maximum lift value is increased by
SO%,
which
is the m inimum increase in the order of magnitude. W hen the AOA is decreased,
the lift increase is greater in percentage terms. For example, at
AOA
= 2 deg, the
lift increase is 250 .
For the CFJ airfoil, a few selected points are recalculated using the refined
mesh of dimensions
480
x
200
around the airfoil and
160
x
30
in the slot.
The refined mesh lift coefficients are shown in Fig. 3 and agree excellently
with the baseline mesh, which indicates that the numerical solutions are con-
verged based on the mesh size.
Figure
4
show s the streamlines at AOA
= 20
deg. The baseline airfoil has a
massive separation, whereas the CFJ airfoil flow is nicely attached. The attached
flow is mainly a result of the turbulent mixing,30 which transfers energy from the
jet to the main flow so that the main flow can overcome the severe adverse
pressure gradient to stay attached.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 304/637
CONTROL METHOD USING CO-FLOW JET
299
AOA
Fig. 3 Lift coefficient against angle of attack.
Figure 5 is the isentropic Mach number distribution on the surface of the
airfoil at AOA = 20 deg. The isentropic Mach number is defined as
The isentropic Mach number is only a function of surface static pressure.
Hence it indirectly gives the surface static pressure. At the same time, the isen-
tropic Mach number also indicates the approximate Mach number outside of
the wall boundary layer assuming that the total pressure loss is small.
Figure
5
show s that the CFJ airfoil creates a very strong suction effect near the
LE and the flow is accelerated from the freestream M ach num ber 0.3 to the peak
Mach number 1.7. The supersonic flow is only in the LE region and smoothly
transits to subsonic flow with no shock wave created. The peak Mach number
of the baseline airfoil is about
0.9.
However, the baseline airfoil cannot sustain
the severe pressure gradient and the massive separation yields small loading on
the aft portion of the airfoil. The CFJ airfoil has much higher LE acceleration
and diffusion on the suction surface and stronger deceleration on the pressure
surface, which results in higher lift and circulation. Figure
5
also shows that
the LE stagnation point of the CFJ airfoil is located more downstream than
that of the baseline airfoil because of higher circulation. The first spikes
near the LE are caused by the CFJ injection, which induces the strong LE
suction through turbulence mixing. The shape of the spike is not necessarily
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 305/637
300
1.6
1.4
5 1.2
f
= 1 -
8
0.8
c
.-
0.6
-
G.-C.HA AND C. D. PAXTON
-
-
-
-
baseline airfoil
Fig. 4 Streamlines at angle of attack
of
20 deg.
_ _ _ _ _ _
aseline
I of low
’.
W h J I I I I I I I I I I I I I I I I I I
0 0.25 0.5
0 75
1
WChord
Fig.
5
Surface isentropic Mach number distribution at angle
of
attack
of
20 deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 306/637
CONTROL METHOD USING CO-FLOW JET
30
accurate and may be created by the numerical boundary condition treatment.
The second spike near the TE is a result of the low-pressure suction at the jet
suction slot.
Figure
6
presents, the Mach number contours in the LE region at
AOA = 20 deg for the baseline and CFJ airfoils. It shows that the CFJ airfoil
has a local supersonic region near the LE. The high-energy jet mixes with the
mainflow through a turbulent shear layer.
It should be noted that the fundamental mechanism of the CFJ airfoil is the
turbulent mixing between the jet and the main flow, which transfers energy
baseline airfoil
Fig. 6 Mach number contours at 01 = 20 deg for the baseline and CFJ airfoil.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 307/637
302
G.-C.HA AND C. D. PAXTON
from the jet to the main flow. A high mixing rate is therefore desirable. As indi-
cated by Gre itzer et al.,27 the adverse pressure gradient enhances je t mixing.
Based on this principle, the injection slot of the CFJ airfoil is located dow nstream
of the LE suction, as shown in Fig.
5 .
After the LE suction, the pressure continu-
ously increases until reaching the suction slot near the TE. This is very different
from the CC airfoil technique, which places the injection right on the geometric
leading position, where the LE suction starts, and a strong local favorable
pressure gradient exists because of the suction.
The other factor that enhances the turbulent jet mixing is having a long enough
distance in which the mixing can occur; that is, the suction slot should be located
as close to the TE as possible, subject to geometric constraint. How ever, the CFD
simulation indicates that the CFJ airfoil performance is more sensitive to the
injection location than to the suction slot location. In general, the closer the injec-
tion to the LE, the more effective the CFJ, but the injection must be located down-
stream of the LE suction peak.
Unlike the C C airfoil, which blows at the LE (near the stagnation point with
high pressure) and at the TE where the pressure is high, the CFJ airfoil has the
injection downstream of the LE suction peak where the pressure is near its
lowest, and has the suction near the TE where the pressure is nearly highest
(except for the stagnation point). The CFJ airfoil therefore creates a more favor-
able pressure condition for injection and suction and may need less energy to
pum p the same amount of jet mass flow than does a CC airfoil.
In this study, although the AO A varies, the CFJ injection total pressure is held
constant to simulate passive flow control. At different AOA, the main flow will
have different static pressure at the location of the je t injection, which determines
the jet mass flow rate of the je t and the je t injection velocity. The je t mom entum
coefficient therefore varies with AOA. The jet momentum coefficient based on
the conventional definition is given by
m
vj
c -
-.5pmU&S
where m is the injection m ass flow rate,
vj
is the injection velocity, p, and
U ,
are the freestream density and velocity, and S is the wing span area. Figure
7
shows the variation of C , with AOA. When AOA varies from - 8 to 22 deg,
C , increases from 0.15 to 0.25.
Figure 8 is the drag polar for the baseline airfoil and the CFJ airfoil. When the
AOA is high, both the lift and drag of the CFJ airfoil are significantly higher than
for the baseline airfoil. When comparing the maximum lift points for the two air-
foils, the drag of the CFJ airfoil is 160% higher than that of the baseline airfoil.
However, when AOA <
4
deg, the lift of the CFJ airfoil is significantly higher
and the drag is significantly lower than that of the baseline airfoil.
When AOA
<
0 deg, the lift coefficient is still very large
Cl =
0.862 at
AOA = 0 deg; see Fig. 3 , but the drag becomes negative and a thrust is gener-
ated. The thrust is primarily generated from the strong LE suction, which is the
same mechanism as a flapping bird wing generating both lift and
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 308/637
CONTROL METHOD USING CO-FLOW JET
303
0.2
c
0
0.05
0
Fig.
7
Jet mom entum coefficient against angle of attack.
2.5 -
2 -
1.5 -
1 -
0.5 -
I
base l ine
cof low
- - - - - .
-0.5
I
I
I \
I
I
I
I
I
I
I
I
I
I
I
I
I
0 0.1 0.2 0.3
Cd
Fig.
8
Drag polar for the baseline and CFJ airfoil.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 309/637
304
0.3
0.25
0.2
0.15
3
0.1
0.05
0 -
-0.05
G.-C.HA AND C. D. PAXTON
-
-
Cdpressure
Cd f r ic t ion
- - - - - .
- _ - _ -
-
-
-
. I . .__
c
I L r r r r ~ ' ' ' ~ ' ' ' ' ~ ' ' ' ' ~ ' ' ' ' ~ ' ' ' ' ~ ' ' '
Fig.
9
Calculated drag coefficients against angle of attack for CFJ airfoil.
The d rag of an airfoil arises from two sources, friction drag and pressure drag
(form drag). The friction drag will always be in the opposite direction of the
flight, that is, always positive. The negative drag must therefore be from the
pressure drag. This can be seen from Fig. 9, which shows the friction drag,
pressure drag, and total drag for the CFJ airfoil. Figure 9 indicates that the friction
drag is fairly constant and decreases slightly near stall. However, the pressure
drag varies largely. The pressure drag is the dominant contribution to the
total drag near stall. When AOA is decreased, the pressure drag also decreases
monotonically. When AOA
<
4 deg, the pressure drag becomes negative, and
the total drag is reduced negative values when AOA
<
0 deg because of the
strong LE suction.
Figure 10shows the drag distribution of the baseline airfoil. Similar to the CFJ
airfoil, the friction drag is also fairly constant compared with the pressure drag.
The pressure drag decreases when the AOA is decreased from the stall region.
However, the pressure drag increases when the decreasing
AOA
passes the
zero lift point and does not become negative. This is because there is no
AOA
that can create a strong enough LE suction for the baseline airfoil.
The negative drag may also be explained from the control volume point of
view. The high-velocity jet transfers the kinetic energy to the main flow
because of turbulent mixing. When the
AOA
is not large, the diffusion is not
severe. The main flow on the suction surface has a large streamwise velocity
past the TE
so
that the streamwise velocity in the wake region is greater than
the freestream velocity. This can be seen in Fig. 11, which shows the wake
shape of the baseline and the CFJ airfoil at one chord length downstream
of the TE. The wake of the baseline airfoil has the usual defect shape,
whereas the wake of the CFJ airfoil has a protruding shape. Figure
12
presents
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 310/637
CONTROL METHOD
USING CO-FLOW JET 305
/
t
0.05
0
AOA
Fig.
10
Calculated drag coefficients against angle of attack for baseline airfoil.
1.1
_ _ _ _ _ _ .aseline
1.075
coflow
1.05
t
1.025
0.975
0.95
.925
Fig. 11 Wake shape for the baseline and CFJ airfoil at (Y =
0
deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 311/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 312/637
CONTROL METHOD USING CO-FLOW JET 307
When U is greater than
Urn,
he drag is negative and becomes thrust. W hen the
AO A is very large, the je t energy is mostly used to diffuse the flow to ma ke the
flow attached. For the current study, with a constant jet inlet total pressure of
1.35P,, at AOA
=
20
deg, the CFJ does not provide enough energy to the
main flow and the wake velocity deficit is very large. Th e pressure d rag is there-
fore overwhelming, which is desirable for short distance landing.
B.
Energy Expenditure
The CFJ airfoil achieves performance enhancement using the powered co-
flow je t, which will involve a certain amount of energy cost. The hypothesis is
that the performance gain from increased lift, reduced drag, and increased stall
margin will outweigh the cost of the energy expenditure of the jet; that is, the
benefit will be realized when the airframe and propulsion are integrated as one
system, because the je t is usually sourced from the engin e. The analysis of this
section is to provide the theoretical foundation of the quantitative analysis of
mission analysis given in the next section.
Assuming a jet engine is used to power the airplane, the power required to
energize the CFJ can be considered as a part of the energy loss of the engine com -
pressor. In othe r words, an extra am oun t of fuel needs to be burned to drive the
compressor with the CFJ pumping system. The loss resulting from the CFJ is
given as
Powercfj
Loss =
Powercompressor
here
yiz fj
is the ratio of the CFJ mass flow rate to the engine mass flow rate,
hcfj= h c f j / h e n g i n e , q&the efficiency, and PR is the total pressure ratio at the
inlet and exit. The hcfj s small and PR,fj is usually also far smaller than
PRcompressor. The penalty to the overall fuel consumption as a result of the loss
of CFJ will therefore be small.
Th e primary penalty to the energy expenditure of the whole aircraft is a result
of the mass flow dumped by the flow control such as in the blowing-only method.
Th e fuel consumption dramatically increases and the thrust is greatly reduced if a
part of the flow is bled from the engine. Th is can be seen by applying a control
volume to an engin e and, assuming that the flow at the engin e exit is expanded to
ambien t pressure, the thrust is given by
F
=
(hinlet + hfue1)vnozzle- inletvinlet
=
h n o zz lev n o zz le - inletvw 5 )
It is obv ious that the bled mass flow from the com pres sor will dec rease hno zz le,
which will directly reduce the thrust. Assuming that the engine is running on the
ground with V = 0, the thrust becomes
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 313/637
308
G.-C.HA AND C. D. PAXTON
where Eq. 6) suggests that the thrust decrease will be directly proportional to the
mass flow dumped if a blowing-only flow control is used. For a recirculating CFJ
airfoil, this serious penalty is avoided.
The specific fuel consumption (SFC) is defined as
It is clear from this that the penalty to the SF C as a result of the dumped flow is
very high.
The only penalty in the CFJ airfoil for aircraft fuel consumption is a result of
the compressor loss given in Eq. (4), which is small and will be easily offset by
the dramatic gain due to the high ratio of
L I D
and reduced w ing weight. These
advantages will be shown in the next section.
C.
The purpose of this section is to conduct a preliminary mission analysis to
study if a CFJ airfoil will be beneficial from the viewpoint of an integrated air-
frame-propulsion system. The military aircraft F-5E is selected as the
example, because a detailed mission analysis has been conducted by Roth
et a1.32-35 and data are available. In Refs. 32 -35 a generalized vehicle thermo-
dynamic loss model is introduced and a loss deck is created for the drag loss of
each component of the aircraft, including the airframe and propulsion systems.
The unification of the airframe drag loss and propulsion system loss makes
it
possible to identify the contribution of each component d rag to the total fuel con-
sumption, which is particularly useful for an aircraft designer in finding and opti-
mizing crucial components to improve the efficiency of the whole aircraft.
The F-5E design mission comprises a subsonic area intercept of 450 n mile
range. The mission includes a maximum power takeoff, climb, subsonic cruise
to the combat zone,
5
min allowance at Mach 1.3, 50,000 ft maximum power
for combat, followed by a subsonic return cruise and 20min reserve loiter,
plus
5
fuel reserve. The aircraft is powered by two J85-GE-21 engines. The
total time for the mission from takeoff to landing is 84 min. Table
1
gives the
detailed breakdown of fuel consumption for the baseline F-5E.33
We created an artificial F-5E using the recirculating CFJ airfoil, named
F-5E-CFJ, to carry out the same mission with the same amount of payload and
fuel. Without actually testing or calculating the F-5E geometry and flowfield,
it is difficult to give the precise values of airfoil performance. Based on the
wind-tunnel tests conducted in Refs. 25 and 28, a conservative estimate of the
fuel consumption may be given.
In the estimation,
CL
is assumed to be twice the baseline
CL
through the
whole mission, with C , = 0.1. The wing lifting surface area therefore only
needs to be half that of the baseline F-5E. The weight of the wing is then also
assumed to be cut by half. The drag coefficient of the CFJ airfoil is assumed to
be one-eighth of the baseline airfoil. Table 2 presents the parameters used to esti-
mate the mission analysis. The integral through the mission for the CFJ F-5E is
then calculated based on the integral results of the baseline aircraft mission
analysis with the assumption that the variation of the penalty
or
benefit is
Mission Analysis of F-5E Aircraft Using CFJ Wing
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 314/637
CONTROL METHOD USING CO-FLOW JET 309
Table
1
Work potential and fuel consumption of
F-5E
intercept mission
based on Roth’s model”
Component Work potential, hp.min Fuel consumption,
Propulsion
Compressor
loss
585
Compressor PR
585
mass flow rate
Total engine
loss
Total engine thrust
Total propulsion
Wave
+
skin friction
Fuselage drag
Tail drag
Wing drag
Induced drag
Structure weight
Propulsion weight
Fixed equip. weight
Stores weight
Fuel
+
misc. weight
Store drag loss
Total airframe
loss
585
Vcompressor
Airframe
27,108
7.8
89%
53.6
kg/m
113,762
191,529
305,291
55,151
19,268
32,504
26,052
9,116
9,572
3,466
26,061
5,446
186,636
8.88
37.3
62.70
100
18.06
6.32
10.65
8.53
2.99
1.14
8.54
1.78
61.11
3.10
linear to the baseline results. Table 3 gives the results
of
the estimated mission
analysis.
The endurance and range are calculated based on the following formu-
l a t i o n ~ * ~ :
Table 2 CFJ wing parameters for F-5E-CFJ
Parameters Value
V C t j
60%
CFJ PR
4
Wing span
1 2
baseline
Wing weight
1 2
baseline
CFJ
CL
wing
2
baseline
CFJ
Cd
wing skin
+
wave
1 8
baseline
CP
0.1
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 315/637
31
G.-C.HA AND C. D. PAXTON
Table 3 Estimation of fuel consumption
of
F-5E intercept mission using
recirculating CFJ wing
Work potential, Fuel consum ption, CFJ benefit,
Component hp.min
Compressor loss 32,003 10.48 - .6
Turbine
loss
17,714 5.8 - .2
Total engine
loss
119,386 39.11 - .81
Total engine thrust 185,905
60.89 1.81
Total propulsion 305,291
100 0
Airframe
Wave + skin
friction
Fuselage drag 55,151 18.06
0
Tail drag
19,268 6.32
0
Structure weight 22,929 7.51 1.02
Propulsion weight 9,116 2.99
0
Fixed equip. weight 9,572 3.10
0
Stores weight 3,466 1.14
0
Fuel
+
misc. 26,061 8.54 0
Store drag loss 5,446 1.78 0
Wing drag 2,03 1.5 0.66 9.98
Induced drag
weight
Total airframe
loss
153,040 50.13 9.17
Endurance 41.3
Range 37.7
and
where
WO
and
W1
are the takeoff gross weight and the weight with empty fuel
tank, and ct is the specific fuel consum ption.
This conservative estimation suggests a benefit of 9.17% fuel consumption
reduction and also 18% total drag reduction for the whole aircraft. Assuming
that the F-5E-CFJ carries the same amount of fuel as the F-5E, the weight
ratio of
Wo/W1
is increased by 1.8%. Because of the fuel consumption reduction,
drag reduction, and the increase in WO/WI, the endurance and range increase by
41 and 37.7%, respectively. The power required to pum p the
CFJ
increases the-
compressor loss by 18%, which generates a penalty to the total fuel consumption
of 1.8%. The largest gain for the fuel consum ption is from the dra g reduction of
the wing, at 9.98%. The gain from the wing weight reduction is 1.02%. The
penalty to the propulsion system is easily offset by the benefits from the reduction
of wing drag and structure weight.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 316/637
CONTROL METHOD USING CO-FLOW JET 311
The following formulations are used to calculate the stall velocity, and takeoff
(TO) and landing distancesz9:
Using the maximum performance wind-tunnel test results of the CFJOO25-
065-196 airfoil with
C
= 0.29,25928he
Vstall
will be decreased by 44%, the
takeoff distance by 68 , and the landing distance will also be reduced by 68
if it is assumed that the resultant force
F 0 7 ~ ~
s the same.
111
Conclusions
A novel airfoil flow control technique using a co-flow jet to achieve superior
aerodynamic performance for subsonic aircraft has been studied numerically by
CFD simulation. The CFJ airfoil opens a slot on the airfoil suction surface near
the LE and TE. A high-energy jet is injected tangentially near the LE and the
same amount of mass flow is sucked in near the TE. The jet can be recirculated
to reduce the energy expenditure of the overall airframe-propulsion system by
avoiding dumping of the jet mass flow, or achieving zero net jet mass flow.
The turbulent shear layer between the main flow and the je t causes strong turbu-
lence diffusion and mixing under a severe adverse pressure gradient, which
enhances lateral transport of energy and allows the main flow to overcome the
severe adverse pressure gradient and stay attached at a high angle of attack.
The CFJ airfoil achieves significantly higher lift because of augmented circula-
tion. The airfoil does not rely on the Coanda effect at the LE or TE. Hence,
the technique can be applied to a modern high-speed thin airfoil, and can be com-
bined with other flow control techniques.
The C FD simulation indicates that the CFJ airfoil performance is more sensi-
tive to injection location than to suction location. The injection location should be
as close to the L E as possible, but must be dow nstream of the L E suction peak to
make use of the adverse pressure gradient as enhance jet mixing with the main
flow.
For the NAC A2415 airfoil studied, at low AOA with moderate m omentum jet
coefficient, the CFJ airfoil will not only significantly enhance the lift, but w ill also
dramatically reduce the drag, or even generate negative drag (thrust). The mech-
anism for this is that the co-flow je t reduces the pressure drag by creating very
strong LE suction, and can generate negative pressure drag greater than the fric-
tion drag. This may allow the wing to generate both lift and thrust, like a flapping
wing, and cruise with very high aerodynamic efficiency. At high AOA, both the
lift and the drag are significantly higher than the a irfoil with no flow control, and
may enhance the performance of takeoff or landing within short distances.
Based on the wind-tunnel test results and a conservative estimate, for a subso-
nic area intercept mission analysis of the military aircraft F-5E, assuming use if a
CFJ airfoil, the fuel consumption is reduced by 9% , and the endurance and range
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 317/637
312
G.-C.HA AND C. D. PAXTON
by 38 and 41%. Based on the maximum performance wind-tunnel test data, Vstall
is reduced by 44%, and the takeoff and landing distances are reduced by
68 .
The engine consumes an extra 1.8% fuel, but the whole system, comprising air-
frame and propulsion, sees a benefit.
The CFJ airfoil concept suggested in this paper appears to have the following
advantages: 1) very effective in enhancing lift and suppressing separation; 2) dra-
matically reduces drag and can achieve very high C L / C D t low AOA (cruise),
and very high lift and drag at high AOA (takeoff and landing); 3) significantly
increases AOA operating range and stall margin; 4) has small penalty regarding
the propulsion system; 5 ) can be applied to any airfoil, thick or thin; 6) can be
used for the entire flying mission, rather than only during takeoff and landing;
7 can be used for low- and high-speed aircraft; and 8) is easy to implement,
with no moving parts.
The aforementioned advantages of the CFJ airfoil may derive the following
superior aircraft features: 1) requirement for extremely short distances for
takeoff and landing; 2) supersonic aircraft to have small wing size matching
cruise need, but also have high subsonic performance (e.g., high lift, low drag
at M
<
1); 3) high maneuverability, high safety, and fast acceleration military
aircraft; 4) very economic fuel consumption; 5 ) small wing span for easy
storage, light weight and reduced skin friction and form drag; 6) low noise
because of no high-lift flap system and reduced wing drag or even wing thrust,
which will require less engine nozzle jet velocity;
7
the possibility of heavy
lift rotorcraft, essentially with no dynamic stall; and
8
stealth aircraft with no
moving control surface.
Acknowledgments
The authors would like to acknowledge NASA Langley Research Center
(LaRC) for supporting the wind-tunnel tests under contract NNL04AA39C of
NRA-03-LaRC-02.25’28We would also like to thank Geoffrey A. Hill at NASA
LaRC for the discussion of possible applications of the CFJ airfoil to supersonic
aircraft.
References
Air Vehicles,” A IAA Paper 2004-3785, June 2003.
‘Sellers, W. L. I., Singer,
B .
A., and Leavitt, L. D., “Aerodynamics for Revolutionary
’Gad-el Hak, M., “Flow Control: The Future,”
Journal of Aircraft,
Vol. 38, 2001,
3Gad-el Hak, M.,
Flow
Control, Passive, Active, and Reactive
Flow
Management,
Cambridge Univ. Press, Cambridge, UK, 2000.
4Anders, S., Sellers, W. L., and Washburn, A., “Active Flow Control Activities at
NASA Langley,” AIAA Paper 2004-2623, June 2004.
5Tilmann, C. P., Kimmel, R. L., Addington, G., and Myatt, J. H., “Flow Control
Research and Application at the AFRL’s Air Vehicles Directorate,” AIAA Paper 2004-
2622, June 2004.
6Miller, D., and Addington, G., “Aerodynamic Flowfield Control Technologies for
Highly Integrated Airframe Propulsion Flowpaths,” AIAA Paper 2004-2625, June 2004.
pp. 402-418.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 318/637
CONTROL METHOD USING CO-FLOW JET
31 3
’Kibens, V., and Bower, W. W., “An Overview of Active Flow Control Applications at
The Boeing Company,” AIAA Paper 2004-2624, June 2004.
‘Modi, V., Fernando, M., and Yokomizo, T., “Drag Reduction of Bluff Bodies Through
Moving Surface Boundary Layer Control,” AIAA Paper 90-0298, 1990.
’Cichy, D., Harris, J., and MacKay, J., “Flight Tests of a Rotating Cylinder Flap on a
North American Rockwell YOV-1OA Aircraft,” NASA Paper CR-2135, 1972.
Englar, R. J., “Circulation Control Pneumatic Aerodynamics: Blown Force and
Moment Augmentation and Modifications; Past, Present and Future,” AIAA Paper
2000-2541, June 2000.
“Bradley, L. C., “An Experimental Investigation of a Sting-Mounted Finite Circulation
Control Wing,” M.S. Thesis, Air Force Inst. of Technology, W right-Patterson Air Force
Base, OH, 1995.
”Wood, N., Robert, L., and Celik,
Z.,
“Control of Asymmetric Vortical Flows over
Delta Wings at High Angle of A ttack,”
Journal
of
Aircraft,
Vol. 27, 1990, pp. 429-435.
13Wood, N., and Robert, L ., “Control of Vortical Lift on Delta Wings by Tangential
Leading-Edge B lowing,”
Journal
of
Aircra f ,
Vol. 25, 1988, pp. 236-243.
14Wood, N., and Nielsen, J., “Circulation Control Airfoils-Past, Present, Future,”
AIAA Paper 85-0204, 1985.
”Englar, R. J., Trobaugh,
L.
A., and Hemmersly, R., “STOL Potential of the Circulation
Control Wing for High-Performance Aircraft,”
Journal
of
Aircraft,
Vol. 14, 1978,
Englar, R. J., “Circulation Control for High Lift and Drag Generation on STOL
17Smith,A., “High-Lift Aerodynamics,”
Journal
of
Aircraf,
Vol. 12, 1975, pp. 501 -530.
“Lin, J., Robinson,
S. ,
McGhee, R., and Valarezo, W., “Separation Control on High
Reynolds Number M ulti-Element Airfoils,” AIAA Paper 92-2636, 1992.
”Wygnanski, I., “The Variables Affecting the Control Separation by Periodic Exci-
tation,’’ AIAA Paper 2004-2625, June 2004.
”McM anus, K., and Magill, J., “Airfoil Performance Enhancement Using Pulsed Jet
Separation Control,” AIAA Paper 97-1971, 1997.
”McManus, K., and Magill, J., “Separation Control in Incompressible and Compressible
Flows Using Pulsed Jets,” AIAA Paper 96-1948, 1996.
”Johari, H., and McM anus, K., “Visulation of Pulsed Vortex Generator Jets for Active
Control of Boundary Layer Separation,” AIAA Paper 97-2021, 1997.
23Whitfield,C., “Airframe System Noise Reduction,”
Proceedings of 2nd NASA Vehicle
System Program Annual Meeting,
July 2005.
24Liu,Y., Sankar,
L.
N., Englar, R. J., Ahuja, K. K., and Gaeta, R ., “Computational Evalu-
ation of the Steady and Pulsed Jet Effects on the Performance of a Circulation Contro l Wing
Section,” AIAA 42nd Sciences Meeting and Exhibit, AIAA Paper 2004-0056, Jan. 2004.
25Zha, G.-C., Carroll, B., Paxton, C., Conley, A., and W ells, A., “High Performance
Airfoil with Co-Flow Jet Flow Control,” AIAA Paper 2005-1260, Jan. 2005; also
AIAA
Journal
(submitted for publication).
26Zha, G.-C., and Paxton, C., “A Novel A irfoil Circulation Augm ent Flow Control
Method Using Co-Flow Jet,” NASA/ONR 2004 Circulation Control Workshop, March
2004; also AIAA Paper 2004-2208, June 2004; also AIAA Book Series,
Progress in Astro-
nautics and Aeronautics.
27Greitzer, E. M ., Tan, C.
S. ,
and Graf, M. B.,
Internal Flow
Cambridge Univ. Press,
Cambridge, UK, 2004.
10
pp. 175-181.
16
Aircraft,”
Journal
of
Aircra f ,
Vol. 12, 1975, pp. 457-463.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 319/637
31
G.-C.HA AND C. D. PAXTON
”Zha, G.-C., Paxton, C., Conley, A., Wells, A., and Carroll, B., “Effect of Injection
Slot Size
on
High Performance Co-Flow Jet Airfoil,”
Journal
ofAircrujl (to appear 2006).
29Anderson, J. D., Introduction to Flight 4th ed. McGraw-Hill Higher Education, New
York, 2000.
30Zha, G.-C., (team m embers Car, D. and Copenhaver, W .), “Super Diffusion Cascades
Using Co-Flow Jet Flow Control,” National R esearch Council Summ er Faculty Final Rept.
Aug. 2002.
31DeLaurier,J., “Work
on
Flapping-Wing Flight,” Lecture,
23rd
AIAA Applied Aerody-
namics Conference, June 2005.
32Roth,B. A., “A Theoretical Treatm ent of Technical Risk in M odem Propulsion System
Design,” Ph.D. Thesis, Department of Aerospace Engineering, Georgia Inst. of Tech.,
Atlanta, GA , March 2000.
33Roth,
B.
A., “Aerodynamic Drag Loss Chargeability and Its Implications in the
Vehicle Design Process,” AIAA Paper 2001-5236, 2001.
34Roth,
B.
A., and Mavris, D., “A Method for Propulsion Technology Impact Evaluation
Via Thermodynamic Work Potential,” AIAA Paper 2000-4854, 2000.
35Roth,B. A., and Mavris, D., “A Generalized Model for Vehicle Thermodynamic Loss
Management and Technology Concept Evaluation,” 2000 World Aviation Conference,
Paper 2000-01-5562, Oct. 2000.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 320/637
Chapter 11
Experimental Development and Evaluation of
Pneumatic Powered-Lift Super-STOL Aircraft
Robert
J.
Englar*
Georgia Institute of Technology, Atlanta, Georgia
and
Bryan A . Campbellt
NASA Langley Research Center, Hampton, Virginia
Nomenclature
A j = blowing slot area
b = wing span, ft
c = chord length, ft
cg = cen ter of gravity, ft
CD= three-dimensional drag coefficient
CL= three-dimensional lift coefficient
CDE= equivalent drag coefficient
C = maximum lift coefficient
CM25,CM quarter chord pitching moment coefficient
CT = thrust coefficient
C, =jet momentum coefficient [see Eq. 2 ) ]
C = et m om entum coefficient, leading-edge blowing
CpChw
=
et momentum coefficient, Channel-Wing blowing
Elev = elevator deflection angle
hslot, j = blow ing je t slot height, in.
iT
= tail incidence angle, deg
m =jet mass flux, slugs/s
q
= freestream dynamic pressure
(=
V2) , psf
Pd
PD ,P , =
duct total pressure
*Principal Research Engineer, Aerospace Transportation Lab., Georgia Tech Research Institute.
'Senior Aerospace Engineer, Configuration Aerodynamics Branch.
Copyright 005 by the authors. Published by the American Institute of Aeronautics and
Associate Fellow AIAA.
Astronautics, Inc., with permission.
315
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 321/637
31
6 R. J. ENGLAR AND B. A. CAMPBELL
s
= wing area, ft2
S
= ground roll, ft
Td
= duct total temperature, OR
TR
=
resultant force, lb
T /
W = thrust/weight ratio
V = freestream velocity, ft/s
v d
=
deflected slipstream velocity, ft/s
vj
=
blowing je t velocity, isentropic, ft/ s
W / S
=
wing loading, psf
x/c = nondimensional chordwise location
x
= moment center location, ft
xTO = takeoff distance, ft
a
=
angle of attack, deg
p = freestream density, slugs/ft3
p = blowing je t density, slugs/ft3
&lipstream = slipstream deflection angle, deg
ACL= incremental lift coefficient
I. Introduction
HE ABILITY to achieve Super-STOL (short takeoff and landing) or V/
T
TOL (vertical/short takeoff and landing) capability with fixed-wing aircraft
has been an attractive goal in the aerospace community for over 50 years.
The impetus toward its achievement has historically been the numerous benefits
associated with very short to zero field length operations of nonrotary-wing air-
craft. Although such capability has direct application for military missions such
as those of a tilt-rotor o r tilt-wing aircraft, there also exists an additional need for
simple/reliable/effective personal and business-sized Super-STOL or VSTOL
aircraft operating from remote or small sites, as well as increasingly dense
urban environments. The development of simple, efficient aeropropulsive tech-
nology and corresponding low-speed control systems to make this possible is a
goal that now seems achievable because of technical breakthroughs in pneumatic
and powered-lift aerodynamic technologies. This chapter, originally presented at
the NASA/ONR CC Workshop in March 2004 (see NASA CP 2005-213509,
2005), will discuss recent progress in the integration of high-lift, propulsive,
and control systems, all employing common pneumatic techniques using circula-
tion control (CC) blowing, into a promising Super-STOL configuration.
Two promising technologies to evolve from earlier STOL/VSTOL research
are the Custer Channel Wing powered-lift configuration and the circulation
control wing (CCW) pneumatic high-lift concept. Through innovative use of
the propeller slipstream, the Channel Wing airplane developed by Willard
Custer (Fig. l)i-3 was able to achieve significant lift coefficient and efficient
downward thrust deflection without varying the high-lift configuration geometry.
This powered-lift technology, tunnel-tested by NACA in 1953,l and then flight-
tested and further developed by Custer in the mid-1960~,~mployed the
Channel Wing concept shown in Fig. 2.3 In essence, the propeller located at
the very trailing edge (TE) of the 180-deg arc circular channel in the wing
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 322/637
PNEUMATIC POWERED-LIFT SUPER-STOL AIRCRAFT
31
7
Three-view
of the Cuetor Chann el Wing CCW-5. Author’r Collection)
Fig. 1 Three-view and in-flight photo of 1960s Custer Channel Wing Aircraft.’-3
further increased the velocity over the channel’s upper surface and augmented the
circulation and lift there in much the sa me manner a s a deflected flap, but perhaps
to a greater extent. Lift was also augmented by the deflected slipstream behind
the channel such that
In-flight lift coefficients nearing
5
were generated by thrust coefficients also
nearing
5
a s dem onstrated by C ~ s t e r . ~owever, the flight-tested Custer
Channel Wing aircraft demonstrated a number of drawbacks associated with
low-speed handling, cruise drag, stability and control, high-incidence operation,
and one-engine-out scenarios, including the following:
1) Much of the high CLwas from redirected thrust, and less from circulation
lift augmentation.
2 )
High cruise dra g could result from the channel’s extra surface area.
3) Asymmetric thrust yields asymm etric mom ents and instability.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 323/637
318
R. J. ENGLAR AND B. A. CAMPBELL
Airfoil Surface in Channel; Replace with
New Pneumatic AirfoilsiAA Turning Surfaces
Fig. 2 Basis of the Channel Wing concept and current pneumatic improvem ents.
4)
Channel LE and TE separation could occur at high angle of attack
a.
5 ) Poor low-speed control is available from conventional aerodynamic sur-
6 There is nose-down pitch from aft propeller loading on the w ing.
7) There is nonuniform flow around the prop at high a.
8)
There is poor lift/drag ratio.
9) High angle-of-attack operation could cause poor visibility and control.
10) There are one-engine-out control problems.
To alleviate these shortcomings, preliminary research has been carried out at
Georgia Tech Research Institute (GTRI), where investigation in adapting CC
pneumatic technology has been made (Fig.
3
and Refs. 4 and 5 for example)
to dramatically improve the Channel Wing configuration. As Fig. 2 shows, the
new pneumatic configuration thus developed combines blowing on curved sur-
faces at the channel TE to greatly augment the lift and thrust deflection
without using high angle of attack. It also employs blown CCW technology on
the outboard wing panels to further augment lift and low-speed controllability
while providing additional drag when needed for slow-speed approaches down
steep glide slopes for Super-STOL.
This channel thrust turning and lift augmentation are based on the CCW/
upper surface blowing USB) oncept of Fig. 4, where tangential blowing on a
highly curved TE behind a je t eng ine augments flowfield entrainment, increases
circulation, and deflects thrust to add more incremental lift. Thrust deflection
angles of 165 deg produced by blowing were measured experimentally on
wind-tunnel This concept provides pneumatic STO L, VSTOL, and
thrust-reversing capabilities without any moving parts. Circulation Control
Wing alone (Fig. 3) employs a similar tangential-blowing configuration, but
without the pneumatic thrust deflection. Such CCW airfoils have generated
faces at low speeds.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 324/637
PNEUMATIC POWERED-LIFT SUPER-STOL AIRCRAFT 319
TANGENTIAL BLOWlNG OVER ROUNDED TRAILING
EDGE
SURFACE
Fig. 3 Basics
of
circulation control pneumatic technology.
measured two-dimensional lift augmentations of
80
times the input blowing
When flight-tested on an A-6 flight demonstrator, CCW showed
a
140
increase in useable high lift, employing only half of the bleed air avail-
able from the aircraft’s standard turbojet engines.* Figure
2
shows how these
blown flow-entrainment devices would be arranged to enhance the effectiveness
of the Pneumatic Channel Wing (PCW) configuration. In addition, the CCW lift
capability can be applied differentially outboard to generate very large rolling
and yawing moments, which are essential for controlled flight at the very low
speeds of Super-STOL.
Based on earlier CCW/USB wind-tunnel and full-scale data (Fig.
4 6,7
and
CCW flight-test data from the A-6 STOL-demonstrator program,8 the predicted
lift and drag capabilities for the pneumatic channel wing configuration were
expected to offer great Super-STOL promise. Reference 9 details these early
mu0
vwt
de?wilm sl
OCHllis
lnd
t u m d m o d d
Fig. 4 Previously developed CCW/upper surface blowing powered-lift concept:
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 325/637
320
R. J. ENGLAR AND B. A. CAMPBELL
predictions before the current wind-tunnel test data were available. These implied
CL alues approaching 9-
10
for a pneumatic channel wing aircraft with blowing
on outboard CCW wing panels at relatively low aircraft angle of attack. Higher
CL
alues were possible at higher thrust coefficients if higher
a
values were used
because of the additional vectored thrust component. Again, for comparison, the
Custer channel wing aircraft generated in-flight CL f 4.9, whereas a conventional
slotted flap on this wing geometry would generate CL values at 2-3. Initial
takeoff predictions’ showed that these PCW capabilities could produce very
short, hot-day takeoff ground rolls for typical mission weights, and even zero
ground roll under certain conditions.
As part of an ongoing program for the NASA Langley Research Center to
develop this PCW concept, GTRI and NASA have teamed together in an experi-
mental development program being conducted at GTRI, which has provided
aerodynamic and propulsive data input for design studies being conducted at
both NASA and GTRI. The current paper will summarize these experimental
results and discuss effects deriving from variations in PCW geometry, propeller
thrust, and channel blowing.
11.
Experimental Apparatus and Test Techniques
A wind-tunnel development/evaluation program was conducted at GTRI on
a generic twin-engine Super-STOL-type transport configuration (Fig. 5) using
the 0.075-scale semispan model shown in Fig.
6.
A variable-speed electric
motor was installed in the nacelle, which could be located at various positions
in the channel, and which drove interchangeable two-, three-, or four-bladed
propellers of various diameters and pitch. Also variable was the height of
the blowing slot located at 95 of the channel chord length, as well as the
blowing momentum coefficient and portions of the slot arc length that were
blown. Behind the slot, the rounded TE curved only 90deg (rather than the
more conventional
180
deg of typical CCW configurations) for an anticipated
maximum thrust deflection of approximately (90 de g
+ a).
It was already
known (Fig. 4) that thrust deflections up to 165 deg yielded by blow ing were a
Fig. 5 Conceptual PCW Super-STOL transport configuration.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 326/637
PNEUMATIC POWERED-LIFT SUPER-STOL AIRCRAFT 321
Fig. 6 PCW/CCW semispan model installation in GTRI Model Test Facility
research tunnel (three-bladed prop with unblown outboard CCW), plus jet flow
turning in channel (black tufts).
possibility. Here, the momentum coefficient is defined as
This semispan model configuration (Fig. 6 ) was mounted on an underfloor
balance with air supplies and automated pitch table in the GTRI Model Test
Facility 30 x 43 x
90
in. test section. The tunnel wall boundary layer near the
test section floor was eliminated by use of tangential floor blowing. In a
follow-on version of this configuration, both the LE and the TE of the outboard
CC W wing section were also blown for separation con trol. The em phasis in the
following data is on the performance of the inboard blown PCW configuration,
but performance of the outboard CCW sections to further augment lift is also
shown.
111. Wind-Tunnel Evaluations and Results
Test techniques employed in the subsonic tunnel evaluation of this pneumatic
powered-lift model are similar to those employed and described in Refs. 10 and
11 for blown airfoil and semispan models, except that special additional tech-
niques were employed to account for the installation of the active propeller in
the channel (see below). Som e 980wind-tunnel runs (including propeller calibra-
tions) have now been conducted during three test programs at GTRI to develop
these blown-configuration geometries and to evaluate their aeropropulsive,
flight-trim, and control characteristics. A typical run consisted of a sweep (incre-
mental variation) of prop thrust or blowing pressure at constant angle of attack
and wind speed. Also, angle-of-attack sweeps or dynamic pressure (velocity)
sweeps were run at constant thrust and blowing coefficients CT andC,. Numerous
runs were made with varying tail configurations to evaluate pitch trim and
control. Typical test results are presented in Secs. 1II.A-1II.C to demonstrate
how these various parameters affected overall performance.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 327/637
322
R. J. ENGLAR AND B. A. CAMPBELL
A. Tunnel Test Results, Outboard Wing
ON
In Figs. 7a and 7 b are shown the effects on lift and drag coefficients of blowing
the channel TE without the prop installed (i.e.,
CT
= 0), but with the engine
nacelle in place (Fig. 6 ) . Note the ability of the blowing to more than double
the CLmax f the unblown configuration with virtually no reduction in the stall
angle, astall he C, values shown are comparable to or greater than those that
would normally be generated by more complex moving mechanical flaps. Note
also the ability of the blowing at a = 0 deg to increase
CL
by a factor of
nearly 10 over the unblown value. At a
=
0 deg, blowing at C,
= 0.30
yields
50 more
C,
than the
C
of the unblown configuration. In Fig. 7b, the drag
polars at constant C, are typically quadratic in
CL.
Earlier in a than where the
stall begins, they follow essentially the same single curve, using blowing to pro-
gress to each successively higher
C,
region.
Addition of the propeller to the channel brings into play the powered-lift
characteristics of the PCW configuration. Figure
8,
for
a
= 0 deg, shows the
variations in
C,
and
C
with thrust coefficient
CT
for fixed values of blowing
coefficient. Here, in order to recognize the direct thrust component to lift and
drag, thrust coefficient is defined as CT
= T / ( q S ) ,
where T is the calibrated unin-
stalled wind-on prop-alone (not-in-the-channel) thrust at the proper advance ratio
that is, representative test dynamic pressure
q.
The reference area S is the wing
semiplanform area. These thrust values were determined prior to installation in
the channel by testing the prop alone in the tunnel at various rpm and tunnel
speeds.
Calibration curves of T (thrust) against rpm were input to the data reduction
program at given test wind speeds. CT,C,, and C are directly comparable on
Fig. 7 Measured blown lift and drag capabilities of the PCW model without the
propeller installed: a) Lift vs
a
) Lift-drag polars.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 328/637
PNEUMATIC POWERED-LIFT SUPER-STOL AIRCRAFT 323
Channel Wing
CT
Fig. 8 Effects of prop thrust variation on lift and drag at constant blowing C,) and
Y
= 0 deg.
a comm on reference basis to determ ine force contributions from installed thrust.
This avoids the difficulty that would be caused by using the standard helicopter
thrust coefficient, based on rotor (or prop) geometry rather than wing area. Also,
note that measured
CD
ncludes the input thrust, which cannot reasonably be sep-
arated from the aerodynamic drag alone once the prop is in the channel. Measured
CD an therefore be (and sometimes is) negative. After the initial low values of
CT
are exceeded, C L ncreases nearly linearly with
CT,
and
CD
educes nearly lin-
early. (This implies that, at a constant C,, the thrust deflection angle is nearly
constant.)
Figure 9 shows that incremental lift augmentation as a result of blowing
(C,)
is much greater than that resulting from CT (Fig. 8).Here, at CT = 2.2, the blown
configuration generates CL of approximately 8.5 at a
=
10deg. Th e flight-tested
Custer Channel Wing3 generated roughly one-third this
CL
at this
CT,
but also
required a =
24-25
deg. Note also that increased blowing at a constant CT
yields increased drag (rather than thrust recovery), which can be quite essential
for Super-STOL approaches and short landings. These lift comparisons in
Figs.
8
and 9 show that lift increases more efficiently by increasing blowing
than by increasing thrust. In Fig. 10 a plot is shown of the variation in lift and
drag with angle of attack for the blown powered-lift configuration in comparison
with the unblown baseline configuration without the prop. Here, flow visualiza-
tion showed that the initial stall
a
=
15-17
deg) seen for most of the lift curves
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 329/637
324
R. J. ENGLAR AND B. A. CAMPBELL
Channel Wing C k
Fig.
9
Effects of blowing variation on lift and d rag at constant
CT
nd Y
= 10
deg.
corresponded to stall of the outboard unblown wing section, whereas the blown
channel wing section then continued on to stall angles of 40-45 deg and C,
values of 8.5-9. Note that CD including thrust) increases from negative to posi-
tive values as incidence increases.
Figure 11 shows the effect on lift and drag of increasing the circular arc length
of the blown slot around the channel at a given prop longitudinal location
x / c
=
0.95), where the ma ximum slot arc of 160 deg was most effective.
Blowing of m ore than 160 deg of channe l arc was not appropriate on this model
because the last 20 deg of inboard arc was along the chan nel right next to the
fuselage, and b lowin g there w ould do little more than bou nce off the fuselage.
Th e effect on increased tail-off pitching m oment caused by suction loading on
the aft of the channel (either by blowing, prop slipstream, or both) is shown in
Fig. 12 as a function of
CT
and
C,,
all at
a
= 0
deg. These moments a re referred
to the ch annel’s quarter-chord location c/4), and confirm the typical trend of this
type of blown configuration: large nosedown CM,which, although does make the
aircraft much more stable longitudinally, causes problem s with pitch trim. It is for
this reason that additional experimental evaluations were conducted tail-on to
investigate increased longitudinal trim capabilities. All data presented
so
far
have been tail-off. A second investigation was conducted with LE blowing
installed on the outboard wing CCW portion to provide counteracting nose-up
pitch for trim, a s well as for L E separation prevention.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 330/637
PNEUMATIC POWERED-LIFT SUPER-STOL AIRCRAFT 325
Angle of Attack, a dcg Angle of Allack, a deg
Fig. 10 Effects of blowing, Ch and a on lift coefficient, stall angle, and drag
coefficient for the P CW model with unblown ou tbo ard wing.
Fig.
11
Effects on lift and drag of varying blown channel slot arc length at
constant C, and a t
a = 0
deg: a) Prop and nacelle installed and b) Prop off, but
nacelle installed.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 331/637
326
R. J.
ENGLAR AND B. A. CAMPBELL
Channel Wing C k or CT
Fig. 12 Effects of prop/nacelle location, blowing and thrust on quarter-chord
pitching m oment, Y =
0
deg.
B.
Tunnel Test Results, Channel Wing Only
Higher nondimensional thrust coefficient values were available when the
channel-only configuration was tested (fuselage, blown channel and prop, but
with no outboard
CCW
panels), because the reference planform area of the
wing was also reduced. This allowed
CT
of x 3 for the channel-only vehicle,
and, as Fig. 13 shows, lift coefficients nearing 11 were measured with a conven-
tional horizontal tail installed at the midvertical location on the aft fuselage.
Needless to say, not all of the lift values show n in Fig. 13 are trimmed long itud-
inally. Furthermore, for the
CT=
case with blowing on, the conventional tail of
the aircraft stalled experimentally over much of the lower
a
ange.
The possible inability to trim these Super STOL aircraft longitudinally has
been highlighted as a problem of blown systems in Refs. 7 and 8. It is further
emphasized in Fig. 13, where the large suction on the aft-loaded blown
channel (and blown wing, if present) produces very large nosedown pitching
moments (compared to the
CT
= 0, C = 0, tail-off curve). Although this can
produce improved longitudinal stability, these moments must also be trimmed.
Horizontal tail investigations were conducted as part of this three-dimensional
model development plan in the hope of determining tail location and configur-
ation to provide enough nose-up pitch to trim the vehicle. Several horizontal
tail configurations [one without an elevator, a second with a 20-deg up elevator
aelev
= +2 0 deg), and a third with an inverted leading edge droop] were
designed and fabricated. As Fig. 14 shows, these could be mounted on a vertical
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 332/637
PNEUMATIC POWERED-LIFT SUPER-STOL AIRCRAFT 327
I I
R e u m a ~ c hannel WmgONLK, Outboard
CCW
OFF,ROD N,Wd Tad
a q r c n
Fig. 13 Effect
of
thrust and/or blowing increase on lift and pitching moment
variation with
a
or channel-wing-only configuration (no outboard wing panels)
with tail at midlocation, i = 0 deg.
center plate yielding variation in both tail incidence iT) and vertical position in
the propeller slipstream. High, midfuselage, and low-tail positions were tested.
Testing of these tail-on configurations over a range of tail parameters revealed
that a low-tail position immersed in the prop slipstream and dynamic pressure
was more effective than the higher tail (Fig.
15),
but the lower tail also experi-
enced more LE stall for the same reason. This tail stall prevents the vehicle
from being trimmed at this higher blowing condition (here with the outboard
CCW wing on again). Considerable videotaping of flow visualization tufts on
the tail revealed these problem areas and led to the development of the
inverted-droop (drooped upward) LE modification for the tail. Keeping the tail
LE attached allows positive nose-up pitch and thus trim to be generated for the
vehicle over a much wider range of lower
a
values. For the channel-wing-only
model with the modified tail, trimmed CL values greater than 9 are therefore
seen (Fig.
16),
but much of these data are still untrimmed, and again the low
tail with no LE modifications is fully stalled. Thus, these data imply that
further tail development (perhaps including LE blowing to prevent the tail stall
without mechanical LE fixes) is needed to trim in this high
CL
range at all
vehicle angles of attack.
C.
Tunnel Test Results: Flow Attachment
An additional series of flow visualizations was conducted to further identify
means to prevent separated flowfields on the wing during high-lift generation.
Figure
17
data show that the flow at the channel LE is entrained to the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 333/637
328
a
R. J. ENGLAR AND B. A. CAMPBELL
Fig. 14 Horizontal tail configurations evaluated, with outboard CCW wing panel
on: a) high tail, b) low tail, and c ) mid-fuselage tail.
point where LE separation is prevented until a = 35-40 deg or more, but that the
outboard CCW is prone to stall there. Leading-edge blowing on this outboard
CCW wing panel greatly entrained this flowfield as well. Figure 18 flow visual-
ization shows this severe separation at a
=
20 deg for the unblown case (Fig.
18a), whereas blowing the LE completely reattached the flowfield there.
An additional means of trim and control was investigated for the PCW. Here,
these large nosedown pitching moments (seen in Fig. 13, 15, and 16) are offset by
moving the aircraft center of gravity aft to trim, with no tail installed. Aft center-
of-gravity movement was previously performed for flight tests of the A-6/CC
Wing aircraft, but with the tail on.8 Figure 19 shows data for the CT = 3 case
of a tail-less PCW without outboard wing. At
C,
=
0,
moving the center-
of-gravity aft from
x / c
=
0.25 to 0.375 gives the aircraft neutral longitudinal
stability but does produce trim over most of the angle of attack range. Similar
reduction in pitching moment can be produced by aft center-of-gravity shift as
blowing is increased (Fig. 19b), but this requires further aft center-of-gravity
to trim at lower
a,
and the C, vs. C curves are now unstable (dC,/dC, =
positive). Some small control surface (such as a blown canard to provide nose-
up pitch and positive lift to trim) could perhaps be incorporated with a
state-of-the-art control system and control laws to make this a feasible pitch-
trim device without lift loss due to tail download.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 334/637
PNEUMATIC POWERED-LIFT SUPER-STOL AIRCRAFT 329
MTF063,
i4
Pitching Moment, Prop ON,
N o Outboard
Blo-g,Tsil Power
CT-2.2,CmuChW-1.0
,
CCW
Rap4 Alpha sweeps
Qusncr-Chord Pitching Moment, C
~ e / 4
Fig. 15 Comparison of high and low tail position on PCW configuration with
unblown outboard wing and horizontal tail.
Tad
Effects.
PneYmshF
Channel WxngONLY,
Outboard
CCW
OFF, Rop
ON Tad
Effects. Pneumahc
Channel WmgONLY.
Ou t h a rd
CCW OFF, Rap ON
Fig. 16 Comparative lift and quarter-chord pitching moment coefficients of PCW,
no outboard CCW, with and without tail
LE
modifications.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 335/637
330
R. J. ENGLAR AND B. A. CAMPBELL
Outboard
CCW Blowing EffectsCCW Flap=O', Tail OFF
Alpha,
deg
Fig.
17
Leading-edge blowing and channel flow entrainment prevent flow
separation over both channel and outboard CCW leading edges.
Fig. 18 Flow attachm ent caused by LE blowing on outboard CCW and channel flow
entrainment at Y
=
20 deg, channel LE not blown: a) Outboard LE slot unblown and
b) outboard LE slot blown.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 336/637
PNEUMATIC POWERED-LIFT SUPER-STOL AIRCRAFT
331
a)
9
8
7
6
d 5
4
3
2
1
0
10
9
8
7
6
5
4
3
2
1
MTF068, Pneumatic Channel
Wng,
Phase
111
CL v s CM, Run
799,
Cp Channel=O.O,
Channel Only, Prop
ON,
CT.3.0. Tai l OFF
1
0
CM
1 2
+xmom=.375
-+Xmom=.SO
I
_ _ _ _ .
I
__ .
I... _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ I
__ .
2 1 0 1
CM
Fig.
19
Effect of a ft center of gravity location on pitching moment curves for the
tail-less
PCW
at
CT
= 3 and two blowing values, x = x,.Jc: a)
CT = 3, C, =
0,
b) CT =
3,
C, =
0.3.
IV.
Comparison of Measurements and Predictions
In Fig. 20 are compared the results of these investigations with previously
predicted lift and drag data, which were estimated from existing CCW/USB
wind-tunnel data and from A-6/CC W flight-test data. W hereas the prop/electric
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 337/637
332
a)
l
9
8
7
cL
6
5
4
3
2
I
0
R. J.
ENGLAR AND B. A. CAMPBELL
. . .
, ,
. . .
,
Solid
Line
Data Derived from CCWNSB, a1
Tunnel Data from
GTRl
MTF055
a=lW
cl=fl.fl
0 1
0.2
0.3 0.4 0.5
0.6
0.7 0.8 0.9 1.0
CCChW
y
olid
Line Data Derived
,rom
CCWIUSB,, a = l W
I
9
Tunnel Data
from GTRI MTF055
0.0 0.1 0.2
0.3
0.4
0.5
0.6 0.7 0.8 0.9 1.0
%hW
Fig. 20 Comparisons of predicted and experimental PCW lift and drag data at
constant Ch
a
= 10 deg, outboard CCW on: a) Measured (symbols) vs predicted
lift (no symbols) and b) measured (symbols) vs predicted (no sym bols) drag.
motor currently available did not allow higher CTvalues than about 2.2 (outboard
wing on), these lower-thrust, wind-tunnel data considerably surpass the predicted
lift data (Fig. 20a).
If
the ratio
of
measured-to-predicted holds linearly up
to
CT=
10, then C, values over
14
are to be expected at a
=
10 deg. The
experimental drag data (Fig. 20b) are similar to the predicted values at lower
NOTE: -17,000
Ib
Increase in TO
Gross Wt over B aseline Tilt Rotor
a t 100 ground roll
Fig. 21
Pneumatic channel wing predicted super-STO L takeoff performance.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 338/637
PNEUMATIC POWERED-LIFT SUPER-STOL AIRCRAFT
333
C, but show less drag than predicted at higher blowing. These estimated data
have been used to predict Super-STOL takeoff distances on a hot day at
3000 ft altitude to be less than 100 ft and, in some instances, 0 ft. (Fig. 21).9
The composition of measured and predicted results in Fig. 20 seems to suggest
that even better takeoff performance might be obtained (higher lift, lower
drag). However, the lower measured drag values indicate that additional attention
will need to be paid to obtaining greater drag values for steeper glide slopes on
STOL approaches (when desired and chosen by the pilot).
V. Potential Applications
Design and mission stud ies conducted at NASA L aRC based on the preceding
tunnel data have led to consideration of several new pneumatic powered-lift
PCW-type configurations. The capability of the PCW to significantly augment
lift, drag, and stall angle to the levels reported herein dem onstrates that this tech-
nology has the potential to enable
simple/reliable/effective
STOL and possibly
VTOL operations of personal and business-sized aircraft operating from remote
or sm all sites as well as increasingly dense urban environm ents. Such capability
now opens the way for alternative visions regarding civilian travel scenarios, as
well as both civilian and military aerial missions. One such vision is represented
by the Personal Air Vehicle Exploration (PAVE) activity at NASA Langley
Research Center. Another vision, a military Super-STOL transport, is discussed
in the mission study of Ref. 9 and Fig. 21.
VI. Conclusions
Results from subsonic wind-tunnel investigations conducted at GTRI on a
0.075-scale powered semispan model of a conceptual PCW transport have con-
firmed the potential aerodynamic payoffs of this possible Super-STOL configur-
ation, including very high lift and overload capability. These results include the
following features. Lift and drag augmentations and/or reductions as desired for
Super-STOL operation have been confirmed, with C, = 9 measured at
a
=
10 deg C,= 10-11 at higher
a),
and drag coefficient (including thrust)
varying between - 2 and +2 , depending on blowing and thrust levels. C,
values nearing 14 are predicted if higher
CT
is available, say on takeoff.
Blowing C, and thrust CT variations were both found to enhance circulation,
thrust deflection significantly, and lift. However, if evaluated as incremental lift
per unit of input thrust or momentum
C,
or
C,),
blowing was far more efficient
than thrust. By varying only C, and/or C,, all the aircraft’s aerodynamic charac-
teristics (forces and moments) can be augmented or reduced as desired by the
Super-STOL aircraft’s pilot or its control system without mechanical moving
parts (such as tilting rotors or wings) and without resorting to high
a
to
acquire larger vertical thrust components for lift.
The blown channel wing itself, without thrust applied, was able to double the
C capability of the baseline aircraft configuration, and multiply its lift at
a = 0 deg by a factor of 10. Addition of blowing on the outboard CCW
section can increase this further, and can also add drag as needed for Super-
STO L approaches.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 339/637
334
R. J. ENGLAR AND B. A. CAMPBELL
Even with the unblown outboard wing stalling at a = 15-17 deg, the blown
and thrusting channel continued to increase lift up to a stall angle of 40-
45
deg as a result of channel flow entrainment. Although this high
a
may not
prove practical as a takeoff/landing operational incidence, it does show signifi-
cant improvement over the asymmetric LE separation of the conventional
channel wing’s stalled channel and the resulting low-speed control problems.
Pneumatic Channel Wing conversion of thrust into either drag decrease or
drag increase without moving parts is also quite promising for S TO L operation.
Large nosedown pitching mom ents are produced by these blown configurations,
and thus longitudinal trim capability needs to be addressed in future evaluations.
Unlike a tilt rotor, in Super-STOL or V/S TO L there is no download on the wing
from prop thrust because the PC W props d o not tilt. The potential of PCW for an
integrated lift, thrust/drag interchange, and control system, all from one se t of
devices, holds promise in terms of simplicity, weight reduction, and
reliability /maintainability.
The projected operational benefits based on these early data suggest Super-
STOL and possible V/STOL capability with significantly increased payload,
reduced noise signatures, and increased engine-out control, all without variable
geometry or mechanical engine/prop tilting. A PCW aircraft thus equipped
could provide a simpler, less costly way of achieving Super-STOL or V/STOL
capability without the complexity, weight, or reliability issues of rotating the pro-
pulsion system, carrying large engines and rotors on the wing tips, or thrusting
downwards on fixed wings during hover. Additionally, the integration of
pulsed-blowing technology with circulation control (currently being investi-
gated)12 may further increase lift efficiency and reduce already low b lowing
requirements by up to
50%
or more, while further enhancing stability and
control. Successful application of these results can lead to positive technology
transfer to personal, business, and military-sized aircraft. In addition to the mili-
tary Super-STOL transport discussed in Fig. 21, these experimental data and
pneumatic technology results have been included in preliminary design studies
of other possible pneumatic powered-lift configurations, including smaller per-
sonal and business-type aircraft.
Future testing, evaluation, and development still need to be accom plished to
address possible pitch-trim problems, performance at higher
CT
nd lower
C,,
and associated stability and control. In the future, the existing model or larger
three-dimensional models should be modified to include blown tail surfaces
and additional improvements to the pneumatic thrust deflection system. The fol-
lowing should be experimentally investigated:
1) Use of pulsed blowing to further reduce required blowing m ass flows (both
inboard on the channel and outboard on the CCW).
2) Higher propulsor solidity for greater thrust and powered lift, or improved
propeller characteristics for greater
CT
vailability.
3)
Further evaluation of low-speed controllability and trim, including evalu-
ation of improved tail surfaces, which might even be blown to reduce tail area
and drag.
4) Further evaluation of low-speed controllability and trim by novel aerody-
namic/pneumatic trim and control devices (blown canards, for example).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 340/637
PNEUMATIC POWERED-LIFT SUPER-STOL AIRCRAFT
335
The earlier mission analyses should be revised to incorporate the experimen-
tally developed aeropropulsive and stability and control characteristics of the
PCW concept. If the projected benefits are confirmed, and further benefits
come to light, then larger-scale, higher-Reynolds-number testing on a full
three-dimensional PCW m odel with variable yaw capability shou ld be conducted
to facilitate greater strides toward this pneumatic powered-lift technology’s
maturation.
Acknowledgments
The primary author would like to thank personnel of the NASA Langley
Research Center for their ongoing support of this powered-lift research a t GTRI.
References
‘Pasamanick, J., “Langley Full-Scale-Tunnel Tests of the Custer Channel Wing Air-
plane,’’ NACA RM L53A09, April 1953.
’Mitchell, K. A., “Mr. Custer and His Channel Wing Airplanes,” Journal o American
Aviation Historical Society, Spring 1998.
3Blick,
E.
F. and Homer, V., “Power-on Channel Wing Aerodynamics,”
Journal
o
Air-
craft,
Vol.
8,
No. 4, 1971, pp. 234-238.
4Englar, R. J., “Circulation Control Pneumatic Aerodynamics: Blown Force and
Mom ent Augm entation and Modification; Past, Present and Future,” AIAA Paper 2000-
2541, AIAA Fluids 2000 Meeting, Denver, CO, June 19-22, 2000.
’Englar, R. J., and Applegate, C. A., “Circulation Control-A Bibliography
of
DTNSRDC Research and Selected Outside References (Jan. 1969 through Dec. 1983),”
DTNSRDC-84/052, Sept. 1984.
6Englar, R. J., “Development of Circulation Control Technology for Powered-Lift
ST OL Aircraft,” NASA CP-2432, Proceedings of the 1986 Circulation Control Workshop.
’Englar, R. J., Nichols, J. H., Jr., Harris, M. J., Eppel, J. C., and Shovlin , M. D., “Devel-
opment of Pneumatic Thrust-Deflecting Powered-Lift Systems,” AIAA Paper 86-0476,
AIAA 24th Aerospace Sciences Meeting, Jan. 1986.
8Pugliese, A. J. (Grumman Aerospace Corporation), and Englar R. J. (DTNSRDC),
“Flight Testing the Circulation Control Wing,” AIAA Paper 79-1791, AIAA Aircraft
Systems and Technology M eeting, Aug. 1979.
’Hines, N., Baker, A., Cartagena, M., Largent, M., Tai, J., Qiu,
S.,
Yiakas, N., Zentner, J.,
and Englar, R. J., “Pneumatic Channel W ing Comparative M ission Analysis and D esign
Study, Phase I,” GTRI Technical Rept., Project A -5942, March 2000.
“Englar, R. J., and Williams, R. M. “Test Techniques for High Lift Airfoils with Bound-
ary Layer and Circulation Control for Application to Rotary Wing Aircraft,”
Canadian
Aeronautics and Space Journal,
Vol. 19, No. 3, 1973, pp. 93-108.
“Englar, R. J., Niebur, C. S., and Gregory, S. D., “Pneumatic Lift and Control Surface
Technology for High Speed Civil Transport Configurations,”
Journal of Aircraft,
Vol. 36,
‘’Jones, G.
S. ,
and Englar, R. J., “Advances in Pneumatic-Controlled High-Lift Systems
Through Pulsed Blowing”, AIAA Paper 2003-341 1, AIAA 21st Applied Aerodynamics
Conference, June 2003.
NO. 2, 1999, pp. 332-339.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 341/637
Chapter 12
Use of Circulation Control for Flight Control
Steven
P.
Frith* and Norman
J.
Woodt
University of Manchester, Manchester, England, United Kingdom
Nomenclature
b
= span, mm
CD
= drag coefficient
C L= lift coefficient
Cl
=
rolling moment coefficient
C = pitching moment coefficient
c
=
chord, mm
c =
standard mean chord, mm
=
mean aerodynamic chord
MAC),
mm
c , = root chord, mm
c , = tip chord, mm
D = drag,
N
h
=
slot height, mm
L
=
lift, N
= rolling moment, Nm
M = Mach number
m
= pitching moment, Nm
riz = mass flow rate, kg /s
p
= rate of roll
pW = freestream pressure, Pa
q = dynamic pressure
r
=
trailing edge radius, mm
CLc0,=
initial lift coefficient
Cl o)=
initial rolling moment coefficient
*Postgraduate Research Student, Fluid Mechanics Research Group, Aerospace Engineering.
'Professor, Head of Department, Aerospace Engineering. Senio r Member
AIAA.
Copyright 005 by the authors. Published by the American Institute of Aeronautics and
Member
AIAA.
Astronautics, Inc., with permission.
337
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 342/637
338
S.P. FRlTH AND N.
J.
WOOD
2
S = reference area, m
vj
=jet velocity, m/s
a
=
angle of attack, deg
p = orifice diameter to internal pipe diameter ratio
V =
freestream velocity, m /s
aC,/aC, = lift augmentation
aC,/at =
rolling moment derivative with respect to aileron deflection
A = wing sw eep angle, deg
p
= density
5= aileron deflection, deg
Subscripts
c /4
= quarter-chord position
D =
drag
= j e t
L
= lift
L = left jet on full-span model
1=
rolling m oment
m = pitching moment
LE
=
leading edge
max
=
maximum
plenum = associated with plenum parameters
R
=
right je t on full-span model
total = combined left and right je ts
trim = trim condition
p =
associated with blowing parameters
6=
aileron deflection
co = freestream
I. Introduction
IRCULA TION control (CC) has been recognized as a technique by which
C ery high lift coefficients can be achieved without the use of mechanical
control devices. It exploits the Coanda effect by blowing a high-velocity jet
over a curved surface, usually a rounded or near-rounded trailing edge (TE),
causing the rear stagnation point to move. In turn, the upper surface boundary
layer is energized, resulting in a delay in separation. As the circulation for the
entire wing is modified, there is an increase in overall lift, often much greater
when compared to more conventional mechanical lift devices.
Earlier researchla2has been focused mainly on two-dimensional unswept
wings, where the flow is predominantly attached to the airfoil. However, in
this work the performance benefits of the application of CC to a low aspect
ratio (AR) wing have been investigated. A delta-wing planform was chosen,
because the regions of separated flow could reveal additional properties of the
technique. Although more recent work3 uses pulsed jets in a bid to reduce the
total jet mass flow rate required, a steady jet was used in this investigation for
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 343/637
USE
OF CC FLIGHT CONTROL
339
model simplicity. With a system with few or no moving parts, the circulation
control wing (CCW) has generated considerable interest, as it is mechanically
simpler, and therefore cheaper to manufacture, and less prone to mechanical
failure in comparison with conventional high-lift devices. Also, lift increments
can be similar to those with conventional high-lift control surfaces, but pitch
increments can be lower, leading to improved aircraft control.
The initial aim of the study was to investigate the effect of various TE con-
figurations with a view to eliminating the cruise drag penalty attributed to
large TEs, while still obtaining high lift augmentation. The lift augmentation is
calculated as the ratio of increase of lift coefficient with blowing. A half-span
cropped-delta model was used to perform a parametric study of TE geometries.
This was then extended to a sting-mounted CC demonstrator consisting of a
generic unmanned air vehicle (UAV) planform with control surfaces with TE
sweep to determine whether there would be an interaction between the two jet s
and also whether CC could be used for roll control, within the limits of pitch
trim and maintaining high lift augmentation.
A lift augmentation of approximately 20 was achieved over all the blowing
ranges tested, with a maximum lift augmentation of 53 recorded. Nosedown
pitching moments were experienced, with a roll authority associated with these
measurements. Roll of the aircraft was possible with differential blowing of
the C C systems.
11. Half-Span Cropped-Delta Model
A. Experimental Procedure
For the preliminary studies to investigate a means of optimizing the CC
system, a half-span model was used to represent a circulation control wing
(CCW). A schematic of the model is shown in Fig.
1.
The CCW consisted of a
generic delta-wing LE section and a plenum/TE section, forming a cropped
delta-wing planform when connected. The LE section comprised a sharp LE
profile with a 50-deg LE sweep angle, incorporating strengthening sections
along the wing root to reduce flexing when under aerodynamic load. As shown
in Fig. 2, the plenum section was manufactured using 2-mm-thick brass sheet
for the lower surface and 3-mm-thick aluminum sheet for the upper surface.
The T E consisted of a 6-mrg-diam brass rod, giving a T E radius-to-mean-aerody-
namic-chord ratio of 0.005C.
A narrow convergent slot provides the je t blowing. This was achieved by man-
ufacturing a “knife-edge” section that ensured that there was a contraction w ithin
the plenum section to ensure the exiting fluid would attach to the Coanda surface.
This was constructed from aluminum and had a spanwise extent of 500mm,
dictating the length of the slot. This was incorporated into the top plate of the
plenum section. A series of pu sh-pgl screws allowzd the slot height to be
adjusted to 0.15 and 0.3 mm (0.00025C 5 h 5
0.0005
C .
The root chord was 853 mm and the tip chord was 254 mm, resulting in an
average chord measurement of 553.5 mm. The half-span measurement was
500
mm, with the slot extending the full distance, although because of the posi-
tioning of the splitter board, an inboard 5-mm section of the slot length was
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 344/637
340
S.P. FRlTH AND N.
J.
WOOD
round transition
\
Pressure
taps
\
a)
,Rectangu lar o I
I
Moun ting Strut
,
\
CCW
trai l ing edge
Fig. 1 Half-span cropped-delta model geometry: a) Upper surface view, b) cross-
sectional view.
permanently sealed off. The aspect ratio of the wing was calculated using
b 2 / S ,
giving a value of approximately 1.7.
The model was mounted from the overhead balance in the Avro
2.74 m x 2.13 m (9 x 7 ft) wind tunnel at the Goldstein Laboratory, M anchester,
UK, as shown in Fig. 3. A splitter board was mounted to ensure that the wind-
tunnel boundary layer did not interfere with measurements and the Coanda jet.
Force and moment data were measured using the six-component balance. The
Fig. 2 Close-up cross-sectional schematic of
TE
section.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 345/637
USE
OF CC FLIGHT CONTROL
34
Fig. 3 Model mounted in wind tunnel.
freestream velocity was set at 25 m/s, corresponding to a freestream Reynolds
number of approximately
8.5
x
lo5 , and m aximum jet ve locities were approxi-
mately 180 m/s.
The air supply was sourced from pressurized receiver tanks fed by an A tlas-
Copco compressor, delivered to the plenum by a flexible hose, such that tare
effects out of the plane of measurement were avoided. The pressure within the
plenum was monitored with a pressure tapping and using a pressure transducer.
The mass flow rate was determined using an orifice plate rig with pressure
transducers and an orifice plate with orifice-to-pipe-diameter ratio
p
of 0.2401.
The pressure and flow temperature data were transferred to the computer via
an A-to-D card.
A computer program was written to accumulate data and calculate the flow
rate. From this the blowing m omentum coefficient C , could be calculated using
Vjriz
c
=
s
where vj is the velocity of the Coanda jet,
r z
is the jet mass flow rate, q is the free-
stream dynamic pressure, and
S
is the model surface area. The jet velocity was
calculated using the isentropic pressure distribution
to avoid errors that can occur using the jet area as a variable. As interest was
directed at low blowing rates, data were recorded at increments of C of
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 346/637
342 S.P. FRlTH AND N.
J.
WOOD
0.0005 up to 0.01 and then using increments of 0.005 up to 0.03 to obtain general
force o r moment curves.
Particle image velocimetry (PIV) was also performed to obtain more infor-
mation on the interaction of the je t with the freestream flow.5 A horizontal light-
sheet was fired at the TE of the CC W using an Nd:YAG laser. A megapixel CC D
camera, positioned under the wind-tunnel floor, captured a sequence of pairs of
images of the seeded freestream flow over the wing. The images were then
time-averaged and analysed using TSI Insight and Tecplot
9
software to obtain
velocity and vorticity data.
As part of a joint project,
BAE
Systems6 calculated computer fluid dynamics
(CFD ) data to compare with the experimental data. There was a reasonable agree-
ment between the computed and experimental data, although it was felt that
additional refinement of the grid would enhance results.
B.
Results
The aerodynamic data are represented as a series of carpe t plots. The lift carpet
plots show the variation of the lift coefficient CLwith blowing at angles of attack
ranging from 0 to 15 deg in 5-deg increments. The blowing values at a particular
angle of attack were offset, with the lowest angle of attack being offset by the
greatest amount, effectively removing the need for a horizontal axis. Lines of
constant blowing coefficient
C
link the lines of data for each angle of attack.
The vertical axis represents the lift coefficient C with dashed horizontal grid-
lines of constant CL sed for reference. An example of this is demonstrated in
Fig.
4.
The lift augmentation is calculated from the gradient of the lift curve
for each angle of attack over a particular blowing coefficient range.
The pitching moment carpet plots are to a large extent very sim ilar in layout to
the lift carpet plots. The main difference is that the offset for the blowing coeffi-
cients is reversed, such that the highest angle of attack is offset by the greatest
amount. The vertical axis represents the pitching moment C about a particular
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fig.
4
Example of a lift data carpet plot.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 347/637
5
USE
OF CC FLIGHT CONTROL
...........................................................
.
Increasing
343
Blowing
............................................................
I
Fig.
5
Example of a p itching mom ent data carpet plot.
reference point (in the case of the half-span model, this is the LE), as shown
in Fig. 5 .
Th e results given in Fig.
6
show the effect of CC on the lift characteristics w ith
a variation in slot height. There is an increase in lift with an increase in C
although the greatest lift increments were found at lower blowing rates. The
level of lift augmentation
aC,/aC,
is of the order 18-25 over the complete
range
of C
tested, although a maximum incremental augmentation of approxi-
mately 53 was recorded. Also, it was found that the smaller slot height yields a
stronger lift augmentation at smaller values of C . This may be due to the
1
0.8
2
0.6
If 0.4
G
J 0.2
4
0,
.-
8
0
0.2
Fig. 6 C L vs C,: effect of slot height on CC.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 348/637
344
S .
P. FRlTH AND N.
J.
WOOD
smaller slot height giving rise to an increase in the ratio between the m omentum
of the jet and that of the freestream flow. It is anticipated, however, that a
minimum slot height will be reached where the jet no longer attaches to the
Coa nda surface. Th is requires further research.
The pitching moment data are characterized by a negative gradient, depicting
a nosedown pitching moment, as shown in Fig. 7. This is typical of a C C system ,
as also seen by Jones and E n g la ~ - .~he rate of pitching moment is greatest at the
lower blowing coefficients, although it levels off with increasing jet momentum.
There are similar trends for each angle of attack, although data for the model at
15 de g revealed that the nosedown p itching moment of the model was less than
that obtained for the model at l0 deg . This can be attributed to the more powerful
nature of the LE vortex at higher angles of attack, the increased suction on the
upper surface giving rise to a slight nose-up pitching moment.
The drag coefficient was also found to increase as the blowing rate is
increased, although the dra g augmentation is significantly less than the equivalent
value for lift, suggesting an overall increase in
LID.
However, drag measure-
ments are not presented in this chapter because of an inconsistency in the data,
which may be du e to fluctuations in the Coanda jet or the accuracy range of
the balance.
Fig.
8
shows the calculated time-averaged velocity vectors obtained from PIV
in the form of a contour plot using the TSI Insight and Tecplot softwares for
different values of C,. It can be seen that the external flow visibly changes at
higher blowing rates, indicated by a downward deflection of the velocity
vectors. Th e data also demonstrate that the downstream exten t of the wake was
reduced. There is also an area of accelerated flow over the upper surface, just
........................................................................
o‘2 T
G O
J
-
E
0” -0.2
a
0,
-0.4
U
C
-0.6
-
-0.8
a -1
E
-1.2
I
3
0
c
C
a
8
p -1.4
.-
r
f -1.6
-1.8
..............................................................................
- 5 d e g r e e s
A
10
degrees
...............................................................
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Fig. 7 Variation of pitching mom ent with blowing for half-span model with
0.15-
mm slot height lower blowing coefficients range:
0
C,
0.01).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 349/637
USE
OF CC FLIGHT CONTROL
345
b)
Fig. 8 PIV velocity contour plots with streamlines obtained for angle of attack
10
deg at different blowing coefficients: a) C , =
0,
b) C , =
0.005,
c) C , =
0.01.
before the je t exit. Because of restrictions w ith the apparatus it was not possible to
seed the jet and investigate the full interaction with the freestream flow.
111. Full-SpanUAV Configuration
A. Experimental Procedure
A full-span model was designed and constructed at the Goldstein Laboratory,
Manchester, to investigate any interaction of the Coanda jet s and exam ine the
possibility of roll control, as well as lift enhancement. A schematic is shown in
Fig.
9.
The main body w as constructed using modelboard. The model had a L E sweep
angle of 55 deg and a TE sweep of
-30
deg, resulting in the diamond-shaped
planform as shown in Fig. 9. The plenum sections, made from aluminum for
the upper su rface and brass for the lower surface, incorporated similar T E dimen-
sions as the previous model: TE diameter of
6
mm and slot height adjustment
from 0.05 to
0.30
mm (this was set at
0.15
mm to compare with previous
results). The spanwise extent of each slot was reduced to 300mm and did not
extend the full length of the TE. The blowing rate was again controlled using
an orifice plate rig for each plenum p
= 0.3),
such that the plenum sections
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 350/637
346
S. P. FRlTH AND N.
J.
WOOD
Balance
Mounting
Plate
Fig. 9
Plan view schematic of full-span model.
could be controlled independently. The air supply was controlled by the use of
two valves for each plenum, allowing finer and more accura te control. The fuse-
lage section of the model was manufactured from aluminum sheet, to create a
theoretical aircraft profile and provide protection for the instrumentation (dual-
axis inclinometer and strain-gauge balance) and air supply within the model.
The model was mounted on a sting in the Avro
9 x
wind tunnel as shown in
Fig. 10, incorporating an internal six-component strain-gauge sting balance to
measure aerodynamic forces and moments.
The air supply was again taken from pressurized tanks and passed through a
series of flexible hoses. Tare effects because of flexing of the hoses when
under pressure were minimized by incorporating highly flexible hose within
the model, adjacent to the calibration center of the balance. Any tare effects
resulting from flexing of hoses were measured wind-off.
Preliminary tests were performed prior to load data acquisition to determine
efficiency of both Coanda surfaces, check for any leakages and uniformity of
both slots. Test runs were made in the wind tunnel to examine model integrity
and performance.
Tests were accomplished at 25 m /s (a freestream Reynolds number of approxi-
mately 1.3
x
lo6) and the angle of attack was varied from
0
to 15deg in 5-deg
increments. The blowing coefficient was varied from zero to 0.004 in increments
of 0.0005. Data were taken for various test parameters: symmetric blowing, in
which the jet momentum from both plenums was identical, and differential
blowing, in which one plenum would maintain a constant C and the other side
would operate over the complete range. Table 1 summarizes the test procedure
undertaken.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 351/637
USE
OF CC FLIGHT CONTROL
347
Fig. 10 Sting-m ounted model in wind tunnel.
B. Results
In quiescent conditions, both C oanda jet s performed a s expected, with the jet s
fully attaching to the Coanda surfaces, verified using tufts. It was possible to
maintain a tunnel velocity up to
40
m/s without the model experiencing signifi-
cant fluctuations, although all tests w ere performed at 25 m/s for consistency.
The load data for the full-span model are again represented in a series of carpet
plots, as described in Sec. 1I.B. In addition to lift and pitching moment plots,
rolling moment data are represented in a similar form to the pitching moment
data, with the vertical axis giving values of
Cl.
Figures 11 to 18 show the e ffectiveness of the full-span model, in the form of
carpet plots with contours of constant C and angles of attack. A lift augmenta-
tion aCL/aC, of 17-24 was achieved, as demonstrated in Fig. 11, in which data
are shown for both Coanda jet s at the same mass flow rate, and therefore the sam e
C (symmetric blowing). Although the lift augmentation achieved is not as
great as that achieved in other studies,’ it is believed that this can be attributed
to the small radius of the Coanda surface. The tradeoff of a lower lift augmenta-
tion is that the drag fo r such a surface is reduced when compared to traditionally
large CC Coanda surfaces.
Table
1
Test proced ure for full-span m odel
Left plenum section Right plenum section
Configuration blowing coefficient,
CI* L)
blowing coefficient, CI* R)
Symmetric
Differential 0
Differential
0.002
Differential
0.004
0- .004
0-0.004
0- .004
0-0.004
0-0.004
0 (constant)
0.002
(constant)
0.004
(constant)
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 352/637
348 S.P. FRlTH AND N.
J.
WOOD
1 ............................................................................
I
-
5 degrees I
Fig. 11 Variation of lift with blowing for full-span model with both system s active
blowing coefficients range: 0
Cp total)
0.008).
Assuming the center of gravity to be at the quarter-chord position, the pitching
moment about this point is nosedown (Fig. 12), which is as expected because the
cente r of lift is located aft of the quarter chord. It is encouraging to see that the C C
device could be used to trim the aircraft, while maintaining high values of lift
augmentation, as the variation in
C
required at various angles of attack is
approximately linear, as shown in Fig. 13. This suggests that the control of this
parameter could be simply transferred to stick control in a real-flight situation.
The investigation in using CC for roll control revealed some interesting
characteristics. The variation of lift with asymmetric blowing (zero blowing
from the right Coanda jet) is shown in Fig. 14. Again, a lift augmentation of
approximately 20-25 is achieved and it was demonstrated that the je t mo mentum
is additive; that is, if the left jet was used at the maximum value of C the acti-
vation of the right jet would result in a similar lift curve to that obtained with
symmetric blowing.
Th e control of rolling moment by C C is demonstrated in Figs. 15 to 18. Figures
15 and 16 demonstrate the effect on the rolling moment of the use of just one
system. It can be seen that a particular rolling moment can be achieved with a par-
ticular value of C independent of the angle of attack, although the leading edge
vortex, particularly effective at angles of attack from approximately 7.5 deg
produces an additional pro-roll moment. This pro-roll moment results fro m a sec-
ondary effect of the blowing that enhances the vortex suction signature ahead of
the blowing slot.4 This can be seen in the kink in the rolling m oment curves.
Th e graphs shown in Figs. 17 and 18 indicate how the differential blowing
affected the rolling moment of the model. The first of these shows the use of
the right system at half the maximum blowing possible
(C, =
0.002) held
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 353/637
USE
OF CC FLIGHT CONTROL
349
0 5 10
15
0.04
5 -0.02
0
c
m
C
0.04
r
m
C
-0.06
h
-0.08
Fig. 12 Variation of pitching mom ent with blowing for full-span model with both
systems active blowing coefficients range:
0 5
Cp total)
0.008 .
constant, while increasing the je t momentum on the left system through the entire
range possible (0 5 C 5 0.004).Th e effect of the vortex is clearly ev ident, with
the most influence with the right system at C
=
0.002 and the left system at
either
C
= 0 or C, = 0.004. Figure 18 shows how the system can be returned
to a state of zero roll from a rolling motion.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 354/637
350
1
0.8
0
0.2
S.P. FRlTH AND N.
J.
WOOD
Fig. 14 Variation of lift with blowing for full-span model with left system active only
blowing coefficients range:
0
5 Ca r)5 0.004).
Adapting ESDU Data Items Aircraft
06.01.01*
and 88013,9 it was possible to
determine the values for the rolling moment derivatives resulting from aileron
input and damping. These are given in Table 2 . From these values, a roll rate
of approximately 403 de g/s is achieved with a single downw ard aileron deflec-
tion of 10 deg.
............................................................................
......................................
/
I 5 degrees
............................................................................
............................................................................
............................................................................
Fig. 15 Variation of rolling moment with blowing for full-span model with left
system active only blowing coefficients range: 0
5
C a ~ )
0.004).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 355/637
0.04
0.03
q 0.02
m
-
0'01
s o
:0.01
a 0.02
.-
0.03
-0.04
USE OF CC FLIGHT CONTROL
.............................................................................
.............................................................................
35
................... ............................ -5 m r e e s ..
-A- 10degrees
,=0.004
.............................................................................
Fig. 16 Variation of rolling moment with blowing for full-span model with right
system active only blowing coefficients range:
0 f i ~ ) .004).
If
the ailerons were substituted with the
CC
system, the parameter
aC,/aC,
would no longer be valid and the parameter aC,/aC, would replace it. This is
essentially the gradient of the rolling moment curves with blowing; a mean
value of approximately
7
is obtained. The parameter
8
would also be replaced
by C . Assuming the response of the rolling moment with blowing is
approximately linear, along with the response due to aileron deflection, it is
0.04
0.03
0 0.02
-
m
-
01
0
5
-0.01
a 0.02
0
.-
-
0.03
-0.04
.............................................................................
........................................-
..
-
l c ,(ToTAL) = 0.002
...........................................................................
Fig. 17 Variation of rolling moment with blowing for right CC system blowing at
constant
C p ~ )
0.002
and increasing blowing on left
CC
system
0
Ca L)
0.004).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 356/637
352
S.
P. FRlTH AND N.
J.
WOOD
Table
2
Calculated values for ro lling moment
derivatives of full-span model
acIm
acIm
- .084
rad-
-0.052
rad-
possible to equate the equivalent blowing coefficient required to generate the
same rate of roll such that
Th is gives a
C
of approximately 0.0021, equivalent to an aileron deflection of
approximately 10 deg. Th e slight negative rolling moment present at an angle of
attack of 0 deg and C = 0 indicates that there is a slight model asymmetry,
although this does not have a significant impact on the effectiveness of the system.
IV. Conclusions
An experimental investigation of
CC
has been successfully modeled, initially
on a single delta-wing configuration with varying T E geometry and then on a full-
span model to investigate the potential for roll control.
The variation of slot height indicated that a smaller slot height yielded a higher
lift augmentation aCL/aC,. However, it is anticipated that there is a limiting
0.04 .............................................................................
0.03
.............................................................................
I
.............................................................................
:-
0.02
-
.-
g
0.01
.............................................................................
O degrees
8 +-lodegrees
= -0.01 .........
0
0.02
..................
......
.......
................. 4 1
0.03
.
TOTAL =
0,004
0.04 .............................................................................
Fig.
18
Variation of rolling moment with blowing for right CC system blowing at
constant
C a ~ )
0 004 and increasing blowing on left CC system 0
C p ~ )
0.004).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 357/637
USE
OF CC FLIGHT CONTROL
353
height, requiring further work. Lift augmentations of approximately 18-25 for
low blowing rates were obtained with both models over the complete lower
blowing range. This suggests that useful lift increments can be obtained with
C
values of the order
0.005,
equivalent to those achieved using existing flap
systems ACLx 0.1). As the CC system is considerably less complex mechani-
cally than other high-lift devices, this may be significantly beneficial when
contemplating maintenance, production costs, and reliability.
The full-span tests demonstrated that the CC system was effective at generat-
ing significant rolling m oments a t low blowing coefficients. Importantly, the pro-
duction of roll moments can be superimposed on the lift generation, suggesting
minimized interaction and simple control development.
More detailed work a t even sm aller increments of C especially in the lower
blowing regions, w ill enable greater understanding of the physics involved in
CC
and the areas of higher lift augm entation. Power requirements for blowing need
to be studied to determine the overall efficiency of the system compared to
conventional systems. Further experimental work using the full-span model
will continue to investigate the application of CC to roll control and pitch trim.
The implementation of pulsed jets will also reduce the required mass flow
bleed, yet provide similar lift augmentation^ ^
Acknowledgments
The authors wish to acknow ledge the contributions of staff and students at the
Goldstein Laboratory at the University of Manchester, especially those of the
technicians for their help with model manufacture. A special mention must
also go to Andrew Kennaugh for his continuous help throughout the project.
References
‘Wood, N. J., and Nielsen, J. N., “Circulation Control Airfoils-Past, Present, Future,”
AIAA, 23rd Aerospace Sciences Meeting, Jan. 1985.
Englar, R. J., and Applegate, C. A., “Circulation Control-A Bibliography
of
DTNSRDC Research and Selected Outside References (Jan. 1969 through Dec. 1983),”
DTNSRDC Rept. 84/052, Sept 1984.
3Jones, G.
S.,
and Englar, R. J., “Advances in Pneumatic-Controlled High-Lift Systems
Through Pulsed Blowing,” 21 st AIAA Applied A erodynamics Conference, June 2003.
4Frith, S. P., and Wood, N. J., “Effect of Trailing Edge Geometry on a Circulation
Control Delta Wing,” 21st AIAA Applied Aerodynamics Conference, June 2003.
’Raffel, M., Willert, C., and Kompenhans, J., “Particle Image Velocimetry-A Practical
Guide,” Springer, Berlin, 1998.
ellars, N. D., Wood, N. J., and Kennaugh, A., “Delta Wing Circulation Control Using
Th e Coanda Effect,” AIAA 1st Flow Control Conference, June 2002.
Englar, R. J., “Circulation Control Pneumatic Aerodynamics: Blown Force and
Moment Augm entation and Modification-Past, Present and Future,” AIAA Fluids
2000 Conference and Exhibit, June 2000.
Stability Derivative, L,, Rolling Moment due to Rolling for Swept and Tapered
Wings,” Engineering Sciences Data Unit, Item A 06.01.01, March 1955.
’“Rolling Moment Derivative, Lg, for Plain Ailerons at Subsonic Speeds,” Engineering
Sciences Data Unit, Item 88013, August 1988.
2
7
8“
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 358/637
1I.C.
Experiments and Applications: Nonaerospace
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 359/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 360/637
358
R. J.
ENGLAR
performance, and safety. As discussed in Ref. 1, for a typical
U.S.
tractor-trailer
rig logging 175,000 miles a year at a fuel price of $lS O /gall on , yearly fuel costs
could average over
40,000
($29,000 if only 125,000 miles are logged). Thus
even a 5-10% increase in fuel econom y could be meaningful. Although
devices that can reduce the HV’s drag coefficient can significantly improve
fuel economy, it is also desirable that additional capabilities result from improved
aerodynamics. These could include increased stability (both lateral and direc-
tional), reduction in side-wind sensitivity, reduction in splash and spray, and
improved traction plus aerodynamic braking. One could also include an aerody-
namic m eans to reduce tire rolling resistance. Any such devices being considered
for these applications should also be sim ple and robust, contain few or no m oving
parts, should not be hampered by weather, and not increase vehicle weight or
external dimensions. This paper discusses pneumatic aerodynamic devices
based on the use of circulation control (CC) aerodynamics, which thus possess
many of these desirable characteristics. These are currently under development
at Georgia Tech Research Institute (GTRI) for the DOE Office of Heavy
Vehicle Technology. First described in the following sections will be the basics
of pneumatic aerodynam ics and application to heavy vehicles, and then details of
wind-tunnel and full-scale programs conducted, their results, and possible future
applications.
11. Basics of Pneumatic Circulation Control Aerodynamics
GTRI researchers have been involved for a number of years in the develop-
ment of pneumatic (pressurized air blowing) concepts to yield efficient yet
mechanically simple means to control, augment, or reduce the aerodynamic
forces and moments acting on aircraft. This was detailed in Refs. 2 to 4,
among others, but will be summarized briefly to familiarize the reader with
this technology. Figure
1
shows the basic pneumatic concept, which has
becom e know n as circulation control (CC) aerodynam ics. Here, an airfoil’s con-
ventional mechanical trailing-edge (TE) device has been replaced with a fixed
curved surface and a tangential slot ejecting a jet sheet over that surface. That
jet remains attached to the curved surface by a balance between subambient
static pressure on the surface and centrifugal force (the so-called Coanda
e f f e ~ t ) . ~his entrains the external flowfield to follow the curving jet, and thus
enhances the circulation around the airfoil and the ae rodynam ic forces produced
by it. The governing parameter is not angle of attack, but rather the blowing
mom entum coefficient:
mV,
c,
=qs
where m is the jet mass flow,
vj
the isentropic je t velocity,
S
is a reference wing
area (or frontal area
A
for a ground vehicle), and q is the freestream dynam ic
pressure
q
= 0.5pV2, with p being the freestream density). At lower
C,
values, augmentation of the aerodynamic lift by a factor of ACl/C, =
80
has
been recorded? representing an
8000%
return on the invested jet momentum
(which in a physical sense is also equal to the jet thrust). Familiarity with
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 361/637
IMPROVING PERFORMANCE OF AUTOMOTIVE VEHICLES 359
POSSIBLE LEADING EDGE B LOWING
TANGENTIAL BLOWING OVER ROUNDED TRAILINQ EDGE SURFACE
t
BOUNDARY LAYER CONTROL
c p
=
IilvJIqs
MOMENTUM COEFFICIENT, Cp
FORCE BALANCE
J E T
SHEEF
Fig. 1 Basics of circulation control pneumatic aerodynam ics on a simple two-
dimensional airfoil.
blown aerodynamic systems will remind the reader that this is quite
extraordinary; thrust-deflecting vertical takeoff and landing (VTOL) aircraft
are fortunate if they recover anything near
100%
of the en gine thrust expended
for vertical lift (which must ex ceed weight), with very little, if any, augmentation
of aerodynamic lift occurring.
It is because of this high return on blowing, o r conversely, b ecause of the very
low required blowing input and associated power required to achieve a desired
lift, that
CC
airfoils appear very promising for a number of applications. The
A-6/CC
Wing short takeoff and landing (STOL) flight demonstrator aircraft
(Fig.
2 *
showed the STOL performance listed, and also suggested capabilities
very useful to ground vehicles: during short takeoff, it demonstrated high lift
with reduced drag,
and in the approach/landing mode, very high lift
with high
drag was shown.
The se advantages led to the application of this pneumatic concept to imp rove
the aerodynamics of an already streamlined car model.5 The resulting large jet
turning over the curved rear of this vehicle is show n in Fig.
3.
Significant but dis-
tinctly different trends were observed during testing, depending upon which
portion of the tangential slot located along the trunk break line was blown.
Blowing the full-width slot produced the large jet turning shown by the striped
tuft in Fig.
3,
with drag increases of greater than
70 ,
showing potential for
pneumatic aerodynamic braking. Blowing only the outside segments of the slot
weakened the comer vortex rollup, attached separated flow, lessened aft
suction, and reduced drag by as much as
35%.
Blowing this aft slot also
yielded a lift increase of 170%.One can envision a similar slot applied to the
lower rear surface that could instead yield negative lift (positive down force).
This concept has been patented by GTRI and verified by a similar installation
on a wind-tunnel model of a European Formula
1
race car.6
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 362/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 363/637
IMPROVING PERFORMANCE OF AUTOMOTIVE VEHICLES 361
Fig.
4
Schematic of application of GTRI pneumatic technology to heavy vehicle
trailer, showing four aft blowing slots and upper
LE blowing slot.
demonstration of an operating pneumatic heavy vehicle (PHV). Figure
4
hows a
schematic of a generic PHV with tangential blowing slots on each of the trailer’s
four curved aft edges, plus blowing on the rounded upper leading edge (LE) of the
trailer. Early portions of that effort, including a preliminary feasibility study and
design of baseline and pneumatic wind-tunnel configurations, are detailed in Ref.
6 .
A. Wind-Tunnel Evaluations of Baseline Unblown HV M odels
To develop a representative PHV configuration prior to full-scale testing,
initial baseline wind-tunnel testing was conducted, which was then followed
by several phases of blown test configurations. For this, an existing generic
HV configuration, the ground transportation systems (GTS) vehicle of Ref. 7,
was used. The model is shown in Fig.
5
before the blowing modifications were
0.95
0.9
0.85
0.8
0.75
0.7
CD
0.65
0.6
0.55
0.5
0.45
0.4
0.35
0
5
10 15
20 25
30
35 40 45
9.
PSf
Fig. 5 Test results for unblown HV models, showing effects of cab height, gap,
wheels, and Reynolds number.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 364/637
362
R. J.
ENGLAR
installed. It is actually representative of a faired cab-over-engine HV based on a
Penske racing team c ar carrier, and is relatively independent of the numerous and
varying cab roof fairings employed on a number of current HV. Tests of this
unblown model configuration did, however, demonstrate the importance of
cab/trailer gap and fairing treatments. These configurations were tested in the
GTRI Model Test Facility research tunnel6 and showed some significant drag
reductions because of changes in the unblown geometry. Figure 5
shows drag
reductions of up to
25
below a low-cab full-open-gap vehicle when the gap
was eliminated (filled in) and the cab top was even with the trailer top (trailer
leading and trailing edges are square here). An additional 15% reduction was
confirmed with a round trailer LE facing into the open gap and a round TE on
the trailer (this is the unblown PHV). These data were taken at a typical tunnel
speed of 70mph. Also very significant is the tremendous increase in
D
n
Fig. 6 (more than a doubling is seen) due to a side wind acting at a yaw angle
on the HV. (In all of the drag data shown herein, D s based on projected
frontal area of the vehicle
A,
including the wheels.)
B.
Based on the preceding unblown configurations with reduced drag , additional
wind tunnel tests were conducted to evaluate aerodynamic improvements
Wind-Tunnel Evaluations
of
Blown HV Configurations
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1 1
CD
1
0.9
0.8
0.7
0.6
-15
-10
5 0 5 loNose right 15
Yaw Angle, y~, eg
Fig. 6 Effects of side wind on drag for various unblown HV configurations.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 365/637
IMPROVING PERFORMANCE OF AUTOMOTIVE VEHICLES 363
resulting from blown configurations. Details of these investigations are presented
in Ref.
8.
Unless otherwise noted, the blowing variations were run at tunnel
(vehicle) wind speeds of approximately 70 -71 mph (dynamic pressure
q
=
11.86 psf and Reynolds nu mber
=
2.5
x
lo 6, based on total tractor
+
railer
length).
C.
Drag Reductions (for Fuel Economy) or Drag Increases
(for Braking Stability)
The blowing slot heights at each aft edge of the trailer could be varied and
tested either unblown or blown in any combination of the four, or even with
LE slots on the trailer front face also blown. Flow visualization tufts in Fig. 7
show jet turning of 90deg on all four aft corners, even the bottom slot
blowing upwards. Figure
7
also shows the results of this jet turning on reducing
or increasing aerodynamic drag by blowing various combinations of these aft
slots. The combination of all four slots blowing together yielded the greatest
drag reduction, more effective than blowing individual slots. Compared to the
typical unblown baseline configuration from above (full gap between cab and
trailer, square trailer LE and TE, and cab fairing slightly lower than the trailer
front), which produced D = 0.824 at this Reynolds number, the blown configur-
ation reduced the drag coefficient to
0.459
at C,
= 0.065.
This is a 44% D
reduction, and the internal plenum blowing pressure required was only
0.5
psig.
A
second blown configuration (labeled 90°/300 TE) used less jet
turning on the upper and lower surfaces to generate even greater drag
reduction-at
0.5
psig,
CD
was reduced by 47 %, and a t 1.0 psig (C,
=
0.13),
D was reduced by
50%.
These data are all for a smoothed bottom tractor-
trailer model with low sides and half-cylinder simulated wheels.
Momealum Coellleieat,
C,
Fig. 7 Drag reduction or augmentation on blown trailer with 90-deg turning
surfaces, plus flow visualization of jet turning.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 366/637
364 R. J.
ENGLAR
Additional evaluation of the effectiveness of the blown configurations was
made. The drag coefficient of the preceding unblown baseline configuration,
but with the tractor-trailer gap filled in, is CD= 0.627 (not shown in Fig. 7).
Addition in Fig.
7
of the unblown pneumatic surfaces onto the trailer TE
reduces CD by 9.7%. Adding blowing at
C
= 0.065 reduces that C D by
another 23.1%. This combination reduces CD to 30.6% less than the square TE
baseline having a smooth fairing filling in the gap.
When only the top slot, the bottom slot, or both of these slots were blown in the
absence of the side jet s, drag was initially reduced slightly, but then significantly
increased with the addition of blowing. This represents an excellent aerodynamic
braking capability to supplement the hydraulic brakes. Blowing efficiency is
plotted in Fig.
8,
where
ACD
s an increment from the blowing-off value (negative
ACD
is reduced drag). Absolute values of
ACD /C ,
greater than
1.0
represent
greater than 100%
return on the input blowing
C,.
It is seen that the 90 deg /
30 deg four-slotted configuration generates values as high as -
.50,
representing
550 of the input blowing momentum recovered as drag reduction. This figure
also shows the opposite trend, with up to 200% of the blowing momentum from
top/bottom slots recovered as increased drag for braking. Obviously, these per-
centages will be modified when the power expended to compress the blowing
air is included, but that will have to await a full-scale sys tems study.
Should additional air be available from an onboard source such as an existing
turbocharger or an electric blower, additional drag reduction is possible, as shown
in Fig. 9. Drag coefficients of less than 0.30 are shown for faired blown
HV
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0
2 . 5
-3.0
-3.5
-4.0
-4.5
-5.0
5 . 5
0.15 0.12
0.09
0.06 0.03 -0.00
-0.03 -0.06
-0.09 -0.12 -0.15
ACD = C D.CDO
Fig. 8 Blowing efficiency and drag increm ents caused by blowing slot configuration.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 367/637
IMPROVING PERFORMANCE OF AUTOMOTIVE VEHICLES 365
0 0.1 0.2 0.3 0.4
0.5
0.6 0.7
0.8
0.9 1
Momentum C oefficient, C p
Fig. 9 Reynolds number effects and increased blowing values, plus LE blowing and
gap plates.
configurations. This is in the arena of streamlined sports cars. The drag coefficient
of a 1999 Corvette coupe is
C, =
0.29. Figure 9, originally intended to show that
the drag curves tend to converge onto one slope independent of Reynolds
number, also shows a measured drag coefficient of 0.13 for the PHV model at
increased C,. This is about half the drag coefficient value of the Corvette or a
Honda Insight hybrid (C, = 0.25). Even though not achieved in the most effi-
cient blowing operation range, this is an 84% drag reduction compared to the
unblown baseline configuration. Note that the tractor cab in Fig. 9 has “gap
plates” (or fairing extensions) instead of the full “no gap” fairing of Fig. 8, and
is thus much closer to an actual tractor/trailer configuration. It also has
blowing on the trailer top LE. Figure 10 shows this alternative m eans of improv-
ing upon the gap problem.
Note that when com paring these data to other experim ents that have been con-
ducted by other researchers on similar GTS models, these GTRI data above and
below include simulated wheels, which, as Fig. 5 shows, add about AC,
=
0.18
to nonwheeled vehicles’
C,
values, perhaps more, depending on how well the
tunnel ground effects are treated experimentally. GTRI’s measured data are
recorded using test-section tangential floor blowing to eliminate floor bound-
a r y - l a y e r i n t e r f e r e n ~ e . ~ ’ ~ , ~
D. Stability and Control
Strong directional instability can be experienced by HVs at yaw angles (i.e.,
when experiencing a side wind) because of large side forces on their flat-sided
trailers (Fig. 11). This yaw sensitivity is confirmed by the unblown (C, =
0)
yawing moment CN shown, where yaw angle as small as
-
deg produces a
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 368/637
366
R. J.
ENGLAR
0 65
0 60
0
s
0 so
CD
0 45
0
40
0 35
0
30
O W
005
0 1 0 015 020
0 2 5 030
Momeolum
Cafl5eieot.Cp
Fig.
10
Side plates and trailer top blowing: a practical solution to the cab gap
problem.
large unblown yawing moment coefficient of
C, =
-2.0. (It should be noted
here that this yawing moment is measured about the rigid model's midpoint at
the ground, whereas on a real articulated tractor-trailer, it would be experienced
at the tires
of
the individual units. However, comparisons of blowing on and
off
Nose
Right 4.0
90°/300 1 / 2 TE ,0.375 R
p 00
O ,
Wheels ON , Left Slot Blowing Only
N = Half Chord Yawing Moment
y O ,
M s e
Straight Ahead
Coefficient
1.5
1 o
0.5
0.0
0.5
-1.0
-1.5
CN
ose Yawed Left
Nose
Lefl
' ' 8 ' ' ' ' ' 8 . .
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20
Momen tum Coemcient, C
Fig. 11 Directional control capability provided by HV configuration with left rear
slot blowing only.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 369/637
IMPROVING PERFORMANCE OF AUTOMOTIVE VEHICLES 367
are being made for the same single HV unit, and apparent benefits should be
valid.) Blowing only one side slot can easily correct this. With the nose straight
ahead (Ic,
=
0 deg), blowing the left slot at C = 0.06 yields the equivalent oppo-
site yawing moment
C N =
+2.0). With the nose yawed left (for example,
Ic,
= - deg), blowing at C = 0.06 returns the unstable yawing moment to
CN=
0.0.
Then, increasing the blowing a bit more causes the nose to yaw in
the opposite direction, to the right. The opportunity for a no-moving-part,
quick-response aerodynamic control system is apparent.
IV. Pneumatic HV Fuel Econom y Testing
The preceding model tests led to the conclusion that a full-scale proof-of-
concept test series should be conducted on a PHV test rig to determine if the
tunnel results would also occur on the road. Based on the above wind-tunnel
results, GTRI team member prototype shop Novatek, Inc. designed and fabri-
cated the PHV blown test trailer, including blowers, drive motors, control
systems, and instrumentation. This configuration is shown in Fig. 12 in compari-
son to a stock (reference) Great Dane trailer. Blown tufts confirming je t turning of
90
deg around the right-side TE curved pneumatic surface are shown in Fig. 13.
A. TuningTests
Test vehicle fabrication and assembly w ere completed at GTRI and the modi-
fied trailer was then picked up by team member V olvo Trucks of North Am erica
and moved to its facility in Greensboro, NC. Here two initial tuning tests were
conducted (Fig. 14). Figure 15 show s a rear view of the pneumatic trailer
with the tufts confirming on-road flow turning. These tests verified the test equip-
ment and data system operations, and indicated unofficial fuel economy increases
from blowing of up to 15.3% over the baseline trailer when on an open highway,
rather than on a test track.
Fig.
12
PHV pneumatic trailer (left) and baseline reference unmodified trailer
(right).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 370/637
368
R. J.
ENGLAR
Fig.
13
Right rear corner view, looking up, with tufts showing 90-deg jet turning.
B.
Full-scale
PHV
On-Track Fuel Economy Tests
On-track testing
of
the PHV test vehicle (tractor and modified trailer) was con-
ducted at the Transportation Research Center (TRC) test track in East Liberty,
Ohio, along with a control vehicle (a stock Volvo/Great Dane rig). Figure
16
shows these two vehicles while in a pit lane fuel station at the 7.5-mile banked
Fig. 14 On-road PH V tuning test near Volvo facility in North Carolina.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 371/637
IMPROVING PERFORMANCE OF AUTOMOTIVE VEHICLES 369
Fig.
15
On-road
PHV
test vehicle rear view with jets blowing and tufts turning.
test track at TRC .
SAE
Type-I1 fuel economy runs were conducted by the TR C/
GTRI/Volvo/Novatek team members in strict accordance with S A E est pro-
cedures (as specified in SAE J1321, Oct. 1986). During these tests, a total of
59 runs was made for the six configurations evaluated, each at three different
Fig. 16 Test and control vehicles in pits at TR C test track.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 372/637
370
R. J.
ENGLAR
Fig. 17
PHV
test vehicle on track at 75 mph.
speeds (55 ,65 , and
75
mph) and with each run covering six laps
(45
miles) of the
TRC test track.g91113 Figure 17 shows the pneumatic test tractor-trailer a t speed
on the TRC track.
The six sets of fuel economy runs were made at different blowing rates,
including zero blowing. This allowed reference comparisons to be made after
the pneumatic test trailer was reconfigured into the baseline trailer and then
tested to provide reference fuel economy of the standard vehicle (all fuel
economy data achieved with the other test configurations were compared to
this one to determine percent fuel efficiency increase, %FEI). Figure
18
shows
0 0.01
0.02 0.03
0.04
0.05 0.06 0.07 0.08
0.09 0.1
Blowing Momentum Coefficient, CF
Fig. 18 Measured
PHV
fuel econom y improvement, with four trailer slots blowing.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 373/637
IMPROVING PERFORMANCE OF AUTOMOTIVE VEHICLES 371
%FEI as a function of blowing coefficient, C The %FEI improvements shown
range from 4 to
5 (5
to 6% if the 1% error band is included) above the fuel
economy of the baseline standard tractor-trailer, but these are seen to reduce
somewhat as blowing increases to values beyond
C
=
0.02-0.03.
Whereas
responses heard from typical HV users indicate this 5-6%FEI to be quite respect-
able, it is less than we had anticipated based on our smaller-scale wind-tunnel
tests.@ Figure 19
compares this on-track data to the predicted fuel efficiency
increase that we had expected from the drag reductions of the blown configur-
ations. Whereas the lower blowing values w ere nearing 70-80% of the expected
values, at greater blowing the payoff was reduced. The test team of GTRI,
Novatek, and American Trucking Associations identified suspected reasons for
this, and we then conducted an experimental test program to confirm these, as
discussed in Sec. V.
V. Updated Wind Tunnel Evaluations
A new series of wind-tunnel runs was made on the 0.065-scale PHV model.
We began with the best blown configuration from Fig.
7
(now seen as Run 205
in Fig. 20 , and then we sequentially downgraded the model by making
changes suspected of being detrimental when installed on the road-test truck. It
was the intent of this new wind-tunnel program to determine if the geometric
differences between the full-scale test vehicle and the wind-tunnel model pro-
duced the aerodynamic and fuel consumption differences discussed above.
Figure 20 shows that as the configurations approached the representation of the
24
13
FEI
12
WT
= GTR l Small-scale Wind Tunnel Tests,
TRC= Full-scale Track Test at TRC
-
(from Figure 1)
LO
9
8
7
5
4
3
2
I
0
, R u n 157, 2 Side
Slots
Only
0 0.01
0.02
0.03
0.04
0.05 0.06 0.07 0.08
0.09 0.1
0.11
Blowing Momentum Coefficient, C
@
Fig.
19
Comparison
of
wind-tunnel results to TRC track results.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 374/637
372
0.85
R. J.
ENGLAR
Comparative CDvs Ck , MTF 065
0.80
0.75
0.70
CD
0.65
0.60
0.55
0.50
0.45
0.40
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16
Momentum Coefficient, C,
Fig. 20 Updated w ind-tunnel test results: Drag change with configuration variation
and with variations in blowing.
full-scale test vehicle (Run 239), both blown and unblown drag increased. These
tests are further detailed in Refs.
14
and 15. Figure 21 compares the percentage
drag reduction resulting fro m eac h configuration ch ang e, whereas Fig. 2 2 shows
the predicted chan ge in %FEI due to each. Major problem s on the full-scale rig
were the lower surface fairing with aft facing step in front of the bottom blowing
slot and the partial gap between tractor and trailer.
A
comparison in Fig. 22 of
Run 239 (model most like the blown full-scale test vehicle) with Run
36
(most
like the standard tractor-trailer vehicle) shows that only a 2% FEI occurs for
the unblown vehicle and only
7%
for the blown one. This confirms the trends
of Figs. 18 and 19, and exp lains the causes of the less-than-expected track-test
results. We have since conducted further testing to improve the final PHV con-
figurations in anticipation of a second on-road fuel economy test at TRC. Note
from Fig. 22 that if we convert the full-scale PHV test vehicle to a blown con-
figuration much more like the one in Run 205, we can expect FEIs of 16%
unblown and 23% blown, which will be very significant results. That plan to
reconfigure the test truck and retest is now under way.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 375/637
IMPROVING PERFORMANCE OF AUTOMOTIVE VEHICLES 373
55
50
45
40
%CD
35
30
25
20
15
10
5
0
36 239 212 210 209 207 205
Test Run No.
Fig.
21
GTR I model drag reductions relative to Run 36 (configuration closest to
baseline
HV)
(positive AC, here is drag reduction).
30
2 5
20
FEI
15
10
5
0
36 239 212 210 209 207 205
Test Run No.
Fig.
22
Equivalent fuel efficiency increase (%FEI) relative to Run
36.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 376/637
374 R. J.
ENGLAR
VI. Pneumatic Sport Utility Vehicles (PSUVs)
A. Background
An analysis of vehicle fuel usage rates in the United States (Fig. 23)16,17shows
that, as of about 2001, S W s and light trucks consumed more total fuel in the
United States than either automobiles or HVs. It thus seemed quite relevant to
determine if this pneumatic technology would be as beneficial to S W s as to
HVs, perhaps even more so in the total picture. To prepare for a full-scale evalu-
ation of the pneumatic concept applied to a sports utility vehicle, we acquired the
use of the Georgia Tech FutureTruck vehicle, a 2000 Chevrolet Suburban SUV.
Preliminary wind-tunnel testing of the conventional SUV was first conducted to
determine baseline aerodynamic characteristics and flow separation point
locations (Figs. 24 and 25). The unmodified baseline GM Suburban
S W
was
installed on the six-component balance of the Lockheed 16 x 23 ft subsonic
wind tunnel in Marietta, GA. Figure 26 shows aerodynamic force and moment
variations for the unblown vehicle as functions of yaw angle, and confirms that
side winds can have a significant effect on the performance and stability of
these large SUVs (much like the HVs).
The conventional Suburban was then modified into the pneumatic SUV con-
figuration for the blowing tests. We had received an additional aft door assembly
for the Suburban, donated by the GM Technical Center. The modification was
conducted at the prototype shop of our team member Novatek, Inc. in Smyrna,
,
1 5 0 , ~
140.0
130.0 -
120.0 -
Source: Refs. 16 and 17
BGY=Billions of Gallons per
i i n n
A n n
1
20.0
w w
10
~ (Class 2b-8) &
Buses
Repon Date ~ (No
Military Vehicles)
o
0.0
L ,
, , ,
' ' ' '
,
. I
,
' I , d
1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 2020
Year
Fig.
23
Highway energy usage comparisons (billions
of
gallons per year) by vehicle
type.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 377/637
IMPROVING PERFORMANCE OF AUTOMOTIVE VEHICLES 375
Fig.
24
GM Suburban test vehicle undergoing smoke flow visualization in the
Lockheed 16
x
23 ft wind tunnel test section.
Georgia. Because it was impractical to cut away the rear sheet metal and door
structure of the Suburban, we simply added blowing plenums, slots and
turning surfaces onto the outside of the donated door. This modification installed
on the vehicle is shown in the Lockheed Low Speed Wind Tunnel in Fig. 27. The
blowing slots were adjustable and the T E je t turning angles could be changed.
Blowing coefficient was variable, and mass flows, pressures, and jet velocities
were measured to enable online calculation and setting of C
Fig. 25
Lockheed wind tunnel.
GM Suburban test vehicle undergoing tuft flow visualization in the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 378/637
376
R. J.
ENGLAR
0.8
0.7
0.6
0.5
0.4
0.3
0.2
CD,
CLfO.1
Cside
0.0
Pneumatic SUV, Lockheed LSWT T1835,10/18/02
Yaw Sweep, q=12.54 psf, V=71.7
mph,
Run 3
I
onventional
GM Suburban
0.25
0.20
0.15
0.10
0.05
-0.00
-0.1
-0.2
-0.05
-0.3
-0.10
-0.4
-0.5
-0.6
-0.7
-0.8
-0.25
-16 -12 -8 -4 0 4 8 12 16
-0.15
-0.20
Nose Right
Yaw
Angle, v, deg
ose Left
Fig. 26 Resulting aero forces and mom ents as functions of yaw angle for baseline
Suburban.
B.
PSUV
Test
Results
Flow visualizations taken with blowing activated on the pneumatic vehicle
showed significant attachment of flow over the new curved a ft surfaces and a con -
tracting of the je t wake behind the vehicle. T he wind-on, blowing-on data showed
different behaviors for the different TE configurations. Greater TE turning-
surface angle produced greater jet turning, but also greater suction on these
TEs, the latter causing an incremental drag force. The resulting total drag is
shown in Fig.
28
for four different blowing configurations. Notice that for
some configurations, initial drag reduction reached a minimum point, followed
by drag increase with higher
C
This drag reduction at lower blowing is of
the ord er of 3 to 4.15 times the input blowing coefficient, representing as much
as a 415% return on the jet momentum invested. Note that when increased
blowing yields a rise in drag for some of the configurations, this represents an
opportunity for an aerodynamic braking system. What is needed, of course, is
an onboard control system to switch f rom drag reduction to braking if requested
by the driver. Note a lso that the configuration with a 45 deg turning surface on all
exposed TEs continued to reduce drag with increased blowing, although at a
lesser rate of reduction. Also, the blowing-off drag coefficient for these
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 379/637
IMPROVING PERFORMANCE OF AUTOMOTIVE VEHICLES
377
Fig. 27 Modified PSUV blown test vehicle in the Lockheed Low Speed Wind Tunnel.
PSUV Drag Variation with Blowing, V=SOmph
Momentum Coefficient, Cb
Fig. 28 PSUV drag coefficient changes with varying C
0.44
0.46
0.48
0.50
0.52
0.54
0.56
CD
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
Momentum Coefficient, Cµ
PSUV Drag Variation with Blowing, V=50mph
Run 12, 45° Top, Bottom & Bottom Sides;90° Top Sides
Run 11, 45° Top & Bottom; 90° Sides
Run 13, 180° All 4 Sides
Run 14, 45° All Sides
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 380/637
378
R. J.
ENGLAR
nonoptimized pneumatic configurations was the same as that of the stock refer-
ence Suburban tested earlier, indicating no blowing-off drag penalty for installing
this system on a typical SUV.
An additional benefit of the blowing system is its ability to provide increased
safety of operation. Aerodynamic braking was already mentioned, but Fig. 29
shows an additional strong potential.
To
counteract the adverse effects of side
winds on both yawing and rolling moments shown in Fig. 26, we tested
blowing of only one side slot, the left side. In Fig. 26, the baseline SUV is direc-
tionally unstable (for instance, nose-left yaw produces nose-left negative yawing
moment, which tends to yaw the vehicle more), but blowing on the left side pro-
duces an aft aerodynamic side force to the left and a restoring yawing moment
that returns the SUV's stability. Figure 29 shows the amount of left-side
blowing required to eliminate the destabilizing yawing moment at each of
three side-wind angles 4. In each case, blowing at a slightly higher rate produced
yaw in the opposite direction, so that the vehicle's stability in either direction
could be controlled by varying blowing alone.
It is to be noted from the above that we have not yet achieved the optimum
configuration to maximize drag reduction and yawing moment generation
while requiring minimum blowing input, but we have otherwise verified that
blowing on SUVs can be a powerful means to reduce or increase drag as
needed, and to increase vehicle stability, all with no external moving parts.
PSUV
Aero Data, LSWT Test 1853, Run 17, Left Side Slot ONLY ,
required
to
offset v=-lOo nblown
C required
to
offset
~ = - 5 '
nblown
0 0.01
0.02 0.03 0.04 0.05 0.06
Left Slot Momentum Coefficient, Cmu
PSUV
Aero Data, LSWT Test 1853, Run 17, Left Side Slot ONLY ,
Cmu Sweep, V=50 mph, All Turning Corners are 45
0.09
. , . , . . . , . , . . . ,
Nose 1
Right
0 08
0.07
0.06
0.05
Cyaw
0 04
0.03
0.02
0.01
O
0 0.01
0.02 0.03 0.04 0.05 0.06
Left Slot Momentum Coefficient, Cmu
Fig. 29 Yawing mom ent control by b lowing the left-side slot only.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 381/637
IMPROVING PERFORMANCE OF AUTOMOTIVE VEHICLES 379
In a related application, GTRI and Novatek are also currently developing a
patented aerodynamic heat exchanger that is based on these pneumatic prin-
c i p l e ~ . ’ ~ ” ~his can be combined with the above devices to further reduce
vehicle drag by reducing the drag associated with the conventional vertical radia-
tor and related cooling system, while also adding favorable aerodynamic and
control characteristics to the vehicle.
VII.
Conclusions
Under DOE-sponsored research programs, GTRI and its teammates on the
DOE Pneumatic Heavy Vehicle project have completed experimental investi-
gations to confirm the use of pneumatic aerodynamics on these vehicles. We
have verified these capabilities by designing, fabricating, and testing small-
scale PHV m odels in three separate wind-tunnel tests, and by designing, fabricat-
ing, and conducting three road or track tests of a full-scale PHV demonstrator.
We have also conducted full-scale, wind-tunnel tests of this technology
applied to a typical
SW . It has thus been verified that these blowing concepts
can reduce aerodynamic drag, favorably modify other aerodynamic character-
istics, and thus improve the performance, stability and con trol, handling qualities,
safety of operation, and economics of both HV and SUV. The very favorable
capability of controlling all aerodynamic forces was shown for the PHV and
pneumatic SUV configurations, as was the ability of a no-external-moving-part
pneumatic control system to restore directional stability by eliminating unstable
yawing moment and providing counter-yaw in the opposite direction.
The preceding test programs and analyses have confirmed the following capa-
bilities for pneumatic aerodynamics applied to HVs or SUVs.
1) Pneumatic devices, using one to four blowing slots and nonmoving
downstream jet turning surfaces on HVs and SUVs, have reduced drag by up
to
84
in tunnel tests. This is a result of the prevention of flow separation and
increase in base pressure on the rear of the vehicle. Recent tunnel tests of a
new PHV configuration soon to be tested full-scale indicate FEI of up to approxi-
mately 23%, corresponding to about 46%
D
eduction at highway speeds.
2) 2)Specific blowing on only certain of the slots can cause a deliberate
increase in drag, which can be used for rapid-response aerodynamic braking.
3) Specific blowing on only one side slot can cause a delibe rate increase in
side force and yawing moment to overcome the directional instability of these
flat-sided vehicles caused by side winds and/or wind gusts.
4)
Blowing on only the top slot can cause a deliberate increase in lift to
reduce tire rolling resistance and thus increase fuel economy; or blowing on
only the bottom slot can deliberately increase down force and thus provide an
increase in load on the w heels to increase both traction and braking.
5 )
Because blowing can be quickly directed to whichever slot it is needed in,
these devices provide a very-rapidly-responding pneumatic control system
without external moving parts. Integrated with an onboard sensor and controls,
a pneumatic system could thus control all aerodynamic forces and moments
acting on HVs and S U V s and increase operational safety.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 382/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 383/637
IMPROVING PERFORMANCE OF AUTOMOTIVE VEHICLES 381
Acknowledgments
The author wishes to thank Sidney Diamond, Jules Routbort, and Rogelio
Sullivan of DOE for their continued support and encouragement of this work,
as well as Victor Suski for the continued very valuable involvement of the Amer-
ican Trucking Associations. The technical assistance of Ken Burdges of Novatek,
Inc., Skip Yeakel of Volvo, and Charlie Fetz of Great Dane is also greatly
appreciated, as are the experimental efforts of GTRI Co-op students Graham
Blaylock, Warren Lee, Chris Raabe, Erik Kabo, and Brian Comer from the
Georgia Tech School of Aerospace Engineering, and researchers Rob Funk
and Paul Habersham of GTRI.
References
‘Hammache, M., Michaelian, M., and Browand, F., “Aerodynamic Forces on Truck
Models, Including Two Trucks in Tandem,” Society of Automotive Engineers Paper
2002-01-0530, Feb. 2002.
’Englar, R. J., Hemmerly, R. A., Taylor, D. W., Moore, U. H., Seredinsky, V.,
Valckenaere, W. G., and Jackson, J. A., “Design of the Circulation Control Wing STOL
Demonstrator Aircraft,” AIAA Paper 79-1842, Aug. 1979.
3Englar, R. J., and Applegate, C. A., “Circulation Control-A Bibliography
of
DTNSRDC Research and Selected Outside References (Jan. 1969 to Dec. 1983),”
DTNSRDC Rept. 84/052, Sept. 1984.
4Englar, R. J., “Circulation Control Aerodynamics: Blown Force and Moment Aug-
mentation and Modification; Past, Present and Future,” AIAA Paper 2000-2541, June
2000.
Englar, R. J., Smith, M. J., Niebur, C. S., and Gregory, S. D., “Development of Pneu-
matic Aerodynamic Concepts for Control of Lift, Drag, and Moments plus Lateral/
Directional Stability of Automotive Vehicles,” Society of Automotive Engineers Paper
960673,26-29 Feb. 1996.
Englar, R. J., “Development of Pneumatic Aerodynamic Devices to Improve the
Performance, Economy and Safety of Heavy Vehicles,” Society of Automotive Engineers
Paper 2000-01-2208, 19-21 June 2000.
’Gutierrez, W. T., Hassan, B., and Rutledge, W. H., “Aerodynamics Overview of the
Ground Transportation Systems (GTS) Project for Heavy Vehicle Drag Reduction,”
Society of Automotive Engineers Paper 960906, June 1996.
‘Englar, R. J., “Advanced Aerodynamic Devices to Improve the Performance,
Economics, Handling and Safety of Heavy Vehicles,” Society of Automotive Engineers
Paper 2001-01-2072, May 2001.
Englar, R. J., “Development and Evaluation of Pneumatic Aerodynamic Devices to
Improve the Performance, Economics, Stability and Safety of Heavy Vehicles,” DOE
Quarterly Progress Rept. No. 14, April-June 2002.
“Englar, R. J., “Preliminary Results of GTR I/DOE Pneumatic Heavy Vehicle Tuning
Tests,” GTRI Rept. A-5871, March 2002.
“Englar, R. J., “Preliminary Results of the Pneumatic Heavy V ehicle SAE Type-I1 Fuel
Economy Test,” GTRI Draft Rept. Sept. 2002.
”Dotson, R., “SAEJ1321 Class-Eight Truck Aerodynamic and Tire Comparison Fuel
Economy Tests,” Transportation Research Center Report, Project 20020465, Sept. 2002.
5
9
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 384/637
382
R. J.
ENGLAR
13Englar, R. J., “Developm ent and Evaluation of Pneumatic Aerodynamic D evices to
Improve the Performance, Economics, Stability and Safety of Heavy Vehicles,” DOE
Quarterly Progress Rept. No. 15, July-Sept. 2002.
14Englar,R. J., “GTRI Updated Wind-Tunnel Investigation
of
Pneumatic Heavy V ehicle
Road-Test Configurations,” GTR I Rept. Projects A-587 1 and A-6395, Jan. 2003.
‘’Englar, R. J., “Drag Reduction, Safety Enhancement and Performance Improvem ent
for Heavy Vehicles and SUVs Using Advanced Pneumatic Aerodynamic Technology,”
2003 SAE International Truck and Bus Meeting and Exhibition, Society of Automotive
Engineers Paper 2003-01-3378, Nov. 2003.
‘6“Transportation Energy Data Book: Edition 19,” DOE/OR NL-6958, Sept. 1999.
”“EIA Annual Energy Outlook 2000,” DOE/EIA-0383 (2000), Dec. 1999; also AEO
2001.
“Gaeta, R.
G.
nglar, R. J., and Blaylock, G., “Wind Tunnel Evaluations
of
an Aero-
dynamic Heat Exchanger,” Proceedings of the
UEF
Conference The Aerodynamics of
Heavy Vehicles: Trucks, Buses and Trains, Dec. 2002.
‘’Gaeta, R. J., and Englar, R. J., “Pneumatically Augmented Aerodynamic H eat
Exchanger,” Paper #18 presented at the NASA/ONR Circulation Control Workshop,
March 2004; also to be published in Workshop Proceedings.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 385/637
Chapter 14
Aerodynamic Heat Exchanger: A Novel Approach to
Radiator Design Using Circulation Control
Richard
J.
Gaeta,* Robert J. Englar,+ and Graham Blaylock*
Georgia Institute
of
Technology, Atlanta, Georgia
Nomenclature
C, =
pressure coefficient or specific heat at constant pressure
CL=
section lift coefficient
CD=
section drag coefficient
C ,
=
momentum coefficient
m = coolant mass flow rate, gal/min
q =
freestream dynamic pressure, psf
Q =
heat energy rejected by coolant, kW
s
= wing reference area, ft2
Tci,=
inlet coolant temperature,
O F
T,,,,
=
outlet coolant temperature,
O F
V , =
freestream velocity, mph
a
=
angle of attack
I. Introduction
A. Motivation
ROPER aerodynamic design of automotive vehicles lead to improved fuel
P
fficiency. This usually means that aerodynamic drag, both profile drag
and friction drag, are minimized. Design strategies for low profile drag include
*Senior Research Engineer, Aerospace and Acoustics Technologies Branch, ATAS, Georgia Tech
'Principal Research Engineer, Aerospace and Acoustics Technologies Branch, ATAS, Georgia
*Undergraduate Student, School of Aerospace Engineering.
Copyright 005 by the authors. Published by the American Institute of Aeronautics and
Research Institute. Associate Member AIAA.
Tech Research Institute. Associate Member AIAA.
Astronautics, Inc., with permission.
383
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 386/637
384
R. J.
GAETA,
R. J.
ENGLAR, AND G. BLAYLOCK
establishing small cross-sectional areas and gentle transitions from the front to
the rear of the vehicle. There is a long history of streamlining automotive
vehicles, which began in earnest in the
1920s
and continues to the present.'
The majority of the work in this area has been aimed at reducing the profile
drag.' In general, vehicle size, and thus its frontal area, is largely determined
by other requirements, such as eng ine size and passenger room.
Heat exchangers used in these veh icles are, typically, finned radiators that are
positioned in the engine compartment away from the freestream flow, which is
rammed into the vehicle body through a duct or an open area at the front. The
radiator must pass a sufficient amount of a ir through its core to remove engine
heat. However, its location and how the vehicle is shaped around are largely gov-
erned by the engine placement and the vehicle styling. A conventional, state-of-
the-art radiator is installed perpendicular to a freestream flow and employs part of
the freestream total pressure to provide a pressure drop that aids the heat transfer
across the radiator. Unfortunately, this also produces large aerodynamic drag
coefficients on the vehicle. The frontal area needed for engine cooling air flow
varies in the extreme from heavy vehicles like tractor-trailer rigs to high-perform-
ance race cars (Fig. 1). Profile drag could potentially be reduced by allowing the
radiator to assume a smaller frontal cross-sectional area relative to the oncom ing
flow. A novel approach to this problem is the aerodynamic heat exchanger (AH E)
concep t, which starts with the notion of housing the heat exchanger in a low-drag
package: a wing.
B.
Aerodynamic Heat Exchanger (AHE) Concept
All current liquid-cooled internal combustion engines used in automotive
vehicles use heat exchangers that rely on stagnation pressure drop across a
porous flat plate. Air flow from the freestream is directed either through a duct
or through louvers to reach the face of the heat exchanger. This pressure differ-
ence is large, but it occurs at the price of a large drag force. A wing is a device
that naturally produces a pressure difference but in a way that produces a low
drag force (Fig. 2). The pressure difference produced by a wing i s not as great
as that produced by the stagnation flow across a conventional radiator, but by
employing established pneumatic flow control techniques, the wing lift (or
Fig.
1
Radiators for large heavy vehicles and passenger cars present large frontal
area to oncoming flow thus contributing to profile drag. High-performance race
cars use pods or ram scoops that are necessary tradeoffs for low profile drag designs.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 387/637
AERODYNAMIC HEAT EXCHANGER USING CC
385
Flat
AP
at
expense of
late
high drag
Wing
P with
low drag
Fig. 2
A
wing has an order of m agnitude lower drag coefficient than a flat plate and
has a mechanism to produce a pressure differential needed for heat transfer.
pressure difference) can be radically augmented. Figure
3
shows the A H E
concept in conjunction with a blown elliptical airfoil, which can generate
suction pressure coefficients of AC, = - 5 to -6 across the center compared
to AC, =
+0.4
to 0.5 across a standard radiator core. This concept is embodied
in a patent that was granted to G TR I and Novatek, Inc. in
2000
and it involves the
use of a very effective high-lift airfoil section to generate the pressure d ifferences
needed across a conventional automotive radiator.’ As the natural pressure differ-
ence is formed by flow over the wing, the difference in static pressure forces air
through the porous thickness of the wing. The greater the lift or pressure differ-
ence, the greater the flow through the wing.
The A H E device embeds the radiator core within an airfoil shape aligned
parallel to the wind flow. Blown airfoils can generate suction rises on the order
of 10 to
15
times the conventional radiator pressure drops (typically 40-50
of freestream total pressure). Thus the potential exists to enhance engine
cooling and reduce aerodynamic drag. This pneumatic-based lift augmentation
concept, also known as circulation control (CC), is based on the physics of the
Coanda e f f e ~ t . ~his effect postulates that a fluid stays attached to a curved
i t o R e s w e
A
Porous
Heat
Conducting
Material
Fig.
3
AHE concept with pneumatic lift augmentation control. Pressure difference
arising from lift provides air to the heat exchanger in a low drag envelope.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 388/637
386
R. J.
GAETA,
R. J.
ENGLAR, AND G. BLAYLOCK
surface as it flows over it by virtue of a balance between the pressure gradient
normal to the surface and the centrifugal force caused by the streamline curva-
ture. A significant amount of study of this effect and how
it
can augment the
lift of a wing can be found in the literature for both fixed and rotating
Figure 4 shows how the lift on a symmetric airfoil can be controlled
by changing the momentum of air blown through a thin slot at the trailing edge
(TE) of the wing. The design challenge for the AHE is to provide enough flow
through the wing (via porosity), but still retain a high enough pressure difference
to create aerodynamic force, if needed. This novel heat exchanger concep t has the
potential to become a dual-use system for automotive vehicles by providing both
aerodynamic advantages and cooling simultaneously.
A proof-of-concept model of the AHE was built and tested in a low-speed
wind tunnel at GTRI to evaluate its feasibility. Several radiator core configur-
ations were tested for their aerodynamic and heat-exchanger performance. The
technical approach and results of this testing are presented in Secs. I1 and
111.
11.
Technical App roach
A. Facilities and Experimental Setup
The testing of the AHE model was performed in GTRI’s Model Test Facility.
This facility is a closed-return wind tunnel with an operating dynamic pressure
range of
5
to
50
psf. The flow is conditioned upstream of the test section with
High Veloc i ty
Jet Sheet
Increasing Mom entum Ratio, C,
Increases Velocity on upp er sur face,
thu s increases l i ft [see Coanda Effect,
Circulat ion]
Circulat ion Control produ ces up to
8000%
increase in CL relative
to m omentum force input
Phenomenal
Inc rease in P
Fig.
4
Controllable lift, thus controllable heat transfer, with pneumatic flow control.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 389/637
AERODYNAMIC HEAT EXCHANGER USING CC
387
honeycomb screens and the nominal freestream turbulence is approximately
1.0 .
The test section is 30 x 30 in. and aerodynamic forces are measured
with a six-component balance attached to a turntable for easy changes of the
model’s angle of attack. The airfoil shape chosen for the AHE concept was ellip-
tical, with a round TE , similar to that shown in Fig. 4. The airfoil has a plenum
near the TE that can be pressurized with air to produce a variable am oun t of flow
through a TE slot for increased circulation or lift. Internal pressure transducers
and flow meters measured blowing air mass flows, velocities, and blowing co-
efficients. Static pressure taps were located on the pressure and suction side of
the airfoil at approximately midchord. Data from these taps were used to
compute an average ACp. The baseline configuration used a nonporous center
section to generate a reference for aerodynamic performance. Three AHE radia-
tor configurations were tested by fabricating a two-dimensional airfoil with a
reconfigurable center section along with three porous center sections.
The elliptical wing with the radiator core was installed vertically in the wind
tunnel and was attached to the force balance. The airfoil was connected via flex-
ible hoses to a three-phase electric
3600
W water heater. Water was heated and
pumped into one side of the wing and in to an inlet reservoir attached to the radia-
tor. After the water made it s way through the radiator core, it exited into an ou tlet
reservoir and into the water heater, closing the coolant loop. Figure
5
shows a
schematic of the coolant flow path. Coolant mass flow was measured with a
water flow meter and thermocouples were placed in both inlet and exit reservoirs
to monitor the temperature drop across the core. The coolant mass flow and temp-
Inlet Reservoir Out let Reservoir
Fig.
5
Schematic of A H E experim ent showing the coolant path through the test airfoil.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 390/637
3aa
R. J.
GAETA,
R. J.
ENGLAR, AND G. BLAYLOCK
eratures were acquired using a LabView program, where a simple heat balance
was used to quantify heat rejection of the coolant to the air passing through the
airfoil into the tunnel. The heat transferred from the coolant can be expressed as
A typical run for a given radiator configuration would include a “sweep” of slot
blowing pressure at constant angle of attack and tunnel speed, to record and evalu-
ate aerodynamic characteristics. Then, for each radiator core installed, the coolant
lines were added (these would have caused balance tares during the aero runs) and
temperature data were taken at constant coolant flow rates for variable blowing
pressures. Variation in tunnel speeds was also conducted for the radiator airfoils
at constant flow rates while varying blowing pressures. For reference, the conven-
tional Visteon radiator was evaluated without blowing or airfoil frame but perpen-
dicular to the freestream flow
so
as to simulate a standard radiator’s cooling
characteristics. All aerodynamic characteristics are based on a wing planform
area of 2.871 ft2 and the blowing momentum coefficient is defined as
riZV,
c
=
s
It should be noted that the model airfoil is mounted inverted in the tunnel, with
negative lift (positive down force) towards the ground as the lifting side of the
airfoil is towards the road, and negative angle of attack a s LE downward.
B. Test Articles
All radiator test articles were made to fit in the center section of the two-
dimensional elliptical wing. The nominal dimensions of the radiators are
8 x 13 x 1.42 in. In addition to a solid wing configuration, the radiator types
of Secs. 1I.B.1-1I.B.3 were tested.
1.
Conventional Aluminum Fin Core
This core was a conventional aluminum finned core used in a Formula SAE
race car operated by the Georgia Tech Motorsports Club. It had relatively low
pressure drop or a high porosity and was produced by the Visteon Company.
Each radiator core had cooling tubing passing through internal channels or
through the foam core. Sealing of coolant leaks was a significant problem for
the Visteon core. Note in Fig.
6 ,
the coolant channels marked in red stripes
were not able to be sealed within the core, and thus were taped over with
metal tape to prevent leakage into the airfoil. Thus the Visteon radiator shown
was tested with only 10.5 of its 18 passages open, o r only 58% flow capacity.
2.
ORNL supplied a radiator that had the same planform dimensions and was
in the same envelope as the Visteon radiator, but was made from solid pieces
of carbon-graphite foam material. This material has phenomenal heat conduc-
ORNL
Very Dense Graphite Foam Core
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 391/637
AERODYNAMIC HEAT EXCHANGER USING CC
389
Fig. 6 Conventional aluminum finned radiator core made by Visteon, showing
coolant passage in lets.
tivity properties. Although it is porous, the bulk density is such that it has a sig-
nificant pressure drop. Brass tubes were press fit into the foam to carry the coolant
through the material for heat exchan ge, as shown in Fig. 7.
3 .
A second
ORNL
supplied radiator core consisted of smaller carbon-graphite
foam fins arranged in such a way that flow could follow the serpentine, a s shown
in Fig. 8. The se were brazed to narrow water channels in a ma nner similar to the
aluminum radiator. The manufacturing of this core was such that some of the
coo lant passages were blocked off,
so
its full heat rejection potential was not rea-
lized. Furthermore, it was made about half an inch thinner than the thickness of
the wing, so a perforated sheet had to cover the wing to maintain smooth flow.
Figure 9 shows two different AHE radiator core configurations installed in the
wind-tunnel test section.
ORNL
PorouslSerpentine Graphite
Foam
Core
111. Results
A. Note on Measurement Uncertainties
The aerodynamic data presented in this chapter were acquired from a six-
component force balance that supported the test article in the wind tunnel. The
accuracy of the load cells are approximately
1
of the reading. All pressure
measurements were acquired with piezoresistive gauges with .5% reading
accuracy. Omega thermocouples were used that were accurate to within a
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 392/637
390
R. J.
GAETA,
R. J.
ENGLAR, AND G. BLAYLOCK
Fig. 7 Dense ORNL carbon-graphite foam radiator core.
deg ree Fahrenheit. Perhaps the largest uncertainty was in the water mass flow rate
measurement.
An
Om ega flow meter using a turbine wheel was used to obtain the
liquid flow rate. Th e manufacturer specified the m eter to have a 0.5% of reading
accuracy. In practice, the repeatability of our data acquisition system was
approximately
10
of reading accuracy. This accounts for most of the observed
scatter in the heat transfer results presented. Weighted curve fits are used to
signify trends in the data.
Section
A-A
Fig.
8
A more porous carbon-graphite radiator core.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 393/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 394/637
392
R.
J. GAETA,
R.
J. ENGLAR, AND G. BLAYLOCK
MTF059 Pressure Increment Across Airfoil Radiator,
q=5 psf, a=O
Conventional Rad iator at
c(=90
0.5
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Cmu
Fig. 10 Average pressure coefficient as a function of blowing momentum for various
AHE configurations.
around the airfoil causes the LE to separate and thus the discontinuity in the lift
curves to occur. This can be corrected by improving the LE shape. There is still
improvement to be realized: the
20
elliptic airfoil of Ref.
5
is a thicker airfoil
(i.e., has a g reater LE radius) version of the current baseline blown ellipse airfoil,
and it show s no sign of separation, reaching a C of or more. Thus great down
force potential is confirmed with blowing (no increase in airfoil angle of attack is
necessary) and this w ill impact the heat transfer potential. The aerodynamic test
also confirms that flow through the radiator core can be varied by controlling the
circulation with trailing edge (TE) blowing.
C. Heat Transfer Results
Results for the conventional radiator core indicated that a maximum coolant
temperature drop of about 5 F was possible for a flow rate of 5 gal/min with
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 395/637
AERODYNAMIC HEAT EXCHANGER USING CC
393
0.5
0.0
-0.5
-1.0
-1.5
-2.0
CL
-2.5
-3.0
-3.5
-4.0
-4.5
-5.0
-5.5
0.00 0.10 0.20
0.30 0.40
0.50
CD
Fig. 11 Drag polar as a function of blowing momentum for various AHE
configurations.
a 6 4 mph freestream velocity. Figure 12 show s coo lant temperature drop for the
Visteon core as a function of blowing coefficient C and coolant mass flow. Note
that for the smaller coolant mass flows, larger temperature drops
are
observed.
This is quite likely caused by the longer exposure of the coolant to the heat
exchanger (longer residence times). It should also be noted that because of fab-
rication anom alies, some (42%) of the coolant flow tubes were blocked off so the
Visteon radiator was not flowing in a evenly distributed manner and it is likely
that its performance was inhibited to some degree (see Fig. 6).
Figure 13 shows the AHE heat removal as a function of C and radiator con-
figuration at a nominal coolant flow rate of 5 gal/min. As expected, the effect
of the pneumatic lift augmentation (the increasing
C
is to increase the heat
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 396/637
394
R. J.
GAETA,
R. J.
ENGLAR, AND G. BLAYLOCK
0.0
-1
.o
-2.0
-3.0
4.0
=
-5.0
6
-
6.0
-7.0
-8.0
-9.0
-1
0
0 0.05 0.1
0.15
0.2
0.25
0.3
Fig.
12
Influence
of
coolant flow rate on coolant temperature differen ce; Visteon
core, V , = 64 mph.
removal rate via increasing air flow through the w ing. A notable exception is the
high-density carbon-graphite configuration.
Figure 14shows the heat removal for the various A HE radiator configurations
at a coolant flow rate and freestream velocity that are close to nominal automotive
vehicle values. It is interesting that the high-density graphite core performs as
well as the Visteon core, which is somewhat surprising because it has little or
no airflow through the core. This performance i s likely because of the superior
conductive performance of the foam; that is, almost all of the heat transfer
takes place in the form of forced convection along the surface of the airfoil
(both upper and low er; see Fig. 7). This result was intriguing and suggests that
the heat removal can be varied and/or augmented by simply varying the turbu-
lence level of the flow over the wing surface. There are many flow control
methods (active and passive) that can aid this forced convection process. The
high-density graphite radiator core was also the best aerodynamic performer.
This makes this configuration all that much more attractive, because the AHE
can function as an effective aerodynamic and heat transfer device.
For comparison, a typical passenger automobile radiator removes about 15
20 kW in normal operation for a full-sized engine. The model AH E tested here
produced roughly half of this heat rejection, but with a radiator core of less
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 397/637
AERODYNAMIC HEAT EXCHANGER USING CC
395
1
9.0
8.0
- 7.0
6.0
8
5.0
m
-
4.0
3.0
t
8
2.0
1.o
0.0
0.2 0.4
0.6 0.8
1
CP
Fig.
13
Rejected heat from three different AHE configurations; V , =
32
mph,
coolant mass
flow
=
5
gal/min.
than half the area. When one accounts for the heat removed per square foot of
radiator, it can be shown that the A HE does the heat transfer jo b the conventional
radiator does with approximately three times less aerodynamic drag.
IV. Conclusions
Initial wind-tunnel evaluations of the aerodynamic heat exchanger concept
employing both conventional and ORNL graphite foam radiator cores have
been performed. This new concept has been shown to adequately transfer heat
at the same or similar rates as convectional radiators at 90-deg to the flow, but
at much lower drag coefficients when enclosed in a lifting surface parallel to
the flowfield. The dense graphite foam core of
ORNL
has been shown to be
both an effective heat transfer medium, employing forced convection and an
excellent aerodynamic surface, and allowing almost no air to pass through the
wing.
The following conclusions can be drawn from this proof-of-concept test of the
AHE:
1)
An aerodynamic heat exchanger (AHE) with pneumatic lift control was
successfully tested in a wind tunnel and the basic concept was validated.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 398/637
396
R. J.
GAETA,
R. J.
ENGLAR, AND G. BLAYLOCK
Fig. 14 Rejected heat from three different AHE configurations; V ,
=
64 mph,
coolant mass flow = 15 gal/min.
Fig. 15 AHE installation into
GT
Formula SAE race car.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 399/637
AERODYNAMIC HEAT EXCHANGER USING CC
397
2) Lift and drag are dramatically affected by the porosity of the radiator core
section, but pneumatic augmentation is still a powerful control.
3) The AHE demonstrated nonoptimized heat rejection performance, but
optimized sizing should further improve results.
4) The AH E has great potential for exhibiting both controllable aerodynamic
force and low drag penalty for eng ine cooling.
5 Carbon-graphite foam enables optimal performance of the radiator core
within the AH E concept.
It is important to note that system integration issues will pose a (surmoun-
table) challenge to designers of cooling systems. Two important issues that
need to be addressed are the production of steady high-pressure air for the
pneumatic system and coolant pump size and ducting for the AHE. It is recog-
nized that any fuel savings obtained from a lower drag configuration will be
offset somewhat by the energy needed to produce the circulation control
blowing air. It is the plan of GTRI to demonstrate this technology on the
GT Motorsports Formula
SAE
race car as a technology demonstrator. Initial
work has highlighted the need for good system integration design. Figure 15
shows one of the Formula SAE student cars with the AHE model being
prepared for installation.
Acknowledgments
The authors would like to thank Jam es Klett and April McM illan of ORNL for
being receptive to the concept of the AHE and providing funds and material for a
part of this work.
References
‘Hucho, W. (ed.), “Aerodynamics of Road Vehicles,” Butterworth-Heinemann,
London, 1990, Chaps. 1, 3-9.
’Burdges,
K. P.,
and Englar, R. J., “Vehicle Heat Exchangers to Augm ent Modify
Aerodynamic Forces,” U.S. Patent No. 6,179,077, Aug.
2000.
3Metral, A. R., “O n the Phenomenon of Fluid Veins and The ir Application, the Coanda
Effect”,
F
Translation, F-TS-786-RE, 1939.
4Cheeseman,
I.
C., and Seed, A. R., “The Application of Circulation Control by
Blowing to H elicopter Rotors,”
Journal ofthe Royal Aeronautical Society,
Vol. 71 , July
1966.
’Williams, R. M., and How e, H. J., “Two-Dimensional Subsonic Wind Tunnel Tests on
a
20
hick,
5
Cambered Circulation Control Airfoil,” NSRDC TN AL-176, Aug. 1970.
Wilkerson, J. B., Reader,
K.
R., and Linck, D. W., “The Application of Circulation
Control Aerodynamics to a Helicopter Rotor Model,” American Helicopter Society
Paper AHS-704, May 1973.
’Englar, R. J., “Experimental Investigation of the High Velocity Coanda Wall Jet
Applied to Bluff Trailing Edge Circulation Control Airfoils,” M.S. Thesis, Dept. of Aero-
space Engineering, U niv. of Maryland, College Park, MD , June 1973.
8Wilkerson, J. B., Barnes, D. R., and Bill, R. A., “The Circulation Control Rotor Flight
Demonstrator Test P rogram,” American Helicopter Society Paper AH S-795 1, May 1979.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 400/637
398
R. J.
GAETA,
R. J.
ENGLAR, AND G. BLAYLOCK
’Pugliese, A. J., and Englar,
R.
J., “Flight Testing the Circulation Control Wing,” AIAA
Paper 79-1791, Aug. 1979.
“Englar, R. J., “Circulation Control Pneumatic Aerodynamics: Blow Force and
Moment Augmentation and Modification; Past, Present, and Future,” AIAA Fluids 2000
Conference, AIAA Paper 2000-2541, June 2000.
“Klett, J., Ott, R., and McMillian, A. “Heat Exchanger for Heavy Vehicles Utilizing
High Thermal Conductivity Graphite Foams,” Society of Automotive Engineers Paper
2000-01-2207, Washington, DC, June 2000.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 401/637
1II.A. Tools for Predicting Circulation Control
Performance:
NCCR
1510
Airfoil Test Case
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 402/637
Chapter 15
Investigation of Turbulent Coanda Wall Jets Using
DNS
and
RANS
Hermann
F.
Fasel,* Andreas Grosst, and Stefan Wernz’
University of Arizona, Tucson, Arizona
Nomenclature
A =
area per unit span, m
B
=
blowing ratio
=
nozzle height, m
c
= chord length, m
cp=
wall pressure coefficient
c p
=jet momentum coefficient
d = cylinder diameter, m
f = frequency, Hz
k = number of spanwise Fourier mode
L
=
domain size, m
M =
Mach number
R
=
gas constant, J/(kg
K)
Re
= Reynolds number
T = temperature, K
p = pressure, kPa
r = radius of curvature, m
v = velocity, m/ s
riz
=
mass flux per unit span, kg/(m s)
x
=
streamwise location (from leading edge), m
y
=
wall-normal location (from chord), m
y2
=
wall-jet half-thickness, m
= spanwise location, m
a
=
angle of attack, deg
*Professor, Department of Aerospace and Mechanical Engineering. Member A I M .
‘Research Associate, Department of Aerospace and Mechanical Engineering. Member AIAA.
Copyright 005 by
the
authors. Published by the American Institute of Aeronautics and Astro-
nautics, Inc., with permission.
40
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 403/637
402
H F. FASEL, A. GROSS, AND S.WERNZ
r
=
circulation, m2/s
y
= ratio of specific heats
6* = displacement thickness, m
8
=
mom entum thickness, m
h = wavelength, m
p =
molecular viscosity, k3/(m
s)
v
=
kinematic viscosit m
s
v
= eddy viscosity, m / s
p
= density, kg/m3
6 = streamwise (azimuthal) angle, deg
w =
vorticity, l/s
Y, /
Subscripts
in
=
inflow inside plenum
jet = nozzle exit
max = wall-normal maximum
wall
=
wall, surface
= span wise direction
6 = streamwise direction
03 = free stream
Superscript
=
wall coordinates
I. Introduction
AL L jets over curved surfaces have great potential for technical appli-
cations. Coanda wall jets over convex surfaces can effectively provide
aerodynamic side forces or change the circulation of an airfoil. An existing appli-
cation is the “No-Tail-Rotor’’ (NOTAR) helicopter. Possible future applications
are the enhancement of low-speed maneuverability of underwater vehicles or
high-lift wings for short take off and landing (STOL) aircraft. However,
without profound understanding of the mechanisms that keep the wall jet
attached to the surface for large downstream distances, any implementation of
Coanda flow technology must rely on empiricism and hence requires excessive
safety margins to account for unknowns. In this paper, results from numerical
investigations of two separate Co anda flow experim ents are presented that may
help to shed som e light on the relevant physical mechanisms.
On e of the most intriguing phen omen a of the C oanda wall jet is the com pe-
tition/interaction of naturally occurring streamwise and spanwise vortical struc-
tures, which are a consequence of a centrifugal, Gortler-type instability (leading
to streamwise coherent structures) and a Kelvin-Helmholtz-type instability
(leading to spanwise coherent structures), respectively. It can be conjectured
that the intensity of these structures, both absolute and relative to each other,
will significantly influence the separation location and, as a consequence, will
have a key effect on the side forces that can be generated and thus on the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 404/637
TURBULENT COANDA WALL JETS AND DNS AND RANS
403
effectiveness and reliability of this technique. T he amp litud es and wavelengths of
the coherent structures will also determine the intensity and the frequency spec-
trum of the associated aerodynamic/hydrodynamic noise. In addition, because
both the streamwise and the spanwise structures are a consequence of hydrodyn-
amic instabilities, instability mechanisms may be exploited advantageously for
active flow control (AF C) strategies.
Tan i was among the first to report on streamwise vortices in a turbulent bound-
ary layer along a concav e wall.’ In his experim ents he observed regularly space d
spanwise modulations of the velocity profiles, which he attributed to a Gortler
instability mechanism. T o com pare with stability theory results for a laminar
boundary layer, he assum ed a constant eddy viscosity v r
=
0 . 0 1 8 ~ ~*,and a dis-
place ment thickness, S*= 1.38( 8 s the momentum thickness). Moser and Mo in2
performed direct numerical simulations (DNS) of a curved turbulent channel flow
to determine the effects of curvatu re in wall-bounded turbulent flows. They found
stationary Gortler vortices, which had a significant impact on the mean Reynolds
shear stresses and which enhanced the asymmetry of the channel flow. Suffi-
ciently close to the wall, the mean velocity profiles followed the law of the
wall. For a curved wall with curvature S * / r= 0.1 the turbulence intensities
and shear stresses were, in som e cases, twice as large as for a plane wall.
In the Reynolds-averaged Navier-Stokes (RA NS ) calculations considered in
this chapter , the prediction of the spreading rate dep ends on the turbulence m odel
employed. Pajayakrit and Kind3 used the Baldwin-Lomax, the Dash et al.
K - E ,
the Wilcox K- -w , and the Wilcox multiscale turbulence mo dels for the calculation
of plane and curved turbulent wall jets. They tuned the model constants to obtain
better agreement with experimental data for the streamwise development of the
skin friction and the half-thickness of the jet. They also pointed out that the
Boussinesq approximation mandates zero shear stress at the velocity peak,
although it is well known that the zero shear stress location in wall jets occurs
substantially closer to the wall. For the cu rved w all je t the nondimensional vel-
ocity profile predicted by the K - E model matched the experimental profile
whereas the profile predicted by the
K - -W
model had the velocity maximum
slightly closer to the wall.
11. Investigated Configurations
At first, in collaboration with an experimental effort by Wygnanski and co-
workers4 a turbulent wall je t on a circular cylinder was investigated. For this con-
figuration extensive numerical simulations, including DNS, large eddy
simulation (LES), and unstead Reynolds-averaged Navier-Stokes (URA NS)
calculations were conducted?” Th e flow parameters in the simulations were
chosen to match the ex eriment, with cylinder diameter
d
= 0.2032 m, nozzle
height
=
2.34
x
10- m, and jet-ex it velocity vjet
=
48 m/ s. Th e Reynolds
number based on jet-exit velocity and cylinder diameter was
Re
= 6.15
x
lo5
Reb= 7,080 based on jet-ex it velocity and nozzle height). The experiment was
conducted in a quiescent environment.
Secondly, the flow around a NCCR 1510-7067 N circulation control airfoil
was co mp uted using steady RAN S. Th is flow configuration was posed as a bench-
mark problem to the CFD community for the 2004 NASA/ONR Circulation
Control Workshop.’ Experiments by Abramson* on this particular airfoil
B
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 405/637
404
H F. FASEL, A. GROSS, AND S.WERNZ
Table 1 Elliptic airfoil
CFD
challenge cases
Case
283
Case
321
a, eg 0 - 8
h j e t , kg/ms 0.196 0.182
CLL
0.209 0.184
served as a reference for the numerical simulations. The airfoil has 15% relative
thickness (maximum thickness to chord length
c
and a Coanda trailing edge
(TE). A blowing slot is located at x/c = 0.967 with slot height b = 0 . 0 0 3 ~ .
The tests were conducted at a freestream Mach number M
=
0.12 for various
angles of attack a.This flow configuration was computed by Slomski et al.'
using the commercial flow softw are Fluent on computational grids with approxi-
mately 1.6
x
lo5points. Computations with the standard and the realizable K - - E
turbulence model only yielded realistic results for the jet momentum coefficient
c p
=
0.026. The jet mom entum coefficient was defined as
with jet-exit velocity vjet, jet-mass flux hjet= pjetvjetb, nd freestream dynamic
pressure 1/2p,vk. For a higher mom entum coefficient, c p
=
0.093, the same
two turbulence models predicted the wall-jet separation slightly farther down-
stream than observed in the experiment. At the even higher momentum coeffi-
cient, c p
= 0.209, the jet wrapped around the entire elliptic airfoil 1.5 times
when the realizable K- -E model was used. Only the full Reynolds stress model
predicted the correct separation locations and hence the correct overall circula-
tion for all momentum coefficients studied. In general, turbulence models
based on the Boussinesq approximation predicted separation too far downstream.
Another simulation for the same flow geometry was carried out by Paterson and
Baker. They studied the two workshop CFD challenge cases (Table 1) using
the incompressible CFDSHOP-IOWA code. For two-dimensional RANS, the
two-equation shear stress transport (SST) turbulence model by Menter was
employed. The predictions of wall-jet separation location and wall-pressure dis-
tribution (and therefore circulation) were in good agreement with the experiment.*
111. Numerical Approach
For the computational results presented in this paper two different numerical
approaches were taken. Each of these approaches is tailored and optimized for
certain subtasks, so computational resources can be focused effectively. When com -
bined, they will help in understanding the different physical mechanisms involved.
A. Direct Num erical Simulations D NS)
was adopted to allow for
highly accurate DNS of turbulent Coanda wall jets for Reynolds numbers in the
An existing incompressible Navier-Stokes
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 406/637
TURBULENT COANDA WALL JETS AND DNS AND RANS
405
range of the laboratory experiments by Wygnanski and coworkers. In this code
the incompressible Navier-Stokes equations in vorticity-velocity formulation
are solved. The governing equations are discretized using fourth-order accurate
compact differences in the streamw ise and wall-normal directions in combination
with a sixth-order compact filter for filtering out disturbances at grid level. The
spanwise direction is assumed to be periodic and is discretized using a pseudo-
spectral decomposition into Fourier cosine or sine series.13 This expansion
reduces the number of spanwise modes by a factor of two when compared
with the full Fourier transform. However, spanwise symmetry is imposed in
addition to periodicity. Metric terms were included to allow for computations
on orthogonal curvilinear grids. The velocity Poisson equations are solved
using an iterative solver with multigrid acceleration.
B.
Reynolds-averaged Navier-Stokes RANS) calculations
A multidomain, compressible, finite-volume Navier-Stokes code with high-
order accurate upwind schemes was developed to allow for robust com putations
of complex geometries. The convective terms are discretized with fifth-order
upwind schemes based on a weighted essentially nonoscillatory (WENO) extra-
polation and the Roe schem e,14 and the v iscous terms are fourth-order accurate.
A second-order accurate Adams-Moulton method is used for time integration.
Various turbulence models were implemented. The standard 1988 and 1998
K-6.1 models and the K- -E model15 can be combined with both a Reynolds
stress based on the Boussinesq approximation, and an explicit algebraic stress
model (EASM).16 The Menter SST and Spalart-Allmaras models were
included as well.
IV. Turbulent Wall Jet on a Circular Cylinder
A. Direct Numerical Simulations DNS)
The objective of our DNS on a segment of the Coanda cylinder from the
experiments4 was to investigate the deve lopm ent of coherent structures in the tur-
bulent flow upstream of separation and the impact of forcing on these structures
and on the mean flow development.
An illustration of the computational dom ain is provided in Fig
1.
At the inflow
boundary
6
=
-5.6
deg), a laminar Glauert wall jet with maximum velocity
v ~ , ~ ~50
m/s and momentum thickness
8
= 3 mm is prescribed
Re0
=
10,000). The laminar flow is transitioned to turbulence at 6
=
0 deg
using a volume forcing technique by which a time-dependent local force field is
applied inside the flow through forcing term s added to the right-hand side of the
Navier-Stokes eq ua tio ns 5 For actively forcing the wall je t to enhance spanwise
or streamwise coherent structures inside the flow, additional time-harmonic or
steady volume forcing is applied at 6 =
0
deg. Inside a buffer domain near the
outflow boundary the turbulent flow is relaminarized to prevent reflections of tur-
bulent fluctuations from the 0u tfl0w .l~Also shown in Fig. 1 is the computational
grid for the present simulations. In the azimuthal direction, 573 points with
constant step-size A 6 = 0.28 deg are used Ax+
RZ
50 in wall-coordinates). An
additional 100 points on a stretched grid are placed inside the buffer domain. In
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 407/637
406
H F. FASEL, A. GROSS, AND S.WERNZ
x [mml
Fig. 1 Com putational grid used for DNS.
the radial direction,
193
points are clustered toward the surface such that, through-
out the domain of interest, the wall-next points are located within y+< 1 from the
surface. The spanwise direction is discretized with 21 modes over a domain of
width
L,
= 20 mm
0.
Id , resulting in Az = 20 between collocation points.
Evidence from experiments4 and from earlier numerical investigations5
suggests that both spanwise and streamwise coherent structures are present in
the turbulent Coanda wall jet . The streamwise structures develop as a result of
a centrifugal, Gortler-type instability while the spanwise structures originate
from an inviscid, Kelvin-Helmholtz-type instability (inflection point of velocity
profile). It may be conjectured that in the natural (unforced) turbulent Coanda
wall je t (under “clean” experimental conditions) the two instability mechanisms
balance each other. For example, the Gortler-type, centrifugal instability and the
resulting Gortler vortices may inhibit the spatial growth of the spanw ise coheren t
structures that result from the inflectional instability.
T o probe this conjecture, DNS were performed, w here deliberate forcing was
introduced to enhance certain structures, or where the simulations were set up
such that certain instability mechanisms were weakened. Three DNS cases will
now be discussed. In the “unforced” case, which serves as a reference, the
flow is transitioned without applying additional forcing. In the second case, the
spanwise rollers are enhanced using time-harmonic volume forcing (frequency
f =
340
Hz) that is two-dimensional, that is, without modulation in the spanwise
direction. In the third case, streamwise vortices w ith a fixed spanwise wavelength
are generated by steady volume forcing with a periodic modulation in the span-
wise direction A, = 20 mm ).
The downstream development of the streamwise structures for the three simu-
lation cases is visualized with the iso-surface plots in Fig. 2 of the time-averaged
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 408/637
TURBULENT COANDA WALL JETS AND DNS AND RANS
407
b) C)
Fig.
2
DNS of tur bu len t Coanda wall je t, showing iso-surface plots of time-averaged
streamwise vorticity (light-shaded surfaces, clockwise rotation; dark-shaded
surfaces, counter-clockwise rotation): a “Unforced” reference; b harmonic two-
dimensional forcing; c) steady three-dimensional forcing.
streamwise vorticity, we. The counter-rotating streamwise vortices are
represented as light- and dark-shaded iso-surfaces w q = 300/s and
wq
=
00/s, respectively).
An
impression of the spanwise vortical structures
is provided with the snapshots in Fig. 3 showing gray scales of instantaneous
spanwise vorticity w, averaged in the spanwise direction. When time-harmonic
two-dimensional forcing is applied, the intensity of the spanwise coherent struc-
tures is strongly enhanced, as seen from a comparison of Figs. 3a and 3b.
However, forcing of the spanwise structures leads to an increase in intensity of
the streamwise vortices, not a decrease as may have been expected (compare
a)
Fig. 3 DNS of turb ulent Coand a w all je t showing instantaneous spanwise vorticity,
spanwise average: a) “Unforced” reference; b) harmonic two-dimensional forcing;
c) steady three-dimensional forcing.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 409/637
408
H F. FASEL, A. GROSS, AND S.WERNZ
Figs. 2a and 2b). It is possible that the strongly enhanced spanwise structures in
Fig. 3b promote the generation of streamwise vortices through a secondary
instability (braid vortices), a conjecture that requires further exploration. On
the othe r hand, by forcing the streamwise structures, the intensity of the naturally
occurring Gortler vortices is significantly increased (compare Figs. 2a and 2c),
whereas the intensity of the spanwise coherent structures is strongly decreased
(compare Figs. 3a and 3c).
Th e time-development of the spanw ise coherent structures ca n be visualized
nicely with their footprint on the wall, namely, fluctuations in the spanwise
wall vorticity qwall.how n in Fig. 4 for the three cases are time-space diagrams
of the spanwise-averaged w ~ , ~ ~ ~ ~lotted versus streamwise angle and time.
Dark lines in the diagrams (amplitude peaks in the wall vorticity) correspond
to propagating spanwise vortices inside the flowfield. A merging of these lines
reflects the pairing of subsequent vortices. These pairings occur repeatedly in a
subharmonic cascade. Regions of local flow separation are indicated by the
black areas (negative wall vorticity) in the downstream part of the flow
domain (Figs. 4a and 4b). Although two-dimesional harmonic forcing enhances
the wall-vorticity fluctuations in the upstream part of the flow and leads to fre-
quent flow separation in the downstream part of the flow (compare Figs. 4a
and 4b), three-dimensional steady forcing strongly reduces both wall-vorticity
fluctuations and local flow separation (com pare Figs. 4a and 4c). T his suggests
that the presence of streamwise vortices indeed inhibits the deve lopm ent of span-
wise coherent structures.
Fig. 4 DNS of turbulent Coanda wall jet showing time-space diagram s for
spanw ise-averaged wall-vorticity: a) “Unforced” reference: b) harmonic two-
dimensional forcing; c) steady three-dimensionalforcing.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 410/637
TURBULENT COANDA WALL JETS AND DNS AND RANS
409
b)
Fig. 5 DNS of turbulent Coanda wall jet showing effect of forcing on the mean-flow
development: a) Inverse-square of stream wise mean-velocity maximum ; and b) wall-
jet half-thickness vs streamwise angle. Experimental data4 are also plotted for
reference.
To
compensate for different initial development near the nozzle,
experimental data in a) are matched at = 35 deg with the unforced case.
The results from the DNS also showed that a strengthening or weakening of
the streamwise o r spanwise structures changes the downstream development of
the Coanda wall jet. For example, both the decay of the streamwise mean vel-
ocity and the radial spreading of the jet in the downstream direction are signifi-
cantly increased in response to the forcing (Fig. 5 . Individually, both
streamwise and spanwise structures facilitate entrainment of low-momentum
fluid from the ambient into the near-wall region of the jet , causing the observed
increase in spreading and velocity decay. Although the separated flow region is
not computed in our DNS, it may be conjectured that the wall jet will separate
from the cylinder surface farther upstream as a direct result of the increased
spreading and decay of the turbulent mean flow. This, in turn, has an effect
on the side force that is being generated. How ever, most of the interacting mech-
anisms between spanwise and streamwise vortical structures have to be investi-
gated in considerably more detail as numerous physical aspects are not yet fully
understood. This understanding is essential for the implementation of the
Coanda technology for practical applications.
B. Reynolds-Averaged Navier-Stokes RANS) Calculations
The applicability of the different available turbulence models for Coan da flow
calculations was scrutinized in two-dimensional RANS calculations of the
Coanda w all jet experiment by W ygnanski and
coworker^.^
The computational
grid used for these investigations is shown in Fig. 6 and consists of three
blocks. The grid sizes for the blocks consist of 200 x
75,
50 x 50, and
150
x 20 cells, respectively. For the turbulence models used in these calculations
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 411/637
41
a)
H F. FASEL, A. GROSS, AND S.WERNZ
b) C)
Fig. 6 Computational grid used for two-dim ensional RANS calculations: a) Entire
grid; b) close-up of cylinder; c) close-up of nozzle region.
the laminar sublayer needed to be resolved. The y+ values of the wall-next grid
points were between
0.2
and 1, and the Ax+ values were between
50
and 300. The
grid resolution in the jet was between 40 and 180 times the local Kolmogorov
length scale. A top-hat velocity profile was prescribed at the nozzle inflow.
The ambient was quiescent. The flow was assumed to be laminar at the nozzle
inflow and in the ambient.
Generally, most turbulence models gave disappointing results, some to a
larger degree than others. Typical results in the form of iso-contours of eddy-
viscosity from such RANS calculations are given in Fig. 7 . The 1988 K-6.1
model facilitates the strongest turbulent mixing across the wall jet and hence
leads to the fastest jet velocity decay and largest jet spreading and the earliest
separation. In contrast, when the K- -E or the Spalart-Allmaras model was
used, the je t wrapped around the cylinder more than once.
For some of these turbulence models the jet-velocity decay and jet-half-
thickness are plotted in Fig. 8 against streamwise angle. When the 1988 K-6.1
model was used in combination with the EASM model, a close match of the
jet-velocity decay with the measured data was achieved. However, even with
this model, the downstream development of the jet-half-thickness was poorly
predicted. The second-best model was the 1988 K - -W model.
Fig. 7 Two-dimensionalRANS computationsof Coanda flow showing eddy viscosity
norm alized by laminar viscosity Note that the K E and S-A results are transient):
a)
1988
K--O model; b) K--E model; c) Spalart-Allmaras model.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 412/637
a)
TURBULENT COANDA WALL JETS AND DNS AND RANS
411
b)
Fig.
8
Two-dimensional RANS computations
of
Coanda flow: a) Jet-velocity decay;
b)
jet half-thickness vs stream wise angle.
The shape of the normalized velocity profiles is predicted best by the K- -E model
(Fig. 9). The second-best results were obtained from the 1988 K- -W model with
EASM. However, because the predicted half-thickness was too small for all
models (Fig. 8), the non-normalized velocity profiles still do not match the experi-
mental velocity profiles. With the 1988
K- -W
model (with Boussinesq or EASM
Reynolds stress), very good predictions of the wall-pressure distribution were poss-
ible (Fig. 9). For the EASM model the separation location was slightly closer to the
experiment. When the K- -E and Spalart-Allmaras models were used, the jet
remained attached to the cylinder for more than
360
deg. To allow for a comparison
with the K- -W model results, the data shown for these two models are not from
steady-state solutions but from transient solutions at a time instant before the drift-
ing separation location had reached
6
=
360 deg.
For all but the 1988 K - -W model with EA SM, jet spreading and velocity decay
were underpredicted. Based on the
DNS
results one may assume that the turbu-
lence models failed to account for (or underpredicted) the additional mixing
facilitated by the strong coherent turbulence structures that are present in the
flow. Because the separation location was predicted within 10% of the exper-
imental result when the standard 1988 K- -W turbulence model was used, this
model was then chosen for subsequent three dimensional RANS stability inves-
tigat io m 6 For these three-dimensional computations, 48 grid cells were used in
the spanwise direction over a domain of width L,
= 0.3d
resulting in Az values
between
50
and 200. A periodicity boundary condition was applied in the span-
wise direction.
In these three-dimensional RANS simulations, several steady perturbations
with a periodic modulation in the spanwise direction were introduced
simultaneously at the nozzle exit, each with a different amplitude and spanwise
wavelength
h, k) = L, /k ,
where
k =
1 , 2 , . represents the spanwise Fourier-
mode number of a perturbation. One such case is illustrated in Fig. 10. The
streamwise development of these perturbations and their interaction was then
studied by plotting the amplitudes of the spanwise Fourier modes representing
the perturbations.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 413/637
41
2
H F. FASEL, A. GROSS, AND S.WERNZ
Fig. 9 Two-dimensional
RANS
computations of Coanda flow: a) Velocity
profiles at three downstream stations;
b)
wall-pressure coefficient
cp
= 2 p
pce)/ pjetv?et)*
Forcing at small amplitudes allows for a comparison with linear stability
theory. From the experiment by Wygnanski and coworkers4
it
was found that
the spanwise wavelength of the locally predominant structures scales roughly
with the local half-thickness of the jet (Fig. 11). This can be confirmed by
computation.
When the streamwise structures were forced with larger (nonlinear) disturb-
ance amplitudes, nonlinear subharmonic resonances could be observed
(Figs. 11 and
12 .
The results obtained for nonlinear amplitudes depend on
the relative phase between the modes. This becomes evident from the total
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 414/637
TURBULENT COANDA WALL JETS AND DNS AND RANS 4 3
Fig. 10
RANS
computation of Coanda wall jet. Spanwise Fourier modes k = 1,
2
forced at nonlinear amplitudes of
0.01
and O lvjet mode
k = 2
phase-shifted by m/2
relative to mode k =
1).
Iso-surfaces of azimuthal velocity component. As the jet
passes along the cylinder in the downstream direction the higher wavenumber
structures disappear, while the lower wavenumber structures emerge.
b)
1o
6 o
1o
10-10
0 50 100 150 200
6
6
Fig. 11
RANS
computation of Coanda wall jet showing amplitude of spanwise
Fourier modes k: a) Linear case, all modes forced at small disturbance amplitudes;
b) Fourier modes k = 1,
2
forced at large, nonlinear amplitudes of 0.01 and 0.lvjet
solid lines). Comparison with linear case dashed lines). In particular, close to the
nozzle 8 0 deg) the growth rates for the nonlinear forcing deviate substantially
from the growth rates for the linear forcing.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 415/637
41 4
a) 10 l
10 1
10 l
10 2
L
0
H F. FASEL, A. GROSS, AND S.WERNZ
no
phase shift
d2 phase shift
100 200
300
6
Fig.
12
Two-dimensional
RANS
computation of Coanda flow for Fourier modes
k
=
1, 2 forced at amplitudes of 0.01 and 0.lvjetand modes 1 and 2 forced in phase
and at a relative phase shift of n/2: a) Total circulation
r 8) 1 ~ 4 1
A; and
b) mode amplitudes.
circulation for 6
>
150deg These preliminary investigations suggest that both
linear instability as well as nonlinear subharmonic resonance are possible
viable mechanisms for the merging of the longitudinal vortices that was observed
in the experiments. Based on our calculations, the linear process appears to be
more likely for the present experimental conditions. However, for possible
control of the C oanda wall jet, the non linear resonance m echanisms might also
be exploited.
Because RANS underpredicted the wall-normal mixing (and hence the
jet-velocity decay and jet spreading), and because our DNS results clearly
indicate that strong turbulent coherent structures play a dominant role in
Fig. 13 Two-dimensional FSM computation of a Coanda wall jet: a) Vorticity;
b) contribution function. Because three-dimensional streamwise vortices are
deliberately excluded, the two-dimensional structures have a high intensity. The
spatial distribution of the contribution function clearly correlates with dominant
flow structures.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 416/637
TURBULENT COANDA WALL JETS AND DNS AND RANS
415
the turbulence mixing, application of our flow simulation methodology
(FSM)'8919appeared to be a logical choice. With FSM, depending on the
local turbulence characteristics and grid resolution, small-scale turbulent
motion is modeled, while large-scale coherent structures are computed in
a time-accurate fashion. Results from a preliminary two-dimensional FSM
are shown in Fig. 13. Large spanwise coherent structures arise as a conse-
quence of the inflectional wall-jet profile (Fig. 9a). The turbulence-model
contribution is clearly linked to the flow structures, as shown in the right
plot of Fig. 13.
V. Circulation Control Airfoil
A. Case Description
The airfoil-chord length was c
=
8 in (or 0.2032 m). The freestream velocity
was v, = 39.18 m/s, the freestream density p, = 1.226 kg/m3, and the free-
stream molecular viscosity p = 1.790
x
kg/ms. Assuming a gas constant
of R = 287.1 J/(kg .
K)
and a ratio of specific heats
y =
1.4, the freestream
temperature can be computed as
T ,
= (v, /w2/(yR)
=
265.21
K.
The Reynolds
number based on freestream velocity and chord length was
Re=--
pwv'ooc
.455 105
PCu
If the assumption p, = pjet is made, the jet-blowing ratio
B = vjet/vW =
c,p,c/(2pj,,b) is 5.90 for case 283 and 5.54 for case 321. How ever, this
i s u l t s in a nozzle-exit Mach number .7 and requires the use of a com-
pressible code. The nozzle-inflow area
is
Ai,/c =
0.03188. The nozzle-area
ratio is 10.2.
B.
Com putational Grid
The com putational grid used for the investigations discussed here is shown in
Fig. 14. The number of cells around the airfoil was 500, and the nozzle interior
Fig. 14 Computational grid for circulation control airfoil: a) Entire grid; b) close-
up of airfoil and block boundaries; and c) close-up of Coanda flow region.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 417/637
416
H F. FASEL, A. GROSS, AND S.WERNZ
was resolved by 100 x 80 cells. The resolutions of the individual blocks were
700
x 80 (block l) , 40 x 40 (block 2), and 400 x 50 (block 3). This results in
a total number of cells of 77,600. The total extent of the grid was 1Oc in both
x
and
y
measured from the center of the airfoil. The
y +
value of the wall-next
grid points was sm aller than one.
C. Boundary Conditions
Following common practice, velocities and temperature were set at the
freestream inflow boundary, while the static pressure was extrapolated. At
the outflow boundary all flow quantities were extrapolated, except for the
static pressure, which was prescribed. A stable and realistic nozzle-inflow
condition was found by extrapolating the static pressure and prescribing
the mass flux hi
= hjet pinvinAin
and the total temperature (the total temp-
erature at the nozzle inlet was chosen to match the total temperature of the
freestream). Inflow velocity
vin
and temperature,
Ti,
were then obtained by
solving
and
1 2
T i n -vin
YR
m+-v; =
- 1
2
y - 1 2
(4)
The wall was considered to be adiabatic and hydraulically smooth.
D. Results
With the 1988
K - -W
model and the Menter SST model the wall jet stayed
attached to the wall for too long (Fig. 15). Show n therefore are transient solutions
for these models. On the other hand, very good results could be obtained when the
EASM model was used.
Case 321 was computed with the 1988
K - -W
model and EASM only
(Fig. 16). For both cases the jet-exit velocity
vjet
6.7v,, resulting in a
jet-exit Mach number
Mjetx
0.85. The nozzle-pressure ratio (nozzle inflow
to nozzle exit) was approximately 1.6 and the nozzle-density ratio was
about 1.4. Wall-pressure distributions are shown in Fig. 17. For both cases
the prediction is in very good agreement with the experiment. When the
1998
K- -W
model with EASM was used, the wall jet separated somewhat
earlier, leading to a slightly smaller circulation augmentation and a slightly
smaller area enclosed by the pressure coefficient curves. The LE stagnation
point moved backward as a result of the increase in total circulation
(Figs. 18 and 19).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 418/637
TURBULENT COANDA WALL JETS AND DNS AND RANS
417
1988K 0
Menter
SST
1988~ 0
EASM
1988~ 0
EASM
Fig. 15 RANS calculation of CC airfoil, Case 283 a
0
deg). Eddy viscosity
norm alized by laminar viscosity is left) and turbulence kinetic energy right)
result for 1988
K 0
and Menter
SST
model are transient).
1988~ 0
EASM
Fig. 16 RANS calculation of CC airfoil, Case 321 a -8 deg). Eddy viscosity is
norm alized by laminar viscosity left) and turbulence kinetic energy right).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 419/637
418
a)
H F. FASEL, A. GROSS, AND S.WERNZ
b)
Fig. 17 RANS calculation of CC airfoil showing the wall-pressure coefficient
cp
=
201 pm)/ pjetvfet):) Case 283 a 0 deg); b) Case 321 a - 8 deg).
a)
b)
c
Fig. 18 RANS calculation
of
CC airfoil showing total velocity and streamlines: a)
Case 283
a 0
deg) 1988 K-W EASM;
b)
Case 283
a 0
deg) 1998 K-W
EASM; c) Case 321 a
8
deg) 1988
K - w
EASM.
Fig. 19 RANS calculationof CC airfoil showing total velocity and streamlines 1988
K-w
model with EASM ): a) Case 283 a
0
deg);
b)
Case 321 a
- 8
deg).
VI. Conclusions
Coanda wall jets for two different configurations were investigated numeri-
cally:
1) The circular cylinder from the experim ents by Wygnanski and coworkers;
2 ) the NCCR 1510-7067 N C C airfoil from the expe rime nts by Abramson
and
(the workshop C FD challenge).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 420/637
TURBULENT COANDA WALL JETS AND DNS AND RANS
419
Configuration 1 was investigated using DNS and
RANS
computations. In the DNS,
both spanwise and streamwise coherent structures were present in the flow. It was
conjectured that in the natural, unforced case both types of structures keep each
other at bay and that if either one was favored or forced by active flow control,
the other one would be weakened. This conjecture was probed by separately
forcing the spanwise and streamwise coherent structures at the nozzle inflow.
Forcing of the spanwise structures indeed strengthened their downstream coher-
ence, but did not noticeably weaken the streamwise structures. The reason for
this is unclear and necessitates further research. Forcing of the streamwise struc-
tures weakened the spanwise structures and strengthened the streamwise structures,
as expected. The downstream development and interaction of both types of struc-
tures and their influence on the turbulent flow are ultimately responsible for the
downstream development of the wall jet. The goal here is to actively control the
je t spreading and velocity decay by application of AFC at the nozzle exit.
Configuration
1
was also used to evaluate turbulence models for steady RA NS
of Coanda wall jets. None of the models tested correctly predicted all relevant
aspects of the flow. Evidently, important physical mechanisms are not modeled
correctly. For example, none of the employed turbulence models had a curvature
correction. Also, the strong turbulent coherent structures that are not captured in
steady and two-dimensional RA NS may significantly contribute to the mean flow
and turbulence characteristics. Relatively speakening, the models based on an
EASM Reynolds stress model performed best. Configuration 1 was also used
for steady RANS stability investigations. Steady streamwise structures were
introduced at the nozzle, and their development in the downstream direction
was investigated. At low disturbance amplitudes (linear case), the local size of
the dominant streamwise structures roughly scales with the local wall jet half-
thickness, an observation that was also made in the experiment. Overall, the
amplification of the streamwise coherent structures by the centrifugal Gortler
instability was rather small. If the streamwise coherent structures observed in
the experiment were of similar strength as in the linear three-dimensional
RANS computation, the vortex mergings observed in the experiment may be
explainable by linear stability mechanisms.
Based on the experience gained from studying configuration 1, the elliptic CC
airfoil (configuration
2)
was then computed using the RANS and by employing
the 1988 and 1998
K-6.1
models and the Menter SST model. In ou r calculations,
only use of the EASM Reynolds stress model resulted in good predictions of the
wall jet separation from the airfoil. For both angles of attack, excellent agreement
with the experimental data could be obtained with this model.
Acknowledgments
The authors gratefully acknowledge the Office of Naval Research for funding
of this research under grant number N00014-01-1-09, with Ronald J o s h serving
as program manager.
References
‘Tani, I., “Production
of
Longitudinal Vortices in the Boundary Layer Along a Concave
Wall,”
Journal
of
Geophysical Research, Vol. 67 , No.
8,
1962, pp. 3075-3080.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 421/637
420
H F. FASEL, A. GROSS, AND S.WERNZ
’Moser, R. D., and Moin, P., “The Effects of Curvature in Wall-Bounded Turbulent
Flows,” Journal
of
Fluid Mechanics, Vol. 175, 1987, pp. 479-510.
3Pajayakrit, P., and Kind, R. J., “Assessment and Modification of Two-Equation
Turbulence M odels,”
AIAA Journal,
Vol. 38, No.
6,
2000, pp. 955-963.
4Neuendorf, R., and Wygnanski, I., “On a Turbulent Wall Jet Flowing Over a Circular
Cylinder,” Journal of Fluid Mechanics, Vol. 381, 1999, pp. 1-25.
’Wernz, S., Valsecchi, P.,
Gross,
A., and Fasel, H. F., “Numerical Investigation
of
Turbulent Wall Jets Over a Convex Surface,” AIAA Paper 2003-3727, June 2003.
6Gross, A., Wernz,
S.,
and Fasel, H. F., “Numerical Investigation of Coherent
Structures in a Turbulent Coanda Wall Jet,” AIAA Paper 2003-4020, June 2003.
7Jones, G., and Joslin, R. D. (eds.), Proceedings
of
the 2004 NA SAIO NR Circulation
Control Workshop, NA SA /CP 2005-213509, June 2005.
‘Abramson, J., “Two-Dimensional Subsonic Wind Tunnel Evaluation of Tw o Related
Cambered 15-Percent-Thick Circulation Control Airfoils,” DTNSRDC Tech. Rept.
ASED-373, Sept. 1977.
’Slomski, J. F., Gorski, J. J., Miller, R. W., and Marino, T. A., “Numerical S imu lation
of Circulation Contro l Airfoils as Affected by Different Turbulence Models,” AIAA Paper
2002-0851, Jan. 2002.
“Paterson, E. G., and Baker, W., “Simulation
of
Steady Circulation Control for
Marinevehicle Control Surfaces,” AIAA Paper 2004-0748, Jan. 2004.
“Menter, F. R., “2-Equation Eddy-Viscosity Turbulence Models for Engineering
Applications,”
AIAA Journal,
Vol. 32, No.
8,
1994, pp. 1598-1605.
”Meitz, H. L., “Numerical Investigation of Suction in a Transitional Flat-Plate Bound-
ary Layer,” Ph.D. Dissertation, Dept. of Aerospace and Mechanical Engineering, Univ.
of
Arizona, Tucson, AZ, 1996.
13Meitz,H. L., and Fasel, H. F., “A Compact-Difference Scheme for the Navier-Stokes
Equations in Vorticity -Velocity Formulation,” Journal of Computational Physics,
‘‘Gross, A., and Fasel, H., “High-Order W E N 0 Schemes Based on the Roe Approximate
Riemann Solver,” AIAA Paper 2002-2735, June 2002.
‘’Wilcox, D. C.,
Turbulence Modeling
for
CFD,
2nd ed., DCW Industries, La Canada,
CA, 2000.
16Rumsey, C. L., and Gatski, T . B. “Recent Turbulence Model Advances Applied
to Multielement Airfoil Computations” Journal
of
Aircraft, Vol. 38, No. 5 , 2001,
17Spalart, P. R., and A llmaras,
S.
R., “A One-Equation Turbulence Model for Aerody-
namic Flows,” AIAA Paper 92-0439, Jan. 1992.
“Fasel, H. F., Seidel, J., and Wernz, S., “A Methodology for Simulations
of
Complex
Turbulent Flows,” Trans. ASME, Journal of Fluid Engineering, Vol. 124, 2002,
‘’Fasel, H. F., von Terzi, D. A., and Sandberg , R. D., “A Methodology fo r Simulating
Compressible Turbulent Flows,” FEDSM 2003-45334, 4th ASMEIJSME Joint Fluids
Engineering Conference, July 2003; also ASME Journal
of
Applied Mechanics (to be
published).
V O ~ . 57, NO. 1, 2000, pp. 371 -403.
pp. 904-910.
pp. 933-942.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 422/637
Chapter
16
RANS and Detached-Eddy Simulation of the
NCCR Airfoil
Eric
G.
Paterson* and W arren J. Bakert
Pennsylvania S tate University, University Park, Pennsylvania
Nomenclature
a
= speed of sound, ft/s
C D
=
section dra g coefficient,
F ~ / ( 1 / 2 ) p U i S
C L= section lift coefficient, F ~ / ( 1 / 2 ) p U i S
C
=
section moment coefficient, M z / (1 / 2 ) p U i S c
C,
= pressure coefficient,
( p
p o o ) / ( 1 / 2 ) p ~ L
C ,
=jet momentum coefficient, r i z ~ j / ( l / 2 ) p ~ L ~
FD = drag force, lbf
FL = lift force, lbf
f
= nondimensional frequency, f c / U ,
g
=
gravitational acceleration, ft/s2
h
=
slot height, in.
k = turbulent kinetic energy, ft/s2
C =
k-w,
or subgrid, length scale, in.
C = DES length scale, in.
c
=
foil chord length, in.
M
=
Mac h number, U / a
M ,
=
moment about the z-axis, ftelbf
m = mass flow rate, pUjhw, lbm/s
p
= pressure, lbf/ft2
*Senior Research A ssociate, Applied Research Laboratory and Associate Professor of Mechanical
'Graduate Research Assistant, Department of Aerospace Engineering. Member AIAA.
Copyright 005 by Eric G . Paterson and Warren J. Baker. Published by the American Institute
and Nuclear Engineering. AIAA member.
of Aeronautics and Astronautics, Inc., with permission.
42
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 423/637
422
E. C. PATERSON AND W .
J.
BAKER
Re =
Reynolds number, p U w c / p
s=
planform area, cw , ft2
U ,
V ,
W
=
velocity components, ft/s
U ,
=
friction velocity,
m
t /s
w = foil span, in.
x , y , z =
Cartesian coordinate
y + =
wall coordinate, U , j / v
p = distance from wall, in.
a = angle of attack, deg
A = maximum dimension of local grid cell
At* =
nondimensional time step,
AtU, / c
S
a 6, = dimensions of local grid cell in each curvilinear
coordinate direction
p
= dynamic viscosity, lbm/ft.s
p
=
density, lbm/ft3
u= DES blending function or cavitation number
T~ = wall-shear stress, lbf/ft2
w
= turbulent dissipation rate, ft2/s3
8 7 = curvilinear coordinates
Subscripts
00 =
freestream
min
=
minimum
=
at je t orifice
Superscripts
r
=
resolved turbulence
s =
subgrid turbulence
tot
=
total
I. Introduction
IRCULA TION control (CC ) for lift augmentation via the Coanda effect has
C een studied for many years.192 n comparison to mechanical means of CC
(e.g., shape change and leading- and trailing-edge flaps), the use of a wall jet
on a convex curved trailing-edge (TE) surface is attractive for many reasons.
Based upon aerospace flow-control applications3 and previous hydrodynamic
assessment^,^ ̂
anticipated benefits for naval vehicles include simplification of
actuation, reduction in weight and number of parts, dual-mode operation (i.e.,
cruise and high-lift scenarios), contribution to novel design options such as
placing control surfaces at nontraditional locations and arrangement of sensors
and payloads on control surfaces, and improved shock resistance.
As with all flow control scheme^,^-^ there are technical as well as economic
and operational issues that must be overcome for systems to be transitioned into
practical application. For example, for CC schemes to be incorporated in the
marine environment, they must address the inherent drag penalty of a blunt TE
at cruise condition, overcome operator reluctance to fixed control surfaces, not
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 424/637
DETACHED-EDDY SIMULATION OF NCCR AIRFOIL
423
suffer from orifice fouling or shock damage, and, for applications w here stealth is
important, have limited impact on the hydroacoustic ~ignature.~he work
presented herein is ultimately motivated by these issues.
Continued development of new actuation methods' potentially leads to novel
solution of issues. Actuators such as high-performance solenoid valves, smart
materials, zero-net-mass actuators, synthetic jet actuators, and plasma control
actuators find application to CC as w ell as other forms of flow control. Of particu-
lar interest to CC are high-performance solenoid
valve^,^ which can achieve effi-
cient pulsed blowing, a mode of CC that has been known to reduce mass-flow
requirem ents for a given performance increment. -12 However, detailed under-
standing of both the unsteady flow physics and their application in water-based
scenarios is lacking.
Even for steady blowing CC, there are important flow physics that compu-
tational fluid dynamics (CFD) models must be able to simulate if such tools are
to be used in design. M ost notable are streamw ise curvature effects on the turbu-
lent boundary layer and spanwise coherence of the wall jet . Nearly the en tire range
of Reynolds-averaged N avier- Stokes (RA NS) turbulence models from algebraic
to full Reynolds-stress transport models (RSTMs) have been modified for curva-
ture effects.13-15 Unfortunately, the s tate of affairs is poor in that modifications to
algebraic and one- and two-equation models are limited in range due to empiri-
cism, whereas RSTMs have yet to convincingly demonstrate capability to
resolve subtleties in the way curvature impacts mean flow and turbulence
s t r u c t ~ r e . ' ~onetheless, numerical experim ents for a C C configuration16 have
demonstrated that baseline RSTMs can improve simulation results in comparison
with baseline two-equation m odels, especially at large je t momentum coefficients.
Moreover, this study showed that simulations using two-equation models dem on-
strated nonphysical behavior with a dramatic reduction in lift and a wall jet
that remained attached to the surface for 1.5 revolutions around the foil.16
Unfortunately, the source (e.g., model limitations or numerical accuracy) of this
discrepancy, and whether or not
it
is flow-code-specific, was not identified.
Detailed understanding of the high-Reynolds-number turbulent wall je t on the
Coanda surface would best be facilitated by direct numerical simulation (DNS),
or possibly large eddy simulation (LES). For the usual reasons, that is, lack of
computer power, this is not yet realizable. Therefore, the approach pursued
here is one based upon the detached-eddy simulation (DES),17 which is a
hybrid RANS/LES method. In this approach, the foil fore body and the near-
wall region is treated as RAN S and the ou ter regions of the after body boundary
layer and near wake are treated as LES. Detached-eddy simulation has been
shown to improve accuracy for massively separated and has been
applied to an active flow control application with zero-net-mass actuation,20
albeit with inconclusive results. The ability of DES to resolve curvature
effects, or the need for curvature modifications in the RANS portion of the
DES model, is unknown.
Although the objective of our research is to develop validated simulation tools
using recently acquired incompressible water-tunnel data for a low-aspect-ratio
tapered control surface21 and wind-tunnel data for a pulsed C C c ~ n f i g u r a t i o n , ~ ~ ' ~
the work presented herein represents our initial efforts to apply RANS and
DES to a sim pler steady-blowing CC configuration.22 It has been selected as a
preliminary validation exercise because of the fact that
it
can be treated as a
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 425/637
424 E. C. PATERSON AND W .
J.
BAKER
two-dimensional geometry and has previously been studied using RAN S CFD.16
Our progress is reported in Secs. 11-VII.
11.
Geom etry, Conditions, and Data
The NCCR-1510-7067N CC foil was tested in a wind tunnel at the David
Taylor Naval Ship Research and Developm ent Center in 1977.22The geometry
was a 15% thick elliptical cambered foil with a s ingle jet orifice on the upper
surface at
x/c =
0.967. The model chord length was
c =
8 in., the slot height-
to-chord ratio
h/c =
0.003, and the Coanda surface a nominal circular arc. A
cross-section of the model is shown in Fig. 1.
Although a wide range of C and a were studied in the original experiment,
two cases are studied here. For the first, designated as Case 283,
C
=
0.209
and a = Odeg. For the second, designated as Case 321,
C
= 0.184 and
a = deg. Both are assumed to have the follow ing comm on param eters: free-
stream velocity U , = 128.54 fps, freestream density
p,
= 0.07654 lbm/ft3, and
kinematic viscosity
=
3.73
x
lo-’ slug/ft-s. This yields a Reynolds number
of
Re =
5.45
x
lo5 and a Mach number of M , = 0.12. Assuming that the jet
is incompressible (i.e., p j / pm
=
l), the nondimensional jet-orifice velocities
can be computed as
v j / U m
= - l - C , =
5.90 and 5.54
;:;
for Cases 283 and 321, respectively. Although this assumption introduces an
unknown modeling error, a posteriori evidence suggests that it is sm all.
Available experimental data are somewhat limited in comparison to modem
experiments, consisting of surface pressure measured via pressure taps placed
at midspan. Experimental lift and moment were computed by integrating the
surface pressure, and drag was evaluated using a wake survey and a m omentum-
deficit method. In addition, estimates of experim ental uncertainty are not avail-
able; however, several possible sources have been identified such as slot-height
TRAILING
EDGE
Fig.
1 Cross-sectional geometry of NCCR 1510-7067N.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 426/637
DETACHED-EDDY SIMULATION OF NCCR AIRFOIL
425
growth because of plenum pressure, and Coanda je t interaction with tunnel walls,
especially at large
C,,
such that the effective a is different from the geometric a
111. Computational Methods
A. Unsteady
RANS
CFD SHIP-IOWA23 is a general-purpose parallel unsteady incompressible
RANS CFD code. The computational approach is based upon the pressure-
implicit split-operator (PISO) approach, which iteratively solves the momentum
and pressure-Poisson equations. Discretization is achieved using structured
overset grids and the finite-difference method, where convective terms are dis-
cretized using a general five-point stencil that permits a user-specified order-
of-accuracy ranging from first-order upwind to fourth-order central. Viscous
and temporal terms are discretized using second-order central and second-
order backward methods, respectively. Turbulence is modeled using a linear
closure and the blended K - W / K - - E SST two-equation Efficient parallel
computing is achieved using coarse-grain parallelism via MPI distributed com-
puting. For time-accurate unsteady simulations, global solution of the pressure-
Poisson equation is achieved using preconditioned GMRES and the PETSc
libraries.25926
B. Detached-Eddy Simulation
Detached-eddy simulation is a three-dimensional unsteady numerical
method using a single turbulence model, which functions as a subgrid-scale
model in regions where the grid density is fine enough for LES, and as a
RANS
model in all other regions. Implementation of DES in CFDSHIP-IOWA was
accomplished by modifying the turbulence model and convective-term
discretization.
The turbulence model is modified by introducing a DES length scale
t
=
min
e k w , C D E S h )
(1)
which compares the subgrid length scale to the local grid size, where the former
can be written as
CDEs
s a model constant with a value between 0.78 and 0.61 weighted by the
Menter k-w1k-E blending function,24 and
A
is based on the largest dimension
of the local grid cell:
3)
=
max ( , a
8,)
The new length scale
t
replaces
t k w
in the destruction term of the k-transport
equation
pk3I2
Dk,,,
=
pp*kw
=
e k w
4)
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 427/637
426 E. C. PATERSON AND W .
J.
BAKER
which results in a new destruction term:
The effect of this modification on the turbulence budget is to shift energy from
subgrid, or modeled, scales to resolved scales as defined by the filter width
CDESA.
Th e second m odification a ims to redu ce numerical dissipation inherent in the
upwind convective-term discretization scheme . The implemented approach is
based upon a hybrid central/upwind approximation of the convective terms (or
fluxes):
where u s defined as
7)
The result is that u smoothly transitions between
1 0
in the RANS regions,
resulting in an “almost upwind” scheme, and
0.0
in the LES regions, resulting
in an “almost centered” scheme. In addition, a Courant-number constraint of
1 0
has been im pose d, which requires that the time step be sufficiently small to
support turbulent eddies. The coefficients n and permit the interface between
RANS and LES regions to be arbitrarily “sharpened”; however, currently we
use n = m = 1 because of the fact that higher-order coefficients have resulted
in unstable simulations.
In CFDSHIP-IOWA the convective terms are discretized with the following
higher-order upwind formula
where
DES implementation is accomplished by redefining these equ ations
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 428/637
DETACHED-EDDY SIMULATION OF NCCR AIRFOIL 427
where W,,, W, W,, W,, and W,, are hybrid coefficients defined as the blending
between second-order upwind and fourth-order central schemes:
w,,
=
(1
) w E
+
ow:
w,= (1 ) w Z
+
mv?
w,
(1 ) w F +mv?
w,=
(1
)wtlf
+
(13)
w,, =
(1
)wtlf,
+ow:
Finally, as discussed in the following section, it is noted that CFDSHIP-IOWA
is an overset-grid capable CFD code w ith an interface to PEGASUS 5.1.27 This
capability will be exploited to perform local grid refinement and flow adaptation
in the wall-jet, wake, and LES regions.
IV. Grid Generation
Overset grids are generated primarily using hyperbolic extrusion, although
transfinite interpolation and elliptic smoothing is used for blocks that do not
lend themselves to that approach, that is, the background mesh and plenum
mesh. Overset interpolation coefficients and holes are computed using Pegasus
5.1.27 CFDSHIP-IOWA employs double-fringe outer and hole boundaries so
that the five-point discretization stencil (i.e., in each curvilinear coordina te direc-
tion) and order-of-accuracy does not have to be reduced near overset boundaries.
Level-2 interpolation capability of PEGASUS 5.1 is also used so as to achieve
optimal match between donor and interpolant meshes.
Grid design is based upon a domain size of
5 x/c 5
4,
5 y / c 5
2, and
0
I
/ c
I
.2, and a near-wall spac ing of 1.0
x
the latter of which aims to
resolve the sublayer of the turbulent boundary layer with a wall spacing of
The grid system used for
RANS
simulations is shown in Fig. 2. Nested orthog-
onal uniform box grids are used for the far-field and a simple 0-grid is used for
the foil. Preliminary solutions were used to locate streamlines, and wake-refine-
ment blocks were built off these streamlines for subsequent higher-fidelity simu-
lations. RANS simulations were computed in a pseudo-two-dimensional fashion
that requires five points in the spanwise direction. The entire grid system consists
of 323,000 points and comprises eight blocks ranging in size from 30,000 to
5 1,000 points.
For DES, the approach is the same as described above, except that the span-
wise resolution must be increased in regions where turbulent eddies are to be
resolved. Overset grids are effectively used to locally refine the simulation. As
shown in Fig. 3, the fore body and far-field, which is in the
RANS
region, is
resolved with five points in the spanwise direction. In contrast, the TE and
near-wake blocks are resolved with 41 points in the spanwise direction. The
wake refinement mesh shown in Fig.
3
is designed for unblown C=
0
y + = 1.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 429/637
428
E. C. PATERSON AND W .
J.
BAKER
a)
Fig. 2 Overset grid system for
RANS
simulation: a) Overall view; b) foil view;
c) plenum and TE view.
simulations and has an isotropic spacing of
=
0.005. The entire grid system
consists of
855,000
points and comprises 15 blocks ranging in size from
31,000 to
67,000
points. It is noted that translational periodicity is imposed in
the spanwise direction and that the extent of the domain in this direction is
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 430/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 431/637
430
E. C. PATERSON AND W .
J.
BAKER
RANS and DES, a cubic polynomial is used to accelerate the foil from rest over a
nondimensional time of 2.0.
No-slip boundary conditions were applied on all surfaces of the foil and the
top and bottom walls of the plenum. On the inlet face of the plenum, a top-hat
velocity profile was prescribed with the magnitude com puted using conservation
of mass, known Uj /U, at the jet orifice, and a plenum contraction of 10.63. For
Cases 283 and 321, this velocity magnitude corresponds to 0.555 and 0.521,
respectively. In addition, it was assumed that the inflow at this location was
laminar. Inlet, far-field, and exit conditions were applied on the outer boundaries
of the largest box grid and translational periodicity was applied on all spanwise
faces. Neumann conditions were used for pressure on all boundaries. As already
mentioned, outer and hole boundary trilinear interpolation coefficients were com-
puted using Pegasus 5.1
. ’
Mathem atical formulation of all boundary conditions
are described in the C FDSH IP-IOWA users’ manual.21 Finally, it is noted that
boundary conditions are set and input file created using the CFDSHIP-IOWA
filter in the GRIDGEN software from Po intwise, Inc.
VI. Results
Research has been undertaken along two paths, both of which are presented.
First, RANS simulations for Cases 283 and 321 will be presented. Second,
DES results for the unblown
C, =
0
case will be shown and discussed.
A. Steady RANS Simulation
A comparison of experimental and simulated surface pressure is shown in
Fig. 4. Relatively good agreement is dem onstrated for both cases. The largest dis-
crepancy is the underprediction of the suction peak aft of the jet orifice. Case 283
shows a strong LE low pressure, relatively uniform loading over the m ajority of
the chord, and a Cp minf 7 and 8 at the TE for the simulation and exper-
iment, respectively. Because of the negative angle of attack, Case 321 lacks the
LE low pressure. It also shows larger error in comparison to the data across the
chord, but especially on the Coanda surface. The predicted and experimental
Cp,min
re 3 and 18, respectively.
Lift, drag, and moment about the z-axis centered at midchord were com puted
by integrating C and T~ on all external surfaces. All plenum surfaces were neg-
lected in the C FD integration process so that comparison could be m ade to exper-
imental va lues. Experimental values were computedz2 by directly integrating the
discrete (and fairly coarse) surface-pressure data. Results are tabulated in
Table 1. The lift coefficient for Case 283 is within 5% of the data, whereas
Case 321 shows a discrepancy of 30% because of the larger underprediction of
the suction peak on the Coanda surface. Drag for both cases shows a very
large difference from the data. The data, which were measured using a wake
profile corrected by the jet momentum, show a negative drag, whereas the
CFD values (which includes both viscous and pressure com ponents) are positive
and substantially larger in magnitude. Moment coefficient C M is positive (LE
down, TE up) for both cases because of the large suction peak on the Coanda
surface. Data for
CM
are not available.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 432/637
DETACHED-EDDY SIMULATION OF NCCR AIRFOIL 4
1
a)
Fig. 4 Comparison of experimental symbols) and computational lines) surface
pressure: a) Case 283,
C
=
0.209,
a
=
0 deg; b) Case 321
C
=
0.184,
a
=
-8 deg.
Figure illustrates the impact of the Co anda effect upon the overall circula-
tion. For both cases, velocity-magnitude contours show a high velocity on the
top surface that is consistent with the surface pressure shown in Fig.
4.
The
streamlines show the effect of the cha nge in angle of a ttack on the overall flow-
field and on the locations
of
stagnation points.
Table 1 Lift, drag, and moment coefficients
CL
D
C M
Data CFD Data CFD Data CFD
Case
283 4.2 4.0 .05
0.18 2.07
Case
321 3.1 2.4 .08 0.12 1.21
CFD
omputational fluid dynamics.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 433/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 434/637
DETACHED-EDDY SIMULATION OF NCCR AIRFOIL
433
a)
Fig. 6 Leading-edge view of velocity magnitude contours and streamlines: a) Case
283,
C
=
0.209,
a
=
0
deg;
b)
Case 321,
C
=
0.184,
a
=
8 deg.
of kinetic energy, both of which correspond to regions of high mean shear. The
first is downstream of the jet-slot knife edge and grows along the wall-jet shear
layer. The second, which is large r in magnitude, starts at the point of wall je t sep-
aration a nd grow s into the wake. It is noted that the max imu m
k
is approximately
0.7
which is two orders-of-magnitude larger than
k
in the turbulent boundary
layer.
To better understand the evolution of the wall jet, profiles of velocity magni-
tude and turbulent kinetic energy a re extracte d at two locations for Case
283,
as
shown in Fig. 9. Location A is slightly aft of the je t orifice, and location B is along
a
y
=
0
line. At location A, the wall je t and its correspondingly strong shear layer
are clearly shown. The turbulent kinetic energy shows spikes downstream of the
plenum walls, the outer of which merges with
k
from the suction-side boundary
layer. At location B, the peak velocity magnitude is close to that at location
A;
however, the wall-jet shape has greatly thickened as a result of viscous and
turbulent stresses near the wall and along the shear layer. The turbulent kinetic
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 435/637
434
a)
E. C. PATERSON AND W .
J.
BAKER
Fig. 7 Trailing-edge view
of
velocity magnitude contours and streamlines: a) Case
283,
C =
0.209,
a = 0
deg;
b)
Case 321,
C
= 0.184,
a
= -8 deg.
energy has significantly grown in both magnitude and thickness, both of which
are consistent with velocity profiles and k contours shown in Fig. 8.
In preparation for future DES of the blown cases, the length scale in Eq. (2)
was computed for Case 283 and is shown in Fig. 10. This shows that the
largest eddies in the boundary layer and near wake are of the order of 0.02~.
However, the length scale is much smaller (i.e.,
,
0.002) in the near
orifice region. Therefore, target grid spac ing in this area should be approximately
=
0.001, which i s five times finer than the grid used in the unblown simulations
discussed in the next section.
B. Detached-Eddy Simulation
Detatched-eddy simulation (DES) was performed for
10,000
time steps with
At = 0.001 (or 10 flow-through periods). Animations of the instantaneous iso-
surface of vorticity shaded by spanwise velocity were made and snapshots are
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 436/637
DETACHED-EDDY SIMULATION OF NCCR AIRFOIL
435
Fig. 8 Contours of turbulent kinetic energy: a) Case 283, C ,
=
0.209,
a
=
0
deg;
b) Case 321,
C ,
=
0.184,
a
=
-8 deg.
shown in Fig. 11. The side view clearly shows the dominant vortex shedding of
spanwise eddies. The overset grid is also shown in the background to illustrate the
effect of switching from high to low, that is, LES-to-RANS, grid resolution in the
near wake (i.e, at about 0 . 4 ~ ownstream of the TE). All spanwise structure is
filtered and only the “two-dimensional” vortex passes through this interface.
The top view clearly displays the longitudinal vortices, which are intertwined
with the spanwise vortices. Again, the impact of switching from high to low
grid resolution is shown. The lack of spurious numerical reflections at this
overset boundary is noted.
Mean and root-mean-square
(RMS)
statistics for all dependent variables were
computed over 6000 time steps. Figure
12
shows the contours of the mean axial
velocity
u,
treamlines through the mean field, and RMS axial velocity fi. he
mean flowfield shows a typical wake with two eddies. The RMS velocity field
also shows a typical wake pattern28’29with two peaks across the wake corre-
sponding to vortices shed off the top and bottom sides of the foil. It is noted
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 437/637
436
a)
0.014
0.012
.
0 01
0
0.006
0.006
0.004
0.002
00
E.
C.
PATERSON AND W.
J.
BAKER
3
5
6
Vel ity mag nit ude (U2+4/2)1’2
6
Fig. 9 Extracted profiles: a) Velocity m agnitude; b) turbulent kinetic energy.
that computed statistics were not yet fully two-dimensional, thus indicating that a
larger integration time is needed to reduce uncertainty in the com puted statistics.
Analysis
of
the turbulent kinetic energy is shown in Fig.
13.
Subgrid turbu-
lence
kS
is computed from the modified k - w turbulence model, whereas
the resolved turbulence is computed from the velocity correlations
k‘
=
(El+
W
+
WW). Total kinetic energy is the sum of these two parts. These
figures show that
kS
s significant only in the boundary layer upstream
of
the sep-
aration. Downstream, total
k
is comprised of resolvable scales only. A region of
particular interest is the potential “gray region” where the solution switches from
RANS to LES nd where the model’s response to the underlying grid does not
yield either a fully LES or a fully
RANS
solution. Contours of total
k
show a
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 438/637
DETACHED-EDDY SIMULATION OF NCCR AIRFOIL
437
Fig.
10 k w
length scale for Case 283.
Fig. 11 Instantaneous iso-surface of vorticity shaded by spanwise velocity
component:
a)
Side view; b) top view.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 439/637
438
a)
E. C. PATERSON AND W .
J.
BAKER
Fig. 12 Statistical analysis of axial velocity: a) Mean velocity; b) root-mean-square
velocity.
slight decrease in magnitude as the T E is approached, and highlights a deficiency
in the overall approach, which is consistent with other recent high
Re
TE DES
sir nu la ti on^.^^
Finally, Fig. 14 shows spectral analysis of velocity at a single point
x / c ,
y / c
= (1.117, 0.016), the location of which was shown in Fig. 12. The time
history and Fourier transform show a shedding frequency at f = c / U , = 3.8.
If a new length scale is defined as the vertical distance between points of mean sep-
aration at the TE, which is
d / c = 0.052,
a more appropriate shedding frequency is
computed to be f
=
d / U ,
=
0.198, which is consistent with a typical Strouhal
number of 0.2. The Fourier transform shows higher harmonics at ff
=
7.5 and
f; = 12, which are 2f$ and 3f$, respectively, and a decay of the higher frequencies
at /3 slope up to a frequency of about 30, the latter of which is consistent with a
grid spacing of
0.005
and the assumption of 10 grid points per wavelength.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 440/637
DETACHED-EDDY SIMULATION OF NCCR AIRFOIL
439
a)
Fig. 13 Analysis
of
turbulent kinetic energy: a) Subgrid, k S ; b) resolved, k ;
c) total,
k S
+
k .
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 441/637
440
E. C. PATERSON AND W .
J.
BAKER
0.75
0.5
0.25
time
(UVC)
Powerspectral density
of
axial velocity
b)
Fig.
14
Frequency analysis
of
axial velocity at (x/c,
y / c )
=
1.117, 0.016): a) Time
history; b) Fourier transform.
C. Cavitation-Free Operating Depth and Speed
Given the low pressure on the Coanda surface, cavitation is a concern fo r ship
hydrodynamics. As a rough estimate, cavitation occurs when the magnitude of
minimum pressure coefficient exceeds the cavitation number:
-cp 2
(T
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 442/637
DETACHED-EDDY SIMULATION OF NCCR AIRFOIL 441
Given that p m = pgz, u increases linearly with depth. Substituting p m into
Eq.
( 1 3 ,
an expression for cavitation-free operation relating
Cp,,,in,
depth
z
and vehicle speed U , can be derived:
3
Using properties of water at
15°C
p
=
1000
kg/m ,pv=
1.7
Wa), a family of
curves can be computed that relates the three variables. Such a figure is shown
in Fig.
15.
It illustrates, for example, that for a
Cp,fin
= -20, cavitation can
be avoided at all depths greater than 50 ft as long as speed remains lower
than
1Okn.
Because
CC
is envisioned for low-speed littoral operation where traditional
control surfaces lose control authority, this is a favorable observation. On the
other hand, a speed of 30
kn
would require a depth of
750
ft to achieve cavita-
tion-free operation, at least for the C studied herein. Fortunately, because
dynamic pressure increases with Urn, ower C and
CL,
and therefore decreased
Cp,min, ould be required at high speed, thus permitting CC to be used throughout
the operation envelope.
VII.
Conclusions
A
CC
foil was studied using incompressible RANS and DES
CFD
methods.
RANS simulations of large jet-momentum coefficient cases demonstrated that
a linear closure with blended
k - W /k E
turbulence model was able successfully
to predict the pressure distribution trends in comparison to benchmark data. This
30
25
20
10
5
10 20 40
50
60
Fig. 15 Cavitation-free operation curves.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 443/637
442 E. C. PATERSON AND W .
J.
BAKER
contrasts w ith other published results,16 which indicate the need for higher-order
curvature-corrected models such as a full Reynolds-stress model. The reason for
this discrepancy is unknown, but recent work by Baker and Paterson3’ indicates
that near-wall grid resolution on the Coanda surface plays an important role when
using two-equation turbulence models. Details of the simulated flow were pre-
sented through analysis of the integral forces and mom ent, velocity field, and tur-
bulent kinetic energy.
Detached-eddy simulation was undertaken for the unblown case, and demon-
strated that the m ethod is capable of resolving turbulent vortex shedding. Statisti-
cal and spectral analysis was undertaken to explain the simulation results;
however, as with the RANS simulations, lack of data precludes validation for
this problem. Nonetheless, results are encouraging and suggest further appli-
cation of DES to both C C studies as well as other TE applications (e.g., propulsor
blades and nozzles).
Future work will focus on validation using modern water-tunnel data for a
low-aspect-ratio ta ered control surface*l and wind-tunnel data for a pulsed
cases will permit study of three-dimensional effects and pulsed blowing, both
of which are important issues for practical application and improved understand-
ing of basic CC flow physics.
CC conf ig~ra t ion .~’In addition to providing high-fidelity flowfield data, these
Acknowledgments
The authors gratefully acknowledge support from both the Office of Naval
Research through Grant Number N00014-03-1-0122 (Program Officer: Ron
Joslin) and NAVSEA SUB-RT (Program Manager: Meg Stout), the latter of
which was in the form of a graduate student fellowship for the second author.
The DoD High Performance Computing Modernization Office (HPCMO) and
Army Research Laboratory-Major Shared Resource Cen ter are acknowledged
for providing computing resources through DoD HPCMO Challenge Project
Number C1E.
References
‘Englar, R., “Circulation Control Pneumatic Aerodynamics: Blown Force and Moment
Augmentation and Modifications; Past, Present, and Future,” AIAA Paper 2000-2541,
June 2000.
*Wood, N., and Nielson, J., “Circulation Control Airfoils Past, Present, and Future,”
AIAA Paper 1985-0204, Jan. 1985.
3McLean, J. D., Crouch, J. D., Stoner, R. C., Sakurai, S . Seidel, G. E., Feifel, W. M.,
and Rush, H. M., “Study of the Application of Separation Control by Unsteady Excitation
to Civil Transport Aircraft,” NASA/CR-1999-209338, June 1999.
4Joslin, R., Kunz, R., and Stinebring,
D.,
“Flow Control Technology Readiness:
Aerodynam ic versus Hydrodynamic,” AIAA Paper 2000-44 12, June 2000.
5Hess, D., and Fu, T., “Impact of Flow Control Technologies on Naval Platforms,”
AIAA Paper 2003-3568, June 2003.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 444/637
DETACHED-EDDY SIMULATION OF NCCR AIRFOIL
443
6Bushnell, D., “Application Frontiers of ‘Designer Fluid Mechanics’ Visions versus
Reality or An A ttempt to Answer the Perennial Question ‘Why Isn’t It Used?’,’’ AIAA
Paper 1997-2110, June 1997.
’Howe, M., “Noise Generated by a Coanda Wall Jet Circulation Control Device,”
Journal
of
Sound and Vibration, Vol. 249, No. 4, 2002, pp. 679-700.
‘Schaffler, N., Hepner, T., Jones, G. nd Kegerise, M., “Overview of Active Flow
Control Actuator Development at NASA Langley Research Center,” AIAA Paper 2002-
3159, June 2002.
’Jones, G. iken, S., Washburn, A., Jenmins, L., and Cagle, C., “An Active Flow
Circulation Controlled Flap Concept for General Aviation Aircraft Applications,” AIAA
Paper 2002-3157, June 2002.
“Oyler, T., and Palmer, W., “Exploratory Investigation of Pulse Blowing for Boundary
Layer Control,” Tech. Rept. NR72H-12, North American Rockwell, Jan. 1972.
Waiters, R., Myer, D., and Holt, D., “Circulation Control by Steady and Pulsed
Blowing for a Cambered Elliptical Airfoil,” Aerospace Engineering TR-32, West Virginia
Univ., Morgantown, WV , July 1972.
12Jones,
G.
nd Englar, R., “Advances in Pneumatic-Controlled High-Lift Systems
Through Pulsed Blowing,” AIAA Paper 2003-341 1, June 2003.
‘3Wallin, S., and Johansson, A., “Modeling Streamline Curvature Effects in Explicit
Algebraic Reynolds Stress Turbulence Models,” International Journal
of
Heat and
Fluid Flow, Vol. 23, 2002, pp. 721-730.
14Patel,V., and Sotiropoulos, F., “Longitudinal Curvature Effects in Turbulen t Boundary
Layers,” Progress in Aerospace Science, Vol. 33, 1997, pp. 1-70.
”Gatski, T., and Speziale, C., “On Explicit Algebraic Stress Models for Com plex Tur-
bulent Flows,” Journal
of
Fluid Mechanics, Vol. 254, 1993, pp. 59-78.
‘6Slomski, J., Gorski, J., Miller, R., and Marino, T., “Num erical S imulation of C ircula-
tion Control Airfoils as Affected by Different Turbulence Models,” AIAA Paper 2002-
0851, Jan. 2002.
”Strelets, M., “Detached-Eddy Sim ulation of Massively Separated Flows,” AIAA Paper
2001-0879, Jan. 2001.
“Squires, K., Forsythe, J., Morton,
S.,
Strang, W., Wurtzler, K., Tom aro, R., Grismer, M.,
and Spalart, P., “Progress on Detached-Eddy Simulation of Massively Separated Flows,”
AIAA Paper 2002-1021, Jan. 2002.
”Forsythe, J., Squires, K., Wurtzler, K., and Spala rt, P., “Detached-Eddy Simulation of
Fighter Aircraft at High Alpha,” AIAA Paper 2002-0591, Jan. 2002.
2oSpalart, P., Hedges, L., Shur, M., and Travin , A., “Simulation of Active Flow C ontrol
on a Stalled Airfoil,” Proceedings
of
IUTAM Symposium on Unsteady Separated Flows
Apr. 2002.
”Rogers, E., and Donnelly, M., “Characteristics of a Dual-Slotted Circulation Contro l
Wing of Low Aspect Ratio Intended for Naval Hydrodynamic Applications,” AIAA
Paper 2004- 1244, Jan. 2004.
”Abramson, J., “Two-Dimensional Subsonic Wind Tunnel Evaluation of Two Related
Cambered 15-Percent Circulation Control Airfoils,” DTNSRDC
ASED-373,
Sept. 1977.
23Paterson, E., Wilson, R., and Stem , F., “General-Purpose Parallel Unsteady RANS
Ship Hydrodynamics Code: CFDSHIP-IOWA,” Tech. Rept. 432, IIHR Hydroscience
and Engineering, Univ. of Iowa, Ames, IA, Nov. 2003.
24Am es, I. A., and Menter, F., “Two-Equation Eddy Viscosity Turbulence Models for
Engineering Applications,” AIAA Journal, Vol. 32, No.
8,
1994, pp. 1598- 1605.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 445/637
444 E. C. PATERSON AND W .
J.
BAKER
25Balay, S., Buschelman, K., Gropp, W. D., Kaushik, D., Knepley, M., McInnes, L. C.,
Smith, B. F., and Zhang, H., “PETSc Users Manual,” Tech. Rept. ANL -95/11-Revision
2.1
S
Argonne National Lab., Jan. 2003.
26Balay,
S.,
Gropp, W. D., McInnes, L. C., and Smith, B. F., “Efficient Management of
Parallelism in Object Oriented Numerical Software Libraries,” Modern Software Tools in
Scient c Computing, edited by E. Arge, A. M. Bruaset, and H. P. Langtangen, B irkhauser
Press, Cambridge, MA, 1997, pp. 163-202.
2 7 S ~ h s, . E., Rogers, S . E., Dietz, W. E., and Kwak, D., “PEGASUS 5: An Automated
Pre-Processor for Overset-Grid CFD,” AIAA Paper 2002-0101, June 2002.
”Blake, W., “A Statistical Description of Pressure and Velocity Fields at the Trailing-
Edges of a Flat S trut,” DTNSRDC Rept. 4241, Dec. 1975.
29Paterson, E. G. nd Peltier, L. J., “Detached-Eddy Simulation of High-Reynolds
Number Beveled-Trailing-Edge Boundary Layers and Wakes,”
ASME Journal
of
Fluids
Engineering, Vol. 127, 2005, pp. 897-906.
30Baker, W. J., and Paterson, E. G. Simulation of Steady Circulation Control for the
GACC Wing,”
Applications of Circulation Control Technologies,
AIAA, Reston, VA,
2005.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 446/637
Chapter 17
Full Reynolds-Stress Modeling
of
Circulation Control Airfoils
Peter A. Chang
III,*
Joseph Slomski,* Thomas M ar h o, +Michael P. Ebert,+
and Jane Abramson*
Naval Sur ace War are Center-Carderock Division,
West Bethesda, Maryland
Nomenclature
A
=
airfoil planform area, m2
c = chord length, m
CL
=
lift coefficient; see Eq. (2)
C
=
pressure coefficient, see Eq. (3)
C = blowing rate; see
Eq.
(1)
h
=
slot height, m
k
=
turbulence kinetic energy, m2/s2
ri =
mass flow rate, kg/s
S
=
span, m
Re =
Reynolds number based on
U,,
c and v,
U , =
freestream velocity, m /s
u
v =
fluctuating horizontal and vertical velocity, respectively, m/s
u
=
friction velocity, m/s
vj
=
mean jet velocity at slot opening, m /s
x , y
=
in-plane coordinates, m
y
=
wall normal distance in viscous units; yu,/vm
a =
angle of attack, rad
=
turbulence dissipation rate, m2 /s3
r )
=
distance from wall, m
w
= specific d issipation rate, 1/s
*Propulsion and Fluid Systems Department. Member AIAA.
'Propulsion and Fluid Systems D epartment.
*Marine and Aviation D epartment (retired). Mem ber A IAA.
This material is declared a work
of
the
U.S.
overnment and is not subject to copyright protection
in the United States.
445
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 447/637
446
P.
A.
CHANG
ET
AL.
p..
= freestream fluid density, kg /m 3
T~ = wall shear stress, kg/(m . s2)
v = free stream kinematic viscosity, m 2/ s
VT
=
turbulence viscosity, m2/s
overbar) time average
I. Introduction
ECENTLY , low-speed maneuverability has become an important design
R quirement for aircraft, ships, and submarines. At low speed, the control
authority (that is, the normal, or lifting force) associated with conventional
hinged control surfaces is often insufficient to perform certain maneuvers. As a
result, designers have begun to investigate the use of circulation control (C C) air-
foils to achieve the required control authority at low speeds.
Circulation control technology has been investigated both experimentally”2
and a n a l y t i ~ a l l y ~ , ~ver the past 25 years. True CC airfoils typically have bluff
trailing edges. These airfoils employ the Coanda effect to obtain lift augmenta-
tion by tangentially ejecting (blowing) a sheet of fluid near the trailing edge
(TE) on the upper surface. Because of the Coanda effect, the je t sheet remains
attached to the bluff TE and provides a mechanism for boundary layer control
(BLC). The blowing can be thought of as a movement of the stagnation point,
producing an increase in circulation around the airfoil. Experimental results for
Coanda-type TE blowing5 have shown lift coefficient increases of as much as a
factor of four when com pared to the case of no blowing.
Because of the difficulty and expense involved in experim entally investigating
different CC configurations for parametric design studies, researchers and
designers have begun to focus on the use of computational fluid dynamics
(CFD) to analyze CC devices. Although most of the computational problem of
the CC airfoil is straightforward, complications arise in the area of the Coanda
jet itself. This jet is bounded by a curved wall on one side and a free shear
layer on the other, and contains very-high-momentum fluid. This high momen-
tum enables the jet to remain attached to the curved TE. The extent to which
the jet remains attached controls the circulation and, hence, the lift generated
by the airfoil. Thus, any computational technique, in order to be successfully
applied to the CC problem, must be able to accurately predict the spreading
rate of the jet and the location at which the Coanda jet finally separates from
the curved T E of the airfoil. To accomplish this, the computational flow solver
must be able to correctly predict the exchange of momentum between the
Coanda jet and the surrounding fluid, the entrained upstream boundary layer,
from the airfoil. Consequently, the computational mesh in the vicinity of the
jet must be fine enough to adequately resolve the boundary layer between the
wall and the jet, and the shear layer between the jet and the surrounding fluid.
In addition, the type of turbulence model chosen for the problem will be
crucial to successful modeling the Coanda jet and its interaction with the sur-
rounding fluid, and subsequent prediction of the lift force generated by the CC
airfoil.
A
recent paper6 reports good results from numerical solutions for CC airfoils
using algebraic7 and o ne equation’ eddy-viscosity turbulence models. However,
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 448/637
FULL-REYNOLDS STRESS MODELING 447
the CC for these airfoils is essentially a blown flap method, where the jet separ-
ates from a sharp, rather than bluff, TE, which fixes the separation point. The
general
CC
airfoil problem requires the jet to separate at some point along a
curved wall (the bluff TE). Figure
1
depicts the streamlines around such an
airfoil at zero degrees angle of attack and some finite free stream velocity. In
the figure, the flow is from left to right, and the jet emerges from a slot above
the curved trailing edge on the right hand side of the airfoil. The jet remains
attached to the TE for some distance before finally separating. Also, the circula-
tion increase caused by the jet has moved the leading edge (LE) stagnation point
to a position below the LE. In general, curved wall jets like those on the CC
airfoil have been problematical for simple eddy viscosity based turbulence
models to predict. Although eddy-viscosity models can often be modified to
improve their predictive accuracy for curved wall jets, these modifications are
largely ad hoc, and cannot be easily enera lized for arbitrary flows and configur-
at ion^ ^ For exam ple, Slomski et al. demonstrate that standard isotropic, two-
equation turbulence models yield nonphysical solutions for a
CC
airfoil as
blowing rate increases, whereas a full Reynolds-stress turbulence model repro-
duces the correct lift/blowin rate behavior for the same airfoil. Recently,
however, Paterson and Baker reported a successful simulation of the highest
blowing rate case reported in Slomski et al.,9 using a blended k -w / k -E SST
(shear stress transport) two-equation turbulence model.
This chapter explores the performance of the Full Reynolds Stress Model
(FRSM) for two-dimensional
CC
airfoils beyond the cases investigated in
Slomski et al.9 and Paterson and Baker. Specifically, a full range of blowing
slot heights, airfoil angles of attack, and two airfoil TE shapes are simulated.
1
Fig. 1 Typical CC airfoil showing Coanda jet and surrounding streamlines. Flow is
from left to right. The jet is depicted by the thick group of streamlines at the trailing
edge of the airfoil.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 449/637
448
P.
A.
CHANG
ET
AL.
Based on the encouraging resu lts reported in Paterson and Baker,” the perform-
ance of the
k -w / k - e
SST model in some of these new conditions is
investigated.
11. Mathematical Development
The steady, two-dimensional Navier-Stokes equations are solved using the
finite volume code, Fluent. The segregated solver, with SIMPLE pressure-
velocity coupling, is utilized. Second-order upwinding is used to discretize the
convective terms in the momentum equations with second-order central differen-
cing used on the viscous terms. First-order upwinding is used on density, energy,
k, E
and Reynolds stress equations.
The effect of turbulent flow on the steady state solution is obtained using the
FRSM of Launder, Reece and Rodi (LR R),” as well as the blended k - w / k - e
SST model. In two d imensions, the FRSM introduces an additional five equations
-three equations for each of the correlations U U U V and VV, and equations for
k
and are solved in order to evaluate at the walls. A wall reflection term is
invoked, which damps the normal stresses at the wall while enhancing the stres-
ses parallel to the wall.
Enhanced wall treatment is utilized, which solves to the wall where y 3
and uses w all functions valid in the buffer region including the effect of pressure
gradients. The wall function is important because of the wide range of velocities
over the foils, where upstream of the slot the grid has y x
1
but in the Coanda
jet, y
RZ
3-10.
Numerical simulations of airfoils with 15% thickness-to-chord ratio, 1
camber, with a slot located at 97% chord, with a 6.7% thickness at the slot
location, and tw o Coanda T E shapes5 are undertaken. B oth the “nominal” circu-
lar T E foil, NCCR 1510-7067N, shown in Fig. 2, and the logarithmic spiral TE
foil, NCCR 15 10-70678, are used. The slot-height-to-chord
( h / c )
atios include
0.0015, 0.0022, and 0.0030. Incidence ang les are 0, -4, and deg.
The logarithmic spiral curve has a constantly increasing radius of curvature
with the smallest radius at the slot.
A
comparison between the circular and
logarithmic-spiral TE geometries is shown in Fig. 3. The rationale for a
LEADING
EDGE
TRAILING
EDGE
Fig.
2
Geometry
of the NCCR 1510 7067 airfoil.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 450/637
FULL-REYNOLDS STRESS MODELING
449
I I I
0.92
0.94 0.96 0.98
1 00
X/C
Fig. 3 Comparison of circular and logarithmic-spiral TE geometry; -, circular
TE;
- -
-, ogarithm ic-spiral TE.
logarithmic spiral TE is that for a given blowing rate the Coanda jet may stay
attached a longer distance around the TE because of the decreasing curvature
where the jet would tend to detach for a circular TE. This would reduce the
power requirement necessary to obtain a given lift augmentation ratio (Rogers,
E., personal communication, March 2004). When computing the solutions for
the logarithmic spiral it was thought that the geometry had h / c
=
0.0015 and,
thus, the
C
values were set to match the
=
0.0015 cases. After the fact,
however, it was found that the geometry actually had
h / c =
0.0020. This is
between the experimental
h / c
values of 0.0015 and 0.0022. For comparison to
results, the
C
values were re-computed and the
=
0.0022 cases closest to
the actual C values are used for comparison.
The computational grids have between 100,000 and 150,000 cells, depending
on slot height. An 0-grid topology is used near the body with an H-grid in the
wake extending approximately 13 chord lengths downstream. The LE and TE
regions are shown in Figs. 3 and 4, respectively. The hybrid mesh consists of
quadrilaterals with triangular elements in the slot exit as shown Fig.
6.
On the body, boundary conditions are specified as no-slip except at the
upstream end of the plenum where rit (mass flow rate) and pressure are specified.
For the incompressible startup conditions, the upstream, outer boundary is set to
a velocity inlet condition where the freestream speed is set to 41.65 m /s ,
vT/v = 5, and k / U k
= 0.05.
Also, for the incompressible startup conditions,
the downstream boundary is set to a pressure outlet with zero pressure. When
the flow is assumed to be compressible, the air is assumed to be governed by the
ideal gas law with the Sutherland law applied to the evaluation of molecular
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 451/637
450
P. A.
CHANG ET
AL.
Fig.
4
Grid for h / c
=
0.0030 airfoil showing detail of LE.
viscosity; the outer boundaries are se t to far-field pressure with M , = 0.12 and
zero pressure. It is assumed that the freestream temperature is 288 K, with a
freestream kinematic viscosity voo
=
1.462 x lOP 5m 2/s and density,
p,
=
1.224 kg/m3. The chord length of the airfoil, c is 0.203 m, giving a free-
stream Reynolds number
Re
=
5.8
x
lo5.
In order to change the angle of
attack a the freestream velocity is rotated appropriately. A negative value of
a denotes that the nose is pitched downward. The m ass flow rate is nondimensio-
nalized as the jet mom entum coefficient
my
1/2p,u2,c
c -
Fig.
5
Grid for
h / c
= 0.0030 airfoil showing detail of C oanda jet region.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 452/637
FULL-REYNOLDS STRESS MODELING
451
Fig. 6 Grid for h / c
=
0.0030 airfoil showing detail of slot region.
where
p,
and V , are the freestream density and velocity, respectively, and c is
the airfoil chord length. The experimental riz values were measured using a cali-
brated venturi m eter that was inserted in the air supply line and the je t velocity
vj
was calculated as an isentropic expansion from duct pressure to freestream static
p r e ~ s u r e . ~
Table
1
lists the cases for the circular arc TE with case numbers corresponding
to those given in Ab ra m ~ o n .~able 2 lists the cases for the logarithmic spiral TE ,
Table
1
Circular arc
TE
runs
293
289
283
311
307
302
330
326
321
60
57
53
229
227
223
0.050
0.092
0.209
0.048
0.093
0.189
0.047
0.090
0.184
0.052
0.104
0.201
0.053
0.103
0.198
0.0030
0.0030
0.0030
0.0030
0.0030
0.0030
0.0030
0.0030
0.0030
0.0015
0.0015
0.0015
0.0022
0.0022
0.0022
0
0
0
- 4
- 4
- 4
-8
-8
-8
0
0
0
0
0
0
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 453/637
452
P.
A.
CHANG
ET
AL.
Table 2 Logarithmic spiral TE runs
case’
56
CFD
361
53
CFD
358
51
CFD
357
0.054
0.041
0.039
0.107
0.080
0.077
0.140
0.105
0.090
0.0015
0.0020
0.0022
0.0015
0.0020
0.0022
0.0015
0.0020
0.0022
showing the experimental C values for h / c = 0.0015 and h / c = 0.022, as well
as the
C
values actually run with h / c
=
0.0020.
The flow is assumed to be compressible in order to validate the wind-tunnel
experiments. Obtaining a well-converged solution is difficult because of the
large range of length and velocity scales (e.g., the ratio of the jet to freestream
velocities is as high as 6 .
Typically, the compressible RSM solutions are
obtained using a m ultistep procedure:
1) Initial Coanda jet development: Incompressible flow, k--E turbulence
model, underrelaxation factors (URFs) less than 0.2, run for several thousand
iterations.
2) Coanda jet development and prediction of approximate separation point:
Incompressible flow, FRSM turned on, URFs lowered to less than 0.1, about
5000
iterations.
3) Incorporation of compressibility effects: Compressible flow,
k -
turbu-
lence model,
URFs
less than 0.1, run for about 10,000 iterations.
4) Final jet development: Com pressible flow, FRSM , URFs less than 0.1 for
Reynolds stress equations, 0.2 for other equations, about 10,000 iterations.
5 Ensuring convergence to stable solution: Compressible flow, RSM, larger
URFs 0.3-0.5, run for 20,000-30,000 iterations.
For the solution with the k-w/k--E SST m odel this procedure is m odified:
1) Initial Coanda jet development: Incompressible flow, k--E turbulence
model, URFs less than 0 .2, run for 10,000 iterations.
2) Incorporation of compressibility effects: Compressible flow,
k -w
turbu-
lence model,
URFs
less than 0.1, run for about 5000 iterations.
3) Turn on k-w/k--E SST model: Com pressible flow, URFs less than 0.1 for
all equations, run for about 2000 iterations.
4) Ensuring convergence to stable solution: Compressible flow,
k -w
SST,
URFs increased to
0.4
run for 10,000-20,000 iterations.
Solutions are considered converged when there is no change in the integrated lift
force over 10,000-20,000 iterations. The lift forces converge to steady-state
values with no transient oscillations.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 454/637
FULL-REYNOLDS STRESS MODELING
453
111. Results
In this section qualitative aspects of the computed flows are shown, then the
integrated lift vs angle of attack curves and surface pressure distributions about
the foil are compared with experimental data.
The lift coefficient is com puted by
where Fy is the force in the y direction, pmand
Urn
are the freestream values of
density and velocity magnitude, respectively, and
A
is the reference surface area
cS,
where
c
is the chord length and
S
the
1
m span (for the two-dimensional cal-
culations, the forces are given in terms of force/unit span). The presure coeffi-
cient
C
is computed by
P
c -
1/2pwU&
(3)
A. C, Variation with h / c = 0.0030, = 0 deg
The Coanda jet changes the location of the detachment point on the trailing
edge (TE) and with it, the circulation around the airfoil.
As
can be seen in
Fig.
7,
the TE detachment point and the LE stagnation point migrate around to
the bottom of the foil as C increases. The FRSM does not predict streamlines
that wrap around to the bottom (referred to henceforth as “trailing edge pressure
drawdown”) as was shown by Slomski et al.9 for isotropic turbulence models.
The lift vs.
C
curve (Fig. S), shows that the lift is underpredicted throughout
the C range, and where the experimental curve has a small amount of curvature,
the predicted curve is almost linear. However, this is a major improvement over
the large drop in lift due to trailing edge pressure drawdown as shown in Slomski
et
a1.9
Surface pressure distributions (Fig.
9)
show that the reason why the predicted
lift is low is because of an underprediction of the midchord pressure differential.
The C = 0.092 case, which has the largest discrepancy in midchord pressure
differential, has the largest discrepancy in the predicted C . The highest C
case (Fig. 9c), matches up with the experimental pressure data very well
across the foil, but has just enough of a discrepancy in the midchord pressure
differential to cause the under prediction of C shown in Fig.
8.
B. Angle-of-Attack Variation with h / c = 0.0030
Figures 10 and 11 show the streamlines for = -4 and 8 deg, respectively.
For both cases it can be seen that at the lower blowing rate the stagnation point is
at the LE or on the upper surface. As
C
is increased, the C oanda jet induces the
stagnation point to migrate around to the bottom, in essence modifying the angle
of attack. The integrated lift coefficients for the circular TE with
h / c
= 0.0030
vs. or = 0, -4, and - 8 deg are shown in Fig. 12. The experimental data
show that as a becomes more negative, the amount of positive lift decreases.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 455/637
454
P.
A.
CHANG
ET
AL.
Fig.
7
Streamlines for
h / c
=
0.0030
and =
0
deg: Upper,
C,
=
0.050;
middle,
C,
= 0.092; bottom,
C,
= 0.209.
5
4
3
0
2
1
0
0.00 0.05
0 10
0.15
0.20
0.25
c
Fig.
8
Lift coefficient vs
C,
at
h / c
= 0.0030,
=
0
deg, comparing FRSM results to
experim ental resu lts: xperiment;
0
FRSM.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 456/637
FULL-REYNOLDS STRESS MODELING
b)
455
Fig. 9 Surface pressure distributions for h l c = 0.0030 at =
0
deg:
-,
FRSM;
0 xperiment-top; +, experiment-bottom. a) C =
0.050,
b) C = 0.093, c)
C = 0.209.
How ever, because of the additional circulation caused by the Coanda jet , nega-
tive angles of attack can still have positive lift. There is a constant difference
between the curves of constant and a slight decrease in slope with increase
in C . The computational FRSM results show similar behavior, although they
are low by about
ACL 0.5.
Figures 13 and 14 show the pressure distributions
for
h / c
=
0.0030 at - 4 deg and deg, respectively. The results are consistent
with the streamline plots (Figs. 1 0 and 11), which show that for the low
C
cases
the stagnation point is on the upper surface, migrating around to the lower surface
as the
C,
increases. In all cases , the TE pressure peak is underpredicted with the
discrepancy decreasing with increasing C and smaller angle of attack. This
seems to indicate that there is a discrepancy in the jet detachment point-that
as angle of attack becomes more negative, the predicted je t detaches relatively
earlier with respect to the experiment, but that as C increases, the je t detachment
points get closer to their experimental values.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 457/637
456
P.
A.
CHANG
ET
AL.
Fig. 10 Streamlines for h / c = 0.0030 and cu = -4 deg: Upper;
C,
= 0.048; middle,
C,
=
0.093; bottom, C, = 0.189.
Fig.
11 Streamlines for h / c = 0.0030 and cu = -8 deg: Upper: C, = 0.047; middle,
C,
= 0.090; bottom, C = 0.184.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 458/637
FULL-REYNOLDS STRESS MODELING
457
5
4
3
0
2
0
0.00 0.05 0.10 0.15 0.20 0.25
c
Fig. 12 Lift coefficient vs
C,
at
h / c
=
0.0030 for three angles of attack, comparing
FRSM results to experimental results. FRSM: -, cu =
0
deg; ---, u
=
-4 deg;
- - - cu
=
-8 deg. Experiment5 symbols: 0 cu = 0 deg; 0 cu
=
-4 deg; H
cu = -8deg .
C. Slot Height Variation with (Y
=
0 deg
Lift coefficient vs C, for three slot he ights, h / c
=
0.0015,0.0022, and
0.0030
are shown in Fig. 15. For a given
C
the product
hvj
is constant
so
that as
h / c
decreases,
vj
must increase in inverse proportion to a decrease in
h.
Thus, jet vel-
ocities for the cases with smaller slot heights and higher C, values are very close
to being supersonic. For
h / c =
0.0015 the two higher
C
cases did not converge.
The experimental results show that at the lower values of C, there is very little
change in
C L
with variation in slot height.
As
C,
increases, the
CL
for
h/c = 0.0030
falls away from the two smaller slot heights. The FRSM results
show that for the lower two values of C, as
h / c
decreases, the discrepancies
between experiment and predicted values decreases, with the smallest slot
height, h/c
=
0.0015, being right on the experimental data. For h/c
=
0.0022
the FRSM values shows the correct trend at higher C,, a decreasing slope as
C, increases. The pressure distributions for
h / c
=
0.0022 are shown in
Fig. 16. They show that as
C
increases, the peak TE pressure as well as the
overall pressure compares increasingly well with experimental data. Fig. 17,
the pressure distribution for
h / c
=
0.0015,
C,
=
0.052 shows very good com-
parison to experimental data, with only a very small underprediction of the
peak TE pressure.
These trends in the p ressure distributions indicate that for the circular TE , for
low values of I he jet detaches early com pared with experiments, resulting in
low midchord pressure differentials and lift. However, as vj increases, the jet
detachment point extends further around, eventually matching up with the exper-
imental location. In these cases, the midchord pressure differential and lift are
well predicted.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 459/637
458
a)
P.
A.
CHANG
ET
AL.
b)
Fig. 13 Surface pressure distributions for
h / c
= 0.0030 at
= -4
deg;
-,
FRSM; 0 experiment-top;
+,
experiment-bottom. a) C ,
=
0.048, b)
C , = 0.093, c) C , = 0.189.
D. Logarithmic-Spiral TE
Figure 18 shows the streamlines for the three
C
cases run on the logarithmic
spiral TE. Figure 18a shows that for the lower
C,
case the detachment point is
well predicted. However as C, increases TE pressure drawdown is predicted
as shown in Figs. 18b and 18c. Figure 19 shows that for the lower
C
case
C, = 0.041 the pressure at the lower TE is correctly predicted. However the
predicted suction side pressures are low. For the
C,
= 0.080 and
C,
= 0.105
cases the pressures on the lower side of the TE do not increase to their constant
suction-side values because of the TE pressure drawdown. The pressure ampli-
tudes at the LE are overpredicted indicating excessive circulation. These
results indicate that with the logarithmic spiral’s increasing radius of curvature
the FRSM is not sensitive enough to predict the correct detachment point.
E. Blended
k - w l k - e
SST Model
Figure 20 compares Case 283 results with the k-w/k-• SST model with
FRSM and experimental results. The k W /k E SST results are similar to
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 460/637
FULL-REYNOLDS STRESS MODELING
b)
459
Fig. 14 Surface pressure distributions for h / c = 0.0030 at = -8 deg;
-,
FRSM;
experiment-top;
+,
experiment-bottom.
a C , = 0.047,
b)
C , =
0.090,
c)
C , =
0.184.
previous computations in that they predict the Coanda je t detachment at the TE ,
rather than the T E pressure drawdown effect typical for other isotropic models.'
In this case, however, the
k-w/k--E
SST results predict low er airfoil circulation
than the experiment, as evidenced by a smaller difference in surface pressure
magnitudes between the upper and lower surfaces of the airfoil.
As
shown in
Fig. 21, this results in a lower
CL=
2.82 as compared with 4.25 from experiments
and 3.81 from FRSM . Paterson and Baker obtained a value of CL
=
4.0, using
the blended
k-w/k--E
SST model. The difference between the results reported
herein and Paterson and Baker's results may be due their use of overset gridding
which allows a finer grid in the Coanda jet region.
Pressure distributions and streamlines for the logarithmic spiral TE using the
k-w/k--E
SST model are shown in Figs. 22 and 23, respectively. Using the
k -w /
k--E SST model, the TE pressure drawdown is not as severe as for the FRSM,
with a pressure distribution on the lower side of the TE much c loser to the exper-
imental results. This generates a midchord pressure distribution much closer
to the experimental values, although the peak pressure at the TE is slightly
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 461/637
460
P.
A.
CHANG
ET
AL.
5 1
4
3
0
2
0
0.00 0.05 0.10 0.15 0.20 0.25
c
Fig. 15 Lift coefficient vs
C
at
(Y =
Odeg for three values of slot height,
hlc ,
comparing FRSM results with experimental results. FRSM:
-, h / c = 0.0030,
. . . . . .,
h / c
= 0.0022; _ _ _ _
, / c
= 0.0015. Experiment5 symbols: 0 / c =
0.0030;
0 h / c
=
0.0022;
A h / c = 0.0015.
underpredicted. Table 3 shows that the experimental C values for the
h / c = 0.0015 and h/c = 0.0022 are 3.86 and 3.62, respectively. The
FRSM
result,
CL
= 3.97 is high because of the circulation induced by the trailing
edge pressure drawdown. The
k-w/k-•
SST value, CL= 3.15, is low, consistent
with the predictions for the circular arc TE.
F. Discussion
It is difficult to say conclusively which turbu lence models are best for the CC
foil problem. The results presented in this paper have not been shown to be grid
independent, for example. H owever, the following trends are evident:
1)
Isotropic turbulence models. The M enter k W / s E SST model appear to
offer the best performance of the isotropic turbulence models. The results
herein and from Paterson and Baker bear this out. Notwithstanding Paterson
Table
3
Lift coefficients for logarithmic-
spiral case with
C,
=
0.105
case' CL
Expt. Case 56
Expt. Case 356
FRSM
k-rn1k-E SST
3.86
3.62
3.97
3.15
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 462/637
FULL-REYNOLDS STRESS MODELING
b)
461
XIC
Fig. 16 Surface pressure distributions for
h / c =
0.0022 at = 0 deg,
-,
FRSM,
0
xperiment: a)
C
=
0.053, b)
C
=
0.103, c)
C
=
0.198.
Fig. 17 Surface pressure distributions for h / c = 0.0015 at
=
0
deg for
C
= 0.052;
-,
FRSM,
0
xperiment.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 463/637
462
P.
A.
CHANG
ET
AL.
b)
Fig. 18 Streamlines
for
logarithmic-spiralTE: a)
C,
= 0.041, b)
C,
=
0.080
and c)
C,
=
0.105.
a)
Fig. 19 Surface pressure distributions for logarithmic spiral cases: -, FRSM-
h / c
=
0.0020,
0 experiment-h/c
=
0.0015; +, experiment-h/c =
0.0020,
a)
C = 0.041, b) C
= 0.080,
c) C
=
0.105.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 464/637
FULL-REYNOLDS STRESS MODELING
-20
463
-1 5
-5
0
0.0 0.2
0.4
0.6 0.8 1.0
xlc
Fig. 20 Surface pressure distributions comparing turbulence models for circular
TE Case 283 h / c
=
0.0030, = 0 deg, C, = 0.209); -, FRSM: A
k -m
SST; 0
experiment.
Fig. 21
Lift coefficient vs C at =
0
deg,
h / c
= 0.0030, comparing FRSM, k -m ,
and experimental results: -, experiment5;0 RSM;
A
k-w SST.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 465/637
464
P.
A.
CHANG
ET
AL.
20
-1 5
-1 0
0
-5
0
0.0 0.2
0.4 0.6 0.8
1.0
X/C
Fig. 22 Surface pressure distributions comparing turbulence models for
logarithmic-spiral TE,
C
= 0.105: FRSM; A k - o SST, 0 xperiment-h/
c
=
0.0015
(Case
51);
0
experiment-h/c
=
0.0022 (Case 356).
Fig. 23 Stream lines for logarithm ic-spiral TE using k - o turbulence model (Case
51, C,
= 0.105 .
and Baker's'' use of overset meshes,
it
is generally accepted that the Menter
k-
W / S - - E SST model provides superior near-wall behavior (this model transitions to
k -w in the near-wall region). The improved near-wall behavior over the
k--E
model may well do a better job of modeling the physics of the turbulent
Coanda wall jet.
2) FRSM . These m odels appear to be better-suited for application to general
CC foil problems. M esh refinement studies are needed to explore fully the per-
formance of these models, however. In addition, only the LRR FRSM was
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 466/637
FULL-REYNOLDS STRESS MODELING
465
exploited. There are other FRSM variants, such as the Launder-Shima12 FRSM ,
which are known to be less dissipative. Such models may offer improved predic-
tive performance.
IV. Conclusions
An extensive series of RANS calculations have been performed on two-
dimensional CC airfoils with circular arc and logarithmic-spiral TEs. It is
shown that for a circular-arc TE, the full Reynolds stress turbulence closure
can predict the Coanda jet detachment point fairly well for a range of angles
of attack, jet slot heights, and jet blowing coefficients. For most cases the lift
is low in comparison to experimental values. However, the trends in lift due to
angle of attack and jet blowing coefficient are correctly predicted. The logarith-
mic-spiral T E is a much more challenging case; for higher jet blowing rates, the
Coanda jet detaches upstream on the pressure (lower) side of the airfoil and the
lift is overpredicted. For lower blowing rates, however, the correct detachment
point is predicted. The
k -w /k -E
SST model is successful in predicting the
detachment point for the circular TE, higher
C
case, and in addition, is able
to come closer to predicting the correct detachment point for the highest C log-
arithmic-spiral case.
Acknowledgments
This work was performed at the Naval Surface Warfare Center-Carderock
Division (NSWCCD), West Bethesda, Maryland. It was sponsored by the
Office of Naval Research, (Ronald D. Joslin, program manager) under work
units 03-1-5400-616 and 04-1-5400-616. Computations were supported by a
grant of High Performance Computing (HPC) time from the Department of
Defense (DoD ) HPC Shared Resource Centers, the U.S. ir Force’s Aeronautical
Systems Center at Wright-Patterson Air Force Base, Ohio (Origin 3900, hpc-
11), and the
U.S.
Army’s Research Laboratory at Aberdeen Proving Ground,
MD (IBM SP-4). The advice of Ernest Rogers is appreciated and duly noted.
References
‘Englar, R., and Huson,
G.
Development of Advanced Circulation Control Wing High
Lift Airfoils,” AIAA Aerospace Sciences Meeting, AIAA Paper 83-1847, Jan. 1983.
’Englar, R., Smith, M., Kelley,
S.,
and Rover, R., “Development of Circulation Control
Technology for Application to Advanced Subsonic Aircraft,” AIAA Aerospace Sciences
Meeting, AIAA Paper 93-0644, Jan. 1993.
3Shrewsbury,
G.
Analysis of Circulation Control Airfoils Using an Implicit Navier-
Stokes Solver,” AIAA Aerospace Sciences Meeting, AIAA Paper 85-0171, Jan. 1985.
4Shrewsbury, G., “Dynamic Stall of Circulation Control Airfoils,” Ph.D. Dissertation,
Aviation and Surface Effects Department, Georgia Inst. of Technology, Atlanta, GA,
Sept. 1990.
’Abramson, J., “Two-Dimensional Subsonic Wind Tunnel Evaluation of Tw o Related
Cambered 15-Percent-Thick Circulation Control Airfoils,” Tech. Rept. ASED-373,
DTNSRDC, Sept. 1977.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 467/637
466 P.
A.
CHANG
ET
AL.
Liu, Y., Sankar, L., Englar, R., and Ahuja,
K.,
“Numerical Sim ulations of the Steady
and Unsteady Aerodynamic Characteristics
of
a Circulation Control Wing Airfoil,” 39th
AIAA Aerospace Sciences Meeting, AIAA Paper 2001-0704, Jan. 2001.
’Baldwin, B. and Lomax, H., “Thin Layer Approximation and Algebraic Model for
Separated Turbulent Flows ” AIAA Aerospace Sciences Meeting, AIAA Paper 78-0257,
Jan. 1978.
‘Spalart, P., and Allmaras, S., “A One-Equation Turbulence Model for Aerodynamic
Flows ” AIAA Paper 92-0439, Jan. 1992.
’Slomski, J. F.
Gorski,
J. J., Miller, R. W., and Marino,
T.
A., “Numerical Simulation
of
Circulation Control Airfoils as A ffected by Different Turbulence M odels,” 40th AIAA
Aerospace Sciences Meeting Exhibit, AIAA Paper 2002-0851, Jan. 2002.
“Paterson, E. G. nd Baker, W. J., “Simulation of Steady Circulation Control for
Marine-Vehicle Control Surfaces,” 42nd AIAA Aerospace Sciences Meeting, AIAA
Paper 2004-0748, Jan. 2004.
“Launder, B., Reece,
G.
nd Rodi, W., “Progress in the Development of a Reynolds-
Stress Turbulence Closure,”
Journal
of
Fluid Mechanics,
Vol. 68, No. 3, 1975, pp.
”Launder, B. and Shim a, N., “Second-M oment Closure for the Near-W all Sublayer:
6
537-566.
Development and Application,”
AIM Journal,
Vol. 27, No. 10, 1989, pp. 1319-1325.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 468/637
1II.B. Tools for Predicting Circulation Control
Performance:
NCCR 103RE
Airfoil Test Case
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 469/637
Chapter 18
Aspects of Numerical Simulation of Circulation
Control Airfoils
R. Charles Swanson,* Christopher
L.
Rumsey,+ and Scott
G.
Anders'
NASA
Langley Research Center, Hampton, Virginia
Nomenclature
A = planform area, ft2
a
=
speed of sound, ft/s
b =
wing span, ft
C
= section drag coefficient,
D / ( q A )
C= surface skin friction coefficient, T w q o o
CL
=
section lift coefficient,
L / ( q . d )
C
=
pressure coefficient, (p oo) /qoo
C =jet momentum coefficient, ( h V , ) / ( q , A )
cr3
= parameter for curvature effects
c = chord length, in.
h
= slot height, in.
k = turbulent kinetic energy per unit mass, ft . b/slug
M = Mach num ber, V / a
m = mass flow rate, slug/s
p
=
pressure, lb/ft2
q
=
dynamic pressure, p V 2 , lb/f t2
R
=
gas constant, ft . bfslug .OR
Re
= Reynolds number,
( pV,c) /p
T = Temperatu re, OR
u,
=
friction velocity,
m,
t /s
V = velocity, ft/s
u
v = Cartesian velocity components, ft/s
*Senior Research Scientist, Computational AeroSciences Branch, Senior Member AIAA.
'Senior Research Scientist, Computational AeroSciences Branch, Associate Fellow AIAA.
'Research Engineer, Flow Physics and Control Branch, Senior Member AIAA.
Copyright 005 by the American Institute of Aeronautics and Astronautics, Inc.
No
copyright is
asserted in the United States under Title 17, U.S. Code. The U.S. Government has a royalty-free
license to exercise all rights under the copyright claimed herein for Governmental purposes. All
other rights a re reserved by the copyright owner.
469
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 470/637
470
R.
C. SWANSON, C. L. RUMSEY, AND S. G. ANDERS
x,y = Cartesian coordinates, in.
y =
normalized coordinate,
y ( u 7 / v )
a =
angle of attack, deg
y
=
specific heat ratio
E
= dissipation rate of k, ft . b/slug/s
p = coefficient of viscosity, lb .s/ft
v
=
kinematic viscosity, ft2/s
p =
density, slug /ft3
T
=
shear stress, lb/ft2
w
= specific dissipation rate of k , k / v t , s-'
Subscripts
c
=
based on chord length
exp = refers to experiment
j =jet condition
ref = reference 00 ) condition
t = turbulent flow quantity
w =
solid surface (wall) condition
0 = total condition
00
=
freestream quantity
I. Introduction
ONVENTIONAL high-lift sys tems use slats and flaps to create the necess-
C ry airfoil camber to achieve the desired circulation, and thus lift. There is a
weight penalty and increased maintenance associated with these systems. For a
number of years,' aerodynamicists have been seeking alternative high-lift
systems that can reduce the weight and complexity of the conventional
systems. One such system for circulation control (CC) involves the Coanda
effect. By controlling a jet discharged from a slot on the upper surface of the
airfoil, the trailing edge (TE) stagnation point is moved toward the lower
surface on a rounded TE, and the leading edge (LE) stagnation point is moved
toward the lower surface as well. In this way the effective camber of the
airfoil can be increased, resulting in the augmentation of lift. Previously, the
weight and operational requirements of such systems have been unacceptable.
The potential benefits of these CC systems in terms of reduced takeoff and
landing speeds as well as increased maneuverability have encouraged aerodyna-
micists to reconsider such systems. Moreover, the benefits of using pulsed jets
offer the genuine possibility of significantly mitigating the obstacles preventing
the implementation of these CC systems.2
Computational methods will play a vital role in designing effective CC con-
figurations. Certainly, detailed experimental data, such as velocity profiles and
Reynolds stresses, will be absolutely essential for validating these prediction
tools. Because of the cost of flow control experiments, design and parametric
studies will strongly depend on accurate and efficient prediction methods.
These methods must have the potential to treat pulsating jets, even multiple
jets, for a broad range of flow conditions (e.g., Mach number, Reynolds
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 471/637
NUMERICAL SIMULATION OF CC AIRFOILS 471
number, angle of attack). In general, the numerical methods must be extendable
to time-dependent and three-dimensional flows.
A number of computational methods have been applied to CC airfoil flows. In
1985 Pulliam et al.3 used ARC2D: an implicit Navier-Stokes solver, to compute
solutions for two of the CC configurations tested by Abramson and R ~ g e r s . ~
spiral grid that begins in the plenum and wraps around the airfoil several times
was used for the computations. Turbu lence modeling of the flow over the airfoil
and Coanda surface was carried out by applying a modified form of the zero-
equation model of Baldwin and lo ma^.^ A term was introduced in the model to
account for streamline curvature effects. The modification includes a constant
C,. This constant was modified for each set of experimental conditions, and a
set is defined by Coanda geometry, freestream Mach number, angle of attack,
and a range of je t momentum coefficient
C,.
The
C
was adjusted
so
that the com -
puted C, matched the experimental value for one of the C, values. Then this C,
was used in computing all of the cases for the given set of conditions. Certainly,
this approach is not satisfactory in general for modeling the turbulence. Neverthe-
less, Pulliam et a l. were ab le to obtain good comparisons with experimental data
for all cases considered. This work demonstrated that accurate Navier-Stokes
simulation of C C airfoil flows is possible, and turbulence modeling is the key issue.
In 2002 S lomski et a1.' considered the effects of turbulence modeling on the
prediction of CC airfoil flows. Calculations were performed for the NCCR
1510-7067 airfoil, which is a cambered, 15% thick, CC airfoil with a jet slot
located on the upper surface just upstream of the TE. The airfoil was at 0 deg
angle of attack. Two variations of a two-equation transport model
( k - ~
odel)
and a Reynolds stress model were used for modeling turbulence. Predictions of
surface pressures with the two-equation model compared favorably with the
experimental data at low blowing rates. At high rates of blowing only the
Reynolds stress model provided predictions that compared well with the data.
A principal conclusion of Slomski et al. is that nonisotropic turbulence models
are probably required for the simulation of CC airfoils or lifting surfaces.
Recently, Paterson and Baker' used an incompressible Navier-Stokes code to
calculate the flow over the same CC airfoil considered by Slomski et al. They
obtained solutions for the high blowing rate case that Slomski et al. computed
and a case with the same freestream conditions but an a of deg. The shear
stress transport (SST) model of Menter was used to model turbulence . Using
this isotropic turbulence model, their predicted surface pressure distributions
compared favorably with experiment, even though an incom pressible simulation
was perform ed. However, it should be pointed out that the variation in the ratio of
the jet density to the freestream density for the
a
of zero degree case can vary
roughly from 0.8 to 1.2. Thus, there are compressibility effects, and these may
be quite important when attempting to predict the characteristics of the jet.
In the current work various aspects of simulating CC airfoil flows are exam-
ined. These aspects include 1) flow conditions, 2) grid density, and
3)
turbulence
modeling. The primary purpose of this paper is to provide some gu idelines for
accurate solutions and to delineate improvem ents needed in numerical techniques
to reliably predict CC flows, eventually including pulsed jets. The two-dimen-
sional, compressible, mass-averaged Navier-Stokes equations are solved with
a finite-volume approach for discretization. The equations are solved on a multi-
block, patched grid, and a multigrid method with an implicit approximate
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 472/637
472
R.
C. SWANSON, C. L. RUMSEY, AND S. G. ANDERS
factorization scheme is used to integrate the equations. Numerical solutions are
obtained for flow over the CC geometry tested by Abramson and Rogers.’
Several turbulence models are considered, including models based on one trans-
port equation and two transport equations. A two-equation explicit algebraic
Reynolds stress model is also considered. The influence of turbulence modeling
is revealed by comparing computed and experimental pressure distributions, as
well as Coanda jet streamlines.
The initial sections of this chapter concern the CC airfoil geometry and flow
conditions, description of grids, numerical method, and boundary conditions.
This is followed by a section on turbulence modeling, where particular emphasis
is given to modifications introduced into the models, and also, implementation
details of the models that can significantly affect their performance. In the
final sections the numerical results are discussed and concluding rem arks are given.
11. Geometry and G rid
The CC geometry for the 2004 Circulation Control W orkshop” held at NASA
Langley Research Center is the CC elliptical airfoil, which is designated NCCR
1510-7067 N. This airfoil has a chord of 8 in., thickness ratio of 15% , and a
camber ratio of 1%. The jet slot height-to-chord ratio is 0.0030, which corre-
sponds to a slot height of 0.024 in.
Previously, we performed calculations for the CC airfoil that was tested by
Abramson and R ogers5 (see also W ilkerson and Montana6). This airfoil, which
is designated as 103RE (and also referred to as 103XW in the literature), has a
chord of 18 in., thickness ratio of 16% , and a cam ber ratio of 1% . The jet slot
height-to-chord ratio is 0.0021, which corresponds to a slot height of
0.0378 in. This CC airfoil is compared with the NCCR 1510-7067 N airfoil in
Fig. 1. The most significant differences between the two configurations are the
0
.~......................................................................................................................
0 3 j
..................
CCR
coarlirrates ...................
:I;
T
----- 103RE coarlimtes
0
. ;
.....................
~
...............................................
~
..............................................
;
0.1
0
- 0 . 1
-0.2
- 0 . 3
~
.....................
~
...............
........,........
............... ................
.......,
......................
:
; ------ i ----
......................
~
.......................
~
........ ...~......... ................
.......,
..........
...
.....
; --
I - - - -
_ _ _ - -
-;- ---- -
L - - -
-_ -
1 -
......................................................................................................................
~
.....................
~
...............
........,........
............... ................
.......,
......................
:
; ------ i ----
......................
~
.......................
~
........ ...~......... ................
.......,
..........
...
.....
; --
I - - - -
_ _ _ - -
-;- ---- -
L - - -
-_ -
1 -
......................................................................................................................
1
i
1
.2 0.4
0.6
0 8 1
0 . 4
x l c
Fig.
1
Geom etry of airfoils.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 473/637
NUMERICAL SIMULATION OF CC AIRFOILS
473
size of the plenum and the je t slot height. Because the computational grid for the
103RE airfoil was available, and this geometry is quite similar to the one of the
workshop, we elected to use the 103RE airfoil in simulating the workshop cases.
In order to compute solutions for the workshop cases, we applied the freestream
conditions for these cases and matched the corresponding jet momentum
coefficients.
The coordinates defining the 103RE airfoil were provided by E. Rogers of the
Naval Surface Warfare Center, Carderock Division (NSWCCD), and they are
given in the Appendix of this chapter. These coordinates include the changes
in the airfoil geometry caused when setting the jet slot height.
In this chapter we consider CC airfoil flows for high and low freestream
Mach numbers. The designated case numbers, which are associated with the exper-
iments, and the flow conditions are given in Table
1.
In addition to these primary
cases, others at M ,
=
0.12 and
a
=
0 deg are computed at different C levels.
The definition of C is given in the nomenclature, and some discussion of C is
given in a later section. For Case 302 the testing was done by Abramson and
Rogers? and for Cases 283 and 321, the experimental data were obtained by Abram-
son.12 Surface pressure distributions are available from the experiments. There are
no velocity profiles or Reynolds stresses to allow a detailed assessment of turbu-
lence models. Nevertheless, pressure data provide an opportunity for initial evalu-
ation of the models. The experimental lift coefficients were determined by
integrating the surface pressures, and the drag coefficients were computed from
wake survey data using a momentum deficit method. Thus, the experimental drag
values include the propulsion effects due to the Coanda jet . There are no data avail-
able specifying the error bounds of the aerodynamic coefficients. Several sources
of
error in the experimental data w ere reported by Abramson. 2 Although the exper-
iments were generally two-dimensional, there were three-dimensional effects pro-
duced at the high blowing rates. Also, there were changes in the slot height
caused by the higher pressures required for the high blowing rates. We have not
accounted for these effects on the experimental data.
For the numerical computations the domain surrounding the CC airfoil
extended 20 chords away from the airfoil. This domain was partitioned with
three blocks. At the interface boundary on the lower airfoil surface the grid is
patched, as seen in Fig. 2, which displays the near-field of a medium-resolution
grid with a total of 17,875 points. This grid includes 235 grid points around the
entire airfoil and 49 points in the normal direction over the forward part of the
airfoil. Over the aft part of the airfoil there are 101 points in the normal direction,
and this number includes the points in the plenum for the jet. For the fine grid
the number of cells in the medium grid is doubled in each coordinate direction,
Table 1 Flow conditions for
CC
airfoil flow
Case M ,
Re, a
eg
CP
302
0.6 5.2 x lo6 0 0.0032
283 0.12 5.45 lo5 0 0.2090
321 0.12 5.45 lo5
-8 0.1840
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 474/637
474
R.
C. SWANSON, C. L. RUMSEY, AND S. G. ANDERS
0
0.5
0.4
0.3
0.2
0.1
u
s o
-0.1
-0 .2
-0 . 3
-0.4
-0.5
0.5 1
0.5
0.4
0.3
0.2
0.1
0
- 0 .1
-0 .2
- 0 . 3
- 0 . 4
-0.5
0 0.5 1
x l c
Fig.
2
Near field
of
medium grid for
CC
airfoil.
resulting in
70,563
points. The clustering of the grid at the airfoil LE and je t slot is
clearly seen in Figs. 3 and 4. In the normal direction the grid is clustered at the
surface so that the normalized distance y+ is less than one for the first point
off
the wall.
0.1
0.08
0.06
0.04
0.02
g o
-0.02
-0.04
-0.06
-0.08
-0 .1
-0.05
0
0.05
0.1
0.1
0.08
0.06
0.04
0.02
0
-0.02
-0.04
-0.06
-0.08
-0.1
-0.05
0
0.05 0.1
XlC
Fig.
3
Leading-edge region of m edium grid for
CC
airfoil.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 475/637
NUMERICAL SIMULATION OF CC AIRFOILS
475
0.95 1 1.05
0.08 0.08
0.06 0.06
0.04 0.04
0.02
0
0
0.02
0
-0.02 -0.02
-0.04 -0.04
-0.06 -0.06
XlC
0.95 1 1.05
Fig.
4
Trailing-edge region of medium grid for
CC
airfoil.
111. Numerical Method
Numerical solutions were computed with CFL3D, a m ultizone mass-averaged
Navier-Stokes code developed at NASA Langley. l 3 It solves the thin-layer form
of the Navier-Stokes equations in each of the (selected) coordinate directions. It
can use one-to-one, patched, or overset grids, and employs local time-step
scaling, grid sequencing, and multigrid to accelerate convergence to steady
state. In time-accurate mode, CFL3D has the option to employ dual-time stepping
with subiterations and multigrid, and it achieves second-order temporal accu racy.
Thus, this code has sufficient flexibility to solve either two-dimensional or three-
dimensional problems with multiple and/or pulsating jets.
The code CFL3D is based on a finite-volume method. The convective terms
are approximated with third-order upwind-biased spatial differencing, and both
the pressure and viscous terms are discretized with second-order central differen-
cing. The discrete scheme is globally second-order spatially accurate. The flux
difference-splitting (FDS) method of Roe is employed to obtain fluxes at the
cell faces. Advancement in time is accomplished with an implicit approximate
factorization method (number of factors determined by number of dimensions).
In CFL3D, the turbulence models are implemented uncoupled from the mean-
flow equations. The turbulent transport equations are solved with the same
implicit approximate factorization approach used for the flow equations. The
advection terms are discretized with first-order upwind differencing. The pro-
duction source term is treated explicitly, while the advection, destruction, and dif-
fusion terms are treated implicitly. For the explicit algebraic Reynolds stress
(EASM-ko) model, the nonlinear terms are added to the Navier-Stokes
equations explicitly.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 476/637
476
R.
C. SWANSON, C. L. RUMSEY, AND S. G. ANDERS
IV. Boundary and Initial Conditions
Boundary conditions are required at the inflow (internal and external),
outflow, and solid surface boundaries. For numerical computations the physical
boundary conditions must be supplemented with numerical boundary conditions,
which generally involve extrapolation of flow quantities or combinations of them
(e.g., Riemann invariants) from the interior of the domain. Discussion of the
numerical boundary conditions is given in the user’s manual for CFL3D.13 At
the far-field inflow boundary a Riemann invariant, entropy, and flow inclination
angle are specified. A Riemann invariant is specified at the far-field outflow
boundary. For the plenum the mass flow rate and flow inclination angle are pre-
scribed. If the mass flow rate is not known from the experiment, it is determined
with an iterative process where it is changed until the experimental
C
at the jet
exit is matched. At the surface boundaries the no-slip and adiabatic wall con-
ditions are specified. Boundary conditions for the various turbulence models con-
sidered herein are given in the CFL3D user’s manual. The initial solution is
defined by the freestream conditions.
V. Turbulence Modeling
Several turbulence models for computing C C airfoil flows are considered. The
three principal models are the one-equation Spalart-Allm aras (SA) model,14 the
Spalart- Allmaras rotation/curvature (SARC) and the two-equation
shear-stress transport (SST) m odel of M enter’0”79’8. In addition, the zero-
equation Baldwin-Lom ax (BL) model’ and the explicit algebraic stress
(EASM ) model in k-w form (EASM -ko)” are used. The three primary models
and the BL model are all linear eddy-viscosity models that make use of the Bous-
sinesq eddy-viscosity hypothesis, whereas the EASM-ko model is a nonlinear
model. The equations describing these four models can be found in their respect-
ive references. H owever, there are certain de tails concerning the implementation
of the SARC and SST models that should be given here in order to facilitate the
discussion of the numerical results.
The SA model can be written in general form as
where V vt, nd
P ,
Vaiff, nd Ddiss are the contributions associated with turbu-
lence resulting from production, diffusion, and dissipation, respectively. The pro-
duction term is given by
P = C&l[l
523WV
(2)
In the SARC model P is replaced by
..*
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 477/637
NUMERICAL SIMULATION OF CC AIRFOILS
477
where the function
r*
is the ratio of scalar measure of strain rate to the scalar
measure of rotation, the function 7 depends on the Lagrangian derivative of the
strain-rate tensor principal axes angle (see Ref. 16 for details), and
rl
=
1
cr2
=
12, and
cr
=
0.6-1.0. As
cr3
is increased, the turbulence pro-
duction decreases near convex surfaces. Later, we will exploit this behavior to
reduce the production of turbulence in the Coanda flow and, in so doing,
explore its local and global effects.
The production term Pk in the turbulent k inetic energy equation of the Menter
SST model can be written as
where the stress tensor
TU
is defined as
and the partial derivatives are strain rates. The production term P n the w
equation of the SST model is proportional to Pk. Generally, in the com putations
with the SST model, the incompressible assumption is imposed, and the turbulent
kinetic energy contribution is neglected. Thus,
where Sij is the strain-rate tensor, and S,S, represents the double dot product of
two tensors. When the strain-rate tensor is used for
Pk,
the SST model will be
designated SST(1994). In some versions of the SST model, also referenced as
SST(base1ine) model herein, the vorticity i s substituted for the strain rate. l In
this case the production term is written as
Pk =
2ptWijwij
=
( U t l f l 2 (8)
where is the magnitude of the vorticity vector. The vorticity is used with the
default SST model in the CFL3D code. Certainly, one would not expect much
difference in boundary-layer-type flows between using strain rate or vorticity
in the production terms.
The eddy viscosity determined with the SST model is defined as
a1
k
max (a1w;RF2
vt =
(9)
where a l is a constant, w is equal to the ratio of the turbulent dissipation rate to
the turbulent kinetic energy, R =
/m
nd
F2
is a blending function. In a
recent paper by Menter et a1.,20 the R in Eq.
(8)
is replaced by S
=Jm
n
the default SST model in CFL3D the
R
is used. Attempts to use S instead of
R
in this work resulted in nonphysical behavior of the solution for high
blowing rates.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 478/637
478
R. C.
SWANSON,
C.
L. RUMSEY, AND S.
G.
ANDERS
VI. Jet Momentum Coefficient
A frequently used parameter in assessing the performance of C C devices is the
je t momentum coefficient. This parameter is defined as
l i j v j =
pjvj2hb
c =
,A
p,Vicb
where usually l i j is a measured quantity. In this definition the jet velocity
vj
is
determined by isentropically expanding the plenum flow to the freestream
static pressure. Thus, vj can be calculated from
In addition, C can be rewritten as
If we assume fixed
h/c
and jet conditions,
Then for
M ,
= 0.12 and
M ,
=
0.6
(two freestream M ach numbers considered in
this chapter)
Thus, for a given C with
M ,
= 0.12, the C corresponding to
M ,
= 0.6 is m ore
than an order of magnitude smaller. One must keep this behavior in mind when
considering C as M , increases.
VII. Num erical Results
The computational method described in previous sections was applied first to
the high-Mach-number flow over the CC airfoil 103RE, which is Case 302 in
Table
1.
Calculations were performed on the medium grid. A comparison of
the surface pressure distributions computed with the BL, SA, SST(baseline),
and the anisotropic EASM-ko models is shown in Fig.
5 .
There is a significant
discrepancy between the calculated and experim ental5 pressures for all of the
turbulence models. Moreover, the predicted lift coefficient is about two times
the experimental C of 0.191 for all models. Because all of the models predict
separation on the Coanda surface downstream of the location indicated by the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 479/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 480/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 481/637
NUMERICAL SIMULATION OF CC AIRFOILS
481
0.968 0.969 0.97 0.971 0.972
0.036
0.035
0.034
0.033
0.032
0.968 0.969 0.97 0.971 0.972
X/C
Fig.
8
Velocity vectors near jet exit computed with SARC model and
cr3
= 9.6
(Moo= 0.6, (Y =
0
deg, Re, = 5.2 X
lo6,
C =
0.0032,
medium grid).
the jet , but still upstream of the TE. This delay in separation results in one of the
vortices normally appearing in the blunt T E region being eliminated.
In the subsequent discussion we consider results for the same airfoil at low
Mach number ( M ,
=
0.12), with several different blowing coefficients. For the
0.85 0.9 0.95 1 1.05 1.1
0.1 0.1
0.05
P
*
0
0.05
0
-0.05 -0.05
-0.1
1 1.05 1.1
-0.1
0.85 0.9 0.95
X/C
Fig. 9 Mach contours at TE computed with SARC model and cr3 = 9.6 (Moo= 0.6,
(Y =
0 deg, Re,
=
5.2
X
lo6,
C
=
0.0032,
medium grid).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 482/637
482
R. C.
SWANSON,
C.
L. RUMSEY, AND S.
G.
ANDERS
0.98 1 1.02 1.04 1.06 1.08
0.06 0.06
0.04 0.04
0.02 0.02
P
*
0 0
-0.02
-0.02
-0.04 -0.04
0.98 1 1.02 1.04 1.06 1.08
XlC
Fig. 10 Streamline pattern at TE computed with SARC model and cr3
=
9.6
( M ,
= 0.6, a
= 0 deg,
Re,
= 5.2 X lo6,
C , =
0.0032, medium grid).
first group of cases, solutions were obtained on the medium grid with the
SA,
SARC(c,3
=
9.6), and SST(base1ine) turbulence models for various C values.
Comparisons are made in Fig.
11
between the computed and experimental
pressure distributions for C
= 0.026.
With the
SA
model there is significant
disagreement with the data on the lower and upper surfaces of the airfoil.
-4
-3
-2
0 -1
0
1
2 o
0.2 0.4 0.6 0.8 1
X/C
Fig. 11 Surface pressures computed with SA, SARC cr3
=
9.6), and SST turbulence
models (Moo= 0.12,
a
= 0 deg,
Re, =
5.45
X
lo5,
C , =
0.026, medium grid).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 483/637
NUMERICAL SIMULATION OF CC AIRFOILS
483
0.85 0.9 0.95 1 1.05 1.1
0.1 0.1
0.05
P
*
0
0.05
0
-0.05 -0.05
-0.1 -0.1
0.85 0.9 0.95 1 1.05 1.1
XlC
Fig. 12 Jet streamlines computed with SARC cr3= 9.6) turbulence model
( M , =
0.12,
a = 0
deg, Re,
= 5.45 X
lo5,
C , =
0.026, medium grid).
There is improvement in the agreement w ith the SST(base1ine) model. T he
solution with the SAR C model and
c,3 =
9.6 exhibits relatively good agreement
with the data. Figure
12
shows the Coanda jet streamlines for the
SARC(c,3 = 9.6) model. The vortex pair usually occurring behind the blunt
TE is conspicuously absent.
0
2500 5000 7500
-4
5
-5 ...................................
-4.5
-5
-5.5
-6
-6.5
-7
-7.5
-8
Fig. 13 Residual histories with
SA
turbulence model, without and with
preconditioning ( M , = 0.12, (Y =
0
deg,
Re,
= 5.45 X lo5, C , = 0.026, medium grid).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 484/637
484
R.
C. SWANSON, C. L. RUMSEY, AND S. G. ANDERS
T o provide some indication of convergence behavior of the computations, the
variation with multigrid cycles in the L2 norm of the residual (for density
equation) is presented in Fig. 13. Roughly 7500 cycles are required to reduce
the residual four orders of magnitude. A major contribution to this slow conver-
gence is the slowly converging plenum solution, which is a consequence of the
very low-s eed flow in the plenum. The implem entation of low-speed precondi-
tioning,21- especially in the plenum, should result in a significant acceleration
of convergence. Recently, we tested preconditioning for this particular case.
Without any attempt to optimize the performance of the preconditioning, the
number of cycles required to attain the same level of convergence obtained pre-
viously was reduced by a factor of two. It should be mentioned that the need for
preconditionin to achieve accurate solutions in very low-speed regions has been
demonstrated.
In Fig. 14 the computed pressures when C
=
0.093 are shown. Generally, the
trends described for C
=
0.026 are exhibited here as well. For this case, sol-
utions with both the S A and SST(base1ine) models indicate je t wraparound
(i.e., Coanda jet moves onto the lower surface of the airfoil), as supported by
the reduced pressures on the airfoil lower surface. These reduced pressures are
associated with the occurrence of recirculation. The jet wraparound with the
SA model is seen in Fig. 15. With the SARC(c,3 = 9.6) model there is generally
good agreement with the data. However, a thin separation region (about 0.01
chord in maximum thickness) occurs just downstream of the airfoil
LE.
This sep-
aration results in a barely discernible plateauing effect on the calculated pressures
in Fig. 14, which i s not consistent with the experimental data. Figure 16 shows the
je t streamlines for the SARC model and the stronger jet penetration (relative to
that in Fig. 12) into the flowfield because of the increased C .
2
25
-1 0
-8
-6
-4
-2
0
0
2
0.2 0.4 0.6 0.8
1
X/C
4 0
Fig. 14 Surface pressures computed
with
SA, SARC cn = 9.6), and SST urbulence
models
( M , =
0.12, (Y
=
0 deg, Re, = 5.45 X lo5,
C =
0.093, medium grid).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 485/637
NUMERICAL SIMULATION OF CC AIRFOILS 485
0.2
0.6
0.7
0.8
0.9 1 1.1
0.2
0.1 0.1
0 0
s
0.1 -0.1
-0.2 -0.2
-0.3 -0.3
-0.4 -0.4
X/C
0.6
0.7 0.8 0.9 1 1.1
Fig. 15 Jet streamlines computed with SA turbulence model ( M , = 0.12,
(Y =
0 deg,
Re, = 5.45 X lo5,
C
= 0.093, medium grid).
The final two cases, Case 283 and Case 321, are those considered in the
2004 Circulation Control Workshop held at NASA Langley Research Center.
Flow conditions for these cases are given in Table
1.
For Case 283, where
C
=
0.209, the computed pressure distributions on the medium grid are
0.85 0.9 0.95 1 1.05 1.1
0.1
0.1
0.05
Y
>,
0
0.05
0
-0.05 -0.05
-0.1 -0.1
0.85
0.9
0.95 1 1.05 1.1
XlC
Fig. 16 Jet streamlines computed with
SARC cr3 =
9.6) turbulence model
( M ,
= 0.12,
a
=
0
deg,
Re,
= 5.45
X
lo5,
C
= 0.093, medium grid).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 486/637
486
R.
C. SWANSON, C. L. RUMSEY, AND S. G. ANDERS
-1 0
-8
-6
-4
0
2
4
0 0.2 0.4 0.6
X/C
0.8
1
Fig. 17 Surface pressures computed with SA, SARC cr3= 9.6), SARC cr3=
0
- 9.6),
and
SST
turbulence models
(Moo=
0.12,
(Y
=
0
deg,
Re,
=
5.45 X
lo5, C
=
0.209,
medium grid).
compared with the experimental data in Fig. 17. There is considerable reduction
in the com puted lower surface pressures with the SA and SST(base1ine) models
relative to the experimental values. Such behavior indicates extensive flow
separation on the lower surface with these models. In fact, the Coanda je t in
these cases wraps around the TE and moves even further upstream than shown
in Fig. 15, a completely unphysical situation. The result with the
SARC(cr3 = 9.6) model exhibits fairly good agreement with the data on the
lower airfoil surface, but it shows a plateau behavior over more than
50%
of
the airfoil on the upper surface. Thus, there is a large separation bubble on the
upper surface. Numerical tests confirmed that this is a consequence of the
large cr3 value being used for the SARC model in the airfoil LE region. By
simply setting cr3 = 9.6 on the Coanda surface and taking it to be zero elsewhere,
relatively good agreement with the data is again obtained for the
SARC(cr3 =
0
9.6) model.
The jet streamlines for the SARC (cr3
=
0-9.6) model on the fine grid are pre-
sented in Fig. 18. In the Mach contours of Figs. 19 and
20
the rearward m ovement
of the LE stagnation point, due to the Coanda effect, and the acceleration of the
Coanda flow are seen. Details of the Mach contours at the je t exit, along w ith the
corresponding fine grid, are displayed in Figs. 21 and 22. The je t flow is acceler-
ated to a Mach number exceeding 0.9, indicating the compressible character of
the jet.
There is only a small effect of mesh refinement on the solution calculated
with the SARC(cr3
=
0-9.6) model. Although not shown, the fine grid solution
for the surface pressures nearly coincides with the medium grid solution. In
addition, the velocity fields for the two grids are quite similar, as evident in
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 487/637
NUMERICAL SIMULATION OF CC AIRFOILS
487
0.85
0.1
0.05
Y
*
0
-0.05
-0 1 0.85
0.9 0.95 1 1.05 1.1
0.1
0.05
0
-0.05
-0.1
0.9 0.95 1 1.05 1.1
XlC
Fig. 18 Jet streamlines computed with
SARC cr3
=
0
- 9.6) turbulence model
( M , =
0.12,
a =
0 deg,
Re, =
5.45
X
lo5,
C =
0.209 , fine grid).
the velocity profiles shown in Figs.
23
and
24.
Table 2 com pares the predicted lift
and drag coefficients with the experimental values. In addition, the changes in
aerodynamic coefficients with further increases in C are indicated. There are
two factors one should keep in mind regarding this table. First, as indicated pre-
viously, the experimental CDvalues include the thrust effects produced by the jet ,
-0.1 0 0.1 0.2
0.2
.2°.2
0.1 0.1
0
0
-0.1 -0.1
-0.2
-0.2 -0.1 0 0.1 0.2
0.2
XlC
Fig. 19 Mach contours computed at LE with
SARC cr3
=
0
- 9.6) turbulence model
( M , = 0.12, a = 0 deg,
Re,
= 5.45 X lo5, C = 0.209 , fine grid).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 488/637
4aa
R. C.
SWANSON,
C.
L. RUMSEY, AND S.
G.
ANDERS
0.85 0.9 0.95
1
1.05 1.1
0.1 0.1
0.05
P
*
0
0.05
0
-0.05 -0.05
-0.1
0.85 0.9 0.95
1
1.05 1.1
XlC
-0.1
Fig. 20 Mach contours computed at
TE
with SARC cn
=
0
-
9.6) turbulence model
(Moo
=
0.12, a
=
0 deg,
Re, =
5.45 X lo5,
C =
0.209, fine grid).
whereas the computed CD alues do not. Secondly, there is some effect, although
it may be small, on these low-speed predictions because of the differences
between the 103RE and the NCCR geometries.
For Case
283
given in Table 2 the calculated CL s about 25 lower than the
experimental
CL.A
rather large increase in the C is needed to attain nearly the
0.966 0.968 0.97 0.972
0.038 0.038
0.036
P
*
0.034
0.036
0.034
0.032 0.032
0.966 0.968 0.97 0.972
XlC
Fig. 21 Fine grid in jet exit region.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 489/637
NUMERICAL SIMULATION OF CC AIRFOILS 489
0.008
0
007
0
006
0
005
0
004
0
003
Y
*
0
002
0 001
0
-0.001
0.966 0.968 0.97 0.972
0.038 0.038
1 1
.........
medium gr id
f ine gr id
.........
I
......................................................
j
......................... ~ ..............................
_ ........................................................ ......................... i..............................
......................... i ............................. ............................ i..............................
1
......................... ............................. ;+...................
1 1
1
- ......................... i.............................. ............................ i................
...............................................................................................
0.036 0.036
Y
*
0.034 0.034
0.032 0.032
0.966 0.968 0.97 0.972
XlC
Fig. 22 Mach contours in the vicinity
of
jet exit computed with SARC cr3 =
0
- 9.6)
turbulence model (Moo= 0.12, (Y =
0
deg, Re, = 5.45
X
lo5, C , = 0.209, fine grid).
Fig. 23 Effect of mesh density on velocity profiles computed at jet exit with
SARC cr3 =
0
-
9.6) turbulence model (Moo
=
0.12, a =
0
deg, Re, = 5.45
X
lo5 ,
C , = 0.209).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 490/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 491/637
NUMERICAL SIMULATION OF CC AIRFOILS 49
1
-1 0
-8
-6
-4
-2
0
0
2
4
0
0.2
0.4
0.6 0.8 1
XlC
Fig. 25 Surface pressures computed with two versions of SST turbulence model
(Moo=
0.12,
a =
0
deg, Re, =
5.45 x lo5,
C =
0.209).
A
comparison of the pressure distributions calculated with the SST(base1ine) and
SST(1994) turbulence models is shown in Fig. 25 for Case 283. Both medium- and
fine-grid results are given. There i s relatively good agreement with the data when
applying the SST( 1994) model, whereas the SST(base1ine) results exhibit poor
1
0.9
0.8
0.7
0.6
0.5
O
0.4
0.3
0.2
0.1
0
0.96 0.97 0.98 0.99 1
X/C
-0.1
Fig. 26 Comparison of surface skin-friction distributions at the TE computed with
SARC cr3=
0
-
9.6) and SST 1994) turbulence models ( M , = 0.12, a =
0
deg,
Re,
= 5.45
X
lo5, C = 0.209).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 492/637
492
R. C.
SWANSON,
C.
L. RUMSEY, AND S.
G.
ANDERS
0.85 0.9 0.95 1 1.05 1.1
0.1 0.1
0.05 0.05
Y
s o
0
-0.05 -0.05
-0.1
-0.1
0.85 0.9
0.95 1 1.05 1.1
X/C
Fig. 27 Jet streamlines and Mach contours computed with SST 1994) turbulence
model ( M , = 0.12,
(Y
= 0 deg, Re, = 5.45 X lo5,
C
= 0.209, fine grid).
agreement. Although use of Eq. 8) for the SST model has proven to be satisfactory
for many aerodynam ic flows of interest, it does not appear to be appropriate for the
Coanda je t flows being considered here; the SST( 1994) model performs better for
these particular low-Mach-number Coanda flows.
There is g reater sensitivity to m esh refinement with the SST( 1994) model
than that experienced with the SARC(cr3
=
0-9.6) model. The effect of
mesh refinement on the Coanda surface skin-friction distributions calculated
with these two models is shown in Fig. 26. Comparing Figs. 18 and 27, the jet
streamlines w ith the SST( 1994) model exhibit less spreading than those with
the SARC(cr3
=
0 - .6) model. Mesh refinement effect on the predicted CL
and C Dwith the SST(1994) model is given in Table
3.
On the fine grid, the pre-
dicted CL for Case 283 is 7.6% below that of the experiment. However, as shown
in Fig.
28,
the lift augmentation (slope of CL vs
C,)
appears to remain about the
Table
3
Comparison
of
computed, [with SST 1994) model] and experimental lift
and drag coefficients for CC airfoil
283 0.209 Medium
4.20
4.19 .050
0.0966
283 0.209 Fine 4.20 3.88
-
.050 0.0746
321 0.184 Medium 3.10 2.96 - .080 0.0655
321 0.184 Fine 3.10 2.41
-
.080 0.0559
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 493/637
NUMERICAL SIMULATION OF CC AIRFOILS
493
4.5
4
3.5
3
2.5
2
0
1.5
1
0.5
0.05 0.1 0.15 0.2 0.25
c,
Fig. 28 Variation
of
lift coefficient with jet momentum coefficientusing
SARC cr3
=
0
-
9.6) and SST 1994) turbulence models
( M , =
0.12,
a =
0 deg,
Re, =
5.45 X lo5).
same for SST(1994) with mesh refinement. In the
C L
predictions with both
models shown in Fig. 28, there is a monotonic increase in CL with increasing
C,.
The two-equation k--E models considered by Slom ski et a1.' result in a
nonphysical decrease in CLbeyond a C , of 0.093 (i.e., jet wraparound predicted).
-20
-1
6
-1 2
-4
0
0.2 0.4
0.6
0.8
1
XlC
4 0
Fig. 29 Surface pressures computed with SARC cr3 =
0
- 9.6) turbulence model
( M , = 0.12,
a
= -8 deg, Re, = 5.45 X lo5,
C
= 0.184).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 494/637
494
R.
C. SWANSON, C. L. RUMSEY, AND S. G. ANDERS
XlC
Fig. 30 Surface pressures computed with SST 1994) turbulence model
( M ,
= 0.12,
a
=
-8 deg, Re,
=
5.45 X lo5, C
=
0.184).
For the second case (Case 321, angle of attack of
-8
deg) of the workshop,
computed surface pressures for the medium and fine grids are presented in
Figs. 29 and 30. Results with both the SST(1994) and SARC(c,3
=
0 9.6)
models compare favorably with the experimental data. Nevertheless, the
0.85 0.9
0.95
1 1.05 1.1
0.1 0.1
0.05
P
*
0
0.05
0
-0.05
-0.05
-0.1
0.85 0.9 0.95 1 1.05 1.1
XlC
-0.1
Fig. 31 Jet streamlines and Mach contours computed at TE with
SARC cr3
=
0
-
9.6) turbulence model ( M , = 0.12, a = -8 deg, Re, = 5.45
x
lo5, C
=
0.184,
fine grid).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 495/637
NUMERICAL SIMULATION OF CC AIRFOILS
495
Fig. 32 Jet streamlines and Mach contours computed at TE with SST 1994)
turbulence model ( M , =
0.12, (Y = -8
deg, Re,
= 5.45
X
lo5,C
=
0.184,
fine grid ).
experimental
C
is underpredicted on the fine grid by more than 22% (see Tab les
2 and 3). As in the previous case (Case 283) one of the effects of grid refinement
seems to be reduced circulation, which results in the pressures on the airfoil
suction surface increasing. This effect appears to be much greater for the
current case because of the deg angle of attack. Paterson and Baker’ obtained
approximately the same value for the
CL
of this case using the SST( 1994) model
and performing an incompressible simulation for flow over the NCCR-1510-
7067 N geometry. With the SARC(c,3 = 0 9.6) model there is again greater
spreading of the jet than with SST(1994), as revealed by comparing Figs. 31
and 32, which depict the jet streamlines and Mach contours. There is an extre-
mely small recirculation region , w hich occurs only for the SS T( 1994) model,
on the lower surface that centers near the 0.92 chord location, but it is not
visible in Fig. 32.
VIII.
Conclusions
A computational method (CFL3D) has been applied to both low- and high-
subsonic Mach number CC airfoil flows. Several turbulence models have been
investigated. These models include the one-equation SA model with curvature
correction (SARC) and two variations of the two-equation shear stress transport
(SST) model of Menter. For the high-subsonic Mach number CC flow
(Case 302), all models have predicted jet separation from the Coanda surface
downstream of the experimental location, resulting in a significant overpredic-
tion of lift. In other words, all of the models have produced near-wall
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 496/637
496
R.
C. SWANSON, C. L. RUMSEY, AND S. G. ANDERS
eddy-viscosity levels that are too high in the Coanda flow. A parameter c,3) in
the curvature correction term of the SAR C model has been used as a vehicle to
explore the effect of reducing the turbulent kinetic energy in the Coanda flow.
In
so
doing, relatively good agreement with the experimental pressure
distribution of Case 302 has been obtained, even though the required c,3
value is unrealistically high.
In the simulation of low Mach number CC airfoil flows a set of calcu-
lations has been performed for a range of values of
C
The two cases of the
2004 Circulation Control Workshop have also been considered. Relatively
good agreement with experimental pressure data has been obtained when
modeling turbulence with the SARC(c,3
=
0
-
.6) and the SST(1994)
models. The SST(1994) model uses principal strain rate for the shear stress in
the modeling of the turbulence production. The SST(base1ine) model, which
uses vorticity in the turbulence production term, has not been satisfactory
when computing Coanda je t flows. An indication of the effects of grid refine-
ment on the results computed with the turbulence models has been given. The
SST( 1994) model has shown greater sensitivity to mesh refinement than the
SAR C(0 9.6) model. Lift and drag coefficients have also been determined
in the calculations.
Clearly, turbulence modeling is the major component in determining the
success of a computational method for predicting C C airfoil flows. Most standard
models, including SA, SARC
(c,3
l.O), SST(baseline), and EASM-ko, have
predicted jet separation too far around the Coanda surface. Accounting for
streamline curvature effects has been shown to be important, although the
SARC model required an artifically high level of its cr3 parameter in order to
produce reasonable results when compared with these particular experiments.
It is appropriate to note that in com parison to a different CC experimentz4 the
SARC model with its recommended value
(cr3
= 1.0) worked reasonably well,
and the SST( 1994) model performed poorly. Further investigation of models is
essential to achieving a reliable prediction technique that can be used for a
broad range of flow conditions.
In addition, improvements in computational efficiency must also be con-
sidered quite important if the prediction method is to be applied on a
routine basis with a high degree of reliability. Some rather straightforward
numerical algorithm features such as low-speed preconditioning should be
included in the method. Potential benefits of this preconditioning have been
indicated in this paper. Another possible improvement in computational
performance can be achieved by full coupling of the fluid dynamic and turbu-
lence transport equations, which is not done currently with the CFL3D code.
These and other improvements in computational efficiency are especially
important as the heiarchy (i.e., complexity) of the turbulence modeling is
increased. For example, if a full Reynolds stress model is used instead of a
two-equation model, such as SST( 1994), one m ust anticipate that there will
be a reduction in computing efficiency, because of a lower degree of numeri-
cal compatibility of the more complex model. In the case of steady flows,
numerical compatibility can be defined as a measure of the effect on solution
convergence of the complete system of flow equations due to turbulence
modeling.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 497/637
NUMERICAL SIMULATION OF CC AIRFOILS
497
Acknowledgments
The authors would like to thank
E.
Rogers
of
the Naval Surface Warfare
Center, Carderock Division, for providing coordinates
of
the
103RE
airfoil and
experimental data.
Appendix: Coordinates of 103RE Airfoil
X I . Y
I.
X I .
Y l C
X I Y
I. X I C
Y
I.
0.91346 0.010565 0.99999 0.0035035
0.093469 -0.0441 14 0.44534 0.087908
0.91436 0.016505 0.99939 -0.0027953 0.072353 -0.039186 0.52873 0.089357
0.91762 0.023061 0.99700 -0.0093482 0.055038 -0.034379 0.60007 0.08861 1
0.92444 0.028929
0.99243 -0.015540 0.040894 -0.029713 0.66105 0.086478
0.93516 0.031569 0.98563 -0.020621 0.029411 -0.025191 0.71316 0.083469
0.94627 0.032378
0.97807 -0.024547 0.020176
-0.020794 0.75764 0.079735
0.95247 0.032618 0.96976 -0.027987 0.012873 -0.016494 0.79558 0.075445
0.95592 0.032815 0.96123 -0.030852 0.0072776 -0.012252 0.82791 0.070840
0.95892 0.033013 0.95121 -0.033717 0.0032691 -0.0080402 0.85543 0.066064
0.96079 0.033051 0.93948 -0.036499 0.00083659 -0.0038889 0.87885 0.061361
0.96232 0.033040 0.92582 -0.039348 0.00027135 -0.0021 13 0.89877 0.056818
0.96419 0.032955
0.90993 -0.042273
0
0
0.91570 0.052486
0.96527 0.032875 0.89147 -0.045275 0.00022742 0.0021185 0.93008 0.048398
0.96646 0.032761 0.87004 -0.048349 0.00075518 0.0039056 0.94228 0.044569
0.96776 0.032607 0.84518 -0.051494 0.0031043 0.0081157 0.95262 0.041005
0.96920 0.032403
0.81636 -0.05471 1 0.0070381
0.012430 0.96137 0.037705
0.97078 0.032117 0.78295 -0.058003 0.012574 0.016818 0.96877 0.034657
0.97250
0.031733 0.74424 -0.061375
0.019838 0.021317 0.96877 0.034491
0.97439 0.031228 0.69939 -0.064825 0.029062 0.025973 0.94527 0.038834
0.97644
0.030569 0.64742 -0.068009
0.040572 0.030832 0.89801 0.038252
0.97866 0.029731 0.58721 -0.070608 0.054791 0.035930 0.83406 0.030760
0.98108 0.028721 0.5 1748 -0.072286 0.072250 0.041298 0.50044 0.030760
0.98367 0.027438 0.43673 -0.072472 0.093598 0.046950 0.50044
-
.041077
0.98643 0.025852 0.35601 -0.070647 0.11963 0.052894 0.73553 -0.041077
0.98933 0.023862 0.28917 -0.067442 0.15131 0.059114 0.91346 0
0.99231 0.021385 0.23385 -0.063413 0.18980 0.065569 0.91346 0.010565
0.99512 0.018159 0.18811 -0.058917 0.23654 0.072089
0.99753 0.014103 0.15032 -0.054098 0.29326 0.078324
0.99925 0.0091942 0.11915 -0.049115 0.36203 0.083837
References
‘Englar, R. J., and Williams, R. M ., “Test Techniques for High Lift Two-Dimensional
Airfoils w ith Boundary Layer and Circulation Control for Application to R otary Wing A ir-
craft,”
Canadian Aeronautics and Space Journal,
Vol. 19, No. 3, 1973, pp. 93-108.
’Jones, G.
S.
Viken,
S.
A, Washburn, A. E., Jenkins, L. N., and Cagle, C. M.,
“An
Active Flow Circulation Controlled Flap Concept for General Aviation Aircraft Appli-
cations,” AIAA Paper 2002-3157, June 2002.
3Pulliam, T. H., Jespersen, D. C, and Barth, T. J., “Navier-Stokes Computations for
Circulation Control Airfoils,” AIAA Paper 85-1587, July 1985.
4Pulliam, T. H., “Euler and Thin Layer Navier-Stokes Codes: ARC2D, A RC3D,”
Notes fo r Computational Fluid Dynamics User’s Workshop, Univ. of Tennessee Space
Inst., Tullahoma, TN , March 1984.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 498/637
498
R.
C. SWANSON, C. L. RUMSEY, AND S. G. ANDERS
’Abramson, J., and Rogers, E., “High-speed Characteristics of Circulation Control Air-
foils,” AIAA Paper 83-0265, Jan. 1983.
6Wilkerson, J. B., and Montana, P. S., “Transonic W ind Tunnel Test of a 16 Percent
Thick Circulation Control Airfoil with 1 Percent Asymmetric Camber,” DTNSRDC
ASED 82/03, April 1982.
’Baldwin, B. S., and Lomax, H., “Thin Layer Approximation and Algebraic M odel for
Separated Flows,” AIAA Paper 78-257, Jan. 1978.
‘Slomski, J. F., Gorski, J. J., Miller, R. W., and M arino, T. A., “Numerical Sim ulation
of Circulation Control Airfoils as Affected by Turbulence Models,” AIAA Paper 2002-
0851, Jan. 2002.
’Paterson, E. G. nd Baker, W. J., “Simulation of Steady Circulation Control for
Marine-Vehicle Control Surfaces,” A IAA Paper 2004-0748, Jan. 2004.
“Menter, F. R., “Two-Equation Eddy-Viscosity Turbulence Models
for
Engineering
Applications,” AIAA Journal, Vol. 32, No.
8,
1994, pp. 1598-1605.
“Jones,
G. .,
and Joslin, R. D. (ed.), Proceedings of the 2004 NASA/ONR Circulation
Control Workshop, NASA/CP 2005-213509, March 2004.
”Abramson, J., “Two-D imensional Subsonic Wind Tunnel Evaluation of Two Related
Cambered 15-Percent Circulation Control Airfoils,” DTNSRDC ASED-373, Sept. 1977.
13Krist,
S.
L., Biedron R. T., and Rumsey, C . L., “CFL3D U ser’s Manual,” NA SA TM
1998-208444, June 1998.
14Spalart, P. R., and A llmaras, S. R., “A One-Equation Turbulence Model for Aerody-
namic Flows,” La
Recherche Aerospa tiale,
Vol. 1, 1994, pp. 5-21.
”Spalart, P. R., and Shur, M., “On the Sensitization of Turbulence Models to Rotation
and Curvature,” Aerospace Science and Technology, Vol. 5 , 1997, pp. 297-302.
16Rumsey,C. L., Gatski, T. B., Anderson, W. K., and Nielsen, E. J., “Isolating Curvature
Effects in Computing W all-Bounded Turbulent Flows,” International Journal of Heat and
Fluid F low, Vol. 22, 2001, pp. 573-582.
”Menter,
F.
R., “Improved Two-Equation k - o Turbulence Model for Aerodynamic
Flows,” NASA TM 103975, Oct. 1992.
18Menter,F. R., “Zonal Two Equation k - o urbulence Model for Aerodynamic Flows,”
AIAA Paper 93-2906, July 1993.
‘’Rumsey, C . L., and Gatski, T. B., “Summary of EASM Turbulence Models in CFL3D
with Validation Test Cases,” NASA/TM-2003-212431, June 2003.
”Menter, F. R., Kuntz, M., and Langtry, R., “Ten Years of Industrial Experience with
the SST Turbulence M odel,” Turbulence, Heat and Ma ss Transfer 4, edited by K. Hanjalic,
Y. Nagano, and M . Tumm ers, Begell House, Redding, CT, 2003, pp. 625-632.
’lTurke1, T., Vatsa, V. N., and Radespiel, R., “Preconditioning M ethods
for
Low-Speed
Flow,” AIAA Paper 96-2460, June 1996.
”Turkel, T., Radespiel, R., and
Kroll,
N., “Assessment of Two Preconditioning Methods
for Aerodynamic Problems,”
Computers and Fluids,
Vol. 26, No. 6, 1997, pp. 613-634.
23Turkel, T., “Preconditioning Techniques in Computational Fluid Dynamics,”
Annual
Review of Fluid Mechanics,
Vol. 31, 1999, pp. 385-416.
24Swanson, R. C ., Rumsey, C. L., and Anders, S. G. Progress Towards Computational
Method
for
Circulation C ontrol Airfoils,’’ AIAA Paper 2005-0089 , Jan. 2005.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 499/637
Chapter
19
Role of Turbulence Modeling in Flow Prediction
of Circulation Control Airfoils
Gregory M cGow an,* Ashok Gopalarathnam,+ Xudong Xiao,*
and Hassan Hassans
No rth Carolina State University, Raleigh, North Carolina
Nomenclature
c =
airfoil chord
C, =
pressure coefficient
Cf
= skin friction coefficient
C ,
=jet momentum coefficients
h =
slot height
k
= turbulence kinetic energy
r z = mass flow rate
M =
Mach number
V = velocity
eff
effective angle of attack
p
= density
w = turbulence frequency
=
enstrophy
Subscripts
j
= j e t
= freestream conditions
*Research Assistant, Department of Mechanical and Aerospace Engineering. Student Member
'Associate Professor, Department of Mechanical and Aerospace Engineering. Senior Member
'Research Assistant, Professor, Department of Mechanical and Aerospace Engineering. Me mbe r
%Professor, Departm ent of Mechanical and Aerospace Engineering. Fellow AIAA.
Copyright 005 by
th
authors. Published by the American Institute of Aeronautics and Astro-
AIAA.
AIAA.
AIAA.
nautics, Inc., with permission.
499
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 500/637
500
G. McGOWAN
ET
AL.
I. Introduction
HERE IS a continuing need to pursue technologies that can address long-
T
erm aerodynamic goals for future aircraft. For example, long-term visions
for aeronau tics predict the need and potential for a dram atic
50
reduction in fuel
bum over the next 20 years, 36% of which is expected to come from improved
aerodynamics. Other system studies by NASA Langley and Boeing2 have ident-
ified that the benefits of flow control are best realized by the development of sim-
plified high-lift systems. Thus there is a long-term need for technology that can
integrate the achievement of significant drag reduction (36%) at cruise with the
achievement of very high lift for short-field operations.
Circu lation control (CC ) is one type of flow control that has received consider-
able attention in recent years. This is because these systems offer the possibility
of reduced takeoff and landing speeds, as well as increased maneuverability. The
flow control is implemented by tangentially injecting a jet over a rounded wing
trailing edge (TE). As a result of the balance between the pressure and the cen-
trifugal force (the Coanda effect), the je t remains attached along the surface of the
wing. Thus, the T E stagnation point is moved towards the lower surface, whereas
the leading edge (LE) stagnation point is moved rearward, resulting in increased
effective camber. This important area of research was the subject of a well-
attended 2004 NASA/ONR workshop on circulation control3 that was held at
NASA Langley Research Center in March 2004. A number of contributors used
different turbulence m odels including algebraic, one-equation, two-equation, and
stress models to try to predict flow characteristics of various C C airfoils. None of
the models employed performed well for all je t mom entum coefficients
C
con-
sidered. The only exception is the Spalart-Allmaras model that includes curva-
ture effects (SARC).4 However, the success of this model came as a result of
adjusting5 one of the model constants, (c,3), which typically lies in the range
of 0.6- 1.0, to 9.6. This adjustment has the effect of reducing the eddy viscosity
throughout the flowfield and may change the character of the flow from turbulent
to laminar or transitional flow over a large portion of the airfoil.
The goal of this investigation i s to consider the flow over the 103RE(103XW)
CC airfoil tested by Abramson and Rogers.6 The tests were conducted to deter-
mine the performance characteristics of CC airfoils at transonic speeds. This
airfoil was considered in Ref. 5 . Two turbulence models are employed in this
investigation: the k - 5 model of Robinson and Hassan7 and the k -w model of
Wilcox.* The latter model is included for comparison purposes because it
yields results similar to the other turbulence models (other than SARC) in
CFL3D.9 Both models were implemented in CFL3D (Version 5 . This version
of CFL3D was modified to incorporate the
k - 5
transitional/turbulence model
of Warren and Hassan.
The k -5 model7, differs from other turbulence models used in Ref. 9 by the
fact that
it
is derived by m odeling the exact equa tions that govern the variance of
velocity, or turbulence kinetic energy,
k,
and the variance of vorticity, or enstro-
phy, 5 As a result, the k - 5 model contains all the relevant physics in the k and
5 equations, is tensorially consistent, Galilean invariant, coordinate-system
independent, and is free of wall or damping functions. It correctly predicts
wall-bounded shear flows and the growth of all free shear layers (je ts, wakes
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 501/637
TURBULENCE MODELING IN FLOW PREDICTION
501
and mixing layers). According to Wilcox,* this is a minimum requirement for any
turbulence model that is proposed for use in com plex flows. It is to be noted that
none of the turbulence models used in Ref. 9 satisfies the requirements suggested
by Wilcox.
The k - j transitional/turbulence model has the option to treat the flow in each
block as laminar, transitional, or turbulent. The model requires that the transi-
tional mechanism and freestream turbulent intensity be specified and is capable
of predicting the onset and extent of transition. In this work, the transition over
the external surface of the airfoil is deemed to be a result of the growth of Toll-
mien-Schlichting waves. The code has no transitional mechanism suited for
internal flows, such as the cavity flow or subsonic nozzle employed here.
11.
Formulation of the Problem
A.
Turbulence Models
Most of the results obtained here employ the k - j turbulence model7 and the
k -
j transitional/turbulence model.’ The governing equations and boundary con-
ditions are detailed in the cited references. C alculations are also presented fo r the
k -w
model.
B. Geometry and Grid
The airfoil under consideration is elliptical in shape, has a chord of 1.5 ft.,
thickness ratio of 16%, and a camber ratio of 1%. The jet slot-height-to-chord
ratio
h/c )
s 0.0021. The near-field of the medium-resolution grid is shown in
Fig.
1.
The fine grid has 235 points around the airfoil and 49 points in the
normal d irection over the forward part (block 2) and 101 points in the aft part
(block 3) including points in the cavity (block 1). In the normal direction, the
grid is clustered at the surface and y + there is less than one. The total number
of grid points is 70,563. For the medium grid the number of cells is halved in
each coordinate direction. The grid is patched at the lower airfoil surface.
C. Num erical Procedure
The numerical solution was computed using the code CFL3D.’ It is based on a
finite volume m ethod. The convective terms are approximated by upwind-biased
spatial differences, and the viscous terms are discretized using central differ-
ences. In this work, the flux difference splitting of Roe is employed. Time inte-
gration is accomplished with an implicit approximate factorization scheme.
The turbulence models are uncoupled from the mean flow equations. Their
advection terms are discretized with first-order upwind differencing, whereas
the source terms were treated implicitly.
Characteristic-type boundary conditions are employed at inflow and outflow
boundaries. For the plenum the mass flow rate and flow inclination angle are pre-
scribed. At the surface of the airfoil, no-slip and adiabatic wall conditions are
employed.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 502/637
502
G. McGOWAN
ET
AL.
Fig. 1 Close-up of the grid employed.
111. Results and Discussion
A
summary of the flow condition employed is given in Table 1, with
C
defined as
mj vj
1/2pmv:c
-
where
j
is the je t m ass flux per unit span, vj is the je t velocity,
poo
and Vm are the
freestream density and velocity, and j is the jet Mach number. The effective
angle of attack a , ~as determined by matching pressure coefficient distribution
forward of midchord with a potential code that used CL and angle of attack as
input^ ^ Because the freestream temperature is constant, it is seen that
C
is inver-
sely proportional to the square of the freestream Mach number. This is why the
C values appear to be small for this case.
Table 1 Summary of flow conditions employed for each case
301 0.0
0.0
.0540
302 0.0032 0.519 .2865
306 0.0110 0.979 .7980
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 503/637
TURBULENCE MODELING IN FLOW PREDICTION 503
0.8
0.6
0.4
0.2
0
0.2
0
0.4
0.6
0.8
1
0
0.25 0.5 0.75
XlC
Fig. 2 Comparison of C using k land
k - o
for Case 301 C
= 0.
As indicated below, grid refinement shows that results for the medium grid are
identical to those for the fine grid. In spite of this, all results presented here
employ the fine grid. Figure 2 com pares calculated and measured pressure distri-
bution in the absence of injection (Case 301). As is seen from the figure, both
model predictions are in good agreement with experiment. Figure 3 compares
predictions with experiment for Case 302.
As
is seen from the figure, the
k - 5
tur-
bulence model predictions are in better agreement with experiment than those
given by the
k-w
model. The reason for this may be observed in Figs. 4 and
5 , which compare the streamline patterns in the injection region.
As
may be
seen from the figures, the flow separation for the
k-w
model is delayed resulting
in higher lift. Figure 6 presents calcu lated skin-friction coefficients using the k-5
model. The transitional behav ior indicated in the figure is a result of a numerical
transition. This is typical of all turbulence models.
Figure
7
com pares the pressure distribution for Case 302 using the
k - 5
model
on the fine and intermediate grids. It is seen that the solutions are grid
independent.
Figure 8 shows the calculated pressure distribution fo r Case 306 using the
k -w
model. We were unable to obtain a steady-state solution using the k -5 model for
this case. This can be seen from a plot of the residual indicated in Fig.
9.
As may
be seen from this figure, one can stop the solution earlier and obtain a rather
reasonable solution or any solution desired depending on when the calculation
is terminated. Because of the above behavior, a time-accurate solution was
attempted. We were unable to detect a statistically steady solution even after
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 504/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 505/637
TURBULENCE MODELING IN FLOW PREDICTION
505
0.06
0.05
0.04
0.03
0.02
0.01
0.01
0.02
-0.03
-0.04
0.05
-0.06
-0.07
0.08
0.09
.
0.95 1 1.05
X/C
Fig.
5
Streamline pattern around separation point
k - o )
for Case 302
C,
= 0.0032.
0.007
0.006
0.005
0.004
u
0.003
0.002
0.001
0
0 0.2 0.4 0.6 0.8 1
XlC
Fig. 6 Calculated skin friction using k -j m od el for Case 302
C
= 0.0032.
x/c
C f
0 0.2 0.4 0.6 0.8 1
0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
lower surface
upper surface
Case 302
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 506/637
506
G. McGOWAN ET AL.
-0.8
0.6
0.4
0.2
0.4
0.6
0.8
1
0 0.2 0.4 0.6 0.8 1
XlC
Fig. 7 Comparison of
C
for k - j o n fine and coarse grid for Case 302
C
=
0.0032.
1.5
-1
0.5
e
0
0.5
1
0
0.2 0.4
0.6 0.8
XlC
Fig. 8
Prediction of
C
using k - o model for C ase 306
C
= 0.011.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 507/637
4
4.5
5.5
6
TURBULENCE MODELING IN FLOW PREDICTION 507
t
Case306
0 5000 10000 15000 20000 25000
I terat ion
Fig.
9
Convergence history using k -j m od el for Case
306 C,
=
0.011.
running the code for 28 periods. This suggests that a more elaborate approach,
such as a large eddy simulation (LES)/Reynolds averaged Navier-Stokes
(RANS),
would be required.
All of the above calculations assumed that the flow is fully turbulent. This is
not necessarily the case. Attention was focused next on the use of the
k
5
ransi-
tional/turbulence model to analyze the flow for Case 306 because such a model
will result in reduced eddy viscosity and earlier separation. The m odel, as coded
in CFL3D, allows the user to specify laminar, transitional, or turbulent flow in
each block. Further, it requires the user to specify the transitional mechanism
and the freestream turbulence intensity.
Th e transitional m echanism considered in the code is a result of the growth of
Tollmien-Schlichting waves. This mechanism is not the correct mechanism for
triggering transition in cavities. As a result, two cases were run. In the first, the
flows in blocks 2 and
3
were specified transitional and turbulent, respectively,
whereas the flow in the cavity was specified to be laminar. In the second case
the flow in the cavity was assumed to be turbulent. It is seen from Figs. 10 and
11 that the results are dependent on whether the flow in the cavity is laminar
or turbulent. Figures 12 and 13 show the streamline patterns in the injection
region. They show that flow separation takes place earlier for the case where
the flow in the cavity is laminar. This result explains why an increased value
for the curvature parameter employed in Ref. 5 which resulted in reduced
eddy viscosity, gave good agreement with experiment.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 508/637
508
G. McGOWAN
ET
AL.
Fig. 10 Prediction of C using transitional k - j for Case 306 C = 0.011 with
laminar cavity.
X/C
Fig. 11 Prediction of C using transitional k - j for Case 306 C = 0.011 with
turbulent cavity.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 509/637
TURBULENCE MODELING IN FLOW PREDICTION
509
Y
*
Fig. 12 Streamline pattern around separation point for Case 306
C,
= 0.011 with
laminar cavity.
Case 306
turbulent cavi ty
Y
*
I I I I I I I I I I L
1 1.05
XlC
Fig. 13 Streamline pattern around separation point for Case 306
C,
= 0.011 with
turbulent cavity.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 510/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 511/637
1II.C. Tools for Predicting Circulation Control
Performance:
GACC Airfoil Test Case
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 512/637
Chapter 20
Simulation of Steady Circulation Control
for the General Aviation Circulation
Control GACC) Wing
Warren J. Baker* and Eric G. Patersont
Pennsylvania State University, University Park, Pennsylvania
Nomenclature
a
speed of sound, ft/sec
CD
section drag coefficient,
F d / l / 2 ) p U k S
CL section lift coefficient, F J 1 / 2 ) p U k S
C, pressure coefficient, p
, ) / 1 / 2 ) p ~ k
C ,
et mom entum coefficient,
~ U ~ / I / ~ ) P U ; S
c
chord length, in.
h slot height, in.
k turbulent kinetic energy (TKE), f t2/s2
reference length used in defining velocity boundary condition
riz mass flow rate, lbm/sec
M
Mach number,
U / a
p pressure, lbm/ft2
ramp Cubic polynom ial used to accelerate the velocity amplitude
from 0 to the final value after a nondimensional time of
1 0
Re
Reynolds number, pU,c/p
s planform area, ft2
U,V,W velocity component in Cartesian coordinates, ft/s
Upoly tenth-order polynomial curve fit for defining velocity
boundary conditions
vjet
steady blowing jet amplitude, ft/s
*Graduate Research Assistant, Department of Aerospace Engineering. Member AIAA.
'Senior Research Associate, Applied Research Laboratory, and Associate Professor of Mechanical
Copyright 005 by Warren J. Baker and Eric G. Paterson. Published by the American Institute
and Nuclear Engineering. Member A I M .
of Aeronautics and Astronautics, Inc., with permission.
513
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 513/637
514 W.
J.
BAKER AND E. G. PATERSON
x,y,z Cartesian coordinates
p
density, lbm /ft3
0, =jet angle applied for velocity boundary conditions at the
jet-slot exit, deg
w
specific dissipation rate, ft2/s3
Subscripts
3 freestream
at jet-slot exit
I. Introduction
HE CONCEPT of circulation control (CC) using the Coanda effect is a
T
henomenon involving a two-dimensional wall bounded jet passing along
a curved surface. The jet itself is introduced via a slot, which expels the jet , typi-
cally, tangentially to the curved surface. This je t adds m omentum to the boundary
layer close to the curved surface. With the curved surface, the Kutta condition is
not applicable, and the rear stagnation point is free to move. T he resultant is a net
change in the circulation, and the flow turning and separation location are altered
based on the rate of mass addition. Accompanying the change in circulation are
changes in certain aerodynam ic values such as lift, total drag, and local skin fric-
tion coefficient. Figure 1 shows an example of a Coanda jet CC setup with a
single slot.
The performance benefits of CC have been show n in many experiments since
the early 1 9 7 0 ~ . l - ~ncreases in lift of as much as 10 times the typical flap system
Fig.
1
Tra iling-edge Coanda jet.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 514/637
STEADY CC SIMULATION FOR GACC WING 515
have been reported. Other possible benefits of the use of circulation control
include elimination of moving parts, part/card decrease, significant weight
decrease, and a less complex high-lift system. Circulation control is very attrac-
tive for certain naval applications, in particular the replacement of current actua-
tion techniques on surface ship and submarine control surfaces with CC schemes.
The current schemes, although robust and efficient, particularly for high-speed
operation, have drawbacks. The force generated by a control surface is a function
of lift coefficient, which in turn i s a function of foil geometry, angle of attack, the
square of the relative velocity of fluid over the control surface, and the fluid
density. At very low vehicle speeds, the control surfaces may not provide suffi-
cient control authority. Also, for marine applications, the density of water
means that very large actuation forces and therefore complicated mechanisms
must be created to move the control surfaces. Because of these drawbacks, and
the desire in the submarine community for effective and safe low-speed littoral
operations, there is motivation to develop alternative technologies for creating
maneuvering forces. Circulation control schemes would provide very high lift
at very low speeds, for example, in littoral operation or for evasive maneuvering,
where the current control surface technologies are insufficient. The placement
of a fixed control surface would increase shock resistance, allow placement of
sensors or payload on the control surface, or even allow for the placement of
the control surface in nontraditional areas previously restricted by the need for
moving surfaces, such as on the outside of the propulsor duct.
The long term objective of the present research is to develop validated simu-
lation tools using m ultiple data sets. These data sets include a two-dimensional
CC experiment using the NCC R 15 10-7067N,* a low-aspect-ratio, tapered,
control surface for marine app lications, CCFOIL,3 and the General A viation Cir-
culation Control (GACC) wing4 the latter two of which are three-dimensional
configurations. The work presented herein is the initial effort to investigate
steady blowing CC of the GACC wing using the Reynolds-averaged Navier-
Stokes (RANS) equations, and knowledge gained here will be combined with
that from previous studies of the NCCR foil5 to continue to develop, validate,
and verify our simulation tools for CC.
The GACC was selected as a validation benchmark because it provides a
modem experiment with computational fluid dynamics (CFD) validation in
mind. Also, other CFD efforts have been initiated for the GACC, and both
steady and pulsed actuation were used in experiment. The geometry itself has
two slots (upper and lower) and has multiple trailing edge (TE) variants.
11.
Geometry,
Conditions, and Data
The GACC was tested in the Basic Aerodynamics Research Tunnel at NASA
Langley Research Center. The GACC section is a modified General Aviation
Wing-1, and is a supercritical 17 thick airfoil, with two slots. The chord
length is 9.40 in. and the freestream velocity for experimentation is 110 fps y ield-
ing a chord Reynolds number of 5.33 x lo5 and a freestream Mach number M
of approximately 0.10. The upper slot is located at
x c
=0.985 and the lower slot
is located at
x c
0.975. The slot height-to-chord ratio
h / c ,
is approximately
0.00106. The circular TE has a radius-to-chord ratio Y C of 2.00 . A cross-
section of the model is shown in Fig.
2.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 515/637
51
6
W.
J.
BAKER AND E. G. PATERSON
UPPER
STEADY
MAMFoLo
ACTUAtOA UPPER
mSm SLOT
Fig. 2 Cross-section of the GAC C wing.
A range of blowing coefficients were investigated, the highest being 0.162.
Equation (1) provides a relation between the jet velocity and the nondimensional
blowing coefficient. The code used for the current work is an incompressible
code. To adjust for this, the density relations during experiment were acquired
in order to obtain the proper jet velocity at the jet-slot exit. Therefore the
maximum jet velocity corresponding to a C, 0.162 is 917 fps and the non-
dimensional je t velocity U j / U W 8 34:
For all cases studied, the angle of attack was 0 deg. Experimental data are
availab le for u er slot steady blowing, lower-slot steady blowing, and dual-
assist blowing. The most recent experimentation completed focuses on pulsed
actuation, and initial data from pulsed testing are available.6 Tab le 1 summ arizes
the experimental data available. Experimental uncertainty has not yet been
provided.
Previous results from CFD simulations using the NASA Fully Unstructured
Navier-Stokes 2D code (FLJN~D)’ave been published? FUN2D uses the
Spalart- Allmaras turbulence model, and all simulations completed assumed fully
turbulent flow. A comparison to experiment of lift and drag data for a range of
steady blowing coefficients has been presented? Tw o slot heights were used in
simulations, 0.01 and 0.02 in. and results showed good trend agreement for the
smaller of the two heights. Figures 3 and 4 show the lift vs. blowing coefficient
curve and the drag polar for
FUN2D
simulations and experiment, respectively?
BP
-
111.
Computational Methods
The flow code used for the current work, CFD SHIP, is a general-purpose, par-
allel, unsteady, incompressible, RANS CF D code. The computational approach is
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 516/637
STEADY CC SIMULATION FOR GACC WING
51
7
Table 1 Available data from GACC experimentation
Baseline (no jet actuation)
1) Surface pressure distribution
2) Lift-curve slope
3)
Drag polar
Steady upper slot blowing
1) Surface pressure distribution
C , 0.059
and
0.162)
2) Lift-curve slope
C ,
0.007, 0.015, 0.025,
0.041, and 0.060)
3) Lift vs blowing coefficient (slot
height
0.01
in. and 0.02 in.)
4)
Drag polar
5) Jet exit Mach number profiles C , 0-0.162)
6) Lift vs mass flow rate
1)
Surface pressure distribution CL 1.2)
2) Lift vs mass flow rate
1) “Negative lift configuration,” lift vs blowing
2) “Negative lift configuration,” drag polar
1) Drag polar (slot height 0.01 in. and 0.02 in.)
2) Drag polar (matched slot
C ,
0.0, 0.004,
3) Drag vs angle of attack
4) Angle
of
attack vs
L I D
Pulsed upper-slot blowing
Steady lower-slot blowing
coefficient
Dual-slot assist steady blowing
0.005, 0.009,
0.021, and 0.0041)
Fig. 3 C , vs C ,
for
previous C FD simulations and experiment
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 517/637
518
W.
J.
BAKER AND E. G. PATERSON
Fig.
4
Drag polar for previous CFD sim ulations and experiment
based upon structured, overset-grid, higher-order finite-difference, and pressure-
implicit split-operator (PISO) numerical methods. Production turbulence model
uses a linear closure and the blended
k -w /k -E
SST two-equation m odel.* Effi-
cient parallel computing is achieved using coarse-grain parallelism via MPI dis-
tributed computing. For time-accurate unsteady simulations, global solution of
the pressure-Poisson equation is achieved using preconditioned GMRES and
the PETSc libraries.
IV.
Grid Generation
Overset grids are generated primarily using hyperbolic extrusion and orthog-
onal box grids, although transfinite interpolation and elliptic smoothing of blocks
can be used when needed. Overset interpolation coefficients are calculated and
holes are cut using PEGASUS 5.1.9 CFDSHIP employs double-fringe outer
and hole boundaries
so
that the five-point discretization stencil (i.e., in each
curvilinear coordinate direction) and order of accuracy does not have to be
reproduced near overset boundaries. The level-2 interpolation capability of
PEGASUS 5.1 is used to achieve an optimal match between donor and inter-
polated meshes.
Two grids were created initially for simulations. One grid included the upper
plenum for modeling of the jet at the diffuser nozzle, whereas the second grid did
not contain the plenum grid and modeled the jet at the orifice. The former of the
grids is shown in Fig. 5, with block numbers noted. The domain size, as marked
by the outermost boundaries of a nested orthogonal box grid shown as block 1,
ranged from - 3
< x / c <
4,
- 3 <
y/c
< 3,
and
0 < z / c <
0.1. Near-wall
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 518/637
STEADY CC SIMULATION FOR GACC WING
a)
519
Fig. 5 Overset computational dom ain including the plenum: a) Overa ll view;
b)
foil
view; c) plenum view.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 519/637
520 W.
J.
BAKER AND E. G. PATERSON
spacing ranged between 2.00 x The finer spacing was
applied to all external surfaces to obtain proper resolution of the sublayer
region of the turbulent boundary layer. The larger spacing was applied for the
internal surfaces of the plenum, such that the boundary layers could be resolved
properly. Two elliptically smoothed blocks span along the TE from upper to
lower slot, denoted as blocks 6 and 7. Then, an O-grid was hyperbolically
extruded around the body and split into four blocks, 2-5. A plenum block was
created, block 8, and finally, an overset grid was placed along the knife edge
of the upper slot, block 9, for investigation of the slot-lip interaction. The
RANS simulations were performed in a pseudo-two-dimensional fashion,
which requires five points in the spanwise direction. The grid consists of nine
blocks containing a total of 394,665 points. Block sizes ranged from 31,000 to
61,000 points, with the plenum block having 33,000 points.
The second grid, which does not include the plenum, totals eight blocks with
381,810 points. Only the T E view is show n in Fig. 6, because the computational
domain i s very similar to that show n in Fig. 5 in all regions except the TE. The
difference in grid point number between the tw o grids is a result of the removal of
a block and modifications to the near-wall spacing at the jet-slot exit to facilitate
the applied boundary conditions. The block numbers coincide w ith those shown
in Fig. 5 , excluding the plenum block.
A
three-point grid study was completed for uncertainty assessment. The
previous grid without the plenum was used as the fine grid for the study.
A
and 2.00 x
0.M
0
4.M
0.96 0.97
0 .a 0.99
1 1.01
1.m
XlC
Fig. 6 Trailing-edge view of grid without plenum.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 520/637
STEADY CC SIMULATION FOR GACC WING
521
Table 2 Total grid points for the
fine, medium, and coarse grids
Grid
Total
grid
points
Fine
381,810
Medium
193,980
Coarse
97,575
J2
refinement process was completed to create a medium and coarse grid. This
process was completed by decreasing the number of grid points by
J2
in each of
the x-and y-directions of the finest grid to create the m edium grid. Therefore, the
near-wall spacing applied for each of the fine, medium, and coarse grids was
2.00
x
lop6,2.83
x lop6 and
4.00
x l op6 respectively. Because of smooth-
ing of some of the com putational domain during grid creation, larger near-wall
spacing occurred. This larger near-wall spacing occurred at the bottom slot and
was
3.48 x lop6 4.44 x lop6,
and
5.67 x lop6
for the fine, medium, and
coarse grids, respectively. The result of the refinement process is a reduction
of grid points by a factor of approximately
1/2
from fine to medium grids.
The same process is carried out to create the coarse grid from the medium.
The coarse grid has approximately
1/2
the total grid points as the medium
grid and approximately
1/4
the total points of the fine grid. Thus, from
the fine to coarse grid, we have what is called “grid halving.” Table
2
shows
the total number of grid points for the fine, medium, and coarse computational
domains.
V.
Initial and Boundary Conditions
Initial conditions for the steady-state RANS simulations were prescribed to be
equal to the freestream velocity, turbulence, and pressure:
where the subscript
00
refers to freestream conditions. No-slip boundary con-
ditions were applied to the upper and low er surface of the airfoil, the round TE
region, and the upper and lower surfaces of the plenum . For each grid, a different
boundary condition was specified for the steady blowing. For all cases, the angle
of attack was zero degrees.
Figure 7 shows the location of the steady blowing boundary condition for the
grid without the plenum. This occurs along the bottom portion of the jet slot. A
no-slip condition is applied to the top portion of the jet slot. A velocity boundary
condition is prescribed, and the velocity profile Upoly s a tenth-order polynomial
curve fit of a typical CC jet profile seen in a previous RANS results for the GACC
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 521/637
522
W.
J.
BAKER AND E. G. PATERSON
S 0 0
Dose
9s7
0.99s Om
Id
Fig. 7 Boundary condition for grid without the plenum.
airfoil4 and is given by Eq.
(3):
Upoly (-
1.2222
x
102*y/1'o) 1.7043
x
102*y/Z9)
+ (1.8036 x 103*y/Zs) 3.4603 x 103*y/17)
+
(2.9482 x 103*y/Z6) 1.0602 x 103*y/15) (3)
9.7236
x
10'*y/Z4)+ (2.2944
x
102*y/13)
8.5386 x 10'*y/Z2)+ (1.4472 x lO'*y/l) +0.0036
where y/1 is the nondimensional distance along the boundary. T o acquire tangen-
tial flow to the round TE, an initial angle of 6
18
deg was enforced. It was
necessary to include this angle because the flow was modeled at the jet-slot
exit. If the plenum flow had been modeled, proper jet attachment would have
already been established at the location of the jet-slot exit. The velocity boundary
condition for the grid without the plenum is given as
U vjet x ramp x cos (6) x UpOly
V vjet ramp x sin (6) x
Upoly
w o
where vjet is the velocity amplitude based on the blowing coefficient and Eq. (l),
and
ramp
is a cubic polynomial used to accelerate the velocity amplitude from
0
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 522/637
STEADY CC SIMULATION FOR GACC WING
523
t
0.001
~ . ' . ~ . ' ' ' ~ . ' ' . L
0
0.5 1 15
U
Fig. 8 Velocity profile prescribed for steady blowing boundary condition at the jet-
slot exit.
to the final value after a nondimensional time of 1.0. The U-velocity profile for
the boundary condition is shown in Fig.
8.
The boundary condition for the grid with the plenum is less complex. Figure 9
shows the upstream face of the plenum where the steady blowing boundary
condition is applied. In this case, a top-hat velocity distribution is used. Also,
no additional flow angle is required to obtain tangential flow. The velocity bound-
ary condition for steady blowing with the grid including the plenum is given
in Eq. (7):
(7)
yetx ramp
VI. Com putational Resources
All simulations were executed on an IBM SP Power 3 machine w ith 64 nodes.
Each node contains sixteen, 375 MH z Power
3
processors. Each C PU has 64 kB
level-1 cache and 8 MB level-2 cache memory along with 1 GBRAM ach pro-
cessor has a maximum sustainable performance of 1.5 GFLOPS, giving each
node 24 GFLOPS peak performance. Scratch space available to users totals
3.2 TB (from ARL MSRC IBM SP Information, http://ww w.arl.hpc.mil/
userservices/ibm.html). As a reference point, a fine grid without the plenum
completed 10,000 iterations (well past convergence for most simulations com-
pleted) in 16.7 wall-clock hours or 133.7 CPU hours.
VII. Results
This section presents the results from three separate studies. The first details
the effects of modeling the Coanda jet vs resolving the internal plenum
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 523/637
524
om
om
OD
W.
J.
BAKER AND E. G. PATERSON
0
am
om M
3dc
Fig.
9
Boundary condition for grid with plenum.
geometry. The second focuses on a performance assessment over a range of
blowing coefficients. Thirdly, the results of a grid study to assess numerical
uncertainty are reported.
A.
Plenum
vs
No
Plenum
Steady RANS simulations of a baseline case at zero degrees angle of attack
were initially completed for the two grids, with and without plenum. The goal
was to determine the efficiency and accuracy for the no-blowing case,
so
as to
choose the method to complete all following simulations. When both simulations
were run to convergence (note that the plenum case is not shown to convergence
for plotting purposes), results showed good agreement, as can be seen in Fig. 10,
which shows the drag coefficient vs time-step number. To further illustrate the
similarity in both solutions, total velocity contours with streamlines for the
grid without the plenum and the grid with the plenum are shown in Fig.
11.
Although both grids converge to a similar value of lift and drag, what is of
importance is the total time to reach convergence. The case without the grid
obtained a converged solution at around
5000
iterations, whereas the grid with
the plenum is not yet completely converged at
20,000
iterations. Performance
parameters such as drag
are
considered converged when the values differ by
less than 0.02 of the previous value. Both grids had similar runtimes per iter-
ation; thus, when calculating the computational costs, one sees at least four
times the CPU runtime, and one extra CPU per simulation as a result of the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 524/637
STEADY CC SIMULATION FOR GACC WING
525
Fig. 10 Convergence comparison of drag coefficient for grids with and without
plenum.
added plenum block. The long time to reach convergence for the grid with the
plenum is caused by a lengthy pressure transient inside the plenum along with
continued slow pressure convergence throughout the simulation, even after the
initial transients.
B. Performance Assessment for Varying Blowing Coefficient
The fine grid w ithout the plenum was chosen for further simulations. A w ide
range of blowing coefficients was studied, and results were compared to exper-
iment and FUN2D simulations. Experimental data included the surface pressure
distribution for C, 0.059. The corresponding results from CFDSHIP are com -
pared to experiment, and are shown in Fig. 12. The simulation compares well to
experiment over the leading 95 of the airfoil. Simulation underpredicts the
magnitude of the maximum positive pressure by a factor of 2 and over predicts
the maximum negative pressure by a factor of 1.5. These locations correspond
to the two slot locations at x / c 0.975 and
0.985,
respectively. More investi-
gation needs to be carried out to further understand the discrepancy, and it
must be noted that experimental uncertainty is high in these regions because of
slow pressure leaks during e~ p er im en ta tio n .~
A plot of mean lift coefficient vs blowing coefficient is shown in Fig. 13.
CFDSHIP fine grid results are compared to experiment and FUN2D solutions.
The plot shows very good agreement of CFDSHIP results with experiment and
FUN2D results for C, .091. At higher values of C, where no experimental
data have been recorded, the results vary from FUN2D solutions. The variations
in FUN 2D and CFDSHIP resuls at the highest blowing coefficient are observable
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 525/637
526
W.
J.
BAKER AND E. G. PATERSON
XlC
x k
Fig. 11 Total velocity contours and streamlines
of
the baseline case for
computational dom ains a) without the plenum and b) with the p lenum.
by investigating the total velocity contours, shown in Figs.
14
and
15,
respect-
ively. FUN2D simulations predict the separation at the lower slot, whereas
CFDSHIP predicts the location of separation on the bottom side of the airfoil
back upstream at about 50 chord, as shown by the streamtraces in Fig. 15.
Initially
it
may seem that the CFDSHIP results are “unphysical.” Yet,
the phenomenon in which the jet reattaches and travels further up towards the
leading edge
(LE)
has been observed in experiment, and has been called the
“ d r a w dow n e f f e ~ t ” . ~ntil more experimental data are obtained,
it
is difficult
to know which of the FUN2D and CFDSHIP simulations is more accurate.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 526/637
STEADY CC SIMULATION FOR GACC WING
527
X
Fig. 12 Surface pressure distribution for experiment and simulation,C = 0.059.
Fig. 13 Lift
vs C
for experim ent and simulations.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 527/637
528
W.
J.
BAKER AND E. G. PATERSON
Fig 14 Mach contours for FUN2D simulations with
C, =
0 162:
Fig 15 Total velocity contours for CFDSHIP simulations with
C,
= 0 162
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 528/637
STEADY CC SIMULATION FOR GACC WING
529
Figure 16 shows the time history of the lift and drag coefficient for a wide
range of blowing coefficients. For
C
5 0.031, forces converge to a single
value. For larger blowing coefficients, forces begin to oscillate. As the blowing
coefficient increases, the amplitude of the oscillations increases, and the wave-
length of the oscillation increases. Figure 17 shows that the surface pressure
NarrdhrerrpknalThre
Fig. 16 a) Lift force and b) drag force h istories for a wide ran ge of C,.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 529/637
530
W.
J.
BAKER AND E. G. PATERSON
x/c
Fig.
17
Surface pressure at different intervals over one oscillation for C =
0 162
changes quite a bit along the TE over one oscillation for C
0.162. Other
blowing coefficients not illustrated in this work,
C
0.041, show similar
trends. This may explain the significant changes in the forces.
The turbulent kinetic energy is shown in Fig. 18 for low, moderate, and high
blowing coefficients. For the lowest blowing coefficient, C
0.021, there exist
two definitive regions of increased turbulent kinetic energy (TKE). The first,
denoted as a) in Fig. 18 is the interaction of the jet shear layer and the incoming
boundary layer from the top half of the airfoil beginning just aft of the je t orifice
and terminating at the jet separation. The second region of high TKE denoted as
b) in Fig. 18, originates near the je t separation and protrudes into the wake. At the
moderate blowing coefficient, C 0.059, the same interaction of the jet shear
layer and shear layer from the top half of the airfoil is observed, a) a smaller
second region of high TKE (hard to see in the figure) arises from the interaction
of the je t passing around the bottom corner of the slot and the recirculation zone
located along the inside comer of the bottom slot b). For the highest blowing
coefficient,
C
0.091, a) is the same as the previous two blowing coefficients,
and the second region of high TKE originates at the location of jet reattachment
past the bottom slot b).
C. Grid Study
A three-point grid study was completed for verification of results. Table
3
shows grid size and runtimes for each of the three grids used in the study.
These values coincide with non-time-accurate
RANS
simulations of 10,000 iter-
ations for each grid. The blowing coefficients used in the earlier work were now
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 530/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 531/637
532
W.
J.
BAKER AND E. G. PATERSON
Table 3 Grid size and runtime characteristics for grid study
Coarse Medium Fine
Grid points
97,575 193,980 381,810
Seconds/time step 1
o
2.8
6.5
W all-clock hours 3.6 9.9 16.7
CPU hours 29.1 79.1 133 .7
investigated using the coarse and medium grids, and results were compared to
each other and experiment. Figure 19 shows a plot of the mean lift coefficient
vs. blowing coefficient for the three grids studied. All three grids show agreement
to experiment for lower values of lift increment gain. At higher lift gain, the
coarse and medium results differ from the fine-grid results. It was determined
that coarse and medium grids were of inadequate fidelity to capture the
Coanda jet physics properly, in particular, the location of separation of the
Coanda jet because of insufficient near-wall spacing, which caused inaccuracies
in the prediction of the TKE in the buffer layer. To illustrate this point, surface
pressure plots for three blowing coefficients, C, 0.021, 0.059, and 0.091, are
shown in Fig. 20. These three cases coincide with instances in which all three
results show similar lift values
(C,
0.021), when the coarse result differs
from the fine and medium results (C, 0.059), and when the coarse and
medium results differ from the fine result C, 0.091). The surface pressure dis-
tributions look similar for all three grids for C, 0.021, and thus the similar
lift predicition is feasible. For
C,
0.059, the “drawdown effect” introduced
c,
Fig.
19
Lift
vs
C, for experiment and sim ulations for grid study.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 532/637
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 533/637
534
W.
J.
BAKER AND E. G. PATERSON
a)
Fig. 21 Plots of
y+
for coarse, medium, and fine grids at varying blowing
coefficients:
a)
C ,
=
0 021;
b)
C ,
=
0.059; c)
C ,
=
0 091
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 534/637
STEADY CC SIMULATION FOR GACC WING
a)
535
Fig. 22 Velocity contours for a) coarse,
b)
medium, and c) fine grids with
C,
= 0.59.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 535/637
536
W.
J.
BAKER AND E. G. PATERSON
previously is visible for the coarse grid, distinguishable by the pressure drop
along the pressure side of the wing along the aft 20 and the effects on the LE
of the airfoil. The same characteristics are seen for the coarse and medium grids
when
C
0.091. This drawdown effect explains the underprediction of the
lift forces. Figure 21 shows plots of y for the three grids and the previous
values of blowing coefficient, C 0.021, 0.059, and 0.091. For the plot with
C
0.021, all three grids show acceptable near wall resolution, y of approxi-
mately 1.00. For
C
0.059, the coarse grid shows a y value much larger
than 1.00 at both the upper and lower slots. For C
0.091, both the coarse
and medium grids have y values much larger than 1.00 at the low er slot. The
lower slot is an important location on which to focus, because the flow can
either reattach aft of the slot or stay separated. The importance of the aft slot is
demonstrated in Fig. 22, which shows total velocity contours and streamlines
for the three grids with C 0.059. Here, the “drawdow n effect” is visible for
the coarse grid, marked by the reattachment of the flow ahead of the lower slot.
The medium and fine grids do not show the drawdown. Recall that it was only
the coarse grid in which the
y
value was greater than 1.00. It is not presented
in this work, but for C 0.091, both the coarse and medium grids show this
jet reattachm ent ahead of the lower slot, whereas the fine grid does not.
To sum up the results from the grid study, the coarse, medium, and fine sol-
utions show monotonic divergence. Determining the proper near-wall spacing
for CC problems is an issue. Typically a flat plate approximation is used when
determining near-wall spacing during grid creation. Adjustments need to be
made to account for the highly curved surfaces. In this case, the flat plate approxi-
mation based on R eynolds number yielded a near-wall spacing of 2.00 x
For the fine grid, near-wall spacing was set at 2.00
x
lo p 6 , with the coarse-
grid near-wall spacing set at 4.00 x lop6.Even with the increased near-wall fide-
lity chosen, the medium and coarse g rids proved to be deficient at higher blowing
coefficients. W ithout a method to determine proper near-wall spacing require-
ments for CC applications, result validation becomes laborious and ineffective
with time and computational resources. A better technique needs to be developed
to determine CFD uncertainty for CC problems, ideally a single grid error esti-
mation p rocess.
VIII.
Conclusions
The GACC wing was studied using non-time-accurate, RAN S CFD. With
careful consideration, computational runtime could be decreased by modeling
the jet at the orifice instead of including the plenum and modeling the jet at
the diffuser nozzle exit, as shown in Figs.
7
and 9, respectively. After choosing
the most efficient and accurate grid, a study of the mean forces on the airfoil
for a wide range of blowing coefficients was completed, and results showed
good agreement with experiment and previous RANS efforts using FUN2D for
blowing coefficients C
0.091. For higher blowing coefficients, where no
experimental data are provided, CFDSHIP results differed from FUN2D
results. CFDSHIP simulations showed the presence of unsteady flow, perhaps
caused by the jet separation and interaction with the wake. A grid study was
performed to verify results, but showed monotonic divergence from the coarse
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 536/637
STEADY CC SIMULATION FOR GACC WING 537
to fine grid solutions. Both the medium and coarse grids had insufficient near-wall
spacing along the lower jet-slot, which affected the separation characteristics.
Future work includes recreating the grid to add in the tunnel walls and opti-
mizing the near-wall spacing. This will determine what effects the interaction
between the wake and the tunnel walls have on the source of unsteadiness.
Some early indication from experiment is that there was interaction between
the wake and tunnel walls, but no quantitative value could be given yet. Other
means to address this include using time-accurate RANS to investigate
whether the oscillations shown are a product of the computational model, that
is, the large dom ain, or a result of non-time-accurate simulations.
Acknowledgments
The authors acknowledge the support of the Advanced Submarine Systems
Development Office of the Naval Sea Systems Command, SEA 073R
(Program Manager; Meg Stout) in the form of a graduate student fellowship
for the first author, and the Office of Naval Research through Grant Number
N00014-03-1-0122 (Program Officer; Ron Joslin) for the second author. Also,
the authors would like to acknowledge the DoD High Performance Computing
Modernization Office (HPCMO) and Army Research Laboratory-Major
Shared Resource Center (ARL-MSRC) for providing computing resources
through a DoD HPCMO Challenge Project.
References
‘Englar, R. J., Stone, M. B., and Hall, M, “Circulation Control-An Updated Bibli-
ography of DTNSRDC Research and Selected Outside References,” DTNSRDC Rept.
77-0076, Sep. 1977.
’Abramson, J., “Two-Dimensional Subsonic Wind Tunnel Evaluation of Tw o Related
Cambered 15-Percent Circulation Control Airfoils,” DTNSRDC A SED-373, Sept. 1977.
3Rogers, E. O., and Donnelly, M. J., “Characteristics of a Dual-Slotted Circulation
Control Wing of Low Aspect Ratio Intended for Naval Hydrodynamic Applications,”
42nd AIAA Aerospace Sciences Meeting Exhibit, AIAA Paper 2004-1244, Jan. 2004.
4Jones, G.
S.,
Viken,
S.
A., Washburn, L. N., Jenkins, L. N., and Cagle, C. M., “An
Active Flow Circulation Controlled Flap Concept for General Aviation Aircraft Appli-
cations,” AIAA Paper 2002-3157, Jan. 2002.
’Paterson, E. G., and Baker, W. J., “Simulation of Steady Circulation Control for
Marine-Vehicle Control Surfaces,” 42nd AIAA Aerospace Sciences Meeting and
Exhibit, AIAA Paper 2004-0748, Jan. 2004.
6Jones, G.
S.,
and Engle, R. J., “Advances in Pneumatic-Controlled High-Lift Systems
Through Pulsed Blowing,” 21 st Applied Aerodynamics Conference, AIAA Paper 2003-
341 1, June 2003.
’Anderson, W.
K.,
and Bonhaus, D. L., “An Implicit Upwind Algorithm for
Computing Turbulent Flows on Unstructured Grids,”
Computers Fluids
Vol. 23, No. 1,
‘Menter,
F.,
“Two-Equation Eddy Viscosity Model for Engineering Applications,”
’Suhs, N., Dietz , W., Rogers,
S.,
Nash,
S.,
and Onufer, J. T., “PEGASU S User’s Guide
1994, pp. 1-21.
AIAA
Journal,
Vol. 32, No.
8,
1994, pp. 1598-1605.
Version 5. lg ,” Tech. Rept., NASA, May 2000.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 537/637
Chapter 21
Computational Study of a Circulation Control
Airfoil Using FLUENT
Gregory M cGowan* and Ashok Gopalarathnamt
No rth Carolina State University, Raleigh, North Carolina
Nomenclature
A
= area
b
=
wing span
c
=
chord
Cd
= drag coefficient
Cl
=
lift coefficient
C ,
=
pitching moment coefficient about quarter chord
C , = momentum coefficient
h
=
slot height
M
=
Mach number
riz = mass flow rate
P
=
pressure
q
=
dynamic pressure
R = gas constant for air
r
=
radius of Coanda surface
Re
=
Reynolds number
= arc length, measured from the slot exit around the upper surface
of the airfoil
T
=
temperature
U
= velocity magnitude
w =
slot width, equal to b for two-dimensional flows
a =
angle of attack
y
=
ratio of specific heats
p
=
viscosity coefficient
*Graduate Research Assistant, Department of Mechanical a nd Aerospace Engineering.
'Associate Professor, Department of M echanical and Aerospace Engineering.
Copyright 005 by
the
authors. Published by the American Institute of Aeronautics and Astro-
nautics, Inc., with permission.
539
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 538/637
540
G. McGOWAN AND A. GOPALARATHNAM
p
=
density
Subscripts
duct
=
stagnation conditions inside plenum
fc
=
conditions at flow-control boundary
3
=
freestream conditions
J
= slot-exit conditions
I. Introduction
ECENT research in the Applied Aerodynamics G roup at the North Carolina
R
tate University (NCSU ) has led to the developm ent of an automated cruise-
flap system.lq2 he cruise flap, introduced by P f e n n i ~ ~ g e r , ~ ~ ~s a small trailing-
edge (TE) flap that can be used to adapt an airfoil and increase the effective
size of the low-drag range of natural-laminar-flow (NLF) airfoils. The au tomation
is achieved by indirectly sensing the leading-edge (LE) stagnation-point location
using surface pressure measurements and deflecting the flap
so
that the stagna-
tion-point location is maintained at the optimum location near the LE of the
airfoil. Maintaining the stagnation point a t the optimum location results in favor-
able pressure gradients on both the upper and lower surfaces of the airfoil. With
such a cruise-flap system, the airfoil is automatically adapted for a wide speed
range. This automated cruise-flap system was successfully demonstrated in the
subsonic wind-tunnel at NCSU.2
Although the use of a cruise flap on an NL F airfoil results in low drag over a large
range of flight speeds, there is a need for a revolutionary approach that integrates
the achievement of significantly lower drag over a large range of operating
speeds with the capab ility for generating very high lift at takeoff and landing con-
ditions. Toward this objective, it is of interest to study an approach that integrates
aerodynam ic adaptation with the well-established high-lift capability of circulation
control (CC) aerodynamics. Circulation control is not a new concept; it has been
around since the late 1930s. The majority of research efforts have focused on
blowing in a positive, or downward, direction at the TE of the airfoil. Early
efforts accomplished this downward inclination using a jet of high-velocity air
blown straight out of the TE at the desired angle.5 This pneumatic-flap concept
has been studied theoretically and experimentally by several researchers over the past
several decades?-'' As time has progressed, more researchers have begun to take
advantage of the Coanda by blowing over a round TE. This Coanda-
based CC is currently attracting significant interest as a means of achieving high
lift.
This aerodynamic adaptation, when achieved using a blown cruise flap, carries
with it the potential for significant skin-friction d rag reductions through ex tensive
laminar flow in addition to the high-lift benefits of CC aerodynamics. Figure 1
illustrates the overall concept. In a manner similar to that of a cruise flap, it is
believed that by utilizing this stagnation-point sensing scheme, an adaptive CC
airfoil, with a blown cruise flap, can achieve extensive laminar flow over a
large lift-coefficient range.
As a first step toward the long-term goal of studying an adaptive CC airfoil,
the current effort was undertaken for establishing and validating computational
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 539/637
STUDY OF CC AIRFOIL USING FLUENT
54
high-speed
cruise
condition
Fig. 1 Illustration of the NCSU concept of an adaptive CC airfoil.
fluid dynamics (CFD) analysis procedures for blown-TE airfoils. The CFD
package used for this work w as the FLUENT flow solver. The results are com-
pared to CFD and experimental data obtained from a recent study by Jones
et al? of a General Aviation CC (GA CC ) airfoil conducted at the NASA
Langley Research Center. Because previous CFD studies on this airfoil did not
include tunnel walls, the current CFD study also includes an investigation of
the effect of tunnel walls on the solution. To provide a foundation for the adap tive
CC airfoil concept, the effects of CC on the LE stagnation-point location were
also exam ined in the current w ork.
The following section gives an explanation of the geometry under examin-
ation and information about the experimental setup. Then a description of the
numerical approach is presented, including grid details, solver settings, and
boundary conditions. Results are then presented for two cases:
1)
solution in
free-air or the infinite domain, and 2) solution with the presence of wind-
tunnel walls. Results are also shown for a stagnation point study, in an effort
to show how the stagnation point moves with changing blowing rates.
11. Configurations and Experiments
The geom etry chosen for the current research w as the GA CC airfoil, designed
by Jones.15 The GACC airfoil was derived from a
17
GAW(
1)
airfoil by mod-
ifying the TE to incorporate a 2
r / c
Coanda surface and is shown in Fig. 2.
Fig. 2 General Aviation Circulation Control (GACC) airfoil geometry
current research.
used in the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 540/637
542 G. McGOWAN AND A. GOPALARATHNAM
The wind-tunnel experimen ts were conducted by Jones et al.15 in the Basic Aero-
dynamic Research Tunnel (BART),which is located at the NASA Langley
Research Center in Hampton, Virginia. The BART tunnel has a physical test-
section size of 28
x
40
x
120 in. The G ACC model chord length was
9.4
in.,
with angle of attack changes made about the half-chord location. Further
details of the experimental setup are given in Ref. 16.
111.
Numerical Approach
The commercial
flow-solver code FLUENT version 6.1 was used in the
current research. Grid generation was performed using GAMBIT, which is the
preprocessor packaged with the FLUENT code. These codes were used to
study two cases. The first case involves the examination of the GACC airfoil
in free air with the objective of comparing the FLUENT two-dimensional
results to CFD and wind-tunnel results presented in Ref. 15. It should be noted
that the CFD solutions obtained in Ref. 15 did not include the effect of wind-
tunnel walls. The second case involves two-dimensional simulations of the
GACC airfoil in the BART facility to examine the influence of tunnel walls on
this particular airfoil. Results from FLUENT were obtained for a matrix of 15
data points for each of the two cases.
A. Grid Details
For the first study, a circular computational domain (Fig. 3 ) was generated that
extends to approximately 20 chord lengths in all directions and is composed of
132,762 cells. For the study of wall effects, a second two-dimensional grid was
generated to include the wind-tunnel upper and lower walls and is shown in
Fig. 4. For the computation with walls, a separate grid was generated for each
angle of attack, each of which comprises 123,602 cells and extends to 20
chord lengths upstream and downstream of the airfoil.
The grids for all of the analyses are hybrid unstructured grids. The domains
consist of an unstructured grid far from the airfoil in order to reduce the
number of cells and a structured grid near the airfoil to maintain good
Fig.
3
Grid generated for the free-air analyses using FLUENT.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 541/637
STUDY OF CC AIRFOIL USING FLUENT
543
Fig. 4 Grid generated for FLUENT study of wall effects.
resolution through the boundary and shear layers. For both cases, minimum wall
spacing was chosen such that
y +
<
1
at the wall.
B. Solver Settings
For the curre nt study the solution is assumed to be steady and is not run time-
accurate. The coupled-implicit solver was chosen with second-order upwind
node-based discretization for both the flowfield and turbulence equations. The
coupled solver was chosen for two reasons. First, compressibility effects need
to be modeled, because the Mach number at the slot exit can often approach
the sonic condition as the blowing rate is increased. Secondly, the FUN2D1'
code has a compressible solver, and because the results from the current study
were compared with FUN2D results, a compressible solver was also used for
the FLU EN T analysis. There was an attempt to run these problems with the seg-
regated (decoupled) solver using very low relaxation factors; however, it was
found that for the cases with larger blowing rates, the solution began to exhibit
an unsteady effect afte r a few thousand iterations. In order to com pare with the
FUN2D results of Ref. 15, the one-equation Spa lart- Allmaras turbulence
model was chosen for the current work. Wall functions were not used in the
FLUENT calculations.
C. Boundary Conditions
FLU EN T does not allow the user to input the freestream Ma ch numbe r and
Reynolds number directly. Instead, the freestream velocity and operating
pressure were calculated using Eqs. (1-3) and provided as inputs for the ana-
lyses. Th e Ma ch and Reynolds numbers were set to 0.l and 533,000, respectively,
to match those used in Ref. 15. The results were used for both cases, with and
without tunnel walls:
Uw
=
MwJ3/RTm
RePW
Pw
=
w c
An approximate method was developed to estim ate the velocity required a t the
flow control boundary U f c ) o achieve a desired
C,,
CPdesired.his method
assumes incompressible flow throughout the duct, and was derived by solving
the continuity equation. The equation for
U f c
rom this approximate method is
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 542/637
544
given in Eq.
(4):
G. McGOWAN AND A. GOPALARATHNAM
Once FL UENT converged, an integration was performed across the slot exit as
shown in Eq. 5 ) to ob tain the actual
C
of the je t at the slot. This C,, however, is
different from
CPdeslred
ecause the Ufc for the latter is set using an approximate
method.
Furthermore, to be consistent with the methods used for calculating C in
Ref.
15,
all of the
C
values presented in this paper were calculated using isen-
tropic flow relations? The equations for this procedure are given in Eqs. (6-8).
To determine how close the isentropic C is to the integrated C the two values
are compared in Fig. 5 for several cases. The C values indicated along the hori-
zontal axis are values calculated using the isentropic relations. Values for C on
the vertical axis were computed by integrating the flow across the slot exit. The
solid line in Fig.
5
indicates where the data points would lie if the two methods
generated the same values for
C,.
The symbols are representative of the actual
values calculated using FLUENT and isentropic relations. Although the differ-
ences are very small, approximately 3 at the highest blowing coefficient,
care must be taken to ensure consistency in the C FD solutions and experiments:
riZ =
PJUJAJ
(6)
Fig.
5
Comparison
of
Cphkgm dith
CpbeotroPic
or
a
=
0;
the stra ight line is included
to indicate deviation from a perfect correlation.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 543/637
STUDY OF
CC
AIRFOIL USING FLUENT 545
IV.
Results
The results from FLUENT predictions for the GACC airfoil are presented
in three parts. In the first part, the prediction for the GACC airfoil in free-
air conditions is compared with the results presented in Ref.
15.
In the
second part, the predicted results for the GAC C airfoil with tunnel walls are pre-
sented and compared with the free-air results. In the third part, the effects of a
and C on the LE stagnation-point location are presented and discussed.
A.
Results for Free-Air Conditions
In this part of the study, FLUENT results for free-air conditions are
compared with CFD and experimental results from Ref. 15. The overall com-
parison between the FLUENT results and experimental results is illustrated
using Cl-a curves in Fig.
6
The results from FLUENT analyses consist of a
matrix of 15 data points for a =
- 5
0, and
5
deg and
C
= 0, 0.008,
0.024
4
3.5
3
2.5
CI 2
1.5
1
0.5
0
1 0
5
0
5 10 15
degrees)
Fig. 6
experimental results from Ref.
15
(data points and curve
fits
for each
Cp .
Comparison of NCSU FLUENT results from the current work with Langley
−10 −5 0 5 10 150
0.5
1
1.5
2
2.5
3
3.5
4
α (degrees)
Cl
Fluent
Calculations
Experimental Results
Curve Fit to
Experimental Data
Cµ= 0.078→
Cµ= 0.047→
Cµ= 0.024→
Cµ= 0.008→
←Cµ= 0.0
←Cµ= 0.060
←Cµ= 0.041
←Cµ= 0.025
←Cµ= 0.015
←Cµ= 0.007
←Cµ= 0.0
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 544/637
546 G. McGOWAN AND A. GOPALARATHNAM
Table
1
FLUENT results for the free-air cases
Lift coefficient,
CL
Blowing coefficient
(C,)
=
-5 d e g a = O d e g
=
5 deg
0.000
0.008
0.024
0.047
0.078
0.090
0.382
1.082
1.979
3.045
0.666 1.193
1.009 1.486
1.646 2.080
2.544 2.7 19
3.206 3.296
0.047, and 0.078, and are presented in Fig.
6
using solid lines and square
markers. The FLUENT data used to generate Fig. 6 are given in Table 1.
The wind-tunnel results from Ref. 15. are presented as circular markers with
the dashed lines in Fig.
6
representing best-fit curves for several angles of
attack and for C = 0, 0.007, 0.015, 0.025, 0.041, and 0.060. The values of
C for the FLUENT results differ from those for the results of Ref. 15
because of the difference between the actual
C
and the desired
C?
when
using the approximate method in Eq. (4) for estimating the Ufc using mcom-
pressible-flow equations.
Although the values of
C
for the FLUENT results do not match those
for the results of Ref. 15, it is clear that the trends and most of the predictions
for the Cl are close to those from Ref. 15. In particular, the FLUENT predic-
tions for
C =
0, 0.008, and 0.047 agree quite well with the results for
similar values of
C
from Ref. 15. Two discrepancies between the FLUENT
predictions and those from Ref. 15 are apparent: 1) for
C
= 0.024 and 2)
for
C
= 0.078. The reason for the first discrepancy in the results is attributed
to the incorrect prediction of the jet-separation location on the Coanda surface
for
C =
0.024. The apparent discrepancy in the results for
C =
0.078 is
attributed to nonlinear effects at the high blowing rates and the fact that the
highest blowing rate in the results of Ref. 15 is for C
=
0.060.
The flowfield data for the FLU ENT results are presented in two parts. In the
first part, the effects of increasing
C
for a constan t angle of attack are presented.
The second part exam ines the effects of angle-of-attack changes and their influ-
ence on the CC airfoil for a constant C,. The flowfield data are presented as
streamline plots; these serve as visual aids in the understanding of the effects
of CC on the flow over the airfoil.
The first part of the flowfield data is shown in Figs. 7a-7c. It can be seen
that as the blow ing rate is increased the streamlines become more curved-an
indication of increased circulation. The second part of the flowfield data is
shown in Figs. 8a-8c and Figs. 9a-9c to illustrate the effects of changing
the angle of attack while holding blowing rates constant. The results are
presented for two blowing rates: the mild blowing case
C
= 0.047 and
the highest blowing rate C = 0.078. The results show that changes to C
have a significant effect on the jet-separation location and the resulting Cl.
In comparison, changes to have a much smaller effect on the jet-separation
location.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 545/637
STUDY OF
CC
AIRFOIL USING
FLUENT
547
Fig. 7 Circulation control effects on the flowfield
at
a
=
0
deg for various values
of
C,:
a)
C =
0.000; b)
C =
0.047;
c) C =
0.078.
B.
Wind Tunnel Wall Effects
In this subsec tion, the FLU ENT results for the GA CC airfoil with the effect of
wind-tunnel upper and lower walls are presented . Figures
10
to 12show the influ-
ence of the walls on the CF D solution. These figures present the predicted Cl as a
function of C for a = 0, 5 , and - deg, respectively. Figure 10 also includes a
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 546/637
548
a)
G. McGOWAN AND A. GOPALARATHNAM
Fig. 8 Circulation control effects on flow field at C, = 0.047 for various values ofa
a)
cu
=
-5
deg;
b)
cu
=
0
deg; c)
cu
=
5
deg.
comparison w ith results for experiment and the FUN2D study15 for
a
= 0 deg,
the only angle of attack for which the FUN2D results were presented in
Ref. 15. The FUN2D simulations in Ref. 15 did not include any wind-tunnel
wall effects. Figures 10-12 indicate that the presence of walls has very
little influence on the CFD solution. Because the study was performed on a
two-dimensional grid, it can be stated that blockage effects are minimal;
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 547/637
STUDY OF CC AIRFOIL USING FLUENT
549
Fig. 9 Circulation control effects on flowfield at C = 0.078 for various values of (Y:
a
(Y = -5
deg;
b) (Y = 0
deg; c
(Y =
5 deg.
however, no conclusion can be drawn for the three-dimensional effects due to
side-wall boundary layer effects and the associated trailing vortices. Because
of the large lift that these configurations produce, it is believed that three-
dimensional effects will be extremely important at larger blowing rates.
The results for the with-walls simulations consistently show that for low
blow ing coefficients, the Cl values are predicted to be lower than those for the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 548/637
550
G. McGOWAN AND A. GOPALARATHNAM
Fig. 10 FLU ENT prediction of wind-tunnel wall effects for varying values of
C,
at
a
=
0
deg.
0 0.01 0.02 0.03 0.04 0
c
i 0.06 0.07 C
Fig.
11
FLU ENT prediction of wind-tunnel wall effects for varying values of C, at
a = 5 deg.
0 0.02 0.04 0.06 0.08 0.10.5
1
1.5
2
2.5
3
3.5
4
Cµ
Cl
Fluent (with walls)Fluent (free−air)FUN2D Jones et al. (free−air)
Experiment Jones et al.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 549/637
STUDY OF CC AIRFOIL USING FLUENT
551
Fig.
12
FLU ENT prediction of wind-tunnel wall effects for varying values of
C
at
a =
-5deg.
free-air simulations. However, at the largest blowing coefficients, the trend
reverses and
Cl
values with walls are predicted to be higher than those without
walls. The FLUENT data accrued for the cases with wind-tunnel walls are
given in Ta ble 2.
C. Stagnation-Point Location
The motivation for examining the LE stagnation-point behavior is that the
stagnation-point location was used successfully in earlier research',* for
closed-loop control of a TE flap. It was, therefore, desirable to examine the
CFD solutions for the CC airfoils to see if there was any evidence that would
suggest that a s imilar approach cou ld be exten ded for use with C C airfoils.
Table 2 FLUEN T resu lts for cases with w ind-tunnel walls
Lift coefficient, CI
Blowing coefficient
cc
=
-5 d e g a = O d e g
=
5 deg
0.000
0.008
0.024
0.047
0.078
0.09 1
0.388
1.063
1.892
2.893
0.702 1.247
1.027 1.491
1.671 2.070
2.475 2.711
3.044 3.178
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 550/637
552
1.15
1.1
SIC
1
OE
1
G. McGOWAN AND A. GOPALARATHNAM
+
=
0.000
+
=
0.008
Cw
=
0.024
-
C =
0.047
+ = 0.078
0 5
1
..............................
//
..........................
Fig.
13
Circulation control effects on
LE
stagnation-point ocation.
Stagnation-point location, measured as an arc length from the jet exit around
the upper surface of the airfoil, as a function of Cl, is presented in Fig.13. Each
line in Fig. 13 represents a different blowing rate and fo r each blowing coefficient
there are three points that correspond to three different angles of attack (- 5,
0,
and 5 deg). From Fig. 13 it can be seen that the stagnation point moves in a pre-
dictable manner, both with angle of attack and with changing blowing rate. This
behavior provides an indication that the stagnation-point location can be used as a
means to develop closed-loop control of the je t C, on CC airfoils.
V. Conclusions
The results from a two-part CFD study using the FLUENT flow solver have
been presented. Results of the first study show that, although the FLUENT pre-
dictions do not match the CFD and experimental results of Ref. 15 exactly, the
overall trends are followed very closely. Throughout the range of blowing coeffi-
cients, with the exception of the no-blowing case C, = O.O> FLUENT consist-
ently predicted a slightly lower overall lift coefficient.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 551/637
STUDY OF CC AIRFOIL USING FLUENT
553
The second study focused on the influence of wind-tunnel walls on the CFD
solution. For low blowing coefficients, it was found that the lift is predicted to
be lower for the cases with walls. The trends are reversed for the higher
blowing coefficients, for which the cases with walls yield a higher predicted
lift. Although the solutions are different, the differences are small, and could
as well be attributed to differences in the grids rather than the actual presence
of walls.
Th e influence of C C on the LE stagnation-point location was examined. I t was
shown that changes in blowing rate and angle of attack result in systematic
changes to the stagnation-point location. Th is observation indicates that it is poss-
ible to use a closed-loop control system that is driven by sensing the stagnation-
point location.
Acknowledgments
Th e authors would like to acknowledge the fund ing for this research through a
grant from the NASA Langley Research Center and the National Institute of
Aerospace. The technical monitor, Greg Jones of NASA Langley, is thanked
for many valuable discussions and for the geometry of the GACC airfoil and
the wind-tunnel test results. In addition, Greg Stuckert from FLUENT Inc. and
Hassan Hassan of NCSU are thanked for their advice regarding the CFD
simulations.
References
‘ M c A v o ~ ,C. W., and Gopalarathnam, A., “Automated Cruise Flap for Airfoil Drag
Reduction over a Large Lift Range,” Journal of Aircraft, Vol. 39, No. 6, 2002, pp.
*MCAVOY,. W., and Gopalarathnam, A ., “Automated Trailing-Edge Flap for Airfoil
Drag Reduction Over a Large Lift-Coefficient Range,” AIAA Paper 2002-2927, June
2002.
3Pfenninger, W., “Investigation on Reductions of Friction on Wings, in Particular
by
Means of Boundary Layer Suction,” NACA TM 1181, Aug. 1947.
4Pfenninger, W., “Experiments on a Laminar Suction Airfoil of 17 Per Cent Thick-
ness,” Journal of the Aeronautical Sciences, April 1949, pp. 227-236.
’Davidson, I. M., “The Jet Flap,”
Journal of the Royal Aeronautical Soc iety, Vol.
60,
No. 1, 1956.
pence, D. A., “The Lift Coefficient of a Thin, Jet-Flapped Wing,” Proceedings of the
Royal Society Series A ,
Vol. 238, No. 121, 1956.
’Spence, D. A., “Some Simple Results for 2-Dimensional Jet-Flap Aerofoils,” The
Aeronautical Quarterly, 1958, pp. 395-406.
‘Garland, D. B., “Jet-Flap Thrust Recovery: Its History and Experimental Realization,”
Canadian Aeronautics and Space Jo urnal, May 1965, pp. 143-151.
’Lissaman, P. B.
S. ,
A Linear Solution f o r the Jet Flap in Ground Effect, Ph.D. Thesis,
California Inst. of Technology, Pasadena, CA, 1965.
“Aiken, T. N., and Cook, A. M., “Results of the Full-Scale Wind Tunnel Tests on the
H-126 Jet Flap Aircraft,” NA SA TN D-7252, April 1973.
981 -988.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 552/637
554 G. McGOWAN AND A. GOPALARATHNAM
l1
Abramson, J., Rodgers, E., and Taylor, D., “High-speed Characteristics of Circulation
Control Airfoils”, AIAA Paper 83-0265, 1983.
‘’Wood, N., and Nielsen, J., “Circulation Control Airfoils Past, Present, Future,” AIAA
Paper 1985-0204, 1985.
13Novak, C. J., and Cornelius,
K.
C., “An LDV Investigation
of
a Circulation Control
Airfoil,” AIAA Paper 86-0503, 1986.
14Novak,C. J., Cornelius,
K.
C., and Roads, R.
K.,
“Experimental Investigations of the
Circular W all Jet
on
a Circulation Control Airfoil”, AIAA Paper 87-0155 , 1987.
15Jones,
G. S.,
Viken,
S.
A., Washburn, A. E., Jenkins, L. N., and Cagle, C. M.,
“An Active Flow Circulation Controlled Flap Concept for General Aviation Aircraft
Applications,” AIAA Paper 2002-3 157, 2002.
“ka gle , C. M., and Jones, G. S., “A Wind Tunnel Model to Explore Unsteady Circula-
tion Control for General Aviation Applications,” AIAA Paper 2002-3240, 2002.
”Jones,
G. S .
and Englar, R. J., “Advances in Pneumatic-Controlled High-Lift Systems
Through Pulsed Blowing,” AIAA Paper 2002-341 1, 2003.
“Anderson, W.
K.,
and Bonhaus, D. L., “An Implicit Upwind Algorithm for Computing
Turbulent Flows on Unstructured Grids,”
Computers Fluids
Vol. 23, No. 1, 1994,
pp. 1-21.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 553/637
1II.D. Tools for Predicting Circulation Control
Performance:
Additional CFD Applications
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 554/637
Chapter 22
Computational Evaluation of Steady and Pulsed Jet
Effects on a Circulation Control Airfoil
Yi
Liu,* Lakshm i
N.
Sankar,+ Robert J. Englar,'
Krishan K. Ahuja,$ and Richard Gaetall
Georgia Institute
of
Technology, Atlanta, Georgia
Nomenclature
a = angle of attack
Aje t=
area of jet slot, ft2
CL,Cl = lift coefficient
C,, Cd= drag coefficient
C, =jet momentum coefficient
f = pulsed jet frequency, Hz
m =jet mass flow rate, slugs/s
s= wing area, ft2
St =
Strouhal number
C,, = averaged jet mom entum coefficient
Lref= length reference, in.
TJet To,,,,
=
temperature and total temperature of the jet, K
Pj,, = pressure of the jet, psia
V , = freestream velocity, ft/s
V,,, = je t velocity, f t/s
pjet,
p,
= densities, slugs/ft3
*Research Scientist, National Institute of Aerospace. Member AIAA.
'Regents Professor, School of Aerospace Engineering. Associate Fellow AIAA.
'Principle Research Engineer, Georgia Tech Research Institute. Associate Fellow AIAA.
%Professor, School of Aerospace Engineering. Fellow A IAA .
TResearch Engineer, Georgia Tech Research Institute. Senior Member AIAA.
Copyright
005
by the American Institute of Aeronautics and Astronautics, Inc. All rights
reserved.
557
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 555/637
558
YI LIU ET AL.
I Introduction
URING
the past several decades, there has been a significant increase in
D
ir travel and a rapid growth in commercial aviation. At the same time,
environmental regulations and restrictions on aircraft operations have become
issues that affect and limit the growth of commercial aviation. In particular,
the noise pollution from aircraft, especially around airports, has become a
major problem that needs to be solved. Reducing aircraft noise has become a
priority for airlines, aircraft manufacturers, and NASA researchers. In response
to this challenge, NASA has proposed a plan to double aviation system capacity
while reducing perceived noise by a factor of two (10 dB) by 201
1,
and to triple
system capacity while reducing perceived noise by a factor of four (20 dB) by
2025.
Large commercial aircraft are dependent on components that generate
high levels of lift at low speeds during takeoff or landing so that they can
use existing runways. Conventional high-lift systems include flaps and slats,
with the associated flap-edges and gaps, are significant noise sources. Since
the mid-l980s, many researchers have pointed out that the airframe noise
predominantly emanates from high-lift devices and the landing gear of subsonic
a i r ~ r a f t . ” ~epending on the type of aircraft, the dominant source varies between
flap, slat, and landing gear.4 Furthermore, these high-lift system s also add to the
weight of the aircraft, and are costly to build and maintain.
An alternative to the conventional high-lift systems is circulation control w ing
(CCW) technology. This technology and its aerodynamic benefits have been
extensively investigated over many ears through experimental studies.596A
limited number of numerical analy~e$~-~ave also been carried out. Work has
also been done on the acoustic characteristics studies of C C wings.’ These
studies indicate that very high CL values (as high as 8.5 at a = 0 deg) may be
achieved with CCW. Because many mechanical components associated with
the high-lift system are no longer needed, the wings can be lighter and less
expensive to build.
lo
Major airframe noise sources such as flap-edges, flap-
gaps, and trailing/leading edge flow separation can all be eliminated with the
use of CCW systems.
Earlier designs of CCW configurations used airfoils with a large-radius
rounded trailing edge (TE) to maximize lift production. These designs also
produced very high drag.” Such high drag levels associated with a blunt,
large-radius TE can be prohibitive under cruise conditions when CC is no
longer necessary. To overcome this difficulty, an advanced CCW section,
called a circulation hinged flap,596 as been developed to replace the traditional
rounded TE CC airfoil. This concept, originally developed by Englar, is shown
in Fig. 1. The upper surface of the CCW flap is a large-radius arc surface, but
the lower surface of the flap is flat. The flap could be deflected from 0 to
90 deg. When an aircraft takes off or lands, the flap is deflected as in a conven-
tional high-lift system, and CC is deployed. The large curvature of the upper
surface produces a large je t turning ang le, leading to high lift. When the aircraft
is in cruise, the flap is retracted and a conventional sharp TE shape results, greatly
reducing the drag. This kind of flap does have some moving elements that
increase the weight and complexity over the earlier CCW design. However,
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 556/637
EVALUATION OF STEADY AND PULSED JET EFFECTS
559
Supsrwidcal Contour C C W
Flap
Fig.
1
Dual radius CCW airfo il with LE b l ~ w i n g . ~
overall, the hinged flap design still maintains most of the advantages of the CC,
while greatly reducing the drag in cruising condition associated with the rounded
TE CCW design.
To understand and quantify the aeroacoustic characteristics and benefits of
the CCW, Munro, Ahuja, and Englar12-15 have recently conducted several
acoustic experiments comparing the noise levels of a conventional high-lift
system with those of an advanced CC wing at the same lift setting. The
present computational fluid dynam ics (CFD) study16 is intended to complement
this work, and numerically investigates the aerodynamic characteristics and
benefits associated with this CC airfoil. Computational fluid dynamic studies
such as the one presented here can also help in the design of future generation
CCW configurations.
The present work is an extension of a previous work where two-dimensional
studies of the effects of steady and pulsed jets on the CCW configuration were
carried 0 ~ t . l ~he objective of this study is to isolate and quantify the effects
of parameters such as leading edge (LE) blowing, freestream velocity, jet slot-
height, and frequency on the performance of two-dimensional steady and
pulsed C C jets . The unsteady Navier-Stokes methodology used here has also
been applied to study a three-dimensional CC wing, and to model tangential
blowing effects.16
11. Mathematical and Numerical Formulation
A. Governing Equations
In the present work, the Reynolds-averaged Navier-Stokes (RANS) equations
were solved using an unsteady three-dimensional viscous flow solver. A semi-
implicit finite-difference scheme based on the Alternating Direction Implicit
(ADI)18,19method was used. This scheme is second- or fourth-order accurate
in space and first-order accurate in time. This solver can model flowfields over
isolated wing-alone configurations. Both time-accurate and local time step
methods can be used in this solver. For the current study, the time-accurate
method is used to predict the unsteady effects. The time step is chosen based
on the Courant-Friedrichs-Lewy (CFL) condition.
This solver has been validated for clean and iced wings by Kwon and Sankar?'
and Bangalore et a1.21 Modifications to this so lver have been made to model
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 557/637
560
YI LIU ET AL.
CC jets. l6 Both three-dimensional finite wings and two-dimensional airfoils may
be simulated with the same solver. The flow around the airfoil is assumed to be
fully turbulent, so currently no transition models are used. Two turbulence
models have been used: the Baldwin-Lomax22 algebraic model and the
Spalart and A11ma1-a~~~ne-equation model. In this work, all the calculations
were done using the Baldwin-Lom ax model. The effects of the turbulence
model are discussed in Ref. 16.
B. Com putational Grid
Construction of a high-quality grid around the CCW airfoil is made difficult
by the presence of the vertical jet slot. In this solver, the jet slot is treated as
part of the airfoil surface, as done by S h r e ~ s b u r y , ~ ~ , ~ ~nd Williams and
Franke.26A hyperbolic three-dimensional C-H grid generator is used to generate
the grid. The single-block three-dimensional grid is constructed from a series
of two-dimensional C-grids with an H-type topology in the spanwise direction.
The normal distance of first grid layer to the airfoil surface is set to lop5
chord length to maintain enough points in the boundary layer. The grid outer
boundaries are set to 10 chord lengths away to satisfy nonreflective boundary
conditions. The grid is also clustered in the vicinity of the jet slot and the TE
to accurately capture the jet behavior over the airfoil surface. From our
studies, the TE spacing should be less than
lop3
chord length in the streamwise
direction, and enough points should be placed in the wake region to accurately
capture the jet flow behavior. Grid studies have been carried out for different
meshes, and results are shown in Ref. 16.
The grid generation and the surface boundary condition routines are general
enough so that one can easily vary the slot location, slot size, blowing velocity
and the direction of blowing.
C. Boundary Conditions
defined as follows:
In CCW studies, the driving parameter is the momentum coefficient C,,
mVjet
1/2p,v:s
Here, the jet mass flow rate is given by
Conventional airfoil boundary conditions are applied everywhere except at
the jet slot exit. Nonreflection boundary conditions are applied at the outer
boundaries of the C grid to allow characteristic waves [for example, Riemann
invariant 2 a / (y 1)
u ]
to leave. On the airfoil surface, adiabatic and
no-slip boundary conditions are applied, and the normal derivative of the pressure
is set to zero.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 558/637
EVALUATION OF STEADY AND PULSED JET EFFECTS
561
At the jet slot exit, the jet is assumed to be subsonic, and the following con-
ditions are specified: total temperature of the jet Tojet, mom entum coefficient
C,
as a function of time, and the flow angle at the exit. In the simulation, the
jet was tangential to the airfoil surface at the exit.
For exam ple with subsonic jets, one characteristic can propagate upwind
into the slot. Thus the pressure at the jet exit is extrapolated from the outside
values. Then the static pressure at the jet slot exit can be obtained as
Pj
=
Pi1
=
(4PQ- Pi3)/3 (3)
From Eqs. (1) and (2), the momentum coefficient can also be expressed as
Pjet
7:tAjet
/ J 1/2pwVLS
-
(4)
From the ideal gas law and the equation of state, the following relations can be
obtained:
Substituting Eq. 5 ) into Eq. (4), another expression for C, with just one
unknown parameter can be obtained:
The only unknown variable is qet, hich can be easily solved from Eq. (6).
After the qet s calculated, the other jet flow variables, such as yetand
pjet, can be obtained from Eq.
5 ) .
These parameters are also nondimen-
sionalized by corresponding reference values before being used in the solver
as the boundary conditions. Formulations for a supersonic jet and for
using total jet pressure as a driven parameter instead of C, can be found in
Ref. 16.
111. Results and Discussion
The CCW configuration and body-fitted grid studied in the present work
are shown in Figs. 1 and 2. The flap-setting angle may be varied both in
the experiments and the simulations. The studies presented here are all for
the 30deg flap setting to take advantages of CC high-lift benefits while
greatly reducing drag. In both the experiments5 and the present studies, the
freestream velocity was approximately 94.3 fps at a dynamic pressure of
10psf and an ambient pressure of 14.2psia. The freestream density is
0.00225 slugs/ft3. These conditions translate into a freestream M ach number
of 0.0836. The airfoil chord was 8 in. and the Reynolds number was
395,000.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 559/637
562
YI LIU ET AL.
Fig. 2 Body-fitted C-grid near the CC airfoil surface.
A. Validation Studies
Prior to its use in studying CC W configurations, the Navier-Stokes solver
was validated by modeling the viscous subsonic flow over a small-aspect-ratio
wing made of NACA
0012
airfoil
section^, ̂
and the results were in good
agreement with the experim ental measurem ents of Bragg and Spring.27 These
validation studies have been previously documented in Refs. 16 and 17, and
are not reproduced here.
Figure 3 shows the variation of lift coefficient with respect to
C
at a fixed
angle of attack
a= 0
deg) for the CCW configuration with a
30
de flap. Excel-
lent agreement with measured da ta from the experimen ts by Englar is evident. It
is seen that very high lift can be achieved by C C technology with a relatively low
C,.
A lift coefficient as high as 4.0 can be obtained at a
C
value of 0.33. And the
lift augm entation
ACl/AC,
is greater than
10
for this
30
deg flap configuration.
Figure
4
shows the computed
Cl
variation with the angle of attack, for a
number of
C
values, along with measured data. It is found that the lift coefficient
increases linearly with angle of attack, just as it does for conventional sharp
TE airfoils. However, the increase of lift with angle of attack breaks down at
high enough angles. This is a result of static stall, and is much like that experi-
enced with a conventional airfoil, but occurs at higher Cl,mawalues, thanks to
the beneficial effects of CC. The calculations also correctly reproduce the
decrease in the stall angle observed in the experiments at higher momentum
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 560/637
EVALUATION OF STEADY AND PULSED JET EFFECTS
563
4 -
-
CI, Measured
I, Com puted
I
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
CP
Fig. 3 Variation of the lift coefficient with the m omentum coefficients ata = 0 deg.
C,=O.1657
/-:
,=0.0740
EXP, C, 0.0
EXP, C, 0.074
EXP, C, = 0.15
-CFD
6
-4
-2 0 2 4 6 8 10 12 14 16
Angle
of Attack
Fig. 4 Variation of lift coefficient with angle of attack for different momentum
coefficients.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 561/637
564
YI LIU ET AL.
Fig. 5 Streamlines over the
CC
airfoil at two instantaneous time levels
C, = 0.1657,
angle
of
attack
= 6
deg).
coefficients. With the turbulence model used in this study, it is found that the
predicted stall angle is less than experimental measurements. However, the lift
prediction is in good agreement with experiments before stall. Unlike conven-
tional airfoils, where experience stall because of the progressive growth of TE
separation, CCW configuration stall is a result of LE separation. Figure 5
shows typical streamlines around the CC airfoil at an angle of attack of
6
deg,
and C = 0.1657 at a typical instance in time. In this case, a LE separation
bubble forms, which spreads over the entire upper surface, resulting in a loss
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 562/637
EVALUATION OF STEADY AND PULSED JET EFFECTS
565
of lift. However, the flow is still attached over the TE because of the strong
Coanda effect.
B. Leading Edge Blowing
Functioning like a slat, LE blowing is an effective way of alleviating LE stall
and achieving the desired performance at high angles of attack. To understand
the effects of LE blowing, a dual-slot CC airfoil was designed, and simulations
of both LE and TE blowing were carried out. Figure 6 shows lift coefficient
variations with angle of attack for three different combinations of LE and TE
blowing. In the first case, there is only a TE blowing with C , = 0.08, and it is
seen that the stall angle is very small, at approximately 5 deg. If a small
amount of LE blowing is used (C,
=
0.04), while keeping the TE blowing at
C , = 0.08 as before, the stall angle is greatly increased from
5
deg to
12
deg.
If even higher levels of LE blowing are used, for example, LE blowing with
C =
0.08 and TE blowing with C , = 0.04, the stall angle is increased to
more than 20 deg, but the total lift is decreased at the same angle of attack
compared to the previous case, even when the total momentum coefficients
(C,,LE
C,,TE)
of both cases are the same, equal to 0.12 here.
In conclusion, LE blowing is seen to increase the s tall angle, replacing the slat,
whereas the TE blowing is effective in producing high levels of lift. Leading-edge
blowing can also reduce the large nosedown pitch moment associated with high
lift and the suction pressure peak in the vicinity of the slot. In general, operating
at high angles of attack is not necessary for CC airfoils because high lift can be
readily achieved with low angles of attack and a moderate amount of blowing.
4
3.5
3
u
2.5
.-
i 2
- - -
LE Blowing,
C
=
0.04
2
LE Blowing,
Cy
.08
TE Blowing,
Cy
.04
5 1.5
Os5
0 2 4 6
8
10 12 14 16 18 20 22 24
Angle of Attack degrees)
Fig.
6
Lift coefficient
vs
angle of attack for the LE blowing case.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 563/637
566 YI LIU ET AL.
However, in situations where the CCW configuration must operate at high
angles of attack, a combination of LE and TE blowing may be necessary to
achieve the best performance.
C.
Effects of Freestream Velocity on Lift Production
As a followup to previous studies,17 num erical simulations have also been
carried out where the freestream velocities (and the Reynolds number) were
systematically varied. The purpose of theses studies was to determine and
isolate how freestream velocities and the Reynolds number affect the beneficial
effects of CC at a fixed momentum coefficient.
In this case, the jet momentum coefficient
C,
is fixed at 0.1657, and the jet slot
height is also fixed at
0.015
in. The freestream velocities vary from
0.5
to
1.8
times the experimental freestream velocity, equal to 94.3 fps, as stated earlier.
The jet velocity also varies with the freestream velocity to maintain a constant
C,. As shown in Figs. 7 and
8,
for a given momentum coefficient, the lift and
drag coefficients are not significantly affected by the variation of the freestream
velocity except at very low freestream velocities. At very low freestream
velocities, degradation of lift and the generation of high drag are seen. This is
because the jet velocity is too low to generate a sufficiently strong Coanda
effect to eliminate TE separation and vortex shedding. At sufficiently high free-
stream velocities, the performance of CC airfoils is independent of the freestream
velocity and the Reynolds number under the fixed C, and fixed jet slot height
conditions. Thus the mom entum coefficient is an appropriate driving parameter
for CC blowing if the jet slot height is fixed.
0
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
(Vm-cfd)
I
(Vm+xp)
Fig. 7 Lift coefficient vs freestream velocity
Cp
= 0.1657, h
=
0.015 in., and
Vm,exp
=
94.3
fps).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 564/637
EVALUATION OF STEADY AND PULSED JET EFFECTS
567
0.15
3
Q
U
0.1
E
8
0.2
-
0 1
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
(V--cfd)/ (V-exp)
Fig.
8
Drag coefficient vs freestream velocity C,
=
0.1657,
h
=
0.015
in., and
Vm,exp= 94.3
fps .
D. Effects of Jet Slot Height
According to recent acoustic measurement^,'^ ^ the jet slot height has a strong
effect on the noise produced by the CC airfoil. These studies indicate that a larger
jet slot will reduce the noise at the same momentum coefficient compared to a
smaller slot. To investigate the effect of jet slot heights on the aerodynamic
characteristics of
CCW
sections, simulations at several slot heights (varying
from 0.006 to 0.018 in.) have been carried out, at a fixed low C, C, = 0.04)
and a fixed high
C, (C,
= 0.1657) value, and at a constant free-stream velocity
From Fig. 9, it is seen that a higher lift coefficient can be achieved with
a smaller slot height even for the same momentum coefficient, and that the lift
coefficient is decreased by 20% as the slot height is increased from 0.006 in. to
0.018 in. A similar behavior is seen for the drag coefficient as shown in
Fig. 10. The
LID
characteristics of the airfoil, which are computed here as
Cl/ Cd
C,)
by adding
C,
to the drag coefficient in order to consider the rate
of change of momentum associated with the jet flow, do not vary much with
the change of the jet slot height. As shown in Fig. 11, when the slot height is
increased, the efficiency decreases approximately by 7.6% for the C, = 0.1657
case, and increases by about 5.3% for the
C, =
0.04 case. However, as shown
in Fig. 12, the jet mass flow rate increases by ~ 6 0 % hen the slot height is
increased from 0.006 in. to 0.018 in., because of the larger jet slot area.
As it is always preferable to obtain higher lift with as low a m ass flow rate as
possible, a thin jet is aerodynam ically more beneficial than a thick jet. How ever,
of 94.3 fps.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 565/637
568
YI LIU ET AL.
1
+ c p = o .o 4
I
- p
=
0.1657
3- -1
I
0.006 0.009 0.012 0.01 5 0.018
Jet Slot Height (inc h)
Fig. 9 Lift coefficient vs jet slot height
V ,
=
94.3 fps).
----
+ p = 0.04
-Cp =
0.1657
0.15
- .
0.006
0.009
0.012 0.015 0.018
Jet Slot Height (inc h)
Fig. 10 Drag coefficient vs jet slot height V ,
=
94.3 fps).
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 566/637
EVALUATION OF STEADY AND PULSED JET EFFECTS
569
0 .006 0.009 0.012 0.01 5 0.01
20
8
Y
5
+ c p = o .o 4
-Cu = 0.1657
+ p = 0.04
v = 0.1657
L
= 0.001
a
a
i?i
0.0005
Fig.
11
Variations
of
the
LID
characteristics with the jet slot height
V ,
=
94.3 fps .
4
0.006
0.009 0.012 0.015 0.018
Jet Slot Height ( inch )
Fig.
12
Mass
flow
rate requirements
of
the jet
vs.
jet slot height
V , = 94.3 fps .
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 567/637
570
YI LIU ET AL.
the large stagnation pressure losses associated with small orifices or slots
means that a higher stagnation pressure is required to generate a jet issuing
through a smaller slot than through a larger slot at the same momentum coeffi-
cient. The higher power consumption of compressors needed to produce the
required high stagnation pressures can negate the beneficial effects of
CC
for
very thin jets.
In sum mary, a smaller je t slot height is preferred from an aerodynamic design
perspective. However, as previously mentioned, a large r jet slot height is pre-
ferred from an aeroacoustic perspective. Thus, an optimum choice must be
made for the jet slot height from aerodynamic, acoustic, and compressor power
consumption considerations.
E. Pulsed Jet Effects
Du ring the past five years, there has been increased interest in the use of pulsed
jets, and “massless” synthetic jets for flow control and performance enhance-
ment. Wygnansky and colleagues28929 tudied the effects of eriodic excitation
on the control of separation and static stall. Lorb er et al e o and W ak e and
Lurie31 have studied the use of directed synthetic jet s fo r dynamic stall alleviation
of the rotorcraft blade. Hassan and Janakiram3’ have studied the use of synthetic
jets placed on the upper and lower surfaces of an airfoil as a way of achieving
desired changes in lift and drag, and offsetting vibratory airloads that otherwise
would occur durin g blade-vortex interactions. Pulsed jets and synthetic jets have
also been used to affect mixing enhancement, thrust vectoring, and bluff body
flow separation control. In 1972 , Oy ler and Palmer33 experimentally studied
the pulsed blowing of blown flap configurations. More recently, some numerical
simulations em ployin g a pulsed je t hav e also been reported fo r separation control
of high-lift systems,34 and traditional rounded T E
CC
airfoils with multiport
blowing.35 Most of the studies abo ve were focused on the use of low mom entum
coefficients or zero-mass blowings to control the boundary layer separation or
static and dyn am ic stall. Only a fe w studies33 considered the use of pulsed jets
for lift augmentation, at smaller mass flow rates co mp ared to steady jets.
In earlier work,17 it has been sho wn that the pulsed jet w ith square-wave
form is more efficient than the traditional sinusoidal form, and that the square-
wave-form pulsed jet can generate the same lift as the steady jet at a much
lower mass flow rate. In this work, we describe the studies done to isolate
the effects of freestream velocity, frequency, and chord length on pulsed jet
behavior.
Figures 13 and
14
show the variation of the time-averaged incremental
lift coefficient
ACl
over and above the baseline unblown configuration at
three frequencies, 40, 120, and 400 Hz. Figure
13
shows the variation with the
average m om entum coefficient; and Fig. 14 the variation with the average
mass flow rate.
At first glance, Figs.
13
and 14 app ear to show opposite trends. Figure 14
appe ars to favor high frequencies; that is, ACl increases as frequency increases,
and the pulsed jet produces a higher
ACl
than a steady jet. This appears to be
consistent with
experiment^.^^
However, Fig. 13 appe ars to show the opposite
trend-the steady je t appears to be always more efficient than a pulsed je t,
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 568/637
3
2.5
2
1.5
1
0.5
0
-
EVALUATION OF STEADY AND PULSED JET EFFECTS
571
-Steady Jet
-Pu lsed Je t , f = 40 Hz
- . Puls ed Jet, f = 120 Hz
~
0.02 0.04 0.06 0.08 0.1 0.12
I
Time-Averaged Mom entum Coef fic ient , CpO
Fig.
13
Incremental lift coefficient
vs
time-averaged mom entum coefficient.
Pulsed J et ,
f
= 40
Hz
Puls ed Jet, f = 120 Hz
Pulsed Jet,
f
= 400 Hz
14
0
0.0002 0.0004 0.0006 0.0008 0.001 0.0012 0.0014 0.0016
Time Averaged Mass Flow Rate (siu g kec )
Fig.
14
Incremental lift coefficient vs tim e-averaged mass
flow
rate.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 569/637
572
YI LIU ET AL.
u
c” 1.2-
.-
E
8
0.8-
E
u
and produces a large
ACl.
To resolve this “apparent” inconsistency between
Figs. 13 and 14, four points A , B, C,
D
are show n in Fig. 13. These points are
all at the same mass flow rate of 0.00088 slug/s. It is seen that point A is
above point
B.
That is, a steady jet is indeed able to produce a higher
ACl
than
a low-frequency
40
Hz jet. This is because the flow separates over a period of
time before a new cycle of blowing begins, destroying the lift generation.
However, ACl at points C and
D
(120 and 400 Hz jets) are higher than point
A. In these cases, bound circulation over the airfoil has not been fully shed
into the wake before a new cycle begins. The time-averaged lift at the same
specified averaged mass flow rate for a higher frequency pulsed jet is thus
higher com pare d to a steady jet. T his is consistent w ith Fig. 14.
It has also been found that high frequencies have the beneficial effect of
decreasing the time-averaged mass flow rate of the pulsed jet.” For exam ple,
as shown in Fig. 15, when the frequency is equ al to 400 Hz, the pulsed je t requires
only 73% of the steady jet mass flow rate while it can achieve
95%
of the lift
achieved with a steady blowing. Examination of the flowfield over an entire
cycle indicates that it takes some time after the jet has been turned off before
all the beneficial circulation attributable to the Coanda effect is completely
lost. If a new blowing cycle could begin before this occurs, the circulation will
almost instantaneously reestablish itself. At high enough frequencies, a s a conse-
quence, the pulsed jet will have all the benefits of the steady jet at considerably
lower mass flow rates.
+Pulsed Jet, Ave. C,=0.04
-Steady Jet, C,=0.04
1
0 4 ” ’ ’ ’ 8 8 8 8 8 8 8 8 8 8 8 8 8 8 1
0
20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400
Frequency
(Hz)
I I I I I I
I
I I I I
0
1.414 2.828
Strouhal Number f Chord /
Vinf)
Fig.
15
Tim e-averaged lift coefficient vs frequency and Strouhal number.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 570/637
EVALUATION OF STEADY AND PULSED JET EFFECTS
573
F. Strouhal Number Effects
For aerodynamic and acoustic studies, the frequency is usually expressed as a
non-dimensional quantity called the Strouhal number. Simulations have been
carried out to calculate the average lift generated by the pulsed jet at fixed
Strouhal numbers. The Strouhal number is defined as
f i e f
VC
St =
7)
In the present study, for the baseline case, Gef s 8 in., and
V ,
is equal to 94.3 fps.
Thus, for a 200 Hz pulsed jet, the Strouhal number is equal to 1.41.
From the preceding equation, besides the frequency, there are two other par-
ameters that could affect the Strouhal number: the freestream velocity V , and Gef
(chord of the CC airfoil). To isolate these effects, as shown in Tables
1
to 3 , three
cases have been studied. In the first case (Table 1), the freestream velocity and the
chord of the CC airfoil are fixed, and the Strouhal number varies with the fre-
quency. In the second case, as shown in Table 2, the Strouhal number is fixed
at 1.41 and the chord of the CC airfoil is also fixed. The frequency varies with
the freestream velocity to achieve the same Strouhal number. In the third case,
as shown in Tab le 3, the S trouhal number is fixed at 1.41 and the freestream vel-
ocity is also fixed, whereas the frequency varies along with the chord of the CC
airfoil. The Mach number and Reynolds number are also functions of the free-
stream velocity and the airfoil chord, and were changed appropriately. The
time-averaged momentum coefficient CF0 is fixed at 0.04 in these studies.
Figure 16 shows the lift coefficient variation with the frequency for these three
cases.
From Tab les 2 and
3,
it is seen that the com puted time-averaged lift coefficient
varies less than 2% when the Strouhal number is fixed, and the chord and/or
the freestream velocity is varied. Table 2 also indicates that the same
Cl
can be
obtained at a much lower frequency with a smaller freestream velocity as long
as the Strouhal number is fixed. Table 3 shows that for a larger configuration
with larger chord lengths, the same Cl can be obtained at a lower frequency pro-
vided the S trouhal number i s fixed. Tab le 1, on the other hand, show s that varying
the frequency and Strouhal number while holding the other variables fixed can
lead to a 12% variation in Cl. Thus, it is concluded the Strouhal number has a
Table 1 Computed time-averaged lift coefficient for the case where
U,
and
LreP
re
fixed, and the Strouhal number is varied with the frequency
Baseline Half frequency Double frequency
Frequency, z
200
100
400
Freestream velocity U,, fps
94.3
94.3 94.3
Chord of the Airfoil befn.
8 8 8
Strouhal number
1.41
0.705 2.82
Computed average lift 1.6804 1.5790 1.8026
coefficient CJ
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 571/637
574 YI LIU ET AL.
Table 2 Com puted time-averaged lift coefficient for t he case where S trou ha l
number and
L,f
ar e fixed, an d
U ,
an d the frequency a re varied
Baseline Half velocity Double velocity
Frequency, Hz
200 100 400
Freestream velocity U,, fps 94.3 47.15 118.6
Chord of the airfoil
Gef,
n.
8 8 8
Strouhal number 1.41 1.41 1.41
Computed average 1.6804 1.6601 1.7112
lift coefficient, Cl
Table
3
Com puted time-averaged lift coefficient for t he case where Str ouh al
number and U , fixed, and Lrefan d frequency a re varied
Baseline Double chord Half chord
Frequency, Hz
200 100 400
Freestream velocity U,, fps 94.3 94.3 94.3
Chord of the airfoil Lef,n. 8 16 4
Strouhal number 1.41 1.41 1.41
Computed average lift coefficient,
Cl
1.6804 1.7016 1.6743
. . A ..
1.24
50 100 150 200 250 300 350 400 450
Frequency
Fig. 16 Time-averaged lift coefficient vs frequency: Case 1: Strouhal numb er not
fixed, V , and
Lref
fixed; Case
2:
Strouhal number and
L,f
fixed, V , not fixed; and
Case 3: Strouhal number and V , fixed; Lrefnot fixed.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 572/637
EVALUATION OF STEADY AND PULSED JET EFFECTS
575
more dominant effect on the average lift coefficient of the pulsed jet than just
the frequency.
IV.
Conclusions
Th e Navier-Stokes simulations are used to model flow over the CCW con-
figurations because of the complexity of the flowfield and the strong viscous
effects. On comparison with experimental measurements, the results indicate
that this approach is an efficient and accurate way of modeling CCW flows
with steady and pulsed jets.
The CC technology is a useful way of achieving very high lift at even zero
angle of a ttack. It can also eliminate vortex shedding in the T E region, a potential
noise source. The lift coefficient of the CC airfoil is also increased with angle
of attack like the conven tional sharp T E airfoil. H owever, the stall angle of the
CC airfoil decreases rapidly with an increase in the blowing momentum
coefficient. This stall phenomenon occurs in the LE region, and may be sup-
pressed by LE blowing. In practice, because high
Cl
values are achievable at
low angles of attack, it may seldom be necessary to operate CC wings at high
angles of attack. However, because there is always a large nosedown pitch
moment for the CC airfoil, LE blowing may be necessary to reduce this
pitch moment at high
C
values, even at zero angle of attack.
At a fixed mom entum coefficient, the performance of the C C airfoil does not
vary significantly with the freestream velocity and the Reynolds number.
However, at a fixed
C
the lift coefficient is influenced by the jet slot height.
A thin jet from a smaller slot is preferred, because it requires much less mass
flow, and has the same efficiency in generating the required
Cl
values as a
thick jet. From a practical perspective, a much higher plenum pressure may be
needed to generate thin jets fo r a given
C
This may increase the power require-
men ts of compressors that provide the high-pressure air.
A square-wave-shape pulsed jet configuration gives larger increments in lift
over the baseline unblown configuration when compared to the steady jet at
the same time-averaged mass flow rate. Pulsed jet performance is improved at
higher frequencies because of the fact that the airfoil has not fully shed the
bound circulation into the wake before a new pulse cycle begins.
The Strouhal number has a more dominant effect on the performance of the
pulsed je t than just the frequency. Thus, the same performance of a pulsed je t
could be obtained at lower frequencies for a larger configuration or at smaller
freestream velocities provided the Strouhal number is kept the same.
Acknowledgment
This work was supported by NASA Langley Research Center under the
Breakthrough Innovative Technology Program, G rant-NAG 1-2146.
References
‘Goldin, D. S., NASA Headquarters, “National Aeronautics and Space Administration
Strategic
Plan,”
NPD
1000.1B,
Sept.
2000,
pp.
42-43.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 573/637
576
YI LIU ET AL.
’Crighton, D. G., “Aircraft Noise in Aeronautics of Flight Vehicles: Theory and
Practice,” Vol. 1: Noise Sources, NASA PR-1258, 1991, pp. 391-447.
3Sen, R., “A Study of Unsteady F ields Near Leading-edge Slats,” AIAA Paper 97-1696,
1997.
4Davy, R., and Remy, H., “Airframe Noise Characteristics on a 1/11 Scale Airbus
Model,” AIAA Paper 98-2335, June 1998.
’Englar, R. J., Smith, M. J., Kelley,
S.
M., and Rover, R. C. III., “Application of Cir-
culation Control to Advanced Subsonic Transport Aircraft, Part I: Airfoil Development,”
Journal OfAircraft,
Vol. 31 No. 5, 1994, pp.1160-1168.
6Englar, R. J., Smith, M. J., Kelley, S. M., and Rover, R. C. III., “Application of
Circulation Control to Advanced Subsonic Transport Aircraft, Part
11:
Transport
Application,”
Journal of
Aircraft, Vol. 31, No. 5, 1994, pp. 1169-1177.
’Shrewsbury, G. D., and Sankar, L. N., “Dynamic Stall of an Oscillating Circulation
Control Airfoil,” Proceedings of International Symposium on Nonsteady Fluid Dynamics,
June 1990, pp. 15-22.
‘Shrewsbury, G. D., and Sankar, L. N., “Dynamic Stall of Circulation Contro l Airfoils,”
AIAA Paper 90-0573, Jan. 1990.
’Salikuddin, M., Brown, W. H., and Ahuja, K. K., “Noise From a Circulation Control
Wing with Upper Surface Blowing,” Journal of Aircraft, Vol. 24, 1987, pp. 55-64.
“McLean, J. D., Crouch, J. D., Stoner, R. C., Sakurai,
S.,
Seidel, G. E., Feifel, W. M.,
and Rush, H. M., “Study of the Application of Separation Control by Unsteady Excitation
to Civil Transport,” NASA /CR-1999-209338, June 1999.
“Eng lar, R. J., and Huson, G. G., “Developm ent of Advanced Circulation Contro l Wing
High Lift Airfoils,” AIAA Applied Aerodynamics Conference, AIAA Paper 83-1847, July
1983.
”Munro, S., Ahuja, K., and Englar, R., “Noise Reduction Through Circulation Control
Technology,” AIAA Paper 2001-0666, Jan. 2001.
‘3Munro, S., and Ahuja, K. K., “Aeroacoustics of a High Aspect-Ratio Jet,” AIAA Paper
2003-3323, May 2003; presented at the 9th AIAA /CEA S Aeroacoustics Conference and
Exhibit, May 2003.
‘‘Munro,
S.,
and Ahuja, K. K., “Fluid Dynamics of a High Aspect-Ratio Jet,” 9th AIAA/
CEAS Aeroacoustics Conference and Exhibit, AlA A Paper 2003-3 129, May 2003.
15Munro, S., and Ahuja, K. K., “Development of a Prediction Scheme for Noise of
High-Aspect Ratio Jets,” 9th AIAA/CEAS Aeroacoustics Conference and Exhibit,
AlAA Paper 2003-3255, May 2003.
16Liu, Y., “Numerical Simulations of the Aerodynamic Characteristics of Circulation
Control Wing Sections,” Ph.D Dissertation, School of Aerospace Engineering, Georgia
Inst. of Technology, Atlanta, G A, 2003.
”Liu,
Y.,
Sankar, L. N., Englar, R. J., and Ahuja, K. K ., “Num erical Sim ulations of the
Steady and Unsteady Aerodynamic Characteristics of a Circulation Control Wing Airfoil,”
AIAA Paper 2001-0704, Jan. 2001.
18Douglas, J., “On the Numerical Integration of ut
=
u,, uyy by Implicit Methods,”
Journal of Society of Industrial and Applied Mathematics, Vol. 3, No. 1, 1955.
”Briley, W ., and McDonald, H., “Solution of Multi-Com ponent Navier-Stokes
Equations by Generalized Implicit Method,”
Journal
of
Computational Physics,
Vol. 24, 1977, p. 372.
”Kwon, J., and Sankar, L. N., “Num erical Study of the Effects of Icing on Finite Wing
Aerodynamics,” AIAA Paper 90-0757, Jan. 1990.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 574/637
EVALUATION OF STEADY AND PULSED JET EFFECTS
577
’lBangalore, A., Phaengsook, N., and Sankar, L. N., “Application of a Third Order
Upwind Scheme to Viscous Flow over Clean and Iced Wings,” AIAA Paper 94-0485,
Jan. 1994.
”Baldwin, B.
S.,
and Lomax, H., “Thin Layer Approximation and Algebraic Model
for
Separated Turbulent Flows,” AIAA Paper 78-257, Jan. 1978.
23Spalart, P. R ., and A llmaras,
S.
R., “A One-Equation Turbulence Model for
Aerodynamic Flows,” AIAA Paper 92-0439, Jan. 1992.
24Shrewsbury,G. D., “Numerical Evaluation of Circulation Control Airfoil Performance
Using Navier-Stokes Methods,” AIAA Paper 86-0286, Jan. 1986.
25Shrewsbury, G. D., “N umerical Study of a Research Circulation Control Airfoil Using
Navier-Stokes Methods,”
Journal ofAircraft,
Vol. 26, No. 1, 1989, pp. 29-34.
26Williams,
S.
L., and Franke, M. E., “Navier-Stokes Methods to Predict Circulation
Control Airfoil Performance,”
Journal
of
Aircraft,
Vol. 29, No. 2, 1992, pp. 243-249.
27Bragg, M. B., and Spring,
S.
A., “An Experimental Study of the Flow Field about an
Airfoil with Glaze Ice,” AIAA 25th Aerospace Science Meeting, AIAA Paper 87-0100,
Jan. 1987.
28Seifert, A., Darabi, A., and Wygnanski, I., “Delay of Airfoil S tall by Periodic
Excitation,” Journal of Aircraft, Vol. 33, No. 4, 1996.
29Wygnanski, I., “Some New Observations Affecting the Control of Separation by
Periodic Excitation,” Fluids 2000 Conference and Exhibit, AIAA Paper 2000-23 14,
June 2000.
30Lorber,P. F., McCormick, D., Anderson, T., Wake , B. E., MacMartin, D., Pollack, M.,
Corke, T., and Breuer, K., “Rotorcraft Retreating Blade-Stall Control,” Fluids 2000
Conference and Exhibit, AIAA Paper 2000-2475, June 2000.
31Wake, B., and Lurie, E. A., “Computational Evaluation of Directed Synthetic Jets
for
Dynamic Stall Control,” 57th American Helicopter Society Annual Forum, Washington
DC, 9-11 May 2001.
32Hassan, A., and Janakiram, R. D., “Effects of Zero-Mass Synthetic Jets on the
Aerodynamics of the NACA 0012 Airfoil,”
Journal
of
the American Helicopter Society,
Vol. 43, No. 4, 1998.
330y le r, T. E., and Palmer, W. E., “Exploratory Investigation of Pulse Blowing
for
Boundary Layer Control,” North American Rockwell Rept. NR72H-12, Jan. 1972.
34Schatz, M., and Thiele, F., “Numerical Study of High-Lift F low w ith Separation
Control by Periodic Excitation,” AIAA Paper 2001-0296, Jan. 2001.
35Sun, M., and Hamdani, H., “Separation Control by Alternating Tangential Blowing/
Suction at Multiple Slots,” AIAA Paper 2001-0297, Jan. 2001.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 575/637
Chapter
23
Time-Accurate Simulations of Synthetic Jet-Based
Flow Control for a Spinning Projectile
Jubaraj
Sahu*
US. rmy Research Laboratory, Aberdeen Proving Ground, Maryland
Nomenclature
D = drag force, N
d =
reference diameter, m
f =
et frequency, Hz
Fy = aerodynamics force in y-direction (lift force)
F, = aerodynamics force in z-direction (side force)
F
=
inviscid flux vector
G = viscous flux vector
H
= vector of source terms
I = impulse, N-s
L = lift force, N
M
=
Mach number
p =
pressure, N/m2
p s = projectile spin rate, Hz
t = time, ms
V ,
=
freestream velocity, m/s
vj = je t velocity, m /s
W = vector of conservative variables
y+ = normal viscous sublayer spacing
x y,
z
= axial, normal (vertical), and horizontal axes
a
=
angle of attack, deg
*Aerospace Engineer. Associate Fellow
AIAA.
This material is declared a work
of
the
U.S.
overnment and is not subject to copyright protection
in the United States.
579
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 576/637
580 J. SAHU
I. Introduction
determination of aerodynamics is critical to the low-cost
A
evelopment of new advanced munitions.
,*
Competent smart munitions
that can more accurately hit a target can greatly increase lethality and enhance
survivability. Desert Storm convincingly demonstrated the value of large-scale
precision-guided munitions. A similar capability for small-scale munitions
would increase the effectiveness of infantry units, reduce collateral damage,
and reduce the weight of munitions that must be carried by individual soldiers.
The Army is, therefore, seeking a new generation of autonomous, course-correct-
ing, gun-launched projectiles for infantry soldiers.
Because of the sm all projectile diameter d = 0.02 to 0.04m), maneuvers by
canards and fins seem very unlikely. An alternative and new evolving technology
is microadaptive flow control through synthetic jets. These very tiny (of the order
of 0.3mm) synthetic microjet actuators have been shown successfully to modify
subsonic flow characteristics and pressure distributions for simple airfoils and
cylinders.394 he synthetic jets (fluid being pumped in and out of the jet cavity
at a high frequency of the order 1000Hz) are control devices (Fig. 1 with
zero net mass flux and are intended to produce the desired con trol of the flowfield
through mom entum effects. M any parameters such as jet location, jet velocity,
and je t actuator frequency, can affect the flow control phenom enon. Until now,
the physics of this phenomenon has not been well understood. In addition,
advanced numerical predictive capabilities or high-fidelity computational fluid
dynamics (CFD) design tools either did not exist or have not been successfully
applied to p ractical real-world problems involving microadaptive flow control.
The present research effort described here is focused on advancing aerodynamic
numerical capability to predict accurately and provide a crucial understanding of
the complex flow physics associated with the unsteady aerodynamics of this new
class of tiny synthetic microjets for control of modem projectile configurations.
High-performance CF D techniques are developed and applied for the design and
analysis of these microadaptive flow control systems for steering a spinning pro-
jectile for infantry operations.
CCURATE
Pulsat ing
Synth et ic Jet
Diaphragm
Fig.
1
Schematic
of
a synthetic jet.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 577/637
FLOW CONTROL ON SPINNING PROJECTILE
581
The control of the trajectory of a 40 m m spinning projectile is achieved
by altering the pressure distribution on the projectile through forced asymmetric
flow separation. Unsteady or time-accurate C FD modeling capabilities are devel-
oped and used to assist in the design of the projectile shape, the placement of the
synthetic actuators, and the prediction of the aerodynam ic force and mom ents for
these actuator configurations. Additionally, the advanced CFD capabilities
provide a simpler way to explore various firing sequences of the actuator
elements. Time-accurate unsteady CFD computations have been performed to
predict and characte rize the unsteady nature of the syn thetic je t interaction flow-
field produced on the M203 grenade launched projectile for various yaw and spin
rates for fully viscous turbulent flow conditions.
Turbulence is usually modeled using a traditional Reynolds-averaged Navier-
Stokes (RANS) approach. RAN S models are easy to use and provide very good
results for many steady flows, especially at supersonic speeds. Although this
approach provides some detailed flow physics, it is not well suited and can be
less accurate for the new class of unsteady flows associated with synthetic jets
at subsonic speeds. In order to improve the accuracy of the numerical simulation,
the predictive capability has been extended to include a higher order hybrid
RANS/LES (large eddy simulation) approach.596This new approach computes
the large eddies present in the turbulent flow structure (in the vicinity of the
microjet) and allows the simulation to capture, with high fidelity, additional
flow structures associated with the synthetic jet interactions (in the projectile
wake or base flow in the present study) in a time-dependent fashion. Modeling
of azimuthally placed synthetic microjets requires adequate grid resolution,
highly specialized boundary conditions for jet activation, and the use of an
advanced hybrid LES approach permitting local resolution of the unsteady turbu-
lent flow with high fidelity. The addition of yaw (angle of attack) and spin while
the projectile is subjected to the pulsating microjets rendered predicting forces
and moments a m ajor challenge.
Both RANS and hybrid RA NS/LES models have been used in the present study.
Although the
RANS
method works well for steady flows, the accuracy of this
method for unsteady flows may be less than desired. Because the large-energy-
containing eddies are computed using the LES method, this technique is expected
to be more capable of handling unsteady shear layers and wakes, and so on.
The advanced CFD capability used here solves the full three-dimensional
Navier- Stokes equations and incorporates unsteady boundary conditions for
simulation of the synthetic jets. The present study investigates the ability of
these advanced techniques with time-accurate computations of unsteady syn-
thetic jets for both nonspinning and spinning projectile cases at low subsonic
speeds. The following sections describe the numerical procedure, the unsteady
jet boundary condition, the hybrid RANS/LES turbulence model, and the com-
puted results obtained.
11 Computational M ethodology
The complete set of three-dimensional time-dependent Navier- Stokes
equations7 is solved in a time-accurate manner for simulations of the unsteady
synthetic jet interaction flowfield on the M203 grenade launched projectile
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 578/637
582 J. SAHU
with spin. The three-dimensional time-dependent RANS equations are solved
using the finite-volume method’:
where W is the vector of conservative variables,
F
and G are the inviscid and
viscous flux vectors, respectively, H is the vector of source terms, V is the cell
volume, and A is the surface area of the cell face.
Second-order discretization was used for the flow variables and the turbulent
viscosity equations. Two-equation9 and higher-order hybrid RANS/LES6 turbu-
lence models were used for the computation of turbulent flows. The hybrid
RA NS/L ES approach based on limited numerical scales (LNS)6 is well suited
to the simulation of unsteady flows and contains no additional empirical con-
stants beyond those appearing in the original RANS and LES subgrid models.
With this method, a regular RANS-type grid is used except in isolated flow
regions where denser, LES-type mesh is used to resolve critical unsteady flow
features. The hybrid model transitions smoothly between an LES calculation
and a cubic
k--E
model, depending on grid fineness. A somewhat finer grid
was placed around the body, and near the je t, the rest of the flowfield being occu-
pied by a coarser, RANS-like mesh. Dual time-stepping was used to achieve the
desired time accuracy. In addition, special jet boundary conditions were devel-
oped and used for numerical modeling of synthetic jets. The grid was actually
moved to take into account the spinning motion of the projectile.
A.
Unsteady Jet Boundary Conditions
One particular boundary condition (BC ) used in the present simulations of the
unsteady jet s is an “oscillating jet” BC . In its basic fo rm, it is a steady inflow/
outflow BC, inwhich the user supplies the velocity normal to the boundary
along with static temperature and any turbulence quantities. When the velocity
provided is negative, it is considered to be an inflow, and when it is positive, it
is treated as an outflow. In the case of inflow, the static temperature and turbulence
quantities are utilized along with the inflow velocity. In the case of outflow, only
the velocity is utilized. At inflow, the tangential component of velocity is set to
zero, and at outflow, the tangential component is extrapolated from the interior.
At outflow, all primitive variables except normal velocity are extrapolated from
the interior. At inflow, the static pressure is taken from the interior.
This BC also has a set of modifiers. The first modifier available for this BC
allows the velocity to oscillate. The base velocity is multiplied by an amplitude
that varies as sin(2@), whe ref is the frequency of the oscillation. Thus, the oscil-
lating velocity can cycle from being positive to being negative and back within
each period (or from being negative to positive and back, based on the sign of
the input for the basic BC formulation). A second modifier permits the steady
or oscillating inflow/outflow to be on over certain time intervals and off
during other intervals. During “on” periods, the basic or the basic multiplied
by the oscillating amplitude multiplier (first modifier), is used. The user provides
the ranges of time during which the je t is on. The user also provides a repetition
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 579/637
FLOW CONTROL ON SPINNING PROJECTILE
583
time period (e.g., the time period corresponding to one spin rotation of the pro-
jectile). Within each time period, therefore, there are sets of start and end times
that define when the jet is on. During “off”periods, the am plitude is set to zero. In
parts of the cycle when the jet is off, the boundary condition thus reverts to the
condition of inviscid surface tangency. This allows slip past the boundary, as
would exist (in the form of a shear layer) if the jet was emanating from a
cavity /hole.
B. Hybrid
RANS/LES
Turbulence Model
Currently, the two most popular forms of turbulence closure, namely ensem -
ble-averaged models (typically based on the
RANS
equations), and LES with a
subgrid-scale model, both face a number of unresolved difficulties. Specifically,
existing LES models have met with problems related to the accurate resolution of
the near-wall turbulent stresses. In the near-wall region, the foundations of large-
eddy simulation are less secure, because the sizes of the (anisotropic) near-wall
eddies approach than of the Kolmogorov scale, requiring a mesh resolution
approaching that of a direct numerical simulation. On the other hand, existing
ensemble-averaged turbulence models are limited by their empirical calibration.
Their representation of small-scale flow physics cannot be improved by refining
the mesh, and over short time scales they tend to be overly dissipative with
respect to perturbations around the mean, often suppressing unsteady motion
altoget her.
Although LES is an increasingly powerful tool for unsteady turbulent flow
prediction, it is still prohibitively expensive. To bring LES closer to becoming
a desi n tool, a hybrid RANS/LES approach based on limited numerical
scales has been recently developed by Metacom p Technologies.’ T his approach
combines the best features of RANS and LES in a single modeling framework.
The hybrid RANS/LES model is formulated from an algebraic or differential
Reynolds-stress model, in which the subgrid stresses are limited by the numeri-
cally computed local length-scale and velocity-scale products. It thus behaves
like its parent RANS model on RANS-type grids, but reverts to an anisotropic
LES subgrid model as the mesh is refined locally, thereby reaching the correct
(DNS) fine-grid limit. Locally embedded regions of LES may be achieved auto-
matically through local grid refinement, whereas the superior near-wall stress
predictions of the RANS model are preserved, removing the need for ad hoc,
topography-parameter-based wall damping.
The hybrid RANS/LES formulation is well suited to the simulation of
unsteady flows, including mixing flows, and contains no additional empirical
constants beyond those appearing in the original RANS and LES subgrid
models. With this method a regular RANS-type grid is used except in isolated
flow regions where denser, LES-type mesh is used to resolve critical unsteady
flow features. The hybrid RANS/LES model transitions smoothly between an
LES calculation and a cubic
k--E
model, depending on grid fineness. A somewhat
finer grid was placed around the body, and near the jet, the rest of the flowfield
being occupied by a coarser, RANS-like mesh.
To date, the hybrid RANS/LES technique has been used successfully on a
number of unsteady flows. Examples include flows over cavities, flows around
k
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 580/637
584 J. SAHU
blunt bodies, flows around airfoils and wings at high angle of attack, separation
suppression using synthetic jets, forced and natural convection flows in a room,
and mixing flows in nozzles.
111. Projectile Geometry and Computational Grid
The projectile used in this study is a 1 %caliber ogive-cylinder configuration
(see Fig. 2). Here, the primary interest is in the development and application of
CFD techniques fo r accurate simulation of projectile flowfield in the presence of
unsteady jets. The first step here was to obtain a converged solution for the pro-
jectile without the jet . The converged jet-off solution was then used as the starting
condition fo r the computation of time-accurate unsteady flowfield for the projec-
tile with synthetic jet s. The je t locations on the projectile are show n in Fig.
3.
The
je t cond itions were specified at the exit of the je t for the unsteady (sinusoidal vari-
ation in jet velocity) jets. The jet conditions specified include the jet pressure,
density, and velocity components. Numerical computations have been made
for these jet cases at subsonic Mach numbers,
M
= 0.11 and 0.24, and at
angles of attack = 0 to 4 deg. The jet width was 0.32 mm, the jet slot half-
angle was 18 deg, and the absolute peak jet velocities used were 3 1 and 6 9 m/s
operating at a frequency f= 1000 Hz.
A
computational grid expanded near the vicinity of the projectile is shown in
Fig. 4. Grid points are clustered near the je t as well as the boundary layer regions
to capture the high gradient flow regions. The computational grid is a single
block; it has 211 points in the streamwise direction, 241 in the circumferential
direction, and 80 in the normal direction. The grid is closeted near the body
surface with grid spacing that corresponds to a
y+
value of approximately 1.0.
The same grid was used for both RANS and hybrid RANS/LES calculations.
The unsteady simulation took thousands of hours of CPU time on Silicon
Graphics Origin and IBM SP3 computers running with 16-24 processors.
More details of the CPU time usage and requirement are iven in Section IV.
The parallel processing capability in C FD ++ code' was designed in
the beginning to be able to run on a wide variety of hardware platforms and
Fig. 2 Projectile geometry.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 581/637
FLOW CONTROL ON SPINNING PROJECTILE
585
Jet
Fig. 3 Aft-end geometry showing the jet location.
communications libraries, including MPI and PVM. MPI was used on various
platforms for communications between different processors. The code runs on
parallel processors and one can switch the use of an arbitrary number of CPUs
at any time. Depending on the number of CPUs being employed, the mesh is
domain-decomposed using the METIS tool developed at the University of
Minnesota.
Fig. 4 Computational grid near the projectile.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 582/637
586 J. SAHU
IV. Results
Time-accurate unsteady numerical computations using advanced viscous
Navier-Stokes methods were performed to predict the flowfield and aerody-
namic coefficients on both a nonspinning and a spinning projectile. Limited
experimental data (from Ref. 10and private communication with J. McMichael,
GTRI) exist only fo r the nonspinning case and were used to validate the unsteady
CFD results. Three-dimensional numerical computations have been performed
for the projectile configuration with jet-interaction using CF D++ code at sub-
sonic Mach numbers,
M =
0.11 and 0.24, and at angles of attack
= 0
4
deg. The preconditioned version of the CFD++ cod e was used to obtain an effi-
cient numerical solution at low speeds. For modeling of the unsteady synthetic
jets, both unsteady RANS and a hybrid RA NS /LE S approach6 were used. For
computations of these unsteady jets, full three-dimensional computations are
performed and no symmetry was used.
A. Nonspinning Projectile
Three-dimensional unsteady CFD results were obtained at a subsonic Mach
number of 0.11 V , = 7 m/s) and several angles of attack from 0 to 4 deg
using both the unsteady
RANS
and the hybrid R AN S/LE S approaches. The syn-
thetic jets are on all the time for these nonspinning cases. These three-dimen-
sional unsteady CFD computations are carried out to provide fundamental
understanding of fluid dynamics mechanisms associated with the interaction of
the unsteady synthetic jets and the projectile flowfields at subsonic speeds.
Many flowfield solutions resulting from the simulation of multiple spin cycles
and, hence , a large number of synthetic je t operations, were saved at regular inter-
mittent time intervals to produce movies to gain insight into the physical
phenomenon resulting from the synthetic jet interactions. The unsteady jets
were discovered to break up the shear layer coming over the step in front of
the base of the projectile. It is this insight that was found to substantially alter
the flowfield (making it unsteady) both near the jet and in the wake region that
in turn produced the required forces and moments even at 0-deg angle of
attack (level flight). Time-accurate velocity magnitude (Fig. 5 and velocity
vectors (Fig. 6 ) confirm the unsteady wake flowfields arising from the interaction
of the synthetic je t with the incoming freestream flow at Mach = 0.11. Figure 7
shows the particles emanating from the jet and interacting with the wake flow,
making it highly unsteady. More important, the breakup of the shear layer is
clearly evidenced by the particles clustered in regions of flow gradients or vorti-
city (evident in computed pressure contours, Fig.
8).
Verification of this con-
clusion is provided by the excellent agreement (Fig.
9)
between the predicted
(solid line) and measured (solid symbols) values of the net lift force due to
the jet. In this case, the solid line represents the results obtained with the
hybrid RANS/LES turbulence model. Also shown in Fig. 9 is a time-averaged
result of the lift force obtained using a RANS turbulence model at 0-deg angle
of attack. It is quite clear that the lift force is underpredicted by the RANS
model and does not compare as well with the experimental data. This indicates
the inability of the
RANS
model to predict accurately the unsteady wake flow-
fields resulting from the synthetic jet flow control.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 583/637
FLOW CONTROL ON SPINNING PROJECTILE
587
Fig. 5 Velocity magn itudes,M = 0.11, Y = 0 deg.
The net lift force F,) was determined by time-averaging the actual time his-
tories of the highly unsteady lift force (an example show n in Fig. 10 for various
ang les of attack) resulting from the jet interaction at zero-degree ang le of attack
and computed with the new hybrid
RANS/LES
turbulence approach. Figure 10
Fig. 6 Velocity vectors, M = 0.11, Y =
0
deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 584/637
5aa
J. SAHU
Fig.
7
Particle traces,
M
=
0.11
x =
0
deg.
shows both low- and high-frequency oscillations in the predicted lift force at
different angles of attack, = 0, 2, and 4 deg. The high-frequency oscillations
(of the order of 1 ms) are a direct result of the jet actuation that corresponds to
the jet frequency of
1000
Hz. The low frequency oscillations observed in the
Fig.
8
Computed pressures, M = 0 . 1 1 , ~ ~
0
deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 585/637
FLOW CONTROL ON SPINNING PROJECTILE
u
c
0 6
Q)
Q)
t
2
0.4
it:
.-
L
Q
5
0
589
-
w EXPERIMENT
+CFD
Hybrid
RANSILES)
CFD RANS)
-
w
o.2A-
w w
o.8
time-histories result from the interaction
of
the jet with wake and the resulting
unsteady wake flowfields.
B.
Spinning Projectile
Of m ore interest is the spinning projectile case for the real-world applications.
Num erical computations have been made in this case for actual flight condition at
Time ms)
Fig.
10
Time-histories of computed lift force at angles of attack cu = 0 2 and 4 deg,
hybrid RANS/LES model, M
=
0.11.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 586/637
590 J. SAHU
Y
t
Jet-on1
a
=3.73ms
t =O .............
ypfa
Fig. 11 Schematic of jet actuation for one spin cycle view from the nose).
a Mach number, of M = 0.24, an angle of attack , of = 0 deg, and a spin rate of
67 Hz. The atm ospheric flight conditions are used here. The jet width was
0.32 mm, the jet slot half-angle was 18 deg, and the absolute peak jet velocities
used were 31 and 69 m /s operating at a frequency of 1000 Hz. In this case, the
projectile (40 mm grenade) spins clockwise at a rate of 67 Hz looking from the
front (Fig. 11). Unlike the nonspinning cases where the jet was on all the time,
here the jet actuation corresponds to one-fourth of the spin cycle from -45 to
+45 deg with 0 deg being the positive y-axis. The jet is off during the remaining
three-fourths of the spin cycle.
The unsteady CFD modeling required about 600 time steps to resolve a full
spin cycle. For the part of the spin cycle when the jet is on, the 1000 Hz et oper-
ated for approximately for four cycles. Time-accurate CFD modeling of each jet
cycle required over 40 time steps. The actual computing time for one full spin
cycle of the projectile was about 50 hours using 16 processors (i.e., 800 pro-
cessor-hours) on an IBM SP3 system for a mesh size of about four million
grid points. Multiple spin cycles and, hence, a large number of synthetic jet oper-
ations were required to reach the desired periodic time-accurate unsteady result.
Some cases were run for as many as 60 spin cycles, requiring over 48,000
processor hours of computer time.
Com puted particle traces emanating from the jet into the wake are shown in
Fig. 12 at four different instants in time for
M
= 0.24 and a = 0 deg.
As
stated
earlier, the 1000 Hz synthetic jet operates for about four jet cycles during one
spin cycle of the rotating projectile. The four different instants of time selected
in Fig. 12 correspond to each of the four jet cycles as the projec tile rotates coun-
terclockwise (looking from the back of the projectile). The particle traces ema-
nating from the jet interact with the wake flow making it highly unsteady. It
also shows the flow in the base region to be asymmetric because of the interaction
of the unsteady jet .
The computed surface pressures from the unsteady flowfields were integrated
to obtain the aerodynamic forces and moments from both unsteady RAN S as
well as the hybrid RA NS /LES solutions. The jet-off unsteady RANS calculations
were first obtained and the jets were activated beginning at time,
t
= 28 ms.
Computed normal or lift force F,) and side force F,) were obtained for two
different jet velocities, j
=
31 and 69 m/s, and are shown in Fig. 13 for the
bigger jet as a function of time. These computed results clearly indicate the
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 587/637
FLOW CONTROL ON SPINNING PROJECTILE 591
Fig. 12 Instantaneous computed particle traces at different times jet-on, M
=
0.24,
= 0 deg.
Time ms)
Fig. 13 Computed lift and side forces, unsteady RANS,
M =
0.24, vj = 69 m /s,
= 0
deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 588/637
592 J. SAHU
unsteady nature of the flowfield. When the je t is on, one can observe a sharp rise
in both the lift and the side forces. The peak levels in the forces remain high until
the je t is turned off. When the jet is turned off, the levels of these forces drop to
the same levels (low-amplitude oscillations) prior to the jet activation corre-
sponding to the jet-off wake flow. The unsteady RANS results clearly show
when the jet is on and when it is off during the spin cycle.
Figure
14
shows the comparison of the predicted lift force using the unsteady
RANS and the hybrid RANS/LES turbulence models for the bigger jet case at
zero-degree angle of attack. As indicated earlier, the unsteady RANS results of
the lift and the side forces clearly show when the jet is on and when it is off
during the spin cycle. The effect due to the jet for the hybrid RANS/LES case
is not as easily seen. It is hidden in these oscillations. However, the mean
value of the lift force seems to be close to zero when the jet is off during the
spin cycle. In general, the levels of the lift force oscillations predicted by the
hybrid RANS/LES model are larger than those predicted by the unsteady
RANS
model. This result can be attributed to the fact that the wake is unsteady
and the hybrid RANS/LES model produces large levels of oscillations for the
unsteady wake flowfield whether the jet is off or on.
As described earlier, the comparisons for the nonspinning cases showed that
the level of lift force predicted by the hybrid RANS/LES closely matched the
data. Here, the addition of spin as well as the jet actuation for part of the spin
cycle further complicates the analysis of the CFD results when the hybrid
RANS/LES model is used. The level of oscillations seen is quite large and the
effect of the jet cannot be easily seen in the instantaneous time histories of the
unsteady forces and moments. In addition, the unsteady wake flowfield is
expected to change from one spin cycle to another. To get the net effect of the
jet, unsteady computations were run for many spin cycles of the projectile with
Time rns)
Fig. 14 Computed lift forces, unsteady RANS and hybrid RANS/LES, M = 0.24
Vj = 69 m/s, 01 =
0
deg.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 589/637
FLOW CONTROL ON SPINNING PROJECTILE
593
0.4
0.3
0.2
t
I
0.1
0
0 5 10
15
Time ms)
Fig. 15 Computed time-averaged lift force over many spin cycles, hybrid
RANS/
LES, Vj = 69 m/s,
M
= 0.24 =
0
deg,
P
= 67 Hz.
the syn thetic jets. The C FD results are plotted over only one spin cycle; each sub-
sequent spin cycle was superimposed and a time-averaged result was then
obtained over one spin cycle. In all these cases, the jet is on for one-fourth of
the spin cycle (time, t = 0-3.73 ms) and is off for the remainder (three-
fourths) of the spin cycle. Figures
15
through 16 show the time-averaged
results over a full spin cycle that corresponds to 15 ms (67 Hz pproximately.
Figure 15 shows the computed lift force, again averaged over many spin
0 4
0.3
j
0.2
l
t
0.1
0
-0.1
0 5 10 15
Time
ms)
Fig. 16 Computed time-averaged lift force over many spin cycles for different jet
velocities, hybrid
RANS/LES,
M = 0.24, =
0
deg,
P
= 67
Hz.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 590/637
594 J. SAHU
0 10
20
30 40
50 60
Number
of Spin Cycles
Fig.
17
Impulse from the lift force vs spin cycles for two jet velocities, hybrid
RANS/LES, M
=
0.24
=
0 deg, P = 67
Hz.
cycles (10 ,2 0, 30, and
40
for the peak jet velocity of 69 m/s . The jet effect can
clearly be seen when the jet is on t
=
0-3.73 ms) even after
10
spin cycles. The
net lift is about 0.17 N because of the jet actuation and seems to have converged
after 20 spin cycles. For the remainder of the spin cycle, the jet is off however,
the effect of the jet on the wake still persists and this figure shows that lift force
(mean value 0.07
N)
is still available. The fact that one can obtain a lift force for
this jet-off portion of the spin cycle is a new result solely caused by the spin effect
of the projectile. Figure 16 show s the com puted time-averaged lift force after 50
and 60 spin cycles for jet velocities 3 1 and 69 m/s, respectively. It clearly show s
that the larger jet produces larger lift force than the smaller jet when the jet is
activated. The lift force can be integrated over time to obtain the impulse I
Figure 17 shows the impulse obtained from the lift force as a function of the
spin cycles for both jets. As seen here, in both cases it takes about 30 to
40
spin cycles before the impulse asymptotes to a fixed value.
The computed lift force along with other aerodynamic forces and moments,
directly resulting from the pulsating jet, were then used in a trajectory analysis
(from private communication with M. Costello, Oregon State University) and
the synthetic microjet was found to produce a substantial change in the cross
range. These results indicate the viability of the use of synthetic microjets to
provide the desired course correction for the projectile to hit its target.
V.
Conclusions
This chapter describes a computational study undertaken to determine the aero-
dynam ic effect of tiny synthetic jets as a means to provide the control authority
needed to maneuver a projectile at low subsonic speeds. Com puted results have
been obtained for a subsonic projectile for both nonspinning and spinning cases
using a time-accurate Navier- Stokes computational technique and advanced
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 591/637
FLOW CONTROL ON SPINNING PROJECTILE
595
turbulence models. The unsteady jet in the case of the subsonic projectile is
shown substantially to alter the flowfield both near the jet and the base region
which in turn affects the forces and moments even at 0-deg angle of attack.
The predicted changes in lift force due to the jet match well with the experimental
data for various angles of attack from 0 to 4 deg in the hybrid RANS/LES
computations. For the spinning projectile cases, the net time-averaged results
obtained over the time period corresponding to one spin cycle clearly showed
the effect of the synthetic jets on the lift as well as the side forces. The jet
interaction effect is clearly seen when the jet is on during the spin cycle.
However, these results show that there is an effect on the lift force (although
reduced) for the remainder of the spin cycle even when the jet is off. This is a
result of the wake effects that persist from one spin cycle to another. The
impulse obtained from the predicted forces for both jets seems to asymptote
after 30 spin cycles.
The results have shown the potential of CFD to provide insight into the jet
interaction flowfields and provided guidance as to the locations and sizes of
the jets to generate the control authority required to maneuver a spinning
munition to its target with precision. This research represents a major increase
in capability for determining the unsteady aerodynamics of munitions in a new
area of flow control and has show n that microadaptive flow control with tiny syn-
thetic jets can provide an affordable route to lethal precision-guided infantry
weapons.
References
‘Sahu ., Heavey, K. R., and Ferry, E. N., “Computational Fluid Dynamics for Multiple
Projectile Configurations”, Proceedings of the 3rd Overset Com posite Grid and Solution
Technology Symposium,
Oct. 1996.
’Sahu
., Heavey,
K.
R., and Nietubicz, C. J., “Time-D ependent Navier-Stokes Com-
putations for Submunitions in Relative Motion,”
6th International Sym posium on Compu-
tational Fluid Dynamics,
Sept. 1995.
3Smith, B. L., and Glezer, A., “The Formation and Evolution of Synthetic Jets,”
Journal of Physics of Fluids, Vol.
10, No. 9, 1998.
4Amitay, M., K ibens, V., Parekh, D., and Glezer, A., “The Dynamics of Flow Reattach-
ment
over
a Thick Airfoil Controlled by Synthetic Jet Actuators,” AIAA Paper 99-1001,
Jan. 1999.
’Arunajatesan, S., and Sinha, N., “Towards Hybrid LES-RANS Computations of
Cavity Flowfields,” AIAA Paper 2000-0401, Jan.
2000.
6Batten, P., Goldberg, U., and Chakravarthy,
S.,
“Sub-grid Turbulence Modeling for
Unsteady Flow with Acoustic Resonance,” 38th AIAA Aerospace Sciences Meeting,
AIAA Paper 00-0473, Jan. 2000.
’Pulliam, T. H., and Steger, J. L., “On Implicit Finite-Difference Simulations of Three-
Dimensional Flow,” AIM
Journal, Vol.
18, No.
2,
1982, pp. 159-167.
‘Peroomian, O., Chakravarthy, S., Palaniswamy, S., and Goldberg, U., “Convergence
Acceleration for Unified-Grid Formulation Using Preconditioned Implicit Relaxation,”
AIAA Paper 98-01 16, June 1998.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 592/637
596 J. SAHU
’Goldberg, U., Peroomian, O., and Chakravarthy, S., “A Wall-Distance-Free k-e
Model With Enhanced Near-Wall Treatment,”
ASME Journal
of
Fluids Engineering,
‘‘finehart, C., McM ichael, J.
M.,
and Glezer, A., “Synthetic Jet-Based Lift Generation
“Sahu, J., “Unsteady Numerical Sim ulations of Subsonic Flow over a Projectile with Jet
V O ~ .20
1998,
pp. 457-462.
and Circulation Control on hisymmetric Bodies,” AIAA Paper 2002-3 168 June 2002.
Interaction,” AIAA Paper 2003-1352 Jan. 2003.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 593/637
IV. Exploring a Visionary Use
of
Circulation Control
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 594/637
Chapter 24
Coanda Effect and Circulation Control for
Nonaeronautical Applications
Terence
R.
Day*
Vortex Dynamics Pty Ltd, Mount Tamborine, Queensland, Australia
I. Introduction
T TH E “Coanda Effect/CC Workshop in Hampton, Virginia (March
16- 17,
A 004)”’
the question was posed, “What are the roadblocks to further devel-
opment?” Those roadblocks may be a result of a failure to address certain
deficiencies or an inability to find solutions. Exam ples of operationa l deficiencies
are insufficient quantity of CC air, heavy, com plicated air pumps, heavy, energy-
wasting plumbing, and
so
on. To address some of these issues the author
describes here a number of practical nonaeronautical devices employing the
Coanda effect or Coanda/Circulation Control (CC), a novel high-volume pum p
and a novel fan to supply C C air.
These projects are proposed com mercial outcomes for the Coanda effect and
CC. The purpose is to describe these novel applications and propose that some
creativity may be beneficial in promotion of the Coanda effect and CC to gain
credibility in a wider arena than only within the Coanda effect/CC scientific
community.
contain ade-
quate history and applications of the Coanda effect as it relates to CC and the
present author will start from this platform of knowledge and show its appli-
cations to novel nonaeronautical situations.
The overview papers in this book and other available
*Consultant.
Copyright
005
by Terence
R.
Day. Published by the American Institute of Aeronautics and
Astronautics, Inc., with permission.
599
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 595/637
6
T.
R
D Y
11. Applications
Oscillating Channel Flow Including Self Oscillating Channel Flow
.
Coanda Effect)
Although this phenom enon has been understood for quite some time,4 it appar-
ently has been a curiosity with little vision for many useful applications. The geo-
metry of a rectangular channel that enables jet self-oscillating flow must be
relatively precise to work at all. Gas jets in a channel w ill oscillate by imposition
of a pressure change alternating either side of the jet . W ith precise geometry, a
round jet will self-oscillate (Fig. 1). It is not difficult to produce either type of
oscillating flow if the air supply is sourced conveniently from the lab compressor.
For some applications including airborne odor treatment, certain chemicals
are coated onto surfaces in order to interact with a turbulent airflow. If the
airflow is laminar, the odor molecules contained in the airflow cannot contact
the chemical coated surfaces. Oscillating channel flow gives the desired turbu-
lence. A second reason for employing oscillating channel flow is that as the je t
skips from wall to wall, a particularly formed passageway is able to accept
each branch of the flow.
The significant breakthrough here is being able to convert a highly turbulent
fan flow into a flow structure that ca n self-oscillate in a channel. The author is
not aware of any previous work describing this. The result is a practical device
employing the Coanda effect (oscillating or self-oscillating jet flow), which is
efficient, easy to manufacture and has higher efficiency distribution of air
throughout a room.
B. Ring Vortex Projection
The vortices shown in Fig. 2 are generated from air slugs such as would be
produced by a piston stroke or the stroke of an acoustic driver, but are far less
expensive to produce as they are fan-flow derived. The geometry required is pro-
prietary, but it can be said that the slug of air is then tripped through an orifice
plate and turned into a ring vortex.
Fig. 1
Wool
tuft enables visualization of self oscillating wall jet.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 596/637
NON ERON UTIC L PPLIC TIONS
601
Fig. 2 Ring vortices containing smoke generated from a proprietary vortex
generator
A ring vortex is able to travel many times the distance of a nozzle discharge
because the ring stores kinetic energy like a flywheel for a short time. Ambient
fluid is entrained from in front of the ring and transported to the rear and
so
the result is propulsion with minimal drag. The strength of the ring vortex is
purpose tuned and the atomized chemical is transported over a large distance
bound within the vortex. Proprietary techniques enable the self-oscillating wall
je t to remain attached to one wall longer than on the opposite wall.
A useful feature is that as the je t oscillates, one side may b e routed through a
labyrinthine pathway with walls coated with a che mical that may possess a large
surface area for longer interaction time and then returned to the inflow to the fan.
Makeu p air is venturied into the recirculating main flow within the system. Only
the smaller part is ejected as a ring vortex. These ring vortices may contain
fragrances or insecticides. They may transport chemicals to foliage in orchards
and the turbulence of the ring enables full wetting of each side of the leaves.
The chemical may be vaporized by pressure reduction, heating, ultrasound, or
any other suitable means.
The self-propelled ring vortex promotes whole room circulation because it
displaces air at a great distance, which must flow back around the room
towards the source. The amount of air in a ring vortex is less than nozzle flow,
but with the same system power it is more effective because the nozzle air is
unable to travel the required distance and can recirculate back through the fan
and so the objective is not achieved.
C. Coanda Vacuum Cleaner
One of several versions of this vacuum cleaner is presented here. Figure
3
shows an underside view of the vacuum while operating over glass with flour
representing the dirt. Viewing the picture from centrally, a ring of small
nozzles is seen.
A
novel high pressure fan (a Jetfan) drives air through these
nozzles, which stirs the carpet pile. Viewed further out is an annular slot
blowing air over a Coanda surface. The Jetfan must generate a significant
pressure differential on both sides to induce a vortex and a jet simultaneously.
This je t entrains dirt and then enters an annular suction slot. The a ir ascends,
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 597/637
602
T.
R
D Y
Fig. 3 Vacuum underside.
but the reduced pressure causes the air to spin while it simultaneously travels
medially
This makes it very difficult for particles to ascend as they have to travel
inwardly while spiraling. The vortex deposits the dirt into a flexible bag
(Fig.
4),
which do es not collapse onto the low-pressure vortex b ecause an even
lower pressure is generated between the bag and the bowl. The vortex flows
inwardly t o form a centr al vortex, which then return s through the fan to recircu-
late. In this way most of the air is recirculated, minim izing the quantity of dirt
needing removal by a filter. Some nondomestic versions need no filter.
Som e other features are proprietary. Th e Coanda effect and the sim ple, low
cost Jetfan, are the main features of this vacuum cleaner.
D.
Coanda Chicken Shed
The Beaudesert Shire Council, a local government authority in Queensland,
Australia, gave approval for a housing estate near to chicken meat production
Fig.
4
Flexible bag in bowl.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 598/637
NON ERON UTIC L PPLIC TIONS
603
sheds. Large fans discharge foul air and dust towards the houses, which caused
the residents to threaten legal action. The company refused to close down,
so
the Shire Council explored various ways to solve the problem.
Their consultants suggested ducting the discharge horizontally and then verti-
cally to dilute with prevailing winds. That is impractical because of the losses
through ducting, especially at the right angle, and the ducting is expensive.
The system resistance causes the fan motors to overheat, which may bum out
in hot weather o r draw excessive current, thereby increasing running costs.
The author proposed a solution, as depicted in Fig.
5
which shows a wool tuft
turning 90 deg around a Coanda surface. The difficulty was how to capture the
turbulent fan flow onto the C oan da surface, especially when the air speed i s rela-
tively low. Once captured , the flow entrains ambient air from the direction of the
housing estate instead of blowing towards it. The Shire Council agrees that this
technique could be a large part of the solution. The author is negotiating with
private enterprise to build these low-cost Coanda surfaces at the end of
chicken sheds where there is a need.
E. Coanda Ceiling Fan
Figure
6
illustrates a smoke-filled air pathway from top side to underside of a
toroidal body. An annular jet exits the top at a certain angle over a step with
particular geometry. The jet trips over the step and three counter-rotating ring
vortices circle the top side (standing ring vortices). These entrain ambient air
and a turbulent flow travels outwardly and circulates to the underside.
The je t is the working fluid and that sam e amount of air reenters the underside
peripheral suction slot. The surplus ambient air entrained into the jet on top is
shed underneath. By altering underside geometry, shed air can be diffused or
alternatively shed as a concentrated plume. Th e body can be translucent with a
circular fluorescent tube inside. Excellent Coanda mixing enhances air-
conditioned air distribution throughout the room.
F. The Jetfan
The Jetfan (Fig. 7) may be a low-cost solution to many applications of
the Coanda effect that use fan-generated flow instead of compressor air. The
Fig.
5
Operating proof of concept prototype.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 599/637
604
T.
R
D Y
Fig. 6 Smoke visualizes
flow.
following results are for the Jetfan water pump performance and demonstrate the
unique characteristics of the impelle r com pared to other pum p impellers; they are
highly indicative of similar characteristics for the fan version, that is, no stall
wi thout s tators or a di ff~ser .~
“The visual inspection of the onset of cavitation ind icated that over the range
of flow rates tested, cavitation first appears at a rotational speed of 33 rpm.
Above this speed, cavitation bubbles were observed to fo rm on the concave
su8ace of each blade near the leading edge (LE)and to be reabsorbed a short
distance inside the blade passage. The point of reabsorp tion corresponds to a
Fig. 7 Injection-moldedJetfan.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 600/637
NON ERON UTIC L PPLIC TIONS
605
line drawn at right angles rom the LE
o
the convex su ace
o
the adjacent
blade. This reabsorption indicates that the pressure is rising as the water
enters the blade passage.
In
comparing the pe ormance
o
the Jetfan water pump with other pump
designs, it must be noted that the pe ormance detailed in this report has been
achieved without the use
o
a complex volute or stator blades, which are
com-
monly used to direct the ow
o
water rom the rotating impeller into the outlet
pipe in many pu mp designs.”
These fans and w ater pumps are useful in producing fan o r pump flows of suf-
ficient power for some CC applications. Potential applications are the NOTAR
(Fig. 8 and some other high-flow but lower-velocity CC applications and
water applications where high-speed water jets may boil. The Jetfan gives a
60
mechanical efficiency in a 5-in.-diam version with a high static efficiency
and enjoys no stators or diffuser.
The Jetfan performance is similar to an efficient mixed-flow fan employing a
stator row and diffuser. It has a “no stall” charac teristic. It has an axial inflow and
discharge.
Significant static pressure is generated within the blade passageways and by
employing no stators and no volute with its tongue o r cutwater, wake collisions
are eliminated and noise reduced. This may have applications for stealth and even
such mundane applications as water pumps for kitchen sinks, and so on, in ship-
ping, including submarines.
The Jetfan, including the water-pump version (Fig. 9), is of complex geometry
with overlapping blades. These fans and water pumps are, however, able to be
made at low cost because a manufacturing method (Fig.
10)
has been invented
to enable them to autorotate from the tooling and can be m ade for approximately
the same price as any low-cost, injection-moulded impeller. (Note, the Jetfan
technology and patents are the property of
DBG
Investments
Ry
Ltd.)
The same manufacturing method enables axial flow fans with overlapping
blades (Fig. 11) to be manufactured at low cost, and metal centrifugal impellers
Fig.
8
NOTAR.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 601/637
6 6
T.
R
D Y
Fig. 9 Jetfan water pumps.
to be made with high-performance geometry and with the ability to be rotation-
ally ex tracted from the m ould instead of employing investment casting and sub-
sequent milling for precision.
G.
Wind Turbines and Orbital
Pump
Full-span and tip b lowing6 is proposed. Wind-tunnel testing has indicated that
turbine efficiency increases of
30-40
are likely after all parasitic losses are
subtracted.
There are two main points here. First, wind turbine power generation is a
potential application for CC which could be revolutionized by a significant
increase in efficiency. The author believes this should be explored fully as
soon as possible before the world trend toward alternative energy sources, includ-
ing wind-power, progresses further, thus making it difficult later to retrofit this
innovation. Secondly, it is likely the practitioners of CC have discovered that
there are few CC applications where adequate air supply can be obtained for
control air. The NOTAR is a successful exception. The V22 tilt-rotor exhaust
deflection i s another good example, but CC there is not, strictly speaking, critical
to the aircraft performance. Many proposed applications, including some suc-
cessfully achieved, are risky, because o ther aircraft systems may be com promised
generally or occasionally. If the only reason that CC development has stagnated
Fig. 10 Manufacturing tooling.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 602/637
NON ERON UTIC L PPLIC TIONS
607
Fig.
11
Axial fan rotational extraction from mould, enab ling blade overlap .
is that of insufficient control air, then CC application to w ind turbines would not
be likely to be any more successful
It is likely that if CC is to be applied to wind turbines that the problem of
inadequate supply of control air be addressed simultaneously. That potential
solution may also apply to other uses of the Coanda effect or CC. Therefore,
this subject has two elements: 1) considering circulation con trol for wind turbines
and 2) examining the air pump needed to provide the CC air.
The basic idea of the Orbitalpump is shown in Fig. 12. It shows how the pins (in
some versions) that support the pistons are activated to allow the pistons to change
over, one replacing the other. The main features of the Orbitalpump are that it is high-
volume, relatively low-speed, low-noise, low-wear, fills and exhausts simul-
taneously, and can function as either a compressor or high-volume air pump, or both.
The O rbitalpump is intended to be the hub of a CC wind turbine (Fig. 13). For
most applications, the Orbitalpump shell, being a hollow toroidal body, remains
stationary while the shaft is turned. In the case of wind turbines the shaft may be
held stationary while the pump body rotates with the blades. The advantage here
is that the pressurized air can be fed almost directly into the hollow blades, thus
eliminating significant amounts of plumbing and the accompanying losses. It also
simplifies air delivery to the blowing slots.
The Orbitalpump may be attached to the hub of fans, including CC centrifugal
fans. Furey and Whitehead show the results of applying CC to a centrifugal fan.
“The better performing combination of these variations was the low solidity
0
0.65) impeller mated with a reduced internal volume volute. This fan
demonstrated a flow rate increase of 100% over that achieved at the design
point, through increasing the flow of control air, while maintaining a constant
head rise. The peak efficiency of this combination was 83% percent.” Notice
the fan achieved a 100% increase in flow over the design point while maintaining
head pressure with 83% efficiency.
Fig.
12
Orbitalpump piston changeover.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 603/637
608
T.
R
D Y
Fig.
13
Nylon O rbitalpump.
It is likely that shifting the rear stagnation point and attenuating o r elimina ting
tip vortices in wind turbines and fan blades is as valid as it is for aircraft wings.
Applying C C to wind turbines may have o ther benefits. A smaller diameter wind
turbine may achieve the sa me efficiency as a larger one. This would reduce man-
ufacturing costs, reduce maintenance, and reduce stress on components. It may
also enable higher efficiency in areas of lower wind speed. A wind turbine and
Orbitalpump combination is now being developed.
The Orbitalpump appears to be the highest volume positive displacement
pump possible. This high capacity is increased by multistaging on one shaft.
Other applications of the Orbitalpump may include a compressor, a pump, a
supercharger, a refrigeration compressor, and low-speed, high-volume water
pumps. A manufacturing license has been granted to apply small versions for
sleep apnea (respiratory support). For CC aircraft applications it can be placed
close to the preslot plenum with minimum plumbing.
H. Hovercraft/WIG
This model hovercraft/wing in ground effect craft (WIG) is aimed at the
hobby market and the entertainment industry. Figures
14
and
15
show existing
W IG craft' and Fig. 16 shows the proposed
X
Hovercraft/WIG. It employs
two methods of blowing generally called the Coanda effect. One method is
upper surface blowing (USB), where a large mass flow scrubs the upper
surface. It also employs CC, which is achieved by a thin wall jet circulating
over the rim. In existing USB applications for wings, USB gas may be supplied
Fig. 14 EkranoplanlWIG.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 604/637
NON ERON UTIC L PPLIC TIONS
609
Fig.
15
ArnphistarlWIG.’
from the engine nozzle, the jet spreading out to scrub the top of the wing.
This USB flow may be induced to coflow with the C C jet around the trailing
edge (TE).
Similarly, this circular planform employs two annular blowing slots. The more
central slot produces “USB” and the more peripheral CC slot entrains the USB
flow over to underneath. Sm all models of 2 ft in diameter cannot carry a com pres-
sor and so the peripheral blowing slot is replaced by several suction slots. These
suction slots serve to reduce the pressure over the rim and return a ir to the internal
fan (a Jetfan having proved the most efficient).
One of several models is shown in Fig. 17 hovering above a table in a still
taken from a video. The two wires seen underneath are restraints in case of
instability. That particular version employs a SuperTigre 90 model aircraft
engine, a tuned pipe, and a Jetfan. The model lifts onto an air cushion by the fol-
lowing mechanisms. The fan (shown in Fig. 18) pumps a large amount of air to
scrub the top surface (USB). The suction generated is by Bernoulli’s principle.
Ideally, a peripheral CC slot would also blow. In the case of this model, as
stated, suction slots are employed instead. This lowers the pressure over the
rim and the USB flow circulates to underneath and pressurizes the underside
by jet stagnation, which lifts the craft onto an air cushion. Suction slots have
been em ployed before for other applications and otherwise have been suggested
by many.
Jacques Cousteau’s yacht the “Halcyon,” employed suction slots each side of a
metal sail (Figs. 19 and 20) with a reported dram atic increase in thrust (available
at http://www.cousteau.org/en/cousteau-world/o~-s~ps/alcyone.php?sPlug=
1). It is claimed that the Turbosail has efficiency 3.5 to 4 times that of a cloth sail.
The disadvantage of using suction slots in this manner is that inflow to the fan
Fig. 16
X
Hovercraft/WIG.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 605/637
61
T.
R
D Y
Fig. 17 Model hovering.
throat is impeded and so efficiency of these CC sails and of the hovercraft/WIG
suffers somewhat. All the proprietq information regarding roll, pitch, and yaw
control of the Hovercraft/WIG cannot be presented here.
It should be noted that with this particular model although roll control was
achieved, pitch control was impaired by asymmetric inflow because of the
tuned pipe positioned in the inlet duct, which distorted the underside plate.
This caused the model to dip on that side, so a small stay was placed under the
edge. As this video was aimed at the movie industry to demonstrate other
skills, that stay and a thin wire preventing countertorque were digitally
removed. Pitch control, countertorque, and yaw control are achieved, but are
not depicted as they are proprietary.
The main point here is that a curious result emerged. When weights were
placed on the model to test lift,
it
supported a
100
payload.
A
paper by
Fig. 18 Top removed, showing fan.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 606/637
NON ERON UTIC L PPLIC TIONS
61 1
Fig.
19 Coanda sails.
Imber and Rogers’ discussed testing performed on a similar configuration. Imber
and Rogers’ work was aimed at other applications such as air and underwater
control surfaces, radome scanning sensors, rotor hub fairings on helicopters,
marine propellers and aircraft wings that have parabolic tips, and towed under-
water arrays. Imber and Rogers showed that by varying positions of azimuthal
blowing, they could achieve roll and pitch moments. This was achieved entirely
pneumatically. They did not address the issue of counter-torque; however, the
author has addressed that with satisfactory results, also achieved pneumatically
without any projectin surfaces.
Imber and Rogers paper reveals achievement of a) roll control, b) pitch
control, c) omnidirectional capability, and d) lift augmentation. In addition, the
author shows a) upper surface blowing of high mass flow (USB), b) rim
blowing slot (CC) or suction slots or both, c) coflow of USB/CC wall jets , and
d) self-contained powerplant and fan. The author has also established propulsion
means. These small models have achieved VTOL through a type of surface effect
or air cushion. It is well understood that to translate from this hovering/loitering
mode into a WIG mode of ground effect travel will require further work and
experimentation on larger models. Indeed, if a manned craft is attempted, like
any other CC applications, a suitable high-flow pump will need to be found to
provide adequate CC air. Perhaps the O rbitalpump will fill that need.
Fig.
20 View of TE slot.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 607/637
612
T.
R
D Y
111. Conclusions
There are many o ther important applications for the Coanda effect and CC in
addition to aeronautical ones. The Coanda effect has proven to be very effective
when applied to the underside of a vacuum cleaner pickup head. Th is may be one
of the first commercial applications. The performance of other smaller domestic
appliances may be improved by employing the Coanda effect as it can simplify
design and reduce production costs. For example, self-oscillating channel flow
eliminates the need for complex and more expensive mechanical and electrical
actuators. This in turn allows for ring vortex propagation, which can transport
a substance much further than any nozzle discharge employed in small appliances
at present, and gives better whole room circulation than present nozzles. Coanda
ceiling fans may be far safer than conventional fans. Results suggest that CC may
make wind turbines more efficient.
The Coanda effect and CC may yield improvements in many industries and
applications. The author believes future research should concentrate on develop-
ing reliable, lightweight, and low-cost portable sources of blowing air instead of
laboratory compressor air. Wind turbine CC blades appear to benefit from the
bluff TE as cruise is not needed. CC aircraft wing TE geometry or mechanical
factors will need to be improved because of the need for cruise capability.
The applications given should stimulate increased interest in solving the very
few but important impediments to being able to incorporate the Coanda effect and
CC into aeronautical, entertainment, industrial, and domestic applications.
Acknowledgments
The author is a member of the International Society of A utomotive Engineers
and is consultant to
1)
the entertainment industry producing special effects
(including on-stage tornados 22ft high) and 2) industry in fluid movement
including Coanda effect applications and ring vortex technology for air-care,
insect control, and odor elimination.
References
‘Jones, G.
S.
and J o s h R. D., (eds.), 2004 NASA/ON R Circulation Control Work-
Englar, R. J., “Development Of The A-G/Circulation Control Wing Flight Demon-
stration Configuration,” David W. Taylor Naval Ship Research And Development
Center Bethesda, MD, Jan. 1979.
3Rogers, E O., Schwartz,. A. W., and Abramson, J.
S.
“Applied Aerodynamics of
Circulation Control Airfoils and Rotors,” 1 th European Rotorcraft Forum, Sept. 1985.
4Murai, K., Kawashima,
Y.,
Nakanishi,
S.,
and Taga, M ., “Self Oscillation Phenomena
of
Turbulent Jets in a Channel,” JSME International Journal, Vol. 30, No. 266, May 1987,
5Dekkers, W., “Performance Tests on a 93 mm JETFAN water pump,” School of
Mechanical, Medical and Manufacturing Engineering, Univ. of Technology,
Queensland, Australia, Rept. No. C 2967 (C), Oct. 1998.
6Taylor, R. M., “Aerodynamic S urface Tip Vortex Attenuation System,” US Patent N o.
5,158,251, Oct. 27, 1992.
shop, NASA CP 2005-213509, M ar. 2005.
2
pp. 1243-1247.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 608/637
NON ERON UTIC L PPLIC TIONS
61
3
’Furey, R. J., and Whitehead, R. E., “Static Evaluation of a Circulation Control
Centrifugal Fan,” David W. Taylor Naval Ship Research and Development Center,
Bethesda, MD, June 1987.
‘Ekranoplans Very Fast Craft by The University of New South Wales, The Institute
of Marine Engineers (Sydney Branch), Univ. of New South Wales (Dept. of Naval Archi-
tecture), Australian Maritime Safety Authority, Australian Maritime Engineering CRC
Ltd., Russian Australian Advanced Technology Group, Dec. 1996, p. 152 (Amphistar),
154 (Ekranoplan).
’Imber, R. D., and Rogers, E. O., “Investigation of a Circular Planform Wing with
Tangential Fluid Ejection,” 34th Aerospace Sciences Meeting
Exhibit, Jan. 1996.
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 609/637
AUTHOR INDEX
Index Terms Links
A
Abramson, J. 69 445
Ahuja, K. K. 167 557
Alexander, M. G. 245
Anders, S. G. 245 469
Angle II, G. 277
B
Baker, W. J. 421 513
Blaylock, G. 383
C
Campbell, B. A. 315Cerchie, D. 113
Chang III, P. A. 445
D
Day, T. R. 599
E
Ebert, M. P. 445
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 610/637
Index Terms Links
Englar, R. J. 23 167 315 357
383 557
F
Fasel, H. F. 401
Frith, S. P. 337
G
Gaeta, R. J. 383 557
Gopalarathnam, A. 499 539
Gross, A. 113
H
Halfon, E. 113
Hammerich, A. 113
Han, G. 113
O’Hara, B. 277
Hassan, H. 499Huebsch, W. 277
I
Imber, R. 69
J
Johnson, S. K. 245
Jones, G. S. 191
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 611/637
Index Terms Links
L
Liu, Y. 557
Loth, J. L. 3
LutzTaubert 113
M
Marino, T. 445
McGowan, G. 499 539
Munro, S. E. 167
O
Owen, F. K. 105
Owen, A. K. 105
P
Paterson, E. G. 421 513
Paxton, C. D. 293
R
Rogers, E. 69
Rumsey, C. L. 469
S
Sahu, J. 579
Sankar, L. N. 567
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 612/637
Index Terms Links
Slomski, J. 445
Smith, J. 277
Swanson, R. C. 469
Lucie-Trouve 113
V
Varghese, P. 113
W
Wernz, S. 401
Wood, N. J. 337
Wygnanski, I. 113
X
Xiao, X. 499
Z
Zha, G.-C. 293
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 613/637
INDEX
Index Terms Links
A
Acoustic optimization, noise reduction
and 174
Active flow control (AFC) 403
Advanced CCW airfoils 40
dual-radius 41
supercritical 41
Aerodynamic heat exchanger (AHE)circulation control and 383
concept of 384
future use of 395
test results 389
aerodynamics 391
heat transfer 392
testing of 386
AFC. See active flow control.
AFSF. See anechoic flight simulation
facility.
AHE. See Aerodynamic heat exchanger.
Airfoil development
CFD techniques 31
circulation control and 31
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 614/637
Index Terms Links
Airfoil development (Cont .)
cruise configuration 228
high-lift mode 216
Airfoils
Bell A821201 279
blowing momentum 110
circulation control
concepts and 106
experiments on 107
measurement and analysis 105
numerical simulation and 469
sample results 107
co-flow jet method 294
conventional flap 118
elliptical 144
GACC design 202
NACA 0015 flapped 125
wake turbulence 111
wake velocities 108 Anechoic flight simulation facility
(AFSF) 171
Annular wing (CC-duct) 79
model specifications 81
Automobiles, pneumatic aerodynamic
technology and 357
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 615/637
Index Terms Links
B
BART, basic aerodynamic research tunnel 207
Basic aerodynamic research tunnel.
See BART.
Bell A821201 airfoil, Coanda effect on 279
computational model and procedure 282
computational results 286
experiment results 285
experimental apparatus and procedure 279
BLC. See boundary layer control.
Blowing coefficient, circulation control
stimulation test results and 525
Blowing momentum 110
Blowing, boundary layer control,
circulation control 115
Blown airfoils, two-dimensional drag 200
Blown airfoils, pneumatic flap
performance and 200
Boundary conditions, circulation control
airfoils and 476
Boundary conditions, FLUENT flow
solver and 543
Boundary conditions, steady and pulsed
jet effects 560 Boundary layer control 3
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 616/637
Index Terms Links
Boundary layer control, suction,
circulation control (CC) high lift
generation 3
history of 4
C
Cavitation 440
CC propeller 53
CC. See circulation control.
CC / jet deflection 51
CC-disc 85
CC-valve 91
CCW airfoils, advanced 40CCW. See circulation control wing 36
CCW / supercritical airfoils 41
CCW / upper surface blowing (USB) concept 318
CCW / USB, powered lift and engine
thrust deflection and 48
CFD techniques 31
CFD. See computational fluid
dynamics.
CFJ. See co-flow jet.
Channel wings, STOL aircraft
wind-tunnel evaluations and 326
Circular
Coanda surface, dual blowing 228
cylinder 405
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 617/637
Index Terms Links
Circular (Cont .)
stopped-rotor aircraft, circulation
control and 28
controlled flow and 150
DNS 405
RANS 409
TE 217
wing (CC-disc) 85
specifications 86
Circulation control
aerodynamic heat exchanger (AHE) 383
airfoil
computational fluid dynamics
(CFD) 106
concepts 106
development, CFD techniques 31
flow prediction, turbulence
modeling 499
FLUENT flow solver 539 full Reynolds-stress modeling and 445
geometry and grid 472
measurement and analysis of 105
experiments on 107
sample results 107
numerical simulation 469
appendix 497
boundary and initial conditions 476
jet momentum coefficient 478
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 618/637
Index Terms Links
Circulation control (Cont .)
numerical method 475
results 478
turbulence modeling 476
pneumatic flap performance 193
appendix 237
results 216
steady and pulsed jet effects 557
transonic mach numbers test 245
configuration tested 247
facilities used 252
instrumentation used 251
procedures and conditions 253
results of 254
turbulent Coanda wall jet and 415
wake turbulence profile 111
wake velocities 108
blowing 20
blowing momentum 110 circular cylinder, controlled flow 150
co-flow jet (CFJ) airfoil method 294
demonstration of 12
elliptical airfoils 4
experiments 113
elliptical airfoil flow 144
flow control 118
GLAS II airfoil 130
NACA 0015 flapped airfoil 125
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 619/637
Index Terms Links
Circulation control (Cont .)
flight control 337
full-span UAV model 345
half-span model 339
flight testing of 12
Grumman Aerospace A-6A 16
larger aircraft 16
high-lift generation 3
noise reduction 167
nonaeronautical applications 599
hovercraft 608
orbital pump 606
wind turbines 606
pneumatic aerodynamics
advanced CCW airfoils 40
airfoil development 31
applications of 28
boundary layer control (BLC) 24
CC propeller system 53 circular cylinder stopped-rotor
aircraft 28
circulation control wing (CCW) 36
Coanda effect 25
Coanda, device 26
elliptic-airfoil CC rotor 28
fixed-wing aircraft applications 23
induced drag reduction 54
introduction to 24
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 620/637
Index Terms Links
Circulation control (Cont .)
microflyer and pulsed blowing 56
moment control 54
nonflying applications 57
other aircraft applications 53
powered lift and engine thrust
deflection 49
stability augmentation 54
X-wing aircraft 35
rounded trailing edge 4
short take-off and landing (STOL) 4
simulation, GACC wing and 515
boundary conditions 521
computational methods 516
computational resources 523
grid generation 518
initial conditions 521
test results 523
blowing coefficient 525 grid study 530
plenum vs. no plenum 524
technology
design capability status 99
workshops
annular wing (CC-duct) 79
circular wing (CC-disc) 85
dual-slotted cambered airfoil
(LSB) 70
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 621/637
Index Terms Links
Circulation control (Cont .)
dual-slotted low aspect ratio wing
(CC hydrofoil) 93
exploratory investigations,
NSWCCD 69
miniature oscillatory valve
(CC-valve) 91
self-driven rotary thruster
(TIPJET) 73
wings, (CCW) 36
conventional wings, noise reduction
comparison 182
demonstrator design 5
noise reduction, experiments 168
Coanda, ceiling fan 603
Coanda, device 26
Coanda effect 25 278
Bell A821201 airfoil and 279
computational model and procedure 282 computational results 286
experiment results 285
experimental apparatus and
procedure 279
nonaeronautical applications 599
ceiling fan 603
jetfan 606
oscillating channel flow 600
ring vortex projection 600
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 622/637
Index Terms Links
Coanda effect (Cont .)
vacuum cleaner 601
slot, setup errors 212
Co-flow jet (CFJ) method 294
advantages of 296
test results 296
energy expenditure 307
F-5E aircraft 308
performance 298
Computational fluid dynamics (CFD) 106
Conventional flap airfoil 118
Conventional wings vs. circulation
control wings, noise reduction
comparison 182
Cruise configuration
airfoil performance and 228
circular Coanda surface, dual blowing 228
pulsed blowing 232
Custer channel wing aircraft 316
D
DES. See detached-eddy simulation.
Detached-eddy simulation (DES) 421
computational methods, unsteady RANS 425
NCCR airfoil
computational methods 424
grid generation 427
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 623/637
Index Terms Links
Detached-eddy simulation (DES) (Cont .)
initial and boundary conditions 429
test conditions 424
test results 430
cavitation 440
RANS simulation 430
Direct numerical simulations. See DNS.
DNS
circular cylinder and 405
direct numerical simulations 403
test calculations, turbulent Coanda wall
jet and 404
turbulent Coanda wall jet and 402
Drag, pneumatic heavy vehicles and 363
Dual blowing, cruise configuration and 228
Dual-radius CCW 41
Dual-slotted cambered airfoil (LSB) 70
Dual-slotted low aspect ratio wing
(CC hydrofoil) 93
E
Elliptic-airfoil CC rotor, circulation
control and 28
Elliptical airfoil flow 4 144
Equal lift condition 182
Equivalent drag 201
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 624/637
Index Terms Links
Exploratory investigations, circulation
control technology workshops,
NSWCCD 69
F
F-5E aircraft, co-flow jet method and 308
Flight control
circulation control
full-span UAV model 345
half-span model 339
wing and 337
Flight testing
circulation control and 12Grumman Aerospace A-6A 16
Flow attachment, STOL aircraft
wind-tunnel evaluations and 327
Flow control, conventional flap airfoil,
circulation control experiments
and 118
Flow prediction, turbulence modeling 499
FLUENT flow solver 539
experiments 541
numerical approach 542
boundary conditions 543
grid details 542
solver settings 543
test results 545
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 625/637
Index Terms Links
FLUENT flow solver (Cont .)
free-air conditions 545
wind-tunnel wall effects 547
Freestream velocity, steady and pulsed
jet effects and 566
Fuel economy, pneumatic heavy vehicles
and 367
Full Reynolds-stress modeling,
best turbulence models 460
circulation control airfoils 445
mathematical development 448
Full-span UAV model
circulation control flight control and 345
experiments results 345
G
GACC
airfoil design 202
BART 207
juncture flow regions 207
solid blockage 206
wake blockage 206
balance limits 208
general aviation circulation control 202
wing, steady circulation control
simulation 513
boundary conditions 521
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 626/637
Index Terms Links
GACC (Cont .)
computational methods 516
computational resources 523
grid generation 518
initial conditions 521
test conditions 515
test results 523
General aviation circulation control.
See GACC.
GLAS II airfoil 130
Grid
computational, projectile geometry
and 585
creation, steady and pulsed jet effects
and 560
details, FLUENT flow solver and 542
generation 427
generation, circulation control
stimulation and 518 generation, GACC wing and 518
study, circulation control stimulation
test results and 530
Grumman Aerospace A-6A 16
H
Half-span CCW model, circulation
control flight control 339
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 627/637
Index Terms Links
Heat transfer, aerodynamic heat
exchanger and 392
Heavy vehicles (HV), pneumatic
aerodynamic technology 357
pneumatic test results 360
blown 362
drag increase 363
drag reduction 363
stability and control 365
unblown 361
wind tunnel evaluations 371
High-lift mode
baseline performance 216
circular TE 217
TE performance comparisons 226
Hovercraft 608
Hybrid RANS / LES turbulence model,
jet-based flow control computer
simulation and 583
I
Induced drag reduction 54
Initial conditions, circulation control
airfoils and 476
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 628/637
Index Terms Links
J
Jet momentum coefficient 478
Jet-based flow control 579
computer simulations 581
hybrid RANS / LES turbulence
model 583
unsteady jet boundary conditions 582
projectile geometry 584
simulation results 586
nonspinning projectile 586
spinning projectile 589
Jet slot height effects, steady and pulsed
jet effects and 567
K
Kutta condition 114
L
Large eddy simulation (LES) 403
Leading edge blowing, steady and pulsed
jet effects and 565
LES. See large eddy simulation.
Lower surface blowing. See LSB.
LSB, lower surface blowing 70
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 629/637
Index Terms Links
M
Mass flow, pneumatic flap performance
and 202
Microflyer and pulsed blowing 56
Miniature oscillatory valve (CC-valve) 91
Moment control 54
N
NACA 0015 flapped airfoil 125
NASA, circulation control wings,
requirements for 193
NCCR airfoil, detached-eddy simulation (DES) 421
computational methods 424
unsteady RANS 425
grid generation 427
initial and boundary conditions 429
test conditions 424
test results 430 cavitation 440
RANS simulation 430
Noise reduction,
acoustic optimization 174
circulation control and 167
circulation control wings vs.
conventional wings 182
experiments 168
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 630/637
Index Terms Links
Noise reduction, (Cont .)
equal lift condition 182
experiments
background information 169
facilities and instrumentation 171
results and discussion 174
technical approach 173
facilities and instrumentation,
anechoic flight simulation facility
(AFSF) 171
Nonaeronautical applications
circulation control and 599
hovercraft 608
orbital pump 606
wind turbines 606
Coanda effect and 599
Coanda ceiling fan 603
Coanda vacuum cleaner 601
jetfan 606 oscillating channel flow 600
ring vortex projection 600
Nonflying applications, circulation
control and 57
Nonspinning projectile, simulation
results 586
NSWCCD, circulation control
technology and exploratory
investigations 69
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 631/637
Index Terms Links
NSWCCD, Naval Surface Warfare
Center, Carderock Division 70
Numerical method, circulation control
airfoils and 475
Numerical simulation 469
boundary and initial conditions 476
circulation control airfoils and, results 478
jet momentum coefficient 478
turbulence modeling 476
O
Orbital pump 606
Oscillating channel flow 600Outboard wing ON, STOL aircraft
wind-tunnel evaluations and 322
P
PCW. See pneumatic channel wing.
PHV. See pneumatic heavy vehicles.
Plenum vs no plenum, circulation control
stimulation test results and 524
Pneumatic
aerodynamic technology
automobiles and 357 heavy vehicles (HV) and 357
sport utility vehicles and 374
aerodynamics
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 632/637
Index Terms Links
Pneumatic (Cont .)
basics of 358
boundary layer control (BLC) 24
CC propeller 53
circulation control
advanced CCW airfoils 40
airfoil development 31
applications of 28
circular cylinder stopped-rotor
aircraft 28
circulation control wing (CCW) 36
elliptic-airfoil CC rotor 28
other aircraft applications 53
powered lift and engine thrust
deflection 49
X-wing aircraft 35
Coanda device 26
channel wing (PCW) 52 319
flap performance 193 airfoil performance 216
blown airfoils, two-dimensional
drag 200
equivalent drag 201
experiments 207
Coanda slot setup errors 212
GACC
airfoil design 202
balance limits 208
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 633/637
Index Terms Links
Pneumatic (Cont .)
mass flow requirements 202
NASA requirements 193
theoretical considerations 195
heavy vehicles
blown test results 362
drag reduction test results 363
fuel economy testing 367
stability and control test results 365
test conclusions 379
test recommendations 380
test results 360
unblown test results 361
wind tunnel evaluations 371
powered-lift super STOL aircraft 315
sport utility vehicles (PSUV) 374
tests on 376
Powered lift and engine thrust deflection 49
CC / jet deflection 51
CCW / USB 48
pneumatic channel wing 52
Projectile geometry 584
PSUV. See pneumatic sport utility
vehicles.
Pulsed blowing, cruise configuration and 232
Pulsed jet effects, test results 570
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 634/637
Index Terms Links
R
RANS
circular cylinder and 409
detached-eddy simulation (DES) 421
Reynolds-averaged Navier-Stokes 403
simulation, NCCR airfoil and 430
test calculations, turbulent Coanda wall
jet and 405
turbulent Coanda wall jet and 402
unsteady 425
RANS / LES turbulence model, hybrid,
jet-based flow control computer
simulation and 583
Reynolds-averaged Navier-Stokes.
See RANS.
Ring vortex projection 600
S
Self-driven rotary thruster (TIPJET) 73
Separation control experiments 113
Sharp trailing edge, circulation control
rounded trailing edge 4
Short take-off and landing. See STOL.
Solid blockage 206
Solver settings, FLUENT flow solver
and 543
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 635/637
Index Terms Links
Spinning projectile
jet-based flow control 579
computer simulations 581
projectile geometry 584
simulations, results 586
Sport utility vehicles (SUV), pneumatic
aerodynamic technology 374
Stability augmentation 54
Steady and pulsed jet effects
boundary conditions 560
circulation control airfoil and 557
grid creation 560
mathematical equations 559
test results 561
freestream velocity 566
jet slot height effects 567
leading edge blowing 565
pulsed jet effects 570
Strouhal number effects 573 validation of 562
STOL
aircraft
CCW / upper surface blowing (USB)
concept 318
Custer channel wing aircraft 316
experiments on
evaluation and techniques 320
predictions vs actual 331
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 636/637
Index Terms Links
STOL (Cont .)
wind-tunnel evaluations 321
future configurations 333
pneumatic channel wing (PCW) 319
wind-tunnel evaluations
channel wings 326
flow attachment 327
outboard wing ON 322
circulation control 4
demonstrator design 6
STOL, short takeoff and landing 316
Strouhal number effects, steady and
pulsed jet effects and 573
Supercirculation 114
Super-STOL aircraft 315
SUV. See sport utility vehicles.
T
TE performance 226
TE. See trailing edge.
TIPJET 73
TIPJET rotor specifications 76
TKE. See turbulent kinetic energy.
Transonic mach numbers 245
Tests using
configuration used 247
facilities used 252
8/19/2019 Applications of Circulation Control Technologies
http://slidepdf.com/reader/full/applications-of-circulation-control-technologies 637/637