AFWAL·TM-84-203

PREDICTION

OF AERODYNAMIC

DRAG

CHARLES

E.

JOBE

AERODYNAl'1ICS

& AIRFRAME

BRANCH

AEROMECHANICS

DIVISION

JULY

1984

19970612 029
RELEASE; DISTRIBUTION

APPROVED FOR PUBLIC IS UNLIMITED

FLIGHT DYNAMICS LABORATORY AIR FORCE WRIGHT AERONAUTICAL LABORATORIES WRIGHT·PATTERSON AIR FORCE BASE, OHIO 45433
DTIC QU.ALITY INSPECTED .l

REPORT DOCUMENTATION
Public reporting burden for this collection of Information Is estimated

PAGE
Including the time for reviewing

Form Approved OMS No. 0704-0188

gathering and maintaining the data needed. and completing and reviewing the collection of Information. Send comments regarding this burden estimate or any other aspect of this collection of Information. including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202·4302, and to the Office of Management and Budget, Paperwork Reduction Project 10704-01B8), Washington, DC 20503.

to average 1 hour per response,

1. AGENCY USE ONLY (Leave blank) 4. TITLE AND SUBTITLE

REPORT DATE

JULy 1984

REPORT TYPE AND DATES COVERED

FINAL REPORT
5.

JULY 1984

FUNDING NUMBERS

PREDICTION OF AERODYNAMIC
6. AUTHOR(S)

DRAG

CHARLES E. JOBE
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION REPORT NUMBER

AERODYNAMICS AND AIRFRAME BRANCH AEROMECHANICS DIVISION FLIGHT DYNAMICS LAB ORA TORY AIR FORCE WRIGHT AERONAUTICAL LABORATORIES :WRTGHLEATTRRSON AFB OH 45433
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)

FLIGHT DYNAMICS DIRECTORATE WRIGHT LABORATORY AIR FORCE MATERIEL COMMAND WRIGHT-PATTERSON AFB OH 45433-7562 POC: Dr Charles E. Jobe. WLIFIGC. 937-255-R4R4: DSN 785-255-8484
11. SUPPLEMENTARY NOTES

10. SPONSORING/MONITORING AGENCY REPORT NUMBER

AFWAL-TM-884-203

128. DISTRIBUTION

AVAILABILITY

STATEMENT

12b. DISTRIBUTION

CODE

APPROVED FOR PUBLIC RELEASE; DISTRIBUTION

IS UNLIMITED

13. ABSTRACT

(Maximum 200 words)

Drag prediction is the most important and challenging problem in aerodynamics. Experimental, empirical, analytical and numerical approaches, singly and in concert, have addressed this problem with varying degrees of success (and notable failures). This report reviews the published material on drag prediction through January 1984., The subject of wind tunnel to flight drag correlation will be reviewed in a later report.

14. SUBJECT TERMS

15. NUMBER OF PAGES

Drag Prediction

210
16. PRICE CODE

17. SECURITY CLASSIFICATION OF REPORT

18. SECURITY CLASSIFICATION OF THIS PAGE

19. SECURITY CLASSIFICATION OF ABSTRACT

20. LIMITATION

OF ABSTRACT

UNCLASSIFIED

UNCLASSIFIED

UNCLASSIFIED

SAR
Prescribed by ANSI Std. 18 Designed using Perform Pro, WHS/DIOR, Oct 94

;;tandard Form 298

Rev. 2J9. 2-89)

(EG)

AFi-JAL-TM-R4-20J

FOREHARD This report was prepared by Charles E. Jobe of the Aerodynamics and Airframe Branch, Aeromechanics Hright-Patterson AFB, Ohio. Division, Flight Dynamics Laboratory,

The work was performed under ;·lork Unit 24041048,

Advanced Hethods for Aerodynamic Prediction. Drag prediction is the most important and challenging problem in aerodynamics. Experimental, empirical, analytical and numerical approaches,

singly and in concert, have addressed this problem with varying degrees of success (and notable failures). This report reviews the published material The subject of wind tunnel to

on drag prediction through January 1984.

flight drag correlation will be reviewed in a later report. This report has been reviewed and is approved.

e~O-~L~
CHARLES E. JOBE Technical Manager Aerodynamic Hethods Group USAF Chief, Aerodynamics and Airframe Branch

PREDICTION OF AERODYNAMIC DRAG 1 Introduction 1.1 Drag 1.2 Level of Drag Prediction Detail 1.2.1 The Preliminary Design Phase 1.2.2 The Detailed Design Phase 1.2.3 The Final Design Phase Subsonic Drag 2.1 Empirical Correlations 2.1.1 Bomber/Transport Aircraft 2.1.2 Fighter/Attack Aircraft 2.2 Detailed Drag Estimates - Component Buildup 2.2.1 Friction Drag 2.2.2 Form Drag 2.2.3 Interference Drag 2.2.4 Camber Drag 2.2.5 Base Drag 2.2.6 Miscellaneous Drag 2.2.7 Drag Due to Lift Transonic Drag 3.1 The Drag Rise 3.2 Detailed Drag Estimates - Component Buildup 3.2.1 Zero-Lift Drag 3.2.2 Drag Due to Lift 3.3 Numerical Transonic Aerodynamics Supersonic Drag 4.1 Friction Drag 4.2 Wave Drag 4.3 Lift-Induced Drag Numerical Aerodynamics
1

5 6 6

11 12 13 14 16 19 21 22 23 23 23

17

30 30
31

31 33 33 36 37 39 42
49

Bibliography

FIGURE
1 2
3

TITLE Test Approach Transport Aircraft Drag Buildup Probable Error During Design Drag Data Scatter About Drag Polar Fairings with Mach Number Typical Aircraft Drag Breakdowns Components of Aircraft Drag Equivalent Skin Friction Subsonic Parasite Drag - Bomber/Transport Subsonic Parasite Drag - Fighter/Attack Relative Wing Size Equivalent Body Shape Equivalent Body Concept Aerodynamic Cleanness

4 5 6
7

8 9

10
11

12 13
14 15
16

Comparison of Empirical Flat Plate Skin Friction Formulae for Incompressible Turbulent Flow Sunnnary of Experimental Research on Flat Plate Skin Friction, Incompressible Speeds Fuselage Profile Drag - Bodies of Revolution Fuselage Shape Factors M ~ 0.70 Airfoil Shape Factors Lift and Speed Regions for Calculation of Drag Due to Lift The Subsonic Drag Polar Drag Predictions Using Panel Methods Effects of Engine Power on Lift and Drag of Various Aircraft Components Historic Correlation of Transonic Drag Rise Transonic Drag Buildup

17
18 19

20

21
22 23 24

;i

25 26 27

F-5 CL versus CD; M = 0.9 Total Forces and Moments Comparison Between Different Computer Programs Correlation of Calculated and Measured Wing-Body Drag Divergence Mach Number, Compressibility Drag, and Drag Rise for Three Supercritical Wings. Supersonic Drag Buildup

28

29
30 31
32

Supersonic Skin Friction Supercruiser Type Configuration - Test Theory Comparison

Experimental and Theoretical Dsag Polars of Supersonic Cruise Airplanes Rc = 4.8 x 10 . Leading Edge Suction Effects, LES 216 Drag Polars, CDAF Configuration 5C3, Drag Performance Supersonic Volumetric Efficiency Zero Lift Drag Buildup Zero Lift Drag Comparison

33

34 35 36
37

iii

TABLE
1 2

TITLE Drag Scatter Aircraft Characteristics - Fighter/Attack Geometry

iv

Prediction of Aerodynamic

Drag

1 INTRODUCTION The state-of-the-art in aircraft drag prediction was defined as of 1973

by the AGARD conference entitled "Aerodynamic Drag.,,20 This chapter draws heavily on that work and attempts to update, to the extent possible, those portions relevant to aircraft drag prediction at subsonic, transonic and supersonic speeds. Special drag prediction problems peculiar to short take-

off and landing (STOL) designs, such as jet flap thrust recovery, and energy efficient aircraft concepts. for example laminar flow control, are beyond the scope of this chapter; as are drag due to speed brakes and fighter aircraft weapons carriage. Certainly, a sound physical understanding of drag will be

required before drag prediction becomes a science. projections and conclusions in this chapter agree. in general, with those of Wood161 who briefly addressed four relevant questions. The assessments,

"1.

Hbw well do ground based estimates for drag polars and engine charac-

teristics correlate to flight test results? 2. In what areas do performance prediction techniques work best or

worst, and why?

3. What are the differences in the way various manufacturers (airframe

could or should be combined to form a better or best method?"

Woodis answer to question one is, very well for up-and-away flight; however, his survey of twelve commercial transport aircraft showed six predictions were low (as much as 22%), four were high (up to 10%), and two were exact. The second question received a similar answer: predictions are An entire

best for up-and-away flight in the clean aircraft configuration.

symposium would be needed to address questions three and four unless the simple answer that no single methodology accepted. is best for every situation is

~Jood also observed that, lithe quality of the estimate is more a

function of the time and care taken to include all the details and higherorder terms than it is of the particular equations used.1I Williams
l158

synthesis of responses to the AGARD-FMP questionnaire

on

prediction techniques and flight correlation is also worthy of detailed study. The five questions asked of experts in six NATO countries were:

"l. niques? 2.

What are the advantage?/disadvantages

of different prediction tech-

What portions of the flight regime cannot/should not be addressed by

ground-based techniques? 3. Are there areas where analytical prediction can be better than wind

tunnel and/or simulation results; or vice-versa? 4. Are there methods of reducing differences between prediction and

flight test results? 5. Are there any new prediction techniques that should be emphasized?"

Williams

section on background to prediction/design

needs and discussion

of increasing technical demands is also especially relevant to the understanding of drag prediction.

The initial sections of this chapter describe drag prediction methods typically used to define the most promising configuration tailed analysis and wind tunnel testing. in origin, although numerical aerodynamics for further, de-

These methods are mainly empirical is presently providing timely data The latter sections of the

for this phase of design in some organizations.

report describe drag prediction methods based on wind tunnel results from models specifically constructed
by

for the particular design project and guided

aerodynamic theory, known as "wind tunnel to flight correlations." and Craig34 have recently published detailed accounts of the

ROOney1l9,120

meticulous testing performed to correlate the wind tunnel and flight measured aerodynamic drag of the Tomahawk Cruise Missile. The size of this missile discrepancies

permitted wind tunnel testing of the full-scale vehicle -- yet remained.

These correlations

and many others for all types of aircraft and

ss i les form the basis of evaluation for a new design. Despite many recent advances, it is generally conceded that accurate drag

predictions, based entirely on the solution of the equations of motion, or some reduced form thereof; are, at present, limited to the prediction of drag due to lift in the linear range and supersonic wave drag for a limited class of slender configurations. empirical correlations. All other drag predictions Progress in computational ultimately depend on

fluid dynamics or, more in the last decade and of new computers,

generally, numerical aerodynamics,

has been tremendous

will continue into the future due to the introduction flew turbulence models.73

faster, more accurate solution algorithms, improved resolution of grids and However, except for the isolated cases of drag-due-

to-lift at small angle of attack and supersonic wave drag for smooth slender bodies, drag prediction is beyond the capability of current numerical aerodynamic methods.

The aerodynamics of the aerodynamic reference mode1147, usually a subscale wind tunnel wing-body model with flow-through nacelles and reference nozzles are considered initially. This is one of many models tested during an

aircraft development program, Fig. 1, and the basic model for drag prediction. The uncorrected airplane lift, drag, and pitching moment characteristics derived from tests of this model. are

Drag corrections to account for the differ-

ences between the data from this model and the actual flight vehicle are discussed in the latter sections of this chapter. The methods of properly

combining these drag data, including thrust effects, are the subject of this book. 1.1 Drag The resultant aerodynamic force caused by a flight vehicle's motion with respect to the atmosphere is the summation of the normal, pressure forces, and the tangential skin friction forces acting on the vehicle's surface. This

resultant force is conventionally resolved into lift and drag components in the vehicle's plane of symmetry. The lift force is the aerodynamic reaction

perpendicular to the free-stream velocity direction which is the level equilibrium flight path direction. Lift is not directly a subject of this

chapter; however, drag, the component of the total force that opposes motion in the equilibrium flight path direction, approximately follows a parabolic variation with lift for heavier-than-air flight as shown in Figure 2. Thus,

the relative state of motion of the vehicle (equilibrium, or accelerated flight) is determined by the lift force, which is an appropriate multiple of the vehicle's weight, and the throttle-dependent thrust force.

Aircraft normally spend ninety percent of their flight time on nominally straight, unaccelerated flight paths where all the forces are in static equilibrium. This is the cruise condition that is usually considered the

standard condition for the design of an airplane.

This chapter is mainly

concerned with cruise drag prediction (Fig. 2) because of the importance of this flight condition. Drag prediction for accelerated flight (take-off,

landing, maneuver) will be treated as an important, although ancilliary, issue. The parabolic variation of drag with lift, or angle of attack, appropriately additive to the zero lift drag is shown in Fig. 2. Drag is produced by

the tangential (skin friction) and normal (pressure) forces acting on the vehicle's surface due to relative fluid motion from a basic fluid mechanics standpoint. Total drag has been dissected into a multitude of various compo-

nents, depending on the phase of the design process, the aircraft mission and configuration, analyst. the experimental data available and the persuasion of the drag

A properly defined, consistent reference condition for the entire

aircraft and propulsion system is crucial to the correct final comparison of theoretical, empirical, ground and flight test data in the latter design stages for performance guarantees. straight-forward The operating reference condition is

for commercial transports with a Single, known cruise design

point; however, nlilitary fighter aircraft with multiple design points pose severe additional problems. It is often difficult to define an operating Additionally, the variation in installed

reference condition in this case.

flight test thrust permitted in military engines of the same family can be an order of magnitude greater than that permitted in engines slated for commercial applications. 1.2 Level of Drag Prediction Detail

Three somewhat distinct levels of drag prediction sophistication and reliability are usually described by authors on this subject.20, 106, 110, 125, 167 In reality, drag predictions are updated continuously throughout the

entire design cycle of an aircraft as data from numerical methods al1d wind tunnel tests become available. However, decisions that affect the final

design are made based on the best available information at many stages of the design process. 1.2.1 The Preliminary Design Phase. This is the beginning of the design process. feasibility20 and conceptual167 design phase. It has also been called the

Primarily empirical methods are

used to assess the relative merits of many design concepts against the mission specifications generated by an apparent market opportunity or military Drag prediction error, as measured with

statement of operational need.

respect to future flight test data, is highest due to method error and geometric uncertainty in the configuration definition13, Fig. 3. Relative accuracy in the methodology is necessary at this stage in order to select the most efficient configuration to meet the design objectives. Absolute accuracy

is desirable in order to give equal consideration areas of the design space or mission profile.

to all new concepts and

Promising new concepts could be

excluded from further consideration at this stage if drag methodology incorrectly predicts that the mission cannot be achi~ved. An overall drag Several

target or maximum drag level is determined during this design phase.

compilations of drag prediction methodology are available4, 63, 101, 107, 122, 135 for rapid performance assessment. limitations. Additionally, Each has its own peculiarities and

each airframe manufacturer

has compiled large drag

handbooks that are highly valued and extremely proprietary. 1.2.2 The Detailed Design Phase. Drag validation becomes a rnore specialized and detailed continuous process. This phase has also been called the development20, definition109 phase. preliminaryl24, and project

Estimates and assumptions are gradually replaced by

predictions supported by intensive wind tunnel test data on the determinate aircraft design, as derived from the clean wing reference configuration aerodynamic reference configuration component of the configuration, assessed and synthesized. models, Fig. 1. and

Drag sub-targets for each

consistent with the overall target, are

In some development programs, prototype flight test reassessments, and analysis if drag

data provides the basis for diagnostics, improvements are found to be necessary. 1.2.3 The Final Design Phase.

Production aircraft performance guarantees with error bounds are made. This is also the pre-production20 or wind tunnel to flight correlationI04 phase.

A detailed analysis and interpretation conducted.

of the prototype test performance is

The drag predictive model is calibrated by reference to the

An excellent description of this process and its difficulties has been published by Rooney119 and Craig34 for the Tomahawk missile. Theory, empirical methods, and ground and flight test measurements continuously intertwined throughout the design, development life. are

prototype data.

and production

phases of an aircraft's strengths and weaknesses.

Each drag prediction method contains its unique

Strictly empirical methods, conventionally

based on flight test data, Reynolds number, gaps, Prob-

include the real aircraft effects of flow separations, steps,' protuberances,

rigging, etc., but are not readily generalized.

lems have even arisen when these methods are applied to aircraft of the same family from which the data base was generated.I5S Current theoretical methods that utilize large computers are capable of immediate generalization to almost any configuration, but are linlited in Prediction of the

accuracy because of the lack of real aircraft effects. drag-due-to-lift

is usually quite accurate up to drag levels where flow

separations become significant. empirical.

Prediction of the zero-lift-drag

remains

Progressive wind tunnel development programs contain a greater degree of reality than theoretical methods, lack some real aircraft effects, and are costly and time consuming. Correct interpretation of the data remains an art,

particularly when wall effects, Reynolds number scaling and propulsion system model data are included. Systematic total drag measurement errors greater

than one percent cannot be tolerated if drag prediction methods for the smaller drag components are to be validated.19
fl..

typical breakdown of flight test drag data is shown in Fig. 2.

The

industry standard for drag measurement (± 3% accuracy) is seldom achieved, even under carefully controlled conditions.IIB From these data, drag

correlations for each drag component are made and drag design charts are constructed as a basis for the preliminary design empirical method. The

pre l inri nary design study is considered as the reference point for judgment on program success. The initial decisions on wing area, aspect ratio, thickness

ratio and matching engine cycle and thrust requirements are based on early lift-to-drag (LID) estimates. A lower bound for L(D must be estimated in

order to assure that the propulsion system will provide the power necessary at the airplane ct'itical flight condition. Misjudgements at this design stage

w i l] result in an inefficient engine-airplane that is noncompetitive. The designer encounters errors from two sources at this stage; errors due to lack of configuration geometric definition, and errors due to inaccurate methods. These errors have been estimated by Bowes13 for long-range subsonic transports and are summarized in Fig. 3. He also suggests an aerodynamicist's The need for an

rating scale, where MDD is the drag divergence Mach Number.

accurate force and moment model as early as possible in the design cycle is clearly evident, from the following tables.
8

LID Level Achieved

MOD
± .002 ± .004
±

Rat i!!.9. Amazing Very Good Average Below Average

3%

± 5% ±7% ± 10%

.006

± .010

Most commercial aircraft flying today fit within this table; however, military programs have been far less successful in achieving these accuracy levels. This may be due, in part, to the less precise flight test data recorded from the older military aircraft as shown in the following table.

TABLE 1 DRAG SCATTER Ai rcraft 8-747 C-SA C-141 B-707 DC-10-30 F-14 F-15 YF-16 F-16 F-18
Tomahawk

Flight Test Data Scatter

Ref 13
86

(drag)

± 3.5% (COp) ± 2% (drag): ± 3.3% (drag)

86 13 44
118

5%

(drag)

3.5% (range factor) 1% 20% 5% 12% (COo) (drag) (drag) (CDmin) (drag) (drag) (drag) (drag)
±

173
18

152

± 20% ±1.5%
±
±

34 1% (COo)
71 124

Tornado Alpha-Jet

3% 5%

The table should not be read such that the data scatter is constant for a given aircraft throughout the flight envelope. representative of the cruise condition. The scatter stated is

Fig. 4, from Arnaiz's3 study of the

XB-70 flight test data, clearly demonstrates this point by showing the actual number of occurrences of scatter and the estimated scatter plotted against Mach number. The drag uncertainty estimates were obtained by combining many

individual errors in weight, Mach number, static pressure, angle of attack, thrust, acceleration, etc., in a root-sum-square calculation.

Performance estimates for military aircraft are more accurate than Table 1 indicates, since the major problems occur when one attempts to isolate the thrust and drag components from the measured performance data. steady-state, Current
±

level flight test data indicate that a repeatability of

1% can

be achieved if meticulous care is exercised during the engine calibration and flight test processes. This level of accuracy must be attained159 in order to obtain useful correlations for the smaller drag components shown in Fig. 2. These component correlations, necessary to develop accurate prediction

methods, are based on convention, tradition, and aerodynamic theory. Flight test data are, however, not presently capable of resolving drag into components. The accuracy of each component cannot be checked since only is available. Additional

total drag, deduced from performance measurements, instrumentation

that could resolve drag into components is superfluous to

aircraft performance guarantees and is usually not included on the flight test aircraft because of the additional cost to the development program. This is

particularly true of military aircraft development programs where budget cuts are traditional and there is little interest in developing drag prediction methodology for the next aircraft program.

10

The relative size of each drag component varies with aircraft type and flight condition. Butler19• Representative examples of the cruise drag and take-off

drag breakdowns for four different aircraft types are shown in Fig. 5 from The drag breakdown also varies with aircraft range and purpose

(military or civilian transport) within each aircraft' type. The conventional approach to drag estimation is to estimate and sum the

zero lift drag of each major physical component of the aircraft with allowances for interference effects and other small contributions due to lift. Their summation is intrinsically and then add the drag

inac·curate since aircraft

components are mounted at incidence to each other so each is not at zero lift simultaneously. But1er19 suggests that this classification of aircraft drag

into components independent of, and dependent on, lift should be rejected and drag could be estimated by compounding elements arising from different basic causes associated with fluid dynamics as shown in Fig. 6. Currently, each manufacturer has a methodology, or drag handbook, by

which a thorough drag estimate is made. drag handbook. Manufacturers

There is no single, industry-wide

may even use several drag accounting breakdowns There exist several basic

(depending on the customer) for the same airplane. sources for fundamental

drag level prediction; however, total drag compari-

sons, between prediction and test, require accurate flight data and expert analysis and are highly proprietary to each manufacturer.161 2 Subsonic Drag The aircraft drag at subsonic speeds, here taken to be less than the drag divergence Mach number, is traditionally minimum drag. decomposed into lift-induced drag and

Minimum drag is further divided into friction drag, profile drag.

drag, and interference

11

.' .

~,

The accuracy of the drag divergence Mach number, MDD prediction varies widely with configuration. Rooney·sl19 detailed assessment of the Tomahawk

flight test data indicates that MOD cannot be accurately predicted while Henne, Dahlin and Peavey57 conclude that a substantial the determination tries. capability exists for geome-

of drag divergence Mach number for transport type

The determination of drag levels and increments is rated marginal. 94 McGeer and Shevel1 were able to correlate transonic drag rise data for peaky type airfoils on older Douglas transport aircraft wings. Their correlation

was not accurate for the Boeing 747 or the F-111A/TACT aircraft, however. 2.1 Empirical Correlations In the preliminary (C.R. James) are used

design phase, gross or overall correlations These

to estimate the size and geometric features of the aircraft. correlations are traditionally

based on wind tunnel and flight test data from The key to this method is to obtain sufficient,

older, successful aircraft.

relevant data from aircraft of the same general type (bomber, transport or fighter, etc.) being designed in order to form trends and extrapolations confidence. Data collections of this type are continuously with

sought and mainThe

tained within the preliminary

design groups of the ~;rcraft industry.

data from a single cirframe manufacturer,

for a particular series of aircraft

similar to the new design, usually produces the most accurate drag prediction. Correlations of flight test data based on aircraft geometry provide

insight as to the relative importance of factors influencing subsonic minimum drag.

This section treats two types of aircraft:

bomber/transport

and

fighter/attack.

Minimum drag is related to wetted area through the use of an Fig. 7 compares flight derived

equivalent skin friction coefficient, values of Cfe with incompressible

flat plate turbulent flow skin friction

12

coefficients for jet aircraft of both types.

This comparison shows that, in

general, about two-thirds of subsonic minimum drag may be attributed to skin friction. The balance is due to form drag and interference. Form drag

results from flow over curved surfaces and flow separation induced by viscous effects. Interference drag is the result of mutual interaction of the flow components. The equivalent skin

fields developed by the major configuration frictior coefficient,

Cfe, is a convenient method to relate total drag to wetted area:

(including form and interference drag)

q Swet where D

Minimum drag, lb

q = Dynamic pressure, lb/sq ft Swet

Total wetted area, sq ft

The quantity Dlq is defined as equivalent parasite area, f. 2.1.1 Bomber/Transport Aircraft

The relationship of Cfe and f to wetted area is illustrated in Fig. 8. Experimental points derived from flight tests of several transport and bombers are indicated. Wetted areas vary from approximately 5800 square feet for the

Breguet 941 to 33,000 square feet for the C-5A.

Equivalent skin friction The

coefficients vary from 0.0027 for the C-SA to 0.0060 for the C-130.

variation in Cfe reflects factors which influence form and interference drag as well as Reynold's number effects resulting from variations in aircraft size.
and

For example, the three turboprop military transports (Br 941, C-130,

C-133) exhibit values of Cfe in the 0.0050 to 0.0060 range. These aircraft have upswept afterbodies characteristic of rear loading arrangements and Interference

consequently incur form drag penalties which increase Cfe•

effects of nacelles and propeller slip streams also contribute to the higher 13

drag levels.

Jet transports

(737, 727, and DC-8) , with a higher degree of The C-SA, because of its

aerodynamic cleanness, are in the 0.0030 range.

size, operates at higher Reynolds numbers and has the lowest Cfe of the transports evaluated. The low Cfe of the C-5A is due to its aerodynamic cleanness and very low aftbody drag increment, despite its military upsweep. 2.1.t Fighter/Attack Aircraft aircraft provides

Correlation of flight test data of fighter/attack

additional insight as to the relative importance of factors influencing subsonic minimum drag. This section illustrates effects of four parameters:

wetted area, relative wing size, fuselage shape, and aerodynamic cleanness. Correlations were developed from flight tests of eight jet fighters.
2 the aircraft and some key parameters

In table

are listed.

Wetted areas vary from

slightly over 1100 square feet for the F-84G to approximately for the F-4E.

square feet

Equivalent fuselage fineness ratios vary from about 4.5 for the These aircraft have single engines with the The relationship between f, Cfe, and points derived from flight

F-86H to 9.3 for the F-18.

exception of the F-4E/J, F-14 and F-18.

wetted area is illustrated on Fig. 9. Experimental

tests of the aircraft listed on Table 2 are indicate~ on this figure. Equivalent skin friction coefficients for the F-4E/J. vary from 0.0032 for the F-8D to 0.005?-

The variation in Cfe results primarily from factors which influence form and interference drag. Significant factors are: (1) relative
(3)

wing size, (2) fuselage shape, and The effect of relative wing size

aerodynamic cleanness. The ratio of mean

is illustrated in Fig. 10.

the equivalent skin friction coefficient,

Cfe, to the incompressible

flat plate value, Cfico' is shown as a function of 0, the ratio of the total wetted area to wing wetted area. The increase in this ratio with increasing

14

TABLE 2 AIRCRAFT CHARACTERISTICS GEOHETRY

FIGHTER/ATTACK

HODEL DESIGNATION

AREAS (SQ. FT. ) WING REFERENCE TOTAL ~VETTED

EQUIVALENT FUSELAGE FINENESS RATIO

A-7A F-4E/J F-8D F-84F F-84G F-86H F-IOOD F-105D F-18 F-14

375 530 375 325 260 313 400 385 400 565

1691 2092 1821 1257 1104 1186 1571 1907 2046 3097

6.43 5.88 8.16 5.92 5.08 4.57 6.61 8.86 9.28 8.22

14.5

o reflects the fact that wing surface tolerances are normally held much
tighter than other components, and protuberances, roughness, and leakage penalties are smaller for the wing than for other components. An important factor affecting fuselage drag is the degree of pressure or form drag developed due to flow separation induced by viscous effects. In

g~neral, flow separation is more pronounced with bodies of low fineness ratio. This effect is illustrated in Fig. 11 by defining an equivalent body which includes frontal areas of the fuselage, canopy, and wing to establish an equivalent fuselage fineness ratio. and defined.
A form factor, FB,

In Fig. 12 this parameter is illustrated

is defined from the empirical expression:

FB

1.02

1+

1.5 (t/d)1.5

Mach number, M, accounts for compressibility

effects, Fig. 11 also includes an

aerodynamic cleanness factor, w, which is related empirically to relative wing size, Fig. 13. The aerodynamic cleanness factor accounts for form drag and interference Examples are form and

effects not treated by fuselage shape parameters.

interference drag for wings and tails, protuberance and leakage drag, manufacturing tolerance effects, and surface roughness. The magnitude of wand the

variation with reiative wing size is shown in Fig. 13. Separate trend lines are identified as a function of the number of inlet and exit flows. line is determined by the single inlet-single nozzle configurations, The lower the

middle line by the data from the F-I05D with two inlets and one exit nozzle, and the upper line by the data from the F-4/J, F-14 and F-18 with two inlets and two nozzles.

15

These correlations suggest an initial approach for estimating subsonic minimum drag for similar fighter/attack component. 2. Calculate an area weighted mean length, 1m' for the complete configconfigurations: length for each

1. Calculate the wetted area and a characteristic

uration or estimate 1m from experience. 3. Use this mean length to calculate Reynold's number at the flight

condition to be evaluated. 4. Determine an incompressible skin friction coefficient, ' using fico equation for turbulent flow or a C

this Reynold's number.

The Karman-Schoenherr

similar expression may be used. 5. Calculate the wetted area ratio, 0, and use Fig. 13 to establish the inlet/nozzle arrangement. fuselage fineness ratio, lid, and determine

value of w for the appropriate 6. the form factor, F • B 7.

Calculate the equivalent

Determine equivalent parasite area, f, from the expression: f

Cf.

A similar approach developed for bomber/transports 2.2 Detailed Drag Estimates - Component Buildup As the design project progresses,

lCO

x w x FB x Swet would highlight those

factors influencing subsonic minimum drag for this type of aircraft.

the geometric definition of the air-

craft is formed and more detailed drag prediction methods are used to assess drag targets or design goals for each aircraft component. These goals must be

consistent with the overall drag target previously determined since the total configuration effects. drag is the sum of the component drags including interference

16

Each component may have an initial run of laminar flow, followed by a transition zone and fully developed turbulent flow thereafter. each flow region is difficult to determine with precision. The extent of

Fully turbulent

flow was usually assumed from the leading edge aft on full-scale flight vehicles at altitudes below about twenty kilometers. The smoothness and

rigidity of newer composite materials may provide longer runs of laminar flow on future flight vehicles. Transition is usually fixed by trip strips or grit The

on subscale wind tunnel models that are tested at lower Reynolds numbers. positioning and selection of boundary-layer trip strips that accurately simulate full-scale conditions is an art. 8 ' 14, 15 2.2.1 Friction Drag There exists little controversy concerning calculation of the average laminar skin friction on one side of a doubly infinite flat plate. It is

given by Blasius,9 formula from his exact solution to the laminar boundary layer equations for zero pressure gradient: Cf where Re = transition

i.32824/Re1/2

Uoo -- 1 and t is the distance fro~ the stagnation point to the v zone.

Transition takes place when the length Reynolds number nominally exceeds 5 3.5 x 10 to 106 for a flat plate.123 This is the critical Reynolds number, often assumed to be one-half million for many flows, although pressure gradients strongly influence the location of transition in the boundary layer. is the objective of boundary-layer stability theory to predict the value of Success in the It

the critical Reynolds number for a prescribed main flow.

calculation of the critical Reynolds number for flows of general aeronautical interest has eluded workers in this field for many decades despite many dedicated efforts.

17

Many correlations of turbulent boundary-layer

data exist.

In Fig. 14

the variation of Cf with Reynolds number, predicted by five different methods, is displayed. Significant
6

changes in Cf can occur when scaling from wind formulas are:

tunnel (Rn = 3 x 10 ) to flight (Rn = 40 x 106) if different correlations are used. Two well-known and widely-used Prandtl-Schlicting:

Cf

0.455
(log R )2.58
e

A
Re

where A depends on the position of transition and Karman-Schoenherr: Ct-1/2 ~ 4.13 log (Re Cfl formula, being explicit in C , has been more f 108 has shown that the Karmanwidely used in the past, although Paterson The Prandtl-Schlichting Schoenherr formula is a good representation
15).

of existing test data (Fig.

Both formulas have been used as the basis for, aircraft drag prediction 63 methods. , 107, 135 A new method, based on explicit Prandtl-Schlichting type relations, has been developed by I~hite and Christoph153, turbulent skin friction. Their explicit approximatipn Cf is accurate to
±
=

154 for compressible

0.42/1n2 (0.056 Re) It is the formula

4% in the Reynolds number range 105 - 109•

selected by Schemensky122

for his complete drag prediction method. is the body

The reference length used in the Reynolds number computation length for near bodies of revolution and MacWilkinson, of the mean aerodynamic chord for lifting surfaces.

et. a1.86 recommend use This choice of reference A

length will be less accurate (about 2% low) for highly swept delta Wings.

strip method, where a reference length is calculated for each streamwise strip of the wing, should be used.63

Compressibility

corrections

to the skin friction are less than 10% at

Mach numbers less than one and are often ignored at subsonic speeds unless continuity with supersonic 2.2.2 Form Drag Subsonic minimum drag is the sum of the friction, form and interference drag when the component buildup method is used. The form drag, or pressure predictions is desired.

drag, resulting from the effects of nonzero pressure gradient (component thickness) is usually accounted for by a multiplicative factor applied to the system for

skin friction drag. example,

Following SchemenskY's122

drag accounting FF * IF

* -S-- * Awet
ref +C Dbase

Components +C Dcamber +C

Eq. 1 Dmisc

wherp Awet is the component wetted area FF is the component form· factor, also called the shape factor, SF IF is the component interference The interference, sections. of the form drag, or FF, is dependent on empirical correlations as given by Hoerner64, Schemensky122, DATCOM63, 0'Conner107, Snodgrass135, MacWilkinson86, and others. Example formulas for the computation of FF are: for bodies: for nacelles: where FR FF FF
=

factor drags are treated in later

camber, base, and miscellaneous

The prediction

1 + 60/FR3 + 0.0025 * FR 1 + 0.35/FR Length

Component

Width * Height and for airfoils:

- --------

-------------------------------------

1 + Kl (tic) + K2 (t/c)2 + K3(t/c)4 where tic is the streamwise thickness to chord ratio and the Ki depend on FF
=

airfoil series.

An additional term, dependent on the section design camber is

added for supercritical airfoils. MacWilkinson, et.al.86 have shown that fuselage profile drag (the sum of skin friction drag and form drag) can be correlated for bodies of revolution by the method of youngl69, Fig. 16. This theory agreed with experimental data

from four transport fuselages that are radical departures from the ideal body of revolution, if the proper fineness ratio is used as shown in Fig. 17. However, excess drag due to upsweep93, which may amount to ten percent of the total cruise drag of military transport aircraft, is not apparent from these correlations. A new prediction method for reasonable estimates of the drag of

afterbodies for military airlifters

has recently been published by Kolesar and May72 to replace the method of Hahn, et.al.50• Methods of estimating a t rfot l shape factors (or form factors) were also compared by MacWilkinson, et.al.86• There is wide scatter in both the theoretical and experimental values of the shape factor, as shown in Fig. 18, particularly at twelve percent thickness to chord ra~io, a nominal value for modern transport wings. The data on the earlier NACA airfoils were rtot suited

to accurate assessment of airfoil form drag because of the testing techniques used. This earlier data, however, forms the basis of many older shape factor

correlations. Formulas for computing the form factors for other aircraft components, canopies, stores, struts, etc., are available from Hoerner64, Schemenskyl22, DATCOM63, O'Conner107, and Snodgrass135, their use for drag estimation. along with examples illustrating

20

Formally, the form drag follows the same dependence on Reynold's

number

and Mach number as the skin friction since the form factors are constant for a given geometry in most drag prediction methods. compressible transforming form factors that explicitly the appropriate Hoerner64 has developed

account for Mach number effects by ratio using the Prandtl-

thickness or slenderness

Glauert transformation

factor, {1 - M2)1/2.

The magnitude of these correc-

tions, however, precludes verification precision in experimental 2.2.3 Interference Aerodynamic Drag data.

of their accuracy due to lack of

interference

in aircraft is the change in flow over given due to the presence of one or more of the can be unfavorable with an attendant drag

components of the configuration other components. Interference

increase, or in the case of skillful design, favorable where the sum of component drags is greater than the total drag of the configuration. importance of interference requirements was delineated important interference to the optimization The

of aircraft to meet performance at a recent AGARD conference167• Generally,

effects involve fluid dynamic phenomena that is far too methods. between for

complex to be analyzed by existing computational Interference

factors, to account for the mutual interference are given constants based on experimental

aircraft components, the smaller components ccmponents. DATCOM63•

evidence

and are usually presented as plotted data for the major examples of this can be found in Schemensky122 and

Representative

The interference

factors, unlike the form factors, are correlated

as

functions of Mach and Reynolds numbers for the major aircraft components. These factors are usually greater than unity at flight Reynolds number since the resultant total drag is usually larger than the sum of the component drags

21

when tested singly. compressibility are constants! 2.2.4 Camber Drag

The factors also increase with Mach number to account for The interference factors for the smaller componen:s

effects.

This increment to the minimum drag accounts for camber and twist in the wing, and, in some drag accounting systems, the fact that all aircraft components are not mounted relative to each other to attain zero lift simultaneously. This increment is also related to the drag due to lift methodology

through the drag accounting.

A correlation consistent with Equation 1 is

C
where e K

DCAMBER

I-e

K ( C )2 L
span effici ency factor

= 1/ (,r *

AR * K)

=
CL

induced drag factor a constant. The lift coefficient for a minimum drag. Often

the design lift coefficient. The factors e and K are determined from the variation of drag with lift, to be considered later. A different drag predicting method has been shown in Fig. 2. In that

system, all changes in profile drag with lift are derived from experimental data correlations and the basic profile drag is not incremented for camber directly. The method of Fig. 2 is most accurate when the aircraft designs are

closely related members of the same series, for example, transports from a single manufacturer. Thus, drag data reduced using one bookkeeping system may not include a11 the interference, camber, etc., effects in the resulting correlations that the drag buildup method assumes are included, unless extreme care is taken to achieve consistency.

22

2.2.5

Base Drag This term is a weak function of Mach number at low speed and is given by

Hoerner 64 as: 8 CDBASE = (0.1 + O.1222M ) SBASE/SREF It obviously gains significance with increasing Mach number, however, it is separately from the body drag. independent of lift and is often not correlated 2.2.6 Miscellaneous Drag between the corrected aerodynamic

This term accounts for the differences

reference model drag and the full-scale airplane drag due to surface irregularities, such as gaps, mismatches, leakage due to pressurization. type,
RS

fasteners and small protuberances23,

and

It is estimated, from experience and aircraft

some percentage of the total friction, form and interference drags in In the later design stages, when configuration

the preliminary design stages. An AGARD conference170,

details are known, each of these small drag terms may be accounted for separately. devoted to this subject, was held in 1981.

The transfer of wind tunnel data to full-scale flight conditions is discussed under wind tunnel/flight correlation of lift, drag, and pitching moment. of each term in

It would be highly desirable to assess the creditability the drag buildup equations. Unfortunately,

for complete aircraft configura-

tions, this would require the ability to determine accurately the variation, with Reynolds and Mach numbers, of the drag forces due to viscous and pressure effects of each aircraft component - an intractable task. 2.2.7 Drag Due to Lift Various methods have been used to correlate subsonic drag due to lift, or induced drag, depending upon: the design project phase, the configuration, of the

the anticipated range of angle of attack, and the sophistication prediction method.

23

The complexity of drag due to lift prediction methodology in Fig. 19, from Schemenskyl22, of Simon
13J •

is illustrated

based on the extensive systematic correlations in each of the regions, one through seven, correlations The following

The drag is predicted

by modified or distinct equations and, perhaps, additional (terms) to account for increasing complexity section describes drag calculation critical Mach number, MeR. Schemensky I s 122 report. Classical aerodynamic

in the flow field.

in regions one, five, nnd six below the

The detdiled drag methodology

in

theory predicts that the drag due to lift is a shed behind the wing (downwash). It is

function of the vortex distribution parabolic with lift and is given by CD.
1

CC /71 AR spanwise. Wing taper and sweep produce

for an elliptic lift distribution

ll i p t i c loadings, thus

C~ u.
where e'
<

2 CL /71 AR e'

1.0 is the wing alone efficiency

factor.

An empirical modiflcation

to e' to account for the presence of the fuselage in the spanwise lift distribution results in
r

where

"D. 1
=

==

c2L /71

AR e

e ' [1. -

(d/b)2] The effects of camber are not

and d/b is the body diameter to wingspan ratio. included in these equations.

Camber causes a shift in the polar so that or tel depending on the Lo illustrated this camber effect Qnd the where flow C

minimum drag occurs at a finite lift coefficient, author of the method. nonparabolic Simon, et.al.131

nature of the "drag polar" at higher lift coefficients

separation effects become important, as shown in Fig. 20.

24

Simon131 found a further empirical modification necessary to account for local supervelocities

to this equation was

at the wing leading edge, while

Schemensky122, using the camber drag term, found that K :: l-R +

CL

---;;r;:R e

Here R is the leading-edge suction parameter and properly bounds K between l/nAR e
o

for full leading-edge suction (R

1.), and the upper bound of drag,

1/CLa, for zero leading-edge suction, or CD.

CL tan

a.

The suction

parameter, R. is significantly affected by airfoil camber, conical camber, leading-eoge radius, Reynolds number, and sweep,
. R-, 1 S d 122 an 1. nvo 1..J VE:u proce ure. It is

The proper determination of

within req i on one, bounded by the onset of leading-edge separation

with reattachment,

i.e., the CL for initial polar break boundary, and the critical Mach number boundary that surface paneling methods are accurate and These methods are all based on the Prandtl-Glauert equation,

can be applied.

a linearized form of the compl~te fluid dynamics equations of motion, and can predict the lift coefficient for minimum drag and the variation of drag with lift for complex, complete aircraft configurations quite accurately. Four panel methods were systematically evaluated in 1976 at an AGARD meetingl40; however, many new methods are currently in use; for example: DOUGLAS-HESS58, 59, MCAEROI6, PAN AIR26, 39, 130, 146, QUADPAN171, USSAER041, Each has its particular advantages and limitations143, equation, because of the 165, and VSAER038, 88.

in addition to those due to the Prandtl-Glauert particular numerical method used.

An evaluation of these six newer methods Drag predic-

against the datum cases of Reference 140 would be enlightening.

tion, not specifically considered during that meeting, should be added as an evaluation criteria.

25

Basically, all newer panel methods use Green's theorem and assume an algebraic form for the source and doublet singularity strengths on each of the panels that are used to represent the surface of the aircraft. Higher-order 16 methods , 26, 59 use polynomials to represent the singularities while lowerorder methods38, 41, 171 assume that the singularities are constant (or vary linearly in one direction) over each panel. type of representation Even though the accuracy of each

increases as the paneling density is increased, so does

the computer time and cost, thus forcing the usual trade-off between accuracy, schedule and cost. Panel methods have primarily been used as flow diagnostic tools3~J 146 by performing systematic comparative studies of alternative geometries and noting regions of high velocity or velocity gradient for further detailed analysis. The proper application of any panel method remains an art due to the myriad of paneling layouts that can be used to represent the configurations26, 149, the impact of various types of boundary conditions28, 142, and the many The calcunumerical ways of enforcing the'Kutta condition59, 143, 146, 171

lation of the induced drag from the velocity potential also presents problems. The near-field surnation of pressure multipl ied by the area projected in the drag direction results in differences of large numbers and a loss of significant figures. Control volume26 and the associated far-field or Trefftz plane 68 wake methods , 74 for drag calculations also present conceptual and numerical problems when applied to complete wing, fuselage and tailplane configurations. The optimum panel method primarily depends upon the skill of the user, the computational resources available, the geometric tG~figuration and the type and accuracy of the results desired. to be analyzed,

26

Although, drag prediction results are not often published in the open literature, two examples have been found to demonstrate the induced drag prediction capabilities of panel methods. The first, from Miller and Youngblood98, is shown in Fig. 21. The intricate paneling required for the advanced fighter configuration is shown, along with the drag polars predicted using PAN AIR130 and a lower-order method, USSAERO-B165• The minimum drag was predicted using the empirical The comparison with experimental data is

methods of the previous sections.

good, with the results favoring use of the higher-order PAN AIR method; however, the computer resources necessary to achieve the increased accuracy were considerably higher. Modeling studies, particularly with respect to the placement of the canard wake, were also necessary to achieve these results9S• The second example is from Chen and Tinoc031. The versatility of the PAN AIR method in predicting the effects of engine power on the lift and drag of a transport aircraft wing, body, strut, and nacelle wind tunnel model is shown in Fig. 22. The computed lift and drag increments on various components of the aircraft due to increasing the fan nozzle pressure ratio from "ram" to "cru i se" shows that the wing, the fuselage, and the strut all had favorable contributions, while the nacelle was the only component that contributed to a drag penalty. When the lost lift was restored and the associated induced drag About two drag

was included, the computed blowing drag was 1.3 drag counts. counts were measured in the wind tunnel.

Additional applications of panel methods include: fighter aircraft with externally carried stores28 (mutual interference effects). transport aircraft with flaps deployed102, and separated flow modeling using free vortex sheets88. Panel methods have also been used with boundary-layer calculation methods31, 70 to simulate viscous displacement effects and the parabolized

27

Navier-Stokes

equationslll.

These hybrid methods, designed to conserve necessary to analyze the flow the fluid Drag

computer resources by using only the equations

each appropriate
anc

zone, have had limited success in predicting

velocities prediction

press~rel11

lift loss due to viscous effects70•

is presently well beyond their grasp. in regions five and six of Fig.
19

Drag due to lift predict~on

is an physical that

empirical art and it is likely to be many years before satisfactory and mathematical models are achieved for the complex separated flows For example, the lift coefficient lift coefficient, CLn
B

occur in these regions61•

for polar break,

CIn r; ,and -':

for thin wings, R.eyno ld s nurn er, or h ichl y swept Wlngs, an d' reglon f'lve d oes not eX1s t122 b . . 191 The lower bound for region five is the lift coefficient at which leadingsepara-

the initial-stall

,coincide

edge separation tions.

and subsequent

reattachment

occur - the bubble-type

The upper bound is reached when stall occurs and the entire upper Blunt, thick airfoils generally exhibit trailingseparation.

surface flow is separated.

edge separ a t t on , wh i le very thin a.irfoils exhibit leading-edge Airfoil~ of moderate leading-edge, coefficients. thickness
by

are likely to separate and reattach at the separation and stall at higher lift produces an increase in drag Above drag break, the

followe~

trailing-edge

The leading-edge

bubble separation leading-edge

due to lift because of decreased flow separates completely

suction.

from the wing and the drag rapidly increases. The d~'ag cor re la t i cns developed by Schemensky 122. S'imon, et.al. 131 , and , Axelson4 can be used for rough estimates during the prellrninary design phase, but lack the logical and ccnsistent tailed design trade-off studies. framework necessary for confident, de-

As shoy.n in Fig. 20, the drag due to lift in The

region five varies with lift at a greater rate than in region one.

28

additional drag in region six is a separation drag increment that has been correlated with the lift coefficient at drag break131 Numerous symposia have been devoted to the prediction of viscous-inviscid interactions, separation and stall. Gilmer and Bristow45 accurately predicted

the lift and drag of three airfoils through stall by combining a twodimensional panel method, an integral boundary-layer method and an empirical

relationship for the pressure in the separated zone and wake recovery region. Maskew, Rao, and Dvorak89, also using a panel method with a simple wake model
wi

th prescribed separation line and wake geometry. were able to predict the over an aspect ratio six, 10° swept wing at 21c angle of

pressure distribution
a t t.ack ,

The subject of leading-edge separation from slender wings has recently been summarized by Hitzel and Schmidt62• Potential flow methods that model the leading-edge vortex and recently-developed Euler methods can predict the The Pohlhamus

flowfield and lift, but drag comparisons are not included.

suction anaiogy in conjunction wi th linear lifting surface theory24 has produced excellent comparisons with experimental wings. drag data for thin slender

However, the method fails for higher aspect ratio wings typical of

subsonic transport aircraft.

use of strakes and other leading-edge devices to reduce the drag at c speeds typ ice l of maneuvering fighter aircraft was the

t.igh 1 ift and transonf

subject of an extensive AGARD conference6l.

Design guidelines and empirical A

rules for strakes were determined from extensive wind tunnel data bases. fllndamental understandin9 of the effects of strakes on the wing flowfield,

lift ana crag was not evident in the proceedings. High angle of attack aerodynamics61 that is not a mature science. is certainly one area of aerodynamics

Considerable theoretical and experimental

29

effort will be required to develop accurate force and moment prediction methods for general configurations. 3 Transonic Drag The format of this section parallels that of Section 2. A rapid, strictly empirical method for determining the overall drag rise for preliminary design is presented, followed by a drag component buildup rr:ethodto assess drag targets for each aircraft component. The section concludes with a

summary of the status of numerical aerodynamics methods for detailed drag cleanup during the later design phases. 3.1 The Drag Rise (C.R. James) A correlation of the point drag difference between Mach numbers 1.2 and 0.8 for many fighter aircraft is shown in Fig. 23. be determined by this method. The increase in minimum drag coefficient at transonic speeds is determined primarily by configuratioh slenderness. Body fineness ratio (length/The shape of the drag

curve between these Mach numbers, where the drag may attain a maximum, cannot

diameter), lifting surface sweeps, and thickness ratio (thickness/chord length) are primary parameters for expressing slenderness of major configuration elements. The area rule concept97, 167 provide~ a convenient approach to combining body and wing cross-sectional areas to define an equivalent body of revolution. Fineness ratio of this equivalent body is used as an index of

configuration slenderness to derive a correlation of transonic drag rise. Historic data from twenty-eight aircraft, where the cross-sectional area

distribution normal to tne body centerline (Mach 1.0 cut) is used to define an equivalent body fineness ratio (Q/d), is compared in Fig. 23. breathing configurations, In all air-

the inlet capture area is subtracted from the body

total area aft of the inlet station to define the equivalent body shupe.

30

Minimum drag coefficient, CD~' is based on the maximum cross-sectional the equivalent body and the difference between drag values at M 0.80 is defined here as the drag rise. Three generations of aircraft are shown:

area of

1.20 and M

(1) early experimental airMost

craft, (2) the Century series aircraft, and (3) subsequent developments.

of the data are taken from flight test results, although data for some aircraft (the NASA LFAX configurations wind tunnel tests. and the XFY-1, for example) are taken from

The theoretical slender body values determined for Sears-

Haack126 minimum drag body shapes, pointed on both ends, is indicated as a lower limit. An overall progression toward this lower limit is discernable.

The first generation supersonic aircraft with fineness ratios of the order of 7.0 had drag rise values about 60% higher than the Sears-Haack second generation about 50%, and the third, about 25%. external store pod is the configuration Unfortunately, limit, the

The B-58 with an

nearest the lower limit. aircraft depart from the

some of the currently operational

trends established by the earlier generations.

The F-15, for example, with a

fineness ratio of 8.7 has a drag rise 200% higher than the Sears-Haack lower limit and the F-18 with a fineness ratio of 9.5 is 225% above this limit. 3.2 3.2.1 Detailed Drag Estimates - Component Buildup Zero-Lift Drag
A schematit diagram of the component buildup method is shown as Fig.

24. The skin friction drag may be computed by the methods of Section 2.2 at the appropriate Reynolds and Mach numbers for the flow. The friction drag is

nearly invariant over the small range of Mach number that is considered transonic, unless the localized details of shock boundary-layer are considered. interactions

31

Whi1 e the form and inter+erence

drags are a1so computed by the methods of Recall

Section 2.2, both are assumed to terminate abruptly at sonic speed.

that some form factors are constants over the entire subsonic range while others are functions of beth Reynolds and Mach numbers. Similarly, the wave

drag, to be treated in Section 4.2, is assumed to begin ~bruptly as Mach one is pxceeded. Since neither assumption is physically realistic, the zero-lift
by

drag curve is forced to be continuous through the first derivative

fitting

a polynomial through the zero-lift drag value at critical, or drag divergence, tljachnumber. Drag creep is accounted for by the variation of form and Thus,

interference factors with subsonic Mach number given in Section 2.2. the drag rise and the interference beyond Mach one.

plus form drag are replaced by wave drag

lhe decrease in critical Mach number with increasing lift for most airfoils causes the subsonic drag polar to begin its rise earlier at higher lift coefficients. The drag rise is usually separated into a minimum drag due to lift to account for this change.
, ")2

contribution and a contribution

The

drag rise due to lifting surfaces begins at MeR while the drag rise due to all ether components beSjins at Mr'

"Ro

as correl ated by Schemensky ": . problems and contradictions are evident in

Certainly. many unresolved this method.

It is merely an acceptable artifice to provide continuous drag For example, the component critical Mach

curves through the transonic regime. ratio122•

numbers depend on airfoil section. thickness ratio, or body slenderness It may be different for each a t rcref t component, causing each to

enter the transorlic regime at a different free steam Mach number, or interference between closely spaced components, such as aft pylon-mounted engi nes , may dominate the entire flow field and drag J ~7 curve fitting also do not account
fei

The drag curves ~enerated by this

the pressure drag slope relationships

32

'f derlved

"'1 rom transonlC Slml arli t y th eory 105 ,an

so on.

Transonic drag

prediction
3.2.2

is not yet a reliable, consistent

art!

Drag Due to Lift In the transonic region three of Fig. 19, bounded by the critical Mach

number and the limit Mach number ML1, where 0.95 ~ ML1 ~ 1.00, the induced drag is computed in the same manner as described for region one, Section 2.2. The drag rise for lifting surfaces from the previous section shifts the basic polar, depending upon MCR•

In the transonic region four above ML1, the drag

polar is calculated two and three.

by interpolation

A newer method for transonic drag due to lift prediction to extremely high angles of uttack has been programmed by Axelson4. This method is based on new forms of compressible flows and requires, ~ling theory covering potential and nonpotential inputs, the limiting The method predicts

in addition to the usual geometric

forward and aft chordwise locations of the shock wave.

the drag due to lift reasonably well for quite general, assumed limit shock positions, as shown in Fig. 25. The drag at zero lift is not calculated by

this program and must be supplied by the user.

Numerical Transonic Aerodynamics The progress in computational transonic aerodynamics has been reviewed in Th e b 00, e di t e d by N" k 1 lxon 105 ,

i .._' severa 1 recen t pu bl lca~lons 105, 145, 167, 168

describes steady transonic flow research from both an industry viewpoint and a research viewpoint and includes recent advances in experimental techniques and prediction methods. Yoshihara's AGARD reports167, 168 review transonic aerodynamics for the applied aerodynamicist engaged in the design of combat and interference167.

airlift aircraft with proper emphasis on aircraft component

33

The numerous authors in the field of numerical transonic aerodynamics have reached a consensus -- transonic drag predictions are currently unreliable by any method. predicted. Neither the polar shape nor drag levels are acceptably in the magnitude and often, even the sign, of

The uncertainty

incremental drag due to small geometric changes precludes the use of drag as an object function in minimization procedures. The surrogate transonic design

criteria is the pressure distribution which places a heavy burden on the aerodynamicist. boundary-layer The target pressure distribution must be chosen such that separation is avoided, wave and vortex drag are minimized, and

the resulting geometry is acceptable. Transonic flow is highly three-dimensional and inherently nonlinear.

heirarchy of approximate equations, in addition to the full Navier-Stokes equations, have been used to analyze these flows with varying degrees of success 168 . Unfortunately, the geometric fidelity of the numerical represen-

tation, necessary to distinguish drag differences due to design changes, diminishes as the higher approximations are used. This is a result of Grid

increased grid generation and computational

storage requirements.

resolution has been a primary cause for poor drag prediction using even the full Navier-Stokes equations with various turbulence models32. The theoretical approximate equations and numerical bases for drag calculations based on the (transonic small disturbance, full potential and Euler) Th e a dd i ,tlon
0f

. . . 'een questlone d103, 167, 172 an d remaln con t roverSla.1 have b

coupled boundary-layer further complicates

solver to the inviscid transonic flow calculation

the problem of accurate drag determination168•

Transonic flow calculations are presently limited in geometric capability to various collections of aircraft components and lack the full configuration

34

geometric cepab i l i ty of panel methods.

The following examples indicate the

current limitations and progress that is being made in this area. Longo, Schmidt, and Jameson84illustrated the improvement in transonic

airfoil drag prediction produced by modifying the original Bauer-GarabedianKorn-Jameson (BGKJ) nonconservative full potential method with weak viscous (VGK) as shown in Fig. flow solver

interaction to account for strong viscous interactions 26. The Dofoil method consists of a conservative,

full-potential

coupled to a set of integral boundary-layer methods with special models for separation bubbles at the trailing edge. Further improvements (unpublished)

have been obtained by substituting a finite volume Euler flow solver for the inviscid method. Henne, Dahlin and Peavey57, also using the Bauer, Garabedian and Korn method, obtained an excellent correlation of drag divergence Mach number. They found significant scatter in the drag level and compressibil ity drag increment results for eight airfoils. Hicks60 extensive review of transonic wing design using both conservative and nonconservative potential flow codes did not use drag as an evaluation

criteria since it was considered to be the least accurate quantity calculated by the aerodynamic codes. pressure distributions bodies. Poor experiment to theory correlations of the wing with low fineness ratio

were found for configurations

Henne, Dahlin, and Peavey57 again found drag divergence Mach number drag increment could

could be correlated while drag level and compressibility

not, for three high aspect ratio transport-type wings, as shown in Fig. 27.
., \~aggoner 150.,uslng . a sn~ 11 dt1S t ur bance transonlC ana 1· YS1S co de coup 1ed with a

two-o imens i ona l boundary-layer

code for the analysis of a supercritical found the method sensitive to design

transport wing-body configuration

changes, but an increment of 41 drag counts was added to the predicted drag

35

l eve l to obtain

agreement ".:ith test

data,

The con.put.at i ona l grid resolutior.

was coarse, particularly on the body nose ~nd bsattail region which may account for this dras increment. The preceding results show that transonic aerodynamics has enjoyed limited success when applied to high aspect ratio transport configuraticns. However, when applied to a high-performance fighter configuration v+th a wi de

complex fuselage. or canards, none of the full-potential straipht-forwaro stations.

methods is capable of

predicti0n of the wing pressures, even at the outboard span

With the possible exception of the development of a new transonic method based on paneling technology39, small disturbance equations. geometric qene ral i ty has been limited to the Shanker and Goebel127 have developed a numerical canard-wing i nt.e rDrag

transformation capable of modeling the closely-coupled actions tyoical of proposed highly maneuverable minimization for these complex configurations using a modified vortex lattice method.

fighter configurations.

has been addressed by Mason90

Transonic aerodynamic theory and the associated numerical methods arE current~y employed for design guidance. It will be some time until transonic

drag predictions are accurate and reliRble enough fur detailed design data. 4 Supersonic Drag Somewhat contrary to the state of transonic drag prediction. supersonic dr2g is firmly based on theoretical a~d numerical methods. In the early

19705, the supersonic transport and the 8-70 programs provided the impetus for the extensive development and numerical implementation of linearized superThe dedicated, pioneering efforts of BonnerIO, 11, 12, Carlson2I, Carmichae125) Harris54, 55, Middleton97• sonic theory that had begun in the 1960:;.

36

Woodward 165 ,and others hctve produced automated numerical design procedures comprised of many individual computer codes under the control of an executive coupling routine. These numerical procedures are continually evolving to

reduce the shortcomings uncovered during the succeeding years of intensive application to fighter, bomber and transport configurations. The drag pre-

dictions produced by the current versions of these programs are excellent for many slender supersonic configurations at low lift coefficients and are widely

used throughout the aircr(lft industry for preliminary design. As shown in Fig. Sa, the supersonic cruise drag components are dramatically different from the subsonic cruise drag components. The large, obvious in

differenc!; i~ the wave drag, a component that does not have a counterpart subsonic flow. The lift-dependent and skin friction drag components are

analogous to their subsonic flow counterparts and are calculated by the following modifications to the previous subsonic methods. The supersonic drag

buildup, using linearized theory, is illustrated in Fig. 28. 4.1 Friction Dr~g The effects of compressibility supersunic speeds. and heat transfer must be accounted for at
by

The mean adiabatic skin friction coefficient decreases

Jb0ut 30%, regardless of Reynolds number, in changing from incompressible speeds to Mach two, as shown in Fig. 29. The relative predictions from the several

four theories vary with both Reynolds and t'lach numbers, illustrating drag prediction problems.

Certainly, the basic full-scale skin friction drag The corre-

pred ic t i on during initial design depends on the theory selected.

lation from wind tunnel model scale to full-scale flight, as illustrated by :he Concorde81 data, depends on the theory selected. Finally, the reduction uf flight test results, in order to define actual drag as a function of Mach number, altitude, ambient temperature and 1ift coefficient, is also theory

37

dependent.

The magnitude of the drag change produced by selection of altPfLeyman and Markham 81 have shown

nate prediction methods is, however, small.

that compared with the later method of Winter and Gaudet, Michel's method predicted lower aircraft drag directly by 2-1/25(, C; (0.8% total drag) and
r

lower aircraft drag from extrapolation of the 1/45th scale model data by 5% C (1.5% total drag) for the Concorde. The Winter and Gaudet method is similar to that of Sommer and Short136, and each reproduced available, acceptable, experimental results more accurately than Michel's method. However, the

scatter in all data precluded changing skin friction prediction methods during the Concorde design process. The same theory should be used for a11 estimates

during the design process to achieve consistent results. The Sommer and Short method136 is used to calculate the compressible skin friction coefficient, CF, from a reference skin friction coefficient, C~, for a selected free stream Mach number, M , and Reynolds number, R , and adiabatic
00 00

wall temperature, Tw' using the .incompressible Karman-Schoenherr (Sect i on 2.2).

formula

The details of the theory and original experimental veriThis method has subsequently been exten-

flcation are given in Reference 136.

sively tested against newer experimental data and found to be accurate by 112 Peterson , Sorrells, Jackson and Czarnecki137, Stallings and Lamb138, and 141 Tendeland for a wide variety of geometries, flow conditions, and heat transfer rates. System-' . Cons istent use of the Sommer and Short method from the pre 1iminery des i9r: phase through flight testing will improve drag correlations. Note that drag
qi

The method has been programmed for many computers and forms

the s~in friction module of the Integl'ated Supersonic DeSign and Analysis

component correlations used during the preliminary design phase are, ultimately, based on older wind tunnel and flight test data. A considerable

38

amount of these data were reduced using the Sommer and Short method as a basis. It is consistent to continue use of this method for subsequent dra£

buildups on new configurations. 4.2 Wave Drag The wave drag produced by standing pressure waves that are not possible in subsonic flow has been traditionally decomposed into two components5: zero-1ift wave drag and a wave drag-due-to-lift. The decomposition is a
a

result of the supersonic area rule method used for wave drag prediction and ;s not permitted if nonlinear methods are used. Zero Lift: Lift-dependent or induced: Do wave These drag components vary as: (volume)2/(length)4

(M2-1)(lift)2/(lifting length)2 Di wave and must be added to the supersonic counterparts of the subsonic skin friction drag and induced or vortex drag. Thus, the supersonic design problem of

obtaining maximum lift-to-drag ratio is more complicated due to constraints ariSing from these additional drag terms. Linear supersonic aerodynamics methods are the mainstay of the aircraft industry and are routinely used for preliminary design because of their simplicity and versatility in spite of their limitations to slender configurations at low lift coefficients. Not surprisingly, most successful supersonic

designs to date have adhered to the theoretical and geometrical limitations of these analysis methods. Linearized theory has not been supplanted since

second-order theories have not been as successful at predicting experimental data, and practical application of nonlinear methods has been precluded by sevpre geometrical restrictions, i.e. ax;syrrrnetricbodies, conical flow, etc. The normal pressure drag components, consisting of form drag, vortex drag, lift-dependent wave drag and volume-dependent wave drag, are shown in
Fi g. b.

39

The super sontr

form

d:-d9.

o r drag

due

to adding the boundary-layer

displacement thickness to the p~ysical areas to obtain the effect of the

on the wave drag, is convent i ona l Iy ignored19, 81. is small (±1%) and dependent
upon

Calculations have shown this ccntribution longitudinal area distribution

for bodies of revolution19.

This assumption is the body alone55,

within the accuracy of the prediction capability

67, particularly for low fineness ratio bodies80• assumption is unknown for wings.

The accuracy of this

As shown in Fig. 28, both near- and far-fleld methods are used to calculate the components of inviscid pressure drag, Fig. 6.
A

surface panel method, such as PAN Jl.IR39 , that represents the confi g_ of

uration's thickness, camber and angle of attack with surface distributions sources
and

doublets and sums the resulting pressure times the streamwise

component of area (near-field method) obtains the sum of the vortex drag, lift-dependent and volume-dependent wave drags. The use of these methods but do not

offer improved geometric representation permit independent consideration purposes.

of the configuration,

of each drag component for optimization

Far-fiel~ methods relate the wave drag to the lateral convection of strearnwise momentum through ~ cylindrical largE distance (several wingspans) control surface, Fig. 28, placed a The vortex drag is

from the configuration.

similarly related to the transverse components of momentum through a crossflow plane far downstream in the wake the Trefttz plane. The supersonic vortex

drag is identical to the induced drag for subsonic flow since the trailing vorticity remains essentially speed. stationary with the fluid, regardless of flight limitations of each approach, a modifica97

Due to the inherent theoretical

comb i na t i on of near-

far-fielG me thods , along with semiempirical

d tlons, are current 1y use·

arag

~. preG1C t 10n

an d

d eSlgn -

op t i - t 10n 5, lmlza -

40

The zero-lift, or volume-dependent,

wave drag is commonly computed using

a nume r i cal implementation of the Supersonic Area RIlle concept, proposed somewhat independently
by

Hayes, Jones, and Whitcomb.

This far-field method

relates the zero-lift wave drag of a wing. fuselage, empennage, nacelles, etc., to a number of developments of the normal components of cross-sectional areas as intersected by inclined Mach planes through the Von KArmAn slenderbody formula (FiS. 28). The use of linear source-sink distributions to

represent the configurat~on precludes representation of both local lift and total integrated lift. The wave drag due to lift and the vortex drag are,

perfurce , calculated by another method, conventionally the near-field t1ach Box method
fOI'

thin (zero thickness) wings.

~Jith an additional restriction on the slenderness of the configuration, the Transfer Area Rule can be obtained that permits optimization of the wing-body for ~inimum wave drag due to volume, independent of the wing. Since

the fuselages of military aircraft are usually not Slender, the use of the Transfer !\rea Rule in this case would represent a greater violation of theory than use of the Supersonic Area Rule. Numerous comparisons of this theory with experimental data from the slender X8-70 and Supersonic Transport models5, 10 have demonstrated that it is remarkab Iy accurate (less than Mach numbers from 1.2 to 3.0. BonnerlO, using an integral similar to Von KArman's slender body formula derived
by ±

10% error in total drag) over a range of

Hayes, obtains both the zero-lift and lift-dependent wave drag.

This procedure, however, requires a priori knowledge of the longitudinal distribution of lift. The use of this procedure is relatively limited since a

41

near-field method capable of obtaining both the vortex and lift-dependent wave drags must be used to supply the required lift distribution and lifting surfaces are much less likely to satisfy the slenderness requirements. The

distant point of view does, however, incorporate lift-volume interference not reflected in many of the theoretical techniques used to estimate the surface pressure.
4.3

Lift-Induced Drag The drag due to lift, both wave and vortex, has conventionally been

calculated using the Mach Box method of Middleton and Carlson96, currently with many semienlpirical modifications to alleviate known shortcomings97• A

comprehensive review of this method, as of 1974, is contained in Ref. 23 and later modifications are contained in Ref. 21 ana Ref. 97. The method sub-

divides the zero-thickness

planform, often with the fuselage outline (Fig.

28), into rectangular panels whose diagonals are aligned with the free-stream

Mach lines of the flow.

Alternatively,

the sides of the element could be Box method

parallel to the Mach lines and the Mach Diamond or Characteristic ~ould result.

Generally, the Mach Box method is more accurate for supersonic

edges, while the Characteristic

The linear theory integral equation relating camber surface slope to lifting pressure difference, ACp' across the planform is numerically evaluated
bl.J
l

considering

AC P a constant within each box or element and reducina_ the

integral equation to an ulgebraic summation.

summation can be solved for (wing design

either the camber surface slope for a known pressure distribution case) or the lifting pressure coefficient slopes (lift analysis case).

in terms of known camber surface

A number of component loadings (up to seventeen)

can be combined in the design mode to determine optimum camber shapes for minimum

42

drag due to lift at a given 1ift coefficient, moment constraints.

subject to optional pitching

The effects of the fuselage and nacelles may be included

in the component loadings and additional constraints may be applied to the d~sign pressure distribution realism. and local camber surface to provide physical overestimates expansion pressures (to the compression pressure114•

Linear theory consistently

less than vacuum pressure!) and underestimates

Proper selection of the component loadings and constraints requires complete understanding of the computer programs backed by critical operational use, otherwise, occasional ~endless loops" may result97• There may also be small

discrerancies between wing loadings and forces determined for an optimized wing dnd the leadings and forces for the same shape upon submittal to the evaluation program23• The preceding supersonic aerodynamic design and analysis methods have been substantiated for numerous wings, wing/body and wing/body/nacelle urations at the lower lift coefficients t;ons. A comparison with wind-tunnel config-

typical of supersonic cruise condidata from the Lightweight Experimental

Supercruise Model (LES) at the design Mach number and lower lift coefficients is shown in FiS. 30. Test-theory comparisons for supersonic transport and

bomber models at typically higher cruise Mach numbers and lower design lift coefficients are equally as accurate10, 116, although only if performed by experienced designers familiar with the methods, Fig. 31. The major unresolved drag prediction problem of thin, highly swept wings typical of efficient supersonic fli9ht is the evaluation of leading-edge thrust or leading-edge suction. The large influence on the drag polar of the

various assumptions used to determine the magnitude of this force is shown in Fig. 32 for the same configuration as shown in Fig. 30. The polar shown

43

in Fig. 32 is for M = 1.35, an off-design condition for the M = 1.8 supercruiser configuration. The leading edge thrust results from the low pressure induced by the hir,r flow velocities around the leading edge from the stagnation pOint on the undersurface of the wing to the upper surface. For the high aspect ratio

wings typical of low subsontc flight speeds, this force largely counteracts the drag from the pressure forces acting on the remainder of the airfoi122,116. The thrust force diminishes with increasing speeds, but exists as long as the leading-edges are subsonicl63• It was found to be negligible at M = 3.0 for the configuration conventionally analyzed by Kulfan76• Thus, thrust effects have

been ignored for supersonic wing design and analysis at Mach numbers approaching threel16, with excellent results, as shown in Fig. 31. The renewed interest in this phenomena resulted from the increased

sophistication

in supersonic wing design required to achieve higher design

lift coefficients for the lower Mach numbers typical of proposed fighter deSigns with sustained supersonic cruise capabi1ityl64. The maneuver require-

ments also demand higher available lift coefficients and the attendant thickness and complex camber shapes to minimize drag. The current complex of NASA computer programs'97 contains four user selected options to estimate leading-edge thrust, Fig. 32. The basic Mach Box

zero-thickness wing design and analysis method provides the zero leading-edge suction drag estimate since the low pressures do not have a forward facing area to act upon. The Polhamus analogy assumes the flow is separated from the

entire leading-edge and has formed the spiral vortex sheets characteristic of slender wings at higher angles of attack76. The Polhamus analogy has been extended, by Kulfan76, to account for airfoil shape, wing warp and

44

planform effects in order to determine the influence of wing geometry on leading-edge vortices. He demonstrates that, except for very slender wings,

lowest drag due to iift is achieved with attached flow. The attainable thrust option (Fig. 32), developed by Carlson. Mack and Barger22, has been found to work well for wings with standard NASA airfoil sections, from which the correlations were developed163• However, the attainmeasured in both

able thrust forces predicted did not agree with those experimentally on wings with sharp and varying leading-edge levels and trends were eVident163. radii; large differences

The magnitude of the full leading-edge thrust is dependent upon the upwash in thF vicinity of the leading edge and the effect on the pressure dist.ribution. coefficient It is determined by calculating the limit of the pressure

(wh ich theoretica1ly

approaches infinity) multiplied by the The evaluation of this limit is difficult as calculated by panel methods, vary from This limit problem and the resulting studied by

distance from the leading edge.

b£cause the pressure distributions, the exact theoretical

values required.

impact on wing design and drag prediction have been extensively Carlson and Mack21. Calculations using near-field surface-panel

methods should predict the integral is sensitive to

full suction drag polar; however, the pressure-area leading-edge paneling density.

Near thin wing tips, the paneling may be

i~sufficient for the accurate evaluation of this integral and the resulting force will be bounded between the full and zero leading-edge The force will thus depend on the numerical representation than aerodynamic assumptions, a disconcerting suction values. of the wing, rather

situation for the designer.

45

As observed by several investigators, a reasonable estimate for the drag polar could be obtained by numerically averaging the zero and full leadingedge suction drag polars. Drag po1ars calculated by two near-field methods for Mach numbers of 1.196, 1.6, and 2.2 are shown in Fig. 3398. The complex canard-wing-fuse1age configuration and paneling arrangement are shown in Fig. 21. The skin friction drag was estimated using the methods of Section 4.1. Considerable

scatter is evident in the zero lift drag which may be attributed to either the skin friction estimate or, more likely, the zero lift wave drag predictions of the near-field methods. The variation in vortex and lift-dependent wave drag

with angle of attack was correctly predicted by the higher-order

panel method,

PAN AIR26.

The lower-order method, USSAERO-B, has recently been updated to

improve supersonic prediction capability by inclusion of the Triplet Singularity41, was 1.8. 165 The wing camber design Mach number for this configuration

Many additional test-theory comparisons. exhibiting much closer

agreement, are contained in Tinoco, Johnson, and Freeman144 and Tinoco and l46 Rubbert . The prediction methods included a pilot version of PAN AIR, the Mach Box method and the Woodward Constant-Pressure
a Mach three Reece-Strike

panel method (USSAERO) for shown in

design and the supercruise configuration

Figs. 30 and 32. Correction procedures to remedy the observed shortcomings of linear theory are under continuous development. Stancil139 has modified near-field

linear theory to improve the prediction of compression and expansion pressures and to eliminate local singularities. His modifications, the use of exact

tangency surface boundary conditi ons and the cor-rected 1oca 1 value of t1jach number, cause the solution technique to become iterative. The modifications

improve the zero-lift wing wave drag prediction capability of linear theory

46

for the F_8C13~.

Shrout and Covell129 compared both Stancil's modified linear

theory and the far-field slender-body theory zero lift wave drag predictions with their experimental data for a series of forebody models having various degrees of nose droop. The far-field method correctly predicted trends for

Mach numbers of 1.2 and 1.47, but was erroneous at 1.8 for the cambered n~dels. Stancil's method139 correctly predicted wave drag trends for Mach

numbers of 1.47, 1.80 and 2.16. Another breakdown of linear theory occurs when the crossflow velocities exceed sonic value, a situation encountered on fighter aircraft at maneuver
1 ift

co€fficients.

The formation of crossflow shock waves causes the experi-

mental drag-dl,;2-to-lift performance of wings designed by linear theory to fail 91 3 to meet predicted values. The supercritical conical camber concept ., (SC ), has been proposed to regain wing performance for wings swept such that the Mach nu~ber normal to the leading edge is transonic. The design method

employs repetitive application of a transonic full potential flow solver, COREL, and a specially adapted version of the Woodward USSAERO linear theory
. pane 1 1ng co d e 92 .

Comparison of this design procedure with data from the wind

tunnel tests of a demonstration wing show that substantial reductions in drag-due-to-lift are possible, Fig. 34, at the maneuver design condition91• Bonner and Gingrich12 approach fighter wing design through the use of va~iable camber and a design compromise between the transonic maneuver configuration and an optimum supersonic cruise design. The unconstrained supersonic

design was initially accomplished by conventional linear theory while both small-disturbance design. and full-potential theories were used for the transonic

These designs established the boundaries for the design cycle.

Experience then guided the selection of the compromise camber shape, followed by alternating constrained optimizations at the supersonic and transonic

47

design points.

This mu lt i po i nt design process illustrates the need f or Spurious high drag regions entirely due to the use

accurate drag prediction.

of linear theory, shown in Fig. 35 at the sonic leading-edge condition, could eliminate the optimum design from consideration on the first cycle!

Significant potential for further improvement in wing design for minimum exists through the use of nonlinear computational methods early in the design cycle46. A comprehensive set of design constraints for efficient, highly swept w i nqs has been determined by Kulfan77. The problems of fighter design and analysis are further compounded by the fact that linear theory does not apply to relatively large fuselages with nonsmooth area distributions, as shown by Wood, et.al.164 Four wings were drag-due-to-lift

df:!signedand tested with a common fuselage in an effort to adapt supersonic technologies to a ~i = 1.8 fighter design. Their test-theory comparisons, Fig.

36. clearly show the inaccuracy in zero-lift drag prediction using the Middleton and Lundy97 complex of computer programs. This fuselage severely

violates the assumptions of linear theory concerning smooth, slender area distribution bodies. Further inaccuracies as the configuration is built up the

are oPPosite in sign and cancel most of the fuselage error -- illustrating somewhat forgiving nature of linear aerodynamics. son for each of the four wings is shown
ill

The zero-lift drag compariThe prediction capaand

Fig. 37.

bilities of linear theory are excellent for the complete configurations vary with slenderness and smoothness of the area distribution. to-lift comparisons were equally
6S

The drag-due-

accurate; however, they were performed

using in-house empirical corrections. Linear theory, as developed by ~1iddleton and Lurldy97, has become the Correction proce-

standard preliminary design tool for supersonic aircraft. dures tu allev i ate
t he

we lI-known res tr+ct tons of the theory are under

48

constant deve 1opment . Higher-order panel methods, necessary for supersonic flow analyses, along with nonlinear numerical methods, are also used during the latter design stages, particularly fighter configurations. An innovative, statistical method for the prediction of fighter aircraft drag-due-to-lift was developed by Simon, et.al.131, for Mach numbers of 0.4 for lower design Mach number, blunt

through 2.5. A 1arge experimental data base was fitted with regression equations based on easily calculated nondimensional geometric parameters. The

accuracy of the resulting regression equations was only limited by the expanse of the data base from whid
. J48 Tomase t t t ' .

they were obtained, as shown by Johnson69 and

5 Numeri ca 1 Aerodynanrics Here, the term numeri cal aerodynami cs is used to de 1 i neate the fu 11 spectrum of aerodynamic methods requiring large computers -- from the linearized theory panel methods to the full Navier-Stokes computation~l equations. The term

fluid dynamics (eFn) is often used in this context, and just as computations, particularly with

often limited to mean only Navier-Stokes respect to future goals.

Numerical aerodynamics now permeates nearly every

stage of the design process~ as well as each speed regime. Many prognostications
and

concerning the possible roles of, primarily, CFD

the wind tunnel have been published (Ref. 27, 29, 30, 42, 48, 53, 73, 79,

83, 99, 110, 115, 121. 124, 125, 132~ 167) since Chapman's 1975 art;cl~29 and the prediction that eFO would supplant traditional wind tunnel tests within a decade. speed. The pacing itenl in eFD technology at that time was computer size and A later article30 recognized turbulence models to be a pacing item.

In addition to the lans-term problem of turbulence modeling, the problem of

49

proper grid generation is of immediate cencern to those engaged in the use of CFD codes in the production (as apposed to research) mode. Comput~r hardware has advanced tremendously in the last decade79, as have flow simulations using the Reynolds averaged Navier-Stokes equations. Still, the wind tunnel is

the primary tool for design verifica.tion and a large portion of design development work. CFD analysis of full three-dimensional configurations capable of flight may still be a decade away 48, 73• The fundamental limitation of CFD during the last decade -- turbulence modeling -- remains the pacing item.
117, 133

Progress in this area in the next 15 years is not projected to

remove this limitation73 and recent reviews of turbulence models52, 82, 87,
. re1n f oree th 1S con tent i _i entl0n.

The fundamental physics of turbulence is

not known well enough to permit accurate mathematical modeling and flow simulation. If the present is any indication of the future, drag prediction

and design optimization with drag as an objective function using CFD are even further in the future. Recent publications recognize the complementary roles of wind tunnels and
CF0 73 ,

while the role of the lower-order approximations

(linear, potential,

Euler) is often understated.

Each of these methods will be found in a compleschedule,

mentary place in the design process dependinq on configuration,

cost, availability of hardware and software, and the persuasion of the designer and his management. Confidence in the accuracy of the predictions is 79 tantamount • 99. Aircraft design de~artments should evolve a full spectrum of numerical aerodynamics methods, coupled through a data management system,
. · . to prav1'd e accurate, tlme 1y, cos t -e ff ... deSlgn dat 110,121, TeC~lve aLa 124, 168

50

BIBLIOGRAPHY 1) Abercrombie, J.M., "Flight Test Verification of F-15 Performance Predictions," AGARD-CP-242, 1977, pp. 17-1 - 17-13. 2) Aidala, P.V., Davis, W.H., and Mason, W.H., "Smart Aerodynamic Optimization," AIAA-83-1863, Applied Aerodynamic Conference, Danvers, MA, July 1983. 3) Arnaiz. H.H., "Flight-Measured Lift and Drag Characteristics of a Large, Flexible, High Supersonic Cruise Airplane," NASA TM X-3532, May 1977. 4) Axelson, John A., "AEROX - Computer Program for Transonic Aircraft Aerodynamics to High Angles of Attack," Vol I, II, and III, NASA TM X-73,208, February 1977. 5) Baals, D.O., Robins, A.W., and Harris, R.V., Jr., "Aerodynamic Design Integration of Supersonic Aircraft," J. Aircraft, September-October 1970. 6) Baldwin, A.IL, Editor, "Symposium on Transonic Aircraft Technology (TACT)," AFFDL-TH-78-100. August 1978.
7) Bills. G.R., Hink, G.R., and Dornfeld, G.M., "A Method for Predicting the Stability Characteristics of an Elastic Airplane," AFFDL-TR-77-55, 1977.

8) Blackwell, J.A., Jr., "Preliminary Study of Effect of Reynolds Number and Boundary Layer Transition Location on Shock Induced Separation,lI NASA TN 0-5003, 1969. 9) Blasius, H., "Boundary Layers in Fluids of Small Viscosity," Dissertation, Gottinger, 1907. 10) Bonner, E., "Expanding Role of Potential Theory in Supersonic Aircraft Design," J. Aircraft, May 1971, pp. 347-353. 11) Bonner, E., Clever, W., and Dunn, K., "Aerodynamic Preliminary Analysis II Part I - Theory," NASA CR 165627, Apri 1, 1981. 12) Bonner, E., and Gingrich, P., "Supersonic Cruise/Transonic Section Development Study," AFWAL-TR-80-3047, June 1980. Maneuver Wing

13) Bowes, G.t·l.,"Aircraft Lift and Drag Prediction and ~1easurement," ACARD-LS-G7, May 1974, pp. 4-1 - 4-44. 14) Braslow, A.L., "Use of Grit-Type Boundary-Layer-Transition Tunnel Models," NASA TN D-3579, September 1966. Trips on Wind

15) Braslow, A.L., and Knox, E.C., "Simplified Method for Determination of Critical Height of Distributed Roughness Particles for Boundary Layer Transition at Mach Numbers from 0 to 5," NACA TN-4363, 1958.

Sl

lSi Br i s tow, D.R., and Hawk , Li.n., "Subsonic Panel ~';ethodfor the Efficient Ana lys i s of Multiple Geometry F'erturbations," NASP. CP.-3528, March 1982.

In Brown, S.W., and Bradley, D., "Uncertainty Ana lys i s for Flight Test Performance Calculations," AIAA-Bl-2390, AIAAjSETP/SAE Flight Testing Conf., Las Vegas, NV, November 1981.
18) Buckner, J.K., and t~ebb. J.B., Tunnel Test Program," J\IAA-74-619, July 1974.
8th

"Selected Results from the YF-IG Wind Aerodynamic Testing Conf • , Bethesda

MA,

19) Butler, S.F.J., "Aircraft Drag Prediction for Project Appraisal and Performance Estimation," AGARD-CP-124, 1973, pp. 6-1 - 6-50, with Appendix.

20) Butler, S.F.J., vii-xv. 21)

"Technical Evaluation Report," AGJl.RD-CP-124, 1973, pp.

Carlson, H.W., and r~ack, R.J., "Estimation of Wing Nonlinear Aerodynamic Characteristics at Supersonic Speeds," NASA TP-1718, November 1980.

22) Carlson, H.W., Mack, R.J., and Barger, R.L., "Estimation of Attainable ~eading-Edge Thrust for Wings at Subsonic and Supersonic Speeds,~ NASA TP 1500, October, 1979.
Ane lys i s

23)

Carlson, H.W., and Miller. D.S., "Numerical Methods for the Design and of vJings at Super son i c Speed," NASA TN 0-7713, December 1974.

24) Carlson, H.W., and Walkley, K.B., "A Computer Program for Wing Subsonic Aerodynamic Performance Estimates Including Attainable Thrust and Vortex Lift Effect, " NASA CR-3S15. ~'aY'ch 1982.

25) Carmichael, R.L .• and Woodward, F.A., "An Integrated Approach to the Analysis and Design of Wings and Wing-Body Combinations in Supersonic Flow," NASA TN D-3685, October 1966.
26) Carmichael. R.L .• and Erickson, L.L., "PAN AIR: A Higher-Order Panel Method for Predicting Subsonic or Supersonic Linear Potential Flows About Arb+trary Configurations." AI.~!~-81-1255, June 1981.
27)
FDA.-80- 07,

1980.

Caughey, D.A., "A (Limited) Perspective on Computational Aerodynamics," Fl u i d Dvnamt cs _u!lel Aerodynami cs Program - Corne 11 Un ivers ity, Ju 1Y
IIPAr~ AIR Applications to Complex Configurations," l\.erospJce Sciences ~;eet1nS'., eno NV, January 1983. R

28)

AIPJl.-83-0007.

Cenko, A.,

Chapman, D.R .• Mark, H" and Pirtle, M.W., "Computers vs. Wind Tunnels Fl ow Simulations." Astronautics and Aeronautics. April 1975, pp. 21-30.
for Aercoynamtc

29)

30) Chapman, D.R., "Comput at i ona l Aerndynami cs Development and Outlook," 1\IAA-79-0129R, 17th Aeros[_9ce Sc i enccs ~1eeti~, January 1979.

52

31) Chen, A.W., and Tinoco, E.N., "PAN AIR Applications to Aero-Propulsion Integration," AIAA-83-1368, 19th Joint Propulsion Conference, Seattle WA, June 1983. Coakley, T.J., "Turbulence Modeling Methods for the Compressible Navier-Stokes Equations," AIAA-83-1693, 16th Fluid and Plasma Dynamics Conf., Danvers MA, July 1983.
32)

33) Cooper, J.M., Hughes, D.L., and Rawlings, K., "Transonic Aircraft Technology - Flight-Derived Lift and Drag Characteristics,1I AFFTC-TR-12, July
1977 .

34) Craig, R.E., and Reich, R.J., "Fl iqht Test Aerodynamic Drag Characteristics Development and Assessment of Inflight Propulsion Analysis Methods for AGM-I09 Cruise Missile," AIAA-81-2423) AIAA/SETP/SFTE/SAE/ITEA/IEEE 1st Flight Testing Conf., Las Vegas NV, November 1981.
35) 36) Crosthwait, E.L., Kennon, I.G., Jr., and Roland, H.L., Design ~1ethodology for Air-Induction Systems," SEG-TR-67-1,

ETAL "Pre liminary January 1967.

Dauqhe rrv , J.C., "Wind Tunnel/Flight Correlation Study of AerodynamiC Characteristlcs of a Large Flexible Supersonic Cruise Airplane (SB-70-1)," NASA TP-1514, November 1979.

37) DeLaurier, J" "Drag of Wings wi th Cambered Airfoils and Partial Landing-Edge Suction,lI J. Aircraft, October 1983, pp. 882-886.

38) Dvorak, F.A., Woodward, F.A., and Maskew, B., "A Three-Dimensional
Viscous/Potential Flow Interaction Analysis Method for Multi-Element NASA CR-152012, 1977.

Wings,"

Erickson, L.L., and Strande, S.M., "PAN AIR Evolution and New Directions • . . Transonics for Arbitrary Configurations." AlAA-83-1831, Applied Aerodynamics Conf., Danvers MA, July 1983.
39) 40) Firmin, M.C,P., and Potter, J.L., IIWind Tunnel Compatibility Test Sections, Cryogenics, and Computer-Wind Tunnel Integration," AGARD-AR-171, April 1982.

Related to

~1) Fornasier, L., "Calculation of Supersonic Flow Over Realistic Configurations by an Updated Low-Order Panel Method,~ AIAA-83-0010, 21st Aerospace Sciences t·1eeting,January 1983.
42) "Future Computer Requirements for Computational Aerodynamics," Workshop Series -- Mixed Papers, NASA CP-2032, October 1977.

NASA

Characteristic Aircraft Excrescences AGARD-CP-124. 1973, pp. 4-1 - 4-12.

~3) Gaudet, L., and Winter, K.G., "Measurements of the Drag of Some

Immersed in Turbulent Boundary Layers,"

44) Geddes, J.P., "Taking an Airliner from Certification to Airline Acceptance - the DC-10-30," INTERAVIA, June 1973, pp. 609-611.

53

45) Gilmer, B.R., and Bristow, D.R., "Analysis of Stalled Airfoils by Simultaneous Perturbations to Viscous and Inviscid Equations," AIAA-81-023, 14th Fluid and Plasma Dynamics Conference, Palo Alto CA, June 1981. 46) Gingrich, P.B., and Bonner, E., "Wing Design for Supersonic Cruise/Transonic Maneuver Aircraft," ICAS-82-5.8.2, 13th Congress of the International Council of the Aeronautical Scie~ces, Seattle WA, August 1982. 47) Glauert, H., The Elements of Aerofoi1 and Airscrew Theory, 2nd ed., Cambridge University Press, New York, 1959. 48) Graves, Randolph A., "Computat ione l Fluid Dynsmi cs - ThE: Coming Revolution," Astronautics and Aeronautics, March 1982, pp. 20-28.
49) Grellmann, H.W., IIYF-17 Full Scale Minimum Drag Predi ct ion," AGARD-CP-242, 1977, pp. 18-1 - 18-12. 50) Hahn, M., Brune, G.W., Rubbert, P.E., and Nark, T.C., "Drag t~easurements of Upswept Afterbodies and Analytical Study on Axisymmetric Separation," AFFDL-TR-73-153, February 1974.

51) Haines, A.B., "Subsonic Aircraft Drag: An Appreciation of Present Standards," The Aeronautical Journal of the Royal Aeronautical Society, Vol. 72, March 1968, pp. 253-266. 52) Haines, A.B •• IITurbulence Modelling," Aeronautical Journal, August/September 1982, pp. 269-277. 53) Hall, M.G., "Compute t ione l Fluid Dynamics - A Revolutionary Aerodynamics,1I AlAA-81-1014, 1981. 54) Harris, R.V., TMX-947, 1964. Force in

"An Analysis and Correlation of Aircraft Wave Drag," NASA

55) Harris, R.V., and Landrum, E.J., "Drag Characteristics of a Series of Low-Drag Bodies of Revolution at Mach Numbers from 0~6 to 4.0," NASA TN D-3163, December 1965. 56) Henne, P.A., "Inverse Transonic Wing Design Method," J. Aircraft, February 1981. 57) Henne, P.A., Dahlin, J.A., and Peavey, C.C., "Applied Computational Transonics - Capabilities and Limitations," Douqlas Paper 7025, Transonic Perspective Symposium NASA Ames, February 1981. 58) Hess, J.L., "A Higher Order Panel Nethod for Three-Dimensional Potential Flow." Naval Air Development Center Report No. NADC 77166-30, June 1979. 59) Hess, J.L. and Friedman, D.M., "An Improved Higher Order Panel Method for Three-Dimensional Lifting Potential Flow," Naval Air Development Center Final Report, NADC -79277-60, December 1981. 60) Hicks, R.M., "Transonic Wing Design Using Potential-Flow Codes -Successes and Failures," 810565, Society of Automotive Engineers, Inc., 1981.

54

61)

"High Angle of Attack Aerodynamics,"

AGARD-CP-124.

1979.

62) Hitzel, S.M., Schmidt, W., "Slender Wings with Leading-Edge Vortex Separation -- A Challenge for Panel-Methods and Euler-Codes," AIAA-83-0562, 21st Aerospace Sciences Meeting. Reno NV. January 1983. 63) 64) Hoak, D.E., USAF Stability and Control DATCOM, Revised. April 1978. Hoerner, S.F., Fluid-Dynamic Drag, published by the author, 1965.

65) Hopps, R.H .• and Danforth, E.C.B., "Correlation of Wind Tunnel And Flight Test Data for the Lockheed L-I0l1 Tristar Airplane," AGARD-CP-242, 1977, pp. 21-1 - 21-12. 66) Hunt, B.L., Gowadia, N.S., "Determination of Thrust and Throttl e-Dependent Drag for Fighter Ai rcraft, AIAA-81-1692, Ai rcraft Systems and Technology Conference, Dayton OH, August 1981.
II

67) Jackson, C.N., Jr., and Smith, R.S., "A Method for Determining the Total Drag of a Pointed Body of Revolution in Supersonic Flow with Turbulent Boundary Layer," NASA TN D-5046, March 1969. 68) James, R.M., "An Investigation of Induced Drag for Field Calculation Methods for Planar Distributions," MDe J7046, December 1975. 69) Johnson, J. Kenneth, "Accuracies and Limitations of a Statistical Method for Computing Lift and Drag Due to Lift of Fighter Type Wing-Body Configurations," AFFDL/FXM-TM-73-46, August 1973. 70) Kjelgaard, S.O •• "Evaluation of a Surface Panel Method Coupled with Several Boundary Layer Analyses," AIAA-83-0011, 21st Aerospace Sciences Meeting, Reno NV, January 1983. 71) Knaus, A., "A Technique to Determine Lift and Drag Polars in Flight," ~ Aircraft. Vol. 20, July 1983, pp. 587-593. 72) Kolesar, C.E., and May, F.W., "An Aftbody Drag Prediction Technique for Military Airlifters," AIAA-83-1787, Applied Aerodynamics Conf., Danvers MA, July 1983. 73) Ko rkeq'i, R.H., "The Impact of CFO on Development Test Facilities _ A National Research Council Projection," AlAA-83-1764, 16th Fluid and Plasma Dynamics Conf., Danvers MA, July 1983. 74) Kuhlman, J.M., and Ku, T.J., "Numerical Optimization Techniques for Bound Circulation Distribution for Minimum Induced Drag of Nonplanar Wings: Computer Program Documentation," NASA CR-3458, August 1981. 75) Kulfan, R.M., "Application of Hypersonic Favorable Aerodynamic Interference Concepts to Supersonic Aircraft," AIAA-78-1458, Aircraft Systems and Technology Conference, August 1978. 76) Kulfan, R.M., IIHing Geometry Effects on Leading-Edge Vorticies,1I AIAA-79-1872 Aircraft Systems and Technology Meeting, NY, August 1979.

55

77) Kulfan, R.M., and Sigaila, A .• "Real Flow Limitations in Supe rson i c I\jr~,laneDesign," AIAP.-78-147, January 1978. Kulfan, R.M., Yoshihara, H., Lord, B.J., and Friebel, G.O., "Application c Favorable Aerodynamic Interference to Fighter Type Aircra~t," AF~DL-TR-78-33, April 1978.
78)
of Supe r-snnf

79)

Kutler, P., IIAPerspective of Theoretical and Applied Computational Fluid Dynamics," AIAA-B3-0037, 21st P,erospace Sciences Meeting, Reno NV, January 1983.

80) Landrum, E.J., and Miller, D.S., "Assessment of Analytic Methods for the Prediction of Supersonic Flow over Bodies," AIAA-BO-0071R, AIAA Journal, February 1981.
81) Leyman, C.S. and Markham, T., "Prediction of Supersonic Aircraft Aerodynamic Characteristics~" AGARD-LS-67, May 1974, pp. 5-1 - 5-52.

82) Liepmann, H.W., liThe Rise and Fall of Ideas in Turbulence," Americdn Scientist, Vol. 67. March-April 1979, pp. 221-228. 83) Lomax. H., "Some Prospects for the Future of Computational Fluid Dynamics," AIAA-81-0991R. AlAA Journal, Vol. 20, No.8, August 1982.
81l) 85)

Longo, J., Schmidt, W .• and Jameson, A., "Viscous Transonic Airfoil Flow Simulation," Z. FlugVliss Wettraumforsh.7, 1983, Heft 1.
l.o tz , ~1., and Friedel, H., "JI.nalysisof Error Sources in Predicted Flight Performance," Performance Prediction Methods, AGARD-CP-242, 1977, pp. 14-1 _

14-11.

86)

t·1acHilkinson, D.G., B-Iackerby. W.T., and Paterson, J.H., "Correlation of Full Scale Drag Prediction with Flight Measurements on the C-141A Aircraft _ Phase II, Wind Tunnel Test, Analysis, and Prediction Techniques. Volume I _ Drag Predictions, Wind Tunnel Data Analysis and Correlation," NASA CR-2333, February 1974.

87) tflarvin,J.G., "Turbulence Modeling for Computational Aerodynamics," AIAA-82-0164, 20th Aerospace Sciences Meeting, Orlando FL, January 1982. 88) Maskew, B., IIPrediction of Subsonic Aerodynamic Characteristics for Low-Order Panel Methods, AIJ\A·-81-02S2,Ja.nuary 1981.
II

- A Case

B9) ":aske\'I, B., Rao, B.M., and Dvorak, F.M., "Prediction of Aerodynamic Characteristics for Wings with Extensive Separations," AGARD-CP-291, February 1981, pp. 31-1 - 31-15. 90) Mason, W.H., "Wino-Canard Aerodynamics at Transonic Speeds - Fundamental Considerations on Mini~um Drag Span Loads," AIAA-82-0097, 20th Aerospace Sciences Meeting, January 1982. 91) t~ason. W.H., I'SC3 A Hing Concept for Supersonic Maneuvering," AIAA-83-1858, July 1983.

56

92) Mason, W.H .• and Rosen, B.S., "The COREL and W12SC3 Computer Programs for Supersonic Wing Design and Analysis," NASA CR-3676, 1983. 93) ~icCluney, B. and ~larshall, J., "Drag Development of the Belfast," Aircraft Engineering, October 1967, pp. 33-37. 94) McGreer, T., and Shevell, R.S., "A Method for Estimating the Compressibility Drag of an Airplane," SUOAAR 535, January 1983. 95) McKinney, L. Wayne, Editor, "Wind Tunnel/Flight Correlation - 1981," NASA Miniworkshop Series, NASA Publication 2225, November 1981. 96) Middleton W.O., and Carlson, H.W., "Numerical method of Estimating and Optimizing Supersonic Aerodynamic Characteristics of Arbitrary Planform Wings," J. Aircraft, July-August 1965.
97) Middleton, W.O., and Lunary, J.L., "A System for Aerodynamic Analysis of Supersonic Aircraft," NASA CR-3351, December 1980.

Design and

98) Miller, S.G., and Youngblood, D.B., ~Applications of USSAERO-B and the PAN AIR Production Code to the COAF Model - A Canard/Wing Configuration," AIAfl-83-1829, Applied Aerodynamics Conf., Danvers MA, July 1983. 99) Miranda, L.R., "A Perspective of Computational Aerodynamics from the Viewpoint of Airplane Design Applications," AlAA-82-0018, Orlando FL, January
1982.

100) Miranda, L.R., Elliott, R.D., and Baker, W.M., "A Generalized Vortex Lattice Method for Subsonic and Supersonic Flow Applications," NASA CR-2865, December 1977.
lUI) Morris, S.J., Neums, W.P .• and Bailey. R.O., "A Simplified Analysis of Propu 1sion Insta 11 ati on Losses for Computeri zed Ai rcraft Des ign, II NASA TM X-73,136, August 1976.

102) Murillo, L.E., and McMasters, J.H., "A Method for Predicting Low-Speed Aerodynamic Characteristics of Transport Aircraft," AlAA-83-1845, Applied Aerodynamics Conf., Danvers MA. July 1983. 103) Murman, E.M., and Cole, J.D., "Inviscid Drag at Transonic Speeds," AlAA-74-540, 7th Fluid and Plasma Dynamics Conference, Palo Alto CA, June 1974. 104) Nelson, D.W., Gornstein, R.J., and Dornfeld, G.M., "Prediction of Maneuvering Drag Polars Including Elasticity Effects for AFTI-F-111," AIAA-81-1658, August 1981. 105) Nixon, D., Transonic Aerodynamics, Aeronautics, AIAA, NY, 1982. Vol. 81, Progress in Astronautics and

106) Norton, D.A., "Airplane Drag Prediction," Annals New York Academy of Sciences, 1968, pp. 306-328.

57

107) O'Conner, W.M., "Lift and Drag Prediction in Computer Aided Design," Vols. I and II, ASD/XR-73-B, April 1973. 108) Paterson, J.H., "Scaling Effects on Drag Prediction,~ AGARD-LS-37-10, June 1970, pp. 4-1 - 4-12. 109) Paterson, J.H., MacWilkinson, D.G., and Blackerby, w. T., "A Survey of Drag Prediction Techniques Applicable to Subsonic and Transonic Aircraft Design," AGARD-CP-124, 1973, pp. 1-1 - 1-38. 110) Perrier, P., "Computational Fluid Dynamics Around Complete Aircraft Configurations," leAS Paper 82.6.11, 13th Congress of the International Council of the Aeronautical SCiences, Seattle WA, August 1982. 111) Paynter, G.C., Vaidyanathan, T.S., Maskew, B., and Dvorak, F.A., "Experience with Hybrid Aerodynamic Methods," AlAA-83-1819, Applied Aerodynamics Conference, Danvers MA, July 1983. 112) Peterson, J.B., "A Comparison of Experimental and Theoretical Results for the Compressible Turbulent-Boundary-Layer Skin Friction with Zero Pressure Gradient," NASA TN 0-1795, March 1963. 113) Peterson, J.B., Jr., "Wind Tunnel/Flight XB-70-1," NASA CP-2225, November 1981. Correlation Program on

114) Pittman, J.L., "Preliminary Supersonic Analysis Methods Including High Angle of Attack," AlAA-82-0938, AIAA/ASME 3rd Joint Thermophysics, Fluids, Plasma, and Heat Transfer Conference, St. Louis MO, June 1982, J. Aircraft, Vol. 20, No.9. September 1983. pp. 784-790. 115) Poisson-Quinton, P., "From Wind Tunnel to Flight, the Rule of the Laboratory in Aerospace Design," J. Aircraft, May-June 1968. pp. 193-214. 116) Robins, A.W., Carlson, H.W., and Mack, R.J., "Supersonic Wings with Significant Leading-Edge Thrust at Cruise," NASA TP-1632. April 1980. 117) Rodi, W., "Progress in Turbulence ~lodeling for Incompressible AIAA-81-0045, 19th Aerospace Sciences Meeting, January 1981. Flows,"

118) Rooney, LC., "Development of Techniques to Measure In-Flight Drag of a U.S. Navy Fighter Airplane and Correlation of Flight Measured Drag with Wind Tunnel Data," AGARD-CP-124, October 1973. 119) Rooney, LC., and Craig, R.E., "Development of Techniques and Correlation of Results to Accurately Establish the Lift/Drag Characteristics of an Air Breathing Missile from Analytical Predictions, Subscale and Full Scale Wind Tunnel Tests and Flight Tests," Performance Prediction Methods, AGARD-CP-242, 1977, pp. 16-1 - 16-18. 120) Rooney, E.C., Craig, R.L, and Lauer, R.F., "Correlation of Full Scale Wi nd Tunnel and Fl ight Measured Aerodynami c Drag," AlAA-77 -996, AIAA/SAE 13th Propulsion Conf., Orlando FL. July 1977.

58

121) Rubbert, P.L, and Tinoco, E.N., "Impact of Computational Methods on Aircraft Design," AIAA-83-2060, Atmospheric Flight t~echanics Conf., Gatlinburg TN, August 1983. 122) Schemensky, R.T., "Development of an Empirically Based Computer Program to Predict the Aerodynamic Characteristics of Aircraft," Vols. I and II, AFFDL-TR-73-144,November 1973. 123) Schlichting, H., Boundary-Layer York, 1968. Theory, Sixth Edition, McGraw-Hill, New

124) Schmidt, W., "Advanced Numerical Methods for Analysis and Design in Aircraft Aerodynamics," Int. J. of Vehicle Design, Technological Advances in Vehicle Design Series, SP6, Applications of New Techniques for Analysis and Design of Aircraft Structures. 1983, pp. 2-37. 125) Schmidt, W., Jameson, A., "Euler Solvers as an Analysis Tool for Aircraft Aerodynamics," Recent Advances in Numerical Methods in Fluids, Vol. 4,1983. 126) Sears, W.R., "On Projectiles of Minimum Wave Drag," Quarterly Journal of Applied Mathematics, January 1947, pp. 361-366. 127) Shankar, V. and Goebel, T., "A Numerical Transformation Solution Procedure for Closely Coupled Canard-Wing Transonic Flows," AIAA-83-0502, Aerospace Sciences Mtg, Reno NV, January 1983. 21st

128) Shrout, B.L., "Aerodynamic Characteristics at Mach 2.03 of a Series of Curved-leading-Edge Wings Employing Various Degrees of Twist and Camber," NASA TN D-3827, February 1967. 129) Shrout, B.L and Covell, P.F., "Aerodynamic Characteristics of a Series of Bodies with Variations in Nose Camber," NASA TP-2206, September 1983. 130) Sidwell, K.W., Baruah, P.K., and Bussoletti, J.E., "PAN AIR - A Computer Program for Predicting Subsonic or Supersonic Linear Potential Flows About Arbitrary Configurations Using a Higher Order Panel Method," NASA CR-3252, May 1980. 131) Simon, W.E., Ely, W.L., Niedling, L.G., and Voda, J.J., "Prediction of Aircraft Drag Due to Lift," AFFDL-TR-71-84, July 1971. 132) Smelt, R. The Aeronautical
t

"The Role of Wind Tunnels in Future Aircraft Development," Journal, November 1978, pp. 467-475. Journal 81-4291,

133) Smith, A.M.O., "The Boundary Layer and I," AIM November 1981.

134) Smith, K.G., "Nethods and Charts for Estimating Skin Friction Drag in Wind Tunnnel Tests with Zero Heat Transfer," Technical Note No. AERO 2980, Royal Aircraft Establishment, Ministry of Aviation, London, United Kingdom, August 1964. 135) Snodgrass, R.R., "A Computerized Method for Predicting Drag and Lift for Aeronautical Systems,1I ASD-ENF-TM-76-1, October 1976.
59

136) Sommer, S.C., and Short, B.J., "Free-Flight ~leasurements of 1urbulent8oundary-Layer Skin Friction in the Presence of Severe Aerodynamic Heating at Mach Numbers from 2.8 to 7.0," NACA TN-3391, March 1955. 137) Sorrells, R.B., Jackson, M.W., and Czarnecki, K.R., "Measurement by Wake ~lomer:tlJm Surveys at Mach 1.61 and 2.01 of Turbulent Boundary-Layer Skin Friction on Five Swept Wings," NASA TN D-3764, December 1966. 138) Stallings, R.L., Jr., and Lamb, M., "vJind Tunnel Measurements and Comparison with Flight of the Boundary Layer and Heat Transfer on a Hollow Cylinder at Mach 3," NASA TP 1789, December 1980. 139) Stancil, R.T., "Improved Wave Drag Predictions Using Modified Linear Theory," Journal of Aircraft, January 1979, pp. 41-46. 140) Sytsma, H.S., Hewitt, B.L., and Rubbert, P.E., "A Comparison of Panel ~;ethods for Subsonic Flow Computati on," AGARD-P.G-241, February 1979. 141) Tendeland, T., "Effects of Mach Number and Wall Temperature Ratio on Turbulent Heat Transfer at Mach Numbers from 3 to 5," NACA TN 4236, April
1958.
142) Thomas, J.L and Miller, D.S., "Numerical Comparisons of Panel Methods at Subsonic and Supersonic Speeds," AlAA-79-0404, 17th Aerospace Sciences Meeting, New Orleans LA, January 1979.

143) Thomas, J.L., Luckring, J.M., and Sellers, W.L. III, "Evaluation of Factors De ternrini nq the Accuracy of Linearized Subsonic Panel Methods,1I AlAA-83-1821, Applied Aerodynamics Conference, Danvers MS, July 1983.

144) Tinoco, E.N., Johnson, F.T., and Freeman, L.M., "Application of a Higher Order Panel Method to Realistic Supersonic Conf i qurat i ons ;' AlAA-79-0274R, Jaruary 1980.
145) Tinoco, E.N., and Rubbert, P.E., "Experience, Issues, and Opportunities in Steady Transonics," AERO-B8111-P82-006, Computational Methods in Potential Aerodynamics (leTS) Short Course, Amalf, Italy, May-June 1982. 14G) Tinoco, E.N., and Rubbert, P.E., "Panel ~lethods: PANAIR," AERO-B8111P82005, Computational Methods in Potential Aerodynamics Short Course, Amalf~ Italy, May-June 1982. 2.47) Tjonneland, E., Payrrter, G., and May, F., "Progress Towar-d Airfrarne/ Propulsion Design Analysis," Paper 8, AGARD Flight Mechanics panel Symposium on Sustained Supersonic Cruise/Maneuver, Brussels, Belgium, October 1983.
]48) Tomasetti9 T.A., "Statistical Determination of the Lift and Drag of Fighter Aircraft," GAM/AE/73-1, June 1972.

149) Towne, M.C., et. a1., "PAN AIR Modeling Studies," AIAA-83-1830, Applied A(rodynamics Conf., Danvers MA, July 1983.

60

150) Waggoner, E.G., "Computational Transonic Analysis for a Supercritical Tra nspo rt ~Ii ng-Body Confi gu ration," AIAA-80-0129, 18th Aerospace Sc iences Meeting, Pasadena CA, January 1980. 151) "~Iall Interference in Wind Tunnels," AGARD-CP-335, May 1982.

152) Webb, T.S., Kent, D.R., and Webb, J.B., "Correlation of F-16 Aerodynamics and Performance Predictions with Early Flight Test Results," AGARD-CP-242, 1977, pp. 19-1 - 19-17. 153) White, F.M., and Christoph, G.H., "A Simple New Analysis of Compressible Turbulent Two-Dimensional Skin Friction under Arbitrary Conditions," AFFDL-TR-70-133, February 1971. 154) White, F.M., and Christoph, G.H., "Rapi~ Engineering Calculation of Two-Dimensional Turbulent Skin Friction," AFFDL-TR-72-136, Sup. 1, November 1972. 155) White, F.M .• Lessman, R.C., and Christoph, G.H., "A Simplified Approach to the Analysis of Turbulent Boundary Layers in Two and Three Dimensions," AFFDL-TR-136, November 1972. 156) ~illiams, J., "Ground/Flight AGARD-CP-339, October 1982. Test Techniques and Correlation,"

157) Williams, J., "Technical Evaluation Report on the Flight Mechanics Panel Symposium on Ground/Flight Test Techniques and Correlation," AGARD-AR-191, June 1983. Also AGARD-CP-339, October 1982. 158) Williams, J., "Synthesis of Responses to AGARD-FMP Questionnaire on Prediction Techniques and Flight Correlation," AGARD-CP-339, 1982, pp. 24-1 24-30. 159) Williams, J., "Aircraft Performance - Prediction Methods and Optimization," AGARD-LS-56, Charlotte St, London. 160) Williams, J., "Prediction Methods for Aircraft Aerodynamic Characteristics," AGARD-LS-67, May 1974.
161) vlood, R.A., "Brief Prepared Remarks for Session II, Performance Correlation," AGARD-CP-339, 1982, pp. 9A-1 - 9A-4.

162) Wood, R.M., Dollyhigh, S.M., and Miller, D.S., "An Initial Look at the Supersonic Aerodynamics of Twin-Fuselage Aircraft Concepts,·· ICAS-82-1.8.3, 13th Congress of the International Council of the Aeronautical Sciences with J..IAA Aircraft Systems and Technology Conference, Seattle WA, August 1982. 163) Wood, R.M. and Miller, D.S., "Experimental Investigation of Leading-Edge Thrust at Supersonic Speeds," NASA TP-2204, September 1983. 164) Wood, R.M., Miller, D.S., Hahne, D.E., Niedling, L., and Klein, J., "Status Review of a Supersonically Biased Fighter Wing-Design Study," AIAA-83-1857, Applied Aerodynamics Conference, Danvers MA, July 1983.

61

165) Woodward, F.A., "Development of the Triplet Singularity for the Analysis of Wings and Bodies in Supersonic Flow," NASA CR-3466, September 1981. 166) Wu, J.C., Hackett, J.E., Lilley, D.E., "A Generalized Wake-Integral Approach for Drag Determination in Three-Dimensional Flows," AIAA-79-0279. Aerospace Sciences Meeting, New Orleans LA, January 1979. 167) Yoshihara, H., "Special Course on Subsonic/Transonic Interference for Aircraft," AGARD-R-712, May 1983. 168) Yoshihara, H. and Spee, 8.M., "Applied Computational Aerodynamics," AGARD-AG-266, August 1982. Aerodynamic Transonic

169) Young, A.D., liThe Calculation of the Total and Skin Friction Drags of Bodies of Revolution at Zero Incidence," A.R.C. R&M 1874, April 1939. 170) Young, A.D., Patterson, J.H., and Jones, J.L., "Aircraft Excrescence Drag, II AGARD-AG-264, July 1981. 171) Youngren, H.H., Bouchard, E.E., Coopersmith, R.M., and Miranda, L.R., "Comparison of Panel Method Formulation and its Influence on the Development of QUADPAN, an Advanced Low Order Method," AIAA-83-1827, Applied Aerodynamics Conference, Danvers MA, July 1983. 172) Yu, N.J., Chen, H.C., Samant, S.S., and Rubbert. P.E., "Inviscid Drag Calculations for Transonic Flows," AIAA-83-1928. Computational Fluid Dynamics Conference, Danvers MA, July 1983, pp. 283-292. 173) Friedman. L.M., "Performance Evaluation of the F-15C," AFFTC-TR-83-32. Dec 1983.

62

......-=:::
~
.~

<:::9
C

CON~IGURATION
• hollttd

CLEAN WING REFERENCE wlnglbody model

.' -

AERODYNAMIC REFERENCE CONF IGURATION • Wlng/bodV model with flow-through nlC.lI .. Ind "'lrenel nozzle.

.....-:::::

<I

PROPULSION REFERENCE CONFIGURATION

• Wing/body model with blown nacl"" Ind ",.,Inc. nozzl ••

CANDIDATE NOZZLE , CONFIGURATION • Winglbody mode' with blown nacelle. Ind c.ndid.t. nOIlI .. ISOLATED NACELLE TEST • Clndld.tl noul ..

Figure 1.

Test Approach et aI, Ref. 147)

(from Tjonneland,

63

c.
+
CRUISE

1,5%
(o,.:av~, Sho ck , ScparatJnn)

I?

LIFT

C
PRESSURE ~

(Cruise)

lncompressjble

FRICTION

5Oi.

Compressihle

CTIP'IHI

figure 2.

Transport Aircraft Drag Buildup (Full Scale, ~

0.8)

64

20

Probable Error
'i / .....

15 10

5 2%

Conceptual Development

Preliminary Design

Design

Flight Test

Figure 3.

Probable Error During Design (from Bowes, Ref. 13)

65

Number of occurrences Deviation from fairing, percent of CD -6

10

Estimated in CD

uncer1ainty

3.0

Figure

4.

Drag Data

Scatter

About Drag Polar (From Arnaiz, Ref.

Fairings 3)

with

~rach Number.

w' "9 wong/body

w,n9

,"CIUd

eng,nlt il'lst." ct.e

'

"9

eody
%V'O

Zero-11ft..

-lift..

wove d ....g

tmpC."no9'"
EMg'''1t
I~Ok~ / / /

inst'""
ctt.e /~/ /

l,.,t..IZ.,.fe,rcnC&

~'fl-d

..p e nd""l

~~?//
a
Cruis e dreg
60slc Ol,.croft..

F'lops

b Toke-off
""0
d~lt..o

drag
M~Z: 0 SI!Z. ... Oe ~ d!Z.lt.o

M=2 Z

S68

"".20
!;w ....

lnuS"oue!

sl..,,,cr

voroob1e s\.~,I< ..

Olrlintt.t'

"t.P'Qt."9·!~ b'J"'bCl'"

'"'9"t.. .. ,.

Figure 5.

Typical Aircraft Drag Breakdowns (from Butler, Ref. 19)

61

tangential total drag

--------------------------

friction

drag

viscous
no rrna l

_[

wake

pressu drag inviscid wave

-L

~:::e:r::ag

lift-dependent

wave drag

volume-dependent

wave drag

Figure (from Butler,

6.

Components Ref. 19;

of Aircraft

Drag 81)

See also

Leyman , Ref.

68

.005
".:
____

• F-4E/ J

..,.
~-~----.-.-.-~., _. __

•.•

....1--

. -of

- ••

•004

. oo'}
.7'.

:-- ~-'-:r i __
C~sA ~

'

:--

.$"")(.
__

~ 1(:;
.()02

:.-

.. If kl", '\/._

_/~ _. C'f'.

.OOl

Reynolds

Number

(Based

on mean length)

Figure

7.

Equivalent

Skin

Friction

69

..... u,
w ee w

0.(XJ20 0.0030 0.0040.......______ ~0.0050.......______~ 0.0060

~200 100 50 40 3D 20

..... ~

'" «
C>.

Z
~

~ :J o w

o
•

t:. BOMBERS
TRANSPORTS REAR-LOADING TRANSPORTS 100

3 4 5 7 10

20 304050

TOTAL WETTED AREA X 10-3 FT2

Figure 8.

Subsonic Parasite Drag-Bomber/Transport (frdm Paterson, et al, Ref. 109)

70

c.;.~

.
~._,

lS"

O .003 0 O.Dl>35

0.004-0 .... O,lX:A·~-·

~ ...\..0 '._

O· 0
+>
Q) Q)

050

4Q)

..
--_
-.
- ------_. --.

So. ttl

V> ttl
Q)

cr

:::::s

So. ttl

+>

Q)

n. +>
Q) r-r-

V> ttl So. ttl

t:

>
cr w
:::::s

ttl

1=UC.H"i

-'.

.1,
(OIJ~I~TlOt-JS:

,.__--/---------

M_:::l_o, R ._ .
-t-·
I

----._
_. ---- _ ..
I

-. --.----

..

•.

-----.. ,

1.5 Total wetted area, thousand square feet

4.0

Figure 9.

Subsonic Parasite Drag-Fighter/Attack

71

2.2

r---------------------------------~-----F-4E/J •

2.0

1.8 • F-IOSD
• F-84G

1.6

1.4 F-84F

1.2L-----~------_+------~------~----~--~ 3.0 2.6 2.8 2.4 2.2 2.0 6

Total Wetted Area Wing Wetted Area

Figure 10.

Relative Wing Size

72

2.4

2.2

1.8
Cfe

Cf. leo
or

1.8 1.6 1.6 1.4 1.4

1.2
-: -,._

.
L'

1.2
.:.; 04

1.0 4

.:. - j-·j.. -L~i--_Jg·>~I~::0:~}.:~L_ ~ ~ L~

- ..;

-! _
I

..

~~~:;~i--1t;:i ~-~-: 8
9

1.0
10

body fineness ratio, ~/d

Figure 11.

Equivalent

Body Shape

73

STATION

ROOT CHORD
MAXIMU~; THICKNESS

(~) d equiv

* ~ /,4 *Ti Aeq.~ )_'f \ )

Figure 12.

Equivalent

Body Concept

74

2.0r-------------------------------------~
1.8

1.6
w

m1EGA

1.4

1.2

F-IOOD

1.0L-------IL.....:-----'------'-----&----""'-~ 2.0 2.2 2.4 2.6 "2.8 3.0

INLETS} ~OZZLES

Total Wetted Area Wing Wetted Area

Figure 13. Aerodynamic Cleanness

75

00044

_0
000 .. 0

Ka~rr.ar.-Sc;J..t.: SpaJGoro~-C:o --Scr.t..:.l:.z-G:-u;P;-ar,Cl.-:': \\ ;;. tt·' -G," .....\,

:-,(~;

-----

---- ---

----

("1

o 003f,

00032

OJ

.002';

£'.2.

b.6

7. 0

7 . ..;

Figure

14.

Comparison of Empirical Flat Formulae for Incompressible (from ~1acWilkinson,

Plate Skin ?riction Turbulent Fl.ow, 36)

et a l , Ref.

76

Hughes

Composite

Curve

1/6 = 41 . 9

12

14

16

10

OO?I, "

~<,"
1-1_

" OO~~ C
~• L.

t--- ""'"""-r---..: J•
--

~ ••

S·. DC ~ f
4'=':

l~~
90 '"'' vv
I

••

Schoenherr

Mean

Line

~~
160

50

60

70

80

12u

RN

14u

r~I200
300

IT

~
40U

x 10

-6

5Cv

1-:---

6CJ

~igure 15.

Summary of Experimental Research on Flat Plate Skin Friction, Incompressible Speeds. (from ~!ac~-Jilkinson, al, Ref. 86) et

77

d/l
150 E:l .170 0.182 V.182 x' .333

711
_".

,,

o .lOO 6 e.

d
7"'1

~
1

-.:
I

......

'

.~~
7"
-/

I I

,
,

,,
/.

.oW ;.-

,
I

:51-

1':
I~

"..

I,
I

,J'

,,I

.004
1,/

,,

,,

'I

f:'"

,, .003

r-,

/
1'1 I 1

L_ ,/
I 1 I

I'

,
I

.~

'

" .=t
.003

"

~.ou.. . DOS

den
Pexpt.

.006

Figure 16.

Fuselage Profile
(from HadJilkinson,

Drag-Bodies of Revolution
e t, 21, Ref. 8G)

78

-- -- :===1--~

_____j

i-

-=~

1.2

CD
PIT; :::

c-1
1.1

1.0 ZO

3n

40 R N

SO

60

70 80

100 on Fuselage Length

i60

300

400

500

x 10-6 Based

Figure 17.

Fuselage Shape Factors, M ~ 0.70 et aL, Ref. 86)

(from HacWilkinson,

79

1.7

"
"

J9

1.6

14

1.5

..
"

u
N

....
1.4
Q.

<,

°e

c:

I.L.

VI

•.

'"0

1.3

Figure 18.

Airfoil Shape Factors et a L, Ref. 86)

(from Hadlilkinson,

80

Figure 19.

Lift and Speed Regions for Calculation (from Schemensky, Ref. 122)

of Drag Due to Lift

81

CD (Wing-body)

~CDIB CLIB

-r

(CL-CLo )

CDr.!?

is is i: C
T

the n i n Lnn.. ll rrofile u the

drag

CDgp

CDo

- KCLo2
CL
2 TINI e

CDo

z.ero lift_ the.)ret12al is

drag

coefficient due to

NCL
K(C;_

drag

lift

ae additional coefficient

pr of i Le drag for drag break, increment, also CLDB also

'"'0

CLB is t,cr'IE is

~he lift the the

interr.:ediate drag variation

drag for

break C L >C

Lerro is
o

~R

Figure 20.

The Subsonic Drag Polar

(from Simon, et aI, Ref. 131)

82

0.6

MACH NUMBER - ..6

0.4

D. USSAERO-B

t:
:::i
Ll-

e
tU

z
L.I.J

0.2

LI-

u::
e

L.I.J

0.0

-0.2

-r---------.,r-------.---------,
0.00 0.02 0.04

0.06

COEFFICIENT OF DRAG

Figure

21.

Drag Predictions (from Miller,

Using Panel Methods Ref. 98)

et aI,

83

o BODY
CO=CO

I
/

I.

114-_~
I ,

-L,/ ..~. I

C2

Theoretical Blowing Dr8g (1.3 Counls)

-0.002 NOTE: AC
I

FAVORABLE

0.,'1) .
I
I I
I

=
=

. L(Crulll

FNPR)

-c

l (Ram FNPR)
FNPR)

UNFAVORABLE -0.004

AC

CO(Crulll

FNPR)-CO(Ram

WING

I
I

I I

TOTAL'; -0.006

?igure

22.

Effects of Engine Power on Lift Aircraft Components (from Chen, et

and Drag of Various

a l , Ref , 31)

84

Figure 23.

Historic

Correlation

of Transonic

Drag Rise

85

CDr.

c~
_ C_L_- 0 .-_CD_

~~J.L
I

MeRor·

1.0 MACH NUMBER

Figure 24.

Transonic

Drag Buildup

(from Schemensky, Ref. 122)

86

1.0

.8

~
.6

..--

3 ..--

.:»:

__ - _- _-

.4

.2

------

EXPERI~ENT ESTIMATE FLOW·ZONE 3

.04

.06

.08

.10

.12

.14

.16

.18

.20

.22

Figure 25.

F-5 CL versus CD; M

0.9

(from Axelson, Ref. 4)

87

CAST 7 / M_· 0 76 fiE : 6 105 r.s.. 007 X/C

IIGI<..I INC)

VGK
Ooloil _

2.0

0.02

0.01

0.6

0.8

0.8

e
- 0.14 ;-.--~ __
....L._~~ __

@..)

:'igure

26.

Total Forces and ~foments Comparison Different Computer Programs (from Longo, et aI, Ref. 84)

Between

88

G.OCII

oco.

.!!!

T. lU' DlTA

.!!!!

ot__

.J
I

~~.S~---L--~0~.'~-~====~O~.7::::~==~_:-~~
... CH HUIISFR

• MNr W1JI.IG OOT Glc*("T-.I[S .. • TWOOI"'(lr..., C ~ 'OiJrrIT'S Ie.

£AC'W GlOII[TR,

CALCUl..lT~

..

OII ~

_[SOlo

"rn.oo

• MlA.5uItUI["'frl ".OM .I~O"'~[

~ T1ST\

0.004 u o o o
ITO O£V: O.DO!.l -'
I'fRFECT COIIRELATIOIIo"-.,.

o o o e

uu CI u

0.002

o o

8
II

070L-070

L-

07S

0.10 Moo MtAS

O~ __ ~ __ -L__~

0.85

ITO O£v : O.lXlO95 __~ __~

0002
':'CO.C MUS

0.004

Figure

27.

Correlation of Calculated and Heasured IHng-Body Drag Divergence Hach Number, Compressibility Drag, and Drag Rise for Three Supercritical 1.Jings. (from Henne, et aI, Ref. 57)

89

SUPERPOSITION

METHOD OF DRAG ANALYSIS

Drag

SKIN FRICTION DRAG DUE TO LIFT AND TRIM DRAG (WAVE & VORTEX)
ZE RO LI FT \'YAVE DRAG

Figure

Supersonic Drag Buildup et aI, Ref. 96)

(from Middleton,

90

c,

fRICTION

COEFFICIENT IN TURaULENT fLOW HEAT TRANSHR

ZERO

---_

"uon

IIAI '110 SCALE IU

'/45 seALE

"'0

CItUISE

~LO·~ __

------IO~J~-------~IO~·~--------JIO' """OlDS IUMIU 'IG.

z·,

Figure 29.

Supersonic Skin Friction et aI, Ref. 81)

(from Leyman,

91

0.28

MACH - 1.80 ]

TEST·THEORY COMPARISON LES-216 WIND TUNNEL MODEL

0.24
-

EXPERIMENTAL THEORETICAL

DATA ESTIMATE

0.20

0.16

CL
0.12

0.08

0.04

o~--~-- __~

~~~~
0.008 0.016 (:.>0.020 0.024
CD

~
0.028

_.
0.032

~
0.036

~ __~
0.040 0.044

0.004

0.006

Figure 30.

Supercruiser

Type Configuration

Test-Theory

Comparison

(from Kulfan, et aI, Ref. 78)

92