Conversion Factors
Density:
1.00 kg/rn3
1.9404 X
slug/ft3
6.2430 X
16.018 kg/rn3
ibm/ft3
1.00 ibm/ft3
3.1081 X
slug/ft3 =
Energy or work:
1.00 cal = 4.187 J = 4.187 N rn 1.00 Btu = 778.2 ft lbf 0.2520 kcal = 1.00 kW h = 3.600 x
Flow rates:
1055
J
j
X
1.00 gal/mm = 6.309
Force:
1.00 N
m3/s = 2.228
X
ft3/s
=
dyne =
16.0 oz
0.2248 lbf
1.00 lbf 1.00 lhf
= 4.4482 N
1.00 U.S. ton = 2000 lbf
2.0 kip
Heat flux:
1.00 W/cm2
=
Length:
3.170
x
0,2388 cal/s cm2 = 0.8806 Btu/ft2. s But/ft2. h
1.OOm
= 3.2808
1.00 km =
1.00 ft 1.00 ft
= =
1.00 mile =
Mass:
ft = 39.37 in. 0.6214 mile = 1093.6 yd 0.3048 m = 30.48 cm 12 in = 0.333 yd 1760yd = 1609.344m 5280 ft
1000 g = 2.2047 ibm 1.00 slug = 32,174 ibm 14.593 kg
1.00 kg
Power: 1.00kW
1.00 kW
1.00 hp
= 1000.0 W = 1000.0
N rn/s =
0.2388 kcal/s
= 1.341 hp
550 ft lbf/s = 0.7457 kW
1.00 Btu/s = 1.415 hp
1.00 Btu/h = 0.29307 W
Pressure:
= 1.00 bar = 1.00 lbf/in.2 = 1.00 lbf/in.2 =
1.00 N/rn2
1.4504 X
lbf/in.2
2.0886 X 102 lbf/ft2
N/rn2 =
6.8947 X
Pa
N/rn2
144.00 lbf/ft2
Specific enthalpy:
1.00 N rn/kg = 1.00 rn/kg = 1.00 Btu/lbrn =
Specific heat:
1.00 J/kg
1.00 rn2/s2
4.3069 X 2.3218 X
Btu/lbm = N rn/kg
1.0781
ft lbf/slug
1.00 N rn/kg. K =
1.00 Btu/lbm
 K = 2.388)< Btu/lbm °R = 5.979 ft . lbf/slug = 32.174 Btu/slug= 4.1879 X N rn/kg. K = 4.1879 X 3/kg K
1.00 J/kg
Temperature: The temperature of the ice point is 273.15 K (491.67°R). 1.00 K = 1.80°R K = °C + 273.15 °R =°F+45967 T°F = 1.8(T°C) + 32 T°R = 1.8(T K)
Velocity: 1.00 rn/s 1.00 km/h 1.00 ft/s
= = = 1.00 mile/h = 1.00 knot =
1.00 kg/rn  s
km/h 0.2778 rn/s = 0.6214 mile/h = 0.9113 ft/s 0.6818 mile/h = 0.59209 knot 1.609 km/h 0.4470 rn/s 1.467 ft/s 1.15155 mile/h
3.60
Viscosity:
lbf s/ft2 = 0.67197 ibm/ft s = 2.0886 X 1.00 ibm/ft. s = 3.1081 X lbf' s/ft2 = 1.4882 kg/rn . s 1.00 lbf s/ft2 = 47.88 N  s/rn2 = 47.88 Pa s ibm/ft s 1.00 centipoise = 0.001 Pa s = 0.001 kg/rn. s = 6.7 197 X 1.00 stoke = 1.00 x rn2/s(kinematic viscosity)
Volume: 1.00 liter
1.00 ft3
=
1000.0 crn3
1.00 barrel =
= = 1.00 gal = 1.00 bushel =
5.6146 ft3
in.3 = 0.03704 yd3 = 7.481 gal 28.32 liters
1728
3.785 liters
=
3.785
x
m3
3.5239 x 102 rn3
AERODYNAMICS
FOR ENGINEERS
Fifth Edition
JOHN J. BERTIN
Professor Emeritus, United States Air Force Academy
and
RUSSELL M. CUMMINGS
Professor, United States Air Force Academy
PEARSON
Prentice
Pearson Education International
If you purchased this book within the United States or Canada you should be aware that it has been wrongfully imported without the approval of the Publisher or the Author.
Vice President and Editorial Director, ECS: Marcia .1. Horton Acquisitions Editor: Tacy Quinn Associate Editor: Dee Betnhard Managing Editor: Scott Disanno Production Editor: Rose Kernan Art Director: Kenny Beck Art Editor: Greg DuIles Cover Designer: Kristine Carney Senior Operations Supervisor: Alexis HeydtLong Operations Specialist: Lisa McDowell Marketing Manager: Tim Galligan
Cover images left to right: Air Force BlB Lancer Gregg Stansbery photo I Boeing Graphic based on photo and data provided by Lockheed Martin Aeronautics FiS I NASA.
p EARSO N
Prentice
© 2009,2002,1998,1989,1979 by Pearson Education, Inc. Pearson PrenticeHall Upper Saddle River, NJ 07458
All rights reserved. No part of this book may be reproduced in any form or by any means, without permission in writing from the publisher.The author and publisher of this book have used their best efforts in preparing this book.These efforts include the development, research, and testing of the theories and programs to determine their effectiveness. The author and publisher make no warranty of any kind, expressed or implied, with regard to these programs or the documentation contained in this book.The author and publisher shall not be liable in any event for incidental or consequential damages in connection with, or arising out of these programs.
ISBN10: 0132355213
ISBN13:
Printed in the
United States of America
10 9 8 7 6 5
Pearson
43
21
Education Ltd., London Pearson Education Singapore, Pte. Ltd Pearson Education Canada, Inc. Pearson Education—Japan Pearson Education Australia PTY, Limited Pearson Education North Asia, Ltd., Hong Kong Pearson Educación de Mexico, S.A. de CV. Pearson Education Malaysia, Pte. Ltd. Pearson Education Upper Saddle River, New Jersey
Contents
PREFACE TO THE FIFTH EDITION PREFACE TO THE FOURTH EDITION
CHAPTER 1
15 17
21
21
WHY STUDY AERODYNAMICS?
1.1
The EnergyManeuverability Technique
1.1.2
Specific Excess Power 24 Using Specific Excess Power to Change the Energy Height 25 1.1.3 John K Boyd Meet Harry Hillaker 26
1.1.1
1.2
Solving for the Aerothermodynamic Parameters 26
Concept of a Fluid 27 Fluid as a Continuum 27 1.23 Fluid Properties 28 1.2.4 Pressure Variation in a Static Fluid Medium 34 1.2.5 The Standard Atmosphere 39
1.2.1 1.2.2
1.3
Summary
42
Problems 42 References 47
CHAPTER 2
FUNDAMENTALS OF FLUID MECHANICS
2.1
48
2.2 2.3 2.4
2.5
2.6 2.7 2.8 2.9
Introduction to Fluid Dynamics 49 Conservation of Mass 51 Conservation of Linear Momentum 54 Applications to ConstantProperty Flows Reynolds Number and Mach Number as Similarity Parameters 65 Concept of the Boundary Layer 69 Conservation of Energy 72 First Law of Thermodynamics 72 Derivation of the Energy Equation 74
2.9.1 2.9.2
59
Integral Form of the Energy Equation 77 Energy of the System 78
5
Contents
2.9.3 2.9.4 2.9.5 2.9.6
Flow Work 78 Viscous Work 79 Shaft Work 80 Application of the Integral Form of the Energy Equation 80
82
2.10
Summary
Problems 82 Ref erences 93
CHAPTER 3 DYNAMICS OF AN INCOMPRESSIBLE, INVISCID FLOW FIELD
3.1
94
Inviscid Flows
94
3.2 3.3
3.4 3.5 3.6 3.7 3.8
3.9
Bernoulli's Equation 95 Use of Bernoulli's Equation to Determine Airspeed 98 The Pressure Coefficient 101 Circulation 103 Irrotational Flow 105 Kelvin's Theorem 106
3.7.1
3.&1
Implication of Kelvin Theorem 107
108
Incompressible, Irrotational Flow
Boundary Conditions 108
3.10
3.11 3.12
Stream Function in a TwoDimensional, Incompressible Flow 108 Relation Between Streamlines and Equipotential Lines 110 Superposition of Flows 113 Elementary Flows 113
Uniform Flow 113 Source or Sink 114 3.12.3 Doublet 116 3.12.4 Potential Vortex 117 3.12.5 Summary of Stream Functions and of Potential Functions 119
3.12.1 3.12.2
3.13
Adding Elementary Flows to Describe Flow Around a Cylinder 119
3.13.1 Velocity Field 119 3.13.2 Pressure Distribution 122 3.13.3 Lift and Drag 125
3.14 3.15
Lift and Drag Coefficients as Dimensionless FlowField Parameters 128 Flow Around a Cylinder with Circulation 133
3.15.1 3.15.2
Velocity Field 133
Lift and Drag 133
Contents
3.16
3.17 3.18
Source Density Distribution onthe Body Surface 135 Incompressible, Axisymmetric Flow
3.17.1
140
Flow around a Sphere 141
143
Summaiy
Problems 144 References 157
CHAPTER 4
VISCOUS BOUNDARY LAYERS
4.1
158
Equations Governing the Boundary Layer for a Steady, TwoDimensional, Incompressible
Flow
159
163
4.2 4.3
Boundary Conditions 162 Incompressible, Laminar Boundary Layer
4.3.1
Numerical Solutions for the FalknerSkan
Problem 166
4.4 4.5
BoundaryLayer Transition 180 Incompressible, Turbulent Boundary Layer
4.5.1
183 Derivation of the Momentum Equation for Turbulent Boundary Layer 184 4.5.2 Approaches to Turbulence Modeling 187 Turbulent Boundary Layer for a Flat Plate 188
4.6 4.7
Eddy Viscosity and Mixing Length Concepts
191
Integral Equations for a FlatPlate Boundary Layer 193
Application of the Integral Equations of Motion to a Turbulent, FlatPlate Boundary Layer 196 4.7.2 Integral Solutions for a Turbulent Boundary Layer with a Pressure Gradient 202
4.7.1
4.8
Thermal Boundary Layer for ConstantProperty
Flows
4.8.1
203 Reynolds Analogy 205 4.8,2 Thermal Boundary Layer for Pr
1 206
4.9
Summary
210
Problems 210 References 214
CHAPTER 5 CHARACTERISTIC PARAMETERS FOR AIRFOIL
AND WING AERODYNAMICS
5.1
215
Characterization of Aerodynamic Forces and Moments 215
5.1.1 5.1.2
General Comments 215 Parameters That Govern Aerodynamic Forces 218
Contents
5.2
Airfoil Geometry Parameters
219
5.2.1 AirfoilSection Nomenclature 219 5.2.2 LeadingEdge Radius and Chord Line 220 5.2.3 Mean Camber Line 220
5.2,4
5.2.5 5.3
Maximum Thickness and Thickness Distribution 221 TrailingEdge Angle 222
5.4
WingGeometry Parameters 222 Aerodynamic Force and Moment Coefficients 229
Lift Coefficient 229 Moment Coefficient 234 Drag Coefficient 236 BoundaryLayer Transition 240 Effect of Surface Roughness on the Aerodynamic Forces 243 5.4.6 Method for Predicting Aircraft Parasite Drag 246
5.4.1
5.4.2 5.4.3 5.4.4 5.4.5
5.5
Wings of Finite Span 5.5.1 Lift 257 5.5.2 Drag 260
5.5.3
256
Lift/Drag Ratio 264
Problems 269 References 273
CHAPTER 6
INCOMPRESSIBLE FLOWS AROUND AIRFOILS OF INFINITE SPAN
6.1
275
6.2
6.3 6.4 6.5
General Comments 275 Circulation and the Generation of Lift
6.2.1
276
Starting Vortex 277 281
General ThinAirfoil Theory 278 Thin, FlatPlate Airfoil (Symmetric Airfoil) Thin, Cambered Airfoil 285
6.5.1 6.5.2
Vorticity Distribution 286 Aerodynamic Coefficients for a Cambered Airfoil 287
6.6 6.7 6.8
6.9
LaminarFlow Airfoils 294 HighLift Airfoil Sections 299 Multielement Airfoil Sections for Generating High Lift 304 HighLift Military Airfoils 312
Problems 314 References 316
Contents
CHAPTER 7
INCOMPRESSIBLE FLOW ABOUT WINGS OF FINITE SPAN

319
7.1
7.2 7.3
General Comments 319 Vortex System 322 LiftingLine Theory for Unswept Wings
7.3.1 7.3.2
323
TrailingVortices and Down wash 326 Case of Elliptic Spanwise circulation
Distribution 328
7.3.3
Technique for General Spanwise Circulation Distribution 333
7.3.4 7.3.5
LiftontheWing 339
VortexInduced Drag 339 Z3.6 Some Final Comments on LiftingLine Theory 346
7.4
Panel Methods 349 7.4.1 Boundary Conditions 350 7.4.2 Methods 351
7.5'
Vortex Lattice Method
7.5.1
352 Velocity Induced by a General Horseshoe Vortex 356 7.5.2 Application of the Boundary Conditions 360 7.5.3 Relations for a Planar Wing 362
7.6
7.7 7.8 7.9
Factors Affecting Drag DuetoLift at Subsonic Speeds 374 Delta Wings 377 LeadingEdge Extensions 387 Asymmetric Loads on the Fuselage at High Angles of Attack 391
7.9.1 7.9.2
Asymmetric Vortex Shedding 392 Wakelike Flows 394
7.10
7.11 7.12
Flow Fields For Aircraft at High Angles of Attack 394 Unmanned Air Vehicle Wings 396 Summary 398 Problems 399 References 400
CHAPTER 8
DYNAMICS OF A COMPRESSIBLE FLOW FIELD
8.1
404
Thermodynamic Concepts
8.1.1 8.1.2 &1.3
405 Specific Heats 405 Additional Relations 407 Second Law of Thermodynamics and Reversibility 408 8.1.4 Speed of Sound 410
8.2
Adiabatic Flow in a VariableArea Streamtube 412
Contents
8.3
8.4
8.5 8.6 8.7
Isentropic Flow in a VariableArea Streamtube 417 Characteristic Equations and PrandtlMeyer Flows 422 Shock Waves 430 Viscous Boundary Layer 440
8.6.1
Effects of Compressibility 442
The Role of Experiments for Generating Information Defining the Flow Field 446
8.7.1 8.7.2
GroundBased Tests 446 Flight Tests 450
8.8
8.9
Comments About The Scaling/Correction Process(Es) For Relatively Clean Cruise Configurations 455 ShockWave/BoundaryLayer Interactions
455
Problems 457 References 464
CHAPTER 9
COMPRESSIBLE, SUBSONIC FLOWS AND TRANSONIC FLOWS
9.1
467
compressible, Subsonic Flow 468 .1.1 Linearized Theory for Compressible Subsonic
Flow About a Thin Wing at Relatively Small Angles of Attack 468
9.2 9.3
Transonic Flow Past Unswept Airfoils 473 Wave Drag Reduction by Design 482
9.3.1 9.3.2
Airfoil Contour Wave Drag Approaches 482 Supercritical Airfoil Sections 482
9.4
Swept Wings at Transonic Speeds
9.4.1
484
Wing—Body Interactions and the "Area Rule" 486 494 9.4.2 SecondOrder AreaRule 9.4.3 Forward Swept Wing 497
9.5 9.6
Transonic Aircraft Summary 503
500
Problems 503 References 504
CHAPTER 10 TWODIMENSIONAL, SUPERSONIC FLOWS AROUND THIN AIRFOILS
10.1
507
Linear Theory
10.1.1 10.1.2 10.1.3
508 Lift 509 Drag 512 Pitching Moment 513
7.9.3 SecondOrder Theory (Busemann's Theory) ShockExpansion Technique 519 516 Problems 524 References 527 CHAPTER 11 SUPERSONIC FLOWS OVER WINGS AND AIRPLANE CONFIGURATIONS 11.7 SingularityDistribution Method 11.5 11.2 The HighSpeed civil Transport (HSCT) 580 11.1 11.12 Slender Body Theory 584 Aerodynamic Interaction 587 Aerodynamic Analysis for Complete Configurations in a Supersonic Stream Problems 591 References 593 590 CHAPTER 12 HYPERSONIC FLOWS 12.7.3 Delta and Arrow Wings 546 11.1 11.4 Classifying HighSpeed Aircraft Designs 582 1L9.2 548 Find the Pressure Distribution Given the configuration sso Numerical Method for Calculating the Pressure Distribution Given the Configuration 559 11.9 Design Considerations for Supersonic Aircraft 575 SomeComments About the Design of the SST and of the HSCT 579 The Supersonic Transport the Concorde 579 11.10 11.6.2 11.9.8 11.4 11.6.7.11 11.2 10.2 11.2 Newtonian Flow Model 597 Stagnation Region FlowField Properties .1 596 600 12.1 528 11.Contents 10.6.9.3 Numerical Method for the Determination of Camber Distribution 572 11.1 11.3 Reducing the Sonic Boom 580 11.3 11.6 General Remarks About Lift and Drag 530 General Remarks About Supersonic Wings 531 Governing Equation and Boundary Conditions 533 Consequences of Linearity 535 Solution Methods 535 ConicalFlow Method 536 Rectangular Wings 537 Swept Wings 542 11.
8 A Hypersonic Cruiser for the TwentyFirst Century? 634 Importance of Interrelating CFD.4 13.3 13.5.1 621 Aerodynamic Heating 628 Similarity Solutions for Heat Transfer 632 12.Contents 12.4 Additional Comments 694 13.4.4.1 13. Wing/Tail.5 12.4 Variable.1. and Tailless Configurations 694 Comments on the FiS Design 699 The Design of the F22 700 The Design of the F35 703 Problems 706 References 708 .1.4.5 Development of an Airframe Modification to Improve the Mission Effectiveness of an Existing Airplane 678 The EA6B 678 The Evolution of the F16 681 13.3 13.3 13.3 External Carriage of Stores 689 13.6 13.5.7 13.6 12.4.5 649 Increasing the Area 650 Increasing the Lift Coefficient 651 Flap Systems 652 Multielement Airfoils 656 PowerAugmented Lift 659 13.1.2 13.1.1 13.4 12.Twist.4 Circulation Control Wing 663 Design Considerations For Tactical Military Aircraft 664 Drag Reduction 669 13.1 649 HighLift Configurations 13.2 13.1 13. GroundTest Data.9 Considerations for Wing/Canard.1.12 13.2 13.3 12.5. VariableCamber Wings 669 LaminarFlow Control 670 Wingtip Devices 673 Wing Planform 676 13. and FlightTest Data 638 BoundaryLayer Transition Methodology 640 Problems 644 References 646 CHAPTER 13 AERODYNAMIC DESIGN CONSIDERATIONS 13.8 13.5.7 Modified Newtonian Flow 605 High LID Hypersonic Configurations — Waveriders 12.
1.1.6 Integrating the Diverse CFD Tools 717 14.5 Computational Techniques that Treat the Entire Flow Field in a Unified Fashion 716 14.5 Establishing the Credibility of CFD Simulations 718 GroundBased Test Programs 720 FlightTest Programs 723 Integration of Experimental and Computational Tools: The Aerodynamic Design Philosophy 724 References APPENDIX A APPENDIX B 725 THE EQUATIONS OF MOTION WRITTEN IN CONSERVATION FORM A COLLECTION OF OFTEN USED TABLES 728 734 742 INDEX .3 14.3 Euler Codes for Inviscid Flow Fields 715 14.4 14.1 711 CFD Tools 713 14.4 TwoLayerFlowModels 715 14.2 14.1.1.1.1 Semiempirical Methods 713 14.1.2 Surface Panel Methods for Inviscid Flows 714 14.Contents 13 CHAPTER 14 TOOLS FOR DEFINING THE AEROD YNA MI C ENVIRONMENT 14.
.
Air Force Academy). we believe that the goal of aerodynamics is to predict the forces and moments that act on an airplane in flight in order to better understand the resulting performance benefits of various design choices. enjoyable. slenderbody theory and sonic boom reduction '(Chapter 11). now included in Chapter 13. New problems have been added to almost every chapter. To help achieve the first goal we provided readers with background for the true purpose of aerodynamics. hypersonid transition (Chapter 12). and wingtip devices. and the impactof boundarylayer transition. as well as numerous. Of special interest in the fifth edition is a description of the aerodynamic design of the F35. readable. This new introduction should help to make it clear to students and engineers alike that understanding aerodynamics is crucial to understanding how an airplane performs. micro UAV and high altitude/lông endurance wing geometries (Chapter 7). These innovations include detailed discussion about laminar flow and low Reynolds number airfoils. and 2) to update the technical innovations and advancements that have taken place in aerodynamics since the writing of the previous edition. as well as example problems showing students how the theoretical concepts can be applied to practical problems. including the impact of scaling data for fullscale aircraft (Chapter 8). as well as modern wing planforms (Chapter 13). the role of experimentation in determining aerodynamics. form' factor.S. Throughout the remainder of the fifth edition we have added new and emerging aircraft technologies that relate to aerodynamics. roughness effects. and readable manner. In addition. additional uptodate references throughout the book. Namely. In order to better accomplish this. and motivational presentation on aircraft performance using material on Specific Excess Power (a topic which is taught to all cadets at the U. Significant new material on practical methods for estimating aircraft drag have also been incorporated into Chapters 4 and 5. Users of the fourth edition of the book will find that all material included in that edition is still included in the 15 . there are 32 new figures containing updated and new information. including methods for' estimating skin friction. and why one airplane may 'be better than another at a specific task. as well as modern highlift systems (Chapter 6). Chapter 1 begins with a fun.Preface to the Fifth Edition There were two main goals for writing the Fifth Edition of Aerodynamics for Engineers: 1) to provide readers with a motivation for studying aerodynamics in a more casual.
Finally. We hope that readers will find the incluslon of all of this additional material helpful and informative.S.S. Mark Drela. Hui Hu of Iowa State University. and their associates at Boeing. Cummings of the U. David Graham. Charles Boccadoro. and their associates from Lockheed Martin. . Finally. Air Force Academy. John McMasters. University of California. Michael Seig. Massachusetts Institute of Technology. Mark Buchholz. We want to especially thank Doug McLean. University of Illinois. and their associates of Northrop Grumman. with the new material added throughout the book to bring a realworld flavor to the concepts being developed. In addition. Based on his significant contributions to both the writing and presentation of new and updated material. he makes a welcome addition to the quality and usefulness of the book. and Gabriel Karpouzian of the U. has been added for the fifth edition of Aerodynamics for Engineers. In order to help accomplish these goals a new coauthor. no major revision of a book like Aerodynamics for Engineers can take place without the help of many people. Davis.Preface to the Fifth Edition fifth edition.S. Rick Baker. Professor Russell M. The authors are especially indebted to everyone who aided in collecting new materials for the fifth edition. Air Force Academy for her unfailing support throughout this project. and Case van Darn. Naval Academy. we also want to thank Shirley Orlofsky of the U. we are very grateful for the excellent suggestions and comments made by the reviewers of the fifth edition: Doyle Knight of Rutgers University.
respectively. Fluid properties and a model for the standard atmosphere are discussed in Chapter 1. "Dynamics of an Incompressible. "Dynamics of a Compressible Flow Field. the basic fluid mechanic principles for compressible flows are discussed in Chapter 8." Thus." Thus. subsonic flows 17 . "Fluid Properties. For the reader who already has had a course (or courses) in fluid mechanics. the linear momentum equation (based on Newton's law of motion). the reader is ready to begin material focused on aerodynamic applications. incompressible flows is the subject of Chapter 3. Chapters 1 through4 present material that covers the principles upon which the aerodynamic applications are based." and Chapter 7 is titled "Incompressible Flow about Wings of Finite Span. is the subject of Chapter 4. Basic fluid mechanic principles are presented in the first four chapters. Modeling inviscid. Techniques for modeling highspeed flows (where density variations cannot be neglected) are presented in Chapters 9 throughl2. Parameters that characterize the geometry of aerodynamic configurations and parameters that characterize aerodynamic performance are presented in Chapter 5." Modeling viscous boundary layers. At this point." Differential and integral forms of the continuity equation (based on the conservation of mass). the material presented in Chapter 8 complements the material presented in Chapters 1 through4. Chapter 6 is titled "Incompressible Flows around Wings of Infinite Span. from a pedagogical point of view. To provide the reader with the necessary background for highspeed aerodynamics. these four chapters provide a comprehensive review of fluid mechanics and an introduction to the nomenclature and style of the present text. "Viscous Boundary Layers.Preface to the Fourth Edition This text is designed for use by undergraduate students in intermediate and advanced classes in aerodynamics and by graduate students in mechanical engineering and aerospace engineering." The equations governing fluid motion are presented in Chapter 2. Inviscid Flow Field. "Fundamentals of Fluid Mechanics." Techniques for modeling the aerodynamic performance of twodimensional airfoils and of finitespan wings at low speeds (where variations in density are negligible) are presented in Chapters 6 and7. Aerodynamic performance for compressible." The next five chapters deal with compressible flow fields. and the energy equation (based on the first law of thermodynamics) are presented. with emphasis on incompressible flows. "Characteristic Parameters for Airfoil and Wing Aerodynamics.
The applications of the theory are illustrated by working one or more problems. should be useful to professionals long after they have completed their academic training. and other practical aspects of contemporary aerodynamic design are also presented. equivalently. One objective is to use the experimental data to determine the limits of applicability for the proposed models.Preface to the Fourth Edition through transonic speeds is the subject of Chapter 9."Supersonic Flows over Wings and Airplane Configurations. chapters have been dedicated to the development of basic models for calculating the aerodynamic performance parameters for each of the possible speed ranges. Smith for his significant contributions to Aerodynamics for Engineers. Furthermore. Because both the International System of Units (Système International d'Unitès. "TwoDimensional Supersonic Flows about Thin Airfoils" and for finitespan wings in Chapter 11. he was recognized as coauthor of the first three editions. For these contributions. and for . shock/boundarylayer interactions. At this point. both are used in this text. Problems at the end of each chapter are designed to complement the material presented within the chapter and to develop the student's understanding of the relative importance of various phenomena. therefore. tions. In each of the the computed aerodynamic parameters are compared with experimental data from the open literature to illustrate both the validity of the theoretical analysis and its limitations (or. The author is indebted to his many friends and colleagues for their help in preparing the first three editions of this text. The text emphasizes practical problems and the techniques through which solutions to these problems can be obtained. "Compressible Subsonic Flows and Transonic Flows. Michael Smith's contributions helped establish the quality of the text from the outset and the foundation upon which the subsequent editions have been based. Chapter 13 is titled "Aerodynamic Design Considerations. compressibility." and Chapter 14 is titled "Tools for Defining the Aerodynamic Environment." Hypersonic flowsare the subject of Chapter 12. I thank for their suggestions. The assumptions and. Advanced material relating to design features of aircraft over more than a century and to the tools used to define the aerodynamic parameters are presented in Chapters 13 andl4. their support. COMMENTS ON THE FIRST THREE EDITIONS The author would like to thank Michael L." Chapter 14 presents an explanation of the complementary role of experiment and of computaadvantages. and roles of computational techniques of varying degrees of rigor are discussed." Supersonic aerodynamics for twodimensional airfoils is the subject of Chapter 10. turbulence modeling. The material presented in Chapters 13 andl4 not only should provide interesting reading for the student but. limitation in defining the aerodynamic environment. Extensive discussions of the effects of viscosity. Conversion factors between SI units and English units are presented on the inside covers. the range of conditions for which the theory is applicable). the restrictions incorporated into the development of the theory are carefully noted. Solutions are obtained using numerical techniques in order to apply the theory for those flows where closedform solutions are impractical or impossible. abbreviated SI) and English units are commonly used in the aerospace industry.
Forsythe. and W H. A. J. Since it was the desire of the author to reflect the current role of computations (limitations. B. Please contact the author at USAFA. R. S. Lemmerman and A. Kanipe of the Johnson Space Center. Capt. THE FOURTH EDITION Rapid advances in software and hardware have resulted in the everincreasing use of computational fluid dynamics (CFD) in the design of aerospace vehicles. Towne. and as suppliers of graphic art. Squire of Cambridge University. Towne were valuable contributors to the changes that have been made to the fourth edition. D. Nipper of the Lockheed Martin Tactical Aircraft Systems. 1. V. Ericsson of Lockheed Missiles and Space. J. G. L. Hillaker (formerly) of General Dynamics. F. W. and J. as editors to the modified text. Spearman. E. Capt. Thomas of Lockheed Georgia. Valdez served as the graphics artist for the first three editions of this text. M. Col. M. The author is indebted to L. 3. . Maydew. E Covell of the Langley Research Center.S. H. Chapter 14 was a major effort. who are active experts in theuse and in the development of CFD in aerodynamic design. J. 2. L. Erickson of Northrop. Marchman III. C. The increased reliance on computational methods has led to three changes unique to the fourth edition. using the postprocessing packages Fieldview and EnSight. transonic flows around an NACA airfoil (Chapter 9).They served as sounding boards before the text was written. Smith of General Dynamics. the author appreciates the many contributions of Maj. J. Lt. Captain J. Forsythe was instrumental in obtaining the appropriate graphics. and flow over the SR71 at three highspeed Mach numbers (Chapter 11) appear for the first time in Aerodynamics for Engineers. Chase of the ANSER Corporation. Not only has T. Towne of the U. R. S. M. G. T. Although these results have appeared in the open literature. intended to put in perspective the strengths and limitations of the various tools that were discussed individually throughout the text. Carison. C. and M. Bradley and C. Major D. Rutledge of the Sandia National Labs. C. but he has regularly located interesting articles on aircraft design that have been incorporated into the various editions. A CD with complementary homework problems and animated graphics is available to adopters. Col. A. McAlees. L. Szebehely of the University of Texas at Austin. McClure. E. and P. W. R. DeJarnette of North Carolina State University. and reference documents. Some very sophisticated numerical solutions for high alpha flow fields (Chapter 7). H. F. R. Periaux of Avions Marcel Dassault. Forsythe. illustrations. and usage) and to present some challenging applications. and Maj. Wierum of the Rice University. Mueller of the University of Notre Dame. Air Force Academy. R. F. 0. Blake. J. LLC. Brandt. C. and Lt. and Dr. of Virginia Polytechnic Institute and State University provided valuable comments as reviewers of the third edition. R.Preface to the Fourth Edition copies of photographs. strengths. Blake. C. The discussion of the complementary use of experiment and computation as tools for defining the aerodynamic environment was the greatest single change to the text. W. the highquality figures were provided by Cobalt Solutions. C. S.
and the Lockheed Martin Tactical Aircraft System for allowing the author to reproduce significant amounts of archival material. Report. and Leland A. Bracken for supplying in formation and photographs regarding the design and operation of military aircraft. E. BERTIN United States Air Force Academy 1The AGARD/NATO material was first published in the following publications: Conf Proc. the Advisory Group for Aerospace Research and Development. CP515. Figliola of Clemson University. Gen. CV437. The author would like to thank John Evans Burkhalter of Auburn University. Carlson of Texas A & M University. North Atlantic Treaty Organization (AGARD/NATO). G. Sept. Peters of the Boeing Company and M. who. June 1989.1. High Lift System Aerodynamics. R. vol. but it also serves as an excellent foundation upon which the reader can explore new topics.1 the Boeing Company. Vèilidation of ('omputational Fluid Dynamics.1988. Marilyn Smith of the Georgia Institute of Technology. Dec. C. Fmally. provided comments that have been incorporated either into the text or into the corresponding CD. The author would also like to thank the American Institute of Aeronautics and Astronautics (AJAA). . R761. 1993. as reviewers of a draft manuscript. thank you Margaret Baker and Shirley Orlofsky. Towne of Lockheed Martin Aeronautics served as points of contact with their companies in providing material new to the fourth edition. This material not only constitutes a critical part of the fourth edition. E. Special Course on Aerothermodynamics of Hypersonic Vehicles. Richard S. Gonf Proc. JOHN J.20 Preface to the Fourth Edition The author would also like to thank M.
1 WHY STUDY AERODYNAMICS? 1. trading potential energy (height) for kinetic energy (velocity). Starting from a superior altitude and converting potential energy to kinetic energy. German aces.. as the successful fighter aces gained a better understanding of the nuances of air combat by building an empirical data base through successful airtoair battles.e. they could dive upon their foe. the sum of the potential energy plus the kinetic energy (i.e. the attacker might be able to destroy his opponent on the first pass. Clearly. realized that. fighter pilots (at least those good enough to survive their first engagements with the enemy) quickly developed tactics that were to serve them throughout the years. These tactics were refined. the pilot of the attacking aircraft could dictate the conditions of the initial phase of the airtoair combat. such as Oswald Boelcke and Max Immelman. A language grew up to codify these tactics: "Check your six. Using the greater speed of his airplane to close from the rear (i.1 THE ENERGYMANEUVERABILITY TECHNIQUE Early in the First World War.. if they initiated combat starting from an altitude that was greater than that of their adversary." This data base of tactics learned from successful combat provided an empirical understanding of factors that are important to aerial combat. the total energy) of the aircraft is one of the factors. 21 . from the target aircraft's "six o'clock position").
kilograms.s2) 0. The equation for the total energy is E = 0. or s2/ft.986 lbf s2/ft or 13..000 E= 2.5 (373 lbf s2) (422. '1 / Why Study Aerodynamics? EXAMPLE Li: The total energy Compare the total energy of a B52 that weighs 450. Even though the total energy of the B52 is so very much greater than that for the F5.000ft) 1.000 lbf) (20. Thus.986 slugs.7329 x 108 ft lbf The total energy of the B52 is 37. it just doesn't seem likely that a B52 would have a significant advantage in airtoair combat with an FS.000 lbf) (20.5 + (450.s2/ft.. A knot is a nautical mile per hour and is equal to 1. Since the mass is given by the equation. the total energy for the B52 is E= E= (139861bf. the B52) is 13. The choice of units often will reflect how mass appears in the application.69 feet per second.5 ft/s.000 feet. the units for velocity should be feet per second rather than knots. slugs. Perhaps the specific energy (i.5mV2 + mgh Solution: To have consistent units. Thus.000 pounds and that is cruising at a true air speed of 250 knots at an altitude of 20.e.22 Chap. . the total energy of the FS fighter is = 0.000 pounds and that is cruising at a true air speed of 250 knots at an altitude of 20.5 times the total energy of the FS.000 feet with the total energy of an FS that weighs 12. the energy per unit weight) is a more realistic parameter when trying to predict which aircraft would have an edge in airtoair combat. 250 knots is equal to 422. ibm. Note that the two aircraft are cruising at the same ifight condition (velocity/altitude combination). Thus.5 + (12.0248 X 1010 ft ibf Similarly. w g Note that the units of mass could be grams. while that for the FS is 373 lbf.e. the difference in total energy is in direct proportion to the difference in the weights of the two aircraft. The mass of the "Buff" (i.
This puts the glare of the sun in the enemy's eyes and makes it difficult to see you and impossible to shoot with any accuracy. or "dicta" [Werner (2005)]. and dive on him swiftly from the rear when the moment to attack is at hand. If we consider only this weight specific energy. it will be given the Since and is called the energy height. Li I The EnergyManeuverability Technique 23 EXAMPLE 1. The first five. Dividing the terms in equation symbol (1. 2. Do not fire the machine guns until the enemy is within range and you have him squarely within your sights.Sec. 1916. 3. However. Solution: The energy height of the B52 is (422. Thus. Captain Oswald Boelcke developed a series of rules: based on his combat experience as a fortyvictory ace by October 19.. Climb before and during the approach in order to surprise the enem. the fact that the energy height is the same for these two aircraft indicates that further effort is needed to provide a more realistic comparison for airtoair combat. the same weight specific energy).y from above. are 1. .e. which deal with tactics. the 1852 and the FS are equivalent.1) by the weight of the aircraft (W in g).5 + 20000 ft He = 22774 ft Since the F5 is cruising at the same altitude and at the same true air speed as the B52. jt has the same energy height (i. Try to place yourself between the sun and the enemy. This is obviously an improvement over the factor of 37. E W v2 2g Compare the energy height of a B52 flying at 250 knots at an altitude of 20. there must be some additional parameters that are relevant when comparing the oneonone capabilities of two aircraft in airtoair combat.when the comparison was made based on the total energy.000 feet with that of an F5 cruising at the same altitude and at the same velocity.5 that the "Buff" had over the F5. Boelcke specified seven rules.5 He = 0.2: The energy height weight specific energy also has units of height. Always try to secure an advantageous position before attacking.
photography. How fast can we. In his role as Director of Academics at the U.4) Multiplying both sides of the Equation (1.S.24 Chap. Air Force brought the needed complement of talents to bear on the problem [Coram (2002)]. 5.4) by V and dividing by W gives (T—D)VVdV W gdt 15 (. Air combat is a dynamic ballet of move and countermove that occurs over a continuum of time. Consider the case where the F5 is flying at a constant altitude.) . Boyd postulated that perhaps the time derivatives of the energy height are more relevant than the energy height itself. 1. it became not only his passion. such as observation. John R.5 ft/s) at an altitude of 20. Now he became obsessed with the challenge of developing the science of fighter tactics. in the target aircraft. they were experiencedbased empirical rules.S. Attack when the enemy least expects it or when he is preoccupied with other duties. Although Boelcke's dicta were to guide fighter pilots for decades to come. If the engine is capable of generating more thrust than the drag acting on the aircraft. 1 / Why Study Aerodynamics? 4.1 Specific Excess Power and If the pilot of the 12. the acceleration of the aircraft can be calculated using Newton's Law: F ru a which for an aircraft accelerating at a constant altitude becomes g dt (1. If you are surprised by an attack on your tail.1. had been learned by experience for the fighter pilot lucky enough to survive his early airtoair encounters with an enemy. Note that the fifth dictum deals with maneuverability. Boyd was driven to understand the physics that was the foundation of the tactics that. Never turn your back and try to run away from an enemy fighter. Boyd taught these tactics in the Fighter Weapons School. Thus. turn and face the enemy with your guns.000 lbf F5 that is flying at a velocity of 250 knots (422. or bombing. Captain John R. We learned from the first two example calculations that predicting the probable victor in oneonone airtoair combat is not based on energy alone. Air Force Fighter Weapons School. how quickly can we increase our energy height and take the offensive? John R. Boyd was an aggressive and talented fighter pilot who had an insatiable intellectual curiosity for understanding the scientific equations that had to be the basis of the "Boelcke dicta". The first dictum deals with your total energy. with an enemy on our six. the sum of the potential energy plus the kinetic energy. It wasn't until almost half a century later that a Captain in the U. Energy AND Maneuverability! The governing equations should include maneuverability as well as the specific energy. his aircraft must have sufficient power either to out accelerate or to out climb his adversary. but his job.000 feet is to gain the upper hand in airtoair combat. until that time. quickly dump energy and allow the foe to pass? Once the enemy has passed.
Thus. change of the potential energy (per unit weight).3). It is the rate of is . one could use equation (1.5 ft/s = W 12000 lbf = 63.83 ft/s 1.000lbf FS as it passes through 20.5 ft/s = 4. while the total drag (D) acting on the aircraft is 1750 lbf.Sec.3.174 ft/s2 422.000 feet.3: The specific excess power and acceleration The lefthand side of equation (1. = = 63.5) and (1.5) to calculate the maximum acceleration for a 12.6) leads us to the conclusion that the specific excess power is equal to the timerateofchange of the energy height.6) represents the rate of change of kinetic energy (per unit weight).1.000 feet at 250 knots.5) is excess power per unit weight.1 / The EnergyManeuverability Technique 25 EXAMPLE 1.e. 1.38 ft/s Rearranging Equation (1.7) to calculate the maximum rateofclimb (for a constant velocity) for the 12. It is a function of the rate of change of the velocity as seen by the pilot (dV'\ The significance of the second tenn is even less cosmic.5 ft/s) at 20.8 ft/mm . the specific excess power (Ps) is (T — D) V [(3550 — 1750) lbf] 422.5) to solve for the acceleration gives dV = = (63. Note also that the vertical com ponent of the velocity [i. one obtains dHeVdV dt — g di' dh + dt The first term on the righthand side of equation (1.000lbf F5 that is flying at 250 knots (422. the rate of climb (ROC)] as seen by the pilot on his altimeter.Thus.2 Using Specific Excess Power to Change the Energy Height Taking the derivative with respect to time of the two terms in equation (1..38 ft/s = 3802. Use equation (1. Performance charts for an FS that is flying at these conditions indicate that it is capable of generating 3550 lbf thrust (I) with the afterburner lit. Air speed and altitude — these are parameters that fighter pilots can take to heart. or specific excess power.38 ft/s) 32. Ps= (T—D)V W VdV dh dHe =—=——+— dt gdt di' Given the specific excess power calculated in Example 1. Combining the logic that led us to equations (1.
the airplane under a given power setting would either slow down. Boyd. observed. Air Force inventory and on their adversaries.e. "The original F16 design had about onethird the drag of an F4 in level flight and onefifteenth the drag of an F4 at a high angleofattack.' Their basic belief was that fighters did not need to overwhelm opponents with speed and size. one for Which the thrust exceeds the drag). "Boyd.." [Hillaker (1997)] In the mid 1960s. he conclUded that maneuvering for position was basically an energy problem." [Grier (2004)] The fighter mafia knew that a small aircraft could enjoy a high thrusttoweight ratio." [Grier (2004)] Hillaker countered that the Fill was designated a fighterbomber. He sought to understand the intricacies of maneuvering flight. Boyd and Harry Hillaker "dated from an evening in the mid1960s when a General Dynamics engineer named Harry Hillaker was sitting in the Officer's Club at Eglin AFB." "He knew that. He also noted that.3 John R.7). Hillaker was annoyed but bantered back. 1 I Why Study Aerodynamics? Clearly.. or both. blustery pilot named John R. as a combat pilot in Korea and as a tactics instructor at Nellis AFB in the Nevada desert.1. his energy was higher than that of his opponentand that he lost that advantage when he allowed his energy to decay to less than that of his opponent.. the greater the bene fits of the available excess power. he (Hillaker) received a call—Boyd had been impressed by Hillaker's grasp of aircraft conceptual design and wanted to know if Hillaker was interested in more organized meetings. and assimilated the relative energy states of his aircraft and those of his opponent's during air combat engageinents. when turning from a steady—state flight condition. Boyd had gathered energymaneuverability data on all of the fighter aircraft in the U." 1. to be able to generate positive values for the terms in equation (1. Hillaker's host introduced him to a tall. In order to predict these aerodynamic forces and moments with suitable accuracy. Experience in Vietnam against nimble Sovietbuilt MiGs had convinced them that technology had not yet turned airtoair combat into a longrange shootout. From these observations. The result meant he was losing energy (the drag exceeded the thrust available from the engine). analyzed.26 Chap. Florida. Winning required the proper management of energy available at the conditions existing at any point during a combat engagement. "A few days later. we need an aircraft with excess power (i. when in a position of advantage.2 SOLVING FOR THE AEROTHERMODYNAMIC PARAMETERS A fundamental problem facing the aerodynamicist is to predict the aerodynamic forces and moments and the heattransfer rates acting on a vehicle in flight. it is necessary to be . since the lighter the aircraft. having an after dinner drink.S. What was it about the airplane that would limit or prevent him from making it do what he wanted it to do? 1." "Thus was born a group that others in the Air Force dubbed the 'fightcr mafia. Boyd Meet Harry Hillaker The relation between John R. lose altitude. who immediately launched a frontal attack on GD's Fill fighter. Small aircraft have less drag. Weight is another important factor.
The resultant flow pattern de pends on the geometry of the vehicle. Therefore. In addition. stress. in analyzing highspeed flows. A gas has no definite volume. one can either define the motion of each and every molecule or one can define the average behavior of the molecules within a given elemental volume. Gas molecules are widely spaced with relatively small cohesive forces.2. the number of . Thus. the temperature variations are so small that they do not affect the velocity field. In fact.1 Concept of a Fluid From the point of view of fluid mechanics. The size of the elemental volume is important. Thus. Furthermore. A fluid is a substance that deforms continuously under the action of shearing forces. if a gas is placed in a closed container. any thermodynamic property may be expressed as a function of two other independent. assumptions aboutthe fluid properties may be introduced. but only in relation to the number of fluid particles contained in the volume and to the physical dimensions of the flow field. if the fluid particles are at rest or if they are all moving at the same velocity. it is often assumed that the density is essentially constant. its orientation with respect to the undisturbed free stream. A liquid is composed of relatively closely packed molecules with strong cohesive forces. for a gas in thermodynamic equilibrium. However. it is possible to formulate the governing equations using the enthalpy and the entropy as the flow properties instead of the pressure and the temperature. As a result. The technical distinction between these two states lies in their response to an applied shear. the density variations cannot be neglected. In some applications.2 Fluid as a Continuum When developing equations to describe the motion of a system of fluid particles. Thus. This zero shear stress condition is known as the hydrostatic stress condition. 1. that is. thermodynamic properties. if it is unconfined. the elemental volume should be large compared with the volume occupied by a single molecule so that it contains a large number of molecules at any instant of time. If a liquid is poured into a container. or tangential. An important corollary of this definition is that there can be no shear stresses acting on fluid particles if there is no relative motion within the fluid. A fluid can be either a liquid or a gas. it may be expressed in terms of these two parameters. In analyzing the various flows that an aerodynamicist may encounter. 1. Since density is a function of pressure and temperature. either solid or fluid. for those applications where the temperature variations have a negligible effect on the flow field. a fluid cannot. it forms an atmosphere that is essentially hydrostatic. A solid can resist a shear stress by a static deformation. matter can be in one of two states. 1. such fluid particles are not deformed. it assumes the shape of the container up to the volume it occupies and will form a free surface in a gravitational field if unconfined from above. and the altitude and speed at which the vehicle is traveling. there are no shear stresses in the fluid. The upper (or free) surface is planar and perpendicular to the direction of gravity. it will expand until it fills the entire volume of the container. Thus. a given mass of liquid will occupy a definite volume of space.2 / Solving for the Aerothermodynamic Parameters 27 able to describe the pattern of flow around the vehicle.Sec.2.
The number of molecules in a cubic meter of air at room temperature and at sealevel pressure is approximately 2. there are only 1. Pressure. therefore. the continuum flow model starts to break down when the Knudsen number is roughly of the order of 0.5 X 1025. Temperature. If the elemental volume is too large.28 Chap. and the speed of sound. There are sufficient molecules in this volume for the fluid to be considered a continuum. the pressure. the pressure . Since a fluid that is at rest cannot sustain tangential forces.2. Two bodies have equality of temperature when no change in any observable property occurs when they are in thermal contact. there are 2. an object feels hot (or cold) to the touch.1. Pressure is the magnitude of this force per unit area of surface. because of the difficulty in quantitatively defining the temperature. we define the equality of temperature. which is the ratio of the mean free path to a characteristic dimension of the body. However. Because of the random motion due to their thermal energy.The fluid in these problems may be considered to be a continuous material whose properties can be determined from a statistical average for the particles in the volume. These collisions occur even though the surface is at rest relative to the fluid. The mean free path at this altitude is 10. that is. contain a large number of molecules. 1. a macroscopic representation.Thus. our primary concern is not with the motion of individual molecules.Thus. but with the general behavior of the fluid. Although there is no definitive criterion. there could be a noticeable variation in the fluid properties determined statistically at various points in the volume. 1 / Why Study Aerodynamics? molecules within the volume will remain essentially constant even though there is a con tinuous flux of molecules through the boundaries.5 X 1010 molecules in a cube 0. However. By Newton's second law. we can describe the gross behavior of the fluid motion using certain properties. the viscosity. Thus. Properties used to describe a general fluid motion include the temperature.01 mm on a side. In problems of interest to this text. we are concerned with describing the fluid motion in spaces that are very large compared to molecular dimensions and that. It follows that an arbitrary scale of temperature can be defined in terms of a convenient property of a standard body. the individual molecules of a fluid would continually strike a surface that is placed in the fluid. The mean free path at sea level is 6. the density. that is. a force is exerted on the surface equal to the time rate of change of the momentum of the rebounding molecules. at this altitude the fluid cannot be considered a continuum.6 x molecules in a cube 1 m on a side. The assumption of a continuous fluid is valid when the smallest volume of fluid that is of interest contains so many molecules that statistical averages are meaningful.3 Fluid Properties By employing the concept of a continuum. Further. two bodies respectively equal in temperature to a third body must be equal in temperature to each other. at an altitude of 130 km.6 x m. A parameter that is commonly used to identify the onset of lowdensity effects is the Knudsen number. We are all familiar with temperature in qualitative terms. and the fluid properties can be determined from statistical averages.2 m.
Furthermore. The density of a fluid at a point in space is the mass of the fluid per unit vol ume surrounding the point.1 Terms used in pressure measurements. The standard atmospheric pressure at sea level is 1. In general. the incremental volume must be large compared to molecular dimensions yet very small relative to the dimensions of the vehicle whose flow field we seek to analyze.595 1 g/cm3 and the acceleration due to gravity is the standard value. the density at a point is defined as p lim 8(mass) (1. In English units. Thus. Standard atmospheric pressure at sea level is defined as the pressure that can support a column of mercury 760 mm in length when the density of the mercury is 13. its temperature. the standard atmospheric pressure at sea level is 14.2 / Solving for the Aerothermodynamic Parameters 29 Positive gage pressure Atmospheric pressure Negative gage pressure Absolute pressure is greater than the atmospheric pressure Absolute pressure is less than the atmospheric pressure Zero pressure Figure 1. p) (1. The relation p(composition.8) The dimensions of density are (mass)/(length)3.9) . which is referred to as gage pressure. 1.22 lb/ft2. we are interested in the difference between the absolute value of the local pressure and the atmospheric pressure. the density of a gas is a function of the composition of the gas. and its pressure. T.Sec. is illustrated in Fig. Density. Many pressure gages indicate the difference between the absolute pressure and the atmospheric pressure existing at the gage. provided that the fluid may be assumed to be a continuum.01325 x N/rn2. the pressure acting at a point in a fluid at rest is the same in all directions. 1.696 lbf/in2 or 2116. This difference.1. on the surface must act in the direction perpendicular to that surface. As is the case when evaluating the other fluid properties. In many aerodynamic applications.
The gas constant for air has the value 287.16 ft2 s2°RJ \. air will be assumed to behave as a perfect gas unless specifically stated otherwise. Since air at this pressure and temperature behaves as a perfect gas.30 is Chap.22— / I 1716.67°R.15 K. °R in English units. One slug is the equivalent of 32.174 Ibm.22— ft2 = ft3 Alternatively. EXAMPLE 14: Density in SI units Calculate the density of air when the pressure is 1. but never in °C or in °F. 1 I Why Study Aerodynamics? known as an equation of state. The temperature in equation (1. 2116.34 ft lbf/lbm °R or 1716.10) should be in absolute units.22 lbf/ft2 and the temperature is 518. Thus. .Note that throughout the remainder of this book.16 ft2/sR .01325 X N/rn2 (287.2250 kg/rn3 EXAMPLE 15: Density in English units Calculate the density of air when the pressure is 2116. Solution: 1. the equation of state is P = (1. we can use equation (1. Solution: 2116. we can use equation (1.10). the temperature is either in K or in °R. Since air at this pressure and temperature behaves as a perfect gas.05 N rn/kg K in SI units and 53.15 K) = 1.10) The gas constant R has a particular value for each substance. I(51867°R) ft2 = ft4 2 The unit lbf s2/ft4 is often written as slugs/ft3.01325 x i05 N/rn2 and the temperature is 288. where slugs are alternative units of mass in the English systern.05 N rn/kg.10). For a thermally perfect gas. K)(288.
11) There are many problems of interest to us in which the effects of viscosity can be neglected. In general. In all real fluids. Thus. the assumption of constant density for velocities below 100 rn/s is a valid approximation because the pressure changes that occur from one point to another in the flow field are small relative to the absolute value of the pressure. We shall use the term inviscid flow in these cases to emphasize the fact that it is the character both of the flow field and of the fluid which allows us to neglect viscous effects. The fluids of interest in this text are Newtonian in nature. In this temperature range. In such problems.4 = 1. the shearing stress is proportional to the rate of shearing deformation.7894 x i05 kg/s m . For temperatures below 3000 K. its temperature.. or less. the viscosity of air is independent of pressure. No real fluid has a zero coefficient of viscosity.15 + 110. Therefore. the coefficient of viscosity is a function of the composition of the gas.6: Viscosity in SI units Calculate the viscosity of air when the temperature is 288. The viscosity of a fluid relates to the transport of momentum in the direction of the velocity gradient (but opposite in sense). We know that the pressure around the vehicle is not constant. that is.458 x 10—6 288. since the aerodynamic forces and moments in which we are interested are the result of pressure variations associated with the flow pattern.2 / Solving for the Aerothermodynamic Parameters For vehicles that are flying at approximately 100 mIs (330 ftls). Solution: (288.10) would require that the pressure and the temperature remain constant (or change proportionally) in order for the density to remain constant throughout the flow field. (1. Viscosity. The constant of proportionality is called the coefficient of viscosity. viscosity is a transport property.15 K.Sec. the den sity of the air flowing past the vehicle is assumed constant when obtaining a solution for the flow field. and its pressure.15)i5 = 1. 1. However. shear stress X transverse gradient of velocity (1.458 x 106 T ±110. the magnitude of the coefficient of viscosity of the fluid and of the velocity gradients in the flow field are such that their product is negligible relative to the inertia of the fluid particles and to the pressure forces acting on them. we could use Sutherland's equation to calculate the coefficient of viscosity: = 1.4 are kg/s m. Rigorous application of equation (1.12a) Here T is the temperature in K and the units for EXAMPLE 1. a shearing deformation is accompanied by a shearing stress.
These comments are made to emphasize the fact that even the basic fluid properties may involve approximate models that have a limited range of applicability. surrounded by an inversepower attractive force. as compared with that for nonattracting rigid spheres. where T is the temperature in °R and the units for EXAMPLE 1. The chief value of Sutherland's formula seems to be as a simple interpolation formula over restricted ranges of temperature. or to take molecular attractions into account to a first order only.7383 x 518. Chapman and Cowling (1960) note that equations (1.6 ft2 1071bfs Equations used to calculate the coefficient of viscosity depend on the model used to describe the intermolecular forces of the gas molecules. the viscosity of air is independent of pressure.27 x 108 = 3.0°F + 459. This model is qualitatively correct in that the molecules attract one another when they are far apart and exert strong repulsive forces upon one another when they are close together. has been used by Svehla (1962) to calculate the viscosity and the thermal conductivity of gases at high temperatures. Solution: First. In this temperature range. so that it is necessary to define the potential energy of the interaction of the colliding molecules.6 are lbf s/ft2.27 X 108T ±198. The coefficients of viscosity for air as tabulated by Svehla are compared with the values calculated using equation (1.67 = 518. 11 Why Study Aerodynamics? For temperatures below 5400°R.7: Viscosity in English units Calculate the viscosity of air when the temperature is 59.12a) and (1.1.12a) in Table 1. They caution.32 Chap. which takes into account both the softness of the molecules and their mutual attraction at large distances. "In general it is not adequate to represent the core of a molecule as a rigid sphere. ." The LennardJones model for the potential energy of an interaction. convert the temperature to the absolute scale for English units. Sutherland's equation for the viscosity of air in English units is = 2. Svehla (1962) noted that the potential for the Sutherland model is described physically as a rigid. 2. that the success of Sutherland's equation in representing the variation of j. and partly to attractive forces which have more than a firstorder effect.t with temperature for several gases does not establish the validity of Suther land's molecular model for those gases. °R. however.0°F. has to be explained as due partly to the 'softness' of the repulsive field at small distances. 59.67°R. The greater rapidity of the experimental increase of with T.12b) clOsely represent the variation of jt with temperature over a "fairly" wide range of temperatures.67 ± 198. impenetrable sphere.
057 5.13) .1 T (K) (kg/m.918 10.219 11.121 6.973 8.179 6.614 4.132 7.016 3.s)t 1. The aerodynamicist may encounter many applications where the ratio p.Sec./p has been replaced by a single parameter.838 4.152 4.s)* 1.579 9. Because this ratio appears frequently. the kinematic viscosity.695 5.734 8. 1.1 2a) Kinematic Viscosity.435 9.12a)] TABLE 1.657 9.272 (kg/m.992 3.488 8.553 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400 3600 3800 4000 4200 4400 4600 4800 5000 From Svehla (1962) 6.422 7.234 8.171 4.531 11.232 9.145 8.252 10.087 tCalculated using equation (1.828 6.625 5.970 7.329 2. The symbol used to represent the kinematic viscosity is v: 1) = p (1.207 9.360 2.285 3.624 200 400 600 800 1000 1200 1400 2.902 11.829 7.670 6. it has been given a special name.456 5.580 10.702 7.512 6.974 9.874 10.373 7.765 8.516 8.878 9.2 I Solving for the Aerothermodynamic Parameters 33 Comparison of the Coefficient of Viscosity for Air as Tabulated by Svehla (1962) and as Calculated Using Sutherland's Equation [Equation (1.
4 PressUre Variation in a Static Fluid Medium In order to compute the forces and moments or the heattransfer rates acting on a vehicle or to determine the flight path (i. the trajectory) of the vehicle. EXAMPLE 1.e. such as Table 1.4 and the speed of sound is given by a= 20.174 ft lbm/lbf s2. from lbf s/ft2.8: Kinematic Viscosity in English units Using the results of Examples 1. = p.02 VT (L14b) where T is the temperature in °R and the units for the speed of sound are ft/s.7383 p. square meters per second or square feet per second).5.5 and 1. Solution: From Example 1.7383 x ft2 (32. we must employ the factor Which is equal to 32.002377 lbf s2/ft4.. The speed at which a disturbance of infinitesimal proportions propagates through a fluid that is at rest is known as the speed of sound.22 lbf/ft2. equivalently. the engineer will often develop an analytic model of the atmosphere instead of using a table. v has the dimensions of L2/T (e. which is designated in this book as a. Thus. to arrive at the appropriate units. p.g. 0. = 3.573 x 104ft2/s Speed of Sound.573 x 10S 0.002377 ft4 If we use the alternative units fr the density.047 VT (1.2. y = 1. p = Example 1. calculate the kinematic viscosity Of air when the temperature is 518. For the range Of temperature over which air behaves as a perfect gas.7. = 3. The speed of sound is established by the properties of the fluid. 1.34 Chap. For a perfect gas a = where yis the ratio of specific heats (see Chapter 8) and R is the gas constant.14a) where T is the temperature in K and the units for the speed of sound are rn/s. ft2 = 3.. In English units a 49.07649 X lbm/ft3 = 0. the force units (or. while. 174ft lbm p ft3 = 1.2.7.7383 p = 1.67°R and the pressure is 2116. 1 / Why Study Aerodynamks? In this ratio. the mass units) canceL Thus. .
05 22 23 218.81847 0.53 328.9145E—01 1.79485 0.55 316.80945 0.Sec.79447 12 13 1.166 0.72 298.S.38 297.07 295.01325 x 1.9328 E—02 2.05 301.6671 E—02 3.0000E+00 '8.0748 E—01 8.700 236.581 0.150 281 .85343 255.61 53 E—01 0.06 299.43 332.07 295.79447 0.07 2.7340 E—02 7.81247 0.80037 295.9304 E—02 8.T51 1.0470 E—02 28 29 30 Reference values: PSL 0.0117 E—01 4 5 0.252 3.4570 E—02 4.79447 0.527 299.1763 E—01 1.552 222.2 1 Solving for the Aerothermodynamic Parameters TABLE 1.29 336.00000 0.7965 E—02 297.1813E—02 = 1.574 219.0397 E—01 2.53 295. 1976 SI Units Pressure (PIPSL) Geometric Temperature (K) Density (P/PsL) Viscosity Altitude (km) 0 1 Speed of Sound (mis) 340.39 221.560 3.85 223.80340 0.07 295.509 N/m2.0854 E—01 5.5185 E—'Ol 5.59 1.87249 0.3985 E—01 1.536 224.8461 E—O1 288.650 21 6.5898 E—01 1.567 4.9204 E—01 6. rn .5029E—02 288.70 296.150 K (continues on next page) = 13894 x 1 kg/s.3917 E—02 2. 225.650 216.58 324.15 295.3756 E—01 2.2168 E—01 1.81461 303.2403 E—01 216.92836 0.2580 E—02 6.79447 16 17 8.3589 E—01 1.544 223.2722 E—02 2.650 14 15 0.650 9.81 547 24 25 26 27 220.215 229.07 295.676 249.1616 E—01 216.4663 E—02 6.83414 0.80643 0.79447 0.79447 0.9945 E—02 3.79447 0.1808 E—02 5.1953 E—01 1.2661 E—02 295.98237 2 3 275.650 217.4215 E—02 2.8317 E02 3.650 0.5950 E—02 216.518 226.774 216.07 295.651 1.90995 0.3341 E—01 7.96456 0.94656 6.5158 E—02 2.0217 E—01 2.4903 E—02 0.659 262.650 21 6.6362 E—01 1.8553 E—02 1.6885 E—01 6.07 295.82147 0.31 308.0000E+00 9.79732 0.3887E—01 4.82446 1.5464E—01 2.154 268.733 320.89133 0.0567 E—01 3.8128 E—01 3.2A 35 U. 1.71 2.2921 E—01 3.650 21 6.8700 E—01 7.1597 E—02 1.11 6 7 8 9 10 11 4. Standard Atmosphere.4894 E—02 7.81 65 E—01 4.3722E—02 1.9780 E—01 299.07 295.6600E—01 4.39 301.7533E—02 1.3829 E—02 0.2258 kg/rn3.4225 E—01 6.72 300.8601 E—01 1.45 312.187 242.79447 18 19 21 6.79447 0.07 20 21 5.650 21 6.
8524 E—02 7.9447 E—01 968.02 1.9447 [—01 7.2267 E—01 2.25 8.7806 E—01 26 28 30 32 34 36 38 40 3.97 389.6402 [—02 7.8609 E—01 7.8707 [—02 4.08 968.4289 E—02 389.08 968.0461 E—01 9.97 389.9539 [—01 7.08 968.8928E—01 9.67 511.3939 [—01 1.3316 E—01 9. 1976 English Units Pressure (P/PSL) Geometric Temperature (°R) Density (P/PsL) Viscosity (/1'//l'SL) Altitude (kft) 0 2 Speed of Sound (ft/s) 1116.16 392.08 968.6462 E—01 4.9463 E—01 7.8951 E—02 5.30 433.2556 E—01 2.5072 E—02 389.0746 [—01 7.54 504.95 411.9447 E—01 42 44 46 48 50 52 54 56 1.97 7.2452 E—01 2.22 977.08 968.5568 [—01 3.08 968.97 6.3300 E—01 4.0092 [—01 8.0142 E—01 7.1223 [—01 9.6268 E—02 7.31 1011.8767 [—01 5.8407 E—02 7.4278E—01 8.4315 [—01 8.3134 E—01 8. 1 / Why Study Aerodynamics? TABLE 1.2667 E—01 1.18 1085.4795 E—01 3.97 389.98 1093.97 389.9447 E—01 7.6763 [—01 9.9447 E—01 7. Standard Atmosphere.53 968.4708 E—01 2.0921 E—01 5.9447 E—01 7.9447 [—01 7.1366 E—02 6.14 1036.76 1100.8811 E—01 8.66 454.8953 [—01 8.2273 E—01 475.3615 E—O1 10 12 518.0402 E—01 1.4585 E—02 5.8783 E—O1 6.1511 E—01 1.94 1.1246 E—02 6.90 8.5311 E—01 1.64 1020.6650 [—01 8.2498 [—01 2.31 1045.73 968.07 418.4569 [—01 9.51 389.21 969.2981 E—01 4 6 8 8.62 390.97 389.S.1945 E—01 1028.33 1077.8576 E—01 8.3461 E—01 9.3590 E—01 7.5991 E—01 4.85 986.8540 E—01 1.5339 [—01 1.7191 E—01 2.90 468.55 970.97 389.73 397.6880 [—01 1.08 968.18 426.7021 E—01 5.8803 [—01 3.4861 E—02 5.08 968.0000 E+00 9.9970 E—01 4.4736 E—O1 2.9649 E—01 7.41 994.53 447.44 1108.97 390.9754 E—01 4.5487 E—01 8.6848 E—01 1.84 404.97 389.5670 E—01 9.08 968.9447 E—01 7.07 391.0000 E+OO 9.9447 [—01 7.0444 E—01 9.9798 E—01 4.42 440.7473 E—01 3.9447 E—01 79447 E—01 7.97 389.7849E—01 9.97 389.38 1053.42 1061.15 483.4286 E—01 6.78 461.9333 E—01 6.5022 E—01 6.6368E—01 8.8565 E—02 7.2346 E—01 14 16 18 20 22 24 5.08 968.9447 E—01 62 64 66 68 70 4.08 968.0443 E—01 1.0000 E+OO 9.97 389.08 2.08 968.3580 E—02 58 60 7.7151 [—01 2.28 490.4224 [—01 4.40 1069.9883 E—01 2.3859 E—O1 1.41 497.1492 E—01 1.3914 E—01 1.2645 E—01 1.36 Chap.9447 [—01 7.08 968.88 1003.0305 E—01 3.9835 E—01 (continues on next page) .4919 E—02 8.9447 E—01 7.97 389.2B U.
28 985.2407E—01 8.7048E—02 2. the term static fluid properties may be applied to situations in which the elements of the fluid are moving.2043 E—01 8.87 400.2943 E—02 Geometric Temperature (CR) Density (P/PsL) Viscosity 8.2589 PSL = 2116.22 405. the pressure at the back face of the element is p — ( at the front face is p + If the fluid is not accelerating. there are no shear forces.62 978.0943 E—01 8.9060 E—02 1.1127 E—01 8.0390 E—01 8.94 980.Thus. 1.67°R PSL = 0.0020 E—O1 8.14 5.96 402. Since there is no relative motion between adjacent layers of the fluid. 1. A firstorder Taylor's series expansion is used to evaluate the pressure that at each face.0909E—02 1. Thus.31 406. are either all at rest or all moving with the same velocity.57 8.49 408.94 984. y.3742 E—02 3.1311 E—01 8. the viscosity of the fluid is of no concern.2. Standard Atmosphere.24 973.22 tbf/ft2 = 518.4460 E—02 3. For equilibrium.1494 E—01 8.2360 E—02 2.Thus. 1976 English Units (continued) Pressure (P/PSL) 4.25 989.0759 E—01 8.6651 E—02 3.2024 x iCY5 lbf s/ft2 = 3. the sum of the forces in any direction must be zero.S.2 / Solving for the Aerothermodynamic Parameters TABLE 1.8505 1.3960 E—02 1.7379 E—02 1.60 397.28 981. the fluid is said to be a static 1'nediun2. For these inviscid flows.2225 E—01 82 84 86 88 90 92 94 96 98 100 Reference values: 2.43 395.57 990. Thus.1860 E—01 8.2044E—02 1.8197 E—02 4. z as shown in Fig.9710 E—02 3.78 399.91 393.0575 E—01 8.6841 E—02 1.05 403.94 988.002377 slugs/ft3 = 1.0339 E—02 1.59 974.7649 E—02 2.0205 E—01 Altitude (kft) 72 74 76 78 80 Speed of Sound (ft/s) 972.1677 E—01 8.0284 E—02 3.40 407.34 394.740 X 1 To do this.60 986. Consider the small fluid element whose center is defined by the coordinates x.93 976. with no relative motion between fluid elements.3121 E—02 4.3355 E—02 3.5331 E—02 1.2B 37 U. I' I 3pLIx\ \ ( \ ( (LiSa) (L15b) .3195 E02 1. the element must be in equilibrium. the only forces acting on the surface of the fluid element are pressure forces.63 982.4589 E—02 2.5850 E—02 1.9761 E—02 2.0997 E—02 404.52 396.0364 E—02 2.5183 E—02 2.Sec. If fluid particles.2755 E—02 2.69 398. provided that there is no relative motion between fluid elements.28 977. let us develop the equations describing the pressure variation in a static fluid medium.6060 E—02 8. as a continuum.
an ordinary derivative may be used and equation (1. That is.18) .16a) (1.16c) may be written dp (1. or shearfree. flow: (1) There is no pressure variation in the horizontal direction. Combining terms and dividing by txz gives us (1. 1 I WhyStudy Aerodynamics? Origin of the cell in the coordinate system Figure 1. the pressure is constant in a plane perpendicular to the direction of gravity.16b)  ay ap (li6c) The three equations can be written as one using vector notation: Vp = p7 = —pgk (1. that is. density. as the pressure is the same on all faces. hydrostatic. pressure at a point in a static fluid is independent of orientation. that is. the body force is gravity.17). —(p + + / z (p — — = 0 (1. For the cases of interest in this book. it is not a function of x or y.2 Derivation of equations (1. it can be seen that Furthermore.38 Chap. —* 0).15) through (1. These equations illustrate two important principles for a nonaccelerating. Since the pressure varies only with z.15c) Note that the coordinate system has been chosen such that gravity acts in the negative z direction. as the element shrinks to zero volume (i. and (2) the vertical pressure variation is proportional to gravity.. and change in depth.17) where 7 represents the body force per unit mass and V is the gradient operator.e.
1. Standard Atmosphere. is equal to 5. separating the variables and integrating between two points yields j — = in — = — j[ P2 fdp g dz g — — Zi) where the integration reflects the fact that the temperature has been assumed constant.Sec. T=ThBz where T0 is the sealevel temperature (absolute) and B is the lapse rate. as taken from the U. Rearranging yields [g(zi z2) p2piexp[ RT — The pressure variation described by equation (1. are reproduced in Table 1.000 m.21) into the relation Jp and integrating.22) The exponent g/RB.20) is a reasonable approximation of that in the atmosphere near the earth's surface.5 The Standard Atmosphere In order to correlate flighttest data with windtunnel data acquired at different times at different conditions or to compute flow fields. "standard" atmospheres have been developed based on the knowledge of the atmosphere at thô time. it is important to have agreedupon standards of atmospheric properties as a function of altitude. Since the earliest days of aeronautical research. which is dimensionless. both of which vary from day to day.0065 K/rn Substituting equation (1. An improved correlation for pressure variation in the earth's atmosphere can be obtained if one accounts for the temperature variation with altitude.10) can be substituted into equation (1. we obtain [dp JRT / — [gdz Bz\g/RB P= (1. . the expression for density given by equation (1. Standard Atmosphere (1976).S. The atmospheric properties most commonly used in the analysis and design of ifight vehicles.15 K and B= 0.2.2 / Solving for the Aerothermôdynamic Parameters 39 Let us assume that the air behaves as a perfect gas. The earth's mean atmospheric temperature decreases almost linearly with z up to an altitude of nearly 11.18) to give dp = —pg = pg — (1.That is.26 for air. The one used in this text is the 1976 U. These are the properties used in the examples in this text. Thus.2.S. 1.19) In those regions where the temperature can be assumed to constant.000 m: T9 = 288.The following standard values will be assumed to apply from 0 to 11.
000 ft. the hydrostatic equation. In reality. A reasonable approximation is that the temperature varies linearly with altitude in some regions and is constant in other altitude regions..01325 x J — 0.14).The calculated value for the ambient pressure is I = 1.26 288.18)J. a standard atmosphere is a valuable tool that provides engineers with a standard when conducting analyses and performance comparisons of different aircraft designs.641 X The comparable value in Table 1.2. and equation (1. for which the density is eliminated through the use of the equation of state for a thermally perfect gas. equation (1.17).21) and (1. variations would exist from one location on earth to another and over the seasons at a given location.0O3565z and that the temperature from 36.15 — 223. respectively.650 X EXAMPLE 1. Assume that the temperature of air from 0 to 36. calculate the temperature and pressure of air at an altitude of 10 km. equation (1.10: Properties of the standard atmosphere in English units Develop equations for the pressure and for the density as a function of altitude from 0 to 65.40 Chap.15 K The tabulated value from Table 1.100 to 65.67 — 0. .252 K. from sounding rockets. Nevertheless. Sutherland's equation.100 ft is given by T= 518.000 mis T= T0 — Bz = 288.12). = 1. 1 I Why Study Aerodynamics? The basis for establishing a standard atmosphere is a defined variation of temperature with altitude.10).01325 X = 2. Given the temperature profile. Viscosity and the speed of sound can be determined as functions of altitude from equations such as equation (1.The analytical model of the atmosphere should make use of the hydrostatic equations [i. and the perfectgas equation of state.2 is 2. are used to derive the pressure and the density as functions of altitude. EXAMPLE 1.15 j N/rn2 N/rn2.2 is 223 . Compare the tabulated values with those presented in Table 1.9: Properties of the standard ahnosphere at 10 km Using equations (1.e.22). This atmospheric temperature profile is developed from measurements obtained from balloons. equation (1.97°R. Solution: The ambient temperature at 10.000 ft is constant at 389.0065(104)15. and from aircraft at a variety of locations at various times of the year and represents a mean expression of these measurements.
34 ftlbf lbrn / lbf = 4.0 — 6.22(1.174 S RT (53.003565°R/ft. T= 518. the density is — — p RT — 2116.67°R and B = 0.000 ft.21).003S6Sz) Dividing by p0.0 6. with the values at 36.100 — z) RI' s2 ( ft•lbf '\ OR)(389.873 X z)526 53.873 X 10—6 472.003565z Thus. multiply by (1/ge).97R) However.9exp[ In English units.34)(518. 12 I Solving for the Aerothermodynamic Parameters 41 Solution: From 0 to 36.67) one obtains the nondimensionalized density: = (1.873 X 106 z)426 Po Since the temperature is constant from 36.22 Po RI'0 — (53.0 — 6.loo = 2116.000 ft.67 — 0. to have the correct units.19 lbf/ft2 Thus. T0 = 518.873 >< z)526 For a thermally perfect gas.20) can be used to express the pressure.22) gives us I = 2116. P=472.67 — 0.100 to 65. Specifically.Sec. — — 2116.0 — 6.8075 X 105/ft . the value of the density at standard sealevel conditions.100 ft serving as the reference values Pi and zi: P36.22(1.22(1. so that g = 32.34(518. Using English unit terms in equation (1. g RT — Ig(36. the temperature varies linearly as described in general by equation (1. equation (1.
. when the drag (or thrust required) curves are presented for aircraft weights of 8. the total drag is equal to the thrust required and the lift balances the weight. X 0. The reader should note that it may be necessary to use alternative relations for calculating fluid properties.7355 — 4. and Yos (1963)]. the drag acting on the vehicle (which is equal to the thrust required to cruise at constant velocity and altitude) reaches a minimum (Dmjn).000 feet.8075 X z) The nondimensionalized density is p p0p0T Since T = 389.g.000lbf T38A can sustain at 20.000 lbf. and speed of sound) have been presented in this chapter. viscosity. P1. the speed of sound is constant for Fig. and the wave drag.7519T0.g. and 12. With the aircraft cruising at a constant altitude (of 20..2231 exp (1.g.000 feet). What is the maximum cruise velocity that a lO. density. The lifttodrag ratio .1.1 through 1. dissociation). The maximum velocity at which an aircraft can cruise occurs when the thrust available with the engines operating with the afterburner lit ("Max") equals the thrust required. Hansen (1957). 10. For instance. Therefore.. More will be said about such curves in Chapter 5. the parasite drag. the total drag is the sum of the induced drag. The thrust available is presented as a function of Mach number for the engines operating at military power ("Mit") or operating with the afterburner ("Max"). When the vehicle is cruising at a constant altitude and at a constant attitude. they reflect the fact that the induced drag depends on the lift. which are represented by the bucket shaped curves.8075 105z) 1.97°R = Po — p T0 0. where the induced drag dominates. 1.1 and the Mach number could be replaced by the i. for the relatively high temperatures associated with hypersonic flight. As will be discussed in Chapter 5. Numerous references present the thermodynamic properties and transport properties of gases at high temperatures and pressures [e.7355 — 4. at the lower velocities. 1 / Why Study Aerodynamics? Thus.000 lbf. = Po 0.42 Chap. it may be necessary to account for realgas effects (e.000 lbf. Moeckel and Weston (1958).2967exp(1.3 SUMMARY Specific values and equations for fluid properties (e. P1. But the lift is equal to the weight. PROBLEMS Problems 1. by the true air speed.5 deal with the EnergyManeuverability Technique for aT38A that is powered by two J85GE5A enginea Presented in Hg.1 are the thrust available and the thrust required for the T38A that is cruising at 20.e.000 feet? As the vehicle slows down. the drag is a function of the weight of the aircraft. Thus..
000lbfT38A can cruise at 20.000 It? 1. C Mach Figure P1.Problems Thrust Required and Thrust Available (2) J85GE5A Engines Aircraft Weights of 12000. while passing through Mach 0. You are operating the aircraft in the region of reverse command. 10000 and 8000 lbs at an Altitude of 20000 ft 43 ''4 >< I.000 It? What is the maximum rateofclimb that our 10. What are the total energy.65). at a Mach number of 0.65 at a constant altitude of 20. more power) to fly slower.1 is. P1. e. a maximum [(LID )max].. More thrust is required to cruise at a slower speed.000lbf T38A is using "Mil" thrust to cruise at a Mach number of 0.000lbfT38A can achieve using "Mil" thrust.000 It? What is the velocity at which the vehicle cruises. when using "Mu" thrust while climbing through 20. therefore. when the lifttodrag ratio is a maximum? As the vehicle slows to speeds below that for [L/Dmju].000 ft? .2. What is the minimum velocity at which a 10.3. What is the maximum acceleration that our 10. the energy height. Eventually.000 ft? Is this minimum velocity due to stall or is it due to the lack of sufficient power? 1. which is equal to [(L/D )max]' it actually requires more thrust (i.65 at 20. and the specific excess power. one of two things happens: either the aircraft stalls (which is designated by the term "Buffet Limit" in Fig.000lbfT38A can achieve at a constant velocity (specifically.000lbf T38A cruising at 20. What is the maximum value of the lifttodrag ratio [(L/D)max] for our 10. if our 10.1) or the drag acting on the aircraft exceeds the thrust available.
16 x 108 and C2 = 183.44 Chap.15 . Nitrogen is often used in wind tunnels as the test gas substitute for air. Compare the values of (L/D)max for aircraft weights of 8. The gas constant.6.000 lbf. 10.10) . The gas constant.000 ft. (1.e.6°R 2. when ourT38A aircraft cruises at 20.70.10) for nitrogen. 1.458 x s• kg K05 and C2 = 110. C1 = 1. where the freestream static pressure is 586 N/m2 and the freestream static temperature is 54. Compare the velocity that is required to cruise at (L/D)max for each of the three aircraft weights. Eqn. which is used in the calculation of the density for a thermally perfect gas.4.6) for nitrogen at a temperature of 350°F and at a pressure of 150 psia with that for air a same conditions. i. i.27 x for air. Similarly. The constants for Sutherland's equation to calculate the coefficient of viscosity. 1 / Why Study Aerodynamics? 1.000 lbf. The constants for Sutherland's equation to calculate the coefficient of viscosity. p= (1. Compare the specific excess power for a 10.6°R for nitrogen. iS.. 1. Similarly. which is used in the calculation of the density for a thermally perfect gas.12a. Compare the value of the kinematic viscosity (1..12b. C1 and C2 = 198.4K f or air. thm• °R 1. Eqn. are: °R is equal to ftlbf for air and to 55.7.000lbf T38A cruising at the Mach number required for while operating at "Mit" thrust with that for the aircraft cruising at a Mach number of 0. with the value for air at the same conditions.e. Compare the value of the kinematic viscosity for nitrogen in a windtunnel test. C1 1.39 X 10_6 ka b s m K°5 and C2 = 102K for nitrogen. are: C1 = 2. and 12.35 and with that for the aircraft cruising at a Mach number of 0. 1.000 Ibf.3 K.
25 lbf/ft2 abs.17. Using equations (1. viscosity.16.12b)] at 35.000 to 20. and velocity? Note that 1.10)] and the viscosity [equation (1. develop metricunit expressions for the pressure. The conditions in the reservoir (or stagnation chamber) of Tunnel B at the Arnold Engineering Development Center (AEDC) are that the pressure (Pti) is 5.5 psia and the temperature (Ti) is 1660°R.2. Using the expressions developed in Problem 1. If the initial temperature is 200°F. and the speed of sound for the ambient atmospheric air at 18 km? Compare these values with the corresponding values tabulated in Table 1. What is the kinematic viscosity at this altitude [equation (1. what is the final temperature? 1.13. The temperature is constant and equal to 216. what are the density and the viscosity in the reservoir? 1. The isentropic expansion of perfect helium takes place such that is a constant. The corresponding values for the freestream pressure and the freestream temperature in the test section are 23.000 m.2. The air in Tunnel B is expanded through a convergent/divergent nozzle to the test section.3 K. The air in Tunnel C accelerates through a convergent/divergent nozzle until the Mach number is 4 in the test section.2.18. Compare the tabulated values with those presented in Table 1.14.2.723 X 106 N/rn2 and the temperature (TA) is 750 K.650 K over this range of altitude. If the pressure decreases to onehalf of its original value.2. the freestream static pressure is 586 N/rn2.Problems 45 is equal to 287. Using the values for the pressure and for the temperature given in Table 1. Using the values for the pressure and for the temperature given in Table 1.18. where the Mach number is 8. calculate the density [equation (1.9.13)]? 1.12. and the density of the atmosphere from 11.10. What is the velocity of the airplane in km/h? In ft/s? 1.13)]? 1. what are the corresponding = values for the testsection density. viscosity. Using the perfectgas relations.8. what are the density and the viscosity in the reservoir? 1.10)] and the viscosity [equation (1. what is the final temperature in in °C? 1.05 for air and to 297 for nitrogen. the density. calculate the density [equation (1.21) and (1. 1.19. A perfect gas undergoes a process whereby the pressure is doubled and its density is decreased to threequarters of its original value. and —65°F. Using an approach similar to that used in Example 1. Compare the calculated values with those given in Table 1.2. Using the perfectgas relations.15. what are the pressure.11.22).000 ft. and velocity in the test section? Note that = Using the values for the static pressure given in Table 1. the viscosity. The conditions in the reservoir (Or stagnation chamber) of Aerothermal Tunnel C at the Arnold Engineering Development Center (AEDC) are that the pressure (Psi) is 24.7. The pilot announces that you are flying at a velocity of 470 knots at an altitude of 35. What would be the advan tage(s) of using nitrogen as the test gas instead of air? 1. Using the perfectgas relations. Compare the calculated values with those given in Table 1. the temperature. calculate the temperature and pressure of the atmosphere at 7000 m.2. 1. What is the kinematic viscosity at this altitude [equation (1. and the freestream temperature is 54. What are the corresponding values for the freestream density.000 ft. respectively. what happens to the temperature? If the initial temperature is 20° C. . what is the pressure altitude simulated in the wind tunnel by this test condition? 1.12a)] at 20 km.
000 ft MSL (mean sea level). What are the atmospheric values of the pressure. Find the values for the pressure altitude (hr). Stagnation point Atmosphere Figure P1. Chap. of the temperature. that define the freestream properties that you would use in the calculations that determine if a proposed design can meet this requirement? 1. Compare these values with the corresponding values presented in Table 1. (b) If the aircraft were flying in a standard atmosphere. How far down must one go before the pressure is 1 atm greater than the pressure at the surface? Use equation (1.46 1.24.18).000 and 65. A Utube manometer is used to measure the pressure at the stagnation point of a model in a wind tunnel.21 and 1. and 1.24). If there is a difference of 2. the other side is open to the atmosphere (Fig. and of the density.22. the temperature altitude (hT). what would be the relationship among hT. if the density of the freestream flow is 0.21. For instance. the temperature. An aircraft flying at geometric altitude of 20.7. Problems 1.2. calculate the pressure. If the water in a lake is everywhere at rest. and the density altitude (ha) to the nearest 500 ft. One side of the manometer goes to an orifice at the stagnation point. It is common to refer to the freestream properties by the altitude in the atmosphericmodel at which those conditions occur.00199 slugs/It3' then the density altitude (ha) would be 6000 ft.22 deal with standard atmosphere usage.000.20. The properties of the standard atmosphere are frequently used as the freestream reference conditions for aircraft performance predictions.000 ft has instrument readings of p = 900 lbf/ft2 andT = (a) 460°R.000 ft. 30. 1.23. P1. One of the design requirements for a multirole jet fighter is that it can survive a maximum sustained load factor of 9g at 15. 1 / Why Study Aerodynamics? Using the expressions developed in Example 1.4 cm in the mercury levels in the two tubes. and the density for the ambient atmospheric air at 10.24 . what is the pressure as a function of the distance from the surface? The air above the surface of the water is at standard sealevel atmospheric conditions. what is the pressure difference in N/rn2. 1.
1960.26. e. Cam bridge: Cambridge University Press Coram R.. is a constant. DC: U. Hansen (1957). and air to 30. During the testing at the maximum design airspeed. The Viper revolution. Government Printing Office Werner J. 1963.g. Composition and thermodynamic properties of air in chemical equilibrium. 1. 2005. 2002. U. 1997. Standard Atmospher& Washington. (a) What is the difference between the pressure that is acting on the inner surface of the window relative to the pressure acting on the outer surface of the window? (b) If the window has a total area of 0.. p = 1 and cDT..25. Boyd. Consult a reference that contains thermodynamic charts for the properties of air. h = where ci. (2) as a calorically perfect gas. Mechanicsburg: Stackpole Books Yos 3M. one on the inner side of the window and the other port on the outside. and delineate the temperature and pressure ranges for which air behaves: (1) as a thermally perfect gas.e.1962. Estimated viscosities and thermal conductivities of gases at high temperatures. The manometer fluid is water. Report R50 Hillaker H. Knight of Germany: Oswald Boelcke — German Ace. A manometer is connected to two pressure ports.5m2. The Mathematical Theory of Nonuniform Gases.References 47 1.000 K. Note 4265 Svehla RA. oxygen. Code One Magazine 12(3) Moeckel WE.Report R132 1976. what is the total force acting on the window due to the pressure difference? REFERENCES Chapman S.S. Weston KC. RADTM637 . 1957.. Cowling TG. i. NASA Tech. Approximations for the thcrmodynamic properties of air in chemical equilibrium. 1958. i. the column of water in the tube that is connected to the outside pressure port is 30 cm higher that the column of water in thetube that is connected to the inside port. NACA Tech. S. Transport properties of nitrogen. 2004. A fairing for an optically perfect window for an airborne telescope is being tested in a wind tunnel.e. Air Force Magazine 87(1):64—69 Hansen CF. The Fighter Pilot Who Changed the Art of War. Boston: Little Brown Grier P. NACA Tech. hydrogen. AVCO Corp. Tribute to John R. Boyd.
Conservation of energy (or the first law of thermodynamics) Because the flow patterns are often very complex. and the altitude and speed at which the vehicle is traveling. it may be necessary to use experimental investigations as well as theoretical analysis to describe the resultant flow. Conservation of mass (or the continuity equation) 2. The theoretical descriptions may utilize simplifying approximations in order to obtain any solution at aH. The fundamental physical laws used to solve for the fluid motion in a general problem are 1. The resultant flow pattern depends on the geometry of the vehicle. Conservation of linear momentum (or Newton's second law of motion) 3. it is necessary to be able to describe the pattern of flow around the configuration. to predict accurately the aerodynamic forces and moments that act on a vehicle in flight.2 FUNDAMENTALS OF FLUID MECHANICS As noted in Chapter 1. Thus. it is important that we understand the fundamental laws that govern the fluid motion so that we can relate the theoretical solutions obtained using approximate flow models with the experimental results.The validity of the simplifying approximations for a particular application should be verified experimentally. which usually involve scale models. 48 . its orientation with respect to the undisturbed free stream.
Sec.1 (Nonsteady) airflow around a wing in the groundfixed coordinate system. which is initially at rest well ahead of the airplane. The flowfield solution can be formulated from the point of view of an observer on the ground or from the point of view of the pilot Provided that the two observers apply the appropriate boundary conditions to the governing equations. The particle. the airplane is flying into a mass of air substantially at rest (assuming there is no wind).. is accelerated by the passing airplane. 2. 2. The description of the flow field in the groundobserverfixed coordinate system must represent the timedependent motion (i.1 / Introduction to Fluid Dynamics 2. To an observer on the ground. both observers will obtain the same values for the aerodynamic forces acting on the airplane. ° particle Initial time. The motion of a typical air particle is shown in Fig. it is necessary to solve the equations governing the flow field about the vehicle. a nonsteady flow).1 49 INTRODUCTION TO FLUID DYNAMICS To calculate the aerodynamic forces acting on an airplane. The neighboring air particles are accelerated and decelerated by the airplane and the reaction of the particles to the acceleration results in a force on the airplane.1. t0 Groundfixed reference \ Time Groundfixed reference Time +2k: Groundfixed reference Figure 2. .e.
t0 w Control volume  Fluid particle which passed through control volume at time t0 ) Time t0 +At I U. Thus. as shown in rig. the velocity and the flow properties of the air particles that pass through a specific location relative to the vehicle are independent of time.2 (Steady) airflow around a wing in a vehiclefixed coordinate system. 2 / Fundamentals of Fluid Mechanics As viewed by the pilot. the terms of the flowfield equations that contam partial denvatives with respect to time are zero in the vehiclefixed coordinate system. If the airplane is flying at constant altitude and constant velocity. Therefore. 2. Note that the subsequent locations the air particle which passed through our control volume at time t0 are mcluded for comparison with Fig 22 In this text we shall use the vehicle (or pilot)fixed coordinate system. instead of descnbmg the fluid motion around a vehicle flying through the air. the air is flowing past the airplane and moves in response to the geometry of the vehiclé. Thus. the equations are usually easier to solve in the vehicle (or Because pilot)fixed coordinate system than in the groundobserverfixed coordinate of the resulting simplification of the mathematics through Galhlean transformation from the grouiidflxedreference coordinate system to the vehiclefixedreference coordinate system.The flow field is steady relative to a set of axes fixed to the vehicle (or pilot).0 w Time t0 + w Figure 2. we will examme air Vehiclefixed reference Initial time. .50 Chap. many problems in aerodynamics are formulated as the flow of a stream of fluid past a body at rest.2.
2 I Conservation of Mass 51 flowing around a fixed vehicle. we shall use a Cartesian coordinate system (x. For convenience.. and the area of that surface.4. z).. those conditions far from the vehicle). since the properties are a function of position. When there is no relative motion between the fluid particles.. Furthermore. we shall treat a twodimensional flow. The massflow rate through a surface bounding the element is equal to the product of the density. in the interest of simplicity. the net outflow of mass per unit time per unit depth (into the paper) is [ pu + —[Pu y + a(pu) — F pv + F r3(pv) zxyl [ 3(pv) ax — LPV — PU r a x Figure 2. As indicated in the sketch of Fig. hence.2 CONSERVATION OF MASS Let us apply the principle of conservation of mass to a small volume of space (a control vol. which is in reality the speed of the vehicle (see Fig.1). Flow patterns are the same for any xy plane. 2. At points far from the vehicle (i. one in which there is no flow along the z axis. y. 2. the component of the fluid velocity in the x direction will be designated by u. 2.2. A firstorder Taylor's series expansion is used to evaluate the flow properties at the faces of the element. Since all the fluid particles in the free stream are moving with the same velocity. that is. and that in the y direction by v. pressure and temperature) are the same for either coordinate system. Flow out of the volume is considered positive..The subscript 00 or 1 will be used to denote the undisturbed (or freestream) flow conditions (i.e. there is no relative motion between them. 2. there are no shearing stresses in the freestream flow.e. 2. the velocity component normal to the surface.Sec. The net outflow of mass through the surface surrounding the volume must be equal to the decrease of mass within the volume.3 Velociiies and densities for the massflow balance through a fixed volume element in two dimensions. and.3. 2. ume) through which the fluid can move freely.g. the fluid is termed a static medium.3. Referring to Fig. The values of the static fluid properties (e. . see Section 1. the undisturbed free stream). the fluid particles are moving toward the vehicle with the velocity (see Fig.2).
. 2 / Fundament&s of Fluid Mechanics which must equal the rate at which the mass contained within the element decreases: —— at tXx Equating the two expressions. incompressible. so that the density is essentially constant. 2. is a function of x.2) As has been discussed.e.) .Chap. the normal component of the velocity is zero at a solid surface). the pressure variations that occur in relatively lowspeed flows are sufficiently small.4.4) for the velocity field.4) Using boundary conditions. such as the requirement that there is no flow through a solid surface (i. the continuity equation becomes au i3v 8w —+—+—=O ax 8y 8z (2. the boundarylayer thickness at a given station. which is called the boundary layer and is discussed at length in Chapter 4. and dividing by 3p a a we obtain lithe approach were extended to include flow in the z direction. we can solve equation (2. In so doing. For these incompressible flows. we would obtain the general differential fonn of the continuity equation: ap + a + a + a = 0 In vector form.3) In vector form. a freestream flow) approaches a flat plate. Is a horizontal line parallel to the plate and a distance from the plate (where is equal to 6 at the downstream station) a streamline? (See Fig. we obtain a detailed picture of the velocity as a function of position. EXAMPLE 2. uniform flow whose (i. the equation is + V (pV) 0 (2.e.. combining terms. the streamwise component of velocity is given by velocity is \ 1/7 u = where 6. In the viscous region near the surface.1: Incompressible boundary layer Consider the case where a steady. this equation is (2.
Thus. incompressible flow. — 0u — 3y ax — 7 Integrating with respect to y yields V= y8"7 d6 8 dx + C 0 where C. and this line is not a streamline. Solution: By continuity for this steady. 2.5 Nomenclature for the integral form of the continuity equation. the mass conservation principle can be applied directly to the entire region. . there is flow across the horizontal line which is above the surface.. Integrating equation (2. 2. there is no flow through the wall). the constant of integration.5) yields Figure 2. when y = Since v is not equal to zero.Sec. äv —+—=0 3y 8u ax Since u = and 6(x).e.4 Flow diagram for Example 2.2) over a fixed finite volume in our fluid space (see Fig. can be set equal to zero since v = when y = 0 (i. If the details of the flow are not of concern.2 I Conservation of Mass Uco 53 Approach flow plate (Y Figure 2.1.
density. using a co ordinate system that is neither accelerating nor rotating. which is called an inertial coordinate system. As the fluid element moves in space. which is = where it dA is a vector normal to the surface dA which is positive when pointing outward from the enclosed volume and which is equal in magnitude to the surface area. Since the time differentiation of equation (2. The volumetric flux Q is the flow rate through a particular surface and is equal to if V•itdA. but its mass is conserved. Since V(x. t) and x(t). The resultant equation is the general integral expression for the conservation of mass: + JpV. the position coordinates x.y.54 '3p Chap. and volume may change. The circle through the integral sign for the area indicates that the integration is to be performed over the entire surface bounding the volume. For a sample problem using the integral form of the continuity equation. z of the fluid particle are themselves a function of time. in general.3 CONSERVATION OF LINEAR MOMENTUM The equation for the conservation of linear momentum is obtained by applying Newton's second law: The net force acting on a fluid particle is equal to the time rate of change of the linear momentum of the fluid particle. 2 / Fundamentals of Fluid Mechanics d(vol) vol + fff V. an explicit function of time t as well as of its position x. (pV)d(vol) = VOl 0 The second volume integral can be transformed into a surface integral using Gauss's theorem.itdA =0 (2. see Example 2.5) In words. Furthermore. Thus. or substantial derivative of V.y(t).7) . y. total.6) follows a given particle in its motion. its velocity.3. dV dt dx 0V dy aV dz av at ax dt äy dt az dt (2. 2. z. we may write 4 (2. z.y. the time rate of change of the mass within the volume plus the net efflux (outflow) of mass through the surface bounding the volume must be zero. and z(t). shape. the derivative is frequently termed the particle.6) The velocity V of a fluid particle is.
The principal forces with which we are concerned are those which act directly on the mass of the fluid element. Problems where the local. the substantial derivative is the sum of the local.3 I Conservation of Linear Momentum 55 (The reader should note that some authors use D/Dt instead of d/dt to represent the substantial derivative. fluid particles can accelerate due to the unbalanced forces acting on them. Note that even for a steadystate flow where ar/at is equal to zero.Sec. The convective acceleration of a fluid particle as it moves to different points in space is by the second term in equation (2. the pressure forces and shear forces.) However. 2.9). timedependent changes that occur at a point in the flow field and of the changes that occur because the fluid particle moves around in space. .9) dt — at total local convective Thus. This is the case for an air particle that accelerates as it moves from the stagnation region to the lowpressure region above the airfoil. 2.6 Nomenclature for the normal stresses and the shear stresses acting on a fluid element. the body forces. —=U dt dx —=V dt 3 dy dz = dt 3 W Therefore. and those which act on its surface. The stress system acting on an element of the surface is illustrated in Fig. timedependent changes are zero.8) or 3 * DV total (2. at are known as steadystate flows. The stress components T acting on the small cube are assigned y z / Figure 2. the acceleration of a fluid particle is dV av au aV aV —=—+u——+v—+w—— dt at ax ay az 3 (2.6.
they produce net forces on the fluid particle. The first subscript indicates the direction of the normal to the surface on which the stress acts and the second indicates the direction in which the stress acts.(Ti) Figure 2. In general. Thus.0) as shown in Schlichting (1968).1. the various stresses change from point to point. 'Txy = Tyx Tyz = Tzy Tzx = 'Txz (2.7. that is. which cause it to accelerate. fluids are isotropic. As a result. The resultant force in the x direction (for a unit depth in the z direction) is + + where is the body force per unit mass in the x direction. denotes a normal stress acting on that surface.7 Stresses acting on a twodimensional element of fluid.. The stresses are described in terms of a righthand coordinate system in which the outwardly directed surface normal indicates the positive direction. The properties of most fluids have no preferred direction in space. To simplify the illustration of the force balance on the fluid particle we shall again consider a twodimensional flow. the resultant force in the x direction is + y + (Tyv) f )i + T_ — + a Ax I Txx + ( 'Tyy 3 — i. Thus. Similarly.56 Chap. . Txy denotes a stress acting in the y direction on the surface whose normal points in the x direction. The forces acting on each surface are obtained by taking into account the variations of stress with position by using the center of the element as a reference point. 2. Including flow in the z direction. The body force for the flow fields of interest to this text is gravity. 2 I Fundamentals of Fluid Mechanics subscripts. as indicated in Fig.
lla) Similarly. since V V = 0 by continuity. For a fluid at rest or for a flow for which all the fluid particles are moving at the same velocity. such as the analysis of the shockwave structure. For fluid particles. the shear stress vanishes).3 I Conservation of Linear Momentum 57 which. The relations between the stress components and the rateofstrain components must be invariant to a coordinate transformation consisting of either a rotation or a mirror reflection of axes. is equal to du —+(V. The friction law for the flow of a Newtonian fluid where T = JL(8u/8y) is a special case of this linear stress/rateofstrain relation.Sec. LV + + = Tzz —p — 2 = — 3 + aw az . there is no shearing stress.V)u Equating the two and dividing by the volume of the fluid particle Ax AyAz yields the linear momentum equation for the x direction: du = + + Tyx + a Tzx (2. one should include the second viscosity coefficieni (A).e. in a more rigorous development. we need to relate the stresses to the motion of the fluid. by equation (2. is presumed to apply See Schlichting (1968).llc) Next. Stress components may be expressed as a linear function of the components of the rate of strain. 3. the stress components must reduce to the hydrostatic pressure. where extremely large changes in pressure and in temperature take place over very short distances.6).p. the stress is related to the rate of strain by a physical law based on the following assumptions: 1. The viscosity /L is more precisely called the first viscosity coefficient. which would appear in the normal stress terms. we obtain the equation of motion for the y direction: dv and for the z direction: = + a + a + a (2. and the normal stress is in the nature of a pressure. For a fluid that satisfies these criteria.1Th) dw = + a + a Tyz + a (2. For other flows.. since a physical law cannot depend upon the choice of the coordinate system. 2. When all velocity gradients are zero (i.The Stokes's hypothesis (A = second viscosity coefficient is of significance in a few specialized problems. The term involving A disappears completely when the flow is incompressible. 2.
therefore. together with the heat due to friction.(—+—}'—— Oy)J Oz \Oz 0YL 0 [ /Ov Ow\l 2 3 Op + — 2. the density and the viscosity are unique functions of pressure and temperature. bring about considerable temperature variations.58 Chap. we obtain 3p 31 3u 2 3 [ (3u 3v\1 3x) J +— 3[ 13w L \ ay + p(V V)v + 3p i3z [ \ 3x 3u'\1 +—l I 3z I J (2. the unknown parameters that appear in the NavierStokes equatiOns are the three velocity components (u. Thus. Note that the viscosity is considered to be dependent on the spatial coordinates. for a fluid of known composition that is in equilibrium. the pressure (p).12b) + p(V V)w + ô[f'äw + +—lp. The temperature dependence of viscosity in the general case should. For a general application.12c) These general. differential equations for the conservation of linear momentum are known as the NavierStokes equations. and w). for a compressible flow.11).12a) 0( 3y\ äv äy I 2 3 0y + —l azL \3Y 3w afl (3w 0z1j 3u (2. there are five primary (or primitive) variables for a general . the changes in velocity and pressure. the density (p). and the viscosity we discussed in Chapter 1. v.a— — Oz OZ\ 3/' Ow V) / (2. be incorporated into the governing equations. This is done since. 2 / Fundamentals of Fluid Mechanics = (3u Tyx + + 8v 3W = Tyz = Tzx = = /3v Tzy + 3w With the appropriate expressions for the surface stresses substituted into equation (2.
For these flows. Since equations (2. twodimensional channel of height h (Fig. Because the channel is infinitely long. the equations for a static medium can be obtained by neglecting the tenns relating to the acceleration of the fluid particles and to the viscous forces. equation (2. EXAMPLE 2. the velocity components do not change in the x direction. we will assume that the viscosity and the density are constant. v.12c). Neglecting these terms in equations (2. we have a system of four independent equations that can be solved for the four unknowns.12a) through (2. and the pressure (p).16c).2). and the three components of the momentum equation.4 I Applications to ConstantProperty Flows 59 flow problem: the three velocity components. Let us assume that the body forces are negligible. . This is known as Poiseuille flow.12) over a volume and using Gauss's Theorem. there are only four unknowns: the three velocity components (u. the reader would obtain equations (1. at present we have only four equations: the continuity equation.4 APPUCATIONS TO CONSTANTPROPERTY FLOWS For many flows. flows.12a) through (2. that is. temperature variations are sufficiently small that the density and vis cosity may be assumed constant throughout the flow field.12c) and assuming that the body force is gravity and that it acts in the z direction. depending upon the application.8). + Fsurface a = pV d(vol) + J V(pV ii dA) (2. We are to determine the velocity profile and the shearstress distribution.16a) through (1. equations (2. Since the flow is low speed. one for which the solution will be Let us consider two obtained using differential equations and one for which the integral equations are used. In this text.12a) through (2. A gas flow is considered incompressible if the Mach number is less than 0. The terms lowspeed and/or incompressible flows will also be used in the description of these flows. 2. such a flow is termed a fully devel aped flow. In vector form. the pressure. the sum of the body forces and of the surface forces equals the time rate of change of momentum within the volume plus the net efflux of momentum through the surface bounding the volume. we would need to introduce the energy equation. Such flows will be termed constantproperty flows in this text. However. lowspeed flow of a viscous fluid in an infinitely long.12c) are the general differential equations for the conservation of linear momentum.Sec. and w). Thus. The integral form of the momentum equation can be obtained by returning to Newton's law. and the temperature. 2.To solve for a general flow involving all five variables. Thus.5.2: Poiseuille flow Consider a steady. 2. the energy equation is not needed to obtain the velocity components and the pressure of a constantproperty flow.3 to 0.13) This equation can also be obtained by integrating equation (2. The sum of the forces acting on a system of fluid particles is equal to the rate of change of momentum of the fluid particles.
all the terms in equations (2. 2 / Fundamentals of Fluid Mechanics ty I u(yonly) + h Figure 2. we know that v 0 everywhere. there is no flow through the walls or.LI—+—J r au aw'\1 I ax)] az L az ax)] (2. z.2. We can also neglect the body forces. or t..The continuity equation for this steadystate. ay Further. az ax a = a2u a /au\ + a  2 + a2u a a2u a a ay ax . w 0 and all the derivatives with respect to z are flow yields zero. Solution. the walls are streamlines) and v does not depend on x or z. and u is a function of y only. Thus.8 Flow diagram for Example 2.14) Because we are considering lowspeed of a simple is constant throughout the flow field. equivalently. the flow is everywhere parallel to the x axis.12b) and (2. we can rewrite the viscous terms of this equation as follows: a/ ax äu 2 3 —\ 1 a [ lau \.60 Chap. since it does not depend on x. = 0 and. Thus. hence. ay 8x ay [ av'\ll+—lpJ—+— a E lau aw ax!] az [ \. For a twodimensional flow. At this point.12c) are zero and need not be considered further.12a). Thus. since v = 0 at both walls (i.e. Expanding the acceleration term of equation (2. we obtain au au au au ap aI 3u 2 a[ ay L (3u ay av\1l+—l a J. au av ax 3y Smce the velocity components do not change m thex direction. w = 0 everywhere. v 0 everywhere.
Sec. we can write equation (2. = = = tOy 0 0 because the flow is steady because a = u(yonly) because v = 82u 0 0 pw— = = Thus. we can further simplify equation (2.4/ Applications to ConstantProperty Flows Noting that 3u 3x 8v tOy 8w 8z and that V V = 0 (since these are two ways of writing the continuity equation for a constant density flow). These two statements can be true These only if d2u dy dp = — = constant dx Integrating twice gives a= I dp aX + C1y + C2 .15) = — + However. 0 t3u 8x =0 because u = u(y only) because body forces are negligible o = —— + /L— tOy tOp 82u tOy 0=——tOz three equations require that the pressure is a function of x only.15) by eliminating terms whose value is zero. 2. Recall that a is a function of y only.14) as + + + tOp (82u + + 82u 02u\ (2.
dp/dx must be negative) to have a velocity in the direction shown.62 Chap. or favorable. we apply the viscousflow boundary condition that the fluid particles at a solid surface move with the same speed as the surface (i.. The negative. the positive momentum efflux at station 2 is balanced by the negative momentum influx at station 1: if = © Figure 2. do not slip relative to the surface). we find that C2 so that ldph2 dx 4 h2 ldp( 2p.e. 2.2. Thus. the velocity profile at the upstream station (i.e. The shear stress at the two walls is du dy y=±h/2 h dp The pressure must decrease in the x direction (i..e. with the maximum velocity at the center of the channel. 2 I Fundamentals of Fluid Mechanics To evaluate the constants of integration...dx\\ 4 The velocity profile is parabolic. Thus.e. station 2). h u=O When we do this. Examination of the integral momentum equation (2.9. Let us verify this by using equation (2.13) verifies that a change in pressure must occur to balance the shear forces. .9 Another flow diagram for Example 2. Since the flow is fully developed. station 1) is identical to that at the downstream station (i.13) on the control volume shown in Fig. pressure gradient results because of the viscous forces.
4 / Applicatiohs to ConstantProperty Flows Thus. Because the flow is steady.. that there are subtle implications regarding the signs in these terms. . euation (2. as shown in the approach using the differential equation. only the control volume the plate will be considered (i. such that velocity measurements at the trailing edge of the plate indicate that the x component of the velocity (above the plate) varies as U— (y\1/7 Below the plate. EXAMPLE 2i: Drag on a airfoil (V A steady. as discussed previously. 2. 1L01). Because of viscosity. Since the pressure is uniform over the control surface and since the body forces are negligible. Neglecting the body forces. which is —d/2.. for this steady flow with negligible body forces. The pressure is uniform over the entire cOntrol surface. since the flow is planar symmetric. what is the drag The drag coefficient is the drag per unit coefficient fOr this flow if <3 span (unit depth into the page) divided by the freestream dynamic pressure times the reference area per unit span (which is the chord length c): Cd= d Solution: Let us apply the integral form of the momentum equation [i. the flow near the plate slows.13)] to this flow. T.Sec. dp/dx < 0). the velocity is a mirror image of this profile. let us cOnsider thex component only. the only force acting on the fluid in the control volume is the retarding fOrce of the plate on the fluid. the shear acts to retard the. the first term on the righthand side of equation (2. For the velocity profile shown. Thus. since that is the direction in which the drag force acts.13) is zero.e.. from y 0 to y = 3). lowspeed flow approaches a very thin. 63 =0 or where the factor of 2 accounts for the existence of shear forces at the upper and lower walls. Finally. fluid motion and the pressure must decrease in the x direction (i. Furthermore.p2—p1h dph Note. flatplate "airfoil" whose length is c. Noting that the momentum equation is a vector equation.e.e.
2. because the flow is low speed. 2. Furthermore. the velocity vector has a y component. (1) Since dA is a vector normal to the surface dA which is positive when pointing outward from the enclosed volume. 2 / Fundamentals of Fluid Mechanics Noting that u and using the values of which are shown in Fig. the density will be assumed constant. ii is in the —x direction and dA per unit span is equal to dy.3. which is a function of y. and the velocity vector along this surface is The volumetric efflux (per unit depth) across this surface is VedX.64 Chap. . (y \117 IL0 F Figure 2.5).10. this y component of velocity does not transport fluid across the area I dy.10. we obtain d 2 = J (1) +J r 1/7 + (2) + I 1 ) 1/7 (3) Let us comment about each of the terms describing the momentum efflux from each surface of the control volume. and the momentum equation becomes = + + To obtain an expression for let us use the integral form of the continuity equation. for the flow of Fig. which will be represented by the symbol Q2.10 Flow diagram for Example 2. (2) Because viscosity slows the air particles near the plate. However. line 2 is not a streamline. (4) There is no flow across this surface of the control volume. equation (2. since it is at a solid wall. (3) Because of viscosity.
To obtain information necessary to develop a flow model that could be used in numerical solutions 2. Some of the objectives of such test programs are as follows: 1. To verify numerical predictions of the aerodynamic characteristics for a particular configuration 4.00389 2. numerous experimental programs have been conducted to measure directly the parameters that define the flow field. To investigate the effect of various geometric parameters on the flow field (such as determining the best location for the engines on a supersonic transport) 3. It is important. either scale models of the complete vehicle or largescale simulations of elements of the vehicle (such as the wing section) have been used in these windtunnel programs. 2.Sec.) for the windtunnel tests were not equal to the values for the flight condition that was to be simulated. the freestream conditions (such as the velocity. the static pressure. in many test programs. etc. then.5 REYNOLDS NUMBER AND MACH NUMBER AS SIMILARITY PARAMETERS Because of the difficulty of obtaining theoretical solutions of the flow field around a vehicle. to determine under what conditions the experimental results obtained for one flow are applicable toanother flow which is confined by boundaries .Olc) 1 Ti2 = 0. To measure directly the aerodynamic characteristics of a complete vehicle Usually.5 / Reynolds Number and Mach Number as Similarity Parameters 65 = f a + + fa + vJ}(Idy)} = + p f VedX + 0 Q2 Substituting this expression into the momentum equation yields d 2 d= Cd d 1 2 7 = 1 = (0. Furthermore.
Similarly. (b) second flow.To do this.66 Freestream conditions Chap. * P Poo.1 * w1= w U001 U2= * U U00.11 Flow around geometrically similar (but different size) configurations: (a) first flow. the freestream value of the property) for each of the two flows.e. that are geometrically similar (but of different size).15) can be written 8u 8u az 3p a2u a2u 3y2 a2u az2 ax ax2 (2. consider the xmomentum equation as applied to the two flows of Fig.i JLoo 1 and for the second flow. for the first flow. let us divide the velocity components by the freestream velocity.2 Note that the freestream values for all three nondimensionalized (*) thermodynamic properties are unity for both cases. For simplicity. for the first flow. * P * P * /1. 2 / Fundamentals of Fluid Mechanics L1 (a) Freestream conditions (ti) Figure 2. Thus. equation (2.2 * w2= w . 2. * U V1= * V U00. Pco.11.2 V2= * V U00.16) Let us divide each of the thermodynamic properties by the value of that property at a point far from the vehicle (i. Thus. Since the bodyforce term is usually negligible in aerodynamic problems..2 * P /i2— * Poo. let us limit ourselves to constantproperty flows. U1= and for the second flow.
the freestream boundary conditions are the same for both flows: that is. the xmomentum equation (2.iUoo. L/UOO is a characteristic time.1 = x Y2 = y z * tU03. at points far from the vehicle and A characteristic dimension L is used to nondimensionalize the independent variables. as indicated by the * quantities. the solutions of the two problems in terms of the dimensionless variables will be identical provided that the differential equations are identical.16) becomes * * * * at1 + * + 3x1 * + 8Yi * az1 * I I '\( 1* + t \pooiU00iJ8xi \poo. the equation of state is p00 p00RT00 . In this case.2 2 ap.2 L2 = L2 In terms of these dimensionless parameters.iLiJ\ POO.i 2 J( ax1 + p1—7 + 3y1 (2.Sec.18) p00U00 Recall that for a perfect gas. ( '\( I Both the dependent variables and the independent variables have been nondimensionalized.5 / Reynolds Number and Mach Number as Similarfty Parameters 67 With the velocity components thus nondimensionalized. Let us examine the first similarity parameter from equation (2. As a consequence. For the second flow. The differential equations will be identical if the parameters in the parentheses have the same values for both problems. * x * y * z * tU00. 2.17). at2 + * ax2 + * + ay2 az2 ( \\ P00.17a) 3z1 I for the first flow. The dimensionless boundarycondition values for the dependent variables are the same for the two flows around geometrically similar configurations. (2. the flows are said to be dynamically similar as well as geometrically similar.
The inverse of the second similarity parameter is written = L (2. This has been done using the values presented in Table 1. and what is the Reynolds number for this flight condition? Solution: The density. Note that 1 knot 1 nautical mile per hour. The freestream Reynolds number is defined by equation (2.20) which is the Reynolds number.Thus. a measure of the ratio of inertia forces to viscous forces. As has been discussed. EXAMPLE Calculating the Reynolds number An airplane is flying at a Mach number of 2 at an altitude of 40.18) yields 1 since Thus.0.84 X i04 ft) and velocities up to 2500 km/h (1554 mi/h or 1350 knots). such as the static pressure and the static temperature. 3600p = = S h = 5280pmi h . the chord of the wing or the diameter of the missile) chosen to be 1.0 m for the correlations of Fig.The correlations represent altitudes up to 30km (9.08 ft/s Since the Mach number is 2. 2 I Fundamentals of Fluid Mechanics and the freestream speed of sound is given by = Substituting these relations into equation (2. 2.2471PsL = 5. the viscosity. once the velocity.20) with the characteristic length L (e. and the speed of sound of the freestream flow at 40. the first dimensionless similarity parameter can be inter= preted in terms of the freestream Mach number.000 ft.8711 = 968.12.. the freestream Mach number and the freestream Reynolds number can be calculated as a function of velocity and altitude.68 Chap.g.2. and the characteristic dimension of the vehicle are defined. the freestream values of the fluid properties.2. what is the velocity in mi/h. are a function of altitude. If the characteristic length for the aircraft is 14 ft. the altitude.000 ft can be found in Table 1.97 13 x >< = °. = = 2.
s2\/ ft4 2.3560 x i07 2. Standard Atmosphere. The corresponding Reynolds number is (5.9713 x ft2 = 5. The velocity of the fluid particles increases from a value of zero (in a vehiclefixed coordinate system) at the wall to the value that corresponds to the external "frictionless" . the flow field may be divided into two regions: (1) a viscous boundary layer adjacent to the surface of the vehicle and (2) the essentially inviscid flow outside the boundary layer. 2.6 CONCEPT OF THE BOUNDARY LAYER For many highReynoldsnumber flows (such as those of interest to the aerodynamicist).S.Sec.12 Reynolds number/Mach number correlations as a function of velocity and altitude for U.8711 x iO4 /too lbf.6 1 Concept of the Boundary Layer 69 I Velocity (kmlh) Figure 2.
2 / Fundamentals of Fluid Mechanics • Relatively thin layer with limited mass transfer Relatively low velocity gradient near the wall Relatively low skin friction y • Thicker layer with considerable mass transport Higher velocities near the surface Higher skin friction y • • The boundary layer thickness I x x Laminar boundary layer Turbulent boundary layer Laminar portion of the boundary layer Turbulent portion of the boundary layer "Effective" inviscid body 'Effects of viscosity are confined to the boundary layer Outside of the boundary layer.13. The solution of the boundarylayer equations and the subsequent determination of a corresponding displacement thickness are dependent on the velocity at the edge of the boundary layer .The "effective" inviscid body (the actual configuration plus the displacement thickness) is represented by the shaded area of Fig.13). the shear forces are relatively large in the boundary layer. the flow may be assumed to be inviscid Figure 2. 2. whose edge is represented by the solid lines in Fig. To generate a solution for the inviscid portion of the flow field. the effect of the viscous terms may be ignored in the solution for the flow field external to the boundary layer.Chap. we require that the velocity of the fluid particles at the surface be parallel to the surface (but not necessarily of zero magnitude).13 Viscous boundary layer on an airfoil. This represents the physical requirement that there is no flow through a solid surface. the velocity gradients become so small that the shear stresses acting on a fluid element are negligible. 2.13. flow outside the boundary layer. Outside the boundary layer. Because of the resultant velocity gradients. The analyst may approximate the effects of the boundary layer on the inviscid solution by defining the geometry of the surface to be that of the actual surface plus a displacement due to the presence of the boundary layer (as represented by the shaded area in Fig. 2. Thus.
the pressure gradient normal to the surface is negligible: 3y which is verified by experiment. For many problems involving flow past streamlined shapes such as airfoils and wings (at low angles of attack). The assumption that the static pressure variation across the boundary layer is negligible breaks down for turbulent boundary layers at very high Mach numbers. Solving for the pressure gradients gives us —— ä1D ( a + Pu a  + pv  a .Sec. the terms on the righthand side of the equation in the second line are typically smaller than the term in the first line. then. The process of determining the interaction of the solutions provided by the inviscidf low equations with those for the boundarylayer equations requires a thorough understanding of the problem [e.—i  1a apla Providing that the boundary layer near the solid surface is thin. (1977) cite data for which the wail pressure is significantly greater than the edge value for turbulent boundary layers where the edge Mach number is approximately 20. We conclude. are 3u 9u 3x au äy 32u 0x 3x2 3y2 and 32v at ax ay ay ax2 ay2 where the x coordinate is measured parallel to the airfoil surface and the y coordinate is measured perpendicular to it. Thus. 2. . the velocity at the surface that corresponds to the inviscid solution).g.e. Let us consider this statement further by studying the x andy components of equations (2. The characteristics distinguishing laminar and turbulent boundary layers are discussed in Chapter 4. (1974)j. Bushnell et aL.. Since the static pressure variation across the boundary layer is usually negligible. the presence of the boundary layer causes the actual pressure distribution to be only negligibly different from the inviscid pressure distribution.13. v < u). 2.. the normal component of velocity is usually much less than the streamwise component of velocity (i. The resultant equations. refer to Brune et aL. that As a result. in effect.12) for a twodimensional incompressible flow.6 / Concept of the Boundary Layer 71 (which is. the pressure distribution around the airfoil is essentially that of the inviscid flow (accounting for the displacement effect of the boundary layer). which define the flow in the boundary layer shown in Fig.
Schetz (1993).12a) through (2. In equation (2.8 FIRST LAW OF THERMODYNAMICS Consider a system of fluid particles. 2 / Fundamentals of Fluid Mechanics When the combined action of an adverse pressure gradient and the viscous forces causes the boundary layer to separate from the vehicle surface (which may occur for blunt bodies or for streamlined shapes at high angles of attack). therefore. for a cyclic process. An example is flow in a heat exchanger. Everything outside the group of particles is called the surroundings of the system. that is. The Reynolds number. For such flows. we derive the energy equation and discuss its application to various flows. The use of lower case symbols to represent the parameters means that we are considering the magnitude of the parameter per unit mass of the fluid.22) . the continuity equation and the momentum equation are independent of the energy equation. the momentum equation. The first law of thermodynamics results from the fundamental experiments of James Joule. for this case. the continuity equation.7 CONSERVATION OF ENERGY There are many flows that involve sufficient temperature variations so that convective heat transfer is important.21). then — 6w = de e2 — e1 (2. — =0 (2. density and viscosity). the reader is referred to Chapter 4 and to Schlichting (1979). Joule found that. This is because. We use the symbols 8q and &w to designate that the incremental heat transfer to the system and the work done by the system are not exact differentials but depend on the process used in going from state 1 to state 2. the temperature field is obtained by solving the energy equation after the velocity field has been determined by solving the continuity equation and the momentum equation.g. 2. one in which the initial state and the final state of the fluid are identical. White (2005). the flow field is very sensitive to the Reynolds number. we have adopted the convention that heat transfer to the system is positive and that work done by the system is positive. Recall the discussion relating to equations (2. Joule has shown that the heat transferred from the surroundings to the system less the work done by the system on its surroundings during a cyclic process is zero. 2. also serves as an indicator of how much of the flow can be accurately described by the inviscidflow equations. if we apply it to a process that takes place between any two states (1 and 2). and Wilcox (1998). For detailed discussions of the viscous portion of the flow field. Equation (2.21) Thus. and the energy equation must be solved simultaneously. In the remainder of this chapter.72 Chap.12c).21) is true for any and all cyclic processes. For compressible flows. Compressible flows are those in which the pressure and temperature variations are sufficiently large that we must account for changes in the other fluid properties (e. but for which the constantproperty assumption is reasonable. Thus. but not vice versa. We must also include the energy equation in the solution algorithm for compressible flows..
Figure 2.. Note that de is an exact differen tial and the energy is. rection of the motion. In an inviscid flow. Chemical.8 I First Law of Thermodynamics 73 where e is the total energy per unit mass of the fluid. 2. which acts normal to the surface. and internal energies. an arbitrary zero energy (or datum) state can be assigned.24b) w= p dv where the work done by the system on its surroundings in going from state 1 to state 2 (a finite process).Sec. In this book. But the product of dA times is just d(vol).14. we will be concerned only with kinetic. we may say that work is done by a system on its surroundings if we can postulate a process in which the system passes through the same series of states as in the original process.24a) where v is the volume per unit mass (or specific volume). It is. and (3) all other energy. The magnitude of the effect is measured by the product of the displacement times the component of the force in the diWork. is positive when dv represents an increase in volume.22) becomes 6q — dke + dpe + dUe (2. as given by equation (2. 2. In mechanics. the only forces acting on a fluid system (providing we neglect gravity) are the pressure forces. as shown in Fig. the change in volume of the system. Equivalently.14 Incremental work done by the pressure force. therefore. there are no dissipative factors such as friction and/or heat transfer. and other forms of energy are normally not relevant to the study of aerodynamics. that is. In terms of the three energy components. the work per unit mass is = +p dv (2. . The internal energy of the fluid is part of the third component. the reciprocal of the density. potential. the work done is p dA ds.e. If this force displaces the surface a differential distance dg in the direction of the force. but in which the sole effect on the surroundings is the raising of a weight. equation (2. work is defined as the effect that is produced by a system on its surroundings when the system moves the surroundings in the direction of the force exerted by the system on its surroundings. (2) potential energy. a property of the fluid. Thus. Consider a small element of the surface dA of a fluid system. nuclear. Thermodynamics deals with phenomena considerably more complex than covered by this definition from mechanics.2 (2. The force acting on dA due to the fluid in the system isp dA. The reader should not confuse this use of the symbol v with the y component of velocity. Since we are normally only concerned with changes in energy rather than its absolute value.24b). Differential displacements are assumed so that the process is reversible. therefore. + j. The energy is usually divided into three components: (1) kinetic energy.23) Note that Ue is the symbol used for specific internal energy (i. Thus. the internal energy per unit mass).
Consider the fluid particles shown in Fig. (b) heat transfer to a twodimensional element.23) in rate form. 2 I Fundamentals of Fluid Mechanics DERIVATION OF THE ENERGY EQUATION 2.15 Heattransfer and flowwork terms for the energy equation for a twodimensional fluid element: (a) work done by stresses acting on a twodimensional element. we are now ready to derive the differential foñn of the energy equation for a viscous. . heatconducting compressible flow. T vrxy — a 2 — a 2 x (a) y ax ax ax 2 4y+ V ax ax ax 2 ay)T x (b) Figure 2.74 Chap.9 Having discussed the first law and its implications. we can describe the energy balance on the particle as it moves ajong in the flow: pq — pw = d = d + d + (2.15. Writing equation (2. 2.25) y 2 + ba T + a —.
the displacement per unit time) at that surface.27). the symbol t. 2. Thus. It is accounted for in the potential energy term. u— + ii— = dx v dx t3Tyx up— dt du — (2. The rate at which work is done by the system on its surrounding is equal to the negative of the product of the forces acting on a boundary surface times the flow velocity (i. we shall again consider a twodimensional flow. if the temperature is increasing in the outward direction. To simplify the illustration of the energy balance on the fluid particle. and Q. 2. Note that the symbol T will be used to denote temperature.26c) + v— OTyy dv vp— — vpf. the total heat flux rate.15b and noting that. using the nomenclature of Fig.15a.15.11). 2.e. of aT\ al aT\ 0 = +—( k— JtIxAy + —1k— OyJ dx! .9 / Derivation of the Energy Equation 75 where the overdot notation denotes differentiation with respect to time. and + iXy Using the constitutive relations for dividing by we obtain —pw = given earlier in this chapter and IOu dv\ + dTyy 2 — + u—) j [IOU L — dv\21 Ox I J dx dTxy OTyx \ 3y Oy dx (226a) From the component momentum equations (2. Q——khA•VT we can evaluate the rate at which heat is added to the system (per unit depth). The work done by the body forces is not included in this term. heat is added to the particle (which is positive by our convention). Recall that the substantial (or total) derivative is d dt 0 —* dt and therefore represents the local. 2. di' From Fourier's law of heat conduction. timedependent changes.. we can evaluate the rate at which work is done by the system (per unit depth): —* = + + Txy.26b) (2. Tyx. Referring to Fig. as well as those due to convection through space.Sec. see equation (2. time. as shown in Fig.
27). dp + —p (2.3).76 Chap. = Substituting t3x) + ayJ (2. 2 I Fundamentals of Fluid Mechanics Therefore. dp dtpdt d(p/p) dt pdp —pV V = For a conservative force field. we obtain I7au\2 (av\21 a / 3T\ a / aT\ —lk—)+—lk—J+2p1 (—I +1—I ax) ay) [\t3xJ J 2 2 F/au av\21 du dv d[(u2 + v2)/2] = dt + d(pe) dt + d(iie) dt (2.25). we obtain ax\ ax) + + + ayj av\21 [\axl d(p/p) dt dp j du 3 [(au +d+pud+PvdPuhPvfY + dv — — dv = Pu du + d(Ue) dt (2.29) . as introduced in equation (3.28a) d(pe) = pV VF — (2. ldp v.27) From the continuity equation.26d) equations (2.28) into equation (2.28b) where Fis the bodyforce potential and VF = —f. Substituting equations (2.26) into equation (2.v=p dt and by definition d(p/p) dt Thus.
a I aT'\ ay 8x) \ äy I L /av\2 I—I +1— \äy + + = — (2.33) That is.1 Integral Form of the Energy Equation The integral form of the energy equation is — = + if epV.hdA (2.30) and combining terms.Sec. it is convenient to introduce a symbol for this sum. . This is the energy equation for a general. 2.32a) + [/'ati L av\2 lay aw'\2 ay I&w az I I J (232b) Equation (2. compressible flow in two dimensions. the net rate heat is added to the system less the net rate work is done by the system is equal to the time rate of change of energy within the control volume plus the net efflux of energy across the system boundary.32b) defines the dissipation function 4). 2. too. Note that the heat added to the system is positive. The process can be extended to a threedimensional flow field to yield p—V(kVT)+çb where + 4) = + (2.30) p where h is called the specific enthalpy. which represents the rate at which work is done by the viscous forces per unit volume.9 / Derivation of the Energy Equation Since the terms Ue and 77 p/p appear as a sum in many flow applications. So.31). Conversely. is the work done by the system. heat transferred from the system or work done on the system is negative by this convention. Let us introduce the definition that h = tie + can write equation (2.29) as (2.9. Using equation (2. we a 7 aT\ ax \.
9. The .34b) Note that the change in kinetic energy during a process clearly depends only on the initial velocity and final velocity of the system of fluid particles. the potential energy per unit mass is given by pe = gz (2. 2.3 Flow Work Flow work is the work done by the pressure forces on the surroundings as the fluid moves through space.34a) Let us further examine the terms that comprise the energy of the system. Let us consider further the term for the rate at which work is done. whereas the changes in the energy components are a function of the states. and shaft work rate (Wa). viscous work rate (Wv).16. Kinetic energy e= ke + pe + ite (2. Consider flow through the streamtube shown in Fig. we obtain — = + gz + ue)d(vol) + if + gz + (2. That is. the total work rate is divided into flow work rate (Wf). The kinetic energy per unit mass is given by ke = V2 (2. For convenience. the amount of heat transferred and the amount of work done during a process are path dependent. 2. Internal energy (ue): energy associated with the internal fields and the random motion of the molecules Thus.2 Energy of the System As noted earlier. They are usually grouped as follows: (ke): energy associated with the directed motion of the mass 2.34) into equation (2. the changes depend not only on the intial and final states but on the process that takes place between these states.Chap.33). Potential energy (pe): energy associated with the position of the mass in the external field 3. the energy of the system may be written as 1.34c) Note that the change in the potential energy depends only on the initial and final elevations. the change in internal energy is a function of the values at the endpoints only. the energy of the system can take a variety of forms.35) It should be noted that. 2 I Fundamentals of Fluid Mechanics 2. Assuming that the external force field is that of gravity. Substituting equations (2. W. Furthermore.9.
The rate at which viscous work is done by the system over some incremental area of the system surface (dA) is = —i•VdA (2. Thus. and the assumed velocity represents movement of the fluid particles in that direction.36). Thus.16. the force acting on the right end surface is —p2n2 dA2. it can be shown that the flow work done on the surrounding fluid at the upstream end (station 1) is Wjç1 ——p1Vj. the velocity and the area vectors are in the same direction.16 Streamtube for derivation of equation (2. Recall that the pressure is a compressive force acting cm the system of particles. the pressure is replaced by the viscous shear. Thus.IuidA1 P1 (2.9 / Derivation of the Energy Equation V2 79 P2 (downstream) (upstream) Figure 2. it is negative by our sign convention.36b) The negative sign results because the pressure force is compressive (acts on the system of particles).9. the dot product is consistent with the convention that work done by the system on the surroundings is positive. the work done is the dot product of the force times the distance: p2ii2dA2'V2At Wf.e.36a) The positive sign is consistent with the assumed directions of Fig. 2. pressure P2 acts over the differential area ii2dA2 at the right end (i. In moving the surrounding fluid through the distance V2 At for the velocity shown in Fig. In the case of viscous work.4 Viscous Work Viscous work is similar to flow work in that it is the result of a force acting on the sur face bounding the fluid system.Sec. since work is done by the surroundings on the system at the upstream end.2 = P2V2 fl2 dA2 = —p2V2 P2 fl2 dA2 (2. that is..36c) . In a similar manner. the downstream end) of the control volume. the system does work on the surroundings (which is posi live by our sign convention). 2.16. however. 2. 2. Thus.
2 / Fundamentals of F'uid Mechanics 2. inviscid. If water drains to the atmosphere at station 2 at the rate of 0. For a steady.37) can be written if p(3.5: A flow where the energy equation is Bernoulli's equation Consider the steady.9.+ gz + where the definition for the enthalpy has been used. and there are no perceptible changes in the internal energy. 2. has been incorporated into the second integral of the righthand side of equation (2. = 0 (2. No heat transfer Q = 0 No shaft work No viscous work Steady flow at = 0 = 0 0 = . as represented by equation (2. 2. the energy equation can be written — = + Jf/ + gz + Ue)d(VOl) + gz + Ue + (2.6 Application of the Integral Form of the Energy Equation Thus.17.5 Shaft Work Shaft work is defined as any other work done by the system other than the flow work and the viscous work.37) Note that the flow work.80 Chap. This usually enters or leaves the system through the action of a shaft (from which the term originates).9. which either takes energy out of or puts energy into the system.36). equation (2. Since a turbine extracts energy from the system.37). the surroundings are doing work on the system and is negative. the system does work on the surroundings and W5 is positive.37) to the control volume that encloses the fluid in the pipe between stations 1 and 2. In the case where the shaft is that of a pump. adiabatic flow (Q = 0) with no shaft work (W3 = 0) and with no viscous work = 0). Solution: We will apply equation (2. Applying the conditions and assumptions in the problem statement. onedimensional flow of water in the curved pipe shown in Fig.38) EXAMPLE 2.OOlir m3/s what is the static pressure at station 1? There is no shaft work or heat transfer.
Thus.ThUS. Thus. It is + 2 + Uel + P1 /112 )P1V1A1 2 + gz2 + Ue2 + P2 By continuity.9 I Derivation of the Energy Equation 5 cm 30cm I (1) Drains to the atmosphere Figure 217 Pipe flow for Example 2. we are told that there are no perceptible changes in the internal energy (i.02)2]/4 10 rn/s (D2\2 V1 = V2 = 0.Sec. steady..uei = Ue2).6 rn/s \ ''11 . p1V1A1 p2V2A2.16(10) = 1. for an incompressible. onedimensional flow are uniform over the plane of each station and since the velocities are perpendicular to the crosssectional area. Pi Ti2 V2 P2 Note that the resultant form of the energy equation for this flow is Bernoul li's equation. Since water is incompressible. 2.e. V2 Q {ir(0. nondissipative flow. fff'v2 if Since the properties for the inviscid.5. we have a mechanical energy equation which simply equates the flow work with the sum of the changes in potential energy and in kinetic energy. the integral can readily be evaluated. Pi = P2• Fur thermore.
1 a(prvr) r +— r 1 a(pv0) =0 2. T11e dependent variables. density. and enthalpy For many applications. is flowing radially outward from a source in a plane such that V = Note that V9 = = 0. thermal conductivity. internal energy. 2 / Fundamentals of Fluid Mechanics Thus. that is. equation (2. simplifying assumptions can be introduced to ehminate one or more of the dependent variables For example.10 SUMMARY Both the differential and integral forms of the equations of motion for a compressible. we can eliminate them as variables and still obtain solutions of suitable accuracy Examples of constantproperty flows will be worked out in subsequent chapters. Presented were the continuity equation.28 2.(pV) —./cm3. or unknowns.3.2) or (2. and the thermal conductivity are constant. temperature. 0 where V = er— + 3r az in cylindrical coordinates. (a) Water.5). the momentum equation.Thus. and the energy equation.12) or (2. assuming that the density.82 Chap. 11 i\2 2 Pi + 9 8066(0 3\ ' + Patrn P1 = + Patm 1000 2 1000 ± (50 — 1. viscous flow have been developed in this chapter.which has a density of 1.0 g. PROBLEMS 2. in cylindrical coordinates. starting with the general vector form t3p + V. equation (2. Which of the following flows are physically possible. V = + u——. in these equations include pressure. equation (2. Note also that.+ 43r rao az .1. gage 2.94)1000 = 2 4.2. a common assumption is that the in the fluid properties as the fluid moves through the flow field are very small when the Mach number is less than 0. velocity.32) or (2.13).33).58 X N/rn2. Note also that and + are not zero. Derive the continuity equation in cylh!drical coordinates. satisfy the continuity equation? Substitute the expressions for density and for the velocity field into the continuity equation to substantiate your answer. the viscosity. viscosity.
3.5.4 m.Prob'ems 83 (b) A gas is flowing at relatively low speeds (so that its density may be assumed constant) where 2xyz (x2 + y2) (x2 — 2U00L y2)z x2 +y Here and L are a reference velocity and a reference length. What type of flow might this velocity field represent? 2.4.2 m. V2.e. d2 = 0. respectively. andd3 = O. that the flow is one dimensional).2. Assuming that the velocities at stations 1. Does this velocity field satisfy the continuity equation for incompressible flow? Transform these velocity components into the polar components Vr and v8 in terms of r and 0. and V3. Consider a onedimensional steady flow along a streamtube. use the integral form of the continuity equation to calculate the velocities. . A twodimensional velocity field is given by Ky U—2 x+y 2 x+y Kx 2 where K is a constant. Two of the three velocity components for an incompressible flow are: u=x2+2xz v=y2+2yz What is the general form of the velocity component w(x.6. For the twodimensional flow of incompressible air near the surface of a flat plate. Differentiate the resultant integral continuity equation to show that dp p dA A dV V For a lowspeed. at a constant volumetric flow rate of 0. as shown in Fig. Water flows through a circular pipe. 2.. 2.7. constantdensity flow.y.6m. and 3 are uniform across the cross section (i. The velocity components for a twodimensional flow are C(y2 — x2) U — —2Cxy (x2 + y2)2 (x2 + y2)2 V — where C is a constant. Does this flow satisfy the continuity equation? 2. the steamwise (or x) component of the velocity may be approximated by the relation U = y — y3 Using the continuity equation. what is the relation between the change in area and the change in velocity? 2. V1.5 m3/s.8. The corresponding diameters are d1 = 0.z) that satisfies the continuity equation? 2. P2. what is the velocity component v in the y direction? Evaluate the constant of integration by noting that v = 0 at y 0.8.
9.31. one measures the velocity distribution well upstream of the airfoil and well downstream of the airfoil. The effects of viscosity are such that the velocity (it) may be considered constant across the cross section at the surface (s) and at station 1. A long pipe (with a reducer section) is attached to a large tank. what are the velocities at s and 1.19 . R2 the radius of the pipe at station 2. The diameter of the tank is 5. If the flow is incompressible. 2 / Fundamentals of Fluid Mechanics d3 Figure P2. two dimensional. what is the total volumetric flow rate (if V  del) across the horizontal surfaces (surfaces 3 and 4)? c I K Figure P2.0 m. and what is the value of U0? S 1 2 Figure P2.9.13 and 2. the diameter to the pipe is 20 cm at station 1 and 10 cm at station 2. as shown in Fig. P2. If the density is 0. Velocity profiles are measured at the upstream end (surface 1) and at the downstream end (surface 2) of a rectangular control volume.10. and r the radial coordinate.8 2.27 through 2. 2. as shown in Fig. The integral equations of motion can be applied to either a rectangular control volume or a control volume bounded by streamlines. but varies with the radius at station 2 such that = u0(i — where U0 is the velocity at the centerline. As part of this exercise.85 g/cm3 and the mass flow rate is 10 kg/s.Chap.10.9 A note for Problems 2.10 through 2. The drag force acting on an airfoil can be calculated by determining the change in the momentum of the fluid as it flows past the airfoil. P2. and steady.
what is the total volumetric flow rate (if V ci dA) across the horizontal surfaces (surfaces 3 and 4)? U= C Figure P2. what is the vertical dimension of the upstream station (He)? Streamline U= k c I ) Figure P2.11. If surfaces 3 and 4 are streamlines. P2.13.12. two dimensional. two dimensional.11 2. as shown in Fig.12 2. what is the vertical dimension of the upstream station (He)? =U(/HD ) C Streamline Figure P2.12. Velocity profiles are measured at the upstream end (surface 1) and at the downstream end (surface 2) of a rectangular control volume. P2. The flow is incompressible. and steady. The flow is incompressible.13 . Velocity profiles are measured at the upstream end (surface 1) and at the downstream end (surface 2) of the control volume shown in Fig. If surfaces 3 and 4 are streamlines. 85 Velocity profiles are measured at the upstream end (surface 1) and at the downstream end (surface 2) of the control volume shown in Fig. two dimensional. and steady. P2. If the flow is incompressible.11.13. and steady.Problems 2.
Water leaves through the upper surface at a rate shown by the parabolic curve.1. 1) at time t = 10? 2. What is the average velocity at this station? 2. Chap.1 m2? 1.18. are given in cubic meters per second per unit length along the duct. 2) at time t = 1? . The maximum values of both flow rates.15. Two of the surfaces of the duct are porous.5 m3Is of water leaves a rectangular duct. As shown in Fig. while it leaves through the front face at a rate that decreases linearly with the distance from the entrance. Water is added through the upper surface at a rate shown by the parabolic curve. shown in the sketch.0 m3/s 0. 1. determine the position along the duct where the average velocity of flow is a minimum.14 2.14. P2.1 m2? ft3 m3/s/unit length 1. The maximum value of both flow rates. What is the average velocity of water leaving the duct if it is 1.16. For the conditions of Problem 2.86 2. what is the total acceleration of a particle at (1.16 2. Given the velocity field V= (6 + 2xy + — (xy2 + lOt)j + 25k what is the acceleration of a particle at (3. What is the average velocity of the water entering the duct if it is 1. while it enters the front face at a rate that decreases linearly with distance from the entrance.16.14. are given in cubic meters per second per unit length along the duct. Consider the velocity field V=i in a compressible flow where p = p0xt.17. P2. 2 I Fundamentals of Fluid Mechanics One cubic meter per second of water enters a rectangular duct as shown in Fig.0 m long and has a cross section of 0. Two of the surfaces of the duct are porous.0 m long and has a cross section of 0.8). Using equation (2.5 ni3ls m3lslunit length Figure P2.14. shown in the sketch.5 m3/s/unit length Figure P2.
O) = f U00( 1 R2\ / — " TI )coSOer 1+ R2\ TIJsinoè9 where is the velocity of the undisturbed stream (and is.19). as shown in Fig. we can express the velocity field for steady. Consider the onedimensional motion of a fluid particle moving on the centerline of the converging channel. the area per unit depth) varies as 2y = 2/i x=O x =L Figure P2. Consider steady twodimensional flow about a cylinder of radius R (Fig. Use equation (2.19 2. therefore.8) and the definition that V= and ar + —— + T80 + + V= Figure P2. incompressible flow around the cylinder as V(r.Problems 87 2. Using cylin drical coordinates. inviscid. thus. P2.20.20 .19. P2. at points where T = R).e.20. Derive the expression for the acceleration of a fluid particle at the surface of the cylinder (i. a constant).The vertical dimension of the channel (and..
. twodimensional channel of height h (i. Panic strikes the crew and other passengers.05 N .24. is in m2. Oxygen leaks slowly through a small orifice from an oxygen bottle." Calculate how long it will take the cabin pressure to drop. 2. 2. The mass flow rate is given by m02 = —0. Assume that the temperature in the tank remains Constant at 18°C and that the oxygen behaves as a perfect gas. If u at x = 0 is 2 mIs. and is in lbmls. we will assume that the viscosity and the density are constant.6847 Po2 . Furthermore. Make the following assumptions: (i) The air in the cabin behaves as a perfect gas: the cabm.5L. A hole is in ft2.21). Furthermore.e. But you leap up and shout. = is in K. shooting a small hole in the airplane. Make the following assumptions: (i) The air in the cabin behaves as a perfect gas: Pc = where the subscript c stands for the cabin." Calculate how long it will take the cabin pressure to drop. ._______ 88 Chap. 2. P2. ft.22. Fig. h is 1 m. is in °R. Assume the body forces to be negligible.1 mm. diameter. "Do not worry! I am an engineering student and I know that it will take seconds for the cabin pressure to drop from 0. The crew refuses the demands of the terrorist and he fires his pistol.25 X N/rn2.5 X N/rn2 to 0. (The units are those of Prob.3 in. and L = 1 m.0 psia to 3. Use the integral form of the continuity equation to describe the velocity along the channel centerline. Consider steady. Panic strikes the crew and other passengers. and is in kg/s. calculate the acceleration when x 0 and when x = Q. shooting a small hole in the airplane. ibm °R (ii) The volume of air in the cabin is 2513 ft3. = 22°C and is constant for the whole time. while the lower plate (which is at y = 0) isstationary. But you leap up and shout. R = 53. 2 / Fundamentals of Fluid Mechanics Assume that the flow is steady and incompressible. The bullet hole is 0. R = 287. 2. You are relaxing on an international flight when a terrorist leaps up and tries to take over the airplane.1 mm and the diameter of the orifice is 0. lowspeed flow of a viscous fluid in an infinitely long.21. the flow is fully developed. The crew refuses the demands of a terrorist and he fires the pistol._____{Ah018] V R02T0. lbf . The bullet hole is 0.34 where the subscript c stands for = 80°F and is constant for the whole time.75 cm in diameter. (iii) Air escapes through the bullet hole according to the equation: = C {Aho!e] where is in lbf/ft2. How long does it take for the pressure in the tank to decrease from 10 to 5 MPa? 2. The volume of the bot tle is 0. (iii) Air escapes through the bullet hole according to the equation: where Pc is in N/rn2. (ii) The volume of air in the cabin is 71.5 psia. The upper plate (which is at y = h) moves in the x direction at the speed V0.23.0 m3. "Do not worry! I am an engineering student and I know that it will take seconds for the cabin pressure to drop by a factor of two from 7.24). Since this is a lowspeed flow. Also determine the corresponding axial acceleration. m/kg K.
U0. P2. and h. equation (2. h. p. v. if u = U = U0 0 at y = h/2.. dp/dx. and the upper half (i.25 2. h/2 y t t___ h/2 Figure P2.26. P2. incompressible flow between two parallel plates. The upper plate moves at velocity U0 to the right and the lower plate is tionary.. as shown in Fig.e. (b) State the conditions that must be satisfied by the fluid velocity at the walls and at the interface of the two fluids.13)] to show how the pressure drop per unit length dp/dx changes if the radius of the pipe were to be doubled while the mass flux through the pipe is held constant at the value rh. (a) State the condition that the shear stress must satisfy for 0 < y < h.e. (c) Obtain the velocity profile in each of the two regions and sketch the result for > (d) Calculate the shear stress at the lower wall. The pressure gradient is zero. Use the integral form of the momentum equation [i...25. Consider the fully developed flow in a circular pipe. y ) Fully developed Figure P2. The velocity u is a function of the radial coordinate only: = uCL(1 — where UCL is the magnitude of the velocity at the centerline (or axis) of the pipe. Neglect the weight of the fluid in the control volume and assume that the fluid properties are constant. The lower half of the region between the plates.26. andy. as shown in Fig.2. 0 y h/2) is filled with fluid of density P1 and viscosity (h/2 y h) is filled with fluid of density P2 and viscosity p.26 . Consider steady. Figure P2. (b) Write the expression for dp/dx in terms of p. and w (which satisfy the boundary conditions) as functions of U0.24 2.25.Problems 89 (a) Develop expressions for u. laminar.
V At the downstream end of the control volume (surface 2). shown in Fig. (This problem is an extension of Problem 2.28 2. You will need to calculate the vertical dimension of the upstream station (Hu).025c and Cd— d The pressure is (a constant) over the entire surface of the control volume.29.025c. what is the drag coefficient fOr the airfoil? The vertical dimension HD is 0.025c F1 I I u = U0..27 2.) At the upstream end (surface 1).10.28. and steady.025c.( ./H ) Figure P2. two dimensional. as.11. . (This problem is an extension of Problem 2. what is the drag coefficient for the airfoil? The vertical dimension H is O. 2 I Fundamentals of Fluid Mechanics Velocity profiles are measured at the upstream (surface 1) and at the downstream end (surface 2) of a rectangular control volume. 0cyH . If the flew is incompressible.27.(—y/HD) j___ Figure P2.29. The pressure is (a constant) over the entire surface of the control volume. P2. The pressure is (a constant) over the entire surface of the (This problem is a variation of Problem 2. P2. Velocity profiles are measured at the upstream end (surface 1) and at the downstream end (surface 2) of a rectangular control volume. Chap. If the flow is incompressible. and steady.90 2.) Streamline :1.27. H Streamline U0. Velocity profiles are measured at the upstream end (surface 1) and at the downstream end (surface 2) of the control volume shown in Fig. what is the drag coefficient? The vertical dimension H = 0.:. Surfaces 3 and 4 are streamlines. P2. If the flow is incompressible.) II Y H= O.27. two dimensional. (ytH) K. two dimensional.U0. as shown in Fig.28. and steady.
022c. )r I H=0. The pressure is control volume.025c p C H=0. If the flow is incompressible. what is the drag coefficient for the airfoil? The ver(a constant) over the entire surface of the tical dimension H is 0. and steady. and steady. two dimensional. two dimensional.30 2. Velocity profiles are measured at the upstream end (surface 1) and at the downstream end (surface 2) of the control volume shown in Fig. Velocity profiles are measured at the upstream end (surface 1) and at the downstream end (surface 2) of a rectangular control volume.25 cos C i 1 "— Streamline Figure P2.30. (This is an extension of Problem 2. as shown in Fig.31.Problems 91 where v(x.31 . Surfaces 3 and 4 are streamlines.25 (1) c 0Figure P2. P2.12. what is the drag coefficient for the i— Streamline U = t11 u P 1 / = — 0. P2. If the flow is incompressible.31.) = ) U [ I ®Y I H=0022c — x 0. y) and v0(x) are y components of the velocity which are not measured.30.025c Figure P2.29 2.
37.21) are path dependent..e. In Problem 2. moves through the standard sealevel atmosphere at 200 ft/s.. (b) A hypersonic transport flies at a Mach number of 6.34 235. etc. Note that all prop erties (p. the entropy change is SCSA(SCSB)+(SBSA) (a) Is the net entropy change (SC — SA) the same for both paths? (ii) Processes AC and ABC were specified to be reversible.8 m.33.0 at an altitude of 30 km.023c.20)] and the free stream Mach number [as given by equation (2.36.e. equation (2. (b) Describe what occurs physically with the piston/cylinder/air configuration during each leg of each cycle.e.34). . What is the freestream Reynolds number for the Mach 3 flight at an altitude of 20 km? (b) What is the characteristic freestream Reynolds number of an airplane flying 160 mi/h in a standard sealevel environment? The characteristic chord length is 4. What are the freestream Reynolds number [as given by equation (2.. following process ii) is SC SA Going via B (i. To illustrate the point that the two integrals in equation (2.e..13. P2. The system of air particles is made to undergo two cyclic processes. the entropy change in going from A to C directly (i. consider a system consisting of air contained in a piston/cylinder arrangement (Fig.4 m. — SA if the processes are irreversible? Does SC — 5A depend on the path if the process is irreversible? 2. whose characteristic length (i.25.92 Chap.32.32a) can be written as ds/dt = 0. since the processes are cyclic. inviscid flow. The pressure is (aconstant) over the entire surface of the control volume. what is the value of (j 6q — (d) Is the first law satisfied for this system of air particles? j Air Process i: Process ii: Figure P2.19)] for the following flows? (a) A golf ball. 2. T. p. (a) Assume that both cycles are reversible and determine (1) 6q and (2) f 6w for each cycle. 2. What assumptions were made in deriving equation (2. What is se. (This problem is an extension of Problem 2. for each cycle? (c) Using the answers to part (a). 2 / Fundamentals of Fluid Mechanics airfoil? The vertical dimension at the downstream station (station 2) HD = 0.0 ft. following process i).32)? 2. (a) An airplane has a characteristic chord length of 10. The characteristic length of the transport is 32.34.7 inches. its diameter) is 1. Show that for an adiabatic. undergo a net change of zero).) return to their original value (i.) 2.
93 Consider the wingleading edge of a Cessna 172 flying at 130 mi/h through the standard atmosphere at 10. compare the total temperature with the freestream static temperature for this flow. = where is the freestream static temperature and is the total (or stagnation point) temperature. onedimensional. NarkTC. Ed. Boundary Layer Analysis. Pti y 1 2 M2 Carefully note the assumptions made at each step of the derivation. Consider the wingleading edge of an SR71 flying at Mach 3 at 80. Under what conditions is this valid? REFERENCES Brune GW. Cary AM. Englewood Cliffs. Viscous Fluid Flow. adiabatic flow: H1 = h + and the equation for the entropy change for a perfect gas: s— = T — Rln— Pt p T1 and develop the expression relating the local pressure to the stagnation pressure: / P —=(1+ \. compare the total temperature with the freestream static temperature for this flow. Presented at Fluid and Plasma Dynamics Conf. If = 0. Rubbert PW. CA Bushnell DM. A new approach to inviwidflow/boundaiy layer matching.38. Calculation methods for compressible turbulent boundary layers. adiabatic flow. La Canada. 2005.2404 Btu/lbm °R. Boundary Layer Theory. 1998. Harris JE. NJ: PrenticeHall Schlichting H. 1974. If cAf. Is convective heating likely to be a problem for this aircraft? 2. NASA SP422 Schetz JA. steady. 7th Ed.000 ft. For a perfect gas. 1979..References 2. New York: McGrawHill White FM. 1993.40. Using the integral form of the energy equation for a steady.000 ft. Ed. 1977. adiabatic flow. New York: McGrawHill Wilcox DC. Start with the integral form of the energy equation for a onedimensional. 7th. onedimensional. 74601.2404 Btu/lbm °R. Is convective heating likely to be a problem for this aircraft? 2. H1 = + and H1 = For a perfect gas. CA: DCW Industries . Palo Alto. Using the integral form of the energy equation for a steady. = where and H1 is the freestream static temperature and T1 is the total (or stagnation point) temperature.39. Turbulence Modeling for CFD. AIAA Pap. = 0.
the inviscid region. and one in which viscous forces cannot be neglected. i. 3.e. once the inviscid flow field has been defined..1 INVISCID FLOWS As noted in Chapter 2. for many applications. viscous boundary layer adjacent to the surface. i.3 DYNAMICS OFAN INCOMPRESSIBLE. However. once the solution for the 94 .e. By using the term inviscid flow instead of inviscid fluid. For the majority of this text. In fact. Furthermore. solutions of the inviscid region can provide important design information. Let us use the term inviscid flow to describe the flow in those regions of the flow field where the viscous shearstresses are negligibly small. it can be used as boundary conditions for a thin. the viscous boundary layer near the surface). INVISCID FLOW FIELD As will be discussed in Chapter 14. There are no real fluids for which the viscosity is zero. there are many situations where the product of the viscosity times the shearing velocity gradient is sufficiently small that the shearstress terms may be neglected when compared to the other terms in the governing equations.. the analysis of the flow field will make use of a tworegion flow model (one region in which the viscous forces are negligible. the shearing stresses may be expressed as the product of the viscosity times the shearing stress velocity gradient. we emphasize that the viscous shear stresses are small because the combined product of viscosity and the shearing velocity gradients has a small effect on the flow field and not that the fluid viscosity is zero.
Let us assume that the viscous boundary layer is thin and thetefore has a negligible influence on the inviscid flow field. equation (2. 3. the equatiOn is —. so these equations apply to a compressible flow as well as to an incompressible one. These equations. are called the Euler equations.e.12) becomes du ph = ap — (3la) (3. In this chapter we develop fundamental conCepts for describing the flow around configurations in a lowspeed stream.2) becomes (for these assumptions) (3.3) Using the vector identity that (U2\ equation (3. The momentum equation is Euler's equation.Sec. the density is essentially constant when the gas particles in the flow field move at relatively low speeds or when the fluid is a liquid.e.2) No assumption has been made about density. the eiigineer may want to solve the boundarylayer equations and calculate the skin friction drag on the configuration. let us consider only body forces that are cOnservative (such as is the case for gravity). Further.2 I Bernoufli's Equation 95 inviscid flow fIeld is Obtained.2 BERNOULLI'S EQUATION As has been discussed.lc) In vector form. 3. in regions where the flow is inviscid). Inregions of the flow field where the viscOus shear stresses are negligibly small (i. the flow outside the boundary layer)..lb) — dw — — 3p (3. f=—VF and flows that are steady (or steady state): (3. 1 (3. (The effect of violating this assumpiioh will be discussed when we compare theoretical results with We will seek the solution for the inviscid portion of the flow field (i. derived in 1755 by Eulef.4) ..
7) The force potential most often encountered is that due to gravity. or (2) for a displacement along a streamline if the flow is rotational. we take the dot product of each term in equation (3. where V x V = 0). + F + constant (3. The result is / 2\I (4i + dF+ — V X (V X V). 3. can be derived from . Let us calculate the change in the magnitude of each of these terms along an arbitrary path whose length and direction are defined by the vector ds. Thus.ds = 0 (3.96 Chap. Therefore.6) Since each term involves an exact differential. and 4.8) u2 2 + gz + p constant (3.4) and the vector ds.9). equation (3. and for which 5. The body forces are conservative. it is not necessary to include the energy equation in the procedure to solve for the velocity and pressure fields.. 2.e. Let us take the z axis to be positive when pointing upward and normal to the surface of the earth. since V x (V x V) is a vector perpendicular to V. which is a form of the momentum equation. The force per unit mass due to gravity is directed downward and is of magnitude g. 3 / Dynamics of an Incompressible. Consideration of the subsequent applications leads us to use U (rather than V). Steady. if the flow is rotational. Inviscid Flow Field In these equations U is the scalar magnitude of the velocity V. Because the density has been assumed constant. for a flow that is 1. we consider only displacements along a streamline).9) 'is known as Bernoulli's equation. To do this.9) Equation (3. Inviscid.3). In fact.5) Note that. the last term is zero (1) for any displacement if the flow is irrotational (i. the first integral of Euler's equation is f u2 2 ± f dF + p p / = constant (3. Irrotational (or. referring to equation (3. so F= The momentum equation becomes gz (3. Incompressible.
and (3) a general point just outside the boundary layer.la. such as viscosity.e. the flow process would be called reversible. there is an interrelation between the total energy of the flow and the flow work.17). the changes in potential energy are negligible.1 that the flow around a wing in a groundfixed coordinate system is unsteady. 3.9) is identical to that given by equation (1. the flow can be made steady through the Gallilean transformation to the vehiclefixed coordinate system of Fig. Neglecting the change in potential energy. . Thus.9). (1) a point far from the vehicle (i. In the Undisturbed (3) (1)° (a) U3 (1)U. Bernoulli's equation is valid only for flows where there is no mechanism for dissipation. As a corollary. For aerodynamic problems. 2.2. Recall from the discussion associated with Fig. The nomenclature for these points is illustrated in Fig._______ Sec. Note that. the other can be uniquely determined provided that the flow does not violate the assumptions listed previously.This is as it should be..2 / Bernoulli's Equation 97 equation (2.9) may be written p+ = constant (3. which is the integral form of the energy equation.. 3. in deriving equation (3. for example. the pressure variation in a static fluid as given by equation (3. As indicated in Example 2.10) This equation establishes a direct relation between the static pressure and the velocity. if the acceleration is zero throughout the entire flow field. a stagnation point) (b) Figure 3. (2) a point where the velocity relative to the vehicle is zero (i. since the five conditions required for Bernoulli's equation are valid for the static fluid.=75!'! (2)U2 = (i. 3. if either parameter is known.1.e. the free stream). Note that. we have assumed that dissipative mechanisms do not significantly affect the flow.e.5. The equation can be used to relate the flow at various points around the vehicle.. we can not apply Bernoulli's equation to the flow depicted in Fig.1 Velocity field around an airfoil: (a) groundfixed coordinate system.37). In thermodynamics. equation (3.. 2.Thus. (b) vehiclefixed coordinate system. However. a stagnation point).
the surface pressure may vary markedly. (2).332N/m2 = 77. and. The purpose of the static ports is to sense the true static pressure of the free stream When the aircraft is operated through a large angle of attack range.lb. is the sum of the freestream static pressure and the freestream dynamic pressure which is designated by the symbol This statement is not true.The stagnation (ortotal)pressure. the static pressure is equal to the total pressure since the velocity at this point is zero. P3 p3  1 2P00U13 rr2 — m — 1 rT2 = 82. (2). the pressure sensed at the static port may be significantly different from the freestream static pressure. 3. The flUid at point 3 moves downstream at 25 rn/s relative to the groundfixed coordinate system. EXAMPLE 3.la moves through the air at 75 mIs at an altitude of 2 km. as a result.501 N/rn2 Point 2: = + = 79. 3.0066 kg/m3)(75 m/s)2 = 82. in Table 1. The velocity at point 3 is 100 m/s relative to is found directly the stationary airfoil.which is the constant of equation (3.2. and (3). The Pitot head has no internal flow velocity.10).1: Calculations made using Bernoulli's equation The airfoil of Fig. The totalpressure and the staticpressure lines can be attached to a differential pressure gage in order to determine . as shown in 3.501 N/rn2 + (1. lnvisdd Flow Field vehiclefixed coordinate system of Figure 3.o = °.2) can be used to obtain a measure of the vehicle's airspeed. and the pressure in the Pitot ttibe is equal to the total pressure of the airstrearn (p.). The resultant flow is steady.299 N/rn2 3. and (3)? Solution: To solve this problem. let us superimpose a velocity of 75 m/s to the right so that the airfoil is at rest in the transformed coordinate system.point (2). What are the values of the static pressure at points (1). In this vehiclefixed coordinate system. we can apply Bernoulli's equation to points (1). 1 — — I rr2 Note that at. 3 I Dynamics of an Incompressible.332 N/rn2 romLJ. the fluid "moves" past the airfoil.7846P5L 79.98 Chap.11) indicates that a Pitotstatic probe (see Fig. if the flow is compressible.3 USE OF BERNOULLI'S EQUATION TO DETERMINE AIRSPEED Equation (3.lb. Point 1: pc.
3 I Use of Bernoulti's Equation to Determine Airspeed Pitotstatic system 99 Pitot with separate static source Pt Total pressure Pt Static pressure ports Pressure indicated by gage is difference between total and static pressure. (The latter correction is included in the calibration of the airspeed instrument dials. and. the static pressure measurement is used to calculate the velocity at the (outside) edge of the boundary layer (i. the value of the static pressure measured at the wall is essentially equal to the value of the static pressure in the inviscid stream (immediately outside the boundary layer).. There can be many conditions of flight where the airspeed indicator may not reflect the actual velocity of the vehicle relative to the air. Indicated airspeed is equal to the Pitotstatic airspeed indicator reading as installed in the airplane without correction for airspeed indicator system errors but including the sealevel standard adiabatic compressible flow correction. incompressible flow. is appropriate. which is valid only for an inviscid. when the aircraft is operated through a large angle of attack range.2. the use of Bernoulli's equation. Indicated airspeed (lAS). 3.2 Pitotstatic probes that can be used to "measure" air speed. The definitions for various terms associated with airspeed are as follows: 1. the surface pressure may vary markedly.) . the measurements of the local static pressure are often made using an orifice flushmounted at the vehicle's surface. For instance. As a result. 3. as a result.e. It is appropriate because (as discussed in Chapter 2) the analysis of the ymomentum equation reveals that the static pressure is essentially constant across a thin boundary layer.Sec. Nevertheless. the airspeed using the value of the freestream density for the altitude at which the vehicle is flying: As indicated in Fig. the velocity of the inviscid stream). — = Figure 3. Although the orifice opening is located on the surface beneath the viscous boundary layer. the pressure sensed at the static port may be significantly different from the freestream static pressure.
From this reasoning.1 presents the EAS and the dynamic pressure as a function of TAS and altitude. standard atmosphere U. we must account for variations in air density. Equivalent airspeed is equal to the airspeed indicator reading corrected for position error.S.Chap. 4.6 464.5 iQ3 3. Position errors are most usually confined to the static source in that the actual static pressure sensed at the static port may be different from the free airstream static pressure. Table 3. 3.37 x 5. TABLE 3.23 x 2. True airspeed (TAS). Inviscid Flow Field 2.0 1.8 600 800 1000 1.73 x 600 800 1000 5.89 x 7.9 107.8 581.4 .55 x i03 232. The position error of the installation must be small in the range of airspeed involving critical performance conditions. To relate EAS and TAS requires consideration that the EAS coupled with standard sealevel density produces the same dynamic pressure as the TAS coupled with the actual air density of the flight condition.S.The freestream properties are those of the U.49 x 53. instrument error.59 >< io3 348.1 Dynamic Pressure and EAS as a Function of Altitude and TAS Altitude Sea level 10. The instrument error may be small by design of the equipment and is usually negligible in equipment that is properly maintained and cared for.2 qcx (N/rn2) EAS (km/h) 200 400 1. The equivalent airspeed (EAS) is the flight speed in the standard sealevel air mass that would produce the same freestream dynamic pressure as flight at the true airspeed at the correct density altitude.20 >< 161.02 x 1.0000psi) TAS (km/h) qcx (N/rn2) EAS (p = qco (N/rn2) (p = EAS (km/h) (km/h) 116.000 m 20.4 1.56 x 200 400 6. The result shows that the EAS is a function of TAS and density altitude. Equivalent airspeed (EAS). CAS is the result of correcting lAS for errors of the instrument and errors due to position or location of the installation.74 x i03 1.6 215. Standard Atmosphere (1976).000 m (p = l.02 x 4. it can be shown that TAS = EAS = airspeed air density PSL = standard sealevel air density. The true airspeed results when the EAS is corrected for density altitude. 3 / Dynamics of an Incompressible. Since the airspeed indicator is calibrated for the dynamic pressures corresponding to air speeds at standard sealevel conditions.70 x i04 3.43 x 269. and for adiabatic compressible flow for the particular altitude. Calibrated airspeed (CAS).38 X 2.
parallel stream of velocity in the room where the tunnel is located. One such dimensionless coefficient is the pressure coefficient. where the local velocity is zero. Nevertheless.. we can express the pressure coefficient in terms of the nondimensionalized local velocity. these experimentally determined static pressures. as shown in Fig.3.4 THE PRESSURE COEFFICIENT The engineer often uses experimental data or theoretical solutions for one flow condi tion to gain insight into the flow field which exists at another flow condition.11). the dimension normal to the paper is infinite). The model is a cylinder of infinite span (i. that is. at the stagnation point. 3. it is most desirable to present (experimental or theoretical) correlations in terms of dimensionless coefficients which depend only upon the configuration geometry and upon the angle of attack. the air accelerates from a reservoir (or stagnation chamber) where the velocity is essentially zero.0 for an inThus. as shown in Figure 3. Therefore.13) = 1.4 / The Pressure Coefficient 3. the region where the model is located is open to the room in which the tunnel is located. In flight tests and windtunnel tests. which is at the stagnation point . Note that the stagnationpoint value is independent of the freestream flow conditions or the configuration geometry. The pressure sensed at orifice 3. are used to gain insight to describe the fullscale flow field at other flow conditions. the test section) in a uniUsing a barometer located on the wall form. Thus.12). 3. = compressible flow.3.There are two pressure orifices flush with the windtunnel walls and two orifices flush with the model surface. pressure orifices. can be presented as the dimensionless pressure coefficient using equation (3. we can write the pressure coef ficient as (3. which are located beneath the viscous boundary layer. 1 Treating point 3 as a general point in the flow field. If we consider those flows for which Bernoulli's equation applies.e.e. Windtunnel data. EXAMPLE 3. we know that the barometric pressure in the room is 29. where scale models are exposed to flow conditions that simulate the design flight environment..2: Flow in an open testsection wind tunnel Consider flow in a subsonic wind tunnel with an open test section. which are located flush mounted in the surface.12)]. through a converging nozzle.Sec. exhausting into the room (i. Rearranging equation (3. which is defined as 1 ir2 (.5 in Hg. sense the local static pressure at the wall [p in equation (3.
Furthermore. we know that the pressure coefficient for point 4 is —1. the pressure in the room is Prnorn = 29.92 in Hg/atm ) Furthermore.51 lbf/ft2 29.22 lbf/ft2/atm\ = 2086. Inviscid Flow Field Converging nozzle Oo Reservoir (or stagnation chamber) Test section a 2 Tunnel exhausts toroomasa free jet Figure 3.1. What is the static pressure at point 4? What is the velocity of an air particle just outside the boundary layer at point 4? In working this example. we will assume that the variation in the static pressure across the boundary layer is negligible. since the nozzle exhausts into the room at subsonic speeds and since the streamlines are essentially straight (so that there is no pressure variation normal to the streamlines). of the cylindrical model. . Thus.5 in Hg. gage. Solution: As discussed in Chapter 1._________ 102 Chap. It is also equal to 2116. is +2.22 lbf/ft2 Since the barometric pressure is 29. What is the pressure sensed by orifice 1 in the stagnation chamber? What is the pressure sensed by orifice 2. 3 / Dynamics of an Incompressible. the pressure in the room is equal to the freestream static pressure for the test section and is equal to the static pressure in the exit plane of the nozzle (p2).3 Open testsection wind tunnel used in Example 3.0 in of water. Thus.2.92 in Hg = 2116. located in the exit plane of the windtunnel nozzle? What is the freestream velocity. tl1e static pressure at the wall is approximately equal to the static pressure at the edge of the boundary layer. an equivalency statement may be written 760 mm Hg 29.2. as will be discussed in Section 4.22 lbf/ft2. the standard atmospheric pressure at sea level is defined as that pressure which can support a column of mercury of 760 mm high.5 in (2116.
14) . Bernoulli's equationis applicable. the circulation is given by T (3.51 Ibf/ft2 We will assume that the temperature changes are negligible and. (2116.22 lbf/ft2/atm gage in H20 is the column of water equivalent to 760 mm Hg if the density of water is 1.00234 lbf• s2/ft4 I = 94.937 slugs/ft3. that is.5 CIRCULATION The circulation is defined as the line integral of the velocity around any closed curve. But since = Pt' we can rearrange Bernoulli's equation: P3 P Pt Pa Equating these expressions and solving for = Since T2(1o.13): Thus.4. the freestream density is reduced proportionally from the standard atmosphere's sealevel value: = (2086. U4 = 139.22 ft/s is the dynamic pressure.387) = 2074. 3.51 + (—1. 3.002376 slug/ft3\ slu I = 0.75 ft/s. it is the difference between the stagnation pressure and the freestream pressure.5 / Circulation Proorn 103 = Pcx P2 2O86.2)(10.00234— \ 2116. 3.22 lbf/ft2 J ft3 Since the pressure measurement sensed by orifice 3 is given as a gage pressure. and we can use equation (3.Sec.387 lbf/ft2) V 0. therefore.05 lbf/ft2 Since we seek the velocity of the air particles just outside the boundary layer above orifice 4. we can rearrange the definition for the pressure coefficient to find the static pressure at point 4: + = 2086. Refering to the closed curve C of Fig. — 2 in H20.51 lbf/ft2)( (0.
Invisdd Flow Field Curve C Integration proceeds so that enclosed area remains on left V x Figure 34 Concept of circulation. —IXI' + (v + — y U y + Vt 4 f + Curve C x x (a) (b) Figure 3. A negative sign is used in equation (3..e. as shown in Fig. where V ds is the scalar product of the velocity vector and the differential vector — length along the path of integration. keeping the area on the left of the path). 3 / Dynamics of an Incompressible. the integration is carried out for the complete closed path. Integrating the velocity components along each of the sides and proceeding counterclockwise (i.5a. Consider the circulation around a small. (b) general curve C.104 y Chap. 3. so the area enclosed by the curve C is always on the left. The path of the integration is counterclockwise.14) for convenience in the subsequent application to liftingsurface aerodynamics.5 Circulation for elementary closed curves: (a) rectangular element. square element in the xy plane. As indicated by the circle through the integral sign. .
(See Fig.15) represents Green's lemma for the transformation from a line integral to a surface integral in twodimensional space.15) Equation (3. 3. The result for this general curve in the xy plane is (3. and equal in magnitude to the incremental surface area. then the line integral of V ds around the closed path is zero. Stokes's theorem is valid when A represents a simply connected region in which V is continuously differentiable.15) is a planar simplification of the more general vector equation. the integral of the normal component of the curl of the velocity vector over any surface A is equal to the line integral of the tangential component of the velocity around the curve C which bounds A.6 IRROTATIONAL FLOW V X V) is zero at all points in the region bounded by C. if the curl of (3.) Note that equation (3. If By means of Stokes's theorem.16). positive when pointing outward from the enclosed volume.5b.Sec.6. (A simply connected region is one where any closed curve can be shrunk to a point without leaving the simply connected region.6 / Irrotational Flow 105 vxv 4 V Figure 3.16) is not valid if the area A contains regions where the velocity is infinite. equation (3.) Thus. Simplifying yields = \\ax — t9yJ This procedure can be extended to calculate the circulation around a general curve C in the xy plane.16) where h dA is a vector normal to the surface. 3. equation (3.6 Nomenclature for Stokes's theorem. it is apparent that. 3. The transformation from a line integral to a surface integral in threedimensional space is governed by Stokes's theorem: = ff (V x i7) dA (3. such as that of Fig.17) . In words. 3.
3 I Dynamics of an Incompressible.19) is a valid representation of the velocity field for an irrotational flow can be seen by noting that Vx (3. the curl of any gradient is necessarily zero.7 KELVIN'S THEOREM Having defined the necessary and sufficient condition for the existence of a flow that has no circulation.. 3. Thus. Expanding equation (3. let us examine a theorem first demonstrated by Lord Kelvin. barotropic flow with conservative body forces. Therefore. A necessary and sufficient condition that be independent of path is that the curl of V is everywhere zero. Inviscid Flow Field and the flow contains no singularities. z) exists for this flow such that the partial derivative of 4) in any direction is the velocity component in that direction. The time derivative of the circulation along a closed fluid line (i. udx + vdy + wdz = —dx + —dy + —dz 84) 84) 84) It is apparent that (3. Stokes's theo rem leads us to the conclusion that For this irrotational velocity field.e. y.20) That is. the flow is said to be irrotational. an irrotational flow is termed a potential flow.19) Thus. A barotropic flow (sometimes called a homogeneous flow) is one in which the density depends only on the pressure. the circulation around a closed fluid line remains constant with respect to time. For an inviscid. However.18) in Cartesian coordinates. That equation (3. a fluid line that is composed of the same fluid particles) is .18) where d4) is an exact differential. Thus. the line integral is independent of path.106 Chap. a line integral can be independent of the path of integration only if the integrand is an exact differential. its value depends only on its limits. a velocity potential 4)(x. = d4) (3.
2). the negative sign is used for convenience.7. ( =0 . as discussed in equation (3. or if the velocity of the fluid in some region is uniform and parallel. the undisturbed. the order of time and space differentiation does not matter. Thus.23) Substituting equations (3. 3. The integral of an exact differential around a closed curve is zero. yields Using the constraint that the body forces are conservative (as is true for gravity. = /ds\ dV (3. all the terms on the righthand side involve exact differentials.22) and (3. (3. although the fluid particles in the subsequent .21) Again.24) Since the density is a function of the pressure only. Kelvin's theorem implies that.25) Or. the body force of most interest to the readers of this text).1 Implication of Kelvin's Theorem If the fluid starts from rest. equation (3. freestream flow is a uniform parallel flow in which there are no shear stresses.14). Kelvin's theorem leads to the important conclusion that the entire flow remains irrotational in the absence of viscous forces and of discontinuities provided that the flow is barotropic and the body forces can be described by a potential function.21) yields d V ds = — I Idp dF — Ic Jc P + Ic V dV (3.23) into equation (3. we have f=—VF and = —VF — (3. the rotation in this region is zero. the circulation remains constant along the closed fluid line for the conservative flow.22) where Fis the bodyforce potential. Since we are following a particular fluid particle.7 / Kelvin's Theorem 107 dr = dl V ds) ds + V (ds) (3. which is the momentum equation for an inviscid flow.Sec. as given in the statement of Kelvin's theorem. In many flow problems (including most of those of interest to the readers of this text). Euler's equation. 3.
3 I Dynamics of an Incompressible.e. so the equation of continuity for an incompressible.4) Note that equation (2.1 Boundary Conditions Within certain constraints on geometric slope continuity. the flow remains irrotational except in those re gions where the dissipative viscous forces are an important factor.4) is valid for a threedimensional flow as well as a twodimensional flow.26) Thus.19). 3. simply connected velocity field is uniquely determined by the distribution on the flow boundaries either of the normal component of the total velocity V4) or of the total potential 4). a bounded. the Neumann formulation will be used since most practical cases involve prescribed normal velocity boundary conditions.The flow need be two dimensional only in the sense that it requires only .8. a barotropic flow around the vehicle will remain irrotational provided that viscous effects are not important. For applications in this book. incompressible flows).8 INCOMPRESSIBLE. Combining equations (2. Thus. twodimensional flow is the necessary and sufficient condition for the existence of a stream function. These boundaryvalue problems are respectively designated Neumann or Dirichlet problems.19) For relatively low speed flows (i. which is known as Laplace's equation. Inviscid Flow Field flow patterns may follow curved paths.4) and (3. since the freestream flow is irrotational. the governing equation. one finds that for an incompressible. 3.108 Chap. V24) =0 (3. V= (3.9 STREAM FUNCTION Just INCOMPRESSIBLE FLOW as the condition of irrotationality is the necessary and sufficient condition for the existence of a velocity potential. irrotational flow. Specifically. the circulation must be constant around a path that moves so as always to touch the same particles and which contains no singularities. the continuity equation is (2. IN A TWODIMENSIONAL. the velocity may be expressed in terms of a potential function. secondorder partial differential equation of the elliptic type. For an irrotational flow. IRROTATIONAL FLOW Kelvin's theorem states that for an inviscid flow having a conservative force field. 3.. the flow tangency requirement associated with inviscid flow past a solid body is expressed mathematically as because the velocity component normal to the surface is zero at a solid surface. is a linear.
Incompressible Flow two spatial coordinates to describe the motion. that stream functions exist for compressible. if they are steady. Since i.Sec. v = dx (3. i/i is a constant along a streamline. the change in cli is zero along a streamline or.e.27b) A corollary to this is the existence of a stream function is a necessary condition for a physically possible flow (i. then.7. it is clear that the product v(—dx) represents the volumetric flow rate per unit depth across AG and the product u dy represents the volumetric flow rate per unit depth across GB. Therefore. The reader might note. that the volumetric flow rate (per unit depth) between any two points in the flow is the difference between the values of the stream function at the two points of interest.28a) and (3. twodimensional flow in Cartesian coordinates. v. 3.. Equating equations (3. It follows. stream functions exist both for plane flow and for axially symmetric flow. the definition of a streamline in a twodimensional flow is dx — dv V U— Rearranging. dcli = so dx + that thji = —v dx + u dy (3. By continuity. one that satisfies the continuity equation).28a) Since a streamline is a curve whose tangent at every point coincides with the direction of the velocity vector. although it is not relevant to this chapter. Referring to Fig. Examining the continuity equation for an incompressible. we find that dcli = 0 (3. the fluid crossing lines AG and GB must cross the .28b) along a streamline.27a) ml. A corollary statement is that lines of constant cl are streamlines of the flow. equivalently. twodimensional flows.9 / Stream Function in a TwoDimensional.li is a point function.28b). Thus. 3. it can be seen that udy—vdx——0 along a streamline.v = components can be calculated as —p du dx + dv 3y =0 for which the velocity It is obvious that the equation is satisfied by a stream function (3.
The velocity components for a twodimensional flow in cylindrical coordinates can also be calculated using a stream function as Vr = (3. we obtain = 0 (3. vxV=o Then writing the velocity components in terms of the stream function. Inviscid Flow Field y I/lB I/IA +dI/J A x Figure 3.The difference dtfi is the = volumetric flow rate (per unit depth) between the two streamlines.10 RELATION BETWEEN STREAMLINES AND EQUIPOTENTIAL LINES If a flow is incompressible. 3.30) Thus. Using the potential function.27). for an irrotational. as defined in equation (3. while a line can be passed through B for which = + dt/i (a different constant). the stream function is also governed by Laplace's equation. irrotational. the velocity field may be calculated using either a potential function or a stream function. The fact that the flow is always tangent to a and has no component of velocity normal to it has an important consequence.________________________ Chap. Any streamline in an inviscid can be replaced by a solid boundary of the same shape without affecting the remainder of the flow pattern. curve AB. incompressible flow. twodimensional.7 The significance of the stream function. 3 / Dynamics of an Incompressible.29a) and VU = — (3. the velocity components in Cartesian coordinates are . and two dimensional. is a measure of the volumetric flow rate per unit depth across AB.29b) If the flow is irrotational. A line can be passed through A for which = (a constant).Therefore.
32) U Refer to the discussion associated with equation (3. u =—— V (dy'\ 1—J (331) Since a streamline is everywhere tangent to the local velocity. the slope of a streamline. Therefore. since 4) = 2 for both points. for lines of constant potential (d4) 0). streamlines (4' = constant) are perpendicular to equipotential lines (4) = constant). EXAMPLE 3. which is a line of constant 4).32) yields 1—) (dy\ =— 1 (3.Sec. where the components vanish simultaneously.10 / Relation Between Streamlines and Equipotential Lines 34) u=— 34) 3y For a potential function.3: Equipotential lines and streamlines for a corner flow Consider the incompressible. (b) Sketch the streamline pattern and discuss the significance of the spacing between the streamlines. 3. except at stagnation points. where the stream function is 4) = 2xy (a) What is the velocity at x = 1. Comparing equations (3. irrotational. y = 1? At x 2. dçb = —dx 34) + —dy = udx + vdy 3y 34) Therefore. is = (3.33) The slope of an equipotential line is the negative reciprocal of the slope of a streamline.28a). twodimensional flow.31) and (3. (c) What is the velocity potential for this flow? (d) Sketch the lines of constant potential. y = Note that both points are on the same streamline. How do the lines of equipotential relate to the streamlines? Solution: (a) The stream function can be used to calculate the velocity components: u = — = 2x 34' v —dx 34) —2y .
Results are presented only for the first quadrant (x positive. the integral form of the continuity equation (2.8. Note that the x 0 and the y = 0 axes represent the = 0 "streamlineS" Since the flow is incompressible and steady. V= 21 — and the magnitude of the velocity is 2. y 4 2 8 0 2 4 x Figure 3. Inviscid Flow Field Therefore. y positive).8 Equipotential lines and streamlines for Example 3.Chap. 3 I Dynamics of an Incompressible. That is.3. y= 1. Mirrorimage patterns would exist in other quadrants.5) indicates that the product of the velocity times the distance between the streamlines is a constant. 0 Therefore. . since p= constant.8284 U= At x = 2. V 2x1 — 2yj 21. y= V = 41 — and the magnitude of the velocity is U= 4. 3.1231 (b) A sketch of the streamline pattern is presented in Fig. At x = 1. the distance between the streamlines decreases as the magnitude of the velocity increases.
e. any one of which may be considered to be a "boundary surface.12 I Elementary Flows 113 (c) Since u = and v= = Also. and a vortex. 3.10). The lines of equipotential are perpendicular to the streamlines. complex flow. 3. has been set equal to zero.12. The total potential function (or stream function) for the singularities and the onset flow are then used to determine the streamlines. There are numerous twodimensional and axisymmetric solutions available through "inverse" methods. and the velocity may be determined independently of the pressure. but instead assume a set of known singularities in the presence of an onset flow. which were suggested by Rankine in 1871. the reader can see that the pressure is a quadratic function of the velocity.8. 3. the equation of motion is not used.. the desired solution has been obtained.26) for the potential function and equation (3. These inverse methods do not begin with a prescribed boundary surface and directly solve for the potential flow. where C. the streamlines are straight and parallel to each other everywhere in the flow field . Bernoulli's equation can be used to calculate the corresponding pressure field. 3. Thus.11 SUPERPOSITION OF FLOWS Since equation (3. include a source.1 Uniform Flow The simplest flow is a uniform stream moving in a fixed direction at a constant speed.12 ELEMENTARY FLOWS 3. the arbitrary constant. the surface is a streamline). a sink.Sec.30) for the stream function are linear. since they are nonlinear functions of the velocity. Once the velocity field has been determined. Thus." If the resultant boundary surface corresponds to the shape of interest. f 2xdx + g(y) = fvdy + f(x) — = —f 2ydy + f(x) + C The potential function which would satisfy both of these equations is x2 — y2 where C is an arbitrary constant. (d) The equipotential lines are included in Fig. functions that individually satisfy these (Laplace's) equations may be added together to describe a desired. the velocity field can be determined using only the continuity equation and the condition of irrotationality. The boundary conditions are such that the resultant velocity is equal to the freestream value at points far from the solid surface and that the component of the velocity normal to the solid surface is zero (i. a doublet. It is important to note that the pressures of the component flows cannot be superimposed (or added together). For a constantdensity potential flow. Referring to equation (3. The singularities most often used in such approaches.
37) ervr + eove .2 Source or Sink A source is defined as a point from which fluid issues and flows radially outward (see Fig. 3 I Dynamics of an Incompressible. Inviscid Flow Field y Figure 3. because its axis extends infinitely far out of and far into the page. Using a Cartesian coordinate system. 3. The velocity field in cylindrical coordinates is V= since V ar + r30 (3. Such a twodimensional souTce is sometimes referred to as a line source. the potential function is 4) = cos a + y sin a) (3. (see Fig.9 is 4) (335a) For a uniform stream inclined relative to the x axis by the angle a.9).34) where is the velocity of the fluid particles.35b) 3. Using a cylindrical coordinate system. the potential function for a uniform flow moving parallel to the x axis is 4) cos 0 (3. 3.Chap.12. the potential function for the uiiiforrn stream of Figure 3. The potential function for the twodimensional (planar) source centered at the origin is 4) = 2ir (336) where r is the radial coordinate from the center of the source and K is the source strength.9 Streamlines for a uniform flow parallel to the x axis.10) such that the continuity equation is satisfied everywhere but at the singularity that exists at the source's center.
fluid flows into a sink along radial streamlines.38) Note that the dimensions of K are (length)2/(time). That is.Sec. and is independent of the radius. K and r öO Note that the resultant velocity has only a radial component and that this component varies inversely with the radial distance from the source. Thus.4: A twodimensional source Show that the flow rate passing through a circle of radius r is proportional to K. EXAMPLE 3. 4> = 2'n (3. for a sink of strength K centered at the origin. A sink is a negative source. 312 I Elementary Flows 115 t / Stream line Figure 310 Equipotential lines and streamlines for flow from a twodimensional source. the strength of the twodimensional source. Solution: th Jo I K p(—)rdo \271rJ / .
In general.Chap.39b) y 'I' = C3 Streamline = C3 x = C6 = C6 Figure 3. The potential for a twodimensional (line) doublet for which the flow proceeds out from the origin in the negative x direction (see Fig. 3 / Dynamics of an Incompressible. The line along which the approach is made is called the axis of the doublet and is considered to have a positive direction when oriented from sink to source.123 Doublet A doublet is defined to be the singularity resulting when a source and a sink of equal strength are made to approach each other such that the product of their strengths (K) and their distance apart (a) remains constant at a preselected finite value in the limit as the distance between them approaches zero. Inviscid Flow Field 3.11 Equipotential lines and streamlines for a doublet (flow proceeds out from the origin in the negative x direction). .11) is 4' = —cosO B r (3. = ——cosacosO B r (3. 3.39a) where B is a constant. the potential function of a line doublet whose axis is at an angle a relative to the positive x axis is 4.
12).Sec. one finds the velocity distribution about an isolated vortex to be Vr — 3r 134" 0 F Thus.12. there is no radial velocity component and the circumferential component varies with the reciprocal of the radial distance from the vortex. 3.40) where F is the strength of the vortex. 3. Differentiating the potential function.4 Potential Vortex A potential vortex is defined as a singularity about which fluid flows with concentric streamlines (see Fig. Note that the dimensions of F are (length)2/(time).12 Equipotential lines and streamlines for a potential vortex. The potential for a vortex centered at the origin is = 2ir (3. The curl of the velocity vector for the potential vortex can be found using the definition for the curl of V in cylindrical coordinates VxV=—— VT a ao a r 9r az Vz rv0 Equipotential lines Streamline Figure 3. . We have used a minus sign to represent a vortex with clockwise circulation.12 I Elementary Flows 117 3.
For solidbody rotation. the circulation is = C1 = Jo I I \ —I' j r1d0e0 = I Jo (—)—dO = 2ir Recall that Stokes's theorem [equation (3. solid body about its axis. we find that 2ir—€ 0 c2 o 2irr1 27r—E 2iir2 or = 0 Thus. which does not enclose the origin. Let us calculate the circulation around a closed curve C1 which encloses the origin. VI =0 V6 = rw . the circulation around a closed curve not containing the origin is zero. We find that vxV=o Thus.Chap. Using equation (3. Iriviscid Flow Field 9=0 (a) (b) Figure 3.e. we must remember that the velocity is infinite at the origin (i. 3 / Dynamics of an Incompressible. which encloses origin. when r = 0). The reader may be familiar with the rotation of a twodimensional. such as the rotation of a wheel. which does not enclose the origin.13 Paths for the calculation of the circulation for a potential vortex: (a) closed curve C1. (b) closed curve C2..14).13a. if we calculate the circulation around a closed curve C2.13b.16)1 is not valid if the region contains points where the velocity is infinite. such as that shown in Figure 3. although the flow is irrotational. However. We shall choose a circle of radius r1. as shown in Figure 3.
Or. Vortex lines (or filaments) will have an important role in the study of the flow around wings. which is taken from Campbell and Chambers (1994).Sec. therefore. Only through the action of viscosity (or some other dissipative mechanism) can they decay or disappear. The circulation around a given vortex line (i. 3. 3. In general. cannot be defined using a potential function.5 Summary of Stream Functions and of Potential Functions Table 3. the strength of the vortex filament) is constant along its length. we can conclude that vortices are preserved as time passes. Therefore. a fluid that is initially irrotational remains irrotational. Because a vortex filament cannot end in a fluid. in the absence of rotational external forces. 3.1 Velocity Field Consider the case where a uniform flow is superimposed on a doublet whose axis is par allel to the direction of the uniform flow and is so oriented that the direction of the . let us summarize the vortex theorems of Helmholtz.12..2 summarizes the potential functions and the stream functions for the elemen tary flows discussed previously. or go to infinity.14.13 ADDING ELEMENTARY FLOWS TO DESCRIBE FLOW AROUND A CYLINDER 3. One vortex can be seen entering the left engine. For a barotropic (homogeneous) inviscid flow acted upon by conservative body forces. end at a boundary. the following statements are true: 1. 3. Substituting these velocity components into the definition rL 3r we find that We see that the velocity field which describes twodimensional solidbody rotation is not irrotational and. Two vortices which are created by the rotational motion of jet engines can be seen in the photograph of Fig. and the downstream ends of the horseshoe vortices representing the loading on a threedimensional wing (see Chapter 7).e. equivalently. Examples of these three kinds of behavior are a smoke ring. A vortex filament cannot end in a fluid. No fluid particle can have rotation. The righthand vortex is in the form of a horseshoe vortex. 3. 2. if it did not originally rotate. a vortex bound to a twodimensional airfoil that spans from one wall to the other in a wind tunnel (see Chapter 6).13. the vortex axis turns sharply and the vortex quickly goes to the ground.13 / Adding Elementary Flows to Describe Flow Around a Cylinder where 119 is the angular velocity. It must form a closed path.
[Taken from Campbell and Chambers (1994). one finds that = = — TABLE 3.120 Chap.] efflux opposes the uniform flow (see Fig.15). Inviscid Flow Field Figure 3.2 Stream Functions and Potential Functions for Elementary Flows 4 Flow Uniform flow Source KO sin 0 cosO 2ir —mr 2w FO K Doublet Vortex (with clockwise circulation) F —i—: 90°cornerflow Solidbody rotation Axy lwr2 Does not exist .39a)] into the expression for the velocity field [equation (3. Substituting the potential function for a uniform flow [equation (3. 3 / Dynamics of an Incompressible.14 Ground vortices between engine intake and ground and between adjacent engines for Asuka STOL research aircraft during static tests. 3.34)] and that for the doublet [equation (3.37)].
= —= 0r B — r which is a constant. axis is perpendicular to the freestream direction. as noted earlier. incompressible flow is that around a cylinder of radius R whose. When B = 0 or ir (points B and A.15 Streamlines for the two elementary flows which. we see that the velocity at the surface of the cylinder is equal to v0 = sin 0 (3.. Since the velocity Note that 0 at every point where r = is always tangent to a streamline. .41a) + 7 yr = cos o( \\ 1 TJ } (3. 113 I Adding Elementary Flows to Describe Flow Around a Cylinder U0.Sec. and v. The resultant twodimensional. violate the noslip requirement). 3.42) Of course. Setting r = R. 121 Streamlines for a uniform flow Streamlines for a doublet Figure 3. Replacing B by R2UcYJ allows us to write the velocity components as V0 = sin R2 (3.16. irrotational (inviscid). Since the solution is for the inviscid model of the flow field. it is not inconsistent that the fluid particles next to the surface move relative to the surface (i. = 0. describe the flow around a cylinder. when superimposed.41b) The velocity field not only satisfies the surface boundary condition that the inviscid flow is tangent to a solid wall. but the velocity at points far from the cylinder is equal to the undisturbed freestream velocity Streamlines for the resultant inviscid flow field are illustrated in Fig.e. the fact that velocity component (vt) perpendicular to a circle of r = R is zero means that the circle may be considered as a streamline of the flow field.
When the air particles in the boundary layer.2 Pressure Distribution Because the velocity at the surface of the cylinder is a function of 0. the actual flow field is radically different from the inviscid solution described in the previous paragraphs. are presented in Fig.e. the local static pres sure will also be a function of 8.13. the flow is decelerated from the freestream velocity to zero velocity at the (windward) stagnation point in the plane of symmetry. 1 — 4 sin2 0 (3. the flow tangent to the leeward surface decelerates to a stagnation point at the surface in the leeward plane of symmetry (at 8 = 0°). not all boundary layers separate when they encounter an adverse pressure gradient. thus calculated. There is a . the surface facing the free stream). 3. we obtain the expression for the 0distribution of the static pressure using dimensional parameters: p= + — sin2 0 (3. respectively. which is presented in equation (3.e. 3.18 clearly illustrates the flow separation.16). 3 / Dynamics of an Incompressible.17 as a function of 0. Recall that. of Figure 3. therefore. in the nomenclature of this chapter. Using Bernoulli's equation [equation (3. which have already been slowed by the action of viscosity. 0 = 180° corresponds to the plane of symmetry for the windward surface or forebody (i. we have: C. Starting with the undisturbed freestream flow and following the streamline that wets the surface.44) The pressure coefficients.10)]. 3.12). The flow then accelerates along the surface of the cylinder.L. the fluid is at rest with respect to the cylinder (i. The photograph of the smoke patterns for flow around a cylinder presented in Fig. reaching a maximum velocity equal in magnitude to twice the freestream velocity. These points are. (Vr = V0 = 0). termed stagnation points.122 Chap. boundarylayer separation occurs. However. even though the viscosity of air is relatively small. Note that separation resuits when the fluid particles in the boundary layer (already slowed by viscosity) en counter an adverse pressure gradient that they cannot overcome.. Once the pressure distribution has been defined.. From these maxima (which occur at 0 = 90° and at 270°). it can be used to determine the forces and the moments acting on the configuration.43) Expressing the pressure in terms of the dimensionless pressure coefficient. However.16 Twodimensional. Inviscid Flow Field Figure 3. inviscid flow around a cylinder with zero circulation. encounter the relatively large adverse pressure gradient associated with the deceleration of the leeward flow for this blunt configuration.
] relation between the characteristics of the boundary layer and the magnitude of the adverse pressure gradient that is required to produce separation.7 x 2 1 0 cp —1 —2 Figure 3.86 X data of Schlichting (1968) data of Schlichting (1968) Supercritical Reynolds number (6.18 Smoke visualization of flow pattern for subcritical flow around a cylinder. 3. which has slowermoving particles near the wall for the same value of the edge velocity.) . courtesy ofT. Figure 3.Sec.17 Theoretical pressure distribution around a circular cylinder. J.44) — — — — Subcritical 123 Reynolds number (1. A turbulent boundary layer. used with permission of McGrawHill Book Company. (Photograph by F. [From Boundary Layer Theory by H. (Boundary layers are discussed in more detail in Chapter 4. the size of the wake. Schlichting (1968). the separation location. M. and the surface pressure in the wake region depend on the character of the forebody boundary layer. which has relatively fast moving particles near the wall.13 / Adding Elementary Flows to Describe Flow Around a Cylinder — . compared with data for a subcritical Reynolds number and that for a supercritical Reynolds number. Brown. Mueller of University of Notre Dame. N. equation (3.) Therefore.— Theoretical solution. would remain attached longer than a laminar boundary layer.
Due to the higher levels of energy for the fluid particles near the surface in a turbulent boundary layer. where the inviscid solution.] The experimentally determined separation locations for a circular cylinder as reported by Achenbach (1968) are presented as a function of Reynolds number in Fig. the forebody boundary layer is turbulent. As discussed in Chapter 2.. the Reynolds number is a dimensionless parameter (in this case.e. If the pressure were actually decreasing in the streamwise direction.124 Chap. the flow is able to run longer against the adverse pressure .e. At subcritical Reynolds numbers (i. [Data from Achenbach (1968). as given by equation (3.17. 3 / Dynamics of an Incompressible.. one for which the staflow. Inviscid Flow Field Subcritical Reynolds ( numbers 30 I 4 Critical Reynolds ) Reynolds numbers Supercritical numbers i t 60I 0100 10 I 00 0 0 0 0 0 0 0 120 1 I I 106 Red Figure 3.44) and presented in Fig. 800 from the windward stagnation point. Above the critical Reynolds number. 3. Note that the occurrence of o separation so alters the flow that separation actually occurs on the windward surface. the occurrence of separation alters the pressure distribution on the forebody (windward surface) of the cylinder. indicates that there still should be a favorable pressure gradient (i.19 Location of the separation points on a circular cylinder as a function of the Reynolds number. less than approximately 3 x tic pressure decreases in the streamwise direction). that is. Thus. separation would not occur. Red = that relates to the viscous characteristics of the the boundary layer on the windward surface (or forebody) is laminar and separation occurs for 1000.19. 3.
However. For supercritical Reynolds numbers. It is conventional to resolve the resultant force on the cylinder into a component perpendicular to the freestream velocity direction (called the lift) and a component par allel to the freestream velocity direction (called the drag). it affects the pressure in the separated region.13 / Adding Elementary Flows to Describe Flow Around a Cylinder 125 gradient. Achenbach observed an intermediate "separation bubble" with final separation not occurring until 8 = 400 (i. The subcriti cal pressurecoefficient distribution is essentially unchanged over a wide range of Reynolds numbers below the critical Reynolds numbers.20. the supercritical pressurecoefficient distribution is independent of Reynolds numbers over a wide range Of Reynolds numbers above the critical Reynolds number.e. the boundary layer is thin.13. because the character of the attached boundary layer affects the separation location. sepa ration occurs in the range 60° < 0 < 70° (The reader should note that the critical Reynolds number is sensitive both to the turbulence level in the free stream and to surface roughness. For the flow upstream of the separation location.5 x 106. 3. See Figure 3. separation is delayed and the pressure in the separated region is higher and closer to the inviscid level. indicating that the supercritical state of flow has been reached. 3.17 3. If the attached boundary layer is turbulent. 140° from the stagnation point). .17 for the cases where the forebody boundary layer is laminar (a subcritical Reynolds number) and where the forebody boundary layer is turbulent (a supercritical Reynolds number). and the pressurecoefficient distribution is essentially independent of the character of the boundary layer for the cylinder.) Experimental pressure distributions are presented in Fig. For Red> 1. the separation bubble no longer occurs.3 Lift and Drag The motion of the air particles around the cylinder produces forces that may be viewed as a normal (or pressure) component and a tangential (or shear) component. Neglecting the shear force dl —pRdOsinO =—pRdO cosO U00 Figure 3. The nomenclature is illustrated in Fig..Sec.20 Forces acting on a cylinder whose axis is perpendicular to the freestream flow. In the critical region. 3. Similarly.
Thus. yields i p2ir which is (dynamic pressure) —0. Instead of using equation (3. as a function of 0. the aerodynamicist might be more likely to use equation (3.47) yields p21T 1= — Jo I — Dividing both sides of this equation by the product times (area per unit span in the x plane).126 Since Chap. we shall consider only the contribution of the pressure to the lift and to the drag.48) Jo Both sides of the equation (3.5 I (1 — 4 sin2 6) sin 0 dO Jo = 0 which..46) It is not surprising that there is zero lift per unit span of the cylinder. note that the net force in any direction due to a constant pressure acting on a closed surface is zero. is the same result as was obtained by integrating the pressure directly.43). As shown in Figure 3.20 and following a similar procedure.5 / sin 0 dO (3.48) are dimensionless.44)] were obtained for an inviscid flow. we can calculate the drag per unit span for the cylinder in an inviscid flow.43) to define the static pressure as a function of 0.50) . As a result. p2w Jo I sin 0 R dO 0 (3. since the pressure distribution is symmetric about a horizontal plane.20. which is the expression for the dimensionless pressure coefficient. To do this.43).44) to define C. 3 I Dynamics of an Incompressible.45) and (3.44). The expression of the lefthand side is known as the section lift coefficient for a cylinder: C1 = (349) Using equation (3.42)] and for the pressure distribution [equation (3. p2w —0. of course. Referring to Figure 3. one finds that /= 0 (3. the drag per unit span is d — Jo / p cos 0 R dO (3.47) Adding equations (3. Inviscid Flow Field the expressions for the velocity distribution [equation (3. the lift per Unit span of the cylinder is p2ir 1= — Jo I p sin 6 R dO (345) Using equation (3. or (3. which is the expression for the static pressure.
the drag coefficient is essentially constant (approximately 1. it was noted that the subcritical pressurecoefficient distribution is essentially unchanged over a wide range of Reynolds number. Note that the actual pressure in the separated. A drag force that represents the streamwise component of the pressure force integrated over the entire configuration is termed pressure (or form) drag. wake region near the leeward plane of symmetry (in the vicinity of 0 = 0 in Fig. As a result. For blunt bodies.17) is much less than thetheoretical value. Recall that when we were discussing the experimental values of presented in Fig.000. (For streamlined bodies at small angles of attaék.. For Reynolds numbers below 300.51) A drag of zero is an obvious contradiction to the reader's experience (and is known as d'Alembert'sparadóx).2).13 / Adding Elementary Flows to Describe Flow Around a Cylinder 127 Substituting equation (3. It is the resultant difference between the high pressure acting near the windward plane of sythmetry (in the vicinity of 0 = 1800. However. Thus.52) Experimental drag coefficients for a smooth circular cylinder in a lowspeed stream [Schlichting (1968)] are presented as a function of Reynolds number in Fig. the friction drag is small.Sec. The drag force that is obtained by integraiing the streamwise component of the shear force over the vehicle is termed skin friction drag. 3.21.) Thus. 3. Since the pressure coefficient distribution for a circular cylinder is essentially independent of Reynolds number belOw the critical Reynolds number. the stagnation point) and the relatively 10w pressuresacthig near the leeward plane of symmetry which produces the large drag component. which is Reynoldsnumber dependent. significant form drag results because of the action of viscosity. Note that in the case of real flow past a cylinder. i. 3.17.e.43) for the local pressure. the difference betiveen the on the foreward surface and that acting on the leeward surface is less turbulent case. The drag coefficient per unit span for a cylinder is Cd = d (3. d= — f + — sin2 o) cos 0 R dO we find that d=0 (3. Cd . the pressure (or form) drag is the dominant drag component. independent of Reynolds number. the dominant component of drag is skin friction. the form drag for a turbulent boundary layer is markedly less than the corresponding value for a laminar (forebody) boundary layer. which causes the boundary layer to separate and therefore radically alters the pressure field. it follows that the drag coefficient would be essentially independent of the Reynolds number. 3. The pressure near the leeward plane of symmetry is higher (and closer to the inviscid values) when the forebody boundary layer is turbulent.
used with permission of McGrawHill Book Company. the effects of viscosity and compressibility cause variations in the force coefficients. Ideally. we recall that the pressure in the separated region is closer to the inviscid level.5  0 00 00 tilt) 4 II ReD I I I I x 106 Figure 3.] for the blunt body. the force coefficient would be independent of size and would be a function of configuration geometry and of attitude only. Such variations are evident in the drag coefficient measurements presented in this chapter. Reviewing the supercritical pressure distribution. The dimples on a golf ball are intended to reduce drag by reducing the form (or.pressure) drag with only a slight increase in the friction drag. 3 / Dynamics of an Incompressible. we can define a force coefficient as CF force (353) area) dynamic pressure Note that. inviscid Flow Field I 1.21 Drag coefficient for a smooth circular cylinder as a function of the Reynolds number. 3. Above the critical Reynolds number (when the forebody boundary layer is turbulent).14 LiFT AND DRAG COEFFiCIENTS AS DIMENSiONLESS FLOWFIELD PARAMETERS The formulas for the drag coefficient [equation (3. a force per unit span would be divided by the reference area per unit span. In a situation where the Reynolds number is subcritical. Examples of such transitionpromoting roughness elements are the dimples on a golf ball. These effects can be correlated in terms of parameters. [From Boundary Layer Theory by H.49)] have the same elements. it may be desirable to induce boundarylayer transition by roughening the surface. However. for a configuration of infinite span. the drag coefficient is significantly lower.0 Chap.128 2. such as the Reynolds number and the Mach number.52)] and for the lift coefficient for a cylin der [equation (3. Thus.0 dD 000 000c000 00000 00000 0 Cd 0. Schlichting (1968). .
Re1 = Same total drag Separation point (ci) Small cylinder with subcritical flow (diameter = d). (b). which are configurations having the same dimension and exposed Separation point Relative drag force (a) Flat plate broadside to the flow (height d). Compare the results for configurations (a).23.] . Red = i05 .r Separation point Skin friction drag component Pressure drag component (b) Large cylinder with subcritical flow (diameter = d). flow condition/configuration geometry combinations are presented in Fig. it is clear that an aerodynamic force is proportional to the square of the freestream velocity.14 / Lift and Drag Coefficients as Dimensionless FlowField Parameters 129 From equation (3. Red = i05 Separation point Larger (e) Large cylinder with supercritical flow (diameter = d).53)..22 Comparison of the drag components for various shapes and flows. Red = Figure 3. The actual drag for several incompressible. and (c). and to the force coefficient. [From Talay (1975).22 and 3.Sec. which are taken from Talay (1975). 3.22. to the size of the object. 3. 3. An indication of the effect of configuration geometry on the total drag and on the drag coefficient is given in Figs. Rea = point (c) Streamlined body (thickness = d). to the freestream density.
2 (d) Small cylinder with subcritical flow (diameter = 0. 3 / Dynamics of an Incompressible.2 (b) Large cylinder with subcritical flow (diameter = d). Red =  Separation point I (e) Large cylinder with supercritical flow (diameter = d). streamlining reduces the drag coefficient. the small cylinder . Streamlining produces dramatic reductions in the pressure (or form) drag with only a slight increase in skin friction drag at this Reynolds number. Inviscid Flow Field Cd = 2.ld). Red = Separation point Cd = 1. Red point (c) Streamlined body (thickness = d).23 comparison of section drag coefficients for various shapes and flows. Thus.] to the same Reynolds number stream. Therefore. in the same freestream flow as configuration (b). Red = Figure 3. [From Talay (1975).130 Chap. Note that the diameter of the small cylinder is onetenth that of the other configurations.0 (a) Flat plate broadside to the flow (height = d). Red = \Separation point I Cd=l.
As a result.5. the drag coefficients are essentially equal (see Fig. World War II airplanes. The air under the hut is at rest and the pressure is equal to the stagnation pressure. 3. 3. as shown in Fig. the denominator of equation (3.21). What is the net lift force acting on the quonset hut? The wind speed is 50 rn/s and the static freestream properties are those for standard sealevel conditions. . As shown in Fig. EXAMPLE 3. Neglect viscous effects and assume that the flow field over the top of the hut is identical to the flow over the cylinder for 0 8 ir.24. I 50 mIs = Pu — . 3. the total drag of the small cylinder is equal to that of the much thicker streamlined shape. Because the size is reduced. mounted on tiedown blocks. and vintage automobiles as well as more classical configurations. the drag coefficient for this condition (e) is only 0.1 hut Air at rest Pi Figure 3. However.24 Inviscid flow model for quonset hut of Example 3. the pressure drag is very large. neglect the presence of the air space under the hut. 3.21). over this range of Reynolds number.Sec. The student can readily imagine how much additional drag was produced by the interplane wire braëing of the World War I biplanes.22. However. whose radius is 5 m.6. In these volumes the reader will find aerodynamic coefficients for flags. Therefore. However. 3. the drag forces for (d) are an order of magnitude less than for (b). When the Reynolds number is increased to (corresponding to the supercritical flow of Fig. The quonset hut may be considered to be a closed (no leaks) semicylinder. even though the dimensional force is increased.5: Forces on a (semicylinder) quonset hut You are to design a quonset hut to serve as temporary housing near the seashore. Note that since the cylinder diameter is the same for both (b) and (e). the two order of magnitude increase in Reynolds number is accomplished by increasing the freestream density and the freestream velocity. which is less than the drag coefficient for the subcritical flow (b) even though the pressure drag is significantly greater. When calculating the flow over the upper surface of the hut. Hoerner (1958) and Hoerner (1975) offer the reader unique and entertaining collections of data.52) increases more than the numerator. the nondimensionalized force coefficient is decreased. There are a variety of sources of aerodynamic data.14 / Lift and Drag Coefficients as Dimensionless FlowField Parameters 131 operates at a Reynolds number of i04.
The lift per unit depth of the quonset hut is fir = — Jo I + = —R f + sin 0 + sin 0 — sin3 o) dO + = + + + = 2 Note that the lift coefficient is i 1  —  3 Since we have assumed that the flow is inviscid and incompressible. The actual lift force is I= kg"7 3 m 5 (5rn) = 40. 3 I Dynamics of an Incompressible.43): = + — sin2 0 The pressure on the lower surface (under the hut) is Equation (3.42): = = and the pressure is given by equation (3. the lift coefficient is independent of the Mach number and of the Reynolds number. which reflects the fact that we neglected the effects of viscosity. since it is a "flat plate" rather than a circular arc.20 can be used to calculate the lifting contribution of the upper surface but not the lower surface.45) and Figure 3. . the velocity is given by equation (3. Inviscid Flow Field Solutiän: Since we are to assume that the flow over the upper surface of the quonset hut is identical to an inviscid flow over a cylinder. the drag is zero.132 Chap.833 N/rn By symmetry.
however. Thus. twodimensional I p(cos6)RdO 0 body in an irrotational.13)].56) The resultant irrotational flow about the cylinder is uniquely determined once the magnitude of the circulation around the body is specified. the resultant potential function also represents flow around a cylinder. we obtain 1 =1 + 0 + (2FR)2] 3. Using the definition for the pressure coefficient [equation (3.55a) and 11 = 1ä4' = —t räO sin 0 — .15 / Flow Around a Cylinder with Circulation 133 3. The velocity at the surface of the cylinder is equal to Again. Forthis flow. the fact that the velocity component (vi) perpendicular to a circle of radius R is zero means that the circle may be considered as a streamline of the flow field. In any real twodimensional flow.55b) which is a constant and will be desig= 0 at every point where r = nated as R.1 Velocity Field Let us consider the flow field that results if a vortex with clockwise circulation is supe imposed on the doublet/uniformflow combination discussed above.2 Liftand Drag If the expression for the pressure distribution is substituted into the expression for the drag force per unit span of the cylinder.15.Sec.15. a . Vr + r — 2ir (3. — 2irjj F\ (3.15 FLOW AROUND A CYLINDER WITH CIRCULATION 3. d Jo The prediction of zero drag may be generalized to apply to any general. the streamline pattern away from the surface is not symmetric about the horizontal plane. steady. incompressible flow. The resultant potential function is = Thus.54) = cos 0 Bcos0 — r2 Bsin0 r (3. Since the velocity is always tangent to streamline. 3. v0 —U = — I' (3.
25) also depend on the circulation. the lift per unit span is directly related to the circulation about the cylinder. the forces acting are independent of the shape of the body and I The locatious of the stagnation points (see Fig. For a closed body. one obtains p2ir 1 = —J0 Thus.134 Chap. This result. Inviscid Flow Field drag force does exist. Since Vr = 0 at every point on the cylinder.58) about closed cylinders of arbitrary cross section. applies to the potential flow p(sinO)RdO (3. =0 (3. consider the circulating flow field around the closed configuration to be represented by the superposition of a uniform flow and a unique set of sources.25). which is known as the KuttaJoukowskj theorem. 3 I Dynamics of an Incompressible. Thus. F= only one stagnation point exists on the cylinder and it exists at 0 For this magnitude of the circulation. 3. 3. They are symmetrically located about the y axis and both are below the x axis (see Fig. the lift per unit span is 7 L ir — — Yr2 — Iiooi_iooi Poou — r The lift coefficient per unit span of the cylinder is I C1 Thus. continuity requires that the sum of the source strengths be equal to the sum of the sink strengths. the stagnation points occur when v9 = Therefore. and vortices within the body. To locate the stagnation points. sinks. I. Integrating the pressure distribution to detennine the lift force per unit span for the cylinder. the distance between the sources and sinks becomes negligible and the flow field appears to be that generated by a single doublet with circulation equal to the sum of the vortex strengths within the body. To see this. If 2700. which pro duce the shear force at the surface and which may also produce significant changes in the pressure field (causing form drag). we need to find where Vr = V0 = 0 0. = = (3.61) . drag is due to viscous effects. For incompressible flow.59) or = sin1 (— If I' there are two stagnation points on the surface of the cylinder. in the limit. When one considers the flow field from a point far from the surface of the body.
all of which must finally become numerical and make use of a computing machine. such as the KuttaJoukowski theorem.where r = 2irR I (1 .. . such as axisymmetric shapes in a uniform stream parallel to the axis of symmetry. can be represented by a source distribution along the axis of symmetry. '—I —4 Ce —8  —12 —16 I I I 180 270 0 6(0) 90 180 (c) Static pressure distributions Figure Stagnating streamlines and the static pressure distribution for a twodimensional circulating flow around a cylinder. 3.Sec. An "exact" solution for the flow around an arbitrary configuration can be approached using a direct method in a variety of ways. Flow fields for other elementary configurations. (a) F = (b) F = The value 4ir represents the maximum lift coefficient that can be generated for a circulating flow around a cylinder unless the circulation is so strong that no stagnation point exists on the body. The reader is referred to Hess and Smith (1966) for an extensive review of the problem. 3.16 / Source Density Distribution on the Body Surface 135 (b)F=4irlJJ? Static pressure for inviscid flow — — — —. using the inverse method.16 SOURCE DENSITY DISTRIBUTION ON THE BODY SURFACE Thus far. we have studied fundamental fluid phenomena.
Referring to equation (3. the velocity potential for the flow resulting from the superposition of the M source panels and the freestream flow is M k P cos a + sin a + (3. The configuration is represented by a finite number (M) of linear segments. we require that the sum of the sourceinduced velocities and the freestream velocity is zero in the direction normal to the surface of the panel at the surface of each of theM panels. or panels. At the control point of the ith panel. a source distribution on the jth panel causes an induced velocity whose potential at a point (x. x in the chordwise direction and y in the spanwise direction) will be used in subsequent chapters on wing and airfoil aerodynamics.62) where is defined as the volume of fluid discharged per unit area of the panel and the integration is carried out over the length of the panel ds1.z) = I mr (3.26 Source density distribution of the body surface.IT J . To determine the strengths of the various sources k1. Note also that r = \/(x or span. all calculations are for a unit length along the y axis.26. Consider a twodimensional configuration in a uniform stream. lnviscid Flow Field fli z for this panel) jth panel Control points Figure 3.The control points are chosen to be the midpoints of the panels. The effect of the jth panel on the flow field is characterized by a distributed source whose strength is uniform over the surface of the panel. z) is given by ç k ds1 4(x.136 Chap. Thus. — x1)2 + (z — z1)2 (3. 3 / Dynamics of an Incompressible.The coordinate system used in this section (i. such as shown in Fig. we need to satisfy the physical requirement that the surface must be a streamline.63) Since the flow is two dimensional.. 3.64) = in ds1 j=1 L.26.36). Each of the M panels can be represented by similar sources. 3. The points at which the requirement that the resultant flow is tangent to the surface will be numerically satisfied are called the control points. as shown in Fig.e.
Sec. 3.16 / Source Density Distribution on the Body Surface
137
where
is the distance from the control point of the ith panel to a point on the jth
=
—
panel.
x1)2
+
(z1 —
(3.65)
Note that the source strength k1 has been taken out of the integral, since it is constant over the jth panel. Each term in the summation represents the contribution of the jth panel (integrated over the length of the panel) to the potential at the control point of the ith panel. The boundary conditions require that the resultant velocity normal to the surface be zero at each of the control points. Thus,
=
0
(3.66)
must be satisfied at each and every control point. Care is required in evaluating the
spatial derivatives of equation (3.64), because the derivatives become singular when the contribution of the ith panel is evaluated. Referring to equation (3.65), we have
rj3 =
0
where j = i. A rigorous development of the limiting process is given by Kellogg (1953). Although the details will not be repeated here, the resultant differentiation indicated in equation (3.66) yields
/c
M
k1
IT
j=1
flj
(3.67)
where is the slope of the ith panel relative to the x axis. Note that the summation is carried out for all values of j except j = i. The two terms of the left side of equation (3.67) have a simple interpretation. The first term is the contribution of the source density of the ith panel to the outward normal velocity at the point (xi, z1), that is, the control point of the ith panel. The second term represents the contribution of the remainder of the boundary surface to the outward normal velocity at the control point of the ith panel. Evaluating the terms of equation (3.67) for a particular ith control point yields a linear equation in terms of the unknown source strengths k3 (for j = 1 to M, including j = i). Evaluating the equation for all values of i (i.e., for each of the M control points) yields a set of M simultaneous equations which can be solved for the source strengths. Once the panel source strengths have been determined, the velocity can be determined at any point in the flow field using equations (3.64) and (3.65). With the velocity known, Bernoulli's equation can be used to calculate the pressure field. Lift can be introduced by including vortex or doublet distributions and by introducing the Kutta condition; see Chapters 6 and 7.
EXAMPLE 3.6: Application of the source density distribution
Let us apply the surface source density distribution to describe the flow
around a cylinder in a uniform stream, where the freestream velocity is
138
Chap. 3 / Dynamics of an Incompressible, Inviscid Flow Field
U0.
x
Figure 3.27 Representation of flow around a cylinder of unit radius by eight surface source panels. The radius of the cylinder is unity. The cylinder is represented by eight, equallength linear segments, as shown in Fig. 3.27. The panels are arranged such
that panel 1 is perpendicular to the undisturbed stream.
Solution: Let us calculate the contribution of the source distribution on panel 2 to the normal velocity at the control point of panel 3. A detailed sketch of the two panels involved in this sample calculation is presented in Fig. 3.28. Referring to equation (3.67), we are to evaluate the integral:
(in
where i
3 and j = 2. We will call this integral 132, Note that
(— 0.38268, + 0.92388)
—
/
/
,..—
fl3
Control point
of panel3
—
(x3
0.00,
z
z3
= ± 0.92388)
r32
(—0.92388, +0.38268)
x
Figure 3.28 Detailed sketch for calculation of the contribution of the source distribution on panel 2 to the normal velocity at the control point of panel 3.
Sec. 316 I Source Density Distribution on the Body Surface
a
139
—
tar32
r32 an3
0n3
(x3
— x2)— + (z3
an3
8x3
az3
—
Z2)
=
8n3
(x3 — x2)2 + (z3
(368)
—
z2)2
where x3 = 0.00 and Z3 0.92388 are the coordinates of the control point of panel 3. Note also that
ax3
an3
= 0.00,
az3
0n3
= 1.00
Furthermore, for the source line represented by panel 2,
=
—0.92388
+ 0.70711s2
= +0.38268 + 0.70711s2
and the length of the panel is
12 = 0.76537
Combining these expressions, we obtain
[0.76537
(0.92388 — 0.38268
—
0.70711s2) ds2
—
'32
—
Jo
(0.92388 — 0.70711s2)2 + (0.92388 — 0.38268
0.76537
0.70711s2)2
This equation can be rewritten
'32 = 0.54120
o
ds 2
1.14645 — 2.07195s2 +
0.76537
—0.70711
J
0
s2as2
1.14645 — 2.07195s2 +
Using the integral tables to evaluate these expressions, we obtain '32 =
—0.70711' tan
—
2.07195s2
+
[
—
[
(2s2 — 2.07195\132=0.76537 \/0.29291 — 320
I
Thus,
'32 = 0.3528
In a similar manner, we could calculate the contributions of source panels 1,4, 5,6,7, and 8 to the normal velocity at the control point of panel 3. Substituting the values of these integrals into equation (3.67), we obtain a linear equation of the form
131k1 + 1321c2 + irk3 +
+ 133k5 + 136k6 + 137k7 + 138k8 = 0.00
(3.69)
The righthand side is zero since a
0 and
=
0.
Chap. 3 / Dynamics of an Incompressible, Inviscid Flow Field
Repeating the process for all eight control points, we would obtain a set
of eight linear equations involving the eight unknown source strengths. Solving the system of equations, we would find that
k1 = 2ITUc,o(+0.3765) k2 = k3 = 0,00
k5 =
= = 0.00
k8 =
Note there is a symmetrical pattern in the source distribution, as should be expected. Also, (370)
as must be true since the sum of the strengths of the sources and sinks (negative sources) must be zero if we are to have a closed configuration.
3.17
INCOMPRESSIBLE, AXISYMMETRIC FLOW The irrotational flows discussed thus far have been planar, two dimensional. That is, the
flow field that exists in the plane of the paper will exist in any and every plane parallel to the plane of the paper. Thus, although sketches of the flow field defined by equations (3.41) though depict the flow around a circle of radius R, in reality they represent the flow around a cylinder whose axis is perpendicular to the plane of the paper. For these flows, w 0 and 3/0z 0. Let us now consider another type of "twodimensional" flow: an axisymmetric The coordinate system is illustrated in Fig. 3.29. There are no circumferential variations in an axisymmetric flow; that is,
Figure 3.29 Coordinate system for an axisymmetric flow.
Sec. 3.17 / Incompressible, Axisymmetric Flow
141
v0—0 and Thus, the incompressible, continuity equation becomes
—+—+—=o 3r r 3z
Noting that r and z are the independent coordinates (i.e., variables), we can rewrite this expression as
+
=
0
(3.71)
As has been discussed, a stream function will exist for an incompressible, twodimen
sional flow. The flow need be two dimensional only in the sense that it requires only two spatial coordinates to describe the motion, The stream function that identically satisfies equation (3.71) is
—=rv,. and
3z
Thus, in the coordinate system of Fig. 3.29,
'Or
=
r3z
and
v = r3r
Z
(3.72)
Note that
3.17.1
= constant defines a stream surface.
Flow around a Sphere
To describe a steady, inviscid,, incompressible flow around a sphere, we will add the
axisymmetric potential functions for a uniform flow and for a point doublet. We will first introduce the necessary relations in spherical coordinates. For a spherical coordinate system,
v,.
=
3r
—
r 3w
v0
=
1
rsina 30
,
(3.73)
for an irrotational flow where V = V4. In equation (3.73), f represents the potential function, and r, 0, and w represent the independent coordinates. By symmetry,
v0=0 and
The velocity potential for an axisymmetric doublet is
4=+
B
where the doublet is so oriented that the source is placed upstream and the doublet axis is parallel to the uniform flow. The potential function for a uniform flow is
142
Chap. 3 / Dynamics of an Incompressible, Inviscid Flow Field
4)
cos
Thus, the sum of the potential functions is
4)
=
+
B
4'nr
The velocity components for this potential function are
aq5
aT
=
B
—
2irr
3cosw
(3.75a)
and
=
r 3w
=
—
B. ,sinw
(3.75b)
As we did when modeling the inviscid flow around a cylinder, we note that
VT
=0
when
r3 =
B
constant = R3
Thus, if B = we can use the potential function described by equation (3.74) to describe steady, inviscid, incompressible flow around a sphere of radius R. For
this flow,
=
and
1 — —i jcosw
rJ
2r
(3.76a)
_uoo(i +
(3.76b)
On the surface of the sphere (i.e., for r = R), the resultant velocity is given by
U=
1171 =
=
—
sin w
(3.77)
The static pressure acting at any point on the sphere can be calculated using equation (3.77) to represent the local velocity in Bernoulli's equation:
p=
+
—
sin2 (0)
(3.78)
Rearranging the terms, we obtain the expression for the pressure coefficient for steady, inviscid, incompressible flow around a sphere:
= I
—
(3.79)
Compare this expression with equation (3.44) for flow around a cylinder of infinite span whose axis is perpendicular to the freestream flow:
= I
—
4sin2O
Summary
143
*
2.0
Measurements for a smooth cylinder (see Fig. 3.21) o Measurements for a smooth sphere
I
Cdl.0
••
•****•.s..
S
0.5
0.1
I
4x103
Red
106
Figure 3.30 Drag coefficient for a sphere as a function of the Reynolds number. [From Schlichting, BoundaryLayer Theory (1968), with permission of McGrawHilL] Note that both 0 and o represent the angular coordinate relative to the axis, one for the twodimensional flow, the other for axisymmetric flow. Thus, although the configurations have the same cross section in the plane of the paper (a circle) and both are described in terms of two coordinates, the flows are significantly different. The drag coefficients for a sphere, as reported in Schlichting (1968), are presented as a function of the Reynolds number in Figure 3.30. The drag coefficient for a sphere is defined as = drag
CD
/4)
(3.80)
The Reynolds number dependence of the drag coefficient for a smooth sphere is similar to that for a smooth cylinder. Again, a significant reduction in drag occurs as the
critical Reynolds number is exceeded and the windward boundary layer becomes turbulent.
3.18
SUMMARY
In most flow fields of interest to the aerodynamicist, there are regions where the product of the viscosity times the shearing velocity gradient is sufficiently small that we may neglect the shear stress terms in our analysis. The momentum equation for these inviscid flows is known as Euler's equation. From Kelvin's theorem, we know that a flow remains irrotational in the absence of viscous forces and discontinuities
144
Chap. 3 / Dynamics of an Incompressible, Inviscid Flow Field
provided that the flow is barotropic and the body forces are conservative. Potential functions can be used to describe the velocity field for such flows. If we assume further that the flow is incompressible (i.e., low speed), we can linearly add potential functions to obtain the velocity field for complex configurations and use Bernouffi's equation to determine the corresponding pressure distribution. The inviscid flow field solutions, thus obtained, form the outer (edge) boundary conditions for the thin viscous (boundary) layer adjacent to the walLThe characteristics of the boundary layer and techniques for analyzing it are described in the next chapter.
PROBLEMS
A truck carries an open tank, that is 6 m long, 2 m wide, and 3 m deep. Assuming that the driver will not accelerate or decelerate the truck at a rate greater than 2 m/s2, what is the maximum depth to which the tank may be filled so that the water will not be spilled? 3.2. A truck carries an open tank that is 20 ft long, 6 ft wide, and 10 ft deep. Assuming that the driver will not accelerate or decelerate the truck at a rate greater than 6.3 ft/s2, what is the maximum depth to which the tank may be filled so that the water will not be spilled? 3.3. What conditions must be satisfied before we can use Bernoulli's equation to relate the flow characteristics between two points in the flow field? 3.4. Water fills the circular tank (which is 20.0 ft in diameter) shown in Fig. P3.4. Water flows out of a hole which is 1.0 in. in diameter and which is located in the side of the tank, 15.0
3.1.
ft from the top and 15.0 ft from the bottom. Consider the water to be inviscid.
PH20 = 1.940 slug/ft3.
(a) Calculate the static pressure and the velocity at points 1,2, and 3. For these calculations you can assume that the fluid velocities are negligible at points more than 10.0 ft from the opening. (b) Having calculated U3 in part (a), what is the velocity U1? Was the assumption of part (a) valid?
1
Surface
15.0
ft
—
2
—

——

flows out through a 1.0 in, diameter hole
110
20.0 ft
Diameter
Figure P3.4
Problems
3.5.
145
Consider alowspeed, steady flow around the thin airfoil shown in Fig. P35. We know the velocity and altitude at which the vehicle is flying. Thus, we know We (i.e., Pi) and have obtained experimental values of the local static pressure at points 2 through 6. At which of these points can we use Bernoulli's equation to determine the local velocity? If we cannot, why not? Point 2: at the stagnation point of airfoil Point 3: at a point in the inviscid region just outside the laminar boundary layer Point 4: at a point in the laminar boundary layer Point 5: at a point in the turbulent boundary layer Point 6: at a point in the inviscid region just outside the turbulent boundary layer.
of boundary
2'
Figure P3.5
3.6. Assume that the airfoil of problem 3.5 is moving at 300 km/h at an altitude of 3 km. The experimentally determined pressure coefficients are
Point
2
3
4
—3.00
5
6
1.00
—3.00
+0.16
+0.16
(a) What is the Mach number and the Reynolds number for this configuration? Assume that the characteristic dimension for the airfoil is 1.5 rn. (b) Calculate the local pressure in N/rn2 and in lbf/in.2 at all five points. What is the per
centage change in the pressure relative to the freestream value? That is, what is Was it reasonable to assume that the pressure changes are suffi(Piocai —
ciently small that the density is approximately constant? (c) Why are the pressures at points 3 and 4 equal and at points 5 and 6 equal? (ci) At those points where Bernoulli's equation can be used validly, calculate the local
velocity.
3.7. A Pitotstatic probe is used to determine the airspeed of an airplane that is flying at an
N/rn2, what is the airspeed? altitude of 6000 m. If the stagnation pressure is 4.8540 X What is the pressure recorded by a gage that measures the difference between the stagnation pressure and the static pressure (such as that shown in Fig. 3.2)? How fast would an airplane have to fly at sea level to produce the same reading on this gage? 3.8. A highrise office building located in Albuquerque, New Mexico, at an altitude of 1 mi is exposed to a wind of 40 mi/h. What is the static pressure in the airstream away from the influence of the building? What is the maximum pressure acting on the building? Pressure measurements indicate that a value of = —5 occurs near the corner of the wall parallel to the wind direction. If the internal pressure is equal to the freestream static pressure, what is the total force on a pane of glass 3 ft X 8 ft located in this region?
146
Chap. 3 / Dynamics of an Incompressible, Inviscid Flow Field
located in a city at sea level is exposed to a wind of 75 km/h. What is the static pressure of the airstream away from the influence of the building? What is the maximum pressure acting on the building? Pressure measurements indicate that a value of = —4 occurs near the corner of the wall parallel to the wind direction. If the internal pressure equals to the freestream static pressure, what is the total force on the pane of glass 1 m X 3 m located in this region? 3.10. You are working as a flighttest engineer at the Flight Research Center. During the lowspeed phase of the test program for the X37, you know that the plane is flying at an
3.9. A highrise office building
altitude of 8 km. The pressure at gage 1 is 1550 N/rn2, gage; the pressure at gage 2 is
—3875 N/rn2, gage.
(a) If gage 1 is known to be at the stagnation point, what is the velocity of the airplane? What is its Mach number? (b) What is the freestream dynamic pressure for this test condition?
(c) What is the velocity of the air at the edge of the boundary layer at the second point relative to the airplane? What is the velocity relative to the ground? What is for
this gage? 3.11. Air flows through a converging pipe section, as shown in Fig. P3.11. Since the centerline of the duct is horizontal, the change in potential energy is zero. The Pitot probe at the upstream station provides a measure of the total pressure (or stagnation pressure). The downstream end of the Utube provides a measure of the static pressure at the second station. Assuming the density of air to be 0.00238 slug/ft3 and neglecting the effects of viscosity, com
pute the volumetric flow rate in ft3/s. The fluid in the manometer is unity weight oil
= 1.9404 slug/ft3).
Flow
Figure P3.11
3.12. An indraft wind tunnel (Fig. P3J2) takes air from the quiescent atmosphere (outside the tunnel) and accelerates it in the converging section, so that the velocity of the air at a point in the test section but far from the model is 60 rn/s. What is the static pressure at this point? What is the pressure at the stagnation point on a model in the test section? Use Table 1.2 to obtain the properties of the ambient air, assuming that the conditions are those for the standard atmosphere at sea level.
Problems
147
Quiescent air
U
68 rn/s
Figure P3.12.
3.13. A venturi meter is a device that is inserted into a pipeline to measure incompressible flow rates. As shown in Fig. P.3.13, it consists of a convergent section that reduces the diameter to between onehalf to onefourth of the pipe diameter. This is followed by a divergent section through which the flow is returned to the original diameter. The pressure difference between a location just before the venturi and one at the throat of the venturi is used to determine the volumetric flow rate (Q). Show that
—
[
A2
—
/2g(p1
—
P2)
(AilA1)2 where Cd is the coefficient of discharge, which takes into account the frictional effects and is determined experimentally or from handbook tabulations.
[Vi
Figure P3.13
3.14. You are in charge of the pumping unit used to pressurize a large water tank on a fire truck. The fire that you are to extinguish is on the sixth floor of a building, 70 ft higher than the truck hose level, as shown in Fig. P3.14.
Fire
3
70 ft
Pressure
Figure P3.14.
148
Chap. 3 I Dynamics of an Incompressible, Inviscid Flow Field
Free
surface
I
I
F
Figure P3.15.
X____
(a) What is the minimum pressure in the large tank for the water to reach the fire? Neglect pressure losses in the hose. (b) What is the velocity of the water as it exits the hose? The diameter of the nozzle is 3.0
in. What is the flow rate in gallons per minute? Note that 1 gallmin equals 0.002228 ft3/s. 3.15. A free jet of water leaves the tank horizontally, as shown in Fig. P3.15. Assuming that the tank is large and the losses are negligible, derive an expression for the distance X (from the tank to the point where the jet strikes the floor) as a function of h and H? What is X, if the liquid involved was gasoline for which o = 0,70?
3.16. (a) What conditions are necessary before you can use a stream function to solve for the flow field?
(b) What conditions are necessary before you can use a potential function to solve for the flow field? (c) What conditions are necessary before you can apply Bernoulli's equation to relate two points in a flOw field? (il) Under what conditions does the circulation around a closed fluid line remain constant with respect to time? 3.17. What is the circulation around a circle of constant radius R1 for the velocity field given as
+
V =
—e0
2'irr
3.18. The velocity field for the fully developed viscous flow discussed in Example 2.2 is
U
=
idp(2
0
y — 4*
h2
v=0
w=
Is the flow rotational or irrotational? Why? 3.19. Find the integral along the path between the points (0,0) and (1,2) of the component of V in the direction of for the following three cases: (a) a straight line. (b) a parabola with vertex at the origin and opening to the right.
Problems
(c)
149
a portion of the x axis and a straight line perpendicular to it. The components of V are given by the expressions
U
= X2 + y2
2xy2
v=
3.20. Consider the velocity field given in Problem 3.12:
(x2 + y2)I + 2xy2]
Is the flow rotational or irrotational? Calculate the circulation around the right triangle
shown in Fig. P3.20.
y
(1,2)
(1,0)
Figure P3.20
What is the integral of the component of the curl
over the surface of the triangle? That is,
Are the results consistent with Stokes's theorem?
3.21. The absolute value of velocity and the equation of the potential function lines in a two
dimensional velocity field (Fig. P3.21) are given by the expressions
JVJ = \/4x2 + 4y2
= — y2 + c Evaluate both the lefthand side and the righthand side of equation (3.16) to demonstrate the validity of Stokes's theorem of this irrotational flow.
y
(2, 1)
Rectangular area
Figure P3.21
Chap. 3 / Dynamics of an Incompressible, Inviscid Flow Field
3.22. Consider the incompressible, irrotational twodimensional flow where the potential function is = K is an arbitrary constant. (a) What is the velocity field for this flow? Verify that the flow is irrotational. What is the magnitude and direction of the velocity at (2,0), at and at (0,2)? (b) What is the stream function for this flow? Sketch the streamline pattern. (c) Sketch the lines of constant potential. How do the lines of equipotential relate to the streamlines? 3.23. The stream function of a twodimensional, incompressible flow is given by
cli =
(a) Graph the streamlines. (b) What is the velocity field represented by this stream function? Does the resultant velocity field satisfy the continuity equation? (c) Find the circulation about a path enclosing the origin. For the path of integration, use a circle of radius 3 with a center at the origin. How does the circulation depend on the
radius? 3.24. The absolute value of the velocity and the equation of the streamlines in a velocity field are given by
IT/I = V4x2
—
4xy + 5y2
4xy
—
y2
=
y2
+ 2xy = constant
Find uandv.
3.25, The absolute value of the velocity and the equation of the streamlines in a twodimensional velocity field (Fig. P3.25) are given by the expressions
=
= xy +
+ x2 + 4xy
= C by two methods.
Find the integral over the surface shown of the normal component of curl
y
(2,1)
Area of interest
X
Figure P3.25
3.26. Given an incompressible, steady flow, where the velocity is
=

xy2)I
+
—
Problems
(a) Does the velocity field satisfy the continuity equation? Does a stream function exist? If a stream function exists, what is it? (b) Does a potential function exist? If a potential function exists, what is it? (c) For the region shown in Fig. P3.26, evaluate
and
to demonstrate that Stokes's theorem is valid.
y
(1,1)
dA is that of the planar triangle
x
Circulation around the triangle
Figure P3,26
3.27. Consider the superposition of a uniform flow and a source of strength K. If the distance from the source to the stagnation point is R, calculate the strength of the source in terms of
and R.
(a) Determine the equation of the streamline that passes through thestagnation point. Let this streamline represent the surface of the configuration of interest. (b) Noting that
1
V0 =
ar
complete the following table for the surface of the configuration.
0
r
—
U
R
C'°
30°
450
90°
1350
1500
1800
3.28. A twodimensional free vortex is located near an infinite plane at a distance h above the paralplane (Fig, P3.28). The pressure at infinity is and the velocity at infinity is lel to the plane. Find the total force (per unit depth normal to the paper) on the plane if
Iriviscid Flow Field the pressure on the underside of the plane is The strength of the vortex is r. A perfect. Find the pressure distribution along the wall as a function of x.15.e. This is the flow field of Section 3. 3 / Dynamics of an Incompressible. 3.152 Chap. To what expression does the force simplify if h becomes very large? + Vortex of strength I' h PC. incompressible irrotational fluid is flowing past a wall with a sink of strength K per unit length at the origin (Fig. 0 = 180°)? Assume that the freestream conditions are those of the standard atmosphere at sea level. Consider the flow formed by a uniform flow superimposed on a doublet whose axis is parallel to the direction of the uniform flow and is so oriented that the direction of the efflux opposes the uniform flow.32. 3. Sketch the resulting pressure distribution.31. Figure P3. Taking the freestream static pressure at infinity to be express the pressure coefficient as a function of x/xo. The fluid is incompressible and perfect. 6 = 90°)? What is the change in pressure from the freestream value to that at the stagnation point (i.1. U.e. show that a circle of radius R.28 3. .29). At infinity the flow is parallel and of uniform velocity Determine the location of the stagnation point x0 in terms of and K. Using the stream functions for these three elementary flows. Consider the flow field that results when a vortex with clockwise circulation is superimposed on the doublet/uniformflow combination discussed in Problem 3..0 Sink of strength K atx=O Figure P3.1. P3.13.31.29.This is the flow field of Section 3. where is a streamline in the flow field. What is the stream function that represents the potential flow about a cylinder whose radius is 1 m and which is located in an air stream where the freestream velocity is 50 mIs? What is the change in pressure from the freestream value to the value at the top of the cylinder (i. where is a streamline in the flow field.. Using the stream functions for these two elementary flows. show that a circle of radius R.29 330..
3.Problems 153 3. Neglecting the body forces in the air and the weight of the cylinder.. due to the airflow)? Air at 100 rn/s = 1.. Neglect variations in the velocity of the wind over the height of the smokestack. Using the data of Fig. Whenever the pressure on the two side holes is equal. P3.=20°C Figure P3...33.34.34. (a) If the orifices of a directionfinding Pitot tube were to be used to measure the freestream static pressure. in which direction does the plug more (i. its pressure is 99 kPa..33 3. The temperature of the air is 30°C. The instrument is called a directionfinding Pitot tube.30 calculate the force and the overturning moment exerted by a 4 rn/s wind on a cylindrical smokestack that has a diameter of 3 m and a height of 50 m.33. = 1. inviscid.0 X N/rn2 T0.5 X N/rn2 Air at 10 rn/s F0. Both air flows may be considered to be steady. as shown in Fig. 335.e. An infinitespan cylinder (twodimensional) serves as a plug between the two airstreams . A cylindrical tube with three radially drilled orifices. can be used as a flowdirection indicator. and incompressible. Figure P3. what is the sensitivity? Let the sensitivity be defined as the pressure change per unit angular change (i.34. as shown in Fig. or a cylindrical yaw probe.e. where would they have to be located if we use our solution for flow around a cylinder? (b) For a directionfinding Pitot tube with orifices located as calculated in part (a). P3. the pressure at the center hole is the stagnation pressure. What is the Reynolds number for this flow? .
what is the pressure at the stagnation points? What will be the velocity and the static pressure at 0 = 90°? at o = 270°? What will be the theoretical value of the lift per spanwise foot of the cylinder? 3. and the static pressure inside the hut is equal to that on the outer surface of the hut. as shown in Figure. and a height of 15 ft. what is the net lift acting on the hut? What is the lift coefficient? (b) Where should the door be located (i. What is the vortex strength required in order to place the stagnation points at O = 300 and 0 = 1500? If the freestream pressure is 2000 lbf/ft2.4 psi. Inviscid Flow Field 3. Assume that velocity distribution for the windward surface of the cylinder is given by the inviscid solution V= sin Calculate the lift and drag coefficients if the base pressure (i.. Calculate the force and the overturning moment exerted by a 45mph wind on a cylindrical flagpole that has a diameter of 6 in.38 3. Door = 175 Figure P3. incompressible. The temperature of the air is 85°F. P3. (a) If the door to the hut is located at ground level (i.. Consider an incompressible flow around a semicylinder. potential flow. Pcomer P base = Pcomer Figure P3. P3. Consider the flow around the quonset hut shown in Fig. The ground plane is represented by the plane of symmetry and the hut by the upper half of the cylinder. the pressure on the flat.39.38 to be represented by superimposing a unifonn flow and a doublet.38. What is the Reynolds number of this flow? 3. Assume steady. A cylinder 3 ft in diameter is placed in a stream of air at 68°F where the velocity is 120 ft/s. its pressure is 14. or leeward.39.36. where the door is located. 3 I Dynamics of an Incompressible.37. Assume that the wall is negligibly thin.e.e.154 Chap. at the stagnation point). the opening is very small compared to the radius Thus.e. the radius R0 of the hut is 6 m. The door is not well sealed. surface) is equal to the pressure at the separation point. Neglect variations in the velocity of the wind over the height of the flagpole.. the pressure on the door is essentially constant and equal to the value of the angle 0o at which the door is located. at what angle relative to the ground) so that the net force on the hut will vanish? For both parts of the problem.39 . The freestream velocity is 175 km/h.
what is the section lift coefficient? 3. as shown in Fig.39. The air under the hut is at rest and the pressure is equal to stagnation pressure. is submerged in a stream of air where = 1. calculate the section lift coefficient for this flow.58) gives the lift force per unit span. Combining equations (3.5 3. the pressure distribution over the top of the hut (the semicircle of the sketch) is represented by the potential function + When calculating the flow over the hut. P3. whose radius is 15 ft. A semicylindrical tube. .What are the lift and drag forces acting on the tube using the equations developed in Problem 3. 3.43. The radius is 0. Using equation (3. estimate the maximum lift that could be developed.3 m. The quonset hut may be considered to be a closed (no leaks) semicylinder.41 See Example 3.45) and (3. thus.56) to define the surface velocity distribution for inviscid flow around a cylinder with circulation. it has been shown that the section lift coefficient for inviscid flow around a cylinder is 2ir ci = (3. Pt(= (a) What is the value of B for the 15ftradius (R) quonset hut? (b) What is the net lift force acting on the quonset hut? (c) What is the net drag force acting on the quonset hut? U.. derive the expression for the local static pressure as a function of 0.Problems 3.42. The design wind speed is 100 ft/s. Tie down p =p1 Tie down Figure P3. Using the definition that C1 = (where 1 is the lift per unit span).49). mounted on tiedown blocks.57) to define the pressure coefficient distribution for inviscid flow with circulation. Substitute the pressure distribution into the expression for the lift to verify that equation (3. The static freestream properties are those for standard sealevel conditions. You are to design quonset huts for a military base in the mideast.39. 3. Consider such a cylinder having a diameter of 1 m and a length of 10 m.44.41. neglect the presence of the air space under the hut.40.5 The flow is such that the velocity distribution and. disregarding end effects. There were early attempts in the development of the airplane to use rotating cylinders as airfoils. as shown in Example 3.48) Using equation (3. If this cylinder is rotated at 100 rpm while the plane moves at a speed of 100 kmlh through the air at 2 km standard atmosphere.22 kg/rn3 and = 75 rn/s.
45. 3 / Dynamics of an Incompressible.6. as shown in Fig. source dis tribution on panel 3 to the normal velocity at the control point of panel 4. Consider air flowing past a hemisphere resting on a flat surface. A hemisphere.e. calculate the contribution of the. P3.47. Inviscid Flow Field 3. or leeward surface) is equal to the pressure at the separation point. What are the lift and drag forces on the hemisphere using the equations developed in Problem 3. Assume that the velocity distribution for the windward surface of the cylinder is given by the inviscid solution V= (3. P3.47. if the internal pressure is find an expression for the pressure force on the hemisphere.e. The radius is 1. Problem 3.46. Using the procedures illustrated in Example 3. P3.) 3. Consider the pressure distribution shown in Fig.47.49 . Neglecting the effects of viscosity.002376 slug/ft3 and = 200 ft/s. At what angular location should a hole be cut in the surface of the hemisphere so that the net pressure force will be zero? Standard atmospheric D Figure P3.27.. P3. Consider an incompressible flow around a hemisphere. How does the drag coefficient for a hemisphere compare with that for a hemicylinder (i.48.47.49.77) Calculate the lift and drag coefficients if the base pressure (i.0 ft. 3. as shown in Fig. 3.156 Chap. as shown in Fig.75 : I I Figure P3.) 3.46 for the windward and leeward surfaces of a thick disk whose axis is parallel to the freestream flow.. What is the corresponding drag coefficient? = 0.46 3.48. The configuration geometry is illustrated in Fig. is submerged in an airstream where = 0. the pressure on the flat.49.
Borst liv. Chambers JR. NASA SP514 Hess JL. Smith AMO. DC: U. 3. NJ: published by the author Hoerner SF. Aeronaut. NJ: published by the authors Kellogg OD. Boundary Layer Theory. 8:1138 Hoerner SF. Distributions of local pressure and skin friction around a circular cylinder in cross flow up to Re = 5 X 106.S. Fundamentals of Potential Theory. Midland Park. Standard Atm osphere.75 in. 157 A major league pitcher is accused of hiding sandpaper in his back pocket in order to scuff up an otherwise "smooth" basebalL Why would he do this? To estimate the Reynolds number of the baseball. Derive the stream functions for the elementary flows of Table 3. NASA SP367 1976.2. U.S. 1968. 6th Ed.References 3.52. Midland Park. Progr. Fluid Dynamic Lift. 3. Government Printing Office . New York: McGrawHill Talay TA. What condition(s) must prevail in order for a velocity potential to exist? For a stream function to exist? REFERENCES Achenbach E. 1975. 1958. 1975. Fluid Mechanics 34:625—639 Campbell J. 1968. assume its speed to be 90 mi/h and its diameter 2. 1967. 1953. New York: Dover Schlichting H. Patterns in the sky: natural visualization of aircraft flow fields.51. Introduction to the aerodynamics of flight. Calculations of potential flow about arbitrary bodies. Washington. Fluid Dynamic Drag. J.50. Sci. 1994.
The principal forces considered were those that act directly on the mass of the fluid element (i. but the velocity for many of the practical applications relevant to this text is such that the Reynolds number is very large. which states that the net force acting on a fluid particle is equal to the time rate of change of the linear momentum of the fluid particle. .e. the flow field may be divided into two regions: (1) a viscous boundary layer adjacent to the surface of the vehicle and (2) the essentially inviscid flow Outside the boundary layer. The resultant equations are known as the NavierStokes equations. reasonable approximations can be introduced to describe the motion of a viscous fluid if the viscosity is either very large or very small.4 VISCOUS BOUNDARY LAYERS The equation for the conservation of linear momentum was developed in Chapter 2 by applying Newton's law. the pressure forces and shear forces). Not only is the viscosity of these fluids very small.. have very small viscosities.. for many high Reynolds number flows. Outside the boundary layer. because the solution of the simplified equation could not be made to satisfy the complete boundary conditions. Nevertheless. it is not permissible simply to omit the viscous terms completely. However. Even in the limiting case where the Reynolds number is large. Even today. the body forces) and those that act on its surface (i. since two important fluids. there are no general solutions for the complete NavierStokes equations.e. The velocity of the fluid particles increases from a value of zero (in a vehiclefixed coordinate system) at the wall to the value that corresponds to the external "frictionless" flow outside the boundary layer. The latter case is of special interest to us. water and air.
By restricting ourselves to such flows. the reader is referred to Schlichting (1979) and White (2005). the coordinate system is fixed to the surface of the. the first step is to solve for the iñviscid portion of the flow field.1. we will consider the flow to be steady. we can concentrate on the development of the solution techniques themselves. INCOMPRESSIBLE FLOW In this chapter. See Fig. and constant property (or. we generated solutions for the inviscid flow field for a variety of configurations. the iterative procedure required to converge to a solution requires an understanding of each region of the flow field and their interactions. we shall assume that we know whether the boundary layer is laminar or turbulent. 4. The stagnation point (or the leading edge if the configuration is a sharp object) is at x = 0. To start the second iteration. and Mach number). the second step is to calculate the boundary layer using the inviscid flow as the outer boundary condition. This coordinate system is used throughout this chapter. However. the inviscid flow field is recalculated. For a more detailed discussion of the parameters that affect transition. In Chapter 3. which is determined by adding the displacerne it thickness of the boundary layer from the first iteration to the surface coordinate of the actual configuration.Sec. The transition process through which the boundary layer "transitions" from a laminar state to a turbulent state is quite complex and depends on many parameters (e. assuming that the inviscid flow field is known. twodimensional. Thus.g.1 I Equations Governing the Boundary Layer for a Incompressible How 159 the transverse velocity gradients become so small that the shear stresses acting on a fluid element are negligibly small. body. 4. The solution for the inviscid portion of the flow field must satisfy the boundary conditions: (1) that the velocity of the fluid particles far from the body be equal to the freestream value and (2) that the velocity of the fluid particles adjacent to the body parallel to the "surface.1 EQUATIONS GOVERNING THE BOUNDARY LAYER FOR A STEADY. for the purpose of t text. A brief summary of the factors affecting transition is presented later in this chapter. replacing the actual configuration by the "effective" configuratiOn. the effect of the viscous terms may be ignored when solving for the flow field external to the boundary layer. The y coordinate is perpendicular to the surface. since the flow is iñviscid. The x coordinate is measured in the streamwise direction along the surface of the configuration. Having solved for the inviscid flow field. 2. it may be necessary to use an iterative process for calculating the flow field. incompressible for a gas flow). equivalently. we discuss techniques by which we can obtain engineering solut when the boundary layer is either laminar or tUrbulent. When using the tworegion flow model to solve for the flow field.As discussed in DeJarnette and Ratcliffe (1996).13. As shown in Fig. pressure gradient. The boundary layer is recalculated using the seconditerate inviscid flow as the boundary condition. 4. To simplify the development of the solution techniques. the velocity component parallel to the surface does not have to be zero. surface temperature. In this chapter we examine the viscous region in detail. TWODIMENSIONAL. Thus. surface roughness. ." The second boundary condition represents the physical requirement that there is no flow through a solid surface.. If the boundary layer is relatively thick.
then v.3) Note that. twodimensional flow is au äx a2u ay (4.The assumption that the static pressure variation across a thin boundary layer is negligible only breaks down for turbulent boundary layers at very high Mach numbers. 4 I Viscous Boundary Layers x is measured along the surface) Figure 4. which were discussed in Chapter 3. This is true whether the boundary layer is laminar. or turbulent. Thus.160 Chap. Euler's equation for a steady . separated regions in the lee side of blunt bodies such as those behind cylinders. the y component of momentum is t3V 3p ax ay (4.4) Thus.1) Referring to equation (2. the local static pressure is a function of x only and is determined from the solution of the inviscid portion of the flow field.2) Similarly. if we compare each term in equation (4. Referring to equation (2. the x component of momentum for this steady. as discussed in Chapter 2. As a result. It is not true in wake flows. transitional.1 Coordinate system for the boundarylayer equations.3) with the corresponding term in equation (4. that is. the differential form of the continuity equation for this flow is ax 8y (4.16) and neglecting the body forces. if the boundary layer is thin and the streamlines are not highly curved.3). The common assumption for thin boundary layers also may be written as (4. we conclude that u Pu— > pu— ax au pV au > pV— äy 3y a2v 8x 82v c3y Thus.2). we conclude that ax ay The essential information supplied by the y component of the momentum equation is that the static pressure variation in the y direction may be neglected for most boundary layer flows.
in addition to the laminar shear stress described in the preceding paragraph. the turbulent boundary layer is relatively thick.. the momentum transfer in a direction perpendicular to the principal flow direction) takes place on a molecular (or microscopic) scale. they tend to accelerate the fluid particles in that layer. such as density p and viscosity are constants. can be used to evaluate the pressure gradient in the viscous region. there is an effective turbulent shear stress that is due to the transverse transport of momentum and that is very large. the transverse exchange of momentum (i.Sec.e. Because slowermoving fluid particles near the wall are transported well upward. the analysis of a compressible boundary layer involves the simultaneous solution of the continuity equation. Thus. also. the shear stress at the wall for a turbulent boundary layer is larger than that for a laminar boundary layer. We could introduce the fluctuating characteristics of turbulent flow at this point and treat both laminar and turbulent boundary layers in a unified fashion.5) into equation (4.6). the shear stress at a point in a Newtonian fluid is that given by constitutive relations preceding equations (2.The assumption that the flow is constant property (or incompressible) implies that fluid properties. 4. the temperature changes in the flow field are sufficiently large that the temperature dependence of the viscosity and of the density must be included. As a result. By limiting ourselves to incompressible flows. and noting that we obtain < (4. As a result of the molecular movement. there is a macroscopic transport of fluid particles. Conversely.6) Pu— + t3u pV äu = dUe dx + 32u Let us examine equations (4. when the fastermoving fluid particles from the upper layer migrate downward.2b. and the energy equation.1) and (4.Thus. it is not necessary to include the energy equation in the formulation of our solution. For a detailed treatment of compressible boundary layers. For a laminar .g. Because the macroscopic transport of fluid introduces large localized variations in the flow at any instant. For a turbulent boundary layer. slowing the particles in the upper layer. slower moving fluid particles from the lower layer (or lamina) of fluid move upward. the changes in pressure and temperature through the flow field are sufficiently small that the corresponding changes in p and 1a have a negligible effect on the flow field. —— = ax = PeUeT dUe ax (45) Substituting equation (4. the x momentum equation. Thus. For compressible (or highspeed) flows.. For lowspeed flows of gases.1 / Equations Governing the Boundary Layer for a Incompressible Flow 161 flow with negligible body forces.2a.2). 4.12). which relates the streamwise pressure gradient to the velocity gradient for the inviscid flow. the reader is referred to other sources [e. This molecular interchange of momentum for a laminar flow is depicted in Fig. they produce relatively high yelocities for the fluid particles near the surface. 4. the values of the fluid properties and the velocity components are (in general) the sum of the "average" value and a fluctuating component. When the boundary layer is laminar. Because fastermoving fluid particles (which are normally located near the edge of the boundary layer) are transported toward the wall. as shown in Fig. Schlichting (1979) and Dorrance (1962)3 and to Chapter 8.
ylarge) Ue(X) (4. we reach the edge of the boundary layer where the streamwise component of the velocity equals that given by the inviscid solution.. Since we are considering that portion of the flow field where the viscous forces are important.O) 0 (4.aty = 0. u(x. we will first discuss laminar flOws and their analysis and then turbulent boundary layers and their analysis. the normal component of velocity must be zero.2 BOUNDARY CONDITIONS Let us now consider the boundary conditions that we must apply in order to obtain the desired solutions.2 Momentumtransport models: (a) laminar boundary layer. those for the inviscid solution). However. Thus.7b) The velocity boundary conditions for porous walls through which fluid can flow are treated in the problems at the end of the chapter. boundary layer the fluctuating components of the flow would be zero.e. In equation form. the subscript e will be used to denote parameters evaluated at the edge of the boundary layer (i. to simplify the discussion.0) = v(x. Furthermore. at y large). .8) Note that throughout this chapter. 0 (4.. at points far from the wall (i.7a) At a solid wall. u(x. the condition of no slip on the solid boundaries must be sat isfied. 4.Thatis. (b) turbulent boUndary layer.e. 4 I Viscous Boundary Layers Inviscid flow external to the boundary layer Ue Motion on a macroscopic scale Inviscid flow external to the boundary layer F Motiononal micrOscopic scale u iLe u(y) (a) (b) Figure 4.__________ 162 Chap.
The streamwise (or x) component of velocity is presented as a function of distance from the wall (the y coordinate). in reality. at each station. The local velocity at the edge of the boundary layer Ue is a function of x only.. 4. Instead of presenting the dimensional parameter u. laminar boundary layers.3 Solutions for the dimensionless streamwise velocity for the FalknerSkan.4 0. similarity flows.e. Instead of using the dimensional y coordinate.Sec. we will use a dimensionless coordinate which is proportional to y/6 for these incompressible.. 4.3.3 I Incompressible.The reader should note that. the velocity varies from zero at y = 0 (i. a logical dimensionless velocity parameter is U/Ue. that "point" at which the boundary layer becomes turbulent). Laminar Boundary Layer 163 43 INCOMPRESSIBLE. let us seek a dimensionless velocity parameter that perhaps can be written as a function of a single variable.2 0. at the wall) to Ue for the inviscid flow outside of the boundary layer. which is a function both of x and y.0 Figure 4. . the boundary layer does not go from a laminar state to a turbulent state at a point but that the transition process takes place over a distance.8 1. The length of the transition zone may be as long as the laminar region. LAMINAR BOUNDARY LAYER In this section we analyze the boundary layer in the region from the stagnation point (or from the leading edge of a sharp object) to the onset of transition (i. Note that. Typical velocity profiles for the laminar boundary layer are presented in Fig.e.The boundarylayer thickness 6 at any x station depends not only on the magnitude — 11 0 0. Thus.6 0. laminar.
The flatplate solution is the classical Blasius solution.ue is a constant (independent of x) and is equal to the freestream velocity upstream of the plate = (4.10) Those readers who are familiar with the transformations used in more complete treat ments of boundary layers will recognize that this definition for is consistent with that commonly used to transform the incompressible laminar boundary layer on a flat plate [White (2005)]. the boundarylayer flow cannot be considered as irrotationaL Therefore. and where s = f dx (4. To do this.164 Chap. thex component of the momentum equation. and the velocity variation from the origin to the point of interest. the continuity equation (4. note that = e \'ayJx \\aYJ'X'\aTlJS (4. potential functions cannot be used to describe the flow in the boundary layer. in terms of only one unknown.) We shall define the stream function such that u=t—I and v=—I '\ax By introducing the stream function.llb) . 4 / Viscous Boundary Layers of x but on the kinematic viscosity.6)] with two unknowns. (Note: Because of viscosity. the velocity components: u and v. the stream function. By using this definition of s as the transformed x coordinate.Thus. we will introduce the coordinate transformation for UeY 77 = (49a) where i' is the kinematic viscosity.9b) Note that for flow past a flat plate. the local velocity at the edge of the boundary layer. Note that we have two equations [equations (4.1) and (4. This tion is also consistent with more general forms used in the analysis of a compressible laminar flow {Dorrance (1962)]. coordinate system.lla) + 1 (4.1) is automatically satisfied.we need to solve only one equation. as defined in Chapter 1.. Let us now transform our equations from the x. we can account for the effect of the variation in Ue on the streamwise growth of the boundary layer. y coordinate system to the s. the necessary and sufficient conditions for the existence of a stream function are satisfied. Since the flow is two dimensional and the density is constant.Thus.
82u äy2 = = a [aiia(uef')l J 2 =( I 2vs where the prime (') denotes differentiation with respect to vj. and equation (4. Using these substitutions. We can rewrite equation (4. 1) coordinate system. Laminar Boundary Layer 165 the streamwise component of velocity may be written in terms of the stream function as = (4.3 1 Incompressible.12b) Comparing equations (4. we have reduced the formulation to one equation with one unknown. the transformed stream function and its derivatives are functions of only. we can develop an expression for the transverse component of velocity: V= —t — + + (4. Thus.As a result. in terms of the transformed stream function.14).(f 1)21  (4. the momentum equation becomes ff" + f" + [1 .6) using the differentials of the variables in the s.Sec.14) In equations (4. is assumed to be constant.12b) and (4. which is represented by the symbol /3. Let us introduce a transformed stream function f. we have written the two velocity components.12a) = Thus.15) As discussed earlier. For example.12a) and (4J2b).16) . which we define so that u= &y1 S (4. the parameter (2S/Ue) ( due/ds).13) Similarly. we see that 1 (4. which were the unknowns in the original formulation of the problem. by using a stream function we automatically satisfy the continuity equation. 4. The assumption that /3 is a constant implies that the s derivatives of f and f' are zero.15) becomes the ordinary differential equation: + fill + [1 — (f')2]/3 = 0 (4. For many problems.
f'(s. . boundarylayer separation may occur. As noted in the discussion of flow around cylinder in Chapter 3. y coordinate system.17) have been generated for —0.17b) Using equations (4. Note that the Reynolds number does not appear as a parameter when the momentum equation is written in the transformed coordinates.166 Chap. when the air particles in the boundary layer encounter a relatively large adverse pressure gradient. 4. the wall.0.17c) is the transformed stream function.This is consistent with the requirement that v(x. there are a variety of welldocumented numerical techniques available to solve it. which results from a favorable pressure gradient [i.e. the surface is streamline). and the solution is that for flow past a flat plate (known as the Blasius solution). Separation results because the fluid particles in the viscous layer have been slowed to the point that they cannot overcome the adverse pressure gradient. this third boundary condition states that the stream function is constant along the wall (i. 4 / Viscous Boundary Layers Because the dimensionless velocity function f' is a function of ij only. 0) = 0. Negative values of /3 correspond to cases where the inviscid flow is decelerating. the boundary condition that the transverse velocity be zero at the wall becomes f(s. The positive values of /3 correspond to an accelerating inviscid flow. the solutions are called similar so lutions. represent the flow around different configurations.1988 tant velocity profiles are presented in Fig..1 Numerical Solutions for the FalknerSkan Problem Numerical solutions of equation (4.3. Substituting the definition that Ue into the boundary conditions given by equations (4.0) = Since f 0 (4. (dp/dx) <0].O) = and far from the wall. which corresponds to an adverse pressure gradient [i. The effect of an adverse pressure gradient is evident in the velocity profiles presented in Fig.1.7). The resulsented by equation (4. 4.. Since = 2s dUe ds these solutions represent a variety of inviscid flow fields and..3. There are no analytical solutions to this thirdorder equation.0 (4. Let us examine the three boundary conditions necessary to solve the equation.7) and (4. Ug = constant. the velocity profiles at one s station are the same as those at another.16) that satisfy the boundary conditions repre +2.3 and Table 4.17a) = 1. which is known as the FalknerSkan equation.8).e. 0 (4. It will appear when our solutions are transformed back into the x. Nevertheless.14) and (4.e. therefore. which results because the component of velocity normal to a streamline is zero. 4. Note that when f3 = 0. (dp/dx) > 0]. Thus.
If the adverse pressure gradient were any larger. 4.9959 4.000 0.4169 0.To calculate the shear force at the wall.6834.9880 0.7959 0.0000 0.3265 0.9214 0.6189 0.8467 0.6244 0.1183 0.6 0..1988.2857 0.6327 0.9955 0.9131 0.0010 0.9529 0.8 3.9568 0.2188 0.9906 0.8 0.9306 0.7 0.180 0.9946 1.6994 0.1876 0.4 4.2790 0.5211 0.Thus.1489 0.1408 0.2 4.6967 0.6635 0.9514 0.0159 0.8985 0.0358 0.4510 0.9947 0.0 2.9727 0.8788 0.2266 0.9531 0.0138 0.0293 0.9927 0.11).4 0.1 Velocity f'(n) for the FalknerSkan.1927 0. the laminar boundary layer would separate from the surface.4 1.3 0.4185 0.0000 0. au/ay at the wall is relatively large.7605 0.2806 0. T = dy (4.2 2.2 3.8717 0.6299 0.3127 0.9 1.3 I incompressible.2 0.1 0.8 4. Laminar.9839 0.5946 0.1094 0.7351 0.5 0.4946 0.969 1 0.9285 0.9617 0.0040 0.0 When /3 = —0.0000 0.e.9929 0.5663 0.7611 0.0487 0.1337 0.9421 0.8633 0.5452 0.2 1.9903 0.0658 0.4145 0.9398 0.9838 0.9959 0.9845 0.1874 0.000 0.7742 0.0 0.0939 0.0 0.0467 0.4895 0.300 1.9915 2. and flow reversal would occur. the shear is T = (4.8325 0.5717 0.Sec.0000 0.2498 0.4 3. Referring to equation (1.0636 0.8 5. the velocity increases rapidly with dis tance from the wall.3802 0.1597 0.2165 0. Similarity Flows /3 —0.9876 0.9848 0.9732 0.6096 0.4671 0.9944 0.3252 0.3494 0.1423 0.9597 0.2342 0.9323 0.9681 0.5219 0.9740 0.9798 0. but the velocity gra dient au/ay is also zero at the wall.lla).6859 0.0000 0.19) .7779 0.7277 0.9011 0.0 3. one would expect that the shear force at the wall would be relatively large. not only is the streamwise velocity zero at the wall.0470 0.9909 0.7032 0.0248 0.8145 0.8 0. For the accelerating flows (i.2979 0.0867 0.4606 0.8377 0.1588 0. 0.0760 0.6 2.6 0. positive /3).8363 0.4167 0.3720 0.000 2.5620 0.0089 0.9151 0.6 3.8791 0.7450 0.0000 0.8167 0.0804 0.9755 0.1988 —0.8606 0.6 1.9804 0.8968 0.18) let us introduce the transformation presented in equation (4. Laminar Boundary Layer Nurnerica I Values of the Dimensionless Streamwise 167 TABLE 4.3462 0.7858 0.5231 0. Thus.0991 0.4099 0.0 4.4 2.
4. f" (0) = 0.4.000 0.4 Transformed shear function at the wall for laminar boundary layers as a function of f3.2. Thus. for the laminar boundary layer on a flat plate.4 and in Table 4.21) where = (4.0 13 Figure 4. Because of its use in equation (4.664 = (4. The value does not depend on the stream conditions. such as velocity or Reynolds number. Conversely.000 1.22) Mentally substituting the values off" (0) presented in Fig.19).180 0. TABLE 4.1988 0.Theoretical values of f" (0) are presented in Fig. T= (4.300 0. When /3 = 0. 4.20) As noted earlier.000 1.1988.4696. we see that the shear is zero when /3 = —0. velocity at the edge of the boundary layer (tie) is equal to the freestream value mensionless skinfriction coefficient We can express the shear in terms of the diT 0.1286 0.168 Chap. we will call f" the transformed shear function. for flow past a flat plate.7748 1.000 —0. this value of /3 corresponds to the onset of separation. when the inviscid flow is accelerating. the shear is greater than that for a zero pressure gradient flow.6872 .2 Theoretical Values of the Transformed Shear Function at the Walt for Laminar Boundary Layers as a Function of f"(O) $ —0. 4 I Viscous Boundary Layers f "(0) 2.2326 2. laminar boundary layers. Note that f" (0) is a unique function of /3 for these incompressible.4696 0. Thus.
1497 0.1874 2.0023 0.9804 0.3.4617 0. Let us define the boundarylayer thickness 6 as that distance from the wall for which u = O.9994 f" 0.5 4.0939 0.3336 0.1 0.0844 0. the dimensionless streamwise velocity (f').9306 0. the thickness of a laminar boundary layer is proportional to and is inversely proportional to the square root of the Reynolds number.7610 0.7955 1.0 2.23) Thus.Sec.8633 0.3 lj 0.8 0.4572 0.0000 0.4696 0.3797 0.4 1.e.1676 0.4344 0.1891 0. 4.2342 0.6 2.9 1.4106 0.2863 2.0069 0.3425 0.2557 0.0 4.0586 0.9905 2.7838 3.6244 0.4686 0. as y increases) the shear goes to zero and the function f' tends asymptotically to 1.6 1. We see that the value of q corresponding to the boundarylayer thickness independent of the specific flow properties of the free stream.2 3.0000 0.2106 0. Converting this to a physical distance.1408 0.4696 0.6967 0.8868 1.2 0.4167 0.1286 0.0464 0.1876 0.3 / Incompressible.4693 0.9010 0.4650 0.9978 0.5829 0. the corresponding boundarylayer thickness (6) is 6 = = or — 6 — 5.4436 0.0 (4. .0951 0.0015 0.0 1.4 2.8 2.2 1.0470 0.9691 0..5 0.4512 0.6 0.3 0.3005 0.0.0244 0.9529 0. and the shear function (f") are presented as a function of for a laminar boundary layer on a flat plate in Table 4.2832 The transformed stream function (f).4 0.0305 0.2 2.5 0.8 3.2330 0.4606 0.0 • f 0.99Ue.0 3.0094 0.8167 0.4673 0.2315 1.3265 0.2806 0. Laminar Boundary Layer Solution for the Lammar Boundary Layer on a Flat Plate 169 TAB LE 4.5452 0.0677 0.9907 0.7288 0.7 0.3720 0. Note that as increases (i.4507 0.6032 1.4 3.0549 1.9880 0.4148 1.1 147 f' 0.0375 0.0211 0.
170 Chap.A more significant measure of the boundary layer is the displacement thickness 8*..5 Velocity profile for a laminar boundary layer on a flat plate illustrating the boundarylayer thickness 6 and the displacement thickness . which is due to the fact that the increasing boundarylayer thickness causes the fluid to be displaced from the wall as it flows along it. Referring to Fig.84 425 This means that at the outer edge there is an outward flow. Referring to equation (4.0 U 1te Figure 4.23)]. we can see that Ve 1 . it is not zero at the edge of the boundary layer. 0.99 was the value used to develop equation (4. which is the distance by which the external streamlines are shifted due to the presence of the boundary layer.fe] Ve (4. 4 / Viscous Boundary Layers Although the transverse component of velocity at the wall is zero. There is no boundarylayer separation for flow past a flat plate. which is chosen as the criterion for the edge of the boundary layer [e.14). Since the streamwise component of the velocity in the boundary layer asymptotically approaches the local freestream value. 4. p8 Petie6* = J0 p(Ue — u) dy y 0. the magnitude of 8 is very sensitive to the ratio of U/tie.5. 0.g.3. since the streamwise pressure gradient is zero.24) Using the values given in Table 4.
The use of the total skinfriction drag coefficient avoids performing the same integration numerous times with different flatplate lengths accounting for different results. we obtain 6* = — fe) 1. 6* = — Using the values presented in Table 4.L o 0. For an incompressible.28) The momentum thickness represents the height of the freestream flow which would be needed to make up the deficiency in momentum flux within the boundary layer due to the shear force at the surface. o — 0.Sec.10) for the laminar boundary layer on a flat plate.The total skinfriction coefficient for laminar boundary layers becomes Cf= 1 1 bJ rdx= o b 1.3 / Incompressible. the displacement thickness 6* is on the order of onethird the boundarylayer thickness 6. The total skinfriction coefficient is defined as Cf (4. for an incompressible boundary layer is given by o = O Ue\ dy (4.664 = 2Cf(L) = 7 28 (4.664 429 Another convenient formulation for skinfriction on a flat plate is found by integrating the "local" skinfriction coefficient.27) Thus.72 (4. The momentum thickness. laminar boundary layer. Cf. found in equation (4.3. for any incompressible boundary layer.26) Note that. for a flat plate at zero incidence in a uniform stream. the upper limit for the integration does not matter providing it is equal to (or greater than) 6. Swet = Lb). 4.30) where D1 is the friction drag on the plate and 5wet is the wetted area of the plate (the wetted area is the area of the plate in contact with the fluid—for one side of the plate. Laminar Boundary Layer Thus.31) . 171 = 6* — dy (4. since the integrand is zero for any point beyond 8. 0.21) to obtain a "total" or "average" skinfriction drag coefficient on the flat plate [White (2005)]. Substituting the transformation of equation (4.
7894 X105kg/m.172 Chap.The freestream velocity is 40 rn/s.328 = (4. s. The density of the air is 1. whose streamwise dimension (or chord c) is 0. The total skinfriction coefficient for laminar flow simply becomes — C1 1.1: A rectangular plate. and x = 0.10 m. N Cf i= 1 Always convert total skinfriction coefficients into drag coefficients (based on a single reference area) and then add the drag coefficients to obtain a total skinfriction drag coefficient: N CD = i=1 CD.. The maximum value of the local Reynolds number.2 m) = 5.33) 'ref It can be tempting to add together the total skinfriction coefficients for various flat plates in order to obtain a total skinfriction drag—this must never be done! Since each total skinfriction coefficient is defined with a different wetted area.225 kg/m3)(40m/s)(0. which occurs when x = c.05 m. it is independent of the spanwise coordinate). Graph the velocity profiles at x = 0. the drag coefficient due to skinfriction is obtained from equation (4. Since drag coefficients are normally nondimensionalized by a reference area rather than a wetted area [see equation (3.53)]. In other words.7894 >< kg/rn. x = 0.s) . x= m. is (1.32) where ReL is the Reynolds number evaluated at x = L. and the absolute viscosity is 1. doing this would result in an incorrect result.2 rn and whose width (or span b) is 1. 4 / Viscous Boundary Layers which is just twice the value of the local skinfriction coefficient evaluated at x = L.34) EXAMPLE 4. let us assume that the flow is two dimensional (i. What is the drag coefficient for the plate? 0.20 Solution: Since the span (or width) of the plate is 9.0 times the chord (or streamwise dimension).is mounted in a wind tunnel.477 X (1.0 m.8 rn. (4.30) as Df S CD = = (4.2250 kg/rn3. Calculate the chordwise distribution of the skinfriction coefficient and the displacement thickness. which is the end of the flat plate.e.
Laminar Boundary Layer y(X1O) 173 Iatx=O.0 + + 2.013 x .3 to calculate the velocity profiles. 4. 4. if the dimensionless velocity (U/tie) is presented as a function of the profile is the same at all stations.6. The resultant profiles are presented in Fig.546 x tIe we can use the results presented in Table 4. However..7X1O4m ue=40m/s + + + + 1.5 X 104m u =40m/s 8=6. These calculations verify the validity of the common assumption that the boundary layer is thin. the velocity is constant (independent of y).The profiles at the other stations illustrate the growth of the boundary layer with distance from the leading edge.e.477 X i05.0 in Fig. the velocity within the boundary layer is a function of x and y. 4.3 x Figure 4. Since the dimensionless profiles are similar at all x stations.Sec. Thus.664 4. Even though the streamwise velocity at the edge of the boundary layer (ue) is the same at all stations. The skinfriction coefficient is 0. the profile is that for /3 = 0.6 Velocity profile for the layer = 5. Note that the scale of the y coordinate is greatly expanded relative to that for the x coordinate.20m 13. The displacement thickness in meters is = 1.3 / Incompressible.3. laminar boundary This Reynolds number is close enough to the transition criteria for a flat plate that we will assume that the boundary layer is laminar for its entire length. Therefore.05m U00=40fl1/S Iatx=O. Specifically. the solutions are termed similarity solutions. at x = 0).6. we will use the relations developed in this section to calculate the required parameters. At the leading edge of the flat plate (i. the inviscid solution obtained neglecting the boundary layer altogether and that obtained for the effective geometry (the actual surface plus the displacement thickness) are essentially the same. Noting that y= = 8.72x = L0394 x The chordwise (or streamwise) distribution of the displacement thickness is presented in Fig. 4.lOm Iatx=O.
and the skinfriction coefficient (Cf). Substituting the expression for the laminar shear forces. Compare the values obtained assuming a linear velocity profile with the more exact solutions presented in this chapter.35). is multiplied by b (the span) and by 2 (since friction acts on both the top and bottom surfaces of the plate).20). Therefore. the integral.2250 kg/rn3 and = 1.36) the edge velocity (ue) is equal to the freestream velocity drag coefficient for the plate is. Calculate the streamwise distribution of the displacement thickness the velocity at the edge of the boundary layer (ye). the standard atmospheric conditions at sea level include = 1. Solution: As given in Table 1. D = (4.174 Chap.32) and computing drag on the top and bottom of the plate.25 x 10_2 Assume that we are trying to approximate the flow of air at standard sealevel conditions past a flat plate where Ue = 2. we have assumed that there is no spanwise variation in the flow. which represents the drag per unit width (or span) of the plate.7894 X kg/s . Using general notation. the drag force acting on the flat plate is due only to skin friction. we see that PC D = 2b / TdX (4. CD = 3. Since the CD D 2. given in equation (4.589 X Alternatively. the pressure contributes nothing to the drag.35) We need integrate only in the x direction.2: The streamwise velocity component for a laminar boundary layer is sometimes assumed to be roughly approximated by the linear relation U = y where = 1.337 rn/s.2. .589x10 Lb EXAMPLE 4.656 = (437) For the present problem. Obviously. since by assuming the flow to be two dimensional. In equation (4. 4 / Viscous Boundary Layers Let us now calculate the drag coefficient for the plate.328 2Lb — —3. using the total skinfriction coefficient equation (4. CD—Cf Sref 1. therefore.
84 Ve — — — x Ue Finally. = 1.00 x .25 x 102 then 3* = 1. Laminar Boundary Layer Thus.60 x 105x Using the definition for the displacement thickness of an incompressible boundary layer. we find that the skin friction for the linear velocity approximation is given by (9u'\ T/Lt1 \&y Thus.27)].Sec.60 X 105x.26).25 x 102 f1 — = 0.430 X 10—2 Using the continuity equation.625 x 102 for the linear profile. since = Ue 8 and S= 1. since we have U/lie jil terms of y/8.10 lie Using the more exact formulation of equation (4. equation (4. we find that and noting that 3* = 0.3 I Incompressible. Using the equation for the more exact formulation [equation (4. we have changed our inde pendent variable from yto y/8.25) yields 0.125 x r vx 2. 175 = = 1. we would find that the linear approximation gives a value for Ve of Ve= 3. the skinfriction coefficient is ILUe c— I — T — PooUeS — — 2 1. 4. for constantproperty flow past a flat plate. Thus.60 X X 102 1. = f8 8* — dy 8 /1(1 — Notice that. We must also change the upper limit on our integral from 8 to 1.
0.664 1.176 1 Chap. if one uses a linear profile.8 to Figufe 4. For the more exact formulation.430 x Cf (3. . 4.4 0.4 0. the shear would be less than that for the exact solution.125 X (1.8 0.66 x Comparing the velocity profiles. the reader should be able to use physical reasoning to determine that these relationships are intuitively correct.10 X (1. where 8* and Ve would be greater for the linear profile.625 X Ve 0.6 (or !i) 0.7. That is.00 x (2.2 0.66 x Summarizing these calculations provides the following comparison: Linear approximation More exact solution 0. 4 / Viscous Boundary Layers 0.6 0.7 Comparison of velocity profiles for a laminar boundary layer on a flat plate. which are presented in Fig.
it clearly does not provide reasonable values for engineering parameters. which tends to delay transition. equation (4. such as 6* and Cf. if we had used the integral approach to determine the value of 6 for a linear profile. which are given in Table 4.25 x 10—2 which is the value obtained using the more exact formulation [i. The coordinates of this airfoil section.8.23)]..16)] for the NACA 65006 airfoil. .4.g. Solution: Using the definition for /3 gives us — — 2s due — 2f tte dx due dx dx ds 1 (Is — But dx = ds — (1 — Therefore. 8* and Cf) would be in closer agreement with those given by the more exact solution. at any chordwise location for a thin airfoil — — (1 — where = x/c.464x = 0.Sec. the values of the other parameters (e.3: Calculate the velocity gradient. A more realistic approximation for the streamwise velocity component in a laminar boundary layer would be = 4(f) (4. Although the linear profile for the streamwise velocity component is a convenient approximation to use when demonstrating points about the continuity equation or about Kelvin's theorem. we assumed that the boundarylayer thickness 6 was 1.38) — EXAMPLE 4.e. Note that the maximum thickness is located relatively far aft in order to maintain a favorable pressure gradient. are given in terms of the coordinate system used in Fig.866 X Although this is considerably less than the assumed (more correct) value.. Laminar Boundary Layer 177 In this example. /3 Calculate the velocity gradient parameter /3.3 / Jncompressible. we would have obtained 3. 4. The /3 distribution is required as an input to obtain the local similarity solutions for a laminar boundary layer. 4. However. which appears in the FalknerSkan form momentum equation [equation (4.
650 0.0000 +0. H.400 0.950 1.0123 0.8 Cross section for symmetric NACA 65006 airfoil of Example 4.100 —0. the negative values of /3 exceed that required for separation of a similar laminar boundary layer.165 —0.163 —0.098 +0. Note that a favorable pressure gradient acts over thO first half of the airfoil. New York: Dover Publications.0300 0.000 0.0051 0.0000 0.019 +0.0285 0.0252 0.143 —0. 4.166 —0.08 1 —0.050 0.005 0.0020 0. 1949 [Abbott and von Doenhoff (1949). von Doenhoff.0248 0. Because of the large streamwise variations in /3.100 0. . Theory of Wing Sections. E.0274 0.155 —0.0295 0. the nonsimilar Note is five times the x Figure 4.900 0.100 —0.6.0194 0.0182 1.3.056 +0. TABLE 4.000 —0.300 0.142 Source: I.0270 0.044 —0.350 0.250 0.000 0.134 —0.500 0.550 0.600 0.01 59 —0.178 Chap.013 0.750 0.0096 0.044 —0..073 —0.850 0.450 0.150 0. 4 / Viscous Boundary Layers Pressur e Distri bution for the NAcA 65006 C.025 0.4 0.700 0.800 0.120 0.145 —0.] The resultant /3 distribution is presented in Fig.0131 0. Abbott and A. 0.209 0.159 —0.124 —0.0290 0.0048 0.149 —0.9.
character of the boundary layer should be taken into account when establishing a separation criteria. Boundarylayer separation would result in significant changes in the flow field. even for this airfoil at zero angle of attack. if the boundary layer were laminar along its entire length. the boundary layer is turbulent over a considerable portion of the airfoil. boundarylayer separation will occur for any Joukowski airfoil that is thicker than 4. and separation is not as likely to occur. the experimental measurements of the pressure distribution indicate that the actual flow field corresponds closely to the inviscid flow field. The edge velocity and therefore the corresponding inviscid pressure distributions are shown in Fig.9 Distribution for NACA 65606 airfoil (assuming that the boundary layer does not separate). This thickening alleviates the adverse pressure gradients.0 x C Figure 4.3 / Incompressible. separation will not occur until a thickness of about 31% has been exceeded. The boundary layer effectively thickens the airfoil. 4. At the relatively high Reynolds numbers associated with airplane flight. However. Incompressible boundarylayer solutions were generated for symmetrical Joukowski airfoils at zero angle of attack. The fact that a turbulent boundary layer can flow without separation into regions of much steeper adverse pressure gradients than can a laminar boundary layer is illustrated in Fig.10. 4. it would separate.6% if the flow is entirely laminar. since creases with distance. a turbulent boundary layer can overcome an adverse pressure gradient longer. boundarylayer separation apparently does not occur at zero angle of attack. Thus. which in turn permits somewhat thicker sections before separation occurs. The reason that separation does not occur is as follows. Nevertheless. To . these calculations indicate that.10.Sec. especially near the trailing edge. As discussed previously. Laminar Boundary Layer 179 /3 1. However. At the conditions indicated. if the boundary layer is turbulent. 4.
The disturbances may be due to surface roughness.10 Thickest symmetrical Joukowski airfoils capable of supporting fully attached laminar and turbulent flows. transition is assumed to occur at the velocity peak. Pressure gradient 2.4 BOUNDARYLAYER TRANSITION As the boundary layer develops in the streamwise direction. 4. For some flows. Results for laminar flow are independent of Reynolds number. thus. such as the following: 1. The onset of transition from a laminar boundary layer to a turbulent layer (if it occurs at all) depends on many parameters. it is subjected to numerous disturbances.for turbulent flow. The change in the laminar case would be negligible. 4 / Vkcous Boundary Layers 2 (Ue\2 1 a Figure 4. Compressibility effects (usually related to the Mach number) 4. the turbulent airfoil would increase to about 33%. one might use vortex generators or other forms of surface roughness. delay or avoid separation altogether. Surface temperature 5. background noise. For turbulent flow. [From Cebeci and Smith (1974). and the Mach number is 0. such as shown in Fig. If displacementthickness effects on pressure distribution were included. Freestream turbulence .11.6%. 4. 31%. and so on. For other flows. these disturbances are damped and the flow remains laminar.] ensure that boundarylayer transition occurs and. The angle of attack is ØO. the disturbances amplify and the boundary layer becomes turbulent. temperature irregularities. Suction or blowing at the surface 6.The turbulent case is calculated for = 1O7.180 Chap. Surface roughness 3. Maximum thickness for laminar flow is about 4.
Obviously. surface roughness.40) Once the critical Reynolds number is exceeded. suction at the surface. a typical transition criterion is = 500. For incompressible flow past a flat plate.4 / BoundaryLayer Transition 181 Figure 4. blowing at the surface.000 (439) Thus. favorable pressure gradients. However. that is. . Stable. increased Mach numbers. Unstable flow containing twodimensional TollmienSchlichting (TS) waves 1.Sec. way of delaying separation (from Ruth Bertin's collection). adverse pressure gradients. the location for the onset of boundarylayer transition would occur at Rextr Xtr = (4. are an effective. no single criterion for the onset of transition can be applied to a wide variety of flow conditions. and surface cooling delay transition. transition criteria incorporate a Reynolds number. as a rule of thumb. but not necessarily an aerodynamically efficient. cause it to occur early.11 Vortex generators. Although the parameters used and the correlation formula for the onset of transition depend on the de tails of the application. and freestream turbulence promote transition. the flatplate boundary layer would contain regions with the following characteristics as it transitioned from the laminar state to a fully turbulent flow: laminar flow near the leading edge 2. which can be seen in front of the ailerons and near the wing leading edge of an A4. 4. Conversely.
182 Chap. such as an adverse pressure gradient and finite surface roughness. When one or more of the transitional regions are bypassed. Stability theory predicts and experiment verifies that the initial instability is in the form of twodimensional TS waves that travel in the mean flow direction. White. Smoke accumulation in the small recirculation regions associated with the TS waves can be seen at the left edge of the photograph. M. Fully turbulent flow 3. 4. 4 / Viscous Boundary Layers Toilmien Schlichting waves Three dimensional vortex breakdown Fully turbulent flow Growth of spanwise vorticity Generation of turbulent spots Figure 4. Downstream of the region where the spots first form. intense local fluctuations occur at various times and locations in the viscous layer. Transitionpromoting phenomena. The den appearance of three dimensionality is associated with the nonlinear growth region of the laminar instability. eliminating one or more of the five transitional regions described previously. A region where vortex breakdown produces locally high shear 5* Fluctuating.12. A vibrating ribbon perturbs the lowspeed flow upstream of the left margin of the photograph. may short circuit the transition process.) A region where threedimensional unstable waves and hairpin eddies develop 4. we term the cause (e. Even though the mean flow is two dimensional. From these local intensities. A sketch of the idealized transition process is presented in Fig. The experimental verification of the transition process is illustrated in the photograph of Fig. (Based on the sketch from Viscous Fluid Flow by F. threedimensional unstable waves and hairpin eddies soon develop as the TS waves begin to show spanwise variations.g.. true turbulence bursts forth and grows into a turbulent spot. . roughness) a bypass mechanism.13. threedimensional flow due to cascading vortex breakdown 6. In the advanced stages of the transition process.12 Idealized sketch of the transition process on a flat plate. the flow becomes fully turbulent. 4. A region where turbulent spots form 7. Copyright © 1974 by McGrawHill Book Company.
However. even when the inviscid (mean) flow is two dimensional. The effects caused by the fluctuations are as if the viscosity were increased by a factor of 10 or more. It should be noted that.13 Flow visualization of the transition process on a flat plate. Lockheed Aeronautical Systems Company. a turbulent boundary layer will be three dimensional because of the threedimensional character of the fluctuations. they have a decisive effect on the overall motion.) 4. Georgia Division. S. A turbulent flow is one in which irregular fluctuations (mixing or eddying motions) are superimposed on the mean flow.. Thus. (Photograph supplied by A. and they affect macroscopic lumps of fluid. Thomas. the reduction in form drag usually dominates the increase in skin friction drag. However. the velocity at any point in a turbulent boundary layer is a function of time. Turbu'ent Boundary Layer 183 Figure 4.The fluctuations occur in the direction of the mean flow and at right angles to it. since a turbulent boundary layer can negotiate an adverse pressure gradient for a longer distance.The size of these macroscopic lumps determines the scale of turbulence. 4. whereas momentum transport occurs on a microscopic (or molecular) scale in a laminar boundary layer.5 / Incompressible. it occurs on a macroscopic scale in a turbulent boundary layer. TURBULENT BOUNDARY LAYER Let us now consider flows where transition has occurred and the boundary layer is fully turbulent.Sec. the shear forces at the wall and the skinfriction component of the drag are much larger when the boundary layer is turbulent. the form drag). W. Therefore.e. although the velocity fluctuations may be only several percent of the local streamwise values. . boundarylayer separation may be delayed or even avoided altogether. For a blunt body or for a slender body at angle of attack. As a result.5 INCOMPRESSIBLE. Delaying (or avoiding) the onset of separation reduces the pressure component of the drag (i.
4 / Viscous Boundary Layers I Figure 4. and u' is the time dependent magnitude of the fluctuating component. turbulent boundary layer. twodimensional. For example. 13.42). 4. the continuity equation is + u') 3(13 + v') + (4. or eddying.5. The time average of any fluctuating parameter or its derivative is zero. the mean value for a steady flow is independent of time. constantproperty.1 Derivation of the Momentum Equation for Turbulent Boundary Layer Let us now derive the x (or streamwise) momentum equation for a steady.184 U Chap. motion. = 0. by definition. we see that U' = 0. When describing a turbulent flow.14. (4. as it should be.01 fundamental importance to turbulent motion is the way in which the fluctuations u'. The integration interval depends on the physics and geometry of the problem.[a(v')]/ax 0. Since the density is constant. it is convenient to express the local velocity components as the sum of a mean motion plus a fluctuating. For example. and ü'.41) u = + ii' where Li is the timeaveraged value of the u component of velocity. The timeaveraged value at a given point in space is calculated as 1 udt (4. 4.43) 0 3v . v'. and w' influence the mean motion ü.14 Histories of the mean component (n) and the fluc tuating component (u') of the streamwise velocity ii for a turbulent boundary layer. as illustrated in Fig. Referring to equation (4. The time average of prod ucts of fluctuating parameters and their derivatives is not zero. but u'v' 0.42) The integration interval should be much larger than any significant period of the fluctuation velocity u'. As a result.
au' ax . we find that pit'— + pu'—— ay au' av' 0 (4.45a) and that a' a' ax 3y (4. As noted when discussing equation (4. that is. 4. we obtain _au _an ay . Turbulent Boundary Layer 185 Expanding yields an ax 3y au' ax äv' äy (4. we learn from the continuity equation that ax + ay 0 (4.We obtain pit— + pu'— + pü— + pu'— = ax _au' au' ax av' ay av' ay 0 Taking the time average of these terms. we have p(ü + it') Expanding gives us + u') ax + P(V + v') + u') ay = du + a2(u + u') ay2 Pu— + Pv— + Pu'— + Pv'— + Pu'— + pv'— + pu— + pv— ax ay ax _au _an an an ay au' ax au' Bu' _au' ay ax = dUe + a2u + a2u' Taking the time average of the terms in this equation.5 I Incompressible.45b) Substituting the fluctuating descriptions for the velocity components into the x momentum equation (4. since their timeaveraged value is zero.45b) by p(i + u').42). ax ay Thus. the terms that contain only one fluctuating parameter vanish. Thus. the timeaveraged value of a fluctuating component is zero.Sec. The first two terms already are timeaveraged values. the time average of terms involving the product of fluctuating terms is not zero.47) .6).au' ay dUe a2u dx ay (4. However. for a turbulent flow.46) Let us now multiply the fluctuating portion of the continuity equation (4.44) Let us take the timeaveraged value for each of these terms.
The flux of momentum in the x direction is given by the product (u)(pv)(dA)(dt). (ä/äx)(u')2 as compared to the transverse gradient. These terms are related not only to physical fluid properties but also to local flow conditions. the momentum equation becomes 8x + 0V 8y = dx + 8y — p—(u'v') a (4.51) Mathematically. etc. and v' go to zero at the wall. and the directions x and z are in the plane of dA.50) — Recall that the first term is the laminar shear stress. geometry. the timeaveraged flux of momentum per unit time is dA = + dA Since the flux of momentum per unit time through an area is equivalent to an equalandopposite force exerted on the area by the surroundings. surface roughness. we can treat the term —pu'v' as equivalent to a "turbulent" shear stress. Thus. The term —pu'v' is the source of considerable difficulties in the analysis of a turbulent boundary layer because its analytical form is not known a priori. Thus." or Reynolds. and the flow for y is basically laminar.49) Let us further examine the last two terms. we can write Txy = — (4. this new variable can be defined only through an understanding of the detailed turbulent structure.186 Chap. (4. stress can be added to the stresses associated with the mean flow. then. This "apparent. Because the wall is a streamline. the turbulent inertia terms behave as if the total stress on the system were composed of the Newtonian viscous stress plus an apparent turbulent stress. It is related not only to physical properties of the fluid but also to the local flow conditions (velocity. The determination of the turbulent shearstress term is the critical problem in the analysis of turbulent shear flows. At points away from the wall. The mass of fluid passing through this area in time dt is given by the product (pv)(dA)(dt). that is. there is no flow through it. the magnitude of —pu'v' depends on the distance from the wall.). The term —pu'v' represents the turbulent transport of momentum and is known as the turbulent shear stress or Reynolds stress. we obtain + = due + 8 — — 2 (4. let us consider a differential area dA such that the normal to dA is parallel to the y axis.48) We will neglect the streamwise gradient of the timeaveraged value of the square of the fluctuating velocity component. For a constant density flow. upstream history. Thus. To evaluate the second term. Since .47) to (4.46) and rearranging the terms. However. 4 I Viscous Boundary Layers Adding equation (4. —pu'v' is the dominant term. Furthermore.
which have been developed from models of varying degrees of rigor.49). the grid resolution requirements are very stringent and increase dramatically with Reynolds number. Zeroequation models are well adapted to simple. It is often written that the Reynolds stress tensor is proportional to the mean strainrate tensor [i. Smith (1991) noted." However. 4.. the effect of flow history on the turbulence can be modeled. with zeroequation. For the RANS approach. Different results can be obtained with different implementation of the same turbulence model.. energy containing scales. empirically based correlations are introduced to model them. One example where modeling the flow history of turbulence is crucial is in the calculation . he further noted. different algebraic models must be applied to the different types of turbulent flows in a single flow problem.This means that all scales of turbulence must be modeled. Zeroequation models use no differential equations and are commonly known as algebraic models.5..Sec. the boundary layer is usually tur bulent. current approaches to turbulence modeling in clude direct numerical simulations (DNS).. more homogeneous scales while resolving the larger. Closure procedures for the turbulent eddy viscosity are generally categorized by the number of partial differential equations that are solved. "In general. and twoequation models being the most popular.2 Approaches to Turbulence Modeling For the flow fields of practical interest to this text. Turbulent Boundary Layer 187 there are no further physical laws available to evaluate these terms. they are more difficult for the user to apply. When this averaging process is applied to the NavierStokes equations (such as was done in Section 4. As discussed by Spalart (2000). oneequation.g.1). algebraic turbulence models remain the most popular choice due to their simplicity. Reynolds stress tensor takes into account the transfer of momentum by the turbulent fluctuations. largeeddy simulations (LES). Recently. Since algebraic models are accurate for a narrow range of flows.5 / Incompressible." By solving one or more differential equations.5. That is. the Reynolds stress tensor (e. the transport of turbulence can be included. 4. while algebraic turbulence models are computationally simple. the result is an equation for the mean quantities with extra term(s) involving the fluctuating quantities. The development of correlations in terms of known parameters is usually termed as the closure problem. Largeeddy simulations attempt to model the smaller.e. is known as the Boussinesq eddyviscosity approximation. There is a hierarchy of techniques for closure. or complex logic must be implemented to automate this process ... Because DNS must model all scales from the largest to the smallest. The RANS approach attempts to solve for the timeaveraged flow. "For solutions of external flows around full aircraft configurations. quantities of interest are time averaged. The user must define in advance which model applies to which region. and Reynoldsaveraged NavierStokes (RANS).e. the local production of turbulence is balanced by the local dissipation of turbulence). hybrid approaches that combine RANS and LES have been proposed in an attempt to combine the best features of these two approaches. The turbulent eddy viscosity is determined using a turbulence model.The direct numerical simulation approach attempts to resolve all scales of turbulence. attached flows where local turbulence equilibrium exists (i. such as that described in equation (4. This makes LES grid requirements less stringent than those for DNS. where is the unknown turbulent eddy viscosity.
" As a result. RANS models lack generality. In the twoequation models. The models are the result of generalizing and applying fundamental experimental observations.3 Turbulent Boundary Layer for a Flat Plate Since is a constant for a flat plate. and kkl models. Even with this simplification. which solves a single partial differential equation for a variable that is related to the turbulent kinematic eddy viscosity.g.5. nonphysical ways of describing the character of the physical situation of turbulence. there is no exact solution for the turbulent boundary layer. they are not governed by the physical principles of turbulence and they are not unique. Smith (1991) notes. zeropressuregradient (flatplate) boundary layers.188 Chap." Neumann (1989) notes. the fluid's physical properties. is one of the most popular turbulence models. One equation models are perhaps the simplest way to model this effect. including kc. for which it has been calibrated." The advantage of the RANS models is that they are relatively cheap to compute and can provide accurate solutions to many engineering flows.) Therefore. However. Very near the wall.. when deciding which turbulence model to use. while for algebraic models the normal stresses are entirely neglected. turbulence modeling becomes more of a factor in the accuracy of the solution. the pressure gradient term is zero. the normal Reynolds stress components are assumed to be equal." For a two equation model. the SpalartAllmaras model. models. Wilcox (1998). A variety of transport equations have been proposed for determining the turbulent length scale. The coefficients in the vari ous models are usually determined by matching the computations to simple buildingblock experimental flows (e. For an indepth review of turbulence models and their applications. Furthermore. "Turbulence models employed in computational schemes to specify the character of turbulent flows are just that . the viscous shear dominates. the errors can be significant." In evaluating computations for the flow over aircraft at high angles of attack. when discussing the oneequation model of SpalartAllmaras (1992). kw.. one transport equation is used for the computation of the specific turbulence kinetic energy (k) and a second transport equation is used to determine the turbulent length scale (or dissipation length scale). the model should have sufficient accuracy and suitable numerical efficiency for the intended applications. and the . 4 / Viscous Boundary Layers • • • of turbulent flow over a multielement airfoil. Smith (1991) notes that "at higher angles of attack. Ludwig Prandtl deduced that the mean velocity in this region must depend on the wall shear stress. the user should take care to insure that the selected turbulence model has been calibrated using measurements from relevant flow fields. noted that "it is especially attractive for airfoil and wing applications. Again. Experimental results show the streamwise Reynolds stress component to be two to three times larger than the normal component. the reader is referred to Wilcox (1998).. the reader is referred to Wilcox (1998) for detailed discussions of these models. the Reynolds shear stress is the dominant stress term in the momentum equations and the two equation models are reasonably accurate. For shear flows with only gradual variations in the streamwise direction. For more complex strain fields. 4.
To a first order. Thus. we obtain (y v)/u* Introducing the definition of the wallfriction velocity. the fluid physical properties.52). The edge of the laminar sublayer corresponds to a of 5 to 10. p.53a) yu* (4. ü is proportional to y. 4. . the velocitydefect law is given by Ue = (4. 189 y).52) = and (4.53c) Note that has the form of a Reynolds number. Specifically. Il is a function of (Tm. Prandtl deduced that the mean velocity in the inner region must depend on the wall shear stress. For the outer region. that is.53b) V where is called the wallfriction velocity and is defined as = (4.55) The outer region of a turbulent boundary layer contains 80 to 90% of the boundarylayer thickness 6.56) for the inner region. y. = Let us define = y (4. Substituting these definitions into equation (4. 6).Sec. the velocity defect (ue — Il) for the outer region is a function of (Tm. Turbulent Boundary Layer distance y from the wall. Thus. Theodor von Kármán deduced that. in the outer region of a turbulent boundary layer. In 1930. the velocities are so small that viscous forces dominate and there is no turbulence. the velocity profile in this region is linear. In 1933. = (4.5 I Incompressible. In the laminar sublayer. it is clear that = (454) +* uu + + + for the laminar sublayer. ü is a function of (Tm. Thus. and the distance y from the wall. the mean velocity ills reduced below the freestream value (Ue) in a manner that is independent of the viscosity but is dependent on the wall shear stress and the distance y over which its effect has diffused. Thus. y). p. p.
e. U Ue (Y As a result. A.Chap.5 The resultant velocity profile is presented in Fig. <400) is fully turbulent. For incompressible flow past a flat plate.41 to 5.15. 4.40 or 0. K 0. 70 < The velocity in the outer region is given by Ue =— in +A (4. the velocity in the inner region is given by = or.15 Turbulent boundary layer illustrating walllayer nomenclature.57a) the equation can be written as = +B (4.5Th) This velocity correlation is valid only in regions where the laminar shear stress can be neglected in comparison with the turbulent stress. 4 I Viscous Boundary Layers Since the velocities of the two regions must match at their interface. in terms of y+ +B (4. Thus. . the flow in this region (i.58) where K. Linear Buffer sublayer layer Viscous sublayer Inner layer 20 ÷ I / Wake" component = u 10  g Loglaw region1 Outer layer Defectlaw region I 1 10 100 + y—p — 1000 u*y Figure 4.. and B are dimensionless parameters.
both because of grid generation considerations and because of the need to develop turbulence models of suitable accuracy for the complex flowfield phenomena that may occur (e. equation (4. Boussinesq introduced the concept of eddy viscosity to model the Reynolds shear stress. pressure gradients. we will use only the timeaverage (or meanflow) properties. therefore. we can now write a relation between the eddy Viscosity and the mixing length: = 3)' (4. Thus.. While there are many turbuletice models of suitable engineering accuracy available in the literature.g. It was assumed that the Reynolds stresses act like the viscous (laminar) shear stresses and are proportional to the transverse gradient of the (mean) streamwise velocity component.59) and (4.Sec. the computational grid shohid.15. Referring to Fig. The distributions of Cm and of 1 across the boundary layer are based on experimental data. we will drop the overbar notation in the subsequent analysis. etc. the analysts should calibrate the partictilar model to be used in their codes against a relevant data base to insure that the model provides results of suitable accuracy for the applications of interest. the computational grid should include points in the.62) From this point on. 4. 4.6 / Eddy Viscosity and Mixing Length Concepts 191 The computation of the turbulent skinfriction drag for realistic aerodynamic applications presents considerable challenges to the analyst. Prandtl proposed the mixing length concept.ltiniinar sublayer. 8m is a function of the flow field (including such factors as surface roughness. has the Units of L2/T. or less. Because the eddy viscosity and the mixing length concepts are based on .51) for the total shear stress may be written T = T1 + Tt = + (au\ (4. In order to determine accurately velocity gradients near the wall.60) Like the kinematic viscosity ii.). 4.59) Having introduced the concept of eddy viscosity. whereby the shear stress is given as p12 (4. However. contain points at a of 5. In an attempt to obtain a more generally applicable relation. The coefficient of the proportionality is called the eddy viscosity and is defined as —p = (4. whereas v is a property of the fluid and is defined once the pressure and the temperature are known.61) Equating the expressions for the Reynolds stress. given by equations (4. viscous/inviscid interactions).6 EDDY VISCOSITY AND MIXING LENGTH CONCEPTS In the late nineteenth century.61).
63) where.02604 1+1I 0.< = 0.67c) where Re6 is the Reynolds number based on momentum thickness. However.e. is used to calculate the eddy viscosity until its value becomes equal to that given by the outer region expression. equation (4. when the eddy viscosy (Cm)oI.66).64b) (u*) dx Thus. is A 26v Nu* (4. . The y coordinate of the interface between the inner region and the outer region is determined by the requirement that the y distribution of the eddy viscosity be continuous.65) For the outer region.192 Chap.26). 4 / Viscous Boundary Layers local equilibrium ideas. equation (4.28).15 the interface value..67a) 0. For the inner region.298z1)] (4. see equation (4. A general conclusion that is drawn from the experimental evidence is that the turbulent boundary layer should be treated as a composite layer consisting of an inner region and an outer region. a= LI 0.66).64a) where is the wall friction velocity as defined by equation (4.55[1 — exp — (4.6Th) and — 1 (4. the mixing length is given as = Ky[(I — exp (—y/A)] (4.65).53c) N= and (1 — (4. the characteristic dimension in the expression for Re0 is the momentum thickness. the eddy viscosity for the inner region becomes (Ky)2[1 (4.66) where is the displacement thickness as defined by equation (4. that is. Thus. where the two expressions for the eddy viscosity are equal [i.64c) . The y coordinate. the original derivations included erroneous physical arguments. ity is calculated using the outer region expression. the inner region expression.exp = aIte8* (4. they are relatively simple to use and provide reasonable values of the shear stress for many engineering applications. In fact. the van Driest damping parameter. they provide only rough approximations to the actual flow and are said to lack generality. equation (4. the eddy viscosity is given as (4.41 (as discussed earlier) and A.
The results are presented in Fig. Following a suggestion by Prandtl.17. For most practical boundarylayer calculations. Since we will use only the timeaveraged (or meanflow) properties in this section. we will drop the overbar notation. We shall use the meanflow properties in the integral form of the equation of motion to develop engineering correlations for the skinfriction coefficient and the bound arylayer thickness for an incompressible. the expressions for the eddy viscosity are multiplied by an intermittency factor. That is. transitional. 4.35 X (4. 4. the turbulent velocity profile will be represented by a powerlaw approximation.16. and turbulent boundary layer for flow past a flat plate. that is. Thus.69a) and (Sm)0 (4. Consider the control volume shown in Fig. turbulent boundary layer on a flat plate.69b) Solutions have been obtained using these equations to describe the laminar.Sec. the boundary layer does not instantaneously change from a laminar state to a fully turbulent profile. 4.7 / Integral Equations for a FlatPlate Boundary Layer 193 Recall that the transition process occurs over a finite length. Although approaches using the differential equations are most common in computational fluid dynamics.68b) The intermittency factor varies from 0 (in the laminar region and at the onset of transition) to 1 (at the end of the transition zone and for fully turbulent flow).68a) where Xtr is the x coordinate for the onset of transition and 3 G= 8. it is necessary to calculate the viscous flow along its entire length. 4. in the transition zone. (Ky)2[1 — exp (4. the integral approach can also be used to obtain approximate solutions for a turbulent boundary layer. for a given pressure distribution (inviscid flow field) and for a given transition criterion. the boundarylayer calculation is started at the leading edge or at the forward stagnation point of the body (where the boundary layer is laminar) and proceeds downstream through the transitional flow into the fully turbulent region. To treat the boundary layerin the transition zone.7 INTEGRAL EQUATIONS FOR A FLATPLATE BOUNDARY LAYER The eddy viscosity concept or one of the higherorder methods are used in developing turbulent boundarylayer solutions using the differential equations of motion. Ytr exp [ G(x — Xtr) (4. Note and the velocity that the freestream velocity of the flow approaching the plate of the flow outside of the boundary layer adjacent to the plate (Ue) are equal and are .
0 0 0 yrn 0.75) for turbulent 1x I I I I I I = Cf = Cf 1 x Transition zone hytr I I I I I 0. tie 114.0 0. 4 / Viscous Boundary Layers TNSBLM (laminar. transitional.4 0.2 0 00 8 a 0.194 Chap. (b) turbulent viscosity/thermodynamic viscosity ratio.5 X I I I 0.228m 0.2 x(m) (a) ox= 1.— .01 DO 0. . 540°R: (a) skin friction distribution.1 ft/s. = 2101.0 1.6 0.1 :1 10 100 (b) Figure 4.2 0.8 1.16 Sample of computed boundary layer for incompressible flow past a flat plate.165m x=0. Te = 542°R.21) for laminar — .320m (near the end of transition) o x1.4  0 0.223m (well into the fully turbulent flow) (near onset of transition) 0. (4.— Eq.8 0 Q 0 0.5 psI..6 .6) Ox=0. (4. turbulent) Eq.
13) yields ry rYo —d= / Jo u(pudy) — Jo I .72) Thus. of course.Sec. The wall (which is. has zero shear force acting across it) will do.5) yields I Jo u dy UeY0 = 0 (4. the continuity equation (2. the integral is independent of the upper limit of integration. Fl dA is zero for both boundary streamlines. Since the integrand goes to zero for y 6. udYJ [Ue+(uue)]dy J0. 4. used interchangeably. provided that it is at. UeY + I (u — Jo lie) dy (4.70) But also. Thus. 0 p1' = thickness. application of the integral form of the momentum equation (2. a streamline) is the inner boundary of the control volume. we have used both 6 and Y as the upper limit for the integration. with Y > Y0. Since V.17 Control volume used to analyze the boundary layer on a flat plate. we have derived the expected result that the outer streamline is deflected by the transverse distance In developing this relation.71) Combining these two equations and introducing the definition for the displacement = we find that — y — = (4. Similarly. Any streamline that is outside the boundary layer (and.7 / Integral Equations for a FlatPlate Boundary Layer 195 x Figure 4. A streamline outside the boundary layer is the outer boundary. the outer boundary is not parallel to the wall. the edge of the boundary layer. Because the viscous action retards the flow near the surface. therefore. or beyond. the streamline is a distance Y0 away from the wall at the initial station and is a distance Y away from the wall at the downstream station.
28) Note that the result is independent of the upper limit of integration provided that the upper limit is equal to or greater than the boundarylayer thickness. Thus. Therefore.7. t3UUe 1 1 ay — 7 6l/7y6/7 . the drag coefficient (for one side of a flat plate of length L) is Cd = (4. note that the velocity gradient for this profile. However.196 Chap.e. ry d= Using equation (4. As discussed earlier. u(y)} in order to achieve "fairly accurate" drag predictions. —=(—) \6J Ue when the local Reynolds number Re. an analytical form for the turbulent shear is not known a priori. FlatPlate Boundary Layer Now let us apply these equations to develop correlations for a turbulent boundary layer on a flat plate..76) 1 x i07.1 Application of the Integral Equations of Motion to a Turbulent. because the integration often averages out positive and negative deviations in the assumed velocity function. 4. 4 1 Viscous Boundary Layers Note that the lowercase letter designates the drag per unit span (d).70)..75) The equations developed in this section are valid for incompressible flow past a flat plate whether the boundary layer is laminar or turbulent.The value of the integral technique is that it requires only a "reasonable" approximation for the velocity profile [i. we find that 1Y — Jo I (pu2 dy) (4. Experimental measurements have shown that the timeaveraged velocity may be represented by the power law.73) d PUe I JO udy — Jo I pu2dy This equation can be rewritten in terms of the section drag coefficient as Cd= 2 d / = —( / —dy L\Jo Ue u u2 — J0 dy Ue (4. Thus.74) Recall that the momentum thickness for an incompressible flow is = (1  (4. we need some experimental information. is in the range 5 x u (4.
(x)hhl]d(x)} / \o. &7 / Integral Equations for a FlatPlate Boundary Layer 197 goes to infinity at the wall. Furthermore. the thickness of a turbulent boundary layer on a flat plate is given by 8 — 0. we obtain fl = which becomes .78) Substituting equations (4.77) yields = 0. we need another piece of experimental data: a correlation for the shear at the wall.77) Differentiating equation (4. although the correlation given in equation (4.3.69) gives us = dx Ue 8 (4.78).J dx 0.79) with the laminar correlation given by equation (4. a total skinfriction coefficient can be found for turbulent flow by integrating equation (4.23).0456( \Ue81 1 (4.1.Sec.77) into equation (4.25 6°25d6 0.76) pro vides a reasonable representation of the actual velocity profile. we see that a turbulent boundary layer grows at a faster rate than a laminar boundary layer subject to the same conditions. at a given x station.0583 (4. Thus. Substitution of equation (4. Blasius found that the skin friction coefficient for a turbulent boundary layer on a flat plate where the local Reynolds number is in the range 5 X i05 to 1 X i07 is given by Cf = 7 1 2 0.80) As with the laminar skinfriction coefficient found in Sec. 4. If we assume that the boundarylayer thickness is zero when x = we find that 8= Rearranging.3747 x (Rex) Comparing the turbulent correlation given by equation (4.23451 .81) ' = 0.0583 dx (4.76) and (4.80) over the length of a flat plate: — Cf = 1 U0 I C1 (x)dx = 1 1L U0 I 0. a turbulent boundary layer is thicker than a laminar boundary layer for the same stream conditions.79) into equation (4.074 (ReL)°2 .
which should usually be used instead of the Prandtl theoretical relation. you start by evaluating the entire plate by assuming that the boundary layer is turbulent along the entire length of the plate. since the drag coefficient is not explicitly represented. Therefore.Chap. equation (4.81). The calculation of the skinfriction drag for a flat plate with transition theoretically would require using the local skinfriction coefficients. known as the Prandtl formula.427 (log10 ReL — 0. since they already rep resent the integrated skin friction over the entire plate from x = 0 to x = L.21) for the laminar care to portion of the flow and equation (4.32) for the laminar portion of the flow and equation (4.82) for the turbulent part of the flow.13 log18 (ReL ±2% accurate (4. when compared with experimental data. is an exact theoretical representation of the turbulent skinfriction drag. when using the total skinfriction coefficients.407)2. which are equation (4. the most accurate relation which is also straightforward to use is the PrandtlSchlichting relation.82) KarmanSchoenherr: VCf SchultzGrunow: — 4. this would be: CD = — Lb — — + — Xtrb (4. Since you want to simulate the plate with both laminar and turbulent boundary layers present. However.83) Cf 0. A number of other empirical and semiempirical turbulent skinfriction coefficient relations also have been developed. 4 I Viscous Boundary Layers This formula. it re quires an iterative solution method to obtain a result.86) .80) for the turbulent part of the only integrate each relation over the laminar and turbulent lengths. In equation form. 4.64 ±7% accurate (4. the process shown in Fig. Care must be taken.58 ±3% accurate (4. respectively. some of which are considerably more accurate than the Prandtl formula [White (2005)1: PrandtlSchlichting: C (log10 Rei )2. In order to properly simulate a flat plate with transitional flow present. equation (4.84) While the KarmanSchoenherr relation is the most accurate of these relationships. however.18 should be used. Since the distance from the leading edge of the plate to the transition location should be evaluated with laminar flow. CD = 111'tr / Cf1 dx + / C1 b dx ) ) (4. it is found to be only ±25% accurate.85) Another approach is to use the total skinfriction coefficients. that portion of the plate should be evaluated with the turbulentflow skinfriction relation and also with the laminarflow skinfriction relation by subtracting the turbulentflow drag from the total plate turbulent results and adding the laminarflow portion.
4.000.A. (1967)].000.000 1.000 500.5 Rextr A 1050 300. A good rule of thumb is to assume that if transition takes place at less than 10% of the length of the plate. in equation (4. varies depending on the transition Reynolds number. Empirical Relations for Transition COrrection [Sch I ichti ng (1979).32).7 / integral Equations for a FlatPlate Boundary Layer b 199 + Figure 4.] TABLE 4.82): — cf=_ 0. through transition.000. in equation (4.455 (log10 A ReL (4. et a!.87) where the correction term reduces the skin friction since laminar boundary layers produce less skin friction than turbulent boundary layers. A more straightforward approach to model transitional flow is to use an empirical correction to the PrándtlSchlichting turbulent skinfriction relation [Dommasch. as shown in Table 4. since it is relatively small.18 Calculation of skinfriction drag coefficient using total skinfriction coefficients. The experimentally determined constant. The value A = 1700 represents the laminar correction for a transition Reynolds number of Rextr = 500. It can be seen from this formulation that if the Reynolds number at the end of the plate is very high.000 1700 3300 8700 . then the laTminar correction term plays a fairly insignificant role in the total skinfriction drag on the plate. and finally to the fully turbulent value. then the laminar correction usually can be ignored.82).Sec.000 3.5.19 shows how the total skinfriction coefficient varies from the laminar value in equation (4. Figure 4.
000. tr = 0.200 Chap.2691 kg/rn3 + 110. assuming that the transition process is completed instantaneously at that location.12344m .] EXAMPLE 4. flat plate.19 Variation of total skinfriction coefficient with Reynolds number for a smooth. et al. 1.556 rn/s We will assume that the transition Reynolds number for this incom pressible flow past a flat plate is 500.01325 X N/rn2 = (278. Thus.458 x 10627815 = 1.001 106 108 1010 ReL = pUL Figure 4.4 = 1. (1967). Solution: For air at atmospheric pressure and 5°C. Compare the velocity distribution for a laminar bound ary layer and for a turbulent boundary layer at the transition point.7404 X and (200 km/h) (1000 m/km) UOO = 3600s/h = 55.15)1.5 = 1.4: Computing the velocity profiles at the "transition point" Air at standard sealevel atmospheric pressure and 5°C flows at 200 km/h across a flat plate. 4 / Viscous Boundary Layers Cf 0. [From Dommasch.
Thus.78 48.64 52.00017 0. (This assumption would be reasonable far downstream of the transition location so that Nevertheless.00000 0.374 8turb = 3.78 33.7 / Integral Equations for a FlatPlate Boundary Layer 201 The thickness of a laminar boundary layer at this point is 5.00067 0.4 Uturb (m) (mis) (mis) 0.00235 0. it is not realistic to use the assumption that the boundary layer is turbulent all the way from the leading edge.6 V Velocity Profiles for Example 4.00302 0.14 46. The macroscopic transport of fluid in the y direction causes both increased shear and increased thickness of the boundary layer. the flow is continuous at the transition location and the boundarylayer thickness does not change instantaneously. Furthermore.00101 0. Note that the streamwise velocity component u increases much more rapidly with y near the wall for the turbulent boundary layer.00 36.32 51. TABLE 4. 4.98 44. assuming that the boundary layer is turbulent all the way from the leading edge.50 0.00201 0.00168 0.56 .Ox v = 8.79 53.00268 0.353 x m In reality. the object of these calculations is to illustrate the charx acteristics of the turbulent boundary layer relative to a laminar boundary layer at the same conditions.00034 0.729 x 4 m For comparison.21 Inviscid flow 39. Thus.6 and Fig.00335 0. since we are at the transition location. The resultant velocity profiles are compared in Table 4. 0.Sec.81 54.00134 0.20.00 17.74 50. 4. the shear at the wall is greater for the turbulent boundary layer even though this layer is much thicker than the laminar boundary layer for the same conditions at a given x station.71 55.33 52. we will calculate the thickness of the turbulent boundary layer at this point for this Reynolds number.
for a given external velocity distribution.7. H.002 0.001 0. the momentum thickness.91) . the displacement thickness. H and Cf.88) where 8.26). is defined as H=— 0 6* (4. 4.202 Chap.2 Integral Solutions for a Turbulent Boundary Layer with a Pressure Gradient If we apply the integral equations of motion to a flow with a velocity gradient external to the boundary layer. we obtain dx Ue dx 2 (4. is defined in equation (4.003 y(m) 0.4.90) where H1 is defined as H1 = 6 3* (4.88) contains three unknown parameters.000 10 20 30 40 50 60 u (mis) Figure 4. 4 / Viscous Boundary Layers 0.28).004 0. Equation (4. was defined in equation (4. the momentum shape factor. 0.89) where 8*. these parameters are interrelated in a complex way. Head (1969) assumed that the rate of entertainment is given by (4.20 Velocity profiles for Example 4. For a turbulent boundary layer.
4 = 0). Thus. lowspeed flow past a flat plate. A curvefit formula given in White (2005) is Cf = (log (4. We will consider flows where the boundary layer is laminar.32) becomes aT + = 32T (4. 0. H. values for two of the three parameters. the viscous dissipation is negligible (i. equation (2.93) provide a relationship between 0 and H.8234(H — 0. which will be defined in Chapter 8.6 1.6778)_3. laminar boundary layer.3 + 3.92) — — O. Cor relations of several sets of experimental data that were developed by Cebeci and Bradshaw (1979) yielded F= and G — J0.0M + 3.3 forH 1. A relation between Cf. and H is needed to complete our system of equations. dp/dt = 0.32).0306(H1 — 30)_06169 (4. Let us examine one such flow. see equations (4.8 to 2. Although it is not possible to define an exact value of H corresponding to the separation. To start the calculations at some initial streamwise station. that is. there are many constantproperty flows for which we are interested in calculating the convective heat transfer. in order to determine the temperature distribution.95) We can numerically solve this system of equations for a given inviscid flow field. Using this method. for a calorically perfect gas.8 THERMAL BOUNDARY LAYER FOR CONSTANTPROPERTY FLOWS As noted earlier.94) where Re0 is the Reynolds number based on the momentum thickness: Re9 = (4.8. the shape factor H can be used as a criterion for separation. the thermal boundary layer for a steady. 4. constantproperty. The solution for the velocity field for this flow has been described earlier in this chapter. Thus. We will now solve the energy equation (2. such as the transition location.90) through (4.94). 4.96) .e. and Cf.8 I Thermal Boundary Layer for ConstantProperty Flows 203 Head also assumed that H1 is a function of the shape factor H.Sec. 0. For flow past a flat plate.6 forH Equations (4. the value of H for separation is usually in the range 1. must be specified at this station. For a lowspeed.The third parameter is then calculated using equation (4.6) through (4. H1 = G(H). the temperature variations in the flow field are sufficiently large that there is heat transfer to or from a body in the flow but are small enough that the corresponding variations in density and viscosity can be neglected..27).
equation (4.e.. where TeTw Note that 0 = 1 at the edge of the thermal boundary the wall (i.98) becomes + Note that pv— = 02u* (4.e. Note that the equations are identical if k/cr = Furthermore.98) Let us replace u in the derivatives by the dimensionless parameter. Let us now change the dependent variable from T to the dimensionless parame ter 0.100) The Prandtl number is an important dimensionless parameter for problems involving con vective heat transfer where one encounters both fluid motion and heat conduction. if k at the the velocity and the thermal boundary layers are identical. U =— where Thus. and 0 layer. . at y = 0). Using 0 as the dependent variable.99) 0 at the wall (i. at y = 0).6).. and 0 = 1 and = 1 at the edge of the boundary layer. the velocity at the edge of the boundary layer (Ue) is constant.97) Since the pressure is constant along the flat plate. see equation (4. Thus. 4 / VIscous Boundary Layers Note that we have already neglected k(82T/8x2) since it is small compared to k(32T/3y2). and the momentum equation becomes pu— + pv— 3y 82u (4.204 Chap.97) and (4. and 1 at the edge of the velocity = 0 boundary layer. Compare equations (4. This ratio is called the Prandtl number (Pr) in honor of the German scientist: jtC Pr = (4. the energy equation becomes 0 at pu — + pv — = 8x 8y 80 80 k820 Ce 8y2 (4.98). the boundary conditions are identical: 0 = 0 and wall. We made a similar assumption about the corresponding velocity gradients when working with the momentum equation.
which is a dimensionless heattransfer coefficient. if the Prandtl number is 1.101) The rate at which heat is transferred to the surface defined as (4.105) Note that if = Pr = 1.1). C 2 (4.106) This relation between the heattransfer coefficient and the skinfriction coefficient is known as the Reynolds analogy.104).104) PUeCp ay Relating the Stanton number. is defined as St Ch = PLteCp(Te T20) (4. we obtain the ratio Cf St au*/ay = k ao/ay (4. the skinfriction coefficient for a flat plate is — T 2. as defined by equation (4.1 Reynolds Analogy The shear at the wall is defined as I T/L \ ay Therefore. tO the skinfriction coefficient.Sec.101).8 / Thermal Boundary Layer for ConstantProperty Flows 205 4. as given by equation (4. 4.8.Läu* 2= is —1 (4.102) = y=o The Stanton number (designated by the symbols St or C. then — 80 3y Thus.103) Combining these last two expressions. we find that the Stanton number is St = (4. .
Solution: Using the results from Example 1.7) over a wide range of flow conditions. k= (288.738 Note that. But we have already obtained the solution for the stream function. 2. the viscosity is 1. that is. as a rule of thumb.108a) + (Pr)fO' = 0 where the ' denotes differentiation with respect to 17.15)15 4.206 Chap.7894 X 105kg/s.2 Thermal Boundary Layer for Pr 1 To solve for the temperature distribution for the laminar.we have f= 'F.8. as defined by equation (4. — (4.15 = 5. 4 / Viscous Boundary Layers EXAMPLE 4.. flatplate boundary layer. that is.K) 2.76 X 400.13). at 288.e.15 K? kg/s m.436 X = k = (1.819 X 105ca1/cmsK Noting that there are 4.436 >< 102J/m•s•K = 0.16) for a flat plate (i.187 J/cal. the thermal conductivity is k= Thus. approximately 2000 K at atmospheric pressure. equation (4. Using the equation to calculate the thermal conductivity.10): Using the transformed stream function f.107) over the range of temperatures below those for which oxygen dissociates.85) becomes (4. What is the Prandtl number for air at 15°C. the Prandtl number for air is essentially constant (approximately 0.7 J/kg K. 13 = 0).5: Calculating the thermal properties of air The thermal conductivity of air can be calculated using the relation k= 4. Referring to equation (4.m)(1004. 4. let us introduce the transformation of equation (4.3.108b) .7894 X The specific heat is 1004.76 X 106T +112 cal/cm s K (4.7J/kg.
0 1. the Stanton St = 0. we find that = 0. Thus.103).113a) In this equation.Sec.111) The heat transfer can be expressed in terms of the Stanton number using equation (4.110) The rate at which heat is transferred to the wall can be calculated using = = k(Te — • ay Using the values of Pohihausen. and (2) for large.332 • (4. we obtain Integrating twice yields 8= C f (fll)Prd + (4. The Nusselt number is defined as (4.0 = 0.4696(Pr)°333 Combining these two relations.332k(Te — (4. which is defined as h (4. 8= TeTw = 1— j 00 (f) .108b) and rearranging.108a) and (4. St = number is /LCp PUeX Using the definitions for the Reynolds number and Prandtl number. 4.8 / Thermal Boundary Layer for ConstantProperty Flows 207 Combining equations (4.. the rate at which heat is transferred from a laminar boundary layer to the wall is given by the relation O.109) where C and 0o are constant of integration. Pr fOo(flP)Prdfl (4.113h) Te . Ii is the local heattransfer coefficient. They can be evaluated by applying the boundary conditions (1) at = 0.112) Another popular dimensionless heattransfer parameter is the Nusselt number.
we can see that the = is given as Thus. equation (4.112) and (4.6938 X = 0. St and for a turbulent flatplate boundary layer.21). 4 / Viscous Boundary Layers Combining this definition with equation (4.6: Calculating the heattransfer rate for a turbulent boundary layer on a flat plate Using the modified Reynolds analogy.9092 kg/rn3 kg/s 'm .114) By dividing the expression for the Stanton number. we shall call this the modified Reynolds analogy.0 m when the airplane is flying at 468 km/h at an altitude of 3 km? The surface temperature is 330 K. we can approximate the Stanton number as St = 0.114).7: Calculating the heat transfer The radiator systems on many of the early racing aircraft were flush mounted on the external surface of the airplane.111) and (4J13a) gives us = (Pr)°333 (4.106). the Nusselt number for turbulent flow past a flat plate can be approximated as (4. by that for the skinfriction coefficient.115) Because of the similarity between this equation and equation (4.112). Solution: Referring to the discussion of turbulent boundary layers.80) Thus.659 K = 1.115) for the mOdified Reynolds analogy. Let us assume that the local heattransfer rate can be estimated using the flatplate relations. Solution: Using Table 1. using equation (4.012 X N/rn2 = 268. EXAMPLE 4. 0292 (4. we obtain Cf St 2(Pr)°667 (4. develop relations for the dimensionless heattransfer parameters.2 to find the freestream flow properties.208 Chap. equation (4.116) Comparing equations (4.0583 (4.117) EXAMPLE 4. = 7. What is the local heating rate for x = 3. we note that C 0.
these values are also the local properties at the edge of the boundary layer at x = 3.8(0. Combining equations (4.8 / Thermal Boundary Layer for ConstantProperty Flows 209 Since we have assumed that the flow corresponds to that for a flat plate.093 x i07 1. the Prandtl number is Pr Thus. In fact.000.0292(Rex)°'8(Pr)°'333k(Te x Ti. = 0. 4.093 X 107)0.5 — To calculate the thermal conductivity for air.738 (0. Since there are 1.506 = 2.306 X cal/cm s K X 102J/msK Since 1 W 1 J/s. Note that Ue = = 468 km/h = 130 rn/s To determine whether the boundary layer is laminar or turbulent. This is as it should be.Om = —8. .944 X i03 W/m2 = —8.9092)(130)(3.6938 X This is well above the transition value. transition would occur at a point 500. proper performance would produce cooling. the calculation of the heating will be based on the assumption that the boundary layer is turbulent over its entire length.113b).738)0.Sec. k= 2.072 m from the leading edge.114 hp/ft2.333(2.117).O) = 2.113a) and (4. since the surface is hotter than the adjacent air.341 hp/kW. q= k 0.944 kW/m2 The minus sign indicates that heat is transferred from the surface to the air flowing past the aircraft. Thus. the heat transfer rate is equivalent to 1. let us calculate the local Reynolds number: = PeUeX = (0.000 Xtr 0.306 x 102 W/m ' K Furthermore.659 — 330)K 3.0292)(2.76 X 10*6 T+112 = 5.306 X 102W/mK)(268. Furthermore.0m. if the transition Reynolds number is assumed to be 500.Thus. NUxk(Te where the is given by equation (4. since the problem discusses a radiator. 4.
4. Techniques have been developed both for a laminar and for a turbulent boundary layer using both integral and differential approaches. incompressible flow over a cylinder..3. Assume that the inviscid external flow over a configuration is given by ue = Ax Thus.e.6. v(x. = 0 for steady. Assume that the wall is porous so that there can be flow through the wall.4 and assuming that the boundary layer is laminar. . the stagnation point) is 6 0 and that 6 increases in the streamwise direction. that is. we will assume that the windward plane of symmetry (i.5. 4 / Viscous Boundary Layers 4. determine if f(0) = —0..14). 4. A very thin.4.5 m)? Assume that the boundary layer is laminar in answering the remaining questions. 0) = 0. What are the remaining two boundary conditions? 4. viscous boundary layer near the surface.9 SUMMARY In this chapter we have developed techniques by which we can obtain solutions for a thin. "flatplate" wing of a model airplane moves through the air at standard sealevel conditions at a velocity of 15 m/s.5 m and its span (length perpendicular to the flow direction) is 5 m. incompressible flow past a flat plate. at x = 0). The inviscid velocity (Ue) is 50 ft/s with standard atmospheric conditions. What is the relation between the shear at a given value of x this flow and that for a flat plate? 4. What are the boundarylayer thickness and the displacement thickness at the trailing edge? What are the local shear at the wall and the skinfriction coefficient at x = 0. at x = 0. We now have reviewed the basic concepts of fluid mechanics in Chapters 1 through 4 and are ready to apply them to aerodynamic problems. 4. determine the value of f"(O).210 Chap. Consider twodimensional. Obtain the expression for /3. 4. Using equation (4. The dimensions of the plate are such that its chord (streamwise dimension) is 0. Transpiration (or injecting gas through a porous wall into the boundary layer) is to be used to reduce the skinfriction drag for steady. laminar flow past a flat plate.4. 4. Thus. that is.5 m.25. Ue = sin 6 and x = RO Determine the values of /3.1. at 0 = 30°. Using the equation developed in Problem 4. show that — f(°) in order to have similarity solutions. the value of the shear function at the wall.5 m? Calculate the total drag on the wing (both sides). We plan to use suction through a porous wall as a means of boundarylayer control. Using the equation developed in Problem 4. Prepare a graph of It as a function of y.2. and at 6 = 90°. at 0 = 45°. Using Fig. For ease of use with the nomenclature of the current chapter. PROBLEMS 4.What is the Reynolds number at the trailing edge (x = 0. where It designates the x component of velocity relative to a point on the ground. that is. determine f(0) if for steady flow = past a flat plate where tie = 10 rn/s at standard sealevel conditions.e. the stagnation point occurs at the leading edge of the configuration (i.
.7. fluid flows through the upper boundary with a velocity Ve which is a function of x.7894 X N/rn2 kg/rn s = 100 m/s = 1.9. P4. calculate (8*/x) Compare these values with those presented in the and chapter that were obtained using the more exact differential technique = 5.Problems 4. If the viscous flow is incompressible.g. 4.3.2250 kg/rn3 . What is the total friction drag acting on the wing? What is the drag coefficient? 4.70)? U Uc. How does the resultant expression compare with equation (4. [e. the outer boundary of our control volume was a streamline outside the boundary layer (see Fig.8. Thus.5 to define 6 when calculating y/6.10. which and (e) = 5. Owing to the growth of the boundary layer.7 4.11. When we derived the integral equations for a flatplate boundary layer.g. 4. 4. For this profile.. For the differential solution.0]. to describe the incompressible flow past a flat (d) plate.. Use the integral momentum analysis and a linear velocity profile for a laminar boundary layer Ue6 U _Y where 8 is the boundary layer thickness. compute (a) (c) (b) Compare these values with those presented in the text. Let us represent the wing of an airplane by a flat plate.0]. the outer boundary is a line parallel to the wall and outside the boundary layer at all x stations. A fiat plate at zero angle of attack is mounted in a wind tunnel where = 1. Prewere obtained using the more exact differential technique [e. U= Ue = (a constant) Figure P4. Use the integral momentum analysis and the assumed velocity profile for a laminar boundary layer: u 3(y'\ 1(y'\3 2U) where 8 is the boundarylayer thickness.01325 x = 1.7.17). Let us now apply the integral equations to a rectangular control volume to calculate the sectional drag coefficient for incompressible flow past a flat plate of length L.The dimensions of the wing are chord = 5 ft and span = 30 ft. The airplane is flying at standard level conditions at 180 mph. as shown in Fig. use 7) = 3. pare a graph comparing this approximate velocity profile and that given in Table 4.
11 4. A Pitot probe is used to determine the velocity profile in the viscous region downstream of the airfoil.13)1 to the flow between the two streamlines bounding this wake.5)] can be used to relate . P4. u/Ue = (y/6)117]. Present the predicted values as (1) The difference between that sensed by the Pitot probe and that sensed by the static port in the wall [i.0 m from the leading edge (Fig. A thin symmetric airfoil section is mounted at zero angle of attack in a lowspeed wind tunnel.e. 4 / Viscous Boundary Layers A Pitot probe is to be used to determine the velocity profile at a station 1.13. Use Table 4.. (a) Using a transition criterion that = 500.y versus Pt(Y) — pstatic] (2) The pressure coefficient Pt(Y) — 1 Note that for part (c) we can use Bernoulli's equation to relate the static pressure and the velocity on the streamline just ahead of the probe and the stagnation pressure sensed by the probe." Thus.11).000. as shown in Fig. assuming that the transition process is completed instantaneously at that location. (d) Is the flow described by this velocity function rotational or irrotational? I— = = 100 rn/S u(y) I Pitot probe x=1.. Even though this is in the boundary layer. P4.__ 212 Chap. we can use Bernoulli's equation. calculate the pressure you should expect to measure with the Pitot probe Pt(Y) as a function of y. since we relate properties on a streamline and since we calculate these properties at "point. The integral continuity equation [equation (2.o If we apply the integral form of the momentum equation [equation (2.13.79) to calculate the thickness of the turbulent boundary layer at a point 1. (c) If the streamwise velocity varies as the th power law [i.12. we can calculate the drag force acting on the airfoil section. where does transition occur? (b) Use equation (4.e. the flow slows down isentropically to zero velocity over a very short dis tance at the mouth of the probe. The resultant velocity distribution in the region —w z +w u(z) = Uc. Compare the streamwise velocity as a function of y for a laminar boundary layer and for a turbulent boundary layer at the transition point.3 to define the laminar profile and the oneseventh power law to describe the turbulent profile. Air at atmospheric pressure and 100° C flows at 100 km/h across a flat plate.OOrn Static port Figure P4. 4.00 m from the leading edge.
The boundary conditions that were used in developing the equation for the laminar thermal boundary layer were that the temperature is known at the two limits (1) 0 = 0 at = 0 and (2) 0 = 1 at —+ large. What is the total heat transferred to the wing if the temperature of the wing is 45°F? 4. At what angle should the walls diverge to maintain a constant velocity between x = 1. What is the Prandtl number of perfect air at this temperature? 4. Perform the estimation again using the approximate method of equation (4.86) to find the total drag coefficient for the flat plate shown in Fig. Derive equation (4. A wind tunnel has a 1rn2.15. find the total skin friction coefficient and the total drag on the wing (both sides) using the method of equation (4.1. 0' = 4.11 is at = 80 rn/s (all other conditions are the same).16.20..97). Convert the resulting relation into a drag relation for the flat plate and simplify the results.19..17.87) which takes into account transition effects. 4.Problems 213 the spacing between the streamlines in the undisturbed flow (2h) to their spacing (2w) at the x location where the Pitot profile was obtained.18. and thus maintain a constant area for the inviscid flow.14.. Represent the wing of an airplane by a flat plate. Now find the skin friction drag for the plate using the approximate formula of equation (4. For the wing dimensions and flow properties of Problem 4. Find the total skin friction drag of the flat plate (both sides) using the PrandtlSchlichting relation (initially assume that the flow is completely turbulent along the length of the plate). I x )r Streamlines Figure P4. The airplane is flying at standard sealevel conditions at 180 mi/h. 6rnlong test section in which air at standard sealevel conditions moves at 70 rn/s. The dimensions of the wing are chord = 5 ft and span = 30 ft. and (2) 0 = 1 at ij large? Hint: From equation (4.13 4. Assume the flow over the flat plate of Problem 4. How accurate was the initial assumption of fully turbulent flow? 4.e. what is the section U.87). Comment on the difference between the two approaches and the accuracy of the approximate method. 4.95).5 m and x = 6 m? .18.. What would be the temperature distribution if the boundary conditions were (1) an adiabatic wall (i. 8' = 0 at rj = 0). If w = drag coefficient Cd? O.86) and the PrandtlSchlichting turbulent skin friction relation. Using equation (4. calculate the thermal conductivity of air at 2000 K.009c. 4. This allows the freestream velocity to remain constant. It is planned to let the walls diverge slightly (slant outward) to compensate for the growth in boundarylayer displacement thickness.
. Presented at AIAA Aerosp. Reno. Defining the aerothermodynamic environment. Vol I: Defining the Hypersonic Environment. CO White FM. 2000. Momentum Transfer in Boundary Layers. Connolly TF. 29th. Orlando:Academic Press DeJarnette FR. Sherby SS. Periaux 3. NewYork: McGrawHill CebeciT. Ed. Boundary Layer Theory.Analysis of Turbulent Boundary Layers.1979. New York: work on entrainment. Ed. Computation of Turbulent Boundary Layers—1968 AFOSRIFPStanford Conference. Alimaras SR. NV Spalart PR. Turbulence Modeling for CFD. CA: DCW Industries . Meet. Reno.. 920439. 30th. NV Spalart PR. Boston: Birkhauser Boston Schlichting H. New York: McGrawHill Wilcox DC. Smith AMO. Viscous Hypersonic Flow. New York: McGrawHill Head MR. New York: McGrawHill Smith BR. von Doenhoff AE.1. 1. 1962. 1989. Viscous Fluid Flow. Theory of Wing Sections. Application of turbulence modeling to the design of military aircraft. Ed. Sci. Sci. 34:35—42 Dommasch DO. Trends in turbulence treatments. 1969. AIAA Pap. 910513. 2005. 1949.. Bertin 33. Ratciffe RA. AIAA Pap. 20002306. Denver. 1998. New York: Dover CebeciT.Chap. 1974. LaCañada. Vol. AIAA . In Proceedings.AIAA Pap. Glowinski R.. 1992. Stanford: Stanford University Press Neumann RD. 4 / Viscous Boundary Layers REFERENCES Abbott IH. Ed. 1991. Pitman Dorrance WH. 1979. Presented at Fluids 2000 Conf. Presented at AIAA Aerosp. A oneequation turbulence model for aerodynamic flows. In Hypersonics. 1996. Meet. 1967.Bradshaw P. Airplane Aerodynamics. Matching inviscid/boundary layer flowfields. 7th Ed.
The pressures and the shear forces can be integrated over the surface Pressure force Shear force Shear force Shear force Pressure force Pressure force Figure 5. which act on the surface due to the motion of air around the vehicle.1 CHARACTERIZATION OF AERODYNAMIC FORCES AND MOMENTS 5.1 Normal (or pressure) and tangential (or shear) forces on an airfoil surface.1. therefore. Although viscosity is a fluid property and. The normal (pressure) forces and the tangential (shear) forces. 5.1 General Comments The motion of air around the vehicle produces pressure and velocity variations through the flow field. the viscous forces acting on the vehicle depend on the velocity gradients near the surface as well as the viscosity itself.1.5 CHARACTERISTIC PARAMETERS FOR AIRFOIL AND WING AERODYNAMICS 5. are shown in Fig. acts throughout the flow field. 215 .
the component of the thrust parallel to the freestream velocity vector is only slightly less than the thrust itself).2.. the lift (L). and a component perpendicular to the flight path. Thus. as shown in Fig. which is the force parallel to the vehicle axis (A). Summing the forces perpendicular to the flight path leads to the conclusion that the weight of the aircraft is balanced by the lift.. for these applications. the equilibrium condition requires that the thrust must equal the drag acting on the airplane.g. unaccelerated level flight in a horizontal plane. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics Weight Figure 5.g. the bodyoriented components are the axial force. on which they act in order to yield the resultant aerodynamic force (R).. its motion is determined by its weight. the thrust produced by the engine. This condition requires (1) that the sum of the forces along the flight path is zero and (2) that the sum of the forces perpendicular to the flight path is zero. Summing the forces along the flight path (parallel to the freestream velocity). which acts at the center of pressure (cp) of the vehicle. as shown in Fig. a tail surface) is needed to produce a nosedown (negative) pitching moment about the cg. the resultant force is divided into components taken relative to the velocity vector (i. when the sum of the moments about the cg is zero. and the normal force. As the airplane moves through the earth's atmosphere. which is the force perpendicular to the vehicle axis (N). the pitch plane). For convenience. 5.2 Nomenclature for aerodynamic forces in the pitch plane.The aircraft is said to be trimmed. the drag (D).216 Chap.. which could balance the . Let us first consider the forces and the moments in the plane of symmetry (i. and the aerodynamic forces acting on the vehicle. it will produce a noseup (positive) pitching moment about the center of gravity (cg). a force from a control surface located aft of the cg (e. the total force vector is usually resolved into components.2.g. 5.Thus. Consider the case of steady. Because the lift force generated by the wing/body configuration acts ahead of the center of gravity. Consider the case where the lift generated by the wing/body configuration acts ahead of the center of gravity.e. For applications such as trajectory analysis. the aerodynamics or the structural dynamics). Let us consider only cases where the angles are small (e. 5.3.e. the resultant force is divided into a component parallel to the flight path.. For the pitchplane forces depicted in Fig. Bodyoriented force components are used when the application is primarily concerned with the vehicle response (e. the flight path).
moment produced by the wing/body lift. the trim drag may vary from 0. referred to the airplane's reference axes. Typically..e. In addition to the force components which act in the pitch plane (i.3. The three moment components are the pitching moment.4 Reference axes of the airplane and the corresponding aerodynamic moments. 5. the pilot's right). which is known as the trim drag.4. 5.1 / Characterization of Aerodynamic Forces and Moments 217 w Figure 5. the rolling moment.The tailgenerated lift force is indicated in Fig. the center of gravity). and the yawing moment. which acts upward perpendicular to the undisturbed freestream velocity. as shown in Fig. but does not include the tail profile drag (which is included in the total drag of the aircraft at zero lift conditions). The side force is the component of force in a direction perpendicular both to the lift and to the drag. The orientation of the tail surface which produces the lift force depicted in Fig. the lift.3 Moment balance to trim an aircraft. . 5. P4 Center x Longitudinal axis Positive rolling moment L Positive yawing moment N Vertical axis Figure 5.Sec. the resulting aerodynamic force usually will not act through the origin of the airplane's axis system (i. As noted earlier..5 to 5% of the total cruise drag for the airplane. and the drag. which acts in the same direction as the freestream velocity) there is a side force. 5. The side force is positive when acting toward the starboard wing (i.The reader should note that the trim drag is associated with the lift generated to trim the vehicle.. Positive pitching moment Lateral axis Y.e.e.3 also pro: duces a drag force. The moment due to the resultant force acting at a distance from the origin may be divided into three components.
where scale models are exposed to flow conditions that simulate the design environment or data from flight tests at other flow conditions. In practice. The pitching moment is the result of the lift and the drag forces acting on the vehicle. 5. such as boundarylayer separation. the results would be independent of all but the first two parameters listed. flow phenomena. Yawing moment. Angle of attack (i.218 1. once in dimensionless form. A positive yawing moment tends to rotate the nose to the pilot's right. the measurements are usually presented in dimensionless form.. Freestream velocity 5. The moment about the lateral axis (the y axis of the airplanefixed coordinate system) is the pitching moment. Mach number (as it relates to compressibility effects) The calculation of the aerodynamic forces and moments acting on a vehicle often requires that the engineer be able to relate data obtained at other flow conditions to the conditions of interest. configuration geometry and angle of attack. the ambient density is relatively low). Personal observations of the aerodynamic forces acting on an arm extended from a car window or on a ball in flight demonstrate the effect of velocity and of configuration.. the engineer often uses data from the wind tunnel. Vehicle size or model scale 4. parameters such as the Reynolds number and the Mach number appear in the correlations for the force coefficient and for the moment coefficients. Rolling moment. A rolling moment is often created by a differential lift. Density of the undisturbed air 6. The moment about the longitudinal axis of the airplane (the x axis) is the rolling moment. Chap. or starboard. wingtip to move downward.2 Parameters That Govern Aerodynamic Forces The magnitude of the forces and of the moments that act on a vehicle depend on the combined effects of many different variables. Pilot manuals advise that a longer length of runway is required if the ambient temperature is relatively high or if the airport elevation is high (i.1. limit the range of flow conditions over which the dimensionless force and moment coefficients remain constant. and compressibility effects. A positive rolling moment causes the right. The parameters that govern the magnitude of aerodynamic forces and moments include the following: 1. A positive pitching moment is in the noseup direction. shockwave/boundarylayer interactions. Reynolds number (as it relates to viscous effects) 7. For such cases. So that one can correlate the data for various freestream conditions and configuration scales. .e. 2. Ideally. The moment about the vertical axis of the airplane (the z axis) is the yawing moment. Thus. generated by some type of ailerons or spoilers. 3. 5 I Characteristic Parameters for Airfoil and Wing Aerodynamics Pitching moment.e. Configuration geometry 2. vehicle attitude in the pitch plane relative to the flight direction) 3.
The NACA investigations were further systematized by separation of the effects of camber and thickness distribution.1 AirfoilSection Nomenclature gradual development of wing theory tended to isolate the wingsection problems from the effects of planform and led to a more systematic experimental approach. will be discussed after a brief introduction to airfoilsection nomenclature. many families of wing sections Quoting from Abbott and von Doenhoff (1949). (3) the maximum thickness and the thickness distribution of the profile. were tested in the laboratories of various countries. most wing sections in common use were derived from more or less direct extensions of the work at Gottingen. the resultant section is called the airfoil section. (2) the mean camber line. The tests made at Gottingen during World War I contributed much to the development of modern types of wing sections. and (4) the trailingedge angle. .5. NACA fivedigit wing sections. 5. and the experimental work was performed at higher Reynolds number than were generally obtained elsewhere.Sec. which are illustrated in Fig. consider the NACA fourdigit wing sections. Geometric parameters that have an important effect on the aerodynamic characteristics of an airfoil section include (1) the leadingedge radius." As a result. Up to about World War II. The effect of these parameters. 5. 5. During this period. There are a variety of classifications. the geometry of many airfoil sections is uniquely defined by the NACA designation for the airfoil. and NACA 6 series wing sections.2 AIRFOIL GEOMETRY PARAMETERS If a horizontal wing is cut by a vertical plane parallel to the centerline of the vehicle. The first integer indicates the z xlocation of maximum thickness Maximum thickness Maximum camber x Chord line (Leading edge) x0 x=c (Trailing edge) Figure 53 Airfoilsection geometry and its nomenclature.2 / Airfoil Geometry Parameters 219 5.2. including NACA fourdigit wing sections. but the work of the NACA was outstanding. As an example.The generated lift and the stall characteristics of the wing depend strongly on the geometry of the airfoil sections that make up the wing.
For many airplanes the chord lines of the airfoil sections are inclined relative to the vehicle axis. 5 / Characteristic Parameters for Airfoit and Wing Aerodynamics maximum value of the mean camberline ordinate (see Fig.220 Chap. sharp leading edge is "0").5) in percent of the chord. one often encounters airfoil sections being developed that are not described by the Standard NACA geometries. The leading edge of airfoils used in subsonic applications is rounded. The leadingedge radius of the airfoil section is the radius of a circle centered on a line tangent to the leadingedge camber connecting tangency points of the upper and the lower surfaces with the leading edge.2. The last two integers indicate the maximum section thickness in percent of the chord. as mea sured perpendicular to the chord line.2. the NACA 0010 is a symmetric airfoil section whose maximum thickness is 10% of the chord. The NACA 4412 airfoil section is a 12% thick airfoil which has a 4% maximum camber located at 40% of the chord. because of the rapid improvements. 5.2 LeadingEdge Radius and Chord Line The chord line is defined as the straight line connecting the leading and trailing edges. A series of "standard" modifications are designated by a suffix consisting of a dash followed by two digits. 5. These modifications consist essentially of (1) changes of the leadingedge radius from the normal value and (2) changes of the position of maximum thickness from the normal position (which is at 0. "freestream" flow. cambered airfoils in a subsonic flow generate lift even when .3c). The second integer of the modification indicates the location of the maximum thickness in tenths of chord. The second integer indicates the distance from the leading edge to the maximum camber in tenths of the chord.3 Mean Camber Line The locus of the points midway between the upper surface and the lower surface.As will be seen in the theoretical solutions and in the experimental data that will be presented in this book. Thus. NACA The first integer indicates the relative magnitude of the leadingedge radius (normal leadingedge radius is "6".The center of the leadingedge radius is located such that the cambered section projects slightly forward of the leadingedge point. with a radius that is on the order of 1% of the chord length. However. both in computer hardware and computer software. 5. The shape of the mean camber line is very important in determining the aerodynamic characteristics of an airfoil section. and because of the broad use of sophisticated numerical codes. The geometric angle of attack is the angle between the chord line and the direction of the undisturbed.The magnitude of the leadingedge radius has a significant effect on the stall (or boundarylayer separation) characteristics of the airfoil section. defines the mean camber line.Thus.
Thus.7 1. zero lift results for sections with positive camber when they are at negative angles of attack. The thickness distribution for an airfoil affects. boundarylayer separation occurs early. As a result.2 / Airfoil Geometry Parameters 221 the section angle of attack is zero.2. an effect of camber is a change in the zerolift angle of attack. Thus. Consider the maximum section lift coefficients for several different thicknessratio airfoils presented in this table. the maximum section lift coefficient for a very thin airfoil section is relatively small. Thus. the pressure distribution and the character of the boundary layer.3 For a very thin airfoil section (which has a relatively small leadingedge radius).48 1. The maximum local velocity to which a fluid particle accelerates as it flows around an airfoil section increases as the maximum thickness increases (see the discussion associated with Fig. 4. The separation phenomena described in the previous paragraph causes the maximum sectionlift coefficients for the relatively thick airfoil sections (i.. those with a thickness ratio of 18% of the chord and of 24% of the chord) to be less than those for medium thickness airfoil sections. It should be noted. Furthermore. If the maximum lift coefficient is high. however. Thus. camber has a beneficial effect on the maximum value of the section lift coefficient. the boundary layer becomes thicker (and is more likely to separate producing relatively large values for the form drag). As a result. the stall speed will be low.e. As the adverse pressure gradient becomes larger. The maximum section lift coefficient increases as the thickness ratio increases from 8% of the chord to 12% of the chord.63 1. one needs to consider the tradeoffs in selecting a design value for a particular parameter. 9) and high twisting moments at high speeds. the adverse pressure gradient associated with the deceleration of the flow from the location of this pressure minimum to the trailing edge is greatest for the thickest airfoil.5 1. 5. 5. the . Airfoil Section NACA 2408 NACA241O NACA 2412 NACA 2415 NACA 2418 NACA 2424 Cimax 1. all other factors being the same. While the symmetric sections have zero lift at zero angle of attack.Sec. aw. the beneficial effects of increasing the maximum thickness are limited. the minimum pressure value is smallest for the thickest airfoil.4 Maximum Thickness and Thickness Distribution The maximum thickness and the thickness distribution strongly influence the aerodynamic characteristics of the airfoil section as well. As the location of the maximum thickness moves aft. that the high thickness and camber necessary for high maximum values for the section lift coefficient produce low critical Mach numbers (see Chapter.65 1. The values are taken from the figures presented in Abbott and von Doenhoff (1949). not far from the leading edge of the upper (leeward) surface.10).
6. In order to fully describe the planform of a wing. Laminar boundary layers produce less skin friction drag than turbulent boundary layers but are also more likely to separate under the influence of an adverse pressure gradient. In addition. 5. attention must be directed to the existence of flow components in the spanwise direction. 5.6 Geometric characteristics of the wing planform.3 WINGGEOMETRY PARAMETERS By placing the airfoil sections discussed in the preceding section in spanwise combinations. The terms that are pertinent to defining the aerodynamic characteristics of a wing are illustrated in Fig. 5 / Characteristic Parameters for AfrfoU and Wing Aerodynamics velocity gradient (and hence the pressure gradient) in the midchord region decreases.The aerodynamic center of thin airfoil sections in a subsonic stream is theoretically located at the quarterchord. canards. several terms are required. When the parameters that characterize the wing planform are introduced. horizontal tails. In other words. wings. This will be discussed in more detail later in this chapter. .5 TrailingEdge Angle The trailingedge angle affects the location of the aerodynamic center (which is defined later in this chapter). airfoil section properties deal with flow in two dimensions while planform properties relate to the resultant flow in three dimensions. 5.2. The resultant favorable pressure gradient in the midchord region promotes boundarylayer stability and increases the possibility that the boundary layer remains laminar. vertical tails. Li b lane Cr line Rectangular wing Unswept trapezoidal wing line Swept wing Delta wing Figure 5.222 Chap. the thicker airfoils benefit more from the use of high lift devices but have a lower critical Mach number. and/or other lifting surfaces are formed.
is simply the plan surface area of the wing. is determined from the equation that the product of the span and the average chord is the wing area (b X = S). is measured at the tip.3 I WingGeometry Parameters 223 1. the aspect ratio is simply C For a nonrectangular wing. Sweep angles of the leading edge or of the trailing edge are often presented with the parameters. If the wing lies below the horizontal plane. 5.5b S JO.0. b. and the wing area. Typical aspect ratios vary from 35 for a highperformance sailplane to 2 for a supersonic jet fighter. A rectangular wing has a taper ratio of 1. 6. AR. For a rectangular wing. The sweep of a wing causes definite changes in the maximum lift. The wing span. and in the effects of compressibility. is the chord at the wing centerline. 5. Considering the wing planform to have straight lines for the leading and trailing edges. the dynamic pressure. The mean aerodynamic chord is given by mac = — I 1 ç+O.0 while the pointed tip delta wing has a taper ratio of 0. is measured tip to tip. in the stall characteristics. 7. 4. since they are of interest for many applications. S. The aspect ratio. it is termed an anhedral angle. The sweep angle A is usually measured as the angle between the line of 25% chord and a perpendicular to the root chord. Thus. b2 The aspect ratio is a fineness ratio of the wing and is useful in determining the aerodynamic characteristics and structural weight. 2.5b [c(y)]2dy 9. The dihedral angle is the angle between a horizontal plane containing the root ch and a plane midway between the upper and lower surfaces of the wing. . Although a portion of the area may be covered by fuselage or nacelles. is the ratio of the span and the average chord.Sec. 8. The wing area. the taper ratio. when multiplied by the product of the average section moment coefficient. The mean aerodynamic chord (mac) is used together with S to nondimensionalize the pitching moments. the pressure carryover on these surfaces allows legitimate consideration of the entire plan area. the mean aerodynamic chord represents an average chord which. gives the moment for the entire wing. The average chord. is the ratio of the tip chord to the root chord: Ct Cr The taper ratio affects the lift distribution and the structural weight of the wing. A. 3. The dihedral angle affects the lateral stability characteristics of the airplane. and the tip chord. The root chord.
These same quantities also have a definite influence on the structural weight and stiffness of a wing. hence. Values of these parameters for a variety of aircraft have been taken from Jane s All the World's Aircraft (1973. tapered wing with geometric twist (wash out). stall) characteristics. and highspeed military aircraft. there is a spanwise variation in the geometric angle of incidence for the sections. Thus. In this case. where the incidence of the airfoil sections relative to the vehicle axis decrease toward the tip. In fact. as noted in Stuart (1978). and are summarized in Table 5. commercial jetliners and transports.7 and 3. the aspect ratio for the fourseat. the taper ratio. is parallel to the longitudinal axis of the vehicle. Note how the values of these parameters vary from one group of aircraft to another. The data presented in Table 5. "the selection of wing aspect ratio represents an interplay between a large value for low dragduetolift and a small value for reduced wing weight. If the angle of incidence increases toward the tip. the wing has "wash in. regardless of the national origin of the specific design. the boundarylayer separation (i..7 Unswept. The chord of the tip section. S / Characteristic Parameters for Airfoil and Wing Aerodynamics Chord of root section Vehicle longitudinal axis Chord of tip section Parallel to the vehicle longitudinal axis Figure 5.e. 10. however." The wings of numerous subsonic aircraft have wash out to control the spanwise lift distribution and." . and the sweep back of a planform are the principal factors that determine the aerodynamic characteristics of a wing and have an important bearing on its stall properties. For example. Data are presented for fourplace singleengine aircraft.8. the twist. whereas the aspect ratio for the supersonic military aircraft is between 1. The chord of the root section of the wing shown in the sketch of Fig. the aspect ratio. singleengine civil aviation designs is approximately 7." The airfoil section distribution. and 1984). which was a case study in aircraft design for the Northrop F5. Geometric twist defines the situation where the chord lines for the spanwise dis tribution of airfoil sections do not all lie in the same plane.1.1966.1 indicate that similar designs result for similar applications. 5.7 is inclined 4° relative to the vehicle axis. the wing has "wash out.224 Chap.
a modem fighter aircraft. Solution: Referring to the sketch of the delta wing in Fig. Notice the grouping by generic classes of aircraft for the correlation between the powertoweight ratio and the wing loading that is presented in Fig.001 Solar Challenger 0.'whereas the F15. a heavy bomber of World War II. weighed 29. Note that a fully loaded B17G Flying Fortess.75 ft). heavier aircraft is evident in this figure. the successful humanpowered aircraft fall in the lower left corner of Fig.0 Supersonic jet fighters — Subsonic jet fighters World War I fighters Powertoweight ratio (hp/lbf) 0. has a maximum takeoff weight of 30.8. It is in this same region that Lockheed's solar highaltitude powered platform (Solar HAPP) operates.Sec. The trend toward larger. with wing load ings (the ratio of takeoff gross weight to wing area) less than 1 lbf/ft2. and more complex with technological improvements. There is a tendency for airplanes to get larger.01 HAPPs Human powered aircraft 0.500 lb) with a wing span of 31.0 10 100 1000 Wing loading (lbflft2) Figure 5. 5. [From Hall (1985). 5. EXAMPLE 5.05 m (42.8.845 kg (68.1: Develop an expression for the aspect ratio Develop an expression for the aspect ratio of a delta wing in terms of the leading edge sweep angle (ALE).700 kg (65.000 ib) with a wing span of 13.81 ft).1  1. Howeyer.] There are other trends exhibited in parameters relating to aircraft performance. heavier.6.62 m (103.8 Historical ranges of power loading and wing loading. the wing area is bc and the tan ALE is given by tan ALE = (b/2) . 53 / WingGeometry Parameters 10 225 World War II fighters 1. 5.1 — Jet transports Turboprops Wright flyer General aviation aircraft Solar 0.
2° t/c = 0.5 at root. 2°30' 7. 4° Airfoil section Speed [km/h (mi/h)] 9.83) inc.90 (32.0) 10.).11 at tip. 4° at (185—200) 21 1—281 Beechcraft Sierra 7.0) tic = 0.42) 32.83 35° at c/4 24° at c/4 32° at c/4 25° at c/4 3° 6° McDonnellDouglas t/c = 0.) Special Boeing sections DC9 Boeing 727 Boeing 737 37. 4° 64—215. t/c = 0. FourPlace SingleEngine Aircraft None 7° 7.35 (93. 23012 at tip.5) 26.92 (35.55 (123.97 (36.30 V35 B (31.N TABLE 5.46) 23016.0) 10. inc. — 1° at tip 458—499 (285—310) 191—235 (119—146) Caravelle 210 8.129 (av.1 WingGeometry Parameters Sweep Wing span Type [m(ft)] a. Model 25 7.67 (35. 1°at tip None NACA 2412. 0.25 7. 20 NACA inc.2) 27. inc. —1°30' at tip (131) 298 (185) Piper Commanche 6.67 8.24 Piper Warrior II 10. 2° NACA 652—415. Commercial Jetliners and Transports 8.35 7. 2° at root. inc. 30 at root.98 (32. Inc.125 at root.37 6° 6° Aspect ratio. Cessna 172 (131—162) 211 9.5) tion (mod. 1°attip root.3 (112. 1°30' at root.0) inc.0) 28. inc.57 7.20 (33.5) 10. inc.52) 9.02 3° 20° at ç/4 NACA 2° (France) BAC 111 (UK) 20° at c/4 NACA cambered 34.00 b. 6.70 2° 70 2°30' forward None None 642A215.25 (89.92 (108.75) 10.97 (88.67 (35. 63A416. AR angle Dihedral 63A41 4(mod).13. inc. Bellanca.) 975 (605) 903 (561) 975 (605) 848 (527) .61 173—245 (108—i 52) None None 325 (202) 2 98—3 2 2 Socata Rallye (France) Ambrosini NF 15 (Italy) Beechcraft Bonanza.116 (av.32 None 6°30' 1°44' 5° 632A415.03 790 (490) 815 (507) Tupolev 154 (USSR) 8.09 to 0.
Mach 1.045—0.81 Thin.70 (38. and 1984).95 7° 59. 5°3' at root llyushin 1L76 (USSR) 50.8) Dassault Mirage III 3. t/c = 0. 958 (595) Lockheed C5A 67.0 35° at c/4 Mitsubishi 12 35°47' at c/4 40° on leading edges NACA 65 series (mod.05 Mach 1. t/c = 0.108 at tip.152 at root.89 (144.93 3.78 45° 35° 8.). 3°30' 815 (507) 667—895 (414—556) Airbus A310 (International) edge) 8.0 Douglas F4 11.080 (outboard).7) 11.134 (inboard).39 t/c Anhedral5° Anhedral5° Anhedral 9° 0.64 (195.82 (France) 8.078 (midspan).5 43.4) LTV F8 10. Mach 2.).23 Northrop F5E 2.0+ Source: Data from Jane's All the World's Aircraft (1973.0 SAAB35 Draken (Sweden) 9. 1966.14 (30.0466 NACA 64A204 General Dynamics F16 9. outer: leading edge 57° Leading edge 60°34' Anhedral 1° None tIc = 0.50 (165.80 (38.40 (30.6 (Japan) 7. inc. —1° Nearly Mach 2 1123 (698) LTVA7 2. inc.8 (mod.94 Central: leading edge 80°.4—2.0) Mach 2.) near midspan.8) 7. inc.75) 4.7) 37°30' at c/4 0.) Over Mach 2.7) Anhedral with wing mounted above the 750—850 (466—528) HighSpeed Military Aircraft 1.77 t/c = 0.22 (27.0) 1.Boeing 747 6.035 65A004.88 (222.67) 24° at c/4 tIc = 0.2 Mach 1.0) at the trailing 25° at c/4 fuselage c.88 (25. inc.051 (av. laminar flow section 65A007. 0.85) t/c = 0.048 Outer panel 12° McDonnell3. 20 t/c = 0. .0 (35.75 25° at c/4 28° at c/4 5°30' 11081 (inboard Anhedral NACA 0011 (mod.13 (26.
trapezoidal wing. the tip chord (ce) is 11. 5.2: Calculate the winggeometry parameters for the Space Shuttle Orbiter To calculate the winggeometry parameters for the Space Shuttle Orbiter. and (d) the mean aerodynamic chord (mac).228 Chap. and the span (b) is 78.9. we obtain S= —htan ALE b2 Substituting this expression for the wing area into the expression for the aspect ratio. . (c) the taper ratio (A). as shown in Fig.44 It.9 Sketch of Space Shuttle Orbiter geometry for Example 5. (b) the aspect ratio (AR).. Using these values which define the reference wing. calculate (a) the wing area (5). For the reference wing of the Orbiter.44 ft Wing glove 'Cf Figure 5. and substituting it into the expression for the wing area. the root chord (ci). Solution: (a) The area for the trapezoidal reference wing is = cr)b = 2690ft2 b= 78. 5 / Characteristic Parameters for and Wing Aerodynamics Solving this second expression for c. is 57. the complex shape of the actual wing is replaced by a swept.056 ft.2.056 ft 57. AR — S — tanALE EXAMPLE 5.48 ft.
we obtain Cr 0. 5.48 Cr b2 (78. it is clear from Fig..20 (d) To calculate the mean aerodynamic chord.48 — 57. to calculate the force in the z direction. trapezoidal wing is AR (c) The taper ratio is 11. which is illustrated in Fig.1 Lift Coefficient Let us develop the equation for the normal force coefficient to illustrate the physical significance of a dimensionless force coefficient. we obtain mac = 39.e.10.44 =0.1 that the lift (and similarly the normal force) results primarily from the action of the pressure forces.4.The required expression is 57. The shear forces act primarily in the chordwise direction (i.44 l.265 57. 5. Therefore. we need consider only the pressure contribution. .57 ft 5.e.028 c(y) 11.10 Pressure distribution for a lifting airfoil section.39. c(y)].44 1.44 mac f [c(y)]2dy 2690! (57.Sec.5b . We choose the normal z) component of the resultant force since it is relatively simple to calculate and it has the same relation to the pressure and the shear forces as does the lift..056)2 = = 2690 = 2.4 AERODYNAMIC FORCE AND MOMENT COEFFICIENTS 5.4 / Aerodynamic Force and Moment Coefficients 229 (b)The aspect ratio for this swept. The pressure force acting on a differential Figure 5.1776y)2dy Integrating this expression.1776y + 39.028 Substituting this expression for the chord as a function of y into the equation for the mean aerodynamic chord. 5. For a relatively thin airfoil section at a relatively low angle of attack. we will first need an expression for the chord as a function of the distance from the plane of symmetry [i. contribute primarily to the drag).
5. the element's length in the direction perpendicular to the plane of the cross section (or spanwise direction).3) and (5. the net force in the z direction is given by J pdxdy surface is zero. the force in the z direction due to a uniform pressure the entire wing area is zero.230 Chap. the resultant force component is = (5.3) Note that the resultant force in any direction due to a constant pressure over a closed acting over if Combining equations (5.6) Since the product eb represents the planform area S of the rectangular wing of Fig.11 Pressure acting on an elemental surface area. (5.4). Since the pressure force acts normal to the surface.11. as shown in Fig.5) To nondimensionalize the factors on the righthand side of equation (5. the wetted length of the element in the plane of the cross section. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics z x Figure 5.4) if (p jj q00 (5. divide by the product which has the units of force. area of the vehicle surface is dF = p ds dy. 5. times dy.1t The elemental surface area is the product of ds. = if . Thus. = \CJ \b (5.2) Integrating over the entire wing surface (including the upper and the lower surfaces). the force component in the z direction is the product of the pressure times the projected planform area: = p dx dy (5.5).
Thus. which is defined as CL = —f— (5. For this twodimensional flow. C1.8) The data from Abbott and von Doenhoff (1949) were obtained in a wind tunnel that could be operated at pressures up to 10 atm. which is the chord length (c). data are presented for a model that had "standard" surface roughness applied near the leading edge. The experimental section lift coefficient is independent of Reynolds number and is a linear function of the angle of attack from —10° to + 10°. Thus. the windtunnel model represented a wing of infinite span.6) depends only on the configuration geometry and on the angle of attack. In addition to the measurements obtained with a smooth model. dC1 = = 0. the resulting dimensionless force parameter. lift is generated at zero angle of attack. The slope of this linear portion of the curve is called the twodimensional liftcurve slope.12 for a NACA 23012 airfoil. the lift measurements are presented in terms of the section lift coefficient C1. we are interested in determining the lift acting on a unit width of the wing [i. As a result. Partly because . the normal force coefficient). max. the Reynolds number ranged from 3 X 106 to 9 x 106 at Mach numbers less than 0. Thus. As the angle of attack is increased above 10°.2°. The section lift coefficient is the lift per unit span (1) divided by the product of the dynamic pressure times the plan area per unit span.7) Data are presented in Fig. that is. C1 max is 1. is independent of model scale and of flow conditions. y/b) is independent of vehicle scale and of the flow conditions. zero lift is obtained at —1.12. C1 = —f (5. Additional comments will be made about surface roughness later in this chapter. the section lift coefficient continues to increase (but not linearly with angle of attack) until a maximum value. the pressure distribution around the airfoil is essentially that of an inviscid flow. A similar analysis can be used to calculate the lift coefficient. 5.104 per degree Note that equations to be developed in Chapter 6 will show that the theoretical value for the twodimensional liftcurve slope is per radian (0. Using the experimental data for this airfoil. the lift per unit span (1)]. Since the NACA 23012 airfoil section is cambered (the maximum camber is approximately 2% of the chord length). the pressure coefficient at. the pressure distribution is independent of Reynolds number and does not depend on whether the boundary layer is laminar or turbulent. Over the range of flow conditions for which the pressure coefficient is a unique function of the dimensionless coordinates (x/c. which is designated or the section angle of attack for zero lift.. is reached.1097 per degree). 5.79 and occurs at an angle of attack of 18°. or force coefficient (in this case. Referring to Fig. In fact. 5.Sec. The lift acting on a wing of infinite span does not vary in the y direction.e. y/b).a particular location on the surface given by the dimensionless coordinates (x/c. the value of the integral in equation (5. When the boundary layer is thin.4 / Aerodynamic Force and Moment Coefficients 231 When the boundary layer is thin.17.
the presence of a separation bubble near the leading edge of the airfoil results in laminar .g. the air particles which have been slowed by the viscous forces cannot overcome the relatively large adverse pressure gradient. Note that the adverse pressure gradient which the air particles encounter as they move toward the trailing edge of the upper surface increases as the angle of attack increases.6 Lift C.1. and the boundary layer separates. the Beechcraft Bonanza.. The study of airfoil lift as a function of incidence has shown that. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics o 3. and the Brewster Buffalo). At angles of attack in excess of 10°.0x106 8. / o 0. 0. deg Figure 5.8  coefficient \.0 C 0.232 Chap. boundarylayer separation has a profound effect on the drag acting on the airfoil.0  Moment o ci coefficient —0.6 liii —20 —10 I 0 10 20 Section angle of attack. and therefore on the Reynolds number. in many instances.0 X 106 (standard roughness) 2. the section lift coefficients exhibit a Reynolds number dependence.2 —0. The separation location depends on the character (laminar or turbulent) of the boundary layer and its thickness.1 1. At these angles of attack.8 DO * —1.0X106 6.12 Section lift coefficient and section moment coefficient (with respect to c14) for an NACA 23012 airfoil. see Table 5. [Data from Abbott and von Doenhoff (1949).8 X 106 (smooth model) 6.] of this relatively high value of C1 max' the NACA 23012 section has been used on many aircraft (e. As will be discussed.
and the lift does not drop abruptly after Cimax but decreases gradually. the experimental values of the section lift coefficient for an NACA 23012 airfoil section range from 0. The reason for this is that the Reynolds number is then large enough for a natural transition to a turbulent boundary layer to occur upstream of the strong pressure rise. We will assume that the lift is a linear function of the angle of attack and that it is independent of the Reynolds number (i. To calculate the conesponding lift force per unit span.e.3 chord length or more downstream.104(4.50 to 0. which develops near the leading edge of the upper surface of the airfoil. Experimental data on twodimensional "peaky" airfoil sections indicate that as limited by laminar stall.12.9) Using the values presented in the discussion associated with Fig..0 — (—1. The test section conditions simulate a density altitude of 3 km. EXAMPLE 5.8) to obtain /= . The separation bubble may not appear at all on relatively thick. it will separate because the relatively low kinetic energy level of the air particles 'near the wall is insufficient to surmount the "pressure hill" of the adverse pressure gradient. This is known as short bubble separation.3 m.57. The separated viscous layer near the pressure peak may not reattach to the surface at all.541 At an angle of attack of 4°. That these are reasonable assumptions can be seen by referring to Fig. the viscous effects are negligible) at these test conditions.4 / Aerodynamic Force and Moment Coefficients 233 section stall. we will assume that the flow is two dimensional.3: Calculate the lift per unit span on a NACA 23012 airfoil section Consider tests of an unswept wing that spans the wind tunnel and whose airfoil section is NACA 23012. strongly cambered profiles operating at high Reynolds numbers. the section lift coefficient is C1 = Gia(a aol) (5. Since the wing model spans the test section. This extended separation region is known as long bubble separation.2)) = 0. If the laminar boundary layer. What is the li. is subjected to a relatively high adverse pressure gradient. or it may reattach within 0. The velocity in the test section is 360 km/h. 5. 5.The separation point moves upstream continuously with increasing angle of attack. The relatively high kinetic energy of the air particles in a turbulent boundary layer permits them to climb the "pressure hill. The separated shear layer that is formed may curve back onto the surface within a very short distance. is strongly dependent on the leadingedge shape and on the Reynolds number.ft per unit span (in N/rn) that you would expect to measure when the angle of attack is 4°? What would be the corresponding section lift coefficient? Solution: First." and boundarylayer separation occurs only a short distance upstream of the trailing edge (trailingedge stall). 5. Thus.Sec. we rearrange equation (5. C1 = 0. The chord of the model is 1. we need to calculate the section lift coefficient.12.
the quarterchord location.Thus.234 Chap. we refer to Table 1. The reference axis may be the leading edge. The significance of these reference axes in relation to the coefficients for thin airfoils will be discussed in subsequent chapters.11). the lift per unit span is 1= m = m ](1. To demonstrate this nondimensionalization. . the aerodynamic center. 360 we need the velocity in rn/s and h 100 kml000m h m S km 3600s = 0.3 m) 1 = (0.) p= Thus.The contribution of the chordwise component of the pressure force and of the skin friction force to the pitching moment is small and is neglected. the net pitching moment is given by M0 = J px dx dy (5. the resultant pitching moment due to this constant pressure is zero. The procedure used to nondimensionalize the moments created by the aerodynamic forces is similar to that used to nondimensionalize the lift.541)[4546.2 to find that PSL (Note: The fact that we are given the density altitude as 3 km does not provide specific information either about the temperature or about the pressure.10) where dx dy is the projected area.11) When a uniform pressure acts on any closed surface. Integrating over the entire wing surface. the pitching moment about the leading edge due to the pressures acting on the surface will be calculated (refer again to Fig. and so on.4.3m) = 5. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics To calculate the dynamic pressure the density in kg/m3. Thus.5](1. 5. the pitching moment about the leading edge due to the pressure force acting on the surface element whose area is ds times dy and which is located at a distance x from the leading edge is dM0 = p dx dy x (5.74225 Given the density altitude is 3 km.2 Moment Coefficient The moment created by the aerodynamic forces acting on a wing (or airfoil) is deter mined about a particular reference axis.
12 and 5. However. divide which has the units of force times length: M0 — q00c2b — if If c '\b Since the product of cb represents the planform area of the wing 5. Since the moment about the aerodynamic center is the product of a force (the lift that acts at the center of pressure) and a lever arm (the distance from the aerodynamic center to the center of pressure). The section moment coefficient is used to represent the dimensionless moment per unit span (mo): m0 Cm0 (5. Section pitching moment coefficients for a NACA 23012 airfoil section with respect to the quarter chord and with respect to the aerodynamic center are presented in Figs. since the pressure coefficient depends only on the dimensionless space coordinates (x/c. Note that the pitching moment coefficient is independent of the Reynolds number (for those angles of attack where the lift coefficient is independent of the Reynolds number).16) since the surface area per unit span is the chord length c. the aerodynamic center is that point along the chord where all changes in lift effectively take place.Sec. as noted previously in this chapter. The section pitching moment coefficient depends on the camber and on the thickness ratio. the mean aerodynamic chord is used together with S to nondimensionalize the pitching moment for a general wing. The quarterchord location is significant. Thus. the center of pressure must move toward the aerodynamic center as the lift increases.12) Combining equations (5. 5.14)}.4 / Aerodynamic Force and Moment Coefficients 235 if M0 (5.13) by To nondimensionalize the factors on the righthand side of equation (5. y/b) {see equation (5.11) and (5.14) (5. the dimensionless moment coefficient is M0 CM0 if x/x\/y'\ (5.12). the resulting pitching moment about the leading edge is if (p — (5. = Thus. the chord c is used. One of the features of the NACA 23012 section is a relatively high Cj. 5. 5.13).15) Since the derivation of equation (5. since it is the theoretical aerodynamic center for incompressible flow about a twodimensional airfoil. The aerodynamic center is that point about which the section moment coefficient is independent of the angle of attack.13.11. respectively.max with only a small .15) was for the rectangular wing of Fig.
[Data from Abbott and von Doenhoff (1949). .236 Chap.241 0.3 N (5.016 0 A A U 0 0 A A A A 0.0 I  —0.13 Section drag coefficient and section moment coeffi cient (with respect to the ac) for a NACA 23012 airfoil.1 11111 0. If t denotes the tangential shear stress at a point on the body surface.18) Drag Coefficient The drag force on a wing is due in part to skin friction and in part to the integrated effect of pressure.4.Therefore. the rolling moment coefficient is = (5. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics ac position (xlc) o 3.0 x 106 Standard roughness III.] The characteristic length (or moment arm) for the rolling moment and for the yawing moment is the wing span b (instead of the chord).0 x 106 <> 8.6 Cl Figure 5.247 A 6. A 0.0 x 106 0.2 II 0.8 X 106 0.8 111111 1.008 0 —0.241 6.17) and the yawing moment coefficient is CN = 5.
on a nozzle boattail). can be approximated as rdx (5. This approach is known as the nearfield method of drag computation. Thus. even a small difference between the computed pressure and the actual pressure can have a significant effect on the accuracy of the predicted drag. is defined as 2 (5.19) where is a unit vector parallel to the free stream and the integration takes place over the entire wetted surface.e. 4. Dividing both sides of equation (5. 5.10. steady.1 and 5.20) by the product yields an expression for the dimensionless force coefficient: sf I j Cf=1 Cfd( —) \cJ (5.21) Cf. The most straightforward approach to calculating the pressure drag is to perform the numerical integration indicated by the second term in equation (5. skin friction is a major component of the chordwise force per unit span (fr). However.18). If the pressure difference is near the middle of the aerodynamic configuration. the pressure distribution is essentially that for an inviscid flow.Sec. the skin friction coefficient. this can be a relatively inaccurate procedure for streamlined configurations at small angles of attack. The first integral represents the friction component and the second integral represents the pressure drag. Referring to Figs. The inaccuracy results because the pressure drag integral is the difference between the integration on forwardfacing and rearwardfacing surface elements. if the pressure difference is near the aft end of the configuration (for instance. In Chapter 3 we learned that zero drag results for irrotational. Equations for calculating the . the drag can be formally expressed as D = if (5. the reader should realize that subtle differences between the computed pressure distribution and the actual pressure distribution can have a significant effect on the validity of the drag estimates. it will have a relatively small effect on the validity of the estimated drag. depending upon where the differences occur. 5.4 / Aerodynamic Force and Moment Coefficients 237 p the static pressure. where the local slope is roughly parallel to the freestream direction. a twodimensional geometry) which is at a relatively low angle of attack so that the boundary layer is thin and does not separate. For an airfoil section (i. Unfortunately. and the external normal to the element of surface dS..22) As was stated in the general discussion of the boundaiylayer characteristics in Chapter 4 (see Fig. Furthermore. this difference being a secondorder quantity for slender bodies. incompressible flow past a twodimensional body. skin friction for a turbulent boundary layer is much greater than that for a laminar boundary layer for given flow conditions.20) where is the chordwise force per unit span due to skin friction.19).
35a) shows that the velocity at the edge of the boundary layer and.074 (5. the local static pressure.26) Also.24) and (5.The total skinfriction coefficient for laminar flow is given by — Cf = 1. the local Reynolds number is defined as Re1 JLoo (5. = 0. Such is not the case for an airfoil section.664 (Re1)°'5 (5.25).29) These total skinfriction coefficients use the lengthbased Reynolds number given by Rer where L is the length of a flat plate. although the PrandtlSchlichting formulation is more accurate: C 0. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics skin friction coefficient were developed in Chapter 4.25) For equations (5.30) .28) which is the Prandtl formulation.21). = 0. the results for a flat plate only approximate those for an airfoil. The potential function given by equation (3. However. as was shown in Chapter 4. let us introduce the correlations for the skin friction coefficient for incompressible flow past a flat plate to gain insight into the force coefficient of equation (5.238 Chap. the analysis will provide useful insights into the section drag coefficient.0583 (5. Referring to Chapter 4.24) If the boundary layer is turbulent.23) for an airfoil at relatively low angles of attack. Of course. Nevertheless. then decelerates to the trailing edge. is constant along the plate.328 (5. total skinfriction coefficients can be deflned. therefore. when the boundary layer is laminar.455 258 (5. = pcxUooL (5. Cd = (5.27) and the total skinfriction coefficient for turbulent flow is = 0. for which the flow accelerates from a forward stagnation point to a maximum velocity.
Cf = T (Ret = 0.959 X TT2 \ — — 2 i—. As reported by Bowes (1974).6601 kg/rn3 = 1.5 m from the leading edge of a flatplate airfoil flying at 60 mIs at a height of 6 km. Cf T = 0.185 N/rn2 As discussed in the text of Chapter 2 and in the homework problems of Chapters 2 and 4. The windtunnel data used in this correlation were also from wake surveys on the 727 wing.26).242 X 106 10 5kg/s'm If the boundary layer is laminar. 5.73 was about 15 percent higher than predicted from windtunnel test data for a smooth airfoil.mjn has been attributed to surface roughness and excrescences on the airplane wing. (3) the drive unit. 5. (0. This quite sizeable difference between the measured and extrapolated values of Cd.522 x = 4. the integral form of the momentum equation can be used to determine the drag acting on an airfoil section. The data were adjusted to fully turbulent flow and extrapolated to flight Reynolds numbers.2 to obtain the static properties of undisturbed air at 6 km: = 0. (2) a rotating arm.Sec. Wingsection profiledrag measurements have been made for the Boeing 727 in flight using the Boeing Airborne Traversing Probe.5949 x = 1. using equation (5. (1 If the boundary layer is turbulent. although the 15percent increase in wingsection profile drag is larger than traditionally allotted in airplane drag estimates. examples of which are presented in Fig.0583 )02 = 3. This is not representative of the entire wing surface. The wing section where this survey was performed was inspected and had numerous steps and bumps due to control devices and manufacturing tolerances which would account for this local level of excrescence drag.664 = 5. The probe consists of four main components: (1) flow sensors.5949 X 105kg/s'm Thus. Solution: Refer to Table 1. "The measured minimum section profile drag at M = .6601 kg/m3)(60 m/s)(0.4 / Aerodynamic Force and Moment Coefficients 239 EXAMPLE 54: Calculate the local skin friction Calculate the local skin friction at a point 0.5 m) 1. and (4) the mounting base. This approach is known as the farfield method of drag determination." .
the boundary layer is laminar. we must know at what point. Disturbances to the flow in the growing viscous layer may be caused by surface roughness.] 5. At higher Reynolds numbers..e.000. if the surface of a flat plate is smooth and if the external airstream has no turbulence. of the surface temperature.. pressure pulses. As the flow proceeds downstream. a temperature variation in the surface.240 Chap. the disturbances will be damped by viscosity and the boundary layer will remain laminar. but it is also used as an indicator of whether the boundary layer is laminar or turbulent. The details of the transition process are quite complex and depend on many parameters. experience has shown that the Reynolds number at which the disturbances will grow and the length over which the transition process takes place depends on the magnitude of the disturbances and on the flow field. turbulent (i. However. The engineer who must develop a transition criterion for design purposes usually uses the Reynolds number. if the Reynolds number is low.e. eventually.14 Airborne traversing probe concept and configurations. Near the forward stagnation point on an airfoil. the boundary layer thickens and the viscous forces continue to dissipate the energy of the airstream. where transition occurs). transition "occurs" at a Reynolds number (Rex) of approximately 500. The Reynolds number not oniy affects the magnitude of Cf.21) depends on the Reynolds number. the boundary layer becomes turbulent (i. and of the local Mach number on transition. Let us consider briefly the effect of surface roughness. the boundary layer may become unstable and. For instance. in such cases. or on a wing. the disturbances may grow.4. Since the calculation of the force coefficient requires integration of Cf over the chord length. 5 / Characteristic Parameters for Airfoi' and Wing Aerodynamics Wake Survey lAreaMomentum Surveyj Mounting 4: Flow sensor Drive unit Axis Arm area Arm base 'Drive unit \ Figure 5. if any. and so on. or near the leading edge of a flat plate. of a pressure gradient in the inviscid flow. . transition will occur). [From B owes (1974).4 BoundaryLayer Transition it is obvious that the force coefficient of equation (5.
Mach number. the inviscid flow is accelerating) delays transition. For larger angles of attack. The transition Reynolds number is usually higher when the flow is compressible (i.. Schlichting (1968).e.. At the very high Reynolds numbers that occur at some flight conditions. Consider the NACA 0009 section and the NACA 66—009 section. For a more detailed discussion of transition.21) would indicate that the drag is lower for the NACA 66—009. already slowed by viscous action. the presence of surface roughness significantly promotes transition (i. When the air particles in the boundary layer. Comparing the cross sections presented in Fig. Since the drag for a streamlined airfoil at low angles of attack is primarily due to skin friction.. The maximum thickness for the NACA 66—009 section is at 0. an adverse pressure gradient promotes transition. a laminarflow section. Note that the section drag curve varies only slightly with C1 for moderate excursions in angle of attack. Mack (1984)..4 / Aerodynamic Force and Moment Coefficients 241 1. 3.15. the static pressure decreases in the streamwise direction or. As the angle of attack and the section lift coefficient increase.] can be an important tool that provides insights into the importance of individual parameters without introducing spurious effects that might be due to flow disturbances peculiar to a test facility (e. Surface roughness. encounter the relatively . Since transition is the amplification of disturbances to the flow. Cooling the surface usually delays transition. As a result. it is difficult to maintain a long run of laminar boundary layer. the section drag coefficient depends both on Reynolds number and on angle of attack. since the skin friction coefficient varies little with angle of attack.] If the skin friction is the dominant component of the drag. 5. the reader is referred to a textbook on boundarylayer theory [e.g. 5. such as the NACA 66—009. the minimum pressure coefficient decreases..3c (see Fig.. The boundarylayer thickness decreases as the surface temperature is decreased. there is a complex relationship between boundarylayer transition and surface cooling. the point of maximum thickness could be moved aft so that the boundary layer is subjected to a favorable pressure gradient over a longer run. the minimum pressure coefficient occurs at x = 0. use of equation (5. Conversely.g. A favorable pressure gradient (i. causes transition to occur at relatively low Reynolds numbers). offers additional benefits.15). "noise" in a wind tunnel). The adverse pressure gradient that results as the flow decelerates toward the trailing edge increases. having a maximum thickness of 0. 4.The subsequent reduction in the friction drag creates a drag bucket for the NACA 66—009 section. To delay transition on a lowspeed airfoil section. the cross section of the NACA 66—009 airfoil provides more flexibility for carrying fuel and for accommodating the loadcarrying structure.The lower local velocities near the leading edge and the extended region of favorable pressure gradient cause transition to be farther aft on the NACA 66—009. while that for the NACA 0009 section is at 0. 5. However.16. transition should be delayed as long as possible to obtain a lowdrag section. 5. 2.e. the minimum pressure occurs near x = 0. Surface temperature. Stability theory [e.45c.09c. For the NACA 0009. for supersonic flows.Sec. especially if surface roughnesses develop during the flight operations.e. How ever. This is verified by the data from Abbott and von Doenhoff (1949) which are reproduced in Fig. Both are symmetric. Pressure gradient. equivalently.6c for the NACA 66—009 and a favorable pressure gradient acts to stabilize the boundary layer to this point.g.lc. as the Mach number is increased).
which are taken from Abbott and von Doenhoff (1949).242 Chap. which has been discussed. where extensive regions of separated flow exist.4 0. See Fig. Adverse Adverse —i pressure gradient for 0009 gradient for 66009 —2 I I 0. the drag coefficient for the NACA 66—009 is greater than that for the NACA 0009. include data for a standard roughness. . the boundary layer thickens and separates. strong adverse pressure gradient. the drag coefficient exhibits a Reynolds number dependence both at the low angles of attack.2 0.6 0. The data illustrate the dependence on Reynolds number and on angle of attack. 5 / Characteristic Parameters for Airfol! and Wing Aerodynamics NACA section (a) 11 I o Cp —*.0 0. 5. The section drag coefficient for a NACA 23012 airfoil is presented as a function of the section lift coefficient in Fig. 5.8 1. Note that measurements. and at high angles of attack.13. at the higher angles of attack (where form drag is important). Note that. However. (b) static pressure distribution. where the boundary layer is thin (and the transition location is important). when the viscous effects are secondary. Because the thickening boundary layer and its separation from the surface cause the pressure distribution to be significantly different from the inviscid model at the higher angles of attack. form drag dominates.16.15 Comparison of geometries and resultant pressure distributions for a "standard" airfoil section (NACA 0009) and for a laminar airfoil section (NACA 66—009): (a) comparison of cross section for an NACA 0009 airfoil with that for an NACA 66—009 airfoil.0 x C (b) Figure 5. Thus. we see that the lift coefficient and the moment coefficient depend only on vehicle geometry and angle Of attack for lowspeed flows.
31) where the subscripts wt and ft designate windtunnel and flight conditions. "The standard leadingedge roughness selected by the NACA for 24in chord models consisted of 0. In many lowspeed wind tunnels.012 Cd ci 0 — a a 0. If one desires to reproduce the Reynolds number for a flight condition in the wind tunnel.This . (5. Abbott and von Doenhoff (1949) present data on the effect of surface condition on the aerodynamic forces. Thus.004 a ODD  0.4. the freestream values for density and for viscosity are roughly equal to the atmospheric values. [Data from Abbott and von Doenhoff (1949).Sec. which obviously would not be a reasonable simulation. The grains were thinly spread to coverS to 10% of the área. the windtunnel value for the freestream velocity would have to be 5 times the flight value.020 0. since the maximum Reynolds number for this "equal density" subsonic windtunnel simulation is much less than the flight value.08c measured from the leading edge on both surfaces. then \ /wt lit (5.000 —1.32) If the windtunnel model is 0.016 0. the tunnel flow would be transonic or supersonic. As a result. 54 / Aerodynamic Force and Moment Coefficients ONACA 0009 O 243 NACA 66009 0.0080 00 0. = 6 X 106.5 Effect of Surface Roughness on the Aerodynamic Forces As discussed in Chapter 2.6 Figure 5.8 0.] 5.2 scale.011in carborundum grains applied to the surface of the model at the leading edge over a surface length of 0. controlled surface roughness is often added to the model to fix boundarylayer transition at the location at which it would occur naturally in flight. respectively.0 Cl 0. the Reynolds number is an important parameter when comparing the viscous character of two fields.6 —0. Thus.8 1.16 Section drag coefficients for NACA 0009 airfoil and the NACA 66—009 airfoil.
This is further illustrated by the data presented in Fig. If the wing surface is sufficiently rough to cause transition near the wing leading edge. 5.002 0. When there is no appreciable separation of the flow. or damage in military combat. The minimum drag increased progressively with forward movement of the roughness strip.011 I roughness roughness roughness I I o O 16 — 1. 5.2  0 0.17.17 Effect of roughness near the leading edge on the maximum section lift for the NACA.8  I I I 0 8 16 24 32 Figure 5.244 Chap. However. large increases in drag are observed. [Data from Abbott and von Doenhoff (1949). the drag on the airfoil is caused primarily by skin friction. mud.004 0. 63(420)422 airfoil. In other test results presented in Abbott and von Doenhoff (1949). but it is considerably less severe than that likely to be encountered in service as a result of accumulation of ice. Smooth o 0. Thus. the location of the roughness strip was systematically varied.] . 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics standard roughness is considerably more severe than that caused by the usual manu facturing irregularities or deterioration in service.12) indicate that the angle of zero lift and the liftcurve slope are practically unaffected by the standard leadingedge roughness. 5.13 for the NACA 23012 airfoil section. the value of the drag coefficient depends on the relative extent of the laminar boundary layer." The data for the NACA 23012 airfoil (Fig. the maximum lift coefficient is affected by surface roughness. as is evident in the data of Fig. A sharp increase in the drag coefficient results when transition is suddenly shifted forward.
will produce a thicker boundary layer at the onset of the pressure recovery as compared to the conditions produced by a turbulent boundary layer which originates further downstream. et al. ice crystals. the required roughness often produces undesirable distortions of the flow. (1966) provides some general guidelines for the use of boundarylayer trips for hypersonic flows. the prime requirements for correct simulation of these transonic shock/boundarylayer interactions include that the boundary layer is turbulent at the point of interaction and that the thickness of the turbulent boundary layer for the model flow is not so large in relation to the fullscale flow that a rear separation would occur in the simulation that would not occur in the fullscale flow. As noted by van Dam and Holmes (1986). the roughness configurations that are required to fix transition in a supersonic flow often cause grit drag. loss of laminar flow can be caused by surface contamination such as insect debris.Sec.4 / Aerodynamic Force and Moment Coeffidents 245 Scaling effects between model simulations and flight applications (as they relate to the viscous parameters) are especially important when the flow field includes an interaction between a shock wave and the boundary layer. A turbulent boundary layer which originates near the leading edge of an airfoil. substantially ahead of the point of minimum pressure. Sterret. surface roughness produced by environmental "contamination" may have a significant (and unexpected) effect on the transition location. As discussed. (2) When surface roughness is used. (1) The reader should note that. The transonic flow field for an airfoil may include a shockinduced separation. et al. the use of boundarylayer trips is intended to compensate for the inability to simulate the Reynolds number in groundtest facilities.. forward movement of transition location and turbulent separation produce a large increase in section drag. Also. moisture due to mist or rain.. (1968). Since roughness heights several times larger than the boundarylayer thickness can be required to fix transition in a hypersonic flow. and another boundarylayer separation near the trailing edge. et a!... considerable care should be taken to properly size and locate the roughness elements in order to properly simulate the desired flow. Data are presented to illustrate problems that can arise from poor trip designs. Fixing transition on windtunnel models becomes increasingly difficult as the Mach number is increased. a change in the turbulent boundary layer conditions . "The surface roughness caused by such contamination can lead to early transition near the leading edge.. Braslow. The previous discussion in this section has focused on the effects of roughness elements that have been intentionally placed on the surface to fix artificially the location of transition. and "innocent" modifications such as the addition of a spanwise paint stripe. 5.. Whereas it is possible to fix boundarylayer transition far forward on windtunnel models at subsonic speeds using grittype transition trips ing little or no grit drag. As noted by van Dam and Holmes (1986). . (1966) provide some general guidelines for the use of grittype boundarylayer transition trips. According to Pearcey. since it is impossible to match the Reynolds number in many scalemodel simulations. The comments regarding the effects of surface roughness were presented for two reasons. With sufficiently steep pressure gradients in the recovery region. surface roughness (in the form of boundarylayer trips) is often used to fix transition and therefore the relative extent of the laminar boundary layer. can lead to premature turbulent separation . a subsequent reattachment to the airfoil surface. However. thus affecting the aerodynamic characteristics and the effectiveness of trailingedge control surfaces. surface damage." .
Accurately predicting parasite drag would seem an a]most impossible task. total aircraft parasite drag is a complex combination of aircraft configuration.g.6 Method for Predicting Aircraft Parasite Drag As you can imagine. fuselage. all of which add confusion about drag and predicting drag.7. caused • • • • • • Interference drag—due to the proximity of two (or more) bodies (e. convert each corrected skinfriction coefficient into an aircraft drag coefficient for that component 5. numerous straightforward estimation methods exist for predicting the parasite drag of aircraft. however. Sum all aircraft parasitedrag coefficients to obtain a total aircraft drag coefficient . wing and fuselage) Trim drag—due to aerodynamic forces required to trim the airplane about the center of gravity Profile drag—the sum of skinfriction and pressure drag for an airfoil section Parasite drag—the sum of skinfriction and pressure drag for an aircraft Base drag—the pressure drag due to a blunt base or afterbody Wave drag—due to shockwave energy losses In spite of these complexities. The basic approaches. skin friction. correct the skinfriction coefficient for surface roughness 3. and semiempirical approaches. especially compared with predicting lift. While the building block methods for predicting skinfriction drag have been presented in Section 4. there are a variety of methods for applying skin friction prediction methods to determine total aircraft drag. 5 I Characteristic Parameters for Airfoil and Wing Aerodynamics 5.) 2.1 and 5. Every aerodynamics group at each aircraft manufacturer has different methods for estimating subsonic aircraft drag. Estimate an equivalent flatplate skinfriction coefficient for each component of the aircraft (wing.To make matters even more difficult. stabilizers. are probably quite similar: 1.3. empirical.. there are a variety of terms associated with drag.4. interference among aircraft components.246 Chap.4. pressure distribution. etc. most of which use a combination of theoretical. and flight conditions (among other things). Some of the common terms used to describe drag are [Mccormick (1979)11: • Induced (or vortex) drag—drag due to the trailing vortex system Skin friction drag—due to viscous stress acting on the surface of the body by flow separation • • Form (or pressure) drag—due to the integrated pressure acting on the body. Apply a form factor correction to each component's skinfriction coefficient to take into account supervelocities (velocities greater than free stream around the component) as well as pressure drag due to flow separation to obtain a parasitedrag coefficient 4.
is the wetted area of each component. a tip chord a leadingedge sweep A. afld Srei is the aircraft reference area (there is only one reference area for the entire aircraft. etc. including rivets. . etc. 5. a different total skinfriction coefficient. fuseis the lage. The zerolift drag coefficient for subsonic flow is obtained by: N KiC1.18) can be defined by a root chord Cr. bolts. the approach provides a good starting point for the estimation of drag. Swet.). however. and a semispan. While this is not true.4 I Aerodynamic Force and Moment Coefficients 247 Of course. b/2. When the airplane drag coefficient is 'defined in this way. each spanwise station would have a different Reynolds number. wires. The following approach to determining subsonic aircraft parasite drag is due to Shevell (1989) and also is presented in Schaufele (2000). is the form factor for each component. Because of this. Since the total skinfriction coefficient is a function of the Reynolds number at the "end of the plate". and hence.Swet. CD0 = i1 (5. A wing with a trapezoidal planform (as shown in Fig. including: • Interference drag • Excrescence drag (the drag due to various small dragproducing protuberances. stabilizers. total skinfriction coefficient for each component.The difficulty comes in applying the flatplate skinfriction analysis to a wing with variable chord lengths along the span. many aerodynamicists refer to drag counts rather than the drag coefficient. which is usually the wing planform area). Most aircraft components fall into one of two geometric categories: (1) winglike shapes and (2) bodylike shapes.) • Engine installation drag • Drag due to control surface gaps • Drag due to fuselage upsweep • Landing gear drag The total aircraft drag coefficient is defined as Df CD = (5.__ Sec. This approach assumes that each component of the aircraft contributes to the total drag without interfering with each other.0100 would be 100 drag counts). an easier approach .0001 (a drag cOefficient of CD = 0. the term "drag count" refers to a drag coefficient of CD = 0. pylons. It would be possible to perform a double integration along the chord and span of the wing to obtain a total skinfriction coefficient for the wing. there are two basic ways to find the equivalent flatplate skinfriction coefficient for the various aircraft components. nacelles. Wing Method.33) where is usually the wing planform area for an airplane. this approach does not take into account a variety of sources of aircraft drag.34) where N is the total number of aircraft components making up the aircraft (wing.
In general.) of the wing as the appropriate flatplate length. so an empirical correction is often used. seams. from equation (4.87): — C1 = 0. The viscous sublayer .455 — 1700 ReL (5. based on the actual flight test data of aircraft compared with the drag prediction using the approach outlined in this section.c. Most subsonic aircraft have a 6 to 9% increase in drag due to surface roughness. such as rivets. Various approaches exist for making this correction. described in Section 5.3.36) It is important to remember that if a portion of the "theoretical" wing is submerged in the fuselage of the aircraft. there is no straightforward method for correcting for surface roughness. However. the skinfriction coefficient should be corrected for surface roughness and imperfections. can be calculated using the following formulation which is valid for a trapezoidal wing m. {Kroo (2003)].000). is to find an equivalent rectangular flat plate using the mean aerodynamic chord (m. including the correction for laminar flow (assuming transition takes place at = 500. = 21 3\ I Cr + CrCt Ct \ 1 Cr+CtJ = 2 3 (A2+A+1\ j 1 (535) where A = Ct/Cr is the taper ratio of the wing. The mean aerodynamic chord can then be used to define a "mean" Reynolds number for the wing: ReL = (5.a. etc. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics b12 Figure 5. Kroo reports that.248 Chap.37) Before proceeding any farther. then that portion of the wing should not be included in the calculation mean aerodynamic chord should be calculated using the root chord at the side of the fuselage! Now the total skinfriction coefficient for the wing can be found using the PrandtlSchlichting formula. some of which include small imperfections in the wing surface.15 for details about the viscous sublayer). gaps.18 Geometry of a wing with a trapezoidal planform." The key factor in correcting for surface roughness is the relative height of the imperfections in the surface compared with the size of the viscous sublayer of the boundary layer (see Figure 4. "carefully built laminar flow. rivets.a. perhaps as low as 2 to 3%. The mean aerodynamic chord.c. composite aircraft may achieve a lower drag associated with roughness. etc.
Finally. as shown in Fig.38) . for mass production spray paint and to k = 6 X in.19.06 x in. the skinfriction coefficient can deviate from the smooth turbulent value by factors as high as 300%. Keeping aerodynamic surfaces as smooth as possible is essential to reducing skinfriction drag! Equiv alent sand grain roughness for different surfaces vary from approximately k = 0. the impact of surface roughness is increased at higher Reynolds numbers. while thinner wings have lower form factors and lower drag. increases in size.2t/C)Sexposed (5. k. 5. Thicker wings have higher form factors and hence higher drag. As the sand grain roughness. since the roughness will emerge from the sublayer and begin to impact the characteristics of the turbulent boundary layer [Hoerner (1965)].001 1010 ReL = UL P Figure 5. 5.19 Effect of surface roughness on skinfriction drag. the skin friction will increase accordingly.0(1 + O. to k = 2 X in. So. as the relative grain size increases. This figure is based on empirical information and shows the correction to the skinfriction coefficient to take into account supervelocities (flow acceleration over the wing which alters the boundary layer properties which are assumed to be based on freestream levels) and pressure drag due to flow separation.20. and eventually remains at a constant value with increasing Reynolds number. usually is contained within a distance from the wall of 10. [From Hoerner (1965)].Sec. if the equivalent "sand" grain roughness of the surface is contained within the viscous sublayer. than the surface is aerodynamically "smooth".010 kIL I 10—6 0. Notice that. An increase in wing sweep also tends to reduce the form factor and the drag coefficient. the wetted area of the wing can be calculated using the following relationship [Kroo (2003)]: Swet 2. 5. Since the thickness of the viscous sublayer actually decreases with Reynolds number.4 / Aerodynamic Force and Moment Coefficients 249 Relative grain size 0. for galvanized metal [Blake (1998)}. for a polished metal surface. The form factor for the wing can be found from Fig.
1.35  40° . 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics I I I 1.20  0 1.14 0.450 1. tic Figure 5.251.08 0. The wing parasitedrag coefficient now can be calculated from equation (5.45 0° — 100 15° — 20° 25° 1.34) CD0 = KCfSwet/Sref (5.18 0.The exposed area is doubled to take into account the top and bottom of the wing.40  1.04 0.] where Sexposed is the portion of the wing planform that is not buried within the fuselage.250 1. The thickness factor in the previous equation takes into account the slight increase in flatplate area due to the fact that the wing thickness increases the arc length of the wing chord. First.21).16 0.12 0.50 Chap. the calculation of the skinfriction coefficient is simpler than for a wing.15  1.34) Fuselage Method.10 0.05 100 0.06 0.20 Thickness ratio.30  50° 0 0 E I.30) . [From Shevell (1989).10 1. Since the fuselage has a single length (as shown in Fig. 5.02 0. find the Reynolds number at the end of the fuselage as ReL = (5.20 Wing form factor as a function of wing thickness ratio and quarterchord sweep angle.
a short. LID Figure 5. The skinfriction coefficient also can be corrected for surface roughness using the same approximation as discussed for the wing.4 I Aerodynamic Force and Moment Coefficients 251 ID '• L Figure 5.40 — 1 I I = 0. [From Shevell (1989).30  I.21 Geometry of a fuselage with circular cross sections. 5. which is defmed as the length of the fuselage divided by the maximum diameter of the fuselage.35  1. 1.. and the smaller the drag coefficient increment due to separation.25  U g 1.22 Body form factor as a function of fuselage fineness ratio.20 0 1.Sec.10  1.50 1. the smaller the form factor. and calculate the total skinfriction coefficient using equation (5. bluff body would have a high form factor due to large amounts of flow separation and hence a hither drag coefficient. Conversely.05 I I I I — 3 4 5 6 7 8 9 10 11 Fineness ratio. The form factor can be found in Fig.] .15 — 1. 0 1.22 and shows that the more long and slender the fuselage. The fuselage form factor is a function of the fineness ratio of the body. LID.37). 5.
the outside of the fuselage!) Total Aircraft Parasite Drag. as shown in Fig. the remaining components of the aircraft must be included.)] . 5.23 Aircraft showing the major components that contribute to drag. Swet = O.39) where the wetted areas can be found assuming that the various portions of the fuselage can be approximated as cones. Vertical and horizontal stabilizers — use wing method • Engine pylons—use wing method • Engines nacelles —use fuselage method • Antenna—use wing method • Figure 5. Now that the drag coefficient for the wing and fuselage have been calculated. using the same methods as have been described previously. cylinders.252 Chap.72lrDLtaii These formulas do not double the exposed area of the fuselage.23.40) O. 5 I Characteristic Parameters for Airfoil and Wing Aerodynamics The total wetted area of the fuselage can be calculated as + Swet + Swet (5.75lrDLnose lrDLbody (5. and conical sections [from Kroo (2003)]. (Namely. hopefully. since air only flows over one side of the fuselage. [(Boeing 777 photograph courtesy of The Boeing Company. For example. Most of the remaining components can be approximated as either wing surfaces or fuselage surfaces.
6 0 6 3 3 '1. Now that these wetted areas have been obtained.Sec. EXAMPLE 5. TABLE 5.10 0.5: Estimate the subsonic parasitedrag coefficient This example will show how the subsonic parasitedrag coefficient for the F16 can be estimated at a specific altitude.04 0.3 23.34) as N KiCfSwet. base drag. 5. It is probably more expedient to use an empirical correction for the remainder of the drag. the parasitedrag coefficient for each surface can be estimated.5 38.2 and 5. i=1 The remaining drag components are due to smaller affects. 5.4 / Aerodynamic Force and Moment Coefficients 253 This approach can be continued for the vast majority of the aircraft.4 7 12. This simplified approximation yields a total wetted area of 1418 ft2. flap gaps. which is very close to the actual wetted area of the F16. ft 14 Ct. (2004) and is reproduced here in Fig.38 and 5.2 F16 WingLike Surface Wetted Area Estimations Span.8 9.24 .4 (to match available flight test data). the subsonic drag coefficient is a function of altitude and Mach number. and other small protuberances can be found in Hoerner's book on drag [Hoerner (1965)].24. which will serve as the reference area for the aircraft: Sref = 300 ft2 Solution: The first task is to estimate the wetted area of the various surfaces of the aircraft.5 *not snown in Fig.5 2 tic 0. as shown in Fig. bolts.06 0.24 and Tables 5.04 0.40).06 112 Wing (1 and 2) Horizontal tail (3 and 4) Strake (5 and 6) Inboard vertical tail (7) Outboard vertical tail (8) Dorsal fins (9 and 1O)* 419. ft '12 Surface Cr. 5. The wetted area for these simplified geometric surfaces is approximated using equations (5. It is important to note that. A good estimate of these areas has been completed in Brandt et al. The theoretical wing area of the F16 is 300 ft2.9 6 2 7. which is 1495 ft2 [Brandt et al. unlike other aerodynamic coefficients. Some of these drag components would be extremely difficult (and time consuming) to calculate.4 117.3 77. In spite of this. 5. since the Reynolds number over the surfaces of the aircraft will vary with altitude and Mach.3. Now the total aircraft zerolift drag coefficient can be found by using equation (5.6 26. and interference drag. semiempirical and empirical methods for estimating the drag of rivets.000 ft and has a Mach number of 0. ft 3. This example will assume that the aircraft is flying at an altitude of 30. such as excrescence drag.The aircraft is approximated with a series of winglike and fuselagelike shapes. which is often done at aircraft manufacturers based on historical data from previous aircraft [Shevell (1989)].5 8 5 1.(2004)].
9 15.1 2 2 4 6. ft2 Net 39 6 ft2 42.3 2.8 2 42.3 F16 FuselageLike Surface Wetted Area Estimations Surface Fuselage (cylinder 1) Nose (cone 1) Boattail (cylinder 2) Side (half cylinder 1 and 2) Length.24 F16 geometry approximated by simple shapes. ft Height.7 2. to calculate the Reynolds number for the equivalent rectangular wing as Surface #1 Half Cylinder #1 Cylinder #1 Cone #1 Surface #5 V Surface #3 Cylinder #2 Half Cone #1 Half Cylinder #3 Half Cone #2 #6 Half Cylinder #2 Surface #4 Surface #2 Surface #8 Surface #7 Half Cylinder #4 Figure 5. ft width. ft 5wet.254 Chap. [From Brandt et al. (2004).1 1.4 62. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics TABLE 5.I .8 67. First.5 2 2 32.3 Wing.3 551.a.8 29.5 6 5 5 6 1 551.c. estimate the mean aerodynamic chord of the wing as 21 3\\ and cc TI \ 2/ 1(14+35 3\ (14)(3.1 Canopy(halfcylinder3) Engine (half cylinder4) Canopyfront(half cone 1) Canopy rear (half cone 2) 2 5 30 2 2.5)\ 14+3.3 4 24 5 0.5 2.5 5.4 62.5) use the m.7 180 3.
800 5.00280 0.00293 0.0 ft >< ReL = — (0.c.00228 From K= 1. so no roughness correction will be applied.20.0/0.06 and the parasite drag coefficient is CD = o KCfSwet 5ref = (1.00026 0.4 / Aerodynamic Force and Moment Coefficients 255 ReL= = 11.08 1.00280 Assume for the sake of this example that the wing is aerodynamically smooth.Sec.631 5.879 4.5*(2.800 ft) — 1107 X Finally.18 X = 0.92 ft/s)(9.472 6.000891 slug/ft3) (397. (requires some extrapolation from the graph) and the parasite drag coefficient is CD = TABLE 5. From Fig.06)(0.91 X 106)2. 5.18 Cf K 1.04 1. However.4 KCfSwet Sref (1. for a leading edge sweep of 40° and a thickness ratio of 0.067.455 1700 (log10 ReL)258 — ReL (log10 11. ft ReL(x106) 9.00280 0.58 — 11..708 4.083 11.05 ratio of L/D = 49.00295 0.00039 0.00415 The other wingLlike components of the F16 have had similar analysis performed.99 6.00082 0.4 ft2) 300 ft2 =0.000891 slug/ft3)(397.00296 0.00280)(419.568 7.4. 5.18 X 106)2.455 1700 — 0.00025 0. K 1.04.5 ft2) 300 ft2 = 0.91 X = 0.107 X 0.a.400 9.00304 0. L= Lnose + + Lboattail 6 + 39 + 4 = 49.06 1.00123 0.91 Cf = (log10 55.18 X 106 (0.0 ft) 3. a form factor correction should be performed.92 ft/s)(49.00524 F16 WingLike Surface ZeroLift Drag Estimations Surface m.303 10.00710 . drag predictions presented in Table 5. the total skinfriction coefficient for the wing is calculated as Cf — — 0.455 — lbs/ft2 1700 = 55.04 C00 Wing (1 and 2) Horjzontal tail (3 and 4) Strake (5 and 6) Inboard vertical tail (7) Outboard vertical tail (8) Dorsal Fins (9 and Total 6.05)(0.00228) (656.58 55. resulting in the zerolift Fuseiage.660 0.0) = 13.04 1.00415 0.06 1.5 + 5.
0 55.5 WINGS OF FINITE SPAN Much of the infonnation about aerodynamic coefficients presented thus far are for two dimensional airfoils (i. If we assume that the other components of drag account for an additional 10%.These results should be considered quite good for a fairly straightforward method that can be used easily on a spreadsheet.00590 1.23 0.256 Chap.25 0.01300 Since the total wettedarea estimate from this analysis was 5. Again. The total aircraft zerolift drag coefficient (assuming aerodynamically smooth surfaces) is CD0 CD0 (Wings) + CD0 (Fuselages) = 0..5 F16 FuselageLike Surface ZeroLift Drag Estimations Surface Length.0151.These streamwise vortices in turn create a downwash.0 24. it would be reasonable to increase the zerolift drag value by 5.0160 and CD0 = 0.00011 1.4% to take into account the simplicity of the geometry model. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics TABLE 5. such as from a CAD geometry).05 The other fuselagelike components of the F16 have had similar analysis performed.5. configurations of infinite span).39 12.0190 after correcting for engine effects and the presence of missiles in the flight test data [Webb.01 0.This would result in a zerolift drag coefficient of CD0 = 0.4% lower than the actual wettedarea of the F16 (something which could be improved with a better representation of the aircraft surfaces. (1977)].00710 + 0.00030 0.0 30.e.55 34. As a result of the induced downwash velocity.15 0. 5.00251 0.91 K CD0 Fuselage + nose + boattail Side (half cylinder 1 and 2) Canopy (front + center + rear) Engine (half cylinder 4) Total 27. which has the effect of "tilting" the undisturbed air. resulting in the zerolift drag predictions presented in Table 5.00244 0. ft ReL(X106) 49. the spanwise variation of lift for the resultant threedimensional flow field produces a corresponding distribution of streamwise vortices. and that there is no interference among the various components of the aircraft. reducing the effective angle of attack.01370.0 11. this result assumes that the surfaces are aerodynamically smooth. the lift generated by the airfoil section of a . et al. For a wing of finite span that is generating lift. the pressure differential between the lower surface and the upper surface causes a spanwise flow.00276 0.00590 = 0. Initial flight test data for the F16 showed that the subsonic zerolift drag coefficient varied between CD0 = 0. then our final estimate for the zerolift drag coefficient would be CD0 = 0.00228 0.00524 1. that there are drag increments due to excrescence or base drag.00025 1. As will be discussed in Chapter 7.
1 Lift The typical lift curve for a threedimensional wing composed of a given airfoil section is compared in Fig. as discussed in Section 5.The liftcurve slope for threedimensional unswept wing (CL.41) The pilot of an F16 wants to maintain a constant altitude of 30.6: An F16C in steady.if).5 / Wings of Finite Span 257 Cl (the lift coefficient for the 2D airfoil) CL for the 2D airfoil (the lift coefficient for the 3D wing) a wing Figure 5. 5.Sec.1.5. EXAMPLE 5. twodimensional airfoil.25 Comparison of the liftcurve slope of a twodimensional airfoil with that for a finitespan wing.25 with that for a twodimensional airfoil having the same airfoil section. known as the vortex drag (or the induced drag).6 and 0. where the angles are . twodimensional airfoil (which will be represented by the symbol C/. Furthermore. 5. the trailing vortex system produces an additional component of drag. one of constant altitude).3c1 (5.. Note that the liftcurve slope for the threedimensional wing (which will be represented by the symbol CLif) is considerably less than the liftcurve slope for an unswept. Recall that a liftcurve slope of approximately 0.1097 per degree is typical for an unswept. finitespan wing which is at the geometric angle of attack a is less than that for the same airfoil section of an infinitespan airfoil configuration at the same angle of attack. Incompressible flows for threedimensional wings will be discussed in detail in Chapter 7. typical values for which are between 0. for flight in a horizontal plane (i. level.if) can be approximated as C/if = ireAR where e is the efficiency factor.95. 5. unaccelerated flight 1+ 57. Recall from the discussion at the start of the chapter.4.000 ft flying at idle power.e.
2 1. 0 4 8 12 16 20 24 28 a(°) Figure 5. Use the lift curves for the F16 aircraft.C.0 — — — — 1. Since the aircraft is to fly in a horizontal plane (one of constant altitude). 5. altitude = 30. = 2. or less.e. which were provided by the General Dynamics Staff (1976) and are presented in Fig.] small. The minimum flight speed (i.35 M.2 0. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics M0. the lift is equal to the weight.26 Trimmed lift coefficient as a function of the angle of attack for the F16C.9—————0. 6HT is deflection of the horizontal tail.258 1 Chap.2 until it reaches its minimum flight speed.35. W= L = (5. is the velocity at which the vehicle must fly at its limit angle of attack in order to generate sufficient lift to balance the aircraft's weight. which for the purposes of this problem will be Mach numbers of 0.—. Solution: Let us first calculate the lower limit for the speed range specified (i.26 for several Mach numbers.750 pounds. the pilot must increase the angle of attack of the aircraft in order to increase the lift coefficient to compensate for the decreasing dynamic pressure. which is 23.Thus. Assume that the lift curve for M = 0.e. as the vehicle slows down.6 1. the stall velocity). the stall speed).G. Therefore. 1. [Provided by the General Dynamics Staff (1976). Prepare a graph of the angle of attack as a function of the air speed in knots (nautical miles per hour) as the aircraft decelerates from a Mach number of 1.000 ft.A.42a) .. = 0. The wing (reference) area S is 300 ft2. the lift must balance the weight of the aircraft.6 — .2 is typical of that for incompressible flow. — — — 0.2 cL Missiles on C.8 0..
= 2W 2 = 1.2 S (5.2 353.4 3.26 to determine the angle of attack required to produce the required lift coefficient at a given Mach number are presented in the following table. Uco (—) (—) (°) (ftls) 1193.4 199. — r— r' — — Solving for Usta11.57 6.338 To calculate the lift coefficient as a function of Mach number from down to the stall value for the Mach number.91 0.125 0. the minimum velocity at which the weight of the F16 is balanced by the lift at 30.1 1. 5.9 2.2 .0008907 lbfs2/ft4)(1.Sec.c. 5.0008907 slugs/ft3.2 0.000 ft is 0. Ustati IA ITT \2 /2W 'I PooCi. the freestream density at 30.281 1.2 knots Thus.5 895.338 0. stall — / 2(237501bf) — (0. Thus. The values obtained using these equations and using the lift curves presented in Fig. The corresponding Mach number is 336.. 5. CL Uc.82 (knots) 706.2b.2 knots with the aircraft at an angle of attack of 27.5 ft/s = a 994.6 0.222 0.499 1.50 471. the lift coefficient must be the maximum attainable value. which corresponds to the limit angle of attack.37 795.8 530.9 0. or 0. From Table 1.26.0008907 lbf s2/ft4.2 27. the maximum value of the lift coefficient (assuming it occurs at an incompressible flow condition) is 1. CL Thus.5 336.8 0.000 ft is 199.85 ft/s = 0.5 I Wings of Finite Span 259 In order to compensate for the low dynamic pressure that occurs when flying at the stall speed (Ustaii).42b) The velocity in knots is given by = where has the units of knots.57)(300ft2) Ustaii = 336.maxS Referring to Fig.50.88 596.5 ft/s = 199.57.
level unaccelerated flight. The correlation between the angle of attack and the velocity (in knots). In fact. such as those due to separation). The angle of attack varies much more slowly with velocity at transonic Mach numbers. the threedimensional flow past a finitespan wing produces an additional component of drag associated with the trailing vortex system.260 Chap. as in determined in this example and as presented in the table. a general expression for the drag acting on a finitespan wing (for which the flow is three dimensional) or on a complete aircraft is — C D.27.mrn + ". Fig. S I Characteristic Parameters for Airfoil and Wing Aerodynamks I) 0 0 0 100 200 300 400 500 600 700 800 Freestream velocity (knots) Figure 5. k' is a coefficient for the inviscid drag due to lift (which is also known as the induced drag or as the vortex drag). This drag component is proportional to the square of the lift coefficient.27 Correlation between the angle of attack and the velocity to maintain an F16C in steady.min) 5 43 In equation (5.43). 5.52 Drag As noted at the start of this section. Similarly. .'—L 'r2 + '—L. Note how rapidly the angle of attack increases toward the stall angle attack in order to generate a lift coefficient sufficient to maintain the altitude as the speed of the F16 decreases toward the stall velocity. k" is a coefficient for the viscous drag due to lift (which includes the skin friction drag and the pressure drag associated with the viscousinduced changes to the pressure distribution. 5.
28 Flight data for a drag polar for F106A/B aircraft at a Mach number of 0.43) correlates these data.02 CD 0. 5. (1961)] Data for a subsonic drag poiar are presented in Fig. which are taken from Piszkin.5 I Wings of Finite Span 0.01 0. Note that the minimum drag occurs when CL is approximately 0.45).28 for F106A/B aircraft at a Mach number of 0.9. 5.28.2 0.min Referring to equation (5.44) (5. + k2CL + CDO (5.07.44) as CD = k1C3. one could also assume.4 261 0. we obtain mm CD = (k' + — + (CDmin + We can rewrite equation (5.00 0. Thus.05 Figure 5. CL is approximately 0. et a!. Thus.45) where = k' + k" I "2 — — '"' "r and CDO — L.3 CL 0. CDO is also known as the parasite drag coefficient.9. CDO CDrnin .03 0.07.0 0. 5.31). CLmm is relatively small. Equation (5. k2 is often neglected. (1961). Expanding the terms in equation (5.1 0.Sec. CDO is the drag coefficient when the lift is zero. [Taken from Piszkin et al.04 0. As noted when discussing Fig. In such cases.
29). approximate breakdowns (by category) for large.29 Contributions of different drag sources for a typical transport aircraft. such as roughness effects and leakage.29 and 5. the greatest contribution arises from turbulent skin friction drag.30. accounts for about 85% of the total drag of a typical transport aircraft (see Fig. one obtains the expression for the total drag acting on an airplane: — '—DO + r' Thus. when added to skin friction drag. which is independent of the lift coefficient and therefore exists when the configuration generates zero lift (CDO or and (3) the (2) the drag associated with lift compressibility related effects that are correlated in terms of the Mach number and are known as wave drag.] . 5.48)]. as constituting the remaining 15%. Roughness Miscellaneous Wave Interference Afterbody Figure 5.46) The reader will note that k1 has been replaced by k (since there is only one constant left). which. 5. Let us introduce an additional term to account for the contributions of the compressibility effects to the drag. interference effects between aerodynamic components. and miscellaneous effects. Thomas cited the pressure drag due to open separation in the afterbody and other regions.45). Although the relative importance of the different drag sources depends on the aircraft type and the mission to be flown. wave drag due to the compressibility effect at nearsonic flight conditions. the total drag for an airplane may be written as the sum of (1) of parasite drag. we can write CD + (5. S / Characteristic Parameters for Airfoil and Wing Aerodynamics Incorporating these comments into equation (5. subsonic transports are presented in Figs. The next most significant contribution arises from liftinduced drag. How one accounts for both the inviscid and the viscous contributions to the liftrelated drag will be discussed under item 2 in a subsequent list [(see equation (5. According to Thomas (1985). [From Thomas (1985). Thus.262 Chap.
Bowes (1974) presented the lift/drag polar which is reproduced in Fig. fuselage. and the tail 10%. Since the wing . and so on. the wingroot boundary layer is more easily separated in the presence of an adverse pressure gradient.020.30 Lift/drag. North Atlantic Treaty Organization (AGARD. Figure 5. the entire flow is subsonic).polar for a large.The original version of this material was first published by the Advisory Group for Aerospace Research and De velopment. [From Bowes(1974). 1. A representative value of CDO would be 0.30. Zerolift drag. The majority of the liftrelated drag is the vortex drag for an elliptic load distribution at subcritical speeds (i.Sec. nacelles.5 I Wings of Finite Span KLL 263 • Friction • Pressure • Interference • Roughness •Nonelliptic load vortex • Elliptic load vortex • Wave • Shock separation 'Friction CL. Adequate filleting and control of the local pressure gradients can minimize the interaction effects.. subsonic transport. When evaluating the zerolift drag.e. tail. a low wing position on a circular fuselage would be sensitive to this interaction. Bowes notes that a good wing design should approach an elliptic loading at the design condition. NATO) in Lecture Series 67. Because of the interaction. Since the upper wing surface has the more critical pressure gradients.] Using slightly different division of the dragcontribution sources. Skin friction and form drag components can be calculated for the wing. 5. one must consider interactions such as how the growth of the boundary layer on the fuselage reduces the boundarylayer velocities on the wingroot surface. 5. the fuselage and the nacelles 40%. of which the wings may account for 50%. May 1974.
Maximum climb angle for jetpowered airplanes 3. by sweeping the wings. Maximum range of propellerdriven airplanes 2.3 Lift/Drag Ratio The configuration and the application of an airplane is closely related to the lift/drag ratio.e. 2. freestream Mach numbers of approximately 0.264 Chap. the compressibility effects) as the third term. Typical values of the airplane efficiency factor range from 0.5. illustrate how the drag coefficient increases rapidly with Mach number in the transonic region. Another factor to consider is the effect of compressibility. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics constitutes such a large fraction of the drag. At high lift coefficients (near CL. e should be changed to account for increased form drag. The aerodynamic characteristics of the F46. As will be discussed in Chapter 9. Wave drag. Maximum endurance for jetpowered airplanes. 5. known as wave drag. Although other factors must be considered. Representative values of the maximum lift/drag ratios for subsonic flight speeds are as follows: Type of airplane Highperformance sailplane Commercial transport Supersonic fighter (L/D)max 2540 1220 4—12 . When the freestream Mach number is sufficiently large so that regions of supersonic flow exist in the flow field (e. The correlation of equation (5. Many important items of airplane performance are obtained in flight at Performance conditions that occur at (L/D)max include 1. or greater). which will be incorporated into Example 5.7. the reasoning implies that an optimum airplane configuration would have a minimum wing surface area and.max).47) includes wave drag (i. The drag due to lift may be represented as = C2 L 'Tr(AR)e (548) where e is the airplane efficiency factor. 3. The deviation of the actual airplane drag from the quadratic correlation. and/or by using area rule. compressibilityrelated effects produce an additional drag component. Maximum poweroff glide ratio (for jetpowered or for propellerdriven airplanes) 4. Drag due to lift. the designer can delay and/or reduce the compressibility drag rise by using low aspect ratio wings.95..6 to 0. highest practical wing loading (N/rn2). is significant for airplanes with low aspect ratios and sweepback.g.7.. reducing the wing area would reduce CDO if all other factors are unchanged. therefore. where e is a constant.
47)].Sec.50 1. horizontal plane) unaccelerated flight.1168 0. . Had we used equation (5.05 1.7: Compute the drag components for an F16 in steady. since we are using equation (5. Recall from the discussion at the start of this chapter. as represented by equation (5. the lift must balance the weight and the thrust supplied by the engine must be sufficient to balance the total drag acting on the aircraft. level (i.3285 0.1168 0. level unaccelerated flight. in a constant altitude. for flight in a horizontal plane. Thus.46) to represent the drag polar. level.42b) for the lift coefficient as a function of the Mach number. The parasite drag can be calculated as Para D = The induced drag can be calculated as md D = where the induceddrag coefficient is given by ''Dirid In order for the aircraft to maintain steady.e. For this exercise. the table would include individual values for the profile drag [the first term in equation (5.47).10 0. 5.47)] and for the compressibility effects [the third term in equation (5.Therefore.0527 0.0208 0.46): ç00 + Consider the following aerodynamic characteristics for the F16: C00 k (—) (—) (—) 0. the tabulated values of CDO include both the profile drag and the compresented in the preceding pressibility effects..84 1. the lift must balance the weight.0479 0.0208 0. the values of table include the two components of the drag not related to the lift.5 / Wings of Finite Span 265 EXAMPLE 5. where the angles are small.4211 Note that.1667 0.0465 0. one can solve equation (5. unaccierated flight The pilot of an F16 wants to maintain steady. we will assume that the total drag coefficient for this aircraft is given by equation (5.80 0.42a).
the induced drag. Referring to equation (5. level. Note. Q) C J o 0. Solution: The first step is to use straight lines to generate values of CDO and of k in Machnumber increments of 0.6 0.0 Wing (reference) area(S) = 300 ft2 Airplane efficiency factor(e) = 0.84 include statement. Note also that .46). as mentioned in the problem 0. and the lifttodrag ratio as a function of Mach number in Machnumber increments of 0.The results are presented in the first three columns of the accompanying table.000 ft. The inclusion of the wave drag causes the drag coefficient CDO to peak at a transonic Mach number of 1. unaccelerated flight at 20.750 pounds in steady.4 0. Use linear fits of the tabulated values to obtain values of CDO and of k at Mach numbers other than those presented in the table. Calculate the parasite drag. level unaccelerated flight at an altitude of 20..0 0.9084 Consider an F16 that weighs 23. the values of for Mach numbers greater a significant contribution of the wave drag to the components of drag not related to the lift.31 Drag components for an F16C flying in steady.266 Chap.1.000 ft (standard atmospheric conditions).8 1.1. the total drag is the sum of the parasite drag and the induced drag.8 Mach number (—) Figure 5.4 1.2 0. the total drag. Additional aerodynamic characteristics for the F46 Aspect ratio (AR) = 3. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics C.6 1.2 1. C C'.05.0 1.
8 1.31. Note that when the total drag is a minimum (which occurs at a Mach number of approximately 0. the lifttodrag ratio is given by Weight — D — Total Drag — L — CL + 549 The computed values for the lifttodrag ratio are presented in the table and in Fig. the parasite drag is equal to the induced drag. .10 is consistent with equation (5. the value of k for a Mach number of 0. 5.6 0.0 1.0 0. and of the total drag are presented in the table and in Fig. Since the weight of the aircraft is fixed.2 1.8 Mach number (—) Figure 5. The freestream density (0. 5.48). of the induced drag. The computed values of the parasite drag.6 1.001267 slugs/ft3) and the freestream speed of sound (1036.) Since the lift is equal to the weight.000 ft. That is. The fact that the induced drag and the parasite drag are equal at this condition is an underlying principle to the solution of Problem 5.32.4 0.94 ft/s) for standard atmospheric conditions at 20.2 0.3. 5.__ Sec.32 LID ratio as a function of the Mach number for an F16C flying at 20.2b.5 I Wings of Finite Span 12 267 10 n 0 8 6 0 4 2 0 0.52).1168 The other tabulated values of k incorporate the effects of compressibility. k 1 1 = 7reAR = = 0.4 1.000 ft are taken from Table 1. the maximum value of the lifttodrag ratio occurs when the total drag is a minimum.
06 5412.13 19625.60 0.0807 0.19 28129.50 8230.0402 0.00 0.99 0.89 30787.0066 657.08 TotD (Ib) LID (—) 27.0479 0.60 0.1168 0.1435 0.70 3224.0012 417.25 61.94 k Vel (fps) 0.10 0.80 0.16 0.0208 0.55 933.25 1036.17 1335.86 829.0465 0.0359 15.79 2206.63 1743.0252 1289.0447 0.80 0.1542 0.1312 0.88 1151.40 1.75 32282.00 1.00 0.3665 4.98 2720.41 1659.0 IndD (Ib) (—) (—) (—) 103.2914 0.49 980.0039 503.93 20037.33 1348.9861 8060.51 0.21 1.92 0.0470 0.10 435.1168 0.02 3964.0021 425.50 0.ind (—) (—) 6.6657 7.1168 0.8858 5.4843 1.31 4965.80 0.68 0.2558 0.17 2082.0968 9.0208 0.75 681.60 1.95 551.0005 358.08 2352.3228 0.0027 447.0208 0.0511 0.75 0.73 2740.0208 0.1948 3582.94 1140.54 680.1816 0.9903 8.7824 1.7909 8.69 207.Aircraft F16C S (sq ft) span (ft) a (fps) 300 30 AR 3 e 0.0501 0.8124 10.2195 0.1853 1.1832 0.3902 0.0014 420.96 0.71 9134.08 1062.4211 414.17 15036.7264 0.00 0.1168 0.30 0.09 1532.85 24796.001267 Pdyn (psf) ParaD (Ib) CL W(Ib) 23750 Mach CD.0593 0.91 17302.9084 rho (s/ft3) 1036.30 22426.02 382.58 31146.72 1555.63 1244.22 2425.10 1762.3594 0.77 42.81 CD.5365 1.0688 0.0616 2015.74 0.46 12907.0017 417.31 108.0208 0.1168 0.9056 1.99 170.7768 32239.0454 0.9431 0.50 1.85 25184.1168 0.1162 0.8443 0.0517 0.39 311.52 0.20 0.0208 0.30 1.0590 0.90 1.50 170.63 0.0007 387.78 518.09 15457.80 1866.3403 1.0208 0.70 1.0009 403.0474 0.10 0.0961 0.67 27756.0208 0.3877 2.09 13325.02 1451.4649 0.70 0.62 1530.0006 372.81 9560.29 245.22 333.7357 2.0490 0.6222 2.0300 0.17 824.1168 0.80 22023.7625 .66 17719.2922 0.56 0.0522 0.3285 0.77 2695.0010 411.1168 0.16 725.2372 0.0122 895.40 0.75 0.41 11.22 0.20 1.47 622.
Problems 269 PROBLEMS 5. For the purposes of this problem.10) through (5.65 m2 and the leadingedge sweep angle is 60°.3.Assume that Rextr = 500.5. an expression for the aspect ratio of a delta wing was developed in terms of the leadingedge sweep angle (ALE).. The lift/drag ratio of a sailplane is 30.000 ft on a standard day = 0.3.3.8 m2 is landing at an airport.001756 slug/ft3) at 130 mi/h.6. If the airplane weighs 2300 lb.1. show that the lift coefficient for max imum lift/drag ratio and the maximum lift/drag ratio for an incompressible flow are given by CL(L/D)max = — 5. Neglect plate edge effects (i. if the airport is at sea level. A Cessna 172 is cruising at 10. level flight to calculate the required lift coefficients.7 5. 5. For this problem. and that the slope of the straightline portion of the CL versus a curve is CL. The sailplane has a wing area of 10.46) and treating CDO and k as constants.8. Using equation (5. Using the results of Problem 5.The measurements were taken from the centerplane of a rectangular . derive an expression for 5. For the delta wing of the F106. P5.000 defines the transition point. Consider a flat plate at zero angle of attack in a uniform flow where = 35 rn/s in the standard sealevel atmosphere (Fig.0 m2 and weighs 3150 N.4. what is the drag coefficient (CD) when the lifttodrag ratio is a maximum? That is.13) make use of pressure measurements taken from Pinkerton (1936) for NACA 4412 airfoil section. the reference area (S) is 58.7. Determine the section drag coefficient. An airplane that weighs 60. what is CD(L/D)max? 5. What is CD when the aircraft is in steady level flight at 170 km/h at an altitude of 1. What are the corresponding values of the aspect ratio and of the wing span? 5. What error would be incurred if it is assumed that the boundary layer is turbulent along the entire length of the plate? = 35 rn/s x Figure P5.000 N and has a wing area of 25. Using the results of Problem 5. In Example 5.a. Assume further that the wing has a positive camber so that its zerolift angle of attack (aoL) is negative. assume twodimensional flow).0 km? Problems (5. develop an expression for the aspect ratio in terms of the sweep angle of the quarter chord 5. (a) Graph CL as a function of the true airspeed over the airspeed range 300 to 180 km/h. Cd. (b) Repeat the problem for an airport that is at an altitude of 2000 m. what CL is required to maintain level flight? 5.7).e. Assume that the lift coefficient is a linear function of for its operating range. assume that the airplane is in steady.2.1.9.
90 —1.00 97.589 0.066 39. which are presented in Table 5.92 64. 5.053 —0.92 1.374 0.896 0.136 0.028 0.52 84.164 0.379 —3.2°) 0. Use equation (3.72 —2.024 —0.178 0.011 —0.94 54.120 0.86 89.100 0.254 0.648 .190 —0.098 0.082 0.426 49.266 —0. Graph as a function of x/c for these three angles of attack.551 0.16 Values of the pressure coefficient.e.54 —2.20 —0.128 0.131 —0.485 0.96 24.989 0.94 9.716 —0.98 14.148 0.90 29.322 0.559 —0.28 —0. Comment on the movemen of the stagnation point.48 —0..6 Experimental Pressure Distributions for an NJACA 4412 Airfoil [Abbott and von Doenhoff Orifice location xStation (percentc c from the leading edge) zOrdinate (percent c above chord) 0 —0.270 Chap.296 0.6 for the midspan section of a 76. having a span of 30 inches and a chord of 5 inches.088 0.10 —2.06 —0.5°) 0.516 0.068 0.105 —0.92 0.22 —0.86 —2. Ci.10 —0.92 0.791 0.111 —0.106 0.84 —1.204 0.627 0.867 —1.179 0.94 2.502 —0.709 —1. the angle that the chord of the model makes with the direction of the flow in the region of the midspan of the model.10.2 cm by 12. the data allow us to develop some interesting graphs.380 —1.365 —0.025 —0.028 —0. are presented both for the physical angle of attack (a) and the effective angle of attack However. TABLE 5.993 0.76 —2.11.38 4.121 0.414 0.60 —1.98 44.66 0.24 0.257 0. 5 I Characteristic Parameters for Airfoil and Wing Aerodynamics planform wing." Thus.90 19.46 —2.36 0 —1.000.713 0.075 0.98 34.181 a = +16° (aeff = 13.. "In order to have true section characteristics (twodimensional) for comparison with theoretical calculations.90 —0.812 —1.980 0.150 —0.146 —0.44 —1.091 0.90 —2.94 74.6.154 94. (cxeff = a = —4° 4°) a = +2° = 1.7 cm model which had a NACA 4412 airfoil section. a determination must be made of the effective angle of attack. i.048 0.96 7.100. the experimentally determined pressure distributions.639 0.053 —0. "The determination of the effective angle of attack of the midspan section entails certain assumptions that are subject to considerable uncertainty?' Nevertheless. average Reynolds number: 3.152 0.010 0.70 0 —1.459 0.264 —1. Pinkerton (1936) spent considerable effort defining the reliability of the pressure measurements.151 0. Comment on changes in the magnitude of the adverse pressure gradient toward the trailing edge of the upper surface. How does this relate to possible boundarylayer separation (or stall)? 5.818 —1.30 —2. The test conditions included an average pressure (standard atmospheres): 21.64 0. Pressure distribution measurements from Pinkerton (1936) are presented in Table 5.13).86 —2.231 100.071 0.
92 3.994 0.16 9.333 0.68 5.90 98.649 —0.13.88 34.44 0.86 74.252 —0.043 —0.95 39.34 7.16 7.076 —0.769 —0.68 0.92 17.414 —0.091 —0.212 —0.50 9.48 0.76 8.438 —1.40 8.183 —0.017 —0.64 4.438 —0.640 —1.693 —0.478 —3.90 69.746 —0. integrate the experimental chordwise pressure distributions of Table 5.88 89.026 —0.961 —0.841 —0.2°) 0.90 7.6 to obtain values of .90 9.609 —0.535 —1.082 —0.939 0.322 —0.139 —0. why not? 5.700 —0.336 —0.52 9.551 —0.381 —0.765 —3. integrate the experimental chordwise pressure distributions of Table 5.116 —0.00 Source: Pinkerton (1935).90 79.191 —0.139 0.004 to calculate the maximum value of the local velocity at the edge of the boundary layer both on the upper surface and on the lower surface for all three angles of attack.240 —2.225 —0.96 22.722 —0.88 94.740 —0.623 —0.74 6.76 9.854 0.62 9.210 —4.54 6. how does the circulation (or lift) change with the angle of attack? 5.364 —0.88 2.786 —0. 1.56 7.92 84.92 59.298 —0. Using the smallangle approximations for the local surface inclinations.64 9.94 1. Using the smallangle approximations for the local surface inclinations.70 2.230 —5.961 4°) 13.147 —0. calculate dC1 a = da Does the section lift coefficient calculated using the pressure measurements for an angle of attack of 16° fall on the line? If not.22 8.016 —0.559 0.90 44.055 —0.76 5.a Va lues Orifice location xStation ('percentc zOrdinate cfromthe (percentc leading edge) above chord) 0 of the pressure coeffident. Assuming that the section lift coefficient is a linear function of a (in this range of a).44 24.329 —0.88 4.Problems TABLE 5.6 to obtain values of the section lift coefficient for a —4° and a = +2°.316 —0.96 12.099 —0.019 —0.148 —0.44 29.091 —0.210 —0.149 —1.319 —0.370 —0.485 —0.58 14.12.50) 0.94 4.78 3.44 19.721 —0.144 —0.069 0.303 —1.355 —0.84 9.262 —5.676 —0.80 49. If these velocities are representative of the changes with angle of attack.322 —0.80 9.143 —0.371 —0.6 271 Experimental Pressure Distributions for an NACA 4412 Airfoil [Abbott and von Doenhoff (1949)1.74 1.92 27.94 64.92 54.952 —1.264 —0. a—4° (aeff 0.378 —0.336 0.16 2.152 —0.269 —1.782 0.440 —2.56 2.568 —0.88 8.758 —1.742 —0.190 —2.68 0.681 a+2° (aeff = 1.471 a+16° (aeff —6.635 —0.709 —2.525 —0.
based on an available thrust of 10. the total drag act ing on the aircraft is equal to the thrust required from the power plant.52.1 760 0.15.0131 0." If the engine of the F16 generates 10. (. assume the following: Mach No. and e = 0. For the jet.0486 0.3515 0.000 ft. at what Mach number will the aircraft stall? Use the lift coefficients presented in the table of Example 5. slower This Mach number range is the region known as "reverse command. 5.0215 0.0321 1.4417 .1725 0. if one wishes to fly at speeds below a Mach number of 0.26). As calculated in Example 5.000 pounds. Thus.272 Chap. the total drag.8000 0. If one wishes to fly at speeds above a Mach num ber of 0.2292 0. (—) C00 k (—) (—) 0.0131 0. x is replaced by (x — 0.16. assume the following: Mach No.1754 0. 5. Calculate the parasite drag.1000 0. from a Mach number of 0. level unaccelerated flight (SLUF) at an altitude of 5 km (standard atmospheric conditions).0277 0.52 in order to sustain the new. Consider the Eurofighter 2000 in steady. what is the maximum Mach number at which the Fi 6 can sustain SLUF? Use the results of the table and of the figures in Example 5. from a Mach number of 0.0510 0.5800 Other parameters for the jet include S = 500 ft2.8.7. the pilot must advance the throttle to obtain more thrust from the engine.1.8500 1.1 to a Mach number of 1.7.1 to a Mach number of 1.2432 0. Calculate the parasite drag. If one seeks to maintain steady.8000 0.000 pounds of thrust when flying at 20. 5.70.1000 0. 5.000 ft (standard atmospheric conditions).8.What is the minimum Mach number at which the Fi6 can sustain SLUF at 20.000 ft. level unaccelerated flight (SLUF).7. and the lifttodrag ratio as a function of the Mach number in Machnumber increments of 0. the total drag.1 100 0. The weight of the aircraft is 32. the induced drag.57 (see Fig. Thus.042 5 0.5000 1. the induced drag.2c.1725 0.. Similarly.000 pounds? If the maximum lift coefficient is 1.52. For the Eurofighter 2000.5000 1. and the lifttodrag ratio as a function of the Mach number in Machnumber increments of 0. in equation (5.—) CDO k (—) (—) 0.10).14.4560 0. b = 36 ft.52. the thrust supplied by the aircraft's engine must balance the total drag acting on the aircraft. level unaccelerated flight (SLUF) at an altitude of 30.1.0220 0.0700 1. Recall that a positive pitching moment is in the noseup direction.0289 0.8600 1. Consider a jet in steady.000 ft exhibits a minimum total drag at a Mach number of 0. the pilot must initially retard the throttle to begin slowing down and then advance the throttle to obtain more thrust from the engine than would be required to cruise at a Mach number of 0. the total drag of an F16 flying at a constant altitude of 20. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics the section pitching moment coefficient relative to the quarter chord for each of the three angles of attack.
The airplane is cruising at a pressure altitude of 17. Calculated and measured pressure distributions over the midspan section of the NACA 4412 airfoil. NACA Report 563 Piszkin F. 2004. Midland Park. Hicks RM.Walacavage R. Harris RV. Santa Ana. The cruise Mach number is 0.500 kg. 1979.5 m.9% and the wing has 38° of sweepback at the 25% chordline. Reston. New York: John Wiley and Sons Pearcey HH. LeClare G.The average weighted airfoil thickness ratio is 9.000 ft on a standard day with a wing loading of 95 lb/ft2. 1961. Boundarylayer linear stability theory. Introduction to Aeronautics: A Design Perspective. AGARD Report 709. Aircraft lift and drag prediction and measurement. Explain why you would not need to correct your results for laminar flow. F16 Air Combat Fighter. Estimate the parasite drag (excluding base drag) on the missile. 2000. 1965. 5. A flying wing has a planform area of 4100 It2.18. VA: AIAA Braslow AL.0 ft. and a diameter of 2. Osborne J. General Dynamics Report F16060 Hall DW. The Elements of Aircraft Preliminary Design. Missile DATCOM. In Special Course on Stability and Transition of Laminar Flow. b = 10. Aeronautics. 1985. pp.4 ft. 1979. To fly on the wings of the sun.4ito 441 Braridt SA.30. CA: Aries Publications . Stiles RJ. 18:60—68 Hoerner SR.17. The weight of the aircraft is 17.The reference area for the missile is given by = lTd2/4 (the crosssectional area of the missile). Applied Aerodynamics:A Digital Textbook. and Flight Mechanics. AGARD Lecture Series 67. von Doenhoff AE. Aerodynamics. Haines AB. 2nd Ed. New York: Dover Blake WB. Note D3579 General Dynamics Staff. Drag. an overall taper ratio of 0. 1984. 5. Use of grittype boundarylayer transition trips on windtunnel models.The body has a length of 20. In Transonic Aerodynamics. Stanford: Desktop Aeronautics Mack LM. and a span of 180 ft. Whitford R. Lockheed Horizons. The interaction between local effects at the shock and rear separation—a source of significant scale effects in windtunnel tests on aerofoils and wings. 1998. Convair Technical Report AD8163 Schaufele RD. Theory of Wing Sections.84. Michel R. 1968. Vol. Determine the following: (a) skin friction drag coefficient (assume a spraypainted surface) (b) pressure drag coefficient (c) induced drag coefficient (d) total drag coefficient REFERENCES Abbott IH. 35. 1936.25. A finless missile is flying at sealevel at 450 mph. Analysis and comparison ofAir Force flight test performance data with predicted and generalized flight test performance data for the F106A and B airplanes. 1966. NJ: published by the author Kroo I. NASA Tech. Ed. AGARD Conference Proceedings. AFRL TR19983009 Bowes GM. 1949. 2003. a root chord at the airplane centerline of 36 It.References 273 Other parameters for the Eurofighter 2000 include S = 50 m2. and e = 0.3—i to 3—81 McCormick BW. 1974. Paper 1 Pinkerton RM. Bertin JJ.
Hicks RM. London: Jane's Yearbooks Taylor JWR. 1978. London: Jane's Yearbooks Shevell RS.. ed. Washington. Couf. Lockheed Horizons 19:22—32 Van Dam CP. 1984—1985. Thomas ASW. Holmes BJ. Presented at Atmospheric Flight Mech. 1966. 1977. NJ: Prentice Hall Sterret JR. 1986. Williamsburg. 1966. In Performance Prediction Methods. Fundamentals of Flight. ed. Webb JB. 862229. 1973. Boundarylayer transition pifects on airplane stability and control. Englewood Cliffs. Whitehead AH. 1989. 1985.274 Chap. 1979. 1984. 19661967. Correlation of F16 aerodynamics and performance predictions with early flight test results. 5 / Characteristic Parameters for Airfoil and Wing Aerodynamics Schlichting H. Kent DR. DC: AIAA Taylor JWR. AIA4 Pap. Morrisette EL. 7th Ed. VA Webb TS. Transition fixing for hypersonic flow. Jane's All the World's Aircraft. NASA Tech. Aircraft viscous drag reduction technology. ed. 1973—1974. 242. Boundary Layer Theory. 19 . AGARD Conference Proceeding. Northrop F5 Case Study in Aircraft Design. New York: McGrawHill Ed. Jane's All the World'sAircraft. Jane's All the World's Aircraft. Note D4129 Stuart WG. London: Jane's Yearbooks Taylor JWR. Vol.
a distortion of the stream that is due to the airfoil thickness. For the rest of this book. For the rest of this book. that the airfoil is a wing of infinite aspect ratio). they approximate the flow per unit width (or per unit span) around the airfoil sections that are elements of a finitespan wing. 275 . it will used for the twodimensional flow fields that would exist when identical airfoil sections are placed side by side so that the spanwise dimension of the resultant configuration is infinite. and a circulatory flow that is related to the lifting characteristics of the airfoil. Thus. the term wing will be used when the configuration is of finite span. That is. To obtain the governing equations.1 GENERAL COMMENTS Theoretical relations that describe an inviscid. these twodimensional airfoil flow fields will be applied to a slice of a wing flow field. The flow around a twodimensional airfoil can be idealized by superimposing a translational flow past the airfoil section. it is assumed that the airfoil extends to infinity in both directions from the plane of symmetry (i. For many applications (see Chapter 7). Neglecting interactions with the tunnelwall boundary layer. Thus.. the flow around the model does not vary in the spanwise direction.e. the term airfoil will be used when the flow field is two dimensional. lowspeed flow around a thin airfoil will be developed in this chapter.6 INCOMPRESSIBLE FLOWS AROUND AIRFOILS OF INFINITE SPAN 6. the flow field around the airfoil is the same for any cross section perpendicular to the wing span and the flow is two dimensional. But the term airfoil will also be used when a finitespan model with identical cross sections spans a wind tunnel from walltowall and is perpendicular to the freestream flow.
the interaction between large pressure gradients and the viscous. Bernoulli's equation for steady. greater than that past the lower surface. the circulation around the line of Fig. Thus.This flow model assumes that the airfoil is sufficiently thin so that thickness effects may be neglected. greater than the pressure on the upper surface. they do provide valuable information about the aerodynamic characteristics for airfoils in incompressible air streams. 6. the pressure on the lower surface of the airfoil is.Thickness effects can be treated using the source panel technique. drag due to skin friction. 6. as represented by two elementary flows. the flow around the airfoil can be represented by the combination of a translational flow from left to right and a circulating flow in a clockwise direction. if one neglects the effects of viscosity. To generate high lift. incompressible flow leads to the conclusion that the velocity over the upper surface is on the average. Furthermore. the line integral of the velocity around any curve completely surrounding the airfoil is zero. . as shown in Fig.e. 6 / Incompressible Flows Around Airfoils of Infinite Span Since it is a twodimensional configuration. The analytical values of aerodynamic parameters for incompressible flow around thin airfoils will be calculated in Section 6. discussed in Chapter 3. the viscous forces produce a velocity gradient near the surface of the airfoil and. 6.1.276 Chap. as discussed in Chapters 4 and 5. However. If the fluid is initially at rest. Although these formulations have long since been replaced by more rigorous numerical models (see Chapter 14).8.2 CIRCULATION AND THE GENERATION OF LIFT For a lifting airfoil.2. on the average. boundary layer produce a complex. one may either place the airfoil at high angles of attack or employ leadingedge devices or trailingedge devices.2a is zero. form drag). an airfoil jn an incompressible stream experiences no drag force. as explained in Section 3. According to Kelvin's theorem for a Figure 6. Thus. The comparison of the analytical values of the aerodynamic parameters with the corresponding experimental values will indicate the limits of the applicability of the analytical models. because the velocity is zero for all fluid particles..6 through 6.15.3 through 6. Reynolds number—dependent flow field. Either way.1 Flow around the lifting airfoil section. hence. the presence of the viscous flow near the surface modifies the inviscid flow field and may produce a significant drag force due to the integrated effect of the pressure field (i.5 using classical formulations from the early twentieth century. The desired characteristics and the resultant flow fields for highlift airfoil sections will be discussed in Sections 6.
2b and c. ds =0 A1+$A2 (c) Figure 6. frictionless flow.2b and c is not zero but equal in magnitude and opposite in sign to the circulation around the other component curve. 6. 6. Thus. the circulation around the line which encloses the lifting airfoil and which contains the same fluid particles as the line of Fig.2 I Circulation and the Generation of Lift 277 =0 (a) Starting vortex A14 tA2 (b) fv. the flow is a potential flow with out circulation. the circulation around each of the component curves of Fig.3a. (b) fluid at time t. in Fig. with a stagnation point occurring on the rear upper surface. 6. the air is required to change . 6. the circulation around this line of fluid particles remains zero if the fluid is suddenly given a uniform velocity with respect to the airfoil. 6. therefore. However. Therefore.Sec. and the streamlines are as shown in Fig.2a should be zero. the lift are not produced instantaneously At the instant of starting. circulation is necessary to produce lift. 6.2.1 Starting Vortex When an airfoil is accelerated from rest. the circulation around it and. (c) fluid at time t + t. as explained in the following paragraphs.2 Circulation around a fluid line containing the airfoil remains zero: (a) fluid at rest. At the sharp trailing edge.
the lift increase progressively. A simple experiment that can be used to visualize the starting vortex requires only a pan of water and a small. the existence of circulation is not in contradiction to Kelvin's theorem. as the circulation around the airfoil and.2.3 GENERAL THINAIRFOIL THEORY The essential assumptions of thinairfoil theory are (1) that the lifting characteristics of an airfoil below stall are negligibly affected by the presence of the boundary layer. a vortex is shed which has the opposite direction of rotation relative to the initial vortex. Thus." If the board is stopped suddenly. However. since the maximum mean camber is small and the ratio of maximum thickness to chord is small.3b. However. since the circulation around the airfoil is equal in strength but opposite in direction to that of the starting vortex.278 Chap. thin board. another vortex is shed which has the same direction as the starting vortex. therefore. another vortex of equal strength but of opposite rotation is generated. and line of action) is only slightly influenced by the airfoil thickness. the gen eration of circulation around the wing and the resultant lift are necessarily accompanied by a starting vortex. if the velocity or the angle of attack is decreased.2. . because of viscosity. If the board is accelerated suddenly at moderate incidence. (b) full circulation. the line integral of the tangential component of the velocity around the curve that encloses area A1 must be equal and opposite to the corresponding integral for area A2. The circulation around the airfoil increases in intensity until the flows from the upper surface and the lower surface join smoothly at the trailing edge. If either the freestream velocity or the angle of attack of the airfoil is increased. A line which encloses both the airfoil and the starting vortex and which always contains the same fluid particles is presented in Fig. a surface of discontinuity emanating from the sharp trailing edge is rolled up into a vortex. which results because of the effect of viscosity.3 Streamlines around the airfoil section: (a) zero circulation. 6. as shown in Fig. direction suddenly. which is called the starting vortex. The stagnation point moves toward the trailing edge. and the air is unable to flow around the sharp trailing edge. stagnation point on the rear upper surface. Referring to Fig. The total circulation around this line remains zero. the large velocity gradients produce large viscous forces. the starting vortex will be seen leaving the trailing edge of the "airfoil. stagnation point on the trailing edge. Place the board upright in the water so that it cuts the surface. 6. 6. 6 I Incompressible Flows Around Airfoils of Infinite Span (a) (b) Figure 6. Instead. (2) that the airfoil is operating at a small angle of attack. 6. direction.Thus. and (3) that the resultant of the pressure forces (magnitude.
For thinairfoil theory.. the circulation that forms places the rear stagnation point exactly on the sharp trailing edge. In addition. the desired flow is obtained by superimposing on a uniform field of flow a field induced by a series of line vortices of infinitesimal strength which are located along the camber line. the sum of the components normal to the surface for these two velocities is zero. The boundary layer is assumed to be thin and. The length of an arbitrary element of the camber line is ds and the circulation is in the clockwise direction. The velocity field around the sheet is the sum of the freestream velocity and the velocity induced by all the vortex filaments that make up the vortex sheet. therefore.4 Representation of the mean camber line by a vortex sheet whose filaments are of variable strength 'y(s). airfoil sections have a maximum thickness of apprOximately. Typically.e. However. 6. there is a reduction in circulation z Leading edge Trailing x Figure 6. When the effects of friction are included.3b.= / y(s)ds JO (6. Furthermore.3 / General ThinAirfoil Theory 279 Note that we assume that there is sufficient viscosity to produce the circulation that results in the flow depicted in Fig.Sec. the Kutta the upper surface and the lower for an inviscid potential condition) requires that 'y 0 at the trailing edge. Thus. A velocity difference across the infinitely thin profile which represents the air foil section is required to produce the liftgenerating pressure difference. the airfoil will be repre sented by its mean camber line in order to calculate the section aerodynamic characteristics. 6. A vortex sheet coincident with the mean camber line produces a velocity distribution that exhibits the required velocity jump. 6. Ideally flow). For the vortex sheet to be a streamline of the flow. the condition that the flows from join smoothly at the trailing edge (i. 12% of the chord and a maximum mean camber of approximately 2% of the chord. we neglect the effect of viscosity as it relates to the boundary layer. Therefore. the boundary layer does not cause the flow to separate when it encounters an adverse pressure gradient. as shown in Fig. The total circulation is the sum of the circulations of the vortex filaments F.4. does not significantly alter the static pressures from the values that correspond to those for the inviscid flow model.1) where y(s) is the distribution of vortiqity for the line vortices. . it is necessary that the resultant velocity be tangent to the mean camber line at each point.
To calculate the resultant vortexinduced velocity at a particular point P. Thus. The chordwise location of the point of interest P will be designated in terms of its x coordinate. the Kutta condition places a constraint on the vorticity distribution that is consistent with the effects of the boundary layer. to calculate the cumulative effect of all the vortex elements.280 z 6 I Incompressible Flows Around Airfoils of Infinite Span Normal to the surface x Figure 6.5 Thinairfoil geometry parameters. it is necessary to integrate over the coordinate from the leading edge = 0) to the trailing edge = c)." Thus. The portion of the vortex sheet designated ds in Fig. 6.5 produces a velocity at point P which is perpendicular to the line whose length is r and which joins the element ds and the point P. one must integrate over all the vortex filaments from the leading edge to the trailing edge.2) The component of the freestream velocity normal to the mean camber line at P is given by = — 8p) . The chordwise location of a given element of vortex sheet ds will be given in terms of its coordinate. Noting that cosf52 = r and ds cos61 the resultant vortexinduced velocity at any point P (which has the chordwise location x) is given bys 1 = —— I (6.The induced velocity component normal to the camber line at P due to the vortex element ds is uvsn 2'rr where the negative sign results because the circulation induces a clockwise velocity and the normal to the upper surface is positive outward. relative to the value determined for an "inviscid flow.
6) . The approximate theoretical solution for a thin airfoil with two sharp edges represents an irrotational flow with finite velocity at the trailing edge but with infinite velocity at the leading edge. and a are small. stall will occur for the actual profile. when the profile is replaced by its mean camber line. As a result. the theoretical values of the lift coefficient (obtained by integrating the cir culation distribution along the airfoil) are in reasonable agreement with the experimental values. FlatPlate Airfoil (Symmetric Airfoil) 281 where a is the angle of attack and Op is the slope of the camber line at the point of in terest P. as will be discussed. For the camber line of the symmetric airfoil. dz —j— ax where z(x) describes the mean camber line. the angles 82. Thus. The desired vorticity distribution must also satisfy the Kutta condition that y(c) 0. dz\ dxi (6.4) The vorticity distribution that satisfies this integral equation makes the vortex sheet (and. the mean camber line) a streamline of the flow. For the actual airfoil. 03.4 I Thin. equation (6. the rounded nose allows the flow to accelerate from the stagnation point onto the upper surface without separation. equation (6.3) Since the sum of the velocity components normal to the surface must be zero at all points along the vortex sheet. Of course. FLATPLATE AIRFOIL (SYMMETRIC AIRFOIL) The mean camber line of a symmetric airfoil is coincident with the chord line. therefore. a flat plate with a sharp leading edge is obtained. Using the approximate trigonometric relations for small angles. 6. Thus.4 THIN. dzidx is everywhere zero. uco sin(a — tan1 (6. the approximate solution does not describe the chordwise variation of the flow around the actual airfoil.5) 6. a region of dead air (or stalled flow) will exist over the upper surface. Within the assumptions of thinairfoil theory.4) becomes 1 I ( \. when the angle of attack is sufficiently large (the value depends on the crosssection geometry). Because it does not account for the thickness distribution nor for the viscous effects. Thus. However. For subsonic flow past a flat plate even at small angles of attack. 1 cos 02 COS 03 2ir j0 I (x — = sin! a — / 1 dz'\ dxj J (6.Sec.5) becomes fc = (6.
) .9) That this is a valid solution can be seen by substituting the expression for 'y(O) given by equation (6. 'IT j0 cos0—cosO0 can be reduced to an identity using the relation nO dO Jo COS 0 — COS = sin nO0 sin 0o (6. the x coordinate transforms to 0o using x= The corresponding limits of integration are — cos0o) and Equation (6. must not only satisfy this integral equation.282 Chap. The resulting equation.10) where n assumes only integer values. and the total circulation. 6 / Incompressible Flows Around Airfoils of Infinite Span It is convenient to introduce the coordinate transformation = — cos0) (6. y(ir) = 0.9) into equation (6. but must satisfy the Kutta condition.9) and the coordinate transformation of equation (6. The solution is 'y(O) = 1 + cos 0 sin 0 . the freestream velocity. Using l'Hospital's rule.8).7). The lift per unit span is = J0 (6. an airfoil has no drag but experiences a lift per unit span equal to the product of freestream density.8) The required vorticity distribution. for twodimensional inviscid flow. the lift per unit span is f (1 + cos 0) dO = — ( .7) Similarly. it can be readily shown that the expression for 'y(O) also satisfies the Kutta condition. The KuttaJoukowski theorem for steady flow about a twodimensional body of and acts perany cross section shows that the force per unit span is equal to pendicular to Thus.11) Using the circulation distribution of equation (6.6) becomes i I y(0)sin0d0 21Tj0 cos0—cos00 = (6. (6. y(O).
"limitations imposed by practically realizable pressures may have a relatively insignificant effect on the normal force but could. known as the leadingedge suction force. twodimensional body in an irrotational. flow accelerates around the leading edge to the upper surface. Referring to the first paragraph in this section.14) because noseup pitching moments are considered positive.13) where a is the angle of attack in radians. The theoretical relation is independent of the airfoil thickness. The lift force." The pressure distribution also produces a pitching moment about the leading edge (per unit span). therefore. Thus. However. at the same time.4. As noted by Carison and Mack (1980). can become much greater than practically realizable values. because the airfoil thickness distribution and the boundary layer affect the flow field. Thus. FlatPlate Airfoil (Symmetric Airfoil) 283 To determine the section lift coefficient of the airfoil. From this stagnation point.e. producing a component of force along the leading edge. produces a nosedown pitching moment about the leading edge. As noted in Section 3. the pitching moment (per unit span) about the leading edge is m0 = — f (1 — cos2 0) dO (6. thinairfoil theory yields a section lift coefficient for a symmetric airfoil that is directly proportional to the geometric angle of attack. inviscid flow about a symmetric airfoil at a small angle of attack. which. the negative sign is used in equation (6. note that the reference area per unit span is the chord. severely limit the attainment of the thrust force. steady. The section lift coefficient is C1 = 2 2'ira (6.15. There is a stagnation point on the lower surface of the airfoil just downstream of the leading edge. therefore. "Linearized theory places no bounds on the magnitude of the peak suction pressure.2. that the resultant force has only a lift component and not both a lift and a drag component). See the discussion in Section 5.4 / Thin. the prediction of zero drag may be generalized to any general.14) 0 The liftgenerating circulation of an element produces an upward force that acts a distance downstream of the leading edge.The geometric angle of attack is the angle between the freestream velocity and the chord line of the airfoil. the approximate theoretical solution for a thin.9)]. resulting in zero drag. The high velocities for flow over the leading edge result in low pressures in this region.Sec. which exactly cancels the streamwise component of the pressure distribution acting on the rest of the airfoil. symmetric airfoil with two sharp edges yields an infinite velocity at the leading edge. 6. using the coordinate transformation [equation (6.7)] and the circulation distribution [equation (6. Consider an incompressible.1. It may seem strange that the net force for an inviscid flow past a symmetric airfoil is perpendicular to the freestream flow rather than perpendicular to the airfoil (i. Again. the actual twodimensional lift curve slope will be less than 2ir per radian. incompressible flow.15) ." However.. which is given by = — J (6.
The difference in the correlation between theory and data for these two airfoil sections is attributed to viscous effects. The theoretical lift coefficient calculated using equation (6. At higher angles of attack. The maximum thickness of the NACA 0012—64 is greater and located farther aft. where the resultant lift force could be placed to produce the pitching moment m0. Data are presented for two different airfoil sections. . it is valid only for angles of attack below stall.13) is in excellent agreement with the data for the NACA 0009 airfoil up to an angle of attack of 12°. = Cl — — (6. 6 / The section moment coefficient is given by Flows Around Airfoils of Infinite Span = ° m0 2 (6. For the symmetric airfoil.. theoretical values would not be expected to agree with the data at high angles of attack.15)] to the product of the lift [equation (6.3c. One.6. the adverse pressure gradients that cause separation of the viscous boundary layer and thereby alter the flow field would be greater for the NACA 0012—64.1: Theoretical aerodynamic coefficients for a symmetric airfoil The theoretical aerodynamic coefficients calculated using the thinairfoil relations are compared with the data of Abbottand von Doenhoff (1949) in Fig. one obtains = (6. Equating the moment about the leading edge [equation (6. Thus. According to thinairfoil theory. the NACA 0009. the viscous effects significantly alter the flow field and hence the experimental lift coefficients. has a maximum thickness which is 9% of chord and is located at x = 0. EXAMPLE 6. Since the theory presumes that viscous effects are small. The measured moments for the NACA 0009 are also in excellent agreement with theory prior to stall.16) Note that the reference area per unit span for the airfoil is the chord and the reference length for the pitching moment is the chord.17) The center of pressure is the x coordinate.284 Chap. the moment about the quarter chord is zero.12)] and the center of pressure yields — 2 2 = 2 Solving for xq.18) The result is independent of the angle of attack and is therefore independent of the section lift coefficient. Thus. 6. The correlation between the theoretical values and the experimental values is not as good for the NACA 0012—64 airfoil section.
19) Recall that this integral equation expresses the requirement that the resultant velocity for the inviscid flow is parallel to the mean camber line (which represents the airfoil).0 X 106 6.. the coordinate transformation = is — cosO) (6. because of camber.7) used..6 —20 —10 0 10 20 —0.JiI11r1I. the actual computations are more involved. 6. deg (a) Section angleofattack. .4 —1. CAMBERED AIRFOIL The method of determining the aerodynamic characteristics for a cambered airfoil is similar to that followed for the symmetric airfoil. [Data from Abbott and von Doenhoff (1949)J.8— —0.4 1.8 0 — a) a) a) 0 a) 0.00 a) 0.6Lift (standard roughness) 2. 1.0 X 106 0 6. Cambered Airfoil Theory 285 Data from Abbott and Doenhoff (1949) Ret: 0 3.6 Comparison of the aerodynamic coefficients calculated using thinairfoil theory for symmetric airfoils: (a) NACA 0009 wing section.0 x 106 9.0 0.0 X 2.Sec. Thus. 6.2 —20 —10 0 10 20 —0.0 o 0.6 L C a. a vorticity distribution is sought which satisfies both the condition that the mean camber line is a streamline [equation (6. deg (b) Figure 6.2 —0.4 11t!.5 THIN.4 Section angleofattack. (b) NACA 001264 wing section. so the integral equation to be solved is y(O)sinOdO 2ir if1T cosO — cosO0 — dx (6.) 08— a) 0. Again.8 —0.5 I Thin. However.0 ° a) —0.5)1 and the Kutta condition.
A Fourier sine series whose terms automatically satisfy the Kutta condition.20) is to represent the vorticity distribution which satisfies the condition that the mean camber line is a streamline. A1. = cos8 — cosO — cosO0 ir Jo cos00 dx .21) can be readily evaluated using equation (6. (6. Thus.5[cos(n Equation (6. .x).e.19) and the Kutta condi tion. To evaluate the series of integrals represented by the second term. = 01.. Since the geometry of the mean camber line would be known for the airfoil of interest.20)1 into equation (6.22) if equation (6.22) and must satisfy from all subsequent equations.. The first integral on the side of equation (6.22) which applies to any chordwise station. Substituting the vorticity distribution [equation (6.. determine the values of the coefficients.10).21) becomes = a — — 1)0 — cos(n + 1)8] A0 + cos nO (6. equation (6.21) This integral equation can be used to evaluate the coefficients A0. the slope is a known function of 0. A1. . 6 / Incompressible Flows Around Airfoils of Infinite Span The vorticity distribution y(0) that satisfies the integral equation makes the vortex sheet (which is coincident with the mean camber line) a streamline of the flow. A2. 6. + n=1 7(0) = Since each term is zero when B = sin0 (6. in terms of the angle of attack and the mean camberline slope. which satisfies equation (6. . The coefficients Thus.e.10) and the trigonometric identity: (sinn0)(sin0) = 0. Since we are evaluating both dz/dx and cos nO0 at the general point (i. A2.19) yields 1 A0(1 + cos0) Jo dO + j.286 Chap.. the coefficients A0.51 Vorticity Distribution The desired vorticity distribution. we have dropped the subscript 0 from equation (6. therefore. One can. one must use equation (6..20) the Kutta condition is satisfied [i. A term of the form for the vorticity distribution for a symmetric airfoil. 2UcoA0 1 + cos8 srn 8 2. may be represented by the series involving 1.. which is known for a given airfoil section.. of the Fourier series depend on the shape of the camber line.
10 / IT (a — A0) cos mO dO + JO / ITOO A. equation (6. as determined using equation (6. note that fir JO / thr any value of n. = 0. Note that P1T JO / A.22).20). cos nO cos mU dO TA.7)] and the expression for y [equation (6.Sec. Thus. by algebraic manipulation of equation (6. cos nO cos mU dO 0 when n in but fir / Thus. the general expression for the cambered airfoil includes the symmetric airfoil as a special case.24) Using equations (6. and integratihg frdni 0 to ir. one obtains JO f IT X cos mO dO ..23) and (6. Therefore.ir dz dx Multiplying both sides of equation (6.24) to define the coefficients. is y(O) = 1+ cos 0 This is identical to equation (6. A. for a symmetric airfoil. Thus. Note that. cos nO cos mO dO The first term on the righthand side is zero for any value of m.20)] gives us .... where m is an unspecified integer. the vorticity distribution for a symmetric airfOil. 6.22) by cos mO. 2 when n m A. The section lift coefficient is given by = 1  2 Jo I Using the coordinate transformation [equation (6. A0 = a. j — 1Tj0 2fdz — cos dx nO dO (6. A1 = A2 = A..5 / Thin) Cambered Airfoil To 287 evaluate A0.9).20) can be used to evaluate the vorticity distribution for a cambered airfoil in terms of the geometric angle of attack and the shape of the mean line.2 Aerodynamic Coefficients for a Cambered Airfoil The lift and the moment coefficients for a cambered airfoil are found in the same manner as for the symmetric airfoil. A0 a— 1 j — dO j '..S. 6.
as shown in the sketch of Fig. results cf2Ao + 2A1 2A0+A1 A2 Noting that C1 = ?r(2A0 + A1).27) Thus. Cm0 = — + A1 — (6. Thus.7 Center of pressure for a thin. 'no xcp The negative sign is used since a positive lift force with a positive moment arm in a nosedown. 6. one obtains + A1) (6. for the cambered airfoil. as well as the magnitude.26) The center of pressure relative to the leading edge is found by dividing the moment about the leading edge (per unit span) by the lift per unit span. upon integration. cambered airfoil. xcp = — mo( Figure 6.25) The section moment coefficient for the pitching moment about the leading edge is given by C.7. The line of action for the lift. Thus. using the coordinate transformation and the vorticity distribution.288 Chap. must be specified for each angle of attack. upon in tegration.fl0 1 2 2 / Jo Again. one obtains. the expression for the center of pressure becomes = + — A2)] (6. . the position of the center of pressure depends on the lift coefficient and hence the angle of attack. or negative moment. 6 / Incompressible Flows Around Airfoils of Infinite Span C1 = 21] A0( 1 + L 0 cos 0) dO + J0 n1 sin nO sin 0 dO 'IT Note that Jo C1 = I sin nO sin 0 dO = 0 for any value of n other than unity.
Thus. the section moment coefficient about the quarterchord point is independent of the angle of attack.5 / Thin. Thus.25c c = f = £  pC Again. If we include the effects of viscosity on the flow around the airfoil.30) Note.29). In order for the section pitching moment to remain constant as the angle of attack is increased. the center of gravity is usually located near the aerodynamic center.Thus. + rn0 (6.29) Since A1 and A2 depend on the camber only. as discussed when comparing theory with data in the preceding section. while the second integral represents the moment per unit span about the leading edge. the moment arm (relative to the aerodynamic center) decreases as the lift increases.24) is used to define A1 and A2. according to the theoretical relations for a thinairfoil section. the quarter chord). the Cmac is zero for symmetric airfoil.. the product of the moment arm (relative to the aerodynamic center) and C1 must remain constant. . the aerodynamic center is at the quarterchord point. The point about which the section moment coefficient is independent of the angle of attack is called the aerodynamic center of the section. Cambered Airfoil 289 If the pitching moment per unit span produced by the pressure distribution is referred to a point O.e. Because of these factors.Sec. which is given in equation (6. If equation (6. the aerodynamic center is the point at which all changes in lift effectively take place. However. This is evident in the expression for the center of pressure.27). Alternatively. Rearranging the relation yields j mo2SC = — J 0 The first integral on the righthand side of this equation represents the lift per unit span.28) The section moment coefficient about the quarterchord point is given by Cm025 = + Cm0 (A2 — A1) (6. the lift due to angle of attack would not necessarily be concentrated at the exact quarterchord point. the moment cois also given by efficient about the aerodynamic center. for angles of attack below the onset of stall. which is given the symbol becomes equation (6.25c downstream of the leading edge (i.Thus. then Cm ac = 2 dX (cos 20 cos 0) dO (6. the actual location of the aerodynamic center for the various sections is usually between the 23% chord point and the 27% chord point. the moment is given by O. 6. the signs are chosen so that a noseup moment is positive.
23) and (6.4ç.125 cos 0 — 0. the location of the maximum camber is 0 = 78.0555[0.24). the second digit defines the location of the maximum camber in tenths of chord. the equation of the mean camber line is — = C fore L C C while aft of the maximum camber position.3694 / (0.025 C Aft of the maximum camber location. the slope is given by \dXJfore 0. Forward of the maximum camber location..463° = 1. the maximum thickness in percent of chord). and A2. — cos8) = 0.3694 rad. and the last two digits represent the thickness ratio (i.e.1 — = 0. it is necessary to evaluate the coefficients A0. The airfoil section selected for use in this sample problem is the NACA 2412. 6 / Incompressible Flows Around Airfoils of Infinite Span EXAMPLE 62 Theoretical aerodynamic coefficients for a cambered airfoil The relations developed in this section will nOw be used to calculate the aerodynamic coefficients for a representative airfoil section. To do this. Therefore. Forward of the maxtmum camber position.0111 Since the maximum camber location serves as a limit for the integrals. Referring to equations (6. the necessary coefficients are A0 = a a 1 — r pl.8(i) — Solution: To calculate the section lift coefficient and the section moment coefficient. the slope is given by (—n 0. it is necessary to convert the x coordinate.11.025) dO + / .0111) dO — 0.125 cos 8 — 0. To evaluate these coefficients it is necessary to integrate an expression involving the function which defines the slope of the mean camber line. The equation for the mean camber line is defined in terms of the maximum camber and its location. A1.0555 cos 0 0. = 0.2 + 0.3694 (0. which is given in equation (6.0555cos0 — 0.004517 . As discussed in Abbott and Doenhoff (1949) the first digit defines the maximum camber in percent of chord. to the corresponding 0 coordinate. which is 0.0444 — c = 0.4c Thus.290 Chap.7). the slope of the mean camber line will be expressed in terms of the 8 coordinate.
08146 A2 = —I I f (0.0111 cosO) do] 13694 (0. the aerodynamic center is at the quar terchord point. Since the theoretical coefficients do not depend on the airfoil section thickness.Sec.OS5ScosOcos2O — 0.095° According to thinairfoil theory. they will be compared with data from Abbott and Doenhoff (1949) for a NACA 2418 airfoil as well as for a NACA 2412 airfoil.4c. Based on the measured lift coefficients for angles of attack for 0° to 10°. For both airfoil sections.O2ScosO)dO = 0. The experimental values of the moment coefficient referred to the aerodynamic center (approximately —0.per radian. 6.8) until the angle of attack becomes so large that viscous phenomena significantly affect the flow field. the section moment coefficient for the moment about the quarter chord is equal to that about the aerodynamic center.0 per radian for the NACA 2412 airfoil and approximately 5.9. The theoretical value of Cj.01387 The section lift coefficient is C1 = 2ir A0 + = 2ira + 0.a is 217.8 and 6. Thus. the maximum camber is 2% of the chord length and is located at x = 0.05309 The theoretical values of the section lift coefficient and of the section moment coefficients are compared with the measured values from Abbott and Doenhoff (1949) in Figs. respectively.2297 Solving for the angle of attack for zero lift.045 for the NACA 2412 section and —0. Cambered Airfoil 2 1 c'3694 L Jo 291 A1 = —I I + 2 (0. 6.0555cos29 — 0.2297 2ir rad = —2.050 for the NACA .5 / Thin.9 per radian for the NACA 2418 airfoil.125cos28 — 0. the experimental value of Cia is approximately 6. The two coefficients are given by Cm = = — A1) = —0. 6. The maximum thickness is 12% of chord for the NACA 2412 airfoil section and is 18% of the chord for the NACA 2418 airfoil section.l2ScosOcos2O — 0. we obtain = — 0. The correlation between the theoretical and the experimental values of lift is satisfactory for both airfoils (Fig.025 cos 20) dO (0. The theoretical value for the zero lift angle of attack agrees very well with the measured values for the two airfoils.0111 cos2O) dO L Jo pIT + 113694 = 0.
8 —0. Also note that.10. the variation in C1 max is less pronounced. Thus.2 0 —0..8 0 U 0. [Data From Abbott and von Doenhoff (1949). The cor relation between the experimental values of the moment coefficient referred to the quarter chord.6 1. (b) NACA 2418 wing section. deg (b) NACA 2418 wing section Figure 6. In Fig. Note that the maximum value of the experimental lift coefficient is consistently greater for the NACA 2412 and that it occurs at a higher angle of attack. the thickness ratio influences the interaction between the adverse pressure gradient and the viscous boundary layer. Note also that the experimentally determined location of the aerodynamic center for these two airfoils is between 0.239c and 0.max decreases rapidly with decreasing thickness.8 I) U 0 U 0 U a) V 0 U 0.0 x 106 6. in turn.0 x 106 (standard roughness) 2.6 c) 0 a) 0.8 —1.23c and 0. 16 —16 0 —1. : o 3. The interaction. 6. and the theoretical value is not as good.8 and 6. except as an implied limit to its applicability.6 —32 I I.] 2418 airfoil) compare favorably with the theoretical value of —0. 6. Above a thickness ratio of 12%. the location is normally between 0. C1 max is presented as a function of the thickness ratio for the NACA 24XX series airfoils. deg (a) NACA 2412 wing section Section angIe of attack.25c calculated using thinairfoil theory.9 show thicknessrelated variations. Ci.247c.292 Chap.4 1.0 0. The data of Abbott and von Doenhoff (1949) and the results of McCormick (1967) are presented. the measured lift coefficients decrease more sharply for the NACA 2412.8 Comparison of the aerodynamic coefficients calculated using thin airfoil theory for cambered airfoils: (a) NACA 2412 wing section. the data of Figs. which vary with the angle of attack. . McCormick notes that below a thickness ratio of approximately 12%. as compared with the value of O.0 x 106 o 6.27c for a real fluid flow.6 —32 —16 0 16 Section angle of attack. affects the aerodynamic coefficients. 6 / Incompressible Flows Around Airfoils of Infinite Span Theory— Data from Abbott and Doenhoff (1949). as the angle of attack increases beyond the maximum lift value.9). Although the thickness ratio of the airfoil section does not enter into the theory.0 0 'U V 0 —0.0 'U 0.. 6. As noted previously.053 (Fig.0 x 106 9.
Cambered Airfoil Theory 293 Data of Abbott and Doenhoff (1949).6 —0. to compute the airfoil lift and pitching moment coefficients for various configurations exposed to a wide range of flow environments. Henderson (1978) notes. if any. At low angles of attack where the boundary layer is thin and there is little.8 0.5 I Thin. the correlation degrades.6 —0.max is important.6 Section lift coefficient.9 Comparison of the theoretical and the experimental for section moment coefficient (about the aerodynamic two cambered airfoils: (a) NACA 2412 wing section. separation. C1 (b) Figure 6.] The correlations presented in Figs. Using repeated application of a panel method (see Chapter 3) to solve for the separated wake displacement surface. especially if knowledge of the maximum section lift coefficient Ci. 6.11. Ret: C 0 3. but as the angle of attack is increased.2 ç Ret: C 2.247 of ac: 0. The lift coefficient calculated using potential flow analysis with no attempt to account for the effects either of the boundary layer or of separation is compared with windtunnel data in Fig.8 1. 6.8 indicate that.7 x 0 8. These effects illustrated in Fig.241 0.1 la. [Data From Abbott and von Doenhoff (1949).1 '1) 0 —1.0 0 0.2 —1. potential flow analysis of the surface alone is a fair approximation to the data. 6.1 x 106 5. Henderson (1978) discussed the relative importance of separation effects.0 0 •0.1 o —0.242 0. at low angles of attack. 6. "Rarely will the boundary layers be thin enough that potential flow analysis of the bare geometry will be sufficiently accurate. it is necessary to include the effects of the boundary layer and of the separated wake.239 0. However.9 0.9 x 106 0.243 .8 x 106 8.9 x 106 0.Sec.9 0 0 X 106 5. (b) NACA 2418 wing section. C1 (a) Theory Data of ref.6 and 6.8 0.0 0. 6." By including the effect of the boundary layer but not the separated wake in the computational flow .239 0.0 Section lift coefficient. the theoretical lift coefficients based on thinairfoil theory are in good agreement with the measured values from Abbott and von Doenhoff (1949).
when one accounts for the boundary layer and the separated wake. culminating in the NACA developing laminarflow airfoils for use on fullscale aircraft (Jacobs.2 Cimax 5x 3x 2X lx 0.8  0.llc.294 Chap.4 — 0. 6. When the angle of attack is increased and separation becomes important. 6.16 0.81) shows that there is a fairly significant reduction in . model.6 = 8 X 1. 2.6 LAMINARFLOW AIRFOILS Airplane designers have long sought the drag reduction that would be attained if the boundary layer over an airfoil were largely laminar rather than turbulent (see Section 5.24 Figure 6. Designers since the 1930s have developed airfoils that could reduce drag by maintaining laminar flow.08 tic 0.5 x 0. used in the example of Fig. 1939). Separation effects must be modeled in order to predict the maximum lift coefficient. the predicted and the measured lift coefficients begin to diverge. 6.4.4 for de tails about boundarylayer transition and its impact on drag).0 1. Comparing equations (4.11. such as the GAW1. as shown in Fig. NACA 24XX series airfoil sections.25) and (4. 6 / Incompressible Flows Around Airfoils of Infinite Span Data of Abbott and von Doenhoff (1949): 06X106 Fairings of McCormick (1967).0 — 0. 6. the agreement between the theoretical lift coefficients and the windtunnel values is good at low angles of attack. max' Thjs will be the case for any gradually separating section.10 Effect of the thickness ratio on the maximum lift coefficient.00 0. As shown in Fig.llb. there is good agreement between theoretical values and experimental values through C1.
6 I I I 0 8 16 24 Angle of attack (deg) (b) 2.11 Relative importance of separation effects: (a) analysis of geometry alone.2 Theory 1.8 a) C) a) 0 1. [From Henderson (1978).0 0. 6.4 Data 1. (c) analysis with boundary layer and separated wake modeled.2 1.6 / LaminarFlow Airfoils 295 •• 1.0 0.] .Sec.8 a) • .6 I I I I 0 8 16 24 Angle of attack (deg) (c) Figure 6.8 a) C) • Data a) 0 C) I I I 0 8 16 24 Angle of attack (deg) (a) 2.4 Data 1. a) 0 1. (b) analysis with boundary layer modeled.
even eliminated. as shown in Fig. the flow transitioned to turbulent.296 Chap. and show a potential drag reduction of up to 25% over the conventional airfoil. 6. and many of the resulting shapes can be found in Theory of Wing Sections by Abbott and von Doenhoff (1949) as the 6digit airfoil series. 6. Unfortunately. and the laminar flow benefits were greatly reduced.81) (ReL)°'2 The early attempts at designing laminarflow airfoils centered around modifications to the airfoil geometry that would maintain a favorable pressure gradient over a majority of the airfoil surface.12. preferably to the midchord or beyond. Notice. — C1 1. 6.13.074 (4. This was accomplished primarily by moving the maximum thickness location of the airfoil further aft. Keeping an airfoil smooth is something .The "bucket" in the drag curve for the laminarflow airfoil occurs at angles of attack that normally might be required for cruise. that when the windtunnel model had typical surface roughness.including airplanes like the Piper Archer. however.328 = 0. [Loftin (1985).12 Shapes of two NACA laminarflow airfoil sections compared with the NACA 23012 airfoil section.] skinfriction drag (at reasonably high Reynolds numbers) if the boundary layers are laminar rather than turbulent. as shown in Fig. In the wind tunnel. 6 / Incompressible F'ows Around Airfoils of Infinite Span NACA 23012 NACA NACA Figure 6. these airfoils initially showed very promising drag reduction at cruise angles of attack.25) C= (4. An entire series of these airfoils were designed and tested. laminarflow airfoils do not function properly if the boundary layer transitions to turbulent. The PSi was the first production aircraft to utilize these laminar flow airfoils in an attempt to improve range by increasing the wing size and fuel volume for the same amount of drag as a turbulentflow airfoil (see Fig.14). These airfoil sections long have been used on general aviation aircraft. which can happen easily if the wing surface is not smooth.
6 / LaminarFlow Airfoils . C.2 1.024 23012 C .008 .2).2 —.14 Restored NACA P51 with laminar flow airfoil sections.004 —1. the use of NACA laminarflow airfoil sections has never resulted in any significant reduction in drag as a result of the achievement of laminar flow" [Loftin (1985)]. .13 Drag characteristics of NACA laminar flow and conventional airfoils sections with both smooth and rough leading edges.020 0 Rough 0 0 .028 . as shown in Section 13.4.] that is relatively easy to achieve with windtunnel models but rarely takes place with production aircraft.8 1.032 297 .6 Section lift coefficient. "As a consequence.016 C 0 0 .This has led to a variety of flowcontrol devices being used to actively maintain laminar flow. Figure 6. [Loftin (1985).] Figure 6. [Courtesy of NASA Dryden Flight Research Center.8 —. which usually does not make them practical as a dragreductiOn concept.Sec. but most of these devices require additional power sources (such as boundarylayer suction or blowing.012 .4 .4 0 . 6.
Laminar flow separates easier than turbulent flow and often leads to separation bubbles. The fact that the uses for laminar airfoils vary a great deal (ranging from UAVs. because induced drag is so much larger than skinfriction drag at higher Reynolds numbers. While transition trips do increase the skinfriction drag.) The relatively high Reynolds numbers of fullscale aircraft flying at common flight altitudes made the realization of lower drag using laminar flow impractical. (2001)]. The difference in design is dictated by the difference in application.16.298 Chap. but the boundary layer usually transitions to turbulent through the separation process. These separated flow regions reattach. but new applications have revived interest in laminar flow airfoils. As the size of the vehicle decreases (and as weight and Reynolds number also decreases). hence the low aspectratio design common for micro UAVs (2003)]. 6 I Incompressible Flows Around Airfoils of Infinite Span Figure 6. aspect ratio no longer is the dominant factor in creating drag— skinfriction drag becomes more important.15). ranging from very low aspectratio "flying discs" to very high aspectratio aircraft. to gliders and to highaltitude aircraft) means that there is no single optimum airfoil: each application requires a different wing and airfoil design in order to optimize performance [Tortes and Mueller (2004)J. Methods to overcome the laminar separation bubble include the tailoring of the airfoil geometry ahead of the bubble formation or by using transition trips. 6.The highaltitude aircraft (such as Hellos) also fly at low Reynolds numbers but require fairly heavy weights in order to carry the solar panels and batteries required for propulsion—the high aspectratio aircraft once again becomes more efficient.15 Candidate laminarflow airfoil aircraft:Black Widow Micro UAV (left) and Hellos highaltitude solarpowered aircraft (right). These aircraft have vastly different configurations. The most important fluid dynamics characteristics for the design of laminar airfoils are laminar separation bubbles and transition. (Black Widow from Grassmeyer and Keennon (2001) and Helios courtesy of NASA Dryden Flight Research Center. Heavier aircraft often have higher aspect ratios. 6. as shown in Fig. with the micro UAV flying at such low Reynolds numbers that typical design thinking about aspect ratio no longer holds.This led many researchers to windtunnel testing of laminarflow airfoils so that designers could choose optimum airfoil sections depending on their . including micro UAVs {Grass meyer and Keennon (2001)] and highaltitude aircraft (see Fig. if used properly they also can lead to a net reduction in drag due to the elimination of the bubble [Gopalarathnam et ad. leading to higher drag due to the bubble and turbulent flow after the bubble.
Sec. highlift airfoils.7 / HighLift Airfoil Sections 299 Figure 6. A good overview of the type of airfoils that work for various uses is presented by Selig (2003).XFOIL [Drela (1989)]. including wind turbines.17). and inverse design methods such as PROFOIL [Selig and Maughmer (1992)] and its applications [Jepson and Gopalarathnam (2004)]. [Selig (2003). and radiocontrolled sailplanes (candidate airfoils are shown in Fig.] requirements [Selig et al.7 HIGHLIFT AIRFOIL SECTIONS As noted by Smith (1975). "The problem of obtaining high lift is that of developing the lift in the presence of boundary layers—getting all the lift possible without causing separation. 6.16 Laminar separation bubble on an airfoil shown by surface oil flow where separation and reattachment are visible. Another approach is to use various numerical prediction methods which have been developed over the years. .These codes partially rely on semiempirical or theoretical methods for predicting laminar separation bubbles and transition and have been found to produce reasonable results.17 Candidate airfoils for radio controlled sailpianes. 6. (1989 and 2001)]. (1980 and 1990)]. [Selig (2003)]. 6. Provided that boundarylayer control is not used. our only means of obtaining Figure 6. airfoils with low pitching moments. including Eppler's code [Eppler et al.
Stevens et al. However. has been used in the work of Liebeck (1973).g. "The process of deceleration is critical." Stratford (1959) has developed a formula for predicting the point of separation in an arbitrary decelerating flow. the circulation around the airfoil section must be increased. A second method uses the panel method for the airfoil analysis [e. a nose flap. used to close the system. the airfoil section is represented by a closed polygon of planar panels connecting the input coordinate pairs. OlsOn et al. and changes in detail shape of a pres sure distribution. (2) the corresponding airfoil shape is practical and realistic. has two components: (1) analysis of the boundary layer. The resultant Stratford pressure distribution. a flap. or. prediction of separatiQn. changed camber. The velocity distribution is presented as a function of s. (1971)]. which recovers a given pressure distribution in the shortest distance. specifies that the upper.18). and determination of the kinds of flows that are most favorable with respect to separation. The science of developing high lift. where s begins at the lowersurface trailing edge and proceeds clockwise around the airfoil surface to the upper. there are sev eral possible means for improvement—changed leadingedge radius. and (2) analysis of the inviscid flow about a given shape with the purpose of finding shapes that put the least stress on a boundary layer. One method uses conformal mapping of the flow to a unit circle domain to generate the airfoil [e. for if it is too severe. Eppler and Somers (1980) and Liebeck (1976)]. when the higher velocities over the upper surface of the airfoil are produced in order to get more lift. once the effect of the boundary layer is included. (1978) note that. equivalently. the distance along the airfoiY surface. there are two options available for calculating the potential flow. if more lift is to be generated. 6 / Incompressible Flows Around Airfoils of Infinite Span higher lift is to modify the geometry of the airfoiL For a onepiece airfoil. a variablecamber leading edge. is presented in Fig.The lowersurface velocity is as low as possible in the interest of obtaining the maximum lift and increases continuously from the leadingedge stagnation point to the trailingedge velocity. since an adverse pressure gradient promotes transition) is used to ease the boundary layer's introduction to the severe initial Stratford gradient. A short boundarylayer transition ramp (the region where the flow decelerates. 6. In the scoordinate system. To develop a class of highlift airfoil sections.surface trailing edge. Once the desired airfoil velocity distribution has been defined. the velocities are negative on the lower surface and positive on the upper surface. 6. referring to [Smith (1975)]. An additional equation. in the potential flow analysis. The uppersurface acceleration region is shaped to provide good offdesign performance. therefore.Chap. modified to obtain a realistic airfoil.. larger adverse pressure gradients are required to de celerate the flow from the maximum velocity to the trailingedge velocity." Thus. the Kutta condition at the trailing edge requires that the uppersurface and the lowersurface velocities assume a value slightly less than the freestream velocity.. Again.and lowersurface velocities have a common limit at .19. separation develops.g. and (3) maximum possible C1 is obtained. the velocity over the upper surface must be increased relative to the velocity over the lower surface. The boundary condition for the inviscid flow—that there be no flow through the airfoil surface—is applied at each of the panel centers." The optimized form of the airfoil velocity distribution is markedly different than that for a typical airfoil section (which is presented in Fig. The "optimum" velocity distribution. Liebeck used a velocity distribution that satisfied "three criteria: (1) the boundary layer does not separate. Hence.
Sec. 6.7 / HighLift Airfoil Sections
u(s)
301
Upper surface
S
L.E. Stagnation point
Ute
S=
Lower surface
Figure 6.18 General form of the velocity distribution around a
typical airfoil. [From Liebeck (1973).]
— Optimum according to variation analysis
 . Modification necessary to obtain a practical airfoil transition ramp
Stratford distribution
U
Optimum according to variational
tile
/
analysis
,*—*———
s=sp
Figure 6.19 "Optimized" velocity distribution for a highlift, singleelement airfoil section. [From Liebeck (1973).]
the trailing edge (i.e., the Kutta condition). The effect of boundarylayer displacement is simulated by piecewise linear source distributions on the panels describing the airfoil contour. Thus, instead of modifying the airfoil geometry by an appropriate displacement thickness to account for the boundary layer, the boundary condition is modified by introducing surface
302
Chap. 6 I Incompressible Flows Around Airfoils of Infinite Span transpiration. Miranda (1984) notes that "The latter approach is more satisfactory because
the surface geometry and the computational grid are not affected by the boundary layer. This means that, for panel methods, the aerodynamic influence coefficients and, for finite difference methods, the computational grid do not have to be recomputed at each iteration." The boundary condition on the surface panels requires that the velocity normal to the surface equals the strength of the known source distribution. Liebeck has developed airfoil sections which, although they "do not appear to be very useful" [the quotes are from Liebeck (1973)1, develop an lid of 600.The airfoil section, theoretical pressure distribution, the experithental lift curve and drag polar, and the experimental pressure distributions for a more practical, highlift section are presented in Figs. 6.20 through 6.22.The pressure distributions indicate that the flow remained attached all the way to the trailing edge. The flow remained completely attached until the stalling angle was reached, at which point the entire recovery region separated instantaneously. Reducing the angle of attack less than 0.5° resulted in an instantaneous and complete reattachment, indicating a total lack of hysteresis effect on stall recovery. Improvements of a less spectacular nature have been obtained for airfoil sections being developed by NASA for light airplanes. One such airfoil section is the General Aviation (Whitcomb) number 1 airfoil, GA(W)1, which is 17% thick with a blunt nose and a cusped lower surface near the trailing edge. The geometry of the GA(W)1 section is
U,0
—3
—2
—1
0
+1
0.0
0.5
1.0
x
C
Figure 620 Theoretical pressure distribution for highlift, singleelement airfoil, = 3 X tmax = 0.125c, C1 = 1.35. [From
Liebeck (1973).]
Sec. 6.7 / HighLiftAirfoil Sections
1.8
303
1.6
Cl
1,4
0.04
0.02
0
—0.02
Cm
0.12/deg
C
—0.04 —0.06 —0.08
vSa
C1vs a
C1 vs Cd
0.03
4
8
I
a
12
16
—0.2
Cd
Figure 6.21 Experimental lift curve, drag polar, and pitching curve
for a highlift, singleelement airfoil,
Liebeck (1973).]
=
3 X 106.
[From
similar to that of the supercritical airfoil, which is discussed in Chapter 9. Experimentally determined lift coefficients, drag coefficients, and pitching moment coefficients, which are taken from McGhee and Beasley (1973), are presented in Fig. 6.23. Included for comparison are the corresponding correlations for the MACA 652—415 and the NACA 653—418 airfoil sections. Both the GA(W)1 and the NACA 653—418 airfoils have the same design lift coefficient (0.40), and both have roughly the same mean thick
ness distribution in the region of the structural box (0.15c to 0.60c). However, the experimental value of the maximum section lift coefficient for the GA(W)4 was approximately 30% greater than for the NACA 65 series airfoil for a Reynolds number of 6 X 106, Since the section drag coefficient remains approximately constant to higher lift coefficients for the GA(W)1, significant increases in the lift/drag ratio are obtained. At a lift coefficient of 0.90, the lift/drag ratio for the GA(W)1 was approximately 70,
304
—
Chap. 6 / Incompressible Flows Around Airfoils of Infinite Span
Potential flow Wind tunnel data
= 0"
—3.0
o =4°
a=
80
oa
120
cp
x
C
Figure 6.22 Comparison of the theoretical potentialflow and the experimental pressure distribution of a highlift, singleelement airfoil, = 3 X 106. [From Liebeck (1973).1
which is 50% greater than that for the NACA 653—418 section. This is of particular importance from a safety standpoint for light general aviation airplanes, where large values of section lift/drag ratio at high lift coefficients result in improved climb performance.
6.8
MULTIELEMENT AIRFOIL SECTiONS FOR GENERATING HIGH LIFT
As noted by Meredith (1993),"IIighlift systems are used on commercial jet transports to provide adequate low speed performance in terms of takeoff and landing field lengths, approach speed, and community noise. The importance of the highlift system is illustrated by the following trade factors derived for a generic large twin engine transport.
Sec.
I Multielement Airfoil Sections for Generating High Lift
NASA GA(W)1 airfoil
305
o NASA standard roughness o NACA standard roughness
NACA airfoils, NACA standard roughness
.
652—415
653—418
Cl
a, degrees
(a)
Figure 6.23(a) Aerodynamic coefficients for a NASA GA(W)1
airfoil, for a NACA 652—415 airfoil, and for a NACA 653—418 airfoil; = 0.20, = 6 X 106: (a) lift coefficient and pitching
moment coefficient curves; (b) drag polars. [Data from McGhee and Beasley (1973).]
1. A 0.10 increase in lift coefficient at constant angle of attack is equivalent to
reducing the approach attitude by about one degree. For a given aft bodytoground clearance angle, the landing gear may be shortened resulting in a weight savings
of 1400 lb.
2. A 1,5% increase in the maximum lift coefficient is equivalent to a 6600 lb increase in payload at a fixed approach speed. 3. A 1% increase in takeoff LID is equivalent to a 2800 lb increase in payload or a 150 nm increase in range.
306
Chap. 6 1 Incompressible Flows Around Airfoils of Infinite Span
NASA GA(W)1 airfoil
o NASA standard roughness
o NACA standard roughness
NACA airfoils, NACA standard roughness
652—415 ———
0,04 —
653—418
0.03
—
Cd
0.02
—
0.01
—
0.00 —
—1.2
I
I
I
I
—0.8
—0.4
0.0
0.4
Cl
0.8
1.2
1.6
2,0
(b)
Figure 6.23(b)
While necessary, highlift systems increase the airplane weight, cost, and complexity significantly. Therefore, the goal of the highlift system designer is to design a highlift system which minimizes these penalties while providing the required airplane takeoff and landing performance." Jasper et a!. (1993) noted, "Traditionally (and for the foreseeable future) highlift systems incorporate multielement in which a number of highlyloaded elements interact in close proximity to each other." Figure 6.24 shows a sketch depicting the cross section of a typical configuration incorporating four elements: a leadingedge slat,
Confluent Boundary Layers
Confluent Boundary Layers
Separatiqn
Reattachmen t
Separation
Separation
Figure 6.24 Sketch of the cross section of a typical highlift
multielement airfoil section. [Taken from Jasper et al. (1993).]
Sec. 6.8 [Multielernent Airfoil Sections for Generating High Lift
307
the mainelement airfoil, a flap vane, and trailingedge flap. Jasper et al. (1993) continued,
"Such configurations generate very complex flowfields containing regions of separated flow, vortical flow, and confluent boundary layers. Laminar, turbulent, transitional, and relaminarizing boundary layers may exist. Although high lift systems are typically deployed at low freestream Mach numbers, they still exhibit compressibility effects due to the large gradients generated. .. . It should be noted that many of the flowfield phenomena (e.g., separation, transition, turbulence, etc.) are areas of intense research in the computational community and are not yet fully amenable to computational analysis." As noted by Yip eta!. (1993), "Twodimensional multielement flow issues include the following:
1. compressibility effects including shock/boundarylayer interaction on the slat; 2. laminar separationinduced transition along the upper surfaces; 3. confluent turbulent boundary layer(s) —the merging and interacting of wakes from upstream elements with the boundary layers of downstream elements; 4. cove separation and reattachment; and S. massive flow separation on the wing/flap upper surfaces."
The complex flowfields for highlift multielement airfoils are very sensitive to Reynolds number—related phenomena and to Mach number—related phenomena. As noted in the previous paragraph, many of the relevant flowfield issues (e.g., separation, transition, and turbulence, etc.) are difficult to model numerically.
The airfoil configuration that is chosen based on cruise requirements determines a lot of important parameters for the highlift devices, such as the chord and the thickness distribution. Only the type of the highlift devices, the shape, the spanwise extensions, and the settings can be chosen by the designer of the highlift system. Even then, the designer is limited by several constraints. As noted by Flaig and Hilbig (1993), "Usually the chordwise extension of the highlift devices is limited by the location of the front spar and rear spar respectively, which can not be changed due to considerations of wing stiffness (twist, bending) and internal fuel volume." These constraints are depicted in the sketches of Fig. 6.25.
Slat TrailingEdge Gap
F/S
A321 Airfoil Section
R'S
Flap Chord
Slat Chord Thin Rear Section
Krueger Flap
A330/A340 Airfoil Section
Figure 6.25 General constraints on the design of highlift multielement airfoil sections. [Taken from Flaig and Hilbig (1993) .1
308
Chap. 6 / Incompressible Flows Around Airfoils of Infinite Span
Flaig and Hilbig (1993) note further, "Especially the required fuel capacity for a
longrange aircraft can be of particular significance in the wing sizing. Moreover, the inner wing flap chord of a typical low set wing aircraft is limited by the required storage space for the retracted main undercarriage. After the chordwise extension of the leading edge and trailing edge devices has been fixed, the next design item is the optimization of their shapes. The typical leading edge devices of today's transport aircraft are slats and Krueger flaps. In the case of a slat, the profile of upper and lower surface is defined by the cruise wing nose shape. Therefore only the shape of the slat inner side and the nose of the fixedwing can be optimized. A Krueger flap with a folded nose or flexible shape, as an example, generally offers greater design freedom to achieve an ideal upper surface shape, and thus gains a little in LID and CL,max. But, tradeoff studies carried out in the past for A320 and A340 have shown that this advantage for the Krueger flap is compromised by a more complex and heavier support structure than required for a slat." The required maximum lift capability for the landing configuration determines the complexity of the highlift system. In particular, the number of slots (or elements) of trailingedge devices has a significant effect on CL max The degrading effect of wing sweep on the maximum lift coefficient necessitates an increase in the complexity of the highlift system. The general trend of the maximum lift efficiency is presented as a function of the
system complexity in Fig. 6.26, which is taken from Flaig and Hilbig (1993). Note that the
maximum value for the coefficient of lift for unpowered highlift systems is approximately 3 (on an aircraft with typical 25degree wing sweep). Powered highlift systems with additional active boundarylayer control may achieve maximum values of the lift coefficient up to 7.
8
SSF = Single slotted flap
DSF = Double slotted flap TSF = Tnple slotted flap LED — Leading edge device
—
(Slat or krueger)
?
—L
+
4
—L.E.D. ±
SSF I DSF*]
FlOOD
A320 ooA34O
0
A300 A321
° °
A310 DC9
727
C17
2
C141
Unpowered highhft systems
F'
COMPLEXITY OF THE
SYSTEM
Figure 6.26 The maximum lift coefficient as a function of the complexity of the highlift system. [Taken from Flaig and Hilbig
Sec. 6.8 / Multielement Airfoil Sections for Generating High Lift
309
The problem of computing the aerodynamic characteristics of multielement airfoils
can be subdivided into the following broad topical areas, each requiring models for the computer program [Stevens, et al. (1971)1:
1. Geometry definition 2. Solution for the inviscid, potential flow 3. Solution for the conventional boundary layer 4. Solution for the viscous wakes and slotflow characteristics
5. Combined inviscid/viscous solution
Stevens et al. (1971) note that the geometric modeling of the complete airfoil, including slots, slats, vanes, and flaps, requires a highly flexible indexing system to ensure that conventional arrangements of these components can be readily adapted to the code. To compute the inviscid, potential flow, Stevens et a!. (1971) and Olson et al. (1978) use distributed vortex singularities as the fundamental solution to the Laplace equation.
Olson et al.(1 978), note that viscous calculations can be separated into three types of flows: conventional boundary layers, turbulent wakes, and confluent boundary layers (i.e., wakes merging with conventional boundary layers).These are illustrated in Fig. 6.27. To obtain a
complete viscous calculation, the conventional boundary layers on the upper and lower surfaces of the main airfoil are first analyzed, These calculations provide the initial conditions to start the turbulentwake analysis at the trailing edge of the principal airfoil.The calculations proceed downstream until the wake merges with the outer edge of the boundary
Confluent boundarylayer
analysis
Turbulent wake
analysis
Turbulent wake
analysis
Conventional laminar/turbulent boundarylayer
analysis
Figure 6.27 Theoretical flow models for the various viscous
regions. [From Olson et al. (1978).]
310
Chap. 6 / Incompressible Flows Around Airfoils of Infinite Span
Velocity profile for confluent boundary layer is initialized at the location where the wing wake merges with the flap boundary layer
Airfoil
Boundary conditions obtained from potentialflow solution at outer edge of viscous zone
N
Wake eddy diffusion model
Conventional boundarylayer eddy viscosity model
Figure 6.28 Flow model for merging of the wake from the principal airfoil with the boundary layer on the flap to form the con
fluent boundary layer on the upper surface of the flap. [From
Olson et al. (1978).] layer on the upper surface of the flap, as shown in Fig. 6.28. The wake from the principal airfoil and the boundary layer of the flap combine into a single viscous layer at this point, a socalled confluent boundary layer. The calculation procedure continues stepwise downstream to the flap trailing edge. At the flap trailing edge, this confluent boundarylayer solution merges with the boundary layer from the lower surface of the flap. The calculation then continues downstream into the wake along a potentialflow streamline. Although the techniques used to calculate the viscous effects differ from those described in the preceding paragraph, the importance of including the viscous effects is
illustrated in Fig. 6.29. Using repeated application of a panel method to solve for the separated wake displacement surface, Henderson (1978) found a significant effect on the pressure distributions both on the principal airfoil and the flap for the GA(W)1 airfoil for which was 12.5° and the flap angle was 400. Although a separation wake occured for both models, the agreement between the calculated pressures and the experimental values was quite good. As aircraft become more and more complex and as computational and experimental tools improve, the highlift design process has matured a great deal. As was stated earlier, including viscous effects in highlift design is important, but even with modern computer systems a highlift design still may require a combination of viscous and inviscid numerical predictions. The Boeing 777 highlift system was designed with various codes at different phases of the design process: a threedimensional lifting surface code was used during preliminary design, twodimensional viscousinviscid coupled codes were used to design the multielement airfoil sections, and threedimensional panel codes were used to evaluate flow interactions, NavierStokes and Euler codes were not used during
Sec. .6.8 / Multielement Airfoil Sections for Generating High Lift
Theorygeometry alone
—8
Theoryseparated
modelled
—6
0
'1)
0 0
—4
0)
—2
0
2
0
0.4
0.8
1.2
1.6
Chord fraction, x/c
NASA GA(W)1 airfoil, 30% Fowler flap, Angle of attack = 12.5° , flap angle = 40°
Flap
Figure 6.29 comparison of experimental and calculated pressure distributions on a twoelement airfoil with separation from both surfaces. [From Henderson (1978).]
the design process [Brune and McMasters (1990) and Nield (1995)]. This approach allowed for a reduction in windtunnel testing and resulted in a doubleslotted flap that was more efficient than the tripleslotted flaps used on previous Boeing aircraft. In fact, as the design process for highlift systems has matured, the systems have become less complex, more affordable, more dependable, and more efficient. Figure
6.30 shows how highlift airfoils designed by both Boeing/Douglas and Airbus have im
proved over the past twenty years. The improvement has been brought about by the increased use of numerical predictions, including the addition of NavierStokes and Euler methods to the predictions of highlift airfoils [Rogers et al. (2001) and van Dam (2002)].This evolution has been from tripleslotted flaps (such as on the Boeing
737), to doubleslotted flaps (such as on the Boeing 777), and now to singleslotted flaps (such as on the Airbus A380 and Boeing 787). However, numerical predictions still re
quire further improvement, including the addition of unsteady effects and improved
312
Chap. 6 / Incompressible Flows Around Airfoils of Infinite Span
r741L300]
1747/400!
!737/300!7371400!
737/500!
Tendency Boeing/Douglas
Tendency Airbus
A330 !A319!
1957
1967
1977
1987
1997
Figure 6.30 Design evolution of highlift trailing edge systems.
[From Reckzeh (2003).]
turbulence models, before highlift design will be as evolved as one might hope [Rumsey and Ying (2002) and Cummings et al. (2004)]. 6.9
HIGHLIFT MILITARY AIRFOILS
As noted by Kern (1996), "There are two major geometric differences that distinguish
modern high performance multirole strike/fighter military airfoils from commercial configurations: (1) leading edge shape and (2) airfoil thickness. Integration of stealth requirements typically dictates sharp leading edges and transonic and supersonic efficiency dictates thin airfoils on the order of 5 to 8% chord .. The Navy also depends on lowspeed highlift aerodynamics, since it enables high performance multirole strike aircraft to operate from a carrier deck." To obtain high lift at low speeds, the advanced fighter wing sections are configured with a plain leadingedge flap and a slotted trailing. .
edge flap. The schematic presented in Fig. 6.31 indicates some of the features of the complex flow field. The sharp leading edge causes the flow to separate, resulting in a shear layer that convects either above or below the airfoil surface. Depending on the angle of attack, the shear layer may or may not reattach to the surface of the airfoil. The flow field also contains cove flow, slot flow, merging shear layers, main element wake mixing, and trailingedge flap separation. Hobbs et al. (1996) presented the results of an experimental investigation using a 5.75% thick airfoil, which has a 14.07% chord plain leadingedge (L. F.) flap, a single slotted 30% chord trailingedge (T. F.) flap, and a 8.78% chord shroud. Reproduced in Fig. 6.32 are the experimentally determined lift coefficients for the airfoil with (the leadingedge flap deflection angle) equal to 34°, with (the trailingedge flap deflection angle) equal to 35°, and with (the shroud deflection angle) equal to 22.94°.This configuration provides the aircraft with the maximum lift required for the catapult and for the approach configurations. Note that, because of the leadingedge flow separation bubble, the lift curve displays
_____
Sec. 6.9 / HighLift Military Airfoils
Moving
313
attachment point
Leading edge bubble separation! transition! reattachment
Laminar/turbulent boundary layer
Offbody wake flow reversal
2layer confluent boundary layer
Free shear
Attachment Shroud Leadingedge flap
ation
Cove
Wake mixing
shear layer
Slot flow
Trailingedge flap
Ledge
Gap
Overhang tt Overhang
Figure 6.31 Sketch of the flow field for a military airfoil in a highlift configuration. [A composite developed from information presented in Kern (1996) and Hobbs (1996).]
=
2.4 2.0
1.6
15.9>< 106
M=
0.20
wingbox
1.2
0.4
0.0
—4
•
I
I
T.E. flap
I
I
—2
0
2
4
6
8
10
12
a(deg)
Figure 6.32 Total and component lift curves, = 22.94°. [Taken from Hobbs (1996).]
= 34°,
35°,
no "linear" dependence on the angle of attack. The maximum lift coefficient of approximately 2.2 occurs at an angle of attack of 2°.The airfoil then gradually stalls, until total separation occurs at an angle of attack of 10°, with a rapid decrease in the section lift coefficient. As noted by Kern (1996), "This behavior seems less surprising when considering splitflap NACA 6% thick airfoils which all stall around a = 4°."
Because the flow field includes trailing viscous wakes, confluent boundary layers, separated flows, and different transition regions, the Reynolds number is an important parameter in modeling the resultant flow field. The maximum values of the measured lift coefficients are presented as a function of the Reynolds number in Fig. 6.33. The data, which were obtained at a Mach number of 0.2, indicate that the maximum lift coefficient is essentially constant for Reynolds number beyond 9 X 106. These results (for
314
Chap. 6 / Incompressible Flows Around Airfoils of Infinite Span
2.0
0
2
4
6
8
10 12 14 16 18
Reynolds Number X
Figure 6.33 The effect of Reynolds number on Cimax, 5,, = 35°, = 22.94°. [Taken from Hobbs (1996).]
34°,
a twodimensional flow) suggest that airfoils should be tested at a Reynolds number of 9 X 106, or more, in order to simulate maximum lift performance at fullscale flight conditions. Conversely, testing at a Reynolds number of 9 X 106 is sufficient to simulate fullscale maximum lift performance.
PROBLEMS
6.1. Using the identity given in equation (6.10), show that the vorticity distribution
y(9) =
1
+ cosO sin 0
satisfies the condition that flow is parallel to the surface [i.e., equation (6.8)]. Show that the Kutta condition is satisfied. Sketch the distribution as a function of x/c for a section lift coefficient of 0,5. What is the physical significance of What angle of attack is required for a symmetric airfoil to develop a section lift coefficient of 0.5? Using the vorticity distribution, calculate the section pitching moment about a point 0.75 chord from the leading edge. Verify your answer, using the fact that the center of pressure is at the quarter chord for all angles of attack and the definition for lift. 6.2. Calculate C, and Cm0 25c for a NACA 0009 airfoil that has a plain flap whose length is O.2c and which is deflected 25°. When the geometric angle of attack is 4°, what is the section lift coefficient? Where is the center of pressure? 6.3. The mean camber line of an airfoil is formed by a segment of a circular arc (having a constant radius of curvature). The maximum mean camber (which occurs at midchord) is equal to kc, where k is a constant and c is a chord length. Develop an expression for the y distribution in terms of the freestream velocity and the angle of attack a. Since kc is small, you can neglect the higherorder terms in kc in order to simplify the mathematics. What is the angle of attack for zero lift (ao,) for this airfoil section? What is the section moment coefficient about the aerodynamic center 6.4. The numbering system for wing sections of the NACA fivedigit series is based on a combination of theoretical aerodynamic characteristics and geometric characteristics. The first integer indicates the amount of camber in terms of the relative magnitude of the design lift coefficient; the design lift coefficient in tenths is three halves of the first integer. The second and third integers together indicate the distance from the leading edge to the location
ProHems
315
of the maximum camber; this distance in percent of the chord is onehalf the number rep
resented by these integers. The last two integers indicate the section thickness in percent of the chord. The NACA 23012 wing section thus has a design lift coefficient of 0.3, has its
maximum camber at 15% of the chord, and has a maximum thickness of 0.12c. The equation for the mean camber line is
= 2.6595[(±)
—
+
0.11471(i)]
for the region 0.Oc x 0.2025c and
C
= 0.022083(1
\.
—
C
for the region O.2025c x 1.000c. Calculate the A0, A1, and A2 for this airfoil section. What is the section lift coefficient, C1? What is the angle of attack for zero lift, a01? What angle of attack is required to develop the design lift coefficient of 0.3? Calculate the section moment coefficient about the theoretical aerodynamic center. Compare your theoretical values with the experimental values in Fig. P6.4 that are reproduced from the work of Abbott and von Doenhoff (1949). When the geometric angle of attack is 3°,what is the section lift coefficient? What is the x/c iocation of the center of pressure?
NACA 23012 Airfoil Section Data from Abbott and von Doenhoff (1949) 3.0 X 106 o 6.0 x 106 o 8.8 X 106
0.0
—0.2
0.0 0.2
2.4
0.4
x
C
0.6
0.8
1.0
0.024
I
1.6
C1
0.016
C(j
°—
0
0.8
Cm,4 0.0
—0.1
0.008
ii
0.0
.000
B
a
a
—0.2 —0.8 —0.3
00
8 00
0.2
—0.4 —1.6
—0.4
—16
—32
16
32
Section angle of attack, a
0.8 —0.8 0.0 Section lift coefficient, C1
1.6
Figure P6.4
Sci. Aircraft 17:890—897 Cummings RM. New York: Dover Brune GW. Spaid FW. Reno. multielement airfoil at high Reynolds numbers. XFOIL:An analysis and design system for low Reynolds number airfoils. Mem. 42nd AIAA Pap. 2003—2902. Hilbig R. Development of the Black Widow micro air vehicle. Ed. 2001—0127. Dayton. MOrton SA. Look at the three airfoil geometries shown in Fig. 1996. Aerodyn. Sci. 6. Washington. Meet. 80210 Eppler R. Conf. New York: SpringerVerlag Flaig A. Forsythe JR. AIAA Pap. Discuss the geometric modifications to the laminar flow airfoils that make them distinct from the typical airfoil (NACA 23012). Huntsville. Why were these modifications successful in creating a laminar flow airfoil? 6. 2001. Keennon MT. Sci. NV Drela M. Data. 2004—1233. Henne PA.5. Presented at AIAA Aerosp. l9th. 96—0057. 6 / Incompressible Flows Around Airfoils of infinite Span 6. 1939. J. In ComputationalAerodynamics. Reno. location of maximum thickness. Mack RJ. AIAA Pap.30)? What are the advantages of these changes to aircraft design? REFERENCES Abbott IH. 1989. Design of low Reynolds number airfoils with trips. Include in your description airfoil geometric parameters such as camber. New York: SpringerVerlag Drela M. 1980. 1949.6. Somers DM..316 Chap.. Presented at AIAA Aerosp. NV Jacobs EN. High lift research program for a fightertype.. NACA WR L—345 . Epstein AR 2003. In Low Reynolds Number Aerodynamics. 2004. Ely WL. 2001—2463. Detachededdy simulation of slat and flap aerodynamics for a highlift wing. Ed. AIAA Pap. 78—156. 39th AIAA Pap. 2001. Presented at Infi.. Computational aerodynamics applied to highlift systems. 1980. What has enabled the evolution of commercial aircraft highlift systems from tripleslotted to double. What is a laminar separation bubble? What impact does it have on airfoil aerodynamics? What airfoil design features could be changed to eliminate (or largely reduce) the separation bubble? 6. NASA Tech.McMasters JH. Meet. AGARD CP 515 Gopalaratimam A. Selig MS. A solution to the 2D separated wake modeling problem and its use to predict CLmax of arbitrary airfoil sections. Meet. Studies of leadingedge thrust phenomena. Theory of Wing Sections. Sci. 1993. OH Eppler R.AIAA Pap. Presented at AIAA Aerosp. McGranahan BD. AL Hobbs CR.12.. 1990. von Doenhoff AE. Anaheim. Goodman WL. Preliminary report on laminarflow airfoils and new methods adopted for airfoil and boundarylayer investigation. Protz JM. In HighLift System Aerodynamics. Mueller TJ. leading edge radius. Meet. Air and Space Symp. DC: AIAA Carison HW. Broughton BA. NV Henderson ML. Presented at Appi. and trailing edge shape.. Presented at AIAA Aerosp.7. 6. Reno. CA Grassmeyer JM.A computer program for the design and analysis of lowspeed airfoils. 1990.or even singleslotted geometries (see Fig. 1978. location of maximum camber. thickness. 16th. Highlift design for large civil aircraft. The role of size in the future of aeronautics. Airfoil Design arid.
. Highlift aerodynamics. Cao HV. J.References Jasper 317 DW. Jn Fixed and Flapping Wing Aerodynamics for Micro Air Vehicle Applications. Mueller TJ. CA Loftin LK. Maughmer MD. Goradia SH. An overview of the Boeing 777 high lift aerodynamic design. NASA CR 1843 . 78—1223. AIAA Pap. NASA Tech. 1995. Sci. Evaluation of turbulence models for high lift military airfoil flowfields. NV Liebeck RH.Viscous phenomena affecting highlift systems and suggestions for future CFD Ddevelopment. 30:1162—1170 Selig MS. Inverse design of adapative airfoils with aircraft performance consideration. Pap. Computation of viscous flow for a Boeing 777 aircraft in landing configuration. 96—0057. Agrawal S. Robinson BA. Beasley WD. AIAA Pap. 1993. Tech. Sci. Aircraft 21:355—369 Nield BN. 1967.. Conf. AGARD CF 515 Jepson JK. Nash SM. 1989. Sci. Prediction of high lift: review of present CFD capability. Aerosp. Donovan JF. J. VKI Lecture Series Smith AMO. New York: Academic Press McGhee RJ. On the design of subsonic airfoils for high lift. 2004. Slotnick JP. 1992.Ying SX. 1975. Conf. 1996. Reno. In Low Reynolds Number Aerodynamics of Aircraft. Progr. Ed. Fraser DB. Quest for performance: the evolution of modern aircraft. 1971. 2001. NavierStokes calculations on multielement airfoils using a chimerabased solver. Meet. Theoretical and experimental study of the drag of multieleinent airfoils. VA: HA Stokely Selig MS. 1993. pp. Seattle. 1 1th. 143—167 Selig MS. Virginia Beach. AIAA J. Low Reynolds number airfoil design. Aircraft 12:501—530 Stevens WA. Roth K. 1978. Presented at AIAA Aerosp. Mathematical model for twodimensional multicomponent airfoils in viscous flow. Gopalaratham A. Aerodynamics of V/S TO L Flight. In HighLift System Aerodynamics. 1985. J. Presented at AIAA Aerosp. 34th. 76—406. In HighLift System Aerodynamics.Aerosp. Reno. The Aeronaut. 1976. Systematic airfoil design studies at low Reynolds number. James WD. Presented at Fluid and Plasma Dyn. A multipoint inverse airfoil design method based on conformal mapping. Sci. Baker D. New York: AIAA. NV Kern S. Whitlock M. Aircraft 10:610—617 Liebeck RH. Presented at Fluid and Plasma Dyn. A class of airfoils designed for high lift in incompressible flows. AIAA Pap. Aircraft 38: 1060—1068 Rumsey CL. McGowan PR. 7:107—119 Rogers SE. San Diego. Application of computational aerodynamics to airplane design. 99:361—371 Olson LE. NASA SP468 McCormick BW. Braden JA. Gopalarathnam A. 2002. Lowspeed aerodynamic characteristics of a 17percentthick section designed for general aviation applications. AGARD CF 515 Miranda LR. J. 1984. J. 2001. 2004—0028. 2003.Aerodynamic design of the highlift wing for a megaliner aircraft. Meet. Note D7428 Meredith PT. 2003. 38:145—180 Selig MS.. Lyon CA. WA Reckzeh D. 1973.. Giguére P. Airfoils at Low Speeds. 1973.
AGARD CP515 . Inflight pressure distributions and skinfriction measurements on a subsonic transport highlift wing section. 38:101444 Yip LP. J. Sd. Lowaspect ratio wing aerodynamics at low Reynolds number. Mueller Ti 2004. Fluid Mech. 42:865—873 Van Dam CE 2002. The prediction of separation of the turbulent boundary layer. In HighLift SystemAerodynamics. Progr.The aerodynamic design of multielement highlift systems for transport airplanes. 5:116 Torres GE. van Dam CR 1993. Aerosp.Vijgen PMHW. AIAA I.318 Chap. Hardin JD. 1959. 6 / Incompressible Flows Around Airfoils of Infinite Span Stratford BS.
there is a chordwise variation in the pressure differential between the lower surface and the upper surface. A procedure that can be used to determine the sectional lift coefficient has been discussed in Chapter 6. Similarly. 7. the flow is two dimensional). As indicated in the sketch of Fig. and therefore the circulation (integrated along the chord length of the section).1 GENERAL COMMENTS The aerodynamic characteristics for subsonic flow about an unswept airfoil have been discussed in Chapters 5 and 6. having both 319 .e.1. the highpressure air beneath the wing spills out around the wing tips toward the lowpressure regions above the wing.la. on the lower surface.lb.. The resultant flow around a wing of finite span is three dimensional. 7. Since the span of an airfoil is infinite. The lift produced by the pressure differences between the lower surface and the upper surface of the airfoil section. As indicated in Fig. 7..7 INCOMPRESSIBLE FLOW ABOUT WINGS OF FINITE SPAN 7. the flow is identical for each spanwise station (i. As a result of the spanwise pressure variation. the lift force per unit span decreases toward the tips. air will tend to flow outward toward the wing tips. The resultant lift force acting on a section (i. there is a spanwise variation in the lift force. Foi a wing of finite span. A sketch of a representative aerodynamic load distribution is presented in Fig. the air on the upper surface flows inboard toward the root. As a consequence of the tendency of the pressures acting on the top surface near the tip of the wing to equalize with those on the bottom surface. a unit span) is obtained by integrating the pressure distribution over the chord length. does not vary along the span.e.
7. In many instances. distributed along the span. The formation of a tip vortex is illustrated in the sketch of Fig.3). and (2) that the trailing vortex sheet remains flat as it extends downstream from the wing. Condensation clearly defines the tip vortices (just inboard of the wing tips) of the Shuttle Orbiter Columbia on approach to a landing at Edwards Air Force Base (see Fig.2d).These small vortices roll up into two large vortices just inboard of the wing tips (see Fig. may be replaced by an infinitesimally thin surface of discontinuity. which is of finite thickness. (b) spanwise lift distribution. Where the flows from the upper surface and the lower surface join at the trailing edge. Air Force Academy's Smoke Tunnel (Fig. chordwise and spanwise velocity components.2).2c and in the filaments of smoke in the photograph taken in the U. 7. water vapor condenses as the air is drawn into the lowpressure flow field of the tip vortices. 7 I Incompressible Flow about Wings of Finite Span Chordwise pressure distribution (differential between lower and upper surface) Relative airflow (a) lift at a section (b) Figure 71 Aerodynamic load distribution for a rectangular wing in subsonic airstream: differential pressure distribution along the chord for several spanwise stations.S. 7. Spreiter and Sacks (1951) note that "it has been . the difference in spanwise velocity components will cause the air to roll up into a number of streamwise vortices.320 Chap. it is customary to assume (1) that the vortex wake. 7. designated the trailing vortex sheet. At this point.Very high velocities and low pressures exist at the core of the wingtip vortices.
Sec. (d) smokeflow pattern showing tip vortex. (Photograph courtesy U.) . (c) formation of the tip vortex. (b) view from trailing edge. 7. Air Force Academy.1 I General Comments Vortex System 321 Relative airflow Leading edge Upper surface flow (inboard) Lower surface flow (outboard) Trailing edge (a) Relatively low pressure on the upper surface ( A— + +++++++ Relatively high pressure on the lower surface (b) ) Tip vortex (c) Figure 72 Generation of the trailing vortices due to the spanwise load distribution: (a) view from bottom.S.
" Thus. relate to changes in lift that might occur at some time. A suitable distribution of vortices would represent the physical wing in every way except that of thickness. 7.2 VORTEX SYSTEM A solution is sought for the vortex system which would impart to the surrounding air a motion similar to that produced by a lifting wing. . The trailing vortex system 3. such that the circulation at the wing tip is zero. an important difference in the threedimensional flow field around a wing variation (as compared with the twodimensional flow around an airfoil) is the in lift. the "starting" vortex is associated with a change in circulation and would.3 Condensation marks the wingtip vortices of the Space Shuttle Orbiter Columbia. there is a corresponding spanwise variation in circulation. The "starting" vortex As stated in Chapter 6.322 Chap.) firmly established that these assumptions are sufficiently valid for the prediction of the forces and moments on finitespan wings. The bound vortex system 2. 7 / Incompressible Flow about Wings of Finite Span Figure 7. The vortex system consists of 1. Procedures that can be used to determine the vortexstrength distribution produced by the flow field around a threedimensional lifting wing are presented in this chapter. Since the lift force acting on the wing section at a given spanwise location is related to the strength of the circulation. (Courtesy NASA. therefore.
we are interested in developing a model that can be used to estimate the aerodynamic characteristics of a wing.1) Orloff (1980) showed that the spanwise lift distribution could be obtained from flow field velocity surveys made behind an airfoil section of the wing only and related to the circulation around a loop containing that airfoil section. At any spanwise location y. The trailing vortex system 7.2). 7. is But vortex filaments cannot end in the fluid. However. The strength of the trailing vortex at any y location is equal to the change in the strength of the boundvortex system. or greater. as discussed in Chapter 3. The strength of the vortex filaments continuing .22. the sum of the strengths of all of the vortex filaments in the bundle at that station is F(y). 1(y) = (7.The velocity surveys employed the integral form of the momentum equation in a manner similar to that used to estimate the drag in Problems 2. 7. is similar to that depicted in Fig. in the x direction).. which is unswept (or is only slightly swept) and which has an aspect ratio of 4.Sec. provided that the spanwise flow is not too great. However.3 LIFTINGUNE THEORY FOR UNSWEPT WINGS For this section. Thus. Thus. The physical load distribution for the wing (which depends on the wing geometry and on the aerodynamic characteristics of the wing sections) 2..0.e.15.18 through 2. we shall place the boundvortex system at the quarterchord line. as shown in Fig.. the change represented in our model by having some of the filaments from our bundle of filaments turn 90° and continue in the streamwise direction (i. the axis of which is normal to the plane of symmetry and which passes through the aerodynamic center of the lifting surface. We will assume that the lift acting on an incremental spanwise element of the wing is related to the local circulation through the KuttaJoukowski theorem (see Section 3. The spanwise variation in lift. the vortex theorems of Helrnholtz state that a vortex filament cannot end in a fluid.e. each section of the finitespan wing generates a section lift equivalent to that acting on a similar section of an infinitespan wing having the same section circulation.e. 1(y). Prandtl and Tietjens (1957) hypothesized that each airfoil section of the wing acts as though it is an isolated twodimensional section. When the lift changes at some spanwise location [i. the idea allows a relation to be established between 1. The strength of the boundvortex system at any spanwise location I(y) is proportional to the local lift acting at that location 1(y). Since the theoretical relations developed in Chapter 6 for inviscid flow past a thin airfoil showed that the aerodynamic center is at the quarter chord. 7.e. which represents the spanwise loading distribution.3 I LiftingLine Theory for Unswept Wings 323 The representation of the wing by a bound vortex system is not to be interpreted as a rigorous flow model.That is.Thus. we shall model the lifting character of the wing by a large number of vortex filaments (i.ib. This is the boundvortex system.. we shall represent the spanwise lift distribution by a system of vortex fil aments.4(a). a large bundle of infinitesimalstrength filaments) that lie along the quarter chord of the wing. Therefore. the total strength of the boundvortex system changes proportionally [i.
The flow around wing/strake configurations will be discussed further in Section 7. Thus.4a.The two strearnwise vortices associated with the flow around the edges of the strakes are easily seen on either side of the fuselage.4a.Thus. which is termed a horseshoe vortex. if the strength of the vortex filaments in the bundle making up the boundvortex system change by the amount a trailing vortex of strength must be shed in the x direction. For steady flight conditions. so that the trailingvortex pair effectively stretches to infinity. "LeadingEdge Extensions. therefore. depends upon parameters such as the wing planform. as shown in Fig. Thus. 7.4b. which the vortex system represents. the geometric twist of the wing. The threesided vortex. the starting vortex is left far behind. 7. the airfoil sections that make up the wing. The streamwise condensation pattern that appears across the wing in Fig. streamwise vorticity filaments originating in the wingleadingedge region can be seen across the whole wing. but turn backward at each end to form a pair of vortices in the trailingvortex system. the vortex filaments that make up the boundvortex system do not end in the fluid when the lift changes. in the boundvortex system depends on the spanwise variation in lift and. the system consists of a the boundvortex system and the related system of trailing vortices. 7. 7. A number of vortices are made visible by the condensation of water vapor in the flow over an F/A18 Hornet in the photograph of Fig. Also included in Fig. etc.324 Chap.4b is not normally .4a is a sketch of a symmetrical lift distribution. for practical purposes.4 (a) Schematic trailingvortex system. 7." In addition. is presented in Fig.8. 7 I Incompressible Flow about Wings of Finite Span r(y) r y axis axis vortex system which represents the spanwise loading distribution Trailing vortex of strength (parallel to the freestream) x Figure 7.
.
0. and an aspect ratio of 4.326 Chap. there will be reasonable agreement between the calculations and the experimental values for a single lifting surface having no sweep.1 Trailing Vortices and Downwash A consequence of the vortex theorems of Helmholtz is that a boundvortex system does not change strength between two sections unless a vortex filament equal in strength to the change joins or leaves the vortex bundle. Conventional Prandtl liftingline theory (PLLT) provides reasonable estimates of the lift and of the induced drag until boundarylayer effects become important. 73.g.5.4a. To calculate the influence of a trailing vortex filament located at y. or greater. It is believed that these streamwise vorticity filaments correspond to the trail ing vortices shed by the spanwise variation in vorticity across the wing that is depicted in the schematic of Fig. Corda. no dihedral. Rasmussen and Smith (1999) and and Snyder (2000)] and in ways of accounting for the nonlinear behavior of the aerodynamic coefficients [e.g. Of course. Anderson. .4) x axis z axis y= —s Figure 7. a semiinfinite vortex of strength from the segment as shown in Fig. 7. Thus. the liftingline theory is still widely used today. yaxisA /aty / I Velocity induced at y1 by the trailing vortex yaxisf y= y=y1  Semiinfinite trailing vortex + I of strength dF 1 dF is negative in this region for the uy rdistribution of Fig 7.. Because improvements continue to be made in calculation procedures [e.. If denotes the strength of the circulatrails tion along they axis (the spanwise coordinate). 7 / Incompressible Flow about Wings of Finite Span observed. operating at relatively low angles of attack. the skin friction component of drag will not be represented in the PLLT calculations at any angle of attack.5 Geometry for the calculation of the induced velocity aty=y1.The strength of the trailing vortex is given by It is assumed that each spanwise strip of the wing (ay) behaves as if the flow were locally two dimensional. 7. and Van Wie (1980)].
7.The vortex at y induces a velocity at a general point y1 on the aerodynamic centerline which is onehalf the velocity that would be induced by an iiifiiiitely long vortex filament of the same strength: = ir + [ a.— Chord line of the airfoil — — — — . in a downward direction (i. which is in the positive z direction. acts normal to the effective flow direction Chord line Undisturbed freestream direction (direction of U. 7..e. negative) and is called the downwash. parallel to the x axis (which is parallel to the free streám.) The resultant velocity for airfoil section — ..dy 1 — Yi) The positive sign results because. as shown in Fig.6 Induced flow.1\ — (7. To calculate the resultant induced velocity at any point y1 due to the cumulative effect of all the trailing vortices.3 / LiftingLine Theory fOr Unswept Wings 327 consider the semiinfinite vortex line.flow) and extending downstream to infinity from the line through the aerodynamic center of the wing (i. they axis).. when both (y — Yi) and dT'/dy are negative.. in general.e.2) The resultant induced velocity at y1 is. .Effective flow direction Undisturbed freestream direction Figure 7.Sec.3) induced drag Lift Effective lift.6. 7. As shown in the sketch of Fig. the downwash angle is = / ( w. the preceding expression is integrated with respect to y from the left wing tip (—s) to the right wing tip (+s): = +— I 1 YYi dy dF/dy (7.5. the trailing Vortex at y induces an upward component of velocity.
7).e. This drag force is a consequence of the lift developed by a finite wing and is termed vortex drag (or the induced drag or the dragduetolift). Since the direction of the resultant velocity at the aerodynamic center is inclined downward relative to the direction of the undisturbed freestream air. for subsonic flow past a finitespan wing. 7 I incompressible Flow about Wings of Finite Span The downwash has the effect of "tilting" the undisturbed air. is less than that for an infinitespan airfoil composed of the same airfoil section and which is at the same angle of attack a. Thus. the threedimensional flow over a finitespan wing generates less lift than the twodimensional flow over an infinitespan airfoil. As a result of the induced downwash velocity.e. is given by the elliptic relation (see Fig. the total lift is given by +s L and the total vortex drag is given by L —f dy (7. both the angle of attack (a) and the downwash angle (s) would be a function of the spanwise position. As a consequence of the trailing vortex system. 5.328 Chap. 7.25.4) Note that. in addition to the skin friction drag and the form (or pressure) drag. 7. a. it is independent of y) and the induced downwash velocity is zero at all points. Integrating over the entire span of the wing. a wing of infinite span) the circulation strength I' is constant across the span (i.3. as discussed in Chapter 6. so the effective angle of attack at the aerodynamic center (i.1) 1(y) = while the vortex drag is (7. the effective lift on the wing has a component of force parallel to the undisturbed freestream air (refer to Fig. the lift on an elemental airfoil section of the wing is (7. = 0. Thus. the quarter chord) is (7.5) The minus sign results because a downward (or negative) value of w produces a positive drag force.6) (7.2 Case of Elliptic Spanwise Circulation Distribution An especially simple circulation distribution.7) Note that for the twodimensional airfoil (i. there is a drag component duetolift.. which is at the geometric angle of attack.e. F(y) =  (Y)2 (7. at a given a.. Thus. if the wing has a geometric twist.8) . the aerodynamic characteristics are modified significantly from those of a twodimensional airfoil of the same section. the effective lift of the section of interest is inclined aft by the same amount. See Fig.6).. Based on the KuttaJoukowski theorem. the lift generated by a finitespan wing composed of a given airfoil section. Thus. which also has significant practical implications. 7.
Since the lift is a function only of the freestream density. only when the wing has a rectangular planform (and c is.3 I LiftingLine Theory for Unswept Wings 329 r(y) = V IH Downwash velocity (wy) = — (a constant) Figure 7. to calculate the section lift coefficient. an elliptic distribution for the circulation would produce an elliptic distribution for the lift.7 Ellipticcirculation distribution and the resultant downwash velocity.8). the section lift force is divided by the product of and the local chord length at the section of interest. Hence. the freestream velocity. therefore. 7. and the circulation. constant) is the spanwise section lift coefficient distribution (C1) elliptic when the spanwise lift distribution is elliptic.9) Integration yields = —(IT + yi') F0 . For the elliptic spanwise circulation distribution of equation (7. the induced downwash velocity is WY1 = dy — Y2(Y — Yi) which can be rewritten as r — [ Lis (yy1)dy Vs2 y2(y — + Yl) y1dy Is Vs2 — y2(Y — Yi) ] (7.Sec. However.
y = 0). Thus. f L= pooUcj0\/1 — cos2 s sin = (7.9). the velocity induced at a point Yi = +a should be equal to the velocity at a point = —a. we can calculate the total vortex (or induced) drag for the wing._ 330 Chap. Referring to equation (7. this can be true if I = 0.12) Similarly. for the elliptic load distribution Wyl = w(y) = (7. = Introducing the coordinate transformation again.e.10) The induced velocity is independent of the spanwise coordinate..11) The lift coefficient for the wing is L — ithF0 c = ')TT I (7.13) and the drag coefficient for the induced component is I. 7 / Incompressible Flow about Wings of Finite Span where j I p+5 dy Js vs—y(y—y1) Since the elliptic loading is symmetric about the pitch plane of the vehicle (i. — 1 r'2 ITI o TT2 ArT2 . The total lift for the wing is L= Using the coordinate transformation y= —scos4 the equation for lift becomes L Thus. we have r2 pIT = = J VI — cos2 4 s sin dçb (7.
2..9.1 has been assumed to be equal to CDO.17) where is the drag coefficient at zero lift and is the liftdependent drag coefficient.12) to solve for gives (7. — (2CLUo0S\2 I 'zrb or Since the aspect ratio is defined as b2 = C2 L irAR (7.18) where CDO. AR2 = 5). 7. If one compares the drag polars for two wings which have aspect ratios of AR1 and AR2. Note also that the trailing vortex drag for an inviscid flow around a wing is not zero but is proportional to The induced drag coefficient given by equation (7.8. which were presented by Schlichting and Truckenbrodt (1969). Equation (7. then for a given value of the lift coefficient. The data from Prandtl (1921) for a series of rectangular wings are reproduced in Fig. These converted drag polars. closely follow the theoretical values up to an angle of attack of 20°.e. The experimentally determined drag polars are presented in Fig. The liftdependent drag coefficient includes that part of the viscous drag and of the form drag..e. which results as the angle of attack changes from These relations describing the influence of the aspect ratio on the lift and the drag have been verified experimentally by Prandtl and Betz.16) Note that once again we see that the induced drag is zero for a twodimensional airfoil (i. 7.3 / LiftingLine Theory for Unswept Wkigs 331 Rearranging equation (7. which . The relatively constant difference between the measuredvalues and the theoretical values is due to the influence of skin friction. as noted in Chapter 5.18) has been used to convert the drag polars for the different aspect ratio wings to the equivalent drag polar for a wing whose aspect ratio is 5 (i. which was not included in the development of equation (7. = CD.16) and the measurements for a wing whose aspect ratio is 5 are compared in Fig. Therefore.15) Thus.16). the drag coefficient for an incompressible flow is typically written as CD = CDO + (7.1 + — AR1) (7. 7. respectively.9a. a wing with an aspect ratio of infinity). 7.Sec. The experimental values of the induced drag coefficient.
we obtain CL ir'AR (719) One can determine the effect of the aspect ratio on the correlation between the lift coefficient and the geometric angle of attack.10)]. we use the equation a2 = ai + — AR1) (7. the downwash velocity for the elliptic load distribution [equation (7. are presented in Fig.lOa are for the same rectangular wings of Fig.0 —0.15)].08 CD 0.00 0.4 0. The data presented in Fig. collapse quite well to a single curve. A similar analysis can be used to examine the effect of aspect ratio on the lift. and the correlation between the lift coefficient and [equation (7. Combining the definition for the downwash angle [equation (7.16) = 2.9.7 X 106. 7.18).332 Chap.9b.2 0.20) Experimentally determined lift coefficients [from Prandtl (1921)] are presented in Fig. data of Schlichting and Truckenbrodt (1969) 1. Thus.16 0.To calculate the geometric angle of attack a2 required to generate a particular lift coefficient for a wing of AR2.3)]. 7. .24 Figure 7. 7.10. if a wing with an aspect ratio of AR1 generates the same lift coefficient at a1. the correlation of the measurements confirms the validity of equation (7. (7.4 0. eq. 7.8 CL 0.8 Experimental drag polar for a wing with an aspect ratio of 5 compared with the theoretical induced drag. 7 I Incompressible Flow about Wings of Finite Span —Theoretical induced drag.
1Db.20) is experimentally verified. the validity of equation (7.Sec. the converted curves collapse into a single correlation.00 0.8  CL 0.20) in terms of a wing whose aspect ratio is 5 (i. Therefore. The results of converting the coefficientoflift measurements using equation (7. 7. 7.24 (a) Figure 7. Again.3 / LiftingLine Theory for Unswept Wings Data from Prandtl (1921) 333 a AR: 1 V 2 3 4 5 6 7 1..3 Technique for General Spanwise Circulation Distribution Consider a spanwise circulation distribution that can be represented by a Fourier sine series consisting of N terms: N F(4) .4  —0. 7.9 Effect of the aspect ratio on the drag polar for rectangular wings (AR from 1 to 7): (a) measured drag polars.4 — 0.3. AR2 = 5) are presented in Fig.e.2 0.08 CD ft16 0.
7.8 CL 0.11.11 is symmetrical. 7. eq.08 CD 0. the physical spanwise coordinate (y) has been replaced by the 4) coordinate: —=—cos4 A sketch of one such Fourier series is presented in Fig.16 0.16) Data from Prandtl (1921) 0 A • 4 0 5 AR: 1 2 3 0 6 V 7 1.9 (continued) (b) drag polars converted to AR = 5.2 0. 7 / Incompressible Flow about Wings of Finite Span — Theoretical induced drag.334 Chap.24 (b) Figure 7.00 0.0 —0. Since the spanwise lift distribution represented by the circulation of Fig. only the odd terms remain. As was done previously. y .4 0. (7.4 0.
A3. sin (7. . it is necessary to determine the circulation at N spanwise locations. .3 / LiftingLine Theory for Unswept Wings Data from Prandtl (1921) 335 AR: ol o2 I . 7. Once this is done.Sec. the Nresultant linear equations can be solved for the coefficients. the series is truncated to a finite series and the coefficients in the finite series are evaluated by requiring the liftingline equation to be satisfied at a number of spanwise locations equal to the number of terms in the series.4 — —12 I I I I I I —8 —4 0 4 8 12 16 20 a. will be developed in this section. known as the collocation method.2 0.22) To evaluate the coefficients A1.e. A2. .This method.10 Effect of aspect ratio on the lift coefficient for rectangular wings (AR from 1 to 7): (a) measured lift coefficients. Typically.4 0.0 —0. deg (a) Figure 7. Recall that the section lift coefficient is defined as C1(4)— — lift per unit span 1 2 .4 05 o6 v7 I 1. The section lift force [i.8 CL 0.. the lift acting on that spanwise section for which the circulation is F(4)] is given by = .AN.
we obtain = C1(çb) (7. Thus. 7.12 for the nomenclature.4 o5 o6 v7 1.24) to form equation (7. Using the local circulation to determine the local lift per unit span.4 V 0 0 0. 7 / Incompressible Flow about Wings of Finite Span Data from Prandtl (1921) AR: ol I o2 I .336 Chap. .8 V5 d CL 0.2  .0 OX S 0 .23) equal to that in equation (7.25).0— 0 0 ——0. 0. . we have = a01) (7.10 (continued) (b) Lift correlations converted to AR = 5.4 —12 —8 —4 0 4 8 12 16  20 (b) Figure 7.24) We now have two expressions for calculating the section lift coefficient at a particular spanwise location We will set the expression in equation (7.23) It is also possible to evaluate the section lift coefficient by using the linear correlation between the lift and the angle of attack for the equivalent twodimensional flow. referring to Fig.
the zero lift angle of attack. Note that since ae = cx — equations (7.23) and (7. 5. the downwash angle.v — Yi dy Using the Fourier series representation for F and the coordinate transformation. five parameters in equation (7. y) at which we will evaluate the terms. Note that 1 dF/dy .25) may depend on the spanwise location 4) (or.11 Symmetric spanwise lift distribution as represented Let the equivalent liftcurve slope (dC. which is illustrated in Fig. which varies with 4) when the wing is twisted (i..3 / LiftingLine Theory for Unswept Wings 337 F = sin n (odd terms only) IT IT y—s by a sine series. we obtain sin nçb sm4) . which varies with 4) for a tapered wing planform.the chord length. .24) can be combined to yield the relation 2F(4)) / * aoj(4))] — U00c(4)) (7. (4) a. the local geometric angle of attack. and (5) a01. 0 +s Figure 7. which depends on the circulation distribution. the local circulation. which varies with 4) when the airfoil section varies in the spanwise direction (which is known as aerodynamic twist). equivalently.7).25) For the present analysis.e. (3) c. The five parameters are (1) F. (2) e.Sec. geometric twist../da)o be designated by the symbol a0. 7.
If we consider only symmetrical loading distributions.12 Nomenclature for wing/airfoil lift. the equation becomes 8s . 7. as shown in the sketch of Fig. the resultant governing equation is a01) sin A. a0 = Cia Cl Incidence Equivalent twodimensional freestream Remote freestream Figure 7. Equation (7. A. only the odd terms of the series need be considered. 7 / Incompressible Flow about Wings of Finite Span Twodimensional lift slope. sin ncb(/Ln + sin (7..338 Chap. sin n4 sin Defining p. = cao/8s. = + A3 sin 34) + A5 sin 54) + .26) which is known as the monoplane equation. That is.25) can be rewritten as 21T ca0 sin n4'.11..
4 Lift on the Wing L= tIT J —s J 0 Using the Fourier series for F(4) gives us L= A sin B = cos(A — f B) — cos(A + B). the integration yields L= + +  The summation represented by the second term on the righthand side of the equation is zero.3.3 / LiftingLine Theory for Unswept Wings 339 7. the coefficient for the vortexinduced drag = 7nAR (7.Sec. the integral expression for the lift becomes L= CL = (7.28) . 7.27) = A1ir AR The lift depends only on the magnitude of the first coefficient. 7.5 VortexInduced Drag f+S Dv_j pcyjwPdy —S sin nq5 PooJ F dy = The integral f f sin sin nç& dçb Thus. Thus. no matter how many terms may be present in the series describing the distribution.3. since each of the terms is zero for n 1.
9. The wing. Since the wing is untwisted. airfoil section NACA 65—210. which is unswept at the quarter chord. the drag is minimum when 6 0. 7.13. The geometry of the wing to be studied is illustrated in Fig. the zerolift angle of attack (a01) is approximately —1. Referring to the data of Abbott and von Doenhoff (1949).. The taper ratio A (i. The aspect ratio (AR) is 9.340 Chap. Ct/Cr) is 0. 7 / Incompressible Flow about Wings of Finite Span Since A1 CL/(ITAR).720 1 (0290 2381 ft 7. AR = A = 0.40. the only term in the series representing the circulation distribution is the first term: Since = which is the elliptic distribution.2° across the span.)b 0.00. equation (7. the geometric angle of attack is the same at all spanwise positions. S= 0..286 m) Figure 7. CDV = where only the odd terms in the series are considered for the symmetric load distribution. is composed of NACA 65—210 airfoil sections.ARL = C2 or •AR (1 + 6) (7. In this case.00. .5(Cr + c.e. CDV= [ 7r.e.29) where 42 P11 42 111 42 6 0.500 ft (2.1: Use the mooplane equation to compute the aerodynamic coefficients for a wing The monoplane equation [i. EXAMPLE 7. Since the wing planform is trapezoidal.26)] will be used to compute the aerodynamic coefficients of a wing for which aerodynamic data are available.5Cr(1 + A)b ALE = 2.13 Planform for an unswept wing.40.
00386 = 0.24933(1 — + (A — 1)cos4)] (7.92388 0.70711 —0. Equation (7. 7.26) is — = + sin4)) + + sin4)) + sin 4)) (731) + A5 sin 54)(Sji + sin 4)) + A7 sin Since there are four coefficients (i. A1.0° 0.92388 1.31) y S Station 1 (/1 (= cos4)) 0. equation (7. = 2(1 = 0.44411A7 TABLE 7.1 Values of the Factor for Equation (7.66154A3 + 0. the 2b Cr + parameter in equation (7.00000 sin4) sin 34) sin 54) sin 74) 2 3 4 22.38268 — 1.e.1.19208 0.31) becomes 0. A5. since we are considering the left side of the wing.38268 —0.14355 0.24933 . Note that.3 / LiftingLine Theory for Unswept Wings and 341 AR= Thus.11112 0.e.Sec.86686A5 + 0.26) becomes Ca0 Ca0 = 4b = 2(AR)Cr(1 + A) Solution: Since the terms are to be evaluated at spanwise stations for which o ir/2 [i. Therefore.e.18897A1 + 0.00000 0.70711 0.5° 45..70711 0.. that for a twodimensional flow over the airfoil section a0) has been assumed to be equal to 2ir.00000 0.38268 0. a fourterm series will be used to represent the spanwise loading.31) must be evaluated at four spanwise locations.0° 67.30) 0.00000 0.92388 —0.70711 0. The resultant values for the factors are summarized in Table 7. It might be noted that numerical solutions for lift and the vortexdrag coefficients were essentially the same for this geometry whether the series representing the spanwise circulation distribution included four terms or ten terms. the y coordinate is negative. equation (7. For a geometric angle of attack of 40. A3.5° 90. —s < y 0 (which corresponds to the port wing or left side of the wing)J.38268 0.92388 0.92388 1.38268 1. so that the reader can perform the required calculations with a pocket calculator.00000 0.6cos4)) where the equivalent liftcurve slope (i.70711 —0.. and A7) to be evaluated.
7. the wing had a dihedral angle of 3°.) .2 Data of Sivells (1947) — 0 1.80451A3 — 1.14 with data for this wing.92388s). the lift coefficient for an angle of attack of 4° is CL A1Tr AR = 0. For the other stations.57407A3 — 0.6459 x A3 = 7.5787 x Using equation (7.00921 = 0.e.24933A1 1.21053A7 0.74799A3 + 2..0 0 —0.00752A5 — 1.0 0.2 —4 0 4 8 12 a.5° (i. 7. In addition to the geometric characteristics already described.4 0.2 0. (Wing is that of Fig. The measurements reported by Sivells (1947) were obtained at a Reynolds number of approximately — Theory 0 1.4 1.60150A1 + 0.24665A5 — 2. y = —0.27).13.3218 x A5 = 8.02263 = 1.03101A1 — 0.72109A5 + 2.74531A7 The solution of this system of linear equations yields A1 = 1.01611 = 1. the equation becomes 0.4654 The theoretically determined lift coefficients are compared in Fig.09577A7 0. deg Figure 714 Comparison of the theoretical and the experimental lift coefficients for an unswept wing in a subsonic stream. 7 / Incompressible Flow about Wings of Finite Span for = 22.342 Chap.8  0 CL 0.6 0.
8 1. is CDV= / = 0.8 0.2 0. The agreement between the theoretical values and the experimental values is very good.0136) = 0.0 0.2 0.6 0.4 0.0 y S Figure 7.4 0. as determined using equation (7.4. The local lift coefficient is given by Cl— which for the trapezoidal wing under consideration is 2(AR)(1 + sin(2n — (7.15 Spanwise distribution of the local lift coefficient. A = 0. 4.00776 .4 X 106 and a Mach number of approximately 0. The spanwise distribution for the local lift coefficient of this wing is presented in Fig.Sec. 7.17.00766(1.3 / LiftingLine Theory for Unswept Wings 343 1. AR = 9.6 CL 0. untwisted wing composed of NACA 65210 airfoil sections.29).0 0.15:As noted by Sivells (1947).32) The theoretical value of the induced drag coefficient for an angle of attack of 4°.0 0. 7. the variation of the section lift coefficIent can be used to determine the spanwise position of initial stall.
13. 7. (Wing is that of Fig. The relatively constant difference between the measured values and the theoretical values is due to the influence of skin friction. Again. the theoretical relations developed in this chapter do not include the effects of skin friction. .16 with the measured drag coefficients for this wing.17.The wings. all had an aspect ratio of 7. The effect of the taper ratio on the spanwise variation of the lift coefficient is iilustrated in Fig. C1 2(1+A)c = sin(2n — 1)4 The values of the local (or section) lift coefficient near the tip of the highly tapered wings are significantly greater than the overall lift coefficient for that planform. 7 / Incompressible Flow about Wings of Finite Span — Theory 0 Data of Sivells (1947) CL 0. Theoretical solutions are presented for untwisted wings having taper ratios from 0 to 1.04 CD Figure 7. this result is important relative to the separation (or stall) of the boundary layer for a particular planform when it is operating at a relatively high angle of attack.) The theoretically determined induced drag coefficients are compared in Fig. 7.16 Comparison of the theoretical induced drag coefficients and the measured drag coefficients for an unswept wing in a subsonic stream. As noted earlier. As has been noted earlier. the local lift coefficient has been divided by the overall lift coefficient for the wings. 7. which were composed of NACA 2412 airfoil sections.28.344 Chap. Thus.
0 0.6 0. all sections will reach stall at essentially the same angle of attack.4 0. the stall pattern is favorable.8 0.2 0.e. since the inboard portion of the wing carries more of the wing's lift than the tip.0 0.2 0. the local lift coefficient is roughly constant across the span).8 1.2 1. The spanwise load distribution for a rectangular wing indicates stall will begin at the root and proceed outward. The desirable stall pattern for a wing is a stall which begins at the root sections so that the ailerons remain effective at high angles of attack.3 I LiftingLine Theory for Unswept Wings 345 1.4 1.17 Effect of taper ratio on the spanwise variation of the lift coefficient for an untwisted wing. Tapering of the wing reduces the wingroot bending moments. Furthermore. 7.. As a result.0 Cl CL 0.18.Sec. 7. Thus.6 0. Sketches of stall patterns are presented in Fig.4 0.4) approximates that of an elliptical wing (i. The spanwise load distribution for a wing with a moderate taper ratio (A = 0. the longer wingroot chord makes it possible to increase the actual thickness of the wing while .0 y S Figure 7.
A = 0.The addition of leading edge slots or slats toward the tip increases the stall angle of attack and is useful in avoiding tip stall and the loss of aileron effectiveness. maintaining a low thickness ratio. which is needed if the airplane is to operate at high speeds also.0. A = 1. or slightly swept. (c) A = 0. (c) pointed wing. the lift coefficients near the tip are higher than those near the root fOr a tapered wing. Therefore. wings. there is a strong tendency to stall near (or at) the tip for the highly tapered (or pointed) wings. liftingline theory is still used to provide rapid estimates of the spanwise load distributions and certain aerodynamic coefficients for unswept.4. 7 / Incompressible Flow about Wings of Finite Span (a) (b) Figure 7. to decrease the local angles of attack at the tip (refer to Table 5.6 Some Final Comments on LiftingLine Theory With continuing improvements.1). the wing may be given a geometric twist. the planform and the twist distributions for general wing configurations are represented explicitly. 7. The spanwise circulation distributiOn F(y) is obtained explicitly in terms of the Fourier coefficients for the chord distribution and the twist distribution.3. . (b) moderately tapered wing. While taper reduces the actual loads carried outboard. In the Fourier series analysis of Rasmussen and Smith (1999).0.18 Typical stall patterns: (a) rec tangular wing. In order to prevent the stall pattern from beginning in the region of the ailerons. Or washout.346 Chap.
016 — 0 U 0. A = [Taken from Rasmussen and Smith (1999). when six. or eight terms are used in the Fourier series. This is shown . Thus. singleengine general aviation airplanes so as to inhibit the onset of stall/spins. such modifications are clearly of practical importance.] 0. Since more than 30% of all general aviation accidents are caused by stall/spins.1. except that many more panels are used to provide better resolution of the spanwise loading. Collocation method of Example 7.1 in this AR = 9 A = 0.013 I 3 4 5 6 7 8 9 Number of terms in series Figure 7.015 :1 // 0.4.2 (see Fig. The method of Rasmussen and Smith (1999) was used to solve for the aerodynamic coefficients for the wing of Example 7. The two methods produce values which are very close. A modification of current interest is an abrupt extension and change in shape of the leading edge along a portion of the wing span — a so called 'drooped' leading edge. seven. The values of 6. 7.3 I LiftingLine Theory for Unswept Wings 0. 7. Anderson.014 — // / 7 — . which were computed using the method of Rasmussen and Smith (1999).4 text 0. Incorporating empirical information into the modeling.19. one can extend the range of applicability of liftingline theory. the method is very much like that used in Example 7.31)." The liftingline theory of Phillips and Snyder (2000) is in reality the vortexlattice method applied using only a single lattice element in the chordwise direction for each spanwise subdivision of the wing. is reproduced in Fig. The induceddrag factor 8. are compared with the values computed using the collocation method of Example 7. as taken from Rasmussen and Smith (1999).Sec. and Van Wie (1980) noted that "certain leadingedge modifications can favorably tailor the highlift characteristics of wings for light. Corda. 7.19 Convergence properties of induceddrag factor 8 for tapered wing. "The method converges faster and is more accurate for the same level of truncation than collocation methods.1.017 347 0. The values of the induceddrag factor are presented as a function of the number of terms in the Fourier series. Rasmussen and Smith (1999) claim.
] schematically in Fig.20. and Van Wie (1980)] The reader is referred to the discussion leading up to equations (7.26). . threedimensional flow with separation. an airplane with a properly designed drooped leading edge has increased resistance toward stalls/spins. 7. As a result. The poststall behavior was modeled by introducing the experimentally determined values of the liftcurve slope in place of a0 [Anderson." As shown in Fig.20. 7. the chord is extended approximately 10% over a portion of the span. the chord and/or the leadingedge shape of the wing change discontinuously. Corda. and the generation of a relatively large value of CL at very high poststall angles of attack.20 Drooped leadingedge characteristic. CL a Figure 7.The authors noted that the greatest compromise in using liftingline theory into the stall angleofattack range and beyond is the use of data for the twodimensional flow around an airfoil. as also shown in Fig.E. Corda. and Van Wie (1980).20.348 Chap. where at a given spanwise location.25) and (7. The net aerodynamic effect of this modification is a smoothing of the normally abrupt drop in lift coefficient CL at stall. 7. 7 / Incompressible Flow about Wings of Finite Span Drooped leading edge Drooped leading edge Cr I I Standard L.The actual flow for this configuration is a complex. [Taken from Anderson.
A variety of methods have been developed to compute the flow about a thin wing which is operating at a small angle of attack so that the resultant flow may be assumed to be steady. when compared with experimental data. as shown in Fig.0834 AR = o ix 5.Sec.4 PANEL METHODS Although liftingline theory (i. and incompressible. The configuration is modeled by a large number of elementary quadrilateral panels lying either on the actual aircraft surface. Once the singularity strengths have been determined. linear.4 / Panel Methods 349 1. [Taken from Anderson. with the use of experimental values for the liftcurve slope of the airfoil section. These singularities are determined by specifying some functional variation across the panel (e..] Nevertheless. (deg) 30 35 40 45 Figure 7.8 0.).g. whose actual value is set by corresponding strength parameters. These strength parameters are determined by solving the appropriate boundary condition equations. The basic concept of panel methods is illustrated in Fig. the velocity field and the pressure field can be computed. and Van Wie (1980).21.314 Experimental data Numerical data 0 0 5 10 15 20 25 a.6 Wing 1.. thin wing of relatively high aspect ratio in a subsonic stream. inviscid. 7. an improved flow model is needed to calculate the lifting flow field about a highly swept wing or a delta wing. there is attached one or more types of singularity distributions. comparison between experiment and ical results. 7.22. or a combination thereof. or on some mean surface. liftingline theory generates reasonable estimates for CL. vortices. . etc.e.6 CL c14 0015 / 0014. To each elementary panel. and doublets. constant.0 0.21 Lift coefficient versus angle of attack for a drooped leadingedge wing. Corda. the monoplane equation) provides a reasonable estimate of the lift and of the induced drag for an unswept. 7. such as sources. 7. irrotational. quadratic.
7 I Incompressible Flow about Wings of Finfte Span Typical surface panel whose effect on the flow can be represented by Source or Doublet or vortex simulation of wake Control point for application of boundary condition Figure 722 Representation of an airplane flowfield by panel (or singularity) methods.The design problem in addition involves such aspects as stream surface lofting (i. and the relationship between a velocity field and its potential. Johnson (1980) noted further that Dirichlet boundary conditions (specification of q') arise in connection with the inverse problem (i. analysis conditions are of the Neumann type (specification of normal velocity). design conditions are fundamentally of the Dirichiet type (specification of potential).350 Chap. the pressure coefficient at each point on the surface of the impermeable boundary can computed using = 1 (733) The reader should note that the tangential velocity at the "surface" of a configuration in an inviscid flow is represented by the symbols U and in equation (3. The specification of cp guarantees a . Analysis conditions are employed on portions of the boundary where the geometry is considered fixed. Once a solution has been found to the boundaryvalue problem. integration of streamlines passing through a given curve). hence.13) and equation (7.1 Boundary Conditions Johnson (1980) noted that.e. the normal component of the resultant velocity must be zero at every point of the surface.. 7.4. as a general rule." Neumann boundary conditions (specification of at every point on the surface) arise naturally in the analysis of fixed configurations bounded by surfaces of known permeability. a boundaryvalue problem associated with LaPlace's equation. If the surface of the configuration is impermeable (as is the case for almost every application discussed in this text). The permeability of the fixed geometry is known. Here a perturbation to an existing tangential velocity vector field is made. Design boundary conditions are used wherever a geometry perturbation is allowed for the purpose of achieving a specific pressure distribution.e. is well posed if either q' or (acp/an) is specified at every point of the surface of the configuration which is being analyzed or designed.26).33). respectively. that of solving for a specified pressure distribution on the surface of the configuration). "Fluid flow boundary conditions associated with LaPlace's equation are generally of analysis or design type. see equation (3. and resultant pressures are desired. hence..
as related through equation (7. The five surface panel methods . A finite set of control points (equal in number to the number of singularity parameters) is selected at which the boulidary conditions are imposed. However. (1985) compare computed aerodynamic coefficients using one vortex lattice method (VLM). accuracy. In such cases. two loworder surface potential distributions. the velocities induced at that control point by the singularities associated with each of the panels of the configuration summed. a predetermined pressure coefficient distribution.To achieve both a specified pressure distribution and a normM flow distributiOn on the surface. and overall code robustness.g. Margason et al. The construction of each network requires developments in three areas: (1) the definition of the surface geometry. computational economy. 7. so that the surface will be a stream surface of the flow The total design problem thus composed Of two problems. the achievement of a desired pressure distribution on the surface is not physically signfficant without restrictions on the flux through the surface.4. and two highorder surface potential distributions.33). A surface doublet distribution can bè•ràplaced by an equivalent surface vortex distribution. For many applications. Numerous codes using panelmethod techniques have been developed [e. their numerical implementation may yield significantly different results from the point of view of numerical stability. (2) the definition of the singularity strengths. The choice of combinations is not a trivial matter.4 / Panel Methods 351 predetermined tangential velocity vector field and. the geOmetric layout of the elementary panels. The first is to find a perturbation potential for the surface that yields the desired distribution for the pressure coefficient and the second is to update the surface geonietry so that it is a stream surface of the resultant flow. The theoretical equivalency between vorticity distributions and doublet distributions does not imply e 4uivalent simplicity in a numerical For. highly loaded surfaces such as the aft portion of a supercritical airfoil. an iterative procedure is required for solution. 7. Johnson (1980) concluded. in general. source strengths with strong gradients can degrade the local velocity calculations. in general. therefore. and the type of boundary condition imposed. Bristow and Grose (1978) note that there is an important equivalence between surface doublet distributions and vorticity distributions. "The two problems are coupled and..Sec. and (3) the selection of the control points and the specification of the boundary conditions. Bristow and Grose (1978)] the variations depending mainly on the choice of type and form of singularity distribution. one source panel method." 7. resulting in a set of linear algebraic equations that express the exact condition of flow tangency on the surface. the aerodynamic coefficients computed using panel methods are reasonably accurate.2 Methods The first step in a panel method is to divide the boundary surface into a number of pan els. respectively. Although many different combinations are in principle mathematically equivalent. be perturbed. Bristow and Grose (1978) discuss some problems with the source panel class of methods when Used on thin.each control point (and there is one control point per panel). The computed values of CL are presented as a function of a for a 45° sweptback and a 45° sweptforward wing in Fig. the position of the surface must.23a and b.
angle of attack: (a) 64A010 section. A 0. 7 / Incompressible Flow about Wings of Finite Span Data . (1985). fortuitously yielding good agreement between the VLM and experiment. Vortex lattice method — .23 Comparison of the lift coefficient as a function of 450. deg (a) Figure 7.352 0 Chap.5 VORTEX LA111CE METHOD The vortex lattice method is the simplest of the methods reviewed by Margason. The VLM represents the wing as a surface on which a grid of horseshoe vortices is superimposed. [From Margason et al. "The VLM predicts the experimental data very well. the effect of viscosity offsets the effect of thickness. (1985) note. aft swept wing.Source panel method Surface potential distributions 0. et al. For most cases. (1985). The velocities induced by each horseshoe vortex at a specified control point are calculated using the law of BiotSavart." 7.6 CL a.A summation is performed for all control points on the wing to produce a set of linear algebraic equations for the horseshoe . due to the fact that vortex lattice methOds neglect both thickness and viscosity effects.— .5.0.] consistently overpredict the experimental data with little difference between the lift coefficients predicted by the various surface panel methods. As Margason et al. AR = 3.
deg (b) = —45°.23 (continued) (b) AR = 3.5 / Vortex Lattice Method 0  353 Data Vortex lattice method — .1 0I 1 I —2 0 2 4 6 8 10 12 a.55. for many engineering applications. 7.2  0. suitable accuracy can be obtained using linearized theory inwhich straightline trailing vortices extend downstream to infinity In the linearized approach.— Source panel method Surface potential distributions 0. they follow a curved path. as shown in Fig. A = 0. This procedure for obtaining a numerical solution to the flow is termed the vortex lattice method. In our approach to solving the governing equation.4  0. therefore.24. the vortex lattice panels are located on the mean camber surface of the wing and. The bound vortex coincides with the quarterchord line of the panel (or element) and is. The vor tex strengths are related to the wing circulation and the pressure differential between the upper and lower wing surfaces. In a rigorous theoretical analysis. when the trailing vortices leave the wing. However. (1985).5.3  0.5 0. the continuous distribution of bound vorticity over the wing surface is approximated by a finite number of discrete horseshoe vortices. [From Margason et al. Figure 7. The pressure differentials are integrated to yield the total forces and moments. forward swept wing.— . the trailing .6 I I I I I I 0. NACA 64A112 section.] vortex strengths that satisfy the boundary condition of no flow through the wing. The individual horseshoe vortices are placed in trapezoidal panels (also called finite elements or lattices).Sec. aligned with the local sweepback angle. 7.
7 / Incompressible Flow about Wings of Finite Span y Bound vortex Control point t____ I The dihedral angle 7 Figure 7. It induces a velocity. Both orientations provide similar accuracy within the assumptions of linearized theory.354 Freestream flow Chap. vortices are aligned either parallel to the free stream or parallel to the vehicle axis. .26. and horseshoe vortices for a typical wing planform in the vortex lattice method. elemental panels. 7. there is no flow through the surface) provides a set of simultaneous equations in the unknown vortex circulation strengths.25. these geometric coefficients do not change as the angle of attack is changed. An indication of why the threequarterchord location is used as the control point may be seen by referring to Fig. This orientation of the trailing vortices is chosen because the computation of the influences of the various vortices (which we will call the influence coefficients) is simpler..24 Coordinate system. 7. Furthermore. Application of the boundary condition that the flow is tangent to the wing surface at "the" control point of each of the 2N panels (i. In this text we shall assume that the trailing vortices are parallel to the axis of the vehicle. as shown in Fig.e. The control point of each panel is centered spanwise on the threequarterchord line midway between the trailingvortex legs. A vortex filament whose strength F represents the lifting character of the section is placed at the quarterchord location.
12). as was discussed in equations (6. If the flow is to be parallel to the surface at the control point.5 I Vortex Lattice Method z 355 y .11) and (6. 2irr at the point c. 1= = Combining the preceding relations gives us F = I U.. . the control point which is a distance r from the vortex filament. 7.25 Distributed horseshoe vortices representing the lifting flow field over a swept wing.26 Planar airfoil section indicating location of control point where flow is parallel to the surface. sin a Figure 7.c/4 x Filled circles represent the control points Figure 7.Sec. the incidence of the surface relative to the free stream is given by a' Sin a U = F But.
Thus.2.33 for the wing of Example 7. the lift is reduced both near the center and near the tips of the wing. Note that the boundvortex filaments for the port (or lefthand) wing are not parallel to the boundvortex filaments for the starboard (or righthand) wing. This will be evident in the spanwise lift distribution presented in Fig.1 Velocity Induced by a General Horseshoe Vortex The velocity induced by a vortex filament of strength and a length of dl is given by the law of Biot and Savart [see Robinson and Laurmann (1956)]: dV — (7.27 Nomenclature for calculating the velocity induced by a finitelength vortex segment. 7.The downwash resulting from the boundvortex system is greatest near the center of the wing.) 7.31. the boundvortex system on one side of the wing produces downwash on the other side of the wing. we see that the control point is at the threequarterchord location for this twodi mensional geometry. Thus.356 Chap.g. This downwash reduces the net lift and increases the total induced drag produced by the flow over the finitespan wing. for a swept wing.25.34) Referring to the sketch of Fig. Consider the flow over the swept wing that is shown in Fig. The use of the chordwise slope at the 0. Falkner (1943) and Kalman et a!. 7. (See Fig. . the magnitude of the induced velocity is = F sinOdl 4irr 2 (735) Vorticity vector B 3 r2 A 3 Ti C Figure 7. (1971)].5. 7. while the downwash resulting from the trailingvortex system is greatest near the wing tips. 7 / Incompressible Flow about Wings of Finite Span Solving for r yields C Thus.. 7.75chord location to define the effective incidence of a panel in a finitespan wing has long been in use [e. for a lifting swept wing.27.
with the vorticity vector directed from A to B. y. the symbol represents the magnitude of the parameter. r0 is the magnitude of the X r21 represents the magnitude of the vector cross product. 7. z) by the horseshoe vortex shown in Fig. and BC. For the bound vortex.34) to calculate the effect of each segment separately. Thus. In this case F = 0 and v= which is the result used in Chapter 6 for the infinitespan airfoils. Let AB be such a segment. 7. let us use equation (7. 7. then 02 = ir. respectively.36) Note that.36) and noting that the direction of the induced velocity is given by the unit vector r1 X r2 yields IT1 X r21 L r2JJ (7. The resultant induced velocity vector will be calculated by considering the influence of each of the elements.37) This is the basic expression for the calculation of the induced velocity by the horseshoe vortices in the VLM.Sec. We shall now use equation (7. segment AB. the nth panel) in Fig.g. The trailing vortices are parallel to the x axis.Then Ti X r2( and designate cosO1= r0r1 cosO2= r0r2 In these equations.28. if a vector quantity (such as is written without a superscript arrow. if the vortex filament extends to infinity in both directions.. as shown in Fig. = = — — + (Y2n — Yin)] + (z2n — = (x (x + (y Yin)J + (z — — + (Y — Y2fl)J + (z — .5 I Vortex Lattice Method 357 Since we are interested in the flow field induced by a horseshoe vortex which consists of three straight segments. AC. Segment AB represents the bound vortex portion of the horseshoe system and coincides with the quarterchord line of the panel element. Let the vectors AB. The horseshoe vortex maybe assumed to represent that for a typical wing panel (e. 1t can be used regardless of the assumed orientation of the vortices. Let C be a point in space whose normal distance from the line AB is We can integrate between A and B to find the magnitude of the induced velocity: V = F [02 / sin 0 dO = " (cos — cos 02) (7.37) to calculate the velocity that is induced at a genera! point in space (x.27. vector Also note that Substituting these expressions into equation (7. 7.24.
7 I Incompressible Flow about Wings of Finite Span y Bound vortex Trailing vortex from B to the x axis) x Figure 7. z) by the vortex filament AB (shown in Figs.y. y. .28 "Typical" horseshoe vortex.z) C(x.29 Vector elements for the calculation of the induced velocities.y.358 Chap. r1 X r2 2 X T2 = {{(y — — (y — Y2n)(Z — — {(x — — — Y2n) — (x — — — + {(x Z2fl) (x — X2n)(y — Yin)]k}/ yin.z) To r1 Zin) Figure 7.29).28 and 7. Using equation (7. VAB = —{FaclAB}{Fac2AB} 4ir Fn (7.38a) where {FaCIAB} = . 7.37) to calculate the velocity induced at some point C(x. C(x.
the induced velocity is —. Since is in the direction of the vorticity vector. finitelength filament that extends from A to D. Thus. let us first calculate the velocity induced by the collinear.5 / Vortex Lattice Method 359 {{(Y — Yin)(Z — (y — — + [(x — + [(x — Xin)(y — Z2n) — — (x — x2fl)(Z — — Y2n) — (x — X2n)(y and {FaC2AB} = r0— — ÷ — r0— + (Y2n — T2 = — + Yin)(Y — Yin) + — — — V(x — — Yin)2 + (z — + (Y2n zin)2 — — [(x2fl — — Yin)(Y — Y2n) + — V'(x + (y — Y2n)2 + (z To calculate the velocity induced by the filament that extends from A to 00. the velocity induced by the vortex filament which extends from A to oo in a positive direction parallel to the . 7. VAD = F where {Facl D} = and (z — Zin)] + (Yin — y)k [(z — + (Yin Y) — {Fac2AD} = — xin)T X3n —X Yin)2 + (z — IV(x — + + (Y 2 XXin V(x — Xin)2 + (Y — Yin)2 + (z — Zin)2 Letting x3 go to 00. the first term of {Fac2AD} goes to 1. 7.Sec.O.Theréfore.29. = = — x3n)i — + (Y — Yin)] + (z = (x as — + (y — Yin)] + (z — Zin)k shown in Fig.
If the flow is tangent to the wing. 7 I Incompressible Flow about Wings of Finite Span x axis is given by — J (z — — Zin)J + — y)k y)2] + — [1. the resultant flow is tangent to the wing at each and every control point (which is located at the midspan of the threequarterchord line of each elemental panel).e.38c) The total velocity induced at some point (x. if the strengths of the 2N horseshoe vortices are known.38). However. To evaluate the induced velocity components. one for each of the control points. 7.5. that for the nth panel) is the sum of the components given in equation (7. the component of the induced velocity normal to the wing at the control point balances the normal component of the freestream velocity.360 Chap. Since the governing equation is linear. the velocity induced by the vortex filament that extends from B to oo in a positive direction parallel to the x axis is given by f (z — 4ir — + 2 — y)k — + + y)2] X 2fl L \/(x — + (Y — + (z 1 — (7. we use the boundary condition that the surface is a streamline. we must introduce at this point our convention that the trailing vortices are parallel to the . Let the point (x.. z) be the control point of the rnth The velocity induced at panel. which we will designate by the coordinates (xrn.2 Application of the Boundary Conditions Thus. the velocities induced by the 2N vortices are added together to obtain an expression for the total induced velocity at the rnth control point: 2N (740) We have 2N of these equations.To compute the strengths of the vortices. lifting flow field of the wing.39) where the influence coefficient depends on the geometry of the nth horseshoe vortex and its distance from the control point of the mth panel. it is possible to determine the resultant induced velocity at any point in space. z) by the horseshoe vortex repre senting one of the surface elements (i.n. we see that (7.38). their strengths are which represent the not known a prion. y.38b) Similarly. That is. the mth control point by the vortex representing the nth panel will be designated as Vm. Examining equation (7.0 + L — * + (y — Yin)2 + (z 1 — (7.
38) is the vehicle axis].42) . (c) section BB.. (b) section AA.41) where 4) is the dihedral angle. and 6 is the slope of the mean camber line at the control point. as shown in Fig. vehicle axis [i.— Mean camber surface (b) (c) Figure 7.30.24. 7. Thus. the tangency requirement yields the relation Um S111 6 cos (1) Vm COS 6 sin 4) + Wm cos 4) cos 6 + sin(a — 6) cos 4) = 0 (7.5 / Vortex Lattice Method z 361 Line in xy plane (reference for dihedral angle) (a) Mean camber z surface .30 Nomenclature for the tangency requirement: (a) normal to element of the mean camber surface. 6 = dx For wings where the slope of the mean camber line is small and which are at small angles of attack. 7. equation (7. Referring to Fig.e. the x axis for equation (7.41) can be replaced by the approximation Wm — Vm tan 4) + ra — L i—) = 0 (7.Sec. 7.
3 Relations for a Planar Wing Equations (7.43) by combining the components into one expression: Ff1 41T 1 (Xm — Xin)(yrn Y2n) Yin) xin)(xm — — + (Y2n — Yin)(Yrn — Yin) L — — V(xrn + + (Y2n + — Yin)2 Yin)(Yrn — Y2n) Xin)(Xm — — — .43a) Tl — k Yin — Ym[ [ il. TherefOre. a planar wing (i. for the planar wing. for our planar wing. Furthermore. 7 / Incompressible Flow about Wings of Finite Span This approximation is consistent with the assumptions of linearized theory. we can simplify equation (7. a downwash). they can be solved to determine the lifting flow for a twist ed wing withdihedral. so that we can learn the significance of the various operations using a geometry which we can readily visualize. For a planar wing.42) are those for the VLM where the trailing vortices are parallel to the x axis. The unknown circulation strengths required to satisfy these tangent flow boundary conditions are determined by solving the systepi of simultaneous equations represented by equation (7. Zm 0 for all the control = points. one that lies in the xy plane). The solution involves the inversion of a matrix. 7.38) through (7.40). 4ir (Xm — Xin)(ym Y2n) — (Xm — X2n)(Ym — — xin)(xm L — — + (Y2n — Yin)(Yrn — Yin) V'(xm (x2n — Xin)(Xm + — Yin)2 — — + (Y2n — Yin)(Ym Y2n)2 Y2n)1 — x2fl)2 + (Ym — J (7.362 Chap. \/(xm — xi8)2 + (Yrn — Yin)2 TI — F 41TY2n k YmL — xrn fl + (Yrn — Note that.e. all three components of the vortex representing the nth panel induce a velocity at the control point of the iiith panel which is in the z direction (i. As such. Let us apply these equations to a relatively simple geometry.. Thus.e.5. = 0 for all the bound vortices. AB —..u+ 7.
The vortex lattice method will be applied using only a single lattice element in the chordwise direction for each spanwise subdivision of the wing. Since we are considering a planar wing in this section. and the trailing edge all have the same sweep. the leading edge. EXAMPLE 7.2: Use the vortex lattice method (VLM) to calculate the aerodynamic coefficients for a swept wing Let us use the relations developed in this section to calculate the lift coefficient for a swept wing. Wm = In Example 7. + (Yin — Y2n)2JJ (7. the quarterchord line. it would not be adequate for a wing with cambered sections or a wing with deflected flaps. 2N Wm = n =1 Wrn. However.e. the resultant flow will be tangent to the wing if the total vortexinduced downwash at the control point of the rnth panel. we will solve for the aerodynamic coefficients for a wing that has a relatively simple planform and an uncambered section.e. Applying the boundary condition that there is no flow through the wing at only one point in the chordwise direction is reasonable for this flatplate wing..47) For small angles of attack.0 — Y2n — Ym[ I 1. Since b2 and since for a swept.45) balances the normal component of the freestream velocity: Wm + sin a Uc.41) and (7. it is a flat plate). 7.31).0 + — xm —xj. = 0 everywhere and = 0. that illustrated in Fig.5 / Vortex Lattice Method 1.2. Thus. 45°. and an uncambered section (i. untapered wing S bc . the threequarterchord line.e. = ce). 7.n (7.. The wing has an aspect ratio of 5..45) Let us now apply the tangency requirement defined by equations (7.oa 0 (7.0 Xrn — 363 + Yin [1. Since the taper ratio is unity.46) (7.44) Summing the contributions of all the vortices to the downwash at the control point of the mth panel. So that the calculation procedures can be easily followed. which is calculated using equation (7. The component of the freestream velocity perpendicular to the wing is sin a at any point on the wing.0 + YmL \/(xm \/(xrn — xjn)2 + (ym Yin)2 1.Sec. let us consider a wing that has a relatively simple geometry (i. a taper ratio of unity (i.42).
2b O. Solution: The flow field under consideration is symmetric with respect to the y = 0 plane (xz plane). it is clear that b = 5c. we must remember to include the contributions of the horseshoe vortices of the port wing to the velocities induced at these control points (of the starboard wing). it is possible to calculate all of the necessary coordinates in terms of the parameter b. that is. AR = 5. A = 45°.np . However.31 Fourpanel representation of a swept planar wing. we need only to solve for the strengths of the vortices of the starboard wing. Using this relation. equation (7. Because of symmetry. Furthermore.e.500b Figure 7.47)] only at the control points of the starboard wing. taper ratio of unity.364 Chap.2b L CP 1 CP2 bound vortex CP3 O. Therefore. we need to apply the tangency condition [i.Thus. 7 / Incompressible Flow about Wings of Finite Span Freestream flow 1450 / O. for this planar symmetric flow. there is no the lift force acting at a point on the starboard wing (+y) is equal to that at the corresponding point on the port wing (—y). equation (7.. the solution does not require that we know the physical dimensions of the configuration.45) becomes N N Wm = Wmns + n1 Wm.
Using equation (7.3533 — 30.y2fl.0 r _I1.0+ L 0. Thus.2125b O.1875b O.0625b)2 1.0 O.1250b 0.4250b 0.0375b)(0.1625b) + (0.1750b O. 0) are the coordinates of a given control point and that and (x2fl.1s — — (0.4375b 0.0625b)2 (0.9335 — 24.0375b)2 + (—O.5 I Vortex Lathce Method 365 where the symbols s and p represent the starboard and port wings. we can calculate the strength of the horseshoe vortices using only a pocket electronic calculator. The control points are designated by the solid symbols in Fig.1750b 0.31. By limiting our selves to only four spanwise panels.3750b 0. The coordinates for a 4 X 1 lattice (four spanwise divisions and one chordwise division) for the starboard (right) wing are summarized in Table 7.3000b O.2 Coordinates of the Bound Vortices and of the Control Points of the Starboard (Right) Wing Panel 1 Xm Ym Yin O.3125b O. As before.4625b O. 1. 7.2319) TABLE 7.3375b O. the bound portion of each horseshoe vortex coincides with the quarterchord line of its panel and the trailing vortices are in the plane of the wing.0625b)2 — 0.37 SOb 0. respectively.0375b 1.1250b)(0.1250b)(—0.1625b)2 + (O.0625b) V(0.1250b)(0.2500b O.1250b O.0625b O.Sec.0625b) L V(0.3000b 0.1625b)2 + (0.S500b 0.1250b)(0. each panel extending from the leading edge to the trailing edge.5875b 2 3 4 O.0000b Y2n 0.0 w1.2. The planform of the starboard wing is divided into four panels.0625b) [(0.O) are the coordinates of the "ends" of the boundvortex filament AB.1625b \/(0.O625b[L0 + V(0. 7.0625b)2 [ = 16.5000b . parallel to the x axis.4250b O.44) to calculate the downwash velocity at the CP of panel 1 (of the starboard wing) induced by the horseshoe vortex of panel 1 of the starboard wing. Recall that (Xm.0375b) + (0.2500b 0. we can more easily see how the terms are to be evaluated.0500b O.0375b)2 + (—0.
1250b)(—0. Adding the components together.0875b)(0. The student should visualize the flow induced by each segment of the horseshoe vortex to verify that a negative value for each of the components is intuitively correct.1250b)(—0.3125b)2 .1625b) + (0.o315b) + (0.0625b) V(0.3125b) V(_o.1875b L r 11.) (—0.0625b)2 + 1.0392 47th 6.1250b)(O. In addition.1250b) ( —0.1625b)(0.1875b) [ — \/(O. as one would expect.1p — to — (O.0 —O.1s = The downwash velocity at the CP of panel 1 (of the starboard wing) induced by the horseshoe vortex of panel 1 of the port wing is w1.0375b V(o. the velocity induced by the vortex trailing from A to is greatest in magnitude.1875b) [(0.0875b) + (0.0625b)2 — = —[—6.12501.o375b)2 + (0. 7 / Incompressib'e Flow about Wings of Finite Span Note that.0625b L'° 1. using equation (7. each of the vortex elements induces a nega tive (downward) component of velocity at the control point.9335] Similarly.0375b)2 + (0.2125b)(—0.1250b)(0.1875b)2 (0.1625b — r —0.366 Chap.1875b)2 O.1625b)2 + (O. we obtain to — 4 l(0.1250b)(0.3125b) — (—0.1625b)2 + (O.1875b) [(—o.0+ 0.1875b)2 (—0.2125b) + (0.3793 + 30.0875b)2 + (—0.1875b) L — V(—0.l2sob)(o.44) to calculate the downwash velocity at the CP of panel 2 induced by the horseshoe vortex of panel 4 of the starboard wing. we find w1.2125b)2 + (—0.0 F1 + V(0.
34371'2 + 1.0+ —0.48341'2 + 0. 7.o875b)2 + (—0.5 I Vortex Lattice Method '1.0757) 4'irb F4 Again. we find that at control point 1 Wi = 71.8792F2 + 20.07571'3 + 0.5887F3 + At CP 2.40061] 4irb —(1.3125bL T4 [ 11.7227F2 + 0.5187F3 + + (+l.3775F4)S + (f18.37761'3 + At CP 4.1742F2 + o.60167 + 3.2174F2 — 71.2125b \/(_0.oo37r1 + 13. the noflow condition of equation (7.07795 — 1.2933F2 + 1. the student should visualize the flow induced by each segment to verify that the signs and the relative magnitudes of the components are individually correct.2933F3 + + 1.6644F3 + 0.28951'3 + Since it is a planar wing with no dihedral.0 367 r —0.3125b)2 —[—0. — 71.0875b + 1.2125b)2 + (—0.1875b)2 O.515OF1 + 2.6434F4 = — 70.4903F3 + + 20.2174F3 — + 0.0 — + V(—0.0504F2 + 0. Evaluating all of the various components (or influence coefficients).5187F2 + 11.3260F4 = .78361'3 + 1.s4Sor'1 + 0.Sec. = + At CP 3.47) requires that Thus —53.34451'2 + 11. = + + 3.5187F1 + 11.
1250b for each panel. the total lift L= 2 J dy (7. 7.50a) 0 or. + 0.48c) (7.0601 per degree Comparing this value CL.1250b To calculate the lift coefficient.48b) (7. Sc for this CL = —f. all the lift is generated by the freestream velocity crossing the spanwise vortex filament.5069173 — Solving for 173.94o1r2 — 71.0287 + 0. the lift of the wing may be calculated. 4 L= Since L (7.a = dCL = 3.3351174 +2.4943r1 + 4. 7 / Incompressible Flow about Wings of Finite Span +5.49) which is also the lift per unit span. Since the flow is symmetric. Therefore. we find that = "2 173 = 174 = (7.50b) = 0.0961Ta Furthermore. a with that for an unswept wing (such as the results pre sented in Fig. the lift acting on the nth panel is = for the wing is (7. since the panels extend from the leading edge to the trailing edge. it is apparent that an effect of sweephack is the reduction in the liftcurve slope.0250)0. recall that S = bc and b = wing.3626f'2 + 20.= 1. and 174.443 per radian 0.1411F3 + n.5112r4 71.368 Chap.48a) (7.14).4272F1 + 2o.0286 + 0.48d) Having determined the strength of each of the vortices by satisfying the boundary conditions that the flow is tangent to the surface at each of the control points. in terms of the finiteelement panels. since there are no sidewash or backwash velocities. . Furthermore. For wings that have no dihedral over any portion of the wing. CL.
31 in a subsonic stream.48) is used . 7.Sec. The theoretical lift curve generated using the VLM is compared inFig.32 with experimental results reported by Weber and Brebner (1958). the value of 1' given in equation (7. the section lift coefficient for the nth panel is 2F — — (7. 7. such as is the case for the 4 X 1 lattice shown in Fig. which was swept 45° and which had an aspect ratio of 5.49). The theoretical lift coefficients are in good agreement with the experimental values. Since the lift per unit span is given by equation (7. 7. The experimentally determined values of the lift coefficient are for a wing of constant chord and of constant section.32 Comparison of the theoretical and the experimental lift coefficients for the swept wing of Fig.31. 7.51) av When the panels extend from the leading edge to the trailing edge.5 / Vortex Lattice Method 369 Data from Weber and Brebner (1958) Inviscid solutiOn using VLM for 4)< 1 lattice o CL 0 2 4 6 8 10 12 Figure 7.
33.51). Jj where Cay is the average chord (and is equal to S/b).8CI ° 0 CL 0.8 1. When there are a number of discrete panels in the chordwise di rection. Jmax( Cay 1 (7. The total lift coefficient is obtained by integrating the lift over the span CL = L (7.4 2y b 0.33 Comparison of the theoretical and the experimental spanwise lift distribution for the wing of Fig. 7.2°. . you should sum (from the leading edge to the trailing edge) the values of I' for those boundvortex filaments at the spanwise location (i.0 Figure 7.e. and j is the index for an elemental panel in the chordwise row.24.6  0.52) \. such as the 10 X 4 lattice shown in Fig. c is the local chord. 7 / Incompressible Flow about Wings of Finite Span Data for a = 4.0 0.2 0.4  0. which was presented by Weber and Brebner (1958).2  0.20 from Weber and l3rebner (1958) Inviscid solution using VLM for 4 X 1 lattice 12 ° o 1.31. 7. 7.0 I 0. For a chordwise row.53) The spanwise variation in the section lift coefficient is presented in Fig.. in equation (7.6 0.0 0.The the oretical distrjbution is compared with the experimentally determined spanwise load distribution for anangle of attack of 4. in the chordwise strip) of interest.370 a Chap.
54) .5 I Vortex Lattice Method 371 y AR =5 A= a 45° 150 =4X 1. 7. but to a reduced extent.%o y 1.0 (a) (b) Figure 7. Boundarylayer fences are often used to break up the spanwise flow on swept wings.0 1. much more important.00. the induceddrag coefficient may be calculated using the relation given by Multhopp (1950): +03b dy k) —O. (b) with fence. This unfavorable behavior is amplified by the fact that the spanwise velocity component causes the already decelerated fluid particles in the boundary layer to move toward the wing tips. 7.Sec.Thus.e.0 \ (9 — \\  C! CI I I 0. If significant tip stall occurs on the swept wing.34. in that the fence divides each wing into an inner and an outer portion. premature separation may occur on the suction side of the wing near the tip. [Data from Schlichting (1960).34 Effect of a boundarylayer fence on the spanwise distribution of the local lift coefficient: (a) without fence. Both transverse flow and boundarylayer separation may be present. shown in Fig. The essential effect of the boundarylayer fence does not so much consist in the prevention of the transverse flow but. The spanwise distribution of the locallift coefficient [taken from Schlichting (1960)] without and with a boundarylayer fence is presented in Fig. This transverse flow results in a large increase in the boundarylayer thickness near the wing tips. Once we have obtained the solution for the section lift coefficient (i. Boundarylayer fences are evident on the swept wings of the Trident.5b (7. at large angles of attack. there is a loss of the effectiveness of the control surfaces and a forward shift in the wing center of pressure that creates an unstable. that for a chordwise strip of the wing). noseup increase in the pitching moment.] The increased loading of the outer wing sections promotes premature boundarylayer separation there.0 o.34.0 1.. 7.
we consider the mth chordwise strip.) where a1.5b C1c a1 = [(y — + C1c 1 (7. Giesing..—:: I = amY!2 + + Cm (7. note that Ym+i = Yin—i = y. and cm.35 Trident illustrating boundarylayer fences on the wing.71 + (em + + em_i) Thus. (Courtesy of British Aerospace. 7 I Incompressible Flow about Wings of Finite Span Figure 7.57) To solve for the coefficients am. bm. fC1c'\ Cm = — 2 — 1 am { dmj( = d 'iii dmo(dmi + dmo)  — + (C1c \CLC + (C1c CLC .55) a symmetrical loading. equation (755) may be written 1 O.56) (y + Following the approach of Kalman. which has a semiwidth of em and whose centerline is located at 77 Ym Let us approximate the spanwise lift distribution across the strip by a parabolic function: I cic\ (. given by 1 C1C (7. which is the induced incidence.___________ ___________ 372 Chap. and Rodden (1970).
(y) CLc — 1 N JY2(Ym + 1 em) am + y2bm + + em) Cm — y2 ern)am — + em)cm y2(ym + y2bm + (yrn — 1 2 (y+ern) 2 iog[Y Y Yrn + + em)]2 + 2emam} (7. we let — em and em+i = 0 at the tip.5 / Vortex Lattice Method and 1 bm 373 1 — = dmidmo(dmi + dmo) — dmj(27)m — dmi)i (C1c — m where dmi em + em_i and dmoem+em+i For a symmetric load distribution.56) and (7. = + + (7. 7.57). Similarly.Sec.59) . we then obtain the numerical form for the induced incidence: a. Substituting these expressions into equations (7. we let and em_i = at the root.58) We then assume that the product also has a parabolic variation across the strip.
afld Cm. For techniques to compute the lift. VLM is only one (and. Therefore. the reader is referred to Margason and Lamar (1971).. 7. Aircraft designed to fly efficiently at supersonic speeds often employ thin. as noted in the introduction to this section. 7. Values of s are presented as a function of the Reynolds . The data presented inFig.6 FACTORS AFFECTING DRAG DUETOLIFT AT SUBSONIC SPEEDS The term representing the liftdependent drag coefficient in equation (7. 7. 7. For an elliptically loaded wing. which result when the angle of attack changes from a01. For the present analysis.36b as a function of the leadingedge sweep angle for several sharpedge wings.374 Chap. which is represented by the symas given in bol CDT. this value would be equation (7.54) is then a generalization of Simpson's rule: The coefficients CDv_ A N —A )I r y. 7. The effect of two of the features (leadingedge flaps and wing warp) is illustrated in Fig. Furthermore. Numerous investigators have studied methods for improving the convergence and the rigor of the VLM. Several features that can increase the effective leadingedge suction can be incorporated into a wing design. bm. Thus. highly swept wings. indeed.16). For 100% suction.4 and whose sharp leading edge is swept 67° in Fig.36 were obtained at the lift coefficient for which (LID) is a maximum. Even for relatively low values for the sweep angle. and the rolling moment for more general wings.37. Zero percent suction corresponds to the condition where the resultant force vector is normal to the chord line. The values of s vary only slightly with the Reynolds number.36a. 7. This lift coefficient is designated CL opt• Values of s are presented as a function of the freestream Reynolds number based on the chord length for a wing with an aspect ratio of 1. Values of the leadingedge suction reported by Henderson (1966) are reproduced in Fig.Values of s are presented in Fig. the dragduetolift coefficient is the potential flow induced vortex drag. there is an everexpanding body of relevant literature with which the reader should be familiar before attempting to analyze complex flow fields. the suction parameter can be presented as a function of the leadingedge sweep angle (independent of the Reynolds number).36. a relatively simple one) of the methods used to compute the aerodynamic coefficients. the pitching moment. i' n1 r The lift developed along the chordwise bound vortices in a chordwise row of horseshoe vortices varies from the leading edge to the trailing edge of the wing because of the longitudinal variation of both the sidewash velocity and the local value of the vortex strength for planforms that have a nonzero dihedral angle. as a result of extensive separation at the wing leading edge. the "effective leadingedge suction" (s) will be used to characterize the dragduetolift. The numerical form of equation (7. suction values no higher than about 50% were obtained.17) includes those parts of the viscous drag and of the form drag. 7 / Incompressible Flow about Wings of Finite Span and can be obtained using au approach identical to that employed to find am.
OPt. As noted by Henderson (1966). 7. At the high lift coefficient (CL 1. derson (1966). A third feature that can be used to improve the effective leadingedge suction is the wing leadingedge radius.38. The total trim drag has two components: (1) the drag directly associated with the tail and (2) an increment in drag on the wing due to the change in the wing lift coefficient required to offset the tail load. the increase in the effective leadingedge suction due to the leadingedge radius accounts for 2/3 of the increase in (L/D)max from 8 to 12. Again.0) which .Sec. and M < 0. Data originally presented by Henderson (1966) are reproduced in Fig.0058c) exhibit large increases in s as the Reynolds number increases. dragduetolift values approaching those corresponding to full leadingedge suction are generally obtained. Therefore.4 • :. sharp leadingedge wing. [Taken from Henderson (1966).0 and whose leading edge is swept 67°.6 I Factors Affecting Drag DuetoLift at Subsonic Speeds A 375 100 'ii 100 AR=1. The full leadingedge suction line and the zero leadingedge suction line are shown for reference. 7.2. which are essentially independent of the Reynolds number. McKinney and Dollyhigh (1971) state "One of the more important considerations in a study of trim drag is the efficiency with which the wing operates.30) typical of 1g flight.30. The values of s for the wing with a rounded leading edge (rLE = 0. the wing with a sharp leading edge generates relatively low values of s.39. At low lift coefficients (CL 0. 7. CL. A reduction in the skin friction drag accounts for the remaining 1/3 of the increase in (L/D)max. (X106) (a) ALE(°) (b) Figure 7.] number for a wing swept 74° for The data [again taken from Henderson (1966)] show that both leadingedge flaps and wing warp significantly increase the values of the effective leadingedge suction relative to those for a symmetrical. The variation of leadingedge suction with Reynolds number and wing geometry is discussed in a paper by Hen. a typical variation of the subsonic dragduetolift parameter (which is a measure of wing efficiency) is presented in Fig. Values of s are presented as a function of the Reynolds number for a wing with an aspect ratio of.36 The variation of the effective leadingedge suction (s) as (a) a function of the Reynolds number and (b) of the leadingedge sweep angle for a wing with a sharp leading edge. The crosshatched area indicates a typical range of values of the dragduetolift parameter for current aircraft including the use of both fixed camber and twist or wing flaps for maneuvering.
%C 0. CL <0.58 8 3CD I s = Constant 50  ac2L C 1 2 4 (X10—6) 10 20 0 1 I I I 2 4 10 20 Figure 7.376 Chap. AR = 2.0 Leading edge 12 TLE.] A U.M < 0. [Taken from Henderson (1966).%c 0.deg  0•— Oi_ • 4 S S• I I I I 2 10 20 1 2 4 10 20 (X106) Figure 7.37 The effect of leadingedge flaps and wing warp on the effective leadingedge suction (s) for CL0I.] .t.30.58 100  rLE. 7 / Incompressible Flow about Wings of Finite Span 100 Symmetrical section = 00 s 50  8f.30. [Taken from Henderson (1966).38 The effect of the leadingedge shape.
it becomes increasingly more difficult to prevent boundarylayer separation. as the sweep angle is increased and the section thickness is decreased in order to avoid undesirable compressibility effects.7 / Defta Wings 377 0. a major aerodynamic consideration in wing design is the pre diction and the control of flow separation." 7. aerodynamics is only one of them. deltawing planforms have aerodynamic characteristics which are substantially different from those of the relatively straight.39 Typical variation of dragdue tolift parameter. . At subsonic speeds. however. sub0. the Me 262.. the boundary layer on the lower surface flows outward and separates as it goes over the leading edge. is recognized. Beginning with the deltawing design of Alexander Lippisch in Germany during World War II. it is often necessary to employ rather complicated variablegeometry devices in order to satisfy a wide range of conflicting design requirements which result due to the flowfield variations for the flight envelope of highspeed aircraft. The final configuration will reflect design priorities and tradeoffs. aspect ratio = 2. However.1 Figure 7.3 0 0.This historical note is included here to remind the reader that many parameters enter into the design of an airplane. [Taken from Henderson (1966).5. Although many techniques have been discussed to alleviate these problems. Thus. It is interesting to note that during the design of the world's first operational jet fighter.40. forming a free shear layer. the outer panels of the wing were swept to resolve difficulties arising when increasingly heavier turbojets caused the center of gravity to move. the introduction of sweepback did not reflect an attempt to reduce the effects of compressibility [Voight (1976)}. eventually rolling up into a core of high vorticity. The shear layer curves upward and inboard.7 DELTA WINGS As has been discussed. as shown in Fig. The need for improving dragduetolift characteristics of wings at the highlift coefficients by means such as wing warp. and so forth. highaspectratio wings designed for subsonic flight. 7. .5 1.Sec.. supersonic aircraft designs have often used thin. highly swept wings of low aspect ratio to minimize the wave drag at the supersonic cruise conditions.0 sonic speeds. Because they operate at relatively high angles of attack. improved maneuver devices.] CL corresponds to the maneuvering case .2 c)c) 0. the dragduetolift typically approaches the zero leadingedge suction line even when current maneuver flap concepts are considered. 7.
(Photo graph courtesy of General Dynamics.] (d) Inclination angle of the vortex trajectory.) (c) Trajectory of the leadingedge vortex.] .378 Chap. 7 / Incompressible Flow about Wings of Finite Span (a) (b) y aCore I I1 x U7<a Vortex core x (d) (c) Figure 7.40 (a) Vortex core that . [Taken from Visbal (1995). (b) Water vapor condenses due to the pressure drop revealing the vortex core for a General Dynamic F16.develops for flow over a delta wing. [Taken from Visbal (1995).
However. The total lift coefficient can be represented by the sum CL = The constant sin a cos2 a + sin2 a cos a (7. A spanwise outflow is induced on the upper surface.0 [Peckham (1958)] to 2. which remains steady throughout the range of practical flight attitudes of the wing.41 Wing geometry nomenclature. delta. The leadingedge suction analogy developed by Poihamus (1971a and 1971b) can be used to calculate the lift and the dragduetolift characteristics which arise when the separated flow around sharpedge delta wings reattaches on the upper surface. Since the vortex flow induces reattachment and since the Kutta condition must. Using the nomenclature for arrow. Furthermore. the correlations apply to thin wings having neither camber nor twist. Data are presented for wings of aspect ratio from 1. Since the analytical method is based on an analogy with potentialflow leadingedge suction.5 are in good agreement with the measured values up to angles of attack in excess of 20°.43. therefore. which requires that flow reattaches on the upper surface inboard of the vortex. respectively. 1955).61) is simply the normalforce slope calculated using the potentialflow liftcurve slope.61) are compared in Fig. beneath the coiled vortex sheet. Values of the lift coefficient calculated using equation (7. 7. The constant be estimated from the potential flow leadingedge suction calculations.44 with experimental values for uncambered delta wings that have sharp leading edges. for the delta wing. I Figure 7.0 and 1. The formation of these vortices is responsible for the nonlinear aerodynamic characteristics that exist over the angleofattack range [Hummel (2004)]. 7. The size and the strength of the coiled vortex sheets increase with increasing incidence and they become a dominant feature of the flow.7 I Delta Wings 379 There is an appreciable axial component of motion and the fluid spirals around and along the axis.Sec. the total lift coefficient consists of a potentialflow term and a vortexlift term. whose aspect ratio is 2. The lift coefficients calculated using equation (7. . Thus. the correlation between theory and data breaks down as flow reattachment fails to occur. and diamondare presented as a function of the planform wings illustrated in Fig.0 (Bartlett and Vidal. and planform parameters in Figs.42 and 7. the method is applicable to wings for which the leading edges are of sufficient sharpness that separation is fixed at the leading edge. significant deviations between the calculated values and the experimental values exist for angles of attack above 15°. still be satisfied at the trailing edge. 7. 7.0.61) for AR = 1. and the flow separates again as it approaches the leading edge [Stanbrook and Squire (1964)].41.
[Taken from Polihamus (1971b).0 1. 7. the maximum value of the lift coefficient and the angle of attack at which it occurs increase as the aspect ratio decreases. The lift coefficients for a series of delta wings are presented as a function of the angle of attack in Fig. the flow pattern is essentially that of potential flow and the peak negative pressures at the section extremity are preserved. increasing the thickness causes a reduction in net lift.45.380 Chap. as noted by Peckham (1958). For a thin flatplate model.0 0.42 Variation of potentialflow lift constant with planform parameters.0 AR Figure 7. 7.0 2. separation occurs well inboard on the wing's upper surface.62) Using equation (7. the resultant force acts normal to the surface. The result is that.0 3. which are taken from Bartlett and Vidal (1955).63) . dCL/da.0 4. outboard of the loci of the separation points.] If the leading edges of the wing are rounded. However. CD = CDO + + (7. The lift coefficients are less for the thicker wings. 7 / Incompressible Flow about Wings of Finite Span ALE 45° 500 550 600 650 700 75° 0. which also have rounded leading edges.46. the induced drag for a flatplate wing would be = CD = CL tan a (7.61) to evaluate CL. However. The liftcurve slope. Thus. becomes progressively smaller as the aspect ratio decreases. The combined effect of the thickness ratio and of the shape of the leading edges is illustrated in the experimental lift coefficients presented in Fig. for all but the smallest aspect ratio wing.
as illustrated by the experimental values from Bartlett and Vidal (1955) presented in Fig.7 / Delta Wings 5.0 0.43 Variation of vortexlift constant with planforrrI parameters. 7. therefore. 7.0 LU 2.0 2. 7. At low speeds (and.0 55° 50° 450 3. The flow field over a delta wing is such that the resultant pressure distribution produces a large nosedown (negative) pitching moment about the apex.0 1.The resultant aerodynamic coefficients present a problem relating to the lowspeed performance of a deltawing aircraft which is designed to cruise at supersonic speeds. [Taken from Pollhamus (1971b). since the location of the aerodynamic center for subsonic flows differs from that for supersonic flows.0 0.0 3.J Experimental values of the drag coefficient from Bartlett and Vidal (1955) are compared with the correlation of equation (7.The magnitude of the negative pitching moment increases as the angle of attack is increased. 7.48.0 4.47). since L 2POOUOSCL . delta wings must be operated at relatively high angles of attack in order to generate sufficient lift.Sec.0 AR Figure 7. The experimental drag coefficient increases with angle of attack (see Fig.0 381 ALE 750 700 65° 60° 4.62) in Fig. at relatively low values of the dynamic pressure).47. The correlation is best for the higher values of the lift coefficient.
the basic delta configuration is often augmented by a lifting surface in the nose region (called a canard). which provides a noseup trimming moment.382 Chap. deg Figure 7. 7 I Incompressible Flow about Wings of Finite Span AR Source of data 2. In addition. the geometric characteristics of closecoupled canard configurations offer a potential for improved longitudinal progression of crosssectional area which could result in reduced wave drag at low supersonic speeds and placement of the horizontal control surfaces out of the high wing downwash and jet exhaust.e. negative camber) to provide the required trimming moment. if the wing is at an angle of attack that is high enough to produce the desired CL. Thus. These ." These benefits are primarily associated with the additional lift developed by the canard and with the beneficial interaction between the canard flow field and that of the wing. or retractable. such as those on the Dassault Mirage.0 X 106 6.6 X 106 6. such as those shown in Fig. However.0 Peckham (1958) 0 1. As noted by Gloss and Washburn (1978).0 X 106 0 1.49 on North American's XB70 Valkyrie.44 Comparison of the calculated and the experimental lift coefficients for thin. a large nosedown pitching moment results. flat delta wings with sharp leading edges. improved pitching moment characteristics. uses a reflexed wing trailing edge (i. An alternative design approach. and reduced trimmed drag. 7. "the proper use of canard surfaces on a maneuvering aircraft can offer several attractive features such as potentially higher trimmedlift capability. which is used on the Space Shuttle Orbiter. The canards may be fixed..0 Bartlett and Vidal CL 0 5 10 15 20 25 a.5 Bartlett and Vidal (1955) V 2.
" . boundary layer separation occurs at the sharp leadingedge. As the vortex reaches the wing trailing edge it is deflected and in a short distance in the near wake aligns itself with the freestream direction. [Data from Bartlett and Vidal (1955). beginning at small angles of incidence. 7.5. t = O. The inclination angle of the vortex trajectory relative to the wing surface (Fig. t = O. and their induced suction on the wing upper surface results in vortexinduced lift which persists at relatively large angles of attack as compared to the 2D wing situation.Sec. deg Figure 7. and results in a 3D shear layer which spirals intO a pair of counterrotating primary vortices above the wing (Fig. 7.40d) also increases linearly with angle of attack.050c O Elliptical leading edge."For a highly swept wing.0 0 5 10 15 20 25 a. and over the range of interest in angle of attack. the vortex sweep angle Acore (see Fig. the trajectory of the leadingedge vortex is straight.40a).] benefits may be accompanied by a longitudinal instability (or pitch up) at the higher angles of attack of the vortex lift developed on the forward canards.7 I Delta Wings o Beveled (sharp) leading edge.45 Effect df the leadingedge shape on the measufed lift coefficient for thin. Over most of the wing extent. 7.lOOc 383 CL 0. t = O. As discussed by Visbal (1995). x 106. 7.40c) is nearly independent of a and this constant value is approximately proportional to ALE.075c kound leading edge. For a fixed leadingedge sweep (ALE). These longitudinal vortices are the mechanisms for the downstream convection of the vOrticity shed at the edge. flat delta wings for which AR = = 6 1.
0 0. Visbal (1995) notes.6 CL 0.2 Chap.This phenomenon is characterized by reverse axial flow and swelling of the vortex core. the coherent fluctuations within the breakdown region can promote a structural response in aircraft surfaces immersed in the vortex path. For maneuvering delta wings.6 —10 20 a.16 3. Vortex breakdown poses severe limitations on the performance of agile aircraft due to its sudden effects on the aerodynamic forces and moments and their impact on stability and control.4 0." In a related comment.8 0. [Data from Schlichting and Truckenbrodt (1969). the interaction between the counterrotating leadingedge vortices increases and may lead to flow asymmetry.2 0.4 —0.] Visbal (1995) also noted.38 3.384 1. "The leadingedge vortices above a delta wing at high angle of attack experience a dramatic form of flow disruption termed 'vortex breakdown' or 'vortex bursting'.12c.97 V 0 0 Figure 7.0 —0. the onset of vortex development with its inherent long time scales results in dynamic hysteresis and lags in the vortex development and aerodynamic loads.2 —0.83 1. This phenomenon becomesmore severe when vortex breakdown is present given the sudden onset and sensitivity and the accompanying swelling of the vortex core. deg A AR = 0." . 7 / Incompressible Flow about Wings of Finite Span 1. An important example is that of 'tail buffet' in twintailed aircraft where the fluid/structure interaction may result in significant reduction of the service life of structural components.61 2. "As the leadingedge sweep and angle of attack increase. t = O. and is accompanied by marked flow fluctuations downstream.46 Lift coefficients for delta wings of various aspect 7x ratios. In addition.
5.2 0. = 6 x 106.Sec. (t = 0.] O Beveled (sharp) leading edge.8 CL 0. t = Round leading edge.0 0. deg for thin.6 0.lOOc) 1.5 0.7 / Delta Wings 0 Beveled (sharp) leading edge.5. 6 X 10. [Data from Bartlett and Vidal (1955).050c 0 —0.8 O.47 Drag correlation for thin.48 Moment coefficient (about the delta wings for which AR 1.4 0.1 0.3 0.7.0 0. flat Figure 7.0 0.2 0.] . flat delta wings for which AR = 1.2 1.2 0 A  I 5 10 15 20 25 a.050c) 385 A Round leading edge. (t = 0.lOOc —0.6 CDCDO Figure 7. t = 0. [Data from Bartlett and Vidal (1955).6  0 A — A CM04 0 A —0.
With the vortex core away from the surface. a delta wing could be designed to maintain its classically good supersonic performance without sacrificing cruise and subsonic performance compared to a more conventional trapezoidal wing. at some distance from the leading edge. or break down. ahead of the trailing edge toward the apex.. When the aircraft is at moderate to high angles of attack. Herbst (1980) noted.52. the leading edge suction can be improved considerably by means of proper leading edge profile. As the angle of attack increases.) As noted in the previous paragraphs and as depicted in Fig. profile. resulting in a lift overshoot. or bursting. or bursts. With the help of an electronic digital control system.49 North American XB70 illustrating the use of canards. The magnitude of the lift overshoot depends on the leadingedge sweep angle. Theoretical tools are available now to refine the wing planform. Note that the diameter of the vortex increases suddenly when bursting. Vortex breakdown. occurs. the flow over a delta wing with sharp leading edges is dominated by vortices inboard of the leading edge. the vortex breaks down. is depicted in the watertunnel flow of Fig. thus improving drag due to lift. . the vortexburst location moves forward.386 Chap. (Courtesy of NASA." . The location of the vortex breakdown during the unsteady pitchup motion is downstream of its steadystate location. Defining the positive attributes of a delta wing. 7. properly designed with modern aerodynamic tools and suitably equipped with a canard control surface. there are large reductions in the peak velocities both in the transverse and in the axial directions. The pressure on the wing's leeward surface beneath these vortices is very low and contributes significantly to the lift. the weakness of any highly swept wing . When vortex breakdown occurs. In particular..40. 7 / Incompressible Flow about Wings of Finite Span Figure 7. "There is a considerable basic aerodynamic potential in terms of lift to drag improvements in delta wings. the axes of the vortices move away from the leeward surface of the wing. The forward movement of the location of vortex breakdown and the attendant increase in the pressure on the leeward surface of the wing beneath the vortices are characteristic of the stall of a delta wing. 7. and twist distribution.
apex regions. Once the angle of attack becomes sufficiently large that strakevortex breakdown progresses ahead of the wing trailing edge. so that this system can no longer cause flow reattachment to occur on Low a Suction Analogy Modeling Flow sketch High a Edge Displaced vortex Figure 7. It should be mentioned that these vortexinduced benefits are realized when the strake vortex is stable and maintaining a wellorganized vortex system over the wing. (2) energizing of the upper surface boundary layer with the resulting flow reattachment on the outer wing panel at moderate to high angle of attack due to the strake vortex. because the strake provides large amounts of vortex lift. it was concluded that the configuration had attached flow away from the leading and side edges. ogee and double delta)."at low angle of a. 7. The strake/wing combinations are also known by different names (e.] . and downstream from the component tips. 7. at the higher values of a. As noted by Lamar and Frink (1982) and illustrated in Fig.8 LEADINGEDGE EXTENSIONS The designs of lightweight fighters that can cruise supersonically and maneuver tran sonically employ additional. fillets.Sec. These forward areas are called strakes. However. Benefits for the strake are (1) upwash from the main wing strengthens the strake vortex and (2) the need for only a small area (hence wetted area and comparatively lightweight structure) to generate its significant contribution to the total lift.50. and leadingedge extensions. these favorable effects deteriorate significantly. Lamar and Frink (1982) note that the mutual benefits derived from strake/wing configurations "include for the wing: (1) minimal interference at or below the cruise angle of attack. and (3) reduced area required for maneuver lift.8 I LeadingEdge Extensions 387 7." A method for estimating the aerodynamic forces and moments for strake/wingf body configurations estimates the vortex flow effects with the suction analogy and the basic potential flow effects with the vortex lattice method [Lamar and Gloss (1975)].. [Taken from Lamar and Frink (1982).50 Theoretical vortex model for strake/wing configuration. gloves. highly swept areas ahead of the main wing.g. vortical flow existed. the strake vortex becomes much larger and tends to displace the wingvortexflow system off the wing.
8 0 0. The CL data are better estimated by the higha theory.6 2. although present.52. "The correlation between modelscale tests and flight tests is surprisingly accurate. (1989) show some discrepancies between data obtained at dif ferent Reynolds and Mach numbers.8 I I 1. et al. most higha flight occurs at relatively low Mach numbers due to structural limitations." The flow field computed for an F18C at an angle of attack of 32° is presented in Fig. [Taken from Lamar and Frink (1982). Although the Reynolds number was relatively low. provided that flow separation occurs from a sharp leading edge.52.51 Predicted and measured CL." Note that. in the caption of Fig.25 and the freestream Reynolds number (based on the model length) is 15 X 106. (1992).53 [Wurtzler and Tomaro (1999)]. This is most likely associated with the exclusion of any vortex lift from the wing in the higha theory.388 Chap.] the wing. the flow separates quite close to the wing leading edge regardless of Reynolds number.8 CM 0.tot and CM for strake/wing configuration at = 0.51. The comparisons of the vortex burst locations on the F/A18 from experiments that spanned the range from 1/48scale models in water tunnels to actual flight tests of the fullscale aircraft are discussed in Komerath et al. Erickson (1982) concludes that "the water tunnel provides a good representation of the wake shed from a wing. The freestream Mach number for this flow is 0. However. At large angles of attack.4 I I —0. it is stated that the Reynolds number is for this test in a water tunnel." The two theoretical solutions along with a potential flow solution are compared with data [as taken from Lamar and Frink (1982)1 in Fig. Erickson. tunnel wall effects. are not primary. there may be an angleofattack range below the a for CL data are underpredicted. A flow visualization photograph of the vortices shed from the sharp leading edges of slender wings and wing leadingedge extensions (LEX5) or wingbody strakes at high angles of attack is presented in Fig. 7.2 1.tot 0. This lack of reattachment causes a large portion of theoretically available aerodynamic effects to be effectively lost to the configuration. Thus. 7.0 deg CL.4 0. tOt Figure 7. It was noted. and model support interference must be considered in each study.2 0. Thus.0 —0. showing that the effects of surface roughness. Also.2. freestream turbulence. the effects of compressibility. 7. The data were obtained in the Northrop Water Tunnel. 7 I Incompressible Flow about Wings of Finite Span 0. 7.8 1.4  CL. the separation point does not vary with Reynolds number. The solution was generated using the turbulence model of Spalart and Ailmaras (1992) .
.
Once the vortex breaks down. Note that. Although strakes (or leadingedge extensions) produce significant increases in the maximum angle of attack and in the maximum lift coefficient.06 0. . [From Stuart (1978).04 0. The higher trimmed angle of attack capability increases air combat effectiveness. The higher drag associated with the higher angle of attack is very useful in producing overshoots of the attacking aircraft.53 are two streamlines of the flow that are affected by the vortex. 7. near the leading edge of the wing. 7.04 SLEx/Swing 0. or bursting. the streamlines exhibit chaotic behavior. so that the reader can visualize the vortex core. the solution exhibits several interesting phenomena. 7 / Incompressible Flow about Wings of Finite Span 'T38 FSE (a) 2. and the FSE LEX is close to this value.8 Area ratio CG at 0. the streamlines form spirals as they are entrained by the vortex. occurs at an x coordinate between the engine inlet and the leading edge of the vertical tail. The surface marking a constantentropy contour is highlighted in Fig.08 0.] in the Cobalt60 code {Strang.Tomaro.0 F5A 0. Nevertheless. the load distribution may concurrently produce large noseup (positive) pitching moments at the high angles of attack.The F5E LEX with an area increase of 4. the original T38 wing planform did not include a leadingedge extension (LEX) at the wing root.54 Effect of wing leadingedge extentions on the max imum angle of attack and the maximum lift for the F5 family of aircraft: (a) configurations.0 0. As noted in Stuart (1978).53.6 I I FSE FSE x j1.54.53 are qualitatively similar. Note that vortex breakdown. Note that the computation of a turbulent separated flow is one of the greatest challenges to the computational fluid dynamicist at the time this fifth edition is written (which is during the summer of the year 2007).52 and of 7.0 1.0 0. (c) maximum lift.06 SLEx/Swing (b) Figure 7.2 x 0. and Grismer (1999)]. the flow fields of Figs.12 M=0. 7. Note that despite the differences in the flow conditions and in the geometry. 7.40 I 0. (b) maximum angle of attack. Northrop studies suggest a practical LEX upper size limit for any given configuration.8 II I 2.3 I I 0.390 Chap. The effects of the LEX in terms of maximum trimmed angle of attack and trimmed CL max are ifiustrated in Fig.4% of the wing reference area provides a CL max increment of 38% of the noLEX value.02 (c) 0. Also included in Fig.02 0.
although the crossflow effects will generate a thick viscous layer on the leeside. As a slender body of revolution is pitched through the angleofattack range from 0 to 90°. Large asymmetric loads can be induced on the body itself. Sketches of these four flow patterns are presented in Fig.At low angles of attack (0° <a <aSv).] 7. 7.55 Effect of angle of attack on the leeside flowfield. the normal force is approximately linear with angle of attack. and the side force is zero. At intermediate angles of attack (asy a <aAy) [i. For these vortexfree flows. They note that experimental results have shown that the vortexinduced side force can exceed the normal force. 7.55. there are four distinct flow patterns that reflect the diminishing influence of the axial flow component. [From Ericsson and Reding (1986).e.9 ASYMMETRIC LOADS ON THE FUSELAGE AT HIGH ANGLES OF ATFACK Ericsson and Reding (1986) noted that it has long been recognized that asymmetric vortex shedding can occur on bodies of revolution at high angles of attack. even at zero sideslip. the angle of attack is equal to or greater than the minimum for the formation of symmetric . the axialflow component dominates and the flow is attached.Sec..9 / Asymmetric Loads on theFuselage at High Angles of Attack 391 No vortices (attached flow) <aAV Symmetric vortices Asymmetric vortices a Vortex wake Figure 7.
7 / Incompressible Flow about Wings of Finite Span Vortex breakdown Higher.1 Asymmetric Vortex Shedding In the angleofattack range where the vortex shedding is asymmetric (QAV a < auv). "This side force is the re sult of surface pressure imbalances around the forebody of the aircraft caused by an asymmetric forebody boundary layer and vortex system. 7.As observed by Cobleigh (1994). At separation. 7. as shown in Fig. the axial flow component is still sufficient to produce steady vortices. but there continues to be no side force. the vortex pattern is asymmetric. inboard vortex Resulting force Figure 7. producing a side force and a yawing moment.56. The normal force increases nonlinearly in response to the added vortex lift. outboard vortex Section AA Later separation corresponds to lower. the boundary layer on each side of the forebody separates at different locations as shown in Fig.The forces on the forebody are generated primarily by . In this scenario. corresponding vortex sheets are generated which roll up into an asymmetrically positioned vortex pair. 7.392 Chap. However. [Taken from Cobleigh (1994).9. the cross flow separates and generates a symmetric pair of vortices.] vortices aSV but less than that for the formation of asymmetric vortices (aAV)]. inboard vortex results in earlier vortex breakdown Earlier separation corresponds to higher.56 Sketch of asymmetric vortices shed from aircraft forebody for aAV a < aUV. outboard vortex results in later vortex breakdown Lower.55c. even at zero sideslip.
NB grit S2 40 30 20 —0.10 Cf0 Figure 7. after which the asymmetric yawing moment coefficient decreased. For the aircraft with these transition strips.00 0. the higher. For the basic aircraft.9 I Asymmetric Loads on the Fuselage at High Angles of Attack 80 393 70 —0— Unmodified forebody 60 —a— FB and NB transition grit —0— a 50 Si strake.040 occurred to Fig. from Cobleigh (1994).The asymmetric forces decreased rapidly with alpha.10 —0.57. Since the center of gravity of the aircraft is well aft of the forebody. Cobleigh (1994) reported.57. the asymmetric forces started to build up.05 strake. A method for determining the yawing moment asymmetry from flight data was developed and an analysis of the various configuration changes completed. is reproduced in Fig. as taken from Cobleigh (1994). FB and NB grit 0. a sizeable yawing moment asymmetry develops.] the boundary layer and to a lesser extent by the vortices. and boundarylayer control. 7. the peak asymmetry increased to of —0. more outboard vortex corresponds to the boundary layer which separated earlier. including physical surfaces. the use of both a strake and boundarylayer transition strips produced significant yawing moments which persisted over a broad angleofattack range. The data indicated that the side forces actually increased when grit strips were used. Fig 7." There are a variety of techniques that can be used to modify the asymmetric vortex pattern. the largest asymmetry occurred in the alpha range from 58° to 61°. The use of a strake along with blunting of the nose tip delayed the initiation of the yawing moment asymmetry up to an AOA of 55° (refer of —0. becoming relatively small by an angle of attack of 66°. transition grit strips were installed on both sides of the forebody (FB) and along the sides of the nose boom (NB). In summary. "Several aerodynamic modifications were made to the X31 forebody with the goal of minimizing the asymmetry. 7. depending on their proximity to the forebody surface.57). 7. beginning at an angle of attack of 48° reaching a peak value of (the yawing moment coefficient at zero sideslip) of —0. more inboard vortex corresponds to a boundary layer which separates later and.56 shows a typical asymmetrical arrangement where the lower. As evident in the data presented in Fig. In an attempt to reduce these asymmetries. Although the largest asymmetry began to build at the same angle of attack as for the basic aircraft (an AOA of 48°). such as blowing or suction at the surface. The baseline aircraft were .063 at an angle of attack of approximately 57°.078.05 0. The peak asymmetric yawing moment coefficient at an angle of attack of 60°. such as strakes. The suction generated by the more persistent boundary layer and the closer vortex combine to create a net force in their direction. conversely.57 Variation of yawing moment asymmetry with angle of attack for an X31 for a 1g maneuver. 7.Sec. The yawing moment asymmetry for an X31 during slow decelerations (essentially ig) to high AOA (angle of attack) conditions.
Installing longitudinal forebody strakes and rounding the sharp nose of the aircraft caused the yawing moment asymmetry magnitude to be reduced. At that 165 knots condition." An article by Dornheim (1994) included the paragraph. the yaw moment from the vortices was estimated to be 2. The problem can be cured by use of nose strakes or other types of body reshaping that cause the forebody vortices to be shed symnietrically. and the X31 should be able to fly at least 300 knots.9. which show the vortexinduced yawing moment to exceed." . where the 6g structural limit comes into play. "Many modern highperformance fighter aircraft have mission requirements which necessitate rapid maneuvers at high angles of attack and large angular rotation rates. the interaction with the asymmetric vortex pattern can produce rolling moments. the dynamic and timedependent effects of these nonlinear flow characteristics contribute significantly to the behavior and maneuvering capability of aircraft. The transition strips and strakes made the asymmetry characteristic of the aircraft more repeatable than the clean forebody configuration. Although no geometric differences between the aircraft were known.5 times more powerful than the thrust vectoring. vortex shedding." 7.2 Wakelike Flows < 90°) the axialflow component has less and less influence. "The X3i's main problem is directional instability from asymmetric nose vortices around 60 deg. so that the vortex shedding becomes unsteady. 7 I Incompressible Flow about Wings of Finite Span found to have significant asymmetries above 45° angle of attack with the largest asymmetry typically occurring around 600 angle of attack. In this angleofattack range. starting on the aft body and progressing toward the nose with increasing angle of attack. and as powerful as therudder itself would be at 10 deg. and vortex lag effects (Fig. the leeside flow resembles the wake of a twodimensional cylinder normal to the flow. During extreme multipleaxis maneuvering conditions. The mean side force decreasestoward zero inthis angleofattack range. AOA. which caused an unintended departure in flight test. by an order of. magnitude. Applying symmetrical bound arylayer transition strips along the forebody sides increased the magnitude of the asymmetry and widened the angleofattack range over which the largest asymmetry acted. Ericsson and Reding (1986) note that "the seriousness of the problem is illustrated by the windtunnel test results for the Fill. the available control capability through full rudder deflection. Fmaily.394 Chap. ship 2 consistently had larger yawing moment asymmetries than ship 1. before the remaining vortices can overpower the thrust vectoring at high AOA.AOA." For axisymmetric bodies with wings or fins. Under these flow conditions. at very high angles of attack 7. A subtle reshaping of the nose and adding vortex control strakes has tamed the vortices.58).10 FLOW FIELDS FOR AIRCRAFT AT HIGH ANGLES OF A1TACK As noted by Mendenhall and Perkins (1996). the vehicle may operate in a flow regime in which the aerodynamic characteristics are dominated by unsteady nonlinear effects induced by flow separation. the effectiveness of the vertical tail and rudder to control the yawing moment falls off as the angle of attack increases. Typically. because the vertical tail lies in the wake of the wing and of the fuselage. 7.
(1992)] and inflight flowfield measurements [e. 7.10 I Flow Fields for Aircraft at High Angles of Attack Vertical tail vortex 395 Horizontal tail vortex Wing trailing vortex Strake vortex Nose vortex Figure 7. have been discussed individually in the previous sections of this chapter.A sketch illustrating the location of the forebody vortex core.. and the nonlinear forces and moments on the vehicle depend on the time history of the motion and the wake development. Numerical predictions [e. of the LEX vortex.] . For example.59.The HARV was a singleplace F18 aircraft built by McDonnell Douglas and Northrop Aircraft and was powered by two General Electric F404GE400 turbofan engines with afterburners. The aircraft featured a midwing with leading." The vortices associated with flow over highly swept wings (such as delta wings).Sec. as taken from Del Frate and Zuniga (1990). Leadingedge extensions extend from the wing roots to just forward of the windscreen.g.58 Vortex wakes near a maneuvering fighter.and trailingedge flaps. 7. with the flow over leadingedge extensions (LEXs) and with the flow over the nose region at high angles of attack. is reproduced in Fig. vorticity shed from the nose of the vehicle will pass downstream to influence the loads on the wing and tail surf aces.] Mendenhall and Perkins continue. "The presence of the vortex wake introduces memory into the flow problem. Del Frate and Zuniga (1990)] have been presented for a modified F18 aircraft called the highalpha research vehicle (HARV). The vortexinduced loads depend on the motion of the vehicle and the vortex wake during the time it takes the vorticity to be transported from the nose to the tail. The pattern is for an angle LEX/forebody vortex interaction Forebody vortex core Figure 759 Wingtip view of fOrebody vortex system.. [Taken from Mendenhall and Perkins (1996). Cummings et al.g. and of the LEX/forebody vortex interaction. [Taken from Del Frate and Zuniga (1990).
7.000 to 10. In an attempt to classify these vehicles and their resulting aerodynamics. Visible evidence of the vortex core breakdown is a stagnation of flow in the core with a sudden expansion in the core diameter. the forebody vortex cores extend downstream until they interact with the LEX vortices. The tried and true methods for designing wings often can fall short when applied to either very large or very small aircraft. This interaction results in the forebody vortex cores being pulled beneath the LEX vortices and then redirected outboard. and the X31 enhanced fighter maneuverability aircraft" were presented by Fisher et al.396 Chap. Pressure distributions obtained on the X29 windtunnel model at flight Reynolds numbers showed excellent correlation with the flight data up to a = 50°. These definitions are used to classify vehicles with similar aerodynamic and technical issues involved in their design.7° and for zero sideslip. Grit strips on the forebody of the X31 reduced the randomness but increased the magnitude of the asymmetry. "These vortex cores are generated at moderate to high angles of attack by the shape of the forebody and the sharp leading edge of each LEX. the X29 forwardswept wing aircraft. 7 / Incompressible Flow about Wings of Finite Span of attack of 38." 7. The detrimental effect of a very sharp nose apex was demonstrated on the X31 aircraft. Above a = 50° the pressure distributions for both flight and wind tunnel became asymmetric and showed poorer agreement. Similarly.000 2. Reynolds number effects were evident on the forebodies at high angles of attack. possibly because of the different surface finish of the model and aircraft. As noted by Del Frate and Zuniga (1990).3 Classification of UAVs [Wood (2002)] Weight (lbs) <1 1 to 2.11 UNMANNED AIR VEHICLE WINGS As was mentioned in Chapter 6. unmanned air vehicles (UAV) of various sizes have led to a revolution in airfoil and wing designs in recent years.000 UAVType Wing Span (ft) Micro Meso Macro Mega <2 2 to 30 3Oto 150 >150 >10.000 . The LEX vortex cores are tightly wound and extend downstream until experiencing vortex core breakdown. as Reynolds number effects change the basic characteristics of the aerodynamics of these vehicles. The correlation between flight and wind tunnel forebody pressure distribution for the F18 HARV were improved by using twin longitudinal grit strips on the forebody of the windtunnel model. Nose strakes were required to reduce the forebody yawing moment asymmetries and the grit strips on the flight test noseboom improved the aircraft handling qualities. "On all three vehicles." "Lessons learned from comparisons between groundbased tests and flight measurements for the highangleangleofattack programs on the F18 High Alpha Research Vehicle (HARV). (1998).60). TABLE 7. Wood (2002) devised four definitions for the size of unmanned air vehicles (see Table 7.3 and Fig.
(2003). aspect ratio no longer is the dominant factor in creating drag (see Fig. as the size of the vehicle decreases (and as weight and Reynolds number also decreases).Sec.] As was discussed in Chapter 6 for laminarflow airfoils.11 / Unmanned Air Vehicle Wings 397 106   102 102  — Weight (ibs) 10° 10° Span (ft) I I 106 108 Figure 7. and Reynolds number. [From Wood (2002).60 Classification of UAVs by weight.] .61)—skin friction drag becomes more 102 102 1o4 Initial Cruise Mass (kg) Figure 7. span.61 The influence of wing aspect ratio on aircraft range as a function of mass (all other factors kept constant). [From Drela et al. 7. 7.
At subsonic speeds. (2003). 7. The planform of a wing is defined by the aspect ratio. more massive root sections are needed to handle the higher bending moments. and taper ratio. and the resulting wings can be quite unusual when compared with typical aircraft. for a given wing area.12 SUMMARY The shape of the wing is a dominant parameter in determining the performance and handling characteristics of an airplane. (2003)].398 Chap. the choice of aspect ratio requires a trade study in which the designer may chose to emphasize one criterion at the expense of another (see the discussion relating to transport aircraft in Chapter 13). as shown in Fig. the use of sweepback may also lead to a growth in wing weight. [From Hammons and Thompson (2006). the liftdependent drag can be reduced by increasing the aspect ratio. A highaspectratio wing produces a highliftcurve slope. Thus. while at small size. 7 / Incompressible Flow about Wings of Finite Span I Q Figure 762 Unusual UAV configuration. . This is related to the well known squarecube law of aircraft design: The weight of the wing varies with the cube of the wing span while the area of the wing varies with the square of the wing span [see for example. Couple the squarecube law with the various amounts of laminar and turbulent flow the wing can experience. which is useful for takeoff and landing but has a weight penalty since. drag is more important so that the increasing friction drag with aspect ratio offsets the reduction in induced drag" [Drela et at. Drela et al.62. McMasters (1998)]. In fact. as the aspect ratio increases. hence the low aspectratio designs common for micro UAVs [Drela et al. a swept wing has a greater structural span than a straight wing of the same area and aspect ratio. 7.] important.This is one of the reasons these wings have such largely different design characteristics.] The highaltitude aircraft also fly at low Reynolds numbers but require fairly high weights in order to carry the solar panels and batteries required for propulsion. Furthermore. sweep. (2003) found that for each aspect ratio examined there was an aircraft size (weight) that would maximize the range: "this is because. the wing is proportionately heavier for a large vehicle than for a smaller vehicle. As a result. The highaspect ratio aircraft once again becomes mOre efficient.
69. Compare the lift coefficient. the downwash velocity the induceddrag coefficient (CDV).000 N and cruises in level flight at 185 km/h at an altitude of 3.29). Assume that the lift coefficient is a linear function of the angle of attack and that = —1.0.37) to calculate the velocity induced at some point C(x.1.0? 7. Consider the case where the spanwise circulation distribution for a wing is parabolic.700 N and cruises in level flight at 300 km/h at an altitude of 3000 m. and a taper ratio of 0.37) to calculate the velocity induced at some point C(x. What would the angle of attack have to be to generate this lift coefficient for a wing whose aspect ratio is 6. the wingtip sections are subjected to higher lift coefficients and tend to stall first (see Fig.38a).Problems 399 Tapering the wing reduces the wingroot bending moments.20). an aspect ratio of 7.18). calculate the value of the circulation in the plane of symmetry (F0).5. If the load distribution is elliptic. Following Example 7. (c) 0. the lift coefficient is 1.0.4. (b) 0.43a). that is.6. PROBLEMS 7.washout) may be used either to reduce local loading or to prevent tip stalling.The wing has a surface area of 16. Consider an airplane that weighs 14. .1. a wing of infinite span) is at an angle of attack of 4°. 7. y.3 m2.no geometric twist) which has a NACA 0012 section and an aspect ratio of 7. use a fourterm series to represent the load distribution. Consider a planar wing (i..2.0 km.e. Assume that the lift coefficient is a linear function of the angle of attack and the airfoil section is a NACA 2412 (see Chapter 6 for the characteristics of this section). 7. what is the relation between the F0 values for the two distributions? What is the relation between the induced downwash velocities at the plane of symmetry for the two configurations? 7. that is. 7.e. 7.5°. Using equation (7.1.2. progressively reducing the incidence of the local section through the use of geometric twist from the root to the tip (i.6. Following Example 7. If the total lift generated by the wing with the parabolic circulation distribution is to be equal to the total lift generated by a wing with an elliptic circulation distributIon..0. calculate the angle of attack at which a wing whose aspect ratio is 7.2°. y. The wing has a surface area of 17. and the spanwise lift distribution for taper ratios of (a) 0. 0) by the vortex filament AB in a planar wing. z) by the vortex filament AB (shown in Fig.3.5°.7.e.52. Thus.5 would have to operate to generate the same lift coefficient.. and (d) 1. 7.5. Use equation (7. The incidence of the root section is + 1. Consider an airplane that weighs 10. While taper reduces the actual loads carried by the outboard sections. there is a geometric twist of —3° (washout). derive equation (7. and the effective angle of attack (ae). derive equation (7. Use equation (7. When a GA(W)1 airfoil section (i. since the inboard por tion of the wing carries mOre of the wing's lift than the portion near the tip.4. Thus.0 m2 and an aspect ratio of 6. the geometric angle of attack (cr). the induceddrag coefficient. the incidence of the tip section is —1. use a fourterm series to represent the load distribution and calculate (a) The lift coefficient (CL) (b) The spanwise load distribution (Cl(y)/CL) (c) The induced drag coefficient (CDV) (d) The geometric angle of attack (a) 7.
g.e.2.23b. Corda S. assume that the wing is planar.8. the threequarterchord line.44.62) to calculate the induced drag for a flat delta wing with sharp leading edges. New York: Dover Anderson SD.11. the threequarter chord line.09 per radian.5 (i.5. it is a flat plate). 7. that computed iii Example 7. and the trailing edge have different sweep angles.Vidal Ri 1955.The delta wing has an aspect ratio of 1. Find the wing loading of the aircraft as a function of wing span. Aeron.14. For this airfoil section a02 —0. Aircraft 17:898—904 Bartlett GE. 7.5. 7. and the quarter chord is swept 45°. calculate the lift coefficient for a swept wing. and the trailing edge all have the same sweep..44? 7.. von Doenhoff AB. 7. is proportional to the cube of the wing span (this is the squarecube law). Using these relationships explain why very large aircraft (like the Boeing 747 or the Airbus 380) have to fly with very large wing loadings. Since the taper ratio is not unity. Following the \'LM approach used in Example 7. = 0. 7. 1949. 1980.15. What is the sweep angle of the leading edge? The fact that the quarterchord and the threequarterchord lines have different sweeps should be taken into account when defining the coordinates of the horseshoe vortices and the control points.2. and an uncambered wing.2. calculate the lift coefficient for a swept = ce). 7.400 7.5. b. Following the VLM approach used in Example 7. 45°.Van Wie DM. Set.e.e. The airfoil section (perpendicular to the quarter chord) is a NACA 64A1 12.. Numerical lifting line theory applied to drooped leadingedge wings below and above stall.94° and = 6.e. Use equation (7. the leading edge. 7.. a taper ratio of unity (i.9. 7. 22:517—533. Assume that the wing area of an airplane is proportional to the square of the wing span and the volume. Following the VLM approach used in Example 7. Compare the solution with the data of Fig. calculate the lift coefficient for a delta wing whose aspect ratio is 1. 7. Following the VLM approach used in Example 7.2)? Is this consistent with our knowledge of the effect of aspect ratio (e. the quarterchord line.55. and the taper ratio 0. an uncambered section. How do the calculated values for the lift coefficient compare with the experimental values presented in Fig. Fig. section (i.000 lbs with a wing span of 75 It. The wing has an aspect ratio of 8. Use equation (7.10.61) to calculate the lift coefficient as a function of the angle of attack for a flat delta wing with sharp leading edges. J. 7 / Incompressible Flow about Wings of Finite Span Calculate the downwash velocity at the CP of panel 4 induced by the horseshoe vortex of panel 1 of the starboard wing for the flow configuration of Example 7. The wing has an aspect ratio of 5. Compare the solution with the data of Fig. For purposes of applying the noflow boundary condition at the control points.13. Prepare a graph of the lift coefficient. what wing span would you need to add an additional 10% to the weight? REFERENCES Abbott IH. Theory of Wing Sections.5. calculate the lift coefficient for the forward swept wing of Fig. and thus the weight. Since the taper ratio is unity.47. the leading edge. If you wanted to redesign an existing aircraft that weighed 92.588 . 7. How does this compare with that of Fig.12.10)? 7. J. The quarter chord is swept forward 45°. The delta wing has an aspect ratio of 1.5 ci).2.This should be taken into account when defining the coordinates of the horseshoe vortices and the control points. Experimental investigations of influence of edge shape on the aerodynamic characteristics of low aspect ratio wings at low speeds. the aspect ratio is 3. Chap.23? 7. the quarterchord line. HOw does the lift coefficient for this aspect ratio (8) compare with that for an aspect ratio of 5 (i. a taper ratio of 0.2..
Pulley CT. Studies of various factors affecting drag due to lift at subsonic speeds. Aerodyn. Hanley RJ. Conf. F16 MATV. 140:46—47 Dornheim MA. Hall RM. 1992. Sci. Ed. Meas. Presented at Appi. 89—2222.. Dayton. Epstein AH. Drela M. Future fighter tcchnologies. Modification of the Douglas Neumann program to improve the efficient of predicting component and high lift charactersitics. Schwartz RJ. X31. 1994. The calculation of aerodynamic loading on surfaces of any shape. ConL. Vortex flow correlation.. Meet. RTOMPAVT111. Ground Test. Experimental investigation of the F/A. DC: AIAA Erickson GE. Kim JM. J. Veh. J. Presented at Intl. Meet. Hemsch MH. 2004. Aircraft 7:574—576 Kalman TP. Zuniga FA. Presented at Congr. WA Falkner VM. AIAA Pap. 2006. Tech. 1989. Spanwise distribution of induced drag in subsonic flow by the vortex lattice method. Albuquerque. Presented at Appi. DW. CO Cummings RM. Seattle. 2006—1265. Air Space Symp. Rodden WE 1970.61. 1986. Chaderjian NM. I. Rizk YM. Presented at Adv. Schiff LB. Reynolds number effects at high angles of attack.The role of size in the future of aeronautics.. Nielsen JN. Reno. DeiFrate JH. NASA CR 3020 Cubleigh BR 1994. Note D3584 Herbst WB. Giesing JP. 1978. 1980. part I: spatial characteristic. Schreiner JA. 941803. NavierStokes prediction for the F18 wing and fuselage at large incidence. Czech Republic. In Tactical Missile Aerodynamics. Presented at NATO RTO Appl. 1992. Aircraft 29:565—574 Del Frate JH. Aircraft 8:406—413 Komerath NMS.. Highangleofattack yawing moment asymmetry of the X31 aircraft from flight test. Tech. 90—0231. Aviat. 1978. Protz JM. A numerical investigation of novel planforms for micro Presented at AIAA Aerosp. 1971. NV Week Space Tech. Johnson FT. Aeron. Effects of boundary layer formation on the vortical flow above a delta wing. AIAA Pap. AIAA Pap. J. Washburn KB. Prague. A general panel method for the analysis and design of arbitrary configurations in incompressible flows. 1966. 2003. Load distribution on a closecoupled wing canard at transonic speeds.. OH Ericsson LE. 1998. Hall RM. 1990. Aircraft 17:561—566 Hummel D. Reno. Flow over a twintailed aircraft at angle of attack. 1982. Application of the doubletlattice method to nonpianar configurations in subsonic flow. Presented at AIAA Aerosp.References 401 Bristow DR.. AIAA Pap. 98—2879. Aircraft 29:413—420 . Banks. 28th. Intl. 20th. Conf. Aircraft 15:234—239 Hammons C. NASA Tech. 1980. Reding JP. Spec. J. WA Erickson GE. Sci. NV Henderson WP.. Meet. Washington. Inflight flow field analysis on the NASA F18 high alpha research vehicle with comparison to the ground facility data. Aerodyn. 2003—2902. Wahls RA. 82—6. Seattle. AIAA Pap. Asymmetric vortex shedding from bodies of revolution. Colorado Springs. Banks DW. Liou SG. Coun. Sci. 1943. Grose GG. ICAS Pap. J. Rodden WE Giesing JP. NASA CR 3079 Kalman TP. FIA18 HARV explore diverse missions. NM Gloss BB. ARC R&M 1910 Fisher DF. 7th. Paper 301. Cobleigh BR. AIAA Pap.18 vortex flows at subsonic through transonic speeds. Thompson DS.
Aerodyn. 1969. Perkins SC. 1960. 1992. AIAA Pap. 1951. Charts for predicting the subsonic vortexlift characteristics of arrow. 2000. Frink NT 1982. 1956. 1971a. Sci. 1975.Aeron. 1971. NASA Tech. Morris CEK. 1950. Meet. 1958.. NASA Note D6142 Margason RJ. 1999. New York: Dover Rasmussen ML. Aircraft 8:193399 Poihamus ED. Sub sonic panel methods—a comparison of several production codes. Some developments in boundary layer research in the past thirty years. Aircraft 8:623—629 McMasters JH. J. New Orleans. 14th. Spanwise lift distribution on a wing from flowfield velocity surveys. Vortexlattice Fortran program for estimating subsonic aerodynamic charactersitics of complex planforms. ARC R&M 3186 Phillips WF. J. The rolling up of the vortex sheet and its effect on the downwash behinds wings. Lowspeed wind tunnel tests on a series of uncambered slender pointed wings with sharp edges. ARC T&M 2884 Orloff KL. Presented at App!.Aircraft 19:639—646 Margason RJ. delta. 38th AIAA Pap.Conf.andAeromechanics. Presented at AJAA Aerosp. J. Aircraft 17:875—882 Peckham DH. A erodynamik des Flugzeuges. Kjelgaard SO. J. Aeron.. 1996. 96—2433. 1971. 1921.. A oneequation turbulence model for aerodynamics flows. 1980. Sellers WL. Aircraft 36:340—348 Robinson A. NACA Tech. Some trim drag considerations for maneuvering aircraft. NV Spreiter JR. Liftingline theory for arbitrarily shaped wings.Tietjens 0G. LA Multhopp H. 64:6479 Schlichting H. J. Reno. Soc. Shields EW. Lamar JE. 30th AIAA Pap. Laurmann JA. 1947.402 Chap. Snyder DO. AIAA Pap. 1957. Walkey KB. NV McKinney LW. NASA Tech. Presented at AIAA Aerosp.Applied Hydro. Sacks. Aircraft Design 1:217—242 Mendenhall MR. Dollyhigh SM. Roy.. Sci.1998. Ailmaras R. Aerodynamic features of designed strakewing configurations. 1971b. 7 / Incompressible Flow about Wings of Finite Span Lamar JE. Advanced configurations for very large transport airplanes. Reno. Gloss BB. 1985. 85—0280. Method for calculating the lift distribution of wings (subsonic lifting surface theory). J. Predicted higha aerodynamic characteristics of maneuvering aircraft. Wing Theory. Smith DE. 18:21—32 . Application of modern hydrodynamics to aeronautic& NACA Report 116 Prandtl L. Experimental and calculated characteristics of three wings of NACA 64210 and 65210 airfoil sections with and without washout. J. Meet. Application of liftingline theory to systems of lifting surfaces. Note 1422 Spalart PR. Sci. AH. NV Polhamus ED. Note D7921 Lamar JF. Sci. Cambridge: Cambridge University Press Schlichting H. Presented at AIAA Aerosp. Subsonic aerodynamic characteristics of interacting lifting sur faces with separated flow around sharp edges predicted by vortexlattice method. and diamond wings.Truckenbrodt E. Meet. Predictions of vortexlift characteristics by a leadingedge suction analogy. Note D6243 Prandtl L. 2000—0653. Reno. Berlin: SpringVerlag Sivells JC.
unstructured EulerlNavierStokes flow solver. 1964. Monterey. 99—0786.153 Weber J. Sci. Lowspeed tests on 45deg sweptback wings. Comp. VA Wurtzler KE. Veh. Grismer MJ. 15:7282 Strang WZ. 37th. Meet. Unsteady aerodynamics of aircraft maneuvering at high angle of attack. User's Group Meet. 2002—3494. The defining methods of Cobalt60: a parallel. 1978. 10:135—139. 1995. The Aeron. Presented at Tech. Computational and physical aspects of vortex breakdown on delta wings. 1976. AIAA Pap.References 403 Stanbrook A. Meet.. 1958. Quart. Air Intern. NV Voight W. Brebner GG. implicit.. A discussion of aerodynamic control effectors for unmanned air vehicles. 95—0585. 2002. NV Stuart WG. CA .Tomaro RF. Presented at AIAA Aerosp. Reno. Portsmouth. Squire LC. DC: AIAA Visbal MR. part I: pressure measurements on wings of aspect ratio 5. Washington.. Workshop Unmanned Aerosp. Reno. 1999. ARC R&M 2882 Wood RM. Presented at AIAA Aerosp.. Gestation of the swallow. Northrop FS Case Study in Aircraft Design.AIAA Pap. Sci. Presented at DoD High Perf. Conf. 33rd AIAA Pap. Possible types of flow at swept leading edges. Tomaro RE 1999.
For our purposes.10)] provides the relation between the pressure distribution about an aircraft and the local velocity changes of the air as it flows around the various components of the vehicle.. the gas obeys the equation of state): p = pRT We will assume that the gas is also calorically perfect. This leads to discrepancies between the actual aerodynamic forces and those predicted by incompressible flow theory. These specific heats are discussed further in the next section. Bernoulli's equation [equation (3.8 DYNAMICS OF A COMPRESSIBLE FLOW FIELD Thus far. However. At low flight Mach numbers (e.e. 404 . the Mach number is the parameter that determines the extent to which compressibility effects are important. (1. changes in the local air density also affect the magnitude of the local static pressure.5)..e. as the flight Mach number increases. that is. air will be assumed to behave as a thermally perfect gas (i. we have studied the aerodynamic forces for incompressible (constant density) flows past the vehicle.10) and of the gas are constant. the specific heats. The term perfect gas will be used to describe a gas that is both thermally and calorically perfect.3 to 0.g. below a freestream Mach number of approximately 0. The purpose of this chapter is to introduce those aspects of compressible flows (i. flows in which the density is not constant) that have applications to aerodynamics Throughout this chapter..
equivalently. consider Ue = Ue(V. where the effects of viscosity and heat conduction are important. 7').2) becomes dUe = (8. As has been true in previous chapters. for any simple substance. equation (8.3) In equation (8. + äv T where the subscript is used to designate which variable is constant during the differentiation process.Sec. the density.1 I Thermodynamic Concepts 405 Even though we turn our attention in this and in the subsequent chapters to compressible flows. by the chain rule of differentiation. we note that. It follows that = Lte2 11e1 dT (8. dUe dT dv (8. one obeying equation (1. and hence depends on the temperature alone. and (2) the external flow. it may be shown that for a thermally perfect gas. we may still divide the flow around the vehicle into (1) the boundary layer near the surface. the specific internal energy is a function of any other two independent fluid properties. 8.1) As an extension of our discussion of fluid properties in Chapter 1. where the effects of viscosity and heat conduction can be neglected.3) we have introduced the definition that (8tte which is the specific heat at constant volume. Then.2) which is equivalent to saying that the internal energy of a perfect gas does not depend on the specific volume or. 8.1. let us consider a system in which there are no changes in the kinetic and the potential energies. the inviscid flow is conservative. that is.1 Specific Heats For simplicity (and without loss of generality). 8. Thus.1 THERMODYNAMIC CONCEPTS Having reviewed the fundamentals of thermodynamics in Chapter 2 and having derived the energy equation therein. Equation (2. From the principles of thermodynamics. let us turn our attention to the related aerodynamic concepts.23) becomes Sq — p dv = dUe (8.4) . Thus.10).
5) is valid even if the process is not one of constant volume. Thus.406 Chap. the terms Ue and pv appear as a sum. 8 / Dynamics of a Compressible Flow Fie'd is constant over a wide range of conditions.T) (8.1) and collecting terms yields + vdp = dli (8.5) is valid for any simple substance undergoing any process where can be treated as a constant. we can write h h(p. the name "specific heat" can be misleading.6) where h is called the specific enthalpy.6) into equation (8. it follows that Ii is also a function of temperature only for a thermally perfect gas.7) which is the first law of thermodynamics expressed in terms of the enthalpy rather than the internal energy. T and equation (8. one sees that is only directly related to the heat transfer if the process is one in which the volume remains constant (i. Substituting the differential forip of equation (8. Thus. Physically.e.1). = (8. Since any property of a simple substance can be written as a function of any two other properties.1717 lbrn°R The assumption that is constant is contained within the more general assumption that the gas is a perfect gas.9) becomes . since both and p/p are functions of the temperature only. can be treated as a constant.6 (or kg K) = 0. Substituting equation (8.9) From the definition of the enthalpy. is the proportionality constant between the amount of heat transferred to a substance and the temperature rise in the substance held at constant volume. dv 0). equation (8. Thus. and so it is convenient to define a symbol for this sum: (8.5) Since and IT are both properties of the fluid and since the change in a property between any two given states is independent of the process used in going between the two states.8) Thus. For air below a temperature of approximately 850 K and over a wide range of pressure. for a perfect gas. In analyzing many flow problems. Thus. The value for air is Experimental evidence indicates that for most gases..3) into equation (8. equation (8. dh = + (8. = 717.
2 Additional Relations A gas which is thermally and calorically perfect is one that obeys equations (1. the kinetic theory of gases shows that n+2 n where n is the number of degrees of freedom for the molecule. K An argument parallel to that used for shows that equation (8.5).Sec. which is equal to 0. Thus.14) Let us introduce another definition. provided that the temperature extremes in a flow field are not too widely separated.7 (8.1. we note that the term specific heat is somewhat misleading. since is only directly related to the heat transfer if the process is isobaric. and R.15) cv For the most simple molecular model. Extremely complex molecules. the value of (or J/kg . depends on the composition of the and its pressure and temperature. for a monatomic gas. 8. such as helium. In general. n = 5 arid y = 1. Again.10) (8. is 1004.400. K). Again. oxygen. n = 3 and y = 1. such as Freon or tetrafluoromethane.10). 8. In many treatments of air at . we write (ah'\ dli due d (p\ (8. and (8.13) rn/kg. In such a case.13) is valid for any simple substance undergoing any process where can be treated as a constant. From (8.2404 Btu/lbm °R. have large values of n and values of y which approach unity. the perfectgas law. For a diatomic gas.11) which is the specific heat at constant pressure.13).1 I Thermodynamic Concepts 407 dh We have introduced the definition dT = dT (8. and the knowledge that h depends upon T alone. is essentially independent of temperature and of pressure over a wide range of conditions. one for the ratio of specific heats: cp y (8. we conclude h2 — = h1 = j' that. or air. such as nitrogen.667. can be treated as a constant so that Ah = For air below a temperature of approximately 850 K. It follows that (812) Ji Experimental evidence indicates that for most gases. there is a simple relation between the definitions of and Ii.
408
Chap. 8 I Dynamics of a Compressible Flow Field
high temperature and high pressure, approximate values of y (e.g., 1.1 to 1.2) are used to
approximate "realgas" effects. Combining equations (8.14) and (8.15), we can write
=
yR
y—l
8.1.3
and
R
(8.16)
Second Law of Thermodynamics and Reversibility
The first law of thermodynamics does not place any constraints regarding what types of processes are physically possible and what types are not, providing that equation (2.22) is satisfied. However, we know from experience that not all processes permitted by the first law actually occur in nature. For instance, when one rubs sandpaper across a table, both the sandpaper and the table experience a rise in temperature. The first law is satisfied because the work done on the system by the arm, which is part of the surroundings and which is, therefore, negative work for equation (2.23), is manifested as an increase in the internal energy of the system, which consists of the sandpaper and the table. Thus, the temperatures of the sandpaper and of the table increase. However, we do not expect that we can extract all the work back from the system and have the internal energy (and thus the temperature) decrease back to its original value, even though the first law would be satisfied. If this could occur, we would say the process was reversible, because the system and its surroundings would be restored to their original states. The possibility of devising a reversible process, such as the type outlined previously, cannot be settled by a theoretical proof. Experience shows that a truly reversible process has never been devised. This empirical observation is embodied in the second law of thermodynamics. For our purposes, an irreversible process is one that involves viscous (friction) effects, shock waves, or heat transfer through a finite temperature gradient. Thus, in regions outside boundary layers, viscous wakes, and planar shock waves, one can treat the flow as reversible. Note that the flow behind a curved shock wave can be treated as reversible only along a streamline. The second law of thermodynamics provides a way to quantitatively determine
the degree of reversibility (or irreversibility). Since the effects of irreversibility are
dissipative and represent a loss of available energy (e.g., the kinetic energy of an aircraft wake, which is converted to internal energy by viscous stresses, is directly related to the aircraft's drag), the reversible process provides an ideal standard for comparison to real processes. Thus, the second law is a valuable tool available to the engineer. There are several logically equivalent statements of the second law. In the remainder of this text we will usually be considering adiabatic processes, processes in which there is no heat transfer. This is not a restrictive assumption, since heat transfer in aerodynamic problems usually occurs only in the boundary layer and has a negligible effect on the flow in the inviscid region. The most convenient statement of the second law, for our purposes, is
ds
0
(8.17)
for an adiabatic process. Thus, when a system is isolated from all heat exchange with its
surroundings, s, the entropy of the system either remains the same (if the process is
Sec. 8.1 / Thermodynamic Concepts
409
reversible) or increases (if it is irreversible). It is not possible for a process to occur if
the entropy of the system and its surroundings decreases. Thus, just as the first law led to the definition of internal energy as a property, the second law leads to the definition of entropy as a property. The entropy change for a reversible process can be written as
.5q = Tds
Thus, for a reversible process in which the only work done is that done at the moving boundary of the system,
T
=
dUe
+
dV
(8.18)
However, once we have written this equation, we see that it involves only changes in properties and does not involve any pathdependent functions. We conclude, therefore, that this equation is valid for all processes, both reversible and irreversible, and that it applies to the substance undergoing a change of state as the result of flow across the
boundary of an open system (i.e., a control volume) as well as to the substance comprising a closed system (i.e., a control mass). For a perfect gas, we can rewrite equation (8.18) as
ds =
This equation can be integrated to give
S2
dT
dv + R—
— Si = cvlntL
J
p1 T2
(8.19a)
Applying the equation of state for a perfect gas to the two states,
V2 —
v1p2p2T1
equation (8.19a) can be written
—
Pi
=R
T1 [k—)
(8.19b)
J P2 —J
Equivalently,
—
=
—
P21
JPiJ
(8.19c)
Using the various forms of equation (8.19), one can calculate the entropy change in
terms of the properties of the end states. In many compressible flow problems, the flow external to the boundary layer undergoes processes that are isentropic (i.e., adiabatic and reversible). If the entropy is constant at each step of the process, it follows from equation (8.19) that p, p, and T are interrelated. The following equations describe these relations for isentropic flow:
= constant
p7
(8.20a)
410
Chap. 8 I Dynamics of a Compressible Flow Field
p
and
= constant
(8.20b)
= constant
8.1.4 Speed of Sound
(8.20c)
From experience, we know that the speed of sound in air is finite. To be specific, the speedofsound is defined as the rate at which infinitesimal disturbances are propagated from their source into an undisturbed medium.These disturbances can be thought of as small pressure pulses generated at a point and propagated in all directions. We shall learn later that finite disturbances such as shock waves propagate at a greater speed than that of sound waves. Consider a motionless point source of disturbance in quiescent, homogeneous air (Fig. 8.la). Small disturbances generated at the point move outward from the point in a spherically symmetric pattern. The distance between wave fronts is determined by the
(a)
(b)
"Mach cone"
Zone of silence Zone of action
Zone
(Note:
(c)
silence aAt
1.
(d)
Figure 8.1 Wave pattern generated by pulsating disturbance of
infinitesimal strength; (a) disturbance is stationary (U = 0); (b) dis
turbance moves to the left at subsonic speed (U < a); (c) disturbance moves to the left at sonic speed (U = a); (d) disturbance moves to the left at supersonic speed (U > a).
Sec. 8.1 I Thermodynamic Concepts
411
frequency of the disturbance. Since the disturbances are small, they leave the air behind
them in essentially the same state it was before they arrived.The radius of a given wave front is given by
r=
at
(8.21)
where a is the speed of propagation (speed of sound) of the wave front and t is the time since the particular disturbance was generated. Now, suppose that the point source begins moving (Fig. 8db) from right to left at a constant speed U which is less than the speed of sound a. The wavefront pattern will now appear as shown in Fig. 8db. A stationary observer ahead of the source will detect an increase in frequency of the sound while one behind it will note a decrease. Still, however, each wave front is separate from its neighbors. If the speed of the source reaches the speed of sound in the undisturbed medium, the situation will appear as shown in Fig. 8.lc. We note that the individual wave fronts are still separate, except at the point where they coalesce. A further increase in source speed, such that U > a, leads to the situation depicted
in Fig. 8.ld.The wave fronts now form a conical envelope, which is known as the Mach cone, within which the disturbances can be detected. Outside of this "zone of action" is the "zone
of silence," where the pulses have not arrived.
We see that there is a fundamental difference between subsonic (U < a) and
supersonic (U > a) flow. In subsonic flow, the effect of a disturbance propagates upstream of its location, and, thus, the upstream flow is "warned" of the approach of the disturbance. In supersonic flow, however, no such "warning" is possible. Stating it another way, dis
turbances cannot propagate upstream in a supersonic flow relative to a sourcefixed observer. This fundamental difference between the two types of flow has significant consequences on the flow field about a vehicle.
We note that the halfangle of the Mach cone dividing the zone of silence from the zone of action is given by
=
1
a
(8.22)
where
is the Mach number. At M =
00, p. 0. and as M (i.e., when U = a), To determine the speed of sound a, consider the wave front in Fig. 8.1 a propagating into still air.A small portion of curved wave front can be treated as planar.To an observer attached to the wave, the situation appears as shown in Fig. 8.2. A control volume is also shown attached to the wave. The boundaries of the volume are selected so that the flow is normal to faces parallel to the wave and tangent to the other faces. We make the key
1
assumption (borne out by experiment) that, since the strength of the disturbance is infinitesimal, a fluid particle passing through the wave undergoes a process that is both reversible and adiabatic (i.e., isentropic). The integral forms of the continuity and the momentum equations for a onedimensional, steady, inviscid flow give
412
Chap. 8 I Dynamics of a Compressible Flow Field
Ambient air
a
I
{
a+dU Air after wave has passed
P
p+dp p+dp
Figure 8.2 Control volume used to determine the speed of sound. (A velocity of equal magnitude and of opposite direction has been
superimposed on a wave of Fig. 8.la so that the sound wave is stationary in this figure.)
padA = (p + dp)(a + dU)dA
(8.23) (8.24)
pdA
—
(p + dp)dA = [(a + dU) — a]padA
Simplifying equation (8.24), dividing equations (8.23) and (8.24) by dA, and combining the two relations gives
dp =
However, since the process is isentropic,
a2
a2
dp
(8.25)
=
(8.26)
S
where we have indicated that the derivative is taken with entropy fixed, as original
ly assumed. For a perfect gas undergoing an isentropic process, equation (8.20a) gives
p=
c is a constant. Thus,
a2 =
=
P
(8.27)
Using the equation of state for a perfect gas, the speed of sound is
a=
8.2
(8.28)
ADIABATIC FLOW IN A VARIABLEAREA STREAMTUBE
For purposes of derivation, consider the onedimensional flow of a perfect gas through a
variablearea streamtube (see Fig. 8.3). Let us apply the integral form of the energy equation [i.e., equation (2.35)} for steady, onedimensional flow. Let us assume that there is no heat transfer through the surface of the control volume (i.e., Q = 0) and that only flow work (pressurevolume work) is done. Work is done on the system by the pressure forces acting at station 1 and is, therefore, negative. Work is done by the system at station 2. Thus,
Sec. 8.2 I Adiabatic Flow in a Variab'eArea Streamtube
413
p1
Properties are uniform over a station
P2
p2
Figure 8.3 Onedimensional flow in a
streamtube.
U2
+ p1U1A1 — p2U2A2 =
Ue1P1U1A1
+
p1U1A1 = p2U2A2 by continuity, and using the definition for
the enthalpy, we obtain H1 =
h1
+
U2
= h2 +
U2
(8.29)
where, as is usually the case in aerodynamics problems, changes in potential energy have
been neglected. The assumption of onedimensional flow is valid provided that the streamtube crosssectional area varies smoothly and gradually in comparison to the axial distance along the streamtube.
For a perfect gas, equation (8.29) becomes
+
or
U2
=
+
u2
(8.30)
u2
(831)
By definition, the stagnation temperature T1 is the temperature reached when the fluid is brought to rest adiabatically.
U2 T1=T+—
(8.32)
Since the locations of stations 1 and 2 are arbitrary,
T11 =
(8.33)
That is, the stagnation temperature is a constant for the adiabatic flow of a perfect gas
and will be designated simply as T1, Note that, whereas it is generally true that the stagnation enthalpy is constant for an adiabatic flow, as indicated by equation (8.29), the
_____—
414
Chap. 8 I Dynamics of a Compressible Flow Field
stagnation temperature is constant only if the additional condition that
is constant
(i.e., the gas behaves as a perfect gas) is satisfied. For any onedimensional adiabatic flow,
T—
2[yR/(y
1)]T 
_1+Y'M2 2
834
Note that we have used the perfectgas relations that
= yR/(y — 1) and that a2 = yRT. It is interesting to note that, when the flow is isentropic, Euler's equation for
onedimensional, steady flow (i.e., the inviscidflow momentum equation) gives the same result as the energy equation for an adiabatic flow. To see this, let us write equation (3.2) for steady, onedimensional, inviscid flow: du
=
dp
— —i
(8.35a)
ip
/
fdp
+
J
/
c
is
du =
0
(8.35b)
Note that for a onedimensional flow u = U. For an isentropic process (which is of
course adiabatic),
Differentiating, substituting the result into equation (8.35b), and integrating, we obtain y
y—l
Thus,
1
+
2
= constant
yp
y—lp
h+
which is equation (8.29).
+
U2
2
= constant
Using the perfectgas equation of state, we obtain
U2
= constant
EXAMPLE 8.1: An indraft, supersonic wind tunnel
We are to design a supersonic wind tunnel using a large vacuum pump to draw air from the ambient atmosphere into our tunnel, as shown in Fig. 8.4.
The air is continuously accelerated as it flows through a convergent/divergent nozzle so that flow in the test section is supersonic. If the ambient air is at the standard sealevel conditions, what is the maximum velocity that we can attain in the test section?
Solution: To calculate this maximum velocity, all we need is the energy equation for a steady, onedimensional, adiabatic flow. Using equation (8.29), we have
Sec. 8.2 I Adiabatic. Flow in a VariableArea Streamtube
Anibient air at standard
conditions serves
415
0
"Sonic throat"
as tunnel reservoir

Test section
To vacuum system
Figure 8.4 Indraft, supersonic wind tunnel.
Since the ambient air (i.e., that which serves as the tunnel's "stagnation chamber" or "reservoir") is at rest, U1 = 0. The maximum velocity in the test sec
tion occurs when h2 =
0
(i.e., when the flow expands until the static
temperature in the test section is zero). (1004.7)288.15J/kg
U2 = 760.9 rn/s
Of course,we realize that this limit is physically unattainable, since the gas would liquefy first. However, it does represent an upper limit for conceptual considerations.
EXAMPLE 8.2: A simple model for the Shuttle Orbiter flow field
During a nominal reentry trajectory, the Space Shuttle Orbiter flies at 3964 ft/s at an altitude of 100,000 ft. The corresponding conditions at the stagnatiÔn point (point 2 Fig. 8.5) are P2 = 490.2 lbf/ft2 and T2 = 1716.0°R. The static pressures for two nearby locations (points 3 and 4 of Fig. 8.5) are p3 = 259.0 lbf/ft2 and = 147.1 lbf/ft2. All three points are outside the boundary layer. What ate the local static temperature, the local velocity, and the local Mach number at points 3 and 4?
Solution: At these conditicins, the air can be assumed to behave approximately as a perfect gas. Furthermore, since all three points are outside the boundary layer and dowhstream of the bow shock wave, we will assume that the flow expands isentropically frOm point 2 to point 3 and then to point 4. (Note that because the shock wave is curved, the entropy will vary through the shock layer. For a further discussion of the rotational flow downstream of a curved shock wave, refer to Chapter 12. Thus, validity of the assumption that the expansion is isentropic should be verified for a given application.) For an isentropic process, we can use equation (8.20b) to relate the temperature ratio to the pressure ratio between two points. Thus,
______
416
Chap. 8 / Dynamics of a Compressible Flow Field
Bow shock wave
layer
Figure 8.5 Shuttle flow field for Example 8.2.
T3 =
Similarly,
(1716.0)(O.83337) = 1430.1°R
= (1716.0)(0.70899) = 1216.6°R
Using the energy equation for the adiabatic flow of a perfect gas [i.e.,
equation (8.31)], and noting that U2 = 0, since point 2 is a stagnation point,
U3 =
—
T3)]°5
= [2(0.2404
= 1855.2 ft/s
Btu ft lbm\ (778.2 ft lbf\28590R)j (32.174 ibm °R) Btu
/
Similarly,
U4
2451.9 ft/s
Using equation (1.14b) for the speed of sound in English units,
=
U3
a3
1855.2
= 1.001
U4 =
a4
2451.9
=
1.434
Thus, the flow accelerates from the stagnation conditions at point 2 to sonic conditions at point 3, becoming supersonic at point 4.
Sec. 83 / isentropic Flow in a VariableArea Streamtube
8.3
417
ISENTROPIC FLOW IN A VARIABLEAREA STREAMTUBE
It is particularly useful to study the isentropic flow of a perfect gas in a variablearea
streamtube, since it reveals many of the general characteristics of compressible flow. In addition, the assumption of constant entropy is not too restrictive, since the flow outside the boundary layer is essentially isentropic except while crossing linear shock waves or downstream of curved shock waves. Using equations (8.20) and (8.34), we can write
(i +
/
p
— 1
M2)
ivi2)
2
1
(8.36)
2
j
(8.37)
where Pti and Pti are the stagnation pressure and the stagnation density, respectively.
Applying these equations between two streamwise stations shows that if T1 is constant and the flow is isentropic, the stagnation pressure Pti is a constant. The equation of state requires that Pti be constant also. To get a feeling for the deviation between the pressure values calculated assuming incompressible flow and those calculated using the compressible flow relations, let us expand equation (8.36) in powers of M2:
p
1 + 1M2 + 0(M4) +
2
(8.38)
Since the flow is essentially incompressible when the Mach number is relatively low, we can neglect higher order terms. Retaining only terms of order M2 yields
p
which for a perfect gas becomes
I +
2
(8.39)
p
Rearranging,
=
+ y u2
2yRT
+
=
+
1
U2
2p/p
(8.40)
Pri =
(8.41)
Thus, for low Mach numbers, the general relation, equation (8.36), reverts to Bernoulli's
equation for incompressible flow. The static pressures predicted by equation (8.36) are compared with those of equation (8.41) in terms of percent error as a function of Mach number in Fig. 8.6. An error of less than 1% results when Bernoulli's equation isused if the local Mach number is less than or equal to 0.5 in air. In deriving equations (8.34), (8.36), and (8.37), the respective stagnation properties have been used as references to nondimensionalize the static properties. Since the continuity equation for the onedimensional steady flow requires that pUA be a constant, the area becomes infinitely large as the velocity goes to zero. Let us choose the area
418
Chap. 8 I Dynamics of a Compressible Flow Field
PC — Pa PC
M
Figure 8.6 Effect of compressibility on the theoretical value for the pressure ratio.
where the flow is sonic (i.e., M = 1): as the reference area to relate to the streamtube area at a given station. Designating the sonic conditions by a (*) superscript, the continüity equation yields A* pU (8.42) p*U* A ptl(p*/ptl)
since M* =
1.
Noting that p*/ptl and
+
are to be evaluated at M = M* =
7
1,
1M2)](Y)/(7)
(8.43)
Given the area, A, and the Mach number, M, at any station, one could compute an A* for that station from equation (8.43). A* is the area the streamtube would have to be if the flow were accelerated or decelerated to = I isentropically. Equation (8.43) is espedaily useful in streamtube flows that are isentropic, and therefore where A* is a constant.
Sec. 8.3 / Isentropic Flow in a VariableArea Streamtube
419
In order to aid in the solution of isentropic flow problems, the temperature ratio [equation (8.34)], the pressure ratio [equation (8.36)], the density ratio [equation (8.37)], and the area ratio [equation (8.43)] are presented as a function of the Mach number in Table 8.1. A more complete tabulation of these data is given in Ames Research Center Staff (1953). The results of Table 8.1 are summarized in Fig. 8.7. In order to determine the massflow rate in a streamtube,
th = pUA =
\PriJ
M
{1 +
v
A
(8.44)
[(y
Thus, the massflow rate is proportional to the stagnation pressure and inversely proportional to the square root of the stagnation temperature. To find the condition of max
imum flow per unit area, one could compute the derivative of (th/A) as given by
equation (8.44) with respect to Mach number and set the derivative equal to zero. At this condition, one would find that M = 1. Thus, setting M = 1 in equation (8.44) yields
— —
2
I'ti
'\A)max
—
R
+ 1)
(8.45)
The maximum flow rate per unit area occurs where the crosssection area is a minimum (designated A*), where the Mach number is one. Thus, the maximum flow rate per unit area occurs at the sonic throat when the flow is choked.
M
Figure 8.7 Property variations as functions of Mach number for isentropic flow for 'y = 1.4.
420
Chap. 8 / Dynamics of a Compressible Flow Field
TABLE 8.1
Correlations for a OneDimensional, Isentropic Flow of Perfect AIR (y
1.4)
M
o
A
P
p
T
Ap
A* Pti
00
A*
00
Pti
1.00000 0.99825 0.99303 0.98441 0.97250 0.95745 0.93947 0.91877 0.89562 0.87027 0.84302 0.81416 0.78400 0.75283 0.72092 0.68857 0.65602
0.62351
Pti
0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85
0.90
11.592 5.8218 3,9103 2.9635 2.4027 2.0351 1.7780 1.5901 1,4487 1.3398 1.2550
11882
1.1356 1.09437 1.06242 1.03823
1.02067
1.00000 0.99875 0.99502 0.98884 0.98027 0.96942 0.95638 0.94128 0.92428 0.90552 0.88517 0.86342 0.84045 0.81644 0.79158 0.76603 0.74000
0.71361
1.00000 0.99950 0.99800 0.99552 0.99206 0.98765 0.98232 0.97608 0.96899 0.96108 0.95238 0.94295 0.93284 0.92208 0.91075 0.89888 0.88652
0.87374
11.571
5.7812 3.8493 2.8820 2.3005 1.9119 1.6336
1.4241
1.2607 1.12951 1.02174 0.93155 0.85493 0.78896 0.73155 0.68110
0.63640
1.00886
1.00214
0.59126
0.55946
0.68704
0.66044
0.95
1.00
1.05 1.10 1.15
1.20 1.25
1.30 1.35
1.00000 1.00202 1.00793
1.01746 1.03044 1.04676
1.06631
0.52828 0.49787 0.46835
0.43983 0.41238
0.63394 0.60765 0.58169
0.55616 0.53114
0.86058 0.84710 0.83333 0.81933 0.80515
0.79083 0.77640
0.59650
0.56066
0.52828 0.49888 0.47206
0.44751 0.42493 0.40411
0.38606
0.36092 0.33697 0.31424 0.29272
0.50670
0.48291
0.76190
0.74738 0.73287 0,71839 0.70397
1.08904
1.1149 1.1440
1.40
1.45
0.45980 0.43742 0.41581
0.38484 0.36697 0.35036 0.33486
1.50
1.1762
1.2115 1.2502 1.2922
0.27240
0.25326
0.39498
0.37496
0.68965
0.67545
0.32039
0.30685
1.55
1.60
1.65 1.70
1.75
0.23527 0.21839
0.20259
0.35573 0.33731
0.31969
0.66138 0.64746
0.63372
0.29414 0.28221
0.27099
1.3376
1.80 1.85
1.90
1.3865 1.4390
1.4952
0.18782 0.17404
0.16120
0.30287 0.28682
0.27153
0.62016 0.60680
0.59365
0.26042 0.25044
0.24102
1.5555
1.6193
0.14924
0.13813
0.25699
0.24317
0.58072
0.56802 (continued
0.23211
0.22367
1.95
on next page)
Sec. 8.3 / Isentropic Flow in a VariableArea Streamtube
421
TABLE 8.1
continued
M
2.00 2.05
A
P
p
T
Ap
A* Pti
0.21567
A*
1.6875
Pti
0.12780
Pti
0.23005 0.55556
2.10 2.15
2.20
2.25 2.30 2.35 2.40 2.45 2.50
2.55
1.7600 1.8369 1.9185 2.0050 2.0964 2.1931 2.2953 2.4031 2.5168
2.6367 2.7630
0.11823 0.10935 0.10113 0.09352 0.08648 0.07997 0.07396 0.06840 0.06327
0.05853
0.21760 0.20580
0.19463
0.54333 0.53135
0.51962
0.20808 0.20087
0.19403
0.18405 0.17404 0.16458 0.15564 0.14720 0.13922
0.13169
0.50813 0.49689 0.48591 0.47517 0.46468 0.45444
0.44444
0.18751
0.05415
2.60 2.65 2.70 2.75 2.80
2.85
2.8960 3.0359 3.1830 3.3376
3.5001
0.05012 0.04639 0.04295 0.03977
0.03685 0.03415 0.03165 0.02935
0.12458 0.11787 0.11154 0.10557 0.09994
0.09462 0.08962 0.08489 0.08043
0.43469 0.42517 0.41589 0.40684 0.39801
0.38941
0.18130 0.17539 0.16975 0.16437 0.15923 0.15432 0.14963 0.14513 0.14083 0.13671 0.13276
0.12897 0.12534 0.12185 0.11850
2.90 2.95
3.00
3.6707 3.8498 4.0376
0.38102 0.37286 0.36490
3.50 4.00
4.50
4.2346 6.7896
10.719
0.02722 0.01311
0.00658
0.07623 0.04523
0.02766
0.35714 0.28986
0.23810
0.11527 0.08902
0.07059
5.00 6.00 7.00 8.00 9.00 10.00
00
16.562 25.000
0.00346 633(10)_6
53.189 104.143 190.109 327.189 535.938
00
242(10)6
474(10y7 236(10)v
0
0.01745 0.01134 0.00519 0.00261 0.00141 0.000815 0.000495
0
0.19802 0.16667 0.12195 0.09259 0.07246 0.05814 0.04762
0
0.05723 0.04725 0.03368 0,02516 0.01947 0.01550 0.01263
0
Figure 8.7 shows that, for each value of A*/A, there are two values of M: one subsonic, the other supersonic. Thus, from Fig. 8.7 we see that, while all static properties of the
fluid monotonically decrease with Mach number, the area ratio does not. We conclude that to accelerate a subsonic flow to supersonic speeds, the streamtube must first converge in an isentropic process until sonic conditions are reached. The flow then accelerates in a diverging streamtube to achieve supersonic Mach numbers. Note that just because a streamtube is convergent/divergent, it does not necessarily follow that the flow is sonic at
422
Chap. 8 / Dynamics of a Compressible Flow Field
the narrowest cross section (or throat).The actual flow depends on the pressure distribu
tion as well as the geometry of the streamtube. However, if the Mach number is to be unity anywhere in the streamtube, it must be so at the throat. For certain calculations (e.g., finding the true airspeed from Mach number and stagnation pressure) the ratio 0.5 is useful.
0.5pU2
Pri
iyp
Pti
—
U2
—
2
Pti yRT
—
yM2 p 2 Pti
(8 46)
Thus,
0.5pU2
Pti
=
7M2 /
2
(\i +
— 1
2
M2)
(847)
The ratio of the local speed of sound to the speed of sound at the stagnation conditions is = + lM2) = = (8.48)
2
Note: The nomenclature used herein anticipates the discussion of shock waves. Since the stagnation pressure varies across a shock wave, the subscript ti has been used to designate stagnation properties evaluated upstream of a shock wave (which correspond to the freestream values in a flow with a single shock wave). Since the stagnation temperature is constant across a shock wave (for perfectgas flows), it is designated by the simple subscript t.
8.4
CHARACTERISTIC EQUATIONS AND PRANDTLMEYER FLOWS
Consider a twodimensional flow around a slender airfoil shape. The deflection of the
streamlines as flow encounters the airfoil are sufficiently small that shock waves are not gen
erated. Thus, the flow may be considered as isentropic (excluding the boundary layer, of course). For the development of the equations for a more general flow, refer to Hayes and Probstein (1966). For the present type of flow, the equations of motion in a natural (or streamline) coordinate system, as shown in Fig. 8.8, are as follows: the continuity equation,
as the smomentum equation,
+ pU— =
0
(8.49)
(8.59)
and the nmomentum equation,
+
as
an
=0
(8.51)
Since the flow is isentropic, the energy equation provides no unique information and is, therefore, not used. However, since the flow is isentropic, the change in pressure with re
8 Supersonic flow around an airfoil in natural (streamline) coordinates.g.53) and introducing the concept of the directional derivative [e. spect to the change in density is equal to the square of the speed of sound. A characteristic is a line which exists only in supersonic flow& characteristics should not be confused with finitestrength waves. we obtain + along the line having the direction dn ds pu2 30 = o (8.. Equivalently. which is the Mach angle. which are so called beause to an observer looking downstream. JL = The characteristics correspond to the leftrunning Mach waves. 8. —=a 2 '3p (8.Sec.53) Combining equations (8. such as shock waves.8).51) and (8. Thus.4 I Characteristic Equations and PrandtlMeyer Flows z 423 echaracteristic x Figure 8. the Mach wave appears to be going downstream in a leftward direction.52) and the continuity equation becOmes äpM2—1300 as PU (8.54a) = tan p = 1 \/M2—1 (i.Wayland (1957)].. The characteristic is inclined to the local streamline by the angle p. the leftrunning characteristic of Fig. . 8.e.
571 1.920 52.5 13.041 2.614 55.5 49.0 18.465 35.631 23.0 24.5 43.418 1.741 2.349 2.415 27.099 67.151 25.) M 1.566 44.764 2.758 1.950 58.5 36.216 27.655 1.932 36.004 2.0 30.0 23.293 35.5 1. (deg) 30.172 2.469 1.383 1.5 22.836 2.5 21.5 31.983 21.5 17.986 p.082 (deg) 90.289 2.847 30.0 11.701 41.961 2.936 2.0 29.066 23.786 21.294 1.177 1.5 33.626 2.847 23.585 22.0 28.0 7.196 24.000 1.554 1.950 1.588 1.180 56.366 1.5 34.329 2.410 2.452 1.0 4.198 1.538 2.874 35.269 2.5 32.0 3.897 1.574 64.348 1.390 2.182 21.5 45.857 44.059 2.287 41.5 35.0 40.0 0.678 27.0 22.491 28.134 2.0 20.400 1.916 19.153 2.5 8.0 27.561 2.812 2.035 38.5 38.486 31.769 28.642 50.290 33.536 30.5 23.0 37.0 33.134 40.451 61.834 32.0 16.053 39.418 23.965 24.256 1.548 33.424 Chap.5 2.672 1.547 38.5 19.706 1.188 32.473 2.5 46.668 25.0 26.520 1.997 22.0 35.218 1.582 2.0 9.5 16.915 33.537 1.065 34.790 22.910 2.431 2.206 22.0 36.5 12.788 2.0 5.0 25.0 41.503 1.382 22.830 20.5 37.330 1.5 29.5 40.639 1.452 2.230 2.915 1.689 1.340 29.827 1.052 28.523 42.7 18 2.306 45.0 45.0 39.435 1.721 v ('deg.0 8.736 24.051 1.0 38.017 20.5 15.0 13.0 32.5 5.0 48.5 44.096 19.5 7.862 1.5 9.369 2.0 49.5 1.560 2.5 42.0 21.622 1.649 2.000 72.658 48.5 18.694 2.287 24.0 6.229 29.312 1.023 2.160 36.894 42.488 32.0 42.096 2.237 1.0 34.0 46.510 24.5 27.0 2.861 2.753 47.619 49.177 43.459 20.207 21.5 28.309 2.205 53.0 19.792 1.724 1.148 31.275 1.5 26.210 2.191 2.997 59.987 (continued on next page) .115 2.155 26.0 44.495 2.968 1.671 2.0 31.644 20.430 25.5 4.277 20.5 48.844 1.398 21.397 26.5 24.164 25.5 M 1.537 39.5 14. 8 / Dynamics of a Compressible Flow Field TABLE 8.0 14.133 1.908 25.605 1.632 29.585 40.810 1.646 26.0 10.5 41.738 51.775 1.0 15.249 2.5 39.673 34.741 20.0 17.928 29.078 2.945 27.611 37.108 1.814 31.886 2.5 6.5 1.486 1.5 47.899 26.591 21.082 46.0 43.073 37.0 1.5 10.896 47.516 2.5 11.738 15.155 1.5 30.0 12.2 Mach Number and Mach Angle as a F unction of PrandtMey er Angle v (deg) 0.879 1.5 3.0 47.
406 3.937 9.727 5.899 10.406 12.623 94:0 7:538 945 96:0 96.0 51.5 64.983 645 650 65.056 9.5 81.426 4.5 57.0 71.5 78.868 15270 15.5 55.433 8.4 I Characteristic Equations and PrandtlMeyer F'ows TABLE 8.201 7.448 6.059 12.913 5.302 7.0 3904 3941 3.955 5.819 6.5 52.852 11.316 3.472 6.729 6.5 53.5 77.0 90.729 76.207 8.414 10.927 12.146 3.0 83.0 77.979 4.006 6.871  18.728 3 762 p.608 4.551 17.881 13.213 19.546 8.911 7.0 87.0 59.768 11.5 60.759 8.0 76. 8. (deg) 11.515 4.126 14.340 6.287 3.5 61.655 4.310 6.173 4.325 13.5 69.0 89.149 12.200 18.5 68.234 9.214 4.777 10.258 3.289 5.530 3.556 6.016 14840 14698 14.0 70.5 100.371 9.232 6.5 91.092 3.5 74.0 80.656 10.470 4.231 p.642 6.0 99.5 51.819 9.375 3.795 5.063 5.467 3.391 17.095 .920 16.660 8.392 11.905 9.0 97 5 7735 8:092 8.5 62.268 11.0 88.233 17.5 75.0 55.347 8.5 79.080 6.535 12.210 9.102 7.627 3.145 11.5 4055 4.Sec.0 68.870 18.006 15.175 5.885 6664 6.417 665 67.348 5.009 5.0 75.498 3.002 13.0 74.0 62 5 M 3.5 89.0 91.801 14278 14.0 58.408 5.0 56.390 6.022 10.562 3.366 18.832 3.0 82.294 10.0 61.5 84.864 5.611 16.5 66.5 70.459 13.661 5.202 3.175 10.5 101.806 4.0 8618 8.013 3.480 7428 7:099 6.5 83.665 12.903 4.0 79.218 8.041 18.039 3.019 5.036 17.191 13.0 84.119 3.594 3.0 81.0 86.708 9.0 50.003 8.005 7.297 4.021 6.0 92.0 3797 3.5 92.752 4.894 .458 16.701 v (deg) M 4.5 80.233 6.0 85.935 6.5 59.5 90.2 (continued) 425 v (deg) 50.703 4.983 7.989 6.532 18.5 87.556 6.596 5.306 16.174 3.561 4.5 82.118 5.5 5.055 9.0 72.5 102.517 11.711 17.0 54.5 56.126 6.386 19.0 100.5 72.5 86.155 6.865 13.660 3.155 16.856 15.320 8.065 3.535 10.701 9.339 4.561 15 415 630 63.133 4.119 9.795 12. (deg) 19.346 3.0 101.0 52.873 17.0 53.0 57.532 5.436 3.708 15.774 8.230 3.5 88.382 4.888 8.536 9.0 73.467 9.5 73.642 11.557 14.5 54.5 97.5 85.5 58.350 9.772 980 98:5 99.765 16.0 78.470 5.5 71.076 16.584 9.277 12.0 60.694 3.
v+0=Q (8. equation (8. let us work through a sample problem .56) But from the adiabaticflow relations for a perfect gas.54b) provide a relation between the local static pressure and the local flow inclination.e.59a) a constant. 8 / Dynamics of a Compressible Flow Field 2 an along the line whose direction is dn = (i. v— 0= R (8.55) yields = dO \/M2 — 1 dM2 2M2{1 + [(y — 1)/2]M2} (8.59b) which is another constant.e.1).57) yields the relation. The use of equations (8. Similarly. Equations (8.58) Tabulations of the PrandtlMeyer angle (v). the rightrunning characteristic of Fig.57) Integration of equation (8.59a) and (8.2. 8. Differentiating equation (8. all the waves are either leftrunning waves or rightrunning waves). Thus. along a leftrunning characteristic. Euler's equation for a steady. the corresponding Mach number. along a rightrunning characteristic.54a) becomes dU U dO VM2_l lM2) (8. To make the application clear. and the corresponding Mach angle (ps) are presented in Table 8.56) and substituting the result into equation (8.8). (U)2 = M2(1 + where is the speed of sound at the stagnation conditions. for the leftrunning characteristic. which can be derived by neglecting the viscous terms and the body forces in the momentum equation (3. which is valid for a leftrunning characteristic: 0= v + constant of integration where i' is a function of the Mach number as given by p= [y+1 Vy—1 arctan Vy+1 (M2 I Iy—i — 1) — arctanVM2 — 1 (8.426 Chap.59b) is simplified if the slope of the vehicle surface is such that one only needs to consider the waves of a single family (i.54a) and (8.. states that dp = —pUdU Thus. inviscid flow..
it will be assumed that both the acceleration of the flow over the upper surface and the deceleration of the flow 2 Free stream M.lOc. 8. For each segment will be 0. Thus. moving through the air at = 2.2c. as shown in Fig. 8.11653 a Uc.Sec.048° —10.. = 2. The leading edge of the airfoil is parallel to the free stream.370° For the freestream flow.145° —3.059.3: Use PrandtlMeyer relations to calculate the aerodynamic coefficients for a thin airfoil Consider the infinitesimally thin airfoil which has the shape of a parabola: x2 C —————(z — Zmax) where = 0. The thin airfoil will be represented by five linear segments.059 4 Zmax + X2 = (ZZmax) z (a) Leftrunning Mach waves Regionco Rightrunning Mach waves (b) Figure 8.9. the slopes of these segments are as follows: Segment 8 a b C d e —1. (b) wave pattern.9 Mach waves for supersonic flow past a thin airfoil: (a) airfoil section. .4 / Characteristic Equations and PrandtiMeyer Flows 427 EXAMPLE 8. = Pu 0.607° —8.3 U Since the turning angles are small.
59b).1018 2.145° and = 2.000° — (—1. Thus.2784 2. Note that the expansion waves on the upper surface diverge as the flow accelerates.0942 0.1091 0.7733 1.0173 1.630° d e 2.952° 17.1.145°) = 26. or = so —dO = p00 (Ow.6070 33. the stagnation pressure is constant throughout the flow field and equal to Pu (which is the value for the freestream flow). we can use equations (8.1244 0.9) to the first segment on the upper surface (region ua).3713 2.1018 Using Table 8. we move along a rightrunning characteristic to cross the leftrunning Mach wave shown in the figure. in going from the free stream to the first segment on the lower surface (region la) we move along a leftrunning characteristic to cross the rightrunning Mach wave shown in the figure.0480 38.o) = 28.370° vj M1 P1 Pti 0.855° 24. 8 / Dynamics of a Compressible Flow Field over the lower surface are isentropic processes.855° Summarizing.1091.1952 2.428 Chap. Since the flow is isentropic. Furthermore.0615 26. 8. — = 28.7400 36.1813 0.2045 b c 2.1604 0.260° 19.59a) and (8.0827 0.000° + (—1.1428 0.9286 1.6940 . = 0. In going from the freestream (region cc in Fig.3930 22.8534 1.1450 31. U pper surface Lo wer surface Pu Segment a Pti 29.4679 0. Thus.145°) = 29. Similarly.0715 0. but the compression waves of the lower surface coalesce. iiO=R or clv = so 1'la = dO + Oc.
0896c Sum Finally. But that is not critical.0832c 0.2460c 0. tan Segment Poo Pco Poo Pco a 1. since we seek the force coefficients.61) = (8.1340c 0.01340c d e 1359 1.710 O.60) Referring again to Fig.759 0.0262c 0. Note that we have not been given the freestream pressure (or. the incremental drag force for any segment is dd1 = (Pu — sin 0.46). = (Pu — dx.380 0. = Thus.9. for a perfect gas. the altitude at which the airfoil is flying) or the chord length of the airfoil.004992c 0. 8..e.2286 0. equivalently.810 0.226 1.24)c dd1 = 0. = — dx1 Similarly. cos 0.0896c = 0.0264c O.939 O. the ith segment) is d11 = (Pu — ds.4 / Characteristic Equations and PrandtlMeyer Flows 429 Let us now calculate the lift coefficient and the drag coefficient for the airfoil.6784c 0.1890c 0. cl= 1 Referring to equation (8.7(4.6784c 0. it is evident that. the incremental lift force acting on any segment (i.Sec.0302 = 0.7(4. C1 = Cd = d4 0. C1 = (8.000524c b c 1.57 .529 0.614 0. 8.0443c 0.24)c dCd = 7.070 0.
Note that whereas the subscript 1 designates the freestream (cc) properties for flows such as those in Fig. For flow conditions where the gas is a continuum. Consider the curved shock wave illustrated in Fig. it designates the local flow properties just upstream of the shock wave when it occurs in the midchord region of a transonic airfoil (see Chapter 9).10. the characteristic curves. the shock wave is oblique and the downstream flow is often supersonic. are patching lines for continuous flows. 8. and the flow downstream of the shock wave is subsonic. At the plane of symmetry. X 10—6 cm) across which there is an almost instantaneous change in the values of the flow parameters.5 SHOCK WAVES The formation of a shock wave occurs when a supersonic flow decelerates in response to a sharp increase in pressure or when a supersonic flow encounters a sudden. . which is stationary in the bodyfixed coordinate system. whereas shock waves are patching lines for discontinuous flows.We will analyze V1 Normal portion of the shock wave U1(= Figure 8. 8. the shock wave is a narrow region (on the order of several molecular mean free paths thick. and temperature. viscous and heatconduction effects are important within the shock wave. Because of the large streamwise variations in velocity.10 Curved shock wave illustrating nomenclature for normal shock wave and for oblique shock wave. pressure. compressive change in direction. or Mach lines. A Mach wave represents a surface across which some derivatives of the flow variables (such as the thermodynamic properties of the fluid and the flow velocity) may be discontinuous while the variables themselves are continuous. is supersonic. Away from the plane of symmetry. Thus. The flow upstream of the shock wave.10. the shock wave is normal (or perpendicular) to the freestream flow. 8 / Dynamics of a Compressible Flow Field 8.430 Chap. The difference between a shock wave and a Mach wave should be kept in mind. A shock wave represents a surface across which the thermodynamic properties and the how velocity are essentially discontinuous.The downstream values are designated by the subscript 2. The velocity and the thermodynamic properties upstream of the shock wave are designated by the subscript 1.
62) 2.Sec. For a steady flow. equation (8.64) 4. Equation (8. these relations can be used to describe the flow across a normal shock wave. Energy: h1 + + = + + (8. and the energy equations for the flow through the ontrol volume. 8. oblique shock waves by writing the continuity.5 / Shock Waves Shock 431 wave Surface of control volume Surface of control volume Figure 8. shown in Fig. Thus. approaches 2. the flow is essentially stagnated (u2 0) behind a normal shock wave.11 Control volume for analysis of flow through an oblique shock wave. or that portiorl of a curved shock wave which is perpendicular to the free stream. Normal component of momentum: Pi + 3. Continuity: — (8. the momentum.0.66) As a cPmax The value of P2 1 — Pi 1 (8.11. Tangential component of momentum: = P2 + P2U2 (8. 8.67) Note that.11.65) In addition to describing the flow across an oblique shock wave such as shown in Fig. at the stagnation point of a vehicle in a supersonic stream is a function .63) becomes P2 — P1 (8.63) can be used to calculate the maximum value of the pressure coefficient in the hypersonic limit as M1 In this case. at the stagnation point of a vehicle in a hypersonic stream. by letting v1 = 0. 8.63) P21L2V2 (8. the integral equations of motion yield the following relations for the flow across an oblique segment of the shock wave: 1.
The tangential component of the velocity is unchanged. the energy equation becomes h1 + = h2 + (8.71) it can be seen that the deflection angle is zero for two "shock"wave angles. and (8. Thus. the property changes across an oblique shock wave are the same as those across a normal shock wave when they are written in terms of the upstream Mach number component perpendicular to the shock. a zerostrength shock wave). For hypervelocity flows where the shock waves are strong enough to cause dissociation or ionization. Note that for a given deflection angle 6. However.e.0. Thus. one can solve these equations numerically using the equation of state in tabular or in graphical form [e.63). Note that equations (8. (8. a Galilean transformation). the entropy must increase as the flow passes through the shock wave. the tangential component of the velocity is constant across the shock wave and we need not consider equation (8.71) are presented in graphical form in Fig.69) There are four unknowns (p2.0 for a lowspeed stream independent of the velocity.71) From equation (8. Moeckel and Weston (1958)]. = V2 (8. and (8.. h2) in the three equations (8. The larger of the two values of 0 is called the strong shock wave.63).68) That is. for a perfectgas flow.69). (1) The flow is not deflected when 0 = since the Mach wave results from an infinitesimal disturbance (i. Solutions to equation (8. Thus.62) with (8.70) Hence. u2.62). since the temperature (and therefore the speed of sound) changes across the shock wave.. 8. we need to introduce an equation of state as the fourth equation.e. Since the flow through the shock wave is adiabatic.e. Comparing equation (8. that the oblique flow is reduced to the normal flow by a uniform translation of the axes (i..e.432 Chap. (2) The flow is not deflected when it passes through a normal shock wave (i. This is the sweep back principle. one finds that for the oblique shock wave. the pressure must increase) as it passes through the shock wave. One obtains the relation between the shockwave (0) and the deflection angle (8). when 0 = 90°). 8 / Dynamics of a Compressible Flow Field of the freestream Mach number and is greater than 1.64) further.g... (8.12a. the flow must decelerate (i. there are two possible values for the shockwave angle 8. [ 1 — 1 I 1) J (8. p=pRT and A comparison of the properties downstream of a normal shock wave as computed using the charts for air in thermodynamic equilibrium with those computed using the perfectgas model for air is presented in Chapter 12. Recall that it is 1.64). Note that the tangential component of the Mach number does change.62).69) involve only the component of velocity normal to the shock wave: = U1 sin 0 (8. provided that the flow is incompressible. while the . P2.
a strong curved shock wave would occur with a complex subsonic/supersonic flow downstream of the shock wave.4: smaller value is called the weak shock wave.72) . Once the shockwave angle 0 has been found for the given values of M1 and 6. in engine inlets. a flatplate airfoil can be inclined 34° to a Mach 3. the weak shock wave typically occurs in external aerodynamic flows. However. In practice. a strong. 8.Sec. the other downstream properties can be found using the following relations: P2 = sin2 0 — ('y 1) Pi y+1 (8. 1. detached shock wave will occur. This is the maximum deflection angle for a weak shock wave to occur.0 stream and still generate a weak shock wave. If the deflection angle exceeds the maximum value for which it is possible that a weak shock can be generated. The high downstream pressure associated with the strong shock wave may occur in flows in wind tunnels. If the airfoil were to be inclined at 35° to the Mach 3. For instance. degrees Figure 8.12 Variation of shockwave parameters with wedge flow deflection angle for various upstream Mach numbers. y (a) shockwave angle.0 stream. or in other ducts. the strong shock wave occurs if the downstream pressure is sufficiently high.5 I Shock Waves 433 a) a) I. a) a) 0 CO 30 Deflection angle. 'a:.
5.8 — 1.4 1.5 4) 0.. 6.2 20 30 50 Deflection angle.8 Chap.6 cc ———— 1.6 '\1.6 —.— Sonic limit (M2 = 1) — — — —. degrees Figure 8.8 a U 4) 0.73) T2 (y — l)][('y — + 2] (y + (y — — (8. P2 P1 ('y + — +2 — (8.434 1.40 —'S 'S 1.—. t Weak shock wave shock wave — .. 8 / Dynamics of a Compressible Flow Field 1.y + 2J (y — 1)] (8.6 M=2.75) — [ (y + — 1Y/(7_1)[ y + :1 — [(. 0.12 (continued) (b) pressure coefficient. Strong 1. U i 1.76) . 1.2 2.2 4) 1.74) +2 ('y — 1)]sin2(O 6) = (8.0 .
The downstream Mach number is presented in Fig. 8. we cannot use Table 8. 8. 8.67) for a normal shock since y —*1 as M1 dissociation of molecules in the air at high Mach numbers. .Sec. and the temperature ratios for an oblique shock wave can be read from Table 8. // . 8.The values for many of these ratios are presented for a normal shock wave in Table 8.3 provided that M1 sin 0 is used instead of M1 in the first column. 8.The values for the pressure ratios. / / Deflection angle. . degrees Figure 8.5 / Shock Waves 435 E I / .12c as a func .12 (continued) (c) downstream Mach number.77) The pressure coefficient is presented in Fig.77) 00 due to the is consistent with equation (8. Equation (8. Note that since it is the tangential component of the velocity which is unchanged and not the tangential component of the Mach number.12b as a function of 6 and M1 .3 to calculate the downstream Mach number. and cp P2P1 sin 0 — 1) + (8.3 and in Fig.13. the density ratios.
3416 1.96972 0.52861 0.5157 2.71 39 1.46012 0.5694 1.47519 0.61650 0.76735 0.59562 0.02965 0.1196 1.00 1.1152 2.000 74.5316 1.50 4.45 1.73971 0.70 2.00 7.54015 0.9596 2.7200 8.97935 0.7704 1.1691 1.5000 4.42355 0.8362 7.99669 0.3966 2.91319 0.9800 9.10 1.5919 2.6450 9.4512 2.48817 0.56907 0.2696 4.00 3.5067 2.2050 3.6635 3.5157 1.00 9.08398 1.74418 0.45115 0.9468 1.8823 2.1490 3.0186 3.2383 2.87598 0.85 2.2683 5.01535 0.2119 3.42359 0.6027 1.388 00 1.5533 6.6363 2.1977 2.13876 0.2262 5.2547 1.20 1.65 2.3538 1.7400 2.40 2.35 2.40416 0.71956 0.1375 2.65396 0.05 1.8569 1.125 18.8761 5.4444 5.2450 1.75 1.8132 1.5133 1.3880 1.00 4.49181 0.6666 2.38980 0.4062 3.55 1.0000 5.03284 1.80 2.51792 0.92978 0.1594 1.41523 0.481 38 0.000 2.7413 6.35 1.43496 0.76175 0.47927 0.30 1.65 1.2762 6.90 1.986 10.333 14.78596 0.00 2.91177 0.458 29.833 57.21295 0.79021 0.06494 1.500 94.00000 0.941 10.2781 2.3592 2.40 1.8262 4.95 2.6133 3.333 116.8000 7.3 Correlation of Flow Properties Across a Normal Shock Wave as a Function of the Upstream Mach Number for air.4286 1.50 1.81264 0.0262 8.75 2.0885 2.84217 0.56128 0.70 1.55 2.7286 1.49563 0.20 2.30 2.15 1.3202 1.9784 5.64055 0.94483 0.3918 3.05 2.6512 5.6562 8.25 2.50387 0.99987 0.35773 0.39736 0.6790 3.09170 0.00 10.00 Do 1.1250 7.60554 0.47821 0. 'y = 1.60570 0.67422 0.00000 1.4583 2.4 M1 PA2 P2 P2 T2 pQ 1.1875 2.57735 0.8119 5.0846 3.3096 9.00 6.38757 0.32834 0.4228 1.6206 2.0403 2.48470 0.2608 4.99280 0.4489 3.2550 1.9014 1.90 2.4800 5.50 00 1.81268 0.436 Chap.65105 0.68410 0.1200 2.00000 1.06172 0.69752 0.5714 4.72088 0.38946 0.53441 0.7363 4.8106 3.2733 3.15 2.1909 1.469 13.00 8.3429 2.6473 1.45 2.95 3.58618 0.56148 0.10 2.4946 1.51299 0.4583 1.54055 0.95819 0.51932 0.0096 3.2901 2.2226 1.95312 0.58331 0.89520 0.62809 0.6079 1.8621 1.37330 0.0317 2.50 2.60 1.85573 0.09657 1.3333 3.7761 1.500 23.6562 1.9473 2.0450 4.441 55 0.39289 0.8119 2.4381 2.8571 4.3762 1.00000 0.87502 0.5652 5.55395 0.62812 0.7396 6.80 1.00496 0.25 1.00304 0 .5047 3.66844 0.70109 0.6119 3.8201 3.3150 4.693 20.2872 1.50831 0.52312 0.2862 2.7629 3.54706 0.4196 7.5632 2.60 2.0469 4.34275 0. 8 / Dynamics of a Compressible Flow Field TABLE 8.83456 0.85 1.00849 0.49965 0.000 41.50 5.5590 3.00000 1.8050 1.98706 0.40622 0.9931 3.37796 1.49902 0.9512 3.6875 1.3383 8.99892 0.1280 1.387 16.6896 1.0050 6.
In this case. Thus.15 as a function of the cone semivertex angle and the freestream Mach number M1.3 (which is the normal component of the Mach number) to the normal component of velocity using T2 to calculate the local speed of sound.0 Upstream Mach number (Mi) Figure 8.13 Property variations across a normal shock wave.0 7.0 5. An alternative procedure to calculate the Mach number behind the shock wave would be to convert the value of M2 in Table 8. oblique shock wave generated when supersonic flow encounters a wedge.0 6.0 4. For supersonic flow past a cone at zero angle of attack.0 2. as shown in the sketch of Fig. The shockwave angle. Then we can calculate the total velocity downstream of the shock wave: U2 = + from which we can calculate the downstream Mach number. the static pressure varies with distance back from the shock along a line parallel to the cone axis. this is not the case for the conical shock wave. . 8. the pressure coefficient cp (where — Pc — Pi is the static pressure along the surface of the cone). and the Mach number of the inviscid flow at the surface of the cone are presented in Fig.Sec.0 3. the shockwave angle Whereas all depends on the upstream Mach number M1 and the cone halfangle properties are constant downstream of the weak. 8.5 / Shock Waves 437 1.14. properties are constant along rays (identified by angle emanating from the vertex of the cone. tion of the deflection angle and of the upstream Mach number. 8.
4: . y (a) shockwave angle.438 Chap.00 Pi (freestream flow) Axial location Figure 8.14 Supersonic flow past a sharp cone at zero angle of attack. 1. 8 / Dynamics of a Compressible Flow Field = = 2. degrees Figure 8.0 39°) = 1. 0 0 C rID 00 10 20 30 40 50 Cone semivertex angle.15 Variations of shockwave parameters with cone semivertex angle for various upstream Mach numbers.
72) as 2(1. Included for comparison is the pressure downstream of the weak. 8.4 2. 30 20 Cone semivertex angle. Note .5150)2 P1 — 0.. as shown in Fig. 8.15b. 8. the shockwave angle is 31°. (I) 3. as can be calculated using Fig. reaching a value of 1 . Using Fig. oblique shock for a wedge with the same turning angle. 8. moving parallel to the cone axis)..e.07 The pressure increases across the shock layer (i.Sec.4 _ 1.29Pi (or at the surface of the cone.14.4: Supersonic flow past a sharp cone at zero angleofattack consider the cone whose semivertex angle is 10° exposed to a Mach 2 stream. 8.5 I Shock Waves 439 U 0 U 3.15a. 40 degrees Figure 8.15 (continued) (b) pressure coefficient. The pressure just downstream of the shock wave is given by equation (8.4)(4)(0.
C) Cl 30 40 degrees Cone semivertex angle. as will be discussed in Chapter 9. 8 / Dynamics of a Compressible Flow Field a) a) I.6 VISCOUS BOUNDARY LAYER In our analysis of boundary layers in Chapter 4. are no longer constant.440 Chap. we considered mostly flows for which the density is constant. which allow the flow to spread around the cone and which do not occur in the case of the wedge. However. that the shockwave angle and the pressure in the shock layer Pw are greater for the wedge. as the freestream Mach number approaches the transonic regime. a) C. Furthermore.15 (continued) (c) Mach number on the surface of the cone. The difference is due to axisymmetric effects. our solution technique must include the energy equation as well as the continuity equation and the momentum equation. shock waves occur at various positions on the configuration. the work of compression and of viscous energy dissipation produces considerable increases in the static temperature in the boundary layer. . when the freestream Mach number exceeds two. such as the density and the viscosity. Figure 8. 8. The correlations for the skin friction coefficient which were developed for these lowspeed flows were a function of the Reynolds number only. Because the temperaturedependent properties. The presence of the boundary layer and the resultant shockwave/boundarylayer interaction can radically alter the flow field.
In this section. 8. the vortical character of the turbulent boundary layer is evident. 8. the laminar boundary layer is thin and "smooth.That a slightly larger angle is observed experimentally is due in part to the displacement effect of the boundary layer. 8. Boundarylayer transition occurs approximately onequarter of the way along the portion of the conical generator which appears in the photograph. Upstream (nearer the apex). we can photographically record = 12°) in a supersonic. Shockwave/boundarylayer interactions .L = 4.5.28 X 106. 11.6 / Viscous Boundary Layer 441 = Figure 8. is 14. as calculated using Fig. we will discuss briefly 1.) ReOO. (Courtesy of Vought Corp.6°. The effects of compressibility 2. the fluid dynamicist must address the problem of heat transfer.16 Hypersonic flow past a slender cone: = 12°. The theoretical value.Sec." Downstream (toward the base of the cone). In addition to the calculations of the aerodynamic forces and moments.16.5 stream is presented in Fig.A shadowgraph of the flow field for a cone Mach 11. The reader might also note that the shockwave angle is approximately 14.15a.3°. the boundary layer. Because the density gradients in the compressible boundary layer affect a parallel beam of light passing through a windtunnel test section.
the heat transfer due to fluid motion). since is 95.84 and at an altitude of 20 km.069 m/s. because heat is transferred from the boundary layer of these flows.1 Effects of Compressibility As noted previously. is For example.e.65[1 + 0.Thus. the designer of vehicles that fly at supersonic speeds must consider problems related to convective heat transfer (i.2(4. A numerical solution for the static temperature distribution across a laminar boundary layer on a flatplate wing exposed to this flow is presented in Fig. This maximum teniperature. it is greater than either the temperature at the wall or at the edge of the boundary layer. .65 K. is 216. Although the maximum value of the static temperature is well below the stagnation temperature. the static temperature does not reach this value. 8 / Dynamics of a Compressible Flow Field 8. is 1428 m/s (or 5141 km/h): 216.095 Tte..34)].6. laminar boundary layer.2. for this flow. considerable varintions in the static temperature occur in the super sonic flow field around a body.7 K = Tte which is the stagnation temperature of the air outside of the boundary layer (where the effects of heat transfer are negligible).84)21 = 1231. 8.17.17 Static temperature distribution across a compressible.Chap. We can calculate the maximum temperature that occurs in the flow of a perfect gas by using the energy equation for an adiabatic flow equation (8. Also. Referring to Table 1. However. which is the stagnation temperature. Thus. 0. we see that the temperature of the air at the stagnation point is sufficiently high that we could not use an aluminum structure. Me = 4. let us calculate the stagnation temperature for a flow past a vehicle flying at a Mach number of 4. We have calculated the stagnation temperature.84 = 0. which exists only where the flow is at rest relative to the vehicle and where there is no heat transferred from the fluid.3 T Tte Figure 8.
4 and 8. the skin friction coefficient depends on the local Reynolds number. which is known as Reynolds analogy and was discussed in Chapter 4. Cf = (8. which is indicated by the char patterns on the recovered spacecraft. where ficient. is clearly illustrated in Fig. For further information about con vective heat transfer. (Courtesy of NASA. given . 8. the reader is referred to Chapter 4 and to texts such as such as Kays (1966) and Chapman (1974).18 Comparison between the oil flow pattern (indicating skin friction) obtained in the wind tunnel and the char patterns on a recovered Apollo spacecraft. and FRe are presented as functions of Me and of in Tables 8.) The correlation between convective heat transfer and the shear forces acting at the wall.6 / Viscous Boundary Layer 443 Figure 8. is the local Reynolds number.Sec. Thus.18. the Mach number of the inviscid flow. Thus. For highspeed flow past a flat plate. and and are correlation parameters which depend only on the Mach number and on the temperature ratio. 8.78) Spalding and Chi (1964) developed a calculation procedure based on the assumption that there is a unique relation between is the skin friction coefand RevFRe. respectively. The streaks in the oilflow pattern obtained in a wind tunnel show that the regions of high shear correspond to the regions of high heating.5. and the temperature ratio.
6180 2.7745 4.50 0.00 0.9839 1.20 0.4036 0.9812 5.7413 300.0576 0.7210 9.2649 5.1278 2.40 0.0645 0.0 2.0137 1.5802 1.00 4.1544 1.9130 2.8956 2.4742 11.3463 0.2618 8.9263 82.05 0.8168 0.1947 0.00 5.9267 1.9584 1.0 1.0970 17.8127 4.6829 0.0 0.6388 0.5409 34.1865 0.4494 1.8660 2.00 4.00 8.0320 0.7145 0.00 10.9247 3.6618 33.1588 7.2576 5.5183 2.7246 3.3239 0.0813 3.5236 0.3336 3.0460 0.0078 0.7999 0.9242 31.4491 1.4331 0.50 0.2309 292.6662 0. 8 / Dynamics of a Compressible Flow Field TABLE 8.0 1e 1.3840 1.0829 0.2581 4.0029 25.9072 7.1061 0.0400 0.3743 0.4777 22.6903 0.0677 0.3304 6.1512 0.0240 0.00 2.60 0.3451 1.7821 0.0294 0.00 10.8523 0.0466 0.9208 0.00 8.5708 278.6642 4.4493 1.0434 1.1713 1.0 5.5022 18.6532 12.1204 0.5326 0.1167 1.5 Values o f FRe as a Function of Me and T?v/Te Me 0.8139 304.4793 16.5477 0.4385 0.6889 3.4884 0.4571 1.6477 3.0114 0.10 0.0349 0.00 3.5733 2.9621 3.10 0.7158 103.1946 0.0 0.1305 4.2384 3.8486 10.3678 2.4867 1.2557 97.0044 3.0586 0.5337 1.6283 0.9526 2.3061 73.60 0.8628 0.0 3.3657 Source: Spalding and Chi (1964).9284 4.0087 0.7580 0.8686 10.7120 2.2556 1.3254 2.5989 0.3562 2.7176 2.6930 4.4393 4.0824 68.6871 6.6355 3.9093 20.2500 2.0834 0.00 1.0248 0.444 Values of Chap.5530 0.0 1.2748 0.0198 0.80 1.00 6.0759 2.7405 1.6631 1.8446 3.1572 2.7286 0.2649 1.0927 3.1185 0.8589 2.6178 0.2318 1.6437 256.4031 1.00 221.0153 0.0 6.6767 1.7986 5.8812 3.6969 2.0540 232.0295 1.7823 4.0143 0.6992 101.1352 0.00 5.1620 0.00 3.0000 1.8972 1.0 4.3445 1.6083 1.8687 3.7117 3.5487 1.05 0.9836 2.7359 3.6957 0.7938 6.6222 0.9971 4. TABLE 8.4203 28.2849 0.2471 0.0715 0.0 0.5588 4.1023 1.1079 0.9499 5.0394 0.0 0.1836 1.0842 1.0246 .0559 1.4 as a Function of Me and Tw/Te Me Tie 'e 0.20 0.6937 5.3611 91.4472 2.0 2.0184 1.7703 9.2410 0.00 6.3608 0.0856 0.0356 0.3658 4.7756 0.80 1.5744 1.7873 0.2796 2.6938 4.3370 1.0127 0.3311 0.2960 7.0 0.0269 0.5076 2.40 0.1036 0.0370 1.2759 1.3863 5.2611 3.0187 0.0465 14.4651 1.0516 0.30 0.4625 0.5126 2.1069 1.9036 1.30 0.0000 1.7015 3.9684 2.9747 3.5958 2.8783 1.
6 0.0040 0. 8.6 Values of FRe as a Function of FReReX FCCf FReRex 0.0085 0. the flow conditions and the surface temperature.0060 0.0010 0.0025 0.Sec. 8.0090 8.610 x 108 4.2 0.679 x 3. Since is known.340 x 106 2.0 0.651 x i07 9. The experimental skin friction coefficients are for adiabatic flows. The ratio of the experimental skin friction coefficient to the incompressible value at the same Reynolds number [as taken from Stalmach (1958)] is presented in Fig.0 0 1 2 3 4 5 Figure 8.0020 0.901 X 2.0055 0.251 x 1.6.0080 0.0015 0.592 x 1.758 x 1010 4.0075 0.796 x 2. the surface temperature was such that there was no heat transferred from the fluid to the wall.19 as a function of the Mach number.0045 0.078 x i04 1.0070 0. — SpaldingChi correlation (1964) o values Stalmach (1958) 1.828 X 2.0050 5.417 x 0. turbulent experimental skin friction coefficient to the incompressible value at the same Reynolds number as a function of Mach number.8 Cf 0.0030 0. we can solve for Cf.062 x 106 4.0065 0.0035 0.19 Ratio of the compressible. Then we can calculate the product FReRex and find the corresponding value of Cj in Table 8.778 X 106 1. we can calculate and FRe.492 X 1.4 0. .The experimental values are compared with those given by the SpaldingChi correlation.697 x 5. that is.006 X Source: Spalding and Chi (1964).
1 GroundBased Tests Despite the remarkable advances in hardware and software for computational fluid dynamics (CFD) tools.7 1.5) Thus. Computational fluid dynamics (CFD) solutions were used primarily to generate numerical values for use in comparisons with the experimental measurements. The second concern was related to the aerodynamics forces and moments during reentry. numerical methods. different facilities are used to address various aspects of the design . we obtain = so that 0. and plate. 8 I Dynamics of a Compressible Flow Field EXAMPLE 85: Skinfriction coefficient for a supersonic. and experimental programs using flighttest programs. = and 2. when two critical "returntoflight" (RTF) concerns were identified in the aerothermodynamic environment of the Space Shuttle.1729 (see Table 8.056 (see Table 8. Windtunnel tests provided the majority of the aerothermal information required by the "returntoflight" advisory teams. ReXFRe = 1.446 Chap.70 x THE ROLE OF EXPERIMENTS FOR GENERATING INFORMATION DEFINING THE FLOW FIELD The tools that are available to the designers of aerospace vehicles include analytical methods. 8. when Me = 2. experimental programs using groundtesting facilities. turbulent boundary layer What is the skinfriction coefficient for a turbulent boundary layer on a flat 6.142 X 106. One RTF concern dealt with the heating in the bipod region during launch.0035 Cf = 8.4) FRe = 0.6. Since there is no single groundbased facility capable of duplicating the highspeed flight environment. Most of the material defining the role of experiments in generating information defining the flow field has been taken from an article written by Bertin and Cummings for the 2006 Annual Review of Fluid Mechanics. very extensive windtunnel programs were conducted.5.OTe? Solution: For these calculations. = 3.062 x Using Table 8.7.
Pressure altitude 5.000 ft). On the other hand. e. and the freestream Reynolds number in conventional windtunnel facilities.480 m (100. Freestream velocity 4. one can simultaneously simulate the freestream velocity. the freestream Mach number. and the heattransfer distribution) are obtained in the following: 1. the relatively long exposure to the hightemperature flight environment causes the surface of the vehicle to become very hot. Most of the measurements that are used during the design process to. Density ratio across the shock wave 7. The temperature measurements that are used to determine the heattransfer rates to the model usually are made early during the test run.7 / The Role of Experiments for Generating Information Defining the Flow Field 447 problems associated with supersonic flight. the total enthalpy of the flow. Thus. Thermochemistry of the fibw field Note that some of the parameters are interrelated (e. Walltototal temperature ratio 9. the freestream velocity. and the walltototal temperature ratio). Conventional windtunnels 2. Shock tubes 4. Shockheated windtunnels 3. .. the temperature of model surface remains relatively cool during the data recording portion of the run. Archeated test facilities 5.5 to 4. the surface ptessure distribution. the aerodynamic forces and rhoments.000 ft) is approximately 950 K (1710°R). (1995)]. Two facilities in which one can match these four parameters are the Supersonic Aerothermal Tunnel C at the Arnold Development Center (AEDC) [Anderson and Matthews (1993)] and the LENS II facility at the CalspanUniversity at Buffalo Research Center (CUBRC) [Holden. during a specific rUn of a windtunnel test program. g. Test gas 8. the model temperature usually starts out at room temperature. freeflight ranges The parameters that can be simulated in groundbased test facilities include: 1. Note also. Total enthalpy of the flow 6.8. The critical heating to the bipod region occurs in the Mach number range from 3..480 m (100. et al.0—just prior to staging at approximately 30. Freestream Mach number 2. the walltototaltemperature ratio in groundbased tests is usually well below the flight value. Ballistic.define the aerothermodynamic environment (i. Because the total temperature for Mach 4 flow at 30. Freestream Reynolds number (and its influence on the character of the boundary layer) 3. To match the flight value of the walltototaltemperature ratio. As a result. that. the freestream Mach number. the total enthalpy of the flow.
448 Chap. even for this relatively benign flow. the model. the test engineer must decide which parameters are critical to accomplishing the objectives of the test program." Thus. such as boundarylayer transition and turbulence modeling. "A precisely defined test objective coupled with comprehensive pretest planning are essential for a successful test program. In fact. Thus. the instrumentation. (1986) noted that "Aerodynamic modeling is the art of partial simulation. Objective 5. In addition to the nine flowfield related parameters that were identified earlier in this section. Obtain data that can be used to develop empirical correlations for phenomena that resist analytical and/or numerical modeling. such as the heat transfer and the drag. one does not match the flight values of all of the flowfield parameters. Types of data available 4. . Some objectives include the following: Objective 1. Even today. et al. et a!. Objective 3. noise. Certify the performance of airbreathing engines. and the test conditions for each program tailored to answer specific questions. to be used in comparison with computed flowfield solutions over a range of configuration geometries and of flow conditions (code calibration). Obtain detailed flowfield data to be used in developing numerical models for use in a computational algorithm (code validation). Test time 3. extensive groundbased testprograms are conducted during the design process in order to define the aerothermodynamic environment for the entire vehicle. Objective 6. Obtain measurements of parameters." There are many reasons for conducting groundbased test programs. (1984). such as the inlet flow field for hypersonic airbreathing engines or the shock/boundarylayer interactions associated with deflected control surfaces (a body flap). Use partial configurations to obtain data defining localflow phenomena. cleanliness. the designers most likely will utilize many different facilities with the run schedule. additional factors must be considered when developing a test plan. and/or the heattransfer distributions for complete configurations whose complex flow fields resist computational modeling. When the access to space study that was conducted by NASA in the early 1990s recommended the development of a fully reusable launch vehicle (RLV) [Bekey et al. As stated by Matthews. Leith Potter. joined an industryled technologydevelopment effort for the X33/RLV. Flow quality (including uniformity. Trimmer. The additional factors include: 1. Objective 4. 8 / Dynamics of a Compressible Flow Field the total temperature of the tunnel flow is reduced below the flight value. during the development of a particular vehicle. Obtain data to define the aerodynamic forces and moments Objective 2. Model scale 2. and steadiness) In a statement attributed to J.
and the criteria for boundarylayer transition to Lockheed Martin in support of X33 development and design. One article [Berry.turbulence modeling." They are Hollis. et a!. (2001). (2001a)} related to objective (5) "Obtain data that can be used to develop empirical correlations for phenomena that resist analytical and/or numerical modeling. the HyperX Research Vehicle also called (HXRV or free flyer) required a booster to deliver the vehicle to the engine test points. et al. et al. such as the heat transfer and the drag. documenting the results from this cooperative effort. (2001a) and Hollis. complete configurations whose complex flow fields resist computational modeling. et al. "The primary goals of the HyperX program are to demonstrate and validate the technologies. (2001) and Murphy. Two articles presented information relating to objective (1) "Obtain data to define the aerodynamic forces and moments and/or the heattransfer distributions for complete configurations whose complex flow fields resist computational modeling. One article presented information relating to objective (4) "Obtain measurements of parameters. a range of configuration geometries and of flow conditions (code calibration). for the X43. et al. et al. et a!.7 / The Role of Experiments for Generating Information Defining the Flow Field 449 As part of the industry/government partnership. and the flighttest program." "In 1996 NASA initiated the HyperX program as part of an initiative to mature the technologies associated with hypersonic airbreathing propulsion. A special section in the September—October 2001 issue of the Journal of Spacecraft and Rockets presented five archival journal articles. One article [Berry.956 rn (95. including information relating to the boundarylayer transition criteria." As described by Woods. Two articles presented information relating to objective (4) "Obtain measurements of parameters. 8. dualmode scramjet propulsion systems. such as boundarylayer transition and turbulence modeling. personnel and facilities at the Langley Research Center (NASA) were assigned the task of providing information regarding the aerodynamic forces and moments.879 N/rn2 (1000 The test conditions for the first flight included = 7 and psf) at an altitude of approximately 28." Although hypersonic airbreathing propulsion systems have been studied in the laboratory environment for over forty years. (2001). such as the heat transfer and the drag.Sec. the development. (2001). (2001) and Holland. A special section in the November—December 2001 issue of the Journal of Spacecraft and Rockets presented seven archival journal articles summarizing the results from this program. Again. personnel and facilities at the Langley Research Center (NASA) were used to define the aerodynamic and surface heating environments. et a!. (2001b). The HyperX Launch Vehicle ." That was by Cockrell.000 ft). Two articles presented information relating to objective (1) "Obtain data to define the aerodynamic forces and moments and/or the heattransfer distributions for . (2001)." They are Horvath." They are Engelund. (2001b)] related to objective (5) "Obtain data that can be used to develop empirical correlations for phenomena that resist analytical and/or numerical modeling. and the computational methods and tools required to design and develop hypersonic aircraft with airframeintegratçd. to be used in comparison with computed flowfield solutions over. et al. et a!. a complete airframeintegrated vehicle configuration had never been flight tested. to be used in comparison with computed flowfield solutions over a range of configuration geometries and of flow conditions (code calibration). as part of the design. the experimental techniques." [Engelund (2001)]. the surface heating. such as boundarylayer transition and. 47. the construction.
" . To demonstrate interactive technologies and to identify unanticipated problems. it is then desirable to make as many measurements as possible during the flight to help verify predictions or explain any modeling inaccuracies. FurthermOre. The instrumentation should measure as directly as possible the things that have been predicted. When the stack reached the desired test conditions and attitude. the Pegasus ignited. 6. 2. To gain knowledge not only from the flights but also from the process of development. Williamson (1992) noted: "Due to the large investment in flight testing. flight tests are critical to our understanding of the hypersonic aerothermodynamic environment. The HXLV was dropped. research vehicle.7. since they provide data which cannot be obtained elsewhere. To validate the overall performance of the system. This understanding was complicated by the unsteady nature of the event. 4. et al. stage separation was critical to reaching the engine test point and. To form a catalyst (or a focus) for technology. a stageseparation sequence of events separated the free flyer from the booster. The freeflying research vehicle then followed a preprogrammed trajectory. Four reasons were suggested by Buck et al. These issues were in three basic areas: unsteady effects. the number of degrees of freedom associated with the booster." Viscous and inviscid computational fluid dynamics (CFD) techniques were used to quantify unsteady effects. Flight tests will never replace groundbased tests or CFD in the design process.2 Flight Tests Flight tests are very expensive.This is often not possible and it is often necessary to infer predictions from related but not direct measurements. Neumann added three reasons of his own: 5. 8. (1963): 1. and limits in the amount of windtunnel data available. several aerodynamic issues remained in fully understanding the dynamics of the stage separation maneuver. it is difficult to obtain quality data at welldefined test conditions. 3. Buning. Neumann (1988) suggests that there are a variety of reasons for conducting flight tests. Nevertheless. hence. 7. To verify groundtest data and/or to understand the bridge between groundtest simulations and actual flight. To demonstrate technology in flight so that it is credible for largerscale applications. Flight tests take a long time to plan and to execute suc cessfully. aerodynamic database extrapolation. was critical to mission success. 8 / Dynamics of a Compressible Flow Field (HXLV) stack was initially carried aloft under the wing of a B52. Although occurring in less than 500 ms. and to identify differences between the windtunnel and the flight environments. (2001) reported that: "Even following the AEDC test. To generate information not available on the ground.450 Chap. and the stack accelerated to the desired test Mach nuniber. and control surfaces. and differences between windtunnel and flight conditions. to examine the cause and the extent of interference between the booster and the research vehicle.
However. g. and Reynolds number.7 / The Role of Experiments for Generating Information Defining the Flow Field Neumann (1989) stated: "Heat transfer is a quantity which cannot be directly measured: It is interpreted within the context of a thermal model rather than measured. several programs successfully generated aerothermodynamic flight data to improve the understanding and interpretationof theoretical and ground test results." Thus. The data reduction program correlates heating with the variables such as angle of attack. NASA (1995)]. sideslip. control surface deflection. (1972) and Sutton (1985)]. The FIRE program provided flight data that was used to understand radiative heattransfer that become important for flight through the earth's atmosphere and when returning from lunar and from planetary missions [Ried. Williamson (1992) notes that "Measurements fall into three groups. Serious problems can occur when the heat sensor is not properly integrated into the flight structure such that minimal thermal distortion is produced. et al. [e. Flighttest data have been obtained on prototype vehicles. As noted by Hodge and Audley (1983)." Project Fire provided calorimeter heating measurements on a largescale blunt body entering the earth's atmosphere at an initial velocity 11. The ASSET and PRIME programs were flown in the early 1960's and provided aerothermodynamic flight data for ablative and metallic thermal protection (TPS) concepts.. "Requirements for the technique include an analytical model for the simulation of the heat transfer to a point on the TPS. and a parameter estimation program to reduce flight thermocouple data. offboard vehicle related sensor measurements. This multiplelayer arrangement provided three distinct time periods when measurements defining the aerothermodynamic environment could be obtained. the flights of the unmanned Apollo Command Modules (017 and 020) as discussed in Lee and Goodrich (1972)] and on operational vehicles [e. . R and D programs are focused on technical issues which drive the design of the vehicle and its flight operations. 8. one must be able to develop a numerical model that describes the relation between the measured temperature and the sensor's design. the forebody of the "Apollolike" reentry capsule was constructed of three beryllium calorimeter shields. This technique could also be used for wind tunnel data reduction.Throckmorton (1993) noted that the concept of using the Shuttle Orbiter as a flight research vehicle as an adjunct to its normal operational mission was a topic of discussion within the research community throughout the 1970s. As discussed by Cornette (1966). which were alternated with phenolicasbestos heat shields." There are two types of flighttest programs: (1) Research and Development (R and D) programs and (2) flights of prototype or operational vehicles. These include atmospheric properties measurements. the Space Shuttle Orbiter.35 km/s (37.24 kft/s). g. Iliff and Shafer (1992) noted that: "In the 1960's. one also must develop a numerical model depicting how the heattransfer sensor responds from its location in the vehicle.. in order to obtain a reasonably accurate value of the experimentallydetermined heat transfer.Sec." .. and onboard vehicle related measurements telemetered to the ground or stored on tape and retrieved postflight. flight test maneuvers which cause the thermocouples imbedded near the surface of the TPS to respond sufficiently above the noise levels. Aerothermodynarnic parameters based on flighttest data obtained from thermocouples embedded in the Space Shuttle Thermal Protection System (TPS) were used to expand the flight envelope for the Orbiter..
the relevant parameters for the flight condition include . Aerodynamic coefficients for the missile. (2) the Mach number of the approach flow. The upstream extent of the interactionproduced perturbations de pends on the size of the subsonic portion of the approach boundary layer and on the strength of the shock wave produced by the turning of the flow. Texas).88 kmls (12.452 Chap. The HYFLEX vehicle [Fujii et al. Using the properties for a standard atmosphere (such as were presented in Chapter 1). the relations developed in this section can be used to calculate the windtunnel test conditions.40 m (14. (2001) investigated the perturbed flow fields due to deflected control surfaces using the experimental values of the heat transfer measured during hypersonic flight. EXAMPLE 8.A twostage launcher provided an initial altitude of 107 km (351 kft) at a velocity of 3." Also flown in the 1960's was the Reentry F vehicle which provided hypersonic boundarylayer transition and is still used today. (2001)] was a smallscale R and D vehicle whose length was 4. 8 / Dynamics of a Compressible Flow Field The interactions between impinging shock waves and the bow shock wave can produce locally severe heattransfer rates. are required for angles of attack from 0° to 55°. The adverse pressure gradient produced by the shock wave causes the boundary layer to thicken and (in many cases) to separate. The Reentry F flight provided boundarylayer transition data at Mach numbers up to 20 and altitudes down to 24. there are numerous flighttest programs where boundarylayer transition was a major focus of the data gathering efforts. (3) the Reynolds number of the approach flow. Solution: 1. Weston and Fitzkee (1963) noted that observed boundarylayer transition Reynolds numbers on the Mercury capsule "agree well with Reynolds numbers obtained for wake transition behind spheres flown in a hypervelocity gun facility. Edney (1968) identified six different shock/shock interactions patterns. Despite the oftmentioned problems with developing boundarylayer transition correlations using flighttest measurements.38 km (80. Upstream of the model shock system. The decision is made to obtain experimental values of the required coefficients in the Vouglit HighSpeed Wind Tunnel (in Dallas. Bertin et a!.0 cm (7. Flight conditions.00 kft) [Wright and Zoby (1977)]. as noted by Bertin (1994). The positive deflection of the Orbiter body flap creates a shock wave which interacts with the boundary layer on the Orbiter.73 kft/s). (5) the deflection angle of the ramp. Thus.5 at an altitude of 27.6: Windtunnel simulation of supersonic missile flow fields Assume that you are given the task of determining the aerodynamic forces and moments acting on a slender missile which flies at a Mach number of 3.000 ft). (4) the surface temperature. the flow in the wind tunnel is isentropic and air.874 in) in diameter and 10 diameters long. Thus.44 ft).432 m (90. at these conditions. which is 20. (1996) and Fujii et al. the parameters that influence the extent of an interaction are (1) whether the ap proach boundary layer is laminar or turbulent. and (6) the chemical state of the gas. behaves as a perfect gas.
6468 in.e.34) and (8. Windtunnel conditions.To ensure that the model is not so large that its presence alters the flow in the tunnel (i.7379 X IO2P5L = 1760.0 x 106/ft.936 X 10 Jwt 2. since the pressure coefficients and the shock interaction phenomena on the control surfaces are Machnumberdependent in this range of freestream Mach number. which is greater than the flight value.183 cm (1. In order to match the flight Reynolds number of 3.The student can use the equations of this section to verify the value for the unit Reynolds number given the conditions in the stagnation chamber and the freestream Mach number.18 K (162. The freestream unit Reynolds number is presented as a function of the freestream Mach number and of the stagnation pressure for a stagnation temperature of 311 K (100°F) in Fig.32°R) and N/rn2 (1.183 cm in diameter. the authors would choose to simulate the Mach number exactly rather than seek a compromise Machnumber[Reynoldsnumber test condition. To obtain the appropriate Reynolds number for the current supersonic we can choose to use a smaller model. 311 K The conditions in the stagnation chamber of the tunnel are (560°R) and Pti = 5. where (as discussed in Chapter 5) the maximum windtunnel Reynolds number is usually much less than the flight value.Thus.20 (which has been taken from Arnold (1968) and has been left in English units). one finds that = 90.868 X 106/ft But as indicated in Fig. Thus.20. Using a model that is 1. the Mach number and the Reynolds number are two parameters which we should try to simulate in the wind tunnel.Sec.231 x . 8.).5 is approximately 9.333 cm in diameter would yield a Reynolds number of 3. Based on the discussion in Chapter 2. 8. If this model is too small. the diameter of the windtunnel model.. the model dimensions are within the allowable blockage area). the lowest unit Reynolds number possible in this tunnel at = 3.936 X i05 in the wind tunnel.7 / The Role of Experiments for Generating Information Defining the Flow Field 453 = 1050rn/s = 1.36) or the values presented in Table 8. = 3. This is much different than the typical subsonic flow.936 x i05 (as desired). will be 4. 8.235 X 106.049 psia). In this case.9 N/rn2 = 224K = d 3.The cold = 7. Information about the operational characteristics of the Vought HighSpeed Wind Tunnel is contained in the tunnel handbook [Arnold (1968)].516 X N/rn2 (80 psia). using either equations (8. the lowest possible tunnel value of is 1.1. if the model is 4.936 X i05 2. we cannot establish a tunnel condition which matches both the flight Mach number and the Reynolds number.
231 X N/rn2 . most highspeed wind tunnels operate at temperatures near liquefaction of oxygen)..g.). the relevant parameters for the windtunnel conditions include = 665 rn/s 7. Thus. 8 / Dynamics of a Compressible Flow Field Constant dynamic pressure (q. the velocity in the test section mIs (2185 ftls).454 Chap.} freestream temperature is typical of supersonic tunnels (e. [From Arnold (1968). even though the is only 665 freestream Mach number is 3. the freestream speed of sound is relatively low and..5.20 Freestream unit Reynolds number as a function of the freestream Mach number for — 100°F in the Vought High Speed Wind Tunnel. psf —. In summary. — Marginal operating limits 36  Constant stagnation pressure (Pu) psia A it 32  28 24 'a x 20 16 12  8— 4 0 0 1 2 3 4 5 6 Figure 8.
21a. and two were right on." Citing the work of other researchers.235 X if d 4. 'surprise' occurs far too often and flow complexity can seriously degrade the 'scaled predictions'. constructed by three major American manufacturers over a 20year period showed that you are just as likely to estimate too high a drag as too low a drag. the Viking aeroshell drag observed during Martian entry agreed within some 2% of the scaled ground results. four predictions were high. Bushnell (2006) noted that "a survey of 12 commercial transport aircraft. and —4% sting and aeroelastic distortion effects. 8. Recall that pressure disturbances can affect the upstream flow only if the flow is subsonic. depends on the character of the upstream boundarylayer. Typical correction levels include 6% wall interference. The extent of the separation. As a result. locally severe heating rates or boundarylayer separation may occur. admittedly for a very bluff body with a fairly well characterized flow field. The thickening boundary layer deflects the external stream and creates a series of compression .5 Because of the significant differences in the dimensional values of the flow parameters (such as and we must again nondimensionalize the parameters so that correlations of windtunnel measurements can be related to the theoretical solutions or to the design flight conditions.333 cm or 1." 8. which can cause a loss of control effectiveness.8 COMMENTS ABOUT THE SCALING/CORRECTION PROCESS(ES) FOR RELATIVELY CLEAN CRUISE CONFIGURATIONS In an article for the Annual Review of Fluid Mechanics.936 X if d 1. 8. Bushnell (2006) noted that similar values/discrepancies were obtained for airbreathing missiles. —5% Reynolds number. +2% roughness. The shock wave generated by a deflected flap will interact with the upstream boundary layer. On the other hand. the boundarylayer thickness increases and the momentum decreases.9 / ShockWave/BoundaryLayer Interactions 455 = 3. The drag predictions were as much as 22% low and 10% high.3.9 SHOCKWAVE/BOuNDARYLAYER INTERAcTIONS problems of locally high heating or premature boundarylayer separation may result due to viscous/inviscid interactions which occur during flight at supersonic Mach numbers. "As the experience from the various appliOations discussed herein indicates. Again.Sec. Total correction(s) from just these issues/concerns/ef fects are the order of + 12%. Six of the twelve (scaling) 'predictions' were low. The pressure rise induced by the shock wave is propagated upstream through the subsonic portion of the boundary layer.183 cm Moc. The basic features of the interaction between a shock wave and a laminar boundary Severe layer for a twodimensional flow are shown in Fig. 8. The interaction will generally cause the upstream boundary layer to separate with locally high heating rates occurring when the flow reattaches. Other viscous interaction problems can occur when the shock waves generated by the forebody and other external components impinge on downstream surfaces of the vehicle.
?/' oundary layer Separated region p Pressure of undisturbed ecompression shock wave Pressure far downstream of upstream flow the interaction (a) Shock structure p4 Pressure of undisturbed upstream flow (b) Pressure far downstream of the interaction x Figure 821 Flow field for shockwave boundarylayer interaction: (a) laminar boundarylayer. the flow reattaches at some distance downstream. If the flow reattaches.21b). the flow may or may not reattach. Immediately downstream of reattachment. the characteristics of the boundary layer. such as an airfoil. Furthermore. a second shock wave (termed the reattachment shock) is formed. For a flat plate. since the subsonic portion of a turbulent boundary layer is relatively thin. the boundarylayer thickness reaches a minimum. and the strength of the shock wave. In the case of a shock interaction with a turbulent boundarylayer (Fig. 8. This results because the air particles near the wall of a turbulent boundarylayer have greater momentum than do those near the wall in a laminar boundarylayer and can therefore overcome a greater adverse pressure gradient. 8 I Dynamics of a Compressible Flow Fields PrandtMeyer fan Ashock //. (b) turbulent boundarylayer. If the shockinduced adversepressuregradient is great enough. the skin friction will be reduced to zero and the boundary layer will separate. waves to form a Alike shock structure. a PrandtlMeyer expansion fan results as the flow turns back toward the surface. . depending upon the body geometry. the length of the interaction is considerably shorter than the interaction length for a laminar boundary layer. In the case of a convex body. It is in this region that the maximum heating rates occur.__________________ 456 Chap. As the flow reattaches and turns parallel to the plate. The subsequent behavior of the flow is a strong function of the geometry.
231 x N/m2. answer the following. A convergentonly nozzle..4. 83. The static temperature should not be allowed to drop below 50 K. The test section of the wind tunnel of Example 8. With this value for the stagnation (or total) temperature. 8. and = 90. which exhausts into a large tank. since at low pressure oxygen begins to liquefy at this temperature. The Mach number in the test section is 8.e.2) if any.1000 K. PROBLEMS 8.75 m. To what range of altitude (as given in Table 1.2) do these pressures correspond? 8.18 K)? (b) What is the volume flow rate of air through the test section? 8. state why. The fluid in the free stream undergoes an isentropic expansion.5 . what is the range of Reynolds number that can be obtained in this facility? Assume that the characteristic dimension is 0.Problems 457 the region through which the shockinduced pressure rise can propagate upstream is limited. Pressure distributions typical of the shockwave/boundarylayer interactions are presented in Fig. As one would expect from the preceding description of the shock interaction. A practical limit for the stagnation temperature of a continuously operating supersonic wind tunnel is . With this as a lower bound for temperature.1. what is the maximum Mach number for the facility? Assume that the air behaves as a perfect gas with y = 1.2. 8.6 has a square cross section that is 1. As a result. Assuming that the air behaves as a perfect gas.90 x i05 N/rn2 to 5. a much greater pressure rise is required to cause a turbulent boundary layer to separate.22m. is used as a transonic wind tunnel (Fig. (a) What is the massflow rate of air through the test section for the flow conditions of the example (i.21 for a laminar boundary layer and for a turbulent one. = 7.5.22m X 1.3. = 3. If the stagnation temperature is 725 K and the stagnation pressure can be varied from 6. P8. which could be the length of a hypersonic waverider model. prepare a graph of the static freestream temperature as a function of the testsection Mach number. (a) What is the (range of the) static temperature in the test section? If there is only one value of the static temperature. the pressure rise is spread over a much longer distance when the boundary layer is laminar. reservoir —3— T. do these temperatures correspond? (b) What is the (range of the) velocity in the test section? Does it depend on the pressure? (c) What is the range of static pressure in the test section? To what range of altitude (as given in Table 1.90 x 106 N/m2.=40°C Atmospheric pressure Large tank Figure P8. Tunnel B at the Arnold Enginering Development Center (AEDC) in Tennessee is often used for determining the flow field and/or the heating rate distributions. Given the flow conditions discussed in Problem 8.4.5).5.
50 X i05 N/rn2. (c) 10 km.7 . Assume that the cabin volume is sufficiently large that the cabin serves essentially as a stagnation chamber and that the conditions in the cabin remain constant for a period of time.01325 X N/rn2). 8 / Dynamics of a Compressible Flow Field (a) If the pressure in the tank is atmospheric (i. the back pres.. T. what is the static temperature in the test stream? for (c) A transonic airfoil with a 30cm chord is located in the test stream. (b) 8 km.6 and 8. 1. independent of the conditions outside the airplane.6.47218 0.458 Chap. the pressure outside the airplane). What is the airfoil? /too (d) What is the pressure coefficient at the stagnation point of the airfoil? 8.. Cabin Nozzle exit plane Figure P86 Sketch for Problems 8.pb. P8. what should the stagnation pressure in the nozzle reservoir be so that the Mach number of the exhaust flow is 0. For the standard atmosphere: (a) Pb = (b) Pb = (c) Pb = 0. as shown in Fig.e.53000Pti Where Pu = Pc = X i05 N/rn2.19399 x N/rn2 = x i05 N/rn2 = x 1O5 N/rn2 = x i05 N/rn2 = 0.35652 0. sure (i. A small hole exists in the body of an airplane and serves as a convergent nozzle..6. Furthennore.26500 (d) Pb = 0. The air in the cabin is at 0.80? (b) lithe stagnation temperature is 40°C. Air outside the airplane p. Since the air expands isentropically from constant reservoir conditions.. assume that the flow in the nozzle is isentropic. Sketch the pressure distribution and calculate the static pressure in the nozzle exit plane when the airplane is at the following altitudes: (a) 6 km. is an important parameter. and (d) 12 km.e.5 °.
The pressure in the test section can be varied by controlling the valve to the vacuum tank. there is no flow. Since the back pressure is equal to the stagnation pressure for condition (a). T.035 times the throat area. Thus. Pi = 100. P8.9. If the temperature in the cabin is 22°C and the exit diameter is 0.7.6 (continued) Solution for Problem 8.117 psia Let us first convert the stagnation temperature to °R. = 659. what is the mass flow rate through the hole of Problem 8.75 cm. Pb Pti (a) (b) Figure P8.000 psia (b) Pb = 97.8.8. U.8. The crosssectional area of the test section is 2. for (a). T1 = 659.6°R. The crosssectional area of the test section is 2.38 psia (b) Pb = 75.6°R. and velocity in the test section for the following back pressures: (a) Pb = 51. M1 = 0. static temperature. (a) Pb = 100. The conditions in the stagnation chamber are Psi = 100 psia and T5 = 200°F. Consider the flow of air through the convergentdivergent nozzle shown in Fig. Mach number.6 (i) pressure distribution: (ii) pressure ratios.86 psia .9. = 0 8. static temperature.000 psia. 8. and velocity in the test section for the following back pressures.. Mach number.Problems 459 PC — Pne Pb (a) Pt' P Pti "C.The conditions in the stagnation chamber are Pti = 100 psia and T5 = 200°F. Thus far we have repeated the condi tions of Problem 8. Calculate the static pressure.250 psia (c) Pb = 93.6 when the altitude is 6 km? 12 km? 8. P8. calculate the static pressure.035 times the throat area.947 psia (d) Pb 9. Assuming isentropic flow in the nozzle. Consider the flow of air through the convergentdivergent nozzle shown in Fig.
x J (second critical ratio) (third critical ratio) d Figure P8.Valve To vacuum tank Back Stagnation chamber Throat tl Test section 1 a c (first critical ratio) x d (third critical ratio) Figure P8. e S..8 Isentropic flow in a convergentdivergent nozzle. S.7) 460 .9 Flow in a convergentdivergent nozzle with a shock wave in the divergent section (dashed lines are from Figure P8. Valve To vacuum tank pressure Stagnation chamber Throat I Test section a b ————— c (first critical ratio) p Pu f S.
12). Air flows through the insulated variablearea streamtube such that it may be considered one dimensional and steady.12. it is necessary to obtain theoretical pressures for comparison with the data.13. What is the flow direction.11. A pitot tube in a supersonic stream produces a curved shock wave standing in front of the nose part.12 8. P8. Derive an expression for in terms of the stagnation (Pr2) and static (P2) pressures sensed by the probe.. that is. P8.0 ft2. and A2 5. >1 I { Figure P8. Consider the flow in a streamtube as it crosses a normal shock wave (Fig.The probe is designed to sense the stagnation pressure behind a normal shock (Pt2) and the static pressure behind a normal shock (P2).2 cm Hg.11. The air flows from right to left. = 500°R.Problems 461 8. If a 30° \vedge is to be placed in a Mach 3. Assume that the probe is at zero angle of attack and that the shock wave is normal in the vicinity of the nose.10.0 X metric pressure is 75. Pt2 = 2116 psfa.0. P8. is the flow from (1) to (2) or from (2) to (1)? 8. At one end of the streamtube.  pressure Wall 'orifice utube manometer Figure P8. (b) What are the limits of A/At as M1 1 and as M1 oo? (c) What is the significance of 4? of M1> 1 (2) Figure P8. P2 = 2101 psfa. as shown in Fig. A1 = 5. To evaluate the experimental accuracy.13 .11 8. You are to measure the surface pressure on simple models in a supersonic wind tunnel. At the other end of the streamtube. calculate (a) The surface pressure in N/rn2 (b) The pressure difference (in cm Hg) between the columns of mercury in Utube manometer between the pressure experienced by the surface orifice and the wall orifice (which is used to measure the static pressure in the test section) (c) The dynamic pressure of the freestream flow N/m2 and the baroOther measurements are that the pressure in the reservoir is 6. (a) Determine the ratio AJ4.13).0 ft2 and M1 = 3.5 stream (Fig.
(a) Determine the final Mach number and net change in entropy. (a) Use the oblique shockwave relations to calculate the static pressure in region (2) in terms of the freestream value (b) Use the PrandtlMeyer relations to calculate the static pressure in region (3) in terms of (c) Calculate C1. P1 Figure P8. or (3) an infinite number of infinitesimal turns. If the radius of the engine inlet is 1.5 (i. such that the conical shock just grazes the lip of the nacelle. the lower surface deflects the flow 5° (Fig..0 m. Do these coefficients depend on the freestream pressure (i.17).e. the altitude)? — Expansion fan 2. P8.The freestream Mach number is 2. The upstream Mach number is 3.15). Chap. can we make any conclusions as to whether it is better to make expansive turns gradually or abruptly. A conical spike whose halfangle is 8° is located in the inlet of a turbojet engine (Fig.. whose length is c. Consider the twodimensional inlet for a turbojet engine. determine 1.17 . 8 / Dynamics of a Compressible Flow Field It is desired to turn a uniform stream of air compressively by 100. when the flow is supersonic? 8. if the turning is accomplished by (1) a single 100 sharp turn.80 stream such that theangle of attack of the spike is zero. is in a Mach 2.The engine is operating in a = 2. P8.0 stream at an angle of attack of 10° (Fig.0 (1) Plate of length c k wave Figure P8.Om S Figure P8. A flatplate airfoil.____ 462 8.17. M1 = 2. the length of the spike extension.e.5) and the pressure Pi is 5. (2) two successive 50 sharp turns. P8. The upper surface deflects the flow 100.15.16. (b) What do you conclude from the results in part (a)? (c) In light of the results in part (b). Cd.14.15 8. Calculate the static pressure. and Cmø5c (the pitching moment about the midchord). r=1.16 8.0 )< i03 N/m2.16).
When the weak. An airplane flies 600 mi/h at an altitude of 35. is at zero angle of attack.19. it is reflected so that the flow in region (3) is parallel to the tunnel wall.4.50 N/rn2. it is not a good assumption for calculating the flow across the shock wave. the static pressure in region (4) is equal to that in region (5).0 wind tunnel (Fig. Use the normal shock relations to calculate the values of the following parameters downstream of the normal shock wave: P2: P12: 7'2: T12: static pressure stagnation pressure Static temperature stagnation temperature M2: Mach number What is the pressure coefficient at the stagnation point. The perfectgas assumption is valid for the freestrearn flow. and 0405 That is.18).000 ft where the temperature is —75° and the ambient pressure is 474 psfa. and (5).18. oblique shock wave generated at the leading edge of the airfoil encounters the wall. If the test section is 30. 8.Problems 463 the Mach number. The airfoil.0 cm high. respectively. there is a region where the shock wave is essentially normal to the freestream flow. Note that since re gions (4) and (5) are divided by a fluid/fluid interface. what is the maximum chOrd length (c) of the airfoil so that it is not struck by the reflected shock wave? Neglect the effects of shockwave/boundarylayer interactions at the wall. P8. However. which has a halfangle 5 of 50. . many of the perfectgas theoretical values for the shockflow properties will not even be close to the actual values. Pt2 — Poo Use the perfectgas relations and assume that y = 1.62 km/s at an altitude of 75. Therefore. (3). (4).0 km. as illustrated in Fig.15 K and 2. What is the temperature and the pressure of the air (outside the boundary layer) at the nose (stagnation point) of the airplane? What is the Mach number for the airplane? 8. Section AA Figure P8. A threedimensional bow shock wave is generated by the Shuttle Orbiter during entry However.18 8.10. See Chapter 12 for further discussions of this problem. and the flow direction in regions (2). that is. A single wedge airfoil is located on the centerline of the test section of a Mach 2. The velocity of the Shuttle is 7. and the flow direction in region (4) is equal to that in region (5).20. The freestream temperature and the static pressure at this altitude are 200. 8.
calculate the static pressure at the surface of the cone (Pc). Repeat Problem 8. in practice. static temperature.15c. 6 12° in Fig. Calculate the stagnation pressure of the flow downstream of the shock wave. 1993. Powell R. 1968. except for the shock wave itself (i. and charts for compressible flow.12). Equations. 8. 8.21 for a planar symmetric wedge that deflects the flow 12° (i. 2001. neglect the bound ary layer on the cone). and velocity of the air behind the shock wave? Moving shock wave Fixed shock wave 1tds — ) Standard sealevel atmosphere P2 U2 U1 = 1000 m/s Ground (a) Ground (b) Figure P8. The cone semivertex angle is 12°. Calculate the Reynolds number at a point 10.23 (a) traveling blast wave. Using Fig. Consider the hypersonic flow past the cone shown in Fig. tables. Using Fig.15b.e.0 cm from the apex of the cone.5 1970K = 1070 N/rn2 Assume that the flow is inviscid. NASA access to space. The atmospheric conditions ahead of the shock wave are those of the standard sealevel atmosphere.23 Blast wave for Problem 8. 8.22. (b) transformed steady flow. Assume further that the gas obeys the perfectgas laws with = 1. In Methodology of Hypersonic Testing. calculate the Mach number at the surface of the cone (Me). 8. Vought Aeronautics Division AEREIR13552B Bekey I.VKI/AEDC Special Course.The freestream flow is defined by 11. What are the static pressure.. Matthews RK. 8. REFERENCES Ames Research Center Staff. Aerodynamic and Aerothermal Facilities.21. I Hypersonic Wind Tunnels.464 Chap. Austin R. RhodeSaintGenèse. 1953. 8 / Dynamics of a Compressible Flow Field 8.4.. NACA Report 1135 Anderson A.e. An explosion generates a shock wave that moves through the atmosphere at 1000 m/s. Aerospace America 32(5):38—43 . which is. Belgium Arnold JW.16. the Mach number at the edge of the boundary layer. High Speed Wind Tunnel Handbook.
Shirozu M. Nugent J.7. Vol. Rockets 38:658—669 Hollis BR. J. Spacecr. Calleja JF. Hypersonic boundarylayer trip development for HyperX. NASA TMX1 035 Edney BE. Ed. 95—6040. Aerodynamic and performance analyses of a superorbital reentry vehicle. Rockets 38:828—835 Hollis BR. Spacecr. Planes Hypers. Woods WC. 2001. Chadwick K. Presented at Intl. Holland SD. X33 aerodynamic computations and comparisons with windtunnel data. Dilley AD. Spacecr. 1968. Alter SJ. 1966. 2001. Berry SA. J. 1970. Rockets 38:853—864 Bertin JJ. Cockrell CE. Bouslog SA. 2001. 36: 111—128 Chapman AJ. DC: AIAA Bertin 11. Auslender H.Audley SR. 2001. Hamilton HH. I. Benson BR. Rockets 38:803—810 Fujii K. Cummings RM. 2001a. Harrison AC. AIAA J. 2006.TN Holland SD. Annu..Watanabe S.2006.Murphy KJ. J. Engelund WC. Probstein RE 1966.Nowak RJ.Wang KC.Annu. Bittner RD. Spacecr. J. Forebody temperatures and calorimeter heating rates measured during project Fire II reentry at 11. New York: John Wiley Buning PG.Alter SJ. Hypersonic A erotherinodynamics. 1974. X33 hypersonic boundarylayer transition.Aerodynamic heating measurements on nose and elevon of hypersonic flight experiment vehicle. Hollis BR. 1996. Recent aerothermodynamic flight measurements during Shuttle Orbiter reentry. IIyperX aerodynamics: The X43A airframeintegrated scramjet propulsion flighttest experiments. Conf. 1983. Computational fluid dynamics prediction of HyperX stage separation aerodynamics.KurotakiT. Frendi. 2001.35 kilometers per second. Spacecr. Horvath TJ. Integrated aeropropulsive computational fluid dynamics methodology for the HyperX flight experiment. 2001. J. Horvath TJ. Dilley AD. 1963. HyperX research vehicle experimental aerodynamics Ttest program overview. Sieron TR. 2001b. In Dynamics of Manned Lifting Planetary Entry. Hamilton HH. 2001a. Engelund WC. Pap. 2001b. Jentink TN.References 465 Berry SA.Tech. Rockets 38:8—14 Hayes WD.J. Rockets 33:457—462 Berlin JJ. Rockets 20:453—460 Holden M. New York: Macmillan Cockrell CE. Campbell CH. Rockets 38:684—691 . Spacecr. Pao JL. Spacecr. Rev. Rev. 1995. Local flow field around a pylonmounted dummy ramjet engine on the X152 airplane for Mach numbers from 2. Wood WA. Rockets 38:801—802 Engelund WC. Note D5638 Bushnell DM. Hypersonic Flow Theory.. Bittner RD. Spacecr. Rockets 38:836—843 Comette ES. Engelund WC. Hypervelocity studies in the LENS facility. Thompson RA. Aerodynamic database development for the HyperX airframeintegrated scramjet propulsion experiments. J. Heat Transfer. Dilley AD. . Spacecr. J. J. Rockets 38:820—827 Burcham FW. Aerothermodynamic parameter estimation from shuttle thermocouple data during flight test maneuvers. Effects of shock impingement on the heat transfer around blunt bodies. X33 com putational aeroheating predictions and comparisons with experimental data. Inviscid Flows. Rogers M.1. Neumann RD. 6:15—21. 1994. Chattanooga. Rockets 38:646—657 Berry SA. Spacecr. 3 8:129—157 Buck ML. Scala SM. Spacecr.Thompson RA.Prabhu R. Kolly J. Aerosp.Thompson RA. Fluid Mech. New York: Academic Press Hodge JK. A. Wong TC.0 to 6. Washington. J. Fluid Mech. Riley CJ. Spacecr. Critical hypersonic aerothermodynamic phenomena. Scaling: wind tunnel to flight. NASA Tech. J.
Presented at Aerodyn. 1988. NASA TMX58091 Spalding DB. Presented at Thermophysics Conf. Nashville. Liechty DS. Ed. France Neumann RD. DiFulvio M. AIAA Pap. Nowak RJ. In Progress in Astronautics andAeronautics Vol. lOth. 86—0739. Fitzkee AL. DRL410 Sutton K. Kidd CT. Composition and thermodynamic properties of air in chemical equilibrium. Nashville. Williamsburg. Princeton.419—441 Throckmorton DA. New York: McGrawHill Lee DB. 92—3988. 1957. at Fluid and Plasmadynarnics Conf. Glowinski R. Space Shuttle hypersonic flight research and the comparison to ground test results. 92—3989. NJ: Van Nostrand Weston KC. West Palm Beach. 1989. 1995. l4th. 1977. 17th. Neuilly sur Seine. Kays WM.AIAA Pap. Ground Test. Spacecr.Wannenwetsch GD. 1966. The drag of a compressible turbulent boundary layer on a smooth plate with and without heat transfer. Spacecr. 1992.. Rockets 30:449465. Rockets 38:670—683 Neumann RD. In Hypersonics.AIAA Pap. HyperX stage separation windtunnel test program. 17th. NACA Tech. 1986. TN. The aerothermodynamic environment of the Apollo Command Module during suborbital entry. Prabhu R. AIAA Pap. 96: Thermal Design ofAeroassisted Orbiital Transfer Vehicles. Shuttle entry aerothermodynamic flight research: the orbiter experiments program. Radiative heating to the Apollo Command Module: engineering prediction and flight measurements. Merski NR. Nelson HF. Special Course on Aerothermodynamics of Hypersonic Vehicles. Ground Test. 18:117—143 Stalmach CJ. Hollis BR. Albuquerque. Hamilton HH. AGARD Report 761. 1964. Ed.. Ed. Milhoan JD.The optimum hypersonic wind tunnel. NASA CP3248 Trimmer LL. The University of Texas at Austin. 85—1003. Volume I Defining the Hypersonic Environment. TN Woods WC. 2001. Cary AM. Spacecr. Hollis BR. Rockets 38:634645 Iliff KW. X33 experimental aeroheating at Mach 6 using phosphor thermography. Experimental investigation of the surface impact pressure probe method of measuring local skin friction at supersonic speeds. Missions and requirements. 2001. AIAA Pap. NM . NASA TM Williamson WE. Developments in aerothermal test techniques at the AEDC supersonichypersonic wind tunnels.1. Rockets 38:881—819 Wright RL. 20th.Voisinet RL. 1993. VA Moeckel WE. Conf. Boston: Birkhaeuser Boston Ried RC.. Differential Equations Applied in Science and Engineering. 1958. Presented at Aerosp. Bertin JJ. . J. 1985. J. Chi SW. Holland SD. FL Wayland H. J. pp. 2001.. Note 4265 Murphy KJ. Boudreau AR 1985. Flight boundary layer transition measurements on a slender cone at Mach 20. Nutt KW. Presented at Aerosp.1 Fluid Mech. 1963. Berry SA. NASA Tech. Orbiter Experiments (OEX) Aerothermodynamics Symposium. Afterbody heat transfer measurements obtained during reentry of the spacecraft of the MercuryAtlas 5 mission. Rochelle WC. Convective Heat and Mass Transfer. Conf. Defense Research Laboratory. Weston KC.. 8 / Dynamics of a Compressible Flow Field Horvath TJ. . Throckmorton DA.466 Chap.Thompson RA. Note D6792 Matthews RK. Hypersonic flight testing. 77—0719. X33 hypersonic aerodynamic characteristics. Defining the aerothermodynamic methodology. Goodrich WD 1972. 1972. 1958. Test. Spacecr. Periaux J. Air radiation revisited. Shafer ME 1992. Zoby EV. Conf.
9. compressible flow. The widening of the streamtubes near the nose and the contraction of the streamtubes Incompressible fluid Figure 9. 467 . flowfield solutions were generated for a variety of configurations using the assumption that the density was constant throughout the flow field. 8. an error of less than 1% results when the incom pressibleflow Bernoulli equation is used to calculate the local pressure provided that the local Mach number is less than or equal to 0.1 Comparison of streamlines for an incompressible flow past an airfoil with those for a subsonic. the streamlines converge as the incompressible flow accelerates past the midsection of the airfoil. and the influence of compressibility can be neglected. if the flight speed is small compared with the speed of sound. As noted when discussing Fig. SUBSONIC FLOWS AND TRANSONIC FLOWS In Chapters 3 through 7.______ 9 COMPRESSIBLE. Thus. the changes in pressure which are generated by the vehicle motion are small relative to the freestream static pressure.6. As shown in Fig.1.5 in air.
9. Increasing the flight speed further. Such mixed subsonic/supersonic flow fields are termed transonic flows. As a result.468 Chap.1. 9 / Compressible. In addition to the absence of significant viscous forces. the compressible flow can be considered to be an irrotational potential motion in many cases. a compressible flow retains a basic similarity to an incom pressible flow. as shown in Fig. In an inviscid flow. the flow may no longer be considered as incompressible.1.The variabledensity flow requires a relatively high velocity and diverging streamlines in order to get the mass flow past the midsection of the airfoil. the components of the momentum equation may be written as U— + V— + w— = U— + v— + w— = 8x 8y au au 1 ap pax (9. Thus.2a) lap pay (9. the existence of potential motion for a compressible flow depends on the existence of a unique relation between the pressure and the density. equivalently. as the velocity increases).. if the density is a function of the pressure only. 9.1 Linearized Theory for Compressible Subsonic Flow About. the fluid elements are accelerated entirely by the action of the pressure gradient.1) In the inviscid region of the flow field (i.1 COMPRESSIBLE. the disturbance caused by the airfoil extends vertically to a greater distance. The expansion of the minimum cross section of the streamtubes forces the streamlines outward so that they conform more nearly to the curvature of the airfoil surface. outside the thin boundary layer). the direction of the pressure gradient will coincide with that of the density gradient at all points.2c) u— + v— 0x 0w 8w + äy w— az = lap paz . three—dimensiona' flow is 0(pu) + 0x ± = (9. the density decreases as the pressure decreases (or. Even though the flow is everywhere subsonic. the name given to the lowest (subsonic) freestream Mach number for which the maximum value of the local velocity first becomes sonic. we reach the critical Mach number. SUBSONIC FLOW For completely subsonic flows. the flow field contains regions of locally subsonic and locally supersonic velocities in juxtaposition. Thus.The force on each element will then be aligned with the center of gravity of the fluid element and the pressure forces will introduce no rotation. In particular. 9. Above the critical Mach number. Subsonic FJows and Transonic Flows in the regions of increased velocity lead to a progressive reduction in the curvature of the streamlines.e.2b) (9. there is a rapid attenuation of the flow disturbance with distance from the airfoil. As the flight speed is increased.a Thin Wing at Relatively Small Angles of Attack The continuity equation for steady.
Thus. v'. = (93) Combining equations (9. the magnitude and the direction of the local velocity are changed only slightly from the freestream velocity. Thus. U = Uco + U' V= w=w. . Thus.1 / Compressible. this may be written in terms of the actual pressure and density changes which result due to the fluid motion.1) through (9. 9. the resultant velocity at any point can be represented as the vector sum of the freestream velocity (a constant) together with the perturbation velocities. Is However. compressible flow past a thin airfoil at a small angle of attack.5) can be simpli fied further if one recalls that the local speed of sound can be determined using the energy equation for adiabatic flow: (9.Sec.3) and noting that the flow is irrotational.4) can be made for the case of a slender body moving in the x direction at the velocity As shown in Fig.4) becomes / \. Equation (9.4) A useful simplification of equation (9.2. Subsonic Flow 469 The speed of sound is defined as the change in pressure with respect to the change in density for an isentropic process. aw (1 ——1+—+—=o u2\öu av (9. 9.. since the flow we are studying actually is isentropic. equation (9. u'.6) U — + U' Resultant local velocity V Figure 9. Thus.5) where u/a is essentially equal to the local Mach number. Since the perturbation velocities are considered to be small in magnitude when compared with the freestream velocity. and w' . one obtains I I a2) ax v2'\av I w2\aw uvau a2 \ a2) ay \ a2) 3z ay a23z — a23x =0 (9.2 Velocity components for subsonic.
The condition of irrotationality allows us to introduce a velocity potential 4'. the binomial theorem can be used to generate the relation a = 1 + 1M2(2U + U'2 + v'2 + w12) To simplify equation (9. one obtains the linearized equation au' 0x 0V' 0w' 0z 0y (9. equation (9. As a result.11) Since the flow is everywhere isentropic.8) and neglecting the higherorder terms yields the expression (9.13) can be reduced to Laplace's equation. note that 1 a (9. irrotational flows. it is irrotational. If the "affine" transformation x' = X (9.5) can be rewritten as ax ay az 2 This equation can be rewritten in terms of the perturbation velocities as (1 — ax + ay + + y az — 2 (9. which we used to describe incompressible. 9 I Compressible. is the linearized potential equation (1 — + cb)J). which applies to a completely subsonic.12) Oz The resultant expression.13) By using a simple coordinate transformation. + =0 (9. Let be the potential function for the perturbation velocity: U' = ax V' = Oy w' = (9. as it is of second order in the perturbation velocity components.14a) . compressible flow.5). Subsonic Flows and Transonic Flows Since only small perturbations are considered.9). equation (9. the term on the righthand side often can be neglected.470 Chap.10) Ox Furthermore. which is a point function with continuous derivatives.9) Using equation (9.
is related The compressible flow over a wing of aspect ratio AR at the Mach number This is illustrated to the incompressible flow over a wing of aspect ratio AR in the sketch of Fig. we see that the transformation in effect changes the ratio of the dimension to they and z dimensions Although the spanwise dimensions are unaltered. compressible flow: (a) compressible flow. Referring to equation (9. the transformed chordwise dimension is. equation (9.) through the correlation (9. the pressure coefficient for the compressible flow is given by — — UOO& (9.14). although the airfoil section of the corresponding wings remain geometrically similar. the linearized compressible flow can be readily obtained. Although the calculation of a compressible flow field from a known:incompressible flow is relatively straightforward.1 I compressible.4. Thus. The potential distribution for an incompressible flow and the corresponding the compressible flow equation) are com"foreshortened" distribution (which pared in Fig.Sec.15) Thus. Using the linearized approximation.13) becomes y z (9. A study of the changes in a completely subsonic flow field around a given wing as the'Mach number is increased corresponds to an investigation of the incompressiblà flow around a series Of wings of progressively reduced aspect ratio.16a) which is related to the pressure coefficient for the corresponding incompressible flow (Ci. Subsonic Flow x 471 x (a) (b) Figure 9.3 Distribution of points having equal values of 4) in the linearized transformation for subsonic. 9.14b) (914c) + + =0 (9. (b) corresponding incompressible flow. 9. care must be taken in the determination of the boundary conditions satisfied by the compressible flqw field.3 at points having the same value of 4). a corresponding solution for. y' = = is introduced.16b) . the aspect ratios for the wings differ. if the potential field for incompressible flow past a given configuration is known. 9.
The effect of compressibility on the flow past an airfoil system is to increase the horizontal perturbation velocities over the airfoil surface by the factor 1/\/1 — The correlation is known as the PrandtlGlauert formula. (b) corresponding wing for incompressible flow. 9.8 1. C' The resultant variation for the liftcurve slope with Mach number is presented for a twodimensional unswept airfoil in Fig.5.5 value for an incompressible flow (C1) by a factor of 1/[(1 — ]. Based on equation (9. We can calculate the lift by integrating the pressure distribution over the airfoil surface from the leading edge to the trailing edge. .0 Mach number with Mach number using PrandtlGlauert formula. Thus.472 Chap.4 Wings for flows related by the linearized transformation: (a) wing for compressible flow. Subsonic Flows and Transonic Flows y (a) (b) Figure 9. dC. dcx 0 0.16b). 9 / Compressible.5 Variation of liftcurve slope 0.6 0. we find that the section lift coefficient for a compressible subsonic flow (C1) also exceeds the corresponding 2 0.4 0.0 02 Figure 9.
He had attained air speeds where new aerodynamic phenomena known as compressibility effects were created. The relative inaccuracy at a particular Mach number depends on parameters such as the section thickness and the angle of attack. The windtunnel correlations. 9. of course. the PrandtlGlauert formula begins to show increasing departures from reality as the Mach number approaches a value of unity. model dragrise characteristics were measured for the first time.Sec." Using a P38 model in the highspeed wind tunnel of the Ames Research Center. 9. . To illustrate the essential changes in the flow. Jones and Cohen (1960) warn that the method generally underestimates the effect of compressibility on the magnitude of disturbances for airfoils of finite thickness. as taken from Foss and Blay (1987). "Major Gilkey had become the first pilot to encounter a new. it is interesting to note that no one was anticipating the nose down pitching moment tendencies. Foss and Blay (1987) note. The airplane was out of control!" As noted by Foss and Blay. 9. Pressure data and flow visualization clearly revealed the shockstall conditions and helped confirm the flighttest findings. mildly at first and then more violently. in the absence of shock waves.6. The control column was shaking and the control forces had become heavy. 9. Attention is called to the section lift coefficient at five particular Mach numbers by the letters a through e).000 ft. Major Gilkey elected to stay with the plane and recovered at 7000 ft. That is. the drag acting on an airfoil in an inviscid. are reproduced in Fig. is true provided that there are no shock waves. "In hindsight. which were hinted at in section on pitching moment data. Major Signa Gilkey peeled off into a steep dive starting at 35. Significant differences exist between the flow fields at these five Mach numbers. and it was impossible to pull the column back to counteract the nosedown tendency.2 / Transonic Flow Past Unswept Airfoils 473 Furthermore. the quantity 1/\/1 — approaches infinity.16b). The variations in the section lift coefficient with Mach number indicate complex changes in the flow field through the transonic speed range. line drawings made from schlieren photographs are reproduced in Fig. subsonic flow is the same as that for an incompressible flow. the section drag is zero. Hence. Although the PrandtlGlauert relation provides a simple method for calculating the flow around an airfoil. compressible. the position of the resultant aerodynamic force for a compressible. the designers turned to the references of the time. This. based on the pressure correlation of equation (9. Foss and Blay (1987) described a test flight of the YP38 in September 1940. causing various perturbation parameters to approach infinity. And of course. the airplane wanted to nose down further and increase the severity of the dive.8. As the freestream Mach number approaches a value of unity. subsonic flow is the same as the drag acting on an airfoil in an inviscid. Thus. 9." In seeking ways to control the aircraft experiencing these compressibility effects.2 TRANSONIC FLOW PAST UNSWEPT AIRFOILS The section lift coefficient measurements presented in Farren (1956) as a function of Mach number are reproduced in Fig. the airplane began to shake. Even worse. uncharted highspeed flight regime.7. The data indicate that the flow is essentially unchanged up to approximately onethird the speed of sound. "As he headed down and the airspeed built up. incompressible flow. there were no data to suggest that hinge moments would become unmanageable.
Subsonic Flows and Transonic Flows Lockheed. 0.6 Highspeed drag variation derived from windtunnel tests. 9 / Compressible.02 .4 0.10 — 0.7 Section lift coefficient as a function of Mach number to illustrate the effect of compressibility.12 — no data available to show variation above M=0.8 1.14 2 I I Maximum CD assumed 0.2 0. [From Farren (1956).08 0 0 0.6 0.] Refer to Fig.0 Mach number Figure 9. P38 Variation of Drag Coefficient with Mach Number Highspeed tunnel data on 1/4 scale model Drag CD A 0.8 for the flow fields corresponding to the lettered points on this graph. .04 0. 9.06 Estimated above / this point 0. [From Foss and Blay (1987).83 0.] CI Figure 9.474 Chap.
The shock waves at the trailing edge remain. (a) When the freestream Mach number is 0.4. As a result. which is (b) At approximately twice the lowspeed value.89. (d) When the freestream Mach number is 0. (e) Mach number [From Farren (1956). 9.98. terminating in a shock wave.The flow around the airfoil is supersonic everywhere except very near the rounded nose.98.8 Flow field around an airfoil in transonic streams based on schlieren photographs: (a) Mach number (b) Mach number = 0. but they have become weaker. The data presented in Figs.81. As indicated in Figs. 9. result. (c) At = 0.81.9a. The corresponding pressure and local Mach number distributions are presented in Fig.7 through 9. Separation at the foot of the upper surface shock wave is more conspicuous and the turbulent wake is wide.. = 0. (e) When the freestream flow is supersonic. = 0. The section lift coefficient is approximately 60% greater than the lowspeed values at the same angles of attack. .89 than at Flow on the upper surface is not greatly different than that for Fig.e. the section lift coefficient reaches its maximum value. 9.2 / Transonic Flow Past Unswept Airfoils Shock wave 475 now Wake 'Flow (a) (b) (c) Flow Shock /1 wave Shock wave (d) (e) Figure 9. Because the viscous flow separates at the foot of the shock wave. The flow on the lower surface is subsonic everywhere. the flow past the upper surface decelerates from local flow velocities which are supersonic without a shock wave.8b and 9.75. The local Mach number is supersonic for most of the airfoil (both upper and lower surfaces). (c) Mach number (d) Mach number = 0. 9.89. 1. The shock wave at the trailing edge of the lower surface effectively isolates the upper surface from the lower surface.75. the shock waves both for the upper surface and for the lower surface have reached the trailing edge.81.8b. flow is supersonic over nearly the entire lower surface and deceleration to subsonic speed through a shock wave at the trailing edge. Parameters such as thickness ratio. the lift is drastically reduced. As a result. flow is supersonic over the first 70% of the surface. the pressure on the upper surface near the trailing edge is greater than that on the lower surface.] = 0.9b. As a = 0. the wake is appreciably wider than for (a). 9. the lower surface pressures are lower at = 0. the detached shock wave in front of the leading edge) is generated.9 illustrate the effect of Mach number for a given airfoil section at a particular angle of attack.Sec. a bow shock wave (i.
Subsonic Flows and Transonic Flows p Prl C. [From Farren (1956).0 0.89.6 1.0 xlc (a) 2.0 1. = 0.5 x/c (b) Figure 9. therefore.0 0.8 1. Transonic flows are very sensitive to the contour of the body surface.4 0. (b) flow at the lower surface trailing edge is supersonic.6 0.8 0. .81.) 0 1. = 0.0 0.9 Pressure distribution and local Mach number distrib ution for transonic flows around an airfoil: (a) Flow at the trailing edge is subsonic.476 Chap.2 p Prl 1. 9 / Compressible. and nose radius also influence the magnitude of the compressibility effects.] camber. the inviscid flow field as well as the downstream boundarylayer.4 1.5 1.) C. since changes in the surface slope affect the location of the shock wave and.
Sec.10. the shockwave/boundarylayer interaction and the possible development of separation downstream of the shock wave are sensitive to the character of the boundarylayer.10 Constantdensity contours for a NACA 0012 airfoil in a Mach 0. since the shockwave location as well as the extent of flow separation can become a function of the artificial tripping.8 stream at 3°.0 X are presented in Fig. it is important to consider the Reynolds number when simulating a flow in the wind tunnel. The turbulence model of Spalart and Allmaras (1992) was used to represent the viscous boundarylayer near the surface of the airfoil. 9.2 I Transonic Flow Past Unswept Airfoils 477 Furthermore. The Mach 1 contours for the upper surface and for the lower surface are highlighted in white. Contours of constant Mach number. as was discussed in Chapter 8.] . et a!. to its thickness. Since a turbulent boundarylayer can negotiate higher adverse pressure gradients than can a laminar one. A large Reynolds number difference between the desired flow and its simulation may produce significant differences in shockwave location and the resultant flow field. (1999)] has been used to compute the flow field around a NACA 0012 airfoil section at an angle of attack of 3°. Thus. as computed by Forsythe and Blake (2000) for a freestream Mach number of 0.8 at a freestream Reynolds number (based on the chord length) of 3. The Cobalt60 code [Strang. The use of artificial trips to force transition to occur at a specified point (as discussed in Chapter 5) may be unsatisfactory for transonic flows. Note that the relatively large area of supersonic flow on the upper surface is terminated by a shock Figure 9. 9. the shockwave/boundarylayer interaction is smaller for a turbulent boundarylayer. and to its velocity profile at the interaction location. [As provided by Forsythe and Blake (2000).
wave at the downstream (or righthand) side of the region of supersonic flow.75 1 x/c Figure 9.0 X 106. and an angle of attack of 2. include a freestream Mach number of 0. and that for the measurements presented by Farren (1956).5 0. the local temperature of the expanding flow can decrease below the dew point. If the airplane is flying through an airstream where the relative humidity is high. et al. During the isentropic expansion of the flow. Note the similarity between the Machnumber dependence of the computed section lift coefficient.10. Water . which are taken from Forsythe. Since the pressure is decreasing. 9. 9 / Compressible. This is evident in the (darkened region) Mach number contours presented in Fig. 9. 9. as presented in Fig. (2000) and reproduced in Fig. The computed flow field for transonic flow over a NACA 0012 airfoil that was presented in Fig.11.11 Pressure distributions for a NACA 0012 in a Mach 0.478 Chap. 9.5 cp Experimental values as taken from Harris (1981) CFD (as provided by Forsythe and Blake (2000) 0. Subsonic Flows and Transonic Flows —1 —0. which are taken from Harris (1981). The sudden increase in pressure due to the shock wave occurs on the upper surface at midchord. 9.The adverse pressure gradient associated with the shockwave/boundarylayer interaction causes the boundary layer to separate from the aftend surface of the airfoil. The shock wave is evident in the sudden increase in pressure evident both in the experimental and in the computed pressure distributions for this airfoil.86°.12. 9. which are reproduced in Fig.799. The section lift coefficient for this twodimensional flow [Forsythe and Blake (2000)] is presented as a function of the freestream Mach number in Fig. a Reynolds number of 9.8 stream at a = 3°. The flow conditions for the experimental data.7. the acceleration of the flow is often termed an expansion.10 depicts regions (both above and below the airfoil) where the flow has accelerated to the point where the flow is locally supersonic.12. there is a corresponding decrease in the temperature.
Prandtl Glauert 0 CFD C1 0.13a. 9. as shown for the transonic flight of an Fill in Fig.5 I I I I 0. The schlieren technique.These data from [Stahl and Mackrodt (1965)] were obtained at Reynolds numbers between 1.3 0. such as that for the F151.l5. Furthermore. Further downstream. The liftcurve slope is seen to be a function of the freestream Mach number. as the flow decelerates (perhaps as it passes throtigh a shock wave that terminates the locally supersonic flow).6 479 0. . 9. the condensed particles evaporate (or vaporize) and are no longer visible.14 as a function of angle of attack.4 — — — — . 9.1 0 0.5 1 1. Under certain circumstances.75 and whose airfoil section is a NACA 65A005 (a symmetric profile for which t = 0.12 Section lift coefficient for a NACA 0012 airfoil as a function of the freesiream Mach number. which is presented in Fig. such as depicted on the cover and in Fig. the pressure and the temperature increase. As the local temperature rises above the dew point.Sec.0 X 106 and 1. 9. Thus. which allows someOne to see shock waves in a windtunnel test.5 2 Mach Figure 9.2 / Transonic Flow Past Unswept Airfoils 0. 9. the relatively large accelerations (expansions) and decelerations (compressions) yield a condensation pattern.8.05c) are presented in Fig.8 X 106.i3b. The corresponding drag polars are presented in Fig. The experimentally determined lift coefficients for an untwisted rectangular wing whose aspect ratio is 2. [As provided by Forsythe and Blake (2000).2 0. these density gradients allow us to "see" the shock waves for an airplane in flight. when an airplane is flying at transonic speeds in a relatively humid atmosphere.] vapor will condense and become visible. 9. is based on density gradients in the flow field that bend light rays passing through the flow. the linear relatiOn between the lift coefficient and the angle of attack remains valid to higher angles of attack for supersonic flows.
.
0. AR = 2.4 A 0.14 Effect of Mach number on the liftcoefficient/angleofattack correlation for a rectangular wing.4 CD 0.2 / Transonic Flow Past Unswept Airfoils Subsonic o o 481 Supersonic A = 0.0 I Figure 9.2 0.2  • •• CL 0.78 I •M= 1. = = 0...] . degrees Figure 9. AR 2.0 I —0.4 — 0.25 • S A 1.4  a Oo 00• a • 0.94 1.Sec.75. [Data from Stahl and Máckrodt (1965).6 [Data from Stahl and Mack rodt (1965).15 Effect of the Mach number on drag polars for a rectangular wing.2  A A A CL 0.50 2.= 2.39 1.50 A M.0 0 8 16 24 a.75.] Subsonic ci Supersonic = 039 Mor.6 — 0. 9.25 = 1.8 0 0 0 0. = 0.78 • Me.
Airfoil section designs which alleviate or delay the onset of the drag rise and buffeting can contribute to higher maximum speeds (transport applications) or better lift performance (fighter applications). unless the flow is locally supersonic at the trailing edge). 9. Whitcomb and his coworkers have developed a "supercritical" airfoil shape which delays the subsonic drag rise.T. including reserves. Subsonic Flows and Transonic Flows Bushnell (2004) noted: "shock waves are usually detrimental. and vortex dragduetolift. re4uiring mitiga The volume and liftengendered drag associated with shock waves is additive to the usual friction. The . 9 / Compressible. comprising onethird of total aircraft drag in the supersonic cruise case. therefore. Shock wave drag is also a major reason there are no economically viable supersonic transports yet extant.3 WAVE DRAG REDUCTION BY DESIGN 9.3.1 Airfoil Contour Wave Drag Approaches Three features which contribute to the location of the shock wave and. The conditions for divergence of the trailingedge pressure correspond to those for a rapid drop in the lift coefficient and to the onset of certain unsteady flow phenomena. Usual SST designs require a fuel fraction approaching 60%. Due to wave drag. The flow in the supersonic region ahead of the shock wave 2.2 Supercritical Airfoil Sections The Machnumber/liftcoefficient flight envelopes of modern jet aircraft operating at transonic speeds are limited by the compressibility drag rise and by the buffeting phenomenon. such as buffeting. The pressure rise across the shock wave itself (which involves considerations of the interaction with the boundary layer) 3. to the pressure distribution [Holder (1964)1 are 1." 9. even for the welldesigned configuration. A 1% overall drag reduction for these designs translates to a 5 to 10% increase in payload.3. the aerodyriamic efficiency of a supersonic transport (SST) is the order of onehalf or less of a conventional (subsonic) transport. Using intuitive reasoning and substantiating experiment Whitcomb and Clark (1965) noted that R. the lower surface flow must adjust itself to produce a similar change in pressure (since the pressure near the trailing edge must be approximately equal for the two surfaces. The reason for the decadeslong Mach .482 Chap. The subsonic flow downstream of the shock wave (which involves considerations of the boundarylayer development between the shockwave and the trailing edge and of the flow in the near wake) If the trailingedge pressure changes as a result of flow separation from the upper surface.8ish cruisespeed plateau associated with conventional longhaul transport aircraft (requiring over ten hours of flight time transpacific) is avoidance of the major drag increases due to strong shock formation on the upper wing surfaces at higher speeds.
M = 0. the induced velocities) in the midchord regions both for the upper and the lower surfaces. particularly on the lower surface.72.Sec. If not. 9. 9. as indicated by the pressure coefficients above the sonic value and by the shaded areas of the flow fields in Fig. M [From Ayers (1972).] 0. The increased lift generated by the concave surface near the trailing edge of the lower surface is evident in the experimental pressure distributions presented in Fig. separation of the lowersurface boundarylayer would occur. which results in increased drag. (b) supercritical airfoil. additional lift must be generated from the region of the airfoil behind the shock wave. principal differences between the transonic flow field for a conventional airfoil and that for a supercritical airfoil are illustrated by the data presented in Fig. especially of the lower surface. 9.3 / Wave Drag Reduction by Design Supersonic flow 483 Strong shock wave Separated boundary Weak shock wave CI.80. At supercritical Mach numbers. .16. when the pressure rise associated with a shock wave is superimposed on the pressure rise caused by the cusp.16. therefore. 9. The much flatter shape of the upper surface of the supercritical airfoil causes the shock wave to occur downstream and therefore reduces its strength. However.17. a broad region of locally supersonic flow extends vertically from both airfoils. To compensate for the loss in lift from the midchord region. the pressure rise across the shock wave is reduced with a corresponding reduction in drag. The region of locally supersonic flow usually terminates in a shock wave. the diminished curvature of the upper surface also results in a reduction of the lift carried by the midchord region of the airfoil section. \\ (b) IC + (a) + Figure 916 Comparison of transonic flow over a NACA 64A series airfoil with that over a "supercritical" airfoil section: (a) NACA 64A series. CI. To minimize the surface curvatures (and. the leadingedge radius is relatively large.Thus. The midchord region of the lower surface should be designed to maintain subcritical flow over the range of operating conditions. The increase of lift in this area is achieved by substantial positive camber and incidence of the aft region of the airfoil.
= 3.0' 0. was the first person to draw attention to the significant reduction in transonic drag which comes when the wing is swept back enough to avoid the formation of shock waves which occur when the flow over the wing is locally supersonic. 0  a.6 0. This simple principle is obviously only true (if at all) on an infinite span wing of constant section. Betz. Adolf Buse mann and Albert Betz. = 0.6 0.17 Experimentally determined pressure distribution for a supercritical airfoil section. so the critical conditions are reached only when the component of the freestream velocity normal to the leading edge has been locally accelerated at some point on the wing to the local sonic speed. the initial suggestion of Betz led to windtunnel tests which substantiated the essence of the theory.0 x/c Figure 9. (1975).4— c(  00 o000o 0 0  —0. Busemann showed that sweepback would reduce drag at supersonic speeds. in his paper on highspeed flight. Nevertheless.2 0' 0' a.8 1.0 I 0.4 SWEPT WINGS AT TRANSONIC SPEEDS In the late 1930s. The basic principle is that the component of the main flow parallel to the wing leading edge is not perturbed by the wing. in 1939.8 0. The . [The interested reader is referred to Miller and Sawers (1970)].] 9.484 Chap. Subsonic Flows and Transonic Flows 0000 000 0 I I  —0.80. 9 / Compressible. discovered that drag at transonic and supersonic speeds could be reduced by sweeping back the wings. 0.0 X 106 (flagged symbols are for the lower surface). C1 = 0.0.2 0. two aerodynamicists who had been taught by Prandtl. At the Volta conference in Rome in 1935. [Data from Hurley et al.4 0.6—0.54.4 0.2 0.
4 0.4 I Swept Wings at Transonic Speeds 485 CL —0. lowaspectratio configuration. 9. (In addition to the data presented in Figs. In the highly swept.0 0.19. One can use a variablegeometry (or swingwing) design to obtain a suitable combination of lowspeed and highspeed characteristics. the liftcurve slope (dCL/da) is greater for the rectangular wing for all the freestream Mach numbers considered.] results of the windtunnel measurements performed at Gottingen in 1939 by H.2 CD 0.18 Comparison of the transonic drag polar for an unswept wing with that for a swept wing. At subsonic speeds.0 0.3 0.) Even in subsonic streams.Sec.3 Figure 9.31 and whose airfoil section is NACA 65A005 (a symmetric profile for which t = 0. the lift is a linear function of the angle of attack up to relatively high inclinations.1 CD 0. over the angleofattack range for which the lift coefficient is a linear function of the angle of attack.19. These data from Stahl and Mackrodt (1965) were obtained at Reynolds numbers between 1. The experimentallydetermined lift coefficients and drag polars for a delta wing whose aspect ratio is 2. 9.0 x 106 and 1. the reader is referred to Chapter 7. . At the opposite end of the sweep range. 9.05c) are presented in Fig.2 0. the variablegeometry wing provides low wave drag and eliminates the need for wing fold in the case of naval aircraft. 9. These data show that the effect of the shock waves that occur on the wing at high subsonic speeds are delayed to higher Mach numbers by sweepback.18.8 X 106. [Data from Schlichting (1960). Ludwieg [as presented in Schlichting (1960)] are reproduced in Fig.14 and 9.1 0. However. the liftcoefficient/angleofattack correlation for the delta wing is markedly different than that for a rectangular wing.
for a straight fuselage.94 CL 0. [Data from Stahl and Mackrodt (1965).] one obtains efficient subsonic cruise and loiter and good maneuverability at the lower speeds.attack correlation. As the Mach number is increased. is presented in Fig.These distortions at the root and at the tip determine the flow pattern. sweptback wing varies across the span such that the maximum local velocity is reached much farther aft at the root and farther forward at the tip compared to the basic infinitewing distribution [Rogers and Hall (1960)]. Therefore.1 Wing—Body Interactions and the "Area Rule" The chordwise pressure distribution on a plane. finitespan.4. Subsonic Flows and Transonic Flows Supersonic = 1. However.39 M.25 a = 0. 9 I Compressible. 9.AR = 2. a shock wave first occurs near the tip. a loss of internal fuel capacity. the Rockwell International Bi. Because of the variation of the component of velocity normal to the wing leading edge. ALE 600: 0 8 16 24 a.50 D = 2. A variablegeometry design.78 Chap.8 Figure 919 Effect of Mach number on the aerodynamic characteristics of a delta wing. the flow at the wing root is constrained to follow the fuselage. On a swept wing with a squarecut tip. a set of compression waves originate in the wingroot .31.20. a rear shock is developed which dominates the flow. This "initial tip shock" is relatively short in extent. = 0. Negative factors in a swingwing design are complexity. degrees (a) (a) liftcoefficient/angleof. streamlines over the wing surface will tend to be curved. 9..486 Subsonic = 0. and the considerable weight of the hinge/pivot structure.
the shockwave formations about relatively complex sweptwing/body combinations at zero lift near the speed of sound are similar to those which occur for a body of revolution with the same axial development of crosssectional area normal to the airstream. Kuchemann (1957> recognized that a properly shaped waistlike indentation on the fuselage could be designed to produce a velocity distribution on the wing near the root similar to that on the corresponding infinitespan wing. As shown in Fig.50 o o = 0.39 = 0.2 I 0 o 0. the compression waves propagate acrOss the wing span and ultimately may coalesce near the tip to form the reat shock. 9.4 0. the wing behaving in some respects like a thinner wing. the zerolift drag ." Whit comb (1954) found that the zerolift drag rise is due primarily to shock waves.6 (b) Figure (continued) (b) drag polar.8  a h. KUchemann noted that "the critical Mach number can be raised and the drag reduced in the transonic and the supersonic type of flow.Whitcomb concluded that "near the speed of sound.94 F 1.Sec. region to turn the flow parallel to the fuselage. 94 / Swept Wings at Transonic Speeds Subsonic 487 Supersonic = 1. it is clear that the interaction between central fuselage and the swept wing has a significant effect on the transonic flow over the wing. Thus. Furthermore.21.0 0. In the early 1950s.2 CD 0. CL • y I I 0.
.
9. would result in a large reduction in the transonic drag rise.S. the Air Force's first supersonic bomber.Sec. Freeflight. so that the combination has nearly the same axial distribution of crosssectional area as the original body alone.4 I Swept Wings at Transonic Speeds 489 Approximate streamlines Fuselage Compression waves I Initial position of tip shock Rear shock Figure 9. which is taken from Shortal (1978). the longitudinal distribution of the crosssectional area. A major redesign effort included the first application of Whitcomb's area rule to reduce the transonic drag. Air Force with its first operational deltawing design. designated MX1626." which is often known as area rule for trarisonic configurations. rocketpowered models of the original Convair design.22. the reader is referred to Taylor (1969). 9.23. are a sketch of the planform of the MX1626. The peak drag coefficient at Mach 1. Whitcomb noted that it would be expected that indenting the body of a wingbody combination. 9. are illustrated in Fig. [From Rogers and Hall (1960). The relation between the distribution of the crosssectional area and the transonic drag rise had a significant effect on the design of the Convair B58. The modified YF102A achieved supersonic flight satisfactorily. and the drag coefficient for the .02 was almost twice as high as that predicted and the model did not achieve supersonic speeds. For additional information. providing the U.] fuselage and the wing) with axial position (as well as a reduction in the relative magnitude of the maximum area). An evaluation of the longitudinal crosssectional area distribution in accordance with the area rule of R. Design applications of this "theory. Early flight tests revealed that the prototypes of the YF102 had serious deficiencies in performance. Presented in Fig.21 Formation of rear shock from compression waves associated with flow near the root. were tested at the Wallops Island Flight Test Range (NAcA). T. Whitcomb along with data obtained in the Helium Gun Facility at Wallops on a body of revolution having the sam e longitudinal distribution of the crosssectional area as the MX1626 provided an explanation of the unexpectedly high drag.
9 / Compressible. [From Goodmanson and Gratzer (1973). (b) sketch of the Convair F 102 A. 0 0 Vehicle station (a) (b) Figure 9. U.] .22 Application of the area rule to the axial distribution of the crosssectional arca: (a) area rule applied to the design of a nearsonic transport. U.490 Chap. Subsonic Flows and Transonic Flows 0 U U.
9 1.Sec.23.23 Planform sketches. included a 3% thick. 9.01 0 0. Furthermore. and the nacelles all add area to the fuselage and the weapons/fuel pod. modified by a 10degree forward swept trailing edge. the longitudinal crosssectional area distribution. [Taken from Shortal (1978).2 0. 60degree delta wing.23c.9 1. 0. Note that the configuration designated the redesigned MX1964 . The resultant planform for the PARD AreaRule Design. To this wing.01 0 — 0. designated PARD AreaRule Design in Fig. 9. longitudinal distributions of crosssectional areas. 9. Convair evaluated four alternative designs. and J. staggered chordwise. Based on these data.] equivalent axisymmetric configuration as a function of the Mach number. Hopko. the Air Force changed the Project Number from MX1626 to MX1964.03 0. The diamondshaped wing provides a more gradual variation in the crosssectional area distribution.0 1. and the measured drag coefficient for the equivalent body of revolution as a function of the Mach number are reproduced in Fig.0 1.1 1. To reflect changes in the design requirements due to the unavailability of the originally planned large jet engines and to changes in the function of the pod.1 1.23b. Hall set out to design a vehiële which met the general requirements of the new MX1964 yet reduced the transonic drag rise. they added four separate nacelles.8 0. N. creating a large total crosssectional area. R. Data for the configuration which had the lowest drag rise of the four Convair configurations are presented in Fig. 9. the placement of these components results in large longitudinal variations in the total crosssectional area.2 Mach number M M Figure 9.02 C0 0. R.01 0 0. the wheel fairings. Piland.03 0.2 0.9 1.03 0.1 1.0 1. and drag coefficients for equivalent bodies of revolution for supersonic bomber models.4 I Swept Wings at Transonic Speeds (a) Original MX1626 491 (c) Redesigned MX1964 (b) PARD area rule design Total 12 X10' Wheel 12 8 12 Total — XlO3es Wing 8 4— 0 o 2 Fuselage & pod I 4 I I ç 10x102 U 0 0 2 4 6 8 4 I 6 8 10 12X102 0 2 4 6 8 10 12X102 I I 0.02 — 0. R.8 0. and made the fuselage a body of revolution. Note how the wing. The design.02 0.
The benefit due to area ruling.27) illustrates these applications of the area rule. as shown in Fig. (From the collection of TIm Valdez. 9.24 Photograph of B58. .5 to 1. While touring the Air Force Museum at the WrightPatterson Air Force Base.95 region. 9.25 through 9. Subsonic Flows and Transonic Flows Figure 9. A slight instability which remained in the windtunnel data and was thought to be Reynolds number dependent was never found during the flight tests. As a result. there is a corresponding increase in transonic acceleration. Continued development tests in the transonic wind tunnel proved that an area ruling concept (as shown in Fig.24. We previously had siamese nacelles (engines sidebyside as in a B52 nacelle) which was lighter in weight but the higher drag more than offset the lower weight.000 ft at speeds from Mach 0. "The arearule requirements of the B58 required that the inboard and outboard nacelles be staggered —the inlet face of the outboard nacelle began on the had same plane as the exhaust face of the inboard nacelle. Area ruling was also applied to the tiptank design for the F5A/B aircraft. since [According to Stuart (1978) and Bradley (1982)] the major and decisive portions of airtoair combat take place at altitudes below 30. considering a straightsided fuselage as a base.25. Hillaker (1994) noted.27. 9 / Compressible. 9. which is very important to this fighter's design.) both a smooth progression of crosssectional area and the minimum value of the maximum crosssectional area.90 to 0.0. Initial prototype flight and windtunnel tests indicated that transonic buffet and longitudinal instabilities existed with the originally proposed wingtip tank. H.492 Chap. The photograph of an F5 (Fig.26) could be used to essentially eliminate the pitch instabilities in the Mach 0." This is illustrated in the photograph of Fig. 9. A cruise drag benefit also existed because of the improvement in wingtip airflow characteristics. is a large decrease in drag at transonic speeds. 9. Applications of the area rule in the design of the FS are illustrated in Figs.
As might be expected. approximately 0. the configuration could be trimmed with zero control deflection near the lift coefficient required for the maximum lifttodrag ratio at a Mach number of 0.6 1.4 / Swept Wings at Transonic Speeds I I I I 493 I Drag I I I I 0. Furthermore.2 1.4 0. This result is due. Figure 925 Reduction in the transonic drag rise due to the application of area rule to the FS fuselage.95.28. in part. to the high sweep angles for the wing and for the tail. this shape is conducive to low transonic drag and the dragrise Mach number is relatively high (i.95). 9. 9. 9. [From Stuart (1978).8 2.0023 Cm() Figure 9.26 Application of area rule to the design of the tip tank for the F5A/B aircraft.] By carefully designing the wing and the tail configurations.Sec.. [From Stuart (1978)] .e.6 0.4 1. CL USE = —0.2 0. As shown in Fig.8 1. the crosssection area distribution is relatively smooth and symmetric about the midbody station.0 1. one can obtain a smooth variation of the crosssection area even without adjusting the fuselage cross section.0 Mach no. Such is the case for the sweptwing cruise missile which is presented in Fig.29.
[From Spearman and Collins (1972). new aircraft are designed to cruise at speeds near Mach 0. (From the author's collection.28 Details of a model of a sweptwing cruise missile.8.2 SecondOrder AreaRute Considerations Optimization of the wing/body/nacelle configuration plays a critical role in the design of business jets.88 (822) 5.) 9. Unlike previous business jets developed for cruising speeds near a Mach number of 0. Added to the requirement of more speed is the desire for significant increases in range. Linear dimensions are in centimeters (inches).27 Illustration of the application of the area rule to the tip tanks of the F5.] .7.4.494 Chap. Subsonk Flows and Transonic Flows Figure 9.03 50 Moment reference Figure 9. 9 / Compressible. 20.
94 / Swept Wings at Transonic Speeds 495 xli 0.04 0 12 8 —— \% (LID)max 4 0 0.4 0. The area rule is essentially a lineartheory concept for zero lift.4 2.2 Cross sectional o..29 Area distribution and various aerodynamic parameters for a sweptwing cruise missile.2 Figure 9. As noted by Galiman.2 L6 20 2. Whitcomb (1976) noted that "to achieve the most satisfactory drag characteristics at lifting conditions the fuselage shape had to. ." Galiman. Full potential analysis of the wingbodynacelle configuration provided an accurate assessment of the influence of the fuselage mounted engines on wing pressures.8 1. "These requirements create an aerodynamic design problem that benefits from modern computational fluid dynamics (CFD) and aerodynamic shape optimization. (1996) used codes similar to those discussed (briefly) in Chapter 14. et at (1996). these highspeed wings must be compatible with typical business jet configurations that have engines mounted on the fuselage near the wing trailing edge. be modified from that defined by the simple application of the area rule as previously described to account for the nonlinearity . the designs employ wings with more sweep and carefuliy designed cross sections that minimize the onset of the compressibility drag.Sec.o 0. [From Spearman and Collins (1972). .8 3.08 CD.08 CL (1___ I) 0.i area 0 0. Fur thermore.] To meet these requirements. et al.
as taken from Whitcomb (1976) is illustrated in Fig.30 Secondorder area rule considerations. To compensate for this effect the fuselage indentation required to eliminate the farfield effects of the wing must be increased. For lifting conditions at near sonic speeds there is a substantial local region of supercritical flow above the wing surface which results in local expansions of the streamtube areas. this expansion is equivalent to an increase in the physical thickness of the wing. "In transonic flow. dissipation of disturbances occurs in the subsonic regions and the streamtube areas are no longer invariant. [From Whitcomb (1976). It is based on the idea of perfect pressure disturbance communication between the wing. In the basic considerations of the area rule concept. Carisen (1995) noted that the communication between the flow on the wing and the flow on the fuselage changes in the transonic regime because of the mixed supersonic and subsonic flow. 9. Carlsen (1995) noted: "That the sonic area rule is a far field method of predicting and understanding the wave drag due to shock losses [Whitcomb (1956)]. It proposes that a body of revolution with the same axial development of crosssectional area will have a wave drag that is similar to the original configuration [Whitcomb (1956)]. Only the volume that will . or other external features. 9 / Compressible. the fuselage crosssectional area needed for storing landing gear and aircraft subsystems and for accommodating passenger seating and headroom. 02 0 C.) Vehicle station Figure 9. As a result of this dissipation. Subsonic Flows and Transonic Flows 0 0 'U 0. conflicts with the arearule requirements in some cases. "The fuselage indentation based on this delay in the corrected crosssectional area distribution resulted in a significant drag rise Mach number compared with that for the indentation based on the zero lift distribution. Part of the pressure changes created by the indentations in the fuselage is dissipated before it reaches the wing by passing through embedded subsonic regions. To account for the majority of this dissipation." However. and the fuselage.496 Chap.30." The additional correction to the crosssectional areas required for a transonic transport. the new transonic method applies a weighting function to the sonic area rule that modifies the volume removed from the fuselage. it is erroneous to subtract the total wing volume from the fuselage.] of the flow at such conditions. as noted by Ayers (1972).
while in the latter case they tend to load the structure further. The procedure is repeated at. recent advances in the development of composite materials for aircraft structures have permitted the design and fabrication of aeroelastically tailored wings which substantially eliminate the divergence problem. the use of FSW in aircraft designs was. The dissipating effect of the subsonic regions in the transonic regime significantly influences the approach to area ruling. with minor exceptions. the FSW constructed of conventional materials has a much lower speed for structural divergence which occurs when the elastic restoring forces can no longer overcome the aerodynamic forces causing the deformation. decreased lift curve slope. along with some of its desirable characteristics. until recently. the tips tend to unload. The twisting behavior occurs because the axis of section centers of pressure along the span of an ASW is behind the structural axis of the wing in typical designs. i. Dragrise delays that match the traditional sonicarearule can be obtained by modifying the aircraft to only 60% of what traditional sonicarearule prescribes. a series of locations along the fuselage axis.4." Carisen (1995) recommended the use of weighting functions that would be applied "during the integration of the area of the wing that is intersected by a given Mach 1 plane. An ASW twists down at the tips as the lift increases while a FSW twists up. (1983)]. The reasons for this will be discussed later.. A swept wing constructed of conventional materials twists under load. while just the opposite is true for the FSW [Uhuad. 9.4 / Swept Wings at Transonic Speeds 497 relieve the flow on the wing should be removed.Sec. the communication between the flow over the fuselage and that over its external parts that accounts for the dissipation due to embedded subsonic regions. has been excluded from the aircraft designer's stable of design options. but.e.The structural axis can be thought of as the locus of points along the wing where a load produces only wing bending with no twist. However." Carisen (1995) concludes that the proposed use of a weighting function adjusts for the effects of mixed flows. Thus. Thus.) resulting from wing sweepback are also present for forward swept wings (FSW) as well. they are constructed of conven tional aircraft materials such as metals. To eliminate this problem using conventional materials requires a prohibitive structural weight penalty.g. the FSW. as an example. Historically. 9. increased critical Mach number. However. et al. which will be discussed in this section.. resulting in a net volume removal from the fuselage. In the former case. there are some comparative advantages to a FSW in relation to an aft swept wing (ASW). etc. Diederich and Budiansky (1948) show that uniform FSWs with sweep angles from —20 to —60° exhibit decreases as high as 78% in aeroelastic divergence dynamic pressures in comparison to straight wings. The weighted wing area is then subtracted from the fuselage area that is intersected by this same Mach plane. Area at the wingtip is given less value than the area near the wing root. precluded by the relatively low (in comparison with ASW) Mach numbers at which they experience aeroelastic flutter and divergence when. This minimizes the volume removed from the fuselage and still maintains the drag rise delay. .3 Forward Swept Wing The aerodynamic effects (e.
As noted by Whitford (1987). an FSW can be designed with a higher aspect ratio than a comparable ASW and produce the same bending moment at the wing root. This. This permits the designer to capitalize on some of the advantages of FSWs compared toASWs without paying an exhorbitant weight penalty. By careful design and fabrication. Typical shock locations for a supercritical wing occur in the vicinity of the 70% chord line along the span. for designs with the same shock sweep angle. Excellent overall discussions of these FSW characteristics are given in Whitford (1987) and in Krone (1980). wing area. careful attention to the aircraft mission will determine the design solution to be adopted in each case.The higher aspect ratio in turn produces lower induced drag. Modem aircraft designs rely on wing sweep and supercritical airfoil sections to delay adverse compressibility effects and reduce their severity when they do occur. Conversely. And although the lower leadingedge sweep will result in higher wave drag at supersonic conditions. We restrict our discussion here to aerodynamic advantages of the FSW. drag due to shockinduced separation has been found to be less. Good design practices force the shock on the upper surface of the wing to occur as close to the trailing edge as possible and to be as weak as possible.498 Chap. aileron effectiveness is preserved up to higher anglesofattack with a consequent improvement in maneuverability. For the FSW the spanwise flow component is toward the wing root leading to root stall first. of course. However. coupled with the more highly loaded tips of the ASW. Flow over an ASW has a component along the span and toward the tips which results in a thicker boundary layer in the tip region than would otherwise exist. This phenomenon can resuit in loss ofaileron control effectiveness at high anglesofattack with a consequent loss of maneuverability. Whitford (1987) also points out that an FSW with the same planform area. Thus. can lead to tip stall. Subsonic Flows and Transonic Flows The solution to this problem for the FSW is to design an aeroelastically tailored wing using composite materials arranged in unidirectional layers or plies which lie at selected an gles relative to one another. this results in a more highly swept shock for an FSW compared to an ASW with the same leadingedge sweep angle. aeroelastic tailoring is not limited to FSWs. and span as an ASW has its aerodynamic center closer to the wing root and hence experiences a lower wing root bending moment. shock sweep. Thus. Key aerodynamic advantages of the FSWfall into two categories: reduced drag and enhanced maneuverability at transonic Mach numbers and high anglesofattack. . results in a lower wave drag penalty. The relative advantages of the FSW over the ASW are. but can be applied to straight and aft swept wings as well [Brown et al. the wing can be made to deform under loading so that the divergence problem just described can be delayed to a much higher Mach number. and taper ratio. with the stall region moving inward toward the root as the stall progresses. as noted earlier. a function of the aircraft mission and fall into aerodynamic and other categories. the FSW requires less leadingedge sweep. 9 / Compressib'e. which resUlts in a higher lift curve slope than the comparable ASW and a reduction in induced drag at subsonic conditions. of course. The more highly swept shock. Further. An additional and undesirable offshoot of thisphenomenon is an overall pitchup tendency since the loss of lift typically occurs aft of the aircraft center of gravity. the FSW with root stall still suffers the pitchup tendency since the loss of lift is once again typically aft of the aircraft center of gravity. (1980)]. Incidentally.
9. These include a threesurface longitudinal control design. and deflectable aftfuselage strake flaps. and the Air Force have funded the design and fabrication of the X29. Croom et al. (1988).Sec. symmetric deflection of wing flaperons. Roll control is provided by differential flaperon deflection (or.35 static margin).4 / Swept Wings at Transonic Speeds Flaperon Strake Nose strake 499 strake Figure 9. To investigate further and to validate by flight test these and other potential advantages of the FSW relative to the ASW. including FSW and closecoupled canards. the allmovable canard and strake flap control surfaces provide significant control authority throughout the trim angleofattack range Moreover. produced by Grumman. NASA.31. equivalently. The X29 configuration. the basic airframe is unflyable without stability augmentation. Flow separation progresses from . However." A major benefit arising from the use of forwardswept wings at high anglesofattack is the favorable progression of the stall pattern. is an experimental flighttest vehicle designed to explore a variety of advanced technologies. (1988) studied the lowspeed. consisting of an allmovable closecoupled canard with large deflection capability. similar to the F5 design. 9. A conventional rudder provides directional control. the Defense Advanced Research Projects Agency (DARPA). "The most dominant characteristic of the configuration is the extreme level of inherent static pitch instability (—0. has several distinctive design features. which produces strong vortical flow in the stall/poststall region. which is very desirable for fighter aircraft. the large control moments in conjunction with the relatively low pitch inertia. As noted by Croom et al. Clearly. highangleofattack flight dynamics of the X29 configuration.31 Threeview sketch of fullscale X29 configuration. This aircraft. aileron deflection). Using data obtained in freeflight windtunnel tests at the Langley Research Center (NASA). associated with the relatively lightweight design indicate the potential for a high level of pitch agility. The X29 has a long slender forebody. The primary objective of the X29 program is to give designers confidence in the FSW approach and to validate the advantages of FSW and other technologies for use in future aircraft designs [Putnam (1984)]. illustrated in the three view sketches of Fig.
In the transonic Mach number range. at poststall angles of attack. the directional stability degrades significantly.17) Comparing equation (9." 9.5 TRANSON1C AiRCRAFT Ayers (1972) also notes that "windtunnel studies have indicated that combining the supercritical airfoil.0. However.e. Past studies have shown that under sideslip conditions..10). computation volume (or cell). for the particular assumptions made during the development of equation (9." A finitedifference equation is derived by applying the divergence theorem to the integral of equation (918) over an elemental. significant roll control moments persist to very high angles of attack and lateral control should be comparatively good throughout the trim angleof—attack range. and wing sweep can push the cruising speed of subsonic aircraft very near Mach 1. rectangular. which requires that the pressure.18) The parameter n reflects the nonuniqueness of the equation and. as the angle of attack is increased above 25deg. which is proportional to be continuous at the trailing edge. Bailey and Ballhaus (1975) used a relaxation procedure to solve the transonic smalldisturbance equation: (1 — + + = 1 (9. numerous investigators have solved the nonlinear equation for the disturbance velocity potential in transonic flow. which is equal to the difference in . angle of attack. Subsonic Flows and Transonic Flows the wingroot region outboard. the area rule. this term cannot be neglected as it was at lower speeds. quoting Bailey and Ballhaus (1975). "can be adjusted to better approximate the exact sonic pressure coefficient.. the flat elliptical crosssection forebody generates an asymmetric vortex system that produces a suction force on the windward side of the forebody resulting in yawing moments into the sideslip (i. one would find that. However. The boundary conditions for the wing include the Kutta condition. see equation (9.. = (9.10). The reduction in stability is due to the vertical tail becoming immersed in the low energy wake of the wing. resulting in the retention of aileron control power beyond the stall angleofattack.500 Chap.Thus. which can be written [refer to Newman and Allison (1971)] (1 — + + 4. 9 / Compressible.10). stabilizing).17) with equation (9. As noted by Croom et al. (1988) "The data show good levels of static directional stability below 20deg. resulting in unstable to neutrally stable values for trim canard incidences." Using computeroriented numerical techniques. the directional stability is reestablished. This fixes the section circulation.
6.60 —0. Thus. Bailey. = 0. The wing has an aspect ratio of 4 and a taper ratio of 0. Thus.2 2y/b 0. Balihaus. The solution of the difference equation is obtained by a relaxation scheme with the iterations viewed as steps in pseudotime.32 Comparison of computed and experimental pressure coefficients for sweptwing/fuselage configuration. The quarter chord of the wing is swept 45° and the strearnwise airfoil section is a NACA 65A006. there is a discrepancy between the experimental values and the theoretical predictions.4 2yIb = 0.80 2y/b = 0.20 CF Figure 9. a = 0°.4 —0. containing crossflow terms which have been previously neglected. The computed results were obtained using a Cartesian grid (x.0 x 106. the wingroot shock propagates laterally to 2y/b 0. y. 9.6. but the experimental shock dissipates before reaching that point. the maximum velocity is reached farther aft forthe stations near the root and moves toward the leading edge at stations nearer the tip. The experimental data were obtained at = 2.32. The combination of new and old values in the difference operators is chosen so that the related timedependent equation represents a properly posed problem whose steadystate solution approaches that of the steadystate equation. As was noted earlier in this chapter.5 I Transonic Aircraft 501 —0.95 2yIb = 0. A = 0. The agreement with experiment on the fuselage centerline and the two inboard panels is good.60.93. In the computed results.Sec. NACA 65A006 streamwise section. z) of 91 X 59 x 27 for the wing as well as for the fuselage. The calculated and the experimental pressure distributions for a sweptwing/fuselage configuration at = 0.93 and a = 0° are compared in Fig. and Prick (1976) note that a modified smalldisturbance equation. 9. a smalldisturbance differential equation which could be used to describe the resultant threedimensional flow is .40 2y/b = 0. would be a suitable approximation to the full potential equation over a wide range of sweep angles.] potential at the section trailing edge linearly extrapolated from points above and below. This deficiency results for wings with moderatetolarge sweep angles.4 —0. [From Bailey and Balihaus (1975).
1 . A structured surfacefitted grid on the configuration surface. as shown in Fig. was Figure 9. (1987) . (1976) note that the use of an improved form of the governing equation alone does not guarantee that the shock waves will be properly represented. [From Kusunose et 2il. to provide an improved approximation to the full potential equation at the critical velocity. surfacefitted grid.502 Chap. (1987) generated solutions for the flow over advanced turboprop (with an ultrahighbypass propulsion system) airplane configurations. 9. Subsonic Flows and Transonic Flows (1 — ± + — ('y + I o'I'y+'xy Y ) ' 2 lvi cY31'X+'XX Note that the terms on the lefthand side of this equation are those of equation (9. Boppe recommends that the higherorder term. The Euler equations for threedimensional flows are presented in Appendix A. et al. Boppe (1978) notes that the crossflow terms. 9 I Compressible.33 Structured surfacefitted grid for an advanced propfan airplane. Ballhaus et a!. The finitedifference scheme must also adequately describe the physics of the problem. Using a structured.18) and those on the righthand side are the additional terms of the modified smalldisturbance formulation.33. Flowfield solutions for complex airplane configurations flying at transonic speeds can be computed using numerical programs that solve the Euler equations. and provide the ability to define shock waves which are swept considerably relative be included to the freestream flow.
A NACA 0006 airfoil section is used at all spanwise stations. (1987) to use the Euler equations in formulating the solution algorithm.Problems 503 generated first and then the flowfield grids were obtained using an elliptic grid gen eration scheme. require the construction of complex computational grids: meshsurface grids for singularity methods andspatial grids for field methods. Starting with the energy equation for steady. dye..29)]. Shockinduced boundarylayer separation can also limit the lift coefficient. derive equation (9. Miranda (1984) notes that the problem of grid generation is a crucial one for the practical application of numerical flow simulation. When discussing Whitcomb's area rule. 'A rectangular wing having an aspect ratio of 3. threedimensional transonic flow computational techniques are developing at a rapid pace. at a point where where . and the horizontal tail as well as the slipstream and 'the rotational flow (with different total pressure and total temperature) produced by the propellor led Kusunose et ai." The presence of trailing vortex wakes produced by the wing. SUMMARY Transonic flows are inherently nonlinear and highly three dimensional for wingbody combinations.. the strut. experimental investigations have provided considerable insights into desirable design practices for transonic configura tions. However. (1978). adiabatic flow of a perfect gas [i.e. PROBLEMS 9." Miranda (1984) indicates that modifying the boundary condition (i.e. it was noted that. (1987) the viscous effects are simulated by using the boundarylayer displacement thickness to create an effective "displacement body. What is the airfoil section and the aspect ratio for the equivalent wing in an incompressible flow? 9.85. Strong shockwave/boundarylayer interactions cause rapid increases in drag and can lead to buffeting. Kusunôse et a!. The nonlinear inviscid flow and the resultant shock waves present a considerable challenge to the analyst. (1987) chose this scheme because it offered "two major advantages: (1) the structural grid requires minimum storage for the flow solution since the grid points are well ordered and (2) the surfacefitted grid allows easier and more accurate implementation of the boundary conditions. the surface transpiration concept) is preferable to the displacement body approach because the surface geometry and the computational grid are 'not affected by the boundary layer. Calculate the distance between two streamlines. For the approach of Kusunose et al.1.5 is flying at = 0. there is a substantial local region of supercritical flow above the wing surface which results in local expansions of the streamtube areas.3." To account for the viscous effects on the wing.85 at 12 km.2. singularity and field. Both methods of solution. trailing vortex wakes come out automatically as part of the solution.7). Furthermore. and the rotational flows are captured naturally through the Euler equation formulation. 9. the Euler solver is coupled with a threedimensional boundary layer method McLean et a!. onedimensional. Consider flow past a twodimensional airfoil in a stream = 0. equation (8. for lifting conditions. "When these are solved.
Private Transmittal Forsythe JR. Astronaut. Aero. Lockheed Horizons 23:2—17 Gailman JW. Bernstorf DJ... Gratzer LB. 96—0554. CO Foss RL.Washington. 347 Ballhaus WF. 1975. 78—104. Recent advances in aerodynamics for transport aircraft. Grafton SB. Aeroelastically tailored wing design. Budiansky B. Blay R. 60:431—449 Forsythe JR. Mem. 1956. 1978.5)]. 1988. 1972. Reno. 2001 Boppe CW. Gringrich PB. Improved computational treatment of transonic flow about swept wings. Hoffmann KA. 2000. Ballhaus WF. Roy.AIAA Pap. 1973. Dayton. 881420. 1982. From propellers to jets in fighter aircraft design. Soc. DC:AIAA BrowniE. Frick 11976. Business jet wing design using aerodynamic shape optimization. NASA Tech. Strang WZ. Astronaut..Twodimensional aerodynamic characteristics of the NACA 0012 airfoil in the Langley 8foot transonic pressure tunnel. 2000—2552.. Aeronaut. AIAA Pap. 16th AIAA Pap. Pfeiffer NJ. NASA Conf Proc. 10(8):32—36 Bailey FR. 1987. is 1. Reuther JJ. The aerodynamic art. SAE Pap.Aircraft 32:1056—1061 Croom MA. Aeronaut. Meet. NASA Spec. Presented at AIAA Aerosp. Development of transonic arearule methodology. AL Bradley RG. Sci. Presented at The Evolution of Wing Design Symp. Denver. 1948. 11(12):30—45 Harris CD. In Transonic Aerodynamics. Murri DG. Computational transonic flow about realistic aircraft configurations. NV Goodmanson LT. Comparisons of computed and experimental pressures for transonic flows about isolated wings and wingfuselage configurations. 1995.20 in terms of the distance between these two streamlines in the undisturbed flow. [equation (2. Highalpha flight dynamics research on the X29 configuration using dynamic model test techniques. Presented at AIAA Aerosp. Blake DC. 80—3046. 34th AIAA Pap. 9 / Compressible. 1996. Huntsville. Proceedings of the Aerosp. Pub.. CA Diederich FW. Price MA. Anaheim. NACA Tech. OH CarlsenWD. REFERENCES Ayers TG. 1980. Subsonic Flows and Transonic Flows the local inviscid Mach number. dYe = Use the integral form of the continuity equation Assume an isentropic expansion of a perfect gas. Supercritical aerodynamics: worthwhile over a range of speeds. Conf. Ed Nixon D. 2000. Forrest WC. Meet. 81927 . Me. Bailey FR. Whipple RD. Validation of several Reynolds averaged turbulence models in a 3D unstructured grid code. Fratello DJ. Sci. 1981. J. Tech. Practical aerodynamic problems—military aircraft. J. Note 1680 Farren WS.504 Chap. Presented at AIAA Fluids 2000 Conf. Divergence of swept wings.
68:501—516 Hurley FX. 1975. 61:37—42 Kusunose K. Sawers D.Aero. Grismer MJ. Collins 1K. Presented at The Evolution of Wing Design Symp. Chen HC. 13:447—453 Strang WZ. P. 80—3047. A new dimension. 3123 Miller R. Roy. Large R. AIAA Pap. (Ed) 1969. and Lasers Conf. 1965.. DC: AIAA Taylor JWR. Cohen D. The defining methods of Cobalt60: a parallel. Dreikomponentenmessungen bis zu grossen anstellwinkeln an fuenf tragfluegeln mit verschieden umrissformen in unterschall und ueberschallstroemung. J. Transonic analysis for complex airplane configurations. Aircraft 21 :355—370 Newman PA. 1984. implicit. The X29 flight research program. unstructured Euler/NavierStokes flow solver. 64:449—464 Schlichting H. 92—0439. Spaid FW. 1994. Honolulu. 37th. Mem.1971. Application of computational aerodynamics to airplane design. The transonic flow past twodimensional aerofoils. 1957. Combat Aircraft of the World.86. New York: G. 1964. Bandettini A. J. HI McLean JTD. Soc. Roy. 1960. J.Plasma Dyn. Roy. NASA Contr. Putnam Sons Uhuad GC. Soc. 1984. Tomaro RF. Mackrodt PA. 30th AIAA Pap. 1999. X2353 Putnam TW.Weeks TM. An annotated bibliography on transonic flow theory. Z. Computer program to calculate threedimensional boundarylayer flows over wings with wall mass transfer. NASA Tech. Some developments in boundary layer research in the past thirty years.. Aero. High Speed Wing Theory. Methods of reducing the transonic drag of sweptback wings at zero lift. 22:2—12.AIAA Pap. Soc. Hall IM. NV Spearman ML. Sci. 1960.References 505 Hillaker H. J. Wind tunnel investigation of the transonic aerodynamic characteristics of forward swept wings. NASA RP1 028 Spalart PR. OH KUchemann D. Meet. AJAA Student J. 1978. The Technical Development of Modern Aviation.. Presented at AIAA Aerosp. X3244 Jones RT. NASA Tech. J. Dayton. Randall JL. Allmaras SR. A oneequation turbulence model for aerodynamic flows. 19th AIAA Pap. Presented at AIAA Fluid Dyn. Princeton NJ: Princeton University Press Krone NJ. 1999 StuartWG. NASA Tech. Roy.. Sci. 1970. Aero. Forward swept wing design. Presented at AIAA Aerosp. 99—0786. Rep. 1978. NV.Allison p0. 1980. 87—1196. Flugwissensch..Yu NJ. 1992. New York: Praeger Publishers Miranda LR. Aerodynamic characteristics of a sweptwing cruise missile at Mach numbers from 0. 1983. Private Conversation Holder DW. J. Aircraft 20:195—202 . Jan. Note D7069 StahlW. Roos FW. Mem. Reno. Stivers LS.. Washington. 1978. 1972. An introduction to the flow about plane sweptback wings at transonic speeds. 1960. Soc. Wallops Island flight test range: the first fifteen years. Reno. 1987. Mascum DL. Aero. Detailed transonic flow field measurements about a supercritical airfoil section. Meet.50 to 2.39 Rogers EWE. Northrop FS Case Study in Aircraft Design. 64:64—79 Shortal JA.
Nov. Presented in NASA CP 2001. 1976. Design forAir Combat. 1273 Whitcomb RT. 4. Clark LR 1965. 1987. "AdvancédTransonic Aerodynamic Technology". Eng. Mciii. 1976 Whitcomb RT. NASA Tech. 9 I Compressible. 1956. Sci. London: Jane's Publishing . X1109 Whitford R. Subsonic Flows and Transonic Flows Whitcomb RT. NACA Rep. A study of the zerolift dragrise characteristics of wingbody combinations near the speed of sound. Adv. An airfoil shape for efficient flight at supercritical Mach numbers.506 Chap.
3.10 TWODIMENSIONAL. the flow downstream of the shock wave(s) 507 . Wave drag can exist without shock waves. it is necessary to use the relations describing the nonisentropic flow through an oblique shock wave.e. The airfoil is assumed to to infinity in both directions from the plane of symmetry (i. there will be a significant drag force on a twodimensional airfoil in a supersonic stream. Thus. the PrandtlMeyer relations were used to calculate the aerodynamic coefficients for supersonic flow past a thin airfoil. even though the relations developed in this chapter neglect the effects of viscosity. This drag component is known as wave drag. the flow field is the same for any cross section perpendicular to the wing and the flow is two dimensional. In Example 8. SUPERSONIC FLOWS AROUND THIN AIRFOILS The equations that describe inviscid supersonic flows around thin airfoils at low angles of attack will be developed in this chapter.) Provided that the assumptions made in the derivations of these techniques are valid. (As will be shown. When relatively large compressive changes in the flow direction occur. it is a wing of infinite aspect ratio). When the leadingedge shock wave(s) is (are) planar.. when the compressive changes in direction are only a few degrees. they can be combined to solve for the twodimensional flow about an airfoil if the shock wave(s) at the leading edge is (are) attached and planar. However. In Chapter 8 we derived the PrandtlMeyer relations to describe the isentropic flow which results when a supersonic flow undergoes an expansive or a compressive change in direction which is sufficiently small that shock waves do not occur. the pressure increase calculated using the PrandtlMeyer relations is essentially equal to that calculated using the oblique shockwave relations.
we mean that the thickness. camber. proffles of the general cross section shown in Fig.1 LINEAR THEORY The basic assumption of linear theory is that pressure waves generated by thin sections are sufficiently weak that they can be treated as Mach waves. 10 / TwoDimensional. or linearized. causing relatively large wave drag. the leadingedge shock wave will be detached and relatively strong. Supersonic Flows Around Thin Airfoils cp <0 x <0 Figure 10. we can apply the method of small perturbations to obtain theoretical approximations to the aerodynamic characteristics of the twodimensional airfoils. theory for thin airfoils and then higherorder theories. the isentropic relations developed in Chapter 8 can be used to describe the subsequent acceleration of the flow around the airfoil.55) to calculate the change in pressure: p— = 0 LJco — (10. The pressure and the velocity changes for a small expansive change in flow direction (an acceleration) have already been derived in Chapter 8 = 0. is isentropic. Thus. First we treat the Ackeret. equation (8. If the leading edge is not sharp (or only slightly rounded).1. therefore. Under this assumption. By "thin" airfoil. 10. 10. we can use equation (8. The coefficients calculated using the linearized and higherorder theories will be compared with the values calculated using the techniques of Chapter 8. For these thin airfoils at relatively small angles of attack. and angle of attack of the sec tion are such that the local flow direction at the airfoil surface deviates only slightly from the freestream direction..e.la) (lOib) . Experience has shown that the leading edge and the trailing edge of supersonic airfoils should be sharp (or oily slightly rounded) and the section relatively thin. Let us define the freestream flow direction to be given by For small changes in 0.1 General features for linearized supersonic flow past a thin airfoil.508 Chap. the flow is isentropic everywhere. We shall consider. [i.55)].
which corresponds to an expansive change in flow direction.2 with the exact values of PrandtlMeyer theory for expansive turns and of oblique shockwave theory for compressive turns.1.lc) and (10. the pressure is greatest at the leading edge. Flow accelerates continuously from the leading edge to the trailing edge for both the lower surface and the upper surface. If the flow is turned toward the upstream Mach waves. 10. 10.2b) The lift. 10.4. equation (10. 10. being greater on the lower surface when the airfoil is at a positive angle of attack.1 Lift Referring to Fig.3. The pressure coefficient is zero (i. cos (10. Downstream.. Thus. Referring to Fig. the pressure coefficient is negative. linear theory provides suitable values for engineering calculations. and the moment coefficients for the section can be determined using equations (10. the local pressure coefficient is positive and is greatest where the local inclination is greatest.lc) which can be used to calculate the pressure on the airfoil surface. we can write = dx dx a (1O. For small deflections. since 0 is known at every point on the airfoil surface. we can set it equal to its tangent.e. the drag. Since the slope of the surface of the airfoil section measured with respect to the freestream direction is small.1 / Linear Theory 509 We will define the angle 0 so that we obtain the correct sign for the pressure coefficient both for leftrunning characteristics and for rightrunning characteristics. The pressure coefficients calculated using the linearized approximation and Busemann's secondorder approximation (to be discussed in the next section) are compared in Fig. A positive pressure coefficient is associated with a compressive change in flow direction relative to the freestream flow.Sec. we see that the incremental lift force (per unit span) acting on the chordwise segment ABCD of the airfoil section is dl Pi ds1 cos 01 — p1. 10.3) Employing the usual thinairfoil assumptions.2). Combining these relations yields = + 20 (1O.4) . 10. the local static pressure is equal to the freestream value) at those points where the local surface is parallel to the free stream.3) can be written as dl (Pi — dx (10.1. for the doubleconvexarc airfoil section shown in Fig.2a) = +a (1O.
Linear theory O 000 Secondorder (Busemann) theory +0.2 —0.0 —0. Supersonic Flows Around Thin Airfoils for compressions.1 —0._ 510 Chap.2 +0. .1 CF 0. degrees Figure 10. PrandtlMeyer —. 10 I TwoDimensional.2 Theoretical pressure coefficients as a function of the deflection angle (relative to the stream) for various twodimensional theories.3 4 Deflection angle.3 +0. 0.3 Detailed sketch of an airfoil section. f Oblique shock theory theory for expansions —— i.(X) x Figure 10. z Z.
we have dC1 (Cr.lc) and (10.1 / Linear Theory z 511 a a Zi(X) Figure 10. 10. and Cmx0.. Cl= 4a (10. the lift coefficient is independent of the camber and of the thickness distribution.5) Using equations (1O. Furthermore. in the linear approximation for supersonic flow past a thin airfoil. we have assumed that positive values both for and 01 represent compressive changes in the flow direction from the freestream flow. without loss of generality.Sec.7a) C and dx =0 C (10. Note that since the 0 at both the leading edge and the = Zi trailing edge. the angle of attack for zero lift is zero. (2a — — (10. We can calculate the total lift of the section by integrating equation (10. — (10.2).8) We see that. x/c (10.4 Thinairfoil geometry for determining C1. Cd. The liftcurve slope is seen to be only a function of the freestream Mach number.6) from 0 to x/c = 1.5) becomes dC. In coefficient form. since dC1 4 (109) .6) = —1 where. equation (10.7b) o Thus.
12) Using equation (10.4 is dd Pi + Pu ds1. we see that for 1.lc) for compressive turns and equation (10.10) Again.2) to approximate the angles.13) yields (10.14) by the average values that they represent. Thus. the liftcurve slope is less than the theoretical value for incompressible flow past a thin airfoil.7). we have dCd = Pi0i dx + dx (10. we can write f jc (dz)2 dx = (10. 10 / TwoDimensional. we find that the integration of equation (10.1. 1 11 [(p) (10.512 Chap.185. sin (10.9).12) yields + 2 + [2a2 dCd dx dx 2 c (dz11 [ —4a (dz1 dz1Vl /x\ + —1 + + — Using equation (10. Supersonic Flows Around Thin Airfoils Examining equation (10. equation (10.lSb) We can use these relations to replace the integrals of equation (10.11) + + (Oi + (10.lSa) i c dx dx = (10.13) 42 Cd 1 d 2 d 2 = + + Note that for the smallangle assumptions commonly used in analyzing flow past a thin airfoil. the coefficient for this frictionless flow model is .2 Drag The incremental drag force due to the inviscid flow acting on the arbitrary chordwise element ABCD of Fig.10) becomes dd = In coefficient form. 10. 1 Similarly. which is 2'n per radian. 10. using the assumptions common to small deflection angles.14) = tan Thus. equation (10.
due to lift + Cd. thickness + Cd. thickness is the Cdo of previous chapters. Equation (10.Sec. Referring to Fig. In coefficient form. parabolic arc airfoil.16) also indicates that. 10. for a given configuration. Such was also the case in Example 8. the wavedrag coefficient decreases with increasing Mach number.1 / Linear Theory d 513 Cd = = 4a2 2  + 1  + 1 — (10. Note also that. it is not necessary that shock waves be present for wave drag to exist.19) Substituting equation (10. as this small perturbation solution shows.20) = — .17a) where Cd 4a2 due to lift = (10. which is not present in subsonic flows.3. friction (10. the incremental moment (taken as positive for nose up relative to the free stream) about the arbitrary point x0 on the chord is (10. we have dCm0 = — Cr1) X (10.lc) into equation (10.19) yields (1Cm0 — (10. we could write Cd = Cd. which examined the shockfree flow past an infinitesimally thin. 10. Let us examine the character of the terms in equation (10.3 Pitching Moment Let us now use linear theory to obtain an expression for the pitching moment coefficient. is known as wave drag. the first term is called the wave drag due to li/I or the induced wave drag and is independent of the shape of the airfoil section. If we were to account for the effects of viscosity.This drag component. Since the lift is directly proportional to the angle of attack and is independent of the section thickness. The second term is often referred to as the wave drag due to thickness and depends only on the shape of the section.16) Note that the drag is not zero even though the airfoil is of infinite span and the viscous forces have been neglected. 10.17c) Note also that Cd.1.16).17b) and Cd = thickness 2 + (10.18) — p1)(x — xo)dx = where we have incorporated the usual smallangle assumptions and have neglected the contributions of the chordwise components of and P1 to the pitching moment.4.
This is In contrast to the thin airfoil in an incompressible flow where the aerodynamic center is at the quarter chord. we can then integrate to find the resultant forces and moments. the angle of attack. For purposes of discussion.4. the flow field has been divided into numbered regions. In each region the flow properties are such that the static pressure and the Mach number are constant. which correspond to each of the facets of the doublewedge airfoil.5.21) as Cm0= x —4a ii C) 4 dz x — dx c x0 / di—) \cj (10. as shown. Therefore. and the geometry of the airfoil neglecting the effect of the viscous boundary layer. and the pitchingmoment coefficient per unit span of the airfoil given the freestream flow conditions. As discussed in Section 5.2) into equation (10. 10 / TwoDimensional. the wavedrag coefficient. once we have determined the static pressure in each region.20) and integrating along the chord gives Cm0= —4a (1 x0'\ 2 1' dzi"\ x — x0 c (x'\ (10. the drag coefficient.22) shows that the aerodynamic center is at midchord for a thin airfoil in a supersonic flow.21) Note that the average of the upper surface coordinate coordinate defines the mean camber coordinate + z1) and the lower surface = ç1 Wecan then write equation (10. and the pitchingmoment coefficient Let us use the linear theory to calculate the lift coefficient. The only forces acting on the airfoil are the pressure forces. although they differ from region to region.22) where we have assumed that = z1 both at the leading edge and at the trailing edge. and the pitchingmoment coefficient for the airfoil section whose geometry is illustrated in Fig 10. the wavedrag coefficient. It may also be considered to be that point along the chord at which all changes in lift effectively take place.Chap. We seek the lift coefficient. Thus.1: Use linear theory to calculate the lift coefficient. the aerodynamic center is that point about which the pitching moment coefficient is independent of the angle of attack. equation (10. EXAMPLE 10.2. Solution: Let us now evaluate the various geometric parameters required for the linearized theory: fxtanio° — x) tan 100 x c zi(x)=s tan 10° forO x 1—(c—x)tanl0° . Supersonic Flows Around Thin Airfoils Substituting equation (10.
865 .0: C1 = 4(107r/180) = 0.1407 2 .8) to calculate the section lift coefficient for a 100 angle of attack at = 2.0 0 wave Expansion fan Figure 10.1 / Linear Theory Leftrunning 515 Expansion fan /11b = 2. 10. Furthermore.4031 Similarly.16): Cd= Therefore.5 Wave pattern for a doublewedge airfoil in a Mach 2 stream. 4(lOir/180)2 2 ri 10 \2 + Cd = 0.Sec.4031 0. JO / and Jo \CJ We can use equation (10.1407 The lift/drag ratio is dCd 1C1 0. the drag coefficient can be calculated using equation (10.
Busemann showed that a more accurate expression for the pressure change resulted if the term were retained in the expansion.. and (10.2 shows that Busemann's theory agrees even more closely with the results obtained from oblique shock and PrandtlMeyer expansion theory than does linear theory.1). (10. is everywhere zero.1 gives C1 and C2 for various Mach numbers in air. . we have Cm05 = 0 This is not a surprising result. 10.516 Chap.22) indicates that the moment about the aerodynamic center of a symmetric (zero camber) thin airfoil in supersonic flow vanishes. Table 10.05c (see Table 5.12). Similarly. the wavedrag coefficient. EXAMPLE 10.23) is always a positive contribution. and the pitchingmoment coefficient Let us calculate the pressure coefficient on each panel of the airfoil in Example 10.23). this air foil section is much thicker (i.. We note that the 02 term in equation (10.2 SECONDORDER THEORY (BUSEMANN'S THEORY) Equation (10.23b) Again.176c) than typical supersonic airfoil sections for which t 0. 0 is positive for a compression turn and negative for an expansion turn.2: Use Busemann's theory to calculate the lift coefficient.23a) = + C202 (10. t = 0.23).5). 10. Cd. and Note that Fig. His result [given in Edmondson. Supersonic Flows Around Thin Airfoils Note that in order to clearly illustrate the calculation procedures.1 using Busemann's theory. the turning angles must be smalL These assumptions imply that the flow is isentropic everywhere. along with equation (10. At midchord. Thus.. equations (10. We will use equation (10. equation (10. The result is a relatively low lift/drag ratio.e. can still be used to find C1. It is important to note that since the pressure waves are treated as Mach waves.19). since equation (10. 10 I TwoDimensional.lc) is actually the first term in a Taylor series expansion of in powers of 0. et aL (1945)] in terms of the pressure coefficient is C "= or 20 +I L 102 J C10 (10.22) for the pitching moment coefficient gives — —4a (i xo since the meancamber coordinate x.
336 1.1 Moo C1 C2 1.789 1.204 1.4 TABLE 10.614 2.56 1.60 1.596 0.748 1.965 3.402 3.30 1.708 1.50 4. For panel 2.18 1.00 2. the airfoil in this sample problem is relatively thick.408 0.300 3.949 1.013 8.316 21.242 2..984 1.333 3.34 1.248 1.2244 As noted.58 1.38 1.20 1.707 0.238 1.096 5.003 1.491 3.1 = 0 as before.90 2.200 Solution: Since panel 1 is parallel to the free stream.80 1.42 1.873 0..28 1.320 1.015 2.364 3.516 0.201 0 2.44 1. 10.0 00 1.930 1. .455 1.755 2.24 1. and therefore the turning angles are quite large.170 2.219 1.408 2.356 4.919 2.48 1.833 1.747 1.Sec.728 2.321 2.904 12.26 1. C. (1.467 1.36 1.313 15.232 1.00 3.618 1.383 2.601 30.00 10.599 3. y 1.609 2.2 1 SecondOrder Theory (Busemann's Theory) 517 Coe ificients C1 and C2 fo r the Busemann Theory for Perfect Air.10 1.16 1.40 1.063 2.635 1.1787 = —0.54 1.4031 2(22 — 1)2 + 0.70 1.103 2.041 1.14 1. As a result.050 6.269 1.12 1.4 + Cp2 — 4(2)2 + 4(2oir\2 = = + —0.109 2.862 2.50 1.32 1.771 4.193 3.307 7.404 10.204 2. For panel 3.155 0.880 1.46 1. the differences between linear theory and higherorder approximations are significant but not unexpected.00 5.654 3.916 3.288 2.52 1.22 1.503 2.50 4.129 2.670 1.529 1.
3846 Cd Similarly. 10 / TwoDimensional. Having determined the pressures acting on the individual facets of the doublewedge airfoil.5.e.4031 + 0.24b) where the signs assigned to the force.5818 For panel 4. Thus. As we have seen.5 (i.25b) Applying this relation to the airfoil section of Fig 10. C1 terms account for the direction of the + 210° = 0.24a) where is the halfangle of the doublewedge configuration.1787 = 0.e.5c). Supersonic Flows Around Thin Airfoils = 0. the theoretical solutions for linearized flow show that the midchord point is the aerodynamic center for a thin airfoil in a supersonic flow. Thus. Cp4=0 the flow along surface 4 is parallel to the free stream. 10.1400 100 (Cr3 sin 20° — sin 20°) Let us now calculate the moment coefficient with respect to the midchord of the airfoil section (i. ..Sc/cos (1O.) = 2 (1O25a) or Cd = 1 2 cos sin U (1O. the section wavedrag coefficient for a = Cd 100 is = 2 = 0. the force acting on a given facet will be normal to the surface and will act at the midpoint of the panel. We can use the fact that the net force in any direction due to a constant pressure acting on a closed surface is zero to get C1 = 1 (1O. relative to x = 0. we can calculate the section wavedrag coefficient: sin O(0.. Since the pressure is constant on each of the facets of the doublewedge airfoil of Fig. in each numbered region).518 Chap.Sc/cos &. let us now determine the section lift coefficient: since C1 = cos O(0.
3: Use the shockexpansion theory to calculate the lift coefficient. 10. Solution: Since the surface of region 1 is parallel to the free stream. such as the static pressure and the Mach number. and the section pitching moment coefficient for the inviscid flow. the downstream flow is isentropic.5. As was true for the approximate theories. Also note that We have accounted for terms proportional to tan2 Since the pitching moment due to a uniform pressure acting on any closed surface is zero. 10.26) Note that. 10. the section drag coefficient. We will calculate the section lift coefficient. the flow properties in each region. the .04329 10. although they differ from region to region. For purposes of discussion. a shock wave is formed as the supersonic flow encounters the twodimensional doublewedge airfoil of the previous example Since the shock wave is attached to the leading edge and is planar. In reality.5.15) for a review of the technique. a noseup pitching moment is considered positive. and the pitchingmoment coefficient. equation (10.11) through (5. as shOwn in Fig.i2 2 = (—p1 + P2 + J33 c2/8 — ii2 2 + (P1 — P2 — + (c2/8) 2 2 (10. the wavedrag coefficient.27) The reader is referred to equations (5. the isentropic PrandtlMeyer relations developed in Chapter 8 can be used to describe the acceleration of the flow around the airfoil. Thus. as usual. Let us usethis shockexpansion technique to calculate the flow field around the airfoil shown in Fig. the flow does not turn in going from the freestream conditions (oo) to region 1. C = 0._________ Sec. Thus.26) can be written as Cm05 (—Cr1 + + Cp3 — + (Cr1 — Cp2 — Cp3 + Cr4) tan2 8 (10. are constant.3 / ShockExpansion Technique 519 Cm05 = 1 i. the flOw field has been divided into numbered regions that correspond toeach of the facets of the doublewedge airfoil.3 SHOCKEXPANSION TECHNIQUE The techniques discussed thus far assume that compressive changes in flow direction are sufficiently small that the inviscid flow is everywhere isentropic. Thus. EXAMPLE 10.
 c Thus. P2Pc 1 TT2 co P2PcO /')\ 2 00 jP2 \\Pco Since the flow over the upper surface of the airfoil is isentropic. 2 \Pt2Pco Using Table 8.520 Chap. a Mach wave (and not a shock wave) is shown as generated at the leading edge of the upper surface. .36).0 v1 = 26.380° 61 = 0° = 0. The pressure coef ficient on the airfoil surface bounding region 1 is zero. 10 / TwoDimensional.83 To calculate the pressure coefficient for region 2. The Mach angle is 30° Since the surface of the airfoil in region 2 "turns away" from the flow in region 1. M2 = 2. the flow accelerates isentropically in going from region 1 to region 2. let us calculate the angle between the Mach wave and the freestream direction. Since the Mach wave is of infinitesimal strength.1. dv = Since the flow direction in region 2 is —dO 62 — —20° '2 is V1 — = 46. Supersonic Flows Around Thin Airfoils properties in region 1 are the same as in the free stream. it has no effect on the flow.0 Furthermore. M1 = 2. To cross the leftrunning Mach waves dividing region 1 from region 2. since the flow is not decelerated in going to region 1. However. we move along rightrunning characteristics. or equation (8.380° Therefore. Ptc'o = Pti = Pz2 Therefore. Therefore. to calculate the pressure ratios given the values for M00 and for M2. for completeness. Thus.
10. Thus. we can use Fig. we find that M3 = 1.3 to calculate the pressure increase across the oblique shock wave: = Poo 2.4(4) = 2 —0. 8. The shock wave decelerates the flow and the pressure in region 3 is relatively high.12b to find the value of directly. it is 0.90 Note that because of the dissipative effect of the shock wave.2588 The fluid particles passing from the free stream to region 3 are turned by the shock wave through an angle of 20°.5.3 . which is 1. M3 1. we must determine the pressure increase across the shock wave. and PL3 = Pr4 the presence of the shock wave causes Pt3 <Ptc. dO = +20° and P4 = 23.0352 = '\ * 1)1.608. Since we know that = 2. As an alternative procedure.3 / ShockExpansion Technique 521 (0.12 to find the shockwave angle. 2 " \poo =0.12c. Having determined the flow in region 3.66. we can use to sin 0 (instead of Mm).0 and 0 = 20°. which is equal = 53. the Mach number in region 4 (whose surface is parallel to the free stream) is less than the freestream Mach number. as discussed in Section 8.558° so M4= 1. dv=dO Since 04 = 0°. To calculate the pressure coefficient for region 3.66 Using Fig. Therefore. 8.20 v3 = 3.5°. One crosses the rightrunning Mach waves dividing regions 3 and 4 on leftrunning characteristics. and the correlations of Table 8. Whereas the flow from region 3 to region 4 is isentropic. 8. we can use Fig.Sec.848 Thus.558° 03 = —20° one can determine the flow in region 4 using the PrandtlMeyer relations.20.
Thus. In some cases.24). Thus. Supersonic Flows Around Thin Airfoils To calculate P . •0 = 0. . The ratio p3/pm has already been found to be 2.848) — 1. — Slfl — IV' 4 Each Mach angle is shown in Fig. it is of interest to locate the leading and trailing Mach waves of the PrandtlMeyer expansion fans at b and d. 10 / TwoDimensional.0108 We can calculate the section lift coefficient using equation (10..0] 1. respectively.4438 Similarly.522 Chap.782 for this airfoil section. 10.25) to calculate the section wavedrag coefficient.4(4. Cd = 2 100 (Cr3 Sjfl 20° — Cp2 sin 20°) Thus.1595 The lift/drag ratio in our example is = 2. using the subscripts land t to indicate leading and trailing Mach waves. we have = sin — '—1 2 — — 1 lv'! = 30° '—1 sin IV' 3 — — .5.848. \P00 JYIVIOO 2 the ratio P4 P3 = P4 Pt3 Pt4 P3 can be determined since both M3 and M4 are known.0) = 0. = (2. 2C 100 (—Cr1 — Cp2 COS 20° + Cp3 COS 20° + Cr4) = 0. . using equation (10.
1400 0.3846 0.3 / ShockExpansion Technique 523 To calculate the pitching moment about the midchord point. such as the section lift and drag coefficients.2.0108 0. Comparison of the Aerodynamic Parameters for the TwoDimensional Airfoil Section of Fig.0.Sec. if the results obtained by the method applied to a variety of airfoils are studied.4031 0. = 2.27) and get Cm05 = 0. The experimental values of the section moment coefficient exhibit the angleofattack dependence of the shockexpansion theory.6. The theoretical values are compared in Table 10. 10. we have used a variety of techniques to calculate the inviscid flow field and the section aerodynamic coefficients fOr the doublewedge airfoil at an angle of attack of 100 in an airstream where = 2. Thus.3 illustrates how to calculate the aerodynamic coefficients using the shockexpansion technique. Thus.2244 +0. a = 100 TABLE 10.0000 0.0000 —0.04329 +0.6.) Otherwise. The airfoil is reasonably thin and the theoretical values for the section lift coefficient and for the section wavedrag coefficient are in agreement with the data. with experimental values taken from Pope (1958). C1 is negative at zero angle of attack. but they differ in magnitude.0000 0.4031 +0.0. (The reader is referred to Table 5. Note that. for the airfoil shown in Fig. A disadvantage of the technique is that it is essentially a numerical method which does not give a closedform solution for evaluating airfoil performance parameters. 10.140] 0. 10. The theoretical values of the section aerodynamic coefficients as calculated using these three techniques are compared in Fig.5818 Shockexpansion technique 0. wave drag becomes prohibitive.2588 +0. the errors in the local pressure coefficients tend to compensate for each other when the aerodynamic coefficients are calculated.5.0000 —0. This is markedly different from the subsonic result.403 1 Cd 0.2 Linearized (Ackeret) theory cp1 Secondorder (Busemann) theory 0. where the section lift coefficient is positive for a cambered airfoil at zero angle of attack. 10.04728 Example 10.This approach is exact provided the relevant assumptions are satisfied.0000 —0. one observes that the most efficient airfoil sections for supersonic flow are thin with little camber and have sharp leading edges.4438 0. However.1595 0.0000 C1 0. Although the airfoil section considered in these sample problems is much thicker than those actually used on supersonic airplanes.1 to see that these features are used on the highspeed aircraft. there is reasonable agreement between the aerodynamic coefficients calculated using the various techniques.04728 . we substitute the values we have found for the pressure coefficients into equation (10. This is another example illustrating that the student should not apply intuitive ideas for subsonic flow to supersonic flows.660 0.
04 0. and the pitchingmoment coefficient about the midchord. The airfoil is symmetric about the chord line. for supersonic flow past an airfoil. The resultant expressions should include the freestream Mach number.04 Cd Figure 10.1 ltL2. 10 / TwoDimensional. and the thickness ratio t/c( = 'i).— Secondorder (Busemann) theory M =2. 0. for a fixed thickness ratio. Supersonic Flows Around Thin Airfoils — — — Linear theory * .L. the drag coefficient. the constants a1 and a2.00 0.6 Comparison of the experimental and the theoretical values of C1. Consider the infinitesimally thin airfoil which has the shape of a parabola: x2 = C 2 ————(z — Zmax) Zfl)ax .08 4 Cm —0. Cd.1. Consider supersonic flow past the thin airfoil shown in Fig.13 Shockexpansion theory 0000 0. Show that.0  U. the wave drag due to thickness is a minimum when a1 = a2 = 0. PROBLEMS 10. and Cm05.5. as taken from Pope (1958) 0.2  CI 0.. Use linearized theory to develop expressions for the lift coefficient. Figure P10.4 — Experimental values. P10.1.524 Chap.
10.17c)] that the section drag coefficient due to thickness is = If T is the thickness ratio.lc.3 and for the same angles of attack. the freestream Mach number is 2. The resultant expressions should include the freestream Mach number and the parameter. the lift/drag ratio.059 and Zmax = 0. (a) Find the expressions for the lift coefficient.7.0°.3.4. Repeat Problem 10.e. show that 2 + 4r2 thickness = —1 for a symmetric. doublewedge airfoil section and that 5.6. Zmax /c. (c) Compare the lift coefficient and the drag coefficient calculated using linearized theory = 2.13._________ Problems 525 The leading edge of the profile is tangent to the direction of the oncoming airstream. The calculations are for = 2. Consider the doublewedge profile airfoil shown in Fig. What is the maximum angle of attack at which this airfoil can be placed and still generate a weak shock wave? 10. 10.3 using the shockexpansion technique. and the moment coefficient about the leading edge. equation (10. In doing this problem.1. and the coefficient for the moment about the leading edge for the airfoil of Problem 10. 10. Repeat Problem 10.5.8. .lc with those calculated in Example 8. For linearized theory. it was shown [i.3.3.3 using secondorder (Busemann) theory. and the pitchingmoment coefficient about the leading edge. if the flow were incompressible subsonic. Also.059 and for Zmax = 0. the drag coefficient. Verify the theoretical correlations presented in Fig. Compare the pressure distributions for linearized theory with those of Example 8. the drag coefficient.04 and the freestream Mach number is 2. P10. Calculate the lift coefficient.3 10. we are verifying the values for K1 given in Table 11.0 at an altitude of 12 km. (b) Graph the pressure coefficient distribution as a function of x/c for the upper surface and for the lower surface. If the thickness ratio is 0.29° and 5. use linearized theory to compute the lift coefficient.6. the drag coefficient.33T2 Cd thickness = for a biconvex airfoil section.3. Make these calculations for angles of attack of 2. compute the static pressure and the Mach number in each region of the sketch.3. Furthermore. Use linearized theory for the following. (3) (5) Figure P10.. for 10. 10. This is the airfoil of Example 8. the lift/drag ratio. Note that for this airfoil section.
(c) Calculate CA. Cd.07L and it occurs at the mid chord.e.20 stream in a wind tunnel. For the wind tunnel. The airfoil is such that = 8°. The angle of attack is 6°. calculate the pressures in regions 2 through 5.9. i. Pti = 125 psia = 650°R. calculate the pressures in regions 2 through 5.00 stream in a wind tunnel (Fig. CN. exposed to a Mach 2. CN. (b) Using the lineartheory relations.. CA is the axial force coefficient for the force along the axis (i. Consider the "cambered.. and for this configuration. 10 / TwoDimensional.40c. normal to the chordline of the airfoil). (c) Calculate CA. where the freestream Mach number in the tunnel test section is 2.9.6c Figure P10..e.4c  0. The airfoil is such that the maximum thickness (t) equals 0.10). diamondshaped airfoil section (Fig.e.07c and is located at x = 0.526 10. normal to the chordline of the airfoil)..00 0. (a) Using the shockwave relations where applicable and the isentropic expansion relations (PrandtlMeyer) where applicable. Chap.e. the maximum thickness t equals to O. calculate the pressures in regions 2 through 5. The angle of attack is 10°.9 10.10 10. CA is the axial force coefficient for the force along the axis parallel to the chordline of the airfoil) and CN is the normal force coefficient (i. (b) Using the lineartheory relations. The angle of attack is 6°. Cd.07c. and C in0 Sc for this configuration. Consider the twodimensional airfoil having a section which is a biconvex profile. The airfoil is such that the maximum thickFor the wind tunnel. C1.11. . 2 = 2. 83 = 2°. all four facets have the same value of = 125 psia = 600°R. P10.2.10.20 Figure P10. parallel to the chordline of the airfoil) and CN is normal force coefficient (i. (a) Using the shockwave relations where applicable and the isentropic expansion relations (PrandtlMeyer) where applicable. = 125 lbf/in2 and = 600°R. Supersonic Flows Around Thin Airfoils Consider the symmetric. calculate the pressures in regions 2 through 5. R(x) = — The model is placed in a supersonic windtunnel." diamondshaped airfoil exposed to a Mach 2. x = 2. as shown in the sketch. ness t equals to 0. P10.
Consider the twodimensional airfoil having a section which is a biconvex profile. Calculate Cd (the section drag coefficient) for the biconvex airfoil section over an angle of attack range from 0° to 10°. The airfoil is such that the maximum thickness (t) equals 0.e. . where the freestream Mach number in the tunnel test section is 2. Corp.References 527 Using the shocklPrandtlMeyer expansion technique to model the flowfield for the biconvex airfoil section. Using the linearized theory relations to model the flowfield for the biconvex airfoil section.. 10. December 1945.The theory and practice of twodimensional su personic pressure calculations.Mumaghan FD. New York: Putnam Publ. The shock/PrandtlMeyer technique is to be used to calculate the pressure distribution for both airfoil sectitns. R(x) = The model is placed in a supersonic windtunnel.13. i. diamondshaped airfoil section. R(x) = — The model is placed in a supersonic windtunnel. Ed.. calculate Cd for the symmetric diamondshaped airfoil section (see Problem 10. What is the effect of airfoil crosssection on the form drag? REFERENCES Edmondson N. 10. = 125 = 600°R.2.07L and it occurs at the mid chord.Aerodynamics of Supersonic Flight. 1958.12. Also. where the freestream Mach number in the tunnel test section is 2. The angle of attack is 6°. calculate the pressure distributions both for the windward (bottom) and for the leeward (top) sides.2. Prepare a graph comparing the section drag coefficient (Cd) for the biconvex airfoil for the angle of attack range from 0° to 10° with that for the symmetric. Use the linearized theory relations to determine the pressure distribution that is required to calculate the form drag. Snow RM. calculate the pressure distributions both for the windward (bottom) and for the leeward (top) sides.07L and it occurs at the mid chord.e. The airfoil is such that the maxand imum thickness (t) equals 0. Present the results in a single graph that compares the pressure distribution for the windward side with that for the leeward side. Consider the twodimensional airfoil having a section which is a biconvex profile. Pope A. = 125 lbf/in2 and = 600°R. 1945. P:. Johns Hopkins Univ Appi Physics Lab Bumblebee Report 26. Present the results in a single graph that compares the pressure distribution for the windward side with that for the leeward side. i. JHU/APL.9).
Techniques by which the aerodynamic forces and moments can be computed can be classed either as panel (or singularity) methods or field methods. A variety of panel methods have been developed.g. the inviscid pressure distribution results in a drag component. that described in Cenko (1981).1. As noted in Chapter 10.e. linearizedtheory concepts also allow one to separate the dragduetolift into two fundamental components: (1) the vortex drag associated with the spanwise distribution of the lifting force and the resultant downwash 528 . neglects the effects of shock waves). approximating the aircraft surface. Furthermore.1 1 SUPERSONIC FLOWS OVER WINGS AND AIRPLANE CONFiGURATIONS The density variations associated with the flow over an aircraft in supersonic flight significantly affect the aerodynamic design considerations relative to those for subsonic flight. which can be used to generate solutions for compressible flow. as illustrated in Table 5. One or more types of singularity distributions (e.. even if one assumes the flow to be isentropic (i.. sources) are assigned to each elementary panel to simulate the effect that panel has on the flow field.Thus. e.g. Panel methods yield good results for slender bodies at small angles of attack.. the wing planform is such that the aspect ratio is relatively small. known as wave drag. the wing sections for supersonic aircraft tend to have a relatively small thickness ratio. the configuration is modeled by a large number of quadrilateral panels. Wave drag represents a significant fraction of the total drag in supersonic flows and is related to the bluntness of the configuration. For panel methods. Furthermore.
Additional rigor is obtained (at the cost of additional computational time and expense) if one solves the Euler equation [e. it is necessary to model the viscous effects and their interdependence on the inviscid flow. (at least) one Euler code was not well suited for the analysis of wings with separated flow at high lift and low supersonic speeds. requiring the use of higherorder prediction techniques which take into account the various nonlinear flow phenomena in determining the configuration aerodynamics. wings of infinite aspect ratio). methods based on lineartheory approximations no longer provide an adequate simulation of the complex flow fields. However. including flows with embedded shock waves and subsonic flows.. . Walkley and Smith (1987) describe a technique which uses a finitedifference formulation based on the characteristics theory of signal propagation to solve the conservative form of the full p0tential equation. in reviewing techniques for predicting wing leadingedge vortex flows at supersonic speeds. However. Wood and Miller (1985) note that. inviscid. supersonic. one should expect reasonable accuracy for the pressure distributions and force coefficients if one solves the flow field using the Euler equations. In this chapter threedimensional effects will be included since the wings are of finite span.Allen and Townsend (1986)1. the drag. associated with the longitudinal distribution of the lift and the resultant disturbance waves propagating into the surrounding air. and the pitching moment for supersonic flows past finite wings. Thus.g. vortex position. as the design lift coefficient is increased. a code based on a linearizedtheory method that was modified to account for both nonlinear attachedflow effects (lower surface) and nonlinear separated flow (upper surface) [Carlson and Mack (1 980)1 provided the best correlation with the experimentally measured vortex strength. Field methods include a wide variety of assumptions in the flow models. Thus.e. even computer codes that are based on the NavierStokes equations employ simplifying assumptions in the solution algorithm. Schiff and Steger (1979)1.g.. At highlift conditions. The highlift flow regime can be modeled using a fully threedimensional. which arises only for supersonic flow. For relatively slender vehicles at small angles of attack (so that the flow is attached and the boundary layer is thin).Chap. it is important to develop an understanding of the general features of supersonic flows and their analysis. Solution techniques employing the Euler equations allow one to incorporate entropy terms that are neglected in the full potential model. Furthermore.. attachedflow computational approach which employs the fullpotential equation. Instead. based on their comparisons between computed flows and experimental data. It is beyond the scope of this text to treat the various methods available by which one can calculate the forces and moments acting on supersonic vehicles. 11 I Supersonic Flows Over Wings and Airplane Configurations 529 behind the wing and (2) the wavedragduetolift. and total lifting characteristics. For such applications. we will continue to concentrate on configurations that can be handled by small disturbance (linear) theories. irrotational flow about wing of finite aspect ratio. NavierStokes formulations are needed [e. In Chapter 10 we evaluated the effect of the section geometry for flows that could be treated as two dimensional (i. The objective is to determine the influence of geometric parameters on the lift. In this chapter we consider steady. the flow becomes extremely nonlinear. To obtain suitable accuracy when computing the flow fields for higher angles of attack or for configurations where viscous/inviscid interactions are important.
As noted in Middleton et al. we proceed to a development of the governing equation and boundary conditions for the supersonic wing problem.7 and to Problems 5.530 Chap.2) where CDO. CD = CDO + (11. Dragduetolift Thus.16. The skin friction drag results from the presence of the viscous boundary layer near the surface of the vehicle (see Chapters 4 and 8). the zerolift wave drag can be calculated using either farfield methods (i. The nearfield method is used as an analysis tool for applications where the detailed pressure distributions are of interest. For values illustrating the relationship between k and the Mach number.friction + CD. 11. also known as the zerolift wave drag 3.. the dragduetolift is itself composed of the vortex drag and of the wavedragduetolift. . that is.1) can be written using the approximation introduced in Chapter 5. and proceed to discuss two solution methods: the conicalflow and the singularitydistribution methods. Experimental evidence indicates that equation (11.e. We close the chapter with discussions of aerodynamic interaction effects among aircraft components in supersonic flight and of some design considerations for supersonic aircraft. CD = CD. the reader is referred to Example 5. The farfield method offers advantages for fuselage optimization according to arearule concepts. thickness For supersonic flows. k (the dragduetolift factor) is a strong function of the Mach number.duetolift (11. Design—to define the wingcambersurface shape required to produce a selected liftingpressure distribution. aircraft drag is composed of 1.1 GENERAL REMARKS ABOUT LIFT AND DRAG A typical lift/drag polar for a supersonic airplane is presented in Fig. (particularly as they pertain to determining drag) of linearity on the part of the equation and the boundary conditions. As noted in Middleton and Lundry (1980a). the supersonic area rule) or nearfield methods (i.. We then outline the consequences. The wingdesign program includes methods for defining an optimum pressure distribution.e. (1980b).1. At supersonic speeds. Wavedragduetothickness (or volume). Example problems are worked using the latter method. 11. the zerolift drag coefficient. the integration of the surface pressure distribution).1) As noted earlier in this chapter. Skinfriction drag 2. The dragduetolift (which includes the trim drag) is computed from liftinganalysis programs. Reference temperature methods can be used in the calculation of skin friction coefficients for compressible flows (see Chapter 12).46). is composed of the sum of CD. equation (5. the wingdesign and the liftanalysis programs are separate liftingsurface methods which solve the direct or the inverse problem of 1. 11 / Supersonic Flows Over Wings and Airplane Configurations After a discussion of the general characteristics of flow about supersonic wings.thickness + CD.15 and 5.
the dragduetolift is about one half of the total.1 Drag buildup using superposition method of drag analysis. 11.2 GENERAL REMARKS ABOUT SUPERSONIC WINGS The unique characteristics of supersonic flow lead to some interesting conclusions about wings in a supersonic stream. [From Middleton and Lundry (1980a).2 / General Remarks About Supersonic Wings ZeroSkin friction lift wave drag 531 Drag duetolift and trim drag Lift Drag IT Uco Drag duetolift Skinfriction drag Zerolift wave drag and trim drag (wave and vortex) Figure 11. as noted in Carison and Mann (1992). 11. Consider the rectangular wing of Fig.Sec. The pressure at a given point is influenced only by pressure disturbances geny) on erated at points within the upstream Mach cone (determined by . For efficient flight at a lift coefficient which maximizes the lifttodrag ratio. 11. Lift analysis—to define the lifting pressure distribution acting on a given wingcambersurface shape and to calculate the associated force coefficients.] 2.2.
> I A \ / / I I 1' F ii / / / / P(x. 1 1/ Supersonic Flows Over Wings and Airplane Configurations Ma. 1 E Figure 113 Wing of arbitrary planform in a supersonic stream. emanating from P As a result.3). A supersonic (subsonic) trailing edge is that portion of the wing trailing edge where the component of the freestream velocity normal to the edge is supersonic (subsonic). .The remainder of the wing (ACDF) is not influenced by the tips and can be treated using the twodimensional theory developed in Chapter 10.532 Chap. we have the following definitions: 1. the wing tips affect the flow only in the regions BAC and DEF. 2. 11.2 Rectangular wing in a supersonic stream. In the case of an arbitrary planform (see Fig. y) B / i'D C\ £ Figure 11. A supersonic (subsonic) leading edge is that portion of the wing leading edge where the component of the freestream velocity normal to the edge is supersonic (subsonic).
As noted therein. Thus. relatively blunt leadingedges were possible without a substantial zerolift wave drag penalty.0 cruise condition.5 has a subsonic leading edge and a supersonic trailing edge. Points on the upper surface within twodimensional regions that are bounded by supersonic edges have flows that are independent of lower surface flow and vice versa. 11. "The inboard wing panel has a leadingedge sweep of 790. Figure 11. . 11. and EB and FC are subsonic trailing edges. In Fig.Sec. B. D. the assumptions made in that derivation are satisfied by thin wings in supersonic flow as well. We also find that the benefits of sweepback (discussed in Chapter 9) in decreasing wave drag are also present in the supersonic regime.4 Delta wing with supersonic leading and trailing edges.3 GOVERNING EQUATION AND BOUNDARY CONDITIONS In Chapter 9 we derived the small perturbation (or linear potential) equation (9. for points on the lower surface. Although the derivation was for the subsonic case. 11.13). AD and BC are supersonic leading and trailing edges. in many cases. The delta wing of Fig.5 Arrow wing with a subsonic leading edge and a supersonic trailing edge. while the arrow wing of Fig. which produces a subsonic normal Mach number at the Mach 3. The conclusion drawn in Chapter 10— that good aerodynamic efficiency in supersonic flight depends on thinairfoil sections with sharp leading and trailing edges—carries over to finite aspect ratio wings. portions of supersonic wings can be treated by the twodimensional theory of Chapter 10.3. An experimental investigation was conducted to determine the aerodynamic characteristics of a potential highspeed civil transport {Hernandez et al. and C are the points of tangency of the freestream Mach cone with the leading and trailing edges. A. respectively. Because of the subsonic leadingedge normal Mach number.4 has supersonic leading and trailing edges. Note that the points. (1993)]. 11.3 / Governing Equation and Boundary Conditions 1 533 1 / / / 'I' I' * p \ \ Figure 11." 11. AE and DF are subsonic leading edges.
the assumptions made in deriving equation (9. This condition is . This is a linear.7) the flow tangency boundary condition since U00 + at' U00 and the surface sponds to z5 0. whereas in a supersonic flow the influence of the disturbance is present only in the "zone of action" defined by the Mach cone emanating in the downstream direction from the disturbance. nor too great (hypersonic regime).41).This fundamental mathematical difference between the equations governing subsonic and supersonic small perturbation flow has already been discussed in Chapter 8. There.5) which is the same as the condition imposed on the flow in the subsonic case. Practically speaking. 111 Supersonic Flows Over Wings and Airplane Configurations In mathematical form. we have (4)z)z=o = as (11. we saw that a small disturbance in a subsonic stream affects the flow upstream and downstream of the disturbance. we have I w' U00 + )surface dz (11. An additional condition that must be applied at a subsonic trailing edge is the Kutta condition. Consistent with the defmition of the perturbation potential.3) cc 00 One observes from equation (11.5) becomes ( UOO + 4)x)surface dx (11. secondorder partial differential equation of the hyperbolic type. These behaviors are characteristic of the solutions to elliptic and hyperbolic equations. A boundary condition imposed on the flow is that it must be tangent to the surface at all points on the wing. equation (11.3) that the assumptions are satisfied for thin wings in supersonic flow provided the freestream Mach number (M00) is not too close to one (transonic regime). this restricts supersonic linear theory to the range 1.13) are I'ul\2 1 1 1 1 1 1 1 00 00 1 1 (11.534 Chap. respectively.13) in standard form (to have a positive factor for the term) yields — — — —0 (114) where 4) is the perturbation potential.2 M00 5. whereas equation (9. Rewriting equation (9. Mathematically. See the equation immediately following equation (7. Note that the equation of the surface is given by z8 = y).13) is of the elliptic type (when M00 < 1).6) Applying the assumption that the flow perturbations are small.
camber. Cpute = (11. and linearly combine the results to arrive at the final flow description. The ability to treat thin wings in this manner greatly simplifies what would otherwise be very difficult problems. for the purpose of analysis. compressible flows could be treated by applying an affine transformation [equation (9. In general. The perturbation potential for each of these components can be determined separately and added together to get the combined potential describing the flow about the actual wing.6): (a) a flat plate of the same planform at the same angle of attack.14)] to [equation (9. a wing can. However. solve each part by methods appropriate to it.5 SOLUTiON METHODS We learned in Chapter 9 that subsonic. 11. (d) resultant wing. . which could be handled by the methods bi Chapters 6 and 7 for two. (d) and thickness are additive in linear theory: (a) angle of attack.6 Effects of angle of attack.8).13)] and the boundary condi— tions. respectively. be replaced by three components (see Fig.14).4)] into [equation (9. respectively. and thickness distribution were additive. the linear nature of the governing equation and the boundary conditions allows us to break the general wing problem into parts..15)]. Physically.8) where the subscripts and stand for upper and lower wing surface at the trailing edge. (b) camber distribution. (10. 11.15)]. no affine transformation exists which can be used to transform [equation (11.8) means that the local lift at a subsonic trailing edge is zero.22). 11. camber. and (c) a wing of the same planform and thickness distribution but with zero camber and zero angle of attack. and (10.4 CONSEQUENCES OF LINEARITY In Chapter 10 it was shown that the effects of angle of attack. 535 (a) + (b) + (c) + Figure 11. the Condition represented by equation (11.This resulted In an equivalent incompressible flow problem [equation (9. Thus.Sec 115 / Solution Methods M. refer to equations (10. (b) a thin plate with the same planform and camber atzerø angle of attack.and threedimensional flows. (c) thickness distribution.
the singularitydistribution method.4). for indepth presentations of conical flow theory and its applications. The interested reader is referred to Shapiro (1954). properties are invariant along rays emanating from a point.. Shapiro (1954). The first. such as the velocity. ji I.it is more widely used today. the conicalWe will discuss two solution methods for flow method. 11 / Supersonic Flows Over Wings and Airplane Configurations equation (11. 11. point P at the wing tip).Thus. 11. P________________ b. (1951). we will present some results that are applicable to simple wing shapes.6 CONICALFLOW METHOD A conical flow exists when 'flow properties. . which is amenable to solution by wellknown methods (complex variable theory.g. However.). However. the static pressure. was first proposed by Busemann (1947) and was used extensively before the advent of highspeed digital computers.__________________ 536 Chap. z coordinate system to a conical coordinate system [as was done in Snow (1948)].. since properties are invariant along rays from the apex of the cone. and Fern (1949)] but was not generally exploited until the highspeed digital computer was commonly available. we will not go through the mathematical details of its development. are invariant along rays (e. If equation (11. y.7 In a conical flow. the solutions generated using the conicalflow method (where applicable) serve as comparison checks for solutions using the computerized singularitydistribution method.g. Region where flow is two dimensional I I I I I I I / ii ii I I jI I I I I I —— I I I A Region influenced by wing tip Figure 11. Fourier series. and the static temperature. Carafoli (1956) and Jones and Cohen (1957). The second. PA in Fig. Fern (1949).4) is transformed from the x. has been known for some time [Lomax et al.7) emanating from a point (e. A further transformation [Shapiro (1954)] results in Laplace's equation in two independent variables. the resulting equation has only two independent variables. The latter method is particularly suited to such computers and is more easily applied to general configurations than the former. etc. Since the conicalflow technique is not generally applicable and is not as adaptable to computers.
8 Examples of regions that can be treated with conical theory: (a) rectangular wing.6 / ConicalFlow Method 537 (a) (b) Shaded regions: can be analyzed with conical theory Unshaded regions: can be analyzed with twodimensional theory (c) (d) Figure 11.9. 11.6.9 represents the equation =—sin 2 . 11.8. 11. (c) swept swing. 11.1 Rectangular Wings Bonney (1947) has shown that the lift inside the Mach cone at the tip of a rectangular wing is onehalf the lift of a twodimensional flow region of equal area. Regions of various wings that can be treated using conicalflow theory are illustrated in Fig. 1 [tanjL' V The analysis can be extended to the interaction of the two tip flows when their respective Mach cones intersect (or overlap) on the wing surface. The curve of Fig. which is presented in Fig. (d) "double" delta wing. The case where the entire trailing edge of the wing is in the overlap region of the tip Mach cones is illustrated in .Sec. This is illustrated in the pressure distribution for an isolated rectangular wing tip. (b) delta wing. 11.
Three cases are shOwn in Fig. The pressure distribution in the region of overlap <when 1 /3 AR 2) is determined by adding the pressures due to each tip and subtracting from them the twodimensional pressure field determined by Busemann's secondorder equation (10. Linear theory applies. AR < 1. . Flow separation does not occur.11.9 Pressure distribution at the tip of a rectangular wing (p = actual pressure due to tip loss.10. AR. Note that /32 = — 1.538 Chap. 11 / Supersonic Flows Over Wings and Airplane Configurations p— P2d — tan tang Figure 11. Tip effects extend to the limits of the Mach cone defined by the freestream Mach number and not to the Mach cone defined by the local Mach number. The extent of the overlap region is determined by the parameter /3 .23).] Fig. 4. [From Bonney (1947). P2—d = corresponding twodimensional pressure). The key assumptions used in this development are as follows: 1. ConicalflOw theory is not applicable in the regions indicated in the Such cases will occur when /3 . 11. 2. 11. Secondary tip effects originating at the point of maximum thickness of a doublewedge airfoil are neglected. 3.
(a) (b) 'l\vodimensional theory applies Conical theory applies Neither twodimensional nor conical theory applies (c) Figure 11. AR> 2: no overlap. (b) 1 /3 'AR 2: overlap along the trailing edge.6 / ConicalFlow Method 539 C Figure 11. . (c) /3 AR < 1: overlap extends beyond the wing tips.11 Regions where conical flow applies for rectangular wings: (a) /3 .Sec.10 Effect of a subsonic wing tip on the pressure distri bution for a rectangular wing for which /3 AR = 1. 11.
4 0.4 a.o cc —0.0 0.1 0. Further. but the moment coefficient about the leading edge decreases. 11. 11 / Supersonic Flows Over Wings and Airplane Configurations A summary of the results from Bonney (1947) for the case of nonoverlapping tip effects is given in Table 11.6 o. Airfoils having the same crosssectional area will have the same center of pressure location. A comparison of conical flow predictions with data obtained from Nielsen et al. Conclusions to be drawn from this analysis are as follows: 1.1. 2. A decrease in the aspect ratio for a given supersonic Mach number and airfoil section results in decreases in the coefficients for the dragduetolift.24 0 4 8 0. and the pitching moment. (1948) is given in Fig. degrees Figure 11. the lift. the coefficients of lift and of dragduetolift for finitespan wings increase slightly. [Data from Nielsen et al.] .540 Chap. When the thickness ratio is increased. (1948). 3. Wing data —— — .75 X tiP 0. The center of pressure will move forward with a decrease in aspect ratio.12 12 0. The thickness drag will vary with the square of the thickness ratio for a given crosssectional shape. 4.12 Comparison of linear theory results with experimental data for a rectangular wing. 5.00 0 _0. the center of pressure moves forward both for airfoils and for wings as the thickness ratio is increased. Airfoils of symmetrical cross section with a maximum thickness at the midchord point will have the least drag for a given thickness ratio.53 0.0 0.Linear theory Isosceles triangle section 5% thick = 1.12 for a doublewedge airfoil for the conditions shown in the figure. Note that the behavior of the dragduetolift here is in direct contrast to its behavior in subsonic flow.04 0.08 8 L CLO•2 CD 0.2 CL 0.
K1: Type of airfoil A' 4 Double wedge Modified double wedge 6 Biconvex Source: Bonney (1947). 2 L} 2 2ARf3—1 in radians.1 ConicalFlow Results for Rectangular Wings Using Busemann SecondOrder Approximation Theory F/atP/ate Wing FlatPlate FiniteThickness Type Airfoil 4a 1 Airfoil FiniteThickness Wing (1 — C3A')] CL s 2AR • 4a(1 ) friction friction 4a2 CD + /3 4a2 friction j + C3A') (1 \ + + C0. L ] — 2)2 airfoil C3 = — 1)3/2 ' A' = chord squared Values of A'. friction — 2AR + + [  1 — C3A'(AR'j3 — C3A')] 2a 2a CM0 32 — i 1)] Xcp 2 + area .TABLE 11. .
13 Pressure distribution over the wing chord for a section of a swept wing with subsonic leading and trailing edges.14. There are penalties associated with sweepback (some of which show up at subsonic speeds) including reduction of the liftcurve slope.13.The tips and center portion of a swept wing can be treated with Shaded regions: conical flow theory applies Unshaded regions: twodimensional flow theory applies Figure 11. As noted earlier. Thus. where the leading and trailing edges of the wing are supersonic. the wing can be treated by the methods of Chapters 7 and 9. increased dragduetolift. Sweepback introduces additional (possibly severe) torsion because the applied loads on the wing act aft of the wing root. Furthermore.2 Swept Wings If the wing leading edge is swept aft of the Mach cone originating at the apex of the wing as shown in Fig.14 Regions of conical flow and of twodimensional flow for a swept wing. There are structural considerations also. 11. . they can be rounded similar to those used for subsonic speeds. and reduced effectiveness of highlift devices.6. tip stalling.542 Chap. even though the flight speed is supersonic. for sufficiently swept leading edges. 11. a leading edge that is swept within the Mach cone is referred to as subsonic leading edge and the flow approaching the wing is similar to subsonic flow. 11 / Supersonic Flows Over Wings and Airplane Configurations Subsonic leading edge cp Subsonic trailing edge a Section aa Figure 11. the disturbances propagate along the Mach lines "warning" the approaching flow of the presence of the wing. Consider the flow depicted in Fig. when the leading edge is subsonic. A swept wing has a greater structural span than a straight wing having the same area and aspect ratio. 11.
] conicalflow theory while the remaining portion of the wing can be analyzed by the twodimensional techniques of Chapter 10.15 NOmenclature for flow around a sweptback wing of infinite aspect ratio: (a) view in plane of wing. 11. (b) view in plane parallel to direction of flight. The freestream Mach number can be broken into the three components. The component tangent to . cosAcosa (c) Figure 11.. if an appropriate coordinate transformation is made. as shown in figure. (c) view in plane normal to the leading edge. [From Shapiro (1954). Refer to Fig. 11.15. (a) Section AA cosa (b) Section BB ccosA M. in which a segment of an infinitely long sweptback wing with sweepback angle A and angle of attack a are presented.Sec.6 / Conicalflow Method / I / / / 543 / / /  COSQ.
cl= and Cie = 2 cos A(1/cos A) (11.544 Chap. Combining equations (11. Thus.12). Cde = d 2 cos A(1/cos A) (11.11) = t ccosA = T cosA where t/c is the thickness ratio. and (11. ignoring viscous effects and noting that the wave drag is normal to the leading edge.13). (11. the total lift per unit span is not changed by the spanwise motion of the observer.14) to get .10) (11.14) where d cos A is the drag component in the freestream direction. The flow at about this section can be treated with the twodimensional theory of Chapter 10.9) and (11. 11. we can see that = or + )03 = Also. we may consider the equivalent freestream Mach number normal to the leading edge.9) = tan ? Te sin a = tan tan a cosA (11. and only generates pressure forces necessary to create lift. we get C1 = Cle(M00e) = Cie(1 — sin2 A cos2 a) (11. — sin2 A cos2 (11.13) while Cd = cos (11.15) For the drag. Referring to the geometry presented in Fig.12) Similarly. Note that the airfoil section exposed to taken in a plane normal to the leading edge.9).15. 11 I Supersonic Flows Over Wings and Airplane Configurations the leading edge is unaffected by the presence of the wing (if we neglect viscous effects). This will be the flow as seçn by an observer moving spanwise at the tangential will be that Mach number sin A cos a. Now. Thus. we combine equations (11.
Sec. 11.6 I ConicalFlow Method
j
'
545
Cd = Cde
I
)
=
COS
A(1
— S1fl A COS
2
(11.16)
The results just derived are true in general. If we restrict ourselves to the assump
tions of the linear (Ackeret) theory, which was discussed in Chapter 10, then
Cie =
(11.17)
Cde
— 1
+
(11.18)
These
results are identical to those obtained in Chapter 10 for infinite aspect ratio
wings with leading edges normal to the freestream flow direction. See equations (10.8) and (10.16). The results of Ivey and Bowen (1947) for flow about sweptback airfoils with doublewedge profiles are presented in Fig. 11.16. Note that significant improvement in performance can be realized with sweep back, The results in Fig. 11.16 are based on the exact relations [i.e., equations (11.15) and (11.16)], the shockexpansion theory (not linear theory), and the assumption that the skin friction drag coefficient per unit span is 0.006.
Cl
Cd
0.00
0.16
0.32
0.48
0.64
0.00
0.16
Cl
0.32 Cl
0.48
0.64
(a)
(b)
Figure 11.16 Theoretical effect of sweepback for doublewedge section airfoils with supersonic leading edges: (a) lifttodrag ratio for = 1.5; (b) lifttodrag ratio for = 2.0;
546
6
Chap. 11 / Supersonic Flows Over Wings and Airp'ane Configurations
0.12
5
a=
= 0.05
4
0.08 = 1.5
c
3
2.0
2
0.04
1
0 0.00
0.16
0.32
Cl
0.48
0
20
40
60
80
A (deg)
(d)
(c)
Figure 11.16 (Continued) (c) lifttodrag ratio for
= 4.0;
(d) liftcurve slope. [From Ivey and Bowen (1947)J
11.6.3
Delta and Arrow Wings
Puckett and Stewart (1947) used a combination sourcedistribution and conicalflow
theory to investigate the flow about delta and arrowshaped planforms (see Fig. 1L17).
Cases studied included subsonic and/or supersonic leading and trailing edges with doublewedge airfoil sections. Stewart (1946) and Puckett (1946) used conicalflow theory to investigate the flow about simple delta planforms. Two significant conclusions about delta and arrow planforms that can be drawn
from these studies are as follows:
1. Fcr wings where the sweepback of both leading and trailing edges is relatively
small, the strong drag peak at Mach 1 (characteristic of twodimensional wings) is
replaced by a weaker peak at a higher Mach number, corresponding to coincidence of the Mach wave with the leading edge. 2. Delta and arrow wings with subsonic leading edges can have lift curve slopes (dCL/da) approaching the twodimensional value (4/f3) with much lower values
of CDIT2 than those characteristic of twodimensional wings of the same thickness.
A theoretical comparison of a rectangular, delta, and arrow wing is given in Table 11.2. As noted in Wright et al. (1978), "One of the prominent advantages of the arrow wing is in
Sec. 11.6 / ConicalFlow Method
547
(a)
(b)
Figure 11.17 Delta and arrowwing planforms with doublewedge sections: (a) delta planform; (b) arrow planform.
Comparison of Aerodynamic Coefficients for Rectangular, Delta, and Arrow Wing Planforms for = 1.50
TABLE 11.2
Wing Plan form
Rectangular
Delta a
Arrowa
A=
700
A=
700
A = 700
b=0.2
b=0.2
f3cotA = 0.4
a=0.25
0.591
2.11
/3cotA = 0.4
AR=1
4 da da
a=0
0.544
1.94
0.554
1.98
Relative area,S
Relative root chord,!
Root thickness ratio, .jCD,thickness
1.00 1.00 0.10
0.0119
1.018 1.69 0.059
0.0048
0.938
1.26
0.080
0.0070
C00
0.0060 0.0179
5.25
0.0060 0.0108
8.6
0.0060 0.0130
9.3
/ CL \
(— J \CDJmax
a See Fig. 11.17 for a definition of a and b. Source: Puckett and Stewart (1947).
548
Chap. 11 I Supersonic Flows Over Wings and Airplane Configurations
Supersonic
leading edge
I
Subsonic leading edge
Delta
A
Figure 11.18 Comparison of induced drag for delta and arrowwing planforms. Note: /XCD [From Wright et al. (1978).]
the area of induced drag. This is illustrated qualitatively in Fig. 11.18, where the planform with a trailing edge cut out or notch ratio is shown to have lower induced drag. The second advantage of the arrow wing is its ability to retain a subsonic round leading edge at an aspect ratio that is of the same level as that of the lesser swept supersonic leading edge delta. The advantages of the subsonic leading edge are a lower wave drag at cruise and a high LID for subsonic flight operations due to increased leading edge suction."
11.7
SINGULARITYDISTRIBUTION METHOD
The second method that can be used to solve equation (11.4) is the singularitydistribution
method. Detailed treatment of the mathematical aspects of the theory as well as applications of it to various wing planforms are presented in Lomax et al. (1951), Shapiro
(1954), Jones and Cohen (1957) and Carison and Miller (1974). For simple planforms, the singularitydistribution method can provide exact analytical closedform solutions to threedimensional wing problems [see Shapiro (1954), Jones and Cohen (1957), and Puckett and Stewart (1946)]. However, the method is quite adaptable for use with computers to solve for flow about complex shapes, and this is where it is most extensively applied today.
In Chapter 3 we learned that the governing equation [i.e., equation (3.26)] for incompressible, irrotational flow is linear even without the assumption of small disturbances. This allowed us to combine elementary solutions (i.e., source, sink, doublet, vortex, etc.) of the governing equation to generate solutions for incompressible flows about shapes of aerodynamic interest. In supersonic flow, where the small disturbance assumption is necessary to linearize the governing equation [e.g., equation (11.4)], there are analogs to the simple solutions for the incompressible case. Owing to a mathematical similarity to
their incompressible counterparts, the supersonic solutions are quite naturally referred to as supersonic sources, sinks, doublets, and vortices. However, the physical relationship to their subsonic counterparts is not quite so direct and will not be pursued here.
Sec. 11.7 I SingularityDistribution Method
549
The supersonic source (recall that a sink is simply a negative source), the doublet, and the horseshoe vortex potentials given by Lomax (1951) are as follows:
Source: & =
—
(11.19a)
Qz/32
Doublet: 4d = +
(11.19b)
where the axis of the doublet is in the positive z direction.
Vortex:
In equation (11.19), Q
is
—
=
(11.19c)
the strength of the singularity and
= {(x
—
P2[(Y
—
—
+ z2]}O5
x
x1
= (y
—
+
One can verify by direct substitution that equations (11.19a) through (11.19c) satisfy equation (11.4). Note that the point (x1, zi) is the location of the singularity. Since the wing is in the z = 0 plane, z1 = 0 for every singularity for this approximation. This is because, in the singularity distribution method, the wing is replaced by a distribution of singularities in the plane of the wing. The hyperbolic radius, is seen to be imaginary outside the Mach cone extending downstream from the location of the singularity in each instance.Thus, the influence of the singularity is only present in the zone of action downstream of the point (xj, Yi 0). Using highspeed digital computers, modern numerical techniques consist of modeling a wing or body by replacing it with singularities at discrete points. These singularities are combined linearly to create a flow pattern similar to that about the actual body. The strengths of the singularities are determined so that the boundary condition requiring that the flow be tangent to the body's surface is satisfied at selected points. For configurations with sharp trailing edges it is also necessary to satisfy the Kutta condition at those sharp trailing edges which are subsonic. Once the singularity distribution is determined, the potential at a given point is obtained by summing the contributions of all the singularities to the potential at that point. Velocity and pressure distributions follow from equations (9.12) and (9.16a). Four types of problems that can be treated by the singularitydistribution method
[see Lomax et al. (1951)} are as follows: Two nonlifting cases:
1. Given the thickness distribution and the planform shape, find the pressure distribution on the wing. 2. Given the pressure distribution on a wing of symmetrical section, find the wing shape (i.e., find the thickness distribution and the planform).
550
Chap. 11 I Supersonic flows Over Wings and Airplane Configurations Two lifting cases:
3. Given the pressure distribution on a lifting surface (zero thickness), find the slope at each point of the surface. 4. Given a lifting surface, find the pressure distribution on it. Here, it is necessary to impose the Kutta condition for subsonic trailing edges when they are present.
Cases 1 and 3 are called "direct" problems because they involve integrations with known integrands. Cases 2 and 4 are "indirect" or "inverse" problems, since the unknown to be found appears inside the integral sign. Thus, the solution of inverse problems involves the inversion of an integral equation. One might expect that cases 1 and 4 would be the only ones of practical interest. However, this is not the case. Many times, a designer wishes to specify a given loading distribution (e.g., either for structural or for stability analyses) and solve for the wing
shape which will give that prescribed loading distribution. Thus, the engineer may
encounter any one of the four cases in aircraft design work. A variation of cases 2 and 3 is to specify the potential on a surface instead of the pressure distribution. Cases 1 and 2 are most conveniently solved using source or doublet distributions, while cases 3 and 4 are most often treated using vortex distributions.
11.7.1
Find the Pressure Distribution Given the Configuration
Consider a distribution of supersonic sources in the xy plane. The contribution to the
potential at any point P(x, y, z) due to an infinitesimal source at F' (x1, Yi, 0) in the plane is, from equation (11.19a),
C(x1,y1)dx1dy1
+ z2i where C(x1, Yi) is the source strength per unit area. Consistent with the linearity assumption, the flow tangency condition [equation (11.7)] gives the vertical (z direction) velocity component in the xy plane as
f32[(y
=
—
(11.20)
w'(x,y,O)
=
z)]
(11.21)
Taking the derivative with respect to z of equation (11.20) gives
a[d/(x,y,z)]
= dw (x, y, z)
C(x1,y1)f32zdx1dy1
{(x — x1)2
— —
yi)2 + z2]}L5
(11.22)
Note that x1 and Yi are treated as constants. Taking the limit of equation (11.22) as 0, we get dw'(x, y, 0) = 0, except very near the point (x1, yi) where the limit is z indeterminant (i.e., of the form 0/0).We conclude that the vertical velocity at a point in the xy plane is due only to the source at that point and to no other sources. In other words, a source induces a vertical velocity at its location and nowhere else. A source
Sec. 11.7 / SingularityDistrtbution Method
does, however,
551
contribute to u' and v' (i.e., the x andy components of the perturbation velocity) at other locations. We still must determine the contribution of the source at the pcint (x1, Yi) to the vertical velocity at P'(x1, 0). Puckett (1946) shows that the latter contribution is
dw'(x1,y1,0) = irC(x1,y1)
However, since this is the entire contribution,
(11.23a)
w'(x1,y1,O) = dw'(x1,y1,0) = irC(x1,y1)
Using equation (11.21), we see that
C(x1, Yi)
Yi) 1
Yi)
(11.23b)
=
d
(1L23c)
where A is the local slope of the wing section. Thus, we obtain the important result that the source distribution strength at a point is proportional to the local surface slope at the point.
Substituting equation (11.23c) into equation (11.20) and integrating gives
4(x, y, 0) =
—
J is [(x —
/ /
f f
A(x1,y1)dx1dy1
x1)2
—
f32(y
—
Yi)2]
05
(11.24)
where S is the region in the xy plane within the upstream Mach cone with apex at P(x, y). Once the potential is known, the pressure distribution follows from equation
(9.16); that is,
C,, =
—
2(a4/ax)
(9.16a)
Since source distributions are used where the airfoil section is symmetric (see the preceding cases 1 and 2), the pressure distribution determined by equations (11.24) and (9.16)
is the same on the upper and lower surfaces. Because linear theory requires that the
deflection
are small, the wavedrag coefficient at zero lift is
=2
JJsu
[f
y) dx dy
(11.25)
where S is the surface of the wing and the subscript u indicates the upper surface. The
formula contains the factor 2 to include the contribution of the lower surface since the section is symmetric.
EXAMPLE 11.1: Determine the pressure distribution for the singlewedge delta wing Let us determine the pressure distribution for the simple wing shape shown in Fig. 11.19, which is a zero angle of attack.This is a singlewedge delta with subsonic leading edges. The leading edge is swept by the angle ALE Granted, this wing is not practical because of its blunt trailing edge. However, neglecting the effects of the presence of the boundary layer, the effects of the trailing
552
Yi
Chap. 11 / Supersonic Flows Over Wings and Airplane Configurations
a
M0.
A
Section aa
a
Figure 11.19 Nomenclature and geometry for singlewedge delta wing of Example 11.1.
edge are not propagated upstream in the supersonic flow. Thus, to obtain the flow about a wing with a sharp trailing edge, we can add (actually subtract) the solutions for the flow around two delta wings of constant slope such that the desired airfoil section can be obtained. This additive process is illustrated in Fig. 11.20.Thus, the simple case considered here can be used as a building block to construct more complex flow fields about wings of more practical shape [see Shapiro (1954) and Stewart (1946)].
Solution: Consider the point P(x, y) in Fig. 11.19.The flow conditions at P are a result of the combined influences of all the sources within the upstream Mach cone from P. From symmetry, the vertical velocity perturbation vanishes ahead of the wing, and the source distribution simulating the wing extends only to the leading edges. Thus, the source distribution which affects P is contained entirely in the region ABPD. The potential at P is given by equation (11.24):
dx1 dy1
ABPD
ff
[(x
—
2
— /3
2
(y —
Yi)1
where we have moved A outside the integral sign, since it is a constant in this example.
Sec. 11.7 / SingularityDistribution Method
I
553
IA
I
'S,
/
/
/
,'
.—
——
—
MG.
B
E
X
Section AA
A
F
D
MG.
B
E,C
t
x
F'
Section AA
Figure 11.20 Geometry for a deltawing planform with a doublewedge section. [From Shapiro (1954).]
To carry out the integration, it is convenient to break ABPD into three separate areas, thus:
=
ADE
If
—
dx1 dy1
2
13(yyi)J
2
20.5
+
AEPC
ii
II
11
{(x
dx1 dy1
 13 (Y
dx1dy1
2
 Yi)]
+
If
[(x
"
— 13
2
(y
554
Chap. 11 / Supersonic Flows Over Wings and Airplane Configurations
To define the limits of integration, we note the following relationships
from geometry:
Along AD:
x1 = —yi tan ALE Along ACB: x1 = +y1 tan AlongBP: x1 = x — P(Yi —
Y)
Along DEP: x1 = x + 13(Yi
Finally, the coordinates of point B and D are
—
y)
+ 13y)tanALE
[
tan ALE + /3
—
x + 13y 'tan ALE + [3
[
Thus, we have
0) =
A
0
/3y)tan ALE —(x — f3y) tan ALE + 13 'tan ALE + /3
x+J3(y_y)
dyif
d
[(x — x1)2
+IdyiI
Jo
çy
px+/3(y1_y)
dx
[(x — x1)2
1xf3(y1_y)
dy1
JyltanALE
—
p2(y
—
Yi)2]
05
(x+J3y)
+
dx1
Jy
{(x —
—
132(y
yi)2]O.S
One can use standard integral tables and relationships involving inverse
hyperbolic functions [Le., see Hodgman (1977)] to show that the result of this integration and subsequent differentiation with respect to x is [see Puckett (1946)]
u'(x,y,O) =
A
2
—
=
05cosh
f(tan ALE'\[
—
1—
(/3y/x)2
10.5
(y2tan2A)/x2j
By equation (9.16a) the pressure distribution on the wing is
= —?Notice that is invariant along rays (y/x = constant) from the apex of the wing. Thus, as we might suspect from the geometry, this is a conical flow.
The wavedrag coefficient can be determined using equation (11.25). Figure 11.21 presents pressure distributions and wave drag for various configurations of single and doublewedge delta wings (see Figs. 11.19 and 11.20).
Sec. 11.7 / SingularityDistribution Method
3.6 3.4 3.2
3.0
555
2.8 2.6
2.4
2.2 2.0
1.8
1.6 1.4 1.2
1.0
0.8
0
0.2
0.4
s
0.6
0.8
1.0
(tan A)y/x
(a)
CDJJ
T2
0.5
b
(b)
Figure 11.21 Theoretical solutions for a delta wing. (a) pressure distribution for a singlewedge delta wing at a = 0. (b) Thickness drag of a doublewedge delta wing with a supersonic leading edge and a supersonic line of maximum thickness. [From Puckett (1946).]
556
Chap. 11 / Supersonic Flows Over Wings and Airplane Configurations
CD!3
T2
0
0.0
0.i
0.2
0.3
0.4
0.5
06
ft7
0.8
0.9
b
(c)
Figure 11.21 (continued) (c)Thickness drag of a doublewedge delta wing with a subsonic line of maximum thickness.
EXAMPLE 11.2: Prepare graphs of the pressure distribution for a singlewedge delta wing
Consider a singlewedge delta wing with a leadingedge sweep angle of 600, flying at a Mach number of 2.2. The wing has a thicknesstochord ratio in the plane of symmetry of 0.04. Thus, referring to Fig. 11.21a, the surface slopeA is 0.02. Prepare graphs of as a function (x — xLE)/c(y) for the four planes shown in the sketch of Fig. 11.22 (i.e., at stations y = 0.125b;
y=
0.250b;
y=
0.375b;
and y =
0.450b).
Solution: Let us first locate the Mach wave originating at the apex of the delta wing.
=
in the sketch of Fig. 11.23, the Mach wave is downstream of the
wing leading edge. Thus, wing leading edge is a supersonic leading edge. In
Sec. 11.7 / Singu'arityDistribution Method
0.450b
y
557
0.375b
M = 2.2
x
Cross section in the plane of symmetry
,'A= 0.02
t = 0.04Cr
I'
Figure 11.22 Sketch for Example 11.2.
y
s=1
= 2.2
X
Figure 11.23 Mach waves and lines of constant s for Example 11.2.
558
Chap. 11 / Supersonic Flows Over Wings and Airplane Configurations
the shaded region representing that portion of the wing between the lead
mg edge and the Mach wave, the pressure must be constant and the same as that for a twodimensional oblique airfoil. In this region,
2u'
P—
—
2A
2A
Uc,o —
=
(11.26)
—
tan2 ALE
The nomenclature used in equation (11.26) is consistent with that used
in